Introduction they physically couple the ECM to the actin

Introduction

The principle function of blood vessels is to transport blood throughout the body, delivering oxygen
and nutrients to tissues and removing waste products such as carbon dioxide. This is fundamentally a
mechanical process, and thus mechanotransduction plays an important role in regulating vascular biology.
Mechanotransduction is defined as the conversion of mechanical stimuli into biochemically detectable
information. Sources of mechanical stimuli can include external forces, stress, or the extracellular
environment, which lead to the induction of signaling cascades. Mechanotransduction is thought to occur
through the deformation of proteins by physical force, resulting in altered function and signal generation.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

Vascular endothelial cells (ECs) are equipped with complex machinery for sensing mechanical forces
induced by pulsatile blood flow and transducing said force into biochemical signals that can affect cellular
processes. The cellular responses of endothelial cells to physical cues are crucial for the regulation of
vascular functions, including control of blood pressure, permeability, and leukocyte recruitment during
inflammation (Tarbell et al., 2014). A cellular process that is strongly influenced by the mechanics of the
microenvironment is endothelial cell morphogenesis. The extracellular matrix (ECM) provides structural
support for angiogenesis, and thus cell-ECM adhesions play an important role in transmitting force from the
extracellular environment to within cells. For example, ECM properties of rigidity and adhesiveness, as well
as cyclic stretch resulting from circulatory pressure, have been shown to influence vascular network
formation (to be discussed in detail later). Vascular ECs can also sense and respond to shear stress in
response to flow (Chien, 2007). With all of these forces, transmembrane integrin receptors serve as key
mechanotransducers because they physically couple the ECM to the actin cytoskeleton, assemble signaling
complexes within focal adhesions, and concentrate forces on focal adhesion sites. Within the cell,
cytoskeletal filaments also play a role in mechanotransduction, as they contribute to the structural support of
the cell and are capable of transmitting force over a distance (Ingber, 2002).

When endothelial cells fail to respond to mechanical cues from their environment, pathophysiological
conditions can ensue. One such condition in which mechanotransduction can have drastic implications on
angiogenesis is cancer. Traditionally, tumor angiogenesis is thought to occur through the activation of
growth signaling pathways by angiogenic factors produced by the cancer, resulting in continuous growth and
extension of new vessels to feed the tumor. For this reason, cancer research and treatments are often focused
on ameliorating the effects of defective cell signaling (Bershadsky et al., 2003). However, this approach
does not necessarily address underlying mechanical mechanism associated disease progression. To
complicate things further, morphogenic endothelium are highly heterogenic and contain areas of expansion,
regression, and involution within the same microenvironment (E.L. Clark & E.R. Clark, 1938). Thus a
deeper understanding of how ECs are differentially regulated by chemical and mechanical stimuli is needed
to better treat and combat progressive cancers, as tissues are capable of sensing multiple inputs
simultaneously. Wound healing is another area that could greatly benefit from a better understanding of
mechanotransduction, as alterations in ECM structure and resulting cytoskeleton contribute at least in part to
angiogenesis during wound repair. Along these same lines, development of engineered vasculature requires
an understanding of the effects of mechanotransduction in order to properly fabricate and condition blood
vessels to improve their physiological relevance and therapeutic effectiveness (Ingber, 2003).

1

This paper will critically discuss publications that explore our evolved understanding of the role of
mechanotransduction within the realm of angiogenesis, upon which a novel extension will be proposed to
elucidate new information in future studies. The schematic below presents a general overview of the
mechanisms that are explored in each publication (Figure 1).

Figure 1: Schematic of paper topics.

Statement of Problems

Flexible ECM facilitates EC traction force generation in angiogenesis

Although angiogenic factors initiate neovascularization, the phenotypic fate of ECs is regulated by
the tissue microenvironment. For example, stimulation with fibronectin growth factor (FGF), a soluble
endothelial mitogen, can either induce growth, differentiation, involution, or quiescence of ECs depending
on the mechanical properties of the ECM (Folkman & Ingber, 1987). In the study “Mechanochemical
Switching between Growth and Differentiation during Fibroblast Growth Factor-stimulated Angiogenesis In
Vitro: Role of Extracellular Matrix,” Inger and Folkman aimed to discern the tension-dependent interplay
between ECs and the ECM in response to FGF-stimulated angiogenesis. To accomplish this, time-lapse
cinematography was used to study various in vitro models of spontaneous EC angiogenesis. On rigid culture
dishes, adhesive matrix tendrils accumulated and applied tension to substrate contact points, resulting in
multicellular retraction and elevation of the ECM web into culture media. The suspended ECM web
functioned as a flexible attachment that deformed under cell-generated tensile forces and permitted capillary
tube formation, whereas ECs that remained adhered to the rigid culture dish did not form tubes. These
results are consistent with previous studies of angiogenesis on tissue culture plastic, which all required
release from contact with the rigid substrate for capillary tubes to form (Folkman & Haudenschild, 1980;
Maciag et al., 1982; Feder et al., 1983; Madri & Williams, 1983). Conversely, ECs cultured on flexible
ECM gels spontaneously formed capillaries without release from the underlying substrate, with tube
formation occurring after days compared to weeks on tissue culture plastic (Nicosia et al., 1982; Madri &
Williams, 1983; Montesano et al., 1983; Schor et al., 1983; Kubota et al., 1988). Taken together, the data
supports the notion that malleable ECM promotes cell-generated tension and EC tube formation.

To further test this hypothesis, nonadhesive dishes were coated with different densities of fibronectin
(FN), an ECM glycoprotein that binds integrins, to yield substrates of varying adhesiveness. Capillary ECs
were then cultured on the FN-coated dishes and stimulated with FGF. Serum (containing fibronectin and
vitronectin) was excluded from cell culture media to limit cell attachment to cell-substrate interactions. On
dishes coated with low densities of FN (500 ng/cm2) promoted extensive cell spreading and growth, with cell
spreading increasing as a function of FN coating density. However, no observable tube formation occurred
at these concentrations. Adhesive forces could be overcome on densely coated FN dishes by plating cells at
a higher density, resulting in increased cell-generated traction force and capillary tube formation. These
results were similarly observed for dishes coated with varying densities of type IV collagen or gelatin (FN
and collagen IV results shown Figure 2 below). Interestingly, under differentiating conditions, cell growth
was suppressed, despite stimulation with a potent endothelial mitogen. This suggests that phenotypic
switches may result from tension-dependent changes in cell signaling pathways that affect sensitivity to
morphogenic factors. Thus, mechanical properties of the ECM, as determined by its adhesivity and rigidity,
may be just as important as chemical properties in governing morphogenesis. This study presents early
evidence that a malleable ECM facilitates traction force during angiogenesis, although the mechanism of
differentiation is not elucidated.

Figure 2: Matrix-dependent control of angiogenesis in medium containing FGF. Nonadhesive dishes coated with 10,
50, 100, or 2500 ng/cm2 (from left to right) of FN (top) or type IV collagen (bottom).

Cyclic strain reorganized integrins

In the study “Cyclic Strain Induces Reorganization of Integrin ?5?1 and ?2?1 in Human Umbilical
Vein Endothelial Cells,” Yano et al. aimed to explore the effect of cyclic strain on integrin activity. Since it
has previously been shown that integrins are capable of tyrosine phosphorylating focal adhesion kinase
pp125FAK, and that pp125FAK is tyrosine phosphorylated under cyclic strain, it was hypothesized that that
integrins are involved in the strain response of ECs. In order to study this, ECs were seeded on fibronectin
(a ligand for ?5?1) or collagen I (a ligand for ?2?1), and the redistribution of ? and ? integrins under 10%
cyclical strain (1 Hz) were observed. Labeling with fluorescein IgG secondary antibody and appropriate
anti-integrin revealed that ?1 integrins redistributed into a linear pattern in the direction of the longest axis of
cells on both materials after stimulation with cyclical strain. ?5 and ?2 integrins also reorganized in a linear
pattern under cyclic strain on fibronectin and collagen, respectively, consistent with their receptor function.
When integrins were labeled on plates that were opposite their receptor function, they maintained a diffuse
pattern in comparison. ?3 integrin, a vitronectin receptor, did not reorganize at all when ECs were exposed
to cyclic strain. Confocal microscopy revealed that ECs under cyclic strain became elongated and aligned
perpendicular to the direction of the strain vector.

3

Immunoprecipitation was performed on biotin-labeled ECs to quality the level of integrin complexes,
and revealed that integrin levels did not change after 24 hours of strain exposure. This indicates that cyclic
strain resulted in redistribution of specific integrins without any significant changes in surface levels of
expression. ?5 integrin was expressed to a greater degree on fibrinogen compared to collagen I plates, while
interestingly, there was no visible difference between ?2 integrin levels on different plate materials. There
was also no visible difference between ?1 integrin levels on different plate materials. Immunoprecipitation
was additionally performed to quantify the level of tyrosine phosphorylation of pp125FAK induced by cyclic
strain, and revealed that tyrosin phosphorylation of pp125FAK increased with cyclical strain in ECs grown on
both plate materials after 4 hours of strain exposure. These results likely a result of spatial redistribution
rather than expression of integrins because integrin reorganization is observed in the same time frame as cell
alignment and increasing tyrosine phosphorylation of pp125FAK. Since integrins have no intrinsic enzymatic
activity themselves, biochemical signaling was likely initiated by the recruitment of cytoskeletal elements
and tyrosine kinases. Although this study could not directly identify the individual roles of ? and ? integrin
subunits in response to cyclic strain, the ? subunit likely contributes to the mechanosensing of specific
ligands since its substrates dictate which integrin complexes are recruited to focal adhesions (Ruoslahti &
Pierschbacher, 1987). Thus this study supports the function of integrins as mechanotransducers that regulate
the conversion of extracellular force to intracellular biochemical signals.

GTPase Rho mediates mechanosensing of the ECM

Just as mechanotransduction is necessary for normal cell structure and function, suppressed
sensitivity at any point in the cascade could lead to aberrant behavior of cells. This could be the case in
tumor blood vessels, which exhibit marked structural and functional abnormalities compared to normal
vasculature. Unlike normal blood vessels, tumor vessels are dynamically changing, irregularly shaped, and
contain abnormal pericyte and basement membranes (Baluk et al., 2005). While VEGF-A, a soluble
angiogenic growth factor, has been historically targeted to normalize vasculature, its effects are transient.
Since force is transmitted by the ECM through integrin receptors, an understanding of how tumor vessel
cells respond to mechanical stimuli during angiogenesis may be necessary to effectively treat cancer
(Bershadsky et al., 2003). In the study “Tumor-derived endothelial cells exhibit aberrant Rho-mediated
mechanosensing and abnormal angiogenesis in vitro,” Ghosh et al. attempted to elucidate the mechanisms
responsible for the formation of structural abnormalities in tumor vessels. They hypothesized that cancer
vasculature results from aberrant mechanosensing mechanisms in tumor capillary ECs. To test how ECs
respond to physical cues in their microenvironment, normal and tumor capillary ECs were cultured on FN-
coated silicon substrates and exposed to 10% uniaxial cyclical strain. Compared to 90% of normal ECs,
only 60% of tumor ECs reoriented their longest axis and actin stress fibers perpendicular to the direction of
applied strain, as visualized using fluorescent microscopy and staining with Alexa Fluor-488 Phalloidin,
respectively. Since cell shape is influenced by the ability of ECM to resist cell-generated traction forces, cell
spreading was measured as an indirect indication of ECM rigidity sensing by culturing normal and tumor
ECs on gelatin gels of varying elasticity (Chicurel et al., 1998). On soft substrates, normal and tumor ECs
exhibited similar shape and size. However, as substrate stiffness increased, normal ECs became elongated
and thin, while tumor ECs maintained their polygonal shape and spread to a much greater extent (in terms of
projected cell area). These results clearly indicate that mechanosensitivity is altered in tumor cells compared
to normal cells.

To test whether the aberrant mechanosensitivity of tumor ECs influences angiogenesis, normal and
tumor ECs were cultured on two-dimensional thrombin-crosslinked fibrin gels at different plating densities.
At low plating densities, normal ECs were quiescent, whereas tumor ECs formed extensive capillary

4

networks. In contrast, at high plating densities, normal cells exhibited capillary formation, whereas tumor
cells experienced multicellular retraction, disruption of the tubular network, and clumping. This was
similarly observed with cells cultured within three-dimensional fibrin gels, instead of on top. Normal cell
behavior is consistent with the results of the fist study study, in that higher cell densities increase cell-
generated traction force and resulting capillary tube formation (Ingber et al., 1989). To confirm that tumor
ECs are indeed more contractile that normal ECs, cells were cultured on thin, FN-coated polyacrylamide
gels containing fluorescent nanobeads and measured with traction force microscopy. Compared to normal
ECs, tumor ECs displayed larger regions of stress and applied greater traction force to their underlying
substrates. This supports the hypothesis that tension-dependent effects on FAs and actin stress fibers

contribute to aberrant adhesion and spreading observed in tumor ECs.
Regulation of cytoskeletal tension generation and focal adhesion formation has been shown to occur

by the small GTPase Rho through its downstream effector, Rho-associated kinase (ROCK), which mediates
phosphorylation of myosin light chain (Amano et al., 1997). Rhotekin pull-down assays and Cyclex Rho-
kinase assays were respectively used to measure Rho and ROCK activity in normal and tumor ECs. Under
normal growth conditions, tumor ECs exhibited elevated baseline Rho and ROCK levels. When subjected to
10% uniaxial cyclic stretch on FN-coated silicon substrates, Rho activity increased in normal ECs, but
remained unchanged in tumor ECs. This implies that Rho requires stretch-activation in normal cells but is
inherently activated in tumor cells. In a prior study by Tzima et al., transient inhibition of Rho was
necessary to relieve cytoskeletal tension in order for cells to realign in response to shear stress. For this
reason, ECs were treated with the ROCK inhibitor Y27632 to determine if lowering baseline level of
cytoskeletal tension in tumor ECs could restore normal cell function under uniaxial cyclic strain. This
resulted in perpendicular reorientation of actin stress fibers to a greater extent in tumor ECs compared to
untreated tumor ECs (83% versus 60%). Additionally, treatment with Y27632 prevented clumping of tumor
ECs at high plating density and instead, promoted capillary tube formation. These results are displayed in
Figure 3 below. The data therefore indicates that the Rho-mediated mechanosensing plays a critical role in
EC response to physical cues from the ECM, and that aberrant mechanosensing may contribute to abnormal
tumor angiogenesis.

Figure 3: (A) Immunofluorescence micrographs of normal and tumor cell reorientation under strain for different
treatment conditions. (B) Quantification of reorientation response in cells in absence of strain (white), presence of
strain (black), and Y27632 pretreatment (gray). (C) Phase contrast micgrographs of control and Y27632-inhibited
angiogenesis.

Extension

As previously stated, cancer therapies should attempt to address abnormal tissue structures and
mechanical behavior in addition to biochemical targets in order to better treat the underlying physical basis
of the disease. Mechanical abnormalities are often not exclusively an effect of cancerous pathology, but
rather a cause. This was seen in the study by Ghosh et al., in which aberrant Rho signaling resulted in

5

insensitivity to mechanical stretch stimuli. While normal ECs exhibited a high degree of perpendicular
alignment to uniaxial stretch, this behavior was significantly suppressed in tumor ECs. Although ROCK
inhibition by Y27632 resulted in recovery of mechanosensitivity to stretch inputs, complete recovery was not
attained. This is likely due to the fact that ROCK is a downstream effector of Rho signaling, so downstream
inhibition will not completely ameliorate the effects of aberrant Rho signaling on other effectors. In addition
to ROCK, Rho signaling has also been shown to regulate effector mDia, which has been implicated in stress
fiber formation, potentially by promoting actin polymerization though profilin and/or FA turnover through
Src-tyrosine-kinase activity (Kaunas, 2005). Additionally, in the study by Gosh et al., tumor cells exhibited
higher baseline Erk 1/2 activity. GEF-H is regulated by the FAK/Ras/ERK signaling pathway in response to
force on integrins, so high ERK activity could also contribute to aberrant downstream signaling and resulting
tumor function.

Thus, I propose targeting a signaling molecule upstream Rho
in order to completely block Rho activity. Since Erk 1/2 has already
been identified to have abnormal levels in tumor ECs, I would start by
investigating that molecule in more detail. First, I would inhibit Erk
1/2 in normal and tumor ECs. This could be done by using the soluble
inhibitor PD184352, which act to block mitogen-activated protein
kinase (MEK), a known activator of ERK. Cells could also be
transfected with siRNAs against the Raf, a viral oncogene that
normally induces ERK phosphorylation. Similar to the methodology
in the Ghosh study, and tumor ECs to 10% uniaxial cyclic strain (1

Hz) under control and both ERK-inhibited conditions in order to test
the percent of cells that reorient perpendicular to the axis of applied
strain. I would expect that inhibiting this pathway upstream of Rho
would increase tumor EC reorientation in response to shear by
blocking all Rho activity downstream, including both ROCK and
mDir activity. To confirm that this treatment does indeed inhibit
molecules downstream in the pathway, a Rhotekin pull-down and
Cyclex Rho-kinase assay could be performed to test for Rho and
ROCK activity, respectively, in response to uniaxial cyclic strain in
both normal and tumor ECs. I would also be interested to perform an angiogenesis assay by culturing cells
on top of two-dimensional and within three-dimensional matrices, then assessing for the extent of capillary
tube formation using phase contrast micrography. A question that I have about this method is the longevity
of inhibitor action. PD184352 is likely only active during treatment and for a given time after, and gene
silencing by siRNA is reversible as well. Since the ERK/MAPK pathway acts as an on/off switch, I am
unsure if this pathway will result in transient inhibition to recover cell reorientation in tumor cells, or if it
will result in even less reorientation by relaxing cytoskeletal tension more than transiently. Thus treatment
longevity is a key question that would need to be assessed. This can be done simply by observing how like it
takes ECs to recover from behavior that deviates from its control. If these preliminary experiments proved
successful, an animal study could be performed to test the viability of the treatment in vivo. If this treatment
were to work, it would provide a novel way to target mechanical perturbations of tumor blood vessels.