Novel Signaling Mechanisms for Endoglin During Angiogenesis

Home , Endoglin, Thrombospondin 1

NOVEL SIGNALING MECHANISMS FOR ENDOGLIN DURING ANGIOGENESIS

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the

Graduate School of The Ohio State University

Christopher Pan, B.S.

Graduate Program in Pharmacy

The Ohio State University

2016

Dissertation Committee:

Professor Nam Lee, Advisor

Professor Dale Hoyt

Professor Keli Hu

Abstract

Endoglin is an endothelial-specific TGF-β co-receptor that is essential for angiogenesis and vascular remodeling. It regulates a wide range of cellular and biological processes, including cell proliferation, migration and adhesion. Endoglin’s importance in endothelial biology is highlighted by the embryonic lethal phenotypes observed in endoglin knockout mice and by the presence of vascular diseases including the autosomal dominant disorder,

HHT1. Aside from normal angiogenesis, endoglin has a critical role in tumor-associated angiogenesis and has become a therapeutic target for the treatment of many solid tumor types. Despite its therapeutic potential and established role in physiology and disease, the underlying mechanisms for endoglin’s actions are poorly defined. In Chapter I, we demonstrate that endoglin inhibits endothelial proliferation by suppressing ERK activation. In Chapter II, we establish endoglin’s evolutionary conserved tyrosine motif as a critical determinant of its stability and function. In Chapter III, we show that endoglin promotes angiogenesis by inducing endothelial autophagy. Lastly, in Chapter

IV, we define a novel mechanism by which the endoglin-targeting antibody, TRC105, neutralizes endoglin function through direct coupling of MMP-14 and endoglin at the cell surface. Taken together, the work presented here provides a greater understanding of endoglin biology, defines several novel signaling mechanisms for endoglin during angiogenesis, and highlights endoglin’s therapeutic potential.

Dedication

Dedicated to my parents, Alice and Y.T Pan, and my brother, Daniel Pan

iii

Acknowledgements

I would like to thank my advisor, Dr. Nam Lee. My graduate training and education would not have been possible without his mentorship, guidance, and patience for the past five years. I would also like to thank the members of my thesis committee members, Dr. Dale Hoyt and Dr. Keli Hu, for their offering their scientific insights and encouragement.

I would also like to thank current and past members of the Lee lab for their invaluable assistance and advice including Jeffery Bloodworth, Dr. Sanjay Kumar, Nirav Shah, Sarah Wheeler, and

Naveed Zaman.

I must thank the Ohio State College Pharmacy and the Department of Pharmacology for supporting my graduate training. Lastly, I would like to thank my family and friends for their endless support and encouragement.

Vita

From San Antonio, Texas

2007-2011 B.S. Biochemistry and Molecular Biology

Rhodes College, Tennessee

2011-present Graduate Research Associate

The Ohio State University, Ohio

Publications

1. Pan CC, Bloodworth JC, Mythreye K, Lee NY (2012). Endoglin inhibits ERK-induced c-Myc and cyclin D1 expression to impede endothelial cell proliferation. Biochem Biophys Res Commun. 424(3): 620-623.

2. Kumar S, Pan CC, Bloodworth JC, Nixon A, Theuer C, Hoyt DG, Lee NY (2013). Antibody- directed coupling of endoglin and MMP-14 is a key mechanism for endoglin shedding and deregulation of TGF-β signaling. Oncogene. 33(30): 3970-3979.

3. Pan CC, Kumar S, Shah N, Hoyt DG, Hawinkels LJ, Mythreye K, Lee NY. (2014). Src- mediated post-translational regulation of endoglin stability and function is critical for angiogenesis. J Biol Chem. 289(37): 25486-96.

4. Pan CC, Kumar S, Shah N, Bloodworth JC, Hawinkels LJ, Mythreye K, Hoyt DG, Lee NY. (2015). Endoglin regulation of Smad2 function mediates beclin1 expression and endothelial autophagy. J Biol Chem. 12:290(24):14884-92.

5. Kumar S, Pan CC, Shah N, Wheeler SE, Hoyt KR, Hempel N, Mythreye K, Lee NY. (2016). Activation of mitofusion2 by Smad2-RIN1 complex during mitochondrial fusion. Molecular Cell. 19:62(4):520-31.

6. Pan CC, Cinti J, Shah N, Kumar S, Wheeler SE, Mythreye K, An M, Beattie CE, Rakotondraibe LH, Lee NY. (2016). Angiostatic actions of capsicodendrin through selective inhibition of AKT signaling and disregulated autophagy. Oncogotarget. DOI: 10.18632/oncotarget.9307

v Fields of Study

Major Field: Pharmacy

vi Table of Contents

Abstract ...... ii

Dedication ...... iii

Acknowledgements ...... iv

Vita ...... v

List of Figures ...... x

Abbreviations ...... xii

Chapter

1. Introduction ...... 1

Angiogenesis 1

Transforming Growth Factor β Signaling ...... 2

Non-Canonical Transforming Growth Factor β Signaling ...... 5

Endoglin ...... 6

Endoglin: Therapeutic Potential ...... 9

Autophagy ...... 11

Thesis aim ...... 13

2. Endoglin Inhibits ERK-induced c-Myc and CyclinD1 Expression ..... 19

to Impede Endothelial Cell Proliferation

Abstract ...... 19

Background ...... 20

Materials and Methods ...... 21

vii Figures ...... 24

Result ...... 27

Discussion ...... 29

3. Src-mediated Post-translational Regulation of Endoglin ...... 31

Stability and Function is Critical for Angiogenesis

Abstract ...... 31

Background ...... 31

Materials and Methods ...... 34

Figures ...... 38

Results ...... 67

Discussion ...... 72

4. Endoglin Regulation of Smad2 Function Mediates Beclin1 ...... 76

Expression and Endothelial Autophagy

Abstract ...... 76

Background ...... 76

Materials and Methods ...... 78

Figures ...... 83

Results ...... 102

Discussion ...... 106

5. Antibody-directed Coupling of Endoglin and MMP-14 is a key ...... 109

Mechanism for Endoglin Shedding and Deregulation of TGF-β

Signaling

Abstract ...... 109

Background ...... 109

Materials and Methods ...... 111

viii Figures ...... 115

Results ...... 137

Discussion ...... 142

6. Conclusions ...... 146

References ...... 150

LIST OF FIGURES

Figure 1.1 A schematic diagram of TGF-β signaling in endothelial cells ...... 15 1.2 A schematic diagram of endoglin structure ...... 16 1.3 A schematic illustration of the autophagy pathway ...... 17 2.1 Endoglin constitutively inhibits ERK activation ...... 23 2.2 Endoglin down-regulation of c-Myc requires β-arrestin2 interaction ...... 24 2.3 Endoglin inhibits cyclinD1 expression ...... 25 2.4 β-arrestin2 mediates endoglin regulation of endothelial cell ...... 26 growth 3.1 Endoglin tyrosine motif is an endocytic signal for ...... 38 down-regulation 3.2 Phosphomimetic mutation of the tyrosine endocytic motif ...... 41 mediates endoglin trafficking to the lysosome 3.3 Endoglin tyrosine motif regulates ALK1 trafficking to the ...... 45 lysosome 3.4 Endoglin tyrosine motif regulates TβRII trafficking to the ...... 48 lysosome 3.5 Endoglin tyrosine motif is a direct substrate for c–Src ...... 51 phosphorylation 3.6 Tyrosine endocytic motif mediates TRC105-induced endoglin ...... 57 internalization 3.7 TRC105 mediates Src-dependent endoglin tyrosine ...... 59 phosphorylation 3.8 Endoglin tyrosine phosphorylation motif is a critical ...... 62 determinant of endothelial migration 3.9 Endoglin tyrosine phosphorylation is critical for endothelial ...... 65 capillary sprouting and stability 4.1 Endoglin promotes autophagy ...... 83

4.2 ALK5 inhibition enhances autophagy ...... 86

4.3 Endoglin promotes autophagy through suppression of Smad2 ...... 89 activity

x 4.4 Inhibition of autophagic responses via TGF-β-induced Smad2 ...... 93 activation

4.5 Smad2 is a BECN-1 transcriptional repressor ...... 96

4.6 Smad2 suppression enhances endothelial capillary ...... 99 sprouting by upregulating Beclin-1 levels and LC3 cleavage

5.1 TRC105 promotes steady-state and ligand-induced Smad2/3 ...... 115 activation while inhibiting Smad1/5/8 signaling

5.2 TRC105 does not induce endothelial growth arrest or apoptosis ...... 117

5.3 Endoglin-targeting Abs enhance sEng production ...... 120

5.4 TRC105 enhances endoglin/MMP-14 association at the ...... 123 cell surface

5.5 TRC105 couples endoglin/MMP-14 into co-patched clusters at ...... 127 the cell surface

5.6 TRC105 promotes MMP-14 gene expression in HUVEC ...... 128

5.7 TRC105 and sEng induce cell migration ...... 132

5.8 TRC105 induces stress fiber formation and dissolution of ...... 135 endothelial cell–cell junctions

xi ABBREVIATIONS

Abbreviation Term

Ab antibody ALK activin receptor-like kianse ATG autophagy-related protein ATP adenosine triphosphate AVM arteriovenous malformation BAEC bovine aortic endothelial cells BMP bone morphogenetic protein CD105 endoglin DAPK death associated protein kinase EC endothelial cells EGF endothelial growth factor EGFR endothelial growth factor receptor ENG endoglin ERK extracellular signal-regulated kinase FGF fibroblast growth factor GIPC GAIP interacting protein, C terminus HHT hereditary hemorrhagic telangiectasia HMEC human microvascular endothelial cells HRE hypoxia response element HUVEC human umbilical vein endothelial cells ID1 inhibitor of DNA binding 1 JNK c-Jun N-terminal kinases KLF6 kruppel-like factor 6 LC3 microtubule-associated protein light chain 3 MAPK mitogen-activated protein kinases

xii MEEC mouse embryonic endothelial cells MIS mullerian inhibiting substance MMP-14 matrix metalloproteinase-14 PAI-I plasminogen activator inhibitor-1 PAS phagophore assembly site PDGF platelet-derived growth factor PE phosphatidylethanolamine PEDGF pigment epithelium-derived factor PI3K phosphatidylinositol-4,5-bisphosphate 3-kinase PTB phosphotyrosines binding p-Tyr phosphotyrosines RTK receptor tyrosine kinase SH2 Src homology-2 SP1 specificity protein 1 TAK transforming growth beta activated kinase TGF-β transforming growth factor beta TNF-α tumor necrosis factor alpha TRC105 TRACON105 T650A threonine at 650 position mutated to alanine TβRII transforming growth factor, beta receptor II ULK unc-51 like kinase VEGF vascular endothelial growth factor VEGFR vascular endothelial growth factor receptor VPS34 vacuolar protein sorting 34 WT wild-type YA tyrosine mutated to an alanine YF tyrosine mutated to a phenylalanine YE tyrosine mutated to a glutamate ZRP-1 zyxin-related protein-1

xiii

Chapter 1: Introduction

Angiogenesis

Angiogenesis is the physiological process in which new blood vessels form from pre-existing blood vessels [1]. During embryonic development, the first vessels formed are through vasculogenesis, the de novo synthesis of endothelial cells and primitive vascular networks.

Following vasculogenesis, angiogenesis predominates and is responsible for most of the vessel growth. Angiogenesis plays a central role in normal growth and development, wound healing, tissue repair, and reproduction [1, 2]; however, it also serves as a critical step for tumor growth and metastasis in cancer. Avascular tumors are limited in their size due to an insufficient blood supply; therefore in order for tumors to grow beyond a certain size, they must develop new blood vessels by recruiting endothelial cells from the surrounding stroma to form a functional microcirculation [3, 4]. Increased blood vessel innervation facilitates oxygen and nutrient delivery to the tumor and provides it with an efficient mechanism for waste removal, allowing it to undergo rapid growth and metastasis to other regions of the body [5]. Endothelial cells play a central role in blood-vessel growth, first by their proteolytic breakdown of the basement membrane, followed by migration, proliferation, and differentiation into capillary tubes to form a fully functioning microvessel [6]. Given that angiogenesis is a prerequisite for tumor progression and metastasis, a number of vascular targets have been explored in recent years for cancer treatment including members of the transforming growth factor β (TGF-β), vascular endothelial growth factor (VEGF), platelet-derived growth factor (PDGF), and fibroblast growth factor

(FGF) families as well as angiostatic agents including angiostatin, endostatin, and pigment 1 epithelium-derived factor (PEDF). Currently, VEGF inhibition is the most well established approach for reducing tumor angiogenesis. Anti-angiogenic drugs that target the VEGF pathway include bevacizumab, sunitinib, pazopanib, axitinib, and sorafenib [7]. Bevacizumab, a known

VEGF inhibitor, binds to VEGF and prevents VEGF from binding to a VEGF receptor (VEGFR), whereas all the other aforementioned inhibitors act as tyrosine kinase inhibitors that block VEGF- induced cellular signaling. Bevacizumab treatment in conjunction to chemotherapy treatment prolonged overall survival for patients suffering from advanced colorectal cancer and non-small- cell lung cancer [8]. In addition, VEGFR tyrosine kinase inhibitor treatments also showed a similar increase in survival in patients suffering from renal cell cancer, hepatocellular cancer, neuroendocrine cancers, and colorectal cancer [8]. Despite this success, the majority of patients treated with either VEGF or VEGFR inhibitors exhibited modest to no effect in the majority of patients. Patients that initially responded well to treatment quickly acquired resistance and suffered from extensive side effects including endothelial dysfunction in healthy tissues, thrombotic events, hypertension, and renal complications [9]. As demonstrated by VEGF inhibition, the benefits of anti-angiogenic therapies are typically minimal and treatment often yields extensive side effects, as blood vessel growth is required for both tumor progression and for normal physiological processes [9, 10]. Therefore, understanding the cellular events governing angiogenesis and how these events are regulated at the molecular level is critical in identifying novel tumor markers that can be used to target tumor angiogenesis for cancer therapy.

Transforming Growth Factor β Signaling

In addition to VEGF, the TGF-β superfamily plays a prominent role in regulating angiogenesis.

The TGF-β superfamily is comprised of over 30 members including TGF-β (TGF-βs 1-3), activins (A, B), inhibins (A, B), Nodals, and some Growth and Differentiation Factors (GDFs),

Bone Morphogenic Proteins (BMPs 1-20), and Mullerian Inhibiting Substance (MIS) [11]. The

2 TGF-β superfamily plays an important role during development as well as in normal and pathophysiological processes by regulating a wide range of cellular and biological processes including proliferation, migration, adhesion, apoptosis, and angiogenesis [12]. In the vasculature,

TGF-β signaling is essential for angiogenesis and vascular homeostasis. In vivo studies in mice demonstrate that targeted deletions of critical TGF-β signaling components such as TGF-β receptors, activin receptor-like kinase (ALK5) and TGF-β receptor II (TβRII), as well as the

TGF-β ligand are embryonic lethal as mice exhibit various cardiovascular defects including impaired formation of the vascular plexus, defective differentiation of the capillary endothelium, and presence of dilated, leaky vessels [13]. In addition, mutations in the TGF-β signaling receptors endoglin and ALK1 in humans result in the autosomal dominant disorder, hereditary hemorrhagic telangiectasia, characterized by arteriovenous malformations (AVM) and numerous hemorrhagic events in multiple tissues and organs including the lung, brain, and the gastrointestinal tract [14-16].

TGF-β family members bind to TGF-β type I (activin like kinase (ALK) 1-7) and type II serine/threonine kinase receptors [17]. Ligand binding to the extracellular domain of the constitutively active TGF-β type II receptor (TβRII) recruits TGFβ type I receptor to form a heteromeric receptor complex consisting of two TβRII and two TGFβ type I receptors [18]. TβRII phosphorylates the TGF-β type I receptor’s Gly-Ser regulatory region (GS domain), resulting in the activation of its serine/threonine kinase activity [19]. The activated type I receptor recruits and phosphorylates a family of nuclear shuttling proteins known as Smads. Smads are divided into three different classes: receptor mediated Smads (R-Smads), which include Smads 1, 2, 3, 5, and 8; common-mediator Smad (co-Smad) which includes Smad4; and the inhibitory Smads (I-

Smads) which includes Smad 6 and 7 [20]. Upon phosphorylation, R-Smads bind to Smad4 and translocate to the nucleus to regulate the expression of TGF-β-targeted genes.

TGF-β has been reported to both stimulate and inhibit endothelial cell migration and proliferation 3 [21]. This dichotomous role of TGF-β is achieved by balancing two opposing TGF-β signaling pathways: ALK1 mediated Smad 1/5/8 activation and ALK5 mediated Smad 2/3 activation

(Figure I-1). In endothelial cells, TGF-β and BMP-9/10 signal through ALK1 and Smad 1/5/8 to promote pro-angiogeneic responses including enhanced endothelial cell proliferation and migration [22-24]. Conversely, TGF-β can also signal through ALK5 and Smad 2/3 to inhibit endothelial cell growth and migration, yielding an overall anti-angiogenic response (Figure I-1)

[25, 26].

The pro- and anti-angiogenic responses mediated by Smad 1/5/8 and Smad 2/3 are attributed to their ability to induce and repress the expression of different genes. TGF-β induces plasminogen activator inhibitor-1 (PAI-1) gene expression through Smad 2/3 activation [27]. PAI-1 is a serine protease inhibitor that exerts both pro and antiangiogenic properties. The proangiogenic effects of

PAI-1 are mediated through the inhibition of proteolysis and stabilization of the extracellular matrix. The antiangiogenic effects of PAI-1 are mediated by its interaction with vitronectin, an extracellular matrix protein that promotes cell adhesion and migration through interactions with integrins and the uPA receptor [28]. By interacting with vibronectin, PAI-1 interferes with vitronectin binding to integrin and the uPA receptor, thereby blocking cell adhesion and migration, resulting in the inhibition of angiogenesis. TGF-β can also exert its anti-angiogenic effects through Smad 2/3-mediated inhibition of c-Myc expression [20]. c-Myc is essential for normal and tumor-associated angiogenesis and is activated by a number of different mitogenic signals and growth pathways including Wnt and MAPK/ERK pathways. In addition, c-Myc promotes numerous biological processes including cell proliferation, cell growth, and cell survival. Deletion of c-Myc is embryonic lethal as c-myc null mice die by embryonic day 10.5 due to improper blood vessel formation and defects in cardiac and neuronal development [29].

Additionally, loss of c-Myc function dramatically impaired tumor formation in immune- compromised mice [29]. BMP-9-stimulated Smad1/5/8 activation induces the expression of Id1, 4 Id2, and Id3 to inhibit proliferation of endothelial cells. Id proteins play a critical role in regulating a variety of different processes including cell growth, differentiation, and angiogenesis.

Loss of Id1 or Id3 results in defective vascularization and angiogenesis, extensive necrosis, and embryonic lethality. Id1 promotes angiogenesis by repressing the transcription of the naturally occurring inhibitor of angiogenesis, thrombospondin-1 (TPS-1) [30].

Non-Canonical Transforming Growth Factor β Signaling

The dichotomous cellular effects of TGF-β can be partially attributed to the growing evidence that TGF-β can signal through Smad-independent pathways. These Smad-independent pathways include the ERK, JNK, and p38 mitogen activated kinase (MAPK) pathways; Rho-like GTPase signaling pathways; and phosphatidylinositol-3-kianse (PI3K)/AKT pathways. The MAPK pathways consist of a highly conserved family of serine/threonine protein kinases that are critical in regulating fundamental cellular processes including proliferation, migration, and apoptosis

[31].

ERK is perhaps the best-characterized MAPK, and activation of the Raf-MEK-ERK signal transduction pathway in endothelial cells is required for angiogenesis [32]. Mutations in genes encoding proteins in the Raf-MEK-ERK pathway that lead to aberrant ERK activation are often found in many human cancers including pancreatic, colorectal, and lung adenocarcinomas. ERK activation is primarily mediated by receptor tyrosine kinases (RTK). Binding of growth factors to

RTK induces dimerization, trans-autophosphorylation of intracellular tyrosine residues, and activation of the RTK. Once phosphorylated, the tyrosine residues serve as docking sites for numerous signaling molecules containing Src homology 2 (SH2) or phospho-tyrosine binding

(PTB) domains such as Grb2 and Shc that lead to ERK activation [33]. Activated ERK phosphorylates numerous transcription factors including Ets-1, c-Jun, and c-Myc, resulting in increased cell proliferation and cell growth [34].

5 TGF-β activation of ERK occurs in a similar manner. TGF-β binding to TGFβRII and ALK5 facilitates tyrosine phosphorylation of both receptors, leading to the recruitment and direct phosphorylation of ShcA on tyrosine and serine residues and formation of the ShcA/Grb2/Sos complex. The ShcA/Grb/Sos complex activates Ras at the plasma membrane, leading to activation of c-Raf, MEK, and ERK [33]. TGFβRII can also undergo autophosphorylation on three tyrosine residues as well as undergo tyrosine phosphorylation by the non-RTK, Src, independently of TGF-β, thereby providing additional mechanisms linking TGF-β and ERK signaling [35]. Although ERK promotes cell proliferation by activating transcription factors such as c-Myc, it can also promote cell proliferation by phosphorylating R-Smads 1, 2, and 3 [36, 37].

Phosphorylation of Smads by ERK inhibits Smad activity and relieves TGF-β-mediated growth arrest.

Endoglin

The dichotomous cellular effects of TGF-β family ligands are balanced by the endothelial- specific co-receptor, endoglin. Endoglin (CD105) is TGF-β superfamily co-receptor that is expressed predominantly in proliferating vascular endothelial cells and plays an important role in regulating angiogenesis and vascular maintenance. A critical role for endoglin in the vasculature is supported by several in vivo findings, including embryonic lethality observed in endoglin knockout mice due to impaired angiogenesis and cardiovascular development of the yolk sac during embryonic development and the presence of endoglin mutations in human patients suffering from vascular diseases including hemorrhagic telangiectasia (HHT1) [38, 39]. Endoglin dysfunction also contributes to a number of other conditions including tumor-associated angiogenesis, inflammation, and preeclampsia [40]. Despite these advances in endoglin biology, the fundamental mechanisms of endoglin during embryonic development and vascular disease progression remain unclear.

6 Endoglin consists of a large extracellular ligand-binding domain, a hydrophobic transmembrane domain, followed by a short/serine/threonine rich cytoplasmic domain [41]. Endoglin exists as a disulfide-linked homodimer and serves as a high-affinity cell surface receptor for TGF-β1, TGF-

β3, and BMPs-2 and -7 when associated with ALKs 1-7. More recently, endoglin and ALK1 have also been shown to directly bind to BMPs-9 and -10 (Figure I-2) [42]. Unlike its counterparts

ALK1 and TβRII, endoglin’s lacks kinase activity in its cytoplasmic domain and does not directly participate in Smad phosphorylation; rather, it functions as an important phosphorylation target for ALK1, ALK5, and TβRII. TβRII and ALK5 phosphorylation of endoglin’s cytoplasmic domain facilitates the release of TβRI but not TβRII from the TGF-β receptor complex, thereby altering Smad 1/5/8 and Smad 2/3 response to TGF-β and BMP9 (Figure I-2) [43]. Furthermore,

ALK5 has been demonstrated to phosphorylate endoglin’s cytoplasmic domain at serine 646 and

649 [43]. Phosphorylation at serine 646 results in enhanced Smad 1/5/8 signaling in response to

TGF-β whereas phosphorylation at both serines 646 and 649 results in enhanced Smad 1/58 signaling in response to BMP-9, suggesting that endoglin’s cytoplasmic domain is important for regulating TGF-β superfamily signaling [43]. Aside from the membrane-bound form, a soluble form of endoglin also exists. Soluble endoglin is strongly elevated in women suffering from preeclampsia and in patients with various solid malignancies including colorectal and breast cancer [44]. Soluble endoglin is produced through the proteolytic cleavage of membrane-bound endoglin by the membrane-type 1 matrix metalloproteinase 14 (MMP-14) and has been shown to antagonize angiogenesis through several mechanisms including receptor down-regulation and by acting as a TGF-β family ligand trap [45].

Endoglin plays a critical role in modulating TGF-β-dependent activation of ALK1 versus ALK5 by regulating their access to TGF-β ligands. Endoglin is required for TGF-β/ALK1 signaling and indirectly attenuates TGF-β/ALK5 signaling by competing with ALK5 for TGF-β [46]. While endoglin mediates the balance of these opposing pathways, much of the current molecular basis

7 for its effects on signaling and vascular biology remains unknown, due to the conflicting reports on the exact roles for endoglin and ALK1. Although the predominant notion suggests that endoglin/ALK1-induced Smad 1/5/8 signaling yields pro-angiogenic responses, there is contradicting in vitro and in vivo evidence suggesting that endoglin and ALK1-induced Smad

1/5/8 signaling results in the inhibition of endothelial cell proliferation and migration [40]. While endoglin is clearly essential for angiogenesis, precisely how endoglin modulates the balance of pro- and anti-angiogenic signaling by TGF-β family ligands remains an active area of investigation.

It is becoming increasingly evident that endoglin alters endothelial cell behavior by signaling through several Smad-independent pathways, including the MAPKs (ERK, JNK, and p38) and

Phosphatidylinositol-3 Kinase (PI3K)/Akt pathways via scaffolding/trafficking adaptor molecules including β-arrestin2, zyxin, zyxin-related protein-1 (ZRP-1), and GAIP interacting protein, C terminus (GIPC). Endoglin interacts with these adaptor molecules through its cytoplasmic domain and these interactions alter both Smad-dependent and Smad-independent signaling. Endoglin, through its interaction with zyxin and ZRP-1, regulates focal adhesions and actin dynamics to inhibit endothelial cell migration. GIPC promotes cell surface retention of endoglin to enhance

ALK-1-induced Smad 1/5/8 activation and facilitates interaction between endoglin and the PI3K subunits, p110α and p85, to recruit and activate PI3K and Akt at the cell membrane.

Endoglin/GIPC-mediated activation of Akt is required for proper endothelial cell function during angiogenesis. In the case of β-arrestin2, binding to endoglin causes internalization of this complex into endocytic vesicles, resulting in the transient suppression of TGF-β-induced ERK activation and endothelial cell migration.

The importance of endoglin in physiological and pathophysiological states has prompted the investigation of factors involved in regulating endoglin gene and protein expression in endothelial cells. Basal endoglin transcription is sustained by specificity protein 1 (Sp1), kruppel-like factor 6

8 (KLF6), and TGF-β. Analysis of the endoglin gene demonstrates that endoglin has a hypoxia- response element (HRE) downstream of the main transcription start site. Hypoxia-induced factor

(HIF-1) recognizes and binds to the HRE to drive endoglin’s promoter activity, transcription, and expression [47]. TGF-β stimulation in conjunction with hypoxic conditions results in the formation of a transcriptional complex consisting of Smad3, Sp1, and HIF-1, leading to a synergistic enhancement of endoglin expression. Endoglin gene and protein expression have also been demonstrated to be upregulated in human umbilical vein endothelial cells (HUVECs) expressing constitutively active ALK1 [48]. TNF-α downregulated endoglin protein levels but had no effect on endoglin transcript levels or promoter activity in endothelial cells [49]. The mechanism underlying TNF-α-mediated downregulation of endoglin has not yet been defined.

While endoglin’s transcriptional regulation is well established, post-translational mechanisms that modulate endoglin expression are virtually unexplored.

Endoglin: Therapeutic Potential

Given that endoglin is essential for angiogenesis and its expression pattern is limited to endothelial cells undergoing active angiogenesis, endoglin serves as an ideal prognostic, diagnostic and therapeutic vascular target in human cancers. Endoglin is highly overexpressed on the plasma membrane of proliferating endothelial cells in a number of solid tumors including lung, prostate, colorectal, pancreatic, and breast tumors [8]. Comparison of between plasma endoglin levels in patients with colorectal cancer and healthy patients demonstrated that patients with colorectal cancer exhibited fourfold higher levels of endoglin. In addition, a positive correlation between plasma endoglin levels and colorectal tumor stage progression was observed in these patients [8]. Microvessel density is a surrogate marker that reflects tumor angiogenesis and has been regarded as a prognostic maker for numerous tumors. In patients with solid tumors, endoglin expression is directly correlated with tumor microvessel density and is associated with

9 poor prognosis [50]. The traditional concept of drug resistance suggests that tumors develop resistance to angiogenic inhibitors after repeated exposure by acquiring mutations in the gene encoding for the drug target or through changes in the influx or efflux of the drug [51]. However, growing evidence suggests that alternate mechanisms may also play a pivotal role in anti- angiogenic therapy resistance. One such mechanism is the compensatory activation or upregulation of pro-angiogenic signaling pathways within the tumors. Endoglin and TGF-β expressions are upregulated in tumor endothelial cells upon inhibition of the VEGF pathway. In addition, endoglin expression is significantly upregulated in human pancreatic cancer xenografts isolated from mice treated with a VEGF neutralizing antibody [8], suggesting that neutralizing endoglin function may circumvent VEGF resistance and effectively inhibit tumor angiogenesis in patients.

In vitro studies have demonstrated that monoclonal antibody treatment against endoglin significantly inhibited proliferation of human microvascular endothelial cells [52]. In vivo studies using severe combined immunodeficient mice harboring human breast tumors demonstrated that systemic administration of radiolabeled or immunotoxin-conjugated monoclonal antibody against endoglin resulted in suppression of tumor growth. In addition, combining traditional chemotherapeutic agents with monoclonal antibody treatment had a synergistic effect in reducing tumor growth and metastasis [52]. In vitro and in vivo studies support the notion that endoglin monoclonal antibodies can specifically localize to the endothelium of the tumor vasculature and effectively inhibit tumor angiogenesis and metastasis with little reactivity to the vasculature of normal tissues. Given endoglin’s promise as a promising therapeutic strategy in limiting tumor vascularization, a growing number of endoglin monoclonal antibodies are under evaluation for clinical applications [53]. TRC105 is a humanized monoclonal antibody that binds to endoglin with high-affinity and neutralizes its function by reducing BMP9-induced Smad 1/5/8 activation and enhancing apoptosis in endothelial cells [42]. In mice, TRC105 treatment results in inhibition

10 of tumor progression and angiogenesis. In humans, TRC105 is currently in phase 2 clinical trials.

As a single agent, TRC105 has shown clinical promise in reducing tumor angiogenesis in patients with multiple solid tumor types such as soft tissue sarcoma, advanced renal cell carcinoma, hepatocellular carcinoma, and metastatic breast cancer. TRC105 is also in clinical trials for its therapeutic potential in combination with VEGF inhibitors such as bevacizumab. Recent phase 1 clinical trials demonstrated that TRC105 combined with bevacizumab exhibited marginal toxicities and significantly reduced tumor volume and size in bevacizumab-resistant patients

[54, 55]. Despite the initial observation that TRC105 reduces BMP9 binding to endoglin, additional underlying molecular mechanisms of TRC105 remain poorly defined.

Autophagy

Autophagy is an evolutionary conserved process in which the cell self-digests its proteins and organelles through the lysosomal degradative pathway. Autophagy plays a critical role in both physiological and pathophysiological processes. Under physiological conditions, autophagy maintains cell homeostasis by facilitating the removal of superfluous and damaged organelles that would otherwise become toxic to the cell. During metabolic stress and nutrient deprivation, autophagy promotes cell survival by recycling long-lived organelles and proteins to help produce

ATP and maintain macromolecular synthesis. Autophagy has recently been implicated in cytoprotection through its positive role in immune response, including the removal of invasive microbes and antigen production [56]. Despite its beneficial role in normal physiology, autophagy is associated with tumor progression, metastasis, and therapeutic resistance. Tumor cells experience higher metabolic demands and stress compared to normal cells due to defects in angiogenesis and inadequate blood supply. Metabolic stress in tumors can also be attribute to their unnaturally high proliferative and migratory nature as well as their increased glycolytic metabolism. To circumvent these abnormalities, tumor cells activate autophagy. Autophagy promotes tumor survival by acting as an alternate source of energy and by bypassing normal

11 apoptotic signals. The first evidence of autophagy’s pro-tumorigenic role came from the observation that induction of autophagy in the hypoxic core of the tumor promoted cancer cell survival [57]. Moreover, genetic inactivation of autophagy through constitutive activation of the

PI3K pathway or allelic loss of key autophagic proteins such as Beclin-1 and autophagy related protein 5 (ATG5) reduces tumor cell survival even in response to metabolic deprivation [58].

Autophagy begins with the formation of the phagophore assembly site (PAS) and the isolation of a double membrane known as the phagophore (Figure I-3). The phagophore assembly is mediated by a complex of proteins including UNC51-like kinase-1, -2 (ULK1, ULK2), ATG13, and the class III PI3K complex consisting of the vacuolar protein sorting 34 (VP34) PI3K and Beclin-1

[59]. The phagophore expands and sequesters targeted cargo such as long-lived organelles or damaged proteins in a double-membrane autophagosome. Phagophore elongation requires two ubiquitin-like conjugation pathways [60]. The first involves the conjugation of ATG5 to ATG12.

The second involves the conjugation of phosphatidylethanolamine (PE) to microtubule-associated protein 1 light chain 3 (LC3) [61]. Following elongation, the autophagosome matures and fuses with the lysosome through LC3-PE recognition and releases its cargo into the lysosome. The autophagosomal contents are degraded by lysosomal proteases, and the resulting constituents are exported out to the cytoplasm where they can be used for metabolism or for macromolecule synthesis (Figure I-3) [59].

Autophagy can be initiated by TGF-β. In human hepatocellular carcinoma cells and mammary carcinoma cells, TGF-β stimulation induced autophagosome formation, lipidation of LC3-I to

LC3-II, and increased gene expression of several autophagy related genes including BECN1,

ATG5, and ATG7 [62]. TGF-β-induced autophagy is suppressed upon knockdown or inhibition of Smad signaling components such as Smad 2/3 and Smad 4, suggesting that TGF-β induces autophagy, at least in part, through the Smad signaling [63]. In addition, knockdown or inhibition of JNK and death associated protein kinase (DAPK) also suppressed TGF-β-induced autophagy,

12 indicating the involvement of both Smad-dependent and Smad-independent pathways. Despite this, the relationship between TGF-β signaling and autophagy remains poorly understood.

Autophagy is a critical mediator of angiogenesis in endothelial cells. Whether autophagy promotes or inhibits angiogenesis is still unclear. AKT3 deficient mice exhibited impaired angiogenesis and increased autophagy compared to control mice, suggesting a negative correlation between autophagy and angiogenesis [64]. However, endothelial-specific deletion of

ATG5 in tumors resulted in smaller and more tortuous capillary formation, suggesting that autophagy is necessary for proper angiogenesis. In addition, induction of autophagy in bovine aortic endothelial cells (BAECs) promoted angiogenesis while inhibition of autophagy suppressed angiogenesis, including VEGF-driven angiogenesis [65]. Elucidating the exact role of autophagy in angiogenesis will further advance our understanding of the underlying mechanisms of cancer progression and chemotherapy resistance as well as provide novel targets for anticancer therapy.

Given that angiogenesis plays a critical role in the growth and spread of tumors, it has become an attractive target for cancer therapy. Current anti-angiogenic therapies have not lived up to their potential, suggesting the need for new anti-angiogenic targets. Endoglin is a novel target for anti- angiogenic therapy because of its essential role in angiogenesis and specific expression pattern.

However, how endoglin promotes angiogenesis is not well defined. Therefore, the specific aims of this thesis are based on the fundamental goal of defining novel signaling mechanisms in which endoglin engages to promote angiogenesis. In the course of completing my thesis, various biochemical and immunofluorescence studies were applied to identify novel functions for endoglin.

The major aims of the thesis research are given below.

Thesis aims:

1. Defining a novel mechanism by which endoglin regulates endothelial growth.

2. Identification of an endogenous mechanism of endoglin downregulation 13 3. Identification of the role of endoglin in autophagy.

4. Characterization of the underlying mechanism by which TRC105 and other endoglin-

targeting antibodies inhibit angiogenesis

Figure 1.1: Schematic Diagram of TGF-β Signaling in Endothelial Cells

TGF-β1 and BMP9 binding to the heteromeric complex consisting of endoglin and ALK1 facilitates phosphorylation and activation of the transcription factors, Smad1/5/8. Alternatively,

TGF-β1 binding to the heteromeric complex consisting of endoglin and ALK5 facilitates phosphorylation of the transcription factors Smad 2/3. Phosphorylated Smad 1/5/8 and Smad 2/3 translocate into the nucleus to drive expression of a variety of TGF-β targeted genes, resulting in pro-angiogenic and anti-angiogenic responses, respectively.

Figure 1.2: A Schematic of Endoglin Structure

Endoglin consists of a large extracellular domain that mediates binding to TGF-β ligands (TGF-

β1 and BMP-9/10), a hydrophobic transmembrane domain, followed by a short serine/threonine rich intracellular domain that serves as a key docking site for an number of adaptor proteins including β-arrestin2 and GIPC.

Figure 1.3: Schematic Illustration of the Autophagy Pathway

Autophagy is initiated by metabolic stress. Upon initiation, pre-autophagosomal structures form within the cytoplasm. Autophagy related proteins (ATG5 and ATG12) are recruited and facilitate the elongation of the pre-autophagosomal structure. The elongated membrane wraps around the targeted cytosolic cargo in a double-membrane autophagosome. The mature autophagosome fuses with the lysosome and releases its cargo. The cargo is degraded by lysosomal hydrolases and proteases. The resulting constituents can be re-used for macromolecule synthesis or ATP generation.

Chapter 2: Endoglin Inhibits ERK-induced c-Myc and Cyclin D1 Expression to Impede Endothelial Cell Proliferation

Abstract

The objective of this thesis aim was to define a novel mechanism by which endoglin promotes angiogenesis. Endothelial cell proliferation is a key step in angiogenesis and has a pivotal role in physiological and pathological processes including wound healing and cancer development.

Previous studies demonstrated that endoglin interacts with the scaffolding protein, β-arrestin2. β- arrestin2 promotes endoglin internalization and attenuates ERK-mediated cell migration. Given that ERK also regulates a number of other cellular processes, we investigated the role of endoglin and β-arrestin2 in endothelial cell proliferation. We demonstrate that endoglin impedes cell growth in a TGF-β-independent manner by inhibiting ERK-induced c-Myc and cyclin-D1 expression. However, this was reversed upon disruption of the endoglin/β-arrestin2 interaction.

This study defines a novel mechanism by which endoglin augments endothelial cell growth by targeting ERK and its downstream mitogenic substrates.

19 Background

Endoglin is a critical mediator of angiogenesis. It is mutated in the autosomal dominant disorder,

HTT1 and its dysfunction contributes to a number of pathophysiological processes including preeclampsia, inflammation, and tumor-associated angiogenesis. Angiogenesis is comprised of a highly coordinated and regulated series of steps that is initiated by activation of endothelial cells by angiogenic factors, followed by degradation of capillary basement membrane, endothelial cell migration, proliferation, stabilization, and capillary vessel formation and maturation. Endoglin is essential for angiogenesis as endoglin null mice die by gestational day 11.5 due to vascular defects and improper blood vessel development. While its pro-angiogenic role is well defined, how endoglin promotes angiogenesis at the cellular and molecular level is largely undefined.

In endothelial cells, TGF-β signaling occurs through serine/threonine kinase receptors, ALK1 and

ALK5, which activate downstream nuclear shuttling transcription factors, Smad 1/5/8 and Smad

2/3, respectively. ALK5-induced Smad 2/3 activation inhibits endothelial cell proliferation and migration whereas ALK1-induced Smad 1/5/8 activation results in enhanced pro-angiogenic responses including increased endothelial cell proliferation and migration. These dichotomous effects of TGF-β are balanced by endoglin. It is becoming more evident that endoglin, through its interaction with various scaffolding and trafficking proteins, signals through Smad-independent mechanisms to regulate angiogenesis. However, how endoglin mediates both Smad-dependent and Smad-independent signaling is still unclear.

Endoglin is comprised of a large extracellular ligand binding domain, a transmembrane domain, and a short serine/threonine-rich intracellular domain. Unlike its TGF-β signaling counterparts

ALK1 and ALK5, endoglin lacks kinase activity. The intracellular domain functions as an important phosphorylation and docking site for multiple scaffolding and trafficking proteins including β-arrestin2, Zyxin, ZRP-1, and GIPC. In the case of β-arrestin2, binding to endoglin

20 promotes internalization of endoglin, transient suppression of ERK activation, and enhanced cell migration. Given that ERK regulates a number of different substrates to induce changes in cellular functions and processes, we examined whether endoglin and β-arrestin2 influence ERK signaling during cell proliferation.

Materials and Methods

Cell Culture, Plasmids, Transfections and Antibodies (Abs) Mouse embryonic endothelial cells

(MEECs) (Eng +/+ and Eng -/-) were derived from wild type and endoglin knockout mice at E9 as previously described [38]. ECs were maintained in MCDB-131 medium (Invitrogen) supplemented with 2 mM L-glutamine, 1 mM sodium pyruvate, 15% fetal bovine serum, 100

μg/ml heparin, and 25 μg/ml endothelial cell growth supplement. HMEC-1 was maintained in

MCDB-131 medium supplemented with 10% fetal bovine serum, 1 μg/ml hydrocortisone

(Sigma), 10 ng/ml epidermal growth factor (Sigma), and 2 mM l-glutamine Ectopic expression in

Eng-/- MEECs was achieved by nucleofection with Amaxa (SF kit). HMEC-1s were nucleofected with pLKO.1 puro vector encoding either scramble control or shRNA to human endoglin sequence (Sigma). The pDisplay expression vectors encoding endoglin wild type (WT) and

T650A were described previously [38]. TGF-β1 was obtained from R&D Systems. Protease and phosphatase inhibitor cocktails were purchased from Sigma. MTT was purchased from Sigma. β- actin was obtained from Sigma. P4A4 antibody was purchased from University of Iowa

Hybridoma. The following Abs were obtained from Cell Signaling: total ERK (no. 9102), phospho-ERK (no. 4370), c-Myc (no. 13987), and cyclin D1 (no. 2978).

Biochemical analysis MEECs were serum starved for 4-5 h. in MCDB-131 media supplemented with L glutamine. Following serum starvation, cells were treated with TGF-β with the appropriate dose and the indicated amount of time and lysed. Cells lysates were normalized for equal protein loading through Bio-rad assay using lysis buffer (20 mM HEPES pH 7.4, 150 mM NaCl, 10 mM

21 NaF, 1 mM EDTA, supplemented with protease and phosphatase inhibitor cocktails).

MTT assay MEECs were plated at 100,000 cells and grown overnight in 6 well plates. Cells were treated with MTT (Thiazolyl Blue Tetrazolium Bromide) in serum free media (MCDB-131 supplemented with L-glutamine) for 2 h at various time points (12, 24, 48 h). Cells were briefly washed with PBS, then the MTT metabolic product (formazan) was extracted in DMSO and the optical density was read at 570 nm in a microplate reader.

Figure 2.1: Endoglin constitutively inhibits ERK activation. (A) Eng +/+ and Eng -/- MEECs were serums starved for 4 h. prior to TGF-β dose response at the indicated concentrations for 15 minutes. Phospho-specific (upper panel) and total ERK (lower panel) immunoblotting are shown.

(B) A time course experiment of Eng +/+ and Eng -/- MEECs treated with 20 pM TGF-β for the indicated time points.

Figure 2.2: (B) Comparison of c-Myc expression in HMEC-1 in non-targeting control (NTC) versus endoglin knockdown (sheng) after 48 h. (C) c-Myc expression of Eng +/+ and Eng -/- ECs treated with 20 pM TGF-β for the indicated amount of time. (D) Comparison of ERK activation and c-Myc levels in Eng -/- MEECs expressing no endoglin (control), WT endoglin, or endoglin point mutant that is unable to bind to β-arrestin2 (T650A).

Figure 2.3: Endoglin inhibits cyclinD1 expression. (A) Comparison of cyclin-D1 expression in

Eng +/+ and Eng -/- MEECs upon TGF-β treatment for the indicated amount of time (0-12 h.).

(B) Comparison of cyclin-D1 levels in Eng -/- MEECs expressing no endoglin (control), WT endoglin, or endoglin mutant T650A.

Figure 2.4: β-arrestin2 mediates endoglin regulation of endothelial cell growth. (A) MTT growth assay of Eng +/+ and Eng -/- MEECs for the indicated time points (0, 12, 24, 48 h.). (B) MTT growth assay of Eng -/- MEECs expressing no endoglin (control), WT endoglin, or endoglin mutant T650A for the indicated time points (0, 12, 24, 48 h.).

Results

TGF-β is a potent inhibitor of proliferation in most cell types, including in endothelial cells [66].

However, the role of endoglin in endothelial cell proliferation is unclear. Various in vivo and in vitro studies have yielded conflicting results. On one hand, it has been reported that endoglin and

ALK1 enhance endothelial cell proliferation through TGF-β-induced Smad1/5/8 signaling [67]

[68]. On the other hand, others have reported the opposite results [69] [70]. In order to elucidate the endoglin’s role in endothelial proliferation, we wanted to determine whether endoglin engages in Smad-independent mechanisms to mediate cell proliferation, particularly the role of endoglin in regulating ERK-mediated mitogenic signaling.

It was previously demonstrated that endoglin rapidly and transiently impairs ERK activation in response to low doses of TGF-β through its interaction with β-arrestin2 [38]. To further elucidate the role of endoglin in ERK signaling, we performed a TGF-β dose response comparing ERK activation in endoglin wild-type (Eng +/+) and endoglin null (Eng -/-) mouse embryonic endothelial cells (MEECs). Endoglin +/+ cells displayed significantly lower basal ERK activation compared to Endoglin -/- cells. In addition, little to no response was observed upon TGF-β treatment (Figure II-1A). To determine whether the observed suppression of ERK activation between was long term, we performed a time course comparing ERK activation in Eng +/+ and

Eng -/- cells over the course of 12 h. of TGF-β treatment. ERK activation was constantly suppressed in Eng +/+ cells despite TGF-β treatment. In addition, there were significantly higher levels of ERK activation in Eng -/- cells compared to Eng +/+ cells over the duration of the time course (Figure II-1B). Given that ERK regulates cellular proliferation and that sustained ERK activation in Eng -/- cells could potentially contribute to cellular proliferation, we examined c-

Myc expression as a downstream mitogenic target of ERK signaling. Basally, c-Myc was higher

27 in Eng -/- cells compared to Eng +/+ cells (Figure II-2A). To demonstrate that the difference in c-

Myc expression was an endoglin-specific effect, we knocked down endoglin in human microvascular endothelial cells (HMEC-1). Knockdown of endoglin resulted in an increase in c-

Myc expression (Figure. II-2B). Next, we observed c-Myc expression in response to TGF-β treatment. Consistent with ERK activation, Eng -/- cells exhibited higher levels of c-Myc expression compared to Eng +/+ cells over the course of 24 h. TGF-β treatment (Figure II-2C).

Overall, our data suggests that endoglin constitutively down-regulates c-Myc.

Next, we tested whether endoglin requires β-arrestin2 to inhibit c-Myc expression. To do this, we rescued endoglin wild type (WT) or the endoglin point mutant that is unable to bind to β-arrestin2

(T650A) in Eng -/- cells. Endoglin WT significantly reduced c-Myc expression levels whereas endoglin T650A had no effect relative to the control (Figure II-2D), suggesting that endoglin’s suppression of ERK activation is β-arrestin2-dependent. Furthermore, ERK activity closely mimicked c-Myc expression as endoglin WT rescue significantly reduced c-Myc expression compared to control whereas rescue with endoglin T650A had no effect (Figure II-2D).

Another downstream target of ERK signaling is cyclin D1. Like c-Myc, cyclin D1 is an oncogene that promotes cell proliferation and cell cycle progression. To determine whether endoglin inhibition of ERK activation also affects cyclin-D1, we performed a TGF-β time course and examined cyclin D1 expression levels. Basal cyclinD1 levels were reduced in Eng +/+ cells compared to Eng -/- cells and TGF-β treatment had little impact over the course of 12 h. (Figure

II-3A). To test whether β-arrestin2 mediates this effect, endoglin WT and endoglin T650A expression were restored in Eng -/- cells. Similar to ERK activity and c-Myc expression, endoglin

WT reduced cyclin-D1 expression whereas endoglin T650A had no effect compared to control

(Figure II-3B), further demonstrating the importance of endoglin/β-arrestin2 interaction in down- regulating downstream targets of ERK.

28 Based on our data that endoglin inhibits ERK activation and the expression of its downstream targets, we examined whether this mechanism contributes to cell growth inhibition. To test this, we performed a MTT assay comparing Eng +/+ and Eng -/- cells. Eng -/- cells proliferated about

30-40% faster than Eng +/+ cells at the 48 h. time point (Figure II-4A). Rescue of endoglin WT in Eng -/- cells attenuated cell growth by approximately 55% compared to control (Figure II-4B).

However, endoglin T650A rescue did not attenuate cell growth as effectively as endoglin WT, suggesting that endoglin inhibits ERK-mediated endothelial cell growth largely through its interaction with β-arrestin2.

Discussion

Endoglin’s pro-angiogenic role in the vasculature is well established. However, how it regulates angiogenesis at the molecular and cellular level is largely unknown. Various in vitro and in vivo studies have yielded conflicting results. Some have demonstrated that endoglin and ALK1 elicit pro-proliferative and pro-migratory responses through activation of Smad 1/5/8 [67, 71]. Others, however, have shown that endoglin and ALK1 signal through Smad 1/5/8 to inhibit endothelial proliferation and migration [69, 70]. Endoglin and other TGF-β superfamily receptors can engage in Smad-independent pathways including the PI3K/Akt and MAPK pathways to regulate angiogenesis. The ability for endoglin to signal through multiple Smad-independent pathways may contribute to these disparate findings.

Our current study is an extension of our previous work, which identified β-arrestin2 as a novel binding partner for endoglin [38]. We demonstrated that endoglin/β-arrestin2 interaction transiently reduced ERK activation and inhibited endothelial migration. In this study, we examined how endoglin/ β-arrestin2 interaction may regulate downstream targets of ERK that are essential for proliferation including c-Myc and cyclinD1. Our results clearly show that endoglin inhibits ERK activation as shown by the difference in ERK activation between Eng +/+ and Eng -

29 /- MEECs. In addition, this inhibition was largely TGF-β-independent as indicated by our dose- response and time-course data. Interestingly, TGF-β is a potent repressor of c-Myc largely through binding of Smad 3 to the TGF-β inhibitory element on the c-Myc promoter [20].

However, endoglin’s role in modulating c-Myc expression has not been characterized. Endoglin’s inhibition of ERK was predominantly dependent on β-arrestin2, as the endoglin point mutant that does not bind to β-arrestin2 (T650A) does not inhibit ERK activation or down-regulate c-Myc and cyclin D1 expression. Its been previously shown that there is significant crosstalk between

Smad-independent and Smad-dependent pathways. Interestingly, ERK can phosphorylate the linker regions of Smad 2 and 3, resulting in their inactivation and the release of their c-Myc repression. There may be a role for endoglin/β-arrestin2 in inhibiting ERK signaling so that Smad

2 and 3 can more effectively repress c-Myc expression.

As shown by our MTT assay, the functional role for endoglin/ β-arrestin2 is to inhibit cell proliferation. While the results of the MTT assay only showed a moderate difference between

Eng +/+ and Eng -/- MEECs, a more pronounced difference was observed upon rescue of Eng -/-

MEECs with Eng WT, suggesting that endothelial cell growth is largely mediated by endoglin.

The inhibition seemed to be partially dependent on β-arrestin2 as rescue with endoglin T650A only moderately enhanced cell growth. This suggests that endoglin/β-arrestin2 may not be the sole mediator of this process. As mentioned earlier, endoglin interacts with other scaffolding proteins including zyxin, ZRP-1, and GIPC. How these interactions contribute to endothelial proliferation remains to be determined.

In summary, we defined a novel mechanism by which endoglin regulates endothelial cell growth.

Our work demonstrates that endoglin, through its interaction with β-arrestin2, inhibits endothelial cell proliferation by suppressing the activation of the mitogenic activated protein kinase, ERK, and its downstream targets, c-Myc and cyclinD1.

Chapter 3: Src-mediated Post-translational Regulation of Endoglin Stability and Function is Critical for Angiogenesis

Abstract

The goal of this thesis aim was to define a mechanism for endoglin down-regulation. Endoglin is a TGF-β co-receptor that is essential for angiogenesis, vascular development, and cancer progression. Despite its established role in physiology and disease progression, the mechanism of endoglin down-regulation is undefined. In this study, we identified two conserved tyrosine residues (612YIY614) that are located distal to the transmembrane segment. These residues serve as critical determinants of angiogenesis and are phosphorylated by the non-receptor tyrosine kinase,

Src, resulting in endoglin’s internalization and degradation via the lysosome. In addition, we identified epidermal growth factor (EGF) and vascular endothelial growth factor (VEGF) as activators of Src that promote 612YIY614 phosphorylation and endoglin turnover. Interestingly, the monoclonal anti-endoglin antibody (TRC105) induces endoglin turnover by phosphorylating

612YIY614 residues through Src activation. Regulation of 612YIY614 phosphorylation is critical for angiogenesis as both the constitutively and unphosphorylatable mutants displayed impaired endothelial migration, proliferation, and capillary tube formation. Taken together, this study

612YIY614 motif is a novel motif for endoglin stability and function.

Background

Endoglin is an endothelial-specific TGF-β accessory receptor that is essential for angiogenesis and vascular homeostasis. Endoglin modulates the balance between pro- and anti-angiogenic activation. Its expression is virtually undetectable in quiescent endothelial cells; however, it is

31 markedly up-regulated in actively proliferating endothelial cells [67]. Its robust expression serves as a maker for tumor vascularization and directly correlates with tumor growth and metastasis

[48, 72, 73]. Given its specific expression pattern and essential role in angiogenesis, endoglin serves as an attractive vascular target for tumor-associated angiogenesis and cancer therapy.

Despite these advances, many aspects of endoglin biology are largely undefined, in particular post-translational mechanisms that mediate endoglin down-regulation.

Endoglin is characterized by a large extracellular ligand binding domain that binds TGF-β1/3 and

BMP-9/10, a single transmembrane domain, and a short cytoplasmic domain that is comprised of many serine and threonine residue, some of which serve as phosphorylation sites and docking sites for variety of signaling/trafficking proteins [74, 75]. In addition, many of these serine/threonine sites are associated with endoglin internalization. For example, PKC, upon activation by thrombin, phosphorylates intracellular serine residues of endoglin and induces its endocytosis [76]. In addition, β-arrestin2 binds to the phosphorylated threonine residue at position 650 of endoglin and promotes its internalization [38]. While serine and threonine phosphorylation appears to have a direct role in endocytic trafficking and downstream signaling, endoglin expression levels remain the same.

In addition to the serine/threonine residues interspersed throughout the cytoplasmic domain, there is an evolutionary conserved juxtamembrane cytoplasmic peptide sequence that contains two tyrosine residues at positions 612 and 614 (612YIY614). To date, no functional role has been assigned to these tyrosine residues. Given their close proximity to the plasma membrane, we hypothesized that, if phosphorylated, these tyrosine residues could induce conformational changes that may affect endoglin function or stability or serve as a platform for tyrosine kinase signaling complexes.

Materials and Methods

Cell Culture, Plasmids, Transfections, and Antibodies (Abs) Mouse embryonic endothelial cells

(MEECs) (Eng+/+ and Eng−/−) were derived from wild type and endoglin knock-out mice at E9 as previously described. MEECs were maintained in MCDB-131 medium (Invitrogen) supplemented with 2 mM L-glutamine, 1 mM sodium pyruvate, 15% fetal bovine serum, 100

μg/ml heparin, 25 μg/ml endothelial cell growth supplement. HMEC-1 was maintained in

MCDB-131 medium supplemented with 10% fetal bovine serum, 1 μg/ml hydrocortisone

(Sigma), 10 ng/ml epidermal growth factor (Sigma), and 2 mM L-glutamine. COS7 cells were maintained in DMEM with 10% fetal bovine serum. Transfections were achieved by using

Lipofectamine 2000 as described according to manufacturer's protocol (Invitrogen). Human endoglin cDNA was used as a template to generate endoglin mutants. Briefly, the Eng-YE

(Y612/614E) and Eng-YF (Y612/614F) double point mutations were created by PCR site- directed mutagenesis using forward primers GCT GCA CTC TGG GAG ATC GAG TCG CAC

ACG CG and GCT GCA CTC TGG TTT ATC TTT TCG CAC ACG CG, respectively. Two shRNA targeting sequences for human endoglin knockdown were as follows:

CCGGGCGAGGTGACATATACCACTACTCGAGTAGTGGTATATGTCACCTCGCTTTTTG and

CCGGCCACTTCTACACAGTACCCATCTCGAGATGGGTACTGTGTAGAAGTGGTTTTTG

. Basement matrigel matrix was obtained from BD Biosciences. TGF-β-1 and BMP-9 were obtained from R&D Systems. Inhibitors for Src (PP2) and the proteasome (MG132) were obtained from Sigma Aldrich. Abs used in this study were: TRC105 (a generous gift from Dr.

Charles Theuer, TRACON Pharmaceuticals), endoglin (H-300, Santa Cruz Biotechnology), endoglin P3D1 (University of Iowa Hybridoma), ALK-1 (Santa Cruz Biotechnology), TβRII

(Santa Cruz Biotechnology), 20s proteasome α6 (Santa Cruz Biotechnology), anti- phosphotyrosine (Millipore), HA (Roche, Applied Sciences), Myc (Sigma-Aldrich), and β-actin 34

(Sigma-Aldrich). The following Abs were all purchased from Cell Signaling: phospho-Smad

1/5/8 (no. 9511), total Smad 1 (no. 6944), phospho-Smad 2/3 (no. 9510), total Smad 2/3 (no.

9510), phospho-Tyr416 Src (no. 6943), total Src (no. 2109), phospho-Ser473 Akt1 (no. 9018), phospho-ERK 1/2 (no. 4370), EEA1 (no. 2411), LAMP1 (no. 9091), phospho-Tyr (no. 9411).

Immunoprecipitation Cells were washed briefly and then lysed on ice with lysis buffer (20 mM

HEPES, pH 7.4, 150 mM NaCl, 2 mM EDTA, 10 mM NaF, 10% (w/v) glycerol, 1% Nonidet P-

40) and supplemented with protease inhibitors (Sigma protease inhibitor mixture) and phosphatase inhibitors (Sigma phosphatase inhibitor mixture). The lysates were precleared by centrifugation and incubated with appropriate primary Abs for 2–4 h, and then with protein agarose G/A for 1–2 h at 4 °C. The immunoprecipitates were collected by centrifugation; pellets were washed with lysis buffer and stored in 2× sample buffer before Western blot analyses. For soluble endoglin, endoglin was immunoprecipitated from the conditioned media with extracellular-targeting antibody (TRC105 or P3D1).

Immunofluorescence Eng−/− MEECs or COS7 cells grown overnight on coverslips were transiently transfected with appropriate constructs using Lipofectamine 2000 (Invitrogen) as described. For ubiquitin staining, cells were treated with MG132 (10 μM) for 2 h prior to fixation.

For TRC105-induced endoglin internalization, MEECs were pre-treated with TRC105 (200 ng/ml) for the indicated duration prior to fixation. 24–48 h following transfection, cells were washed with PBS and then fixed with 4% paraformaldehyde. Cells were permeabilized in 0.1%

Triton X-100 in PBS for 3–10 min, then blocked with 5% bovine serum albumin in PBS containing 0.05% Triton X-100 for 20 min. All primary Abs were incubated at room temperature for 1 h unless noted otherwise. Eng-WT, Eng-YE, and Eng-YF expression was detected using

TRC105. EEA1, LAMP1, and ubiquitin Abs were used to detect early endosomes, lysosomes, and ubiquitin clusters, respectively. Proteasome was detected using 20 S Proteasome α6. ALK1,

TβRII, and ALK5 were detected using ALK1, TβRII, and HA Abs, respectively. Following

primary antibody incubation, cells were incubated with appropriate flurophore conjugated secondary antibodies (Alexa-Fluor) at room temperature for 30 min. Cells were co-stained with

DAPI (Sigma) immediately before immunofluorescence microscopy analyses (Olympus FV1000 confocal system). Pearson correlation coefficient analysis was used with ImageJ to measure co- localization.

Transwell Migration Assays Eng+/+ and Eng−/− MEECs were transfected with appropriate endoglin constructs. 24 h following transfection, cells were seeded in the upper chamber of a transwell filter in complete growth media, coated both at the top and bottom with gelatin. Cells were allowed to migrate for 12 h toward the lower chamber containing growth media alone or growth media containing EGF (5 × 10−4 mg/ml) or VEGF (5 × 10−4 mg/ml). Cells that migrated to the bottom surface of the filter were fixed, stained, and then digitally imaged and counted.

Crystal Violet Cell Growth Assay Eng−/− MEECs were plated at 15,000 in 12-well plates and transfected with the appropriate endoglin constructs. Following transfection, cells were fixed at different time points (4% paraformaldehyde in PBS for 15 min). Following fixation, cells were washed with 1× water and stained with 0.1% crystal violet for 20 min. Cells were washed 3× with water and allowed to air dry for 30 min. Cells were destained using crystal violet destaining

solution (10% acetic acid, 50% methanol, 40% H20) for 20 min, and the optical density was read at 590 nm in a microplate reader.

Endothelial Tube Formation Eng−/− ECs were transfected with the appropriate endoglin constructs. 24 h following transfection, cells were trypsinized and plated on a 24-well plate coated with 200 μl of matrigel basement matrix (BD Biosciences) at 160,000 cells/well. 1 h following plating, growth medium was removed and 200 μl of matrigel basement matrix was added. 30 min following the addition of the matrigel basement matrix, 300 μl of growth medium was added. Endothelial tubes were digitally imaged and quantified by measuring the relative tube length and counting the number of branches per node. Tube length and number of branches per

node for each endoglin construct were normalized to the value in Eng-WT.

Figure 3.1: Endoglin tyrosine motif is an endocytic signal for down-regulation. (A) Endoglin schematic shows the extracellular domain containing 6 tyrosine sites, a transmembrane segment, and an intracellular domain containing 2 tyrosine residues as indicated. The two intracellular tyrosine residues were mutated to glutamate (Eng-YE), phenylalanine (Eng-YF), or alanine (Eng-

YA). (B) Peptide sequence alignment shows the conserved intracellular tyrosine motif

(highlighted) for endoglin across species for endoglin and with TβRIII.

continued

Figure 3.1 continued

(C) A representative Western analysis of endoglin expression in Eng−/− ECs (upper panel). A quantitative densitometry of relative endoglin expression normalized to Eng-WT and β-actin.

Statistical analysis was based on four independent experiments. *p < 0.05.

continued

Figure 3.1 continued

(D) Endoglin localization upon expression of human Eng-WT, Eng-YE, Eng-YF, or Eng-YA in

Eng−/− MEECs and stained for endoglin (green) using TRC105. DAPI co-staining (blue) indicates non-transfected Eng−/− MEECs. (E) Western analysis of endoglin cleavage upon immunoprecipitation of sEng from conditioned media of Eng−/− MEECs expressing Eng-WT,

Eng-YE, or Eng-YF. 40

Figure 3.2: Phosphomimetic mutation of the tyrosine endocytic motif mediates endoglin trafficking to the lysosome. (A) Eng-YE expression in Eng−/− MEECs is detected with TRC105

(green) and co-stained for EEA1 (red). The numbered squares within each image represent typical ROIs used for the quantification of co-localization. (B) Eng-YE expression (red) is shown to co-localize with Rab9-GFP (green). (C) Eng-YE expression (green) is shown to co-localize with LAMP1 (red).

continued

Figure 3.2 continued

(D) Eng-YE expressing ECs (green) are pretreated with MG132 (4 h) prior to fixation and counterstaining for ubiquitin (red), and the proteasome (blue).

continued 42

Figure 3.2 continued

(E) Quantification of Eng-YE co-localization with various markers using Image J and Pearsons correlation coefficient analysis. Two to three random regions (ROI) away from the nuclear regions were chosen from each Eng-YE-expressing cell for correlation analyses with the indicated markers. Error bars represent the standard error of the mean of correlation coefficients between Eng-YE and each marker. Data are based on the analyses of at least 15 Eng-YE-positive cells (3 ROIs each) from three independent experiments.

continued

Figure 3.2 continued

(F) Immunoblot of Eng-YE expression in Eng−/− MEECs treated with either chloroquine (100

μM, 200 μM) or MG132 (10 μM) for 9 h. (G) COS7 cells expressing Eng-YE (green) are pretreated with chloroquine (100 μM) for 18 h and stained for LAMP1 (red).

Figure 3.3: Endoglin tyrosine motif regulates ALK1 trafficking to the lysosome. (A) Eng−/− ECs expressing either Eng-WT (green; upper panels) or Eng-YF (green; lower panels) and HA-ALK1

(red) and stained for LAMP1 (blue). (B) Quantification of Eng-WT or Eng-YF with ALK1 and

LAMP1 using ImageJ and Pearsons correlation coefficient analysis. Random regions were chosen from each Eng-WT or Eng-YF-positive Eng−/− MEEC and analyzed for correlation with ALK1 and LAMP1. Data are based on the analyses of at least 5 Eng-WT and Eng-YF-positive cells (2

ROIs each) from three independent experiments. continued 45

Figure 3.3 continued

Eng−/− MEECs expressing Eng-YE, ALK1, and Rab9-GFP (green) and stained for ALK1 (red).

(E) Western analysis of Eng−/− ECs expressing HA-ALK1 alone, HA-ALK1 with Eng-WT, HA- 46

ALK1 with Eng-YE, HA-ALK1 with Eng-YF. (F) Quantification of Eng-YE co-localization in a complex with ALK1 and Rab9 using ImageJ and Pearsons correlation coefficient analysis.

Random regions of each ALK1-positive cell were analyzed for correlation with ALK1 and Rab9.

Data are based on the analyses of at least 15 Eng-YE-positive cells (3 ROIs each) from three independent experiments.

Figure 3.4: Endoglin tyrosine motif regulates TβRII trafficking to the lysosome. (A) COS7 cells expressing Eng-WT or Eng-YF (green) and TβRII (red) and stained for LAMP1 (blue). (B) quantification of Eng-WT or Eng-YF with TβRII and LAMP1 using ImageJ and Pearsons correlation coefficient analyses. Random regions were chosen from each Eng-WT or Eng-YF- positive Eng−/− MEEC and analyzed for correlation with TβRII and LAMP1. Data are based on the analyses of at least 5 Eng-WT and Eng-YF-positive cells (2 ROIs each) from four independent experiments. continued 48

Figure 3.4 continued

Eng−/− MEECs expressing Eng-YE, myc-TβRII, and Rab9 (green) and stained for TβRII (red).

(E) Quantification of TβRII co-localization with Eng-YE and GFP-Rab9 using Image J/Pearsons correlation coefficient. Data are based on the analyses of at least 15 Eng-YE-positive cells (3

ROIs each) from three independent experiments.

continued

Figure 3.4 continued

(F) representative subcellular localization of ALK5 alone in Eng−/− MEECs (red). (G) representative subcellular localization of ALK5 (red) and Eng-YE (green) in Eng−/− MEEC

Figure 3.5: Endoglin tyrosine motif is a direct substrate for c –Src phosphorylation. (A) COS7 cells expressing endoglin mutants were immunoprecipitated with p-Tyr Ab, then immunoblotted for endoglin, endogenous p-Src and total Src. Src activity was inhibited with PP2 (15 μM for 2 h) prior to p-Tyr immunoprecipitation (lane 5).

continued 51

Figure 3.5 continued

(B) Effect of Src 5overexpression on expression of Eng-WT versus Eng-YF in COS7 cells.

continued

Figure 3.5 continued

(C) Detection of endogenous interaction between endoglin and Src in human microvascular endothelial cells (HMEC-1). HMEC-1 was transfected with either scramble control or endoglin- targeting shRNAs for 30 h. Endoglin was immunoprecipitated with TRC105 (for control and sh-

Eng lysates) or IgG (control lysate), then immunoblotted for endogenous total Src and endoglin.

The upper two panels show short and long film exposures for endogenous endoglin immunoprecipitation, respectively.

continued

Figure 3.5 continued

(D) Detection of endoglin/Src interaction by co-immunoprecipitation. Various endoglin constructs were expressed along with Src, followed by immunoprecipitation of endoglin

(TRC105) and immunoblotted for endoglin (top panel), phosphotyrosine (second panel), and Src

(lower two panels).

continued 54

Figure 3.5 continued

(E) Eng−/− MEECs expressing Eng-WT was treated for 2 h with TGF-β (200 pM), BMP-9 (1 nM), insulin (200 nM), VEGF (500 ng/ml), and EGF (500 ng/ml).

continued

Figure 3.5 continued

(F) Immunoblot of endogenous endoglin expression in HMEC-1 stimulated with either VEGF

(500 ng/ml) or EGF (500 ng/ml) alone, or in the presence of PP2 (15 μM) for 2 h.

Figure 3.6: Tyrosine endocytic motif mediates TRC105-induced endoglin internalization. (A and

B) Eng−/− MEECs expressing human Eng-WT or Eng-YF were treated with TRC105 (200 ng/ml) for various time points (0 to 6 h) to allow antibody-induced receptor internalization.

Following treatment, cells were fixed, permeabilized, and stained with a fluorescently labeled secondary Ab. continued 57

Figure 3.6 continued

TRC105-endoglin-containing vesicles per cell. Over 100 endoglin-positive cells were counted per condition (Eng-WT versus Eng-YF) in three independent experiments. *, p < 0.03; **, p < 0.05.

Figure 3.7: TRC105 mediates Src-dependent endoglin tyrosine phosphorylation. (A) Eng −/−

MEECs transiently expressing Eng-WT were treated with TRC105 (200 ng/ml) for 2 h and stained for endoglin (green), EEA1 (red), and LAMP-1 (red).

continued 59

Figure 3.7 continued

(B) COS7 cells transiently expressing Eng-WT were treated with TRC105 for 15 min before immunoprecipitation with p-Tyr Ab, and immunoblotted for endoglin. Lysates were immunoblotted for Src, Eng, and β-actin for expression.

continued 60

Figure 3.7 continued

(C) COS7 cells expressing Eng-WT were treated with TRC105 (200 ng/ml) alone, or pre-treated with PP2 (15 μm) for 15 min before treatment with TRC105 and PP2. Cells were immunoprecipitated for p-Tyr and immunoblotted for Eng. Cell lysates were immunoblotted for

Eng and β-actin using appropriate antibodies.

Figure 3.8: Endoglin tyrosine phosphorylation motif is a critical determinant of endothelial migration. (A) Quantification of transwell migration for Eng+/+, Eng−/−, and Eng−/− MEECs expressing Eng-WT, Eng-YE, or Eng-YF. MEECs were plated into transwells in growth media.

*, p < 0.05; **, p < 0.04 (Eng-WT or Eng-YF compared with Eng−/−).

continued

Figure 3.8 continued

(B) Representative images of Eng-WT or Eng-YF transwells upon VEGF (500 ng/ml) or EGF

(500 ng/ml) stimulation for 16 h. (C) Quantification of transwell migration for Eng−/− MEECs expressing Eng-WT or Eng-YF in growth media, or growth media plus VEGF (500 ng/ml) or

EGF (500 ng/ml) stimulation (16 h). *, p < 0.04; **, p = 0.01 (compared with Eng-WT plus EGF treatment). continued 63

Figure 3.8 continued

(D) Quantification of transwell migration for Eng−/− MEECs expressing Eng-WT or Eng-YF upon VEGF (500 ng/ml) or EGF (500 ng/ml) stimulation (16 h) in the presence of PP2 (5 μm). *, p = 0.02 (compared with Eng-WT plus EGF treatment); **, p = 0.03 (compared with Eng-WT plus VEGF treatment).

Figure 3.9: Endoglin tyrosine phosphorylation is critical for endothelial capillary sprouting and stability. (A) Representative images of three-dimensional Matrigel-induced capillary tubules for

Eng−/− MEECs, and Eng−/− MEECs expressing Eng-WT, Eng-YE, or Eng-YF for 72 h. (B)

Quantification of overall capillary tubule formation and branching by measuring the average number of branches per node at 72 h in the presence or absence of PP2 treatment (5 μm) 72 h following plating. *, p < 0.05 (Eng−/−, Eng-YE, or Eng-YF compared with Eng-WT), **, p =

0.01 (Eng-WT + SRC inhibitor compared with Eng-WT). (C) Time course measurements of capillary tubule branching for Eng-WT, Eng-YE, and Eng-YF at 24 and 72 h. continued

Figure 3.9 continued

(D) Crystal violet growth assay of Eng−/− and Eng−/− MEECs expressing Eng-WT, Eng-YE, or

Eng-YF 12 h, 24 h, and 36 h following transfection. *, p = 0.002 (Eng−/− compared with Eng-

WT at 36 h); **, p = 0.00005 (Eng-YE compared with Eng-WT at 36 h).

Results

The human endoglin sequence reveals that endoglin is comprised of eight tyrosine residues, six of which reside in the extracellular domain and two are located immediately below the transmembrane segment (Tyr612 and Tyr614) (Figure III-1A). As demonstrated by sequence alignment, the intracellular tyrosine residue motif is evolutionarily conserved and is also present in another TGF-β co-receptor, TβRIII (Figure III-1B). To test the function of this motif, we performed site directed mutagenesis and generated cDNAs encoding for phosphomimetic mutant

(both tyrosine residues were mutated to glutamate; Eng-YE) and for an unphosphorylatable mutant (both tyrosine residues were mutated to phenylalanine; Eng-YF) and performed rescue experiments in Eng -/- cells. Eng-YE expression was noticeably reduced compared to Eng-WT and Eng-YF, suggesting that phosphomimetic mutant either didn’t express correctly or it is constitutively degraded (Figure III-1C). Analysis of Eng-YE and Eng-YF subcellular localization demonstrated that Eng-YF and Eng-WT displayed membrane and cytoplasmic distribution, which is typical for all TGF-β receptors including endoglin. Eng-YE, however, displayed a punctate endocytic distribution (Figure III-1D). To rule out the possibility of protein aggregation of Eng-

YE due to structural perturbation, we tested the subcellular distribution of Eng-YA, an endoglin construct that has both tyrosine residues mutated to alanine. Eng-YA distribution pattern was the same as Eng-WT and Eng-YF, suggesting that the phosphomimetic mutation likely alters Eng-YE expression. Given that endoglin can undergo ectodomain shedding through cleavage by the matrix metalloproteinase-1 (MMP-14) to release soluble endoglin (sEng), we compared sEng levels of Eng-WT, Eng-YF, and Eng-YA to demonstrated that the reduced expression levels and atypical subcellular localization of Eng-YE were not due to improper folding or trafficking to the cell surface. sEng levels were relatively the same for all endoglin constructs (Figure III-1E), suggesting that Eng-YE is capable of trafficking to the cell membrane and undergoing ectodomain shedding despite the punctate cellular distribution and reduced protein expression 67

level.

To determine the subcellular characteristics of Eng-YE, we co-stained Eng-YE with variety of different endosomal markers, including the early endosome (EEA1), late endosome (Rab-9), lysosome (LAMP1, ubiquitin, and the proteasome (Figure III-2) Eng-YE showed significant colocalization with EEA1, Rab9, and LAMP1 (Figures III-2A, B, C), but showed no significant colocalization with ubiquitin and proteasome (Figure III-2D). Consistent with these results, Eng-

YE expression was significantly increased upon treatment with a lysosomal inhibitor

(chloroquine) (Figure III-2F). However, this was not observed upon treatment a proteasome inhibitor (MG132) (Figure III-2F). Moreover, chloroquine treatment enhanced Eng-YE punctate size, which strongly co-localized with LAMP1 (Figure III-2G). Taken together, these findings suggest that Eng-YE down-regulation occurs primarily through the lysosomal degradative pathway.

Given that endoglin can form a heteromeric complex with other TGFβ receptors including ALK1

TβRII, and ALK5, we examined whether the endoglin’s YIY motif influences their subcellular trafficking. ALK1 localization was not affected upon co-expression with Eng-WT and Eng-YF

(Figures III-3A and B). However, ALK1 localization was dramatically altered upon co- expression with Eng-YE (Figure III-3C) as it was recruited into endocytic vesicles by Eng-YE where they colocalized with late endosomal markers such as Rab9 (Figures III-3C, D, F).

Supporting the notion that endoglin’s YIY motif alters ALK1 stability, global ALK1 levels were reduced when co-expressed with Eng-YE whereas ALK1 levels remained the same in the presence of Eng-WT and Eng-YF (Figure III-2E). Similarly to ALK1, TβRII subcellular localization was not altered when co-expressed with Eng-WT and Eng YF but was dramatically altered when co-expressed with Eng-YE (Figures III-4A, B, C). TβRII co-localized with Eng-YE in endocytic vesicles in late endosomal compartments as well (Figures III-4C and D).

Interestingly, ALK5 maintained its normal cytoplasmic distribution when co-expressed with Eng-

YE (Figure III-4G). Taken together, these findings suggest that endoglin phosphomimetic motif selectively mediates the lysosomal degradation of ALK-1 and TβRII, but not ALK-5.

To test whether the YIY motif becomes tyrosine phosphorylated, we immunoprecipitated for phosphorylated tyrosine upon expression of Eng-WT, Eng-YE, or Eng-YF in endoglin-null cells

(Figure III-5A). Tyrosine phosphorylation was only detected for Eng-WT despite the tyrosine residues located on the extracellular domain of Eng-YE and Eng-YF, indicating that the YIY motif is major site of tyrosine phosphorylation on endoglin. To identify the major tyrosine kinase that is responsible for endoglin tyrosine phosphorylation, we searched for peptide substrate motifs that are recognized by various protein tyrosine kinases. Previous studies have demonstrated that c-Src and related family kinases recognize the YIY sequence with a hydrophobic residue in position 2 and a basic amino acid in position 7 [77, 78]. Given that endgolin’s YIY motif is followed by a conserved basic residue, arginine, at position 7 (Figure III-1B), we tested c-Src as a potential tyrosine kinase by inhibiting its catalytic activity and immunoprecipitating for phosphorylated tyrosine in Eng-WT-expressing cells. Interestingly, phosphorylated tyrosine levels were dramatically reduced upon c-Src inhibition (Figure III-5A). To further determine c-

Src regulates endoglin expression through tyrosine phosphorylation, Eng-WT and Eng-YF were expressed in the presence of increasing Src expression. Eng-WT levels decreased with increasing levels of Src whereas Eng-YF levels were unaffected (Figure III-5B), suggesting that Src targets endoglin YIY motif to alter its stability and reduce its expression. To determine whether endoglin is a Src substrate, we performed endogenous co-immunoprecipitation studies. Src endogenously interacted with endoglin and this interaction was abolished upon endoglin knockdown (Figure III-

5C). We next tested for Src interaction with the endoglin mutants to determine whether Src binding specifically required the YIY motif. Src interacted with Eng-WT and Eng-YF, but did not interact with Eng-YE, suggesting that the dual aromatic side chains of WY612IY614 motif are an 69

essential structural recognition motif for the Src catalytic domain (Figure III-5D). More importantly, Eng-WT/Src interaction resulted in endoglin tyrosine phosphorylation whereas Eng-

YF/Src interaction did not, further supporting endoglin-612YIY614 motif as a novel Src phosphorylation motif and not a SH2 domain-binding site (Figure III-5D).

We next screened for various Src-activating cytokines that mediate endoglin tyrosine phosphorylation and degradation [79, 80]. Treatment with VEGF and EGF for 2 h. caused notable endgolin turnover whereas treatment with TGF-β, BMP-9, and insulin had no effect, suggesting a selective Src-dependent response (Figure III-5E). To further test Src as a mediator of this process, we examined the effects of VEGF and EGF stimulation on endoglin expression in the presence or absence of Src inhibition (Figure III-5F). Consistent with our previous results, VEGF and EGF- induced endoglin degradation was blocked upon Src inhibition. Overall, the results demonstrate that VEGF and EGF mediated Src activation results in endoglin down-regulation.

Antibody-induced receptor internalization is a common process by which receptors are degraded.

Endoglin targeting antibodies suppress angiogenesis and tumor growth by binding to endoglin and promoting its internalization and degradation. However, the mechanisms underlying antibody-induced internalization and degradation are not well defined. As part of our ongoing studies on the anti-angiogenic properties of TRC105, we tested the role of tyrosine motif in

TRC105-induced endoglin internalization. To test this, Eng-W- and Eng-YF-expressing Eng -/-

MEECs were treated with TRC105 for various time intervals (0 to 6 h). Internalization of the receptor-antibody complex was monitored at each time point through labeling with a fluorescent secondary antibody after cell fixation and permeabilization (Figure III-6A). At early time points, there was minimal TRC105-induced internalization of cell surface labeled Eng-WT and Eng-YF.

However, after 30 min. of treatment, there was significantly more TRC105-Eng-WT present in endocytic vesicles compared to TRC105-Eng-YF and this became more apparent with increased

treatment time, suggesting that the tyrosine motif plays a major role in TRC105 induced endoglin internalization (Figures III-6A, B, C). Co-staining with early endosomal and lysosomal makers suggested that the internalized receptor complex is degraded (Figure III-7A). Interestingly, there was a dose dependent increase in endoglin tyrosine phosphorylation upon TRC105 treatment and this increase was abolished upon Src inhibition (Figures III-7B, C), further supporting the role of

Src-induced tyrosine phosphorylation in endoglin degradation.

We began characterizing the cellular effects of the tyrosine motif by comparing the effects of

Eng-WT, Eng-YE, and Eng-YF expressing Eng -/- MEECs on various endothelial cell functions including migration, proliferation, and capillary tube formation. Consistent with our previous findings, endoglin inhibited cell migration by 30-40% when comparing the number of migrated cells between Eng +/+ and Eng -/- MEECs (Figure III-8A). In addition, overexpression of Eng-

WT in Eng -/- MEECs further inhibited migration. Unlike Eng-YF and Eng-WT, Eng-YE phenocopied Eng -/- MEECs in that it failed to suppressed migration, most likely due to its constitutive turnover (Figure III-8A). We also examined the effects of EGF and VEGF stimulation on migration of Eng-WT and Eng-YF-expressing MEECs. EGF and VEGF normally enhance endothelial cell migration through mitogenic signaling [81]. As expected, Eng-WT- expressing MEECs displayed enhanced migration compared to control in response to VEGF and

EGF. Eng-YF expression, however, impaired the EGF and VEGF-induced migratory response

(Figures III-8B, C). Src inhibition typically results in reduced migratory responses in many cell types, including ECs. Here, Src inhibition dramatically reduced EGF and VEGF-induced migratory response in Eng-WT expressing MEECS, but had no effect on Eng-YF-expressing

MEECs (Figure III-8D), suggesting that endoglin turnover following tyrosine phosphorylation by

Src is a critical process by which VEGF and EGF enhance endothelial cell migration. In addition to endothelial cell migration, we also examined the role of tyrosine motif on endothelial capillary tube formation. Consistent with our previous results, Eng -/- MEECs displayed unstable capillary 71

tubes that regressed over time compared to those that were formed in Eng-WT-expressing Eng -/-

MEECs (Figures III-9A, B). Rescue with Eng-YE or Eng-YF only partially restored efficient capillary branching and stability relative to Eng-WT (Figure III-9B), suggesting that a balance of tyrosine phosphorylation is essential for proper capillary formation. Consistent with this, Src inhibition reduced capillary tube formation in Eng-WT-expressing MEECs but had no effect on capillary tubes formed by Eng-YE and Eng-YF-expressing MEECs (Figure III-9B). The deficiency in capillary formation in Eng-YE and Eng-YF-expressing MEECs, while initially surprising, can be attributed to the fact that Eng-YE-expressing MEECs may have formed branches that regressed over time, whereas Eng-YF-expressing MEECs failed to form branches in the first place (Figure III-9C). Eng-YF-expressing MEECs not only had impaired VEGF and

EGF-induced cell migration, but also proliferated slower than Eng-YE-expressing MEECs

(Figure III-9D). Our results demonstrate that phosphorylation of endoglin’s tyrosine motif plays critical roles in endothelial functions during angiogenesis.

Discussion

Our studies on endoglin’s tyrosine motif demonstrated the first post-translational mechanisms by which endoglin is regulated. In addition, we identified a mechanism by which Src and potentially other family members regulate endoglin function and stability. Very few studies have examined post-translational mechanisms mediating endoglin down-regulation whereas post-translational regulations of TβRII and ALK5 have been more thoroughly characterized. A study identified

TNF-α as a key post-translational mediator of endoglin expression. Long-term treatment with

TNF-α elicited a significant suppression of endoglin protein levels, but the underlying mechanism by which it does so is unknown. TβRII and ALK5 are both internalized in clathrin- coated pits and lip raft endosomes, and ALK5 is degraded through the proteasome by the ubiquitin-associated salt-inducible kinase (SIK) and Smad7. While we were not able to detect

endoglin ubiquitination and degradation through the proteasome, it remains plausible that only a subset of the TGF-β heteromeric receptor complex is degraded by the proteasome. In addition, interaction with other binding partners could promote endoglin’s proteasomal degradation. Our results clearly demonstrate that chloroquine enhanced Eng-YE accumulation in lysosomal compartments and restored its expression, supported a novel function for endoglin tyrosine phosphorylation in mediating endoglin’s lysosomal degradation.

The fact that only VEGF and EGF, and not other Src-activating ligands, induced endoglin down- regulation suggests that Src targeting of endoglin tyrosine motif is highly specific and tightly regulated. Consistent with this, stimulation with insulin yielded the longest period of Src activation despite having little to no effect on endoglin stability. It is unclear whether specific receptor tyrosine kinases such as VEGFR2 and EGFR activate Src and recruit it to endoglin or whether other signaling complexes are involved.

Given that Src targets endoglin, it is also likely that endoglin may be regulated by tyrosine phosphatases. Interestingly, deletion of the endothelium-expressed receptor phosphatase, CD148, was embryonic lethal at day 11.5 due to growth retardation and disorganized vascular structures and resulted in a noticeable decrease in endoglin expression throughout the entire vasculature whereas the expression VEGFR2, ALK1, and Tie2 were unaffected. This suggests that tyrosine phosphatases play a key role in endoglin stability and signaling whereas tyrosine kinases promote endoglin degradation and antagonize its normal signaling. The interplay between endoglin- specific tyrosine phosphatases and tyrosine kinases may be critical for angiogenesis.

The defects in capillary tube formation in Eng-YE and Eng-YF ECs were initially surprising.

Given that endoglin is essential for proper tube formation, we expected a less stable form (Eng-

YE) of endoglin to lead to impaired capillary development and a more stable form (Eng-YF) of endoglin to lead to enhanced tube formation. However, Eng-YE and Eng-YF expressing ECs both

exhibited impaired tube formation. A closer inspection of our data indicated that Eng-YE and

Eng-YF mutants impaired angiogenesis through distinct mechanisms. While Eng-YE expressing capillaries initially formed, they were unstable and regressed over time. In contrast, Eng-YF expressing ECs were not able to form capillaries in the fist place, possibly due to impaired

VEGF-or EGF-induced migratory response. Taken together, a critical balance between tyrosine phosphorylation and dephosphorylation is required for proper capillary formation. As we have reported previously, rescue with Eng-WT in Eng -/- ECs inhibited endothelial proliferation.

However, it is unclear as to why Eng-YF expression increased proliferation relative to Eng-WT over time. Given that endoglin modulates numerous signaling pathways that are involved in migration and proliferation such as Smads, ERK, and PI3K/Akt, it is highly likely that endoglin’s tyrosine motif may regulate these pathways to induce changes in endothelial cell function.

As part of our ongoing studies on TRC105, we tested whether TRC105 had an effect on endoglin tyrosine phosphorylation. Our results indicate that TRC105 promotes endoglin tyrosine phosphorylation, internalization, and lysosomal degradation through Src activation in a concentration and time dependent manner. Whether through exogenous (TRC105) or endogenous

(VEGF, EGF) factors, Src phosphorylation of the 612YIY614 motif is required for endoglin down- regulation.

Overall our study defines a role for the evolutionarily conserved 612YIY614 motif. Our work indicates that this motif acts as a direct substrate for the non-receptor tyrosine kinase, Src, and serves as a key determinant for endoglin down-regulation and angiogenesis. In addition, we identified VEGF and EGF as negative regulators of endoglin. We demonstrate that VEGF and

EGF promote endoglin tyrosine phosphorylation through rapid Src activation. This regulation may be part of a negative feedback mechanism that allows for endothelial cells to migrate and proliferate out of the basement membrane to form stable capillaries. Finally, our work provides

clinically relevant data by elucidating a novel mechanism by which TRC105 inhibits endoglin function and signaling.

Chapter 4: Endoglin Regulation of Smad2 Function Mediates Beclin1 Expression and Endothelial Autophagy

Abstract

Autophagy is a catabolic process by which the cell self-digests its organelles and proteins to maintain macromolecular synthesis and ATP production. Although autophagy can be induced by a number of different stimuli in different cell types including nutrient deprivation, growth factor depletion, and hypoxia, the relationship between TGF-β signaling and autophagy is poorly understood Importantly, how TGF-β controls autophagic responses in endothelial cells during angiogenesis is unknown. In this study, we identified endoglin as a key determinant of autophagy. In addition, we also demonstrate that Smad2, but not Smad3, is a major transcriptional regulator of autophagy. Smad2 binds upstream of the beclin1 (BECN1) promoter and suppresses its gene expression, thereby attenuating autophagic responses. Overall, endoglin promotes autophagy by suppressing Smad2’s repression of BECN1 gene expression and this pro-autophagic role is essential for angiogenesis. Taken together, our results identify endoglin as a key mediator of autophagy in endothelial cells, establish a novel transcriptional mechanism by which Smad2 inhibits angiogenesis, and demonstrate that autophagy is a critical promoter of angiogenesis.

Background

Autophagy is an evolutionary conserved catabolic process by which damaged organelles and long-lived proteins are degraded and recycled to maintain cell homeostasis. Autophagy is upregulated in response to a variety of different stimuli including starvation, growth factor

deprivation, and hypoxia to promote cell survival [82]. Despite its importance in maintaining cell survival, dysfunctional autophagy plays a pivotal role in a number of human pathologies including cancer, neurodegneration, and infectious disease.

Autophagy begins with the de novo synthesis of an isolation membrane, which sequesters cytoplasmic components into a vesicle called the autophagosome. The autophagosome fuses with the lysosome, leading to the breakdown of its contents and the generation of amino acids that can be used for macromolecule synthesis or ATP production. Autophagy is regulated by a variety of proteins including autophagy-related gene (Atg) proteins and microtubule-associated protein light chain 3 (LC3) that help facilitate the initiation, elongation, and fusion of the autophagosome. In particular, Beclin-1 is a critical activator of autophagy. Through its interaction with the class III phosphoinositide-3-kinase (PI3K), Vsp34, Beclin-1 forms a core complex consisting of Beclin-1-

Vps34-Vps15 that mediates the localization of autophagic proteins to a pre-autophagosomal structure. Given Beclin-1’s central role in autophagy, there are numerous transcription and post- translation mechanisms that tightly regulate Beclin-1 function and expression. Transcriptional factors such as FoxO3, NFkB, HIF1α, c-Jun, and E2F1 drive Beclin-1 gene expression whereas

Bcl-2, Bcl-XL, DAPK control its function.

Despite recent progress, the role of autophagy in angiogenesis is unclear. Some studies have shown that autophagy promotes angiogenesis by enhancing endothelial tube formation and cell migration, while others have shown the opposite. Similarly, the role of TGF-β in autophagy is equally unclear. TGF-β activates autophagy in some hepatocellular carcinoma cells whereas TGF-

β counteracts this process in fibroblasts. TGF-β’s dichotomous role in autophagy can be attributed to the fact that TGF-β simultaneously activates a number of cellular pathways that regulate pro- and anti-autophagic responses including TGF-β activated kinase 1 (TAK1) and Jun N-terminal kinase (JNK) which induce autophagy in epithelial cells and Akt and mTOR which inhibit

autophagy in fibroblast. In addition to Smad independent pathways, Smads have also been implicated in regulating autophagy. Whether they have direct transcriptional effects remains unclear.

To determine the role of TGF-β in endothelial autophagy, we investigated how endoglin regulates this process during angiogenesis. Endoglin is an endothelial-specific TGF-β co-receptor that is essential for both normal and tumor-associated angiogenesis. It modulates the critical balance between pro- and anti-angiogenic pathways by mediating Smad 1/5/8 and Smad 2/3 signaling, respectively. In addition to Smad dependent pathways, endoglin also regulates Smad independent pathways such as the MAPK pathway to elicit changes in endothelial biology. Here we identify endoglin as a key mediator of autophagy in endothelial cells and demonstrate Smad 2 as a novel repressor of Beclin-1 gene transcription.

Materials and Methods

Cell Culture, Plasmids, Transfections, and Antibodies (Abs) Mouse embryonic endothelial cells

(ECs) (Eng+/+ and Eng−/−) were derived from wild type and endoglin knock-out mice at E9 as previously described (29). ECs were maintained in MCDB-131 medium (Invitrogen) supplemented with 2 mM L-glutamine, 1 mM sodium pyruvate, 15% fetal bovine serum, 100

μg/ml heparin, 25 μg/ml endothelial cell growth supplement. Human umbilical vein endothelial cells (HUVECs) were maintained in EBM-2 basal medium (Lonza) supplemented with EGM-2

SingleQuot Kit Suppl. and Growth Factors (Lonza). Transfections were achieved by using

Lipofectamine 2000 as described according to manufacturer's protocol (Invitrogen). shRNA constructs for mouse and human Smads Smad2 were all purchased from Sigma. The shRNA targeting sequence for mouse Smad2 were: 5′-

CCGGTGGTGTTCAATCGCATACTATCTCGAGATAGTATGCGATTGAACACCATTTTTG

-3′ and 5′-

CCGGCTAAGTGATAGTGCAATCTTTCTCGAGAAAGATTGCACTATCACTTAGTTTTT-

3′. The shRNA targeting sequence for mouse Smad3 was 5′-

CCGGCTGTCCAATGTCAACCGGAATCTCGAGATTCCGGTTGACATTGGACAGTTTTTG

-3′. The shRNA targeting sequences for human Smad2 were: 5′-

CCGGGCGTTGCTCAAGCATGTCATACTCGAGTATGACATGCTTGAGCAACGCTTTTTG

-3′ and 5′-

CCGGCAAGTACTCCTTGCTGGATTGCTCGAGCAATCCAGCAAGGAGTACTTGTTTTTG

-3′. The shRNA targeting sequence for human Smad3 was: 5′-

CCGGGGATTGAGCTGCACCTGAATGCTCGAGCATTCAGGTGCAGCTCAATCCTTTTT

G-3′. Stable Smad2 knockdown MEECs were generated by transfecting each mouse shRNA construct into Eng+/+ ECs and selecting with puromycin (3 μg/ml). Individual puromycin- resistant colonies were isolated and scaled up as clones upon validation of endogenous > 80%

Smad2 depletion. Although different clones of sh-Smad2 MEECs were used interchangeably, the experimental data shown is based on ECs originally derived from transfection with te construct:

5′-

CCGGTGGTGTTCAATCGCATACTATCTCGAGATAGTATGCGATTGAACACCATTTTTG

-3′. Basement matrigel matrix was obtained from BD Biosciences. TGF-β1 and BMP-9 were obtained from R&D Systems. Inhibitors of ALK1 (LDN193189), ALK5 (SB431542), PI3K

(LY294002), TAK1 ((5Z)-7-Oxozeaenol), JNK (SP600125), the lysosome (chloroquine), proteasome (MG132), and protein translation (cyclohexamide) were obtained from Sigma

Aldrich. Abs used in this study were: TRC105 (TRACON Pharmaceuticals), endoglin (H-300,

Santa Cruz Biotechnology), flag (Sigma-Aldrich), and β-actin (Sigma-Aldrich). The following abs were all purchased from Cell Signaling: total Beclin1 (no. 3738), total Atg5 (no. 12994), total

Atg12 (no. 4180), total LC3 A/B (no. 4108), total p62 (no. 5114), phospho-Smad1/5 (no. 9516),

total Smad1 (no. 6944), phospho-Smad2/3 (no. 9510), total Smad2 (no. 3122), total Smad2/3 (no.

8685), phospho-Akt (no. 13038), phospho-JNK (no. 9255), and GAPDH (no. 2118).

Immunofluorescence Eng+/+ ECs or Eng−/− ECs were grown overnight on coverslips and transiently transfected with appropriate constructs using Lipofectamine 2000 (Invitrogen) as described. 24–48 h following transfection, cells were washed with PBS and then fixed with 4% paraformaldehyde. Cells were permeabilized in 0.1% Triton X-100 in PBS for 3–10 min, then blocked with 5% bovine serum albumin in PBS containing 0.05% Triton X-100 for 20 min. All primary Abs were incubated at room temperature for 1 h unless noted otherwise. Flag-Smad2 expression was detected using anti-flag Ab. LC3 A/B Ab was used to detect autophagosome clusters. Following primary antibody incubation, cells were incubated with appropriate flurophore conjugated secondary antibodies (Alexa-Fluor) at room temperature for 30 min. Cells were co- stained with DAPI (Sigma) immediately before immunofluorescence microscopy analyses (Nikon

Eclipse Ti). Autophagy was quantified by counting the number of autophagosome positive cells and the number of autophagosome vesicles per cell. Statistical significance is presented as mean

± S.E.

Endothelial Tube Formation Eng+/+ ECs stably expressing non-targeting control vector and

Eng+/+ ECs stably expressing Smad2 knockdown (sh-Smad2) were plated on a 24-well plate coated with 200 μl of matrigel basement matrix (Corning) at 140,000 cells/well. 30 min following plating, growth medium was removed and 200 μl of matrigel was added. 30 min following the addition of matrigel, 500 μl of growth medium was added. Endothelial tubes were digitally imaged and quantified by counting the number of branches per node. Following imaging, endothelial tubes were lysed using 2× sample buffer and immunoblotted for Beclin1 and

LC3A/B.

Reverse Transcriptase, Real-time PCR Eng+/+ ECs were grown overnight on 6-cm dishes and

transiently transfected with appropriate constructs using Lipofectamine 2000 (Invitrogen) as described. 24 h following transfection, cells were treated with ALK5 inhibitor (SB431542, 30

μM) for 6 h. Total RNA was extracted from the cells with Trizol reagent (Invitrogen), and 2 μg

RNA was converted to cDNA through the use of High Capacity cDNA Reverse Transcription Kit

(Applied Biosystems). BECN1 was quantified by real-time reverse transcriptase PCR (Applied

Biosystems) using SYBR green assay reagent and gene specific primers (Forward: 5′-

GGCCAATAAGATGGGTCTGA-3′. Reverse: 5′-GCTGCACACAGTCCAGAAAA-3′). Primers for MMP-2 were: (Forward: CTTCGCTCGTTTCCTTCAAC) and (Reverse:

AGAGTGAGGAGGGGAACCATR). Relative amplification was quantified by normalizing gene-specific amplification to that of 18S rRNA (5′-GCTCTAGAATTACCACAGTTATC-3′) and Reverse (5′-AAATCAGTTATGGTTCCTTTGGT-3′) in each sample. Changes in mRNA

−ΔΔC abundance were calculated using 2( T) method. Quantitative PCR were run in triplicates.

Statistical significance is presented as mean ± S.E.

Chromatin Immunoprecipitation Assay Eng+/+ ECs stably expressing Smad2 knockdown were grown overnight in 10-cm dishes and transiently transfected with either control vector or Flag-

Smad2. 20 h following transfection, cells were serum starved using MCDB-131 medium

(Invitrogen) supplemented with 2 mM L-glutamine, 1 mM sodium pyruvate for 3 h and treated with 200 pM TGFβ for 30 min. Following treatment, formaldehyde was added directly to the media at a concentration of 0.75% v/v and rotated at room temperature for 10 min. Following rotation, glycine was added to the media to a final concentration of 125 mM and rotated at room temperature for 5 min. Cells were washed two times with PBS, harvested into a 50 ml tube by scraping, and centrifuged for 5 min at 1,000 × g. The resulting pellet was resuspended in 750 μl of ChIP lysis buffer. ChIP lysis buffer and other buffers were made according to the recipes provided by Abcam. Samples were sonicated using 20-s bursts 9 times at 30% power. After

sonication, samples were centrifuged for 30 s, 4 °C, 8,000 × g. Supernatants were harvested and immunoprecipitated using anti-flag (Sigma) and Protein G Plus-Agarose Suspension (Invitrogen) at 4 °C. Following immunoprecipitation, beads were washed three times with ChIP wash buffer and one time with ChIP final wash buffer. 120 μl of elution buffer was added to the beads and rotated for 15 min. at room temperature. Following elution, beads were centrifuged for 1 min at

2000 × g. Supernatants were harvested and incubated at 65 °C overnight.

Figure 4.1: Endoglin promotes autophagy. (A) Western analysis of autophagic markers in

Eng+/+ and Eng−/− MEECs, and in HUVECs transiently expressing a scrambled, non-targeting shRNA (control) or shEng.

continued

Figure 4.1 continued

(B) LC3 A/B staining in Eng+/+ and Eng−/− MEECs followed by the percentage of LC3 A/B positive cells and number of autophagosome vesicles in Eng+/+ and Eng−/− MEECs. Bars indicate mean ± S.E. *, p value = 1.1 × 10−6; **, p value = 0.01.

continued

Figure 4.1 continued

Figure 4.2: ALK5 inhibition enhances autophagy. (A) Western analysis of LC3 cleavage in

HUVECs treated with ALK1 inhibitor (LDN193189, 1 μM) and ALK5 inhibitor (SB431542, 30

μM) for 6 h and 12 h.

continued 86

Figure 4.2 continued

(B) Western analysis of LC3 cleavage and Beclin-1 levels in Eng+/+ and Eng−/− ECs treated with ALK1 inhibitor (LDN193189, 0.5 μM), ALK5 inhibitor (SB431542, 15 μM), PI3K inhibitor

(LY294002, 15 μM), TAK inhibitor ((5Z)-7-Oxozeaenol, 5 μM), and JNK inhibitor (SP600125,

15 μM) for 16 h.

continued 87

Figure 4.2 continued

(C) Western analysis of LC3 cleavage and Beclin-1 levels in Eng+/+ and Eng−/− ECs treated with ALK5 inhibitor (SB431542, 30 μM) for 2, 6, and 16 h.

Figure 4.3: Endoglin promotes autophagy through suppression of Smad2 activity. (A) Western analysis of LC3 I/II, p-Smad1/5/8 and Smad1 in HUVECs treated with TGF-β (100 pM) and

BMP-9 (1 nM) in the presence or absence of ALK1 inhibitor (LDN193189, 1 μM) and ALK5 inhibitor (SB431542, 30 μM).

continued 89

Figure 4.3 continued

(B) Western analysis of LC3 cleavage and Beclin-1 levels in Eng+/+ and Eng−/− ECs treated with TGF-β (100 pM) and BMP-9 (1 nM) in the presence or absence of ALK1 inhibitor

(LDN193189, 0.5 μM) and ALK5 inhibitor (SB431542, 15 μM).

continued 90

Figure 4.3 continued

continued

Figure 4.3 continued

(D) Immunofluorescence of Smad2, Smad3, and LC3II in Eng+/+ MEECs transiently transfected with Flag-Smad2/3 and stained for Flag (red) and endogenous LC3 II (green), and graph quantification (*, p value = 0.0007).

Figure 4.4: Inhibition of autophagic responses via TGF-β-induced Smad2 activation. (A)

Comparison of LC3 cleavage and Beclin-1 levels upon stable Smad2 knockdown (sh-Smad2) and transient Smad3 knockdown (sh-Smad3) in Eng+/+ MEECs and HUVECs.

continued

Figure 4.4 continued

(B) Western analysis of LC3 cleavage and Beclin-1 levels in Eng+/+ and Eng−/− ECs treated with increasing doses of TGF-β (50 pM, 200 pM) for 16 h, and densitometry quantification of the relative ratio of LC3 II to GAPDH for three independent experiments (*, p = 0.002).

continued 94

Figure 4.4 continued

Smad 2) in Eng+/+ ECs treated with increasing doses of TGFβ (50 pM, 200 pM) for 16 h, and densitometry quantification of the relative ratio of LC3 II to GAPDH for three independent experiments (*, p value = 0.001; **, p value = 0.04).

Figure 4.5: Smad2 is a BECN-1 transcriptional repressor. (A) Western analysis of Beclin-1 protein levels in Eng+/+ and Eng−/− ECs treated with MG132 (10 μM), chloroquine (100 μM), and cyclohexamide (3 μg/ml) for 6 h.

continued 96

Figure 4.5 continued

(B) Quantification of Beclin-1 mRNA (BECN1) in Eng+/+ and Eng−/− MEECs by quantitative

PCR analyzed by delta-delta-CT (ddCT) methods using 18S rRNA as internal control. Fold change was calculated by setting mean fractions of Eng+/+ MEECs as one. Bars indicate mean ±

S.E. *, p value = 0.0029. (C) Quantification of BECN1 in Eng+/+ MEECs transiently transfected with Flag-Smad2, two separate sh-Smad2 constructs, or treated with ALK5 inhibitor (SB431542,

15 μM) for 6 h. Fold change was calculated by setting mean fractions of non-treated Eng+/+

MEECs as one. Bars indicated mean ± S.E. *, p value<0.04; **, p value<0.01. continued

Figure 4.5 continued

(D) ChIP assay upon expression of Flag-Smad 2 or control pcDNA in Eng+/+ ECs treated with or without TGF-β for 30 min following 4 h of serum deprivation. The graph represents the relative binding by Smad2 to the specified DNA binding elements.

Figure 4.6: Smad2 suppression enhances endothelial capillary sprouting by upregulating Beclin-

1 levels and LC3 cleavage. (A) Representative images and quantification of average number of capillaries per node of three-dimensional Matrigel-induced capillary tubules upon stable Smad2 knockdown (sh-Smad2) in Eng+/+ ECs at 16 h. *, p value<0.0001. (B) Western analysis of LC3 cleavage and Beclin-1 levels in three-dimensional Matrigel-induced capillary tubules upon stable

Smad 2 knockdown (sh-Smad2) in Eng+/+ ECs.

continued

Figure 4.6 continued

100

Matrigel-induced capillary tubules in Eng+/+, Eng−/−, and Eng−/− transiently expressing Smad2 shRNA (sh-Smad2) ECs at 16 h. *, p value = 0.0006. (D) Western analysis of LC3 cleavage and

Beclin1 levels of three-dimensional Matrigel-induced capillary tubules in Eng+/+, Eng−/−, and

Eng−/− transiently expressing Smad2 shRNA (sh-Smad2) ECs and densitometry quantification of relative ratio of LC3 II to GAPDH for three independent experiments (*, p value = 0.0001; **, p value = 0.04).

101

Results

In order to test the effects of TGF-β signaling on autophagy in the vascular endothelium, we tested the role of the endothelial specific TGF-β co-receptor, endoglin, in autophagy utilizing two types of endothelial cells (EC) – endoglin knock-out (Eng -/-) and control MEECs and primary human umbilical vein EC (HUVEC). A variety of autophagy makers were used to assess changes in autophagy in Eng +/+ and Eng -/- ECs. Eng +/+ EC exhibited increased levels of Beclin-1,

Atg12, and LC3I to LC3II conversion; compared to Eng -/- ECs, suggesting that endoglin regulates basal autophagy (Figure IV-1A: MEEC). To further assess endoglin’s role in autophagy, we knocked down endogenous endoglin HUVECS, and observed a noticeable decrease in autophagy markers such as a Beclin-1, LC3II, and Atg5 whereas an increase in p62 indicated a decrease in autophagy (Figure IV-1A: HUVECS). To supplement our biochemical data, we performed immunofluorescence staining of LC3II in Eng +/+ and Eng -/- ECs. Eng +/+ ECs exhibited noticeably more autophagasome activity as indicated by the significantly greater percentage of LC3-positive cells and the average number of LC3 vesicles per cell (Figure IV-1B).

We previously demonstrated that TRC105, the first humanized monoclonal endoglin antibody for the treatment of metastatic tumors, neutralizes endoglin function by promoting its degradation and ectodomain cleavage [53]. Treatment with TRC105 neutralized endoglin’s ability to promote autophagy as shown by the decrease in Beclin-1 levels and LC3I conversion to LC3II, suggesting that endoglin is a key mediator of endothelial autophagy (Figure 1C).

Endoglin regulates the critical balance between the ALK1/Smad 1/5/8 and ALK5/Smad2/3 signaling pathways. To test these pathways as mediators of endothelial autophagy, we treated

HUVECS with ALK1 inhibitor and ALK5 inhibitor over the course of 12 h. to inactivate Smad

1/5/8 and Smad 2/3, respectively (Figure IV-2A). Interestingly, there was a significant increase in

LC3 II levels upon ALK5 inhibition and a subsequent Smad 2/3 inactivation over the course of 6

102

h. and 12 h., suggesting that ALK5-induced Smad 2/3 signaling may inhibit autophagy in ECs

(Figure IV-2A). Given that TGF-β signaling through ALK5 can mediate the activation of a number of different proteins that are regulate autophagy including PI3K/AKT, TAK1, and JNK, we tested their involvement in endothelial autophagy by treating Eng +/+ and Eng -/- MEECs with inhibitors against these proteins for 16 h. (Figure IV-2B). Consistent with our results in

HUVECs, greater levels of LC3 II and Beclin1 were observed in Eng +/+ ECs relative to Eng -/-

MEECs under no treatment and upon ALK5 inhibition (Figure IV-2B). Surprisingly, JNK and

TAK1 inhibition resulted in minimal changes, suggesting that these autophagy inducers have a minor role in endothelial cells. As expected, PI3K inhibition resulted in an increase in LC3II and

Beclin1 levels, conforming its canonical inhibitory role in autophagy through AKT and mTOR signaling (Figure IV-2B). ALK5 inhibition time-course revealed a greater level of Beclin1 expression and LC3 cleavage in Eng +/+ ECs compared to Eng -/- ECs (Figure IV-2C). Taken together, our results suggest that endoglin regulates ALK5/Smad2/3 activity to promote endothelial autophagy.

TGF-β can activate both ALK1/Smad1/5/8 and ALK5/Smad2/3. While endoglin does not directly participate in their activation, it modulates the balance of these two opposing pathways, favoring

ALK1/Smad1/5/8 signaling and suppressing ALK5/Smad2/3 signaling through TGF-β and

BMP9. To test the role of TGF-β and BMP9 in endothelial autophagy, HUVECs were treated with TGF-β and BMP9 in the presence or absence of ALK1 or ALK5 inhibitor (Figure IV-3A).

LC3II levels increased only upon ALK5 inhibition irrespective of TGF-β or BMP-9, suggesting that autophagy is Smad 1/5/8-independent. To determine endoglin’s role, we treated Eng +/+ and

Eng -/- ECs with TGF-β or BMP-9 in the presence or absence of ALK5 inhibitor. Basally, Eng

+/+ ECs displayed significantly higher levels of LC3II and Beclin1. In addition, irrespective of

TGF-β and BMP-9, ALK5 inhibition resulted in a significant increase in LC3 II levels, suggesting that endoglin promotes autophagy by suppressing Smad 2/3 activation (Figure IV-3B). 103

To test this, we analyzed relative basal levels of Smad 1/5 and Smad2 activation in Eng +/+ and

Eng -/- ECs (Figure IV-3C). As expected, there was more Smad 2/3 activation and less Smad 1/5 activation in Eng -/- ECs compared to Eng +/+ ECs (Figure 3C). Interestingly, immunofluorescence studies revealed a noticeable loss in LC3II staining in Smad2, but not

Smad3, overexpressing Eng +/+ ECs, suggesting that Smad2 is the major inhibitor endothelial autophagy (Figure IV-3D). Similar results were observed in HUVECs and MEECs, as knockdown of Smad2, but not Smad3, resulted in an increase in Beclin1 and LC3II levels (Figure

IV-4A). Next we analyzed the effects of TGF-β on autophagy by treating Eng +/+ and Eng -/-

ECs with increasing doses of for 16 h. A dramatic concentration-dependent decrease in Beclin1 and LC3II levels were observed with TGF-β treatment (Figure IV-4B). To determine whether

TGF-β’s autophagy inhibition was through Smad2, we treated Eng +/+ ECs and Eng +/+ ECs stably expressing Smad2 shRNA with increasing concentrations of TGF-β. As expected, TGF-β induced a significant loss of beclin1 and LC3II in control, but not in Smad2 knockdown ECs, indicating that Smad2 is the key regulator of endothelial autophagy (Figure IV-4C).

Given the link between endoglin and beclin1, we hypothesized that endoglin regulates beclin1 expression either at the protein level or transcriptionally. To test whether endoglin regulates beclin1’s turnover, Eng +/+ and Eng -/- ECs were treated with MG132 and chloroquine to block protein degradation by the proteasome and lysosome, respectively. In addition, ECs were also treated with cyclohexamide to block protein translation. Beclin1 levels remained relatively constant after MG132 and chloroquine treatment, suggesting that endoglin has minimal impact on beclin1’s stability (Figure IV-5A). Treatment with cyclohexamide, however, resulted in a significant decrease in beclin1 levels, suggesting that endoglin control beclin1 biosynthesis

(Figure IV-5A). To further support this, we measured BECN1 levels in Eng +/+ and Eng -/- ECs using qPCR. Eng +/+ ECs displayed 5-fold higher BECN1 levels compared to Eng -/- ECs

(Figure IV-5B). To determine whether Eng +/+ promotes Beclin1 transcription by impeding 104

Smad2 function, we measured BECN1 levels following Smad2 overexpression, transient Smad2 knockdown using two different shRNA sequences, and ALK5 inhibition in Eng +/+ ECs. As expected, ALK5 inhibition resulted in an increase in BECN1 levels compared to control (Figure

IV-5C). More importantly, knockdown of Smad2 using two different shRNA targeting sequences resulted in a dramatic increase in BECN1 levels compared to control whereas Smad2 overexpression had relatively no effect on BECN1 levels compared to control, suggesting that

Smad2 inhibits BECN1 transcription (Figure IV-5C). To determine whether Smad2 serves as a direct BECN1 transcriptional repressor, we performed chromatin immunoprecipitation (ChIP) assay in sh-Smad2 ECs transiently expressing Flag-Smad2 or pcDNA control (Figure IV-5D).

Cells were stimulated with or without TGF-β then immunoprecipitated with Flag antibody prior to RT-PCR analysis using primers that covered up to 10 kb upstream of BECN1 start site while

MMP-2 was used as a control [83]. There was a 10-fold increase in Smad2 binding near the promoter region of MMP2 after TGF-β treatment, which is consistent with the previous finding that MMP2 is a target gene for Smad2 [83]. Importantly, we identified that Smad2 binds to DNA elements spanning ∼6.6 kb upstream (approximately -7729 to -1076). However, no Smad2 interaction was observed in the first 1.1 kb upstream (-1090 to -25) of the BECN1 start site or (-

9950 to -7700) of BECN1. Overall, our results indicated that Smad2 is a direct transcriptional repressor of BECN1 by binding upstream of the BECN1 promoter region (Figure IV-5D).

Lastly, we tested the role of Smad2’s regulation of Beclin1 in angiogenesis. To test this, we performed matrigel-induced capillary tubule assay in Eng +/+ ECs and Eng +/+ ECs stably expressing Smad2 shRNA. Smad2 knockdown produced greater capillary branching compared to control Eng +/+ MEECs, supporting a novel inhibitory function for Smad2 in angiogenesis

(Figure IV-6A). To demonstrate that Smad2’s inhibitory role was through suppression of autophagy, we measured the levels of Beclin1 and LC3I/II in control and sh-Smad2 ECs from cell

105

lysates derived directly from the matrigel. sh-Smad2 ECs displayed higher LC3I/II conversion and Beclin1 levels compared to control ECs (Figure IV-6A and B). To further test Smad2’s contribution to angiogenesis, we knocked down Smad2 in Eng -/- MEECs and observed a significant rescue in endothelial capillary tubes and branching as well as increased LC3I/II and

Beclin1 levels compared to control Eng -/- MEECs (Figure IV-6C and D and graph quantification). Overall, these findings define a link between autophagy and angiogenesis and establish a novel mechanism by which endoglin promotes angiogenesis.

Discussion

The role of autophagy in angiogenesis is highly controversial, as some studies have demonstrated that induction of autophagy promotes angiogenesis whereas others have shown the opposite. In the present study, we examined the role of endoglin mediated TGF-β signaling in autophagy and how this relationship regulates angiogenesis. While TGF-β control numerous mediators of autophagy under different cellular contexts, our pharmacologic screening results indicate that many of previously reported mediators of autophagy such as TAK1, JNK, and Smad 1/5 had little impact on endothelial autophagy. Surprisingly, we identified that ALK5 signaling acts as a critical determinant of endothelial autophagy. In addition, we also identified that endoglin suppresses ALK5 signaling to induce autophagy by disrupting Smad2’s inhibition of BECN1.

Previous studies have demonstrated that Smad2 inhibits angiogenesis primarily through inhibition of endothelial cell proliferation. Additional anti-angiogenic mechanisms for Smad2 have not been well defined. We believe that our findings represent an important new mechanism by which

Smad2 impairs angiogenesis.

Endoglin binds to TGF-β with high affinity to activate ALK1-induced Smad 1/5/8 signaling. By binding to TGF-β, endoglin sequesters TGF-β away from TGF-βRII and ALK5 thereby indirectly inhibiting ALK5-induced Smad2/3 signaling [67]. Given a recent study that showed that Smad

106

1/5 promotes autophagy, in hepatocarcinoma cells, we hypothesized that endoglin promotes autophagy through a similar mechanism. However, ALK/Smad1/5 had minimal effects on autophagy and stimulation with BMP-9, a more potent activator of ALK1/Smad1/5 compared to

TGF-β, had negligible effects on autophagy (Figure 2 and 3). Smad1/5 promotes endothelial migration and proliferation by up regulating Id-1 gene expression, which negatively regulates basic helix-loop-helix transcription factors by forming heterodimers and inhibiting their DNA binding and transcriptional activity. ALK5, on the other hand, promotes the expression of PAI-1, which inhibits endothelial cell proliferation and migration [84]. ALK5’s anti-proliferative signals are predominantly through Smad2/3. Aside from PAI-1, Smad2/3 mediates its anti-angiogenic effects by inhibiting the expression of cell cycle promoters such as c-Myc and cylinD1 and promoting the expression of cell cycle inhibitors such as p15 and p21 [85]. Our current study defines an additional yet novel mechanism by which Smad2 suppresses angiogenesis through the inhibition of autophagy.

Interestingly, studies in hepatocellular carcinoma and breast cancer cells have shown that TGF-β activates autophagy by enhancing mRNA expression levels BECLIN1, ATG5, ATG7, and death- associated protein kinase (DAPK). More importantly, knockdown of Smad2/3 as well as Smad4 suppressed TGF-β induced autophagy. However, whether these Smads play a direct role in transcriptional regulation of autophagy related proteins has not been demonstrated. In the EC system, our results provide compelling evidence that Smad2 acts as a BECN1 transcriptional repressor by binding to upstream of the BECN1 promoter region. Whether Smad2 directly interacts with the canonical Smad binding element or requires additional co-factors remains to be determined.

Aside from inhibiting Smad2 function, endoglin also engage in other TGF-β related pathways to regulated endothelial autophagy. Although we show that endoglin expression correlates with

107

autophagy, we failed to see a proportional increase in autophagy upon rescue of endoglin in Eng

-/- MEECs. Rather, we observed a further reduction in autophagy with endoglin overexpression

(data not shown). While this was unexpected, it suggests that endoglin’s role in autophagy is more complex than anticipated. This is consistent with our previous findings that show endoglin enhances BMP-9 induced Akt activation. Considering that Akt is a potent inhibitor of autophagy in most cell types, endoglin overexpression may have led to Akt hyperactivation and suppression of autophagy. Collectively, our work defines a novel role for endoglin in endothelial autophagy and identifies Smad2 as a novel transcriptional repressor of BECN1. Lastly, we demonstrate that endoglin-mediated autophagy enhances capillary tube formation, thereby providing a new functional mechanism by which endoglin promotes angiogenesis.

108

Chapter 5: Antibody-directed Coupling of Endoglin and MMP-14 is a key Mechanism for Endoglin Shedding and Deregulation of TGF-β Signaling

Abstract

Endoglin is a TGF-β superfamily co-receptor that plays a critical role in regulating normal and tumor-associated angiogenesis. Given its essential role angiogenesis, endoglin serves as an ideal therapeutic vascular target in human cancer. A number of endoglin ectodomain-targeting antibodies have shown clinical efficacy in suppressing tumor-associated angiogenesis; however, their underlying molecular mechanisms remain poorly understood. In the present study, we define a key mechanism by which TRC105 neutralizes endoglin’s function. TRC105, along with several other endoglin antibodies, enhanced endoglin shedding by coupling endoglin with the membrane- type 1 matrix metalloproteinase (MMP)-14 at the cell surface, resulting in the release of the anti- angiogenic factor, soluble endoglin (sEng). TRC105 treatment abrogated both stead-state and

TGF-β-induced Smad1/5/8 activation while enhancing Smad 2/3 activation. Interestingly,

TRC105 augmented endothelial migration by enhancing stress fiber formation and disrupting cell-to-cell junction. Overall, our study defines endoglin shedding and deregulated TGF-β signaling as major mechanisms by which TRC105 neutralizes endoglin function and inhibits angiogenesis.

Background

Angiogenesis is the formation of new blood vessels from pre-existing blood vessels. While it plays a central role in normal growth and development, angiogenesis also serves as a critical step for the growth and spread of cancers. Tumors require blood supply to grow beyond a few

109

millimeters in size. Increased blood innervation facilitates oxygen and nutrient delivery as well as

waste removal, allowing tumors to grow and metastasize to distant organs. Given its dependency on angiogenesis, reducing tumor vascularization is a promising strategy in limiting cancer progression. Endoglin is a TGF-β co-receptor that is emerging as a promising vascular target in anti-cancer therapy. Numerous studies have shown that endoglin is required for both normal and tumor-associated angiogenesis and serves as a gold standard biomarker for tumor vascularization and survival rate. A growing number of endoglin monoclonal antibodies are currently under evaluation for their clinical efficacy; however their underlying mechanisms are poorly defined.

The critical role of endoglin in angiogenesis is well defined, as endoglin knockout mice die during midgestation due to impaired cardiovascular development of the yolk sac and target inhibition of endoglin expression impairs tumor development and vascularization. Endoglin regulates angiogenesis by mediating two canonical TGF-β signaling pathways: ALK1 mediated

Smad1/5/8 activation and ALK5 mediated Smad 2/3 activation. TGF-β can drive both pro- and anti-angiogenic response depending on its association with either ALK1 or ALK5. Binding of

TGF-β or the structurally related TGF-β ligand, BMP9, to ALK1 promotes pro-angiogenic responses by activating Smad1/5/8 transcription responses whereas TGF-β binding to ALK5 suppresses angiogenesis through Smad2/3 activation. Although the precise mechanisms by which

TGF-β and BMP-9 regulate angiogenesis are under active investigation, endoglin is considered a critical component in mediating the balance between ALK1 and ALK5 signaling to exert either pro- or anti-angiogenic signals.

Cell surface receptor expression and function are often regulated through ectodomain shedding.

Ectodomain shedding converts membrane-associated proteins into soluble effectors and reduces the receptor cell surface expression. A recent study has demonstrated that endoglin undergoes ectodomain shedding through cleavage by MMP-14, resulting in the release of a soluble form of

110

endoglin (sEng) in to the circulation. Although membrane-bound endoglin promotes angiogenesis, sEng antagonizes this process and acts a potent anti-angiogenic factor by sequestering BMP-9, but not TGF-β, from binding to cell surface endoglin or ALK1. Overall, endoglin shedding has a critical role in regulating in TGF-β signaling during angiogenesis.

A number of endoglin-targeting antibodies have shown pre-clinical efficacy in several solid tumors by inducing the regression of tumor size and inhibiting metastasis and formation of new tumors. Of particular interest is TRC105; the first humanized monoclonal antibody currently in phase I/II clinical trials for the treatment of advanced or metastatic solid tumors. While TRC105 has shown immense promise in limiting tumor growth and progression, its underlying molecular mechanism has not been defined. Therefore, in the present study, we examined the mechanistic basis for how TRC105 and other endoglin-targeting antibodies inhibit angiogenesis.

Materials and Methods

Cell culture, transfection and antibodies HUVEC were purchased from Lonza and maintained in endothelial growth medium (EGM)-2 growth medium (Lonza, Walkersville, MD, USA). HUVEC from passages 2 to 9 were used for experiments. HMEC-1 were maintained in MCDB-131 medium (Invitrogen, Grand Island, NY, USA), supplemented with 10% fetal bovine serum,

1 µg/ml hydrocortisone (Sigma), 10 ng/ml epidermal growth factor (Sigma, St Louis, MO, USA) and 2 mm L-glutamine. COS-7 cells were maintained in DMEM with 10% fetal bovine serum.

Lipofectamine 2000 was used to transfect COS-7 cells as described according to manufacturer's protocol (Invitrogen). HMEC-1 were either transfected with Lipofectamine or Amaxa 4D nucleofection system according to manufacturer's protocol. Primary Abs used in this study were:

TRC105 (TRACON Pharmaceuticals, San Diego, CA, USA), endoglin (H-300, Santa Cruz

Biotechnology, Dallas, TX, USA), endoglin P4A4, P3D1, and PECAM (University of Iowa

Hybridoma, Iowa city, IA, USA), HA-Ab (Roche, Applied Sciences, Indianapolis, IN, USA),

111

MMP-14 (Abcam, Cambridge, MA, USA) and β-actin (Sigma-Aldrich, St Louis, MO, USA). The following Abs were all purchased from Cell Signaling (Danvers, MA, USA): total Smad2/3 (no.

8685), total Smad 1 (no. 6944) and phospho-Smad2/3 (no. 9510), phospho-Smad 1/5/8 (no.

9511), VE-Cadherin (no. 2500), total SAPK/JNK (no. 9258), phospho-SAPK/JNK (no. 9665), caspase-3 (no. 9665) and caspase-9 (no. 9508).

Immunopreciptation Cells were washed with PBS, then lysed on ice with lysis buffer (20 mm

HEPES, pH 7.4, 150 mm NaCl, 2 mm EDTA, 10 mm NaF, 10% (w/v) glycerol, 1% Nonidet NP-

40) and supplemented with protease inhibitors (Sigma protease inhibitor cocktail) and phosphatase inhibitors (Sigma phosphatase inhibitor cocktail). The lysates were precleared by centrifugation and incubated with appropriate Abs and protein agarose G for 4–6 h at 4 °C. The immunoprecipitates were collected by centrifugation; pellets were washed with lysis buffer, and stored in 2 × sample buffer before western blot analyses. For immunoprecipitation of sEng from the conditioned media, typically media from 10-cm plates of HUVEC and HMEC-1 were collected and concentrated by Amicon Ultra Centrifugal Filters (Millipore, Billerica, MA, USA) before immunoprecipitation with appropriate Abs.

Reverse transcriptase, real-time PCR and shRNAs Total RNA was extracted from the cells with

Trizol reagent (Invitrogen), and 2 µg RNA was then converted to cDNA by using the High

Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Grand Island, NY, USA). Gene expression of MMP-14 was quantified by real-time reverse transcriptase PCR (Applied

Biosystems, Step One) using the SYBR green assay reagent and gene-specific primer (Forward

5′-CCTGCCTGCGTCCATCA-3′ and Reverse 5′-TCCAGGGACGCCTCATCA-3′). Relative amplification was quantified by normalizing the gene-specific amplification to that of 18s rRNA

Forward (5′-GCTCTAGAATTACCACAGTTATC-3′) and Reverse (5′-

AAATCAGTTATGGTTCCTTTGGT-3′) in each sample. The specific products were confirmed

112

by SYBR green single-melting curve and a single correct-size product when running in 2% agarose gel. Changes in mRNA abundance were calculated using 2(−ΔΔCT) method.

Quantitative PCR were run in triplicate. Statistical significance is presented as mean±s.d. The knockdown of human endoglin in HMEC-1 was achieved by transfecting cells with a shRNA vector purchased from Sigma (Mission shRNA) (5′-

CCGGCCACTTCTACACAGTACCCATCTCGAGATGGGTACTGTGTAGAAGTGGTTTTTG

-3′) and stably selected using puromycin.

Immunofluorescence HMEC-1s and HUVECs grown on coverslips were serum starved for 3–4 h, washed with PBS and then fixed with 4% paraformaldehyde. Cells were permeabilized in 0.1%

Triton X-100 in PBS for 4 min, then blocked with 5% bovine serum albumin in PBS containing

0.05% Triton X-100 for 20 min. All primary Abs were incubated at room temperature for 1 h unless noted otherwise. Cytochrome c was detected using cytochrome c Ab (Invitrogen). VE- cadherin (Cell Signaling) and PECAM Abs (University of Iowa Hybridoma) were used to detect endothelial cell–cell junctions. Actin was detected using Alexa Fluor-conjugated phalloidin for

30 min. Wherever appropriate, cells were co-stained with DAPI (Sigma) immediately before immunofluorescence microscopy analyses. For immunofluorescence co-patching, cells expressing endoglin and MMP-14 were washed with cold PBS, blocked with 5% bovine serum albumin in PBS before 1 h incubation with appropriate Abs (TRC105 and Myc or HA) at 4 °C to allow cell surface labeling while preventing internalization. Cells were then washed three times with cold PBS, then incubated with secondary fluorophore-conjugated Abs (for example, anti- human for TRC105 and anti-mouse for HA or Myc-Ab) for 30 min at 4 °C. To analyze the co- patching, Pearson correlation coefficient software

(http://www.wessa.net/rwasp_correlation.wasp/) was used with ImageJ

(http://rsbweb.nih.gov/ij/index.html) to calculate the degree of co-localization of two

113

fluorophores.

Transwell migration assays Cells were seeded in the upper chamber of a transwell filter in growth media, coated both at the top and bottom with gelatin to assess cell migration. Cells were allowed to migrate for 16 h at 3 °C through the gelatin-coated toward the lower chamber containing growth media with TRC105 (0.02–2 μg/ml). Migrated cells on the bottom surface of the filter were fixed, stained and then digitally imaged before counting.

Cell surface biotinylation assay Cells were washed briefly with cold PBS before biotinylation with membrane-impermeable biotinylation reagent (Sulfo-NHS-LC-Biotin, Thermo Scientific,

Pierce, Rockford, IL, USA) according to manufacturer protocol. The biotinylation reaction was neutralized and cells were washed three times with cold PBS, then lysed and prepared for immunoprecipitation with MMP-14 (HA-Ab). The immunoprecipitated MMP-14 along with interacting proteins were resolved on SDS–PAGE and immunoblotted with streptavidin- horseradish peroxidase.

114

Figure 5.1: TRC105 promotes steady-state and ligand-induced Smad2/3 activation while inhibiting Smad1/5/8 signaling. Western blot analyses show endogenous expression of phosphorylated (p-) Smad2/3 (activated), total Smad2/3, p-Smad1/5/8 and total Smad1 in cell lysate following a 24 h treatment with TRC105 (0–2 μg/ml) in HMEC-1 (A) and HUVEC (B). β- actin levels indicate equal loading of lysates in both cell types.

continued 115

Figure 5.1 continued

β1 (100 pM) and TRC105 (0.2 μg/ml) plus TGF-β (100 pM) for 30 min. Lower panels show total

Smad1 and Smad2/3 from the same cell lysate. (D) Western blot shows Smad1/5/8 activation in response to BMP-9 treatment (30 min) in HUVEC in the presence or absence of TRC105

(0.2 μg/ml) for 30 min. Data are representative of at least three independent experiments.

116

Figure 5.2: TRC105 does not induce endothelial growth arrest or apoptosis. (A) MTT assay showing the HUVEC growth pattern following treatment with either control IgG or TRC105

(2 μg/ml) for 12, 24, 48 h (left graph). A parallel MTT assay showing the effects of control and stable endoglin depletion through shRNA (shEng) in HMEC-1 (right graph). *P=0.001 comparing control versus shEng at 24 h.

continued

117

Figure 5.2 continued

(B) HMEC-1 cells were treated with different concentrations of TRC105 (0–2 μg/ml) and TGF-

β1 (400 pM) for 24 h, and assessed for cytochrome c release via immunofluorescence. Shown are representative images of cells treated with TRC105 and TGF-β1. Arrow identifies a cell from which cytochrome c was released. Data are mean±s.d. of at least 30 cells counted for each condition (**P<0.01 compared with control 0 μg/ml).

continued 118

Figure 5.2 continued

(C) The western blot shows caspase-9 and -3-cleaved products upon treatment of HMEC with either IgG control or TRC105 (2 μg/ml) for 16 h.

119

Figure 5.3: Endoglin-targeting Abs enhance sEng production. (A) Western blot of sEng immunoprecipitated from conditioned media (Abs P3D1 or H-300) after 24 h treatment with

TRC105 (0–2 μg/ml) in HMEC-1. Lower panels show expression of full-length endoglin, MMP-

14 and β-actin in cell lysates. All of the western blots are representative of at least four independent experiments.

continued 120

Figure 5.3 continued

(B) Western blot of sEng immunoprecipitated from conditioned media (Abs P3D1 or H-300) after

24 h treatment with TRC105 (0–2 μg/ml) in HUVEC. Lower panels show expression of full- length endoglin, MMP-14 and β-actin in cell lysates. All of the western blots are representative of at least four independent experiments.

continued 121

Figure 5.3 continued

(C) Endogenous sEng immunoprecipitated from the conditioned media upon 24 h treatment with indicated Abs directed against different epitopes on endoglin extracellular domain (200 ng/ml).

Lower panels show expression of full-length endoglin, MMP-14 and β-actin in cell lysates. All of the western blots are representative of at least four independent experiments.

122

Figure 5.4: TRC105 enhances endoglin/MMP-14 association at the cell surface. (A) Western blot analysis shows co-immunoprecipitation of endoglin and MMP-14. COS-7 cells expressing HA- tagged MMP-14 (MMP), endoglin (Eng), MMP and Eng in the presence or absence of TRC105

(200 ng/ml) pretreatment for 10 min. Lysates were immunoprecipitated with MMP-14-specific

Ab and immunoblotted for co-immunoprecipitated endoglin, MMP-14, total endoglin, and β- actin.

continued 123

Figure 5.4 continued

(B) Western blot analysis shows co-immunoprecipitation of Eng and MMP-14 following immunoprecipitation with HA-Ab (MMP-14).

continued

124

Figure 5.4 continued

(MMP), Eng, MMP and Eng in the presence or absence of TRC105 pretreatment (200 ng/ml) at

4 °C. Cell lysates were prepared following biotinylation, and MMP-14 was immunoprecipitated with HA-Ab. The western blot shows biotinylated endoglin and MMP-14 detected with streptavidin-horseradish peroxidase (HRP). Data are representative of at least three independent experiments.

continued

125

Figure 5.4 continued

(D) HUVECs were treated with TRC105 alone (200 ng/ml), MMP-14 blocking Ab alone

(2 μg/ml), or co-treated for 24 h. Graph represents a densitometry analysis of sEng present in each condition as a ratio compared with no treatment. Each error bar represents the s.d. derived from three independent experiments normalized to no treatment. *P<0.001, **P=0.0014, ***P<0.001.

126

Figure 5.5: TRC105 couples endoglin/MMP-14 into co-patched clusters at the cell surface. (A)

Green immunofluorescence indicates cell surface TRC105-mediated clusters of endoglin. (B) Red immunofluorescence indicates cell surface HA-Ab-mediated clusters of MMP-14. (C) Shown is the overlay image of endoglin/MMP-14 clusters. (D) Enlarged overlay images of endoglin/MMP-

14 co-patched clusters in COS-7 cells treated with TRC105, or overlay images of MMP-14 and

Myc-tagged ALK3 in cells probed with MMP-14 and anti-Mycspecific Abs, respectively. (E)

Graph represents the quantification co-patched clusters of endoglin/MMP-14 or ALK3/MMP-14

(random). Cells were treated with TRC105 and HA-Ab (MMP-14), TRC105 and MMP-14- specific Ab, or P3D1 (endoglin) and MMP-14-specific Ab. For random quantification, Myc and

MMP-14-specific Abs were used to induce patching. % Co-localization was quantified based on

Pearson correlation coefficient measurements of at least 20 cells per condition. *P<0.01,

**P<0.04 compared with control (random) 127

Figure 5.6: TRC105 promotes MMP-14 gene expression in HUVEC. (A) Cells treated with

TRC105 (200 ng/ml) for 24 h were quantified by SYBR green based quantitative PCR and analyzed by delta-delta-CT (ddCT) methods using18S rRNA as internal control. Fold changes were calculated by setting the mean fractions of untreated cells as one. Bars indicate mean±s.d. in cells from TRC105-treated and -untreated cells.

continued 128

Figure 5.6 continued

F (B) Graph shows the effect of the ALK5 inhibitor (SB431542, 5 µm) and/or TRC105 (200 ng/ml) for 24 h on MMP-14 gene expression in HUVEC. Inset of western blot shows endogenous expression of sEng immunoprecipitated from the conditioned media of the same cells that were used to isolate RNA for MMP-14 gene expression study.

continued

129

Figure 5.6 continued

(SP600125, 5 μm) and TRC105 with JNK inhibitor.

continued

130

Figure 5.6 continued

(D) Western blot analysis shows TRC105 concentration-dependent phosphorylation (activation) of JNK (upper panel) with t (lower panel) as loading control. *P<0.05, **P<0.001.

131

Figure 5.7: TRC105 and sEng induce cell migration. (A) HUVEC and (B) HMEC-1 were plated on transwells coated with 0.02% gelatin and treated with TRC105 (0 to 2 μg/ml) in growth media for 16 h. Cells that migrated to the bottom side of the membrane were fixed, stained for their nuclei and imaged for counting by using ImageJ software..

continued

132

Figure 5.7 continued

(C) The graph represents the effect of TRC105 treatment (200 ng/ml with or without MMP-14- neutralizing Ab (1:500 v/v)) treatment on migration of HMEC-1.

continued 133

Figure 5.7 continued

(D) The graph shows migration of HMEC-1 expressing the secreted endoglin extracellular domain (Eng-ECD), treated with or without TRC105 or MMP-14 Ab. *P<0.05, **P<0.001,

***P<0.08. Relative cell migration is represented as a percentage compared with untreated cells, from triplicates for each of the three independent experiments. The error bars indicate the s.e.m. of the percentage of the cells that migrated.

134

Figure 5.8: TRC105 induces stress fiber formation and dissolution of endothelial cell–cell junctions. (A) Alexa-phalloidin staining shows the detection of actin stress fiber organization in

HMEC-1 (A, B) and HUVEC (C, D) in control versus treatment with TRC105 (0.2 μg/ml) for 16 to 24 h.

continued 135

Figure 5.8 continued

(B) HUVEC were grown to near confluent monolayer (∼80–90%) before either no treatment (I–

III) or TRC105 treatment (IV–VI) (0.2 μg/ml) for 16–24 h. Cells were fixed, and then co-stained for VE-cadherin (green), PECAM (red) and DAPI (blue).

136

Results

To determine the mechanisms by which TRC105 inhibits angiogenesis, we used two types of human endothelial cells, a human microvascular endothelial cell line (HMEC-1) and primary human umbilical vein endothelial cells (HUVECs). We first examined the effects of TRC105 on steady state TGF-β signaling. 24 h. treatment with TRC105 in growth media caused a significant enhancement in Smad 2/3 activation, whereas Smad 1/5/8 activation was dramatically reduced in a concentration-dependent manner (Figure V-1A and B). We next tested whether TRC105 affected ligand responsiveness by subjecting cells to serum deprivation, followed by a brief pretreatment with TRC105 (10 min) before stimulation with either TGF-β or BMP-9.

Interestingly, TRC105 treatment abrogated TGF-β-induced Smad 1/5/8 activation, whereas its effect on Smad 2/3 was enhanced (Figure V-1C). Given that TGF-β requires ALK1/endoglin complex to induce Smad 1/5/8 activation, these results suggest that TRC105 prevents endoglin from interacting with TGF-β and allows greater access of TGF-β to ALK5, leading to increased

Smad 2/3 activation. In the case of BMP-9, TRC105 had relatively no effect on BMP-9 induced

Smad 1/5/8 activation (Figure V-1D). Taken together, these results suggest that TRC105 inhibits angiogenesis by inhibiting TGF-β, but not BMP-9, signaling and this inhibition favors activation of Smad 2/3 over Smad 1/5/8.

To determine the cellular effects of TRC105, we tested its impact on endothelial cell growth through MTT assay. We have previously reported that endoglin inhibits endothelial growth, partially through suppression of ERK activation and its downstream targets, c-Myc and cyclinD1.

Interestingly, TRC105 had minimal effect on cell growth compared to control over the course of

48 h. (Figure V-2A). Endoglin depletion in HMEC-1 resulted in enhanced ERK activation and enhanced c-Myc compared to control, supporting our previous findings that endoglin inhibits endothelial cell growth though suppression of ERK and its downstream targets, c-Myc and

137

cyclinD1, whereas treatment with TRC105 had minimal effects. Aside from endothelial cell proliferation, we also analyzed whether TRC105 had an effect on cell apoptosis through cytochrome c release and caspases cleavage. Consistent with our cell proliferation data, TRC105 had no significant effect on cytochrome c release compared to untreated cells (∼3–5%) (Figure

V-2B). In comparison, TGF-β, a known inducer of apoptosis, resulted in 25-30% cytochrome c release. In addition, there was no detectable difference in caspase cleavage relative to control IgG

(Figure V-2C), indicated that TRC105 has no direct effect on endothelial cell growth or apoptosis.

Although previous studies have established sEng as an antiangiogenic factor, whether endoglin targeting antibodies regulate sEng production is not known. To test whether TRC105 has a role sEng production, we measured endogenous sEng production in the conditioned media of

HUVECs and HMEC-1 after TRC105 treatment for 24 h. TRC105 treatment resulted in a concentration-dependent increase in sEng in both cell types, whereas the endogenous cellular endoglin remained constant (Figure V-3A). Given that TRC105 promotes endoglin shedding rather than degradation, we explored the possibility that TRC105 may regulate MMP-14 activity.

MMP14 expression was markedly increased with concentrations of TRC105 (Figure V-3A and

B). To determine whether this was an epitope specific effect, we tested whether other endoglin- targeting antibodies resulted in an increase in sEng. Similar to TRC105, all endoglin-targeting antibodies tested significantly enhanced sEng production compared to no treatment (Figure V-

3C), suggesting that sEng production is a general mechanism for all endoglin-targeting antibodies. Moreover, three of the four antibodies tested, including TRC105, promoted MMP-14 expression relative to control (Figure V-3C).

To define the mechanism of antibody-induced endoglin shedding, we examined the potential interaction between MMP-14 and endoglin. We hypothesized that endoglin-targeting antibodies

138

promote the coupling of MMP-14 and endoglin at the cell surface. To test this, cells expressing myc-tagged endoglin and HA-tagged MMP-14 were pretreated with TRC105 at 4 °C to allow

TRC015 to react with membrane-localized endoglin while preventing their endocytosis. Basally, there was very little interaction between endoglin and MMP-14 at the cell surface. However, after

TRC105 treatment there was a dramatic increase in MMP-14 and endoglin interaction at the cell surface, suggesting that TRC105 couples MMP-15 and endoglin at the cell surface (Figure V-4A and B). To exclude the possibility of TRC105 non-specifically targeting MMP-14, cells expressing MMP-14 alone or MMP-14 and endoglin were treated pre-treated with TRC105 and then immunoprecipitated with TRC105. As expected, TRC105 failed to react with MMP-15 irrespective of pretreatment and only co-immunoprecipitated MMP-14 when endoglin was co- expressed. To specifically examine whether the antibody-induced coupling occurred at the cell surface, cells expressing MMP-14 and endoglin were pretreated with or without TRC105 at 4 °C before cell surface biotinylation. Following biotinylation, cells were immunoprecipitated with

HA-MMP-14 and immunoblotted for biotin-streptavidin. Consistent with our previous results, endoglin/MMP-14 complex was virtually basally undetectable at the cell surface. However, after

TRC105 pretreatment, biotinylated endoglin/MMP-14 complex was detected, confirming

TRC105 action at the cell surface (Figure V-4C). Finally, to demonstrate that endoglin shedding requires MMP-14 catalytic activity, we used an MMP-14 blocking antibody (α-MMP-14) that neutralizes its catalytic function. Although TRC105 treatment alone dramatically enhanced endoglin shedding relative to basal state, inhibition of MMP-14 catalytic activity resulted in only a modest increase in sEng production (Figure V-4D). The slight increase in sEng production is most likely due to indirect proteolytic actions of other MMPs. More importantly, MMP-14 catalytic inhibition reduced TRC105’s increase in endoglin shedding (Figure V-4D). Taken together, these biochemical studies suggest that TRC105 induces endoglin shedding by promoting its interaction with MMP-14 for proteolytic cleavage.

139

To supplement our biochemical approaches, we employed immunofluorescence co-patching to visualize TRC105-induced cell surface clustering of the endoglin/MMP-14 complex. In control experiments, we expressed another TGF-β superfamily receptor, myc-tagged ALK3, and HA- tagged MMP-14. Each protein was allowed to aggregate into patched clusters by probing with its respective primary and then fluorophore conjugated secondary antibodies at 4 °C before fixation.

The antibody-induced aggregation yielded mostly distinct green and red patches, indicating that

TRC105 did not induce aggregation of ALK3 and MMP-14 at the cell surface (Figure V-5D and

E). In contrast, endoglin and MMP-14 co-expression followed by immunoprobiing with TRC105 and anti-HA or MMP-14 specific antibodies resulted in significant co-patching and co- localization (Figure V-5A-C) compared to the random control (60 versus 10%) regardless of whether HA or MMP-14-specific antibodies was used (Figure V-5E). To test whether or not this coupling was TRC105 specific or a general endoglin-antibody-direct effect, we probed cells with

P3D1 antibody. Similar to TRC105, P3D1 also resulted in s-Eng production (Figure V-3C). P3D1 and MMP-14-specific antibodies yielded 40% co-patching (Figure V-5E), suggesting a general role of endoglin-targeting antibodies in coupling endoglin and MMP-14 at the cell surface. Based on our biochemical and immunofluorescence co-patching experiments, endoglin-targeting antibodies mediate endoglin/MMP-14 coupling at the cell surface to promote endoglin proteolytic cleavage

Given that TRC105 also promotes MMP-14 expression (Figure V-3) to potentiate endoglin shedding, we measured the effect of TRC105 on MMP-14 gene expression. We found that

TRC105 enhanced MMP-14 gene expression 1.5 fold compared to control (Figure V-6A). To elucidate the mechanism by which TRC105 enhances MMP-14 gene expression, we inhibited

ALK5-induced Smad2/3 activation as TGF-β has been previously shown to regulate MMP-14 gene expression through Smad2/3 in other cell types. Contrary to expectations, Smad2/3

140

inhibition markedly enhanced MMP-14 gene expression, suggesting that TRC105-induced MMP-

14 expression is Smad2/3 independent. Next, we screened small molecular inhibitors in order to identify potential signaling effectors mediating this process. Inhibition of JNK resulted in a decrease in MMP-14 mRNA level and blocked TRC105-induced MMP-14 expression (Figure V-

6B). Consistent with this finding, there was a noticeable concentration-dependent increase in JNK activation by TRC105 (Figure V-6D), supporting a novel mechanism by which TRC105 promotes MMP-14 gene expression through JNK activation. s-Eng is a well-established antiangiogenic factor in vivo. However, its underlying molecular and cellular mechanisms have not been fully elucidated. Our previous data suggests that s-Eng acts independently of growth inhibition and apoptosis to inhibit angiogenesis. Therefore, we tested whether TRC105 and s-Eng production disrupt cell motility. HUVEC and HMEC-1 were treated with either low or high concentrations of TRC105 and allowed to migration in a transwell system.

Interestingly, TRC105 enhanced endothelial cell migration in a concentration-dependent manner

(Figure V-7A and B). This is consistent with previous findings demonstrating endoglin’s anti- migratory role using endoglin knockout and knockdown systems. Next, given that MMP-14 is required for migration in many cell types, we examined whether TRC105 enhances migration through MMP-14 by treating cells with TRC105 in the presence or absence of MMP-14 neutralizing antibody. Consistent with previous results, MMP-14 neutralizing antibody reduced migration compared to control. More importantly, TRC105-induced migration was reduced in the presence of MMP-14 inhibition (Figure V-7C), suggesting that MMP-14 activity mediates

TRC105-induced migration. Finally, to specifically test the role of s-Eng in cell motility, we compared the motility of cells treated with TRC105 to cells expressing a secreted form of endoglin in conditioned media (Eng-ECD: extracellular domain). As predicted, expression of

Eng-ECD enhanced cell migration, similar to that of TRC105 treatment. In addition, the effects of

141

Eng-ECD were not significantly reduced even in the presence of MMP-14 neutralizing antibody

(Figure V-7D).

To determine the mechanisms by which increased migration might reduce angiogenesis, we examined several cellular properties including actin cytoskeleton organization and epithelial- mesenchymal transition. Given that α-smooth muscle actin (α-SMA) is a marker of endothelial- to-mesenchymal transition, we tested for α-SMA expression in HUVEC and HMEC-1. Although we couldn’t detect α-SMA expression through immunofluorescence or immunoblotting (data not shown), we did notice a striking difference in actin stress fiber formation in HMEC-1 and

HUVECs treated with TRC105 (Figure V-8A). TRC105 enhanced stress fiber formation compared to control cells, suggesting that the enhanced contractile forces from stress fibers likely contribute to the increase in migration. In addition to stress fibers, we also measured how

TRC105 influences endothelial cell-cell contacts during maturation phase of angiogenesis. Near confluent monolayer of HUVECs were treated with or without TRC105 for 12-24 h. and then stained with endothelial-specific adheren junction markers (Figure V-8B). Untreated cells formed a complete monolayer and exhibited prominent vascular endothelial (VE)-cadherin and platelet endothelial cell adhesion (PECAM) staining at the cell-cell junctions. TRC105 treated cells, however, exhibited reduced VE-cadherin and PECAM localization along cell membranes (Figure

V-8B). Taken together, our data suggests that TRC105 perturbs normal endoglin regulation of endothelial migration and endothelial cell-cell junctions.

Discussion

Endoglin shedding is an important regulatory process in angiogenesis and vascular homeostasis.

Shedding reduces the cell surface levels of endoglin and promotes the production of sEng, a potent antiangiogenic factor, which binds and sequesters circulating BMP-9. The discovery that of MMP-14 as a major protease response for endoglin shedding has raised an important question

142

as to whether MMP-14’s role in endoglin shedding is actively regulated or acts as a general housekeeping mechanism. Our current study suggests that endoglin-targeting antibodies have a significant role in endoglin shedding, which is now supported clinically. Results from the first human phase I trial at Duke University revealed that out of the 37 plasma-based protein biomarkers tested, there is a dramatic dose increase in sEng levels in patients treated with

TRC105.

Our data indicates that TRC105 induces endoglin shedding through two mechanisms. Through the use of various biochemical and immunofluorescence techniques, we identified that TRC105 couples MMP-14 and endoglin at the cell surface, leading to the production of sEng. This process likely involves the stabilization of preformed endoglin and MMP-14 complex, as our initial experiments demonstrate that TRC105 specifically targets endoglin and does not directly react with MMP-14. Secondly, our small-molecule kinase inhibitor treatment demonstrated that

TRC105 induced MMP-14 gene expression through JNK activation rather than through Smad2/3 activation. This was quite surprising; as TGF-β mediated Smad signaling is a known inducer of

MMP-14 expression and treatment with TGF-β resulted in a dose-dependent increase in Smad2/3 activation. Instead, the rapid increase in Smad2/3 activation may contribute to the observed pro- migratory phenotype as Smad2/3 is a known regulator of mediators of migration, including PAI-

1. Given that ALK5 can promote cell migration through activation of pro-migratory factors such as TGF-β-activated kinase (TAK1) and JNK, our data is also consistent with the role of TRC105 in stimulating cell migration through ALK5/JNK-induced stress fiber formation. Endoglin forms heteromeric complexes with not only ALK1 and ALK5 but also TGF-β superfamily receptors,

ALK3 and ALK6, to regulate vascular homeostasis and promote angiogenesis [24, 72, 86].

Whether endoglin-targeting antibodies such as TRC105 alter these heteromeric complexes remains to be seen.

143

Although it’s clear that endoglin-targeting antibodies facilitate the release of endoglin’s extracellular domain, the fate of its intracellular domain is unknown. As a co-receptor, endoglin’s cytoplasmic domain lacks catalytic function. However, it serves as an important docking site for a number of adaptor proteins including zyxin, zyxin-related protein 1, β-arrestin2, Tctex2β and

GIPC [38, 40, 87-89]. Given that these proteins differentially regulate endoglin trafficking and signaling during angiogenesis, it is critical that we understand how endoglin-targeting antibodies affect these interactions.

Our data suggests that TRC105’s effect on endoglin is much more complex than anticipated.

Given that TRC105 had no effect on ERK activation, cell proliferation, or c-Myc expression suggests that TRC105 alters, rather than neutralizes, endoglin function. This notion is consistent with our data showing that endoglin expression is not altered even though a subpopulation of endoglin undergoes ectodomain shedding. Another interesting, yet surprising, result was that

TRC105 had no effect on apoptosis and cell proliferation, which are cell processes regulated by

TGF-β signaling through the ALK5/Smad2/3 pathway. These findings contradict pre-clinical data in which TRC105 induced apoptosis in HUVECs [90]. Although our results show no such effects, a noticeable difference between the two studies is the markedly higher concentration of TRC105 used to induce apoptosis (50–100 μg/ml) in preclinical studies compared with our experimental conditions (0.02–2 μg/ml). Preclinical pharmacokinetic studies suggest that TRC105 binds to

human endoglin with a Kd of 4.6 ng/ml (31 pm), and achieves saturation at 200 ng/ml in proliferating HUVEC. Although we also did observe growth inhibition and apoptosis at significantly higher doses (500–1000 μg/ml), these effects were relatively small (∼10%) and statistically similar to that of IgG control (data not shown). Still there are discrepancies that cannot be explained, such as the recent study demonstrating that high doses of TRC105 (6 μg/ml or higher) blocks BMP-9-induced Smad 1/5/8 activation in HUVEC [91]. In contrast, our data

144

indicate little to no effects on BMP-9/Smad1/5/8 activation either at low (0.2 μg/ml) or high concentrations (10 μg/ml; data not shown).

The role of endoglin in endothelial migration is highly controversial, as earlier studies have demonstrated that endoglin and ALK1 together promote migration whereas others have reported the opposite [67, 86, 87]. While these disparities can be partially explained by differences in experimental conditions, it is notable that in these studies sEng’s role had not been examined. Our data suggests that varying degrees of sEng could contribute to these disparate findings and that s-

Eng may have a key role in fine-tuning cell motility.

In summary, we have examined the multifunctional role of TRC105 to provide clinically relevant mechanistic insights for current and existing anti-angiogenic therapies. Our results demonstrate a novel role of TRC105 in enhancing sEng production to deregulate normal TGF-β signaling, endothelial migration, and angiogenesis.

145

Chapter 6: Conclusion

The notion that tumors require angiogenesis to grow and metastasize has sparked interest in identifying novel angiogenic targets. Current FDA-approved anti-angiogenic agents that inhibit

VEGF signaling have yielded only moderate results, as patients initially respond favorably but suffer serious toxicity and develop resistance to treatment over time. In light of this, new therapeutics that target signaling pathways essential for tumor angiogenesis are of interest in the treatment of tumor growth and metastasis. Endoglin has emerged as an attractive therapeutic target as it is required for tumor-associated angiogenesis and is strongly correlated with tumor progression, survival rate, and metastasis. Despite endoglin’s critical role in normal and tumor- associated angiogenesis, the underlying molecular mechanisms by which endoglin promotes angiogenesis are not fully defined. Understanding these molecular aspects will provide valuable insight as to how targeting endoglin therapeutically reduces tumor growth and progression.

The ultimate goal of this thesis was to identify new mechanisms of endoglin signaling during angiogenesis.

Angiogenesis is a highly regulated process that involves the coordination of numerous steps.

Determining the molecular basis of how each of these steps is regulated is critical for understanding the pathophysiology of tumor progression and how targeting these steps may therapeutically reduce tumor-associated angiogenesis. The role of endoglin in endothelial proliferation is highly controversial, as some have reported a pro-proliferative role for endoglin whereas others have reported the opposite. Given that endoglin can signal through Smad- independent pathways to regulate endothelial biology, we examined whether endoglin signals

146

through Smad-independent pathways to regulate endothelial proliferation. In Chapter I,

wedemonstrate that endoglin augments endothelial cell growth by suppressing ERK activation and its downstream targets, cyclinD1 and c-Myc, through its interaction with β-arrestin2. Our study is the first to link endoglin to ERK-mediated endothelial proliferation and defines a novel

Smad-independent mechanism by which endoglin augments endothelial growth. More importantly, our study provides mechanistic evidence in support of targeting endoglin as a therapeutic approach for the treatment of solid cancers.

Despite endoglin’s established role in tumor-associated angiogenesis and its potential in tumor therapy, the underlying mechanisms for its turnover are poorly understood. Uncovering these mechanisms will provide clinically relevant insight as to how endoglin function and expression can be negatively regulated to suppress tumor angiogenesis and tumor progression. Therefore, in

Chapter II, we identified a novel mechanism by which endoglin undergoes endogenous down- regulation. Specifically, we demonstrated that endoglin’s conserved tyrosine motif serves as a key determinant for its stability and angiogenesis. Prior to this, the majority of studies investigating regulatory mechanisms for endoglin expression were primarily at the transcriptional level. One study, however, did correlate endoglin turnover with TNF-α, but the underlying mechanism for this was never elucidated, and the authors did not link the decrease in endoglin expression to any functional outcome. Overall, the data presented in Chapter II provides critical information on the structure and function of endoglin and provides clinically relevant insight as to how endoglin can be post-translationally regulated for the treatment of many solid tumors.

Another aspect of endoglin biology that is poorly understood is its role in autophagy. Growing evidence suggests that autophagy is a novel therapeutic target whose modulation presents new avenues for cancer treatment. While several clinically used drugs inhibit autophagy, most of these drugs lack anti-tumor effects and specificity, resulting in immense off-target effects. Therefore,

147

identification of new regulators of autophagy as well as more potent and specific inhibitors of autophagy is needed. Therefore, in Chapter III, we investigated the role of endoglin in mediating endothelial autophagy during angiogenesis and demonstrated that endoglin promotes endothelial autophagy by relieving Smad2’s repression of Beclin-1 gene transcription. Our study demonstrates a novel mechanism by which endoglin promotes angiogenesis and provides additional mechanistic insight supporting endoglin’s therapeutic potential in the treatment of tumor progression and metastasis.

TRC105 is a novel antibody against endoglin that has completed multiple phase I and phase II clinical trials as a single agent for the treatment of a number of solid tumor types. Although

TRC105 has shown good tolerability, high clinical efficacy, and minimal side effects, its underlying mechanism of action is unknown. Recent evidence has shown that endoglin undergoes ectodomain shedding through a direct interaction with membrane-localized MMP-14, resulting in the release of the anti-angiogenic factor, s-Eng. In Chapter IV, we investigated whether

TRC105’s anti-angiogenic effects are through a similar mechanism. Our results indicate that

TRC105 promotes endoglin ectodomain shedding in a dose-dependent manner by coupling endoglin and MMP-14 at the cell surface and enhancing MMP-14 gene expression through JNK activation. s-Eng binds BMP-9 and -10 ligands and prevents their interaction with type II receptors, thereby blocking multiple signaling complexes at once and providing an advantage over other therapeutic agents that only target one component of the TGF-β pathway. Aside from ectodomain shedding, TRC105 also alters endoglin’s function by promoting its degradation. In

Chapter II, we demonstrate that TRC105 promotes endoglin tyrosine phosphorylation by inducing

Src activation, resulting in the internalization and degradation of endoglin through the lysosomal degradative pathway. Collectively, our results indicate that TRC105 has multifunctional roles in modulating endoglin function and offer mechanistic insight for the improvement of existing and future antiangiogenic therapies. 148

Taken together, the extensive work presented in this thesis provides a greater understanding of endoglin biology, defines novel signaling mechanisms by which endoglin engages in to promote angiogenesis, and provides substantial evidence highlighting endoglin’s therapeutic potential as an anti-angiogenic target for the treatment of many human malignancies.

149

REFERENCES

1. Otrock, Z.K., et al., Understanding the biology of angiogenesis: review of the most important molecular mechanisms. Blood Cells Mol Dis, 2007. 39(2): p. 212-20. 2. Potente, M., H. Gerhardt, and P. Carmeliet, Basic and therapeutic aspects of angiogenesis. Cell, 2011. 146(6): p. 873-87. 3. Nishida, N., et al., Angiogenesis in cancer. Vasc Health Risk Manag, 2006. 2(3): p. 213-9. 4. Nishida, T., Angiogenesis, which is essential for cancer growth, is a diagnostic and therapeutic target. J Gastroenterol, 2005. 40(3): p. 320-1. 5. Wang, Z., et al., Broad targeting of angiogenesis for cancer prevention and therapy. Semin Cancer Biol, 2015. 35 Suppl: p. S224-43. 6. Carmeliet, P., Mechanisms of angiogenesis and arteriogenesis. Nat Med, 2000. 6(4): p. 389-95. 7. Korpanty, G., E. Smyth, and D.N. Carney, Update on anti-angiogenic therapy in non-small cell lung cancer: Are we making progress? J Thorac Dis, 2011. 3(1): p. 19-29. 8. Rosen, L.S., et al., Endoglin for targeted cancer treatment. Curr Oncol Rep, 2014. 16(2): p. 365. 9. Kubota, Y., Tumor angiogenesis and anti-angiogenic therapy. Keio J Med, 2012. 61(2): p. 47-56. 10. Bisacchi, D., et al., Anti-angiogenesis and angioprevention: mechanisms, problems and perspectives. Cancer Detect Prev, 2003. 27(3): p. 229-38. 11. Gordon, K.J. and G.C. Blobe, Role of transforming growth factor-beta superfamily signaling pathways in human disease. Biochim Biophys Acta, 2008. 1782(4): p. 197-228. 12. Pardali, E. and P. Ten Dijke, TGFbeta signaling and cardiovascular diseases. Int J Biol Sci, 2012. 8(2): p. 195-213. 13. Goumans, M.J., Z. Liu, and P. ten Dijke, TGF-beta signaling in vascular biology and dysfunction. Cell Res, 2009. 19(1): p. 116-27. 14. Bourdeau, A., D.J. Dumont, and M. Letarte, A murine model of hereditary hemorrhagic telangiectasia. J Clin Invest, 1999. 104(10): p. 1343-51. 15. Oh, S.P., et al., Activin receptor-like kinase 1 modulates transforming growth factor-beta 1 signaling in the regulation of angiogenesis. Proc Natl Acad Sci U S A, 2000. 97(6): p. 2626-31. 16. Yang, X., et al., Angiogenesis defects and mesenchymal apoptosis in mice lacking SMAD5. Development, 1999. 126(8): p. 1571-80.

150

17. Wrighton, K.H., X. Lin, and X.H. Feng, Phospho-control of TGF-beta superfamily signaling. Cell Res, 2009. 19(1): p. 8-20. 18. Goumans, M.J., et al., Activin receptor-like kinase (ALK)1 is an antagonistic mediator of lateral TGFbeta/ALK5 signaling. Mol Cell, 2003. 12(4): p. 817-28. 19. Xu, P., J. Liu, and R. Derynck, Post-translational regulation of TGF-beta receptor and Smad signaling. FEBS Lett, 2012. 586(14): p. 1871-84. 20. Massague, J., J. Seoane, and D. Wotton, Smad transcription factors. Genes Dev, 2005. 19(23): p. 2783-810. 21. Pepper, M.S., Transforming growth factor-beta: vasculogenesis, angiogenesis, and vessel wall integrity. Cytokine Growth Factor Rev, 1997. 8(1): p. 21-43. 22. Wang, W., D. Rigueur, and K.M. Lyons, TGFbeta signaling in cartilage development and maintenance. Birth Defects Res C Embryo Today, 2014. 102(1): p. 37-51. 23. Cheifetz, S., et al., Endoglin is a component of the transforming growth factor- beta receptor system in human endothelial cells. J Biol Chem, 1992. 267(27): p. 19027-30. 24. Barbara, N.P., J.L. Wrana, and M. Letarte, Endoglin is an accessory protein that interacts with the signaling receptor complex of multiple members of the transforming growth factor-beta superfamily. J Biol Chem, 1999. 274(2): p. 584-94. 25. Ten Dijke, P., et al., Regulation of cell proliferation by Smad proteins. J Cell Physiol, 2002. 191(1): p. 1-16. 26. Goumans, M.J., et al., Balancing the activation state of the endothelium via two distinct TGF-beta type I receptors. EMBO J, 2002. 21(7): p. 1743-53. 27. Dong, C., et al., Upregulation of PAI-1 is mediated through TGF-beta/Smad pathway in transplant arteriopathy. J Heart Lung Transplant, 2002. 21(9): p. 999-1008. 28. Isogai, C., et al., Plasminogen activator inhibitor-1 promotes angiogenesis by stimulating endothelial cell migration toward fibronectin. Cancer Res, 2001. 61(14): p. 5587-94. 29. Baudino, T.A., et al., c-Myc is essential for vasculogenesis and angiogenesis during development and tumor progression. Genes Dev, 2002. 16(19): p. 2530-43. 30. Volpert, O.V., et al., Id1 regulates angiogenesis through transcriptional repression of thrombospondin-1. Cancer Cell, 2002. 2(6): p. 473-83. 31. Zhang, W. and H.T. Liu, MAPK signal pathways in the regulation of cell proliferation in mammalian cells. Cell Res, 2002. 12(1): p. 9-18. 32. Murphy, D.A., et al., Inhibition of tumor endothelial ERK activation, angiogenesis, and tumor growth by sorafenib (BAY43-9006). Am J Pathol, 2006. 169(5): p. 1875-85. 33. Zhang, Y.E., Non-Smad pathways in TGF-beta signaling. Cell Res, 2009. 19(1): p. 128-39. 34. McCubrey, J.A., et al., Roles of the Raf/MEK/ERK pathway in cell growth, malignant transformation and drug resistance. Biochim Biophys Acta, 2007. 1773(8): p. 1263-84. 151

35. Galliher, A.J. and W.P. Schiemann, Src phosphorylates Tyr284 in TGF-beta type II receptor and regulates TGF-beta stimulation of p38 MAPK during breast cancer cell proliferation and invasion. Cancer Res, 2007. 67(8): p. 3752-8. 36. Kretzschmar, M., J. Doody, and J. Massague, Opposing BMP and EGF signalling pathways converge on the TGF-beta family mediator Smad1. Nature, 1997. 389(6651): p. 618-22. 37. Matsuura, I., et al., Identification and characterization of ERK MAP kinase phosphorylation sites in Smad3. Biochemistry, 2005. 44(37): p. 12546-53. 38. Lee, N.Y. and G.C. Blobe, The interaction of endoglin with beta-arrestin2 regulates transforming growth factor-beta-mediated ERK activation and migration in endothelial cells. J Biol Chem, 2007. 282(29): p. 21507-17. 39. Li, D.Y., et al., Defective angiogenesis in mice lacking endoglin. Science, 1999. 284(5419): p. 1534-7. 40. Lee, N.Y., et al., Endoglin regulates PI3-kinase/Akt trafficking and signaling to alter endothelial capillary stability during angiogenesis. Mol Biol Cell, 2012. 23(13): p. 2412-23. 41. Guerrero-Esteo, M., et al., Extracellular and cytoplasmic domains of endoglin interact with the transforming growth factor-beta receptors I and II. J Biol Chem, 2002. 277(32): p. 29197-209. 42. Jonker, L., TGF-beta & BMP receptors endoglin and ALK1: overview of their functional role and status as antiangiogenic targets. Microcirculation, 2014. 21(2): p. 93-103. 43. Ray, B.N., et al., ALK5 phosphorylation of the endoglin cytoplasmic domain regulates Smad1/5/8 signaling and endothelial cell migration. Carcinogenesis, 2010. 31(3): p. 435-41. 44. Castonguay, R., et al., Soluble endoglin specifically binds bone morphogenetic proteins 9 and 10 via its orphan domain, inhibits blood vessel formation, and suppresses tumor growth. J Biol Chem, 2011. 286(34): p. 30034-46. 45. Hawinkels, L.J., et al., Matrix metalloproteinase-14 (MT1-MMP)-mediated endoglin shedding inhibits tumor angiogenesis. Cancer Res, 2010. 70(10): p. 4141-50. 46. Huang, F. and Y.G. Chen, Regulation of TGF-beta receptor activity. Cell Biosci, 2012. 2: p. 9. 47. Sanchez-Elsner, T., et al., Endoglin expression is regulated by transcriptional cooperation between the hypoxia and transforming growth factor-beta pathways. J Biol Chem, 2002. 277(46): p. 43799-808. 48. Nassiri, F., et al., Endoglin (CD105): a review of its role in angiogenesis and tumor diagnosis, progression and therapy. Anticancer Res, 2011. 31(6): p. 2283-90. 49. Li, C., et al., TNF alpha down-regulates CD105 expression in vascular endothelial cells: a comparative study with TGF beta 1. Anticancer Res, 2003. 23(2B): p. 1189-96. 50. Svagzdys, S., et al., Microvessel density as new prognostic marker after radiotherapy in rectal cancer. BMC Cancer, 2009. 9: p. 95.

152

51. Bergers, G. and D. Hanahan, Modes of resistance to anti-angiogenic therapy. Nat Rev Cancer, 2008. 8(8): p. 592-603. 52. Fonsatti, E. and M. Maio, Highlights on endoglin (CD105): from basic findings towards clinical applications in human cancer. J Transl Med, 2004. 2(1): p. 18. 53. Kumar, S., et al., Antibody-directed coupling of endoglin and MMP-14 is a key mechanism for endoglin shedding and deregulation of TGF-beta signaling. Oncogene, 2014. 33(30): p. 3970-9. 54. Paauwe, M., et al., Endoglin targeting inhibits tumor angiogenesis and metastatic spread in breast cancer. Oncogene, 2016. 55. Gordon, M.S., et al., An open-label phase Ib dose-escalation study of TRC105 (anti-endoglin antibody) with bevacizumab in patients with advanced cancer. Clin Cancer Res, 2014. 20(23): p. 5918-26. 56. Murrow, L. and J. Debnath, Autophagy as a stress-response and quality-control mechanism: implications for cell injury and human disease. Annu Rev Pathol, 2013. 8: p. 105-37. 57. Viry, E., et al., Autophagy: an adaptive metabolic response to stress shaping the antitumor immunity. Biochem Pharmacol, 2014. 92(1): p. 31-42. 58. Mathew, R., V. Karantza-Wadsworth, and E. White, Role of autophagy in cancer. Nat Rev Cancer, 2007. 7(12): p. 961-7. 59. Glick, D., S. Barth, and K.F. Macleod, Autophagy: cellular and molecular mechanisms. J Pathol, 2010. 221(1): p. 3-12. 60. Kaur, J. and J. Debnath, Autophagy at the crossroads of catabolism and anabolism. Nat Rev Mol Cell Biol, 2015. 16(8): p. 461-72. 61. Luo, S. and D.C. Rubinsztein, Apoptosis blocks Beclin 1-dependent autophagosome synthesis: an effect rescued by Bcl-xL. Cell Death Differ, 2010. 17(2): p. 268-77. 62. Yang, X., et al., The role of autophagy induced by tumor microenvironment in different cells and stages of cancer. Cell Biosci, 2015. 5: p. 14. 63. Zarzynska, J.M., Two faces of TGF-beta1 in breast cancer. Mediators Inflamm, 2014. 2014: p. 141747. 64. Mizumura, K., et al., The emerging importance of autophagy in pulmonary diseases. Chest, 2012. 142(5): p. 1289-99. 65. Belkacemi, L. and S.X. Zhang, Anti-tumor effects of pigment epithelium-derived factor (PEDF): implication for cancer therapy. A mini-review. J Exp Clin Cancer Res, 2016. 35(1): p. 4. 66. Huang, S.S. and J.S. Huang, TGF-beta control of cell proliferation. J Cell Biochem, 2005. 96(3): p. 447-62. 67. Lebrin, F., et al., Endoglin promotes endothelial cell proliferation and TGF- beta/ALK1 signal transduction. EMBO J, 2004. 23(20): p. 4018-28. 68. Pardali, E., M.J. Goumans, and P. ten Dijke, Signaling by members of the TGF- beta family in vascular morphogenesis and disease. Trends Cell Biol, 2010. 20(9): p. 556-67. 69. Park, S.O., et al., ALK5- and TGFBR2-independent role of ALK1 in the pathogenesis of hereditary hemorrhagic telangiectasia type 2. Blood, 2008. 111(2): p. 633-42. 153

70. Pece-Barbara, N., et al., Endoglin null endothelial cells proliferate faster and are more responsive to transforming growth factor beta1 with higher affinity receptors and an activated Alk1 pathway. J Biol Chem, 2005. 280(30): p. 27800-8. 71. Lebrin, F., et al., TGF-beta receptor function in the endothelium. Cardiovasc Res, 2005. 65(3): p. 599-608. 72. Bernabeu, C., J.M. Lopez-Novoa, and M. Quintanilla, The emerging role of TGF- beta superfamily coreceptors in cancer. Biochim Biophys Acta, 2009. 1792(10): p. 954-73. 73. Perez-Gomez, E., et al., The role of the TGF-beta coreceptor endoglin in cancer. ScientificWorldJournal, 2010. 10: p. 2367-84. 74. Ehrlich, M., et al., Oligomeric interactions of TGF-beta and BMP receptors. FEBS Lett, 2012. 586(14): p. 1885-96. 75. Bernabeu, C., B.A. Conley, and C.P. Vary, Novel biochemical pathways of endoglin in vascular cell physiology. J Cell Biochem, 2007. 102(6): p. 1375-88. 76. Tang, H., et al., Thrombin induces endocytosis of endoglin and type-II TGF-beta receptor and down-regulation of TGF-beta signaling in endothelial cells. Blood, 2005. 105(5): p. 1977-85. 77. Okada, M., Regulation of the SRC family kinases by Csk. Int J Biol Sci, 2012. 8(10): p. 1385-97. 78. Lam, K.S., J. Wu, and Q. Lou, Identification and characterization of a novel synthetic peptide substrate specific for Src-family protein tyrosine kinases. Int J Pept Protein Res, 1995. 45(6): p. 587-92. 79. Schliess, F., et al., Involvement of integrins and Src in insulin signaling toward autophagic proteolysis in rat liver. J Biol Chem, 2004. 279(20): p. 21294-301. 80. Park, S.S., et al., Involvement of c-Src kinase in the regulation of TGF-beta1- induced apoptosis. Oncogene, 2004. 23(37): p. 6272-81. 81. Lamalice, L., F. Le Boeuf, and J. Huot, Endothelial cell migration during angiogenesis. Circ Res, 2007. 100(6): p. 782-94. 82. He, C. and D.J. Klionsky, Regulation mechanisms and signaling pathways of autophagy. Annu Rev Genet, 2009. 43: p. 67-93. 83. Phanish, M.K., et al., The differential role of Smad2 and Smad3 in the regulation of pro-fibrotic TGFbeta1 responses in human proximal-tubule epithelial cells. Biochem J, 2006. 393(Pt 2): p. 601-7. 84. Ota, T., et al., Targets of transcriptional regulation by two distinct type I receptors for transforming growth factor-beta in human umbilical vein endothelial cells. J Cell Physiol, 2002. 193(3): p. 299-318. 85. Shimoe, M., et al., Overexpression of Smad2 inhibits proliferation of gingival epithelial cells. J Periodontal Res, 2014. 49(3): p. 290-8. 86. Lopez-Novoa, J.M. and C. Bernabeu, The physiological role of endoglin in the cardiovascular system. Am J Physiol Heart Circ Physiol, 2010. 299(4): p. H959-74. 87. Conley, B.A., et al., Endoglin controls cell migration and composition of focal adhesions: function of the cytosolic domain. J Biol Chem, 2004. 279(26): p. 27440-9. 154

88. Sanz-Rodriguez, F., et al., Endoglin regulates cytoskeletal organization through binding to ZRP-1, a member of the Lim family of proteins. J Biol Chem, 2004. 279(31): p. 32858-68. 89. Meng, Q., et al., Identification of Tctex2beta, a novel dynein light chain family member that interacts with different transforming growth factor-beta receptors. J Biol Chem, 2006. 281(48): p. 37069-80. 90. Tsujie, M., et al., Anti-tumor activity of an anti-endoglin monoclonal antibody is enhanced in immunocompetent mice. Int J Cancer, 2008. 122(10): p. 2266-73. 91. Nolan-Stevaux, O., et al., Endoglin requirement for BMP9 signaling in endothelial cells reveals new mechanism of action for selective anti-endoglin antibodies. PLoS One, 2012. 7(12): p. e50920.

155