Reconstructing the Vascular Regenerative Niche with Oxygen-Controllable Hydrogels

Home , Tunica intima, Tunica media, Vasa vasorum

RECONSTRUCTING THE VASCULAR REGENERATIVE NICHE WITH OXYGEN-CONTROLLABLE HYDROGELS

By Michael R. Blatchley

A dissertation submitted to Johns Hopkins University in conformity with the requirements for

the degree of Doctor of Philosophy

Baltimore, Maryland July 2019

1 Abstract

A thorough understanding of the molecular and microenvironmental regulators of blood vessel formation is key to designing new therapeutic strategies to treat two of the leading causes of death worldwide, cardiovascular disease and cancer. Two key parameters of the vascular regenerative niche that play broad and potent roles in endothelial cell (EC) morphogenesis are matrix stiffness and hypoxia

(defined as <5% O2). To recapitulate these cues, we designed hydrogels comprised of gelatin and dextran backbones to study the effects of O2 and matrix stiffness on ECs. For both of these hydrogel backbones, we conjugated phenolic moieties, which can be enzymatically crosslinked in an O2-consuming reaction to form a hypoxic hydrogel. By altering the concentration of each polymer component and adding or omitting a secondary crosslinker, we generated a set of hydrogels that spanned both a range of physiological stiffnesses and hypoxic conditions.

We studied the mechanism of cluster-based vasculogenesis that had previously been reported in vivo, but never studied in vitro because of the limitations with existing cell culture platforms. Using our

O2-controllable hydrogels, we observed and quantified the kinetics of cluster formation, and determined key regulators of the process. EC exposure to hypoxia led to upregulation of reactive oxygen species and proteases that facilitated rapid matrix degradation and cluster formation. ECs sprouted from clusters via cell-matrix interactions to form expansive vascular networks, which could be further enhanced through secondary crosslinking to increase cell-matrix interactions. In vivo studies in a mouse model confirmed the proposed mechanism. Finally, in depth analysis of specific signaling involved in cluster formation included genes associated with cell cycle, cell survival, and metabolism. This work has illuminated key regulators of a complex mechanism for vasculogenesis, and represents a powerful in vitro tool to study cells in an environment that closely matches the niche in which they reside in vivo.

Thesis advisor: Dr. Sharon Gerecht Thesis committee members: Dr. Warren Grayson and Dr. Akrit Sodhi

Acknowledgements

I would first like to thank my thesis advisor, Dr. Sharon Gerecht, for providing me with exemplary mentorship. From the first meeting I had with her to discuss a rotation in the lab, she listened to my interests and helped guide me to learn the skills necessary to become a successful PhD student, as well as set myself up well for the next stages of my career. She believed in me as a scientist and writer from the early stages of my time in the lab, and has been invaluable in helping me refine and expand my skillset both in the lab and in disseminating my work. There were high points and low points during my time as a PhD student, and it was great to know I would always have Dr. Gerecht’s support to work through the difficult times and celebrate the successes.

I would also like to thank my thesis committee members, Dr. Warren Grayson and Dr. Akrit

Sodhi, for helping guide and refine my lines of inquiry. They have challenged me to think critically about my work, have asked thoughtful questions, and have suggested a number of important experiments. Our discussions have also helped me refine the best ways to present my work, and have helped me improve how I design an experimental plan.

Letters of recommendation are a crucial component in attaining fellowships, awards, and future jobs. I have applied for many fellowships and awards over the last six years and am incredibly thankful for all those professors that have written letters for me since applying for graduate school: Dr. Kinam

Park, Dr. Clint Chapple, Dr. Irina Petrache, Dr. Jennifer Elisseeff, Dr. Sharon Gerecht, Dr. Hai-Quan

Mao, Dr. Honggang Cui, Dr. Jason Burdick, and Dr. Michael Miller.

I also want to thank the past and current members of the Gerecht lab. We are a wonderfully diverse and ceaselessly entertaining and wonderful collection of young scientists. Our conversations in the office range from incredibly productive discussions of science and experimental design, to venting about reviewers, to talking about new restaurants and movies. It is great to have a group of people so willing talk science as well as life outside of science.

iii

When I first joined the lab, I learned from many lab members, but Dr. Kyung Min Park served as my primary mentor. Kyung was a great mentor to me and served as a role model through his work ethic and his commitment to always better himself as a scientist, communicator, and husband and father. I have had many productive scientific discussions with members of the lab and am specifically thankful for fruitful discussions with Dr. Xin Yi Chan, Dr. Tom Shen, Dr. Quinton Smith, Dr. Hawley Pruitt, Dr. Zhao

Wei, Dr. Dan Lewis, Matt Davenport, Bria Macklin, Justin Lowenthal, Morgan Elliott, John Jamieson,

Eugenia Volkova, and Franklyn Hall. The members of the lab have become good friends of mine and I have really enjoyed getting to know everyone better through game nights and happy hours.

It has also been wonderful to have a competitive outlet throughout my time in graduate school, so

I am thankful for the many opportunities to play soccer, basketball, football, and softball with friends from the lab, the BME and ChemBE programs, and outside of Hopkins in various city leagues.

I have had the opportunity to mentor a number of students: Brooke Smith, Songnan Wang,

Franklyn Hall, Arianne Papa, and Vidur Kailash, over the last six years and am thankful for the opportunity to improve myself as a mentor and very thankful for their patience as I improved my mentorship skills.

The INBT and BME administration have been instrumental to my success in graduate school.

Hong Lan has been a wonderful resource in the BME department, and I am grateful to have had the opportunity to work closely with her as president of the PhD council. Conversations with Hong never fail to be entertaining and she is an incredibly positive presence in the department and Hopkins community. I am also thankful to the INBT staff, including Kierra Suggs, Ellie Boettinger-Heasley, Gregg Nass, Carla

Dodd, Christine Duke, Jon French, Camille Mathis, Lateai Jones, Ada Simari, Sky Tharp, Luke

Thorstenson, and Gina Wadas.

I will always be grateful to have received such wonderful mentorship from my time as an undergraduate until now. Dr. Kinam Park, John Garner, Dr. Clint Chapple, Dr. Kelly Schweitzer, Dr.

Irina Petrache, and Dr. Jennifer Elisseeff all provided amazing research opportunities and helped me develop new research skills each step of the way.

I would be nowhere without the continued support of my family. I am extremely fortunate to have a family who has an empathetic view towards the rigors of a PhD program. My dad was my first research mentor, and I really want to thank him for giving me the opportunity to work in his lab in high school, and will be forever thankful for our discussions of science and how to best set myself up for career success on our long bike rides. I have learned the importance of a long bike ride, run, hike, or walk from my dad, and these times have been valuable in organizing my thoughts, planning a paper or figures, or simply clearing my head after a long and frustrating day.

Finally, I want to give thanks to my wonderful wife, Kaylee. She has supported me through my time in graduate school and has encouraged me to take a break or two from working and email for a number of long weekend trips from Shenandoah to the Adirondacks. I am grateful for her patience and understanding through the busy grant and paper writing times, and for listening to my complaints. There have been times that I have been overwhelmed with work, and she is always there to support me and encourage me. She always challenges me to become a better person, and I couldn’t ask for a better partner in life.

Thank you!

Table of Contents

ABSTRACT ...... ii

ACKNOWLEDGEMENTS ...... iii-v

LIST OF TABLES ...... xii

LIST OF FIGURES ...... xiii-xiv

1 Reconstructing the Vascular Developmental Milieu In Vitro ...... 1 1.1 Teaser ...... 1 1.2 Introduction ...... 1 1.3 Evolution of the vascular system ...... 2 1.4 Early vascular development – vasculogenesis ...... 3 1.4.1 Formation of the primary vascular plexus ...... 3 1.4.2 Aterio-venous fate specification ...... 7 1.5 Expanding and remodeling the vascular system – angiogenesis...... 8 1.5.1 Tip cell selection and initiation of sprouting ...... 10 1.5.2 Effectors of sprouting and branching ...... 10 1.6 Lumen formation ...... 13 1.6.1 Cell hollowing ...... 14 1.6.2 Cord hollowing ...... 15 1.7 Human stem-cell derived endothelial cells and perivascular cells ...... 16 1.8 Environmental control of EC differentiation and network assembly ...... 19 1.8.1 ECM organization and composition ...... 19 1.8.2 2D matrix manipulation ...... 21 1.8.3 3D matrix manipulation ...... 23 1.8.4 Hypoxia ...... 28 1.8.5 Hemodynamic flow ...... 30 1.8.6 Gradients ...... 32 1.9 Future perspectives and conclusions ...... 34

2 Acellular Implantable and Injectable Hydrogels for Vascular Regeneration ...... 36 2.1 Teaser ...... 36 2.2 Introduction ...... 36 2.3 Incorporating biological factors in hydrogels to modulate vasculogenesis and angiogenesis ...... 40 2.3.1 VEGF ...... 41 2.3.2 bFGF ...... 43 2.3.3 Other growth factors/cytokines ...... 45 2.3.4 Other biological factors ...... 46 2.3.5 Combinations of biological factors for angiogenesis/vasculogenesis ...... 49 2.4 Altering physical properties of hydrogels to regulate vasculogenesis and angiogenesis ...... 54 2.4.1 Geometry and ultrastructure (pore size, connectivity) ...... 54 2.4.2 Degradation ...... 56 2.4.3 Hydrogel stiffness ...... 57 2.5 Other microenvironmental factors affecting vasculogenesis: future opportunities ...... 58 2.5.1 Hypoxia ...... 58 2.5.2 Reactive oxygen species ...... 59 2.5.3 Hyperoxia ...... 60 2.6 Conclusions and future directions for the field ...... 61 3 Hypoxia and Matrix Manipulation for Vascular Engineering ...... 62 3.1 Teaser ...... 62 3.2 Introduction ...... 62 3.3 Concepts in the Regulation of the Vasculature by Oxygen and the ECM ...... 66 3.3.1 The influence of oxygen tension on vascularization ...... 66 3.3.2 The in vivo consequences of oxygen gradients...... 66 3.3.2.1 Oxygen availability in the body ...... 66 3.3.2.2 Oxygen-sensing mechanisms of vascular cells ...... 71 3.3.3 Cellular responses to different oxygen concentrations ...... 74 3.3.3.1 Metabolism and oxygen uptake rate ...... 74 3.3.3.2 Transcription of angiogenic genes ...... 76 3.3.3.3 Cell death and survival ...... 77

vii

3.3.4 Cell pluripotency and differentiation ...... 78 3.4 Vascular responses to ECM ...... 80 3.4.1 Types of ECM found participating in vascularization ...... 80 3.4.2 Properties of the ECM that affect vascular morphogenesis ...... 84 3.4.2.1 Cell adhesion regulates neovascularization ...... 86 3.4.2.2 Scaffold degradation regulates vascular morphogenesis ...... 87 3.4.2.3 Physical orientation of the ECM ...... 89 3.4.2.4 Regulating matrix mechanics ...... 91 3.5 The effects of oxygen availability on the ECM ...... 94 3.5.1 Varying oxygen tensions in the ECM of tissue and matrix scaffolds: measuring/modeling ...... 94 3.5.1.1 Oxygen measurement techniques and challenges ...... 94 3.5.1.2 Modeling oxygen transport in tissues ...... 96 3.5.1.2.1 Static models ...... 96 3.5.1.2.2 Dynamic and in vivo models ...... 99

3.5.2 Targeted cellular responses to O2 availability in matrix hydrogels ...... 100 3.6 Future Directions ...... 105 4 Overview of Materials and Experimental Methods ...... 108 4.1 Materials ...... 108 4.2 Polymer synthesis and characterization ...... 110 4.2.1 Synthesis of gelatin-g-ferulic acid (Gtn-FA) ...... 110 4.2.2 Synthesis of aminated dextran-g-poly(ethylene glycol)-tyramine (DexE-PT) ...... 110 4.2.3 Characterization of hydrogel precursor solutions ...... 111 4.3 Hydrogel formation and characterization...... 112

4.3.1 Preparation of hypoxia-inducible (O2-controllable) hydrogels ...... 112

4.3.2 O2 measurements...... 113 4.3.3 Measurement of viscoelastic properties ...... 113

4.4 Cell culture and analysis in hypoxia-inducible (O2-controllable) hydrogels ...... 114

4.4.1 ECFC encapsulation in hypoxia-inducible (O2-controllabe) hydrogels ...... 114

4.4.2 Mimicking hypoxic conditions in nonhypoxic geometries (1% O2 flush, DMOG, COCl2) ...... 115 4.4.3 siRNA transfection ...... 116

viii

4.4.4 ROS inhibition (DPI) ...... 116 4.4.5 Altering matrix stiffness with mTG ...... 116 4.4.6 Dynamic matrix stiffening ...... 117 4.4.7 Protease activity assay ...... 117 4.4.8 Proteome profiler array ...... 118 4.4.9 Protease inhibition (GM6001, Z-FF-FMK, PI, EDTA) ...... 118 4.4.10 Addition of exogenous MMP ...... 118 4.4.11 Observation of cell-cell interactions and vascular network formation ...... 119 4.5 Cell culture and analysis in layered hypoxia-inducible (O2-controllable) hydrogels ...... 119 4.5.1 ECFC encapsulation in layered hydrogels ...... 119 4.5.2 CellROX for oxidative stress detection ...... 120 4.5.3 RNA extraction ...... 121 4.5.4 RNA-sequencing ...... 121 4.5.5 Analysis of RNA-seq data ...... 122 4.6 In vivo mouse model ...... 122 4.6.1 Plug assay in vivo ...... 122

4.6.2 O2 measurements in vivo ...... 123 4.7 Image and data analysis ...... 124 4.7.1 Image quantification ...... 124 4.7.2 Statistical analysis ...... 124 5 Designer Hydrogels for Precision Control of Oxygen Tension and Mechanical Properties ...... 125 5.1 Teaser ...... 125 5.2 Introduction ...... 125 5.3 Results and Discussion ...... 128 5.3.1 Hybrid Gtn-Dex hydrogel formation and characterization ...... 128 5.3.2 Hybrid Gtn-Dex hydrogels for control of oxygen tension and mechanical properties ...... 130 5.3.3 Decoupling TH and mechanical properties through addition of non-conjugated phenol-containing molecules ...... 134 5.3.4 Decoupling mechanical properties and TH through secondary crosslinking ...... 134 5.3.5 Varying oxygen tension and mechanical stiffness alters vasculogenesis in vitro ...... 138

5.4 Conclusions ...... 142 5.5 Supplementary information ...... 144

6 Applications of O2-Controllable Hydrogels ...... 145 6.1 Teaser ...... 145 6.2 Introduction ...... 145 6.3 Methods to culture under hypoxic conditions ...... 147 6.4 Development of the protocol...... 151 6.5 Applications of the protocol ...... 154 6.5.1 Cancer ...... 154 6.5.2 Vascular applications ...... 154 6.6 Limitations of the platform ...... 155 6.7 Experimental design ...... 156 6.8 Anticipated results ...... 157 7 Hypoxia and Matrix Viscoelasticity Sequentially Regulate Endothelial Progenitor Cluster-Based Vasculogenesis ...... 162 7.1 Teaser ...... 162 7.2 Introduction ...... 162 7.3 Results and Discussion ...... 165 7.3.1 Hypoxia mediates endothelial cluster formation ...... 165 7.3.2 ECFC cluster formation is governed by reactive oxygen species (ROS) production and dependent on matrix stiffness ...... 168 7.3.3 Protease-mediated matrix degradation is required for hypoxic cluster formation and stabilization 171 7.3.4 Increased viscoelasticity promotes vascular sprouting from clusters ...... 176 7.3.5 ECFCs form clusters in vivo ...... 180 7.4 Defining a new mechanism for cluster-based vasculogenesis ...... 184 7.5 Supplementary information ...... 187

8 Layered O2-Controllable Hydrogels Facilitate Study of Uniform Cell Behavior to Identify a Detailed Mechanistic Understanding of Cluster-Based Vasculogenesis ...... 198 8.1 Teaser ...... 198 8.2 Introduction ...... 198 8.3 Results and Discussion ...... 200

8.3.1 Development and characterization of layered hydrogels ...... 200 8.3.2 Analysis of uniform cell behavior in layered hydrogels ...... 201 8.3.3 Confirmation of a conserved mechanism for cluster formation between two-layer and conventional hypoxic hydrogels ...... 204 8.3.4 Global view of RNA sequencing of two-layer hypoxic and nonhypoxic hydrogels ...... 205 8.3.5 Genes associated with oxidative stress are differentially expressed over the experimental time course ...... 208 8.3.6 A vast array of genes encoding for proteases are differentially expressed over the experimental time course, necessitating broad protease inhibition to block cluster formation ...... 209 8.3.7 Cell-cell interactions stabilize ECFC clusters ...... 213 8.3.8 Cell survival, apoptosis, and cell cycle progression are differentially regulated concomitant with cluster formation ...... 214 8.3.9 Carbohydrate metabolism is upregulated in hypoxic conditions ...... 217 8.4 Conclusions ...... 219 9 Conclusions and future directions ...... 222 9.1 Conclusions ...... 222 9.2 Future directions ...... 224 9.2.1 Acellular therapeutics ...... 224 9.2.2 Vascular biology and vascular tissue engineering ...... 225 References ...... 227 Curriculum Vitae ...... 257

List of Tables

Table 2.1 Combinations of Growth Factors for Therapeutic Angiogenesis ...... 51 Table 5.1 Concentrations of Crosslinking Moieties ...... 129 Table 6.1 Definitions of Oxygen Concentration and Usual In Vivo Distribution ...... 147

xii

List of Figures

Figure 1.1 Vertebrate vascular development ...... 9 Figure 1.2 Effectors of branching morphogenesis ...... 13 Figure 1.3 Human stem cell derived vascular cells...... 20 Figure 1.4 Environmental regulators of EC fate and platforms/methods to incorporate them in vitro ...... 31 Figure 1.5 Effect of gradients on gene expression, cell fate, and cell phenotype ...... 34 Figure 2.1 Treatment of blood vessel occlusion by injection or implantation of hydrogels ...... 37

Figure 3.1 Controlling ECM and O2 in vitro...... 107 Figure 5.1 Crosslinking chemistry for hybrid Gtn-FA and DexE-PT hydrogels ...... 131 Figure 5.2 Dissolved oxygen (DO) and rheological characterization for hybrid hydrogels ...... 133 Figure 5.3 Addition of non-conjugated TA enables control over TH with minimal effects on G′ ...... 136 Figure 5.4 Dissolved oxygen (DO) and rheological characterization for hybrid hydrogels with secondary crosslinking microbial transglutaminase (mTG) ...... 138 Figure 5.5 Vascular morphogenesis in an array of microenvironmental conditions ...... 141 Figure S5.1 1H-NMR spectra for Gtn-FA and DexE-PT ...... 144 Figure 6.1. Schematic representation of the gelatin-based hypoxia-inducible (Gel-HI) hydrogels for in vitro and in vivo applications ...... 150 Figure 6.2 Tubulogenesis in Gel-HI ...... 160 Figure 6.3 Subcutaneous and intramuscular in vivo oxygen sensing ...... 161 Figure 7.1 ECFC clusters form only in hypoxic conditions ...... 167 Figure 7.2 Cluster formation is dependent on hypoxia-induced ROS and matrix stiffness ...... 170 Figure 7.3 MMP-1 -mediated matrix degradation is required for cluster formation and stabilization ..... 175 Figure 7.4 Sprouting is enhanced via ECM interactions and as matrix stiffness increases ...... 179 Figure 7.5 Gel-HI hydrogels facilitate ECFC cluster formation in vivo ...... 183 Figure 7.6 A mechanism for cluster-based vasculogenesis ...... 186

Figure S7.1 Hydrogel height controls O2 gradients ...... 187 Figure S7.2 Cluster size over time ...... 188

Figure S7.3 Low O2 tension, rather than diffusional limitations or nutrient deprivation, facilitates cluster formation ...... 189 Figure S7.4 Cluster formation is HIF-independent and ROS-mediated ...... 190

xiii

Figure S7.5 Experimental setup for timelapse monitoring of increases in fluorescence upon proteolytic degradation of DQ-gelatin ...... 191 Figure S7.6 Protease activity and inhibition ...... 192 Figure S7.7 Vascular sprouting from clusters ...... 194 Figure S7.8 Dynamic matrix stiffening accelerates vascular network formation ...... 195 Figure S7.9 Increased matrix viscoelasticity influences vascular network formation ...... 196

Figure S7.10 In vivo O2 measurements...... 197 Figure 8.1 Three-layered hydrogel design and characterization ...... 201 Figure 8.2: Two-layer hypoxic hydrogels enable uniform ECFC cluster behavior, and match conventional hypoxic hydrogel cluster formation kinetics and O2 ...... 203 Figure 8.3 Oxidative stress is upregulated at early timepoints in ECFCs cultured in two-layer hypoxic hydrogels compared to nonhypoxic hydrogels ...... 205 Figure 8.4 Differential gene expression between ECFCs encapsulated in hypoxic and nonhypoxic hydrogels ...... 207 Figure 8.5 Oxidative stress-associated differential gene expression...... 209 Figure 8.6 Numerous proteases are upregulated in hypoxic conditions, thereby requiring broad inhibition of proteases to block cluster formation ...... 212 Figure 8.7 Significantly differentially expressed genes associated with binding of ECs ...... 213 Figure 8.8 Cell survival, apoptosis, and cell cycle progression are differentially regulated as ECFC clusters form...... 216 Figure 8.9 Carbohydrate metabolism is upregulated in hypoxic conditions ...... 218 Figure 8.10 Proposed mechanism for cluster formation ...... 221

xiv

CHAPTER 1

1 Reconstructing the Vascular Developmental Milieu In Vitro

1.1 Teaser

Insights from model organisms and 2D stem cell culture have guided our understanding of human vascular development. Incorporating characteristics of the developmental milieu into in vitro culture platforms enhances biomimicry to study vascular development and morphogenesis.

1.2 Introduction

The collective guiding principles informing our understanding of human development are an amalgamation of knowledge gained examining model organisms as well as studying human cells in conditions that are ingeniously designed for ease of analysis, but ultimately fall short in recapitulating the multi-factorial nature of the developmental microenvironment. Advancement in any field requires identification of current limitations, concurrent with development of novel tools.

As our knowledge of development continues to grow alongside an ever-expanding library and complexity of biomaterials, an exchange of ideas between these two areas of research has been, and will continue to be, instrumental in driving innovation. To generate new ideas and test new hypotheses, the requisite background knowledge is key. In this review, we describe details of vascular development from a biological perspective, then provide insight into the approaches by which biological principles can be accurately studied and perturbed in vitro towards identifying therapeutics targets, as well as improve our methods for engineering and building vascularized tissues. In particular, we will focus on the development and organization of stem-cell derived endothelial cells into vascular networks. We point the reader to reviews of perivascular/mural cells, including smooth muscle cells and pericytes in [1-3] for additional detail regarding the role of

1 these other cell types. Further, we aim to describe methods and platforms that enable such vascular networks to self-assemble, with a focus on top-down approaches, where cells are exposed to a biomimetic microenvironment that is amenable to endothelial fate specification and vascular network organization. Advances in bottom-up approaches, including advanced bioprinting of blood vessels and microfluidic platforms have been described elsewhere [4-7].

1.3 Evolution of the vascular system

The evolutionary expansion of organisms beyond the simplest unicellular and 2D multicellular organisms to complex, 3D organisms led to the requisite formation of an O2 delivery system [8, 9]. The size of larger organisms outgrew the transport capacity of passive diffusion and, in order to meet metabolic demands for growth, development, and homeostasis, an efficient delivery system for O2 and other nutrients developed [10]. Over time, the capacity for complex branching to reach all tissues was optimized and this system continually evolved to contain mechanisms for fluid flow by ciliary function or muscle contraction, eventually leading to the formation of the heart. In vertebrate organisms, an endothelial lining developed distinct advantages over previous delivery networks to: provide a smooth surface for flow in order to permit higher blood pressures required to pump O2-containing blood and nutrients in larger organisms, enable the capacity for thrombogenicity in response to vascular injury, establish mechanisms for vasoregulation, and facilitate selective permeability [8, 9]. The continued advancement and optimization of the O2 delivery system is fascinating to track, and many evolutionarily conserved phenomena persist in the model organisms used to study the mechanisms for tubular oxygen delivery in Drosophila and other insects, as well as astounding conservation of mechanistic and morphological organization in vertebrate vascular development across model organisms such as zebrafish, amphibians, avian species, and mice, with homologous genes and proteins governing

2 many related processes [11-14]. It is however, critically important to recognize the prevalence of key differences between the developmental and cell biology of model organisms and human biology [15, 16]. Notably, an understanding of the complex and dynamic intricacies of vascular development in these model organisms has facilitated the design of advanced systems to study and perturb human cells in vitro to understand vascular development and regeneration.

1.4 Early vascular development - vasculogenesis

Studies in vertebrate embryos, most notably in mice, amphibians, avian species, and zebrafish have illuminated many key regulators of early embryonic vascular development, which often manifest as events involving pairs of receptor tyrosine kinases and their associated ligands

[17]. Numerous excellent reviews have described these regulators in great detail [12, 18-21]. Here, we identify the main regulators and key morphogenetic events in the formation of the extraembryonic primary vascular plexus (Fig. 1.1), as well as the formation of the initial blood vessels within the embryo, both of which contribute to the formation and development of the vertebrate vascular system. Key aspects of vascular development rely on intrinsic genetic pre- patterning and regulation, while remaining highly plastic in order to adapt to extrinsic parameters of the dynamic developmental microenvironment. Identifying and controlling these environmental cues has garnered increasing interest in recent years.

1.4.1 Formation of the primary vascular plexus

Post-gastrulation, the vascular system is the first functional organ system to develop in vertebrate organisms [22]. This early development facilitates transport of nutrients and removal of waste required by the rapidly growing embryo. Cells of the endothelial lineage are derived from mesodermal precursors (Fig. 1.1A). During gastrulation, embryonic epiblast cells are recruited to

3 the primitive streak, where they undergo epithelial to mesenchymal transition (EMT), then migrate between the visceral endoderm and epiblast to form mesoderm or definitive endoderm [23].

Mesodermal specification is reliant on numerous cell signaling pathways, including fibroblast growth factor 2 (FGF-2) [24], bone morphogenetic protein 4 (BMP-4) [25], Activin-A [26], and

Wnt signaling [27]. Initially, mesodermal cells are isolated from one another and migrate through a soft (<20 Pa) [28], glycosaminoglycan (GAG)-rich microenvironment. The onset of mesoderm is concurrent with increasing concentration of one specific GAG, hyaluronic acid (HA) [29].

Notably, mesodermal lineage specification is also enhanced in low O2 (hypoxic) microenvironments [30], as well as soft substrata [31, 32]. The latter studies highlight the recent interest in regulating activity of mechanosensitive signaling via the Wnt/β-catenin pathway that can be manipulated by controlling the mechanics of the local microenvironment. Matrix mechanics can drive the cell rearrangements that are required for EMT-driven mesodermal specification.

Within the mesodermal population, cells of endothelial and hematopoietic lineage arise.

Debate continues regarding the origins of these two cell populations during primitive hematopoiesis, with one theory hypothesizing a common bipotent progenitor, the hemangioblast, gives rise to both cell types. Another theory suggests that these two cell populations arise in parallel from distinct mesodermal precursors [33]. Regardless of their origin, these cells are further specified to generate angioblasts (endothelial precursors) surrounding hematopoietic populations of cells in blood islands (Fig. 1.1B). However, this subpopulation of cells is not the only source of endothelial cells (ECs), as they may also arise in isolation from hematopoietic cells [34]. Broadly speaking, in the earliest stages of vasculogenesis, these two cell populations give rise to extraembryonic vasculogenesis through blood island formation, blood island fusion, and the development of the primary vascular plexus, while isolated angioblasts aggregate to form the basis

4 for the first two blood vessels intraembryonically, the dorsal aorta and the cardinal vein [19, 35].

From here, an elegant set of papers have confirmed the presence of the hemogenic endothelium, which is a small population (<2%) of the endothelial cells within the early vascular plexus, including the yolk sac, the placenta, and the aorta-gonad-mesonephros (AGM), that possess the capacity for definitive hematopoiesis [36-42]. These modern studies have confirmed Florence

Sabin’s observations nearly a century ago, of cells migrating, or ‘budding’ from a group of ECs to generate morphologically and functionally distinct hematopoietic precursors through endothelial to hematopoietic transition [41, 43]. The cells originating from the hemogenic endothelium establish the lineal basis for nearly the entire composition of the adult hematopoietic system. These cells ultimately migrate from the hemogenic endothelial to other sites of definitive hematopoiesis, including the fetal liver and the bone marrow, where they expand [33].

Angioblasts express vascular endothelial growth factor receptor-2 (VEGFR-2/KDR/Flk-1)

[19], which enables them to respond to production of vascular endothelial growth factor (VEGF) in their surrounding microenvironment. This is a key component in the differentiation schema, as pluripotent cells will not undergo endothelial specification in the presence of VEGF without first undergoing mesodermal specification [16]. The endoderm serves a key role in EC fate specification and maturation, as it produces key factors, including VEGF and hedgehog (Hh) signaling [16, 44-46]. Further, complex reciprocal signaling between mesoderm and endoderm advances both EC proliferation and maturation. While VEGF is critical to functional vascular development, it is not a requirement for EC specification, as VEGFR-2 (-/-) mice can form angioblasts/ECs, but are embryonically lethal, as they do not form blood islands or the primary vascular plexus, indicating deficiencies in both migration and expansion [47, 48].

A balance between integrin subunit αvβ3 and vitronectin (VN; an ECM protein) signaling, which promotes proliferation of ECs, and integrin α5β1 interactions with another ECM protein fibronectin (FN), inhibits EC proliferation, is required to enable maturation of ECs in the developing vasculature. Deposition of FN between endodermal and mesodermal populations is crucial for maintaining the endoderm and the capacity to produce factors contributing to the promotion of EC fate and remodeling. An imbalance between these cell-ECM interactions can lead to uncontrolled proliferation of ECs or inadequate vascular remodeling [49]. As a result, highly specific cell-ECM interactions through integrins, and the presence of competing signals as dependent on integrin specific binding lead to tight regulation of cell growth, differentiation and morphogenesis. Taken together, the effects of dynamic nature of the developmental ECM manifest through manipulation of an impressively broad range of regulators of cell fate and function, and thus represent an understudied and underutilized driver of differentiation and vascular morphogenesis.

Another example of this type of signal competition exists between VEGFR-1 and VEGFR-

2. The positive regulation of EC fate and proliferation via VEGFR-2/VEGF signaling has previously been discussed. However, in the absence of a competing pathway(s), the balance required for physiological development is lost and a disorganized vascular plexus develops.

VEGFR-1 acts to modulate the proliferative effects of VEGF to guide organization of the vascular plexus [50].

One other guiding principle important to consider is the temporal regulation of lineage specification. The ETS family of transcription factors is known to regulate EC fate, specifically through Etv2 [51]. Mesodermal cells activate Etv2 to promote EC fate, then downregulate Etv2 to

6 allow maturation of ECs [20, 52]. The transient nature of this signaling again highlights the dynamic nature of the signaling pathways governing vascular development.

1.4.2 Aterio-venous fate specification

An initial step in maturation of the vascular system is arterio-venous specification. While initially thought to be exclusively governed by hemodynamic flow, numerous studies have shown that arterio-venous (A/V) specification occurs prior to the onset of flow, as early as in the blood island stage (Fig. 1.1B) [53, 54]. Analyses of the development of intraembryonic vessels, most notably the dorsal aorta and cardinal vein, have revealed that A/V fate appears to be genetically pre-determined with Ephrin B2, an Eph family transmembrane ligand, marking arterial cells, and its receptor EphB4 marking venous cells [55]. Accordingly, these two molecules are required for the progression and organization of the primordial vascular system. VEGF, as well as Notch signaling, including Notch-1 and its receptor Delta-like ligand 4 (Dll4) are upstream regulators of arterial fate, with high levels of VEGF driving an arterial phenotype in human pluripotent stem cells [56, 57]. This signaling axis also acts to inhibit EphB4 and venous fate, while COUP-TFII acts in the opposite manner, inhibiting another upstream regulator of Notch and arterial fate,

Neuropilin-1, to promote Eph4 expression [17]. While genetic predisposition governs initiation of these fate decisions, hemodynamic flow later in development highlights the plasticity and adaptability of developmental blood vessels. Interestingly, vascular identity can be switched in the embryo, as dependent on the microenvironment [58], and altering flow regimes, in particular by disrupting flow in the developing embryo, can disturb arterial differentiation, as well as regulate patterning and morphogenesis of the early vasculature [59, 60]. The plasticity of A/V fate switching is robust early in development, but reduced over time, with changes in fate resulting from alterations in flow patterns [58, 60]. While hemodynamic flow is not a requirement for

7 arterial fate specification, exposure to flow is crucial for maintenance of several arterial markers, suggesting that fate maturation and phenotype maintenance of fate are flow dependent [59]. The role of flow towards guiding A/V fate has been tested in a number of in vitro platforms, with physiological shear stress upregulating markers of arterial fate in endothelial progenitor cells, mouse embryonic stem cells, and human pluripotent stem cell-derived ECs [61-63]. Shear stress can also lead to upregulation of mural cell recruitment factors [64]. Once recruited, these support cells strengthen the arterial wall, thereby altering the local EC microenvironment. These studies are complemented by work describing the effect of matrix stiffness on A/V fate, with murine EC progenitors adopting an arterial fate on stiffer matrices, and adopting a venous fate on softer matrices [65]. In addition, exposure to hypoxic conditions during EC fate specification can upregulate markers of arterial fate [66]. For additional details regarding A/V fate specification and phenotype maintenance, we point the reader to excellent reviews [67-69].

1.5 Expanding and remodeling the vascular system – angiogenesis

Following the formation of the primary vascular plexus by vasculogenesis (Fig. 1.1C), the vascular network expands through complex branching morphogenesis, develops perfusable lumen, and stabilizes through recruitment of mural support cells by way of angiogenesis (Fig. 1.1D). This process relies on genetic regulation to stereotype patterning of initial branches, but is also stringently reliant on environmental changes to optimize O2 and nutrient delivery to increasingly complex and metabolically demanding organs and tissues. It is also during this process that organ and tissue specific vasculature develop in response to tissue-derived signals to meet the specific demands required for optimized functionality. Further details regarding tissue specific vasculature are discussed in [70-73]. Numerous mechanisms for branching morphogenesis and lumen formation are conserved between higher vertebrates and humans, but also in less evolutionarily

8 advanced organisms, such as Drosophila. Next, we focus on the vascular sprouting and branching and the formation of lumen. Stabilization of blood vessel by mural cell interactions and signaling is reviewed here [74].

Figure 1.1: Vertebrate vascular development. (A) During gastrulation, the three germ layers

(ectoderm, mesoderm, and endoderm) develop. Mesodermal specification is guided by soluble factors (e.g. FGF-2, BMP-4, Activin-A, and Wnt), as well as properties of the local environment.

The mesodermal microenvironment is soft, hypoxic, and contains a high concentration of HA.

Integrin binding specificity to the dynamic ECM components is critical to proper development.

(B) Endothelial specification is driven by endodermal signaling through Hh as well as Etv2, leading to the formation of endothelial precursors, which are KDR+ cells that can respond to

VEGF produced by the neighboring endodermal cells. Angioblasts organize in a hypoxic, soft

9 matrix to surround hematopoietic cells in the formation of blood islands. Blood island fusion accompanies A/V specification. Arterial cells are specified by EphrinB2, Notch-1, and Dll4, while venous cells are specified by EphB4 and COUP-TFII. Early A/V specification is plastic and adaptable to changes in hemodynamic flow. High flow leads to a propensity for an arterial phenotype. (C) Further blood island fusion leads to he formation of the primary vascular plexus, which (D) expands, forms lumen, and remodels to establish the functional vascular system.

1.5.1 Tip cell selection and initiation of sprouting

Initiation of sprouting angiogenesis is dependent on the selection of a tip cell from the primary vascular plexus. VEGF gradient signaling guides tip cell formation. Tip cells then lead the migration of new vascular sprouts into the surrounding matrix. Stalk cells trail behind and expand the new vascular branch. Notch signaling determines tip vs stalk cell fate, with Notch ligand Dll4 upregulated in tip cells, which activates Notch in stalk cells to laterally inhibit tip cell specification, which acts to ensure directed migration of the new branch is achieved, without overdevelopment of tip cells and unnecessary branch formation. Tip and stalk specification is an intricate and elegant process defined by constant and reiterative shuffling to ensure persistent tip and stalk elongation in the direction towards the highest concentration of VEGF [18, 75].

1.5.2 Effectors of sprouting and branching

While the Drosophila trachea system is a non-endothelialized epithelial tubular system, it nonetheless carries the burden of O2 delivery to all organs and tissues. In many ways similar to mammalian and vertebrate development, the early development of the tracheal system is highly stereotyped and dependent largely on genetic pre-programming, in particular by Branchless (a homolog of FGF), Breathless (a homolog of FGF receptors), and Pointed (an ETS-domain

10 transcription factor) [76-78]. Three levels of branching exist in the tracheal system, primary, secondary, and terminal branching. While all tracheal cells express generalized tracheal markers that are required to initiate and develop the primary branches, not all tracheal cells branch, indicating the importance of spatial signaling in secondary and terminal branching [76]. The last several decades of research have implicated both neighboring tissue and cell signaling, as well as spatiotemporal presentation of soluble factors and the importance of the surrounding microenvironment. Starting in the 1930s, the role of one critical environmental cue, O2, began to explored as a regulator of terminal branching [79], with more definitive data revealing a specific role for low O2, hypoxia, in regulating the non-stereotyped, or plastic, terminal branching [80].

These low levels of O2 have since been shown to upregulate production of Branchless in neighboring cells. Terminal branches migrate along these O2 and growth factor gradients to provide O2 and nutrients to areas of need (Fig. 1.2A, B) [80, 81]. This process is mechanistically conserved through higher vertebrates with endothelialized vascular systems. O2 gradients stimulate stabilization of hypoxia inducible factor (HIF) to upregulate numerous effectors of angiogenic sprouting, including VEGF [82, 83].

Migration facilitating branching morphogenesis is also regulated by the surrounding ECM.

The organization of ECM proteins can either accelerate branching along organized, aligned fibers

(Fig. 1.2C) or inhibit migration by presenting a dense fibrous matrix composition adjacent to non- sprouting cells [84]. In some cases, remodeling of the tracheal or vascular networks require migration into these inhibitory matrices. In response to signaling cues, including hypoxia, cells upregulate both membrane-bound and soluble proteases, such as matrix metalloproteases (MMPs), to degrade the surrounding matrix and allow for cell migration through the newly void spaces [85-

87] (Fig. 1.2D). Degradation of the matrix must be controlled by inhibitors of these proteases, such

11 as tissue inhibitors of MMP (TIMPs), in order to regulate cell sprouting [87]. In both cases these migratory events are largely controlled by cell-ECM interactions through integrins [78, 88].

Integrins can impart critical information to regulate migration, but also cell proliferation and cell fate. Specificity of integrin binding can be dependent on ECM composition, as different ECM proteins can bind different integrins [88]. At least 9 sets of integrin heterodimers have been linked to angiogenesis [88, 89]. These integrins can promote angiogenesis through distinct pathways guided by different growth factor signaling. For example, integrin αvβ5 promotes angiogenesis via VEGF signaling, while integrin αvβ3 signals through an FGF mediated pathway [90]. Integrin specific binding can also influence vascular patterning and organization [91].

Often directed migration involves cell-ECM interactions as a highway for movement, but gradients of signals are typically required to steer the cells in the right direction. Classically, this type of migration, chemotaxis, has been guided by soluble factor gradients, but can also be regulated by gradients of ECM proteins in a process known as haptotaxis [92, 93]. Additionally,

ECM can be deposited by migrating cells, so the local microenvironment is in constant flux with regard to ECM composition [87]. Taken together, these data provide evidence of the importance of ECM protein concentration, organization, and composition throughout the branching process.

Figure 1.2: Effectors of branching morphogenesis. (A) O2 gradients attract new branches. Cells within low O2 environments increase production of (B) soluble factors, which guide the formation of new branches. (C) ECM organization can facilitate migration of cells to form new branches.

Aligned ECM fibers provide a permissive environment for rapid migrations required to form new branches. (D) Cell-mediated degradation of the surrounding matrix can facilitate migration to generate new branches.

1.6 Lumen formation

The formation of biological tubes requires a complex set of parameters working in concert to mediate both cellular and matrix remodeling. Evolutionary conservation of mechanisms of tubulogenesis between organisms can lead to the construction of lumen diameters between less

13 than 0.1 μm in unicellular capillaries to 20 cm in the gut of the adult Asian elephant [94]. Several mechanisms are employed to generate tubular structures in living organisms, but in vascular tubulogenesis, two main mechanisms, cell hollowing and cord hollowing, are used. While mechanistically distinct, these two methods for the formation of lumen both appear to contribute, sometimes in overlapping fashion, to vascular network maturation.

1.6.1 Cell hollowing

Observations of cell hollowing date back more than 100 years when Florence Sabin described vacuole formation in angioblasts in the developing chick embryo [95]. These observations remained in relative obscurity until the 1980s when endothelial cells were first cultured in vitro and accompanied by the presence of ‘longitudinal vesicles’ that could connect between cells as well as generate unicellular lumen [96]. Confirmation of the relevance of this method for vascular lumen formation was confirmed in three-dimensional cell culture assays alongside details of many key regulators of this process involving integrin-dependent pinocytotic vacuole formation [97]. These vacuoles are highly dynamic structures that change rapidly over time to eventually enlarge and coalesce to form lumen in the process of cell hollowing.

Following the vasculogenic formation of the dorsal aorta and cardinal vein in the developing embryo, sprouting from the vascular cords to form the intersegmental vessels (ISVs) in the zebrafish provides a highly reproducible model to study lumen formation. Analyzing the formation of these vessels revealed the formation of vacuoles, followed by vacuole fusion to generate perfusable lumen in vivo [98]. Interestingly this process does not involve apoptosis, but rather rearrangements and remodeling of endothelial cells, as well as cell proliferation [14]. Lumen formation by cell hollowing can also contribute to tubuologenesis in lower organisms [99]), and this mechanism can be utilized to generate multicellular tubes, as in the previous example, but also

14 can generate unicellular lumen that provide flow through the smallest capillaries of the vascular tree.

1.6.2 Cord hollowing

Vascular cords sprout from the primary vascular plexus and initially do not contain lumen.

These nascent cords are held together by cell-cell junctions. Cord hollowing follows a three-step process, initiated by establishing apicobasal polarity [18]. Shuttling and spatial redistribution of cell-cell junctional proteins, (e.g. VE-cadherin) to the peripheral regions of the vascular cord as a result of basal surface cell-ECM binding via integrin β1, leads to endothelial cell polarization

[100]. A delicate balance between cell-cell and cell-ECM binding is necessary to establish cell polarity and ultimately to form patent lumen. Disruptions of cell-ECM binding or downstream effectors of GTPase signaling, such as Ras interacting protein 1 (Rasip1), severely limits lumen formation [101]. Collecting data form numerous studies indicates that inhibiting integrin binding as a whole can result in more severe phenotypes than can knockdowns of single integrins [101].

This phenomenon is likely due, in part, to the compensatory and overlapping roles of numerous integrin heterodimers. Following polarization, another set of proteins and glycoproteins, termed deadhesive proteins, including CD34 and negatively charged podocalyxin (PODXL), are recruited to the apical region [102]. Electrostatic repulsions enabled by deadhesion proteins initiate separation of adjacent ECs. Finally, to further separate cells, the F-actin cytoskeleton and actomyosin complexes organize in response to VEGF to induce changes in cell shape and generate the force required for further separation of cells and lumen formation [103].

1.7 Human stem-cell derived endothelial cells and perivascular cells

While the bulk of the insightful work highlighted in the previous sections originated through observations in living organisms, the field of stem cell biology has significantly expanded since culture methods were developed for mouse and human embryonic stem cells [104] and have accelerated further since reprogramming pluripotency of somatic cells has become widespread

[105]. Broadly, three differentiation methods have persisted in the literature: (1) pluripotent stem cell (PSC) culture with stromal/feeder cells, (2) embryoid body (EB) formation, and (3) 2D differentiation on protein-coated culture plates with highly specified and temporally regulated growth factor and small molecule media formulations (Fig. 1.3A). A main focus has been on the development of pure, functional endothelial cell (EC) populations to be used in disease modelling or to develop methods to achieve clinically relevant cell numbers for therapeutic purposes. To achieve this level of purity, functionality, and proliferative potential, success has been seen mainly utilizing method (3) above. Studies using this platform have been developed ECs with specified

A/V fate [106], tissue specific function [107-111], hemogenic endothelium [112], EC-mural cell co-differentiation [113-115], and accurate recapitulation of pathophysiology of numerous disease states [116-118]. The current state-of-the-art for differentiation of ECs has aimed to use fully defined media conditions to ensure reproducibility and robustness of EC specification. These protocols have often been developed in response to new findings from the study of embryological development. Many protocols first specify mesodermal populations by activation of canonical Wnt signaling via small molecule GSK inhibitors (e.g. CHIR99021) with pluripotent cells seeded on

ECM proteins (e.g. Matrigel, vitronectin) in 2D culture plates (Fig. 1.3A, B) [111, 114, 119, 120].

It is interesting to note that while Wnt activation alone can induce a robust population of mesodermal cells which can all be differentiated into vascular lineages, patterning of mesodermal

16 subpopulations with varying concentrations of Activin-A/BMP-4 can alter subsequent specification to cardiogenic or hemogenic mesoderm [111, 121].

In any case, mesodermal precursors must be further specified to vascular lineages. A number of different EC differentiation media exist, but many include VEGF, FGF-2, TGFβ inhibitor, often added to commercially available EC growth media to obtain a population of ECs as well as progenitors of mural cells (Fig. 1.3C-E). [111, 113, 114, 119, 120, 122]. Cell sorting can purify these populations with commonly used EC markers (e.g. CD31, KDR) (Fig. 1.3D).

However, heterogeneities exist even in these populations, owing to their immaturity or progenitor status, in other words they maintain a plasticity that contains some variability, highlighting the need for further A/V fate and tissue specification [119, 123, 124]. An interesting technical note here is the critical importance of cell seeding density, which ultimately manifests as a delicate balance between cell-cell and cell-ECM signaling alongside soluble factor availability to have a profound effect on cell survival and fate decisions. The absence of numerous other signaling parameters in these 2D differentiation protocols, including 3D interactions with the surrounding dynamic ECM, as well as extracellular stresses, such O2 and hemodynamic flow, may contribute to the immaturity and heterogeneity of stem cell derived ECs.

While the field has shifted much of its focus to 2D protocols, interesting insights have been gained using EBs. EBs are aggregates of pluripotent stem cells that are classically formed in suspension culture, in methylcellulose semisolid culture, or via hanging drops (Fig. 1.3A) [125-

127]. Initial experiments relied on spontaneous differentiation to achieve the three germ layers, then cells of numerous different tissues, including the vasculature [128, 129]. More refined methods for EC differentiation have since utilized GFs, such as VEGF, to specify EC fate, and complex 3D vascular structures have been observed (Fig. 1.3F) [130-133]. While EBs do

17 recapitulate more aspects of the native embryological microenvironment, including cell-cell interactions, the presence of all 3 germs layers, and three-dimensionality, difficulty in controlling outcomes to the extent which is possible in 2D and low yields of target cell populations, has limited their utility.

Interestingly, EBs share many similarities with organoids, which have seen a recent boom in popularity, as methods to control cell populations with GFs or small molecules in these organoids have become more controllable with insights gained from developmental biology and

2D stem cell protocols. These insights coupled with designer biomaterials that can incorporate numerous environmental cues, such as matrix stiffness, integrin binding specificity, and cell- specific matrix degradation, to inform not just cell fate, but also cell morphology, which is critical in many stages of development, in particular in the formation of functional vascular networks.

Indeed, an understanding of how environmental cues inform cell phenotype and fate represents a nascent sect of stem cell research and is the focus of the following sections.

The above text described methods for derivation of ECs from human PSCs. While not explicitly stated, some of these studies were developed and carried out using embryonic stem cells

(ESCs) and some with induced pluripotent stem cells (iPSCs). Broadly, ESCs and iPSCs share many similarities, including morphology, marker expression, and capacity to form the three germ layers, as classically defined by teratoma formation. However, recent work has illuminated differences between these two cell types, including variations in mRNA [134], miRNA expression

[135], chromatin structure [136], and DNA methylation [137]. Additionally, the somatic cell of origin appears to contribute to some of these differences, with epigenetic memory driving variations in gene expression [138]. These studies have identified evidence for the molecular underpinnings driving differences in differentiation capacity and functional potential of cells

18 derived from ESCs and iPSCs [139, 140]. Better understanding of the regulatory mechanisms guiding differences and developing approaches to more precisely match gene expression profiles between ESCs and iPSCs will be a major step forward towards the translational potential of iPSCs as a therapeutic modality.

1.8 Environmental control of EC differentiation and network assembly

In the previous sections, several key factors of the developmental microenvironment were described, including ECM composition and organization, matrix mechanics, matrix remodeling, integrin binding specificity, cell tension, hypoxia, and hemodynamic flow. Targeting and manipulating these environmental factors, which can facilitate control of a broad spectrum of signaling cascades to regulate vascular morphogenesis, represents a fascinating new arm of research in developmental biology. Engineered systems currently in use in the field of vascular developmental biology will be discussed. Additionally, concepts and platforms developed through studying other tissues will be used to inform potential for novel study of vascular development in vitro, with a particular focus on their utility for studying human stem cells.

1.8.1 ECM organization and composition

Conventional wisdom for many years described the ECM as a simple structural component existing passively between the cells of organs and tissues [141]. The structural role of the ECM imparted by its spatially organized rigidity and elasticity is still critical to tissue homeostasis, tissue organization, and regulation of sprouting and branching, but many other important roles of the

ECM have been identified, including its ability to serve as a highly specific adhesive substrate, control migration by either organizing migratory tracks or enabling cell-mediated degradation to promote specified directional migration, sequester growth factors, serve as a co-factor for

Figure 1.3: Human stem cell derived vascular cells. (A) Three distinct methods are utilized for culture of pluripotent stem cells in preparation for differentiation: defined, directed differentiation of PSCs cultured atop ECM proteins, co-culture of PSCs with stromal/feeder cells, and embryoid

20 body formation. (B) Environmental and soluble signaling factors coax PSCs towards development of mesodermal populations, as identified with markers such as KDR, Brachyury, and MESP1.

Commons soluble factors include Wnt activation via GSK inhibition (e.g. CHIR99021), Activin-

A, and BMP-4. Variations in concentrations of these signaling factors regulate mesodermal subpopulations. (C) EC specification can be directed by both environmental and soluble factors.

Common soluble factors include VEGF, FGF-2, and TGFβ inhibitor within EC growth media. (D)

Cell sorting via FACS or MACS can yield pure populations of ECs (CD31+/VE-cad+). (E)

Without sorting, a mixed population arises comprised of ECs and mural cells, including SM22α+ and PDGFRβ+ cells. Small populations of hematopoietic (CD45+) and mesenchymal stem cells

(CD90+/CD73+/CD44+) may also be present. (E) Complex vascular structures arise from EB differentiation protocols. A low yield of ECs develop along with numerous other cell types.

numerous signaling molecules, and inform signal transduction via dynamic mechanosensory mechanisms [141, 142]. Alterations in signal transduction pathways can manifest as changes in cytoskeletal dynamics to regulate cell shape and morphology, as well as propagate signaling required for mesodermal and EC fate decisions.

1.8.2 2D matrix manipulation

Perhaps the most basic form of in vitro study of cell-ECM interactions is through 2D stem cell culture on ECM-coated cell culture plates. A variety of proteins or combinations of proteins, including Matrigel and Geltrex (combination basement membrane matrices extracted from mouse sarcoma), fibronectin, collagen I, collagen IV, and vitronectin have been used to culture stem cells.

Composition of the ECM is variable throughout development, and has been shown to influence

21 not only maintenance of pluripotency, but also fate decisions [143], as well as morphological changes leading to EC migration, as well branching and sprouting [144]. Microarray patterning of combinatorial ECMs has further elucidated the effects of more complex mixtures of ECM proteins, which aim to recapitulate the makeup of the basement membrane during early development, to determine their effect on cell fate, revealing that combinations, such as collagen IV/heparan sulfate/laminin or collagen IV/gelatin/heparan sulfate, enhance the efficiency of EC differentiation

(Fig. 1.4A) [145]. Other array-based platforms, which have been designed to test the effects of tissue-specific decellularized ECMs, combinations of native ECMs, or libraries of biomaterials could also be used for optimization of in vitro EC fate specification [146-148]. These micropatterning and printing technologies enable analysis of many combinations of proteins in parallel and are thus powerful tools to increase throughput and tease out key proteins of the developmental basement membrane, which contains a complex combination of proteins, to assess their specific role in fate decisions.

Beyond protein composition, protein organization, including ECM fiber alignment and orientation, as well as spatial distribution of proteins, play a role in EC development [84, 142].

Using a related technology, micropatterning of ECM proteins into different patterns and shapes, which can be seen as a 2D stand-in for some aspects of ECM organization, can regulate cell fate independent of chemical (soluble signaling) cues. One specific study showed that pattering of

PSCs facilitated control of cytoskeletal tension and indicated that lower tension in these cells coaxed them down and an EC lineage pathway, with slightly higher tensions resulting in pericyte specification (Fig. 1.4B). Interestingly, the higher efficiency of EC differentiation was accompanied by increased cell-cell contact, suggesting that the delicate balance between cell-cell and cell-ECM interactions can indeed mediate stem cell fate through alterations in cytoskeletal

22 tension [149]. Micropatterning has also been used to manipulate mesenchymal stem cell (MSC) shape through differential cytoskeletal tension, eventually informing stem cell fate decisions [150].

In addition to patterning of the ECM, matrix mechanics can control cell tension. When cells interact with their surrounding microenvironment, they pull on the matrix. This pulling force generates intracellular tension that is dependent on the substrate stiffness, which can serve to transduce mechanical characteristics of the local microenvironment to the cell. Much has been made of this phenomenon, beginning with Engler and Discher’s pioneering work demonstrating

MSC fate specification was dependent on substrate stiffness [151]. Since then, additional work has provided data relevant to EC differentiation as a result of differential matrix stiffness. On compliant (soft) matrices, PSCs showed enhanced mesodermal specification by upregulation of

Yes-associated protein (YAP) and Wnt signaling [31, 32]. Further, the mesodermal populations gathered from this differentiation scheme showed enhanced EC fate specification (Fig. 1.4C) [32].

In addition to governing cell fate, soft matrices also promote enhanced vascular network formation in 2D (Fig. 1.4C) [152].

1.8.3 3D matrix manipulation

Studying cell fate and morphogenesis in vitro in 3D progresses experimental work one step closer to recapitulating the native microenvironment. While work in 3D in the field of stem cell biology is still in relative infancy, insights from many other biological morphogenetic events can help demonstrate the utility of the tools at our disposal to deepen our understanding of differentiation and vascular network assembly.

The groundbreaking work of Bissell demonstrated the importance of ECM context in determination of cell behavior. Providing not just a 3D platform, but the proper ECM composition

23 facilitated functional and structural tissue composition of mammary epithelial cells in culture

[153]. Moving forward, many of these complex phenomena in mammary gland biology, cancer biology, and regenerative biology observed using reconstituted basement membrane (Matrigel) were difficult to recapitulate using alternative proteins or synthetic biomaterials, further supporting the importance of not just dimensionality, but composition and organization. Additional components of contextual importance include matrix viscoelasticity, integrin binding specificity, degradability and degradation products, and growth factor sequestration.

Numerous naturally derived proteins (e.g. collagen, gelatin, fibrin) and engineered synthetic materials (e.g. PEG, HA) have served as the basis for 3D hydrogel matrices to culture cells. Hydrogels are water-swelled polymeric networks that can mimic many physical and mechanical properties of native tissue. Hydrogel precursor solutions are liquid and must be crosslinked to form viscoelastic materials. Methods for crosslinking are diverse, ranging from pH or temperature dependent self-assembly to enzyme-mediated to complex conjugation of multiple moieties to synthetic polymeric backbones. Hydrogels have served as platforms to study stem cell differentiation in 3D. Although not nearly as ubiquitous as 2D differentiation, a handful of papers describe differentiation towards an endothelial lineage, of either isolated cells or EBs, in 3D hydrogel matrices, including PLLA/PLGA [131], HA [154], fibrin [155], and gelatin [156].

Towards studying the effects of mechanical signals on stem cell fate in a 3D context, the matrix sandwich method has been described for differentiation of cardiomyocytes, as well as to study amniogenesis [157, 158] and could also be utilized to study EC differentiation. In this system, cells grown atop a matrix are subsequently covered by an additional matrix to facilitate 3D cell-ECM interactions.

Studies of vascular network assembly in 3D have been much more widespread, and have included culture of mature and stem-cell derived ECs as isolated cells undergoing vasculogenesis

[97], in 2D monolayer culture atop hydrogels to study angiogenic sprouting [159], and on coated beads encapsulated in hydrogel materials to study angiogenesis [160].

Matrix stiffness has become a relatively easy parameter to modulate in biomaterials.

Simply increasing polymer concentrations or crosslinker concentrations can increase stiffness

(Fig. 1.4C). However, in conventional hydrogels, alterations to these parameters change other material properties, such as pore size, which can alter accessibility of nutrients and signaling molecules. To combat this limitation, decoupling of these two parameters has been shown using alginate hydrogels, where the crosslinking is dependent on salt concentration, which reinforces crosslinking sites, while maintaining crosslinking density and pore size [161]. Other hydrogels, both synthetically and naturally-derived have been developed to combat this limitation.

In addition to matrix stiffness, another material property, stress relaxation, has recently been identified as a potent regulator of cell fate and morphogenesis. As discussed previously, when cells bind to their surrounding matrix, they exert a force on the matrix. This force causes a displacement in the ECM. In an elastic material, this deformation is maintained. However, in a viscoelastic material, this displacement is subject to remodeling through stress relaxation. This change in the matrix has a profound effect on how the cell interprets the mechanical cues of the surrounding ECM. Interestingly, many hydrogel materials are covalently crosslinked and, as such, do not exhibit stress relaxation behavior. This is a noticeable mismatch to the behavior of the native

ECM of most tissues [162]. Methods for controlling this behavior precisely in both naturally derived hydrogel materials, as well as synthetic biomaterials are emerging [163, 164].

Differentiation of PSCs in these materials will be interesting to track in the coming years.

Inextricably related to matrix mechanics, the ultrastructure and organization of the ECM can influence cell function, in particular as related to cell locomotion. The most ubiquitous ECM protein, collagen, can serve as an illustrative example. Collagen is comprised of complex hierarchy of self-assembled fibers/fibrils. The organization, density, and alignment of these fibers can influence cell migration [84, 142]. Rapid cell movement can occur along aligned cell fibers and dense fibrillar structures prevent cell migration. Techniques to manipulate these properties have begun to be utilized for studying cell behavior in vitro [165].

Not all tissues contain fibrillar highways for cell trafficking. In many cases, cell migration and morphology are dependent on matrix remodeling. Cell-mediated degradation by production of proteases (e.g. MMPs, ADAMS, cathepsins) facilitates this migration [87]. Naturally derived materials contain inherent capacity for this mechanism of remodeling. However, control and patterning of degradation is difficult in natural matrices in the absence of chemical inhibitors of proteases. In user-controlled synthetic matrices, specific concentrations of sites of degradation can be incorporated into hydrogel design (Fig. 1.4D). Protease sensitive peptides can be conjugated to polymer backbones to facilitate cell-mediated degradation [166, 167]. To control this process further, inhibitors of proteases, such as TIMPs, can be incorporated as a controlled release component [168]. Other signaling molecules, such as angiogenic GFs, can be incorporated in the same way to control temporal regulation. This process mimics a natural function of the ECM, which is to serve as a repository for soluble cues or cryptic binding sites that can be exposed upon degradation [141]. Further, while matrix degradation can be utilized to release supplemented signaling factors, in the native ECM, degraded components of the ECM can act as signaling factors. These molecules, termed matrikines, could potentially be incorporated into polymer design as another method for temporal regulation of cell function. In addition to matrix remodeling, new

26 insights into the effects of cell-secreted ECM have revealed their important role in regulated cell fate [169].

As discussed previously, specificity of integrin binding is vital to physiological developmental fate decisions and vascular morphogenesis. Cells can bind to a variety of ECM proteins through multiple integrins, often with compensatory mechanisms ensuring the capacity to bind is conserved. In synthetic hydrogels, cell binding motifs must be included (Fig. 1.4E). A particular focus has been placed on RGD, a peptide sequence which can bind numerous integrins

[170]. This is often satisfactory for many applications, but choosing a peptide binding sequence should correspond with the integrins associated with the ECM protein of interest for each particular application, and a number of such sequences have been designed, such as laminin associated bioactive epitope IKVAV [171]. Additional control of the concentration of these motifs can alter capacity for and rate of cell migration. Migration is dependent on a bell-curve association of cell adhesivity, with a requirement for a ‘Goldie-locks’ level of adhesion to promote migration [93].

In synthetic systems, the concentration of these binding sites can be decoupled from other ECM properties, so its effect can be better characterized. In natural materials, binding sites are linked specifically to the polymer backbone, so to increase binding sites would require an associated increase in protein concentration, as well as matrix stiffness.

Each of these properties of the ECM have all been studied in isolation. However, many are interrelated with overlapping pathways controlling each. Studying interplay between ECM properties can identify key regulatory mechanisms that will continue to drive our knowledge of cell behavior in 3D systems [172]. To bookend this section, systems that have been developed to provide the proper 3D context, including the organization and composition of the ECM for culture of organoids will be briefly discussed. A popular approach to design materials to study complex

27 structures, such as organoids, is to begin with a hydrogel ‘blank slate’ approach. The most ubiquitous choice here is polyethylene glycol (PEG), which can be readily modified to control matrix stiffness, binding motif concentration, and degradability. Growth factor sequestration and matrikines could also theoretically be included. Optimizing all of these components can result in construction of organoid niches that have the capacity to facilitate both cell differentiation and morphology, resulting in a high level of mimicry to the native developmental microenvironment

[173, 174]. Moving forward, continued development of dynamic, 4D materials that can match compositional changes in the embryo with controllable changes in vitro will likely emerge.

1.8.4 Hypoxia

Hypoxia is often studied in the context of its role as a pathological environmental cue in cancer biology and ischemic disease. However, mammalian development occurs under physiologically hypoxic conditions (~1-5% O2) [175, 176], and a clear indication for its role in vascular development has been made (Fig. 1.4F) [177-179]. HIF-1 can indeed transcriptionally regulate over 200 genes in ECs [180], highlighting its importance as a potent cue for development and vascular assembly. HIF-2 is also an important O2-dependent regulator of transcription in ECs.

Interestingly, the roles of HIF-1 and HIF-2 have both overlapping and divergent roles, further highlighting the specificity with which cells can sense and respond to changes in environmental

O2 [181, 182]. Less attention has been paid to HIF-independent pathways regulated by mammalian target of rapamycin (mTOR) [183] and production of reactive oxygen species (ROS) [184], both of which are sensitive to changes in O2. Regulation of these signaling cascades is complex, and crosstalk between O2 sensitive signaling pathways can lead to regulation of NF-κB signaling [185], as well as influence both Notch and Wnt/β-catenin pathways to regulate differentiation and morphogenesis [83, 186].

Traditional experimental platforms to modulate environmental O2 have relied on mixed gases and hypoxia chambers or incubators to culture cells (Fig. 1.4Fi). Great strides have been made utilizing this setup to uncover hypoxia-induced changes in gene expression and cell fate. In particular, low O2 has been shown to enhance EC fate decisions in 2D [66], as well as promote enhanced vascular network formation in 3D [187]. Recent advances in microfluidic devices have facilitated culture of cells in more physiological conditions, including the inclusion of O2 gradients

(Fig. 1.4Fi) [188-190]. 3D culture of cells within hypoxic gradients has also been facilitated by physical orientation of hydrogels, via rolling a composite biomaterial to generate layers and an O2 gradient (Fig. 1.4Fii) [191]. Another hydrogel platform has been developed to generate a hypoxic gradient through enzymatic crosslinking (Fig. 1.4Fiii) [192]. This hydrogel has been used to study the effects of hypoxia on vascular network formation in 3D [193, 194], as well as to uncover novel mechanisms for blood vessel formation that closely recapitulate developmental vasculogenesis

[195]. A further discussion of the importance of developing gradients in biomaterials systems is provided later in this manuscript.

As mentioned, low O2 can induce production of ROS. In fact, ROS has been shown to be a key regulator of EC fate and EC network formation [66, 187, 195]. Hydrogels have been recently developed to produce ROS (H2O2) to enhance EC proliferation and improve the formation of new blood vessels in vivo [196]. This system provides an interesting spring-board for discussion. We have described hypoxia as a potent regulator of many signaling pathways. Targeting specific components of these pathways (e.g. HIFs, ROS, mTOR) has advanced our understanding of EC behavior. However, the interplay between these pathways, and the overlapping functions they elicit may provide insight into limitations for drugs targeting single or even multiple factors in these pathways. As is the case with perturbing factors related to the ECM, the presence of compensatory

29 mechanisms limits the potency of these therapeutics, further highlighting the need for control of the most upstream regulators in the cell or tissue’s local microenvironment.

1.8.5 Hemodynamic Flow

The importance of hemodynamic flow in determination of A/V fate has been previously discussed. However, hemodynamic flow can regulate further maturation of blood vessels and is an important factor in several common pathological conditions [197]. Studying the effects of hemodynamic flow in vitro has advanced rapidly with microfluidic device design to manipulate flow patterns and shear stress (Fig. 1.4G). In 2D systems these flow patterns are relatively easy to control, and the effect of shear stress on EC biology has been studied extensively [198]. Many methods for transduction of flow induced signaling have been studied, highlighting the importance for the capacity of ECs to sense these changes in development as well as homeostasis [199]. In particular, variations in stem cell derived EC’s ability to sense these changes can significantly impact their morphology, including alignment to shear and coordinated migration [200]. These findings illustrate the importance in analyzing a full gamut of functionality tests for stem cell derived ECs, as traditional methods for functional maturity have included formation of 3D networks, which is still permitted in ECs lacking the capacity to sense changes in shear stress.

Studying the effects of shear stress in 2D can serve as an in vitro stand-in for the effects of shear on larger diameter vessels, where the effects of curvature are minimized. These studies are important in identifying novel signaling pathways related to hemodynamic flow. However, in smaller diameter vessels, hemodynamic flow has been more difficult to manipulate in vitro. New platforms to test these effects have been developed for small diameter engineered vessels [201] and perfusable networks based on vasculogenic network formation [202].

Figure 1.4: Environmental regulators of EC fate and platforms/methods to incorporate them in vitro. (A) Micropatterning of specific combinations of ECM components in an array can be used to aid in the identification of the impact of the organization and composition of both natural and synthetic biomaterials on EC fate and morphology in a high throughput manner. (B)

Micropatterning can be used to alter cell shape and size to control cytoskeletal tension, which can lead to manipulation of cell fate decisions. (C) Culture of cells with ECMs of different matrix stiffness can alter both cell fate and vascular network formation. (D) Incorporation of protease

(MMP) sensitive moieties (e.g. GCRDGPQGI↓WGQDRCG) can facilitate cell-mediated degradation and migration to promote vascular network formation. (E) Conjugation of integrin specific binding motifs (e.g. IKVAV and RGD, laminin and fibronectin mimetic peptides,

31 respectively) can alter cell fate and vascular network organization. (F) Low O2 can regulate gene expression through HIFs, ROS, and mTOR. (i) Cells can be subjected to low O2 by using mixed gases and hypoxia chambers or microfluidic devices. (ii, iii) Rollable scaffolds and O2-controllable hydrogels can be used to culture cells in hypoxic gradients. (G) Hemodynamic flow can be imparted using microfluidics to influence cell fate and maturity.

1.8.6 Gradients

Several examples previously described have posited the presence and importance of gradients in regulating cell behavior. Over a half century ago, Turing hypothesized that signaling molecules, which he termed morphogens, could establish a graded distribution to elicit varied cellular response [203]. The concept of positional morphogenesis was later theorized to inform a cell’s concept of its location within a morphogen gradient as a method to pattern tissue development [204]. These concepts laid the groundwork for study in tissue patterning and identification of morphogen gradients in development. While these ideas have been identified using model organisms and have studied the effect of gradients in mesodermal induction, the importance of gradients in EC development is an understudied and emerging field of study.

Cells have established sensitive machinery to interpret small changes, as small as 2-3 fold, in morphogen concentration to alter levels of expression of one or more genes and regulate cell fate [205-207]. An illustrative example of this concept has been observed with a morphogen gradient of nodal to facilitate differentiation of the three germ layers during development. Cells close to the source of nodal, the vegetal pole, have an increased number of surface receptors occupied, which leads to elevated transcription to coax the cells to form the endoderm.

Intermediate levels of nodal receptor occupancy lead to formation of mesoderm; the absence of nodal results in ectodermal specification [208]. While the three germ layers are discrete, a

32 continuous morphogen gradient suggests differentiation populations within each germ layer (Fig.

1.5). This has been further tested in vitro to pattern specific subpopulations of mesoderm by subjecting cells to different levels of Activin-A/BMP-4 and Wnt/BMP-4, which resulted in variable enhancements in definitive subpopulations in accordance with highly specific patterning

[121, 209]. Methods to pattern morphogens using diffusion have been studied [210, 211], but with advances in the resolution and biocompatibility of 3D printing and micropatterning, the specific effects of gradients of a variety of morphogens identified during model organism development will likely be studied to further our understanding of differentiation of human stem cells towards EC populations, perhaps further delineating how tissue-specific and functionally mature ECs arise.

While patterning signaling molecules in these 2D systems has recapitulated complex patterning in the developing embryo, simple diffusion does not tell the complete story of regulation of morphogen gradients. Some of these gradients are indeed regulated primarily by simple diffusion, but components of the ECM can differentially bind to morphogens to change gradient shapes via restricted diffusion [212-214].

In addition to soluble proteins, molecular O2 and stiffness can be can precisely control cell behavior [215, 216]. Thus, establishing methods to control these two types of gradients can provide extensive control over tissue and developmental patterning. O2 gradients can be controlled in both microfluidic and 3D hydrogel systems [188-192]. Methods to establish matrix properties have begun to be explored [216, 217] and will likely continue to advance into 3D systems to more precisely and profoundly control cell fate.

Figure 1.5: Effect of gradients on gene expression, cell fate, and cell phenotype. Cells can sense gradients in matrix stiffness (G’), local O2, as well as morphogen concentrations. These three parameters can influence receptor occupancy on cells within a specified location along the gradient to precisely regulate gene expression, cell fate decisions, and cell phenotype. The cell’s ability to sense small changes informs its positional awareness, which is critical to proper tissue patterning and development.

1.9 Future Perspectives and Conclusions

A common theme of this manuscript has been learning from the past. Often observations and conclusions developed in the absence of requisite technology or parallel data limit their utility and reach. Re-examining these phenomena in a new light has led to some of the most important insights in the stem cell field to date. Moving forward our increasing ability to design spatiotemporal cues into biomaterials in vitro will propel the work in this field to new heights. 4D and dynamic materials which can regulate cell-mediated release of signaling molecules, alter

34 matrix stiffness, matrix organization and composition, ECM deposition, and O2 will continue to be refined to more closely mimic the embryological milieu. Continually referencing the expanding literature regarding the dynamics of these properties in model organisms will inform biomaterials design.

With an impressive library of biomaterials that can be designed to manipulate a remarkable number of properties with inherent instructive capacity, careful consideration must be taken into which of these properties are sufficient to elicit the desired effect. Two methods for designing and utilizing biomaterials platforms can be employed. (1) A case can be made for simply giving stem cells an environment amenable to their capacity to self-organize. Learning which cues are needed in this theoretical system will continue to be explored. (2) On the other end of the spectrum, tight control over multiple cues can specifically guide cells to behave in a user-defined way, largely circumventing their capacity for self-assembly. These types of materials will need to employ dynamic characteristics. Perhaps more information can be gained from a biological perspective using method (1), including identifying novel therapeutic targets and signaling pathways using human stem cell-derived ECs, and even patient specific cells, in vitro for personalized medicine or disease modeling. Method (2) can establish reproducible approaches for building human tissues that, ultimately, may be used in the clinic. Feedback loops between these two methods will drive our knowledge to accurately recapitulate the developmental microenvironment.

CHAPTER 2 2 Acellular Implantable and Injectable Hydrogels for Vascular Regeneration 2.1 Teaser

A library of biomaterials has been developed for delivery of angiogenic therapeutics.

Incorporating multiple factors and spatiotemporal delivery will enhance the efficacy of these therapeutics.

2.2 Introduction

A plethora of vascular diseases and disorders affect millions of people each year. These diseases span from minor disorders with mild symptoms (such as Raynaud’s disease) to severe ones with significant reduction in quality of life (chronic wounds, peripheral artery disease) and mortality (coronary artery disease). Although the risk factors for these disorders vary widely, there are two main mechanisms for healing and regeneration of vasculature: vasculogenesis and angiogenesis. These complex processes are vital throughout life from embryonic development through adulthood and play important roles, both positive and negative, in a number of diseases states including wound healing, tissue ischemia, and cancer.

Vasculogenesis, the formation of de novo blood vessels, largely occurs during embryogenesis, but may also occur in the adult through accumulation and network formation of endothelial progenitor cells (EPCs) and subsequent differentiation to the vascular lineage through growth factor gradients and cues from the existing extracellular matrix [12, 21, 218]. In order to form networks these EPCs follow a specific morphogenesis process, as outlined in Fig. 2.1. First, vacuoles form within ECs as a result of cues from the microenvironment. Those ECs with vacuoles then merge with neighboring cells to form open lumen structures. Tubulogenesis, in which the ECs branch and sprout, then occurs. Finally, these new networks are stabilized and become functional through attachment of mural cells [98, 219-222].

A much more common process, and a process that has become increasingly understood recently, is angiogenesis. Angiogenesis is formation of new blood vessels by way of sprouting from existing vessels [12, 21, 218] (Fig. 2.1). The life of an endothelial cell (EC) as it matures into part of a mature, functional vessel passes through three main stages: quiescence, activation, and branching [21]. Quiescent ECs survive by way of maintenance signals, including vascular endothelial growth factor (VEGF) [223], NOTCH [224], angiopoietin-1 (ANG-1) [225, 226], and fibroblast growth factors (FGFs) [227]. They are interconnected through junctional molecules.

Supporting cells, such as pericytes, suppress differentiation potential and contribute to cell survival through release of additional VEGF and ANG-1 [21]. ECs and supporting cells in the quiescent state form and attach to a shared basement membrane, composed of a variety of proteins and proteoglycans including collagen IV, fibrin, and laminin [228]. Components of this basement membrane have additional functions once the ECs are activated.

Figure 2.1: Treatment of blood vessel occlusion by injection or implantation of hydrogels.

(A) Lower limb artery occlusion. Upon hydrogel injection (i) or implantation (ii), the occluded blood vessels are surrounded by pro-angiogenic, crosslinked (shown with dashed lines), growth factor bound (ovals) hydrogel. These hydrogels promote both (B) Vasculogenesis and (C)

Angiogenesis (D) Clinical success of hydrogel therapy results in a degraded hydrogel with blood flow restored by functional, stable neovascular networks. B is adapted from [219] ; C is adapted from [229] and influenced by [21]

Activation of ECs and pericytes occurs when a resting state vessel senses an angiogenic signal. These ‘help’ signals may include VEGF, ANG-2, or FGFs, which are released by inflammatory cells and cells which encounter specific microenvironmental conditions, including hypoxia [21]. Malignant cells may release these factors, as the tumor microenvironment is often hypoxic [21]. Additionally, a hypoxic microenvironment may be met by non-cancerous cells following injury or various pathological conditions including both peripheral and coronary artery disease [230]. When pericytes sense ANG-2, they react by detaching from the vessel wall [225] and are released from the basement membrane through proteolytic degradation, largely mediated by matrix metalloproteinases (MMPs). In response to the angiogenic signals, ECs loosen their junctions thereby increasing the permeability of the EC layer [21]. New ECM is deposited where there is increased permeability in the EC layer, which attracts ECs to form a new vessel branch

[21]. Proteases reveal additional angiogenic molecules stored within the new ECM [231, 232] including VEGF and FGF, which help remodel the new ECM into a pro-vascular microenvironment. In order to allow perfusion and control growth of the new branch, a tip cell leads the way along the growth factor (VEGF [233], NOTCH [224, 234], etc) gradient. Cells

38 connected to the tip cell are termed stalk cells. These cells divide to increase the branch size and form a lumen in response to a number of molecules including NOTCH [224], placental growth factor (PlGF) [235] and FGFs [227]. Myeloid bridge cells connect new branches with pre-existing portions of the vascular network, enabling perfusion. As blood flow begins, ECs return to their resting state and become stabilized by pericytes in response to platelet-derived growth factor B

(PDGF-B) [74, 236, 237], ANG-1 [74, 225, 236], transforming growth factor-β (TGF- β) [74, 236], ephrin-B2 [237], and NOTCH [74, 224]. Basement membrane is then deposited to further stabilize the vessel and junctions between endothelial cells are re-established [21].

As the mechanisms by which these two processes work have been revealed over the last

20 years, new treatment methods have been developed to promote or enhance aspects of each pathway. An emerging technology that has received much attention in recent years is utilization of hydrogel technologies for regeneration of vasculature and ultimately healing of a range of vascular diseases and disorders. Hydrogels are an ideal choice for this type of clinical application because of their inherent similarities to the native ECM [238], which is critical to both angiogenesis and vasculogenesis as described above. Hydrogels may be composed of either synthetic or natural polymer backbones. As hydrogel technologies have evolved, a large library of potential materials has been developed, leading to many possibilities for development of hydrogels for highly specific applications [239, 240]. Many properties of these materials can be controlled including mechanical stiffness, degradability, ultrastructure, and cross-linking density. Functionalization of these materials is largely dependent on material choice; however, many materials allow addition of peptide sequences (e.g. for enhanced degradation or cell adhesion) or conjugation of bio-active growth factors and cytokines, with control over release kinetics. Hydrogels have been used to gain an understanding of the mechanism by which vascular regeneration occurs and for bio-mimicking

39 different aspects in the vascular niche to activate the desired signaling pathway. These hydrogels, specifically acellular implantable and injectable hydrogel systems, will be discussed in detail in this review. For a review of cell-based therapies and pre-vascularized constructs we point the reader to [229, 241-243].

2.3 Incorporating biological factors in hydrogels to modulate vasculogenesis and angiogenesis

Much of the work done in developing hydrogels for vascular regeneration has been done by addition of biological factors to enhance re-vascularization and make new vessels more robust.

This work has been made possible by the elucidation of the mechanisms of both angiogenesis and vasculogenesis. Limited clinical success has been found by bolus injection or systemic delivery of growth factors [244]. Localized or targeted therapies are necessary for optimal delivery because of the rapid clearance of proteins upon systemic injection [245]. Localized and highly controlled release is also vital, as excess growth factor delivery may also be detrimental, leading to vascular leakage and hypotension, as well as potential tumor formation [246-248]. Additional unwanted consequences of angiogenic growth factor delivery include unwanted angiogenesis at non-targeted sites, such as the retina [246-248]. Sustained levels of VEGF are required, as transient signaling from VEGF leads to new vessel regression [249]. Precise spatial and temporal delivery are also necessary for effective therapeutic angiogenesis. Physiologic angiogenesis is insufficient to fully heal various clinical presentations including chronic wounds and tissue ischemia. Therefore, additional therapeutic angiogenesis is necessary for a positive therapeutic effect [250]. Many hydrogel technologies have employed biological factors to establish therapeutic angiogenesis strategies. These hydrogels are able to more precisely control delivery of growth factors to the target region and sustain release to elicit a robust response.

2.3.1 VEGF

VEGF is a vital component of many of the steps of angiogenesis, as described above. With its importance recognized, many of the early attempts to create hydrogels incorporating growth factors for vascular regeneration, involved encapsulation or conjugation of VEGF. In fact, as the complexity of hydrogels has increased and combination conjugations attempted, VEGF remains a primary component.

An important hydrogel system was developed in 2001 as one of the early attempts in regulating growth factor release for controlled and localized delivery. This was a relatively simple system of covalent incorporation of VEGF into fibrin-based hydrogels. The choice of including

VEGF is clear, but the choice of fibrin hydrogels was made because fibrin is a clinically acceptable material, a natural component of the endothelial cell basement membrane, to which VEGF may be bound, and can be derived from a patient’s own blood, thus avoiding the potentially severe inflammatory and immune responses elicited by xenogenic or allogenic hydrogel materials, such as gelatin and collagen [251]. Importantly, the rate of release of VEGF was determined by the degradation of fibrin by invading cells upon injection in vivo. Because the rate of degradation was largely cell-based and thus localized, a low but continuous level of VEGF release could be achieved. Additionally, a dose dependent response could be achieved by simply increasing the amount of VEGF conjugated in the hydrogel. Another important design parameter this group included was both bound and unbound VEGF. The unbound VEGF could act as a recruiter of ECs, and bring them toward the bound VEGF to mature and form functional vascular networks [251].

As a vital step to development of more advanced hydrogel systems, the work described above laid the ground work for a rapidly expanding field.

Progression has been significant, as recapitulation of the complex interactions that take place during angiogenesis have been attempted. The use of both synthetic and natural polymer backbones for hydrogels has provided many unique solutions. For example, the use of a polyethylene glycol (PEG) backbone with incorporated cell adhesion (RGD) sites, MMP substrates, and VEGF aimed to mimic the cell-ECM interactions presented during angiogenic revascularization. Cells attached to RGD sites and as cell MMPs degraded the injected hydrogels and cell associated proteolysis released bound VEGF, remodeling allowed for 3D cell migration and network formation. Again, cell-mediated release of VEGF resulted in controlled, local release of VEGF, eliminating many of the negative effects of systemic growth factor release [246].

More recent studies have also sought to control degradation of natural hydrogels to precisely control release of VEGF. Recently, fibrin matrices bound with a fibrinolysis inhibitor were able to control and sustain release of VEGF to form functional vessels for up to 3 months after injection. The fibrin backbone allows cell-mediated hydrogel degradation and thus control of release based on tissue need, as cell infiltration is more likely when resident cells have already begun signaling to attract angiogenic cells [250].

Another study that aimed to control the release of VEGF and maintain prolonged local release involved use of alginate-based hydrogels. Alginate is a natural polymer that has been used in a number of applications and can be easily tuned to obtain desired properties. By partial oxidation and control of molecular weight distribution and because aligante can reversibly bind to

VEGF, controlled release of VEGF was obtained. Alginate also increased the bioactivity of VEGF

[249].

Other more complex technologies have also been developed to enhance therapeutic angiogenesis. Hydrogel systems in combination with microspheres containing VEGF were shown

42 to enhance angiogenesis through sustained and controlled release of VEGF. Hydrogel systems alone weren’t as effective because of less control of drug release, while microspheres alone were subject to phagocytosis by macrophages. Thus, the combination system provided protection for the microspheres and sustained VEGF release [252, 253]. Another hydrogel system utilized microparticles containing VEGF with high affinity for cross-linking pyrrole molecules in alginate hydrogels. Because the microparticles were bound to cross-linking molecules, a truly localized release was achieved and this release was sustained over relatively long time periods [254].

Similarly, encapsulation of nanoparticles within 3D hydrogel matrices, allowed for an increased protection against VEGF degradation and controlled, sustained release. VEGF was encapsulated in nanoparticles by a high efficiency complex formation with dextran sulfate and chitosan. Dextran sulfate complexed with the binding site of VEGF, protecting it from degradation and contributing to sustained release, while chitosan stabilized the complex and also contributed to protection and controlled release of VEGF [255].

The controlled release, local release, and sustained release achieved by these systems led to improvement of therapeutic angiogenesis by preventing unwanted systemic effects and bolstering the angiogenic response to form mature, functional vessels. The neovasculature was able to maintain functionality and stability even when the pool of growth factors had been exhausted, indicating the potential for translation of this technology to clinical applications [250].

2.2 bFGF

Basic fibroblast growth factor (bFGF) plays an important role in angiogenesis. It has been a target of study, because of its influence on therapeutic angiogenesis; however, many of the same issues that arise with clinical delivery of VEGF, exist with bFGF. This being said, the approaches

43 for delivery of bFGF encapsulated in or conjugated to hydrogels have significant overlap with those for delivery of VEGF. Highlighted here are several unique hydrogel systems.

Similar to several of the strategies described above, much work has been done to slow the release rate and extend the release time of therapeutic bFGF. As discussed above, release of bound protein from a fibrin hydrogel is mediated by fibrin degradation, mostly due to cell infiltration mediated proteolysis. Once released from the bound state, bFGF release rate is mediated by diffusion; in hydrogels with higher cross-linking density, release rate is slowed. These facts were all utilized in the design of a controlled-release fibrin hydrogel. First, because bFGF binds to heparin, addition of heparin to fibrin hydrogels slows the release rate. Addition of thrombin and fibrinogen increases the cross-linking density of the hydrogel and slows the degradation rate and diffusion, so the release rate is further reduced. With a decreased release rate, more mature vessels form upon injection [256]. To further improve the desired release rate and further increase the release period of bFGF, heparin-conjugated poly(L-lactide-co-glycolide) (PLGA) nanospheres were suspended in fibrin hydrogels. Utilizing these nanospheres as well as optimizing fibrinogen concentration (as discussed previously) allowed controlled zero order release of bFGF for up to one month. Such long release periods significantly enhance the angiogenic effects of these injectable hydrogels [257]. Another system for prolonged release of bFGF was peptide-amphiphile

(PA) based hydrogels. PAs are able to self-assemble into nanofibers with similar properties to the fibrous proteins of the natural ECM. These nanofibers are able to form hydrogels in vivo, and by incorporating bFGF, are able to improve revascularization through prolonged growth factor release

[258].

2.3.3 Other growth factors/cytokines

Several other pro-angiogenic growth factors and cytokines have been encapsulated within hydrogels to improve therapeutic angiogenesis. These factors have been less studied, but remain important in angiogenesis, so their study is still vital. PDGF-BB is not involved in the initial steps of angiogenesis, but is an important recruiter of vessel stabilizing pericytes and increases VEGF expression as well as cell migration collagenases, which enable EC migration for further vessel growth. By conjugating PDGF-BB to a synthetic polymer, PEG, growth factor stability and solubility were increased, while the immune response was decreased, owing to the inert nature of

PEG. Incorporation of both soluble and PEG-immobilized PDGF-BB increased angiogenesis in vivo. Another advantage presented here was the synthetic biomaterial “blank slate” approach. In other words, the synthetic material can be easily tuned and its properties accurately reproduced, while preventing unwanted interactions in vivo [259]. As an aside, a similar PEG system was developed with incorporation of cell adhesion and proteolytic peptides, as well as VEGF. Growth factor was sequestered and thus allowed for sustained delivery of growth factors during remodeling of the hydrogel, which enhanced angiogenesis [260].

Stromal derived factor-1 (SDF-1) helps recruit endothelial progenitor cells and pro- angionenic bone marrow derived cells to aid in angiogenic healing [21]. Immobilized SDF-1 in gelatin-based hydrogels is released as the gelatin is enzymatically degraded. This again presents a situation in which release is dependent on cell invasion and subsequent degradation and remodeling of the hydrogel. In the case of SDF-1, pro-angiogenic cells are directly recruited to drive angiogenesis [261]. Again, the elusive sustained, controlled release was sought, and the immobilization of SDF-1 in starPEG-heparin hydrogels help ameliorate and sustain release rates,

45 thus improving the therapeutic effect. These hydrogels created sustainable concentration gradients of SDF-1 which aided in recruitment of essential endothelial progenitor cells [262].

Eph receptors and their ephrin ligands are vital to angiogenesis by mediating cell adhesion and by contributing to vascular assembly and remodeling. Immobilization of ephrin-A1 within poly(ethylene glycol)-diacrylate (PEGDA) hydrogels contributed to the predicted enhancement of cell adhesion and spreading. These effects occurred in a prolonged, dose-dependent manner, and ultimately led to an augmented angiogenic response [263]. Further, covalently immobilized ephrin-A1 conjugated to PEG in PEGDA hydrogels was able to regulate deposition of important

ECM proteins (collagen IV and laminin) in a dose-dependent manner as well as contribute to angiogenic cell adhesion and remodeling, thereby improving upon the previously described ephrin-A1 PEGDA hydrogels [264].

Sphingosine-1-phosphate (S1P), a lipid factor, regulates endothelial barrier function and contributes to angiogenic vessel stabilization [21, 265]. Importantly, S1P is involved throughout the angiogenic process, being involved in promoting EC migration and proliferation and stabilizing newly created vessels. Cross-linking PEG-vinyl sulfone with albumin, allowed for controlled release and delivery of S1P. Albumin was used as a carrier for S1P, as a protein carrier is necessary due to the highly hydrophobic nature of lipid factors such as S1P. This hydrogel system was able to promote EC migration and angiogenesis [265].

2.3.4 Other biological factors

In addition to the use of the many of the important angiogenic growth factors/cytokines, several other important biological mediators of angiogenesis have been utilized in conjunction with hydrogel technologies.

Rather than using natural growth factors and cytokines, peptide mimetic versions of growth factors have been studied for their regenerative capabilities. Advantages of peptides over their counterpart proteins include lower cost and enhanced stability. One specific example of this technology with potential for therapeutic vascular regeneration is the use of a 15 amino acid VEGF mimetic peptide (QK). QK is able to bind to VEGF receptor 2 and elicit angiogenic activity.

Immobilization of QK to mixing-induced two-component hydrogels shows sustained release and prolonged angiogenic activity upon injection [266].Another method that has been utilized to provide sustained local delivery without unwanted systemic side effects that may arise with administration of growth factors is DNA plasmid delivery. Plasmid DNA has inherent bioactivity longer than that of many growth factors. Prolonged delivery may be achieved by condensing DNA with cationic polymers in order to decrease the size of DNA particles, protect from DNA degradation, and promote interactions with cells in order to make transfection more effective. DNA may also be encapsulated in microspheres or nanospheres to protect the DNA and prolong its release. Genes encoding for pro-angiogenic factors, such as VEGF, bFGF, and PDGF-BB may be delivered from collagen-gelatin hydrogels [267]. Delivery of plasmid DNA may also be regulated by degradation of hydrogels. Specifically, alginate hydrogels with degradable cross-linking segments may be used to control local delivery of plasmid DNA encoding for VEGF and improve re-vascularization in vivo [268]. Macroporous PEG hydrogels were able to deliver lentivirus encoding for VEGF and effectively form vascular networks within the pores. Slow-degrading PEG hydrogels were used, as rapid degradation may lead to transient transfection and maintenance of hydrogel stability may help unsupported ECs maintain functional phenotypes [269]. Similarly, hyaluronic acid hydrogels (discussed below) have also been utilized to delivery non-viral pVEGF

47 by use of poly(ethylene imine). Delivery could be controlled by both initial pore size and degradation [270].

A class of synthetic hydrogels with inherent wound healing and angiogenic properties was developed a decade ago and has been modified and refined since then. Hyaluronic acid (HA) hydrogels are biodegradable with controllable degradation rates, easy to modify (contain reactive groups), may serve as protein delivery constructs, and are naturally non-immunogenic and non- adhesive [271, 272]. Incorporating heparin binding of growth factors in HA hydrogels regulates and prolongs release rate of growth factors [273]. HA hydrogels may be used for therapeutic angiogenesis and vasculogenesis because of HA regulation of stimulation of cytokines and EC proliferation. Peptides for cell adhesion (RGD) and degradation (MMP sensitive peptides) allow control of adhesion to and degradation of HA hydrogels for EC morphogenesis into functional vasculature. By controlling each step in vascular morphogenesis, more robust networks can be obtained. Integrin mediated adhesion contributed to vacuole and lumen formation, while degradation enabled branching and sprouting as vascular structures matured [219]. Further spatial organization of both pro and anti degradation cues provide a more robust system for organizing neovascular networks [274].

Another new class of hydrogels has been developed recently. Peptide-based hydrogels provide more precise control over hydrogel properties than traditional hydrogel materials

(discussed throughout the manuscript). The use of peptide-based materials has been studied over the last decade, but has only been optimized for use as hydrogels recently. Specifically, a self- assembling leucine zipper based hydrogel has been developed for use in tissue engineering and revascularization therapies. These hydrogels exhibited highly tunable mechanical properties and pore size, simple conjugation of important cell-active peptides (such as RGD), did not elicit a

48 foreign body response, and were able to support dynamic cellular interaction including promoting vascularization by inducing secretion of pro-angiogenic ECM proteins without the use of additional bound growth factors or cytokines [275].

Most of the hydrogel systems discussed thus far aim to promote production or increase local levels of biological factors. However, because angiogenesis is such a complex process, inhibition of certain factors has also been shown to improve vascular repair and remodeling.

Degradation of the ECM by MMPs is a vital aspect to vascular remodeling. However, in many disease states, this process is rampant. Excessive levels of MMPs persist and may lead to continued disease progression. Tissue inhibitors of MMPs (TIMPs) regulate MMP activity and help maintain tissue homoeostasis. Systemic administration of TIMPs encountered many of the same problems that plagued systemic injection of growth factors. In order to control TIMP release in a manner which is able to control remodeling, an MMP-degradable HA hydrogel capable of polysaccharide conjugation (to mimic TIMP-ECM binding) has been developed to control release of TIMPs by local MMP activity. TIMPs are encapsulated within hydrogels through electrostatic interactions, allowing their release upon cross-link degradation by MMPs. This hydrogel system was able to effectively reduce adverse modeling in a porcine myocardial infarction model [168]. Similarly, this hydrogel could be applied to other vascular disorders in which excessive levels of MMPs must be controlled.

2.3.5 Combinations of biological factors for angiogenesis/vasculogenesis

Although potential positive therapeutic effects are clear by utilization of many of the hydrogel systems described above in vitro and in animal disease models, in order to form even more robust vascular networks capable of maintaining long-term stability and effective treatment of vascular disease, design of hydrogel systems with complementary and synergistic biological

49 factors may prove paramount. While the design of hydrogels that maintain prolonged release rate, local delivery, and other optimal design parameters for vascular morphogenesis should always be considered and were discussed in sections 2.3.1-2.3.4, here we focus on the novelty and efficacy of combining biological factors, which aim to establish networks that accommodate for the complexity of angiogenic and vasculogenic pathways.

A number of combinations of growth factors with other biologically active molecules conjugated or encapsulated in hydrogels, both natural and synthetic have been tested and synergistic effects have been elucidated. For example a simple combination of VEGF with HA (an

ECM component with a range of physiologic functions including important angiogenic functions described in section 2.4) hydrogels produced significantly more vessel formation than just VEGF or just HA, indicating a synergistic effect [276].

Study of the effects of controlled release of many combinations of growth factors has been shown to enhance angiogenesis. Most of these hydrogel systems combine an angiogenic initiator

(such as VEGF) and a growth factor that helps stabilize new blood vessels by recruitment of mural cells (PDGF-BB) or vessel maturation (KGF). A summary of some of these combinations can be seen in Table 2.1.

The many technologies listed in Table 2.1 work through simultaneous (or nearly simultaneous) release of each growth factor. Because often timing of growth factor release and uptake is vital to angiogenesis, as angiogenesis is a complex, multi-step process, controlling the release of multiple growth factors may facilitate closer mimicry of physiological angiogenesis.

Utilizing differential heparin binding affinities of growth factors has hinted at temporal release, as independent control of binding of multiple growth factors has been achieved [277, 278]. Further, using the heparin binding domain of a major ECM component (fibrinogen), a range of affinities

Table 2.1: Combinations of Growth Factors for Therapeutic Angiogenesis Hydrogel material Biological factor(s) Results References HA VEGF, Ang-1 Enhanced angiogenic [279] response by combinatorial approach HA VEGF, KGF Combination of [280] ubiquitous growth factor (VEGF) and late stage growth factor (KGF) for long term vessel maintenance HA/heparin/gelatin VEGF, KGF; VEGF, Improved growth [281] PDGF-BB factor retention and release kinetics through heparin binding Dextran/PEGDA VEGF, IGF, SDF-1, Controlled hydrogel [282] Ang-1 properties with conjugation of multiple growth factors leads to rapid, efficient, functional vascular networks Alginate + PLGA VEGF, Ang-1 Encapsulation of [283] microspheres GFs within microspheres led to sustained delivery for enhanced angiogenesis Alginate microparticles VEGF, MCP-1 Combination [284] + collagen/fibronectin delivery of hydrogels microparticles containing VEGF and MCP-1 (a mural cell recruiter) Gelatin bFGF, G-CSF Controlled delivery [285] of bFGF and G-CSF improved perfusion

and neovascularization Gelatin bFGF, BDNF Co-delivery of bFGF [286] and BDNF (an EC survival factor) improved angiogenesis Gelatin bFGF, platelet-rich A vast array of pro- [287] plasma growth factor angiogenic factors mixture (VEGF, enhanced therapeutic PDGF-BB, TGFβ-1, angiogenesis EGF) Star-PEG/heparin VEGF, bFGF Difference in heparin [288] binding affinity led to independent release of each growth

Star-PEG/heparin VEGF, bFGF, SDF-1 Sustained [278] independent delivery of three growth factors enhanced network formation

of a number of growth factors could be achieved. Those growth factors with less affinity were released more quickly, and those with higher affinity were maintained within the hydrogels for a longer time, increasing their healing potential. Importantly, this illustrates the different binding affinities of ECM components, which often act to sequester growth factors and release them when specific microenvironmental cues are activated [289].

As more binding relationships between growth factors and ECM protein components have been discovered, more complex hydrogel systems may be constructed. For example, placenta growth factor-2 (PlGF-2) has been shown to bind with high affinity to many ECM proteins. By binding additional growth factors to PlGF-2 in fibrin hydrogels, enhanced wound healing and

52 improved angiogenesis could be achieved. A major issue with VEGF delivery could also be resolved using high affinity bound VEGF, rather than wild type VEGF. VEGF has been shown to induce vascular permeability, which has led to hypotension and edema clinically. By binding

VEGF with high affinity, lower permeability of new vessels was achieved [290]. This study, [290], explored the binding affinities of many growth factors to a range of ECM proteins. By utilizing the different binding affinities of growth factors for ECM proteins, temporal control of growth factors may be achieved to more closely mimic the in vivo mechanisms for vascular regeneration.

Successful temporal regulation of growth factor release was achieved by coaxial electrospinning of chitosan hydrogel/PELCL for a VEGF loaded inner layer and PELGA emulsion/PELCL electrospinning for a PDGF-BB loaded outer layer. Faster release of VEGF helped initialize the angiogenic response, while slower release of PDGF-BB helped support and stabilize new vessels [291].

While much work has been done controlling release of many pro-angiogenic factors, inhibition of certain factors is also a necessity in the angiogenic pathway. The Notch pathway modulates VEGF signaling. The Notch pathway is important throughout the lifetime of ECs, but most notably for applications in vascular regeneration, it limits EC proliferation and maintains the resting state while downregulating VEGFR2 expression. By inhibition of Notch signaling through

DAPT (a gamma secretase inhibitor), the effects of VEGF in the early stages of angiogenesis are amplified. Delivery of VEGF and DAPT by alginate hydrogels showed improved healing effects in a mouse hind-limb ischemia model [292].

Many therapies incorporating biological factors with known pro-angiogenic activity have been developed. Accounting for limitations of previous hydrogel systems and issues that have arisen as a result of failures of treatments with biological factors has resulted in many potential

53 treatment options with a prospective efficacy in the clinic. Control of growth factor release to prolong the effects and mature new blood vessels to a stable, functional form may be achieved by current technologies as they proceed to clinical trials. However, additional temporal and spatial control of multiple growth factor release to more precisely mimic angiogenesis may contribute to optimization of hydrogel therapies in the future.

2.4 Altering physical properties of hydrogels to regulate vasculogenesis and angiogenesis

In parallel with biological factors, many physical factors contribute to maturation of new blood vessels. Just as the addition of pro-angiogenic biological factors aims to recapitulate the microenvironment most conducive to vascular regeneration, designing hydrogels to account for physical factors of that microenvironment has been shown to be another vital factor towards effective treatment of vascular diseases. Some of these important parameters have been mentioned above, but as decoupling the pro-angiogenic effects of physical properties from those of biological properties is difficult, specific hydrogel systems have been developed to study these physical factors. A discussion of these hydrogels and the important results relating to improvement of therapeutic angiogenesis is provided below.

2.4.1 Geometry and ultrastructure (pore size, connectivity)

Advances in photolithographic techniques in the past decades have enabled the development of many novel devices in a wide range of fields. As it pertains to biological systems and specifically study of angiogenesis, micropatterning has proved beneficial. Geometric micropatterning of ECM proteins has been studied for many years and has been shown to influence

EC morphogenesis [293, 294]. By creating well-defined structures through micropatterning, geometric influence on EC organization has become clear. While much work has been done

54 micropatterning ECM proteins, these proteins often have additional functions which may be cryptic and thus difficult to account for. By using PEGDA hydrogels with micropatterned adhesion

(RGDS) sites of a variety of concentrations and a variety of widths, influence of physical cues through geometry were elucidated. Wide areas of cell adhesion RGDS were not conducive to vascular network morphogenesis (cells remained in quiescent state), but thinner (50 µm wide) micropatterned areas showed vascular structure formation. Importantly, this work showed that precise control of binding area provided cues for EC to form (or not to form) networks [295]. In the future, degradation sites (MMP sensitive) and incorporation of cell recruiting growth factors and other pro-angiogenic factors may lead to effective therapeutic constructs. Another group has developed a more complex patterning of ECM components in a variety of concentrations to regulate tube size [296]. As new vessels mature following angiogenesis, their orientation and organization may be controlled by tissue-derived forces. Control over these forces by applying patterning cues throughout the development of new blood vessels can lead to a desired blood vessel alignment and directional patterning [297]. While the systems in [295-297] were designed as pre- vascularized constructs, the concepts of each may be applied to acelluar implants. Proper geometry and manipulation of hydrogel forces coupled with incorporation of biological factors may be able to elicit function of vessels of a desired size and organization, based on patterned implantable hydrogels. Similar to geometry, the ultrastructure of the hydrogel, has been shown to influence and direct angiogenesis. Of particular importance is the pore size of the gel. When pore size promotes cell infiltration and the correct spatial cues, functional networks may be obtained.

Typically, pore sizes between 50 and 300 µm have been shown to permit mature vascular network formation [270, 298, 299]. Smaller pore size limits cell infiltration, thereby limiting the formation of vascular networks [270, 298, 299]. Within the effective range, there are competing theories as

55 to the optimal pore size. In [270], it was hypothesized that a smaller pore size (60 µm) within the range established above (50-300 µm) provided a stiffer structure that may help guide the initial steps of angiogenesis. These smaller pore sizes can then degrade to allow additional infiltration and remodeling. The increase in cell infiltration and thus more cellular control over angiogenesis by remodeling and degradation is thought to be the reason for the better results with larger pore sizes in [298] and [299]. It is likely that specific optimal pore size is dependent on hydrogel material. For example, in highly degradable hydrogels, a smaller pore size may eventually permit network formation, as pore sizes increase with degradation.

2.4.2 Degradation

As mentioned above, degradation plays an important role in angiogenesis. Degradation allows for important remodeling of the microenvironment, which permits vascular network formation. Most hydrogel systems require a degradation component, as ECs often migrate along the path of degradation to form networks [300]. Degradation of natural hydrogels happens along with cell infiltration. However, in synthetic hydrogels, which do not inherently have sites of degradation, protease sensitive peptide sequences (often MMP-sensitive) allow for invasion and remodeling by ECs and mural cells [301] (use of MMPs is nearly ubiquitous with synthetic hydrogels). Control of cell-mediated degradation has been achieved through patterning of regions that either permit or inhibit degradation. This system has enabled specific study of the effects of degradation on vascular tube formation[274]. As with all other topics discussed, combination of degradation with other important pro-angiogenic factors, whether they be biological or physical, are vital for clinical success. Degradation is also a fundamental property that facilitates growth factor release, which, as discussed, is of upmost importance for effective hydrogels.

2.4.3 Hydrogel stiffness

One of the many reasons hydrogels have been widely used as scaffolds for tissue regeneration is their ability to mimic many of the properties of the native ECM. One of the important factors that may be regulated and finely tuned in many hydrogels is mechanical stiffness.

It has become clear through the work of many research groups that hydrogel stiffness plays a major role in therapeutic angiogenesis. For example, softer (storage modulus ~1 kPa) hydrogels with loose ultrastructure (as discussed) permit increased proliferation and survival, while hydrogels with slightly higher elasticity (storage modulus ~2.5 kPa) were more conducive to tube formation.

Importantly, it was not stiffness that modulated release rate of growth factors in this experiment, rather amount of binding molecule (heparin). [288]. While the change in stiffness provides important information about in vitro characteristics of ECs, optimizing in vivo properties is a necessity for clinical use. With this, use of softer hydrogels has been shown to be more beneficial, as they permit increased EC infiltration. Many groups have shown with softer hydrogels with loose ultrastructure, vascular cells form more robust vascular networks [278, 282, 302-304]. While the coupling of hydrogel stiffness and hydrogel degradation is clear in most hydrogel systems (with degradation, hydrogel stiffness decreases further enabling organization and network formation of

ECs), one group has recently decoupled the two and revealed a potential mechanobiological signaling that may contribute to the advantage of compliant versus stiff hydrogels. The importance of the RGD cell adhesion peptide sequence has been shown time and again. EC survival and proliferation are partially dependent on RGD. It is also known that β1 integrins are important in tubulogenesis. These integrins have a known binding site on RGD, indicating there may be an interplay between the two that enables RGD bound cells to sense the stiffness of the substrate to which they are bound [305]. Further investigation into the mechanism by which hydrogel stiffness

57 regulates angiogenesis is necessary for advanced understanding and potential optimization of new hydrogel materials. However, it has become clear stiffness provides critical cues to facilitate formation of vascular networks.

2.5 Other microenvironmental factors affecting vasculogenesis and angiogenesis: future opportunities

While there have been significant advances in hydrogel technologies designed for vascular regeneration over the last 15 years that have accounted for many of the vital factors contributing to formation of functional vasculature, several important factors have recently been taken into consideration, with vast potential for clinically relevant and potentially effective therapies. These therapies represent microenvironmental factors that have effects upstream of pro-angiogenic growth factor expression. As they act upstream, these factors have the potential to elicit a magnified physiologic response, with increases in expression of many pro-angiogenic growth factors and other signals.

2.5.1 Hypoxia

Hypoxia, defined as low partial pressure of oxygen (<5% O2) has been shown to play a vital role in activation of angiogenesis. Hypoxia plays a role in a number of disease states including tissue ischemia, coronary artery disease, and cancer. Hypoxia inducible factors (HIFs) are transcription factors which act to regulate oxygen homeostasis. Under normoxic conditions, HIFs are degraded, but as oxygen levels become hypoxic, HIFs are stabilized and activated. Increases in HIF expression cause numerous downstream effects including increased expression in many pro-angiogenic growth factors, including VEGF, SDF-1, PlGF, Ang-1, Ang-2, PDGF-BB, and stem cell factor. In addition, HIF expression recruits bone-marrow derived angiogenic cells [83,

230, 306]. Taken together, the list of factors here effectively summarizes the list of growth factors that have been conjugated to hydrogels and tested for their potential as therapeutics. With this motivation, our group has developed hydrogels that induce a local hypoxic environment through an oxygen consuming, laccase-mediated cross-linking reaction. Laccase has a high affinity for phenol groups, thus gelatin-based polymers were functionalized with phenol groups. Laccase was able to successfully cross-link gelatin polymers to form hydrogels, all the while consuming oxygen to produce a hypoxic micronenvironment. These hydrogels have been shown to be effective in promoting formation of vascular networks and are able to establish neovasculature, which is functional and able to anastomose with the host network [307]. In order to enhance the effects of hypoxia, we have developed another, Dextran-based, hydrogel, which is able to maintain a prolonged hypoxic microenvironment (>10 hours) enabling increased stabilization of HIFs and likely more potent angiogenic effects [308].

2.5.2 Reactive radical species

Both reactive oxygen species (ROS) and reactive nitrogen species (RNS) have been implicated in pro-angiogenic pathways. While ROS and RNS have been historically associated with negative effects such as cytotoxicity and mutagenicity, appropriate (low) levels of ROS and

RNS aid in stimulation of gene expression, cell growth, migration, and differentiation. In addition, exogenous ROS and RNS may stimulate VEGF expression, cytoskeletal reorganization, and tubular morphogenesis. Effects of VEGF and Ang-1 are also enhanced by ROS and RNS.

Neovascularization has been shown to be dependent on ROS and RNS production, as inhibitors of

ROS are shown to decrease formation of new vessels in several disease states[309-311]. Vessel maturation by recruitment of mural cells is mediated by ROS and RNS [312]. ROS and RNS may also be produced through hypoxia and reoxygenation cycling. By mediating expression of HIFs,

59 these ROS and RNS influence myriad angiogenic activites [313]. Currently, no hydrogel systems have aimed to directly influence ROS or RNS production. However, this may prove an interesting area of research, and as an additional mechanism for influencing “upstream” effects that provide the full range of pro-angiogenic responses. As a potential pitfall to this design, optimal ROS and

RNS levels must be determined and obtained as to permit healing, rather than the potential negative effects that may lead to cell death.

2.5.3 Hyperoxia

Hyperbaric oxygen treatment for enhanced wound healing presents a potential paradox in oxygen regulation of angiogenic pathways. Hypoxia has been clearly shown to be a potent mediator of vascular regeneration. Interestingly, hyperbaric oxygen treatment employs similar mechanisms to hypoxia through production of ROS and RNS which, in turn, increase expression of HIFs and their many downstream pro-angiogenic effects. Hyperbaric oxygen treatment increases circulating progenitor cell (EPC) mobilization to the site of injury and increases growth factor production by these EPCs [314, 315]. Additional interplay between hypoxia and ROS/RNS production, as discussed above ([313]), may prove an interesting area of research. Because

ROS/RNS production can be mediated by hyperoxia, development of precise control of oxygen

(both hypoxic and hyperoxic) could elicit a synergistic response with increased accumulation of

HIFs. Local hyperoxic microenvironments may also prove to be an interesting area of research.

Although positive therapeutic effects have been achieved by systemic delivery of increased oxygen

(hyperbaric oxygen treatment), investigation of local delivery of high oxygen levels has not been performed.

2.6 Conclusions and future directions for the field

Much work has been done to develop implantable and injectable hydrogels to promote vascular regeneration. Many approaches have sought to develop hydrogels with controlled release of pro-angiogenic growth factors for sustained angiogenesis and ultimately maturity of neovasculature. Combined effects of multiple growth factors have enhanced recapitulation of the natural angiogenic microenvironment, but has failed to account for the immense complexity of the angiogenic healing pathway. Spatial and temporal release of growth factors has enhanced therapeutic angiogenesis and holds great potential. However, additional parameters, including matrix stiffness, degradability, ultrastructure, and geometry have also proved vital. Developing a hydrogel that optimizes and accounts for as many of these parameters as possible will produce the most favorable hydrogels for treatment of vascular diseases and disorders. As more complex hydrogels are developed, it is important to maintain the ability to simultaneously decouple a range of parameters to study additive or synergistic effects. In the coming years, it is likely that promoting upstream effectors of angiogenesis will lead to more effective therapies, as control of upstream effect factors leads to activation and increased expression of many therapeutic factors and may activate more pro-angiogenic factors than previously possible.

CHAPTER 3 3 Hypoxia and Matrix Manipulation for Vascular Engineering 3.1 Teaser

Hypoxia and ECM properties are two broad and potent regulators of vascular development and regeneration, and both have recently been engineered into cell culture platforms.

3.2 Introduction

The permanency of multicellular organisms on Earth is stringently reliant on the ability of cells to respond and adapt to their surrounding milieu. Cells respond to a complex array of biological, biophysical, and biochemical cues to develop and regenerate functional tissues.

Alterations to these properties can catastrophically result in impaired development or healing, as well as tumor development and metastasis. At the most basic level, two of the most critical components of the cellular microenvironment are oxygen (O2) and the composition of the extracellular matrix (ECM).

According to the most recent findings, O2 reached sufficient levels (estimated to be 0.2 to

2% O2) for aerobic organisms to be able to survive between 2.45 and 2.2 billion years ago in the oceans [316, 317] and between 540 and 600 million years ago in Earth’s atmosphere [317-319].

Ever since, O2 has been a highly available potential source of energy for multicellular organisms to commence, survive, and multiply. Multicellular organisms require specialized systems to enable sufficient amounts of O2 to reach their cells. For instance, insects regulate the transport of O2 into their tissues with a special respiratory system consisting of spiracles and trachea. Around their tissues, they retain the relatively low O2 levels (1.4 mmHg) thought to be equivalent to the atmospheric O2 concentrations at the time of their evolution [320, 321]. In vertebrates, O2 is carried by proteins in the blood, particularly by hemoglobin, and is transported to tissues through

62 endothelial cells (ECs). The cells throughout the body are highly dependent on the dynamics of

O2. In humans, O2 concentrations vary between 1 and 10% in tissues (other than the lungs) and between 5 and 13% in blood vessels [182, 322]. Therefore, the ability for cells to sense and respond to O2 is critical, and O2 acts as a signaling molecule for cells, regulating their metabolism, survival, cell-cell interactions, migration, and differentiation.

Besides O2 availability, the transition from unicellular to multicellular organisms requires that cells be connected together in a way that allows them to interact with each other as parts of the same system. This interconnectedness could happen either by having junctions at the cell peripheries or by having connecting cement between the cells. Many multicellular organisms connect their cells in both ways: they use cellular junctions to allow direct signaling between cells, and they use the ECM to regulate the transport of molecules (e.g., O2, glucose, and signaling proteins) between the cells by remodeling the components of the ECM. Thus, both cell-cell and cell-ECM interactions are significant for determining the fate of cells in tissues. Transmembrane proteins known as integrins are responsible for signaling from the ECM to the cell. Therefore, cells have different responses with respect to the composition and structure of the surrounding ECM. In particular, vascular morphogenesis is regulated by endothelial cell (EC) interactions with the ECM through integrins and is highly dependent on the ECM context [323].

In the field of vascular engineering, the effects of both O2 tension and the ECM on blood vessel formation continue to be extensively investigated. Blood vessel formation essentially occurs by angiogenesis or vasculogenesis. Angiogenesis is the formation of blood vessels from pre- existing vasculature, orchestrated with the proliferation, migration, and assembly of ECs, as well as the remodeling of the ECM [324, 325]. Most of the ECs comprising the blood vessel walls are in a state of quiescence in physiological conditions. A stimulus is required for ECs to switch from

63 their resting state to their navigating state, where they are activated to produce angiogenesis- promoting proteins [326]. Angiogenesis occurs in several situations, such as wound healing, arthritis, cardiovascular ischemia, and solid tumor growth [327, 328]. In all of these situations, the tissue or vasculature is deprived of oxygen, leading to hypoxic conditions that promote angiogenesis. For vessel sprouting, the ECM surrounding the vasculature needs to be degraded so that ECs can easily navigate into the tissue and proliferate. Hypoxia is known to promote the production of ECM-degrading enzymes via secretion from activated ECs [329-331]. Thus, EC sprouting is more favorable toward hypoxic regions in the ECM through which the secretion of enzymes is upregulated by hypoxia, whereas the invasion of vessels into the ECM is not favored in the direction of sufficiently oxygenated regions.

An oxygen gradient emerges in early development, which guides cellular differentiation and morphogenesis [332]. While the O2 uptake of early embryonic cells relies on the simple diffusion of oxygen, hypoxia starts to be observed in different regions as the embryo expands [332,

333]. The initial vascularization, vasculogenesis, starts with the differentiation of angioblasts

(embryonic progenitors of ECs), which surround hemopoietic cells to form blood islands. Blood islands ultimately fuse as angioblasts differentiate into endothelial cells to form the primary capillary plexus, then undergo further tubulogenesis and vascular network formation throughout the yolk sac [22, 334, 335]. This process of vasculogenesis has been suggested to occur in hypoxic conditions [332, 334]. Hypoxia also stimulates microvascularization and the capillary network to form around the developing organs. Vasculogenesis in adult organs has been demonstrated to originate from endothelial progenitor cells (EPCs) circulating in the blood [336]. The migration of

EPCs and their recruitment to the appropriate sites to induce the formation of new blood vessels depends on complex cell signaling. Circulating EPCs home to hypoxic regions along both O2 and

64 growth factor gradients, in particular gradients of stromal derived factor 1 (SDF-1) [337, 338].

Investigations of tumor growth and wound healing have revealed that hypoxia occurs in both situations, inducing EPCs to migrate from the circulating blood through the ECM. Hypoxia also plays a role in the recruitment of EPCs by promoting receptor expression on the tissue that recognizes EPCs [339], as well as on the EPCs themselves [340], which is followed by their differentiation into mature ECs [341]. Moreover, vascular endothelial growth factor (VEGF), a key regulatory protein known to induce vasculogenesis and angiogenesis, was found to be upregulated in hypoxia [342]. These processes take place in the milieu of the ECM, which is mostly composed of fibronectin during early development [323, 343]. In adult tissue, on the other hand, collagen becomes abundant and controls the cellular fate.

The formation of new blood vessels through angiogenesis or vasculogenesis depends on the dynamic effects and interplay between the ECM and oxygen tension. A thorough understanding of the mechanisms involving the ECM and O2 during angiogenesis and vasculogenesis is essential for the fundamental understanding that can be harnessed for developing vascular engineering applications. Indeed, the effects of these two factors on vascular cells are being investigated extensively. The in vitro vascularization of primary vascular cells has been studied using many different biomaterials [152, 344, 345] as three-dimensional (3D) matrix components, and they were shown to influence various aspects of angiogenesis and vasculogenesis. Similarly, a considerable amount of work has focused on the effects of hypoxia- inducible factors (HIF1α, HIF2α, HIF3α) on the regulation of genes that induce vasculature network formation [342, 346, 347]. In addition, some researchers have also investigated the effects of hypoxia and the ECM context together [348, 349]. Success in engineering blood vessels from

65 primary vascular cells or stem cells relies on understanding the influence of all critical parameters and controlling them in targeted directions.

The main focus of this chapter is a review and discussion of how the cells in the body respond to variations in oxygen tension and ECM components, leading to new vasculature formation.

3.3 Concepts in the Regulation of the Vasculature by Oxygen and the ECM

3.3.1 The influence of oxygen tension on vascularization

Variations in oxygen concentrations at every stage of embryogenesis and in different regions of adult tissues lead to diverse vascular responses, depending on the cell type and microenvironment. Many cell types respond differently, but also collectively, to the changes in O2 equilibrium through specialized sensing mechanisms and effectors in order to maintain homeostasis. In this section, we will first discuss the formation and location of poorly oxygenated regions in the body, as well as the mechanisms that cells utilize to sense changes in oxygen levels.

Then, we will then focus on several responses of pluripotent and vascular cells to low O2 tensions in terms of gene regulation, differentiation, oxygen consumption, and cell survival.

3.3.2 The in vivo consequences of oxygen gradients

3.3.2.1 Oxygen availability in the body

In vertebrates, O2 transport to the tissues relies on three main processes: the oxygenation of the blood in the alveoli in the lungs, the convectional transport of oxygen in blood along the veins, and the diffusion of oxygen across the vessel walls followed by penetration of O2 to the

66 deeper tissues. There are three distinct resistances to the mass transfer of the O2 molecule, which result in O2 gradients throughout the body.

O2 deprivation has been observed early in the development of mouse embryos [332].

Additionally, polarographic oxygen measurements in the human placenta have shown that O2 levels are 1.3 to 3.5% in the first 8 to 10 weeks and reach between 7.2 and 9.5% in weeks 12 to 13 of pregnancy [350, 351]. Oxygen levels measured in the gestational sac revealed even lower O2 levels in earlier stages of embryogenesis, where O2 is only transported by simple diffusion [352].

Diffusion, as opposed to convection, transports nutrients between cells very slowly. Before vasculogenesis begins, the maximum diameter that a spherical embryo can reach without having any anoxic cells was calculated to be 2 mm [353]. This value varies with the embryo’s geometry and, most importantly, with the O2 consumption of the animal cells. The results of in vivo imaging of various animal embryos show that the maximum diameter remains below 1 mm, which agrees with the theoretically estimated value [335, 353].

Vasculogenesis is crucial to facilitate cell proliferation and for the embryo to grow larger.

In mouse embryos, vasculogenesis commences after day seven, with the differentiation of the mesoderm into angioblasts, which then assemble to form a simple circulatory system consisting of a heart, dorsal aorta, and yolk sac by day eight [354, 355]. Afterwards, spatial increases are observed in O2 levels throughout the course of embryonic development [332]. These profound spatiotemporal O2 level changes in the embryo can be accepted as evidence for vascular formation during embryogenesis. The large existing vasculature then sprouts and proliferates to supply O2 and nutrients to cells located in poorly oxygenated regions. Hypoxia, considered the most critical factor controlling the angiogenesis process, works via numerous protein-signaling pathways. The mechanism determining the directionality of angiogenesis and the complex networking of

67 endothelial capillaries around the tissues is manipulated by several other parameters, including hemodynamic forces and cytokines; this mechanism will be discussed later in the chapter [356].

Once embryonic development is complete and sufficient concentrations of O2 and nutrients are supplied to the tissues, the oxygen gradient still persists in some tissues, providing several benefits to specific cell types. In adults, O2 distribution ranges from 1 to 13 in normal tissues.

Although the formation of blood vessels and capillary networking is complete, some tissues still lack vasculature, such as the bone marrow niche [357-359]. In such tissues, diffusion is the controlling mechanism for nutrient transport, thus resulting in a wide range of O2 distributions from the internal hypoxic region to the external regions, which remains at physiological O2.

The discovery of circulating EPCs in blood vessels revealed that neovascularization in adults is directed not only by angiogenesis, but also by the vasculogenesis process, which depends on the renewal, mobility, recruitment, and differentiation of EPCs [360-363]. The bone marrow

(BM) provides a host microenvironment for a variety of cells, including hematopoietic stem cells

(HSCs), mesenchymal stem cells (MSCs), and EPCs. The development of EPCs occurs in the BM, which has a unique structure that allows severe hypoxic regions to exist. Although the BM is inaccessible for noninvasive oxygen measurements, both simulation studies and qualitative measurements have demonstrated the existence of hypoxic regions. Several theoretical models have been developed in order to simulate the distribution of oxygen throughout the BM [358, 364,

365]. Chow et al. used homogeneous Kroghian models to estimate oxygen levels in the BM [364].

Their simulations suggested that both HSCs and EPCs are exposed to low O2 tensions in the BM.

There are various BM architectural organizations possible depending on parameters such as the spatial arrangement of vasculature and the distribution of many different cell types populating the

BM. Therefore, in the absence of supporting evidence from in vivo quantitative measurements,

68 model predictions must be used to assess the effects of different parameters on the O2 tension distribution in the BM. The model described by Kumar et al. considered three possible vessel arrangements to simulate oxygen level variations under various conditions [358]. They suggested that hypoxic, and even anoxic, regions could be found in the BM, assuming that the cells’ oxygen consumption is constant and that the density of arterioles in the BM is low.

On the other hand, qualitative observations in the study by Parmar et al. demonstrated that

HSCs are distributed according to oxygen availability in the BM [359]. Staining with pimonidazole and sectioning revealed the oxygen gradient throughout the BM, showing that HSCs more likely reside at the lower end of the gradient. These results are in agreement with other in vitro studies suggesting that hypoxia supports the maintenance of stemness [366-368]. Moreover, BM transplantation studies have shown that BM-derived EPCs enhance neovascularization and the formation of arteries [369, 370]. The renewal of EPCs in the BM depends on the differentiation dynamics of HSCs, which are regulated by the microenvironments (i.e., the niches) they reside in.

Osteoblasts, bone cell progenitors, bind to each other and to HSCs via adhesion molecules to form the osteoblastic niche that is located far from the sinusoidal arteries. Researchers have discovered the existence of another type of niche within the BM, the vascular niche, which is located closer to the sinusoidal arteries than the osteoblastic niche. The differences in physicochemical factors within the various niches play fundamental roles in controlling the dynamics of HSC migration and differentiation. Since the vascular niche’s close proximity to arteries means that it is richer in

O2 than the osteoblastic niche, Heissnig’s group hypothesized that HSCs are in a quiescent state in the osteoblastic niche’s severe hypoxic conditions [371]. When vasculogenesis is necessary in neighboring tissues, specific cell signaling stimulates the migration of HSCs from the osteoblastic niche to the more oxygenated vascular niche, where HSCs can switch from their quiescent state to

69 a proliferative state. The proliferation and differentiation of HSCs reconstitute the EPC pool in the vascular niche before they enter the circulation.

Wound healing, another situation where tissue hypoxia is prevalent, consists of a series of events that includes new vasculature formation, which is regulated by varying O2 levels. Platelets interfere with microcirculation in the wounded tissue, followed by the release of coagulation factors to reinforce the clotting process. Histamine and bradykinin, secreted by mast cells, also influence the microcirculation by enhancing vascular permeability and arteriolar vasodilation, thereby increasing the blood flow rate [372, 373]. Recruitment of leukocytes and macrophages into the damaged tissue is followed by their activation in response to several growth factors (GFs) and integrins. High rates of O2 consumption in activated macrophages, along with perturbation of the microcirculation, leads to a further decrease in O2 levels and results in hypoxia [374], which leads to the accumulation of HIF1α at the wound site [375]. Albina et al. [376] showed that the

HIF1α mRNA of inflammatory cells peaks about six hours after injury. On the other hand, HIF1α protein levels could be detected between one and five days after wounding. More recently, Zhang et al. [375] demonstrated that, during the burn wound healing process, the accumulation of HIF1α increases the number of circulating angiogenic cells, as well as smooth muscle actin-positive cells, in the wounded tissue. These hypoxic conditions — either directly or indirectly through the accumulation of HIF— stimulate angiogenesis during wound healing.

Ischemic tissues, including those affected by myocardial infarction or peripheral artery occlusion (e.g. limb ischemia) have also been a hotbed for study of the effects of low O2 on cellular recruitment and tissue regeneration. In particular, lack of O2 delivery to these diseased tissues results in HIF stabilization and subsequent upregulation of recruitment chemokines, perhaps most importantly SDF-1, as well as transmembrane proteins integrin β2 and ICAM-1 that facilitate

70 adhesion of circulating cells to the damaged endothelium [337-340]. Importantly, hypoxic conditions also facilitate ECM remodeling through upregulation of proteases, such as cathepsins and matrix metalloproteinases [377-379]. These factors, in concert with HIF-induced production of other pro-angiogenic factors, such as VEGF, lead to robust formation of neovasculature.

3.3.2.2 Oxygen-sensing mechanisms of vascular cells

Most cell types in the body respond to variations in O2 tensions [380]. Gene expression, viability, metabolism, and the oxygen uptake rate of the cells change with alterations in O2 levels, in order to maintain homeostasis. When cells experience a change in extracellular O2 levels, they adapt to the new conditions, which may occur rapidly. Hence, O2 sensing in cells is expected to be controlled by well-organized, highly sensitive mechanisms.

Several mechanisms have been proposed in the literature to account for O2 sensing in cells.

Although their sensitivities may differ from one another, more than one such mechanism can co- exist in a cell, resulting in various cellular responses. Within the cell, the O2 molecule mainly engages in two distinct processes: it is involved directly in biosynthesis reactions; or it participates in metabolic processes, such as the electron transport chain occurring in mitochondria. Any change in the concentration of O2 extensively perturbs these processes and, following a sequence of events, may have a number of different effects on the cell. Therefore, O2 sensors in cells can be mainly categorized as mitochondria-related sensors (bio-energetic) and biosynthesis-related sensors (bio-synthetic) — although they can be linked to each other in some cases, making the distinction not completely clear [380].

Among the several effectors of O2-sensing mechanisms, HIFs are the most essential in terms of the diversity of their influences. The family of HIFα subunits (HIF1α, HIF2α, and HIF3α)

71 has been shown to be responsible for regulating expression of a large number of genes, including those coding for key regulatory proteins of angiogenesis and vasculogenesis. Although HIFα is expressed at every oxygen tension, it is rapidly ubiquitinated in normoxic conditions, resulting in its degradation. Thus, the amount of intracellular HIFα protein depends on the balance between its expression and degradation. In conditions of low O2 availability, all HIFα proteins heterodimerize with HIFβ (ARNT) and form a transcriptional complex which regulates the transcription of numerous genes [342]. Stabilization of HIFα in the cell is controlled by two main O2 sensing proteins, prolyl hydroxylase domain (PHD) and factor-inhibiting HIFα (FIH), which belong to the previously mentioned bio-synthetic sensors category. Three isoforms of PHDs are present in all mammals [381]. Specific proline residues on the oxygen-dependent domain of HIFα are hydroxylated by PHDs at separate hydroxylation sites, leading to HIFα degradation. The activity of PHDs in the cytoplasm is controlled by various O2-dependent molecular events and, directly, by the concentration of the O2 molecule [346]. All three PHDs remain partially active in normoxia.

PHD activity is expected to be very sensitive to small changes in cytoplasmic O2 levels since Km, the Michaelis-Menten parameter for the activation of PHDs, is approximately 230 to 250 µM, which is much higher than physiological oxygen concentration (approximately 60 µM) [382].

Besides, mitochondria are also involved in the PHD activation process through their consumption of O2, regulation of reactive oxygen species (ROS), and production of nitric oxide (NO). While the stabilization of HIFα depends on PHD activity, the expression of HIFα is controlled by FIHs.

Therefore, when O2 levels are lowered, both the stabilization and transactivation of HIFα increase, resulting in several angiogenic responses that will be discussed in the following section.

NO and ROS not only contribute to the HIFα stabilization process, but they also have several direct effects on vascular cells and blood vessels. A number of studies have shown that NO

72 induces angiogenesis, hyperpermeability, and vasodilation [383]. Moreover, NO also perturbs EC respiration through the inhibition of cytochrome c oxidase, which causes lower mitochondrial O2 consumption [384]. Mitochondrial ROS are also increased as a consequence of electron transport chain inhibition, which then contributes to the deactivation of PHDs via oxidizing cofactor Fe (II) and helps to stabilize HIFα. ROS production, in respect to hypoxia, is proportional to the concentrations of intracellular O2 and electron donors. Under hypoxia, the amount of O2 required to form superoxides is decreased, whereas the concentration of the electron donors increases as a consequence of the reduction in the proximal electron transport chain. Therefore, ROS production can change in both manners, depending on the variations in these molecules’ concentrations [380].

Ushia-Fukari et al. [385] showed that ROS influence the expression of surface adhesion molecules of ECs and stimulate EC proliferation and vessel permeability. Moreover, the hypoxia-induced decrease in ROS production leads to the inhibition of K+ channels of pulmonary artery smooth muscle cells (SMCs), whereas an increase in ROS production leads to intracellular Ca+ release from ryanodine-sensitive stores [380]. Another molecular path found between mitochondrial energy generation and K+ channel inhibition occurs through AMP kinases. The energy of the cell is generated by the conversion of ADP to one molecule of ATP and AMP. Hence, AMP kinase becomes highly dependent on the ADP/ATP ratio, which is very sensitive to changes in

+ cytoplasmic O2 concentrations. AMP kinases were shown to inhibit K channels through the regulation of Ca+ release in pulmonary arterial SMCs and also to induce cellular survival in tumor cells when exposed to severe hypoxia [386, 387].

Moreover, heme oxygenases (HOs) and NADPH oxidases (NOXs) play important roles in the biosynthetic oxygen sensing of cells. NOX-2, one of the three isoforms of NOX, is used for superoxide production from molecular O2. Hypoxic conditions can cause a decrease in NOX-2-

+ derived ROS concentrations, due to the low Km values (18 µM) of NOX-2; this helps Ca release in pulmonary artery SMCs [388]. However, some studies also suggest that hypoxia increases

NOX-2 activity, therefore causing the generation of a greater amount of ROS [380]. On the other hand, Ca+-activated K+ channels in glomus cells were shown to be related to the activity of HO-2, an isoform of HO which can convert heme to CO, biliverdin, and Fe(II) using O2 and NADPH

[389].

The effectiveness of an oxygen sensor can be determined by evaluating (a) its sensitivity to small changes in intracellular O2 levels and (b) the subsequent diversity of triggered cellular responses. Taking this considerations into account, PHDs and FIHs appear to be the most critical oxygen sensors responsible for controlling HIF activity [390]. Deactivation of these two sensors leads to HIFα stabilization, initiating the regulation of hundreds of different genes. Using O2 as a controlling parameter to engineer vascular tissues demands a clear understanding of the biochemical events that follow changes in O2 tension, as well as the net response of the cells and how O2 affects their collective behaviors.

3.3.3 Cellular responses to different oxygen concentrations

3.3.3.1 Metabolism and oxygen uptake rate

Several studies have observed that the O2 consumption of cells depends on O2 availability

[387, 391-393]. We have recently shown that the O2 uptake rates (OURs) of EPCs and human umbilical vein endothelial cells (HUVECs) are similar, but not identical, to each other and that both decrease when O2 availability is lowered [393]. Many mechanisms have been proposed to explain the relationship between mitochondrial O2 consumption and variations in O2 levels. HIF1α was found to be responsible for inducing the enzymes required for glycolysis [387]. It also plays

74 a role in activating pyruvate dehydrogenase kinase-1, which reduces the amount of pyruvate that flows into the TCA cycle and therefore decreases aerobic respiration in mitochondria [387]. In addition, the increases in both the transcription and expression of glucose transporter protein 1

(GLUT-1) were shown to be HIF1α-dependent in hypoxic conditions [394]. In other words, deactivation of PHDs and FIHs at low O2 levels leads to the stabilization of HIFα, which then reduces O2 aerobic respiration by inducing pyruvate degradation while, at the same time, promoting glycolysis by increasing the expression of glucose transporter proteins. Another proposed mechanism involves the inhibition of cytochrome oxidase by NO, which is known to be regulated by shear stress and O2 tension. NO influences mitochondrial respiration by the competitive inhibition of cytochrome oxidase with O2 and by inhibiting electron transfer between cytochrome b and c, therefore increasing ROS production [392].

The effects of blood flow and O2 tension are crucially important for the ECs comprising the vessel walls, since these conditions can be perturbed in many pathophysiological situations in the body. Some studies have shown mitochondrial respiration of ECs to be lower than other cell types, suggesting that most of the O2 consumption is non-mitochondrial [395, 396]. Helmlinger et al. demonstrated that ECs consume O2 during capillary formation, whereas they also preserve and expand the capillary structures, even under severe hypoxia (about 0.6% O2), by upregulating

VEGF expression [397]. It is not surprising that ECs possess a special type of metabolism — aerobic glycolysis in their resting state (physiological conditions) and anaerobic glycolysis in their navigating state (hypoxic conditions) — since O2 is transported through ECs to other tissues and thus they must possess the ability to survive and commence angiogenesis under hypoxic conditions

[326].

Moreover, when ECs are exposed to excess glucose, their ATP generation shifts to glycolysis, and lactate levels, increased as a by-product of glycolysis, contribute to the inactivation of PHDs and, therefore, the stabilization of HIFα [398]. Where blood flow is perturbed, such as in ischemia and wound healing, both NO and O2 levels are changed in blood vessels, and all of the metabolic variations discussed become more important.

3.3.3.2 Transcription of angiogenic genes

Manalo et al. showed in their study of ECs that 245 genes are upregulated and 325 genes are downregulated at least 1.5-fold in response to hypoxia and HIF1α. These genes are responsible for the expression of collagens, GFs, receptors, and transcription factors, all of which are significant for the processes of angiogenesis and vasculogenesis. This wide range of hypoxia- related transcription factors also indirectly affects HIF1α. The genes directly regulated by HIF1α include VEGF-A, VEGFR-1, Flt1-1, and erythropoietin (EPO). Examples of indirectly regulated genes include fibroblast growth factor (FGF); placental growth factor (PLGF); platelet-derived growth factor (PDGF); angiopoietins (ANG-1 and 2); and Tie-2, the receptor of ANGs [346].

Although VEGF is the major GF that stimulates blood vessel formation, when it alone was transgenically overexpressed in mice, defective blood vessels formed, which then led to tissue edema and inflammation [12]. On the other hand, overexpressing both VEGF and ANG-1, which is important for maintaining vascular integrity, has been shown to induce hypervascularity without imperfections in mice [399]. ANG-2 is responsible for EC apoptosis and vascular regression in the absence of VEGF, whereas, when combined with VEGF expression, it enhances angiogenic responses by destabilizing the blood vessels [400, 401]. More recently, ANG-4 was shown to function similarly to ANG-1 and to induce angiogenesis by binding the ANG receptor TIE-2, which is also upregulated by HIF1α [347]. We have recently shown that VEGF and ANG-2 genes

76 are upregulated in hypoxic (1% O2) cultures of EPCs and HUVECs [393], and the fold differences in upregulation levels of VEGF and ANG-2 in EPCs were shown to vary during the three-day exposure period, where no significant change was observed for HUVECs. How hypoxia affects the regulation of these angiogenic genes depends on the cell type; for instance, VEGF is upregulated in ECs, SMCs, cardiac fibroblasts, and myocardiocytes, whereas ANG-2 is induced only in ECs [342]. Therefore, from a tissue engineering perspective, co-culturing of different cell types under controlled hypoxic conditions should be considered, since a combination of hypoxia- induced angiogenic proteins is required to obtain vascular formation without excessive permeability.

3.3.3.3 Cell death and survival

Hypoxia influences the proliferation and viability of many cell types [366, 393, 402, 403].

The wide spectrum of HIF1α-dependent genes also includes proapoptotic and prosurvival genes.

BH3-only proapoptotic genes, a subfamily of BCL-2 that includes BNIP3, BNIP3L, NOXA,

RTP801, HGTP-P, are directly activated by HIF1α [404]. Although these genes play important roles in cellular apoptosis, a growing body of evidence suggests that hypoxia mediates cellular survival in many cell types [402, 403, 405]. Programmed cell death is, of course, a very critical step for cells and is most likely taken only after all possible survival mechanisms have been exhausted. One of these mechanisms, autophagy, is a cellular catabolic process where cytoplasmic organelles are degraded to provide ATP generation in nutrient deprivation. Hypoxia was found to induce mitochondrial autophagy via both HIF1α-dependent and HIF1α-independent pathways

[402, 403]. Small interfering RNA silencing of BNIP3 and BNIP3L together suppresses autophagy to a greater extent than silencing only one of them at a time [406]. Zhang et al. have shown that mitochondrial autophagy is induced by HIF1α-dependent upregulation of BNIP3 incorporated into

77 the constitutive expression of BECLIN-1 and ATG-5 [402]. On the other hand, the neuron-derived orphan receptor (NOR-1), which is overexpressed in ECs exposed to hypoxia, mediates cellular survival as a downstream effector of HIF1α signaling [405]. CD105, one of the EC markers also shown to play a role in cellular survival, is significantly upregulated under hypoxia [407]. In vivo studies of rats subjected to hypoxia also found the induction of mitochondrial autophagy by overexpression of BNIP3 [408]. In addition, Papandreou et al. propose that hypoxia induces autophagy in tumor cells through AMP kinase, which is activated by hypoxia independently of

HIF1α, as discussed previously in the O2 sensing section [403].

3.3.4 Cell Pluripotency and Differentiation

Vasculogenesis takes place in low O2 environments, such as the early development of the embryo, EPC regeneration in the BM, or EPC attachment and differentiation into mature ECs at neovascularization sites. All of these processes rely on pluripotent/unipotent cells differentiating into the endothelium, where O2 tension is a crucial parameter regulating their differentiation characteristics. As already discussed, EPC regeneration in the BM depends on cellular dynamics between the osteoblastic niche (low O2) and vascular niche (high O2); HSCs are quiescent in the osteoblastic niche and differentiate into EPCs in the vascular niche before joining the circulation

[409]. Therefore, it is important to understand the effect of O2 tension on the differentiation of cells into EPCs/ECs as a primary step of vasculogenesis. Hypoxia enhances human embryonic stem cell (hESC) pluripotency via the upregulation of Oct-4, NANOG, and SOX-2, which are pluripotent markers [366-368, 410]. HIF2α is responsible for the overexpression of Oct-4, SOX-

2, and NANOG, while HIF3α also plays a role in the process by inducing HIF2α transcription

[366, 368]. Prasad et al. demonstrated that hypoxic conditions (5% O2) prevent the spontaneous differentiation of hESCs, whereas the inhibition of Notch activation revoked this effect, suggesting

78 that hypoxia-induced pluripotency occurs via Notch signaling [411]. On the other hand, the efficiency of the process of reprogramming mouse and human somatic cells into induced pluripotent stem cells (iPSC) was shown to be improved in 5% O2 cultures, compared to atmospheric O2 cultures [412]. In contrast, other studies have demonstrated that hypoxia induces the expression of early cardiac genes in spontaneously differentiating embryoid bodies (EBs) [413,

414]. In a more recent study, Lopez et al. showed that EPCs/ECs can be obtained from hESCs more efficiently when cultured in 5% O2, compared to previous methods that induce EB formation in atmospheric O2 [415]. Additionally, simply priming EBs in hypoxic conditions mediated suppression of pluripotent marker Oct4 and upregulated VEGF [416]. When hPSCs are differentiated towards an endothelial lineage, hypoxia has also been shown to enhance EC differentiation through changes in the early stages (mesodermal specification) of EC lineage commitment. Interestingly this affect, which was dependent on low O2 tension, was driven by production of ROS [66]. More in-depth investigations have uncovered the specific role of NADPH oxidase 2 (Nox2)-produced ROS in upregulating NOTCH signaling to facilitate differentiation towards arterial endothelial cells [417]. Another group showed a biphasic regulation of EC fate specification via a HIF1α-mediated pathway in mESCs. Hypoxia led to upregulation of the transcription factor Etv2 in the early stages of differentiation, which resulted in development of endothelial progenitor cells. Continued exposure to hypoxia led to HIF1α-induced upregulation of

Notch1 signaling and formation of functional arterial endothelial cells, with the capacity to contribute to revascularization of ischemic tissues [418]. Similarly, HIF1α also induces the differentiation of peripheral blood mononuclear cells into EPCs, and hypoxia stimulates the further differentiation of EPCs into mature ECs [419, 420].

All of these findings highlight the significance of O2 tension as a critical parameter to control vascular differentiation of pluripotent or multipotent cells. Although some of these studies suggest contrary hypotheses, the importance of O2 considerations in cell culture environments cannot be overstated, as O2 tension can be manipulated to prevent spontaneous differentiation of pluripotent cells and to enhance the efficiency of the differentiation into EPCs and mature ECs.

3.4 Vascular responses to ECM

In the human body, vascular cells are surrounded by diverse components of the ECM, the unique spatial and temporal distribution of which affects GF availability and matrix properties which, in turn, regulate vasculogenesis and angiogenesis. Just like oxygen tension, which varies throughout vascular development, ECM components are also uniquely distributed; for example, hyaluronic acid (HA; also known as hyaluronan) levels were found to be highest during embryogenesis and to be replaced by fibronectin and then collagen, which remains abundant throughout adulthood. In this section we will discuss ECM distribution and its effects on vascular development and maintenance. Then, we will discuss various ECM components that affect vascular morphogenesis. Lastly, we will describe strategies for manipulating the ECM using synthetic biomaterials and emerging technology.

3.4.1 Types of ECM found participating in vascularization

The ECM surrounding blood vessels contributes significantly to their diverse functions and complexity. This ECM diversity encompasses different vascular development periods (i.e., embryonic versus adult) and specialized vessels at various locations in the body (i.e., capillary, arteriole, and venule) or tissues in the body (i.e. heart, kidney, lung, etc). During early vascular development, the ECM provides informational cues to the vascular cells, thus regulating their

80 differentiation, proliferation, and migration. Fibronectin and HA, which are major components of the embryonic ECM, have been shown to be vital regulators for vascularization during embryogenesis [421]. Fibronectin, a unique glycoprotein, contains cell adhesion and heparin- binding sites that synergistically modulate the activity of VEGF to enhance angiogenesis [422].

Various lineage studies have found developmental abnormalities in embryonic hearts and vessels in fibronectin-null mice, suggesting its crucial role in mediating EC interactions [423, 424]. The levels of hyaluronan, a nonsulfated linear polysaccharide, are greatest during embryogenesis and then decrease at the onset of differentiation [425], where it plays a crucial role in regulating vascular development [426]. Hyaluronan and its receptor, CD44, have been shown to be essential in the formation and remodeling of blood vessels [426-428]. We have previously reported that a completely synthetic HA hydrogel can maintain the self-renewal and pluripotency of hESCs [429,

430]. Interestingly, when VEGF is introduced into the culture media, this unique HA microenvironment can direct the differentiation of hESCs into vascular cells, as indicated by positive staining for α-smooth-muscle-actin and an early stage of the endothelial cell marker,

CD34. More recent studies have employed higher throughput methods to explore the effects of

ECM composition on EC fate. As the ECM of the developing embryo consists of multiple components, culturing ECs on combinatorial ECM arrays revealed optimal conditions for EC survival, in response to low O2 and low nutrient availability [431], as well as enhancements in EC fate which were regulated by ECM composition, at least partially through upregulation of integrin

β3 and its associated signaling pathway [145].

In contrast, the adult ECM consists mostly of a laminin-rich basement membrane, which maintains the integrity of the mature endothelium, and interstitial collagen I, which promotes capillary morphogenesis [432]. Although collagen I is present during development, its role

81 becomes increasingly important in postnatal angiogenesis, after its reactive groups have been cross-linked to further stabilize the interstitial matrix [433]. EC integrins, which interact with collagens and fibrin, are key receptors in EC activation, proliferation, and tubular morphogenesis.

The collagen-I-mediated activation of Src and Rho, and the suppression of PKA, promote the formation of prominent actin stress fibers, which mediate EC retraction and capillary morphogenesis. Moreover, the activation of Src also disrupts VE-cadherin from cell junction and cell-cell contact which, in turn, facilitates multicellular reorganization. Conversely, basement membrane laminin-1 is responsible for maintaining the mature endothelium. During the proliferative stage of morphogenesis, the laminin-rich basal lamina is degraded, exposing the tips of sprouting ECs to the underlying interstitial collagens and activating signaling pathways that drive cytoskeletal reorganization and vascular morphogenesis. This sharp difference in how ECM components affect capillary morphogenesis is responsible for controlling the delicate balance between vascular sprouting and maturation.

Once nascent vessels are formed, ECM components regulate their maturation and specialization into capillaries, arteries, and veins. Capillaries, the most abundant vessels in our body, consist of ECs surrounded by pericytes and basement membrane. Exchanges of nutrients and oxygen occur through diffusion between blood and tissue in these regions, due to the capillary’s thin wall structure and large surface-area-to-volume-ratio. Maturation of the vessel wall involves the recruitment of mural cells, development of the surrounding matrix, and organ-specific specialization [236]. ECM distribution in various tissues dictates the specialization of these capillaries to support the functions of specific organs. The capillary endothelial layer is continuous in most tissues (e.g., muscle), while it is fenestrated in exocrine and endocrine glands (e.g., kidney and pancreas). Moreover, the enlarged sinusoidal capillaries of the liver, spleen, and BM are

82 discontinuous, allowing increased exchange of hormones and metabolites between the blood and the surrounding tissues. In contrast, where the excess exchange of molecules is not desirable, such as at the blood-brain-barrier and the blood-retina-barrier, the interendothelial connection is further reinforced with tight junctions, such as occludin and ZO-1 [434].

Compared with capillaries, arterioles and venules have an increased coverage of mural cells and ECM components. Arterioles are completely surrounded with vascular SMCs that form a closely packed basement membrane. The walls of larger vessels are composed of three layers: the tunica intima, the tunica media, and the tunica adventitia. The EC layer of blood vessels is anchored to a basement membrane, which is the major component of the tunica intima [435]. The basement membrane contains network-organizing proteins, such as collagen IV, collagen XVIII, laminin, nidogen, entactin, and the proteoglycan perlecan. The tunica media contains vascular

SMCs (v-SMCs) and elastic tissue composed of elastin, fibrillins, fibulins, emilins, and microfibril-associated proteins. The tunica adventitia contains fibroblasts and elastic laminae and has its own blood supply, known as the vasa vasorum [435]. SMCs and elastic laminae contribute to the vessel tone and regulate vessel diameter and blood flow. This generic blood vessel architecture is modified with various ECM components to fulfill their individual tasks. Arteries, which function to deliver oxygenated blood, usually have a thick tunica media with numerous concentric layers of v-SMCs, whereas veins have a thick tunica adventitia layer enriched in ECM components with elastic properties, such as elastin and fibrillin.

As described, the composition of the ECM is inherently dynamic throughout development as well as vascular regeneration, positing the importance of remodeling and deposition of new

ECM as these processes progress. Additionally, stability of mature vessels requires a different

ECM composition than developing or regenerating vasculature. Several studies have highlighted

83 these changes in ECM deposition and have identified regulators of these important mechanisms.

Much of the work to date has established the role of perivascular cells, including pericytes and smooth muscle cells, in ECM production [436]. Crucially, ECs also produce ECM as blood vessels form. Of particular interest, endothelial progenitor populations and mature ECs produce ECM differently; EPCs produce collagen IV, fibronectin and laminin, while mature ECs have limited

ECM production in standard cell culture conditions. However, when subjected to hypoxic conditions, mature ECs adopt an ECM secretome similar to the pro-regenerative EPCs, wherein they secrete collagen IV, fibronectin, and laminin at low O2 (1%). At moderate hypoxia (5% O2), both cell types produce collagen I [437]. When developing engineered vasculature, these factors are critical to consider to obtain mature, long-lasting blood vessels, as ECM composition is an important parameter governing vascular stability.

3.4.2 Properties of the ECM that affect vascular morphogenesis

Recent decades have vastly expanded our understanding of how ECM properties affect vascular assembly, primarily due to newly available, well-defined in vitro models. The most common models are cultures of ECs in gels made of different ECM components, such as collagen, fibrin, fibronectin, and Matrigel. These ECM components contain instructive physical and chemical cues that direct vascular morphogenesis, which involves several steps: (i) proteolytic degradation of basement membrane proteins by both soluble and membrane-bound matrix metalloproteinases (MMPs); (ii) cell activation, proliferation, and migration; (iii) vacuole and lumen assembly into a tube with tight junctions at cell-cell contacts; (iv) branching and sprouting;

(v) synthesis of basement membrane proteins to support the formation of capillary tube networks; and (vi) tube maturation and stabilization by pericytes. These complex processes require a delicate balance between various immobilized and soluble GFs, as well as endothelial and perivascular cell

84 interactions. Gels made from ECM components, engineered to have properties resembling those of native tissues, have been widely explored as a tool to study the molecular regulation underlying vascular development [323] and as a scaffold to transplant vascular progenitor cells [438-440].

However, their manipulation for vascular tissue engineering has been narrowly limited by their inherent chemical and physical properties. Therefore, a great need exists to chemically modify these ECM components [441, 442] or to utilize biomaterials to form scaffolds from hydrogels, which are xeno-free and instructive for vascular tissue engineering [238]. Hydrogels are crosslinked polymer networks which can store a large amount of fluid and which have biophysical properties similar to many soft tissues [443]. Hydrogels can be engineered from natural biomaterials (including ECM components), artificial protein polymers, self-assembling peptides, and synthetic polymers to form scaffolds which mimic the native ECM. For example, dextran and chitosan, natural biomaterials with similar structures, do not possess any inherent cross-linking ability [444, 445]. However, a simple chemical modification, such as introducing double bonds into the repeating unit, allows the cross-linking of these polysaccharides to form hydrogels.

Alginate is another natural material which can be physically cross-linked by adding cations (e.g.,

Ca2+ or Mg2+) [446]. Another approach utilizes a purely synthetic polymer, like polyethylene glycol (PEG) or poly-[lactic-co-glycolic acid] (PLGA), whose physical and chemical properties can be easily manipulated. A simple modification can turn PEG, a cell-resistant material, into an instructive scaffold designed to promote vascularization [447-450]. Furthermore, the synthetic material of choice must be biodegradable and biocompatible, and such physical properties as pore size, degradation kinetics, and matrix mechanical properties must be easily tunable to favor vascular morphogenesis. Bioactive molecules — like GFs, cell adhesion motifs, such as arginine- glycine-aspartic acid (RGD), and MMP-sensitive peptides — must be presented with correct

85 spatial and temporal distributions within the synthetic biomaterials. Next, we will discuss several strategies for manipulating the chemical and physical properties of synthetic biomaterials.

3.4.2.1 Cell Adhesion Regulates Neovascularization

In order to support vascular cells and instruct them to undergo vascular morphogenesis, synthetic biomaterials must first be able to provide cell adhesion. Instead of incorporating ECM components to make such materials bioactive, certain synthetic peptides important for vascular morphogenesis can be incorporated into these inert synthetic materials. The most common template is the integrin-binding domain of fibronectin, RGD [451], and the laminin-derived peptide IKVAV [452]. The first crucial step in vascular morphogenesis occurs when vascular cells utilize integrin receptors to sense their surrounding microenvironments. Integrins are transmembrane receptors which not only maintain cell adhesion to ECM, but also control cell proliferation, migration, differentiation, and cytoskeletal organization. Since blood vessels must be able to assemble in diverse tissue environments (e.g., adult versus embryo and muscle versus kidney), which have different distributions of ECM components (as discussed in the previous section), it is evident that both β1 and αv integrins can support vascular morphogenesis. For example, αvβ3 and α2β1 integrins associate with vascular morphogenesis in collagen-rich ECM, like adult tissue, while α5β1 and α6β1 integrins involve fibronectin- and fibrin-rich ECM, like in embryonic tissue and healing wounds [432]. The binding of integrins onto RGD triggers several downstream signaling events mediated by Rho GTPase, particularly Rac1 and Cdc42 [323].

Extensive work by Davis and his colleagues revealed the molecular mechanism that regulates this

EC morphogenesis in fibrin and collagen. This mechanism has also been observed and controlled in synthetic (HA-based) hydrogels [219].

To further substantiate the role of cell-ECM interactions, particularly those mediated by

86 integrin engagement, several groups have identified the importance of integrin specificity in vascular regeneration. In tumor vessels, αvβ3 is preferentially expressed, leading to formation of new, albeit disorganized, leaky vasculature [453]. In order to establish organized, mature neovessels, engagement of α3/α5β1, rather than αvβ3, was necessary [91]. While RGD peptides facilitate cell adhesion in synthetic matrices, it is important to consider the non-specific integrin engagement potential of these peptides, which may influence vascular regeneration.

The number of RGD adhesion sites and the method of their presentation to the vascular cells are also crucial in affecting cell migration [454] and vascular morphogenesis [455]. Using an in vitro angiogenesis model, Folkman and Ingber were able to show that, when cultured on a moderate coating density that only partially resisted cell traction forces, ECs could retract and differentiate into branching capillary networks [455, 456]. High ECM density was saturated with

RGD adhesion peptide, which allowed the ECs to spread and proliferate, while low ECM density resulted in rounded and apoptotic cells. Interestingly, in medium ECM density, with the appropriate RGD adhesion peptide, ECs collectively retracted and differentiated into branching capillary networks with hollow tubular structures. It is evident that the ECs exerted mechanical forces on the surrounding ECM to create a pathway for migration and branching in forming vascular structures [457]. Hence, both the quantity of RGD peptide and the method of presentation within the engineered synthetic biomaterials determine the initial morphogenetic events in angiogenesis.

3.4.2.2 Scaffold degradation regulates vascular morphogenesis

Scaffolds made from ECM components, like collagen and fibrin gels, contain proteolytic degradable sequences which can be degraded by the MMPs and other proteases (e.g. cathepsins) secreted by vascular cells. This cell-mediated degradation controls both structural integrity and

87 temporal mechanical properties, which dictate the presentation of chemical and mechanical cues at various stages of angiogenesis. However, the degradation kinetics of these ECM-based scaffolds is determined by their inherent cross-linking density which, in turn, limits their manipulation for vascular tissue engineering. In contrast, synthetic biomaterials can be engineered to have degradation profiles ranging from days to months, in order to suit the specific needs of the engineered vascularized tissue constructs [445]. The polymer backbone can be cross-linked using a nondegradable cross-linker that provides structural integrity and/or a degradable cross-linker that allows directed cell migration and vascular morphogenesis. Hydrolytic degradation by the body fluid can break down the ester bonds within the polymer backbone, allowing tissue infiltration over time [444, 445]. MMP-sensitive peptides can also be used to cross-link hydrogels, allowing cell-mediated degradation, leading to a rapid response of vascular growth. Overall, by adjusting the percentages of nondegradable and degradable cross-linkers, scaffold degradation can be tuned to allow cellular infiltration, lumen formation, and ECM synthesis and distribution.

In order for the intracellular vacuoles to coalesce into a lumen, ECs require adhesive ligands for traction [238] and utilize membrane-type-1-MMPs (MT1-MMPs) to create physical spaces which facilitate the directed migration of cells to align with neighboring cells [457-459].

Therefore, ECs can only invade this synthetic scaffold if the minimal pore size is larger than the cell diameter (e.g., a soft self-assembling peptide) [460] or if the scaffold bears an MMP- degradable sequence [166]. The Hubbell research group has pioneered this approach by incorporating an MMP-degradable sequence as a cross-linker into PEG scaffolds to promote vascular healing and therapeutic angiogenesis [461, 462]. When grafted in vivo, ECs were able to invade, remodel, and vascularize this MMP-sensitive scaffold [461, 463]. Using concepts from this work, synthetic (HA-based) biomaterials utilized spatial control of degradation through

88 photopatterning to organize vascular morphogenesis [274]. Hence, incorporating MMP- degradable peptides is essential for directing vascular morphogenesis in 3D synthetic biomaterials.

3.4.2.3 Physical orientation of the ECM

The native ECM provides an instructive template for ECs and perivascular cells to orient, interact, and organize into tubular structures. Studies have demonstrated that a stable vasculature could be achieved by co-transplantation of ECs and perivascular cells, such as MSCs or SMCs

[438, 439, 464-466]. Recent studies showed that engineering a stable vascularized tissue construct requires the triculture of ECs, fibroblasts, and tissue-specific cells, such as cardiac or skeletal muscle cells [465, 467-469]. Perivascular cells, such as fibroblasts, stabilize the developing vascular tube through physical support, by differentiating into v-SMCs and wrapping around the nascent tube [470, 471], and chemical support, by secreting Ang-1, PDGF-BB, and tissue inhibitor of metalloproteinase-3 (TIMP-3) [472, 473]. These perivascular cells are also responsible for laying down ECM components in early embryogenesis and continue to do so throughout adulthood. Many studies using fibroblast-derived matrices have further revealed the 3D complexity of these ECM networks [474-476]. A study by Soucy and Romer showed that fibroblast-derived matrix alone is sufficient to induce HUVECs to undergo vascular morphogenesis, independent of any angiogenic factors. Further analysis of protein colocalization suggested that fibronectin with a distinct structure and organization was uniquely distributed among other secreted matrix components, such as collagen, tenascin-C, versican, and decorin. Cell matrix adhesions and MT1-MMP activities were reported to orient and localize within this fibrous fibronectin, which is indicative of integrin-mediated vascular morphogenesis [477]. In fact, ECs initiate neovascularization by unfolding soluble fibronectin and depositing a pericellular network of fibrils that serve as a structural scaffolding on a mechanically ideal substratum for vessel

89 development [478]. We have studied how such fibronectin organization influences endothelial tube formation by patterning fibronectin on cell culture surfaces to optimize vasculogenic potential and understand how microstructure influences vascular tube formation [479]. Alignment of other important ECM proteins, such as collagen, has also been shown to guide vascular regeneration by enhancing EC organization and migration [480]. Interestingly, similar effects in vascular organization are observed when tensile forces are incurred upon vascular fibrin-based constructs, where vascular network alignment was induced by application of force. Aligned microvasculature was shown to enhance vascular integration upon implantation in abdominal muscle [481]. This last study suggests a potential mechanism for organization of ECM components to guide vascular network organization through force-induced remodeling. It is likely such a mechanism is coupled with ECM degradation and secretion of new ECM components to establish a microenvironment amenable to the formation of new blood vessels.

The unique orientation, organization, and nanotopography of fibrous fibronectin represent features that can be integrated into synthetic scaffolds. Synthetic polymers, like PLGA and polycaprolactone (PCL), can be electrospun to produce various fiber sizes with micro- to nanoscale features that resemble fibrous fibronectin. We previously showed that surface nanotopography enhanced the formation of capillary-like structures (CLSs) in vitro [482]. Growing EPCs on grooves that were 600 nm wide reduced their proliferation and enhanced their migration without changing the expression of EC markers. Moreover, after six days of culture, the EPCs organized into superstructures along the nanogrooves, in significant contrast to the EPCs grown on planar surfaces. The addition of Matrigel further induced the formation of CLSs, with enhanced alignment, organization, and tube length compared to a flat surface. This underscores the increasingly important role of nanotopography in guiding and orienting vascular assembly. When

90 integrated into the tissue-engineered construct — for instance, using filamentous scaffold geometry [483] and micropatterning [295, 484, 485] — the orientation and structure of the engineered vasculature can be controlled.

3.4.2.4 Regulating matrix mechanics

It has become increasingly evident that the biomechanical properties of the ECM, such as matrix orientation and mechanics, profoundly influence the control of vascular morphogenesis.

Due to their versatility with respect to mechanical properties (e.g., cross-linking density, pore sizes, and topography), synthetic biomaterials have powerful features that can be exploited to further direct vascularization. Changes in ECM mechanics can lead to changes in GF availability

[442, 455], drive capillary morphogenesis [486], and stimulate angiogenesis in vivo [487]. By altering matrix adhesive characteristics and mechanics, Ingber and Folkman illustrated how bFGF- stimulated ECs can be switched between growth and differentiation during angiogenesis [455].

Recently, biomechanical cues from the ECM and signals from GF receptors have been implicated in regulating the balance of activity between TFII-I and GATA2 transcription factors, which govern the expression of VEGFR2 to instigate angiogenesis [488]. Matrix stiffness not only regulates the cell’s response to soluble GFs, but also cell morphogenesis during angiogenic sprouting. Primarily due to MMP activity, the tip of a new capillary sprout becomes thinner, locally degrading the basement membrane proteins. This region, with its high rate of ECM turnover and thin basement membrane, becomes more compliant and stretches more than the neighboring tissue.

Consequently, the decrease in matrix stiffness changes the balance of forces across the cell integrin receptors, increases cell tension, and results in cytoskeletal arrangement to form branching patterns that are characteristic of all growing vascular networks [486].

The pioneering work by Deroanne et al. showed that a decrease of matrix stiffness

91 increased capillary branching and the elongation of tubes. A reduced tension between ECs and

ECM, accompanied by a profound remodeling of the actin-FAP complex, is sufficient to trigger an intracellular signaling cascade leading to tubulogenesis [489]. This observation has been further confirmed in collagen gels [489, 490], fibrin gels [491], self-assembling peptides [460], and HA- gelatin hydrogels [152].

Although ECM-based gels, such as collagen, fibrin, and Matrigel, have been widely used in angiogenesis assays, their inherent physical properties have limited their usage when studying the effects of matrix mechanics on angiogenesis. The stiffness of ECM-based gels can be increased either by increasing their concentration, which also alters their ligand and fibril density [492], or by altering the cross-linking of ECM proteins in a narrow range using a microbial transglutaminase

[493]. Therefore, examining the effects of matrix stiffness alone on angiogenesis requires the use of synthetic hydrogels, the stiffness of which can be easily adjusted over a wide range of moduli without altering other chemical properties. Unlike naturally available ECM-based gels, the elasticity of which is limited to their inherent cross-linking density, synthetic HA-hydrogels can be used to study a physiologically relevant range of matrix elasticity [152]. When the cross-linking density of the HA-gelatin hydrogels was further reduced, the matrix elasticity became relatively compliant, resulting in an increase of capillary branching, elongated tubes, and enlarged lumen structures [152]. On a relatively compliant matrix, EPCs can produce fewer MMPs than a stiffer matrix would require and still degrade, exert mechanical tension on, and contract the matrix to enable vascular morphogenesis. On the other hand, EPCs must produce more MMPs on a stiffer matrix, to overcome the extra mechanical barriers; even then, this local decrease in substrate stiffness cannot support vascular morphogenesis [152]. This model also explains the rapid appearance of large functional vessels in granulation tissue, as a response to the wound-healing

92 mechanism [487].

In addition to the effects of matrix stiffness on post-natal vascular regeneration, matrix stiffness has been probed as an important regulator of stem cell fate. Beginning with the pioneering work of Engler et al. [151], studies examining the effects of substrate stiffness and mechanical signaling transduction pathways on stem cell fate have proven instrumental in enhancing our collective knowledge of differentiation schema. To this end, our group has shown that substrate stiffness can govern EC fate through alterations in mesodermal precursors. Similar to the enhancements in EC fate observed upon culture in low O2 environments [66], compliant substrates enhance mesodermal differentiation, which results in robust EC differentiation [32].

A recent illuminating study identified stress relaxation as an important, yet understudied, regulator of mechanical signal transduction. Specifically, in alginate-based hydrogels with the same matrix stiffness and pore size, altering stress relaxation modulated MSC cell fate [162].

While studies of the effect of stress relaxation on EC fate and vascular morphogenesis have not been published, stress relaxation is an important parameter to bear in mind for biomaterials design, particularly because covalently crosslinked hydrogels do not exhibit stress relaxation behavior similar to that of the native ECM.

These studies underline the importance of engineering a tissue construct with a matrix stiffness amenable to promote in vivo vascularization. However, investigating how matrix stiffness may affect in vivo vascularization remains challenging due to the complexity of the system, which involves matrix remodeling, host capillary ingrowth, as well as anastomosis of the vascular construct and contributions from other cell types. For example, in vivo vascular ingrowth into

Matrigel scaffolds was found to be optimal at intermediate matrix stiffness, in sharp contrast to the observed in vitro ingrowth [488]. Elegant work by Yoder’s research group also found that

93 increasing the collagen concentration yielded stiffer scaffolds, which in turn promoted host capillary ingrowth in vivo. Compared to stiffer scaffolds, softer scaffolds might have experienced excessive in vivo remodeling and failed to retain the vascular constructs. Moreover, in vitro angiogenesis studies have found that ECM-based gels produce a much narrower range of stiffness

[440, 488] than synthetic hydrogels [152, 460, 488]. Future investigations are needed to evaluate vascularization by both the host capillary and the engineered vascular construct over a wider range of physiologically relevant matrix elasticities. Despite the differences in scaffold composition

(ECM-based gels versus synthetic hydrogels), culture conditions (in vitro versus in vivo), assay type (2D versus 3D), and ranges of matrix stiffness, all of these studies highlight the relevance of engineering scaffolds with mechanical elasticity suited to the specific needs of tissue vascularization.

3.5 The Effects of Oxygen Availability and the ECM

In this section, we will consider O2 tension and the ECM as two interdependent factors determining the efficiency of vasculature formation. We will review currently available O2 measurement techniques and challenges, along with the mathematical modeling approaches used to overcome some of these challenges in describing O2 gradients in 3D environments. Then, we will discuss cellular adaptations and responses to O2 availability in 3D ECM constructs and the possible outcomes of variations in O2 distribution in 3D cultures of vascular cells.

3.5.1 Varying oxygen tensions in the ECM of tissue and matrix scaffolds: measuring/modeling

3.5.1.1 Oxygen measurement techniques and challenges

Manipulation of oxygen, in order to direct pluripotent or vascular cells to form blood vessels, requires knowing the precise O2 tension that the cells are exposed to under varying

94 conditions. Many different O2 measurement techniques have been used in vitro and in vivo. The accuracy of these measurements is fundamental to confidently describe the cellular responses under various O2 availabilities, as well as to controlling the O2 tension in order to direct angiogenesis and vasculogenesis. An O2 measurement method needs several properties to be considered superior, including accuracy, sensitivity, repeatability, rapidity, and noninvasiveness.

Although some methods are used more commonly in a broader range of applications, no ‘gold standard’ exists for all applications, since the method chosen usually depends on the purpose of the measurement. In vivo O2 measurement methods can be divided into two main categories: (i) direct measurements, where the concentration or the partial pressure of O2 is directly measured, and (ii) indirect measurements, where levels of O2-indicative molecules (e.g. hemoglobin, cytochrome) are detected and correlated to relative O2 concentrations.

The most common direct measurements are electrodes, phosphorescent probes, electron paramagnetic resonance (EPR) oximetry, and nuclear magnetic resonance (NMR). Some of the indirect measurement methods involve monitoring of hemoglobin/myoglobin, mitochondrial cytochromes, and NADH/FADH [494]. Springett’s paper thoroughly reviews the benefits and limitations of the most recent methods [494].

In addition, in vitro studies have applied these currently available methods to monitor O2 levels quantitatively, such as by measuring O2 tensions at the cellular level in 2D monolayer cell cultures or O2 gradients in 3D gels or scaffolds. Two major methods used to measure O2 levels during in vitro cultures are polarographic and fluorescence quenching techniques. The latter has been shown to surpass the polarographic technique, which consumes O2 during the measurements

[495]. When an implemented measurement technique, like the polarographic technique, consumes

O2, it more likely generates even greater inaccuracies and leads to incorrect conclusions in low O2

95 environments, as occurred in studies investigating the effect of hypoxia in 3D scaffolds [496-498].

Fluorescence quenching technology is available both for invasive applications, using an electrode probe with a very thin (approximately 5 µm) tip, and for noninvasive applications, using a sensor patch composed of a ruthenium-based metal complex that can be excited by an external fluorescent light source.

3.5.1.2 Modeling oxygen transport in tissues

The limitations of these measurement techniques, caused mostly by the difficulties in measuring spatial O2 concentrations in tissues or scaffolds, raise a need for predictive mathematical models. Transport of O2 in vivo is controlled by several parameters, including blood flow rate, degree of vascularization in the tissue, physiological distance of the cells from the microvasculature, and, depending on cell type, the cells’ rate of O2 consumption. These factors affect O2 distribution in the tissue, and some can also have an impact on O2 transport in 3D in vitro cultures of pluripotent or vascular cells. Additional factors that in vitro studies should consider are the geometry of the scaffold, the available surface area for O2 transport from the environment to the system, and controlled dissolved O2 levels in the culture media.

In general, fundamental mathematical models estimating O2 distribution in 3D constructs can be classified into: (i) static models, where O2 is only transported via diffusion, and (ii) dynamic models, where convectional transport of O2 is also incorporated using perfusion systems, such as microfluidic devices, or microcirculation in the tissues.

3.5.1.2.1 Static Models

In tissues cultivated under static conditions within 3D scaffolds, using different types of biomaterials, spatial O2 concentration can be defined with a one-dimensional (1D), unsteady state species continuity equation:

휕퐶 휕퐶2 푂2 = 퐷 푂2 − 푅 (Eq. 1) 휕푡 푂2 휕푧2

Where CO2 is the spatial O2 concentration in the scaffold changing with time (t) and axial position

(z), DO2 is the diffusion coefficient of O2 in the scaffold material, and R is the oxygen consumption rate of cells. This form of the transport equation has been used in many studies attempting to predict the O2 gradients in 3D scaffolds [499-501]. The equation implies that O2 changes both with time and depth, while being consumed by the cells as it diffuses from the environment into the scaffold. Boundary conditions, which are critical for O2 distribution, depend on the O2 equilibrium between the environment (media/air) and the boundaries of the scaffold. Therefore, for a 3D scaffold with a depth of L and open boundaries from both sides, the boundary conditions can be given as:

At z=0 and z= L, CO2=S.PO2 (Eq. 2)

Thus, the solubility (S) of O2 in the scaffold material is one of the determining parameters of O2 distribution. Although the diffusion coefficient can also be considered a critical factor in relatively stiff scaffolds of the sort usually used for cartilage and cardiomyocyte tissues [499], it has been shown to be less significant for the natural hydrogel scaffolds commonly used for vascular tissues, such as collagen and HA. For instance, the diffusion coefficient of O2 in collagen gels was found to be 99% of that in water [502]. Therefore, modeling studies usually assumed that it has the same

-5 2 O2 diffusion coefficient as water or cell media (3.3x10 cm /s at 37 °C) [503]. The consumption rate of O2 (R) given in Eq.(1) is a function of both CO2 and ρcell and is governed by the Michaelis-

Menten equation, which states that the O2 uptake rate of each cell increases with O2 availability, reaching a maximum at a point, Vmax:

푉푚푎푥퐶푂2 푅 = 휌푐푒푙푙 (Eq. 3) 퐾푚+퐶푂2

Where Km is the O2 concentration at which the O2 uptake rate is half of its maximum value and

ρcell is the cell density as a function of time and position. Different groups have reported the Vmax and Km parameters of many vascular cell types at various cell seeding densities [396, 504]. For

6 example, the Vmax and Km of HUVECs, at a density of 1x10 cells/ml, are found to be 22.05 ± 1.92

(pmol s-1 10−6 cells) and 0.55 ± 0.02 (µM), respectively [396]. It should be noted that these parameters are estimated as to mitochondrial consumption of O2. However, as already discussed,

ECs also consume O2 for ROS production, and the theoretical models should also take this additional O2 consumption into account by, for example, including a linear correlation in the O2 consumption rate equation (Eq. 3). Besides, all estimations of the Vmax and Km parameters for the

O2 consumption of different cell types are carried out in 2D cultures. The literature currently lacks studies investigating whether or not, depending on the composition of the extracellular matrix, encapsulating cells in 3D gels changes their consumption of O2.

Vascular cells proliferate, die, migrate, and assemble during 3D cultivation, which affects their spatial and temporal density and, therefore, O2 distribution. Models developed for 3D cultures of cardiomyocytes take into account the cellular proliferation and changes in the dimensions of the cells during nutrient transport in scaffolds [501, 505]. However, we need more detailed models, that consider how capillary formation affects O2 transport, to achieve more reliable estimations of spatial O2 concentration. Tube formation and the networking of ECs in 3D gels have been simulated by more complicated numerical models [506, 507], although the effects of O2 concentrations on tube formation dynamics still need to be incorporated.

3.5.1.2.2 Dynamic and in vivo models

The models used for static cultures in 3D scaffolds can also be used to describe O2 distributions in vivo when combined with a fluid perfusion model that considers the convectional

O2 transfer to the tissues. The velocity profile of a fluid in capillaries or in an engineered microchannel system can be calculated using the simplified Navier-Stokes equation with cylindrical coordinates given for a laminar, one-dimensional, steady-state, and fully developed flow of an incompressible fluid:

푑푃 1 푑 푑푉 = 휇 [ (푟 푧)] (Eq. 4) 푑푧 푟 푑푟 푑푟

Where P is the total pressure in the fluid changing in an axial direction, µ is the viscosity of the fluid, and Vz is the axial velocity of the fluid changing in a radial direction. After estimating the blood velocity profile, the species continuity equation, which involves both diffusional and convectional transfers of O2, can be used to obtain the O2 distribution inside the capillary or microchannel:

휕퐶 1 휕 휕퐶 휕2퐶 푉 푂2 = 퐷퐵푙표표푑 [ (푟 푂2) + 푂2] (Eq.5) 푧 휕푧 푂2 푟 휕푟 휕푟 휕푧2

The technical difficulties of making quantitative O2 measurements in BM have led many researchers to develop mathematical models to describe BM O2 distribution [364, 365, 508].

Additional parameters that need to be considered in vivo are the vascularization of the tissue and the transport of O2 via hemoglobin proteins, making the concentration of hemoglobin another essential factor for determining the oxygenation of the tissue. Studies take these additional factors into account using the following equation:

휕 퐷푗 [∇2퐶푗 ] = 푉 [퐶푗 + 푁휑 ] (Eq.6) 푂2 푂2 푧 휕푧 푂2 푂2

The superscript j denotes each sinusoid/arteriole around the BM. N is the O2 carrying capacity of the blood and is the concentration of O2 bound to hemoglobin, which depends on the plasma O2 concentrations [509]. Finally, spatial and temporal CO2 in tissue can be estimated in a similar manner to the in vitro models, using Eq. (1) incorporated with the continuity of fluxes assumption at the ECM-vessel interface as a boundary condition.

3.5.2 Targeted cellular responses to O2 availability in matrix hydrogels

Engineering vascular tissues in a 3D ECM is well-orchestrated process combining proliferation, apoptosis, migration, activation, and assembly of vascular or precursor cells inside the construct. As discussed in the previous section, the composition of the biomaterial used to encapsulate the cells is critical for cellular fate and vessel formation. In addition to the effects of the chemical and physical properties of the ECM material on blood vessel formation, temporal and spatial levels of O2 and other nutrients are also crucial for various targeted cellular responses. A number of studies have investigated the effects of matrix content and stiffness on angiogenesis/vasculogenesis [486], and many others have proposed using different types of biosynthetic materials to develop more precise blood vessels [238, 345]. However, only a few studies have highlighted how O2 gradients occurring in the matrix contribute to the angiogenic process [348, 349, 397]. The availability of O2 and other nutrients decreases at the center of the gel compared to the periphery, especially in engineered vascular tissues, whichrequirement a high cell seeding density for sufficient vascular tissue generation or repair. Hence, cells that reside along various layers of the matrix respond differently to the nonuniform distribution of O2 and nutrients.

For primary vascular cells to form blood vessels, they require survival, activation, and the induction of angiogenesis by GFs, cell signaling, and regression. All of these responses, necessary for blood vessel formation, are controlled by ECM properties, as well as by O2 availability.

100

Therefore, the influences of both the ECM and dissolved O2 distribution should be considered simultaneously.

Cell assembly and tube formation in the ECM require a sufficient cell density. Deprived of

O2 and nutrients, vascular cells can undergo apoptosis or necrosis [329]. These two cellular death mechanisms should be distinguished; apoptosis contributes to the process of angiogenesis at any

O2 tension, whereas necrosis usually results in the collapsing and deformation of tubes [510]. The critical issues to consider to prevent cellular necrosis during 3D vascular cell cultures are the permeability of the ECM material to O2 and glucose, the cell seeding density, and the thickness of the gel. Thus, cell seeding density is constrained by an upper limit, above which the cells undergo necrosis due to nutrient deprivation, and a lower limit, below which the cells cannot assemble sufficiently to form tube-like structures. Both limits depend on the equilibrium O2 levels in the environment.

Cells may also die as a result of apoptosis after their encapsulation in the gel. Interestingly, some groups have demonstrated that programmed cellular death is necessary for angiogenesis/vasculogenesis [510]. Segura et al., having studied tube formation of ECs in both 2D

Matrigel and 3D collagen, concluded that a considerable number of cells undergo apoptosis at the initial stages of cultivation and that, once angiogenesis is induced and tube formation has started, no further apoptosis occurs throughout the process. Inhibition of proapoptotic proteins has been shown to correlate with defective tube formations, suggesting that apoptosis is important for avoiding imperfections during blood vessel growth. Hypoxia, as already discussed, induces angiogenic responses and also regulates proapoptotic gene expressions. Thus, spatial variations in

O2 levels may alter the apoptotic responses in the gel and therefore regulate vascular tube morphogenesis.

101

MMPs are promoted by integrin-ligand interactions between cells and the ECM, leading to the degradation of the ECM and facilitating the migration of the cells [511]. It is hypothesized that

ECM fragmentation, orchestrated by the secretion of MMPs, can mediate caspase activity through the rebinding of ECM protein fragments to unligated integrins, namely death receptors [512].

Therefore, the survival of ECs depends on the balance between cell survival promoters, such as

FAK, Src, and Raf, and cellular apoptosis promoters, such as caspase 8 and caspase 3. Hypoxia may again play a critical role here, affecting both sides of the equilibrium, by upregulating MMPs and VEGF at the same time [329, 397]. Hypoxia, accompanied by nonuniform distribution of O2 throughout the gel, can result in spatial differences of cellular viability, which may subsequently disrupt vascular networking.

Overall, blood vessel growth requires remodeling of the ECM, which is based on two distinct mechanisms: (i) degradation of the ECM by secreted proteases, and (ii) production of new

ECM to support the invading vasculature. Many studies have shown that hypoxia can regulate the degradation, maintenance, and synthesis of the ECM [330, 513]. ECM degradation is important for cellular migration into and blood vessel invasion of tissue. MMPs, as mentioned above, are a major family of proteinases that participate in the degradation of the ECM during angiogenesis. In particular, MMP-2 and MMP-9, both members of the gelatinase subgroup of MMPs, have been shown to contribute to the process of angiogenesis [514]. MMP-2 secreted by the cells is activated through membrane MT1-MMPs where the activation can be avoided in the presence of tissue inhibitor of MMP-2 (TIMP-2) at high levels [515]. Furthermore, hypoxia was shown to influence the expression of MMP-2, as well as of MT1-MMP and TIMP-2, in ECs [329]. Lahat’s group demonstrated the upregulation of MMP-2 expression in hypoxic (0.3% O2) cultures of HUVECs,

102 whereas MT1-MMP and TIMP-2 are downregulated, enhancing migration and tube formation

[329].

ECM degradation is accompanied by ECM production and the secretion of cells. Once the quiescent state of the ECs composing the blood vessel walls is perturbed and angiogenesis is induced, ECs start to proliferate and invade the neighboring ECM by using proteinases. At the same time, they start to remodel the existing ECM by synthesizing new ECM. In healing wounds,

ECs produce transitional ECM proteins, including fibrinogen and fibronectin, and temporarily deposit them in the ECM in order to provide available ligands during vessel growth [512].

Moreover, ECs also produce such matricellular proteins as tenascin C and SPARC in the ECM to mediate angiogenesis [512]. Clearly, the new ECM synthesis of cells is crucial for angiogenesis, and hypoxia, through HIF1α, has been shown to regulate the expressions of many different types of ECM proteins [516]. For example, many in vivo and in vitro studies have shown that hypoxia enhanced the synthesis of collagen, the most abundant protein in mammalian tissues [517-519].

Moreover, proliferation of the cells during angiogenesis/vasculogenesis in 3D scaffolds is regulated by basic fibroblast growth factor (bFGF) and VEGF, which are known to be hypoxia- dependent proteins [348]. In most studies of the vascularization of 3D scaffolds, both GFs are broadly used as soluble factors that supplement cell growth medium to induce proliferation and migration [324, 520]. In addition, Shen et al. demonstrated that immobilization of VEGF into a

3D collagen scaffold promotes EC viability, proliferation, and vascularization [521]. VEGF has been shown to promote blood vessel formation, not only by inducing cellular proliferation and migration, but also by directly regulating elongation and capillary networking in 3D ECM constructs deprived of O2 and nutrients [397]. Helminger et al. used a sandwich system to seed

HUVECs inside a collagen gel [397]. The transfer of O2 and nutrients was accomplished with only

103 simple diffusion through the edges of the collagen, so that O2 and nutrient levels decreased towards the center. They found that, in a short time period (about nine hours), VEGF intensity increased in the interior regions deprived of O2, which correlates well with cell elongation and branching.

VEGF promoted capillary networking independently of proliferation, highlighting the role of autocrine VEGF in the reorganization of vascular networks in hypoxic regions of solid tumors.

Another study focusing on quantitative measurements of O2 gradients in 3D collagen also showed that increased VEGF concentrations correlated well to decreasing O2 levels throughout the 3D constructs during a ten day period of cultivation [496].

A few studies considering the induction of angiogenesis by hypoxia in 3D scaffolds have emphasized that lowering the O2 tension in 3D gels improved cellular branching and tube formation of ECs [348, 349]. As made abundantly clear throughout this chapter, both O2 and matrix mechanics act as potent upstream regulators of a variety of signaling pathways that contribute significantly to the regulation of vascular differentiation as well as morphogenesis. To better understand these processes, engineers and biologists together have built an impressive library of materials and assays to examine an array of signaling cascades guiding the formation of blood vessels. There are several interesting platforms that can be used to study the effect of hypoxia and

O2 gradients on cellular function, including hydrogels [191-193, 522] and microfabrication or microfluidic devices [190, 523-527]. Stiffness gradients can also be controlled to study ranges of viscoelasticity and their effect on cell fate, migration, and other parameters [217]. Finally, O2 and matrix mechanics can be independently controlled in the same system using gelatin and dextran- based hydrogels. These polymers can be both enzymatically crosslinked by an O2-consuming reaction to create hypoxic conditions, and crosslinked by a secondary non-O2-consuming reaction, to create a stiffer microenvironment without appreciably affecting microenvironmental O2 [194].

104

A growing body of publications have both investigated the influence of the ECM composition on angiogenesis/vasculogenesis and suggested the crucial roles of hypoxia in blood vessel formation. In addition, the evidence discussed above is sufficient to suggest that the ECM composition and O2 tension are coupled factors that need to be taken into account concurrently when developing and repairing vascular tissues in 3D microenvironments.

3.6 Future Directions

Understanding the simultaneous effects of the ECM and O2 tension on the processes of angiogenesis/vasculogenesis will enable researchers to control these two factors and thereby manipulate cellular responses in desired directions. Recent developments in many different fields of research, such as smart biomaterials and microfluidics, have made it possible to design and construct novel in vitro microenvironments for cells. Smart biomaterials have been developed that can dynamically respond to external stimuli, such as light [429], pH [528], temperature [529], and cytokines [530]; these materials can truly mimic the complexity of a native ECM environment.

The ability to control the physical and chemical properties of the gels at different spaces and times will provide better control over different stages of angiogenesis. Light-sensitive hydrogels can be used to create biomaterials with distinct cross-linking densities to promote and inhibit cell spreading and migration [531], which in turn can be used to pattern complex vascular networks.

Since vascular morphogenesis is sensitive to tissue stiffness [489] [152], orientation [482], and polarity [94, 532], researchers could also induce vascular assembly into a tube by creating elasticity, GFs, adhesion peptide, and oxygen gradients along the 3D scaffold [192, 533, 534].

The development of photodegradable hydrogels, as well as the control of cell-mediated degradation in synthetic hydrogels, whose mechanical and chemical properties are controllable during the timescale of cellular development [535], has enabled control of vascular assembly [219,

105

274]. On the other hand, creating smart biomaterials that can shrink, swell, or degrade in response to oxygen tension would also be desirable to prevent the formation of anoxic regions inside the gels. More precise temporal control of O2 gradients inside the constructs could also be beneficial to explain various phenomena taking place in the body, such as EPC regeneration in the BM and embryonic development, where the O2 gradient plays a critical role in differentiation and migration dynamics.

Fig. 3.1 illustrates two proposed approaches for controlling oxygen distribution and ECM properties. One proposal for regulating O2 gradients inside the gel would be to incorporate microfluidic technology [504, 536-538]. Although this approach provides better O2 control over

3D microenvironments, the problem of spatial variations in O2 levels throughout the gel, due to the cells’ O2 consumption, must still be addressed. Advancements in microfluidic technology could enable spatial O2 control over 3D microenvironments; for instance, the gel could be prepared around a microtube, which would supply O2 by flushing growth media containing a desired amount of O2 (Fig. 3.1A). Hence, different O2 gradients could be generated via the manipulation of O2 concentrations in the outside environment and inside the microtube.

Another method for controlling and improving O2 transport in the gel would be to microencapsulate O2 carrier liquids, such as perfluorocarbons (PFCs). Due to their high capacity to dissolve O2, PFCs have been used as a blood replacement to improve O2 delivery to tissues [539,

540]. Based on the high oxygen-carrying capacity of PFCs, Radisic et al. [540] developed a PFC- perfused system to supply sufficient levels of oxygen to 3D cardiomyocyte cultures. A study by

Chin et al. [541] made a similar attempt, developing hydrogel-PFC composite scaffolds to improve oxygenation throughout the gel. In a similar manner, taking advantage of the high O2 solubility of

PFCs, controlled release of O2 in 3D microenvironments could be improved via

106 microencapsulation of PFCs (Fig. 3.1B). Polymeric microspheres loaded with O2 have also been shown to be effective in enhancing cell viability in anoxic microenvironments [542].

The continued development of novel biomaterials technologies, that enable study of human cells in highly biomimetic settings in vitro, will continue to guide discovery of cell behavior and vascular morphogenesis. These powerful systems, coupled with continued innovation throughout all areas of biotechnology, including gene editing and stem cell technology, have positioned the field of vascular tissue engineering in a fascinating arena, where new therapeutic targets and more robust vascularized constructs continue to be discovered and translated to the clinic. Combining expertise in biology, materials science, engineering, and medicine will continue to inform our understanding of the complex cell-cell and cell-matrix interactions that drive tissue formation and regeneration.

Figure 3.1: Controlling ECM and O2 in vitro. (A) 3D gel prepared around a micro-tube supplying O2 (insert: white arrows indicate oxygen transport; blue arrows indicate the direction of airflow); (B) Microencapsulated O2 carriers, such as PFCs, embedded within 3D gel. Drawing not to scale.

107

CHAPTER 4

4 Overview of Materials and Experimental Methods

4.1 Materials

Gelatin (Gtn, from porcine skin gel strength 300, Type A; G2500), trans-4-hydroxy-3- methoxycinnamic acid (FA; 128708), N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC; E6383), N-hydroxysuccinimide (NHS; 56480), dimethyl sulfoxide (DMSO;

276855), Dextran (Dex, MW 70 000; 31390), 2-bromoethylamine hydrobromide (BEAHB; B65705), triethylamine (TEA; 471283), 4-dimethylaminopyridine (DMAP; 107700), p-nitro- phenylchloroformate (PNC; 160210), poly(ethylene glycol) (PEG, MW 4000; 81240), tyramine (TA;

T90344), anhydrous dichloromethane (DCM; 270997), diethyl ether (296082), sodium chloride

(310166), deuterium oxide (D2O; 151882), fluorescamine (F9015), laccase (lyophilized powder from mushroom, ≥4.0 U/mg; 75117), DMOG (D3695), CoCl2 (C8661), DPI (D2926), cathepsin L inhibitor I (Z-FF-FMK; 219421), CaO2 (466271), Ca(OH)2 (239232), 4′,6-diamidine-2′- phenylindole dihydrochloride (DAPI; 10236276001), bovine serum albumin (BSA; A3059), paraformaldehyde (PFA; P6148), and Hoescht Stain solution (Hoechst; H6024) were purchased from Sigma-Aldrich (St. Louis, MO) and used as obtained without purification. Laccase (976 U/ml from C. gallica) was provided by collaborator Dr. Rafael Vazquez-Duhalt, Centro de Nanociencias y Nanotecnología UNAM. Microbial transglutaminase (mTG) (Activa-TI) was obtained from

Ajinomoto Inc or mTG (TI formula) from Moo Gloo. Dulbecco’s phosphate-buffered saline

(DPBS; 14190250), trypsin-EDTA 0.05% (25300120), DQ Gelatin from Pig Skin Fluorescein

Conjugate (DQ-gelatin; D12054), Collagenase Type IV powder (17104019), Alexa Fluor 488 phalloidin (phalloidin; A12379), Isolectin GS-IB4 Alexa Fluor 568 Conjugate (I21412), Halt

Protease and phosphatase inhibitor cocktail (PI) (100×) and EDTA (78440), formaldehyde 37%

108 by weight (formaldehyde; F79), Triton X-100 (85111), Tween-20 (BP337), Donkey anti-Goat

Secondary Antibody Alexa Fluor 488 (gt488; A11055), Donkey anti-Mouse Secondary Antibody

Alexa Fluor 546 (ms546; A10036), Antibody diluent (003218), BD Lo-Dose U-100 insulin needles (28 gauge) (BD329461), Trizol reagent (15596018), and Molecular Probes CellROX

Green Reagent (C10444) were all purchased from Thermo Fisher Scientific (Waltham, MA).

Dialysis membranes (molecular mass cutoff = 3500 Da) (132724) and (molecular mass cutoff =

6000-8000 Da) (132660) were purchased from Spectrum Laboratories (Rancho Dominguez, CA).

Endothelial growth media–2 (EGM2; CC-3162) was purchased from Lonza (Walkersville, MD) and supplemented with additional characterized HyClone FBS from GE Healthcare Life Sciences

(Hyclone FBS; SH30071.03) (Logan, UT) and used to culture ECFCs (provided by M. Yoder,

Indiana University School of Medicine) on collagen I, rat tail (Col I; 354236) from Corning

(Corning, NY) coated cell culture plates. O2 (1%) was purchased from Airgas (Radnor, PA). Envy

Green fluorescent microspheres (FS06F) were purchased from Bangs Laboratories Inc. (Fishers,

IN). GM6001 InSolution (GM6001; 364206) and GM6001 (364205) were purchased from EMD

Millipore (Burlington, MA). VE-cadherin antibody (sc-9989) was purchased from Santa Cruz

Biotechnology (Dallas, TX). ICAM-1 antibody (BBA3), integrin-β2 antibody (AF1730), SDF-1α

(350-NS-010), and the Human Protease Proteome Profiler Array Kit (ARY021B) were purchased from R&D Systems (Minneapolis, MN). SMARTpool:siGENOME HIF1A siRNA (M-004018-

05-005), siGENOME Non-Targeting siRNA 1 (D-001210-01-05), and DharmaFECT2 (T-2002-

02) were purchased from Dharmacon Inc. (Lafayette, CO). Direct-zol RNA Miniprep Kits (R2052) were purchased from Zymo Research (Irvine, CA).

109

4.2 Polymer synthesis and characterization

4.2.1 Synthesis of gelatin-g-ferulic acid (Gtn–FA)

Gelatin-g–ferulic acid (Gtn-FA) was synthesized using EDC and NHS as coupling reagents. A mixture of DMSO and distilled (DI) water (1:1 volume ratio) was prepared as a solvent. Gtn (1.0 g) was dissolved in 50 ml of the solvent at 40°C. FA (0.777 g, 4.0 mmol) was dissolved in 20 ml of the solvent and reacted with EDC (0.92 g, 4.8 mmol) at room temperature for 15 min and then with NHS (0.64 g, 5.6 mmol) at room temperature for 15 min to activate the terminal carboxyl groups of FA (carboxyl/EDC/NHS = 1:1.2:1.4). The activated solution was then added to the Gtn solution, and a conjugative reaction was conducted at 40°C for 24 hours. Following completion of the reaction, the solution was dialyzed against DI water for 5 days (molecular mass cutoff = 3500

Da) and then lyophilized.

4.2.2 Synthesis of aminated dextran-g-poly(ethylene glycol)-tyramine (DexE–PT)

Aminated-dextran (DexE) was synthesized by coupling of BEAHB and Dex using TEA as a catalyst.

Briefly, Dex (1.5 g, 0.02 mmol) was dissolved in 60 ml of anhydrous DMSO at 50 °C and reacted with

TEA (6.2 ml, 44.5 mmol) for 30 min. BEAHB (4.6 g, 22.5 mmol) was dissolved in 50 ml of DMSO then added to the Dex solution and the reaction was conducted at 50 °C for 24 hours. Following completion of the reaction, the solution was dialyzed against water for 5 days (molecular weight cutoff

= 6000–8000 Da) and then lyophilized. For the synthesis of amine-reactive PEG, the terminal hydroxyl groups of PEG were activated with excess PNC. PEG (10 g, 2.5 mmol) was dissolved in 100 ml of anhydrous DCM at room temperature under a nitrogen atmosphere. DMAP (0.916 g, 7.5 mmol) and

TEA (0,759 g, 7.5 mmol) were dissolved in 20 ml of DCM and then added to the PEG solution. The

110 mixture was reacted at room temperature for 15 min to activate the terminal hydroxyl groups. The PNC solution (1.511 g, 7.5 mmol) dissolved in DCM was added dropwise to the activated PEG solution, and the reaction was conducted under a nitrogen atmosphere at room temperature for 24 hours. The molar ratio of PEG : DMAP : TEA : PNC was 1 : 3 : 3 : 3. Following completion of the reaction, the solvent was evaporated using a rotary evaporator at 30 °C. The concentrated solution was precipitated in cooled diethyl ether. The precipitate was filtered and dried overnight under vacuum to give a white powder of amine-reactive PEG (PEG–(PNC)2). Finally, DexE–PEG–TA (DexE–PT) was synthesized by coupling

DexE and TA using PEG–(PNC)2. For this synthesis, PEG–(PNC)2 (3.2 g, 0.8 mmol) was dissolved in

30 ml of anhydrous DMSO at room temperature under a nitrogen atmosphere. The TA (0.11 g, 0.8 mmol) solution dissolved in 30 ml of anhydrous DMSO and was added dropwise to the PEG–(PNC)2 solution. The reaction was performed at room temperature under a nitrogen atmosphere for 6 hours to give mono-TA conjugated PEG (PNC–PEG–TA). The molar ratio of PEG–(PNC)2 : TA was 1 : 1. The

PNC–PEG–TA solution was applied to DexE (0.25 g, 3.6 μmol) solution and dissolved in 50 ml of anhydrous DMSO. The reaction was conducted at room temperature under a nitrogen atmosphere for

24 hours. Following completion of the reaction, the solution was dialyzed against a saltwater solution

(NaCl 0.01 nM) to remove PNC salt and unconjugated molecules for 5 days (molecular cutoff = 6000–

8000 Da), then further purified by dialysis against distilled water (molecular cutoff = 6000–8000 Da) for 1 day to remove NaCl molecules. The purified solution was lyophilized to give the DexE–PT polymer.

4.2.3 Characterization of hydrogel precursor solutions

The degrees of substitution (DS) of FA and TA were measured using a UV/vis spectrometer

(SpectraMax; Molecular Devices, Sunnyvale, CA). Gtn–FA polymer (10 mg) was dissolved in 1 ml of

111 a DMSO : DI water (1 : 1) solution, and the absorbance was measured at 320 nm. DexE–PT polymer

(10 mg) was dissolved in 1 ml of DMSO, and the absorbance was measured at 275 nm. The concentrations of the conjugated FA and TA molecules were calculated from a calibration curve given by monitoring the absorbance of known concentrations of FA or TA. The availability of primary amines for mTG crosslinking was analyzed using fluorescamine. Following manufacturer instructions, polymer solutions of 10 mg/ml PBS (for Gtn–FA and DexE–PT) were measured following reaction (15 min) with fluorescamine with a fluorescence microplate reader (SpectraMax Gemini XPS; Molecular

Devices, Sunnyvale, CA) at excitation/emission 380/470 nm. Polymers were compared to a calibration curve of glycine with known concentrations. The chemical structures of Gtn–FA and DexE–PT were characterized using a 1H NMR spectrometer (Bruker AMX-300 NMR spectrometer, Billerica, MA).

Gtn–FA and DexE–PT polymer solutions (10 mg/ml of D2O) were prepared and analyzed.

4.3 Hydrogel formation and characterization

4.3.1 Preparation of hypoxia-inducible (O2-controllable) hydrogels

Hydrogel precursor solutions (Gtn–FA, DexE–PT, laccase, mTG) were prepared in DBPS. Enzymes were maintained at final concentrations of 25 U/ml (laccase) and 0.15-0.6 U/ml (mTG). Hypoxia- inducible (HI) hydrogels were prepared by mixing aqueous Gtn–FA and/or DexE–PT polymers and laccase and/or mTG solutions. Hydrogels were prepared in 1.5 ml vials in a 3 : 1 polymer : enzyme ratio. Polymer solutions (Gtn–FA 4.0 wt%, DexE–PT 13.3 wt%) and enzyme solutions (100

U/ml laccase and/or 0.6-2.4 U/ml mTG) were mixed by pipetting to form hydrogels. All gels were formed at 37 °C.

112

4.3.2 O2 measurements

We measured dissolved O2 (DO) levels noninvasively in both acellular and cell-encapsulated hydrogels at the bottom of hydrogels using commercially available sensor patches (Oxygen Sensor

Spot; SP-Pst3) and a multichannel fiber-optic oxygen meter (OXY-4 mini) from Presens

(Regensburg, Germany). To measure O2 levels at the bottom of hydrogels, the hydrogels were added on top of the sensors, which were immobilized in each well of a 96-well plate. All experiments were conducted in a controlled environment at 37°C and 5% CO2 in a standard incubator.

O2 gradients were measured in preformed cell-encapsulated hydrogels at specified time points

(continuous gradient measurements are not possible with our O2 sensors). Commercially available needle-type oxygen microsensors (Oxygen Microsensor; NTH-Pst1) and a microfiber optic oxygen transmitter (Microx TX3) from PreSens were used to measure O2 gradients. Sensors were calibrated using atmospheric and anoxic (N2 flush) conditions at 37°C. O2 sensors were precisely controlled using a Manual Micromanipulator MM (PreSens). Starting at the bottom of the hydrogel, measurements were recorded (when the reading stabilized) every 250 μm within the hydrogel and every 500 μm or 1 mm within the media.

4.3.3 Measurement of viscoelastic properties

Rheological analysis of the HI hydrogels was performed using a rheometer (RFS3 or AR1500ex, TA

Instruments, New Castle, DE) with a 25-mm or 8-mm parallel plate geometry. For analysis of viscoelastic properties of preformed hydrogels in their equilibrium swelling state, hydrogel discs were prepared and swelled in DPBS for >12 hours. For dynamic time sweep, we monitored the elastic

113 modulus (G′) and viscous modulus (G′′) at 1% strain and a frequency of 0.1 or 10 Hz at 37 °C with a constant force of ∼10 gm or 1.3-mm gap.

4.4 Cell culture and analysis in hypoxia-inducible (O2-controllable) hydrogels

All cells were cultured using standard, humidified cell culture incubators at 37 °C and 5% CO2, unless otherwise specified.

4.4.1 ECFC encapsulation in hypoxia-inducible (O2-controllable) hydrogels

Endothelial colony forming cells (ECFCs) (Lonza, Walkersville, MD or Yoder Lab, Indiana

University School of Medicine) were cultured in EGM2 (Lonza) prepared according to the manufacturer’s instructions with an additional 10% Hyclone FBS, on standard tissue culture plates coated with type I collagen (Corning). For all hydrogels, polymer solutions were dissolved in 1×

DPBS (pH 7.4) and mixed with ECFC pellets to provide a cell suspension (4 million cells/ml), and then laccase solution (100 U/ml) was added at a volume ratio of 3:1 (polymer solution/laccase solution) and gently mixed at 37°C for a predetermined “preincubation time.” Preincubation times for each batch of polymer were based on the gelation time, as measured by the vial-tilt method.

On the basis of this gelation time, a preincubation time was calculated (gelation time minus 1 min) to prevent all cells from falling to the bottom of the hydrogel, thus ensuring a homogeneous distribution of cells in all conditions. Each mixture [cells, polymer, cross-linker(s), non-conjugated phenolic molecules] was pipetted into a 96-well plate and incubated at 37°C for 20 min. After 20 min, 200 μl of EGM2 was added to hypoxic hydrogels and 100 μl of EGM2 was added to nonhypoxic hydrogels, and cells were cultured under standard cell culture conditions (37°C, 5%

CO2) for up to 5 days. The culture medium was replaced daily. Bright-field images were captured

114 at predetermined time points to monitor cell morphology using an Olympus IX50 (Olympus;

Center Valley, PA). Additional details regarding cell encapsulation in Gel-HI hydrogels can be found in [192]. ECFCs in Gel-HI were observed via time-lapse microscopy for up to 48 hours using a Nikon Eclipse Ti (Nikon; Melville, NY) or a Zeiss AXIO Observer Z.1 (Carl Zeiss;

Oberkochen, Germany) under standard cell culture conditions (37°C, 5% CO2).

4.4.2 Mimicking hypoxic conditions in nonhypoxic geometries (1% O2 flush, DMOG, CoCl2)

To mimic hypoxic conditions in nonhypoxic geometries, a custom hypoxia chamber was designed using AutoCAD (Autodesk; San Rafael, CA) and machined at the Johns Hopkins University

Whiting School of Engineering Machine Shop. This custom hypoxia chamber facilitated flow of mixed gases and real-time monitoring of O2 using commercially available O2 sensors (PreSens).

In particular, 1% O2 (Airgas; Radnor, PA) was continually flushed through the chamber commencing immediately following ECFC encapsulation, and O2 was monitored along the culture period. Bright-field images were captured at predetermined time points to monitor cell morphology

(Olympus IX50).

DMOG and CoCl2 act as PHD inhibitors, which up-regulate HIFs. Stock solutions of DMOG were made in diH2O (according to the manufacturer and previous literature [543]). Stock solutions of

CoCl2 were made fresh before each use in diH2O (according to the manufacturer and previous literature [544]). DMOG (1 mM) or CoCl2 (200 μM) was co-encapsulated with ECFCs. Bright- field images were captured at predetermined time points to monitor cell morphology (Olympus

IX50). ImageJ (public domain) was used to quantify cell clusters. Graphs were generated using

GraphPad Prism 6 (GraphPad Software Inc., La Jolla, CA).

115

4.4.3 siRNA transfection

ECFCs were transfected with SMARTpool:siGENOME HIF1A and Non-Targeting siRNA 1 (scr) according to the manufacturer’s protocol and previous literature [193]. Briefly, cells were seeded on a six-well plate and treated with 50 nM siRNA. We confirmed knockdown via quantitative real- time fluorescence polymerase chain reaction after 24 hours and used transfected cells in experiments. ImageJ (public domain) was used to quantify cell clusters. Graphs were generated using GraphPad Prism 6 (GraphPad Software Inc.).

4.4.4 ROS inhibition (DPI)

DPI is a potent inhibitor of ROS. DPI stock solutions were prepared at 3.5 mM in DMSO. ECFC medium (EGM2) was supplemented with 20 μM DPI, and we captured bright-field images at predetermined time points to monitor cell morphology (Olympus IX50). ImageJ (public domain) and a custom MATLAB script (MathWorks) were used to quantify cell clusters. Graphs were generated using GraphPad Prism 6 (GraphPad Software Inc.).

4.4.5 Altering matrix stiffness with mTG

To increase matrix stiffness, mTG was added as a secondary cross-linker to laccase stock solutions to a concentration of 0.6-2.4 U/ml to achieve a final working concentration of 0.15-0.6 U/ml.

ECFCs were encapsulated within these stiffer gels and monitored by both bright-field imaging

(Olympus IX50) and time-lapse microscopy (Nikon Eclipse Ti or a Zeiss AXIO Observer Z.1) under standard cell culture conditions (37°C, 5% CO2). ImageJ (public domain) was used to quantify cell clusters.

116

4.4.6 Dynamic matrix stiffening

At predetermined time points (24 hours), media changes were accompanied by supplemental addition of mTG (0.6 to 2.4 U/ml). Rheological analysis (AR1500ex; TA Instruments) was performed as described above using an immersion cup. We assume uniform secondary crosslinking by the addition of supplemental mTG. While diffusion of mTG into the existing matrix potentially leads to a nonuniform matrix stiffness, we assume uniformity through a measurement of bulk matrix properties. ECFCs encapsulated in Gel-HI were monitored up to day 3 by capturing bright-field images (Olympus IX50), at which point they were fixed and stained with phalloidin and DAPI as described above. Vascular networks and lumen were observed using confocal microscopy (LSM 780; Zeiss). Networks with and without the addition of mTG were analyzed.

4.4.7 Protease activity assay

DQ-gelatin (Thermo Fisher Scientific) fluoresces in the presence of proteases. Stock solutions of

DQ-gelatin were made according to the manufacturer’s instructions (1 mg/ml in diH2O). DQ- gelatin was co-encapsulated with Gel-HI hydrogels to a final concentration of 100 μg/ml.

Fluorescent microspheres (Bangs Laboratories Inc.) or ECFCs were encapsulated in Gel-HI, and cluster formation was monitored via time-lapse microscopy (Zeiss AXIO Observer Z.1). DQ- gelatin fluorescence (488/515 nm excitation/emission) signal was continuously monitored in fluorescent microsphere experiments and quantified using ImageJ (public domain). Cell experiments required different hydrogel geometries; thus, fluorescence was measured at time 0

(T0) and time 24 hours (T24) using a SpectraMax Gemini XPS (Molecular Devices; Sunnyvale,

CA) and graphed using GraphPad Prism 6 (GraphPad Software Inc.).

117

4.4.8 Proteome profiler array

Medium was collected after 24 hours of culture from two hypoxic and four nonhypoxic hydrogels per experiment to normalize for cell number. Immediately after collection, protease inhibitors

(Halt Protease and phosphatase inhibitor cocktail) were added to the medium, and then the medium was stored at −80°C until use. Arrays were prepared according to the manufacturer’s protocol.

Films were exposed for 20 min and then analyzed using ImageJ (public domain) and graphed using

GraphPad Prism 6 (GraphPad Software Inc.). Values of mean pixel density were reported.

4.4.9 Protease inhibition (GM6001, Z-FF-FMK, PI, EDTA)

Stock solutions were prepared according to manufacturer’s instructions in the appropriate solvent.

Hydrogel precursor solutions and/or ECFC medium (EGM2) were supplemented with GM6001

(final concentration 1 mM), Z-FF-FMK (final concentration 10-500 μM), PI (final concentration

0.1X), EDTA (final concentration 0.1X), or PI/EDTA (final concentration 0.1X), and we captured bright-field images at predetermined time points to monitor cell morphology and cluster formation or inhibition (Olympus IX50).

4.4.10 Addition of exogenous MMP

Following ECFC encapsulation in Gel-HI, ECFC medium (EGM2) was supplemented with exogenous MMP (collagenase type IV; Thermo Fisher Scientific) at a range of concentrations.

Bright-field images were captured at predetermined time points to monitor cell morphology

(Olympus IX50). ImageJ (public domain) and a custom MATLAB script (MathWorks) were used to quantify cell clusters. Graphs were generated using GraphPad Prism 6 (GraphPad Software Inc.,

La Jolla, CA).

118

4.4.11 Observation of cell-cell interactions and vascular network formation

After 24 hours in culture (for observations of cell-cell interactions) or 72 hours in culture (for observations of vascular network formation), ECFCs within Gel-HI hydrogels were fixed with 2%

FA for 20 min at room temperature. Hydrogels were then washed three times with 1× DPBS with

10 min in between each wash. Encapsulated ECFCs were then permeabilized with 1% Triton X-

100 for 15 to 20 min and then washed three times with 1× DPBS with 10 min in between each wash. Next, hydrogels were incubated in 5% BSA blocking solution for 1 hour at room temperature and then washed with 0.05% Tween-20 in 1× DPBS three times with 10 min in between. Hydrogels were then stained with primary antibody in antibody diluent solution overnight at 4°C and then washed with 0.05% Tween-20 in 1× DPBS three times with 10 min in between. Hydrogels were incubated in secondary antibody in antibody diluent solution for 2 hours at room temperature and then washed with 0.05% Tween-20 in 1× DPBS three times with 10 min in between. Last, hydrogels were incubated in DAPI solution for 15 min at room temperature and then washed with

1× DPBS three times with 10 min in between. Hydrogels were analyzed using confocal microscopy

(LSM 780; Zeiss). Primary antibodies ICAM-1 and ITG-β2 (R&D Systems) were diluted in sterile

DBPS to 500 and 200 μg/ml, respectively, and were used at a final concentration of 15 μg/ml in antibody diluent (Thermo Fisher Scientific), according to the manufacturer’s protocol. VE- cadherin antibody (Santa Cruz Biotechnology) was used at a 1:50 dilution in antibody diluent.

Secondary antibodies, gt488 and ms 546, were used at 1:250 in antibody diluent, and phalloidin was used at 1:500 in antibody diluent.

4.5 Cell culture and analysis in layered hypoxia-inducible (O2-controllable) hydrogels

4.5.1 ECFC encapsulation in layered hydrogels

119

The method for encapsulation in layered hydrogels is nearly identical to the method outlined in section 4.4.1. Here, we generated both 2- and 3-layer hydrogels. For both the 2- and 3-layer hydrogels, ‘bottom’ layers containing cells (ECFCs or GFP-ECFCs) were generated in the same manner as the nonhypoxic hydrogels described previously, but rather than adding media after the

20 min incubation, an additional 50 μl (for the 2-layer system) or 100 μl (for the 3-layer system) layer of acellular hydrogel was added, then incubated for 20 min at 37°C for 20 mins. Following incubation, 100-200 μl EGM-2 was added. To generate the ‘top’ layer, an acellular hydrogel (50

μl for the 2-layer system; 100 μl for the 3-layer system) was generated and incubated at 37°C for

20 mins, then a 50 μl cellularized layer was added atop the acellular hydrogel, then incubated at

37°C for 20 mins. Following incubation, 100-200 μl EGM-2 was added. To generate the ‘middle’ layer in the 3-layer hydrogels, a 50 μl acellular hydrogel was generated and incubated at 37°C for

20 mins, then a 50 μl cellularized layer was added atop the acellular hydrogel, then incubated at

37°C for 20 mins, then a 50 μl acellular layer was added atop the hydrogel, then incubated at 37°C for 20 mins. Following incubation 100-200 μl EGM-2 was added. Cells were cultured under standard cell culture conditions (37°C, 5% CO2). The culture medium was replaced daily. Bright- field images were captured at predetermined time points to monitor cell morphology using an

Olympus IX50 (Olympus; Center Valley, PA).

4.5.2 CellROX for oxidative stress detection

CellROX is used to identify oxidative stress in live cells. CellROX stock solutions were prepared at 2.5 mM in DMSO. CellROX (5 μM) was then co-encapsulated with ECFCs in layered hydrogels and fluorescence intensity was monitored over 24 hours at pre-determined time points using an

Olympus IX50 (Olympus; Center Valley, PA) for brightfield and fluorescent images.

120

4.5.3 RNA extraction and purification

RNA was extracted from ECFCs cultured in ‘bottom’ (hypoxic) layers or nonhypoxic conditions.

EGM-2 was removed and 200 ul Trizol reagent was added atop the hydrogels and incubated at room temperature for 5 min. Hydrogels were then mixed using a P1000 pipette tip and transferred to a 1.5 ml microfuge tube containing an additional 300 ul Trizol reagent. Samples were then homogenized using a pellet pestle homogenizer (FisherBrand) for 1-2 min. Samples were then centrifuged at 12G for 10 min at 4°C. The supernatant was then transferred to a fresh 1.5 ml microfuge tube and stored at -80°C until ready for purification. Samples were thawed on ice prior to purification. RNA was purified using Direct-zol RNA Miniprep kits according to the manufacturer’s suggestion (Zymo Research; Irvine, CA), with a few minor exceptions and clarifications identified in the following text. All centrifugation steps were performed at 13G for

30 seconds at 4°C, unless another speed was specified. DNase I treatment was used. Samples were eluted in 15 μl DNase/RNase-free water. Following elution, the samples were reloaded and eluted a second time. Following purification, samples were stored at -80°C.

4.5.4 RNA-sequencing

RNA sequencing was performed at the Johns Hopkins University School of Medicine

Transcriptomics and Deep Sequencing Core, guided by the direction of Dr. Haiping Hao. RNA- seq library for Illumina platform sequencing is prepared using Illumina TruSeq stranded total RNA

Sample kit following manufacturer’s recommended procedure. Briefly, 100ng of total RNA is first depleted of ribosomal RNA with Ribozero Gold magnetic beads and further purified with

Agencourt AMPure XP beads. RNA is fragmented at 94 °C for 8 minutes and primed with random primer. The fragmented RNA was then converted to double strand cDNA, end repaired, A tailed, and ligated with Unique Dual Indexed adaptors. The adaptor added cDNA library was then PCR

121 amplified using the following conditions: 94 °C for 30 seconds, 15 cycles of 98 °C for

10 seconds, 60 °C for 30 seconds, 72 °C for 30 seconds, 72 °C for 5 minutes. The PCR amplified library was purified using Agencourt AMPure XP magnetic beads and run out on Agilent High

Sensitivity DNA Chip for quality check. Individual library was then further quantified using

KAPA library quantification qPCR kit and pooled. The pooled library was sequenced on Illumina

NovaSeq S1 200cycle kit for 2X100bp sequencing.

4.5.5 Analysis of RNA-seq data

RNA sequencing analysis was performed at the Johns Hopkins University School of Medicine

Transcriptomics and Deep Sequencing Core, by Conover Talbot. In brief, differential gene expression was determined by one-way ANOVA of log2(FC) between hypoxic and non-hypoxic conditions at each of the three time points. Significantly differentially expressed genes were identified by the following: [SD (σ) Binned Max^-0.8 Hypoxic vs Non-hypoxic log2(FC)]. Genes of interest ≥ǀ2σǀ were assessed for relevancy. Ingenuity Pathway Analysis (IPA) was used to gain an understanding of the relationship between differentially expressed genes over the time course and to identify new canonical pathways and biological functions of interest in guiding cluster- based vasculogenesis.

4.6 In vivo mouse model

4.6.1 Plug assay in vivo

Acellular hydrogels were injected into the flank of nu/nu female mice (8 to 10 weeks old) (nu/nu

088 homozygous; Charles River) using insulin needles (28 gauge). Hydrogel (200 μl) was injected with CaO2(nonhypoxic) or Ca(OH)2 (hypoxic), as described previously (31). Briefly,

CaO2 decomposes and releases O2 when it comes into contact with H2O. As such, we mixed

122

CaO2 as an O2-releasing compound for nonhypoxic hydrogels. For hypoxic hydrogels, we mixed

Ca(OH)2, a by-product of CaO2 decomposition, to minimize the differences in hydrogel composition between the two groups. We mixed sterilized CaO2 or Ca(OH)2 with the polymer solution immediately before hydrogel injection to achieve a final concentration of 0.02 wt%

CaO2 or Ca(OH)2. In addition, GFP-ECFCs (provided by M. Yoder and transfected by K. Eisinger and K. Pak; University of Pennsylvania) were injected via intracardiac injection (1 million cells per 100 μl of injection) with insulin needles (28 gauge). SDF-1α (100 ng/ml; R&D Systems) was co-encapsulated within hydrogels to enhance recruitment of circulating EPCs, including host cells and injected GFP-ECFCs [262, 338]. DPI (20 μM) or vehicle control (DMSO) was co- encapsulated where indicated. At 12 hours, mice were euthanized and hydrogels were fixed overnight in 2% FA. Following fixation, hydrogel was washed with 1× DPBS three times, stained with Isolectin GS-IB4 Alexa Fluor 568 Conjugate (1:200 in antibody diluent) (Thermo Fisher

Scientific) for 2 hours at room temperature, and then washed with 1× DPBS three times with 10 min in between. Hydrogels were then incubated in Hoechst solution (1:2000 in DPBS) for 10 min at room temperature and then washed with 1× DPBS three times with 10 min in between. Stock solutions of Isolectin GS-IB4 and Hoechst were prepared according to the manufacturer’s instructions. Recruited cells were analyzed using confocal microscopy (LSM 800; Zeiss). All animal studies were performed following approval by the Johns Hopkins University Institutional

Animal Care and Use Committee (MO15A154, MO18A244).

4.6.2 O2 measurements in vivo

Commercially available needle-type oxygen microsensors (Oxygen Microsensor; NTH-Pst1) and a microfiber optic oxygen transmitter (Microx TX3) from PreSens were used to measure in vivo

O2. After 12 hours, needle-type sensors, controlled by a micromanipulator and concurrent with

123 temperature readings using a manufacturer-provided thermometer, were inserted into the hydrogel, and O2 was recorded when the reading stabilized. All animal studies were performed following approval by the Johns Hopkins University Institutional Animal Care and Use Committee

(MO15A154, MO18A244).

4.7 Image and data analysis

4.7.1 Image quantification

Analysis of cells and clusters during time-lapse experiments was performed using ImageJ (public domain) and a custom MATLAB script (MathWorks; Natick, MA). Analysis of vessel characteristics was performed using ImageJ (public domain) and MetaMorph (Universal Imaging,

Downingtown, PA). Analysis of stress fibers was performed using ImageJ (public domain) and a custom MATLAB script developed in the laboratory (MathWorks; Natick, MA) [200]. Analysis of GFP-ECFCs in vivo was performed using Imaris (Bitplane, Belfast, UK) and ImageJ (public domain). All graphs were made using GraphPad Prism 6, 7, or 8 (GraphPad Software Inc.).

4.7.2 Statistical analysis

We performed statistical analysis using GraphPad Prism 6, 7, or 8 (GraphPad Software Inc.). We also used this software to perform t tests to determine significance. Replicates are indicated throughout the figure captions. All graphical data are reported as means ± SD. Significance levels were set at *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. All graphical data were reported.

124

CHAPTER 5

5 Designer Hydrogels for Precision Control of Oxygen Tension and Mechanical Properties

5.1 Teaser

A set of hydrogels was developed to precisely control O2 and matrix stiffness, enabling study of the impact of these two parameters on vascular morphogenesis.

5.2 Introduction

The ever increasing complexity and precision control of hydrogel biomimetic properties have enabled great strides to be taken in a broad range of interdisciplinary fields, perhaps most importantly at the intersection of materials engineering and biomedical sciences. Applications for hydrogel technologies have expanded to include controlled cell [545-549] and drug/growth factor delivery [249-251, 256, 259, 260, 276-278], engineered tissue [238, 239, 550], acellular therapeutics [551], and more recently, as scaffolds to study the complex interplay between cells and their microenvironment. Many important instructive interactions between cells and their microenvironment have been studied and defined utilizing hydrogel systems, which have sought to isolate specific variables. Important parameters of the cellular microenvironment that have been incorporated into hydrogel design and subsequently shown to affect cell function, phenotype, and fate include the effects of mechanical properties [151, 152, 552], degradation [219, 274, 553], protein tethering [554], cell adhesion [295, 555], geometry [296, 479, 556], topography [557, 558], oxygen tension [66, 307], and the presence of signaling molecules including growth factors and cytokines [559]. Gaining a fundamental understanding of the factors which control the most downstream effects and elicit the most robust responses will likely lead to important steps towards the development and translation of a wide range of novel therapies.

125

Substrate stiffness and oxygen tension have emerged as two vital parameters that have influenced tissue engineering and regenerative medicine, as well as for numerous basic science studies. Indeed, these two factors elicit numerous downstream effects and have proven to be two of the most potent regulators of cellular function. Of particular interest to our group, oxygen tension and substrate stiffness play vital roles in vascular morphogenesis (including angiogenesis and vasculogenesis)[66, 152, 219, 557] and cancer progression and metastasis, specifically tumor angiogenesis [560-563].

Extremely low levels of oxygen have proven detrimental to the success of tissue engineered constructs, owing to the diffusion limit of oxygen and nutrients and ultimately the necrosis imparted by the lack of these two key regulators of cell survival. However, low levels of oxygen, defined as hypoxia (<5% O2) have been shown to be a potent promoter of the angiogenic program both for the body’s natural healing of injured tissue and for tumor angiogenesis [230, 564].

Hypoxia inducible factors (HIFs) are transcription factors that regulate numerous downstream effects including the production of important pro-angiogenic signaling molecules that aid in the development of neovasculature [230]. In addition to its effect on more mature endothelial and endothelial progenitor cells, hypoxia also plays an important role in determining stem cell fate. In the developing embryo, dissolved oxygen levels range form 2-9% [83]. As the vasculature develops under hypoxic conditions, exact mimicry of the embryonic differentiation conditions necessitates a hypoxic microenvironment. Indeed, priming pluripotent stem cells in a hypoxic microenvironment enhances their differentiation along a vascular lineage and improves their ability to form vascular networks [66].

The importance of substrate stiffness on cell function and stem cell fate decisions has been made abundantly clear over the last ten years. Studies pertaining to angiogenesis and

126 vasculogenesis are largely limited to mature and progenitor cells indicating that that soft substrates

(< 1 kPa) are generally more conducive to the formation of functional vascular networks. Stiffer substrates often limit cell migration and the eventual coalescing of endothelial cells to form perfusable, lumenal structures [152]. Further, the elasticity of the ECM has proven a vital signaling cue for metastatic cancer [561].

Others and us have utilized hydrogels as an important tool to study the interactions between cells and their microenvironment. Recently, we have developed both gelatin (Gtn) and dextran

(Dex) based oxygen-controlling hydrogels, which we have termed hypoxia-inducible (HI) hydrogels [307, 565]. Ferulic acid (FA), a phenol containing molecule that also contains a carboxyl end group, was conjugated to the amine groups readily present on Gtn via a carboiimide-mediated reaction. Dextran was first aminated (DexE), then phenol containing tyramine (TA) was conjugated through amine reactive polyethylene glycol (PEG). The hydrophilic PEG linker was used to enhance the crosslinking reactivity. The crosslinking groups are FA and NH-PEG-TA for the Gtn and Dex-based hydrogels, respectively. Hydrogels were formed individually using either

Gtn or Dex polymers through an enzymatic laccase-mediated crosslinking reaction. Gtn and Dex hydrogels were formed by formation of di-ferulic acid and di-tyramine bonds, respectively.

In both of these hydrogel systems, the time which hypoxia is maintained and the hydrogel mechanical properties are directly related to the composition of polymer and the concentration of laccase used. Specifically, as the concentration of polymer is increased, both stiffness and time hypoxic (TH) increase accordingly in a highly predictable fashion. As the importance of both of these parameters in cellular responses have been described individually, we propose a hydrogel system which can precisely control each of these parameters individually, while also retaining the ability to study potential synergistic effects between the two variables.

127

Here we show that by incorporating non-conjugated phenol containing molecules prior to hydrogel crosslinking, TH is significantly increased with only mild effects on hydrogel stiffness.

Further, by incorporating a secondary crosslinker, effects on mechanical properties can be induced with negligible effects on TH. Transglutaminase provides an ideal secondary crosslinker, as it is able to enzymatically crosslink the amine groups readily present on the gelatin backbone, as well as the unconjugated amine groups on the dextran backbone [493, 566]. By controlling these two variables independently, we can gain an increased understanding into the effects of two parameters vital to vascular morphogenesis and cancer progression.

5.3 Results and Discussion

5.3.1 Hybrid Gtn-Dex hydrogel formation and characterization

Gelatin-based and dextran-based polymers were synthesized as previously described [307,

565]. The chemical structures of the conjugated crosslinking groups (FA and NH-PEG-TA) are provided in Fig. 5.1 and were confirmed with 1H-NMR (Fig. S5.1). In order to determine the degree of substitution (DS) of phenol containing moieties, UV/vis spectroscopy was performed.

The DS values obtained here were similar to those obtained in previous studies, indicating the reproducibility of these synthesis protocols. The DS of FA for gelatin-based polymers (Gtn-FA) was 55.31±5.67 µmol/g of polymer (DS 55), while the DS of TA for aminated, PEG-linked dextran-based polymers (DexE-PT) was 174.81±3.27 µmol/g of polymer (DS 175). Additional small batch-to-batch variability was accounted for when separate batches of polymer were used.

These results are summarized in Table 5.1.

128

Table 5.1. Concentrations of Crosslinking Moieties

Polymer backbone DS of Phenol Moiety Available –NH2 Total Available (µmol/g polymer) (µmol/g polymer) Crosslinking Molecules (µmol/g polymer) Gtn-FA 55.31±5.67 394.69±6.73 ~450.00 DexE-PT 174.81±3.27 0.64±0.31 ~175.45

In order to form hybrid Gtn-Dex hydrogels, a laccase-mediated enzymatic crosslinking reaction was employed (Fig. 5.1). A constant concentration of laccase (25 U/ml) was used in all hydrogels, as this concentration exhibited high crosslinking efficiency and good biocompatibility.

Hydrogels were formed at 37°C in order to enhance the enzymatic activity, as well as to recapitulate the in vivo setting and further display the potential for the formation of these hydrogels in situ. Further, reaction conditions, including pH and ion strength, were chosen to mimic the in vivo microenvironment to enable study of these hydrogels in a highly biomimetic fashion. By combining the two polymer backbones, homo-crosslinking (Gtn-FA to Gtn-FA; DexE-PT to

DexE-PT) and hetero-crosslinking (Gtn-FA to DexE-PT) occurs yielding hybrid hydrogels (Fig.

5.1A). As the oxygen consumption is highly dependent on the phenol concentration of the constituents of the hydrogel, and laccase crosslinking is inherently linked to phenol concentration, it follows that with increased crosslinking, increases in TH and mechanical properties occur in concert. In order to account for this, we utilized non-conjugated, phenol-containing molecules,

TA, to increase TH without increasing the mechanical properties of the hydrogel. As shown in Fig.

5.1B, non-conjugated molecules either bind to other non-conjugated molecules, or interrupt some of the crosslinking by binding to conjugated phenolic moieties. As such, the TH can be controlled precisely by the addition of given concentrations of TA in a hydrogel and the mechanical properties will show a slight decrease to account for the decrease in crosslinking. Oppositely, increases in

129 mechanical properties may be achieved by employing a secondary crosslinking reaction. Both polymers used here (Gtn-FA and DexE-PT) contain unconjugated primary amines. These primary amines can be crosslinked enzymatically using microbial transglutaminase (mTG).

Transglutaminase, like laccase, has been used extensively in the food industry and has recently found a niche in biomedical applications through crosslinking of primary amine containing polymers, such as collagen and gelatin [493, 567]. This secondary crosslinking has theoretically insignificant effects on hydrogel oxygen levels, and can act to increase crosslinking density and thus increase the mechanical properties of the hydrogel (Fig. 5.1C). To determine the number of available crosslinking primary amines on each polymer, primary amines were tagged with fluorescamine and subsequently quantified using a fluorescence microplate reader. Predictably, there were a large number of excess primary amines on Gtn-FA polymers (394.69±6.73 µmol/g of polymer), indicating potential for a more significant increase in mechanical properties of these hydrogels compared to the relatively low unconjugated primary amine concentration on DexE-PT polymer (0.64±0.31 µmol/g of polymer). With only trace concentrations of available primary amines to crosslink through mTG on DexE-PT polymers, the secondary crosslinking will likely be dominated by the Gtn-FA portion of hybrid hydrogels. The results presented here are summarized in Table 5.1.

5.3.2 Hybrid Gtn-Dex hydrogels for control of oxygen tension and mechanical properties

In order to examine the hypothesis that both oxygen tension and elastic modulus will increase as the concentration of phenolic moieties increases, a range of hybrid Gtn-Dex polymers were synthesized and analyzed. The gelatin backbone of Gtn-FA provides many important components often present in natural polymers. The biodegradability and presence of cell adhesion motifs are both of the utmost importance when studying cells in vitro, and are critical to vascular

130 morphogenesis [222]. One major drawback of the use of a Gtn based backbone in our system is the limited conjugation efficiency of phenolic moieties and the subsequent limitation of TH.

Prolonged hypoxia could only be maintained in this system by way of cell oxygen consumption in vitro [307], or by suspected oxygen consumption of surrounding tissue upon injection in vivo. To account for these limitations, a dextran based hydrogel (DexE-PT) was developed. Conjugation

131

Figure 5.1: Crosslinking chemistry for hybrid Gtn–FA and DexE–PT hydrogels. (A) Hypoxia inducible hydrogels are formed via laccase-mediated dimerization of phenolic moieties. Laccase catalyzes the reduction of atmospheric O2 to water, resulting in the oxidation of phenol-containing ferulic acid (FA, ) and tyramine (TA, ) conjugated to Gtn and Dex, respectively, forming crosslinked polymer networks. (B) Additional non-conjugated phenol-containing molecules (e.g.

TA, ) are added prior to the reaction, resulting in decreased cross-linking density, and thus slightly decreased stiffness. The same reaction kinetics are employed, resulting in nearly identical oxygen consumption and a decoupling of oxygen concentration and mechanical properties. (C) Secondary crosslinking via transglutaminase-mediated dimerization of primary amines ( ) present on both polymer backbones leads to increased stiffness, while not affecting oxygen consumption. By controlling the number of phenolic moieties (both conjugated and unconjugated), as well as controlling the crosslinking density by secondary crosslinking, precise control over both oxygen concentration and mechanical stiffness may be achieved.

efficiency of DexE-PT was 3-4 times higher than for Gtn-FA (Table 5.1), leading to prolonged hypoxia controlled solely by the crosslinking reaction [565]. The major limitation of dextran-based hydrogels is the lack of cell adhesion binding sites. By combining these two hydrogels, a highly bioactive hydrogel with prolonged hypoxia is achieved.

Hybrid hydrogels were formed by ranging Gtn-FA and DexE-PT concentrations. Bounds of our region of study were defined by previous results reported in [307, 565]. Gtn-FA 3 wt% was shown to be an appropriate hydrogel composition for studies both in vitro and in vivo, while DexE-

PT 10 wt% yielded the highest TH of any hydrogel tested. Combinations of hydrogels with total concentrations of phenolic moieties falling between these two hydrogels (Gtn-FA 3 wt% and

132

DexE-PT 10 wt%) were tested. As shown in Fig. 5.2A, increasing concentrations of phenol containing molecules led to predictable increases in TH. Further, we determined the dynamic elastic modulus (G’) and found that as phenol concentrations increased, G’ increased accordingly

(Fig. 5.2B). When normalizing G’ to relative total number of phenolic molecules, we found that increasing phenol concentration resulted in both increased TH and G’ (Fig. 5.2C). This type of predictability can enable a user-defined TH and G’ to develop hydrogels for highly specific applications.

Figure 5.2: Dissolved oxygen (DO) and rheological characterization for hybrid hydrogels. (A) DO levels were measured at the bottom of hydrogels (3.13 mm thick, 25 U/ml laccase) as a function of time.

(B) Rheological characterization of hydrogels as measured by elastic modulus (G′) using dynamic time sweep. (C) DO and elastic modulus are coupled: as phenol concentration increases, both stiffness and time hypoxic increase.

133

5.3.3 Decoupling TH and mechanical properties through addition of non-conjugated phenol- containing molecules

In order to decouple TH and G’, addition of non-conjugated TA to a range of hybrid polymers was performed. As a demonstration of the precision of this approach, concentrations of

TA were determined as to match each of the oxygen graphs (with higher TH) in Fig. 5.2A. In other words, specific concentrations of TA were added to Gtn-FA 3 wt% to match TH for each of the hybrid hydrogels. The concentrations of TA added to the Gtn-FA 3 wt% polymer enabled precision control over TH (Fig. 5.3A(i)). By matching the phenol concentrations with the addition of TA,

TH was matched for a variety of hydrogel compositions (Fig. 5.3B(i);D(i)). We next examined the effect of this non-conjugated TA on G’, and found that as predicted, decreases in G’ result from interruption of crosslinks by non-conjugated TA (Fig. 5.3A(ii);D(ii)). The data here are presented as a normalized elastic modulus (G’) (normalized to G’ of Gtn-FA 3 wt%), in order to make this data easier to compare between rheological tests. With the addition of TA to Gtn-FA, the already low crosslinking density was decreased further, resulting in a highly viscous fluid, rather than a hydrogel. The importance of a secondary crosslinker for these types of hydrogels is discussed below. Notably, highly predictable TH may be achieved by simply adding non- conjugated phenol containing molecules. These molecules have minimal effect on mechanical properties of the hydrogels, enabling the potential study of a range of TH, while restricting mechanical properties to a small range of values which are not strictly dependent on phenol concentration (Fig. 5.3A(iii);D(iii)).

5.3.4 Decoupling mechanical properties and TH through secondary crosslinking

In order to further decouple oxygen tension and mechanical properties, the addition of a secondary crosslinker, transglutaminase (TG), was examined. By maintaining a consistent oxygen tension

134 and providing additional crosslinks between polymer chains, the opposite effect of addition of non- conjugated TA may be achieved. In other words, stiffer hydrogels with shorter TH can theoretically be developed. TG is a naturally occurring crosslinker that catalyzes the formation of

N ε-(γ-glutamyl) lysine amide bonds between gelatin or collagen polymer chains [493, 567]. TG has been used in a number of collagen or gelatin containing hydrogels due to its improved cytocompatibility over other crosslinkers and potential as an in situ crosslinker, owing to its relatively short gelation time at body temperature and pH [493]. Its crosslinking potential can be expanded beyond just gelatin and collagen to include crosslinking of primary amine containing polymer chains, such as DexE-PT. Mammalian TG inherently relies on the presence of proenzymes and calcium ions to reach its maximum potency. However, innocuous microbial TG

(mTG) has seen widespread use in the food industry, is U.S. Food and Drug Administration approved, and functions independently of proenzymes and calcium ions [493], thus further indicating its vast potential in biomedical applications.

The effects of a mTG on oxygen tension are shown in Fig. 5.4A. Hydrogels were formed with the same concentration of laccase as described above (25 U/ml), but with the addition of mTG at 0.3 U/ml, as described in [566]. As expected, the addition of mTG had minimal effect on TH.

However, the addition of mTG only had insignificant or neglible effects on the mechanical properties of the hybrid hydrogels (Fig. 5.4B,C). Predicted increases in G’ were not observed, particularly in those hydrogels with high concentration of the Gtn-FA component. These hydrogels have a significant increase in the number of potential crosslinking sites upon the addition of secondary crosslinker, mTG (Table 5.1). Hydrogels with less or no gelatin component only have a slight increase in potential crosslinking sites. This predicted difference was not evident in the

135

136

Figure 5.3: Addition of non-conjugated TA enables control over TH with minimal effects on G′.

(A(i)–D(i)) DO levels were measured at the bottom of hydrogels for hydrogels with addition of TA:

A(Gtn–FA 3%), B (Gtn–FA 2%, DexE–PT 3.33%), C (Gtn–FA 1.5%, DexE–PT 5%), C (Gtn–FA

1%, DexE–PT 6.67%). (A(ii)–D(ii)) Rheological characterization of the same hydrogel formulations.

(A(iii)–D(iii)). DO and G′ are no longer coupled and precision control over each variable is shown.

Black triangles represent hybrid hydrogel TH (no TA); red triangles represent hybrid hydrogel TH with additional non-conjugated TA. *indicates P < 0.005.

rheological characterization of these hydrogels (Fig. 5.4B). It is possible that the faster reaction kinetics of the laccase enzyme led to decreased availability of crosslinking sites for mTG due to steric hindrance and decreased polymer chain mobility. In order to combat these minor effects in future studies, increases in mTG concentration will be investigated, along with their accompanying cytotoxicity tests, as well as a two-step crosslinking in which mTG is added first, then laccase is added so as both enzymes complete their respective catalytic crosslinking reactions at the same time and are both minimally hindered in their reactions.

Another important function of mTG component becomes clear when Gtn-FA hydrogels with high TA concentration are studied. These combinations were essentially viscous fluids rather than hydrogels (data not shown). By adding mTG along with non-conjugated TA, hydrogel formation will likely occur, leading to an enhanced breadth of possible studies for these hybrid hydrogels, as well as enhanced therapeutic potential as a stand-alone treatment option.

137

Figure 5.4: Dissolved oxygen (DO) and rheological characterization for hybrid hydrogels with secondary crosslinking microbial transglutaminase (mTG). (A) DO levels were measured at the bottom of hydrogels (3.13 mm thick, 25 U/ml laccase, 0.3 U/ml mTG) as a function of time. (B)

Rheological characterization of hydrogels as measured by viscoelastic modulus (G′) using dynamic time sweep and normalized to Gtn–FA 3% in order to provide a direct comparison between hydrogels. (C)

DO and viscoelastic modulus remain coupled, but less strictly. * indicates P < 0.05.

5.3.5 Varying oxygen tension and mechanical stiffness alters vasculogenesis in vitro

In order to establish the viability of these hydrogels for biological experimentation, in particular the effects of varying oxygen tension and mechanical stiffness on vascular morphogenesis, we utilized endothelial colony forming cells (ECFCs). The cytocompatibility of both polymers (Gtn and Dex), which contain free phenol groups in the same concentrations as the non-conjugated phenol groups tested here, and both enzymes (laccase and mTG) have been established previously [307, 493, 565]).

138

Encapsulation of ECFCs within hydrogels with a range of properties was tested. As described previously [307], soft hydrogels with moderate hypoxic conditions (Gtn-FA 3%) are amenable to vascular network formation (Fig. 5.5). Upon investigation of other hybrid hydrogels, several interesting conclusions can begin to be drawn. As a proof of concept, hybrid hydrogels

(Gtn-FA 2%, DexE-PT 3.33%) were tested (Fig. 5.5). Interestingly, following the same encapsulation protocol, the ECFCs in the hybrid hydrogels seemed to aggregate much more readily than in the Gtn-FA 3% hydrogels. This increased aggregation is likely a result of the decreased availability of cell adhesion sites in the hybrid hydrogels, likely leading to more cell-cell interactions in the initial encapsulation. The well documented steps of vascular morphogenesis

[274] were not seen in the Gtn-FA 2%, DexE-PT 3.33% hybrid hydrogels. Vacuole formation was noticeably absent, and sprouting cells appeared to bridge from one ‘island’ of cell aggregates to another ‘island’.

To further prove the novelty and efficacy of the hybrid hydrogels described, Gtn-FA 3% hydrogels with additional TA (0.054%) to match the oxygen tension of the Gtn-FA 2%, DexE-PT

3.33% hybrid hydrogels were examined (see Fig. 5.3; Fig. 5.5). Robust structures were formed even in these extreme hypoxic conditions (~0% O2) indicating the ability for ECFCs to survive and even thrive in low hypoxic conditions. Further indications for an interplay between oxygen tension and mechanical properties are evidenced here, as few vascular morphogenetic structures were formed in the severely hypoxic ‘stiff’ hybrid gels (Gtn-FA 2%, DexE-PT 3.33%), while extensive network formation was present in the softer hydrogels (Gtn-FA 3%, TA 0.054%).

Finally, hybrid hydrogels with the same TH as Gtn-FA 3%, but approximately two times the elastic modulus, were tested (see Fig. 5.4). By adding mTG, stiffer hydrogels were formed

(Gtn-FA 3%, mTG; Fig. 5.5). As previously observed [152], ECFCs in these hydrogels formed

139 much less robust networks. They followed the initial stages of vasculogenesis (vacuole formation), but were unable to progress to form sprouts and vascular networks. This may reveal an additional specification about the importance of the mechanical stiffness on vascular network formation and display the precision with which ECFCs sense their surrounding microenvironment. The small, but clearly significant, change in elastic modulus prevented completion of vascular network formation. These results may also indicate that mechanical stiffness plays an important role in the later stages of vascular morphogenesis, as the increase in stiffness did not impede the initial stages of the process. In stiff gels with longer TH, sprouts were able to form, while in stiff gels with briefer TH, vasculogenic potential was limited.

Taken together, these results indicate the sensitivity of ECFCs to form vascular networks, highlighting the need for a system such as the one presented here to tease out the exact mechanisms of vasculogenesis that permit the most robust and ultimately functional vascular networks.

140

Figure 5.5: Vascular morphogenesis in an array of microenvironmental conditions. Gtn–FA 3% hydrogels (soft, moderate hypoxia) were amenable to network formation, which followed the well understood vascular morphogenetic process of vacuole formation, followed by lumen formation and ultimately branching and sprouting to form complex networks. Hybrid hydrogels (Gtn–FA 2%, DexE–

PT 3.33%) (stiff, severe hypoxia) permitted sprouting, but not extensive network formation.

Interestingly, cells seemed to aggregate and form sprouts between groups of cell aggregates without the

141 initial stage of vascular morphogenesis (vacuole/lumen formation). Gtn–FA 3%, TA 0.054% hydrogels

(soft, severe hypoxia) matched the oxygen tension in Gtn–FA 2%, DexE–PT 3.33%, but were soft.

Again, cell aggregates led to sprouting, but in these softer hydrogels, more robust networks formed.

Gtn–FA 3%, 0.3 U ml−1 mTG (stiff, moderate hypoxia) permitted the initial stages of vascular morphogenesis, but completely inhibited branching and sprouting. Vacuoles/lumen indicated by arrowheads ( ), sprouting indicated by arrows ( ). Scale bars 100 μm.

5.4 Conclusions

By varying the composition of hydrogel backbone components, predictable changes in TH and mechanical properties were achieved. Choosing a specific composition of polymers allows for user-defined control of both of these parameters, indicating the potential for such hybrid hydrogel systems in a range of cellular studies in vitro and as potential treatment options in vivo. However, these studies are restricted by the short TH and low elastic modulus achievable by Gtn-FA hydrogels and the limited bioactivity of the DexE-PT hydrogels. By decoupling both TH and elastic modulus, significant enhancement of control of these properties is achieved. Addition of non-conjugated TA allows for significant increases in TH, while minimally affecting the mechanical properties of hydrogels. Utilizing a secondary crosslinker (mTG), which has negligible effects on TH, while increasing the elastic modulus will also aid in controlling these hydrogels.

Perhaps most importantly, this secondary crosslinker will help when designing hybrid hydrogels with high concentration of TA and low inherent crosslinking density (e.g. Gtn-FA). Secondary crosslinking will also allow prolonged hypoxia beyond the 10-12 hour range exhibited in this study.

142

Choosing specific polymer compositions also enables control over the concentration of cell adhesion motifs, a vital regulator of cellular function. Because DexE-PT does not contain cell binding sites, the cell adhesion is strictly regulated by the Gtn-FA component. Varying this component will allow investigation of the potential synergistic effects of TH and elastic modulus as they pertain to cell adhesion. Additionally, because two degradation mechanisms are necessary to degrade hybrid hydrogels (collagenase for Gtn, dextranase for Dex), control over degradation of these hydrogels can be incorporated into the design. The range of elastic moduli presented in this study, from <20 Pa to 1 kPa, will allow study of the range of mechanical properties experienced by the cells of the vasculature, from development to maturity [152, 568]. Such potential studies may help elucidate the mechanisms controlling myriad downstream effects on angiogenesis and vasculogenesis, towards the development of potential novel therapeutics.

Another important aspect of these hybrid hydrogels is the ability to control the oxygen gradient. Oxygen gradients exist throughout the body. As such, cellular function is highly dependent on specific oxygen tension, indicating the importance of studying oxygen gradients in both tissue engineering and tumor biology. While not measured here, there are likely a range of oxygen gradients that may be developed utilizing these hydrogels. For example, while most of the hybrid hydrogels shown here reach 0% oxygen tension, there are likely significant differences in the steepness of the oxygen gradient and how long this oxygen gradient is maintained. Altering the crosslinking density will change the pore size and ultimately the oxygen diffusion through these hydrogels, which may enable precision control of oxygen gradients.

Varying TH and the mechanical properties of these hydrogels without changing any aspects of polymer synthesis enables a process that is highly repeatable and has potential for scale up in

143 the future, as these hydrogels are optimized for high throughput studies and as disease treatment therapies.

5.5 Supplementary information

Figure S5.1: 1H-NMR spectra for Gtn-FA (A) and DexE-PT (B).

144

CHAPTER 6

6 Applications of O2-Controllable Hydrogels

6.1 Teaser

O2-controllable hydrogels are a powerful platform to study tumor and vascular biology both in vitro and in vivo.

6.2 Introduction

Oxygen (dioxygen, O2) is a key molecule vital for the existence of all multicellular organisms. Variations in O2 are known to control signaling cascades that lead to metabolic and phenotypic changes. As partial pressure of O2 falls below 5%, herein defined as hypoxia (see Table

6.1), a myriad of cellular and systemic adaptations commence.

During tumor development and progression, a low O2 environment manifests when a tumor outgrows the existing O2 supply, leading to the upregulation of numerous important factors, perhaps most notably hypoxia inducible factor 1 alpha (HIF-1). HIF-1 upregulation has been correlated with negative patient outcomes in many cancers such as sarcoma [569], prostate cancer

[570], and squamous cell head and neck cancer [571]. To exacerbate such negative outcomes, tumors with high levels of HIFs have resistance to chemotherapy and radiation therapy, due to limited perfusion [572]. On the cellular level, hypoxia has been shown to induce cancer cell migration [573, 574], extracellular matrix (ECM) modification [575], and recruitment of blood vessels [576]. We have shown that low O2 leads to faster cell migration in sarcoma [573] via HIF-

1, which upregulates collagen modifiers such as lysloxidase (LOX) and procollagen-lysine, 2- oxoglutarate 5-dioxygenase 2 (PLOD2) [522, 575]. Hypoxic conditions in the tumor environment have also been shown to upregulate the secretion of angiogenic factors such as vascular endothelial

145 growth factor (VEGF) [218]. Overall, the hypoxic microenvironment regulates multifaceted pathways that are crucial in the development and progression of cancer.

A similar mechanism by which hypoxia upregulates VEGF and other effectors of cancer cell function in the tumor microenvironment has also been shown to affect various vascular diseases and play a role in wound healing and tissue regeneration. For example, hypoxic upregulation of VEGF secretion has been documented in macular degeneration, wherein high levels of VEGF cause uncontrolled blood vessel growth resulting in poorly formed or leaky vessels, which can lead to blindness [577-579]. In wound healing, hypoxia plays a crucial role in the enhancement of blood vessel recruitment and inflammatory responses. Here, a reduced blood flow at the injury site leads to a locally low O2 environment, which in turn initiates angiogenesis and wound healing. Further, the hypoxic environment in wounded blood vessels increases the expression of intercellular adhesion molecule 1, a key factor in recruiting endothelial progenitor cells [580, 581] and immune cells [582] to the injury site. Diseases which commonly present with

HIF-1 overexpression can lead to chronic wounds and fibro-proliferative disorders, such as hypertrophic scars [583]. Conversely, low levels of HIF-1 are commonly seen in diabetic ulcers and in elderly patients, which may lead to impaired wound healing [584].

Responses to hypoxia are required for the normal development of the vascular system [585,

586]. In early embryonic development, fast growing tissues must coordinate growth with O2 delivery. O2 in this environment is critically low and acts as a signal to stimulate vascular growth

[587] and endothelial cell differentiation [588]. It has also been suggested that during embryonic vessel development, regions of relative hypoxia attract growing blood vessels [332]. During the early stages of embryogenesis, a primitive vascular pathway is formed through vasculogenesis, where progenitor cells form an immature vascular plexus [587]. Once this primitive vascular

146 network is formed, endothelial cells form new capillaries via angiogenesis. This pathway stimulates VEGF and other angiogenic factors via low O2, which facilitates blood vessel formation

[589]. Overall, our current understanding suggests that vasculogenesis and angiogenesis in mammalian tissues occur in hypoxic conditions [590-592]. The vital roles of hypoxia and O2 gradients are made clear through their critical role in development, tissue regeneration, and cancer/tumor growth and progression.

Table 6.1: Definitions of Oxygen Concentration and Usual In Vivo Oxygen Distribution

As shown in ref [182]

Term %O2 Non-hypoxia 21 Hypoxia <5 Anoxia <0.1 Organ %O2 Aterial blood 13.2 Brain 4.4 Liver 5.4 Kidney 9.5 Muscle 3.8 Bone marrow 6.4

6.3 Methods to culture under hypoxic conditions

Variations in O2 tension in healthy or diseased tissues occur as gradients ranging from anoxic (0%) to normoxic (~13%) depending on blood supply and tissue O2 demands (Table 6.1).

Nonetheless, while hypoxia plays a pivotal role during various cellular processes, the design and utilization of a 3D hypoxic-gradient microenvironment to mimic the in vivo niche has only recently been realized.

147

To control O2, alternative in vitro research platforms use specialty equipment such as hypoxic chambers, hoods and incubators, which don’t allow for precise control over dissolved oxygen (DO) concentration in the cell’s local microenvironment, thus limiting the ability to study cell responses to O2 gradients. Numerous methods exist in the literature to combat this issue. These include many 2D systems with precision spatial and temporal control of O2 [593, 594] and even

O2 gradients [190, 523]. Additionally, methods for control of O2 tension and O2 gradients compatible with 3D cell encapsulations have been previously developed [189, 191, 524, 595].

While these platforms may be preferred for some experimental setups and should be considered in an application dependent manner, several general limitations exist. For example, many of the methods describe complex platforms requiring microfabrication techniques [189] and intricate pump and flow setups [596], the use of hypoxic incubators and supplemental gas tanks [597], the need for high cells concentrations (107-108 cells/ml) [598], and prohibit real-time monitoring and imaging [599]. To address the above-mentioned limitations and the increasing interest in study of cell responses to O2 gradients, we developed O2-controllable hydrogels that can serve as 3D hypoxic-gradient microenvironments.

Our hydrogel polymeric network is formed by the di-ferulic acid (diFA) formation in laccase-mediated crosslinking reaction with O2 consumption (Fig. 6.1A). Following completion of the reaction, the low dissolved oxygen level is maintained through O2 consumption by the cells or tissue [600]. The crosslinking reaction allows for the cells to be rapidly exposed to a hypoxic environment, without the need to wait for a spontaneous (cell-mediated) gradient to form. This is beneficial because it takes time for these cell-mediated gradients to form spontaneously and the construct cannot be guaranteed to still be mechanically stable. A biomimetic extracellular environment is crucial for research into vascular differentiation and development, in which 3D

148 networks are formed though cell-matrix interactions including adhesion and degradation [219, 220,

601]. In the case of cancer cells, O2 gradients can form in sufficiently thick 3D tissue but the gradients are transient and very difficult to control. Finally, it should also be noted that rapid, predictable O2 gradients cannot form when encapsulating tissue grafts or when transplanting non-

O2 controllable materials in vivo. Our platform circumvents these pitfalls by rapid O2 consumption during hydrogel formation via laccase-mediated crosslinking reaction. When the chemical reaction is completed, the O2 consuming reaction is terminated and atmospheric O2 diffusion begins. In vitro, the O2 levels within the gel can be simply controlled by varying the height of the hydrogel construct. The gradients in the gel can last up to 1-2 weeks, allowing for long term hypoxia experiments, while minimizing the generation of reactive oxygen species (ROS) that may be elicited through rapid re-oxygenation. These gels can be fabricated in traditional polystyrene 96- well plates, and once cells are encapsulated or microtissues are embedded, the constructs are cultured with traditional cell culture media in a standard incubator. The DO levels and gradients are controlled by varying hydrogel thickness in a volume-dependent manner in a well in 96 well- plates (e.g., nonhypoxic hydrogels, 45 L; hypoxic hydrogels, 90 L). In addition to using conventional cell culture equipment, these gels use a small volume of polymer solution, typically

45-100 L, and therefore a small volume of reagents can be used for molecular biology assays, such as PCR, histology, and immunofluorescent analysis. These gels are optically clear and therefore are ideal for the real-time measurement of O2 and cell migration, as well as in-situ staining for immunofluorescent imaging. Furthermore, the platform can be used to screen small molecular inhibitors [522]. In vivo experiments utilizing hydrogels as therapeutics or to study the effects of O2 gradients in disease models may also be designed and performed.

149

Cells encapsulated in hydrogels can be administered to mice. O2 gradients and severity of hypoxia in these cell grafts are dependent on two main factors in vivo: (1) hydrogel volume and

(2) surrounding tissue O2 concentration. Hydrogels with larger relative volumes will have a more severely hypoxic core, and thus a steeper O2 gradient. If hydrogels are injected or implanted in tissues with high relative O2, steep gradients may be obtained. Those tissues with lower relative

O2 will exhibit shallower gradients [600]. We have demonstrated that insights into the role of varying O2 during tumor growth, vascular development[602] and angiogenesis [194, 600] can be acquired using these hydrogels.

Figure 6.1: Schematic representation of the gelatin-based hypoxia-inducible (Gel-HI) hydrogels for in vitro and in vivo applications. The gelatin-g-ferulic acid (Gtn-FA) precursor solution is dissolved in DPBS at 37 C. It can be mixed with either cells or tissues to provide artificial hypoxic microenvironments with oxygen gradients (see inserted computer simulation of oxygen tension and gradients). The precursor solution can also be directly injected into the animal as a hypoxia-inducible acellular matrix that induces temporal hypoxia and an oxygen gradient in the body (see inserted computer simulation of oxygen tension and gradients).

150

6.4 Development of the protocol

Gelatin hypoxia-inducible (Gel-HI) hydrogel is prepared by chemical functionalization of porcine gelatin. Gelatin has numerous inherent advantages for use as a tool for biological experimentation. Gelatin is produced through denaturation (partial hydrolysis) of collagen, a crucial ECM protein found in nearly all tissues throughout the human body. Thus, gelatin possesses bioactive characteristics with broad implications for many biological applications of numerous cell and tissue types. In particular, the presence of integrin mediated cell adhesion motifs and the amenability to cell-mediated proteolytic degradation highlight the underlying advantages for use of a gelatin backbone. Such properties alleviate the need for user-defined optimization of concentration of cell adhesion motifs and degradable add-ins, leading to a highly biomimetic culture system. Because gelatin is naturally derived, there is a certain degree of batch-to-batch variability, as with any natural polymer. However, widespread use and characterization of gelatin in both 2D culture systems[603], and as 3D platforms [493, 566, 604], along with a high level of bioactivity motivate its use in our system.

Control over O2 tension in our Gel-HI hydrogel necessitates O2 consumption. Additionally, in order to develop gelatin-based culture systems, chemical functionalization to achieve a crosslinking mechanism is necessary, as gelatin only thermally crosslinks below physiological temperature. These two needs define our design approach to develop this system. Importantly, a large number of molecules can be conjugated due to a high number of chemically active functional groups being available, making chemical functionalization simple. In particular, for our application, primary amines are abundant. Less abundant are biocompatible reactions that include physiologically relevant O2 consumption. Laccase, an enzyme derived from mushroom (among other select fungi and trees), has particular specificity in catalyzing reactions between phenol-

151 containing molecules [605]. Specifically, laccase can catalyze the four-oxygen reduction of O2 to

H2O, which results in oxidation of two phenol-containing molecules. Similar to laccase, other enzymes can be adopted to create hypoxia-inducible hydrogel matrices (e.g., tyrosinase and glucose oxidase, which consume molecular O2 during crosslinking of phenolic molecules. Our choice of phenolic moiety required the presence of a carboxyl group opposite the phenol to facilitate chemical conjugation to the gelatin backbone. Ferulic acid (FA) satisfies our design criteria, and thus was conjugated to the gelatin backbone by employing EDC/NHS chemistry. The

FA-conjugated gelatin polymers form hydrogels through diFA formation in a laccase-mediated reaction, resulting in bioactive and O2-controllable Gel-HI hydrogels.

It is crucial to understand how O2 tension and O2 gradients are regulated in Gel-HI hydrogels. As discussed above, Gel-HI hydrogels are used in standard incubators and culture plates. In such conditions, passive diffusion of atmospheric (~21%) O2 is an important consideration. Such diffusion acts to increase the O2 tension in the culture microenvironment. With that, two elements of the system act to decrease O2 tension: (1) laccase-mediated O2 consumption, and (2) cell-mediated O2 consumption. An initial decrease is induced by the laccase-mediated crosslinking reaction, and in acellular experiments, following completion of the reaction, the hydrogel microenvironment returns to atmospheric conditions within hours. Hydrogels with cells encapsulated maintain low O2 (< 5% is considered hypoxic) conditions for 3-7 days in culture.

Maintenance of low O2 is highly dependent upon cell concentration and cell type[522, 606]. In general, higher cell concentrations lead to prolonged maintenance of hypoxia. Additionally, cell types vary in rates of O2 consumption, and thus understanding cell O2 needs is critical to experimental design. These parameters should be considered as a key factor to accurately mimic the native tumor (or ischemic) microenvironment, as cellular activities are highly variable within

152 the dynamic O2 microenvironment. Growing evidence has demonstrated that transcription factors

(e.g., HIFs), which regulate cellular activities, accumulate when the cells or tissues are exposed to hypoxia, [10]. For example, cells may change their metabolism to reduce O2 consumption by upregulating GLUT gene expression in hypoxic environments. In addition, HIFs activate various cellular processes, including metabolism, proliferation, invasion and metastasis, and tumor- angiogenesis. In our previous study, we also demonstrated that mouse sarcoma cultured within 3D hypoxic matrices upregulated numerous hypoxia-related genes including GLUT-1, MMPs, and

VEGF. Furthermore, there are variations in the cellular O2 consumption rate depending on the cell type and number within the artificial hypoxic microenvironments. Thus, we should consider the variations in the cellular metabolism within the artificial hypoxic microenvironments as a critical design parameter throughout experimental in this hydrogel system. Lastly, because of the constant and delicate balance between active O2 decreases and passive O2 increases, the volume of the hydrogel plays a significant role in in vitro assays. Steepness of O2 gradients and maintenance of hypoxic conditions rely heavily on the volume, and perhaps more specifically, the height of the hydrogel. In smaller volume (decreased height) hydrogels, atmospheric concentrations of O2 diffuse more quickly throughout the gel, resulting in higher O2 with a shallow gradient. Larger volume (increased height) hydrogels more slowly diffuse O2 throughout the entire volume, resulting in hypoxic conditions with significant O2 gradients.

In addition to controlling O2 tension, other important ECM parameters, such as stiffness and degradability, can be controlled by simply adjusting polymer or enzyme concentration. Further adjustment is described through secondary crosslinking with an additional enzyme, microbial transglutaminase (mTG) in [30][194]. In this manuscript we describe the detailed protocol for synthesis, characterization, manipulation, and application of our versatile Gel-HI hydrogels.

153

Limitations in 3D hypoxic cell culture and O2 gradient studies are alleviated by use of Gel-HI hydrogels, with a wide breadth of use both in vitro as a culture platform, and in vivo as a cell delivery vehicle and as an acellular therapeutic.

6.5 Applications of the protocol

6.5.1 Cancer

Hypoxia is a key regulator affecting cell migration, ECM modification, and blood vessel recruitment during tumor development, growth and metastasis. Gel-HI hydrogels allow us to simultaneously study the effects of a wide-range of O2, whereas most studies focus on a fixed O2, typically drawing comparisons between experiments conducted within the severe hypoxic range

(0.5-1%) and those conducted at atmospheric (21%) O2. One of the key benefits of this platform is the DO gradient along the hydrogel depth, which uniquely provides opportunities to study how gradients regulate cell behavior in vitro in three dimensions. Moreover, this platform not only allows for the use of conventional cell and molecular biology assays, but it also permits real time imaging to track single cancer cell proliferation and study motility/migration. Conventional hypoxic platforms, such as chambers, don’t allow for a controlled gradient to be formed, and are prohibitive for live cell imaging due to the chamber’s size, as it cannot fit conventional microscope set ups. Tumor sections can also be encapsulated in the hydrogel and the migration of cells out of the tumor can be documented. This allows the user to study cells in pathophysiologically relevant

O2 gradients using real-time imaging to elucidate pathways that lead to cancer cell migration and potential metastasis.

6.5.2 Vascular applications

Not surprisingly, the vascular regeneration niche has been studied in great detail, owing to the high mortality and morbidity of cardiovascular disease in the developed world [607]. Complex

154 mechanisms have been uncovered, which have described the interplay between multiple cell types

(including interactions between endothelial cells mural cells [74]), growth factors [21], ECM properties [220], and O2 tension [608]. Specifically, as detailed above, hypoxic conditions have been shown to trigger vascular formation and growth during embryonic development, various vascular-related disorders and cancer. However, studying the effects of O2 tension and O2 gradients in a 3D microenvironment were not been possible until the development of Gel-HI hydrogels. Gel-

HI hydrogels facilitate studies in a highly biomimetic experimental system by harnessing the EC’s own vasculogenic capacity. Many platforms which are used to study vascular network formation, require the addition of hyperphysiological concentrations of pro-angiogenic growth factors or an engineered ‘angiogenic’ or ‘vasculogenic’ cocktail of factors [159, 219]. However, encapsulation of vascular cells in Gel-HI leads to cell-mediated production of angiogenic growth factors and subsequent vascular network formation. Thus Gel-HI provides a system which more closely mimics the native regenerative niche than other culture systems. For example, our lab has shown that early vascular cells derived from induced pluripotent stem cells of diabetic patients can generate lumenized networks in the hypoxic Gel-HI hydrogels, indicating their responsiveness to hypoxic pathways[602].

6.6 Limitations of the platform

No biomaterial is without limitation, whether in function or application. Several limitations have been briefly mentioned in the preceding text, which we will again describe here. First, batch- to-batch variability exists with these hydrogels, as with any naturally derived polymeric hydrogel.

Even with a relatively simple polymer synthesis protocol, variability can be seen in an academic lab setting. As such, it is crucial to maintain consistency in technique throughout every step of the process, including synthesis, purification, storage, and particularly cell encapsulation. To ensure

155 reproducibility, we recommend careful analysis of polymer properties prior to cell/tissue encapsulation experiments. Enzymatic crosslinking is particularly sensitive to temperature and timing, both of which are crucial to maintain experimental reproducibility. Hydrogel precursor solutions, particularly laccase, may become cytotoxic outside of the concentration ranges suggested by this protocol. Laccase, is sensitive to storage conditions and may lose activity with extended storage in aqueous solution (> 2-3 months). Additionally, because the matrix can be degraded in a cell-mediated fashion, over the course of the experiment mechanical properties are subject to cell-mediated change. This property of the hydrogels is crucial to many applications, but can serve as a limitation to culture time. If left in culture long enough, the matrix may completely degrade (usually > 5 days). Further, cell culture time may vary with cell type and encapsulation concentration.

6.7 Experimental design

A vast array of powerful experiments can be developed by following the steps described in this protocol. Examples of several endpoint analyses are included, which can be utilized or omitted as the researcher tailors the experimental design to fit their needs. Polymer synthesis, purification, and characterization are key first stages to any and all experiments following this protocol. Upon initial synthesis, we recommend the gel components undergo a full array of characterization strategies including NMR, UV/vis, mechanical/rheological testing, and O2 monitoring. O2 may be monitored using commercially available O2 sensors (PreSens), which measure local O2 in real time via a fluorescence quenching methodology. Successful hydrogel synthesis requires several benchmarks. First, NMR will confirm conjugation of ferulic acid to the gelatin backbone and UV/vis will determine the concentration of conjugated ferulic acid (40-60

μmol/g polymer). Rheological characterization provides verification of hydrogel formation and

156 measurement of mechanical properties. Finally, acellular O2 measurements in 96 well plates are critical to ensure O2 is reduced to <5% (in 90-100 μl hydrogels), confirming successful establishment of a hypoxic microenvironment. Characterization may be streamlined and strategies omitted once the user has mastered the synthesis portion of the protocol. Following successful synthesis, determination of gelation time is key to ensure a homogeneous encapsulation. This may vary slightly with cell type and concentration, so gelation time measurements should be done on a per-experiment basis. Monitoring of cell morphology may be done via light microscopy or timelapse microscopy and can be coupled with DO monitoring (either over the entire time-course or gradient measurements at specified time points) to gain an enhanced understanding of the microenvironmental properties over time. End point analysis can involve RNA extraction or fixation and immunofluorescent staining for in vitro experiments. In vivo experiments can also incorporate cells and DO monitoring (although only at specified time points) and endpoint experiments may include lectin tail vein injection for EC applications or tissue excision and histology for numerous other applications. Indeed, Gel-HI hydrogels provide an impressive potential for breadth of study and may be adapted to many applications (see Fig. 6.1).

Expertise in cell culture techniques is a necessity to perform experiments using this protocol. Additionally, experience with microscopy, immunofluorescent staining, and microbiological techniques will aid in the range of experiments possible with this protocol.

Background in materials science and polymer synthesis is suggested, but not required, as the synthesis protocol is relatively simple.

6.8 Anticipated Results

Gel-HI hydrogels provide a uniquely biomimetic microenvironment that can be used for a variety of applications. In particular, the ability to control hypoxic gradients and overall O2 tension

157 may be applied to vascular applications [194, 600] (see Fig. 6.2). Vasculogenesis and angiogenesis often rely on the hypoxic microenvironment to facilitate formation or invasion of new blood vessels. We anticipate Gel-HI hydrogels may identify novel insights into such vascular development and regenerative processes. This knowledge may be used to inform targets for drug discovery, as well as aid in the design of vascularized tissue constructs. Overall, the ability to mimic the physiological, or pathophysiological, microenvironments in vitro will likely provide a platform for continued innovation.

For cancer research, Gel-HI hydrogels provide a powerful platform to study single-cell responses to hypoxic gradients, such as matrix remodeling and cell migration [573]. As a display of the potential breadth of study, cellular mechanisms can be analyzed with standard cell and molecular biology techniques. The Gel-HI hydrogels can also provide a platform for analyzing tumor biopsies that can be grafted into the gel to analyze cell migration out of the graft into the surrounding hydrogel, as an in vitro corollary to tumor growth, migration, and potential metastasis.

As a tool in cancer biology, Gel-HI hydrogels may identify novel pathways for tumor growth and migration induced only in hypoxic conditions, as well as provide a method to screen cancer therapeutics and determine potential pitfalls or unintended consequences that cannot be studied with standard in vitro platforms.

During in-vivo implantation of the hydrogel real-time O2 measurements can be observed using the needle type O2 sensor (see Fig. 6.3). After implantation of the hydrogel an analysis of the vasculature can be performed to look at vascular recruitment to the wound as well as the number of vessels. If the proper O2 is maintained recruitment should be clear under histological analysis and fluorescent staining. The translational potential of Gel-HI hydrogels is immense, as hypoxia governs production of many signaling molecules that may influence cell behavior.

158

Overall, the Gel-HI hydrogels have a great potential either as 3D artificial hypoxic cellular microenvironments or as hypoxia-inducible acellular matrices for the treatment of vascular disorders. The HI hydrogels also provide an innovative platform as an engineered tumor microenvironment to study basic cancer biology and to screen the drug candidates for better clinical outcome in cancer treatment. In addition to applications discussed, looking forward, we anticipate the use of Gel-HI hydrogels will provide a new platform for innovation in the stem cell and developmental biology fields. We anticipate studying the effects of developmentally relevant

O2 gradients will lead to conclusions previously unattainable with available research tools. Such revelations may lead to advances in tissue engineering and regenerative medicine, as well as accelerate the understanding of the role of hypoxic gradients in development and tissue regeneration.

159

Figure 6.2: Tubulogenesis in Gel-HI. (a) Light microscope images of early vascular cells

(EVCs)-derived from induced pluripotent stem cells in the Gel-HI hydrogels (hypoxic vs. nonhypoxic) encapsulated at 2 million cells/ml. Scale bars 100 μm. (b) Dissolved oxygen (DO) levels measured in EVC-Gtn-HI constructs encapsulated at 2 million cells/ml. Measurements were taken immediately following encapsulation. (c) Fluorescent microscope images of EVCs after 6 days of culture. Phalloidin (red) merged with DAPI (blue). Scale bars 500 μm. (d) Quantitative analysis of vascular tube formation. Results are shown as averages ± standard deviation. T-test significance levels were defined as * p < 0.05 and **p < 0.01.

160

Figure 6.3: Subcutaneous and intramuscular in vivo oxygen sensing. Commercially available needle type oxygen sensors (PreSens) were utilized to measure partial pressure of oxygen over time both (a) subcutaneously and (b) intramuscularly in 10-12-week-old nude mice. Sensors remained stationary and readings equilibrated after approximately 60 minutes to 7.4% and 3.0% in subcutaneous and intramuscular regions, respectively. A temperature probe was used concomitantly to ensure the accuracy of the oxygen readings. The animal study was performed using a protocol (MO16A284) as approved by The Johns Hopkins University Institutional

Animal Care and Use Committee.

161

CHAPTER 7

7 Hypoxia and Matrix Viscoelasticity Sequentially Regulate Endothelial Progenitor

Cluster-Based Vasculogenesis

7.1 Teaser

A broadened perspective on neovascularization is achieved by mimicking the vascular regenerative microenvironment in vitro.

7.2 Introduction

Functional vasculature is critical for tissue homeostasis. Thus, the formation of neovasculature, vascular morphogenesis, is a hallmark of tissue development and regeneration, as well as cancer growth and metastasis. An in depth understanding of the mechanisms governing vascular morphogenesis is critical to the identification of novel therapeutic targets and refinement of therapeutic strategies. Numerous studies have elegantly uncovered many key regulators of angiogenesis and vasculogenesis. A mechanistic understanding of ‘classical’ single cell vasculogenesis has been defined and refined over the last two decades by employing intricately designed in vivo models, including those in both chick and mouse embryos [354, 609-611], as well as zebrafish [98, 612, 613]. Hydrogels, which can mimic the native three-dimensional (3D) extracellular matrix (ECM), have been used successfully to analyze and control vascular regeneration from human cells [91, 97, 160, 219, 614] and to understand ‘classical’ single cell vasculogenesis from human cells [97, 219]. In brief, ‘classical’ single-cell vasculogenesis commences when ECs establish interactions with their surrounding ECM through integrin engagement [221, 222, 615]. These cells then respond to an array of pro-angiogenic cues, including biological factors and physical factors, such as matrix stiffness [152, 616], to re-organize their cytoskeleton to form intracellular void spaces, termed vacuoles. These vacuoles then coalesce

162 intercellularly and intracellularly to form lumen [98, 221, 222, 615]. Sprouting and branching concurrent with ECM degradation and ECM remodeling culminate in the formation of perfusable, nascent vasculature [97, 221, 222, 615]. Both membrane-type and soluble matrix metalloproteinases (MMPs) are crucial throughout this process for matrix degradation and cell migration [219, 458]. While these studies have uncovered many details of vascular morphogenesis that have transformed the understanding of how blood vessels form and re-form, only a handful of studies describe another vasculogenic mechanism, termed herein ‘cluster’ vasculogenesis.

Cluster vasculogenesis from circulating endothelial progenitor cells (EPCs) has been documented in animal models, but the underlying mechanism has not been studied. Interestingly,

EPC-cluster vasculogenesis has only been observed in hypoxic settings in vivo, including during development [617], ischemia [618], and tumor vascularization [619], suggesting the importance of low oxygen (O2) tension as a cue for this novel mechanism. Here, circulating EPCs are mobilized and home to hypoxic/ischemic tissue through cytokine and growth factor gradients, particularly of stromal derived factor 1 (SDF-1), as well as gradients of O2 [337]. Recruited EPCs then attach to the activated endothelium local to the defect through hypoxia-upregulated integrin

β2 and intercellular adhesion molecule-1 (ICAM-1) [339, 379]. Finally, EPCs extravasate through the endothelium to the interstitial tissue, form clusters, then sprout to anastomose with existing blood vessels, thus revascularizing the local microenvironment [618, 619]. While the recruitment of circulating EPCs to regions of hypoxia/ischemia has been well defined, it is not understood how

‘clusters’ form and what drives subsequent vascular sprouting.

Hypoxia, defined as low O2 tension (<5%), has become recognized as a potent regulator of cell function and morphogenesis. Hypoxic microenvironments, accompanied by hypoxic gradients, exist throughout embryonic development [83], in tissue regeneration [620], and in tumor

163 angiogenesis and progression [218], and are thus a critically important parameter to consider to fully understand these processes. Hypoxia stabilizes the transcription factors hypoxia inducible factors (HIFs), which can upregulate numerous pro-angiogenic factors, including vascular endothelial growth factor and MMP [181, 193, 608, 621]. Further, hypoxia can act in a HIF- independent manner to promote production of reactive oxygen species (ROS) [622, 623]. ROS can influence vascular morphogenesis and endothelial fate commitment [66, 187], thus providing a basis for their role in cluster formation. Additionally, ROS are present in the inflammatory microenvironment in which neovessel formation is common [624].

Current approaches to study the effect of hypoxia on EPCs fail to recapitulate the range of

O2 tensions present in pro-angiogenic microenvironments, and thus may not capture the same responses observed in vivo. We developed oxygen-controllable hydrogels, which provide a means to establish and study the effects of hypoxic gradients in vitro [192-194, 522, 565]. These hydrogels provide a 3D system to examine EPC responses to a controllable microenvironment that recapitulates key parameters of the pro-regenerative milieu in which endothelial cells establish new blood vessels. These hydrogels have enabled us to elucidate a novel two-step mechanism for cluster vasculogenesis, including EPC hypoxic cluster formation and stabilization, followed by vascular sprouting into the stiffer, un-degraded matrix and network formation.

164

7.3 Results and Discussion

7.3.1 Hypoxia mediates endothelial cluster formation

Gelatin-based hypoxia-controllable hydrogels (Gel-HI) are prepared by conjugating phenol-containing ferulic acid to a gelatin backbone, followed by enzymatic crosslinking using a laccase-mediated reaction that consumes O2, resulting in hypoxic conditions within the hydrogel

[192-194, 522, 565]. Gelatin, hydrolyzed collagen, provides intrinsic bioactivity, which enables cellular adhesion, migration, degradation, and remodeling, all of which are necessary for vascular regeneration [97]. Interestingly, because these hydrogels are analyzed in conventional incubators

(5% CO2, atmospheric O2), O2 diffusion from the atmosphere competes with the O2-consuming crosslinking reaction to create a gradient throughout the hydrogel. Depending on the height of the hydrogel, the O2 gradient can be varied to create both hypoxic and non-hypoxic conditions (Fig.

S7.1A, B). Following completion of the crosslinking reaction, O2 returns to atmospheric conditions as O2 diffuses into the gel (Fig. S7.1B).

We found that encapsulation of a subtype of EPCs, endothelial colony forming cells

(ECFCs), at a high concentration in Gel-HI hydrogels leads to cluster formation in hypoxic, but not non-hypoxic hydrogels (Fig. 7.1 A-C). By 6-12 hours after encapsulation, ECFCs commence cluster formation (Fig. 7.1B). At this timepoint, when clusters begin to form, an O2 gradient is present within the hydrogel (Fig. S7.1C, D). Thus, analyzing discrete z positions within our hydrogels revealed cluster formation was consistently initiated at specific z positions, namely at

~250 μm above the bottom of the plate, corresponding to ~1% O2 (Fig. 7.1D, E; S7.1D, S7.2).

Through 24 hours and up to 48 hours, clusters increase in size in hypoxic conditions (in terms of number of ‘cells in clusters’) and fall towards the bottom of the hydrogel. Accordingly, the number of ‘single cells’ decreases as the number of ‘cells in clusters’ increases (Fig. 7.1E, S7.2). We

165 observed consistent cluster size up to 48 hours, suggesting the clusters we observe are the requisite size for this novel cluster-based vasculogenesis mechanism. Cells that participated in cluster formation appear to remain spherical throughout the 48 hour experiment (Videos S1-S4; included in published manuscript [195]). In this case, we postulated that encapsulated ECFCs degrade their surrounding matrix and passively migrate to the space voided by degradation. In non-hypoxic hydrogels, clusters do not form, and cells remain isolated as ‘single cells’ with cell elongation and vascular sprout formation (Fig. 7.1F, G; Videos S5-S8 [195]). Video S5 observations in non- hypoxic conditions (at z=0) show classical endothelial sprout formation by 24 hours. A comparison of this mechanism with the mechanism governing cluster formation displays a clear distinction between the two methods for cell movement and morphology.

To ensure nutrient deprivation is not the determining factor guiding cluster formation, we generated hypoxic conditions in non-hypoxic hydrogels using a custom gas flush chamber with O2 sensing. Indeed, exposing ECFCs encapsulated in hydrogels with non-hypoxic geometry to 1% O2 resulted in similar kinetics of cluster formation and nearly identical O2 measurements (Fig. S7.3A,

C). ECFCs in non-hypoxic control hydrogels did not form clusters (Fig. S7.3B). Quantification of percent area covered by clusters confirms that hypoxia, rather than nutrient deprivation, drives cluster formation (Fig. S7.3D). To summarize, ECFCs exposed to hypoxic conditions form cell clusters after 6-12 hours, thereby recapitulating the first step of an alternative mechanism for vascular morphogenesis.

166

167

Figure 7.1: ECFC clusters form only in hypoxic conditions. (A) Schematic for hypoxic and non-hypoxic cell encapsulation. (B) Bright field images of cell morphology in hypoxic and non- hypoxic hydrogels up to 48 hours. Hypoxic hydrogels exhibit cluster morphology starting at approximately 6-12 hours in culture. Clusters expand up to 48 hours. ECFCs encapsulated in non- hypoxic hydrogels do not form clusters, but rather branch and sprout. (C) O2 tension, measured at the bottom of the gel over 4 days in culture. (D) Time-lapse microscopy and (E) subsequent quantification of number of cells in clusters (upper panel) and single cells (lower panel) confirm observations of cluster formation kinetics in hypoxic hydrogels. (F), (G) Clusters do not form in non-hypoxic hydrogels.

7.3.2 ECFC cluster formation is governed by reactive oxygen species (ROS) production and dependent on matrix stiffness

To determine whether HIFs play a role in cluster formation, we co-encapsulated ECFCs with either DMOG or CoCl2. These potent PHD inhibitors, which act to stabilize HIFs in non- hypoxic conditions, have been used extensively in vitro and in pre-clinical models [543, 544, 625].

Interestingly, neither DMOG nor CoCl2 treatment resulted in cluster formation in non-hypoxic geometries (Fig. 7.2A; S7.4A). ECFCs subjected to DMOG treatment remained rounded up to 48 hours, while those exposed to CoCl2 exhibited classical sprouting morphology, although to a similar extent as those cells in control non-hypoxic hydrogels. Small interfering RNA (siRNA) studies further showed that suppression of HIF-1α does not impact cluster formation (Fig. 7.2B).

These experiments suggest that HIF1α does not play a major role in cluster formation. To determine if ROS production induces cluster formation, we utilized a ROS inhibitor, DPI [66,

187]. Addition of DPI inhibited cluster formation in hypoxic hydrogels (Fig. 7.2C; S7.4B). In non- hypoxic controls, DPI inhibited endothelial sprouting, as seen in previous studies in 3D hydrogels

168

(Fig. S7.5) [187]. In summary, hypoxia is a crucial factor which acts to promote ECFC clustering through HIF-independent ROS production.

In addition to hypoxia, another important parameter that guides vascular morphogenesis is matrix stiffness [152, 194, 616]. The oxygen-controllable hydrogels can be designed to vary initial matrix stiffness by adding a secondary crosslinker, microbial transglutaminase (mTG) [194]. mTG is a biocompatible crosslinker that has been used in numerous other hydrogel applications [493,

566]. By crosslinking unconjugated primary amines on the gelatin backbone (Fig. 7.2D), addition of mTG can significantly increase the initial matrix viscoelasticity (Fig 7.2E). Encapsulating

ECFCs within dual-crosslinked Gel-HI hydrogels yields a profound reduction in cluster formation in hypoxic hydrogels (Fig. 7.2F-G). While we attribute a reduction in cluster formation to an increase in initial matrix stiffness, addition of mTG may also influence the matrix microstructure, including pore size, which may also impact cluster formation.

169

Figure 7.2: Cluster formation is dependent on hypoxia-induced ROS and matrix stiffness.

(A) Treatment with PHD inhibitors (DMOG, CoCl2), which act to stabilize HIFs, did not result in cluster formation in hydrogels with the same dimensions as non-hypoxic (NH) hydrogels. NC – no clusters. Scale bars 100 μm. (B) siRNA knockdown of HIF1α did not affect cluster formation.

170

Scale bars 100 μm. (D, E) Addition of mTG increases the stiffness of hydrogels. (F, G) Upon addition of mTG, robust morphological changes can be observed. Most notably, decreased cluster size indicate the importance of matrix mechanics. (F) Timelapse microscopy and (G) quantification of number of cells in clusters (upper panel) and single cells (lower panel) confirms observations of decreased cluster size. All graphical data are reported as mean ± SD. **** p<

0.0001.

7.3.3 Protease-mediated matrix degradation is required for hypoxic cluster formation and stabilization

Proteolytic degradation is a key mediator of cell migration and morphogenesis. To understand the role of matrix degradation in cluster formation, we established an assay utilizing

DQ-gelatin, which fluoresces as it is degraded by proteolysis, and can identify a broad range of protease activity with high sensitivity [626]. We began by encapsulating fluorescent microspheres in our Gel-HI hydrogels, with or without MMP (collagenase IV) and tracked cluster formation and relative fluorescence with timelapse microscopy (Fig. S7.5A-D). Clusters formed in the MMP- supplemented hydrogels only (Fig. S7.5C, D), where clusters began to form concurrent with increases in fluorescence (Fig. S7.5B-D). It is important to note that the microspheres used here may interact with one another through the hydrophobic effect, suggesting that this is not simply a random degradation-clustering event, but rather a more complex phenomenon requiring interactions between participating particles. As cells are not hydrophobic, this further highlights the importance of cell-cell interactions throughout the cluster formation process. We next used the

DQ-gelatin assay in the ECFC-Gel-HI hydrogels. We observed cluster formation in hypoxic, but not in non-hypoxic hydrogels (Fig. S7.5E, F) with a significant increase in fluorescence in hypoxic

171 conditions after 24 hours, indicating the importance of protease production for cluster formation

(Fig. 7.3A). It should be noted that an increase in fluorescence in non-hypoxic conditions supports the importance of protease production in ‘classical’ vasculogenesis. [219, 458]. Moreover, fluorescence steadily increases starting at approximately 6 hours post-encapsulation and continues to increase until it reaches saturation at approximately 18 hours, corroborating our observations of cluster formation at 6-12 hours post-encapsulation (Fig. 7.3B). Rather than quantitatively identifying concentrations of proteases present, the trend of increased fluorescence informs our understanding of the temporal regulation of cluster formation via matrix degradation.

As matrix degradation is critical for cluster formation, we next sought to identify which proteases specifically regulate this process. We used a proteome profiler human protease array to identify relative quantities of soluble proteases in media collected from hypoxic and non-hypoxic hydrogels. Interestingly, many proteases were present in the media in both conditions, including a bevy of cathepsins and MMPs, as well as ADAMTS1 and others (Fig. 7.3C; S7.6A). Both cathepsins and MMPs have been inextricably linked to vasculogenesis [97, 627]. Only one protease, MMP-1, was significantly more prevalent in hypoxic versus non-hypoxic conditions. To confirm the importance of MMP-1, we added a broad-spectrum inhibitor of MMPs, GM6001, and saw inhibition of clusters in the presence of this inhibitor (Fig 7.3D; S7.6B). It should be noted that a high concentration (1 mM) of GM6001 was required to inhibit cluster formation. Lower concentrations, closer to those previously published (5-500 μM) [219], did not inhibit cluster formation (data not shown).

There are a number of potent upstream regulators of MMP, but perhaps most notably, ROS can rapidly induce MMP production in ECs [628]. This rapid induction is of particular interest here, as cluster formation commences at 6-12 hours post-encapsulation. To identify the

172 contribution of ROS to MMP upregulation, we performed the same experiment, but with control

(no treatment) hypoxic hydrogels and hypoxic hydrogels with the addition of DPI. We observed cluster formation in control hydrogels, but not DPI-treated hydrogels (Fig. S7.6C), alongside a reduction in the fluorescence of DPI-treated hydrogels, suggesting MMP upregulation is at least partially regulated by ROS production (Fig. S7.6C). A significant increase in fluorescence in both conditions indicates that protease production is not completely blocked by DPI.

To generate clusters in non-hypoxic conditions, we supplemented the cell media with MMP and found that with increasing concentrations of MMP, cellular clusters formed more rapidly (Fig.

7.3E; S7.6D). With the highest concentration (100 μg/ml) the hydrogels were almost completely degraded by 24 hours. The reduction in area covered by clusters between 24 and 48 hours is due to an increase in cell cluster density between those two timepoints. In other words, cells are packed more tightly, thus covering a decreased percentage of the field of view. With the intermediate concentration (10 μg/ml), the kinetics and morphology of clusters more closely resembled hypoxic conditions. The lowest concentration of MMP (1 μg/ml) yielded sprouting reminiscent of

‘classical’ single-cell vasculogenesis, but no cluster formation. These results suggest that an unprecedented concentration of MMPs required for cluster formation compared with those analyzed from patient samples and those typically associated with ‘classical’ single-cell vasculogenesis [629-633].

Through this process, cell clusters remain intact as they grow in size and migrate through the ECM, indicating an active interaction between cells. ICAM-1 and integrin β2 are vital towards attachment of circulating EPCs to the local activated, hypoxic endothelium [339, 379]. Indeed, both proteins were visible at the membrane of ECFCs within hypoxic clusters but cytosolic in non- hypoxic conditions, indicating cell-cell interactions within the hypoxic clusters (Fig. 7.3G).

173

Vascular endothelial cadherin (VE-cad) is crucial for vessel stabilization and barrier function, and can control a plethora of other EC signaling pathways [634-636]. Here, we observe VE-cad expression localized at the cell membrane in hypoxia, indicative of cell-cell interactions stabilizing

ECFC clusters (Fig. 7.3F). F-actin (phalloidin) staining does not show intracellular stress fibers, but rather is localized at the cell membrane with structures resembling filopodia, revealing that, in addition to stabilization, cells on the periphery of the clusters begin to send protrusions into the surrounding ECM (Fig. 7.3F), preparing to sprout.

ECFC cluster formation is dependent on hypoxia-induced MMP-1-mediated matrix degradation. As clusters grow in terms of number of cells in clusters, they are stabilized by cell- cell interactions governed by intercellular proteins such as VE-cad, ICAM-1, and ITG-β2.

174

175

Figure 7.3: MMP-1 -mediated matrix degradation is required for cluster formation and stabilization. (A) DQ quantification along culture period. Increasing fluorescence indicates the presence of proteases. H n=6, NH n=4 (B) Representative quantification of fluorescent signal over the course of the experiment. (C) Proteome profiler protease array showing only MMP-1 is increased in hypoxic compared to non-hypoxic conditions. Results are presented as the mean pixel density of two arrays. (D) Media supplemented with 1 mM GM6001 inhibited cluster formation in hypoxic hydrogels. (E) Exogenous MMP supplementation facilitates dose dependent cluster formation in non-hypoxic hydrogels. (F) Cluster stabilization is shown in hypoxic conditions by

VE-cad localization between some, but not all, cells in clusters, indicating cell-cell interactions

(arrowheads). F-actin phalloidin staining localized at the cell membrane shows structures resembling filopodia (arrows). Clusters are rare in non-hypoxic hydrogels, but some cellular sprouting is present. (G) ICAM-1 and ITG-β2 are present at the cell membrane within hypoxia- induced clusters, indicating cell-cell interactions. Only a few cell-cell interactions are observed in non-hypoxic conditions. Scale bars are 20 μm. Graphical data in A, D, and E are reported as mean

± SD. * p < 0.05, **p < 0.01, and **** p< 0.0001.

7.3.4 Increased viscoelasticity promotes vascular sprouting from clusters

As anticipated in this hydrogel system, prolonging the cell culture period beyond 48 hours reveals extensive sprouting from clusters, which rapidly establish vascular networks extending to up to 500 μm in length by 72 hours (Day 3) (Fig. S7.7A). Critically, cluster-based vasculogenesis yields vascular networks much larger than those that materialize through single-cell vasculogenesis in non-hypoxic hydrogels (Fig. S7.7A). Orthogonal projections of confocal z- stacks show two important additional characteristics of these vascular networks. First, both multicellular and single cell luminal structures are present, thereby confirming the three-

176 dimensionality of these structures, as well as their potential as nascent functional vasculature (Fig.

7.4Ai). Second, vascular networks extend in the XY, XZ, and YZ planes, but only to Z planes below clusters (lower Z plane) (Fig. 7.4Aii). Because vascular sprouting requires cell-ECM interactions [97, 637], it is possible that cluster formation, which necessitates extensive matrix degradation, limits network formation in the degraded matrix above (higher Z plane) the areas of

ECFC clusters. There is also the possibility that vascular networks extend to the hypoxic regions.

Nonetheless, increased cell-ECM interactions can be observed where after 24 hours in culture

(D1), cells do not contain clear intracellular cytoskeletal structures, but rather diffuse F-actin stress fibers, with filopodia extending into the surrounding ECM. By 72 hours in culture (D3), cells establish cytoskeletal structures and contain an increased number of stress fibers, confirming cell-

ECM interactions (Fig. 7.4B; S7.7B, C). To accelerate vascular network formation and maximize potential for cell-ECM interactions, we supplemented cell media with mTG thus enabling user- controlled increases in matrix viscoelasticity. We found that supplementation with mTG resulted in an approximately two-fold increase in matrix viscoelasticity (~200 Pa prior to addition of mTG,

~400 Pa after addition of mTG; Fig S7.8A). Addition of mTG following cluster formation (D1) results in accelerated sprouting on the second day (D2) of culture and ultimately results in extensive vascular network formation (Fig. 7.4C; S7.8B). Indeed, while the length, area, and thickness of vascular tubes remains similar in both control and mTG treated samples (Fig. S7.8C-

E), the area covered by vascular networks is enhanced in mTG treated hydrogels (Fig. 7.4D).

These changes are accompanied by increased cytoskeletal tension, as indicated by an increase in the number of stress fibers per cell, highlighting the elevated role of cell-ECM interactions following dynamic increases in matrix stiffness (Fig. 7.4E; Fig. S7.9). Complementary to the increase in vessel coverage, we also noticed that lumen widths anecdotally appear larger in the

177 mTG-treated hydrogels (Fig. S7.9A-B). Cluster-based vasculogenesis results in the rapid formation of extensive vascular networks in Gel-HI hydrogels. These vascular networks can be controlled by user-defined increases in matrix stiffness.

178

Figure 7.4: Sprouting is enhanced via ECM interactions and as matrix stiffness increases.

(A) Orthogonal projections, XY (bottom left panel), YZ (top left panel), and XZ (bottom right panel), show (i) both multicellular (arrow) and unicellular lumen (arrowheads) are present in vascular networks, and (ii) vascular networks spread in three dimensions (XY, YZ, and XZ), but only in the (-) Z direction. (B) Cell-ECM interactions increase from 24 to 72 hours. (i) At 24 hours

(D1), ECFCs exhibit membrane or diffuse cytosolic F-actin (phalloidin) localization. ECFCs begin to show filopodia extending into the surrounding matrix (arrowheads). At 72 hours (D3), ECFCs

179 establish cytoskeletal F-actin fibers indicative of cell-ECM interactions (arrows). (ii) Stress fiber quantification reveals an increased number of stress fibers per cell between 24 and 72 hours. (iii)

Representative quantification data for a cell at 24 hours (gray) and 72 hours (black). Graphical data in B are reported as mean ± SD of quantification of 123 cells (24 hrs) and 100 cell (72 hrs).

**** P<0.0001. (C) Vascular network formation is accelerated with addition of mTG as shown by

(i) bright field images on day 2 of culture (sprouts indicated by arrowheads) and (ii) subsequent quantification. Ctl n=9 gels. mTG n=18 gels. (D). Vascular tubes cover a higher percentage of the analyzed fields of view with the addition of mTG. Ctl n=5 gels. mTG n=8 gels. D reported as mean

± SD. * p < 0.05. (E) Stress fiber quantification for control (red) and mTG treated hydrogels (gray), measured at 72 hours. Graphical data in E are reported as mean ± SD of quantification of 120 cells

(ctl) and 118 cells (mTG). **** P<0.0001.

7.3.5 ECFCs form clusters in vivo

To evaluate the relevancy of our proposed model in vivo, we used a plug assay, utilizing the oxygen-controllable hydrogel system. As EPC recruitment to ischemic tissues has been well defined [337, 339, 379, 618], we focused on cell-behavior in a controlled environment post- recruitment. Using this approach, we were able to mask the variable effects of cell recruitment as dependent on O2, by injecting hydrogels loaded with SDF-1α [262, 338]. To determine early stages of EPC behavior in the hypoxic environment, we also injected green fluorescence protein (GFP) tagged-ECFCs via intra-cardiac injection (Fig. 7.5A) to follow their organization once recruited to the hydrogels. To control the O2 microenvironment in vivo, we injected the hydrogels with either

CaO2, an O2 releasing molecule, for non-hypoxic hydrogels, or Ca(OH)2 a decomposition product of CaO2, for hypoxic hydrogels (Fig. S7.10A) [193]. Within 12 hours we could detect GFP-ECFCs recruited to both hypoxic and non-hypoxic hydrogels. O2 measurements in vivo confirmed

180 differences between the two conditions up to the 12 hour timepoint (Fig. S7.10B). GFP-ECFCs recruited to the non-hypoxic hydrogels appeared more singular with some adopting a spindle-like shape (Fig. 7.5Bi) and some remaining rounded (Fig. 7.5Bii). In contrast, GFP-ECFCs recruited to the hypoxic hydrogels where more often grouped in clusters with some already sprouting (Fig.

7.5Biii-iv). Mouse lectin (GS-IB4) was used to identify host cells recruited to the hydrogels.

Lectin-positive cells did not cluster as often in non-hypoxic hydrogels (Fig. 7.5Bi-ii) as in hypoxic conditions (Fig. 7.5Biii-iv). Importantly, the number of GFP cells in both conditions were similar, suggesting differences in morphology arise from changes in O2 (Fig. 7.5C). The area covered by clusters was slightly higher and a statistically significant increase in the mean size of clusters was observed in the hypoxic hydrogels compared to non-hypoxic hydrogels (Fig. 7.5D, E).

Inhibition of ROS with DPI resulted in complete blocking of cluster formation in hypoxic conditions (Fig. 7.2C). As such, we co-encapsulated 20 μM DPI within hypoxic hydrogels, then injected the gels into the mouse flank. DPI-treated hydrogels were compared to hypoxic (ctl) hydrogels (Fig. 7.5F). GFP-ECFCs were again injected via intra-cardiac injection (Fig. 7.5F) and tracked 12 hours post-injection. GFP-ECFCs were recruited to both DPI-treated and hypoxic (ctl) hydrogels (Fig. 7.5G). In both conditions, cells exhibited both single cell (Fig. 7.5Gii-iii) and cluster morphology (Fig. 7.5Gi, iv). However, clusters were more prevalent in hypoxic (ctl) hydrogels than in DPI-treated hydrogels. Interestingly, DPI appeared to have a multi-potent effect, both reducing the total number of GFP+ cells recruited (Fig. 7.5H) and reducing the total area covered by clusters (Fig. 7.5I). While the percent area covered by clusters was decreased by addition of DPI, the mean cluster size was not significantly different between the two groups (Fig.

7.5J). While DPI does not completely block cluster formation in vivo, it does have a significant effect on both cell number and area covered by clusters, confirming the role of ROS in cluster

181 formation, even in a more complex in vivo microenvironment. This result supports the highly biomimetic nature of our in vitro platform as a method to identify regulators of complex biological phenomena.

The existence of cluster-based vasculogenesis has been documented in numerous animal development and disease models [617-619]. While each of these animal model systems were different in a variety of ways, they all shared the important characteristic of a hypoxic microenvironment in which ECs interacted to form neovessels. By establishing a controllable environment to study cluster-based vasculogenesis in vivo, we were able to observe cluster formation at a similar timescale in vivo to that which we observed in vitro, as well as confirm the importance of a key regulator of the mechanism.

182

Figure 7.5: Gel-HI hydrogels facilitate ECFC cluster formation in vivo. (A) Intra-cardiac injection of GFP-ECFCs, followed by injection of SDF-1α loaded non-hypoxic and hypoxic hydrogels into the flank of nu/nu mice. (B) 12 hours post-injection, GFP-ECFCs exhibited single cell spindle (i) or rounded (ii) morphology in non-hypoxic hydrogels. Host-cells (GS-IB4 lectin) were also present as isolated, rounded cells. In hypoxic conditions, GFP-ECFCs were present in

183 clusters (iii, iv), with some clusters containing sprouting cells (iii). Host cells also exhibited cluster morphology in hypoxic conditions (iii, iv). (C) Quantification revealed a similar number of cells in both conditions and (D) a slight increase in % area covered by clusters in hypoxic compared to non-hypoxic conditions. (E) Cluster were larger in hypoxic compared to non-hypoxic conditions.

(F) Intra-cardiac injection of GFP-ECFCs followed by injection of SDF-1α loaded hypoxic (ctl) and hypoxic (DPI) hydrogels into the flank of nu/nu mice. (G) 12 hours post-encapsulation, GFP-

ECFCs were present as both single cells (ii, iii) and clusters (i, iv) in both conditions. (H) An increased number of GFP+ cells were present in the control group, and (I) the % area covered by clusters was increased in the control vs DPI-treated group. (J) Mean cluster sizes between the two groups were not statistically significantly different. n=6 nu/nu mice per experiment. Graphical data in C, D, H, and I are reported as box and whisker plots from minimum to maximum. Graphical data in E and J are reported as mean ± SD, with all points denoted by dots. *p<0.05.

7.4 Defining a new mechanism for cluster-based vasculogenesis

The bulk of the literature describing post-natal vasculogenesis has focused on the study of single-cell vasculogenesis. In this mechanism, cells initiate interactions with their surrounding

ECM, then proceed to form vacuoles and lumen through cytoskeletal rearrangements in response to an assortment of angiogenic growth factors. Cell-mediated local matrix degradation then permits vascular branching and sprouting to form vascular networks [97]. This mechanism has informed countless studies ranging from those in vascular biology to the clinic and should never be discounted or neglected. However, an understudied, parallel mechanism appears to contribute to vasculogenesis in hypoxic microenvironments, which has been documented in numerous animal development and disease models [617-619]. An understanding of this mechanism is critical to expand upon our understanding of vascular morphogenesis, as hypoxia is a key regulator in most

184 niches that induce vascular growth. To re-establish vasculature and aid tissue regeneration, EPCs are recruited along O2 and chemokine gradients to regions in need of neovascularization [337].

These EPCs leave the pre-existing vasculature to occupy the hypoxic niche (Fig. 7.6A). From here, we have defined a mechanism by which EPCs (ECFCs) form vasculature in response to their surrounding hypoxic microenvironment. EPCs rapidly respond to the new hypoxic microenvironment and produce ROS that upregulates proteases, which degrade the surrounding

ECM to facilitate EPC cluster formation (Fig. 7.6B). As clusters continue to degrade the matrix, the number of cells per cluster increases, while clusters are stabilized through ITG-β2, ICAM-1, and VE-cad (Fig. 7.6C). Finally, EPCs in clusters engage with the surrounding stiffer ECM to sprout and form vascular networks (Fig. 7.6D). Overall, using an oxygen-controllable 3D environment we were able to observe cluster formation at a similar timescales both in vitro and in vivo, allowing us to recapitulate and understand the hypoxic conditions that activate the signaling pathway of cluster-based vasculogenesis.

185

Figure 7.6: A mechanism for cluster-based vasculogenesis. (A) EPCs (ECFCs) are recruited to injured or diseased tissues in need of new vasculature. These cells then migrate to hypoxic regions.

(B) Hypoxia induces production of ROS, which leads to upregulation of MMP-1 and other proteases, followed by matrix degradation, resulting in cluster formation. (C) Clusters are stabilized by cell-cell interactions (VE-cad, ICAM-1, ITG-β2). (D) Cells interact with the surrounding ECM to form vascular networks. Increases in matrix viscoelasticity result in accelerated and enhanced vascular networks.

186

7.5 Supplementary information

Figure S7.1: Hydrogel height controls O2 gradients. (A) Taller hydrogels establish hypoxic conditions within the hydrogel, while shorter hydrogels maintain non-hypoxic conditions. (B) O2 readings at the bottom of the hydrogel reveal hypoxic conditions until completion of the crosslinking reactions, at which point O2 begins to return to atmospheric conditions. Non-hypoxic hydrogels experience a reduction in O2, but not to hypoxic levels (<5% O2). (C) Non-hypoxic hydrogels (~1.5 mm height) and hypoxic hydrogels (~3 mm height). (D) Representative gradient measurements with commercially available needle-type O2 sensors. Closed markers r represent measurements within the hydrogel. Open markers represent measurements within media.

187

Figure S7.2: Cluster size over time. Quantification of brightfield timelapse imaging shows that clusters form after ~6 hrs hours in culture at 250 μm above the bottom of the hydrogel, corresponding to ~1% O2 (red box). Up to 48 hours, clusters expand in terms of number of cells in clusters and migrate to the bottom of the gel. Graphs are presented as number of single cells (gray bars) + number of cells in clusters (black bars). The total height of each bar is the sum of the two cell populations. Representative quantification from n=3 images per z-location and time.

188

Figure S7.3: Low O2 tension, rather than diffusional limitations or nutrient deprivation, facilitates cluster formation. (A) Hydrogels with the same volume (50 μl) as used for non- hypoxic conditions were subjected to a 24 hr 1% O2 flush in a custom hypoxia chamber. (i) Cluster formation is observed when cells are exposed to these conditions and, indeed (ii) O2 readings match those observed for hypoxic hydrogels (Cii). (B) ECFCs in non-hypoxic control hydrogels do not form clusters, while (C) ECFCs in hypoxic control hydrogels form clusters. Bii and Cii are reproduced from Figure 2 here for ease of comparison. (D) Quantification of % area covered by clusters confirms cluster formation in non-hypoxic (NH) geometries when exposed to 1% O2 is similar to cluster formation in hypoxic (H) conditions. NC=no clusters.

189

Figure S7.4: Cluster formation is HIF-independent and ROS-mediated. (A) Addition of hypoxia mimetic molecules (DMOG and CoCl2) do not facilitate cluster formation in non-hypoxic hydrogels. DMOG limits sprouting events and cells remained rounded up to 48 hrs. CoCl2 permits sprouting in non-hypoxic hydrogels, but does not enhance sprouting. Clusters form, as expected, in hypoxic hydrogels. Arrows indicate sprouting. (B) Addition of DPI prevents sprouting in non- hypoxic hydrogels and blocks cluster formation in hypoxic hydrogels. In control hypoxic hydrogels, large clusters form by 24 hrs and sprouts (arrows) are observed in non-hypoxic hydrogels.

190

Figure S7.5: Experimental setup for timelapse monitoring of increases in fluorescence upon proteolytic degradation of DQ-gelatin. (A) Schematic for DQ-gelatin and microsphere encapsulation. Following encapsulation, MMP was added to degrade the matrix. Addition of MMP resulted in an increase in relative fluorescence. (B-D) Microspheres cluster as DQ fluorescence increases in the presence of MMP, as observed through timelapse imaging with merged brightfield and 488 (for DQ) images. Scale bars 200 μm (C, D). (E, F) Timelapse merged images (brightfield and 488 nm) of ECFCs encapsulated in (E) non-hypoxic and (F) hypoxic hydrogels showing clusters begin to form in the hypoxic hydrogels at approximately the 300 μm z-plane. As the fluorescence increases in the presence of proteases, the hydrogel degrades and clusters migrate towards lower z-planes. Scale bars 100 µm.

191

Figure S7.6: Protease activity and inhibition. (A) For the proteome profile assay, media was collected after 24 hours of culture from hypoxic and non-hypoxic hydrogels to examine secreted soluble proteases. Results are presented as the mean pixel density of the average of two arrays.

Only MMP-1 is significantly increased in hypoxic compared with non-hypoxic conditions.

Cathepsin A, Cathepsin D, Cathepsin X/Z/P, and MMP-10 were significantly increased in non-

192 hypoxic vs hypoxic conditions. All graphical data are reported as mean ± SD. $ P<0.01, &

P<0.001, # P<0.0001. (B) Media supplemented with 1 mM GM6001 inhibited cluster formation in hypoxic hydrogels. (C) DPI inhibits cluster formation, but does not completely abrogate protease activity. Brightfield images and DQ quantification showing that addition of DPI blocks cluster formation in hypoxic conditions. H (ctl) n=6, H (DPI) n=6. (D) Addition of exogenous

MMP facilitates dose dependent cluster formation in non-hypoxic hydrogels. As the concentration increases, clusters form more rapidly, with 10 μg/ml MMP most closely mimicking the behavior observed in hypoxic hydrogels. Hydrogels completely degraded by 24 hours with 100 μg/ml and by 48 hrs with 10 μg/ml. At 1 μg/ml, MMP contributes to enhanced sprouting (arrows) verses non- hypoxic controls, which do not appreciably sprout.

193

Figure S7.7: Vascular sprouting from clusters. (A) Non-hypoxic hydrogels exhibit classical

‘single-cell’ based vasculogenesis, while hypoxic hydrogels yield larger networks through cluster- based vasculogenesis. (B, C) Vascular sprouting is dependent on cell-ECM interactions. (B) At 24 hours, ECFCs exhibit membrane or diffuse cytosolic F-actin (phalloidin) expression. ECFCs begin to show filopodia extending into the surrounding matrix (arrowheads). Scale bars 20 μm. (C) At

72 hours, ECFCs establish cytoskeletal F-actin fibers indicative of cell-ECM interactions (arrows).

Scale bars 20 μm.

194

Figure S7.8: Dynamic matrix stiffening accelerates vascular network formation. (A)

Rheological characterization of temporal addition of mTG. Gel-HI hydrogels were generated and maintained in cell medium for 24 hours. Addition of mTG (1.2 U/ml) resulted in an increase in

G’. (B) Vascular network formation is accelerated with addition of mTG. Day 2 images are reproduced from Fig. 4 here. (C-E) Vascular tubes remain similar in terms of length, area, and thickness. Ctl n=9 gels. mTG n=18 gels. B-D reported as mean ± SD.

195

Figure S7.9: Increased matrix viscoelasticity influences vascular network formation. (A, B)

Vascular networks form in control and mTG treated conditions, all of which contain lumen (A-

Bii, iii). Comparison of F-actin (phalloidin) reveals cytoskeletal changes result from an increase in matrix stiffness (A-Biv, v).

196

Figure S7.10: In vivo O2 measurements. (A) Injections of hypoxic and non-hypoxic hydrogels.

Acellular, in vitro O2 measurements of hydrogels. With the same geometry, O2 can be controlled by addition of CaO2 (non-hypoxic) or Ca(OH)2 (hypoxic). Results presented as minimum O2 after

~1 hr. (B) After 12 hours, in vivo O2 sensing reveals lower O2 in hypoxic hydrogels than in non- hypoxic hydrogels. Graphical data are reported as mean ± SD of measurements from 3 hydrogels per condition (A) and mean ± SD of measurements from 3 hydrogels per group (3 mice). **

P<0.01.

197

CHAPTER 8

8 Layered O2-Controllable Hydrogels Facilitate Study of Uniform Cell Behavior to Identify a Detailed Mechanistic Understanding of Cluster-Based Vasculogenesis

8.1 Teaser

Discretized O2 gradients enable study of uniform cell behavior in 3D hypoxic microenvironments to expand our understanding of the mechanisms driving cluster-based vasculogenesis.

8.2 Introduction

Uncovering the details describing the mechanisms of blood vessel development and regeneration is critical for the advancement of therapeutics to promote neovessel formation to treat diseases of cardiovascular origin, including coronary and peripheral artery disease. Additionally, identifying inhibitors of blood vessel formation is crucial as a treatment modality for preventing tumor angiogenesis and metastasis. To understand how to both promote and inhibit the formation of new blood vessels requires an intricate understanding of each step guiding endothelial cells to build new vascular networks.

Angiogenesis, or the formation of new blood vessels by sprouting from existing vasculature, was long considered the primary mechanism for postnatal neovascularization.

However, the discovery of circulating endothelial progenitor cells (EPCs) shifted this perceived dogma to include a secondary mechanism for postnatal vascular regeneration, termed vasculogenesis, wherein EPCs form new vasculature de novo [638]. Using 3D hydrogel culture platforms and model organisms, such as zebrafish, a detailed mechanism for single-cell based postnatal vasculogenesis has been described [97, 98]. In brief, endothelial cells (ECs) attach to their surrounding extracellular matrix (ECM) through integrin binding. In response to an array of

198 proangiogenic cues, these cells then rearrange their cytoskeleton to form intracellular vacuoles.

These vacuoles then fuse to form lumen, and vascular networks expand through matrix degradation-mediated cell sprouting concurrent with ECM deposition. Ultimately, new vascular networks are stabilized by recruitment of support cells, such as pericytes, and localization of membrane proteins, such as vascular endothelial-cadherin (VE-cad), to the cell-cell junctions. This mechanism, single-cell based vasculogenesis, has been shown to contribute to embryological, as well as adult, blood vessel formation. However, in addition to this mechanism, a secondary mechanism, termed cluster-based vasculogenesis, has been observed in vivo in hypoxic microenvironments, including during wound healing and tumors [618, 619]. Additionally, we have recently used a hydrogel platform to recapitulate cluster-based vasculogenesis in vitro, and have identified several key regulators of the mechanism [195].

We designed gelatin-based O2-controllable hydrogels to mimic the hypoxic vascular regenerative microenvironment [192-194]. Importantly, these hydrogels contain a biomimetic O2 gradient and the capacity to dynamically tune matrix stiffness, both of which serve as key regulators of vasculogenesis. Our previous work described rapid production of reactive oxygen species (ROS) upon EPC exposure to hypoxic gradients. Hypoxia and ROS led to upregulation of proteases that degraded the surrounding matrix to facilitate cluster formation. Clusters were then stabilized by cell-cell interactions via membrane localization of VE-cad, intercellular adhesion molecule 1 (ICAM-1), and integrin subunit β2 (ITG-β2). EPCs sprouted from clusters via cell- matrix interactions to form expansive vascular networks, containing both single cell and multicellular lumen, and sprouting in three dimensions. Enhancements in vascular network formation were achieved through dynamic matrix stiffening to accelerate vascular network formation [195].

With the capacity to recapitulate this complex biological process in vitro, the aim of the current

199 work was to uncover additional details specifically related to the formation and stabilization of

EPC clusters. Based on our previous work, we anticipate that the mechanisms governing sprouting from clusters will closely match mechanisms defined for vascular sprouting from EC spheroids

[97] or EC-coated beads [160]. Here we develop a method to study uniform cell behavior in 3D hypoxic microenvironments, thus enabling in-depth analysis of the molecular mechanisms guiding cluster formation. Using RNA-sequencing, we confirmed our findings for ROS and protease mediated cluster formation, and identified additional pathway enrichment related to cell survival, inhibition of apoptosis, cell cycle upregulation, and increased carbohydrate metabolism.

8.3 Results and Discussion

8.3.1 Development and characterization of layered hydrogels

As a proof of concept, we encapsulated ECFCs within three distinct layers in O2- controllable hydrogels (Fig. 8.1A). O2 measurements were recorded using needle-type O2 sensors.

In the bottom layer, cells were exposed to severe hypoxia (0-0.2% O2), in the middle layer, cells were exposed to moderate hypoxia (2.2-6.6% O2), and in the top layer, cells were exposed to nonhypoxic conditions (7.5-14.4% O2) (Fig. 8.1A). In each layer, a distinct cell morphology was observed. In the bottom (severely hypoxic) layer, ECFC clusters formed, with no sprouting (Fig.

8.1Bi). In the middle (moderately hypoxic) layer, ECFC clusters formed, with extensive sprouting from clusters (Fig. 8.1Bii), and in the top (nonhypoxic) layer, vascular sprouting resembling single-cell vasculogenesis was observed (Fig. 8.1Biii).

200

Figure 8.1: Three-layered hydrogel design and characterization. (A) Cells were encapsulated in bottom (B; severely hypoxic), middle (M; moderately hypoxic), and top (T; nonhypoxic) layers.

(B) Distinct cell morphology was observed in each layer. (i) Cells in the T layer exhibit single cell vasculogenesis; (ii) cells in the M layer exhibit cluster formation and vascular sprouting from clusters; and (iii) cells in the B layer exhibit cluster formation. All scale bars 100 μm.

8.3.2 Analysis of uniform cell behavior in layered hydrogels

In our previous work, we studied cell behavior in hypoxic gradient hydrogels, and compared the behavior of those cells to cells in nonhypoxic hydrogels [195]. While this analysis allowed for accurate recapitulation of cluster-based vasculogenesis in hypoxic conditions, nonuniform cell behavior resulted from the presence of the O2 gradient in hypoxic conditions, rendering conventional molecular biology techniques (e.g. qPCR, Western blot) difficult to use given the necessity to collect RNA and/or protein from all the cells within the hydrogel to perform these analyses. To combat this limitation, we generated a simplified version of our layered hydrogel platform (Fig. 8.2A). Here, we discretized the O2 gradient to culture cells in a 3D hypoxic

201 environment, but with uniform cell behavior. We compared this two-layer hypoxic hydrogel system, where we cultured cells in the bottom layer, with our conventional hypoxic hydrogels and observed uniform cell seeding throughout the z-plane at an early time point, 40 mins after encapsulation, in both conditions (Fig. 8.2A). After 24 hr in culture, ECFC clusters were observed in both conditions (Fig. 8.2B). However, in the two-layer hypoxic condition, nearly all cells participated in cluster formation (at the lowest z-plane, z1), with no cells in focus in the higher z- planes (z2 and z3). In the conventional hypoxic hydrogel, clusters appeared at the lowest z-plane

(z1), no cells were in focus at the intermediate z-plane (z2), and cells appeared in isolation at the highest z-plane (z3), revealing heterogeneity in cell behavior (Fig. 8.2B).

We also confirmed the kinetics of cluster formation in the two-layer hypoxic hydrogels matched cluster formation kinetics in conventional hypoxic hydrogels. In both conditions, cells were seeded uniformly in a single cell morphology, then clusters formed by 10 hrs in culture, and ultimately expanded in terms of number of cells in clusters by 24 hrs in culture (Fig. 8.2C).

Accordingly, measurements of O2 at the bottom of the hydrogel matched along the entirety of the culture period, with rapid exposure to hypoxia (<5% O2) in both conditions, and maintenance of hypoxia throughout the 24 hr culture period (Fig. 8.2D).

202

Figure 8.2: Two-layer hypoxic hydrogels enable uniform ECFC cluster behavior, and match conventional hypoxic hydrogel cluster formation kinetics and O2. (A) ECFCs were encapsulated in the bottom layer of two-layer hypoxic hydrogels and in conventional hypoxic hydrogels. 40 mins after encapsulation, uniform cell seeding was confirmed in both conditions, as measured by images taken at three distinct z-planes [z1 (lowest) → z3 (highest)]. (B) After 24 hr

203 in culture, clusters formed in both hydrogels. In the two-layer hypoxic hydrogels, uniform cell behavior was observed, with all cells in focus at z1, and no cells in focus at z2 or z3. In the conventional hypoxic hydrogel, clusters were in focus at z1, with no cells in focus at z2, but cells in single cell morphology at z3, indicating non-uniform cell behavior. (C) Cluster formation kinetics were matched in the two hydrogel platforms, with clusters forming at 10 hr after encapsulation in both conditions, and clusters growing in terms of number of cells in clusters by

24 hr. Scale bars are 100 μm. (D) O2 measurements show rapid exposure to hypoxia (<5% O2, grayed area) in both conditions and exposure to hypoxia over the entire 24 hr culture period.

8.3.3 Confirmation of a conserved mechanism for cluster formation between two-layer and conventional hypoxic hydrogels

Rapid upregulation of ROS facilitates cluster formation in conventional hypoxic hydrogels.

To confirm the role of ROS and oxidative stress in the two-layer hypoxic hydrogels, we co- encapsulated CellROX, which fluoresces upon oxidation with ROS, to identify oxidative stress in live cells. While cells cultured in two-layer hypoxic hydrogels and nonhypoxic hydrogels exhibit

CellROX+ fluorescence (Fig. 8.3A), there was a statistically significant increase in fluorescence at early timepoints in two-layer hypoxic conditions compared to nonhypoxic conditions (Fig.

8.3B). We hypothesize that early exposure to hypoxia results in rapid oxidative stress to facilitate cluster formation.

204

Figure 8.3: Oxidative stress is upregulated at early timepoints in ECFCs cultured in two- layer hypoxic hydrogels compared to nonhypoxic hydrogels. (A) ECFCs in both conditions were CellROX+ over the time course of the experiment. (B) Quantification reveals significant upregulation of ROS, and associated oxidative stress, at early time points in two-layer hypoxic conditions compared to nonhypoxic conditions. n=3 independent experiments per condition.

Graphical data are reported as mean ± SD. *p<0.05.

8.3.4 Global view of RNA sequencing of two-layer hypoxic and nonhypoxic hydrogels

Following confirmation of uniform cell behavior and accurate mimicry of cluster formation kinetics and O2, as well as a conserved guiding mechanism for cluster formation between two-

205 layer and conventional hypoxic hydrogels, we performed RNA sequencing on cells encapsulated in two-layer hypoxic hydrogels (herein referred to as hypoxic hydrogels) and cells encapsulated in nonhypoxic hydrogels, which do not exhibit cluster formation (Fig. 8.4A). Measurements of O2 at the bottom of the gels revealed rapid exposure to hypoxic levels of O2 that were maintained over the culture period (Fig. 8.4B). Interestingly, cells in nonhypoxic hydrogels experienced delayed and transient exposure to hypoxic conditions, indicating the importance of rapid exposure to hypoxia in cluster formation (Fig. 8.4B). At pre-determined time points, 40 mins after encapsulation, 10 hr after encapsulation, and 24 hr after encapsulation, we collected and purified

RNA for RNA sequencing. Principle component analysis (PCA) confirmed clustering of experimental conditions at each time point, verifying differential gene expression between the two conditions at all 3 time points (Fig. 8.4C). A global view of the sequencing data clearly identifies thousands of statistically significant differentially expressed genes between cells exposed to hypoxic or nonhypoxic conditions at each time point (Fig. 8.4D).

206

Figure 8.4: Differential gene expression between ECFCs encapsulated in hypoxic and nonhypoxic hydrogels. (A) ECFCs encapsulated in hypoxic hydrogels exhibited cluster formation, while cells in nonhypoxic hydrogels remained isolated as single cells. (B) O2 measurements at the bottom of hypoxic hydrogels indicated rapid and sustained exposure to hypoxia. O2 measurements at the bottom of nonhypoxic hydrogels indicated delayed and transient exposure to hypoxia. (C) PCA identifies 3 PCs contribute to 67.1% of the variance in the data.

Biological replicates for each condition (hypoxic vs nonhypoxic) cluster together at each time points. Outliers, as identified by statistical analysis, were omitted from subsequent analysis. n=3-

4 for each condition at each time point. (D) Statistically significant differentially expressed genes in hypoxic vs nonhypoxic conditions at each time point, as identified by 1-way ANOVA.

207

8.3.5 Genes associated with oxidative stress are differentially expressed over the experimental time course

86 genes associated with oxidative stress were analyzed, and 27 genes were statistically significantly differentially expressed (>ǀ2σǀ) at one or more time point (Fig. 8.5A). Of those genes, six were highly differentially expressed at one or more time point (>ǀ3σǀ) (Fig. 8.5B). Two of these genes, UCP2 (-6σ) and SOD2 (-3σ), were downregulated at 40 mins and 10 hr, respectively. All other highly differentially expressed genes, TXNRD2, BNIP3, DHCR24, and HSPA1A, were upregulated (>3σ) at the 10 hr or 24 hr time points. Of particular interest to cluster formation are those genes that are significantly upregulated at the early (40 mins) time point. GPX7, CAT, and

MGST3 were all significantly upregulated at the early time point (Fig. 8.5C). Interestingly, these three genes act to regulate the response and control of ROS in an antioxidant role [639]. Even small changes in the concentration of ROS can lead to vastly different cell behaviors [628], ranging from cell death to enhancing angiogenesis [184]. As such, we hypothesize that these genes are upregulated to protect against the early exposure of ECFCs in hypoxic hydrogels to oxidative stress to facilitate downstream signaling leading to cluster formation.

208

Figure 8.5: Oxidative stress-associated differential gene expression. (A) Heat map for all statistically significant (>ǀ2σǀ) genes associated with oxidative stress. (B) Genes with the highest fold change (>ǀ3σǀ) at one or more time points. (C) Genes with differential expression (>ǀ2σǀ) at the early (40 mins) time point.

8.3.6 A vast array of genes encoding for proteases are differentially expressed over the experimental time course, necessitating broad protease inhibition to block cluster formation

Our previous work established the importance of matrix degradation in facilitating cluster formation. To deepen our understanding of the role of proteases, we investigated the gene expression of all proteases and observed differential expression of >200 genes encoding proteases

(Fig. 8.6A). We then plotted only those genes with >3σ differential expression, and saw upregulation of proteases with a range of functions, and saw upregulation of various proteases at all time points, suggesting a temporal and multifactorial effect of proteases on cluster formation

209

(Fig. 8.6B-D). Notably, upregulated genes include ASAH1, which can facilitate cell survival and rescue from oxidative stress, as well as act as an activator of another protease, CTSB, which can degrade the surrounding ECM [640-642]. In addition, the upregulated proteases themselves can degrade the surrounding ECM, including CTSH [643]. PLAT and PLAU can both activate other proteases, as well as degrade the surrounding ECM [644]. While not significantly upregulated between hypoxic and nonhypoxic conditions at the RNA level, MMP1, MMP2, MMP10, MMP14, and MMP16 were present and may serve an important role in degradation-mediated cluster formation. Taken together, these results present an array of proteases that may contribute to cluster formation.

In our previous work, we analyzed an array of soluble proteases and observed many proteases were present in both hypoxic and nonhypoxic conditions, but we only saw significant upregulation of MMP-1 in hypoxic conditions. Even with a broad spectrum MMP inhibitor

(GM6001), we did not see complete inhibition of cluster formation. This result is not surprising in light of the vast array of genes encoding for proteases that were upregulated in hypoxic conditions at the RNA level. To confirm our hypothesis that numerous genes are involved in matrix degradation that leads to cluster formation, we knocked down MMP1 with siRNA, and observed no inhibition of cluster formation compared to the scrambled control (Fig. 8.6E). We further tested inhibition of CTSL, which has been shown to play a crucial role in EPC-mediated matrix degradation [379]. Concentrations of CTSL inhibitor (Z-FF-FMK) ranging from 10-500 μM did not inhibit cluster formation (Fig. 8.6F). Finally, we utilized a protease and phosphatase inhibitor cocktail (PI) to broadly inhibit proteases, including serine and cysteine proteases, and saw partial inhibition of cluster formation (Fig. 8.6G). We saw similar results with metalloprotease inhibitor

EDTA (Fig. 8.6G), but saw marked reduction in cluster formation with a combination inhibitory

210 cocktail containing both PI and EDTA (Fig. 8.6G). Single factor protease inhibition is not sufficient to inhibit cluster formation, but rather requires broad protease inhibition. These results provide insights into the potent nature of hypoxia as an inductive cue regulating cluster formation through upregulation of numerous proteases involved as either primary matrix degrading proteases, such as CTSH, or secondary players, such as ASAH1, which activate other proteases with the capacity to degrade the surrounding ECM.

211

212

Figure 8.6: Numerous proteases are upregulated in hypoxic conditions, thereby requiring broad inhibition of proteases to block cluster formation. (A) >200 proteases are statistically significantly differentially expressed (>ǀ2σǀ) in hypoxic vs nonhypoxic conditions. (B-D) Cysteine,

Metallo, and Serine proteases are highly upregulated (>3σ) over the experimental time course. (E) siRNA knockdown of MMP1 did not inhibit cluster formation. (F) A range of concentrations of

CTSL inhibitor Z-FF-FMK did not inhibit cluster formation. (G) PI and EDTA both partially inhibit cluster formation when used in isolation, but result in substantial reduction in cluster formation when used as a combination inhibitor (P+E).

8.3.7 Cell-cell interactions stabilize ECFC clusters

Localization of membrane proteins to the cell-cell junction stabilizes ECFC clusters through ICAM1, ITGβ2, and VE-cad. Interestingly, none of the genes associated with these proteins were upregulated in our RNA-seq analysis. Rather, it is likely that the localization of these proteins shifted to allow for cell-cell interactions. Other genes associated with cell-cell binding were upregulated, such as VCAM1 and SELPLG (Fig. 8.7).

Figure 8.7: Significantly differentially expressed genes associated with binding of ECs.

213

8.3.8 Cell survival, apoptosis, and cell cycle progression are differentially regulated concomitant with cluster formation

Ingenuity Pathway Analysis (IPA) is a powerful tool designed to gain an understanding of how the relationships between differentially expressed genes may regulate a variety of biological functions, or how they play a role in canonical pathways. For our purposes, we utilized IPA to identify and analyze new pathways that contribute to the formation of ECFC clusters.

At the 40 min time point, genes associated with cell survival were downregulated in hypoxic conditions, but genes associated with this pathway were then upregulated at the 10 and 24 hr time points (Fig. 8.8A, B). We predict this switch is due to the initial microenvironment causing upregulation in cell stress, but then cluster formation facilitating upregulation of cell survival signaling to overcome the high stress microenvironment. A subset of genes was selected based on differential gene expression and relevance to cluster formation, including VCAM1, SMAD6,

HSPA6, and EGLN3 (Fig. 8.8C). VCAM1 is endothelial cell specific, and is associated with cell- cell adhesion, activation of MMPs, and acts as a regulator of the proangiogenic effects of oxidative stress [628]. SMAD6, as well as SMAD7, protect cells from apoptosis and can be altered in response to oxidative stress [645]. HSPA6, and more broadly the family of heat shock proteins, alter their gene expression in response to cell stress. HSPA6 can specifically alter expression based on oxidative stress [646]. Finally, EGLN3 plays an important role in mediating cell survival, in particular when cells are exposed to hypoxic conditions [647].

Analyzing a related pathway, apoptosis, we saw a nearly opposite trend (Fig. 8.8D). At the early time point, genes associated with apoptosis are upregulated, then as clusters form, apoptosis is downregulated. The 24 hr time point is slightly puzzling, as cell survival is increased, but some apoptosis is required for vascular lumen formation, so apoptosis may be a requirement for cluster-

214 based vascular network lumen formation. This is an interesting line of inquiry outside the scope of the work presented here.

Genes associated with cell cycle progression follow a similar trend to those associated with cell survival, with many genes upregulated at the 10 and 24 hr time points (Fig. 8.8E, F).

Additionally, a variety of biological functions are activated to contribute to nearly all stages of the cell cycle at both the 10 and 24 hr time points (Fig. 8.8G).

215

Figure 8.8: Cell survival, apoptosis, and cell cycle progression are differentially regulated as

ECFC clusters form. (A) There is a statistically significant overlap in the genes from our dataset and the set of genes regulating cell survival. Bars are the -log (p-value), with the number of relevant genes listed above each bar. The activation Z-score predicts that cell survival is downregulated at the 40 min time point and upregulated at 10 and 24 hr. (B) Significantly differentially expressed genes associated with cell survival over the time course of the experiment (≥ǀ6σǀ at one or more

216 time point). (C) Selected genes of interest associated with cell survival. *ǀ3σǀ, **ǀ6σǀ, *** ≥ǀ6σǀ.

(D) Statistically significant overlap between genes in our dataset and those associated with apoptosis. This set of genes is initially upregulated (40 mins), then downregulated (10 hr), then upregulated again (24 hr). (E) Statistically significant overlap between genes in our dataset and those associated with cell cycle progression. This set of genes is initially downregulated (40 mins), then upregulated at 10 and 24 hr. (F) Significantly differentially expressed genes associated with cell cycle progression over the time course of the experiment (≥ǀ6σǀ at one or more time point).

(G) Many biological functions associated with cell cycle progression are significantly upregulated at the 10 and 24 hr time points.

8.3.9 Carbohydrate metabolism is upregulated in hypoxic conditions

ECs are known to be highly glycolytic. The role of metabolism in angiogenesis has become a topic of widespread interest, pioneered by Carmeliet’s group, where EC metabolism has been established as a driver rather than a bystander in angiogenesis [648, 649]. Analysis of biological functions associated with carbohydrate metabolism revealed enrichment and activation at both the

10 and 24 hr time points (Fig. 8.9A). Assessment of the specific genes associated with the

‘metabolism of polysaccharide’ function revealed numerous differentially expressed genes (Fig.

8.9B), with several genes exhibiting highly differential expression (6σ), including APLN, DKK1,

NR4A1, STBD1, PPP1R3C, PLAT, and PPP1R3G (Fig. 8.9C). An inspection of the differentially expressed genes related to glycolysis (Fig. 8.9D) revealed several key genes regulating glycolysis were highly differentially expressed, including ENO2, HK2, PFKFB4, and SLC2A1 (Fig. 8.9E).

HK2 represents a potential target gene because of its significant differential expression, as well as its upstream role in a rate-limiting step in glucose metabolism [650].

217

218

Figure 8.9: Carbohydrate metabolism is upregulated in hypoxic conditions. (A) Many biological functions associated with carbohydrate metabolism are significantly upregulated at the

10 and 24 hr time points. (B) Significantly differentially expressed genes associated with metabolism of polysaccharide over the time course of the experiment (≥ǀ3σǀ at one or more time point). (C) Highly differentially expressed genes (ǀ6σǀ at one or more time point). (D) Significantly differentially expressed genes associated with glycolysis over the time course of the experiment

(≥ǀ3σǀ at one or more time point). (E) Highly differentially expressed genes (ǀ6σǀ at one or more time point). *ǀ3σǀ, **ǀ6σǀ

8.4 Conclusions

Utilizing layered hydrogels with discretized O2 gradients, we were able to design a hydrogel platform that facilitated study of uniform ECFC cluster formation in a 3D platform. This novel design accurately recapitulated the cluster formation mechanism we have previously defined, and allowed for in-depth analysis of the molecular regulators of cluster formation using

RNA sequencing. We have outlined a detailed mechanism for cluster formation in Fig. 8.10. Using our layered hydrogels, we confirmed the role of early upregulation of oxidative stress and reactive oxygen species. RNA-seq analysis revealed an early response to oxidative stress through upregulation of antioxidants GPX7, CAT, and MGST3, which act to protect cells exposed to this stress-inducing environment. We then confirmed upregulation of numerous proteases, which we hypothesize have direct effects on matrix degradation, and indirect effects on matrix degradation through coordinated activation of other matrix-degrading proteases to facilitate cluster formation.

Using broad spectrum protease inhibitors, but not single factor inhibitors, we successfully blocked cluster formation. IPA identified new pathways and potential target genes involved in cluster

219 formation, including cell survival, anti-apoptosis, and cell cycle progression. Finally, upregulation of carbohydrate metabolism and cell-cell adhesion were observed at later time points. It is unknown whether alterations to carbohydrate metabolism influence cluster formation, but we predict they will impact vascular sprouting from clusters at time points outside the scope of the current work. Localization of membrane proteins, ICAM1, ITGβ2, and VE-cad, to the cell-cell junction was previously shown to stabilize ECFC clusters. Here, we show increases in genes associated with cell-cell adhesion, most notably VCAM1 and SELPLG.

Moving forward, we will confirm the relevancy of the identified pathways through PCR and Western blot, as well as siRNA knockdown or pharmacological inhibition. Following confirmation of the impact of the identified pathways on inhibition of cluster formation, in vivo analysis using our previously developed murine hydrogel plug model, or a murine model of a hypoxic tumor will be used to assess the effects of our identified inhibitors of cluster formation in vivo.

220

Figure 8.10: Proposed mechanism for cluster formation. Hypoxia leads to upregulation of

ROS, which can be inhibited by DPI to block cluster formation. In response to ROS, genes encoding antioxidants are upregulated to protect cells from oxidative stress. Proteases are then upregulated to degrade the matrix and clusters form. Increases in cell survival and cell cycle progression, as well as decreases in apoptosis are associated with cluster formation. Upregulation of carbohydrate metabolism and cell-cell interactions occur at later time points.

221

CHAPTER 9

9 Conclusions and Future Directions

9.1 Conclusions

Diseases with a vascular component are among the leading causes of death worldwide. A detailed understanding of how these clinical pathologies arise requires sophisticated experimental platforms to recapitulate the complex biological processes of angiogenesis and vasculogenesis.

Insights gained from model organisms have informed our understanding of neovascularization, and coupled with in vitro studies of human vascular development and regeneration have established an impressive wealth of knowledge towards a more complete understanding of how blood vessels form. These two areas of research have converged to incorporate human cells in highly biomimetic in vitro platforms to study their behavior with a previously unprecedented level of biomimicry.

In Chapter 1, we described vascular development from a biological perspective and presented methods for current and future reconstruction of the vascular developmental niche by employing engineered top-down approaches to facilitate vascular network self-assembly. Chapter

2 focused on the design of hydrogels for therapeutic delivery of biological factors to promote vascular regeneration. The necessity for spatiotemporal, multifactorial delivery highlighted by limited clinical success of single factors therapies, drove our work towards identifying and modulating upstream regulators of the local vascular niche. To this end, in Chapter 3 we described the role of two such upstream and potent regulators of vascular development and regeneration, hypoxia and properties of the ECM, such as stiffness. In Chapters 5 and 6, we described the design and implementation of gelatin and dextran-based O2-controllable hydrogels with the capacity to independently control O2 and matrix stiffness. These hydrogels have a broad reaching utility, as

222 hypoxia and matrix stiffness influence numerous signaling pathways in nearly all cell types.

Specific examples were provided in studying vascular network formation, as well as tumor and cancer cell behavior in hypoxic conditions within O2-controllable hydrogels.

Gelatin-based O2-controllable hydrogels were used to study cluster-based vasculogenesis in Chapters 7 and 8. Prior to our work, cluster-based vasculogenesis was an observed, but understudied mechanism for revascularization of hypoxic microenvironments in vivo. In Chapter

7, we effectively recapitulated EPC cluster-based vasculogenesis and defined several key regulators of the mechanism. EPCs (ECFCs in our study), were subjected to a biomimetic O2 gradient, and in response, upregulated production of ROS and proteases to degrade the ECM to promote the formation of EPC clusters. Cell clusters were stabilized by cell-cell interactions through junctional localization of proteins VE-cad, ICAM-1, and ITG-β2. EPCs then sprouted from clusters into the surrounding matrix via cell-matrix interactions to form expansive vascular networks. Network formation was accelerated and enhanced by dynamic matrix stiffening, which coaxed an increase in cell-matrix interactions. We then confirmed the kinetics of cluster formation in a murine model, and verified the relevance of ROS as a target for inhibition of cluster formation.

We refined our hydrogel platform in Chapter 8 to enable in-depth analysis of uniform cell behavior in a 3D hypoxic culture system. RNA-seq confirmed the importance of oxidative stress, protease production, and cell-cell binding in cluster formation and stabilization, and identified additional key players in the process, including upregulation of cell survival and cell cycle progression, accompanied by inhibition of apoptosis. Upregulation of carbohydrate metabolism was also observed, which may play an important role in subsequent vascular sprouting from clusters. Our O2-controllable hydrogel platform is a powerful research tool to recapitulate

223 biological processes and identify key regulators and therapeutic targets for cluster-based vasculogenesis.

9.2 Future Directions

9.2.1 Acellular therapeutics

Acellular hydrogel therapeutics represent a highly translational approach for tissue regeneration. In particular, our lab has had success utilizing dextran-based hydrogels for treatment of burn wounds, even in the absence of cellular or drug/protein payloads [551, 651]. Hydrogels are an ideal platform to treat wounds, as they keep the wound bed hydrated, allow for transport of biomolecules, and their properties can be easily manipulated and tuned.

Several papers have identified an optimal stiffness, ~300 Pa, that provides physical support to the wound, but also allows cellular remodeling and infiltration [551, 652]. We have developed dextran-based, O2-controllable hydrogels, with which we can tune the stiffness to the optimal 300

Pa, as well as control O2 in the wound bed. O2 can alter infiltrating cell phenotype, in particular to tune immune cell polarization to a regenerative rather than fibrotic phenotype, as well as alter the composition of newly deposited ECM to further coax the healing response to regeneration [653].

Deferoxamine, which can facilitate accumulation of HIF1α, can accelerate wound healing in mice, further supporting our hypothesis of hypoxia enhanced wound healing [654]. Additionally, there is evidence of hypoxic signaling in blastema formation, further supporting its role in promoting regeneration [653]. By specifically controlling O2 in a transplantable biomaterial, we hypothesize that we will enhance wound healing towards full skin regeneration.

224

9.2.2 Vascular biology and vascular tissue engineering

We have identified a number of regulators guiding cluster-based vasculogenesis, but need to confirm their relevancy and potency moving forward. We have shown broad spectrum inhibition is required to block cluster formation as dependent on downstream signaling, such as protease production, so it is likely multifactorial inhibition will be required. Alternatively, identifying the most upstream regulators of the process may prove the most beneficial. We anticipate these upstream regulators will be associated with production of or response to oxidative stress induced by hypoxia.

Following confirmation of an optimal therapeutic target, testing its effect in vivo will be critical towards its potential translational impact. Two candidate animal models for such work are a vascularized hypoxic tumor in a mouse or a developing zebrafish exposed to hypoxic conditions.

In our previous in vivo work, we observed similar kinetics of cluster formation compared to our in vitro work, with some effect of cluster inhibition using a ROS inhibitor, DPI. It is likely that the host immune cells, which may contribute to matrix degradation and other hypoxia-induced responses, somewhat mitigated the striking differences we saw in our in vitro model of hypoxic vs nonhypoxic conditions, in which we only studied EC behavior.

We have been able to closely match the regenerative microenvironment, in terms of hypoxia and matrix mechanics. However, addition of other cell types, such as immune cells that also may be recruited to hypoxic tissues in vivo, may improve the relevancy of our in vitro model.

Immune cells play an important role in post-natal vasculogenesis, so studying their effect on cluster-based vasculogenesis will likely improve our ability to accurately identify therapeutic targets.

225

Further down the pipeline, we can apply our knowledge of cluster-based vasculogenesis to tissue engineering applications. Cluster-based vasculogenesis results in expansive vascular networks in the absence of hyperphysiological growth factor concentrations after just three days in culture. Manipulating two potent cues will likely minimize the dependence of tissue function on highly tissue-specific media formulations, enabling co-culture of multiple cells types within one engineered platform. The presence of an O2 gradient may also be interesting to harness to study the behavior of a variety of tissue types, as cells of different tissues respond differently to hypoxia.

Finally, 3D stem cell differentiation protocols and organoid development are becoming popular and impactful subsets of research at the intersection of the fields of biomaterials and stem cell biology. Because embryological development occurs under hypoxic conditions, it would be fascinating to understand the role of O2 gradients and hypoxia on tissue development in vitro.

226

References

1. Armulik, A., G. Genove, and C. Betsholtz, Pericytes: Developmental, Physiological, and Pathological Perspectives, Problems, and Promises. Developmental Cell, 2011. 21(2): p. 193- 215. 2. Wanjare, M., S. Kusuma, and S. Gerecht, Perivascular cells in blood vessel regeneration. Biotechnology Journal, 2013. 8(4): p. 434-+. 3. Wang, G., et al., Origin and differentiation of vascular smooth muscle cells. Journal of Physiology-London, 2015. 593(14): p. 3013-3030. 4. Song, H.H.G., et al., Vascular Tissue Engineering: Progress, Challenges, and Clinical Promise. Cell Stem Cell, 2018. 22(3): p. 340-354. 5. Smith, Q., M. Blatchley, and S. Gerecht, Chapter 30 - Engineering Niches for Blood Vessel Regeneration, in Biology and Engineering of Stem Cell Niches, A. Vishwakarma and J.M. Karp, Editors. 2017, Academic Press: Boston. p. 479-497. 6. Smith, Q. and S. Gerecht, Going with the flow: microfluidic platforms in vascular tissue engineering. Current Opinion in Chemical Engineering, 2014. 3: p. 42-50. 7. Rouwkema, J. and A. Khademhosseini, Vascularization and Angiogenesis in Tissue Engineering: Beyond Creating Static Networks. Trends in Biotechnology, 2016. 34(9): p. 733-745. 8. Strilic, B., T. Kucera, and E. Lammert, Formation of cardiovascular tubes in invertebrates and vertebrates. Cellular and Molecular Life Sciences, 2010. 67(19): p. 3209-3218. 9. Monahan-Earley, R., A.M. Dvorak, and W.C. Aird, Evolutionary origins of the blood vascular system and endothelium. Journal of Thrombosis and Haemostasis, 2013. 11: p. 46-66. 10. Lammert, E. and J. Axnick, Vascular Lumen Formation. Cold Spring Harbor Perspectives in Medicine, 2012. 2(4): p. 9. 11. Isogai, S., M. Horiguchi, and B.M. Weinstein, The vascular anatomy of the developing zebrafish: An atlas of embryonic and early larval development. Developmental Biology, 2001. 230(2): p. 278-301. 12. Risau, W., Mechanisms of angiogenesis. Nature, 1997. 386(6626): p. 671-674. 13. Weinstein, B.M., Plumbing the mysteries of vascular development using the zebrafish. Seminars in Cell & Developmental Biology, 2002. 13(6): p. 515-522. 14. Blum, Y., et al., Complex cell rearrangements during intersegmental vessel sprouting and vessel fusion in the zebrafish embryo. Developmental Biology, 2008. 316(2): p. 312-322. 15. Ginis, I., et al., Differences between human and mouse embryonic stem cells. Developmental Biology, 2004. 269(2): p. 360-380. 16. Kelly, M.A. and K.K. Hirschi, Signaling Hierarchy Regulating Human Endothelial Cell Development. Arteriosclerosis Thrombosis and Vascular Biology, 2009. 29(5): p. 718-U209. 17. Red-Horse, K., et al., Endothelium-microenvironment interactions in the developing embryo and in the adult. Developmental Cell, 2007. 12(2): p. 181-194. 18. Herbert, S.P. and D.Y.R. Stainier, Molecular control of endothelial cell behaviour during blood vessel morphogenesis. Nature Reviews Molecular Cell Biology, 2011. 12(9): p. 551-564. 19. Coultas, L., K. Chawengsaksophak, and J. Rossant, Endothelial cells and VEGF in vascular development. Nature, 2005. 438(7070): p. 937-945. 20. Marcelo, K.L., L.C. Goldie, and K.K. Hirschi, Regulation of Endothelial Cell Differentiation and Specification. Circulation Research, 2013. 112(9): p. 1272-1287. 21. Carmeliet, P. and R.K. Jain, Molecular mechanisms and clinical applications of angiogenesis. Nature, 2011. 473(7347): p. 298-307. 22. Risau, W. and I. Flamme, Vasculogenesis. Annual Review of Cell and Developmental Biology, 1995. 11: p. 73-91. 23. Tam, P.P.L. and R.R. Behringer, Mouse gastrulation: the formation of a mammalian body plan. Mechanisms of Development, 1997. 68(1-2): p. 3-25.

227

24. Cox, C.M. and T.J. Poole, Angioblast differentiation is influenced by the local environment: FGF-2 induces angioblasts and patterns vessel formation in the quail embryo. Developmental Dynamics, 2000. 218(2): p. 371-382. 25. Winnier, G., et al., BONE MORPHOGENETIC PROTEIN-4 IS REQUIRED FOR MESODERM FORMATION AND PATTERNING IN THE MOUSE. Genes & Development, 1995. 9(17): p. 2105-2116. 26. Asashima, M., et al., MESODERMAL INDUCTION IN EARLY AMPHIBIAN EMBRYOS BY ACTIVIN-A (ERYTHROID-DIFFERENTIATION FACTOR). Rouxs Archives of Developmental Biology, 1990. 198(6): p. 330-335. 27. Lindsley, R.C., et al., Canonical Wnt signaling is required for development of embryonic stem cell-derived mesoderm. Development, 2006. 133(19): p. 3787-3796. 28. Moore, S.W., R.E. Keller, and M.A.R. Koehl, THE DORSAL INVOLUTING MARGINAL ZONE STIFFENS ANISOTROPICALLY DURING ITS CONVERGENT EXTENSION IN THE GASTRULA OF XENOPUS-LAEVIS. Development, 1995. 121(10): p. 3131-3140. 29. Solursh, M. and G.M. Morriss, Glycosaminoglycan synthesis in rat embryos during the formation of the primary mesenchyme and neural folds. Developmental Biology, 1977. 57(1): p. 75-86. 30. Ramirez-Bergeron, D.L., et al., Hypoxia affects mesoderm and enhances hemangioblast specification during early development. Development, 2004. 131(18): p. 4623-4634. 31. Przybyla, L., J.N. Lakins, and V.M. Weaver, Tissue Mechanics Orchestrate Wnt-Dependent Human Embryonic Stem Cell Differentiation. Cell Stem Cell, 2016. 19(4): p. 462-475. 32. Smith, Q., et al., Compliant substratum guides endothelial commitment from human pluripotent stem cells. Science Advances, 2017. 3(5): p. 9. 33. Hirschi, K.K., Hemogenic endothelium during development and beyond. Blood, 2012. 119(21): p. 4823. 34. Vogeli, K.M., et al., A common progenitor for haematopoietic and endothelial lineages in the zebrafish gastrula. Nature, 2006. 443(7109): p. 337-339. 35. Risau, W., DIFFERENTIATION OF ENDOTHELIUM. Faseb Journal, 1995. 9(10): p. 926-933. 36. Zovein, A.C., et al., Fate Tracing Reveals the Endothelial Origin of Hematopoietic Stem Cells. Cell Stem Cell, 2008. 3(6): p. 625-636. 37. Eilken, H.M., S.-I. Nishikawa, and T. Schroeder, Continuous single-cell imaging of blood generation from haemogenic endothelium. Nature, 2009. 457: p. 896. 38. Lancrin, C., et al., The haemangioblast generates haematopoietic cells through a haemogenic endothelium stage. Nature, 2009. 457: p. 892. 39. Bertrand, J.Y., et al., Haematopoietic stem cells derive directly from aortic endothelium during development. Nature, 2010. 464: p. 108. 40. Boisset, J.-C., et al., In vivo imaging of haematopoietic cells emerging from the mouse aortic endothelium. Nature, 2010. 464: p. 116. 41. Kissa, K. and P. Herbomel, Blood stem cells emerge from aortic endothelium by a novel type of cell transition. Nature, 2010. 464: p. 112. 42. Chen, M.J., et al., Runx1 is required for the endothelial to haematopoietic cell transition but not thereafter. Nature, 2009. 457: p. 887. 43. Sabin, F.R., Studies on the origin of blood-vessels and of red blood-corpuscles as seen in the living blastoderm of chicks during the second day of incubation. Contributions to Embryology, 1920. 9(27/46): p. 215-U56. 44. Goldie, L.C., M.K. Nix, and K.K. Hirschi, Embryonic vasculogenesis and hematopoietic specification. Organogenesis, 2008. 4(4): p. 257-263. 45. Damert, A., et al., Insufficient VEGFA activity in yolk sac endoderm compromises haematopoietic and endothelial differentiation. Development, 2002. 129(8): p. 1881-1892. 46. Vokes, S.A., et al., Hedgehog signaling is essential for endothelial tube formation during vasculogenesis. Development, 2004. 131(17): p. 4371-4380.

228

47. Shalaby, F., et al., Failure of blood-island formation and vasculogenesis in Flk-1-deficient mice. Nature, 1995. 376(6535): p. 62-66. 48. Schuh, A.C., et al., In vitro hematopoietic and endothelial potential of flk-1(-/-) embryonic stem cells and embryos. Proceedings of the National Academy of Sciences of the United States of America, 1999. 96(5): p. 2159-2164. 49. Bohnsack, B.L., et al., Signaling hierarchy downstream of retinoic acid that independently regulates vascular remodeling and endothelial cell proliferation. Genes & Development, 2004. 18(11): p. 1345-1358. 50. Fong, G.H., et al., ROLE OF THE FLT-1 RECEPTOR TYROSINE KINASE IN REGULATING THE ASSEMBLY OF VASCULAR ENDOTHELIUM. Nature, 1995. 376(6535): p. 66-70. 51. Salanga, M.C., et al., ETS family protein ETV2 is required for initiation of the endothelial lineage but not the hematopoietic lineage in the Xenopus embryo. Developmental dynamics : an official publication of the American Association of Anatomists, 2010. 239(4): p. 1178-1187. 52. Wareing, S., et al., The Flk1-Cre-Mediated Deletion of ETV2 Defines Its Narrow Temporal Requirement During Embryonic Hematopoietic Development. Stem Cells, 2012. 30(7): p. 1521- 1531. 53. Herzog, Y., N. Guttmann-Raviv, and G. Neufeld, Segregation of arterial and venous markers in subpopulations of blood islands before vessel formation. Developmental Dynamics, 2005. 232(4): p. 1047-1055. 54. Zhong, T.P., et al., Gridlock signalling pathway fashions the first embryonic artery. Nature, 2001. 414(6860): p. 216-220. 55. Wang, H.U., Z.F. Chen, and D.J. Anderson, Molecular distinction and angiogenic interaction between embryonic arteries and veins revealed by ephrin-B2 and its receptor Eph-B4. Cell, 1998. 93(5): p. 741-753. 56. Rufaihah, A.J., et al., Human induced pluripotent stem cell-derived endothelial cells exhibit functional heterogeneity. American journal of translational research, 2013. 5(1): p. 21-35. 57. Zhang, J., et al., Functional characterization of human pluripotent stem cell-derived arterial endothelial cells. Proceedings of the National Academy of Sciences, 2017. 114(30): p. E6072. 58. Moyon, D., et al., Plasticity of endothelial cells during arterial-venous differentiation in the avian embryo. Development, 2001. 128(17): p. 3359-3370. 59. Chong, D.C., et al., Stepwise arteriovenous fate acquisition during mammalian vasculogenesis. Developmental Dynamics, 2011. 240(9): p. 2153-2165. 60. le Noble, F., et al., Flow regulates arterial-venous differentiation in the chick embryo yolk sac. Development, 2004. 131(2): p. 361-375. 61. Obi, S., et al., Fluid shear stress induces arterial differentiation of endothelial progenitor cells. J Appl Physiol (1985), 2009. 106(1): p. 203-11. 62. Masumura, T., et al., Shear Stress Increases Expression of the Arterial Endothelial Marker EphrinB2 in Murine ES Cells via the VEGF-Notch Signaling Pathways. Arteriosclerosis, Thrombosis, and Vascular Biology, 2009. 29(12): p. 2125-2131. 63. Sivarapatna, A., et al., Arterial specification of endothelial cells derived from human induced pluripotent stem cells in a biomimetic flow bioreactor. Biomaterials, 2015. 53: p. 621-633. 64. Mitsumata, M., et al., Fluid shear stress stimulates platelet-derived growth factor expression in endothelial cells. Am J Physiol, 1993. 265(1 Pt 2): p. H3-8. 65. Xue, C., et al., Substrate stiffness regulates arterial-venous differentiation of endothelial progenitor cells via the Ras/Mek pathway. Biochimica et Biophysica Acta (BBA) - Molecular Cell Research, 2017. 1864(10): p. 1799-1808. 66. Kusuma, S., et al., Low Oxygen Tension Enhances Endothelial Fate of Human Pluripotent Stem Cells. Arteriosclerosis Thrombosis and Vascular Biology, 2014. 34(4): p. 913-920. 67. Niklason, L. and G. Dai, Arterial Venous Differentiation for Vascular Bioengineering. Annual Review of Biomedical Engineering, 2018. 20(1): p. 431-447.

229

68. Fish, J.E. and J.D. Wythe, The molecular regulation of arteriovenous specification and maintenance. Developmental Dynamics, 2015. 244(3): p. 391-409. 69. Swift Matthew, R. and M. Weinstein Brant, Arterial–Venous Specification During Development. Circulation Research, 2009. 104(5): p. 576-588. 70. Herron, L.A., C.S. Hansen, and H.E. Abaci, Engineering tissue-specific blood vessels. Bioengineering & Translational Medicine, 2019. 0(0): p. e10139. 71. Gomez-Salinero, J.M. and S. Rafii, Endothelial cell adaptation in regeneration. Science, 2018. 362(6419): p. 1116. 72. Rafii, S., J.M. Butler, and B.-S. Ding, Angiocrine functions of organ-specific endothelial cells. Nature, 2016. 529: p. 316. 73. Nolan, Daniel J., et al., Molecular Signatures of Tissue-Specific Microvascular Endothelial Cell Heterogeneity in Organ Maintenance and Regeneration. Developmental Cell, 2013. 26(2): p. 204-219. 74. Gaengel, K., et al., Endothelial-Mural Cell Signaling in Vascular Development and Angiogenesis. Arteriosclerosis Thrombosis and Vascular Biology, 2009. 29(5): p. 630-638. 75. Jakobsson, L., et al., Endothelial cells dynamically compete for the tip cell position during angiogenic sprouting. Nature Cell Biology, 2010. 12(10): p. 943-953. 76. Samakovlis, C., et al., Development of the Drosophila tracheal system occurs by a series of morphologically distinct but genetically coupled branching events. Development, 1996. 122(5): p. 1395-1407. 77. Affolter, M. and E. Caussinus, Tracheal branching morphogenesis in Drosophila: new insights into cell behaviour and organ architecture. Development, 2008. 135(12): p. 2055-2064. 78. Hayashi, S. and T. Kondo, Development and Function of the Drosophila Tracheal System. Genetics, 2018. 209(2): p. 367-380. 79. Ruhle, H., Larval spiracle systems of drosophila melanogaster particles and its variability. Zeitschrift Fur Wissenschaftliche Zoologie, 1932. 141(2): p. 159-245. 80. Wigglesworth, V.B., GROWTH AND REGENERATION IN THE TRACHEAL SYSTEM OF AN INSECT, RHODNIUS-PROLIXUS (HEMIPTERA). Quarterly Journal of Microscopical Science, 1954. 95(1): p. 115-&. 81. Sutherland, D., C. Samakovlis, and M.A. Krasnow, Branchless encodes a Drosophila FGF homolog that controls tracheal cell migration and the pattern of branching. Cell, 1996. 87(6): p. 1091-1101. 82. Gerhardt, H., et al., VEGF guides angiogenic sprouting utilizing endothelial tip cell filopodia. Journal of Cell Biology, 2003. 161(6): p. 1163-1177. 83. Simon, M.C. and B. Keith, The role of oxygen availability in embryonic development and stem cell function. Nature Reviews Molecular Cell Biology, 2008. 9(4): p. 285-296. 84. Egeblad, M., M.G. Rasch, and V.M. Weaver, Dynamic interplay between the collagen scaffold and tumor evolution. Current Opinion in Cell Biology, 2010. 22(5): p. 697-706. 85. Ismat, A., A.M. Cheshire, and D.J. Andrew, The secreted AdamTS-A metalloprotease is required for collective cell migration. Development, 2013. 140(9): p. 1981-1993. 86. Wang, Q., M. Uhlirova, and D. Bohmann, Spatial Restriction of FGF Signaling by a Matrix Metalloprotease Controls Branching Morphogenesis. Developmental Cell, 2010. 18(1): p. 157- 164. 87. Bonnans, C., J. Chou, and Z. Werb, Remodelling the extracellular matrix in development and disease. Nature Reviews Molecular Cell Biology, 2014. 15(12): p. 786-801. 88. Avraamides, C.J., B. Garmy-Susini, and J.A. Varner, Integrins in angiogenesis and lymphangiogenesis. Nature Reviews Cancer, 2008. 8(8): p. 604-617. 89. Rupp Paul, A. and D. Little Charles, Integrins in Vascular Development. Circulation Research, 2001. 89(7): p. 566-572. 90. Friedlander, M., et al., Definition of Two Angiogenic Pathways by Distinct α_v Integrins. Science, 1995. 270(5241): p. 1500-1502.

230

91. Li, S.R., et al., Hydrogels with precisely controlled integrin activation dictate vascular patterning and permeability. Nature Materials, 2017. 16(9): p. 953-+. 92. Bronnerfraser, M., DISTRIBUTION AND FUNCTION OF TENASCIN DURING CRANIAL NEURAL CREST DEVELOPMENT IN THE CHICK. Journal of Neuroscience Research, 1988. 21(2-4): p. 135-147. 93. Palecek, S.P., et al., Integrin-ligand binding properties govern cell migration speed through cell- substratum adhesiveness. Nature, 1997. 385(6616): p. 537-540. 94. Lubarsky, B. and M.A. Krasnow, Tube morphogenesis: Making and shaping biological tubes. Cell, 2003. 112(1): p. 19-28. 95. Downs, K.M., Florence Sabin and the mechanism of blood vessel lumenization during vasculogenesis. Microcirculation, 2003. 10(1): p. 5-25. 96. Folkman, J. and C. Haudenschild, Angiogenesis in vitro. Nature, 1980. 288: p. 551. 97. Davis, G.E., et al., MOLECULAR BASIS FOR ENDOTHELIAL LUMEN FORMATION AND TUBULOGENESIS DURING VASCULOGENESIS AND ANGIOGENIC SPROUTING, in International Review of Cell and Molecular Biology, Vol 288, K.W. Jeon, Editor. 2011, Elsevier Academic Press Inc: San Diego. p. 101-165. 98. Kamei, M., et al., Endothelial tubes assemble from intracellular vacuoles in vivo. Nature, 2006. 442(7101): p. 453-456. 99. Berry, K.L., et al., A C-elegans CLIC-like protein required for intracellular tube formation and maintenance. Science, 2003. 302(5653): p. 2134-2137. 100. Zovein, A.C., et al., beta 1 Integrin Establishes Endothelial Cell Polarity and Arteriolar Lumen Formation via a Par3-Dependent Mechanism. Developmental Cell, 2010. 18(1): p. 39-51. 101. Xu, K., et al., Blood Vessel Tubulogenesis Requires Rasip1 Regulation of GTPase Signaling. Developmental Cell, 2011. 20(4): p. 526-539. 102. Strilic, B., et al., Electrostatic Cell-Surface Repulsion Initiates Lumen Formation in Developing Blood Vessels. Current Biology, 2010. 20(22): p. 2003-2009. 103. Strilic, B., et al., The Molecular Basis of Vascular Lumen Formation in the Developing Mouse Aorta. Developmental Cell, 2009. 17(4): p. 505-515. 104. Thomson, J.A., et al., Embryonic stem cell lines derived from human blastocysts. Science, 1998. 282(5391): p. 1145-1147. 105. Takahashi, K., et al., Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell, 2007. 131(5): p. 861-872. 106. Zhang, J., et al., Functional characterization of human pluripotent stem cell-derived arterial endothelial cells. Proceedings of the National Academy of Sciences of the United States of America, 2017. 114(30): p. E6072-E6078. 107. Katt, M.E., et al., Human Brain Microvascular Endothelial Cells Derived from the BC1 iPS Cell Line Exhibit a Blood-Brain Barrier Phenotype. Plos One, 2016. 11(4): p. 18. 108. Qian, T.C., et al., Directed differentiation of human pluripotent stem cells to blood-brain barrier endothelial cells. Science Advances, 2017. 3(11): p. 12. 109. Zhao, J.G.J. and N.A. Afshari, Generation of Human Corneal Endothelial Cells via In Vitro Ocular Lineage Restriction of Pluripotent Stem Cells. Investigative Ophthalmology & Visual Science, 2016. 57(15): p. 6878-6884. 110. Songstad, A.E., et al., Connective Tissue Growth Factor Promotes Efficient Generation of Human Induced Pluripotent Stem Cell-Derived Choroidal Endothelium. Stem Cells Translational Medicine, 2017. 6(6): p. 1533-1546. 111. Palpant, N.J., et al., Generating high-purity cardiac and endothelial derivatives from patterned mesoderm using human pluripotent stem cells. Nature Protocols, 2017. 12(1): p. 17. 112. Uenishi, G.I., et al., NOTCH signaling specifies arterial-type definitive hemogenic endothelium from human pluripotent stem cells. Nature Communications, 2018. 9: p. 14.

231

113. Kusuma, S., et al., Self-organized vascular networks from human pluripotent stem cells in a synthetic matrix. Proceedings of the National Academy of Sciences of the United States of America, 2013. 110(31): p. 12601-12606. 114. Orlova, V.V., et al., Generation, expansion and functional analysis of endothelial cells and pericytes derived from human pluripotent stem cells. Nature Protocols, 2014. 9(6): p. 1514-1531. 115. Orlova, V.V., et al., Functionality of Endothelial Cells and Pericytes From Human Pluripotent Stem Cells Demonstrated in Cultured Vascular Plexus and Zebrafish Xenografts. Arteriosclerosis Thrombosis and Vascular Biology, 2014. 34(1): p. 177-+. 116. Gu, M.X., et al., Patient-Specific iPSC-Derived Endothelial Cells Uncover Pathways that Protect against Pulmonary Hypertension in BMPR2 Mutation Carriers. Cell Stem Cell, 2017. 20(4): p. 490-+. 117. Chan, X.Y., et al., Three-Dimensional Vascular Network Assembly From Diabetic Patient- Derived Induced Pluripotent Stem Cells. Arteriosclerosis Thrombosis and Vascular Biology, 2015. 35(12): p. 2677-2685. 118. Hamauchi, S., et al., Cellular Functions and Gene and Protein Expression Profiles in Endothelial Cells Derived from Moyamoya Disease-Specific iPS Cells. Plos One, 2016. 11(9): p. 14. 119. Lian, X.J., et al., Efficient Differentiation of Human Pluripotent Stem Cells to Endothelial Progenitors via Small-Molecule Activation of WNT Signaling. Stem Cell Reports, 2014. 3(5): p. 804-816. 120. Gu, M., Efficient Differentiation of Human Pluripotent Stem Cells to Endothelial Cells. Current Protocols in Human Genetics, 2018. 98(1): p. e64. 121. Palpant, N.J., et al., Inhibition of beta-catenin signaling respecifies anterior-like endothelium into beating human cardiomyocytes. Development, 2015. 142(18): p. 3198-U212. 122. James, D., et al., Expansion and maintenance of human embryonic stem cell-derived endothelial cells by TGF beta inhibition is Id1 dependent. Nature Biotechnology, 2010. 28(2): p. 161-U15. 123. Rufaihah, A.J., et al., Human induced pluripotent stem cell-derived endothelial cells exhibit functional heterogeneity. American Journal of Translational Research, 2013. 5(1): p. 21-U122. 124. Paik, D.T., et al., Large-Scale Single-Cell RNA-Seq Reveals Molecular Signatures of Heterogeneous Populations of Human Induced Pluripotent Stem Cell-Derived Endothelial Cells. Circulation Research, 2018. 123(4): p. 443-450. 125. Rungarunlert, S., et al., Embryoid body formation from embryonic and induced pluripotent stem cells: Benefits of bioreactors. World journal of stem cells, 2009. 1(1): p. 11-21. 126. Vittet, D., et al., Embryonic stem cells differentiate in vitro to endothelial cells through successive maturation steps. Blood, 1996. 88(9): p. 3424-3431. 127. Kurosawa, H., Methods for inducing embryoid body formation: In vitro differentiation system of embryonic stem cells. Journal of Bioscience and Bioengineering, 2007. 103(5): p. 389-398. 128. Doetschman, T.C., et al., The <em>in vitro</em> development of blastocyst-derived embryonic stem cell lines: formation of visceral yolk sac, blood islands and myocardium. Journal of Embryology and Experimental Morphology, 1985. 87(1): p. 27. 129. Risau, W., et al., VASCULOGENESIS AND ANGIOGENESIS IN EMBRYONIC-STEM-CELL- DERIVED EMBRYOID BODIES. Development, 1988. 102(3): p. 471-478. 130. Levenberg, S., et al., Endothelial cells derived from human embryonic stem cells. Proceedings of the National Academy of Sciences of the United States of America, 2002. 99(7): p. 4391-4396. 131. Levenberg, S., et al., Differentiation of human embryonic stem cells on three-dimensional polymer scaffolds. Proceedings of the National Academy of Sciences of the United States of America, 2003. 100(22): p. 12741-12746. 132. Levenberg, S., et al., Isolation, differentiation and characterization of vascular cells derived from human embryonic stem cells. Nature Protocols, 2010. 5(6): p. 1115-1126. 133. Nourse, M.B., et al., VEGF Induces Differentiation of Functional Endothelium From Human Embryonic Stem Cells Implications for Tissue Engineering. Arteriosclerosis Thrombosis and Vascular Biology, 2010. 30(1): p. 80-U189.

232

134. Chin, M.H., et al., Induced Pluripotent Stem Cells and Embryonic Stem Cells Are Distinguished by Gene Expression Signatures. Cell Stem Cell, 2009. 5(1): p. 111-123. 135. Wilson, K.D., et al., MicroRNA Profiling of Human-Induced Pluripotent Stem Cells. Stem Cells and Development, 2009. 18(5): p. 749-757. 136. Guenther, M.G., et al., Chromatin structure and gene expression programs of human embryonic and induced pluripotent stem cells. Cell stem cell, 2010. 7(2): p. 249-257. 137. Lister, R., et al., Hotspots of aberrant epigenomic reprogramming in human induced pluripotent stem cells. Nature, 2011. 471: p. 68. 138. Ghosh, Z., et al., Persistent Donor Cell Gene Expression among Human Induced Pluripotent Stem Cells Contributes to Differences with Human Embryonic Stem Cells. PLOS ONE, 2010. 5(2): p. e8975. 139. Narsinh, K.H., et al., Single cell transcriptional profiling reveals heterogeneity of human induced pluripotent stem cells. The Journal of Clinical Investigation, 2011. 121(3): p. 1217-1221. 140. Feng, Q., et al., Hemangioblastic Derivatives from Human Induced Pluripotent Stem Cells Exhibit Limited Expansion and Early Senescence. STEM CELLS, 2010. 28(4): p. 704-712. 141. Rozario, T. and D.W. DeSimone, The extracellular matrix in development and morphogenesis: A dynamic view. Developmental Biology, 2010. 341(1): p. 126-140. 142. Lu, P.F., et al., Extracellular Matrix Degradation and Remodeling in Development and Disease. Cold Spring Harbor Perspectives in Biology, 2011. 3(12): p. 24. 143. Schenke-Layland, K., et al., Collagen IV Induces Trophoectoderm Differentiation of Mouse Embryonic Stem Cells. STEM CELLS, 2007. 25(6): p. 1529-1538. 144. Dye, J.F., et al., Distinct patterns of microvascular endothelial cell morphology are determined by extracellular matrix composition. Endothelium-Journal of Endothelial Cell Research, 2004. 11(3-4): p. 151-167. 145. Hou, L.Q., et al., Combinatorial Extracellular Matrix Microenvironments for Probing Endothelial Differentiation of Human Pluripotent Stem Cells. Scientific Reports, 2017. 7: p. 12. 146. Beachley, V.Z., et al., Tissue matrix arrays for high-throughput screening and systems analysis of cell function. Nature Methods, 2015. 12(12): p. 1197-+. 147. Flaim, C.J., S. Chien, and S.N. Bhatia, An extracellular matrix microarray for probing cellular differentiation. Nature Methods, 2005. 2(2): p. 119-125. 148. Mei, Y., et al., Combinatorial development of biomaterials for clonal growth of human pluripotent stem cells. Nature Materials, 2010. 9(9): p. 768-778. 149. Smith, Q., et al., Cytoskeletal tension regulates mesodermal spatial organization and subsequent vascular fate. Proceedings of the National Academy of Sciences of the United States of America, 2018. 115(32): p. 8167-8172. 150. McBeath, R., et al., Cell shape, cytoskeletal tension, and RhoA regulate stem cell lineage commitment. Developmental Cell, 2004. 6(4): p. 483-495. 151. Engler, A.J., et al., Matrix elasticity directs stem cell lineage specification. Cell, 2006. 126(4): p. 677-689. 152. Hanjaya-Putra, D., et al., Vascular endothelial growth factor and substrate mechanics regulate in vitro tubulogenesis of endothelial progenitor cells. Journal of Cellular and Molecular Medicine, 2010. 14(10): p. 2436-2447. 153. Barcelloshoff, M.H., et al., FUNCTIONAL-DIFFERENTIATION AND ALVEOLAR MORPHOGENESIS OF PRIMARY MAMMARY CULTURES ON RECONSTITUTED BASEMENT-MEMBRANE. Development, 1989. 105(2): p. 223-&. 154. Gerecht, S., et al., Hyaluronic acid hydrogel for controlled self-renewal and differentiation of human embryonic stem cells. Proceedings of the National Academy of Sciences, 2007. 104(27): p. 11298. 155. Zhang, S., et al., The influence of a spatiotemporal 3D environment on endothelial cell differentiation of human induced pluripotent stem. Biomaterials, 2014. 35(12): p. 3786-3793.

233

156. Zhang, F.Z., et al., Optimizing mesoderm progenitor selection and three-dimensional microniche culture allows highly efficient endothelial differentiation and ischemic tissue repair from human pluripotent stem cells. Stem Cell Research & Therapy, 2017. 8: p. 15. 157. Zhang, J.H., et al., Extracellular Matrix Promotes Highly Efficient Cardiac Differentiation of Human Pluripotent Stem Cells The Matrix Sandwich Method. Circulation Research, 2012. 111(9): p. 1125-1136. 158. Shao, Y., et al., Self-organized amniogenesis by human pluripotent stem cells in a biomimetic implantation-like niche. Nature Materials, 2017. 16(4): p. 419-+. 159. Bayless, K.J., H.I. Kwak, and S.C. Su, Investigating endothelial invasion and sprouting behavior in three-dimensional collagen matrices. Nature Protocols, 2009. 4(12): p. 1888-1898. 160. Nakatsu, M.N. and C.C.W. Hughes, An optimized three-dimensional in vitro model for the analysis of angiogenesis, in Angiogenesis: In Vitro Systems, D.A. Cheresh, Editor. 2008, Elsevier Academic Press Inc: San Diego. p. 65-+. 161. Shin, J.W. and D.J. Mooney, Extracellular matrix stiffness causes systematic variations in proliferation and chemosensitivity in myeloid leukemias. Proceedings of the National Academy of Sciences of the United States of America, 2016. 113(43): p. 12126-12131. 162. Chaudhuri, O., et al., Hydrogels with tunable stress relaxation regulate stem cell fate and activity. Nature Materials, 2016. 15(3): p. 326-+. 163. Lou, J.Z., et al., Stress relaxing hyaluronic acid-collagen hydrogels promote cell spreading, fiber remodeling, and focal adhesion formation in 3D cell culture. Biomaterials, 2018. 154: p. 213- 222. 164. Vining, K.H., A. Stafford, and D.J. Mooney, Sequential modes of crosslinking tune viscoelasticity of cell-instructive hydrogels. Biomaterials, 2019. 188: p. 187-197. 165. Walters, B.D. and J.P. Stegemann, Strategies for directing the structure and function of three- dimensional collagen biomaterials across length scales. Acta Biomaterialia, 2014. 10(4): p. 1488-1501. 166. Lutolf, M.P., et al., Synthetic matrix metalloproteinase-sensitive hydrogels for the conduction of tissue regeneration: Engineering cell-invasion characteristics. Proceedings of the National Academy of Sciences of the United States of America, 2003. 100(9): p. 5413-5418. 167. Patterson, J. and J.A. Hubbell, SPARC-derived protease substrates to enhance the plasmin sensitivity of molecularly engineered PEG hydrogels. Biomaterials, 2011. 32(5): p. 1301-1310. 168. Purcell, B.P., et al., Injectable and bioresponsive hydrogels for on-demand matrix metalloproteinase inhibition. Nature Materials, 2014. 13(6): p. 653-661. 169. Ferreira, S.A., et al., Bi-directional cell-pericellular matrix interactions direct stem cell fate. Nature Communications, 2018. 9: p. 12. 170. Plow, E.F., et al., Ligand binding to integrins. Journal of Biological Chemistry, 2000. 275(29): p. 21785-21788. 171. Huettner, N., T.R. Dargaville, and A. Forget, Discovering Cell-Adhesion Peptides in Tissue Engineering: Beyond RGD. Trends in Biotechnology, 2018. 36(4): p. 372-383. 172. Darnell, M., et al., Material microenvironmental properties couple to induce distinct transcriptional programs in mammalian stem cells. Proceedings of the National Academy of Sciences of the United States of America, 2018. 115(36): p. E8368-E8377. 173. Glorevski, N., et al., Designer matrices for intestinal stem cell and organoid culture. Nature, 2016. 539(7630): p. 560-+. 174. Cruz-Acuna, R., et al., Synthetic hydrogels for human intestinal organoid generation and colonic wound repair. Nature Cell Biology, 2017. 19(11): p. 1326-+. 175. Okazaki, K.M. and E. Maltepe, Oxygen, epigenetics and stem cell fate. Regenerative Medicine, 2006. 1(1): p. 71-83. 176. Lee, Y.M., et al., Determination of hypoxic region by hypoxia marker in developing mouse embryos in vivo: A possible signal for vessel development. Developmental Dynamics, 2001. 220(2): p. 175-186.

234

177. Abbott, B.D. and A.R. Buckalew, Placental defects in ARNT-knockout conceptus correlate with localized decreases in VEGF-R2, Ang-1, and Tie-2. Developmental Dynamics, 2000. 219(4): p. 526-538. 178. Maltepe, E., et al., Abnormal angiogenesis and responses to glucose and oxygen deprivation in mice lacking the protein ARNT. Nature, 1997. 386(6623): p. 403-407. 179. Duan, L.J., Y.H. Zhang-Benoit, and G.H. Fong, Endothelium-intrinsic requirement for Hif-2 alpha during vascular development. Circulation, 2005. 111(17): p. 2227-2232. 180. Manalo, D.J., et al., Transcriptional regulation of vascular endothelial cell responses to hypoxia by HIF-1. Blood, 2005. 105(2): p. 659-669. 181. Krock, B.L., N. Skuli, and M.C. Simon, Hypoxia-Induced Angiogenesis: Good and Evil. Genes & Cancer, 2011. 2(12): p. 1117-1133. 182. Koh, M.Y. and G. Powis, Passing the baton: the HIF switch. Trends in Biochemical Sciences, 2012. 37(9): p. 364-372. 183. Wouters, B.G. and M. Koritzinsky, Hypoxia signalling through mTOR and the unfolded protein response in cancer. Nature Reviews Cancer, 2008. 8(11): p. 851-864. 184. Kim, Y.W. and T.V. Byzova, Oxidative stress in angiogenesis and vascular disease. Blood, 2014. 123(5): p. 625-631. 185. Perkins, N.D., Integrating cell-signalling pathways with NF-kappa B and IKK function. Nature Reviews Molecular Cell Biology, 2007. 8(1): p. 49-62. 186. Sainson, R.C.A. and A.L. Harris, Hypoxia-regulated differentiation: let"s step it up a Notch. Trends in Molecular Medicine, 2006. 12(4): p. 141-143. 187. Abaci, H.E., et al., Unforeseen decreases in dissolved oxygen levels affect tube formation kinetics in collagen gels. American Journal of Physiology-Cell Physiology, 2011. 301(2): p. C431-C440. 188. Brennan, M.D., et al., Oxygen control with microfluidics. Lab on a Chip, 2014. 14(22): p. 4305- 4318. 189. Oppegard, S.C. and D.T. Eddington, A microfabricated platform for establishing oxygen gradients in 3-D constructs. Biomedical Microdevices, 2013. 15(3): p. 407-414. 190. Chen, Y.A., et al., Generation of oxygen gradients in microfluidic devices for cell culture using spatially confined chemical reactions. Lab on a Chip, 2011. 11(21): p. 3626-3633. 191. Rodenhizer, D., et al., A three-dimensional engineered tumour for spatial snapshot analysis of cell metabolism and phenotype in hypoxic gradients. Nature Materials, 2016. 15(2): p. 227-+. 192. Lewis, D.M., et al., O-2-controllable hydrogels for studying cellular responses to hypoxic gradients in three dimensions in vitro and in vivo. Nature Protocols, 2017. 12(8): p. 1620-1638. 193. Park, K.M. and S. Gerecht, Hypoxia-inducible hydrogels. Nature Communications, 2014. 5: p. 12. 194. Blatchley, M., K.M. Park, and S. Gerecht, Designer hydrogels for precision control of oxygen tension and mechanical properties. Journal of Materials Chemistry B, 2015. 3(40): p. 7939-7949. 195. Blatchley, M.R., et al., Hypoxia and matrix viscoelasticity sequentially regulate endothelial progenitor cluster-based vasculogenesis. Science Advances, 2019. 5(3): p. eaau7518. 196. Lee, Y., et al., Hydrogen Peroxide-Releasing Hydrogels for Enhanced Endothelial Cell Activities and Neovascularization. Acs Applied Materials & Interfaces, 2018. 10(21): p. 18372-18379. 197. Chiu, J.J. and S. Chien, Effects of Disturbed Flow on Vascular Endothelium: Pathophysiological Basis and Clinical Perspectives. Physiological Reviews, 2011. 91(1): p. 327-387. 198. Baeyens, N., et al., Endothelial fluid shear stress sensing in vascular health and disease. The Journal of Clinical Investigation, 2016. 126(3): p. 821-828. 199. Ando, J. and K. Yamamoto, Flow detection and calcium signalling in vascular endothelial cells. Cardiovascular Research, 2013. 99(2): p. 260-268. 200. Smith, Q., et al., Differential HDAC6 Activity Modulates Ciliogenesis and Subsequent Mechanosensing of Endothelial Cells Derived from Pluripotent Stem Cells. Cell Reports, 2018. 24(4): p. 895-+.

235

201. Bogorad, M.I., et al., Review: in vitro microvessel models. Lab on a Chip, 2015. 15(22): p. 4242- 4255. 202. Chen, M.B., et al., On-chip human microvasculature assay for visualization and quantification of tumor cell extravasation dynamics. Nature Protocols, 2017. 12(5): p. 865-880. 203. Turing, A.M., THE CHEMICAL BASIS OF MORPHOGENESIS. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences, 1952. 237(641): p. 37-72. 204. Wolpert, L., Positional information and the spatial pattern of cellular differentiation. Journal of Theoretical Biology, 1969. 25(1): p. 1-47. 205. Green, J.B.A., H.V. New, and J.C. Smith, RESPONSES OF EMBRYONIC XENOPUS CELLS TO ACTIVIN AND FGF ARE SEPARATED BY MULTIPLE DOSE THRESHOLDS AND CORRESPOND TO DISTINCT AXES OF THE MESODERM. Cell, 1992. 71(5): p. 731-739. 206. Ericson, J., et al., Pax6 controls progenitor cell identity and neuronal fate in response to graded shh signaling. Cell, 1997. 90(1): p. 169-180. 207. Wilson, P.A., et al., Concentration-dependent patterning of the Xenopus ectoderm by BMP4 and its signal transducer smad1. Development, 1997. 124(16): p. 3177-3184. 208. Zhang, J., et al., The role of maternal VegT in establishing the primary germ layers in Xenopus embryos. Cell, 1998. 94(4): p. 515-524. 209. Rao, J., et al., Stepwise Clearance of Repressive Roadblocks Drives Cardiac Induction in Human ESCs. Cell Stem Cell, 2016. 18(3): p. 341-353. 210. Nguyen, E.H., M.P. Schwartz, and W.L. Murphy, Biomimetic Approaches to Control Soluble Concentration Gradients in Biomaterials. Macromolecular Bioscience, 2011. 11(4): p. 483-492. 211. Manfrin, A., et al., Engineered signaling centers for the spatially controlled patterning of human pluripotent stem cells. Nature Methods, 2019. 16(7): p. 640-648. 212. Danchakoff, V., Cellpotentialities and differential factors, considered in relation to erythropoiesis. American Journal of Anatomy, 1918. 24(1): p. 1-35. 213. Williams, P.H., et al., Visualizing long-range movement of the morphogen Xnr2 in the Xenopus embryo. Current Biology, 2004. 14(21): p. 1916-1923. 214. Yan, D. and X.H. Lin, Shaping Morphogen Gradients by Proteoglycans. Cold Spring Harbor Perspectives in Biology, 2009. 1(3): p. 16. 215. Jiang, B.H., et al., Hypoxia-inducible factor 1 levels vary exponentially over a physiologically relevant range of O-2 tension. American Journal of Physiology-Cell Physiology, 1996. 271(4): p. C1172-C1180. 216. Raab, M., et al., Crawling from soft to stiff matrix polarizes the cytoskeleton and phosphoregulates myosin-II heavy chain. Journal of Cell Biology, 2012. 199(4): p. 669-683. 217. Hadden, W.J., et al., Stem cell migration and mechanotransduction on linear stiffness gradient hydrogels. Proceedings of the National Academy of Sciences of the United States of America, 2017. 114(22): p. 5647-5652. 218. Carmeliet, P. and R.K. Jain, Angiogenesis in cancer and other diseases. Nature, 2000. 407(6801): p. 249-257. 219. Hanjaya-Putra, D., et al., Controlled activation of morphogenesis to generate a functional human microvasculature in a synthetic matrix. Blood, 2011. 118(3): p. 804-815. 220. Davis, G.E. and D.R. Senger, Endothelial extracellular matrix - Biosynthesis, remodeling, and functions during vascular morphogenesis and neovessel stabilization. Circulation Research, 2005. 97(11): p. 1093-1107. 221. Davis, G.E. and C.W. Camarillo, An alpha 2 beta 1 integrin-dependent pinocytic mechanism involving intracellular vacuole formation and coalescence regulates capillary lumen and tube formation in three-dimensional collagen matrix. Experimental Cell Research, 1996. 224(1): p. 39-51. 222. Bayless, K.J., R. Salazar, and G.E. Davis, RGD-dependent vacuolation and lumen formation observed during endothelial cell morphogenesis in three-dimensional fibrin matrices involves the

236

alpha(v)beta(3) and alpha(5)beta(1) integrins. American Journal of Pathology, 2000. 156(5): p. 1673-1683. 223. Lee, S., et al., Autocrine VEGF signaling is required for vascular homeostasis. Cell, 2007. 130(4): p. 691-703. 224. Phng, L.K. and H. Gerhardt, Angiogenesis: A Team Effort Coordinated by Notch. Developmental Cell, 2009. 16(2): p. 196-208. 225. Augustin, H.G., et al., Control of vascular morphogenesis and homeostasis through the angiopoietin-Tie system. Nature Reviews Molecular Cell Biology, 2009. 10(3): p. 165-177. 226. Saharinen, P., et al., Angiopoietins assemble distinct Tie2 signalling complexes in endothelial cell-cell and cell-matrix contacts. Nature Cell Biology, 2008. 10(5): p. 527-537. 227. Beenken, A. and M. Mohammadi, The FGF family: biology, pathophysiology and therapy. Nature Reviews Drug Discovery, 2009. 8(3): p. 235-253. 228. Kalluri, R., Basement membranes: Structure, assembly and role in tumour angiogenesis. Nature Reviews Cancer, 2003. 3(6): p. 422-433. 229. Park, K.M. and S. Gerecht, Harnessing developmental processes for vascular engineering and regeneration. Development, 2014. 141(14): p. 2760-2769. 230. Semenza, G.L., Hypoxia-Inducible Factors in Physiology and Medicine. Cell, 2012. 148(3): p. 399-408. 231. Bergers, G., et al., Matrix metalloproteinase-9 triggers the angiogenic switch during carcinogenesis. Nature Cell Biology, 2000. 2(10): p. 737-744. 232. Rundhaug, J.E., Matrix metalloproteinases and angiogenesis. Journal of Cellular and Molecular Medicine, 2005. 9(2): p. 267-285. 233. Tvorogov, D., et al., Effective Suppression of Vascular Network Formation by Combination of Antibodies Blocking VEGFR Ligand Binding and Receptor Dimerization. Cancer Cell, 2010. 18(6): p. 630-640. 234. Phng, L.K., et al., Nrarp Coordinates Endothelial Notch and Wnt Signaling to Control Vessel Density in Angiogenesis. Developmental Cell, 2009. 16(1): p. 70-82. 235. Carmeliet, P., et al., Synergism between vascular endothelial growth factor and placental growth factor contributes to angiogenesis and plasma extravasation in pathological conditions. Nature Medicine, 2001. 7(5): p. 575-583. 236. Jain, R.K., Molecular regulation of vessel maturation. Nature Medicine, 2003. 9(6): p. 685-693. 237. Foo, S.S., et al., Ephrin-B2 controls cell motility and adhesion during blood-vessel-wall assembly. Cell, 2006. 124(1): p. 161-173. 238. Lutolf, M.P. and J.A. Hubbell, Synthetic biomaterials as instructive extracellular microenvironments for morphogenesis in tissue engineering. Nature Biotechnology, 2005. 23(1): p. 47-55. 239. Lee, K.Y. and D.J. Mooney, Hydrogels for tissue engineering. Chemical Reviews, 2001. 101(7): p. 1869-1879. 240. Slaughter, B.V., et al., Hydrogels in Regenerative Medicine. Advanced Materials, 2009. 21(32- 33): p. 3307-3329. 241. Rafii, S. and D. Lyden, Therapeutic stem and progenitor cell transplantation for organ vascularization and regeneration. Nature Medicine, 2003. 9(6): p. 702-712. 242. Kusuma, S. and S. Gerecht, Engineering blood vessels using stem cells: innovative approaches to treat vascular disorders. Expert review of cardiovascular therapy, 2010. 8(10): p. 1433-1445. 243. Sieveking, D.P. and M.K.C. Ng, Cell therapies for therapeutic angiogenesis: back to the bench. Vascular Medicine, 2009. 14(2): p. 153-166. 244. Gupta, R., J. Tongers, and D.W. Losordo, Human Studies of Angiogenic Gene Therapy. Circulation Research, 2009. 105(8): p. 724-736. 245. Lazarous, D.F., et al., Pharmacodynamics of basic fibroblast growth factor: route of administration determines myocardial and systemic distribution. Cardiovascular Research, 1997. 36(1): p. 78-85.

237

246. Zisch, A.H., et al., Cell-demanded release of VEGF from synthetic, biointeractive cell-ingrowth matrices for vascularized tissue growth. Faseb Journal, 2003. 17(13): p. 2260-+. 247. Hariawala, M.D., et al., VEGF improves myocardial blood flow but produces EDRF-mediated hypotension in porcine hearts. Journal of Surgical Research, 1996. 63(1): p. 77-82. 248. Lee, R.J., et al., VEGF gene delivery to myocardium - Deleterious effects of unregulated expression. Circulation, 2000. 102(8): p. 898-901. 249. Silva, E.A. and D.J. Mooney, Spatiotemporal control of vascular endothelial growth factor delivery from injectable hydrogels enhances angiogenesis. Journal of Thrombosis and Haemostasis, 2007. 5(3): p. 590-598. 250. Sacchi, V., et al., Long-lasting fibrin matrices ensure stable and functional angiogenesis by highly tunable, sustained delivery of recombinant VEGF(164). Proceedings of the National Academy of Sciences of the United States of America, 2014. 111(19): p. 6952-6957. 251. Zisch, A.H., et al., Covalently conjugated VEGF-fibrin matrices for endothelialization. Journal of Controlled Release, 2001. 72(1-3): p. 101-113. 252. Lee, J. and K.Y. Lee, Local and Sustained Vascular Endothelial Growth Factor Delivery for Angiogenesis Using an Injectable System. Pharmaceutical Research, 2009. 26(7): p. 1739-1744. 253. Lee, J., et al., Active Blood Vessel Formation in the Ischemic Hindlimb Mouse Model Using a Microsphere/Hydrogel Combination System. Pharmaceutical Research, 2010. 27(5): p. 767-774. 254. DeVolder, R., E. Antoniadou, and H. Kong, Enzymatically cross-linked injectable alginate-g- pyrrole hydrogels for neovascularization. Journal of Controlled Release, 2013. 172(1): p. 30-37. 255. des Rieux, A., et al., 3D systems delivering VEGF to promote angiogenesis for tissue engineering. Journal of Controlled Release, 2011. 150(3): p. 272-278. 256. Jeon, O., et al., Control of basic fibroblast growth factor release from fibrin gel with heparin and concentrations of fibrinogen and thrombin. Journal of Controlled Release, 2005. 105(3): p. 249- 259. 257. Jeon, O., et al., Long-term and zero-order release of basic fibroblast growth factor from heparin- conjugated poly(L-lactide-co-glycolide) nanospheres and fibrin gel. Biomaterials, 2006. 27(8): p. 1598-1607. 258. Hosseinkhani, H., et al., Enhanced angiogenesis through controlled release of basic fibroblast growth factor from peptide amphiphile for tissue regeneration. Biomaterials, 2006. 27(34): p. 5836-5844. 259. Saik, J.E., et al., Covalently immobilized platelet-derived growth factor-BB promotes antiogenesis in biomirnetic poly(ethylene glycol) hydrogels. Acta Biomaterialia, 2011. 7(1): p. 133-143. 260. Phelps, E.A., et al., Bioartificial matrices for therapeutic vascularization. Proceedings of the National Academy of Sciences of the United States of America, 2010. 107(8): p. 3323-3328. 261. Kimura, Y. and Y. Tabata, Controlled Release of Stromal-Cell-Derived Factor-1 from Gelatin Hydrogels Enhances Angiogenesis. Journal of Biomaterials Science-Polymer Edition, 2010. 21(1): p. 37-51. 262. Prokoph, S., et al., Sustained delivery of SDF-1 alpha from heparin-based hydrogels to attract circulating pro-angiogenic cells. Biomaterials, 2012. 33(19): p. 4792-4800. 263. Moon, J.J., S.H. Lee, and J.L. West, Synthetic biomimetic hydrogels incorporated with Ephrin-A1 for therapeutic angiogenesis. Biomacromolecules, 2007. 8(1): p. 42-49. 264. Saik, J.E., et al., Biomimetic Hydrogels with Immobilized EphrinA1 for Therapeutic Angiogenesis. Biomacromolecules, 2011. 12(7): p. 2715-2722. 265. Wacker, B.K., et al., Delivery of sphingosine 1-phosphate from poly(ethylene glycol) hydrogels. Biomacromolecules, 2006. 7(4): p. 1335-1343. 266. Mulyasasmita, W., et al., Avidity-controlled delivery of angiogenic peptides from injectable molecular-recognition hydrogels. Tissue engineering. Part A, 2014. 20(15-16): p. 2102-14. 267. Storrie, H. and D.J. Mooney, Sustained delivery of plasmid DNA from polymeric scaffolds for tissue engineering. Advanced Drug Delivery Reviews, 2006. 58(4): p. 500-514.

238

268. Kong, H.J., et al., Design of biodegradable hydrogel for the local and sustained delivery of angiogenic plasmid DNA. Pharmaceutical Research, 2008. 25(5): p. 1230-1238. 269. Shepard, J.A., et al., Hydrogel macroporosity and the prolongation of transgene expression and the enhancement of angiogenesis. Biomaterials, 2012. 33(30): p. 7412-7421. 270. Tokatlian, T., C. Cam, and T. Segura, Non-viral DNA delivery from porous hyaluronic acid hydrogels in mice. Biomaterials, 2014. 35(2): p. 825-835. 271. Leach, J.B., et al., Photocrosslinked hyaluronic acid hydrogels: Natural, biodegradable tissue engineering scaffolds. Biotechnology and Bioengineering, 2003. 82(5): p. 578-589. 272. Leach, J.B. and C.E. Schmidt, Characterization of protein release from photocrosslinkable hyaluronic acid-polyethylene glycol hydrogel tissue engineering scaffolds. Biomaterials, 2005. 26(2): p. 125-135. 273. Pike, D.B., et al., Heparin-regulated release of growth factors in vitro and angiogenic response in vivo to implanted hyaluronan hydrogels containing VEGF and bFGF. Biomaterials, 2006. 27(30): p. 5242-5251. 274. Hanjaya-Putra, D., et al., Spatial control of cell-mediated degradation to regulate vasculogenesis and angiogenesis in hyaluronan hydrogels. Biomaterials, 2012. 33(26): p. 6123-6131. 275. Huang, C.C., et al., 3-D self-assembling leucine zipper hydrogel with tunable properties for tissue engineering. Biomaterials, 2014. 35(20): p. 5316-5326. 276. Peattie, R.A., et al., Stimulation of in vivo angiogenesis by cytokine-loaded hyaluronic acid hydrogel implants. Biomaterials, 2004. 25(14): p. 2789-2798. 277. Zieris, A., et al., Dual independent delivery of pro-angiogenic growth factors from starPEG- heparin hydrogels. Journal of Controlled Release, 2011. 156(1): p. 28-36. 278. Chwalek, K., et al., Glycosaminoglycan-based hydrogels to modulate heterocellular communication in in vitro angiogenesis models. Scientific Reports, 2014. 4. 279. Riley, C.M., et al., Stimulation of in vivo angiogenesis using dual growth factor-loaded crosslinked glycosaminoglycan hydrogels. Biomaterials, 2006. 27(35): p. 5935-5943. 280. Peattie, R.A., et al., Dual growth factor-induced angiogenesis in vivo using hyaluronan hydrogel implants. Biomaterials, 2006. 27(9): p. 1868-1875. 281. Hosack, L.W., et al., Microvascular maturity elicited in tissue treated with cytokine-loaded hyaluronan-based hydrogels. Biomaterials, 2008. 29(15): p. 2336-2347. 282. Sun, G., et al., Functional neovascularization of biodegradable dextran hydrogels with multiple angiogenic growth factors. Biomaterials, 2011. 32(1): p. 95-106. 283. Shin, S.H., et al., Co-delivery of Vascular Endothelial Growth Factor and Angiopoietin-1 Using Injectable Microsphere/Hydrogel Hybrid Systems for Therapeutic Angiogenesis. Pharmaceutical Research, 2013. 30(8): p. 2157-2165. 284. Jay, S.M., et al., Dual delivery of VEGF and MCP-1 to support endothelial cell transplantation for therapeutic vascularization. Biomaterials, 2010. 31(11): p. 3054-3062. 285. Layman, H., et al., Co-delivery of FGF-2 and G-CSF from gelatin-based hydrogels as angiogenic therapy in a murine critical limb ischemic model. Acta Biomaterialia, 2009. 5(1): p. 230-239. 286. Liu, Y., et al., Application of bFGF and BDNF to improve angiogenesis and cardiac function. Journal of Surgical Research, 2006. 136(1): p. 85-91. 287. Matsui, M. and Y. Tabata, Enhanced angiogenesis by multiple release of platelet-rich plasma contents and basic fibroblast growth factor from gelatin hydrogels. Acta Biomaterialia, 2012. 8(5): p. 1792-1801. 288. Zieris, A., et al., FGF-2 and VEGF functionalization of starPEG-heparin hydrogels to modulate biomolecular and physical cues of angiogenesis. Biomaterials, 2010. 31(31): p. 7985-7994. 289. Martino, M.M., et al., Heparin-binding domain of fibrin(ogen) binds growth factors and promotes tissue repair when incorporated within a synthetic matrix. Proceedings of the National Academy of Sciences of the United States of America, 2013. 110(12): p. 4563-4568. 290. Martino, M.M., et al., Growth Factors Engineered for Super-Affinity to the Extracellular Matrix Enhance Tissue Healing. Science, 2014. 343(6173): p. 885-888.

239

291. Zhang, H., et al., Dual-delivery of VEGF and PDGF by double-layered electrospun membranes for blood vessel regeneration. Biomaterials, 2013. 34(9): p. 2202-2212. 292. Cao, L., et al., Promoting angiogenesis via manipulation of VEGF responsiveness with notch signaling. Biomaterials, 2009. 30(25): p. 4085-4093. 293. Chen, C.S., et al., Micropatterned surfaces for control of cell shape, position, and function. Biotechnology Progress, 1998. 14(3): p. 356-363. 294. Dike, L.E., et al., Geometric control of switching between growth, apoptosis, and differentiation during angiogenesis using micropatterned substrates. In Vitro Cellular & Developmental Biology-Animal, 1999. 35(8): p. 441-448. 295. Moon, J.J., et al., Micropatterning of Poly(Ethylene Glycol) Diacrylate Hydrogels with Biomolecules to Regulate and Guide Endothelial Morphogenesis. Tissue Engineering Part A, 2009. 15(3): p. 579-585. 296. Raghavan, S., et al., Geometrically Controlled Endothelial Tubulogenesis in Micropatterned Gels. Tissue Engineering Part A, 2010. 16(7): p. 2255-2263. 297. Chang, C.C., et al., Determinants of Microvascular Network Topologies in Implanted Neovasculatures. Arteriosclerosis Thrombosis and Vascular Biology, 2012. 32(1): p. 5-U55. 298. Chiu, Y.C., et al., The role of pore size on vascularization and tissue remodeling in PEG hydrogels. Biomaterials, 2011. 32(26): p. 6045-6051. 299. Artel, A., et al., An Agent-Based Model for the Investigation of Neovascularization Within Porous Scaffolds. Tissue Engineering Part A, 2011. 17(17-18): p. 2133-2141. 300. Lutolf, M.P., et al., Cell-responsive synthetic hydrogels. Advanced Materials, 2003. 15(11): p. 888-+. 301. Miller, J.S., et al., Bioactive hydrogels made from step-growth derived PEG-peptide macromers. Biomaterials, 2010. 31(13): p. 3736-3743. 302. Yee, D., et al., Hyaluronic Acid Hydrogels Support Cord-Like Structures from Endothelial Colony-Forming Cells. Tissue Engineering Part A, 2011. 17(9-10): p. 1351-1361. 303. Rao, R.R., et al., Matrix composition regulates three-dimensional network formation by endothelial cells and mesenchymal stem cells in collagen/fibrin materials. Angiogenesis, 2012. 15(2): p. 253-264. 304. Wu, Y., M.A. Al-Ameen, and G. Ghosh, Integrated Effects of Matrix Mechanics and Vascular Endothelial Growth Factor (VEGF) on Capillary Sprouting. Annals of Biomedical Engineering, 2014. 42(5): p. 1024-1036. 305. Forget, A., et al., Polysaccharide hydrogels with tunable stiffness and provasculogenic properties via alpha-helix to beta-sheet switch in secondary structure. Proceedings of the National Academy of Sciences of the United States of America, 2013. 110(32): p. 12887-12892. 306. Pugh, C.W. and P.J. Ratcliffe, Regulation of angiogenesis by hypoxia: role of the HIF system. Nature Medicine, 2003. 9(6): p. 677-684. 307. Park, K.M. and S. Gerecht, Hypoxia-inducible hydrogels. Nature Communications, 2014. 5. 308. Park, K.M., M.R. Blatchley, and S. Gerecht, The Design of Dextran-Based Hypoxia-Inducible Hydrogels via In Situ Oxygen-Consuming Reaction. Macromolecular Rapid Communications, 2014: p. n/a-n/a. 309. Ushio-Fukai, M. and R.W. Alexander, Reactive oxygen species as mediators of angiogenesis signaling - Role of NAD(P)H oxidase. Molecular and Cellular Biochemistry, 2004. 264(1-2): p. 85-97. 310. Ushio-Fukai, M. and Y. Nakamura, Reactive oxygen species and angiogenesis: NADPH oxidase as target for cancer therapy. Cancer Letters, 2008. 266(1): p. 37-52. 311. Xu, W.M., et al., The role of nitric oxide in cancer. Cell Research, 2002. 12(5-6): p. 311-320. 312. Kvietys, P.R. and D.N. Granger, Role of reactive oxygen and nitrogen species in the vascular responses to inflammation. Free Radical Biology and Medicine, 2012. 52(3): p. 556-592. 313. Dewhirst, M.W., Y. Cao, and B. Moeller, Cycling hypoxia and free radicals regulate angiogenesis and radiotherapy response. Nature Reviews Cancer, 2008. 8(6): p. 425-437.

240

314. Thom, S.R., Hyperbaric Oxygen: Its Mechanisms and Efficacy. Plastic and Reconstructive Surgery, 2011. 127(1): p. 131S-141S. 315. Goldman, R.J., Hyperbaric Oxygen Therapy for Wound Healing and Limb Salvage: A Systematic Review. Pm&R, 2009. 1(5): p. 471-489. 316. Bekker, A., et al., Dating the rise of atmospheric oxygen. Nature, 2004. 427(6970): p. 117-20. 317. Lyons, T.W., C.T. Reinhard, and N.J. Planavsky, The rise of oxygen in Earth's early ocean and atmosphere. Nature, 2014. 506(7488): p. 307-315. 318. Scott, C., et al., Tracing the stepwise oxygenation of the Proterozoic ocean. Nature, 2008. 452(7186): p. 456-9. 319. Frei, R., et al., Fluctuations in Precambrian atmospheric oxygenation recorded by chromium isotopes. Nature, 2009. 461(7261): p. 250-3. 320. Massabuau, J.C., From low arterial- to low tissue-oxygenation strategy. An evolutionary theory. Respir Physiol, 2001. 128(3): p. 249-61. 321. Massabuau, J.C., Primitive, and protective, our cellular oxygenation status? Mech Ageing Dev, 2003. 124(8-9): p. 857-63. 322. Ma, T., et al., Hypoxia and stem cell-based engineering of mesenchymal tissues. Biotechnol Prog, 2009. 25(1): p. 32-42. 323. Davis, G.E., W. Koh, and A.N. Stratman, Mechanisms controlling human endothelial lumen formation and tube assembly in three-dimensional extracellular matrices. Birth Defects Res C Embryo Today, 2007. 81(4): p. 270-85. 324. Koh, W., et al., In vitro three dimensional collagen matrix models of endothelial lumen formation during vasculogenesis and angiogenesis. Methods Enzymol, 2008. 443: p. 83-101. 325. Stupack, D.G. and D.A. Cheresh, ECM remodeling regulates angiogenesis: endothelial integrins look for new ligands. Sci STKE, 2002. 2002(119): p. pe7. 326. Fraisl, P., et al., Regulation of angiogenesis by oxygen and metabolism. Dev Cell, 2009. 16(2): p. 167-79. 327. Shweiki, D., et al., Induction of vascular endothelial growth factor expression by hypoxia and by glucose deficiency in multicell spheroids: implications for tumor angiogenesis. Proc Natl Acad Sci U S A, 1995. 92(3): p. 768-72. 328. Colville-Nash, P.R. and D.L. Scott, Angiogenesis and rheumatoid arthritis: pathogenic and therapeutic implications. Ann Rheum Dis, 1992. 51(7): p. 919-25. 329. Ben-Yosef, Y., et al., Hypoxia of endothelial cells leads to MMP-2-dependent survival and death. Am J Physiol Cell Physiol, 2005. 289(5): p. C1321-31. 330. Erler, J.T., et al., Hypoxia-induced lysyl oxidase is a critical mediator of bone marrow cell recruitment to form the premetastatic niche. Cancer Cell, 2009. 15(1): p. 35-44. 331. Fahling, M., et al., Regulation of collagen prolyl 4-hydroxylase and matrix metalloproteinases in fibrosarcoma cells by hypoxia. Comp Biochem Physiol C Toxicol Pharmacol, 2004. 139(1-3): p. 119-26. 332. Lee, Y.M., et al., Determination of hypoxic region by hypoxia marker in developing mouse embryos in vivo: a possible signal for vessel development. Dev Dyn, 2001. 220(2): p. 175-86. 333. Adelman, D.M., E. Maltepe, and M.C. Simon, Multilineage embryonic hematopoiesis requires hypoxic ARNT activity. Genes Dev, 1999. 13(19): p. 2478-83. 334. Maltepe, E. and M.C. Simon, Oxygen, genes, and development: an analysis of the role of hypoxic gene regulation during murine vascular development. J Mol Med, 1998. 76(6): p. 391-401. 335. Walls, J.R., et al., Three-dimensional analysis of vascular development in the mouse embryo. PLoS One, 2008. 3(8): p. e2853. 336. Urbich, C. and S. Dimmeler, Endothelial progenitor cells: characterization and role in vascular biology. Circ Res, 2004. 95(4): p. 343-53. 337. Ceradini, D.J., et al., Progenitor cell trafficking is regulated by hypoxic gradients through HIF-1 induction of SDF-1. Nature Medicine, 2004. 10(8): p. 858-864.

241

338. De Falco, E., et al., SDF-1 involvement in endothelial phenotype and ischemia-induced recruitment of bone marrow progenitor cells. Blood, 2004. 104(12): p. 3472-3482. 339. Yoon, C.H., et al., Intercellular adhesion molecule-1 is upregulated in ischemic muscle, which mediates trafficking of endothelial progenitor cells. Arteriosclerosis Thrombosis and Vascular Biology, 2006. 26(5): p. 1066-1072. 340. Chavakis, E., et al., Role of beta 2-integrins for homing and neovascularization capacity of endothelial progenitor cells. Journal of Experimental Medicine, 2005. 201(1): p. 63-72. 341. Ceradini, D.J. and G.C. Gurtner, Homing to hypoxia: HIF-1 as a mediator of progenitor cell recruitment to injured tissue. Trends Cardiovasc Med, 2005. 15(2): p. 57-63. 342. Manalo, D.J., et al., Transcriptional regulation of vascular endothelial cell responses to hypoxia by HIF-1. Blood, 2005. 105(2): p. 659-69. 343. Limper, A.H. and J. Roman, Fibronectin. A versatile matrix protein with roles in thoracic development, repair and infection. Chest, 1992. 101(6): p. 1663-73. 344. Bianchi, F., et al., Microfabrication of fractal polymeric structures for capillary morphogenesis: applications in therapeutic angiogenesis and in the engineering of vascularized tissue. J Biomed Mater Res B Appl Biomater, 2007. 81(2): p. 462-8. 345. Almany, L. and D. Seliktar, Biosynthetic hydrogel scaffolds made from fibrinogen and polyethylene glycol for 3D cell cultures. Biomaterials, 2005. 26(15): p. 2467-77. 346. Fong, G.H., Regulation of angiogenesis by oxygen sensing mechanisms. J Mol Med, 2009. 87(6): p. 549-60. 347. Yamakawa, M., et al., Hypoxia-inducible factor-1 mediates activation of cultured vascular endothelial cells by inducing multiple angiogenic factors. Circ Res, 2003. 93(7): p. 664-73. 348. Phillips, P.G., L.M. Birnby, and A. Narendran, Hypoxia induces capillary network formation in cultured bovine pulmonary microvessel endothelial cells. Am J Physiol, 1995. 268(5 Pt 1): p. L789-800. 349. Ottino, P., et al., Hypoxia activates matrix metalloproteinase expression and the VEGF system in monkey choroid-retinal endothelial cells: Involvement of cytosolic phospholipase A2 activity. Mol Vis, 2004. 10: p. 341-50. 350. Rodesch, F., et al., Oxygen measurements in endometrial and trophoblastic tissues during early pregnancy. Obstet Gynecol, 1992. 80(2): p. 283-5. 351. Siggaard-Andersen, O. and R. Huch, The oxygen status of fetal blood. Acta Anaesthesiol Scand Suppl, 1995. 107: p. 129-35. 352. Jauniaux, E., B. Gulbis, and G.J. Burton, The human first trimester gestational sac limits rather than facilitates oxygen transfer to the foetus--a review. Placenta, 2003. 24 Suppl A: p. S86-93. 353. Burggren, W.W., What is the purpose of the embryonic heart beat? Or how facts can ultimately prevail over physiological dogma. Physiol Biochem Zool, 2004. 77(3): p. 333-45. 354. Drake, C.J. and P.A. Fleming, Vasculogenesis in the day 6.5 to 9.5 mouse embryo. Blood, 2000. 95(5): p. 1671-1679. 355. Huber, T.L., et al., Haemangioblast commitment is initiated in the primitive streak of the mouse embryo. Nature, 2004. 432(7017): p. 625-30. 356. Lucitti, J.L., et al., Vascular remodeling of the mouse yolk sac requires hemodynamic force. Development, 2007. 134(18): p. 3317-26. 357. Harrison, J.S., et al., Oxygen saturation in the bone marrow of healthy volunteers. Blood, 2002. 99(1): p. 394. 358. Kumar, R., F. Stepanek, and A. Mantalaris, An oxygen transport model for human bone marrow microcirculation. Food and Bioproducts Processing, 2004. 82(C2): p. 105-116. 359. Parmar, K., et al., Distribution of hematopoietic stem cells in the bone marrow according to regional hypoxia. Proc Natl Acad Sci U S A, 2007. 104(13): p. 5431-6. 360. Hirschi, K.K., D.A. Ingram, and M.C. Yoder, Assessing identity, phenotype, and fate of endothelial progenitor cells. Arteriosclerosis, Thrombosis, and Vascular Biology, 2008. 28(9): p. 1584-1595.

242

361. Asahara, T. and A. Kawamoto, Endothelial progenitor cells for postnatal vasculogenesis. Am J Physiol Cell Physiol, 2004. 287(3): p. C572-9. 362. Asahara, T., et al., Bone marrow origin of endothelial progenitor cells responsible for postnatal vasculogenesis in physiological and pathological neovascularization. Circ Res, 1999. 85(3): p. 221-8. 363. Timmermans, F., et al., Endothelial progenitor cells: Identity defined? Journal of Cellular and Molecular Medicine, 2009. 13(1): p. 87-102. 364. Chow, D.C., et al., Modeling pO(2) distributions in the bone marrow hematopoietic compartment. I. Krogh's model. Biophys J, 2001. 81(2): p. 675-84. 365. Kumar, R., et al., Coupled oxygen-carbon dioxide transport model for the human bone marrow. Food and Bioproducts Processing, 2008. 86(C3): p. 211-219. 366. Forristal, C.E., et al., Hypoxia inducible factors regulate pluripotency and proliferation in human embryonic stem cells cultured at reduced oxygen tensions. Reproduction, 2010. 139(1): p. 85-97. 367. Ezashi, T., P. Das, and R.M. Roberts, Low O2 tensions and the prevention of differentiation of hES cells. Proc Natl Acad Sci U S A, 2005. 102(13): p. 4783-8. 368. Covello, K.L., et al., HIF-2alpha regulates Oct-4: effects of hypoxia on stem cell function, embryonic development, and tumor growth. Genes Dev, 2006. 20(5): p. 557-70. 369. Walter, D.H., et al., Statin therapy accelerates reendothelialization: a novel effect involving mobilization and incorporation of bone marrow-derived endothelial progenitor cells. Circulation, 2002. 105(25): p. 3017-24. 370. Werner, N., et al., Intravenous transfusion of endothelial progenitor cells reduces neointima formation after vascular injury. Circ Res, 2003. 93(2): p. e17-24. 371. Heissig, B., et al., Recruitment of stem and progenitor cells from the bone marrow niche requires MMP-9 mediated release of kit-ligand. Cell, 2002. 109(5): p. 625-37. 372. Hopf, H.W. and M.D. Rollins, Wounds: an overview of the role of oxygen. Antioxid Redox Signal, 2007. 9(8): p. 1183-92. 373. Artuc, M., et al., Mast cells and their mediators in cutaneous wound healing--active participants or innocent bystanders? Exp Dermatol, 1999. 8(1): p. 1-16. 374. IA, S., Cellular microenvironment in healing and non-healing wounds in Soft and hard tissue repair, H. TK, Editor. 1984, Praeger: New York. p. 50-66. 375. Zhang, X., et al., Impaired angiogenesis and mobilization of circulating angiogenic cells in HIF- 1alpha heterozygous-null mice after burn wounding. Wound Repair Regen, 2010. 376. Albina, J.E., et al., HIF-1 expression in healing wounds: HIF-1alpha induction in primary inflammatory cells by TNF-alpha. Am J Physiol Cell Physiol, 2001. 281(6): p. C1971-7. 377. Huang, P.H., et al., Matrix Metalloproteinase-9 Is Essential for Ischemia-Induced Neovascularization by Modulating Bone Marrow-Derived Endothelial Progenitor Cells. Arteriosclerosis Thrombosis and Vascular Biology, 2009. 29(8): p. 1179-U58. 378. Abbott, J.D., et al., Stromal cell-derived factor-1 alpha plays a critical role in stem cell recruitment to the heart after myocardial infarction but is not sufficient to induce homing in the absence of injury. Circulation, 2004. 110(21): p. 3300-3305. 379. Urbich, C., et al., Cathepsin L is required for endothelial progenitor cell-induced neovascularization. Nature Medicine, 2005. 11(2): p. 206-213. 380. Ward, J.P., Oxygen sensors in context. Biochim Biophys Acta, 2008. 1777(1): p. 1-14. 381. Bruick, R.K. and S.L. McKnight, A conserved family of prolyl-4-hydroxylases that modify HIF. Science, 2001. 294(5545): p. 1337-40. 382. Hirsila, M., et al., Characterization of the human prolyl 4-hydroxylases that modify the hypoxia- inducible factor. J Biol Chem, 2003. 278(33): p. 30772-80. 383. Fukumura, D., S. Kashiwagi, and R.K. Jain, The role of nitric oxide in tumour progression. Nat Rev Cancer, 2006. 6(7): p. 521-34.

243

384. Jones, C.I., 3rd, et al., Endothelial cell respiration is affected by the oxygen tension during shear exposure: role of mitochondrial peroxynitrite. Am J Physiol Cell Physiol, 2008. 295(1): p. C180- 91. 385. Ushio-Fukai, M. and Y. Nakamura, Reactive oxygen species and angiogenesis: NADPH oxidase as target for cancer therapy. Cancer Lett, 2008. 266(1): p. 37-52. 386. Evans, A.M., et al., Does AMP-activated protein kinase couple inhibition of mitochondrial oxidative phosphorylation by hypoxia to calcium signaling in O2-sensing cells? J Biol Chem, 2005. 280(50): p. 41504-11. 387. Papandreou, I., et al., HIF-1 mediates adaptation to hypoxia by actively downregulating mitochondrial oxygen consumption. Cell Metab, 2006. 3(3): p. 187-97. 388. Wolin, M.S., M. Ahmad, and S.A. Gupte, Oxidant and redox signaling in vascular oxygen sensing mechanisms: basic concepts, current controversies, and potential importance of cytosolic NADPH. Am J Physiol Lung Cell Mol Physiol, 2005. 289(2): p. L159-73. 389. Williams, S.E., et al., Hemoxygenase-2 is an oxygen sensor for a calcium-sensitive potassium channel. Science, 2004. 306(5704): p. 2093-7. 390. Hirota, K. and G.L. Semenza, Regulation of hypoxia-inducible factor 1 by prolyl and asparaginyl hydroxylases. Biochem Biophys Res Commun, 2005. 338(1): p. 610-6. 391. Steinlechner-Maran, R., et al., Oxygen dependence of respiration in coupled and uncoupled endothelial cells. Am J Physiol, 1996. 271(6 Pt 1): p. C2053-61. 392. Boveris, A., et al., Regulation of mitochondrial respiration by oxygen and nitric oxide. Ann N Y Acad Sci, 2000. 899: p. 121-35. 393. Abaci, H.E., et al., Adaptation to oxygen deprivation in cultures of human pluripotent stem cells, endothelial progenitor cells, and umbilical vein endothelial cells. Am J Physiol Cell Physiol, 2010. 394. Airley, R., et al., Glucose transporter glut-1 expression correlates with tumor hypoxia and predicts metastasis-free survival in advanced carcinoma of the cervix. Clin Cancer Res, 2001. 7(4): p. 928-34. 395. Tsai, A.G., P.C. Johnson, and M. Intaglietta, Oxygen gradients in the microcirculation. Physiol Rev, 2003. 83(3): p. 933-63. 396. Garedew, A., U. Kammerer, and D. Singer, Respiratory response of malignant and placental cells to changes in oxygen concentration. Respir Physiol Neurobiol, 2009. 165(2-3): p. 154-60. 397. Helmlinger, G., et al., Formation of endothelial cell networks. Nature, 2000. 405(6783): p. 139- 41. 398. Xu, W., et al., Alterations of cellular bioenergetics in pulmonary artery endothelial cells. Proc Natl Acad Sci U S A, 2007. 104(4): p. 1342-7. 399. Thurston, G., et al., Leakage-resistant blood vessels in mice transgenically overexpressing angiopoietin-1. Science, 1999. 286(5449): p. 2511-4. 400. Holash, J., et al., Vessel cooption, regression, and growth in tumors mediated by angiopoietins and VEGF. Science, 1999. 284(5422): p. 1994-8. 401. Holash, J., S.J. Wiegand, and G.D. Yancopoulos, New model of tumor angiogenesis: dynamic balance between vessel regression and growth mediated by angiopoietins and VEGF. Oncogene, 1999. 18(38): p. 5356-62. 402. Zhang, H., et al., Mitochondrial autophagy is an HIF-1-dependent adaptive metabolic response to hypoxia. J Biol Chem, 2008. 283(16): p. 10892-903. 403. Papandreou, I., et al., Hypoxia signals autophagy in tumor cells via AMPK activity, independent of HIF-1, BNIP3, and BNIP3L. Cell Death Differ, 2008. 15(10): p. 1572-81. 404. Webster, K.A., Puma joins the battery of BH3-only proteins that promote death and infarction during myocardial ischemia. Am J Physiol Heart Circ Physiol, 2006. 291(1): p. H20-2. 405. Martorell, L., et al., The hypoxia-inducible factor 1/NOR-1 axis regulates the survival response of endothelial cells to hypoxia. Mol Cell Biol, 2009. 29(21): p. 5828-42.

244

406. Bellot, G., et al., Hypoxia-induced autophagy is mediated through hypoxia-inducible factor induction of BNIP3 and BNIP3L via their BH3 domains. Mol Cell Biol, 2009. 29(10): p. 2570-81. 407. Li, C., et al., CD105 prevents apoptosis in hypoxic endothelial cells. J Cell Sci, 2003. 116(Pt 13): p. 2677-85. 408. Band, M., et al., Hypoxia-induced BNIP3 expression and mitophagy: in vivo comparison of the rat and the hypoxia-tolerant mole rat, Spalax ehrenbergi. FASEB J, 2009. 23(7): p. 2327-35. 409. Iwasaki, H., The Niche Regulation of Hematopoietic Stem Cells, in Regulatory Networks in Stem Cells 2009, Humana Press. p. 165-175. 410. Ji, L., et al., Self-renewal and pluripotency is maintained in human embryonic stem cells by co- culture with human fetal liver stromal cells expressing hypoxia inducible factor 1alpha. J Cell Physiol, 2009. 221(1): p. 54-66. 411. Prasad, S.M., et al., Continuous hypoxic culturing maintains activation of Notch and allows long- term propagation of human embryonic stem cells without spontaneous differentiation. Cell Prolif, 2009. 42(1): p. 63-74. 412. Yoshida, Y., et al., Hypoxia enhances the generation of induced pluripotent stem cells. Cell Stem Cell, 2009. 5(3): p. 237-41. 413. Niebruegge, S., et al., Generation of human embryonic stem cell-derived mesoderm and cardiac cells using size-specified aggregates in an oxygen-controlled bioreactor. Biotechnol Bioeng, 2009. 102(2): p. 493-507. 414. Koay, E.J. and K.A. Athanasiou, Hypoxic chondrogenic differentiation of human embryonic stem cells enhances cartilage protein synthesis and biomechanical functionality. Osteoarthritis Cartilage, 2008. 16(12): p. 1450-6. 415. Prado-Lopez, S., et al., Hypoxia Promotes Efficient Differentiation of Human Embryonic Stem Cells to Functional Endothelium. Stem Cells, 2010. 416. Lee, S.W., et al., Hypoxic priming of mESCs accelerates vascular-lineage differentiation through HIF1-mediated inverse regulation of Oct4 and VEGF. Embo Molecular Medicine, 2012. 4(9): p. 924-938. 417. Kang, X.L., et al., Nox2 contributes to the arterial endothelial specification of mouse induced pluripotent stem cells by upregulating Notch signaling. Scientific Reports, 2016. 6: p. 13. 418. Tsang, K.M., et al., Embryonic Stem Cell Differentiation to Functional Arterial Endothelial Cells through Sequential Activation of ETV2 and NOTCH1 Signaling by HIF1 alpha. Stem Cell Reports, 2017. 9(3): p. 796-806. 419. Akita, T., et al., Hypoxic preconditioning augments efficacy of human endothelial progenitor cells for therapeutic neovascularization. Lab Invest, 2003. 83(1): p. 65-73. 420. Jiang, M., et al., Angiogenesis by transplantation of HIF-1 alpha modified EPCs into ischemic limbs. J Cell Biochem, 2008. 103(1): p. 321-34. 421. Toole, B.P., Hyaluronan: from extracellular glue to pericellular cue. Nat Rev Cancer, 2004. 4(7): p. 528-39. 422. Wijelath, E.S., et al., Heparin-II Domain of Fibronectin Is a Vascular Endothelial Growth Factor-Binding Domain: Enhancement of VEGF Biological Activity by a Singular Growth Factor/Matrix Protein Synergism. Circ Res, 2006. 99(8): p. 853-860. 423. Francis, S.E., et al., Central roles of alpha5beta1 integrin and fibronectin in vascular development in mouse embryos and embryoid bodies. Arterioscler Thromb Vasc Biol, 2002. 22(6): p. 927-33. 424. Astrof, S., D. Crowley, and R.O. Hynes, Multiple cardiovascular defects caused by the absence of alternatively spliced segments of fibronectin. Developmental Biology, 2007. 311(1): p. 11-24. 425. Toole, B.P., Hyaluronan in morphogenesis. Semin Cell Dev Biol, 2001. 12(2): p. 79-87. 426. Banerjee, S.D. and B.P. Toole, Hyaluronan-binding protein in endothelial cell morphogenesis. Journal of Cell Biology, 1992. 119(3): p. 643-652.

245

427. Evanko, S.P., W.T. Parks, and T.N. Wight, Intracellular hyaluronan in arterial smooth muscle cells: association with microtubules, RHAMM, and the mitotic spindle. J Histochem Cytochem, 2004. 52(12): p. 1525-35. 428. Camenisch, T.D., et al., Disruption of hyaluronan synthase-2 abrogates normal cardiac morphogenesis and hyaluronan-mediated transformation of epithelium to mesenchyme. Journal of Clinical Investigation, 2000. 106(3): p. 349-360. 429. Gerecht, S., et al., Hyaluronic acid hydrogel for controlled self-renewal and differentiation of human embryonic stem cells. Proc Natl Acad Sci U S A, 2007. 104(27): p. 11298-303. 430. Hanjaya-Putra, D. and S. Gerecht, Vascular engineering using human embryonic stem cells. Biotechnology Progress, 2009. 25(1): p. 2-9. 431. Hou, L.Q., et al., Combinatorial extracellular matrix microenvironments promote survival and phenotype of human induced pluripotent stem cell-derived endothelial cells in hypoxia. Acta Biomaterialia, 2016. 44: p. 188-199. 432. Davis, G.E. and D.R. Senger, Extracellular matrix mediates a molecular balance between vascular morphogenesis and regression. Current Opinion in Hematology, 2008. 15(3): p. 197- 203. 433. Robins, S.P., Biochemistry and functional significance of collagen cross-linking. Biochemical Society Transactions, 2007. 35(5): p. 849-852. 434. Wolburg, H. and A. Lippoldt, Tight junctions of the blood-brain barrier: Development, composition and regulation. Vascular Pharmacology, 2002. 38(6): p. 323-337. 435. Eble, J.A. and S. Niland, The Extracellular Matrix of Blood Vessels. Current Pharmaceutical Design, 2009. 15: p. 1385-1400. 436. Wanjare, M., S. Kusuma, and S. Gerecht, Defining Differences among Perivascular Cells Derived from Human Pluripotent Stem Cells. Stem Cell Reports, 2014. 2(5): p. 561-575. 437. Kusuma, S., S. Zhao, and S. Gerecht, The extracellular matrix is a novel attribute of endothelial progenitors and of hypoxic mature endothelial cells. Faseb Journal, 2012. 26(12): p. 4925-4936. 438. Au, P., et al., Bone marrow derived mesenchymal stem cells facilitate engineering of long-lasting functional vasculature. Blood, 2008. 111(9): p. 4551-4558. 439. Melero-Martin, J.M., et al., Engineering robust and functional vascular networks in vivo with human adult and cord blood-derived progenitor cells. Circulation Research, 2008. 103(2): p. 194- 202. 440. Critser, P.J., et al., Collagen matrix physical properties modulate endothelial colony forming cell- derived vessels in vivo. Microvascular Research. In Press, Uncorrected Proof. 441. Kreger, S.T. and S.L. Voytik-Harbin, Hyaluronan concentration within a 3D collagen matrix modulates matrix viscoelasticity, but not fibroblast response. Matrix Biology, 2009. 28(6): p. 336-346. 442. Chiu, L.L.Y. and M. Radisic, Scaffolds with covalently immobilized VEGF and Angiopoietin-1 for vascularization of engineered tissues. Biomaterials. 31(2): p. 226-241. 443. Langer, R. and D.A. Tirrell, Designing materials for biology and medicine. Nature, 2004. 428(6982): p. 487-492. 444. Sun, G. and C.-C. Chu, Synthesis, characterization of biodegradable dextran-allyl isocyanate- ethylamine/polyethylene glycol-diacrylate hydrogels and their in vitro release of albumin. Carbohydrate Polymers, 2006. 65(3): p. 273-287. 445. Sun, G., et al., Functional groups affect physical and biological properties of dextran-based hydrogels. Journal of Biomedical Materials Research Part A, 2009. 9999(9999): p. NA. 446. Gerecht-Nir, S., et al., Three-dimensional porous alginate scaffolds provide a conducive environment for generation of well-vascularized embryoid bodies from human embryonic stem cells. Biotechnol Bioeng, 2004. 88(3): p. 313-20. 447. Ehrbar, M., et al., Cell-demanded liberation of VEGF121 from fibrin implants induces local and controlled blood vessel growth. Circ Res, 2004. 94(8): p. 1124-32.

246

448. Ehrbar, M., et al., Endothelial cell proliferation and progenitor maturation by fibrin-bound VEGF variants with differential susceptibilities to local cellular activity. J Control Release, 2005. 101(1-3): p. 93-109. 449. Phelps, E.A., et al., Bioartificial matrices for therapeutic vascularization. Proceedings of the National Academy of Sciences. 107(8): p. 3323-3328. 450. Moon, J.J., et al., Biomimetic hydrogels with pro-angiogenic properties. Biomaterials. 31(14): p. 3840-3847. 451. Pierschbacher, M.D. and E. Ruoslahti, Cell attachment activity of fibronectin can be duplicated by small synthetic fragments of the molecule. Nature, 1984. 309(5963): p. 30-33. 452. Silva, G.A., et al., Selective differentiation of neural progenitor cells by high-epitope density nanofibers. Science, 2004. 303(5662): p. 1352-5. 453. Desgrosellier, J.S. and D.A. Cheresh, Integrins in cancer: biological implications and therapeutic opportunities. Nature Reviews Cancer, 2010. 10(1): p. 9-22. 454. Gobin, A.S. and J.L. West, Cell migration through defined, synthetic extracellular matrix analogues. FASEB J., 2002: p. 01-0759fje. 455. Ingber, D.E. and J. Folkman, Mechanochemical switching between growth and differentiation during fibroblast growth factor-stimulated angiogenesis in vitro: Role of extracellular matrix. Journal of Cell Biology, 1989. 109(1): p. 317-330. 456. Folkman, J., C.C. Haudenschild, and B.R. Zetter, Long-term culture of capillary endothelial cells. Proc Natl Acad Sci U S A, 1979. 76(10): p. 5217-21. 457. Davis, G.E. and C.W. Camarillo, Regulation of endothelial cell morphogenesis by integrins, mechanical forces, and matrix guidance pathways. Experimental Cell Research, 1995. 216(1): p. 113-123. 458. Stratman, A.N., et al., Endothelial cell lumen and vascular guidance tunnel formation requires MT1-MMP-dependent proteolysis in 3-dimensional collagen matrices. Blood, 2009. 114(2): p. 237-247. 459. Sage, E.H. and R.B. Vernon, Regulation of angiogenesis by extracellular matrix: The growth and the glue. Journal of Hypertension, 1994. 12(SUPPL. 10). 460. Sieminski, A.L., et al., The stiffness of three-dimensional ionic self-assembling peptide gels affects the extent of capillary-like network formation. Cell Biochem Biophys, 2007. 49(2): p. 73- 83. 461. Zisch, A.H., et al., Cell-demanded release of VEGF from synthetic, biointeractive cell ingrowth matrices for vascularized tissue growth. FASEB J, 2003. 17(15): p. 2260-2. 462. Seliktar, D., et al., MMP-2 sensitive, VEGF-bearing bioactive hydrogels for promotion of vascular healing. Journal of Biomedical Materials Research - Part A, 2004. 68(4): p. 704-716. 463. Zisch, A.H., M.P. Lutolf, and J.A. Hubbell, Biopolymeric delivery matrices for angiogenic growth factors. Cardiovasc Pathol, 2003. 12(6): p. 295-310. 464. Au, P., et al., Differential in vivo potential of endothelial progenitor cells from human umbilical cord blood and adult peripheral blood to form functional long-lasting vessels. Blood, 2008. 111(3): p. 1302-1305. 465. Levenberg, S., et al., Engineering vascularized skeletal muscle tissue. Nat Biotechnol, 2005. 23(7): p. 879-84. 466. N., K., et al., Tissue engineering: creation of long-lasting blood vessels. Nature, 2004. 428: p. 138. 467. Hanjaya-Putra, D.G., Sharon, Mending the Failing Heart with a Vascularized Cardiac Patch Cell Stem Cell, 2009. 5 (6): p. 575-576. 468. O., C., et al., Tissue engineering of vascularized cardiac muscle from human embryonic stem cells. Circ Res, 2007. 100: p. 263.

247

469. Caspi, O., et al., Tissue engineering of vascularized cardiac muscle from human embryonic stem cells. Circ Res, 2007. 100(2): p. 263-72. 470. Jain, R.K., et al., Engineering vascularized tissue. Nat Biotech, 2005. 23(7): p. 821-823. 471. Vo, E., et al., Smooth-Muscle-Like Cells Derived from Human Embryonic Stem Cells Support and Augment Cord-Like Structures In Vitro. Stem Cell Rev, 2010. 472. Hirschi, K.K. and P.A. D'Amore, Pericytes in the microvasculature. Cardiovascular Research, 1996. 32(4): p. 687-698. 473. Hirschi, K.K., S.A. Rohovsky, and P.A. D'Amore, PDGF, TGF-Î², and heterotypic cell-cell interactions mediate endothelial cell-induced recruitment of 10T1/2 cells and their differentiation to a smooth muscle fate. Journal of Cell Biology, 1998. 141(3): p. 805-814. 474. Soucy, P.A. and L.H. Romer, Endothelial cell adhesion, signaling, and morphogenesis in fibroblast-derived matrix. Matrix Biology, 2009. 28(5): p. 273-283. 475. Sorrell, J.M., M.A. Baber, and A.I. Caplan, A self-assembled fibroblast-endothelial cell co- culture system that supports in vitro vasculogenesis by both human umbilical vein endothelial cells and human dermal microvascular endothelial cells. Cells Tissues Organs, 2007. 186(3): p. 157-68. 476. Sottile, J. and D.C. Hocking, Fibronectin polymerization regulates the composition and stability of extracellular matrix fibrils and cell-matrix adhesions. Mol Biol Cell, 2002. 13(10): p. 3546-59. 477. Romer, L.H., K.G. Birukov, and J.G. Garcia, Focal adhesions: paradigm for a signaling nexus. Circ Res, 2006. 98(5): p. 606-16. 478. Zhou, X., et al., Fibronectin fibrillogenesis regulates three-dimensional neovessel formation. Genes and Development, 2008. 22(9): p. 1231-1243. 479. Dickinson, L.E., M.E. Moura, and S. Gerecht, Guiding endothelial progenitor cell tube formation using patterned fibronectin surfaces. Soft Matter, 2010. 6(20): p. 5109-5119. 480. Lai, E.S., et al., Aligned nanofibrillar collagen regulates endothelial organization and migration. Regenerative Medicine, 2012. 7(5): p. 649-661. 481. Rosenfeld, D., et al., Morphogenesis of 3D vascular networks is regulated by tensile forces. Proceedings of the National Academy of Sciences of the United States of America, 2016. 113(12): p. 3215-3220. 482. Bettinger, C.J., Zhang, Z., Gerecht, S., Borenstein, J., Langer R. , Enhancement of in vitro capillary tube formation by substrate nanotopography. Adv. Mater. , 2008. 20: p. 99-103. 483. Gafni, Y., et al., Design of a filamentous polymeric scaffold for in vivo guided angiogenesis. Tissue Eng, 2006. 12(11): p. 3021-34. 484. Igarashi, S., J. Tanaka, and H. Kobayashi, Micro-patterned nanofibrous biomaterials. J Nanosci Nanotechnol, 2007. 7(3): p. 814-7. 485. Dickinson, L.E., et al., Functional surfaces for high-resolution analysis of cancer cell interactions on exogenous hyaluronic acid. Biomaterials, 2010. 31(20): p. 5472-8. 486. Ingber, D.E., Mechanical signaling and the cellular response to extracellular matrix in angiogenesis and cardiovascular physiology. Circulation Research, 2002. 91(10): p. 877-887. 487. Kilarski, W.W., et al., Biomechanical regulation of blood vessel growth during tissue vascularization. Nature Medicine, 2009. 15(6): p. 657-664. 488. Mammoto, A., et al., A mechanosensitive transcriptional mechanism that controls angiogenesis. Nature, 2009. 457(7233): p. 1103-1108. 489. Deroanne, C.F., C.M. Lapiere, and B.V. Nusgens, In vitro tubulogenesis of endothelial cells by relaxation of the coupling extracellular matrix-cytoskeleton. Cardiovascular Research, 2001. 49(3): p. 647-658. 490. Sieminski, A.L., R.P. Hebbel, and K.J. Gooch, The relative magnitudes of endothelial force generation and matrix stiffness modulate capillary morphogenesis in vitro. Experimental Cell Research, 2004. 297(2): p. 574-584. 491. Stephanou, A., et al., The rigidity in fibrin gels as a contributing factor to the dynamics of in vitro vascular cord formation. Microvasc Res, 2007. 73(3): p. 182-90.

248

492. Roeder, B.A., et al., Tensile mechanical properties of three-dimensional type I collagen extracellular matrices with varied microstructure. Journal of Biomechanical Engineering, 2002. 124(2): p. 214-222. 493. Yung, C.W., et al., Transglutaminase crosslinked gelatin as a tissue engineering scaffold. Journal of Biomedical Materials Research Part A, 2007. 83A(4): p. 1039-1046. 494. Springett, R. and H.M. Swartz, Measurements of oxygen in vivo: overview and perspectives on methods to measure oxygen within cells and tissues. Antioxid Redox Signal, 2007. 9(8): p. 1295- 301. 495. Shaw, A.D., et al., Assessment of tissue oxygen tension: comparison of dynamic fluorescence quenching and polarographic electrode technique. Crit Care, 2002. 6(1): p. 76-80. 496. Cheema, U., et al., Spatially defined oxygen gradients and vascular endothelial growth factor expression in an engineered 3D cell model. Cell Mol Life Sci, 2008. 65(1): p. 177-86. 497. Lewis, M.C., et al., Heterogeneous proliferation within engineered cartilaginous tissue: the role of oxygen tension. Biotechnol Bioeng, 2005. 91(5): p. 607-15. 498. Kellner, K., et al., Determination of oxygen gradients in engineered tissue using a fluorescent sensor. Biotechnol Bioeng, 2002. 80(1): p. 73-83. 499. Brown, D.A., et al., Analysis of oxygen transport in a diffusion-limited model of engineered heart tissue. Biotechnol Bioeng, 2007. 97(4): p. 962-75. 500. Landman, K.A. and A.Q. Cai, Cell proliferation and oxygen diffusion in a vascularising scaffold. Bull Math Biol, 2007. 69(7): p. 2405-28. 501. Galban, C.J. and B.R. Locke, Effects of spatial variation of cells and nutrient and product concentrations coupled with product inhibition on cell growth in a polymer scaffold. Biotechnol Bioeng, 1999. 64(6): p. 633-43. 502. Guaccio, A., et al., Oxygen consumption of chondrocytes in agarose and collagen gels: a comparative analysis. Biomaterials, 2008. 29(10): p. 1484-93. 503. Mamchaoui, K. and G. Saumon, A method for measuring the oxygen consumption of intact cell monolayers. Am J Physiol Lung Cell Mol Physiol, 2000. 278(4): p. L858-63. 504. Mehta, G., et al., Quantitative measurement and control of oxygen levels in microfluidic poly(dimethylsiloxane) bioreactors during cell culture. Biomed Microdevices, 2007. 9(2): p. 123- 34. 505. Galban, C.J. and B.R. Locke, Analysis of cell growth kinetics and substrate diffusion in a polymer scaffold. Biotechnol Bioeng, 1999. 65(2): p. 121-32. 506. Lanza, V.A., D. Preziosi, L., Exogenous control of vascular network formation in vitro: a mathematical model Networks and Heterogeneous Media, 2006. 1(4): p. 621-638. 507. Daphne, M., A mechanochemical model of angiogenesis and vasculogenesis. Modelisation mathematique et analyse numerique, 2003. 37(4): p. 581-599. 508. Kumar, R., F. Stepanek, and A. Mantalaris, A conceptual model for oxygen transport in the human marrow. Modelling and Control in Biomedical Systems 2003 (Including Biological Systems), 2003: p. 365-370 538. 509. Popel, A.S., Theory of oxygen transport to tissue. Crit Rev Biomed Eng, 1989. 17(3): p. 257-321. 510. Segura, I., et al., Inhibition of programmed cell death impairs in vitro vascular-like structure formation and reduces in vivo angiogenesis. FASEB J, 2002. 16(8): p. 833-41. 511. Giannelli, G., et al., Induction of cell migration by matrix metalloprotease-2 cleavage of laminin- 5. Science, 1997. 277(5323): p. 225-8. 512. Cheresh, D.A. and D.G. Stupack, Regulation of angiogenesis: apoptotic cues from the ECM. Oncogene, 2008. 27(48): p. 6285-98. 513. Pollard, P.J., et al., Regulation of Jumonji-domain-containing histone demethylases by hypoxia- inducible factor (HIF)-1alpha. Biochem J, 2008. 416(3): p. 387-94.

249

514. Hagemann, T., et al., Enhanced invasiveness of breast cancer cell lines upon co-cultivation with macrophages is due to TNF-alpha dependent up-regulation of matrix metalloproteases. Carcinogenesis, 2004. 25(8): p. 1543-9. 515. Hofmann, U.B., et al., Expression of integrin alpha(v)beta(3) correlates with activation of membrane-type matrix metalloproteinase-1 (MT1-MMP) and matrix metalloproteinase-2 (MMP- 2) in human melanoma cells in vitro and in vivo. Int J Cancer, 2000. 87(1): p. 12-9. 516. Myllyharju, J. and E. Schipani, Extracellular matrix genes as hypoxia-inducible targets. Cell Tissue Res, 2010. 339(1): p. 19-29. 517. Tajima, R., et al., Hypoxic enhancement of type IV collagen secretion accelerates adipose conversion of 3T3-L1 fibroblasts. Biochim Biophys Acta, 2001. 1540(3): p. 179-87. 518. Horino, Y., et al., Prolonged hypoxia accelerates the posttranscriptional process of collagen synthesis in cultured fibroblasts. Life Sci, 2002. 71(26): p. 3031-45. 519. Berg, J.T., et al., Alveolar hypoxia increases gene expression of extracellular matrix proteins and platelet-derived growth factor-B in lung parenchyma. Am J Respir Crit Care Med, 1998. 158(6): p. 1920-8. 520. Saunders, W.B., K.J. Bayless, and G.E. Davis, MMP-1 activation by serine proteases and MMP- 10 induces human capillary tubular network collapse and regression in 3D collagen matrices. J Cell Sci, 2005. 118(Pt 10): p. 2325-40. 521. Shen, Y.H., M.S. Shoichet, and M. Radisic, Vascular endothelial growth factor immobilized in collagen scaffold promotes penetration and proliferation of endothelial cells. Acta Biomater, 2008. 4(3): p. 477-89. 522. Lewis, D.M., et al., Intratumoral oxygen gradients mediate sarcoma cell invasion. Proceedings of the National Academy of Sciences, 2016. 523. Adler, M., et al., Generation of oxygen gradients with arbitrary shapes in a microfluidic device. Lab on a Chip, 2010. 10(3): p. 388-391. 524. Funamoto, K., et al., A novel microfluidic platform for high-resolution imaging of a three- dimensional cell culture under a controlled hypoxic environment. Lab on a Chip, 2012. 12(22): p. 4855-4863. 525. Oppegard, S.C., et al., Precise control over the oxygen conditions within the Boyden chamber using a microfabricated insert. Lab on a Chip, 2010. 10(18): p. 2366-2373. 526. Abaci, H.E., et al., Recapitulating physiological and pathological shear stress and oxygen to model vasculature in health and disease. Scientific Reports, 2014. 4: p. 9. 527. Abaci, H.E., et al., Design and development of microbioreactors for long-term cell culture in controlled oxygen microenvironments. Biomedical Microdevices, 2012. 14(1): p. 145-152. 528. Park, T.G. and A.S. Hoffman, SYNTHESIS AND CHARACTERIZATION OF PH- AND OR TEMPERATURE-SENSITIVE HYDROGELS. Journal of Applied Polymer Science, 1992. 46(4): p. 659-671. 529. Sun, G.M., X.Z. Zhang, and C.C. Chu, Formulation and characterization of chitosan-based hydrogel films having both temperature and pH sensitivity. Journal of Materials Science- Materials in Medicine, 2007. 18(8): p. 1563-1577. 530. Klumb, L.A. and T.A. Horbett, DESIGN OF INSULIN DELIVERY DEVICES BASED ON GLUCOSE SENSITIVE MEMBRANES. Journal of Controlled Release, 1992. 18(1): p. 59-80. 531. Khetan, S., C. Chung, and J.A. Burdick, Tuning hydrogel properties for applications in tissue engineering. Conf Proc IEEE Eng Med Biol Soc, 2009. 1: p. 2094-6. 532. Chung, S. and D.J. Andrew, The formation of epithelial tubes. Journal of Cell Science, 2008. 121(21): p. 3501-3504. 533. Lutolf, M.P., Biomaterials: Spotlight on hydrogels. Nat Mater, 2009. 8(6): p. 451-453. 534. Vega, S.L., et al., Combinatorial hydrogels with biochemical gradients for screening 3D cellular microenvironments. Nature Communications, 2018. 9: p. 10. 535. Kloxin, A.M., et al., Photodegradable hydrogels for dynamic tuning of physical and chemical properties. Science, 2009. 324(5923): p. 59-63.

250

536. Toh, Y.C., et al., A novel 3D mammalian cell perfusion-culture system in microfluidic channels. Lab Chip, 2007. 7(3): p. 302-9. 537. Figallo, E., et al., Micro-bioreactor array for controlling cellular microenvironments. Lab Chip, 2007. 7(6): p. 710-9. 538. Korin, N., et al., Periodic "flow-stop" perfusion microchannel bioreactors for mammalian and human embryonic stem cell long-term culture. Biomed Microdevices, 2009. 11(1): p. 87-94. 539. Iyer, R.K., et al., Synthetic oxygen carriers in cardiac tissue engineering. Artif Cells Blood Substit Immobil Biotechnol, 2007. 35(1): p. 135-48. 540. Radisic, M., et al., Mathematical model of oxygen distribution in engineered cardiac tissue with parallel channel array perfused with culture medium containing oxygen carriers. Am J Physiol Heart Circ Physiol, 2005. 288(3): p. H1278-89. 541. Chin, K., et al., Hydrogel-perfluorocarbon composite scaffold promotes oxygen transport to immobilized cells. Biotechnol Prog, 2008. 24(2): p. 358-66. 542. Cook, C.A., et al., Oxygen delivery from hyperbarically loaded microtanks extends cell viability in anoxic environments. Biomaterials, 2015. 52: p. 376-384. 543. Asikainen, T.M., et al., Stimulation of HIF-1 alpha, HIF-2 alpha, and VEGF by prolyl 4- hydroxylase inhibition in human lung endothelial and epithelial cells. Free Radical Biology and Medicine, 2005. 38(8): p. 1002-1013. 544. Wu, D.L. and P. Yotnda, Induction and Testing of Hypoxia in Cell Culture. Jove-Journal of Visualized Experiments, 2011(54): p. 4. 545. Park, K.M., et al., Thermosensitive chitosan-Pluronic hydrogel as an injectable cell delivery carrier for cartilage regeneration. Acta Biomaterialia, 2009. 5(6): p. 1956-1965. 546. Park, J.S., et al., In vitro and in vivo test of PEG/PCL-based hydrogel scaffold for cell delivery application. Journal of Controlled Release, 2007. 124(1-2): p. 51-59. 547. Park, H., et al., Injectable biodegradable hydrogel composites for rabbit marrow mesenchymal stem cell and growth factor delivery for cartilage tissue engineering. Biomaterials, 2007. 28(21): p. 3217-3227. 548. Ford, M.C., et al., A macroporous hydrogel for the coculture of neural progenitor and endothelial cells to form functional vascular networks in vivo. Proceedings of the National Academy of Sciences of the United States of America, 2006. 103(8): p. 2512-2517. 549. Kraehenbuehl, T.P., et al., Cell-responsive hydrogel for encapsulation of vascular cells. Biomaterials, 2009. 30(26): p. 4318-4324. 550. Seliktar, D., Designing Cell-Compatible Hydrogels for Biomedical Applications. Science, 2012. 336(6085): p. 1124-1128. 551. Sun, G.M., et al., Dextran hydrogel scaffolds enhance angiogenic responses and promote complete skin regeneration during burn wound healing. Proceedings of the National Academy of Sciences of the United States of America, 2011. 108(52): p. 20976-20981. 552. Wen, J.H., et al., Interplay of matrix stiffness and protein tethering in stem cell differentiation. Nature Materials, 2014. 13(10): p. 979-987. 553. Khetan, S., et al., Degradation-mediated cellular traction directs stem cell fate in covalently crosslinked three-dimensional hydrogels. Nature Materials, 2013. 12(5): p. 458-465. 554. Trappmann, B., et al., Extracellular-matrix tethering regulates stem-cell fate. Nature Materials, 2012. 11(7): p. 642-649. 555. Wang, X., et al., Effect of RGD nanospacing on differentiation of stem cells. Biomaterials, 2013. 34(12): p. 2865-2874. 556. Kusuma, S., et al., Micropattern size-dependent endothelial differentiation from a human induced pluripotent stem cell line. Journal of Tissue Engineering and Regenerative Medicine, 2015: p. n/a-n/a. 557. Dickinson, L.E., et al., Endothelial cell responses to micropillar substrates of varying dimensions and stiffness. Journal of Biomedical Materials Research Part A, 2012. 100A(6): p. 1457-1466.

251

558. Lipski, A.M., et al., The effect of silica nanoparticle-modified surfaces on cell morphology, cytoskeletal organization and function. Biomaterials, 2008. 29(28): p. 3836-3846. 559. Blatchley, M.R. and S. Gerecht, Acellular implantable and injectable hydrogels for vascular regeneration. Biomedical materials (Bristol, England), 2015. 10(3): p. 034001. 560. Shen, Y.I., et al., Hyaluronic acid hydrogel stiffness and oxygen tension affect cancer cell fate and endothelial sprouting. Biomaterials Science, 2014. 2(5): p. 655-665. 561. Levental, K.R., et al., Matrix Crosslinking Forces Tumor Progression by Enhancing Integrin Signaling. Cell, 2009. 139(5): p. 891-906. 562. Wilson, W.R. and M.P. Hay, Targeting hypoxia in cancer therapy. Nature Reviews Cancer, 2011. 11(6): p. 393-410. 563. Keith, B. and M.C. Simon, Hypoxia-inducible factors, stem cells, and cancer. Cell, 2007. 129(3): p. 465-472. 564. Gilkes, D.M., et al., Hypoxia-inducible Factor 1 (HIF-1) Promotes Extracellular Matrix Remodeling under Hypoxic Conditions by Inducing P4HA1, P4HA2, and PLOD2 Expression in Fibroblasts. Journal of Biological Chemistry, 2013. 288(15): p. 10819-10829. 565. Park, K.M., M.R. Blatchley, and S. Gerecht, The Design of Dextran-Based Hypoxia-Inducible Hydrogels via In Situ Oxygen-Consuming Reaction. Macromolecular Rapid Communications, 2014. 35(22): p. 1968-1975. 566. Paguirigan, A.L. and D.J. Beebe, Protocol for the fabrication of enzymatically crosslinked gelatin microchannels for microfluidic cell culture. Nature Protocols, 2007. 2(7): p. 1782-1788. 567. McDermott, M.K., et al., Mechanical properties of biomimetic tissue adhesive based on the microbial transglutaminase-catalyzed crosslinking of gelatin. Biomacromolecules, 2004. 5(4): p. 1270-1279. 568. Wozniak, M.A. and C.S. Chen, Mechanotransduction in development: a growing role for contractility. Nature Reviews Molecular Cell Biology, 2009. 10(1): p. 34-43. 569. Brizel, D.M., et al., Tumor oxygenation predicts for the likelihood of distant metastases in human soft tissue sarcoma. Cancer research, 1996. 56(5): p. 941-943. 570. Mori, R., et al., The relationship between proangiogenic gene expression levels in prostate cancer and their prognostic value for clinical outcomes. The Prostate, 2010. 70(15): p. 1692-1700. 571. Koukourakis, M.I., et al., Hypoxia-inducible factor (HIF1A and HIF2A), angiogenesis, and chemoradiotherapy outcome of squamous cell head-and-neck cancer. International Journal of Radiation Oncology* Biology* Physics, 2002. 53(5): p. 1192-1202. 572. Bertout, J.A., S.A. Patel, and M.C. Simon, The impact of O2 availability on human cancer. Nature Reviews Cancer, 2008. 8(12): p. 967-975. 573. Lewis, D.M., et al., Intratumoral oxygen gradients mediate sarcoma cell invasion. Proceedings of the National Academy of Sciences, 2016: p. 201605317. 574. Chaturvedi, P., et al., Hypoxia-inducible factor–dependent breast cancer–mesenchymal stem cell bidirectional signaling promotes metastasis. The Journal of clinical investigation, 2013. 123(1): p. 189-205. 575. Eisinger-Mathason, T.K., et al., Hypoxia-dependent modification of collagen networks promotes sarcoma metastasis. Cancer discovery, 2013. 3(10): p. 1190-1205. 576. Carmeliet, P., et al., Role of HIF-1α in hypoxia-mediated apoptosis, cell proliferation and tumour angiogenesis. Nature, 1998. 394(6692): p. 485-490. 577. Kvanta, A., et al., Subfoveal fibrovascular membranes in age-related macular degeneration express vascular endothelial growth factor. Investigative ophthalmology & visual science, 1996. 37(9): p. 1929-1934. 578. Lin, C.-H., et al., Silibinin inhibits VEGF secretion and age‐related macular degeneration in a hypoxia‐dependent manner through the PI‐3 kinase/Akt/mTOR pathway. British journal of pharmacology, 2013. 168(4): p. 920-931. 579. Stefánsson, E., Á. Geirsdóttir, and H. Sigurdsson, Metabolic physiology in age related macular degeneration. Progress in retinal and eye research, 2011. 30(1): p. 72-80.

252

580. Lewis, D.M., et al., Endothelial progenitor cell recruitment in a microfluidic vascular model. Biofabrication, 2015. 7(4): p. 045010. 581. Arnould, T., C. Michiels, and J. Remacle, Increased PMN adherence on endothelial cells after hypoxia: involvement of PAF, CD18/CD11b, and ICAM-1. Am J Physiol, 1993. 264(5 Pt 1): p. C1102-10. 582. Luster, A.D., R. Alon, and U.H. von Andrian, Immune cell migration in inflammation: present and future therapeutic targets. Nature immunology, 2005. 6(12): p. 1182-1190. 583. Craig, R., J. Schofield, and S. Jackson, Collagen biosynthesis in normal human skin, normal and hypertrophic scar and keloid. European journal of clinical investigation, 1975. 5(1): p. 69-74. 584. Botusan, I.R., et al., Stabilization of HIF-1α is critical to improve wound healing in diabetic mice. Proceedings of the National Academy of Sciences, 2008. 105(49): p. 19426-19431. 585. Ramirez-Bergeron, D.L. and M.C. Simon, Hypoxia-inducible factor and the development of stem cells of the cardiovascular system. Stem Cells, 2001. 19(4): p. 279-86. 586. Licht, A.H., et al., Inhibition of hypoxia-inducible factor activity in endothelial cells disrupts embryonic cardiovascular development. Blood, 2006. 107(2): p. 584-90. 587. Krock, B.L., N. Skuli, and M.C. Simon, Hypoxia-induced angiogenesis good and evil. Genes & cancer, 2011. 2(12): p. 1117-1133. 588. Kusuma, S., et al., Low oxygen tension enhances endothelial fate of human pluripotent stem cells. Arteriosclerosis, thrombosis, and vascular biology, 2014. 34(4): p. 913-920. 589. Liu, Y., et al., Hypoxia regulates vascular endothelial growth factor gene expression in endothelial cells identification of a 5′ enhancer. Circulation research, 1995. 77(3): p. 638-643. 590. Covello, K.L. and M.C. Simon, HIFs, hypoxia, and vascular development. Curr Top Dev Biol, 2004. 62: p. 37-54. 591. Pugh, C.W. and P.J. Ratcliffe, Regulation of angiogenesis by hypoxia: role of the HIF system. Nat Med, 2003. 9(6): p. 677-84. 592. Marti, H.J., et al., Hypoxia-induced vascular endothelial growth factor expression precedes neovascularization after cerebral ischemia. Am J Pathol, 2000. 156(3): p. 965-76. 593. Oppegard, S.C., et al., Modulating Temporal and Spatial Oxygenation over Adherent Cellular Cultures. Plos One, 2009. 4(9): p. 8. 594. Rexius-Hall, M.L., et al., Microfluidic platform generates oxygen landscapes for localized hypoxic activation. Lab on a Chip, 2014. 14(24): p. 4688-4695. 595. DelNero, P., et al., 3D culture broadly regulates tumor cell hypoxia response and angiogenesis via pro-inflammatory pathways. Biomaterials, 2015. 55: p. 110-118. 596. Funamoto, K., et al., A novel microfluidic platform for high-resolution imaging of a three- dimensional cell culture under a controlled hypoxic environment. Lab on a Chip, 2012. 12(22): p. 4855-4863. 597. Wu, D. and P. Yotnda, Induction and testing of hypoxia in cell culture. JoVE (Journal of Visualized Experiments), 2011(54): p. e2899-e2899. 598. Lin, X., et al., Oxygen-induced cell migration and on-line monitoring biomarkers modulation of cervical cancers on a microfluidic system. Scientific reports, 2015. 5: p. 9643. 599. Allen, C.B., B.K. Schneider, and C.W. White, Limitations to oxygen diffusion and equilibration in in vitro cell exposure systems in hyperoxia and hypoxia. American Journal of Physiology-Lung Cellular and Molecular Physiology, 2001. 281(4): p. L1021-L1027. 600. Park, K.M. and S. Gerecht, Hypoxia-inducible hydrogels. Nat Commun, 2014. 5. 601. Davis, G.E., K.J. Bayless, and A. Mavila, Molecular basis of endothelial cell morphogenesis in three-dimensional extracellular matrices. Anatomical Record, 2002. 268(3): p. 252-275. 602. Chan, X.Y., et al., Three-Dimensional Vascular Network Assembly From Diabetic Patient- Derived Induced Pluripotent Stem Cells. Arteriosclerosis, Thrombosis, and Vascular Biology, 2015. 603. Higuchi, A., et al., Biomimetic Cell Culture Proteins as Extracellular Matrices for Stem Cell Differentiation. Chemical Reviews, 2012. 112(8): p. 4507-4540.

253

604. Loessner, D., et al., Functionalization, preparation and use of cell-laden gelatin methacryloyl- based hydrogels as modular tissue culture platforms. Nature Protocols, 2016. 11(4): p. 727-746. 605. Minussi, R.C., G.M. Pastore, and N. Duran, Potential applications of laccase in the food industry. Trends in Food Science & Technology, 2002. 13(6-7): p. 205-216. 606. Park, K.M. and S. Gerecht, Hypoxia-inducible hydrogels. Nature Communications, 2014. 5(May): p. 1-12. 607. Mozaffarian, Heart Disease and Stroke Statistics-2016 Update: A Report From the American Heart Association (vol 133, pg e38, 2016). Circulation, 2016. 133(15): p. E599-E599. 608. Semenza, G.L., Vascular Responses to Hypoxia and Ischemia. Arteriosclerosis Thrombosis and Vascular Biology, 2010. 30(4): p. 648-652. 609. Ferkowicz, M.J. and M.C. Yoder, Blood island formation: longstanding observations and modern interpretations. Experimental Hematology, 2005. 33(9): p. 1041-1047. 610. McGrath, K.E., et al., Circulation is established in a stepwise pattern in the mammalian embryo. Blood, 2003. 101(5): p. 1669-1676. 611. Francis, S.E., et al., Central roles of alpha(5)beta(1) integrin and fibronectin in vascular development in mouse embryos and embryoid bodies. Arteriosclerosis Thrombosis and Vascular Biology, 2002. 22(6): p. 927-933. 612. Ellertsdottir, E., et al., Vascular morphogenesis in the zebrafish embryo. Developmental Biology, 2010. 341(1): p. 56-65. 613. Lawson, N.D. and B.M. Weinstein, In vivo imaging of embryonic vascular development using transgenic zebrafish. Developmental Biology, 2002. 248(2): p. 307-318. 614. Blinder, Y.J., et al., Vasculogenic dynamics in 3D engineered tissue constructs. Scientific Reports, 2015. 5: p. 8. 615. Davis, G.E. and K.J. Bayless, An integrin and rho GTPase-dependent pinocytic vacuole mechanism controls capillary lumen formation in collagen and fibrin matrices. Microcirculation, 2003. 10(1): p. 27-44. 616. Sieminski, A.L., et al., The stiffness of three-dimensional ionic self-assembling peptide gels affects the extent of capillary-like network formation. Cell Biochemistry and Biophysics, 2007. 49(2): p. 73-83. 617. Proulx, K., A. Lu, and S. Sumanas, Cranial vasculature in zebrafish forms by angioblast cluster- derived angiogenesis. Developmental Biology, 2010. 348(1): p. 34-46. 618. Tepper, O.M., et al., Adult vasculogenesis occurs through in situ recruitment, proliferation, and tubulization of circulating bone marrow-derived cells. Blood, 2005. 105(3): p. 1068-1077. 619. Vajkoczy, P., et al., Multistep nature of microvascular recruitment of Ex vivo-expanded embryonic endothelial progenitor cells during tumor angiogenesis. Journal of Experimental Medicine, 2003. 197(12): p. 1755-1765. 620. Ceradini, D.J. and G.C. Gurtner, Homing to hypoxia: HIF-1 as a mediator of progenitor cell recruitment to injured tissue. Trends in Cardiovascular Medicine, 2005. 15(2): p. 57-63. 621. Hirota, K. and G.L. Semenza, Regulation of angiogenesis by hypoxia-inducible factor 1. Critical Reviews in Oncology Hematology, 2006. 59(1): p. 15-26. 622. Quintero, M., et al., Mitochondria as signaling organelles in the vascular endothelium. Proceedings of the National Academy of Sciences of the United States of America, 2006. 103(14): p. 5379-5384. 623. Majmundar, A.J., W.H.J. Wong, and M.C. Simon, Hypoxia-Inducible Factors and the Response to Hypoxic Stress. Molecular Cell, 2010. 40(2): p. 294-309. 624. Taniyama, Y. and K.K. Griendling, Reactive oxygen species in the vasculature - Molecular and cellular mechanisms. Hypertension, 2003. 42(6): p. 1075-1081. 625. Rey, S. and G.L. Semenza, Hypoxia-inducible factor-1-dependent mechanisms of vascularization and vascular remodelling. Cardiovascular Research, 2010. 86(2): p. 236-242.

254

626. Jedeszko, C., et al., Visualizing Protease Activity in Living Cells: From Two Dimensions to Four Dimensions. Current protocols in cell biology / editorial board, Juan S. Bonifacino ... [et al.], 2008. 0 4: p. Unit-4.20. 627. van Hinsbergh, V.W.M., M.A. Engelse, and P.H.A. Quax, Pericellular proteases in angiogenesis and vasculogenesis. Arteriosclerosis Thrombosis and Vascular Biology, 2006. 26(4): p. 716-728. 628. Deem, T.L. and J.M. Cook-Mills, Vascular cell adhesion molecule 1 (VCAM-1) activation of endothelial cell matrix metalloproteinases: role of reactive oxygen species. Blood, 2004. 104(8): p. 2385-2393. 629. Hire, J.M., et al., Variance of matrix metalloproteinase (MMP) and tissue inhibitor of metalloproteinase (TIMP) concentrations in activated, concentrated platelets from healthy male donors. Journal of Orthopaedic Surgery and Research, 2014. 9: p. 6. 630. Maeda, S., et al., DETERMINATION OF INTERSTITIAL COLLAGENASE (MMP-1) IN PATIENTS WITH RHEUMATOID-ARTHRITIS. Annals of the Rheumatic Diseases, 1995. 54(12): p. 970-975. 631. Rodrigues, M., et al., VEGF Secreted by Hypoxic Muller Cells Induces MMP-2 Expression and Activity in Endothelial Cells to Promote Retinal Neovascularization in Proliferative Diabetic Retinopathy. Diabetes, 2013. 62(11): p. 3863-3873. 632. Kudo, Y., et al., Matrix Metalloproteinase-13 (MMP-13) Directly and Indirectly Promotes Tumor Angiogenesis. Journal of Biological Chemistry, 2012. 287(46): p. 38716-38728. 633. Lai, S.L., et al., In Vitro and In Vivo Anti-Angiogenic Activities of Panduratin A. Plos One, 2012. 7(5): p. 14. 634. Gory-Faure, S., et al., Role of vascular endothelial-cadherin in vascular morphogenesis. Development, 1999. 126(10): p. 2093-2102. 635. Wang, Y., et al., Moesin1 and Ve-cadherin are required in endothelial cells during in vivo tubulogenesis. Development, 2010. 137(18): p. 3119-3128. 636. Dejana, E., Endothelial cell-cell junctions: Happy together. Nature Reviews Molecular Cell Biology, 2004. 5(4): p. 261-270. 637. Hynes, R.O., Cell-matrix adhesion in vascular development. Journal of Thrombosis and Haemostasis, 2007. 5: p. 32-40. 638. Asahara, T., et al., Isolation of putative progenitor endothelial cells for angiogenesis. Science, 1997. 275(5302): p. 964-967. 639. Cai, H., Hydrogen peroxide regulation of endothelial function: Origins, mechanisms, and consequences. Cardiovascular Research, 2005. 68(1): p. 26-36. 640. Beckham, T.H., et al., Acid ceramidase-mediated production of sphingosine 1-phosphate promotes prostate cancer invasion through upregulation of cathepsin B. International Journal of Cancer, 2012. 131(9): p. 2034-2043. 641. Mai, J., et al., Degradation of extracellular matrix protein tenascin-C by cathepsin B: an interaction involved in the progression of gliomas. Biol Chem, 2002. 383(9): p. 1407-13. 642. Sugano, E., et al., Overexpression of acid ceramidase (ASAH1) protects retinal cells (ARPE19) from oxidative stress. J Lipid Res, 2019. 60(1): p. 30-43. 643. Platt, M.O. and W.A. Shockey, Endothelial cells and cathepsins: Biochemical and biomechanical regulation. Biochimie, 2016. 122: p. 314-323. 644. Mahmood, N., C. Mihalcioiu, and S.A. Rabbani, Multifaceted Role of the Urokinase-Type Plasminogen Activator (uPA) and Its Receptor (uPAR): Diagnostic, Prognostic, and Therapeutic Applications. Frontiers in oncology, 2018. 8: p. 24-24. 645. Kiyono, M. and M. Shibuya, Inhibitory Smad transcription factors protect arterial endothelial cells from apoptosis induced by BMP4. Oncogene, 2006. 25(54): p. 7131-7137. 646. Ito, Y.A., et al., Dysfunction of the stress-responsive FOXC1 transcription factor contributes to the earlier-onset glaucoma observed in Axenfeld-Rieger syndrome patients. Cell death & disease, 2014. 5(2): p. e1069-e1069.

255

647. Walmsley, S.R., et al., Prolyl hydroxylase 3 (PHD3) is essential for hypoxic regulation of neutrophilic inflammation in humans and mice. J Clin Invest, 2011. 121(3): p. 1053-63. 648. Draoui, N., P. de Zeeuw, and P. Carmeliet, Angiogenesis revisited from a metabolic perspective: role and therapeutic implications of endothelial cell metabolism. Open Biology, 2017. 7(12): p. 9. 649. Wong, B.W., et al., Endothelial cell metabolism in health and disease: impact of hypoxia. Embo Journal, 2017. 36(15): p. 2187-2203. 650. Fitzgerald, G., I. Soro-Arnaiz, and K. De Bock, The Warburg Effect in Endothelial Cells and its Potential as an Anti-angiogenic Target in Cancer. Frontiers in cell and developmental biology, 2018. 6: p. 100-100. 651. Shen, Y.-I., et al., Acellular Hydrogels for Regenerative Burn Wound Healing: Translation from a Porcine Model. Journal of Investigative Dermatology, 2015. 135(10): p. 2519-2529. 652. Griffin, D.R., et al., Accelerated wound healing by injectable microporous gel scaffolds assembled from annealed building blocks. Nature Materials, 2015. 14(7): p. 737-+. 653. Seifert, A.W. and K. Muneoka, The blastema and epimorphic regeneration in mammals. Developmental Biology, 2018. 433(2): p. 190-199. 654. Bonham, C.A., et al., Deferoxamine can prevent pressure ulcers and accelerate healing in aged mice. Wound Repair and Regeneration. 0(0).

256

CURRICULUM VITAE FOR Ph.D. CANDIDATES The Johns Hopkins University School of Medicine

______Michael Rowland Blatchley______7/25/2019______Name Date of this Version

EDUCATIONAL HISTORY Ph.D. expected 2019 Program in Biomedical Engineering Johns Hopkins School of Medicine

Mentor: Sharon Gerecht, PhD

B.S. 2013 Biomedical Engineering Purdue University

RESEARCH/PROFESSIONAL EXPERIENCE

PhD Research Spring 2014 - present Johns Hopkins University, Baltimore, MD Principle Investigator: Dr. Sharon Gerecht - Developed methods to independently modulate oxygen and matrix stiffness in hydrogel biomaterials. - Utilized these hydrogels to uncover the basis for a mechanistic understanding of endothelial cluster-based vasculogenesis. Currently investigating more in-depth details of this mechanism to identify potential therapeutic targets to either promote or inhibit vascular network formation for treatment of cardiovascular disease and tumor angiogenesis, respectively. - Currently investigating the therapeutic effect of oxygen concentration within hydrogels in a murine wound healing model.

PhD Lab Rotation Fall 2013 Johns Hopkins University, Baltimore, MD Principle Investigator: Dr. Jennifer Elisseeff - Worked with a post doc to develop a novel high throughput extracellular matrix (ECM) microarray chip for testing properties of ECM particles from a variety of tissue sources

NIH Funded Clinical Internship Summer 2012 IU School of Medicine, Indianapolis, IN Principle Investigator: Dr. Irina Petrache - Worked with clinicians, graduate students, and clinical fellows on a project in which the objective was to determine the effects of bone marrow conditioned media and cigarette smoke on mouse endothelial lung cells (MLECs)

257

HHMI Undergraduate Research Experience Summer 2011 Purdue University, West Lafayette, IN Principle Investigator: Dr. Clint Chapple - Investigated how mutations in the Mediator complex alter the expression of phenylpropanoid biosynthetic genes in Arabidopsis thaliana

Research Intern, Akina, Inc. Summer 2010 Akina, Inc., West Lafayette, IN Principle Investigator: Dr. Kinam Park, Manager: John Garner - Worked with a team of research interns on several contract research projects including a re- shapeable tissue expanding hydrogel and a biodegradable ureteral stent

Research Assistant, Civil Engineering Summer 2008 Purdue University, West Lafayette, IN Principle Investigator: Dr. Ernest Rowland Blatchley III - Partnered with a CE graduate student on a project in which the objective was to define volatile disinfection byproduct (DBP) concentrations in chlorinated public pools using membrane introduction mass spectrometry and DPD/KI spectrophotometry

AWARDS AND HONORS

• Young Investigator’s Day Awardee (Paul Talalay Award), Johns Hopkins University School of Medicine (2019) • BME Student Seminar Series, 2nd Place Seminar, Johns Hopkins University Biomedical Engineering (2018) • Ruth L. Kirschstein National Research Service Award (NRSA) Individual Predoctoral Fellowship (F31) (2017-2020); stipend and research support • American Heart Association Predoctoral Fellowship (2017-2019; awarded, declined) • 1st Place Poster Award, Society for Biological Engineering, International Conference on Stem Cell Engineering (2016) • NSF Graduate Research Fellowship Program Honorable Mention (2014) • NIH-Funded Clinical Research Intern (2012) • Howard Hughes Medical Institute Undergraduate Research Experience (2011) • Second Team American Collegiate Rowing Association Academic All-American (2010, 2011, 2012, 2013) • Dean’s List and Semester Honors (Fall 2009-2012, Spring 2010-2013) • Hoosier Scholar Award (2009) • Presidential Scholarship (2009-2013)

PUBLICATIONS 1. Wei, Z.; Volkova E.; Blatchley, M.R.; Gerecht, S. “Hydrogel vehicles for sequential delivery of protein drugs to promote vascular regeneration,” Advanced Drug Delivery Reviews, invited, in revision. 2. Blatchley, M.R.; Gerecht, S. “Re-constructing the vascular developmental milieu in vitro,” submitted.

258

3. Xiao, Y.; Chang, L.; Chen, Z.; Blatchley, M.R.; Zhou, J.; Xu, M. Gerecht, S.; Fan, R. (2019) “Senescent cells with augmented cytokine production for microvascular bioengineering and tissue repairs,” Advanced Biosystems, 1900089. 4. Blatchley, M.R.; Hall, F.; Wang, S; Pruitt, H.C.; Gerecht, S. (2019) “Hypoxia and matrix viscoelasticity sequentially regulate endothelial progenitor cluster-based vasculogenesis,” Science Advances, 5 (3), eaau7518. 5. Blatchley, M.R.; Abaci, H.E.; Hanjaya-Putra, D.; Gerecht, S. (2018) “Hypoxia and Matrix Manipulation for Vascular Engineering,” Biophysical Regulation of Vascular Differentiation and Assembly, Ch. 4, 73-119. 6. Cho, H.; Blatchley, M.R.; Duh, E.J.; Gerecht, S. (2018) “Acellular and cellular approaches to improve diabetic wound healing,” Advanced Drug Delivery Reviews. 7. Blatchley, M.R.; Gerecht, S. (2017) “Integrin binding: Sticking around vessels,” Nature Materials, 16 (9), 881-883.

8. Lewis, D.M.*; Blatchley, M.R.*; Park, K.M.; Gerecht, S. (2017) “O2-controllable hydrogels for studying cellular responses to hypoxic gradients in three dimensions in vitro and in vivo,” Nature Protocols, 12 (8), 1620-1638. 9. Smith, Q; Blatchley, M.R.; Gerecht, S. (2017) “Engineering niches for blood vessel regeneration,” Biology and Engineering of Stem Cell Niches, Ch. 30, 479-497. 10. Beachley, V.Z.; Wolf, M.T.; Sadtler, K.; Manda, S.S.; Jacobs, H.; Blatchley, M.R.; Bader, J.S.; Pandey, A.; Pardoll, D; Elisseeff, J.H. (2015) “Tissue matrix arrays for high-throughput screening and systems analysis of cell function,” Nature Methods, 12, 1197-1204. 11. Blatchley, M.R.; Park, K.M; Gerecht, S. (2015) “Designer hydrogels for precision control of oxygen tension and mechanical properties,” Journal of Materials Chemistry B, 3, 40, 7939-7949. 12. Blatchley, M.R.; Gerecht, S. (2015) “Acellular implantable and injectable hydrogels for vascular regeneration,” Biomedical Materials, 10, 3, 034001. 13. Park, K.M.; Blatchley, M.R.; Gerecht, S. (2014) “The design of dextran-based hypoxia-inducible hydrogels via in situ oxygen consuming reaction,” Macromolecular Rapid Communications, 35, 22, 1968-1975. 14. Bonawitz, N.D.; Soltau, W.L.; Blatchley, M.R.; Powers, B.L., Hurlock, A.K.; Seals, L.A.; Weng, J.; Stout, J.; Chapple, C. (2012) “The REF4 and RFR1 subunits of the eukaryotic transcriptional coregulatory complex Mediator are required for phenylpropanoid homeostasis in Arabidopsis,” The Journal of Biological Chemistry, 287, 8, 5434-5445. 15. Weaver, W.A.; Li, J.; Wen, Y.L.; Johnston, J.; Blatchley, M.R.; Blatchley III, E.R. (2009) “Volatile disinfection by-product analysis from chlorinated indoor swimming pools,” Water Research, 43, 13, 3308-3318. *Indicates authors contributed equally.

ORAL PRESENTATIONS 1. Blatchley, M.R.; Gerecht, S. (2019) “Re-constructing the vascular regenerative niche to investigate blood vessel formation,” Young Investigators’ Day Ceremony Lecture, Baltimore, MD.

259

2. Blatchley, M.R.; Hall, F.; Wang, S; Pruitt, H.C.; Gerecht, S. (2018) “Hypoxia and matrix viscoelasticity sequentially regulate endothelial progenitor cluster-based vasculogenesis,” BME Department Seminar, Baltimore, MD. 3. Blatchley, M.R.; Gerecht, S. (2018) “Designing Hydrogels to Guide Cellular Responses During Development, Vascular Morphogenesis and Tumor Growth,” The 7th International Symposium of Green MAP Center and LPIC, Yamagata University, Yonezawa, Japan. 4. Blatchley, M.R.; Wang S.; Hall, F.; Gerecht, S. (2017) “Oxygen-controllable hydrogels to study hypoxic, cluster-based vasculogenesis,” Biomedical Engineering Society Annual Meeting, Phoenix, AZ. 5. Blatchley M.R.; Park, K.M.; Gerecht, S. (2015) "Designer Polymeric Hydrogels for Independent Control of Oxygen Tension and Mechanical Properties," Materials Research Society Fall Meeting, Boston, MA. POSTER PRESENTATIONS 1. Blatchley, M.R.; Hall, F.; Wang, S.; Pruitt, H.C.; Gerecht, S. (2019) “Re-constructing the vascular regenerative niche to investigate blood vessel formation,” Society for Biological Engineering, International Conference on Bioengineering and Nanotechnology, Baltimore, MD. 2. Blatchley, M.R.; Hall, F.; Wang, S.; Pruitt, H.C.; Gerecht, S. (2019) “Re-constructing the vascular regenerative niche to investigate blood vessel formation,” INBT Translation of Nano and Bio Research Symposium, Baltimore, MD. 3. Blatchley, M.R.; Hall, F.; Wang, S.; Pruitt, H.C.; Gerecht, S. (2019) “Hypoxia and matrix viscoelasticity sequentially regulate endothelial progenitor cluster-based vasculogenesis,” Society for Biomaterials Annual Meeting, Seattle, WA. Poster accepted as ‘late breaking’ submission. 4. Blatchley, M.R.; Hall, F.; Wang, S.; Pruitt, H.C.; Gerecht, S. (2018) “Hypoxia regulates cluster-based vasculogenesis,” Biomedical Engineering Society Annual Meeting, Atlanta, GA. 5. Blatchley, M.R.; Hall, F.; Wang, S.; Pruitt, H.C.; Gerecht, S. (2018) “Hypoxia regulates cluster-based vasculogenesis,” Gordon Research Seminar & Conference: Signal Transduction by Engineered Extracellular Matrices, Andover, NH. 6. Blatchley, M.R.; Hall, F.; Wang, S.; Pruitt, H.C.; Gerecht, S. (2018) “Hypoxia regulates cluster-based vasculogenesis,” INBT Advanced Biomanufacturing Symposium, Baltimore, MD. 7. Blatchley, M.R.; Hall, F.; Wang, S.; Pruitt, H.C.; Gerecht, S. (2018) “Hypoxia regulates cluster-based vasculogenesis,” Johns Hopkins Cardiovascular Research Retreat, Baltimore, MD. 8. Blatchley, M.R.; Wang, S.; Hall, F.; Gerecht, S. (2017) “Elucidating a mechanism for hypoxic, cluster- based vasculogenesis,” INBT Vascularization Symposium, Baltimore, MD. 9. Blatchley, M.R.; Wang, S.; Hall, F.; Gerecht, S. (2016) “Elucidating a mechanism for cluster vasculogenesis using in vitro engineered hydrogels,” Society for Biological Engineering, International Conference on Stem Cell Engineering, Toronto, ON. 10. Blatchley, M.R.; Wang, S.; Gerecht, S. (2016) “Hypoxia-Inducible Hydrogels to Study Vascular Morphogenesis,” INBT Precision Medicine Symposium, Baltimore, MD. 11. Blatchley, M.R.; Schweitzer,K; Petrache, I. (2012) “Evidence of a Paracrine Crosstalk Between the Bone Marrow and the Lung,” SEED Poster Session, Indianapolis, IN. 12. Blatchley, M.R.; Bonawitz, N.D.; Soltau, W.L.; Chapple, C. (2011) “Mutations in the Mediator Complex Alter the Expression of Phenylpropanoid Biosynthetic Genes in Arabidopsis thaliana,” Howard Hughes Medical Institute Summer Undergraduate Research Poster Session. West Lafayette, IN.

260

TEACHING EXPERIENCE

Teaching assistant, Tissue Engineering Spring 2017 Teaching assistant, BME Modeling and Design Fall 2016 Guest lecturer, Bioengineering in Regenerative Medicine (Biomaterials lecture) Spring 2016

MENTORING

Graduate students • Franklyn Hall, BME PhD student, 2017-present • Lin Lu, MS Chemical and Biomolecular Engineering, Alternative laccase sources for crosslinking hypoxia-inducible hydrogels, 2015-2016 Undergraduate students • JHU: Vidur Kailash, Biophysics undergraduate student, 2016-present • JHU: Songnan Wang, BS Chemical and Biomolecular Engineering; Provost’s Undergraduate Research Award recipient, 2015-2017; current position Research Associate Gladstone Institutes/UCSF • JHU: Arianne Papa, BS Biomedical Engineering, 2016-2017; current position PhD student Columbia University • REU: Franklyn Hall, Mississippi State University; Summers 2015, 2016; current position PhD student Johns Hopkins University • BME graduate student mentoring group: New organization developed by graduate students to mentor undergraduates and provide information about applying to graduate school and succeeding in graduate school High school student • Brooke Smith, Baltimore Polytechnic Institute, Women Serious About Science Program, 2014- 2015 Laboratory technician • Michael Cho, laboratory technician, 2018-present; will be a PhD student at Johns Hopkins University, Fall 2019

OUTREACH AND LEADERSHIP

Organizer, Gerecht Lab Journal Club 2017 - present

Spearheaded the inception of a bi-monthly journal club for the members of the Gerecht lab, and serve as the discussion leader. Lab members discuss recent papers relating to biomaterials, stem cell biology, cancer biology, and cardiovascular tissue engineering with accompanying background information.

Biomedical Engineering PhD Council President 2014 - 2015

The BME PhD Council aims to foster an environment within the BME department at Johns Hopkins, in which students have open opportunities to discuss their research through student-run seminars and get involved with outreach programs. The PhD council also aims to solve inter-departmental issues, organize

261 social events, and put together discussion panels to help prepare fellowship applications and prepare for qualifying exams.

Science Outreach Program Summer 2014

The Science Outreach Program (SOP) is comprised of graduate students in the Johns Hopkins University School of Medicine. I developed lesson plans for a summer science program at a Baltimore City community center for elementary school aged children. The main goal of SOP is to get the kids excited about science through interactive learning and hands on activities.

Thread Mentoring Program 2013 - 2018

The Thread Mentoring Program has a significant presence within three Baltimore City high schools with the goal to provide both academic help and positive role models for underachieving students. Mentors tutor the high school students and help the students set goals including high school graduation and college acceptance. Mentors also work to provide the students with a positive life influence and encourage participation in productive activities outside of school.

Alpha Eta Mu Beta (The Biomedical Engineering Honor Society) 2011 - 2013 Tau Beta Pi (The Engineering Honor Society) 2010 - 2013 Purdue Rowing Team 2009 - 2013

262