This item has the following downloads:
1 INTEGRATED CELLULAR BIOPHYSICAL MIGRATION APPROACH By SHEN SHIU HUNG A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHI LOSOPHY UNIVERSITY OF FLORIDA 2013
2 2013 Shen Hsiu Hung
3 To my family
4 ACKNOWLEDGMENTS Recalling the six year life in University of Florida, it was f ull with challenge and hap piness Here I appreciate lots of advisors in my life. I would like to thank to my Ph.D advisor, Dr. Yiider Tseng. For his guidance and patience, I learn lots knowledge from him, not only in academic research, but also the attitude of life. Also, I am tha nkful to my committee members, Dr. Richard Dickinson, Dr. Tanmay Lele and Dr. Brandi Ormerod for their valuable suggestions and expertise to make my work done well. I wound put special thanks to my all group membranes for their assistance in many ways. In managing the lab, I wound thank for Jun Yin and Stephens help to make the lab work well. In my experiment, Jacob Li Kai Cheng, Jeff Kao, Ellie Fan, and Qian Peng work lots on my research. I also appreciate lots undergraduates help, Weijie, Omar, Justin and Jose. Without your help, these works cannot be done smoothly. Further I want to thank to my parents and my girlfriend for their love and support in these years. I wound share this achievement with you because this work could not have been accompl ished without your support Finally, I truly appreciate everyone who helps me during these years.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 7 LIST OF FIGURES .......................................................................................................... 8 LIST OF ABBREVIATIONS ........................................................................................... 1 0 ABSTRACT ................................................................................................................... 11 CHAPTER 1 INTRODUCTION .................................................................................................... 12 Cell Migration and Subcellular Activities ................................................................. 12 The Importance of Cell Migration in Cell Biology .................................................... 13 Quantitative Approach to Cell Migration Wound Disclosure Approach ................. 15 Quantitative Approach to Cell Migration Single Cell Trajectory ............................ 16 The Relationship between Cell and Nucleus .......................................................... 18 Rho GTPases Signaling Pathway in Cell Migration ................................................ 19 Integrated Concept for Cell Migration ..................................................................... 21 2 MATERIAL AND METHODS .................................................................................. 26 Cell Cultures and Plasmid Transfection .................................................................. 26 Plasmid Delivery and Drug Treatment .................................................................... 26 Microscopy and Image Acquisition for Single Cell .................................................. 27 Wound Healing Assay ............................................................................................. 28 Single cell Trajectory Analysis ................................................................................ 28 Immunoblotting ....................................................................................................... 29 Rho GTPases Activity Assay .................................................................................. 30 3 RESULTS AND CONCLUSIONS ........................................................................... 31 The Change in Consistence between the C ell and Its Coupled Nucleus Trajectories under Different Time Scales ............................................................. 31 The Consistence between Cell and Nucleus Displacement Suggests Different Migration Patterns ............................................................................................... 31 Highly Correlated Cell and Nucleus Reveals Longer Displacement ....................... 32 A CellNucleus (C N) Displacement Correlation Setup ........................................... 33 Biophysical Cellular Activities on CN Correlation ................................................... 34 Colorful Barcode Gives a Quick Screening ............................................................. 35 Specific Subcellular Activity in Cell Migration .......................................................... 36 C N correlation as A Cellular Parameter ................................................................. 37
6 Cell Migration Potential Index ( CMPI) ..................................................................... 39 Cell Trajectories Was Set as a Gold Standard of Cell Migration ............................. 40 Persistence Time Cannot Give a Stable Estimation on C ell Migration Ability ......... 41 CMPI, a Better Estimation on Mesenchymal Cell Migration .................................... 42 The Stability of CMPI .............................................................................................. 42 Collective Approach, Wound Closure Assay, Used to Judge Cell Migration and Consistency with CMPI ........................................................................................ 43 CellCell Interaction in Wound Closure approach Can Be Illustrated by CMPI ....... 44 Cell Density, another Potential Issue for Wound Closure Approach ....................... 45 CMPI Has Distinct Advantages over Existing Migration Evaluations ...................... 46 A Cellular Parameter for Distinguishing Cells ......................................................... 47 Cellular Activities Disrupted by Drug Tr eatment Alter the C N Correlation ............. 48 The Contribution of This Work ................................................................................ 50 4 FUTURE WORK ..................................................................................................... 85 A Potential Approach for miRNA Assay .................................................................. 85 Pattern Recognition ................................................................................................ 86 Application of Pattern Recognition: Rho GT Pases Activity under PI3K Inhibitors ... 88 LIST OF REFERENCES ............................................................................................... 93 BIOGRAPHICAL SKETCH .......................................................................................... 101
7 LIST OF TABLES Table page 3 1 The composition of occurrence and average cell displacement in region I and II of different cell lines and their related CMPI. ................................................... 83 3 2 Student T test between cell lines in occurrence profile ....................................... 84
8 LIST OF FIGURES Figure page 1 1 Cell migration in embryonic development ......................................................... 23 1 2 Cell migration in wound healing ......................................................................... 24 1 3 Cell migration in metastasis. ............................................................................... 25 3 1 The consistence between cell and nucleus trajectories at different time period.. ............................................................................................................... 52 3 2 Consistence of cell and nucleus trajectories suggests long er cell displacement. ..................................................................................................... 53 3 3 Highly correlated cell and nucleus reveals longer displacement. ........................ 54 3 4 Correlation between cell and nucleus displacement. .......................................... 55 3 5 The relations between the CCD and NCD// can be built in a NCD// vs. CCD plot. ..................................................................................................................... 56 3 6 Colorful barc ode representing migration behaviors. ........................................... 57 3 7 Colorful barcodes with their related cell migration behaviors. ............................. 58 3 8 Trailing end detachment on C N correlation. ...................................................... 59 3 9 Protrusion on C N correlation. ............................................................................ 60 3 10 Membrane ruffling on C N correlation. ................................................................ 61 3 11 Complexity subcellular activities on C N correlation. .......................................... 62 3 12 Retraction on C N correlation. ............................................................................ 63 3 13 Sketch of calculation of CMPI. ............................................................................ 64 3 14 Trajectory as gold standard for cell migration. .................................................... 65 3 15 T he correlation between < 10hNCD > and persistence time. ............................ 66 3 16 The correlation between < 10hNCD > and CMPI. ............................................. 67 3 17 The stabi lity of CMPI. ......................................................................................... 68 3 18 Wound recovery image and recovery rate of epithelial cells. .............................. 69
9 3 19 Well correlation between CMPI and recovery rate. ............................................. 70 3 20 Recovery rate at different time point suggests cell cell contact effect. ............... 71 3 21 Correlation between recovery rate and CMPI at different time points. ................ 72 3 22 Different migration pattern during wound disclosure. .......................................... 73 3 23 Western blotting s upports the cell cell contact issue. ......................................... 74 3 24 The introduction of E cadherin into MDA 231 cells hindered the woundlike gap recovery process. ........................................................................................ 75 3 25 Cell density plays roles in wound healing assay altering migration ability. ......... 76 3 26 C N correlation suggests a highly correlated between two NIH 3T3 groups. .... 77 3 27 Similar angular distribution but significant different in magnitude of the CCD in CN correlation. ............................................................................................... 78 3 28 The distinguishing between protrusion and ruffling in Swiss 3T3 and OVCAR 3. ........................................................................................................................ 79 3 29 Cellular activities disrupted by Y27362 alter the C N correlation. ....................... 80 3 30 Cellular activities disrupted by Nocodazole alter the C N correlation. ................ 81 3 31 Cellular activities disrupted by Arp2/3 shRNA alter the C N correlation. ............ 82 4 1 The correlation between persistence time or CMPI and 10 hour nucleus distance (<10 hNCD>). ...................................................................................... 90 4 2 R ho GTPases pattern recognition. ..................................................................... 91 4 3 Rho GTPases activity estimation through pattern recognition. ........................... 92
10 LIST OF ABBREVIATIONS CCD Cellular centroid displacement h CCD One hour cell displacement m CCD One minute cell displacement NCD Nucleus centroid displacement m NCD One minute nucleus displacement Rx, Ry Correlation between cell and nucleus in X and Y axes Rxy Summary of Rx and Ry NCD// P rojection of nucleus displacement on its related cell displacement NCD Nuc leus displacement is perpendicular to the direction of CCD CMPI Cell migration potential index Occur Occurrence in regions C N Cell and nucleus correlation profile LD Low cell density condition M D Medium cell density condition HD High cell density condition PPMCC Person product moment correlation coefficient
11 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy INTEGRATED CELL ULAR BIOPHYSICAL MIGRATION APPROACH By SHEN HSIU HUNG August 2013 Chair: Yiider Tseng Major: Chemical Engineering Cell migration has been considered as a critical and necessary component for cell physiology and pathology. However, current assessments to such cell activity are incomprehensive and greatly limit our understanding of cells through this aspect. Here, a cell migration analysis, utilizing the corresponding nuclear displacements as a reference to judge the effectiveness of the cellular displacem ent, is proposed to measure the cell migration potential. Through this approach, cell migration can be described much more comprehensive.
12 CHAPTER 1 INTRODUCTION Cell Migration and Subcellular Activities Cell migration is a fundamental cellular activity th at includ es detachment, protrusion and translocation, and has been studied since 1970 (Abercrombie et al., 1970; Abercrombie et al., 1971; Chen, 1981a; Chen, 1981b) C ell migration is typically characterized by a c ombination of three stages (Kaverina et al., 2002) : 1) cells form lamellipodia and filopodia f orm substrate sampling to create or reinforce cell substrate adhesions; 2) cells polarize via nuclear relocation, which is thought to help align the cytoskeletal components necessary for support and proper movement; and 3) cells detach their adhesions at the trailing end to release the cell from its former anchorage sites allow ing it to move forward. Considering the cycle of these three stages, there emerges certain unique migration patterns reflecting different migration strategies, such as random or persistent movement. Further, these motility phenomena can be explained by molecular mechanisms. The regulation of integrin which mediates cell adhesion to extracellular matrices and react s to actincytoskeleton rearrangement, plays a critical role during cell detachment and anchorage during protrusion (Buck and Horwitz, 1987; Palecek et al., 1998; Tamkun et al., 1986) Hall, Ridley and Nobes identified that Rho GTPases, RhoA, Rac1 and Cdc42, play regulatory roles to control cytoskeleton remodeling (Nobes and Hall, 1995; Ridley and Hall, 1992; R idley et al., 1992) This later led to progress in understanding the molecular mechanisms for migrationrelated subcellular activities protrusion, detachment and translocation (Ballestrem et al., 2001; Machacek et al., 2009; Mullins et al., 1998; Ren et al., 1999) These studies construct an integrative
13 frame that tightly fuses subcellular activities and its underlying molecular mechanism together. This integrative concept started a new era in the study of cell m igration. The Importance of Cell Migration in Cell Biology Cell migration is an essential cell activity for a broad range of physiological and pathological events, such as embryonic development (Richardson and Lehmann, 2010) wound healing following an immune response (Martin and Parkhurst, 2004) and cancer metastasis (Bacac and Stamenkovic, 2008) In early embryo, three major parts of the germ, the endoderm, mesoderm, and ectoderm are classified by functional ability. Endoderm, in the internal layer of germ, is related to stomach, pancreas and thyroid development; mesoderm and ectoderm, middle and external layers of germ, separately mainly develop into muscle and neural crest. During the development process, chemotaxis dominated cell migration plays an important role in tissue and organ development (Figure 1 1) For example, the dors al migration of endodermal in gastrulation of zebrafish is regulated by the Sdf1/Cxcr4 signaling pathway (Mizoguchi et al., 2008) Furth er, actin remodeling through Rac1 and Prex1 during endodermal cell migration has been suggested by Nodal signaling (Woo et al., 2012) Adhesion regulation, different peptide motif coatings, (Winklbauer, 1990) and PDGF chemtaxis (Nagel et al., 2004) all also play an essential role in the guidance of mesoderm migration in Xenopus gastrula. For muscle precursor cell migration, from mesoderm, the mechani sm of EphA4/ephrinA5 and Pax 3 are essential for well developed muscle (Bober et al., 1994; Swartz et al., 2001) Furthermore, inhibition of Hif suggests a clue f or epithelial to mesenchymal transition study (Barriga et al., 2013)
14 Thus, during embryonic development, a highly coordinated cell migration defines the correct cell shape and positions in organogenesi s, whereas defects from such activity will result in fatal damage (Kurosaka and Kashina, 2008) Wound healing is a regular process in life, comprised of three basic phases: inflammation, proliferation and maturation (Figure 1 2) Inflammation happens immediately after the bleeding stops when fluids that cont ain plasma proteins, antibodies and blood cells are sent to wound regions. When inflammation is subsiding, fibroblasts start to enter and gather around the wounds and granulation tissue will be seen to gradually fill up the cavity Subsequently, remodeling occurs where collagen fibers are laid down, nerve ends are regrown and tissue rearranges itself During the inflammation process, the immune system is activated and migrates to the wound region to work against extracellular materials. This migration mechanism was mainly suggested as being regulated by Rho GTPases and PI3K activities (Jones, 2000) Current research further suggests that PI3K inhibitors are potential agents i n inflammation, especially in gamma and delta isoforms (Ellard et al., 2012; Ghigo et al., 2010) Further, under PDGF regulation, fibro cells will move toward the wound region and release collagen, glycosaminoglyc an, and collagenase (Lynch et al., 1987) In addition to PDGF, other growth factors, such as epidermal growth factor (EGF) and IGF I, have also been suggested to play an essential role in cell migration during wound healing (Buckley et al., 1985; Lynch et al., 1989) Thus, the wound healing process involves not only growth factor stimulated cell proliferation, but also chemokineinduced cell migration (Greenhalgh, 199 6; Grose et al., 2002; Nath and Gulati, 1998)
15 In addition, metastasis, a process where cancer cells migrate from the original position to a distant position, mainly involved four steps (Figure 13). 1) Detachment from the original position, in which, the ability of cellcell contact is greatly reduced. 2) Migration between cell cell intervals. The cell has a more dynamic strategy to overcome the cell cell contact issue during the migration, including MMP. 3) Intravasation, where the cancer cell reaches t he blood or lymphatic vessels to penetrate into the circular system. And 4) extravasation, where under certain conditions, the cells penetrate through the wall of circular system and into the tissue or organ. Even though the cancer metastasis contains many different processes, cell migration is the critical process that changes cancer cells from their original anchorage positions to new positions to settle in a distal environment (MacDonald et al., 2002) Currently, it is thought that these migration processes adopt either mesenchymal like or amoeboidal like moving strategies to migrate (Wolf et al., 2003; Yoshida and Soldati, 2006) Quantitative Approach to Cell Migration Wound Disclosure Approach Cell migration is essential in several physiology activities and the next research milestone is to syste mically qualify these activities and integrate our understanding with various molecular mechanisms. To access cell migration, two quantitative strategies have been developed that view cells as simple objects. One approach is to use a small tip to scratch a n artificial wound to monitor healing during wound recovery, and to note especially the recovery rate. This woundlike gap recovery process (generally called woundhealing assay) has been developed to study collective cell migration for over 40 years (Lipton et al., 1971; Todaro et al., 1965) This approach quantifies the progress of cell migration into a denuded area after a woundlike gap has been generated on a monolayer of confluent cells, mimicking the wound healing process in the body. Hence,
16 the different time extents required for a cell type to recover a woundlike gap under different extracellular stimuli (drugs) or intracellular alternations (mutations or expression level changes in proteins), can be used to identify disruptive factors on the wound recovery capacity of the tested cluster of cells (Lee et al., 2000) Quantitative Approach to Cell Migration Single Cell Trajectory The other quantitative approach for studying cell migration is to use a mathematical method that captures cell trajectories. Through timelapse microscopy, the distribution of cell trajectories can give direct visual information to suggest a migration pattern. Over a lengthy observation, cells can be easily seen as a single particle and its migration can be addressed through physical concepts. To investigate the trajectory information, mean square displacement (MSD) and its derived persistent random walk model, reflects the capacity for a cell to consistently mov e along a certain direction due to such a capacity may possess greater determination over the capacity of a cell to hop a greater displacement at a time interval In 1970, a diffusion constant, D, of a two dimensional random walk was suggested as a cell migration characteristic when observed over a suitably long period (Gail and Boone, 1970) This study initiated an era of mathematical modeling of cell migration. Dunn and Brown put forth a quantitative analysis with stochastic process on single cell migration and emphasized several benefits of this quantitative analysis on cell motility: 1) a concise description of the motile process 2) supplying a well communicated platform to make determinations, 3) i ntroduction of a comprehensive discipline integrated with a mathematical model concept s (Dunn, 1983; Dunn and Brown, 1987) This mathematical concept was widely applied in cell migration studies, including modeling analysis of endothelial cell migration under chemotaxis (Stokes et
17 al., 1991) and adhesion mechanics d uring cell migration (DiMilla et al., 1991; Schmidt et al., 1993) and further in combination with statistical concepts to investigate cell migration (Dickinson et al., 1993) Concurrently, another approach was pursued utilizing directionality, defined as the division of the final cellular displacement (D) during a time course by the accumulated transient displacements (T). This gives an alternative and simple method to study cell movement patterns, whether persistence on direction or when random without directionality. If the D over T is close to 1, it suggests a persistent movement. Instead, if the ratio is far away from 1, this reveals a random movement. This approach was adopted by Yamada to examine PTEN downstream signaling, with Shc and Fak, and reveals different capacities for directional persistence (Gu et al., 1999) With the linkage between cellular activities and underlying signaling pathways widely investigated (Huttenlocher et al., 1995; Oliver et al., 1994; Ridley and Hall, 1992; Theriot, 1994) t he overall information and knowledge led to an integrated discipline platform to investigate cell migrat ion from both molecul ar and physical viewpoints (Lauffenburger and Horwitz, 1996) These integrative concepts affected later research in cell migration a nd further clarified the connection between molecular mechanisms and endpoint cellular activities. For example, a persistent zig zag migration searching strategy in Eukaryotic cells was revealed when there is a deficiency in nutrients (Li et al., 2008) The stage s of maturation (double positive or single positive) of developing thymocytes was quantitatively suggested by its migration capacity (Witt et al., 2005) Also, t he persist ence of cell migration quantified by directionality was suggested to be linked with
18 Rac1 activity (Pankov et al., 2005) More integrative approaches adopting microarray techniques t hat take abundant protein expressions into account have suggested that the migration capacity of cells is related to the expression level of HER2 proteins (Wolf Yadlin et al., 2006) These various approaches, however, have yet to reveal the relation between the contributed subcellular activities and the underlying signaling pathways T herefore, these approaches cannot effectively unveil multi factor induced pathological problems, su ch as those occurring in cancer. In addition, as cancer cells can adopt either mesenchymal like or amoeboidal like mov ement strateg ies to migrate (Wolf et al., 2003; Yoshida and Soldati, 2006) and mesenchymal like movement is known to include several different cell migration patterns (Sahai and Marshall, 2003; Sanz Moreno et al., 2008; Yamazaki et al., 2009) the contents of the subcellular activities assembling the cell migration patterns cannot be overlooked when pursuing cellmigration characterization. The Relationship b etween Cell a nd Nucleus Cell migration is a highly dynamic process, in which a fine force balance is required among all subcellular compartments, includi ng the nucleus that occupies a significant portion of the cell volume and weight, to properly maintain the overall coherence of the cell when it migrates. It has been understood that Rho GTPases are also involved in nuclear positioning during neuron devel opment (Heasman and Ridley, 2008) Recent studies further suggest that the nucleus and the cytoskeleton have a tight physical association, which is a common cell phenomenon and not limited to neurons. It has been found that the actin network can form a cap, surrounding the apical nuclear and perinuclear
19 region to provide structural support to the nucleus that further guides the nuclear shape and regulates its position (Khatau et al., 2009) Further, it has also been shown that the actin cap is composed of actomyosin filaments on the top nuclear region that connect s to the nucleoskeleton (Khatau et al., 2009) through the LINC complex (Crisp et al., 2006) including nesprins and SUN proteins. Since the nuclear shape and movement, i.e., translocation and rotation, are tightly regulated by the nucleoskeletoncytoskeleton connection (Lee et al., 2005) it is clear that the activities of Rho GTPases would also be involved in nuclear positioning during cell migration. Hence, a deeper study of nuclear movem ent, in particular the correlation between nuclear movement and cell movement, provides us another reference to understand cell migration. Another Rho GTPases RhoA, works locally at the trailing edge of cells where the nucleus has been observed to be loc ated due to polarization. Since it was shown that the actin cap contains actomyosin filaments in the apical nuclear region and it is believed that the nucleus is connected to the cytoskeleton through the LINC complex (Crisp et al., 2006) nucleus movement could be controlled by RhoA activity through ROCK (Khatau et al., 2009) This hypothesis could be supported by the recent observation that the rearward movement of the nucl eus was enhanced by the LPA treatment in the edge of wounds (Gomes et al., 2005) and the LINC complex could transmit a force to facilitate nucleus movement (Luxton et al., 2 010) Rho GTPase s S ignaling P athway in C ell M igration A multi scaled cell activity such as cell migration is usually accomplished through a complicated signaling network controlled by proteins at certain nodes within the network (Gimona, 2008; Zhang et al., 2011) A nodal protein could exist that responds
20 to its collective upstream signals and distributes the ensemble activity accordingly to the pathways downstream from the nodal point. In the case of cell motilit y, the Rho GTPase s family of proteins, RhoA, Rac1 and Cdc42, act as the three critical nodal proteins in cytoskeletal remodeling (Small et al., 1999) They usually work together to accomplish a cell remodeling event, such as cell protrusion (Machacek et al., 2009) or cell division (Yoshizaki et al., 2003) Studies using fluorescent resonance energy transfer (FRET) demonstrated that t he activities of RhoA, Rac1 and Cdc42 reach a dynamic spatiotemporal harmony to coordinately achieve cell sampling (Machacek et al., 2009) However, the signaling crosstalk between RhoA and Rac1 is far more complic ated (Bu rridge and Wennerberg, 2004) It has been suggested that the activity of Rac1 can activate RhoA for serum starved cells (Ridley et al., 1992) and inhibit RhoA through the oxygen radical p190RhoGAP pathway (Nimnual et al., 2003) or through downstream signals of Rac 1 (Sander et al., 1999; van Leeuwen et al., 1999) Because RhoA and Rac 1 activity have a specific local effect in cells and they mutually regulate each other through multiple signaling pathways (Scita et al., 2000) it is difficult to precisely predict overall cell behavior such as its migration status based only on the activity of one protein or even the activity profiles of a group of protei ns that belong to the pathways of the Rho GTPases. D espite elucidation over the past two decades and abundant knowledge about the structures and functions of proteins and the interactions among proteins that belong to the Rho GTPases signaling pathways, a full understanding of overall cell migration at the cellular level is still unavailable.
21 Integrated Concept for Cell Migration An integrative concept for studying cell migration emerged in the 1980s This approach tried to use molecular mechanism s and mathematic al modeling to decipher cell migration. From a molecular viewpoint, cell migration can be tightly managed from its collective subcellular activities, which can be smoothly explained by upstream molecular interactions, such as the dynamics of Arp2/3 (Mullins et al., 1998) or integrins (Tamkun et al., 1986) and actin remodeling, and further link s to migration nodal proteins, such as Rho GTPases (Hall, 1998; Nobes and Hall, 1995; Ridley and Hall, 1992; R idley et al., 1992) However, th is complex signaling crosstalk cannot be effectively utilized to unveil certain issue s in cell migration. For example, the feedback mechanism in molecular interaction is difficult to detect and may play a determining step i n the regulation of cell migration. M athematic al modeling in cell migration may also overlook subcellular activities and cannot completely describe true migration behaviors. Unfortunately, apart from qualitatively distinguishing movement patterns, there i s no advanced approach relevant to biological aspects that can quantitatively analyze these movement patterns. Hence, an integrative approach is needed which has the capacity to decipher cell migration patterns to their subcellular activities. This type of approach can further connect subcellular activities with the underlying signaling pathways to fill the huge gap between cellular and molecular knowledge. Closing this gap would have a tremendous impact on biomedical research, as evidenced by the highly ci ted numbers associated with those articles focusing on cell migration. Therefore, we are developing a tool to directly connect cell migration and its underlying signaling pathway to address the gap between molecular and cellular considerations This approach utilizes individual nuclear displacements as indices to
22 screen out effective coupled cellular displacements as the contribution components for cell migration.
23 Figure 11. Cell migration in embryonic development (modified from Richardson, B. and Lehmann, R. (2010). Mechanisms guiding primordial germ cell migration: strategies from different organisms. Nature Reviews Molecular Cell Biology 11, 3749 ).
24 Figure 12. Cell migration in wound healing ( adapted from Stein, C. and Kchler, S. (2013). Targeting inflammation and wound healing by opioids. Trends Pharmacol Sci 34, 303 12).
25 Figure 13. Cell migration in metastas is ( modified from Bacac, M. and Stamenkovic, I. (2008). Metastatic cancer cell. Annual Review of Pathology Mechanisms of Disease 3 221 247 ) .
26 CHAPTER 2 MATERIAL AND METHODS Cell C ultures and Plasmid T ransfection NIH 3T3, Swis s 3T3 and human foreskin fibroblasts, MDA 231 cells (all purchased from ATCC, Manassas, VA), and MDA 231 derived E cadherinintroduced stable cells (a generous gift from Denis Wirtz, The Johns Hopkins University) were cultured in DMEM containing 10% fetal bovine serum (FBS) and 1% Lglutamine (Mediatech, Manassas, VA). OSE 10, OVCAR 3 and SKOV 3 cells ( generous gifts from LeMing Shih, The Johns Hopkins Medical Institute) were cultured in RPMI 1640 medium with 10% FBS, 1% Lglutamine and 1% penicillin strep tomycin (Mediatech). U 2 OS cells (ATCC) were cultured in McCoys 5a medium with 10% FBS and 1% Lglutamine (Mediatech). Cultured cells were maintained in a humidified incubator at 37 C and 10% CO2. Plasmid D elivery and D rug T reatment Cells were cultured to around 70% confluency before plasmid delivery. The transfection procedure was followed by a general protocol with Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Plasmid, pRFP R CS (Origene, Rockville, MD) containing a red fluorescent protein gene, was used to confine the cell boundary. HuSH RhoGTPasse RNA interference constructs against RhoA, Rac1, and Cdc42 in the pRFP R CS plasmids were purchased from Origene. The other shRNA constructs were separately cloned in the vector, pLKO.1, to be against Tiam1 (target sequence, GACATCAAGGAGACAGACATC), HOXD10 (target sequence, GCAAGTGATCTGTAATCCCTA) and Arp2 (target sequence, GCAGTTATCCATTATAGGGAT) mRNA (from The Open Biosystems Expression Arrest
27 TRC library, Thermo Scientific, Waltham, MA) and were provided by the Interdisciplinary Center for Biotechnology Research (ICBR), University of Florida. The miRNA precursor hsa miR10b 5p was purchased from Invitrogen. For pLKO.1 shRNA plasmids and miRNA transfection, an additional one tenth of pRFP R CS was applied. Y27 632 (Millipore, Billerica, MA) and Nocodazole (Millipore) were applied separately with 20 M and 1 g/ml (final concentration) on cell for 30 minutes before image taking. For the separately 5 M and 0.1 M, and both were purchased from Chemdea (Ridgewood, NJ). Afterward, the transfected cells were seeded on fibronectin (BD B iosciences, San Jose, CA) coated glass bottom dishes for 24 hours prior to image acquisition. For livecell nuclear stain, H33342 nucleus dye (SigmaAldrich, St. Louis, MO) was applied to the probed cells 5 minutes prior the image acquisition. Microscopy and Image Acquisition for Single C ell A Nikon TE 2000 microscope (Nikon, Melville, NY), equipped with a X Cite 120PC fluorescent light source (EXFO, Ontario, Canada), a Cascade: 1K CCD camera (Roper Scientific, Tucson, AZ) and an onstage incubator with a CO2 supplementary system (In Vivo Scientific, St. Louis, MO), was used to acquire cell and nucleus images. The image acquisition environment was kept at 10% CO2 and 37 C during experiments. The images of the red fluorescent cells and their corresponding blue fluorescent nuclei were sequentially acquired using a 20 objective lens (Nikon) at a frequency of oneminute per cycle. For CMPI determination, the cell and nuclear images were acquired for one hour and for persistence time calculation, the nuclear i mages were
28 acquired for ten hours with a 10X objective lens. The parameter settings for microscopy were 100 ms exposure time, 3 3 bin size and 25% light source power. Wound H ealing A ssay Cells were loaded to the fibronectincoated glass bottom dishes at a loading density of 1105 cells/per well with a cultureinsert (Ibidi, Verona, WI) to create the standard width of a wound gap. When the cells were grown to confluence after 16 hours, the cultureinsert was removed and the culture medium was replaced with fresh complete medium. Thereafter, the closure process of the woundlike gap was recorded every 30 minutes for 12 hours using differential interference contrast (DIC) microscopy. To examine the cell density effect on wound recovery, cells were loaded at 1 105, 3105, and 5105 cells/per well and the nuclei were stained using Hoechst 33342 to record nucleus images. Cell density was estimated by counting the number of nuclei in the field of view for each frame. Single cell T rajectory A nalysis The geometric c enters of the imaged nucleus were determined using a custom made program in the MATLAB platform (The Mathworks, Natick, MA). The nucleus trajectories were determined when the geometric centers from the same nucleus were linked by the order of the imaging acquisition time. Those nucleus trajectories further served as the source information to calculate persistence time using another custom made program in MATLAB. The persistence time of each cell type was gained by fitting the mean square displacement, < r2 >, with speed, S, using the persistent random walk equation: < r2 > = 2S2P[t P(1 et/P)] (Dunn, 1983; Othmer et al., 1988) where P and t are the persistence time and time interval, respectively.
29 Immunoblotting Cultured SKOV 3, OSE 10, OVCAR 3, MDA 231, and U 2 OS cells were separately scraped from the tissue culture dishes and the suspensions were subjected to centrifugation. Then a lysis buffer (The Cytoskeleton, Denver, CO) was added to the cell pellets to lyse the cells. The lysates were subject to centrifugation to obtain the supernatant, for which a DC Protein Assay Kit (BioRad Laboratories, Hercules, CA) was used to determine the total protein concentrations. The proteins samples were mixed with Laemmli Sample Buffer (BioRad) with 100 mM DTT (Thermo Fisher Scientific, Waltham, MA) before being boiled for 5 minutes. Then 10 g of total protein from each supernatant was individually loaded to a sample well of a 12% pol yacrylamide gel for electrophoresis. The proteins in the gel were then transferred to PVDF membranes (BioRad). The membranes were blocked by 5% nonfat milk in a TBST buffer and then separately probed with mouseraised monoclonal E cadherin (BD Biosciences ) or N cadherin (Invitrogen) antibodies, which were diluted 1,000 fold by TBST buffer, containing 2% nonfat milk. After washing with TBST buffer, these membranes were incubated with horseradish peroxidase (HRP) conjugated goat anti mouse IgG (Jackson Immun oResearch, West Grove, PA), which was in 10,000 dilution. In the blotting, GADPH was chosen as the reference protein and immunoblotted by HRP conjugated GADPH (Santa Cruz Biotechnology, Santa Cruz, CA), which was in 500 dilution. A Super Signal West Pico Kit (Thermal Scientific, Rockford, IL) was used to enhance the signal of probed proteins for 10 minutes. Finally, the protein amounts were measured by the signals and developed using Biomax XAR films (Carestream Health, Woodbridge, CT).
30 Rho GTPase s A ctivi ty A ssay Cell lysate was prepared by scraping cultured cell from tissue culture dishes with the addition of a lysis buffer (Cytoskeleton, Inc., Denver, CO). After the lysate was centrifuged, the total protein concentration of the supernatant was accessed by a DC Protein Assay kit (BioRad Laboratories, Hercules, CA). For individual Rho GTPases pull down assays, a 200g protein sample at 1 g/l concentration was mixed with 20 g PAKPBD conjugated beads or 30 g Rhotekin RBDconjugated beads (The Cytoskele ton) and further processed following the manufacturers protocol. Then 2X Laemmli Sample Buffer (Bio Rad) and 100 mM DTT (Thermo Fisher Scientific, Waltham, MA) were added to the protein samples or active Rho G TP ase s binding beads before the mixture was boiled for 5 minutes. Proteins were then separated using a 12% polyacrylamide gel and transferred to the PVDF membrane (BioRad). The membrane was blocked by a 5% nonfat milk TBST buffer and separately probed with anti mouse RhoA (Santa Cruz Biotechnology, S anta Cruz, CA), Rac1 (Millipore), and Cdc42 (Cytoskeleton, Inc.) antibodies, which were diluted by a 5% nonfat milk TBST buffer at 100 500 and 100 dilution, respectively. After TBST washing, the membrane was incubated with horseradish peroxidase (HRP) conjugated goat anti mouse IgG (Santa Cruz), in 2000 dilution by a TBST buffer containing 5% nonfat milk. A Super Signal West Pico Kit (Thermal Scientific) was applied for 10 minutes to enhance signals.
31 CHAPTER 3 RESULTS AND CONCLUSIONS The Change in C o nsistence between the C ell a nd I ts C oupled N ucleus T rajectories under Different Time Scales For a long time period, during the cell translocation, there is a higher consistence between the nucleus and cell The nucleus cannot be wipe d out f ro m the cell mig ration. T his consistence however, may be altered over a short time period. Take SKOV3, which is involved in invasive ovarian cancer, as an example (Figure 31 ). The g reen and blue lines separately represent the cell and nucleus trajectories, in which the red empty circle marked the end points of each particular time frame. High consistence between the cell and nucleus displacement is shown in the ten hour trajectories. With the monitored time reduced, h owever, there was a reduced consistence as show n in th e onehour cell and nucleus trajectories. Further, while taking tenminute time periods, significant nonconsistence appears in the cell and nucleus trajectories. This indicates that carefully study of the consistence between cell and nucleus displacement at proper time s may reveal more information about the regulation of subcellular activities during cell migration. The C onsistence b etween C ell a nd N ucleus D isplacement Suggests Different Migration Patterns Cell migration is a slow process; even a fast m igratory cell, such as MDA 231, a breast cancer cell, generally still moves at a speed of less than 1 m/min (or ~ 0.8 pixels/min in our camera setup). Hence, in this study, single cells and their coupled nucleus were alternatively recorded as a cycle at 1 frame/min temporal resolution for a onehour span to compose the cellular and coupled nuclear trajectories. Accordingly, three outstanding trajectories of NIH 3T3 fibroblasts and OSE 10 ovarian cells were
32 analyzed since they displayed visually distinguishable patterns (Figure 32). Green and blue lines separately represent the cell and nucleus trajectories, in which the red empty circles mark end points for a 60minute time period. There was a m ore straight forward displacement and highly consistence between cell and nucleus displacement in NIH 3T3. In contrast random movement and reduced consistence between cell and nucleus showed up in OSE10. This implies that the consistence between the cell and nucleus could be a potential index for cell migration abi lity, and directional or hindered movement. To further examine the correlation between cell translocation and the consistence between cell and nucleus displacement. The average of a one hour cell centroid displacement, < hCCD >, of 20 randomly picked NIH 3T3 fibroblasts, calculated as 10.92 m, was much greater than that of the same amount of randomly picked OSE 10 cells, which was 4.17 m. Highly Correlated Cell and Nucleus Reveals Longer Displacement Previous results suggest that the consistence between cell and nucleus could help determine cell migration capacity. To access this issue, X and Y components of the cell and nucleus were examined through a correlation coefficient. Here, the person product moment correlation coefficient (PPMCC) was used to j udge the correlation between these two variables ( cell and nucleus) and give a value, R. Meanwhile, the correlation between the minute cell centroid displacements, m CCDs, and their nuclear counterparts, m NCDs (NCD refer s to nuclear centroid displacement, m NCD is at a oneminute time interval), in the x and y coordinate direction were independently assessed using the correlation coefficient, denoted as Rx and Ry, respectively. Afterwards, the overall correlation between m CCDs and m NCDs could
33 be defined as Rx + Ry, or Rxy. In this regard, the mean value of Rxy, < Rxy >, for those randomly picked NIH 3T3 and OSE 10 cells were determined as 1.4175 and 0.4837, respectively. The results indicated that a cell type with a greater < Rxy > might have a greater < h CCD > as well. The relationship between the Rxy and hCCD at the level of individual cells within the same cell type was also separately examined. The Rxy value and the corresponding h CCD value of each cell were plotted against the ascending ranking o f its h CCD magnitude among either NIH 3T3 fibroblasts or OSE 10 cells (Figure 33). Based on the observation that both Rxy and hCCD have an inclined trend, a correlation between the coordination of cell and nuclear motion and cell migration capacity was clearly presented. A C ellN ucleus (C N) D isplacement C orrelation S etup Following the analysis, the correlation between a m CCD and the coupled m NCD was systematically investigated. Here, m CCD, indicated by the green arrow, is set as a reference direction for its concomitant m NCD (blue arrow), which is divided into two orthogonal components O ne is projected on the m CCD direction, denoted as m NCD// (black dash arrow) and the other is perpendicular to the m CCD direction (black dash arrow) (Figure 34 Left ). In this case, the physical interaction between the cell and its contained nucleus is through the connection of the cytoskeleton, in which the correlation between the m CCD and m NCD// reveals direct meaning. Meanwhile, a m CCD vs. m NCD// plot can b e used to describe a correlation between a m CCD and its concomitant m NCD// (C N correlation) in cell migration events ( Figure 34 Right ). Red solid dots represent the m CCDs and their coupled m NCD//s over a monitoring time of 60 minutes. Each cell was m onitored for 60 minutes with a cell sample size of 20.
34 Biophysical C ellular A ctivities o n C N C orrelation Using the concept of mesenchymal migration, the likely relation of an instantaneous cellular and nuclear displacement could be roughly described and l inked with a region of the CCD vs. NCD// plot (Figure 35). For example, when the trailing edge detaches, the rear of the cell abruptly loses its substrate adhesions and the cell centroid moves rapidly in the opposite direction of the detachment. The nucleus shows a similar centroid displacement because it is typically found at the rear of polarized cells before loss of adhesion, and is propelled forward along with the rest of the cells trailing end. Hence, the area corresponding to positive cellular and nuclear displacement (CCD < NCD//) in the CCD vs. NCD// plot was termed as Region I which reflects the trailing end detachment event of cell migration. However, abrupt forward cell movement due to the loss of trailing end substrate adhesion may not always occur in a drastic fashion. Cells maintain several simultaneous substrate adhesions of various strengths and can manage forward motility while avoiding the rapid motion reflected in Region I of the CCD vs. NCD// plot. In this instance, the nuclear displac ement is expected to be not as large as the cellular displacement, although the two could still be correlated and lie within the first quadrant of the CCD vs. NCD// plot. We termed this Region II, which reflects a forward moving cell that has some loss of trailing end adhesions, but maintains several other connections with the substrate so that the overall nuclear displacement is damped (CCD > NCD//). Sometimes, migratory cells were found to exhibit a sampling behavior that involved the development of nond irected, broad, flat lamellipodia. Cells were slightly less elongated in this period of time and any transient protrusion affected the
35 immediately apparent direction of the CCD, but this sampling activity did not significantly displace the nuclear centroid. Therefore, the area of the CCD vs. NCD// plot near to the CCD axis, where the absolute magnitude of NCD// is very small, was assigned as Region III, which reflects the state of cell sampling. Finally, since cells do not perpetually move in straight lines, but often turn or pause to reposition their nuclei, the remaining area of the CCD vs. NCD// plot in the second quadrant could be associated with turning behaviors. During a turn, the cell developed a protrusion with a significant perpendicular component to its previous trajectory, causing the greater portion of the NCD to be perpendicular to the CCD at that moment. Simultaneously, forward protrusions were in a state of retraction, and together the abrupt change in CCD direction caused an NCD// that was n egative, or opposed to the CCD. Furthermore, cells were not always moving and were frequently stationary for a period of time during tracking. However, the nucleus was actively repositioned inside the cell during this time and exhibited NCD// that appeared to counteract the CCD, but this was in reality an effect of the small transient changes in the stationary cell boundary. Together, turning behavior and a more stationary status were classified into Region IV of CCD vs. NCD// plots, which is defined as mos t of the second quadrant Here phi , donates the angle turning from the first quadrant to the second quadrant. Colorful B arcode G ives a Q uick S creening Cell migration is a continuous process, which was mixed up by different cellular activities, including protrusion, cell body translocation and detachment. To further understand the time sequence issue and hav e a quick screening tool for cellular behavior, a colorful barcode was designed. Videos of individual cells with their C N correlations predominantl y located on one specific region of the CCD vs. NCD// plot
36 should be the best resources to identify whether a region could be mapped to a genuine subcellular event of cell migration. Hence, a barcode was designed for the purpose of quickly screen ing a spe cial attribute of the C N correlations of individual cells so that the corresponding cell videos can be used to associate the biophysical trait to a region. In a barcode (an example shown in Figure 36), each C N correlation datum was presented as a bar, w hich was sequentially aligned by its occur rence time from left to right and placed at the row corresponding to the region it belongs to. The bars at the same row showed by the same color and the rows from top to the bottom presented Region I to Region IV, respectively. Specific S ubcellular A ctivity i n C ell M igration Thereafter, the barcodes of randomly picked NIH 3T3 fibroblasts were surveyed to identify special barcodes having the data mainly fall on a special region of the CCD vs. NCD// plot (Figure 3 7 first column, from left to right). Hence, the subcellular activity associated to a certain region could be elucidated from the corresponding video. As we mentioned before, cell migration is a slow process. With an increase in time interval, the outstandi ng features of these subcellular activities can be magnified. W hen the time lag was increased from 1 minute to 3 minutes (Figure 37, second column), a simple single color barcode sequence was revealed. The corresponding fluorescent images of the cell (lef t) and its coupled nucleus (right) at a 20min interval (at 1st, 21st, 41st, and 61st minute) were show n from top to bottom and the boundaries (red, green, yellow and black traces, respectively) and centroids (black circles) of these images were indicated by arrows (red: cell, blue: nucleus) show ing the cellular behavior progression (Figure 3 7 third and fourth column).
37 The cell trajectories from the stack images of the videos were further analyzed (Figure 3 7, fifth column). The trajectories of the cell (red) and nucleus (blue) for each r egion support arrows by the boundary sketches. This revealed that a cell mainly moves along a persistent direction with continual trailing edge detachments when the C N correlations are mainly located on Region I, and wit h a continuous development of rich, dynamic lamellipodia features at both the front edge and the trailing edge when the data was mainly located on Region II. F urther the C N correlations in Region III and IV correspond to a cell undergoing cell sampling and making a turn, respectively. From the arrows and the related trajectories, the biophysical meanings of the CN correlation located in each region were elucidated. C N correlation as A Cellular Parameter To further carefully investigate these images, i n a pure trailing edge detachment event we found that the amount of a NCD// is always greater than that of a coupled CCD; hence, a C N correlation generated from such an event will locate in the region between 10 40 when using the spherical coordinate (e.g., the NCD// axis is along the 0 direction and the CCD axis is along the 90 direction) to describe the direction (Figure 3 8). A red cross and a yellow star separately represent the center of the nucleus and cell. The red and yellow arrows represent the magnitude and direction of nucleus and cell. For a pure protrusion event, the nucleus usually possesses a minimal forward movement (NCD// ~ 0); while a considerable CCD is contributed by a sheet of lamellipodium and/or the extension of filopodia, which significantly is in the leading edge; therefore, C N correlation data that corresponded to this event is located in the first quadrate and close to the CCD axis (~ 90 ) (Figure 3 9). When cells perform pure
38 membrane ruffling, the quick fluctuation of the CC D value and direction makes the small NCD// swing between positive and negative values and their C N correlations cover a small semicircle region with the center of diameter at the origin of the CCD vs. NCD// plot (Figure 310) Always, cell migration is a continuous process and frequently more than one dynamic feature is present in a cell at a certain time (Figure 311). For example, a cell could have a predominant trailing edge detachment event occur simultaneously with a mild protrusion around the leading edge. In this case, the extra event will enhance or reduce the CCD value, respectively, but wont change its direction from a pure trailing edge detachment accompanied with cell body translocation; also the NCD// will retain around the same value. Hence, the C N correlation presenting this mix ed event will shift upward (with a same direction protrusion), up leftward (with a large side protrusion), and leftward (with small side protrusion) (Figure 311, Right). In other cases, for example if the predomina nt feature during a time interval is the formation of cell protrusion, a minor trailing edge detachment will increase the CCD value. More importantly, it will add a nonnegligible value to the NCD// because of the cell body translocation, which is the cons equence of the detachment event. As a result, the C N correlation will shift the location up rightward from that in the pure protrusion event (Figure 311, left). Unlike a protrusion event, a sole significant retraction event almost always wont happen to a mesenchymal cell. As we observed from the large amount of samples the retraction events are most frequently accomplished with a side protrusion; whereas the overall CCD most of the time will be redirected to an obtuse angle compared to the
39 NCD because the nucleus still carries the inertia to move toward the original direction. Hence a considerable size of negative NCD// is yielded accordingly to the new CCD direction. A less frequent case is when a predominant retraction event occurs coincidently with a trailing edge detachment event, in which the amount of CCD would be smaller than that of the retraction event alone. Since the NCD// is negative, the location of the C N correlation is shifted toward a greater angle than where a pure retraction would be. In both cases, the C N correlations with a predominant retraction event are located in the second quadrate ( > 90 ) of the CCD vs. NCD// plot (Figure 3 12). Cell Migration Potential Index (CMPI) The time that a cell spends at each stage of movement would dictate how many points would lie in each corresponding region of the CCD vs. NCD// plot; thus, C N correlations directly relate to the cells migration behavior. Still, if migration is mainly concerned with effective cell displacement, then migration events like sampling or turning might only obscure the true potential of a cell to effectively move. From that perspective, a cells migration potential might be better measured by focusing on events that effectively displace the cell. Since the C N correlati ons located in Region I and II positively contribute to the RCN of a migratory cell, and since a greater < h CCD > is associated with a larger < RCN >, the occurrences (denoted as Occur ) and the average values of m CCDs (< m CCD >) for both regions were us ed to evaluate the cell migration potential using a weighted sum (Figure 3 13) Thus, only the portions of the cell trajectory that directly contributes to the overall cellular displacement are used to create the cell migration potential index (CMPI), defi ned as:
40 CMPI = i = Region I Region II Occuri < m CCD >i Here, we suggest a cell with a greater CMPI should migrate faster than a cell having a smaller CMPI. Three fibroblasts, NIH 3T3, Swiss 3T3, HF, and five epithelial cell lines, including one normal, OSE 10, and four cancer cell lines, SKOV3, OVCAR 3, MDA 231, U 2 OS, were used to examine this approach (Table 31). Th e results indicate d that MDA 231 had the highest migration potential ability follow ed by NIH 3T3, SKOV3, OSE 10, Swiss 3T3, U 2 OS, OVCAR 3, and HF. Cell Trajectories Was Set as a Gold Standard of Cell Migration Cell migration potential predicts the migratory ability of a cell type during a reasonable time frame. Currently the most acceptable method to evaluate such potential for single cells is persistence time, in which the extracted results need to be ensured as being much shorter than the monitored time for the probed cell trajectories. Usually a persistence time study utilizes nuclear centroid trajectories that are documented at around a 10hour length. With this time frame, the correlations between a 10hour CCD (10 h CCD) and a 10hour NCD (10 h NCD) for any cell type studied here are generally greater than 0.97 (data not shown). Hence, the average values of the 10h NCDs (or < 10 h NCD >) were chosen here as the gold standard for the comparison of CMPI and persistence time. The 10 hour nuclear trajectories (10hNCD) of at least 10 cells from each of the 8 probed cell types exhibited distinct migration patterns (Figure 314 ). Each c olor trajectory represents a 10hour cell trajectory where a red empty color represents the end point of each cell trajectory. A bl ack solid circle represents the average of a 10 hour cell displacement. The 10hour displacement distribution of each cell li ne is shown
41 below each trajectory pattern. The black solid circle represents the 10hour displacement of each cell, where the bl ack and red dash ed line separately represent the mean and the standard value of the 10hour displacement respectively For all cell types, most of the cells traveled a distance within one standard deviation of the < 10hNCD >, but some do not, this suggests that the 10hNCD is not a normally distributed parameter. Of the fibroblasts, NIH 3T3 fibroblasts demonstrated straight er tr ajectories and were one of the most migratory cell types. Human foreskin fibroblasts, in contrast, demonstrated trajectories that remained near to the initial position. Meanwhile, the trajectories of Swiss 3T3 fibroblasts were usually not as straight as that of NIH 3T3 fibroblasts but still were more extensive than that of human foreskin fibroblasts. Similar differences in migration patterns were observed in the other 5 cell types probed in this study as well. Invasive MDA 231, SKOV 3 and U 2OS cancer cell s migrated more directionally while normal OSE 10 cells and noninvasive OVCAR3 cancer cells appeared to have hindered mobility and did not demonstrate significant migration away from their initial positions. Persistence Time Cannot Give a Stable Estimat ion on Cell Migration Ability Among the 8 cell types, NIH 3T3 fibroblasts possessed the greatest < 10hNCD >, followed by MDA 231, SKOV 3, OSE 10, Swiss 3T3, U 2 OS, OVCAR 3, and human foreskin fibroblasts, which had the lowest < 10hNCD > value. The persi stence time was gained from persistent random walk theory with mean square displacement and speed fit. The < 10hNCD >, represented by the solid blue bar, of each cell type was plotted with an averaged persistence time, represented by the empty circle (Fig ure 31 5 ). The correlation between < 10hNCD > and 10 hour persistence time had a correlation coefficient value (R value) of 0.51. However, if the persistence time was extracted from
42 the same trajectories in a 9hour period instead of 10hour, the value could alter significantly for some cell types (such as SKOV 3 cells and Swiss 3T3 and human foreskins fibroblasts). In addition, the R value between the < 10hNCD > and the 9hour persistence time also dropped to 0.39, indicating the instability of an approach using persistence time to evaluate cell migration. CMPI, a Better Estimation on Mesenchymal Cell Migration The c ell migration potential ability of eight cell lines were suggest ed by CMPI (Table 31). In order to evaluate its ability, the < 10hNCD > o f the various cell lines were plotted against its CMPI. T he linear fitting reveal ed a slightly stronger correlation value of 0.59 for all cell types. With further investigati on of cell migration patterns the correlation value was seen as increasing up to 0.87 if only considering the cell types that predominantly undergo mesenchymal migration (noninvasive OVCAR3 cancerous cell types and normal epithelial OSE 10 have been excluded) (Figure 31 6 ). Hence, CMPI grants an improved estimation of a cells 10hou r migration ability over 10hour persistence times Importantly, the CMPI only requires one hour of cell tracking. This provides a time advantage and also improves the probability that a cell would not migrate out of the frame during tracking, drastically increasing the sample size that c an be acquired per hour. The Stability of CMPI To further investigate the variance and stability of CMPI, a random sample was picked up from each pool of the cell lines. For variance estimation, the average of CMPI was determined by choosing a random pickup sample size twenty times. Further, for stability, the sample size was increased up to 50. The CMPI of the topthree fastest migratory cell types (MDA 231, NIH 3T3 and SKOV 3) and one slow migratory cell type
43 (OVCAR 3) shows a diminishing standard deviation as the sample size is increased with the OVCAR 3 demonstrating the quickest convergence. This suggests that reliable CMPI values could be attained quickly and easily to rapidly assess the migration potential of a cell ty pe (Figure 31 7 ). In conclusion, CMPI can quantitatively evaluate the change in cell migration without the interference of cell cell interactions. In contrast to persistence time, which simply provides a quantity to distinguish cell migration without integ rating direct insight about the cell migration process, the C N correlation defines a solid foundation to characterize a cell migration pattern. A detailed cell migration mechanism can be described by C N correlations, which can be further linked to the mo lecular mechanisms that contributed to the cellular activities such as lamellipodia formation, actomyosin contraction and cell polarity. This will significantly promote the current effort in taking a systems approach to biomedicine and to appropriately describ ing the intrinsic cell migration potential for different cell types. Collective Approach, Wound Closure Assay, Used to Judge Cell Migration and Consisten cy with CMPI To evaluate whether the CMPI is an effective index to describe a cells migration pot ential, five different nonfibroblast cell types, MDA 231, OSE 10, OVCAR 3, SKOV 3, and U 2 OS cells, were subjected to the woundlike gap closure assay (Figure 3 18, top). Cells were loaded with optimal density to reach the confluent monolayer cell after 12 hour in two wells of insert, which has a 500m gap between two wells. Ten hours after a 500 m gap was created in a confluent cell monolayer, the order of the average width of the remaining gap for an individual cell type from the smallest to the large st was MDA 231, SKOV 3, OSE 10, U 2 OS, and OVCAR 3 (Figure 318, bottom ). The
44 recovery percentage was calculated by the difference between the original gap area and remaining area divided by the original gap area. This recovery order was consistent to tha t of CMPI, calculated from the C N correlation data of each cell type over a onehour period (Figure 31 9 ). Further, the gap recovery rates during the gap closure process were not constant, as revealed from the gap recovery plot. When the slopes of each g ap recovery curve were calculated by its 2th, 6th and 10th hours, and normalized by the initial slope, it showed that the evolution of the slopes could be used to classify the cell types into two groups: MDA 231, OSE 10 and U 2 OS cell types, which had their slopes decline; and SKOV 3 and OVCAR 3, which had their slopes incline (Figure 320). Cell Cell Interaction in Wound Closure approach Can Be Illustrated by CMPI The most likely reason for the change in the gap recovery rates should be from interference of the cell cell interactions. On the one hand, cells in a crowd could push neighboring cells toward the void space, therefore enhancing the cell migration rate. On the other hand, the activities of cadherin family proteins could produce different degrees of interaction strength in hinder ing cell migration (Kim et al., 2005; Panorchan et al., 2006) To better illustrate the effects of cell cell interactions, the percentages of gap recovery of the 5 cell types at dif ferent time extents were plotted against their CMPI (Figure 3 21). Curve fitting revealed that the regressions were consistently improved by a longer time extent, in which the influences of cell cell interactions were gradually downplayed. Furthermore, dur ing the woundlike gap closure process, MDA 231 cells moved toward the void via a single cell format, while the OVCAR 3 cells moved as a group (Figure 322). The other three cell types mainly moved in a similar pattern as the
45 OVCAR 3 cells but occasionall y displayed singl e cell migration These indicate the cell cell interaction play s a critical role in wound closure. To further examine the cell cell contact issue, western blotting using antibodies against the cadherin family proteins, E and N cadherin, could provide an explanation for the difference in the migration patterns among the five cell types. The results suggested that all cell types, except MDA 231, contain cadherin proteins that can initiate cellcell interactions (Figure 323). Also, in order to demonstrate that cell cell interactions through the cadherin proteins could conceal the intrinsic cell migration potential, a woundlike gap closure experiment was applied again to a stable MDA 231 cell type derivative, in which E cadherin was recombi nantly introduced into the MDA 231 cells. The discrepancy between the MDA 231 derivative and its original counterpart for the gap closure at the same time frame clearly addressed the effect of cell cell interactions on the gap closure process (Figure 324) The recovery percentage at 10 hours dropped to around 50% due to the introduction of E cadherin. Cell Density, a nother Potential Issue for Wound Closure Approach T he cell density effect on woundlike gap closure was also evaluated. All different cell de nsities (from high, medium to low density) reached a confluent monolayer cell before the insert was removed. Two ovarian cell lines were chosen, OSE10, the normal epithelial ovarian cell, and OVCAR 3, the benign ovarian cancer. The results showed that OVCA R 3 cells in a more dense state possess a better woundlike gap closure effect than the otherwise faster moving OSE 10 cells (Figure 32 5, left ). The nucleus was dyed with H33342 and fluorescent images of the nucleus was used to represent the
46 condition of cell dens ity ( Figure 32 5, right). These results suggest that the wound closure approach should be processed with well density control. Taken together, the involvement of cell cell interactions and cell density in a woundlike gap recovery assay can complicate the interpretation of the experimental results and obscure a simple explanation. In contrast, CMPI provides a compatible and much clearer cell migration index that is assessed in a single cell context without the interference of cell cell interactio ns. Hence, CMPI can fully replace the woundlike gap recovery assay to probe the change in the intrinsic cell migration capacities between a cell type and its derivative. CMPI Has Distinct Advantages o ver Existing Migration Evaluations Taken together, the involvement of cell cell interactions and the physical crowding effect in a woundlike gap recovery assay can complicate the interpretation of the experimental results. Meanwhile, persistence time provides a quantity to distinguish cell migration without integrating direct insight s about the cell migration process. In contrast, CMPI is assessed in a single cell context without cell cell contact to deliver a compatible cell migration index with much clearer migration details. Hence, CMPI can be used independently and in a complementary manner with the woundlike gap recovery assay to obtain the full spectrum of migration information, including intrinsic cell migration potential and the effect of cell cell interactions on collective migration for various cel l types. The C N correlation during cell migration also defines a solid foundation to characterize cell migration patterns thru a detailed descri ption of the cell migration mechanism, which can be further linked to the molecular mechanisms that contributed to the cellular activities such as lamellipodia formation, actomyosin contraction and cell polarity. Hence, this will significantly promote
47 the current effort in taking a systems approach to biomedicine t hat appropriately describes the intrinsic cell mig ration potential for different cell types. A Cellular Parameter for Distinguishing Cells C o existing cellular dynamic features represented by occurrence and timesequence comprises the cell migration patterns For example, a ballistic cell migration is as sembled with a sequence of continuous leading edge protrusions and trailing edge detachments, working harmonically to allow a cell body to move forward. Thus, the C N correlation profile of a ballistic cell migration will have C N correlation data that is mainly distributed in the first quadrate of the CCD vs. NCD // p lot. Cell sampling is composed of membrane ruffling, pure protrusion and pure retraction. Hence, C N correlation data for cell sampling is primarily distributed along the CCD axis of the CCD vs NCD// plot During a cell turning event, the sequence evolves from a lamellipodia rich, directional migration to predominant cell sampling events, and ends up with another directional migration in a different direction. During this process, retraction wi ll be a predominant event in the mixture of dynamic features at certain time so that the occurrence of the C N correlations at the time will be located in the second quadrat away from the CCD axis. Hence, a turning event always contains a sequence of C N c orrelations switching back and forth between the first and second quadrate, but away from the CCD axis Taken together, the distinct distribution pattern of C N correlation data of a probed cell type can be mapped with its unique migration pattern once am p le amounts of C N correlations are available We further tested whether the C N correlation profile can serve as a cellular parameter for a cell type. First, two independent groups of NIH 3T3 fibroblasts, 50 cells each, were probed for one hour at the sam e culture condition to construct their C N
48 correlation profiles. Since the C N correlations for similar cellular dynamic features are located around the same conical region classified by similar angles in the CCD vs. NCD// plot, the Student T tests were subjected to these two profiles to compare the occurrence (Figure 326 Top) and the CCD quantities (Figure 326, Bottom and Insert) by different distribution angles. The results suggested that these two profiles are highly consistent We further constructed the C N correlation profiles for 8 different cell types (3 fibroblasts: NIH 3T3, Swiss 3T3 and human foreskin fibroblast 3 mesenchymal invasive cancer cells: MDA 231, U2 OS and SKOV3, 1 epithelial noninvasive cancer cell: OVCAR 3, and 1 normal epitheli al cell: OSE 10) and the occurrences of these C N correlations against the distribution angles were compared by pairs (Table 32). The results suggested that eight pairs of those cells have s similar angle distribution for their C N correlations. Hence, th e quantities of the CCDs of the C N correlations were further crosscompared by their angular groups using student T tests, which showed that, even two cell types might have the same angular distribution, and the magnitude of the CCD in those C N correlati ons were still different (Figure 327). It is noteworthy that the main dynamic features of Swiss 3T3 fibroblasts, which are protrusion (red crosses), and OVCAR 3 cells, which are membrane ruffling (blue crosses), can be easily distinguished by the averaged CCDs along 90 as discussed in the previous section ( Fig ure 328). Cellular Activities Disrupted b y Drug Treatment Alter t he C N Correlation To support this concept, we separately applied Y 27632, which inhibits ROCK and further prevents actomyosin contraction, and nocodazole, which disrupt s microtubules, and used a RNA interference technique on knockdown Arp2/3, which is
49 essential for lamellipodial formation, in NIH 3T3 fibroblasts. The effects of these on NIH 3T3 fibroblasts were studied comprehensively H ence, we examined whether the changes in the dynamic features of NIH 3T3 fibroblasts under these conditions can be reflected i n their C N correlation profiles. For each condition, twenty cells were documented to construct the corresponding CCD vs. NCD// plots. When the CCD vs. NCD// plot of the NIH 3T3 fibroblasts under the presence of Y 27632 to inhibit the actomyosin contraction was compared to that of the normal NIH 3T3 fibroblasts (Figure 329), the C N correlations locating on the area where the dyna mic features associated to the actomyosin contraction, such as the trailing edge detachment or large angle cell turning, were completely depleted from their original positions. In addition, the magnitudes of the CCD of the C N correlations mapped to the pr otrusion and retraction regions was decreased, suggesting that the leading edge protrusive event after the contraction force is diminished, and tested at a one minute interval, is not as aggressive as the normal condition. It was reported that the inhibiti on of ROCK causes the fibroblasts to increase membrane ruffling in the active lamella (Omelchenko et al., 2002) but the focal adhesions and stress fibers are unable to form to stabilize the protrusion (Rottner et al., 1999; Uehata et al., 1997) Hence the small magnitudes of the CCD along the CCD axis were in agreement with the literature and mapped well with the ruffling events. Further, the knockdown of the expression level of Arp2/3 i mpeded lamellipodial formation but still allow ed filopodial formation, which is critical for lamellipodial formation, and the knockdown of the expression level of Arp2/3 in the NIH 3T3 fibroblasts will demolish lamellipodial formation. A comparison of the locations of the C -
50 N correlations in the CCD vs. NCD// plot between the normal cells and the cells with the Arp2/3 knockdown indicated that the C N correlation data disappears from the protrusion and retraction region (Figure 330); hence, it suggested ag ain that the distribution profile of the C N correlations in the CCD vs. NCD// plot can faithfully describe the dynamic activities of cells (Suraneni et al., 2012) Further, after nocodazole disrupted the microtubulecytoskeleton in the NIH 3T3 fibroblasts, the occurrence of the C N correlations declined in the locations associated to the trailing edge detachment event and increased in the locations associated to the protrusion and retraction events ; also, the CCD magnitudes of the C N correlations associated to the protrusion and retraction were elevated (Figure 331). This observation strongly supported the conclusion of the study suggesting that Calpain6, a microtubulestabilizing protein, is released from the surface of a microtubule and interact s with guanine nucleotide exchange factor, GEF H1, to activate Rac1 upon microtubule depolymerization at the leading edge of NIH 3T3 fibroblasts (Tonami et al., 2011) The Contribution of This Work Cell migration has attracted interest since 1970; however, the current measurement approach, which is ir relevant to biological aspects, is incapable of qualitatively analyz ing these moving patterns and can only quantitativ ely assess them Furthermore, the role played by the nucleus in regulating cell migration is still obscured since the importance of the nucleus in actin remolding has been emphasized. In this work, a relationship was successfully identified between physical measurement during cell migration and the underlying signaling of subcellular activities which offer s a multi scale perspective on cell migration. This approach avoids
51 overlooking the essence of subcellular activities during cell migration since thi s in formation has been overlooked by the other two quantitative approaches for decades. Through this bridging, a straightforward analysis can be achieved that connect s physiological activities and molecular reactions. Further more a new quantitative approach for estimating cell migration capacity was extracted to enrich the current approaches. A substantial amount of cell migration capacity can be deciphered in minutes instead of hours and at a high level of quality This quantitative information emphasizing the essence of these cellular characteristics during cell migration promotes our recognition of cell migration. Moreover, through this correlation, unique migration pattern assembled by subcellular activities which represented different cell lines can be u sed to distinguish between cells, especially in cancer cell research. It is difficult to directly access molecular changes in living cells and to clarify the simple role of protein under complex signaling crosstalk; however, this work may provide some clues to estimate the protein activity level, because these characteristic are directly contributed from their upstream molecular mechanism. Furthermore, the regulatory nodal protein in signaling pathways can be elucidated. This study sets forth a framework of top down integrative biology that only monitors cell and nucleus movements to understand the status of the controlling pathways. Through such integrative work a new era in cell biology can be expected.
52 Figure 31. The consistence between cell and nucleus trajectories at different time period. A highly correlation exists between 10 hour cell and nucleus trajectory (Left panel). With the time scale reduced to 10 minutes, the consistence between cell and nucleus is significant decreased (Right panel) 10m 10 hour trajectories Cell Nucleus 5m 1 hour trajectories 2.5m 10 minutes trajectories
53 Figure 3 2 Consistence of cell and nucleus trajectories suggests longer cell displacement. NIH 3T3 revealing higher cell and nucleus consistence has longer displacements. Instead, OSE 10 having less consistent in cell and nucleus results in short translocation. Three representing cells were used for NIH 3T3 and OSE 10 separately. Average displacement of each cell line was calculated from 20 random pickup samples. -20 0 20 -20 -10 0 10 20 NIH 3T3 X (m) Y (m) < h CCD > ~ 11 m -20 0 20 -20 -10 0 10 20 OSE 10 X (m) Y (m) < h CCD > ~ 4 m
54 h CCD (m) 0 5 10 15 20 0 20 40 0 5 10 15 20 -1 0 1 2 < Rxy> = 1.4175RxySample (N)NIH 3T3Sample (N) 0 5 10 15 20 0 5 10 15 0 5 10 15 20 -1 0 1 2 < Rxy> = 0.4837h CCD (m) RxyOSE 10 Figure 33 Highly correlated cell and nucleus reveals longer displacement. The correlation of cell and nucleus in X and Y axes was represented as Rxy. Comparison of Rxy and its cell displacement reveals that higher correlation has longer displacement in NIH 3T3 and OSE 10. From 20 random pickup cell s amples, NIH 3T3 has larger average correlation than OSE 10. The corr elation coefficient (R) is determined by PPMCC.
55 Figure 34 Correlation between cell and nucleus displacement. A nuclear centroid displacement (NCD) is divided into two orthogonal components, NCD//, which is paralleled to the direction of cellular centroid displacement (CCD), and NCD which is perpendicular to the direction of CCD. Meanwhile, the CCD and it corresponding NCD// can be used to construct a correlation (C N correlation) through a Cartesian correlation system to describing cell migration events (Right). Red solid dots represent CCD and their coupled NCD// over a monitoring time of 60 minutes. C ell was monitored for 60 minutes with a cell sample size of 20. -2 0 2 0 1 2 3 CCD NCD NCD NCD// CCD: cell centroid displacement NCD: nucleus centroid displacement NCD // (m) CCD (m)
56 M 10 = ( / / ) NCD//CCD NCD//> CCD 0 < NCD//< CCD NCD//~ 0 NCD//< 0 Figure 35 The relations between the CCD and NCD// can be built in a NCD// vs. CCD plot. When is defined as the angle from the x coordinate to a line in the counterclockwise direction, four regions can be defined: Regi on I: 0 < Region II: 45 < Region III: 85 < and Region IV: 95 < For Regions I and II, it means the NCD are contributed to the whole cell displacement; for Region III, NCD ~ 0 and the cells are sampling; and for Region IV, nucleus motion is against cell motion.
57 Figure 36 Colorful barcode representing migration behaviors. A barcode was created to present a group of CN correlation data generated from monitoring a cell for one hour. Individual datu m in the group was presented as a bar, aligned following both its occurrence order in the group, from left to right, and its location in the CCD vs. NCD// plot, colored as red, yellow, blue, and green for Region I, II, III and IV from top to bottom, respec tively. RegionTime III II I IV
58 Cell: Red line Nucleus: Blue line Cell : Red arrow Nucleus: Blue arrow X (m)Y (m) X (m)Y (m) X (m)Y (m) X (m)Y (m) Nucleus Cell Fluorescent image Trajectories Barcode = 3min = 1min Region II. Hindered nucleus movement Region IV. Cell confined movement Region I. Cell body translocation Region III. Cell sampling Figure 37 Colorful barcodes with their related cell migration behaviors. The analysis results of the four special cases were presented by individual panels, listed from Region I to IV to correspond to each barcode. These special ca ses were used to identify the unique cell migration pattern associated to the four regions. When the time interval was increased to 3 min to construct those barcodes, the predominant effects were further signified compared to 1 min interval time.The fluorescent images of the cell (left) and its nucleus (right) at a 20min interval (at 1st, 21st, 41st, and 61st minute) were showed from top to bottom. Scale bar is 10 m. The progressions of the boundaries (red, green, yellow and black traces, respectively) and centroids (black circles) of these images were indicated by arrows (red: cell, blue: nucleus). The trajectories of cell (red) and nucleus (blue) for each Region support arrows aside by the boundary sketches. From the arrows and its related trajectories, the biophysical meanings of the CN correlation located in each region were elucidated.
59 Figure 38. Trailing end detachment on C N correlation. Through screening images, pure detachment events can be seen from the continuous images and it corresponding cell and nucleus displacement can be mapped on C N correlation profile between 10 and 40 Red and Blue crosses separately represent the left and right images. The bottom sketch was used to describe the event, where a red cross and a yellow star separately represent the center of the nucleus and cell. The red and yellow arrows represent the magnitude and direction of nucleus and cell. Footnote in images represent the time series.
60 Figure 39. Protrusion on C N correlation Pure protrusion events can be seen from the longer continuous images and their corresponding cell and nucleus displacements can be mapped on C N correlation profile between 80 and 100. Red and Blue cross es separately represent the top an d bottom images. The bottom sketch was used to describe the event, where a red cross and a yellow star separately represent the center of the nucleus and cell. The red and yellow arrows represent the magnitude and direction of nucleus and cell. Footnote in images represent the time series.
61 Figure 310. Membrane ruffling on C N correlation. Ruffling events shows high frequency of dynamic protrusion and detachment at peripheral edge of cell with stationary nucleus movement. It corr esponding cell and nucleus displacements can be mapped on C N correlation profile with smaller magnitude but all range from 0 and 18 0 The bottom sketch was used to describe the event, where a red cross and a yellow star separately represent the center o f the nucleus and cell. The red and yellow arrows represent the magnitude and direction of nucleus and cell. Footnote in images represent the time series.
62 Figure 31 1 Complexity subcellular activities on C N correlation A forw ard movement usually includes a mixture of detachment and protrusion activities. Left and right panel separately shows the protrusion and detachment dominating migration process. Their corresponding cell and nucleus displacements mapping on C N correlation separately represented as blue and red cross es and between 80 and 100. The bottom sketch was used to describe the event, where a red cross and a yellow star separately represent the center of the nucleus and cell. The red and yellow arrows represent the magnitude and direction of nucleus and cell. Footnote in images represent the time series.
63 Figure 31 2 Retraction on C N correlation Retraction is not a solely effect and always happens a fter protrusion. Three image series show a retraction effect after protrusion and their corresponding cell and nucleus displacements mapping on C N correlation separately represented as blue (Left top) red (Right top) and green (Left bottom) crosses and between 110 and 13 0 The bottom sketch w as used to describe the event, where a red cross and a yellow star separately represent the center of the nucleus and cell. The red and yellow arrows represent the magnitude and direction of nucleus and cell. Footnote in images represent the time series.
64 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 Region I Region 2Occur (X100%) X X+ CMPI < m CCD> (m) Figure 31 3 Sketch of calculation of CMPI. CMPI was determined by summary of occurrence (Occur) plus average cell displacement (
65 SKOV 3 (46.60; 16) Swiss 3T3 (36.53; 10) MDA 231 (54.31; 26) (< 10 h NCD >; Sample size) NIH 3T3 (58.17; 21) OSE 10 (7.65; 12) U 2 OS (23.48; 17) OVCAR 3 (11.21; 12) HF (27.34; 15) 20 m Figure 314. Trajectory as gold standard for cell migration. Displacements started at the center of the reference circle (yellow), the radius of which represents the average 10 hour displacement. The horizontal black and dotted red lines indicate the < 10hNCD > and one standard deviation, respectively, for all the cells, and the points plotted correspond to the individual 10hNCD values of each cell. The values of < 1 0hNCD > and sample size are in d i cated in the parenthesis, respect ively, following by the name of each cell type.
66 Figure 315. The correlation between < 10hNCD > and persistence time. < 10h NCD > of each cell type (bars) is compared against the extracted persistence time from 9 (green circles) and 10 (red circles) hours of trajectories. The correlation shows 0.39 and 0.51 separately. The correlation coefficient (R) is determined by PPMC.
67 Figure 316. The correlation between < 10hNCD > and CMPI. The correlation coefficient (R valu e) between < 10hNCD > and CMPI improves from 0.59 to 0.87 when only considering cell types that mainly rely on mesenchymal mode to migrate. The correlation coefficient (R) is determined by PPMC.
68 Figure 31 7 The stability of CMPI. CMPI variance reduces as sample size increases for MDA 231, NIH 3T3, SKOV 3 and OVCAR 3, supporting that CMPI stabilizes when sample size increase to at least 20 cell. Error bar represents the S TD .
69 Figure 318. Wound recovery image and recovery rate of epithelial cells. The woundlike gap closure processes of 5 different epithelial cells were analyzed by their videos (examples on left panels) and the results were presented as the percentage of gap recovery over a 10hour period (right panel). Error bar represents the STD. U 2 OS SKOV 3 OVCAR 3 MDA 231 OSE10 T= 0 hours T= 10 hours T= 5 hours 0 5 10 0 25 50 75 100 Recovery (%)HoursMDA 231SKOV 3 OSE 10 U 2 OS OVCAR 3 Error bar: STD
70 Figure 319. Well correlation between CMPI and recovery rate. Well correlation between CMPI and recovery rate. The cell migration potentials of the 5 epithelial cells directly determi ned by their CMPI were highly consistent to the results gained from 10 hours after the woundlike gap processes. Error bar represents the STD.
71 Figure 320. Recovery rate at different time point suggests cell cell contact effect. Eac h of the woundlike gap recovery process of the 5 epithelial cells was not at constant rate. The rates of 3 cell types were inclined, while those of 2 cell types were declined. 0 0.5 1 1.5 2 Vi /V2 2 hour 6 hour 10 hour
72 Figure 321. Correlation between recovery rate and CMPI at different time points. The linear correlation curves of the woundlike gap recovery percentages at 2, 6 and 10 hours vs. CMPI values for the 5 probed epithelial cells (left panel). The linear regressions were 0.88, 0.95 and 0.99, respectively (rig ht panels). The correlation coefficient (R) is determined by PPMC. 0 10 20 30 40 50 0 25 50 75 100 CMPIRecovery (%)OVCAR 3 U 2 OS SKOV 3 OSE 10 MDA 231 T = 10 hours T = 6 hours T = 2 hours
73 OVCAR 3 MDA231 Figure 322. Different migration pattern during wound disclosure. MDA 231 moves toward single instead of collective movement, such as OVCAR 3.
74 N cadherin E cadherin GADPH Figure 323. Western blotting supports the cell cell contact issue. The western blotting of the 5 cell types against E and N cadherin revealed the expression levels of these proteins in the cell samples. The same blotting used housekeeping protein, GADPH, as a reference protein.
75 25 50 75 100 Recovery (%)10thhourError bar: STD Figure 324. The introduction of E cadherin into MDA 231 cells hindered the woundlike gap recovery process. After E cadherin delivered into MDA 231, MDA 231 (E cad) was culturing in insert under general process (see material and method). MDA 231 with E cad h erin delivery shows less recovery rate comparing with wild type (around 50%) Error bar represents the STD.
76 Figure 32 5 Cell density plays roles in wound healing assay altering migration ability. Wound like gap recovery at different initial c ell densit ies alters the apparent migration ability of OVCAR 3 and OSE 10 cells ( L eft) Cell density was estimated through staining and counting the number of nuclei in the initial frame. All scale bar s represent 200 m and error bar s represent standard deviation (Right) LD, MD and HD separately represent different seeding densities.
77 Figure 326. C N correlation suggests a highly correlated between two NIH 3T3 groups The occurrence profile ( T op) between two NIH 3T3 groups are correlated with a p value around 0.37 with 1 degree spanning. The magnitude of cell displacement ( < m CCD profile >, B ottom)) also gives a good correlated result between two NIH 3T3 groups with 5 degree spanning The corresponding student t test in every 5 degree spanning was shown in the right top corner. The red dash line represents the p value is 0.05.
7 8 Figure 32 7 Similar angular distrib ution but significant different in magnitude of the CCD in CN correlation. The above paired cells cannot not be distinguished by occurrence profile but show a significant difference through cell displacement profile ( < m CCD> ). This suggests that C N corre lation has the cell distinguishing ability as one of cellular parameter.
79 Figure 328. The distinguishing between protrusion and ruffling in Swiss 3T3 and OVCAR 3. The significant difference between Swiss 3T3 and OVCAR 3 is the magni tude along the Y axes on the C N correlation. From the continuous images of Swiss 3T3 ( T op) and OVCAR 3 ( B ottom), Swiss 3T3 has large magnitude displacement at the peripheral edge of cell as protrusion activity. In contract, OVCAR 3 has minor peripheral edge displacement but more dynamic activities there.
80 Figure 329. Cellular activities disrupted by Y27362 alter the C N correlation To investigate if the C N correlation can respond to the change in the subcellular activities, Y 27632, the ROCK inhibitor was applied. From the C N correlation mapping, Y27632 treated NIH 3T3 (green dots) shows a decrease density between 10 and 80 comparing with wild type (black dots) ( L eft panel) Further, it shows a significant decrease in magnit ude between 10 and 90, which may mainly due to the loss of focal adhesion.
81 Figure 330. Cellular activities disrupted by Nocodazole alter the C N correlation. Also, Nocodazole, the microtubule inhibitor was further applied to examine the correlation. From the C N correlation, it shows no significant difference between wild type (black dots) and Nocpdazole applied NIH 3T3 (red dots). However it shows a n increase in magnitude between 8 0 and 10 0 which may due to the disruption of microtubule turnover
82 Figure 331. Cellular activities disrupted by Arp2/3 shRNA alter the C N correlation. With Arp 2/3 shRNA delivery, the C N correlation shows a reduced density (blue dots) between 40 and 80 comparing to wild type (black dots) (Left panel) Further it shows a de crease in magnitude between 80 and 100, which may due to the disruption of lamellipodia (Right panel)
83 Table 3 1 The composition of occurrence and average cell displacement in region I and II of different cell lines and their related CMPI. MDA 231 NIH 3T3 SKOV 3 OSE 10 Swiss 3T3 U 2 OS OVCAR 3 HF OccurRegion I (%) 24.05 28.93 18.03 17.93 18.50 25.54 17.60 23.52 < m CCD >Region I (m) 0.57 0.32 0.28 0.27 0.19 0.17 0.21 0.14 OccurRegion II (%) 35.81 34.63 38.56 38.60 32.50 35.27 32.40 31.67 < m CCD >Region II (m) 0.76 0.59 0.56 0.56 0.40 0.34 0.38 0.24 CMPI 0.41 0.30 0.27 0.27 0.17 0.16 0.16 0.11
84 Table 3 2 Student T test between cell lines in occurrence profi le P value NIH 3T3 OVCAR 3 SKOV 3 MDA 231 U 2 OS Swiss 3T3 HF OSE 10 NIH 3T3 1 < 0.0001 0.0134 0.5931 0.5257 < 0.0001 0.3381 < 0.0001 OVAR 3 1 < 0.0001 < 0.0001 < 0.0001 0.3990 < 0.0001 0.0052 SKOV 3 1 0.003 0 0.0026 < 0.0001 0.1658 0.0701 MDA 231 1 0.9126 < 0.0001 0.1459 < 0.0001 U 2 OS 1 < 0.0001 0.1256 < 0.0001 Swiss 3T3 1 < 0.0001 0.0163 HF 1 0.0023 OSE 10 1
85 CHAPTER 4 FUTURE WORK A Potential Approach for miRNA Assay MicroRNAs (or miRNAs) play an important role as a master regulator to the cell. Each type of miRNA utilizes up to 8 of its total ~22 nucleic acids to recognize and bind to the complementary sequence appearing in messenger RNAs (mRNAs) for post transcriptional regulation. T he working mechanism of miRNAs occur at the post transcriptional level, and the proteins, modulated by a specific type of miRNA via such a mechanism, do not necessarily possess similar functions or coexist at the same signaling pathway. Hence, miRNAs can simultaneously alter the activi ties of several different signaling pathways that have no known phy sical connection. Currently, at least 800 miRNAs have been identified in the human genome (Bentwich et al., 2005) For example, miR 10b has been reported to promote cancer metastasis, in which cancer cells undergo epi thelial mesenchymal transition (EMT) and migrate to a remote site (Gee et al., 2008) Hence, proteins encoded from the mRNAs that are the post transcriptional targets of miR 10b have been heavily studied in an attempt to understand their effects on promoting metastasis. Two of those proteins, HOXD10 and Tiam1, have been separately investigated in breast cancer cells using cell migration (wound healing) and cell invasion assays (Ma et al., 2007; Moriarty et al., 2010) Th e se studies have suggested that the knockdown of HOXD10 increases cell m igration speed while that of Tiam1 hinder s migration. Hence, th e se two proteins give rise to opposite effects on miR 10bs role in metastasis and how those effects would join other similar effects from other proteins, of which the mRNAs are the targets of miR10b, to reach to a collective outcome for metastasis. However,
86 this result seems to disobey the notion that two or more proteins are regulated by the same mechanism and the same macromolecules, but adopt opposite effects toward the same biological acti vity to maintain a balance d state for a biological process. Here we used miR 10b, Tiam1 shRNA and HOXD10 shRNA to examine migration capacity through persistence time and CMPI on NIH 3T3. Following the previous described methods, the results showed both as sessments on cell migration are well estimated, with around a 0.97 and 0.90 correlation coefficient to a 10hour nucleus distance. HOXD10 shRNA and Tiam1 shRNA separately promotes and hinders cell migration capacity. The introduction of miR 10b revealing t he slowest migration pattern implied that Tiam1 and HOXD10 may not be the only two downstream signaling proteins of miR 10b. This strategy may suggest a potential approach for studying miRNA. Pattern Recognition The distinct subcellular activities in cell migration are all governed by Rho GTPases but individually with different activity profiles. For example, the trailing edge detachment is mainly governed by RhoA activity with the involvement of microtubules (Kaver ina et al., 1999; Rid et al., 2005) P rotrusion is a result of all three Rho GTPases activities with precise spatial separation and timing (Machacek et al., 2009) Mean while membrane ruffling is produced mainly by Rac1 and Cdc42 activities. Hence, the changes in the C N correlation profiles possessed by a cell type can also reveal the changes in the mean activity profile of the Rho GTPases (or that of their pathways) for the cell type. This capacity can reveal the instantaneous changes in the Rho GTPases activity profile for their response to remote extracellular and/or intracellular signal(s) and bypass the complex signaling crosstalk. Hence, the three Rho GTPases might also
87 serve as an antennae to understand the response of a cell to perturbations once the connections between the activity profile of the Rho GTPases and other major cell activities such as cell proliferation and apoptosis are built. The d irect prediction of activity changes in individual Rho GTPases from the redistribution of the C N correlation data might be not straightforward. The previous example blocking the ROCK activity is easy to understand because the perturbation happens downstream of RhoA. Due to the crosstalk among the Rho GTPases (Cdc42 can activate Rac1 (Nobes and Hall, 1995) and Rac1 and RhoA mutually inhibit each other (Arthur and Burridge, 2001; Nimnual et al., 2003; Ohta et al., 2006; Rottner et al., 1999) and the crosstalk between individual Rho GTPases and other signaling pathways, a upstream signaling event leading to the changes in one of the Rho GTPases will most likely affect its cou nterparts. Therefore, the resulting changes in the C N correlation profile from a remote perturbation might not be directly linked to the changes in the activity profile of Rho GTPases. In order to solve this issue, a pattern recognition approach was adop ted. The RhoA, Rac1 and Cdc42 were separately knocked down using specific shRNAs in NIH 3T3 fibroblasts and the corresponding C N correlation profiles were constructed to compare to that of normal NIH 3T3 fibroblasts (Figure 42). The results showed that t hese manipulations individually possess unique C N correlation profiles that can be used for pattern recognition to identify the changes of individual members of the Rho GTPases. The pattern recognition principles obtained from comparing the redistributed C N correlation patterns of the individual Rho GTPases knockdown to the normal patterns
88 are: (1) l ower Cdc42 activity will shift the distribution of the C N correlation data from being skewed toward a smaller angle to more symmetry at 90 (a strong C dc42 activity will make polarity clear with more straightforward directional migration) ; (2) a higher CCD peak around 90 means weaker the Rac1 activity ; and (3) stronger RhoA activity is signified by having more C N correlation data with larger NCD// valu e between 10 and 40. Application of Pattern Recognition: Rho GTPases A ctivity u nder PI3K Inhibitors Using previous principles, we assessed the C N correlation profiles of NIH 3T3 fibroblasts with a treatment of 0.1M IC87114 and 5M Ly294002, which i s a specific a broad spectrum PI3K inhibitor, respectively (Figure 43). The decrease the activities of RhoA, Rac1 and Cdc42, while the inhibit ion of all PI3 Kinases can increase the activity of RhoA and Rac1, but decreases Cdc42. We further probed NIH 3T3 fibroblasts under the same conditions using Western blotting and pull down assays against Rho GTPases (Figure 43, right panels). The results validated and confirmed our biophysical analysis. T hese results are also in agreement with various studies in the literature (Ferreira et al., 2006; Papakonstanti et al., 2007; Polizio et al., 2011) Therefore, our novel approach is an effective method that can bypass the signaling crosstalk and faithfully reflect the changes in the activity profile of Rho GTPases under various and instantaneous perturbations. The above studies show that analysis thru a C N cor relation profile is a novel and extremely powerful topdown cellular approach, enabling us to tightly link cell migration capacity to the underlying Rho GTPases pathways. In practical cases such as
89 disease states, the expression level of the three Rho GTPases could be much more complicated. Nevertheless, the C N correlation profile can link the phenotype and molecular information and facilitate an effective comparison between abnormal cells and the normal counterparts.
90 Figure 41. The correlation between persistence time or CMPI and 10 hour nucleus distance (<10 hNCD>) To examine the effect of miR10b and its downstream proteins, HOXD10 and Tiam1, on cell migration, two approaches, persistence time and CMPI, are used to examine the migration capacity. Both estimation on migration capacity show good correlation, 0.97 and 0.9, with the 10 hour nucleus distance. Red circle and blue cross separately represent persistence time ( Left ) and CMPI (Right). The correlation coefficient (R) is determined by PPMC.
91 Figure 42. Rho GTPases pattern recognition To eliminating the complexity in signaling pathways, a pattern recognition through Rho GTPases, the nodal proteins in cell migration, may give a clue to solve t his issue. NIH 3T3 separately delivered with Rho GTPases shRNAs, including RhoA, Rac1, and Cdc42 shRNAs, was used to examine their single cell migration profile. Black solid line represents the wild type and the red, blue, and green dash lines separately r epresent the Rac1, RhoA, and Cdc42 shRNAs.
92 Figure 43 Rho GTPases activity estimation through pattern recognition. Through Rho GTPase pattern recognition, the change in Rho GTPase activity by PI3K inhibitor can be rev eled through C N correlation. From the C N correlation, it shows that lower Rac1 activity has large magnitude around 80100 in m CCD profile (Top panel) This suggests that the IC87114 treated NIH 3T3 has the less Rac1 activity comparing with wild type and Ly294002 treated NIH3T3. For RhoA activity, it also suggested a higher magnitude around 10 40 in m CCD profile corresponding to lower RhoA activity (middle panel). This suggests that Ly294002 and IC87114 separately promote and reduced the RhoA activit y. For Cdc42 estimation, l ower Cdc42 activity will shift the distribution of the C N correlation data from being skewed toward a smaller angle to more symmetry at 90 (Bottom panel) This suggests that both Ly294002 and IC87114 treated NIH 3T3 show less Cdc42 activity comparing with wild type.
93 LIST OF REFERENCES Abercrombie, M., Heaysman, J. E. and Pegrum, S. M. (1970). The locomotion of fibroblasts in culture. 3. Movements of particles on the dorsal surface of the leading lamella. Exp Cell Res 62, 38998. Abercrombie, M., Heaysman, J. E. and Pegrum, S. M. (1971). The locomotion of fibroblasts in culture. IV. Electron microscopy of the leading lamella. Exp Cell Res 67 35967. Arthur, W. T. and Burridge, K. (2001). RhoA inactivation by p190RhoGAP regulates cell spreading and migration by promoting membrane protrusion and polarity. Mol Biol Cell 12, 271120. Bacac, M. and Stamenkovic, I. (2008). Metastatic cancer cell. Annual Review of Pathology Mechanisms of Disease 3 221 247. Balles trem, C., Hinz, B., Imhof, B. A. and Wehrle Haller, B. (2001). Marching at the front and dragging behind: differential alphaVbeta3integrin turnover regulates focal adhesion behavior. J Cell Biol 155 1319 32. Barriga, E. H., Maxwell, P. H., Reyes, A. E. and Mayor, R. (2013). The hypoxia factor Hif J Cell Biol 201 75976. Bentwich, I., Avniel, A., Karov, Y., Aharonov, R., Gilad, S., Barad, O., Barzilai, A., Einat, P., Einav, U., Meiri, E. e t al. (2005). Identification of hundreds of conserved and nonconserved human microRNAs. Nat Genet 37 766 70. Bober, E., Franz, T., Arnold, H. H., Gruss, P. and Tremblay, P. (1994). Pax 3 is required for the development of limb muscles: a possible role for the migration of dermomyotomal muscle progenitor cells. Development 120 603 12. Buck, C. A. and Horwitz, A. F. (1987). Cell surface receptors for extracellular matrix molecules. Annu Rev Cell Biol 3 179205. Buckley, A., Davidson, J. M., Kamerath, C. D., Wolt, T. B. and Woodward, S. C. (1985). Sustained release of epidermal growth factor accelerates wound repair. Proc Natl Acad Sci U S A 82 73404. Burridge, K. and Wennerberg, K. (2004). Rho and Rac take center stage. Cell 116, 167 79. Chen, W. T. (1981a). Mechanism of retraction of the trailing edge during fibroblast movement. J Cell Biol 90 187 200. Chen, W. T. (1981b). Surface changes during retractioninduced spreading of fibroblasts. J Cell Sci 49 1 13.
94 Crisp, M., Liu, Q., Roux, K., Rattner J. B., Shanahan, C., Burke, B., Stahl, P. D. and Hodzic, D. (2006). Coupling of the nucleus and cytoplasm: role of the LINC complex. J Cell Biol 172 41 53. Dickinson, R. B., McCarthy, J. B. and Tranquillo, R. T. (1993). Quantitative characterization of cell invasion in vitro: formulation and validation of a mathematical model of the collagen gel invasion assay. Ann Biomed Eng 21 67997. DiMilla, P. A., Barbee, K. and Lauffenburger, D. A. (1991). Mathematical model for the effects of adhesion and mechanics on cell migration speed. Biophys J 60, 1537. Dunn, G. A. (1983). Characterising a kinesis response: time averaged measures of cell speed and directional persistence. Agents Actions Suppl 12, 14 33. Dunn, G. A. and Brown, A. F. (1987). A unified approach to analysing cell motility. J Cell Sci Suppl 8 81 102. Ellard, K., Sunose, M., Bell, K., Ramsden, N., Bergamini, G. and Neubauer, G. Bioorg Med Chem Lett 22, 45469. Ferreira, A. M., Isaacs, H., Hayflick, J. S., Rogers, K. A. and Sandig, M. (2006). The p110delta isoform of PI3K differentially regulates beta1 and beta2 integrinmediated monocyte adhesion and spreading and modulates diapedesis. Microcirculation 13 439 56. Gail, M. H. and Boone, C. W. (1970). The locomotion of mouse fibroblasts in tissue culture. Biophys J 10, 980 93. Gee, H. E., Camps, C., Buffa, F. M., Colella, S., Sheldon, H., Gleadle, J. M., Ragoussis, J. and Harris, A. L. (2008). MicroRNA 10b and breast cancer metastasis. Nature 455 E8 9; author reply E9. Ghigo, A., Damilano, F., Braccini, L. and Hirsch, E. (2010). PI3K inhibition in inflammation: Toward tailored therapies for specific diseases. Bioessays 32 185 96. Gimona, M. (2008). The microfilament system in the formation of invasive adhesions. Semin Cancer Biol 18 23 34. Gomes, E. R., Jani, S. and Gundersen, G. G. (2005). Nuclear movement regulated by Cdc42, MRCK, myosin, and actin flow establishes MTOC polarization in migrating cells. Cell 121 451 63. Greenhalgh, D. G. (1996). The role of growth factors in wound healing. J Trauma 41, 159 67.
95 Grose, R., Werner, S., Kessl er, D., Tuckermann, J., Huggel, K., Durka, S. and Reichardt, H. M. (2002). A role for endogenous glucocorticoids in wound repair. EMBO Rep 3 57582. Gu, J., Tamura, M., Pankov, R., Danen, E. H., Takino, T., Matsumoto, K. and Yamada, K. M. (1999). Shc and FAK differentially regulate cell motility and directionality modulated by PTEN. J Cell Biol 146 389403. Hall, A. (1998). Rho GTPases and the actin cytoskeleton. Science 279 50914. Heasman, S. J. and Ridley, A. J. (2008). Mammalian Rho GTPases: new i nsights into their functions from in vivo studies. Nat Rev Mol Cell Biol 9 690701. Huttenlocher, A., Sandborg, R. R. and Horwitz, A. F. (1995). Adhesion in cell migration. Curr Opin Cell Biol 7 697 706. Jones, G. E. (2000). Cellular signaling in macrophage migration and chemotaxis. J Leukoc Biol 68, 593 602. Kaverina, I., Krylyshkina, O. and Small, J. V. (1999). Microtubule targeting of substrate contacts promotes their relaxation and dissociation. J Cell Biol 146 103344. Kaverina, I., Krylyshkina, O. and Small, J. V. (2002). Regulation of substrate adhesion dynamics during cell motility. Int J Biochem Cell Biol 34, 746 61. Khatau, S. B., Hale, C. M., Stewart Hutchinson, P. J., Patel, M. S., Stewart, C. L., Searson, P. C., Hodzic, D. and Wirtz, D. (2009). A perinuclear actin cap regulates nuclear shape. Proc Natl Acad Sci U S A 106 1901722. Kim, Y. J., Sauer, C., Testa, K., Wahl, J. K., Svoboda, R. A., Johnson, K. R., Wheelock, M. J. and Knudsen, K. A. (2005). Modulating the strength of cadherin adhesion: evidence for a novel adhesion complex. J Cell Sci 118 3883 94. Kurosaka, S. and Kashina, A. (2008). Cell biology of embryonic migration. Birth Defects Res C Embryo Today 84, 102 22. Lauffenburger, D. A. and Horwitz, A. F. (1996). Cell migration: a physically integrated molecular process. Cell 84 35969. Lee, H., Goetzl, E. J. and An, S. (2000). Lysophosphatidic acid and sphingosine 1 phosphate stimulate endothelial cell wound healing. Am J Physiol Cell Physiol 278 C6128. Lee, J. S., Chang, M. I., Tseng, Y. and Wirtz, D. (2005). Cdc42 mediates nucleus movement and MTOC polarization in Swiss 3T3 fibroblasts under mechanical shear stress. Mol Biol Cell 16, 87180.
96 Li, L., Nrrelykke, S. F. and Cox, E. C. (2008). Persistent cell motion in the absence of external signals: a search strategy for eukaryotic cells. PLoS One 3 e2093. Lipton, A., Klinger, I., Paul, D. and Holley, R. W. (1971). Migration of mouse 3T3 fibroblasts in response to a serum factor. Proc Natl Acad Sci U S A 68 2799801. L uxton, G. W., Gomes, E. R., Folker, E. S., Vintinner, E. and Gundersen, G. G. (2010). Linear arrays of nuclear envelope proteins harness retrograde actin flow for nuclear movement. Science 329 956 9. Lynch, S. E., Colvin, R. B. and Antoniades, H. N. (198 9). Growth factors in wound healing. Single and synergistic effects on partial thickness porcine skin wounds. J Clin Invest 84 640 6. Lynch, S. E., Nixon, J. C., Colvin, R. B. and Antoniades, H. N. (1987). Role of platelet derived growth factor in wound healing: synergistic effects with other growth factors. Proc Natl Acad Sci U S A 84, 7696700. Ma, L., Teruya Feldstein, J. and Weinberg, R. A. (2007). Tumour invasion and metastasis initiated by microRNA 10b in breast cancer. Nature 449 6828. MacDonal d, I. C., Groom, A. C. and Chambers, A. F. (2002). Cancer spread and micrometastasis development: quantitative approaches for in vivo models. Bioessays 24, 885 93. Machacek, M., Hodgson, L., Welch, C., Elliott, H., Pertz, O., Nalbant, P., Abell, A., Johns on, G. L., Hahn, K. M. and Danuser, G. (2009). Coordination of Rho GTPase activities during cell protrusion. Nature 461 99 103. Martin, P. and Parkhurst, S. (2004). Parallels between tissue repair and embryo morphogenesis. Development 131 30213034. Mi zoguchi, T., Verkade, H., Heath, J. K., Kuroiwa, A. and Kikuchi, Y. (2008). Sdf1/Cxcr4 signaling controls the dorsal migration of endodermal cells during zebrafish gastrulation. Development 135 25219. Moriarty, C. H., Pursell, B. and Mercurio, A. M. (20 10). miR 10b targets Tiam1: implications for Rac activation and carcinoma migration. J Biol Chem 285, 205416. Mullins, R. D., Heuser, J. A. and Pollard, T. D. (1998). The interaction of Arp2/3 complex with actin: nucleation, high affinity pointed end capping, and formation of branching networks of filaments. Proc Natl Acad Sci U S A 95 6181 6. Nagel, M., Tahinci, E., Symes, K. and Winklbauer, R. (2004). Guidance of mesoderm cell migration in the Xenopus gastrula requires PDGF signaling. Development 131 2727 36.
97 Nath, C. and Gulati, S. C. (1998). Role of cytokines in healing chronic skin wounds. Acta Haematol 99, 175 9. Nimnual, A. S., Taylor, L. J. and Bar Sagi, D. (2003). Redox dependent downregulation of Rho by Rac. Nat Cell Biol 5 236 41. Nobes, C. D. and Hall, A. (1995). Rho, rac, and cdc42 GTPases regulate the assembly of multimolecular focal complexes associated with actin stress fibers, lamellipodia, and filopodia. Cell 81 53 62. Ohta, Y., Hartwig, J. H. and Stossel, T. P. (2006). FilGAP, a Rho and ROCK regulated GAP for Rac binds filamin A to control actin remodelling. Nat Cell Biol 8 803 14. Oliver, T., Lee, J. and Jacobson, K. (1994). Forces exerted by locomoting cells. Semin Cell Biol 5 139 47. Omelchenko, T., Vasiliev, J. M., Gelfan d, I. M., Feder, H. H. and Bonder, E. M. (2002). Mechanisms of polarization of the shape of fibroblasts and epitheliocytes: Separation of the roles of microtubules and Rhodependent actinmyosin contractility. Proc Natl Acad Sci U S A 99 104527. Othmer, H. G., Dunbar, S. R. and Alt, W. (1988). Models of dispersal in biological systems. J Math Biol 26, 263 98. Palecek, S. P., Huttenlocher, A., Horwitz, A. F. and Lauffenburger, D. A. (1998). Physical and biochemical regulation of integrin release during r ear detachment of migrating cells. J Cell Sci 111 ( Pt 7) 929 40. Pankov, R., Endo, Y., Even Ram, S., Araki, M., Clark, K., Cukierman, E., Matsumoto, K. and Yamada, K. M. (2005). A Rac switch regulates random versus directionally persistent cell migration. J Cell Biol 170 793 802. Panorchan, P., Thompson, M. S., Davis, K. J., Tseng, Y., Konstantopoulos, K. and Wirtz, D. (2006). Single molecule analysis of cadherinmediated cell cell adhesion. J Cell Sci 119 66 74. Papakonstanti, E. A., Ridley, A. J. a nd Vanhaesebroeck, B. (2007). The p110delta isoform of PI 3kinase negatively controls RhoA and PTEN. EMBO J 26, 305061. Polizio, A. H., Chinchilla, P., Chen, X., Kim, S., Manning, D. R. and Riobo, N. A. (2011). Heterotrimeric Gi proteins link Hedgehog s ignaling to activation of Rho small GTPases to promote fibroblast migration. J Biol Chem 286, 1958996. Ren, X. D., Kiosses, W. B. and Schwartz, M. A. (1999). Regulation of the small GTP binding protein Rho by cell adhesion and the cytoskeleton. EMBO J 18 57885.
98 Richardson, B. and Lehmann, R. (2010). Mechanisms guiding primordial germ cell migration: strategies from different organisms. Nature Reviews Molecular Cell Biology 11, 37 49. Rid, R., Schiefermeier, N., Grigoriev, I., Small, J. V. and Kaverina I. (2005). The last but not the least: the origin and significance of trailing adhesions in fibroblastic cells. Cell Motil Cytoskeleton 61 161 71. Ridley, A. J. and Hall, A. (1992). The small GTP binding protein rho regulates the assembly of focal adhesions and actin stress fibers in response to growth factors. Cell 70, 38999. Ridley, A. J., Paterson, H. F., Johnston, C. L., Diekmann, D. and Hall, A. (1992). The small GTP binding protein rac regulates growth factor induced membrane ruffling. Cell 70, 401 10. Rottner, K., Hall, A. and Small, J. V. (1999). Interplay between Rac and Rho in the control of substrate contact dynamics. Curr Biol 9 6408. Sahai, E. and Marshall, C. J. (2003). Differing modes of tumour cell invasion have distinct requirement s for Rho/ROCK signalling and extracellular proteolysis. Nat Cell Biol 5 7119. Sander, E. E., ten Klooster, J. P., van Delft, S., van der Kammen, R. A. and Collard, J. G. (1999). Rac downregulates Rho activity: reciprocal balance between both GTPases determines cellular morphology and migratory behavior. J Cell Biol 147 100922. Sanz Moreno, V., Gadea, G., Ahn, J., Paterson, H., Marra, P., Pinner, S., Sahai, E. and Marshall, C. J. (2008). Rac activation and inactivation control plasticity of tumor cell movement. Cell 135, 510 23. Schmidt, C. E., Horwitz, A. F., Lauffenburger, D. A. and Sheetz, M. P. (1993). Integrin cytoskeletal interactions in migrating fibroblasts are dynamic, asymmetric, and regulated. J Cell Biol 123 97791. Scita, G., Tenca, P., Frittoli, E., Tocchetti, A., Innocenti, M., Giardina, G. and Di Fiore, P. P. (2000). Signaling from Ras to Rac and beyond: not just a matter of GEFs. Embo J 19 23938. Small, J. V., Rottner, K. and Kaverina, I. (1999). Functional design in the actin cyt oskeleton. Curr Opin Cell Biol 11 54 60. Stokes, C. L., Lauffenburger, D. A. and Williams, S. K. (1991). Migration of individual microvessel endothelial cells: stochastic model and parameter measurement. J Cell Sci 99 ( Pt 2) 419 30.
99 Suraneni, P., Rubi nstein, B., Unruh, J. R., Durnin, M., Hanein, D. and Li, R. (2012). The Arp2/3 complex is required for lamellipodia extension and directional fibroblast cell migration. J Cell Biol 197 239 51. Swartz, M. E., Eberhart, J., Pasquale, E. B. and Krull, C. E. (2001). EphA4/ephrinA5 interactions in muscle precursor cell migration in the avian forelimb. Development 128 4669 80. Tamkun, J. W., DeSimone, D. W., Fonda, D., Patel, R. S., Buck, C., Horwitz, A. F. and Hynes, R. O. (1986). Structure of integrin, a g lycoprotein involved in the transmembrane linkage between fibronectin and actin. Cell 46, 271 82. Theriot, J. A. (1994). Regulation of the actin cytoskeleton in living cells. Semin Cell Biol 5 1939. Todaro, G. J., Lazar, G. K. and Green, H. (1965). The initiation of cell division in a contact inhibited mammalian cell line. J Cell Physiol 66 325 33. Tonami, K., Kurihara, Y., Arima, S., Nishiyama, K., Uchijima, Y., Asano, T., Sorimachi, H. and Kurihara, H. (2011). Calpain6, a microtubulestabilizing protein, regulates Rac1 activity and cell motility through interaction with GEF H1. J Cell Sci 124 121423. Uehata, M., Ishizaki, T., Satoh, H., Ono, T., Kawahara, T., Morishita, T., Tamakawa, H., Yamagami, K ., Inui, J., Maekawa, M. et al. (1997). Calcium sensitization of smooth muscle mediated by a Rhoassociated protein kinase in hypertension. Nature 389 9904. van Leeuwen, F. N., van Delft, S., Kain, H. E., van der Kammen, R. A. and Collard, J. G. (1999). Rac regulates phosphorylation of the myosinII heavy chain, actinomyosin disassembly and cell spreading. Nat Cell Biol 1 242 8. Winklbauer, R. (1990). Mesodermal cell migration during Xenopus gastrulation. Dev Biol 142 155 68. Witt, C. M., Raychaudhur i, S., Schaefer, B., Chakraborty, A. K. and Robey, E. A. (2005). Directed migration of positively selected thymocytes visualized in real time. PLoS Biol 3 e160. Wolf Yadlin, A., Kumar, N., Zhang, Y., Hautaniemi, S., Zaman, M., Kim, H. D., Grantcharova, V ., Lauffenburger, D. A. and White, F. M. (2006). Effects of HER2 overexpression on cell signaling networks governing proliferation and migration. Mol Syst Biol 2 54. Wolf, K., Mazo, I., Leung, H., Engelke, K., von Andrian, U. H., Deryugina, E. I., Strong in, A. Y., Brcker, E. B. and Friedl, P. (2003). Compensation mechanism in tumor cell migration: mesenchymal amoeboid transition after blocking of pericellular proteolysis. J Cell Biol 160, 267 77.
100 Woo, S., Housley, M. P., Weiner, O. D. and Stainier, D. Y (2012). Nodal signaling regulates endodermal cell motility and actin dynamics via Rac1 and Prex1. J Cell Biol 198 941 52. Yamazaki, D., Kurisu, S. and Takenawa, T. (2009). Involvement of Rac and Rho signaling in cancer cell motility in 3D substrates. O ncogene 28 157083. Yoshida, K. and Soldati, T. (2006). Dissection of amoeboid movement into two mechanically distinct modes. J Cell Sci 119 383344. Yoshizaki, H., Ohba, Y., Kurokawa, K., Itoh, R. E., Nakamura, T., Mochizuki, N., Nagashima, K. and Mat suda, M. (2003). Activity of Rhofamily GTPases during cell division as visualized with FRET based probes. J Cell Biol 162 22332. Zhang, J., Xu, X. K., Li, P., Zhang, K. and Small, M. (2011). Node importance for dynamical process on networks: A multiscale characterization. Chaos 21, 016107.
101 BIOGRAPHICAL SKETCH ShenHsiu was born in 1980 in Taipei, Taiwan. He is senior in his family with two younger sisters. After graduating from Yanping high school in 1998, he entered C hemical Engineering D epartment, National Chung Hsing University (NCHU). He earned his B.S. in 2002 and study graduate program Chemical E ngineering, in National Cheng Kung University ( NCKU) from 2002 to 2004. Later, he served as second l ieutenant in Taiwan Army Artillery and Missile School and had teaching assistant position in NCKU till 2007. In 20 07 fall, he enrolled in the Master program in the Department of Chemical Engineering at Univer sity of Florida. He joined Dr. Yiider Tsengs research group and focused on cellular biophysical approach from 20092013. He earned his Doctor of Philosophy degree in chemical engineering in August 2013