Mechanobiology of Brain-Derived Cells During Developmental

MECHANOBIOLOGY OF BRAIN-DERIVED CELLS DURING DEVELOPMENTAL

STAGES

GAUTAM MAHAJAN

Bachelor of Engineering in Biotechnology Engineering

Panjab University

May 2013

Submitted in partial fulfillment of requirements for the degree

DOCTOR OF PHILOSOPHY IN APPLIED BIOMEDICAL ENGINEERING

CLEVELAND STATE UNIVERSITY

December 2019

We hereby approve this dissertation for Gautam Mahajan Candidate for the Doctor of Philosophy in Applied Biomedical Engineering degree For the Department of Chemical and Biomedical Engineering and CLEVELAND STATE UNIVERSITY’s College of Graduate Studies by ______Chandra Kothapalli, Ph.D., Dissertation Committee Chairperson, Chemical and Biomedical Engineering, 12/9/19 ______Moo-Yeal Lee, Ph.D., Dissertation Committee Member, Chemical and Biomedical Engineering, 12/9/19 ______Nolan B. Holland, Ph.D., Dissertation Committee Member, Chemical and Biomedical Engineering, 12/9/19 ______Xue-Long Sun, Ph.D., Dissertation Committee Member, Chemistry, 12/9/19 ______Parthasarathy Srinivasan, Ph.D., Dissertation Committee Member, Mathematics, 12/9/19

This student has fulfilled all requirements for the Doctor of Philosophy degree. ______

Chandra Kothapalli, Ph.D., Doctoral Program Director, 12/9/19

November 25, 2019 ______Student’s Date of Defense DEDICATION

This dissertation is dedicated to my mother Poonam Mahajan, my father Inderjeet Mahajan, my brother Gaurav Mahajan, my uncle Bhushan Mahajan, my aunt Madhu Mahajan, my sister in law Rashmi Mahajan, my fiancé Akshata Datar, and my nephews Aayush & Vansh

Mahajan.

ACKNOWLEDGMENTS

First and foremost, I thank my PhD advisor Dr. Chandra Kothapalli for providing me the opportunity to take a lead on this and numerous other projects. He has always been an incredible advisor, and without his mentoring, I could not have done what I was able to do.

I would not have come this far without his constant guidance and support which helped me to develop as a researcher in the best possible way. I would also like to thank Dr. Moo-

Yeal Lee, Dr. Nolan Holland, Dr. Parthasarathy Srinivasan and Dr. Xue-Long Sun, all of whom served on my committee, helped me with research questions as they came up, without which these experiments would not be possible.

I would like to thank Soo-Yeon Kang for her remarkable support throughout the project. I would like to thank Dr. Tatiana Byzova from Cleveland Clinic for the opportunity to work on Kindlin mechanics project (Aim 2). Dr. Tejasvi Dudiki from Cleveland Clinic to provide cells, tissues and constant support during my PhD. I would like to acknowledge the help from Dr. Huan Liu, Dr. Jyotsna Joshi, Sean Moore, Brian Hama, Marissa Sarsfield,

Akshata Datar, Jennifer Vasu, and Dr. Alex Roth. I would like to thank Dr. Pranav Joshi and Dr. Kurt Farrell for guidance and support during the initiation of this project. I would also like to thank Rebecca Laird and Darlene Montgomery and acknowledge other staff throughout the department and college who helped along the way.

This work would not have been possible without the generous financial support from NIH

RO1, NSF CBET, Graduate Student Research Awards (GSRA) and College of Graduate

Studies. I am grateful for the depth of knowledge that I have gained and the skills that I have acquired during my time at CSU. MECHANOBIOLOGY OF BRAIN-DERIVED CELLS DURING DEVELOPMENTAL

STAGES

GAUTAM MAHAJAN

ABSTRACT

Development of nervous system has been greatly explored in the framework of genetics, biochemistry and molecular biology. With the growing evidence that mechanobiology plays a crucial role in morphogenesis, current studies are geared towards understanding the role of mechanical cues in nervous system development and progression of neurological disorders. Formation, maturation and differentiation of various development related cells are sensitive to extrinsic and intrinsic perturbations. Based on this hypothesis, the objective of this study was to investigate the effects of environmental toxicants, mutations in molecular clutch proteins, and matrix stiffness cues on the biophysical, biomechanical, and phenotypic changes in brain-derived neural progenitor cells (NPCs) and microglia.

In the first aim, we established the utility of biophysical and biomechanical properties of

NPCs as indicators of developmental neurotoxicity. Significant compromise (p < 0.001) in

NPC mechanical properties was observed with increase in concentration (p < 0.001) and exposure duration (p < 0.001) of four distinct classes of toxic compounds. We propose the utility of mechanical characteristics as a crucial maker of developmental neurotoxicity

(mechanotoxicology). In the second aim, we elucidated the critical role of molecular clutch proteins, specifically that of kindlin-3 (K3) in murine brain-derived microglia, on the cell membrane mechanics and physical characteristics. Using genetic knockouts of K3 and

AFM analysis, we established the role of K3 in regulating microglia membrane mechanics.

Mutation at the K3-β1 integrin binding site revealed that the connection serves as the major contributor of membrane to cortex attachment (MCA). Finally, in aim 3, we identified the molecular mechanisms (non-muscle myosin II) by which NPCs transduce mechanical input from external substrate into fate decisions such as differentiation and phenotype. We established cell mechanics as a label-free marker of differentiation and mechano- adaptation as possible mechanism of differentiation.

Mechanical, phenotypical, and genotypical characteristics of brain-derived cells and molecular mechanisms of mechanotransduction established in this dissertation will provide important insights in various cell processes such as morphogenesis, brain plasticity, wound healing, cancer metastasis, and disease progression.

vii

TABLE OF CONTENTS

Page

ABSTRACT……….…………………………………………….…………….………….....….....vi

LIST OF TABLES.……………………………………………….……….……….…………...... xii

LIST OF FIGURES…………………………………………….…...……………………………xiii

CHAPTER

I. BACKGROUND AND INTRODUCTION...………………………………..….……...…1

1.1. Role of substrate stiffness……………………...……….……………………….5

1.2. Role of integrins.….……….…………………………………..………………..5

1.3. Role of nucleus………………..……………………………………………….11

1.4. CNS and mechanobiology………….…….…………….……………………...14

1.4.1. CNS mechanical microenvironment…………………………………………..14

1.4.2. Mechanobiology of NSCs/NPCs……………………………….……………...15

1.4.3. Mechanobiology of neurons and neurogenesis………………………………..18

1.4.4. Mechanobiology of glial cells…………………………………………………20

1.4.5. CNS injury, diseases, and biomechanics………………………………………23

1.5. Approaches to study cell mechanics…………………………………………...24

1.5.1. Micropipette aspiration………………………………………………………..25

1.5.2. Optical tweezers……………………………………………………………….27

1.5.3. Magnetic twisting cytometry…………………………………………………..28

1.5.4. Microfluidics…………………………………………………………………..29

1.5.5. Atomic force microscopy……………………………………………………...31

1.6. Mechanical properties of cells…………………………………………………35

1.6.1. Cellular traction forces………….……………………………………………..35

viii

1.6.2. Cellular adhesive forces……………………………………………………….36

1.6.3. Modulus of elasticity…………………………………………………………..37

1.6.4. Tether force……………………………………………………………………38

1.6.5. Cellular deformability…………………………………………………………39

1.7. Scope of the dissertation………………………………………………………40

II. OPTIMIZATION OF AFM PROTOCOLS FOR CHARACTERIZING CELL,

TISSUE AND SOFT SUBSTRATES……………………………………………..……..41

2.1. Utility of MFP-3D-Bio AFM to study mechanobiology….…...…...... 41

2.2. Beading of AFM cantilever tips...…………………………....………….…….44

2.3. Spring constant calibration…...…….……………………………...……..……45

2.4. Characterization of PDMS films………………………………………………46

2.4.1. Experimental methods…………………………………………………………46

2.4.2. Results………………………………………………………………………....47

2.5. Characterization of rat-tail derived type I collagen gels...... …...... 48

2.5.1. Experimental methods…………………………………………………………48

2.5.2. Results…………………………………………………………………...…….49

2.6. Characterization of fixed and fresh mouse retinal tissues……………………..50

2.6.1. Methods………………………………………………………..…..………….50

2.6.2. Results………………………………………………..…………...…………...51

2.7. Characterization of fresh rat spinal cord tissues……………………………….53

2.7.1. Methods………………………………………………………..…………...…53

2.7.2. Results………………………………………………..…………...…………...55

2.8. Characterization of healthy and diseased human SMCs………………………56

2.8.1. Methods…………………………………………………………………….…56

2.8.2. Results…………………………………………………………………………57

2.9. Characterization of human pediatric glioblastoma-derived cells………..……..60

2.9.1. Methods……………………………………………………………………….60

2.9.2. Results…………………………………………………………………...... 60

2.10. Summary…………………………………………………………………...... 61

III. BIOPHYSICAL AND BIOMECHANICAL PROPERTIES OF NEURAL

PROGENITOR CELLS AS INDICATORS OF

DEVELOPMENTAL NEUROTOXICITY……………………………………………...62

3.1. Introduction…………...……………………………..…..……………..……...62

3.2. Materials and Methods……………………….………..….…...……..…..……70

3.3. Results…………..…………………………………….……...……...………...76

3.4. Discussion.………………………………………….…….…………………...89

3.5. Conclusions…………………...…………....……..………………...……...... 101

IV. ROLE OF KINDLIN3 IN MEMBRANE TO CORTEX ATTACHMENT AND

FORCE TRANSMISSION…………………………...……………...……...... 102

4.1. Introduction……………...………………...……….……….…………...... 102

4.2. Materials and Methods……………….………………....…..…..…...... 105

4.3. Results…………………………………...…….....……………….…...... 108

4.4. Discussion……………………………………………………….…………...119

4.5. Conclusions…………………………………………………….…………….123

V. SUBSTRATE STIFFNESS INDUCED MECHANOTRANSDUCTION

REGULATES NPC PHENOTYPE, DIFFERENTIATION, AND MECHANICS…….124

5.1. Introduction…………………...………....……………….…………...... 124

5.2. Materials and Methods……………………………………….…………...... 126

5.3. Results……………………………………….…………………….…………132

5.4. Discussion………………………………………….………………….……..146

5.5. Conclusions……………………………………….……………………….…149

VI. CONCLUSION AND FUTURE DIRECTIONS……………………………………….151

6.1. Conclusions………………………….………………….……………………151

6.2. Future Directions………………………………….……….………………....154

REFERENCES………………………………………………………………………..………...156

APPENDICES…………………………………………………………………………………..192

LIST OF TABLES

Table Page

1. Comparison of different techniques for mechanical characterization of cell…………....25

2. Average values of mechanical properties of microglia measured by AFM…….………112

3. Average values of mechanical properties of macrophage measured by AFM………....119

4. Average values of mechanical properties of NPCs measured by AFM………….……..145

5. Summary of results from individual aims………………………………………………153

xii

LIST OF FIGURES

Figure Page

1.1. Mechanical response of the molecular clutch….………………………………………….7

1.2. Kindlin-Paxillin and integrins regulate outside-in signaling…...……………...…..……...8

1.3. Model of a migrating cell………………………………………………..………………10

1.4. Cortical neurogenesis……………………...…………………………………………….17

1.5. Various experimental approaches to study cell mechanics…………………..………….27

1.6. Microfluidic device to characterize deformability of cells………………...…………….30

1.7. Tip selection for indentation of cells and scaffolds……………………………………...32

1.8. Single force spectroscopy………………………………………….…………………….38

1.9. Microfluidics-based RT-DC device……………………………………………………..39

2.1. Beading of AFM probe………………………………………………………………….45

2.2. Mechanical characterization of PDMS films…………………………………………....47

2.3. Mechanical characterization of collagen gels……………………………………………49

2.4. Mechanical characterization of different layers of fixed retina…....………………….....51

2.5. Mechanical characterization of fresh retina ……………...……………………………..52

2.6. AFM mapping of fresh spinal cord……………………………………………………...55

2.7. Young’s modulus of SMCs and AAA-SMCs……………………………………………59

2.8. Young’s modulus of pediatric glioblastoma-derived cells……..………………………..61

3.1. Mechanisms of NSC toxicity…………………..………………………………………..65

3.2. Dose-response curves of NPCs exposed to rotenone…………………………...... 77

3.3. Biomechanical characteristics of NPCs exposed to rotenone…………………………...78

3.4. Dose-response curves of NPCs exposed to digoxin…………………………………..…80

3.5. Biomechanical characteristics of NPCs exposed to digoxin ...……...…………..………81

3.6. Dose-response curves of NPCs exposed to AEA…………………….………………….82

xiii

3.7. Biomechanical characteristics of NPCs exposed to AEA…. …………………………...83

3.8. Dose-response curves of NPCs exposed to chlorpyrifos…...……………………………84

3.9. Biomechanical characteristics of NPCs exposed to chlorpyrifos…………..……….…...86

3.10. Immunofluorescence images of NPCs under different exposure conditions...…………..87

3.11. Immunofluorescence images of SOX2 expression……………………………………....92

3.12. Comparison of IC50 values from different assays………………………………………..93

3.13. Young’s modulus, adhesion force and tether force of hNPCs…………………………...95

3.14. Correlation of elastic modulus and cell death…………………………………….……...98

4.1. Integrin binding specificity of kindlins……………...……………………………...….104

4.2. Western blot analyses of K3 in microglia…...…………………………………………109

4.3. Young’s modulus of microglia… ……………………………………………………...110

4.4. Histograms of Young’s modulus and tether forces in microglial cells……………..….110

4.5. K3 deficiency and loss in microglial membrane tension…………………...... 111

4.6. Western blot of kindlin expression in macrophages……………………………………114

4.7. Young’s modulus of macrophage cells…………………………………………..…….115

4.8. K3 deficiency and loss in microglial membrane tension…………………...……...... 116

4.9. Histograms of Young’s modulus and tether forces in macrophages cells……………...118

5.1. Graphical summary of the experimental design………………………………………..129

5.2. Mechanical properties of GeltrexTM…………………………………………………….132

5.3. Phase contrast and immunofluorescence images of NPCs after 3 days of on TCPS…..133

5.4. Mechanical characterization of NPCs after 3 days of priming on TCPS………………133

5.5. Phase contrast and immunofluorescence images of NPCs after 3 days on TCPS……...134

5.6. Phase contrast images of cells after 3 days on G-25 & G-100…………………………135

5.7. Immunofluorescence images of cells after 3 days on G-25 & G-100……...…………...135

5.8. Mechanical characterization of NPCs after 3 days on TCPS, G-25 & G-100………….136

5.9. Phase contrast images of cells after 9 days on G-25 & G-100……...………………….138

xiv

5.10. Immunofluorescence images of cells after 9 days on G-25 & G-100……...…………...138

5.11. Phase contrast images of cells after 9 days on TCPS……………………………….….139

5.12. Immunofluorescence images of cells after 9 days on TCPS……………………………139

5.13. Mechanical characterization of NPCs after 9 days on TCPS, G-25 & G-100………….141

5.14. Actin images of cells on G-25 & G-100………………………………………………..141

5.15. YAP-immunofluorescence images of cells on TCPS & G-100………………………...142

5.16. Phase contrast images and mechanical properties of blebbistatin-treated cells………...143

5.17. Immunofluorescence images of blebbistatin-treated cells……..……………………….144

CHAPTER I

BACKGROUND AND INTRODUCTION

One of the major milestones in biological research over the past three decades is the discovery that cells and tissues respond to forces. With the recognition of the fact that cells not only sense biochemical signals but also mechanical signals such as matrix elasticity, force, and geometry, scientists are beginning to explore the role these critical physical forces play in defining morphology, regeneration, and functions of cells and tissues.

Mechanobiology is defined as the science focusing on how physical forces and changes in the mechanical properties of cells and tissues contribute to development, cell differentiation, physiology and disease[1]. It is an interdisciplinary study that relates the biological effects of mechanical forces on cells and mechanotransduction mechanisms of conversion of these forces into biochemical signal cascades[1]. The process of transduction of mechanical stimuli into cellular signals and responses is termed mechanotransduction[2]. Reciprocity of signals is important to maintain homeostasis, which is a feedback between target of signaling and origin of the signals. Signal reciprocity is not limited to just electrochemical signals, as in neurons, but mechano-reciprocity also occur for mechanical signals[3]. Changes in matrix deposition or degradation, or cytoskeleton integrity, or decreasing cellular integrity can significantly alter the

1 environment[4]. These changes in turn can lead to reciprocal response within the cell or generate a passive response in nearby cell population. Any mechanical change in the physiological milieu causes changes in temporal and spatial organization of tissues, suggesting mechanics as means of tissue communication[5]. This implies that study of cellular and tissue mechanics extends our knowledge of the relationship between biochemistry and their physical effects. The aim is to understand how biological and mechanical interactions contribute in development and homeostasis. The knowledge gained could lead to advances in the fields of tissue engineering and regenerative medicine, and a better understanding of developmental biology.

In normal physiology, cells exert and generate mechanical forces depending on the environment they live in. Cellular capability of sensing mechanical forces leads to underlying changes in cytoskeleton which in turn affects their motility and shape, subsequently influencing cell and tissue function[6]. With the help of integrin-based adhesion complexes, cells not only sense biochemical signals, but also respond to the physical attributes such as elasticity of extracellular surroundings[2],[7]. For instance, during embryogenesis, various external and internal forces contribute to mechanical stresses. At the cellular level, the contractile forces generated by the actomyosin cytoskeleton of cell are called ‘internal forces’ and the forces generated outside of the cell responding to the force are ‘external forces’[8],[9]. In X. laevis dorsal non-involuting and involuting marginal zones, cellular internal forces regulating embryogenesis was demonstrated[5]. Cultured explants of these tissues still extend and converge showing that these movements in gastrulation are actively regulated by cellular forces generated in the tissue and not by the external forces generated in different place of embryo[10]. These

2 internal and external forces influence many processes such as embryogenesis, metastasis, would healing, and angiogenesis[11],[12]. The study of these forces at cellular level can be useful for clinical diagnostics by indicating the state of cell health and assisting in developing therapeutic strategies for human diseases. Such measurements extend our knowledge on cellular mechanotransduction under normal and pathologically-aberrant conditions and offers potential to better understand cellular and tissue remodeling within these microenvironments.

Chemical and mechanical cues also play an integral role in regulating cell morphologies, movements, and metabolism[13],[14]. The formation of complex tissue geometries of the organism requires coordinating massive matrix remodeling and movement of cells[4]. There is no doubt on the role of mechanotransduction in these processes. Biomechanical and biophysical characterization of such cells and tissues helps in understanding and exploiting the role mechanobiology plays at different hierarchical levels. At the molecular level, mechanobiology is involved in the control of gene transcription through changes in protein conformation or interaction of proteins[15]. At cellular level, mechanobiology operates through protein complexes and molecular clutches that regulates various processes such as force transmission, cell migration, and transport of materials between and within cells[16], [17]. These molecular and cellular forces subsequently shape cells and tissues. The physical and chemical signals between cells and with ECM determines the multifaceted architecture of tissues such as central nervous system.

One of the least explored application of mechanobiology is in the field of toxicology and drug development. The interplay of mechanical microenvironment and

3 cellular response plays an important role in determining safety and efficacy of lead compounds during drug development. The primary reason for differences in preclinical

(e.g., cell cultures, animal models) and clinical (e.g., in humans) pharmacology is the use of xenogeneic species to test efficacy[18]. When the same targets are explored in human cells the efficacy is different as compared to what is observed in preclinical studies[19].

This complication arises from the differences in behavior of drug target in the assay systems compared with the target system in the subject/patient. The efficacy analysis of a drug is generally quantified in overgeneralized cell cultures where it lacks the mechanical characteristics of pathophysiological conditions[20]. Besides, endpoints in cellular outcomes such as stiffness, deformation, and contraction should be actively incorporated in an ideal assay[21].

It has been known for several decades that cell biophysical changes are important during development and provide great insights in progression of cardiovascular diseases, cancer and infectious diseases[22], [23]. These processes require a comprehensive understanding of dynamic functions such as adhesion[24], polarization[25] and migration[26]. Development and disease progression are associated with structural and physical characteristics along with their functional and biological changes. Several studies have highlighted the use of cells intrinsic mechanical properties as important marker of disease state/progression[27] and therapeutic activity[28]. CNS development and disease progression is a highly complex series of mechanically derived processes at interplay of cells, substrates and mechanotransducers[29]–[32].

1.1 Role of Substrate Stiffness

The integration of intrinsic and extrinsic forces, biochemical factors and microenvironment regulates many processes such as metastasis of tumor cells and differentiation of cells[33], [34]. It has been shown that mechanical signals regulate differentiation of mesenchymal stem cells (MSCs)[35],[36]. It was found that MSCs, when cultured on soft substrates that mimic brain’s stiffness, commit to neurogenic phenotypes.

However, when cultured on stiff (muscle mimic) and rigid (bone mimic) substrates, MSCs showed myogenic and osteogenic commitment, respectively[35]. Lang et al. investigated the effects of increase in substrate stiffness on invasion of breast carcinoma cells into 3-D collagen gels[37], and noted that 3-D cell invasion is enhanced by higher matrix stiffness if the pore size does not fall below a critical value. Mechanobiology provides important insights into the effects of physical forces from within and microenvironment on stem cell fate and cancer metastasis, and can lead to potential therapeutics in regenerative medicine and early detection of diseases.

1.2. Role of integrins

Integrins are transmembranous heterodimers composed of alpha () and beta () subunits, and connects the intracellular actin cytoskeleton with ECM, leading to mechanical integration of extracellular and intracellular compartments[2]. The extracellular domains of integrins are responsible for binding specificity that recognize diverse matrix ligands such as laminin (α2β1, α3β1, α6β1), collagen (α1β1, α2β1), and fibronectin (α5β1, α4β1, αvβ3)[7]. Integrins also recognizes the cell surface receptors such

5 as ICAM-1 (αLβ2, αMβ2) or VCAM-1 (α4β1)[7]. Within the matrix proteins, integrins bind to specific motifs which is evident from the fact that nine different integrins have the potential to bind to fibronectin[2], [7]. Therefore, cell adhesion dynamics and cellular motility have differential effects coming from different integrins that helps cells adhere to fibronectin. For instance, α5β1-mediated integrin adhesion are more dynamic than αvβ3- mediated integrin adhesion that are associated with more persistent migration on fibronectin[38]. Changes in integrin repertoire have been shown to result in changes in integrin-mediated mechano-transduction and migration. For example, expression of αvβ3 is being correlated with the tumor invasion ability of melanomas and expression of α2β1 with metastasis of rhabdomyosarcoma[39].

Intracellular molecules (proteins) such as talin and vinculin are responsible for dynamic coupling of integrin with actomyosin, and belong to the molecular clutch[40].

Integrin activation caused by the conformational change in the integrin ectodomain is the first step in integrin-mediated adhesion, leading to a shift from low to high affinity state for ligand binding. Integrin-ligand binding recruits numerous proteins at the short cytoplasmic tails of the integrins[40]. This leads to the assembly of various adhesion structures of different mechanical properties and protein conformation with differential morphologies and subcellular localization[2].

Molecular clutch is defined as the mechanical linkage formed by the dynamic association between integrin and actomyosin cytoskeleton as shown in Fig. 1.1[41]. Due to the lack of actin binding sites (ABS) in integrin cytoplasmic domains, the clutch is mediated by molecules such as vinculin and talin[42]. Vinculins strengthen the connection/ adhesion between integrin and F-actin facilitated by talin. Other proteins such as kindlins

6 and -actinin also contribute towards the molecular clutch. This clutch is a highly tunable system; on stiff substrates in the presence of high-force transmission, recruitment of vinculin to bind to vinculin-binding sites (VBS) reinforces the clutch. However, on soft substrates, the low mechanical rate deformation on talin is not sufficient to recruit vinculin- based reinforcement of clutch[41].

Fig. 1.1. Model depicting the mechanical response of the molecular clutch and integrin ligand bonds on the soft or rigid ECM. Adapted from Sun et al[2].

Integrin-binding proteins such as talins and kindlins have been shown to be absolutely necessary for integrin activation[40]. Two paralogues of talin expressed by mammals have >80% homology. Talin-1 is expressed by a wide range of cell types whereas talin-2 is expressed by only muscle and neuronal tissues[2]. It consists of an N-terminal

FERM (four-point-one, ezrin, radixin, moesin) domain known as talin head domain (THD).

The long C-terminal rod domain consists of 13 helical bundles (R1-R13) followed by a dimerization motif. The talin rod contains two ABS, 11 VBS, KN motif and ankyrin repeat

7 domains (Kank) and binding sites for regulatory proteins. An intramolecular interaction between R9 domain and THD which masks the integrin-binding site of the THD cause autoinhibited conformation of talin in the cytosol[3]. Convergence of chemokine signaling on Rap1 GTPase and its effector Rap1-GTP-interacting adapter molecule (RIAM) binds and actives talin in leukocytes[2]. For integrin binding activity, the N-terminal FERM is enough. The other domain such as actin-binding rod domain is necessary for cell adhesion and spreading, demonstrating the role of actomyosin-generated forces[40]. Many evidences support the hypothesis of role of talin in force transmission to integrins[7].

Fig 1.2. Kindlin-Paxillin Bridges Integrin and FAK to Regulate Integrin Outside-In Signaling[43].

In cells such as mouse embryonic fibroblasts, talin seems to play a low-key role in force transmission on substrates less than 10 kPa in stiffness as compared to stiffer matrices[44]. Less important role of talin suggests the presence of other F-actin-binding and integrin- binding proteins such as kindlins which contribute to the molecular clutch

8 under different rigidity regimes[45]. Kindlins- 1, 2 and 3 belong to the family of FERM domain-containing proteins. Kindlin-1 is highly expressed by epithelial cells, kindlin-2 is expressed by a variety of cells outside the hematopoietic cells, whereas kindlin-3 is expressed by hematopoietic cells and some nervous system-derived cells[24]. The main function of kindlin is to act as a protein-protein interaction hub by recruiting the integrin- linked pseudo kinase-PINCH-parvin complex, paxillin and the Arp2/3 complex to integrins

(Fig. 1.2). Activation of FAK, Arp2/3 and Rac1 via paxillin-beta-pix axis by kindlin drives the cell spreading[46], [47]. All these reports suggest that kindlin connects integrin with actomyosin and cluster integrins through dimerization. There is growing evidence that kindlins play an important role in the stabilization of cellular adhesion by regulating actin dynamics[45], [46].

During integrin-mediated cell migration, talins and kindlins activate the nucleating multiple ligand-bond which leads to emergence of nascent adhesion (NAs) at the leading edge of cell protrusions[2] (Fig. 1.3). The pathways for the recruitment of adhesion proteins such as vinculin are dependent on the presence of tension. In tension-independent conditions, adhesion proteins are recruited via paxillin, whereas recruitment takes place via talin in tension-dependent manner[48]. The dynamic coupling of NAs and polymerizing branched actin network is through the molecular clutch proteins such as vinculin and talin.

The dynamic coupling helps convert the retrograde movement of polymerizing actin- network into protrusive force at leading edge and rearward traction force on the ECM. In lamella, a small number of NAs matures into large focal adhesions (FAs). Within these

FAs the molecular clutch forms strong attachment to F-actin via binding to the talin and vinculin binding to vinculin-binding sites (VBS). This strong attachment facilitates the

9 transmission of high traction force across integrins leading to formation of catch bond between ligand and integrin. Kank2 is recruited behind the lamella to keep the talin in its integrin-bind state and diminish the F-actin binding to talin. Slip bind formation is characterized by the decrease in force transmission by Kank2. High traction forces to detach the cell rear is applied at the trailing edge FAs at the rear end of the migrating cells[2].

Fig. 1.3. Model of a migrating cell containing diverse integrin-based adhesion structures that transmit different levels of traction forces[2].

1.3. Role of nucleus

The role of nucleus in mechanosensing and genome function is gaining a lot of interest[38]. Nucleus is the largest and stiffest organelle, one of the most important components of the cells, and itself is a membrane-enclosed organelle. Nucleus contains the genetic material which is DNA molecules and protein like histones in a tightly packed structure. Nucleus is composed of nuclear interior which contains chromatin, nuclear bodies, intranuclear elements and nuclear envelope[49]. This nuclear envelope consists of outer nuclear membranes and inner nuclear membranes (ONM and INM). Entry of large molecules into the nuclear membrane is controlled by the nuclear pore complexes (NPCs).

Nuclear lamina, a filamentous protein network, lies underneath the INM and consists of A- type (lamin A is encoded by the LMNA gene) and B-type lamins[50]. LMNA mutations were observed in various diseases such as cardiomyopathy and muscular dystrophy which highlights the role of nucleus as a mechanosensor and importance of studying the role of nucleus in mechanotransduction[15]. The fact that external forces transmit to cell through cytoskeleton is well established, but forces transmitted across the cytoskeleton to the nucleus remains incompletely understood[51].

One of the potential mechanisms of mechanotransduction through nucleus is by modifying chromatin’s physical organization[52]. The non-random organization of DNA within the nucleus is important regulation of transcription and corresponding cellular functions. The presence of heterochromatic DNA towards the periphery of nucleus promotes gene silencing whereas delocalization or repositioning towards the interior leads to gene activation[51]. Therefore, force induced changes in the localization of gene in nuclear periphery could lead to transcription of a particular gene which contributes towards

11 nuclear mechanotransduction[49]. The contributions of biophysical properties of nucleus in regulating the genome function and cytoplasmic scaffold-chromatin assembly coupling is still unknown. Using laser ablation and cellular perturbations, it was found that heterochromatin ablation leads to shrinkage of the nucleus[1]. Experiments involving the depolymerization of microtubules and actin showed that cytoskeleton exerts forces on nucleus. Taken together, changes in cytoplasm govern the nuclear size in a pre-stressed cell[42]. This opens the discussion to investigate the role of nucleus’s biophysical properties in governing genome function.

Studies have shown that if cells are cultured on micropatterned substrates, it leads to changes in nuclear shape and gene expression[53]. It is still unclear whether it is a direct impact of upstream signaling pathways that are sensitive to cytoskeleton reorganization or an effect of change in cytoskeletal force acting on nucleus[50]. It has also been reported that the state of chromatin directly effects the mechanical properties of the nucleus. De- condensation of chromatin increases the deformability of nucleus whereas the chromatin condensation leads to decrease in the deformability of the nucleus[49]. Therefore, changes in the organization of nucleus in the presence or absence of other downstream pathways lead to nuclear deformation, causing abnormal/modified mechanotransduction.

Cellular tension directly regulates the translocations of nuclear regulators such as transcription factors in the process of cellular mechanotransduction or mechanosensing.

The most characterized transcription factors for mechanotransduction are YAP (Yes- associated protein) and TAZ (Tafazzin)[15], [54]–[60]. YAP and TAZ activate the canonical Hippo pathway and localize in the nucleus of the cells cultured on stiff substrates.

The nuclear co-localization of YAP and TAZ leads to a downstream of signaling cascades

12 regulating complex cellular process such as differentiation and migration[55], [56]. YAP and TAZ are also well characterized for their ability to store memory of the past mechanical microenvironment. Studies have shown that stem cells possibly retain information from the past environment, and this influences their fate[54], [57], [59], [60]. This phenomenon, termed as mechanical memory, was first attributed to lineage commitment of human mesenchymal stem cells in a matrix of tunable stiffness. It was concluded that stem cells store information from past physical environment with YAP/TAZ domain acting as an intracellular rheostat and playing a decisive role in stem cell fate[61]. YAP/β-catenin mediated mechanosensitive fate of rat NSCs was shown by using reversible-switching of substrate elasticity determining the temporal window for mechanosensing[62], [63].

Recently, mechanical priming of epithelial cells was shown to regulate mechanical memory dependent migration, in which the mechanical memory was represented by YAP activity[59]. YAP/TAZ acts as nuclear relays of mechanical signals exerted by cell shape and ECM rigidity. Markers staining YAP/TAZ proteins showed the localization of these proteins in nucleus during activated state whereas changes in localization of YAP/TAZ expression from nucleus to cytoplasm shows the deactivated state. The activated and deactivated state signifies the ability of cells to remember their past mechanical state and act upon the new signals[55], [57].

Nuclear mechanosensing is accomplished via several pathways such as protein conformations, nuclear translocation of transcription factors, and membrane dilation, to name a few[49]–[51]. Aberration of these pathways lead to abnormal mechanotransduction eventually leading to disorders.

1.4. CNS and mechanobiology

1.4.1. CNS mechanical microenvironment

Human tissues vary widely in mechanical stiffness ranging from few hundred pascals for soft tissues like brain to 5-10 MPa for stiff calcified bone tissues[30]. CNS and soft tissues inside CNS are protected by bony skull, vertebral column, superficial dura mater, central arachnoid mater, deep pia mater, and serum-like cerebrospinal fluid (CSF) which makes CNS the most protected organ system in the body[30]. CNS tissues are complex and highly-organized structures, with heterogenous and anisotropic ECM containing fibronectin, collagen, laminin but mainly hyaluronic acid[64]. Distinct anatomical structures (brain stem, diencephalon, cerebral hemisphere, and cerebellum) crisscrossed by blood vessels defines a highly-organized brain parenchyma. This defined structure is complemented with an organized and complex neuroarchitecture of gray matter and white matter. Stiffer grey matter is composed of neuron cells bodies and non-neuron brain cells (glial cells) whereas white matter is composed primarily of axon tracts insulated by fatty layer called myelin[64]. Similarly, spinal cord is organized into distinct anatomical structures (ascending and descending motor and sensory white matter tracts).

Longitudinally-oriented heterogenous spinal cord consists of butterfly-shaped gray matter in the center, surrounded by white matter tracts[30], [63], [64], providing a unique set of biochemical and biomechanical cues to cells residing in the brain and spinal cord.

Brain is a perfect example of a viscoelastic material which shows relatively small differences among mammals[65]. Most of the biomechanical studies characterized stiffness of brain post-mortem, due to tissue accessibility limitations[65]. With aging, there is an increase in lipid content, which is developmentally related to dendritic arbors and

14 myelination of rapidly branching axonal shafts and a decrease in water content. The nonlinear stress-strain relationship of the CNS tissue was found to be surprisingly preserved even after post-mortem and post fixation[63]. Such nonlinear stress-strain relationship of the CNS tissue was similar to the characteristics of collagenous soft tissues and forms the basis for models mimicking the remodeling and traumatic injury. However, their role in cellular mechanobiology remains incompletely understood[63].

1.4.2. Mechanobiology of NSCs/NPCs

Neural stem cells (NSCs)/ neural progenitor cells (/NPCs) are self-renewing, multipotent cells which can differentiate into neuronal and glial lineages. NSCs are of quiescent phenotype with an ability to proliferate, migrate and differentiate upon receiving exogenous signals from their microenvironment. NSCs have been shown to maintain their undifferentiated state in the presence of basic fibroblast growth factor (bFGF) and epidermal growth factor (EGF), and removal of these mitogens usually leads to their differentiation into progenies[66],[67]. During CNS development, a large population of

NSCs exist in the ventricular zone (VZ) and subventricular zone (SVZ), but sharply decline in SVZ with maturation (i.e., aging). Using rat embryogenesis model, it has been shown that the large population of NSCs present in neural crest and neural spinal tube reduce to ≤

1% postnatal. These studies highlight the importance of NSCs and their microenvironment in proper embryonic development[68], [69].

Various cellular properties such as motility, matrix remodeling, differentiation, proliferation, and cell polarity are directly influenced by the mechanical properties of the

ECM[30]. Gastrulation is the process of dynamic orchestration of cellular migration and differentiation leading to physical reorganization of a single sheet of embryonic cells into

15 three germ layers: ectoderm, mesoderm and endoderm[10]. Gastrulation is followed by organogenesis, where cells with the three germ layers differentiate into early tissue and then organs[70]. During development, millions of resident cells in the proliferative tissues of brain migrate and differentiate into specialized cells to establish highly-organized structures with distinctive internal architecture and shape[71]. During embryogenesis,

NSCs/NPCs are responsible for the formation of brain, whereas in adults, their role is limited to memory and learning with no contribution in repair and regeneration[69].

The development of CNS begins with the invagination of the neural plate to form neural tube. The neural tube consists of a single layer of neuroepithelial cells, the most primitive form of NSCs as shown in Fig. 1.4[72]. The U-shaped structure pinched off to form neural tube – the primitive form of CNS, with the remaining cells on the outside of this tube migrating to form peripheral nervous system (PNS)[73]. Prior to these cellular rearrangements and migration activity, cells require epithelial-mesenchymal transition

(EMT)[30]. During EMT, cells undergo transition for collective static phenotype to individual migratory phenotype. Once cells reach their predefined location, they undergo mesenchymal-epithelial transition (MET)[30]. EMT and MET is accompanied by changes in cytoskeleton reorganization and cell-to-cell and cell-to-ECM contact.

During neurulation (neural plate to neural tube transformation), tissue level mechanotransduction leads to cycles of EMT and MET by alteration in cell-cell and cell-

ECM contact, and cytoskeleton rearrangement[70]. Closure of neural tube requires stiffening of dorsal tissues caused by actomyosin-driven contraction[74]. Neural tube defects occur due to abnormal progression of mechanically-regulated processes such as cellular migration from the neural crest and improper cellular adhesion in neural folds[74].

Fig. 1.4. Cortical neurogenesis in the mouse[75].

Embryogenesis is a process which progresses within specific matrices with mechanical composition facilitating development and differentiation of specific cell types[76]. During nervous system development, NSCs give rise to temporal and regional specified neural progenitors. These neural progenitors first give rise to large projection neurons and later glia and small interneurons[68], [72]. The remaining small population of

NSCs in the adult brain slowly proliferate and produce neurons if the need arises (e.g., injury, disease), mainly for replacement of cells in olfactory bulb and hippocampus. These adult stem cells maintain their ability of self-renewal based on their tissue-specific niche or environment.

During development, the composition of SVZ changes over time and relative profusion of proteoglycans, collagens, and glycoproteins maintain the differentiation and

17 self-renewal ability of NSCs[68]. Various extracellular signals (biochemical and biomechanical) and intracellular signals are responsible in deciding the fate of these cells.

Innate stiffness of niches and biophysical properties of cells and tissues play an important role in deciding cell fate. Recently, using AFM as a characterization tool, it was reported that there is a gradual increase in stiffness of SVZ during embryonic development[77], although the bulk elastic modulus was found to remain constant during development and postnatally. Any perturbation in regulation of NSCs/NPCs would lead to various diseases ranging from psychiatric disorders to neurodegenerative disorders to glioblastomas[74],

[78]. With the advancements in the field of regenerative medicine, NSCs/NPCs show potential in applications such as cell-based therapies and neurotoxicant testing. NSC cultures provide in vitro models to enhance our knowledge about nervous system development and diseases. NSC based therapies are also being explored for nervous system injury and neurodegenerative disorders[30], [79], [80].

1.4.3. Mechanobiology of neurons and neurogenesis

One of the major mechanically relevant events during nervous system development is formation of neurons from NSCs/NPCs[16]. NSCs/NPCs and radial glial cells differentiate into neurons and in later stages into glial cells[16]. It has long been established that mechanosensing is one of the major functions of neurons. Neurons such as somatosensory neurons are one of widely studied cells for their mechanotransduction potential, which transduce sound and tactile cues[81]. It was uncertain for many years if differentiated neurons which are protected to mechanical damage hold the same level of awareness and sensitivity to mechanical niche[81], [82]. Several groups reported that

18 mature neurons in the CNS respond to mechanical cues as dynamically as cells in other tissues which experience regular mechanical loading[16], [81].

Explanting CNS tissues onto substrates with mechanical properties mimicking that of CNS tissues supports optimal growth of both neuronal and glial population[65], [83].

Softer substrates support neuronal growth while stiffer matrices support growth and proliferation of glial lineages. Studies have reported that mimicking neurite extension and branching in vitro on substrates that are too soft or stiffer than the CNS tissues is very difficult[63]. The interplay of substrate geometry, cell source, and composition might be responsible for this regulation of neurite outgrowth[84]. The Young’s modulus of reactive astrocytes following injury was ~ 2-fold lower than that of naïve cultured astrocytes[85].

This softening of reactive astrocytes in the injury area provides an appropriate complaint substrate which stimulates neurite extension[85]. Application of retinoic acid (RA) to induce neurite extension and reduce proliferation is strongly dependent on ECM stiffness[84]. Process of neurite extension exerts mechanical stress on ECM substrate which requires ECM substrate to be compatible to support generation of contractile forces within extending projections.

It has been shown that cytoskeleton dynamics impact the directionality of the axonal growth[86]. Biochemical modification of ECM and cytoskeleton lead to changes in neurite mechanics. Perturbation of developing Xenopus brains by manipulating the ECM composition using chondroitin sulfate led to changes in mechanical properties[30], [65], although the stiffness gradient was found to be similar to that of untreated developing

Xenopus brain (control). It was found that the axons in these softened brains lack

19 directionality and fasciculation with axons dispersed widely than their normal trajectory.

This behavior was similar to axon behavior on soft substrates in vitro[87].

Organization of neurofilaments networks within axons was dependent on the repulsive forces between phosphate groups on neurofilament sidearm domains, showing the contribution of cytoskeleton networks to the shape and mechanics of mature axons[88].

It has been reported that traumatic nerve injury leads to changes in organization of neurofilaments and other cytoskeleton proteins[89], [90]. Microtubule-associated proteins by similar mechanisms cause organization of microtubule bundles found in axons and dendrites[91]. Modulation of actin cytoskeleton such as activation of myosin motors and

Rho-family GTPases play an important role in regulation of axon growth dynamics[86]. It was noted that dysfunctional Rho family GTPase leads to development of various adult and congenital neurological disorders[63]. Factors such as mechanosensitive ion channels govern the neurite growth kinetics. In neurons explanted from Xenopus laevis spinal cord tissue, inhibition of Ca2+ influx through stretch-activated ion channels lead to a dramatic growth of neurite extension[63].

1.4.4. Mechanobiology of glial cells

Non-neuronal cells commonly known as glia, came from Greek word “glue”, act as support cells in regulating and maintaining CNS function[16]. Glial cells perform a variety of functions such as axon myelination by oligodendrocytes, secretion and cilia-driven circulation of CSF by ependymal cells, and structural support by astrocytes[68], [92]. With the advances in biophysical techniques, glial cells are no longer stereotyped as the glue of the nervous system[83]. For instance, astrocytes are shown to be involved in regulation of

20 adult neurogenesis, plasticity and organizing host response to injury, homeostasis maintenance, and blood-brain barrier maintenance[16], [63], [78].

AFM analysis revelated that glial cells are softer than neurons[83], questioning the function of glial cells to provide good structural support. Similarly, it was noted that elastic forces are more dominating than viscous forces, questioning the glue ability of glial cells[16], [83]. In situ tensile testing of spinal cord explants and those from disrupted glial matrix highlighted the significant contribution of glia to the spinal cord[64], [93]. The compliance of glial cells as compared to other cells in CNS might allow them to protect neurons by cushioning during any traumatic event[64].

One of the many cell types in CNS, microglia is particularly well positioned to coordinate both mechanosensing and tissue organization[92]. Originating from yolk sac, microglia populate neural tissues early in development prior to the formation of blood vessels and help shape the neural circuits[16]. In adults, the key function of microglia is tissue surveillance and defense, since it is the only cell type constantly moving in the CNS and thereby encountering changes in microenvironment[94]. Interactions of microglia with its microenvironment and other cells are often mediated by cell adhesion receptors, including integrins[68]. It has also been hypothesized that astrocyte activation may be the result of the architectural disruption[27], [63]. Abnormal cell number and abnormal cell size known as hyperplasia and hypertrophy, respectively, are two main events during astrogliosis/reactive gliosis[16], [63]. Astrogliosis is the CNS response to the many neurological disorders and traumatic events leading to scar formation. Astrocytes become activated during events of mechanical deformation such as stress due to trauma, mass effect of a tumor, buildup of fluid following tissue insult, and increasing pressure due to

21 edema[94]. Astrocytes activation and action is a highly synchronized event of biochemical and biomechanical communication[90].

Astrocytes plays an important role in mechanosensing; they directly convert mechanical signals into chemical signals using stretch activated and inactivated ion channels[65]. Transmembrane influx of calcium which propagates to neighboring cells as a transcellular wave in a confluent culture, has been demonstrated using micropipette indentation of astrocytes at a single cell level[95],[96]. This calcium wave signaling acts as an important mechanism to organize functions such as gene expression, proliferation, gene expression, and guidance of growth cones[63]. Similar effects were observed with chemical stimuli such as glutamate which leads to initiation and propagation of calcium signaling[63].

In addition to regular mechanotransductive pathways, cellular contractile machinery via gap junctions regulate calcium signaling[97], along with cytoskeleton[97].

It was observed that underdeveloped cytoskeleton was unable to propagate calcium signaling in neonatal astrocytes despite the presence of extensive gap junctional coupling[98]. The radius of propagated calcium waves is found to be proportional to the number of cells with well-defined and defined actin cytoskeleton[98]. Inhibition of myosin light chain kinase activity and actin depolymerization not microtubules disruption using active pharmaceutical ingredients lead to suppression in propagation of calcium wave[99].

Mechanobiological characteristics of glial cells are typical of cells found outside the CNS. Glial cells respond to mechanical stress by reorganizing their actin cytoskeleton and intermediate filaments[92]. Astrocyte syncytium helps transmission of mechanical signals over a long range which plays a vital role in CNS injury and pathogenesis[63].

1.4.5. CNS injury, diseases, and biomechanics

Neurodegenerative disorders are characterized by loss in myelin content, change in the ECM, loss of adult neurogenesis[100]. These factors play a role in loss of brain stiffness as compared to increase in stiffness with age[101]. Aging causes slowing of CSF flux through the brain which affects the cellular functions dependent on shear flow[101]. Even after decades of establishment that aging and diseased states show changes in biomechanics, it is still not clear to what extent these changes drive pathology[30]. CNS diseases such as Alzheimer’s are characterized by ECM loss which directly or indirectly leads to neuronal and synaptic loss[100]. These changes in ECM loss are accompanied by regional changes in brain stiffness, documented using non-invasive techniques such as three-dimensional magnetic resonance elastography[74]. Parkinson’s disease is characterized by changes in stiffness/ elasticity of substantia nigra which can be detected before the motor impairment[18], [74]. Elucidating the reasons for these changes and how these changes impact cellular mechanotransduction might provide insights on better approaches to diagnosis and treatment.

Extensive work has been done in the field of CNS tissue biomechanics owing to a spike in sports injuries, automobile accidents, and developmental disorders[16]. The field of cellular mechanobiology in the CNS is comparatively less explored. NSCs and differentiated progenies are involved in morphogenesis of CNS tissue, which requires a spatial and temporal control of processes such as cell division, differentiation, migration and apoptosis[62], [69]. The study of mechanobiology of CNS is multi-disciplinary in nature and is at interface of engineering, medicine, physics and biology[42]. The primary objective of this work is to investigate the role of biomechanics on the development of the

23 nervous system using atomic force microscopy (AFM) techniques. The primary objectives are to uncover the mechanisms that enable cells to sense, transduce and respond to mechanical and chemical stimuli, as well as to characterize the mechanical properties of molecules and cells. The outcomes from such research could provide important insights into the pathogenesis of neurological disorders (e.g., multiple sclerosis, Alzheimer’s and

Parkinson’s disease) and lead to development of novel tools and techniques for early detection, diagnosis, and treatment.

1.5. Approaches to study cell mechanics

Recent advances in soft matter physics and mechanical engineering paved way for understanding the cell mechanics by applying various experimental methods and theoretical models[102]. Various techniques have been developed to study the cell biomechanics, which unravel how the cells feel under transformations and perturbation due to mechanical forces as shown in Fig 1.5[23]. Micropipette aspiration, optical tweezers, magnetic twisting cytometry, microfluidics, atomic force microscopy (AFM), top the most widely used techniques, and have been used over the years for mechanical characterization of cells and tissues[103]. Table 1 summarizes the pros and cons of various techniques for mechanical characterization of cells.

Table 1: Comparison of different techniques for mechanical characterization of cells. Techniques Advantages Disadvantages References Micropipette Simple, inexpensive, Cause damage of the [104]–[106] Aspiration force measurement as sample low as piconewton

Optical tweezers Constant stress state Overheating of cells, to cell, contamination low throughput, [107]–[110]

free measurements limited force range

Magnetic twisting Attachment of bead to Contamination free [22], [111] cytometry measurements, non- sample, size of the bead invasive method Microfluidics Simple concept, no Clogging of the bulky set up, channel, time [22], [104], inexpensive, minimal consuming, air [105], [112] experience required, bubbles and leaking perfect for problems, neglecting deformation analysis effects of viscosity and cell membrane

Atomic force Indentation depth, Specific sample [102], [113]– microscope young’s modulus, preparation, requires a [121] adhesion forces, skilled worker, indentation phase requires noise free angle, loss modulus conditions and storage modulus

1.5.1. Micropipette aspiration

Micropipette aspiration is one of the earliest experimental approaches applied to study cell mechanics which opened research in this area. Micropipette is a rigid tube generally made of glass which narrows down to a small diameter at the tip. When suction is applied and micropipette is brought closer to the cell, the negative pressure causes cell to form a seal. This leads to cell drawn into the micropipette forming a protrusion. When the pressure applied exceeds the threshold pressure, there is continuous deformation of cell

25 into the micropipette as shown in Fig. 1.4c. This principle explains the use of liquid-drop model, which considers cell as a Newtonian viscous fluid, with its outer cover under constant surface tension. By applying liquid-drop model, the micropipette aspiration can be analyzed using laws of Laplace. Laws of Laplace relates the difference between the inside and outside of a thin pressure walled vessel with the surface tension within the vessel wall[105],[122],[123]. By using this, we can relate the suction pressure to the morphology of the cell. Using this model, the cells show instability when the radius of the pipette and the radius of the aspirated protrusion are equal[105],[122].

Using micropipette aspiration, mechanical properties of neutrophils were compared to endothelial cells and chondrocytes. Neutrophils showed elasticity in the range of 100 Pa, as compared to the endothelial cells and chondrocytes which showed elasticity in the range of 0.5-0.6 kPa[123]. In another study, the effect of cholesterol depletion on membrane deformability of bovine aortic endothelial cells was measured using micropipette aspiration. The results suggested that altering the properties of the submembrane F-actin and cholesterol depletion caused an increase in the stiffness of the membrane of endothelial cells[124].

Despite its simplicity, ease of use, and ability to visualize the cells during deformation, a basic limitation of using the model in micropipette aspiration experiments is ignoring the friction between the micropipette and cell membrane. The techniques used to create the negative pressure in these experiments are susceptible to vibration and temperature changes. The open environment working of nature of micropipette aspiration measurements are affected by the evaporation causing fluid loss which necessitates regular recalibration of the device[123],[124]. Besides, modulating the relatively low suction

26 pressure also requires sophistication and only one cell could be interrogated at a single time, making it a time-consuming process.

Fig. 1.5. Various experimental approaches to study cell mechanics [adapted from Bao et al[125]].

1.5.2. Optical tweezers

Optical tweezers were originally developed to study the individually trapped atoms, bacteria, and viruses. It utilizes a microscope and an infrared laser to trap and control the motion of a single molecule. Based on the refractive index of the sample, there is a change in the direction of photons when they pass through the sample. This change in direction causes a change in the momentum of the sample leading to generation of a force in the sample[108], [109]. For the single cell studies (Fig. 1.5d), a small bead is used, and the trapping force is calculated using the refractive index of culture medium, refractive index of the bead, and the intensity gradient of laser[105],[108], [109], [122].

In a study using single beam optical tweezer experiments, bone cells were mechanically stimulated to observe the calcium channel response. A small 7 pN force was applied using an optical tweezer on rat and human osteoblasts, and calcium levels were observed using Fluo-3 labelling. Results suggest that rat cells had less increase in intracellular calcium levels compared to the human osteoblasts. It also showed that stimulating different locations of cells show differential increase in the levels of calcium[109].

Optical tweezers provide contaminant-free measurements by applying a constant stress state to cell. By using optical tweezers, the limitations associated with the physical contact-based techniques can be overcome[108], [109]. The limitation of using optical tweezers include overheating of the cells which could change the mechanical properties of cells in focus when exposed for a long time, and confinement of the analysis to one or two samples at a time[108].

1.5.3. Magnetic twisting cytometry

To study the rheological properties of the cells, magnetic twisting cytometry is an important and widely used technique (Fig. 1.5b). The quantification of the rheological properties depends on the characteristics of the ferromagnetic beads, which can be placed intracellularly by injection or phagocytosis, or bound to the cell membrane using ligand- receptor binding. Weak magnetic field is used for the rotation of these magnetic beads; both angular rotation of beads and torque are quantifiable. Various available models[122],

[126] are applied to these angular rotation and torque data to obtain important mechanical properties such as elastic modulus.

In a study using magnetic twisting cytometry, the responses of individual human airway smooth muscle cells to contractile and relaxing agonists were measured[126]. The experiment used a novel optical detection system and the ferromagnetic beads were bound to the integrins on cell membrane. The bead mediated rotation was recorded using phase- synchronized video camera to measure and quantify the mechanical characteristics such as storage and loss moduli. The study also quantified these mechanical properties at the muscle tissue level and concluded that these mechanical properties were in line but was on lower side as compared to the intact tissue. These cell properties characterized using magnetic twisting cytometry showed clear although heterogeneous responses to contractile and relaxing agonists[126].

The magnetic twisting cytometry offers advantages such as non-invasive and low- contamination measurements. There are certain limitations associated with using this technique: the size of the magnetic beads must be large enough to compare with the sample of interest, and the expertise required to find a way to bind these beads to the cell membrane for non-invasive approach[105],[126].

1.5.4. Microfluidics

Microfluids are being used for processing and manipulation of fluid on a microlevel scale; considering this size advantage, it is being used in the analysis of single cell analysis.

Various types of microfluidic techniques such as constricted geometry, fluid stretching, and optical stretching are available for single cell mechanical properties analysis. In optical stretching, two laser beams are used for serial stretching the cell flowing in a microfluidic channel to quantify the cell mechanics. In fluid stretching, the variation in the geometry of

29 the channels provide the deformation force to the cells and subsequently utilized for mechanical properties characterization[122], [127], [128].

In a study using constricted geometry microfluidics, tumor cells were characterized for repeated deformability with and without taxol treatment[129]. The study concluded that the cells treated with taxol required more transit time than the untreated cells when travelling through first constriction (Fig. 1.6). This difference in transit time decreased with the long-term travel through many constrictions. There are many models available which can be used to convert the transit time to the commonly used mechanical terms as Young’s modulus.

Fig. 1.6. Schematic of device and operations used to characterize the study the deformability of cells with and without taxol treatment[129].

Despite its simplicity of use and other advantages, there are some limitations associated with the use of microfluidics for the characterization of mechanical properties of cells. Some of the commonly encountered problems include clogging of the channel, duration of experiments, air bubbles, and leaking problems. The major limitation is the whole experiment is based on neglecting the effects on viscosity and cell membrane[105].

1.5.5. Atomic force microscopy (AFM)

The unique capabilities of AFM, such as minimal sample preparation, non-invasive characterization, precise control on the magnitude and frequency of the applied forces, and high spatial control, makes it the most suitable technique to image and mechanically characterize either fixed or live cells or tissues[130], [131]. This technique provides information on cell membrane structure, cell topography, and mechanics of cells under physiologically-relevant conditions[105].

AFM utilizes a tip microfabricated on a cantilever and an optical detection system for the structural and mechanical characterization of samples (Fig. 1.5a). In AFM, a laser beam is reflected at the back of the cantilever and falls onto the position set on the photodiode; this setup is used to determine the changes in the deflection with respect to the position of the laser on interacting with a sample[120]. During AFM studies, some of the sample variables such as temperature and pH, could be controlled which makes it the best available tool to study mechanical and structural properties under physiological mimicking-conditions.

To calculate the stiffness and Young’s modulus, the force-distance curves generated by AFM are used [20],[120] . To generate the force curves, the tip is moved

31 towards the sample until it makes a contact and experience a repulsive force due to interaction with the sample. The amount of indentation can be preset by either fixing the amount of z-distance the tip will travel or fixing the amount of maximum force to be applied to the sample. To generate an appropriate force-distance curve, the sensitivity and stiffness of the cantilever are very important. The main factor to attain these is using a cantilever with appropriate spring constant. It is very important to have a good match between the sample stiffness and spring constant of the cantilever[31],[132]. This allows the conversion of force from deflection using Hooke’s law. If the cantilever is too soft as compared to the sample, it will cause bending of the cantilever which will generate data of no use. On the other hand, if cantilever is too hard, there will be less deflection than it should be, which consequently affects the data. Based on the need, it is possible to modify the geometry and size of the tips of the cantilever to analyze different areas of sample (Fig.

1.7). For biological samples such as cells or cell-seeded scaffolds, it is recommended to use spherical probes as there are fewer chances of damaging the samples compared to a sharp tip[133],[134].

Fig. 1.7. Schematic illustrating tip selection for indentation of cells and cells-seeded scaffold[135]

AFM can determine the mechanical modulus ranging from few tens of Pa to GPa.

From the force curves, a variety of information can be generated, such as indentation depth,

Young’s modulus, and adhesion forces. By applying the dynamic drive signal during indentation, phase angle, loss modulus and storage modulus can be calculated. Using these parameters and applying a proper fit model, it is possible to calculate viscosity, reduced modulus and equilibrium modulus[131]. AFM provides a unique platform to present the data in the best possible way. For example, it is possible to overlay the topographical scan image onto the force map of the same area to visualize the change in mechanical modulus over the surface[132],[136]. The use of AFM in single cell analysis is gaining attention.

AFM offers a unique advantage to perform the measurements in the cell culture conditions[102]. AFM is equipped with the technologies required to maintain the temperature at 37 °C, and to maintain the growth environment for the cells. The use of optical (or epi-fluorescence) microscope in conjunction with AFM allows simultaneous visualization of the fluorescent labeling of cell organelles to better understand their role during biomechanical characterization[133],[115].

In a recent study where benign and metastatic tumor cells were separated based on their mechanical properties, metastatic tumor cells showed less modulus of elasticity compared to benign cells[26]. The cell mechanics was characterized using AFM in that study, and the metastatic cells showed a decrease in adhesive forces relative to benign cells.

In another study, the effect of paclitaxel on cancer cells was investigated using AFM[113].

The anticancer activity of paclitaxel was observed using AFM on HeLa and Ishikiwa cells.

The anticancer activity leading to apoptosis of cells were correlated to changes in mechanical properties of cells. A decrease in elasticity and increase in roughness was

33 observed when cells were exposed to paclitaxel. These measurements help in evaluating the anticancer activity of drugs. There are various diseases that are known to lead changes in cellular mechanics.

The study of external factors such as toxicants to elucidate the biophysical effects at a single cell level is a relatively new field. In one of the studies, the biophysical effects of diesel exhaust particles (DEPs) on human aortic endothelial cells were characterized using AFM[133]. The changes in elasticity were correlated to expression of cytoskeletal protein; however, the biochemical effects of DEP were not studied. In a similar study, the cytotoxic and mechanical effects of carbon-based nanomaterials on fibroblasts were evaluated[137]. Exposure of graphene flakes lead to a 20% decrease in Young’s modulus compared to that in control fibroblasts. It was also noted that exposure of multiwalled carbon nanotubes caused a 10% decrease in Young’s modulus than in unexposed cells. The effects of these exposures were correlated to changes in disruption of actin fibers ( 50%).

It was noted that such low-level exposure of these carbon-based nanomaterials lead to 20% cell death. However, the study did not correlate the changes in biophysical and biochemical properties of cells[137].

In a separate study, the effects of Bordetella pertussis adenylate cyclase toxin

(CyaA toxin) on integrin-mediated adhesion and mechanics in alveolar epithelial cells were characterized using multiple-bond force spectroscopy with an AFM[138]. It was observed that even at very low CyaA concentration (0.5 nM) with minimal effect on cell viability (>

95%), there was a significant increase in Young’s modulus of alveolar epithelial cells as compared to controls. Inactive variant CyaAE5 has no effect on mechanics of alveolar epithelial cells. It was also noted that the number of integrin bonds was reduced to half

34 upon exposure to CyaA. The effects of amyloid beta oligomers (Aβ40 and Aβ42) on neuronal elasticity responsible for various neurotoxic mechanisms relevant to Alzheimer’s disease pathology was quantified using AFM[139]. It was found that exposure of amyloid based toxins affected the elasticity of neurons in an age-dependent manner. Among the young, mature, and aged neurons treated with Aβ species for a short duration, aged neurons showed the largest decrease in Young’s modulus.

1.6. Mechanical properties of cells

1.6.1. Cellular traction forces

Quantification of cellular traction forces is one of the most used terminologies to define cell mechanics, dating back to 1980’s. The simplest way to measure cell traction forces is using ultrathin film approach[140]. Traction force of the cells cultured on these films were quantified by examining the wrinkling of the film[141]. Despite its heavy use, the complexity of the quantification of traction forces from such wrinkles led to use of more accurate fluorescent-bead approach. Embedding fluorescent beads in PAA gels provides more accurate measurements of traction force by measuring the displacement of beads. In one such study, traction force of focal adhesions was quantified from a single migrating 3T3 fibroblast cultured on collagen conjugated PAA gels embedded with fluorescent beads[142]. It was reported that propulsive thrust for fibroblast locomotion was

~ 0.2 dyn.

With the advances in material sciences and imaging technologies, utility of

BioMEMS in calculating traction forces is gaining attention. Use of flexible pillars coated

35 with ECM matrix to culture cells and measuring deflection of the pillars captured by high- resolution cameras are widely used to measure traction forces[143]. Using microcantilever micropads and flexible substrate approach, traction force for fibroblasts and fish keratocytes was measured as 100 nN and 20 nN, respectively.

1.6.2. Cellular adhesive forces

Adhesion of cells to any substrate is mediated by transmembrane glycoprotein, integrins etc[144]. Cell-substrate adhesion is an active process which involves a cascade of events leading to integrin mediated adhesion[145]. Various techniques such as AFM

(Fig 1.5b), high-speed centrifugation technique, manipulation force microscope and traction force microscope are used to measure cell adhesion forces[146]. Adhesion forces for human cervical carcinoma cells, epithelial cells, murine fibroblast cells and rat cardiac fibroblasts have been reported to be 190 – 210, 100, 300 – 400, and 10 nN, respectively[147]. These quantified adhesion forces represent the force required to overcome the attachment of cells and substrate[148]. Adhesion force quantification is sensitive to various parameters such as temperature and composition of protein substrate[144], [148]. Microfluidic platforms are currently being explored for cellular sorting based on their adhesion strength. In one such study, cells were exposed to constant fluid flow inside a microfluidic device, significant difference between adhesion strength of somatic, pluripotent and differentiated progenies was observed[149]. Use of this approach allowed authors to sort out pure populations of live cells with 95% purity.

1.6.3. Modulus of elasticity

There are different techniques summarized in this dissertation that are used to study the mechanical properties of the cells. In order to define cell mechanics in a technique or method or model independent manner, elastic modulus is widely used[150]. Elastic modulus help define the mechanics in terms of quantitative parameters assigned to a material in a technique-independent way[150]. To describe the elastic moduli of a cell, three primary moduli of elasticity can be used: shear, bulk and Young’s (tensile) moduli[151]. In the field of cell mechanics, a cell is assumed as a homogenous and isotropic material which allows cell mechanical properties to be defined by only two parameters - elastic modulus, and Poisson’s ratio[150]. In this dissertation, elastic modulus and Young’s modulus are interchangeably used. It is of concern in the cell mechanics field that definition of Young’s modulus needs redefining at the nanoscale. Poisson’s ratio of soft materials are unknown and possibly ranges between 0.3 – 0.5, which leads to a maximum of < 10% error in elastic modulus[152]. Therefore, studies quantifying mechanical properties of cells in term of modulus of elasticity characterize cell mechanics with just one parameter, the Young’s modulus.

Fig. 1.8. Single force spectroscopy: Schematic of working of AFM (a) and a typical force- distance curve obtained on a cell (b)[153].

1.6.4. Tether force

Intrinsic deformability of the plasma membrane influences the dynamics of cell shape change which is a hallmark of various cell phenomena[154]. Tether force which depends on the connection between cytoskeleton and plasma membrane is studied as a measure of plasma membrane tension (Fig 1.8b). Techniques such as AFM, optical tweezers, magnetic tweezers, and micropipette are typically utilized to measure tether forces[155], [156]. For example, micropipette is used to hold lipid vesicles or cells, and a second pipet is used to pull membrane tethers. Recently AFM based tether extraction was utilized to elucidate the role of ezrin in maintaining membrane tension in MDCK II cells.

It was noted that tether force of WT cells (~ 60 pN) decreased upon depletion of ezrin

(siRNA treated, ~ 48 pN)[157]. Using various techniques, tether forces of a variety of vesicles and cells have been reported in the range of 40 – 300 pN[107], [110], [156].

1.6.5. Cellular deformability

Deformation of cell can act an indicator of cellular processes such as growth, polarization, and differentiation. Various approaches have been developed and utilized to study the cellular deformability, such as microfluidics based real-time deformability cytometry (RT-DC)[158] and electrical stimulations[159]. RT-DC based mechanical characterization of cells relies on the deformation of cells by hydrodynamic shear and normal stresses observed using high-speed microscopy images (Fig 1.9). Using custom based algorithms images, cell deformation was quantified from projected area, in real-time.

Deformation in RT-DC based microfluidics can be calculated as:

2√π Area Deformation = 1 − Perimeter

Fig 1.9. Operating principle of microfluidics-based RT-DC device[158].

Recently, a new technique utilizing electrical stimulation to characterize cell deformation was used. Piezoelectric PbZrxTi1-xO3 nanoribbons were fabricated to measure deformation of neuronal cells under electrical excitations. On application of 120 mV stimulus, 1 nm of cell deformation was observed in agreement with a theoretical model of a depolarized cell membrane experiencing tension[159].

1.7. Scope of the dissertation

The overall goal of this project is to investigate and correlate the effects of external and internal cues responsible for phenotypic and genotypic changes during nervous system development. To elucidate the role of mechanobiology on the formation, maturation, differentiation, and migration of various cells in the CNS during development, this research focuses on three major areas: In Aim 1, I investigated the cytotoxic, biochemical, biophysical, and biomechanical characteristics of human fetal NPCs exposed to a variety of environmental toxicants. In Aim 2, I elucidated the critical role of mechanosensory complex (sub-cellular features of a cell) proteins, specifically Kindlin-3, in microglia on the cell membrane mechanics and physical characteristics. Finally, in Aim 3, I identified the molecular mechanisms by which human developmental stages-derived NPCs transduce mechanical input from external substrate into fate decisions such as differentiation and phenotype expression. Investigation of such biomechanical outcomes at the cellular and molecular levels will help better elucidation of the mechanisms underlying disease progression, brain plasticity, morphogenesis, wound healing, and cancer metastasis.

CHAPTER II

OPTIMIZATION OF AFM PROTOCOLS FOR CHARACTERIZING CELL,

TISSUE AND SOFT SUBSTRATES

In this chapter, I will first discuss the characteristics of an MFP-3D-Bio AFM and its utility in biomedical applications, specifically in mechanobiology. Then I will provide protocols, which I designed and optimized using this AFM, to characterize synthetic substrates and biological specimens. I utilized various cell types, soft tissues, and soft materials to optimize the AFM probes and working conditions before using them in our mechanobiology studies. Specifically, I will detail the results from characterization of live human aortic smooth muscle cells, live human glioblastoma cells, mouse fixed and fresh retinal tissues, fresh rat spinal cord tissues, polydimethylsiloxane (PDMS) films, and rat- tail derived type I collagen gels of various concentrations to optimize the operating conditions of the MFP-3D-Bio AFM. I compared the results I obtained to that reported in literature, where relevant and applicable.

2.1. Utility of MFP-3D-Bio AFM to study mechanobiology

Major global players in AFM market include Bruker Corporation, Hitachi High-

Technologies Corporation, JPK Instruments AG, Park Systems Corp., Oxford Instruments

(formerly Asylum Research Corp.), Nanonics Imaging Ltd., Nanosurf AG, and Anasys

Instruments Corp. Despite their relative advantages and limitations, not all the AFMs available currently in the market are designed for characterizing mechanobiology.

Specifically, most of the biomedical projects require high-resolution imaging (typically <

100 nm), usually at a nanometer scale, to characterize sub-cellular features, hierarchical protein structures, single polymer chains, nanomaterials, biosensors, and surface topography of single crystals. In addition, some of the projects demand low-noise, high- precision force spectrometry capabilities, to investigate cellular responses to mechanotransduction, mechanical properties of soft matter (hydrogels, colloids, cellular surfaces), dispersion forces due to protein adsorption, and surface charge density-mapping. For some specific projects, live imaging of stem cells under mechano-transduction and simultaneous measurements of intracellular proteins (usually via transfection) might offer key insights into the biological pathways involved in the observed phenomenon. To meet the broad array of aforementioned requirements, we utilized a high-performance MFP-3D-BIO atomic force microscope (Oxford Instruments; formerly Asylum Research, Santa Barbara,

CA, USA) integrated with a Nikon Eclipse Ti inverted epi-fluorescence microscope

(Melville, NY, USA), to provide the desired imaging and force measurement capabilities.

Compared to its predecessor model (MFP-1D), MFP-3D-BIO eliminates optical interferences and ensures straight and linear z-travel (> 15 microns) with its sensor-based feedback flexure stage design. Furthermore, compared to the MFP-3D Stand Alone (MFP-

3D-SA) AFM model, the MFP-3D-BIO allows combining high-resolution optical imaging and spectroscopy measurements simultaneously with the AFM scanning on the sample.

For integrating the instrumentation, the sample stage of the Nikon Eclipse Ti microscope

42 was replaced with a mechanical stage, allowing sub-micron positioning of the sample in two dimensions. The following are the specifications of this instrument:

• MFP-3D ARC2 Controller with following operating modes: contact, AC with Q-

control, dual AC, force spectroscopy, nanolithography and nano-manipulation

• A 90-micron closed-loop x-y stage for precise positioning of samples

• Microscope stage compatible with different sample supports, including glass slides,

coverslips, or petri dishes, from 35 mm to 85 mm in diameter

• A 3D cantilever holder compatible with both air and liquid

• Data acquisition and control software for analysis, image processing and rendering

software

• Low-coherence (860 nm) super-luminescent diode source for optical lever deflection

measurement, enabling flat-baseline, low-noise force and topography measurements.

A blocking filter is provided to eliminate infrared light from the diode.

• A stand-alone environment controller for use with heating and cooling stages to

manipulate the sample microenvironment. This controller can be fully operated through

the graphical interface of the AFM software.

• A vibration isolation system with 2'  2'  3'' granite top (Active 1.2 - 200 Hz, Passive

> 200 Hz).

• An acoustic noise enclose with window, to provide quiet environment for low-noise

AFM imaging.

• A temperature-controlled closed fluid cell (Bioheater®) for manipulating specimen

temperatures up to 80 C, with a precision of 0.1C and accuracy to 0.2 C. This fluid

cell requires environment controller listed above.

• Nikon Eclipse Ti inverted optical microscope with 0.55 H/Ph1/2/3/DIC LD condenser,

and three objectives (10, 40, and 60).

• A Himamatzu sCMOS Ocra FLASH 2.8 camera kit, for excellent high-resolution video

capture and display. This model comes with no fan or water-cooling system, which is

highly desired for low-noise and low-vibration measurements and imaging.

2.2. Beading of AFM cantilever tips

Tip-less AFM cantilevers (Arrow TL1, NanoWorld, Watsonville, CA, USA) were modified by attaching a bead using two-part epoxy (5-min setting time) as explained earlier[160], [161]. A small droplet of epoxy was placed on the one end of the glass slide using a p10 pipette tip and an extremely diluted suspension of 2, 4.5, 35 and 80 µm polystyrene beads (Polybead® Microspheres, Polysciences, Inc., Warrington, PA) was placed on the other end of the slide. Using a cantilever-moving technique[162], the cantilever mounted onto AFM head was first moved to pick up a tiny amount of glue and after 3-4 min, it was moved to pick the polystyrene bead. The beading process was observed using a 40 objective on an inverted optical microscope (Fig. 2.1).

A B C

Fig. 2.1. Representative beading of AFM probes with micron-sized polystyrene spheres for non-destructive characterization of gels, cells and soft substrates. AFM tip glued with a microsphere (4.5 µm) and cured for one hour.

2.3. Spring constant calibration

The spring constant of beaded/regular cantilevers was calculated using thermal

calibration method built into the Asylum Research software as explained earlier[160]. First,

the deflection sensitivity was measured using a glass slide (stiffer material than cantilever).

A force-deflection curve with flat baseline and positive slope showing cantilever deflection

by the glass surface was obtained using a trigger point of 1 V. The slope of this line was

calculated using inbuilt option which yielded the deflection sensitivity. Using thermal tune

menu in software, thermal fluctuations of the cantilever were calculated against time. A

Lorentzian curve was fitted and integrated area under the curve was calculated which

yielded the cantilever spring constant[163]. The spring constant for the Arrow TL1

(NanoWorld, Watsonville, CA, USA; used for mechanical characterization of collagen

45 gels, fixed and fresh tissues and live-cells) was found to be 0.029 – 0.037 N/m which is in the range (0.004 - 0.54 N/m; nominal spring constant ~0.03) provided by the manufacturer.

The spring constant for the AC 160 TS-r3 (NanoWorld, Watsonville, CA, USA; used for mechanical characterization of PDMS films) was found to be 24.8 – 27.2 N/m which is in the range (24 - 28 N/m; nominal spring constant ~ 26 N/m) provided by the manufacturer.

2.4. Characterization of polydimethylsiloxane (PDMS) films

2.4.1. Experimental Methods

PDMS films of different crosslinking ratios (1:10, 1:20, 1:30; elastomer curing agent and elastomer base, respectively) were prepared using Sylgard 184 Silicone

Elastomer Base and Sylgard 184 Silicone Elastomer Curing Agent. For instance, 4 g of

Sylgard 184 Silicone Elastomer Curing Agent and 40 g of Sylgard 184 Silicone Elastomer

Base were weighed and mixed to achieve 1:10 ratio. Using a glass stirring rod, both components were mixed together for a minimum of 1 minute. The mix was placed in a desiccator and ran under vacuum for 20 min to remove any bubbles. The mix was placed in a drying oven for 2 h at 75 °C in a circular mold. PDMS films were removed from the mold and were placed in boiling water for 45 min to remove any un-crosslinked polymer.

PDMS films were attached to 50 mm petri-dishes (Falcon™ Tight Fit Lid Dishes, catalog number 08-757-105) containing PBS using a double-sided tape.

AFM cantilever, AC 160 TS-r3, capable of characterizing stiff materials (100 kPa to 3 MPa), was used for this indentation study. Force-displacement curves were obtained at a rate of 5 µm/s and force distance of 5 µm. For each crosslinking, 2-3 force maps (1024

46 indentations per map) from 3 independent PDMS films were obtained in force-volume mode and resulting force vs indentation curves were analyzed by fitting Hertz model for pyramidal tip using proprietary software (Igor Pro 6.37).

푡푎푛훼 퐸 퐹 = δ2 √2 (1 − 휇2) where F is indentation force, E is Young’s modulus, α is face angle 36°, µ is Poisson’s ratio

(~0.3)[164], [165], and δ is indentation depth.

2.4.2. Results

A B B

C D Fig. 2.2. Representative force map (A) of PDMS substrate crosslinked at 1:20 ratio. The Young's modulus (B) for different cross-linking PDMS films were calculated from force- indentation curves ranged between 0.1 – 2.3 MPa, * denotes p<0.05.

We noted the Young's moduli of PDMS films to decrease monotonously with change in crosslinking. The average Young's moduli were 2.3 ± 0.25, 1.44 ± 0.10 and 0.19

± .01 MPa, for 1:10, 1:20 and 1:30 cross-linking ratios, respectively, with significant differences between each group (p < 0.05; ANOVA followed by Tuckey’s post-hoc) (Fig.

2.2). 1:10 ratio is one of the most characterized PDMS films due to its application in

47 microfluidics platforms. Several studies report the Young’s modulus of 1:10 ratio PDMS film in the range of 1.7 - 2.6 MPa at similar curing temperature[166],[167],[168],[169].

Elastic moduli of 1:20 and 1:30 ratio was also found to be similar to reported values in literature[165],[170],[169],[168]. I was able to establish the utility of MFP-3D-Bio AFM for mechanical characterization of PDMS films typically used in mechanobiology studies for cell cultures and optimized the protocol for mechanical characterization of substrates in hydrated conditions.

2.5. Characterization of rat-tail derived type I collagen gels

2.5.1. Experimental Methods

Tip-less AFM cantilevers (model Arrow TL 1) modified by attaching a 35 μm polystyrene bead using epoxy (details in 2.2.) was used for indentation of collagen gels.

The size of the bead was carefully selected based on the past studies characterizing collagen or similar type of hydrogels (such as Matrigel)[84],[171]. Besides, 35 μm polystyrene bead allows more indentation area as compared to 2 or 4.5 μm polystyrene bead. The spring constant for each probe was precisely calibrated using thermal calibration method before use (explained in 2.3.). Collagen samples for AFM tests were cold-pipetted into custom- made wells (n=3 wells/concentration), allowed to polymerize under humidified conditions for one hour, and maintained in PBS at 37 °C during testing. Force-displacement curves were obtained at a rate of 5 µm/s and a trigger force of 4 nN was applied. For each concentration, 18-21 force curves were obtained at random locations on each gel (n = 3

gels per concentration) and resulting force vs indentation curves were analyzed by fitting

Hertz model using proprietary software (Igor Pro 6.37).

4 퐸 퐹 = √푅훿3 3 (1 − 휇2)

where F is indentation force, E is Young’s modulus, µ is Poisson’s ratio (~0.33)[172], R is

radius of spherical bead, and δ is indentation depth.

2.5.2. Results

A B 3.5 3.5 Power-law fit (p = 0.0027) 3.0 E ∝ c1.22 1500 3.0 2.5

2.0 1000

2.5 1.5 Force, nN Force, 1.0 500

2.0 0.5 0 1 2 3 4 5 100 200 300 400 500 600 700 Young's Modulus, Pa 1.5 Indentation, nm Collagen (mg/ml)

Force, nN Force, 1.0 Fig. 2.3. Representative force-indentation curves from AFM analysis were shown for collagen gels prepared over a wide range of concentrations. A systematic increase in slope 0.5of the curves is evident with increasing concentration. The Young's modulus calculated from these curves ranged between 50 – 1,400 Pa for the gel concentrations tested here. The data were fit to a power-law fit (p = 0.0027) which showed that E ∝ c1.22. 100 200 300 400 500 600 700 Indentation, nm The results in this section have been accepted for publication[173]. We varied the

concentration of collagen to examine Young’s modulus of these gels of various

concentrations (Fig. 2.3). The average Young's moduli were 113 ± 24.7 Pa, 547.1 ± 79.1

Pa, 732.4 ± 50.6 Pa, 922.4 ± 84.4 Pa, and 1395 ± 172.1 Pa, for 1, 2, 3, 4, and 5 mg/mL

gels, respectively, with significant differences between each group (p < 0.05; using

ANOVA followed by Tuckey’s post-hoc). Our results for 2 mg/mL gels are similar to those reported earlier (~ 511 ± 142 Pa)[84]. We noted the Young's moduli of collagen gels to increase monotonously with concentration (1 to 5 mg/mL), with a power-law dependence

(E ∝ c1.22; p = 0.002), as evident from AFM indentation tests. Prior studies suggest that the storage moduli of collagen gels increased with concentration (G′∼ c2.1, at 37 °C)[174]; the shear moduli (G′) of 1, 2, and 3 mg/ml collagen gels at low strain (0–0.1) were reported as

3, 44.5, and 97 Pa, respectively, while their respective tensile modulus (E) at 0.1 mm/min strain rate were reported in 10–17 KPa range[175],[176]. I was able to establish the utility of MFP-3D-Bio for mechanical characterization of collagen gels, which is one of the most widely used hydrogel for various 3D encapsulation platforms. These studies helped me streamline the protocol for mechanical characterization of substrates mimicking biological tissues which are less stiff than PDMS films and validate the results I obtained with those reported in literature by others.

2.6. Characterization of fixed and fresh mouse retinal tissues

2.6.1. Methods

For characterization of fixed and fresh mouse retinal tissues, a 4.5 μm or 35 μm bead (refer to Fig. 2.1), respectively, was glued to an ArrowTM TL1 tip-less cantilever

(details in 2.2.). A smaller bead (diameter ~ 4.5 μm) was selected for fixed tissue, as the size of the indenter is small enough to probe the specific layer of retina (OPL is typically

~ 5 μm). The tissues were isolated from mouse models and kindly provided by Dr. Tatiana

Byzova’s group at the Lerner Research Institute [Dudiki et al. Micrglia control vascular

50 architecture by a novel mechanosensory mechanism, Nature Communications, In Revision,

(2019)]. For characterizing fresh retinal tissues, a 35 μm bead was selected based on the heterogeneity of the tissue sample[177]. The probe was submerged in PBS prior to each experiment to stabilize, minimize drift, and achieve thermal/mechanical equilibrium.

Retinal sections (40 μm thick) were prepared from the molds with a cryotome. Force- indentation curves (n ≥ 15 in each region for 3 fixed retinal tissues, and n ≥ 55 from 4 fresh retinal tissues) were obtained with a trigger force of 4 nN at an approach speed of 5 μm/s.

Step sizes of 20 μm for retinal sections immersed in PBS, and 50 μm for live tissues immersed in media were used. The force-indentation curves were fitted to a Hertz model

(explained in 2.5.1) to calculate the elastic modulus using Igor Pro 6.37 software.

2.6.2. Results

DAPI Isolectin CX3CR1-GFP

(INL) INL ~ 2 kPa ~ 0.7 kPa (OPL)

ONL ~ 6 kPa (ONL)

30μm 0 6 (kPa)

Fig. 2.4. 3D reconstitution of whole-mount P16 mouse retina with CX3CR1-GFP- expressing microglia (green) stained with isolectin to detect vasculature (red) and DAPI for nuclei (blue). The Young’s modulus of the inner nuclear layer (INL), the outer

51 plexiform layer (OPL) and the outer nuclear layer (ONL) were measured by AFM and their respective average Young’s modulus are shown as the Young’s modulus map of retinal layers (n = 3 retinas). [Immunofluorescence image provided by Dr. Dudiki from CCF]

The retina has a highly conserved layered structure with three vascular layers that are positioned exactly between the respective nuclear layers. As shown in Fig. 2.4, while the deep vasculature is positioned on top of the outer nuclear layer (ONL), the intermediate vascular network grows between the outer and inner nuclear layers (INL), suggesting the presence of mechanical barriers between these vascular layers. A Young’s modulus map

(Fig. 2.4) of the retina generated with AFM revealed that vascular networks are separated by the stiffer INL and ONL, characterized by elastic modulus of 2 ± 0.2 kPa and 6 ± 0.8 kPa, respectively. At the same time, the Young’s modulus of the vascularized outer plexiform layer (OPL) was 0.7 ± 0.03 kPa. We observed a similar pattern of mechanics for different layers of retinal tissue with different cell body density as reported earlier[177],[178]. This study helped establish the utility of this AFM for characterizing the mechanical properties of fixed tissue samples.

1500 p < 0.001

1000

500

0 Young'sModulus, Pa Control Treated

Fig. 2.5. CS treatment of fresh P16 mouse retinas ex vivo for 6 h led to significantly decreased retinal modulus as measured by AFM (Mann-Whitney test; P < 0.001, 55 n 59 measurements on four mice per group). All measurements were taken within 700-1200 μm from the optic nerve.

To establish the characterization capabilities of MFP-3D-Bio AFM on fresh tissues

(not fixed with paraformaldehyde), we adopted a new technique developed for the CNS tissues that allows softening live tissues without damaging or disrupting the matrix and cells[179]. Results suggest that treatment of fresh retinas with chondroitin sulfate (CS) for six hours decreased the average retinal Young’s modulus from ~900 Pa to ~400 Pa (Fig.

2.5). Fresh-retina modulus was found to be in the range and follow trends reported by others, after chondroitin sulfate treatment[178],[180],[87]. This study helped optimize the protocol for mechanical characterization of live tissue samples, which are less stiff than the fixed tissues, without damaging the tissue.

2.7. Characterization of fresh rat spinal cord tissues

2.7.1. Methods

In collaboration with Dr. Nic Leipzig’s group at the University of Akron, we characterized the mechanical properties of adult rat spinal cord tissues. Spinal cords were collected at 28 weeks after perfusion with cold artificial cerebral spinal fluid (aCSF: 119 mM NaCl, 26.2 mM NaHCO3, 2.5 mM KCl, 1.3 MgCl2, 10 mM glucose bubbled 15 min with 5% CO2/95% O2 then sterile-filtered) to ensure preservation of tissue integrity. Spinal cords were immediately stored in cold aCSF before embedding in 4% low-melting point agarose, then sectioned in cold aCSF to 500 μm sections using a vibratome (Leica).

Sections were mounted onto a petri dish with agarose that hardened around and secured

53 the sample to the plate. To secure the tissue further, a plastic Thermanox 13-mm round coverslip with a 412 mm biopsy punched window was placed over the sample with superglue. Samples were then submersed in aCSF and stored on ice until AFM measurements, usually < 8 h post animal sacrifice.

Tip-less AFM cantilevers (model Arrow TL 1) were modified by gluing an 80-μm polystyrene bead using epoxy (detailed in 2.2). For mechanical characterization of fresh spinal cord tissues (n = 3), an 80 μm bead was selected based on past studies characterizing

CNS tissues[64] and considering the heterogeneity of the tissue sample[30]. Before each experiment, the cantilever spring constant was determined using thermal noise method in a clean culture dish containing PBS (explained previously in 2.3). Prior to each experiment, probes were submerged in PBS maintained at 37 C to achieve thermal equilibrium and to stabilize and minimize drift. Using an optical microscope (AmScope SM-1TSZZ-144SS-

10M Digital Professional Stereo Zoom Microscope), each tissue was marked for reference and force-indentation curves were obtained at an approach velocity of 5 µm/s and a maximum force of 20 nN, with a 100 µm step size on a motorized stage. The analysis of force-indentation curves was automated using proprietary software (Igor Pro 6.37). Using

Hertz’s contact model (explained in 2.5.1.), Young’s modulus was determined from these force curves, using Poisson’s ratio of 0.5[181], radius of the tip as 40 µm, and indentation depth of 10 µm. An indentation depth of 10 µm and radius of 40 µm provides an effective contact radius (√푅훿) of ~ 20 µm.

2.7.2. Results

AFM micro-indentation was used to measure the EY of naïve white and grey matter at 28 weeks. Spatial distribution of mechanical properties and the interphase between white and grey matter were captured with EY mapping (Fig. 2.6A). Significant differences in EY was evident at random locations tested in naïve white and grey matter, and grey matter (~

420 Pa) had an EY nearly double that of white matter (~ 177 Pa) (Fig. 2.6B-C). Reported values of Young’s modulus of naïve white (150 – 200 Pa) and grey tissues (350 – 430 Pa) of rat spinal cords were in similar range to our observations[84],[64]. I established the utility of MFP-3D-Bio AFM for mechanical characterization of live tissue samples which are larger than the previously optimized gels and retinal tissues. These studies helped me in optimizing protocols for minimally-invasive characterization of live tissues without damaging the tissue, which could pave way for live cell characterization without damaging/rupturing them. Another advantage of such characterization with AFM is that the same samples could be further processed for immunofluorescence imaging and/ or matrix quantification as they were gently handled.

B • Longitudinal segment of spinal cord White A 800 Grey matter • 18-weeks timepoint matter 600

11 400

200 1 2 3 4 5 6 7 9

0 10 Young'sModulus, Pa 1 2 3 4 5 6 7 8 9 10 11 Location C 1 mm 800 p < 0.001 Pa 600 600 White 400 matter 200 300 Grey 0 matter Young's Modulus, Pa Grey matter White matter 0

Fig. 2.6. AFM mapping reveals spatial mechanical properties of the naïve spinal cord tissue. Representative naïve spinal cord tissue, and the distinct differences in elastic moduli (EY) in white and grey matter (separated by dotted black line) as evident from the heat map (A). Young’s moduli of tissue at random regions (marked in B) was characterized in white and grey matter (C). Data represented as average  SE. n = 3 independent tissues; and statistically compared via t-test.

2.8. Characterization of healthy and diseased human smooth muscle cells

2.8.1. Methods

Healthy adult abdominal aortic Hu-SMCs were obtained from Life Technologies

Corp. (Carlsbad, CA, USA) and passaged using Medium 231 (ThermoFisher Scientific,

USA). AAA-SMCs (kindly provided by Dr. Pinet’s group from Inserm-France) were cultured on 2D laminin-coated substrates. Cells were cultured for 21 days in SMC media, containing 100 nM GSNO (Sigma, Saint-Louis, MO, USA), a NO donor. Control cultures received no GSNO (termed 0 nM). The culture media contained recombinant human insulin-like growth factor-I (2 μg/mL) and glucose (D-Glucose; levels unknown).

Single cell indentation measurements were obtained to quantify the elastic modulus

(E) of live healthy and abdominal aortic aneurysm (AAA) tissues derived human smooth muscle cells (SMCs) under different culture conditions. Healthy- and AAA- SMCs were cultured on laminin-coated petri dishes (50 mm diameter) and maintained at 37 °C throughout live-cell indentation assay, over a one-hour experiment each time. Tip-less

AFM cantilevers (model Arrow TL 1) were modified by attaching a 4.5-μm polystyrene bead using epoxy (as shown in Fig. 2.1). The bead size was selected considering the size of the cell and scope of the study, to obtain heat maps of the whole cell. The spherical bead indenter reduces the possibility of destructive deformation of cell due to smaller local

56 strains applied, compensates for the x-y drift (usually < 1 μm), and yields higher spatially averaged measurement with minimal variability, relative to conventional sharp-tip probes.

The actual spring constant was determined from a force-distance curve and using thermal calibration method in a clean culture dish containing DMEM (detailed in 2.3.). Drifting minimization and thermal/ mechanical equilibrium of AFM was achieved by submerging probe in cell medium for stabilization.

For each experiment, at least 30 cells were randomly selected, indented between nuclei and cell edges, and 150 - 200 force curves were obtained at random locations on each cell in force-volume mode, at a rate of 0.25 Hz with approach/ retraction velocity of

2.5 µm/s. Using Hertz’s contact model (explained in 2.5.1.), Young’s modulus was determined from these force curves, using Poisson’s ratio (~ 0.5 for cells[116], [182],

[183]), radius of spherical bead (2.25 µm) and indentation depth (300–400 nm). The indentation data was analyzed using Igor Pro 6.37 software assuming cells behave as linearly elastic at small strain regimes. The mean elastic modulus values were calculated for each cell from the elastic modulus heat maps generated over randomly selected 90  90

µm2 area, and then averaged across multiple cells for each culture condition group. A two- tailed Student’s t-test was used for statistical analysis and p < 0.05 were considered significant.

2.8.2. Results

The results in this section have been accepted for publication[28]. Representative heat maps of elastic moduli (Fig. 2.7A) and force-indentation curves (Fig. 2.7B) were shown for healthy and AAA SMCs. The average Young’s modulus for AAA-SMCs (17.41

± 0.34 kPa) was significantly higher (1.6-fold; p < 0.001) than for healthy SMCs (10.42 ±

0.29 kPa), indicating an increase of stiffness in AAA-SMCs (Fig. 2.7C). Upon exposure to GSNO, the average modulus of elasticity for healthy SMCs decreased by 1.3-fold

(healthy SMCs: 8.67 ± 0.25 kPa, p < 0.05; AAA-SMCs: 13.03 ± 0.35 kPa, p < 0.05). In addition to alterations in composition and functionality of vascular endothelium and ECM, changes in cell stiffness might also contribute to vascular tissue stiffness in AAAs, atherosclerosis and hypertensive conditions.

The Young’s modulus (~30 kPa) of hypertensive aorta-derived SMCs, as measured by AFM nano-indentation, was reported to be two-fold higher than that of normal aorta

(~14 kPa), and such stiffness was mediated by cytoskeletal proteins[184]. Vascular SMCs from old male monkeys exhibited much higher modulus than their younger counterparts

(42 kPa vs. 13 kPa)[185]. The elastic modulus of primary adult male rat arteriole-derived

SMCs (~13.5 kPa) decreased by 38% upon exposure to NO donor PANOate, most likely due to remodeling of actin cytoskeleton network[186]. Here, we reported for the first time on the Young’s modulus of human aortic SMCs from healthy and AAA tissues, using AFM nano-indentation of live cells.

A B Hu-SMCs AAA-SMCs

C * Hu-SMCs + GSNO AAA-SMCs + GSNO * * 40 * 30

0 Young'sModulus, kPa - + - + GSNO Hu-SMCs AAA-SMCs

Fig. 2.7. (A) Representative force maps of healthy Hu-SMCs (left panels) and AAA-SMCs (right panels) treated with 100 nM GSNO (lower panels). Background from Petri dish was subtracted in these images. Representative force-indentation curves (B) and average Young's Modulus (C) of these cells cultured on 2D substrates, in the presence or absence of 100 nM GSNO. The black dotted lines in (C) represent the arithmetic averages in respective cases, i.e., 10.33 ± 0.2624 KPa for untreated Hu-SMCs (n=605), 8.674 ± 0.2491 KPa for Hu-SMCs receiving GSNO (n=550), 17.41 ± 0.3405 KPa for untreated AAA- SMCs (n=524), an 13.03 ± 0.3498 KPa for AAA-SMCs receiving GSNO (n=469). Data were analyzed using one-way ANOVA followed by Tukey's HSD post-hoc testing, assuming unequal variance and differences deemed significant for p < 0.05.

Our results show that (i) elastic modulus of healthy Hu-SMCs is lower than that in rats and monkeys; (ii) human AAA-SMCs have higher elastic modulus compared to healthy Hu-SMCs; and (iii) human AAA-SMCs had significantly lower elastic modulus compared to hypertensive rat SMCs or aged monkey SMCs. This study helped streamline the protocol for mechanical characterization of live cells without damaging/ rupturing the cells in force-volume mode, to obtain force maps of the whole cell and quantify changes in cell mechanics in the presence of physiological compounds.

2.9. Characterization of human pediatric glioblastoma-derived cells

2.9.1. Methods

Pediatric glioblastoma multiforme cells (GBM; SJ-GBM-2) were obtained from

Xenograft Repository at Texas Tech University Health Science Center School of Medicine and Children’s Oncology Group. SJ-GBM-2 were obtained from a five-year-old female patient. SJ-GBM-2 were cultured in 50-mm uncoated petri dishes and maintained at 37 °C throughout indentation assay. Tip-less AFM cantilevers (model Arrow TL 1) were modified by attaching a 2 μm polystyrene bead using epoxy (detailed in 2.2.). These beads were selected considering the epithelial morphology of the SJ-GBM-2 cells, to selectively indent the peri-nuclear area. Force-displacement curves (n = 39) were obtained using 2 nN force at 5 µm/s approach/retraction velocity. Using Hertz’s contact model (explained in

2.5.1.), Young’s modulus was determined from these force-indentation curves, using

Poisson’s ratio (~ 0.5 for cells), radius of spherical bead (1 μm) attached to cantilever and indentation depth (400 – 500 nm).

2.9.2. Results

Using single cell nanoindentation assay in the peri-nuclear region of the live-cell, we are first to characterize the elastic modulus of SJ-GBM-2 cells. Average Young’s modulus of SJ-GBM-2 cells was computed as 1.4 ± 0.75 kPa (Fig. 2.8). The Young’s modulus of SJ-GBM-2 was found to be in range of 1.1 – 2.0 kPa as reported for various types of glioblastomas cells such as LN-229 cell line (right frontal parieto-occipital glioblastoma), U87MG (hypodiploid human cell line), A172 (Glioblastoma cell line),

BT145 (Primary GBM CICs), obtained using similar AFM setup[65],[187]. These outcomes helped streamline the protocol for mechanical characterization of live cells to

60 obtain force-indentation curves on the specific location of the cells (peri-nuclear region), which will be used in mechanobiology studies of this dissertation work.

2.0

1.5

1.0

0.5

0.0 Young'sModulus, kPa SJ-GBM-2

Fig. 2.8. Average Young’s modulus 1.4 ± 0.75 kPa (n=39) for SJ-GBM-2 cell line was calculated from force-indentation curves by fitting Hertz model.

2.10. Summary

In summary, using various synthetic and biological substrates, biological tissues, and mammalian cells, I was able to establish and refine the protocols for implementing the

MFP-3D-Bio AFM to obtain force maps, and biomechanical characteristics under non- destructive indentation of cells, tissues, and soft substrates. I was able to optimize various operating parameters of the AFM, as well as sample preparation techniques unique for each sample. These experiences and outcomes helped in the mechanobiology studies of brain- derived cells from developmental stages, detailed in subsequent chapters.

CHAPTER III

BIOPHYSICAL AND BIOMECHANICAL PROPERTIES OF NEURAL

PROGENITOR CELLS AS INDICATORS OF DEVELOPMENTAL

NEUROTOXICITY

3.1. Introduction

During CNS development, the large reservoirs of NPCs in the ventricular and subventricular zones decline with maturation and aging. For instance, the large NPC population in neural crest and neural spinal tube dramatically decline to ≤ 1% postnatal in a rat embryogenesis model, highlighting the role of NPCs and their microhabitat in normal embryogenesis[67],[60]. NPCs are self-renewing, multi-potent cells which can differentiate into neuronal and glial lineages. CNS development is a complex process comprising several highly coordinated events such as migration, differentiation, proliferation, cell death, myelination, and synthesis of neurotransmitters[69], which occur in a well-defined time-frame, making each event differentially vulnerable on exposure to harmful compounds. Any perturbation during CNS development could lead to permanent damage with little chance to repair. Neurotoxicity could be defined as any adverse effects

62 on the structure and/or function of the developing or mature nervous system.

Developmental neurotoxicity (DNT) is the least studied effect of chemicals; only 150 compounds have been screened for their potential DNT, and only a handful of them (e.g. lead, arsenic) have been classified as developmental neurotoxicants[188],[189].

Epidemiological studies have detailed the role of environmental chemicals and clinical drugs on neurological disorders such as learning disabilities, lowered IQ, cognitive dysfunction, impaired verbal skills, and poor perceptual and motor skills[190],[191].

During fetal development, transporters such as influx transporters for glutamate, copper, and zinc are more active, and mechanisms such as specific transport of plasma protein takes place, that makes the blood-CSF barrier immature and more permeable to toxins in the environment[78],[192]. There is strong evidence that the developing nervous system is more susceptible to damage than adult nervous system[66], most likely due to the sensitivity of developmentally-relevant NSCs and the immature/ permeable blood- brain barrier. The contrast between the small library of known neurotoxicants and vulnerability of developing nervous system suggests an urgent need for obtaining DNT data[191].

Neurological insults during the first trimester could manifest during later stages of adolescence as developmental disorders, which currently affects one in every six children[193],[188]. This has prompted the need to methodically screen a range of widely- known environmental and pharmaceutical compounds for their potential neurodevelopmental toxicity. Animal models have been widely used to flag compounds for toxicological identification. In the field of reproductive and developmental toxicology, higher number of experimental animals are typically used, mostly for statistical

63 significance purposes[194]. Since animal models are xenogeneic, expensive, poorly predictive of human outcomes, and ethically and morally contentious, demand for alternative in vitro test methods has been growing. In the absence of developmentally- relevant fresh primary brain cells, immortalized cell lines such as embryonic stem cells

(ESCs), induced pluripotent stem cells (iPSCs), primary trophoblast cells, NPCs, and primary neurons are being explored to elucidate neurotoxicity of various classes of compounds[195],[196]. However, current in vitro tests focus mostly on biochemical assays to assess the toxicity, while important changes in biophysical and biomechanical characteristics of these progenitor cells were rarely studied[20],[133].

Various studies have reported the use of NPCs to study the mechanisms of neurotoxicity with special emphasis on proliferation, cell death, and differentiation. Cell death can be broadly categorized as apoptosis or necrosis (Fig. 3.1); apoptosis being programmed cell death[197],[198]. Molecular mechanisms driving apoptosis include membrane preservation, lack of cytokine/ chemokine production, nuclear fragmentation, reduction in cellular volume, to name a few. On the other hand, necrosis is typically characterized by loss of membrane integrity, increase in production of cytokines and chemokines that cause inflammation, and increase in cellular volume[197], [199]. These characteristics identify the cause and physical features associated with the type of cell death. In addition, there are caspase-dependent and independent pathways, and genetic/ biochemical markers both in the field of disease models and toxicology which shed further light on the cause of neural death[199].

Currently, there are many proteins and enzymes under investigation for their role in cell death. Caspase is one of the most widely studied enzyme for its role in apoptosis.

Caspase-3 is one of the fourteen known caspases; caspase-3 dependent apoptosis is characterized either by compromise of surface receptors or internally by mitochondrial damage[198]. To investigate the specific vulnerability of developing nervous system, it is the important to elucidate the mechanisms underlying apoptosis under those toxicant- aberrant conditions. Studying apoptosis helps predict the effects of toxicants with comparable underlying mechanisms of toxicity[188],[197].

Apoptosis and/or Necrosis Passive Drug Diffusion Influx Transporter Genetic Unmodified Drug Caspase/Cathepsis changes Activation Phase I DNA Metabolism (CYP) Damage Mitochondrial damage Reactive Metabolite ROS/RNS Generation

Phase II H2O2 OH Metabolism (GST) Lipid Fluctuations of Peroxidation ATP levels Conjugated Metabolite Cl- Protein Membrane Compromise Efflux Adduction Ion Transporter channel (MDR) blocking Extracellular Excretion ions Ca++ Na+ Exposure to a toxic compound

Fig. 3.1. Mechanisms of NSC toxicity ultimately leading to either apoptosis or necrosis. Red outline indicates drug metabolism effects[200].

In the presence of toxicants, the normal metabolic process produces reactive oxygen species (ROS) at an upregulated rate. These ROS are toxic to any cell type, especially for brain tissues, as the brain cells are highly susceptible to ROS-mediated damage. ROS causes inability for self-repair and leads to high metabolic and oxygen

65 consumption rate. The presence of free radicals within NSCs could cause alteration of

DNA, proteins and lipids which eventually leads to apoptosis[201],[198]. Lower levels of glutathione (GSH) signifies the compromise in antioxidant defense functions. This is associated with increased concentrations of glutathione disulfide (GSSG), which leads to a decrease in GSH/GSSG ratio, which is a critical indicator of cellular health[188].

Studies have shown the effect of ROS on NSCs under toxicant-aberrant conditions.

NSC death caused by exposure to 100 μM ketamine was characterized by an increase in

ROS production and mitochondrial fission. Upon exposure to manganese (Mn), NSCs experienced ROS-mediated compromise in viability. Mn accumulation in mitochondria disrupt the oxidative phosphorylation which leads to an increase in ROS production. It is evident that many toxicants cause damage to NSCs by oxidative stress, but the underlying process needs further evaluation[66],[189].

DNA and RNA denaturation has been linked to apoptosis and oxidative stress- induced cell damage. DNA damage can be observed by the fragmentation of nuclei. One of the most common and widely used markers for characterization of DNA denaturation is

Hoechst. In one such study, NSCs were exposed to trimethyltin (organotin compound) and the damage to DNA was observed to be the most sensitive mechanism of toxicity[197]. In another study, exposure of MeHg led to 15-20% nuclear fragmentation, which induced adult NSC death. Exposure of 10-fold lower concentration of MeHg showed same amount of nuclear fragmentation in developing NSCs, demonstrating the nuclear fragmentation- based susceptibility of developing CNS. This highlights the need to study the pathways involved in NSC survival upon toxicant insult[189].

The signature move of both necrosis and apoptosis is the cell membrane compromise, either by swelling or degradation. The use of cell impermeable stains such as propidium iodide (PI) shows the dead cells or the effect of necrosis. To determine the compromise in membrane integrity, stains such as Sytox® Green have been used by researchers. In one such study which evaluated the effects of heavy metals on NSC membrane integrity, it was noted that with increasing concentration of heavy metals, there was an increase in the percentage of cells with a compromised cell membrane. This membrane compromise showed a strong correlation with the percentage of apoptotic cells using YO-PRO-1 assay[66],[199]. Adult NSCs exposed to MeHg (0.5 to 2 μM) resulted in

45-50% compromise in membrane integrity, whereas exposure of ten-fold lower concentration to embryonic NSCs showed comparable damage to membrane. Various studies summarizing the effect of toxicants on membrane compromise followed similar trends as apoptosis and nuclear fragmentation discussed above[189].

Mitochondria is one of the key organelles best known for ATP generation and its role in apoptosis. It is also involved in essential processes such as calcium, iron, copper homeostasis, synthesis of steroids, and fever response. Mitochondrial morphology ranges from highly fragmented to highly networked, depending on factors such as cell type, environment, and developmental stage[201],[197]. Recognition of mitochondrial damage due to drugs and environmental toxicants is a relatively recent development. A recent study elucidating the effects of heavy metal ions on NSCs used JC-1 membrane potential stain to access mitochondrial function. The study showed the progressive transformation from healthy to disrupted mitochondria with increasing concentration of each metal ion[66]. In a recent study, Cisplatin (CDDP) - one of the most widely used cancer drugs with a known

67 blood-barrier crossing capability - was evaluated for its neurotoxic potential. It was reported that CDDP induced mitochondrial damage, increased oxidative stress in cultured hippocampal neurons and NSCs[198]. These studies show the importance of studying mitochondrial dysfunction-based assays for proper and efficient screening of compounds with neurotoxic potential.

Majority of compounds which show adverse effects on developing brain even at minute dosages show no such effect in adults. It is important to develop in vitro methodologies for assessing the damage caused even at extremely lower dosages of toxicants[196],[189],[202]. Exposure to 2.5 – 5 nM concentrations of MeHg caused impairment in neuronal formation from differentiating NSCs. The concentration range tested in this study is well below the amount of MeHg found in the umbilical cord of pregnant women in Sweden[189]. In contrast to MeHg, non-cytotoxic nanomolar concentrations of polychlorinated biphenyls (PCBs) promoted neuronal differentiation in differentiating NSC cultures[190]. In a recent study, nanomolar concentrations of heavy metals (Cd, Hg, Pb) were investigated for their effects on differentiation of embryonic murine stem cells. Within 7 days of cultures, significant decrease in neural differentiation was observed at as low as 10 pM concentration of Hg and Pb. By day 14, all three heavy metals showed inhibition in neural and glial differentiation[66]. However, the effects of toxicants on cellular differentiation was not the focus of the study.

CNS development is a tightly-regulated process, from maturation of neurons to folding of the brain, and relies heavily on mechanical forces and biochemical cues[93]. For instance, radial glial cells clonally linked to NPCs act as a mechanical scaffold for cell migration during brain formation[30], highlighting the importance of intrinsic cellular

68 mechanical characteristics such as membrane tension in organizing motility, cell shape, and mechanotransduction[203]. Perturbations to cellular biophysical aspects could change the coupling between cellular intrinsic forces and matrix mechanical properties, causing abnormal mechanotransduction[3]. Cell mechanics is gaining traction as an important biomarker of cell differentiation, pathophysiology, and cancer progression[26],[116],[20],[185]. The biomechanics of various cell types has been explored using optical tweezers, micropipette aspiration, magnetic twisting cytometry, and AFM, among others[204],[205],[206],[207]. The utility of AFM to study the mechanical properties of individual cells under pathological and toxicant-aberrant conditions is gaining attention[138],[104],[113],[137]. However, characterization of the changes in biophysical and biomechanical properties and correlation of the biomechanical and biochemical outcomes after toxicants exposure remain unexplored.

Since biochemical and biomechanical cues play an integral role in regulating fetal development[5], in this work, we used human fetal NPCs to evaluate the cytotoxic potential of various classes of compounds on developmental neurotoxicity. We evaluated the subcellular mechanisms of action of rotenone, digoxin, chlorpyrifos, and arachidonoyl- ethanolamide (AEA) over a wide range of concentrations. These four compounds have been selected for their toxic potential in various in vitro and in vivo conditions, although the extent of their prior testing was confined to quantifying IC50 or LD50 levels[194],[208],[18],[202],[188]. Using AFM, we not only quantified the changes in physical and mechanical properties of NPCs treated with these compounds, but also correlated the biochemical outcomes with biomechanical characteristics of cells.

3.2. Materials and methods

3.2.1. Cell culture and compound exposure

Human NPCs (ReNcell VM; SCC008; EMD Millipore, Burlington, MA, USA) were maintained in an undifferentiated state by culturing in a complete medium (ReNcell maintenance medium; SCM005; EMD Millipore) supplemented with 20 ng/mL each of bFGF and EGF, and 1% penicillin/streptomycin, on laminin-coated flasks. Stock solutions of rotenone, digoxin, and chlorpyrifos were prepared in DMSO at 10 mM, while AEA was prepared in ethanol at 10 mM. Unless specified otherwise, all the compounds, growth factors, and solvents were purchased from Sigma-Aldrich (St. Louis, MO, USA). The working concentrations of the compounds were prepared by serial dilution of stock solutions in DMSO and followed by dilution in the complete medium.

3.2.2. Cell viability measurement

ReNcell VM was cultured in 96-well tissue-culture plates at a density of 15  103 cells/well and exposed to a range of concentrations of the four individual compounds for

24 h. A Live/Dead® viability/cytotoxicity kit (Thermo Fisher Scientific, Waltham, MA,

USA) was used to evaluate the effect of each compound on cell survival (n = 4 wells/ concentration/ compound). Images of cells in individual wells were acquired using an automated fluorescent microscope (S+ scanner, Samsung Eletro-Mechanics, Co.

(SEMCO), South Korea). Calcein AM (0.25 µM) and ethidium homodimer (0.5 µM) solutions in PBS were freshly prepared before every experiment. NPC media was removed from each well, washed 3 times with PBS for 5 min each, and incubated with staining solution for 30 min. After incubation, cells were washed with PBS three times for 5 min

70 each. Cells in PBS were imaged (5 images/well) using a S+ scanner at 4 using Ex: 500 -

560 nm & Em: 580 - 660 nm filter for Calcein AM, and Ex: 540 - 610 nm & Em: 600 - 650 nm filter for Ethidium homodimer.

3.2.3 Assays for cellular mechanisms of toxicity

All assays were performed with an initial seeding density of 15  103 cells/ well.

3.2.3.1. DNA damage

After 24 h of compound exposure, surviving cells (n = 4 wells/ concentration/ compound) were stained with 15 µM of Hoechst 33342 (Thermo Fisher Scientific; Ex-352 nm/Em-416 nm) in 1 PBS (pH ~ 7.2) for 30 min to assess changes in nucleic DNA content. Nuclei were imaged using the S+ scanner with a 4 objective and a blue fluorescence filter from Semrock (DAPI-5060C-000). Hoechst 33342 binds to the double stranded DNA of viable and fixed cells. The media was removed from the well, washed 3 times with PBS for 5 min each. The cells were incubated with 15 µM of Hoechst 33342 for 30 min and washed 3 times with PBS for 5 min each. The wells were imaged (5 images/well) using S+ scanner at 4 using Ex: 340 - 405 nm & Em: 410 - 481 nm filter.

3.2.3.2. Mitochondrial membrane potential assay

Mitochondrial impairment, i.e., changes in mitochondrial membrane potential, in the presence of the compounds was assessed using tetramethyl rhodamine methyl ester

(TMRM; Thermo Fisher Scientific; Ex-545 nm & Em-600 nm). The cells were stained with 0.5 µM of TMRM after 24 h of compound exposure and imaged using the S+ scanner with a 4 objective and an orange filter from Semrock (TxRed-4040C-000). Cell permeant

TMRM dye (0.5 µM) was added to wells after removing media and washing 3 times for 5 min each with warm PBS. Cells were incubated for 30 min with TMRM and were washed

3 times with PBS for 5 min each after incubation. Cells were imaged (5 images/well) using

S+ scanner using Ex: 540 - 610 nm & Em: 600 - 650 nm filter, at 4 magnification.

3.2.3.3. YO-PRO®-1 assay to identify apoptotic cells

YO-PRO®-1 (Thermo Fisher Scientific; Ex-491 nm/Em-509 nm) identifies apoptotic cells via nucleic acid binding. Apoptotic cells are permeant to YO-PRO®-1, but non- permeant to propidium iodide which could stain only dead cells. After 24 h of compound exposure, live cells (n = 4 wells/ concentration/ compound) were incubated with 5 µM of

YO-PRO®-1 for 30 min. Using a Zeiss Axio Vert.A1 inverted fluorescence microscope

(Thornwood, NY, USA), at least 10 images were taken at random locations in each well.

The fold change in number of apoptotic cells in toxicant-exposed cases was calculated by normalizing it to controls.

3.2.3.4. Intracellular glutathione levels

ReNcell VM (n = 4 wells/ concentration/ compound) was stained with 50 µM of monochlorobimane (mBCl; Thermo Fisher Scientific; Ex-394 nm/Em-490 nm) after 24 h compound exposure to determine the levels of intracellular glutathione (GSH), which provides an indirect measure of cellular defense against oxidative stress. Lower GSH levels are associated with increased concentrations of glutathione disulfide (GSSG) leading to a decrease in GSH/GSSG ratio, a critical indicator of cellular health. The cells were imaged at 4 magnification using the S+ scanner with the blue filter. mBCl is a non-fluorescent cell permeant dye and only gives blue fluorescence when it conjugates with a thiol. Cells

72 were incubated in 50 µM of mBCl for 30 min after washing 3 times for 5 min each with

PBS. After incubation, cells were washed 3 times with PBS for 5 min each before imaging.

Immediately after washing, cells were imaged (5 images/well) using S+ scanner at 4 magnification using Ex: 340 - 405 nm & Em: 410 - 481 nm filter.

3.2.4. Biophysical and biomechanical measurements using AFM

ReNcell VM at a seeding density of 1.5105 cells were cultured in laminin-coated

50 mm AFM-specific petri dishes for 24 h and exposed to compounds for up to 36 h. All measurements were made using a MFP-3D-Bio AFM mounted on an inverted fluorescence microscope. Tip-less AFM cantilevers were modified by attaching a 4.5-μm polystyrene bead using epoxy (as detailed in section 2.2). The spherical bead indenter reduces the possibility of destructive deformation of cell due to smaller local strains applied and enables the force-indentation curves to be obtained over a bigger contact area compared to conventional AFM cantilever. The actual spring constant was determined using thermal calibration method in a clean culture dish containing DMEM (details in 2.3). The probe was submerged in media for 30 min to reach thermal and mechanical equilibrium. After thermal and mechanical stabilization, the drift was < 2 µm in x, y, and z directions. The 4.5

µm bead was used for excess contact area to compensate the 1-2 µm drift and yield higher

-spatially averaged measurements. Using an optical microscope, target cells were identified for indentation.

Force-indentation profile provided information for important mechanical properties such as elasticity, adhesion and tethers. The analysis of force-indentation curves was automated using proprietary software (Igor Pro 6.37). The cells were maintained at 37 °C

73 throughout the live-cell nanoindentation assay. For each experiment, 50-60 cells were randomly selected, and force curves were obtained between nucleus and cell margin at approach/retraction velocity of 5 µm/s. At least 60 force-indentation curves were obtained for each group of the cells (controls and compound-treated; 4, 12, 24 and 36 h). Using

Hertz’s contact model, the Young’s modulus was determined from the force-indentation curves using Poisson’s ratio as 0.5 for cells, tip radius of 2.25 µm, and an indentation depth of ~500 nm (details in 2.5.1).

The force required to separate AFM tip and cell surface describes the adhesion or force of adhesion (Fad). The tether forces (FT) were directly calculated from the series of force steps in the retraction curve. The average noise of 50 pN was set as a threshold of tether force and only forces higher than threshold was considered as actual rupture events.

The apparent membrane tension (TM), the force needed to deform a membrane, was

2 2 calculated from tether forces using 푇푀 ≈ 퐹푇 ⁄8휋 퐾퐵, where KB is the bending stiffness

, which lies in the range of 0.1-0.3 pN.µm[203], [209] [210]. Similarly, RT (tether radius), which describes the connection between plasma membrane and cytoskeleton, was calculated from the tether forces as 푅푇 ≈ 2휋퐾퐵⁄퐹푇.

3.2.5. SOX2 staining for undifferentiated ReNcell VM

After 24 h of compound exposure, ReNcell VM was washed twice with 1 PBS, fixed with 4% paraformaldehyde for 30 min, washed twice with 1 PBS for 5 min each, and incubated with a permeabilizing/blocking buffer solution containing 5% goat serum and 0.5% triton-X in PBS for 30 min. After removing the permeabilizing/blocking buffer solution, the cells were incubated with mouse monoclonal anti-SOX2 (Thermo Fisher

Scientific) for 4 h. The cells were washed twice with 1 PBS and incubated with goat anti- mouse FITC secondary antibody (1:100; Santa Cruz Biotechnology) for 2 h at room temperature. The cells were washed twice with PBS, incubated with DAPI for 10 min, and washed once with PBS before imaging. At least five images were taken at random locations in each well using a 10 objective on the inverted fluorescence microscope.

3.2.6. Actin staining

After 4, 12, 24 and 36 h of compound exposure, ReNcell VM was washed once with 1 PBS, fixed with 4% PFA for 40 min, washed twice with PBS for 5 min, and incubated with 0.1% Triton X-100 in PBS for 5 min. The cells were washed twice with

PBS and incubated with actin-staining Alexa Flour 488 Phalloidin (Thermo Fisher

Scientific) for 25 min at room temperature. After removing the staining solution, the cells were washed twice with PBS for 5 mins and incubated with DAPI for 10 min before imaging. Images were taken at random locations using a 20 objective on the inverted fluorescence microscope. The images were used to calculate changes in cell area using

ImageJ.

3.2.7. Statistical analysis

ANOVA, based on the number of groups compared, followed by Tukey multiple comparison or Bonferroni post-hoc test, to find statistically significant differences between

75 the groups. p < 0.05 deemed statistically significant. Multivariate logistic regression of the parameters contributing to NPC health was performed using R software (version 3.5.2).

3.3. Results

3.3.1. Biochemical and biomechanical effects of rotenone exposure

To investigate the cytotoxic potential of rotenone on human NPCs, ReNcell VM was exposed to a range of rotenone concentrations (0.1 – 25 µM) for 24 h. A dose- dependent decrease in cell viability was observed with a calculated IC50 value of 0.27 ±

0.01 µM (Fig. 3.2A). To elucidate the subcellular mechanisms of rotenone toxicity, damage to nucleic DNA (IC50  0.28 ± 0.01 µM) and mitochondria (IC50  0.22 ± 0.01

µM), as well as changes in glutathione levels (IC50  0.33 ± 0.02 µM) were assessed (Fig.

3.2B). An increase in the number of apoptotic cells by YO-PRO®-1 staining was noted at

 0.78 µM concentration, beyond which necrosis might be dominating the mechanism of cell death (Fig. 3.2C). Based on these IC50 values, mitochondrial impairment seems to be the dominant mechanism by which rotenone influences a compromise in NPC health (p <

0.01 vs. other mechanisms; Fig. 3.12). However, no significant changes in SOX2 expression was noted in the surviving and adherent ReNcell VM upon rotenone exposure

(Fig. 3.11).

A B C 4 125 100 apoptosis necrosis 100 TMRM 3 75 Hoechst 75 mBCl 2 50 50 1 25 25

% live cells 0

0 0 5 6 7 8 apoptotic cells 5 6 7 8 5 6 7 8 Fold-change in

% stained cells Log [Rotenone (pM)] Log [Rotenone (pM)] Log [Rotenone (pM)]

Fig. 3.2. ReNCell VM were exposed to a range of rotenone concentrations (0.1 – 25 µM) for 24 h. Concentration-dependent cell survival (A), DNA damage, mitochondrial impairment, and changes in intracellular glutathione levels (B), and apoptosis-induced cell death (C) in rotenone-exposed cultures. The transition from apoptosis to necrosis as the mechanism of death is marked for visualization.

No significant changes in the biophysical and biomechanical characteristics (EY,

Fad, FT) of ReNcell VM were noted in control cultures at 4, 12, 24 and 36 h time points

(Fig. 3.13). The baseline properties of control ReNcell VM were: EY = 5.04  0.16 kPa, Fad

= 2.47  0.03 nN, FT = 329  3.2 pN, TM = 13.9  0.27 nN/m, and RT = 1.9  0.01 nm.

The mechanical properties of control ReNcell VM at the 24 h timepoint was used as a reference vs. test cases. Significant decreases in EY, Fad, FT, and TM with concomitant increases in RT were noted, as a function of increasing exposure duration (p < 0.001 vs. controls) and toxicant concentration (p < 0.001 vs. controls) in ReNcell VM exposed to rotenone (Fig. 3.3A-E). Significant concentration-dependent differences in biomechanical and biophysical characteristics of ReNcell VM were noted at every time-point, while such differences were more pronounced at higher concentrations with increasing exposure time.

A B 10 25 Control Rotenone 20 Rotenone exposed 0 µM 3.5 15 0.06 µM 8 10 Force, nN Force, 3.0 5 0.13 µM 0 0.27 µM 0 200 400 600 800 1000 1200 14001600 2.5 6 Indentation, nm

, nN 2.0

, kPa Y

4 ad 1.5 F E 2 1.0 0.5 0 0.0 C 4 12 24 36 4 12 24 36 4 12 24 36 C 4 12 24 36 4 12 24 36 4 12 24 36 Time,Time, hours h Time,Time, hours h

C D 20 400 184 pN 250 pN 271 pN 2 2

Force T = F /8 K

m M T B

300 15 

Indentation

200 10

, pN

, nN/ F

100 M 5 T 0 0 C 4 12 24 36 4 12 24 36 4 12 24 36 C 4 12 24 36 4 12 24 36 4 12 24 36 Time,Time, hours h Time, h

E 15

F = 2K /R

10 T B T , nm ,

T 5 R

0 C 4 12 24 36 4 12 24 36 4 12 24 36 Time,Time, hours h

Fig. 3.3. Biomechanical characteristics such as elastic modulus (A), force of adhesion (B), tether force (C), apparent membrane tension (D), and radius of tether (E) were quantified at various time points (4, 12, 24, and 36 h) and rotenone concentrations (0, 0.06, 0.13, and 0.27 µM) using AFM, and reported as average ± standard error. Representative force- indentation curves obtained from control and rotenone-exposed cells were shown in the inset in panel A. Elastic modulus was calculated by applying Hertz model to force- indentation curves (n  45 cells/condition) obtained from control and rotenone-exposed cells. Membrane tether forces (FT; n  50 cells/condition), apparent membrane tension 2 2 (푇푀 ≈ 퐹푇 ⁄8휋 퐾퐵), and tether radius (푅푇 ≈ 2휋퐾퐵⁄퐹푇) were measured by retraction of beaded-AFM probe from the cell surface. Here KB is the bending modulus,  0.1 pN.m

78 for lipid layers. Representative adhesive and tether forces obtained from force-indentation curves were shown in the insets in panels B and C.

ReNcell VM were immunolabeled for actin to detect potential alterations to cytoskeletal structure in the cells treated with varying rotenone up to 36 h (Fig. 3.10A).

While NPCs appeared more spread with distinguishable actin filaments in control cultures, even after 36 h, an exposure duration-concentration dependent compromise in cellular morphology was evident in rotenone-exposed cultures. Significant actin reorganization with more protein aggregation at the periphery and depletion in the cytoplasm was noted with rotenone addition. Similarly, the average cellular area of ReNcell VM (~1650 m2) significantly decreased at all rotenone concentrations tested and at every time point (Fig.

3.10B).

3.3.2. Biochemical and biomechanical effects of digoxin exposure

A concentration-dependent decrease in ReNcell VM viability was noted after 24 h exposure to a range of digoxin concentrations (0.2 – 50 µM), with IC50  0.56 ± 0.01 µM

(Fig. 3.4A). The mechanisms by which digoxin induced cellular damage was assessed by quantifying changes in DNA (IC50  2.02 ± 0.09 µM) and mitochondria (IC50  0.94 ± 0.02

µM), and intracellular glutathione levels (IC50  0.84 ± 0.02 µM) (Fig. 3.4B). Based on

® these IC50 values, plasma membrane integrity (Live/Dead assay) seems to be the dominant mechanism by which ReNcell VM is influenced by digoxin (p < 0.01 vs. other mechanisms; Fig. 3.12). An increase in the number of apoptotic cells was observed at 

0.39 µM digoxin (Fig. 3.4C), and no changes in SOX2 expression was observed within

79 surviving and undetached ReNcell VM upon digoxin exposure at  IC50 concentrations

(Fig. 3.11).

A B C 3 125 100 apoptosis necrosis 100 TMRM 75 Hoechst 2 75 mBCl 50 50 1 25 25 % live cells 0 0 0 5 6 7 8 5 6 7 8 apoptotic cells 5 6 7 8 Fold-change in % stained cells Log [Digoxin (pM)] Log [Digoxin (pM)] Log [Digoxin (pM)]

Fig. 3.4. ReNCell VM were exposed to a range of digoxin concentrations (0.2 – 50 µM) for 24 h. Concentration-dependent cell survival (A), DNA damage, mitochondrial impairment, and changes in intracellular glutathione levels (B), and apoptosis-induced cell death (C) in digoxin-exposed cultures. The transition from apoptosis to necrosis as the mechanism of death is marked for visualization.

Similar to the trends noted for rotenone exposure, significant decreases in EY, Fad,

FT, and TM with concomitant increases in RT was noted with increasing digoxin concentration (p < 0.001 vs. controls) and exposure time (p < 0.001 vs. controls) (Fig.

3.5A-E). At a fixed digoxin concentration, significant differences in the observed properties were noted with increasing exposure duration; similarly, at every time point tested, significant differences in mechanical and physical characteristics was evident with increasing digoxin concentration. Significant digoxin concentration and exposure time dependent loss in actin meshwork and cytoskeleton reorganization were evident in ReNcell

VM (Fig. 3.10A). Except for 0.14 M concentration and 4 h treatment, digoxin exposure significantly compromised the cellular area at all concentrations and time points tested vs. controls (Fig. 3.10C).

A B 10 3.5 Digoxin 8 0 µM 3.0 0.14 µM 2.5 0.28 µM 6 0.56 µM

, nN 2.0 , kPa ad 1.5

Y 4 F

E 1.0 2 0.5 0 0.0 C 4 12 24 36 4 12 24 36 4 12 24 36 C 4 12 24 36 4 12 24 36 4 12 24 36 Time, h Time, h

C D 400 20

300 m 15 

200 10

, pN

, nN/ F

100 M 5 T 0 0 C 4 12 24 36 4 12 24 36 4 12 24 36 C 4 12 24 36 4 12 24 36 4 12 24 36 Time, h Time, h

E 15

10 , nm ,

T 5 R

0 C 4 12 24 36 4 12 24 36 4 12 24 36 Time, h

Fig. 3.5. Biomechanical characteristics such as elastic modulus (A), force of adhesion (B), tether force (C), apparent membrane tension (D), and radius of tether (E) were quantified at various time points (4, 12, 24, and 36 h) and digoxin concentrations (0, 0.14, 0.28, and 0.56 µM) using AFM, and reported as average ± standard error. Elastic modulus was calculated by applying Hertz model to force-indentation curves (n  45 cells/condition) obtained from control and digoxin-exposed cells. Membrane tether forces (FT; n  50 2 2 cells/condition), apparent membrane tension (푇푀 ≈ 퐹푇 ⁄8휋 퐾퐵), and tether radius (푅푇 ≈ 2휋퐾퐵⁄퐹푇) were measured by retraction of beaded-AFM probe from the cell surface.

3.3.3. Biochemical and biomechanical effects of AEA exposure

A concentration-dependent decrease in cell viability with IC50  8 ± 0.13 μM was noted (Fig. 3.6A) upon ReNcell VM exposure to range of AEA concentrations (0.11 – 30

µM). The mechanisms by which AEA influences ReNcell VM was assessed by quantifying changes in intracellular glutathione levels (IC50  6.4 ± 0.33 µM), DNA damage (IC50 

11.4 ± 0.23 µM), and mitochondrial damage (IC50  4.3 ± 0.16 µM) (Fig. 3.6B). Based on these IC50 values, changes in mitochondrial membrane potential seems to be the dominant mechanism by which AEA affects ReNcell VM (p < 0.01 vs. other mechanisms; Fig. 3.12).

Apoptosis seems to be dominant mechanism up to 0.48 µM of AEA (Fig. 3.6C). NPC differentiation was not noted in AEA exposed cultures, as evident from SOX2 expression in surviving cells (Fig. 3.11).

A B C 125 100 4 100 apoptosis necrosis 75 3 75 2 50 50 TMRM Hoechst 1

25 25 mBCl % live cells 0 0 0

5 6 7 8 apoptotic cells 5 6 7 8 5 6 7 8 Fold-change in Log [AEA (pM)] % stained cells Log [AEA (pM)] Log [AEA (pM)]

Fig. 3.6. ReNCell VM were exposed to a range of AEA concentrations (0.11 – 30 µM) for 24 h. Concentration-dependent cell survival (A), DNA damage, mitochondrial impairment, and changes in intracellular glutathione levels (B), and apoptosis-induced cell death (C) in AEA-exposed cultures. The transition from apoptosis to necrosis as the mechanism of death is marked for visualization.

Significant decreases in EY, Fad, FT, and TM and increase in RT over 36 h (p < 0.001 vs. controls) and higher AEA concentrations (p < 0.001 vs. controls) were noted (Fig. 3.7A-

E). At every time point tested, significant compromise in mechanical properties of ReNcell

VM was noted with increasing AEA concentrations. Similarly, significant changes in

82 biomechanical characteristics of ReNcell VM was observed with increasing exposure duration at a fixed AEA concentration. AEA exposure led to time and concentration dependent upregulation in F-actin degradation and cytoskeleton reorganization (Fig.

3.10A), and significant compromises in the cell area (Fig. 3.10D).

A B 10 3.5 AEA 0 µM 3.0 8 2 µM 4 µM 2.5 6 8 µM

, nN 2.0 , kPa ad 1.5

Y 4 F

E 1.0 2 0.5 0 0.0 C 4 12 24 36 4 12 24 36 4 12 24 36 C 4 12 24 36 4 12 24 36 4 12 24 36 Time, h Time, h

C D 400 20

m 15

300 

200 10

, pN

, nN/ F

100 M 5 T

0 0 C 4 12 24 36 4 12 24 36 4 12 24 36 C 4 12 24 36 4 12 24 36 4 12 24 36 Time, h Time, h

E 15

10 , nm ,

T 5 R

0 C 4 12 24 36 4 12 24 36 4 12 24 36 Time, h

Fig. 3.7. Biomechanical characteristics such as elastic modulus (A), force of adhesion (B), tether force (C), apparent membrane tension (D), and radius of tether (E) were quantified at various time points (4, 12, 24, and 36 h) and AEA concentrations (0, 2, 4, and 8 µM) using AFM, and reported as average ± standard error. Elastic modulus was calculated by

83 applying Hertz model to force-indentation curves (n  45 cells/condition) obtained from control and AEA-exposed cells. Membrane tether forces (FT; n  50 cells/condition), 2 2 apparent membrane tension (푇푀 ≈ 퐹푇 ⁄8휋 퐾퐵), and tether radius (푅푇 ≈ 2휋퐾퐵⁄퐹푇) were measured by retraction of beaded-AFM probe from the cell surface

3.3.4. Biochemical and biomechanical effects of chlorpyrifos exposure

When ReNcell VM were treated with chlorpyrifos (0.1 – 50 μM), a concentration- dependent decrease in cell viability was observed with IC50  9.9 ± 0.17 μM (Fig. 3.8A).

The subcellular mechanisms of chlorpyrifos toxicity was assessed by quantifying changes in mitochondrial membrane potential (IC50  5.5 ± 0.23 µM), nucleic DNA damage (IC50

 11.6 ± 0.45 µM), and intracellular glutathione levels (IC50  8.25 ± 0.25 µM) (Fig. 3.8B).

Based on these IC50 values, mitochondrial damage seems to be the dominant mechanism by which chlorpyrifos influences a compromise in ReNcell VM health (p < 0.01 vs. other mechanisms; Fig. 3.12). An increase in the number of apoptotic cells was noted up to 0.78

µM chlorpyrifos, after which necrosis might be dominant (Fig. 3.8C). No changes in SOX2 expression was evident upon chlorpyrifos exposure (Fig. 3.11).

A B C 125 3 100 apoptosis necrosis 100 75 2 75 50 50 TMRM Hoechst 1

25 25 mBCl % live cells 0 0 0

5 6 7 8 5 6 7 8 apoptotic cells 5 6 7 8 Fold-change in Log [Chlorpyrifos (pM)] % stained cells Log [Chlorpyrifos (pM)] Log [Chlorpyrifos (pM)]

Fig. 3.8. ReNCell VM were exposed to a range of chlorpyrifos concentrations (0.1 – 50 µM) for 24 h. Concentration-dependent cell survival (A), DNA damage, mitochondrial impairment, and changes in intracellular glutathione levels (B), and apoptosis-induced cell death (C) in chlorpyrifos-exposed cultures. The transition from apoptosis to necrosis as the mechanism of death is marked for visualization.

A concentration-time dependent decrease (p < 0.001 vs. controls) in EY, Fad, FT, and TM and concomitant increase in RT (p < 0.001 vs. controls) of ReNcell VM were observed upon exposure to chlorpyrifos (Fig. 3.9A-E). At a fixed chlorpyrifos concentration, a significant drop in biophysical and biomechanical characteristics of

ReNcell VM was observed with increasing time. Similarly, at every exposure time point tested, a significant decrease in ReNcell VM biomechanics was observed with increasing concentration. Finally, gradual changes in cytoskeleton organization with actin down- regulation (Fig. 3.10A) and shrinkage of the cell area (Fig. 3.10E) were observed with increasing chlorpyrifos concentration and exposure time.

A 10 B 3.5 Chlorpyrifos 0 µM 3.0 8 2.5 µM 4.9 µM 2.5 6 9.8 µM

, nN 2.0 , kPa ad 1.5

Y 4 F

E 1.0 2 0.5 0 0.0 C 4 12 24 36 4 12 24 36 4 12 24 36 C 4 12 24 36 4 12 24 36 4 12 24 36 Time, h Time, h

C D 400 20

300 m 15 

200 10

, pN

, nN/ F

100 M 5 T 0 0 C 4 12 24 36 4 12 24 36 4 12 24 36 C 4 12 24 36 4 12 24 36 4 12 24 36 Time, h Time, h

E 15

10 , nm ,

T 5 R

0 C 4 12 24 36 4 12 24 36 4 12 24 36 Time, h

Fig. 3.9. Biomechanical characteristics such as elastic modulus (A), force of adhesion (B), tether force (C), apparent membrane tension (D), and radius of tether (E) were quantified at various time points (4, 12, 24, and 36 h) and chlorpyrifos concentrations (0, 2.5, 4.9, and 9.8 µM) using AFM, and reported as average ± standard error. Elastic modulus was calculated by applying Hertz model to force-indentation curves (n  45 cells/condition) obtained from control and AEA-exposed cells. Membrane tether forces (FT; n  50 2 2 cells/condition), apparent membrane tension (푇푀 ≈ 퐹푇 ⁄8휋 퐾퐵), and tether radius (푅푇 ≈ 2휋퐾퐵⁄퐹푇) were measured by retraction of beaded-AFM probe from the cell surface.

Fig. 3.10. (A, B) Representative immunofluorescence images of ReNCell VM showing morphological changes, i.e., reorganization of F-actin mesh network, in the presence of various toxicants, over 36 h exposure. Cellular area was quantified at various concentrations of rotenone (C), digoxin (D), AEA (E), and chlorpyrifos (F), over a 36 h period. Scale bar: 25 µm

3.4. Discussion

For many years, the scientific and medical community believed that uterus provides a fortress to the developing offspring against the attack of environmental and pharmaceutical toxicants. During 1959-1961, this claim was disproved when mothers on morning sickness medication, thalidomide, gave birth to babies with severe ear, heart and limb malformations[211]. Around 10,000-12,000 babies were affected by the toxic potential of thalidomide. This event highlighted the need for better screening of pharmaceutical compounds and chemicals, and their metabolites, and investigate the mechanisms by which they could affect embryos during developmental stages[212]. A

89 white paper issued by PPTOX III conference in 2012 highlights the role of environmental exposures and nutrients during development on subsequent diseases/dysfunctions later in life, and alludes to the developmental phase as plastic, i.e., allows the organism to respond and adapt to the surrounding environment[211]. Plasticity is more profound during early stages of development when cells are differentiating, and tissues are forming in all the three layers (endoderm, mesoderm, ectoderm). Since the sensitivity of cells and tissues to subtle changes in microenvironment during this phase is more profound compared to a fully developed tissue, any interference during these early stages could lead to adverse consequences during later stages of life. The in utero and early postnatal development phases are the most sensitive to toxic environmental exposures and nutritional factors.

Therefore, fundamental research on developmental neurotoxicity and disease prevention strategies are critically required to protect growing fetuses from developmental toxicity[211],[212].

Recent studies have demonstrated vulnerability of a developing fetus when it encounters even trace amounts of environmental toxicants which are harmless to most adults[196]. The developing CNS is susceptible to damage by exposure to toxicants primarily due to the still immature blood-brain barrier. Since NPCs are among the most sensitive cell types, they are commonly used as model cells to screen compounds for their neurotoxic potential. A few recent studies, including from our group, have evaluated the role of various toxicants (e.g., heavy metals, pharmaceutical compounds, PCBs) on the sensitivity of immortalized and primary NPCs isolated from various species (e.g., murine, human)[188],[66]. Since insulating NPCs from toxic insult during embryogenesis is crucial not only for proper CNS development but also during later stages of life, the goal of this

90 study is to screen a variety of compounds used in daily lives for their cytotoxic potential on NPCs, evaluate the sub-cellular mechanisms by which these compounds induce damage to these cells, elucidate the biomechanical and biophysical changes in these cells upon toxicant exposure, and finally correlate the biochemical and biophysical changes. To our knowledge, we report here for the first time on the biophysical and biomechanical changes in human NPCs upon exposure to toxic compounds, including the four compounds tested.

Specifically, we here show that although most cells survived for more than 24 h at very low dosages of the compounds (e.g., 0.25  IC50), their mechanical properties were severely compromised even after 4 h of incubation.

Rotenone, a plant-derived product and commonly used as an insecticide, blocks electron flow from NADH to co-enzyme Q. Prior studies reported on the IC50 values of rotenone (0.02 - 200 M) depending on the cell type[47], and rotenone-induced apoptosis on various in vitro platforms. We observed a concentration-dependent NPC death upon rotenone exposure, and found NPCs to be the most sensitive to mitochondrial damage as reported by others[197],[192]. Similar results were observed when neuroblastomas were exposed to rotenone, leading to reduction in intracellular ROS and decreased intracellular glutathione levels, both contributing to apoptosis[213]. NPCs exposed to rotenone showed caspase 9/3 independent apoptosis, a time-dependent release of cytochrome c and apoptosis-inducing factor, following mitochondrial depolarization-based increase in ROS generation, leading to apoptosis[192]. Despite such information, the biophysical and biomechanical changes induced by rotenone on NPCs (or other cells) remained unexplored.

Fig. 3.11. Representative immunofluorescence images of SOX2 expression in ReNcells VM exposed to various concentrations (0, 0.25IC50, 0.5IC50, IC50) of rotenone, digoxin, AEA, and chlorpyrifos for 24 h. Cultures were counterstained with DAPI for cell identification. Primary antibody for SOX2 was used in combination with appropriate secondary antibody. Scale bar: 100 µm.

Digoxin is one of the most widely used cardiac medication recommended by the

American Heart Association, although it does not have a favorable efficacy and safety profile. Various studies have evaluated digoxin’s neurotoxic potential[19],[214] and found that to be toxic to human NPCs but not rat NPCs; similar data was obtained for other cardiac glycosides. Digoxin was also reported to be toxic to iPSCs, neurons and fetal astrocytes[214]. Cells exposed to digoxin showed a distinct morphology which can be

92 correlated to its mechanisms of toxicity. Digoxin was found to have an inhibiting action on neuroblastoma growth in vitro, and on the growth of murine or human neuroblastoma in vivo[214]. In agreement with previous studies[215],[216], our results have shown a compromise in membrane integrity upon digoxin exposure, leading to dose-dependent cytotoxicity. This specificity of digoxin on cells of CNS origin suggests an urgent need for in-depth testing of its developmental neurotoxicity[19].

0.40

A M] B M]

 2.0 

, 0.35

, 50

50 1.5 0.30 1.0 0.25 * 0.5 * 0.15

Digoxin [IC Digoxin 0.3

Rotenone [IC 0.00 0.0 L/D Hoechst TMRM mBCl L/D Hoechst TMRM mBCl

C D M] 12 

, 12

M] 50  10

, 10 50 8 8 6 6 * 4 * AEA [IC 3 4 0 0 L/D Hoechst TMRM mBCl Chlorpyrifos[IC L/D Hoechst TMRM mBCl

Fig. 3.12. The IC50 values from the four assays were compared to identify the most sensitive mechanism by which rotenone (A), digoxin (B), AEA (C), or chlorpyrifos (D) induces toxicity in human NPCs. All assays were done at the 24 h time point.

Chlorpyrifos is one of the most widely investigated organophosphates for developmental neurotoxicity[190], and its regulation and usage remain highly controversial. Despite much work done to evaluate the toxic effects of chlorpyrifos, the underlying mechanism remains unknown, which projects an urgent need for better assays

, to elucidate the toxic nature of chlorpyrifos[18] [217]. The IC50 value of chlorpyrifos we

93 obtained is in the range of previous studies (0.8 – 56 M) noted in human brain cells[217],[18]. Studies have shown the effects of chlorpyrifos not only during early stages of brain development in murine models, but also in the later stages of brain development.

For instance, chlorpyrifos was shown to induce neurobehavioral abnormalities during second and third postnatal weeks in rats, which corresponds to human neonatal stages[194].

Chronic exposure of chlorpyrifos has been linked to neurocognitive and neurobehavioral deficits, with glial cells more susceptible than neurons[196]. It has been reported that oxidative stress, inflammation, and/or irreversible neuropathies might be the mechanisms involved in such chronic exposure. Experimental studies utilizing rodent models showed that pre- and post-natal exposure of chlorpyrifos leads to alternation in many cellular processes, including proliferation and DNA replication[18]. Various studies have shown the differential effect of chlorpyrifos exposure; in vitro models suggested an increase in neuronal apoptotic process and a decrease in glial proliferation[18],[218]. In our study, mitochondrial impairment and glutathione depletion by oxidative stress are the main mechanisms of chlorpyrifos-induced cytotoxicity along with apoptosis, in agreement with previous studies[219],[220].

A 15 B 4

10 , nN

, kPa 2

Y ad

5 F E 1

0 0 4 12 24 36 4 12 24 36 Time, h Time, h C 450 400

350 , pN

T 300 F 250 200 4 12 24 36 Time, h

Fig. 3.13. Average (± standard error) elastic modulus (A), adhesion force (B) and tether force (C) of human NPCs in control cultures for 4, 12, 24 and 36 h (n  45 cells in each case). No significant changes were observed in the biomechanical properties in control cultures over time.

In recent years there has been a spike in the use of recreational drugs and other cannabis. The biologically active compound of cannabis, Δ9-tetrahydrocannabinol, affects by binding to cannabinoid receptors. The long-term motor defects in the offspring and involvement of these compounds in neural differentiation was reported[221], [222]. The use (abuse) of these compounds during pregnancy poses an existential threat to the developing fetus. However, lack of pertinent data dictates an urgent need to elucidate the mechanisms underlying the cytotoxic potential of AEA[221], [222]. Most of these recreational drugs are cationic lipophilic molecules which can readily cross placenta and blood-brain barrier to influence amniotic fluid and fetal tissues[202]. On the other hand, some studies have shown the neuroprotective nature of endocannabinoids, both in vivo and in vitro. AEA offered protective biological effects after brain injury and in neuronal cultures mimicking ischemic conditions, making them a valuable target for drug discovery.

Our results show a dose-dependent AEA-induced cell death in NPC cultures, with mitochondrial damage driving the mechanism of cell sensitivity to AEA below IC50 levels, followed by intracellular glutathione levels. Apart from their neuroprotective effects, studies have shown the cytotoxic effects of AEA in cortical and hippocampal neuron cultures, mainly mediated by caspase-3 dependent apoptosis and ROS generation[223].

Using YO-PRO®-1 assay which only stains for apoptotic cells, we observed an increase in apoptotic cells up to a certain concentration for each toxicant tested, followed by a decrease in cells staining positive for apoptotic marker, although cell death continued to monotonically rise with increasing toxicant concentration. While the crosstalk between signaling networks involved in the mechanisms of cell death (apoptosis vs. necrosis) could not be completely decoupled, numerous studies attest to the role of apoptosis as the dominant mechanism at low concentrations and necrosis taking over that role at higher dosages of a variety of toxicants[224], [225], similar to that noted in our studies. Exposure of cigarette smoke condensate led to activation of p53-mediated activation of apoptotic signaling at low concentrations and interruption in apoptotic signaling leading to necrosis at higher concentrations[225]. Similarly, low taxol concentrations led to blocking of mitosis and onset of apoptosis whereas higher levels led to microtubule polymerization and necrosis induction[224]. A hallmark of necrosis is the loss of membrane integrity[226],[227]. Our AFM analysis showed that apparent membrane tension decreased with both increasing concentration and exposure time (Fig. 3.3D, 3.5D, 3.7D, 3.9D), indicating loss of membrane integrity and induction of necrosis.

To our knowledge, there is no existing literature on the characterization of cellular biomechanics of human NPCs, nor the effect of these four compounds on biomechanical

96 characteristics of any cell types. Therefore, our study addresses a critical gap in literature and establishes a baseline for biochemical and biophysical changes in NPCs in the presence of toxicants. Our results showed a significant effect of toxicant exposure on cellular elasticity, adhesion, tether forces, radius of tether, and membrane tension, compared to untreated NPCs. The elastic modulus of NPCs decreased by ~80% upon exposure to 0.27

µM of rotenone for 24 h. In select cases, prolonged exposure of lower toxicant concentrations showed similar effects on biomechanical properties as short-term exposure at higher concentrations, suggesting the need to extend the extent of toxicity testing beyond

IC50 values. Two-way ANOVA revealed significant primary effects of concentration (p <

0.001), exposure time (p < 0.001), and interaction between concentration and exposure time (p < 0.001) on NPCs for all the compounds tested.

We correlated the change in Young’s modulus to cell death measured from the most sensitive mechanistic assay (e.g., mitochondrial impairment), pooled for all four different neurotoxicants (Fig. 3.14). A strong negative correlation (R2 = 0.92) between the normalized elastic modulus (to controls) of surviving cells and the percentage of living cells in that environment was noted, suggesting that (a) the health of organelles in a toxic environment strongly influences cell mechanics, (b) cell biomechanics could also be correlated to biochemical outcomes, in addition to biophysical changes, and (c) biomechanical characteristics (modulus, tether forces, adhesion) are key markers and predictors of neurotoxicity. In studies exploring role of silver nanoparticles on human embryonic kidney cells, a similar strong negative correlation between DNA damage and factor of viscosity was noted[228].

Multivariate logistic regression analysis[229],[230] suggested that elastic modulus

(p < 0.001), adhesion force (p < 0.0001), and tether force (p < 0.0001) were significant predictors of NPC toxicity. The receiver operating characteristic (ROC) was analyzed to gauge the utility of elastic modulus, adhesion force, and tether force as reliable indicators of NPC health, resulting in the model given by:

M = 1.672 – 0.033  EY – 0.109  Fad – 0.003  FT

where EY (elastic modulus) is in kPa, Fad (adhesion force) is in nN, and FT (tether force) is in pN units. Based on the Akaike information criterion (AIC) calculated from this model[231], changes in tether force appears to be the strongest predictor of neurotoxicity.

2.0 Linear Regression 2 1.5 (r = 0.92)

1.0 change in in change

- 0.5 Fold

Young’s Modulus Young’s 0.0 Young'sModulus, kPa C 12.5 25 50 Cell Death, %

Fig. 3.14. Changes in elastic modulus of human NPCs was correlated to compromise in cell death calculated from most sensitive assay upon toxicants exposure. A strong negative correlation (R2 = 0.92) was noted suggesting that the biomechanical properties can be correlated to sub-cellular mechanism of cell death induced by toxicants.

The time-concentration dependent decrease in adhesion force upon toxicant exposure suggests an alteration of cell membrane, possibly the modification of surface adhesion molecules. The polysaccharides on cell surface and secreted ECM proteins are primarily responsible for cellular adhesion. Typically, an increase in adhesion force suggests an aggregation of polysaccharides, while a decrease in adhesion force suggests denaturation or detachment of polysaccharides from cell surface. Cell adhesion molecules such as integrins, proteoglycans, selectins, cadherins, oligosaccharides, and immunoglobulins mediate cell-cell and cell-ECM interactions[121]. Exposure to heavy metals (e.g., lead, methylmercury) and organic sulfates led to altered expression and function of cell adhesion molecules such as cadherins and IgCAMs[232]. Upon exposure of glyphosate (a herbicide), a concentration-dependent reduction in cell adhesion was observed in MC3T3-E1 cells, via blocking of RGD-specific integrin[233]. Inhibiting fibronectin-integrin binding leads to a strong decrease in adhesion force of primary gastrulating cells as detected by AFM based analysis[146]. Adhesion force is an encouraging biomarker for changes in cell surface due to inflammation or differentiation[234],[235]. Our observations on changes in adhesion forces of NPCs are in line with reports on human aortic endothelial cells exposed to diesel exhaust particles[133].

Various techniques (e.g., micropipette aspiration, optical-tweezers, magnetic twisting cytometry, microfluidics, AFM) have been developed to study cell biomechanics, to understand how cells feel under transformations and perturbation due to mechanical forces. Optical tweezer analysis of erythrocytes and leukemic cells exposed to doxorubicin showed a decrease in cell modulus, demonstrating the need for biomechanical assays to screen the toxic concentration of drugs and prevent the risk of vascular complications due

99 to high dosages[236]. The unique capabilities of AFM such as non-invasive characterization, measurements under physiological conditions, precise control on the magnitude and frequency of the applied forces, and high spatial control, makes it the most suitable technique to image and mechanically characterize either fixed or live cells or tissues.

From the retraction profile of force-indentation curve, we quantified the forces involved in individual and multiple membrane tethers on NPCs in the presence and absence of the compounds (50 – 400 pN). Our results are in the range of tether forces measured using techniques such as AFM or optical tweezers, for a variety of different cells and vesicles[110],[155]. Toxicant-exposed NPCs showed higher values of tether radius as compared to control cells, indicating a weak connection between cytoskeleton and plasma membrane upon toxicant exposure. The average apparent membrane tension decreased from 13.9 nN/µm in control cells to 0.5 nN/µm upon exposure to 0.56 µM digoxin. Such compromise in apparent membrane tension follows the similar patterns observed with cellular stiffness and cellular adhesion. However, changes in adhesive forces, tether forces, or membrane tension would have less effect on cell elastic modulus compared to alterations in actin cytoskeleton.

Cell shape is maintained by the cortical tension, cell-matrix adhesion, cell-cell contact, and the microenvironment. Various developmental stages require changes in cell shape which contribute to morphogenetic processes. CNS diseases (e.g., Parkinson’s,

Alzheimer’s) and disorders (e.g., neural tube defects) have been shown to be associated with the rearrangement of the cytoskeleton[74]. Therefore, the significant time- concentration dependent changes in NPC cytoskeleton upon toxicant exposure is of interest

100 as it could lead to developmental defects in the short term (changes in NPC migration, differentiation) and neurodegenerative diseases in the long-term. Although genetic and epigenetic changes could be expected in NPCs exposed to toxicants, it was not the primary goal of this study and therefore will be investigated in future studies.

3.5. Conclusions

We investigated and correlated the cytotoxic, biophysical and biomechanical aspects of a variety of classes of compounds on human NPCs. We highlighted the role of cell biomechanics as a crucial marker and predictor of developmental neurotoxicity by demonstrating a compromise in cellular mechanics upon exposure to four different classes of toxicants. A negative linear correlation between the levels of cell death in vitro upon toxicant exposure and the intrinsic changes in cell elastic modulus was noted. We propose the term mechanotoxicology to ascribe cellular mechanics as an important toxicity endpoint and validated the utility of developmentally-relevant human NPCs as an appropriate cell model for compound screening. We also elucidated the underlying mechanisms of cell death and deterioration of cell health associated with toxicant exposure. Our current studies are geared towards investigating genetic changes corresponding to perturbation of cellular mechanics, as well as investigating the effects of these compounds on NPCs differentiation.

101

CHAPTER IV

ROLE OF KINDLIN3 IN MEMBRANE TO CORTEX ATTACHMENT AND

FORCE TRANSMISSION

4.1. Introduction

CNS is mechanically heterogenous due to varied cellular composition and cell density[177]. Microglia are specialized macrophages of the CNS and are involved in immune regulation, tissue development, homeostasis and wound repair[29], [237], [238].

Microglia respond to variations in tissue stiffness by altering their membrane and cytoskeletal tension to induce changes in their morphology, activation status, and migration[29], [239], [239]–[241]. Macrophages, on the other hand, play a significant role in various diseases and could experience mechanical forces ranging from a few pN in suspensions to several orders of magnitude higher in tissues[47]. Macrophages are less adherent while migrating through soft and healthy tissues, but could become adherent and activated, and change their shape when they sense stiff or damaged tissue regions, owing to their mechanosensing ability. Such migration or matrix attachment is primarily mediated by transmembrane integrins that induce changes in both cortical and deep cytoskeleton in response to their environment[5], [90], [242]. Elongation stress or stretch also regulates cell tension leading to changes in macrophage shape and phagocytosis[242].

102

Mechanosensing via membrane-to-cortex attachment (MCA) proteins appears to be a major regulator of microglia and macrophage activation, cytokine production, phagocytosis, and proliferation, though the exact mechanism and specific proteins involved is still emerging.

Cell membrane could be approximated as a heterogenous bilipid layer with glycocalyx on the surface and numerous proteins embedded within (e.g., integrins), that undergoes compression, expansion, bending, shear, or viscoelasticity under external forces.

A key factor of mechano-transduction is membrane tension, generated from the in-plane tension within the lipid bilayers of the membrane, in response to external or internal forces on the membrane surface[243]. It plays a key role in cellular characteristics and functions including cell shape, size, migration, cytokinesis and intracellular trafficking[203], [244],

[245]. The membrane tension deforms the plasma membrane and the extracellular force is transmitted from surface integrins to the actin cytoskeleton[246] via focal adhesion (FA) complex, made up of a diverse array of proteins and their complexes, including a dynamic set of receptors, regulators and adapters (e.g., talins and kindlins) which bind integrins at the primary level[2], [3], [95], [112], [145],[247]. Talins have been shown to contribute to

MCA by forming a relatively weak structural link between β-integrins and actin cytoskeleton thereby regulating the dynamics of adhesion proteins [40], [45], while the specific role of kindlins in this process is only recently being elucidated[248].

Kindlins are a family of membrane-bound cytoplasmic proteins consisting a C- terminal FERM domain, a PH (pleckstrin homology) domain inserted in the middle of the

FERM domain, and an N-terminal domain[46] (Fig. 4.1). The FERM domain (a highly conserved domain of 4.1, ezrin, radixin, and moesin) is commonly present in cytoskeleton-

103 associated proteins that interface with plasma membrane[24], [40]. Kindlins directly bind integrins through their FERM domain, specifically through the highly conserved

QW614/615 residues on F3 subdomain[249]. Mutations in QW614/615 on kindlins result in the lack of kindlin-integrin interactions and leads to diminished integrin-mediated cell adhesion, spreading, and migration[47]. The lack of PH domain disrupts the functions of kindlins analogous to the QW614/615 mutation[45]. The F0 domain on kindlins interacts with cytoskeletal proteins F-actin[24], [46] and paxillin[2]. The mechanical forces from the ECM are transmitted via the integrin-kindlin-actin “molecular clutch” to modulate the rate of actin polymerization[2].

Fig. 4.1. Integrin-binding specificity of kindlins[24].

The kindlin family constitutes three members that have differential tissue-specific expression. Kindlin1 is expressed specifically in keratinocytes, while kindlin2 (K2) is ubiquitous except for hematopoietic cells which express kindlin3 (K3)[40], [247]. In humans, the importance of K3 is underscored by the lack of leukocyte adhesion and integrin activation in hematopoietic cells leading to life-threatening complications.

Deficiency of K3 in hematopoietic cells ablates activation of three families of integrins

104

(1, 2, 3) causing devastating bleeding and immune disorder in humans[249]. Defective integrin activation on platelets and blood cells did not account for the neurological complications observed in K3-deficient patients[249], indicating a role for microglia – the only cells in the CNS to express K3.

Therefore, in this study, using AFM and K3-deficient cell models, we investigated the role of integrin-K3 complex in membrane tension and MCA of hematopoietic cells

(microglia & macrophages). To aid these studies, we used rescued K3 knockout (K3KO) raw cells, and rescued K3KO cells generated by CRISPR-Cas9 technology via overexpression of hK3 using lentivirus infection. Results helped establish K3 as a main regulator of membrane mechanics and MCA, and K3-integrin binding at the membrane interface as responsible for K3-mediated MCA and membrane tension.

4.2. Methods

4.2.1. Animal models and microglial cells

The cells used in this study were provided by Dr. Tejasvi Dudiki at the Cleveland

Clinic from respective animal models – WT, K3 gene knockouts, and K3 gene knockins.

Cells containing mutation on K3 gene were generated from K3hypo mice as previously described[250] and hitherto would be referred to as K3hypo cells. The K3hypo mice were crossed with the FLP1 expressing mice from Jackson laboratories to remove the neomycin cassette, and the resulting mice referred to as K3mut (K3 mutant knock-in) showed increased expression of mutant K3 matching the endogenous K3 levels of WT mice. The cells isolated from these mice were hitherto referred to as K3mut cells. The K3KO mice (K3- deficient) were generated by crossing K3 floxed mice with CX3CR1-cre (tamoxifen

105 inducible) mice obtained from Jackson Laboratory. The cells derived from K3KO mice are hitherto referred to as K3KO cells.

Primary microglia were isolated from the brains of WT, K3hypo, K3KO and K3mut at postnatal day 1 (P1) mice. The cerebral hemispheres were minced and dissociated into a slurry in ice cold DMEM/F12 (with 20% FBS, 100 U/mL penicillin and streptomycin, 0.25

μg/mL amphotericin B and supplementation of non-essential amino acids (NEAA)). Any undissociated tissue pieces were removed by passing the slurry through a 70 µm sieve. The slurry was then centrifuged, pellet resuspended in DMEM/F12 and plated into T75 flasks.

Media was changed after three to four days to remove dead cells and the cultures grown for one week before microglia harvest. The microglia (WT, K3hypo, K3KO and K3mut) were harvested by collecting the conditioned media and centrifugation to pellet the microglia.

Western blot analysis was performed on these cells to confirm the presence or absence of

K3 protein expression, using primary antibodies against Rabbit anti-Kindlin-3 (Catalog# ab68040, Abcam), and Rabbit anti-GAPDH (D16H11; Catalog# 5174, Cell Signaling

Technology), and was performed by Dr. Duduiki at Cleveland Clinic.

4.2.2. Generation of K3-deficient macrophages cell lines

CRISPR-Cas9 technology was used to generate kindlin-3 knockout in Raw 264.7 cell line. Two independent sgRNAs were designed by the CRISPR Design Tool[251] and resulting cells were designated as K3KO1 & K3KO2 cells. Human kindlin3 was overexpressed in these cells using lentivirus expression and the cases studied here include:

K3KO1hk3, K3KO2hk3, K3KO1hk3int, and K3KO2hk3int. The K3 knockouts and overexpression of hk3 was verified by western blotting. All these studies were done by

Drs. Huan Liu and Tejasvi Dudiki at Cleveland Clinic.

106

4.2.3. Atomic force microscopy

The microglia and macrophages cells were plated onto cover glass and allowed for overnight attachment and spreading. The cover glass was adhered onto a 50-mm petri dish, and a Petri Dish Heater was used to maintain the cells at 37 °C in microglia conditioned media throughout the live-cell nanoindentation assay. All measurements were made using a MFP-3D-Bio AFM (Oxford Instruments) mounted on an inverted optical microscope using TR 400, PSA cantilevers (nominal spring constant ~ 0.02 N/m). The actual spring constant was determined from the force-distance curves and thermal calibration method using a clean culture dish containing DMEM, prior to each experiment. The probe was submerged in cell medium prior to each experiment to stabilize, minimize drift, and achieve thermal/ mechanical equilibrium. Around 35n 70 cells were randomly selected for each experiment and the force curves were obtained at various locations on each cell at approach/retraction velocity of 5 µm/s and a constant applied force of 2 nN. The force- indentation curves were analyzed using Igor Pro 6.37 software and Young’s modulus was calculated using Hertz’s contact model (assuming Poisson’s ratio of 0.5), after accounting for 15 nm tip radius and 500-650 nm indentation depth.

As per tether extraction theories based on free-energy minimization[252], the apparent membrane tension, TM experienced during bending under small strains was

2 2 calculated as 푇푀(≅ 퐹푇 ⁄8휋 퐾퐵). Here, KB is the bending stiffness (i.e., energy required to sustain altered molecular interactions on membrane under small extensional strains), and

FT is the measured tether force approximated as 2휋퐾퐵⁄푅푇, where RT is the radius of curvature of tether.

107

4.2.4. Statistics

All data histograms are expressed as mean ± SEM. Unpaired student t-test were performed to compare two groups and when relevant a between-groups analysis of variance

(ANOVA) was ran followed by appropriate post hoc analysis. A p-value less than 0.05 was considered significant.

4.3. Results

4.3.1. K3 in microglia is crucial for MCA

Cell responsiveness to matrix stiffness is tightly linked to the plasma membrane tension. Knowing that membrane tension increases during spreading, we compared K3 expression levels in WT, K3 hypomorph (K3hypo), K3 knockout (K3KO) and K3-integrin mutant knockin (K3mut) primary microglia cells prior to their spreading. The K3hypo and

K3KO microglia are K3 deficient used to determine the importance K3 protein expression.

The K3mut microglia expressed K3 bearing the mutation QW to AA (in the integrin binding domain) at levels similar to WT (Fig. 4.2). Young’s modulus was derived from the initial part of the force-displacement curves (Fig. 4.3a) obtained from AFM analysis. K3 deficiency in K3hypo and K3KO or K3 mutation in integrin-binding site (K3mut) had diminished Young’s modulus (5-fold in K3hypo/K3KO and 2-fold in K3mut) compared to

WT (Fig. 4.3b) reflecting low stiffness. The EY values had wider distribution from 750-

2500 Pa in WT cells, but a significant narrowing and shift to lower values in K3mut (400-

900 Pa), and K3hypo and K3KO (25-550 Pa) cases (Fig. 4.4a). Similarly, membrane tethers

(both single and multiple tethers) during retraction phase of the AFM tip from the cell surface showed that the absence of K3 or lack of K3 interaction with integrins resulted in

108 substantially lower tether forces compared to WT (Fig. 4.5a). The narrower distributions of tether forces and elastic moduli values (measured at multiple locations on numerous cells in each case) in K3KO cells (including K3hypo, K3mut) stand in contrast to their WT counterparts (Fig. 4.4), attesting to the homogeneity in the decoupling of membrane and cytoskeleton. Thus, the lack of K3 (in K3KO and K3hypo) or its interaction with integrin - subunit in K3mut substantially diminishes membrane to cortex attachment (MCA), indicating that the direct connection between K3 and integrin is crucial for membrane mechanics.

Kindlin3

GAPDH

Fig. 4.2. Western blot analyses of Kindlin3 expression in knockout and rescue cells. The western blots were probed with K3 antibody followed by GAPDH as loading control for primary microglia isolated from WT, K3hypo, K3KO and K3mut mice.

p 0.0001 p 0.0001 a b p 0.0001 2500 2.0 WT K3hypo 2000 ) 1.5 K3KO 1500 nN p 0.0001 mut 1.0 K3 p 0.0001 1000 p= n

Force( 0.5 500

0 Modulus (Pa) Young’s 0 0 500 1000 1500 2000 WT K3hypo K3KO K3mut Indentation (nm)

109

Fig. 4.3. K3 deficiency results in decrease of microglia Young’s modulus. a. Representative force-indentation curves showing significant decrease in force required for indentation of K3 deficient cells. b. A dot plot showing the cell stiffness (mean ± SE) represented as Young’s modulus (E, Pa) for WT, K3hypo, K3KO and K3mut cells. n = 68 (WT), 69 (K3hypo), 69 (K3KO) and 45 (K3mut) for the analyzed force-indentation curves.

a b 30 50 WT WT mut K3mut K3 25 40 hypo K3hypo K3 K3KO 20 K3KO 30 15

20 Count Count 10 10 5

0 0 0 500 1000 1500 2000 2500 0 50 100 150 200 Young's modulus, Pa Tether force, pN

Fig. 4.4. Histograms of Young’s modulus (a) and tether forces (b) in microglial cells.

110

a p 0.0001 p 0.0001 p 0.0001 p 0.0096 b 800 p 0.0001

150 p 0.0001 p 0.0017 p 0.0001 )

600 pN

= /µm) 100 p ns p=ns 400

pN p=ns

( p=ns M 50 T 200

Tether force( Tether 0

WT K3hypo K3KO K3mut

p=ns c 800 d p 0.0001 Tb  p 0.0001 25 p 0.0001 p=ns 600 p 0.0001

20 /µm) 400

pN 15 (

M 200 10 T

0 5

-100 radius (nm) Tether 0

Fig. 4.5. K3 deficiency results in loss of microglia cell membrane tension. a. Membrane tether forces (퐹푇 = 2휋퐾퐵⁄푅푇) of primary microglia measured by retraction of AFM tip from the cell surface (i.e., tether), where KB is the bending modulus, and RT is the tether radius. The average force is shown by the black bar in each case. b. The mean apparent 2 2 cell membrane tension (푇푀 = 퐹푇 ⁄8휋 퐾퐵) for each genotype. c. The bar graph on right shows the individual contributions from the bilipid layer tension (Tb) and MCA due to K3 hypo mut () that make up the apparent membrane tension TM. d. The K3 , K3KO and K3 cells have greater tether radius compared to WT. n = 106 (WT), 109 (K3hypo), 93 (K3KO) and 119 (K3mut) for the analyzed membrane tether forces, cell membrane tension and tether radius.

111

4.3.2. K3 in microglia regulates membrane tension

Considering typical KB values to range from 0.1 to 0.2 pN.μm for lipid bilayers[253], the average apparent membrane tension was three- and two- fold lower in

K3hypo/K3KO and K3mut respectively, compared to WT (Fig. 4.5b). In these cells, the apparent membrane tension (TM) is generated by a combination of the lipid bilayer tension

(Tb) and MCA () due to K3, i.e., TM = Tb + . Since K3KO cells completely lack K3 mediated MCA (i.e.,  = 0), their membrane tension is contributed solely by lipid bilayer tension, i.e., Tb = TM (K3KO). Thus, in WT cells,  = TM (WT) – TM (K3KO)  385 pN/µm, suggesting that it contributes to >65% of the overall MCA due to K3 (Fig. 4.5c). The radius of curvature of tethers (RT) from these cells, calculated from the respective tether forces, showed significantly higher values for K3hypo, K3KO and K3mut compared to WT cells

(Fig. 4.5d), indicating lower resistance to deformation and a weaker MCA in the absence of K3 (parameter values shown in Table 2).

Table 2. Average values of membrane parameters of microglia measured by AFM. Mean ± St. Error of all the parameters calculated as described in the experimental procedures. The top panel represents effects of K3 depletion in K3hypo & K3KO, or K3 mutation in K3mut primary microglial cells. Bottom panel shows fold-change in values in comparison to WT. Here E is the Young’s modulus, FT the tether force, TM the apparent membrane tension, RT the radius of tether.

WT K3hypo K3KO K3mut

E (Pa) 1308 ± 47.7 236.3 ± 14.6 249.3 ± 14.1 619.4 ± 17.4

FT (pN) 57.98 ± 3.5 33.55 ± 1.3 34.60 ± 1.5 44.26 ± 1.44

TM (pN/µm) 592.4 ± 91.1 164.4 ± 12.2 206.5 ± 33.2 279.3 ± 23.1

RT (nm) 14.48 ± 0.74 21.75 ± 0.8 20.84 ± 0.8 15.63 ± 0.42

112

4.3.3. Rescue of MCA and membrane tension with K3

We investigated whether K3 plays a similar role in macrophages, by generating two independent K3 knockout clones of RAW 264.7 cells (K3KO1 & K3KO2) using

CRISPR/Cas9 system. Absence of K3 protein in these cell lines was confirmed by western blot analysis (Fig. 4.6a). Similar to microglia, both K3KO1 and K3KO2 cells exhibited substantial changes in membrane mechanics. Representative force indentation curves for

WT (control), K3KO and K3KO1-hK3 were also shown (Fig. 4.7a). The average Young’s modulus and tether forces were reduced by ~70% and ~35% in the K3KO1 and K3KO2 cell lines respectively, compared to control cells (Fig. 4.7b, 4.8a), indicating loss of MCA.

This translated to a decrease in membrane tension by more than 2-fold in both the K3KO cell lines compared to the control (Fig. 4.8b). The tether radius, calculated from their respective tether forces, was ~35 % higher in K3KO cell lines (Fig. 4.8c). To confirm the role of K3, we re-expressed DsRed tagged-human Kindlin-3 (hK3) in both the K3KO lines

(K3KO1-hK3 & K3KO2-hK3) at protein levels similar to that in WT control (Fig. 4.6b).

As anticipated, re-expression of hK3 in K3KO1 and K3KO2 cells rescued the cell phenotype including the membrane characteristics; the membrane stiffness, tether forces

(MCA), apparent membrane tension, and tether radius, establishing K3 as a main regulator of membrane tension and MCA (Fig. 4.7b, 4.8a-c).

113

a b DsRed-hKindlin3 Kindlin3

Kindlin3 GAPDH GAPDH

Figure 4.6. Western blot analyses of Kindlin expression in knockout and rescue cells. The western blots were probed with K3 antibody followed by GAPDH as loading control. a. WT and two clones of K3 knockout RAW 264.7 cells – K3KO1 & K3KO2 – generated by CRISPR Cas9. b. RAW WT cells transfected with vector alone (control), K3 knockouts (K3KO1 and K3KO2), K3KO1 and 2 expressing DsRed-Tagged-human K3 (K3KO1-hK3 and K3KO2-hK3 respectively), and K3KO1 and 2 expressing DsRed-Tagged-human K3 integrin mutant (K3KO1-hK3-int and K3KO2-hK3-int respectively).

114

a 2.1 WTWT 1.8 K3KO12-7-2K3KO--hK3hK3 1.5 K3KO12-7K3KO 1.2 0.9 0.6 Force, nN 0.3 0.0 0 500 1000 1500 2000 Indentation, nm b * * * 1500 * * * * * * 1000 * *

500 Young'sModulus (Pa) 0

WT K3KO1 K3KO2

K3KO1-hK3 K3KO2-hK3 K3KO1-hK3-int K3KO2-hK3-int

Fig. 4.7. K3 deficiency results in decrease of microglia Young’s modulus. a. Representative force-indentation curves showing significant decrease in force required for indentation of K3 deficient cells and rescue K3 deficient cells with hK3. b. A dot plot showing the cell stiffness (mean ± SE) represented as Young’s modulus (E, Pa) for WT, K3KO1, K3KO1-hK3, K3KO1-hK3-int, K3KO2, K3KO2-hK3, K3KO2-hK3-int cells. n = 53 (WT), 44 (K3KO1), 50 (K3KO1-hK3), 26 (K3KO1-hK3-int), 30 (K3KO2), 30 (K3KO2-hK3) and 19 (K3KO2-hK3-int) for the analyzed force-indentation curves. * denotes p < 0.001

115

* * a * b * * * 120 * * * 600 * * * * * * * 80 * 400 * * *

40 (pN/µm)

M 200

T Tether force (pN) force Tether

0 WT K3KOK3KO-hK3K3KO1-hK3-intK3KO2K3KO2-hK3K3KO2-hK3-int 0 WT K3KO K3KO-hK3K3KO1-hK3-intK3KO2K3KO2-hK3K3KO2-hK3-int

c d * * * Tb Tb * 400 400 *   20 * *

* 300 300 /µm) 15 /µm)

200 200

( (

M T T 100 100 5

Tether radius Tether (nm) 0 0 0 -50 -50 WT K3KO K3KO-hK3K3KO1-hK3-intK3KO2K3KO2-hK3K3KO2-hK3-int

Fig. 4.8. Rescue of membrane tension with human K3 in macrophages. Two clones of K3 knockout RAW 264.7 cells K3KO1 and K3KO2 were generated using CRISPR/Cas9. Both the clones were then transfected with human K3 (K3KO1-hK3 & K3KO2-hK3) to rescue K3 function. a. Tether forces (FT) were calculated from the retraction part of force- indentation curves. b. The apparent membrane tension was calculated from the tether forces assuming KB, bending modulus, as 0.1 pN.µm. c. Membrane tether force was used to calculate the tether radius (RT). For a-c, n = 81 (WT), 59 (K3KO1), 77 (K3KO1-hK3), 32 (K3KO1-hK3-int), 40 (K3KO2), 39 (K3KO2-hK3) and 20 (K3KO2-hK3-int) for the analyzed force-indentation curves. * denotes p < 0.001

116

4.3.4. K3-integrin interaction at the membrane is a major contributor to MCA

It was previously shown that mutation in QW614/615 of F3 domain of K3 results in the lack of K3-integrin interactions and leads to diminished integrin-mediated cell functioning[47]. This mutation in microglia results in significantly lowered Young’s modulus, membrane tension, tether forces and tether radius were observed in the K3mut microglia compared to WT (Fig. 4.3b, 4.5a-d). To confirm the role for K3-integrin interactions in MCA, we re-expressed the DsRed tagged-F3 domain mutant human

Kindlin-3 (hK3int- bearing the mutation QW614/615 to AA614/615 in the F3 domain) in both the K3KO macrophage lines at similar protein levels (Fig. 4.6b). Re-expression of hK3int in K3KO cells partly rescued mechanical defects of K3KO cells. The distributions of EY and FT values in control, K3KO1 and K3KO2 RAW cells, and in their respective hK3 and hK3int genotypes were shown in Fig. 4.9. The average Young’s modulus lowered by ~18% and ~13% respectively, in K3KO1 & K3KO2 cells re-expressed with hK3int compared to those re-expressed with hK3 (Fig. 4.7b). The loss in membrane tension of hK3int cells was ~15% and ~23%, comparable to the effect of total loss of K3 in K3KO cells (Fig. 4.8b). This was mirrored by the measurements of the tether forces and tether radius (Fig. 4.8c). However, re-expression of hK3int recovered  to only 48 pN/µm. This shows that the K3-integrin binding contributes to ~ 55% of K3-mediated MCA at the membrane interface (parameter values shown in Table 3).

117

a b 40 40 Control Control K3KO1 35 K3KO2 35 K3KO1hK3 K3KO2hK3 K3KO1hK3int 30 K3KO2hK3int 30 25 25

20 20 Count Count 15 15 10 10 5 5 0 0 0 200 400 600 800 1000 1200 1400 1600 1800 0 200 400 600 800 1000 1200 1400 1600 1800 Young's Modulus, Pa Young's Modulus, Pa

c d

40 40 Control Control 35 K3KO1 35 K3KO2 K3KO1hK3 K3KO2hK3 30 K3KO1hK3int 30 K3KO2hK3int 25 25

20 20 Count Count 15 15 10 10 5 5 0 0 20 40 60 80 100 120 0 20 40 60 80 100 120 Young'sTether Modulus, Force, pN Pa TetherControl Force, pN

Fig. 4.9. Histograms of Young’s modulus (a, b) and tether forces (c, d) in macrophage cells.

118

Table 3. Average values of membrane parameters of RAW 264.7 measured by AFM. Mean ± St. error of all the parameters measured by AFM for RAW calculated as described in the experimental procedures.

Genotype EY, Pa FT, pN TM, pN/μm RT, nm

Macrophages Control 813.8 ± 26.9 61.7 ± 1.9 518.4 ± 31.1 11.7 ± 0.4

(RAW 264.7 K3KO1 235.0 ± 6.1 39.7 ± 1.5 215.8 ± 18.3 17.0 ± 0.6

cells) K3KO2 245.2 ± 6.2 40.8 ± 1.4 220.0 ± 16.1 16.0 ± 0.5

K3KO1hK3 615.6 ± 33.5 48.6 ± 1.3 317.1 ± 19.7 13.6 ± 0.3

K3KO1hK3pxn 380.8 ± 19.6 40.0 ± 1.2 209.0 ± 10.1 16.3 ± 0.5

K3KO1hK3int 504.8 ± 22.7 44.8 ± 1.6 269.1 ± 18.2 14.9 ± 0.5

K3KO2hK3 576.9 ± 3.6 50.8 ± 1.3 334.6 ± 17.7 12.7 ± 0.3

K3KO2hK3pxn 402.4 ± 10.1 41.8 ± 1.5 234.8 ± 20.8 15.7 ± 0.4

K3KO2hK3int 504 ± 17.8 43.97 ± 1.5 258.5 ± 17.9 15.1 ± 0.5

4.4. Discussion

The changes in intrinsic tissue stiffness is an early indicator of pathophysiological changes within the CNS. Cells constantly sense the mechanical forces exerted by their environment and accordingly adjust their gene and protein expression as well as cellular functions such as proliferation, migration, differentiation and apoptosis, vital for embryogenesis and morphogenesis. Cells do this by converting the mechanical stimuli from the extracellular domain to the nucleus via a combination of biochemical signals, a process commonly termed mechanotransduction[2], [3]. The extent of mechanical stimuli dictates the plethora of changes cells can undergo within, including rearrangement of their

119 glycocalyx, synthesis and crosslinking of cytoskeletal proteins, activation and recruitment of stress-sensitive molecules, transduction of signal from cell surface to nuclei via multiple pathways, force-induced protein aggregates, and biomolecular shuttling intracellularly to zones of high demand.

Mechanotransduction typically involves integrins and various receptors on cell surface to initiate the focal adhesion (FA) site-mediated extracellular matrix-integrin- cytoskeleton-nucleus nexus. Although the role of kindlins in many physiological processes has been well-established[24], [46], the specific roles of K3 in membrane tension has not been measured quantitatively. Using AFM, we here show mechanistically that the mechanosensory cell functioning in hematopoietic cells depends on K3, which coordinates membrane mechanics and MCA. MCA proteins simultaneously bind to cytoskeleton and to membrane phospholipids, PIP2, a hub for most MCA proteins including ERM (ezrin, radixin, moesin)[209]. K3 serves as a prime candidate for this function, as it contains three

FERM domains that are similar to ERM proteins, a PH domain anchoring K3 to PIP2, and interacts with numerous cytoskeletal proteins and adaptors[47], [254]. Besides the docking sites characteristic for MCA proteins, K3 binds to integrin cytoplasmic domain, an additional anchoring point enabling integration of ECM signaling mediated by integrins.

Indeed, our results show that K3 deficiency results in a loss of ~65% of the apparent membrane tension. Importantly, this cannot be explained by the role of K3 in integrin activation preceding cell spreading since the measurements were performed on non-spread cells with equal footprint. Interestingly, K3 deficiency diminished membrane tether force even greater than actin polymerization inhibitor, latrunculin A[255]. This defect was rescued by K3 re-expression with human K3 emphasizing the importance of K3 and

120 cytoskeleton in MCA. The K3-mutant (K3mut), which does not interact with integrins, hinders MCA resulting in ~50% loss of membrane tension. This implies that the K3- integrin interactions serve as the major anchor for membrane binding while the PH domain of K3 only has a minimal role (~15%). Finally, we show that K3 interaction with the cytoskeleton protein paxillin accounts for almost all the K3-mediated membrane tension.

Microglial morphology changes with the location, differentiation, and activation and is reciprocally related to its function[89], [239]. For instance, microglia acquire bipolar rod shape as a result of injury or pathology indicating its highly proliferative and low pro- inflammatory status[13],[90], [213]. The pathways controlling microglial shape expectedly affect a variety of their physiological and pathological functions. Studies exploring the crosstalk between chemical and mechanical signaling support the notion that microglial mechanosensing is just as critical as biochemical signaling in the CNS[16], [241]. Cells with mechanosensing function, including microglia, respond to stiffness gradients by migrating towards the stiffer side of the substrate, in a process termed durotaxis[29]. Stiff implants trigger significantly enhanced levels of glial cell activation and subsequent inflammation in vivo compared to softer implants. Neural implants such as electrodes are used to treat patients suffering from Parkinson’s disease and clinical depression. Microglial cells are activated by these implants and migrate toward this stiffer implant, which significantly contributes to the progression of foreign body reactions. This intimate crosstalk between chemical and mechanical signaling supports the view that mechanosensing is just as critical as biochemical signaling microglial response in the CNS.

However, the mechanism of mechanosensing and force transmission during microglial response was not well understood.

121

We show that the loss of apparent membrane tension resulting from K3 deficiency is profound and comparable to the effect of actin inhibitors or defects in ERM proteins known to mediate membrane to cytoskeleton attachment[255],[157]. Together, with our results, we can deduce that the spatiotemporal arrangements between its members and the consequences of their individual interactions are critical for MCA and membrane tension, not the composition of the integrin adhesome.

Membrane tension is widely measured from the tether forces obtained using optical tweezers or AFM. The tether forces of wild-type microglia and control macrophages in our study are in close ranges to that reported for primary macrophages and microglia isolated from neonatal mice brains measured using optical tweezers, and those extracted from a variety of cell types[256]. Our results also show that TM was 2- to 3- fold greater in control macrophages, when cell membrane was connected to the cytoskeleton via K3. Similar reductions in TM were evident when membrane-cytoskeleton attachment proteins (e.g., ezrin, class I myosins), cytoskeleton, or membrane composition are perturbed[256].

In a recent study, it was found that ezrin-mediated connection between actomyosin cortex and plasma membrane is required to maintain the normal membrane tension in epithelial cells[157]. AFM based tether pulling analysis showed that ezrin silencing cause a 28% decrease in overall membrane tension in epithelial cells[157]. Looking at a global level of changes in membrane tension, K3 knockout in microglia and macrophage causes a more dramatic decrease in membrane tension as compared to ezrin silencing in epithelial cells and inhibition of actin polymerization in Chinese hamster ovary cells. K3 deficient cells are characterized by highly contractile core, which is disconnected from the “loose” membrane with a number of characteristic ruffles and folds. This defect, however, is

122 rescued not only by K3 re-expression, but also by forced activation of β1 integrin signaling by LIBS antibody or by myosin inhibition, emphasizing an importance of the link between

β1 integrin, K3 and cytoskeleton. It is important to note that integrin activation or inside- out signaling, previously believed to be the main and sole function of K3 was not sufficient to achieve a rescue. Microglia-specific deletion of β1 but not αm or β2 integrins mirrors hyper-vascularization caused by the lack of K3, thereby emphasizing the key role for β1-

K3 interaction in microglia.

4.5. Conclusion

We found that CNS-derived microglial cells regulate the membrane tension through

Kindlin-3 mediated connections between the actomyosin cortex and the inner leaflet of the plasma membrane. Cytoskeleton remodeling upon K3 knockout renders the knockout cells less stiff compared with control cells. Interestingly, the cells largely restore the initial mechanical properties after reintroducing of K3 in macrophages emphasizing the role of tension homeostasis in these cells.

123

CHAPTER V

SUBSTRATE STIFFNESS INDUCED MECHANOTRANSDUCTION

REGULATES NPC PHENOTYPE, DIFFERENTIATION, AND MECHANICS

5.1. Introduction

Cells can monitor their microenvironment via mechanosensing which keeps them informed of their surroundings (i.e., niche) to maintain homeostasis[30],[3].

Mechanotransduction empowers cells to detect, decipher, and retort the physical properties of niche by converting them into biochemical signals[81]. Such interplay between biochemical and biomechanical signals is more pronounced in stem and progenitor cells during various phases of development, starting with embryogenesis up to later stage of life.

For instance, NPCs migrate from their niche in various CNS regions to proliferate and differentiate into neuronal and glial subtypes, which experience and architect mechanically heterogenous tissue microenvironment[93]. NPCs respond to substrate stiffness by biasing towards neuronal differentiation on softer substrates and astrocyte lineage on relatively stiffer substrates[63]. Stem cells based therapies for drug delivery, tissue engineering, cell replacement, immunotherapy, and regenerative medicine are being developed to provide

124 desirable cues in a spatiotemporal manner, to influence the gene expression of interacting cells and promote the regeneration by mimicking biological and mechanical properties of native tissues[67], [80], [84]. Studies exploring the effects of mechanotransduction on

NPCs and differentiated progenies provide important insights in designing such strategies

(e.g., delivery vehicles).

Central nervous system (CNS) formation and maturation is a highly mechanically coordinated process[69]. Stiffness of the brain and microenvironment of residing cells increases with aging from neonate to adult to aged tissue, which coincides with the loss in regenerative ability of the adult stem cells and progenitor cells inhabiting CNS[257], establishing a strong link between cell functionality and matrix composition and rigidity.

However, in the case of neurodegenerative disorders, stiffness of brain decreases owing to the loss of adult neurogenesis and changes in the ECM[100]. Although changes in ECM during injury or disorders lead to abnormal mechanotransduction, the extent of contribution from such changes in mechanical characteristics is disease pathology and progression is still relatively unknown[64]. Translocation of transcriptional factors such as Yes- associated protein (YAP)[15], [56], [57] and transcriptional coactivator with PDZ-binding motif (TAZ)[55], [58], [60], expression of PIEZO1[87], [257] and other ion channels[258], and RhoA[56], [62] activation during development and disorders, have been shown to play important roles in converting ECM niche signals into functional outcomes. Elucidation of molecular mechanisms underlying mechanobiology can provide greater insights in the progression of changes at the cellular level during normal CNS development vs. pathological-aberrant conditions, opening new doors for mechanotherapy.

125

Cellular mechanics is an important parameter contributing to fate decisions by orchestrating biophysical changes-based development of tissues[86], [95]. Biophysical changes due to dynamic evolution of cytoskeleton has been shown as a hallmark of stem cell differentiation[183]. Studying the changes in deformability of cells during and after differentiation is of great interest in the field of mechanotransduction[210], [259]. Analyses of cytoskeleton rearrangement, deformability, and stress fiber-formation, along with genetic analysis might contribute to the translational potential of stem cell-based therapies.

In this work, we investigated the mechanosensing ability of human NPCs (hNPCs) on biological scaffolds prepared at different protein concentrations. We also investigated the effect of mechanosensing on hNPCs mechanics and fate. In particular, we report for the first time on the temporal evolution of mechanical properties of hNPCs over a 9-day period, modulated by the underlying matrix niche, to elucidate mechano-adaptation as a possible mechanism and indicator of differentiation (Fig. 1). We investigated the signaling pathways such as YAP and myosin motors responsible for mechanotransduction, leading to the changes in hNPCs intrinsic mechanical properties and lineage commitment. We address the changes in biophysical, functional, and phenotypical characteristics of hNPCs as a result of mechanotransduction, exploring niche-activated sensors.

5.2. Methods

5.2.1. Substrate preparation

GeltrexTM LDEV-free reduced growth factor basement membrane matrix was obtained from ThermoFisher Scientific (catalog number A1413302). As-received matrix was used to make 100% GeltrexTM solution (referred hither to as G-100), while the as-

126 received matrix was diluted in ReNcell maintenance medium (SCM005; EMD Millipore) to prepare 25% GeltrexTM solution (referred as G-25). The calculated volume of solutions

(for 70 µm thick gel) was pipetted onto a 24-well plate and AFM specific petri dishes, maintained on ice. To uniformly spread the small volumes (μL) in a well-plate or petri dish to obtain required height of the gel, we used cover glass coated with Sigmacote (Sigma-

Aldrich) as explained earlier[260]. Briefly, coverslips treated with Sigmacote were placed on the solutions which were allowed to gel for 30 min at 37 C, and the coverslips were removed to obtain a ~ 70 µm thick gel. Sigmacote prevents attachment of cover glass with

Geltrex and allows removal of cover glass without disturbing polymerized gels[260]. TCPS was functionalized with laminin for 2 hours before use.

5.2.2. Cell culture

Human NPCs (ReNcell VM; SCC008; EMD Millipore, Burlington, MA, USA) cultured on laminin-coated flasks were maintained in undifferentiated state by using complete medium (ReNcell maintenance medium; SCM005; EMD Millipore) containing

20 ng/mL each of bFGF and EGF. In select cultures, hNPC differentiation on TCPS was induced by culturing in medium without bFGF and EGF. Blebbistatin (Catalog # 72402,

STEMCELLTM) was added to the culture medium (5 µM), in select cases, for non- muscle myosin II inhibition.

5.2.3 Experimental design

To elucidate the mechanosensing ability of hNPCs, we first primed the cells on

TCPS for 3 days in the presence of growth factors to keep them in undifferentiated state, which was confirmed by the expression of SOX2. After these 3 days, mechanical properties

127 of hNPCs were evaluated using AFM. These hNPCs were detached and reseeded on different substrates, and cultured for further studies as detailed below:

1. hNPCs were cultured on TCPS for 3 days or 9 days in the presence of growth factors.

At the end of 3 and 9 days, mechanical properties and expression of stemness marker

were characterized (Fig 5.1A).

2. hNPCs were seeded on TCPS in the absence of growth factors. At the end of 9 days,

mechanical properties and differentiation markers expressed by the cells were

evaluated (Fig 5.1B). This experiment serves as a control to compare the mechanical

properties of prospective differentiated cells.

3. hNPCs were cultured on G-100 substrates for 3 or 9 days and evaluated for changes in

phenotype using differentiation markers. Also, at the end of these 3 and 9 days,

mechanical properties of the cells were evaluated (Fig 5.1C).

4. hNPCs were cultured on G-25 substrates for 3 or 9 days. At the end of these 3 or 9

days, mechanical properties and differentiation of cells was quantified (Fig 5.1D).

5. Cells were cultured on G-100 (Fig 5.1E), G-25 (Fig 5.1F) and TCPS (Fig 5.1G)

substrates for 9 days in the presence of blebbistatin to study the role of myosin II. At

the end of 9 days, mechanical properties and differentiation of cells was quantified.

128

Fig. 5.1. Graphical summary of the experimental design implemented in this study to characterize mechanosensing effects on hNPCs fate and mechanics.

5.2.4. Atomic force microscopy

Young’s moduli of the gels (G-100 and G-25) were obtained using tip-less AFM cantilevers (Arrow™ TL1, Nanoworld, nominal spring constant 0.03 N/m) modified by attaching a 35-μm polystyrene bead using epoxy as explained earlier (details in section

2.2). Geltrex solutions for indentation assays were cold-pipetted on to custom-made PDMS wells (n=3 wells/concentration) and allowed to gel at 37°C in incubator. Force-indentation curves were obtained at random location on the gels and were analyzed by fitting Hertz model (details in 2.3.2) for a spherical indenter using Poisson’s ratio µ  0.49[172]

(typically used for similar scaffolds).

129

hNPCs at a seeding density of 0.5  105 cells/well were cultured on three different substrates in AFM-specific petri dishes: 2D laminin-coated surface, G-100 or G-25 gel surfaces for 3 and 9 days in presence and absence of growth factors (details in 5.2.3., Fig

5.1). The cells were maintained at 37 °C throughout the live-cell indentation assay. For each experiment, 30–50 cells were randomly selected, and force curves were obtained between nuclei and cell margin at approach/retraction velocity of 5 µm/s using TR 400

PSA cantilever (nominal spring constant ~0.05 N/m). Using Hertz’s contact model, the

Young’s modulus was determined from the force–indentation curves at indentation depth

(~ 400 nm). The tether forces (FT) were directly calculated from the series of force steps in the retraction curve. The apparent membrane tension (TM) was calculated from tether

2 2 forces using 푇푀 ≅ 퐹푇 ⁄8휋 퐾퐵, where KB is the bending stiffness which lies in the range of

0.1–0.3 pN.µm[203], [244]. Similarly, RT (tether radius) was calculated from the tether forces as 푅푇 ≅ 2휋퐾퐵⁄퐹푇.

5.2.5. Immunofluorescence labeling

On days 3 and 9, the following cultures were processed to identify and quantify the distinct neural and glial lineages: ReNcell in absence of growth factors on laminin-coated

TCPS, ReNcell on G-100, and ReNcell on G-25. Cells were washed once in sterile 1 PBS, fixed with 4% PFA for 20 min, washed twice with 1PBS (5 min), and incubated with blocking buffer (0.5% Triton-X, 5% serum, 1 PBS) for 2 h at room temperature. Serum selection was based on primary and secondary antibody host species. After removing the blocking buffer, cells were incubated with respective primary antibodies (4 °C, 24 h): mouse polyclonal anti-GFAP (Abcam), mouse monoclonal anti-SOX2 (Thermo Fisher

Scientific), mouse monoclonal anti-3 tubulin (Thermo Fisher Scientific), mouse

130 monoclonal YAP antibody, and actin-staining Alexa Flour™ 488 Phalloidin (Thermo

Fisher Scientific). Cells were washed four times in PBS for 10 min each, incubated with appropriate secondary antibodies (Santa Cruz Biotechnology, Dallas, TX) at room temperature for 4 h, washed again three times with PBS for 10 min each, and then counterstained with 4,6-diamidino-2-phenylindole (DAPI, Sigma).

Cells were imaged using the Zeiss AxioVert A1 fluorescence microscope under both phase contrast and fluorescence channels using a digital camera (Axiocam C1, Carl

Zeiss) and Axiovision data acquisition software. Total number of cells per well was quantified using ImageJ by manually counting all the cells in the well emitting DAPI signal and comparing that to the total number of cells in that same well positively stained for the aforementioned antibody markers (> 200 cells counted per condition).

5.2.6. Statistical Analysis

Data were represented as mean ± standard error from at least n=4 wells/condition, with at least three independent repeats of each assay, and statistical analysis was performed using GraphPad Prism 5. Data analysis was performed using one-way ANOVA, followed by Tukey multiple comparison to find statistically significant differences between the groups. p < 0.05 deemed statistically significant.

131

5.3 Results

5.3.1. Characterization of mechanical properties of GeltrexTM

GeltrexTM based scaffolds are being used in mechanobiology studies exploring nervous system development[261],[262]. Young’s modulus was calculated from force- indentation curves by applying Hertz model (Fig. 5.2A). AFM measurements indicate that the average modulus of as-received GeltrexTM (G-100) was 892 ± 148 Pa (n = 30 locations/concentration; stiffness  7 mN/m). Dilution of GeltrexTM to 25% concentration

(G-25) drastically decreased the average Young’s modulus to 78 ± 23 Pa (Fig. 5.2B; stiffness  0.7 mN/m; p < 0.001 vs. G-100 for both modulus and stiffness).

Fig. 5.2. Polystyrene microsphere (35 µm diameter) mounted on tip-less cantilever was used as the mechanical probe. Representative force-indentation curves from AFM analysis of GeltrexTM gels (A). Moduli values were obtained from the force-indentation curves generated from at least 30 random locations on each gel specimen (B). * denotes p < 0.001.

5.3.2. Effect of substrate (TCPS) priming on hNPC mechanics and phenotype

hNPCs were initially cultured and primed on TCPS for three days (details in 5.2.3.,

Fig 5.1). After 3 days of priming, hNPCs maintained their phenotype (Fig 5.3A) and

132 expressed stemness marker SOX2 (Fig 5.3B). As expected, we did not observe expression of differentiation markers such as β-III tubulin and GFAP (data not shown here). Young’s modulus (Fig 5.4A), tether force (Fig 5.3B), apparent membrane tension (Fig 5.4C) and tether radius (Fig 5.4D) of hNPCs quantified were 3.48 ± 0.06 kPa, 91 ± 1.12 pN, 1110 ±

28 pN/µm, and 6.1 ± 0.38 nm, respectively. These values establish the baseline mechanical properties and phenotypical characteristics of hNPCs cultured on TCPS for 3 days in presence of growth factors.

Fig. 5.3. Phase contrast images of cells in presence of growth factors after 3 days (A). Representative immunofluorescence images of cells in the presence of bFGF and EGF, after 3 days in culture (B). Cultures were counterstained with DAPI for cell identification. Primary antibody for SOX2 was used, with appropriate secondary antibody. Scale bar: 50 μm.

133

Fig. 5.4. Mechanical characterization results of primed cells. Young’s modulus (A), tether force (B), tether radius (C) and apparent membrane tension (D) of cells in the presence of bFGF and EGF after 3 days in culture.

5.3.3. Characteristics of hNPCs after 3-day culture on TCPS and 3D substrates

To establish the standalone effects of mechanosensing on hNPCs, we cultured

NPCs on gels and TCPS for 3 days in the presence of growth factors (details in 5.2.3., Fig

5.1 – A, C, D). hNPCs cultured on TCPS showed homogenous morphology (Fig. 5.5A) and expression of SOX2 marker (Fig. 5.5B), similar to that observed during their priming.

However, hNPCs cultured on G-100 and G-25 gels evolved into a heterogenous mixture of cells binned into three distinct morphologies, denoted hitherto as NSC-like, star-shaped, and elongated cells (Fig. 5.6), and expressed staining markers for mixed lineages (Fig.

5.7).

Fig. 5.5. Representative phase contrast (A) and immunofluorescence (B) images of cells cultured on TCPS for 3 days in the presence of growth factors (bFGF and EGF). Cultures were counterstained with DAPI for cell nuclei identification. Primary antibody for SOX2 was used with an appropriate secondary antibody. Scale bar: 50 μm.

134

Fig. 5.6. Representative phase contrast images of NPC-like, star-shaped, and elongated cells on G-25 (A) and G-100 (B) after 3 days. Scale bar: 50 μm.

Fig. 5.7. Representative immunofluorescence images of hNPCs cultured on G-25 (A) and G-100 (B), after 3 days in culture, and stained for 3-tubulin (neuronal) and SOX2 (stemness) markers and counterstained with DAPI (nuclei). Scale bar: 50 μm.

135

A B * * * * * * * * 4 * 100 * * * * * 3 * 80 * * * 60 2 *

* , pN * , kPa

* T * Y 40 * F E 1 20 0 0

NPC NPC-like Elongated Star-Shaped NPC NPC-like Elongated Star-Shaped

C D * * * * * * * 30 * * 1500 * * * * * 20 * 1000 * * *

, nm *

* pN/µm T

* M, R 10 500 *

T * * 0 0

Elongated Star-Shaped Elongated Star-Shaped NPC NPC-like NPC NPC-like

Fig. 5.8. Mechanical characterization results: Young’s modulus (A), tether force (B), tether radius (C), and apparent membrane tension (D) of cells on TCPS, G-100, and G-25, after 3 days in culture, in the presence of growth factors. * denotes p < 0.01.

We quantified mechanical properties of hNPCs after 3 days of culturing on TCPS,

G-100, and G-25 (Fig. 5.8, Table 1). We observed that hNPCs on TCPS had almost similar values of EY, FT, TM, and RT as those primed on TCPS (Fig. 5.4), but had significantly (p

< 0.001) higher Young’s modulus, tether force, and apparent membrane tension than the heterogenous population of cells on G-100 and G-25. Tether radius follows opposite trends as expected, expectedly. The average EY of NPC-like, elongated, and star-shaped cells on

G-100 was ~ 1.4 to 1.9 -fold higher (p < 0.01 in all cases) compared to their counterparts

136 on G-25. Similarly, the FT of NPC-like, elongated, and star-shaped cells on G-100 were ~

1.2 to 1.35 -fold higher (p < 0.05 in all cases) than their counterparts on G-25. Likewise, significant differences (p < 0.01) in FT and TM of cells cultured on G-100 and G-25 were noted. Finally, significant differences (p < 0.001) in EY, FT, TM and RT were noted among the various cell types (i.e., NPC-like, elongated and star-shaped) within each substrate conditions (i.e., G100 or G25).

5.3.4. Characteristics of hNPCs after 9-day culture on G-25 and G-100

To study the evolution of mechanics and maturation of differentiation markers, hNPCs were cultured on G-100 and G-25 up to 9 days in the presence of growth factors

(details in 5.2.3., Fig 5.1-A, C, D). In contrast to the three distinct morphologies observed after 3 days of culture on G25 and G100 substrates (Fig. 5.6), we observed only two distinct morphologies after 9 days of culture – star-shaped and elongated cells (Fig. 5.9). Higher percentage of cells expressed β-III tubulin compared to GFAP on G-25, while the converse was true on G-100 gels (Fig. 5.10). SOX2 expression was absent in both these cultures at day 9 indicating the lack of stemness phenotype in these cells and their maturation into astrocytic and neuronal lineages.

137

Fig. 5.9. Phase contrast images of star-shaped and elongated cells on G-25 and G-100 after 9 days. Scale bar: 50 μm.

Fig. 5.10. Representative immunofluorescence images of cells on G-25 and G-100 after 9 days of culture (A). Cultures were counterstained with DAPI for cell nuclei identification. Primary antibodies for SOX2, β-III tubulin, and GFAP were used, with appropriate secondary antibodies. Scale bar: 50 μm. (B) The number of stained cells for each marker were counted manually from the images; n > 200 cells/case; * p < 0.01.

138

5.3.5. Role of growth factors on hNPCs cultured on TCPS for 9 days

We cultured hNPCs on laminin-coated TCPS in the presence and absence of growth factors (bFGF and EGF) for 9 days (details in 5.2.3., Fig 5.1 - A, B). With growth factors, hNPCs maintained their phenotype (Fig. 5.11A) and stemness (SOX2 expression, Fig.

5.12A), similar to their status after initial priming and subsequent 3-day culture. Removal of growth factors led to formation of elongated and star-shaped cells (Fig. 5.11 - B, C), positively staining for β-III tubulin & GFAP and negatively staining for SOX2 markers

(Fig. 5.12 - B, C), indicating their differentiation and lineage commitment.

Fig. 5.11. Representative phase contrast images of cells after 9 days of culture on TCPS in presence of growth factors (A). Elongated cells (B) and star-shaped cells (C) in absence of growth factors after 9 days on TCPS. Scale bar: 50 μm. Distinct morphologies of the cells in both the cases was evident.

139

Fig. 5.12. Representative immunofluorescence images of cells after 9 days of culture, in the presence (A) and absence (B) of bFGF and EGF. Cultures were counterstained with DAPI for cell identification. Primary antibodies for SOX2, β-III tubulin, and GFAP were used with appropriate secondary antibodies. Scale bar: 50 μm.

We quantified mechanical properties of cells after 9 days of culturing on TCPS, G-

100 and G-25 (Fig. 5.13, Table 1). The mechanical properties of NPCs (EY, FT, TM) in the presence of growth factor after 9 days (Fig. 5.13) were (i) similar to those of NPCs after 3 days under similar conditions (Fig. 5.8), and (ii) significantly higher (p < 0.01) than those of cells (elongated, star-shaped) cultured under all other conditions (TCPS, G100, G25) at the end of 9 days. The mechanical properties (EY, TM, FT) of cells on TCPS in the absence of growth factors (denoted by TCPS# in Fig. 5.13) were significantly higher than those on

G-25 and G-100. For elongated and star-shaped cells, the EY, FT and TM values are of the

# order TCPS > TCPS > G-100 > G-25. Finally, the EY, FT and TM values of elongated cells are higher than those of star-shaped cells on respective substrates (Table 1, Fig 5.13).

140

Fig. 5.13. Mechanical characterization: Young’s modulus (A), tether force (B), tether radius (C), and apparent membrane tension (D), of resulting cells after 9 days in culture on TCPS, G-100 and G-25. Cells were cultured on TCPS in the presence or absence of growth factors. * denotes p < 0.01, # denotes without growth factors.

We observed more prominent stress-fibers in cells cultured on G-100 (Fig. 5.14 -

B, D) than G-25 (Fig. 5.14 - A, C) on day 9, which might explain the differences in their mechanical properties. To explore one potential sensor dictating mechanosensing, we stained cells on both stiff (TCPS) and compliant (Geltrex) substrates for YAP. However, we observed no expression of YAP protein in these cells (Fig. 15).

141

Fig. 5.14. Representative immunofluorescence images of cells showing morphological changes, i.e., reorganization of F-actin mesh network on G-25 (A, C) and G-100 (B, D). Scale bar: 50 µm.

Fig. 5.15. Representative immunofluorescence images of cells stained for YAP antibody on TCPS (A) and Geltrex (G-100) (B). Cultures were counterstained with DAPI for cell identification. Scale bar: 50 μm. * p < 0.01.

142

5.3.5. Inhibiting non-muscle myosin-II in hNPC cultures regulate their phenotype and mechanical characteristics

Blocking non-muscle myosin-II with addition of blebbistatin relaxes cellular actin network[263]. NPCs were treated with blebbistatin throughout the 9-day cultures on TCPS,

G-100 and G-25 substrates, and assessed for changes in cell mechanics and differentiation

(details in 5.2.3., Fig 5.1 - E, F, G). The presence of blebbistatin caused polarization of cells (Fig. 5.16A). Young’s modulus and tether forces of blebbistatin-treated cells on TCPS are significantly higher (p < 0.01) than cells on G-100 and G-25. There was no statistically significant difference (p > 0.05) in the EY and FT values of blebbistatin-treated cells on G-

100 vs G-25 substrates (Table 1; Fig. 5.16 - B, C). The EY values of blebbistatin-treated cells were significantly lower than their non-treated counterparts on respective substrates, while the FT values were higher (Fig. 5.13, A-B). This indicated that myosin-II plays an important role in sensing mechanical cues which dictates cellular mechanics.

Fig. 5.16. Representative phase contrast images of blebbistatin-treated cells on TCPS, G- 100 and G-25 after 9 days in culture (A). Young’s modulus (B) and tether force (C) of cells on TCPS, G-100 and G-25 after 9 days in culture. Scale bar: 50 μm. *denotes p< 0.05

143

NPCs maintained their stemness even after blebbistatin treatment after 9 day

(5.17A). Compared to dense, network pattern formed by untreated cells on G-100 and G-

25 substrates (Figs 5.10, 5.14), blebbistatin treatment led to formation of polarized, independent cells (Fig 5.16A, 5.17A). No statistically-significant difference (p > 0.05) was noted between the β-III tubulin positive and GFAP positive cells on G-100 vs. G-25, after blebbistatin treatment for 9 days (Fig 5.17). Blebbistatin treatment led to reduced differentiation potential of hNPCs on Geltrex compared to their untreated counterparts

(Fig. 5.10), which indicates actomyosin-mediated mechanosensing as an important regulator of cell fate.

144

Fig. 5.17. Representative immunofluorescence images of blebbistatin-treated cells on TCPS, G-25 and G-100 after 9 days in culture (A). Percentages of stained cells were calculated by manually counting positively-stained cells in respective conditions and normalizing to total cell density (n > 200) (B). Cultures were counterstained with DAPI for cell identification. Primary antibodies for SOX2, β-III tubulin, and GFAP were used, with appropriate secondary antibodies. Scale bar: 50 μm. * indicates p < 0.01.

Table 4. Average values of mechanical characteristics of cells measured by AFM. Mean ± st. error of all the parameters measured by AFM for cells calculated as described in the experimental procedures. TCPS (tissue culture polystyrene), GF (growth factors), D3 (day 3), D9 (day 9), B (blebbistatin-treated), EY (Young’s modulus), FT (tether force), TM (apparent membrane tension), RT (radius of tether).

Culture Phenotype EY, kPa FT, pN TM, pN/μm RT, nm

D3 TCPS (+GF) NPCs 3.48 ± 0.06 91 ± 1.12 1110 ± 28 6.1 ± 0.38

D9 TCPS (+GF) NPCs 3.31 ± 0.08 95 ± 1.86 1208.4 ± 12 5.88 ± 0.43

D9 TCPS (-GF) Elongated 2.67 ± 0.04 69.3 ± 1.23 630.1 ± 35.2 9.4 ± 0.55

D9 TCPS (-GF) Star-shaped 1.81 ± 0.06 54.2 ± 1.4 410.1 ± 30.3 12 ± 0.4

D3 G-100 NPC-like 2.53 ± 0.05 61.2 ± 1.3 582.2 ± 24.7 7.89 ± 0.5

D3 G-100 Elongated 1.21 ± 0.06 43.1 ± 1.2 234.0 ± 15 13.5 ± 0.55

D3 G-100 Star-shaped 0.85 ± 0.04 34.1 ± 1.5 139.9 ± 31.9 18.9 ± 0.45

D3 G-25 NPC-like 1.75 ± 0.06 51.3 ± 1.3 290.0 ± 30.1 11.2 ± 0.55

D3 G-25 Elongated 0.82 ± 0.07 34.8 ± 1.2 149.8 ± 11.8 17 ± 1.2

D3 G-25 Star-shaped 0.45 ± 0.05 24.9 ± 0.08 75 ± 18.9 23.5 ± 0.55

D9 G-100 Elongated 2.1 ± 0.06 50.7 ± 1.13 250 ± 15.1 15.5 ± 0.55

D9 G-100 Star-shaped 1.52 ± 0.04 41.3 ± 1.5 151 ± 31.2 21.1 ± 0.45

D9 G-25 Elongated 1.72 ± 0.07 42.1 ± 1.25 165.4 ± 12.2 19.1 ± 1.2

D9 G-25 Star-shaped 1.12 ± 0.05 35.1 ± 1.81 98.2 ± 19.8 26.3 ± 0.55

D9 TCPS (+B) Polarized 1.56 ± 0.02 69.9 ± 1.20 663.1 ± 11.8 9.2 ± 0.62

D9 G-100 (+B) Polarized 1.01 ± 0.22 61.3 ± 1.85 571.1 ± 22.5 8.01 ± 0.56

D9 G-25 (+B) Polarized 0.92 ± 0.03 55.2 ± 1.31 570.9 ± 20.7 8.05 ± 0.86

145

5.4. Discussion

Mechanotransduction is regarded as one of the most important factors driving and guiding the CNS development[63]. Any perturbation by intrinsic or extrinsic agents cause abnormal mechanotransduction leading to neurological disorders[30]. During development, primitive cells such as NPCs are spatiotemporally regulated by mechanical cues[160]. To understand the contribution of mechanotransduction in CNS development, recent focus has been geared towards elucidating the changes in biophysical and biomechanical properties during fate commitment. Here, we characterized and quantified the mechanical properties of human NPCs and differentiated progenies, primarily as a result of mechanosensing. In particular, we explored the utility of AFM to monitor the evolution of mechanical forces associated with substrate-mediated differentiation of hNPCs. We observed that hNPCs are stiffer than their differentiated progenies (glial cells and neurons), and differentiated neurons are stiffer than glial cells. In this regard, this is the first study to evaluate the evolution of mechanics during differentiation of hNPCs.

We found that process of differentiation (day 9) of NPCs to neurons and glial cells, has an intermediate state (day 3) of softer cells which matures into much stiffer progenies.

NPCs and their differentiated progenies on softer substrate (G-25) were more compliant than cells on stiffer substrates (G-100, TCPS). We observed that NPCs on softer matrices

(G-25) preferentially differentiates towards neural lineages compared to those on stiffer matrices (G-100). In support to our observations, murine NSCs have been reported to differentiate into neurons on softer substrates but towards glial cells on stiffer substrates[62], [63], and mesenchymal stem cells (MSCs) cultured on soft substrates mimicking brain stiffness committed to neurogenic phenotypes[264].

146

Translocation of transcriptional factors such as YAP and TAZ act as mechanosensors[15], [54], [57] by converting mechanical signals into functional fate.

NSCs were shown to have maximum sensitivity towards substrate stiffness in the first 24 h, which drive their lineage commitment[62], [265]. Overexpression and silencing of YAP in the first 24 h led to inhibition and enhancement of neurogenesis, respectively[265].

Decline in regenerative capacity of stem cells and progenitor cells with aging can be linked to altered mechanical signaling from the ECM. PIEZO1 (mechanoresponsive ion channel) was shown to mediate conversion of these mechanical signals from the ECM into differentiation of oligodendrocyte progenitor cells (OPCs)[257]. Functional activity of

OPCs was restored in vivo by inhibiting PIEZO1 which defies the mechanical input during aging.

RNA sequencing data available in public domain on GEO (Gene Expression

Omnibus) for ReNCell VM (series: GSE89623) pointed that ReNcell VM express low levels of YAP, TAZ, ROCK and PIEZO1 genes (compared to SOX2 expression)[262].

With such low expression of these widely explored sensors (YAP/TAZ[15], [54]–[56],

[58]–[60] & PIEZO1[87], [257]), it is interesting to explore the transducers driving niche- dependent changes in phenotype and functional fate of these specific hNPCs. It is difficult to decouple the effects of substrate mechanical properties and composition of biological scaffolds. Our results indicate that intrinsic mechanical properties play an important role during niche-mediated differentiation of hNPCs, establishing mechano-adaptation as a possible mechanism of differentiation. We thus explored the role of actomyosin contraction in this process by its inhibition using blebbistatin. Results suggest that myosin plays an important role in regulating mechano-adaptability which ultimately influences cell fate.

147

Comparing differentiation and cell mechanics outcomes before and after blebbistatin treatment established substrate mechanosensing as one of the main factors driving cell fate.

Intrinsic mechanical properties of cells have shown great potential as a label-free biomarker in many processes. Mechanical properties of stem cells have been explored to indicate their stemness. In one study, lower Young’s modulus was correlated with greater differentiation potential; pluripotent cells were found to be softer compared to multipotent and unipotent stem cells[266]. In a recent study, limbal stem cells were sorted from a heterogenous population of corneal cells using AFM-based mechanical characterization[267]. While these studies utilized Young’s modulus for mechanical sorting of cells, our study explored Young’s modulus, tether force, membrane tension, and tether radius to distinguish NPCs, glial cells, and neurons in a heterogenous cell population on substrates with varying stiffness.

In support of our observation that differentiated progenies are more compliant than progenitor cells, Urbanska et al. reported on the formation of softer differentiated NPCs from stiffer iPSCs[158], and a positive correlation between Young’s modulus and pluripotency[158]. Similarly, a 2-fold decrease in Young’s modulus was observed during

MSC differentiation to osteogenic lineage[268]. Others have reported that MSCs are stiffer than osteoblasts[269], [270], and glial cells are more compliant than neighboring neurons[271], strengthening our own observations that hNPCs are less compliant than differentiated progenies (neurons and glial cells). Glial cells, being more compliant, serve as mobile immune cells in the CNS and provide protection during trauma. In contrast to most prior studies and our observations, murine embryonic stem cells (ESCs) were reported

148 to be less stiff compared to differentiated progenies[272] possibly arising from mesodermal differentiation of ESCs.

During development, cells such as MSCs, NPCs, and ESCs experience heterogenous mechanical microenvironments which directly influence their biophysical characteristics, accompanied with genetic and epigenetic changes[30], [69]. We observed that cells on stiffer matrices show higher EY compared to cells on softer matrices. Similar results were reported with other cells cultured on matrices of varying stiffness elucidating the potential mechano-adaptability of development relevant cells[264],[62],[273].

Tension-based morphogenesis proposed by David Van Essen in 1997 suggested that tension along glial cell extension, axons, and neurites is sufficient to generate folding patterns of the cortex, compactness of neural circuitry, and many other structural features of the mammalian CNS[274]. In support of this theory, application of tensile forces using glass microneedles on neurites caused active elongation when tension is maintained above threshold, followed by active retraction upon tension withdrawal[275]. However, the intrinsic membrane tension of progenitor cells and their differentiated progenies forming mechanically-heterogenous CNS was not explored. We quantified for the first time the en route membrane tension of hNPCs to differentiated progenies. We observed that membrane tension of hNPCs decreased after initial 3 days of differentiation, followed by an increase upon maturation at day 9.

5.5. Conclusion

We investigated the effect of mechanosensing on hNPCs using substrates of varying stiffness. We highlighted the role of cell biomechanics as a label-free marker of

149

NPC differentiation. We identified the role of mechano-adaptation as a possible mechanism of differentiation, specifically for these cells. We explored potential mechanosensors and elucidated the role of myosin motors in sensing and responding to the stiffness of their environment. Taken together, our findings establish the en route mechanical properties of hNPCs to differentiated progenies, trigged by substrate stiffness exploring niche-activated mechanosensors. Our future studies are geared towards correlating the mechanical properties of these cells with migration ability during durotaxis.

150

CHAPTER VI

CONCLUSIONS AND FUTURE DIRECTIONS

6.1. Conclusions

Current understanding of nervous system development and neurological disorders has been drive from the molecular biology perspective. However, studying the interplay of biochemical and biomechanical processes will help in developing a comprehensive understanding of the process of CNS development and disease progression. Elucidating the contribution of mechanical forces is still in its infancy, but with the advances in material science and biophysics, current studies are geared towards exploring neuromechanics. As detailed in the dissertation, mechanical properties of cells provide important insights on mechano-adaptability in different environments during development. Furthermore, here we demonstrate the potential of cell mechanics as indicators of developmental neurotoxicity and neurological disorders. The approach and results presented in this work can be implemented in the field of disease biophysics research and regenerative medicine.

This dissertation details the phenotypical characteristics of brain-derived cells by focusing on changes in mechanical properties due to external and internal factors. In aim

1, using an AFM, we established the role of both intrinsic and extrinsic mechanical forces dictating various events during CNS development. For example, sensitivity of CNS cells

151 to toxicants was demonstrated using biochemical and biomechanical endpoints of human brain-derived NPCs. The key correlation of mechanical properties of surviving cells and percentage of living cells in the respective cultures provided important insights which were not addressed in conventional cytotoxicity literature, and therefore merits inclusion as a crucial endpoint of toxicology assessments.

Molecular clutch is known to play a crucial role in the normal functioning of microglia during CNS development, and alterations to the proteins influencing microglial responses have been identified as determining factors of various CNS pathologies such as multiple sclerosis, Alzheimer’s, and Parkinson’s diseases. In aim 2 of this study, I specifically established that K3 is a main contributor to membrane mechanics and K3-β1 integrin interactive domain is the main connection of MCA. In the final aim 3, we elucidated how human NPCs sense the mechanical cues provided by matrix, and how the substrate stiffness influences mechanical characteristics of these cells and their fate. Specifically, I demonstrated the role of myosin II and mechano-adaptability as possible factors driving substrate stiffness-mediated differentiation of NPCs.

The outcomes from these studies address some of the long-term knowledge gaps in the field of neuroscience on how intrinsic mechanical forces are influenced by internal and external cues responsible for phenotypical and genotypical changes during nervous system development. In summary, the results and experimental framework presented here could provide insights into cellular processes such as morphogenesis, wound healing, brain plasticity, and metastasis. Table 5 summarizes the key findings of each aims.

152

Table 5. Summary of results from individual aims

Specific Aim 1: Establish the utility of biophysical and biomechanical characteristics as indicators of development neurotoxicity

Key Results:

• Utilized AFM to quantify biophysical and biomechanical characteristics of human NPCs in the presence of a variety of toxicant types • Evaluated the subcellular mechanisms of action for different classes of toxicant compounds, and identified the most dominant and sensitive mechanism • Identified a strong, negative, linear correlation between the elastic modulus of surviving cells and number of living cells in that environment • Using multivariable logistic regression, established the change in tether forces as the strongest predictor of neurotoxicity • Established the utility of cell mechanics as a crucial marker of developmental neurotoxicity (mechanotoxicology).

Aim 2: Elucidate the critical role of Kindlin-3 on microglial membrane mechanics and physical characteristics

Key Results:

• Utilized AFM to investigate membrane mechanics for genetic knockout or mutated cells • Established K3 as a major controller of membrane mechanics in microglia • Identified K3-β1integrin binding in the molecular clutch domain as the major contributor of MCA • Demonstrated the critical role of K3 mediated membrane mechanics in macrophage cell line • Validated the biomechanical functions of K3 by re-expressing human K3 in knockout cells.

Aim 3: Investigate the substrate-dependent mechanotransduction of brain-derived NPCs and identify the underlying mechanisms.

Key Results:

• Identified the mechanosensing ability of NPCs using biologically-relevant substrate stiffness and composition • Utilized AFM to quantify the mechanical properties of hydrogels and cells seeded on soft hydrogels • Established the use of mechanical properties as label-free marker of NPC differentiation

153

• Investigated the en route mechanical properties of NPCs into differentiated progenies • Elucidated the changes in mechanical properties and differentiation patterns of NPCs and differentiated progenies stimulated by substrate stiffness • Identified myosin II as major regulator of mechanosensing ability of NPCs

6.2. Future Directions

1. Although the biochemical and biomechanical effects of different classes of

toxicants were studied, the genetic and epigenetic changes in NPCs exposed to

toxicants should be investigated in the future.

2. The results obtained in this study establishing mechanical properties as indicator of

neurotoxicity should be expanded to study toxicity related to other tissues such as

liver.

3. The effects of K3 knockout and mutation (K3-β1 binding) on membrane mechanics

in microglia and macrophage identified in this study should be expanded to include

mutations at paxillin and talin binding sites.

4. Future studies should explore developing high-throughput technologies for cell

sorting based on mechanical properties.

5. En route mechanical properties of NPCs quantified in this study could be expanded

to investigate the mechanical properties during reprogramming of cells to other

phenotypes and genotypes.

6. Effect of mechanosensing on NPCs mechanics and fate established in this study

could be expanded to study the influence of mechanosensing on NPCs migration

using substrate gradient.

154

7. Due to limited expression of mechanosensors by ReNcell VM, future studies could

focus on exploring mechanosensors using primary NPCs.

8. The mechanosensing results obtained in this study could be expanded to elucidate

the metastasis ability of cancer cells by exploring substrates of varying stiffness.

155

REFERENCES

[1] C. T. Lim, A. Bershadsky, and P. Sheetz, “Mechanobiology,” no. April, pp. 291–

293, 2010.

[2] Z. Sun, S. S. Guo, and R. Fässler, “Integrin-mediated mechanotransduction,” vol.

215, no. 4, 2016.

[3] K. S. Kolahi and M. R. K. Mofrad, “Mechanotransduction: A major regulator of

homeostasis and development,” Wiley Interdiscip. Rev. Syst. Biol. Med., vol. 2, no.

6, pp. 625–639, 2010.

[4] P. Lu, K. Takai, V. M. Weaver, and Z. Werb, “Extracellular matrix degradation

and remodeling in development and disease,” Cold Spring Harb. Perspect. Biol.,

vol. 3, no. 12, p. 10.1101/cshperspect.a005058 a005058, Dec. 2011.

[5] Michele A. Wozniak and Christopher S. Chen, “Mechanotransduction in

development: a growing role for contractility,” Nat. Rev. Mol. Cell Biol., vol. 10,

no. 1, pp. 34–43, 2009.

[6] E. Paluch and C. P. Heisenberg, “Biology and Physics of Cell Shape Changes in

Development,” Curr. Biol., vol. 19, no. 17, pp. R790–R799, 2009.

[7] T. V. Malinin, N.L., Pluskota, E., and Byzova, “Integrin signaling in vascular

function,” Curr. Opin. Hematol., vol. 19, pp. 206–211, 2012.

[8] A. Pietuch, B. R. Brückner, T. Fine, I. Mey, and A. Janshoff, “Elastic properties of

cells in the context of confluent cell monolayers: Impact of tension and surface

area regulation,” Soft Matter, vol. 9, no. 48, pp. 11490–11502, 2013.

156

[9] E. Fischer-Friedrich, A. A. Hyman, F. Jülicher, D. J. Müller, and J. Helenius,

“Quantification of surface tension and internal pressure generated by single mitotic

cells,” Sci. Rep., vol. 4, pp. 4–11, 2014.

[10] S. W. Moore, R. E. Keller, and M. A. R. Koehl, “The dorsal involuting marginal

zone stiffens anisotropically during its convergent extension in the gastrula of

Xenopus laevis,” Development, vol. 121, no. 10, pp. 3131–3140, 1995.

[11] L. Li, Y. He, M. Zhao, and J. Jiang, “Collective cell migration: Implications for

wound healing and cancer invasion,” Burn. trauma, vol. 1, no. 1, pp. 21–26, Jun.

2013.

[12] Y. T. Shiu, J. A. Weiss, J. B. Hoying, M. N. Iwamoto, I. S. Joung, and C. T.

Quam, “The role of mechanical stresses in angiogenesis,” Crit. Rev. Biomed. Eng.,

vol. 33, no. 5, pp. 431–510, 2005.

[13] M. C. Zalis, S. Johansson, F. Johansson, and U. E. Johansson, “Exploration of

physical and chemical cues on retinal cell fate,” Mol. Cell. Neurosci., vol. 75, pp.

122–132, 2016.

[14] S. A. Abdellatef, A. Ohi, T. Nabatame, and A. Taniguchi, “The effect of physical

and chemical cues on hepatocellular function and morphology,” Int. J. Mol. Sci.,

vol. 15, no. 3, pp. 4299–4317, Mar. 2014.

[15] T. Panciera, L. Azzolin, M. Cordenonsi, and S. Piccolo, “Mechanobiology of YAP

and TAZ in physiology and disease,” Nat. Rev. Mol. Cell Biol., vol. 18, no. 12, pp.

758–770, 2017.

157

[16] K. B. Bernick, “Cell biomechanics of the central nervous system.,” Diss. Abstr.

Int. Sect. B Sci. Eng., vol. 73, no. 2-B, p. 1074, 2012.

[17] Q. T. Nguyen, T. D. Jacobsen, and N. O. Chahine, “Effects of Inflammation on

Multiscale Biomechanical Properties of Cartilaginous Cells and Tissues,” ACS

Biomater. Sci. Eng., vol. 3, no. 11, pp. 2644–2656, 2017.

[18] G. Giordano and L. G. Costa, “Developmental Neurotoxicity: Some Old and New

Issues,” ISRN Toxicol., vol. 2012, pp. 1–12, 2012.

[19] S. Hoffmann, A. Kinsner-Ovaskainen, P. Prieto, I. Mangelsdorf, C. Bieler, and T.

Cole, “Acute oral toxicity: Variability, reliability, relevance and interspecies

comparison of rodent LD50 data from literature surveyed for the ACuteTox

project,” Regul. Toxicol. Pharmacol., vol. 58, no. 3, pp. 395–407, 2010.

[20] Y. X. Liu et al., “Single-Cell Mechanics Provides an Effective Means to Probe in

Vivo Interactions between Alveolar Macrophages and Silver Nanoparticles,” J.

Phys. Chem. B, vol. 119, no. 49, pp. 15118–15129, 2015.

[21] R. Krishnan, J. A. Park, C. Y. Seow, P. V. S. Lee, and A. G. Stewart, “Cellular

Biomechanics in Drug Screening and Evaluation: Mechanopharmacology,” Trends

Pharmacol. Sci., vol. 37, no. 2, pp. 87–100, 2016.

[22] G. Y. H. Lee and C. T. Lim, “Biomechanics approaches to studying human

diseases,” Trends Biotechnol., vol. 25, no. 3, pp. 111–118, 2007.

[23] T. C. Lim, “Single Cell Mechanics and its connections to Human Diseases,” Asia.

Pac. Biotech News, vol. 09, no. 14, pp. 674–675, 2005.

158

[24] E. Rognoni, R. Ruppert, and R. Fässler, “The kindlin family: Functions, signaling

properties and implications for human disease,” J. Cell Sci., vol. 129, no. 1, pp.

17–27, 2016.

[25] J. P. Campanale, T. Y. Sun, and D. J. Montell, “Development and dynamics of cell

polarity at a glance,” J. Cell Sci., vol. 130, no. 7, pp. 1201–1207, 2017.

[26] Q. S. Li, G. Y. H. Lee, C. N. Ong, and C. T. Lim, “AFM indentation study of

breast cancer cells,” Biochem. Biophys. Res. Commun., vol. 374, no. 4, pp. 609–

613, 2008.

[27] D. E. Ingber, “Mechanobiology and diseases of mechanotransduction,” Ann. Med.,

vol. 35, no. 8, pp. 564–577, 2003.

[28] K. Farrell et al., “Alterations in phenotype and gene expression of adult human

aneurysmal smooth muscle cells by exogenous nitric oxide,” Exp. Cell Res., vol.

384, no. 1, p. 111589, 2019.

[29] L. Bollmann et al., “Microglia mechanics: Immune activation alters traction forces

and durotaxis,” Front. Cell. Neurosci., vol. 9, no. September, pp. 1–16, 2015.

[30] J. M. Barnes, L. Przybyla, and V. M. Weaver, “Tissue mechanics regulate brain

development, homeostasis and disease,” J. Cell Sci., vol. 130, no. 1, pp. 71–82,

2017.

[31] K. Franze, “Atomic force microscopy and its contribution to understanding the

development of the nervous system,” Curr. Opin. Genet. Dev., vol. 21, no. 5, pp.

530–537, 2011.

159

[32] V. Lulevich, C. C. Zimmer, H. S. Hong, L. W. Jin, and G. Y. Liu, “Single-cell

mechanics provides a sensitive and quantitative means for probing amyloid-β

peptide and neuronal cell interactions,” Proc. Natl. Acad. Sci. U. S. A., vol. 107,

no. 31, pp. 13872–13877, 2010.

[33] N. Evans et al., “Substrate stiffness affects early differentiation events in

embryonic stem cells,” Eur. Cells Mater., vol. 18, pp. 1–14, 2009.

[34] K. S. Kolahi et al., “Effect of substrate stiffness on early mouse embryo

development,” PLoS One, vol. 7, no. 7, 2012.

[35] A. J. Engler, S. Sen, H. L. Sweeney, and D. E. Discher, “Matrix Elasticity Directs

Stem Cell Lineage Specification,” Cell, vol. 126, no. 4, pp. 677–689, 2006.

[36] H. Lv et al., “Mechanism of regulation of stem cell differentiation by matrix

stiffness,” Stem Cell Res. Ther., vol. 6, no. 1, p. 103, May 2015.

[37] N. R. Lang et al., “Biphasic response of cell invasion to matrix stiffness in three-

dimensional biopolymer networks,” Acta Biomater., vol. 13, pp. 61–67, 2015.

[38] Y. S. Liu and O. K. Lee, “In search of the pivot point of mechanotransduction:

Mechanosensing of stem cells,” Cell Transplant., vol. 23, no. 1, pp. 1–11, 2014.

[39] R. G. Fehon, A. I. McClatchey, and A. Bretscher, “Organizing the cell cortex: The

role of ERM proteins,” Nat. Rev. Mol. Cell Biol., vol. 11, no. 4, pp. 276–287,

2010.

[40] Z. Sun, M. Costell, and R. Fässler, “Integrin activation by talin, kindlin and

mechanical forces,” Nat. Cell Biol., vol. 21, no. 1, pp. 25–31, 2019.

160

[41] N. Nakao, K. Maki, M. R. K. Mofrad, and T. Adachi, “Talin is required to increase

stiffness of focal molecular complex in its early formation process,” Biochem.

Biophys. Res. Commun., no. xxxx, 2019.

[42] A. W. Holle et al., “Cell-Extracellular Matrix Mechanobiology: Forceful Tools

and Emerging Needs for Basic and Translational Research,” Nano Lett., vol. 18,

no. 1, pp. 1–8, 2018.

[43] L. Zhu, H. Liu, F. Lu, J. Yang, T. V. Byzova, and J. Qin, “Structural Basis of

Paxillin Recruitment by Kindlin-2 in Regulating Cell Adhesion,” Structure, vol.

27, no. 11, pp. 1686-1697.e5, 2019.

[44] J. J. Thottacherry et al., “Mechanochemical feedback control of dynamin

independent endocytosis modulates membrane tension in adherent cells,” Nat.

Commun., vol. 9, no. 1, 2018.

[45] T. Ni et al., “Structure and lipid-binding properties of the kindlin-3 pleckstrin

homology domain,” Biochem. J., vol. 474, no. 4, pp. 539–556, 2017.

[46] T. V. Malinin, N.L., Plow, E.F., and Byzova, “Kindlins in FERM adhesion.,”

Blood, vol. 115, pp. 4011–4017, 2010.

[47] et al. Meller, J., Chen, Z., Dudiki, T., Cull, R.M., Murtazina, R., Bal, S.K.,

Pluskota, E., Stefl, S., Plow, E.F., Trapp, B.D., “Integrin-Kindlin3 requirements

for microglial motility in vivo are distinct from those for macrophages.,” JCI

Insight, vol. 2, 2017.

[48] D. a Fletcher and R. D. Mullins, “Cell mechanisms and cytoskeleton,” Nature, vol.

161

463, no. 7280, pp. 485–492, 2010.

[49] T. J. Kirby and J. Lammerding, “Emerging views of the nucleus as a cellular

mechanosensor,” Nat. Cell Biol., vol. 20, no. 4, pp. 373–381, 2018.

[50] S. Cho, J. Irianto, and D. E. Discher, “Mechanosensing by the nucleus: From

pathways to scaling relationships,” J. Cell Biol., vol. 216, no. 2, pp. 305–315,

2017.

[51] J. Lammerding, “Mechanics of the nucleus,” Compr. Physiol., vol. 1, no. 2, pp.

783–807, 2011.

[52] M. Lherbette, Á. Dos Santos, Y. Hari-Gupta, N. Fili, C. P. Toseland, and I. A. T.

Schaap, “Atomic Force Microscopy micro-rheology reveals large structural

inhomogeneities in single cell-nuclei,” Sci. Rep., vol. 7, no. 1, pp. 1–13, 2017.

[53] S. R. Polio, S. Stasiak, R. Krishnan, and H. Parameswaran, “Extracellular matrix

stiffness regulates force transmission pathways in multicellular ensembles of

human airway smooth muscle cells.,” bioRxiv, p. 402842, 2018.

[54] S. Lee, A. E. Stanton, X. Tong, and F. Yang, “Hydrogels with enhanced protein

conjugation efficiency reveal stiffness-induced YAP localization in stem cells

depends on biochemical cues,” Biomaterials, vol. 202, no. January, pp. 26–34,

2019.

[55] S. Dupont et al., “Role of YAP/TAZ in mechanotransduction,” Nature, vol. 474,

no. 7350, pp. 179–184, 2011.

[56] G. Nardone et al., “YAP regulates cell mechanics by controlling focal adhesion

162

assembly,” Nat. Commun., vol. 8, no. May, 2017.

[57] D. Han et al., “YAP/TAZ enhance mammalian embryonic neural stem cell

characteristics in a Tead-dependent manner,” Biochem. Biophys. Res. Commun.,

vol. 458, no. 1, pp. 110–116, 2015.

[58] G. Sorrentino et al., “Metabolic control of YAP and TAZ by the mevalonate

pathway,” Nat. Cell Biol., vol. 16, no. 4, pp. 357–366, 2014.

[59] S. Nasrollahi, C. Walter, A. J. Loza, G. V. Schimizzi, G. D. Longmore, and A.

Pathak, “Past matrix stiffness primes epithelial cells and regulates their future

collective migration through a mechanical memory,” Biomaterials, vol. 146, pp.

146–155, 2017.

[60] J. Baek et al., “Distinct Mechanosensing of Human Neural Stem Cells on

Extremely Limited Anisotropic Cellular Contact,” ACS Appl. Mater. Interfaces,

vol. 10, no. 40, pp. 33891–33900, 2018.

[61] and K. S. A. Chun Yang, Mark W. Tibbitt, Lena Basta, “Mechanical memory and

dosing influence stem cell fate,” Nat. Mater., vol. 13, no. 6, pp. 645–652, 2014.

[62] A. J. Keung, E. M. De Juan-Pardo, D. V. Schaffer, and S. Kumar, “Rho GTPases

mediate the mechanosensitive lineage commitment of neural stem cells,” Stem

Cells, vol. 29, no. 11, pp. 1886–1897, 2011.

[63] T. A. Ulrich and S. Kumar, “Mechanobiology in health and disease in the central

nervous system,” Mechanobiol. Handb., pp. 391–412, 2011.

[64] E. Moeendarbary et al., “The soft mechanical signature of glial scars in the central

163

nervous system,” Nat. Commun., vol. 8, pp. 1–11, 2017.

[65] K. Pogoda et al., “Compression stiffening of brain and its effect on

mechanosensing by glioma cells,” New J. Phys., vol. 16, pp. 1–17, 2014.

[66] S. Tasneem, K. Farrell, M. Y. Lee, and C. R. Kothapalli, “Sensitivity of neural

stem cell survival, differentiation and neurite outgrowth within 3D hydrogels to

environmental heavy metals,” Toxicol. Lett., vol. 242, pp. 9–22, 2016.

[67] I. Ladran, N. Tran, A. Topol, and K. J. Brennand, “Neural stem and progenitor

cells in health and disease,” Wiley Interdiscip. Rev. Syst. Biol. Med., vol. 5, no. 6,

pp. 701–715, 2013.

[68] F. T. Merkle and A. Alvarez-Buylla, “Neural stem cells in mammalian

development,” Curr. Opin. Cell Biol., vol. 18, no. 6, pp. 704–709, 2006.

[69] W. Xu, N. Lakshman, and C. M. Morshead, “Building a central nervous system:

The neural stem cell lineage revealed,” Neurogenesis, vol. 4, no. 1, p. e1300037,

2017.

[70] A. Sivakumar and N. A. Kurpios, “Transcriptional regulation of cell shape during

organ morphogenesis,” J. Cell Biol., vol. 217, no. 9, pp. 2987–3005, 2018.

[71] W. J. Tyler, “The mechanobiology of brain function,” Nat. Rev. Neurosci., vol. 13,

no. 12, pp. 867–878, 2012.

[72] T. Bergström and K. Forsberg-Nilsson, “Neural stem cells: Brain building blocks

and beyond,” Ups. J. Med. Sci., vol. 117, no. 2, pp. 132–142, 2012.

[73] T. Duval et al., “Axons morphometry in the human spinal cord,” Neuroimage, vol.

164

185, pp. 119–128, 2019.

[74] H. May-Simera and C. Liu, “Neuronal Polarity and Neurological Disorders,” J.

Neurol. Transl. Neurosci., vol. 2, no. 1, p. 1026, 2013.

[75] R. Shparberg, H. Glover, and M. B. Morris, “Modelling mammalian commitment

to the neural lineage using embryos and embryonic stem cells,” Front. Physiol.,

vol. 10, no. MAY, 2019.

[76] B. R. Brückner and A. Janshoff, “Elastic properties of epithelial cells probed by

atomic force microscopy,” Biochim. Biophys. Acta - Mol. Cell Res., vol. 1853, no.

11, pp. 3075–3082, 2015.

[77] M. Iwashita, N. Kataoka, K. Toida, and Y. Kosodo, “Systematic profiling of

spatiotemporal tissue and cellular stiffness in the developing brain,” Dev., vol. 141,

no. 19, pp. 3793–3798, 2014.

[78] N. R. Saunders, S. A. Liddelow, and K. M. Dziegielewska, “Barrier mechanisms

in the developing brain,” Front. Pharmacol., vol. 3 MAR, no. March, pp. 1–18,

2012.

[79] S. Bauer, B. J. Kerr, and P. H. Patterson, “The neuropoietic cytokine family in

development, plasticity, disease and injury,” Nat. Rev. Neurosci., vol. 8, no. 3, pp.

221–232, 2007.

[80] K. Farrell, G. Mahajan, P. Srinivasan, M. Y. Lee, and C. R. Kothapalli, “Pediatric

glioblastoma cells inhibit neurogenesis and promote astrogenesis, phenotypic

transformation and migration of human neural progenitor cells within cocultures,”

165

Exp. Cell Res., vol. 362, no. 1, pp. 159–171, 2018.

[81] M. Chalfie, “Neurosensory mechanotransduction,” Nat. Rev. Mol. Cell Biol., vol.

10, no. 1, pp. 44–52, 2009.

[82] L. Ponce and a Berquand, “Combining Atomic Force Microscopy and Live Cell

Imaging to Study Calcium Responses in Dorsal Root Ganglion Neurons to a

Locally Applied Mechanical,” Microsc. Sci. …, pp. 530–536, 2010.

[83] Y. B. Lu et al., “Viscoelastic properties of individual glial cells and neurons in the

CNS,” Proc. Natl. Acad. Sci. U. S. A., vol. 103, no. 47, pp. 17759–17764, 2006.

[84] C. R. Kothapalli and R. D. Kamm, “3D matrix microenvironment for targeted

differentiation of embryonic stem cells into neural and glial lineages,”

Biomaterials, vol. 34, no. 25, pp. 5995–6007, 2013.

[85] W. J. Miller, I. Leventhal, D. Scarsella, P. G. Haydon, P. Janmey, and D. F.

Meaney, “Mechanically induced reactive gliosis causes ATP-mediated alterations

in astrocyte stiffness.,” J. Neurotrauma, 2009.

[86] K. Franze and J. Guck, “The biophysics of neuronal growth,” Reports Prog. Phys.,

vol. 73, no. 9, 2010.

[87] D. E. Koser et al., “Mechanosensing is critical for axon growth in the developing

brain,” Nat. Neurosci., vol. 19, no. 12, pp. 1592–1598, Dec. 2016.

[88] J. T. Kevenaar and C. C. Hoogenraad, “The axonal cytoskeleton: from

organization to function,” Front. Mol. Neurosci., vol. 8, p. 44, Aug. 2015.

[89] T. Li and S. Zhang, “Microgliosis in the Injured Brain: Infiltrating Cells and

166

Reactive Microglia Both Play a Role,” Neuroscientist, vol. 22, no. 2, pp. 165–170,

2016.

[90] C. K. Donat, G. Scott, S. M. Gentleman, and M. Sastre, “Microglial activation in

traumatic brain injury,” Front. Aging Neurosci., vol. 9, no. JUN, pp. 1–20, 2017.

[91] P. Date, “UC Berkeley UC Berkeley Electronic Theses and Dissertations Cellular

Mechanobiology of Glioblastoma Multiforme,” 2011.

[92] K. Pogoda and P. A. Janmey, “Glial Tissue Mechanics and Mechanosensing by

Glial Cells,” Front. Cell. Neurosci., vol. 12, p. 25, Feb. 2018.

[93] K. Franze, “The mechanical control of nervous system development,” Dev., vol.

140, no. 15, pp. 3069–3077, 2013.

[94] E. Lee and W. S. Chung, “Glial control of synapse number in healthy and diseased

brain,” Front. Cell. Neurosci., vol. 13, no. February, pp. 1–8, 2019.

[95] D. E. Ingber, N. Wang, and D. Stamenović, “Tensegrity, cellular biophysics, and

the mechanics of living systems,” Reports Prog. Phys., vol. 77, no. 4, pp. 1–42,

2014.

[96] Y. Fujii, S. Maekawa, and M. Morita, “Astrocyte calcium waves propagate

proximally by gap junction and distally by extracellular diffusion of ATP released

from volume-regulated anion channels,” Sci. Rep., vol. 7, no. 1, p. 13115, Oct.

2017.

[97] N. S. M. Jones T.J., “Mechanosensory Calcium Signaling.,” Islam M. Calcium

Signaling. Adv. Exp. Med. Biol. Springer, Dordr., vol. 740, 2012.

167

[98] L. Heimfarth et al., “Calcium signaling mechanisms disrupt the cytoskeleton of

primary astrocytes and neurons exposed to diphenylditelluride,” Biochim. Biophys.

Acta, vol. 1860, no. 11 Pt A, p. 2510—2520, Nov. 2016.

[99] S. Burdyga, T. & Wray, “The effect of inhibition of myosin light chain kinase by

Wortmannin on intracellular [Ca2+], electrical activity and force in phasic smooth

muscle,” Pflügers Arch, vol. 436, no. 801, 1998.

[100] G. F. Winner B, Kohl Z, “Neurodegenerative disease and adult neurogenesis.,”

Eur. J. Neurosci., vol. 33, no. 6, pp. 1139–1151, 2011.

[101] M. C. Murphy et al., “Decreased brain stiffness in Alzheimer’s disease determined

by magnetic resonance elastography,” J. Magn. Reson. Imaging, vol. 34, no. 3, pp.

494–498, Sep. 2011.

[102] K. Haase and A. E. Pelling, “Investigating cell mechanics with atomic force

microscopy,” J. R. Soc. Interface, vol. 12, no. 104, pp. 6–9, 2015.

[103] E. Moeendarbary and A. R. Harris, “Cell mechanics: Principles, practices, and

prospects,” Wiley Interdiscip. Rev. Syst. Biol. Med., vol. 6, no. 5, pp. 371–388,

2014.

[104] M. Li, D. Dang, L. Liu, N. Xi, and Y. Wang, “Atomic force microscopy in

characterizing cell mechanics for biomedical applications: A review,” IEEE Trans.

Nanobioscience, vol. 16, no. 6, pp. 523–540, 2017.

[105] M. L. Rodriguez, P. J. McGarry, and N. J. Sniadecki, “Review on cell mechanics:

Experimental and modeling approaches,” Appl. Mech. Rev., vol. 65, no. 6, 2013.

168

[106] K. Lykov, Y. Nematbakhsh, M. Shang, C. T. Lim, and I. V. Pivkin, “Probing

eukaryotic cell mechanics via mesoscopic simulations,” PLoS Comput. Biol., vol.

13, no. 9, pp. 1–22, 2017.

[107] Z. Li, B. Anvari, M. Takashima, P. Brecht, J. H. Torres, and W. E. Brownell,

“Membrane tether formation from outer hair cells with optical tweezers,” Biophys.

J., vol. 82, no. 3, pp. 1386–1395, 2002.

[108] S. J. Koch MD, “Introduction to Optical Tweezers. Optical Tweezers,” Methods

Mol. Biol., vol. 3, no. 24, 2016.

[109] et al. Walker LM, Holm Å, Cooling L, “Mechanical manipulation of bone and

cartilage cells with ‘optical tweezers’.,” FEBS Lett., vol. 459, no. 1, pp. 39–42,

1999.

[110] B. Pontes, N. B. Viana, L. T. Salgado, M. Farina, V. M. Neto, and H. M.

Nussenzveig, “Cell cytoskeleton and tether extraction,” Biophys. J., vol. 101, no.

1, pp. 43–52, 2011.

[111] J. P. Butler and S. M. Kelly, “A model for cytoplasmic rheology consistent with

magnetic twisting cytometry,” Biorheology, vol. 35, no. 3, pp. 193–209, 1998.

[112] I. L. Ahmad and M. R. Ahmad, “Technological advancements in characterizing

single cell’s stiffness properties- A review,” IECBES 2014, Conf. Proc. - 2014

IEEE Conf. Biomed. Eng. Sci. “Miri, Where Eng. Med. Biol. Humanit. Meet,” pp.

187–192, 2014.

[113] K. S. Kim, C. H. Cho, E. K. Park, M. H. Jung, K. S. Yoon, and H. K. Park, “AFM-

169

Detected apoptotic changes in morphology and biophysical property caused by

paclitaxel in Ishikawa and HeLa cells,” PLoS One, vol. 7, no. 1, 2012.

[114] M. Lekka, “Discrimination Between Normal and Cancerous Cells Using AFM,”

Bionanoscience, vol. 6, no. 1, pp. 65–80, 2016.

[115] N. I. Nikolaev, T. Müller, D. J. Williams, and Y. Liu, “Changes in the stiffness of

human mesenchymal stem cells with the progress of cell death as measured by

atomic force microscopy,” J. Biomech., vol. 47, no. 3, pp. 625–630, 2014.

[116] X. Liang, X. Shi, S. Ostrovidov, H. Wu, and K. Nakajima, “Probing stem cell

differentiation using atomic force microscopy,” Appl. Surf. Sci., vol. 366, pp. 254–

259, 2016.

[117] B. Cappella and G. Dietler, “Force-distance curves by atomic force microscopy,”

Surf. Sci. Rep., vol. 34, no. 1–3, pp. 1–3, 1999.

[118] M. Krieg et al., “Atomic force microscopy-based mechanobiology,” Nat. Rev.

Phys., vol. 1, no. 1, pp. 41–57, 2019.

[119] L. Zang, “Basics of Atomic Force Microscopy,” pp. 1–55, 2013.

[120] K. Haase and A. E. Pelling, “Investigating cell mechanics with atomic force

microscopy,” J. R. Soc. Interface, vol. 12, no. 104, 2015.

[121] D. J. Müller and Y. F. Dufrêne, “Force nanoscopy of living cells,” Current

Biology. 2011.

[122] A. M. Ahmad IL, “Trends in characterizing single cell’s stiffness properties.,”

Micro Nano Syst. Lett., vol. 2, no. 1, 2014.

170

[123] Hochmuth RM., “Micropipette aspiration of living cells.,” J. Biomech., vol. 33, no.

1, pp. 15–22, 2000.

[124] L. I. Byfield FJ, Aranda-Espinoza H, Romanenko VG, Rothblat GH, “Cholesterol

Depletion Increases Membrane Stiffness of Aortic Endothelial Cells.,” Biophys. J.,

vol. 87, no. 5, pp. 3336–3343, 2004.

[125] G. Bao and S. Suresh, “Cell and molecular mechanics of biological materials,”

Nat. Mater., vol. 2, no. 11, pp. 715–725, 2003.

[126] et al. Fabry B, Maksym GN, Shore SA, “Selected contribution: time course and

heterogeneity of contractile responses in cultured human airway smooth muscle

cells.,” J Appl Physiol, vol. 91, no. 2, pp. 986–994, 2001.

[127] et al. Zhang G, Fan N, Lv X, “Investigation of the Mechanical Properties of the

Human Osteosarcoma Cell at Different Cell Cycle Stages.,” Micromachines., vol.

8, no. 3, p. 89, 2017.

[128] M. R. Fletcher DA, “Cell mechanics and the cytoskeleton.,” Nature, vol. 463, no.

7280, pp. 485–492, 2010.

[129] M. Mak and D. Erickson, “A serial micropipette microfluidic device with

applications to cancer cell repeated deformation studies,” Integr. Biol. (United

Kingdom), 2013.

[130] R. Wilson, Ha and Bullen, “Introduction to Scanning Probe Microscopy (SPM)

Basic Theory Atomic Force Microscopy (AFM),” pp. 1–8.

[131] Y. M. Efremov, W. H. Wang, S. D. Hardy, R. L. Geahlen, and A. Raman,

171

“Measuring nanoscale viscoelastic parameters of cells directly from AFM force-

displacement curves,” Sci. Rep., vol. 7, no. 1, pp. 1–14, 2017.

[132] S. Sen, S. Subramanian, and D. E. Discher, “Indentation and adhesive probing of a

cell membrane with AFM: Theoretical model and experiments,” Biophys. J., vol.

89, no. 5, pp. 3203–3213, 2005.

[133] Y. Wu, T. Yu, T. A. Gilbertson, A. Zhou, H. Xu, and K. T. Nguyen, “Biophysical

assessment of single cell cytotoxicity: Diesel exhaust particle-treated human aortic

endothelial cells,” PLoS One, vol. 7, no. 5, pp. 1–10, 2012.

[134] J. R. Staunton, B. L. Doss, S. Lindsay, and R. Ros, “Correlating confocal

microscopy and atomic force indentation reveals metastatic cancer cells stiffen

during invasion into collagen i matrices,” Sci. Rep., vol. 6, no. January, pp. 1–15,

2016.

[135] D. M. Ebenstein and L. A. Pruitt, “Nanoindentation of biological materials,” Nano

Today, 2006.

[136] M. C. Evans and V. H. Barocas, “The modulus of fibroblast-populated collagen

gels is not determined by final collagen and cell concentration: Experiments and an

inclusion-based model,” J. Biomech. Eng., vol. 131, no. 10, pp. 1–7, 2009.

[137] H. F. Pastrana, A. X. Cartagena-Rivera, A. Raman, and A. Ávila, “Evaluation of

the elastic Young’s modulus and cytotoxicity variations in fibroblasts exposed to

carbon-based nanomaterials,” J. Nanobiotechnology, vol. 17, no. 1, p. 32, Feb.

2019.

172

[138] D. I. Christelle Angely, Ngoc‐Minh, Nguyen Sofia, Andre Dias, Emmanuelle

Planus, Gabriel Pelle, Bruno Louis, Marcel Filoche, Alexandre Chenal, Daniel

Ladant, “Exposure to Bordetella pertussis adenylate cyclase toxin affects integrin‐

mediated adhesion and mechanics in alveolar epithelial cells,” Biol. cell, vol. 109,

no. 8, pp. 293–311, 2017.

[139] A.-A. Ungureanu et al., “Amyloid beta oligomers induce neuronal elasticity

changes in age-dependent manner: a force spectroscopy study on living

hippocampal neurons,” Sci. Rep., vol. 6, p. 25841, May 2016.

[140] M. Unal et al., “Micro and Nano-Scale Technologies for Cell Mechanics,”

Nanobiomedicine. 2014.

[141] M. T. Yang, J. Fu, Y. K. Wang, R. A. Desai, and C. S. Chen, “Assaying stem cell

mechanobiology on microfabricated elastomeric substrates with geometrically

modulated rigidity,” Nat. Protoc., 2011.

[142] M. Dembo and Y. L. Wang, “Stresses at the cell-to-substrate interface during

locomotion of fibroblasts,” Biophys. J., 1999.

[143] C. G. Galbraith and M. P. Sheetz, “A micromachined device provides a new bend

on fibroblast traction forces,” Proc. Natl. Acad. Sci. U. S. A., 1997.

[144] R. D. Kamm, “Cell membrane mechanics and adhesion,” Book, pp. 3–17, 2015.

[145] M. Arpin, D. Chirivino, A. Naba, and I. Zwaenepoel, “Emerging role for ERM

proteins in cell adhesion and migration,” Cell Adhes. Migr., vol. 5, no. 2, pp. 199–

206, 2011.

173

[146] P. H. Puech, A. Taubenberger, F. Ulrich, M. Krieg, D. J. Muller, and C. P.

Heisenberg, “Measuring cell adhesion forces of primary gastrulating cells from

zebrafish using atomic force microscopy,” J. Cell Sci., 2005.

[147] L. Andolfi et al., “Investigation of adhesion and mechanical properties of human

glioma cells by single cell force spectroscopy and atomic force microscopy,” PLoS

One, vol. 9, no. 11, 2014.

[148] G. Sagvolden, I. Giaever, E. O. Pettersen, and J. Feder, “Cell adhesion force

microscopy,” Proc. Natl. Acad. Sci. U. S. A., vol. 96, no. 2, pp. 471–476, Jan.

1999.

[149] A. Singh et al., “Adhesion strength-based, label-free isolation of human

pluripotent stem cells,” Nat. Methods, 2013.

[150] N. Guz, M. Dokukin, V. Kalaparthi, and I. Sokolov, “If Cell Mechanics Can Be

Described by Elastic Modulus: Study of Different Models and Probes Used in

Indentation Experiments,” Biophys. J., 2014.

[151] I. Sokolov and M. Dokukin, “Mechanics of biological cells studied with atomic

force microscopy,” in Microscopy and Microanalysis, 2014.

[152] A. A. Soufivand, M. Navidbakhsh, and M. Soleimani, “Is it appropriate to apply

Hertz model to describe cardiac myocytes’ mechanical properties by atomic force

microscopy nanoindentation?,” Micro Nano Lett., 2014.

[153] J. Helenius, C.-P. Heisenberg, H. E. Gaub, and D. J. Muller, “Single-cell force

spectroscopy,” J. Cell Sci., vol. 121, no. 11, pp. 1785 LP – 1791, Jun. 2008.

174

[154] A. Pietuch, B. R. Brückner, and A. Janshoff, “Membrane tension homeostasis of

epithelial cells through surface area regulation in response to osmotic stress,”

Biochim. Biophys. Acta - Mol. Cell Res., vol. 1833, no. 3, pp. 712–722, 2013.

[155] M. Sun et al., “Multiple membrane tethers probed by atomic force microscopy,”

Biophys. J., vol. 89, no. 6, pp. 4320–4329, 2005.

[156] M. P. Dai, J., and Sheetz, “Membrane tether formation from blebbing cells.,”

Biophys. J., vol. 77, pp. 3363–3370, 1999.

[157] B. Rouven Brückner, A. Pietuch, S. Nehls, J. Rother, and A. Janshoff, “Ezrin is a

Major Regulator of Membrane Tension in Epithelial Cells,” Sci. Rep., vol. 5, pp.

1–16, 2015.

[158] M. Urbanska et al., “Single-cell mechanical phenotype is an intrinsic marker of

reprogramming and differentiation along the mouse neural lineage,” Dev., vol.

144, no. 23, pp. 4313–4321, 2017.

[159] T. D. Nguyen et al., “Piezoelectric nanoribbons for monitoring cellular

deformations,” Nat. Nanotechnol., 2012.

[160] G. Mahajan, M.-Y. Lee, and C. Kothapalli, “Biophysical and biomechanical

properties of neural progenitor cells as indicators of developmental neurotoxicity,”

Arch. Toxicol., no. 0123456789, 2019.

[161] B. Hama, G. Mahajan, and C. Kothapalli, “Characterizing viscoelasticity of

unhydrolyzed chicken sternal cartilage extract suspensions: Towards development

of injectable therapeutics formulations,” J. Mech. Behav. Biomed. Mater., vol. 72,

175

no. April, pp. 90–101, 2017.

[162] Y. Gan, “Invited Review Article: A review of techniques for attaching micro- and

nanoparticles to a probe’s tip for surface force and near-field optical

measurements,” Rev. Sci. Instrum., vol. 78, no. 8, pp. 3–10, 2007.

[163] J. L. MacKay and S. Kumar, “Measuring the Elastic Properties of Living Cells

with Atomic Force Microscopy Indentation,” in Cell Imaging Techniques:

Methods and Protocols, D. J. Taatjes and J. Roth, Eds. Totowa, NJ: Humana Press,

2013, pp. 313–329.

[164] Royal society of Chemistry, “PDMS stiffness with different mixing ratios -

supplementary material,” vol. 633, no. June 2004, pp. 21–23, 2007.

[165] E. Sollier, C. Murray, P. Maoddi, and D. Di Carlo, “Rapid prototyping polymers

for microfluidic devices and high pressure injections,” Lab Chip, vol. 11, no. 22,

pp. 3752–3765, 2011.

[166] R. N. Palchesko, L. Zhang, Y. Sun, and A. W. Feinberg, “Development of

Polydimethylsiloxane Substrates with Tunable Elastic Modulus to Study Cell

Mechanobiology in Muscle and Nerve,” PLoS One, vol. 7, no. 12, p. e51499, Dec.

2012.

[167] I. D. Johnston, D. K. McCluskey, C. K. L. Tan, and M. C. Tracey, “Mechanical

characterization of bulk Sylgard 184 for microfluidics and microengineering,” J.

Micromechanics Microengineering, vol. 24, no. 3, 2014.

[168] Z. Wang, A. A. Volinsky, and N. D. Gallant, “Crosslinking effect on

176

polydimethylsiloxane elastic modulus measured by custom-built compression

instrument,” J. Appl. Polym. Sci., vol. 131, no. 22, pp. 1–4, 2014.

[169] Z. Wang, “Polydimethylsiloxane Mechanical Properties Measured by Macroscopic

Compression and Nanoindentation Techniques,” p. 68, 2011.

[170] M. Ochsner et al., “Micro-well arrays for 3D shape control and high resolution

analysis of single cells,” Lab Chip, vol. 7, no. 8, pp. 1074–1077, 2007.

[171] C. T. McKee, J. A. Last, P. Russell, and C. J. Murphy, “Indentation versus tensile

measurements of young’s modulus for soft biological tissues,” Tissue Eng. - Part

B Rev., vol. 17, no. 3, pp. 155–164, 2011.

[172] W. Han et al., “Oriented collagen fibers direct tumor cell intravasation,” Proc.

Natl. Acad. Sci. U. S. A., 2016.

[173] J. Joshi, G. Mahajan, and C. R. Kothapalli, “Three-dimensional collagenous niche

and azacytidine selectively promote time-dependent cardiomyogenesis from

human bone marrow-derived MSC spheroids,” Biotechnol. Bioeng., vol. 115, no.

8, 2018.

[174] Y. L. Yang, L. M. Leone, and L. J. Kaufman, “Elastic moduli of collagen gels can

be predicted from two-dimensional confocal microscopy,” Biophys. J., vol. 97, no.

7, pp. 2051–2060, 2009.

[175] M. D. C. Lopez-Garcia, D. J. Beebe, and W. C. Crone, “Young’s modulus of

collagen at slow displacement rates,” Biomed. Mater. Eng., vol. 20, no. 6, pp. 361–

369, 2010.

177

[176] B. Lee et al., “A Three-Dimensional Computational Model of Collagen Network

Mechanics,” PLoS One, vol. 9, no. 11, p. e111896, Nov. 2014.

[177] I. P. Weber, S. H. Yun, G. Scarcelli, and K. Franze, “The role of cell body density

in ruminant retina mechanics assessed by atomic force and Brillouin microscopy,”

Phys. Biol., vol. 14, no. 6, 2017.

[178] K. Franze et al., “Spatial mapping of the mechanical properties of the living retina

using scanning force microscopy,” Soft Matter, vol. 7, no. 7, pp. 3147–3154, 2011.

[179] J. T. Morgan, V. K. Raghunathan, S. M. Thomasy, C. J. Murphy, and P. Russell,

“Robust and artifact-free mounting of tissue samples for atomic force

microscopy,” Biotechniques, vol. 56, no. 1, pp. 40–42, Jan. 2014.

[180] A. J. Thompson, E. K. Pillai, I. B. Dimov, S. K. Foster, C. E. Holt, and K. Franze,

“Rapid changes in tissue mechanics regulate cell behaviour in the developing

embryonic brain,” Elife, vol. 8, p. e39356, 2019.

[181] C. J. Sparrey, G. T. Manley, and T. M. Keaveny, “Effects of white, grey, and pia

mater properties on tissue level stresses and strains in the compressed spinal cord,”

J. Neurotrauma, vol. 26, no. 4, pp. 585–595, 2009.

[182] Y. Zhao, G. Mahajan, C. R. Kothapalli, and X.-L. Sun, “Sialylation status and

mechanical properties of THP-1 macrophages upon LPS stimulation,” Biochem.

Biophys. Res. Commun., no. xxxx, pp. 1–6, 2019.

[183] J. M. Maloney et al., “Mesenchymal stem cell mechanics from the attached to the

suspended state,” Biophys. J., vol. 99, no. 8, pp. 2479–2487, 2010.

178

[184] N. L. Sehgel et al., “Increased vascular smooth muscle cell stiffness: A novel

mechanism for aortic stiffness in hypertension,” Am. J. Physiol. - Hear. Circ.

Physiol., vol. 305, no. 9, pp. 1281–1287, 2013.

[185] H. Qiu et al., “Short communication: Vascular smooth muscle cell stiffness as a

mechanism for increased aortic stiffness with aging,” Circ. Res., vol. 107, no. 5,

pp. 615–619, 2010.

[186] Z. Hong, K. J. Reeves, Z. Sun, Z. Li, N. J. Brown, and G. A. Meininger, “Vascular

smooth muscle cell stiffness and adhesion to collagen I modified by vasoactive

agonists,” PLoS One, vol. 10, no. 3, pp. 1–12, 2015.

[187] A. Beliveau, G. Thomas, J. Gong, Q. Wen, and A. Jain, “Aligned Nanotopography

Promotes a Migratory State in Glioblastoma Multiforme Tumor Cells,” Sci. Rep.,

vol. 6, no. May, pp. 1–13, 2016.

[188] C. Tamm and S. Ceccatelli, “Mechanistic insight into neurotoxicity induced by

developmental insults,” Biochem. Biophys. Res. Commun., vol. 482, no. 3, pp.

408–418, 2017.

[189] C. Tamm, J. Duckworth, O. Hermanson, and S. Ceccatelli, “High susceptibility of

neural stem cells to methylmercury toxicity: Effects on cell survival and neuronal

differentiation,” J. Neurochem., vol. 97, no. 1, pp. 69–78, 2006.

[190] Y. S. Lee et al., “Repeated exposure to neurotoxic levels of chlorpyrifos alters

hippocampal expression of neurotrophins and neuropeptides,” Toxicology, vol.

340, pp. 53–62, 2016.

179

[191] N. Malik et al., “Compounds with species and cell type specific toxicity identified

in a 2000 compound drug screen of neural stem cells and rat mixed cortical

neurons,” Neurotoxicology, vol. 45, pp. 192–200, 2014.

[192] K. Seoposengwe, J. J. van Tonder, and V. Steenkamp, “In vitro neuroprotective

potential of four medicinal plants against rotenone-induced toxicity in SH-SY5Y

neuroblastoma cells,” BMC Complement. Altern. Med., vol. 13, pp. 1–11, 2013.

[193] E. J. Ross, D. L. Graham, K. M. Money, and G. D. Stanwood, “Developmental

consequences of fetal exposure to drugs: What we know and what we still must

learn,” Neuropsychopharmacology, vol. 40, no. 1, pp. 61–87, 2015.

[194] F. S. Sanchez, “AFM investigations of cellular response to environmental and

local chemo-mechanical By Fernando Suárez Sánchez,” p. 172, 2011.

[195] A. D. Ebert, P. Liang, and J. C. Wu, “Induced pluripotent stem cells as a disease

modeling and drug screening platform,” J. Cardiovasc. Pharmacol., vol. 60, no. 4,

pp. 408–416, 2012.

[196] A. K. Bal-Price, H. T. Hogberg, L. Buzanska, P. Lenas, E. van Vliet, and T.

Hartung, “In vitro developmental neurotoxicity (DNT) testing: Relevant models

and endpoints,” Neurotoxicology, vol. 31, no. 5, pp. 545–554, 2010.

[197] J. Li, M. L. Spletter, D. A. Johnson, L. S. Wright, C. N. Svendsen, and J. A.

Johnson, “Rotenone-induced caspase 9/3-independent and -dependent cell death in

undifferentiated and differentiated human neural stem cells,” J. Neurochem., vol.

92, no. 3, pp. 462–476, 2005.

180

[198] N. Lomeli, K. Di, J. Czerniawski, J. F. Guzowski, and D. A. Bota, “Cisplatin-

induced mitochondrial dysfunction is associated with impaired cognitive function

in rats,” Free Radic. Biol. Med., vol. 102, pp. 274–286, 2017.

[199] S. Fujisawa et al., “Evaluation of YO-PRO-1 as an early marker of apoptosis

following radiofrequency ablation of colon cancer liver metastases,”

Cytotechnology, vol. 66, no. 2, pp. 259–273, 2014.

[200] K. Farrell, P. Joshi, A. Roth, C. Kothapalli, and M.-Y. Lee, “Chapter 3 High-

throughput Screening of Toxic Chemicals on Neural Stem Cells,” in Human Stem

Cell Toxicology, The Royal Society of Chemistry, 2016, pp. 31–63.

[201] R. A. Roberts et al., “Nitrative and oxidative stress in toxicology and disease,”

Toxicol. Sci., vol. 112, no. 1, pp. 4–16, 2009.

[202] V. A. Movsesyan et al., “Anandamide-induced cell death in primary neuronal

cultures: Role of calpain and caspase pathways,” Cell Death Differ., vol. 11, no.

10, pp. 1121–1132, 2004.

[203] A. Diz-Muñoz, D. A. Fletcher, and O. D. Weiner, “Use the force: Membrane

tension as an organizer of cell shape and motility,” Trends Cell Biol., vol. 23, no.

2, pp. 47–53, 2013.

[204] M. P. Lins et al., “Association between biomechanical alterations and migratory

ability of semaphorin-3A-treated thymocytes,” Biochim. Biophys. Acta - Gen.

Subj., vol. 1862, no. 4, pp. 816–824, 2018.

[205] A. Pillarisetti, J. P. Desai, H. Ladjal, A. Schiffmacher, A. Ferreira, and C. L.

181

Keefer, “Mechanical phenotyping of mouse embryonic stem cells: Increase in

stiffness with differentiation,” Cell. Reprogram., vol. 13, no. 4, pp. 371–380, 2011.

[206] E. K. F. Yim, E. M. Darling, K. Kulangara, F. Guilak, and K. W. Leong,

“Nanotopography-induced changes in focal adhesions, cytoskeletal organization,

and mechanical properties of human mesenchymal stem cells,” Biomaterials, vol.

31, no. 6, pp. 1299–1306, 2010.

[207] M. Yokokawa, K. Takeyasu, and S. H. Yoshimura, “Mechanical properties of

plasma membrane and nuclear envelope measured by scanning probe microscope,”

J. Microsc., vol. 232, no. 1, pp. 82–90, 2008.

[208] M. Dubovický, P. Kovaèovský, E. Ujházy, J. Navarová, I. Brucknerová, and M.

Mach, “Evaluation of developmental neurotoxicity: Some important issues focused

on neurobehavioral development,” Interdiscip. Toxicol., vol. 1, no. 3–4, pp. 206–

210, 2008.

[209] A. Diz-Muñoz et al., “Control of directed cell migration in vivo by membrane-to-

cortex attachment,” PLoS Biol., vol. 8, no. 11, 2010.

[210] R. M. Hochmuth, J. Y. Shao, J. Dai, and M. P. Sheetz, “Deformation and flow of

membrane into tethers extracted from neuronal growth cones,” Biophys. J., vol. 70,

no. 1, pp. 358–369, 1996.

[211] K. Shrader-Frechette, “Taking action on developmental toxicity: Scientists duties

to protect children,” Environ. Heal. A Glob. Access Sci. Source, vol. 11, no. 1, p. 1,

2012.

182

[212] S. L. Makris et al., “Current and future needs for developmental toxicity testing,”

Birth Defects Res. Part B - Dev. Reprod. Toxicol., vol. 92, no. 5, pp. 384–394,

2011.

[213] I. Hristovska and O. Pascual, “Deciphering resting microglial morphology and

process motility from a synaptic prospect,” Front. Integr. Neurosci., vol. 9, no.

JAN2016, pp. 1–7, 2016.

[214] Å. Svensson, F. Azarbayjani, U. Bäckman, T. Matsumoto, and R. Christofferson,

“Digoxin inhibits neuroblastoma tumor growth in mice,” Anticancer Res., vol. 25,

no. 1 A, pp. 207–212, 2005.

[215] N. Arispe, J. C. Diaz, O. Simakova, and H. B. Pollard, “Heart failure drug

digitoxin induces calcium uptake into cells by forming transmembrane calcium

channels,” Proc. Natl. Acad. Sci. U. S. A., vol. 105, no. 7, pp. 2610–2615, 2008.

[216] J. D. Gardner, D. R. Kiino, T. J. Swartz, and V. P. Butler, “Effects of digoxin

specific antibodies on accumulation and binding of digoxin by human

erythrocytes,” J. Clin. Invest., vol. 52, no. 8, pp. 1820–1833, 1973.

[217] R. D. Burke et al., “Developmental neurotoxicity of the organophosphorus

insecticide chlorpyrifos: from clinical findings to preclinical models and potential

mechanisms,” J. Neurochem., vol. 142, no. Suppl 2, pp. 162–177, 2017.

[218] E. Buse et al., “The Placenta in Toxicology. Part I,” Toxicol. Pathol., vol. 42, no.

2, pp. 314–326, 2014.

[219] M. Salama, D. El-Morsy, M. El-Gamal, O. Shabka, and W. M. Y. Mohamed,

183

“Mitochondrial complex I inhibition as a possible mechanism of chlorpyrifos

induced neurotoxicity,” Ann. Neurosci., vol. 21, no. 3, pp. 85–89, 2014.

[220] S. Yamada, Y. Kubo, D. Yamazaki, Y. Sekino, and Y. Kanda, “Chlorpyrifos

inhibits neural induction via Mfn1-mediated mitochondrial dysfunction in human

induced pluripotent stem cells,” Sci. Rep., vol. 7, no. December 2016, pp. 1–12,

2017.

[221] E. Contassot et al., “Arachidonylethanolamide induces apoptosis of human glioma

cells through vanilloid receptor-1,” J. Neuropathol. Exp. Neurol., vol. 63, no. 9,

pp. 956–963, 2004.

[222] E. Contassot, M. Tenan, V. Schnüriger, M. F. Pelte, and P. Y. Dietrich,

“Arachidonyl ethanolamide induces apoptosis of uterine cervix cancer cells via

aberrantly expressed vanilloid receptor-1,” Gynecol. Oncol., vol. 93, no. 1, pp.

182–188, 2004.

[223] I. Cernak et al., “The ‘Dark Side’ of Endocannabinoids: A Neurotoxic Role for

Anandamide,” J. Cereb. Blood Flow Metab., vol. 24, no. 5, pp. 564–578, 2004.

[224] T. K. Yeung, C. Germond, X. Chen, and Z. Wang, “The mode of action of Taxol:

Apoptosis at low concentration and necrosis at high concentration,” Biochem.

Biophys. Res. Commun., 1999.

[225] B. Messner et al., “Apoptosis and necrosis: Two different outcomes of cigarette

smoke condensate-induced endothelial cell death,” Cell Death Dis., 2012.

[226] Z. Li and Z. Zhou, “How are necrotic cells recognized by their predators?,” Worm,

184

2016.

[227] M. Sachet, Y. Y. Liang, and R. Oehler, “The immune response to secondary

necrotic cells,” Apoptosis. 2017.

[228] X. Jiang et al., “Nanotoxicity of Silver Nanoparticles on HEK293T Cells: A

Combined Study Using Biomechanical and Biological Techniques,” ACS Omega,

vol. 3, no. 6, pp. 6770–6778, 2018.

[229] E. C. Alexopoulos, “Introduction to multivariate regression analysis.,”

Hippokratia, 2010.

[230] B. A. Kozel et al., “Biomechanical properties of the skin in cutis Laxa,” Journal of

Investigative Dermatology. 2014.

[231] B. Wang, “Variable selection in ROC regression,” Comput. Math. Methods Med.,

vol. 2013, 2013.

[232] K. R. Reuhl, L. A. Lagunowich, and D. L. Brown, “Cytoskeleton and cell adhesion

molecules: Critical targets of toxic agents,” in NeuroToxicology, 1994.

[233] I. Szekacs, E. Farkas, B. L. Gemes, E. Takacs, A. Szekacs, and R. Horvath,

“Integrin targeting of glyphosate and its cell adhesion modulation effects on

osteoblastic MC3T3-E1 cells revealed by label-free optical biosensing,” Sci. Rep.,

2018.

[234] J. Pi et al., “Detection of lipopolysaccharide induced inflammatory responses in

RAW264.7 macrophages using atomic force microscope,” Micron, vol. 65, pp. 1–

9, 2014.

185

[235] F. Xing, J. Wang, M. Hu, Y. Yu, G. Chen, and J. Liu, “Comparison of immature

and mature bone marrow-derived dendritic cells by atomic force microscopy,”

Nanoscale Res. Lett., vol. 6, pp. 1–9, 2011.

[236] K. Fraczkowska et al., “Alterations of biomechanics in cancer and normal cells

induced by doxorubicin,” Biomed. Pharmacother., 2018.

[237] B. Salter, M.W., and Stevens, “Microglia emerge as central players in brain

disease.,” Nat. Med., vol. 23, no. 625, pp. 1018–1027, 2017.

[238] B. Schafer, D.P., Lehrman, E.K., Kautzman, A.G., Koyama, R., Mardinly, A.R.,

Yamasaki, R., Ransohoff, R.M., Greenberg, M.E., Barres, B.A., and Stevens,

“Microglia sculpt postnatal neural circuits in an activity and complement-

dependent manner.,” Neuron, vol. 74, pp. 691–705, 2012.

[239] K. Kierdorf and M. Prinz, “Microglia in steady state,” J. Clin. Invest., vol. 127, no.

9, pp. 3201–3209, 2017.

[240] K. M. Lenz and L. H. Nelson, “Microglia and beyond: Innate immune cells as

regulators of brain development and behavioral function,” Front. Immunol., vol. 9,

no. APR, 2018.

[241] C. A. Mosser, S. Baptista, I. Arnoux, and E. Audinat, “Microglia in CNS

development: Shaping the brain for the future,” Prog. Neurobiol., vol. 149–150,

pp. 1–20, 2017.

[242] N. R. Patel et al., “Cell Elasticity Determines Macrophage Function,” PLoS One,

vol. 7, no. 9, pp. 1–10, 2012.

186

[243] A. L. Le Roux, X. Quiroga, N. Walani, M. Arroyo, and P. Roca-Cusachs, “The

plasma membrane as a mechanochemical transducer,” Philos. Trans. R. Soc. B

Biol. Sci., vol. 374, no. 1779, 2019.

[244] A. Diz-Muñoz et al., “Membrane Tension Acts Through PLD2 and mTORC2 to

Limit Actin Network Assembly During Neutrophil Migration,” PLoS Biol., vol.

14, no. 6, pp. 1–30, 2016.

[245] A. Colom et al., “A fluorescent membrane tension probe,” Nat. Chem., vol. 10, no.

11, pp. 1118–1125, 2018.

[246] K. Keren, “Membrane tension leads the way,” Proc. Natl. Acad. Sci. U. S. A., vol.

108, no. 35, pp. 14379–14380, 2011.

[247] D. A. Calderwood, I. D. Campbell, and D. R. Critchley, “Talins and kindlins:

Partners in integrin-mediated adhesion,” Nature Reviews Molecular Cell Biology.

2013.

[248] A. Vahabikashi et al., “Probe Sensitivity to Cortical versus Intracellular

Cytoskeletal Network Stiffness,” Biophys. J., 2019.

[249] M. Moser, B. Nieswandt, S. Ussar, M. Pozgajova, and R. Fässler, “Kindlin-3 is

essential for integrin activation and platelet aggregation,” Nat. Med., 2008.

[250] Z. Xu et al., “Direct interaction of kindlin-3 with integrin αiibβ3 in platelets is

required for supporting arterial thrombosis in mice,” Arterioscler. Thromb. Vasc.

Biol., 2014.

[251] P. D. Hsu et al., “DNA targeting specificity of RNA-guided Cas9 nucleases,” Nat.

187

Biotechnol., vol. 31, p. 827, Jul. 2013.

[252] I. Derényi, F. Jülicher, and J. Prost, “Formation and Interaction of Membrane

Tubes,” Phys. Rev. Lett., 2002.

[253] D. Vorselen, F. C. Mackintosh, W. H. Roos, and G. J. L. Wuite, “Competition

between Bending and Internal Pressure Governs the Mechanics of Fluid

Nanovesicles,” ACS Nano, 2017.

[254] et al. Malinin, N.L., Zhang, L., Choi, J., Ciocea, A., Razorenova, O., Ma, Y.Q.,

Podrez, E.A., Tosi, M., Lennon, D.P., Caplan, A.I., “A point mutation in

KINDLIN3 ablates activation of three integrin subfamilies in humans.,” Nat. Med.,

vol. 15, pp. 313–318, 2009.

[255] M. Sun, M., Graham, J.S., Hegedus, B., Marga, F., Zhang, Y., Forgacs, G., and

Grandbois, “Multiple membrane tethers probed by atomic force microscopy.,”

Biophys. J., vol. 89, pp. 4320–4329, 2005.

[256] B. Pontes, P. Monzo, and N. C. Gauthier, “Membrane tension: A challenging but

universal physical parameter in cell biology,” Seminars in Cell and Developmental

Biology. 2017.

[257] M. Segel et al., “Niche stiffness underlies the ageing of central nervous system

progenitor cells,” Nature, vol. 573, no. 7772, pp. 130–134, 2019.

[258] K. Franze et al., “Neurite branch retraction is caused by a threshold-dependent

mechanical impact,” Biophys. J., vol. 97, no. 7, pp. 1883–1890, Oct. 2009.

[259] P. F. Davies, “Hemodynamic shear stress and the endothelium in cardiovascular

188

pathophysiology,” Nat. Clin. Pract. Cardiovasc. Med., vol. 6, no. 1, pp. 16–26,

Jan. 2009.

[260] and L. M. M. Conleth A. Mullen, Ted J. Vaughan, Kristen L. Billiar, “The Effect

of Substrate Stiffness, Thickness, and Cross-Linking Density on Osteogenic Cell

Behavior,” Biophys. J., vol. 108, no. 7, pp. 1604–1612, 2015.

[261] M. Winiecka-Klimek et al., “SOX2 and SOX2-MYC Reprogramming Process of

Fibroblasts to the Neural Stem Cells Compromised by Senescence,” PLoS One,

vol. 10, no. 11, p. e0141688, Nov. 2015.

[262] P. Joshi, “Three-Dimensional Human Neural Stem Cell Culture for High-

Throughput Assessment of Developmental Neurotoxicity.,” Electron. Thesis or

Diss. Clevel. State Univ., vol. 1, no. 1, 2019.

[263] T. Lourenço et al., “Modulation of oligodendrocyte differentiation and maturation

by combined biochemical and mechanical cues,” Sci. Rep., vol. 6, no. February,

pp. 1–17, 2016.

[264] A. J. Engler, S. Sen, H. L. Sweeney, and D. E. Discher, “Matrix Elasticity Directs

Stem Cell Lineage Specification,” Cell, 2006.

[265] D. V. S. SEBASTIAN RAMMENSEE, MICHAEL S. KANG, KATERINA

GEORGIOU, SANJAY KUMAR, “Dynamics of Mechanosensitive Neural Stem

Cell Differentiation,” Stem Cells, pp. 2792–2800, 2016.

[266] K. E. Hammerick et al., “Elastic properties of induced pluripotent stem cells,”

Tissue Eng. Part A, vol. 17, no. 3–4, pp. 495–502, Feb. 2011.

189

[267] T. Bongiorno, J. L. Chojnowski, J. D. Lauderdale, and T. Sulchek, “Cellular

Stiffness as a Novel Stemness Marker in the Corneal Limbus,” Biophys. J., vol.

111, no. 8, pp. 1761–1772, 2016.

[268] I. Titushkin and M. Cho, “Modulation of cellular mechanics during osteogenic

differentiation of human mesenchymal stem cells,” Biophys. J., vol. 93, no. 10, pp.

3693–3702, 2007.

[269] T. Bongiorno et al., “Mechanical stiffness as an improved single-cell indicator of

osteoblastic human mesenchymal stem cell differentiation,” J. Biomech., vol. 47,

no. 9, pp. 2197–2204, Jun. 2014.

[270] Q. CHEN et al., “AFM Studies of Cellular Mechanics during Osteogenic

Differentiation of Human Amniotic Fluid-derived Stem Cells,” Anal. Sci., vol. 26,

no. 10, pp. 1033–1037, 2010.

[271] Y.-B. Lu et al., “Viscoelastic properties of individual glial cells and neurons in the

CNS,” Proc. Natl. Acad. Sci., vol. 103, no. 47, pp. 17759 LP – 17764, Nov. 2006.

[272] A. Pillarisetti, J. P. Desai, H. Ladjal, A. Schiffmacher, A. Ferreira, and C. L.

Keefer, “Mechanical Phenotyping of Mouse Embryonic Stem Cells: Increase in

Stiffness with Differentiation,” Cell. Reprogram., vol. 13, no. 4, pp. 371–380, Jul.

2011.

[273] D. R. Overby et al., “Altered mechanobiology of Schlemm’s canal endothelial

cells in glaucoma,” Proc. Natl. Acad. Sci. U. S. A., 2014.

[274] D. C. Van Essen, “A tension-based theory of morphogenesis and compact wiring

190

in the central nervous system,” Nature, vol. 385, no. 6614, pp. 313–318, 1997.

[275] J. Zheng, P. Lamoureux, V. Santiago, T. Dennerll, R. E. Buxbaum, and S. R.

Heidemann, “Tensile regulation of axonal elongation and initiation,” J. Neurosci.,

vol. 11, no. 4, pp. 1117 LP – 1125, Apr. 1991.

191

APPENDICES

192

193

194

195

196

197

198