ABSTRACT
Understanding angiogenesis increases comprehension of pathological conditions such as
tumor growth and peripheral artery disease. Angiogenesis is the growth of new vessels
from pre-existing vessels and commonly associated with two modes: capillary sprouting
and capillary splitting. Past observations in our laboratory provide evidence of vascular
islands in the adult microcirculation. Vascular islands are defined as endothelial cell
segments disconnected from nearby networks. The two objectives of this specific aim
were to (1) determine if vascular islands are involved in angiogenesis during
microvascular network growth, and (2) determine whether vascular islands associated
with microvascular regression are involved in microvascular remodeling. Mesenteric
tissues were harvested from adult male Wistar rats according to the experimental groups:
unstimulated, post stimulation (3, 10 and 70 days), and 70 days post stimulation +
restimulation (3 and 10 days). Stimulation was induced by mast cell degranulation via
intraperitoneal injections of compound 48/80. Tissues were immunolabeled for PECAM
(endothelial cells), NG2 (pericytes), collagen IV (basement membrane), and BrdU
(proliferation). On day 3, the percentage of islands with at least one BrdU-positive cell
increased compared to the unstimulated level and was equal to the percentage of capillary
sprouts with at least one BrdU-positive cell. At day 10, the number of vascular islands
per vascular area dramatically decreased compared to unstimulated and day 3 levels. Data
collected independently for both aims showed that percent vascular area per tissue area
and length density increased by day 10 post stimulation compared to the unstimulated
group. At day 70, vascular area and length density were then decreased, indicating
vascular regression compared to the day 10 levels. During regression at day 70, the
number of islands increased. The disconnected endothelial cells were commonly bridged
to surrounding networks by collagen IV labeling. NG2-positive pericytes were observed
both along the islands and the collagen IV tracks. At 3 days post restimulation, vascular
islands contained BrdU-positive cells. By day 10 post restimulation, when vascular area
and length density were again increased, and the number of vascular islands was
dramatically reduced. The result of this study suggest that (i) segment proliferation
correlates with sprouting and network remodeling, (ii) blood vessel segments could
provide a novel mode of endothelial cell presence in angiogenesis, (iii) blood vessel
segments may be a reserve to connect with existing vasculature, and (iv) vascular islands
originating during microvascular regression are capable of undergoing proliferation and
incorporation into nearby networks during network regrowth.
ACKNOWLEDGEMENTS
First and foremost, I would like to thank my mentor, Dr. Walter L. Murfee, for his
continued instruction, guidance, support, and patience during my years in his laboratory.
I also owe many thanks to the rest of my support in the laboratory; Ming Yang, Peter
Stapor, Rick Sweat, and Sadegh Azimi all contributed to my education and training
during the completion of this project. Specifically, I extend my gratitude to my good
friend Erica Winterer for her continued support and research contributed to this study. I
would also like to express my appreciation for my committee members, Dr. Tabassum
Ahsan and Dr. Aline Betancourt. In addition to my committee, I am sincerely grateful for
the faculty and staff of the Tulane University Department of Biomedical Engineering.
The knowledge, wisdom, and support they have given me will serve as a strong base
from which I will continue my education and career. Finally, I would like to thank the
Louisiana Board of Regents and the Tulane University Newcomb College Institute for
funding contributed to this project.
ii
TABLE OF CONTENTS
ACKNOWLEDGEMENTS……………………....………………………………...……. ii
LIST OF FIGURES…………………………………………………………………..….. v
1. BACKGROUND
1.1 INTRODUCTION AND SPECIFIC AIMS………………………………..... 1
1.2 MICROVASCULAR REMODELING…………………………………….... 3
1.3 ENDOTHELIAL CELL DYNAMICS DURING ANGIOGENESIS……….. 4
1.3.1 SPROUTING ANGIOGENESIS……………………………...…… 5
1.3.2 INTUSSUSCEPTIVE ANGIOGENESIS………………………..… 7
1.4 PRESENCE OF VASCULAR ISLANDS IN UNSTIMULATED
TISSUES....................................................................................... 9
1.5 RELATED ENDOTHELIAL CELL DYNAMICS........................................ 10
1.6 THERAPEUTIC SIGNIFICANCE OF ISLANDS DURING
REGRESSION.......................................................................................... 12
2. MATERIALS AND METHODS
2.1 VASCULAR ISLAND PROLIFERATION AND INTEGRATION…….… 15
2.1.1 MAST CELL DEGRANULATION MODEL OF
ANGIOGENESIS......................................................................... 15
2.1.2 TISSUE HARVESTING................................................................. 16
2.1.3 IMMUNOHISTOCHEMISTRY..................................................... 16
2.1.4 MICROSCOPY AND IMAGE ACQUISITION............................. 17
2.1.5 QUANTIFICATION OF MICROVACSULAR REMODELING.. 17
2.1.6 QUANTIFICATION OF ENDOTHELIAL CELL DYNAMICS... 18
iii
2.1.7 STATISTICAL ANALYSIS18....................................................... 18
2.2 VASCULAR ISLANDS IN REGRESSION AND REGROWTH
2.2.1 MAST CELL DEGRANULATION MODEL OF
ANGIOGENESIS......................................................................... 19
2.2.2 TISSUE HARVESTING................................................................. 20
2.2.3 IMMUNOFLUORESENCE............................................................ 21
2.2.4 MICROSCOPY AND IMAGE ACQUISITION............................. 22
2.2.5 QUANTIFICATION OF MICROVACSULAR REMODELING.. 23
2.2.6 QUANTIFICATION OF ENDOTHELIAL CELL DYNAMICS... 23
2.2.7 STATISTICAL ANALYSIS........................................................... 24
3. RESULTS
3.1 VASCULAR ISLANDS IN PROLIFERATION AND INTEGRATION...... 26
3.2 VASCULAR ISLANDS IN REGRESSION AND REGROWTH................. 32
4. DISCUSSION
4.1 VASCULAR ISLANDS IN PROLIFERATION AND ANGIOGENESIS... 41
4.2 VASCULAR ISLANDS IN REGRESSION AND REGROWTH................. 45
5. CONCLUSION............................................................................................................. 51
REFERENCES................................................................................................................. 53
BIOGRAPHY................................................................................................................... 58
iv
LIST OF FIGURES
Figure 1:
Angiogenesis is a dynamic and multi-cellular, multi-system
process.
Figure 2:
Sprouting, splitting, and vascular island reincorporation represent
different modes of angiogenesis.
Figure 3:
Sprouting angiogenesis in microvascular networks.
Figure 4:
Intussusceptive angiogenesis in microvascular networks.
Figure 5:
Representative images of vascular islands without an open lumen
connection to nearby vessels.
Figure 6:
Experimental timeline for first specific aim.
Figure 7:
Experimental timeline for second specific aim.
Figure 8:
Representative examples of vascular islands in a rat mesenteric
microvascular network.
Figure 9:
Quantification of microvascular network growth and capillary
sprout activity in response to stimulation of mast cell degranulation
during island proliferation study.
Figure 10:
Vascular island proliferation during microvascular network
growth.
Figure 11:
Quantification of proliferation and endothelial cell branch points
along vascular islands during microvascular network growth.
Figure 12:
The presence of vascular islands over the time course of
microvascular network growth.
Figure 13:
Microvascular network area and length density changes post
stimulation and restimulation.
Figure 14:
The presence of vascular islands associated with network
regression.
Figure 15:
Confocal projection of optical sections throughout the thickness of
a vascular island at day 70 post-stimulation.
Figure 16:
Pericytes bridge and extend along vascular islands during
regression.
v
Figure 17:
Collagen IV tracks connect vascular islands to other vessel
segments.
Figure 18:
Connecting basement membrane tracks contain pericytes.
Figure 19:
Proliferation of an island 3 days post-restimulation
vi
1. BACKGROUND
1. 1 INTRODUCTION AND SPECIFIC AIMS
Understanding the cellular dynamics involved in microvascular network growth is critical
for future development of cell-specific therapies targeted at manipulating the
microcirculation during tumor growth, diabetic retinopathy, myocardial infarction and
other pathological conditions. Microvascular network growth in adult tissues is largely
attributed to angiogenesis, defined as the growth of new vessels from pre-existing
vessels. Angiogenesis is commonly associated with two modes: capillary sprouting and
intussusception (Carmeliet et al., 2005). Capillary sprouting involves the proliferation
and migration of endothelial cells from an existing vessel. Intussusception is
characterized by vessel splitting via the extension of endothelial cell filopodia to form an
intra-luminal pillar and subsequent lumen division (Cliff, 1965)(Burri et al., 2004).
Previous work completed in our laboratory identified vascular islands, defined as
endothelial cell segments disconnected from neighboring microvascular networks
(Robichaux et al., 2010). While vascular islands have been seen during vascular
regression (Hughes et al., 2000), their involvement during network growth is unknown. A
potential role for blood vascular islands during angiogenesis is suggested by observations
in the lymphatic system. Lymphatic vascular islands have been associated with cell
proliferation, migration, and recruitment associated with lymphangiogenesis (Benest et
al., 2008; Parsons-Wingerter et al., 2006; Rutkowski et al., 2006; Ji, 2005). Based on the
current evidence regarding lymphatic vascular islands during lymphangiogenesis, we
hypothesize that blood vascular islands can contribute to the growth of a microvascular
network during angiogenesis.
1
The specific aims of this study were twofold:
(1) Determine if vascular islands present in unstimulated tissues are
involved in angiogenesis during microvascular network growth, and
(2) Determine if vascular islands associated with microvascular
regression are involved in microvascular remodeling.
We demonstrate that vascular islands are involved in angiogenesis in the adult rat
mesentery. We show that vascular islands are multi-cellular and associated with cell
proliferation during angiogenesis. Further, the number of vascular islands present in a
network decreases during later stages of network growth, but reappears at day 70 after
network pruning and regression. Our observations implicate vascular island proliferation
and incorporation as an alternative mode of growth during the initial stages of
angiogenesis in adult microvascular networks.
2
1.2 MICROVASCULAR REMODELING
Microvascular remodeling occurs in three different modes throughout an organism’s
lifespan: vasculogenesis, angiogenesis, and arteriogenesis. In the embryo, the de novo
formation of a primary vascular plexus occurs during the process of vasculogenesis via
mesoderm differentiation of angioblasts (Carmeliet, 2000). Primitive vasculature is
generated from the non-luminal endothelial networks during network maturation.
Quiescent microvasculature is then formed through interactions between endothelial cells
and with extracellular matrix, other cells such as smooth muscle cells and pericytes, and
other signals such as shear stress and hyperoxia (Risau, 1997). Interestingly enough,
vasculogenesis has also been implicated in the adult (Drake, 2003). While the source of
cells and their role in adult vascularization are poorly defined, bone marrow derived
circulating endothelial progenitor cells have been shown to actively contribute to
postnatal vasculogenesis (Asahara et al., 1999).
In vivo blood vessel formation also occurs in the adult through angiogenesis.
Angiogenesis is the creation of new blood vessels form a pre-existing network. This
process can occur by two different processes – capillary sprouting and capillary splitting
– that are discussed in detail below (Risau, 1997). Finally, as the vasculature begins to
stabilize, nascent vessels increase in diameter and obtain a viscoelastic and vasomotor
coat of smooth muscle cells (Carmeliet, 2000). This process is known as arteriogenesis
and is important for the consistent perfusion of a tissue. Together, these three processes
are required for healthy tissue function. Understanding the mechanisms and dynamic
interplay of the three processes is imperative to the development of angiogenic and antiangiogenic therapies.
3
1.3 ENDOTHELIAL CELL DYNAMICS DURING ANGIOGENESIS
Endothelial cells are already thought to contribute to angiogenesis through two modes:
sprouting angiogenesis and intussusception. Sprouting angiogenesis is primarily
characterized by degradation of the surrounding extracellular matrix, migration and
proliferation of endothelial cells, and the formation of a lumen to promote functional
maturation of vasculature. Intussusceptive angiogenesis is best defined by transcapillary
pillars of extracellular matrix that splits the lumen of a single capillary into two vessels
with a fully-developed endothelium (Risau, 1997). Both processes are multicellular and
involve significant interaction with the environment down to the capillary level, including
the maintenance or degradation of Collagen IV basement membrane (Figure 1). This
thesis proposes a potential third mode of angiogenesis, vascular island reincorporation,
which is dynamically similar to traditional angiogenesis, but novel for the involvement of
vascular islands (Figure 2).
Figure 1: Angiogenesis is a dynamic, multi-cellular process that includes interaction with
the surrounding basement membrane.
4
1.3.1 SPROUTING ANGIOGENESIS
Sprouting angiogenesis begins notably early in embryo development and persists
throughout adult life. Initially, sprouting angiogenesis is characterized by vasodilation via
nitric oxide in part to increase vascular permeability to vascular endothelial growth factor
(VEGF) (Carmeliet, 2000).
VEGF is known to both stimulate and direct sprouting
angiogenesis (Figure 3a). This process is led by an endothelial tip cell that extends
filopodia to guide the direction of sprouting (Gerhardt et al., 2003). Degradation of the
basement membrane, also termed basal lamina, is mediated by activated endothelial cells
after exposure to VEGF and further allows for extension of endothelial cells into
extracellular space (Figure 3b). Proliferating and migrating endothelial cells then undergo
intercalation and fusion to subsequently acquire a lumen (Figure 3c) (Carmeliet et al.,
2000). Through this process, new capillary segments are able to grow and mature across
their surrounding tissue space until they reach quiescence (Figure 3d).
Figure 2: Angiogenesis in three modes. Endothelial cell dynamics in angiogenesis
include (1) sprouting, (2) intussusception, and, proposed here, (3) vascular island
incorporation. Image credit: Richard Sweat.
5
Figure 3: Sprouting angiogenesis in microvascular networks. Capillaries (a) are excited
by the presence of a growth factor. This causes (b) the degradation of the basal lamina,
also called basement membrane, and activation of endothelial cell proliferation and
migration. (c) Endothelial cells develop into the surrounding tissue to form a capillary
sprout, up the growth factor gradient. (d) The sprout stabilizes into a mature vessel.
Image credit: Erica Winterer.
6
1.3.2 INTUSSUSCEPTIVE ANGIOGENESIS
Intussusceptive angiogenesis, known also as splitting angiogenesis, involves the
separation of a whole capillary into two by the extension of a luminal column.
Endothelial cells within a vessel will proliferate to produce a lumen wide enough to allow
the fusion of the splitting capillary (Figure 4a). This “pinching off” of one vessel into two
is traditionally only seen in thick-walled vasculature, like that seen in the lung or heart
(Risau, 1997). Intussusception is frequently described in three phases: microvascular
growth, arborization, and branch remodeling (Makanya et al., 2009). During
microvascular growth, pillar creation and expansion increases the capillary surface area
begins longitudinal division via a central perforation, often aided by local fibroblasts
(Figure 4b-c). Arborization encompasses the delineation of generations of feeding and
draining vessels to maintain network hierarchy. Merging of septa and horizontal pillar
folds create definitive vessel entities that are separated by collagen fibrils (Figure 4d).
Finally, branch remodeling involves any vascular architectural changes that are done to
meet local blood flow demands (Makanya et al., 2009). Modifications include branch
angle adjustment, bifurcation relation, and vessel dilation or pruning. Overall, this
process is mediated by VEGF promotion of vessel dilation and expansion. TIE-2 is also
implicated with inducing endothelial cell narrowing to begin intussusception (Risau,
1997).
7
Figure 4: Intussuceptive angiogenesis in microvascular networks. (a) Longitudinal view
of a capillary. (b) Endothelial cells begin division by extension of lumenal pillars. (c)
Arborization begins as the two sides of the vessel are fully merged at a central,
perforating bilayer. (d) Basement membrane develops between two daughter vessels for
structural support and branch hierarchy development. Image credit: Erica Winterer.
8
1.4 PRESENCE OF VASCULAR ISLANDS IN UNSTIMULATED TISSUES
Isolated endothelial cell segments have been observed in vascular networks in recent
history. Specifically, vascular islands have been identified in adult rat mesentery as being
disconnected from the surrounding network (Robichaux et al., 2010). It was later shown
that vascular islands have no luminal connection to the network and are not capable of
perfusion (Kelly-Goss et al., 2012). The lack of a vascular connection to surrounding
microvessels was confirmed by intravascular injection of either FITC-albumin or FITCdextran (Figure 5). FITC was observed in all perfused microvessels, but was completely
absent in nearby vascular islands. Further, these Lectin- and PECAM-positive vascular
islands have similar labeling and morphology compared to nearby capillary sprouts,
implicating potential mechanistic similarities between both vessel segment types.
A
B
Figure 5: Representative images of vascular islands without a luminal connection to the
surrounding network. (A) BSI-Isolectin shows previous observations of disconnected
endothelial cell segments, termed vascular islands, which are morphologically similar to
nearby capillary sprouts. (B) Fully perfused FITC-albumin shows vascular islands have
no luminal connection to the surrounding vasculature.
9
1.5 RELATED ENDOTHELIAL CELL DYNAMICS
The potential role of vascular islands as a new mode of angiogenesis is supported by
endothelial cell dynamics observed in lymphangiogenesis, vasculogenesis during
development, and inosculation after implantation of microvessel fragments.
The lymphatic system also provides compelling evidence for the dynamic
involvement of vascular islands in the microvasculature. While the presence of vascular
islands in other tissues is unknown, their involvement in angiogenesis is supported by
similar phenomena in lymphangiogenesis. Lymphatic islands have a long history of being
implicated in network establishment and remodeling. Past work has shown local isolated
lymphatic
endothelial
cell
being
recruited
and
playing
an
active
role
in
lymphangiogenesis (Parsons-Wingerter et al., 2006; Rutkowski et al., 2006).
Disconnected lymphatic islands have also been implicated in lymphangiogenesis in the
adult. In both scenarios, lymphatic endothelial cells form small networks that are able to
proliferate and incorporate into the preexisting lymphatic network. These interconnected
island networks are considered an active proliferating plexus that is required for normal
physiological and pathological function (Ji, 2005; Ji et al., 1997). Due to the shared
mechanisms between lymphangiogenesis and blood vessel angiogenesis, as well as
similar morphology between lymphatic and blood vessel islands, this phenomenon brings
greater attention to the potential role of blood vascular islands in angiogenesis.
Other endothelial cell scenarios support the integration of vascular islands into
remodeling networks during angiogenesis. For example, the connection of disconnected
endothelial cell segments into a perfusion-capable network is observed during embryonic
vasculogenesis (Drake, 2003). During this process, “chord-like formations” of progenitor
10
cells, morphologically similar to vascular islands, aggregate to create a primary plexus
that is capable of integrating multiple new segments during network growth (Drake,
2003). This mechanism reflects the potential vasculogenic-like capabilities of vascular
islands in the adult. Further, circulating endothelial progenitor cells (EPCs) are associated
with incorporation of exogenous endothelial cell sources into adult remodeling
microvasculature (Asahara et al., 1999; Asahara et al., 1997; Tepper et al., 2003;
Grunewald et al., 2006). Finally, inosculation – the successful connection of microvessel
fragments to an active network – has been observed in adult rat subcutaneous
implantations (Hoying et al., 1996; Nunes et al., 2010). When implanted, these
microvessel fragments are capable of becoming fully perfused. These past studies
demonstrate a precedent for endothelial cell incorporation in microvasculature networks.
Vascular islands may share many phenotypic and mechanistic traits with these forms of
endothelial cell integration during angiogenesis.
11
1.6 THERAPEUTIC SIGNIFICANCE OF ISLANDS DURING REGRESSION
An increased understanding of this potential new mode of angiogenesis could have
significant therapeutic applications. Aberrant angiogenesis occurs in one of two ways:
excessive vascularization and insufficient vascularization. In either case, the delicate
balance between pro- and anti-angiogenic factors, as well as their effect on endothelial
cell dynamics, needs to be restored.
Looking to the former case, angiogenesis is hyperactive in a myriad of disorders
ranging from tumor metastasis to inflammatory or ophthalmic disorders (Carmeliet,
2005). Specifically, tumor growth is dependent on the growth of a vascular network to
provide blood and oxygen once it reaches over 1-2 cubic mm to increase oxygen and
nutrient supply (Hillen et al., 2007). Tumor growth takes advantage of bFGF, VEGF,
PLGF, and members of the angiopoiten family to promote proliferation and motility and
suppress anti-angiogenic factors. The TGF-β signaling pathway has also been shown to
suppress the immune system, increase endothelial cell migration, and differentiate
stromal cells (Elliott et al., 2005). The manipulation of endothelial cell survival factors
after radiation also creates a need for greater understanding of vascular islands. Vessel
fragments that survive initial tumor response and regression after chemotherapy have
been suggested to be sufficient to quickly reestablish a functional vasculature during
tumor regrowth (Kozin et al., 2012). By understanding the dynamic interplay between
vascular islands and their environment, therapies may be able to target these endothelial
cells and take advantage of a tumor’s pro-angiogenic environment.
Further, Mancuso et al. demonstrated that molecular and cellular players left
behind after vessel regression provided spatial tracks for rapid regrowth of the
12
vasculature (Mancuso et al., 2006). Collagen IV basement membrane sleeves associated
with pericytes remained after Vascular Endothelial Growth Factor Receptor – 2
(VEGFR-2) inhibition of tumor growth. Upon removal of an antibody inhibitor, the
tumor microvasculature rapidly regrew along the preexisting tracks (Mancuso et al.,
2006); the network had left behind a roadmap of collagen IV to aid rapid regrowth. Based
on these results, we hypothesized that the endothelial cells left behind during regression
are reservoirs along the collagen IV tracks and that these vascular islands are also
capable of being manipulated during regrowth of a microvascular network, potentially in
a tumor-angiogenic environment.
Conversely, angiogenesis is often deficient during preeclampsia and myocardial
infarction. In pregnant woman with preeclampsia, an imbalance in anti- and proangiogenic factors was recently evaluated. Women with preeclampsia had an increase in
anti-angiogenic factors, such s-FLT1 and soluble endoglin, but a significant decrease in
placental growth factor (Reyes et al., 2012). This causes erratic and insufficient
vascularization that can often leads to miscarriages or, when not treated, stroke. Further,
the anti-angiogenic VEGF isoforms that are observed in preeclampsia are involved in
both cancer are other pathologies. One such form is VEGF165b, where the 8th exon of
VEGF has been spliced (Qiu et al., 2009). Overexpression of VEGF165b has been
implicated in glomerular dysfunction and renal failure, as well as local endothelial cell
loss in diabetic retinopathy (Qiu et al., 2009). Specifically in myocardial infarction,
VEGF165b has been implicated in vascular remodeling in both normal and ischemic
tissues to induce insufficient blood flow for sustainable survival (Furlani et al., 2009).
Also in infarcted mature tissues, VEFF-mediated vascular permeability increases edema
13
formation while attempting to promote neovascularization from the infarct border zone.
As a result, vessel leakage causes increased tissue injury before network outgrowth and
remodeling can occur (Weis et al., 2005). Understanding the mechanisms behind
endothelial permeability and motility can help refine the pro-angiogenic therapies that are
used to stabilize the struggling vasculature. Specifically, the use of vascular islands
therapeutically could aid in the revascularization of a network. By providing a source of
both endothelial cells and growth factors, vascular islands may help crease stable
microvascular outgrowth and selectively block or bind to the anti-angiogenic isoform of
VEGF.
14
2. MATERIALS AND METHODS
2.1 VASCULAR ISLAND PROLIFERATION AND INTEGRATION
2.1.1 MAST CELL DEGRANULATION MODEL OF ANGIOGENESIS
All experimental protocols were approved by Tulane University’s Institutional Animal
Care and Use Committee. Similar to a previously established protocol, a single 2 mL
dose of compound 48/80 (Sigma-Aldrich, St. Louis, MO, USA) was administered via
intraperitoneal injections twice a day for 3 days in increasing concentrations (200, 400,
600, 800, and 1000 μg/mL in saline) into adult male Wistar rats (350–450 g body weight,
Charles River Laboratories, Wilmington, MA, USA) (Murfee et al., 2006; Stapor et al.,
2012). Tissues were harvested and prepared for immunolabeling in three experimental
groups: unstimulated (n = 4 rats), 3 days post-stimulation (n = 4 rats), and 10 days poststimulation (n = 4 rats) (Figure 6). The compound 48/80 inflammatory stimulus was used
for this study because it produces a robust angiogenic response across the hierarchy of
mesenteric networks within a relatively short time period(Murfee et al., 2006; Stapor et
al., 2012; Norrby et al., 1990).
Figure 6: Experimental time line for first set of aims. Rats were harvested at three time
points. The first group was at day 0, when unstimulated. Then compound 48/80, a mast
cell degranulator, was administered for three days before rats were harvested at the two
other time points: experiment day 6 (3 days after stimulation) and experiment day 13 (10
days after stimulation).
15
2.1.2 TISSUE HARVESTING
Mesenteric windows, defined as the thin translucent connective tissues between the
mesenteric arterial/venous vessels feeding the small intestine, were labeled for
BrdU + PECAM. Rats were anesthetized with intramuscular injections of ketamine (80
mg/kg bw) and xylazine (8 mg/kg bw). The peritoneal cavities were then injected with
BrdU (1 mg/ml; 30 ml). After a two-hour period, the rats were euthanized via intracardiac injection of Beuthanasia (Schering-Plough Animal Health Corp. Union,
Kenilworth, NJ, USA). 8–10 mesenteric windows were carefully dissected starting from
the ileum and immediately placed in 10 mM Phosphate Buffered Saline (PBS; SigmaAldrich, St. Louis, MO, USA). Next, the windows were whole mounted on positivelycharged slides, fixed in 100% methanol for 30 minutes at −20 °C.
2.1.3 IMMUNOHISTOCHEMISTY
After methanol fixation, tissues were labeled for PECAM (endothelial cell marker) and
BrdU (proliferating nuclei marker). Tissues underwent the following immunolabeling
protocol:
PECAM + BrdU: 1) 6 N HCl at 37°C for 1 hour; 2) 1:100 monoclonal mouse antibromodeoxyuridine (BrdU) (Dako, Denmark) with 5% NGS in ABS overnight at 4°C; 3)
1:100 goat anti-mouse CY2 (GAM-CY2, Jackson Immunochemcals Inc., PA) in ABS at
room temperature for 1 hour; 4) 1:200 PECAM in ABS at room temperature for 1 hour;
5) 1:500 strep-CY3 in ABS at room temperature for 1 hour.
Following each step, tissues were washed for three ten-minute intervals with PBS + 0.1%
saponin. All tissues were sealed and preserved in a 50:50 glycerol:PBS solution.
16
2.1.4 MICROSCOPY AND IMAGE ACQUISITION
Images were digitally captured by the following systems: an inverted microscope
(Olympus IX70, Olympus America, Inc., Melville, NY, USA) coupled with a PixelFly
camera (PCO, Kelheim, Germany) and a Olympus 4x dry, 10x dry, 20x oil or 60x oil
objective; a digital camera (FUJIFILM FinePix S1 Pro) mounted on an inverted
microscope (Olympus IX70) with a Cooke 5x dry, Olympus 20x dry, or Olympus 60x oil
immersion objective; or on an inverted microscope (Leica DM IRE2) with a Leica SP2
AOB confocal microscopy system with a 20x dry or 63x oil objective.
2.1.5 QUANTIFICATION OF MICROVACSULAR REMODELING
Two vascularized tissues were randomly selected from each rat per experimental group.
Thus, a total of 8 tissues were analyzed per experimental group: unstimulated (n = 8
tissues; 2 tissues x 4 rats), 3 days post-stimulation (n = 8 tissues; 2 tissues x 4 rats), and
10 days post-stimulation (n = 8 tissues; 2 tissues x 4 rats). Network montages, generated
by overlaying 4x images, were used to measure vascular area per tissue area. Vascular
area was defined as the cumulative area circumscribed around every microvascular
network in a tissue. Vascular length density was measured for two randomly chosen
representative 4x fields of view per tissue (ImageJ, U.S. National Institutes of Health,
Bethesda, MD, USA). Vascular length density was calculated by the sum of vessel
segment lengths in a field of view divided by the corresponding circumscribed vascular
area in that field. Vascular area and vascular length metrics were based on measurements
of PECAM-positive microvessels along intact networks and did not include vascular
island segments.
17
2.1.6 QUANTIFICATION OF ENDOTHELIAL CELL DYNAMICS
For the unstimulated and day 3 groups, the total number of vascular islands, the number
of vascular islands with at least one proliferating cell, the branch points per vascular
island, the number of blind-ended capillary sprouts, and the number of capillary sprouts
with proliferating cells were measured. Vascular islands were defined as blood
endothelial cell segments that are disconnected from the neighboring blood microvascular
network. The disconnection based on PECAM labeling from the nearby network per
island was confirmed by focusing throughout the full tissue thickness, which is
approximately 20 – 40 μm (Barber et al., 1987). BrdU-positive nuclei along vascular
islands and sprouts were identified based on an elongated nucleus morphology and
location within the PECAM-positive vessel segment.
2.1.7 STATISTICAL ANALYSIS
Measurements are presented as means +/− SEM. Microvascular network growth metrics
were compared across the 10 day time course using a one-way ANOVA or one-way
ANOVA on Ranks followed by a Student-Newman-Keuls pairwise comparison test.
Capillary sprout proliferation was compared between unstimulated and day 3 groups
using a Student’s t-test. The vascular island proliferation and branch point metrics were
compared between unstimulated and day 3 groups using a Mann–Whitney Rank Sum
Test. All statistical comparisons were made using SigmaStat (Systat Software, Inc.,
Chicago, IL, USA). p < 0.05 was regarded as statistically significant.
18
2.2 VASCULAR ISLANDS IN REGRESSION AND REGROWTH
2.2.1 MAST CELL DEGRANULATION MODEL OF ANGIOGENESIS
All experiments were performed under the guidelines and in accordance with Tulane
University’s Animal Care and Use Committee. Adult male Wistar rats (350-450 grams;
Charles River Laboratories, Wilmington, MA) were treated with a protocol previously
established by our laboratory (Methods Section 2.1.1). Individual 2.5 mL doses of
compound 48/80 (Sigma-Aldrich, St. Louis, MO) were administered via intraperitoneal
injections twice a day for 3 days in increasing concentrations (140, 280, 420, 560, and
700 μg/mL in sterile saline). Tissues were harvested from the following experimental
groups 1) unstimulated (n = 4 rats), 2) 10 days post stimulation (n = 4 rats), 3) 70 days
post stimulation (n = 4 rats), and 4) 70 days post stimulation + 10 days restimulation (n =
4 rats). A timeline for the experimental protocol is displayed in Figure 7. For
restimulation groups, animals received a repeat of the 48/80 injection protocol starting on
day 70. Following the repeated 3 day injection protocol, tissues were harvested after 10
days. Compound 48/80 is a mast cell degranulator compound that has been previously
shown to stimulate a robust multi-factorial angiogenic response in the adult rat
mesenteric networks over a relatively short time period (Norrby et al., 1990; Ponce et al.,
2003; Anderson et al., 2008). The rat mesentery and the compound 48/80 stimulation
model were selected for this study to be consistent with previous work that identified the
presence of vascular islands during angiogenesis. In this preceding study, it was
demonstrated that endothelial cells along vascular islands undergo proliferation at day 3
post the initial 48/80 stimulation. In the current study, additional tissues were harvested
from 2 rats at day 3 post restimulation to qualitatively determine whether cells along
19
vascular islands associated with vascular regression were also capable of undergoing
proliferation.
Figure 7: Experimental timeline for second set of aims. For the unstimulated group,
tissues were harvested on the equivalent of day 0. Initial stimulation involved 3 days of
48/80 injections. On day 13 from the start of the experiment, tissues were harvested for
the 10 days post stimulation group. On day 73, tissues were either harvested for the day
70 post stimulation group or re-stimulated with a repeat of the 3 day 48/80 injection
protocol. On day 86, tissues were harvested for the 70 days + 10 day re-stimulation
group.
2.2.2 TISSUE HARVESTING
Rats were anesthetized by an intramuscular injection of ketamine (80 mg/kg bw) and
xylazine (8 mg/kg bw). After shaving and sterilizing the rats abdomen, a single incision
between the sternum and pelvic bone was made down the linea alba. The rats were
euthanized via 0.15 mL intra-cardiac injection of Beuthanasia (Shering-Plough Animal
Health Corp. Union, Kenilworth, NJ). For the 70 days stimulation + 3 days restimulation
group, peritoneal cavities were injected with bromodeoxyuridine (BrdU) (1 mg/ml; 30
ml) and left for two hours. After the incubation time, the rats were euthanized via the
same method as the other groups. The first 8-10 vascularized mesenteric windows per rat
were dissected starting from the ileum and placed directly in 10 mM Phosphate Buffered
Saline (PBS). Mesenteric windows were defined as the thin, translucent connective
20
tissues between the mesenteric arterial/venous vessels feeding the small intestine. Then,
the tissues were whole mounted on positively charged slides. Tissues were then fixed in
100% methanol for 30 minutes at -20 °C. Rat mesenteric tissue windows were again
approximately 20 – 40 μm thick (Barber et al., 1987) enabling the whole-mounting and
visualization of intact networks across their hierarchy. This characteristic of mesentery
windows is necessary for the identification of vascular islands in relation to basement
membrane and pericytes.
2.2.3 IMMUNOFLUORESCENCE
After methanol fixation, tissues were labeled for PECAM (endothelial cell marker), NG2
(pericyte marker), collagen IV (basement membrane marker), or BrdU (proliferating
nuclei marker). Tissues underwent one of the following immunolabeling protocols:
PECAM + NG2 + Collagen IV: 1) 1:100 rabbit polyclonal neuron-glial antigen 2 (NG2)
(Millipore, Billerica, MA), 1:200 mouse monoclonal anti-CD31 (PECAM)
(BD
Pharmingen, San Diego, CA), 1:50 goat polyclonal anti-type IV collagen (COLIV)
(Millipore, Billerica, MA), and 1:20 normal donkey serum (NDS) (Jackson
ImmunoResearch Laboratories, West Grove, PA) in antibody buffer solution (ABS)
(PBS + 0.1% saponin + 2% BSA) at room temperature for 1 hour; 2) 1:500 CY3conjugated streptavidin secondary (Strep-CY3) (Jackson ImmunoResearch Laboratories),
1:50 DyLight-405 AffiniPure donkey anti-goat IgG (H+L) (Jackson ImmunoResearch
Laboratories), and 1:20 NDS in ABS at room temperature for 1 hour; 3) 1:20 NGS in
ABS blocking step for 1 hour at room temperature; 4) 1:100 Alexa Fluor-488 goat antirabbit IgG (GAR-488) (Jackson ImmunoResearch Laboratories) and 1:20 normal goat
serum (NGS) in ABS for 1 hour at room temperature.
21
PECAM + Collagen IV: 1) 1:200 PECAM, 1:50 collagen IV, and 1:20 NDS in ABS for 1
hour at room temperature; 2) 1:500 strep-CY3, 1:50 Alexa Fluor-488 AffiniPure donkey
anti-goat IgG (H+L) (DAG-488) (Jackson ImmunoResearch Laboratories) and 1:20 NDS
in ABS for 1 hour at room temperature.
PECAM + NG2: 1) 1:200 PECAM, 1:100 NG2, and 1:20 NGS in ABS for 1 hour at
room temperature; 2)1:500 strep-CY3, 1:100 GAR-488, and 1:20 NGS in ABS for 1 hour
at room temperature.
PECAM + BrdU: 1) 6 N HCl at 37°C for 1 hour; 2) 1:100 monoclonal mouse antibromodeoxyuridine (BrdU) (Dako, Denmark) with 5% NGS in ABS overnight at 4°C; 3)
1:100 goat anti-mouse CY2 (GAM-CY2, Jackson Immunochemcals Inc., PA) in ABS at
room temperature for 1 hour; 4) 1:200 PECAM in ABS at room temperature for 1 hour;
5) 1:500 strep-CY3 in ABS at room temperature for 1 hour.
Following each step, tissues were washed for three ten-minute intervals with PBS + 0.1%
saponin. All tissues were sealed and preserved in a 50:50 glycerol:PBS solution.
2.2.4 MICROSCOPY AND IMAGE ACQUISITION
Images were digitally captured by the following systems: an inverted microscope
(Olympus IX70, Olympus America, Inc., Melville, NY) coupled with a Photometrics
CoolSNAP EZ camera using 4x (dry; Numerical Aperture = 0.1), 10 x (dry; NA = 0.3),
20x (oil; NA = 0.8 or dry; NA = 0.75), and 60x (oil; NA = 1.4) objectives. The projection
image shown in Figure 15 was obtained with a Ziess LSM 510 META confocal
microscope using a Ziess 40x/NA = 1.3 oil objective.
22
2.2.5 QUANTIFICATION OF MICROVACSULAR REMODELING
The percent vascular area per tissue area, vascular length density, and the number of
vascular islands per vascular length density were quantified per experimental group: 1)
unstimulated, 2) 10 days post stimulation, 3) 70 days post stimulation, 4) 70 days post
stimulation + 10 days restimulation. From the harvested tissues, four vascularized tissues
per animal were randomly selected (n = 4 rats per group) for analysis. Measurements for
each metric were made per tissue. The 4 tissues per animal were then averaged and this
average was used as the value per animal. A montage of each tissue was generated by the
overlaying of 4x images and used for measurement of tissue and vascular area. Tissue
area was defined as the entire area of the mesenteric window. Vascular area was defined
as the cumulative area circumscribed around all microvascular networks in the tissue. For
vascular density measurements, two 4x fields of view were randomly selected per tissue
and the total length of PECAM positive vessels was divided by the corresponding
circumscribed vascular area. If a single 4x field of view was avascular, then it was not
used and another field of view was again randomly selected. Measurements were made
using ImageJ (US National Institutes of Health, Bethesda, MD). The number of vascular
islands per vascular area for each tissue was counted under the microscope using a 10x
objective.
2.2.6 QUANTIFICATION OF ENDOTHELIAL CELL DYNAMICS
Vascular islands were defined as PECAM-positive endothelial cell segments
disconnected from the nearby blood microvascular network. Disconnected segments were
confirmed based on discontinuous PECAM labeling by focusing through the thickness of
23
the tissue with the 20x or 60x objective. Previous characterization of vascular islands
confirmed that the discontinuous PECAM labeling was associated with a lack of luminal
connection with the nearby network (Methods Section 2.1.5). Vascular islands were not
included in the vascular area or vascular length measurements. For unstimulated and 70
days post stimulation groups, the percentage of vascular islands containing a NG2
positive pericyte, the percentage of vascular islands bridged by a NG2 positive pericyte to
a another blood vessel segment, and the percentage of vascular islands bridged by
collagen IV labeling to a another blood vessel segment were quantified in each tissue (n=
4 rats per group). Pericytes were identified based on positive NG2 labeling and a
characteristic elongated cell morphology.
2.2.7 STATISTICAL ANALYSIS
Percent vascular area per tissue area, vascular length density, and the number of vascular
islands per vascular length density metrics were compared across the experimental
groups using a one-way ANOVA on Ranks followed by a Student-Newman-Keuls
pairwise comparison test. An ANOVA on Ranks was used for the comparisons because
the data was not normally distributed. A Student’s t-test was used to compare the
percentage of vascular islands containing a NG2-positive pericyte, the percentage of
vascular islands bridged by a NG2-positive pericyte to a another blood vessel segment,
and the percentage of vascular islands bridged by collagen IV labeling to another blood
vessel segment
between the unstimulated and 70 day post stimulation groups. All
statistical comparisons were made using SigmaStat (Systat Software, Inc., Chicago, IL).
24
Statistical significance was accepted for p < 0.05. Values are presented as means +/SEM.
In order to evaluate inter-animal variances the averages of the tissues analyzed for
each metric (percent vascular area per tissue area, vascular length per vascular area, and
the number of islands per vascular length density) were compared across rats within the
unstimulated and day 10 experimental groups using a one-way ANOVA. No significance
difference across rats within each group was found suggesting that metrics were not
animal dependent. Also, using a Mann-Whitney U-test, the averages for each metric were
compared between the two aims for both unstimulated and day 10 groups. The averages
were also not significantly different.
25
3. RESULTS
3.1 VASCULAR ISLANDS IN PROLIFERATION AND INTEGRATION
In unstimulated microvascular networks, PECAM labeling served to identify endothelial
cells along all hierarchies of microvascular networks down to the capillary level.
PECAM-positive labeling also revealed numerous vascular islands. Vascular islands had
both blood and lymphatic vessel morphologies. Lymphatics were distinguishable from
blood microvessels by their larger diameter, their irregular vessel diameter, and a
decreased PECAM label intensity along endothelial junctions (Murfee et al., 2007).
Blood vascular islands displayed comparable diameters to blood capillary sprouts and
lacked the lymphatic marker, LYVE-1, expression (data not shown). Vascular islands
were identified as disconnected from neighboring microvascular networks based on the
lack of continuous PECAM labeling as determined by focusing up and down across the
full thickness of the tissue (Figure 8).
Similar to previous reports from our laboratory (Stapor et al., 2012), compound
48/80 stimulation caused mesenteric microvascular networks to undergo extensive
angiogenesis. By day 10, networks displayed significantly increased vascular area per
tissue area compared to unstimulated levels (Figure 9A). Vascular length density
increased by day 3 compared to the unstimulated level. By day 10, vascular length
density further increased (Figure 9B). The increases in vascular area and length density at
day 10 were preceded by an increase in capillary sprouting (Figure 9C). At day 3, the
number of capillary sprouts per vascular area significantly increased compared to
unstimulated levels. By day 10 capillary sprouting returned to an unstimulated level.
26
During peak angiogenesis, the percentage of blind-ended capillary sprouts associated
with proliferating cells was increased (Figure 9D).
During the initial stages of microvascular growth, BrdU-positive cells were
commonly observed along vascular islands (Figure 10). On day 3, the percentage of
blood vascular islands associated with at least one BrdU-positive cell increased 5 fold
compared to the percentage in unstimulated tissues (Figure 11A). Blood vascular islands
often contained multiple endothelial cells. As another indicator of growth, the percentage
of vascular islands with branch points increased during angiogenesis (Figure 11B). The
location of proliferating cells within a vascular island or capillary sprout was confirmed
by relative positions of BrdU-positive nuclei and PECAM labeling in 1 μm optical
sections using confocal microscopy. Proliferation was not evaluated along vascular
islands at day 10 after stimulation since at this time point the number of vascular islands
had dramatically decreased and too few were available for analysis (Figure 12).
Figure 8: Representative examples of vascular islands in a rat mesenteric microvacular
network. Positive PECAM labeling identifies lymphatic vessels (L) and blood capillaries
27
(c). PECAM labeling also identifies vascular islands (arrows) defined as disconnected
endothelial cell structures having blood capillary-like morphology.
Figure 9: Quantification of microvascular network growth in response to stimulation of
mast cell degranulation by compound 48/80. (A) Vascular area per tissue area. (B)
Vascular length per vascular area. (C) Total number of sprouts per vascular area. (D)
Percentage of sprouts with at least one proliferating cell. *p < 0.05 compared to
Unstimulated. +p < 0.05 compared to day 3.
28
Figure 10: Vascular island proliferation during microvascular network growth. (A-I)
BrdU-positive cells (arrowheads) along PECAM labeled vascular islands (arrows) were
observed on day 3 post compound 48/80 stimulation. Vascular islands in some cases
contained multiple proliferating cells and endothelial cell branches. (J-L) BrdU colocalized with DAPI-positive nuclei along the disconnected PECAM-positive endothelial
cell segments.
29
Figure 11: Quantification of proliferation and endothelial cell branch points along
vascular islands during microvascular network growth. (A) Percentage of vascular islands
with at least one BrdU-positive cell. (B) Percentage of vascular islands with at least one
branch point. *p < 0.05 compared to Unstimulated.
30
Figure 12: The presence of vascular islands over the time course of microvascular
network growth. (A) On day 3 vascular islands (arrows) were commonly located along
the periphery of a network. Compared to the unstimulated scenario, day 3 vascular
islands commonly contained at least one branch point (arrowhead). (B) By day 10 post
stimulation, the presence of vascular islands was dramatically decreased. (C) Number of
vascular islands per vascular area. *p < 0.05 compared to unstimulated. +p < 0.05
compared to Day 3.
31
3.2 VASCULAR ISLANDS IN REGRESSION AND REGROWTH
In all tissues, PECAM labeling identified vessel segments along the hierarchies of
microvascular networks (Figure 13A-D). Similar to the previous results shown here and
in our laboratory (Stapor et al., 2012; Sweat et al., 2012), compound 48/80 stimulation
caused a significant increase in vascularized area and vascular length density at day 10
compared to unstimulated tissues (Figure 13E-F). At day 70 post stimulation,
microvascular networks displayed significant decreases in area and vessel density
compared to day 10 (Figure 13B-C). The decreases in angiogenic metrics indicate that by
this time point networks have undergone regression. Restimulation of the day 70 tissues
caused microvascular network regrowth (Figure 13C-D) indicated by a significant
increase in both vascular area and density (Figure 13E-F).
Consistent with the initial characterization of vascular islands during angiogenesis
(Results Section 3.1), the number of vascular islands dramatically decreased at day 10
post stimulation versus the unstimulated level (Figure 14C). During the state of
regression at day 70, the total number of vascular islands per tissue and the normalized
number of vascular islands per vascular density increased (Figure 14A-C). This increase
suggests that the presence of vascular islands temporally correlates with regression. The
majority of vascular islands at day 70 were located at the periphery of a regressing
network (Figure 14B). The endothelial cells along vascular islands at day 70 were
confirmed to be disconnected from nearby networks using confocal microscopy. A
projection of optical sections through a vascular island identifies gaps in PECAM
labeling at both ends of the endothelial cell segment (Figure 15). Vascular islands at day
70 were commonly associated with NG2-positive pericyte wrapping (Figure 16). Also, all
32
of the PECAM-positive vascular islands at day 70 displayed collagen IV labeling and
most of the islands were bridged to the nearby network via a collagen IV track (Figure
17). Similar to Mancuso et al.’s description of collagen sleeves left behind after VEGF
inhibition in tumors (Mancuso et al., 2006), the collagen IV bridges associated with
regressing networks in the rat mesentery commonly contained pericytes (Figure 18). The
percentages at day 70 were significantly increased compared to unstimulated levels
(vascular islands with pericyte wrapping at 58% ± 3% vs. 39% ± 4%; vascular islands
with pericyte bridges at 41% ± 4% vs. 10% ± 3%; and vascular islands with collagen IV
bridges at 83% ± 2% vs. 42% ± 3%, respectively; p<0.01 for all comparisons).
After network regrowth, the number of vascular islands again dramatically
decreased. Day 10 post restimulation represents a time point at which a network has just
undergone angiogenesis. Evidence for the vascular islands present at day 70 being
involved in this regrowth is supported by observation of positive BrdU labeling of
endothelial cell nuclei within PECAM positive vascular islands at day 3 post
restimulation (Figure 19).
33
Figure 13: Microvascular network area and length density changes post stimulation and
restimulation. (A-D) Representative montages of PECAM-labeled mesenteric tissues
from the following experimental groups: unstimulated, 10 day post stimulation, 70 day
post stimulation, and 70 day post stimulation + 10 day restimulation. Lymphatic
vasculature is indicated by “L”. (E) Quantification of vascular area per tissue area. (F)
Quantification of vascular length per vascular area. * p<0.05 compared to the
unstimulated group, + p<0.05 compared to the previous time point. Values are means ±
SEM.
34
Figure 14: The presence of vascular
islands is associated with network
regression. (A-B) Representative
images of the periphery of vascular
networks labeled for PECAM at day
10 (A) and at day 70 post stimulation
(B). At day 70 during microvascular
network regression, the number of
vascular
islands
(arrows)
is
increased. Blind ended capillary
segments still connected to a network
are indicated by “c”. Lymphatic
vessels are indicated by “L”. (C)
Quantification of the number of
vascular islands per vascular length
density over the time course of
growth, regression, and regrowth. *
p<0.05 compared to the unstimulated
group, + p<0.05 compared to the
previous time point. Values are
means ± SEM.
35
Figure 15: Confocal projection of optical sections throughout the thickness of a
PECAM-positive vascular island at day 70 post stimulation. The disconnection of the
endothelial cell segment from nearby vasculature is confirmed by the voids in PECAM
labeling at either end of the vascular island (arrows).
36
Figure 16: Pericytes bridge and extend along vascular islands during regression. (A-C)
Examples of PECAM-positive endothelial cells along vascular islands (filled arrows)
with positive NG2 pericyte association. NG2 positive pericytes were observed
associating with islands by either wrapping along (open arrows) or bridging (arrowheads)
endothelial cells. (D-F) Higher magnification images of PECAM and NG2 labeling
within the area defined by the rectangle. Blind ended capillary segments still connected to
a network are indicated by “c”.
37
Figure 17: Collagen IV tracks connect vascular islands to nearby networks and other
vascular islands. (A-C) An example of PECAM-positive endothelial cells along vascular
islands (arrows) with positive collagen IV connections to nearby networks (arrowheads).
(D-F) Higher magnification images of PECAM and collagen IV labeling within the area
defined by the rectangle. Blind ended capillary segments still connected to a network are
indicated by “c”. Lymphatic vessels are indicated by “L”.
38
Figure 18: Connecting basement membrane tracks contain pericytes. (A-C) Examples of
PECAM positive endothelial cells along a vascular island (filled arrows) connected to a
nearby vessel segments via an NG2/collagen IV track. Open arrows indicate pericytes;
arrowheads indicate basement membrane. (D-F) Higher magnification images of
PECAM, NG2, and collagen IV labeling within the area defined by the rectangle. Blind
ended capillary segments still connected to a network are indicated by “c”.
39
Figure 19: Endothelial cells along vascular islands undergo proliferation during
angiogenesis post restimulation. (A-C) Example BrdU-positive nucleus (filled arrow)
along a PECAM positive vascular island. (D-F) Higher magnification images of PECAM
and BrdU labeling of the same vascular island.
40
4. DISCUSSION
4.1 VASCULAR ISLANDS IN PROLIFERATION AND ANGIOGENESIS
The primary findings for Specific Aim 1 are that 1) during the initial stages of
angiogenesis vascular islands undergo proliferation comparable to capillary sprouting and
2) during later stages of angiogenesis vascular islands are no longer present, in line with
the hypothesis that these segments are able to incorporate into growing microvascular
networks. In addition, vascular islands are associated with vascular pericytes, which have
been attributed to play a role in capillary growth and stabilization (Gerhardt et al., 2003).
The mesentery was selected because it offers an overview of a complete
microvascular network (as compared to tissue cross-sections) with a resolution down to
the single cell level. Use of the mesentery to investigate cellular dynamics during
angiogenesis has served to identify endothelial cell phenotypic changes along capillary
sprouts (Anderson et al., 2008), the relative positioning of pericytes along capillary
sprouts (Ponce et al., 2003), and angiogenic pericyte phenotypes (Murfee et al., 2006;
Stapor et al., 2012). The current study takes advantage of a robust model of angiogenesis
stimulated by injections of compound 48/80, a mast cell degranulator (Norrby et al.,
1990). In the original description of this model, Norrby et al. demonstrated the ability of
compound 48/80 to dramatically increase vascularized area, vascular density, and the
number of vessels (Norrby et al., 1990). The quantification of angiogenesis metrics in the
current study also demonstrates these dramatic effects on microvascular network growth.
Our characterization of vascular islands at different time points during growth further
demonstrates the usefulness of this model and suggests a potential new mode of
angiogenesis in an adult tissue involving endothelial cell proliferation and incorporation.
41
The origin of the endothelial cells along vascular islands is currently unknown
emphasizing the need for future lineage studies. Potential sources could be attributed to
the migration from existing vessels, a resident population of endothelial precursor cells,
or vascular regression. Support for regression is provided by the observation of increased
isolated vascular islands during vascular pruning. In rat juvenile retinas hyperoxia was
shown to increase the number of disconnected vascular segments. These segments were
associated with endothelial cell apoptosis (Hughes et al., 2000). In contrast, the vascular
islands in rat mesenteric networks exhibited no evidence for positive TUNEL labeling
(data not shown). Regardless of their origin, our results suggest that vascular islands can
be triggered to enter a proliferative state.
Since proliferation along vascular islands was assessed with a single BrdU pulse,
the cells in the S-phase of proliferation were only captured at the time of tissue
harvesting. At day 3, the percentage of vascular islands with at least one proliferating cell
was 33%. This value most likely underestimates the cellular proliferation associated with
the vascular islands given that cells along the vascular islands are presumably in different
stages of the cell cycle. Capillary sprouting is generally associated with endothelial cell
proliferation. During peak capillary sprouting, the percentage of sprouts with a BrdUpositive cell was comparable to the percentage of vascular islands with a BrdU-positive
cell. Thus, cells along vascular islands undergo proliferation similarly to cells along
capillary sprouts.
Previous dye injection experiments in the rat mesentery confirmed that capillary
sprouts have lumens and endothelial cell segments commonly extend well past them
(Stapor et al., 2013). The structure of blood vascular islands is similar to these extending
42
endothelial cell segments. Based on this and the dye injection data presented in earlier
work done by our laboratory, we hypothesize that vascular islands do not form lumens
prior to connection to a nearby network.
A limitation of the current study is that individual vascular islands were not
tracked over the time course of angiogenesis in the same tissue. This type of time lapse
investigation is required to confirm the fate of vascular islands, to determine whether
vascular islands increase their length, or whether the number of endothelial cells
increases along a vascular island. During this study it was determined that the average
length of a vascular island was not different between the unstimulated and 3 days post
stimulation scenario (data not shown). This lack of difference can be attributed to a high
variability in both vascular island length and cell number. In spite of this heterogeneity,
vascular islands consistently form new branch points during angiogenesis. The increase
in cellular extensions indicates that the vascular islands are dynamic and capable of
changing their structure. Finally, recent work in our laboratory has been able to image
vascular islands during angiogenesis in an adult rat mesentery tissue culture model
(Stapor et al., 2013). By using a unique tissue culture model developed in the lab, Sadegh
Azimi has been able to show up to 75% of vascular islands are able to extend and connect
to the nearby network during vascular remodeling. This supports the observations listed
about that vascular islands potentially increase length and number of branch points during
vascular island incorporation.
The ability of vascular islands to connect to a surrounding vascular network is
further supported by the endothelial cell dynamics observed during embryonic
vasculogenesis. Progenitor cells aggregate and elongate into chord like formations
43
subsequently producing vascular islands composed of endothelial cells (Drake, 2003).
Over time these islands connect with each other and eventually to the surrounding
vasculature, highlighting the ability of disconnected endothelial cell segments to connect
to an existing network (Risau, 1997; Drake, 2003). In the adult, circulating endothelial
progenitor cells (EPCs) have been suggested to incorporate into remodeling vessels
(Asahara et al., 1999; Asahara et al., 1997; Tepper et al., 2003; Grunewald et al., 2006).
Additionally, microvessel fragments isolated from multiple adult tissues in vitro are able
to develop intact networks after implantation (Hoying et al., 1996; Nunes et al., 2010).
These examples combined with our observations that the number of vascular islands
decreases as vascular area and density increase support the hypothesis for incorporation
of vascular islands into growing adult networks.
The involvement of vascular islands in angiogenesis is in line with a similar
growth mechanism seen in lymphangiogenesis, as discussed above. Lymphatic vascular
islands, identified as lymphatic endothelial cell segments disconnected from the
surrounding network have been associated with proliferation, migration, and recruitment
of cells during lymphangiogenesis (Rutkowski et al., 2006; Ji, 2005; Ohtani et al., 2001).
These lymphatic islands have the ability to eventually connect with the existing
lymphatic network in which they are located (Ji, 2005; Stacker et al., 2008). The
overlapping mechanisms of lymphangiogenesis and angiogenesis include common cell
phenotypes and responses to growth factors (Ji, 2005; Stacker et al., 2008). The
involvement of lymphatic islands in lymphangiogenesis implicates a role for blood
vascular islands in angiogenesis.
44
4.2 VASCULAR ISLANDS IN REGRESSION AND REGROWTH
The results for Specific Aim 2 suggest that vascular islands, defined as endothelial
segments disconnected from a nearby network, originate during microvascular
regression. The results of Specific Aim 1 showed that vascular islands are capable of
undergoing proliferation and branching during angiogenesis. In another study in our
laboratory, Stapor et al. established the culturing of rat mesenteric tissues as a tool for
investigating angiogenesis and demonstrated that vascular islands are capable of
physically connecting to nearby networks by comparing images of the same network
before and after serum stimulated network growth (Stapor et al., 2013). This specific aim
adds new information by addressing the origin of the endothelial cells for the vascular
islands.
Logical cell origins include endothelial cells left behind during vascular
regression and the differentiation of new cells from a tissue resident progenitor cell
population. These results implicate vessel regression as a source for the disconnected
endothelial cell segments (Figures 13 and 14) and further suggest that these vascular
islands associated with regression are still capable of undergoing proliferation and
incorporation into a network (Figures 14 and 19).
This project is the first to show that blood vascular islands, capable of
contributing to angiogenesis, exist in an adult tissue. Work by Hughes and Chang-Ling,
described the presence of apoptotic isolated endothelial cell segments during vascular
regression in the retina (Hughes et al., 2000). The results presented here indicate that
vascular islands in rat adult mesenteric microvascular networks are also associated with
regression, and further suggest that at least a population of these disconnected segments
can be recruited to contribute to network growth. The therapeutic relevance of this
45
dynamic might be important for better understanding the source of newly formed vessels
after the removal of an anti-angiogenic treatment. In a recent commentary, Kozin et al.
focused on the issue of tumor neovascularization after non-curative radiotherapy (Kozin
et al., 2012). Questions regarding the source of the new vessels in relapsing tumors and
whether or not endothelial cells that survive are able to re-establish a network remain
unanswered.
During rapid regrowth of tumor microvessels after the removal of a VEGFR-2
inhibitor, endothelial cell sprouts have been shown to track along pericyte-lined, collagen
IV-positive basement membrane sleeves left behind by vessel regression (Mancuso et al.,
2006; Inai et al., 2004). We show similar pericyte-lined collagen IV tracks in regressed
rat mesenteric networks post an initial inflammatory stimulus. Our results suggest that
endothelial cells might also be left behind during regression and that these cells can also
be reused by networks. During the initial stages of angiogenesis post network
restimulation, endothelial cells along vascular islands are proliferative. By 10 days, a
time point after angiogenesis has occurred, the number of vascular islands dramatically
decreases suggesting that the vascular islands have reincorporated with the network.
Since vascular pericytes have been reported to both stabilize and potentially guide
capillary sprouts (Gerhardt et al., 2003; Ponce et al., 2003), we speculate that pericytes
both stabilize and guide vascular islands along the existing collagen IV tracks. However,
the molecular mechanisms of the vascular incorporation process remain to be
investigated and likely include the extension of endothelial cells from both the vascular
island and a nearby capillary sprout.
46
Day 70 was selected to capture network regression post compound 48/80
stimulation based on the previous characterization of 48/80 stimulated angiogenesis in
adult rat mesentery networks (Jakobsson, 1994; Amos et al., 2008). Both studies indicate
that vascular length density decreases post 48/80 stimulation by day 60. Consistent with
these results, we show in the current study that at day 70 both density and area have
significantly decreased compared to day 10. We also show that vascular islands at day 70
displayed increased pericyte coverage, pericyte bridging and collagen IV bridging to
nearby networks compared to vascular islands in unstimulated tissues. We speculate that
the majority of vascular islands in unstimulated versus day 70 tissues are associated with
different stages of the regression process. Work discussed under the first aim noted that
vascular islands in unstimulated networks are not TUNEL-positive. While we cannot rule
out the possibility that a percentage of vascular islands at day 70 are apoptotic, our results
do indicate that they are able to proliferate. Future studies will be needed to determine
whether the presence of existing collagen IV, pericyte-positive tracks influences the rate
of reconnection and to identify the unknown dynamics related to vascular regression and
the fate of vascular islands when a network is not stimulated to undergo growth.
Consistent with Specific Aim 1, a limitation of Specific Aim 2 was that individual
vascular islands were not shown to directly connect to a nearby network. While the
observation of endothelial cell proliferation and the decrease in the number of vascular
islands during network regrowth support vascular island reincorporation, definitive
evidence is not provided. We do know that 1) vascular islands can connect to networks
(Stapor et al., 2013) and 2) that at 10 days post 48/80 stimulation the hierarchy of the
remodeled network, including blind-ended capillary segments, are perfused (Sweat et al.,
47
2012)(Stapor et al., 2013). In light of these findings, the results from Specific Aim 2
specifically demonstrate that vascular islands originating from network regression are
capable of reincorporating into restimulated networks and one can speculate that this
process involves the re-endothelization of the collagen IV tracks.
The rat mesentery and compound 48/80 stimulation model were selected to be
consistent with our previous characterization of vascular islands in the first half of this
study. The rat mesentery is a tissue that is actively remodeling, causing even
unstimulated tissues to contain blind ended capillary vessels, as seen in Figure 13A.
Evidence that at least a population of these blind ended vessels is associated with network
growth is supported by adult rats having larger vascular areas than young rats (HansenSmith et al., 1994). Still, presumably their presence can be attributed to either vessel
growth or regression. The latter case might explain the observation of vascular islands in
the unstimulated tissues. In unstimulated tissues, we previously demonstrated that
vascular islands were disconnected based on the comparison of PECAM labeling with
injected fixable dextran labeling. In unstimulated tissues, 40 kDa dextran injected via the
femoral vein identified vessels throughout adult rat mesenteric microvascular networks,
including the lumens of blind ended capillary segments, but not the disconnected
endothelial cell segments associated with vascular islands. In our current study,
disconnected segments in day 70 post stimulation tissues were confirmed by confocal
optical sections (Figure 15). An advantage of the rat mesentery is that it allows for
observation of cells at different locations across the hierarchy of intact networks and
enables analysis of vascular network area and length density (Figure 13), the presence of
vascular islands (Figure 14) and cellular proliferation (Figure 19).
48
Rat mesenteric tissues can contain: blood vessels only, lymphatic vessels only,
both blood and lymphatic vessels, or no vessels at all. Our criterion for using a
mesenteric tissue in the current study was that it had blood vessels. Thus, only a subset of
the tissues used for our study contained lymphatic vessels (Figure 13C). PECAMpositive lymphatic vessels were distinguishable from blood vessels based on 1) a
decreased labeling intensity, 2) larger relative diameters, and 3) more-uniform diameters
across the hierarchy of a branching network. Also, at higher magnifications the PECAM
labeling is more discontinues along lymphatic versus blood endothelial cell junctions
(Robichaux et al., 2010; Benest et al., 2008; Sweat et al., 2012; Baluk et al., 2007). The
lymphatic vessels in the rat mesentery identified based on the above PECAM labeling
characteristics also label for typical lymphatic markers, including LYVE-1, podoplanin,
and Prox1 (Robichaux et al., 2010; Benest et al., 2008; Sweat et al., 2012). Interestingly,
vascular regression in this study seemed to be specific to blood vascular networks. In
those tissues that contained lymphatics at day 70 post compound 48/80 stimulation,
lymphatic vessels were commonly observed to still cover the entire tissue space (Figure
13C). Our laboratory’s previous characterization of lymphangiogenesis and angiogenesis
in response to 48/80 indicated that both processes occur by day 30, yet by day 30, blood
vascular networks began to regress, while lymphatic vessel density remained the same
(Sweat et al., 2012). The qualitative observations from our current study suggest that
lymphatic vessels continue to occupy the entire tissue space at day 70 (Figure 13C). The
idea that lymphatic vessels persist longer than blood vessels is consistent with previous
work by Baluk et al. that characterized lymphatic growth in chronic airway inflammation
49
(Baluk et al., 2007). Thus, it is not surprising that blood vascular networks significantly
regressed by day 70 while lymphatic vessels did not.
50
5. CONCLUSION
Vascular islands, defined as endothelial cell segments disconnected from nearby
microvascular network, are a dynamic part of angiogenesis in the adult rat. This study
reports that vascular islands are associated with cellular proliferation and increased
branching after mast cell degranulation stimulation. In addition, the number of vascular
islands decreases as microvascular area and length density increase. While future studies
are required in other tissues and during other angiogenic scenarios, these results implicate
a new mode for adult microvascular network growth involving vascular island
proliferation and incorporation.
Further, the results presented here suggest that vascular islands originate during
microvascular network regression and are a reservoir of endothelial cells for network
regrowth. Basement membrane and pericytes along tracks left behind during regression
implicate their role in reconnection of vascular islands to the nearby networks. While
future work is required to understand the molecular mechanisms that regulate this process
and its contribution to angiogenesis, this study adds to our fundamental understanding
regarding endothelial cell dynamics involved in microvascular network growth. The next
step in confirming that vascular islands contribute to network growth requires a
longitudinal study in the same tissue that allows for comparing the location of a vascular
island before and after angiogenesis in vivo or, alternatively, demonstrating that vascular
islands that reincorporate during tissue culture with our lab’s current model are perfusioncompetent. While it has been demonstrated that vascular islands are able to connect to
their nearby network during angiogenesis, how these vascular islands contribute to
51
network growth and determine their fate remains unclear. Additional lineage and time
lapse studies will need to be conducted to further define the origin of vascular islands.
52
References
Amos, P. J., H. Shang, A. M. Bailey, A. Taylor, A. J. Katz, and S. M. Peirce. 2008.
IFATS collection: The role of human adipose-derived stromal cells in inflammatory
microvascular remodeling and evidence of a perivascular phenotype. Stem Cells
(Dayton, Ohio) 26 (10) (Oct): 2682-90.
Anderson, C. R., N. E. Hastings, B. R. Blackman, and R. J. Price. 2008. Capillary sprout
endothelial cells exhibit a CD36 low phenotype: Regulation by shear stress and
vascular endothelial growth factor-induced mechanism for attenuating antiproliferative thrombospondin-1 signaling. The American Journal of Pathology 173
(4) (Oct): 1220-8.
Asahara, T., H. Masuda, T. Takahashi, C. Kalka, C. Pastore, M. Silver, M. Kearne, M.
Magner, and J. M. Isner. 1999. Bone marrow origin of endothelial progenitor cells
responsible for postnatal vasculogenesis in physiological and pathological
neovascularization. Circulation Research 85 (3) (Aug 6): 221-8.
Asahara, T., T. Murohara, A. Sullivan, M. Silver, R. van der Zee, T. Li, B. Witzenbichler,
G. Schatteman, and J. M. Isner. 1997. Isolation of putative progenitor endothelial
cells for angiogenesis. Science (New York, N.Y.) 275 (5302) (Feb 14): 964-7.
Baluk, P., J. Fuxe, H. Hashizume, T. Romano, E. Lashnits, S. Butz, D. Vestweber, et al.
2007. Functionally specialized junctions between endothelial cells of lymphatic
vessels. The Journal of Experimental Medicine 204 (10) (Oct 1): 2349-62.
Barber, B. J., J. Oppenheimer, D. C. Zawieja, and H. A. Zimmermann. 1987. Variations
in rat mesenteric tissue thickness due to microvasculature. The American Journal of
Physiology 253 (4 Pt 1) (Oct): G549-56.
Benest, A. V., S. J. Harper, S. Y. Herttuala, K. Alitalo, and D. O. Bates. 2008. VEGF-C
induced angiogenesis preferentially occurs at a distance from lymphangiogenesis.
Cardiovascular Research 78 (2) (May 1): 315-23.
Burri, P. H., R. Hlushchuk, and V. Djonov. 2004. Intussusceptive angiogenesis: Its
emergence, its characteristics, and its significance. Developmental Dynamics : An
Official Publication of the American Association of Anatomists 231 (3) (Nov): 47488.
Carmeliet, P. 2005. Angiogenesis in life, disease and medicine. Nature 438 (7070) (Dec
15): 932-6.
———. 2000. Mechanisms of angiogenesis and arteriogenesis. Nature Medicine 6 (4)
(Apr): 389-95.
53
Carmeliet, P., and R. K. Jain. 2011. Molecular mechanisms and clinical applications of
angiogenesis. Nature 473 (7347) (May 19): 298-307.
———. 2000. Angiogenesis in cancer and other diseases. Nature 407 (6801) (Sep 14):
249-57.
Carmeliet, P., and M. Tessier-Lavigne. 2005. Common mechanisms of nerve and blood
vessel wiring. Nature 436 (7048) (Jul 14): 193-200.
Cliff, W. J. 1965. Kinetics of wound healing in rabbit ear chambers, a time lapse
cinemicroscopic study. Quarterly Journal of Experimental Physiology and Cognate
Medical Sciences 50 (Jan): 79-89.
Drake, C. J. 2003. Embryonic and adult vasculogenesis. Birth Defects Research.Part C,
Embryo Today : Reviews 69 (1) (Feb): 73-82.
Elliott, R. L., and G. C. Blobe. 2005. Role of transforming growth factor beta in human
cancer. Journal of Clinical Oncology : Official Journal of the American Society of
Clinical Oncology 23 (9) (Mar 20): 2078-93.
Furlani, A. P., R. A. Kalil, I. Castro, A. Canedo-Delgado, M. Barra, P. R. Prates, R. T.
Sant'Anna, and I. A. Nesralla. 2009. Effects of therapeutic angiogenesis with
plasmid VEGF165 on ventricular function in a canine model of chronic myocardial
infarction. Revista Brasileira De Cirurgia Cardiovascular : Orgao Oficial Da
Sociedade Brasileira De Cirurgia Cardiovascular 24 (2) (Apr-Jun): 143-9.
Gerhardt, H., and C. Betsholtz. 2003. Endothelial-pericyte interactions in angiogenesis.
Cell and Tissue Research 314 (1) (Oct): 15-23.
Grunewald, M., I. Avraham, Y. Dor, E. Bachar-Lustig, A. Itin, S. Jung, S. Chimenti, L.
Landsman, R. Abramovitch, and E. Keshet. 2006. VEGF-induced adult
neovascularization: Recruitment, retention, and role of accessory cells. Cell 124 (1)
(Jan 13): 175-89.
Hansen-Smith, F. M., G. R. Joswiak, and J. L. Baustert. 1994. Regional differences in
spontaneously occurring angiogenesis in the adult rat mesentery. Microvascular
Research 47 (3) (May): 369-76.
Hillen, F., and A. W. Griffioen. 2007. Tumour vascularization: Sprouting angiogenesis
and beyond. Cancer Metastasis Reviews 26 (3-4) (Dec): 489-502.
Hoying, J. B., C. A. Boswell, and S. K. Williams. 1996. Angiogenic potential of
microvessel fragments established in three-dimensional collagen gels. In Vitro
Cellular & Developmental Biology.Animal 32 (7) (Jul-Aug): 409-19.
54
Hughes, S., and T. Chang-Ling. 2000. Roles of endothelial cell migration and apoptosis
in vascular remodeling during development of the central nervous system.
Microcirculation (New York, N.Y.: 1994) 7 (5) (Oct): 317-33.
Inai, T., M. Mancuso, H. Hashizume, F. Baffert, A. Haskell, P. Baluk, D. D. Hu-Lowe, et
al. 2004. Inhibition of vascular endothelial growth factor (VEGF) signaling in cancer
causes loss of endothelial fenestrations, regression of tumor vessels, and appearance
of basement membrane ghosts. The American Journal of Pathology 165 (1) (Jul):
35-52.
Jakobsson, A. E. 1994. Angiogenesis induced by mast cell secretion in rat peritoneal
connective tissue is a process of three phases. Microvascular Research 47 (2) (Mar):
252-69.
Ji, R. C. 2005. Characteristics of lymphatic endothelial cells in physiological and
pathological conditions. Histology and Histopathology 20 (1) (Jan): 155-75.
Ji, R. C., and S. Kato. 1997. Enzyme-histochemical study on postnatal development of rat
stomach lymphatic vessels. Microvascular Research 54 (1) (Jul): 1-12.
Kelly-Goss, M. R., E. R. Winterer, P. C. Stapor, M. Yang, R. S. Sweat, W. B. Stallcup,
G. W. Schmid-Schonbein, and W. L. Murfee. (2012). Cell proliferation along
vascular islands during microvascular network growth. BMC Physiology 12 (Jun 21),
7,6793-12-7.
Kozin, S. V., D. G. Duda, L. L. Munn, and R. K. Jain. 2012. Neovascularization after
irradiation: What is the source of newly formed vessels in recurring tumors? Journal
of the National Cancer Institute 104 (12) (Jun 20): 899-905.
Makanya, A. N., R. Hlushchuk, and V. G. Djonov. 2009. Intussusceptive angiogenesis
and its role in vascular morphogenesis, patterning, and remodeling. Angiogenesis 12
(2): 113-23.
Mancuso, M. R., R. Davis, S. M. Norberg, S. O'Brien, B. Sennino, T. Nakahara, V. J.
Yao, et al. 2006. Rapid vascular regrowth in tumors after reversal of VEGF
inhibition. The Journal of Clinical Investigation 116 (10) (Oct): 2610-21.
Murfee, W. L., J. W. Rappleye, M. Ceballos, and G. W. Schmid-Schonbein. 2007.
Discontinuous expression of endothelial cell adhesion molecules along initial
lymphatic vessels in mesentery: The primary valve structure. Lymphatic Research
and Biology 5 (2): 81-9.
Murfee, W. L., M. R. Rehorn, S. M. Peirce, and T. C. Skalak. 2006. Perivascular cells
along venules upregulate NG2 expression during microvascular remodeling.
Microcirculation (New York, N.Y.: 1994) 13 (3) (Apr-May): 261-73.
55
Norrby, K., A. Jakobsson, and J. Sorbo. 1990. Quantitative angiogenesis in spreads of
intact rat mesenteric windows. Microvascular Research 39 (3) (May): 341-8.
Nunes, S. S., L. Krishnan, C. S. Gerard, J. R. Dale, M. A. Maddie, R. L. Benton, and J. B.
Hoying. 2010. Angiogenic potential of microvessel fragments is independent of the
tissue of origin and can be influenced by the cellular composition of the implants.
Microcirculation (New York, N.Y.: 1994) 17 (7) (Oct): 557-67.
Ohtani, Y., and O. Ohtani. 2001. Postnatal development of lymphatic vessels and their
smooth muscle cells in the rat diaphragm: A confocal microscopic study. Archives of
Histology and Cytology 64 (5) (Dec): 513-22.
Parsons-Wingerter, P., T. L. McKay, D. Leontiev, M. B. Vickerman, T. K. Condrich, and
P. E. Dicorleto. 2006. Lymphangiogenesis by blind-ended vessel sprouting is
concurrent with hemangiogenesis by vascular splitting. The Anatomical Record.Part
A, Discoveries in Molecular, Cellular, and Evolutionary Biology 288 (3) (Mar): 23347.
Ponce, A. M., and R. J. Price. 2003. Angiogenic stimulus determines the positioning of
pericytes within capillary sprouts in vivo. Microvascular Research 65 (1) (Jan): 458.
Qiu, Y., C. Hoareau-Aveilla, S. Oltean, S. J. Harper, and D. O. Bates. 2009. The antiangiogenic isoforms of VEGF in health and disease. Biochemical Society
Transactions 37 (Pt 6) (Dec): 1207-13.
Reyes, L. M., R. G. Garcia, S. L. Ruiz, D. Broadhurst, G. Aroca, S. T. Davidge, and P.
Lopez-Jaramillo. 2012. Angiogenic imbalance and plasma lipid alterations in women
with preeclampsia from a developing country. Growth Factors (Chur, Switzerland)
30 (3) (Jun): 158-66.
Risau, W. 1997. Mechanisms of angiogenesis. Nature 386 (6626) (Apr 17): 671-4.
Robichaux, J. L., E. Tanno, J. W. Rappleye, M. Ceballos, W. B. Stallcup, G. W. SchmidSchonbein, and W. L. Murfee. 2010. Lymphatic/Blood endothelial cell connections
at the capillary level in adult rat mesentery. Anatomical Record (Hoboken, N.J.:
2007) 293 (10) (Oct): 1629-38.
Rutkowski, J. M., K. C. Boardman, and M. A. Swartz. 2006. Characterization of
lymphangiogenesis in a model of adult skin regeneration. American Journal of
Physiology.Heart and Circulatory Physiology 291 (3) (Sep): H1402-10.
Stacker, S. A., and M. G. Achen. 2008. From anti-angiogenesis to antilymphangiogenesis: Emerging trends in cancer therapy. Lymphatic Research and
Biology 6 (3-4): 165-72.
56
Stapor, P. C., M. S. Azimi, T. Ahsan, and W. L. Murfee. 2013. An angiogenesis model
for investigating multicellular interactions across intact microvascular networks.
American Journal of Physiology.Heart and Circulatory Physiology 304 (2) (Jan 15):
H235-45.
Stapor, P. C., and W. L. Murfee. 2012. Identification of class III beta-tubulin as a marker
of angiogenic perivascular cells. Microvascular Research 83 (2) (Mar): 257-62.
Sweat, R. S., P. C. Stapor, and W. L. Murfee. 2012. Relationships between
lymphangiogenesis and angiogenesis during inflammation in rat mesentery
microvascular networks. Lymphatic Research and Biology 10 (4) (Dec): 198-207.
Tepper, O. M., B. A. Sealove, T. Murayama, and T. Asahara. 2003. Newly emerging
concepts in blood vessel growth: Recent discovery of endothelial progenitor cells
and their function in tissue regeneration. Journal of Investigative Medicine : The
Official Publication of the American Federation for Clinical Research 51 (6) (Nov):
353-9.
Weis, S. M., and D. A. Cheresh. 2005. Pathophysiological consequences of VEGFinduced vascular permeability. Nature 437 (7058) (Sep 22): 497-504.
57
BIOGRAPHY
Molly Kelly-Goss is a 5th Year student, completing her combined Bachelors and Masters
of Science degree from Tulane University in the coming weeks. She anticipates attending
the University of Virginia to complete her PhD in Biomedical Engineering and continue
research on microvascular dynamics. Originally from Minneapolis, Minnesota, Molly
enjoys interesting food, and live music and theater. Molly also enjoys gallivanting around
the world limited only by her backpack, bank account, and relatively short stature.
58