Engineering 3D Systems with Tunable Mechanical Properties to Mimic the Tumor Microenvironment Shalini Raj Unnikandam Veettil Iowa State University

Iowa State University Capstones, Theses and Graduate Theses and Dissertations Dissertations

2018 Engineering 3D systems with tunable mechanical properties to mimic the tumor microenvironment Shalini Raj Unnikandam Veettil Iowa State University

Follow this and additional works at: https://lib.dr.iastate.edu/etd Part of the Chemical Engineering Commons

Recommended Citation Unnikandam Veettil, Shalini Raj, "Engineering 3D systems with tunable mechanical properties to mimic the tumor microenvironment" (2018). Graduate Theses and Dissertations. 17339. https://lib.dr.iastate.edu/etd/17339

This Thesis is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected].

Engineering 3D systems with tunable mechanical properties to mimic the tumor microenvironment

Shalini Raj Unnikandam Veettil

A thesis submitted to the graduate faculty

in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

Major: Chemical Engineering

Program of Study Committee: Ian C Schneider, Major Professor Kaitlin Bratlie Michael Bartlett

The student author, whose presentation of the scholarship herein was approved by the program of study committee, is solely responsible for the content of this thesis. The Graduate College will ensure this thesis is globally accessible and will not permit alterations after a degree is conferred.

Iowa State University

Ames, Iowa

2018

DEDICATION

This thesis is dedicated to my family and friends who have been a great source of support. iii

TABLE OF CONTENTS

Page

ABBREVIATIONS ...... vi

ACKNOWLEDGMENTS ...... vii

ABSTRACT ...... viii

CHAPTER 1. INTRODUCTION ...... 1 1.1Motivation ...... 1 1.2 Background and Introduction ...... 2 1.2.1 Metastasis requires directed migration ...... 2 1.2.2 Cells produce, remodel and degrade extracellular matrix ...... 3 1.2.3 Cell - ECM interactions and regulation of cellular responses ...... 4 1.2.4 Techniques used to tune stiffness ...... 6 1.2.5 Cell morphology and migration in 2D systems with tunable mechanical properties ...... 8 1.2.6 3D systems with tunable mechanical properties ...... 9 1.2.7 Durotaxis ...... 11

CHAPTER 2. RESEARCH OBJECTIVES ...... 13 2.1 Develop and Characterize a TME Mimicking 3D System with Tunable Mechanical Properties ...... 13 2.2 Evaluate How Cell Migration and Morphology Vary in 3D as a Function of Gel Thickness and Surface Chemistry Linking the Soft and Stiff Interface ...... 14 2.3 Examine if Durotactic Responses Could Be Elicited in the Surface Tunable 3D Systems ...... 15

CHAPTER 3. TUNING SURFACE FUNCTIONALIZATION AND COLLAGEN GEL THICKNESS TO REGULATE CANCER CELL MIGRATION ...... 16 3.1 Abstract ...... 16 3.2 Highlights ...... 17 3.3 Introduction ...... 18 3.4 Material and Methods ...... 21 3.4.1 Surface modifications generating high and low collagen binding surfaces ...... 21 3.4.2 Characterization of amine density on glass surfaces ...... 21 3.4.3 Contact angle measurement ...... 22 3.4.4 Adhesion characterization ...... 22 3.4.5 Cell culture and characterization of cell morphology on 2D functionalized glass surfaces ...... 23 3.4.6 Preparation of collagen gel in 3D chambers ...... 24 3.4.7 Fixing and staining of cells ...... 25 3.4.8 Data Analysis ...... 25 3.4.9 Statistical analysis ...... 26

3.5 Results ...... 27 3.5.1 Surface modification of glass generating high and low collagen binding surfaces ...... 27 3.5.2 Cell spreading on 2D functionalized glass surfaces ...... 28 3.5.3 Adhesion of collagen to functionalized glass surfaces ...... 28 3.5.4 Quantitative cell morphology in 3D collagen chambers with different glass functionalization ...... 29 3.5.5 Cell motility in 3D collagen chambers with different glass functionalization ..31 3.6 Discussion ...... 33 3.7 Conclusions ...... 37 3.8 Acknowledgements ...... 38 3.9 Competing Interests ...... 38 3.10 Contributions ...... 38 3.11 References ...... 39 3.12 Figure Legends ...... 44 4.1 Investigating the Biophysical Effects of Hyaluronan Molecular Weight and Concentration on Breast Cancer Cell Migration in Collagen Matrices ...... 54 4.2 Investigating the Influence of Surface Topographical Cues to Elicit Durotaxis in 3D Systems Among Breast Cancer Cells ...... 56

REFERENCES ...... 59

APPENDIX. MATLAB CODE ...... 68

LIST OF FIGURES

Page

Figure 1 Characterization of surface properties on functionalized glass surfaces ...... 44

Figure 2 Characterization of cell spreading on functionalized glass surfaces...... 45

Figure 3 Characterization of adhesion of collagen to functionalized glass surfaces...... 46

Figure 4 Characterization of cell morphology in 3D collagen chambers of different thicknesses and glass functionalization...... 47

Figure 5 Characterization of cell shape in 3D collagen chambers of different thicknesses and glass functionalization ...... 49

Figure 6 Motility characterization in 3D collagen chambers with functionalized glass surfaces...... 50

Figure 7 Motility characterization in 3D collagen chambers with step change in thickness ...52

Figure 8 The structure of hyaluronan consisting of repeating disaccharides of glucuronic acid and N-acetylglucosamine ...... 56

Figure 9 Schematic of pyramidal 3D structure used to provide soft-stiff interfaces (scale in mm) ...... 58

ABBREVIATIONS

TME Tumor Microenvironment

ECM Extracellular Matrix

2D Two-Dimensional

3D Three-Dimensional

DI Directionality Index

HA Hyaluronan

MMP Matrix metalloproteinase

vii

ACKNOWLEDGMENTS

I would like to thank Prof. Ian C Schneider for allowing me to work under his supervision. This thesis would not have been possible without his support, guidance and encouragement. I would like to thank my committee members, Prof. Michal Bartlett, and Prof.

Kaitlin Bratlie for their immense help throughout the course of this research. I am extremely thankful to my lab members Jacob Nuhn and Juan Wang for mentoring me. I would also like to thank my collaborator Dohgyu Hwang for his help and support in carrying out some of the studies.

In addition, I thank my family for their endless support. I would also like to thank my friends, colleagues, the department faculty and staff at Iowa State University. viii

ABSTRACT

The extracellular matrix (ECM) network provides biophysical cues that regulate many physiologically relevant cellular functions, particularly cell migration. Cells are known to sense and respond to biophysical cues including mechanical and structural properties of the tumor microenvironment (TME). This includes understanding how ECM stiffness, in general or when presented in spatial gradients alters cell behavior. The latter is critical given that cell migration can be directed along gradients of stiffness in a process called durotaxis or mechanotaxis.

Durotaxis has been implicated in contributing to cell migration that governs cancer invasion.

Research probing the influence of ECM stiffness in controlling cancer cell migration has mostly been conducted by tuning the bulk elastic modulus of the 2D substrates. Soft-stiff interfaces have also been used to control mechanical properties since cells have the ability to sense the interfaces as far as hundreds of microns away from them. This approach can be used to both create uniform as well as spatial gradients of stiffnesses. Soft-stiff interfaces are useful in manipulating the effective local stiffness without altering the bulk properties. In addition, they allow non-expert end users to generate gradients of stiffness, because the surface, which is manufactured before introduction of the ECM and cells, controls the stiffness gradients and not some patterning process that occurs co-temporal to the introduction of cells. Much work has been carried out in

2D to study cellular responses to stiffness cues leveraging the interactions at the soft-stiff interfaces. However, 3D in vitro models mimicking TME are becoming a standard for studying cell behavior and how stiff-soft interfaces operate in 3D environments is unknown. Finally, there is an opportunity to tune the stiff-soft interfacial interaction through surface chemistry, but the design principles associated with this surface chemistry is currently unknown. In this thesis, I present a novel method to tune the mechanical properties of the 3D microenvironment by ix leveraging the interactions at the soft-stiff interface. We examine the migration of cells embedded in a collagen gel sandwiched between two stiff substrates (glass). We study how surface attachment of collagen to the stiff substrate and the thickness of the collagen fiber network influence the cell migration speed and cell morphology. The findings of this study suggest that these perturbations result in different cell responses. Cell migration appears to be controlled by the gel thickness, whereas cell morphology is tuned by the surface chemistry at the stiff-soft interface and the gel thickness. Our study also suggests that durotaxis can be elicited by leveraging the thickness of the gels. In the future, we will design 3D structures to elicit durotactic responses using soft-stiff interfaces. 1

CHAPTER 1. INTRODUCTION

1.1Motivation

Cancer is regarded as a key health issue globally, causing high patient morbidity and mortality rates and huge economic loss. This disease is a condition in which some of the cells uncontrollably divide and invade the surrounding tissues. In 2015, one in six deaths was caused by cancer and the economic impact of cancer was estimated to be 1.16 trillion USD in 2010[1,2].

Cancer patients are prone to develop metastasis. Metastasis is caused by a multi-step biological process, known as the invasion-metastasis cascade which results in the transformation of a local tumor into a lethal disease. In metastasis, the epithelial cells from a primary tumor invade into the extracellular matrix (ECM) and intravasate into the blood/lymph vessels where the cells survive in circulation. At a distant site, they extravasate into a new tissue environment and form micro-metastases. They adapt and colonize in the foreign microenvironment by re-initiating proliferation, thereby forming detectable macro-metastases [3–7]. Metastatic ability can lead to a rather incurable disease for various cancer types. Blocking invasion and metastasis is a good therapeutic strategy to cure cancer, since these are important steps in cancer progression.

Metastasis causes >90% of mortality due to cancer, whereas a confined primary cancer is often cured by a surgery or irradiation. Although there are biological mechanisms to ensure normal cell function and apoptosis, cancer overcomes these natural barriers through the evolutionary selection of traits, a process deemed to be one of the hallmarks of metastasis [8]. Despite the large amount of work that has been conducted on cancer invasion, there are still questions to be answered pertaining to how cells carry out this process, particularly with respect to the biophysical cues as they are presented in the TME. For instance, while gradients in stiffness are known to exist in the TME and stiff-soft interfaces are present there as well, particularly in 2 malignant tumors, how these affect cancer cell morphology and migration are unknown. This understanding of the biophysical cues could potentially be used to devise therapeutic strategies that specifically target mechanisms that control biophysical sensing. Engineered platforms mimicking the TME have been shown to elicit directed cell migration, a key step in tumor invasion. These are immensely useful since they can be tightly controlled to provide specific mechanical cues and used to probe their influence on cell migration for different cell types.

Durotactic cues could be provided to study how different cell types engage in durotaxis.

Designing environments with tunable mechanical properties, including durotactic cues, could be developed as a high-throughput screening technique for patient derived cells, based on differential metastatic potential. Migration of cells through these platforms could be used to screen cells for therapeutic purposes.

1.2 Background and Introduction

1.2.1 Metastasis requires directed migration

Metastasis involves the translocation of cancer cells from a primary site to a new location and hence is above all, an end-product of cell migration. Cell migration is important in a variety of physiological and pathological processes such as wound healing [9], immunological responses

[10] and cancer metastasis [11,12]. Cells inside the body can undergo random migration or directed migration. The micro-environmental cues affect the way in which cells migrate, by activating signaling pathways that can alter the actomyosin cytoskeleton. Cells migrate in a random fashion in a spatially homogenous local environment. However, spatial heterogeneity leads to a bias in the motility machinery and results in directed migration. Directed cell migration is important in highly invasive cancer and is hypothesized to be a key indicator of metastatic cancer. The tumor 3 microenvironment is known to have altered mechanical properties. Cancer lesions are stiffer than the surrounding tissue. This in turn leads to a bias in the migrational cues originating from the spatial inhomogeneity in stiffness[13]. Durotaxis is the mode of directed migration in response to gradients of mechanical cues in the ECM[14]. Thus, it has been hypothesized that durotaxis[15] is involved in the initial steps of invasion-metastasis cascade.

1.2.2 Cells produce, remodel and degrade extracellular matrix

The extracellular matrix (ECM) is a vital non-cellular component with a plethora of dynamic macromolecules that are responsible for protective and organizational functions [16–

18]. The ECM is a heterogeneous network of biopolymers, primarily constituted of proteoglycans (PGs) and fibrous proteins which are chemically and physically cross-linked. The major components in the ECM include collagen, fibronectin, elastin, laminin, hyaluronic acid, keratan sulfate, chondroitin sulfate and heparan sulfate, to name a few. There are about 300 components in the ECM responsible for providing biophysical and biochemical properties to the

ECM. Some of these components like collagen, elastin, and fibronectin are capable of forming long filamentous fibers with different mechanical properties which serve as structural support.

For example, collagen strains very little before fracture and is non-linearly elastic, whereas elastin undergoes high levels of straining before fracture. Fiber density and alignment are also central to determining the ECM material properties. The other ECM components also have functional properties. The proteoglycans in the ECM are highly negatively charged, and are capable of resisting compressive stress [19]. The biopolymeric network of the ECM is a dynamic system which gets modified by generation, remodeling and degradation of its constituents.

Cells are capable of altering the ECM by synthesizing ECM proteins, changing the extent of 4 crosslinking, remodeling or degrading the matrix. Collagen is the main structural component in the ECM. Therefore, degradation of collagen is an important process in tissue remodeling, development and repair. Matrix metalloproteinases (MMPs) are enzymes that facilitate ECM degradation. There are different groups of MMPs, namely the collagenases, the gelatinases, the stromelysins, and the membrane-type MMPs. Among the known MMPs, MMP-1, 8, 13 and 18 are collagenases whereas MMP-2 and 9 are gelatinases that degrade different types of collagens and gelatins respectively [20–22]. Also, myosin mediated contractility is important in remodeling the ECM. The myosin motors generate contractility that allow cells to exert forces which remodel the ECM [23]. Cell-ECM interactions and force transmission are discussed in the following section. Increased matrix alignment, collagen density and consequently higher stiffness have been observed and characterized in the tumor microenvironment in cancer specimens from clinical biopsies and in vitro samples .This is potentially caused by increased matrix crosslinking and ECM component density[19,24–26]. These changes are hallmarks of cancer progression.

1.2.3 Cell - ECM interactions and regulation of cellular responses

Micro-environmental and cell intrinsic factors play a major role in cell morphology and migration [27–29]. With numerous studies, it has become increasingly evident that by modulating its components, the ECM provides different biochemical and biophysical signals that regulate cellular processes like survival, proliferation, differentiation, and migration [14,30–

32].There are bi-directional cell-cell and cell-matrix interactions, which result in a coordinated response to orchestrate various cellular functions. Cells adhere to the ECM using focal adhesions. These are assemblies of intracellular signaling and structural proteins that function to 5 link the cytoskeleton to the ECM through transmembrane receptors present at the cell surface.

Cells constantly experience mechanical forces produced by neighboring cells through the ECM and exert forces on their surroundings through these focal adhesions, which serve as mechanical linkage between the cells and ECM [33,34]. The spatially controlled dynamics of the focal adhesion sites create an imbalance in the traction stresses. Spatial heterogeneity in the ECM causes imbalance in the traction stress experienced by the cell which guides cell migration

[13,35]. Integrins are known to be the cell receptors responsible for sensing the forces and physical signals in the matrix. The interconnected systems of adhesion receptors, focal adhesion proteins, cytoskeletal networks and motor proteins produce responses to the mechanical forces experienced by the cells. Thus, focal adhesions are important in cell migration and force transmission between the cell and the surrounding. [25,27,34,36–38].

While the adhesion points between the cell and the ECM are central to biochemical signaling, force transmission and mechanosensing allow the cells to detect the biophysical properties of the ECM. Biophysical cues due to ECM properties like rigidity, porosity and topography influence anchorage-associated responses like cell migration and cell division [39–

46]. For example, cell number was observed to reduce with decreasing pore size. Pore size of the network fiber is a quantifiable measure of the 3D ECM structure. If the pore is too small, the structure poses restriction for cell motility and limits diffusion of nutrients and waste removal[47].

It has been widely accepted that cells can feel and respond to the external micro- environmental cues like rigidity, chemical and topographical properties [48,49]. Furthermore, researchers have shown that in vitro cells migrate in response to gradients of chemical concentration [50], light intensity [51], electrostatic potential [52], and rigidity [53] in the ECM. 6

However, ECM stiffness has been central to researchers probing the influence of ECM on cancer cell behavior, because the tumor-associated ECM is observed to be stiffer than the normal tissues, as high as ten-fold. Increased stiffness is linked to the enhanced cell proliferation and reduced cell apoptosis during tumorigenesis. Additionally, tumor cells are also known to increase the matrix deposition, and thereby increasing the matrix stiffness [54].Consequently, numerous studies examining the morphological and migrational response to uniform differences in stiffness in 2D and 3D as well as durotactic responses to spatial gradients of stiffness have been carried out and will be reviewed in the following three sections.

1.2.4 Techniques used to tune stiffness

ECM stiffness has been shown to increase during cancer progression. Understanding the effects of matrix stiffness on cell behavior is central to developing therapeutic strategies to fight this disease. In vitro models with stiffness cues have been extensively used to study cell responses among different cell lines. The most common method used to elicit cellular responses arising from stiffness is to control the bulk modulus of substrates. This is achieved by tuning the polymerization parameters such as temperature, concentration and polymerization time.

Controlled polymerization of synthetic polymers like polyacrylamide (PA) is widely used to create platforms for studying cellular responses on 2D substrates. PA gels have been used to study the influence of matrix stiffness on migration, proliferation, differentiation, adhesion and force generation in 2D. PA gels provide an ideal platform these studies, because the ratio of the monomer (acrylamide) to cross-linker (bis-acrylamide) can be varied to generate the required

Young’s Modulus (10 Pa – 100 kPa). However, PA gels are not compatible for cell encapsulation as it is known to induce cytotoxicity, thereby limiting 3D studies [55]. Controlling 7 bulk modulus has been used in 3D as well. Increasing collagen density has been shown to increase the bulk modulus. For instance, Hadjipanay et al 2009 examined 3D gels with stiffness gradients using collagen hydrogel [56]. But this altered the ligand density along with the stiffness gradient. Matrix protein crosslinking using glutaraldehyde and glycation are used to tune stiffness of the collagen gel. However, this can often change the chemical composition or architecture of the matrix [57]. Introducing glutaraldehyde in the matrix is known to alter the pore size, induce calcification, consequently limit the diffusion of nutrients and trigger toxicity.

Thus, bulk changes may be not the most appropriate way to study the effects of stiffness of the matrix. Controlling the microscale environment has emerged as a technique used to generate 3D scaffolds with stiffness cues.

While differential polymerization has been used to study the effects of substrate stiffness on cell migration, geometry and topography of the substrate, distance from the boundary and thickness of the substrate can also control the local stiffness experienced by cells. For example,

Ayaka Ueki and Satoru Kidoaki in 2015 showed that by designing curvature of the elastic boundary, they could manipulate cell mechanotaxis by controlling the effective stiffness [58].

Researchers have also used techniques based on manipulation of hydrogel thickness to tune mechanical properties, particularly effective modulus. Leong et al., 2010 showed that, in systems with interfacing soft-stiff substrates, along with the intrinsic mechanical properties of the substrates, the thickness also contribute to controlling the effective stiffness. This study demonstrated that in Human mesenchymal stem cells (hMSCs), cellular response was induced even at hundreds of microns away from the rigid support [59]. Later, in one of the early works based on the inherent mechanical gradients produced by edge effects, Rao SS et al. 2012, studied the influence of mechanical gradients on glioblastoma multiforme (GBM) tumor cells at the level 8 of morphology, actin organization and migration, by interfacing soft-stiff substrates [60]. Even though soft-stiff interfaces have been used to regulate stiffness, most of these studies are restricted to 2D systems.

1.2.5 Cell morphology and migration in 2D systems with tunable mechanical properties

Historically, the influence of ECM stiffness on cell behavior has been studied in vitro on

2D substrates [61]. Soft materials undergo large deformations under large forces, whereas stiff substrates can maintain the force without large deformations as it can store more mechanical energy. Consequently, studies show that cells spread less on soft substrates as compared to stiff substrates and the cell spreading area has been observed to increase with the substrate rigidity due to the reinforced adhesion on stiffer substrates. In one of the early studies, fibroblasts and epithelial cells on ECM coated substrates were shown to respond differently to soft and stiff substrates, at the level of cell morphology and migration. On soft substrates, the epithelial cells showed increased lamellipodial protrusion/retraction activities and fibroblasts migrated at a faster rate [30]. This was attributed to the destabilization of adhesion on soft/flexible substrates.

Later, in 2000 Lo et al. reported an increased cell spreading area for fibroblast cells on stiffer substrates[53]. Similar results were reported in smooth muscle cells [62] and A7r5 SMCs (aorta- derived cell line) [63]. Furthermore, the cell spreading area is also correlated to the size of the focal adhesions. Kim, Dong-Hwee and Wirtz Denis in 2013 showed that the cell spreading increased linearly with the focal adhesion size [64]. In 2017, Yeh, Yi-Chun et al showed that soft matrices impeded cell spreading and focal adhesions in Mouse mammary gland epithelial cells

(NMuMG) and collecting duct tubular epithelial cells (M1) [65]. These studies have established 9 that the cell spreading area and focal adhesions increase with increasing stiffness among different cell lines.

In addition to morphology, several studies have reported an increase in cell speed with increasing substrate Young’s Modulus. From a mechanistic perspective, cell migration is an outcome of net translocation arising from unbalanced forces acting on the cell. Cell migration in

2D has three major stages: i) extension of a protrusion in the direction of motion ii) adhesion at the leading edge and the de-adhesion at the trailing end of the cell on the substrate and iii) generation of contractile forces at the rear end and the cell body to pull itself in the direction of motion [66]. Vascular smooth muscle cells[67] and MCF10A epithelial cells [68] have been shown to follow this trend on 8–72 kPa and 3–35 kPa, respectively. However, in 1997 Pelham and Wang reported a decrease in cell migration speed with an increase in Young’s modulus of polyacrylamide substrates. Similarly, Ghosh et al 2007 showed a similar trend in primary adult human dermal fibroblasts (AHDFs)[69]. T24 carcinoma cells[70] and neutrophils cells[71] have also been shown to have a decreased migration speed on stiffer substrates. Later, studies showed that the migration speed has a biphasic dependence on the stiffness of the substrates and a variety of cell types exhibit a stiffness optimum, at which the cell speed is maximal thus unifying these seemingly contradictory results. A model based on the motor-clutch hypothesis for force transmission also exhibited a maximum in traction force with respect to substrate stiffness and predicted the biphasic dependence of the speed for stiffness sensitive cell migration.[72,73].

1.2.6 3D systems with tunable mechanical properties

Most of our understanding about cellular responses to mechanical cues are from studies on 2D substrates. In 2D, cells adopt morphologies and migration modes that are physiologically 10 irrelevant. The difference in dimensionality between the 2D and 3D systems itself can alter cell behavior. 2D and 3D cultures are distinct at the level of receptors anchored and consequently, force distributed to the ECM. Additionally, differences in 3D ECM structure can alter cell mechanotransduction. Focal adhesions and integrins present at the cell surface act as mechano- sensors and mechano-transducers. These proteins sense and transduce the mechanical signals into biochemical signals. It has been widely acknowledged that the mechanical cues can alter focal adhesion dynamics and integrin activation. Mechano-activated proteins like integrin have been shown to have an increased activation and clustering in response to an increase in matrix stiffness [65]. Integrin activation then triggers the organization of various cytoskeletal proteins like talin, paxillin etc. and recruits signaling molecules into the focal adhesion points [74]. This process is central to cellular responses caused by mechanical cues. In fibroblasts, Beningo et al. showed that by an increase in dimensionality, even in its simplest form, could elicit differences in cell adhesion and morphology. The cells cultured between two layers of polyacrylamide coated with ECM changed the morphology into a bipolar or stellate morphology which is found in vivo, whereas the cells cultured on 2D polyacrylamide substrate coated with ECM were more spread [75].

The type of 3D ECM used can also alter matrix architecture and composition. Collagen, fibrin and matrigel are commonly used polymerizable ECM proteins. An increase in ECM concentration has been shown to increase stiffness in collagen type I gels [23]. Polymerization temperature can alter the ECM microstructure, fibril size, pore size, fibril stiffness and modulus of the gel [76]. Due to the complexity of the physical environment, cell migration in 3D is still poorly understood. However, there are studies which characterize cell behavior in 3D systems with tunable steric and mechanical properties mimicking the tumor microenvironment. For 11 example, Fraley SI et al. 2010, showed that in 3D matrices, the focal adhesion proteins control the protrusion and matrix deformation, thereby affecting the cell migration speed and persistence time, despite the absence of detectable focal adhesion structures. Thus, this study provided evidence that 3D migration mode cannot be completely understood by simply stretching our understanding about 2D migration, in which cell motility depends upon the force transmitted to the ECM through focal adhesions [77].Furthermore, studies have shown that the dependence of cell speed on elastic modulus is cell line dependent. Although there are numerous studies demonstrating durotactic cell responses in 2D, very few papers have shown this in 3D systems

[78]. This could be attributed to the difficulty in finely controlling the mechanical properties of the matrix. The studies probing the effect of matrix parameters on cell response is often challenging, because the change in one of the factors often results in the change of another factor and hence independently controlling the mechanical cues is difficult.

1.2.7 Durotaxis

Heterogeneity in the stiffness around the cell leads to differences in force experienced by the cell at different focal adhesions. This imbalance experienced by the cell causes the directed migration due to mechanical cues [13]. Directional cell locomotion towards a stiffer environment is commonly observed in cell biology, a process known as durotaxis or mechanotaxis, which originates from the spatial heterogeneity in the stiffness and is dependent on the strength of the stiffness gradient. The term was coined in the paper by Lo et al. 2000 which studied the cell locomotion guided by physical interactions among fibroblast cells due to soft-stiff interfaces[53].

Numerous cell types have been shown to demonstrate the ability to sense and crawl towards the stiffer environments. In 2003, Wong Y.J et.al showed that vascular smooth muscle cells on 12 radial-gradient-compliant substrata exhibited durotaxis and cells accumulated in the stiffer region after 24 hrs [79]. Further, Isenberg, C. Brett and coworkers showed that in vascular smooth muscle cells, durotaxis with respect to the gradient increased with increasing magnitude of gradient, but remained independent of the absolute modulus [80]. The dependence of durotaxis on stiffness gradient strength was established among mesenchymal stem cells as well [81]. With more studies among a variety of cell types, it is universally appreciated that cells migrate towards a stiffer environment by mechanosensing the surroundings [82–90]. However, much of this work is confined to ECM coated 2D substrates. Stiffness mapping on clinical biopsy samples have shown that there is a higher stiffness in the local microenvironment associated with the tumor. In addition to the tissue stiffness gradients, there are soft-stiff interfaces which also provide mechanical gradients. Although it is unclear how these cues orchestrate directed migration, there are studies which show that durotaxis can be elicited in 3D [91]. 13

CHAPTER 2. RESEARCH OBJECTIVES

2.1 Develop and Characterize a TME Mimicking 3D System with Tunable Mechanical

Properties

Physical characteristics of the TME have been identified to influence cell function. A mechanical signal is induced by the physical properties of the microenvironment. This is then converted into a chemical signal which manifests a cell response. The transmembrane integrin receptors on the cell membrane receive the mechanotransduction signals, integrated over time and space. 3D tumor mimicking systems can be used to provide specific mechanical cues to understand the cellular responses in a physiologically relevant TME mimicking environment.

The influence of mechanical cues on cell behavior has been extensively studied in the past on 2D substrates. It has been widely acknowledged that cell behavior in 3D environments is characteristically distinct from 2D. Further, 3D assays have been shown to mimic in vivo cell responses. Tuning mechanical properties in a 3D system is often challenging since a change in one of the matrix parameters could alter other parameters. For example, collagen density can be used to control the matrix stiffness. But this changes the ligand density in the matrix. However, techniques using different heights soft gel interfaced on top of a stiff substrate are commonly used to create stiffness gradients in a 3D environment.

Soft-stiff interfaces have been frequently used to examine cellular responses to stiffness associated cues without altering the bulk properties. This is particularly important for cancer cell studies, because of the soft-stiff interfaces present in the TME. Although a handful of papers have leveraged this technique, most of this work is limited to 2D substrates. Also, surface chemistry at the soft-stiff interfaces have not been used to manipulate the local mechanical properties in 3D gels. A novel method using chemical functionalization of surfaces is used to 14 tune the stiff-soft interfacial mechanical linkage, possibly creating different stiffness properties in 3D in vitro gel above. By creating chambers with chemically functionalized coverslips and sandwiching cells embedded in collagen, cells are expected to experience differential forces depending on the strength of the attachment of collagen with the functionalized substrates. The differential attachment at the soft-stiff interface is expected to alter the mechanical properties.

Furthermore, by varying the collagen gel thickness, the influence of gel thickness on cell response is also probed. As the cells are embedded in the collagen squeezed between the glass coverslips, they feel a 3D in vivo like microenvironment. We want to probe how confinement and interfacial mechanical cues in the TME influence cell migration and morphology, by varying the collagen gel thickness and interfacial adhesion. The technique provides a way to generate mechanical cues without patterning or controlled polymerization or complex fabrication methods, which require expertise.

2.2 Evaluate How Cell Migration and Morphology Vary in 3D as a Function of Gel

Thickness and Surface Chemistry Linking the Soft and Stiff Interface

The migratory and morphological discrepancies in 2D models and in vivo have shifted the research focus to 3D in vitro models. The mechanism and the influence of mechanical cues on cell migration, morphology, adhesion, force generation and contractility have been extensively studied in 2D. 3D systems mimicking TME are becoming a standard to study cellular responses with varying mechanical cues. In our study, the 3D system developed is used to examine how cell responses varied with perturbations that provided cues arising from adhesion interactions at the soft-stiff interface. We expect the surface chemistries at the soft-stiff interface to change the adhesion between the collagen gel and the glass substrate, and consequently alter 15 the local mechanical properties felt by the cells. Additionally, thickness of the gels are known to alter the effective stiffness. Thick gels behave like soft substrates whereas thin gels behave like stiff substrates. We are interested in investigating how these perturbations affect the cellular responses, particularly cell migration and morphology. While confinements and soft-stiff interfaces are known to exist in the TME, how these affect cancer cell morphology and migration are poorly understood. This study will provide us insights about the effects of matrix parameters arising from adhesion interactions due to the surface chemistry present at the soft-stiff interface and the thickness of the collagen gel.

2.3 Examine if Durotactic Responses Could Be Elicited in the Surface Tunable 3D Systems

Directed migration is important in metastasis. Also, the cancer associated ECM is known to undergo mechanical and organizational transformation to form stiffer microenvironments.

Thus, durotaxis, the process by which cells migrate in the direction of increasing stiffness, is theorized to play a key role in cancer invasion. Durotaxis on 2D substrates has been established among many cell lines and it is universally accepted that cells migrate towards stiffer environments. Much of the physical understanding of the initial steps of the invasion-metastasis cascade is derived from these studies. However, there are very few studies which examine the durotactic migration of cells in a physiologically relevant 3D microenvironment. In our study, we examine whether durotactic migration could be observed in 3D environments with different collagen fiber network thicknesses. This could potentially help in devising a screening technique to separate patient derived cells based on metastatic potential. 16

CHAPTER 3. TUNING SURFACE FUNCTIONALIZATION AND COLLAGEN GEL

THICKNESS TO REGULATE CANCER CELL MIGRATION

Shalini R. Unnikandam Veettil1, Shawn M. Van Bruggen1, Doh-Gyu Hwang3, Michael D.

Bartlett3 and Ian C. Schneider1,2‡

1Department of Chemical and Biological Engineering, Iowa State University

2Department of Genetics, Development and Cell Biology, Iowa State University

3Department of Materials Science and Engineering, Soft Materials & Structures Lab, Iowa State

University

Present address: Iowa State University, Department of Chemical and Biological Engineering,

2114 Sweeney Hall, Ames, IA, 50011-2230

‡Author for correspondence (phone: (515) 294-0450, fax: (515) 294-2689, e-mail:

[email protected])

Modified from a manuscript submitted to Colloids and Surfaces B: Biointerfaces

3.1 Abstract

Cells have a tremendous ability to sense and respond to biophysical cues in their environment. This is particularly important in the context of cancer invasion and metastasis, where cells sense the extracellular matrix stiffness of the tumor microenvironment. The magnitude of the sensed stiffness can either promote or inhibit the migration of cancer cells out of the primary tumor into surrounding tissue. Much work has been done on examining the role of stiffness in tuning cancer cell migration by controlling elastic modulus in the bulk. To 17 complement this approach others have leveraged interactions between soft-stiff interfaces to understand cell responses to changes in stiffness. Cells appear to sense these interfaces at reasonable length scales (10-200 µm), providing a complementary approach for conveying biophysical signals to cells. In this paper, we examine the migration of cells embedded in a collagen fiber network sandwiched between two flat plates. We examine the role of both surface attachment of the collagen network to the stiff interface as well as thickness (50-540 m) of the collagen gel in driving cell morphology as well as cell migration speed. We find that surface attachment and thickness do not operate overlapping mechanisms, because they elicit different cell responses. While thickness and surface chemistry appear to control morphology, only thickness regulates cell migration speed. This suggests that surface attachment and thickness controls cell behavior through mechanisms beyond local stiffness regulation in confined fiber- forming networks.

3.2 Highlights

• Surface functionalization results in tunable adhesion between collagen gels and glass interfaces.

• When collagen is attached to the glass surface, projected cell area decreases, when moving

further into a collagen gel from the collagen gel-glass interface.

• When collagen does not attach to the glass surface, projected cell area is independent of

distance from the collagen gel-glass interface.

• Cell speed only depends on collagen gel thickness and not surface functionalization.

• Collagen gel thickness can be used to direct cell migration from areas where the collagen gel is

thick to areas where it is thin, opening the use of stiff-soft interfaces to control durotaxis, or

directed migration in response to stiffness gradients. 18

3.3 Introduction

Metastasis is the major cause of death due to cancer [1,2]. Invasion of the local extracellular matrix (ECM) is the first step of metastasis and is driven by cell migration. The

ECM influences fundamental aspects of cell migration by providing a scaffold on which to migrate and presenting promigratory ligands that the cell can recognize through receptors. Both biophysical and biochemical interactions between cells and the ECM influence cell adhesion, morphology and migration, and thereby play a key role in metastasis [3–6]. Collagen is arguably one of the most important components of the ECM in the tumor microenvironment (TME).

Collagen is a fiber forming protein that assembles into an entangled network that can be differentially crosslinked. The density and crosslinking of the collagen network determine the mechanical properties of the network, which are known to change as the tumor progresses [7].

The stiffness in and around tumors increases as the tumor progresses, brought on by enhanced collagen deposition and crosslinking. Cells sense the stiffness of their surroundings by anchoring onto the ECM and exerting traction forces using focal adhesions [8]. Engineered platforms that can control the mechanical properties of the ECM are tremendously useful in understanding how biophysical properties regulate cell migration and other processes germane to cancer metastasis.

Significant research has focused on the characterization of ECM stiffness and its influence on cell behavior, particularly cell migration [9–12]. Most of the studies examining the role of stiffness on cell behavior control the bulk elastic modulus of the ECM by tuning the polymerization parameters such as temperature, concentration, and polymerization time.

Furthermore, many of these have examined cell migration in 2D. Several studies have reported an increase in cell speed with increasing ECM elastic modulus. For example, vascular smooth muscle cells [13] and MCF10A epithelial cells [14] have been shown to follow this trend. 19

However, contrasting results have been shown in 3T3 fibroblasts [10]. Thus, it has been theorized that cell migration speed has a biphasic dependence on the stiffness of the ECM [15] with a maximal speed at an optimum elastic modulus, which varies for different cell lines.

Furthermore, it was found that cells preferentially move from less stiff region to a stiffer environment, a phenomenon known as durotaxis or mechanotaxis [10,11,13,16]. Although 2D experimental studies are often easier to conduct and are helpful in broadly answering the fundamental aspects of cell behavior by simplifying the intricacies arising from dimensionality, it has been established that 2D and 3D cell responses are characteristically distinct [17,18].

Cells embedded in a 3D system are exposed to a more complex environment with a variety of signals, compared to a 2D monolayer [18]. During 3D migration, the traction forces are transmitted to the matrix through focal adhesions, allowing the cell to remodel the surrounding ECM. Common 3D in vivo mimicking systems are composed of collagen, matrigel, hyaluronan, alginate, gelatin and poly(ethylene-glycol) to name a few [19–24]. In line with the

2D studies, the dependence of cell speed on elastic modulus depends on the cell type [25,26]. In addition, durotaxis can be elicited in 3D environments [27]. Collagen networks represent important environments in which to study the effects of mechanical properties on cell migration due to collagen’s abundance and ability to drive cancer cell invasion. However, altering collagen mechanical properties can be challenging. For instance, collagen networks are stiffer at higher densities, but the ligand density increases along with the elastic modulus. Collagen can also be crosslinked using glutaraldehyde or transglutaminase or glycated, stiffening the collagen gel.

However, glutaraldehyde crosslinking is not compatible with already embedded cells, transglutaminase crosslinking is difficult to control, and glycation leads to the formation of advanced glycation end products, thereby changing the chemical composition of the matrix [28]. 20

Other approaches are needed to tune the stiffness of collagen networks. Cells sense stiffness changes due to stiff-soft interfaces and ECM topographical features have been widely used to elucidate the influence of mechanical properties of the local TME [29–31]. Cell behavior depends on the thickness of the ECM network or the distance from a soft-stiff interface. For instance, cell spreading area is known to decrease with distance from the surface. However, many of these studies have focused on 2D cell behaviors. Only a handful of papers have assessed the role of soft-stiff interfaces in altering cell behavior in 3D. This includes reports of using this stiff-soft interface to control spread area and random migration [26], directed migration [32], focal adhesion formation [33] or myosin activity [27]. While this regulation of stiffness through soft-stiff interfaces has been used frequently in literature, most of this work is confined to examining cells on 2D substrates. Those in 3D have not leveraged control over stiffness by altering the surface chemistry at soft-stiff interfaces.

In this study we alter the surface chemistry of glass surfaces as well as collagen gel thicknesses to create different biophysical conditions in 3D collagen matrices. By controlling surface chemistry, we altered the strength of adhesion between the functionalized glass and collagen, presumably changing the local stiffness experienced by cells in the matrix close to the surface. The effect of gel thickness on stiffness is governed by the fiber structure of the gel and in fibrous collagen gels, the cell-mediated forces can travel up to a few hundred microns from the surface. We controlled the thickness of the collagen gel from 50 µm to 500 µm in addition to controlling the surface chemistry. Cell morphology and migration characteristics were determined under conditions with different surface chemistry and thickness to understand the biophysical effects of matrix parameters arising from polymerization of collagen in 3D chambers. Varying the collagen gel thickness parameters as well as the interfacial adhesion 21 allows us to probe how confinement and stiff-soft interfaces present in the tumor microenvironment influences cell migration.

3.4 Material and Methods

3.4.1 Surface modifications generating high and low collagen binding surfaces

Glass coverslips (Corning) were cleaned using the squeaky-clean procedure described elsewhere [34]. To prepare glutaraldehyde-treated surfaces, the squeaky cleaned coverslips were immersed in a piranha solution, 3:1 H2SO4 (Fisher): 30 wt% H2O2 (Fisher) v/v, for one hr at room temperature. Next, the coverslips were rinsed three times with nano-pure water and then immersed in a 1% (v/v) 3-aminopropyltriethoxysilane (APTES) (Acros Organics) in a 1 mM aqueous acetic acid (Fisher) solution for two hrs. After the silane coupling reaction, the coverslips were rinsed three times with nano-pure water baked at 100 °C. To treat the coverslips with glutaraldehyde (Electron Microscopy Scinces), they were immersed in a 6% glutaraldehyde solution (v/v) in 1x phosphate buffered saline (PBS) (Gibco) for one hr. Non-binding surfaces were created by treating the squeaky clean glass coverslips with 250 μg ml-1 poly(L-lysine)- poly(ethylene glycol) (PLL-PEG) (Alamanda Polymers) in PBS. Coverslips were immersed in

PLL-PEG solution and placed on a shaker for five mins and then placed in an incubator (37 °C) overnight (> 12 hrs).

3.4.2 Characterization of amine density on glass surfaces

The condensation reaction between primary amines and 4-nitrobenzaldehyde (Sigma

Aldrich) in anhydrous ethanol was used as a method to quantify the attachment of APTES to the glass coverslips [35]. The glass coverslips were immersed into a solution containing 0.4 mg ml-1 22

4-nitrobenzaldhyde and 20 μl acetic acid in 25 ml of anhydrous ethanol at 50 °C for three hrs.

After the reaction, the glass coverslips were washed with absolute ethanol and air dried. After drying, the glass coverslips were crushed and immersed in 10 ml 0.2% aqueous acetic acid solution for one hr at 30 °C. The absorbance of the 4-nitrobenzaldehyde in 0.2% acetic acid was used to quantify the concentration of APTES on the surface using a calibration curve prepared from absorbance data collected for different concentrations of 4-nitrobenzaldehyde.

3.4.3 Contact angle measurement

To test for the successful completion of a glass surface treatment, water contact angles were measured on each surface using a video camera (Javelin) and µManager 1.4 software [36].

A droplet of nano-pure water was placed on the surface and imaged through an objective lens

(10x, NA = 0.2, Thor Labs, Newton, NJ, USA) attached to the video camera. The collected images were analyzed using the contact angle plug-in in ImageJ (National Institutes of Health)

[37].

3.4.4 Adhesion characterization

The pull-off force of a glass indenter from an elastic gel as a function of displacement was measured employing the custom-built adhesion apparatus as shown in Fig. 3a. The adhesion instrument consists of a functionalized hemi-spherical glass indenter with a diameter of 5 mm, a heating stage and an optical microscope. The glass indenter was glued (cyanoacrylate adhesive) to a small piece of a glass slide that is glued to the head of a screw. Then the screw was inserted into a uniaxial load cell (FUTEK Advanced Sensor Technology, LSB200), which was connected to a piezo-controlled linear actuator (Physik Instruments (PI), N-565). Collagen at 2 mg ml-1 23 concentration in imaging media (see Section 2.5), was placed within a circular polydimethylsiloxane (PDMS)(Dow Corning Corporation) ring mounted on 25x75x1 mm glass microscope slide to prevent the specimen from spreading during polymerization. In addition, a dome-shaped glass case with a small circular hole was placed above the glass slide to reduce evaporation. The heating stage (Warner Instruments) on which the glass side was mounted has a circular hole at the center, which enables optical observation of the interface between the indenter and gel throughout the experiments.

Adhesion experiments were performed by bringing the indenter into contact with collagen. The probe was held in contact with the substrate for 30 mins at room temperature, followed by an additional 30 mins at elevated temperature of 37 °C, thereby polymerizing the collagen in situ and crosslinking collagen with the chemically treated surface of the indenter.

After the polymerization, the indenter was retracted at a constant displacement rate of 10 m s-1 until complete separation between the probe and collagen gel occurred. The load data were collected with a DAQ (National Instruments, NI USB-6002) in LabVIEW (National

Instruments).

3.4.5 Cell culture and characterization of cell morphology on 2D functionalized glass surfaces

The cells were cultured at 37 °C, 5% CO2 in Dulbecco’s Modified Eagle’s medium

(DMEM), (Sigma Aldrich) with 10% fetal bovine serum (FBS) (Gibco), 2% GlutaMAX (Gibco), and 1% penicillin streptomycin (Gibco). MDA-MB-231 cell morphology on functionalized glass coverslips was determined using a solution of 800,000-1,000,000 cells ml-1 in cell imaging media made from confluent cultures (70-90%). Imaging media consisted of DMEM without 24 phenol red, supplemented as above with 12 mM HEPES (Sigma Aldrich). Images were taken at

2, 4 and 16 hrs at 10x magnification (NA = 0.3, Nikon) and captured with a CCD (The Imaging

Source, DMK 41AU02 or Photometrics CoolSNAP HQ2). Aspect ratio, a parameter calculated by dividing the length of a cell by the width of that same cell and cell spreading area were measured using the ImageJ software.

3.4.6 Preparation of collagen gel in 3D chambers

MDA-MB-231 cells (human mammary basal/claudin low carcinoma cells, ATCC) were embedded (750,000-1,000,000 cell ml-1) between two glass coverslips to form a chamber in a 2 mg ml-1 rat collagen type I solution (Corning) prepared by mixing the imaging media (mentioned above) and collagen. Both the bottom and top coverslip were modified with a specific glass surface treatment (glutaraldehyde, squeaky clean glass, or PLL-PEG). The thickness of the chamber was controlled by placing spacers between the two glass coverslips. Additionally, a chamber with a step change in height was created by placing a glass strip on a double-sided tape in a chamber of 300 µm (see Fig. 7a). The solution of collagen and cells was mixed well before being sandwiched between the glasses. The samples were then flipped once every min for thirty mins at room temperature to keep the cells evenly distributed within the chamber as the collagen polymerized. The samples were placed in an incubator (37 °C, 5% CO2) for 30 min to allow the collagen to polymerize further. Finally, imaging media was added to each sample and placed back in the incubator for 24 hrs. Live cell images were taken in the middle of the chamber on a heating stage at 37 °C and imaged for 8 hrs at an interval of 2 mins. The transmitted images were taken with a 10x objective lens. At least three samples over at least two different days compiled a complete data set. 25

3.4.7 Fixing and staining of cells

To analyze the cell behavior in a 3D environment, cells were fixed and stained for F-actin as done previously [38]. Both fluorescent and transmitted images were collected from each sample at various heights within the collagen matrix using the z-stack feature on µManager 1.3 software. Images were taken every 5 µm for the 100 µm chambers and every 10 µm for the 240

µm chambers. The images were taken with a 20 x objective lens (NA = 0.5, Nikon) with a charge-coupled device (CoolSNAP HQ2, Photometrics) attached to an inverted microscope

(Eclipse Ti, Nikon). Image equipment was controlled by µManager. The images were analyzed for cell body area, protrusion length, F-actin intensity, and cell shape using ImageJ software.

3.4.8 Data Analysis

Cell migration was analyzed using the MTrackJ plugin [39] in ImageJ and cell speed was calculated as we have done previously [38]. The motility coefficient (µ) and persistence time

(Pm) were calculated using the following equations:

t  d22( t) = nS P t − P1 − ePm (1) mm 

and

SP2  = m (2) n

where n is the dimension of cell motility and Pm, persistence time, characterizes the average time over which there is no change in the direction and speed of the cell movement. 26

In the 3D gel with step change in chamber thickness, the cell orientation was measured using  (0 ≤  ≤ /2), the angle between the orientation of the cell and the direction pointing from the thick side to the thin side of the chamber. In order to quantify the orientation of cells with respect to the change in the gel thickness, an orientation directionality index (DI) defined as

DI = cos 2 (3)

was calculated using the measured angles. DI was calculated as a function of distance from the boundary of the step change for both the thick and thin sides after 2 hrs and 16 hrs. For the study examining cell migration from the thick side to the thin side of the gel, a directional migration parameter, DI over a time interval of 12 mins was calculated as we have done previously [40].

Additionally, a ratio was calculated based on the number of steps in the direction and opposite direction of the change in the thickness of the gel. This parameter was then averaged over the population to get the average ratio.

3.4.9 Statistical analysis

Analyses of variance (ANOVA) and student t-tests were carried out using MATLAB to investigate comparisons when the data sets were normally distributed. The Mann–Whitney U test was run in RStudio, when the normality of the data sets could not be assumed. For all the comparisons, the significance level (α) was set at 0.05 unless otherwise specified. Connecting lines over the conditions indicate the statistical differences in the distributions calculated using the methods mentioned above.

3.5 Results

3.5.1 Surface modification of glass generating high and low collagen binding surfaces

High binding surfaces that reacted with collagen were generated through functionalization with APTES and subsequent reaction with a bifunctional aldehyde

(glutaraldehyde). Amine density after APTES functionalization was measured by UV absorption of 4-nitrobenzaldehyde (4-NB) as described in materials and methods (Fig. 1a&b) [35]. The number density was calculated from the amount of 4-NB and the known area of the glass surface. A higher density of 4-NB molecules was recovered from the APTES-treated slides than the unfunctionalized and aldehyde treated surfaces (Fig. 1c). Treating APTES functionalized coverslips with glutaraldehyde abolished the free amines on the surface available for the condensation reaction and brought the number density to background level. The topographical area of APTES is 53.7 Å2 [41].Since APTES is attached to the surface, we estimated the surface area covered by a molecule to be approximately 7.2 Å2 .Thus, the average fractional surface area covered by APTES molecule was calculated to be between 0.05-0.1 (based on topographical area, which is the higher limit) and 0.007-0.02 (based on the estimated area, which is the lower limit). This indicates that there is low amine surface coverage. To further characterize the surfaces, contact angle measurements were taken. The aldehyde surfaces along with the amine surfaces had much larger contact angles (Fig. 1d) compared to the other surface chemistries, indicating a more hydrophobic surface. All the surface treatments increased the contact angle when compared to the unfunctionalized glass coverslips (Fig. 1e) and this increase indicates surface modification.

3.5.2 Cell spreading on 2D functionalized glass surfaces

Given that our functionalization appears to chemically alter the surface, we assessed the spreading of cells on these high and low binding surfaces. The observed cell spreading area for glutaraldehyde treated coverslips was greater than the conditions with unfunctionalized and PLL-

PEG surfaces after 4 and 16 hrs (Fig. 2a). Furthermore, PLL-PEG acted to inhibit spreading in comparison to unfunctionalized surfaces (Fig. 2a). Cell aspect ratio changed similarly to spread area, albeit with fewer significant differences. While aldehyde surfaces caused cells to adopt an elongated cell morphology as compared to either unfunctionalized or the PLL-PEG surface, the aspect ratio was comparable to the aldehyde surface with PLL-PEG. This is perhaps not surprising, given that the contact angle of aldehyde and aldehyde with PLL-PEG were also comparable (Fig. 2b). While not definitive, differences in the spreading of cells on the surfaces likely indicate differences in serum protein binding, as PLL-PEG is known to block protein binding and aldehyde surfaces can react with protein.

3.5.3 Adhesion of collagen to functionalized glass surfaces

Given changes in surface properties and cell adhesion that were dependent on surface treatment, we measured the adhesion interactions between a collagen gel and functionalized glass hemispheres with three different surface treatments: unfunctionalized, PLL-PEG functionalized, aldehyde functionalized. In these experiments the indenter and collagen gel were brought into contact for a predefined time and temperature and cured in-situ. The indenter was then retracted until complete separation occurs (see Fig. 3a for setup). We defined two parameters, the maximum adhesive force (Fmax) which is the maximum tensile load during separation [42], and the work of separation, which is the area under the force – displacement 29 curve from the starting point of retraction until the complete separation from the intender and gel

(a representative plot is provided in Fig. 3b). Fmax was measured for three different surface treatments (Fig. 3c). The PLL-PEG functionalization showed the smallest adhesion force between the indenter and the collagen whereas the aldehyde functionalization was the highest.

The work of separation for different indenters are shown in Fig. 3d. The observed trends are similar to Fig. 3c, such that the aldehyde-coated indenter showed the highest work of separation.

The Fmax value of the aldehyde functionalized intender was roughly 25% higher than the PLL-

PEG and the unfunctionalized glass chemistries. However, the work of separation for the aldehyde condition was only 15 % higher than the unfunctionalized intender. We hypothesize that even though the aldehyde-coated indenter creates covalent bonding at the coating-gel interface and is expected to show an increase in pull-off adhesion, because the low amine coverage on the indenter, a less pronounced enhancement of adhesion is seen.

3.5.4 Quantitative cell morphology in 3D collagen chambers with different glass functionalization

After MDA-MB-231 cells were embedded in a 3D collagen chambers (Fig. 4a), cells were fixed and stained for F-actin. The 3D collagen chambers had a thickness of 100 µm or 240

µm, and the glass coverslips were either unfunctionalized or aldehyde functionalized.

Fluorescence images were taken and analyzed for cell body area, protrusion length and F-actin intensity at various positions within the chamber. The cell body area was significantly higher in chambers with aldehyde functionalized glass coverslips (Fig. 4b). It was found that the area of the cell does not depend on the thickness of the chamber regardless of the surface chemistry.

However, this was an average measure and cells can be various distances from the surface, so the 30 area was calculated as a function of distance from the surface (Fig. 4c).The area of cells embedded in chambers with aldehyde-treated glass coverslips was higher near the surface of the chamber and decreases as the distance from the surface increases. On the other hand, MDA-MB-

231 cells in unfunctionalized chambers had a significantly lower area near the glass surface that remained relatively constant as the distance from the surface increases. This result indicated that the cells can sense the attachment of the aldehyde functionalized glass with collagen fibers in the chamber, but only near the stiff-soft interface. Once the cell is far away from the surface, the effect is not observed. Cell protrusion length was also analyzed (Fig. 4d). Cell protrusions in 240

µm chambers were longer than those in 100 µm chambers, but protrusion length was less sensitive to the chemical modification of the coverslips. This was different than cell spreading area, which was primarily dependent on surface chemistry. Finally, measurements of F-actin intensity in the cell body were taken under all four chamber conditions (Fig. 4e). The intensity of cells embedded in the 100 µm chambers was higher than the intensity of cells in the 240 µm chambers, forming an opposite to that of the cell protrusion length.

In addition to quantitative metrics of cell morphology, MDA-MB-231 cells embedded in the collagen chambers were also analyzed for their shape. Common cell morphologies were observed within the chambers and were placed into three general categories (Fig. 5a-c). A cell with a round body, no protrusions, and no net polarity in any direction was one cell morphology identified in the chambers (Fig. 5a). Alternatively, a cell with a polarity as it has a single protrusion on one side of the cell was also observed (Fig. 5b). The third type of cell morphology found in the chamber was a cell with an elongated body and two primary protrusions on both ends of the cell body (Fig. 5c). The fraction of cells displaying the three morphologies in the various collagen chamber conditions was analyzed (Fig. 5d-f). The thickness of the chamber 31 regulated the presence of polar (Fig. 5b) and non-polar cells (Fig. 5a&c). Collagen chambers of the 240 µm thickness had a larger fraction of cells displaying the polar morphology (Fig. 5b&e) than the non-polar, rounded morphology (Fig. 5a&c&d&f) regardless of the surface chemistry used. In addition, the fraction of cells with an elongated, non-polar morphology (Fig. 5c&f) did not dramatically change under different surface modifications or chamber thicknesses.

3.5.5 Cell motility in 3D collagen chambers with different glass functionalization

Cell migration was observed in the middle of the chambers for different gel thicknesses.

Cell speed and motility coefficient were dependent on the gel thickness, but the data did not provide evidence to suggest that surface treatment affected these motility properties. The cell speed and motility coefficient as both were found higher in the thick chambers (300 µm and 540

µm) in comparison to the thin chambers (50 µm and 100 µm) (Fig. 6a&b). However, persistence time calculations provided a differential response to cues arising from surface chemistries (Fig.

6c). The persistence time was found to be greater in the PLL-PEG at a chamber height of 50 µm, whereas in the 100 µm chamber it was larger in the unfunctionalized condition when compared with the other conditions. In the 300 µm chamber, the persistence time in the unfunctionalized condition was found to be larger than the aldehyde treated condition (Fig. 6c). Taken together, cell speed and motility coefficient are not dependent on surface functionalization, but are dependent on collagen gel thickness, whereas persistence time tends to be more dependent on surface functionalization, but in a collagen gel thickness manner.

The differential cell response with gel thickness, observed in cell speed and the motility coefficient led us to examine cell behavior in a system with a step change comprising a thick side

(300 µm) and a thin side (50 µm) (Fig. 7a). The orientation of cell alignment near the boundary 32 of the step change was measured and directionality index for static, but oriented cells was quantified at 2 and 16 hrs after embedding cells in the collagen gel. It was observed that the directionality index increased after 16 hrs for the thick and the thin sides when compared to the cell orientation after 2 hrs (Fig. 7b). Further, the cells at the thick side had a higher directionality index than the thin side after 16 hrs. In addition to this, when the cell orientation was examined as a function of distance from the boundary of the step, a higher directionality index was observed closer to the boundary at 16 hrs for the thick side (Fig. 7c). However, this trend was not observed at 2 hrs, suggesting that it takes time for cells to develop an orientation. At a distance of

125 µm from the boundary, the orientation observed was random (directionality = 0) for cells examined at 2 and 16 hrs in the thick and the thin sides (Fig. 7c). Along with orientation, we examined cell migration. The migration studies showed that a higher migration speed and motility coefficient were observed in the thick side of the chamber (Fig. 7d & 7e). The migration speeds in the thick and the thin sides of the chamber were comparable with the migration speeds observed in the chambers of gel thicknesses of 50 µm and 300 µm, respectively (Fig. 7d vs. 6a &

7e vs. 6b). We also examined directional migration. Directional cell migration was observed in the thick side of the chamber. These cells showed a positive, non-zero directionality (Fig. 7f).

However, cells on the thin side showed a more random migration in comparison to the thick side.

The persistence time of cell migration in a direction from the thick side to the thin side was also calculated. Although, the median of the observed persistence time was higher in the thick side, the data did not suggest that there was a difference between the distributions for the thin and thick sides statistically (Fig. 7g). Interestingly, though the median ratio of steps taken by cells in the direction of the projected vector (pointing from the thick side to the thin side) was positive for both the sides (300 µm and 50 µm), in the thin side there were cells which took large number 33 of steps in the opposite direction (Fig. 7h). These results indicate that thickness can result in directional sensing, reorienting cells towards interfaces over time. However, migration in the direction of the interface only occurs on the thick side, presumably where the stiffness is lowest.

3.6 Discussion

Collagen is an important and abundant protein in the TME and consequently, many studies have focused on methods for altering the mechanical properties within collagen gels including altering the concentration, temperature of polymerization and gelation time [43]. In addition, crosslinking through glutaraldehyde, transglutaminase, lysl oxidase or glycation has been shown to increase the stiffness of collagen gels [44,45]. However, these techniques have drawbacks including changes in collagen gel properties other than stiffness, difficulty in controlling the stiffness and incompatibility with already embedded cells. As an alternative approach, local stiffness can be controlled based on the distance between the cell and the stiff- soft interface, a parameter frequently controlled by the thickness of the soft material. Several studies have examined the regulation of thickness or position with respect to a stiff-soft interface to alter local stiffness, but none have examined the role of surface attachment in regulating local stiffness. Our primary interest was to assess the role of surface attachment of collagen within the collagen gel to a stiff glass surface, thus controlling local stiffness within the 3D collagen gel.

We assessed the adhesion of the collagen gel to the stiff glass surface under different functionalization methods by polymerizing the collagen gel in the presences of differentially functionalized glass beads. While we did not see dramatic changes in the adhesion force (~30% increase), this is likely due to the low coverage of the amine and consequently, aldehyde functionalization. Although, the aldehyde functionalized surface showed a larger pull-off force 34 and work of separation as expected, there was no significant difference between the PLL-PEG and unfunctionalized conditions. This approach of controlling attachment of the soft material to a stiff interface complements other approaches that modulate bulk stiffness.

Given that surface functionalization changed the adhesion between the collagen and the glass surface, we perturbed surface functionalization and collagen gel thickness in order to assess their influence on cell morphology and cell migration. Our original hypothesis was that one could tune the local stiffness in the collagen gel through either the surface attachment or position with respect to the stiff-soft interface, the latter being controlled by the thickness of the collagen gel. Stronger attachment and smaller distances from the interface would result in higher local stiffnesses. Cell area and fraction of cells that were extended and polarized were both dependent on surface chemistry and distance from the surface (or gel thickness) in a way that was consistent with the hypothesis that we were controlling local stiffness through either surface attachment or thickness. Surprisingly, other morphological and migration parameters showed no such trend.

Protrusion length, F-actin content and the fraction of polarized cells were not dependent on distance from the stiff-soft interface or surface functionalization, yet they were primarily dependent on the thickness of the collagen gel. In addition, migration speed and motility coefficient showed no dependence on surface functionalization, but rather thickness only.

Finally, persistence time depended only on surface functionalization and not gel thickness. While surface functionalization and collagen gel thickness jointly regulate local stiffness, other mechanisms could influence cell behavior. For instance, cells secrete diffusible pro-migratory factors into the medium. The local concentrations of these factors depend on both thickness of the collagen gel as well as distance to the stiff-soft interface. Different thicknesses represent different volumes in which these diffusible signals could accumulate, where thinner collagen 35 gels could attain higher concentrations than thicker collagen gels. Furthermore, since the concentration boundary condition at the stiff-soft interface is a no flux boundary condition, this enhances the concentration at short times when cells due to the reflective nature of the boundary.

However, absent of binding of the factors to the interface, surface chemistry should not affect this mechanism. If cell morphology or migration is dependent on secreted diffusible factors, a thickness or distance dependence would occur. Alternatively, collagen fiber assembly could depend on the surface functionalization of the interface, altering the structure of the collagen fibers close to the interface. Presumably this effect would be absent further away from the interface. If cell morphology or migration is dependent upon collagen structure, a surface functionalization dependence would occur. These represent possible additional mechanisms that may control cell morphology and migration through either surface chemistry or collagen gel thickness rather than through unified control over local stiffness.

While we are unaware of cell studies that have examined the role of surface attachment of soft materials to stiff materials, several groups have examined the role of gel thickness or distance from the surface. Numerous 2D studies have shown that the thickness of soft gels attached to a stiff surfaces alter cell spreading area, where thick gels result in small spread areas and thin gels result in large spread areas [46–51]. Interestingly, fiber forming matrices like those composed of fibrin or collagen appear to affect cell behavior further away from the surface as compared to gels like polyacrylamide, even when the bulk modulus is similar [52]. Fiber forming gels can exert changes on cell spreading areas up to about 150 µm away from the surface. The distance over which the area decreases by 50% for fiber forming gels appears to be around 80

µm, whereas that for polyacrylamide is about 4 µm [52]. Migration speed does not show such simple behavior. Migration speed increased in mesenchymal cells and decreased in fibroblasts 36 with thicker gels [53]. This is to be expected as migration has a biphasic response to stiffness and depending on the cell type and the elastic modulus of the gel, the stiffer environment could either act to increase or decrease speed. Finally, gel thickness appears to modulate collective cell migration and durotactic movement during the clustering of cells [54]. Fewer studies on the effect of stiffness have been conducted in 3D. Glioblastoma cells embedded in matrigel were shown to decrease their area and aspect ratio as the distance between the interface and the cell increased [26]. Similar to 2D, the penetration depth of the effect appears to be on the order of

150 µm. Furthermore, migration speed was fast, close to the stiff-soft interface and slower further away from the interface. Fibroblasts in collagen also showed the same area dependence as glioblastoma cells in matrigel, however, they did not appear to alter their migration speed

[32]. Finally, while surface attachment has not been quantitatively altered, attached vs. floating collagen gels have been compared and fibroblasts appear to decrease their migration speed in collagen gels that are not attached to a stiff interface [55]. This change in migration may possibly be due to the change in focal adhesions, which seem to disappear when more than 200 µm from the surface [56]. The area dependence of MDA-MB-231 cells in collagen on distance seems to be a bit blunted as compared to glioblastoma, fibroblasts or mesenchymal cells. We only observed differences over ~60 µm and only in situations where collagen was covalently attached to the glass. This could be a function of diminished mechanosensing in MDA-MB-231 cells or the presence of additional mechanisms beyond local stiffness modulation described in the previous paragraph. In addition, MDA-MB-231 cells appear to increase their migration speed in thicker collagen gels, but because this affect was not altered as a function of surface attachment of collagen, perhaps migration too depends on mechanisms beyond local stiffness modulation. 37

This change in stiffness as a function of distance from a stiff-soft interface can also be leveraged to induce durotaxis, directed migration in response to a stiffness gradient. Again, some work has been conducted on 2D substrates, where the surface stiffness has been controlled by underlying topographical features [11,57]. In 3D, durotactic gradients have been formed in constant thickness collagen gels attached to polyacrylamide gels with gradients of stiffness [27] or in gels with step changes in surface features [53]. Within the collagen gels formed over step changes, durotaxis occurred from soft to stiff (thick to thin section) with the relevant changes occurring from 100 to 40 µm thick collagen gels. The directionality found in our cells moving from 300 to 50 µm is somewhat larger but matches well with those for fibroblasts and mesenchymal stem cells. Patterning surface attachment of the soft material with a stiff interface or designing complex topographical structures is an interesting way to guide cell migration through durotaxis. These approaches afford the ability to fabricate surfaces well before cell embedding in the soft material. These techniques complement current approaches that employ spatially altering bulk modulus in 3D durotactic environments.

3.7 Conclusions

In this study we examined the role of glass surface chemistry and collagen gel thickness in controlling cell morphology and migration of MDA-MB-231 breast cancer cells in 3D collagen gels. Aldehyde functionalized glass in comparison to PLL-PEG or unfunctionalized glass is more hydrophobic, induces better 2D cell spreading and is more adhesive to collagen, presumably increasing the observed stiffness close to the glass-collagen gel interface. Cell spreading area in 3D collagen gels depended on the proximity of the cell to the glass-collagen interface, but only when glass was aldehyde functionalized and glass-collagen adhesion was largest. Unfunctionalized glass showed no area dependence on distance from the glass-collagen 38 interface. Cell migration differed. Cells migrated with higher speeds in thick collagen gels and appeared to show no dependence on the glass surface chemistry. Finally, directional migration could be induced by leveraging step changes in the thickness of the collagen. This indicates that while surface chemistry and collagen gel thickness can be used to alter the stiffness, they affect cell properties differently, suggesting additional mechanisms that may cooperate with stiffness in driving cell migration in confined ECM.

3.8 Acknowledgements

This work was supported by a seed grant from the College of Engineering at Iowa State

University as well as start-up funds from the Department of Materials Science and Engineering.

Shawn Van Bruggen was supported by a National Science Foundation Research Experience for

Undergraduates grant [1560012].

3.9 Competing Interests

The authors declare no competing interests.

3.10 Contributions

SRUV conducted the experiments in figures 1, 2, 6 and 7 and conceived the experiments in figure 7. SMVB conducted the experiments in figures 1, 2, 4 and 5 and conceived the experiments in figure 1. DGH conducted and conceived the experiments in figure 3. ICS conceived the experiments in all figures. MDB conceived the experiments in figure 3. ICS,

SRUV, SMVB, DGH and MBD wrote the manuscript.

3.11 References

[1] S. Valastyan, R.A. Weinberg, Tumor metastasis: Molecular insights and evolving paradigms, Cell. 147 (2011) 275–292. doi: 10.1016/j.cell.2011.09.024.

[2] P. Mehlen, A. Puisieux, Metastasis: A question of life or death, Nat. Rev. Cancer. 6 (2006) 449–458. doi:10.1038/nrc1886.

[3] F. Gattazzo, A. Urciuolo, P. Bonaldo, Extracellular matrix: A dynamic microenvironment for stem cell niche, Biochim. Biophys. Acta - Gen. Subj. 1840 (2014) 2506–2519. doi:10.1016/j.bbagen.2014.01.010.

[4] J.K. Mouw, G. Ou, V.M. Weaver, Extracellular matrix assembly: A multiscale deconstruction, Nat. Rev. Mol. Cell Biol. 15 (2014) 771–785. doi:10.1038/nrm3902.

[5] C. Bonnans, J. Chou, Z. Werb, Remodelling the extracellular matrix in development and disease, Nat. Rev. Mol. Cell Biol. 15 (2014) 786–801. doi:10.1038/nrm3904.

[6] S.L.K. Bowers, I. Banerjee, T.A. Baudino, The extracellular matrix: At the center of it all, J. Mol. Cell. Cardiol. 48 (2010) 474–482. doi:10.1016/j.yjmcc.2009.08.024.

[7] K. Polyak, I. Haviv, I.G. Campbell, Co-evolution of tumor cells and their microenvironment, Trends Genet. 25 (2009) 30–38. doi:10.1016/j.tig.2008.10.012.

[8] U.S. Schwarz, M.L. Gardel, United we stand – integrating the actin cytoskeleton and cell–matrix adhesions in cellular mechanotransduction, J. Cell Sci. 125 (2012) 3051– 3060. doi:10.1242/jcs.093716.

[9] A. Pathak, S. Kumar, Independent regulation of tumor cell migration by matrix stiffness and confinement, Proc. Natl. Acad. Sci. 109 (2012) 10334–10339. doi:10.1073/pnas.1118073109.

[10] C.M. Lo, H.B. Wang, M. Dembo, Y.L. Wang, Cell movement is guided by the rigidity of the substrate, Biophys. J. 79 (2000) 144–152. doi:10.1016/S0006-3495(00)76279-5.

[11] D.S. Gray, J. Tien, C.S. Chen, Repositioning of cells by mechanotaxis on surfaces with micropatterned Young’s modulus, J. Biomed. Mater. Res. - Part A. 66 (2003) 605–614. doi:10.1002/jbm.a.10585.

[12] D.E. Discher, Tissue Cells Feel and Respon to the Stiffness of Their Substrate, Science (80-.). 310 (2005) 1139–1143. doi:10.1126/science.1116995.

[13] B.C. Isenberg, P.A. DiMilla, M. Walker, S. Kim, J.Y. Wong, Vascular smooth muscle cell durotaxis depends on substrate stiffness gradient strength, Biophys. J. 97 (2009) 1313–1322. doi:10.1016/j.bpj.2009.06.021. 40

[14] M.R. Ng, A. Besser, G. Danuser, J.S. Brugge, Substrate stiffness regulates cadherin- dependent collective migration through myosin-II contractility, J. Cell Biol. 199 (2012) 545–563. doi:10.1083/jcb.201207148.

[15] B.L. Bangasser, S.S. Rosenfeld, D.J. Odde, Determinants of maximal force transmission in a motor-clutch model of cell traction in a compliant microenvironment, Biophys. J. 105 (2013) 581–592. doi:10.1016/j.bpj.2013.06.027.

[16] J.Y. Wong, A. Velasco, P. Rajagopalan, Q. Pham, Directed movement of vascular smooth muscle cells on gradient-compliant hydrogels, Langmuir. 19 (2003) 1908–1913. doi:10.1021/la026403p.

[17] A.S. Meyer, S.K. Hughes-Alford, J.E. Kay, A. Castillo, A. Wells, F.B. Gertler, D.A. Lauffenburger, 2D protrusion but not motility predicts growth factor-induced cancer cell migration in 3D collagen, J. Cell Biol. 197 (2012) 721–729. doi:10.1083/jcb.201201003.

[18] B.M. Baker, C.S. Chen, Deconstructing the third dimension – how 3D culture microenvironments alter cellular cues, J. Cell Sci. 125 (2012) 3015–3024. doi:10.1242/jcs.079509.

[19] S. Geraldo, A. Simon, N. Elkhatib, D. Louvard, L. Fetler, D.M. Vignjevic, Do cancer cells have distinct adhesions in 3D collagen matrices and in vivo?, Eur. J. Cell Biol. 91 (2012) 930–937. doi:10.1016/j.ejcb.2012.07.005.

[20] S.T. Kreger, S.L. Voytik-Harbin, Hyaluronan concentration within a 3D collagen matrix modulates matrix viscoelasticity, but not fibroblast response, Matrix Biol. 28 (2009) 336– 346. doi:10.1016/j.matbio.2009.05.001.

[21] Z. Li, S. Huang, Y. Liu, B. Yao, T. Hu, H. Shi, J. Xie, X. Fu, Tuning Alginate-Gelatin Bioink Properties by Varying Solvent and Their Impact on Stem Cell Behavior, Sci. Rep. 8 (2018). doi:10.1038/s41598-018-26407-3.

[22] M. Cavo, M. Caria, I. Pulsoni, F. Beltrame, M. Fato, S. Scaglione, A new cell-laden 3D Alginate-Matrigel hydrogel resembles human breast cancer cell malignant morphology, spread and invasion capability observed “in vivo,” Sci. Rep. 8 (2018) 5333. doi:10.1038/s41598-018-23250-4.

[23] E. Gontran, M. Juchaux, C. Deroulers, S. Kruglik, N. Huang, M. Badoual, O. Seksek, Assessment of the ability of poly(l-lysine)–poly(ethylene glycol) (PLL–PEG) hydrogels to support the growth of U87-MG and F98 glioma tumor cells, J. Appl. Polym. Sci. 135 (2018) 46287. doi:10.1002/app.46287.

[24] K.M. Hakkinen, J.S. Harunaga, A.D. Doyle, K.M. Yamada, Direct Comparisons of the Morphology, Migration, Cell Adhesions, and Actin Cytoskeleton of Fibroblasts in Four Different Three-Dimensional Extracellular Matrices, Tissue Eng. Part A. 17 (2011) 713– 724. doi:10.1089/ten.tea.2010.0273. 41

[25] M.H. Zaman, L.M. Trapani, A.L. Sieminski, D. MacKellar, H.Y. Gong, R.D. Kamm, A. Wells, D.A. Lauffenburger, P. Matsudaira, A. Siemeski, D. MacKellar, H.Y. Gong, R.D. Kamm, A. Wells, D.A. Lauffenburger, P. Matsudaira, Migration of tumor cells in 3D matrices is governed by matrix stiffness along with cell-matrix adhesion and proteolysis, Proc. Natl. Acad. Sci. 103 (2006) 10889–10894. doi:10.1073/pnas.0604460103.

[26] S.S. Rao, S. Bentil, J. DeJesus, J. Larison, A. Hissong, R. Dupaix, A. Sarkar, J.O. Winter, Inherent Interfacial Mechanical Gradients in 3D Hydrogels Influence Tumor Cell Behaviors, PLoS One. 7 (2012) e35852. doi:10.1371/journal.pone.0035852.

[27] M. Raab, J. Swift, P. Dingal, P. Shah, J.W. Shin, D.E. Discher, Crawling from soft to stiff matrix polarizes the cytoskeleton and phosphoregulates myosin-II heavy chain, J. Cell Biol. 199 (2012) 669–683. doi:10.1083/jcb.201205056.

[28] B.N. Mason, C.A. Reinhart-King, Controlling the mechanical properties of three- dimensional matrices via non-enzymatic collagen glycation., Organogenesis. 9 (2013) 70–5. doi:10.4161/org.24942.

[29] M. Nikkhah, F. Edalat, S. Manoucheri, A. Khademhosseini, Engineering microscale topographies to control the cell-substrate interface, Biomaterials. 33 (2012) 5230–5246. doi:10.1016/j.biomaterials.2012.03.079.

[30] E. Hadjipanayi, V. Mudera, R.A. Brown, Guiding cell migration in 3D: A collagen matrix with graded directional stiffness, Cell Motil. Cytoskeleton. 66 (2009) 121–128. doi:10.1002/cm.20331.

[31] H. Jeon, S. Koo, W.M. Reese, P. Loskill, C.P. Grigoropoulos, K.E. Healy, Directing cell migration and organization via nanocrater-patterned cell-repellent interfaces, Nat. Mater. 14 (2015) 918–923. doi:10.1038/nmat4342.

[32] C.H. Feng, Y.C. Cheng, P.H.G. Chao, The influence and interactions of substrate thickness, organization and dimensionality on cell morphology and migration, Acta Biomater. 9 (2013) 5502–5510. doi:10.1016/j.actbio.2012.11.024.

[33] S.I. Fraley, Y. Feng, R. Krishnamurthy, D.-H. Kim, A. Celedon, G.D. Longmore, D. Wirtz, A distinctive role for focal adhesion proteins in three-dimensional cell motility, Nat. Cell Biol. 12 (2010) 598–604. doi:10.1038/ncb2062.

[34] C.M. Waterman-Storer, Microtubule/organelle motility assays., Curr. Protoc. Cell Biol. Chapter 13 (2001) Unit 13.1. doi:10.1002/0471143030.cb1301s00.

[35] S. Xiang, G. Xing, W. Xue, C. Lu, J.-M. Lin, Comparison of two different deposition methods of 3-aminopropyltriethoxysilane on glass slides and their application in the ThinPrep cytologic test, Analyst. 137 (2012) 1669. doi:10.1039/c2an15983j.

[36] A. Edelstein, N. Amodaj, K. Hoover, R. Vale, N. Stuurman, Computer control of microscopes using manager, Curr. Protoc. Mol. Biol. (2010) 1–17. doi:10.1002/0471142727.mb1420s92. 42

[37] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, NIH Image to ImageJ: 25 years of image analysis, Nat. Methods. 9 (2012) 671–675. doi:10.1038/nmeth.2089.

[38] J. Wang, I.C. Schneider, Myosin phosphorylation on stress fibers predicts contact guidance behavior across diverse breast cancer cells, Biomaterials. 120 (2017) 81–93. doi:10.1016/j.biomaterials.2016.11.035.

[39] E. Meijering, O. Dzyubachyk, I. Smal, Methods for cell and particle tracking, Methods Enzymol. 504 (2012) 183–200. doi:10.1016/B978-0-12-391857-4.00009-4.

[40] I.C. Nuhn, J.A.M.; Perez, A.M.; Schneider, J.A.M. Nuhn, A.M. Perez, I.C. Schneider, Contact Guidance Diversity in Rotationally Aligned Collagen Matrices, Acta Biomater. accepted (2017).

[41] National Center for Biotechnology Information, (n.d.). https://pubchem.ncbi.nlm.nih.gov/compound/13521.

[42] M.D. Bartlett, A.J. Crosby, Scaling normal adhesion force capacity with a generalized parameter, Langmuir. 29 (2013) 11022–11027. doi:10.1021/la4013526.

[43] E.E. Antoine, P.P. Vlachos, M.N. Rylander, Tunable collagen I hydrogels for engineered physiological tissue micro-environments, PLoS One. 10 (2015). doi:10.1371/journal.pone.0122500.

[44] B.A. Harley, J.H. Leung, E.C.C.M. Silva, L.J. Gibson, Mechanical characterization of collagen-glycosaminoglycan scaffolds, Acta Biomater. 3 (2007) 463–474. doi:10.1016/j.actbio.2006.12.009.

[45] M.G. Haugh, C.M. Murphy, R.C. McKiernan, C. Altenbuchner, F.J. O’Brien, Crosslinking and Mechanical Properties Significantly Influence Cell Attachment, Proliferation, and Migration Within Collagen Glycosaminoglycan Scaffolds, Tissue Eng. Part A. 17 (2011) 1201–1208. doi:10.1089/ten.tea.2010.0590.

[46] A.J. Engler, L. Richert, J.Y. Wong, C. Picart, D.E. Discher, Surface probe measurements of the elasticity of sectioned tissue, thin gels and polyelectrolyte multilayer films: Correlations between substrate stiffness and cell adhesion, in: Surf. Sci., 2004: pp. 142– 154. doi:10.1016/j.susc.2004.06.179.

[47] S. Sen, A.J. Engler, D.E. Discher, Matrix strains induced by cells: Computing how far cells can feel, Cell. Mol. Bioeng. 2 (2009) 39–48. doi:10.1007/s12195-009-0052-z.

[48] J.M. Maloney, E.B. Walton, C.M. Bruce, K.J. Van Vliet, Influence of finite thickness and stiffness on cellular adhesion-induced deformation of compliant substrata, Phys. Rev. E. 78 (2008) 041923. doi:10.1103/PhysRevE.78.041923.

[49] A. Buxboim, K. Rajagopal, A.E.X. Brown, D.E. Discher, How deeply cells feel: Methods for thin gels, J. Phys. Condens. Matter. 22 (2010) 194116. doi:10.1088/0953- 8984/22/19/194116. 43

[50] C.A. Mullen, T.J. Vaughan, K.L. Billiar, L.M. McNamara, The effect of substrate stiffness, thickness, and cross-linking density on osteogenic cell behavior, Biophys. J. 108 (2015) 1604–1612. doi:10.1016/j.bpj.2015.02.022.

[51] W.S. Leong, C.Y. Tay, H. Yu, A. Li, S.C. Wu, D.H. Duc, C.T. Lim, L.P. Tan, Thickness sensing of hMSCs on collagen gel directs stem cell fate, Biochem. Biophys. Res. Commun. 401 (2010) 287–292. doi:10.1016/j.bbrc.2010.09.052.

[52] M.S. Rudnicki, H.A. Cirka, M. Aghvami, E.A. Sander, Q. Wen, K.L. Billiar, Nonlinear strain stiffening is not sufficient to explain how far cells can feel on fibrous protein gels, Biophys. J. 105 (2013) 11–20. doi:10.1016/j.bpj.2013.05.032.

[53] P. hsiu G. Chao, S.C. Sheng, W.R. Chang, Micro-composite substrates for the study of cell-matrix mechanical interactions, J. Mech. Behav. Biomed. Mater. 38 (2014) 232–241. doi:10.1016/j.jmbbm.2014.01.008.

[54] C.G. Tusan, Y.H. Man, H. Zarkoob, D.A. Johnston, O.G. Andriotis, P.J. Thurner, S. Yang, E.A. Sander, E. Gentleman, B.G. Sengers, N.D. Evans, Collective Cell Behavior in Mechanosensing of Substrate Thickness, Biophys. J. 114 (2018) 2743–2755. doi:10.1016/j.bpj.2018.03.037.

[55] M. Miron-Mendoza, J. Seemann, F. Grinnell, The differential regulation of cell motile activity through matrix stiffness and porosity in three dimensional collagen matrices, Biomaterials. 31 (2010) 6425–6435. doi:10.1016/j.biomaterials.2010.04.064.

[56] K.E. Kubow, A.R. Horwitz, S.I. Fraley, Y.F. Feng, D. Wirtz, G.D. Longmore, Reducing background fluorescence reveals adhesions in 3D matrices, Nat Cell Biol. 13 (2011) 5- U254. doi:10.1038/ncb0111-5.

[57] D. Joaquin, M. Grigola, G. Kwon, C. Blasius, Y. Han, D. Perlitz, J. Jiang, Y. Ziegler, A. Nardulli, K.J. Hsia, Cell migration and organization in three-dimensional in vitro culture driven by stiffness gradient, Biotechnol. Bioeng. 113 (2016) 2496–2506. doi:10.1002/bit.26010.

3.12 Figure Legends

Figure 1. Characterization of surface properties on functionalized glass surfaces a) UV absorbance spectra for 4-nitrobenzaldehyde (4NB) at different concentrations (0.10, 0.21, 0.41,

0.83, 1.7, 3.3, 6.6, 13 M) for one experiment. The gray arrow indicates increasing 4NB concentrations. b) Standard curve obtained from the absorbance at 270 nm (Nexperiments = 2). c) The number of density 4NB on unfunctionalized, amine functionalized and aldehyde treated glass (Nsamples > 4). d) Contact area between water and glass surfaces with different surface functionalization (Nsamples > 5). The error bars represent the 95% confidence intervals. Lines over bars indicate that conditions were statistically significantly different as determined by a student’s t-test with p = 0.05. 45

Figure 2. Characterization of cell spreading on functionalized glass surfaces a) Cell spreading area and b) cell aspect ratio of MDA-MD231 cells observed after 2, 4 and 16 hrs for different chemical functionalization on glass surfaces (Nsamples > 3, Ncells > 226). The error bars represent the 95% confidence intervals. Lines over bars indicate that conditions were statistically significantly different as determined using analysis of variance ANOVA test with p

= 0.05.

Figure 3. Characterization of adhesion of collagen to functionalized glass surfaces a)

Custom-built adhesion apparatus used to measure the pull-off force of the glass indenter as a function of displacement. b) A representative image of the force-displacement curve obtained from the experiment. c) The corrected pull-off force and d) the work of separation observed when the functionalized glass indenter was in contact with polymerizing collagen. The in situ polymerization of collagen and simultaneous cross-linking of chemically treated glass surfaces with the collagen were carried out with heating stage (Nsamples > 3). The error bars represent the 95% confidence intervals. Lines over bars indicate that conditions were statistically significantly different as determined by a student’s t-test with p = 0.05 and * indicates p = 0.1.

Figure 4. Characterization of cell morphology in 3D collagen chambers of different thicknesses and glass functionalization a) Schematic representation of imaging set up with cells embedded in collagen gel. b) Cell spreading area, d) protrusion length and e) F-actin intensity averaged over thickness for 100 μm and 240 μm chambers of unfunctionalized and aldehyde functionalized coverslips. c) Cell spreading area observed at intervals of 5 μm and 10

μm for 100 μm and 240 μm collagen chambers with unfunctionalized and aldehyde functionalized glass surfaces, respectively (Nsamples = 3, Ncells > 99 for b,d & e, Nsamples = 3,

Ncells > 9 for c). The error bars represent the 95% confidence intervals. In the box plots, the 48 middle line indicates median, the top and bottom of the box indicate 75th percentile and 25th percentile respectively, and the whiskers indicate the 90th and 10th percentiles. Lines over bars indicate that conditions were statistically significantly different as determined by an ANOVA test with p = 0.05. + indicates that there are values outside the range of y-axis.

Figure 5. Characterization of cell shape in 3D collagen chambers of different thicknesses and glass functionalization Different cell morphologies were found for all surface conditions and chamber thicknesses. The cells were grouped according to the following cell morphologies: a) round and non-polar, b) round and polar, c) skinny and non-polar. Roundness is determined by the cell body while polarity is determined by the protrusions. Fractional cell morphology observed in 100 µm and 240 µm chambers with unfunctionalized and aldehyde treated conditions for d) round and non-polar, e) round and polar and f) skinny and non-polar morphologies (Nsamples = 3, Ncells > 99). All scale bars are 10 µm.

Figure 6. Motility characterization in 3D collagen chambers with functionalized glass surfaces Migration analysis of MDA-MB-231 cells embedded in 2mg ml-1 collagen gel chambers with different gel thicknesses and surface chemistries. a) Cell migration speed, b) motility coefficient and c) persistence time of MDA-MD231 cells observed in the middle of the chamber (Nsamples > 3, Ncells > 44). The cell migration speed and motility coefficient were found to be larger in the thick chambers (300 µm and 540 µm) than the thin chambers (50 µm and 100 µm). In the box plots, the middle line indicates median, the top and bottom of the box 51 indicate 75th percentile and 25th percentile respectively, and the whiskers indicate the 90th and

10th percentiles. Lines over bars indicate that conditions were statistically significantly different as determined by a Mann-Whitney U test with p = 0.05. + indicates that there are values outside the range of y-axis.

Figure 7. Motility characterization in 3D collagen chambers with step change in thickness a) Schematic representation of a step change in the collagen gel. Migration and directionality 53 analysis of MDA-MB-231 cells embedded in 2mg ml-1 collagen gel chambers, observed at the boundary of a step change in the thickness (Nsamples>4, Ncells > 75 for b, Ncells > 12 for c,

Nsamples= 4, Ncells = 50 for d-h). b) Directionality index averaged over distance from the boundary of the step change for 50 µm and 300 µm gel thicknesses. c) Directionality index as a function of distance from the boundary. d) The cell migration speed and e) motility coefficient are found to be larger in the thick side of the chamber (300 m) than the thin side (50 m).A directed motility of the cells was found from the thick side of the step change to the thin side f).

Persistence time g) and ratio of the direction (towards+ / away-) of each step of cell movement h) in a direction oriented from the thick side to the thin side of the chamber. The error bars represent the 95% confidence intervals. In the box plots, the middle line indicates median, the top and bottom of the box indicate 75th percentile and 25th percentile respectively, and the whiskers indicate the 90th and 10th percentiles. Lines over bars indicate that conditions were statistically significantly different as determined by ANOVA test with p = 0.05 or determined by a Mann-Whitney U test with p = 0.05. + indicates that there are values outside the range of y- axis. 54

CHAPTER 4. FUTURE WORK

4.1 Investigating the Biophysical Effects of Hyaluronan Molecular Weight and

Concentration on Breast Cancer Cell Migration in Collagen Matrices

In our study, collagen matrices in 3D confined chambers have been developed as a technique to mimic the in vivo environment with tunable mechanical properties. We are furthering our studies by incorporating another important ECM component into the 3D systems.

Glycosaminoglycans (GAGs) are complex polysaccharide molecules present in the ECM with repeating unit of specific disaccharides. GAGs are known to modulate cellular process like cell proliferation, cell migration, cell-matrix adhesion and cell-matrix interactions, which are critical to tumorigenesis. There are four sub families of GAGs: Heparin (HS), chondroitin sulfate/dermatan sulfate (CS), keratan sulfate (KS), and hyaluronan (HA) (Fig 8). Researchers have been interested in probing the influence of hyaluronan in cancer invasion and metastasis because a high level of hyaluronan has been correlated to high mortality rates among certain carcinomas[92,93]. The interplay between HA producing hyaluronan synthases and hyaluronidases balances the HA levels in normal tissues. Numerous studies have shown that the expression of hyaluronan synthases (HASs) is high during cancer progression, particularly at the invading ends of certain breast carcinomas. Consequently, an overproduction of HA is observed

[94–97]. Furthermore, matrix elastic modulus has been shown to with increase in the concentration of HA in HA-collagen composite matrices in vitro[98]. In vivo and in vitro studies have also revealed that an increased expression of the HAS proteins resulted in an increased tumor progression[99]. Currently, many clinical trials are probing the possibility of degrading

HA in the tumor microenvironment using hyaluronidases as a potential therapy for cancer.

However, the degradation would lead to formation of smaller fragments of HA with low 55 molecular weights. This could alter biophysical properties and biochemical signaling pathways.

For example, high molecular weight HA increased but low molecular weight HA reduced the

CD44 clustering among different cell lines [100]. However, the impact of these changes and the mechanism behind the cell responses largely remain obscure. There are limited number of studies exploring the influence of molecular weights of HA on cancer cell behavior.

Furthermore, a number of HA receptors have been identified, of which CD44 and RHAMM are known to influence cell behavior [101,102]. HA in the ECM interacts with the cell by binding itself to these receptors at the cell-surface. Studies have shown that knockdown of CD44 and

RHAMM receptors have been shown to reduce proliferation and migration. Thus, it is expected that there are two possible ways by which HA composition and concentrate on affect cell behavior. One possible mechanism is by altering the mechanical properties of the matrix and the other through the hyaluronan induced signaling through the receptors. It could also be an interplay between the mechanical cues and bio-chemical signals. Although HA-induced biophysical mechanisms are known to alter the cell behavior by imparting localized variations in the ECM, it is still poorly understood. Analyzing the function of HA and the underlying mechanisms that elicit cellular responses in collagen-HA composites mimicking the TME will provide insights into potential therapeutic strategies for cancer.

Figure 8. The structure of hyaluronan consisting of repeating disaccharides of glucuronic acid and N-acetylglucosamine [92]

4.2 Investigating the Influence of Surface Topographical Cues to Elicit Durotaxis in 3D

Systems Among Breast Cancer Cells

Engineered surface topographical cues in a 3D scaffold designed to manipulate physical cues to the cells could be useful to elicit specific cellular responses. The cell response is thought to be a function of the strength and persistence of these signals. Micro-fabrication techniques have been widely used to manipulate physical cues by generating surface topographies to study cellular responses. Surface topographies with defined physical features can be precisely generated using a variety of techniques, such as nanoimprint lithography, photolithography, embossing, colloidal lithography, fused deposition modeling, micro-stereolithography, and 3D printing. Furthermore, studies have reported that by administering artificial biomaterials to present surface textures like groves, pillars and pillars affected cell responses among stem cells, particularly proliferation and differentiation [103,104]. Although these fabrication techniques offer precise control and reproducibility, most of the cell studies are on 2D substrates with surface topographies. Surface micro-textures have been shown to affect cellular responses such as cell attachment, orientation, adhesion, cell growth, migration and differentiation[105–107]. 57

For example, Alvaro Mata et.al in 2002 showed that by designing channel textures on PDMS, cell alignment, proliferation and morphology were significantly different from the non-textured

PDMS surface in Growth of human connective tissue progenitor cells (CTPs) [108]. 3D gels with embedded cells capable of remodeling the microenvironment and force transmission through focal adhesions all over the cell surface, are physiologically more relevant. In the more recent studies, 3D systems are becoming a standard for cell biology studies. In one of the early studies in 3D, micro-rods incorporated into matrigel matrices were reported to affect cell proliferation among fibroblasts [109]. Stiffness or mechanical cues, in fiber forming matrices like collagen or fibronectin, are capable of affecting cell behavior even hundreds of microns away from the cells. We are interested in manipulating the surface topographies using 3D printing to provide physical cues to direct cell migration in 3D. Previously, we observed durotactic response in confined 3D environment by spatially altering the mechanical cues by presenting a step change in thickness of the collagen gels, among breast cancer cells. Thus, in our study we are devising 3D structures (Fig. 9) and leveraging the cell responses to soft-stiff interfaces to direct cell migration. 58

Figure 9. Schematic of pyramidal 3D structure used to provide soft-stiff interfaces (scale in mm) 59

REFERENCES

[1] WHO | Cancer, WHO. (2017). http://www.who.int/mediacentre/factsheets/fs297/en/ (accessed August 9, 2017).

[2] R.L. Siegel, K.D. Miller, A. Jemal, Cancer statistics, 2018, CA. Cancer J. Clin. 68 (2018) 7–30. doi:10.3322/caac.21442.

[3] G.P. Gupta, J. Massagué, Leading Edge Review Cancer Metastasis: Building a Framework, (n.d.). doi:10.1016/j.cell.2006.11.001.

[4] S. Valastyan, R.A. Weinberg, Tumor Metastasis: Molecular Insights and Evolving Paradigms, (n.d.). doi:10.1016/j.cell.2011.09.024.

[5] P.S. Steeg, METASTASIS SUPPRESSORS ALTER THE SIGNAL TRANSDUCTION OF CANCER CELLS Women’s Cancers Section, Nat. Rev. | CANCER. 3 (2003) 5. doi:10.1038/nrc967.

[6] S. Valastyan, R.A. Weinberg, Tumor metastasis: Molecular insights and evolving paradigms, Cell. 147 (2011) 275–292. doi:10.1016/j.cell.2011.09.024.

[7] I.J. Fidler, The pathogenesis of cancer metastasis: The “seed and soil” hypothesis revisited, Nat. Rev. Cancer. (2003). doi:10.1038/nrc1098.

[8] D. Hanahan, R.A. Weinberg, Hallmarks of Cancer: The Next Generation, Cell. 144 (2011) 646–674. doi:10.1016/j.cell.2011.02.013.

[9] P. Martin, Wound Healing—Aiming for Perfect Skin Regeneration, (n.d.).

[10] A.D. Luster, R. Alon, U.H. Von Andrian, Immune cell migration in inflammation: present and future therapeutic targets, Nat. Immunol. 6 (2005). doi:10.1038/ni1275.

[11] H. Jin, J. Varner, Integrins: roles in cancer development and as treatment targets INTEGRINS REGULATE CELL SURVIVAL AND MIGRATION, Br. J. Cancer. 90 (2004) 561–565. doi:10.1038/sj.bjc.6601576.

[12] D.A. Lauffenburger, A.F. Horwitz, Cell migration: A physically integrated molecular process, Cell. (1996). doi:10.1016/S0092-8674(00)81280-5.

[13] L.L. Rodriguez, I.C. Schneiderw, Directed cell migration in multi-cue environments, Integr. Biol. Integr. Biol. 5 (2013) 1306–1323. doi:10.1039/c3ib40137e.

[14] C.M. Lo, H.B. Wang, M. Dembo, Y.L. Wang, Cell movement is guided by the rigidity of the substrate, Biophys. J. 79 (2000) 144–152. doi:10.1016/S0006-3495(00)76279-5.

[15] M.J. Paszek, N. Zahir, K.R. Johnson, J.N. Lakins, G.I. Rozenberg, A. Gefen, C.A. Reinhart-King, S.S. Margulies, M. Dembo, D. Boettiger, D.A. Hammer, V.M. Weaver, Tensional homeostasis and the malignant phenotype,(2005).doi: 0.1016/j.ccr.2005.08.010.

[16] F. Gattazzo, A. Urciuolo, P. Bonaldo, Extracellular matrix: A dynamic microenvironment for stem cell niche, Biochim. Biophys. Acta - Gen. Subj. (2014). doi:10.1016/j.bbagen.2014.01.010.

[17] S.L.K. Bowers, I. Banerjee, T.A. Baudino, The extracellular matrix: At the center of it all, J. Mol. Cell. Cardiol. 48 (2010) 474–482. doi:10.1016/j.yjmcc.2009.08.024.

[18] C. Bonnans, J. Chou, Z. Werb, Remodelling the extracellular matrix in development and disease, Nat Rev Mol Cell Biol. 15 (2014) 786–801. doi:10.1038/nrm3904.

[19] A. Malandrino, M. Mak, R.D. Kamm, E. Moeendarbary, Complex mechanics of the heterogeneous extracellular matrix in cancer, Extrem. Mech. Lett. 21 (2018) 25–34. doi:10.1016/j.eml.2018.02.003.

[20] C.J. Malemud, Matrix metalloproteinases (MMPs) in health and disease: an overview., Front. Biosci. 11 (2006) 1696–701. http://www.ncbi.nlm.nih.gov/pubmed/16368548 (accessed October 23, 2018).

[21] M.M.H. Van Marion, Matrix Metalloproteinases and Collagen Remodeling A Literature Review, 2006. http://www.mate.tue.nl/mate/pdfs/7435.pdf (accessed October 23, 2018).

[22] H.K. Kleinman, G.R. Martin, Matrigel: Basement membrane matrix with biological activity, Semin. Cancer Biol. 15 (2005) 378–386. doi:10.1016/j.semcancer.2005.05.004.

[23] K. Wolf, M. te Lindert, M. Krause, S. Alexander, J. te Riet, A.L. Willis, R.M. Hoffman, C.G. Figdor, S.J. Weiss, P. Friedl, Physical limits of cell migration: Control by ECM space and nuclear deformation and tuning by proteolysis and traction force, J. Cell Biol. 201 (2013) 1069–1084. doi:10.1083/jcb.201210152.

[24] P.P. Provenzano, K.W. Eliceiri, J.M. Campbell, D.R. Inman, J.G. White, P.J. Keely, Collagen reorganization at the tumor-stromal interface facilitates local invasion, BMC Med. 4 (2006). doi:10.1186/1741-7015-4-38.

[25] K.R. Levental, H. Yu, L. Kass, J.N. Lakins, M. Egeblad, J.T. Erler, S.F.T. Fong, K. Csiszar, A. Giaccia, W. Weninger, M. Yamauchi, D.L. Gasser, V.M. Weaver, Matrix Crosslinking Forces Tumor Progression by Enhancing Integrin Signaling, Cell. 139 (n.d.) 891–906. doi:10.1016/j.cell.2009.10.027.

[26] M. Binnewies, E.W. Roberts, K. Kersten, V. Chan, D.F. Fearon, M. Merad, L.M. Coussens, D.I. Gabrilovich, S. Ostrand-Rosenberg, C.C. Hedrick, R.H. Vonderheide, M.J. Pittet, R.K. Jain, W. Zou, T.K. Howcroft, E.C. Woodhouse, R.A. Weinberg, M.F. Krummel, Understanding the tumor immune microenvironment (TIME) for effective therapy, (2018). doi:10.1038/s41591-018-0014-x. 61

[27] D. Wirtz, K. Konstantopoulos, P.C.P.P.C. Searson, The physics of cancer: the role of physical interactions and mechanical forces in metastasis, Nat. Rev. Cancer. 11 (2011) 522. doi:10.1038/nrc3080.

[28] P. Friedl, S. Alexander, Cancer Invasion and the Microenvironment: Plasticity and Reciprocity, Cell. 147 (2011) 992–1009. doi:10.1016/j.cell.2011.11.016.

[29] P. Bitter, K.L. Beck, P. Lenz, A mechanical model for guided motion of mammalian cells, Europhys. Lett. 58002 (2015). doi:10.1209/0295-5075/112/58002.

[30] R.J. Pelham, J. And, Y.-L. Wang, Cell Locomotion and Focal Adhesions Are Regulated by the Mechanical Properties of the Substrate, Biol. Bull. 194 (1998) 348–350. http://www.journals.uchicago.edu/doi/pdfplus/10.2307/1543109 (accessed September 5, 2017).

[31] A. Skardal, D. Mack, A. Atala, S. Soker, Substrate elasticity controls cell proliferation, surface marker expression and motile phenotype in amniotic fluid-derived stem cells, (n.d.). doi:10.1016/j.jmbbm.2012.10.001.

[32] F. Gattazzo, A. Urciuolo ⁎, P. Bonaldo, Extracellular matrix: A dynamic microenvironment for stem cell niche ☆, (2014). doi:10.1016/j.bbagen.2014.01.010.

[33] D.E. Discher, P. Janmey, Y.-L. Wang, Tissue Cells Feel and Respond to the Stiffness of Their Substrate, n.d. http://science.sciencemag.org/ (accessed August 14, 2018).

[34] S. Kumar, V.M. Weaver, Mechanics, malignancy, and metastasis: The force journey of a tumor cell, (n.d.). doi:10.1007/s10555-008-9173-4.

[35] J.R. Lange, B. Fabry, Cell and tissue mechanics in cell migration, Exp Cell Res. 100 (2013) 130–134. doi:10.1016/j.pestbp.2011.02.012.Investigations.

[36] D.A. Fletcher, R.D. Mullins, Cell mechanics and the cytoskeleton., Nature. 463 (2010) 485–92. doi:10.1038/nature08908.

[37] F. Kai, H. Laklai, V.M. Weaver, Force Matters: Biomechanical Regulation of Cell Invasion and Migration in Disease, Trends Cell Biol. 26 (2016) 486–497. doi:10.1016/j.tcb.2016.03.007.

[38] F. Broders-Bondon, T. Huong, N. Ho-Bouldoires, M.-E. Fernandez-Sanchez, E. Farge, Mechanotransduction in tumor progression: The dark side of the force, (2018). doi:10.1083/jcb.201701039.

[39] A.L. Bauer, T.L. Jackson, Y. Jiang, Topography of extracellular matrix mediates vascular morphogenesis and migration speeds in angiogenesis, PLoS Comput. Biol. 5 (2009). doi:10.1371/journal.pcbi.1000445. 62

[40] G.P. Raeber, M.P. Lutolf, J.A. Hubbell, Molecularly engineered PEG hydrogels: A novel model system for proteolytically mediated cell migration, Biophys. J. 89 (2005) 1374– 1388. doi:10.1529/biophysj.104.050682.

[41] J. Zeltinger, J.K. Sherwood, D.A. Graham, R. Müeller, L.G. Griffith, Effect of pore size and void fraction on cellular adhesion, proliferation, and matrix deposition., Tissue Eng. 7 (2001) 557–72. doi:10.1089/107632701753213183.

[42] A. Cretu, P. Castagnino, R. Assoian, Studying the Effects of Matrix Stiffness on Cellular Function using Acrylamide-based Hydrogels, JoVE. 42 (2010). doi:10.3791/2089.

[43] D.E. Discher, Tissue Cells Feel and Respon to the Stiffness of Their Substrate, Science (80-. ). 310 (2005) 1139–1143. doi:10.1126/science.1116995.

[44] W.-H. Guo, M.T. Frey, N.A. Burnham, Y.-L. Wang, Substrate Rigidity Regulates the Formation and Maintenance of Tissues, (n.d.). doi:10.1529/biophysj.105.070144.

[45] M. Nikkhah, F. Edalat, S. Manoucheri, A. Khademhosseini, Engineering microscale topographies to control the cell-substrate interface, Biomaterials. 33 (2012) 5230–5246. doi:10.1016/j.biomaterials.2012.03.079.

[46] I.A. Janson, A.J. Putnam, Extracellular matrix elasticity and topography: material-based cues that affect cell function via conserved mechanisms, (n.d.). doi:10.1002/jbm.a.35254.

[47] C.M. Murphy, M.G. Haugh, F.J. O’brien, The effect of mean pore size on cell attachment, proliferation and migration in collagen-glycosaminoglycan scaffolds for bone tissue engineering, Biomaterials. 31 (2009) 461–466. doi:10.1016/j.biomaterials.2009.09.063.

[48] F. Spill, D.S. Reynolds, R.D. Kamm, M.H. Zaman, Impact of the physical microenvironment on tumor progression and metastasis, Curr. Opin. Biotechnol. 40 (2016) 41–48. doi:10.1016/j.copbio.2016.02.007.

[49] S. Kidoaki, H. Sakashita, Rectified Cell Migration on Saw-Like Micro-Elastically Patterned Hydrogels with Asymmetric Gradient Ratchet Teeth, PLoS One. 8 (2013). doi:10.1371/journal.pone.0078067.

[50] P. Roca-Cusachs, R. Sunyer, X. Trepat, Mechanical guidance of cell migration: Lessons from chemotaxis, Curr. Opin. Cell Biol. (2013). doi:10.1016/j.ceb.2013.04.010.

[51] C.-C. Lan, E.Y. Lu, H.-J. Pan, C.-H. Lee, H. Hsu, P.-Y. Hsu, W.-J. Yang, P.-L. Kuo, C.- H. Lee, 12 V J Forrest, Y.H. Kang, D.E. Mcclain, D.H. Robinson, N. Ramakrishnan, 13 H J Sung, Y. Kim, H. Kang, J.W. Sull, Y.S. Kim, S.-W. Jang, J. Ko, Directional migration of cancer cells induced by a blue light intensity gradient, Integr Biol Proc. Natl. Acad. Sci. U.S.A. Opt. Express J.-L. Xiao, T.- Photochem. Photobiol. B Trends Cell Biol. 6 (2005) 1139–1143. doi:10.1364/BOE.6.002624.

[52] D.J. Cohen, W.J. Nelson, M.M. Maharbiz, Galvanotactic control of collective cell migration in epithelial monolayers, (2014). doi:10.1038/NMAT3891. 63

[53] C.-M. Lo, H.-B. Wang, M. Dembo, Y.-L. Wang, Cell Movement Is Guided by the Rigidity of the Substrate, Biophys. J. 79 (n.d.) 144–152. doi:10.1016/S0006- 3495(00)76279-5.

[54] K. Polyak, I. Haviv, I.G. Campbell, Co-evolution of tumor cells and their microenvironment, Trends Genet. 25 (2009) 30–38. doi:10.1016/j.tig.2008.10.012.

[55] C.M. Kraning-Rush, C.A. Reinhart-King, Controlling matrix stiffness and topography for the study of tumor cell migration, Cell Adhes. Migr. (2012). doi:10.4161/cam.21076.

[56] E. Hadjipanayi, V. Mudera, R.A. Brown, Guiding cell migration in 3D: A collagen matrix with graded directional stiffness, Cell Motil. Cytoskeleton. 66 (2009) 121–128. doi:10.1002/cm.20331.

[57] B.N. Mason, C.A. Reinhart-King, Controlling the mechanical properties of three- dimensional matrices via non-enzymatic collagen glycation, Organogenesis. 9 (n.d.) 70– 75. http://dx.doi.org/10.4161/org.24942 (accessed June 25, 2018).

[58] A. Ueki, S. Kidoaki, Manipulation of cell mechanotaxis by designing curvature of the elasticity boundary on hydrogel matrix, Biomaterials. (2015). doi:10.1016/j.biomaterials.2014.11.030.

[59] W.S. Leong, C.Y. Tay, H. Yu, A. Li, S.C. Wu, D.H. Duc, C.T. Lim, L.P. Tan, Thickness sensing of hMSCs on collagen gel directs stem cell fate, Biochem. Biophys. Res. Commun. 401 (2010) 287–292. doi:10.1016/j.bbrc.2010.09.052.

[60] S.S. Rao, S. Bentil, J. DeJesus, J. Larison, A. Hissong, R. Dupaix, A. Sarkar, J.O. Winter, Inherent Interfacial Mechanical Gradients in 3D Hydrogels Influence Tumor Cell Behaviors, PLoS One. 7 (2012) e35852. doi:10.1371/journal.pone.0035852.

[61] D.E. Discher, P. Janmey, Y.-L. Wang, Tissue Cells Feel and Respond to the Stiffness of Their Substrate, (n.d.).

[62] A. Engler, L. Bacakova, C. Newman, A. Hategan, M. Griffin, D. Discher, Substrate Compliance versus Ligand Density in Cell on Gel Responses, 2004. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1303831/pdf/617.pdf (accessed August 29, 2018).

[63] A.J. Engler, L. Richert, J.Y. Wong, C. Picart, D.E. Discher, Surface probe measurements of the elasticity of sectioned tissue, thin gels and polyelectrolyte multilayer films: Correlations between substrate stiffness and cell adhesion, Surf. Sci. 570 (2004) 142–154. doi:10.1016/j.susc.2004.06.179.

[64] D.-H. Kim, D. Wirtz, Predicting how cells spread and migrate: focal adhesion size does matter., Cell Adh. Migr. 7 (2013) 293–6. doi:10.4161/cam.24804. 64

[65] Y.-C. Yeh, J.-Y. Ling, W.-C. Chen, H.-H. Lin, M.-J. Tang, Mechanotransduction of matrix stiffness in regulation of focal adhesion size and number: reciprocal regulation of caveolin-1 and β1 integrin OPEN, (n.d.). doi:10.1038/s41598-017-14932-6.

[66] B.M. Baker, C.S. Chen, Deconstructing the third dimension – how 3D culture microenvironments alter cellular cues, J. Cell Sci. 125 (2012) 3015–3024. doi:10.1242/jcs.079509.

[67] C.D. Hartman, B.C. Isenberg, S.G. Chua, J.Y. Wong, Vascular smooth muscle cell durotaxis depends on extracellular matrix composition, (n.d.). doi:10.1073/pnas.1611324113.

[68] M.R. Ng, A. Besser, G. Danuser, J.S. Brugge, Substrate stiffness regulates cadherin- dependent collective migration through myosin-II contractility, J. Cell Biol. 199 (2012) 545–563. doi:10.1083/jcb.201207148.

[69] K. Ghosh, Z. Pan, E. Guan, S. Ge, Y. Liu, T. Nakamura, X.-D. Ren, M. Rafailovich, R.A.F. Clark, Cell adaptation to a physiologically relevant ECM mimic with different viscoelastic properties, (n.d.). doi:10.1016/j.biomaterials.2006.09.038.

[70] D. Ambrosi, A. Duperray, V. Peschetola, C. Verdier, Traction patterns of tumor cells, J. Math. Biol. 58 (2009) 163–181. doi:10.1007/s00285-008-0167-1.

[71] P.W. Oakes, D.C. Patel, N.A. Morin, D.P. Zitterbart, B. Fabry, J.S. Reichner, J.X. Tang, Neutrophil morphology and migration are affected by substrate elasticity., Blood. 114 (2009) 1387–95. doi:10.1182/blood-2008-11-191445.

[72] B.L. Bangasser, G.A. Shamsan, C.E. Chan, K.N. Opoku, E. Tüzel, B.W. Schlichtmann, J.A. Kasim, B.J. Fuller, B.R. Mccullough, S.S. Rosenfeld, D.J. Odde, ARTICLE Shifting the optimal stiffness for cell migration, Nat. Commun. 8 (2017). doi:10.1038/ncomms15313.

[73] B.L. Bangasser, S.S. Rosenfeld, D.J. Odde, Determinants of Maximal Force Transmission in a Motor-Clutch Model of Cell Traction in a Compliant Microenvironment, Biophysj. 105 (2013) 581–592. doi:10.1016/j.bpj.2013.06.027.

[74] W.-C. Wei, H.-H. Lin, M.-R. Shen, M.-J. Tang, Mechanosensing machinery for cells under low substratum rigidity, Am J Physiol Cell Physiol. 295 (2008) 1579–1589. doi:10.1152/ajpcell.00223.2008.-Mechanical.

[75] K.A. Beningo, M. Dembo, Y. Wang, Responses of fibroblasts to anchorage of dorsal extracellular matrix receptors., Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 18024–9. doi:10.1073/pnas.0405747102.

[76] A.D. Doyle, N. Carvajal, A. Jin, K. Matsumoto, K.M. Yamada, ARTICLE Local 3D matrix microenvironment regulates cell migration through spatiotemporal dynamics of contractility-dependent adhesions, (2015). doi:10.1038/ncomms9720. 65

[77] S.I. Fraley, Y. Feng, R. Krishnamurthy, D.-H. Kim, A. Celedon, G.D. Longmore, D. Wirtz, A distinctive role for focal adhesion proteins in three-dimensional cell motility, (2010). doi:10.1038/ncb2062.

[78] M. Raab, J. Swift, P. Dingal, P. Shah, J.W. Shin, D.E. Discher, Crawling from soft to stiff matrix polarizes the cytoskeleton and phosphoregulates myosin-II heavy chain, J. Cell Biol. 199 (2012) 669–683. doi:10.1083/jcb.201205056.

[79] J.Y. Wong, A. Velasco, P. Rajagopalan, Q. Pham, Directed Movement of Vascular Smooth Muscle Cells on Gradient-Compliant Hydrogels, (n.d.). doi:10.1021/la026403p.

[80] B.C. Isenberg, P.A. Dimilla, M. Walker, S. Kim, J.Y. Wong, Vascular Smooth Muscle Cell Durotaxis Depends on Substrate Stiffness Gradient Strength, Biophysj. 97 (n.d.) 1313–1322. doi:10.1016/j.bpj.2009.06.021.

[81] L.G. Vincent, Y.S. Choi, B. Alonso-Latorre, J.C. Del Álamo, A.J. Engler, Mesenchymal Stem Cell Durotaxis Depends on Substrate Stiffness Gradient Strength, (n.d.). doi:10.1002/biot.201200205.

[82] D.S. Gray, J. Tien, C.S. Chen, Repositioning of cells by mechanotaxis on surfaces with micropatterned Young’s modulus, J. Biomed. Mater. Res. - Part A. 66 (2003) 605–614. doi:10.1002/jbm.a.10585.

[83] S. Kidoaki, T. Matsuda, Microelastic gradient gelatinous gels to induce cellular mechanotaxis, J. Biotechnol. 133 (2008) 225–230. doi:10.1016/j.jbiotec.2007.08.015.

[84] G. Charras, E. Sahai, Physical influences of the extracellular environment on cell migration, Nat. Publ. Gr. 15 (2014). doi:10.1038/nrm3897.

[85] E.A. Novikova, M. Raab, D.E. Discher, C. Storm, Persistence-Driven Durotaxis: Generic, Directed Motility in Rigidity Gradients, Phys. Rev. Lett. 118 (2017). doi:10.1103/PhysRevLett.118.078103.

[86] D. Missirlis, J.P. Spatz, Combined Effects of PEG Hydrogel Elasticity and Cell-Adhesive Coating on Fibroblast Adhesion and Persistent Migration, (2013). doi:10.1021/bm4014827.

[87] P. Grace Chao, S.-C. Sheng, W.-R. Chang, Micro-composite substrates for the study of cell-matrix mechanical interactions, J. Mech. Behav. Biomed. Mater. 38 (2014) 232–241. doi:10.1016/j.jmbbm.2014.01.008.

[88] P.-Y. Wang, W.-B. Tsai, N.H. Voelcker, Screening of rat mesenchymal stem cell behaviour on polydimethylsiloxane stiffness gradients, Acta Biomater. 8 (2012) 519–530. doi:10.1016/j.actbio.2011.09.030.

[89] M. Raab, J. Swift, P.C.D.P. Dingal, P. Shah, J.W. Shin, D.E. Discher, Crawling from soft to stiff matrix polarizes the cytoskeleton and phosphoregulates myosin-II heavy chain, J. Cell Biol. (2012). doi:10.1083/jcb.201205056. 66

[90] J.R. Tse, A.J. Engler, Stiffness Gradients Mimicking In Vivo Tissue Variation Regulate Mesenchymal Stem Cell Fate, PLoS One. 6 (2011). doi:10.1371/journal.pone.0015978.

[91] P. hsiu G. Chao, S.C. Sheng, W.R. Chang, Micro-composite substrates for the study of cell-matrix mechanical interactions, J. Mech. Behav. Biomed. Mater. (2014). doi:10.1016/j.jmbbm.2014.01.008.

[92] B.P. Toole, Hyaluronan: From extracellular glue to pericellular cue, Nat. Rev. Cancer. (2004). doi:10.1038/nrc1391.

[93] M. Wu, M. Cao, Y. He, Y. Liu, C. Yang, Y. Du, W. Wang, F. Gao, A novel role of low molecular weight hyaluronan in breast cancer metastasis, FASEB J. • Res. Commun. FASEB J. 29 (2015) 1290–1298. doi:10.1096/fj.14-259978.

[94] K.L. Schwertfeger, M.K. Cowman, P.G. Telmer, E.A. Turley, J.B. McCarthy, Hyaluronan, inflammation, and breast cancer progression, Front. Immunol. 6 (2015). doi:10.3389/fimmu.2015.00236.

[95] A. Herrera-Gayol, S. Jothy, Effects of hyaluronan on the invasive properties of human breast cancer cells in vitro, 2001. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2517708/pdf/iep0082-0193.pdf (accessed October 1, 2018).

[96] M. Viola, K. Brüggemann, E. Karousou, I. Caon, E. Caravà, D. Vigetti, & B. Greve, C. Stock, G. De Luca, A. Passi, M. Götte, MDA-MB-231 breast cancer cell viability, motility and matrix adhesion are regulated by a complex interplay of heparan sulfate, chondroitin−/dermatan sulfate and hyaluronan biosynthesis, (n.d.). doi:10.1007/s10719- 016-9735-6.

[97] M.I. Tammi, R.C. Savani, J.W. Fischer, Hyaluronan Stabilizes Focal Adhesions, Filopodia, and the Proliferative Phenotype in Esophageal Squamous Carcinoma Cells * □ S Sö ren Twarock, (2010). doi:10.1074/jbc.M109.093146.

[98] S.S. Rao, J. Dejesus, A.R. Short, J.J. Otero, A. Sarkar, J.O. Winter, Glioblastoma Behaviors in Three-Dimensional Collagen-Hyaluronan Composite Hydrogels, (2013). doi:10.1021/am402097j.

[99] P. Li, T. Xiang, H. Li, Q. Li, B. Yang, J. Huang, X. Zhang, Y. Shi, J. Tan, G. Ren, Hyaluronan synthase 2 overexpression is correlated with the tumorigenesis and metastasis of human breast cancer, Int. J. Clin. Exp. Pathol. (2015).

[100] C. Yang, M. Cao, H. Liu, Y. He, J. Xu, Y. Du, Y. Liu, W. Wang, L. Cui, J. Hu, F. Gao, The High and Low Molecular Weight Forms of Hyaluronan Have Distinct Effects on CD44 Clustering * □ S, (2012). doi:10.1074/jbc.M112.349209.

[101] E. Karousou, S. Misra, S. Ghatak, K. Dobra, M. Götte, D. Vigetti, A. Passi, N.K. Karamanos, S.S. Skandalis, Roles and targeting of the HAS/ hyaluronan/CD44 molecular system in cancer Hyaluronan metabolism in cancer, (2016). doi:10.1016/j.matbio.2016.10.001.

[102] T.A. Martin, G. Harrison, R.E. Mansel, W.G. Jiang, The role of the CD44/ezrin complex in cancer metastasis, Crit. Rev. Oncol. Hematol. 46 (2003) 165–186. doi:10.1016/S1040- 8428(02)00172-5.

[103] M.J. Dalby, N. Gadegaard, R. Tare, A. Andar, M.O. Riehle, P. Herzyk, C.D.W. Wilkinson, R.O.C. Oreffo, The control of human mesenchymal cell differentiation using nanoscale symmetry and disorder, (2007). doi:10.1038/nmat2013.

[104] A. Hart, A.E. Nikolaj, G. Ae, C.D.W. Wilkinson, A.E. Richard, O.C. Oreffo, A.E. Matthew, J. Dalby, Osteoprogenitor response to low-adhesion nanotopographies originally fabricated by electron beam lithography, (n.d.). doi:10.1007/s10856-007-0157-7.

[105] A. Curtis, C. Wilkinson, Topographical control of cells, 1997. https://ac.els- cdn.com/S0142961297001440/1-s2.0-S0142961297001440-main.pdf?_tid=99c3cb9f- 76d7-406e-ad6d- 64c4474ae3e5&acdnat=1538609994_71ba7ad0fb122be831fab04d120f03e5 (accessed October 3, 2018).

[106] I.A. Janson, A.J. Putnam, Extracellular matrix elasticity and topography: Material-based cues that affect cell function via conserved mechanisms, J. Biomed. Mater. Res. - Part A. 103 (2015) 1246–1258. doi:10.1002/jbm.a.35254.

[107] D.W. Hamilton, D.M. Brunette, The effect of substratum topography on osteoblast adhesion mediated signal transduction and phosphorylation, Biomaterials. 28 (2007) 1806–1819. doi:10.1016/j.biomaterials.2006.11.041.

[108] A. Mata, C. Boehm, A.J. Fleischman, G. Muschler, S. Roy, Analysis of Connective Tissue Progenitor Cell Behavior on Polydimethylsiloxane Smooth and Channel Micro-Textures, n.d. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1428792/pdf/nihms-3095.pdf (accessed October 3, 2018).

[109] J.J. Norman, J.M. Collins, S. Sharma, B. Russell, T.A. Desai, Microstructures in 3D Biological Gels Affect Cell Proliferation, (n.d.). doi:10.1089/tea.2007.0077.

APPENDIX. MATLAB CODE

[For migration directionality index] function DirectionalityCellAnalysisStepChange214011(home_directory,excel_file,image_fi lename,downsamplingfold,NumberofCells) cellfiber = imread([home_directory filesep image_filename]); sz=size(cellfiber) imshow(cellfiber,[]); %shows the image hold on downsamplingfold = 8;%Every Pixel will have its own vector h1 = impoint(gca,[]); position1 = wait(h1) h2 = impoint(gca,[]); position2 = wait(h2) NumberofCells = 40; cx=[position1(1) position2(1)] cy=[position1(2) position2(2)] cc = [[1; 1] cx(:)]\cy(:); slope = cc(2) %slope of the line intercept = cc(1); %intercept of the line p=[300 400];ix=1500; iy=1500; [x,y]=meshgrid(1:ix, 1:iy); A=[slope -1; 1 slope]; b=[-intercept; slope*p(2)+p(1)]; X=linsolve(A,b); dirx= -(X(1)-p(1))/sqrt((X(1)-p(1))^2+ (X(2)-p(2))^2); diry=-(X(2)-p(2))/sqrt((X(1)-p(1))^2+ (X(2)-p(2))^2); U2 = dirx*ones(ix,iy);%Gives x part of the vector V2 = diry*ones(ix,iy);%Gives y part of the vector imax = log2(downsamplingfold); for i=1:imax U2 = dyaddown(U2,0,'m'); V2 = dyaddown(V2,0,'m'); end quiver(1:downsamplingfold:downsamplingfold*length(U2(1,:)),1:downsamplingfold :downsamplingfold*length(U2(:,1)),U2,V2);

Cellx = xlsread(excel_file, 'Directionality','D2:D24137'); Celly = xlsread(excel_file, 'Directionality','E2:E24137'); Cellx = round(Cellx); Celly = round(Celly); VectorLength = length(Cellx);

for i=1:VectorLength %new array which has radial vector values RadialLocationArray(i) = asind(diry/sqrt(dirx^2+ diry^2)); end

RadialLocationArray = transpose(RadialLocationArray);%reorganize it RLALENGTH = length(RadialLocationArray); 69

CellNumber = xlsread(excel_file, 'Directionality','B2:B24137'); xpos = xlsread(excel_file, 'Directionality','D2:D24134'); ypos = xlsread(excel_file, 'Directionality','E2:E24134');

CellIdentityandDifferences = horzcat(CellNumber, Cellx,Celly,RadialLocationArray);%combine infomration CellIdentityandDifferences = [CellIdentityandDifferences zeros(size(CellIdentityandDifferences,1),1)];%populates the cell array with zeros for area we will need later CellIdentityandDifferences = [CellIdentityandDifferences zeros(size(CellIdentityandDifferences,1),1)];%populates the cell array with zeros for area we will need later CellIdentityandDifferences = [CellIdentityandDifferences zeros(size(CellIdentityandDifferences,1),1)];%populates the cell array with zeros for area we will need later for i=1:NumberofCells %split each cell into its own matrix in cell array

EachCellCellArray{1,i} = CellIdentityandDifferences(CellIdentityandDifferences(:,1)==i,:); end for i=1:NumberofCells % truncate all the info into steps every 12 or 24 %EachCellCellArray{2,i} = EachCellCellArray{1,i}(1:6:end,:); EachCellCellArray{2,i} = EachCellCellArray{1,i}(1:6:end,:); EachCellCellArray{3,i} = EachCellCellArray{1,i}(1:24:end,:); end

for i=1:NumberofCells%arctangent in degrees of cell movement for 12 step rotatingmatrixsize2 = size(EachCellCellArray{2,i},1); for m = 2:rotatingmatrixsize2 x1=((EachCellCellArray{2,i}(m,2)-EachCellCellArray{2,i}(m-1,2))); y1=((EachCellCellArray{2,i}(m,3)-EachCellCellArray{2,i}(m-1,3))); EachCellCellArray{2,i}(m,5) = asind(y1/sqrt(x1^2+y1^2));

end end for i=1:NumberofCells%arctangent in degrees of cell movement for 24 step rotatingmatrixsize3 = size(EachCellCellArray{3,i},1); for m = 2:rotatingmatrixsize3 x1=((EachCellCellArray{3,i}(m,2)-EachCellCellArray{3,i}(m-1,2))); y1=((EachCellCellArray{3,i}(m,3)-EachCellCellArray{3,i}(m-1,3))); EachCellCellArray{3,i}(m,5) = asind(y1/sqrt(x1^2+y1^2));

end end 70

for i=1:NumberofCells %subtracts theatam from theataf of 12 step rotatingmatrixsize2 = size(EachCellCellArray{2,i},1); for m = 2:rotatingmatrixsize2

EachCellCellArray{2,i}(m,6) = EachCellCellArray{2,i}(m,4)- EachCellCellArray{2,i}(m,5);

end end for i=1:NumberofCells%subtracts theatam from theataf of 24 step rotatingmatrixsize3 = size(EachCellCellArray{3,i},1); for m = 2:rotatingmatrixsize3

EachCellCellArray{3,i}(m,6) = EachCellCellArray{3,i}(m,4)- EachCellCellArray{3,i}(m,5);

end end for i=1:NumberofCells%changes the difference between theatas to between 0 and 90 rotatingmatrixsize2 = size(EachCellCellArray{2,i},1); for m = 2:rotatingmatrixsize2 EachCellCellArray{2,i}(m,6) = abs(EachCellCellArray{2,i}(m,6));%absolute value of all the theatas

end end for i=1:NumberofCells%changes the difference between theatas to between 0 and 90 rotatingmatrixsize3 = size(EachCellCellArray{3,i},1); for m = 2:rotatingmatrixsize3 EachCellCellArray{3,i}(m,6) = abs(EachCellCellArray{3,i}(m,6));%absolute value of all the theatas end end k=0; for i=1:NumberofCells%Writes the DI for the 12 step rotatingmatrixsize2 = size(EachCellCellArray{2,i},1); for m = 2:rotatingmatrixsize2 EachCellCellArray{2,i}(m,7) = cosd(EachCellCellArray{2,i}(m,6)); EachCellCellArray{4,i} = (nansum(EachCellCellArray{2,i}(:,7)))/(rotatingmatrixsize2-1) end end for i=1:NumberofCells%Writes the DI for the 24 step rotatingmatrixsize3 = size(EachCellCellArray{3,i},1); for m = 2:rotatingmatrixsize3 EachCellCellArray{3,i}(m,7) = cosd(EachCellCellArray{3,i}(m,6)); 71

EachCellCellArray{5,i} = nansum(EachCellCellArray{3,i}(:,7))/(rotatingmatrixsize3-1- sum(isnan(EachCellCellArray{3,i}(:,7)))); end end

EachCellCellArray{6,1} = vertcat( EachCellCellArray{4,:}); %EachCellCellArray{6,1} = mean(EachCellCellArray{6,1,:});

EachCellCellArray{6,2} = vertcat( EachCellCellArray{5,:}); %EachCellCellArray{6,2} = mean(EachCellCellArray{6,2,:});

end