\IHIUWIWlIllHlHHUIHIUHlllllllllllllllllllllullll

3 1293 010152

This is to certify that the

dissertation entitled

Evidence that formation of neurofibrosarcomas
in individuals with type 1 neurofibromatosis is
associated with activation of growth factor
pathways

presented by
Jewry W. Wray

has been accepted towards fulfillment
of the requirements for

P11 b degree in i‘Xicroloiologb‘
Ph. D. Microbiology

mm (M

Major professor

($3
Datg‘t‘iflrc 5/

MS U is an Affirmative Action/Equal Opportunity Institution

 

 

LIBRARY
Michigan State
University

 

 

 

PLACENREI'URN BOXtoromovoihbchodt

ouiromywnoord.
To AVOID FINES Mummerbdondmduo.

DATE DUE ‘ DATE DUE DATE DUE
LEE 2 7 396‘ I

g

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EVIDENCE THAT FORMATION OF NEUROFIBROSARCOHAS IN
INDIVIDUALS HITH TYPE I NEUROFIBROHATOSIS IS
ASSOCIATED NITH ACTIVATION OF GROHTH FACTOR PATHNAVS

By

Jeremy H. Nray

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Microbiology
1994

ABSTRACT
EVIDENCE THAT FORHATION OF NEUROFIBROSARCOHAS IN PATIENTS

AFFECTED HITH TYPE 1 NEUROFIBROHATOSIS IS ASSOCIATED
HITH ACTIVATION OF GROHTH FACTOR PATHHAYS

By
Jeremy w. Hray

One of the features of Type 1 Neurofibromatosis (NFl) is that affected
individuals commonly form numerous benign neurofibromas, which arise from
Schwann cells that surround peripheral nerves. Rarely, a neurofibroma
cell undergoes malignant transformation giving rise to a
neurofibrosarcoma. I used a reverse transcription-polymerase chain
reaction (RT-PCR) assay to examine whether tumor cell lines derived from
NFl-associated neurofibromas and neurofibrosarcomas express TGF-a and
PDGF-B mRNA. The results of these experiments indicate the
neurofibrosarcoma cells express elevated levels of TGF-a and PDGF-B mRNA
relative to neurofibroma cells. Neurofibroma cells also express PDGF-B
mRNA but at a lower level than that neurofibrosarcoma cells. Normal rat
Schwann cells and human Schwann cells derived from neurofibromas do not
express receptors for serum growth factors such as PDGF, therefore, I
tested whether neurofibrosarcoma cells express such receptors by'measuring
proliferation of neurofibrosarcoma cells in response to PDGF, TGF-a and
bFGF in vitro. These results indicate that some neurofibrosarcomas
constitutively express one or more of the receptors for PDGF, EGF/TGF-a,

and bFGF. This work also demonstrates that one of two neurofibrosarcoma

cell lines capable of growth in medium lacking exogenous growth factors is
completely inhibited from growth in these conditions in the presence'of 50
ug/ml of suramin. These results support a role for activation of growth
factor pathways, including elevated expression of growth factors and
growth factor receptors, in the abnormal proliferation of Schwann cells

that occurs in NFl patients.

This work is dedicated to my wife, Christine, and to our children.

iv

ACKNOHLEDGHENTS

I have only been able to accomplish this work with the unending
support and patience of many individuals. I would like to specifically
thank a few of these individuals for all they have done.

My parents, Eugene and Jackie Nray, taught me correct principles,
expected me to live by those principles, and gave me their examples to
follow.

I am deeply indebted to Dr. J. Justin McCormick, mw'major professor,
for his financial support and scientific guidance. But more than this, I
thank Dr. McCormick for exemplifying one who truly has good will towards
all of his fellow women and men.

Dr. Veronica M. Maher has provided invaluable thought, insight, and
criticism that has aided my progress all along the way.

Finally, I have been privileged to have association with all my
colleagues in the Carcinogenesis Laboratory, and would like to give
special thanks to several individuals including Dr. Chien Cheng Lin, Dr.
DaJun Yang, Dr. Dave Reinhold, Dr. Jeanette Scheid, Dr. Glenn McGregor,
Dr. Rebecca Odenwaller, Lonnie Milam, and Suzanne Kately-Kohler, for their

friendship and their frequent advice.

TABLE OF CONTENTS

Page
LIST OF TABLES ........................... x
LIST OF FIGURES ........................... xi
INTRODUCTION ............................ 1
List of References ........................ 4
CHAPTER 1. LITERATURE REVIEW
I. Neoplasia is a consequence of alterations
in DNA and/or DNA expression .............. 5
A. Tumors have many changes ............. 5
8. Early observations ................ 6
C. Terminology of in Vitro transformation ...... 7
D. DNA is the target of oncogenic changes ...... 7
II. In Vitro Studies that Support Multistep
Carcinogenesis ..................... 10
A. Investigation of Rodent Cells In Vitro ...... 10
1. Spontaneous Transformation ......... 10
2. Carcinogen-induced Transformation ..... 11
3. Oncogene-induced Transformation ....... 12
B. Investigations of Human Cells In Vitro ...... 14
1. Carcinogen Experiments Using
Normal Human Cells ............. 14
2. Transfection of Normal Human Cells
with Oncogenes ............... 16
3. Infinite Life-Span Human Cell Strains . . . . 16
4. Neoplastic Transformation of Infinite
Life-Span Human Cell Strains ........ 18
III. In Vivo Models of Multistep Carcinogenesis ....... 20
A. Mouse Skin Carcinogenesis ............. 20
1. The Two-Step Model ............. 20
2. Oncogenes in Mouse Skin Carcinogenesis . . . 23

vi

8. Human Colorectal Carcinogenesis .......... 24

1. Familial Adenomatous Polyposis Coli ..... 25
2. Hereditary Nonpolyposis Colon Cancer . . . . 28
3. Other Genetic Changes Associated

with Colon Cancer .............. 30

IV. Ras Proteins are Key Mediators of Mitogenic Signals

in Normal and Neoplastic Cells ............. 31
A. The GDP/GTP cycle ................. 33
B. GTPase Activating Proteins ............ 34

1. p120-GAP .................. 35

2. Neurofibromin ................ 36
C. Guanine Nucleotide Releasing Factors ....... 38
D. Mitogenic Growth Factors ............. 40

V. Evidence of Autocrine Mechanisms in Transformation:
The Role of PDGF-B in Neoplasia ............. 43

A. The Transforming Gene of SSV is
Homologous to the c-sis Gene ........... 43

8. Structure and Function of PDGF-B and
Its Receptor ................... 45

C. The Role of PDGF-B Expression in Neoplasia . . . . 46

1. Autocrine Activation of the PDGF-receptor . 47
2. Experiments with Suramin .......... 48
3. PDGF-B Expression in Human Tumor Cells . . . 50
4. Paracrine Mechanisms in Neoplasia ...... 52
LIST OF REFERENCES ......................... 55

CHAPTER II. EVIDENCE THAT FORMATION OF NEUROFIBROSARCOMAS
IN INDIVIDUALS HITH TYPE 1 NEUROFIBROMATOSIS
IS ASSOCIATED WITH ACTIVATION OF GROWTH

FACTOR PATHWAYS ..................... 72
Abstract ........................... 73
Introduction ......................... 75

vii

Materials and Methods ..................... 78

Culture of tumor-derived cells ............. 78
Assay for tumorigenicity ................ 78
RNA isolation ...................... 79
Determination of level of expression of mRNA ...... 79
Assay for growth factor independence .......... 80
Assay of cellular response to specific growth factors . . 81
Results ............................ 82
Establishing cell lines from
tumors from NF1 patients ................ 82
Verification of the presence of neoplastic cells
in the tumor-derived cell population .......... 83
Neurofibromas ................... 83
Neurofibrosarcomas ................ 85
RT-PCR assay to detect PDGF-B and TGF-a mRNA ...... 89
Response of neurofibrosarcoma cells to
exogenous growth factors ................ 96
Effect of suramin on growth factor independence . . . . 100
Discussion ......................... 105
LIST OF REFERENCES ........................ 110

CHAPTER III.

CHARACTERIZATION OF A NEOPLASTIC
HUMAN FIBROUS HISTIOCYTONA CELL LINE
INCLUDING EVIDENCE SUGGESTING A

MESENCHYMAL ORIGIN ............... 114
Abstract .......................... 115
Introduction ........................ 116
Materials and Methods .................... 117
Culture of MFH cells ................. 117
Cloning Assay ..................... 117
Saturation Density .................. 118
Tumorigenicity Assay ................. 118
Isolation of RNA ................... 118
Detection of mRNA by RT-PCR .............. 118
Response of MFH cells to
PDGF-B and TGF-a ................... 119
Results ........................... 119

viii

MFH cells in culture ................. 119

Saturation density of MFH-HSU-l cells ......... 121
Tumorigenicity assay ................. 125
Growth factor mRNA expression ............. 125
Response of MFH-HSU-l to
mitogenic growth factors ............... 129
Discussion ......................... 129
LIST OF REFERENCES ........................ 135

CHAPTER IV. FIBROBLASTS DERIVED FROM NORMAL SKIN
BIOPSIES FROM A CHILD HITH INFANTILE
MYOFIBROMATOSIS HAVE ABNORMAL GROWTH

CHARACTERISTICS .................... 137
Abstract .......................... 138
Introduction ........................ 139
Case Report ......................... 140
Materials and Methods .................... 141
Culture of IM cells .................. 141
Focus assay ...................... 141
Tumorigenicity assay ................. 142
Results ........................... 142
IM fibroblasts in culture ............... 142
Foci from 1M fibroblasts in 1% serum ......... 143

IM fibroblasts form benign tumors
in athymic mice .................... 143
Discussion ......................... 147
LIST OF REFERENCES ........................ 153
APPENDIX ............................. 155

ix

LIST OF TABLES

Table Page
Chapter II
1. Tumorigenicity of neurofibrosarcoma-derived

cell lines .......................... 86
2. Human cell lines and strains ................. 91
3. Results of RT-PCR ....................... 99
Appendix
1. Tumor cell lines ...................... 156

LIST OF FIGURES

Figure Page
Chapter II
1. Hematoxylin and eosin stained

sections of neurofibromas ................... 84
2. Hematoxylin and eosin stained

sections of neurofibrosarcomas ................ 88
3. PCR Primers .......................... 93
4. Neurofibrosarcomas: Expression of GAPDH mRNA

and PDGF-B mRNA determined by RT-PCR ............. 94
5. Neurofibrosarcomas: Expression of GAPDH mRNA

and TGF-a mRNA determined by RT-PCR .............. 95
6. Neurofibromas: Expression of GAPDH mRNA

and PDGF-B mRNA determined by RT-PCR ............. 97
7. Stimulation of the growth of neurofibrosarcoma

cells by exogenous growth factors .............. 102
8. Growth factor free proliferation of neurofibrosarcoma

cells in the presence or absence of suramin ......... 104

Chapter III

1. Early passage MFH-HSU-l cells in culture .......... 120
2. Morphology of infinite life span cell line

derived from a human MFH .................. 123
3. Saturation density assay .................. 124
4. Hematoxylin and eosin stained tumor sections ........ 127
5. Hematoxylin and eosin stained tumor section

demonstrating infiltration of mononuclear cells ....... 128
6. Expression of GAPDH, PDGF-B, and TGF-a

mRNA determined by RT-PCR .................. 130
7. Growth response of MFH-HSU-l cells to

PDGF-BB and TGF-a ...................... 131
Chapter IV
1. Focus formation of IM fibroblasts in 1% serum ........ 144
2. Focus of cells that formed when IM fibroblasts

were plated in medium containing 1% serum .......... 145
3. Hematoxylin and eosin stained sections of

myofibromas ......................... 148

xi

INTRODUCTION

Malignant cancer cells have acquired a number of characteristics
including the ability to proliferate abnormally, invade tissue locally
and/or form distant metastases (Vogelstein et al., 1988). These abnormal
characteristics occur as a result of changes in the quantity or quality of
expression of specific genes. The predominate theory to explain the
acquisition of such changes is that they arise sequentially, and each
change confers some growth advantage on the cell in which it occurs.
Early observations recognized the association between certain behaviors or
environmental exposures and specific types of cancer. However, it was not
until the early 1980's that the genes which are altered in human cancers
began to be identified. Much of the current research is aimed toward
understanding the normal function of such genes and the mechanisms by
which these genes become subverted in neoplasia.

Many of the genes associated with cancer that have been identified
encode proteins that are involved in the transduction of mitogenic
signals. Classes of proteins involved in signal transduction include: (1)
growth factors, (2) growth factor receptors, (3) cytoplasmic proteins that
transduce growth signals to cellular effector sites and, (4) nuclear
factors that regulate the expression of genes involved in the mitogenic
response. One of the primary goals of cancer research is to identify how
the various mediators and regulators of signal transduction interact in
normal cells and the mechanisms of abnormal signal transduction in cancer

cells.

2

The growth of cells in multicellular organisms is highly regulated
(reviewed by Cooper, 1992). In human adults, cells in some tissues divide
continuously, whereas cells in other tissues divide rarely, if ever. For
example, superficial cells of the gastrointestinal tract and the skin are
continuously sloughed off, and constant, rapid cell division is required
to replace these cells. Blood cells also divide continuously to replenish
aging and dying cells. Conversely, neurons and cardiac muscle rarely if
ever divide in mature adults. Understanding the mechanisms that control
cell growth and division is central to understanding the uncontrolled
growth characteristic of neoplastic cells.

Type 1 neurofibromatosis provides an opportunity to investigate
neoplasia in cells that are predisposed to abnormal activation of the ras
signal transduction pathway(s). Individuals affected with this disease
have inherited one defective allele of the NF1 gene, which encodes a
GTPase activating protein that is known as neurofibromin (DeClue et al.,
1992). Neurofibromin, like other GTPase activating proteins, is involved
in negative regulation of ras proteins. Therefore, individuals affected
with NF1 are predisposed to neoplasia as a consequence of the inability to
regulate ras normally.

NF1 patients typically develop many benign neurofibromas which can
undergo malignant transformation to form a neurofibrosarcoma. Although
the occurrence of neurofibrosarcomas is significantly elevated in NF1
patients compared to the general population, malignant transformation of
a neurofibroma is nevertheless a rare event, given the number of
neurofibromas from which they can arise. By studying cell lines that I
have derived from neurofibromas and neurofibrosarcomas that occurred in
individuals with NF1, I have been able to identify specific changes that
are associated with nalignant transfomatim of namfibrmas into newcfibrosarcams.

3

The review of the literature in this dissertation has two aims. The
first is to describe the evidence that has contributed to the current
understanding of multistep carcinogenesis. The second aim will focus on
signal transduction pathways that are mediated by ras proteins and the
manner in which ras proteins are regulated in normal and abnormal cells.
The second chapter is a paper that describes the evidence from my research
indicating that specific growth factors, including PDGF-B and TGF-a, have
a role in the pathogenesis of neurofibrosarcomas that occur in individuals
with Type 1 neurofibromatosis, and is written in the format appropriate
for Carcinogenesis. This paper (chapter 2) is the result of the majority
of the research that I have done in my Ph.D. program. The last two
sections are composed of minor chapters of additional research that I have
done in The Carcinogenesis Laboratory. The third chapter describes the
characterization of a cell line derived from a human malignant fibrous
histiocytoma. The fourth chapter describes abnormal characteristics
displayed by normal skin fibroblasts that were obtained from a 6 year old
male with congenital myofibromatosis, a rare disease in which affected
individuals are born with benign tumors. Finally, there is an appendix
that describes the human tumor cell lines that I have established in

culture in the course of my research.

LIST OF REFERENCES

Cooper, G.M. (1992) Elelents in hulan cancer. Jones and Bartlett
publishers, Boston.

DeClue, J.E., Papageorge, A.G., Fletcher, J.A., Diehl, S.R., Rather, N.,
Vass, H.C., and Lowy, D.R. (1992) Abnormal regulation of mammalian p21"‘
contributes to malignant tumor growth in van Recklinghausen (Type 1)
neurofibromatosis. Cell, 69: 265-273.

Vogelstein, 8., Fearon, E.R., Hamilton, S.R., Kern, S.E., Preisinger,
A.C., Leppert, M., Nakamura, Y., Hhite, R., Smits, A.M.M., and 805, J.L.
(1988) Genetic alterations during colorectal-tumor development. N. Engl.
J. M., 319: 525-537.

CHAPTER 1

LITERATURE REVIEH

 

A. Tuners Have Many Changes

Each tumor that occurs in humans or other vertebrates is considered
to be the consequence of qualitative and/or quantitative changes in the
expression of specific genes. Dne predominate theory to explain the
presence of multiple changes in tumor cells is that the changes arise
sequentially, and each of change confers some growth advantagelon the cell
in which it occurs. Proliferation of a cell that has an advantageous
change results in a large population of cells possessing the change. Such
a population of cells can serve as a target for further rounds genetic
change and subsequent selective growth. The accumulation of specific
genetic changes in a single cell is considered to result in a neoplastic
cell (Fearon and Vogelstein, 1990). Another theory explains the presence
of multiple changes by assuming that infidelity of DNA replication is
responsible for tumor oncogenesis and progression (Loeb et al., 1974).
The following sections present evidence that has contributed to this

current understanding of multistep carcinogenesis.

8. Early Observations

Evidence for a relationship between environmental and/or behavioral
factors and the occurrence of cancer was recognized long before there was
any understanding of the mechanisms of chemical carcinogenesis. For
example, in 1775 Pott described a high frequency of scrotal cancer in
males who had been employed as chimney sweeps when they were boys (Pott,
1775). Pott hypothesized that "The disease, in these people, seems to
derive its origin from la lodgement of soot in the rugae of the
scrotum,..". Nearly two centuries later benzo(a)pyrene>was isolated from
coal tar that is present in chimneys used to burn coal (Cook et al.,
1933). 8enzo(a)pyrene is a potent carcinogen and could have been the
agent responsible for the high frequency of scrotal cancer that Pott
observed in London chimney sweeps. Even prior to these observations by
Pott, the Italian physician Ramazzini had recognized that nuns have a
higher incidence of breast cancer thaniwomen in the general population (cf
Boyland, 1980). Investigators now have some understanding of the
protective effect that pregnancy-induced hormones exert on breast tissue.
Another example of early insights into carcinogenesis came from Hill, who
in his 1761 publication entitled, Cautions Against the Immoderate Use of
Snuff, reported that users of tobacco snuff developed cancer of the nasal
passage much more frequently than those who abstained (Redmond, 1970). The
presence of carcinogens in tobacco and tobacco smoke, and their role in
various types of cancer, is now well established. These investigators
were some of the first to recognize that agents related to environment
and/or behavior could elevate the risk for developing specific types of

cancers .

7

C. Terlinology of In Vitro transforlation

The two general categories of changes that occur in multistep
carcinogenesis have been classified as those that are genetic and those
that are epigenetic (Barrett, 1987). By definition, genetic alterations
describe "any change in phenotype which results from an aJteration in
primary DNA sequence. This change may be a single base pair change, a
deletion, an insertion, a rearrangement or duplication of one~or more base
pairs, or loss or gain of an entire chromosome". An epigenetic change
includes "any change in phenotype‘which does not result from an alteration
in DNA sequence. This change may be stable and heritable, and includes
alterations in DNA.methylation, transcriptional activation, translational
control, and post-translational modifications" (Barrett, 1987). However,
these terms are often used imprecisely in the cancer literature. For
example, abnormal elevation in transcription of a growth factor gene,
defined as an epigenetic change according to Barrett, may in fact be the
result of alteration of the DNA sequence in the promoter or enhancer
region of the gene, and in that case it should be defined as a genetic
change. To avoid confusion I will use the term genetic in reference to
changes in gene sequence and changes in the level of gene expression, and

I will describe the mechanism(s) of such changes when they are known.

D. DNA is the Target of Oncogenic Changes

Hypotheses that cancer originates from alteration of the genetic
material were proposed prior to the discovery that genes are composed of
DNA. The somatic mutation theory of carcinogenesis was first purposed by
T.H. Boveri, considered by many to be the father of the somatic mutation
theory of carcinogenesis (Boveri, 1929). His theory was based on the

observation that cancer cells typically have abnormal, and often bizarre

8

chromosomes. Furthermore, he observed that the entire population of
neoplastic cells in a given tumor usually shared the same chromosomal
defects, which suggested that the defects were passed fromia parental cell
to its progeny. Based on cytogenetic analysis he proposed that heritable
alterations in the chromosomes are responsible for cancer. In 1937 Furth
and Kahn demonstrated that a single mouse leukemia cell injected into a
syngeneic mouse caused the development of leukemia in rapid fashion,
indicating that the malignant phenotype is stably heritable. These early
studies suggested that cancer was the consequence of irreversible
alteration of the genetic material.

The characterization of DNA as the genetic material suggested that
the permanent effects caused by carcinogens occur at the DNA level. To
test the hypothesis that DNA was the target of carcinogens, Brookes and
Lawley (1964) investigated whether a positive correlation existed between
the degree of binding of various carcinogenic hydrocarbons to DNA, RNA,
and proteins and the degree of carcinogenicity of such compounds. They
found no association between thelcarcinogenicity and binding of any of the
hydrocarbons to protein or RNA. However, they found significant positive
correlation between the binding of specific hydrocarbons to DNA and the
carcinogenicity of the hydrocarbons (Brooks and Lawley, 1964).

In 1968 investigations by Maher et al. demonstrated a correlation
between the degree of carcinogenic potency of various derivatives of 2-
acetylaminofluorene (AAF) and N-methyl-4-aminobenzene (MAB) and the
ability of each derivative to induce mutations in DNA. In these
experiments DNA mutation was measured as a function of inactivation of
transforming DNA isolated from $8 19, a strain of Bacillus subtilis. SB
19 DNA containing an intact tryptOphan gene confers thelability to grow in

the presence of indole to the recipient bacterial strain known as T3.

9

Treatment of strain SB 19 DNA with specific carcinogens reduced the
ability of transforming DNA to confer growth proficiency to T3 recipient
bacteria. Maher et al. found that when transforming DNA was treated with
various derivatives of AAF and MAB, those compounds known to be the most
carcinogenic in vivo were the most potent in: (1) reducing the
transforming ability of transforming DNA, (2) elevating the frequency of
mutations in transforming DNA, and (3) decreasing the buoyant density of
transforming DNA as a result of covalent binding of the carcinogen
residues to the DNA. Miller and Miller had earlier proposed that pro-
carcinogens become ultimate carcinogens by metabolic conversion to
electron-pair-deficient electrophilic reactants. In light of the studies
of Maher et al. (1968), they extended that hypothesis to include the fact
that such electrophilic metabolites bind readily to the genetic material
to cause genetic damage.

One of the most convincing links between DNA damage and human
carcinogenesis resulted from studies of cells obtained from patients
affected with the autosomal recessive disease xeroderma pigmentosum (XP).
These individuals are highly sensitive to sunlight exposure, and normal
levels of exposure to sunlight results in early age, non-melanoma skin
cancer (Johnson, 1982). Investigations of XP cells in culture
demonstrated that XP skin fibroblasts are defective in repair of damage
caused by UV irradiation (Cleaver, 1969). Treatment of XP cells with UV
radiation resulted in a significant increase in mutation frequency
compared to that of normal fibroblasts (Maher and McCormick, 1976). The
conclusion from these studies is that the cells of individuals affected
with XP have a decreased rate of UV-induced excision repair which in turn
leads to an increased frequency of UV-induced mutations. This suggests

‘that the genes responsible for skin cancer undergo an increased mutation

10
frequency, and this results in skin cancer in sunlight exposed areas of
the body. In normal persons of light complexion who are exposed to
sunlight for many years, skin cancer histologically similar to that found
in XP patients occurs, but the cancer arises in these normal individuals

much later in life than it does in XP patients.

HIMWMWB

In vitro transformation experiments have been valuable in the
identification of those genes altered in carcinogenesis as well as in
understanding the biochemical role of such genes in normal and neoplastic
cells. The following terminology used in in vitro transformation was
agreed upon by the Tissue Culture Association (Schaeffer, H.I., 1990) and
will be used in this dissertation. "In vitro transformation" refers to
any heritable phenotypic change that can be observed and/or measured.
This differs from "in vitro neoplastic transformation“ which refers to in
vitro acquisition of the ability to form benign or malignant tumors in the
appropriate experimental animal. The term "in vitro nnJignant
transformation" is reserved to describe cells that have acquired, in
culture, the ability to form a tumor with malignant characteristics,
including the ability to invade host tissue locally and/or metastasize

within the host animal.

A. Investigations of Rodent Cells In Vitro

l. Spontaneous transforlation

Gey was the first to report that spontaneous malignant
transformation of rodent cells could occur in vitro. He observed that rat

fibroblasts propagated in cell culture for a period of time formed

11
sarcomas when they' were injected into rats (Gey, 1941). Shortly
thereafter, Earle reported that mouse fibroblasts undergo spontaneous
malignant transformation in culture, albeit at a lower frequency than
those treated with carcinogen (Earle, 1943). Later studies demonstrated
that the in vitro environments used by Earle and Gey were inherently
genotoxic as a result of factors such as high oxygen concentrations and
the presence of visible light (Parshad et al., 1977). Improvements in
cell culture techniques have reduced the frequency of spontaneous
transformation that occurs in rodent cell culture systems (Sanford and

Evans, 1982).

2. Carcinogen-Induced Transforlation

Carcinogen-induced in vitro transformation of Syrian hamster embryo
(SHE) cells was first described by Berwald and Sachs (1963, 1965). Nhen
SHE cells plated at cloning density were treated with either 3-
methylcholanthrene or benzo(a)pyrene, some of the clones were found to
consist of morphologically altered cells. Nhen populations of
carcinogen-treated SHE cells were continuously propagated in cell culture
some of the cells escaped senescence. Untreated SHE cells did not escape
senescence in these experiments.

Barrett and Ts’o demonstrated that when SHE cells treated with
benzo(a)pyrene were routinely subcultured, they expressed properties of
transformed cells, including morphological transformation, fibrinolytic
activity, anchorage independence (growth in soft agar), and the ability to
form tumors in athymic mice (1978). Furthermore, these changes occurred
in a sequential manner. Morphological transformation occurred within one
week of carcinogen treatment. Enhanced f ibrinolytic activity followed two

to three weeks after carcinogen treatment. Six to 15 weeks after (32 to

12
75 population doublings) after carcinogen treatment, cells acquired the
ability to form large colonies in soft agar.

The ability of the transformed SHE cells to grow in soft agar
correlated with malignant transformation. Eight different populations of
carcinogen-treated SHE cells that had become morphologically transformed,
but had not yet become anchorage independent, did not produce tumors in
neonatal hamsters. However, carcinogen-treated SHE lines that had
attained anchorage independence formed tumors in 100% of the neonatal
hamsters into which they were injected. SHE cells that had spontaneously
immortalized in cell culture (occurs at a low frequency in control cells),
but were not anchorage independent, did not form tumors in athymic mice.
However, one spontaneously immortalized SHE cell line became anchorage
independent at passage 37, and when injected into neonatal hamsters, this
cell line produced tumors in three of six animals (Barrett and Ts’o,
1978). The observation that SHE undergo several discernable phenotypic
changes on the path towards nalignancy suggests that multiple genetic

changes are required to achieve the malignant state.

3. Oncogene-Induced Transforaation

Investigations of the RNA virus known as the Rous sarcoma virus
(RSV), gave the first direct evidence that alteration of a normal
cellular gene could contribute to neoplastic transformation. RSV is a
transforming retrovirus and can neoplastically transform the~cells that it
infects. Bishop and his colleagues showed that homologs of the
transforming sequence of RSV were present in the genomes of normal avian
(Stehelin et al., 1976) and mammalian (Spector et al., 1978) Cells. The
finding that diverse species carry normal genes homologous to the viral

src transforming gene of RSV suggested that normal genes have the

13

potential to participate in carcinogenesis if they are altered, and that
RSV obtained the src gene by transduction of the host src sequence after
infection of a host cell (Bishop, 1983). Other investigations revealed
that many retroviral oncogenes contain sequences homologous to normal
genes present in several animal species including humans (reviewed by
Bishop, 1983). The normal cellular genes that correspond to the
transforming sequences of retroviruses were called protooncogenes to
distinguish them from the oncogenes of retroviruses.

The identification of protooncogenes in human cells led
investigators to ask whether such genes played a causal role in human
cancers. One strategy to detect whether activated protooncogenes were
present in human tumors was to introduce DNA isolated from a human tumor
into non-cancer cells and assess whether the tumor DNA was able to
transform the recipient cells. Such an assay had been used to identify
the oncogenic sequences of transforming retroviruses. The most common
assays utilized the immortalized murine cell line NIH 3T3 (Andersson et
al., 1979; Copeland and Cooper, 1979, Lowy et al., 1978). Heinberg and
his colleagues demonstrated that transfection of DNAs isolated from human,
murine, rabbit, and rat tumor cell lines induced transformation of NIH 3T3
cells (Shih et al., 1981). Transfection of DNAs from two human bladder
carcinoma cell lines also induced transformation of NIH 3T3 cells
(Krontiris and Cooper, 1981). Analysis of the NIH 3T3 cells transformed
by DNA prepared from these human tumors revealed that the sequences
responsible for transformation of the NIH 3T3 cells were homologous to the
ras genes of the Harvey and Kirsten sarcoma viruses (Der et al., 1982;
Parada et al., 1982 Dahr et al., 1982). Subsequently, transforming DNA
sequences from numerous human cancers were characterized and matched to

their transforming retroviral homologs (Land at al., 1983; Slamon et al.,

14
1984; Bishop, 1987). These studies demonstrated that activated
protooncogene sequences were present in human malignancies and strongly
suggested that alterations in specific protooncogene sequences contributed

to the genesis of neoplastic cells in humans.

8. Investigations of Hulan Cells In Vitro

Unlike rodent cells, normal diploid human fibroblasts have never
been observed to undergo spontaneous transformation in vitro . Hhen
normal human foreskin fibroblasts are cultivated in vitro they undergo
approximately 50 population doublings (Hayflick and Morehead, 1961) after
which they cease cell division. Normal diploid epithelial cells divide
only a few times in cell culture and also have not been observed to
undergo spontaneous transformation in vitro (Chang, 1986). Although
rodent cell. models have been an indispensable tool in advancing the
understanding of carcinogenesis, the differences between rodent cells and
human cells suggested the need for human cell systems to investigate

carcinogenesis in vitro.

1. Carcinogen Experiaents using Norlal Hulan Cells

In contrast to the numerous studies that describe in vi tro malignant
transformation of normal rodent cells, there are only a few reports of
carcinogen-induced or spontaneous in vitro malignant transformation of
normal human cells, and these have not been verified (see review by
McCormick and Maher, 1988 and McCormick et al., 1988). The most cited
example is that of Kakunaga (1978), who reported that treatment of the
normal human fibroblast cell line K0 with a single dose of carcinogen,
resulted in malignantly transformed cells. Because similar experiments

failed to reproduce the results of Kakunaga, attempts were made to

15

determine if the~malignant cells arose from the normal parental cell line.
Investigations of enzyme patterns, HLA antigens, restriction fragment
length polymorphisms, and marker chromosomes clearly showed that the
malignant cells could not have arisen from cell line KD. Further analysis
revealed that the isoenzyme pattern of the malignant cells was identical
to the human fibrosarcoma cell line 8387 (McCormick et al., 1988). These
results strongly suggest that the cell line 8387 cross-contaminated the
culture of carcinogen treated KD cells.

In 1978 Milo and Dipaolo selected carcinogen-treated normal human
fibroblasts for their ability to form colonies in soft agarose (Milo and
Dipaolo, 1978), and reported that such cells formed progressively growing
undifferentiated mesenchymal tumors when injected into athymic mice.
However, similar attempts by other laboratories to generate malignant
cells from normal human cells by carcinogen treatment in vitro have not
been successful (McCormick and Maher, 1989).

Investigators have achieved in vitro morphological transformation,
but not malignant transformation, of normal, finite life-span human cells
by carcinogen treatment in vitro. In each of these studies, normal cells
were exposed to a carcinogen, then cultured under conditions in which
transformed cells could be distinguished from non-transformed cells. For
example, when normal human cells were treated with MNNG are subsequently
grown in low serum conditions, they gave rise to areas of focus formation
(focal piling up of cells) (DeMars and Jackson, 1977). Though the focus-
derived cells were morphologically transformed, they did not form
progressively growing tumors when they were injected into athymic mice.
Similarly, normal human foreskin fibroblast treated with propane sultone
attained the ability to form colonies in soft agar, but failed to form
tumors in athymic mice (Silinskas et al., 1981). Other investigators

16
generated anchorage independent cells in vitro by treatment with °°Co
radiation (Wang et al., 1986) or heavy metals (Biedermann and Landolph,
1988), but anchorage independent cells in these investigations also
produced no tumors in athymic mice. One recent study demonstrated that
human mammary cells became malignantly transformed after repeated
exposures to ionizing radiation followed by suitable selection after each
treatment (Hazer et al.,1994). This is the first report of in vitro
generation of malignant cells from normal human cells by carcinogen

treatment alone.

2. Transfection of Norlal Hulan Cells with Oncogenes

Experiments designed to transform normal, finite life-span cells
into neoplastic cells by oncogene transfection have also generally been
unsuccessful. Transfection of ras oncogenes did not transform normal
finite life-span human fibroblasts to the malignant state even though
Southern analysis demonstrated the presence of the activated ras oncogene
in the genome of the transfectants (Sager et al., 1983). Nhen an
oncogenically activated, over-expressed H-ras gene was transfected into
normal human fibroblasts, the cells formed foci of morphologically
transformed cells, but such cells did not form malignant tumors in athymic
mice (Hurlin et al., 1987). Investigations in which other oncogenes
including v-sis, c-sis, H-ras (Fry et al., 1986, 1988), or N-ras (Wilson
et al., 1989) were transfected into normal human diploid fibroblasts also

generated no malignant cell strains.

3. Infinite Life Span Hulan Cell Strains
One possible explanation for the inability of ras oncogenes to

malignantly transform normal human cells is that the cells did not have

17

the ability to continuously replicate even though they were
morphologically transformed. This hypothesis suggested that what was
needed to obtain malignantly transformed human cells in vitro by
transfection of ras oncogenes was an infinite life-span human cell line
that could continue replicating indefinitely. Cell lines with an infinite
life-span could theoretically survive long enough to acquire all of the
changes necessary to become malignant. Namba and his colleagues were able
to generate infinite life-span fibroblast lines by repeated “Co gamma
exposure or repeated chemical carcinogen treatment of normal human
fibroblasts (Namba et al., 1981, 1985, 1988). The resulting cell lines
were not malignant, but did exhibit many features tumor-derived cells,
including altered cell morphology and reduced requirements for serum or
growth factors in cell culture. Similar experiments generated imortal
cell lines from normal human mamary epithelial cells (Stampfer and
Bartley, 1985). Such experiments have not been repeated by other
investigators. However, there is no doubt of their validity since
molecular techniques have demonstrated that the immortalized cells arose
from the parental cells (McCormick and Maher, 1988)

The infinite life-span human cell strains just described have
already acquired many features of transformation, such as anchorage
independence and the ability to grow in the absence of exogenous growth
factors. Because of their partial transformation, they are of limited
utility in investigating malignant transformation. McCormick and his
colleagues took another approach in order to create an infinite life-span,
human cell strain which was otherwise normal and therefore more useful for
studying in vitro malignant transformation. They transfected normal human
foreskin fibroblasts with the v-myc oncogene (Morgan et al., 1991). This

strategy was used because of other investigations which demonstrated that

18
rat embryo fibroblasts undergo spontaneous immortalization at a very low
frequency, but when such fibroblasts were transfected with myc oncogenes
the frequency of immortalization increased significantly (Land et al.,
1986). The v-myc construct employed by Morgen et al. (1991) also
contained a gene coding for resistance to geneticin, so transfectants from
the human fibroblasts could be selected for drug resistance. Five
independent clonal populations of geneticin resistant transfectants were
selected and were serially passaged to the end of their life-span. All
five clonal populations of v-myc, geneticin resistant cells eventually
sensesced. However, in one of the senescing populations a few cells
(perhaps a single clone) continued to proliferate. This cell strain has
continued to divide beyond 200 population doublings and is known as MSU-
1.1. The finding that the vast majority of that clonally-derived
population senesced indicates that an additional change or changes are
required, in conjunction with v—myc expression, to cause a normal human
fibroblast to become immortal. Isozyme and restriction fragment length
polymorphism patterns of the immortal MSU-1.1 strain are identical to that
of the normal parental fibroblast cell line from which MSU-1.1 was
generated demonstrating that MSU-1.1 arose from a normal human fibroblast

(Morgan et al., 1991).

4. Neoplastic Transforlation of Infinite Life-Span Hulan Cell Strains

There are a number of studies iniwhich immortalized human cell lines
have been neoplastically or malignantly transformed by oncogene
transfection. O’Brien et al. (1986) found that when SV40 immortalized
fibroblasts were infected with the Kirsten murine sarcoma virus, which
carries the K-ras oncogene, they acquired the ability to form small,

static nodules when injected into athymic mice. Similar experiments by

19

Rhim and his colleagues (1985) generated cells that formed malignant
tumors in athymic mice. Namba et al. (1988) found that a carcinogen
immortalized human cell line became malignantly transformed when it was
transfected with the K-raS' oncogene. (Namba et al., 1988). These
investigations could be interpreted to mean that some immortalized cells
need only a single change to become neoplastic. However, SV40
immortalized cell lines typically express features of transformation in
addition to infinite life-span, such as anchorage independence and growth
factor independence, and the immortalized fibroblasts generated by Namba
occurred only after several exposures to carcinogen treatments. The
implication is that each of these cell strains has already acquired many
of the changes necessary to become a neoplastic cell in addition to the
change(s) leading to imortality. Furthermore, expression of viral
oncogenes by transfected cells can be interpreted to consist of more than
a single change. This point is illustrated by studies comparing
constructs that express either the H-ras (Hurlin et al., 1989) or the N-
ras (Hilson et al., 1990) oncogene at either high or low levels. When the
imortal human fibroblast line MSU-1.1 was transfected with low expression
mutant ras oncogene constructs, the cells were not transformed to
malignancy, whereas transfection with mutated ras engineered for high
expression caused MSU-1.1 cells become malignantly transformed. These
observations suggest that at least two changes, high expression and
mutation, are necessary to cause a c-ras protooncogene to acquire the
transforming capability of a v-ras oncogene.

There are'a few’examples of infinite life-span human cell lines that
have been malignantly transformed by carcinogen treatment. SV40-
immortalized human fibroblasts treated with various carcinogens resulted

in strains that were malignantly transformed (Rhim et al., 1985, 1990;

20
Reznikoff et al., 1988). When the infinite human fibroblast line, MSU-
1.1, was treated with a single treatment of (+/-)-7B,8a-dihydroxy-9a,10a-
epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (BPDE) , the ultimate metabolite of
benzo[a]pyrene, then replated in low serum conditions and allowed to grow
to confluence, foci of morphologically transformed cells appeared (Yang et
al., 1992). A number of cell strains from these independent foci, were
assayed for various features of transformation including the ability to
form malignant tumors in athymic mice. Three of the eight focus-derived
strains formed progressively-invasive malignant tumors in athymic mice,
one strain formed a benign tumor, and others did not produce tumors at
all. Yang et al. demonstrated that focus formation alone does not predict
neoplastic transformation. Only focus-derived cells that were also
anchorage independent and growth factor independent were neoplastically
transformed. These results demonstrate that two or more changes are

necessary for malignant transformation of fibroblast strain MSU-1.1.

111. W

A. House Skin Carcinogenesis

1. The Two-Step Model

In 1941 Berenblum reported that co-treatment of mouse skin with
croton oil and the carcinogenic hydrocarbon, benzo(a)pyrene, resulted in
a significant increase in the frequency of benign and malignant neoplasms
compared to that found following benzo(a) pyrene treatment alone.
Berenblum tested croton oil and other non-carcinogenic agents as co-
carcinogens because of the studies of Rous and his associates which

suggested a causative 'role for "irritation", by chemical agents or

21

physical injury in carcinogenesis (Rous and Kidd, 1941; Friedwald, and
Rous, 1944). In 1944 Mottram showed that a single, sub-carcinogenic
application of 7,12-dimethylbenz(a)anthracene (DMBA), another carcinogenic
hydrocarbon, followed by ‘repeated treatments with croton oil, also
resulted in production of multiple skin tumors in mice. These and other
experiments led to the two-stage hypothesis of carcinogenesis: initiation
followed by promotion (Boutwell, 1974; Slaga, 1983). In mouse epidermal
carcinogenesis, initiation consists of the application of a sub-
carcinogenic dose'of carcinogen to the skin. Promotion refers to multiple
treatments of the skin with a tumor promoter after initiation has
occurred. Promoters are not carcinogens and do not cause tumors even when
applied at high doses over a long period of time. The promotor in croton
oil was found to be the phorbol ester, 12-o-tetradecanoyl-phorbol-13-
acetate (TPA) (Hecker,1968).

In the mouse skin model, initiation followed by successive rounds of
promotion results in the development of multiple benign papillomas, a few
of which develop into carcinomas after 20-60 weeks of TPA treatment
(Slaga, 1983). These observations imply that the initial carcinogen
treatment results in a number of initiated cells, each of which has an
irreversible genetic change. Promoter-induced proliferation results in
clonal expansion of these initiated cells, resulting in papillomas. If
promotor treatment is stopped, the papillomas usually regress. However,
if promotor treatment is continued some of the papillomas develop into
carcinomas. As a result of this observation the term progression has been
added to initiation and promotion to describe the events following
promotion which cause benign tumors (papillomas) to give rise to malignant

tumors (Burns et al., 1978; Albert et al., 1979). Therefore, the mouse

22
skin model suggests that there are at least three steps in the genesis of
a malignant cell from a normal cell.

Subsequent investigations using the mouse model further support the
hypothesis that mouse'epidermal carcinogenesis consists of multiplevsteps.
Shubik (1950) described four distinct phenotypic stages in progression
toward malignancy. These stages included (1) small tumors that appeared
and were present for at least four weeks but then regressed and did not
reappear, (2) papillomas that appeared and remained at a fixed size during
the extent of the experiment, (3) papillomas that grew progressively
during the extent of the experiment but did not attain malignancy and (4)
those that progressed to malignancy. Hennings et al. (1983) used a
variation in the standard mouse skin carcinogenesis protocol. In this
work initiation and promotion steps were followed by an additional
carcinogen treatment, resulting in a significant increase incidence of
carcinomas. This implies that carcinomas arise from benign papillomas as
the result of an additional DNA damaging event(s). In the classical
model, the additional genetic event(s) is spontaneous and occurs during
TPA induced proliferation. The rarity of a spontaneous mutation is
illustrated by the low frequency of papillomas that progress to
carcinomas. The second carcinogen treatment in the modified protocol
magnifies the number of carcinomas that occur by inducing DNA damaging
events in papilloma cells that already contain the first genetic
change(s). This work to refine the mouse skin model supports the
hypothesis that multiple steps are required for the genesis of a malignant

cell.

23
2. Oncogenes in Mouse Skin Carcinogenesis

The discovery of the rolei of retroviral oncogenes in animal
carcinogenesis and the finding that protooncogenes are often activated in
human and animal malignancies led investigators to hypothesize that known
oncogenes might become activated during mouse epidermal carcinogenesis.
Nhen mouse epidermal cells were stably infected with a retrovirus that
contained the H-ras oncogene, they did not express any phenotypic change
in the absence of promoter treatment. But when the infected cells were
treated with TPA, papillomas occurred, some of which progressed to
malignancy (Brown et al., 1986). Therefore, H-ras seems capable of
initiating mouse skin cells but other changes are necessary for the
initiated cells to become neoplastic. Examination of the viral
integration sites of these epidermal papillomas demonstrated that the
papillomas were polyclonal whereas the carcinomas were monoclonal in
origin. Since each monoclonal carcinoma arises within a papilloma
population, the multistep model of carcinogenesis is validated. In this
model, a single cell undergoes a genetic change which confers that cell
with a selective advantage, leading to clonal expansion of that cell.
Such a clone of cells then serves as a target for further genetic change
and selection.

Evidence that introduction of ras genes could initiate mouse skin
cells led to investigations to determine whether ras or other known cancer
genes were activated in papillomas and/or carcinomas arising in
carcinogen-induced mouse skin tumors. Quintanilla et al. (1986) found
that papillomas induced by DMBA and promoted with TPA nearly always
contain an A --> T transition at the second position of codon 61 of the H-

ras gene. Bianchi et al. (1990) and Brember and Balmain (1990)

24
demonstrated that amplification and/or over-expression of an activated H-
ras gene, possibly in association with loss of the remaining normal H-ras
allele, was associated with progression of papillomas to carcinomas. It
therefore'appears that ras activation can be responsible for two different
aspects of carcinogenesis in mouse skin.

Recent studies demonstrate that tumor suppressor genes also have a
role in mouse skin carcinogenesis. Alterations in the structure or
expression of the p53 tumor suppressor gene in benign or malignant cells
derived from carcinogen-induced mouse skin tumors suggest that it is
involved at a later stage in mouse skin carcinogenesis. p53 expression
was normal in papilloma-derived cells and well differentiated carcinoma
cells, whereas a high frequency of undifferentiated carcinomas were found
to contain p53 gene alterations including loss of heterozygosity at the
p53 locus or point mutations in one or both p53 alleles (Ruggeri et al.,
1991). Many of the undifferentiated carcinomas lacked the p53 protein,

which is in agreement with the gene data.

8. Hanan Colorectal Carcinogenesis

The mouse skin carcinogenesis model provided the first well—
characterized example of in vivo multistep carcinogenesis. Similar
studies in humans could not be carried out because of the obvious ethical
considerations. Nevertheless, the accessibility of normal and neoplastic
colorectal tissue via colonoscopy has provided investigators with the
opportunity to study genetic changes that have occurred in tissues
obtained from each of the identified stages in human colorectal
tumorigenesis. Analysis of the histological stages that occur in human
colorectal carcinogenesis have identified a progression of distinct

histological phenotypes including: (1) hyperplastic growth of apparently

25

normal colonic epithelial cells, (2) dysplastic hypertrophic lesions, (3)
small benign adenomas that increase in size and abnormal cellular
features, (4) malignant foci embedded within adenomas, (5) fully invasive
malignant lesions, and (6) carcinomas that have invasive and metastatic
capability (Haggitt and Reid, 1986). It is thought that most colorectal
carcinomas, sporadic or otherwise, arise from hyperplastic colonic
epithelial cells in a sequence similar to that just described (Sugarbaker
et al., 1985).

Several inherited syndromes have been identified that predispose
affected individuals to colorectal carcinoma including familial
adenomatous polyposis coli, hereditary nonpolyposis colorectal cancer
syndrome, Peutz-Jeghers syndrome, juvenile polyposis syndrome, Gardner’s
syndrome, Muir-Torre syndrome and Cowden’s disease (Haggitt and Reid,
1986; Hamilton, 1993). Furthermore, inflammatory bowel diseases,
including Crohns disease and ulcerative colitis, elevate the risk of
colorectal carcinoma in affected individuals, presumably through exposure
of the colonic epithelial cells to the effects of chronic inflammation
(Glickman, 1987). Investigation of neoplastic tissues obtained from
individuals affected with familial adenomatous polyposis coli (FAP),
hereditary nonpolyposis colorectal cancer (HNPCC) syndromes, as well as
sporadic colorectal cancers, have been valuable in identifying and
understanding genetic changes that occur in human colorectal

carcinogenesis.

1. Falilial Adenoaatous Polyposis Coli
Individuals affected with the syndrome known as familial adenomatous
polyposis coli (FAP) develop multiple benign colorectal adenomas early in

life, some of which inevitably progress to colorectal carcinomas if the

26

colon is not removed (Okamoto et al., 1988). This syndrome is transmitted
in an autosomal dominant fashion to offspring, and is relatively rare,
accounting for only about one percent of colorectal malignancies.
However, since each of the various stages observed in colorectal
carcinogenesis can often be identified in the colon of an affected
individual, and because clinical specimens are readily available, FAP
provides a unique opportunity to study in viva human multistep
carcinogenesis.

The pattern of inheritance and disease progression of FAP suggests
that affected individuals have inherited a constitutional genetic defect
that predisposes their colonic epithelial cells to the initial phenotypic
changes (such as hyperplasia) on the path toward colon carcinoma (Fearon
et al., 1987). Leukocytes obtained from individuals with FAP were
immortalized by Epstein Barr virus, then linkage analysis was performed,
localizing the germ-line defect to the long arm of chromosome five
(Leppert et al, 1987). Investigations of germ-line mutations in numerous
patients in several different affected families led to mapping and
sequencing of the gene, which is known as the APC gene (Nishisho et al.,
1991, Groden et al., 1991).

Prior to the cloning of the APC gene, Vogelstein (1988) had
hypothesized that the loss of a single APC allele, either as a germ-line
or somatic mutation, results in widespread epithelial hyperplasia, and
that such hyperplasia is the foundation from which hundreds to thousands
of benign adenomas arise. These predictions were based on studies of
allelic deletions in adenomas and carcinomas of the colon in normal
individuals as well as those affected with FAP. Once the APC sequence was
determined, sequencing experiments of the entire APC coding region in

immortalized leukocytes from individuals affected with FAP demonstrated

27

that 59 of 79 unrelated kindreds carried germ line APC gene mutations
(Miyoshi et al., 1992), confirming it as the defective gene in the
majority of individuals with a clinical diagnosis of FAP. Sequencing
experiments of colorectal neoplasms in normal individuals demonstrated
that 63% of adenomas and 60% of the carcinomas that occurred sporadically
also contained mutations in the APC gene (Powell et al., 1992).
Therefore, APC gene defects are associated with carcinogenesis in a
majority of sporadic colorectal tumors as well as those that occur in FAP
affected individuals. Furthermore, since the frequencies of APC‘mutations
in colorectal adenomas and carcinomas are essentially equal, these
findings also suggest that .APC' gene defects are early events in
colorectal carcinogenesis.

Analysis of the predicted protein structure of the APC gene product
at the time that the sequence became known showed no specific homology
with any known protein (Nishisho et al., 1991, Groden et al., 1991).
However, sequencing of the APC gene in individuals affected with FAP
revealed that greater than 90%iof such patients have mutations that result
in a messenger RNA that prematurely terminates during translation ,
suggesting that the carboxy terminus of the APC gene product is required
for normal function (Miyoshi et al., 1992). Antibodies targeted against
various regions of the predicted amino acid sequence of the APC protein
demonstrated that normal cells in individuals affected with FAP contain
both normal and truncated APC proteins. Cell lines derived from sporadic
and FAP colorectal neoplasms contained no normal APC protein, and 75% of
such cell lines contained truncated APC protein (Smith et al., 1993).
Normal APC protein was detected in 40/40 human tumor cell lines derived
from neoplasms that occurred in organs other than the colon in normal

individuals. These studies demonstrate that the loss of the normal APC

28
gene product is associated with the formation of most, if not all,
colorectal malignancies, but do not appear to be involved in sporadic
malignancies arising in other organs.

Recent studies have given some insight into understanding the normal
and abnormal function of the APC protein. Vogelstein and his colleagues
showed that fusion proteins containing the first 171 amino acids of the
APC protein bind to the normal APC protein in vitro (Su et al., 1993a).
Experiments designed to detect proteins that immunoprecipitate with the
APC protein revealed that o- and B- catenin coimmunoprecipitate with the
APC protein (Rubinfeld et al., 1993; Su et al., 1993b). The catenins are
proteins that interact with the cell adhesion molecule known as E-
cadherin. The results suggest two important points: (1) that defective
APC protein may be able to bind and to some degree interfere with normal
APC protein function (dominant-negative effect) and (2) that the normal

APC protein is involved in cell adhesion.

2. Hereditary Noinolyposis Colon Cancer

Hereditary non-polyposis colorectal carcinoma (HNPCC) is an
autosomal dominantly inherited predisposition to colorectal cancer that
was first described by Narthin in 1913. This disease is clinically
diagnosed in individuals who develop a colorectal carcinoma and fulfill
each of the following clinical criteria including: (1) three or more
relatives with histologically verified colorectal carcinoma, one of whom
is a first degree relative of the other two; (2) colorectal carcinoma
involving at least two generations; and (3) one or more colorectal
carcinoma cases diagnosed before age 50 (Vasen et al., 1991). Other
characteristics of this disease include a tendency to develop carcinomas

in the~proximal (or right) colon, occurrence of malignancies at an early

29
age, and predisposition to malignancy in other organs including the
ureter, kidney, stomach, small bowel, ovary, pancreas, biliary tract,
skin, larynx and blood (Lynch et al., 1993). This syndrome is much more
common than FAP and is thought to account for five to 15% of all
colorectal carcinomas (Hamilton, 1993).

Data from linkage analyses demonstrated that the HNPCC phenotype is
closely linked to a marker on chromosome two in some families affected
with HNPCC (Peltomaki, et al.. 1993) and to a marker on chromosome three
in other families affected with HNPCC (Lindblom et al., 1993).
Comparison of the types of genetic alterations in colorectal neoplasms
that occurred in individuals affected with HNPCC and in those that
occurred in non-affected individuals revealed a startling difference in
the nature of the mutations. Analysis of oncogenes or tumor suppressor
genes known to be frequently altered in colorectal cancers, including K-
ras, P53, and APC, demonstrated that each of these genes were mutated
slightly more frequently in HNPCC colorectal neoplasms than in non-HNPCC
neoplasms. What was most interesting and unexpected was that errors in
DNA replication appeared to account for an unusually high percentage of
the mutations that occurred during colorectal carcinogenesis in HNPCC
individuals (Aaltonen et al., 1993). Based on the known sequences of
mismatch repair genes in other species, investigators found the human
HNPCC gene on chromosome two to be homologous to the yeast hMSH2 gene
(Fishel et al., 1993; Leach et al., 1993). A short time thereafter,
similar strategies were applied in the identification of the hMLHI gene on
human chromosome three (Bronner et al., 1994; Papadopoulos et al., 1994).
The implications of these findings is that germ-line defects in either of

these genes cause defective mismatch repair during DNA synthesis.

30
Mutations in oncogenes or tumor suppressor genes that occur as the result

of this defective mismatch repair lead to colorectal neoplasia.

3. Other genetic Changes Associated with Colorectal Neoplasia

Mutations in ras genes have also been implicated in colorectal
tumorigenesis .. Approximately 50% of benign adenomas larger than one
centimeter and 50% of colorectal malignancies contain ras gene mutations,
whereas fewer than 10% of adenomas smaller than one centimeter have ras
mutations (805 et al., 1987, Forrester et al., 1987, Farr et al., 1988).
Therefore, it does not appear that ras mutations are often the initial
change in colorectal carcinogenesis. Nevertheless, since the frequency'of
ras mutations does not increase with the occurrence of malignant tumors,
it appears that ras mutations are usually a relatively early change in
colorectal carcinogenesis. Further evidence that ras mutations play a
role in the>early stages of colorectal carcinogenesis comes from the study
of a cell line established from a colorectal adenoma that occurred in a
person affected with FAP. This cell line, designated PC/AA, has a
mutation in K-ras codon 12 (Farr et al., 1988). Both early and late
passage PC/AA cells were not tumorigenic when injected into athymic nude
mice, even though colorectal carcinoma cell lines ‘typically formed
progressively growing malignant tumors in athymic mice.

Investigations of p53 gene mutations suggest that loss of p53
function is a late event in colorectal tumorigenesis. Vogelstein and his
colleagues used restriction fragment length polymorphism analysis to
detect loss of heterozygosity at the p53 locus on chromosome 17p (Fearon
et al., 1987). They obtained both normal and neoplastic tissues from a
number of patients. Tumor samples included in the study were those from

individuals whose normal cells were heterozygous at the p53 locus. Their

31
results showed that 76% (25/33) of informative colorectal carcinomas had
allelic deletions of the p53 gene, whereas only 3% (1/30) of the precursor
colorectal adenomas showed loss of heterozygosity at the p53 locus..
These findings suggest that loss of p53 function is often involved in
progression of colorectal adenomas to carcinomas.

Other genetic changes frequently associated with human colorectal
carcinogenesis include allelic deletion of chromosome 18q (Vogelstein et
al., 1988), elevated expression of the c-myc gene (Erisman et al., 1988)
and DNA hypomethylation (Goelz et al., 1985). A number of other, less
frequent changes have also been described including the allelic deletions
of other chromosomal sites (Fearon & Vogelstein, 1990). Although there
appears to be a preferred order in the sequence in which the genetic
events occur in colorectal carcinogenesis, it is the accumulation of
changes that results in a neoplastic cell. Furthermore, it does not
appear that any of the specific genetic defects identified so far are an
absolute prerequisite in colorectal carcinogenesis. For example, although
p53 gene defects have been identified in a majority of colorectal
carcinomas, a portion of such tumors have normal p53 alleles (Fearon et
al., 1987). Presumably there are other genes that participate in

colorectal carcinogenesis.

 

A common characteristic of the genes implicated in cancer is that

most, if not all, encode proteins that promote or regulate cell growth and

division. Growth of cells in multicellular organisms is highly regulated.

32

In human adults, cells in some tissues divide continuously, whereas cells
in other tissues divide rarely, if ever. For example, superficial cells
of the gastrointestinal tract and the skin are continuously sloughed off,
and constant, rapid cell division is required to replace the cells that
are lost. Blood forming cells also divide continuously to replenish aging
and dying cells. Conversely, neurons and cardiac muscle cells rarely if
ever divide in mature adults (reviewed by Cooper, 1992). Understanding
the mechanisms that control cell growth and division is central to
understanding the uncontrolled growth characteristic of neoplastic cells.
Many of the proteins involved in the mitogenic signal transduction have
been identified, but the precise mechanisms of signal transduction are not
completely understood. It is hoped that the discovery of the details of
mitogenic signalling mechanisms in normal and neoplastic cells will reveal
differences that can be used as targets for novel cancer treatments.

Classes of proteins involved in signal transduction include: (1)
growth factors, (2) growth factor receptors, (3) cytoplasmic proteins that
transduce growth signals to cellular effector sites and, (4) nuclear
factors that regulate the expression of genes involved in mitogenic
responses. The replication of normal cells is initiated by nfitogenic
growth factors (Deuel et al., 1988). Growth factors bind to specific
cytoplasmic or transmembrane receptors that become activated, and as a
result can transduce the mitogenic signal to effector proteins. The
signal eventually reaches the nucleus where specific genes are either
turned on or off in such a fashion that the cell becomes committed to DNA
synthesis and cell division. The multistep hypothesis of carcinogenesis
considers that a neoplastic cell is the result of a series of genetic
alterations. It is therefore plausible that synergistic or additive

defects in proteins involved in one or more signal transduction pathway,

33
including genes encoding proteins that inhibit growth and those that
promote growth, contribute to cells becoming neoplastic. The following
sections describe some of the details of signal transduction, with an
emphasis on the regulation and function of ras proteins and their pivotal
role in growth factor-initiated signalling pathways.

Viral ras oncogenes were first identified as the transforming
sequences of retroviruses. Shortly thereafter, cellular ras genes were
identified, and investigation of human tumor DNA identified transforming
sequences that were altered forms of c-ras genes (for review see Bishop,
1983). Ras genes are highly conserved in all eukaryotic cells including
yeast. In mammalian cells, the major function of ras proteins is to
transduce mitogenic signals, initiated at the cell surface or in the
cytoplasm, to the nucleus. In the ten or so years since ras oncogenes
were first characterized, many details of how ras proteins function in
signal transduction in normal and neoplastic mammalian cells have been

identified.

A. The GDP/GTP cycle

The ras proteins implicated in human carcinogenesis are part of a
large family of proteins known as guanine nucleotide binding proteins or
G-proteins (Simon et al., 1991). The activity of ras proteins is
dependent upon which guanine nucleotide is bound to the ras protein
(review by Santos and Nebreda, 1989). Ras proteins are in their inactive
conformation when they are bound by GDP and become activated when GDP is
replaced by GTP. The molecules that regulate whether ras is GDP- or GTP-
bound thereby determine the activation state of ras proteins.

In quiescent, normal cells, nearly all of the ras proteins are in

the inactive GDP-bound form (Haubruck and McCormick, 1991). Upstream

34

mitogenic signals cause ras to release GDP. Since the majority of free
guanine nucleotide in the cytoplasm is GTP, the empty guanine nucleotide
binding site of ras proteins quickly become filled by GTP, resulting in
the activation of ras. Normal ras proteins have intrinsic GTPase
activity, but the intrinsic activity is so low that hydrolysis of bound
GTP to GDP by the intrinsic ras GTPase activity alone is very slow (Hall
and Self, 1986). However, the ras GTPase activity can be enhanced by
proteins known as GTPase activating proteins or GAP proteins (Trahey and
McCormick, 1987). Hhen GAP proteins interact with ras, the elevated
GTPase activity causes rapid hydrolysis GTP to GDP, and as a result ras
becomes inactivated.

Mutations in ras genes have been implicated in more than 30% of
human malignancies (Barbacid, 1987). The result of activating ras gene
mutations is that the proportion of GTP-bound ras protein in a cell at any
one time is elevated, turning on the ras-mediated signal transduction
pathway(s). It is thought that this is the mechanism whereby mutations in
ras genes contribute to the unregulated growth that is characteristic of

cancer cells.

B. GTPase Activating Proteins

Two of GTPase activating proteins (GAPs) have been identified that
regulate as proteins in mammalian cells. p120-GAP was the first
discovered and is present in all mammalian cells (Trahey and McCormick,
1987). Data indicates that p120-GAP directly regulates ras proteins by
facilitating intrinsic ras GTPase activity. Hhen the type 1
neurofibromatosis gene was cloned and sequenced it was found to have
considerable homology to the gene encoding p120-GAP (Xu et al., 1990a).
Subsequent efforts demonstrated that the NF1-GAP protein, neurofibromin,

35
also binds to and stimulates the intrinsic GTPase activity of ras proteins

(Xu et al., 1990b).

1. p120~GAP

It was initially considered that the difference between mutant and
wild type ras proteins was that the activated mutant ras proteins were
deficient in their intrinsic GTPase activity and therefore, the level of
mutant ras bound to GTP was elevated relative to that of wild type ras.
However, in vitro studies of mutant and wild type ras proteins did not
reveal any significant differences between the GTPase activity of normal
and mutant ras (Trahey and McCormick, 1987). ‘These investigators analyzed
wild type N-ras, which has a glycine residue at amino acid position 12
(Gly-12 p21), and two mutant N-ras proteins, Asp-12 p21 and Val-12 p21.
xenopus oocytes provided a useful in vivo assay because of the previous
observation that activated (GTP-bound) human ras proteins can induce
xenopus oocyte'maturation, a change that is easily'observed (Birchmeier et
al., 1985). Upon injection into the oocytes, the mutant ras proteins
induced oocyte maturation much more efficiently than wild type ras. Hhen
:”P-labeled mutant or wild type ras proteins were recovered from oocytes,
mutant ras proteins remained mostly bound to GTP, whereas wild type ras
was bound mainly to GDP. When a non-hydrolyzable analogue of GTP was used
in these experiments, the wild type and mutant ras proteins were equally
potent inducers of oocyte maturation. Furthermore, when GOP was attached
to mutant ras proteins or the wild type ras protein, all were very
inefficient inducers of maturation. The ability of both mutant and wild
type ras to induce oocyte maturation depended upon whether GDP or GTP was

bound to the protein. The results indicated that some factor may be

36
present in the oocyte milieu that was able to stimulate GTP hydrolysis on
wild type ras but not on mutant ras protein.

To investigate whether such a ras-regulating factor might be present
in Xenopus oocytes, purified wild type or mutant ras proteins were
co-injected into oocytes with 3"’P-phosphate (Trahey and McCormick, 1987).
After 3 hours the proteins were extracted and immunoprecipitated with an
anti-ras antibody. Following immunoprecipitation, the radiolabeled
nucleotides were separated from the ras proteins and analyzed by thin
layer chromatography, revealing that wild type ras was almost entirely
GDP-bound, whereas mutant ras protein was mainly GTP-bound. A time course
demonstrated that the t,” of GTP bound to wild type ras was 2—3 minutes
whereas that of mutant ras was greater than 1000 minutes. The regulating
factor participating in the phenomenon, p120-GAP, enhanced the intrinsic
GTPase activity of wild type ras proteins but was unable to enhance the
GTPase activity of any of the mutant ras proteins used in the assay.
p120-GAP was subsequently isolated from oocytes as well as mamalian cells
(Trahey and McCormick, 1987) and within a year, the GAP gene was cloned
(Trahey et al., 1988).

2. Neurofibrouin

Since GTPase-activating proteins regulate ras activity, defects in
such proteins could cause elevation of the amount of activated ras in
cells and in that way contribute to neoplasia. In other words, normal
GAPs may act as tumor suppressors. Evidence to support a role for GAPs as
tumor suppressor proteins came from studies of cells derived from tumors
that occur at an elevated frequency in type 1 neurofibromatosis.
Individuals affected with NF1 develop multiple benign neurofibromas.

Neurofibrosarcomas, which occur as a result of malignant transformation of

37
a neurofibroma, occur at an elevated frequency in NF1 patients (Riccardi
and Eichner, 1986). Declue'et al. (1992) studied three cell lines derived
neurofibrosarcomas that occurred in NFI patients. Immunoprecipitation and
Western blot analysis demonstrated that the neurofibrosarcoma cells
produce almost no neurofibromin, whereas the control neoplastic (HeLa
cells) and nonneoplastic cells (NIH-3T3 cells) produce considerable
neurofibromin. The ratio of ras-GTP to ras-GDP was found to be elevated
in each of the three neurofibrosarcoma cell lines, similar to the high
levels of ras-GTP in v-H-ras-transformed NIH-3T3 cells. Since it was
possible that ras mutations, and not the absence of neurofibromin, was
responsible for the high proportion of GTP bound to ras in the
neurofibrosarcoma cell lines, they devised a strategy to distinguish
whether the ras proteins in each cell line behaved like wild type or
mutant ras proteins. Ras proteins were isolated from the three
neurofibrosarcoma cell lines and control cell lines and were bound in
vitro with [a-“Pl-GTP, and then were incubated to determine the intrinsic
GTPase activity of ras proteins from individual cell lines. The results
showed that the ras proteins isolated from v-H-ras-transformed NIH-3T3
cells remained greater than 80% GTP-bound, whereas ras proteins from all
other cell lines, including the neurofibrosarcoma cell lines, was only 30%
GTP-bound. This indicates that ras proteins from neurofibrosarcoma cells
retained their intrinsic GTPase activity and the ability to interact with
normal GAP proteins in the incubation milieu. When one of these
neurofibrosarcoma lines was transfected with a plasmid carrying the
catalytic domain of p120-GAP and a selectable'marker, the amount of ras in
the GTP-bound form in transfectant cells was reduced from 38% to 10%
(DeClue et al., 1992). The findings that neurofibrosarcoma-derived ras

proteins had normal intrinsic ras-GTPase activity in vitro, and that the

38
amount of ras-GTP in such cells was significantly reduced by the addition
of a normal GAP catalytic domain, strongly suggests that the ras proteins
in the neurofibrosarcomas were normal and that the elevated level of ras-
GTP in neurofibrosarcoma cells was due to deficient GTPase activity.

Comparison of the ratio of ras-GDP verses ras-GTP in sub-confluent
and confluent ras-transformed NIH 3T3 cells in culture suggests that the
activity of GAP proteins is effected by cell density (Zhang et al., 1992).
The activity was measured by incubating recombinant ras proteins bound to
[a-”P]GTP in buffer containing cell extracts obtained at high or low
density. The cell extract obtained from cells at high density had four
times the GTPase activity as cell extracts isolated from sub-confluent
cells, suggesting that cell-cell contact induces elevation of p120-GAP
and/or NFI (or other GAP proteins) GTPase activity.

One question regarding ras regulation that continues to be explored
is the role of neurofibromin and p120-GAP in the regulation of ras
proteins in various cell types. Both of these GAPs are expressed in all
normal cells that have been examined (Golubic et al., 1992), although
neurofibromin is expressed at a ten-fold higher level in brain cells than
in cells from kidney, liver, lung, or spleen (Daston et al., 1990 ).
Neurofibromin appears to regulate ras activity in Schwann cells, the cell
of origin of neurofibromas and neurofibrosarcomas. It is unclear why
p120-GAP does not compensate for the deficiency of neurofibromin GTPase

activity in such neurofibrosarcoma cells.

C. Guanine Nucleotide Releasing Factors
The ras-GDP complex is extremely stable in the presence of
physiologic concentrations of Hg“ in vitro (Nolfmaniand Macara, 1990), and

nearly all ras is bound to GDP in normal quiescent cells despite the

39
relative abundance of GTP in the cytosol (Haubruck and McCormick, 1991).
However, ras rapidly releases GDP and becomes GTP-bound in response to
specific external stimuli, suggesting the presence of some factor or
factors that have the ability to catalyze the release of GDP from the ras-
GDP complex.

Guanine nucleotide releasing factors (GRFs) were first identified in
non-mammalian systems. The first was the CDC25 gene product, Cdc25p, of
Saccharomyces cerevisiae (Jones et al., 1991). Interest in the CDC25 gene
came about as a result of its observed role in the regulation of ras-
mediated cyclic AMP activation in yeast (Camonis et al., 1986). In yeast,
GTP-bound ras stimulates adenylate cyclase, which is necessary for
entrance into the cell cycle, and the CDC25 gene product is necessary for
ras activation (Broek et al., 1987). The $05 gene of Drosophila
melanogaster is homologous to yeast CDC25, and it also triggers release of
GDP from ras (Simon et al., 1991). Because the structure and function of
ras is so well conserved among diverse species, it is likely that proteins
similar to the C0025 product function in mammalian cells to mediate the
release of GDP from ras proteins in response to signals that activate ras.

To test for the presence of a factor that promotes the dissociation
of GDP from mamnalian ras, Holfman and Macara (1990) incubated ras
proteins bound to [o’zPl-GDP in rat brain homogenate containing excess
unlabeled GDP and Mg“. Under these conditions the half-life of labeled
GDP bound to ras was two minutes, compared to more than 60 minutes in the
absence of brain cytosol (homogenization buffer only). To ensure that GAP
proteins in the homogenate were not the factors responsible for
dissociation of GDP, GAP proteins were also added to the homogenization
buffer (cytosolic fraction) in a control incubation and were found to have

no effect on the release of 3"'P--labeled GDP from ras. This suggested that

40
some factor in the homogenate other than GAP proteins caused the release
of labeled GDP from ras proteins.

Shou et al. (1992) isolated a mamalian (rat) cONA encoding a
guanine nucleotide releasing factor by using a set of degenerate
oligonucleotide primers based on highly conserved regions of the yeast
homolog. The rat gene showed approximately 30% homology to yeast and
fruit fly guanine nucleotide releasing factor genes. Guanine nucleotide
releasing factors have also been identified or isolated from human
placenta (Downward et al., 1990) and bovine brain (Huang et al., 1990).
It remains unclear whether the guanine nucleotide releasing factors
isolated from different types of mamalian cells represent a single
protein encoded by homologous genes in various species and cell types, or
whether more than one factor is involved. Characterization of this
protein or proteins is critical to fully understand ras-mediated signal
transduction pathways, particularly the mechanisms(s) whereby ras proteins

become activated .

D. Mitogenic Growth Factors

The importance of ras proteins in growth factor induced mitogenesis
in mammalian cells was first illustrated by studies in which neutralizing,
anti-ras antibodies were micro-injected into quiescent NIH-3T3 cells prior
to stimulation by fetal bovine serum. Microinjection of approximately 3
x 10‘ molecules of ras antibodies (0.2% of total cell protein) into the
cells, prevented DNA synthesis and subsequent cell division in response to
fetal bovine serum stimulation (Mulcahy et al., 1985). Similar amounts of
control antibodies had no effect on DNA synthesis and cell proliferation.

In addition, the accumulation of GTP-bound ras in rodent fibroblasts

41
occurred in response to the addition of specific growth factors, such as
PDGF (Satoh et al., 1990a) and EGF (Satoh et al., 1990b).

Hhen mitogenic growth factors such as PDGF and EGF bind with the
external domain of their respective receptors, the internal domains of the
receptors become autophosphorylated on tyrosine residues. The activated
internal domains interact with cytoplasmic substrates and phosphorylate
tyrosine residues on these substrates (Heldin and Nestermark, 1991).
Examples of such cytoplasmic substrates include phospholipase C-y (PLC-y)
(Meisenhelder et al., 1989), p120-GAP (Kaplan et al., 1990; Ellis et al.,
1990) and phoshatidylinositol 3-kinase (PI3K) (Coughlin et al., 1989;
Kavanaugh et al., 1992). Although GAPs have been implicated in the
regulation of ras protein activity, the possible role of PLC-v or PI3K in
ras activation remains to be elucidated.

Elevation of ras-GTP levels as a result of growth factor stimulation
of normal cells may occur because phosphorylation of p120-GAP abrogates
the ability of p120-GAP to enhance the GTPase activity of ras proteins
(Molloy et al., 1992). In non-transformed NIH 3T3 cells, or those
transformed by over-expressed c-ras, stimulation by PDGF significantly
increases the proportion of GTP-bound ras (Molloy et al., 1992). However,
when v-ras-transformed NIH 3T3 cells are stimulated with PDGF there is
little or no increase in the already elevated proportion of ras bound to
GTP. One interpretation of these findings is that some ras mutations
result in ras proteins that are no longer subject to GAP regulation, and
as a result do not require growth factor-induced phosphorylation of GAP to
increase the amount of activated ras proteins. Alternatively, some types
of mutant ras may have constitutionally high guanine nucleotide releasing

activity, thereby increasing the amount of ras-GTP.

42

Much less is know about guanine nucleotide releasing factors (GRFs)
than GAP proteins in regard to tyrosine kinase signaling pathways. Hork
by Zhang et al (1992) suggests that either PDGF or serum induced elevation
of GTP—bound ras in NIH 3T3 cells is a consequence of increased GRF
activity. Two functionally different c-H-ras mutations were employed to
arrive at this conclusion. One mutant ras gene encodes a protein that has
a single amino acid substitution at amino acid 116 (c-H-rasu‘), the site
of guanine nucleotide binding. Ras proteins with this mutation retain
their normal sensitivity to regulation by p120-GAP and NFI, but have an
10-fold increase in the intrinsic guanine nucleotide exchange rate (Der et
al., 1988). With c-H-ras‘“ transfectants, if the serum/PDGF-induced
elevation of ras-GTP depends on inactivating p120-GAP or NFI, the'addition
of serum or PDGF should increase the level of ras-GTP. Conversely, if
serum/PDGF-induced increase in GTP-bound ras is dependent upon activation
of a GRF, then serum or PDGF stimulation should have no effect on the
level of ras-GTP. Nhen confluent NIH 3T3 cells transfected with c-H-ras“‘
were treated with serum or PDGF, the PDGF receptors were activated, yet
there was no increase in the level of GTP-bound Ras. To verify that the
effect of serum is to stimulate GRF activity, another ras protein mutated
at amino acid 36 was used in similar experiments. Mutant c-H-ras“‘has
about the same transforming ability as c-H-ras‘“ as judged by the
frequency of foci formed in NIH 3T3 cells and the level of GTP bound to
ras in each of the transfected cell strains. However, in contrast to NIH
3T3 cells transfected c-H-ras‘“, those transfected with c-Horas36 are
resistant to regulation by p120-GAP and neurofibromin, and the cells
release GDP at the same rate as cells with non-mutated ras. Nhen
confluent NIH 3T3-ras” transfectants in serum free conditions were

stimulated with serum or PDGF, there was a three to four-fold increase in

43

the level of GTP bound to ras. Since the c-H-ras" protein is not
responsive to GAP or neurofibromin regulation, the increase in GTP-bound
ras was very likely the result of growth factor mediated stimulation of
guanine nucleotide releasing activity. This suggests that the mechanism
whereby ras becomes GTP-bound in response to serum or growth factor
stimulation at least partially involves stimulation of guanine nucleotide
releasing factor activity. Recent investigations have shown that
activation of the epidermal growth factor receptor in manmalian cells also
leads to activation of guanine nucleotide releasing factor, leading to
subsequent activation of ras proteins (Gale et al., 1993; Buday and
Downward, 1993).

A combination of GTPase activity and guanine nucleotide releasing
activity is responsible for ras protein regulation. Oncogenic synergy
could occur as a consequence of elevated growth factor expression and the
loss of normal GAP protein activity. Endogenous production of mitogenic
growth factors increases the formation of GTP-ras by promoting the release
of GDP from ras proteins. Loss of GTPase activity results in decreased
hydrolysis of GTP—ras. The overall effect of these combined changes would
be to increase the fraction of ras that is bound to GTP. This observation
demonstrates both oncogenic synergy and the importance of mitogenic signal

transduction to maintaining normal cell growth.

 

A. The Transforaing Gene of SSV is Hoaologous to the c-sis (PDGF-B)
Gene

44

By the early 1980’s investigators had identified several oncogenes
and were attempting to discover the mechanisms whereby each of these
accomplished its neoplastic effect. Sporn and Todaro (1980) demonstrated
that when grown in agar, normal cells grow like tumor cells if a fraction
of the spent medium from tumor cells is added to the normal cells. From
one of these tumor cell lines, which was derived from a human
osteosarcoma, a PDGF-like growth factor'was isolated (Heldin et al, 1980).
Based on these facts, Sporn and Todaro (1980) purposed the autocrine
hypothesis of carcinogenesis, namely that cancer cells produce a mitogenic
growth factor(s) and the appropriate receptor(s) for the growth factor,
resulting in continuous growth and division of such cells. Although it
appeared that growth factors might play a causal role in the development
of human and animal cancers, there was no evidence to indicate a direct
relationship between any known transforming gene and a growth factor.

Support for the autocrine hypothesis came from different groups of
scientists who were investigating what appeared to be unrelated questions.
First, the transforming gene of the simian sarcoma virus (SSV), was
identified and isolated from a sarcoma that occurred in a woolly monkey,
and was sequenced by investigators at the National Cancer Institute
(Devare et al., 1983). About three~ months later, Antoniades and
Hunkapiller (1983) reported the partial amino acid sequence'of PDGF. When
they published their work they were unaware of the homology between the
protein sequence of PDGF and that of the SSV product. Less than two
months after the sequence of PDGF was published, a protein database was
used to demonstrate the sequence homology between the amino acid sequence
of PDGF and the SSV encoded protein. (Haterfield et al., 1983; Doolittle
et al., 1983). This was the first direct evidence that a growth factor

gene could act as a transforming oncogene.

45
8. Structure and Function of PDGF-B and Its Receptor

PDGF isolated from human platelets consists of two distinct but
related polypeptides designated the PDGF-A chain and the PDGF-B chain.
The active protein is a dimer including either of the homodimers (PDGF-AA
and PDGF-BB) or the heterodimer (PDGF-AB) (reviewed by Hannink and
Donoghue, 1989). The human PDGF-A chain (Betsholtz et al., 1986) and
PDGF-B chain (Swan et al., 1982) are encoded on chromosomes 7 and 22
respectively, and are approximately 50% homologous. The viral and normal
vertebrate genes that encode PDGF-8 are* known as v-sis and c-sis,
respectively.

Although their sequences are nearly identical, there are significant
differences between the mature c-sis and the v-sis mRNAs. Mature c-sis
mRNA includes unusually long 5’ and 3’ untranslated sequences (Chiu et
al., 1984). Studies in which the 5’ untranslated sequence of c-sis was
deleted demonstrated that the 5’-untranslated sequence exerts a negative
regulatory effect on the translation efficiency of the mature c-sis
transcript (Rao et al., 1988; Fen and Daniel, 1991; Franklin et al.,
1991). The SSV-derived v-sis gene is under virally encoded
transcriptional and translational control and lacks the 3’ and 5’
untranslated sequences that inhibit translation efficiency of c-sis mRNA.
Comparison of v-sis and c-sis coding regions has shown that there are only
a few amino acid differences (Chiu et a., 1984). Since the c-sis gene of
the woolly monkey, the animal from which SSV originated, has not been
sequenced it is not known whether the differences between the coding
regions of v-sis and c-sis are mutations that contribute to the
transforming activity of v-sis or whether they are due to species specific
differences between the woolly monkey and the human c-sis gene. Nhen the

c-sis coding sequence was inserted into the genome of a non-transforming

46
retrovirus, the recombinant retrovirus was able to transform cells that
subsequently formed sarcomas when injected into mice (Pech et al., 1989),
suggesting that the differences in the coding sequences between v-sis and
c-sis are unimportant, and that high levels of v-sis or c-sis gene
products generated by viral vectors account for the transforming ability
of sis genes.

The PDGF protein initiates mitogenesis by binding to cell surface
receptors which leads to dimerization of the receptors. There are two
different PDGF receptor peptides, known as the PDGF-o and PDGF-B
receptors, which form receptor dimers when they are bound by PDGF (Yarden
et al., 1986; Gronwald et al., 1988) There are three possible receptor
dimer combinations including on, a8 and 88. The PDGF-AA growth factor
dimer can bind only to the PDGF-on receptor, whereas PDGF-AB can bind to
either PDGF—an or PDGF-o8. PDGF-BB can bind to any of the receptor
dimers, but binds to the PDGF-BB receptor dimer with significantly higher
affinity than either PDGF-AB and PDGF-AA dimers (Heldin, 1992). Each of
the receptor dimers (any combination) can transduce mitogenic signals, but
only receptor dimers that include the PDGF-B receptor peptide participate
in chemotaxis and actin organization (Nister et al., 1988; Eriksson et
al., 1992). Furthermore, the PDGF-B/c-sis protooncogene is much more
efficient than PDGF-A oncogene as a transforming gene in animal and human
cells (Bywater et al, 1988; Beckmann etial., 1988). The full significance
of the various types of platelet-derived growth factors and receptor

dimers and their interactions is still not completely understood.

C. The Role of PDGF-B Expression in Neoplasia
Following the discovery of near complete sequence homology between

v-sis and the human c-sis gene, investigators sought to understand the

47
mechanism whereby the v-sis oncogene transforms cells, and to determine
whether such a mechanism is activated in human tumor cells that express
POGF-B/c-sis. The simplest assumption was that production and secretion
of PDGF by cells that express the PDGF receptor resulted in continuous
receptor activation, and therefore continuous transduction of downstream
mitogenic events, including cell division. ‘The following sections discuss
the role of PDGF-B expression in neoplastic cells generated in cultureiand

those which have been derived from benign and malignant human tumors.

1. Autocrine Activation of the PDGF-receptor

In response to PDGF stimulation, PDGF receptors become
phosphorylated on cytoplasmic tyrosine residues (reviewed by Heldin,
1992). Anti-phosphotyrosine antibodies have therefore been useful in
detecting activation of PDGF receptors in inmunoblot assays (Ek and
Heldin, 1984; Zippel at al., 1986). Nhen NIH-3T3 cells or normal human
fibroblasts in culture are placed in medium containing no exogenous growth
factors, PDGF receptors remain present on the cell surface and are not
phosphorylated on tyrosine residues (Ross et al, 1986). Under similar
conditions, PDGF-B receptors of SSV-transformed NIH-3T3 cells and SSV-
transformed human fibroblasts areldown-regulated (i.e., are not present on
the cell surface due to internalization) and constitutively phosphorylated
on tyrosine residues (Huang and Huang, 1988; Fleming et al., 1992b;
Keating and Hilliams, 1988). Hhen exogenous PDGF or serum is added to
quiescent, normal human fibroblasts or NIH-3T3 cells in culture, their
PDGF receptors become tyrosine phosphorylated and down-regulated like the
receptors of SSV/v-sis transformed cells (Ross et al., 1986). Since PDGF-
B receptors are constitutively phosphorylated in SSV/v-sis transformed

cells, and the phosphorylation of PDGF-B receptors is associated with

48
POGF-B-induced cell division in normal cells, SSV/v-sis could bring about
transformation of cells by continuously producing PDGF-B proteins, and in
that way activating PDGF-B receptors by continuous tyrosine
phosphorylation.

Hhen PDGF is added to quiescent, normal fibroblasts in serum free
culture conditions they enter into cell division (Heldin et al., 1981;
Johnsson et al, 1985). But when anti-PDGF neutralizing antibodies are
added to the quiescent fibroblasts in addition to the exogenous PDGF, the
PDGF-induced growth of normal fibroblasts is prevented. If SSV/v-sis
transforms cells solely through an autocrine mechanism by secreting PDGF-B
and activating PDGF receptors extracellularly, then neutralizing
antibodies should inhibit the effects of autocrine activation as they do
for normal cells. When neutralizing anti-PDGF antibodies were used to
inhibit the growth of SSV transformed cells in serum free medium, only
partial inhibition (Johnsson et al., 1985) or no inhibition (Huang et al.,
1984; Huang and Huang, 1988) was achieved. These findings suggest either
that the autocrine interaction of PDGF with its receptor occurs in an
internal compartment that is unaccessible to neutralizing antibodies, or
that some other mechanism, other than external receptor activation, is
responsible for the ability of SSV transformed cells to grow in medium

lacking exogenous growth factors.

2. Experilents with Suraain
To examine whether autocrine activation involving PDGF and the PDGF
receptor' occurs in an internal compartment, investigators used the

antifilarial drug, suramin, which inhibits interactions between PDGF and

49

the PDGF receptor (Williams et al., 1984; Garrett et al., 1984), as well
as interactions between other growth factors and their receptors (Minniti
et al., 1992, Sato and Rifkin, 1988; Betsholtz et al., 1986; Mills et al.,
1990). Although suramin has been reported to affect various enzymes such
as DNA polymerase (Jindal et al., 1990) and reverse transcriptase (Levine
et al, 1986), the effects of suramin at lower concentrations (up to
100ug/ml) appear to be relatively specific for interferring with growth
factor/growth factor receptor interactions in culture systems using cell
lines such as NIH-3T3, MSU-1.1, and strains derived from these cell lines.

Concentrations of 50-100ug/ml of suramin inhibit the transformed
phenotype of SSV transformed NIH-3T3 cells in culture (Garrett et al.,
1984, Betsholtz et al., 1986) and reverse the down regulation of the PDGF-
receptor in such cells. Within a few minutes of suramin treatment, PDGF
receptors in SSV-transformed NIH 3T3 cells are no longer tyrosine
phosphorylated (Huang and Huang, 1988). However, when NIH 3T3 cells are
transformed by oncogenes that are not dependent upon growth factor
activation, suramin has no effect. For example, suramin has no inhibitory
effect on erbB-2-transformed NIH-3T3 cells (Fleming et al., 1989).
Whether suramin exerts its inhibitory effect on the cell surface and/or in
the cytoplasm is not known. However, it is likely that suramin can
inhibit growth factor interactions both intracellularly and on the cell
surface. Like neutralizing antibodies, suramin inhibits the proliferative
effects that TGF-a and PDGF-B exert on normal fibroblasts in culture.
Furthermore, suramin, unlike neutralizing antibodies against growth
factors, completely inhibits the growth of SSV/v-sis transformed cells in
culture (Huang and Huang, 1988; Keating and Williams, 1988).

Such observations have been validated in the NSU-1.1 lineage

developed in this laboratory by' Morgan et al. (1991). When the

50
immortalized, non-neoplastic human fibroblast line, MSU-1.1, was
transfected with a plasmid carrying the v-sis gene, transformants formed
large colonies in soft agar and grew rapidly in serum free medium. The
addition of 50ug/ml suramin to the experimental conditions completely
prevented the MSU-1.1-v-sis cells from forming colonies in soft agar and
from proliferating in serum free medium (Yang et al., 1994). If the
effect of suramin is to interrupt growth factor/receptor interactions,
oncogenes such as H-ras and erbB-2, which act downstream of growth factor
initiated signalling pathways, would not be expected to be affected by
suramin. When serum free growth and anchorage independence of MSU-1.1-v-H-
ras cells were studied, they were unaffected by levels of up to IOOug/ml
suramin (Yang et al., 1994). This data supports the hypothesis that
SSV/v-sis-transformed cells depend upon PDGF activation of PDGF receptors
in an autocrine manner to attain the transformed phenotype. Because
suramin completely inhibits growth of SSV/v-sis transformed cells and
anti-PDGF antisera causes slight or no inhibition of such cells, PDGF-B
appears to have the capability to activate the PDGF receptors through an

internal autocrine mechanism (Keating and Williams, 1988).

3. PDGF-B Expression in Nulan Tulor Cells

There are several reports describing the co-expression of PDGF-B and
the PDGF-B receptor in fixed human tumor tissue and in human tumor-derived
cell lines (for review see Silver, 1992). Some of the types of human
neoplasms that commonly express PDGF-B and its receptor include glial cell
tumors (Maxwell, 1990; Fleming et al., 1992a; Hermanson, 1992), tumors
that arise in mesenchymal tissues (Richter and Graves, 1988; Fleming et

al., 1992b; Malik et al., 1991), gastric carcinomas (Chung and Antoniades,

51
1992), and lung carcinomas (Antoniades et al., 1992). ‘These studies
suggest that autocrine mechanisms involving PDGF-B have a causal role in
the formation of various tumors, but do not directly proove that an
autocrine mechanism contributes to the neoplastic phenotype.

Heldin and Westermark and their colleagues studied a human
osteosarcoma cell line and attempted to demonstrate the presence of
autocrine activation involving PDGF. This cell line, known as U-2 OS,
expresses PDGF-B and its receptor (Betsholtz et al., 1984). Efforts to
abrogate in vitro features of transformation exhibited by cell line U-2 OS
(and other human tumor cell lines) using receptor antibodies or suramin
have failed to show any inhibitory effect (Heldin and Westermark, 1991).
For example, suramin reversed PDGF receptor down regulation in U-2 05
cells, but caused no inhibition of the DNA synthesis associated with cell
division (Richter and Graves, 1988). Since suramin was shown to
significantly inhibit DNA synthesis of SSV/v-sis transformed cells in
culture (Fleming et al., 1989), induction of DNA synthesis and subsequent
cell division in U-2 OS cells in serum free medium is not dependent upon
an autocrine mechanism involving PDGF-B and its receptor. Suramin also
does not stop the serum free growth of several established human
fibrosarcoma cell lines, such as HT-1080 (Dave Reinhold, unpublished
results) that have previously been shown to express PDGF-B mRNA and
protein (Pantazis et al., 1985). However, one recent report (Fleming et
al., 1992b) showed that the growth of other human tumor cell lines are
adversely affected by suramin. In this study 100 ug/ml of suramin
inhibited the serum free growth of a number of human sarcoma and glial
cell tumor lines to various degrees. Because suramin is not a specific
inhibitor of PDGF-B, it is possible that other activated growth factor

pathways in these tumors are being inhibited by suramin. Therefore, there

52
is still no conclusive, direct evidence that PDGF-B has any effect on the

neoplastic phenotype of any cell line derived from a human tumor.

4. Paracrine Mechanisas in Neoplasia

Failure to demonstrate that PDGF-B expression had a role in any
feature of transformation in human tumor-derived cells led to the
speculation that PDGF-B expression might be is an early change in
multistep carcinogenesis. For example, PDGF-B expression might fulfill a
role similar to that of TPA in mouse skin carcinogenesis, namely, by
causing proliferation of an abnormal cell, thereby providing a much larger
population of cells, which are the target for further oncogenic changes.
In such a scenario it is possible that additional oncogenic changes
eventually render the production of PDGF-B unnecessary for the malignant
phenotype. Another possibility is that PDGF-B secreted by tumor cells
acts on non-neoplastic cells in the tumor environment. Some human cancer
cells produce PDGF-B but have no PDGF receptors (Rozengurt et al., 1985;
Westermark et al., 1986; Peres et al., 1987), suggesting that paracrine
effects of growth factors, i.e., contribute to neoplasia. Using
immunohistochemical methods and in situ hybridization, Heldin and
Westermark and their colleagues showed that non-neoplastic cells in the
tumor environment, including hyperplastic stromal cells and vascular
endothelial cells, express high levels of PDGF-B receptor mRNA and
protein, whereas the PDGF-B expressing neoplastic cells express low levels
of PDGF-B receptor protein and mRNA (Hermanson et al., 1992). Similar
studies showed that AIDS-related Kaposi sarcomas consist of two
populations of cells, a minority population of cells that express PDGF-B
and the PDGF-B receptor, and a majority population that express only the
PDGF-B receptor (Sturzl et al., 1992). The cells that express only PDGF-B

53

receptors appeared to be normal fibroblasts and/or smooth muscle cells
rather than neoplastic cells. These observations support the hypothesis
that neoplastic cells use normal cells from the tumor environment for some
undefined role in neoplasia.

PDGF-B is a strong chemo-attractant for fibroblasts (Seppa et al.,
1982) and endothelial cells (Bar et al., 1989). Proliferation of
fibroblasts and endothelial cells is required for stroma formation, which
is the formation of connective tissue and new blood vessels
(neovascularization) in the tumor environment. The ability to stimulate
neovascularization is a characteristic of malignant tumors and is
considered necessary to provide nutrients for rapidly growing tumors. To
examine whether the expression of PDGF-B contributes to stroma and blood
vessel proliferation in vivo, Forsberg et al (1993) transfected human WM9
melanoma cells, which do not express PDGF-B, with a plasmid carrying the
coding sequence of human PDGF-B and a neomycin resistance gene. Control
cells were injected with a construct that differed from the experimental
construct in that it lacked the PDGF-B gene. PDGF-B- expressing
transfectants and control transfectants were injected into five athymic
mice each. All of the mice developed progressively growing malignant
tumors. The PDGF expressing cells produced one cubic centimeter tumors in
five weeks. In contrast, control-transfected cells required two
additional weeks to reach an equivalent size (not considered to be
significant by the authors). Examination of histological sections of the
tumors demonstrated a dramatic difference in stromal formation. PDGF-B
expressing cells formed a tumor with a well developed network of
connective tissue and blood vessels surrounding and penetrating the
tumors. In addition, the vasculature and connective tissue of the PDGF-B

producing cells were organized in nodular, connective tissue encapsulated

54
regions. Mock-transfected cells and untreated WM9 cells produced tumors
that lacked substantial connective tissue and blood supply. Furthermore,
large areas of necrosis were present in tumors produced by mock-
transfected or non-transfected WM9 cells. Similar areas of necrosis were
not found in the tumors formed by the PDGF-B transfected cells. Tissue
necrosis presumably results from a blood supply that is insufficient to
meet the nutritional and waste disposing demands of highly metabolic
cancer cells. The results of these experiments indicate that PDGF-B can

participate in stroma formation in human cells in vivo.

LIST OF REFERENCES

Aaltonen, L.A. Peltomaki, P., Leach, F.S., Sistonen, P., Pylkkanen, L.,
Mecklin, J-P., Jarvinen, H., Powell, S.M., Jen, J., Hamilton, S.R.,
Petersen, G.M., Kinzler, K.W., Vogelstein, 8., and de al Chapelle, A.
(1993) Clues to the pathogenesis of familial colorectal cancer. Science,
260:812-816.

Albert, R.E., Burns, F.J. and Altshuler, B. (1979) Reinterpretation of the
linear non-threshold dose-response model in terms of the initiation-
promotion mouse skin tumorigenesis. Adv. Mod. Toxicol., 1:89-85.

Andersson, P., Goldfarb, M.P. and Weinberg, R.A. (1979) A defined
subgenomic fragment of in vitro synthesized Moloney sarcoma virus DNA can
induce cell transformation upon transfection. Cell, 16:63-75.

Antoniades, H.N. and Hunkapillar, M.W. (1983) Human platelet-derived
growth factor (PDGF): amino terminal amino acid sequence. Science,
220:963-965.

Antoniades, H.N., Galanopoulos, T., Neville-Golden, J. and O’Hara, C.J.
(1992) Malignant epithelial cells in primary human lung carcinomas
coexpress in vivo platelet-derived growth factor (PDGF) and PDGF receptor
mRNAs and their protein products. Proc. Natl. Acad. Sci. USA, 89:3942-
3946.

Bar, R.S., Boes, M., Booth, B.A., Dake, B.L., Henly, S. and Hart, H.N.
(1989) The effects of platelet-derived growth factor in cultured
microvessel endothelial cells. EndocrinolOOY. 124:1841-1848.

Barbacid, M. (1987) ras genes. Annu. Rev. Biochel., 56:779-827. 1987.

Barrett, C.J. (1987) Genetic and epigenetic mechanisms of carcinogenesis.
In: Mechanisas of Environaental Carcinogenesis, Barrett C.J. (Ed.), CRC
Press Inc, Boca Raton, Fl, 1-15.

Barrett, C.J. and Ts’o, O.P. (1978) Evidence for the progressive nature'of
neoplastic transformation in vitro. Proc. Natl. Acad. Sci. USA, 75:3761-
3765.

Basu, T.N., Gutmann, D.H., Flethcher, A., Glover, T.W., Collins, F.S. and

Downward, J. (1992) Aberrant regulation of ras proteins in malignant
tumour cells from type 1 neurofibromatosis patients. Nature, 356: 713-715.

55

56

Beckmann, M.P., Betsholtz, C., Heldin, C.H-., Westermark, 8., Di Marco,
E., Di Fiore, P.P., Robbins, K.C. and Aaronson, S.A. (1988) Comparison of
biological properties and transforming potential of human PDGF-A and PDGF-
8 chains. Science, 241: 1346-1349.

Berenblum, I. (1941) The mechanism of carcinogenesis. A study of the
significance of co-carcinogenesis and related phenomena. Cancer Res.,
1:807-814.

Berwald, Y. and Sachs, L. (1965) In vitro transformation of normal cells
to tumor cells by carcinogenic hydrocarbons. J. Natl. Cancer Inst.,
35:641-657.

Berwald, Y. and Sachs, L. (1963) In vitro cell transformation with
chemical carcinogens. Nature, 200:1182-1184.

Betsholtz, C., Johnsson, A., Heldin, C.-H. and Westermark, B. (1986)
Efficient reversion of simian sarcoma virus-transformation and inhibition
of growth factor-induced mitogenesis by suramin. Proc. Natl. Acad. Sci.
USA. 83:6440-6444.

Betsholtz, C., Westermark, 8., Ek, Bo. and Heldin, C.-H. (1984)
Coexpression of a PDGF-like grow factor and PDGF receptors in a human
osteosarcoma cell line: Implications for autocrine receptor activation.
Cell, 39:447-454.

Bianchi, A.B., Aldaz, C.M. and Conti, C.J. (1990) Nonrandom duplication of
the chromosome bearing a mutated Ha-ras-I allele in mouse skin tumors.
PY‘OC. ht]. Md. SCI. “SA, 87:6902’6906.

Biedermann, K.A. and Landolph, J.R. (1988) Induction of anchorage
independence in human diploid foreskin fibroblasts by carcinogenic metal
salts. Cancer Res., 47:3815-3823, 1988.

Birchmeier, C., Broek, M. and Wigler, M. (1985) RAS Proteins Can Induce
Meiosis in Xenopus Oocytes. Cell, 43:615-621.

Bishop, J.M. (1983) Cellular* oncogenes and retroviruses. Ann. Rev.
Biochel. , 52:301-354, 1983.

Bishop, M.J. (1987) The molecular genetics of cancer. Science, 235:305-
311, 1987.

Bos, J.L., Fearon, E.R., Hamilton, S.R., Verlaan-de Vries, M., van Boom,
J.H., van der Eb, A.J. and Vogelstein, B. (1987) Prevalence of ras gene
mutations in human colorectal cancers. Nature, 327:293-297.

Boutwell, R.K. (1974) The function and mechanism of promoters of
carcinogenesis. CRC Crit. Rev. Tbxicol., 2:419-447.

Bovary, T.H. (1929) The Origin of Malignant Tumors, Williams and Wilkins,
Baltimore, 1929.

Boyland, E. (1980) The History and future of chemical carcinogenesis, Dr.
M. “1].. 36:5’10.

57

Bremner, R. and Balmain. (1990) A. Genetic changes in skin tumor
progression: Correlation between presence of a mutant ras gene and loss of
heterozygosity on mouse chromosome 7. Cell, 61:407-417.

Broek, D., Toda, T., Michaeli, T., Levin, L., Birchmeier, C., Zoller, M.,
Powers, S. and Wigler, M. (1987) The S. cerevisiae CDCZS gene product
regulates the Ras/adenylate cyclase pathway. Cell 48:789-799.

Bronner, C.E., Baker, S.M., Morrison, P.T., Warren, Gwynedd, W., Smith,
L.G., Lescoe, M.K., Kane, M., Earabino, C., Lipford, J., Lindblom, A.,
Tannergard, P., Bollag, R.J., Godwin, A.R., Ward, D.C., Nordenskjold, M.,
Fishel, R., Kolodner, R., and Liskay, R.M. (1994) Mutation in the DNA
mismatch repair gene homologue hMLHI is associated with hereditary non-
polyposis colon cancer. Nature, 368:258-261.

Brooks, P. and Lawley, P.D. (1964) Evidence for the binding of polynuclear
aromatic hydrocarbons to the nucleic acids of mouse skin; relation between
carcinogenic power of hydrocarbons and their binding to deoxyribonucleic
acid, Nature, 202, 781-784.

Brown, K., Quintanilla, M., Ramsden, M., Kerr, 1.8., Young, S. and
Balmain, A. (1986) v-ras genes from Harvey'and BALB murine sarcoma viruses
can act as initiators of two-stage mouse skin carcinogenesis. Cell,
46:447-456.

Buday, L. and Downward, J. (1993) Epidermal growth factor regulates p21"‘
through the formation of a complex of receptor, Grb2 adapter protein, and
Sos nucleotide exchange factor. Cell, 73:611-620.

Burns, F.J., Vanderlaan, M., Snyder E. and Albert, R.E. (1978) Induction
and progression kinetics of mouse skin papillomas. In: Nechanisns of “nor
"caution and Cocarcinogenesis, Slaga, T.J., Sivak, A. and Boutwell, R.K.
(Eds.), Raven Press, New York, 91-96.

Bywater, M., Rorsman, F., Boncam-Rudloff E., Mark, G., Hammacher, A.,
Heldin, C.H., Westermark, B. and Betsholtz, C. (1988) Expression of
recombinant platelet-derived growth factor A- and 8- chain homodimers in
Rat-1 cells and human fibroblasts reveals differences in protein
processing and autocrine effects. Mol. Cell. Biol., 8:2753-2762.

Camonis, H.J., Calekine, M., Gondre, 8., Garreau, H., Boy-Marcotte, E.,
Jacquet, M. (1986) Characterization, cloning and sequence analysis of the
C0025 gene which controls the cyclic AMP level of saccharomyces
cerevisiae. EMBO J. 5:375-380.

Cawthon, R.M., Weiss, R., Xu, G., Viscochil, D., Culver, M, Stevens, J.,
Robertson, M., Dunn, 0., Gesteland, R., O’Connell, P. and White, R.A.
(1990) major segment of the neurofibromatosis type I gene: cDNA sequence,
genomic structure, and point mutations. Cell, 62:193-201.

Chang, S.E. (1986) In vitro transformation of human epithelial cells.
Biochel. Biophys. Acta., 823:161-194.

Chiu, I.-M. Reddy, E.P., Givol, D., Robbings, K.C., Tronick, S.R. and
Aaronson, S.A. (1984) Nucleotide~sequenceianalysis identifies the human c-

58

sis proto-oncogene as a structural gene of platelet-derived growth factor.
Cell, 37:123-129.

Chung, C.K. and Antoniades, H.N. (1992) Expression of c-sis/platelet-
derived growth factor B, insulin-like growth factor I, and transforming
growth a messenger RNAs and their respective receptor messenger RNAs in
primary human gastric carcinomas: In vitro studies with in situ
hybridization and immunocytochemistry. Cancer Res., 52:3453-3459.

Cleaver, J.E. (1969) Xeroderma pigmentosum. a human disease in which an
initial stage of DNA repair is defective, Proc. Natl. Acad. Sci. USA. 63,
428-435.

Cook J.W., Hewett C.L., and Heiger, I. (1933) The isolation of a cancer-
producing hydrocarbon from coal tar. Parts I,II and 111, J. Chen. Soc.,
395-405.

Cooper, G.M. (1992) Elelents of Hunan Cancer. Jones and Bartlett
publishers, Boston.

Copeland, N.G. and Cooper, G.F. (1979) Transfection by exogenous and
endogenous murine retrovirus DNAs. Cell, 16:347-356.

Coughlin, S.R., Escobedo, J.A. and Williams, L.T. (1989) Role of
phosphatidylinositol kinase in PDGF receptor signal transduction. Science,
243:1191-1194.

Dahr, R., Ellis, R.W., Shih, T.Y., Oroszian, S., Shapiro, 8., Maizel. J.,
Lowy, D. and Scolnick, E. (1982) Nucleotide sequence of the p21
transforming protein of Hervey murine sarcoma virus. Science, 217:934-937.

Daston, M.M., Scrable, H., Nordlund, M., Strubaum,.A.K., Nissen, L.M., and
Ratner, N. (1992) The product of the neurofibromatosis type 1 gene is
expressed at highest abundance in nerurons, schwann cells, and
oligodendrocytes. Neuron 8:415-28.

DeClue, J.E., Papageorge, A.G., Fletcher, J.A., Diehl, S.R., Ratner, N.,
Vass, W.C. and Lowy, D.R. (1992) Abnormal regulation of mammalian p21"‘
contributes to malignant tumor growth in van Recklinghausen (type I)
neurofibromatosis. Cell, 69:265-273.

Demars, R. and Jackson, J.L. (1977) Mutagenicity detection with human
cells. J. Environ. Pathol. Toxicol. 1:55-57.

Der, C.J., Krontiris, T.G. and Cooper, G.M. (1982) Transforming genes of
human bladder and lung carcinoma cell lines are homologous to the ras
genes of Harvey and Kirsten sarcoma viruses. Proc. Natl. Acad. Sci. USA,
79:3637-3640.

Der, C.J., Weissman, M.J., and MacDonold, M.J. (1988) Altered guanine
nucleotide binding and H-ras transforming and differentiating activities.

Oncogene 3:105-112.

59

Oeuel, T.F., Silverman, N.J. and Kawahara, R.S. (1988) Platelet-derived
growth factor: a multifunctional regulator of normal and abnormal cell
growth. BioFactors 1:213-217.

Devare, S.G., Reddy, E.P., Law, J.D., Robbins, K.C. and Aaronson, S.A.
(1983) Nucleotide wequence of the simian sarcoma virus genome:
Demonstration that its acquired cellular sequences encode the transforming
gene product p28‘“; Proc. Natl. Acad. Sci. USA 80:731-735.

Doolittle, R.F., Hunkapiller, M.W., Devare, S.G., Robbins, K.C., Aaronson,
S.A. and Antoniades, H.N. (1983) Simian sarcoma virus one gene, v-sis, is
derived from the gene (or genes) encoding a platelet-derived growth
factor. Science, 221:275-277.

Downward, J., Riehl, R., Wu, L., Weinberg, R.A. (1990) Identification of
a nucleotide exchange-promoting activity for p21‘“. Proc. Natl. Acad. Sci.
USA, 87:5998-6002.

Earle, W.R. (1943) Production of malignancy in vitro. IV. The mouse
fibroblast cultures and changes seen in the living cells. J Natl. Cancer
Inst. 4:165-212.

Ek, B. and Heldin, C.-H. (1984) Use of an antiserum against
phosphotyrosine for the identification of phosphorylated components in
human fibroblasts stimulated by platelet-derived growth factor. J. Biol.
Chel., 259:11145-11152.

Ellis, C., Moran, M., McCormick, F. and Pawson, T. (1990) Phosphorylation
of GAP and GAP-associated proteins by transforming and mitogenic tyrosine
kinases. Nature, 343:377-381.

Eriksson, A., Siegbahn, A., Westermark, 8., Heldin, C.-H. and Claesson-
Welsh, L. (1992) PDGF a- and B-receptors activate unique and common signal
transduction pathways, Elbo J., 11:543-550.

Erisman, M.D., Scott, J.K., Watt, R.A. and Astrin, S.M. (1988) The c-myc
protein is constitutively expressed at elevated levels in colorectal
carcinoma cell lines. Oncogene, 2:367-378.

Enzinger, F.M. and Weiss, S.W. (1989) Soft Tissue Saree-as, Dysmsyhid, G.,
Elliott, K.C., Gregory. P.L. (Eds), C.V. Mosby Co.

Farr, C.J., Marshall, C.J., Easty, D.J., Wright, N.A., Powell, S.C. and
Christos P. (1988) A study of ras gene mutations in colonic adenomas from
familial polyposis coli patients. Oncogene, 3:673-678.

Fearon, E.R. and Vogelstein, B. (1990) A genetic model for colorectal
tumorigenesis. Cell, 61:759-767.

Fearon, E.R., Hamilton, S.R. and Vogelstein, B. (1987) Clonal analysis of
human colorectal tumors. Science, 238:193-196.

Fen, Z. and Daniel, T. (1991) 5’ untranslated sequences determine
degradative pathway for alternatie PDGF B/c-sis mRNA’s. Oncogene, 6:953-
959.

60

Fishel, R., Lescoe, M.K., Rao, M.R.S., Copeland, N.G., Jenkins, N.A.,
Garber, J., Kane, M., and Kolodner, R. (1993) The human mutator gene
homolog MSH2 and its association with hereditary nonpolyposis colon
cancer. Cell, 75:1027-1038.

Fleming, T.P., Matsui, T., Molloy, C.J., Robbins, K.C. and Aaronson, S.A.
(1989) Autocrine mechanism for v-sis transformation requires cell surface
localization of internally activated growth factor receptors. Proc. Natl.
Acad. Sci. USA, 86:8063-8067.

Fleming, T.P., Matsui, T., Heidaran, M.A., Molloy, C.J., Artrip, J. and
Aaronson, S.A. (1992a) Demonstration of an activated platelet-derived
growth factor autocrine pathway and its role in human tumor cell
proliferation in vitro. Oncogene, 7:1355-1359.

Fleming, T.P., Saxena, A., Clark, W.C., Robertson, J.T., Oldfield, E.H.,
Aaronson, S.A. and Ali, I.U. (1992b) Amplification and/or overexpression
of platelet-derived growth factor receptors and epidermal growth factor
receptor in human glial tumors. Cancer Res., 52:4550-4553.

Forrester, K., Almoguera, C., Han, K., Grizzle, W.E. and Perucho, M.
(1987) Detection of high incidence of K-ras oncogenes during human colon
tumorigenesis. Nature, 327:298-303.

Forsberg, K., Valyi-Nagy. I., Heldin, C.-H. Herlyn, M. and Westermark, B.
(1993) Platelet-derived growth factor (PDGF) in oncogenesis: Development
of a vascular connective tissue stroma in xenotransplanted human melanoma
producing PDGF-BB. Proc. Natl. Acad. Sci. USA, 90:393-397.

Franklin, G.C., Donovan, M., Adam, G.I.R., Holmgren, L., Pfieifer-Ohlsson,
S. and Ohlsson, R. (1991) Expression of the human PDGF-B gene is regulated
by both positively and negatively acting cell type-specific regulatory
elements located in the first intron. Elbo J., 10:1365-1373.

Friedwald, W.F. and Rous, P. (1944) The initiating and promoting elements
in tumor production. J. Exp. Med. 80:101-125.

Fry, D.G., Milam, L.D., Maher, V.M. and McCormick, J.J. (1986)
Transformation of diploid human fibroblasts by DNA transfection with the
v-sis oncogene. J. Cell. Physiol., 128:313-321.

Fry, D.G., Hurlin, P.J., Maher, V.M. and McCormick, J.J. (1988)
Transformation of diploid human fibroblasts by transfection with the v-
sis, PDGFZ/c-sis, or T24 H-ras genes. Mutat. Res. 199:341-351.

Furth, J. and Kahn, M.C. (1937) The transmission of leukemia of mice with
a single cell, AI. J. Cancer, 31:276-282.

Gale, N.W., Kaplan, S., Lowenstein, E.J.,Schlessinger, J. and Bar-Sagi, D.
(1993) Grb2 mediates the EGF-dependent activation of guanine nucleotide
exchange on Ras. Nature, 363:88-92.

Garrett, J.S., Coughlin, S.R., Niman, H.L., Tremble, P.M., Giels, G.M. and
Williams, L.T. (1984) Blockade of autocrine stimulation in simian sarcoma

61

virus-transformed cells reverses down-regulation of platelet-derived
growth factor receptors. Proc. Natl. Acac. Sci. USA, 81:7466-7470.

Gey, 6.0. (1941) Cytological and cultural observations on transplantable
rat sarcomata produced by the inoculation of altered normal cells
maintained in continuous culture. Cancer Res., 1:737.

Glickman, R.M. (1987) Inflammatory Bowel Disease. In: Harrison’s
Principles of Internal Medicine. Braunwald, E., Isselbacher, K.J.,
Petersdorf, R.G. Wilson, J.D., Martin, J.B., and Fauci, A.S. Eds. McGrah-
Hill , Inc., New York, 11th Edition, 1277-1290.

Goelz, S.E. and Vogelstein, B. (1985) Hypomethylation of DNA from benign
and malignant human colon neoplasms. Science, 228:187-190.

Golubic, M., Roudebush, M., Dobrowolski, S., Wolfman, A., Stacey, D.W.
(1992) Catalytic properties, tissue and intracellular distribution of
neurofibromin. Oncogene 7:2151-2159.

Groden, J., Thliveris, A., Samowitz, W., Carlson, M., Gelbert, L.,
Albertsen, H., Joslyn, G., Stevens, J., Spirio, L., Robertson, M.,
Sargeant, L., Krapcho, K., Wolff, E., Burt, R., Hughes, J.P., Warrington,
J., McPherson, J., Wasmuth, J., Le Paslier,D., Abderrahim, H., Cohen, 0.,
Leppert, M., and White, R. (1991) Identification and Characterization of
the Familial Adenomatous Polyposis Coli Gene. Cell, 66:589-600.

Gronwald, R.G.K., Brant, F.J., Haldeman, B.A., Hart, C.E., O’Hara, P.J.,
Hagen, F.S., Ross, R., Bowen-Pope, D.F. and Murray, M.F. (1988) Cloning
and expression of a cDNA coding for the human platelet-derived growth
factor receptor: Evidence for more than one receptor class. Proc. Natl.
Acac. Sci. USA, 85:3435-3439.

Haggitt, R.C. and Reid, B.J. (1986) Hereditary gastrointestinal polyposis
syndromes. “0 Jo surge P.th°]og 10:871-887.

Hall, A. and Self, A.J. (1986) The Effect of Mg”’on the'Guanine Nucleotide
Exchange Rate of p21”‘”. J. Biol. Chen. 261:10963-10965.

Hamilton, S.R. (1993) The Molecular Genetics of Colorectal Neoplasia.
Gastroenterology. 105:3-7.

Hannink, M. and Donoghue, B.J. (1989) Structure and function of platelet-
derived growth factor (PDGF) and related proteins. Biochel. Biophys. Acta.
989:1-10.

Haubruck, H. and McCormick, F. (1991) Ras p21: effects and regulation.
Biochel. Biophys. Acta. 1072:215-229.

Hayflick, L., and Moorhead, P.S. (1961) The serial cultivation of human
diploid cell strains. Exp. Cell Res., 25:585-621.

Hecker, E. (1968) Cocarcinogenic principles from the seed oil of Croton
tiglium and from other Euphorbiaceae. Cancer Res., 28:2338-2349, 1968.

62

Heldin, C.-H. (1992) Structural and functional studies on platelet-derived
growth factor. EMBO J., 11:4251-4259.

Heldin, C.-H. and Westermark, B. (1991) Platelet-derived growth factor and
autocrine mechanisms of oncogenic processes. Oncogenesis, 2:109-124, 1991.

Heldin, C.-H., Westermark, B. and Wasteson, A. (1980) Chemical and
biological properties of a growth factor from human cultured osteosarcoma
cells: resemblance with platelet-derived growth factor. J. Cell Physiol.
105:235-246.

Heldin, C.-H., Westermark, B. Wasteson, A. (1981) Demonstration of an
antibody against platelet-derived growth factor. Expc Cell Res., 136:255-
261.

Hennings, H., Shores, R., Wenk, M.L., Spangler, E.F., Tarone, R. and
Yuspa, S.H. (1983) Malignant conversion of mouse skin tumours is increased
by tumour initiators and unaffected by tumour promoters. Nature, 304:67-
69.

Hermanson, M., Funa, K., Hartman, M., Claesson-Welsh, L., Heldin, C.-H.,
Westermark, B. and Nister, N. (1992) Platelet-derived growth factor and
its receptors in human glioma tissue: Expression of messenger RNA and
protein suggests the presence of autocrine and paracring loops. Cancer
Res., 52:3213-3219.

Huang, J.S., Huang, S.S. and Oeuel, T.F. (1984) Transforming protein of
simian sarcoma virus stumulates autocrine growth of SSV-transformed cells
through PDGF cell-surface receptors. Cell, 39:79-87.

Huang, S.S. and Huang, J.S. (1988) Rapid turnover of platelet-derived
growth factor receptor in sis-transformed cells and reversal by suramin.

Huang, Y.K., Kung, H.F., and Kamata, T. (1990) Purification of a factor
capable of stimulating the guanine nucleotide exchange'of ras proteins and
its effect on ras-related small molecular mass G proteins. Proc. Natl.
“do SC". USA. 87:8008'8012.

Hurlin, P.J., Fry, D.G., Maher, V.M. and McCormick, J.J (1987).
Morphological transformation, focus formation, and anchorage independence
induced in diploid human fibroblasts by expression of a transfected H-ras
oncogene. Cancer Res., 47:5752-5757.

Hurlin, P.J., Maher, V.M., and McCormick, J.J. (1989) Malignant
transformation of human fibroblasts caused by expression of a transfected

Jindal, H.K., Anderson, C.W., Davis, R.G. and Vishwanatha, J.K. (1990)
Suramin affects DNA sythesis in HeLa cells by inhibition of DNA
polymerases. Cancer Res. 50:7754-7757.

Johnson, J.A. (1982) Biochemical aspects of cancer-associated
genodermatoses. In: Cancer Associated Genoderaatoses. Lynch, H.T. and
Fusaro, R.M. (Eds.) Van Norstrand Reinhold Co., New York, 440-540.

63

Johnsson, A., Betsholtz, C., Heldin, C.-H. and Westermark, B. (1985)
Antibodies agaist platelet-derived growth factor inhibit acute
transformation by simian sarcoma virus. Nature 317:438-440.

Jones, S., Vignais, M-L. and Broach, J.R. (1991) The CDC25 protein of
Saccharomyces cerevisiae promotes exchange of guanine nucleotides bound to
Ras. Mol. Cell. Biol., 11:2641-2646.

Kakunaga, T. (1978) Neoplastic transformation of human diploid fibroblast
cells by chemical carcinogens. Proc. Natl. Acac. Sci. USA, 75:1334-1338.

Kaplan, D.R., Morrison, E.K., Wong, G., McCormick, F. and Williams, L.T.
(1990) PDGF-B receptor stimulates tyrosine phosphorylation of GAP and
association of GAP with signaling complex. Cell, 61:125-133.

Kavanaugh, W.M., Klippel, A., Escobedo, J.A. and Williams, L.T. (1992)
Modification of the 85-kilodalton subunit of phosphatidylinositol-3 kinase
in platelet-derived growth factor-stimulated cells. Mol. Cell. Biol.,
12:3415-3424.

Keating, M.T. and Williams, L.T. (1988) Autocrine stimulation of
intracellular PDGF receptors in v-sis-transformed cells. Science, 239:914-
916.

Krontiris, T.G. and Cooper, G.M. (1981) Transforming activity of human
tumor DNAs. Proc. Natl. Acad. Sci. USA, 78:1181-1184.

Land, H., Chen, A.C., Morgenstern, J.P., Parada, L.F., and Weinberg, R.A.
(1986) Behavior of myc and ras oncogenes in transformation of rat embryo

Land, H., Parada, L.F. and Weinberg, R.A. (1983) Tumorigenic conversion of
primary embryo fibroblasts requires at least two cooperating oncogenes.
Nature, 304:596-602.

Leach, F.S., Nicolaides, N.C., Papadopoulos, N., Liu, 8., Jen, J.,
Parsons, R., Peltomai, P., Sistonen P., Aaltonen, L.A., NystromrLahti, M.,
Guan, X-Y., Zhang, J., Melzer, P.S., Yu J-W., Kao, F-A., Chen, D.,
Cerosaletti, K.M., Fournier,R.E.K., Todd,S., Lewis, T., Leach, R.J.,
Naylor, S.L., Weissenbach, J., Mecklin, J-P., Jarvinen, H., Petersen,
G.M., Hamilton, S.R., Green, J., Jass, J., Watson, P., Lynch, H.T., Trent,
J.M., de la Chapelle, A., Kinzler, K.W., and Vogelstein, B. (1993)
Mutations of a mutS homolog in hereditary nonpolyposis colorectal cancer.
Cell, 75:1215-1225.

Leppert, M., Dobbs, M., Scambler, P., O’Connell, P., Nakamura, Y.,
Stauffer, 0., Woodward, S., Burt, R., Hughes, J., Gardner, E., Lathrop,
M., Wasmuth, J., Lalouel, J.-M. and White, R. (1987) The gene for familial
polyposis coli maps to the long arm of chromosome 5. Science, 238:1411-

1413.

Levine, A., Gill, L.S., Cohen, J., Hawkins, J.G., Formenti, S.V., Aguilar,
S., Meyer, P.W., Krailow, M., Parker, J. and Rasheed, S. (1986) Suramin
antiviral therapy in the aquired immunodeficiency syndrome. Ann. Intern.
Med., 105:32-36.

64

Lindblom, A., Tannergard, P., Werelius, B., and Nordenskjdld, M. (1993)
Genetic mapping of a second locus predisposing to hereditary nonpolyposis
colon cancer. Nature Genet. 5, 279-282.

Lowy, D.R., Rands, E. and Scolnick, E.M. (1978) Helper-independent
transformation by nintegrated Harvey Sarcoma‘Virus DNA. J. Virol., 26:291-
298.

Lynch, H.T., Smyrk, T.C., Watson, P., Lanspa, S.J., Lynch, J.E., Lynch,
P.M., Cavalieri, R.J., and Boland, R. (1993) Genetics, natural history,
tumor spectrum, and pathology' of hereditary nonpolyposis colorectal
cancer: An updated review. GastroenterolOOY. 104:1535-1549.

Maher, V.M. and McCormick, J.J. (1976) Effect of DNA repair on the
cytotoxicity and mutagenicity of UV irradiation and of chemical
carcinogens in normal and xeroderma pigmentosum cells. In: Biology of
Radiation Carcinogenesis, Yuhas, J.M., Tennant R.W., and Regan J.B. (Eds.)
Raven Press, New York, 129-145, 1976.

Maher, V.M., Miller, E.C., Miller, J.A., and Szybalski, W. (1968)
Mutations and decreases in density of transforming DNA produced by
derivatives of the carcinogens 2-acetyl-aminofluorene and N-methyl-4-
aminobenzene. Mol. Pharaacol., 4: 411-426.

Malik, R., Gope, M., Womer, R.B., Nagashunmugam,‘T., Margolin, J.E., Basu,
Amitiabha, B. and Scher, C.D. (1991) Structure and expression of the B-
platelet-derived growth factor receptor gene in human tumor cell lines.
Cancer Res. 51:5626-5631.

Maxwell, M., Naber, S.P., Wolfe, H.J., Galanopoulos, T., Hedley-Whyte, E.,
Black, P. and Antonaides, H.N. (1990) Coexpression of platelet-derived
growth factor (PDGF) and PDGF-receptor genes by primary human astrocytomas
may contribute to their development and maintainance. J. Clin. Invest.,
86:131-140.

McCormick, J.J. and Maher, V.M. (1991) Malignant transformation of human
fibroblasts in vitro. In: Neoplastic Transforlation in Hulan Cell Culture.
Rhim, J.S. and Dritschilo, A. (Eds.) The Human Press, Totowa, New Jersey,
347-357.

McCormick, JJ. and Maher, V.M. (1989) Malignant transformation of
mammalian cells in culture, including human cells. Environ. Mol.
Mutagenesis 14: 105-113.

McCormick, J.J. and Maher, V.M. (1988) Towards an understanding of the
malignant transformation of diploid human fibroblasts. Mutat. Res.
199:273-291.

McCormick, J.J., Yang, 0., Maher, V.M., Farber, R.A., Neuman, W.,
Peterson, W.D. and Pollack, M.S. (1988) The Hut series of "carcinogen-
transformed" human fibroblast cell lines are derived from the human
fibrosarcoma cell line 8387. Carcinogenesis 9:2073-2079.

65

Meisenhelder, J., Suh, P.-G., Rhee, S.G. and Hunter, T. (1989)
Phospholipase C-y is a substrate for the PDGF and EGF receptor protein
kinases in vivo and in vitro. Cell, 57:1109-1122.

Miller, J.A. and Miller E.C. (1969) The metabolic activation of
carcinogenic aromatic amines and amides. Prog. Exp. Tulor Res., 11:273-
301.

Milo, G.E. and Dipaolo J.A. (1978) Neoplastic transformation of human
diploid cells in vitro after chemical carcinogen treatment. Nature
275:130-132.

Mills, G.B., Zhang, N., May, C., Hill, M. and Chung, A. (1990) Suramin
prevents binding of interleukin 2 to its cell surface receptor: A possible
mechanism for immunosupprossion. Cancer Res., 50:3036-3042.

Minniti, C.P., Maggi, M. and Helman, L.J. (1992) Suramin inhibits the
growth of human rhabdomyosarcoma by interrupting the insulin-like growth
factor II autocrine growth loop. Cancer. Res., 52:1830-1835.

Miyoshi, Y., Ando, H., Nagase, H., Nishisho, I., Horee, A., Miki, Y.,
Mori, T., Utsunomiya, J., Baba, S., Petersen, 6., Hamilton, S., Kinzler,
K.W., Vogelstein, 8., and Nakamura, Y. (1992) Germ-line mutations of the
APC gene in 53 familial adenomatous polyposis patients. Proc. Natl. Acad.
Sci. USA, 89:4452-4456.

Molloy, C.J., Fleming, T.P., Bottaro, O.P., Cuadrado, A. and Aaronson,
S.A. (1992) Platelet-derived growth factor stimulation of GTPase-
activating protein tyrosine phosphorylation in control and c-H-ras-
expressing NIH 3T3 cells correlates with p21”“ activation. Mol. Cell.
Biol., 12:3903-3909.

Mottram, J.C. (1944) A developing factor in experimental blastogenesis. J.
Pathol. Bacteriol., 56:181-187.

Morgan, T.L., Yang, 0., Fry, D.G., Hurlin, P.J., Kohler, S.K., Maher, V.M.
and McCormick, J.J. (1991) Characteristics of an infinite life span
diploid human fibroblast cell strain and a near-diploid strain arising
from a clone of cells expressing a transfected v-myc oncogene. Exp. Cell
Res., 197:125-136.

Mulcahy, L.S., Smith, M.R. and Stacey, D.W. (1985) Requirement for ras
proto-oncogene function during serum-stimulated growth of NIH 3T3 cells.
Nature, 313:241-243.

Namba, M., Nishitani, K., Fukushima, F., Kimoto, T., Utsumoniya, J. and
Hayflick, L. (1981) Neoplastic transformation of human diploid fibroblasts
treated with chemical carcinogens and ”Co gamma rays. Gann Monogr.,
27:221-230.

Namba, M., Nishitani, K., Fukushima, F. and Kimoto, T. (1985) Neoplastic
transformation of human diploid fibroblasts (KMST-6) by treatment with ”Co
gamma rays. Int. J. Cancer, 35:275-280.

66

Namba, M., Nishitani, K., Fukushima, F., Kimoto, T. and Yuasa, Y. (1988)
Multistep process of neoplastic transformation of normal human fibroblasts
by “Co gamma rays and H-ras oncogenes. Mutat. Res., 199:427-435.

Nishisho, I., Nakamura, Y., Miyoshi, Y., Miki, Y., Ando, H., Horii, A.,
Koyama, K., Utsunomiya, J., Baba, S., Hedge, P., Markham, A., Krush, A.J.,
Petersen, G., Hamilton, S.R., Nilbert, M.C., Levy, D.B., Bryan, T.H.,
Preisinger, A.C., Smith, K.J., Su, L-K., Kinzler, K.W., and Vogelstein, B.
(1991) Mutations of Chromosome 5q21 Genes in FAP and Colorectal Cancer
Patients. Science, 253: 665-669.

Nister, M., Hammacher, A., Mellstrom, K., Siegbahn, A., Ronnstrand, L.,
Westermark, B. and Heldin, C.-H. (1988) A glioma-derived PDGF A chain
homodimer has different functional activities than a PDGF A8 heterodimer
purified from human platelets. Cell, 52:791-799.

O’Brien, W., Stenman, G. and Sager, R. (1986) Suppression of tumor growth
by senescence in virally transformed human fibroblasts. Proc. Natl. Acac.
Sci. USA, 83:8659-8663.

Okamoto, H., Sasaki, M., Sugio, K., Sato, C., Iwama, T., Ikeuchi, T.,
Tonomura, A., Sasazuki, T. and Miyaki, M. (1988) Loss of constitutional
heterozygosity in colon carcinoma from patients with familial polyposis
coli. Nature, 331:273-277.

Pantazis, P., Pelicci, R.G., Dalla-Fauera, R., and Antoniades, H.N. (1985)
Synthesis and secretion of proteins resembling platelet-derived growth
factor by human glioblastoma and fibrosarcoma cells in culture. Proc.
Mt]. Md. SCI. USA, 82: 2404-2408-

PapadOpoulos, N., Nicolaides, N.C., Wei, Y-F., Ruben, S.M., Carter, K.C.,
Rosen, C.A., Haseltine, W.A., Fleischmann, R.D., Fraser, C.M., Adams,
M.D., Venter, J.C., Hamilton, S.R., Petersen G.M., Watson, P., Lynch,
H.T., Peltomaki, P., Mecklin, J-P., de la Chapelle, A., Kinzler, K.W., and
Vogelstein, B. (1994) Mutation of a mutt homolog in hereditary colon
cancer. Science, 263:1625-1628.

Parada, L.P., Tabin, C.J., Shih, C. and Weinberg, R.A. (1982) Human EJ
bladder carcinoma oncogene is homologue of Harvey sarcoma virus ras gene.
Nature, 297:474-478.

Parshad, R., Sanford, K.K., Jones G.M., Price P.M., Taylor, N.G. (1977)
Oxygen and light effects on chromosomal aberrations in mouse cells in
vitro. Exp. Cell Res., 104:199-205.

Pech, M., Gazit, A., Arnstein, P. and Aaronson, S.A. (1989) Generation of
fibrosarcomas in vivo by a retrovirus that expresses the normal 8 chain of
platelet-derived growth factor and mimics the alternative splice pattern
of the v-sis oncogene. Proc. Natl. Acad. Sci. USA, 86:2693-2697.

Peltomaki, P., Aaltonene, L.A., Sistonen, P., Pylkkanen, Mecklin, J-P.,
Jarvinen, H., Green J.S., Jass, J.R., Weber, J.L., Leach, F.S., Petersen,
G.M., Hamilton, S.R., de la Chapelle, A., and Vogelstein, B. (1993)
Genetic mapping of a locus predisposing to human colorectal cancer.
Science, 260:810-812.

67

Peres, R., Betsholtz, C., Westermark, C. and Heldin, C.-H. (1987) Frequent
expression of growth factors for mesenchymal cells in human mammary
carcinoma cell lines. Cancer Res., 47:3425-3429.

Pott, P. (1987) Chirurgical observations relative to cancer of the
scrotum, 1775. Republished in: Classics in Oncology. Holleb, A.I. and
Randers-Pehrson, M.B. (Eds.) American Cancer Society Inc., New York.

Powell, S.M., Zilz, N., Beazer-Barclay, Yasmin, Bryan, T.M., Hamilton,
S.R., Thibodeau, S.H., Vogelstein, 8., and Kinzler, K.W. (1992) APC
mutations occur early during colorectal tumorigenesis. Nature, 359:235-
237.

Ouintanilla, M., Brown, K., Ramsden, M. and Balmain M. (1986) Carcinogen-
specific mutation and amplification of Ha-ras during mouse skin
carcinogenesis. Nature, 322:78-80.

Rao, C.D., Pech, M., Robbins, K.C. and Aaronson, S.A. (1988) The 5’-
untranslated sequence' of the c-sis/platelet-derived growth factor' 2
transcript is a potent translational inhibitor. Mol. Cell Biol., 8:284-
292.

Ratner, N., Lieberman, M.A., Riccardi, V.M., and Hong, D. (1990) Mitogen
accumulation in van Recklinghausen neurofibromatosis. Ann. Neurol.,
27:496-501.

Redmond, 0.E. (1970) Tobacco and cancer: the first clinical report, 1761.
'0 [M]. J. M., 282:18‘23.

Reznikoff, C.A., Loretz, L.J., Christian, B.J., Wu, S. and Meisner, L.F.
(1988) Neoplastic transformation of SV40-immortalized human urinary tract
epithelial cells by in vitro exposure to 3-methylcholanthrene.
Carcinogenesis, 8:1427-1436.

Rhim, J.S., Jay, G., Arnstein, P., Price, E.M., Sanford, K.K. and
Aaronson, S.A. (1985) Neoplastic transformation of human epidermal
keratinocytes by A012-SV4O and Kirsten sarcoma viruses. Science, 227:1250-
1252.

Riccardi, V. and Eichner, J.(1986) Neurofibrolatosis: Phenotype, Natural
History, and Pathogenesis, Riccardi, V. and Eichner, J. (Eds) Johns
Hopkins University Press, Baltimore, Md.

Richter, M.R. and Graves, D.T. (1988) DNA synthesis in U-2 OS human
osteosarcoma cells is independent of PDGF binding to functional cell
surface receptors. J. Cell. Physiol., 135:474-480.

Ross, R., Raines, E.W. and Bowen-Pope, D.F. (1986) The biology of platelet
derived growth factor. Cell, 46:155-169.

Rozengurt, E., Sinnett-Smith, J. and Taylor-Papadimitriou, J. (1985)
Production of PDGF-like growth factors by breast cancer cell lines. Int.
J. Cancer, 36:247-252.

68

Rous, P. and Kidd, J.G. (1941) Conditional neoplasms and subthreshold
neoplastic states. A study of tar tumors in rabbits. J Exp. Med., 73:365-
390.

Rubinfeld, 8., Souza, 8., Albert, 1., Miiller, O, Chamberain, S.H.,
Masiarz, F.R., Munemitsu, S., and Polakis, P. (1993) Association of the
APC Gene Product with B-Catenin. Science, 262:1731-1734.

Ruggeri, 8., Caamano, J., Goodrow, T., DiRado, M., Bianchi, A., Trono, 0.,
Conti, C.J. and Klein-Szanto, A.J.P. (1991) Alterations of the p53 tumor
suppressor gene during mouse skin tumor progression. Cancer Res., 51:6615-
6621.

Sager, R., Tanaka, K., Lau, C.C., Ebina, Y. and Anisowicz, A. (1983)
Resistance of human cells to tumorigenesis induced by cloned transforming
genes. Proc. Natl. Acad. Sci. USA, 80:7601-7605.

Sanford, K.K. and Evans, V.J. (1982) A quest for the mechanisms of
"spontaneous" malignant transformation in culture with associated advances
in culture technology. J. Natl. Cancer Inst., 68: 895-913.

Santos, E. and Nebreda, A.R. (1989) Structural and functional properties
of ras proteins. FASEB J., 3:3151-3163.

Sato, Y. and Rifkin, 0.8. (1988) Autocrine acivities of basic fibroblastic
growth factor: regulation of endothelial cell movement, plasminogen
activator synthesis, and DNA synthesis. J. Cell Biol. 107: 1199-12058.

Satoh, T., Endo, M., Nakafuku, M., Nakamura, S. and Yoshito Kaziro.
(1990a) Platelet-derived growth factor stimulates formation of active
p21“”-GTP complex in Swiss mouse 3T3 cells. Proc. Natl. Acad. Sci. USA,
87:5993-5997.

Satoh, T., Endo, M., Nakafuku, M., Akiyama, T., Yamamoto, T. and Kaziro.
Y. (1990b) Accumulation of p21"‘-GTP in response to stimulation with
apidermal growth factor and oncogene) products with tyrosine kinase
activity. Proc. Natl. Acad. Sci. USA, 87:7926-7929.

Schaeffer, W.I. (1990) Terminology associated with cell, tissue and organ
culture, molecular biology and molecular genetics. In Vitro Cell. Dev.
Biol. 26: 97-101.

Seppa, H.E.J., Grotendorst, G.R., Seppa, S.I., Shiffman, E. and Martin,
G.R. (1982) Platelet-derived growth factor is chennnotactic for fibroblasts.
J. Cell Biol., 92:584-588.

Sheela, S., Riccardi, V.M., and Ratner, N. (1990) Angiogenic and Invasive
Properties of Neurofibroma Schwann Cells, J. Cell Biol., 645-653.

Shih, C., Padhy, L.C., Murray, M. and Weinberg, R.A. (1981) Transforming
genes of carcinomas and neuroblastomas introduced into mouse fibroblasts.
Nature, 290:261-2641.

69

Shou, C., Farnsworth, C.L., Neel, 8.6. and Feig, L.A. (1992) Molecular
cloning of cONAs encoding a guanine-nucleotide-releasing factor for Ras
p21. Nature, 351-354.

Shubik, P. The growth potentialities of induced skin tumors in mice: The
effects of different methods of chemical carcinogenesis. Cancer Res.,
10:713-717, 1950.

Silinskas, L.C., Kately, S.A., Tower, J.E., Maher, V.M. and McCormick,
J.J. (1981) Induction of anchorage independent growth in human fibroblasts
by propane sultone. Cancer Res., 41:1620-1627.

Silver, B.J. (1992) Platelet derived growth factor in human malignancy.
BioFactors, 3:217-227.

Simon, M.A., Bowtell, D.D.L., Dodson, S.G., Laverty, T.R. and Rubin, G.M.
(1991) Rasl and a putative guanine nucleotide exchange factor perform
crucial steps in signaling by the sevenless protein tyrosine kinase. Cell,
67:701-716. 1991.

Simon, M.I., Strathmann, M.P. and Guatam, N. (1991) Diversity of G
proteins in signal transduction. Science, 252:802-808.

Slaga, T.J. (1993) Overview'of tumor promotion in animals. Environ. Health
Perspec., 50:3-14.

Slamon, D.J., deKernion, J.B., Verna, I.M. and Cline, M.J. (1984)
Expression of cellular oncogenes in human malignancies. Science, 224:256-
262.

Smith, K.J., Johnson, K.A., Bryan, T.M., Hill, 0.E., Markowitz, S.,
Willson, J.K.V., Paraskeva, C., Petersen, G.M., Hamilton, S.R.,
Vogelstein, 8., and Kinzler, K.W. (1993) The APC gene product in normal
and tumor cells. Proc. Natl. Acad. Sci. USA, 90:2846-2850.

Spector, D.H., Varmus, H.E. and Bishop, J.M. (1978) Proc. Natl. Acad. Sci.
USA, 4102-4106.

Sporn, M.B. and Todaro, G.J. (1980) Autocrine secretion and malignant
transformation of cells. N. Engl. J. Med. 303:878-880.

Stampfer, M.R. and Bartley, J.C. (1985) Induction of transformation and
continuous cell lines from normal human mammary epithelial cells after
exposure to benzo[a]pyrene. Proc. Natl. Acac. Sci. USA, 82:2394-2398.

Stehelin, 0., Varmus, H.E., Bishop, J.M. and Vogt, P.K. (1976) DNA related
to the transforming gene(s) of avian sarcoma viruses is present in normal
avian DNA. Nature, 260:170-173.

Sturzl, M., Roth, H.K., Brockmeyer, N.H., Zietz, C., Speiser, B. and
Hofscneider, P.H. (1992) Expression of platelet-derived growth factor and
its receptor in AIDS-related Kaposi sarcoma in viva suggests paracrine and
autocrine mechanisms of tumor maintenance. Proc. Natl. Acac. Sci. USA,
89:7046-7050.

70

Su, L-K., Johnson, K.A., Smith, K.J., Hill, 0.E., Vogelstein, B., and
Kinzler, K.W. (1993a) Association between Wild Type and Mutant APC Gene
Products. Cancer Res., 53:2728-2731.

Su, L-K., Vogelstein, 8., and Kinzler, K.W. (1993b) Association of the.APC
Tumor Suppressor Protein with Catenins. Science, 262:1734-1737, 1993b.

Sugerbaker, J.P., Gunderson, L.L., and Wittes, R.E. (1985) In: Cancer:
Principles and Practice of Oncology. Devita, V.T., Hellman, S., and
Rosenberg, Eds. Philadellphia, J.P. Lippincot, 800-803.

Swan, D.C., McBridge, O.W., Robbins, K.C., Keithly, B.A., Reddy, E.P. and
Aaronson, S.A. (1982) Chromosomal mapping of the simian sarcoma virus
oncogene analogue in human cells. Proc. Natl. Acac. Sci. USA, 79:4691-
4695.

Trahey, M. and McCormick, F. A. (1987) cytoplasmic protein stimulates
normal N-ras p21 GTPase, but does not affect oncogenic mutants. Science,
238:542-545.

Trahey, M., Wang, G., Halenbeck, R., Rubinfeld, 8., Martin, G.A., Ladner,
M., Long, C.M., Crosier, H.J., Watt, K., Koths, K. and McCormick, F.
(1988) Molecular cloning of two types of GAP complementary DNA from human
placenta. Science, 242:1697-1700.

Vasen, HFA, Mecklin, J.P., Meera, K.P., and Lynch, H.T. (1991) Hereditary
non-polyposis colorectal cancer. Lancet, 338:887.

Vogelstein, B., Fearon, E.R., Hamilton, S.R., Kern, S.E., Preisinger,
A.C., Leppert, M., Nakamura, Y., White, R., Smits, A.M.M. and Bos, J.L.
(1988) Genetic alterations during colorectal-tumor development. N. Engl.
J. Med., 319:525-532.

Wang, Y., Kateley-Kohler, S., Maher, V.M. and McCormick, J.J. (1986) ”C0
radiation-induced transformation to anchorage independence of fibroblasts
from normal persons and patients with an inherited predisposition to
retinoblastoma. Carcinogenesis, 7:1927-1929.

Warthin, A.S. (1913) Heredity with reference to carcinoma. Arch. Intern.
Med., 12:546-555, 1913.

Waterfield, M.D., Scrace, G.T., Whittle, N., Stroobant, P., Johnsson, A.,
Wasteson, A., Westermark, B., Heldin, C.-H. Huang, J.S. and Oeuel, T.F.
(1983) Platelet-derived growth factor is structurally related to the
putative transforming protein p28‘" of simian sarcoma virus. Nature,
304:35-39.

Wazer, 0.E., Chu, 0., Liu, X-L., Gao, Q. Safaii, H., and Band, V. (1994)
Loss of p53 Protein during Radiation Transformation of Primary Human
Mammary Epithelial Cells. Mol. Cell Biol., 14:2468-2478.

Westermark, 8., Johnsson, A., Paulsson, Y., Betsholtz, C., Heldin, C.-H.,
Herlyn, M., Rodeck, U. and Koprowski, H. (1986) Human melanoma cell lines
of primary and metastatic origin express the genes encoding the chains of

71

platelet-derived growth factor (PDGF) and produce a PDGF-like growth
factor. Proc. Natl. Acad. Sci. USA, 83:7197-7200.

Williams, L.T., Tremble, P.M., Lavin, M.F. and Sunday, M.E. (1984)
Platelet-derived growth factor receptors form a high affinity state in
emebrane preparations. J. Biol. Chel., 259:5287-5294.

Wilson, 0.M., Fry, D.G., Maher, V.M. and McCormick, J.J. (1989)
Transformation of diploid human fibroblasts by transfection of N-ras-
oncogene. Carcinogenesis, 10:635-640.

Wilson, 0.M., Yang, 0., Dillberger, J.E., Dietrich, S.E., Maher, V.M. and
McCormick, J.J. (1990) Malignant transformation of human fibroblasts by a
transfected N-ras oncogene. Cancer Res., 50:5587-5593.

Wolfman, A. and Macara, I.G. (1990) A cytosolic protein catalyzes the
release of GDP from p21'“. Science, 248:67-69.

Xu, G., O’Connell, P., Viskochil, 0., Cawthon, R., Robertson, M., Culver,
M., Dunn, 0., Stevens, J., Gesteland, R., White, R. and Weiss, R. (1990a)
The neurofibromatosis type I gene encodes a protein related to GAP. Cell,
62:599-608.

Xu, G.F., Lin, 8., Tanaka, K., Dunn, 0., Wood, 0., Gesteland, R., White,
R., Weiss, R. and Tamanoi, F. (1990b) The catalytic domain of the
neurofibromatosis type I gene product stimulates ras GTPase and
complements IRA mutants of S. cerevisiae. Cell, 63:835-841.

Yang, 0., Louden, C., Reinhold, 0.5., Kohler, S.K., Maher, V.M. and
McCormick, J.J. (1992) Malignant transformation of human fibroblast cell
strain MSU-1.1 by (+/-)-7B,8a-dihydroxy-9a,10o-epoxy-7,8,9,10-
tetrahydrobenzo[a]pyrene. Proc. Natl. Acac. Sci. USA, 89:2237-2241.

Yarden, Y., Escobedo, J.A., Kuang, W.-J., Yang-Feng, T.L., Daniel, T.O.,
Tremble, P.M., Chen, E.Y., Ando, M.E., Harkins, R.N., Fancke, U., Fried,
V.A., Ullrich, A. and Williams, L.T. (1986) Structure of the receptor for
platelet-derived growth factor helps define a family of closely related
growth factor receptors. Nature, 323:226-232.

Zhang, K., Papageorge, A.G. and Lowy, D.R. (1992) Mechanistic aspects of
signaling thrugh Ras in NIH 3T3 cells. Science, 257:671-674.

Zippel, R., Sturani, E., Toschi, L., Naldini, L., Alberghina, L. and
Comoglio, P.M. (1986) In viva phosphorylation and dephosphorylation of the
platelet-derived growth factor receptor studied by immunoblot analysis
with phosphotyrosine antibodies. Biochel. Biophys. Acta, 881:54-61.

CHAPTER II

Evidence that Forlation of Neurofibrosarcoaas in
Individuals with Type 1 Neurofibroaatosis is Associated
with Activation of Growth Factor Pathways

Jeremy W. Wray‘, Veronica M. Maher‘, and J. Justin McCormick1

‘Carcinogenesis Laboratory - Fee Hall
Department of Microbiology and Departement of Biochemistry

Michigan State University, East Lansing, MI 48824-1316

72

One of the features of Type 1 neurofibromatosis, a dominantly
inherited syndrome, is that affected individuals comnonly form numerous
benign neurofibromas, which arise from Schwann cells that surround
peripheral nerves. Rarely, a neurofibroma cell undergoes malignant
transformation giving rise to a neurofibrosarcoma. Although
neurofibrosarcomas are rare even in NF1 patients, they are much nore
cannon in such individuals than in the general population. We used a
reverse transcription-polymerase chain reaction (RT-PCR) assay to examine
whether tumor cell lines derived from NF1-associated neurofibromas and
neurofibrosarcomas expressed TGF-a and PDGF-B mRNA. Five of six
neurofibrosarcoma cells expressed high levels of TGF-a and PDGF-B nRNA
similar to the levels seen in human fibrosarcoma-derived cells.
Neurofibroma cells also express PDGF-B mRNA but at a lower level than that
of neurofibrosarcoma cells. Since normal rat Schwann cells and human
Schwann cells derived from neurofibromas do not express receptors for
serum growth factors such as PDGF , we tested whether the neurofibrosarcoma
cells in culture proliferate in response to PDGF, TGF-a and bFGF. Our
results indicated that two of the three neurofibrosarcoma cell lines
tested respond to one or more of these growth factors. The third, which
expresses high levels of PDGF and TGF-a, grows extremely rapidly in serum-
free medium and is not inhibited by suramin. We also demonstrate that one
of two neurofibrosarcoma cell lines capable of growth in medium lacking

73

74
exogenous growth factors is completely inhibited from growth in such
conditions in the presence of 50 ug/ml of suramin. Our results support a
role for activation of growth factor pathways including elevated
expression of growth factors and growth factor receptors in the formation

of neurofibrosarcomas that occurs in individuals with NF1.

INTRINHNITION

Type I neurofibromatosis (NF1) is an autosomal dominantly inherited
disease that affects approximately one in 3000 individuals and is
characterized by multiple neurofibromas (benign tumors), cafe’ au lait
spots, lisch nodules, bony dysplasia, learning disabilities, and a
significant predisposition to the development of malignant tumors (1).
Neurofibromas arise from Schwann cells, which form the sheaths around
peripheral nerves (2). The presence>of numerous neurofibromas in affected
individuals suggests that very few changes, or possibly only a single
change, in addition to the constitutional NF1 defect, are required to
cause a Schwann cell to proliferate and form a neurofibroma. The change
(or changes) that causes the development of neurofibromas in affected
individuals is unknown.

Neurofibrosarcomas, also known as malignant schwannomas or malignant
peripheral nerve sheath tumors, occur as a consequence of malignant
transformation of a neurofibroma cell and are the most cannon of the
malignancies that occur at an elevated frequency in NF1 patients (2).
Although only a minority of NF1 patients develop neurofibrosarcomas, the
frequency of these malignancies is significantly higher in affected
individuals than in the general population. In fact, more than half of
all neurofibrosarcomas that are diagnosed occur in NF1 patients, even

though such individuals compose only a small segment of the population

75

76
(3). Nevertheless, the chance of a neurofibroma giving rise to a
neurofibrosarcoma is relatively small, suggesting that multiple changes
are required for this to occur.

The gene that is constitutively defective in individuals with NF1
has been cloned and sequenced (4), and the gene product has been termed
neurofibromin. Analysis of the NF1 coding sequence revealed that one
segment of the gene encodes a domain homologous to the catalytic domain of
the GTPase-activating protein, p120-GAP (5). GTPase-activating proteins
enhance the intrinsic GTPaseiactivity of ras proteins, and therefore favor
hydrolysis of active GTP-ras to inactive~GDP-ras (6). In normal quiescent
cells, the majority of ras is in the inactive GDP-bound form. However,
ras rapidly becomes activated in response to signals initiated by
mitogenic growth factors such as PDGF-B (7). The proteins involved in
mediating this growth factor-initiated signal are guanine nucleotide
releasing factors (8, 9), which stimulate the release of GDP from GDP-ras.
Since the majority of free guanine nucleotide in the cytosol is GTP, ras
not bound by a guanine nucleotide quickly becomes activated by binding to
GTP. The finding that the NF1 gene, one allele»of which constitutively is
defective in NF1 patients, encodes a protein with a domain homologous to
p120-GAP, suggested that defects in an allele of this gene predisposes
affected individuals to farming neurofibromas because of a deficiency in
inactivation of ras. If so, NF1 would represent a tumor suppressor gene
and the cells in neurofibrosarcomas can be expected to have defects in
both alleles of the NF1 gene. Evidence that this is the case, and that
neurofibromin functions as a GTPase-activating protein, comes from study
by DeClue et al. (10). They demonstrated that human neurofibrosarcoma

cell lines from NF1 patients have normal ras proteins, but are virtually

77
devoid of neurofibromin and have high levels of ras-GTP, as expected if
the cells lack normal regulation of ras protein activity.

As noted above, growth factors stimulate the release of GDP from
ras, resulting in binding of ras to GTP and therefore to ras activation.
If a cell carrying a mutation in one or both alleles of the NF1 gene were
to upregulate expression of a growth factor gene, such as the gene
encoding PDGF, this would enhance the effect of their decreased GTPase
activity by increasing the level of ras in the activated form, i.e., GTP-
ras. It has been shown that extracts from neurofibromas from NF1
individuals contain mitogenic activity (11, 12) and that cells from
neurofibromas produce angiogenic factors (13). Several growth factors,
including PDGF, could account for these activities. Certain growth
factors can stimulate the growth of rat Schwann cells (14) and human
neurofibroma-derived schwann cells (12) in standard cell culture.
Furthermre, addition of cyclic AMP or exogenous stimulation of the
production of this molecule has been shown to cause rat Schwann cells to
respond to additional growth factors (15, 16). The question of the status
of growth factor expression by cells from neurofibrosarcomas and its
possible effect on the growth of such cells has not been examined. To
determine how neurofibrosarcomas differ from neurofibromas, we have
compared cells which we derived from such tumors, taken from individuals
with NF1, for their expression of growth factors, for the effect of growth
factors on the proliferation of such cells, and for evidence of dependence

of the cells on such growth factors.

78

Materials and lethods

Culturing tumor-derived cells

The tumor material was fragmented into smaller pieces and mechanically
agitated using a sterile syringe and cannula to free individual cells from
the tumor tissue. Some of the resulting cells and small tumor fragments
were placed in 75 cu? tissue culture flasks (Corning Glass Works, Corning
N.Y.) and cultured in Eagle’s minimum essential medium supplemented with
0.2 mM aspartic acid, 1.0 mM sodium pyruvate, 0.2 mM serine, penicillin
(100 U/ml), streptomycin (100 ug/ml), hydrocortisone (10 pg/ml), and 5%
and/or 10% fetal bovine serum (culture medium). Cultures were incubated

at 37'C in a humidified atnosphere of 5% CO2 and air.

Assay for tumorigenicity

Cells (1 x 10‘ to 10 x 10‘) in exponential growth were injected
subcutaneously in the scapular region of six week old BALB/c athymic mice.
Alternatively, small pieces of tumor tissue‘were implanted subcutaneously
into these animals. The mice were examined weekly for tumors. Tumors
that resulted were excised, fixed in formalin, embedded in paraffin,
sectioned, and stained with hematoxylin and eosin prior to histological

evaluation.

79

RNA isolation

Cells were grown in 150 mm-diameter dishes in culture medium containing
supplemented calf serum rather than fetal bovine serum. Total RNA was
isolated from cells just prior to their reaching confluence, using the
procedure of Chomczynski and Sacchi (17) essentially as described. The
RNA was resuspended in diethylpyrocarbonate-treated H,0, and stored at
-80°C. The concentration of RNA was determined from the optical density
at 260nm, and the integrity of the RNA was qualitatively determined from

the appearance of ribosomal RNA bands on 1% agarose/formaldehyde gels.

Determination of the level of expression of mRNA

Synthesis of cONA using Moloney murine leukemia virus reverse
transcriptase (New England Biolabs) and amplification of specific cONAs
using Taq polymerase (Perkin-Elmer Cetus, Norwalk, CT), RT-PCR, was
carried out using the procedure described by Yang et al. (18) with the
following modifications. One microgram of total RNA was added to a series
of 500 pl Eppendorf tubes, and an equal volume of triple distilled water
was added to an additional tube to be used as a negative control. Random
hexamer primers (0.1 mg/ml, Pharmacia/LBK, Piscataway, NJ) were
substituted for oligo (dT),,-,,, and 20 units/ml RT were used instead of 2.5
units/ml. Five pl of reverse transcriptase cocktail were added to each
tube, and the samples were incubated at 37'C for 1 hour. A Taq polymerase
cocktail was made containing 2.22 nM MgCl,, 66.67 m KCl, 16.67 M Tris-HCL
(pH 8.55), 0.44 nM deoxynucleotide triphosphate, 1.11 ng/pl of each
primer, and 0.055 units/pl of Taq polymerase and the specified sets of
primers. Once the RT assay was completed, 45 pl of the Taq cocktail was
added to each tube followed by 100 pl of paraffin oil. The tubes were
then placed in the DNA thermal cycler (Perkin-Elmer Cetus), using an

80
initial 5 min denaturation step at 94°C, followed by up to 30 cycles of
denaturation for 1 min at 94°C, annealing at 50°C for 1 min, and
polymerization at 72°C for 2 min. After 30 cycles, a 7 min extension at
72°C was performed. The primers used for PDGF-B were those of Wang et al.
(19); those for TGF-a‘were chosen by us using the published sequence (20).
They are located in separate exons to eliminate amplification of the
genomic DNA of that gene. The PCR primers were synthesized in the
Macromolecular Structure Facility at Michigan State University. The GAPDH
primers were used as an internal control to insure that amplification had

occurred equally in each reaction tube as described (21).

Assay for growth factor independence
Cells growing in culture medium were trypsinized, counted, and plated in
culture medium containing only 2% fetal bovine serum into 96-well
microtiter plates (Corning Glass Works, Corning, NY) at a density of 500
to 3000 cells/microwell (depending upon plating efficiency and rate of
growth of individual cell lines). After 24 h of incubation, the medium
was replaced with McM medium (18) modified to contain only 0.1 mM calcium,
rather than the usual 1.0 nM calcium, and containing the serum replacement
supplements specified by Ryan et al. (22) but lacking any exogenous
protein growth factors (designated SR,). In some instances suramin was
added. It was obtained as a sodium salt in lyophized form (FBA, Germany,
distributed by Miles Inc., West Haven, CT) and used at the designated
concentration. The cells were fed with the appropriate medium every two
to three days.

To determine the amount of cell proliferation that occurred in each
microwell, we used a soluble tetrazolium/formazan solution to determine

the relative number of cells in each microwell from the change in

81
absorbance, essentially as described (23, 24). Briefly, a 1 mg/ml
solution of 2-3-bis(2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenyamino)car-
bonyl]-2H-tetrazolium hydroxide (XTT) was made in McM medium containing
only 0.1 mM calcium and with 10 mM Hepes (pH 7.2) to maintain the pH, and
lacking phenol red. .A 1.53 mg/ml solution of phenazine methosulfate (PMS)
was made in phosphate buffered saline and was added to the XTT solution to
obtain a 0.05 mM solution of PMS in 1 mg/ml XTT (150 pl of 1.53 mg/ml PMS
per 6 ml of 1 mg/ml XTT). The medium was removed from the microwells and
replaced with 200 pl of this modified McM medium and 50 pl of the PMS/XTT
solution. The color reaction was allowed to proceed for 4 h at 37C° in a
humidified atmosphere of 5% CO2 and air. After the 4 h incubation, the
microwell plate was placed on a shaker for 30 min and then the absorbance
of each well was determined with a Zmax Precision Microplate Reader

(Molecular Devices Corp., Menlo Park CA).

Assay of cellular response to specific growth factors

Cells growing in culture medium were trypsinized and plated at 5 X 10‘ to
75 X 10‘ cells per 60 mm-diameter culture dish in 5 ml of culture medium
with 5% fetal bovine serum. After 24 h, the medium was removed and
replaced with 4 ml of McM medium containing only 0.1 mM calcium and SR,.
After another 24 h, the medium was changed to the specified medium The
medium was changed on day 3 and cell growth was determined on day 6 by

counting cells in triplicate 60 mm culture dishes.

82

Results

Establishing cell lines from tumors from NF1 patients
To investigate the possible role of protein growth factors in the
neoplastic transformation of human Schwann cells into neurofibromas and
neurofibrosarcomas, it was important to derive cell lines from fresh
tumors at the time of surgical resection in order to have relatively early
passage cells for comparative studies. A total of twelve neurofibromas
were obtained. Eleven were surgically removed from an NF1 patient with
multiple benign tunrors and provided to us by Dr. William Mercer of
Michigan Capital Medical Center, Lansing, MI and the twelfth was obtained
from another NF1 patient and provided to us by Dr. Mark Lebwohl, a
dermatological surgeon, from New York, NY. The tumor tissue was
transported in Lebovitz medium containing 100 pg/ml Gentamicin sulfate.
Upon arrival, approximately one third of the tissue was innnnediately fixed
in 10% formalin for histopathological analysis. The rest was minced, and
the cells were separated as described in the materials and methods
section. Half of the remaining material was cryopreserved at -135°C
or lower for future studies; the rest was placed into culture as described
in the methods section. The time between reuoval from the patient and
introduction into cell culture was always less than 20 h. The
neurofibroma cell lines derived from the first NF1 patient were designated
NFb-l through Nfb-ll; that from the second NF 1 patient was designated NFb-
12.

Three neurofibrosarcomas from unrelated NF1 patients were similarly
obtained from surgeons through our colleague Dr. Mark Zalupski of the
Division of Hematology Oncology, Wayne State University School of

Medicine, Detroit, MI. These three tumors were handled in the same way as

83
the neurofibroma tissue. The cell lines derived from these three patients

are referred to as NF1-WSU-4, NF1-WSU-10, and NF1-WSU-17.

Verification of the presence of neoplastic cells in the tumor-derived cell

population

Neurofibromas

The histopathology report from the pathologist at Michigan State
University on the material we fixed in formalin from the original twelve
human neurofibromas confirmed the diagnosis that the tumors were
neurofibromas. We tested tumor fragments that had been placed into cell
culture, as well as early passage cells in culture, for tumorigenicity in
athymic mice using the techniques described in the materials and methods
section. In virtually every case, small nodules developed at the site of
injection or implantation after one month. If not removed, the majority
persisted for several months, but eventually regressed. The nodules that
were removed were fixed and stained. Every single tumor analyzed
demonstrated a histological pattern consistent with the diagnosis of a
human neurofibroma (Figure 1). These results verify that a significant
percentage of the early passage neurofibroma cells that we used in the
studies of the expression of growth factors to be described below were
composed of neoplastic Schwann cells. ‘This finding is significant because
it has reported that neoplastic Schwann cells in neurofibroma-derived cell

populations rapidly become overgrown by fibroblasts.

84

 

Figure 1. Hematoxylin and eosin stained sections of neurofibromas.

(A) Neurofibroma NFb-10 that occurred in patient with Type 1
neurofibromatosis. (B) Neurofibroma that occurred when neurofibroma cell

line NFb-10 was injected subcutaneously into an athymic mouse.

85

Neurofibrosarcomas

Similar studies were carried out with early passage populations of the
cell lines derived from the three neurofibrosarcomas obtained from Wayne
State University. The cells from these three tumors gave rise to
vigorously growing cell lines that have continued to replicate
indefinitely (i.e., >100 population doublings). Therefore, we also
examined the tumorigenicity of later passage cells from each of these
three cell lines. Cells from both the early passage and the late passage
papulations from two of the cell lines, i.e., NF1-WSU-10 and NF1-WSU-17
produced progressively-growing tumors at the site of injection with a
relatively short latency (see Table I). The histopathology report from
the pathologists at Michigan State University on the original tissue that
we fixed in formalin on the day the resected tumors were received agreed
with the pathology reports subsequently received from Wayne State
University indicating that all three of the tumors were
neurofibrosarcomas. The histopathology of the tumors that developed in
the athymic mice after injection of the cell lines referred to above
indicated that they were neurofibrosarcomas. These tumors strikingly
resembled the histopathology of the original human tumors from*which they
were derived (see Figure 2).

For the sake of comparison, three additional neurofibrosarcoma cell
lines were obtained from Dr. Thomas Glover from the Department of
Pediatrics of the University of Michigan. One, designated NF1-90.8, was
from one NF1 patient; the other two, designated NF1 88.3 and 88.14, are

cell lines derived from a primary and a recurrent neurofibrosarcoma from

86

Table I. Tumorigenicity of neurofibrosarcoma-derived cell lines

 

 

Cell Frequency “Latency
line of Tumors

NF1-WSU-4 0/4 NAb
NF1-WSU-10 4/4 5
NF1-WSU-17 4/7 36
NF1-90.8 6/8c 18
NF1-88.3 0/2d NA
NF1-88.14 0/2‘ NA

 

’Number of weeks for tumor to reach 0.6 cm in diameter.

bNA, not applicable.

cOriginal neurofibrosarcoma from which cell line NF1-90.8 is derived was
not available to be compared with that of the tumors that occurred when
NF1-90.8 was injected into athymic mice.

dBecause the cells did not continue to grow in culture, cells were not

available for additional experiments.

87

Figure 3. Hematoxylin and eosin stained sections of neurofibrosarcomas.

(A) Neurofibrosarcoma NF1-WSU-10 that occurred in patient with Type 1
neurofibromatosis. (B) Neurofibrosarcoma that occurred in athymic mouse
when NF1-WSU-10 cells were injected subcutaneously into an athymic mouse.
(C) Neurofibrosarcoma NF1-WSU-17 that occurred in patient with Type 1
Neurofibromatosis. (0) Neurofibrosarcoma that occurred in athymic mouse
when NF1-WSU-17 cells were injected subcutaneously into an athymic mouse.
(E) Neurofibrosarcoma that occurred in athymic mouse when NF1-90.8 cells

were injected subcutaneously into an athymic mouse.

88

r

L
”4..
w

,.

, ’11.]

.4
r.

 

89
a second NF1 patient, respectively. These three cell lines were also
tested for the ability to make tumors in athymic miceu NF1-90.8 produced
progressively growing tumors with histological features consistent with

that of human neurofibrosarcomas (see Figure 2).

RT-PCR assay to detect PDGF-8 and TGF-a mRNA

Human neurofibromas have been shown to produce mitogenic growth factors
including bFGF (11). However, bFGF and several other growth factors are
not mitogenic for rat Schwann cells or Schwann cells derived from human
neurofibromas unless cAMP is artificially elevated. Since
neurofibrosarcomas have elevated GTP-ras, and in several cell types ras
activation stimulates adenylate cyclase which subsequently elevates the
level of cAMP, we reasoned that growth factors such as bFGF, PDGF-B, and
other growth factors could be mitogens in neurofibrosarcoma cells. We
have also previously observed in this laboratory that PDGF-B and TGF-o are
both expressed in human fibrosarcomas (unpublished studies). Because of
these observations, we investigated whether neurofibrosarcoma cells
express PDGF-B and TGF-a at elevated levels.

Normal human cells used as controls included two normal human
foreskin fibroblast lines, an adult snooth muscle cell line, and a
fibroblast line that was derived from normal skin of the NF1 patient that
developed the neurofibrosarcoma that gave rise to cell line NF1-90-8

(Table II). We also included cell lines derived from human malignancies

90

'Provided by Dr. Peter Libby, Brigham and Women's Hospital, Boston, MA
”Provided by Dr. Thomas Glover, University of Michigan, Ann Arbor, MI
cTumors provided by Dr. Mark Zalupski, Wayne State Univ., Detroit, MI
dTumors provided by Dr. William Mercer, Michigan Capital Medical Center,
Lansing, MI
’Tumor provided by Dr. Mark Lebwohl, dermatologic surgeon, New York, NY.
'Cell lines from The American Type Culture Collection, Rockville, MD.
“Justin McCormick, unpublished studies

91

Table II. Human cell lines/strains

 

Cell line or strain

Cell type - origin

 

SL-80, SL-68

HSVSMC.13

90.8 NF

NF1-NSU-4, NF1-WSU-10,
NF1-HSU-17

NF1 90.8, NF1 88.3.
NF1 88.14

SC #1-11

NY-l

HT-lOBO, VIP-FT

MFH-NSU-l

MSU-1.1-v-sis

MSU-1.1-H-ras

Human foreskin fibroblast cell lines
derived from normal infants

Human smooth muscle cells derived from a
normal adult’

Phenotypically normal fibroblasts derived
from biopsy in NF1 patient that developed
neurofibrosarcoma NF1 90.8b

Human Neurofibrosarcoma cell lines
established at Michigan State Universityc

Cell lines derived from neurofibrosarcomas
that occurred in patients with NF1b

Neurofibroma cell lines derived from
multiple neurofibromas that occurred on an
NF1 patientd

Neurofibroma cell line derived from
neurofibroma that occurred on an NF1
patient’

Human fibrosarcoma-derived cell lines'

Human malignant fibrous histiocytoma cell
line established at Michigan State
Universityc

Malignant cell strain generated in vdtro
from immortal human fibroblast line, MSU-
1.1 (Morgan et al) by transfection of a v-
sis constructo

Malignant cell strain generated in vitro by
transfection of H-ras construct into MSU-
1.1“

92

that arose from mesenchymal cells including HT-1080, VIP-FT, and MFH-WSU-
1. Other cells used as controls were two malignant strains generated from
the same immortal human fibroblast strain MSU-1.1. MSU-1.1 v-sis was
produced by transfection with a v-sis oncogene and, as such, expresses
high levels of PDGF-B mRNA. MSU-1.1 v-ras was generated by transfection
with a v-H-ras oncogene and does not express either TGF-a or PDGF-B mRNA.

The ideal control cells to assess changes that have occurred to
cause a neurofibroma would be the normal human Schwann cells.
Unfortunately, such cells are not readily available and do not grow in
standard cell culture. In any case, we doubt that normal Schwann cells
express significant levels of growth factors, since Ratner et al. (11)
demonstrated that extracts prepared from normal human peripheral nerves
did not contain any mitogens, whereas neurofibroma extracts did. The>most
appropriate control cells to assess changes that cause neurofibrosarcomas

are the neoplastic Schwann cells in neurofibroma cell populations.

RT-PCR

To determine if the neurofibrosarcoma and neurofibroma cell lines
expressed TGF-o and PDGF-B mRNA, and if so, the relative level of
expression, we used an RT-PCR assay. The primers that we used are shown
in Figure 3. One out of the six neurofibrosarcoma cell lines tested
expressed high levels of PDGF-B mRNA, and TGF-a mRNA equivalent to that of
human fibrosarcoma cell line HT-1080 (Figures 4 and 5), which has
previously been shown to express elevated levels of PDGF-B mRNA and
protein (25) as well as TGF-a mRNA (26). Early passage neurofibromas

cells were found to express intermediate or low levels of PDGF-B mRNA

93

Figure 3: PCR PRIMERS

 

GAEDH_E£1m££S:

ri

5’ sense primer
5' -GGTGAAGGTCGGAGTCAACG- 3'
(exon 2)

3' antisense primer
5' -CTTCTGCATGGTGGTGAAGA- 3'
(exon 5)

cONA product = 313 bp

5' sense primer
5' -GAAGGAGCCTGGGTTCCCTG- 3'
(exon 3 - exon 4 junction)

3' antisense primer
5' -TTTCTCACCTGGACAGGTCG- 3'
(exon 4 - exon 5 junction)

cDNA product = 226 bp

5’ sense primer
5' -GTATTGTGTTGGCTGCGTGC- 3'
(exon 2)

3' antisense primer
5' -CACTGTTTCTGAGTGGCACG- 3'
(exon 3)

cONA product = 437 bp

 

94

 

 

 

22 cycles <- GAPDH

26 cycles <- GAPDH

‘— PDGF-B

30 cycles __ ' -_ 1. . ‘— GAPDH
--. 3* -'.._-'°° '5 'Mfiélnl o ”T:

. . — .- - - <— PDGF-B

 

12 345678910

Fig. 4. Expression of GAPDH mRNA and PDGF-B mRNA determined by RT-

PCR.

(A) Sampling of PCR reaction after 22 cycles showing GAPDH product. (B)
Sampling of PCR product after 26 cycles also showing PDGF—B PCR product. (C)
Sampling after 30 cycles of PCR. Lane 1, HT—1080; Lane 2, MFH-WSU-l; Lane
3, NF1-WSU-4; Lane 4, NF1-WSU-10; Lane 5, NF 1-WSU-17; Lane 6, NF1-88.3;

Lane 7, NF1-88.14; Lane 8, NF1-90.8; Lane 9, 90.8 NF; Lane 10, SL-80.

95

22 cycles (— GAPDH

26 cycles <- TGF—Ol
('- GAPDH

30 cycles (— TGF-Ot

COCO-C'III. <—GAPDH

1234567891011

 

Fig. 5. Expression of GAPDH mRNA and TGF-a mRNA determined by RT-PCR.

(A) Sampling of PCR reaction after 22 cycles showing GAPDH product. (B)

Sampling of PCR product after 26 cycles also showing TGF-oz PCR product. (C)

Sampling after 30 cycles of PCR. Lane 1, HT-1080; Lane 2, MFH-WSU-1; Lane

3, VIP-FT; Lane 4, NF1-WSU-4; Lane 5, NF1-WSU-10; Lane 6, NF1-WSU-17;

Lane 7, NF1-88.3; Lane 8, NF1-88.14; Lane 9, NF1-90.8; Lane 10, 90.8 NF;

Lane 11, SL-BO.

96

(Figure 6). Only one neurofibroma cell line, NFb-l, was found to express
both TGF-a mRNA and PDGF-B mRNA (data not shown). To exclude the
possibility that neurofibroma cell line NFb-l, which expresses elevated
TGF-a mRNA, became cross-contaminated by a TGF-a expressing tumor cell
line during cell culture, we reestablished this neurofibroma cell line
from tissue and cells that had been frozen at the time that tumor tissue
was received from the hospital. Our findings confirmed that NFb-l
expresses elevated TGF-o mRNA.

As shown, normal human foreskin fibroblasts and human adult smooth
muscle cells express low'or undetectable levels of PDGF-B, and none of the
normal cells expressed detectable TGF-a mRNA. All experiments were
carried out at least twice and consistent results were obtained. The data
from RT-PCR are representative of the data we obtained in several

experiments. The results are summarized in Table III.

Response of neurofibrosarcoma cells to exogenous growth factors

In order to determineiwhether production of PDGF-B and/or TGF-a could have
a direct effect on the growth of neurofibrosarcoma cells, we have used an
assay in which cells are plated in medium lacking exogenous growth factors
and then are stimulated over a period of 6 days, and at the end of the
assay cell numbers in each culture dish are determined. If no growth was
detected above that which normally occurs in growth factor free medium,
this indicated either that the cells do not have functional receptors for
that growth factor or that the growth of the cells is completely
independent of exogenous growth factors. If significant growth occurred

in response to a particular growth factor, it indicated that the cell

97

  

(— GAPDH

26 a a 00.6.-

<- PDGF-B
30 cycles <— GAPDH

(- PDGF-B

 

Fig. 6. Expression of GAPDH mRNA and PDGF-B mRNA determined by RT-
PCR.

(A) Sampling of PCR reaction after 26 cycles showing GAPDH product and
PDGF-B products. (B) PCR product after 30 cycles of PCR amplification. Lane
1, SL-80; Lane 2, NF1-WSU-17; Lane 3, NFb-1; Lane 4, NFb—2; Lane 5, NFb—5;

Lane 6, NFb-7; Lane 7, NFb-9; Lane 8, NFb-10.

98

(data is based on several experiments)

°++

no PCR product detected after 30 cycles of amplification in

any experiment

light band of PCR product visualized at 30 cycles in some

experiments

a light band was visualized at 26 cycles in some experiments

bright band of PCR product visualized at 26 cycles of
amplification

PCR product visualized at 22 cycles in some experiments

tar"

use

m8

Table III. Results of RT-PCR

99

 

Cells

TGF-a

PDGF-B

 

Narmal_cells
SLBO

SL68
HSVSMC.13
90.8

Sarcoma_cells
HT 1080
VIP-FT
MSU-1.1-v-sis
MSU-1.1-H-ras

Neuratibrasarcoma.cells
NF1-WSU-4

NF1-WSU-10

urn-usu-17

NF1 88.3

NF1 88.14

NF1 90.8

Neuratibrama.cells
urn-1

urn-2

urn-3

urn-4

NFb-5

urn-7

urn-9

NFb—lO

NFb-12

11111.

iiiii.

+ii++++11

 

100

lines expressed functional receptors for that growth factor. We used the
three tumorigenic human neurofibrosarcoma cell lines that grow

continuously in standard culture medium containing serum, and grow to some
degree in medium lacking exogenous growth factors. NF1-WSU-10
proliferates rapidly in the absence of growth factors, whereas NF1-90.8
grows moderately and NF1-WSU-17 grows slowly in these conditions. We also
included bFGF in addition to PDGF and TGF-a in the assay because of
reports that bFGF is found in extracts obtained from human neurofibromas
(11) and is mitogenic for rat Schwann cells in the presence of elevated
cAMP (16). The results demonstrate that neurofibrosarcoma cell line NF1-
90.8 responds markedly to TGF-a, PDGF, and bFGF, indicating that receptors
for these growth factors are expressed (Figure 7). NF1-WSU-17 responds to
TGF-a and bFGF, but is not stimulated by PDGF. None of the growth factors
that we used had a stimulatory effect on neurofibrosarcoma cell line NF1-
WSU-IO (data not shown). These results indicate that neurofibrosarcoma
cell lines NF1-90.8 and NF1-WSU-17, in contrast to normal rat Schwann
cells and human neurofibroma-derived Schwann cells, constitutively express

one or more of the EGF/TGF-a, PDGF, and bFGF receptors.

Effect of suramin on growth factor independence

If growth factors such as PDGF-B and bFGF are produced by
neurofibrosarcomas and have a direct role in the abnormal growth of such
cells, then inhibiting the interaction between growth factors and their
receptors in such cells should have an inhibitory effect on cell growth.
Because two of the neurofibrosarcoma cell lines, NF1-WSU-1O and NF1-90.8,

continue to divide in medium lacking growth factors, we could examine

101

Figure 7. Stimulation of the growth of neurofibrosarcoma cells by
exogenous growth factors. The cells were grown in McM with 0.1 mM calcium
and SR, (no mitogen), or in McM with 0.1 mM calcium and SR, plus 10% FBS
(FBS), or in McM with 0.1 mil calcium and SR, plus 10 ng/ml PDGF (PDGF), or
McM with 0.1 mM calcium and SR, plus 10 ng/ml TGF-a (TGF-a), or McM with
0.1 nM calium and SR, plus 10 ng/ml bFGF (bFGF). The number of cells at the
end of the assay are shown in cell/cm,. The bold horizontal line
demonstrates the number of cells present at the time cells were changed to

experimental conditions.

102

NF1-WSU-l7

cells/cm: x 10"

 

no mitogen FBS PDGF TGF-o bFGF

NF1-90.8

cells/cm2 x 103

 

no mitogon FBS PDGF TG F-o bFGF

103

whether inhibiting growth factor initiated mitogenesis would have a
negative effect on the serum free growth of these cells. The other
neurofibrosarcoma cell lines and all neurofibroma cell lines do not

divide, or divide very slowly (NF1-WSU-17), under these conditions (data
not shown). Suramin is an antifilarial drug that is known to inhibit cell
division that is dependent on growth factor pathways by interrupting
interactions between growth factors and their receptors (27, 28). Since
neurofibrosarcoma cell lines NF1-WSU-10 and NF1-90.8 both express TGF-a
and PDGF-B mRNA, and divide in the absence of exogenous growth factors, we
treated the cells with suramin to determine whether either of these cell
lines depends upon growth factor pathways for growth in the absence of
exogenous growth factors. As control cells we used two tumorigenic cell
lines derived from the same parental human fibroblast cell line, one
generated by transfection with an activated H-ras gene (MSU-1.1-H-ras) and
the other generated by transfection with a v-sis gene (MSU-1.1-v-sis)
(29). Both of these malignantly transformed cell lines grow rapidly in
serum free medium. However, as shown in Figure 8, MSU-1.1-v-sis, is
completely inhibited from growing in serum free medium in the presence of
50pg/ml suramin. The serum free growth of cell strain MSU-1.1-H-ras, is
unaffected by 50pg/ml suramin. Like MSU-1.1-H-ras, neurofibrosarcoma cell
line NF 1-WSU-10 is unaffected by 50pg/ml of suramin. However, the growth
of neurofibrosarcoma line NF1-90.8 in growth factor free medium is
completely inhibited by 50 pg/ml of suramin (Figure 8). These findings
suggest that neurofibrosarcoma-derived cell line NF1-90.8 is dependent
upon endogenous production of a growth factor(s) to proliferate in the

absence of exogenous growth factors.

104

- (-) suramin D (+) suramin

2.50'/

2.02“"
1.54 -
1.06 -

0.58 4

absorbance 450-650nm

 

 

 

 

 

 

 

 

 

 

0.10‘

 

Figure 8. Growth of tumor cell lines in McM medium with 0.1 11M calcium
and SR, (black bar), and in McM medium with 0.1 nM calcium and SR, plus 50
ug/ml of suramin (white bar). (A), MSU-1.1-H-ras. (B), MSU-1.1-v-sis. (C),
NF1-WSU-10. (0), NF1-90.8. Growth is shown as a function of absorbance
calculated by subtracting the absorbance determined at day zero from the
absorbance determined at the end of the assay using the colorimetric assay

described in materials and methods.

105
Discussion

In the present study, we have shown evidence that supports a role
for activation of growth factor pathways, i.e., PDGF-B and TGF-o, in the
pathogenesis of neurofibrosarcomas that occurs in NF1 patients. Our
results clearly demonstrate that the expression of TGF-a mRNA is
associated with malignant transformation of neurofibroma into
neurofibrosarcomas (Table III), since with only one exception,
neurofibroma cell lines do not express TGF-a mRNA but neurofibrosarcoma
cells do. One possible explanation for this result is that neurofibroma
cell line NFb-l is actually derived from a tumor that is at an
intermediate stage between a neurofibroma and a neurofibrosarcoma.

It is clear that the majority of neurofibrosarcomas have elevated
expression of PDGF-B mRNA relative to neurofibroma cell lines. However,
we cannot exclude to possibility that the neoplastic Schwann cells in
neurofibroma cell populations also express levels of PDGF-B mRNA similar
to that of neurofibrosaroma cells, but that the concentration of PDGF-B
mRNA has become diluted due to the presence of the RNA from non-neoplastic
cells that are present in neurofibroma cell populations. In addition,
since we could not compare the level of PDGF-B mRNA to that of normal
human Schwann cells, it is not possible to determine whether the level of
PDGF-B mRNA expression by neurofibroma cells represents upregulation or a
steady-state level of expression in normal human Schwann cells.
Nevertheless, it seems unlikely that normal Schwann cells produce PDGF-B,
since other workers have demonstrated that soluble extracts from normal
human nerve do not contain any factors that are mitogenic for NR6-3T3
fibroblasts (11).

Further evidence that growth factor expression can play a critical

role in the abnormal growth of neurofibrosarcomas is based on studies with

106

suramin. Suramin has been used to inhibit the growth of malignant cells
generated in cell culture by growth factor oncogenes (27, 28). There are
few examples in which inhibition of a growth factor pathway has been shown
to have an inhibitory effect on growth of human tumor-derived cells (30,
31). Our results demonstrate that NF1-90.8 cells, just like those
malignantly transformed by v-sis, are inhibited by 50 pg/ml of suramin.
These results demonstrate that NF1-90.8 cells depend upon endogenous
production of a mitogenic growth factor(s) to proliferate in the absence
of exogenous growth factors.

Lemke (32) has proposed that the abnormal phenotype of neoplastic
Schwann cells in NF1 patients could in part be due to abnormal regulation
of ras and subsequent upregulation of growth factor receptor(s).
Defective regulation of ras proteins in yeast leads to elevation of cAMP
(33). Activation of adenylate cyclase'and subsequent increase in cAMP can
have a causal role in some types of human tumors. For example, a group of
human pituitary tumors have been identified that have constitutively
elevated levels of cAMP (3). The mutations that cause elevation in cAMP
are in the a, subunit of G., a ras-related protein that functions in
activation of adenylate cyclase in pituitary and other types of cells.
Mutations in the a, subunit, which is the GTP-binding domain of G,
proteins, results in decreased GTPase activity and therefore to elevation
of GTP bound to G,,‘which is the active form. Consequently, adenylate
cyclase is constitutively activated and continuously generates cAMP in
these pituitary tumors. If a similar mechanism occurs in human Schwann
cells from NF1 individuals who that express reduced amounts of
neurofibromin, as suggested by Lemke, then the constitutional NF1 defect
alone does not appear to be sufficient to result in constitutively

activated adenylate cyclase, since neurofibroma Schwann cells do not

107

respond to growth factors in the absence of forskolin (11, 12). On the
other hand, the human neurofibrosarcoma cell lines studied by DeClue et
al. (10) have elevated ras-GTP. If adenylate cyclase is activated in
neurofibrosarcoma cells, as a result of their constitutively activated ras
proteins, then it is likely that such cells will have upregulated growth
factor receptor expression as a result of adenylate cyclase-induced
elevation of cAMP.

In vitro studies suggest that rat Schwann cells are similar to human
Schwann cells in their regard to their expression of growth factor
receptors, since Schwann cells derived from human neurofibromas also
require elevated cAMP to respond to specific growth factors. Glial growth
factor (GGF) was the first Schwann cell mitogen identified and stimulates
rat Schwann cells in culture in the absence of elevated cAMP (16). Ridley
et al (14) demonstrated that TGF-Bl and TGF-BZ, like GGF, also stimulate
rat Schwann cells to proliferate in the absence of elevated cAMP. Davis
and Stroobant (16) screened several mitogens to determine which, if any,
could stimulate rat Schwann cells in culture in the absence or presence of
cAMP. Their results verified that GGF, TGF-Bl and TGF-BZ stimulate growth
of rat Schwann cells in the absence of cAMP, and they found that PDGF-BB,
PDGF-AB, bFGF, and aFGF could also stimulate the growth of rat Schwann
cells, but only in the presence of forskolin-induced elevation of cAMP.
cAMP induces Schwann cells to become sensitive to the mitogenic effects of
PDGF-B by elevating the expression of PDGF receptors (15). In the light of
these facts, our results imply that two significant changes have occurred
to neurofibrosarcomas compared to the neoplastic Schwann cells in
neurofibromas. First, we predict that neurofibrosarcoma cells do not
require exogenous stimulation of cAMP to express growth factor receptors,

suggesting that such cells have elevated levels of cAMP. Second, since

108
neurofibrosarcomas can (to various degrees) replicate in the absence of
serum or growth factors, our results suggest either that
neurofibrosarcomas produce endogenous growth factors or that they have
acquired other changes which preclude the need for growth factors in cell
division.

Expression of a growth factor(s) by neurofibrosarcoma cells can be
expected to would likely have two consequences in regard to signal
transduction. First, growth factor expression would further elevate the
level of ras-GTP by activation of guanine nucleotide releasing factors.
It is possible that the effect of growth factor-induced activation of ras
would be magnified in neurofibrosarcoma cells, since these cells lack
neurofibromin (10) required to inactivate ras-GTP. Second, by elevating
the level of ras-GTP and subsequently that of cAMP, production of a
specific growth factor, such as PDGF-B, would further elevate the level of
expression of its own receptor and possibly other growth factor receptors.

To sunnmarize, our findings provide evidence consistent with the
findings of others suggesting a role for mitogenic growth factors in the
pathogenesis of neurofibromas and neurofibrosarcomas that occur in NF1
patients. Furthermore, our results add to the understanding of how Schwann
cells in individuals affected with NF1 can become malignant. Previously,
other workers have shown that loss of the remaining normal NF1 allele is
associated with malignant transformation of neurofibromas (10). This
finding is consistent with that of Menon et al (34), who demonstrated that
the majority of neurofibromas retain the normal NF1 allele. Loss of
heterozygosity of the p53 allele has also been associated with the
formation of neurofibrosarcomas (34). Our findings suggest that
expression of growth factors, including PDGF-B and TGF-a, and the ability

to respond to such growth factors, is a characteristic of some human

109
neurofibrosarcoma cells. It is clear that future progress in
understanding the molecular details involved in the formation of
neurofibromas and neurofibrosarcomas depends on understanding the
mechanism whereby growth factor receptors become upregulated in human
Schwann cells, and the connection, if any, between ras activation and

activation of adenylate cyclase in such cells.

10.

LIST OF REFERENCES

Rubenstein, A.E., (1986) A review of the clinical problem. In A.H.
Rubenstein, R.P. Bunge, and 0.E. Houseman, 0.E. (eds),
Neurofibromatosis. New York Academy of Sciences, New York, pp. 1-
13.

Enzinger, F.M. and Weiss, S.W. (1989) Malignant tumors of peripheral
nerves. In F. M. Enzinger, and S. W. Weiss, (eds), Soft Tissue
Tumors. C.V. Mosby Co., St. Louis. PP. 781-815.

Bade, J.L.(1986) Neurofibromatosis and cancer. In A.E. Rubenstein,
R.P. Bunge, and 0.E. Houseman (eds), Neurofibromatosis, New York
Academy of Sciences, New York, pp. 57-65.

Cawthon, R.M., Weiss, R., Xu, G., Viskochil, 0., Culver, M.,

Stevens, J., Robertson, M., Dunn, 0., Gesteland, R., O’Connel, P.

and White, R. A. (1990) Major segment of the neurofibromatosis type

ézgene:3 c%N1A sequence, genomic structure, and point mutations. Cell,
, 19 -2 .

Xu, G., O’Connel, P., Viskochil, 0.,Cawthon, R., Robertson, M.,
Culver, M., Dunn, 0., Stevens, J., Gesteland, R., White, R. and
Weiss, R. (1990) The neurofibromatosis type 1 gene encodes a protein
related to GAP. Cell, 62, 599-608.

Haubruck, H. and McCormick, F. (1991) Ras p21: effects and
regulation. Biochem. Biophys. Acta., 1072, 215-229.

Satoh, T., Nakafuku, M. and Kiziro, Y. (1992) Function of ras as a
molecular switch in signal transduction. J. Biol. Chemu, 267, 24149-
24152.

Downward, J., Riehl, R., Wu, L., and Weinberg, R.A. (1990)
Identification of a nucleotide exchange-promoting activity for
p21"”. Proc. Natl. Acad. Sci., 87, 5998-6002.

Huang, Y, K., Kung, H.-F., Kamata, T. (1990) Purification of a
factor capable' of stimulating the guanine nucleotidei exchange
reaction of ras proteins and its effect on ras-related small
molecular mass G proteins. Proc. Natl. Acad. Sci., 87, 8008-8012.

Declue, J.E., Papageorge, A.G., Fletcher, J.A., Diehl, S.R., Ratner,
N., Vass, W.C. and Lowy, D.R. (1992) Abnormal regulation of

110

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

111

mammalian p21"“ contributes to malignant tumor growth in van
Recklinghausen (type 1) neurofibromatosis. Cell, 69, 265-273.

Ratner, N., Lieberman, M.A., Riccardi, V.M. and Hong, D-M H. (1990)
Mitogen accumulation in van Recklinghausen neurofibromatosis. Ann.
Neurol., 27, 298-303.

Riccardi, V.M. (1986) Growth-promoting factors in neurofibroma crude
extracts. In Neurofibromatosis, New York Academy of Sciences, New
York, pp. 206-226.

Sheela, S., Riccardi, V.M. and Ratner, N. (1990) Angiogenic and
invasive properties of neurofibroma Schwann cells. J. Cell Biol.,
111, 645-653.

Ridley, A.J., Davis, J.B., Stroobant, P. and Hartmut, L. (1989)
Transforming growth factors-81 and B2 are mitogens for rat Schwann
Cells. J. Cell Biol., 109, 3419-3424.

Weinmaster, G. and Lemke, G. (1990) Cell-spedific cyclic AMP-
mediated induction of the PDGF receptor. EMBO J., 9, 915-920.

Davis, J.B. and Stroobant, J.B. (1990) Platelet-derived frowth
factors and fibroblast growth factors are mitogens for rat Schwann
cells. J. Cell Biol., 110, 1353-1360.

Chomczynski, P. and Sacchi, N. (1987) Single step method of RNA
isolation by acid guanidinium thiocyanate-phenol-chloroform
extraction. Anal. Biochem., 162, 156-159.

Yang, J-L., Maher, V.M. and McCormick, J.J. (1989) Amplification and
direct nucleotide sequencing of cONA from the lysate of low numbers
of diploid human cells. Gene, 83, 347-354.

Wang, A.M., Doyle, M.V. and Mark, D.F. (1989) Quantitation of mRNA
by the polymerase chain reaction. Proc. Natl Acad. SCi. USA, 86,
9719-9721.

Vassar, R., Hutton, M.E. and Fuchs, E. (1992) Transgenic
overexpression of transforming growth factor a bypasses the need for
c-H-ras mutations in mouse skin tumorigenesis. Mol. Cell. Biol. 12,
4643-4653.

Noonan, K.E., Beck, C., Holzmayer, T.A., Chin, J.E., Wunder, J.S.,
Andrulis, I.L., Gazdur, A.F., Willianu C.L., Griffith, 8., Von Hoff,
0.0. and Robinson, I.-B. (1990) Quantitation analysis of MDRI
(multidrug resistance) gene expression in human tumors by polymerase
chain reaction. Proc. Natl. Acad. Sci. USA, 87, 7160-7164.

Ryan, P. A, McCormick, J. J. and Maher, V. M. (1987) Modification of

MCDB- 110 medium to support prolonged growth and consistent high
ilgning efficiency of human diploid fibroblasts. Exp. Cell. Res. ,
7 , 318-328.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.
33.

134.

112

Scudiero, 0.A., Shoemaker, K.D., Paull, K.D., Monks, A., Tierney,
S., Nofzinger, T.H., Currens, M.J., Seniff, D. and Boyd, M.R. (1988)
Evaluation of a soluble tetrazolium/formazan assay for cell growth
and drug sensitivity in culture using human and other tumor cell
lines. Cancer Res., 48, 4827-4833.

Cory, A.H., Owen, T.C., Barltrop, J.A. and Cory, J.G. (1991) Use of
an aqueous soluble tetrazolim/formazan assay for cell growth assays
in culture. Cancer Commun., 3, 207-212.

Pantazis, P., Pelicci, P.G., Dalla-Fauera, R. and Antoniades, H.N.
(1985) Synthesis and secretion of proteins resembling platelet-
derived growth factor by human glioblastoma and fibrosarcoma cells
in culture. Proc. Natl. Acad. SCi. USA, 82, 2404-2408.

Derynk, R., Goeddel, 0.V., Ullrich, A., Gutterman, J.V., Williams,
R.D., Bringman, T.S. and Berger, W.H. (1987) Synthesis of messenger
RNAs for transforming growth factor a and 8 and the epidermal growth
factor receptor by human tumors. Cancer Res., 47, 707-712.

Garret, J.S., Coughlin, S.R., Niman, H.L., Tremble, P.M., Giels,
G.M. and Williams, L.T. (1984) Blockade of autocrine stimulation in
simian sarcoma virus-transformed cells reverses down-regulation of
platelet-derived growth factor receptors. Proc. Natl. Acad. SCi.
USA, 81, 7466-7470.

Huang, S.S. and Huang, J.S. (1988) Rapid Turnover of the Platelet-
derived growth factor receptor in sis-transformed cells and reversal
by suramin. J. Biol. Chem., 263, 12608-12618.

Yang, 0., Kohler, S.K., Maher, V.M., and McCormick, J.J. (1994) v-
sis oncogene-induced transformation of human fibroblasts into cells
capable of forming benign tumors. Carcinogenesis, (in press).

Heldin, C.-H. and Westermark, B. (1991) Platelet-derived growth
gactor and autocrine mechanisms of oncogenic processes. Oncogenesis,
, 109-123.

Fleming, T.P., Matsui, T., Heidaran, M.A., Molloy, C.J., Artrip, J.
and Aaronson, S.A. (1992) Demonstration of an activated platelet-
derived growth factor autocrine pathway and its role in human tumor
cell proliferation in vitro. Oncogene, 7, 1355-1359.

Lemke, G. (1990) Mitogein Signal. Nature, 348: 201.

Dumont, J.E., Jauniaux, J.-C. and Roger, P.P. (1989) The cyclic AMP-
mediated stimulation of cell proliferation. Trends Biochem. SCi.,
14, 67-71.

1989.

Landis, C.A., Masters, 5.8., Spada, A., Pace, A.M., Bourne, H.R. and
Vallar, L. (1989) GTPase inhibiting mutations activate the a chain
of G, and stimulate adenylyl cyclase in human pituitary tumours.
Nature, 340, 692-696.

35.

113

Menon, A.G., Anderson, K.M., Riccardi, V.M., Chung, R.Y., Whaley,
J.M., Yandell, D.W., Farmer, G.E., Freiman, R.N., Lee, J.K., Li,
F.P., Barker, D.F., Ledbetter, D.H., Kleider, A., Martuza, R.L.,
Gusella, J.F. and Seizinger, B.R. (1990) Chromosome 17p deletions
and p53 gene mutations associated with the formation of malignant
neurofibrosarcomas in van Recklinghausen neurofibromatosis. Proc.
Natl. Acad. SCi. USA, 87, 5435-5439.

CHAPTER III

Characterization of a Neoplastic Huaan Fibrous Histiocytoaa
Cell Line Including Evidence Suggesting a Mesenchylal Origin

Jeremy W. Wray, Veronica M. Maher, and J. Justin McCormick

Carcinogenesis Laboratory, Fee Hall
Department of Microbiology and Department of Biochemistry
Michigan State University, East Lansing, MI, 48824-1316

114

In this study we characterize a human malignant fibrous histiocytoma
(MFH) cell line that was derived from a recurrent MFH that occurred in a
72 year old male. Several morphological types of cells were observed in
the primary culture. After several population doublings a homogeneous
infinite life span cell line, designated MFH-WSU-l, which consisted of
ovoid-spindle shaped cells and frequent multi-nucleated giant cells,
remained. When injected into athymic mice, the MFH-WSU-l cells produced
a tumor histologically identical to the original tumor indicating that the
cells are tumor-derived. This cell line expressed platelet-derived growth
factor B-chain (PDGF-8) mRNA at a level similar to that of human
fibrosarcoma cell line HT-1080 and several other human fibrosarcomas.
Furthermore, MFH-WSU-l proliferates in response to exogenous transforming
growth factor alpha (TGF-a). Since these growth factors and receptors are
frequently associated with tumors of mesenchymal origin and not associated
with cells of monocyte/macrophage origin, these results support a

mesenchymal origin for this MFH cell line.

115

INTRODUCTION

Malignant fibrous histiocytoma (MFH) is the most common soft tissue
malignancy (4,8). ‘This diagnosis describes a histologically heterogeneous
group of sarcomas. The majority of MFHs are of the storiform-pleomorphic
pattern (4). Other less common forms include the myxoid, giant cell,
inflammatory, and angiomatoid types. Despite the fact that MFH is the
most commonly diagnosed soft tissue malignancy, there are very few reports
describing and characterizing established human MFH cell lines.

The cell type from which MFHs arise has been the subject of much
controversy. The designation of MFH was originally used because
investigators observed that MFH explants give rise to cells in culture
that resemble histiocytes (9). Later passage cells appeared fibroblastic
and it became generally accepted that histiocytes were able to assume the
function and appearance of fibroblasts. Conflicting results have been
reported in investigations that used immunocytochemical methods to
demonstrate the presence of mesenchymal or histiocytic markers in MFH-
derived cells. Several reports conclude that MFH cells are of mesenchymal
origin (11,12,7,18). Conversely, other reports indicate that MFHs are of
histiocytic origin (16,14,15,1). Each of these studies relies on the
specificity of cell markers. However, one of the characteristics of MFHs
is that they are composed of a diverse cell types, including neoplastic
and nonneoplastic cells. Therefore, data from studies of fixed tumor
tissue or primary cell cultures must be interpreted with caution, since it

116

117
is often unclear which cells are neoplastic and which are not. In the
present study we characterize a newly established MFH cell line and we
provide data on the malignant cells that suggest that the tumorigenic

cells are of mesenchymal origin.

MATERIALS AND METHODS

Culture of MFH cells

Tumor tissue was obtained from a fresh human tumor that had been
surgically excised and was transported to our laboratory in Lebovitz
medium containing 100 pg/ml Gentamicin sulfate. Upon arrival at this
laboratory, the tissue was fragmented into smaller pieces and mechanically
agitated using a sterile syringe and cannula to break down the tissue and
free individual cells. Some of the resulting cells and small tumor
fragments were placed in 75 cm’ tissue culture flasks (Corning Glass Works,
Corning N.Y.) and cultured in Eagle’s minimum essential medium
supplemented with 0.2 mM aspartic acid, 1.0 mM sodium pyruvate, 0.2 mM
serine, 100 U/ml penicillin, 100 pg/ml streptomycin, 10 pg/ml
hydrocortisone, and 5% or 10% fetal bovine serum (FBS) (culture medium).
Cultures were incubated at 37°C in a humidified atmosphere of 5% C0, and

air.

Cloning Assay

Cells were plated at 100, 200, or 500 cells per 100 mun-diameter
culture dish in McM medium (13) containing 10% supplemented calf serum,
100 U/ml penicillin, 100 pg/ml streptomycin, and 10 pg/ml hydrocortisone.

Medium was changed weekly. Cell clones were stained after 3 weeks.

113 ‘
Saturation Density Assay
Cells were plated at 5 x 10‘ cells per 60 mm-diameter culture dishes
in 5 ml of culture medium containing 5% FBS. Medium was changed every 3
days. Duplicate culture dishes for each cell line were counted on the

indicated days after plating.

Tumorigenicity Assay
Cells (1 x 10‘ to 10 x 10‘) in exponential growth were trypsinized,

suspended in a total volume of 0.2 ml of Eagle’s medium and injected
subcutaneously in the scapular region of six week old BALB/c athymic mice.
The mice were examined weekly for tumors. Tumors that were removed were
fixed in formalin, embedded in paraffin, sectioned, and stained with

hematoxylin and eosin prior to histological evaluation.

Isolation of RNA

Cells were grown in 150 min-diameter dishes in culture medium
containing supplemented calf serum rather then fetal bovine serum. Total
RNA was isolated from cells just prior to their reaching confluence, using
the procedure of Chomczynski and Sacchi (2) essentially as described. The
RNA was resuspended in diethylpyrocarbonate-treated H,O, and stored at -
80°C. The concentration of RNA was determined from the optical density at
260nm, and the integrity of the RNA was qualitatively determined from the
appearance of ribosomal RNA bands on 1% agarose/formaldehyde ethidium

bromide gels .

Detection of ilNA by RT-PCR

Synthesis of c0NA using Moloney murine leukemia virus reverse

transcriptase (New England Biolabs) and amplification of specific cDNAs

119
using Taq polymerase (Perkin-Elmer Cetus, Norwalk, CT), RT-PCR, was
carried out using the procedure described by Yang et al (19). The
primers used for glyceraldehyde phosphate dehydrogenase (GAPDH) and TGF-a
cDNA have been described by Vassar et al. (17) and primers for PDGF-B cDNA
are described by Yang et al. (19).

Response of ll-‘H cells to PDGF-B and TGF-a

To determine the response of cells to specific growth factors, cells
growing exponentially in culture medium were trypsinized and plated at 7.5
x 10“ cells per 60 mum-diameter culture dish in 5 ml of culture medium with
5% fetal bovine serum. After 24 h, the medium was removed and replaced
with 5 ml of medium containing McM medium modified to contain only 0.1 mM
calcium, rather than the usual 1.0 mM calcium, and containing the serum
replacement supplements (designated SR,) specified by Ryan et al. (13), but
lacking any exogenous protein growth factors plus 1% FBS. Fetal bovine
serum was added to positive growth control conditions to achieve a
concentration of 10% FBS. Growth factors including PDGF-B (Oncogene
Science, Uniondale, N.Y.) and TGF-o (Oncogene Science) were added to
achieve a final concentration of 2 ng/ml or 10 ng/ml. Cells in triplicate

60 mun-diameter culture dishes were counted on day 8.

RESULTS
ifH Cells in Culture
Cells and tissue from the MFH gave rise to a primary cell culture
that consisted of a heterogeneous population of cells (Fig. 1). We
reduced the concentration of serum present in culture medium to 5% and

found that this allowed a more homogeneous population of cells to overgrow

120

 

Figure 1. Early passage (passage two) MFH-WSU-l cells in culture

demonstrating variety of morphological types of cells in primary culture.

121

the rest the cells in primary cultures. These cells appeared
morphologically transformed and had an ovoid-spindle shaped appearance in
cell culture (Fig. 2). These cells have continued to divide in cell
culture for more than 200 population doublings. To verify that the
malignant cells were present in the primary population of cells we thawed
cells and tissue that had been frozen at the time the human tumor was
received and placed the material in culture medium that contained only 1%
supplemented calf serum. The predominant cell type that grew out under
these conditions had cell morphology identical to the infinite life span
cells that eventually grew'out of the cultures in standard culture medium.
These results suggest that the malignant cells are present in the primary
culture.

Several clonal papulations were isolated from the heterogeneous
primary cultures, and passaged as individual cell strains. These strains
demonstrated a variety of morphologies, including fibroblastic and
histiocytic strains. However, each of the cell strains senesced in
culture when passaged continuously. Cloning assays of the homogenous late
passage cell line have demonstrated that the cloning efficiency of these
cells is less than one percent, and explains why we did not isolate these

cells from cloning assays of the early passage culture primary cells.

Saturation Density of MFH-HSU-l

We used a saturation density assay to compare the growth of this MFH
cell line, designated MFH-WSU-l, to that of normal human fibroblasts and
human fibrosarcoma-derived cell line HT-1080. Under the conditions used,
the MFH-WSU-1 cells proliferated considerably more slowly than HT-1080,
but over time reached a density only 2-fold lower than that of HT-1080
(Fig. 3). MFH-WSU-l grew at a rate similar to that of the normal

122

Figure 2. Morphology of infinite life span cell line derived from human
MFH. (A) MFH cells in culture including avoid to fibroblastic cells and
several multi-nucleated giant cells (arrows). (8) Higher power of MFH

cells in culture including a multi-nucleated giant cell (arrow).

123

...C ...”)..an

J}

.0

011 ha“

 

ells an:

 

124

 

10

cells / cm” x 10"

 

 

 

 

OJ
0 1 4 7 10 13 18 19 30

DAYS

Figure 3. Saturation density assay. Cells were plated and cultured as
described. (+) represents growth of human f ibrosarcoma cell line HT-1080.
(0) represents normal human foreskin fibroblasts. (4) represents human MFH

cell line MFH-WSU-l. Cell growth is calculated as the number of cells per

cm2 .

125
fibroblasts until the cells became newly confluent. Under these
conditions, the~MFH-WSU-1 cells were able to continue dividing at the same
rate, whereas the rate of growth of the normal fibroblasts slowed as the
density increased. As a result, MFH-WSU-l had achieved a 5-fold greater
density than the normal fibroblasts by the time the assay was ended on day

30.

Tuaorigenicity Assay

Both early or late passage MFH-WSU-l cells that were injected
subcutaneously into athymic mice produced tumors. However, if such tumors
were not removed within approximately two months they eventually
regressed. A total of 8 mice were injected with these cells. Tumors
formed in seven of the eight mice and the tumors grew to approximately 0.5
cm in diameter. The two tumors that were removed had a histological
pattern similar to the original human tumor (Fig. 4). Microscopic
examination of hematoxylin and eosin stained sections demonstrated
numerous mitotic figures in the tumor mass. We also noted the presence of
numerous mononuclear cells at the periphery of the tumor mass that
appeared to be invading tumor tissue (Fig. 5), suggesting that the
regression of these tumors occurred as the result of residual

immunological response of the athymic mice.

Growth Factor IRNA Expression

Investigators have previously reported that human fibrosarcoma-
derived cell lines, such as HT-1080, express elevated levels (relative to
normal human fibroblasts) of mitogenic growth factors including TGF-a (3)
and PDGF-B (10). McCormick and his colleagues have also found that human

fibrosarcoma cell lines frequently express both of these growth factors

126

Figure 4. Hematoxylin and eosin stained tumor sections. (A) Original
human tumor that occurred in a 72 year old male and gave rise to cell line
MFH-WSU-I in culture. (8) Tumor that occurred when MFH-WSU-l cells were

injected subcutaneously into an athymic mouse, as described.

127

. 7,. i 54;“? I I ‘
.4 14'. dfg’r'
‘ ’IQ‘Q’D"

 

128

 

Figure 5. Hematoxylin and eosin stained section of tumor produced by MFH-
WSU-l when cells from culture were injected subcutaneously into an athymic
mouse as described. This field demonstrates infiltration of host

mononuclear cells into the neoplastic tissue.

129
(unpublished results). The results showed that MFH-WSU-l expresses levels
of PDGF-B mRNA similar to or higher than that of HT-1080, but WSU-1 do not
express TGF-a mRNA (Fig. 6).

Response of MFH-WSU-l to aitogenic growth factors

MFH-WSU-l divide slowly in the absence of exogenous growth factors
or serum. MFH-WSU-l cells responded markedly to 2 ng/ml or 10 ng/ml TGF-o
in the absence of any other growth factor, indicating that this MFH cell
line expresses functional receptors for this growth factor (Fig. 7).
However, MFH-WSU-l did not respond to recombinant PDGF-B, suggesting that
this cell line does not express the appropriate receptor for this growth

factor.

DISCUSSION

Studies of this MFH cell line have demonstrated that several
different morphological cell types were present in primary cultures
prepared from the original tumor. These results are similar to those in
the early studies of other investigators who identified several cell types
including those with histiocytic features (9). None of the clonally-
derived independent cell strains isolated from primary MFH-WSU-l cell
populations, and chosen because they represented the various predominant
cell types making up that population, proved to have an infinite life-
span. The cells that eventually grew out of the tumor-derived tissue and
continued to divide indefinitely did not resemble any of the cell strains
isolated in various cloning experiments. These findings suggest that
several types of cells present in primary cultures do not represent the

neoplastic cells that formed the human tumor. ‘The infinite life span cell

GAPDH --->

GAPDH --->
PDGFB --->
GAPDH --->
PDGFB —-->

A

 

 

130

,, "
cycles w A 4'

26

cycles

 

30

cycles

 

<--- GAPDH
<--- TGF-a
<-—— GAPDH
<--- TGF-a
<——— GAPDH

Figure 6. Expression of GAPDH, PDGF-B, and TGF-a mRNA determined by RT-

PCR. PDGF-B and GAPDH mRNA expression shown in panel A. TGF-a and GAPDH

mRNA expression shown in panel 8.

Lane (1), HT-1080. Lane (2), MFH-WSU-l.

Lane (3), Normal human foreskin fibroblasts.

131

 

cells/cm2 x 10’

 

Figure 7. Growth response of MFH-WSU-l cells to PDGF-BB and TGF-a. Growth
is shown as cells per cm’. The bold horizontal line represents the number
of cells present at the time that the cells were changed to the
experimental conditions as described. All conditions include McM medium
containing 0.1 mM calcium and SR, plus the following as listed: (A) 10%
FBS, (8) no growth factor added, (C) 2 ng/ml PDGF-BB, (D) 10 ng/ml PDGF-
80, (E) 2 ng/ml TGF-a, (F) 10 ng/ml TGF-a.

132
line that eventually grew out of the primary cell culture population
formed tumors in athymic mice, and these tumors had similar histological
features similar to that of the original human tumor, indicating that
these cells are the bane fide MFH cells.

Cells of mesenchymal origin and those of the monocyte/macrophage
lineage differ in the specific growth factor receptors that they express
and those growth factors that they respond to. Histiocytic cells and
other cells of hematopoietic origin respond to mitogens such as
granulocyte/macrophage-colony stimulating factor. They have not been
shown to respond to TGF-a (6). Fibroblasts and other mesenchymal cells
have receptors for growth factors such as epidermal growth factor (EGF)
and TGF-a, PDGFs, and fibroblast growth factors (FGFs). Our results
indicate that these MFH-WSU-l cells respond to TGF-o indicating that the
receptor for this growth factor, the EGF/TGF-a receptor, is present on
these cells. These receptors have been identified on many types of solid
tumors, including those of mesenchymal origin, but have not been
associated with cells of the monocytic/macrophage lineage or tumors that
arise from these cells. We have also found that MFH-WSU-l cells express
elevated levels of PDGF-B mRNA, which. is a characteristic of human
fibrosarcoma-derived cell line HT-1080 and several other human
fibrosarcoma-derived cell lines. However, we found that MFH-WSU-l cells
did not respond to PDGF-B, a potent mitogen for normal human fibroblasts.
One possible explanation for this observation is that MFH-WSU-l cannot
respond to exogenous PDGF-B because the level of PDGF-B produced
endogenously is already at a saturating level. Alternatively, MFH-WSU-l
may not have receptors for PDGF-B and the elevated production of PDGF-B,
which is suggested by RT-PCR results, could exert some role in stromal

formation or neovascularization (5).

133

The tumors that occurred when MFH-WSU-I cells were injected
regressed if they were not removed. When static or regressing tumors were
removed and evaluated microscopically, several mitotic figures *were
observed in each high power field, indicating that despite the shrinking
size of the tumor mass, the malignant cells continued to divide. This
observation suggests that residual immunological factors in athymic mice
can effectively destroy MFH-WSU-l cells (20). We also observed
mononuclear cells invading the tumors that occurred in athymic mice (Fig.
5), suggesting that MFH-WSU-l cells secrete chemotactic factors and that
such mononuclear cells contribute to destruction of the tumor cells. The
infiltration of immune cells into the tumor mass is a known characteristic
of MFH (20).

The overall slow rate of growth of MFH-WSU-l cells in viva relative
to other human-derived malignant cell lines, such as HT-1080 (Fig. 3),
could also contribute to the ability of mouse factors to cause regression
of these tumors. However, since in this laboratory we have observed that
other human tumor-derived cell lines, not of MFH origin, also grow slowly
in culture and in athymic mice, but produce progressively growing tumors
capable of killing the mice, slow growth alone is not enough to account
for regression of the tumors produced by MFH-WSU-l cells in athymic mice.

In conclusion, we have established and characterized a cell line
derived from a human MFH, identified the malignant cells and demonstrated
these cells produce tumors in athymic mice that are consistent with the
diagnosis of a MFH. The numerous studies that have relied an
immunohistochemical markers in fixed human MFH tissue to determine the
cell of origin have two weaknesses. First, MFH tissue contains several
types of cells including many non-neoplastic cells, and as a result

studies of fixed tissue are difficult to interpret. Second, the majority

134
of the immunocytochemical markers routinely do not distinguish between
mesenchymal and histiocytic cells. We have demonstrated that MFH-WSU-l
cells produce PDGF-B mRNA and respond to exogenous TGF-a. The receptors
for these and other growth factors are commonly associated with cells that
give rise solid tumors including sarcomas and carcinomas (6) but are not
usually associated with tumors that arise from hematopoietic cells. It is
possible that these or other receptors may be useful markers in
distinguishing between tumors originating from mesenchymal cells and those

of histiocytic origin.

10.

LIST OF REFERENCES

Binder, S.W., Said, J.W., Shintaku, P., and Pinkus, G.S. A
histiocyte-specific marker in the diagnosis of malignant fibrous
histiocytoma: use of monoclonal antibody KP-l (C068). Am J Cflin
Pathol 1992; 97: 759-763.

Chomczynski, P., and Sacchi, N. Single step method of RNA isolation
by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal
Biochem 1987; 162: 156-159.

Derynk, R., Goeddel, 0.V., Ullrich, A., Gutterman, J.V., Williams,
R.D., Bringman, T.S., and Berger, W.H. Synthesis of messenger RNAs
for transforming growth factor a and B and the epidermal growth
factor receptor by human tumors. Cancer Res 1987; 47: 707-712.

Enzinger, F.M., and Weiss, S.W. Sbft Tissue sarcomas 1989; St.
Louis, Missouri: C.V. Mosby Co.

Forsberg, K., Valyi-Nagi, I., Heldin, C.-H., Herlyn, M., and
Westermark, B. Platelet-derived growth factor (PDGF) in oncogenesis:
Development of a vascular connective tissue stroma in
xenotransplanted human melanoma producing PDGF-BB. Proc Natl Acad
SCi USA 1993; 90: 393-397.

Heldin, C.-H., and Westermark, B. Platelet-derived growth factor and
autocrine mechanisms of oncogenic processes. Oncogenesis 1991; 109-
123.

Iwasaki, H., Isayama, T., Johzaki, H., and Kikuchi, M. Malignant
fibrous histiocytoma. Am J Pathol 1987; 128: 528-537.

Kanzaki, T., Kitajima, S., and Suzumori, K. Biologcial behavior of
cloned cells of human malignant fibrous histiocytoma in vivo and in
vitro. Cancer Res 1991; 2133-2137.

Ozello, L., Stout, A.P., Murray, M.R. Cultural characterization of
malignant histiocytomas1and fibrous xanthomas. Cancer, 1963; 16:331-
344.

Pantazis, P., Pilicci, P.G., Dalla-Fauera, R., and Antonaides, H.N.
Synthesis and secretion of proteins resembling platelet-derived
growth factor by human glioblastoma and fibrosarcoma cells in
culture. Proc Natl Acad Sci USA 1985; 82: 2404-2408.

135

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

136

Roholl, P.J.M., Kleyne, J., Elbers, H., Van Der, Vegt, M.C.D.,
Albus-Lutter, C., and Van Unnik, J.A.M. Characterization of tumour
cells in malignant fibrous histiocytomas and other soft tissue
tumours in comparison with malignant histiocytes. I.
Immunohistiochemical study on paraffin sections. J Pathol 1985a;
147: 87-95.

Roholl, P.J.M., Kleyne, J., Van Unnick, J.A.M. Characterization of
tumor cells in malignant fibrous histiocytomas and other soft-tissue
tumors, in comparison with malignant histiocyes. Am J Pathol 1985b;
121:269-274.

Ryan, P.A., McCormick, J.J., and Maher, V.M. Mocification of MCDB-
110 medium to support prolonged growth and consistent high cloning
efficiency of human diploid fibroblasts. Exp Cell Res 1987; 172:
318-328.

Sieber, S.C., Lopez, V., Rosai, J., Buckley, P.J. Primary tumor of
spleen with morphologic features of malignant fibrous histiocytoma.
Am J Surg Pathol 1990; 14: 1061-1070.

Shirasuna, K., Masaru, S., Miyazaki, T. Establishment and
characterization of neoplastic cells from a malignant fibrous
histiocytoma. Cancer 1985; 55: 2521-2532.

Strauchen, J.A., and Dimitriu-Bona, A. Malignant fibrous
histiocytoma: expression of monocyt/macrophage differntiation
antigens detected with monoclonal antibodies. Am J Pathol 1986; 124:
303-309.

Vassar, R., Hutton, M.E., and Fuchs, E. Transgenic overexpression of
transforming growth factor a bypasses the need for c-H-ras mutations
in mouse skin tumorigenesis. Mol Cell Biol 1992; 12: 4643-4653.

Wood, G.S., Beckstead, J.H., Turner, R.R., Hendrickson, M.R.,
Kempson, R.L., and Warnke, R.A. Malignant fibrous histiocytoma tumor
cells resemble fibroblasts. Am J Sprg Pathol 1986; 10: 323-335.

Yang, 0., Kohler, S.K., Maher, V.M., and McCormick, J.J. v-sis
oncogene-induced transformation of human fibroblasts into cells
capable of forming benign tumors. Carcinogenesis 1994;(in press)

Yoshida, M., Matasuzaki, H., Hata, H., Matsuno, F., Takeya, M.,
Okabe, H., and Takatsuki, K. Neutrophil chemotactic factors produced
by malignant fibrous histiocytoma cell lines. Br J Cancer 1993; 67:
508-513.

CHAPTER IV

Fibroblasts Derived frol Norlal Skin Biopsies
froa a Child with Infantile Myofibroaatosis have
Abnoraal Growth Characteristics

Jeremy W. Wray, Saroj Kapur, and J. Justin McCormick

Carcinogenesis Laboratory, Fee Hall
Department of Microbiology and Department of Biochemistry
Michigan State University, East Lansing, MI, 48824-1316

137

Infantile myofibromatosis is a rare disease in which neonates are
born with benign lesions composed of fibroblastic or myofibroblastic
cells. In most cases, the prognosis is good with eventual spontaneous
regress of the tumors. In some instances, the lesions may involve the
viscera and possibly cause death. This report describes a patient who was
diagnosed with infantile myofibromatosis shortly after birth. This case
in unusual because the lesions have neither been life threatening nor have
they regressed. The patient is currently eight years old. We obtained
two biopsies of apparently normal skin at the site of a surgical
procedure. This report describes the studies of fibroblasts derived from
these biopsies and demonstrates that such cells have abnormal growth
characteristics including the ability to form foci in medium containing 1%
serum and the ability to form benign tumors when injected into athymic
mice. Fibroblasts derived from the skin of normal humans do not exhibit

similar characteristics.

138

INTRODUCTION

Infantile myofibromatosis (IN) is an uncommon congenital syndrome
that is characterized by the presence of benign tumors that occur in the
skin, subcutaneous tissues, skeletal muscle, bone, and internal organs
(2). The clinical course is variable. Typically, more than half of the
tumors are present at birth, and the occurrence of new tumors is common
only in the first two years of life. The majority of cases in which
lesions are limited to soft tissues and bone generally carry a good
prognosis including eventual spontaneous regression of the tumors. When
there is considerable involvement of the viscera the prognosis is poor,
and death often occurs at birth or shortly thereafter.

The etiology of IN is not known. The majority of cases of IM are
sporadic. However, there are several reports of a familial pattern of
inheritance (12,1), suggesting that the disease is either an autosomal
dominantly inherited disease with reduced penetrance or an autosomal
recessive inherited disease, with the former being the most favored (8).
In either case, it appears that affected individuals carry an underlying
genetic defect that predisposes them to the abnormal proliferation of
fibroblasts and/or myofibroblasts.

This report includes a short clinical case report of an affected
individual with an atypical clinical course. In addition, we describe
studies of skin fibroblasts derived from two normal biopsies that were
obtained at the time of an orthopedic procedure from an unaffected area of

139

140
skin of this patient. The fibroblasts from this patient exhibited
abnormal growth in culture and produced benign nodules when injected into

athymic mice.

CASE REPORT

This Caucasian male infant was delivered by caesarean section to a
healthy mother after an uneventful term pregnancy. Apgar scores were 9
and 10. In the post delivery period hard lumps were noted on his back,
legs and tongue. His right foot was flaccid and in severe dorsiflexion.
Otherwise, the infant appeared healthy and in no distress.

A lesion was removed from the left lateral chest wall for
histological diagnosis. Microscopic examination of hematoxylin and eosin
stained sections revealed a well circumscribed nodule composed of a
proliferation of spindle shaped cells in whorls, fascicles and bundles,
which in many areas infiltrate through skeletal muscle. The spindle
shaped cells were relatively uniform and elongate, but in some areas had
distinctly plump, blunt ended nuclei and relatively abundant amphophilic
cytoplasm. There were up to 2-3 mitoses per 10 high power fields with no
atypical mitoses noted. Immunoperoxidase staining performed for S-100
protein was negative except for neural elements within the nodule.
Electron microscopy examination of the proliferating fusiform cells
demonstrated numerous profiles of slightly dilated reticulum as well as
prominent Golgi zones. External lamina were sparse. The periphery of
many of the cells contained myofilaments in association‘with Bence bodies.
The nuclei were ovoid with dispersed chromatin and large active-appearing
nucleoli. These findings demonstrate that the spindle shaped cells have

features of fibroblasts and myofibroblasts.

141

This child, who is now 8 years old, continues to have new lesions
arise. Magnetic resonance imaging has been used to follow the occurrence
of these lesions. These have developed on the right breast and beside the
trachea. The patient also has cutis marmorata (mottled atrophic appearing
change) over extensive areas of the skin, especially on the right side of
the face and neck, back, and legs. Biopsy of these regions revealed no
specific findings. Other health-related problems include recurrent otitis
media, tympanic membrane perforations, and recurrent sinusitis. The

patient has undergone surgery for the deformities of his right foot.

MATERIALS All) nennoos

Culture of IM cells

Skin biopsies were obtained at the time that the patient underwent surgery
for the deformity of his right foot. Two biopsies were taken from the
margin of the surgical incision. The skin at this site appeared normal.
The biopsies were fragmented and then agitated using a sterile syringe and
canula to free individual cells. The resulting mixture of tissue
fragments and individual cells was plated in culture medium that consisted
of Eagle's minimum essential medium supplemented with 0.2 mM aspartic
acid, 1.0 mM sodium pyruvate, 0.2 mM serine, 10% supplemented calf serum
(SCS), 100 U/ml penicillin, 100 pg/ml streptomycin, and 10 pg/ml
hydrocortisone. Cultures were incubated at 37°C in a humidified

atmosphere of 5% CO, and air.

Focus assay
Focus assays were performed by plating 5 x 10‘ cells per 100 mm

culture dish in culture medium as described, except that the medium

142
contained 5% fetal bovine serum (FBS) rather than 10% SCS. Normal
foreskin-derived fibroblasts from a normal neonate were used as negative
control cells in focus assays. Twenty four hours after plating, the
medium was changed to medium identical to culture medium prepared as
described except that it contained 1% FBS. Cells were examined
microscopically each week for evidence of focus formation (a localized
piling up of cells resulting from continued cell division). Some foci
were picked and cultivated as individual cell strains. Other plates were

stained four to six weeks after plating.

Tuaorigenicity assay

The ability of cells to form tumors in mice was tested by injecting
1 x 10‘ to 10 x 10‘ exponentially growing cells (suspended in 0.2 ml Eagles
medium with no additives) subcutaneously in the scapular region of six
week old BALB/c athymic mice. In some instances cells were injected into
collagen sponges that had been implanted subcutaneously at the injection
site one week prior to injection of cells. The mice were examined weekly
for tumors. Small nodules that occurred were excised, fixed in formalin,
embedded in paraffin, sectioned, and stained with hematoxylin and eosin

prior to histological evaluation.

RESULTS

IN fibroblasts in culture
The fibroblasts that arose from biopsies of phenotypically normal
skin of this IM patient appeared similar to normal human fibroblasts in

cell cultureu Cultures were passaged serially as described and eventually

143
the rate of growth slowed down and the cells ceased to divide at about 35

population doublings.

Foci froa IM fibroblasts in 1% seru-

Focus assays were performed to determine whether fibroblasts derived
from phenotypically' normal skin ‘from 'this infantile' myofibromatosis
patient could form foci in low serum. Normal foreskin fibroblasts used as
controls did not form foci in 1% FBS. However, fibroblasts from the
infantile myofibromatosis patient consistently formed approximately two to
four foci per 100 mm culture dish (Fig. 1). These results were observed
in multiple culture dishes in two independent experiments. Microscopic
examination revealed that the foci were composed of piled up spindle-
shaped cells on a monolayer of spindle-shaped cells (Fig. 2). Cells were
isolated from foci and were serially passaged in standard culture medium

and were found to have a finite life span in culture.

IM fibroblasts fora benign tuners in athylic lice

Early passage infantile myofibromatosis fibroblasts cell lines or
focus-derived cell strains derived from infantile myofibromatosis
fibroblasts were injected in athymic mice to determine whether such cells
could form tumors in athymic mice. In the first of these tumorigenicity
experiments, IM fibroblasts were injected into collagen sponges that had
been implanted subcutaneously into two athymic mice one week prior to
injection of cells. We used these collagen sponges because of reports
that the sensitivity of the tumorigenicity assay is improved when cells
are injected into collagen or some other matrix. Each mouse was implanted
with two collagen sponges which resulted in a total of four injection

sites in this experiment. One month after injection small nodules

144

 

Figure 1. Focus formation of IM fibroblasts in 1% serum. (A) IM
fibroblasts demonstrating dark staining foci. (8) Normal human foreskin

fibroblasts that did not form foci.

  
    
      
  
 
 

: . ' ‘ I. V ‘l'1
‘. .“ xii/(- " 447/7 1,1,
3:» 1 -. . AMI/w
.. —’ "- r’ . , , ° ‘ A”
o’er; ' . ..iII/Z’W
’. 11¢"

3" h190“477%,1
.-,r'16;;92é

  

..oau
121°C\\:‘ ‘1'.
\3“

n. N).

‘

-
ax

\

-

    
    
     
  
 

» .- l .- .W'v '
. 1 \V-‘n-s‘.
‘11“ ~13 ‘\1f;*<§’-‘9*»:;

‘
\. .,,\\—G._fl
, v. \
11“. . .1.
1
. \ \

  
 
 

 
  

 
   
 
 

   

', 1 - inc-“71:;
_ 1 5' -. ‘2
.1 14;”;1 ~‘1r a
I Y 3’ {Q
\_
l

1 ‘ V54;
' Y.
° I ° % .\ 3"“,
1 A ‘44 A ’V'
1 ° 7"” 1*. I

1"
p

 
   

D
~ ‘ ‘11
. . - 4’ 1: \ 1
‘1. _\\.—’- 'x 1. ,1 '. r .59 ' ‘ 11
1. - .1, 11 «.1 - l
B 2 ~ " .. :1’ . , -.- ‘.' .JN‘rkx‘...
‘ *' ‘ - :1 - 1 1.,”). .-
‘ \

D

’ I“: - 223141311» -
1” " - nan
. .- [Thf’ m

Figure 2. Focus of cells that formed when IM fibroblasts that were plated

in medium containing 1% serum as described. (A) low power magnification
demonstrating focus on background of non-focus forming cells. (8) high

power magnification demonstrating dividing cells within focus.

146
persisted at the injection sites. One mouse was sacrificed and the
tumors were removed from both sponge sites. Histological examination
revealed what appeared to be neoplastic cells with remnants of collagen
present in the tissue, but the presence of the remnants of collagen made
histological diagnosis difficult. ‘The tumors in the1other mouse regressed
within about one week after the first mouse was sacrificed.

The ability of IM fibroblasts to form small tumors in athymic mice
was reassessed by injecting IM cell lines and focus-derived IM strains
into athymic mice. In this experiment, no sponges were used and the mice
were injected subcutaneously at only one site on each mouse (between the
scapula). Two focus-derived infantilei myofibromatosis strains were
injected into a total of five athymic mice. The two early passage primary
IM fibroblast lines were injected into three athymic mice each. We did
not detect formation of nodules in any of the mice that were injected with
focus-derived IM strains. Previously'we have determined that when normal
fibroblasts from several humans, including neonates, children, and1adults,
were injected into athymic mice, the swelling at the site of injection
caused by the cells and medium persists only for seven to ten days
(slightly longer when cells are injected into the collagen sponges) and
tumors are never detected either with or without sponges. These results
have been confirmed in more than 50 athymic mice (data not shown).
Nodules did not persist beyond two weeks in the mice that were injected
with focus-derived strains. However, small nodules persisted for three to
four weeks in three of five mice that were injected with the primary IM
fibroblasts cultures. One of the mice injected with a primary IM
fibroblast strain was sacrificed after 15 days. Microscopic examination
of a hematoxylin and eosin stained section from this nodule demonstrated

that the nodule was composed cells identical to the cells that were

147
present in the nodules that occurred when IM fibroblasts were injected
into sponges in athymic mice and identical to the cells of the tumor that

was removed from this IM patient shortly after birth (Fig. 3).

Discussion

Extensive experiments from this laboratory demonstrate that
fibroblasts from normal skin of neonates, children, or adults do not form
foci when assayed in 1% serum as described in this paper (J. McCormick,
unpublished studies). The fact that this occurs repeatedly at a high
frequency in fibroblasts obtained from apparently normal skin from a child
with IM strongly suggests that some fibroblasts of this individual have
abnormal growth characteristics. In our experiments we have found that
expression of various growth factor related oncogenes can be responsible
for such foci (see below). The fact that IM fibroblasts form fibromas in
athymic mice is a clear indication that the IM cells, or at least a sub-
p0pulation of them, are neoplastic.

Since only a fraction of the cells in these IM fibroblast
populations were capable of forming foci, i.e., approximately two to four
distinguishable foci formed per 5 x 10‘ cells, injection of 1 x 10‘ to 10
x 10‘5 cells into an athymic mouse would include 40 to 800 cells capable of
focus formation. Based on experiments of various frankly malignant human
mesenchymal cells studied in this laboratory (J. McCormick), it is very
unlikely that less than 1 x 10‘ neoplastic cells, even in the presence of
1 x 106 to 10 x 10‘ carrier cells, could give rise to a tumor within a
three month period. Therefore, it appears that non-focus forming 1M
fibroblasts that were injected along with the few cells capable of focus

formation, somehow contributed to the formation of the nodules that

148

Figure 3. (A) Hematoxylin and eosin stained section of a tumor removed
from IM patient shortly after patient was born. (8) Hematoxylin and eosin
stained section of a tumor that occurred in an athymic mouse as a result

of subcutaneous injection of IM fibroblasts as described.

149

 

150

occurred in athymic mice. One possible explanation is that the cells
capable of focus formation in low serum represent the cells that cause the
nodules in mice, but that these focus forming cells can recruit other
cells (non-neoplastic) in the formation of these small tumors.
Recruitment of non-neoplastic cells in a neoplastic role occurs, for
example, when neoplastic cells secrete a growth factor that causes non-
neoplastic cells in the tumor environment to proliferate.

Cells derived from foci gave rise to strains that did not have any
growth or morphological characteristics that distinguished them from the
IM fibroblast lines from which they were derived, although they could not
be extensively studied because of their short life span. The focus-
derived strains stopped dividing within a few population doublings after
being expanded to 10 x 10‘ to 20 x 10‘ cells and senesced in a typical
pattern (10). Since clonally derived populations senesce 10-15 population
doublings before bulk populations, this is not unexpected. This possibly
explains why the focus-derived cells did not form nodules in athymic mice.

A. number' of experiments have shown that neoplastic, but non-
malignant human cells can form benign tumors in athymic mice that
eventually regress. We have observed that cells established in culture
from benign human tumors, including several neurofibromas and a benign
myxoma, form small nodules in athymic mice with a histological appearance
typical of the tumor of origin (J.Wray, unpublished studies). These
regress if not removed. Sheela et al. (11) reported similar results with
human neurofibroma tissue. Yang et al. (15) demonstrated that when a
human fibroblast strain was transfected with a v~sis oncogene, clonal
populations expressing the v-sis protein formed fibromas in athymic mice

that frequently regressed.

151

Experiments from this laboratory designed to transform human
fibroblasts in vitro have generated cells with characteristics similar to
the IM fibroblasts in this experiment. When an mutationally activated,
over-expressed H-ras oncogene was transfected into normal diploid human
fibroblasts, transfected cells formed foci in low serum, but the focus-
derived cells failed to malignant tumors when they were injected into
athymic mice (6). Similar experiments in which other oncogenes including
v-sis, c-sis, (4,5), or N-ras (13) were transfected into normal human
diploid fibroblasts produced similar results. When these same genes were
transfected into an infinite life span human fibroblast cell strain, the
cells expressing these genes were malignantly transformed (7,14). These
results suggest that immortalization is a critical step in the
transformation process and that cells that do not have this property may
form lesions in athymic mice that fail to grow progressively and
eventually regress.

It will be of interest to examine the cells in culture from
myofibromas from this individual as well as others with myofibromatosis.
We predict that all or most of these cells will grow to high density
(i.e., like a focus) when in 1% serum, which is indicative of a defect in
a growth control pathway. The underlying defect in such individuals can
only be guessed at at this time. Since the underlying genetic defects
that predispose individuals to neoplasia have inevitably been shown to
encode a protein that functions as a suppressor gene (3), we assume that
an IM individuals inherit a defective suppressor gene which controls
fibroblast cell growth. In the course of normal development, the other
allele is inactivated by chance mutational processes in some fibroblasts.
Since background mutations rates in human cells are 1-100 per 107 cells per

generation, many cells with such a phenotype would already be present in

152
a full term IM infant. This could account for the tumors seen at birth or
later. However, if the situation were this simple, tumors would be
expected to be far more common than they are in IM individuals. We
postulate that IM fibroblasts that have lost the function of both copies
of a suppressor gene also need the activation of an oncogene to form a
fibroma. Such a model agrees better with studies carried out to model the

carcinogenesis process in human fibroblasts in this laboratory (9).

10.

LIST OF REFERENCES

BraCko, M., Cindro, C., and Rastko, G. Familial Occurrence of
infantile myofibromatosis. Cancer 1992; 69: 1294-1299.

Chung, E.B., and Enzinger, F.M. Infantile myofibromatosis. Cancer
1981; 48: 1807-1818.

Evans , H.J., and Prosser, J. Tumor-suppressor genes: cardinal
factors in inherited predisposition to human cancers. Envir. Health
Perspect 1992: 98: 25-37.

Fry, D.G., Milam, L.D., Maher, V.M., and McCormick, J.J.
Transformation of diploid human fibroblasts by DNA transfection with
the v-sis oncogene. J Cell Physiol 1986; 128: 313-321.

Fry, D.G., Hurlin, P.J., Maher, V.M., and McCormick, J.J.
Transformation of diploid human fibroblasts by transfection with the
v-sis, PDGF2/c-sis, or T24 H-ras genes. Mutat Res 1988; 199: 341-
351.

Hurlin, P.J., Fry, D.G., Haher, V.M., and McCormick, J.J.
Morphological transformation, focus formation, and anchorage
independence induced in diploid human fibroblasts by expression of
a transfected H-ras oncogene. Cancer Res 1987; 47: 5752-5757.

Hurlin, P.J., Maher, V.M., and McCormick, J.J. Malignant
transformation of human fibroblasts caused by expression of a
transfected T24 HRAS oncogene. Proc Natl Acad.Sci USA 1989; 86: 187-
191.

Jennings, T.A., Duray, P.H., Collins, F.S., Sabetta, J., and
Enzinger, F.M. Infantilelmyofibromatosis. Evidence for an autosomal
dominant disorder. Am. J. SUrg. Pathol. 1984; 8: 529-538.

McCormick, J.J. and Maher, V.M. Analysis of the1mulitstep process of
carcinogenesis using human fibroblasts. Risk Analysis 1994; 14: 257-
263.

Ryan, P.A., Maher, V.M., and McCormick, J.J. Failure of infinite
life span human cells from different immortality complementation
groups to yield finite life span hybrids. J Cell Physiol 1994; 159:
151-160.

153

11.

12.

13.

14.

15.

154

Sheela, S., Riccardi, V.M., and Ratner, N. Angiogenic and invasive
properties of neurofibroma Schwann cells. J Cell Biol 1990; 111:
645-653.

Venencie P.Y., Bigel, P., Desgruelles, C. Lortat-Jacob, S., Dufier
J.L., and Saurat, J.H. Infantile myofibromatosis. Report of two
cases in one family. British J. Derm. 1987; 117: 255-259.

Wilson, 0.M., Fry, D.G., Maher, V.M., and McCormick, J.J.
Transformation of diploid human fibroblasts by transfection of N-ras
oncogene. Carcinogenesis 1989; 10: 635-640.

Wilson, 0.M., Yang, E., Dillberger, J.E., Dietrich, S.E., Maher,
V.M., and McCormick, J.J. Malignant transformation of human
fibroblasts by a transfected N-ras oncogene. Cancer Res 1990; 50:
5587-5593.

Yang, 0., Kohler, S.K., Maher, V.M., and McCormick, J.J. v-sis
oncogene-induced transformation of human fibroblasts into cells
capable of forming benign tumors. Carcinogenesis 1994;(in press)

APPENDIX

APPENDIX

Establish-ant of cell lines froa hulan soft tissue tulors

As part of my research, 1 established several cell lines from benign and
malignant human soft tissue tumors. The table on the following page provides a
summary of the results of this work. Those cell lines that are listed in bold
letters are those that have been discussed in the text of this dissertation.
Some basic information on each of the tumors and the cell line that resulted is
listed in the columns. All of these cell lines were established in the
Carcinogenesis Laboratory at Michigan State University'with the exception of the
three neurofibrosarcoma cell lines that were provided from Dr. Thomas Glover of
the University of Michigan. Dr. Mark Zalupski, of Wayne State University,
provided the majority of the human tumor material with the exception of the
neurofibromas and the normal biopsies from the patient with infantile
myofibromatosis.

The information that is given in this table includes the age1and sex of the
patient from which each tumor was obtained. Fresh tumor material was implanted
into athymic mice in some instances and the results of these experiments are
listed as yes if a tumor occurred and no if a tumor did not occur. The next
column lists the results of experiments in which cells from culture were injected
into athymic mice. Finally, the last column lists the results of continuous

culture of each of the cell lines.

155

156
TUMOR CELL LINES

 

 

Cell Line Age of Sex of Tumor by Tumor by Population
Designation Patient Patient Implant Cell Line Doublings
m

USU-1 72 it yes yes 150 +

WSU-S 70 N yes no 30 ---> senesce
WSU-ll ro no 20 ---> senesce
WSU-13 not done no 35 ---> senesce
WSU-14 rot done no 35 ---> senesce
WSU-16 no no

13.1mm

WSU-7 72 M yes not done few --> senesce
1150-12 55 F yes no 30 ---> senesce
1150-15 64 M not done yes 35 ---> senesce
WSU-18 43 F no no 50 ---> senesce
N i

lrl—HSIH 44 N no no 150 +
irl—WSU-IO 17 F yes yes 200 +
il-‘l—liSU-17 33 F yes yes 150 +

Til-88.? not canine no I!) --> senesce
til-88.14’ not rknne no I) --> senesce
til-m8 not mine yes 110+

m

WSU-2 66 N no no few --> senesce
WSU-3 70 F yes yes 30 ---> senesce
WSU-B no no 30 ---> senesce
WSU-9 no no 40 ---> senesce
11.1mm

x; 32 F yes (represses) yes (rewesses) Ill --> senesce
inn-3

M-‘b—4

NFb-S

lib-6

lFb-7

lib—8

nth-9

lib-10

lFb—ll

lib-12’

W

TIL; 8 It not mine yes (messes) 50 --> senesce
1P.

 

‘ = Cell lines provided by Dr. T. Glover of The University of Michigan.
’ = NFb-l througn NFb-ll are from same 32 year old female with NF1.
’ = NF1 patient of unknown sex and age. Tumor provided by Dr. Mark Lebwohl, New York.

nIcwIan $1an UNIV. LIBRARIES
1111111111 lllllllllllll lllllllll
31293010152035