Growth Factor-Induced Cell Migration: A Quantitative and Mechanistic
Analysis
by
Margaret Faye Ware
B.S. Chemical Engineering
B.A. Chemistry
North Carolina State University, 1992
M.S. Chemical Engineering
University of Illinois at Urbana-Champaign, 1995
SUBMITTED TO THE DEPARTMENT OF CHEMICAL ENGINEERING IN PARTIAL
FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY IN CHEMICAL ENGINEERING
AT THE
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
FEBRUARY 1998
@ 1998 Massachusetts Institute of Technology. All rights reserved.
Signature of Author:
Departrfient of Chemical Engineering
September 9, 1997
Certified by:
Dulas A. L'auffenburger
J.R. Mares Professor of Chemical Engineering
Thesis Supervisor
Accepted by:
Robert Cohen
Engineering
of
Chemical
St. Laurent Professor
Chairman, Committee for Graduate Students
ap
3r ',
1.
Growth Factor-Induced Cell Migration: A Quantitative and Mechanistic
Analysis
by
Margaret Faye Ware
Submitted to the Department of Chemical Engineering
on September 9, 1997 in Partial Fulfillment of the
Requirements for the Degreee of Doctor of Philosophy in
Chemical Engineering
Abstract
While it has been established that epidermal growth factor (EGF) elicits motility
responses in fibroblastic cells, little is understood concerning the biophysical processes by
which biochemical signals are translated into migratory behavior. We apply individual cell
tracking techniques to quantitatively characterize the linear migration speed (S { m/min })
and directional persistence (P {min}) for NR6 fibroblasts on varying concentrations of the
extracellular matrix substratum Amgel in response to EGF. We find that at an intermediate
Amgel concentration, EGF stimulation leads to a substantial increase in migration speed
compared to control conditions; however at low or high Amgel concentrations EGF fails to
induce cell motility above basal cell movement. EGF stimulation simultaneously results in a
substantial decrease in directional persistence at the same intermediate Amgel concentration.
This translates to an increase in both the mean path length (SP {.m}) and the mean
dispersion (S2p {(pm 2 /min}) due to EGF stimulation. The speed effect was abrogated by
-. inhibiting the PLCy signaling pathway (using the pharmacological agent U73122 and c'973
EGFR truncation mutant), consistent with previous observations implicating PLCy in EGF
receptor-mediated motility responses. However, the persistence effect was only altered by
4 the truncation mutant but not by the drug. The biophysical process of membrane extension
was also examined under the same conditions as speed and persistence. The DiMilla model
implies the rate of membrane extension and cell speed should be directly proportional.
Membrane extension increases concomitantly with EGF-induced speed, and also depends
on the substratum concentration. There is a positive correlation between speed and
extension rate while there is a negative correlation between persistence and extension rate.
Simultaneously we find a negative correlation between speed and cell spread area and a
positive correlation between persistence and cell spread area. Thus, fibroblast migration
responses to EGF are strongly dependent on substratum conditions, and change linear
speed and directional persistence in qualitatively opposite manners. The result that speed
and persistence are affected differently under similar conditions has important implications
for therapy targeting strategies.
Thesis Supervisor: Douglas A. Lauffenburger
Title: J.R. Mares Professor of Chemical Engineering
Acknowledgments
I would like to acknowledge my advisor, Doug Lauffenburger, for his mentoring
and encouragement over the course of my thesis. I would also like to thank Alan Wells for
his gift of cell lines and many ideas. Experimentally, I would like to thank Doug Osborne
for generating all the data for the five minute time points and Gargi Maheshwari for the
streamlined Excel macro in Appendix 1 for calculating mean squared displacements. Many
thanks to Eric Boder for his help in the final stages of this thesis. And, most of all, I
would like to thank "The Muddy" for my sanity.
The funding for this work was provided by NIH CA69213.
Table of Contents
7
List of Tables ................................................................................
List of Figures .........
8
....................................................................
Chapter 1: Introduction ..........................................................
...............
1.1 Cancer and EGFR ......................................................................
1.2 EGFR System ...........
9
................................................................ 11
1.3 EGF and Cell Physiology .................
1.4 EGF Motility Signaling ......................
1.5 Cell Migration Mechanisms .......................
......................................... 12
................................ 13
....
..................................... 14
1.6 Experimental System................................................................. 17
1.7 Thesis Overview.......................................................................... 18
Tables ......
...............................................
20
Figures ..................................................................................................
21
..........
Chapter 2: Materials and Methods ....................................................................
24
2.1 M aterials ..............................................................................
24
2.2 Cell Lines and Culture .......................................................
24
....
2.3 Substratum Preparation .............................................................. ..
24
2.4 Cell Migration Assays ................................................................
25
2.5 Pharmacological PLC Inhibitors ...................................................
30
2.6 Membrane Extension Determinations ................................................. 30
Figures ...................................................................
.......... 31
C hapter 3: R esults ...........................................................................
3.1 EGF Increases Individual Cell Locomotion ...........
..
36
......................... 36
3.2 Full EGF Motility Response Requires an Induction Time .......................... 36
3.3 Level of EGF Response Depends on Substratum Concentration ................. 37
3.4 EGF Causes a Substratum Dependent Increase in Cell Speed and a
Decrease in Directional Persistence .................................................
37
3.5 EGF Causes an Increase in Path Length and Dispersion ........................ 39
3.6 PLC Abrogation Reduces EGF Effect .............................................. 39
3.7 Membrane Extension is Increased by EGF ........................................... 41
3.8 Membrane Extension Correlates with Cell Persistence in the Presence
of EGF, but with Cell Speed without EGF ......................................... 42
T ables ....................................................................................
..
44
Figures ...................................................................................
..
45
Chapter 4: Discussion .................................................................
75
Appendix 1: Microsoft Excel Macro for Mean Squared Displacement Calculations ....... 85
Appendix 2: Analysis Comparison................................................................ 87
Appendix 3: Individual Cell Speed and Persistence Times ..................................... 88
Appendix 4: Speed Histograms.................................................................... 97
A4.1 WT Basal and EGF-Induced Speed Comparisons ................................ 97
A4.2 Effect of Amgel on WT Speed Histograms ...................................... 99
A4.3 c'973 Basal and EGF-Induced Speed Comparisons ............................ 100
A4.4 Effect of Amgel on c'973 Speed Histograms .................................. 102
A4.5 Effect of PLC Abrogation on Speed Histograms ................................. 103
Appendix 5: Persistence Histograms ............................................................. 104
A5.1 WT Basal and EGF-Induced Persistence Comparisons ........................ 104
A5.2 Effect of Amgel on WT Persistence Histograms ................................ 106
A5.3 c'973 Basal and EGF-Induced Persistence Comparisons ........................ 107
A5.4 Effect of Amgel on c'973 Persistence Histograms ................................ 109
A5.5 Effect of PLC Abrogation on Persistence Histograms ............................ 110
Appendix 6: Membrane Extension Histograms ................................................. 111
A6.1 WT Basal and EGF-Induced Extension Comparisons ............................ 111
A6.2 Effect of Amgel on WT Extension Histograms .................................. 113
A6.3 Effect of PLC Abrogation on Extension Histograms ............................ 114
Appendix 7: Cell Spread Area Histograms ......................................................
115
A7.1 WT Basal and EGF-Induced Area Comparisons ................................ 115
A7.2 Effect of Amgel on WT Area Histograms ........................................ 117
A7.3 Effect of PLC Abrogation on Area Histograms................................ 118
References ..................................................................................
119
List of Tables
Chapte r 1:
Introduction
Table 1.1
Chapte r 3:
Amgel Components................................
......... 20
Results
Table 3.1
Percent M otile Cells..........................................44
List of Figures
Chapter 1: Introduction
Figure 1.1 EGFR Structure.................................................................21
Figure 1.2 EGFR Motility Signaling ....................................................
22
Figure 1.3 Physical Components of Cell Migration ...................................
23
Chapter 2: Materials and Methods
Figure 2.1 EGF Induction Period ........................................................
31
Figure 2.2 Mean Squared Displacement versus Time ................................. 32
Figure 2.3 Membrane Extension Calculation ............................................. 35
Chapter 3: Results.
Figure 3.1 EGF-induced Morphology ................................................... 45
Figure 3.2 Cell Tracks, "Wind-rose" Plots ............................................... 46
Figure 3.3 EGF-induced Speed and Persistence, WT EGFR .......................... 47
Figure 3.4 Speed and Persistence Histograms ....................................... 49
Figure 3.5 EGF-induced Path Length and Dispersion............................... 52
Figure 3.6 Receptor Truncation Effect on Speed and Persistence .................... 54
Figure 3.7 PLC Abrogation Effect on Speed and Persistence .......................... 60
Figure 3.8 EGF-induced Membrane Extension ......................................
62
Figure 3.9 PLC Abrogation Effect on Membrane Extension and Cell Area......... 63
Figure 3.10 EGF Effect on Interdependence of Speed, Persistence, Extension and
Substratum Concentration .................................................. 65
Figure 3.11 Interdependence of Speed, Persistence and Extension .................. 69
Figure 3.12 Interdependence of Speed, Persistence and Cell Spread Area ......... 73
Chapter 1: Introduction
Cell migration plays an important role in many normal and pathological processes.
Physical mechanisms of adhesion, contraction and protrusion can regulate cell migration
parameters such as speed and persistence time.
Understanding these biophysical and
biochemical steps will provide valuable insight into potential therapies and applications in
the areas of wound healing, tissue engineering and cancer. The focus of this thesis is to
determine how epidermal growth factor affects fibroblast migratory characteristics,
specifically speed, persistence time and membrane extension.
1.1
Cancer and EGFR
The transformation of a normal cell to a cancerous cell can be initiated by normal
cellular genes, proto-oncogenes, that have been rendered transforming by mutations,
deletions, or insertions. Amplification of normal cellular genes with no other modifications
is also capable of transforming cells.
Altered genes can cause a change in the growth
characteristics of cells in a variety of ways. Cellular proto-oncogenes are found throughout
the cell as growth factors, growth factor receptors, intracellular transducers, and nuclear
proteins.
The ubiquitous presence of these proteins suggests functions in growth,
development, and differentiation (Burck, Liu et al. 1988).
oncogenes
may control
autocrine
In normal cells, proto-
growth factor production,
signal transduction,
transcription, translation, and cell surface properties along with any number of other
functions.
Many known proto-oncogenes encode growth factors and growth factor
receptors.
One well studied system involved in cellular proliferation that has been
implicated in the transformation of human cells is the epidermal growth factor system.
Mortality from human cancers is largely attributable to invasive spread and
metastasis. Localized tumors are often curable by removal or targeted radio- or chemo4 therapy. The epidermal growth factor receptor (EGFR) is the growth factor receptor most
commonly associated with human cancers (Aaronson
1991).
Overexpression
or
dysregulation of EGFR signaling correlates with poor prognosis in glioblastomas, breast,
bladder, prostate and other cancers (LeRiche, Asa et al. 1996).
An important aspect of
tumor invasion is cell motility. In vitro motility has been correlated with malignancy in
human gliomas (Chicoine and Silbergeld 1995). Cell motility signals have been shown to
be critical for invasion of prostate tumor xenografts (Xie, Wang et al. 1995; Turner, Chen
et al. 1996).
Treatment of a variety of cells in vitro with EGF has been shown to stimulate
invasion.
Both rat mammary adenocarcinoma cells and PC-3 prostate carcinoma cells,
when treated with EGF show increased invasion (Li, Nagayasu et al. 1993; Jarrad, Blitz et
al. 1994). When human glioblastoma biopsies were tested for invasion of normal brain
tissue, enhanced invasion occurred in 7 of 8 biopsies when they were treated with EGF,
but not with PDGF, NGF or bFGF (Engebraaten, Bjerkvig et al. 1993).
EGF enhances
the migration and invasion of follicular and papillary thyroid cancer both in culture and in
nude mice (Hoelting, Siperstein et al. 1994).
DU-145 prostate cells or fibroblasts
expressing the full-length migratory EGFR have also been shown to transmigrate the ECM
with no concurrent alterations in protease production (Xie, Wang et al. 1995).
Metastasis appears to depend in part on tumor cell motility. Also, the presence of
EGFR correlates more closely with tumor invasion rather than tumor formation. Even
though there are a great number of pharmaceuticals for breast cancers, the death rate has not
decreased (Dickson, Johnson et al. 1996).
These chemotherapeutic agents are generally
growth inhibitors and do not deal directly with the invasive potential.
Amplification of
erbB-2, a relative of the EGFR, is a poor prognostic indicator of breast cancer (Dickson,
Johnson et al. 1996).
There is increased EGFR expression and message in -40% of
glioblastomas but not in the lower grade gliomas (Libermann, Razon et al. 1984; Collins
1993). A much higher proportion of advanced (38/130) compared to early (0/26) gastric
carcinomas have enhanced EGFR immunostaining (Yasui, Sumiyoshi et al. 1988). Also,
the metastatic potential of human colon cells has been shown to correlate directly with the
level of EGFR functioning (Radinsky, Risin et al. 1995).
1.2
EGFR System
EGFR is a 170 kDa transmembrane protein tyrosine kinase that becomes activated
via phosphorylation upon ligand binding, is the initial step in eliciting cell movement,
growth, and differentiation responses of epithelial as well as non-epithelial cell types
(Carpenter 1985; Wells and Bishop 1988; Nanney and King 1996).
Activated EGFR is
able to bind the src homology two (SH2) domains of many proteins including Grb2,
PLC'y, and Shc (Pawson and Schlessinger 1993). Binding to activated EGFR can serve to
modify the activity of these proteins (e.g., PLCy) or simply to recruit them to the cell
surface making them available for interactions with other proteins (e.g., Grb2). The ability
of EGF to elicit these different responses, and the association of known components of cell
regulatory pathways with activated EGFR, indicate an intimate involvement of the EGF
system with physiological function of a variety of cell types. However, it has never been
clear which dysregulated cell function proximally underlies malignant progression.
The gene encoding the EGFR was one of the first cellular genes found associated
with tumors and identified as an oncogene (Kung, Chang et al. 1994). Additionally, the
oncogene erbB-2 (neu/HER-2), which is found amplified or overexpressed in a large
number of human cancers, is a close relative to erbB, the gene encoding the EGFR (Hynes
and Stem 1994). The EGFR and the erbB-2 gene product are very similar structurally. A
single mutation in the juxtamembrane region of the EGFR transforms a cell in an erbB-2
like fashion (Hynes and Stem 1994). The v-erbB oncogene product is also similar to the
EGFR, homologous to the transmembrane and protein kinase domains of the protein, but
lacking the regulatory extracellular
ligand binding domain and carboxyl-terminal
autophosphorylation sites (Gill, Bertics et al. 1987). The structure of the EGFR is shown
in Figure 1.1.
The receptor was originally isolated from a human carcinoma cell line (A-
431) expressing large numbers of the receptor (Carpenter and Wahl 1990).
The
extracellular domain contains two cysteine-rich regions that are likely to flank the ligand
binding region. A short, hydrophobic sequence anchors the receptor in the cell membrane
and possibly transmits the ligand binding signal to the cytoplasmic domain (Gill, Bertics et
al. 1987). The juxtamembrane region contains key residues that are major phosphorylation
sites and have been shown to play a part in receptor modulation through binding to protein
kinase C and other mechanisms (Carpenter and Wahl 1990; Kung, Chang et al. 1994).
The mutation site for conferrance of the erbB-2 phenotype, Arg 662, is also contained
within this region (Hynes and Stern 1994).
The remaining cytoplasmic region of the
receptor contains the tyrosine kinase domain and several autophosphorylation sites
(Carpenter and Wahl 1990). In the unphosphorylated inactive state, the carboxyl- terminal
regulatory region is thought to "fold-up" and interact with the protein kinase domain
(Kung, Chang et al. 1994).
Upon ligand binding and receptor activation, the
phosphorylated tyrosine residues become potential binding sites for SH2 containing
proteins, many of which are involved in tyrosine kinase signaling pathways (Pawson
1995).
1.3
EGF and Cell Physiology
The EGFR is expressed in a variety of human tissues, and its normal functioning is
essential in a variety of physiologies. When the receptor system is dysfunctional, for any
reason, pathologies such as invasion and metastasis can arise as discussed in section 1.1.
While EGF has traditionally been considered a purely proliferative agent, its effects on cell
morphogenesis and motility have begun to be studied in depth. EGFR signaling can lead to
a variety of cellular processes that include motility events as well as proliferation. The
motility signaling pathway will be discussed further below. The EGFR itself is able to
bind actin through residues 984-996 (den Hartigh, van Bergen en Henegouwen et al.
1992).
Addition of EGF to A431 cells causes changes in cell morphology including
extensive membrane ruffling (Diakonova, Payrastre et al. 1995) and an increase in actin
filaments (Dadabay, Patton et al. 1991; Rijken, Post et al. 1995). EGF has also been
found to increase migration for corneal epithelial cells (Wilson, He et al. 1994; Maldonado
and Furcht 1995), mammary epithelial cells (Matthay, Thiery et al. 1993), keratinocytes
(Ando and Jensen 1993), as well as fibroblasts (Chen, Gupta et al. 1994). In addition to
these in vitro studies, EGF has been implicated in a variety of normal physiological
functions including embryogenesis and wound healing (Kurachi, Morishige et al. 1994;
Nanney and King 1996). The EGFR is essential for normal development; EGFR null mice
are viable to day 8, but they are born with multiple tissue deficiencies due to impaired
epithelial development (Miettinen, Berger et al. 1995).
1.4
EGF Motility Signaling
Increased movement of fibroblast and epithelial cells is induced by EGF (Chen,
Gupta et al. 1994, Cha, 1996; Maldonado and Furcht 1995). This EGF-effect has been
shown to be associated with increased membrane activity in fibroblast cells (Ridley,
Paterson et al. 1992; Segall, Tyerech et al. 1996). At a molecular level, roles have been
implicated for the actin-modifying protein gelsolin and the membrane phospholipidmodifying enzyme PLCy in the EGF motility response. Cellular events triggered by PLCy
activation affect the cell motility machinery (Figure 1.2).
PLCy has been implicated
specifically in the EGF-induced motility response (Chen, Xie et al. 1994) as well as in
motility responses to PDGF and IGF-1 (Bornfeldt, Raines et al. 1994; Kundra, Escobedo
et al. 1994).
PLC hydrolyzes phosphoinositide bisphosphate (PIP2) generating
diacylglycerol (DAG) and inositol trisphosphate (IP3). Actin modifying proteins such as
profilin and gelsolin can bind PIP2 and are released upon its hydrolysis (Cunningham,
Stossel et al. 1991; Goldschmidt-Clermont, Kim et al. 1991; Noh, Shin et al. 1995).
These proteins serve to modify actin and therefore cell movement (Carpenter and Cantley
1996).
Gelsolin has been shown to affect cell movement (Cunningham, Stossel et al.
1991), and it has recently been shown that EGF-induced cell movement can be reproduced
by gelsolin displacement from PIP2 (Chen, Murphy-Ullrich et al. 1996). The hydrolysis
products of PIP2 can also serve to modify cell motility. DAG activates protein kinase C
(PKC), which modulates cell-substratum interactions, and IP3 causes intracellular calcium
levels to rise with resultant effects on actin filament generation (Preston, King et al. 1990).
Elevated intracellular calcium levels in tumor cells are associated with increased cell motility
(Savarese, Russell et al. 1992) and transmigration of an extracellular matrix barrier (Fong,
Sutkowski et al. 1992).
1.5
Cell Migration Mechanisms
This work is based on a model developed previously (DiMilla, Barbee et al. 1991),
and which has had a number of its major assumptions and predictions validated by
experimental tests (DiMilla, Stone et al. 1993; Schmidt, Horwitz et al. 1993; Wu, Hoying
et al. 1994). This model schematizes locomotion as a cycle of lamellipodal extension,
uropodal detachment and cell body translocation, and relaxation (Figure 1.3) (Lackie
1986). Movement of a cell over time is determined both by the rate of locomotion (speed of
motility) and the time period over which the leading edge remains in the same direction
(directional persistence of motility). Thus, an observed increase in cell movement may be
the result of either speed or persistence. Which of these two parameters is altered may
depend on the underlying molecular and physical bases.
Moreover, two different aspects of active force generation operate in the locomotion
cycle. One, due to actin polymerization, causes lamellipodal extension; the other, due to
actin/myosin contraction, results in uropodal detachment and cell body translocation; both
are likely generated by the cytoskeleton (Bray 1992). The model basically poses the
fundamental conceptual relationship between adhesion and migration in a quantitative
framework, focusing on the ratio of intracellular motile force to cell/substratum traction
during the cell body translocation phase following lamellipodal extension.
Dynamic cell/substratum adhesion processes during migration are mediated in many
cells by specific reversible interactions between transmembrane receptors and substratumbound extracellular matrix (ECM) ligands. Foremost among these receptors are the
integrins (Buck and Horwitz 1987; Hynes 1987; Ruoslahti and Pierschbacher 1987).
Because integrins bind both ECM ligands and cytoskeletal elements with low affinity they
are attractive candidates for the role of translating intracellular stresses to extracellular
traction (Burridge, Fath et al. 1988). Other classes of membrane-anchored molecules
present in adhesion
plaques,
including members of the
syndecam
family
of
glycosaminoglycans, may also serve to bridge the cytoskeleton to the ECM substratum
(Turley 1992). Thus, a number of different classes of transmembrane proteins may serve to
transmit the forces necessary for cell motility.
By investigating cell adhesiveness and
movement on a biologically active, complex ECM, such as Amgel (Siegal, Wang et al.
1993), one may obtain a relevant picture of cell motility.
Several studies have demonstrated that variations in either the adsorbed density of
substratum-bound ligands for integrins or the integrin/ligand bond affinity can affect
motility. Goodman et al. (Goodman, Risse et al. 1989) found a biphasic relationship
between the movement of murine skeletal myoblasts and the adsorbed density of both
laminin and the cell-binding laminin fragment E8. Duband et al. (Duband, Dufour et al.
1991) observed that the extent of migration for neural crest cells decreased with increasing
surface density of high-affinity antibodies against the 11 integrin subunit but was enhanced
by increasing densities of corresponding low-affinity antibodies. An optimal strength of
cell/substratum adhesion exists for maximal cell migration speed (DiMilla, Stone et al.
1993; Wu, Hoying et al. 1994); adhesiveness above or below that optimum results in
decreased cell motility and even immotility.
Thus, cell migration as well as morphology may depend on the strength of transient
cell-substratum attachments (Stein and Bronner 1989), and three regimes of motile and
morphological behavior can be envisioned for a cell interacting with a surface. On weakly
adhesive surfaces cell/substratum interactions cannot provide traction, so that the cell
spreads poorly and no locomotion is possible. On strongly adhesive surfaces the cell is
well-spread and immobilized, so regular dynamic disruption of cell/substratum attachments
is difficult and locomotion again does not occur. For an intermediate strength of
cell/substratum interactions, however, cell body translocation may be possible. The terms
weak and strong adhesion are relative to the level of motile force generated within the cell
and transmitted to the cell/substratum attachments. The frequency of lamellipod extension
concomitantly governs the rate at which a cycle of extension and contraction occurs.
A crucial unanswered question is what are the specific physical processes by which
EGFR-mediated signaling leads to enhanced cell migration.
spectrum
of
processes,
including
cytoskeletal
Alterations in a diverse
organization,
integrin-mediated
cell/substratum linkages, and intracellular force generation, can be plausibly envisioned to
be involved (Lauffenburger and Horwitz 1996).
At present, however, it is not known
whether EGF-enhanced cell movement occurs by means of effects on linear translocation
speed or on directional persistence, the two major components of random motility
(Lauffenburger and Linderman 1993) -- or both.
I therefore have used individual cell
tracking to examine how EGF stimulation affects speed and persistence. Moreover, I do
this over a range of extracellular matrix levels, for the human amnion extract Amgel
(Siegal, Wang et al. 1993), in order to ascertain whether the EGF effect is modulated by
cell/substratum interactions.
1.6
Experimental
System
A cell type in which EGF receptors are normally present is preferred for
investigating EGFR-mediated signaling. An in vitro cell culture system allows isolation of
the motility response. Established cell lines minimize cell variability of biologic responses;
however, the target cell line should not be transformed as epigenetic changes may result in
spurious signaling pathways. Lastly, a cell line which does not express native EGFR is
essential, as the background signaling may confound analyses; overcoming background
signals by extreme supraphysiologic expression runs the risk of recruiting physiologically
irrelevant responses (DiFiore, Pierce et al. 1987; Velu, Beguinot et al. 1987; McCubrey,
Steelman et al. 1991).
The murine 3T3-derived NR6 cells (Pruss and Herschman 1977) are contactinhibitable, non-transformed, devoid of endogenous EGF receptors and receptor message,
and do not demonstrate any measurable response to EGF or TGFa. Use of this line avoids
the confounding background of signaling from, or heterodimeric complexing (SpivakKroizman, Rotin et al. 1992; Hack, Sue-A-Quan et al. 1993) with endogenous wild-type
EGFR. This EGF non-responder variant line was derived from 3T3 cells that express
EGFR and proliferate in the presence of ligand. It is assumed that these cells possess the
normal fibroblast response machinery, except for the EGF receptor. This assumption is
supported by the observation that cell responses are similar to other 3T3 cells and
fibroblasts when challenged by PDGF and insulin (A. Wells, unpublished data) and
conditioned media (Cook, Pittelkow et al. 1991); these responses include cell proliferation,
altered morphology, and de novo transcription of early response genes. Furthermore, NR6
cells reacquire the normal fibroblast responses to EGF after expression of exogenous
EGFR (Jiang and Schindler 1990; Wells, Welsh et al. 1990; Masui, Wells et al. 1991;
Welsh, Gill et al. 1991).
The selection of a substratum is also an important consideration. I have chosen to
use Amgel, a biologically active extra-cellular matrix isolated from human amnions. Amgel
is very similar to the commercially available product Matrigel, but lacks detectable levels of
PDGF, TGFa, or EGF, any of which could confound our assays (Siegal, Wang et al.
1993). The exact composition of Amgel is shown in Table 1.1.
An important aspect of these experiments is the time scale. Many studies have
examined various types of membrane activity and cellular motility in response to growth
factors, but most examine a time course immediately following addition of factor. I have
observed an 8 hour induction period for the full growth factor migratory effect. For this
reason, I not only examined the specific motility parameters of speed and persistence time
on this time scale, but also the underlying biophysical parameter of membrane extension.
1.7
Thesis Overview
Early detection has resulted in a significant reduction in the mortality rate associated
with cancer (Boring, Squires et al. 1993). The ability to inhibit migration of tumor cells
would have a profound effect on the mortality rate due to cancers of the breast, prostate,
and other organs. The key to such prevention is the identification of early steps in the
complex processes leading to metastasis.
One such process is the induction of cell motility signaled by growth factor
receptors. I have defined one physical property of cell locomotion that is altered by EGFRmediated signaling. The mechanism characterized here is likely to be utilized by many
growth factor receptors.
This identification of a critical physical alteration and its
underlying molecular mechanism may present a unique target for therapeutic intervention.
The overall objective of this thesis is to examine epidermal growth factor mediated
cell motility. Previous studies involving growth factors and cell motility have generally
looked at population behaviors rather than individual cells, using assays such as the
Boyden chamber, wound healing, or cell scattering (Chen, Gupta et al. 1994; Cha, O'Brien
et al. 1996; Klemke, Shuang et al. 1997) While these approaches can yield data on overall
cell movement responses, they do not give information about characteristics of cell
movement that bear more directly on mechanistic issues. Thus, I have used individual cell
tracking to determine the parameters of speed and directional persistence time, along with
the fraction of cells locomoting. And I have done so over a range of substratum protein
concentrations, in order to discern possible influences of the substratum on EGF effects. I
have examined NR6 cells, mouse 3T3 derivatives lacking endogenous EGFR, that can be
transfected with various human EGFR constructs.
These cells were tested for their
responses to saturating levels of EGF (25 nM). I examined effects of perturbing the PLC
pathway by utilizing the pharmacological agent U73122 along with the receptor truncation
mutant, c'973. In order to determine which biophysical processes may be important in the
cells migratory response to EGF, I examined interactions with the substratum. Due to the
large amount of literature devoted to examining EGF-induced ruffling and membrane
effects, I chose to concentrate on membrane extension.
Amgel
Matrigel
collagen
460
collagen IV
380
450
entactin
250
120
heparan sulfate
10
25
laminin
130
810
tenascin
75
(all concentrations are listed in mg/ml)
Table 1.1:
Components of the substratum Amgel compared to the commercially
available Matrigel.
Y*
622-
-992-
TM
644
2
TM
K72 L
I
N
A
S
E
957 -
R
E
G
Y*
-
1086
3
U
L
A
T
R*
y
1186
- 1068
Y*
_
Y
-
1114
-
1148
-1173
1186-
Figure 1.1: Structure of the EGF-receptor. The regulatory tyrosines are labeled. The
c'973 truncation mutant serves to eradicate all regulatory tyrosines.
AMP
XCa++
PI3-K
A8h
siC
G-actin
ras (Grb2/sos)
F-actin
________________
membrane activity
adhesion
Figure 1.2: The EGFR-mediated signaling cascade implicated in the motile response.
Lamella extension
and adhesion
Rear release and
retraction
Regulatory
Components
CDC42
rac
rho
PLCy?
PI-kinases
Regulatory
Components
Ca++/calcineurin
rho
tyrosine kinase ?
Effectors
actin modifying proteins(AMP)
(binding, crosslinking,
Structural
Components
myosin II (contraction)
severing, capping)
focal adhesion components
Figure 1.3: Physical components of cell migration. Adapted from Lauffenburger and
Horwitz, 1996.
Chapter 2: Materials and Methods
2.1 Materials
MEMa, sodium pyruvate, non-essential amino acids, penicillin/streptomycin, Lglutamine, geneticin/G418 sulfate, trypsin EDTA, DPBS, fetal bovine serum, and
recombinant hEGF were all obtained from Gibco BRL (Grand Island, NY). Dialyzed fetal
bovine serum (dFBS) was from Sigma (St. Louis, MO). Tissue culture flasks and 35 mm
suspension culture dishes were obtained from Coming (Cambridge, MA). U73122 and
U73343 were from BIOMOL Research Labs, Inc. (Plymouth Meeting, PA).
2.2
Cell Lines and Culture
The NR6 cells expressing EGFR constructs were generated as described previously
(Chen, Gupta et al. 1994). Briefly, the full length EGFR cDNA was derived from human
placental isolate (Welsh, Gill et al. 1991) and the c'973 mutant is an EGFR in which a stop
codon was introduced downstream of the 973 codon (Chen, Gupta et al. 1994).
This
construct was introduced into NR6 cells, 3T3 derivatives which are devoid of endogenous
EGFR (Pruss and Herschman 1977), using retroviral-mediated transduction (Wells and
Bishop 1988).
Both cell lines were routinely passaged in MEMa supplemented with 7.5% FBS
and 350pg/ml G418 to maintain plasmid selection. Pre-confluent 75 cm 2 tissue culture
flasks were split 1:5 every two days. Cells were frozen in MEMa supplemented with 10%
DMSO and 30% FBS.
2.3
Substratum Preparation
Non-tissue culture suspension dishes were coated with varying levels of Amgel.
Amgel is a biological extracellular matrix derived from human placental amniotic
membranes containing type IV collagen, laminin, entactin, tenascin, and heparan sulfate
proteoglycan (Table 1.1) (Siegal, Wang et al. 1993). It does not contain detectable levels
of EGF, TGFoc, or PDGF, all of which could introduce confounding signals in our assays.
Various concentrations of Amgel in PBS (0.014 to 48 pg/ml) were incubated in 35
mm dishes at room temperature for 1 hour and 15 minutes. The solution was aspirated,
and the plates were blocked for 1 hour with 1% BSA. The plates were then washed with
PBS and stored in PBS at 4*C for up to 10 days.
2.4
Cell Migration Assays
Experiments were conducted in 35 mm non-tissue culture plates that had been
coated with Amgel. Approximately 4000 cells were seeded in normal growth medium and
allowed to attach overnight. The low cell density was chosen to minimize any cell/cell
interactions for the course of the experiment. The cells were then switched to serum-free
medium for 24 hours to induce quiescence.
For control experiments cells were then
switched to MEMao medium without bicarbonate, but with 20 mM Hepes (pH 7.4) for the
air atmosphere, 1% BSA, (M/H/B) and 1% dFBS. For experiments involving growth
factor, a saturating concentration of EGF (25 nM) was also added. Due to the high EGF
concentration and low cell density, depletion of growth factor over the course of the
experiment is avoided. Cells were then allowed to incubate for 8 hours in a humidified,
non-C02 environment before tracking.
The 8 hour incubation is due to the induction effects caused by the addition of
growth factor. The average mean squared displacements of all the cells for subsequent 30
minute time intervals were examined. The resultant data is shown in Figure 2.1. It is clear
that migration increases over a period up to 8 hours after addition of growth factor. For
this reason I pre-incubated the cells for 8 hours in EGF before I began capturing centroids
for analysis.
Motility parameters were determined by tracking single cells for a minimum of 10
hours and a maximum of 20. Cells were placed into a motorized stage (LUDL electronics,
Hawthorne, NY) and were observed using a Zeiss Axiovert 35 inverted phase contrast
microscope. A 10X objective and 10X eyepiece were used for a final magnification of
100X. A 37C environment was maintained by circulating warm water through the stage.
To prevent evaporation of the medium over the course of the experiment, mineral oil was
overlaid on the medium to provide a sealed environment that still allows gas exchange.
Cell centroid data were acquired using a nuLogic stage controller connected to a power
Macintosh running Labview software (National Instruments, Bloomfield, CT). In a given
experiment, centroids from an average of 50 cells were obtained every 15 minutes from 12
different fields. One set of experiments utilized 5 minute centroid measurements to insure
an accurate measurement of persistence time. Visual data were acquired using a Sony
CCD-IRIS camera and Panasonic AG-6040 time-lapse VCR.
Centroid positions were
transferred to a spreadsheet program and mean squared displacements (<d 2 >) over time
were calculated using the method of non-overlapping intervals (Dickinson and Tranquillo
1993).
Aspects of cell migration can be quantified in a variety of ways both experimentally
and mathematically.
As discussed in the Introduction, methods such as the Boyden
chamber are used to give a measure of the total number of cells traversing a membrane
under given conditions. As described above, I have chosen to utilize single cell tracking as
the experimental method to describe cell motility. There are many ways to mathematically
define cell motility from the obtained centroid data, but I have chosen the parameters of
speed and persistence time to fully characterize the motility response. Speed is simply the
average distance a cell travels over time, defined by the shortest time interval examined,
either 15 or 5 minutes. Each cell had a minimum of 40 such intervals. This speed will be
the true speed of the cell if the cell travels in a straight line during the interval examined.
However, if the direction of the cell is altered significantly during the time period, the speed
would be underestimated. Mathematically the speed can be defined as
lim
< 2 >1/2
r-0
=S
"
which is the root mean square value where t represents time intervals of equal duration and
5 is the displacement of the cell centroid (Dunn 1983).
By measuring cell persistence time as well as speed, I am able to determine the
duration of directed cell movement. Persistence time is a measure of the time a cell moves
in a given direction and is defined similarly to speed as
lim
22
=P
r-40 <
2
where
4 is the difference
between the directions of the displacements during adjacent time
intervals (Dunn 1983). These definitions of speed and persistence are utilized to describe a
kinesis by generating an equation describing a persistent random walker
< d 2 (t) >= 2S 2 P[t - P(1- e-'/P)
(Dunn 1983; Othmer, Dunbar et al. 1988). The physical interpretation of persistence time
is less intuitive than that for speed. It is helpful to note that a portion of the persistent
random walk equation has the same form as the exponential autocorrelation function for
persistence time, G(t) = e- P (Alt 1989). This suggests relative magnitudes of t and P for
any given loss in correlation. For example, a 90% loss in correlation corresponds to G(t) <
0.1 which in turn gives t > 2.3P. Therefore, after 2.3 persistence times, only 10% of the
cell ensemble will remain correlated to the original direction of individual cell movement._
At a time equal to one persistence time, the ensemble will exhibit a 63% loss of correlation._
To directly measure persistence time, a measure of the angle between three centroid
positions would need to be taken. One of the analysis methods I examined did exactly that.
However, the data generated did not yield adequate information and so will not be
discussed further.
To determine speed and persistence time from the cell centroid data, I utilized a
variety of methods. One method involved utilizing the data from the entire population of
cells before a number was generated. To do this, the mean squared displacements for each
time interval were calculated for each cell and then averaged over the population. Two
options exist at this point for determination of speed and persistence.
The first method
(Method 1) involves performing a two parameter fit to the persistent random walk equation.
The second method (Method 2) involves utilizing the definition of speed at the shortest time
interval to calculate speed. This assumes the cell shows no changes in direction during that
time interval. The speed is inserted in the persistent random walk equation, and persistence
time is determined by performing a single parameter fit on the equation. As long as the
assumptions are valid, it is preferable to perform a single rather than a double parameter fit.
In addition to utilizing population data, I determined speed and persistence by
calculating individual cell parameters before averaging the population. Again, there are two
options for analysis.
The first method (Method 3) is fitting speed and persistence
concurrently and the second method (Method 4) is calculating speed based on the definition
and fitting persistence only to the persistent random walk equation. Method 3 yielded
disappointing data, with many cells resulting in an unreasonable fit to the equation. The
numbers generated using this method disagreed completely with visual interpretation of the
video tapes and plotted cell tracks (data not shown). The parameters I report in this thesis
were generated using Method 4. Not only did this method produce values of speed and
persistence that were consistent with visual interpretations, but the trends were the same as
those produced by Methods 1 and 2 and the numbers were virtually identical to the
population parameters generated by Method 1. This method also has the advantage of
generating the entire range of population parameters so that a more thorough interpretation
of alterations in parameters is possible.
Data generated by Methods 1, 2 and 4 are
compared in Appendix 2.
For the speed and persistence values presented in this thesis, the speed was
determined by the 15 minute mean squared displacement, and persistence times were
obtained by fitting these to the persistent random walk equation (Dunn 1983; Othmer,
Dunbar et al. 1988) using a non-linear least squares regression analysis. Parameters were
determined for each individual cell by fitting between 1/4 and 1/2 of the data obtained. Due
to reduced data points over time, after a certain amount of time the noise in the data was too
great to allow curve fitting.
A representative plot is shown in Figure 2.2.
Arrows
represent possible data cutoffs. Both options are fit (Figures 2.2b and 2.2c) and the
resulting parameters were examined. Parameters for choosing a fit were an R value of at
least 0.9 and at least 1/4 of the data giving that fit. Once the individual parameters were
known, histograms were generated for each condition. By examining the full histogram of
the data, more information is gathered than if the mean squared displacements were first
averaged and then fit to the equation. The mean and 95% standard error were calculated for
each histogram and reported as the speed and persistence times for each of the conditions.
95% standard error of the mean provides an estimate of the accuracy of our mean, and one
can be (1-)100%
confident that the error will be less than (stdev)(tinv --
where tiny
= f (a, n-1) is the inverse of the students t distribution, stdev is the standard deviation, and
n is the sample size.
The assumption that the cell shows no change in direction during the appropriate
time interval needed to be confirmed. Examination of the video tapes of cell movement
supported the assumption, but initial determinations of persistence time put it on a time
scale with the minimum time interval, 15 minutes. For this reason, experiments were
performed with a 5 minute interval and speed and persistence times were calculated
according to method (4).
The persistence time was not altered due to the shortened
sampling time. The speed decreased slightly, although this is probably due to a difference
in the substratum rather than a true difference in speed. These experiments were performed
many months after the original 15 minute experiments, and Amgel will degrade in that
period, altering the substratum. The individual speed and persistence measurements are
shown in Appendix 3.
2.5
Pharmacological PLC Inhibitors
For experiments involving pharmacological inhibition of the PLC pathway,
U73122 (1-(6-((17P-3-methoxyestra- 1,3,5(10)-trien-17-yl)amino)hexyl)-1H-pyrrole-2,5dione) and the inactive cogener U73343 (1-(6-((17P-3-methoxyestra-1,3,5(10)-trien-17yl)amino)hexyl)-2,5-pyrrolidine-dione) were added at 1pM (Bleasdale, Thakur et al.
1990).
The U-drugs were introduced concurrently with the EGF.
To prepare the
solutions, 5 mg of each compound was dissolved into 10.8 ml of DMSO.
1ll of the
DMSO solution was then added to each ml of M/H/B, 1% dFBS, 25 nM EGF solution.
2.6
Membrane Extension Determinations
Membrane extension experiments were performed as described for the migration
experiments with minor modifications. Cells were observed under a 32X objective in order
to observe any membrane movement. Fewer cells were examined per experiment due to
the smaller focal area. Also, each cell had to be analyzed manually so fewer data points per
condition were gathered. Individual cell images were captured using Scion Image from
video tapes with time signatures. Each cell was observed for 60 to 120 minutes at 15
minute intervals. Individual cells were outlined and then overlaid one at a time to determine
two dimensional cell protrusion and retraction areas (Figure 2.3).
120
100
80
60
40
20
-200
-200
0
200
400
600
800
1000
1200
1400
time (min)
Figure 2.1: Induction period for the onset of full EGF-induced migration. Mean
squared displacement for 15 minute intervals is calculated at each time for 36 cells. Cells
expressed WT EGFR and were plated on an Amgel coating concentration of 4.8 [ig/ml.
EGF concentration was 25 nM. EGF was added 60 minutes prior to the beginning of the
experiment. The time represents the time since image capture began.
4 104
I
I
i
I
I
I
*I
3.5 10 4
*
3 104
2.5 104
4
2 10
1.5 10 4
4
4
1 10
0
q
5000
-200
-200
--
"
-
•,
* ** **
mmlli--
0
200
400
600
800
1000
1200
1400
time interval (min)
Figure 2.2a: Mean squared displacement versus time. The time interval represents the
average of all 15, 30, 45, ..... , 1200 minute intervals during the course of the experiment.
Arrows represent possible limits of data suitable for fitting.
110 4
I
I
I
I
I
I
' I
I
~I~~I
I
'
'
I
I
'
'
II
'I
'I
I
'
II
I
'
*
'
'I
I
'
I
'
^
y = 2*681 2*m2*(mO-m2*(1-e...
Value
8000
-
S
8
Error
0.64338
NA
18.148
0.94034
persistence
R
*O
9
6000
P)
4000
S.
°
U,
Es
go
°
-
0
2000
01
100
1
*
200
1 1 1 1
300
time interval (min)
Figure 2.2b: 600 minute fit of data in Figure 2.2a.
400
500
600
6000
5000
a)
4000
8b.~
5
a)
3000
a)
C,
2000
1000
0
1
0
7
I
50
I
I
I
I
100
1I I 1 I I
I
150
I
l I I
200
I
250
time interval (min)
Figure 2.2c: 400 minute fit of data in Figure 2.2a.
I
,
a ,
I
300
i a
I
350
a
i
I
I
400
A
B
t=0
t=15
C
D
protrusion
retraction
Figure 2.3: Schematic for calculation of membrane extension. Each cell is outlined at
15 minute intervals using Adobe Photoshop (a, b). The outlines are then overlaid, with the
later time represented in gray. Protrusion area (c) and retraction area (d) are filled in and
the resultant area is calculated utilizing Scion Image.
Chapter 3: Results
3.1
EGF Increases Individual Cell Locomotion
We first wanted to confirm a stimulatory effect of EGF on individual cell
locomotion (Matthay, Thiery et al. 1993; Chen, Gupta et al. 1994; Wilson, He et al. 1994;
Maldonado and Furcht 1995). Figure 3.1 shows the effect of 25 nM EGF on behavior of
NR6 cells expressing the wild type (WT) EGFR at an Amgel coating concentration of 0.48
g/ml. As can be seen in this figure, the cells in the presence of EGF show a highly
polarized morphology, generally more indicative of movement, whereas the untreated cells
are more isotropically spread. EGF treatment causes cells to move 1 or 2 cell lengths in
this one hour time period, while untreated cells remain almost stationary despite exhibiting
membrane extension activity.
A first way to quantify the EGF effect is by the fraction of cells that are stimulated
to locomote. Here, I have defined non-motile cells as cells that never move one cell length
(< 50 pm) in a single hour over the 20 hour course of an experiment. Motile cells move at
least one cell length (> 70 pm) in one hour at some time during the experiment. I have also
assigned an ambiguous movement condition between motile and non-motile, representing
the range of average cell lengths The results for these calculations are shown in Table 3.1.
It is very clear from these data that EGF increases cell motility under these conditions, with
80% of the treated cells being counted as motile compared to only 4% of untreated cells.
Other than the presence of EGF, these experiments were performed identically.
3.2
Full EGF Motility Response Requires an Induction Time
The EGF-induced cell motility requires time for the full response. Upon addition of
EGF, cells show a profound morphological change, rounding and showing a large increase
in membrane activity (Welsh, Gill et al. 1991; Chen, Murphy-Ullrich et al. 1996).
However, this dramatic morphological response is less obvious at later times. This caused
us to think the EGF response could be changing with time, so I examined the 30 minute
mean squared displacements over the course of an experiment. Essentially, I compared the
mean squared displacement for the first 30 minutes of the experiment with the same
measurement from the last 30 minutes. The parameter slowly increased for the first 400
minutes of the experiment but plateaued for the remainder of the time (Figure 2.2). For this
reason, I only used data after an initial 8 hour incubation in EGF to calculate the motility
parameters.
3.3
Level of EGF Response Depends on Substratum Concentration
I further examined EGF motogenic effects over a range of concentrations of the
substratum Amgel, a biological extracellular matrix isolated from human amniotic
membranes. It shares many of the same components as the EH sarcoma derived Matrigel,
but does not contain detectable levels of EGF, TGFoc, or PDGF (Table 1.1) (Siegal, Wang
et al. 1993). I utilized this complex ECM instead of individual components as a closer
mirror of the in vivo situation.
Figure 3.2 shows the difference in cell tracks as a result of the substratum
concentration. Five random, individual cell tracks from each condition were superimposed
to a common starting point. These "wind-rose" plots cover three orders of magnitude of
Amgel coating concentrations from 0.048 pg/ml to 4.8 pg/ml. It can be clearly seen the
paths in the presence of EGF are much longer and result in a more disperse cell population.
In the presence of 25 nM EGF, the cells show a maximal migratory response at an
intermediate Amgel concentration. Cells not in the presence of EGF also appear to migrate
differently, albeit to a lesser extent, depending upon the Amgel concentration.
3.4
EGF Causes a Substratum Dependent Increase in Cell Speed and a
Decrease in Directional Persistence
In order to more rigorously quantify individual cell movement responses to EGF, I
calculated speed and persistence times for a range of Amgel concentrations.
Individual
cells were tracked and the displacement data were fit to a persistent random walk model as
described in Materials and Methods. This analysis allows me to observe many individual
cells and then derive population parameters from their behaviors. Figure 3.3 shows the
effect of substratum concentration on cell motility. Basal cell speed is essentially constant
at all Amgel concentrations (S, Figure 3.3a, closed circles), with a slight increase at the
highest concentration (4.8 pg/ml). However, the basal speed is significantly greater than
zero, with cells moving over 25 jim/hr.
surprising (P, Figure 3.3b,
The behavior of basal persistence is also
closed circles).
Persistence
depends upon Amgel
concentration, with an intermediate substratum concentration (0.48 ptg/ml) corresponding
to an increase in persistence time. In the presence of EGF, cell speed (S, Figure 3.3a,
open circles) increases and has a maximum at an intermediate substratum concentration.
Cell persistence time (P, Figure 3.3b, open circles), however, decreases in the presence of
EGF and has a minimum under the same conditions. Neither of these motility parameters
exhibits the same magnitude of EGF response at every Amgel concentration, suggesting
some sort of interaction between the EGF and substratum signals.
The mean cell speed and persistence times were generated from a large number of
individual cell parameters. Example histograms of these data are shown in Figure 3.4 (the
remainder of the histograms can be seen in Appendices 4 and 5). Examining the entire
population of parameters yields more information than the mean alone. For example, by
examining the speed histograms in Figure 3.4a, it is clear that EGF has its effect on cell
speed by increasing both the mean and variance of the population; this occurs at each
Amgel concentration to varying degrees (Appendix 4).
The effects on the persistence
histogram are quite different. EGF has an entirely opposite effect here, shifting the mean to
the left and causing a decrease in the variance of the population. Changing substratum
concentration also exhibits this behavior, with Amgel concentrations of 0.048 and 4.8
pg/ml causing similar shifts as compared to 0.48 jlg/ml (Figure 3.4c). Examining only the
mean data limits the physical understanding of how EGF or the substratum is altering the
relevant parameter.
If the only effect EGF had was in moving the population, it is
reasonable to assume the same signaling pathway is being utilized in a magnified fashion.
By altering the variance of the population, EGF could serve to modify the signaling
pathway.
3.5
EGF Causes an Increase in Path Length and Dispersion
The combination of the motility parameters, S and P, can yield important physical
understanding of the system. Examining the product, SP, with units of length, yields
information about the path length of the cells. S2 P, with units of square length per unit
time, will suggest a cell diffusion constant, or an idea of the final distribution of the cells.
Figure 3.5 shows the comparison of these parameters. The path length is constant over
this range of substratum concentrations both with and without EGF.
However, the
presence of EGF causes an increased path length (SP) and final distribution (S2 P) as
compared to basal. If these numbers are compared with the rose plots in Figure 3.2, a
physical interpretation becomes clear. Even though EGF causes the cells to turn more
frequently through a decreased persistence time, their final distribution away from their
starting point is much greater.
3.6
PLC Abrogation Reduces EGF Effect
It has been suggested in previous studies that the EGF-induced migration pathway
is regulated through the PLCy signaling molecule (Chen, Gupta et al. 1994; Chen, Xie et
al. 1994). Here I wanted to specify this pathway as well and investigate what effects
abrogation of the PLC pathway would have on EGF-induced migration on Amgel. To do
this I utilized the pharmacological agent U73122 and its inactive cogener U73343, along
with the receptor truncation mutant, c'973.
The receptor truncation mutant c'973 was utilized to prevent any interaction of the
regulatory region of the EGFR with downstream signaling molecules such as PLCy. The
migratory behavior of these cells is very similar to that of the WT expressing cells. Figure
3.6 a and b show how basal speed and persistence of the c'973 expressors compares with
the WT expressors over the range of substratum concentrations. The truncation mutant
exhibits a different basal motility than the WT, with an increased speed at lower Amgel
concentrations, and a decreased speed at the high concentration. The basal persistence of
the truncation mutant is decreased at each amgel concentration, but it shows the same trend
as WT expressing cells with a slight increase of persistence time at the intermediate Amgel.
The effect of saturating levels of EGF on the c'973 expressing cells is greatly reduced
compared to that of the WT expressing cells (Figure 3.6 c, d, e and f). While the speed of
the c'973 expressors is increased at each amgel concentration with the addition of EGF
(Figure 3.6 c), the response is well below that of the WT expressors (Figure 3.6 d). The
effect becomes even less when the increase in speed due to EGF is compared to the
respective basal response (Figure 3.7a).
EGF also alters the persistence time in c'973
expressors in a similar fashion to WT expressors. Where there is a maximum in the basal
case there is a minimum in the stimulated case (Figure 3.6 e, compare with Figure 3.3 b),
although in this case the two curves overlap, exhibiting similar persistence times. Again,
the c'973 response is diminished compared to the WT response to growth factor (Figure
3.6 f).
The pharmacological agent U73122 has been shown to have activity against PLC
(Bleasdale, Thakur et al. 1990). I used this drug at a concentration of 1pM which is nontoxic to the cells but still has an effect on PLC. This concentration decreases EGF-induced
PLC activity by about half (Chen, Xie et al. 1994). The inactive cogener, U73343, was
also used at 1pM as a negative control. I investigated the effects these drugs had on the
EGF-induced motility parameters of speed and persistence time at the Amgel concentration
where EGF exhibited maximum effectiveness (0.48 gg/ml Amgel). As shown in Figure
3.7, the presence of U73122 decreased the EGF response for speed but U73343 showed
no effect (data not shown).
Persistence was affected only minimally. In the U73122
experiment, the speed was only 179% of the baseline EGF response as compared to 312%
with only EGF. The persistence time, however was 48% of the baseline response with
only EGF giving 41% of baseline.
The extent of these changes reflects the partial
inhibition of induced PLC activity.
3.7
Membrane Extension is Increased by EGF
It has been shown previously that EGF stimulates lamellipod extension in
adenocarcinoma cells (Segall, Tyerech et al. 1996) and induces membrane ruffling in A431
cells (Diakonova, Payrastre et al. 1995).
However, the long-term effect of EGF on
membrane activity has not been fully elucidated. I chose to examine membrane extension
on a time scale relevant to EGF-induced cell migration; >8 hours after addition of EGF.
Each protrusion measurement was normalized to the overall cell area to give a number
proportional to the total spread area of a cell. At short times (<1 hour), cells treated with
EGF were observed to round up and exhibit a great deal of membrane activity.
After
approximately an hour, however, this membrane activity slowed, the cells re-spread, and
only then began migrating.
At each Amgel concentration where cell speed increased in response to EGF, there
was also an increase in membrane extension, although to a much lesser extent (Figure 3.8).
While cell speed increases by 311% (0.48 gg/ml Amgel coating concentration), protrusion
rate increases only 137%. Based on the DiMilla model (DiMilla, Barbee et al. 1991) the
rate of extension and cell speed should be directly proportional. This suggests membrane
extension plays some role in EGF-induced cell motility, but other factors are involved as
well.
The effect of U73122 on membrane extension at the Amgel concentration with the
maximal EGF motility response was also examined. At a substratum coating concentration
of 0.48 glg/ml, the membrane protrusion rate as increased by EGF was not greatly affected
by the presence of U73122 (Figure 3.9a). It is important to note the cell area was affected.
The treatment of these cells with EGF caused a dramatic decrease in average cell area. The
cells that were treated with the combination of EGF and U73122 exhibited similar cell areas
to that of untreated cells (Figure 3.9b).
3.8
Membrane Extension Correlates with Cell Persistence in the Presence
of EGF, but with Cell Speed without EGF
To determine which specific motility parameter was affected by the change in
membrane protrusion rate, speed and persistence time were examined as functions of
protrusion rate both with and without EGF (Figure 3.10). Basal speed shows a very clear
similarity with the dependence of protrusion rate on Amgel coating concentration.
However, the overlay plot of persistence and protrusion rate versus Amgel concentration
shows no apparent similarities between the two dependent variables (Figure 3.10 a, b).
The opposite correlation is true in the case of treatment with EGF. When
persistence time and protrusion rate are plotted together versus Amgel there is comparable
dependence (Figure 3.10 c). However the comparison of protrusion rate and cell speed in
relation to Amgel show no real similarity (Figure 3.10 d).
In order to determine if membrane extension is the physical mechanism governing
EGF-induced cell motility, speed and persistence were each plotted versus membrane
extension. Figure 3.11 shows there are two distinct populations of extension rates. The
two populations consist of a set of low extension rates (0.250 ± .001) and high rates
(0.381 ± .028). Low extension rates correspond to low speeds and high persistence times
and vice versa. These trends can be more clearly seen by grouping the appropriate points
(Figure 3.11 c and d).
There is a positive correlation between speed and membrane
extension and a negative correlation between persistence time and membrane extension.
While speed and persistence form groups with membrane extension, the parameters
exhibit a broader relationship with overall cell area. There is a distinct decrease in speed
with increasing cell area. Once again the opposite trend is seen with persistence time, with
increasing persistence correlating with increasing area. These trends are seen clearly in
Figure 3.12. In the case of membrane extension, treatment with U73122 (Figure 3.11,
open square) or use of the receptor truncation mutant (Figure 3.11, open triangle) did not
serve to modify the general trend. Cell area was altered by addition of U73122 so that the
point no longer remains within the general trend (Figure 3.12, open square). Once again,
the c'973 receptor mutant did not modify the trend (Figure 3.12, open triangle).
WT
0 nM EGF 25 nM EGF
c'973
25 nM EGF
non-motile
(d2 <2500 pm 2 )
79%
2%
26%
17%
17%
19%
4%
80%
54%
intermediate
(2500 pm 2 <d 2 <4800 pm 2 )
motile
(d 2>4800 pm 2 )
Table 3.1: Percent motile cells at 0.48 jg/ml Amgel. A cell was determined to be motile
if it moved at least 70 pm (one cell length) in any given hour over the course of the
experiment. The duration of each experiment was between 12 and 20 hours. Total number
of cells examined for each case: WT, 0 nM EGF: 24 cells; WT, 25 nM EGF: 46 cells;
c'973, 25 nM EGF: 57 cells.
60 min
30 min
0 min
Figure 3.1: NR6 cells display a marked change in morphology upon treatment with 25
nM EGF. The right three panels are cells without EGF and the left three panels are cells in
25 nM EGF after at least 8 hours of incubation. Untreated cells are isotropically spread.
whereas treated cells show a more rounded morphology and generally appear more
migratory. Note the membrane extensions, cell polarity, and cell body translocation over
the 60 minute period in the EGF-treated cells. Bar is 100tpm.
I
I
I
25nM EGF
I
I
I
I
I
I
I
I
I
OnM EGF
0.048[pg/ml
Amgel
0.48pg/ml
Amgel
I
I
I
I
I
I
4.8[tg/ml
Amgel
Figure 3.2: Cell tracks were superimposed with their origins at 0,0. Each plot is a
representative 5 cell tracks from each condition. The cells all express the WT EGFR.
Tracks were after a minimum of 8 hours EGF incubation. Distance between hatch marks is
50 pm.
100
20.
0
-
0.1
1
Amgel coating concentration (Vlg/ml)
Figure 3.3a: Effect of EGF on cell speed for cells expressing the WT EGFR. Mean
speeds are reported for a variety of substratum concentrations. Error bars represent the
95% confidence interval based on the standard error of the mean.
I
I
I
i "la I
I
I
I
I
I
I
I
I
I
I
rI I ..
I
..-I
I
I I
I
WT
OnM EGF
WT
25nM EGF
20
10
I
ai
I
I I
I
i
I
I
I I fieli
I
I
a I
il
0.01
Amgel coating concentration (Vg/ml)
Figure 3.3b: Effect of EGF on persistence time for cells expressing the WT EGFR.
Mean persistence times are reported for a variety of substratum conditions. Error bars
represent the 95% confidence interval based on the standard error of the mean.
0.7
0.6
0.5
0.4
0.3
0.2
-50
0
50
100
150
200
speed (pm/hr)
Figure 3.4a: Effect of EGF on population histogram for cell speed at 0.48 tg/ml
Amgel. The number of cells with a given speed is divided by the total number of cells in
the population to give the normalized number.
0.4
'
I I I
I
I
I I I
I I I
I l I
II l
m
0.48g/ml Amgel
25nM EGF
0.3
0.2
SOnM EGF
*
1
0
-50
-~
~~
~~
a I I I
II I a
-0.1
I
0
50
100
150
-
200
250
persistence (min)
Figure 3.4b: Effect of EGF on population histogram for persistence time at 0.48 glg/ml
Amgel. The number of cells with a given persistence is divided by the total number of cells
in the population to give the normalized number.
0.7
0.6
-
-'
0.5
I
0.4
0.3
0.48 pg/ml
S0.048
pg/ml
4.8 pg/ml
-0.1
0
50
100
150
200
250
persistence (min)
Figure 3.4c: Effect of substratum on population histograms for basal persistence. The
number of cells with a given persistence is divided by the total number of cells in the
population to give the normalized number.
I
I
~ 1.~I
--
............ . ...
I IIII
1
13
10
OnM EGF
5
I
U
0.01
0.1
.I
.I
I
mi l
1
I
I
10
Amgel coating concentration (pig/ml)
Figure 3.5a: Effect of EGF on cell path length, SP. Individual cell path lengths are
calculated and the mean for each condition is reported. Error bars represent the 95%
confidence interval based on the standard error of the mean.
.1
0
-e
5
0
0.01
0.1
1
Amgel coating concentration (pg/ml)
Figure 3.5b: Effect of EGF on cell dispersion, S2P. Individual cell dispersions are
calculated and the mean for each condition is reported. Error bars represent the 95%
confidence interval based on the standard error of the mean.
- I
I
I
I-
I
-
I
OnM EGF
c'973
WT
K
A
.
0.01
. . ..
...
.I
0.1
. . . . . .. . . . . . ..
..
-
I
1
10
Amgel coating concentration (jig/ml)
Figure 3.6a: Effect of receptor truncation on basal speed. Error bars represent the 95%
confidence interval based on the standard error of the mean.
.
..........
.......
OnM EGF
I
,
WT
4t--
c'973
100
.1
.01
0.01
' ' ' '"~
Amgel coating concentration (pg/ml)
Figure 3.6b: Effect of receptor truncation on basal persistence. Error bars represent the
95% confidence interval based on the standard error of the mean.
-
c'973
25nM EGF
30
OnM EGF
20
A
10
fr•I
' ' ' ""
0
0.01
0.1
1
10
Amgel coating concentration ([pg/ml)
Figure 3.6c: Effect of EGF on cell speed for cells expressing the c'973 EGFR. Mean
speeds are reported for a variety of substratum concentrations. Error bars represent the
95% confidence interval based on the standard error of the mean.
100
_
_ _____
_______
_______
25nM EGF
WT
80
9-7
20
A
l
.1
I |i|
s
|
0.01
Amgel coating concentration (Vtg/ml)
Figure 3.6d: Effect of receptor truncation on EGF-induced speed. Error bars represent
the 95% confidence interval based on the standard error of the mean.
I
Il~r
~_rr
_ _ I I_ _
c'973
25nM EGF
OnM EGF
10
nI
ai
I
l,
,I
.
I
I
B
RI |•
I
l
•
•
i
II
0.01
Amgel coating concentration (VLg/ml)
Figure 3.6e: Effect of EGF on cell persistence time for cells expressing the c'973
EGFR. Mean persistence times are reported for a variety of substratum concentrations.
Error bars represent the 95% confidence interval based on the standard error of the mean.
.
-
1.
I..
25nM EGF
c'973
WT
10
0.01
0.1
1
1
Amgel coating concentration (pg/ml)
Figure 3.6f: Effect of receptor truncation on EGF-induced persistence. Error bars
represent the 95% confidence interval based on the standard error of the mean.
I
350
0.48 jg/ml Amgel
11111111111-
312%
300
250
200
T 179%
150
141%
100
50
0
baseline
WT
25nM EGF
WT
25nM EGF
U73122
c'973
25nM EGF
Figure 3.7a: Effect of abrogating the PLC pathway on cell speed at 0.48 plg/ml Amgel.
WT and WT with U73122 (1 IpM) are compared to basal motility of cells expressing the
WT EGFR. c'973 is compared to basal motility of cells expressing the c'973 EGFR.
Error bars represent the propagated 95% confidence interval based on the standard error of
the mean.
120
_
0.48 Vg/ml Amgel
100
a)
82%
C,,
U,.
C.
a.)
Gn
a.)
T 41%
U,
T 48%
20
n
baseline
WT
25nM EGF
WT
25nM EGF
U73122
c'973
25nM EGF
Figure 3.7b: Effect of abrogating the PLC pathway on cell persistence time at 0.48
jg/ml Amgel. WT and WT with U73122 (1 pM) are compared to basal persistence of cells
expressing the WT EGFR. c'973 is compared to basal motility of cells expressing the
c'973 EGFR. Error bars represent the propagated 95% confidence interval based on the
standard error of the mean.
0.5
0.45
WT
25nM EGF
0.4
0.35
0.3
WT
OnM EGF
0.25
S0.1
0.2
.
0.0
Amgel coating concentration (gtg/ml)
Figure 3.8: Effect of EGF on membrane protrusion rates of WT EGFR expressing
cells. Protrusion areas were divided by the overall cell area to produce a normalized value.
Error bars represent the 95% confidence interval based on the standard error of the mean.
0.5
-0.48I
g/ml
I
Amgel
.
0.48 pgg~ml Amgel
I
I
1
0.4
.
1
0.372
10.344
0.3
1 0.250
0.2
0.1
_ l .
WT
OnM EGF
WT
25nM EGF
WT
25nM EGF
U73122
Figure 3.9a: Effect of utilizing U73122 (1 pM) to abrogate PLC on EGF-induced cell
extension rates. U73122 was used at a submaximal concentration to avoid cell toxicity.
Error bars represent the 95% confidence interval based on the standard error of the mean.
1400
1200
I I I I
I
T
I
I
1189
1189
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I,
I
.
0.48 /g/ml Amgel
1139
1000
CI
M
el
.
r(U,.
800
T 702
600
== -k
U===
400
200
0
WT
OnM EGF
WT
25nM EGF
WT
25nM EGF
U73122
Figure 3.9b: Effect of utilizing U73122 (1 pM) to abrogate PLC on EGF-induced
changes in cell area. U73122 was used at a submaximal concentration to avoid cell
toxicity. Error bars represent the 95% confidence interval based on the standard error of
the mean.
I
I
I
I
ll
I
I
I
I
I
I
I
I
I
I
0.5
i l
OnM EGF
7 0.45
40
o
30
-
0.4
,-=-
speed
0.35
20
0.3
10
protrusion
rate
0 a 1-
-
/
--- l
- 0.25
0
0.01
0.2
0.1
1
14
Amgel coating concentration (pg/ml)
Figure 3.10a:
Comparison of basal speed and protrusion rates for WT EGFR
cells.
Error
bars represent the 95% confidence interval based on the standard
expressing
error of the mean.
0.5
50
OnM EGF
0.45
40
-
-
persistence
0
0.4
0.35
0.3
10
0
0.C)1
/
protrusion
.
rate
-
f.....
Ic.
0.25
.1••
- - - ----
0.2
Amgel coating concentration (pg/ml)
Figure 3.10b: Comparison of basal persistence time and protrusion rates for WT EGFR
expressing cells. Error bars represent the 95% confidence interval based on the standard
error of the mean.
100
0.45
OR
0.4
o
0.35
0.3
20
0
0.01
0.25
0.2
0.1
1
Amgel coating concentration ([g/ml)
Figure 3.10c: Zomparison of EGF-induced speed and protrusion rates for WT EGFR
expressing cells. Error bars represent the 95% confidence interval based on the standard
error of the mean.
0.45
0
0.4
0.35
0.3
10 0
E
0.25
0.2
0.01
0.1
1
10
Amgel coating concentration (pg/ml)
Figure 3.10d: Comparison of EGF-induced persistence time and protrusion rates for
WT EGFR expressing cells. Error bars represent the 95% confidence interval based on the
standard error of the mean.
0
O
r
a
WT, no EGF
WT, 25nM EGF
WT, 25nM EGF U-drug
c'973, 25nM EGF
120
100
80
60
A
-
40
0.1
0.2
0.3
0.4
0.5
0.6
membrane extension per 15 minutes
Figure 3.11a: Dependence of cell speed on membrane extension rate. The general
experimental conditions are explained in the legend. Error bars represent the 95%
confidence interval based on the standard error of the mean.
0
10
0
0.1
0.2
0.3
0.4
0.5
0.6
membrane extension per 15 minutes
Figure 3.11b: Dependence of cell persistence on membrane extension rate. The general
experimental conditions are explained in the legend. Error bars represent the 95%
confidence interval based on the standard error of the mean.
I
80
I
70
avg = 59.9
60
50
40
30
avg = 28.6
20
10
0
low
high
membrane extension rate
Figure 3.11c: Populations of speed dependence on membrane extension rate. Error
bars represent the standard deviation.
I_________
~_____
RT
avg = 32.7
avg = 19.6
low
high
membrane extension rate
Figure 3.11d: Populations of persistence dependence on membrane extension rate.
Error bars represent the standard deviation.
0
O
O
WT, no EGF
a
c'973, 25nM EGF
WT, 25nM EGF
WT, 25nM EGF, U-drug
120
100
S
I
' I
'
'
'
I
I
I
'*
*
I
a
I
I
-
80
±
60
IL
k
40
20 I
I
400
I
I
600
.
.
I
800
I
I
I
1000
overall cell area (Mrn
1200
I
a
a
I
.
1400
2)
Figure 3.12a: Dependence of cell speed on overall cell spread area. The general
experimental conditions are explained in the legend. Error bars represent the 95%
confidence interval based on the standard error of the mean.
10 --
400
600
800
1000
1200
1400
overall cell area (pm 2)
Figure 3.12b: Dependence of cell persistence on overall cell spread area. The general
experimental conditions are explained in the legend. Error bars represent the 95%
confidence interval based on the standard error of the mean.
Chapter 4:
Discussion
Activation of receptors with intrinsic tyrosine kinase activity, such as those for
EGF, PDGF, and IGF-1, results in enhanced cell motility (Bornfeldt, Raines et al. 1994;
Chen, Gupta et al. 1994; Kundra, Escobedo et al. 1994).
There is evidence that this
growth factor enhanced migration behavior requires the PLC'Y pathway (Chen, Xie et al.
1994). Other biochemical signaling pathways have also been implicated. However, the
biophysical characteristics of specific, EGF-induced motility have yet to be well defined.
Our work here describes the EGF-induced motility behavior of NR6 fibroblasts transfected
with EGFR constructs.
A time dependence for the EGF-induced migration was described in Section 3.2.
EGF has been shown previously to cause an acute increase in membrane activity and
lamellipod extension (Ridley, Paterson et al. 1992; Segall, Tyerech et al. 1996).
These
observations have all been at relatively short times after the addition of EGF. I observed a
dramatic morphological change in the cells upon addition of EGF that decreased with time.
Immediately after EGF addition, the cells retract, become less spread and exhibit membrane
blebbing activity, in agreement with earlier scanning electron microscopy studies of acute
cell responses (Welsh, Gill et al. 1991). After a few hours in the continual presence of
EGF, the cells once again are spread. The experiments were performed under supersaturating conditions of EGF and extremely sparse cell concentrations, so the decrease in
membrane activity is not due to depletion of factor. Even though this obvious membrane
activity slows down, the motility parameters take hours to increase to their final EGFinduced values (Figure 2.2). This suggests that the physical parameters involved in the
motility response need to be examined at long times after EGF or other factor stimulation of
cells. This gradual increase in motility insinuates the involvement of new protein synthesis
or gene induction in the EGF-induced response.
Temporal and spatial regulation of growth factors could then be very important in
cellular responses in physiological and pathological situations. If the cell necessary for the
response is not activated in time, it could result in an abnormal response. The role of EGF
in wound healing is specific enough that it must be regulated both temporally and spatially
in keratinocyte populations relevant in burn wound healing (Wenczak, Lynch et al. 1992).
However, previous in vitro investigations of growth factor induced motility have not taken
into account the possible time effects. Examination of ruffling activity of cells treated with
EGF is generally on the time scale of minutes.
Maximum increase in a mammary
carcinoma cell area due to extension is seen within 5 minutes of EGF treatment (Segall,
Tyerech et al. 1996). The manner in which this correlates with long-term membrane
activity and cell migration is unknown. The reported association of PLCy-1 with EGFR in
membrane ruffles disappears 5-10 minutes after treatment with growth factor (Diakonova,
Payrastre et al. 1995). The manner in which these phenomena translate to continuing
migration would be of interest for eventual treatment strategies.
As shown in Figures 3.1 and 3.2, the addition of EGF can indeed increase NR6
fibroblast migration. It is apparent in both the general migratory morphology (Figure 3.1)
and in the total path length and dispersion of the EGF treated cells (Figure 3.2). Table 1
also shows the difference in the number of migrating cells, with fully 80% of the EGF
treated cells being considered migratory as compared to only 4% of the non-treated case.
This would imply that EGF does not only affect a few cells, increasing the overall average
migratory behavior, but rather the entire population. This is confirmed by examining the
histogram data shown in Figure 3.4a. The definition I used for motile translates to a cell
speed of around 50-70 nm/hr. By examining the histogram for cell speed with and without
EGF, it is obvious the majority of the 0 nM population lies below this value while the 25
nM population is above this value. Also, it is important to note that EGF does not serve
purely to increase the mean cell speed, but also to broaden the population of speeds. This
effect is seen in every population for the cell speed. In the case of persistence time, EGF
has the opposite effect at this amgel concentration (Figure 3.4b). The mean is decreased,
and at the same time the cell population spread is reduced. This is not the case for every
substratum condition. In fact, changing the substratum concentration yields the same
effects. In the persistence case, the difference is apparently a substratum effect on the basal
response more than an EGF effect. EGF certainly alters the cell persistence, but it is only
distinct at the intermediate Amgel concentration. The character of the histograms implies
EGF causes all the cells in a population to be modified, not just a certain subpopulation. If
that were the case, the histograms should show some kind of discontinuity.
The
histograms for the remainder of the conditions utilized in this study can be found in
Appendices 4 and 5. If the mean were the only information available, there could be
unobserved subpopulations.
The traditional techniques for examining motile behavior of epithelial and fibroblast
cells include the Boyden chamber, wound closure, and cell scattering or under-agarose
assays. While these assays all give a general idea of the migratory potential of a cell, they
do not describe fully the migratory parameters of the cells. There is a tendency to retrieve
information that is dependent upon the assay. Instead, a rigorous measure of migration
parameters could be used to compare results from many researchers.
Fibroblast and
epithelial cells migrate as persistent random walkers. Over a short time period they will
appear to move in a straight line. However, as time increases they exhibit motion similar to
Brownian motion of inert particles (Lauffenburger and Linderman 1993). Two important
parameters to describe this motion are speed and persistence time as described in the
Introduction. (It is after a few multiples of the persistence time that cell movement takes on
characteristics of Brownian motion.)
I have utilized videomicroscopy single cell tracking
on the cells described above to specifically determine these parameters.
Previous work states simply that EGF causes cell migration (Blay and Brown 1985;
Matthay, Thiery et al. 1993; Wilson, He et al. 1994). There is no specificity, implying that
the simple addition of EGF will cause cells to be more motile.
I have shown here,
however, that the EGF response is very dependent upon the substratum concentration. The
EGF response is not simply a linear addition at each Amgel concentration. Rather, there is
a measured response, dependent upon the concentration of the underlying substratum
(Figure 3.3). The biphasic dependence of cell speed on cell-substratum adhesiveness has
been seen previously for integrin mediated migration on extracellular matrix proteins
(Palecek, Loftus et al. 1997), but this is the first time it has been observed for growth
factor induced migration. The specific parameters of cell speed and persistence time were
examined under varying substratum concentrations and saturating levels of EGF (25 nM).
Cell speed, when analyzed using the persistent random walk equation and data taken at
short time intervals (15 min), is a measure of how fast a cell moves with time, rather than
just a measure of final position as is the case with the Boyden chamber assay or cell
scattering. Cell persistence time is a measure of how far a cell will travel before altering
direction. At an intermediate Amgel coating concentration (0.48 pg/ml), treatment of WT
expressing cells with EGF caused an increase in cell speed and a concomitant decrease in
persistence time. At both higher and lower substratum concentrations, the effect of EGF is
more limited or absent. These data suggest that growth factor induced cell motility is not
accomplished by simply altering cells substratum adhesiveness, but increases cell intrinsic
properties required for motility.
Cell motility requires both actin and myosin. Actin-based protrusion must make a
productive contact with the substrate (binding of integrins to the extracellular matrix) in
order for a cell to move forward. If adequate contact is not made, retrograde flow of actin,
dependent upon myosin activity, causes retraction of the membrane (Schwarzbauer 1997).
Traction force is stronger than protrusion force, and is probably more important in strongly
adhered cells (Mitchison and Cramer 1996). Protrusion in this case would be useful as a
mechanism for the placement of an actin "track" for the generation of the traction force.
Asymmetry of the cell is also an important property to allow net locomotion. Myosin I is
present in initial protrusions of migrating fibroblasts (Conrad, Giuliano et al. 1993), but
myosin II localizes at the rear of the cell (Huttenlocher, Sandborg et al. 1995).
Cortical
actin is anchored to the external matrix more strongly at the front of the cell than at the rear
in fibroblasts, and the leading edge is the site of preferential attachment of the cytoskeleton
to crosslinked glycoproteins, including integrins (Sheetz 1994).
I examined the biophysical ramifications of the EGF-effect in an attempt to
determine what specific mechanisms are at work. Differences between basal and EGFinduced cell movement are also apparent in protrusion rates. Examining plots of speed,
persistence, and protrusion versus Amgel yields a variety of information. The basal speed
parallels protrusion with Amgel (Figure 3.10 a). This suggests protrusions are limiting in
basal cell motility; cell speed depends directly upon rate of membrane extension. The EGFinduced response is quite different, however.
In this case, persistence parallels the
protrusion dependence on Amgel (Figure 3.10 d). This switch in correlation is possibly
due to a switch in internal mechanism. In the case of basal cell motility, the driving force
for migration could be distinct for that of EGF-induced motility. It has been shown that
PLCy activity is required for EGF-induced migration but not for basal (Chen, Xie et al.
1994). This difference in signaling requirements could translate into a different physical
response. Possibly basal migration operates purely through integrin mediated migration,
utilizing the more direct links to the cytoskeleton involved in focal adhesions, as well as
signaling molecules such as PKC and FAK.
There are two distinct populations in the comparisons of speed and persistence
directly to membrane extension. In both cases, there is a set of low protrusion values and a
set of high protrusion values. Although there is not a true spread in the data, general trends
are still apparent. There is a positive correlation of speed and a negative correlation of
persistence with membrane extension. Both results are consistent with physical intuition.
As membrane extension increases, overall cell speed also increases. The more membrane
successfully extends, the more it is able to displace. Concurrently, with more membrane
extension, there are increased chances for the cell to change direction and its persistence
time should decrease. Perturbation of the PLC pathway through the use of U73122 or the
c'973 EGFR does not alter the interpretation. While these conditions altered the EGF-
induced speed, they altered it in such a way as to stay within the trend of extension. This is
to be expected as long as perturbation of the PLC pathway does not serve to divert the
EGF-induced response to a different signaling pathway that utilizes an alternate physical
mechanism to affect cell migration.
An additional measure of physical mechanism is cell area. Cell area appears to be a
more appropriate independent variable than membrane extension.
There is a greater
variation in cell area than in membrane extension, resulting in a clearer trend over the range
of substratum concentrations tested.
These results are also consistent with physical
intuition. As cell area increases, the cell speed decreases and persistence time increases
(Figure 3.12). The adhesiveness could be increasing with the area, or there could be a
decrease in asymmetry. Either case would result in a low cell speed. The increase in area
also corresponds to a high persistence. This agrees with the idea of increased adhesion.
Decreased asymmetry would be reasonable if it were due to the high adhesion.
implies an additional biophysical mechanism underlying cell migration.
This
Earlier it was
stated the increase in cell speed due to EGF did not correspond exactly to the increase in
membrane activity.
An additional mechanism for migration, such as an alteration of
adhesion strength, would be consistent with these observations.
Treatment with U73122 causes a drastic change in cell area that does not fit with the
correlations between speed or persistence and area as shown in Figure 3.12.
Cells
expressing the c'973 EGFR, however, fall within the correlations. This suggests treatment
with U73122 may abrogate signaling in a manner different from the truncation mutant.
NR6 cells expressing the c'973 EGFR are able to stimulate the ras pathway through
transmodulation of the erbB-2 receptor (Sasaoka, Langlois et al. 1996).
U73122 could
abrogate PLC in such a way that the signaling pathway is not changed. c'973 could be
signaling through a pathway that does not have the same potential for migration whereas
U73122 causes a decrease in the WT signaling pathway.
Modification of the EGF effect was accomplished by targeting the PLCy pathway in
a variety of ways. The addition of a pharmacological inhibitor of PLC had the effect of
reducing the increased speed due to EGF (Figure 3.7a).
The effect was not completely
eradicated possibly due to the sub-maximal concentrations of U73122 used (Chen, Xie et
al. 1994).
The truncation of the receptor also causes a dramatic decrease in the EGF-
induced speed as shown in Figure 3.7a with the c'973 truncation mutant. This suggests
that PLCy is indeed involved in EGF-induced motility and may provide the cell-intrinsic
motive force. However, even untreated cells demonstrate some motility, by pathways yet
to be defined. This movement may represent integrin-mediated motility. It must be noted
that PLC activity abrogation does not significantly impair basal cell motility, but rather
EGF-induced motility (Chen, Xie et al. 1994). Furthermore, gelsolin, required for PLCy
mediated motility, is not required for basal cell movement (Witke, Sharpe et al. 1995), but
rather for EGF-induced (Xie, et al., submitted).
Both EGF and PLCy have been implicated in embryogenesis (Kurachi, Morishige
et al. 1994; Schlessinger 1997). Activated EGFR and VEGFR (Flt-1) both phosphorylate
PLCy (Hernandez-Sotomayor and Carpenter 1992; Thomas 1996), and PLCy-1 colocalizes
with EGFR and F-actin at membrane ruffles of A431 cells upon treatment with EGF
(Diakonova, Payrastre et al. 1995). Abrogation of the PLCy pathway also reduces the
PDGF-BB motile response in NIH 3T3 fibroblasts, although it has no effect on random
migration (Kundra, Escobedo et al. 1994).
The SH2 region of PLCy-1 targets it to
tyrosine phosphorylated receptors such as EGFR, and the SH3 region of the molecule
targets PLCy-1 to the actin microfilament network (Rhee and Bae 1997).
Binding of
profilin, an actin modifying protein, to PIP 2 inhibits hydrolysis of unphosphorylated but
not phosphorylated PLCy-1 (Goldschmidt-Clermont, Kim et al. 1991).
Therefore,
phosphorylation of PLCy- 1 by a receptor tyrosine kinase receptor could be the first step in
releasing PIP 2 from profilin, thereby making profilin available for binding to g-actin and
causing a local freezing of F-actin formation. The production of DAG and IP 3 would then
have their effects on PKC and Ca2+. Micromolar Ca2+ makes gelsolin sever actin and then
cap the filaments (Cunningham, Stossel et al. 1991). Gelsolin has also been shown to play
a role in the depolymerization of actin in adherent human neutrophils (Wang, Coburn et al.
1997). This localized breakdown of F-actin causes the osmotic pressure to increase,
thereby allowing extension of the plasma membrane (McCarthy, Iida et al. 1996).
PIP 2
could then be regenerated, perhaps due to rho (Chong, Kaplan et al. 1994). PIP 2 is also
generated by adhesion to fibronectin in C3H10T1/2 cells (McNamee, Inger et al. 1993).
This final state would enable re-formation of an actin network, serving to stabilize the
newly formed protrusion (McCarthy, Iida et al. 1996). Connection of the growth factor
induced PLC pathway directly with membrane protrusion is one of the purposes of these
experiments.
The addition of EGF and its effect on fibroblast migration was examined in the
parameters SP and S2P, terms which can describe the physical results of the induced
migration. SP has units of length and is a measure of the distance actually traveled by a cell
during each polarized traverse.
S2P has units of length squared over time, and is
essentially a cell diffusion constant, being a measure of dispersion of a cell population. As
seen in Figure 3.5, EGF causes an increase in overall path length, SP, and in cell
dispersion, S2 P. This could have implications in the mechanism through which cells
overexpressing EGFR are able to form metastasis and cause a poor prognosis in many
cancers. The implications for wound healing are obvious; growth factor induction would
increase the rate and extent of cellular repopulation, but the direction and final location of
cells would be dependent on chemotactic and matrix directional signals akin to axonal
growth, guidance and collapse (Komuro and Rakic 1995).
Many growth factors have shown activity in physiological and pathological
processes such as wound healing, embryogenesis, angiogenesis, the immune response,
and metastasis. Also involved is the process of cell motility which requires a variety of
biochemical and biophysical conditions be met in order to proceed. Our work here has
described the system of epidermal growth factor induced cell motility. Not only have I
examined whether migration is induced, I have made specific measurements of the absolute
level of components of migration including speed, persistence time and membrane
extension. I have also examined the role of the PLC pathway in these various biophysical
parameters.
Many growth factors including EGF, TGFa and FGF promote restitution, the
initial phase of mucosal repair characterized by a rapid migration of the epithelium
(Podolsky 1997). Some cytokines, EGF, TGF, and PDGF are also found secreted from
wound borders (Moulin 1995). Accumulation of EGF and overexpression of the EGFR in
the area of a gastrointestinal ulcer contributes to the healing process by encouraging
migration of cells from the ulcer margin, formation of granulation tissue and angiogenesis
(Konturek, Konturek et al. 1995). Application of EGF topically also causes extensive
acceleration of epithelialization in partial thickness wounds across the dorsal surface of pigs
(Nanney 1990). The initial stage in the wound healing response involves the formation of
a blot clot that serves as a provisional matrix for cell migration. In addition to the matrix,
there are also a variety of cell types in the area that are secreting growth factors and
cytokines to help direct new cells into the wound.
Some of the new cells are platelets
which then adhere, aggregate and release additional adhesive proteins and growth factors.
Neutrophils and monocytes must also emigrate into the injured tissue along with
fibroblasts. Once epithelialization has ensued, tissue formation and remodeling must occur
(Clark 1995).
The degree of EGFR expression is also correlated with the malignant
phenotype in certain epithelial tumors (Fujii, Dousaka-Nakajima et al. 1995). Indeed, the
EGFR is the receptor most commonly implicated in the pathology of all human cancers
(Khazaie, Schirrmacher et al. 1993).
Monoclonal antibodies directed against the EGFR
have been used in squamous lung cancer cases (Zumkeller and Schofield 1995).
All of
these processes have an obvious dependence upon not only cell migration but growth
factors as well. Unfortunately, the actual mechanisms governing these process have yet to
be fully elucidated. By examining EGF-induced cell migration in a systematic, mechanistic
fashion I provide a deeper understanding that can be utilized for novel therapeutic
strategies. Interventions targeting to the EGFR may inhibit only one pathological response.
If the cause of a pathology is of a more fundamental nature, such as a dysregulated motility
machinery that is perhaps initiated by EGF, then a therapy targeting that specific aspect
would be of greater use.
I have shown here that EGF-induced cell migration is not a simple effect. Not only
does EGF increase cell motility, but it does so in a specific, substratum dependent manner.
Overall cell speed is increased while the persistence time is decreased, but only at an
intermediate substratum concentration. There are morphological differences in EGF treated
cells, but much of the early membrane activity lessens by the time relevant migration
parameters are measured. The early membrane activity does not seem to correlate to a later
dependence of cell motility on protrusion.
EGF does cause an overall increase in
protrusion, but the difference in protrusion compared with either speed or persistence is
minimal. Membrane extension does, however, appear to play a more important role in
basal cell motility. Speed and persistence are more dependent upon cell area. In addition,
EGF-induced migration has an associated induction time. These data show specific effects
of EGF on fibroblast migration parameters. These patterns of cell characteristics provide
testable modulation for both the biochemistry and biology of growth factor-induced
motility.
Appendix 1: Microsoft Excel Macro for Mean Squared Displacement
Calculations
This macro was written in visual basic within Microsoft excel. It calculates the mean
squared displacements for each time interval. The 15 and 30 minute time interval
displacements are printed out as a result of the macro to aid in interpretation. The macro
also calculates superimposable coordinates for use in "wind-rose" plots of cell tracks.
'must declare these variables public so they can be used by all subroutines
Public delta, maxtime, displacementsq, driftx, drifty, mark, i, r
'subroutine to set the data sheet up
Sub first_initial_calculationsO
'user inputs
'set time increment (in minutes)
delta = 15
'set drift rate in pixels/min
driftx = 0
drifty = 0
'figure out the maximum row number to perform calculations upon and write drift
corrected columns
i=0
Do Until Cells(2 + i, "e").Value = 0
Cells(2 + i, "m").Value = Cells(2 + i, "e").Value - (driftx * delta * (i + 1))
Cells(2 + i, "n").Value = Cells(2 + i, "f').Value - (drifty * delta * (i + 1))
Cells(2 + i, "e").Select
i=i+
Loop
'calculate maximum time length
maxtime = (i - 1) * delta
Cells(3, "l").Value = maxtime
Cells(7, "l").Value = i
'calculate # of intervals for each time point
p= 1
Cells(2, "o").Value = 0
Do Until p = i
Cells(2 + p, "o").Value = Int(maxtime / (delta * p))
Cells(2 + p, "v").Value = Int(maxtime / (delta * p)) - 1
p=p+ 1
Loop
'calculate coordinates for rose plot
r=0
Do Until Cells(2 + r, "m").Value = 0
Cells(2 + r, "s").Value = Cells(2 + r, "m").Value - Cells(2, "m").Value
Cells(2 + r, "t").Value = Cells(2 + r, "n").Value - Cells(2, "n").Value
r=r+
Loop
'subroutine to calculate the squared displacments
Cells(2, "p").Value = 0
mark = 2
Do Until mark = i + 1
work
Cells(mark + 1, "p").Value = (displacementsq * ((1)
"o").Value
mark = mark + 1
Loop
End Sub
^ 2))
/ Cells(mark + 1,
'actual code to calculate squared displacement
Sub work()
iterl = 0
displacementsq = 0
rowid = 1 + mark
subtract = 2
d2sum = 0
Do Until iterl = Cells(mark + 1,"o").Value
d2sum = (Cells(rowid,"m").Value - Cells(subtract, "m"))
"n").Value - Cells(subtract, "n"))
^
2
If mark = 2 Then Cells(rowid, "q").Value = d2sum
If mark = 3 Then Cells(rowid, "r").Value = d2sum
rowid = rowid + mark - 1
subtract = subtract + mark - 1
displacementsq = displacementsq + d2sum
iterl = iterl + 1
Loop
End Sub
A
2 + (Cells(rowid,
Appendix 2:
Analysis Comparison
The average speed and persistence times calculated by each of three methods described in
Materialsand Methods are shown below. Notice the trends between experiments are very
similar, however the magnitude of the speeds for Method 1 are much greater than either
Method 2 or 4. Also, the persistence times reported in Method 1 in the presence of EGF
are not physically consistent with visual observations.
Method 1 (Population, double parameter fit)
Sped (pm/hr)
Aigld (pg/m1)
EGE (nM)
195
0.048
25
270
0.48
64
4.8
0
0.048
0.48
4.8
Method 2 (Population, single parameter fit)
gel (tg/.i)
S(M). .....
0.048
25
0.48
4.8
0
0.048
0.48
4.8
Method 4 (Individual Cells, single parameter
Amgel ( il) :
EGF (nM)
0.048
25
0.48
4.8
0
0.048
0.48
4.8
Persistence (min)
2
1
10
2
19
20
no fit
90
68
Sped (Pm/br)
77
95
54
Persistence (min)
12
9.5
14
35
35
51
9
23
10
fit)
)r
Sped (
72
90 (66 for 5 min)
46
28
29
41
Persistenco (min)
19
16 (12 for 5 min)
26
25
40
21
Appendix 3:
Individual Cell Speed and Persistence Times
Cell speeds were calculated from the 15 minute mean squared displacements as described in
MaterialsandMethods. Persistence times were then obtained by fitting the persistent
random walk equation with the calculated speed utilizing a non-linear least squares
algorithm in Kaleidagraph. All curve fits possessed an R value above 0.9.
WT EGFR Basal Motility
pgml
048 ig/mI
t0.048
pgt/I
speed
(um/hr)
persistence
(min)
speed
persistence
speed
persistence
10
13
14
26
8
31
7
23
50
39
21
36
76
33
16
90
12
31
14
52
43
35
26
19
69
11
10
22
6
18
19
19
15
6
62
15
4
6
6
31
7
7
47
23
5
10
17
5
10
17
15
6
8
9
43
25
12
40
11
77
4
218
7
49
22
27
56
18
27
30
20
21
38
10
40
16
18
38
24
31
98
16
18
24
18
11
17
56
38
15
28
45
16
10
41
29
25
67
62
18
19
30
8
40
15
49
24
116
88
47
79
8
45
50
6
69
15
29
50
40
8
84
26
41
9
127
9
42
16
7
12
78
10
9
3
8
15
22
8
13
40
3
10
8
18
12
20
13
11
13
42
18
9
44
85
15
105
56
22
90
68
35
15
EGF-Induced WT Motility
0048
0,014
.,
0A48
48
I
4.80
_I__:
persistence
persistence
(min)
14
69
78
68
50
53
73
34
64
51
38
70
51
16
18
30
10
20
8
13
8
22
8
16
16
11
8
27
8
23
11
14
21
38
16
33
18
19
14
16
29
26
38
17
30
52
51
53
30
15
10
31
8
34
10
32
5
48
18
32
15
38
27
20
7
21
11
50
16
110
107
43
80
48
21
129
67
72
60
106
75
73
104
68
86
71
66
88
50
57
68
55
32
66
127
63
28
70
85
110
70
54
41
120
62
persistence
47
134
103
102
105
72
82
115
144
132
99
85
146
102
115
73
33
108
71
97
87
64-1
116
41
84
89
131
36
66
132
40
67
55
27
113
14
18
20
4
7
4
8
4
9
5
12
5
2
16
19
44
16
7
14
6
10
5
61
17
10
5
56
43
17
40
10
22
34
6
I
s
persistence
24
34
62
22
49
97
69
56
96
88
76
26
102
39
59
61
65
47
73
17
74
107
55
65
19
22
53
62
12
53
41
57
44
50
7
26
37
26
19
4
32
6
100
28
9
3
13
2
43
16
21
16
11
17
26
4
8
9
45
8
192
25
22
18
8
22
25
12
4
58
EGF-Induced WT Motility, cont.
contpee
eed
persistence
17
42
40
8
33
37
42
28
23
45
31
10
9
59
60
17
15
11
93
52
31
47
20
22
22
30
82
35
52
44
15
6
17
26
10
27
25
36
11
16
10
22
ncnt
eed
82
87
128
67
160
59
159
94
82
19
101
cnt
persistence
persistence
5
22
17
22
3
5
3
20
14
45
12
72
18
19
54
52
41
33
10
9
53
14
17
42
77
35
23
14
20
18
3
5
28
31
3
5
104
11
38
20
42
25
21
2
56
4
22
30
16
EGF-Induced WT Motility, 5 minute intervals
0748
[
speed (pm/hr)
persistence (min)
speed (cont.)
persistence (cont.)
63
134
74
130
45
84
79
28
138
42
97
70
270
51
78
56
57
58
92
43
57
59
84
33
84
57
50
60
101
75
67
130
73
76
66
76
32
1
3
13
12
1
9
3
4
14
3
6
1
15
12
13
17
12
6
2
4
6
25
4
1
11
27
3
5
11
10
7
4
16
29
23
68
24
61
62
46
42
51
37
54
66
57
21
188
37
18
95
40
28
18
46
67
44
22
80
53
21
63
55
76
44
35
10
5
16
8
4
73
3
10
10
13
2
3
5
36
2
4
14
2
1
10
31
9
6
28
28
6
12
58
16
22
5
Basal c'973 Motility
I
A6_
speed
.1
.
.
________
-
____
seed
persistence
(min)
(um/hr)
41
94
94
32
57
68
89
12
29
17
45
25
63
42
27
44
91
58
79
21
66
69
71
57
24
59
86
8
74
55
61
51
38
25
41
53
17
32
71
___
__
_
_
I
8
71
51
20
21
66
16
47
74
13
54
40
20
36
20
54
15
57
81
19
27
133
52
13
15
57
64
6
25
9
90
23
11
60
24
31
79
59
40
___
_
7777i
]___94.___I
_14
persistence
19
9
19
7
6
3
18
9
3
27
8
22
3
12
50
29
12
22
9
13
249
22
7
8
20
5
5
41
6
30
8
6
9
13
8
32
12
5
19
£
7
,
___77
persistence
.1
18
12
7
25
10
7
8
10
8
14
6
62
11
18
25
20
7
7
6
11
14
16
11
7
14
7
8
9
17
12
8
20
7
19
7
6
12
14
6
3
28
6
17
5
8
8
6
6
7
12
5
5
5
6
7
5
8
8
4
3
35
23
12
6
21
6
7
33
4
5
7
4
6
3
10
38
3
39
Basal c'973 Motility, cont.
0.048 pg:
Il
speed
86
26
24
72
56
47
57
55
60
94
63
47
36
47
70
63
69
69
10
20
29
32
54
43
86
83
69
7
8
I
11
_______
4.8pg/nml
cent.
persistence
&
9
9
6
41
9
9
17
4
20
6
9
32
32
19
9
10
12
9
19
16
8
18
9
8
7
11
11
38
30
speed
I
J
persistence
8
14
4
4
4
14
6
11
171
6
23
6
5
43
3
A
I
£
I
EGF-Induced c'973 Motility
AS
speed
(um/hr)
90
70
86
86
55
53
36
43
34
75
20
48
76
27
32
41
66
19
92
31
24
61
127
29
98
9
7
17
43
16
21
49
59
29
23
61
52
22
38
13
39
persistence
(min)
17
6
19
18
7
21
13
23
14
8
15
35
8
43
10
10
25
38
7
23
36
24
6
9
7
79
46
17
31
59
15
11
23
16
18
3
94
21
16
14
36
Ij'
persistence
16
105
10
90
88
98
85
68
43
39
108
57
74
54
61
128
46
73
35
41
79
23
61
55
56
51
71
58
57
62
85
59
55
19
61
16
53
70
39
62
78
46
20
22
5
18
22
15
15
23
9
11
9
53
18
19
6
7
8
9
12
17
32
12
22
6
19
6
6
8
23
26
10
19
22
7.5
9
5
17
29
15
11
speed
10
9
13
13
37
47
9
27
8
27
37
7
22
7
22
10
9
14
11
12
9
47
14
26
20
16
7
8
13
16
20
14
13
11
47
15
22
9
14
55
15
I4_______
persistence
22
55
20
62
8
14
89
20
62
17
25
103
42
65
14
31
55
108
29
17
54
15
37
42
40
58
105
60
111
59
30
24
23
46
37
31
38
38
61
25
40
EGF-Induced c'973 Motility, cont.
e~it
___
~~~~~~048~
~
~
~
cet
__
__
.
1 pg~.14
_
i>
_
__
_
et
-8p/~
speed
persistence
speed
persistence
speed
persistence
40
67
25
45
26
27
26
8
4
22
184
34
6
6
50
40
56
31
10
51
54
89
75
53
70
5
53
34
9
6
24
11
13
7
24
7
12
73
7
31
11
22
17
30
21
45
16
25
65
12
15
6
6
10
8
7
25
13
23
6
18
12
6
24
17
22
24
70
50
34
7
20
7
9
8
54
8
95
4
6
7
9
43
41
7
11
21
22
21
43
31
12
44
12
18
9
21
35
8
63
54
6
4
4
8
24
25
17
14
16
EGF-Induced WT Motility, 1 pM U73122
Appendix 4:
Speed Histograms
A4.1 WT Basal and EGF-Induced Speed Comparisons
The following plots are comparisons of basal and EGF induced responses of WT EGFR
expressing cells. Speeds were all calculated from the 15 minute average mean squared
displacement of an individual cell. The number of cells corresponding to a given speed
were divided by the total number of cells at the given Amgel concentration to give the
normalized number. Basal speed is represented by closed circles, EGF-induced speed is
represented by open circles.
0.048 pg/ml Amgel
-0.1
-50
0
50
100
150
200
150
200
speed (pm/hr)
0.48 jig/ml Amgel
-50
0
50
100
speed (mn/hr)
4.8 pg/ml Amgel
WT
25nM EGF
WT
OnMEGFo
OnM EGF
-0--
S--
0
-0.1
0
0
50
100
speed (gm/hr)
150
2C
A4.2:
Effect of Amgel on WT Speed Histograms
The following two plots compare the speed response of WT EGFR expressing cells on
varying concentrations of Amgel. Both basal and EGF-induced speed were examined.
Note the addition of EGF causes an increase in the variation of the population in addition to
an increase in the mean.
WT, OnM EGF
-0.1
-50
0
50
100
150
200
150
200
speed (pm/hr)
WT, 25nM EGF
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-0.1
0
50
100
speed (jpm/hr)
A4.3: c'973 Basal and EGF-Induced Speed Comparisons
The following plots are comparisons of basal and EGF induced responses of c'973 EGFR
expressing cells. Speeds were all calculated from the 15 minute average mean squared
displacement of an individual cell. The number of cells corresponding to a given speed
were divided by the total number of cells at the given Amgel concentration to give the
normalized number. Basal speed is represented by closed triangles, EGF-induced speed is
represented by open triangles.
0.048pg/ml Amgel
0
-0.2
0
50
100
150
200
150
200
speed (pm/hr)
0.481ig/ml Amgel
0
-0.2
-50
0
50
100
speed (jm/hr)
100
4.8jig/ml Amgel
0.8
.
.
I
0.6
0.4
-0.2
c'973
OnM EGF
c'973
•
-50
0
50
100
speed (Lmn/hr)
101
150
200
A4.4:
Effect of Amgel on c'973 Speed Histograms
The following two plots compare the speed response of c'973 EGFR expressing cells on
varying concentrations of Amgel. Both basal and EGF-induced speed were examined.
Note the presence of subpopulations in the basal speed at the intermediate Amgel
concentration. The addition of EGF apparently causes a shift of the majority of the cells to
the subpopulation with the higher speed.
c'973, OnM EGF
-0.2
-50
0
50
100
150
200
150
200
speed (pm/hr)
c'973, 25nM EGF
-0.
-50
0
I ,
50
i '
100
speed (gim/hr)
102
.
A4.5:
Effect of PLC Abrogation on Speed Histograms
The basal and EGF-induced WT response are in closed circles and open circles,
respectively. The addition of 1pM U73122 (open squares) causes a decrease in the EGF
induced response. The variance of the population is not as broad, and the mean is not as
high. Truncation of the EGFR, however, has a different effect (open triangles). The
variance of the population is very similar to the basal WT variance, but the mean is shifted.
This suggests addition of U73122 and truncation of the EGFR affect the PLC pathway in
different ways.
0.48gig/ml Amgel
0.7
WT
0.6
0.5
EGF
OnM EGF
-
0.4
S
c'973
,25nM
0.3
25nM EGF
'
'
1pM U73122
I
0.2ia
0.2
.WT
-
25nM EG
0.1
-0.1
.
-50
0
50
100
speed (gm/hr)
103
150
200
Appendix 5:
A5.1:
Persistence Histograms
WT Basal and EGF-Induced Persistence Comparisons
The following plots are comparisons of basal and EGF induced responses of WT EGFR
expressing cells. Persistence times were determined by fitting the persistent random walk
equation with the appropriate speed inserted. The number of cells corresponding to a given
persistence were divided by the total number of cells at the given Amgel concentration to
give the normalized number. Basal persistence is represented by closed circles, EGFinduced persistence is represented by open circles. Note the level of response due to EGF
is greatly reduced compared to the case of cell speed.
0.048gtg/ml Amgel
WT
OnM EGF
0.5
0.4
0.3
-WT
25nM EGF
L
0.2
0.1
0
m
I.
-
.
-0.1
0
)
50
100
150
200
250
150
200
250
persistence (min)
0.48pg/ml Amgel
.
-0.1
-50
0
50
100
persistence (min)
104
4.8gg/ml Amgel
.- WT
OnM EGF
,
WT
25nM EGF
-4
0
.
-0.1
-50
0
.
.
.
-
50
.
-
-
-
100
persistence (min)
105
2
150
2
2
•
'
200
|
.I I
250
A5.2:
Effect of Amgel on WT Persistence Histograms
The following two plots compare the persistence response of WT EGFR expressing cells
on varying concentrations of Amgel. Both basal and EGF-induced persistence were
examined. Note the addition of EGF generally causes an increase in the variation of the
population rather than a direct shift of the mean.
WT, OnM EGF
-0.1
0
-20
40
20
60
80
100
60
80
100
persistence (min)
WT, 25nM EGF
.
-0.1
-20
0
40
20
persistence (min)
106
A5.3: c'973 Basal and EGF-Induced Persistence Comparisons
The following plots are comparisons of basal and EGF induced responses of c'973 EGFR
expressing cells. Persistence times were determined by fitting the persistent random walk
equation with the appropriate speed inserted. The number of cells corresponding to a given
persistence were divided by the total number of cells at the given Amgel concentration to
give the normalized number. Basal persistence is represented by closed triangles, EGFinduced persistence is represented by open triangles.
0.048gg/ml Amgel
0.6
0.5
0.4
0.3
0.2
0.1
. .. .
-0.1
-50
0
50
100
150
200
250
150
200
250
persistence (min)
0.48pg/ml Amgel
Sc'973
25nM EGF
0.4
c'973
OnM EGF
0.3
0.2
0.1
0
-0.1
50
100
persistence (min)
107
4.8gg/ml Amgel
c'973
"
OnM EGF
c'973
25nM EGF
A-
0
,I
-0.1
)
•....
0
.I
I
50
...
.
100
persistence (min)
108
I
150
...
I
200
...
250
A5.4:
Effect of Amgel on c'973 Persistence Histograms
The following two plots compare the persistence response of c'973 EGFR expressing cells
on varying concentrations of Amgel. Both basal and EGF-induced persistence were
examined. Note the Amgel effects are much clearer in the presence of EGF.
c'973, OnM EGF
0.1
-0.1
• • '
-20
0
20
40
60
80
100
60
80
100
persistence (min)
c'973, 25nM EGF
-0.1
-20
0
20
40
persistence (min)
109
A5.5:
Effect of PLC Abrogation on Persistence Histograms
The basal and EGF-induced WT response are in closed circles and open circles,
respectively. The addition of 1pM U73122 (open squares) and truncation of the EGFR
(open triangles), have very similar effects. The variance of the populations are much
smaller than the basal WT variance, and the means are smaller.
0.48gtg/ml Amgel
0
/I
-0.1
-20
0
25nM EGF
1pM U73122
20
40
persistence (min)
110
60
80
100
Appendix 6:
Membrane Extension Histograms
A6.1 WT Basal and EGF-Induced Extension Comparisons
The following plots are comparisons of basal and EGF induced responses of WT EGFR
expressing cells. Extension rates were calculated as described in Materialsand Methods.
The number of cells corresponding to a given extension were divided by the total number
of cells at the given Amgel concentration to give the normalized number. Basal extension is
represented by closed circles, EGF-induced extension is represented by open circles.
0.048[tg/ml Amgel
0
-0.1
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.2
1.4
normalized extension per 15 min
0.48 pg/ml Amgel
0
-0.1
0
0.2
0.4
0.6
0.8
1
normalized extension per 15 min
111
4.8gg/ml Amgel
0.4
WT
OnM EGF
WT
,
0.3
25nM EGF
SFr
0.2
-o
0.1
0
.
-0.1
0
...
0.2
...
......
0.4
0.6
0.8
1
normalized extension per 15 min
112
a..
1.2
1.4
A6.2:
Effect of Amgel on WT Extension Histograms
The following two plots compare the extension response of WT EGFR expressing cells on
varying concentrations of Amgel. Both basal and EGF-induced extensions were examined.
Note both the substratum and the addition of EGF cause differences in membrane
extension. The addition of EGF has a greater effect directly on the mean than the variance.
WT, OnM EGF
0
-0.1
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.2
1.4
normalized extension per 15 min
WT, 25nM EGF
0
-0.1
0
0.2
0.4
0.6
0.8
1
normalized extension per 15 min
113
A6.3:
Effect of PLC Abrogation on Extension Histograms
The basal and EGF-induced WT response are in closed circles and open circles,
respectively. Neither the addition of U73122 (open squares) or the truncation of the EGFR
(open triangles) causes a large effect on the EGF-induced extension rate. The curves are
essentially superimposed.
0.48Cpg/ml Amgel
0.4
0.3
-
OnM GF
0.2
\
I
/
0.1
Z
WT
25nM EGF
I,:
1 ,c'973
WT
25nM EGF
25nM EGF
1pM U73122
0
I
...
-0.1
0
0.2
0.4
.
0.6
normalized extension per 15 min
114
0.8
1
Appendix 7:
Cell Spread Area Histograms
A7.1 WT Basal and EGF-Induced Area Comparisons
The following plots are comparisons of basal and EGF induced responses of WT EGFR
expressing cells. Cell spread areas were calculated as described in MaterialsandMethods.
The number of cells corresponding to a given area were divided by the total number of cells
at the given Amgel concentration to give the normalized number. Basal area is represented
by closed circles, EGF-induced area is represented by open circles.
0.048 pg/ml Amgel
500
1000
cell area
1500
2000
1500
2000
(pm 2)
0.48pig/ml Amgel
0
-0.1
500
1000
cell area (gm 2)
115
4.81tg/ml Amgel
0.4
WT
25nM EGF
0.3
WT
0.1
00
-0.1
0
500
1000
cell area (Pm 2)
116
1500
2000
A7.2:
Effect of Amgel on WT Area Histograms
The following two plots compare the cell spread area response of WT EGFR expressing
cells on varying concentrations of Amgel. Both basal and EGF-induced areas were
examined. Note the addition of EGF causes a decrease in the variance of the population
along with a decrease in the mean.
WT, OnM EGF
0
-0.1
0
500
1000
cell area
1500
2000
1500
2000
(pm 2)
WT, 25nM EGF
0.4
0.3
0
-0.1
0
500
1000
cell area (pm 2)
117
A7.3:
Effect of PLC Abrogation on Area Histograms
The basal and EGF-induced WT response are in closed circles and open circles,
respectively. The addition of U73122 (open squares) shows a bimodal area histogram.
The truncation of the EGFR (open triangles) along with addition of EGF, causes an even
larger decrease in variance than addition of EGF alone.
0.48pg/ml Amgel
0.4 -........
04 ,
'-- c'973
25nM EGF
WT
25nM EGF
1pM U73122
0.3
WT
25nM EGF
0.2
O\
OnM EGF
-0.1
0
1000
500
cell area
118
(pm 2)
1500
2000
References
Aaronson, S. A. (1991). "Growth Factors and Cancer." Science 254: 1146.
Alt, W. (1989). Correlation analysis of two-dimensional locomotion paths. Lecure Notes
in Biomathematics. W.Alt and G. Hoffmann, Springer-Verlag: 89.
Ando, Y. and P. J. Jensen (1993). "Epidermal growth factor and insulin-like growth factor
I enhance keratinocyte migration." Journaloflnvestigative Dermatology 100(5): 633.
Blay, J. and K. D. Brown (1985). "Epidermal growth factor promotes the chemotactic
migration of cultured rat intestinal epithelial cells." Journalof CellularPhysiology 124(1):
107.
Bleasdale, J. E., N. R. Thakur, et al. (1990). "Selective inhibition of receptor-coupled
phospholipase C-dependent processes in human platelets and polymorphonuclear
neutrophils." Journalof PharmacologyandExperimental Therapeutics255: 756.
Boring, C. C., T. S. Squires, et al. (1993). Cancer Statistics. CA. 43: 7.
Bornfeldt, K. E., E. W. Raines, et al. (1994). "Insulin-like growth factor-1 and plateletderived growth factor-BB induce directed migration of human arterial smooth muscle cells
via signalling pathways that are distinct from those of proliferation." Journalof Clinical
Investigation 93: 1266.
Bray, D. (1992). Cell Movement, Garland Publishing.
Buck, C. and A. F. Horwitz (1987). "Cell receptors for extracellular matrix molecules."
Annual Review of Cell Biology 3: 179.
Burck, K. B., E. T. Liu, et al. (1988). Oncogenes: An Introduction to the Concept of
Cancer Genes. New York, Springer-Verlag.
Burridge, K., K. Fath, et al. (1988). "Focal adhesions: transmembrane junctions between
the extracellular matrix and the cytoskeleton." Annual Review of Cell Biology 4: 487.
Carpenter, C. L. and L. C. Cantley (1996). "Phosphoinositide kinases." Current Opinion
in Cell Biology 8: 153.
Carpenter, G. (1985). "Epidermal Growth Factor: Biology and Receptor Metabolism."
Journalof Cell Science Supplement 3: 1.
Carpenter, G. and M. I. Wahl (1990). The Epidermal Growth Factor Family. Peptide
Growth Factors and their Receptors I. M. B. Sporn and A. B. Roberts, Springer-Verlag:
69.
Cha, D., P. O'Brien, et al. (1996). "Enhanced modulation of keratinocyte motility by
transforming growth factor-alpha (TGF-alpha) relative to epidermal growth factor." Journal
of Investigative Dermatology 106(4): 590.
Chen, P., K. Gupta, et al. (1994). "Cell Movement Elicited by Epidermal Growth Factor
Receptor Requires Kinase and Autophosphorylation but Is Separable from Mitogenesis."
Journalof Cell Biology 124(4): 547.
119
Chen, P., J. E. Murphy-Ullrich, et al. (1996). "A Role for Gelsolin in Actuating Epidermal
Growth Factor Receptor-mediated Cell Motility." Journalof Cell Biology 134(3): 689.
Chen, P., H. Xie, et al. (1994). "Epidermal Growth Factor Receptor-mediated Cell
Motility: Phospholipase C Activity Is Required, but Mitogen-activated Protein Kinase
Activity Is Not Sufficient for Induced Cell Movement." Journalof Cell Biology 127(3):
847.
Chicoine, M. R. and D. L. Silbergeld (1995). "The in vitro motility of human gliomas
increases with increasing grade of malignancy." Cancer75(12): 290.
Chong, L. D., A. T. Kaplan, et al. (1994). "The small GTP binding protein Rho regulates
a phosphatidylinositol 4-phosphate 5-kinase in mammalian cells." Cell 79: 507.
Clark, R. (1995). Wound Repair: Overview and General Considerations. The Molecular
and Cellular Biology of Wound Repair. R. A. F. Clark. New York, Plenum Press: 3.
Collins, V. P. (1993). "Amplified genes in human gliomas." Seminars in CancerBiology
4: 27.
Conrad, P. A., K. A. Giuliano, et al. (1993). "Relative distribution of actin, myosin I, and
myosin II during the wound healing response of fibroblasts." Journalof Cell Biology
120(6): 1381.
Cook, P. W., M. R. Pittelkow, et al. (1991). "Growth factor-independent proliferation of
normal human neonatal keratinocytes: Production of autocrine- and paracrine-acting
mitogenic factors." Journalof CellularPhysiology 146: 277.
Cunningham, C. C., T. P. Stossel, et al. (1991). "Enhanced motility in NIH 3T3
fibroblasts that overexpress gelsolin." Science 251(4998): 1233.
Dadabay, C. Y., E. Patton, et al. (1991). "Lack of correlation between changes in
polyphosphatidylinositol 4,5-bisphosphate levels and actin/gelsolin complexes in A431
cells treated with epidermal growth factor." Journal of Cell Biology 112: 1151.
den Hartigh, J. C., P. M. P. van Bergen en Henegouwen, et al. (1992). "The EGF
Receptor is an Actin-binding Protein." Journalof Cell Biology 119(2): 349.
Diakonova, M., B. Payrastre, et al. (1995). "Epidermal growth factor induces rapid and
transient association of phospholipase C-yl with EGF-receptor and filamentous actin at
membrane ruffles of A431 cells." Journalof Cell Science 108: 2499.
Dickinson, R. and R. T. Tranquillo (1993). "A stochastic model for adhesion-mediated cell
random motility and haptotaxis." Journalof MathematicalBiology 31: 563.
Dickson, R. B., M. D. Johnson, et al. (1996). "Anti-invasion drugs." Breast Cancer
Research and Treatment 38: 121.
DiFiore, P. P., J. H. Pierce, et al. (1987). "Overexpression of the human EGF receptor
confers an EGF-dependent transformed phenotype to NIH 3T3 cells." Cell 51: 1063.
120
DiMilla, P. A., K. Barbee, et al. (1991). "A mathematical model for the effects of adhesion
and mechanics on cell migration speed." BiophysicalJournal60: 15.
DiMilla, P. A., J. A. Stone, et al. (1993). "Maximal migration of human smooth muscle
cells on fibronectin and type IV collagen occurs at an intermediate attachment strength."
Journalof Cell Biology 122: 729.
Duband, J.-L., S. Dufour, et al. (1991). "Neural crest cell locomotion is induced by
antibodies to 31 integrins: A tool for studying the roles of substratum molecular avidity
and density in migration." Journalof Cell Science 98: 517.
Dunn, G. A. (1983). "Characterizing a kinesis response: Time averaged measures of cell
speed and directional persistence." Agents and Actions [Supplement] 12: 14.
Engebraaten, O., R. Bjerkvig, et al. (1993). "Effects of EGF, bFGF, NGF an dPDGF(bb)
on cell proliferative, migratory and invasive capacities of human brain-tumor biopsies in
vitro." InternationalJournalof Cancer53: 209.
Fong, C. J., D. M. Sutkowski, et al. (1992). "Utilization of the Boyden chamber to further
characterize in vitro migration and invasion of benign and malignant human prostatic
epithelial cells." Invasion and Metastasis 12: 264.
Fujii, K., N. Dousaka-Nakajima, et al. (1995). "Epidermal growth factor enhancement of
HSC-1 human cutaneous squamous carcinoma cell adhesion and migration on type I
collagen involves selective up-regulation of oa20 1 integrin expression." Experimental Cell
Research 216: 261.
Gill, G. N., P. J. Bertics, et al. (1987). "Epidermal growth factor and its receptor."
Molecularand CellularEndocrinology 51: 169.
Goldschmidt-Clermont, P. J., J. W. Kim, et al. (1991). "Regulation of Phospholipase Cyl by Profilin and Tyrosine Phosphorylation." Science 251(4998): 1231.
Goodman, S., G. Risse, et al. (1989). "The E8 subfragment of laminin promotes
locomotion of myoblasts over extracellular matrix." Journalof Cell Biology 109: 799.
Hack, N., A. Sue-A-Quan, et al. (1993). "Expression of human tyrosine kinase-negative
epidermal growth factor receptor amplifies signaling through endogenous murine epidermal
growth factor receptor." Journalof Biological Chemistry 268: 26441.
Hernandez-Sotomayor, S. M. T. and G. Carpenter (1992). "Epidermal growth factor
receptor: Elements of intracellular communication." Journalof Membrane Biology 128:
81.
Hoelting, T., A. E. Siperstein, et al. (1994). "Epidermal Growth Factor Enhances
Proliferation, Migration, and Invasion of Follicular and Papillary Thyroid Cancer in Vitro
and in Vivo." Journalof ClinicalEndocrinology and Metabolism 79(2): 401.
Huttenlocher, A., R. R. Sandborg, et al. (1995). "Adhesion in cell migration." Current
Opinion in Cell Biology 7: 697.
121
Hynes, N. E. and D. F. Stern (1994). "The biology of erbB-2/neu/HER-2 and its role in
cancer." Biochemica et BiophysicaActa 1198: 165.
Hynes, R. (1987). "Integrins: A family of cell surface receptors." Cell 48: 549.
Jarrad, D. F., B. F. Blitz, et al. (1994). "Effects of epidermal growth factor on prostate
cancer cell line PC3 growth and invasion." Prostate24: 46.
Jiang, L. W. and M. Schindler (1990). "Nucleocytoplasmic transport is enhanced
concomitant with nuclear accumulation of epidermal growth factor (EGF) binding activity
in both 3T3-1 and EGF receptor reconstituted NR-6 fibroblasts." The Journalof Cell
Biology 110: 559.
Khazaie, K., V. Schirrmacher, et al. (1993). "EGF receptor in neoplasia and metastasis."
Cancerand Metastasis Reviews 12: 255.
Klemke, R. L., C. Shuang, et al. (1997). "Regulation of Cell Motility by Mitogenactivated Protein Kinase." The Journalof Cell Biology 137(2): 481.
Komuro, H. and P. Rakic (1995). "Dynamics of granule cell migration: A confocal
microscopic study in acute cerebellar slice preparations." The JournalofNeuroscience
15(2): 1110.
Konturek, P. C., S. J. Konturek, et al. (1995). "Epidermal growth factor and transforming
growth factor-alpha: role in protection and healing of gastric mucosal lesions." European
Journalof Gastroenterologyand Hepatology 7(10): 933.
Kundra, V., J. A. Escobedo, et al. (1994). "Regulation of chemotaxis by the plateletderived growth factor receptor-b." Nature 367: 474.
Kung, H. J., C. M. Chang, et al. (1994). Structural Basis of Oncogenic Activation of
Epidermal Growth Factor Receptor. Biochemical and Molecular Aspects of Selected
Cancers, Academic Press, Inc. 2: 19.
Kurachi, H., K. Morishige, et al. (1994). "Expression of epidermal growth factor and
transforming growth factor-alpha in fallopian tube epithelium and their role in
embryogenesis." Hormone Research 41(Suppl. 1): 48.
Lackie, J. (1986). Cell Movement and Cell Behavior, Allen & Unwin, Ltd.
Lauffenburger, D. A. and A. F. Horwitz (1996). "Cell Migration: A Physically Integrated
Molecular Process." Cell 84: 359.
Lauffenburger, D. A. and J. J. Linderman (1993). Receptors: Models for Binding.
Trafficking, and Signaling. New York, Oxford University Press.
LeRiche, V. K., S. L. Asa, et al. (1996). "Epidermal growth factor and its receptor (EGFR) in human pituitary adenomas: EGF-R correlates with tumor aggressiveness." Journalof
ClinicalEndocrinologyand Metabolism 81(2): 656.
Li, C. B., H. Nagayasu, et al. (1993). "Enhancement of tumorigenicity and invasion
capacity of rate mammary adenocarcinoma cells by epidermal growth factor and
transforming growth factor-b." JapaneseJournalof CancerResearch 84: 1145.
122
Libermann, T. A., N. Razon, et al. (1984). "Expression of epidermal growth factor
receptors in human brain tumors." CancerResearch 44: 753.
Maldonado, B. A. and L. T. Furcht (1995). "Epidermal Growth Factor Stimulates
Integrin-Mediated Cell Migration of Cultured Human Corneal Epithelial Cells on
Fibronectin and Arginine-Glycine-Aspartic Acid Peptide." Investigative Ophthalmology
and Visual Science 36(10): 2120.
Masui, H., A. Wells, et al. (1991). "Enhanced tumorigenesis of NR6 cells which express
non-downregulating epidermal growth factor receptors." CancerResearch 51: 6170.
Matthay, M. A., J. P. Thiery, et al. (1993). "Transient effect of epidermal growth factor on
the motility of an immortalized mammary epithelial cell line." Journalof Cell Science 106:
869.
McCarthy, J. B., J. lida, et al. (1996). Mechanisms of parenchymal cell migration into
wounds. The Molecular and Cellular Biology of Wound Repair. R. A. F. Clark. New
York, Plenum Press.
McCubrey, J. A., L. S. Steelman, et al. (1991). "Growth-promoting effects of insulin-like
growth factor-i (IGF-1) on hematopoietic cells: Overexpression of introduced IGF-1
receptor abrogates interleukin-3 dependency of murine factor-dependent cells by a liganddependent mechanism." Blood 78: 921.
McNamee, H. P., D. E. Inger, et al. (1993). "Adhesion to fibronectin stimulates inositol
lipid synthesis and enhances PDGF-induced inositol lipid breakdown." The Journalof Cell
Biology 121(3): 673.
Miettinen, P. J., J. E. Berger, et al. (1995). "Epithelial immaturity and multiorgan failure
in mice lacking epidermal growth factor receptor." Nature 376(6538): 337.
Mitchison, T. J. and L. P. Cramer (1996). "Actin-Based Cell Motility and Cell
Locomotion." Cell 84: 371.
Moulin, V. (1995). "Growth factors in skin wound healing." EuropeanJournalof Cell
Biology 68(1): 1.
Nanney, L. B. (1990). "Epidermal and dermal effects of epidermal growth factor during
wound repair." Journalof Investigative Dermatology95(5): 624.
Nanney, L. B. and L. E. King (1996). Epidermal Growth Factor and Transforming
Growth Factor-a. The Molecular and Cellular Biology of Wound Repair. R. A. F. Clark.
New York, Plenum Press: 171.
Noh, D., S. H. Shin, et al. (1995). "Phosphoinositide-specific phospholipase C and
mitogenic signaling." Biochimica et BiophysicaActa 1242: 99.
Othmer, H. G., S. R. Dunbar, et al. (1988). "Models of dispersal in biological systems."
JournalofMathematicalBiology 26(3): 263.
Palecek, S. P., J. C. Loftus, et al. (1997). "Integrin-ligand binding properties govern cell
migration speed through cell-substratum adhesiveness." Nature 385: 537.
123
Pawson, T. (1995). "Protein modules and signalling networks." Nature 373: 573.
Pawson, T. and J. Schlessinger (1993). "SH2 and SH3 domains." CurrentBiology 3(7):
434.
Podolsky, D. K. (1997). "Healing the epithelium: solving the problem from two sides."
Journalof Gastroenterology32(1): 122.
Preston, T. M., C. A. King, et al. (1990). The Cytoskeleton and Cell Motility. New York,
Chapman and Hall.
Pruss, R. M. and H. R. Herschman (1977). "Variants of 3T3 cells lacking mitogenic
response to epidermal growth factor." Proceedingsof the NationalAcademy of Science
USA 74: 3918.
Radinsky, R., S. Risin, et al. (1995). "Level and function of epidermal growth factor
receptor predict the metastatic potential of human colon carcinoma cells." Clinical Cancer
Research 1(1): 19.
Rhee, S. G. and Y. S. Bae (1997). "Regulation of phosphoinositide-specific
phospholipase C isozymes." The Journalof Biological Chemistry 272(24): 15045.
Ridley, A. J., H. F. Paterson, et al. (1992). "The small GTP-binding protein rac regulates
growth factor-induced membrane ruffling." Cell 70: 401.
Rijken, P. J., S. M. Post, et al. (1995). "Actin polymerization localizes to the activated
epidermal growth factor receptor in the plasma membrane, independent of the cytosolic free
calcium transient." Experimental Cell Research 218: 223.
Ruoslahti, E. and M. Pierschbacher (1987). "New perspectives in cell adhesion: RGD and
integrins." Science 238: 291.
Sasaoka, T., W. J. Langlois, et al. (1996). "Involvement of ErbB2 in the signaling
pathway leading to cell cycle progression from a truncated epidermal growth factor receptor
lacking the c-terminal autophosphorylation sites." The Journalof Biological Chemistry
271(14): 8338.
Savarese, D. M. F., J. T. Russell, et al. (1992). "Type IV Collagen stimulates an
increasing intracellular calcium: Potential role in tumor cell motility." The Journalof
Biological Chemistry 267: 977.
Schlessinger, J. (1997). "Phospholipase Cy activation and phosphoinositide hydrolysis are
essential for embryonal development." Proceedingsof the NationalAcademy of Science,
USA 94: 2798.
Schmidt, C. E., A. F. Horwitz, et al. (1993). "Integrin/cytoskeleton interactions in
migrating fibroblasts are dynamic, asymmetric, and regulated." Journalof Cell Biology
123: 977.
Schwarzbauer, J. E. (1997). "Cell migration: May the force be with you." Current
Biology 7(5): R292.
124
Segall, J. E., S. Tyerech, et al. (1996). "EGF stimulates lamellipod extension in metastatic
mammary adenocarcinoma cells by an actin-dependent mechanism." Clinical and
ExperimentalMetastasis 14: 61.
Sheetz, M. P. (1994). "Cell migration by graded attachment to substrates and contraction."
Seminars in Cell Biology 5(3): 149.
Siegal, G. P., M.-H. Wang, et al. (1993). "Development of a novel human extracellular
matrix for quantitation of the invasiveness of human cells." CancerLetters 69: 123.
Spivak-Kroizman, T., D. Rotin, et al. (1992). "Heterodimerization of c-erbB2 with
different epidermal growth factor receptor mutants elicits stimulatory and inhibitory
responses." The Journalof Biological Chemistry 267: 8056.
Stein, W. and F. Bronner, Eds. (1989). Cell Shape: Determinants, Regulation, and
Regulatory Role, Academic Press.
Thomas, K. A. (1996). "Vascular endothelial growth factor, a potent and selective
angiogenic agent." The Journalof Biological Chemistry 271(2): 603.
Turley, E.A. (1992). "Hyaluronan and cell locomotion." Cancerand MetastasisReviews
11: 21.
Turner, T., P. Chen, et al. (1996). "EGF receptor signaling enhances in vivo invasiveness
of DU-145 human prostate carcinoma cells." Clinicaland ExperimentalMetastasis 14(4):
409.
Velu, T. J., L. Beguinot, et al. (1987). "Epidermal Growth Factor-Dependent
Transformation by a Human EGF Receptor Proto-Oncogene." Science 238: 1408.
Wang, J.-S., J. P. Coburn, et al. (1997). "Role of gelsolin in actin depolymerization of
adherent human neutrophils." MolecularBiology of the Cell 8: 121.
Wells, A. and J. M. Bishop (1988). "Genetic determinants of neoplastic transformation by
the retroviral oncogene v-erbB." Proceedingsof the NationalAcademy of Science USA
85(20): 7597.
Wells, A., J. B. Welsh, et al. (1990). "Ligand-Induced Transformation by a
Noninternalizing Epidermal Growth Factor Receptor." Science 247: 962.
Welsh, J. B., G. N. Gill, et al. (1991). "A Negative Feedback Loop Attenuates EGFinduced Morphological Changes." The Journalof Cell Biology 114(3): 533.
Wenczak, B. A., J. B. Lynch, et al. (1992). "Epidermal growth factor receptor distribution
in bum wounds. Implications for growth factor-mediated repair." Journalof Clinical
Investigation 90(6): 2392.
Wilson, S. E., Y. G. He, et al. (1994). "Effect of epidermal growth factor, hepatocyte
growth factor, and keratinocyte growth factor, on proliferation, motility and differentiation
of human corneal epithelial cells." ExperimentalEye Research 59(6): 665.
Witke, W., A. H. Sharpe, et al. (1995). "Hemostatic, inflammatory, and fibroblast
responses are blunted in mice lacking gelsolin." Cell 81(1): 41.
125
Wu, P., J. Hoying, et al. (1994). "Soluble integrin-binding competitor either inhibits or
enhances endothelial cell migration, predictably from adhesion effect." Annals of
Biomedical Engineering22: 144.
Xie, H., M.-H. Wang, et al. (1995). "In vitro invasiveness of DU-145 human prostate
carcinoma cells is modulated by EGF receptor-mediated signals." Clinicaland Experimental
Metastasis 13: 407.
Yasui, W., H. Sumiyoshi, et al. (1988). "Expression of epidermal growth factor receptor
in human gastric and colon carcinomas." CancerResearch 48: 137.
Zumkeller, W. and P. N. Schofield (1995). "Growth factors, cytokines and soluble forms
of receptor molecules in cancer patients." Anticancer Research 15(2): 343.
126