In Vivo Roles of Endothelial TGF-Beta Receptors in Vascular Development and Malformations

Permanent Link: http://ufdc.ufl.edu/UFE0042741/00001

Material Information

Title: In Vivo Roles of Endothelial TGF-Beta Receptors in Vascular Development and Malformations
Physical Description: 1 online resource (132 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011


Physiology and Pharmacology (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Abstract: Haploinsufficiency of the TGF-beta type I and type III receptors, Activin receptor-like kinase-1 (ALK1) and Endoglin (ENG), cause the autosomal dominantly-inherited vascular disorder Hereditary Hemorrhagic Telangiectasia (HHT), which is characterized by vascular lesions, such as telangiectases and visceral arteriovenous malformations (AVMs). Current treatment for HHT treat the symptoms, thus understanding the pathological mechanism is essential for creating more effective or preventative treatment. The central dogma of the HHT field is that ALK1 and ENG work in concert to propagate TGF-? signals in a linear SMAD-dependent manner during angiogenesis and to maintain vascular integrity. Thus, AVMs form when either is disrupted; however, this relationship has not been clearly defined in vivo. Thus, two Eng conditional knockout (cKO) mouse models were generated and characterized using the same two cre-deleter lines (endothelial-specific L1cre and tamoxifen-inducible R26-CreER, iKO) previously used against Alk1, in which mice consistently developed AVMs. Although, many of the same organs developed AVMs or displayed AV-shunting in Eng iKO and Alk1 iKO models, phenotypes in Eng iKO mice varied widely in terms of development, severity and location. The results indicate that there may be differing pathogenetic mechanisms underlying the two major types of HHT. Additionally, there is still debate over what roles, if any, the TGF-beta type II and type I receptors (TGFBR2 and ALK5) play in endothelial cells (ECs). Contradictions in previous in vivo and in vitro data indicate that TGF-? superfamily members have spatial and temporal roles in ECs. Thus, Alk5 and Tgfbr2 were deleted specifically in ECs using a novel ALK1 cre-knockin deleter line (Alk1GFPcre) at E9.5. Alk1GFPcre;Alk5(fl/fl) and Alk1GFPcre;Tgfbr2(fl/fl) mice were embryonic lethal by E13.5 and E14.5, respectively. Mice displayed specific cerebral hemorrhages at E11.5 as previously seen in knockout models of alpha(v) integrin, alpha(v)beta(8) and Tgfb1/3. Thus, it was suggested paracrine neuroepithelium-specific integrin-mediated activation of TGF-beta signaling within ECs is essential for the midgestational establishment of the cerebrovasculature.Taken together, the data presented provide resourceful insight in the formation of vascular lesions in vivo. Our mouse models produced more consistent phenotypes than currently available mouse models and will be vital to future studies in HHT.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by HA-LONG PHUOC NGUYEN.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Oh, Suk P.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042741:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042741/00001

Material Information

Title: In Vivo Roles of Endothelial TGF-Beta Receptors in Vascular Development and Malformations
Physical Description: 1 online resource (132 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011


Physiology and Pharmacology (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Abstract: Haploinsufficiency of the TGF-beta type I and type III receptors, Activin receptor-like kinase-1 (ALK1) and Endoglin (ENG), cause the autosomal dominantly-inherited vascular disorder Hereditary Hemorrhagic Telangiectasia (HHT), which is characterized by vascular lesions, such as telangiectases and visceral arteriovenous malformations (AVMs). Current treatment for HHT treat the symptoms, thus understanding the pathological mechanism is essential for creating more effective or preventative treatment. The central dogma of the HHT field is that ALK1 and ENG work in concert to propagate TGF-? signals in a linear SMAD-dependent manner during angiogenesis and to maintain vascular integrity. Thus, AVMs form when either is disrupted; however, this relationship has not been clearly defined in vivo. Thus, two Eng conditional knockout (cKO) mouse models were generated and characterized using the same two cre-deleter lines (endothelial-specific L1cre and tamoxifen-inducible R26-CreER, iKO) previously used against Alk1, in which mice consistently developed AVMs. Although, many of the same organs developed AVMs or displayed AV-shunting in Eng iKO and Alk1 iKO models, phenotypes in Eng iKO mice varied widely in terms of development, severity and location. The results indicate that there may be differing pathogenetic mechanisms underlying the two major types of HHT. Additionally, there is still debate over what roles, if any, the TGF-beta type II and type I receptors (TGFBR2 and ALK5) play in endothelial cells (ECs). Contradictions in previous in vivo and in vitro data indicate that TGF-? superfamily members have spatial and temporal roles in ECs. Thus, Alk5 and Tgfbr2 were deleted specifically in ECs using a novel ALK1 cre-knockin deleter line (Alk1GFPcre) at E9.5. Alk1GFPcre;Alk5(fl/fl) and Alk1GFPcre;Tgfbr2(fl/fl) mice were embryonic lethal by E13.5 and E14.5, respectively. Mice displayed specific cerebral hemorrhages at E11.5 as previously seen in knockout models of alpha(v) integrin, alpha(v)beta(8) and Tgfb1/3. Thus, it was suggested paracrine neuroepithelium-specific integrin-mediated activation of TGF-beta signaling within ECs is essential for the midgestational establishment of the cerebrovasculature.Taken together, the data presented provide resourceful insight in the formation of vascular lesions in vivo. Our mouse models produced more consistent phenotypes than currently available mouse models and will be vital to future studies in HHT.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by HA-LONG PHUOC NGUYEN.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Oh, Suk P.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042741:00001

This item has the following downloads:

Full Text




2 2011 Ha Long Phuoc Nguyen


3 To my parents, Nguyen Phuoc Thanh and Nguyen Thi Kim Truc


4 ACKNOWLEDGMENTS I wish to acknowledge those that hel ped me throughout my graduate career. Firstly, I would like to thank my mentor Dr. Suk Paul Oh, for his encouragement and guidance over the course of my studies. I would also like to thank my committee members Drs. Jim Resnick, Peter Sayeski and Maria Gran t for their time and comments during my studies. I am very thankful for Dr. Charles Wood for providing the NIH T32 training grant in Endocrine, Metabolic, and Pre Natal Basis of Chronic Kidney Disease and Hypertension by which I was supported for two year s. I would also like to acknowledge Dr. Helen Arthur at Newcastle University in Newcastle, UK, for providing the Endoglin floxed mice. I would like to express my gratitude to former and present Oh lab members. I would like to thank Dr. Young Jae Lee for p roviding help in designing some cre primers and resolving various technical problems. I would like to acknowledge Naime Fleiss for her advice on immunohistochemistry and how she provided a friendly environment. I give thanks to Jairo Tabora for editing aid of writings, such as abstracts, and for his assistance in tissue preparation. I am indebted to Dr. Chastity Bradford for providing constructive criticism in writings and presentation preparation. I am forever grateful to my parents, Nguyen Phuoc Thanh an d Nguyen Thi Kim Truc, for their support in all my endeavors and efforts to ensure I can focus on my career. I also appreciate my siblings (and nieces and nephews) and my friends for being my cheerleaders along the way and providing the laughter that alway s helped reduce stress.


5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF FIGURES ................................ ................................ ................................ .......... 9 LIST OF ABBREVIATIONS ................................ ................................ ........................... 11 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 BACKGROUND ................................ ................................ ................................ ...... 14 The Vasculature and E ndothelium ................................ ................................ .......... 14 The Vasculature ................................ ................................ ............................... 14 The Endothelium ................................ ................................ .............................. 17 Transforming Growth Factor (TGF Signaling Pathway ................................ .... 19 TGF ................................ ................................ ................................ 20 Receptors for TGF ................................ ................................ ........... 20 Intracellular Pathways ................................ ................................ ...................... 22 TGF ................................ ........................ 23 Hereditary Hemorrhagic Tel angiectasia (HHT) ................................ ....................... 25 Genetics of HHT ................................ ................................ ............................... 26 Clinical Aspects ................................ ................................ ................................ 27 Curr ent Treatments ................................ ................................ .......................... 31 Significance ................................ ................................ ................................ ............ 32 2 METHODS AND MATERIALS ................................ ................................ ................ 35 Trans genic Murine Models ................................ ................................ ...................... 35 Mouse Breeding ................................ ................................ ............................... 35 Timed Matings ................................ ................................ ................................ .. 35 Polyme rase Chain Reaction (PCR) Analysis ................................ .......................... 36 X Gal Staining ................................ ................................ ................................ ......... 36 Histology and Immunohistochemistry ................................ ................................ ..... 37 Tissue Preparation ................................ ................................ ........................... 37 Hematoxylin and Eosin (H&E) Staining ................................ ............................ 38 Immunohistochemistry (IHC) ................................ ................................ ............ 39 Wound Induction in Endoglin ( Eng) iKO models ................................ ..................... 40 Latex Vascular Casting ................................ ................................ ........................... 40 Southern Blot Analysis ................................ ................................ ............................ 41 Hemoglobin Measurement ................................ ................................ ...................... 42


6 3 CHARACTERIZATION OF CONDITIONAL ENDOTHELIAL SPECIFIC AND GLOBAL ENDOGL IN MURINE KNOCKOUT MODELS ................................ ......... 46 Hereditary Hemorrhagic Telangiectasia (HHT): HHT1 vs. HHT2 ............................ 46 Endoglin (ENG) ................................ ................................ ................................ ....... 47 Activin Like Kinase Receptor 1 (ALK1) ................................ ................................ ... 52 Mouse Models for HHT ................................ ................................ ........................... 53 Results ................................ ................................ ................................ .................... 55 Generation of Conditional (cKO) and Tamoxifen Inducible (iKO) Eng Knockout Mice ................................ ................................ ............................... 55 Endothelial Specific Deletion of Vascular Eng (L1cre (+); Eng 2f/2f ) were Viable ................................ ................................ ................................ ............ 56 X gal and Immunohistochemistry Confirmed Cre was Active in ECs of Expected Organs ................................ ................................ .......................... 58 Confirmed Loss of Vascular Endothelial Eng did not Affect Vascular Development ................................ ................................ ................................ 58 Eng iKO Mice Displayed Variable Phenotypes ................................ ................. 59 Multiple Lower Doses of Tamoxifen (2.5mg/40g BW) Leads to More Consistent Appearances of Phenotypes ................................ ....................... 62 The 2f to 1f Allelic Conversion Efficiency at the Different Doses Contributed to the Observed Variable Phenotypes ................................ ........................... 65 Discussion ................................ ................................ ................................ .............. 66 4 SPATIOTEMPORAL ROLE OF ENDOTHELIAL TGF DEVELOPMENT ................................ ................................ ................................ ..... 85 Neurovascular Development ................................ ................................ ................... 85 Integrin Activation of TGF ................................ ...................... 86 Results ................................ ................................ ................................ .................... 89 Characterization of an Endothelial Specific Alk1 cre Knock in Line ................. 89 Endothelial Specific Deletion of Tgfbr2 and Alk5 Resulted in Cerebral Hemorrhaging Beginning at E11.5 and Embryonic Lethality by E15.5 and E14.5, Respectively ................................ ................................ ...................... 90 Alk1 GFPCre ;Tgfbr2 fl/fl and Alk1 GFPcre ;Alk5 fl/fl Embryos Form Glomeruloid like Vascular Structures in the Ganglionic Eminence ................................ .......... 92 Alk1 GFPCre ;Tgfbr2 fl/fl Embryos Die by E14.5 with CNS Specific Vascular Defects, but all Other Organ Systems are Largely Unaffected ...................... 93 Alk1 GFPCre ; Alk5 fl/fl Embryos Die by E13.5 with Evidence of Gross Cerebral Vessel and Cardiac Defects ................................ ................................ .......... 93 Discussion ................................ ................................ ................................ .............. 94 5 CONCLUSIONS AND PERSPECTIVES ................................ ............................... 107 Endothelial and Glob al Eng cKO Mice Suggest Divergent Pathogenetic Mechanisms Underlying HHT1 and HHT2 ................................ ......................... 107 Endothelial TGF Development ................................ ................................ ................................ ..... 110


7 Perspectives ................................ ................................ ................................ ......... 111 LIST OF REFERENCES ................................ ................................ ............................. 115 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 132


8 LIST OF TABLES Table page 2 1 Conditional transgenic mouse lines used ................................ ........................... 44 2 2 Transgenic cre deleter mou se lines used ................................ ........................... 44 2 3 Mating schemes for cKO models ................................ ................................ ........ 45 2 4 Primers used for PCR reactions ................................ ................................ ......... 45 5 1 Comparison of Eng vs Alk1 cKO mouse models ................................ .............. 114 5 2 Summary of Alk5 and Tgfbr2 cKO models ................................ ....................... 114


9 LIST OF FIGURES Figure page 1 1 The two main TGF signaling pathways in vascular cells are the TGF BMP pathways ................................ ................................ ................................ ... 34 3 1 L1cre(+); Alk1 2f/2f mice developed AVMs in the brain and lungs and were lethal by PN5 ................................ ................................ ................................ ..... 70 3 2 Global deletion of Alk1 in adult mice result ed in phenotypes in the lungs, liver, heart uterus, and in areas surrounding i nduced wounds ......................... 71 3 3 Schematic diagram of the Eng 2f conditional allele and subsequent deletion of exons 5 6 in the presence of cre. ................................ ................................ ....... 72 3 4 L1cre(+); Eng 2f/2f ( Eng cKO) mice were viable and vasculature in various organs were comparable to controls. ................................ ................................ .. 73 3 5 Eng cKO lung vasculature was comparable to controls at various s tages of development and adult mice ................................ ................................ .............. 74 3 6 Only two old (15 month) Eng cKO mice displayed any vascular phenotype. ...... 75 3 7 As the mice also contained the ROSA26 gene, X gal staining of various organs was performed to confirm L1cr e was active in expected organs ............ 76 3 8 Histological evaluation of X gal stained Eng cKO orga ns confirm cre was active specifically in ECs. ................................ ................................ ................... 77 3 9 Colocalization staining of PECAM 1 and ENG ................................ ................... 78 3 10 Eng iKO mice given a si ngle IP injection of high dose TM were viable and phenotypes developed much more slowly than Alk1 iKO mice. .......................... 79 3 11 Eng iKO TM2 or 3/high dose died within seven days of treatment. .................... 80 3 12 Eng iKO mice give two to three low doses of TM developed the same pattern of vascular lesions in a shorter time period. ................................ ....................... 81 3 13 Eng iKO TM2 or 3/low dose developed vascular phenotypes in more consistent manner but with variable severity ................................ ...................... 82 3 14 Kaplan Meier survival curve for each Eng iKO treatment and Hgb measurement of E ng iKO/low dose mice. ................................ ........................... 83 3 15 Testing efficiency of conversion of the 2f to 1f allele conversion between the different TM treatments in Eng iKO mice to Alk1 iKO mice ................................ 8 4


10 4 1 Synthesis and activation of the TGF ligand. ................................ .................. 100 4 2 Alk1 GFPcre a novel EC specific cre deleter line. ................................ .............. 101 4 3 Tgfbr2 and Alk5 cKO embryos display specific cerebral hemorrhaging beginning at E11.5. ................................ ................................ ........................... 102 4 4 Histological Sections of E10.5 and E11.5 embryos. ................................ ......... 103 4 5 Endothelial specific deletion of Tgfbr2 during midgestation specifically affected cer ebral vessel formation at E13.5 ................................ ..................... 104 4 6 Alk5 cKO emb ryos exhibited abnormal cerebral vasculature in the GE and cardiac defects. ................................ ................................ ................................ 105 4 7 Proposed mechanism of cerebrovascular development in the telencephalon. 106


11 LIST OF ABBREVIATION S ALK Activin Like Kinase Receptor AVM Arteriovenous malformation cre cre recombinase cKO Conditional knockout EC Endothelial Cell ECM Extracellular matrix ENG Endoglin EPC Endothelial Progenitor Cell GI Gastrointestinal HHT Heredi tary Hemorrhagic Telangiectasia iKO Inducible Knockout KO Knockout PECAM 1 Platelet Endothelial Cell Adhesion Molecule 1 RGD Arginine Glycine Aspartic acid SMA Smooth muscle actin SMC Smooth Muscle Cell TGF Transforming Growth Factor TGFBR2 Transformin g Growth Factor


12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy IN VIVO ROLES OF ENDOTHELIAL T RANSFORMING GROWTH F ACTOR RECEPTORS IN VASCULA R DEVELOPMENT AND MA LFORMATIONS By Ha Long Phuoc Nguyen Ma y 2011 Chair: Suk Paul Oh Major: Medical Sciences Physiology and Pharmacology Haploinsufficiency of the TGF like kinase 1 (ALK1) and Endoglin ( ENG ), cause the autosomal dominantly inherited vascular disorder Hereditary H emorrhagic Telangiectasia (HHT) which is characterized by vascular lesions, such as telangiectases and visceral arteriovenous malformations (AVMs). Current treatment for HHT treat the symptoms, thus understanding th e pathological mechanism is essential for creating more effective or preventative treatment. The central dogma of the HHT field is that A LK1 and E NG work in concert to pro pagate TGF dependent manner during angiogenesis and to maintain vascular integrity. Thus, AVMs form when either is disrupted; however, this relationship has not been clearly defined in vivo Thus, t wo Eng conditional kno ckout (cKO) mouse models were generated and characterized using the same two cre deleter lines (endothelial specific L1cre and tamoxifen inducible R26 Cre ER iKO) previously used against Alk1 in which m ice consistently developed AVMs Although, many of the same o rgans developed AVMs or displayed AV shunting in Eng iKO and Alk1 iKO models, phenotypes in Eng iKO mice varied widely in terms of development, severity


13 and location. The results indicate that there may be differing pathogenetic mechanisms underlying the t wo major types of HHT. Additionally, t here is still debate over what roles if any, the TGF TGF (ECs) C ontradiction s in previous in vivo and in vitro data indica te that TGF members have spatial and temporal roles in ECs. Thus, Alk5 and Tgfbr2 were deleted specifically in EC s using a novel ALK1 cre knockin deleter line ( Alk1 GFP c re ) at E9.5 Alk1 GFPcre ;Alk5 fl/fl and Alk1 GFPcre ;Tgfbr2 fl/fl mice were embryonic lethal by E13.5 and E14.5, respectively. Mice displayed specific cerebral hemorrhages at E11.5 as previously seen in knockout models of integrin, and Tgfb1/3 Thus, it was suggested paracrine neuroepithelium specific integrin mediated activation of TGF signaling within ECs is essential for the midgestational establis hment of the cerebrovasculature Taken together, the data presented provide resou rceful insight in the formation of vascular lesions in vivo Our mouse models produced more consistent phenotypes than currently available mouse models and will be v ital to future studies in HHT.


14 CHAPTER 1 BACKGROUND The Vasculature and Endothelium The Vasculature There are three main types of blood vessels ( the arteries, capillaries and veins ) that have distinct features, from their structure to specific factors expressed within each. Arteries and veins have three cellular layers: the innermost layer (intima) consists of a single layer of endothelial cells (ECs); the media consists of smooth muscle; and the tunica adventia is the outermost layer consisting primari ly of fibroblasts. Arterial vascular walls are thicker and their tunica media have multiple layers of smooth muscle cells (SMCs), elastin, and extracellular matrix (ECM). This contributes to the high elasticity necessary for these vessels to withstand the high shear stress, pressure and pulsatile flow as a result of blood circulation Veins have thinner smooth muscle layers and are not as flexible; consequently veins have a low tolerance to high pressure and high blood flow [1] Additionally, many veins have valves that prevent black flow of blood. Arteries and veins are connected via capillaries, with smaller arterioles and venu les. Capillaries consist of a single layer of ECs occasionally surrounded by pericytes and ECM This feature is essential for the exchange of oxygen, waste and nutrients between blood and tissues. Capillaries account for the majority of the urface area and possess the most phenotypic differences between vascular beds and highest adaptability to the local tissue [2, 3] Formation and m aintenance of the v asculature The vasc ular system is one of the firs t organs to form during embryogenesis. The vasculature is established and maintained mainly via two mechanisms: vasculogenesis and angiogenesis [4]


15 Vasculogenesis, the de novo establishment of the primary capillary plexus, is the initial process [5] The process i s initiated as Sonic and Indian hedgehog (SHH/IHH) proteins signal for endothelial progenitor cells (EPCs) within a population of mesodermal cells to differentiate into hemangioblasts, the common progenitor of endothelial cells and hempatopoietic stem cell s (HSCs) [1, 6] The hemangioblast becomes the angioblast, which pools into blood islands that fuse into the primary capillary plexus [7, 8] Vasculogenesis is followed by an giogenesis, which is defined as the sprouting of new vessels from pre existing ones, leading to branching and remodeling of vessels to form the basic vascular architecture of arteries, capillaries and veins [9] It is not until the heart has developed and blood flow commences that a functional circulatory system is established Vasculogenesis mainly takes place during embryogenesis, however, studies with EP Cs suggests that it may occur to a certain degree in adults [5, 10] Angiogenesis continues after birth and over the lifetime of an organism to maintain the endothelium. It is tightly regulated during physiological conditions (wound healing, ovulation, and pregnancy) and aberrant angiogenesis is responsible for several pathological conditions (certain genetic vascular diseases, rheumatic arthritis cancer, retinopathies) [9, 11 12] There are two phases of angiogenesis: the activation phase and resolution phase. The activation phase consists of the proliferation and migration of ECs. During this stage a group of ECs within the capillary bed, termed tip cells, begin to sprout. The tips cells then elongate and grow directionally by extending filipodia that sense attractive and repulsive cues within the microenvironment [13] Lying behind the tip cells are proliferating and differentiating stalk cells. In the resolution phase, there is an arrest of


16 EC proliferation and migration, deposition of the extracellular matrix, recruitment of the smooth muscle cells, and the eventual stabilization of the nascent vessel. When the tip cells contact another tip cell or vessel, they fuse with and form a bridge at the cell to cel l contact point. The vascular lumen is established by vacuole fusion in stalk cells or during tip cell fusion. The establishment of the lumen allows blood flow to begin, and consequently, leads to reduced pro angiogenic factors and stabilization of the ves sel, e.g. by recruitment of mural cells (pericytes and SMCs) [14] More recently, it has been discovered that vascular growth can occur through other modes. Arteriogenesis is now recognized as a third phenomenon by which ves sels can form that takes place in adult stages [ 15] Mechanistically it appears similar to angiogenesis in that it involves neovascularization from pre existing vessels; however, it is more specifically the formation of a vessel from an arteriole to restore blood flow in response to an arterial occlus ion. Another major distinction between the two is that angiogenesis is initiated by physiological circumstances such as wound healing and hypoxia, while arteriogenesis results from a persistent state of inflammation and monocyte invasion To a lesser deg ree in adult stages, bone marrow derived EPCs have been shown to either incorporate directly into the EC layer (thereby enlarging the vessel) or be recruited to active tip cells and participate in the angiogenic process [16] Vascular anomalities Vascular anomalies in humans, which often appear congenitally, are localized pathol ogical conditions that result from aberrant vasculogenesis or angiogenesis and phenotypically appear as focal increases of tortuous and enlarged vessels [17, 18] Vascular anomalities often spontaneously appear but several rare genetic diseases are associated with certain types [19] These


17 are separated into two broad categories: vascular tumors and vascular malformations. The vascular tumors consist primarily of hemangiomas, which is an overgrowth of ECs typically seen as red lesions in small children. It is believed these arise as a result of hyperproliferation of progenitor cells, but the trigger for proliferation is unknown. Vascular malformations (VM) are further classified into the vessel type affected (e.g. venous capillary, arteriovenous, and lymphatic), each possessing a particular characteristic and, in some cases, associated w ith a specific genetic mutation. A s such, some insight in the mechanisms underlying pathological/physiological angiogenesis can be deriv ed from the genetic diseases [20] Some examples are familial venous malformations (mutations in TIE2), familial juvenile hemangioma (VEGFR2), and Hypotrichosis lymphedema telangiectasia/LM (SOX18). Interestingly, t he inherited vascular malformations are typically autosomal dominantly inherited, multifocal, systemic, begin small in size but enlarge over time, and manifestation severity vary among patients, even within the same family [17, 18] The majority of the vascular malformations affects the EC layer of vessels (opposed to SMCs), and are associated with mutations in endothelial specific factors. The Endothelium The endothelium is the innermost single layer of cells lining the circulatory and lymphatic systems [1, 21] It is the intermed iate barrier between blood and tissues and the loca tion of many critical processes such as the maintenance of blood homeostasis, vascular tone and d elivery of necessary nutrients to the different organs [12] The functionally adaptive endothelium is highly heterogenous and its vessel identity (e.g. artery vs. vein) is dependent upon expression of particular molecular surface markers. EC factors expressed by arteries inclu de ephrinB2 (EPHB2), Jagged1, no t ch1, Delta


18 like 4 (DII4), ALK 1, endothelial PAS domain protein I (EPAS1), Hairy/enhancer of split related with YRPW motif protein ( H EY) 1, HEY 2, neurophilin 1 (NRPI), and decidual protein induced by progest erone (DEPP). In veins, EphB4, NRP2 Chicken Ovalbumin Upstream Promoter Transcription Factor ( COUP TF ) tubulin are expressed, although the last one is expressed only in the tip of venous valves. The lymphatics include vascular endothelial growth fact or receptor ( VEGFR ) 3, lymphatic vessel endothelial hyaluronan receptor 1 ( LYVE1 ) and Podoplanin [14, 22] The determination of arterial/venous identity is seemingly pre programmed as several endothelial factors are expressed even during vasculogenesis. At the hemangioblast stage, ECs expressing bone morphogenetic protein ( BMP ) 4, VEGF120, and VEGF164 are influenced into becoming arteries, while VEGF160 expressing ECs to veins or lymphatic system. ECs destined for arterial specification express DII4 and the Notch1 receptor at the angioblast stage then HEY1 and HEY2 at the final stage, after the capillary plexus is established. Likewise, venous/lymphatic identity is determined by the expression of COUP TFII Furtherm ore, the lymphatic ECs, derived from jugular vein ECs, express of VEGF C, Sox18 and Prox1 [6, 23] The actions and function of the endothelium is the driving force of angiogenesis There are two main states of ECs: active and quiescence [24] The adult endothelium is largely quiescent, in which ECs appear to be anti coagulant, anti adhesive, and coagulant/adhesive and vasoconstrictive roles [12] S everal signaling pathways are vital for maintaining this balance. EC dysfunction, the ECs inability to adapt to a pathophysiological sti mulus, is the basis of many pathological diseases, either as the primary cause or secondary


19 response [12] Among the most studied endothelial pathways are the vascular endothelial growth factor (VEGF) and angiopoietin/Tie signaling pathways [25 28] Other pathways include the platelet derived growth factor (PDGF), ephrin, and Notch pathways [22, 29, 30] Not only are these involved in embryonic vascular development, they are also known to be invol ved in physiological and patholog ical angiogenesis after birth [12, 19] Add ing to the complexity of vascular signaling is the fact that expression s of certain molecules are restricted to certain tissues, cell types or vessels and pathways may cross talk The Tra nsforming Growth Factor (TGF interest in this study. Transforming Growth Factor (TGF Signaling Pathway The Transforming Growth Factor beta (TGF wide range of physiological and pathological cel lular processes, such as inflammation, proliferation, apoptosis, development, tumor progression, and cancer [31 34] Cellular functions are dependent on the interaction between over 40 various highly conserved membe rs of this superfamily, in mammals, including ligands, Type I and II Serine Threonine (Ser Thr) kinase receptors, and intracellular signaling transport molecules, mainly SMADs [35] The classical mechanism of TGF to dimeric Type II receptors, which in turn phosphorylate serine and threonine residues within corresponding dimeric Type I receptors. The activated Type I receptor then recru its intracellular receptor activated (R )SMADs to the membrane, where these form heteromeric complexes. The heteromeric SMAD complexes then interact with a common mediator SMAD4. The complex is now allowed to translocate to the nucleus where it can regulat e gene transcription of various downstream targets [31 33, 36]


20 [Summarized in Figure 1 1, though showing those superfamily members pertaining to vascular cells] TGF L igands Over 30 multifunctional TGF have been identified in mammals. The ligands are categorized into five subfamilies that can be structurally distinguished by the number conserved cysteines within the C terminus of the mature polypeptide [31] The subfamilies are the 1) TGF ) ; 3) growth and the differentiation factors (GDFs); 4) nodal, inhibins, and activins; and 5) anti M llerian hormone (A MH). Furthermore, the each subfamily of ligands may have multiple isoforms. There are three TGF 2, and 3); about nine BMPs (BMP2 7, 8A, 8B, 9, 10); eleven GDFs (GDF1, 3, 5 9, 9b, 10, 11); three nodals (Nodal, LEFTY1, LEFTY2); five act ivins and inhibins (inhibin inhibin A, B, ); and one AMH ligand [37] Receptors for TGF L igands For the multitude of ligands, there have been only five TGF seven TGF promiscuity among the TGF superfamily members [38] The type II receptors are: 1) TGF B R2), 2) Activin type II receptor (ACVR2A), 3) BMP type II receptor (BMPR2), 4) Activin type II receptor (ACVR2B), and 5) AMH type II receptor (AMHR2). The type I receptors are activin receptor like kinases (ALK ) 1 through 7) [31] The type II and I receptors share high homology and have large intracellular domains, with the key feature of cytoplasmic kinase domains with strong Ser Thr kinase activity. T h e TGF tors are constitutively active and bind ligands with high affinity but do not possess properties to further transduce signals [35]


21 Structurally, Type II and Type I receptors differ in that th e Type I receptors lack a Ser Thr rich cytoplasmic tail. Additionally, TGF aa GS domain that contains the Ser Thr residues targeted by the type II receptor and a non activating non downregulating (NANDOR) box. An L45 loop se quence determines t he types of downstream intracellular SMADs recruited to the Type I receptor [39] TGF signaling can be divided into two major signaling pathways, based on the ligand and the L45 loop sequence in the Type I receptor, thus the SMADs involved [24, 40] The TGF signali ng cascade is activated by TGF 1, 2, 3, nodal, inhibins A, 1, 3, 8 and 9 ligands and their subsequent interaction with any Type II receptor, excluding AMHRII, and the Type I receptors, ALK4, 5, and 7. Signals are propagated intracellul arly via SMAD2/3 (Figure 1 1A). The BMP signaling pathway involves all other ligands and the Type I receptors in combination with any Type II receptor. The signal is further processed by SMAD1/5/8 (Figure 1 1B). Signaling of the Type II Type I receptor com plexes can involve auxiliary type III receptors. There are two type III receptors, Endoglin/CD105 (ENG) and glycan), which share highly similar sequences [41 43] Structurally, the Type III rece ptors have large extracellular domains but lack the kinase domain typical of the Type II and I receptors, thus Type III receptors can aid in enhancing or inhibiting TGF but cannot signal on their own. The receptors glycan is ubiquitous l y expressed and reportedly more involved in the TGF signaling pathway [41] (Figure 1 1A ). Meanwhile, ENG is predominantly found in ECs and involved in both the TGF and BMP signaling cascades [44] (Figure 1 1A and B).


22 Intracellular Pathways SMAD dependent The dominant manner that TGF are pr opagated to the nucleus is via SMAD proteins. SMAD s were initially discovered and studied in C. elegans and the Drosophila ( SMA from C. elegan s + MAD [Mother against Decapentaplegic] proteins from Drosphila ) and are highly evolutionarily conserved. Eight S MAD subtypes are categorized into three classes: 1) receptor activated, 2) co mediator, and 3) inhibitory SMAD [45, 46] There are five receptor activated (R ) SMAD s: SMAD s 1, 2, 3, 5, and 8. When activated the R SMAD s are recruited to an activated type I receptor, specific R SMAD s group to f orm homo/heteromeric complexes, which form in any combination between SMAD 2/3 or SMAD 1/5/8 (Figure 1 1A and B) Co mediator SMAD 4 will then combines with the R SMAD complex and aid in translocating the TGF heterocomplex can control various transcriptional targets by interacting with specific response elements, such as SMAD transcription regulatory regions [36, 46, 47] (Figure 1 1C). SMAD signaling can be inhibited at various points in the signaling cascade [33, 48] The Inhibitory (I ) SMADs [ 6 or 7] may also preven t TGF propagation by competing with the R SMAD s for the binding site within the TGF receptor. I SMAD s quench the signal propagation by calling for the degradation of the type I receptor by recruiting ubiquitin ligases or inviting i nactivating phosphatases (Figure 1 1D) A signal may also be inhibit ed if the R SMAD s are directly degraded by SMURF1 and SMURF2 within the cytoplasm [48] SMAD independent pathway TGF signaling pathways, most notably components of the Mitogen Activated Protein Kinase


23 (MAPK), Rho like GTP ase, and phosphatidylinositol 3 kinase (PI3K)/Akt pathway s. The cross talk can be initiated at the membrane, with the Type II receptor, or intracellularly with the SMAD s [49] (Figure 1 1E) TGF Several TGF within the vasculature, some specifically in the endothelium. These are the TGF 1, 3, BMP9 and BMP10 ligands, Type II receptors TGFBR2 and BMPR2 Type I receptors ALK1 and ALK5, and all SMADs [40, 50] These were considered major players in the vasculature because there are several human cardio vascular diseases linked to certain genetic mutations and further confirmed in various in vivo mod els. Such examples are Marfan Syndrome (TGFBR2 mutations) and Loeys Dietz S yndrome ( ALK5 and TGFBR2) [51, 52] Other members may be essential within the vasculature based on in vitro and in vivo studies, despite not being associated with human diseases. It was found that mice null for the TGF ligand, Tgfbr2 Alk5 Eng and Smad5 were commonly embryonic lethal at midgestation (from E9.5 11.5) due to defective yolk sac vasculogenesis [53 59] Alk1 was also embryonic lethal at E10.5 but due to defective angiogenesis. Additionally, Eng knockout (KO) and Alk5 KO mice presented cardiac defects and Tgfb1 KOs were affected with inflammation and autoimmunity. There is debate over wha t role s the differing TGF vasculature, and many in vivo and in vitro studies are contradictory. Both ALK1 and ALK5 are thought to be both expressed in EC. It is reported from in vitro studies that both propagated the TGF class o f ligands via TGFBR2 but ALK5 signals via SMAD 2/3 and ALK1 via SMAD 1/5/8. Goumans et al. suggested that ALK5 kinase is essential for ALK1 activation


24 [60] The a u thors found that ECs deficient in ALK5 were defective in both TGF and TGF angiongenic responses. TGF (activ ation phase) while TGF extracellular matrix deposition (ECM) and inhibit s EC migration and proliferation [61] However, another group performed a study in which ALK1 was constituitively activated in ECs. They f ound that ALK1 appeared in be involved in the resolution phase of angiogenesis [62] Additionally, t here is not a clear consensus as to which cell types the Type I receptors are expressed. It is accepted that ALK1 is expressed on ECs. I mmunohistochemical expression experiments of human colon demonstrated that both ALK1 and ALK5 are expressed in ECs [60] but murine expression studies indicate that ALK5 is expressed only in SM Cs, not ECs [63] T he murine finding was further supported in an Alk5 cKO mouse model in which A lk 5 was specifically deleted in ECs using an Alk1 cre deleter line (L1cre). The L1cre(+); Alk5 fl/fl mice were completely unaffected, while Alk1 cKO mice using the same cre deleter line consistently developed vascular malformations and died shortly after birth [64] T here has been debated over which pathway ALK1 may activate, having originally implicated in the TGF pathway based largely on i n vitro data assuming the physiological ligand for ALK1 was TGF 1. However, there is growing evidence that ALK1 is more of a BMP signaling member and that its natural ligands are BMP9 and BMP10 [65, 66] Of the tw o signaling cascade, the BMP signaling pathway seemingly plays a larger role in the endothelium [24, 50] The cross talk with other signaling pathways, such as the Notch, Shh, and WNT, is essential for the regulatio n and function of this pathway. Of


25 the ligands, BMP2 and BMP4 are mediators of endothelial differentiation and function. Two other endothelial specific ligands, BMP10 and BMP9, promote a quiescent endothelium [66, 67 ] Targets of the quiescent factors include inhibitor of differentiation (ID) 1, ID2, interleukin 8, and ENG. Certain BMP antagonists (including noggin, chordin, and BMPER) play critical roles in regulating the expression of the various BMP ligands, typic ally in a dose dependent manner. For example, low concentrations of BMPER activate BMPs, while high concentrations have the opposite effect. Gremlin is another BMP antagonist that is induced in hypoxic conditions and works by preventing the BMP ligand from binding to the receptor [50] The importance of TGF within the endothelium is exemplified in the genetic vascular dysplasia Hereditary Hemorrhagic Telangiectasia because mutations in Alk1 Eng and Smad4 cause the disease. Hereditary Hemo rrhagic Telangiectasia (HHT) Hereditary Hemorrhagic Telangiectasia (HHT), or Rendu Osler Weber Syndrome, is an autosomal dominantly inherited vascular disorder. Though still considered rare, it is one of the more common inherited diseases, affecting 1:5,00 0 to 10,000 people worldwide, with no clear preference to gender geography or race [68, 69] Clinical diagnosis of HHT is based on the following Cura ao Criteria: 1) an affected immediate family member, 2) epistaxi s, 3) mucocutaneous telangiectasia, and 4) arteriovenous malformations (AVMs) within certain major organs [70] A patient is a giv en a suspected diagnosis if there are only two of the four Criteria, but a confirmed diagnosis if he/she have at least three. The vascular malformations can leave the vessels quite fragile and prone to rupturing and hemorrhaging. HHT is a largely unpredict able disease in that the severity and location of symptoms varies amongst patients, even family members, and


26 symptoms often appear spontaneously. Interestingly, only predilection sites of vascular beds are frequently affected in a highly variable pattern Age of onset of symptoms also varies, although there seems to be a degree of age dependent prevalence, appearing more frequently in the older population, typically older than 40 years of age [71] HHT is often diagnosed late or overlooked as the symptoms are very broad and many non specialized physicians may not properly identify an HHT patient [68] The l evel of concern caused by hemorrhaging depends upon bleeding severity, tissue location and environment. Acute hemorrhaging is of concern if it affect s hemodynamics, occur s in a closed space (i.e. cerebral/spinal system), or compromises organ functio n (e.g. oxygen exchange in the lungs). Chronic h emorrhaging can lead to anemia; a small population of HHT patients (6 7%) may experience thromboemboli [72] Genetics of HHT HHT is caused by mutations that result in the haploinsufficiency of five genes, each of which correspond to a type of the disease (HHT1 4, JP HHT); however, only three of the genes have been definitively identified. HHT type I (HHTI) is associated with mutations in ENG and HHT type II (HHT2) is d ue to ALK1 [73, 74] HHT1 and HHT2 account for >80% of cases. SMAD4 is linked to a phenotypically combined syndrome of HHT and another autosomal dominantly inherited disease called Juvenile Polyposis (JP HHT) [75] The JP HHT patients (which comprise of only 1 2% of cases) experience the formation of polyps within the gastrointestinal (GI) tract and colon during childhood, hallmarks of JP, in addition to HHT phenotype. It is r eported that a quarter of JP HHT were results of de novo mutations [76] HHT3 and HHT4 are reportedly due to mutations in regions within chromosomes 5 and 7p14, respectively, however, the specific genes are still unknown [77, 78] A suggested candidate gene for HHT4 is


27 BMPER [78] Based on recent fine mapping sequencing, genes that have been excluded for HHT3 are the angiogenic proteins vascular endothelial (VE) cadherin, Sprouty4, and fibroblast growth factor (FGF) 1 [79] The paucity of information of HHT3 and HHT4 indicate symptoms are milder compared to the other types of HHT [77, 79] In gener al, i t is largely believed that patients are heterozygous for the genetic mutations as there has not been a confirmed case of a homozygous mutated patient [71] Also, conventional animal knockout models for each HHT gene are embryonic lethal, supporting the notion that patient s may be heterozygous for the dysfunctional gene. Additionally, it has been accepted that the vascular malformations develop as a result of haploinsufficiency of the functioning gene [80, 81] Clinical Aspects The most common symptom (afflicting > 90% of patients) is recurrent, spontaneous epistaxis, lasting from a few minutes up to hours. Though seldom life threatening, it often lead s to a reduced quality of life and is the most frequent reason a patient would seek tre atment and possibly a diagnosis [82] Many p atients typically experience the first nosebleed by age 10 then their condition may worsen by which they may experience nosebleeds daily and multiple times a day [72] In general, the bleeding seen in HHT patients is presumably a consequence of abnormalities in the vascular bed with endothelial dysfunction being the most likely culprit. The nosebleeds are believed to result from the formation and subsequent rupturing of telangiectases within the nasal passage. Telangiectases, the second most common symptom (at 8 0% of cases), are focal lesions in which the postcapillary venules have dilated [83] Consequently, the vessel wall is compromised and highly susceptible to rupturing. Telangiectases can appear as red spots if they are near the


28 surface of the skin. These are often visible within the tongue, oral cavity, lips, and surface of the skin (seen in ~75% cases ), though they have also been found in visceral organs, like the l iver and brains and have been shown to cause arteriovenous (AV) shunting a process when a rterial blood flows directly in the veins While telangeictases inside the nose or GI tract frequently bleed, they are seldom dangerous if these are superficially located on the skin or oral cavity [84] However, facial lesions can be stigmatizing. When telangiectases have progressed in size enough the arteries will connect directly into the veins, an abnormality known as an arteriovenous malformation (AVM) [83, 85] AVMs are the most problematic of the HHT symptoms especially if formed in major visceral organs, such as the lungs, liver, and brain. AVMs are treatable when found; however, in some cases the patient may be unaware of its existence These years, then suddenly cause catastrophic complications. AVMs within the pulmonary and cerebral vasculature can cause severe neurological problems and are the leading cause of death i f they are not detected and treated in a timely manner. When patients are screened for AVMs, it is for these two types of AVMs [86] Interest ingly, 80 90% of all reported pulmonary AVMs (PAVMs) are associa ted with HHT, although only 15 5 0% of HHT patients develop PAVMs. Most PAVMs develop later in life and more HHT1 patients over HHT2. Lung a fflictions that may directly result from PAVMs include hypoxaemia, cyanosis, poor exercise tolerance /decompres sion illness and dyspnoea [87] Secondary effects to the cerebral system include


29 abscesses due to paradoxic embolism [87] In fact, 30 40% of patients with a PAVM develop strokes, seizures and cerebral abscesses. The size and number of the PAVMs plays a big factor in the severity of secondary symptoms. R upturing of a PAVM and death from it can happen though rare, with only ~10% of patients experiencing hemothorax or hemoptysis [87] Another commonly affected organ system is the GI tract, with 25 33% of HHT patients found with vascular malformations in areas such as the stomach, duodenu m, jejunum and colon Telangiectases are more often see n and bleeding can lead to severe, transfusion dependent anemia and having patients undergo more fr equent blood transfusions occur, particularly in patients older than 60 years old [87] Patients with SMAD4 mutations should additionally be screened for gastric polyps as these may progressively become malignant [76] It is believed that 32 .5 74% patients experience the hepatic manifestations [88] Fortunately, most liver vascular malformatio ns are not life threatening and HHT patients typically remain asymptomatic because the heart can compensate for minor shunting. On the other hand, in very rare cases complications from the severe liver AVMs can be among the most dangerous and difficult to treat [89] There are 3 categories of HHT liver disease: Garcia types 1 3. Garcia type 1 leads to heart failure (most common but treatable) and post capillary pulmonary hypertension ; Garcia type 2 to portal hyperten sion (the most consistently difficult to manage) ; an d Garcia type 3 to varietal bili ary disease (with varia ble outcomes) [71] Liver complications appear to affect more women than men, an d those over the age of 40 experience heightened risk. Additionally, vascular lesions may also form in the bladder/ureter, conjunctiva,


30 pancreas, bowel or spleen, however, hemorrhaging at these sites are seldom of relevance [72] Occurring in about 10 20% of HHT patients one of t h e most critical types of AVM for HHT patients are the cerebral AVMs (CAVMs) because treatments are not standardized and can be high risk [72, 87] HHT related CAVMs can be classified into three types: micro AVMs (<1 cm), small nidus AVMs (1 3 cm) and fistulous AV shunts, each with a prevalence of 40%, 30%, and 30%, respectively. Most CAVMs are superficial, rarely penetrating the white tissue matter. Symptoms from CAVMs are similar to the neurological affects of PAVMs mentioned previously. Interestingly, only 28% cases of the neurological complications are due directly from cereb ral hemorrhaging from the CAVMs; instead 61% are indirectly from PAVM emboli [87] Additionally, CAVMs can cause epilepsy and there are cases that CAVMs in children could lead to high output cardiac failure. Spinal AVMs are even rarer (>1%) and hemorrhaging fro m thes e can cause patients to be become paraplegics [72] Women with HHT are typically capable of becoming pregnant and carry ing the child to term if extra precautions are taken [ 71, 72, 84, 87] PAVMs pose the biggest treat as these have a tendency to rapidly grow in size and consequently result in hemothorax or hemoptysis Precutaneous embolotherapy is sometimes performed during the second trimester as a preventative measure Ad ditionally, a ntibiotics can be given before delivery to help prevent infections. If spinal AVMs are present, epidurals are largely avoided during delivery. Cases of major PAVMs hemorrhaging (1.3%) and major stroke (1%) is very rare. The chances of a miscar riage are increased if the mother is hypoxic


31 A subset of HHT patients, predominantly HHT2, may also develop another rare but potentially fatal vascular disease called pulmonary arterial hypertension (PAH) [90] PAH may result from maladaptive pulmonary vascular remodeling, including exaggerated proliferation and dy sfunction of both ECs and vascu lar smooth muscle cells (vSMCs). This can lead to hypertrophied SMC layers in the pulmonary vascular system and cause increased vascular resistance. The consequent increase in pulmonary blood pressure leads to the PAH charact eristic mean pulmonary arterial pressure >25 mmHg at rest or >30mmHg during exercise [91] There are multiple categories of PAH. Familial PAH (fPAH) occurs when the cause is genetic. Confirmed genetic cause of PAH are attributed to the TGF Type II receptor BMPR2 Secondary PAH arises in response to another disorder (e.g. scleroderma), and idiopathic (IPAH) PAH is when the cause is not known [91] Symptoms of PAH which overlap with HHT include exercis e intolerance, fatigue, dyspnea, and weakness [92] Current Treatments Current treatments for HHT alleviat e the inconvenience of symptoms [86] Recently, more focus has been on antiangiogenic targ ets. Clinical trials for drugs treating HHT patients include using anti angiogenic agents such as beviacizumab (a neutralizing antibody against all VEGF 2b in attempts to stop bleeds or the formation of AVMs [93 9 5] In terms of treating epistaxis, it is believed that nasal humidification by topical ointments and sprays should be initially attempted [86] Pharmacological t reatments for epistaxis include the drugs mentioned above antifibrinolytics (e.g. epsilon aminocaproic acid and tranesamic acid), female sex hormones (e.g. estrogen) applied either systemically or topically or partial hormone antagonist distributed orally, such as tamoxifen or raloxifen [96, 97] Sur gical options


32 encompass coagulation of the endonasal telangiectases with electrical or laser cautery, the replacement of fragile respiratory nasal mucosa with skin grafts from the thigh or pedicled oral mucosa grafts as m odifications of a septodermoplasty ( first described by Sanders ), and, i n extreme cases the nose may be surgically closed (via a modified [98 100] Preliminary attempts of gene therapy for HHT pat ients show promise For one, a plasmid construct was created in which the endothelial specific gene ICAM 2 and Eng were inserted upstream of human Eng cDNA After injecting the expression vectors systemically or locall y it was found that there was indeed E NG expression in the vesse ls of the lung, liver and skin [101] Another method showing promise as therapy is the use of isolating blood outgrowth endothelial cells (BOECs) from the circulation, expanding them in cul ture then re incorporating the cells in vivo The initial reports of this were done in 2005 in which BOECs from HHT1 and HHT2 patients were collected and characterized. The matured ECs expressed the expected EC markers (e.g. PECAM 1 1, ENG, FLK1, ALK1, von Willebrand Factor), but there was a drastic decrease of ENG in both HHT1 and HHT2 derived cultures [102] Significance HHT is a systemic complex and unpredictable disease. For one, the vascular malformations seen i n HHT appear as focal lesions, thus not all vascular beds are affected. Braverman et al. first describe the growth of a telangiectasia into an AVM in 1990. They described how the telangeictasia begins as a dilatation of the postcapillary venule. The SMC la yer of the venule thickens and its diameter enlarges. Though there are still intervening capillaries conneting the growing venule to arterioles, eventually the arterioles begin to grow as well and the capillaries eventually disappear [83] Because


33 the three known genes responsible for HHT are part of the TGF interestingly those expressed specifically on ECs, defective TGF the EC is primarily affected appears to be a cause of the HHT lesions. Hence, examining TGF signaling in ECs to understand the underlying pathogenic mechanism is important for effective treatment. There are currently no treatments that genetically target the HHT genes, as the mechanisms of these are still not clearly understood. To determine the in vivo role of Eng in physiological and pathologic angiogenesis and how the loss of Eng can lead to formation of vascular malformations, two Eng cKO (an endothelial specific and a global KO) were generated and the phenotypes compared to Alk1 cKO phenotype s (using the same cre deleter line). Then to clarify why EC specific deletion of ALK5 and TGFBR2 did not result in vascular malformations, in our previous models [64] EC specific Alk5 and Tgfbr2 ablation models wer e characterized using a novel cre knockin line to determine whether TGF superfamily members have spatiotemporal roles in ECs.


34 Figure 1 1. The two main TGF signaling pathways in vascular cells are the TGF BMP pathways A) In the TGF 2, 3) ligands bind to TGFBR2, which phosphorylates a TGF ALK5. There is still debate whether ALK1 interacts or is involved in this glycan or ENG may aid in li gand binding. The TGF SMADs 2/3. B) In the BMP signaling pathway, a BMP ligand (BMP2, 4, 9, 10) binds to the type II receptor BMPR2, which recruits and activates ALK1 or ALK2. ENG is the only type III receptor th at is involved in this pathway. The signal is transduced intracellularly through SMAD1/5/8. C) The pathways converge as the pathway corresponding R SMADs complex with the co mediator SMAD4. The heterocomplex translocates into the nucleus and regulate trans cription of downstream targets. D) Signal transduction can be interfered by I SMAD6 or 7, which can compete with R SMADs to bind the activated type I receptor or recruit ligases that degrade the type I receptor or signal. E) Additionally, either pathway c an cross talk with other signaling pathways throughout the intracellular cascade, most notably TAK, JNK, and notch signaling. These can interact directly with the activated type I receptor, with the SMADs, or intercept and transduce the signal to the nucle us.


35 CHAPTER 2 METHODS AND MATERIAL S Transgenic Murine Models Mouse Breeding All procedures performed on animals were reviewed and approved by the University of Florida Institutional Animal Care and Use Committee. The generation of transgenic mouse lines: L1cre, bi floxed Endoglin ( Eng 2f/2f ), floxed Tgfbr2 ( Tgfbr2 fl/fl ), and floxed Alk5 ( Alk5 fl/fl ) were previously described [55, 64, 103, 104] R26 Cre ER was purchased from Jackson Laboratories. To generate L1cre(+);En g 2f/2f mice, Eng +/2f mice were mated to L1cre(+) mice, resulting in litters heterozygous for the floxed Eng allele. L1cre(+);En g + /2f were crossed with En g + /2f mice. Eventually Eng 2f/2f were selected to mate with L1cre(+);En g + /2f to maintain colonies. Mice were on a mixed breed background of C57BL6 and 129Sv. To generate R26 Cre ER (+) ;Eng mice, R26 Cre ER (+) ;Eng 2f/2f mice were mated with Eng 2f/2f mice to generated offspring that were either R26 Cre ER (+) ;Eng 2f/2f or Eng 2f/2f R26 Cre ER activity was induced by an intraperitoneal (IP injection) of tamoxifen (TM), dissolved in corn oil, in adult mice (older than 2 months of age). Concentrations and number of injections used were at 2.5mg TM/25 g mouse body weight either once, twice or thrice, or at 2.5mg TM/40 g body weight either two or three times. Timed Matings Tgfbr2 fl/fl and Alk5 fl/fl mice were mated with respective Alk1 GFPcre ;Tgfbr2 +/fl and Alk1 GFPcre ;Alk5 +/fl mice then the female was checked for a copulation plugs the next morning. The presence of a plug was designated embryonic day (E)0.5. At the embryonic day of desire the pregnant female was euthanized by exposure to 100%


36 isoflurane (Sun Surgical) followed by cervical dislocation. The embryos were removed and collected into cold 1X phosphate buffered sa line (PBS). For E10.5 13.5 embryos, the heads were detached from the body then the embryos and corresponding placentas were collected. For E15.5 and older embryos, whole organs were removed and collected. Embryonic tissues were fixed in 4% Paraformaldehyde (PFA) in 1X PBS (pH 7.0) for 2 12 hrs at 4C, with shaking. Polymerase Chain Reaction ( PCR ) Analysis Two mm tail biopsies were collected from mice at three weeks of age or older; for embryos, the yolk sac was collected. Tissues were digested in Lysis buf fer [50 mM Tris (pH 8.0), 0.5% Triton X] with 1 mg/ml proteinase K (EMD Biosciences) for 8 16 hr at 55C. The tissues were centrifuged for 10 min at 12000 rpm, 2 l of the supernatant was added to the following PCR reaction: 5X PCR Buffer (Promega), 25 m M MgCl 2 (Promega), 25 pM dNTP, 0.5 l of 25 pmol primers (Integrated DNA Technologies), volume to 25 l with sterile ddH 2 O. To avoid evaporation, a layer of mineral oil (Sigma) or 10 min, then to 72C during which Taq polymerase (Promega) was added. The reaction was allowed to undergo 36 cycles of: 95C for 45 sec, 6 0 C for 45 sec, and 72C for 1 min; followed by an additional 72C elongation for 10 min. The products were run in a 2% agarose (Lorenza) and visualized by a ultraviolet (UV) transilluminator. See Table 2 1 for the list of PCR primers used. X Gal Staining 1 2 mm fragments of various organs were collected into 1X PBS. Tissues were incubated in X gal fixative solu tion [1% formaldehyde, 0.2% glutaraldehyde, 2 mM MgCl 2 5 mM EGTA, 0.02% NP 40) for 15 min at room temperature (RT), with rocking.


37 Tissues were rinsed twice with ddH 2 O for 5 min each, with rocking. After washing the tissues were incubated in X gal staining solution [5 mM K 3 Ke(CN) 6 5 mM K 4 Ke(CN) 6 2 mM MgCl 2 0.01% Sodium Deoxycholate, 0.5 mg/ml X gal], X gal (Fisher Scientific)] for 6 16 hrs at 37C in the dark, with rocking. The stained specimens were then washed in 1X PBS at RT then analyzed using a ligh t microscope. Afterwards, these were prepped for histology. Histology and Immunohistochemistry Tissue Preparation Embryos or 1 5 mm biopsies from various organs from adult mice (1.5 months or older) were collected into cold 1X PBS. The left lobe of adult mice was inflated for better resolution of the vessels. First, the lungs were flushed with 1 ml 10U/ml heparin (1:10 in 1X PBS,100U/ml; Abbott Laboratories) by injection into the right ventricle of the heart with a 26 G needle (BD Precision). The whole hea rt and lungs were collected into 1X PBS, then lung left lobe and the connected portion of the primary bronchi to the trachea branching point were excised. A blunted 25 mm gauge (G) butterfly needle was inserted into the bronchi and tightly tied with a sutu re. The lung was inflated with 1X PBS for 7 min following by 4% paraformaldehyde (PFA) for an additional 7 min, both via gravity flow, in which the fluid level was 20 cm higher than the lung. The inflated lung was pulled off the needle, immediately and com pletely ligated at the bronchi, then fixed with 4% PFA. For paraffin embedding of tissue adult organs, the whole or at least 2 to 5 mm biopsies were collected into 1X PBS and fixed in 10% buffered Formalin (Fisher Scientific) or 4% PFA overnight (O/N) at RT, with shaking. For embryos, whole embryos were collected in cold 1X PBS, decapitated and fixed in 4% PFA O/N at 4, with


38 shaking. All fixed tissues were then washed in 1X PBS then dehydrated in a graded series of ethanol (for adult organs: 2x70% to 2x9 5% to 3x100%; for embryos: 1x25% to 1x50% to 2x70% to 1x80 to 2x95% then 3x100%), followed by clearing in CitriSolV (Fisher Scientific). The tissues were incubated in twice paraffin at 60C then once at 60C with vacuum pressure and finally embedded in para ffin wax. Using a microtome, tissues were cut into 5 m thick sections onto positively charged slides (Fisherbrand SuperfrostPlus) and heated for at least 20 min to fix the tissue onto the slide. For further histological processing, paraffin embedded secti ons were cleared and rehydrated in CitriSolV (Fisher Scientific), a degraded ethanol series (100% to 95%) and ddH 2 O. For frozen embedding of tissue whole organs or 2 to 5 mm biopsies were fixed in 4% PFA overnight at 4C, with shaking. Fixed tissues were briefly washed in 1X PBS and ddH 2 O, and incubated in 15% sucrose (Sigma) for 3 6 hrs then 30% sucrose O/N at 4C, with shaking. Tissues were soaked in a 1:1 mix of OCT:30% sucrose for 2 hrs, then embedded in OCT (Tissue Tek), then stored at 80C until nee ded. Tissues were cut into 8m sections onto positively charged slides (Fisherbrand SuperfrostPlus) using a cryostat. Hematoxylin and Eosin (H&E) Staining Tissues were incubated in filtered hematoxylin (Richard Allan Scientific) for 1.5 to 2 min, rinsed in ddH 2 O, rinsed in clearing solution for 8 secs, bluing solution for 15 sec, then 95% ethanol for 1 min. Slides were placed in eosin (Y with phloxine; Richard Allan Scientific) for 15 25 sec, dehydrated/cleared in 95% and 100% ethanol then CitriSolV and m ounted with Permount (Fisher Scientific)


39 Immunohistochemistry (IHC) After rehydration, tissues were blocked in 3% hydrogen peroxide (30%; Fisher). For staining of smooth muscle, tissues were stained using the Vector Mouse on Mouse (M.O.M) Peroxidase Kit ( The primary antibody us SMA (1:800; Clone 1A; Sigma) and visualization was carried out using the DAB Substrate kit (Vector Laboratories). For staining of endothelial cells, tissues were digested using trypsin (Carezyme I; BioCare Medical) followed by b locked with Rodent Block M (BioCare Medical) then incubated in PECAM 1/CD31 (1:100; BioCare Medical), diluted in Da Vinci Green diluent (BioCare Medical) O/N at 4C. The protocol continued using the Rat on Mouse AP Polymer kit (Bio Care Medical), with 20 min incubation in both solutions. Chromogen reaction was carried out using the Vulcan Fast Red chromogen (BioCare Medical). Slides were counterstained in hematoxylin, dehydrated and cleared, then mounted with Permount. For ENG stain ing, frozen sections were air dried and fixed in cold acetone. Tissues were blocked with 5% Normal donkey serum (CalBioChem) then incubated in rat anti ENG (1:100; Clone MJ7/18; eBioscience) O/N at 4C. This was followed by incubation in AlexaFluor 594 donkey anti rat IgG (1:500; Molecular Probes). Sections were doubled stained with PECAM 1 re fixing tissue in 4% PFA, blocking in 5% n ormal goat seru m and incubating tissues with a PECAM 1 FITC conjugated mouse monoclonal (1:250; Clone 390; Millipore) O/N at 4C. Tissues were mounted using Vectashield mounting medium with DAPI (Vector Laboratories) diluted 1:1 in 1X PBS. Staining was vis ualized by a Leica confocal microscope.


40 Wound Induction in Endoglin ( Eng ) iKO models On the last day of TM injection, mice were anesthetized with 5% isoflurane/95% O 2 then maintained under 1 3% isoflurane/O 2 delivered via nose cone connected to an anesthe tization machine. A 2 mm hole was induced in the untagged ear with a 2 mm biopsy punch. For the dorsal skin wound, the dorsal fur was shaved off with an electric razor and the area briefly sprayed with alcohol. A 5 7 mm wound was induced and the wound imme diately treated with an antiseptic, such as betadine. Latex Vascular Casting To visualize potential AVMs, a latex dye vascular casting was done in which latex was injected into the arterial circulation. Because latex is a thick liquid it cannot travel com pletely through the fine capillary beds thus in normal conditions remains solely in the arteries. However, if there is AV shunting, the latex can be seen in the veins (often as a lighter shade of blue than the arteries). Mice were anesthetized with a mix of 100 mg ketamine and 10 mg xylazine per kg body weight, injected intraperitoneally. A sagittal incision was made from the abdominal to the neck and the chest was opened to reveal the heart. For visualization of the systemic circulation, a 25 G butterfly needle (BD Precision) was inserted into the left ventricle. The circulation was perfused with a succession of three solutions (10 ml each) at a rate of 120 ml/hr using a KD Scientific syringe pump: 1) Heparin (20U/ml; Sigma) to unclot blood; 2) freshly pre pared vasodilator solution [10U/ml Heparin (Sigma), 0.4 mg/ml Papaverine, 100 M s odium nitrop russide in 1X PBS] to dilate vessels; and 3) 10% buffered formalin (Fisher Scientific) 400 700 l of blue latex (Connecticut Valley Biological Supply) was immedi ately injected using a 26G 3/8 syringe (BD Precision) inserted into the same injection site. For latex injection of the pulmonary circulation,


41 latex was injected into the right ventricle of the heart and the left atrium was nicked to allow dye to flow out. The injected mouse was washed of excess dye and fixed O/N in 10% buffered formalin. To better visualize vessels, the lungs, skin and ears were further process by dehydration in a graded methanol/ ddH 2 O series (50% to 70% to 90% to 100%) followed by clearin g in an organic solvent solution (1: 1 benzyl alcohol:benzyl benzoate) Southern Blot Analysis For the southern blot probe, a 760 bp PCR product was generated using conditions mentioned previously. The isolated PCR product was then labeled with DIG using ei labeling kit, 3 mm tissue biopsies were harvested and genomic DNA (gDNA) was isolated using QIAGEN Tissue and Blood kit, according to the manufactur Beckman Coulter spectrophotometer. 10 g of gDNA was digested with 5U/g BamHI restriction enzyme (NEB) for at least 16 hrs at 37C. If total volume of digestion reaction were loaded directly into the 1% agarose gel. If total volumes were > 25l, sample volumes were to brought up to 100 l by the addition of ddH 2 O. Samples were precipitated by adding 2X sample volume and 10% sample volume of 100% ethanol and 3M sodium acet ate, pH 5.2, respectively, then incubating at 20C for at least 20 min. Samples were centrifuged at RT for 25 min, washed with 3X original sample volume of 70% ethanol, centrifuged for another 10 min at RT, the pellet air dried, then resuspended in 20 l autoclaved ddH 2 O. 1 l of 6X DNA loading dye was added to all samples. 1% agarose gels were loaded with a DIG labeled molecular marker, with a well skipped between the marker and experimental samples. The


42 digested gDNA was slowly run in the gel at 35 35V u ntil the dye has migrated at least 7 cm in the gel, indicating sufficient separation or gDNA. After electrophoresis, the gel was incubated in depruination solution (250mM HCl), a denaturation solution (NaOH, NaCl) then neutralization solution (Tris base, N aCl). The gDNA was transferred onto a positively charged nylon membrane (Roche) via capillary action for at least 12 hrs, using 20X SSC (NaCl, Sodium citrate, pH 7.0) as a transfer buffer. Post transfer, gDNA was fixed onto the membrane by UV crossing by exposure to low UV wavelengths of a UV transilluminator for 5 min. The membrane was rinsed briefly in ddH 2 labeled probe was allowed to hybridize to target gDNA for at least 16 hrs in a 4 2C water bath, with shaking. Next, the membrane was washed in low stringency buffer (2X SSC, 0.1% SDS) at RT and high stringency buffer (0.1X SSC, 0.1% SDS) at 68C, DIG AP conjugated secondary anti body (1:10,000 in 1X blocking solution; Roche). Chemiluminescent bands were visualized using CDP Star (Roche), then exposing the membrane to autoradiography film (GeneMate). Quantification of the southern blots was performed with ImageJ software (NIH, Best hesda). Hemoglobin Measurement To monitor potential internal hemorrhaging within cKO mice, the hemoglobin levels of tamoxifen treated mice were measured. Two mm of the tip of the mouse tail was cut off and the first drop of blood was discarded. A second dr op of blood was extracted, collected into a microcuvette (HemoPoint H2 n*x*t microcuvette, Stanbio Laboratories), then placed into a HemoPoint H2 Hemoglobin Photometer (Stanbio Laboratories). The


43 normal hemoglobin levels for mice is 10 20 g/dL, thus levels below would be indicative of hemorrhaging ( http://www.fauvet.fau.edu/oacm/VetData/Handouts/mouseHO.htm ).


44 Table 2 1 Conditional transgenic mouse lines used Mouse line Description Eng 2f loxP site flank exons 5&6 which code the extracellular domain Tgfbr2 fl loxP site flank exon 2, which contains the transmembrane domain Alk5 fl loxP site flank exon 3, which contains the transmembrane domain R26R Reporter gene used to visualize cre activity after X gal staining Alk1 2f loxP site flank exons 4 6 of Alk1 gene, which includes the transmembrane domain Alk1 3f loxP site flank exons 4 6; an additional neomycin resistant cassette and loxP site is inserted in intron 6 Table 2 2. Transgenic cre deleter mouse lines used Cre line Time of cre activation M echanism Cell type affected L1cre Beginning at E13.5 Cre driven by 9.2 kb region of Alk1 promoter ECs of lungs, GI tract, brain, reproductive organs, eye; NOT skin, liver, kidneys R26 Cre ER Adult mice (aged >2months) Cre activated by IP injection of tamoxifen Ubiquitously expressed Alk1 GFPcre Beginning at E9.5 A GFPcre construct inserted into intron 3 of Alk1 gene ECs




46 CHAPTER 3 CHARACTERIZATION OF CONDITIONAL ENDOTHELIAL SPECIFIC AND GLOBAL ENDOGLIN MURINE KNOCKOUT MODELS Hereditary Hemorrhagic Telangiectasia (HHT): HHT1 vs. HHT 2 The majority of HHT cases are attribut ed to mutations in Eng (HHT1) and Alk1 (HHT2). Eng was the first identified HHT gene, reported by two independent groups in 1994 [73, 74] ; however, even within one of the reports it was suspected that more than one gene may be responsible. It took another two years, in 1996, for Alk1 to be identified by the Marchuk group [105] Interestingly, it would take another ten years for other HHT causative genes to be found, although t wo have still not been definitively identified [75, 77, 78] The different types of HHT are clinically indistinguishable, but g eographically, there are reportedly higher incidences of HHT2 around the Mediterranean ( e.g., Italy, France and Spain) and HHT1 is more prevalent in North America and northern Europe [106 110] Adding to the complexity of understanding this disorder is the fact there is not one specific mutation respo nsible for causing the disease. Each reported HHT family has a unique mutation and de novo mutations have also been discovered, though rare. Thus far, there have been almost 330 and about 270 distinct ENG and ALK1, respectively, mutations reported (hhtmuta tion.org). Genotype:phenotype correlation studies (mostly performed on HHT1 and HHT2 patients) suggests that certain symptoms and visceral AVMs occur more frequently in HHT1 over HHT2, and vice versa. For instance, more HHT1 patients develop PAVMs and CAVM s, while HHT2 patients develop more hepatic and pancreatic AVMs and have a higher risk for PAH. Additionally, HHT2 patients seem to develop mucocutaneous telangiectases at an earlier age [72 111]


47 Fervent attempts have been made in understanding the functions and expression of both ALK1 and ENG in physiology and the HHT pathological states, with the overall goal of creating ef ficient treatment for the disease. ENG and ALK1 are unique compared to TGF superfamily members because their expression is almost exclusive to ECs. Furthermore, ALK1 is expressed specifically in arterial endothelial cells [112] Endoglin (ENG) Endoglin/CD105 is a homodimeric disulphi de linked plasma membrane receptor located on the human chromosome 9q34 [113] and chromosome 2 in mice [114] Structurally, Eng consists of 15 exons and has a large extracellular domain. ENG has five glycosylation sites within its N terminus and an O glycan domain rich in serine and threonine residues near the transmembrane domain [115, 116] ENG is conserved among higher mammals; however, o ne feature that is found only in human ENG is an Arginine Glycine Aspartate (RGD) motif that serves as a binding site for adhesion molecules in the extracellular matrix (ECM) [117] The short cytoplasmic domain act s as a target for the TGF [41] There are two splice forms of ENG: long (L ) and short (S ENG) forms, which differ in 33 amino acids (aa) in size [118] L ENG is a 658 aa long polypeptide, with a 47 aa long cytoplasmic tail. The L ENG gene contains 15 exons, numbered 1 9a, 9b, 10 14. S EN G involves the inclusion of an additional 135 nucleotides between exon 13 and exon 14 (effectively called exon 13A) introducing a premature stop codon that leads to a shortened 625 aa protein produc t with a 14 aa cytoplasmic tail [119] The four domains of ENG are domains: a short 25 aa signal peptide domain (part of exon 1), a large extracellular domain (part of exon 1 through exon 13), a small 25 aa transmembrane domain (part of exon 13) and a 14 or 47 aa long cytoplasmic domain


48 (exons 13 14). There is also a third albeit soluble form of ENG (s ENG ) that lacks the transmembrane domain. ENG is predominantly expressed in all vascular endothelial cells but has also been found in erythroid precursors and stromal cells of the fetal and adult bone marrow, syncytiotrophoblasts, hematopoietic stem cells, very low levels in smooth muscle cells, and monocytes [120 122] ENG expression is high during development and i s expressed at basal levels in quiescent ECs [123] ENG expression is increased again when ECs are activated, such as in angiogenesis, inflammation, wound healing and vascular injury [124 126] ENG is first expressed at E6.5 in the extraembryonic ectoderm, and seen at E8.5 on primitive ECs of the yolk sac. By E9.5 ENG expression is on all ECs (particularly in endocardium but not myocardium). By E12.5 the strongest expression is within capillaries but weak in veins and intermediate arteries [126] The L ENG isoform is the dominant form within the vasculature and the form that is the most studied in the literature. The functional differences between L ENG and S ENG have not been extensively studied. Interesting, it appears that the two lesser forms, S ENG and sENG, have the same opposing /antiangiogenic response to L ENG. Many studies focusing on S ENG have been performed in vitro The S ENG expression is increased in cultured senescent human ECs and older mice, suggesting S ENG is more active within the aged vasculature [127, 128] An in vivo model by Perez Gomez and colleagues, in which Eng +/ mice overexpressed S ENG ( Eng +/ ;S Eng + ) via adenovirus, was previously characterized. It was found that the two membrane bound ENG isoforms were highly co e xpressed in the lungs, heart and liver. Also, overabundance of S ENG could not rescue the Eng null mice, as no surviving Eng / ; S


49 Eng + were found. Further supporting opposing functions between L ENG and S ENG, when injected with Lewis Lung Carcinoma cell, S Eng + mice developed tumors that were 35% smaller in weight and contained 65% less hemoglobin than control littermates [129] ENG has received more attention lately as excessive s ENG is linked to another pathology preeclampsia [130] Preeclampsia is a serious complication in 3 5% pregnancies and is characterized by hypertension, proteinurea, and edema [40, 131] The disproportionate ratio of angiogenic and anti angiogenic circulating factors is believed to be the major cause of preeclampsia [132, 133] Sera taken from preeclampsic women have elevated levels of the anti angiogenic circulating so luble fms like tyrosine kinase (sFlt1) and sENG, but reduced pro angiogenic factors like placental growth factor (PIGF) and VEGF [134, 135] Additionally, the levels of sENG predict the severity of the disease [136 138] A rat model of preeclampsia was created by injecting rats with adenovirus expression of sENG [138] Additionally, elevated sENG levels are also found in cancer, atheros clerotic lesions, coronary artery disease, hepatitis, diabetes, biliary atresia, and sickle cell anemia [119, 139, 140] Functionally, ENG has been found to interact directly with either Type II or T ype I receptor s. It is believed that E NG interacts with the two different TGF T ype I receptors, ALK1 and ALK5, and induces opposing angiogenic responses. ENG preferentially interacts and activates the pro angiogenic response via the ALK1/SMAD1/5/8 signaling. Conversel y, it can stimulate an anti angiogenic response by inducing ALK5/SMAD2/3 signaling [141 143] During embryogenesis, ENG


50 expression overlaps with ALK1 (especially at E6.5 13.5), TGF 1 and TGFBR2, but not with ALK5, TGF 2, nor TGF 3 [126, 144] Many in vitro HHT studies have been performed using human umbilical vein endothelial cells (HUVECs) derived from HHT patients [145, 146] Inde ed, HUVECs from newborn HHT1 overall showed reduced ENG surface expression. Further studies also showed reduced ENG levels in the activated monoctyes and HUVECs collected from HHT1 patients. When cells were transfected with a mutated transcript of ENG, sur face ENG expression was not detectable (due to a truncated protein or unstable mRNA transcript ) [111 147] There had been three independent Eng conventional mouse knockout (KO) models, with similar phenotypes: embryonic lethality at E10 1 1.5, defective yolk sac angiogenesis and heart development [56, 57, 148] Analysis of the Eng / mice at E9.5 revealed the KOs had abnormal vasculature and yolk sac anemia. Eng / embryos were capable of forming hig hly vascularized primary plexuses, but with retarded vessel branching. Thus, the models revealed that ENG plays a role in angiogenesis, but apparently not vasculogenesis. The in vivo finding corroborated with embryonic stem cells studies derived from one o f the Eng / models. Perlingeiro found that Eng +/+ and Eng / embryonic stem cells expressed the same amounts of the vascular precursors (VE Cadherin/FLK 1 and VE Cadherin/T IE 2), indicating vasculogenesis is not affected in Eng / In addition, ENG express ion is essential within the early hemangioblast and in promoting HSC specification [149] Other in vivo approaches in deleting Eng in mice take advantage of the cre loxP system. Recently, Mancini et al. executed complementary studies in wh ich ENG


51 function was activated specifically within ECs or SMCs [150] The authors created an expression strain of Eng (TgEng fl ) by which Eng recombinase is located as it will remove a stop codon flanked by loxP sites within the construct. In order to express Eng specifically in ECs or SMCs, they used the Tie2 cre and cre deleter lines, respectively, and bred them into an Eng / background. Thus, Tie2 cre;TgEngNull express ed ENG only within ECs and cre;TgEngnull expressed ENG only in VSMCs. Both models saw a rescue of VSMC recruitment to the layers. The study provided evidence that ENG plays a valuable role in recruitment of smooth muscle cells. Interestingly, activation of Eng within either cell type was not sufficient to rescue the Eng / embryos, indicating that ENG may be required earlier in vascular development, namely in the angio blast progenitor population. One possible function may be in arteriovenous identity. A conditional bi floxed Eng mouse was first described in 2007 by the Arthur lab and brought promise of specifically deleting Eng in different cells and tissues temporally [103] However, it is now too early to report any findings although several approaches are currently being attempted, including the projects within this chapter. The only published report so far of an Eng conditional (cKO) mouse is by the Arthur lab by which Eng was specifically ablated with in ECs using a tamoxifen inducible Cdh5(PAC) Cre ERT2 deleter line [151] The Cdh5(PAC) Cre ERT2 ; Eng 2f/2f mice were created at postnatal day (PN)2 (examined at PN7) and adult mice. The Cdh5(PAC) Cre ERT2 ; Eng 2f/2f mice failed to develop gross AVMs or consistent HHT like phenotypes. Though, closer examination revealed reduced angiogenic potential and venomegaly in adult mice. Retina


52 angiogenesis studies in the postnatal mice showed Cdh5(PAC) Cre ERT2 ; Eng 2f/2f mice had dela yed vascular remodeling of capillary plexus, increased proliferation, localized AVMs, and areas of AV shunts underwent muscularization, which may have resulted from the increased blood flow. Activin Like Kinase Receptor 1 (ALK1) ALK1/ACVRL1 ( Activin recept or like kinase 1 ) is located o n the human chro m osome 12q11 q14. ALK1 is a TGF Alk1 ge ne consists of ten exons (1 10), a lthough only 9 represent the actual coding sequence as exon 1 is noncoding. Because it has two transcriptional start sites, there are two mRNA isoforms. Structur ally, ALK1 has six domains: a 21 aa leader peptide (exon 2), extracellular (exons 2 4), transmembrane (exon 4), GS rich (exons 4 5), serine threonine kinase (exons 5 10), and a NANDOR box (within exon 10) [147, 152] Characterization of HUVECs and monocytes collected from HHT2 patients revealed that missense mutations in ALK1 that lead to amino acid substitutions account for about 60% of HHT2 cases [145, 153] Interestingly, t hree of the in vitro HUVEC and transfection studies demonstrated that not only was ALK1 still expressed on the cell surface, but in two ALK1 was behaving in a dominant negative manner. However, two of the three HUVEC lines were derived from families at ris k for developing HHT associated PAH, which could complicate the study and affect results. The drawback from the commonly used HUVECs in HHT research are the limited supplies collected from the newborns and that these cells are collected before any phenotyp es have the chance to manifest. Another fundamental problem is that cell culture conditions vary among groups of investigators


53 As with the Eng / mice, Alk1 / are embryonic lethal at midgestation due to defects in angiogenesis [154, 155] Supporting its importance in angiogenesis, a zebrafish model, in which Alk1 ortholog violet beauregard was suppressed, resulted in the formation of arteriovenous malformations (AVMs) [156] Additional murine studies targeting the expression of arterial marker EphrinB2 showed that Alk1 / mice lost this expression within the dorsal aorta. Interestingly, Eng / mice do not demonstrate this phenotype [154 157] Mouse M odels for HHT One of the most striking findings from the Eng and Alk1 null mouse models was that 30 70% of heterozygous mice ( Eng +/ and Alk +/ ) faithfully phenocopied the human disease [57, 158] Unfortunately, the appearance of HHT like phenotypes in mice is as variable and unpredictable as the human disease. Nonetheless, most labs studying HHT still utilize heterozygous mice. It has been suggested that the variability in phenotypes may be influen ced by modifier genes [159] Bourdeau and colleagues showed that there were strain dependent penetra nce in Eng +/ mice, as 30%, 50% and 70% of Eng +/ mice on pure C57BL/6, mixed, and pure 129/Ola (respectively) str ain background developed HHT like symptoms [160] The Alk1 conditional mouse was generated by the Oh lab and first described in 2008 [64] Simultaneously, the novel endothel ial specific cre deleter line L1cre, in which cre recombinase was driven by a 9.2 kb region of the Alk1 promoter [161] was introduced [64] This line is unique in that cre i s active beginning late in gestation, at E13.5 with spotty EC specific X gal staining. By E15.5 there is very strong X gal staining of ECs. This delay in cre is essential as it by bypasses the midgestation lethality of many conventional knockout models. An other unique characteristic of L1cre


54 is activity is restricted to the brain, intestine, yolk sac, and strongly in the lungs. It was previously reported that L1cre(+); Alk1 3f/3f and L1cre(+); Alk1 2f/2f mice consistently developed dilated, disorganized and tor tuous vessels in the lungs, GI tract and brain starting at E15.5 or E17.5, respectively, and death during late gestation (by E18.5) or shortly after birth, at PN5, respectively (Figure 3 1). Pulmonary vessels displayed irregular SMC layers. When the L1cre line was used to delete Tgfbr2 fl/fl and Alk5 fl/fl the cKO mice were unaffected, suggesting that ALK1 signals independently of TGFBR2 and ALK5. It further supports that ALK1 may actually signal via the BMP driven signaling pathway [64] Further studies were done with the Alk1 conditional mice that provided more ALK1 signaling insight. Alk1 was deleted globally in adult mice using a tamoxifen inducible R26 Cre ER line ( Alk1 iKO) [162] Adult mice consistently showed pulmonary and GI bleeding. Latex vascular casting revealed evidence of AV shunting in the lungs, with Alk1 3f/3f iKO mouse livers (Figur e 3 2). When acute wounds were given in the ears and dorsal skin of Alk1 iKO mice, de novo AVMs formed, suggesting that a second environmental injury, in addition to the genetic deficiency, is sufficient to inflict AVMs (Figure 3 2). The underlying patholo gical mechanism underlying HHT is still largely unknown. Interestingly, the genes responsible for HHT are known, their physiological functions have not been clearly defined. Based on in vitro and some in vivo studies, it is assumed that ENG and ALK1 work i n concert in a linear TGF /SMAD dependent manner within ECs to induce angiogenesis and maintain vascular tone so reduced function of either receptor can lead to the vascular defects seen in HHT. However, this has not been


55 tested directly in vivo thus we attempted to accomplish th is goal by generating Eng cKO mouse models. Although, the heterozygous models recapitulate HHT like phenotype, the vascular malformations are highly variable in formulation, location, and severity. The L1cre(+); Alk1 cKO and Alk1 iKO mice our lab previously characterized are the first murine models that displayed highly consistent AVMs in specific organs (among those, the lungs, GI tract, and brain). This affords the opportunity to test the ENG/ALK1 signaling hypothesis by generating Eng cKO mice using the s ame cre deleter lines. The expectation is that if these, indeed, signal in a linear fashion, the Eng cKO and Eng iKO mice would produce the same or overlapping phenotypes. Results Generation of C onditional (cKO) and T amoxifen I nducible (iKO) Eng K nockout M ice Eng bifloxed mice were previously generated by and obtained through the Arthur lab. In this conditional murine model, LoxP sites flank exons 5 and 6, which are part of the extracellular domain of the Eng gene (Figure 3 3A). When cre recombinase is pres ent, homologous recombination between the loxP site leads to the removal of these exons ( Eng 6 ). Consequently, Eng 6 results in a frameshift in the sequence that introduces a premature stop codon in exon 7 when translated into protein. The truncated pr otein (consisting of the extracellular domain from exons 1 4 and part of exon 7) is too short be functional and assumed to not even reach the plasma membrane, thus is a null allele [103] Mice were genotyped via PCR for the floxed allele using primers 5+6 (Figure 3 3B). There was approximate ly 80 bp nucleotides introduced


56 into introns 4 and 6 associated with the loxP sites, however, these did not interfere with normal function of the gene. Endothelial S pecific D eletion of V ascular Eng ( L1cre(+); Eng 2f/2f ) were V iabl e The first Eng cKO model g enerated was an endothelial specific ablation, utilizing the L1cre line. It was surprising that all Eng cKO (hereby designated cKO) mice were viable and without signs of reduced longevity or exhibiting any outward health problems, compared to control mice (littermates that were cre negative). Immunostaining of the endothelial layer with PECAM 1 (red) and the SMC layer (brown) of the vessels of the uterus, intestine, kidney, and liver revealed Eng cKO mice were developmentally indistinguishable (Figure 3 4). As L1cre is most active in the lungs and L1cre (+) ; Alk1 3f/3f displayed abnormal pulmonary vasculature in a highly consistent manner, the pulmonary vessel development was compared between control and Eng cKO mice at E15.5, E18.5 and adult (aged 2 months o r older) stages (Figure 3 5). The Eng cKO mice were developmentally indistinguishable from control littermates (Figure 3 5A no indications of hemorrhaging or dilated vessels at any stage (Figure 3 5). In adult lungs, neither the arteriole v essels no arteries exhibited any evidence of vessel dilation (Figure 3 5E,G,F,H). Additionally, Eng cKO pulmonary vessels developed uniform smooth muscle layers similar to Controls (Figure 3 In a population of mice aged 4 to 15 months, hem oglobin (Hgb) levels were measured to indicate whether any Eng cKO mice may be bleeding internally. Surprisingly, almost all the Eng cKO had Hgb levels that were normal for mice (10 20 g/dL) and to Control littermates (Figure 3 6A). Eventually, only two En g cKO mice, at an older age of 15 month, displayed any visceral phenotype. Latex vascular casting of


57 lungs taken from mice aged 1.5 years showed Eng cKO mice developed proper pulmonary vessels (Figure 3 6B), with no signs of AVMs as in L1cre(+); Alk1 2f/2f ( Figure 3 1J) It should be noted that one of the two mice that measured below 10 g/dL is the hemorrhagic mouse. Interestingly, the two Eng cKO mice both displayed abnormal livers, despite L1cre not being active in the liver. In one mouse, the liver was reg ressed and some vessels were visibly infused with latex (Figure 3 4C). In control mice, the latex dye is not normally visible. In the second Eng cKO mouse, the liver developed normally but there were five AVMs present, exemplified in Figure 3 fied view (Figure 3 Eng cKO mouse aged 15 months (one of the two mice in Figure 3 6A with Hgb levels below 10 g/dL) exhibited local hemorrhaging at two sites in the intestine (Figure 3 6D). There was mild AV shunting in few vascular beds, evidenced by incomplete penetration of the latex into the vein, but one telangiectasia in one hemorrhage site (Figure 3 In Figure 3 6E, PCR analysis for the null allele was performed using the indicated primer set 4+6 (Figure 3 1A) on DNA extracted from various organs from control and Eng cKO mice. The results show that L1cre was active and exons 5 and 6 were deleted in the expected organs of Eng cKO mice, such as the lungs, heart, intestine appendix, and uterus, but not in the liver or kidney and in any of the control organs. Reverse transcriptase (RT) PCR was also done but results were not as clear because mRNA was collected from whole tissues, thus the true reduction of mEng in vascular ECs may be difficult to detect.


58 X gal and I mmunohistochemistry C onfirmed C re was A ctive in ECs of E xpected O rgans The Eng cKO model also possesses the ROSA26 locus, which allows visualization of cre activity by X gal staining. It was confirmed that cre wa s active in the Eng cKO lungs, heart, eye, brain, intestine and female reproductive organs, indicated by the blue X gal precipitate (Figure 3 Eng cKO liver and kidney and all Control samples were not stained blue pr oving cre was not active in these locales (Figure 3 7K L, A Also, shown as an additional control is that lung and brain from L1cre(+); Eng +/2f mice were also stained blue but not the liver or kidney (Figure 3 5Q,R and S,T, respectively), further estab lishing cre is active and restricted to certain organs. Histological analysis of X gal stained Eng cKO lung (Figure 3 strongest cre activity. However, because the lung used for X gal staining was not inflated and the X gal staining is s o strong, it was difficult to prove in the pulmonary microvasculature that the EC staining was EC specific (compared to SMC specific) (Figure 3 Co staining of X gal positive cells in the uterus, and intestine with either PECAM 1 or SMA (Figure 3 8/ E, F respectively), indicate cre was active in ECs. Confirmed L oss of V ascular E ndothelial Eng d id n ot Affect V ascular D evelopment Immunofluorescence (IF) staining of ENG in Control and Eng cKO lungs reveal that ENG was overall reduced (Figure 3 9A,E). An arterial vessel was selected from each sample (Figure 3 9B,F) and the colocalization of ENG and PECAM 1 show that the number of ENG expressing cells were reduced in Eng cKO mice compared to control (Figure 3 9C,G). Quantification of the coloca lization rate confirmed there was a


59 ~50% decrease the numbers of ECs expressing ENG in the Eng cKO mice (Figure 3 9D,H,I). Thus, the data from the Eng cKO model surprisingly suggests that Eng cKO mice were largely unaffected by the late gestational loss o f Eng in vascular ECs. Eng iKO M ice D isplayed V ariable P henotypes The second knockout model was a global silencing of Eng in adult mice using the tamoxifen (TM) inducible cre line, R26 Cre ER In this line, cre is fused with a mutated form of the estrogen r eceptor and remains inactive until TM is introduced and binds to the mutated estrogen receptor, activating it. Furthermore, the construct is inserted into the ubiquitous ROSA26 locus, thus induction of cre is a global event. The Alk1 iKO model was created to determine what effects the loss of ALK1 in adult tissues would occur. A single IP injection of 2.5mg TM per 25g mouse body weight in R26 Cre ER (+) ; Alk1 2f/2f ( Alk1 iKO) was sufficient to induce the aforementioned AVMs and hemorrhaging (Figure 3 2) within 7 days of TM treatment, with death as early as 9 days but 100% fatality by 21 days after TM [162] Thus, the same dose of TM was first given to the R26 Cre ER (+) ; Eng 2f/2f ( Eng iKO TM1) mice that were at least 2 mont hs of age. Control mice were Eng 2f/2f littermates, which were R26 Cre ER negative, and injected with the same dose of TM. Surprisingly, the Eng iKO TM1 mice were completely viable and did not seem affected by the loss of Eng Eng iKO TM1 were able to surviv e over 12 month after TM injection, without overt health concerns ( Figure 3 14A ). The only noticeable outward phenotype was the appearance of abnormal vasculature in the ear tissue surrounding the ear identification (ID) tag commonly given to mice at the t ime of weaning (age 21 days) (Figure 3 10A). The appearance of AVMs was confirmed by vascular (blue) latex casting of the systemic vasculature system. The ear phenotype, which was seen in all Eng iKO TM1 mice, was slow to progress as dilated vessels were


60 n ot noticeable until 1.5 to 2 months after TM treatment (Figure 3 10G) and worsened over time, as seen at 6 months (Figure 3 10E, compared to control, C). When further studies with the Alk1 iKO mice were done in which 2 mm holes were induced within the ea r and dorsal skin, Alk1 iKO mice developed de novo AVMs in the change, in addition to genetic loss of Alk1 can cause AVMs (Figure 3 2G I). However, when 2 mm holes were induced within the ID tag free ear and dorsal skin of the Eng iKO TM1 mice, AVMs never formed (Figure 3 10,D,F). In support of the slow progression of phenotype (as seen in the ID tagged ear), autopsies of Eng iKO TM1 mice 1 or 2 month after TM did not reveal any abnormalities (data not shown). However, autopsies of 5 and 6 month post TM Eng iKO TM1 mice revealed grossly tortuous, disorganized vessels in the ovaries and appendix, but the phenotype appearance was not in a consistent manner. Two females w ere hemorrhaging and possessed abnormal ovarian vessels (Figure 3 month Eng iKO TM1 case, the ovarian vessels appear to connect to the intestine (Figure 3 as demonstrate in Figure 3 10J. Besides the ID tag associated phenotype, the other consistent phenotype seen in mice was mild AV shunting in the intestine, particularly in the large intestine (Figure 3 10K). Mice displayed cardiomegaly, shown here 5 months af ter TM (Figure 3 10L), beginning at 2 months post TM. However, hemorrhages were not found as in Alk1 iKO mice. Lung hemorrhaging, often seen in Alk1 iKO mice, was absent in Eng iKO TM1. Additionally, some Eng iKO TM1/high dose mice excreted red


61 feces, indi cative of blood in the feces. This would suggest internal bleeding; however, it was surprising that no hemorrhages were found during the necr opsy. A question of whether the single dose of TM was sufficient to use in the Eng iKO TM1 was raised. Compared to other literature utilizing TM inducible CreER lines, the Alk1 iKO was unique in that a single injection was enough to lead to a phentotype. However, the typical number and dose of injections average three to five times at 2mg TM [151, 163] Thus, two to three injections of the same concentration of TM was administered ( Eng iKO TM2 or 3/high dose ). Eng iKO TM2 or 3/high dose had high mortality rates ( Figure 3 14A), with mice very sickly or close to death at an averag e of 7 days after the first TM injection. Eng iKO TM2 or 3/high dose mice became hunched, lethargic and had scruffy fur, compared to control littermates (Figure 3 11A). Dilated vessels in tissue surrounding the ID tag were noticeable at 4 days after the fi rst TM treatment but AVMs did not form (Data not shown), probably because death occurred too soon. Cardiomegaly and lung hemorrhaging were not seen in Eng iKO TM2 or 3/high dose mice for the same reasons (not shown). There was 3/17 mice that had an abnorma lly small liver ( Figure 3 11B ). The GI tract appeared to be the most affected in the Eng iKO TM2 or 3/high dose mice. 6/17 mice exhibited a distended GI tract, however only 3 had major general intestinal hemorrhages. All mice have AV shunting in the intest ines and appendix ( Figure 3 11 C E ). A small AVM was found in the intestine of one Eng iKO TM2 or 3/high dose mouse ( Figure 3 11 E), however, global vascular defects as seen in Alk1 iKO mice was not found. The rapid death of Eng iKO TM2/3, but survival of En g iKO TM1 mice with a slow phenotype progression, suggests that multiple injections of TM at 2.5 mg/25 g mouse


62 BW can lead to complete ablation of Eng in all cells, not just the vasculature, which have detrimental effects on Eng iKO mice. Though TM can be toxic at high doses and 2.5 mg TM/25 g BW is high compared to what is typically used, TM toxicity seems unlikely to be the cause of de ath as control mice do not die. Multiple L ower D oses of T amoxifen (2.5mg/40g BW) L eads to M ore C onsistent A ppearances of P henotypes A lower dose (2.5 mg TM/40 g body weight) was then used, with two or three TM injections given to mice ( Eng iKO TM2 or 3/low dose ). Overall, there was an about 75% survival rate for either Eng iKO TM2 or 3/low dose with 25% dying at three weeks after the first TM injection (Figure 3 14A) As consistently seen in Eng iKO TM1/high dose mice, the vessels in the ear tissue around the ear ID tag became dilated and disorganized. The progression of the phenotype occurred sooner, with the appearance of possible telangiectases at 12 days post TM (Figure 3 12B). AVMs were clearly visible if left for longer periods, as seen at 52 days after TM treatment (Figure 3 12C). In the wound induction studies, AVMs did not form around induced wounds in the Eng iKO TM 2/low dose ear, 12 days after (Figure 3 12E). Even at 52 days after TM treatment and 50 days after the ear punch, AVMs did not form; although, vessels were more disorganized and in the there may have been telangiectases but it was not clear (Figure 3 12F). It should be noted that the images represented is the most extreme case of the induced ear wound. Other Eng iKO TM2 or 3/low dose ear punch mice were similar to controls. AVMs were never seen in wounds induced in the dorsal skin, seen here 12 days after wounds induction (Figure 3 12E,F). Wounds were capable of healing in the Eng iKO TM3/low dose dorsal skin. A fresh wound was given 42 days later and latex


63 injection was performed 10 days later, but there were still no AVMs present. Instead there was an inc rease of microvessels around the fresh wound (Figure 3 12I). The highly consistent ear ID tag associated AVMs among all the treatments in the Eng iKO mice mirrors the de novo AVMs that form in the wounds in Alk1 iKO mice; however, this phenotype arises fro m opposing stimuli. The ear ID tag could be considered a chronic stimulus and the wound can be an acute stimulus. Interestingly, as the Eng iKO mice did not form AVMs around the induced wounds, the Alk1 iKO mice did not form AVMs in the ear tag (Figure 3 2 J). A necropsy revealed Eng iKO TM3/low dose mice kept longer than 1.5 months after TM treatment experienced cardiomegaly (Figure 3 13AB), which is more likely a secondary affect to internal hemorrhaging or possible hepatic vascular malformation. Assessme nt of vascular malformations by latex injection shows that that, as seen in the Eng iKO at all TM doses, the GI tract was affected. The appendix also shows mild AV shunting in all Eng iKO/low dose mice (Figure 3 13D and E), and 1/17 Eng iKO TM3 mouse devel oping an AVM in the appendix three months after TM (Figure 3 13E). The liver of Eng cKO was largely similar to control mice (Figure 3 13F&G), however, 2/17 Eng iKO TM3/low dose mice had a regressed liver (Figure 3 13H). There was varying degrees of, though consistent, AV shunting within intestinal vessels compared to control mice (Figure 3 13J&K,I, respectively); of note, not all vascular beds were affected. 1/6 of Eng iKO TM2/low dose and 4/17 of Eng iKO TM3/low dose had hemorrhages in the intestine (Figur e 3 13J&K). Two Eng iKO TM3/low dose mouse deve loped AVMs specifically in the p similar finding as in the Alk1 iKO model. Eng iKO TM2 or 3/low dose female mice


64 consistently hemorrhaged and developed AVMs specifically in ovarian vessels, but not in uterine vessels (as seen in Alk1 iKO mice) (Figure 3 13M). Interestingly, two Eng iKO TM3/low dose males had visibly enlarged scrotums with bloody feces (Figure 3 13O, compared to control N ). T he necropsy revealed the appendix had been moved in the scrotal area, and there was no vascular defects in the male reproductive organs, nor was there ever found in either Eng cKO model. As was found in the Eng iKO TM1/high dose mice, there was not intesti nal bleedi ng. H owever, AVMs in the rectum were discovered (Figure 3 in the feces without any signs obvious internal hemorrhaging. Lastly, as the vessels of the head are often affected in HH T patients we decided to look the head va sculature of the Eng iKO mice. Examination revealed that abnormally dilated vessels in the turbinate of Eng iKO mice (Figure 3 13R). In two mice (one treated 3 times, the other 2 times of TM) telangiectases were p resent in the gums of the front two lower teeth (Figure 3 13S). Using multiple injections of the lower (2.5 mg TM/40 g BW) dose seems the most appropriate dose, with three times injection the best. Though, the appearance of vascular malformations in common organs were much more consistent in the Eng iKO/low dose mice, the phenotype is still heterogenous in terms of severity and location of AVM formation. Hgb measurements were taken of Eng iKO TM2 or 3/low dose mice to test for possible internal bleeding t wo to four months after TM treatment (Figure 3 14B). It was found that more Eng iKO TM3/low dose (>50%) mice gave measurements around 10%, indicating internal hemorrhaging. In two of the mice, the feces were bright red in color, suggestive of blood in the feces. However, despite these f indings the mice with the low


65 hgb did not display any signs of reduced health quality compared to controls or Eng iKO TM2 or 3/low dose mice with h gb values within the normal range. The 2f to 1f A llelic C onversion E fficiency at the D ifferent D oses C ontributed to the O bserved V ariable P henotypes To test whether the variability in the Eng iKO phenotype is a true phenotype or a technical issue concerning the use of tamoxifen, the conversion of the conditional (2f) to null (1f) a llele genomic copy of each allele was quantified by genomic southern blot (Schematic in Figure 3 15A) An Alk1 iKO southern blot was also performed to show the 2f to 1f allele conversion in the Alk1 iKO lungs and liver to act as a comparison for the Eng iK O samples. There was >95% conversion in the Alk1 iKO lungs. However, the liver samples taken from the same mouse varied. One mouse had complete conversion to the null allele but the other showed ~90% conversion (Figure 3 15B). However, due to the high back ground in the former sample on the blot, this may not reflect a true measurement. More Alk1 iKO southern blots will need to be run. The single high dose of TM appeared to have the same conversion rate in the liver samples as the Alk1 iKO mice (Figure 3 15 D). T here was a general trend in that there was more efficien t 2f to 1f allele conversion in Eng iKO mice given increasing numbers of TM injections. The high dose of TM effective deleted Eng there was >75% of the 2f to 1f allele in all Eng iKO mice. In th e triple injected mice, there was 100% conversion, indicating deletion of Eng in the mice (Figure 3 15D). Eng iKO lung samples will have to be repeated to confirm this trend but th ere is indications that complete ablation of Eng has detrimental consequenc es. O rgan s pecific differences in TM efficiency are seen as gDNA from the liver has more complete 2f to 1f conversion at the equivalent dose compared to the lungs (Figure 3 15 D F ). There a low 2f to 1f conversion rate in the


66 lungs of Eng iKO mice given the lower TM dose, which may account for the lack of vascular phenotype in the lungs of Eng iKO/low dose mice (Figure 3 15 D ). However, there was >50% conversion in the livers of the same mice, with >80% deletion in the triple injected mice (Figure 3 15F ). D iscussion ALK1 and ENG are assumed to work in a linear, canonical BMP driven signaling pathway during angiogenesis and in maintenance of the endothelium because both receptors are restricted to ECs, mutations in Alk1 Eng and Smad4 are related to HHT, and mice heterozygous for either gene can develop HHT like symptoms [57, 73, 75, 105, 158, 164] Thus, it was very surprising that the confirmed loss of Eng in ECs expressing ALK1 had no obvious developmental impact on Eng cKO mice. In fact, any vascular anomalities seen in Eng cKO mice were slow to progress as only two mice, months, presented mild vascular anomalities in the intestine and liver. Interestingly, despite L1cre not being expressed in the liver, bot h mice display abnormal livers. Additionally, one mouse exhibited two intestinal hemorrhages with a few vascular beds showing AV shunting at each site and only one visible telangiectasia in one site. This contrasts with the Alk1 cKO mice as all vascular b eds are affected. It was even more surprising that global deletion of Eng under the same conditions as the Alk1 iKO model did not severely impact the Eng iKO mice. A majority (75% of iKO mice injected thrice wi th the low dose of TM) survived without any m any health concerns. Though testing varying tamoxifen dosage and the number of injections in addition to the genomic southern blot quantification of the genomic copy of the null allele implies technical hindrances influenced the variability in the appearan ce of vascular abnormalities in the Eng iKO, the organs affected and the type of phenotype


67 seen were consistent. Additionally, though rare, similar pathologies were seen in the Eng cKO mice. This suggests the AV shunting, GI hemorrhages, and AVMs that form ed are legitimate phenotypes resulting from the loss of Eng Several organs (the ear, GI tract, female reproductive system, heart, liver) were commonly affected in the Alk1 and Eng iKO models There were only few cases where overlap seen. For example, two Eng the intestine, a structure in which Alk1 iKO mice commonly form AVMs However, in many cases the phenotype seen differed in severity, location, and in response to a stimulus. In the ear, AVMs formed in t he tissue surrounding the ear ID tag but not when a wound was given in the untagged ear (nor dorsal skin); Alk1 iKO mice formed de novo AVMs around the wound site. In the female reproduction system, it is quite curious that the ovary and uterus is specifi cally affected in the Eng iKO and Alk1 iKO model, respectively. The Alk1 3f/3f iKO liver often exhibits AVMs however Eng iKO livers did not consistently form the hepatic abnormalities These discrepancies indicate there may be disparate pathological mechani sms underlying HHT1 and HHT2. One possible contributing factor to the difference in phenotypes in the Alk1 and Eng iKO models is in the expression of ALK1 vs. ENG. ALK1 is more specifically expressed on arterial ECs while ENG is on all vascular ECs and in circulating endothelial cells (CECs), such as EPCs and HSCs. Additionally, it was previously reported in expression studies that in the pulmonary vasculature, ALK1 and ENG overlap only in the distal arterioles, venules and capillary bed [165] In the L1cre line cre is active in arterial ECs but as Eng is also expressed on vein ECs, loss of arterial Eng may not have as a severe affect as Alk1 loss. Also, the milder phenotype in Eng cKO


68 mice may be CECs expressing ENG may be compensating for the loss of vascular Eng [166] The variable phenotypes seen support the finding that genetic loss alone of either Eng or Alk1 is not always sufficient for the formation of AVMs. In the pre vious studies utilizing heterozygous mice, local treatment with VEGF, induction of repeated mechanical stress with magnets and induced inflammation have lead to the formation of AV shunting and abnormal vessels [167 170] The preferential formation of AVMs in response to chronic vs. acute stimulus (the ear ID tag vs. induced wound) in the Eng and Alk1 iKO models, respectively, strongly indicate Eng deficient ECs may form AVMs in response to an increase in local infl ammation at the injury site, while AVMs in Alk1 deficient ECs may form in response to an increase in pro angiogenic factors/angiogenesis at the site of injury. The finding that inflammation may contribute to AVMs is supported by previous findings in that v ascular lesions often form at sites of inflammation in Eng +/ mice (both secondary to spontaneous bleeding in the GI tract, eye or ear and induced by dextran sulfate in the intestine [169, 171] It is believed that monocyte infiltration at the site of injury and the presence of certain cytokines, such as tumor necrosis factor vascular ENG [151] Additionally, the Letarte g roup proposed that Eng is important in regulating endothelial nitric oxide synthase (eNOS) and that loss of Eng leads to increased superoxide production, instead of nitric oxide [172] This inflammatory response has not been tested or confirmed in Alk1 +/ mice. There is evidence that the activation of angiogenesis in Alk1 deficent conditions may is sufficient to lead to AVM formation. It was seen that treatment of Alk1 +/ brains with VEGF was sufficient to cause


69 dysp lasia [167] In comparing to HHT patients, there has been no study focusing on how specific or equivalent stimuli in humans, e.g. piercings, affect HHT patients. Although, there is a mention that a volleyball player had a large number of telangiectases on the tips of their fingers, indicating repetitive stress leads to telangiecta sia formation but a clear relationship to genotype has not been determined [100] The organs affected in the murine models are the same organs affected in HHT patients (discussed in detail in Chapter 1 ), establishing the importance ALK1 and ENG within the vasculature of these organs. One of the most striking findings is the lack of any phenotypes in the lungs in either Eng ablation models, even though Alk1 cKO models consistently do and there are higher incidences of PAVMs found in HHT1 patients [64, 162] The lack of phenotype in the pulmonary vasculature was also seen in another, albeit endothelial specific (using Cdh5(PAC) Cre ERT 2 ), Eng iKO. This may be explain ed, in part, from the genomic southern results in that there was low 2f to 1f conversion in Eng iKO/low dose mice. However there is lack of any reported evidence of PAVMs in any Eng or Alk1 heterozygous mice. The main reported pulmonary phenotype in adult Eng +/ mice was truncated pulmonary vasculature. It was previously suggested that this may due to the shorter pulmonary branches in mice [165] thus indicating a limitation in using mice as true model for HHT.


70 F igure 3 1. L1cre(+); Alk1 2f/2f mice developed AVMs in the brain a nd lungs and were lethal by PN5 A) Overview of control brain vessels at PN5. B) Overview of PN5 control brain vessels after latex vascular casting. In the adjacent panel, a magnified view sho ws very fine microvessels in which the latex remained within arterial vessels. C) Overview of PN5 mutant shows grossly dilated vessels in the brain. D) Latex vascular casting revealed AV shunting and formation of AVMs in cerebral vessels, as evidenced with latex within the veins. E) Control PN5 lung. F) Magnified view of PN5 control pulmonary artery and vein. G) Latex vascular casting shows proper pulmonary vascular development and fine branching of vessels. H) PN5 Alk1 cKO lungs were hemorrhagic and showed obvious vascular defects throughout the lung (exemplified by white arrows). I) Close up up a dilated and disorganized PN5 Alk1 cKO pulmonary artery and vein. J) Latex vascular casting revealed improper vascular branching of pulmonary vessels and large are a AVMs (magnified in inset). K) Histological view of PN5 control lung. L) Histological view of PN5 Alk1 cKO lung revealed blood cells throughout the tissue, not properly contained in vessels (denoted by asterisks). A, artery; V, vein; H, heart; LPa, left p ulmonary artery, RPa, right pulmonary artery; LPv, left pulmonary vein.


71 Figure 3 2. Global deletion of Alk1 in adult mice resulted in phenotypes in the lungs, liver, heart, uterus, and in areas surrounding induced wounds, and death within 21 days after T amoxifen introduction A) Alk1 iKO mice GI tracts were hemorrhagic. Latex vascular casting showed signs of AV shunting. B) Alk1 iKO mice developed cardiomegaly. C) Alk1 3f/3f iKO mice, hepatic AVMs developed (denoted by white arrow). D) The uterine vessels (purple arrows) were severely impacted in Alk1 iKO females. E) Interestingly, the vascular malformations were restricted to uterine vessels. (white box) magnified view of uterine wall shows AV shunting and dilated vessels in all vascular beds. (purple box) However, the ovarian vessels are not affected. F) A second example of the Alk1 iKO female reveals the generalized uterine vascular malformations are consistent (white box). (Purple box) Though it appears the ovarian vessels appear affected, a magnified vi ew shows that AVMs are present on the ovarian bursa, not within the ovary itself. G) Latex vascular casting of unwounded Alk1 iKO dorsal shows no AVMs developed; however, de novo AVMs did arise after wound induction (H). I) AVMs formed in wound induced ear s of Alk1 iKO mice. J) AVMs did not form in Alk1 iKO ear tissue surrounding the ear ID tag. Scale bar denotes 3 mm. A, artery; V, vein


72 Figure 3 3. Schematic diagram of the Eng 2f conditional allele and subsequent deletion of exons 5 6 in the presence of cre. A) When in the presence of cre recombinase, homologous recombination between the loxP sites results in the loss of exons 5 and 6. Consequently, the null allele would be translated into a truncated, nonfunctional protein because a frameshift in the seq uence in exon 7 would lead to a premature stop codon. B) Mice were PCR genotyped for the presence of the bifloxed allele using the primers 5+6. If the loxP site was present, PCR produced a product 490 bp. Otherwise the wild type allele would have a 420 bp product


73 Figure 3 4. L1cre(+); Eng 2f/2f ( Eng cKO) mice were viable and vasculature in various organs were comparable to controls. Immunostaining of the endothelial layer PECAM 1 (aged 2 months) Eng cKO mice were capable was developing proper vessels. A) PECAM 1 Eng Eng Eng Eng cKO liver. G, glomerulus.


74 Figure 3 5. Eng cKO lung vasculature was c omparable to controls at various stages of development and adult mice L1cre is highly active in the lungs so more attention was given to this organ. PECAM 1 and revealed comparable pulmonary vasculature in Eng cKO mice to Control littermates. A, E15 .5 Eng Eng the Eng cKO lungs. G) Control lung bronchiole and artery. H) Eng cKO pulmonary bronchiole and artery. Art, arteriole; A, artery; Br, bronchiole.


75 Figure 3 6 Only two old (15 month) Eng cKO mice displayed an y vascular phenotype. A) Hemoglobin (Hgb) levels measured in adult mice (>4 months) show the majority of Eng cKO mice had Hgb levels within the normal range, thus not experiencing internal hemorrhaging. One of the Eng cKO below 10 g/dL Hgb is the mouse rep resented in (D). B) Latex vascular casting of 15 month old Eng cKO mouse revealed properly formed pulmonary vasculature, with no signs of hemorrhaging or AVMs. C) The lobe of one Eng cKO liver was regressed and the visibility of the blue latex suggests vas cular malformations; The second Eng cKO liver developed normally overall. However, 5 AVMs were found exemplified by the white arrow. D ) One Eng cKO mouse exhibited two focal hemorrhages in the GI tract (purple a rrows). ) Latex vascular casting revealed mild AV shunting in a few, not all vascular beds. (inset) One clear AVM was found that may have been the source of one hemorrhage site. E) PCR for the null allele using primers 4+6 produce a 550 bp product if the null allele was produced. Otherwise the PCR would be too large and unstable to produce. Testing for the null allele in various organs confirmed Eng was deleted in the lung, heart, intestine, a drenal gland, spleen appendix and uterus, but not in the liver and kidney of Eng cKO mice. Control mice, which lack L1cre, did not produce any bands, indicating the null allele was not present.


76 Figure 3 7 As the mice also contained the ROSA26 gene, X gal staining of various organs was performed to confirm L1cre wa s active in expected organs. Control A) lung, B) heart, C) liver, D) kidney, E) eye, F) brain, G) intestine, and H) female reproductive organ were not stained by X gal, confirming the absence of cre. Cre activity was visualized in Eng cKO as I) lung, J) he art, M) eye, N) brain, O) intestine, and P) female reproductive organs were stained blue with X gal. As expected the K) liver and L) kidney were not stained. X gal staining of select L1cre (+) ; Eng + /2f organs show the same expected X gal staining trend. The Q) lung and T) brain stained blue, but not the R) liver or S) kidney.


77 Figure 3 8. Histological evaluation of X gal stained Eng cKO organs confirm cre was active specifically in ECs The strong X gal staining indicates that L1cre is highly active in the l ung. As this is an uninflated lung section and the X gal staining is so strong in the lung, it is difficult to differentiate between EC (A) gal is specific to the endothelium as the SMCs are not stained blue in the represented art eriole vessel 1 and SMA staining of uterine vessels show L1cre is active in ECs. E) Another view of a uterine vessel reveals EC specific X 1 and SMA staining of intestinal vessels reveal EC specific X gal staining. F ) Another example of the Eng cKO intestinal vessel. A, artery; Art, arteriole


78 Figure 3 9. Colocalization staining of PECAM 1 and ENG The number of ECS expressing ENG were reduced in adult Eng cKO lungs. Overview of a lung ssion was reduced in Eng cKO mice (E) compared to Control littermates (A). An arterial vessel was selected from the Control (B) and Eng cKO (F) lungs and magnified. Colocalization between PECAM 1 and ENG were compared (denoted by white pixels, C, G) and qu antitated. D,H) graphical representation of colocalization rate (area between two straight lines) confirmed that the number ECs expressing ENG was reduced in Eng cKO. I) Numerical representation show there was ~50% colocalization in the Eng cKO mice.


79 Fig ure 3 10. Eng iKO mice given a single IP injection of high dose TM were viable and phenotypes developed much more slowly than Alk1 iKO mice. A) The only obvious phenotype was the development of telangiectases in ear tissue surrounding the ear ID tag, shown here 6 months after tamoxifen treatment. B) Latex injection of control uninjured ears act as control of how the latex dye does not readily enter the microvasculature of the ear. C) Vascular latex casting of the systemic vessels after 6 months demonstrates many defined microvessels in the ear tissue surrounding the ear tag. E) Conversely, dilated, tortuous vessels and AVMs were seen in Eng iKO TM1/high dose mice. G) A noticeable dilation of vessels surrounding the tagged ear tissue was slow to progress, wit h the abnormal vessels visible by 2 months. D) Control and F) Eng iKO TM1/high dose mice given a 2 mm wound in the untagged ear did not develop AVMs. H) Control female reproductive organ after latex. Latex casting displayed gross vascular defects in the ov aries and fallopian tubes, but not uterus, in females 5 months (I) and 6 AV shunting was commonly seen in the appendix (J) and intestine (K), there were two Eng iKO TM1/high dose mice that demonstrated severe AVMs in the append ix. Eng iKO TM1/high dose collected more than 2 months after TM displayed cardiomegaly (L).


80 Figure 3 11. Eng iKO TM2 or 3/high dose died within seven days of treatment. A) Eng iKO TM3/high dose mice were lethargic, hunched, and in some cases close to dea th by day seven after TM. B) 3/17 Eng iKO TM3/high dose mice had abnormal livers. C,D,E) Eng iKO TM3/high dose mice often displayed distended intestines, and in some case, not all, hemorrhaging and AV shunting. (inset of D) Magnified view of vessels reveal ed dilated arteries and veins, in addition to massive AV shunting, but no obvious vascular malformations. E) A small AVM in the intestine was found in one mouse.


81 Figure 3 12 Eng iKO mice give n two to three low doses of TM developed the same pattern of v ascular lesions in a shorter time period. A) Fine microvessels properly form in control ID tagged ear tissue. B) Eng iKO TM2, low dose mice were capable of developing dilated, tortuous vessels by twelve days after TM treatment. C) Eng iKO TM3/low dose mice displayed gross AVMs in 52 days in a similar manner of Eng iKO TM1/high dose mice. No AVMs developed in control ear punch (D) nor skin wound (G), Eng iKO TM2, low dose ear punch (E) nor skin wound (H), or Eng iKO TM3/low dose ear punch (F) nor skin wound (I). Although, there appears to be mild telangiectases in the Eng iKO TM3/low dose ear punch (magnification of F). It should be noted that images are representations of the most extreme phenotypes. Other Eng iKO, low dose do not show this phenotype. Aste risks denote wound site.


82 Figure 3 13. Eng iKO TM2 or 3/low dose developed vascular phenotypes in more consistent manner but with variable severity. Eng iKO/low dose mice developed cardiomegaly (B), as compared to A) control littermates. The appendix (C E ), liver (F H), GI tract (I K), female (L M) and male (N O) reproductive organs, and head vessels (O S) were affected in Eng iKO/low dose mice. C) Control appendix. D) In many cases, there was mild AV shunting in the appendix of Eng iKO/low dose mice. E) 1 /17 Eng iKO TM3, low dose mice developed an AVM 3 months after tamoxifen. In many cases, the G) iKO TM2/3/low dose livers were similar to F) controls. H) However, 2/17 iKO patch present. J K) Varying degrees of multifocal AV shunting was found in all Eng iKO/low dose mice intestines. K) In 2/17 Eng iKO TM3/low dose mice, reproductive organ. M) AVMs developed specificall y in ovarian, not uterine vessels of all Eng iKO/low dose females. N) Control scrotum, with measurement from penis to anus. O) Eng iKO TM3/low dose scrotum is (magnified view in yellow box). P) Anatomy of the control head. Q) Control turbinate has picked up some latex. R) Two examples of Eng iKO/low dose mice that developed AVMs within the turbinate. S) 1/17 Eng iKO TM3/low dose mouse developed vascular malformations in gums.


83 Figure 3 14. Kaplan Meier survival curve for each Eng iKO treatment and Hgb measurement of Eng iKO/low dose mice. A) A summary Kaplan Meier survival curve for each Eng iKO treatment. Controls were pooled together as there was no difference in longevity in response to TM among the Control groups. B ) Hgb levels in Eng iKO TM2 or 3/ low dose and Controls were compared 2 to 3 months after TM treatment. Interestingly, it was observed that despite having low Hgb counts in a few Eng iKO TM3/low dose mice, not all presented any overt health concerns besides the appearance of white paws.


84 Figure 3 15 Testing efficiency of conversion of the 2f to 1f allele conversion between the different TM treatments in Eng iKO mice to Alk1 iKO mice. A) Schematic for genomic southern blot o f Eng. The gDNA was digested with BamHI (denoted by black lines). A probe was generated that would recognize any fragments containing exons 7 8. For the 2f allele, the probe would recognize a 4.1 kb fragment. For the 1f allele, a 3.4 kb fragment would be d etected B) Southern blot of Alk1 iKO lung and liver samples provides a comparison point of the efficiency of the Alk1 2f to 1f conversion with the single high dose of TM. C ) Genomic PCR using primers 5+6 (floxed allele) and 4+6 (1f allele) was done to con firm the genotype of each sample used. At least three samples each of Controls and each TM treatments were collected. The lung and liver were tested. D ) Southern blot of Eng iKO/low dose lungs with graphical representation of quantification of the 2f to 1f allele from each sample. Multiple injections of the low dose do not appear to effective convert the 2f to 1f allele, which may explain the lack of pulmonary phenotype. E ) The use of the higher dose of TM leads to >75% conversion to the 1f allele in the li ver. F ) TM appears to be much more effective in the liver overall as there is a higher 2f to 1f conversion rate in liver samples taken from the same Eng iKO/low dose mice as in C. Additionally, increasing the number of injections lead to higher conversion rates.


85 CHAPTER 4 SPATIOTEMPORAL ROLE OF ENDOTHELIAL TGF SIGNALING DURING DEVELOPMENT In early studies and for a long time, investigations on the TGF signaling mechanism in ECs have focused on the interaction between the type I receptors, ALK1 and ALK5 [148, 155] Thi s relationship is of interest to those in the HHT field, despite there never being a connection between ALK5 and HHT, because there was a belief that an aberrant balancing act between ALK1 and ALK5 within ECs contributed to the development of HHT lesions However, we previously provided evidence that ALK1 signaling is independent of ALK5 and, more surprisingly, TGFBR2 as mice in which Tgfbr2 and Alk5 were ablated using the EC specific L1cre line were viable [64] The indication that ALK1 signaling is separate from the TGFBR2, and instead prefers BMP signaling [65, 66] raises the question of whether TGF superfamily ligands are the physiological ligands for ALK,1 as previously believed. It was quite curious that the cKO mice were seemingly completely unaffected by the loss of Alk5 and Tgfbr2 within the ECs. Thus, another question that arises is whether different TGF actually have spatiotemporal roles during embryogenesis, particularly in the endothelium. To examine this we investigated the mechanism of cerebrovascular development. Neurovascular D evelopment Along with the vasculature the neural system is one of the first organ systems to develop in vivo [173] The development of the neurovasculature is driv en solely through angiogenesis. It was initially believed that angiogenesis in the central nervous system (CNS) was a passive process, in which vessels devel op in response to the needs of the developing brain. However, there is mounting evidence suggesting that CNS


86 angiogenesis is indeed a dynamic process [173] Additionally, although it is accepted that there ar e overlapping transcription factors that regulate both neurogenesis and CNS angiogenesis, it is still unknown whether neuronal development and cerebral angiogenesis occurs independently. There are two categories of brain vessels: the pial vessels (which de velop into the venous sinuses) and the periventricular vessels (the arterial networks). Pial vessels are present more widely throughout the brain at E9, whilst periventricular vessels are the dominant vessels within in the telencephalon /forebrain region [174] For this study the periventricular region is of interest. Within the telencephalon EC migratio n of periventricular vessels begins directionally from dorsal to ventral then lateral to medial (Figure 4 7D, top image). This mechanism is still largely unknown, thus we sought to see whether endothelial TGF signaling pathway may be involved. Nishimura reported from in vitro studies that paracrine astrocyte EC signaling was essential in maintaining vascular integrity. Specifically, they found that the integrin expressed on astrocytes w ere responsible for activating the latent TGF then induce TGF downstream targets [175] However, this has yet to be confirmed in vivo Integrin A ctivation o f TGF S ubfamily L igands Synthesis and activation of the TGF subfamily of ligands is a tightly regulated process [176, 177] (Described in detail in Figure 4 1). As mentioned, TGF are initially secreted as late nt proteins bound by the LAP that may be deposited in the ECM until released, typically in a context and tissue dependent manner. There are several modes by which the TGF may be freed. Environmental changes, such as heat and pH changes, have been shown a s examples [178] Thrombospondin 1 (Tsp 1) is a major


87 proteolytic activator of the ligand. Tsp 1 acts by binding the LAP, causing a conformational change in the LAP that forces it to lose its hold of the TGF Integrins are also well known TGF ligand activators [178] sub subunits ( 1 9, IIb, D, E, L, M, R, V, X) subunits ( 1 8) make up at least 24 different integrins [179] Integrins activate the ligands by reco gnizing and binding to specific Arginine Glycine Aspartate acid (RGD) sequences found within the LAP With the integrin bound, the LAP is removed from the TGF a conformational change [179, 180] TGF mediated activation is relevant only for TGF TGF [178] Functionally, integrins are involved in a variety of cellular processes, most notably cell adhesion, but has been found to be important in angiogenesis as some are found, particularly v, to be upregulated during angiogenesis in vivo Of the 24 types of integrins identified, eight have been associated with endothelial cells [180, 181] Two classes of embryos were found from conventional knockout itgav / ) and of itgb8 / ) [182, 183] The majority (~80%) of Class I embryos from each knockout model died at midgestation. The remaining 20 30% of embryos developed cerebral hemorrhagi ng starting around E10 12. These mice were born with brain hemorrhages and cleft palates (as previously reported in Tgfb3 / mice [184] ) and died shortly after birth. A double knockout of the TGF 3 ( Tgfb1 RDE/RDE /Tgfb3 / ) ligands resulted in similar cerebral hemorrhaging within the forebrain at midgestation. In this model the aspartic acid of the TGF LAP RGD sequence wa s changed to


88 glutamic acid and the TGF hemorrhaging seen in the general knockouts suggests TGF mediated activation of the TGF development, particularly in the periventricular vessels [185] Additionally, the period when the phenotype was seen in the various models from E10.5 13.5, indicates this action occurs specifically during midgestation Furthermore, conditional knockout models of itgav and itgb8 usin g Nestin cre and GFAP cre ), not ECs (deleted using Tie2 cre), are required [186, 187] More recently, a conditional murine knockout model of Tgfb3 cre; Alk5 fl/fl was characterized in which the authors reported cerebr al hemorrhaging in embryos. H owever, is not clear when the cerebral hemorrhaging first appeared in the Tgfb3 cre; Alk5 fl/fl as E14.5 was the earliest reported stage [188] Though, the phenotype correlation among the establishment of cerebral vasculature by activating the TGF ligands, it is not clear if the mechanism continues onto another neuroepithelial cell or another cell type. Furthermore, the consistent and close timing of the specific phenotype occurring during late mid gestation also suggests that this establishment is essential during a specific time period. Thus, we tested in vivo whether paracrine astrocytic integrin activation of TGF signaling o n ECs during midgestation is essential for proper cerebral vessel development. We focus more downstream of the TGF ALK5, and expect that deleting each from ECs at midgestation would result in t he same cerebral hemorrhaging.


89 Results Characterization of an E ndothelial S pecific Alk1 cre K nock in L ine TGFBR2 and ALK5 are involved in a variety of cellular processes and expressed on many cell types. There have been previous reports of conditional murine knockouts wherein each receptor has been ablated from ECs using different EC specific cre deleter lines. Tie1 cre ;Alk5 fl/fl and Tie1 cre ;Tgfbr2 fl/fl recaptitulated the null phenotypes [55, 177] Although, these models may not reflect the true role of ALK5 and TGFBR2 in ECs or angiogenesis because cre activity in the Tie1 cre line begins around E8 [189] which is before any embryonic vascular events commence at E8 9. We previously described mouse models in which Alk5 and Tgfbr2 w ere deleted in ECs using the L1cre line in which cre is driven by a 9.2 kb region of the Alk1 promoter. In this line cre is active beginning at late gestation, at E13.5. The knockout mice were viable and seemingly unaffected by the loss of either receptor in the ECs, but the lack of phenotype could be due to the late cre activity. The use of Tie2 cre, in which cre begins by E8.5 [190] yielded variable results. Tie2 cre(+); Alk5 fl/fl mice were able to survive past mi dgestation; however, the Tie2 cre ; Alk5 fl/fl mice displayed cardiac edema that resulted in cardiac failure and death by E13.5 [191] As for Tie2 cre Tgfbr2 fl/fl embryos, 65% mimicked the conventional KOs, and the rem aining ~35% that survived past midgestation were phenotypically indistinct from control littermate up to E12.5 but by were embryonic lethal by E13.5. The authors reported, but did not show, that mutant embryos were edemic and hemorrhagic, and stated the c ause of death was presumably [192] More recently, the tamoxifen inducible endothelial specific VE Cadherin (Cdh5(PAC) Cre ERT2 ) line was used to delete Tgfbr2 fl/fl at E11.5, which resulted in embryonic lethality between E15 18 due to cardiac defects [193]


90 Interestingly, the authors of the Tie2 cre and Cdh5(PAC) Cre ERT2 mediated ablation models discussed cerebral hemorrhaging, however, did not further elaborate on the phenotype as the studies focused more on cardiac development [151, 191, 193] To assess the importance of TGFBR2 and A LK5 on ECs during midgestation as well as re solve timing issues related to the activation of cre, we generated an d used a nov el EC specific cre knockin line ( Alk1 GFPc re ), in which a GFPCre fusion gene was inserted into the Alk1 locus (Figure 4 2 A) The GFPcre cassette is inserted into intron 3 and replaces exons 4 8. Cre is still expressed in Alk1 expressing endothel ial cells because a 9.2 kb region of the Alk1 promoter, up to exon 2, is sufficient to drive Alk1 expression. In t h e Alk1 GFPc re line cre is expressed beginning at E9.5 in a patchy manner throughout the embryo (Figure 4 1B), including the head (Figure 4 2 B is drastically increased at E10.5 and active in ECs of all organs (Figure 4 1C), in a similar fashion to previously reported Alk1 expression studies. Further analysis of transverse sections of the X gal stained head revealed Alk1 GFPc re r evealed cre was active in ECs invading the neuroepithelium (Figure 4 Endothelial S pecific D eletion of Tgfbr2 and Alk5 R esulted in C erebral H emorrhaging B eginning at E11.5 and E mbryonic L ethality by E15.5 and E14.5, R espectively It has been repo rted that itgb8 / embryos that survived to birth, Tgfb 1 RDE/RDE ; Tgfb 3 / embryos and surviving itgav / embryos exhibited specific cerebral hemorrhaging within the ganglionic eminence (GE) during late mid gestation, indicating a vital role of TGF signali ng in proper development of the neurovasculature. Furthermore, it was shown that specifically deleting Itgb8 and itgav in ECs, using Tie2 cre, did not result in any vascular defects, but doing so in the neuroepithelium, using Nestin cre, did. This suggests v integrins within ECs are not essential in cerebrovasular


91 development However, it is not clear whether the activated ligand binds to another neuronal or a vascular cell. When our collaborators, the McCarty lab in Texas, used Nestin cre to ablate Tgfbr2 in neuroepithelial cells, Nestin cre(+); Tgfbr2 fl/fl mice survived with no obvious phenotypes or other developmental defects, forgoing the possibility that another neuronal cell is involved (Data not shown). Conversely, when Tgfbr2 or Alk5 wa s silenced in ECs, there were no live Alk1 GFP cre ;Alk5 fl/fl (Alk5 cKO ) nor Alk1 GFP cre ; Tgfbr2 fl/fl (Tgfbr2 cKO ) pups were born, indicating these were embryonic lethal. E mbryos were collected at embryonic day ( E)10.5, E11.5, E13.5, E14.5, E15.5 and E18.5 to determine whether the cerebral hemorrhagi ng occurred and when, as well of time and possible causes of death. At E10.5 Tgfbr2 and Alk5 cKO embryos were indisintguishable from control littermates (Figure 4 3A B ). It should be noted that controls were litter mates that lack the Alk1 GFP cre ;Alk 5 + /fl or Alk1 GFP cre ;Tgfbr2 + /fl At E11.5 all Alk1 GFP cre ;Tgfbr2 fl/fl and Alk1 GFP cre ;Alk 5 fl /fl exhibited cerebral bleeding within the forebrain region (Figure 4 3 C D ). The hemorrhaging remained largely within defined areas o f the brain likely the GE and midbrain then progressively worsened during gestation. Interestingly, hemorrhages were only seen in the central nervous system (CNS), and no other organs appeared to be affected. By E13.5 the head of Alk1 GFP cre ;Tgfbr2 fl/fl emb ryos were edemic ( Figure 4 3E ). Alk1 GFP cre ;Alk 5 fl /fl mice died a day earlier but did not exhibit the same level of edema as Alk1 GFP cre ; Tgfbr2 fl / fl embryos (Figure 4 3F). Additionally, the chest cavity of Alk1 GFP cre ;Alk 5 fl /fl embryos appeared distended, sug gesting Alk1 GFP cre ;Alk 5 fl /fl may have had cardiac defects. Alk1 GFP cre ; Tgfbr2 fl /fl embryos died by E14.5, with littermates undergoing various stages of death, from having very faintly beating hearts and close to


92 death or already dead and undergoing desorptio n (Figure 4 3G). There was bilateral cerebral bleeding in the cKO models starting at E11.5 (Figure 4 3H) and the bilateral bleeding also worsened over time, represented here at E13.5 (Figure 4 3I). In some, not all, Alk1 GFP cre ; Tgfbr2 fl /fl and Alk1 GFP cre ;Al k 5 fl /fl embryos, hemorrhages were seen along the spinal cord (Figure 4 3J and K). Alk1 GFPCre ;Tgfbr2 fl/fl and Alk1 GFPcre ;Alk5 fl/fl E mbryos F orm G lomeruloid like V ascular S tructures in the G anglionic E minence H&E staining and PECAM 1 (red) staining for ECs o f transverse sections of the E10.5 embryonic head revealed that Tgfbr2 fl/fl and Alk5 fl/fl cKO embryos (a representative cKO section shown in Figure 4 comparably to control littermates (Figure 4 taining shows the heads of Alk1 GFP cre ; Tgfbr2 fl /fl and Alk1 GFP cre ;Alk 5 fl/fl have perforations within the GE (Figure 4 4G and K), as compared to the control (Figure 4 4C). Interestingly, PECAM 1 staining revealed that vascular endothelial cells were clumped into glomeruloid like structures, which are not seen in control littermates (Figure 4 4D) in both Alk1 GFP cre ; Tgfbr2 fl /fl and Alk1 GFP cre ;Alk 5 fl /fl models (Figure 4 4H and L) The aggregate of EC cells suggest the lack of proper EC migration. The aberrant v essel development appears to be specific to the neuroepithelium as the vessels lying just outside the neuroepithelial tissue are not dilated (Figure 4 4I and M) and are comparable to the control vessels (Figure 4 4E). At E11.5 the neurovasculature appears to be the only organ system to be affected as the heart and lungs develop as control littermates (Figure 4 4F, J, and N).


93 Alk1 GFPCre ;Tgfbr2 fl/fl E mbryos D ie by E14.5 with CNS S pecific V ascular D efects, but all O ther O rgan S ystems are Largely U naffected H&E stained sections of E13.5 heads confirmed much of the hemorrhages within Alk1 GFPC re ;Tgfbr2 fl/fl embryos were around the abnormal GE and tele n cephalon. The fourth and third ventricles of th e brain were grossly distended (Figure 4 5B ). Further examination o f the vascular ECs of the GE and telencephalon reveal the vessels to be disorganized, tortuous and disassociated with surrounding neural tissues as in the control Vessels of the dicephalon are dilated and irregular, which made the vessels more vulnerable to rupturing and leaking out. Nucleated blood cells were found within the tissue and cavities, not contained within the vessels. Phenotypically, the brain appeared to be the only organ affected in Alk1 GFPC re ; Tgfbr2 fl/fl embryos. Further examination of t h e heart and lung development showed these did not seem affected by loss of endothelial TGFBR2 (Figure 4 5C and F). To analyze proper blood vessel formation, the smooth muscle layer was additionally stained with smooth muscle actin (SMA). The vessels of Alk 1 GPFCre ;Tgfbr2 fl/fl lungs were similar to control vessels in terms of the development of the smooth muscle and endothelial layer layers (Figure 4 5D and G) being unaffected. Likewise, SMA staining of the dorsal aorta was similar in both the Tgfbr2 cKO and controls (Figure 4 5E and H). These findings suggest that deleting Tgfbr2 in ECs specifically during midgestation has a grave impact on brain development, particular angiogenesis, but not as much on cardiac and lung development. Alk1 GFPCre ; Alk5 fl/fl E mbry os Die by E13.5 with Evidence of Gross Cerebral Vessel and Cardiac D efects In the Alk1 GFPC re ;Alk5 fl/fl t he GE is perforated, as seen in the Alk1 GFPC re ;Tgfbr2 fl/f l and blood cells are seen within the third ventricle and tissue (Figure 4 6A and B)


94 However, a magnified view shows that the GE nor di ence phalon is as severely damaged in Alk1 GFPC re ;Alk5 fl/fl embryos as in the Alk1 GFPC re ;Tgfbr2 fl/fl embryos (Figure 4 6A and B, panels two and three) There is no evidence of edema of the head as in the Tgfbr2 cKO. P ECAM 1 staining of ECs presents vessels that are disorganized in the GE but the dience phalon vessels are not as impacted by the endothelial loss of Alk5 (Figure 4 6A and B). The development of the pulmonary vasculature was also not affected but the Alk1 GFPC re ;Alk5 fl/fl heart appears undeveloped (Figure 4 6C and E) The ventricles are trabeculated but the ventricular septum does not appropriately fuse at the atrioventricular cushions. The dorsal aorta is not round but irregular in shape (Figure 4 6D and E).The endothelial layer of the dorsal aorta is developed as in the control (data not shown), however, SMA staining reveals the smooth muscle layer was diffused, compared to the control, suggesting a failure in the resolution phase of establishing the vess el. Thus, the vSMCs are more diffuse around the aorta instead of established around the vessel. Discussion The angiogenic processes involving TGF ng are still poorly understood. In many cases in vivo models did not correlate to in vitro finding, nor are conditional KO models phenotypes in agreement. For one example, e ndothelial specific Tie1 mediated a blation of the generally expressed receptors Tgfbr2 and Alk5 and the majority of Tie2 cre(+); Tgfbr2 fl/fl are embryonic lethal, like conventional knockouts, at E10 10.5 due to aberrant yolk sac vasculogenesis and angiogenesis [55, 192] In contrast, Tie2 cre; Alk5 fl/fl 35% Tie2 cre; Tgfbr2 fl/fl and Cdh5(PAC) Cre ERT2 ; Tgfbr2 fl/fl embryos survive past midgestation, however, these died of cardiac defects during late gestation [191]


95 Contradictory, still, our lab reported that conditional kno ckouts of both receptors using another endothelial specific transgenic line L1cre, were not developmentally affected [64] The timing of c re activity cou ld be a contributing facto r in the contrasting phenotypes. In the Tie1 cre line, Cre activity begins at E8, around the same time the vasculature forms (E8 9), and slightly before the reported first expression of TGFBR2 [194] Hence, the null nor Tie1 cre deletion models would reflect the true function of these receptors in angiogenesis at midgestation. As for the L1cre line, Cre is not initiated until late gestation, at E13.5 [64] The Tie2 cre (E8.5) and Cdh5(PAC) Cre ERT2 (induced at E11.5) models have slightly later cre expression [190] however, the phenotypes were variable or, as in the case of the Tie2 cre(+); Tgfbr2 fl/fl descriptions were vague [192, 193] The differing phenotypes imply a temporal role for ALK5 and TGF BR2 in ECs. Thus, w e generated a novel Alk1 cre knock in line Alk1 GFPCre The slight delay in the initiation of cre activity at E9.5 provi ded an opportunity to more accurately examine the roles endothelial TGFBR2 and ALK5 play in angiogenesis during midgestation than previous models. To test specifically examine the spatiotemporal role of TGFBR2 and ALK5 during embryogenesis, we focused on t he development of the cerebral vessels. Nishimura et al postulated from in vitro of a secreted TGF [175] Several independent knockout models, such as the itgav / [182] itgb8 / [183] Tgfb1 RDE/RDE /Tgfb3 / [185] Tgfb3 cre;Alk5 fl/fl [188] and Id1/3 / [TGF downstream targets] [195] strongly support this hypothesis and all develop the same cerebral hemorrhaging within the forebrain region. Conditional KO of itgav / and itgb8 /


96 in neuroepithelial, but not endothelial, cells also result in the hemorrhagic phenotype, confirming neuronal derived [186, 187] However, it is not clear what type of cell may be the target cell. We present in vivo confirmation that endothelial TGFBR2 and ALK5 are the targets as each developed the same cerebral hemorrhaging. Further establishing TGF on on ECs is involved, our collaborators found Nestin cre(+); Tgfbr2 fl/fl had no phenotype; though, the data was not shown here. We show that EC specific deletion of Alk5 and Tg f br2 using the Alk1 cre knock in resulted in a consistent and highly specific ce rebral hemorrhage phenotype that ini tially manifested at E11 11.5 Though there has been previous endothelial specific KOs of each receptor that reported cerebral hemorrhaging in some embryos, we are the first to show that the GE is perforated and vessels disorganized in Alk1 GFPCre ; Tgfbr2 fl/fl and Alk1 GFPCre ; Alk5 fl/fl embryos in a similar, consistent fashion as the itgav / itg b 8 / Tgfb1 RDE/RDE /Tgfb3 / models Interestingly, there was a slight delay in the appearance of the cerebral phenotype in the su rviving Tie2 cre ; Alk5 fl/fl and Tie2 cre;Tgfbr2 fl/fl compared to our Alk1 GFPCre lines, however, the results were still consistent in that a phenotype was seen in the ALK5 cKO than the TGFBR2 cKO. Additionally, the previous endothelial specific knockout mod els report embryos died due to cardiac defects. We show that Alk1 GFPCre ; Alk5 fl/fl display defects in the ventricular septum, but in a manner similar to the Cdh5(PAC) Cre ERT2 ; Tgfbr2 fl/fl instead of the Tie2 cre(+); Alk5 fl/fl There was a failure of complete ventricular septation and failure to fuse with the atrioventricular cushions in our Alk5 cKO and the Cdh5(PAC) Cre ERT2 ; Tgfbr2 fl/fl KOs. The Tie2 cre(+); Alk5 fl/fl embryos did not properly form a ventricular septum. Interestingly, the only organ system


97 appa rently affected in the Alk1 GFPCre ; Tgfbr2 fl/fl embryos is the CNS, as no obvious developmental cardiac defects were seen. Considering deleting Tgfbr2 with Cdh5(PAC) Cre ERT2 two days later (at E11.5) than Alk1 GFPCre (at E9.5) resulted in a cardiac phenotype e nforces the essential temporal role of endothelial TGFBR2 during development. CNS angiogenesis heavily relies on the migration and sprouting of ECs, particularly in EC tip cells providing directionality in response to various factors in the microvascular environment [173, 174, 196] Our data suggests that TGF signal activation within EC tip cells by the TGF 1/3 ligands are essential for EC migration of periventricular vessels. However, downstream targets or inter actions with other signaling pathways may be required for further sprouting and directionality. There are several murine KO models of other transcriptional factors (for example, ETS family member, Fli1) and members of other signaling family members (Wnt/ catenin) important in angiogenesis of the brain in which cerebral hemorrhaging has been observed which may provide further insight into the mechanism underlying neurovascular development [197, 198] The strongest c andidate of an interacting signaling pathway ( with the endothelial TGF signaling) involves the orphan G protein coupled receptor GPR124/tumor endothelial marker 5 (TEM5) [199] It was reported that GFR124 also appears to have spatial temporal role during neurovascular development, sp ecifically in the ECs of vessels within the forebrain. GFR124 KO mice, particularly in the ECs, resulted in the same cerebral phenotypes as our Tgfbr2 and Alk5 cKO models, with other organ systems unaffected. Furthermore, GPR124 requires Cdc42 Par6 [196]


98 a transcript ion factor important for filipodia formation and determining cell polarity, in angiogenic migration and directionality to remain within the forebrain environment [200] A more novel view of cerebral vessel developm ent states that the ECs of periventricular vessels migrate (beginning at E10) from the ventral to dorsal forebrain to establish an angiogenic gradient, in which specific transcription factors are expressed in different regions of the forebrain [201] Nkx2.1 and Dxl1/2 are expressed in the ventral region, while Pax6 is expressed in the dorsal region [201] Loss of either of these factors results in reduced periventricular vesse l migration, particularly in the dorsal forebrain. As Pax6 is a morphogenetic factor important in patterning throughout the body but also in axonal guidance and neuronal guidance [202] it is tempting to speculate t hat the begin to activate the TGF The migration of ECs is initiated by activation of TGF signaling in the tip cells, which regulate the ligand for GPR124 or inter acts with the GPR124 signaling pathway. Sprouting and directionality of EC migration is then controlled by GPR124 Cdc42 Par6. This model explains how the cerebral phenotype was specific to the GE of the forebrain (which is in the ventral region). ECs are a ble to sprout and perhaps migrate within the GE, however, the loss of any components leads to failure of migration directionality into the dorsal forebrain. The periventricular vessels aggregate within GE, the vessel walls weaken/rupture, and the perforati ons in the ventral brain form due to apoptosis. However, the mechanisms will need to be further evaluated, for example, by examining the expression of GPR124 or Cdc42 in our cKO models. [Summarized in Figure 4 7]


99 We show that endothelial TGF ia the receptors TGFBR2 and ALK5 is indispensable for establishment of the neurovasculature and for cardiac development during midgestational embryogenesis, and the vascular beds of the rest of the body are largely unaffected. The results also suggest that TGF temporal roles during development, which correspond with previously reported expression patterns of TGFBR2. This finding helps rectify the seemingly opposing phenotypes seen in various conditional knockout models of variou s TGF family members. It may also provide insight in maintenance of cerebral vessels in adult stages and help understand cerebral vessel defects, such as AVMs, that may appear.

PAGE 100

100 Figure 4 1 Synthesis and activation of the TGF The TG F prodomain signaling peptide, a latency associated peptide (LAP), and the TGF TGF Golgi appar atus, where furin enzymatically cleaves the signaling peptide. As the ligand prepares to exit the cell, the LAP encompasses the TGF and acts a chaperone. When secreted from the cell, the LAP still enclosed the TGF th the latent TGF (LTBP). This large complex is deposited in the extracellular matrix of a cell until it is activated by various means, such as pH change, heat, specific enzymes, or integrins. In these cases the TGF the LAP and allowed to bind and activate TGF 1, Thrombospondin 1.

PAGE 101

101 Figure 4 2 Alk1 GFPcre a novel EC specific cre deleter line. A) A GFPcre fusion construct was inserted into the endogenous Alk1 gene, replacing exons 4 8. B) Th e cre in this knockin line is active beginning at E9.5 of mouse embryogenesis in all ECs expressing Alk1 however, in a punctate fashion A gal staining covers the entire embryo, in a head confirms more robust X gal staining. D) A transverse section of the brain reveals strong endothelial Alk1 GFPcre is active in the ECs penetrating the neuroepithelium thus should be appropri ate to use in the study. SD/SA, SV40 splicing donor/acceptor signal; IRES, internal ribosomal entry sequence; pA, poly A signal.

PAGE 102

102 Figure 4 3 Tgfbr2 and Alk5 cKO embryos display specific cerebral hemorrhaging beginning at E11.5. Tgfbr2 (A) and Alk5 (B) mu tant embryos are indistinguishable from control littermates at E10.5. Beginning at E11.5, both Tgfbr2 and Alk5 cKO embryos begin exhibiting the same hemorrhaging in the forebrain region, shown by the white arrows. The cerebral hemorrhaging worsened over ti me in both cKO models (C and D). E) The head region of the Tgfbr2 cKO embryos becoming edemic but with the same hemorrhages seen; however, there were no hemorrhages seen in any other organ system besides the central nervous system. F) Alk5 cKO embryos die d by E13.5. In addition to CNS hemorrhages, the chest area of Alk5 cKO embryos were distended (white arrow head), indicating there may have been cardiac defects associated with the Alk5 cKOs. G) Tgfbr2 cKO embryos died a day later, at E14.5, with embryos p resenting severe brain hemorrhaging and very faintly beating heart, or already dead and in the process of resorption. H) and I) the forebrain bleeding was seen in both hemispheres of the brain. In some, not all, Tgfbr2 (J) and Alk5 (K) embryos, there were hemorrhages in the spinal cord beginning at E12.5.

PAGE 103

1 03 Figure 4 4 Histological Sections of E10.5 and E11.5 embryos. At E10.5, as compared brain vessels in the mutant mice, as see n in a representative section in (B) vessels that are closely associated with the surrounding cerebral tissue in the GE (closer view in D). E) Vessels in the tissue outside of the forebrain region. Conversely, Tgfbr2 (G) and Alk5 (K) cKO embryos begin displaying cavitations in the GE, specifically at E11.5. A closer view of PECAM 1 stained sections revealed that ECs clustered into glomeruloid like structures (H and L). However, it appears that only vessels within the neuroepithelium were affected as vessels lying outside of the GE were normal, as compared to the control, and blood cells were confined to the vessels. There does not seem to be any cardiac defects in the Tgfbr2 (J) or Alk5 (N) cKO embryos, as mutant hearts were comparable to control hearts (F). GE, ganglionic eminence; lu, lung; RA, right atrium; RV, right ventricle; LA, left atrium; LV, left ventricle.

PAGE 104

104 Figure 4 5 Endothelial specific deletion of Tgfbr2 during midge station specifically affected cerebral vessel formation at E13.5. Overall transverse view of the E13.5 control head (A) shows a properly developed head with various parts of the head labeled. A closer view of the GE, DC, and TC show EC layers of vessels (P ECAM 1 stained red) are closely associated with the surrounding tissue. However, the Tgfbr2 mutant head (B) displays hemorrhages in the forebrain region and the fourth ventricle and third ventricles are enlarged. A closer view of the GE shows large cavitat ions and disorganized, poorly formed vessels. The vessels within the DC are dilated and there are some cavitations. In the TE, vessels are dilated. Development of the lungs and heart of the Tgfbr2 cKO (F) did not appear to be affected and were comparable t o control controls (C). PECAM 1 and SMA (brown) staining of the Tgfbr2 confirms no developmental differences between control (E) and Tgfbr2 mutant (H) dorsal aortas. FV, fourth ventricle; TV, third ventricle; TE, telancephalon; DC, dicephalon; GE, ganglion eminence; lu, lung; RA, right atrium; RV, right ventricle; LA, left atrium; LV, left ventricle. DA, dorsa aorta.

PAGE 105

105 Figure 4 6 Alk5 cKO embryos exhibited abnormal cerebral vasculature in the GE and cardiac defects. A) H&E stain of a transverse view of the control head. Panels two (GE) and three (DC) show PECAM 1 stained ECs (red arrows) of the neuroepithelial vessels. These vessels are fine and closely associated with the surrounding tissue. B) H&E stain of Alk5 cKO head reveals the GE is perforated and there is free blood cells within the tissue. The ECs of the GE are still clumped in glomeruloid like structures, as in E11.5 brains, in the GE (B, panel one). In panel two, the vessels within the DC is similar to control heads. C) Control E12.5 heart. D) A magnified view of the SMA stained control dorsal aorta E) The Alk5 cKO heart is underdeveloped and the ventricular s eptum does not properly fuse in the atrioventricular cushion (*). F) The Alk5 cKO dorsal aorta is irregularly shaped and the smooth muscle layer appears diffuse. FV, fourth ventricle; TV, third ventricle; TE, telancephalon; DC, dicephalon; GE, ganglion em inence; lu, lung; RA, right atrium; RV, right ventricle; LA, left atrium; LV, left ventricle. DA, dorsa aorta.

PAGE 106

106 Figure 4 7 Proposed mechanism of cerebrovascular development in the telencephalon. A) GPR124 is active before E10 and involved in EC sprouting in the ventral telencephalon. Other factors, such as Nkx2.1 may facilitate in EC migration within this region B) At E10, the ECs begin to migrate from the ventral to dorsal telencephalon. This is initiated by the activation of the latent TGF 1/3 ligand by neural LAP of the ligand. C) A conformational change in the LAP causes it to release the ligand, which subsequently activates TGF likely tip cells), specifically thr ough the receptor pair TGFBR2 and ALK5. This leads to downstream regulation migration. Cdc42 Par6 is required for GPR124 migration directionality, however, it is unclear whether Cdc42 Par6 would associate before or after TGF of CNS angiogenesis suggests that periventricular vessels migrate from the ventral to dorsal telancephalon to establish an angiogenic gradient by which specific transcription factors are expressed (top image). Pax6 is an axon guidance factor, hence it seem s that that vessel would migrate to the dorsal region if driven or enticed by a neuronal cell. Thus, if this model is implicated in our proposed mechanism, the loss of TGFBR2 or ALK5 means there is no active migration in the EC tip cells. The consequence i s the sprouting ECs remain in the dorsal telancephalon, which includes the GE. The ECs accumulate and aggregate. The cells begin to under apoptosis and cause cavitations in the tissue (bottom panel). Dotted green lines denote EC migration path.

PAGE 107

107 CHAPTER 5 CONCLUSIONS AND PERSPECTIVES In the present study, the focus was to elucidate the role various TGF superfamily members play in ECs to determine pathogenetic etiology underlying the formation of vascular malformations, such as those exemplified in HHT. The focus was specifi cally on the in vivo roles of the Type III receptor Endoglin, Type II receptor TGFBR2, and Type I receptor ALK5. The functions of these receptors in vascular endothelial biology are still largely unknown and debated, as previous findings have been contradi ctory or unclear. In Chapter 3, the interaction between the two major genetic causes of the vascular disorder HHT were examined by generating two different Eng cKO models using the same two cre deleter lines (L1cre and R26 Cre ER ) previously used in Alk1 cK O models, then comparing the phenotypes. Then in Chapter 4, we sought to determine whether TGFBR2 and ALK5 have spatiotemporal roles during embryogenesis by selectively silencing each receptor in endothelial cells using a novel transgenic Alk1 cre knockin line ( Alk1 GFPCre ). Endothelial and G lobal Eng cKO Mice Suggest Divergent Pathogenetic Mechanisms U nderlying HHT1 and HHT2 It is assumed in the HHT field that ALK1 and ENG are intimately involved in ECs, likely transducing signals via the BMP signaling path way (Figure 1 1), thus it would be expected that deletion of either gene would lead to the HHT vascular malformations (e.g. AV shunting, AVMs, etc). The Oh lab had previously generated and characterized the first Alk1 cKO murine models that featured AVMs, as seen in HHT, at high, consistent frequencies. This was achieved using the endothelial specific cre deleter line, L1cre, and the tamoxifen inducible, globally expressed R26 Cre ER line. It was quite

PAGE 108

108 surprising that neither the Eng cKO mice nor the Eng iK O mice given the same treatment as the Alk1 iKO model were severely impacted by the loss of Eng Though it appeared that there may be a technical issue responsible (the efficiency at which the 2f allele is converted to the null 1f allele by different tamox ifen treatments) for the more varied phenotypes seen in the Eng iKO model, there were common phenotypes seen in similar organs in both models that prove these were true phenotypes. Many of the same organ systems were affected in the Eng and Alk1 ablation models but there were few shared patterns of pathology. Necroscopies of the Eng and Alk1 iKO model revealed local hemorrhaging, AV shunting, and AVMs in the intestine. a s econdary response to the hemorrhages, in cKO models of each receptor. Many more distinct differences were observed in the two models. Death occurred early in both endothelial specific and global deletion of Alk1 (by E17.5/PN5 or 3 weeks, respectively), whi le Eng cKO and Eng iKO are viable, with only a quarter of Eng iKO dying by 21 days. It was previously published that Alk1 iKO mice exhibited de novo AVMs in response to acute wounds given in the ear and dorsal skin. However, we found that the Eng iKO mice rarely do; instead, AVMs always appeared in tissue surrounding the ear ID tag, which can be considered a chronic injury. Also, AVMs formed specifically in the uterine vessels of the female reproductive organs of Alk1 iKO, while Eng iKO mice exhibited ovar ian AVMs. More surprising is that the pulmonary vasculature was not obviously affected in either Eng deletion model even though the lungs presented phenotypes at high frequencies in Alk1 deletion models. These distinct phenotypes

PAGE 109

109 bring unanticipated eviden ce that there are separate pathogenetic mechanisms underlying HHT1 and HHT2. This seems plausible considering the expression pattern of ENG versus ALK1. ALK1 is more restricted to arterial ECs, while ENG is more pan endothelial. ENG expression has also bee n found in other cell types, such as the smooth muscle layer (in low levels) and several types of circulating cells, such as monocytes, HSCs, and stromal cells. Additionally, it was reported in the pulmonary vasculature that ENG and ALK1 expression overlap s only in the distal vessels. In the Eng cKO mice, as Eng is silenced in arterial ECs, venous Eng is still present. The lack of phenotype in this model could be explained in that circulating cells expressing Eng could be compensating for the loss of vascul ar Eng Future studies examining the bone marrow or circulating endothelial cells (CECs) must be performed to confirm this possibility. As for the Eng iKO model, quantification of the genomic copy of the 1f allele suggests that the dosage and number of ta moxifen affects how much Eng is deleted. In contrast to the endothelial specific cKO models or the Alk1 iKO models, it is less predictable which vessel and cell type Eng is actually silenced in the Eng iKO mice. This can account for the variable phenotype. For example, for the mice in which there was overlap in the Alk1 and Eng iKO mice, it could mean that more arterial ENG was silenced. More comprehensive expression studies in the Eng iKO model should b e conducted either by isolating ECs and CECs and quantifying expression of Eng or histological analysis by immunofluorescence to determine whether one vessel type or cell type (other than ECs) may be more impacted.

PAGE 110

110 Endothelial TGF S ignaling ha s S patiotemporal R oles in C erebro vascular D evelopment The understanding of the in vivo mechanisms of TGF angiogenesis has been complicated by contradictory phenotypes in conventional and conditional knockout mouse models. Because it was believed that a balance between TGFBR2 ALK1 and TGFBR2 ALK5 was essential in maintaining vascular homeostasis, ALK1 involvement with TGFBR2 and ALK5 has been of interest in the HHT field. In Chapter 4 we tested the hyp othesis that endothelial TGF temporal roles in vascular development, focusing on the tightly regulated mechanism of cerebrovascular development. Using a novel cre deleter line ( Alk1 GFPCre ) to silence Alk5 and Tgfbr2 in ECs at E9 .5, we provide in vivo evidence that responsible for activating latent ligands TGF ALK5 within ECs. Alk1 GFPCre ; Tgfbr2 fl/fl and Alk1 GFPCre ; Alk5 fl/fl displayed the similar specific cerebra l hemorrhaging in the forebrain region particularly in the GE, as seen in GPR124 / itgav / itg b 8 / and Tgfb1 RDE/RDE ;Tgfb3 / mice. The only organ system affected in the Alk1 GFPCre ; Tgfbr2 fl/fl mice was the CNS, particularly the cerebral vessels. Thi s differs slightly from the Cdh5(PAC) Cre ERT2 ; Tgfbr2 fl/fl by which Tgfbr2 was deleted 48 hrs later, but developed ventricular septal defects. This delay seemingly accentuates the temporal aspect of TGF signaling. The CNS manifestation was less severe in Alk1 GFPCre ; Alk5 fl/fl embryos However, mutants also had cardiac development defects and died sooner. A potential mechanism is that d uring midgestation EC sprouting may to regulate migration. Because TGF the Cdc42 to direct migration towards the dorsal region of the telencephalon It is

PAGE 111

111 probable that the loss of ALK5 and TGFBR2 lead to a defect in the activation of EC migration. Thus, EC sprouting is taking place but the lack of proper migration out of the GE occur results in the glomeruloid like vasc ular aggregates. It would be interesting to see whether this mechanism would be important in maintenance of the cerebral vasculature in adult tissue and whether it may contribute to BAVMs. It should be noted that the study confirms that ALK5 is in fact exp ressed and essential in ECs, in contrast to expression studies stating it is expressed only in SMCs. The earlier results can be clarified in that ALK5 is typically expressed at such low levels in ECs that it was not gal stainin g. Perspectives The use of the murine models has the advantage over in vitro models in that the signal transduction can be examined in a physiological mammalian model that is similar to humans, without the bias of conditional or artificially supplemented m edium. A drawback to our models is that the mice were mixed genetic background (containing 129Sv/J, C57BL/6, and FVB), which may influence the frequency by which a phenotype may be seen due to modifier genes. This was an issue that was addressed in the cur rently used Eng +/ HHT1 mouse model, by which the C57BL/6 backbred line is now often used. Backcrossing the Eng cKO mice to either a pure C57BL/6 or 129Sv/J background may resolve the problem. As the three known genes associated with HHT as part of the TGF pathway, it is strongly believed that ALK1, ENG, and SMAD4 signal in a linear pathway. However, the findings here that the Eng cKOs, behaved distinctly from Alk1 cKO mice, in addition to findings that L1cre(+); Smad4 fl/fl mice were also viable ( personal observations), have proven the HHT mechanism may be much more complicated than

PAGE 112

112 anticipated. To rectify potentially independent signaling pathways indicated from the findings in the three ( Alk1, Eng, Smad4 ) EC cKO models, a double L1cre mediated en dothelial specific knockout mouse ( L1cre(+); Smad4 fl/fl ; Eng fl/fl ) may need to be generated to determine how essential ENG and SMAD4 truly are for ALK1 signaling. As mentioned earlier in this chapter, the mild phenotype in Eng cKO mice may be due to the wide r expression of ENG in vascular endothelial cells in addition to compensation of the loss of arterial ENG with ENG expressing CECs. The same phenomena may be relevant for SMAD4, although SMAD4 is expressed on other tissue and cells types, not only ECs. It would be expected that if ENG and SMAD4 were essential for ALK1, the double KO would likely recapitulate the L1cre(+); Alk1 2f/2f mice. Thus, settle whether there are differing mechanisms underlying HHT. The goal of generating the HHT murine models and u nderstanding the physiological and pathological signal transduction in the endothelium is to ultimately create and test therapeutic treatments for patients. Although, the current models provided some interesting insight into the in vivo mechanisms of the d ifferent TGF receptors in ECs and potential mechanisms in AVM formation, the Eng cKO studies seem to raise even more questions. For example, do the disparate mechanisms of HHT1 and HHT2 (or even HHT3, HHT4, and JP HHT when more information arises) conclu de that the various types of HHT must be approached or treated separately in the clinical setting? One way to decipher this is to perform a microarray analysis of different tissues taken from each Eng cKO model and compare the gene expression profile deriv ed from the Alk1 cKO models, such as those we previously accu mulated from lung samples. As the finding from Chapter 4 suggests it may be essential to evaluate each

PAGE 113

113 organ separately as there are time and organ specific expression of different TGF signaling members during angiogenesis. Additionally, there are poten tial organ specific molecules, such as the orphan GPR124 that should be given more attention because they may contribute to vascular malformations. The differences in response to chronic and acute injuries between the Alk1 iKO and Eng iKO were an interest ing finding. As Eng iKO mice constantly developed AVMs because of the ear ID tag, it may be possible that AVMs in the Eng deficient vessels may be a response to the recruitment of inflammatory cytokines, such as tumor necrosis factor inflammation in Eng +/ for example by dextran sulfate, lead to aber rant angiogenesis. A future study in the Eng iKO mice would be to infuse a nanoparticle with TNF embed the infused nanoparticle into the dorsal skin of the mouse. The potential formation of AVMs to the nanoparticle would be observed. In conclusion, we generated several cKO models that provided interesting in vivo insight into TGF Eng iKO, the appearance of vascular malformations was consistent, making it a valuable tool in studying HHT. Additionally, the T gfbr 2 and A lk 5 cKO models indicated TGF signaling have essential spatial and temporal roles in angiogenesis. Finally, a novel transgenic Alk1 GFPCre was introduced and has proven to be an effective and useful line to study and evaluate the specific functions of angiogenic molecules at midgestation. Overall, our mouse models produced more consistent phenotypes than currently available mouse models and will be vital to future studies in HHT.

PAGE 114

114 Table 5 1. Comparison of Eng vs Alk1 cKO mous e models cKO Genotype Lethality Time of visible phenotype Affected organs L1cre(+); Eng 2f/2f Survived Aged 15 months GI tract Liver 1) L1cre(+); Alk1 2f/2f 2) L1cre(+); Alk1 3 f/ 3 f 1) PN5 2) E17.5 E15.5 Lung GI tract Liver Brain R26 Cre ER(+) ; Eng 2f/2f +TM (5 treatments): 1) 2.5 mg/25 g BW x 1 2) 2.5 mg/25 g BW x 2 3) 2.5 mg/25 g BW x 3 4) 2.5 mg/40 g BW x 2 5) 2.5 mg/40 g BW x 3 [Time after TM given] 1) Survived 2) 10 days 3) 10 days 4) at 3 wks, variable 5) at 3 wks, variable 1) 2 months 2) 4 days 3) 4 day s 4) 4 days 5) 4 days Ear (ID tag) GI tract : ovarian vessels Liver 1) R26 Cre ER(+) ; Alk1 2f/2f 2) R26 Cre ER(+) ; Alk1 3 f/ 3 f + 2.5 mg TM/25 g BW x1 By 21 days 7 days Lungs GI tract : uterine vessels Liver ( Alk1 3 f/ 3 f iKO ) Acute wound Ear Dorsal skin Table 5 2. Summary of Alk5 and Tgfbr2 cK O models cKO Genotype Lethality Time of visible phenotype Affected organs Alk1 GFPCre ; Tgfbr2 f l/ f l E14.5 E11.5 Brain (edemic), Ganglionic eminence Diencephalon Telencephalon Spine (variable) Alk1 GFPCre ; Alk5 f l/ f l E13.5 E11 Forebrain, Gan glionic eminence Spine (variable) Heart

PAGE 115

115 LIST OF REFERENCES 1. A ir d W C. Phenotypic heterogeneity of the endothelium: I. Structure, function, and mechanisms. Circ Res. 2007 ; 100 (2):158 73. 2. Jain RK. Molecular regulation of vessel ma turation Nat Med 2003; 9 (6):685 93. 3. Rocha, SF, Adams R H. Molecular differentiation and specialization of vascular beds. Angiogenesis. 2009 ; 12 (2):139 47. 4. Flamme I Frolich T Risau W. Molecular mechanisms of vasculogenesis and embryonic angiogenesis. J Cell Physiol 1997 ; 173 (2):206 10. 5. Risau W, Flamme I. Vasculogenesis Annu Rev Cell Dev Biol 1995; 11 :73 91. 6. Hirashima M Suda T. Differentiation of arterial and venous endothelial cells and vascular morphogenesis Endothelium 2006 ; 13 (2):137 45. 7 LaRue AC Lansford R Drake C.J. Circulating blood island derived cells contribute to vasculogenesis in the embryo proper Dev Biol 2003 ; 262 (1):162 72. 8. Kovacic JC, Moore J, Herbert A, Ma D, Boehm M, Graham RM Endothelial progenitor cells, angioblast s, and angiogenesis -old terms reconsidered from a current perspective Trends Cardiovasc Med 2008 ; 18 (2):45 51. 9. Risau W. Mechanisms of angiogenesis Nature 1997 ; 386 (6626):671 4. 10. Drake CJ. Embryonic and adult vasculogenesis Birth Defects Res C Emb ryo Today 2003 ; 69 (1):73 82. 11. Baldwin HS. Early embryonic vascular development Cardiovasc Res 1996 ; 31 Spec No :E34 45. 12. Aird WC. Endothelium in health and disease Pharmacol Rep 2008 ; 60 (1):139 43. 13. Adams RH, Alitalo K. Molecular regulation of an giogenesis and lymphangiogenesis. Nat Rev Mol Cell Biol 2007; 8 (6):464 78. 14. Adams RH. Molecular control of arterial venous blood vessel identity J Anat 2003 ; 202 (1):105 12. 15. Buschmann I, Schaper W. Arteriogenesis Versus Angiogenesis: Two Mechanisms of Vessel Growth News Physiol Sci. 1999 ; 14 :121 125. 16. Cai W, Schaper W. Mechanisms of arteriogenesis Acta Biochim Biophys Sin (Shanghai) 2008 ; 40 (8):681 92.

PAGE 116

116 17. Boon LM, Ballieux F, Vikkula M. Pathogenesis of vascular anomalies Clin Plast Sur g 2011; 3 8 (1):7 19. 18. Brouillard P, Vikkula M. Vascular malformations: localized defects in vascular morphogenesis. Clin Genet 2003 ; 63 (5):340 51. 19. Brouillard P Vikkula M. Genetic causes of vascular malformations Hum Mol Genet 200 7 ; 16 Spec No. 2 :R14 0 9. 20. Marchu k DA, Srinivasan S, Squire TL, Zawistowski JS Vascular morphogenesis: tales of two syndromes Hum Mol Genet 2003 ; 12 Spec No 1 :R97 112. 21. Cleaver O Melton D A. Endothelial signaling during development Nat Med 2003 ; 9 (6):661 8. 22. Adams RH Wilk inson GA, Weiss C, Diella F, Gale N W, Deutsch U, Risau W, Klein R Roles of ephrinB ligands and EphB receptors in cardiovascular development: demarcation of arterial/venous domains, vascular morphogenesis, and sprouting angiogenesis Genes Dev 1999 ; 13 (3): 295 306. 23. Aird W C. Phenotypic heterogeneity of the endothelium: II. Representative vascular beds Circ Res 2007 ; 100 (2):174 90. 24. Pardali E Goumans MJ ten Dijke P. Signaling by members of the TGF beta family in vascular morphogenesis and disease Tr ends Cell Biol 2010; 2 0 (9):556 67. 25. Breier G, Risau W. T he role of vascular endothelial growth factor in blood vessel formation. Trends Cell Biol 1996; 6 (12):454 6. 26. Sato T N, Tozawa Y, Deutsch U, Wolburg Buchholz K, Fujiwara Y, Gendron Maguire M, Gri dley T, Wolburg H, Risau W, Qin Y Distinct roles of the receptor tyrosine kinases Tie 1 and Tie 2 in blood vessel formation Nature 1995; 376 (6535):70 4. 27. Augustin HG Koh GY, Thurston G, Alitalo K Control of vascular morphogenesis and homeostasis thr ough the angiopoietin Tie system Nat Rev Mol Cell Biol 2009; 10 (3):165 77. 28. Carmelie t P, Ferreira V, Breier G, Pollefeyt S, Kieckens L, Gertsenstein M, Fahrig M, Vandenhoeck A, Harpal K, Eberhardt C, Declercq C, Pawling J, Moons L, Collen D, Risau W, N agy A Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele Nature 1996; 380 (6573):435 9. 29. Hofmann JJ, Iruela Arispe M L. Notch signaling in blood vessels: who is talking to whom about what? Circ Res 2007; 100 (11):1556 68.

PAGE 117

117 30. Rossant J, Howard L. Signaling pathways in vascular development Annu Rev Cell Dev Biol 2002 ; 18 :541 73. 31. Moustakas A, Heldin C H. The regulation of TGFbeta signal transduction Development 2009 ; 136 (22):3699 714. 32. Zhu HJ, Burgess A W. Regulat ion of transforming growth factor beta signaling Mol Cell Biol Res Commun 2001; 4 (6):321 30. 33. Shi Y, Massague J. Mechanisms of TGF beta signaling from cell membrane to the nucleus Cell 2003 ; 113 (6):685 700. 34. Bobik A. Transforming growth factor beta s and vascular disorders Arterioscler Thromb Vasc Biol 2006 ; 26 (8):1712 20. 35. Derynck R, Feng X H. TGF beta receptor signaling Biochim Biophys Acta 1997 ; 1333 (2):F105 50. 36. ten Dijke P, Hill C S. New insights into TGF beta Smad signalling Trends Bioch em Sci 2004 ; 29 (5):265 73. 37. M assague J. The TGF beta family of growth and differentiation factors Cell 1987 ; 49 (4):437 8. 38 Wrana JL Attisano L, Crcamo J, Zentella A, Doody J, Laiho M, Wang XF, Massagu J TGF beta signals through a heteromeric prot ein kinase receptor complex. Cell 1992 ; 71 (6):1003 14. 39. Persson U Izumi H, Souchelnytskyi S, Itoh S, Grimsby S, Engstrm U, Heldin CH, Funa K, ten Dijke P The L45 loop in type I receptors for TGF beta family members is a critical determinant in specif ying Smad isoform activation. FEBS Lett 1998 ; 434 (1 2):83 7. 40. Walshe TE. TGF beta and microvessel homeostasis. Microvasc Re s 2010; 8 0 (1):166 73. 41. Gatza CE, Oh SY Blobe GC. Roles for the type III TGF beta receptor in human cancer. Cell Sign al. 2010; 2 2 ( 8):1163 74. 42. Bernabeu C, Lopez Novoa JM Quintanilla M. The emerging role of TGF beta superfamily coreceptors in cancer. Biochim Biophys Acta 2009 ; 1792 (10):954 73. 43. Jovine L, Qi H, Williams Z, Litscher E, Wassarman PM The ZP domain is a conserved module for polymerization of extracellular proteins. Nat Cell Biol 2002 ; 4 (6):457 61.

PAGE 118

118 44. Barbara NP, Wrana JL Letarte M. Endoglin is an accessory protein that interacts with the signaling receptor complex of multiple members of the transforming growth f actor beta superfamily. J Biol Chem 1999 ; 274 (2):584 94. 45. Lin H, Wang H Zhu C. The Smad pathway in transforming growth factor beta signaling. Sci China C Life Sci 2003 ; 46 (5):449 63. 46. Feng XH Derynck R. Specificity and versatility in tgf beta signa ling through Smads. Annu Rev Cell Dev Biol 2005 ; 21 :659 93. 47. Moustakas A Heldin CH. From mono to oligo Smads: the heart of the matter in TGF beta signal transduction. Genes Dev 2002 ; 16 (15):1867 71. 48. Itoh S ten Dijke P. Negative regulation of TGF beta receptor/Smad signal transduction. Curr Opin Cell Biol 2007 ; 19 (2):176 84. 49. Zhang YE Non Smad pathways in TGF beta signaling. Cell Res 2009 ; 19 (1):128 39. 50. Lowery JW de Caestecker MP BMP signaling in vascular development and disease. Cytokine Growth Factor R ev 2010; 21 ( 4):287 98. 51. Judge DP Dietz HC Marfan's syndrome. Lancet 2005 ; 366 (9501):1965 76. 52. Loeys BL, Chen J, Neptune ER, Judge DP, Podowski M, Holm T, Meyers J, Leitch CC, Katsanis N, Sharifi N, Xu FL, Myers LA, Spevak PJ, Camero n DE, De Backer J, Hellemans J, Chen Y, Davis EC, Webb CL, Kress W, Coucke P, Rifkin DB, De Paepe AM, Dietz HC A syndrome of altered cardiovascular, craniofacial, neurocognitive and skeletal development caused by mutations in TGFBR1 or TGFBR2. Nat Genet 2005 ; 37 (3):275 81. 53. Dickson MC, Martin JS, Cousins FM, Kulkarni AB, Karlsson S, Akhurst RJ Defective haematopoiesis and vasculogenesis in transforming growth factor beta 1 knock out mice. Development 1995 ; 121 (6):1845 54. 54. Martin JS, Dickson MC, Cou sins FM, Kulkarni AB, Karlsson S, Akhurst RJ Analysis of homozygous TGF beta 1 null mouse embryos demonstrates defects in yolk sac vasculogenesis and hematopoiesis. Ann N Y Acad Sci 1995 ; 752 :300 8. 55. Larsson J, Goumans MJ, Sjstrand LJ, van Rooijen MA, Ward D, Leven P, Xu X, ten Dijke P, Mummery CL, Karlsson S A bnormal angiogenesis but intact hematopoietic potential in TGF beta type I receptor deficient mice. EMBO J 2001 ; 20 (7):1663 73. 56. Li DY Sorensen LK, Brooke BS, Urness LD, Davis EC, Taylor DG Boak BB, Wendel DP Defective angiogenesis in mice lacking endoglin. Science 1999 ; 284 (5419):1534 7.

PAGE 119

119 57. Bourdeau A, Dumont DJ Letarte M. A murine model of hereditary hemorrhagic telangiectasia. J Clin Invest 1999 ; 104 (10):1343 51. 58. Oshima M, Oshima H Taketo MM. TGF beta receptor type II deficiency results in defects of yolk sac hematopoiesis and vasculogenesis. Dev Biol 1996 ; 179 (1):297 302. 59. Chang H Huylebroeck D, Verschueren K, Guo Q, Matzuk MM, Zwijsen A Smad5 knockout mice die at mid gestat ion due to multiple embryonic and extraembryonic defects. Development 1999 ; 126 (8):1631 42. 60. Goumans MJ, Valdimarsdottir G, Itoh S, Rosendahl A, Sideras P, ten Dijke P. Balancing the activation state of the endothelium via two distinct TGF beta type I r eceptors. EMBO J 2002 ; 21 (7):1743 53. 61. Goumans MJ, Valdimarsdottir G, Itoh S, Lebrin F, Larsson J, Mummery C, Karlsson S, ten Dijke P. Activin receptor like kinase (ALK)1 is an antagonistic mediator of lateral TGFbeta/ALK5 signaling. Mol Cell 2003 ; 12 (4 ):817 28. 62. Lamouille S, Mallet C, Feige JJ, Bailly S. Activin receptor like kinase 1 is implicated in the maturation phase of angiogenesis. Blood 2002 ; 100 (13):4495 501. 63. Seki T, Hong KH Oh SP. Nonoverlapping expression patterns of ALK1 and ALK5 rev eal distinct roles of each receptor in vascular development. Lab Invest 2006 ; 86 (2):116 29. 64. Park SO Lee YJ, Seki T, Hong KH, Fliess N, Jiang Z, Park A, Wu X, Kaartinen V, Roman BL, Oh SP ALK5 and TGFBR2 independent role of ALK1 in the pathogenesis o f hereditary hemorrhagic telangiectasia type 2. Blood 2008 ; 111 (2):633 42. 65. Scharpfenecker M, van Dinther M, Liu Z, van Bezooijen RL, Zhao Q, Pukac L, Lwik CW, ten Dijke P. BMP 9 signals via ALK1 and inhibits bFGF induced endothelial cell proliferation and VEGF stimulated angiogenesis. J Cell Sci 2007 ; 120 (Pt 6):964 72. 66. David L, Mallet C, Mazerbourg S, Feige JJ, Bailly S. Identification of BMP9 and BMP10 as functional activators of the orphan activin receptor like kinase 1 (ALK1) in endothelial cell s. Blood 2007 ; 109 (5):1953 61. 67. Luo J, Tang M, Huang J, He BC, Gao JL, Chen L, Zuo GW, Zhang W, Luo Q, Shi Q, Zhang BQ, Bi Y, Luo X, Jiang W, Su Y, Shen J, Kim SH, Huang E, Gao Y, Zhou JZ, Yang K, Luu HH, Pan X, Haydon RC, Deng ZL, He TC. TGFbeta/BMP ty pe I receptors ALK1 and ALK2 are essential for BMP9 induced osteogenic signaling in mesenchymal stem cells. J Biol Che m 2010; 285 (38):29588 98.

PAGE 120

120 68. Daina E, D'Ovidio F Sabba C. Introduction: hereditary hemorrhagic telangiectasia as a rare disease. Curr Ph arm Des 2006 ; 12 (10):1171 2. 69. Sabba C Pasculli G, Cirulli A, Gallitelli M, Virgilio G, Resta F, Guastamacchia E, Palasciano G Hereditary hemorrhagic teleangiectasia (Rendu Osler Weber disease). Minerva Cardioangiol 2002 ; 50 (3):221 38. 70. Geisthoff UW Hille K, Ruprecht KW, Verse T, Plinkert PK. Prevalence of ocular manifestations in hereditary hemorrhagic telangiectasia. Graefes Arch Clin Exp Ophthalmol 2007 ; 245 (8):1141 4. 71. Lenato GM Guanti G. Hereditary Haemorrhagic Telangiectasia (HHT): genetic and molecular aspects. Curr Pharm Des 2006 ; 12 (10):1173 93. 72. Shovlin CL Hereditary haemorrhagic telangiectasia: pathophysiology, diagnosis and treatment. Blood R ev 2010; 24 ( 6):203 19. 73. McAllister KA, Grogg KM, Johnson DW, Gallione CJ, Baldwin MA, J ackson CE, Helmbold EA, Markel DS, McKinnon WC, Murrell J, et al. Endoglin, a TGF beta binding protein of endothelial cells, is the gene for hereditary haemorrhagic telangiectasia type 1. Nat Genet 1994 ; 8 (4):345 51. 74. Shovlin CL, Hughes JM, Tuddenham EG Temperley I, Perembelon YF, Scott J, Seidman CE, Seidman JG. A gene for hereditary haemorrhagic telangiectasia maps to chromosome 9q3. Nat Genet 1994 ; 6 (2):205 9. 75. Gallione CJ, Repetto GM, Legius E, Rustgi AK, Schelley SL, Tejpar S, Mitchell G, Drouin E, Westermann CJ, Marchuk DA. A combined syndrome of juvenile polyposis and hereditary haemorrhagic telangiectasia associated with mutations in MADH4 (SMAD4). Lancet 2004 ; 363 (9412):852 9. 76. Gallione C, Aylsworth AS, Beis J, Berk T, Bernhardt B, Clark R D, Clericuzio C, Danesino C, Drautz J, Fahl J, Fan Z, Faughnan ME, Ganguly A, Garvie J, Henderson K, Kini U, Leedom T, Ludman M, Lux A, Maisenbacher M, Mazzucco S, Olivieri C, Ploos van Amstel JK, Prigoda Lee N, Pyeritz RE, Reardon W, Vandezande K, Waldman JD, White RI Jr, Williams CA, Marchuk DA Overlapping spectra of SMAD4 mutations in juvenile polyposis (JP) and JP HHT syndrome. Am J Med Genet A 2010 Feb;152A(2):333 9. 77. Cole SG, Begbie ME, Wallace GM, Shovlin CL. A new locus for hereditary haemorrha gic telangiectasia (HHT3) maps to chromosome 5. J Med Genet 2005 ; 42 (7):577 82. 78. Bayrak Toydemir P, McDonald J, Akarsu N, Toydemir RM, Calderon F, Tuncali T, Tang W, Miller F, Mao R A fourth locus for hereditary hemorrhagic telangiectasia maps to chrom osome 7. Am J Med Genet A 2006 ; 140 (20):2155 62.

PAGE 121

121 79. Govani FS Shovlin CL. Fine mapping of the hereditary haemorrhagic telangiectasia (HHT)3 locus on chromosome 5 excludes VE Cadherin 2, Sprouty4 and other interval genes. J Angiogenes R es 2010; 2 : 15. 80. Pece Barbara N, Cymerman U, Vera S, Marchuk DA, Letarte M Expression analysis of four endoglin missense mutations suggests that haploinsufficiency is the predominant mechanism for hereditary hemorrhagic telangiectasia type 1. Hum Mol Genet 1999 ; 8 (12):217 1 81. 81. Pece N Vera S, Cymerman U, White RI Jr, Wrana JL, Let arte M Mutant endoglin in hereditary hemorrhagic telangiectasia type 1 is transiently expressed intracellularly and is not a dominant negative. J Clin Invest 1997 ; 100 (10):2568 79. 82. Hoag J B, Terry P, Mitchell S, Reh D, Merlo CA. An epistaxis severity score for hereditary hemorrhagic telangiectasia. Laryngosc ope 2010; 120 (4):838 43. 83. Braverman IM, Keh A Jacobson BS. Ultrastructure and three dimensional organization of the telangiectases of hereditary hemorrhagic telangiectasia. J Invest Dermatol 1990 ; 95 (4):422 7. 84. Pasculli G Resta F, Guastamacchia E, Di Gennaro L, Suppressa P, Sabb C Health related quality of life in a rare disease: hereditary hemorrhagic telangiectasia (HHT) or Re ndu Osler Weber disease. Qual Life Res 2004 ; 13 (10):1715 23. 85. Jacobson BS Hereditary hemorrhagic telangiectasia: A model for blood vessel growth and enlargement. Am J Pathol 2000 ; 156 (3):737 42. 86. Faughnan ME, Palda VA, Garcia Tsao G, Geisthoff UW, M cDonald J, Proctor DD, Spears J, Brown DH, Buscarini E, Chesnutt MS, Cottin V, et al. International guidelines for the diagnosis and management of hereditary haemorrhagic telangiectasia. J Med G enet 2 011;48(2):73 87. Epub 2009. 87. Brady AP, Murphy MM O' Connor TM Hereditary haemorrhagic telangiectasia: a cause of preventable morbidity and mortality. Ir J Med Sci 2009 ; 178 (2):135 46. 88. Sabba C Pompili M. Review article: the hepatic manifestations of hereditary haemorrhagic telangiectasia. Aliment Pharm acol Ther 2008 ; 28 (5):523 33. 89. Dupuis Girod S, Chesnais AL, Ginon I, Dumortier J, Saurin JC, Finet G, Decullier E, Marion D, Plauchu H, Boillot O. Long term outcome of patients with hereditary hemorrhagic telangiectasia and severe hepatic involvement af ter orthotopic liver transplantation: a single center study. Liver Tran spl 2010; 16 (3 ):340 7. 90. Abdalla SA Gallione CJ, Barst RJ, Horn EM, Knowles JA, Marchuk DA, Letarte M, Morse JH Primary pulmonary hypertension in families with hereditary haemorrhag ic telangiectasia. Eur Respir J 2004 ; 23 (3):373 7.

PAGE 122

122 91. Stringham R Shah NR. Pulmonary arterial hypertension: an update on diagnosis and treatment. Am Fam Physic ian 2010; 82 (4):370 7. 92. Girerd B Montani D, Coulet F, Sztrymf B, Yaici A, Jas X, Tregouet D, Reis A, Drouin Garraud V, Fraisse A, Sitbon O, O'Callaghan DS, Simonneau G, Soubrier F, Humbert M. Clinical outcomes of pulmonary arterial hypertension in patients carrying an ACVRL1 (ALK1) mutation. Am J Respir Crit Care Me d 2010; 181 (8):851 61. 93. Ch en S T Karnezis T Davidson TM Safety of intranasal Bevacizumab (avastin) treatment in patients with hereditary hemorrhagic telangiectasia associated epistaxis. Laryngosco pe 2010 Nov 11. [Epub ahead of print] 94. Davidson TM, Olitsky SE Wei JL. Heredita ry hemorrhagic telangiectasia/avastin. Laryngosc ope 2010; 120 (2):432 5. 95. Lebrin F, Srun S, Raymond K, Martin S, van den Brink S, Freitas C, Brant C, Mathivet T, Larrive B, Thomas JL, Arthur HM, Westermann CJ, Disch F, Mager JJ, Snijder RJ, Eichmann A, Mummery CL. Thalidomide stimulates vessel maturation and reduces epistaxis in individuals with hereditary hemorrhagic telangiectasia. Nat M ed. 2010; 16 ( 4):420 8. 96. Fernandez L A, Garrido Martin EM, Sanz Rodriguez F, Ramirez JR, Morales Angulo C, Zarrabei tia R, Perez Molino A, Bernabu C, Botella LM. Therapeutic action of tranexamic acid in hereditary haemorrhagic telangiectasia (HHT): regulation of ALK 1/endoglin pathway in endothelial cells. Thromb Haemost 2007 ; 97 (2):254 62. 97. Jameson JJ Cave DR. Hor monal and antihormonal therapy for epistaxis in hereditary hemorrhagic telangiectasia. Laryngoscope 2004 ; 114 (4):705 9. 98. Gluckman JL Portugal LG Modified Young's procedure for refractory epistaxis due to hereditary hemorrhagic telangiectasia. Laryngos cope 1994 ; 104 (9):1174 7. 99. Illum P Bjerring P. Hereditary hemorrhagic telangiectasia treated by laser surgery. Rhinology 1988 ; 26 (1):19 24. 100. Geisthoff UW, Fiorella ML Fiorella R Treatment of recurrent epistaxis in HHT. Curr Pharm Des 2006 ; 12 (10) :1237 42. 101. Velasco B, Ramrez JR, Relloso M, Li C, Kumar S, Lopez Bote JP, Prez Barriocanal F, Lpez Novoa JM, Cowan PJ, d'Apice AJ, Bernabu C. Vascular gene transfer driven by endoglin and ICAM 2 endothelial specific promoters. Gene Ther 2001 ; 8 (12) :897 904.

PAGE 123

123 102. Fernandez L A, Sanz Rodriguez F, Zarrabeitia R, Prez Molino A, Hebbel RP, Nguyen J, Bernabu C, Botella LM. Blood outgrowth endothelial cells from Hereditary Haemorrhagic Telangiectasia patients reveal abnormalities compatible with vascula r lesions. Cardiovasc Res 2005 ; 68 (2):235 48. 103. Allinson KR, Carvalho RL, van den Brink S, Mummery CL, Arthur HM. Generation of a floxed allele of the mouse Endoglin gene. Genesis 2007 ; 45 (6):391 5. 104. Chytil A Magnuson MA, Wright CV, Moses HL. Condi tional inactivation of the TGF beta type II receptor using Cre:Lox. Genesis 2002 ; 32 (2):73 5. 105. Johnson DW, Berg JN, Baldwin MA, Gallione CJ, Marondel I, Yoon SJ, Stenzel TT, Speer M, Pericak Vance MA, Diamond A, Guttmacher AE, Jackson CE, Attisano L, K ucherlapati R, Porteous ME, Marchuk DA. Mutations in the activin receptor like kinase 1 gene in hereditary haemorrhagic telangiectasia type 2. Nat Genet 1996 ; 13 (2):189 95. 106. Lesca G, Genin E, Blachier C, Olivieri C, Coulet F, Brunet G, Dupuis Girod S, Buscarini E, Soubrier F, Calender A, Danesino C, Giraud S, Plauchu H; French Italian HHT Network. Hereditary hemorrhagic telangiectasia: evidence for regional founder effects of ACVRL1 mutations in French and Italian patients. Eur J Hum Genet 2008 ; 16 (6):7 42 9. 107. Olivieri C, Pagella F, Semino L, Lanzarini L, Valacca C, Pilotto A, Corno S, Scappaticci S, Manfredi G, Buscarini E, Danesino C. Analysis of ENG and ACVRL1 genes in 137 HHT Italian families identifies 76 different mutations (24 novel). Comparis on with other European studies. J Hum Genet 2007 ; 52 (10):820 9. 108. Sadick H, Hage J, Goessler U, Stern Straeter J, Riedel F, Hoermann K, Bugert P. Mutation analysis of "Endoglin" and "Activin receptor like kinase" genes in German patients with hereditary hemorrhagic telangiectasia and the value of rapid genotyping using an allele specific PCR technique. BMC Med Genet 2009 ; 10 :53. 109. Fernandez LA, Sanz Rodriguez F, Zarrabeitia R, Perez Molino A, Morales C, Restrepo CM, Ramirez JR, Coto E, Lenato GM, Bern abeu C, Botella LM. Mutation study of Spanish patients with hereditary hemorrhagic telangiectasia and expression analysis of Endoglin and ALK1. Hum Mutat 2006 ; 27 (3):295. 110. Bayrak Toydemir P, McDonald J, Markewitz B, Lewin S, Miller F, Chou LS, Gedge F, Tang W, Coon H, Mao R. Genotype phenotype correlation in hereditary hemorrhagic telangiectasia: mutations and manifestations. Am J Med Genet A 2006 ; 140 (5):463 70.

PAGE 124

124 111. Thomas B, Eyries M, Montagne K, Martin S, Agrapart M, Simerman Franois R, Letarte M, Soubrier F. Altered endothelial gene expression associated with hereditary haemorrhagic telangiectasia. Eur J Clin Invest 2007 ; 37 (7):580 8. 112. Seki T, Yun J Oh SP Arterial endothelium specific activin receptor like kinase 1 expression suggests its rol e in arterialization and vascular remodeling. Circ Res 2003 ; 93 (7):682 9. 113. Fernandez Ruiz E, St Jacques S, Belln T, Letarte M, Bernabu C Assignment of the human endoglin gene (END) to 9q34 ->qter. Cytogenet Cell Genet 1993 ; 64 (3 4):204 7. 114. Qures hi ST, Gros P, Letarte M, Malo D. The murine endoglin gene (Eng) maps to chromosome 2. Genomics 1995 ; 26 (1):165 6. 115. Bernabeu C, Conley BA Vary CP Novel biochemical pathways of endoglin in vascular cell physiology. J Cell Biochem 2007 ; 102 (6):1375 88. 116. Koleva RI, Conley BA, Romero D, Riley KS, Marto JA, Lux A, Vary CP. Endoglin structure and function: Determinants of endoglin phosphorylation by transforming growth factor beta receptors. J Biol Chem 2006 ; 281 (35):25110 23. 117. Gougos A Letarte M. Primary structure of endoglin, an RGD containing glycoprotein of human endothelial cells. J Biol Chem 1990 ; 265 (15):8361 4. 118. Bellon T, Corb A, Lastres P, Cals C, Cebrin M, Vera S, Cheifetz S, Massague J, Letarte M, Bernabu C. Identification and exp ression of two forms of the human transforming growth factor beta binding protein endoglin with distinct cytoplasmic regions. Eur J Immunol 1993 ; 23 (9):2340 5. 119. ten Dijke P, Goumans MJ Pardali E. Endoglin in angiogenesis and vascular diseases. Angioge nesis 2008 ; 11 (1):79 89. 120. Buhring HJ, Mller CA, Letarte M, Gougos A, Saalmller A, van Agthoven AJ, Busch FW. Endoglin is expressed on a subpopulation of immature erythroid cells of normal human bone marrow. Leukemia 1991 ; 5 (10):841 7. 121. Gougos A, St Jacques S, Greaves A, O'Connell PJ, d'Apice AJ, Bhring HJ, Bernabeu C, van Mourik JA, Letarte M. Identification of distinct epitopes of endoglin, an RGD containing glycoprotein of endothelial cells, leukemic cells, and syncytiotrophoblasts. Int Immunol 1992 ; 4 (1):83 92. 122. Rokhlin OW, Cohen MB, Kubagawa H, Letarte M, Cooper MD. Differential expression of endoglin on fetal and adult hematopoietic cells in human bone marrow. J Immunol 1995 ; 154 (9):4456 65. 123. Jonker L Arthur HM. Endoglin expression i n early development is associated with vasculogenesis and angiogenesis. Mech Dev 2002 ; 110 (1 2):193 6.

PAGE 125

125 124. Ma X, Labinaz M, Goldstein J, Miller H, Keon WJ, Letarte M, O'Brien E. Endoglin is overexpressed after arterial injury and is required for transform ing growth factor beta induced inhibition of smooth muscle cell migration. Arterioscler Thromb Vasc Biol 2000 ; 20 (12):2546 52. 125. Piao M Tokunaga O. Significant expression of endoglin (CD105), TGFbeta 1 and TGFbeta R 2 in the atherosclerotic aorta: an i mmunohistological study. J Atheroscler Thromb 2006 ; 13 (2):82 9. 126. Torsney E, Charlton R, Parums D, Collis M, Arthur HM. Inducible expression of human endoglin during inflammation and wound healing in vivo. Inflamm Res 2002 ; 51 (9):464 70. 127. Velasco S, Alvarez Muoz P, Pericacho M, Dijke PT, Bernabu C, Lpez Novoa JM, Rodrguez Barbero A. L and S endoglin differentially modulate TGFbeta1 signaling mediated by ALK1 and ALK5 in L6E9 myoblasts. J Cell Sci 2008 ; 121 (Pt 6):913 9. 128. Blanco FJ, Grande MT, Langa C, Oujo B, Velasco S, Rodriguez Barbero A, Perez Gomez E, Quintanilla M, Lpez Novoa JM, Bernabeu C. S endoglin expression is induced in senescent endothelial cells and contributes to vascular pathology. Circ Res 2008 ; 103 (12):1383 92. 129. Perez Go mez E, Eleno N, Lpez Novoa JM, Ramirez JR, Velasco B, Letarte M, Bernabu C, Quintanilla M. Characterization of murine S endoglin isoform and its effects on tumor development. Oncogene 2005 ; 24 (27):4450 61. 130. Stepan H, Geipel A, Schwarz F, Krmer T, We ssel N, Faber R. Circulatory soluble endoglin and its predictive value for preeclampsia in second trimester pregnancies with abnormal uterine perfusion. Am J Obstet Gynecol 2008 ; 198 (2):175 e1 6. 131. Young BC, Levine RJ Karumanchi SA. Pathogenesis of pre eclampsia. Annu Rev Patho l 2010; 5 :17 3 92. 132. Hladunewich M, Karumanchi SA Lafayette R. Pathophysiology of the clinical manifestations of preeclampsia. Clin J Am Soc Nephrol 2007 ; 2 (3):543 9. 133. Grill S, Rusterholz C, Zanetti Dllenbach R, Tercanli S, Holzgreve W, Hahn S, Lapaire O. Potential markers of preeclampsia -a review. Reprod Biol Endocrinol 2009 ; 7 :70. 134. Luft FC. Soluble endoglin (sEng) joins the soluble fms like tyrosine kinase (sFlt) receptor as a pre eclampsia molecule. Nephrol Dial Tran splant 2006 ; 21 (11):3052 4.

PAGE 126

126 135. Levine RJ, Lam C, Qian C, Yu KF, Maynard SE, Sachs BP, Sibai BM, Epstein FH, Romero R, Thadhani R, Karumanchi SA; CPEP Study Group. Soluble endoglin and other circulating antiangiogenic factors in preeclampsia. N Engl J Me d 2006 ; 355 (10):992 1005. 136. Salahuddin S, Lee Y, Vadnais M, Sachs BP, Karumanchi SA, Lim KH. Diagnostic utility of soluble fms like tyrosine kinase 1 and soluble endoglin in hypertensive diseases of pregnancy. Am J Obstet Gynecol 2007 ; 197 (1):28 e1 6. 1 37. Lopez Novoa JM. Soluble endoglin is an accurate predictor and a pathogenic molecule in pre eclampsia. Nephrol Dial Transplant 2007 ; 22 (3):712 4. 138. Venkatesha S, Toporsian M, Lam C, Hanai J, Mammoto T, Kim YM, Bdolah Y, Lim KH, Yuan HT, Libermann TA, Stillman IE, Roberts D, D'Amore PA, Epstein FH, Sellke FW, Romero R, Sukhatme VP, Letarte M, Karumanchi SA. Soluble endoglin contributes to the pathogenesis of preeclampsia. Nat Med 2006 ; 12 (6):642 9. 139. Wipff J, Avouac J, Borderie D, Zerkak D, Lemarech al H, Kahan A, Boileau C, Allanore Y. Disturbed angiogenesis in systemic sclerosis: high levels of soluble endoglin. Rheumatology (Oxford) 2008 ; 47 (7):972 5. 140. Calabro L, Fonsatti E, Bellomo G, Alonci A, Colizzi F, Sigalotti L, Altomonte M, Musolino C, Maio M. Differential levels of soluble endoglin (CD105) in myeloid malignancies. J Cell Physiol 2003 ; 194 (2):171 5. 141. Lastres P, Letamenda A, Zhang H, Rius C, Almendro N, Raab U, Lpez LA, Langa C, Fabra A, Letarte M, Bernabu C. Endoglin modulates cel lular responses to TGF beta 1. J Cell Biol 1996 ; 133 (5):1109 21. 142. Lebrin F, Goumans MJ, Jonker L, Carvalho RL, Valdimarsdottir G, Thorikay M, Mummery C, Arthur HM, ten Dijke P. Endoglin promotes endothelial cell proliferation and TGF beta/ALK1 signal t ransduction. EMBO J 2004 ; 23 (20):4018 28. 143. Lee NY, Ray B, How T, Blobe GC. Endoglin promotes transforming growth factor beta mediated Smad 1/5/8 signaling and inhibits endothelial cell migration through its association with GIPC. J Biol Chem 2008 ; 283 ( 47):32527 33. 144. Dallas NA, Samuel S, Xia L, Fan F, Gray MJ, Lim SJ, Ellis LM. Endoglin (CD105): a marker of tumor vasculature and potential target for therapy. Clin Cancer Res 2008 ; 14 (7):1931 7. 145. Abdalla SA, Pece Barbara N, Vera S, Tapia E, Paez E, Bernabeu C, Letarte M. Analysis of ALK 1 and endoglin in newborns from families with hereditary hemorrhagic telangiectasia type 2. Hum Mol Genet 2000 ; 9 (8):1227 37.

PAGE 127

127 146. Ota T, Fujii M, Sugizaki T, Ishii M, Miyazawa K, Aburatani H, Miyazono K. 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):299 318. 147. Abdalla SA Letarte M. Hereditary haemorrhagic telangiectasia: current view s on genetics and mechanisms of disease. J Med Genet 2006 ; 43 (2):97 110. 148. Arthur HM, Ure J, Smith AJ, Renforth G, Wilson DI, Torsney E, Charlton R, Parums DV, Jowett T, Marchuk DA, Burn J, Diamond AG. Endoglin, an ancillary TGFbeta receptor, is require d for extraembryonic angiogenesis and plays a key role in heart development. Dev Biol 2000 ; 217 (1):42 53. 149. Perlingeiro, RC. Endoglin is required for hemangioblast and early hematopoietic development. Development 2007 ; 134 (16):3041 8. 150. Mancini ML, T erzic A, Conley BA, Oxburgh LH, Nicola T, Vary CP. Endoglin plays distinct roles in vascular smooth muscle cell recruitment and regulation of arteriovenous identity during angiogenesis. Dev Dyn 2009 ; 238 (10):2479 93. 151. Mahmoud M, Allinson KR, Zhai Z, Oa kenfull R, Ghandi P, Adams RH, Fruttiger M, Arthur HM. Pathogenesis of arteriovenous malformations in the absence of endoglin. Circ Res 2010; 106 (8 ):1425 33. 152. Fernandez LA, Sanz Rodriguez F, Blanco FJ, Bernabu C, Botella LM. Hereditary hemorrhagic tel angiectasia, a vascular dysplasia affecting the TGF beta signaling pathway. Clin Med Res 2006 ; 4 (1):66 78. 153. Sanz Rodriguez F, Fernandez L A, Zarrabeitia R, Perez Molino A, Ramrez JR, Coto E, Bernabeu C, Botella LM. Mutation analysis in Spanish patient s with hereditary hemorrhagic telangiectasia: deficient endoglin up regulation in activated monocytes. Clin Chem 2004 ; 50 (11):2003 11. 154. Urness LD, Sorensen LK Li DY Arteriovenous malformations in mice lacking activin receptor like kinase 1. Nat Genet 2000 ; 26 (3):328 31. 155. Oh SP, Seki T, Goss KA, Imamura T, Yi Y, Donahoe PK, Li L, Miyazono K, ten Dijke P, Kim S, Li E. Activin receptor like kinase 1 modulates transforming growth factor beta 1 signaling in the regulation of angiogenesis. Proc Natl Aca d Sci U S A 2000 ; 97 (6):2626 31. 156. Roman BL, Pham VN, Lawson ND, Kulik M, Childs S, Lekven AC, Garrity DM, Moon RT, Fishman MC, Lechleider RJ, Weinstein BM. Disruption of acvrl1 increases endothelial cell number in zebrafish cranial vessels. Development 2002 ; 129 (12):3009 19.

PAGE 128

128 157. Sorensen LK, Brooke BS, Li DY, Urness LD Loss of distinct arterial and venous boundaries in mice lacking endoglin, a vascular specific TGFbeta coreceptor. Dev Biol 2003 ; 261 (1):235 50. 158. Srinivasan S, Hanes MA, Dickens T, P orteous ME, Oh SP, Hale LP, Marchuk DA. A mouse model for hereditary hemorrhagic telangiectasia (HHT) type 2. Hum Mol Genet 2003 ; 12 (5):473 82. 159. Bourdeau A, Faughnan ME Letarte M. Endoglin deficient mice, a unique model to study hereditary hemorrhagic telangiectasia. Trends Cardiovasc Med 2000 ; 10 (7):279 85. 160. Bourdeau A, Faughnan ME, McDonald ML, Paterson AD, Wanless IR, Letarte M. Potential role of modifier genes influencing transforming growth factor beta1 levels in the development of vascular de fects in endoglin heterozygous mice with hereditary hemorrhagic telangiectasia. Am J Pathol 2001 ; 158 (6):2011 20. 161. Seki T, Hong KH, Yun J, Kim SJ, Oh SP. Isolation of a regulatory region of activin receptor like kinase 1 gene sufficient for arterial en dothelium specific expression. Circ Res 2004 ; 94 (8):e72 7. 162. Park SO, Wankhede M, Lee YJ, Choi EJ, Fliess N, Choe SW, Oh SH, Walter G, Raizada MK, Sorg BS, Oh SP. Real time imaging of de novo arteriovenous malformation in a mouse model of hereditary hem orrhagic telangiectasia. J Clin Invest 2009 ; 119 (11):3487 96. 163. Sohal DS, Nghiem M, Crackower MA, Witt SA, Kimball TR, Tymitz KM, Penninger JM, Molkentin JD. Temporally regulated and tissue specific gene manipulations in the adult and embryonic heart us ing a tamoxifen inducible Cre protein. Circ Res 2001 ; 89 (1):20 5. 164. Sabba C, Cirulli A, Rizzi R, Pasculli G, Gallitelli M, Specchia G, Liso V. Angiogenesis and hereditary hemorrhagic telangiectasia. Rendu Osler Weber disease. Acta Haematol 2001 ; 106 (4): 214 9. 165. Mahmoud M, Borthwick GM, Hislop AA, Arthur HM. Endoglin and activin receptor like kinase 1 are co expressed in the distal vessels of the lung: implications for two familial vascular dysplasias, HHT and PAH. Lab Invest 2009 ; 89 (1):15 25. 166. va n Laake LW, van den Driesche S, Post S, Feijen A, Jansen MA, Driessens MH, Mager JJ, Snijder RJ, Westermann CJ, Doevendans PA, van Echteld CJ, ten Dijke P, Arthur HM, Goumans MJ, Lebrin F, Mummery CL. Endoglin has a crucial role in blood cell mediated vasc ular repair. Circulation 2006 ; 114 (21):2288 97.

PAGE 129

129 167. Hao Q, Zhu Y, Su H, Shen F, Yang GY, Kim H, Young WL. VEGF Induces More Severe Cerebrovascular Dysplasia in Endoglin than in Alk1 Mice. Transl Stroke Res 2010; 1 (3) :197 201. 168. Hao Q, Su H, Marchuk DA, Rola R, Wang Y, Liu W, Young WL, Yang GY. Increased tissue perfusion promotes capillary dysplasia in the ALK1 deficient mouse brain following VEGF stimulation. Am J Physiol Heart Circ Physiol 2008 ; 295 (6):H2250 6. 169. Jerkic M, Peter M, Ardelean D, Fine M, Konerding MA, Letarte M. Dextran sulfate sodium leads to chronic colitis and pathological angiogenesis in Endoglin heterozygous mice. Inflamm Bowel Dis 2010; 16 ( 11):1859 70. 170. Xu B, Wu YQ, Huey M, Arthur HM, Marchuk DA, Hashimoto T, Young WL, Yang GY Vascular endothelial growth factor induces abnormal microvasculature in the endoglin heterozygous mouse brain. J Cereb Blood Flow Metab 2004 ; 24 (2):237 44. 171. Torsney E, Charlton R, Diamond AG, Burn J, Soames JV, Arthur HM. Mouse model for hereditary h emorrhagic telangiectasia has a generalized vascular abnormality. Circulation 2003 ; 107 (12):1653 7. 172. Toporsian M, Gros R, Kabir MG, Vera S, Govindaraju K, Eidelman DH, Husain M, Letarte M. A role for endoglin in coupling eNOS activity and regulating va scular tone revealed in hereditary hemorrhagic telangiectasia. Circ Res 2005 ; 96 (6):684 92. 173. Vasudevan A Bhide PG. Angiogenesis in the embryonic CNS: a new twist on an old tale. Cell Adh Migr 2008 ; 2 (3):167 9. 174. Gerhardt H Betsholtz C. How do endo thelial cells orientate? EXS 2005 ;94 :3 15. 175. Cambier S, Gline S, Mu D, Collins R, Araya J, Dolganov G, Einheber S, Boudreau N, Nishimura SL. Integrin alpha(v)beta8 mediated activation of transforming growth factor beta by perivascular astrocytes: an an giogenic control switch. Am J Pathol 2005 ; 166 (6):1883 94. 176. ten Dijke P Arthur HM. Extracellular control of TGFbeta signalling in vascular development and disease. Nat Rev Mol Cell Biol 2007 ; 8 (11):857 69. 177. Leveen P, Larsson J, Ehinger M, Cilio CM Sundler M, Sjstrand LJ, Holmdahl R, Karlsson S. Induced disruption of the transforming growth factor beta type II receptor gene in mice causes a lethal inflammatory disorder that is transplantable. Blood 2002 ; 100 (2):560 8. 178. Worthington JJ, Klemento wicz JE Travis MA TGFbeta: a sleeping giant awoken by integrins. Trends Biochem Sci 2010; 36 (1):47 54.

PAGE 130

130 179. Hodivala Dilke KM, Reynolds AR Reynolds LE. Integrins in angiogenesis: multitalented molecules in a balancing act. Cell Tissue Res 2003 ; 314 (1):1 31 44. 180. Sheppard D. Endothelial integrins and angiogenesis: not so simple anymore. J Clin Invest 2002 ; 110 (7):913 4. 181. Silva R, D'Amico G, Hodivala Dilke KM, Reynolds LE. Integrins: the keys to unlocking angiogenesis. Arterioscler Thromb Vasc Biol 2008 ; 28 (10):1703 13. 182. Bader BL, Rayburn H, Crowley D, Hynes RO. Extensive vasculogenesis, angiogenesis, and organogenesis precede lethality in mice lacking all alpha v integrins. Cell 1998 ; 95 (4):507 19. 183. Zhu J, Motejlek K, Wang D, Zang K, Schmidt A, Reichardt LF. beta8 integrins are required for vascular morphogenesis in mouse embryos. Development 2002 ; 129 (12):2891 903. 184. Taya Y, O'Kane S Ferguson MW Pathogenesis of cleft palate in TGF beta3 knockout mice. Development 1999 ; 126 (17):3869 79. 1 85. Mu Z, Yang Z, Yu D, Zhao Z, Munger JS. TGFbeta1 and TGFbeta3 are partially redundant effectors in brain vascular morphogenesis. Mech Dev 2008 ; 125 (5 6):508 16. 186. McCarty JH, Lacy Hulbert A, Charest A, Bronson RT, Crowley D, Housman D, Savill J, Roes J, Hynes RO. Selective ablation of alphav integrins in the central nervous system leads to cerebral hemorrhage, seizures, axonal degeneration and premature death. Development 2005 ; 132 (1):165 76. 187. Proctor JM, Zang K, Wang D, Wang R, Reichardt LF. Vasc ular development of the brain requires beta8 integrin expression in the neuroepithelium. J Neurosci 2005 ; 25 (43):9940 8. 188. Li WY, Dudas M Kaartinen V. Signaling through Tgf beta type I receptor Alk5 is required for upper lip fusion. Mech Dev 2008 ; 125 ( 9 10):874 82. 189. Gustafsson E, Brakebusch C, Hietanen K, Fssler R. Tie 1 directed expression of Cre recombinase in endothelial cells of embryoid bodies and transgenic mice. J Cell Sci 2001 ; 114 (Pt 4):671 6. 190. Kisanuki YY, Hammer RE, Miyazaki J, Willi ams SC, Richardson JA, Yanagisawa M. Tie2 Cre transgenic mice: a new model for endothelial cell lineage analysis in vivo. Dev Biol 2001 ; 230 (2):230 42. 191. Sridurongrit S, Larsson J, Schwartz R, Ruiz Lozano P, Kaartinen V. Signaling via the Tgf beta type I receptor Alk5 in heart development. Dev Biol 2008 ; 322 (1):208 18.

PAGE 131

131 192. Jiao K, Langworthy M, Batts L, Brown CB, Moses HL, Baldwin HS. Tgfbeta signaling is required for atrioventricular cushion mesenchyme remodeling during in vivo cardiac development. Dev elopment 2006 ; 133 (22):4585 93. 193. Robson A, Allinson KR, Anderson RH, Henderson DJ, Arthur HM. The TGFbeta type II receptor plays a critical role in the endothelial cells during cardiac development. Dev Dyn 2010; 239 (9) :2435 42. 194. Wang YQ, Sizeland A Wang XF, Sassoon D. Restricted expression of type II TGF beta receptor in murine embryonic development suggests a central role in tissue modeling and CNS patterning. Mech Dev 1995 ; 52 (2 3):275 89. 195. Lyden D, Young AZ, Zagzag D, Yan W, Gerald W, O'Reil ly R, Bader BL, Hynes RO, Zhuang Y, Manova K, Benezra R. Id1 and Id3 are required for neurogenesis, angiogenesis and vascularization of tumour xenografts. Nature 1999 ; 401 (6754):670 7. 196. Dejana E Nyqvist D. News from the brain: the GPR124 orphan recept or directs brain specific angiogenesis. Sci Tran sl Med 2010; 2 (58 ):58ps53. 197. Tanaka Y, Naruse I, Hongo T, Xu M, Nakahata T, Maekawa T, Ishii S. Extensive brain hemorrhage and embryonic lethality in a mouse null mutant of CREB binding protein. Mech Dev 2000 ; 95 (1 2):133 45. 198. Daneman R, Agalliu D, Zhou L, Kuhnert F, Kuo CJ, Barres BA. Wnt/beta catenin signaling is required for CNS, but not non CNS, angiogenesis. Proc Natl Acad Sci U S A 2009 ; 106 (2):641 6. 199. Kuhnert F, Mancuso MR, Shamloo A, Wang HT Choksi V, Florek M, Su H, Fruttiger M, Young WL, Heilshorn SC, Kuo CJ. Essential regulation of CNS angiogenesis by the orphan G protein coupled receptor GPR124. Scien ce 2010; 330 (6006):985 9. 200. Johansson A, Driessens M Aspenstrom P. The mammalian hom ologue of the Caenorhabditis elegans polarity protein PAR 6 is a binding partner for the Rho GTPases Cdc42 and Rac1. J Cell Sci 2000 ; 113 ( Pt 18) :3267 75. 201. Vasudevan A, Long JE, Crandall JE, Rubenstein JL, Bhide PG. Compartment specific transcription factors orchestrate angiogenesis gradients in the embryonic brain. Nat Neurosci 2008 ; 11 (4):429 39. 202. Simpson TI, Pratt T, Mason JO, Price DJ. Normal ventral telencephalic expression of Pax6 is required for normal development of thalamocortical axons in embryonic mice. Neural Dev 2009 ; 4 : 1 9.

PAGE 132

132 BIOGRAPHICAL SKETCH Ha Long Nguyen, the youngest of five children, was born and raised in Clearwater, Florida. After graduating from Clearwater High School in 2001, she attended the University of Central Florida i n Orlando, FL, where she majored in m olecular b iology and m icrobiology. During summer 2003 her first scientific research experience was in the laboratory of Dr. My Lien Dao at the University of South Florida, looking at the role of Streptococcus mutans in tooth decay. During her junior and senior years she joined the laboratory of Dr. Antonis Zervos and researched apoptotic processes in cardiovascular disease, including screening for apoptotic inhibitors from natural extracts. After receiving her Bachelor of Science in 2005, she worked for Dr. Steven Ebert and studied the role of catecholamines in cardiac development. In 2006 she entered the Interdisciplinary Program in Biomedical Sciences at the University of Florida. In spring 2007, she joined the laborat ory of Dr. S. Paul Oh, where her dissertation research involved the generation and characterization of conditional knockout mouse models of different TGF endothelial cells.