Modulation of Endothelial Cell Responses by Different TGFβ. Isoforms and Oxygen Levels and Their Potential Impact on. Angiogenesis

Transcription

1 Modulation of Endothelial Cell Responses by Different TGFβ Isoforms and Oxygen Levels and Their Potential Impact on Angiogenesis By Meghan Danielle Doerr A Thesis presented to The University of Guelph In partial fulfillment of requirements for the degree of Master of Science in Biomedical Sciences Guelph, Ontario, Canada Meghan Doerr, December, 2014

2 ABSTRACT Modulation of Endothelial Cell Responses by Different TGFβ Isoforms and Oxygen Levels and Their Potential Impact on Angiogenesis Meghan Danielle Doerr University of Guelph, 2014 Advisor: Dr. A.M Viloria-Petit Angiogenesis is vital for both normal tissue homeostasis and tumor progression via tumor angiogenesis. This process is regulated by many factors including VEGF, TGFβ and oxygen levels. All of these factors are present in the tumor microenviroment, suggesting they might interact to govern tumor angiogenesis. Endothelial to Mesenchymal Transition (EnMT) may also play a role in angiogenesis because some of the same factors also regulate this process. In this study, it was first confirmed that TGFβ treatment with either isoform 1 or 2 at a 5 ng/ml dose preferentially activates SMAD2 and not SMAD1, the effector proteins of the TGFβ pathway, under normal or low oxygen conditions. This suggests that high doses of either isoform favor the TGFβRII/ALK5 pathway. Next it was determined how activation of this pathway affected protein levels of endothelial cell and EnMT markers. It was found that TGFβ treatment increased the levels of a key EnMTassociated transcription factor, Snail, and decreased the levels of the adhesion protein, VE-cadherin. This response was consistent with an EnMT response. Lastly, it was determined how changes in these factors affected the angiogenic response of the bovine aortic endothelial cell line under study. It was found that treatment with either isoform of TGFβ in both oxygen conditions inhibited endothelial cell migration and cord formation, suggesting that TGFβ induced EnMT-like changes might not necessarily reflect the migratory and invasive components of an angiogenic response. More studies are necessary to establish the role of Snail in angiogenic response to TGFβ.

3 ACKNOWLEDGEMENTS First I would like to thank my advisor, Alicia, for taking me on as an undergrad student and also for this opportunity to pursue an M.Sc in a topic I find so fascinating. Your mentorship and support have helped me develop into the researcher I am today. Also, thank you Brenda, Jim and Nina for agreeing to be on my committee. You are all excellent in your area of research and I was truly blessed to have your input throughout this process. Specifically Brenda, thank you for all of your help during the optimization of the cord assay. Without your invaluable help I don t think this vital piece of this project would have been completed. I would also like to thank members of the Viloria-Petit, Coomber and Mutsaers labs, with special mention to Stacey Butler, Nelson Ho, Sean Masson, Raishard Haynes, Adam Andrade, Jon Asling, and Saedeh Dadgar, you guys were always there not only for a good laugh but also to help me get all of this work done. Thank you also to all of the technicians I ve had the honor to work with. Monica Antenos, Allison McKay, Helen Coates and Jodi Morrison, you technical abilities amaze and inspire me. Thank you for your countless hours of troubleshooting. Lindsay Bergeron, thank you for all of your help with the hypoxia work. This project would not have been possible without your vital training and equipment. Dr. Susan Lolle, I would not even be here today if you didn t see my potential as a researcher. You were the one who initially exposed me to research and it was in your lab that I fell in love with the wonderment of science. iii

4 To my family, your endless piles of support throughout my academic career have made this all possible. Thank you for listening to my random blurbs about my project; I can only imagine how weird it was for me to talk about a snail in that context. Lastly, to my love, Sinan Baltacioglu- you have been there through the best and worst times of both of my degrees. Always with much-needed coffee and most importantly, love. Seni seviyorum canim, I can t wait to start the next chapter of my life with you! I have truly stood on the shoulders of giants iv

5 DECLARATION OF WORK PERFORMED I declare that with the exception of the items below, all work reported in this thesis was performed by me. Wimasis (IBIDI) performed analysis on cord formation assay images that I had captured. v

6 TABLE OF CONTENTS ACKNOWLEDGEMENTS... iii DECLARATION OF WORK PERFORMED... v TABLE OF CONTENTS... vi LIST OF FIGURES... viii LIST OF ABBREVIATIONS... ix INTRODUCTION...1 LITERATURE REVIEW...4 Overview Basics of Angiogenesis Physiological and Pathological Angiogenesis Angiogenesis Mediators Stages of Sprouting Angiogenesis Tip and Stalk Cells Angiogenesis in Cancer Transforming Growth Factor-Beta TGFβ Ligands Canonical TGFβ Signaling TGFβ in Angiogenesis TGFβ in Disease Hypoxia Hypoxia s Interaction with TGFβ Hypoxia in Angiogenesis Hypoxia in Cancer Endothelial to Mesenchymal Transition (EnMT) EnMT Markers and EnMT-Promoting Signaling EnMT-Associated Transcription Factors EnMT in Disease EnMT in Angiogenesis...25 RATIONALE...28 MATERIAL AND METHODS...30 RESULTS...38 Characterization of TGFβ Receptors and Endothelial Cell Markers in Bovine Aortic Endothelial Cells TGFβ1 and 2 Signaling Kinetics in Normoxia Versus Hypoxia Time-Dependent Effects of TGFβ1 and 2 on EnMT Markers in Normoxia Versus Hypoxia Time-Dependent Effects of TGFβ1 and 2 on EC Markers in Normoxia Versus Hypoxia Smad2 Subcellular Localization Study EnMT-associated Transcription Factor Localization Study Changes in Adhesion Protein Levels in Response to TGFβ Functional Angiogenic Assays Study vi

7 Figures...50 DISCUSSION TGFβ Treatment Induces EnMT-Like Changes and it Signals though TGFβRII/ALK EnMT in Angiogenesis SUMMARY AND CONCLUSIONS FUTURE DIRECTIONS...82 LITERATURE CITED...84 APPENDICES I: Supplementary Figures Hypoxia Control Localization of β Catenin in Response to TGFβ1 and 2 in Normal and Low Oxygen Conditions...98 II: Preparation of Materials III: Chemical List and Suppliers vii

8 Figure: LIST OF FIGURES Page: 1 Characterization of TGFβ Receptors and Endothelial Cell Markers in BAECs in Response to TGFβ1 2 Characterization of TGFβ Receptors and Endothelial Cell Markers in Bovine Aortic Endothelial Cells in Response to TGFβ TGFβ1 and 2 Signaling Kinetics in Normoxia 52 4 TGFβ1 and 2 Signaling Kinetics in Hypoxia 53 5 Time-Dependent Effects of TGFβ1 and 2 on EnMT Markers in Normoxia 54 6 Time-Dependent Effects of TGFβ1 and 2 on EnMT Markers in Hypoxia 55 7 Time-Dependent Effects of TGFβ1 and 2 on EC Markers in Normoxia 56 8 Time-Dependent Effects of TGFβ1 and 2 on EC Markers in Hypoxia 57 9 Real Time PCR Analysis of VEGFR2 and Snail Expression Levels in Response to TGFβ1 and TGFβ2 Treatment Under Normal and Low Oxygen Smad2 Subcellular Localization in Response to TGFβ Smad2 Subcellular Localization in Response to TGFβ Smad2 Subcellular Localization-Immunofluorescence EnMT-Associated Transcription Factor Localization in Response to TGFβ EnMT-Associated Transcription Factor Localization in Response to TGFβ Snail Subcellular Localization in Response to TGFβ1 and Changes in Adhesion Protein Levels in Response to TGFβ Changes in Adhesion Protein Levels in Response to TGFβ TGFβ Treatment Effect on EC Migration/Proliferation in Normoxia TGFβ Treatment and Oxygen Level Effect on EC Cord Formation 65 viii

9 LIST OF ABBREVIATIONS AJs ALK APS BMP BSA DAPI DMEM DTT ECM EnMT FBS HIF HRE HRP IF PBS PCR PMSF PVDF (rh) TGFβ SDS Smad TβR TBS TEMED TJs VEGF Adherens Junctions Activin-like Kinase Ammonium Persulfate Bone-Morphogenic Protein Bovine Serum Albumin 4, 6 -Diamidino-2-phenylindole Dulbelcco s Modified Eagles Medium Dithiothreitol Extracellular Matrix Endothelial-to-Mesenchymal Transition Fetal Bovine Serum Hypoxia-inducible Factor Hypoxia Response Element Horseradish Peroxidase Immunofluorescence Phosphate-buffered Saline Polymerase Chain Reaction Phenylmethanesulfonyl Fluoride Polyvinylidine Fluoride (Recombinant human) Transforming Growth Factor beta Sodium Dodecyl Sulfate Small Body and Mothers Against Decapentaplegic Homolog Transforming Growth Factor Beta Receptor Tris-buffered Saline Tetramethylethylenediamine Tight Junctions Vascular Endothelial Growth Factor ix

10 Introduction Angiogenesis is a crucial process that regulates the level of oxygen and nutrients delivered to the tissue. It is a highly complex phenomenon, orchestrated by the interaction of many growth factors and environmental factors including, VEGF, TGFβ and hypoxia [1]. The VEGF signaling pathway is one of the major regulators of angiogenesis and thus this pathway is typically inappropriately activated during cancer progression [2]. Within the tumor microenvironment, all of the aforementioned regulators of angiogenesis are aberrantly activated, enabling cancer cells to recruit blood vessels through the process of tumor angiogenesis. Judah Folkman was the first to discover that in order for solid tumors to grow past 1 mm 3 they need to acquire a stable blood supply [3]. Unfortunately once a tumor is connected to a blood vessel, not only does this enable tumor growth by providing the cancer cells with more nutrients and oxygen, it also provides a means for them to escape and colonize distant sites through metastasis [1]. Further, anti-angiogenic therapies have been proven inefficient in preventing cancer progression, highlighting the critical need to further understand the mechanism of angiogenesis in hopes to identify more suitable targets [4]. As mentioned previously, TGFβ is one of the many regulators of angiogenesis. This pathway is highly context dependent in endothelial cells, whereby its effect can be modulated by many factors including the abundance of its different isoforms (TGFβ1, 2 and 3) and also by which TGFβ receptor type 1 gets incorporated into the signaling pathway [5]. Also, this complexity could arise from TGFβ being part of a super family containing many ligands. All three TGFβ isoforms, through the use of murine knock-out 1

11 studies in vivo, have been demonstrated to play non-overlapping roles during angiogenesis [6]. However, the specific roles of each individual TGFβ isoform in adult angiogenesis have not been fully elucidated. Due to this pathway being context dependent, TGFβ exerts a biphasic effect on endothelial cells that line the blood vessels, where low doses of TGFβ promote endothelial cell migration and proliferation but at high doses it inhibits these processes. This has been attributed to the presence of two TGFβ type 1 receptors, the widely expressed ALK5 and the endothelial cell specific receptor ALK1 [7]. HIF, the main effector protein of the hypoxia pathway, is one pathway responsible for driving the expression of VEGF-A to promote angiogenesis. Previous research has identified an active hypoxia response element (HRE) within its promoter and also implicated that TGFβ can work with hypoxia to enhance this transcriptional regulation [8]; providing evidence that these two factors present in the tumor microenvironment can work together to drive angiogenesis. In addition, a few studies have suggested that the hypoxia pathway itself regulates the expression and/or activation of TGFβ1, 2 and 3 [9-11], implicating there could be a positive feedback mechanism in how these factors regulate angiogenesis, but this theory has not been fully explored. Another cellular process that has been relatively unexplored in the context of angiogenesis is EnMT. Two theories have been proposed as to where this cellular phenomenon could be playing a role in angiogenesis. The first states that EnMT could be occurring within the tip cells of the nascent vessel while the other suggests that EnMT might be how mural cells arise from the endothelium [12]. The goal of this thesis was to 2

12 begin to address these unsolved questions in hopes to provide more insight into how TGFβ and hypoxia can induce EnMT and how this fits into the process of angiogenesis. 3

13 Overview Literature Review The tumor microenvironment is comprised of many different factors including signaling proteins and variable ph, glucose and oxygen levels, among others. Tumor cells and surrounding stroma cells secrete pro-angiogenic growth factors, such as Vascular Endothelial Growth Factor (VEGF), angiopoietins [13] and Transforming Growth Factor-Beta (TGFβ) [14], all of which influence the endothelial cell phenotype. Within the tumor, regions of low oxygen form due to the rapid increase of proliferation coupled with inadequate vascularization [15]. An in vitro study conducted in embryonic microvascular cells (HMEC-1) demonstrated that TGFβ and hypoxia work synergistically to induce VEGF expression [8]. This is just an example of how these two different components of the tumor microenvironment are able to work together to promote angiogenesis. In the next sections, the concept of angiogenesis and how TGFβ signaling and hypoxia modulate this process will be reviewed. Also, evidences of the role of these two microenvironmental components in modulating a transition from an endothelial to mesenchymal (EnMT) phenotype are presented, and ultimately the possibility that this EnMT contributes to angiogenesis was explored. 1. Basics of Angiogenesis 1.1 Physiologic and Pathologic Angiogenesis Angiogenesis is the formation of new blood vessels from a pre-existing parental vessel. It plays an important part in many physiological and pathological conditions. The 4

14 endothelium is normally quiescent in the adult, but it can be activated upon specific stimuli. Angiogenesis plays an important role in inflammation, tissue repair and female menstruation and pregnancy. Most importantly, angiogenesis is responsible for managing the oxygen and nutrient demand of the tissue [16]. Angiogenesis also plays a role in the etiology of many diseases; this can be due to insufficient or excessive angiogenesis. Insufficient angiogenesis can cause endothelial cell dysfunction, which in turn leads to vessel malformation or regression. It can also prevent revascularization, healing and regeneration [17]. When there are not enough blood vessels present, it can lead to diseases such as ischemic heart disease [18]. Some diseases caused by excessive angiogenesis include psoriasis [19] and tumor growth and metastasis [16]. Thus, angiogenesis is crucial for normal tissue function but when it functions improperly, it constitutes the basis for many diseases. 1.2 Angiogenesis Mediators: Emphasis on Vascular Endothelial Growth Factor (VEGF) Angiogenesis is regulated by the homeostatic balance between pro- and antiangiogenic factors [20]. The most notable pro-angiogenic factors are members of the vascular endothelial growth factor (VEGF) signaling pathway. The VEGF family consists of many different homologous factors, with the most studied being vascular endothelial growth factor-a (VEGF-A) [21]. All of the VEGF polypeptides exert their physiological effect by binding to several type V receptor tyrosine kinases (RTKs), such as VEGFR1 (Flt-1) and VEGFR2 (Flk-1) [22,23]. The complex interplay of VEGF family proteins with VEGF receptors and co-receptors is strictly required for shaping and maintaining the 5

15 functionality of blood vessels, and the coordinated signal output initiated by distinct VEGFs are central for proper vessel organization [24]. VEGF-A and VEGFR2 have the most documented role in angiogenesis. VEGF-A controls many endothelial cell behaviors such as endothelial cell differentiation, migration, proliferation and survival. It also controls endothelial cell-cell junction formation. Due to VEGF-A regulating these cellular behaviors it comes as no surprise that it is a critical factor for the initiation and direction of sprouting angiogenesis [2]. VEGF-A mostly binds to and signals through the VEGFR2, but has been known to signal through the other receptor VEGFR1 [25]. Expression of VEGF-A is regulated by several factors including hypoxia and growth factors, such as EGF-family members [26], Insulin-like growth factor-1 (IGF-1) [27] and TGFβ [8]. I will expand on VEGF modulation by TGFβ and hypoxia in later sections. 1.3 Stages of Sprouting Angiogenesis: The complex process of sprouting angiogenesis can be broken down into broadly two phases, the activation phase and the resolution/maturation phase [28]. Upon proangiogenic stimulation with VEGF-A, the normally quiescent endothelium is activated leading to enhanced vessel permeability and vasodilation. Permeability is mediated by the re-localization of platelet endothelial cell adhesion molecule (PECAM)-1 and vascular endothelial (VE) cadherin from the cell membrane to the cytoplasm [29]. During the next step of the activation phase of angiogenesis, endothelial cells prepare to migrate in response to the VEGF gradient, for which they extend filopodia studded with VEGFR2 [24]. The gradient is established by VEGF-A binding to components of extracellular 6

16 matrix (ECM), which establishes a path for the endothelial cell to follow. This gradient determines the overall directional growth of endothelial sprouts [30]. In order to accomplish migration, the mature parent vessel needs to destabilize by loosening endothelial cell contacts with each other and with the surrounding support structures [31,32], which is in part mediated by PECAM-1 and VE-cadherin re-localization, mentioned above [29]. Now that the endothelial cells are free to move, they need to clear a path through the surrounding sub-endothelial basement membrane. Here, endothelial cells utilize proteinases, such as matrix metalloproteinases (MMPs) to break down and reorganize the surrounding ECM [33]. The break down of the ECM, releases and subsequently activates a plethora of growth factors (e.g. bfgf, VEGF and TGFβ) [34,35]. Now that a path has been cleared, proliferating endothelial cells are able to migrate into the now open space according to the VEGF gradient, utilizing growth factors that were released from the ECM [36]. As the endothelial cells migrate, they begin to form capillary sprouts through the process of tubulogenesis [2]. These newly formed capillaries will eventually join with nearby sprouts or vessels to form a primitive network. Once a connection is made, this nascent vessel begins lumen formation. There are two proposed methods of lumen formation, the first entails the formation and joining of vacuoles within the endothelial cells [37], while the second involves endothelial cell s disruption of adhesion complex proteins (such as VE-cadherin and ZO-1) to form a slit lumen that eventually widens from the shear force of blood flow and growth [38]. These alternative processes of lumen formation have been both experimentally demonstrated and might not be mutually exclusive [39]. 7

17 With the primitive vessel formed, it now enters the maturation phase of angiogenesis. Here, there is a cessation of endothelial cell proliferation and migration, and the basement membrane is reconstituted. Next, there is remodeling and pruning of the new vessel, specification of the vessel into veins and capillaries, and the recruitment of mural cells [30]. Although the process of vessel maturation is not yet fully understood, several angiogenesis modulators have been shown to play critical roles in vessel stability and/or mural cell recruitment. These include angiopoietin 1 (Ang-1), plateled-derived growth factor (PDGF), TGFβ and thrombospondin-1 (TSP-1) [40]. TGFβ1 is activated upon contact between endothelial cells and mesenchymal cells both in vitro and in vivo [41,42,40]. This activation regulates many actions associated with vessel maturation including, inhibition of endothelial cell proliferation and migration, induction of pericyte differentiation, and production of basement membrane, highlighting TGFβ1 has a critical role in vessel remodeling and stability in both in vitro and in vivo models [41,43,44]. Upon completion of the maturation phase, the endothelium returns back to its quiescent state. 1.4 Tip and Stalk Cells interplay between VEGF and Notch Signaling Pathways: During vessel branching, the nascent vessel is comprised of a heterogeneous population of endothelial cells denoted, leading tip and trailing stalk cells [45]. Stimulated by VEGF-A, tip cells produce long, dynamic filopodia, which they use to probe the environment for directional cues. These cells are highly motile, lumen-less and do not proliferate. The endothelial cells that follow the tip cells, termed stalk cells, 8

18 produce fewer filopodia and instead proliferate when stimulated with VEGF-A [30]. Stalk cells also form the vascular lumen [46], and they establish adherens and tight junctions to maintain the integrity of the new sprout [47]. Whether a cell becomes a tip or stalk cell is largely determined by the interplay between VEGF and Notch signaling. The Notch signaling pathway is a key regulator of cell fate, differentiation and patterning processes in a large variety of tissues and organisms [48,49]. This pathway is thought to work along the VEGF signaling pathway to determine endothelial cell fate within the newly forming vessel. In response to VEGF, activation of VEGFR-2 upregulates Deltalike ligand-4 (DLL4), a Notch ligand, expressed in tip cells [50,51]. Expression of DLL4 inhibits surrounding cells ability to become tip cells because it interacts with the Notch receptor on the outside membrane of the cells in direct contact. Upon Notch activation, VEGFR2 is downregulated. Hence, these cells become less responsive to the sprouting activity of VEGF and become stalk cells [52]. Research has demonstrated endothelial cells continuously compete for the tip-cell position by fine-tuning their expression of VEGFR-2 and DLL4, indicating that this signaling cross talk is constantly re-evaluated as cells meet new neighbors [53]. As tip cells form connections with other sprouts or nearby capillaries they get incorporated into the endothelial lining of new vessels. Interestingly, the Notch pathway can act in a very transient, oscillating manner [54], which may promote the dynamic interconversion of tip and stalk cells. Thus, tip and stalk cell determination is transient, reversible and depends on the balance between pro-angiogenic factors (e.g. VEGF) and signals suppressing endothelial cell sprouting and proliferation (e.g. Dll4 Notch activity). When the endothelium is transitioning from active to quiescent, tip cells adopt a phalanx phenotype that is non-proliferative, lumenized, and 9

19 immobile. Phalanx cells are similar to stalk cells, but differ from these cells by their reduced mitogenic response to VEGF and increased quiescence [55]. This phenotype promotes vessel integrity and stabilization by increasing cell adhesion and tapering the response to VEGF [56]. Recently research cumulating in silico, in vitro and in vivo studies have also implicated the adhesion protein VE-cadherin has a new role during sprouting angiogenesis. Endothelial cells along the growing vasculature display two different types of cell-cell junction organization that can be either active or inhibited. Active junctions displayed irregular, serrated distribution of VE-cadherin, whereas inhibited junctions were characterized by straight, continuous staining of VE-cadherin. When junctions are active, there is a high turnover of VE-cadherin and weaker cell-cell contact, which enables higher cell motility and is preferentially associated with a tip cell phenotype. When VE-cadherin stability is increased, the junctions are inhibited causing the cells to have a more static phenotype, just like stalk cells [57]. This research suggests that VEcadherin is another important factor governing how tip and stalk cells behave. 2. Angiogenesis in Cancer: The crucial role of angiogenesis during tumor progression and metastasis is well documented. Although angiogenesis is not an initiating factor of tumorigenesis, it is vital for promoting tumor progression. In order for a mass to grow past 1 mm 3 it requires the formation of new blood vessels to support the increase in oxygen and nutrient needs [58,59,3]. These vessels are formed in response to enhanced production of pro- 10

20 angiogenic factors (e.g., VEGF) and/or reduced expression of anti-angiogenic factors (e.g. TSP-1, endostatin, etc.) that are secreted from tumor and/or stroma cells within the mass [60]. Inappropriate secretion of these angiogenic mediators causes the angiogenic switch; whereby there are enough pro-angiogenic factors present to override the local inhibitory forces. This promotes the angiogenic activity of endothelial cells making up the newly formed vessel [13,61,59] and self-perpetuation of the activation phase of angiogenesis. Endothelial cells are under constant angiogenic sprouting and are never able to enter the maturation phase, causing the blood vessels to be highly disorganized and leaky [62]. These vessel characteristics, in addition to the anti-tumor effects of the administration of endogenous (e.g., endostatin, Thrombospondin) and other chemical or biologic angiogenesis inhibitors (e.g., VEGF-targeting antibodies or small molecule inhibitors [4]) suggest that angiogenesis plays an important role in tumor progression. Thus, a better understanding of this process could lead to more efficient, anti-cancer therapies. 3. Transforming Growth Factor-Beta (TGFβ) 3.1 TGFβ Ligands: TGFβ is part of a large superfamily composed of 30 structurally related growth and differentiation factors including, activins, nodal and BMPs [63,64]. TGFβ is a potent pleiotropic cytokine that regulates mammalian differentiation, development, and homeostasis. Three TGFβ isoforms have been identified, which are expressed in nearly 11

21 all cell types and tissues [65]. Separate genes encode each isoform, which generates the pre-protein sequence consisting of a signal peptide (~30 kda) and the mature TGFβ sequence (~12.5 kda) [66]. After proteolytic processing, the TGFβ monomers utilize disulfide bonding to form homodimers; this results in the formation of the precursor molecule, pro TGFβ. The pro TGFβ dimer is then further processed to yield a small latent complex (SLC) consisting of the mature TGFβ protein and a latency associate peptide (LAP) [67]. The SLC then binds onto the large latent TGFβ binding protein (LTBP) to yield the large latency complex (LLC) [68]. Once secreted, the LLC interacts with various extracellular matrix (ECM) components to allow TGFβ to become deposited in the ECM. Within the ECM, latent TGFβ is stored, where it remains biologically unavailable until its activation [69,34]. Latent TGFβ is activated by several factors, including proteases [70], thrombospondin-1[71], reactive oxygen species [72], and integrins (αvβ6 and αvβ8) [73]. Mature TGFβ is released by these factors through several means including, separating it from the latency-associated proteins, modification of latent complex conformation or degradation of latent TGF-β binding protein [34]. Knockout (KO) studies of TGFβ signaling mediators have elucidated their diverse function in vascular development. Specifically, KO studies on the different TGFβ isoforms have demonstrated they have different, non-overlapping roles [6]. Mice deficient in TGFβ1 die in utero due to vascular defects, including the formation of weak, aberrantly formed capillaries, and this was attributed to problems with endothelial cell differentiation [74]. TGFβ2-null embryos display, among other developmental defects, severe cardiac malformations, which were found to result from defective EnMT [75]. TGFβ3-null mice displayed delayed pulmonary development and cleft palate [76]. These 12

22 KO studies have demonstrated each isoform s importance during development, but have not fully expanded it to the adult body. 3.2 Canonical TGFβ Signaling: TGFβ signaling is mediated through canonical (SMAD-dependent) and noncanonical (SMAD-independent) pathways to regulate transcription, translation, protein synthesis, and post-translational modifications [77]. Even though the downstream effects of TGFβ are heavily context dependent, its signaling pathway is partially conserved in many cell types [5]. During the canonical signaling pathway, TGFβ ligands bind to the type 2 TGFβ receptor (TβRII) with high affinity, this in turn, recruits the type 1 TGFβ receptor (TβRI). These receptors both contain transmembrane serine/threonine kinase domains. TβRII s kinase domain is constitutively active but TβRI kinase activation depends on being close to TβRII for activation. Once in close proximity, TβRII is able to phosphorylate the GS domain of TβRI [78]. There are 2 different types of TβRI, activinlike kinase-5 (ALK5) and activin-like kinase-1 (ALK1). ALK5 is widely expressed across all cell types, whereas expression of ALK1 is restricted to the endothelium. Upon TβRI phosphorylation, nuclear effector proteins called receptor-smads (R-smads), can now bind. This association enables their activation via phosphorylation of SMAD2/3 by ALK5 or SMAD1/5 by ALK1. The now phosphorylated SMAD proteins dissociate from the SMAD Anchor for Receptor Activation (SARA) protein and hetero-oligomerize with co-smad4. Association with SMAD4 enables nuclear translocation and interaction with different transcriptional modulators. Depending of which transcriptional modulator is 13

23 incorporated into this complex, it can serve as either an activator or suppressor of transcription [5]; highlighting another example of how context dependent this pathway is. 3.3 TGFβ in Angiogenesis: TGFβ has a well-documented role in angiogenesis due to its ability to regulate endothelial cell behavior. TGFβ signaling has a complex, biphasic effect on endothelial cell migration and proliferation. Research has demonstrated that low doses of TGFβ promote endothelial cell proliferation and migration, whereas high doses are inhibitory [79]. Goumans et al., further explored this theory though the use of an in vitro scratch wound and proliferation assay on Bovine Aortic Endothelial cells (BAECs). They found that activation of ALK1 and ALK5 have opposite effects on cell migration and proliferation. By performing a dose response experiment it was noted that low doses of TGFβ (0.25 ng/ml 0.5 ng/ml) promoted, whereas high doses ( 5 ng/ml) of TGFβ inhibited migration and proliferation. They attributed this biphasic effect to the activation of ALK1 at lower concentrations and ALK5 at higher concentrations [7]. This study was further expanded to theorize that due to ALK1 s permissive effect on endothelial cell behavior, this receptor could have a more pronounced role during the activation phase of angiogenesis. In contrast, due to ALK5 s inhibitory effect, this receptor is though to be more important during the resolution phase [7]. Another attribute contributing to the complexity of the pathway is TGFβ is part of a very large superfamily. By cells having a multitude of ligands and receptors to signal through, they are able to fine-tune the signaling outcome. For example, ALK5 is activated by TGFβ whereas the physiological 14

24 ligands for ALK1 also include BMP9 and BMP10 [80]. It is also important to note that these ALK receptors do not act independent of each other. ALK5 is needed for the incorporation and activation of ALK1 within the signaling complex, but in turn ALK1 s presence inhibits ALK5 s activity within this complex [81]. Research has also implicated endoglin as being partially responsible for the activation of ALK1 over ALK5. Endoglin is a type III TGFβ receptor, which serves as a co-receptor for type I/II receptors and is primarily expressed in endothelial cells [82]. These initial findings were complemented by overexpression and knock down studies, where it was observed that endoglin over expression promoted ALK1 s signaling but inhibited ALK5 s. When Endoglin was knocked down, there was impaired ALK1 signaling [83]. Overall, TGFβ s role in angiogenesis is highly complex, as depicted by the biphasic effect it has on endothelial cell behavior, thus more research is still needed to understand this mechanism. 3.4 TGFβ in Disease Aberrant TGFβ signaling has been implicated in many different diseases, such as Hereditary Haemorrhagic Telangiectasia (HHT) [5] and cancer [84,85]. HHT, a human vascular disorder, has been directly linked to mutations in ALK1 or the endoglin receptor. In contrast, TGFβ role in cancer is much more complex. In the early stages of tumorigenesis, it acts as a potent inducer of growth arrest, but in the later stages this pathway becomes severely deregulated causing TGFβ to promote tumor growth and progression [84-86]. This functional switch has been termed the TGFβ paradox [87]. It has been proposed, that this switch is due to cancer cells mutating at the later stages to the 15

25 point they no longer respond to the tumor-suppressive effects of TGFβ. Cancer cells can also secrete large amounts of TGFβ into their surrounding environment, and utilize its positive signaling outcomes to promote their proliferation, survival and invasiveness. TGFβ in the later cancer stages can promote tumor cell invasion and metastasis through the process of epithelial to mesenchymal transition (EMT). It can also cause endothelial cells to undergo endothelial to mesenchymal transition (EnMT), which could help explain the increase in cancer-associated-fibroblasts observed in most tumors [88]. TGFβ can also promote tumor angiogenesis indirectly [89,90], and create an immunosuppressive tumor microenvironment [91]. TGFβ s role in tumor angiogenesis was illustrated by in vivo experiments with kinase mutants of ALK5. It was found that ALK5 signaling in tumor cells upregulated MMP-9, a valuable MMP for the angiogenic process [92]. Overall, it is apparent that TGFβ has many roles in tumor progression; hence it is imperative that these mechanisms are further studied. 4. Hypoxia Hypoxia, defined as low oxygen conditions, is a key regulator of both angiogenesis and EnMT [93,94]. Low oxygen conditions are sensed through the regulation of hypoxia-inducible factors (HIF). HIF1 is a heterodimer comprising an oxygen-labile alpha-subunit (HIF-1α) and a constitutively expressed beta subunit (HIF1β or ARNT) [95,96]. Under well-oxygenated conditions, the Von Hippel-Lindau (VHL) protein binds HIF-1α to target it for proteasomal degradation [97]. VHL binding is dependent upon hydroxylation of a specific proline residue within HIF-1α by a member 16

26 of the Prolyl Hydroxylase (PHD) family, PHD2. PHD2 uses oxygen as a substrate, and hence, its activity is inhibited under hypoxic conditions [98]. This inhibition of PHD2 allows for the stabilization of HIF1α, enabling HIF1 to translocate into the nucleus and modulate transcription of hypoxia-response genes that contain hypoxia-response elements (HRE) [99]. Apart from hypoxia, HIF1α protein can be stabilized by the activation of oncogenes such as, RAS and SRC [100] and inactivation of tumor suppressor genes such as, VHL and PTEN [101,100]. Functional HIF1 has been shown to activate the transcription of multiple factors pertaining to angiogenesis, glycolysis and TGFβ signaling [9,8,102,11]. 4.1 Hypoxia s Interaction with TGFβ: Hypoxia regulates the expression of TGFβ ligands both indirectly and directly. For example, hypoxia has been linked to the increased expression and activation of TGFβ2 in Human Umbilical Vein Endothelial Cells (HUVEC) through its interaction with thrombospondin-1, a known activator of latent TGFβ [9]. Also, hypoxia can regulate the expression of TGFβ3 in human trophoblast as depicted by antisense inhibition of HIF-1α expression causing inhibited expression of TGFβ3[11]. Although in the case of TGFβ3, it was undetermined if this was a direct or indirect effect, suggesting future work is needed to expand on this. Most of the work investigating HIF1 s interaction with TGFβ1 has also alluded to their cooperation. Research has shown that smooth muscle cell conditioned media with the addition of TGFβ1 lead to an increase in the angiogenic potential of Bovine Capillary Endothelial (BCE) cells as demonstrated by the increase in 17

27 tube formation. Further, pre-incubation of this conditioned media with a TGFβ IgG prevented this effect, indicating that TGFβ1 is a critical pro-angiogenic factor secreted by smooth muscle cells under low oxygen conditions. Lastly, this research group demonstrated that hypoxia does not directly increase the expression of TGFβ1 but rather affects its level post-transcriptionally [103]. Which could suggest hypoxia s effect on TGFβ1 works in a mechanism similar to what has been seen between TGFβ2 and hypoxia. In contrast to this work, another research group demonstrated in 293T cells that HIF1α directly drives the expression of TGFβ1 by binding to a HRE within its promoter [10]. This discrepancy between cell lines again highlights how context dependent this pathway is. Interestingly, it has also been previously reported that HIF1 and TGFβ cooperate with each other to promote angiogenesis by driving the expression of VEGF-A and Endoglin. To drive expression of endoglin, HIF is able to complex with Smad3 and Sp1 (another transcription factor). It was proposed that in this case Sp1 serves as a docking site for Smad3 because in this region of the promoter there are no Smad binding elements [104]. Cooperation of HIF1 and TGFβ to drive VEGF-A expression in Human Microvascular Endothelial Cells (HMEC-1) has been attributed to a direct interaction between HIF1 and Smad3 on the VEGF-A promoter, which contains active binding sites for both transcription factors [8]. Taken together, this data suggest that HIF1 is able to regulate the levels of TGFβ isoforms present and that this synergistic behavior can subsequently modulate angiogenesis via regulation of VEGF-A and Endoglin. 18

28 4.2 Hypoxia in Angiogenesis: It has long been observed that hypoxia is tightly linked to angiogenesis. When a tissue is experiencing low oxygen conditions it sends out signals to recruit new blood vessels to help bring the oxygen levels back to normal. In vivo studies in mice have further alluded to this close relationship. Mice lacking functional HIF-1α die by embryonic day 10.5 with vascular defects, cardiac malformations, and impaired erythropoiesis. This demonstrates that all three components of the circulatory system are dependent upon HIF-1 for normal development [105]. Angiogenesis is a complex process, involving multiple gene products expressed by different cell types, all contributing to an integrated sequence of events [106]. Consistent with a major role for hypoxia in the overall process of angiogenesis, a large number of genes involved are independently responsive to hypoxia in vitro. For example, VEGF and DLL4 contain HREs within their promoters [ ]. Due to hypoxia regulating many genes associated with angiogenesis, it comes as no surprise that hypoxia also regulates individual cellular processes that together contribute to angiogenesis. For instance, hypoxia has an important role in regulating migration of endothelial cells during angiogenesis. As mentioned previously, in order to migrate, endothelial cells must degrade the surrounding ECM and this is achieved through the upregulation of MMPs such as MMP-2, which is a direct transcriptional target of HIF-1α [110]. 19

29 4.3 Hypoxia in cancer: Research over the years has demonstrated hypoxia s inevitable role in tumor progression. In cancers where it is possible to directly measure tumor oxygenation by Eppendorf microelectrode, such as cervical cancer [111] and soft tissue sarcoma, [112] the average oxygen level is about 1.5% compared to 3-5% in normal adult tissue. Solid tumors contain hypoxic regions due to high rates of cell proliferation coupled with the formation of vasculature that is structurally and functionally abnormal [3]. Experimental manipulations that increase HIF1α expression result in increased tumor growth, whereas loss of HIF activity results in decreased tumor growth [113]. As mentioned above, aberrant HIF activation is caused by genetic alterations in human cancers, most notably VHL loss of function [101]. HIFs activate transcription of genes that play key roles in critical aspects of cancer biology, including stem cell maintenance [114], endothelialmesenchymal transition [115], vascularization [116], invasion and metastasis [117], and radiation resistance [118]. 5. Endothelial to Mesenchymal Transition (EnMT): As an integral part of the vascular system, the endothelium can be described as a single-cell layer of mostly squamous epithelium that provides the inner cell lining of blood vessels. Endothelial cells can exhibit a wide range of morphological variability depending on the local tissue s physiologic needs [119]. Furthermore, in pathologic states, the endothelium can be affected in a number of different ways; the most 20

30 interesting is an extreme form of endothelial plasticity known as endothelial-tomesenchymal transition (EnMT). Endothelial-to-mesenchymal transition may be initiated by autocrine and/or paracrine signals originating from within the surrounding tissue, such as TGFβ or microenvironmental factors such as hypoxia [115] in response to vascular injury. EnMT has been described as a specialized form of epithelial to mesenchymal transition (EMT), a cellular process that has been widely studied due to its role in tumor metastasis. Recent research has demonstrated these cellular processes work through similar pathways given they both give rise to cells that have a similar mesenchymal phenotype. In this review the focus will be on EnMT, but will refer to EMT s mechanism if an EnMT equivalent has not yet been documented. 5.1 EnMT Markers and EnMT-Promoting Signaling During EnMT, endothelial cells separate from an organized cell layer, become fibroblastoid, and invade the underlying tissue. This mesenchymal phenotype can be characterized by loss of endothelial markers, such as CD31 (also known as platelet endothelial cell adhesion molecule-1 (PECAM-1)) and the loss of cell cell junctions (VE-cadherin-dependent adherens junction, and ZO-1/2 and Occludin-dependent tight junctions). Also this phenotype is characterized by the acquisition of invasive and migratory properties, and gain of mesenchymal markers, such as fibroblast-specific protein 1 (FSP1), vimentin, fibronectin, collagen types I and III and/or α-smooth muscle actin (αsma) [120,121,88,122]. 21

31 Endothelial cells undergo EnMT when exposed to TGFβ, Notch ligands, and hypoxia in vitro [122,88, ,93]. These studies have led to a clearer picture of the mechanisms driving this cellular process. The role of TGFβ and notch signaling in EnMT was further elucidated in in vivo studies, where they found that endothelial cells ability to undergo EnMT was modulated in response to manipulations of the TGFβ or Notch pathways [122,88,123,6]. It is also likely that other signaling pathways interact with TGFβ and Notch to mediate EnMT. For example, VEGF, Wnt/β-catenin and BMP have been implicated in EnMT during cardiac development [120]. Different TGFβ isoforms have been demonstrated to have different roles during EnMT. TGFβ2 has a betterdocumented role in driving EnMT [ ], whereas TGFβ1 seems to be the main effector of EMT [130]. Further, TGFβ1 is more likely to cause endothelial cell proliferation due to Endoglin promoting ALK1 as opposed to ALK5 signaling [83]. This highlights how these isoforms can have different roles during EnMT, making it important to study both of them. 5.2 EnMT-Associated Transcription Factors: Many transcription factors (TF) have been implicated in regulation of EnMT, these include members of the Zinc finger family (Snail and Slug), Zeb1, Zeb2, and the WNT family (β-catenin) [131,132]. Hence these factors have been called EnMTassociated transcription factors. Research has demonstrated an increased expression of the Snail family of transcriptional repressors, Snail and Slug, within endothelial cells that undergo EnMT [123,133]. Complementing this research, TGFβ is known to increase the 22

32 expression of Snail, and this is mediated via a SMAD-Binding Domain (SBD) in the Snai1 promoter [134]. The Snail promoter also contains a hypoxia response element (HRE), suggesting that hypoxia also regulates the expression of this factor and could work with TGFβ to regulate Snail-driven EnMT [135]. The subcellular localization of Snail is critical for its transcriptional function [136]. Research has demonstrated that its localization is controlled via phosphorylation by GSK-3β and that this phosphorylation causes Snail s nuclear export into the cytoplasm where it can be targeted for proteosomal degradation [137]. Snail upregulation during EMT is involved in disrupting cell cell junctions, via downregulation of E-cadherin and other junctional proteins [ ]. During EnMT, Snail and Slug are similarly believed to downregulate VE-cadherin, the endothelial cell equivalent to E-cadherin [141]. This downregulation in turn, leads to the disruption of adheren junctions, allowing endothelial cells to delaminate and migrate. Another role of Snail during EMT is to promote cell survival, although this has not yet been documented during EnMT. Research has suggested that the TGFβ paradox could occur in part due to TGFβ regulation of Snail. In an elegant study it was found that alteration of the expression of Snail could modulate hepatocyte cell death in response to TGFβ, where forced expression of Snail conferred apoptosis-resistance but silencing of this gene restored the cell death response [142]. The role of Zeb proteins has yet to be defined in EnMT, but during EMT both ZEB1 and 2 proteins repress the E-cadherin gene, promoting tumor invasion in epithelial tumors [143,144]. Importantly, previous research demonstrated that Zeb2 conditional knock-out in mice resulted in variable heart defects, such as ventricle wall thinning. This was attributed to failure in hematopoietic stem cells differentiation and migration programs, highlighting the importance of Zeb2 23

33 during vascular development [145]. During EMT, β-catenin disassociates from E- Cadherin and enters the cytoplasm. There, it can either translocate into the nucleus to drive EMT-related transcription or be targeted for degradation by GSK3β [146]. Similarly in endothelial cells, β-catenin interacts with VE-Cadherin [147] and although undefined this may suggest a similar role for this protein during EnMT. 5.3 EnMT in Disease: Previous studies of EnMT have focused mainly on embryonic development of the heart, where this process was first discovered [148]. However, recent evidence suggests that EnMT can occur in the adult body under pathological settings, such as cardiac fibrosis and cancer [122,88]. Using two different mouse models, studies by the Kalluri group [88] concluded that EnMT accounts for a significant proportion (up to 30-40%) of cancer-associated fibroblasts (CAFs). The endothelial origin of these CAFs was demonstrated by their co-expression of PECAM-1 along side αsma. These experiments suggest that EnMT is an important mechanism for CAF recruitment to the tumor stroma. CAFs can alter the tumor microenvironment, and play an important role in tumor progression. Specifically, CAFs can deposit various extracellular matrix molecules and also secrete paracrine factors that directly affect how different cell types behave within the tumor. In addition, CAFs release potentially oncogenic signals, such as TGFβ, and are also a main source of VEGF [149]. Hypoxia has also been hypothesized to have a role in EnMT-dependent formation of fibroblasts [150]. Hypoxia has previously been linked to EMT and metastasis [151]. Therefore, it is likely that as tumors grow larger and become 24

34 more hypoxic, a higher incidence of CAF formation through EnMT will occur [152]. Since both TGFβ and hypoxia have roles during CAF formation individually, it is important to see if these factors are working together to cause this phenomenon. Interestingly, TGFβ signaling is also known to regulate EnMT during heart formation [153], where it plays a role in EnMT-dependent cardiac cushion thickening [122]. This, together with TGFβ s abundance in tumors [88], suggests that TGFβ-induced EnMT may potentially play a role in tumor progression via its participation in tumor angiogenesis. Nevertheless, the molecular mechanism of EnMT in tumors has not yet been specifically studied, but is likely to involve similar pathways as described in the setting of cardiac fibrosis [122]. 5.4 EnMT in Angiogenesis: It has been previously hypothesized that EnMT may be involved in angiogenic sprouting by enabling tip cells to migrate into adjacent tissue [88]. This theory is backed by several key similarities between tip cells and cells undergoing EnMT, such as motility, and morphology. Also, at the angiogenic front, tip cells are exposed to growth factors, such as TGFβ and interstitial matrix molecules, which differ from their normal vascular basement membrane components [154]. It is possible that in response to these environmental factors, a small subset of endothelial cells undergo EnMT. In addition, previous studies have demonstrated that pericytes may arise from the endothelium itself [155]. Pericytes are an important component of mature vessels and have the ability to transdifferentiate into vascular smooth muscle cells; hence EnMT may be important in 25

35 stabilizing the nascent vessel during the late stage of angiogenesis. Specifically, EnMT may be another mechanism for recruiting such mural cells during angiogenesis [12,126,155]. Lastly, Slug has been demonstrated to play a role in angiogenesis due to sirna-mediated knockdown of this factor causing impaired sprouting angiogenesis [156]. This research was the first to demonstrate that a transcription factor previously known to be important for EMT/EnMT plays a role in angiogenesis, highlighting a potential link between these two processes. The endothelium is highly dynamic and heterogenous by nature, thus future studies will need to address EnMT in the context of this inherent variation in endothelial phenotypes. In summary, angiogenesis is a highly complex process that is affected by many different signaling pathways including, VEGF, Notch, TGFβ and hypoxia. VEGF causes endothelial cells to migrate or proliferate depending on which other pathways it is interacting with. The role of TGFβ in angiogenesis is not as clearly defined as VEGF because TGFβ has a biphasic effect on endothelial cell migration and proliferation. Similar to TGFβ, hypoxia also has a critical role in governing the angiogenic response. Low oxygen conditions have been linked to increased expression of VEGF-A and TGFβ isoforms, suggesting hypoxia can affect both angiogenesis and TGFβ signaling, but this relationship needs to be further elucidated. Interestingly, both TGFβ and hypoxia have been implicated separately in the regulation of factors involved in EnMT, such as Snail. Although little work has been done investigating if these factors work together to govern this response. This study was designed to address some of the missing pieces within our current knowledge highlighted within this literature review. Specifically, this research 26

36 aimed to investigate how TGFβ and hypoxia present within the microenvironment can work together to promote EnMT and how this might impact angiogenic response to TGFβ. 27

37 Rationale Angiogenesis has a well-documented role in cancer progression through it being aberrantly induced by factors found within the tumor microenvironment [157]. Tumor angiogenesis can be induced by both hypoxia and TGFβ separately, but little work has been done investigating how these factors could interact with each other to modulate an angiogenic response. The TGFβ pathway is highly context dependent and has manifested a biphasic effect on how endothelial cells respond [7]. This context dependency could stem from not only TGFβ being able to signal through multiple isoforms and receptors but also could be through its interaction with the hypoxia pathway. Also, although angiogenesis has been studied extensively the full process is not completely understood. Other mechanisms, such as EnMT could be occurring during angiogenesis as suggested by them both causing endothelial cells to behave similarly [12]. With this is mind, I hypothesize that the EnMT response to TGFβ and its impact in angiogenesis depends on the isoform type and oxygen level present in the tumor microenvironment Objective 1: Evaluate the signaling patterns of endothelial cells in response to TGFβ isoforms 1 and 2 in the presence of low and high oxygen levels. 28

38 Objective 2: Examine the expression of endothelial cell markers and EnMT-associated transcription factors in response to TGFβ isoforms 1 and 2 in the presence of low and high oxygen levels. Objective 3: Investigate how conditions determined by the TGFβ isoform and oxygen level modulate angiogenic response using established in vitro assays. 29

39 Material and Methods Antibodies and Growth Factors The source, catalog number and dilutions for each specific antibody were as follow: from Cell Signaling Technology: SMAD2 (#3103) 1:1000, phospho-smad2 (S465/467, #3101) 1:500, SMAD1 (6944) 1:1000, phospho-smad1/5 (Ser463/465, #9516) 1:1000, β catenin (#8480) 1:20,000, Snail (#3879) 1:1000, Slug (#9585) 1:1000, Zeb1 (#3396) 1:1000, Lamin A/C (#2032) 1:2000, VEGFR2 (#2472) 1:1000; from Santa Cruz Biotechnologies: ZO-1 (sc-33725) 1:4000, Zeb2 (sc-48789) 1:2000, ALK5 (sc-398) 1:500, PECAM (sc-1505) 1:200, VE-cadherin (sc-6458) 1:2000: from Sigma-Aldrich: α- tubulin (#T5168) 1: Secondary antibodies used were goat-anti mouse HRP (A0168) 1:10,000, mouse anti-rabbit HRP (A0545) 1:10,000, rabbit anti-goat (A5420) 1:25,000 or 1:50,000, all from Sigma-Aldrich; goat anti-rat HRP (AP136P) 1:4000 (Millipore) and from Molecular Probes, Life Technologies: Alexa-Fluor-conjugated secondary antibodies, donkey anti-rabbit (A-21206) and donkey anti-mouse (A10036) (5 µg/ml concentration). Growth factors/hormones included: rhtgfβ1, rhtgfβ2, rhvegfa, and rhbmp-9 (all from Invitrogen, Life Technologies). Cell Lines and Culture Conditions Primary Bovine Aortic Endothelial Cells (BAECs) were isolated from the aorta (macrovessel) of a single donor at the Ontario Veterinary College, and were grown and 30

40 maintained as monolayers in high glucose DMEM (Thermoscientific) supplemented with 5% fetal bovine serum (FBS; Invitrogen) and 1 mm Sodium Pyruvate (Sigma). Media was changed every 3-4 days. Cells were used in experiments between passages Once confluency was reached, cells were either passaged or plated into an experiment. To passage BAECs, monolayers were washed once with PBS and then trypsinized using a 0.05 % Trypsin and 0.53 mm EDTA (Invitrogen) solution. Cells were collected through centrifugation at 400 g for 4 min, and resuspended in fresh complete media. Cells were re-plated with a split ratio of 1:4. Cells to be used for experiments were serum starved for 24 hours prior to treatment, by washing the monolayer with PBS and then adding starve media (DMEM supplemented with 0.2% FBS, 1 mm sodium pyruvate). Cells were maintained in a humidified incubator at 37 C in the presence of 5% CO 2 and 95% atmospheric air (21% O 2 ). The representative normoxia oxygen concentration of 21% was chosen based on this standard tissue culture condition, which is used for most in vitro studies worldwide; although normal tissue oxygenation in vivo is estimated to be between 3-5% O 2. The oxygen levels to represent hypoxia was chosen because 1% oxygen was within the range of what has been previously documented within tumor cells in vivo [111,112]. To achieve hypoxic culture conditions, cells were incubated in a hypoxia modulator chamber (Stemcell Technologies) purged with gas (1% oxygen, 5% CO 2, 90% nitrogen balance) at a rate of 20 L/min for 10 minutes. At 10 minutes, the chamber was sealed and placed in the incubator. 31

41 Immunofluorescence Staining Endothelial cells were maintained under standard culture conditions as aforementioned. Sub-confluent monolayers were trypsinized in a solution of 0.05% Trypsin/0.53 mm EDTA (Invitrogen), trypsin was neutralized by the addition of DMEM plus 5% FBS 1 mm sodium pyruvate. Cells were then centrifuged 400 g for 4 minutes, resuspended in complete media, and plated as a single cell suspensions on 1:1 PBS diluted growth factor reduced Matrigel (BD Biosciences) in 8-chamber slides (BD Biosciences). At 85% confluency cells were washed with PBS and starve media was added. After 24 hours, media was replaced with fresh starve media plus treatment for 6 hours. Treatments consisted of 5 ng/ml of TGFβ1 or TGFβ2. Immunofluorescence (IF) was performed following a modified version of the Cell Signaling protocol [ Briefly, 3D cultures in 8- well glass chamber slides (BD Biosciences) were washed twice with ice-cold PBS per chamber, after which cultures were fixed with 4% Paraformaldehyde (EMS) in PBS for 15 minutes at room temperature. The fixed cultures were then washed three times with PBS and permeabilized with 0.2% Triton-X (Bio-Rad) in PBS for 10 minutes, followed by three washes with PBS. Permeabilized cultures were blocked for one hour at room temperature using blocking buffer consisting of 5% normal donkey serum (Sigma) and 0.2% Triton X-100 (Bio-Rad) solubilized in PBS. Primary antibody was then diluted in antibody dilution buffer consisting of 1% BSA (Santa Cruz) and 0.3% Triton X-100, and added to the fixed cultures overnight at 4 C. Next day, fixed cultures were washed 3 times with PBS then incubated with Alexa-Fluor-conjugated secondary antibodies diluted 32

42 in antibody dilution buffer for 1 hour at room temperature in the dark. Following incubation with secondary antibody, cultures were washed three times in PBS, followed by a 5 minute incubation with 0.3 µm 4', 6 -diamidino-2- phenylindole (DAPI) (Sigma- Aldrich) in PBS. The chambers were then removed, and slides were mounted with coverslips using Prolong Gold (Promega) to preserve the fluorescence. All slides were analyzed using an Olympus FV1200 confocal microscope (Olympus). Treatment and Western Blotting For analysis of protein expression, 80% confluent endothelial cells were serum starved with reduced serum media for 24 hours, after which the spent media was changed to fresh reduced serum media containing TGFβ1 (0.5 and 5 ng/ml) or TGFβ2 (0.5 and 5 ng/ml) for 6 hours in normal (21% oxygen) or hypoxic (1% oxygen). Two collection methods were employed, nuclear cytosolic fractionation and whole cell lysis. Nuclear/Cytoslic fractionation was carried out as per manufacture s instructions (BioVision Research Products). For whole cell lysis, cells were lysed with protein lysis buffer containing 20 mm Tris-HCl, 150 mm NaCl, 1 mm Na 2 EDTA, 1 mm EGTA, 1% Triton, 2.5 mm sodium pyrophosphate, 1 mm β-glycerophosphate, 1 mm Na 3 VO 4, and 1 µg/ml leupeptin (Cell Signaling) supplemented with 1 mm PMSF, 2 µg/ml aprotinin (all from Sigma-Aldrich), 1 mm Na 3 VO 4 (New England Biolabs), and 1 % phosphatase inhibitor cocktail 2 (Sigma-Aldrich). Protein samples were resolved on 7.5%, or 10% polyacrylamide (Bio-Rad) gels and then transferred to a PVDF membrane (Roche), washed twice with Tris Buffered Saline plus 0.01 % Tween (TBS-T) and then blocked 33

43 for one hour with either 5% milk in TBS-T, 5% BSA in TBS-T or 1 % BSA/ 5% milk in TBS-T. Membranes were incubated with primary antibody diluted in blocking solution overnight at 4 C. Next day, membranes were washed 3 times per 10 min with TBS-T, and then incubated with a HRP-secondary antibody for 1 hour at room temperature. Membranes were then washed 3 times per 10 min in TBS-T following secondary antibody incubation. Immunoglobulin-antigen complexes were incubated with Luminata chemiluminescence reagent (Roche) for 1 min and exposed on a Chemidoc MP Image System (Bio-Rad). Images captured were quantified using Image Lab software (Bio- Rad). Protein loading was normalized using bands for α-tubulin or lamin A/C. Relative phosphorylation levels were normalized to native protein bands. Quantitative Real-Time PCR (qrt-pcr) RNA was isolated from BAEC cells by homogenization in Trizol (Life Technologies) on ice followed by the Aurum total isolation RNA kit (Bio-Rad). cdna synthesis was performed using the iscript cdna synthesis kit (Bio-Rad) with 1 µg total RNA. The cdna that was obtained was diluted 1:4 in nuclease free water and 4 µl of diluted cdna was added to reaction mixes (10 µl) containing 5 µl SsoFast Eva Green Supermix (Bio- Rad) and 500 nm of each primer. Primer sequences were obtained from primer blast (NCBI), and are as follows: bovine VEGFR2 F 5 - TCCGAAGGCTCAAACCAGAC - 3, VEGFR2 R 5 - ATCGGAGGAGAGCTCAGTGT-3, Snail F 5 - GCAGCTATTTCAGCCTCCTG -3, Snail R 5 - TTCTGGGAGACACAGTGGTG -3. Primers were optimized using an eight point standard curve and temperature gradient to 34

44 determine primer efficiency. Acceptable efficiency was deemed between 90% and 105%. Amplicon product size was confirmed by agarose gel electrophoresis. qrt-pcr amplifications were carried out using a CFX-96 (Bio-Rad) as follows: an initial denaturation step (2 minutes at 95 C), followed by 40 cycles of 5 seconds at 95 C and 5 seconds at 62 C. Data was expressed as relative gene expression normalized to HPRT mrna, which was determined to be a suitable housekeeping gene. Matrigel Cord- Formation Assay Analysis of endothelial cell s ability to form cord like structures was performed using µ- angiogenesis slide (Ibidi), and according to manufacture s instructions. Briefly, µ- angiogenesis slides were coated with full growth factor Matrigel (1:1 diluted in Endothelial Growth medium (EGM; Lonza) and placed in incubator to solidify. Once gels were formed (~ 30 minutes), 8,000 cells/well were resuspended in EGM supplemented with 5% FBS, Bovine Brain Extract (BBE), human Endothelial Growth factor (hegf), hydrocortisone and gentamicine (All from Lonza, as a kit) plus the specific growth factors used for treatment. Cells were then seeded on top of the solidified Matrigel. Treatments included, 5 ng/ml VEGF-A, 5 ng/ml TGFβ1 or 2. Slides were then placed in the incubator either exposed to 21% oxygen or hypoxia (1% O 2 ) for 6 hours, after which the slides were imaged using phase-contrast microscopy (Olympus CKX41) at a low magnification (4X objective) for analysis. The whole growth surface was captured in one view. Analysis of captured images was performed using WimTube (Wimasis). 35

45 Scratch Wound Migration Assay For analysis of cells migratory ability, endothelial cells were plated as monolayers in 24 well plates. Once cells were 80% confluent, they were serum starved with reduced serum media (0.2% FBS, 1 mm Sodium Pyruvate) for 24 hours. Following starvation, the monolayer was wounded using a P100 tip, washed with PBS to remove cellular debris and the different treatments were added in fresh starve media. Treatments included, 5 and 50 ng/ml VEGF-A, 5 ng/ml TGFβ1 or 2. Images were taken at 0, 4, 6, 8 and 12 hours. Three images were taken per well at the same location for all time points. Analysis of cell migration was done by measuring the width of the wound using Image J software (NIH). Wound width was normalized to t=0 time point to determine the relative distance cells had migrated at each given time. Statistical Analysis Means were calculated and plotted along with standard error bars. All statistical analyses were done using GraphPad Prism software version 5.0 (GraphPad), and only in the cases where at least 3 replicate experiments were performed. Data were first analyzed by a Kruskal Wallis non-parametric one-way ANOVA. Significant differences between means were subsequently determined using the Dunn's multiple comparisons test and were considered statistically significant when the p value was less than Three biological replicates were included for all experiments, with the exception of immunofluorescence data, which were tested in one independent experiment and in the time course study, 36

46 where TGFβ2 has only 2 replicates. qrt-pcr, scratch wound and cord formation assay experiments additionally included three technical replicates per run. 37

47 Results Characterization of TGFβ Receptors and Endothelial Cell Markers in Bovine Aortic Endothelial Cells To demonstrate that the primary cells generated were indeed Bovine Aortic Endothelial Cells (BAEC) and were appropriate to address the hypothesis, the expression of endothelial markers and TGFβ receptors (TβRs) were examined. The cells expressed Alk5 (TβRI), TβRII and Endoglin (an endothelial-specific TβRIII), and no significant changes in these proteins were observed in response to TGFβ1 treatment in normoxia (Figure 1A, B). The cells also expressed endothelial cell (EC) markers VEGFR2 and PECAM/CD31 [56]. In normoxia, 6-hour TGFβ1 treatment causes a non-significant reduction in both of these markers at the highest concentration (5 ng/ml). Similar expression patterns of TβRI (ALK5), TβRII, and Endoglin were observed under hypoxia (Figure 1C,D). However, PECAM shows no reduction in response to 5 ng/ml TGFβ1, while VEGR2 shows nearly 50% reduction in protein levels at this concentration of TGFβ1 (p<0.05 for VEGFR2). Similar results were seen when the cells were treated with the second isoform of TGFβ (TGFβ2). No changes in TβRI, TβRII or Endoglin were observed in normoxia. A significant reduction of about 50% in both EC markers, VEGFR2 and PECAM was observed when BAEC were treated with 5 ng/ml TGFβ2 for 6 hours (Figure 2A, B, p<0.05). In hypoxia, TβRII protein levels remain unchanged in response to TGFβ2 treatment (Figure 2C,D). PECAM protein levels show no significant change in hypoxia, whereas VEGFR2 shows a 50% reduction in protein level compared to the untreated control when treated with 5 ng/ml TGFβ2 (p<0.05 for VEGFR2). These 38

48 results confirm the endothelial origin of the cells and their appropriateness as a system to study the effect of hypoxia on endothelial cell response to TGFβ isoforms 1 and 2. TGFβ 1 and 2 Signaling Kinetics in Normoxia Versus Hypoxia Levels of SMAD 1 and 2 activation in BAEC were analyzed at 2, 6 and 24 hours in response to 5 ng/ml TGFβ1, TGFβ2 in both normoxic (21% oxygen) and hypoxic (1% oxygen) conditions (see Materials and Methods for details). To ensure BAECs were indeed exposed to hypoxic culture conditions a qrt-pcr was performed on samples exposed to high and low oxygen in the absence of TGFβ treatment to investigate the expression of GLUT-1, a known gene that is activated under low oxygen conditions [102]. There was a significant increase in GLUT-1 mrna under hypoxic culture conditions, demonstrating treatment effectiveness (Appendix Figure 1, p<0.05). BMP-9 (10 ng/ml) was used as a positive control for SMAD1 activation [80]. Also, due to the small sample size with TGFβ2 (n=2), statistical analysis was not performed during this study and the results with this isoform were graphed separately. In normoxia, both TGFβ isoforms strongly induce the activation of SMAD2 compared to the untreated control and 10 ng/ml BMP-9 (Figure 3A, B, C, p<0.05 for TGFβ1). TGFβ2 was a relatively stronger activator of SMAD2 at the earlier time point (2 hours), and had a relatively longer lasting effect on SMAD2 activation, when compared to TGFβ1 at the 24 hours time point. BMP- 9 induced SMAD2 activation only at the 2-hour time point, but this effect was negligible compared to the induction observed with TGFβ1/2. As expected, the reverse was observed for SMAD1 activation, which occurs primarily in response to 10 ng/ml BMP-9 39

49 at all time points, and to a much lesser extent in response to TGFβ1 and TGFβ2, but only at the 2-hour time point. BMP9 s effect on SMAD1 can be seen up to 24 hours but weakens over time (Figure 3A, B). Similar results were obtained under hypoxia. The only difference was enhanced SMAD2 activation by TGFβ1 as compared to TGFβ2 at the 6 hours time point (Figure 4A, B, C, p<0.05 for TGFβ1). These data suggest that there are no major differences in the general kinetics of SMAD2 and SMAD1 activation in response to TGFβ isoforms 1 and 2 in normoxic versus hypoxic conditions, but hypoxia enhances the response to TGFβ1, particularly at the 6-hour time point. Results from this section and from the following studies up to those presented in Figure 8 are summarized in Table 1. Time-Dependent Effects of TGFβ 1 and 2 on EnMT Markers in Normoxia Versus Hypoxia To begin addressing the possibility that an EnMT response influences the outcome of TGFβ signaling in endothelium, I examined protein levels of EnMTassociated transcription factors Snail and Slug and their downstream target VE-cadherin after 2, 4, and 6 hours of treatment with 5 ng/ml TGFβ1, TGFβ2 and 10 ng/ml BMP-9 in both normal and low oxygen conditions. Due to the small sample size with TGFβ2 treatment (n=2), statistical analysis was only performed on TGFβ1 and BMP-9. Snail levels increase in response to all treatments under normoxia, at all time points, but this effect was only significant for TGFβ1 at the 6-hour time point (Figure 5A and 5B, p<0.05). In low oxygen conditions, all treatments enhanced Snail levels at the 2- and 6-40

50 hour time points. This effect was significant for BMP-9 at 2 hours, but at 6 hours it was only significant for TGFβ1 (Figure 6A and 6B, p<0.05). As compared to Snail, Slug induction by both TGFβ isoforms under normoxia was weaker or virtually nonexistent at the 2- and 6-hour time points and was only significant for TGFβ1 at the 2-hour time point, while the most pronounced increase in Slug was seen in response to 10 ng/ ml BMP-9 at 24 hours (Figure 5A and 5C p<0.05). Neither TGFβ1 nor TGFβ2 significantly enhanced Slug expression (~2 fold) under hypoxia, while BMP-9, similarly to normoxia, induces Slug significantly (~5 fold) at the 24-hour time point (Figure 6A and 6C, p<0.05). Under both oxygen conditions, VE-cadherin protein levels significantly increase in response to 10 ng/ml BMP-9 but only at 2 hours. However, VE-cadherin protein levels significantly decrease in response to TGFβ1 under hypoxia (Figure 5A, D and 6A, D, p<0.05) at the 6-hour time point. Time-Dependent Effects of TGFβ 1 and 2 on EC Markers in Normoxia Versus Hypoxia To investigate whether any of the aforementioned changes in EnMT markers paralleled changes in the EC markers and EC receptors that can potentially influence TGFβ signaling outcome, the protein levels of PECAM-1, VEGFR2 and TβRIII (Endoglin) were measured under the same treatment conditions and time points. Again, due to the small sample size with TGFβ2 (n=2), statistical analysis was only performed on TGFβ1 and BMP-9. Neither TGFβ isoform affects PECAM-1 expression under normoxia or hypoxia, but PECAM levels decreased significantly in response to BMP-9 41

51 after 24 and 6 hours treatment in normoxia and hypoxia, respectively (Figure 7A, 7B, 8A and 8B, p<0.05). In normoxia, Endoglin levels show no significant change in response to any treatment. In contrast, when treatment is administered under hypoxic conditions 10 ng/ ml BMP-9 treatment for 2 hours causes a significant increase in endoglin protein levels (Figure 7A, 7C, 8A and 8C, p<0.05). VEGFR2 protein levels showed a reduction in response to both TGFβ isoforms in both normal and hypoxic oxygen conditions at 6 and 24 hours. Although the statistical significance of this effect could only be verified for TGFβ1 at 6 hours under normal oxygen conditions (Figure 7D, p<0.05), it should be noted that TGFβ2 caused a more pronounced reduction (~70%) in VEGFR2 at the 24 hours time point under both normal and low oxygen conditions (Figure 7A, 7G, 8A and 8G). Taken together with the time-dependent SMAD2 activation, and the expression of EnMT markers, these results suggest that the 6-hour time point shows the most obvious differences between TGFβ isoforms, as well as the most significant responses to TGFβ1 and/or TGFβ2 when normoxic versus hypoxic conditions are compared. To verify results obtained by western blot at this time point, VEGFR2, and Snail were selected for quantitative real time-pcr (qrt PCR) analysis after 6 hours of treatment with 5 ng/ml of either TGFβ isoform under normoxic and hypoxic conditions. Similar to what was observed at the protein level, VEGFR2 mrna expression decreased by ~50% and 60% in response to TGFβ1 and TGFβ2, respectively. This effect was equivalent in both oxygen levels and was significant for TGFβ2 (Figure: 9A TGFβ2 p<0.05). Snail mrna level significantly increased in response to both TGFβ1 and TGFβ2 regardless of oxygen conditions, but this effect was only significant for TGFβ1 and was slightly enhanced by 42

52 hypoxia (Figure: 9B, p<0.05). Based on all aforementioned results, all subsequent treatments were done for 6 hours. SMAD2 Subcellular Localization Study: To explore if the significant increase in activated SMAD2 seen previously correlates with changes in its subcellular localization, nuclear/cytosolic fractionation was performed on samples treated with either TGFβ isoforms in both normal and hypoxic oxygen conditions. Nuclear localization of psmad2 is imperative for it transcriptional function [64]. In normoxia, an increasing dose response was observed, but only 5 ng/ml TGFβ1 caused a significant increase (~30 fold) in activated SMAD2 nuclear signal when compared to control (Figure 10A, B, p<0.03). Interestingly at this same dose, an even stronger increase in psmad2 nuclear signal was observed in hypoxia (~50 fold increase) when compared to untreated control (Figure 10C, D, p<0.03). A similar, yet weaker SMAD2 activation was seen when BAEC were treated with the second isoform, TGFβ2 under the same conditions. In normal oxygen conditions there was a ~10 fold increase in psmad2 nuclear localization compared to the untreated control when treated with 5ng/mL TGFβ2 (Figure 11A, B, p<0.05). The same increase in nuclear signal of about 10 fold was observed with the same treatment under low oxygen conditions (Figure 11C, D, p<0.05). Here it seems TGFβ2 does not have any different effects on SMAD2 localization between each oxygen condition. Further, it is important to note that regardless of treatment psmad2 is only ever detected in the nucleus and never the cytoplasm. This is true even in untreated, although weakly, suggesting there could be 43

53 mild autocrine TGFβ signaling occurring. Overall, these data also suggest that TGFβ1 is more efficient at increasing SMAD2 nuclear translocation than TGFβ2 in both oxygen conditions, particularly at the 5 ng/ml dose. To confirm SMAD2 nuclear translocation, an immunofluorescence assay was performed. In line with what was observed via nuclear-cytoplasmic fractionation, treatment of 5 ng/ml TGFβ1 or 2 under both oxygen conditions resulted in SMAD2 (green) moving from the cytoplasm into the nucleus as depicted by the overlap of SMAD2 signal with DAPI (blue), a nuclear stain (Figure 12A, B). EnMT-Associated Transcription Factor Localization Study: Similar to psmad2, localization of EnMT-associated transcription factors is also important for their function as transcriptional modulators and these factors require nuclear translocation in order to function properly [131]. Hence, subcellular localization of EnMT-associated transcription factors was analyzed in nuclear/cytosolic fractioned samples in response to either TGFβ isoform in both normoxia and hypoxia at the aforementioned time point. High dose (5 ng/ml) TGFβ1 treatment, under normal oxygen conditions resulted in a 3.5 fold increase in nuclear localized Snail (Figure 13A, B, p<0.05). In contrast to Snail, Slug nuclear localization remained unchanged regardless of treatment dose in normoxia (Figure 13A, C). Another member of the EnMT-associated transcription factor family, Zeb1, showed a similar trend to Snail when treated with TGFβ1 under normal oxygen conditions. 5ng/mL treatment resulted in a 2.5 fold increase in its nuclear signal (Figure 13A, D, p<0.05). In contrast, Zeb2 nuclear protein levels 44

54 were unaffected by TGFβ1 treatment in normoxia (Figure 13A, E). Snail nuclear localization in hypoxia was similar to what was seen in normoxia. 5 ng/ml TGFβ1 treatment resulted in a 4-fold increase compared to the untreated control (Figure 13F, G, p<0.05). In contrast to what was observed in normoxia, Slug nuclear signal showed a dose-dependent, albeit no statistically significant increase in nuclear localization in response to TGFβ1 (Figure 13F, H), while the TGFβ1-induced increase in Zeb1 nuclear localization was similar to that observed in normoxia but lacked statistical significance (Figure 13F, I). Interestingly, in the case of Zeb2, hypoxic conditions resulted in a significant decrease at the highest dose of TGFβ1 (Figure 13F, J, p<0.05). These data suggests that the oxygen level present influences TGFβ1 s regulation of both Slug and Zeb proteins localization. Also, with respect to Snail, it seems that TGFβ1 s effect is enhanced by low oxygen conditions. Analysis of nuclear/cytosolic samples treated with TGFβ2 resulted in similar results as compared to the isoform 1. Snail nuclear localization significantly increased in response to 5 ng/ml TGFβ2 in both oxygen conditions, (4 fold increase for normal and 3.5 fold increase in hypoxia) (Figure 14A, B, F, G, p<0.05). Slug nuclear levels did not change under normoxia, but they increased up to 2 fold in response to TGFβ2 treatment under low oxygen conditions (Figure 14A, C, F, H). In the case of Zeb1, TGFβ2 treatment of 0.5 or 5 ng/ml in either oxygen condition did not significantly change the level of nuclear translocation, although there was a weak dose response under both oxygen conditions (Figure 14A, D, F, I). In contrast to Zeb1, Zeb2 still shows a decreasing trend in response to TGFβ2, and this effect was significant under low oxygen conditions, similar to what was seen with TGFβ1 (Figure 14A, E, F, J, p<0.05). Taken together this data suggest that both TGFβ isoforms cause similar effects on Snail, Slug, 45

55 Zeb1 and Zeb2 s nuclear translocation. Lastly, it is important to note that all of these transcription factors are expressed, albeit at lower levels, under basal conditions, and are only being detected in the nucleus. Although the basal levels are unexpected, the nuclear localization could be explained by the presence of nuclear localization sequences that favor their nuclear accumulation in the absence of cytoplasmic sequestration/degradation signals like what has been demonstrated for Snail [136]. Nuclear translocation of Snail in response to both TGFβ isoforms and oxygen levels was confirmed through the use of immunofluorescence (IF). Under low oxygen levels, Snail staining (green) was cytoplasmic in untreated cells, but upon TGFβ stimulation of either isoform Snail staining co-localized with DAPI (blue) in addition to the cytoplasm, suggesting Snail was being shuttled into the nucleus (Figure 15A, B). These data further demonstrates that both TGFβ1 and 2 treatments increase Snail nuclear translocation. One important difference between the IF and western blot data is that there was cytoplasmic staining seen in the Snail IF but this was not detected in the western blot. This could be caused by the presence of Matrigel used during the IF. Matrigel provides a basement membrane-like support structure that can change how the cells behave due to it mimicking in vivo more appropriately [158]. Alternatively, this could be explained by these assays having different sensitivities. Localization of another EMT-associated transcription factor, β- catenin [146], was also assessed by immunofluorescence. It was observed that in the absence of treatment, β-catenin localized solely to the membrane under both normoxic and hypoxic oxygen conditions. Treatment with 5 ng/ml of either TGFβ isoform caused β-catenin to dissociate from the membrane and increase its cytoplasmic localization (Appendix Figure 2A, B). Interestingly, β-catenin was never observed in the nucleus, 46

56 suggesting it is not acting as a transcriptional modulator in this case. Still, the morphological change seen in the endothelial cells where they lost their typical cobblestone morphology and took on a more elongated, separated phenotype is still consistent with an EnMT response. Due to β-catenin s lack of nuclear staining it is apparent that this TF is not driving this response and its change in localization may be a byproduct of the dissolution of adherens junctions. Changes in Adhesion Protein Levels in Response to TGFβ: In the previous section, changes in EnMT-associated transcription factors in response to both TGFβ isoforms and under different oxygen levels were observed. To further investigate how these changes translate into an EnMT response, it was next investigated how cell-cell adhesion protein levels change under the same conditions. In order to do this, levels of VE-cadherin and ZO-1 were analyzed because previous work has demonstrated an inverse relationship between Snail transcriptional activity and these adhesion proteins expression level [141]. Under normal oxygen conditions, treatment with either TGFβ1 dose caused no significant changes in VE-cadherin or ZO-1 (Figure 16A, B, C). In contrast, during hypoxic oxygen conditions both adhesion proteins show a 50% decrease in protein level in response to 5 ng/ml TGFβ1 when compared to the untreated control. However, only VE-cadherin showed a significant reduction (Figure 16D, E, F, p<0.05). This reduction in VE-cadherin protein level is consistent with an EnMT response and again highlights the potential cooperation between TGFβ1 and hypoxia. Treatment with the second TGFβ isoform yielded similar results as TGFβ1 47

57 under normal oxygen conditions, whereby there was no significant change in either VEcadherin or ZO-1 protein levels in response to treatment of 0.5 or 5 ng/ml TGFβ2 (Figure 17A, B, C). Also, this treatment under low oxygen conditions caused no significant change in ZO-1 protein levels (Figure 17D, F). Interestingly in hypoxia, 5 ng/ml TGFβ2 caused a significant increase in VE-cadherin (Figure 17D, E, p<0.05), suggesting that TGFβ1 and 2 might have different roles in the regulation of this protein under low oxygen conditions. Overall, this data suggests that TGFβ treatment of either isoform mainly affects VE-cadherin protein levels instead of ZO-1 and that this response was only seen when treatment was administered under low oxygen conditions. Functional Angiogenic Assays Study: Angiogenesis is a complex process comprised of many different steps including cell migration and tube formation [2]. Thus, to see how all of the previous data translated into an angiogenic response, a scratch wound-migration and Matrigel cord formation assays were performed. BAEC s ability to migrate was measured after 6, 8 and 12 hours of treatment with either TGFβ isoform under normal oxygen conditions, VEGF-A treatment was included as a positive control because it is a known inducer of endothelial cell migration and tubulogenesis [28]. At 6, 8 and 12 hours, 5 ng/ml TGFβ1 and 2 in normoxia caused a reduction in migration, but this effect was only significant in response to TGFβ1 at 8 and 12 hours (Figure 18A, C, D, E, p<0.01). These data suggest that in normoxia, both TGFβ isoforms inhibit endothelial cell migration, but TGFβ1 is more effective than TGFβ2. A similar inhibitory effect by TGFβ was seen during the Matrigel 48

58 cord formation assay performed at 6 hours in both normal and low oxygen conditions. In order to assess cord formation two parameters were measured, total cord length and total number of branching points. These parameters were chosen because they analyze the activity of tip and stalk cells, where total cord length measures stalk cell elongation of the cord and total number of branching points measures how often an endothelial cell branches off and starts a new cord. In normoxia, significant reductions in both parameters were seen in response to 5 ng/ml TGFβ1 and 2 (Figure 19A, B, C, p<0.05 for TGFβ1 and p<0.01 for TGFβ2). Treatment in hypoxia rendered similar results to what was seen in normoxia, whereby there was still a reduction in both cord length and number of branching points. For total cord length both isoforms caused a reduction, but this was only significant for 5 ng/ml TGFβ1 (Figure 19A, D, p<0.05). Also in hypoxia, the number of branching points was significantly reduced in response to both TGFβ isoforms (Figure 16A, E, p<0.01 for TGFβ1 and p<0.05 for TGFβ2). These data suggest that under normal oxygen condition TGFβ2 is a stronger inhibitor of cord formation, but when treatment is administered under low oxygen conditions TGFβ1 is more effective. Again, this highlights that hypoxia mainly enhances the effect TGFβ1 has on endothelial cell behavior. Overall this functional data suggest that high doses of TGFβ inhibits angiogenesis, which is consistent with what has been observed previously [159,7]. 49

59 Figure 1: Characterization of TGFβ Receptors and Endothelial Cell Markers in Bovine Aortic Endothelial Cells (BAEC) in Response to TGFβ1. Monolayer cultures of BAEC were treated with 0.5 and 5 ng/ml TGFβ1 for 6 hours. (A) Representative Western blot and (B) Densitometry analysis shows the effect of TGFβ1 on TβRII, Activin-like Kinase 5 (ALK5), Endoglin (TβRIII), Vascular Endothelial Growth Factor Receptor 2 (VEGFR2) and PECAM under normal oxygen conditions (n=3) at 6 hrs. (C) Representative Western blot and (D) Densitometry analysis shows effect of TGFβ1 on ALK5, Endoglin, VEGFR2 under hypoxic oxygen conditions (n=3) at 6 Hrs. p<0.05 by a nonparametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. All densitometry is normalized to loading control α-tubulin, then normalized to untreated to give relative changes. 50

60 Figure 2: Characterization of TGFβ Receptors and Endothelial Cell Markers in Bovine Aortic Endothelial Cells in Response to TGFβ2. Monolayer cultures of BAEC were treated with 0.5 and 5 ng/ml TGFβ2 for 6 hours. (A) Representative Western blot and (B) Densitometry analysis shows effect of TGFβ2 on TβRII, ALK5, Endoglin (TβRIII), VEGFR2 and PECAM under normal oxygen conditions (n=3) at 6 hrs. (C) Representative Western blot and (B) Densitometry analysis shows effect of TGFβ1 on ALK5, Endoglin, VEGFR2 under hypoxic oxygen conditions (n=3) at 6 Hrs. p<0.05 by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. All densitometry is normalized to loading control α-tubulin, then normalized to untreated control to determine relative changes. 51

61 Figure 3: TGFβ1 and 2 Signaling Kinetics in Normoxia. Monolayer cultures of BAEC were treated with 5 ng/ml TGFβ1 (T1) and 2 (T2), and 10 ng/ml BMP-9 (B9) for 2, 6 and 24 hours. (A) Representative Western blot depicting psmad2, Smad2, psmad1 and Smad1 and (B) Densitometry analysis shows Smad1 and 2 activation in response to treatment of TGFβ1 and BMP-9 under normal oxygen conditions at 2, 6, and 24 hours (n=3). p<0.05 by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. (C) Densitometry analysis depicting Smad1 and Smad2 activation in response to the TGFβ2 treatment, under normal oxygen conditions at the same time points (n=2). The level of Smad activation was determined by first standardizing it to normalized Smad2 levels and then compared between control and treatments. UT= untreated, T1= 5 ng/ml TGFβ1, T2=5 ng/ml TGFβ2, B9= 10 ng/ml BMP-9. 52

62 Figure 4: TGFβ1 and 2 Signaling Kinetics in Hypoxia. Monolayer cultures of BAEC were treated with 5 ng/ml TGFβ1 (T1) and 2 (T2), and 10 ng/ml BMP-9 (B9) for 2, 6 and 24 hours. (A) Representative Western blot depicting psmad2, Smad2, psmad1 and Smad1 and (B) Densitometry analysis shows Smad1 and 2 activation in response to treatment of TGFβ1 and BMP-9 under low oxygen conditions at 2, 6, and 24 hours (n=3). p<0.05 by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. (C) Densitometry analysis depicting Smad2 and Smad1 activation in response to the TGFβ2 treatment, under hypoxic oxygen conditions at the same time points (n=2). The level of Smad activation was determined by first standardizing it to normalized Smad2 levels and then compared between control and treatments. UT= untreated, T1= 5 ng/ml TGFβ1, T2=5 ng/ml TGFβ2, B9= 10 ng/ml BMP-9. 53

63 Figure 5: Time-dependent effects of TGFβ1 and 2 on EnMT Markers in Normoxia. Monolayer cultures of BAEC were treated with 5 ng/ml TGFβ1 (T1) and 2 (T2), and 10 ng/ml BMP-9 (B9) for 2, 6 and 24 hours. Representative Western blot (A) and densitometry analysis shows Snail (B), Slug (C) and Vascular Endothelial- Cadherin (VE-Cadherin) (D) response to TGFβ1 and BMP-9 treatment under normal oxygen conditions (n=3) at 2, 6 and 24 hrs. p<0.05 by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. Densitometry analysis shows Snail (E), Slug (F) and Vascular Endothelial- Cadherin (VE-Cadherin) (G) response to TGFβ2 treatment under normal oxygen conditions (n=2) at 2, 6 and 24 hrs. All densitometry is normalized to loading control α-tubulin, then compared to untreated control to obtain relative fold changes. UT= untreated, T1= 5 ng/ml TGFβ1, T2=5 ng/ml TGFβ2, B9= 10 ng/ml BMP 54

64 Figure 6: Time-dependent effects of TGFβ1 and 2 on EnMT Markers in Hypoxia. Monolayer cultures of BAEC were treated with 5 ng/ml TGFβ1 (T1) and 2 (T2), and 10 ng/ml BMP-9 (B9) for 2, 6 and 24 hours. Representative Western blot (A) and densitometry analysis shows Snail (B), Slug (C) and Vascular Endothelial- Cadherin (VE-Cadherin) (D) response to TGFβ1 and BMP-9 treatment under hypoxic oxygen conditions (n=3) at 2, 6 and 24 hrs. p<0.05 by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. Densitometry analysis shows Snail (E), Slug (F) and Vascular Endothelial- Cadherin (VE-Cadherin) (G) response to TGFβ2 treatment under low oxygen conditions (n=2) at 2, 6 and 24 hrs. All densitometry is normalized to loading control α-tubulin, then compared to untreated control to obtain relative fold changes. UT= untreated, T1= 5 ng/ml TGFβ1, T2=5 ng/ml TGFβ2, B9= 10 ng/ml BMP 55

65 Figure 7: Time-Dependent Effects of TGFβ1 and 2 on EC Markers in Normoxia. Monolayer cultures of BAEC were treated with 5 ng/ml TGFβ1 (T1) and 2 (T2), and 10 ng/ml BMP-9 (B9) for 2, 6 and 24 hours. (A) Representative Western blot and densitometry analysis shows PECAM (B), Endoglin (C) and VEGFR2 (D) in response to TGFβ1 and BMP-9 treatment under normal oxygen conditions (n=3). p<0.05 by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. Densitometry analysis PECAM (E), Endoglin (F) and VEGFR2 (G) in response to TGFβ2 treatment under normal oxygen conditions (n=2). All densitometry is normalized to loading control α- Tubulin, then compared to untreated control to yield fold changes. UT= untreated, T1= 5 ng/ml TGFβ1, T2=5 ng/ml TGFβ2, B9= 10 ng/ml BMP-9. 56

66 Figure 8: Time-Dependent Effects of TGFβ1 and 2 on EC Markers in Hypoxia. Monolayer cultures of BAEC were treated with 5 ng/ml TGFβ1 (T1) and 2 (T2), and 10 ng/ml BMP-9 (B9) for 2, 6 and 24 hours. (A) Representative Western blot and densitometry analysis shows PECAM (B), Endoglin (C) and VEGFR2 (D) in response to TGFβ1 and BMP-9 treatment under low oxygen conditions (n=3). p<0.05 by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. Densitometry analysis PECAM (E), Endoglin (F) and VEGFR2 (G) in response to TGFβ2 treatment under hypoxic oxygen conditions (n=2). All densitometry is normalized to loading control α-tubulin, then compared to untreated control to yield fold changes. UT= untreated, T1= 5 ng/ml TGFβ1, T2=5 ng/ml TGFβ2, B9= 10 ng/ml BMP-9. 57

67 Table 1: Summary of Time Course Experiments. Table summarizing changes in protein levels of activated SMAD (Figure 3, 4), EnMT (Figure 5, 6) and endothelial cell (Figure 7, 8) markers over time in response to TGFβ1, TGFβ2 and BMP-9 in both oxygen conditions. Each directional arrow represents at least a 2-fold change. N.C. = No change and * denotes p<

68 Figure 9: Real Time PCR Analysis of VEGFR2 and Snail Expression Levels in Response to TGFβ1 and TGFβ2 Treatment Under Normal and Low Oxygen Conditions. Graphs depicting relative CT values shows VEGFR2 mrna levels in (A) Normoxia and Hypoxia. Graphs depicting relative CT values shows Snail mrna levels in (B) Normoxia and Hypoxia. Ct values were normalized to HPRT house keeping gene. Relative CT Values we calculated using the Ct method of analysis. p<0.05 as determined by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. 59

69 Figure 10: Smad2 Subcellular Localization in Response to TGFβ1. Nuclear/Cytosolic samples depicting BAEC monolayer cultures response to TGFβ1 treatment under normal and low oxygen conditions at 6 hours. (A) Representative Western blot and (B) Densitometry analysis shows effect of TGFβ1 on psmad2 localization under normal oxygen conditions (n=3) at 6 hrs. (C) Representative Western blot and (B) Densitometry analysis shows effect of TGFβ1 on psmad2 localization under hypoxic oxygen conditions (n=3) at 6 Hrs. p<0.05 by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. All densitometry is normalized to loading control Lamin A/C for the nuclear fractions and α-tubulin for the cytoplasmic fractions. 60

70 Figure 11: Smad2 Subcellular Localization in Response to TGFβ2. Nuclear/Cytosolic samples depicting BAEC monolayer cultures response to TGFβ2 treatment under normoxic and Hypoxic oxygen conditions at 6 hours. (A) Representative Western blot and (B) densitometry analysis shows effect of TGFβ2 on psmad2 localization under normal oxygen conditions (n=3) at 6 hrs. (C) Representative Western blot and (D) densitometry analysis shows effect of TGFβ2 on psmad2 localization under hypoxic oxygen conditions (n=3) at 6 Hrs. p<0.05 by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. All densitometry is normalized to loading control Lamin A/C for the nuclear fractions and α-tubulin for the cytoplasmic fractions. 61

71 Figure 12: Smad2 Subcellular Localization-Immunofluorescence. Immunofluorescence images (60X objective) depicting native Smad2 nuclear translocation in response to 5 ng/ml TGFβ1 and TGFβ2 under normal (A) and low oxygen (B) conditions after 6 hours. Smad2 nuclear translocation occurs regardless of the oxygen level present, but in hypoxia there is an increase in cytoplasmic staining. 62

72 Figure 13: EnMT-Associated Transcription Factor Localization in Response to TGFβ1. Nuclear/Cytosolic samples depicting BAEC monolayer cultures response to TGFβ1 treatment under normoxic and Hypoxic oxygen conditions at 6 hours (A) Representative Western blot and densitometry analysis shows effect of TGFβ1 on Snail (B), Slug (C), Zeb1 (D) and Zeb2 (E) localization under normal oxygen conditions (n=3). (F) Representative Western blot and densitometry analysis shows effect of TGFβ1 on Snail (G), Slug (H), Zeb1 (I) and Zeb2 (J) localization under hypoxic oxygen conditions (n=3). p<0.05 as determined by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. All densitometry is normalized to loading control Lamin A/C for the nuclear fractions and α-tubulin for the cytoplasmic fractions. 63

73 Figure 14: EnMT-Associated Transcription Factor Localization in Response to TGFβ2. Nuclear/Cytosolic samples showing BAEC monolayer cultures response to TGFβ2 treatment under normoxic and Hypoxic oxygen conditions at 6 hours. (A) Representative Western blot and densitometry analysis shows effect of TGFβ1 on Snail (B), Slug (C), Zeb1 (D) and Zeb2 (E) localization under normal oxygen conditions (n=3). (F) Representative Western blot and densitometry analysis shows effect of TGFβ2 on Snail (G), Slug (H), Zeb1 (I) and Zeb2 (J) localization under hypoxic oxygen conditions (n=3). p<0.05 as determined by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. All densitometry is normalized to loading control Lamin A/C for the nuclear fractions and α-tubulin for the cytoplasmic fractions. 64

74 Figure 15: Snail Subcellular Localization in Response to TGFβ1 and 2. Immunofluorescence images (60X objective) depicting native Snail nuclear translocation in response to 5 ng/ml TGFβ1 and TGFβ2 under low oxygen conditions after 6 hours. 65

75 Figure 16: Changes in Adhesion Protein Levels in Response to TGFβ1. Monolayer cultures of BAEC were treated with 0.5 and 5 ng/ml TGFβ1 6 hours under normal and low oxygen conditions. Representative Western blot (A) and densitometry analysis shows effect of TGFβ1 on VE-cadherin (B) and ZO-1 (C) under normal oxygen conditions (n=3). Representative Western blot (D) and densitometry analysis shows effect of TGFβ1 on VE-cadherin (E) and ZO-1 (F) under hypoxic oxygen conditions (n=3). p<0.05 as determined by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. All densitometry is normalized to loading control α-tubulin, then compared to untreated to determine relative fold change. 66

76 Figure 17: Changes in Adhesion Protein Levels in Response to TGFβ2. Monolayer cultures of BAEC were treated with 0.5 and 5 ng/ml TGFβ2 6 hours under normal and low oxygen conditions. Representative Western blot (A) and densitometry analysis shows effect of TGFβ2 on VE-cadherin (B) and ZO-1 (C) under normal oxygen conditions (n=3). Representative Western blot (D) and densitometry analysis shows effect of TGFβ1 on VE-cadherin (E) and ZO-1 (F) under hypoxic oxygen conditions (n=3). p<0.05 as determined by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test. All densitometry is normalized to loading control α-tubulin, then compared to untreated to determine relative fold change. 67

77 Figure 18: TGFβ Treatment Effect on EC Migration/Proliferation in Normoxia. Scratch wound assay on BAECs depicting migration after 6, 8 and 12 hours of treatment with 5 and 50 ng/ml VEGF-A, 5 ng/ml TGFβ1 and 2 in normoxia. Representative phase-contrast images using a 4X objective, depicting the migration of BAECs in response to treatment (A). Graph showing relative migration at 6 (B), 8 (C), and 12 hours (D) of treatment in normal oxygen conditions. Relative migration was obtained by comparing wound width at the given time point versus the initial wound width and dividing it by two to account for both wound margins. Data is also represented as relative migration over time in response to treatments (E). p<0.01 as determined by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test (n=3). UT= untreated, 5V= 5 ng/ml VEGF-A, 50V= 50 ng/ml VEGF-A, T1= 5 ng/ml TGFβ1, T2=5 ng/ml TGFβ2. 68

78 Figure 19: TGFβ Treatment and Oxygen Level Effect on Endothelial Cell Cord Formation. Cord formation assay performed on matrigel in response to 5 ng/ml VEGF-A and 5 ng/ml TGFβ1 and 2 for 6 hours in both normoxic and hypoxic oxygen conditions. (A) Representative image of Wimasis analysis output (A). Graphical representation of total cord length in response to treatment coupled with normal (B) and hypoxic (D) oxygen conditions and graphs depicting the total number of branching points in response to treatments under normal (C) and low (E) oxygen conditions. *p<0.05 and **p<0.01 as determined by a non-parametric analysis of variance (ANOVA) coupled with a Dunns comparison of multiple means post test (n=3). 69

79 Discussion TGFβ Induces EnMT-Like Changes and Signals Primarily Through TGFβRII/ALK5 in BAECs In a quiescent state endothelial cells demonstrate apical-basal polarity and are tightly bound by adherens junctions (AJs) and tight junctions (TJs). Two major proteins of these junctions are VE-cadherin and ZO-1 in AJs and TJs, respectively. VE-cadherin is a vital component of the AJ, where it interacts with other anchor proteins, such as β- catenin, to initiate cell-cell contacts and promote their maturation and maintenance. In contrast, ZO-1 is a main component of TJs, which control the passage of solutes and ions between cells [160,161]. Also in this state endothelial cells express unique markers such as, PECAM, TIE1, TIE2, von Willebrand factor (vwf), and cytokeratins [134]. During EnMT the levels of endothelial cell markers and adhesion proteins are drastically reduced. Their expression is replaced by mesenchymal markers such as, fibroblastspecific protein-1 (FSP-1), alpha-smooth muscle actin (α-sma), vimentin, and N- cadherin. These changes in marker expression cause cells to lose the apical-basal polarity, adhesion and organized cobblestone morphology and take on a more elongated, migratory phenotype [162,12]. In this study, changes in the expression of endothelial cell adhesion proteins and markers in response to TGFβ1 and 2 under normal and low oxygen conditions were explored. It was found that only high doses of TGFβ1 under low oxygen conditions caused a significant decrease in the adhesion protein VE-cadherin, and a similar, albeit non-significant reduction in ZO-1 levels. This reduction in VE-cadherin 70

80 and ZO-1 levels is consistent with an EnMT response. Previous work by our research group has found that 5 ng/ml TGFβ1 under normoxic conditions caused a significant reduction in ZO-1 expression after 144 hours of treatment [163]. This suggests that TGFβ1 treatment could affect TJ proteins more dramatically after a longer exposure. Changes in PECAM, an endothelial specific marker, were observed in response to both TGFβ isoforms under normal oxygen conditions, indicating that these cells lose their endothelial cell characteristics when exposed to TGFβ. This additionally suggests that an EnMT-like response was occurring in BAECs because loss of endothelial cell characteristics is also a hallmark of EnMT. An essential mediator of angiogenesis expressed by endothelial cells is VEGFR2. Although this is not necessarily a marker of EnMT, a reduction in VEGFR2 in response to TGFβ1 and 2 under both oxygen conditions was consistently observed, suggesting that as these cells undergo EnMT-like changes, this vital receptor is also lost. This reduction is in agreement with previous findings showing a dose-dependent lost of VEGFR2 in response to TGFβ1 under normoxic conditions in BAEC [164]. One possible explanation for the loss of this receptor is that VEGF-A signaling inhibits EnMT [165,166], but the underlying mechanism as to how remains unknown [152]. To further explore TGFβ induction of EnMT, EnMT-associated transcription factors Snail, Slug, Zeb1, Zeb2 and β-catenin levels and/or localization were assessed. The basal level of EnMT-associated TF in the nucleus could be an artifact of cell culture. In culture it is possible that I am selecting for cell survival inadvertently and one of the effects of these factors is to promote cell survival [12]. Regardless, TGFβ treatment is still causing a significant increase in some of these factors above the basal levels. Recent 71

81 evidence has pointed towards Snail being critical during TGFβ-induced EnMT [128]. In this work, it was consistently observed that both TGFβ isoforms drastically increased Snail mrna and protein levels and also its nuclear translocation. In absence of an optimized Snail promoter activity assay and time, Snail s activity in the nucleus was indirectly measured by analyzing VE-cadherin protein levels. It is well known that Snail is able to inhibit the expression of cell-cell adhesion molecules during EnMT, especially VE-cadherin. In fact, in Human Dermal Microvascular Endothelial Cells (HDMEC), both Snail and Slug have the ability to bind the VE-cadherin promoter and inhibit its expression [141]. An increase in Snail was indeed seen with a significant decrease in VEcadherin at 6 hours in response to TGFβ1, and coupling that with the increase in Snail nuclear localization suggests that this protein is also serving its functional role as a transcriptional modulator to govern TGFβ-induced EnMT. In contrast, Slug protein levels or nuclear translocation only increased under hypoxia in response to TGFβ1/2 treatment, and this increase was half of that observed for Snail. This could be due to Slug being mainly activated by Notch signaling pathway as opposed to TGFβ [167]. Given the recently demonstrated role of Slug in angiogenesis [156], this suggests that the role played by TGFβ in angiogenesis might be more dependent on Snail as compared to Slug. Zeb1/2 are known repressors of E-cadherin in epithelial cells undergoing EMT, but little is known about their role in endothelial cells. In this study it was found that in response to both TGFβ isoforms Zeb1 increases in both normoxia and hypoxia, while Zeb2 protein levels significantly decrease in hypoxia. Interestingly, the Zeb proteins have previously been found to have opposing roles in regulation of TGFβ signaling in epithelial cells, through their direct interaction with Smad1, 2 and 3. It was found that 72

82 Zeb-1 synergizes with Smads to regulate a number of functions including transcription, whereas Zeb-2 repressed Smad functions [168]. This could explain why TGFβ treatment is causing a significant reduction in Zeb2 protein nuclear localization. Although unable to assess the protein levels of β-catenin via western blotting in this study, changes in its localization were observed via immunofluorescence. β-catenin has been demonstrated to interact directly with both the C-terminal region of VEcadherin s cytoplasmic tail and α-catenin, to aid in the structural organization and function of VE-cadherin by linking it to actin [169]. The immunofluorescence performed showed that upon treatment with either TGFβ isoform and oxygen condition, β-catenin s membrane localization was lost and staining became only cytoplasmic. This could be due to the decrease in VE-cadherin protein available at the AJ site. Interestingly, β-catenin was never seen in the nucleus, indicating it may not serve as an EnMT-associated transcription factor, as previously reported during EMT [146]. Even so, its disassociation from the membrane could be consistent with an EnMT response because upon the adherens junction s dissolution, membrane associated β-catenin is freed and able to join the preexisting cytoplasmic β-catenin pool where it is either degraded or translocated into the nucleus, depending on the activity of GSK3β [169]. Alternatively a time point of 6 hours could be either too early or too late for capturing β-catenin nuclear translocation and different time points could be more appropriate. To elaborate on the type of TGFβ signaling observed during the aforementioned changes in endothelial cell adhesion proteins, EnMT markers and EnMT-associated transcription factors, as well as the phospho-active status of TGFβ signaling proteins 73

83 were analyzed. By demonstrating the expression of TβRI (ALK5), TβRII and the endothelial-specific TβRIII (Endoglin) by the BAECs used in this study, it was first confirmed that these cells are capable of fully responding to TGFβ. The activation of SMAD2 in response to TGFβ1 and 2 under both oxygen conditions demonstrates that treatment is inducing the TβRII/ALK5 signaling arm of the TGFβ pathway [81]. Although unable to find a suitable ALK1 antibody to determine its expression, the activation of SMAD1 in response to a well-documented ligand, BMP9 [80], indirectly suggests that this receptor is expressed and functional. Further, the non-overlapping nature of the EnMT-like responses discussed above and ALK1 activation, suggests that these responses are mainly driven by TGFβRII/ALK5. This points to a similar mechanism between TGFβ-induced EnMT and EMT [87]; the later relying on ALK5 signaling in the absence of input by ALK1, an endothelial specific receptor that is not present on epithelial cells [7]. Overall, the results suggest that EnMT-like changes occur in BAECs and they are caused by TGFβRII/ALK5 as opposed to TGFβRII/ALK1 signaling. A direct way to demonstrate this in follow-up experiments will be to use specific inhibitors of ALK1 and ALK5, and examine the impact of these inhibitors on the expression of EnMT markers in response to TGFβ1/2. Finally, it is important to point out that in the BAECs under study, TGFβ1 is a stronger inducer of EnMT markers, which is contradictory to what was previously believed [128,130]. This discrepancy could come from the fact that these research groups never looked at TGFβ1 or 2 s effect on EnMT under hypoxia. In fact, findings from this study suggest isoform-specific responses, some of which become only evident under the influence of hypoxia. For instance, TGFβ1 caused a significant decrease in VE-cadherin 74

84 protein levels, whereas TGFβ2 treatment caused a significant increase in this protein at 6 hours in hypoxic conditions. In most cases observed in this study TGFβ1 effect was enhanced by hypoxia, whereas TGFβ2 was less affected. It is possible that HIF1 may be directly or indirectly affecting the expression of TGFβ1 ligand or its signaling mediators (through transcriptional or post-translational means) to amplify TGFβ1 s response. Previous work in human kidney cells (293T) has demonstrated that HIF-1α directly regulates the expression of TGFβ1 through an HRE located within its promoter [10]. Alternatively, other work has demonstrated HIF1 has an indirect role in regulating the expression of other TGFβ isoforms (TGFβ2 and 3) in HUVEC and trophoblast respectively [9,11]. Overall, this indicates that it is possible for HIF1 to regulate TGFβ1 s expression directly or indirectly in endothelial cells, but more research is needed to confirm this. Importantly, the possibility of a positive impact of hypoxia on TGFβ2 signaling shouldn t be discarded in more complex scenarios, where positive modulators of TGFβ2 signaling might be present. For instance, it has been shown that hypoxia is able to enhance Thrombospondin-1 (TSP1) induced TGFβ2 activation, leading to enhanced signaling [9]. Whether this scenario is likely to occur in a pro-angiogenic environment, where TSP1 is usually downregulated due to its inhibitory effects on angiogenesis [170,171] remains to be determined. Taken together, all results provide evidences for both functional redundancy and specificities of TGFβ isoforms 1 and 2 in adult vascular endothelium, and thus complement the initial discoveries of such specificities in the developing vasculature, via knockout studies [6]. 75

85 EnMT in Angiogenesis One avenue that has been relatively unexplored in the context of mechanisms underlying angiogenesis is EnMT. The few studies where this cellular process has been investigated, primarily built up a body of generic evidence suggesting a link [134,12,155], which remains to be demonstrated. Two theories that have been proposed to address this potential link are, that tip cells are undergoing some form of EnMT during sprouting angiogenesis [134] or, that EnMT may play a role in vessel stabilization through its role in mural cell differentiation and activity [12,155]. The first theory of tip cells utilizing EnMT comes from the fact that, endothelial cells undergoing EnMT share several characteristics with tip cells involved in angiogenesis. For example, morphologically and behaviorally they are very similar in the sense that they are both elongated fibroblast-like cells that are highly motile. Also, similar pathways have been implicated in the regulation of their phenotype, such as TGFβ [93] and hypoxia [151]. TGFβ regulates EnMT transcription factors such as Snail and Slug [134] and these transcription factors play a role in down-regulating VE-cadherin; an essential component of adherens junctions. This down-regulation of VE-cadherin is critical for both EnMT [122] and sprouting angiogenesis, since instability of VE-cadherin has been attributed to an active cell-cell junction, which enables tip cell migration [57]. In this study we observed a reduction in VE-cadherin that is compatible with a tip cell phenotype. Granted, future studies looking at the localization of VE-cadherin in an in vitro or in vivo model of angiogenesis, where the tip versus stalk cells could be identified, 76

86 would be a better model for confirming this. The major issue with this theory is that tip cells require VEGFR2 in order to respond to the VEGF gradient created during angiogenesis. This gradient is crucial to govern the movement of the tip cell to the area that needs better vascularization. The data presented here demonstrates that cells undergoing TGFβ-induced EnMT have a significant reduction in VEGFR2 protein levels, and this might make them non-responsive to the VEGF-A gradient. Since the decrease in VEGFR2 observed in response to TGFβ1 was seen in monolayer cultures, it remains to be determined whether this occurs during actual angiogenesis. Nevertheless, the negative correlation found between SMAD2 activation and VEGFR2 expression in vessels of colorectal cancer xenografts suggests this mechanism might operate in vivo [164]. In agreement with this, the involvement of ALK5 signaling in VEGFR2 downregulation [164], and the demonstrated role of EMT in promoting cell survival via Snail [172], one might speculate that ALK5-driven signaling, including EnMT, might still contribute to the sprouting phase of angiogenesis possibly by impeding the excessive generation of tip cells, while preventing cell death. Endothelial cells undergoing EnMT typically lose their endothelial characteristics and take on a more mesenchymal phenotype, which includes the expression of smooth muscle actin. Another cell type that is involved in angiogenesis and expresses smooth muscle actin are vascular smooth muscle cells and pericytes, which leads us into the next theory. This second theory is supported by the documented role of TGFβ during the final maturation phase of angiogenesis, and is more in line with the results obtained in this study. Several pieces of evidence from previous studies have collectively demonstrated 77

87 that mural cells necessary for vessel stabilization can arise from endothelial cells, and this could occur via EnMT [12,126,155]. Prior to these, other studies demonstrated that TGFβ1 is activated upon the interaction of endothelial and mesenchymal cells [40,173,42] and that this mediates many processes involved in vessel maturation, such as inhibition of endothelial cell proliferation and migration, and induction of pericyte differentiation [41,43,44]. A similar inhibitory effect was observed with TGFβ on endothelial cell migration in vitro in a scratch wound assay but pericyte recruitment was not assessed in this study. This data is also in line with TGFβ treatment being more so affiliated with the resolution phase of angiogenesis as demonstrated by a few key findings. It is known that TGFβ/ALK5 signaling has a more important role during the resolution phase of angiogenesis, whereas ALK1 is more so involved in the activation phase [134]. High doses of TGFβ are known to signal through ALK5 and inhibit endothelial cell migration and proliferation, whereas low doses of TGFβ activate ALK1 and promote these cellular behaviors [7]. Another piece of evidence from this study that suggest TGFβ s role in the resolution phase was seen by TGFβ treatment inhibiting cord formation. Finally, the downregulation of VEGR2 caused by TGFβ also supports this factor s role in the resolution phase. It has been demonstrated that pericytes utilize Platelet Derived Growth Factor (PDGF) and not VEGFR2 to participate in endothelium stabilization. To stabilize the newly formed endothelial vessels, endothelial cells release PDGF-B to chemoattract PDGF receptor-β (PDGFR-β)-expressing pericytes [1]. The lack of VEGFR2 s role in pericyte recruitment coincides with high doses of TGFβ inhibiting VEGFR2 expression and protein levels. Overall, this points to TGFβ-induced 78

88 EnMT having a role in vessel stabilization during the maturation/resolution phase of angiogenesis. The significant increase in Snail levels and nuclear localization coupled with a reduction in cord formation ability also further points to TGFβ-induced EnMT s role during the maturation phase of angiogenesis. This could be due to TGFβ-induction of Snail occurring though TβRII/ALK5, which was previously mentioned to be more active during the later stages of angiogenesis. Also, the inexistent and weak increase in Slug nuclear localization induced by TGFβ under normoxic and hypoxic conditions, respectively, suggest that the EnMT-like changes observed in this study are probably not mediated by this EnMT-associated transcription factor. Coupled with TGFβ inhibiting cord formation, this suggests that Slug also has no role in this inhibitory effect and thus no role during the resolution phase of angiogenesis. In agreement with this, previous findings indicates that Slug is involved in the activation phase of angiogenesis, as demonstrated by impaired cord formation resulting from its knock out [156]. Overall, these data suggest that Snail might play a role during the resolution phase, whereas Slug plays a role during the activation phase, possibly indicating opposite roles of these factors in angiogenesis. 79

89 Summary and Conclusions Angiogenesis has long been known to enable tumor growth and metastasis, but the specific mechanisms governing this process, are only partially understood. Cancer cells are able to secrete pro-angiogenic factors into the tumor microenvironment to promote angiogenesis [157]. Clinically, endothelial cells are believed to be a more suitable target to stop tumor growth and spread because compared to cancer cells, they are genomically stable. Thus, are less likely to develop resistance to anti-angiogenic therapies [17]. Due to the significance of VEGFR2 in angiogenesis, many researchers have designed anti-angiogenic therapies around blocking signaling mediators of the VEGF pathway. Unfortunately, this strategy has been highly inefficient in the clinics [1]. Due to cancer cells being highly unstable, anti-vegf/vegfr therapies can select for cancer cells that express other pro-angiogenic factors besides VEGF, such as basic fibroblast growth factor or epidermal growth factor (a potent inducer of additional proangiogenic factors) [174], and VEGF-C [175,176]. In addition, recent studies have highlighted other signaling ligands that are inappropriately overexpressed in tumors and can modulate angiogenesis both independently and due to signaling crosstalk with VEGF, such as TGFβ. Aberrant TGFβ signaling is able to decrease the expression of VEGFR2 in endothelial cells and may yield anti-vegfr2 therapies obsolete [164]. Thus, it is crucial to further understand how TGFβ regulates angiogenesis, in hopes to enhance the efficacy of current anti-angiogenic strategies or to discover alternatives. With this goal in mind, this study investigated the effect of hypoxia, a known inducer of angiogenesis, on the expression of EnMT-associated markers by TGFβ isoforms 1 and 2 in BAECs. 80

90 This study demonstrated that treatment with TGFβ isoforms 1 and 2 under both normal and low oxygen conditions preferentially leads to the activation of SMAD2, one of the effector SMADs of the TGFβRII/ALK5 signaling pathway. It was observed that activation of this pathway parallels EnMT-compatible changes in EnMT-associated transcription factors and endothelial cell markers, including the adhesion proteins VEcadherin and ZO-1. This suggests that TGFβ treatment causes BAECs to undergo EnMTlike changes. It was also found that hypoxia preferentially enhances the effects of TGFβ1 treatment and not TGFβ2. Hypoxia enhancing TGFβ1 s effect is in support of previous work implicating HIF1 and TGFβ can act synergistically together to regulate angiogenesis, but the lack of cooperation with TGFβ2 is against other research that alluded to their positive interaction [9,103]. Lastly, this study looked at how these changes relate to the angiogenic response of BAECs, using a scratch/wound assay and a cord formation assay. The inhibitory effect of TGFβ in these assays suggests that TGFβinduced EnMT, and particularly Snail, potentially play a role during the resolution phase of angiogenesis. Granted, these results do not necessarily rule out the possibility that EnMT could also play a role during the activation phase, by preventing excessive tip cell generation while still promoting cell survival, hence future studies are still needed. This is the first study to systematically address the role of hypoxia in modulating endothelial cell response to TGFβ isoforms 1 and 2 in the context of EnMT, and the first to show hypoxia-modulated isoform specificity, as well as the significantly enhanced upregulation of Snail as compared to Slug, by TGFβ isoforms 1 and 2 in endothelial cells. 81

91 Future Directions The findings summarized above highlight the necessity of a number of complementary experiments that need to be performed to elucidate both the mechanisms involved in TGFβ-induced EnMT and the role that this EnMT plays in angiogenesis. First, it is important to investigate how other mesenchymal markers are being affected during this putative TGFβ-induced EnMT, because not all the spectrum of EnMTassociated markers were evaluated in this study. This would include looking at the expression levels of Vimentin, N-cadherin and α-smooth Muscle Actin [121], as well as other tight junction proteins, to demonstrate whether or not TGFβ-induced EnMT is cadherin-specific. In this regard, Claudin-5, an endothelial cell specific tight junction protein, would be a worthwhile contender to look at [169]. Further, it would be beneficial to investigate how the levels of adhesion proteins, including ZO-1, change in response to both TGFβ isoforms and oxygen levels at later time points (e.g., from hours), to expand on previous work from our laboratory [163]. Another place to expand this work would be to further investigate VE-cadherin changes. It would be important to see if EnMT-associated transcription factors are actually regulating this protein at the mrna level, because previous work has demonstrated that it contains a Snail binding site within its promoter [141]. It is also imperative to follow up on my β-catenin data by performing a co-localization immunofluorescence to see how VE-cadherin and β-catenin change in response to treatments. This would confirm the proposed theory as to why β- catenin is dissociating from the membrane into the cytoplasm. 82

92 Apart from expanding the data on cell-cell adhesion molecules, it is necessary to elucidate the mechanism as to how TGFβ1 and hypoxia cooperate to enhance TGFβ1 s effect. It is still unknown whether this is a direct effect like what has been demonstrated for TGFβ1 in epithelial cells[10], or an indirect effect like TGFβ2 or 3 [9,11]. To begin addressing the mechanism could use in silico methods to locate putative hypoxia response elements within TGFβ1 s promoter and them mutate them to see if this synergistic response is still seen. Lastly, to complement this study it would be optimal to confirm if the TGFβ-induced reduction observed in VEGFR2 actually occurs during angiogenesis in vivo. This could be demonstrated via double VEGFR2/PECAM immunofluorescence of tumor xenografts [164] obtained from mice with conditional modifications to TGFβ1 expression/signaling. Finally, to address the hypothesis that Snail and Slug play opposite roles during TGFβ-induced angiogenesis, our laboratory is currently knocking down Snail in BAEC, with the expectancy that this might result in enhanced TGFβ-dependent cord formation. This will be complemented with similar experiments involving Slug. An alternative in vivo study could involve the manipulation of Slug and Snail expression/activity using morpholinos [177] in a TGFβ-induced chorioallantoic membrane model [178]. All together, these strategies will ultimately aid in establishing the stage-specific role of TGFβ-induced EnMT in angiogenesis, as well as the particular meaning of VEGR2 downregulation and Snail upregulation during this process. 83

93 Literature Cited: 1. Carmeliet P, Jain RK (2011) Molecular mechanisms and clinical applications of angiogenesis. Nature 473 (7347): doi: /nature Potente M, Gerhardt H, Carmeliet P (2011) Basic and therapeutic aspects of angiogenesis. Cell 146 (6): doi: /j.cell Folkman J (1971) Tumor angiogenesis: therapeutic implications. The New England journal of medicine 285 (21): doi: /nejm Cesca M, Bizzaro F, Zucchetti M, Giavazzi R (2013) Tumor delivery of chemotherapy combined with inhibitors of angiogenesis and vascular targeting agents. Frontiers in oncology 3:259. doi: /fonc Massague J (2012) TGF- beta signaling in development and disease. FEBS letters 586 (14):1833. doi: /j.febslet Sanford LP, Ormsby I, Gittenberger- de Groot AC, Sariola H, Friedman R, Boivin GP, Cardell EL, Doetschman T (1997) TGFbeta2 knockout mice have multiple developmental defects that are non- overlapping with other TGFbeta knockout phenotypes. Development 124 (13): Goumans MJ, Valdimarsdottir G, Itoh S, Rosendahl A, Sideras P, ten Dijke P (2002) Balancing the activation state of the endothelium via two distinct TGF- beta type I receptors. The EMBO journal 21 (7): doi: /emboj/ Sanchez- Elsner T, Botella LM, Velasco B, Corbi A, Attisano L, Bernabeu C (2001) Synergistic cooperation between hypoxia and transforming growth factor- beta pathways on human vascular endothelial growth factor gene expression. The Journal of biological chemistry 276 (42): doi: /jbc.m Zhang H, Akman HO, Smith EL, Zhao J, Murphy- Ullrich JE, Batuman OA (2003) Cellular response to hypoxia involves signaling via Smad proteins. Blood 101 (6): doi: /blood Hung SP, Yang MH, Tseng KF, Lee OK (2013) Hypoxia- induced secretion of TGF- beta1 in mesenchymal stem cell promotes breast cancer cell progression. Cell transplantation 22 (10): doi: / x Caniggia I, Mostachfi H, Winter J, Gassmann M, Lye SJ, Kuliszewski M, Post M (2000) Hypoxia- inducible factor- 1 mediates the biological effects of oxygen on human trophoblast differentiation through TGFbeta(3). The Journal of clinical investigation 105 (5): doi: /jci Potenta S, Zeisberg E, Kalluri R (2008) The role of endothelial- to- mesenchymal transition in cancer progression. British journal of cancer 99 (9): doi: /sj.bjc Tammela T, Enholm B, Alitalo K, Paavonen K (2005) The biology of vascular endothelial growth factors. Cardiovascular research 65 (3): doi: /j.cardiores Derynck R, Goeddel DV, Ullrich A, Gutterman JU, Williams RD, Bringman TS, Berger WH (1987) Synthesis of messenger RNAs for transforming growth factors alpha and beta and the epidermal growth factor receptor by human tumors. Cancer research 47 (3):

94 15. Semenza GL (2010) Defining the role of hypoxia- inducible factor 1 in cancer biology and therapeutics. Oncogene 29 (5): doi: /onc Carmeliet P (2003) Angiogenesis in health and disease. Nature medicine 9 (6): doi: /nm Carmeliet P (2005) Angiogenesis in life, disease and medicine. Nature 438 (7070): doi: /nature Jesmin S, Hattori Y, Togashi H, Ueno K, Yoshioka M, Sakuma I (2005) Age- related changes in cardiac expression of VEGF and its angiogenic receptor KDR in stroke- prone spontaneously hypertensive rats. Molecular and cellular biochemistry 272 (1-2): Leong TT, Fearon U, Veale DJ (2005) Angiogenesis in psoriasis and psoriatic arthritis: clues to disease pathogenesis. Current rheumatology reports 7 (4): Carmeliet P, Jain RK (2000) Angiogenesis in cancer and other diseases. Nature 407 (6801): doi: / Takahashi H, Shibuya M (2005) The vascular endothelial growth factor (VEGF)/VEGF receptor system and its role under physiological and pathological conditions. Clin Sci (Lond) 109 (3): doi: /cs Shibuya M, Yamaguchi S, Yamane A, Ikeda T, Tojo A, Matsushime H, Sato M (1990) Nucleotide sequence and expression of a novel human receptor- type tyrosine kinase gene (flt) closely related to the fms family. Oncogene 5 (4): Terman BI, Carrion ME, Kovacs E, Rasmussen BA, Eddy RL, Shows TB (1991) Identification of a new endothelial cell growth factor receptor tyrosine kinase. Oncogene 6 (9): Ruhrberg C, Gerhardt H, Golding M, Watson R, Ioannidou S, Fujisawa H, Betsholtz C, Shima DT (2002) Spatially restricted patterning cues provided by heparin- binding VEGF- A control blood vessel branching morphogenesis. Genes & development 16 (20): doi: /gad Waldner MJ, Neurath MF (2012) Targeting the VEGF signaling pathway in cancer therapy. Expert opinion on therapeutic targets 16 (1):5-13. doi: / Petit AM, Rak J, Hung MC, Rockwell P, Goldstein N, Fendly B, Kerbel RS (1997) Neutralizing antibodies against epidermal growth factor and ErbB- 2/neu receptor tyrosine kinases down- regulate vascular endothelial growth factor production by tumor cells in vitro and in vivo: angiogenic implications for signal transduction therapy of solid tumors. The American journal of pathology 151 (6): Kim HJ, Kim TY (2005) Regulation of vascular endothelial growth factor expression by insulin- like growth factor- II in human keratinocytes, differential involvement of mitogen- activated protein kinases and feedback inhibition of protein kinase C. The British journal of dermatology 152 (3): doi: /j x 28. Eilken HM, Adams RH (2010) Dynamics of endothelial cell behavior in sprouting angiogenesis. Current opinion in cell biology 22 (5): doi: /j.ceb

95 29. Eliceiri BP, Paul R, Schwartzberg PL, Hood JD, Leng J, Cheresh DA (1999) Selective requirement for Src kinases during VEGF- induced angiogenesis and vascular permeability. Molecular cell 4 (6): Gerhardt H, Golding M, Fruttiger M, Ruhrberg C, Lundkvist A, Abramsson A, Jeltsch M, Mitchell C, Alitalo K, Shima D, Betsholtz C (2003) VEGF guides angiogenic sprouting utilizing endothelial tip cell filopodia. The Journal of cell biology 161 (6): doi: /jcb Ausprunk DH, Folkman J (1977) Migration and proliferation of endothelial cells in preformed and newly formed blood vessels during tumor angiogenesis. Microvascular research 14 (1): Esser S, Lampugnani MG, Corada M, Dejana E, Risau W (1998) Vascular endothelial growth factor induces VE- cadherin tyrosine phosphorylation in endothelial cells. Journal of cell science 111 ( Pt 13): van Hinsbergh VW, Koolwijk P (2008) Endothelial sprouting and angiogenesis: matrix metalloproteinases in the lead. Cardiovascular research 78 (2): doi: /cvr/cvm Annes JP, Munger JS, Rifkin DB (2003) Making sense of latent TGFbeta activation. Journal of cell science 116 (Pt 2): Hawinkels LJ, Zuidwijk K, Verspaget HW, de Jonge- Muller ES, van Duijn W, Ferreira V, Fontijn RD, David G, Hommes DW, Lamers CB, Sier CF (2008) VEGF release by MMP- 9 mediated heparan sulphate cleavage induces colorectal cancer angiogenesis. Eur J Cancer 44 (13): doi: /j.ejca Ferrara N (1999) Role of vascular endothelial growth factor in the regulation of angiogenesis. Kidney international 56 (3): doi: /j x 37. Kamei M, Saunders WB, Bayless KJ, Dye L, Davis GE, Weinstein BM (2006) Endothelial tubes assemble from intracellular vacuoles in vivo. Nature 442 (7101): doi: /nature Herwig L, Blum Y, Krudewig A, Ellertsdottir E, Lenard A, Belting HG, Affolter M (2011) Distinct cellular mechanisms of blood vessel fusion in the zebrafish embryo. Current biology : CB 21 (22): doi: /j.cub Strilic B, Kucera T, Eglinger J, Hughes MR, McNagny KM, Tsukita S, Dejana E, Ferrara N, Lammert E (2009) The molecular basis of vascular lumen formation in the developing mouse aorta. Developmental cell 17 (4): doi: /j.devcel Walshe TE, Saint- Geniez M, Maharaj AS, Sekiyama E, Maldonado AE, D'Amore PA (2009) TGF- beta is required for vascular barrier function, endothelial survival and homeostasis of the adult microvasculature. PloS one 4 (4):e5149. doi: /journal.pone Hirschi KK, Rohovsky SA, D'Amore PA (1998) PDGF, TGF- beta, and heterotypic cell- cell interactions mediate endothelial cell- induced recruitment of 10T1/2 cells and their differentiation to a smooth muscle fate. The Journal of cell biology 141 (3): Antonelli- Orlidge A, Saunders KB, Smith SR, D'Amore PA (1989) An activated form of transforming growth factor beta is produced by cocultures of endothelial 86

96 cells and pericytes. Proceedings of the National Academy of Sciences of the United States of America 86 (12): Neubauer K, Kruger M, Quondamatteo F, Knittel T, Saile B, Ramadori G (1999) Transforming growth factor- beta1 stimulates the synthesis of basement membrane proteins laminin, collagen type IV and entactin in rat liver sinusoidal endothelial cells. Journal of hepatology 31 (4): Sato Y, Rifkin DB (1989) Inhibition of endothelial cell movement by pericytes and smooth muscle cells: activation of a latent transforming growth factor- beta 1- like molecule by plasmin during co- culture. The Journal of cell biology 109 (1): Affolter M, Zeller R, Caussinus E (2009) Tissue remodelling through branching morphogenesis. Nature reviews Molecular cell biology 10 (12): doi: /nrm Iruela- Arispe ML, Davis GE (2009) Cellular and molecular mechanisms of vascular lumen formation. Developmental cell 16 (2): doi: /j.devcel Dejana E, Orsenigo F, Molendini C, Baluk P, McDonald DM (2009) Organization and signaling of endothelial cell- to- cell junctions in various regions of the blood and lymphatic vascular trees. Cell and tissue research 335 (1): doi: /s Phng LK, Gerhardt H (2009) Angiogenesis: a team effort coordinated by notch. Developmental cell 16 (2): doi: /j.devcel Roca C, Adams RH (2007) Regulation of vascular morphogenesis by Notch signaling. Genes & development 21 (20): doi: /gad Liu ZJ, Shirakawa T, Li Y, Soma A, Oka M, Dotto GP, Fairman RM, Velazquez OC, Herlyn M (2003) Regulation of Notch1 and Dll4 by vascular endothelial growth factor in arterial endothelial cells: implications for modulating arteriogenesis and angiogenesis. Molecular and cellular biology 23 (1): Lobov IB, Renard RA, Papadopoulos N, Gale NW, Thurston G, Yancopoulos GD, Wiegand SJ (2007) Delta- like ligand 4 (Dll4) is induced by VEGF as a negative regulator of angiogenic sprouting. Proceedings of the National Academy of Sciences of the United States of America 104 (9): doi: /pnas Moya IM, Umans L, Maas E, Pereira PN, Beets K, Francis A, Sents W, Robertson EJ, Mummery CL, Huylebroeck D, Zwijsen A (2012) Stalk cell phenotype depends on integration of Notch and Smad1/5 signaling cascades. Developmental cell 22 (3): doi: /j.devcel Jakobsson L, Franco CA, Bentley K, Collins RT, Ponsioen B, Aspalter IM, Rosewell I, Busse M, Thurston G, Medvinsky A, Schulte- Merker S, Gerhardt H (2010) Endothelial cells dynamically compete for the tip cell position during angiogenic sprouting. Nature cell biology 12 (10): doi: /ncb Jakobsson L, Bentley K, Gerhardt H (2009) VEGFRs and Notch: a dynamic collaboration in vascular patterning. Biochemical Society transactions 37 (Pt 6): doi: /bst De Smet F, Segura I, De Bock K, Hohensinner PJ, Carmeliet P (2009) Mechanisms of vessel branching: filopodia on endothelial tip cells lead the way. Arteriosclerosis, 87

97 thrombosis, and vascular biology 29 (5): doi: /atvbaha Ribatti D, Crivellato E (2012) "Sprouting angiogenesis", a reappraisal. Developmental biology 372 (2): doi: /j.ydbio Dejana E, Lampugnani MG (2014) Differential adhesion drives angiogenesis. Nature cell biology 16 (4): doi: /ncb Hanahan D, Weinberg RA (2000) The hallmarks of cancer. Cell 100 (1): Hanahan D, Christofori G, Naik P, Arbeit J (1996) Transgenic mouse models of tumour angiogenesis: the angiogenic switch, its molecular controls, and prospects for preclinical therapeutic models. Eur J Cancer 32A (14): Onimaru M, Yonemitsu Y (2011) Angiogenic and lymphangiogenic cascades in the tumor microenvironment. Front Biosci (Schol Ed) 3: Moserle L, Amadori A, Indraccolo S (2009) The angiogenic switch: implications in the regulation of tumor dormancy. Current molecular medicine 9 (8): Jain RK (2005) Normalization of tumor vasculature: an emerging concept in antiangiogenic therapy. Science 307 (5706): doi: /science Mishra L, Derynck R, Mishra B (2005) Transforming growth factor- beta signaling in stem cells and cancer. Science 310 (5745): doi: /science Derynck R, Zhang YE (2003) Smad- dependent and Smad- independent pathways in TGF- beta family signalling. Nature 425 (6958): doi: /nature Roberts AB, Sporn MB (1993) Physiological actions and clinical applications of transforming growth factor- beta (TGF- beta). Growth Factors 8 (1): Munger JS, Sheppard D (2011) Cross talk among TGF- beta signaling pathways, integrins, and the extracellular matrix. Cold Spring Harbor perspectives in biology 3 (11):a doi: /cshperspect.a ten Dijke P, Arthur HM (2007) Extracellular control of TGFbeta signalling in vascular development and disease. Nature reviews Molecular cell biology 8 (11): doi: /nrm Rifkin DB (2005) Latent transforming growth factor- beta (TGF- beta) binding proteins: orchestrators of TGF- beta availability. The Journal of biological chemistry 280 (9): doi: /jbc.r Taipale J, Miyazono K, Heldin CH, Keski- Oja J (1994) Latent transforming growth factor- beta 1 associates to fibroblast extracellular matrix via latent TGF- beta binding protein. The Journal of cell biology 124 (1-2): Ge G, Greenspan DS (2006) BMP1 controls TGFbeta1 activation via cleavage of latent TGFbeta- binding protein. The Journal of cell biology 175 (1): doi: /jcb Crawford SE, Stellmach V, Murphy- Ullrich JE, Ribeiro SM, Lawler J, Hynes RO, Boivin GP, Bouck N (1998) Thrombospondin- 1 is a major activator of TGF- beta1 in vivo. Cell 93 (7): Barcellos- Hoff MH, Dix TA (1996) Redox- mediated activation of latent transforming growth factor- beta 1. Mol Endocrinol 10 (9): doi: /mend

98 73. Sheppard D (2005) Integrin- mediated activation of latent transforming growth factor beta. Cancer metastasis reviews 24 (3): doi: /s Dickson MC, Martin JS, Cousins FM, Kulkarni AB, Karlsson S, Akhurst RJ (1995) Defective haematopoiesis and vasculogenesis in transforming growth factor- beta 1 knock out mice. Development 121 (6): Bartram U, Molin DG, Wisse LJ, Mohamad A, Sanford LP, Doetschman T, Speer CP, Poelmann RE, Gittenberger- de Groot AC (2001) Double- outlet right ventricle and overriding tricuspid valve reflect disturbances of looping, myocardialization, endocardial cushion differentiation, and apoptosis in TGF- beta(2)- knockout mice. Circulation 103 (22): Kaartinen V, Voncken JW, Shuler C, Warburton D, Bu D, Heisterkamp N, Groffen J (1995) Abnormal lung development and cleft palate in mice lacking TGF- beta 3 indicates defects of epithelial- mesenchymal interaction. Nature genetics 11 (4): doi: /ng Mu Y, Gudey SK, Landstrom M (2012) Non- Smad signaling pathways. Cell and tissue research 347 (1): doi: /s y 78. Wrana JL, Attisano L, Wieser R, Ventura F, Massague J (1994) Mechanism of activation of the TGF- beta receptor. Nature 370 (6488): doi: /370341a0 79. Pepper MS (1997) Transforming growth factor- beta: vasculogenesis, angiogenesis, and vessel wall integrity. Cytokine & growth factor reviews 8 (1): Scharpfenecker M, van Dinther M, Liu Z, van Bezooijen RL, Zhao Q, Pukac L, Lowik CW, ten Dijke P (2007) BMP- 9 signals via ALK1 and inhibits bfgf- induced endothelial cell proliferation and VEGF- stimulated angiogenesis. Journal of cell science 120 (Pt 6): doi: /jcs Goumans MJ, Valdimarsdottir G, Itoh S, Lebrin F, Larsson J, Mummery C, Karlsson S, ten Dijke P (2003) Activin receptor- like kinase (ALK)1 is an antagonistic mediator of lateral TGFbeta/ALK5 signaling. Molecular cell 12 (4): Letamendia A, Lastres P, Botella LM, Raab U, Langa C, Velasco B, Attisano L, Bernabeu C (1998) Role of endoglin in cellular responses to transforming growth factor- beta. A comparative study with betaglycan. The Journal of biological chemistry 273 (49): Lebrin F, Goumans MJ, Jonker L, Carvalho RL, Valdimarsdottir G, Thorikay M, Mummery C, Arthur HM, ten Dijke P (2004) Endoglin promotes endothelial cell proliferation and TGF- beta/alk1 signal transduction. The EMBO journal 23 (20): doi: /sj.emboj Padua D, Massague J (2009) Roles of TGFbeta in metastasis. Cell research 19 (1): doi: /cr Massague J (2008) TGFbeta in Cancer. Cell 134 (2): doi: /j.cell Inman GJ (2011) Switching TGFbeta from a tumor suppressor to a tumor promoter. Current opinion in genetics & development 21 (1): doi: /j.gde

99 87. Wendt MK, Tian M, Schiemann WP (2012) Deconstructing the mechanisms and consequences of TGF- beta- induced EMT during cancer progression. Cell and tissue research 347 (1): doi: /s Zeisberg EM, Potenta S, Xie L, Zeisberg M, Kalluri R (2007) Discovery of endothelial to mesenchymal transition as a source for carcinoma- associated fibroblasts. Cancer research 67 (21): doi: / can Roberts AB, Wakefield LM (2003) The two faces of transforming growth factor beta in carcinogenesis. Proceedings of the National Academy of Sciences of the United States of America 100 (15): doi: /pnas Derynck R, Akhurst RJ, Balmain A (2001) TGF- beta signaling in tumor suppression and cancer progression. Nature genetics 29 (2): doi: /ng Jarnicki AG, Lysaght J, Todryk S, Mills KH (2006) Suppression of antitumor immunity by IL- 10 and TGF- beta- producing T cells infiltrating the growing tumor: influence of tumor environment on the induction of CD4+ and CD8+ regulatory T cells. J Immunol 177 (2): Safina A, Vandette E, Bakin AV (2007) ALK5 promotes tumor angiogenesis by upregulating matrix metalloproteinase- 9 in tumor cells. Oncogene 26 (17): doi: /sj.onc Piera- Velazquez S, Jimenez SA (2012) Molecular mechanisms of endothelial to mesenchymal cell transition (EndoMT) in experimentally induced fibrotic diseases. Fibrogenesis & tissue repair 5 (Suppl 1 Proceedings of Fibroproliferative disorders: from biochemical analysis to targeted therapiespetro E Petrides and David Brenner):S7. doi: / S1- S7 94. Shweiki D, Itin A, Soffer D, Keshet E (1992) Vascular endothelial growth factor induced by hypoxia may mediate hypoxia- initiated angiogenesis. Nature 359 (6398): doi: /359843a0 95. Wang GL, Semenza GL (1995) Purification and characterization of hypoxia- inducible factor 1. The Journal of biological chemistry 270 (3): Wang GL, Jiang BH, Rue EA, Semenza GL (1995) Hypoxia- inducible factor 1 is a basic- helix- loop- helix- PAS heterodimer regulated by cellular O2 tension. Proceedings of the National Academy of Sciences of the United States of America 92 (12): Kaelin WG, Jr., Ratcliffe PJ (2008) Oxygen sensing by metazoans: the central role of the HIF hydroxylase pathway. Molecular cell 30 (4): doi: /j.molcel Epstein AC, Gleadle JM, McNeill LA, Hewitson KS, O'Rourke J, Mole DR, Mukherji M, Metzen E, Wilson MI, Dhanda A, Tian YM, Masson N, Hamilton DL, Jaakkola P, Barstead R, Hodgkin J, Maxwell PH, Pugh CW, Schofield CJ, Ratcliffe PJ (2001) C. elegans EGL- 9 and mammalian homologs define a family of dioxygenases that regulate HIF by prolyl hydroxylation. Cell 107 (1): Arany Z, Huang LE, Eckner R, Bhattacharya S, Jiang C, Goldberg MA, Bunn HF, Livingston DM (1996) An essential role for p300/cbp in the cellular response to hypoxia. Proceedings of the National Academy of Sciences of the United States of America 93 (23):

100 100. Giaccia A, Siim BG, Johnson RS (2003) HIF- 1 as a target for drug development. Nature reviews Drug discovery 2 (10): doi: /nrd Majmundar AJ, Wong WJ, Simon MC (2010) Hypoxia- inducible factors and the response to hypoxic stress. Molecular cell 40 (2): doi: /j.molcel Chen C, Pore N, Behrooz A, Ismail- Beigi F, Maity A (2001) Regulation of glut1 mrna by hypoxia- inducible factor- 1. Interaction between H- ras and hypoxia. The Journal of biological chemistry 276 (12): doi: /jbc.m Sakuda H, Nakashima Y, Kuriyama S, Sueishi K (1992) Media conditioned by smooth muscle cells cultured in a variety of hypoxic environments stimulates in vitro angiogenesis. A relationship to transforming growth factor- beta 1. The American journal of pathology 141 (6): Sanchez- Elsner T, Botella LM, Velasco B, Langa C, Bernabeu C (2002) Endoglin expression is regulated by transcriptional cooperation between the hypoxia and transforming growth factor- beta pathways. The Journal of biological chemistry 277 (46): doi: /jbc.m Kotch LE, Iyer NV, Laughner E, Semenza GL (1999) Defective vascularization of HIF- 1alpha- null embryos is not associated with VEGF deficiency but with mesenchymal cell death. Developmental biology 209 (2): doi: /dbio Conway EM, Collen D, Carmeliet P (2001) Molecular mechanisms of blood vessel growth. Cardiovascular research 49 (3): Manalo DJ, Rowan A, Lavoie T, Natarajan L, Kelly BD, Ye SQ, Garcia JG, Semenza GL (2005) Transcriptional regulation of vascular endothelial cell responses to hypoxia by HIF- 1. Blood 105 (2): doi: /blood Forsythe JA, Jiang BH, Iyer NV, Agani F, Leung SW, Koos RD, Semenza GL (1996) Activation of vascular endothelial growth factor gene transcription by hypoxia- inducible factor 1. Molecular and cellular biology 16 (9): Diez H, Fischer A, Winkler A, Hu CJ, Hatzopoulos AK, Breier G, Gessler M (2007) Hypoxia- mediated activation of Dll4- Notch- Hey2 signaling in endothelial progenitor cells and adoption of arterial cell fate. Experimental cell research 313 (1):1-9. doi: /j.yexcr Krock BL, Skuli N, Simon MC (2011) Hypoxia- induced angiogenesis: good and evil. Genes & cancer 2 (12): doi: / Hockel M, Schlenger K, Aral B, Mitze M, Schaffer U, Vaupel P (1996) Association between tumor hypoxia and malignant progression in advanced cancer of the uterine cervix. Cancer research 56 (19): Brizel DM, Scully SP, Harrelson JM, Layfield LJ, Bean JM, Prosnitz LR, Dewhirst MW (1996) Tumor oxygenation predicts for the likelihood of distant metastases in human soft tissue sarcoma. Cancer research 56 (5): Guan Y, Reddy KR, Zhu Q, Li Y, Lee K, Weerasinghe P, Prchal J, Semenza GL, Jing N (2010) G- rich oligonucleotides inhibit HIF- 1alpha and HIF- 2alpha and block tumor growth. Molecular therapy : the journal of the American Society of Gene Therapy 18 (1): doi: /mt Tsai CC, Chen YJ, Yew TL, Chen LL, Wang JY, Chiu CH, Hung SC (2011) Hypoxia inhibits senescence and maintains mesenchymal stem cell properties through 91

101 down- regulation of E2A- p21 by HIF- TWIST. Blood 117 (2): doi: /blood Mak P, Leav I, Pursell B, Bae D, Yang X, Taglienti CA, Gouvin LM, Sharma VM, Mercurio AM (2010) ERbeta impedes prostate cancer EMT by destabilizing HIF- 1alpha and inhibiting VEGF- mediated snail nuclear localization: implications for Gleason grading. Cancer cell 17 (4): doi: /j.ccr Liao D, Johnson RS (2007) Hypoxia: a key regulator of angiogenesis in cancer. Cancer metastasis reviews 26 (2): doi: /s y 117. Chan DA, Giaccia AJ (2007) Hypoxia, gene expression, and metastasis. Cancer metastasis reviews 26 (2): doi: /s Moeller BJ, Richardson RA, Dewhirst MW (2007) Hypoxia and radiotherapy: opportunities for improved outcomes in cancer treatment. Cancer metastasis reviews 26 (2): doi: /s Chi JT, Chang HY, Haraldsen G, Jahnsen FL, Troyanskaya OG, Chang DS, Wang Z, Rockson SG, van de Rijn M, Botstein D, Brown PO (2003) Endothelial cell diversity revealed by global expression profiling. Proceedings of the National Academy of Sciences of the United States of America 100 (19): doi: /pnas Armstrong EJ, Bischoff J (2004) Heart valve development: endothelial cell signaling and differentiation. Circulation research 95 (5): doi: /01.res da 121. Arciniegas E, Frid MG, Douglas IS, Stenmark KR (2007) Perspectives on endothelial- to- mesenchymal transition: potential contribution to vascular remodeling in chronic pulmonary hypertension. American journal of physiology Lung cellular and molecular physiology 293 (1):L1-8. doi: /ajplung Zeisberg EM, Tarnavski O, Zeisberg M, Dorfman AL, McMullen JR, Gustafsson E, Chandraker A, Yuan X, Pu WT, Roberts AB, Neilson EG, Sayegh MH, Izumo S, Kalluri R (2007) Endothelial- to- mesenchymal transition contributes to cardiac fibrosis. Nature medicine 13 (8): doi: /nm Timmerman LA, Grego- Bessa J, Raya A, Bertran E, Perez- Pomares JM, Diez J, Aranda S, Palomo S, McCormick F, Izpisua- Belmonte JC, de la Pompa JL (2004) Notch promotes epithelial- mesenchymal transition during cardiac development and oncogenic transformation. Genes & development 18 (1): doi: /gad Noseda M, McLean G, Niessen K, Chang L, Pollet I, Montpetit R, Shahidi R, Dorovini- Zis K, Li L, Beckstead B, Durand RE, Hoodless PA, Karsan A (2004) Notch activation results in phenotypic and functional changes consistent with endothelial- to- mesenchymal transformation. Circulation research 94 (7): doi: /01.res c Ishisaki A, Hayashi H, Li AJ, Imamura T (2003) Human umbilical vein endothelium- derived cells retain potential to differentiate into smooth muscle- like cells. The Journal of biological chemistry 278 (2): doi: /jbc.m

102 126. Frid MG, Kale VA, Stenmark KR (2002) Mature vascular endothelium can give rise to smooth muscle cells via endothelial- mesenchymal transdifferentiation: in vitro analysis. Circulation research 90 (11): Romano LA, Runyan RB (2000) Slug is an essential target of TGFbeta2 signaling in the developing chicken heart. Developmental biology 223 (1): doi: /dbio Kokudo T, Suzuki Y, Yoshimatsu Y, Yamazaki T, Watabe T, Miyazono K (2008) Snail is required for TGFbeta- induced endothelial- mesenchymal transition of embryonic stem cell- derived endothelial cells. Journal of cell science 121 (Pt 20): doi: /jcs Azhar M, Runyan RB, Gard C, Sanford LP, Miller ML, Andringa A, Pawlowski S, Rajan S, Doetschman T (2009) Ligand- specific function of transforming growth factor beta in epithelial- mesenchymal transition in heart development. Developmental dynamics : an official publication of the American Association of Anatomists 238 (2): doi: /dvdy Kalluri R, Weinberg RA (2009) The basics of epithelial- mesenchymal transition. The Journal of clinical investigation 119 (6): doi: /jci Peinado H, Olmeda D, Cano A (2007) Snail, Zeb and bhlh factors in tumour progression: an alliance against the epithelial phenotype? Nature reviews Cancer 7 (6): doi: /nrc Aisagbonhi O, Rai M, Ryzhov S, Atria N, Feoktistov I, Hatzopoulos AK (2011) Experimental myocardial infarction triggers canonical Wnt signaling and endothelial- to- mesenchymal transition. Disease models & mechanisms 4 (4): doi: /dmm Carmona R, Gonzalez- Iriarte M, Macias D, Perez- Pomares JM, Garcia- Garrido L, Munoz- Chapuli R (2000) Immunolocalization of the transcription factor Slug in the developing avian heart. Anatomy and embryology 201 (2): van Meeteren LA, ten Dijke P (2012) Regulation of endothelial cell plasticity by TGF- beta. Cell and tissue research 347 (1): doi: /s Luo D, Wang J, Li J, Post M (2011) Mouse snail is a target gene for HIF. Molecular cancer research : MCR 9 (2): doi: / mcr Dominguez D, Montserrat- Sentis B, Virgos- Soler A, Guaita S, Grueso J, Porta M, Puig I, Baulida J, Franci C, Garcia de Herreros A (2003) Phosphorylation regulates the subcellular location and activity of the snail transcriptional repressor. Molecular and cellular biology 23 (14): Zhou BP, Deng J, Xia W, Xu J, Li YM, Gunduz M, Hung MC (2004) Dual regulation of Snail by GSK- 3beta- mediated phosphorylation in control of epithelial- mesenchymal transition. Nature cell biology 6 (10): doi: /ncb Nieto MA (2002) The snail superfamily of zinc- finger transcription factors. Nature reviews Molecular cell biology 3 (3): doi: /nrm Cano A, Perez- Moreno MA, Rodrigo I, Locascio A, Blanco MJ, del Barrio MG, Portillo F, Nieto MA (2000) The transcription factor snail controls epithelial- 93

103 mesenchymal transitions by repressing E- cadherin expression. Nature cell biology 2 (2): doi: / Batlle E, Sancho E, Franci C, Dominguez D, Monfar M, Baulida J, Garcia De Herreros A (2000) The transcription factor snail is a repressor of E- cadherin gene expression in epithelial tumour cells. Nature cell biology 2 (2): doi: / Lopez D, Niu G, Huber P, Carter WB (2009) Tumor- induced upregulation of Twist, Snail, and Slug represses the activity of the human VE- cadherin promoter. Archives of biochemistry and biophysics 482 (1-2): doi: /j.abb Nieto MA (2011) The ins and outs of the epithelial to mesenchymal transition in health and disease. Annual review of cell and developmental biology 27: doi: /annurev- cellbio Grooteclaes ML, Frisch SM (2000) Evidence for a function of CtBP in epithelial gene regulation and anoikis. Oncogene 19 (33): doi: /sj.onc Comijn J, Berx G, Vermassen P, Verschueren K, van Grunsven L, Bruyneel E, Mareel M, Huylebroeck D, van Roy F (2001) The two- handed E box binding zinc finger protein SIP1 downregulates E- cadherin and induces invasion. Molecular cell 7 (6): Goossens S, Janzen V, Bartunkova S, Yokomizo T, Drogat B, Crisan M, Haigh K, Seuntjens E, Umans L, Riedt T, Bogaert P, Haenebalcke L, Berx G, Dzierzak E, Huylebroeck D, Haigh JJ (2011) The EMT regulator Zeb2/Sip1 is essential for murine embryonic hematopoietic stem/progenitor cell differentiation and mobilization. Blood 117 (21): doi: /blood Kim K, Lu Z, Hay ED (2002) Direct evidence for a role of beta- catenin/lef- 1 signaling pathway in induction of EMT. Cell biology international 26 (5): Dejana E, Orsenigo F, Lampugnani MG (2008) The role of adherens junctions and VE- cadherin in the control of vascular permeability. Journal of cell science 121 (Pt 13): doi: /jcs Markwald RR, Fitzharris TP, Smith WN (1975) Sturctural analysis of endocardial cytodifferentiation. Developmental biology 42 (1): Kalluri R, Zeisberg M (2006) Fibroblasts in cancer. Nature reviews Cancer 6 (5): doi: /nrc Higgins DF, Kimura K, Iwano M, Haase VH (2008) Hypoxia- inducible factor signaling in the development of tissue fibrosis. Cell Cycle 7 (9): Jiang J, Tang YL, Liang XH (2011) EMT: a new vision of hypoxia promoting cancer progression. Cancer biology & therapy 11 (8): Medici D, Kalluri R (2012) Endothelial- mesenchymal transition and its contribution to the emergence of stem cell phenotype. Seminars in cancer biology 22 (5-6): doi: /j.semcancer Nakajima Y, Yamagishi T, Hokari S, Nakamura H (2000) Mechanisms involved in valvuloseptal endocardial cushion formation in early cardiogenesis: roles of transforming growth factor (TGF)- beta and bone morphogenetic protein (BMP). The Anatomical record 258 (2):

104 154. Davis GE, Senger DR (2005) Endothelial extracellular matrix: biosynthesis, remodeling, and functions during vascular morphogenesis and neovessel stabilization. Circulation research 97 (11): doi: /01.res e Armulik A, Abramsson A, Betsholtz C (2005) Endothelial/pericyte interactions. Circulation research 97 (6): doi: /01.res d Welch- Reardon KM, Ehsan SM, Wang K, Wu N, Newman AC, Romero- Lopez M, Fong AH, George SC, Edwards RA, Hughes CC (2014) Angiogenic sprouting is regulated by endothelial cell expression of Slug. Journal of cell science 127 (Pt 9): doi: /jcs Folkman J, Merler E, Abernathy C, Williams G (1971) Isolation of a tumor factor responsible for angiogenesis. The Journal of experimental medicine 133 (2): Staton CA, Reed MW, Brown NJ (2009) A critical analysis of current in vitro and in vivo angiogenesis assays. International journal of experimental pathology 90 (3): doi: /j x 159. Pepper MS, Vassalli JD, Orci L, Montesano R (1993) Biphasic effect of transforming growth factor- beta 1 on in vitro angiogenesis. Experimental cell research 204 (2): doi: /excr Gonzalez- Mariscal L, Tapia R, Chamorro D (2008) Crosstalk of tight junction components with signaling pathways. Biochimica et biophysica acta 1778 (3): doi: /j.bbamem Bazzoni G, Dejana E (2004) Endothelial cell- to- cell junctions: molecular organization and role in vascular homeostasis. Physiological reviews 84 (3): doi: /physrev Thiery JP (2003) Epithelial- mesenchymal transitions in development and pathologies. Current opinion in cell biology 15 (6): Zours S (2012) The Characterization of Endothelial- Mesenchymal- Transition in Response to TGF- beta and its Potential Role in Angiogenesis. University of Guelph, Guelph 164. Kuczynski EA, Viloria- Petit AM, Coomber BL (2011) Colorectal carcinoma cell production of transforming growth factor beta decreases expression of endothelial cell vascular endothelial growth factor receptor 2. Cancer 117 (24): doi: /cncr Medici D, Shore EM, Lounev VY, Kaplan FS, Kalluri R, Olsen BR (2010) Conversion of vascular endothelial cells into multipotent stem- like cells. Nature medicine 16 (12): doi: /nm Paruchuri S, Yang JH, Aikawa E, Melero- Martin JM, Khan ZA, Loukogeorgakis S, Schoen FJ, Bischoff J (2006) Human pulmonary valve progenitor cells exhibit endothelial/mesenchymal plasticity in response to vascular endothelial growth factor- A and transforming growth factor- beta2. Circulation research 99 (8): doi: /01.res c 167. Niessen K, Fu Y, Chang L, Hoodless PA, McFadden D, Karsan A (2008) Slug is a direct Notch target required for initiation of cardiac cushion cellularization. The Journal of cell biology 182 (2): doi: /jcb

105 168. Postigo AA (2003) Opposing functions of ZEB proteins in the regulation of the TGFbeta/BMP signaling pathway. The EMBO journal 22 (10): doi: /emboj/cdg Lampugnani MG, Corada M, Caveda L, Breviario F, Ayalon O, Geiger B, Dejana E (1995) The molecular organization of endothelial cell to cell junctions: differential association of plakoglobin, beta- catenin, and alpha- catenin with vascular endothelial cadherin (VE- cadherin). The Journal of cell biology 129 (1): Jimenez B, Volpert OV, Crawford SE, Febbraio M, Silverstein RL, Bouck N (2000) Signals leading to apoptosis- dependent inhibition of neovascularization by thrombospondin- 1. Nature medicine 6 (1): doi: / DiPietro LA, Nebgen DR, Polverini PJ (1994) Downregulation of endothelial cell thrombospondin 1 enhances in vitro angiogenesis. Journal of vascular research 31 (3): Gras B, Jacqueroud L, Wierinckx A, Lamblot C, Fauvet F, Lachuer J, Puisieux A, Ansieau S (2014) Snail family members unequally trigger EMT and thereby differ in their ability to promote the neoplastic transformation of mammary epithelial cells. PloS one 9 (3):e doi: /journal.pone Hirschi KK, Rohovsky SA, Beck LH, Smith SR, D'Amore PA (1999) Endothelial cells modulate the proliferation of mural cell precursors via platelet- derived growth factor- BB and heterotypic cell contact. Circulation research 84 (3): Viloria- Petit AM, Kerbel RS (2004) Acquired resistance to EGFR inhibitors: mechanisms and prevention strategies. International journal of radiation oncology, biology, physics 58 (3): doi: /j.ijrobp Bocci G, Man S, Green SK, Francia G, Ebos JM, du Manoir JM, Weinerman A, Emmenegger U, Ma L, Thorpe P, Davidoff A, Huber J, Hicklin DJ, Kerbel RS (2004) Increased plasma vascular endothelial growth factor (VEGF) as a surrogate marker for optimal therapeutic dosing of VEGF receptor- 2 monoclonal antibodies. Cancer research 64 (18): doi: / can Bianco F, Basini G, Grasselli F (2005) Angiogenic activity of swine granulosa cells: effects of hypoxia and vascular endothelial growth factor Trap R1R2, a VEGF blocker. Domestic animal endocrinology 28 (3): doi: /j.domaniend Braasch DA, Corey DR (2002) Novel antisense and peptide nucleic acid strategies for controlling gene expression. Biochemistry 41 (14): Ribatti D, Nico B, Vacca A, Roncali L, Burri PH, Djonov V (2001) Chorioallantoic membrane capillary bed: a useful target for studying angiogenesis and anti- angiogenesis in vivo. The Anatomical record 264 (4):

106 Appendix I: Supplementary Figures Appendices Appendix Figure 1: Hypoxia Control. Glut1 mrna expression in normoxia compared to hypoxia in untreated BAECs. p<0.05 as determined by a Wilcoxon matched-pairs signed rank test. Ct values were normalized to HPRT house keeping gene. Relative CT Values we calculated using the Ct method of analysis. 97

107 Appendix Figure 2: Localization of β Catenin in Response to TGFβ1 and 2 in Normal and Low Oxygen Conditions. Immunofluorescence images (60X objective) depicting β-catenin cytoplasmic localization in response to 5 ng/ml TGFβ1 and TGFβ2 under normal (A) and low oxygen (B) conditions after 6 hours 98