23‘?qu09

STATE UN VERSITY LIBPAWES

 

 

”‘T‘fifl l ”M. l: ‘ll‘ ll ll ‘5}‘1‘ W l‘ '
mm lulllmxlggllllllll .
Mkhigan State

University

 

 

This is to certify that the

dissertation entitled

MUTATIONAL ANALYSIS OF THE H-ras ONCOCENE IN SPONTANEOUS
B6C F MOUSE LIVER TUMORS AND TUMORS INDUCED WITH A
aENOTOXIC AND NONGENOTOXIC HEPATOCARCINOGENS

presented by

Tony R. Fox

has been accepted towards fulfillment
of the requirements for

. M
Ph D degree in icrobiology

sci/M ‘
as w

Major professor

 

 

Date

 

MSU is an Affirmative Action/Equal Opportunity Institution 042771

 

.4 _ ‘ --, f4 7, J-

PLACE IN RETURN BOX to remove thle checkout from your record.
TO AVOID FINES return on or before date due.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Adieu/Equal Opportunity lrltltution

 

MUTATIONAL ANALYSIS OF THE H-ms ONCOCENE IN SPONTANEOUS
B5C3F1 MOUSE LIVER TUMORS AND TUMORS INDUCED WITH A
GENOTOXIC AND NONGENOTOXIC HEPATOCARCINOGENS

By

Tony R. Fox

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Microbiology

c:_‘

LOOSGIEI

ABSTRACT

MUTATIONAL ANALYSIS OF THE H-ras ONCOCENE IN SPONTANEOUS
BsC3F1 MOUSE LIVER TUMORS AND TUMORS INDUCED WITH A
GENOTOXIC AND N ONGENOTOXIC HEPATOCARCINOGENS

BY

Tony R. Fox

The frequency and mutational profile of H-ras gene activation was
determined in spontaneous liver tumors of male BéC3F1 mice and in tumors
induced with the genotoxic hepatocarcinogen benzidine-ZHCI or the
nongenotoxic hepatocarcinogens phenobarbital, chloroform or ciprofibrate.
DNA sequence analysis of the H-ras gene from representative tumors
revealed that 32 of 50 (64%) spontaneous tumors and 13 of 22 (59%)
benzidine-ZHCl induced tumors contained a point mutation in codon 61.
Tumors induced with the nongenotoxic agents had a much lower frequency
of codon 61 mutations, i.e., phenobarbital 1 of 15 (7%), chloroform 5 of 24
(21%) and ciprofibrate 8 of 39 (21 %). No mutations were observed at codons
12, 13 or 117 in tumors from any of the groups. Only three base pair

substitutions within codon 61 were found. The one most frequently detected

in all of the groups was a CC to AT transversion at the lst nucleotide
position, occurring at a 59%, 85%, 100%, 80%, and 88%, frequency in the
spontaneous tumors and in the tumors induced with benzidine 2-HC1,
phenobarbital, chloroform and ciprofibrate respectively. In these same groups
a A-T to CC transition or a A-T to T-A transversion at the second nucleotide
position occurred at a frequency of 34%, 8%, 0%, 0%, 12% and 6%, 8%, 0%,
20%, 0% respectively. The number of tumors carrying an activated H-ras gene
in the nongenotoxic treatment groups is within the range that would be
expected if those animals had not received any treatment. This indicates that
the activation of the H-ras gene in those tumors is probably the result of a
spontaneous event. The data suggest that these toxicologically and
pharmacologically diverse nongenotoxic hepatocarcinogens increase the
frequency of liver tumors but do not induce mutations in the H—ras gene.
Instead these agents appear to interact with a population of cells that do not
contain an activated H-ras gene. This suggests that the mechanisms of tumor
development by these nongenotoxic carcinogens differs at least partially from
the mechanisms responsible for the development of spontaneous tumors or

those induced by a typical genotoxic agent.

Dedicated

to the memory of my father, Robert L. Fox, who recognized and
kindled my natural curiosity of the unknown and showed me how
knowledge could provide the means to satisfy it; and to my mother,

Doris L. Fox, for her love, guidance and support over the years.

Acknowledgements

I wish to express my appreciation to Dr. I. Justin McCormick for serving
as the director of this research and also thank Dr. Veronica M. Maher for her
encouragement and support. I would also like to thank the other members
of my graduate committee: Dr. Alan Schumann, Dr. Susan Conrad and Dr.
Leeland Velicer for their invaluable time.

A special debt of gratitude is extended to Dr. Iia-ling Yang and Dr. Peggy
Schott for their technical advice on DNA sequencing and DNA
oligonucleotide synthesis respectively. I would also like to thank Sherry
Pagels for her excellent technical assistance.

I am especially grateful to Dr. Philip Watanabe for providing me with a
leave of absence from my job so that I could retum to the university in order
to fulfill the requirements for this degree and to The Dow Chemical Company
for their financial support. I would also like to thank Mrs. Gywen Knight for
her gracious hospitality while I lived in her home during my stay in East
Lansing.

I owe an unrepayable debt of appreciation to my wife, Terra D. Fox, for
here love, patience, understanding and enduring support. This was
especially important during the year and a half that we lived apart while I

stayed in East Lansing and she maintained our home in Midland.

iii

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LIST OF TABLES vi
LIST OF FIGURES vii
INTRODUCTION 1
CHAPTER I. LITERATURE REVIEW 7
A. The Process of Carcinogenesis 7
1, Initiation 7
2. Promotion 9
3. Progression 12
4. Properties of benign and malignant mmm-c 1';
5. Tumor metastasis 18
6. Tumor heterogeneity 19
B, Chemical Cardnngpnmie 77
1. Genotoxic carcinogens 73
2. Nongenotoxic carcinngpne 75
C. The Role of Oncogenes and Suppressor Genes in Carcim o ’79
1. Retroviral oncogenesis 30
a. The acutely transforming retroviruses 30
b. The latent transforming retroviruses 34
2. DNA Tumor Viruses 36
a. Oncogenic Papovavirnsns 36
b. Oncogenic Adenovirnses 40
C. Oncogenic Herpesvirucec 41
d. DNA tumor viruses without oncogenes ..................... 42
3. Cellular Oncogenes in Carcin- o ' 44
a. Activation by point mutation 45
b. Activation by gene rearrangement 46
c. Activation through gene amplification 48
4. Oncogenes that Act as Suppressors to Neoplastic Growth ................. 50
a. The retinoblastoma paradigm an
b. Other possible tumor suppressor genes 52
D. Oncogene Protein Products and their Role in Carcinogenesis ..................... 54
1. Extracellular oncogene proteins 55
2. Oncogene protein products as growth factor receptors ........................ 56
3. Oncogene proteins with phosphokinase activity ................... 58

4. Oncogene proteins located within the nucleus

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E. The ras Genes and Their Protein Products 63
1. The molecular biology of the ras genes 63
2. The ras protein structure 67
3. Posttranslational modification of the ras proteins 71
4. The biochemical activity of the ras proteins 73
F. Chemical Carcinogenic Risk A uent 76
1. Methods used in risk ‘ 76
a.) Epidemiological analysis in risk assessment ............................ 77
b.) Short-term tests for the evaluation of chemical
carcinogenic potential 80
c.) The animal bioassay 84
i.) Basic conduct 95
ii) Problems with interpretation 85
G. Oncogene Activation in B6C3F1 Mouse Liver Tumors 90
1.) Activated oncogenes in spontaneous liver tumors 90
2.) Activation of cellular oncogenes in chemically induced
36C3F1 mouse liver tumors 94
Reference: 101
CHAPTER II. MUTATIONAL ANALYSIS OF THE H-ras ONCOCENE
IN SPONTANEOUS 35C3F1 MOUSE LIVER TUMORS
AND TUMORS INDUCED WITH A GENOTOXIC AND
NONGENOTOXIC HEPATOCARCINOGENS ........................... 12
ABSTRACT 12%
INTRODUCTION 124
MATERIALS AND METHODS 127
Tumor Induction 127
Oligonucleotide Primer Preparation. 127
PCR Amplification. 128
Sequencing 129
RESULTS 130
Tumor Induction 13o
H-ras Gene Activation 1 '31
DISCUSSION 133
ACKNOWLEDGEMENTS 140
REFERENCES 141
CHAPTER III. A MODEL FOR 85C3F1 MOUSE LIVER TUMOR
DEVELOPMENT: IMPLICATIONS FOR RISK
ASSESSMENT 1 '53
APPENDIX 162

 

LIST OF TABLES

W

 

 

TABLE 1. Factors effecting cancer incidences 7
TABLE 2. Selected oncogenes: function and cellular localization ...................... 33
TABLE 3. Transforming genes in spontaneous hepatocellular tumors of the
B(,C3F1 mouse 91
TABLE 4. Oncogene activation in B6C3F1 mouse liver tumors ......................... 93

TABLE 5. Mutation spectrum in codon 61 (CAA) of the H-ras gene in B(,C3F1
mouse liver tumors 9:;

TABLE 6. Distribution of activating mutations within the H-ras gene from
86C3F1 mouse liver tumors 98

 

W

TABLE 1. Incidence of Spontaneous and chemically induced liver tumors in
male 136C3F1 mice 149

 

TABLE 2. Frequency of activation of the H-ras gene at codon 61 in male
36C3F1 mouse liver tumors 1'50

 

TABLE 3. Frequency of H-ras gene activation in hepatocellular adenomas and
carcinomas in male B6C3F1 mice 151

 

TABLE 4. Mutation spectrum in codon 61 (CAA) of the H-ras gene in male
B6C3F1 mouse liver tumors 1‘52

 

LIST OF FIGURES

W
FIGURE 1. Schematic representation of the carcinogenic process ........................ 8

FIGURE 2. Structural and functional domains of the mammalian ras gene...69

W

FIGURE 1. Nucleotide sequence of the coding strand for the H-ras gene which
was PCR amplified and sequenced 147

FIGURE 2. Representative DNA sequence of the H-ras gene codon 61 region
derived from PCR amplified liver tumor DNA .............................. 148

W

FIGURE 1. Representative chromatogram of HPLC purification for DNA
oligonucleotide primers 166

FIGURE 2. Agarose electrophoretic gel analysis demonstrating the utilility of
Bam H1 digestion on increasing PCR amplification efficiency ....168

FIGURE 3. Agarose electrophoretic gel analysis from a representative PCR
amplification of a 110 bp region within the H-ras gene from liver
tumor tissue 169

FIGURE 4. Representative DNA sequence analysis of the H-ras gene codon
12, 13 and codon 117 regions 170

vii

INTRODUCTION

Epidemiological studies suggest that the majority of human cancers are
caused by environmental factors (Doll and Peto, 1981). Table 1 highlights a
number of general factors responsible for the development of cancer in
humans. A closer examination of this list reveals that these general factors
represent variations of only three specific etiologic agents,- radiation, viruses
and chemicals. Of these three, over 75% of induced human cancers are
caused by the exposure to chemicals (the majority coming from natural
sources). Because such a high percentage of human cancers are the result of
chemical exposure, extensive efforts have been put forth by the scientific
community to understand the mechanisms of chemical carcinogenesis and to
develop experimental methodologies for the evaluation of a chemical's
carcinogenic potential.

Methodologies currently used to evaluate a chemicals carcinogenic
potential include human epidemiological analysis, in vitro and in viva
mutagenicity assays and the animal bioassay for carcinogenicity. Although
data from all of these sources are used for chemical carcinogenic risk
assessment, the results from the animal bioassay have the most significant
influence on this decision making process. The development of tumors in
these animals is not however definitive evidence that a chemical is a human
carcinogen. There are a number of factors which complicate the extrapolation

of bioassay results in determining potential human risk.

Table 1. Factors effecting cancer incidences 1

Percent of all cancer deaths

 

Factor or class of factors Best estimate Range of
acceptable
estimates

Diet 35 10 - 70
Tobacco 30 25 -40
Reproductive and sexual behavior 7 1 - 13
Occupation 4 2 - 8
Alcohol 3 2 - 4
Geophysical factors 3 2 — 4
Pollution 2 < 1 - 5
Medicines and medical procedures 1 0.5 - 2
Food additives < 1 -53 _ 2
Industrial products <1 < 1 - 2
Infection 10 ? 1 - ?

1 From Doll and Peto, 1981

These complicating factors include such issues as differences in metabolism
and pharmacodynamics between the test species and man, inherent problems
with the bioassay design, and the occurrence of tissue specific high
spontaneous tumor incidences within the inbred rodent strains.

The issue of high spontaneous tumor incidences has been of particular
concern with the B§C3F1 mouse. This mouse is used extensively in many
carcinogenicity bioassay programs, including those conducted by the National
Toxicology Program. The male 86C3F1 mouse has a characteristically high
spontaneous liver tumor incidence. Increases in B6C3F1 mouse hepatic
tumors are frequently observed in many bioassays after administration of
either genotoxic or nongenotoxic chemicals. Evaluating the significance of an
increase in these liver tumors to potential human risk after exposure to
nongenotoxic agents has been a very controversial issue of the current
bioassay program (Interdisciplinary Panel on Carcinogenicity, 1984; Nutrition
Foundation, 1983; Task Force of Past Presidents- SOT, 1982). Greater
confidence in the interpretation of bioassay results using this strain of mouse
requires a better understanding of how these spontaneous tumors arise and
how various chemicals enhance this tumor incidence.

Efforts to understand the molecular mechanisms of carcinogenesis
have been aided significantly by the recent detection of activated cellular
oncogenes (Shilo et al., 1981). These genes appear to be involved in the
control of cellular proliferation and of differentiation. These oncogenes have
been found in a number of human and experimentally induced animal

tumors.

Recently, activated oncogenes have been identified in spontaneous
liver tumors of the B(,C3F1 mouse and in tumors induced with several
genotoxic chemicals (Fox et al., 1987; Reynolds et al., 1986, 1987; Wiseman et
al., 1986). The most frequently detected oncogene in these tumors was the
activated H-ras gene. Mutational analysis of a limited number of
spontaneous tumors and tumors induced with several genotoxic carcinogens
has revealed specific mutational spectra within the H-ras gene.

One important parameter in evaluating a chemicals' carcinogenic
potential is determining whether the chemical has in vivo genotoxic activity.
The existence of mutational specificity in the H-ras gene suggests that a
comparitive analysis between spontaneous tumors and tumors induced with
various carcinogenic agents might be useful for making this type of
evaluation. Such analysis would be particularly valuable for determining
whether nongenotoxic hepatocarcinogens have any cryptic in vivo genotoxic
activity since the current assignment of a nongenotoxic status is based
primarily on in vitro mutagenicity assays. It may also provide important
mechanistic information into how this class of carcinogens cause liver
tumors in the B6C3F1 mouse. Information of this type could be very useful in
the interpretation of bioassay results with this strain of mouse. This
ultimately should provide for a more accurate evaluation of potential human
risk from exposure to these chemicals.

To address these issues, the objectives of this thesis are to (1) provide
additional data on the frequency of H-ras gene activation in spontaneous

B(,C3F1 mouse liver tumors (2) evaluate the mutational spectra of the H-ras

gene in spontaneous liver tumors (3) determine the frequency of H-ras gene
activation in tumors induced with several functionally diverse nongenotoxic
hepatocarcinogens and (4) evaluate the mutational spectra in the hepatic
tumors induced with these nongenotoxic agents.

Chapter I provides a review of the scientific literature relevant to the
proposed objectives. It includes discussions of the general process of
carcinogenesis, chemical carcinogenesis, oncogenes and suppressor genes, a
comprehensive review of the ras genes, chemical carcinogenic risk
assessment and oncogene activation in the BGC3F1 mouse liver.

Chapter II consists of an original research paper excepted for
publication in the journal Cancer Research. Within this paper the objectives
of the current thesis are addressed. The design and execution of the
experimental work, including the development of a modification in the
polymerase chain reaction (PCR) technique for amplifying designated regions
of the H-ras gene from liver genomic DNA, presented in this paper was
conducted by myself with the guidance of Dr. I. Justin McCormick.
Histopathological evaluation of the liver tumors was conducted by Dr. Barry
Yano. Instructional assistance on the synthesis of DNA oligonucleotide
primers and modification in the DNA sequencing of PCR amplified H-ras
genes was provided by Dr. Peggy Schott and Dr. Iia-ling Yang respectively.

Chapter III presents a discussion of a theoretical model for the

development of liver tumors in the 86C3F1 mouse.

An Appendix has been included to provide additional information
regarding specific methodological material that was not discussed in the

Cancer Research manuscript.

W

AIheEmcessnLCarsinngenesis

Cancer is a complex, dynamic disease which has been observed to occur
in nearly every tissue of the body. It is a disease characterized by abnormal
cellular proliferation and differentiation. These perturbations of growth and
development give rise to a tumor (except in cancers of the reticuloendothelial
system) which is the end result of a series of cellular changes that may have
taken many years to develop. Although these changes occur as a continuous
event, carcinogenesis can functionally be divided into three distinct stages,
designated initiation, promotion and progression (Figure 1). Much of our
understanding about the events occurring during these stages of
carcinogenesis has been obtained from the development of experimental
animal models using radiation and chemicals to induce specific cancers (Pitot,

1986).

1. I 'l' l'
Initiation is considered to be a primary, irreversible change which
predisposes the cell to further neoplastic development (Boutwell, 1964). The
most likely site for this irreversible event is the genetic material (DNA) where
the induction of mutations is thought to be a key step. This conclusion is
based on the large amount of data associating the induction of mutations with

the development of cancer.

8:238 88:80.22 . 32898. 88:88.8 .

$9.: 0588.an . maotgafiogafi . 30.299: .mccoc .
29992: . c.9955. . =8 29% .
5:205 SEE . :meoaxo econ 0923 . engage: .
c9.mu__nmficv 260.com . coacaaxo =8 . 89:3 <20 .
\ 553565 T 228.55 T cozaEc. / «45>
8E3 E2982 :8 .232 Arlcozmfimm
”$2820

$305 2:32.830 05 Lo 5:25.332 czancom ._. 959”.

Experimental evidence indicates that these initiating mutations may
occur spontaneously or they can be induced by exposure to chemicals
(naturally occurring and synthetic), radiation, and viruses. For many years the
genetic target for these mutational events was unknown. However, over the
past ten years considerable data have been generated implicating two groups of
cellular genes as the likely targets. These two groups are represented by a set of
dominant and recessive genes referred to as oncogenes and suppressor genes
respectively.

Druckey (1967) used the initiating carcinogen 4-dimethylaminostilbene
to investigate the relationship between dose of initiator and tumor
development in the rat. He observed a linear relationship between these two
parameters with the line passing thru the origin. This study, along with data
from many other studies where chemicals and radiation have been used as
initiating agents, suggests that initiation is a "one hit, no threshold" process
(Smets, 1980). Consequently, multiple divided doses of the initiator achieves
the same result as a comparable single dose.

Initiation alone is not sufficient for tumor development. The initiated
cell may remain dormant indefinitely or grow very slowly. These
"transformed" cells remain latent until they are acted on by some promoting

event.

2. Protection
Promotion is a complex process thought to involve several steps, some

irreversible and some reversible. Experimental evidence indicates that

10

promotion does not involve mutational changes in the genetic material
(Pitot, 1980). Instead promotion involves an increase in cellular proliferation
and the alteration of gene expression (Angel et al., 1986; Krieg et al., 1988).

The end result of this activity is the expansion of the initiated cell population.

Cellular proliferation appears to be a necessary but not sufficient aspect
of promotion. The requirement for cell division probably involves the
fixation of the initiating event. The sooner the proliferative stimulus occurs
after initiation the greater the probability that a neoplasm will develop. This
is a reasonable observation if initiation is a mutagenic process because a delay
of cell division would allow this damage to be repaired eliminating the
initiating event.

Promotional proliferation has been experimentally induced using both
physical and chemical methods. In the rodent liver the rapid regeneration
following surgical partial hepatectomy has been shown to induce tumors in
the chemically initiated rat liver where chemical treatment or partial
hepatectomy alone will not cause tumor development ( Pitot, 1980a).

Chemical agents used to promote rodent skin and liver tumors also
induce cell division (O'Connell et al., 1987). The increased cellular
proliferation by these agents may be the result of a degenerative-regenerative
response due to the cytotoxic effects of the chemical or to a specific
pharmacologic stimulus inducing the activation of genes involved in cellular
division. These agents include a diverse group of materials such as

hormones, drugs, plant products and industrial chemicals.

11

Chemical tumor promoters can be divided into specific and nonspecific
classes. The specific promoters alter gene expression by forming complexes
with unique receptors or receptor-like molecules. These receptor-ligand
complexes can activate a series of biochemical reactions leading to changes in
transcriptional activity. Examples of this class of promoter include steroid or
polypeptide hormones (Pitot et al., 1980a) and the potent tumor promoter
TCDD (Pitot et al., 1980b; Poland et al., 1975). Nonspecific promoters do not
interact with specific receptors but induce changes in gene expression by
indirect mechanisms that are not clearly understood.

Experimental studies have demonstrated that chemical promoters, like
chemical initiators, exhibit a dose response relationship. However unlike
initiating agents, tumor promoters appear to have a distinct threshold
concentration below which tumors will not occur and a maximum
concentration where a further increase in tumor development is not
observed. This concept has been demonstrated using either a mouse skin or
rat liver model with TPA or phenobarbital as the promoting agent
respectively (Verma and Boutwell, 1980; Peraino et al., 1980). Pitot (1982) and
his associates, using quantitative measurements of enzyme altered foci as a
parameter to study chemical promotion in the liver, showed that a
maximum number of foci occurred at doses above 0.01% phenobarbital in the
diet. At the low dose of 0.0001% no difference in the number of foci was
detected between the treatment and the control groups. Studies such as these
clearly exemplify the distinct differences between chemical initiation and

promotion. This concept is not only important to an academic understanding

12

of the process of carcinogenesis but also has practical implications for

chemical risk assessment.

3- mm

After the promotional phase of neoplastic development the growing
tumor consists primarily of a clone of cells which morphologically are very
similar to the surrounding normal cells. These early tumors do however
exhibit biochemical differences that can be exploited to distinguish them from
the surrounding background of normal cells. Using various histological
staining techniques they can be identified microscopically as areas of altered
focal development (Pitot, 1980). These developing tumors are pathologically
classified as benign. During the period of progression these benign tumors
become more malignant in character. This involves a process of tumor
evolution where the sequential selection of variant sub-populations occurs
which exhibit progressive phenotypic changes that are of a more neoplastic
character.

One of these characteristics is a tendency for the developing tumor to
exhibit an increasing growth rate. This does not represent a shortening of the
cell cycle but rather involves an increase in the proportion of cells which are
actively dividing instead of progressing towards terminal differentiation
(Ford, 1986). In some cases this release from normal growth control may
involve an altered response to growth factors through the loss of specific
receptors or by an autocrine stimulation of growth caused by the unregulated

overproduction of a growth factor (Lippman, 1986).

13

As the tumor becomes more malignant a decrease or disappearance of
cellular organelles and metabolic functions necessary for specialized activity is
observed within cells of the evolving tumor (Croce, 1985). At the same time
significant increases in the synthesis of specific proteins may occur which
contributes to the ability of the tumor to invade or metastasize. These
proteins can include proteolytic enzymes, plasminogen activators,
fibronectin, and tumor angiogenic factors (Nicolson, 1984). Decreased
antigenicity and the acquisition of drug resistance may also appear within the
advanced neoplastic cell population which could provide those cells with a
selective growth advantage (Nowell,1986).

One of the more pronounced changes occuring during progression is
an increase in the frequency of cells with chromosomal abnormalities.
Histological studies of isolated tumor masses have demonstrated that tumor
cell populations become more genetically unstable with increasing time
(Oksala et al., 1979; Ling et al., 1985). These cells show higher numbers of
chromosomal breaks, nondisjunctions, increased sister chromatid exchanges
and changes in ploidy. Some limited data also suggest that the rate of these
chromosomal changes increases with tumor progression (Ling et al., 1985).

These chromosomal changes are probably not occuring randomly.
Evidence has been accumulating suggesting that specific chromosomal
changes are associated with particular malignancies. The most striking
examples of this can be found in several human leukemias. For example in
Burkitt's lymphoma, an aggressive Bocell neoplasm, approximately 75% of

the patients exhibit a chromosomal translocation in which the distal part of

14

chromosome 8 is translocated to the long arm of chromosome 14 (Manolova
and Manolova, 1972). Variant translocations involving chromosome 8 with
chromosome 22 and chromosome 2 are found in the remaining 25% of these
patients (16% and 9% respectively) (van den Berghe et al., 1986). Most cases of
chronic myelogenous leukemia (CML) involve a translocation between
chromosome 9 and chromosome 22 producing a marker chromosome which
has been named the Philadelphia chromosome (Nowell, 1960). Recently the
myc and abl oncogenes have been located on the fragments of chromosomes 8
and 9 involved in the translocations of Burkitt's lymphoma and CML
respectively (Dalla Favera, 1982; Groffen, 1984). Alterations in the expression
or in the protein coding region of these genes as a result of the translocations
is thought to play an important function in generating these specific
malignancies.

Single gene mutations could initiate the chromosomal aberrations
observed during tumor progression. Defects in DNA repair enzymes or
abnormalities resulting in more error prone DNA synthesis or mutations in
mitotic spindle proteins could precipitate some of the changes described
above. Once a single chromosomal abnormality occurs it in itself may
contribute to further genomic instabilities.

As cells in the developing tumor begin to accumulate increasing
numbers of genetic changes, eventually a cell(s) may emerge which has
acquired a completely malignant phenotype. As this cell proliferates its
progeny begin to invade the surrounding tissue. Some of these cells may find

their way into the vascular or lymphatic systems where they can be

15

disseminated throughout the body giving rise to metastatic tumors leading to
the ultimate death of the organism.

The multi-stage cancer model of initiation, promotion and progression
serves as a paradigm to recognize the functional importance of the major
critical events in tumorigenesis. This model depicts tumor development as a
series» of distinct steps occuring temporally one after the other. It is important
to remember that the formation of a tumor from a single initiated cell to the
appearance of a malignant mass is a continuous process representing many
biological changes that may overlap each other. In spite of these limitations,
the multi-stage cancer model has made an important contribution in

developing our current understanding of the cancer process.

4. W

The classification of a tumor as benign or malignant is based on a
subjective criterion incorporating specific growth and morphological
characteristics. If one uses these criteria most tumors fall into one of these
two categories. However, there are some borderline cases which cannot be
categorized with certainty. All tumors, benign or malignant, have two basic
things in common; (1) they consist of proliferating neoplastic cells of
parenchymal origin and (2) have a supporting stroma consisting of
connective tissue, blood vessels and possibly lymphatics (Cotran et al., 1989).
Beside these two common parameters, differences distinguishing benign and
malignant tumors exist in the level of differentiation, rate of growth and type
of growth.

16

In general, cells comprising a benign tumor are well differentiated.
Cytologically the tumor cells do not differ substantially from the surrounding
normal cells within the organ. In many cases they may present themselves as
outgrowths of the normal tissue. For example, bone tumors may produce
nodules that are indistinguishable from normal bone. Benign epithelial
tumors of the skin form local growths which contain all of the components of
normal skin but are compacted to form a solid nodule (Franks, 1986). An
example of this type of tumor is the common wart. As the growing tumor
expands it may compress and damage the normal tissue. This can lead to the
severe loss of organ function. Depending on the particular organ and the
extent of damage some benign tumors, although rarely, can be fatal.

On the other hand malignant tumors are composed primarily of
undifferentiated cells. It is thought that they may originate from the pool of
undifferentiated stem cells found in all specialized tissues (Cotran et al., 1989).
The cells within the tumor characteristically display morphological
pleornorphisms. Cells much larger and much smaller than normal can be
found. These cells can be multi—nucleated or contain a single giant
polymorphic nucleus. Large numbers of cells in mitosis are frequently
observed indicative of the increased proliferative activity associated with
malignancy. In general, the malignant tumor shows little resemblance to the
surrounding tissue.

The growth characteristics of the cells within a tumor are another
important parameter used to distinguish a benign tumor from malignant

one. Most benign tumors grow slowly over a period of years at a steady pace.

17

They may however enter periods of dormancy where they do not grow at all
or even decrease in size. Factors such as hormone dependency and blood
supply can influence this growth rate ( Cotran et al., 1989). Growth of a
benign tumor is nearly always confined within a fibrous capsule. This
contained, localized growth is a key factor distinguishing benign from
malignant neoplasms. However, the lack of a capsule does not imply that the
tumor is malignant.

The growth rate of cells within a malignant tumor is usually rapid and
sometimes erratic. The pattern of growth is much more unorganized than
that found in a benign tumor. Malignant tumors have the capacity to invade
the nearby surrounding tissue, although carcinomas in situ can still be
diagnosed as malignant based on cellular characteristics.

The unequivocal diagnosis of malignancy is the detection of a
metastasis. A metastasis is a secondary tumor arising from the primary
tumor at some other location in the body. Metastases can occur via several
different mechanisms (1) embolization through the blood vessels (2)
lymphatic permeation and (3) direct seeding of the body through open
cavities (e.g. peritoneal cavity) (Cotran et al., 1989)). The properties necessary
for tumor cells to metastasize are most likely acquired during the period of

tumor progression.

18

5. Wasted:

The ability of a tumor to invade and metastasize ultimately results in
the death of the cancer patient. It makes treatment difficult and if metastasis
could be eliminated most cancers would be curable by surgery.

Tumor metastasis is a complex, dynamic process involving a number
of specific sequential events resulting in the dissemination of the primary
tumor to various organs around the body. Following tumor growth and
development, usually during the period of progression, metastasis begins
with the local invasion and infiltration of tumor cells into the surrounding
normal tissue (Cotran et al., 1989). Some of these cells penetrate the small
vessels of the lymphatic or vascular system.

For a metastatic cancer cell to enter the lymphatic or vascular vessels it
must first traverse the extracellular matrix (ECM). This matrix is composed of
a highly cross-linked meshwork of type IV collagen, specific glycoproteins
such as laminin, and several types of proteoglycans (Cotran et al., 1989).
Studies have shown that the invading cells actively secrete enzymes which
breakdown the ECM components (Pauli and Lee, 1988).

Once the tumor cells have penetrated the vasculature, individual cells
or clumps of cells are released into the circulation (Kaminski et al., 1988). The
circulatory system is a hostile environment for the cancer cell and many cells
do not survive. This is the result of several factors which include
unfavorable growth conditions and the action of immunological factors
which tend to control the spread of the tumor cells. Once into the circulation

the tumor cells tend to aggregate with themselves and with platelets to form

19

small emboli. The aggregation of the tumor cells with the platelets may
increase survival by protecting them from immunological attack.

Some of the circulating emboli eventually come to rest within the capillary
beds of various organs. Here they exit the vasculature and begin growing as a
secondary tumor completing the metastatic process.

Studies have shown that some primary tumors form metastases
within specific tissues (Liotta, 1988). It is not entirely clear what factors
influence where these secondary tumors will arise. This selectivity may in
part be a function of the anatomical location of the primary tumor. The
likelihood of whether circulating tumor cells find their way to a particular
organ may also be influenced by the patterns of blood flow from the primary
tumor site. These natural patterns of blood flow cannot entirely explain
metastatic specificity. This is exemplified by the fact that muscle tissue and
the kidneys are both well vascularized yet these organs seldom are the sites
for metastases (Cotran et al., 1989). Cell surface factors, most likely involving
common membrane receptors between the tumor cells and the cells where
the metastases occur, probably play a more important role (Pauli and Lee,

1988).

l 6. W
As discussed previously, the stage of tumor progression is characterized
as a time when the cells of the developing tumor become increasingly
heterogeneous. A number of sub-populations will arise within the

developing tumor which have their own characteristic phenotype. This

20

heterogeneity involves an extensive number of phenotypic changes
including cellular morphology, tumor histology, karyotype, growth rate, cell
products, receptors, immunological characteristics, metastatic ability and
sensitivity to therapeutic agents (Heppner, 1984). Much of the experimental
work attempting to explain the source of this heterogeneity has focused on
the inherent genetic instability of the transformed cell. The work of Cifone
and Fidler (1981) demonstrated that in fibrosarcoma cells the rate of
spontaneous mutation was greater in metastatic cells than in nonmetastatic
subpopulations. Mutations such as genomic rearrangements, chromosomal
losses, gene amplification and base pair substitutions are involved in the
increased phenotypic varibility.

It is important to recognize that heterogeneity is not a process restricted
to tumor cells. Phenotypic variability also occurs in normal tissues.
Mechanisms generating tumor heterogeneity may therefore be similar to
those involved in the development of normal tissues. This is exemplified in
the observation that normal cells differ in their susceptibility to various
carcinogens and oncogenic viruses. For example, the heterogeneity observed
in different SV40-transformed clones reflects the inherent heterogeneity
associated within the normal cell population prior to transformation
(Kakunaga, 1980; Omar, 1983).

. This normal variability probably arises as a consequence of epigenetic
events involving changes in gene expression. Evidence implicating the
importance of epigenetic processes is provided by Harris et al. (1982) and

Peterson et al. (1983) where they reported the rates of variation / generation to

21

be 10'2 t010'5 which is several orders of magnitude greater than the rates of
mutation (10'6 to 10'8).

Unequivocal proof that tumor heterogeneity is a real phenomenon of
tumorigenesis and not an artifactual observation is still unavailable. The
concept of heterogeneity is however generally accepted by most pathologists
and minor biologists. One corollary of the tumor heterogeneity hypothesis,
which is far more controversial, states that metastases are derived from one
or more of the heterogeneous sub-populations within a single primary
tumor. Recent work by Kerbal and his associates has provided convincing
evidence that metastases are derived from variant sub-populations within a
single tumor (Wanghorne et al., 1988). These investigators generated a
number of transformed clones from single tumors each carrying a selectable
marker on a different DNA restriction fragment. Injecting these cells into
animals they were able to generate primary tumors that eventually gave rise
to tumor metastases. By comparing the restriction fragments of the marker
gene in the primary and metastatic outgrowths they could trace the cell
lineage of the secondary tumors. Using this approach they demonstrate that
the metastatic tumors carried the marker gene on the same restriction
fragment as was found in the primary tumor. Unexpectedly, it was also
observed that the metastatic cells eventually achieved a dominant status

within the heterogeneous cell population of the developing tumor.

EQemicaLCm'mcgrnesis

One of the earliest indications that chemicals can be carcinogenic comes
from the now classical studies of Percival Potts, the eighteenth century
physician, who made the astute observation that the increased incidence of
scrotal cancer in chimney sweeps was associated with exposure to soot. Early
in this century two Japanese workers demonstrated that coal tar in soot could
induce tumors when directly applied to the skin of rabbits (Yamagiwa and
Ichikawa, 1918). Later Kennaway chemically fractionated coal tar and showed
that the pure components dibenz[a,h]anthracene and benzo[a]pyrene had
carcinogenic potential (Williams and Weisburger, 1986).

During the late 1800's the German physician Rehn observed that
workers in the dye industry had an abnormally high incidence of bladder
cancers. It was not, however, until nearly 50 years later that Hueper
demonstrated that 2—naphthylamine was the causative agent by reproducing
these lesions in dogs (Williams and Weisburger, 1986).

Since these early observations several hundred chemicals have been
identified as having carcinogenic activity. Many of these carcinogens are
synthetic and include the general classes of chemicals such as the polycyclic
aromatic hydrocarbons, heterocyclic amines, nitro alkyls, halogenated
hydrocarbons, aldehydes, carbamates, and N-nitroso compounds (Williams
and Weisburger, 1986).

In addition to these synthetic chemicals, a number of natural materials

are potent human and/ or animal carcinogens. These carcinogens can be

found occuring naturally in our food or as a result of cooking. For example
the plant product cycasin found in certain foods eaten by peoples in the Pacific
islands was found to have potent carcinogenic activity (Williams and
Weisburger, 1986). Raw mushrooms contain various hydrazines and certain
bracken ferns have been shown to cause cancer of the alimentary tract
(Wigley, 1986). Other naturally occurring carcinogens can be formed as a
result of the metabolizing activity of micro-organisms. Aflatoxin Bl, a
mycotoxin produced by a strain of Aspergillus flavus, is one of the most
potent known carcinogens and has been implicated in the high incidence of
liver cancer present in many third world countries. (Williams and
Weisburger, 1986).

Once sufficient evidence had been accumulated implicating chemicals
as an important etiologic agent of cancer causation efforts were made to
understand the mechanisms of chemical carcinogenesis. From these efforts
chemical carcinogens are now grouped into two general categories based on
their mechanism of action. These categories are referred to as genotoxic and

nongenotoxic carcinogens.

1. Warns

Genotoxic carcinogens are defined as having the ability to react with
and cause damage to the DNA. By definition these agents are also mutagens.
This class of carcinogens consists mainly of materials that are chemically

electrophilic. A few of the genotoxic carcinogens, for example the alkylating

24

and acylating agents dimethyl sulfate, ethyl nitrosourea,
dimethylnitrosamine and sodium azide, are electrophilic enough to react
directly with the DNA (Mch et al., 1975). However, for the majority of
chemical carcinogens metabolic activation is required to convert the parent
material to a more reactive electrophilic species (Miller and Miller, 1981).
This metabolic conversion is accomplished primarily by a group of enzymes
located in the microsomal fraction of the endoplasmic reticulum referred to
as the cytochrome P-450 dependent mono-oxygenases. Electrophilic
activation may occur as a single step reaction or may involve multiple steps.
At the same time other reactions may be taking place to detoxify the parent
material or its metabolic intermediates. Therefore the potency of any
particular genotoxic carcinogen involves the balance between these
competing reactions.

Regardless of the particular mode of activation the end result is the
production of an electrophilic molecule that can react with DNA. No single
unique alteration of the DNA is responsible for the carcinogenic action of a
chemical carcinogen/mutagen. Chemical interaction with the DNA can
produce the formation of specific DNA base adducts that can lead to base pair
substitutions. Some chemicals can intercalate between the bases causing
frameshift mutations, additions or deletions. Other carcinogens induce gross
alterations of the genetic material producing chromosomal abnormalities
such as translocations or chromosomal breaks. All of these mutagenic

changes can have carcinogenic consequences.

1W

Not all carcinogenic chemicals have mutagenic activity. There is an
important class of chemical carcinogens, generally referred to as epigenetic or
nongenotoxic carcinogens, which do not interact directly with DNA but still
can cause the formation of tumors in animals and humans. This class of
carcinogens is composed of a diverse group of pharmacologically and
toxicologically distinct chemicals and materials. Examples include certain
chelating agents, hormones, pharmaceutical drugs, solid state materials
(plastics and metals) and various synthetic organic chemicals. Little is known
about the specific mechanisms by which these agents cause tumor
development. It has been suggested that these agents may exert their
carcinogenic activity by promoting previously initiated cells or through the
induction of somatic mutations by indirectly damaging DNA as a
consequence to some abnormally induced physiologic response (Schulte-
Hermann et al., 1983).

Studies conducted a number of years ago may have provided a possible
explaination as to how nongenotoxic agents can be carcinogenic. These early
studies demonstrated that chronic tissue irritation can cause the development
of tumors in rodents at the site of the irritation (Berenblum, 1929). This
irritation can be brought about through repeated freezing or by repeated
subcutaneous injections of nonreactive materials such as glucose, saline, and
distilled water. This continuous irritation can cause repetitive tissue damage

resulting in an increase in focal cellular proliferation. It is believed that the

26

increasein cellular proliferation is associated with the development of these
tumors.

Increased cellular proliferation is also frequently observed at
tumorigenic doses for many animal nongenotoxic chemical carcinogens.
This is particularly true for a number of small chlorinated hydrocarbons such
as chloroform, tetrachloroethylene and trichloroethylene (Loury et al., 1987)
This increase in cellular division appears to be a consequence of a toxicity
induceddegenerative/ regenerative response within the target organ.

It is not entirely clear how increased cellular proliferation can result in
the enhancement of tumor development. It has been well established that
mutations occur during normal DNA synthesis. As reviewed by Loeb et al.
(1978) and Hartman (1980), base mismatches may arise from polymerase base
selection errors or as a result of induced errors by proof reading enzymes.
Estimates of in viva somatic cell mutation rates vary from 10'9 to 10'12 events
per nucleotide per cell division (Drake et al., 1969). It stands to reason that an
increase in the number of replicating cells will result in an increase in the
mutation frequency. This in turn could lead to an increase in the number of
cells carrying mutations within important cancer genes (oncogenes or
suppressor genes). The end result could be an increase in the population of
initiated cells or the induction of additional mutational changes required for
the further neoplastic progression of previously initiated cells. Either one of
these mechanisms could lead to an increase in tumor development.

'As was discussed earlier, an important component of tumor

promotion is the expansion of the initiated cell population. Increases in

27

cellular proliferation as a result of exposure to some nongenotoxic agents may
increase tumor development by promoting previously initiated cells. Thus,
nongenotoxic compounds could increase tumor yield without causing any de
novo initiation.

Other nongenotoxic carcinogens do not appear to exert their
tumorigenic action through increases in cellular proliferation. Alterations in
gene transcription, perturbations in signal transduction pathways, inhibition
of intercellular communication and peroxisome proliferation have also been
proposed as mechanisms of nongenotoxic carcinogenicity (Butterworth and
Slaga, 1987). For example, the very potent rodent hepatocarcinogen
tetrachlorodibenzo-p—dioxin (TCDD) has a specific cytosolic receptor protein
with a high affinity for TCDD. This protein / TCDD complex translocates to
the nucleus of the liver cell and increases the transcriptional activity of
certain genes (Poland and Knutson, 1982).

Tetradecanoyl phorbol acetate (TPA) has been classically considered a
tumor promoter based on the two-stage mouse skin model. However TPA by
itself can cause a low incidence of skin tumors and when applied to hairless
mice elicits a 100% incidence of skin cancer (Inversen and Inversen, 1979).
These carcinogenic effects may in part be due to the ability of TPA to bind
directly to and activate protein kinase C. This inappropriate activation of
protein kinase C results in the perturbation of signal transduction pathways
altering gene expression which could have transforming consequences.

Some of these nongenotoxic agents may exert their carcinogenic effect

by causing indirect damage to the DNA. It has been known for some time

that many tumor promoters can activate polymorphonuclear leukocytes
(PMNs) and that this activation leads to the formation of large amounts of
active oxygen radicals (Goldstein et al., 1981). When these cells are activated
by nonphagocytic stimuli such as tumor promoters, these oxygen radicals are
free to react with neighboring cells and can damage cellular macromolecules
such as DNA. Several investigators have demonstrated that these PMN
generated oxygen species can induce DNA strand breaks and oxidized bases
such as 5-hydroxymethyl-2'-deoxyuridine and thymidine glycol in
coincubated cells (Dutton and Bowden, 1985; Frenkel and Chrzan, 1987).

A similar mechanism of action may be involved with another class of
nongenotoxic animal liver tumorigens which are commonly referred to as
peroxisome proliferators. Some of these peroxisome proliferators include the
hypolipodemic drugs clofibrate, ciprofibrate, fibric acid, fenofibrate, the
plasticizer di(2-ethylhexyl)phthalate and the organic solvent 1,1,2
trichloroethylene (Elcombe et al., 1984; Reddy et al., 1983; Kluwe et al., 1985). A
common physiologic response to exposure from these chemicals is a large
increase in the production of intracellular peroxisomes (Reddy et al., 1980).
These cytoplasmic organelles are involved in lipid metabolism. One
consequence of lipid metabolism is the production of H202 which is a highly
reactive molecule. It has been proposed that the increase in the number of
peroxisomes results in an excess production of hydrogen peroxide that may
escape normal detoxification mechanisms and react with cellular lipids to
initiate the auto-oxidation of fatty acids. Some of the products of this fatty acid

metabolism include a number of reactive oxygen species (ie. aldehydes,

29

organoperoxides and conjugated dienes) that could react and damage DNA
(Reddy and Lalwani, 1983).

Although the mechanisms of nongenotoxic carcinogenesis appear to be
many and diverse, the one obvious similarity is that they all seem to exert
their carcinogenic activity at dose levels which are high enough to elicit some
abnormal physiologic response (ie. toxicity, peroxisome proliferation etc).
This has important implications in the evaluation of chemical carcinogenic
risk for humans. It suggests that a carcinogenic threshold exists for these
compounds which implies that if the exposure is low enough so that the
physiologic response is not induced, then the development of tumors should

not occur.
C. W

In the late 1960's Huebner and Todaro (1969) proposed that specific
genes (which they named oncogenes) are involved in the genesis of cancer.
Support for the existence of these genes has been provided by research
demonstrating that (1) cells can be transformed in vitro or tumors induced in
viva by exposure to radiation and mutagenic chemicals, (2) damage to DNA is
associated with cellular transformation, (3) a hereditary prediposition exists
for certain types of cancer, and (4) defects in certain DNA repair processes are
associated with specific cancer predispositions.

Since the conception of this hypothesis, cancer genes have been

identified in a variety of mammalian cancer cells. The first oncogenes were

30

however originally identified in a specific class of animal tumor viruses, the

retroviruses.

' 1. W

The type-C retroviruses represent a group of RNA tumor viruses that
can induce leukemias, lymphomas, sarcomas, and carcinomas in many
different animal species including man (Bishop, 1983). The genome of these
viruses consists of a single strand of RNA approximately 10 kb in length
(Bister and Duesberg, 1983). Three genes, required for viral replication (gag,
pal, env), are bound on each side by several hundred bases of untranslated,
repetitive RNA referred to as long terminal repeats (LTR's). These LTR's
contain transcriptional promoter and enhancer sequences in addition to a 5'
cap and 3' poly -A tail. The gag gene codes for a structural protein that is
associated with the RNA; the pol gene makes the reverse transcriptase used to
synthesize a complementary DNA from the RNA; and the env gene codes for
a glycoprotein which resides in the lipoprotein envelope of the viral particle.
Based on their transforming properties these viruses are divided into 2
general groups referred to as (1) the acutely transforming retroviruses and (2)

the latent transforming viruses.

a. The acutely transforming retroviruses

The acutely transforming viruses are characterized by the ability to
induce rapid tumor development in infected animals. These viruses can

produce tumors because they contain an oncogene within their viral genome

31

(Bister and Duesberg, 1982). All of the acutely transforming viruses (except
the Rous sarcoma virus) have deletions in one or more of the three essential
genes for replication. As a result these viruses are replication defective. Viral
progeny can however be produced if these viruses are co-infected with non-
transforming helper viruses.

Once the virus enters the cell, the RNA is converted into DNA which
integrates within the infected cell's genome. Expression of the viral oncogene
results in the rapid transformation of the cell and eventual tumor
development. The different transforming retroviruses are distinguished from
each other by the particular oncogene residing within the virus. There are
approximately 20 different retroviral oncogenes which can induce most of the
major forms of neoplasms (Bishop, 1987).

These viral oncogenes have evolved from normal cellular genes. They
have been. acquired by the virus as a result of a viral transduction event
(Bishop, 1983). During the evolution of the virus, specific structural changes
(such as base substitutions, additions and deletions) have taken place within
the normal cellular gene to produce the transforming properties of these viral
oncogenes. The transforming capabilities of these viruses rests not only with
the acquisition of an oncogene but with the ability to produce very high levels
of the oncogene product. This capability for the high level of gene expression
is due to the presence of the transcriptional enhancers located within the
LTR's of the viral genome (Bister and Deusberg, 1982).

A variety of structurally diverse proteins are encoded by the

transforming retrovirus oncogenes. These genes can be grouped using

32

several different criteria. One that is particularly useful is based on location
and biochemical activity of the protein products. Table 2 provides a partial
list of some of the more common oncogene protein products.

In general, the products and transforming function of the viral
oncogene proteins are not well understood. The ability of these proteins to
transform cells is undoubtedly related to their ability to disrupt the regulation
of normal cellular proliferation and/ or differentiation. A more detailed
discussion of the biological function of some of these oncogene protein

products is provided in a later section.

33

Table 2. Selected oncogenes: function and cellular localizationa

 

Oncogene Retrovirus Cellular Property or
location function of protein
Growth factors
sis simian sarcoma extracellular related to PDGF
' . . virus
int-2 - - extracellular related to fibroblast
growth factor
Protein kinases
src Rous sarcoma plasma tyrosine kinase
virus membrane
yes avian sarcoma plasma tyrosine kinase
virus membrane
abl Abelson murine plasma tyrosine kinase
leukemia virus membrane
fes feline sarcoma plasma tyrosine kinase
virus membrane
fps feline sarcoma plasma tyrosine kinase
virus membrane
mos Moloney murine cytoplasm serine/ threonine
sarcoma virus kinase
raf avian sarcoma serine/threonine
virus kinase
Growth factor
receptors
erb-B avian transmembrane EGF receptor
erythroblastosis tyrosine kinase
neu - transmembrane similar to EGF
receptor
fms feline sarcoma transmembrane CSF-1 receptor
tyrosine kinase
ros avian sarcoma transmembrane structure related to

growth factor

 

Table 2 (cont)
Oncogene Retrovirus Cellular Property or
location function of protein
GTP binding
H-ras Harvey murine plasma GTP binding/GTPase
sarcoma virus membrane binds with GAP
K-ras Kirsten murine plasma GTP binding/GTPase
sarcoma virus membrane binds with GAP
N-ras - plasma GTP binding/GTPase
membrane binds with GAP
Nuclear
myc avian nucleus binds d.s. DNA
myelocytomatosis
L-myc - nucleus binds d.s. DNA
N-myc nucleus binds d.s. DNA
ski avian SKV770 nucleus ? '
myb avian nucleus ?
. myeloblastosis
ets E26 nucleus ?
fos murine nucleus AP-1 transcription
osteosarcoma factor
jun avian sarcoma nucleus AP-I transcription
virus-17 factor

a Derived from Cotran et al., 1989; Land et al., 1983; Hunter T., 1984.

b. The latent transforming retroviruses

These viruses are replication competent and have the typical retroviral

genomic organization discussed above. They do not have transforming

oncogenes, but can still cause the development of tumors after a long latency

35

period following infection. The mechanisms by which they cause neoplastic
transformation is frequently referred to as "insertional mutagenesis" because
it is the insertion of the proviral DNA at critical sites within the host genome
that leads to cellular transformation (Bishop, 1987). If this insertion occurs
within a cellular gene, structural modifications may take place leading to the
oncogenic conversion of the gene. Alternatively, the presence of strong viral
promoters near a cellular gene may lead to the production of high levels of
the protein or result in the inappropriate expression of the gene. Thus, both
genetic damage and / or altered gene expression may underly the neoplastic
transforming properties of these viruses.

Many of the latent transforming animal retroviruses are associated
with the development of chronic leukemias. One human retrovirus, HTLV-
1, appears to be linked to the development of adult T-cell leukemia (ATL) in
humans (Ehrlich and Poiesz, 1988). Evidence associating this virus with the
etiology of ATL in humans includes the fact that (1) this virus has repeatedly
been isolated from patients with ATL, (2) HTLV-l proviral sequences have
been found in leukemic cells and not in normal cells from these patients, (3)
cultured neoplastic cells release H'TLV-l that can immortalize T cells in vitro,
and (4) antibodies against HTLV-l are found in over 90% of patients with
ATL .

The genomic structure of HTLV-I is typical of other latent
transforming retroviruses in that it contains the 3 essential genes flanked by
the LTR sequences. This virus differs in that it contains a fourth coding
region located between the env gene and the 3' LTR which has been named

tat (Nerenberg et al., 1988). The tat region codes for at least 3 proteins and it is
believed that these proteins are partially responsible for the transforming
activity of the virus. One protein causes an increase in HTLV-l transcription
by acting on the viral 5' LTR. The other two proteins induce the activation of
interleukin-2 (IL-2) and the IL-2 receptor. Thus, infected cells produce
increased amounts of a growth factor and its receptor which would act as an
autocrine stimulation for cellular proliferation. This process alone cannot
cause cellular transformation indicating that additional changes must also
take place, once again demonstrating the multi-step process of tumor

development.

2. W

Several DNA viruses have been associated with the development of
cancer in animals and recent evidence suggests they are involved in some
human cancers. These viruses belong to several different taxonomic families
which include the Papovavirus, Adenovirus, Herpesvirus, and

Hepadenovirus groups.

a. Oncogenic Pamvaviruses

Members of the Papovavirus group are of particular interest because
they include two viruses, SV-40 and polyoma, of which more is probably
known about their mechanism of transformation than about that of any
other DNA tumor virus. In addition, this family includes the human

papilloma virus (HPV) which has been associated with development of

37

several types of human cancers including in situ and invasive cancer of the
cervix (Bonfiglio,1988).

The genomes of both the SV-40 and polyoma viruses are composed of a
small, circular, double strand of DNA approximately 5 kb in length (Dulbecco
and Ginsberg, 1980). Infection by the virus is divided into two phases: an
early phase .in which the viral DNA is replicated and a late phase during
which viral proteins are synthesized and assembly of viral particles takes
place. Each phase of the infection cycle is accompanied by a specific program
of gene expression.

These viruses can produce a productive or nonproductive infection
depending on the species of the infected cell. Productive infection occurs in
cells referred to as permissive. In these cells expression of both the early and
late genes occurs resulting in the synthesis of large numbers of viruses which
eventually kill the cell. Each type of papova virus has a narrow range of
permissive cells in which expression of both early and late genes results in a
lytic infection. For SV-40 and polyoma viruses, monkey and mouse cells
respectively are the permissive hosts (Delbecco and Ginsberg, 1980).

If infection occurs in a nonpermissive cell, viral DNA synthesis does
not occur and the late genes are not expressed (Darnell et al., 1986). The
expression of the early genes however still takes place which induces the cell
to enter the S phase of the cell cycle. As long as the expression of the early
viral genes occurs the cell is continually stimulated to proceed through the
cell cycle. Most of the infected nonpermissive cells eventually revert to a

normal phenotype as a result of the degradation or dilution of the viral

 

38

genome. If the viral DNA integrates into the genome of the host cell in a
manner resulting in the constitutive expression of the early genes with the
subsequent loss of late gene expression, then those cells become permanently
transformed (Delbecco and Ginsberg, 1980). This integration involves a
nonspecific type of recombination with the sites of integration occurring
within the viral late gene region and occurring randomly within the recipient
DNA. '

These viruses therefore owe their transforming capabilities to the
proteins of the early genes. The early gene region of SV-40 codes for two
proteins; large T and small t, and the polyoma early gene region codes for
three proteins; large T, middle T, and small t. The early region of the
polyoma genome is transcribed as one large mRN A molecule (3000 bp) that is
spliced in three different ways to produce the appropriate transcripts for the
three early proteins (Rassoulzadegan et al., 1982)

Using molecular cloning techniques, plasmids have been constructed
containing sequences coding for each of the early region polyoma proteins.
Transfecting these plasmids into various cell lines has established that the
polyoma large T protein can induce cellular immortalization whereas the
middle T protein can transform previously immortalized cells (Gross, 1983).
The mechanisms by which these proteins elicit their phenotypic response are
not entirely understood. Recent advances have provided some important
insights into the molecular mechanisms associated with the transforming
properties of these proteins. The middle T protein of polyoma has been

shown to associate with the cellular src protein at the inner aspect of the

 

39

cytoplasmic membrane (Courtneidge and Smith, 1983). This association
increases the specific tyrosine kinase activity of the src protein by as much 50-
fold. The middle T protein accomplishes this activation apparently by
blocking the phOSphorylation of specific tyrosine residues on the c-src protein
that when phosphorylated down regulates its biochemical activity.

A similar process appears to be involved with the large T protein of the
SV40 virus. Based on structural and genetic studies, the large T protein has
two domains which are involved in the transformation of cells in culture
(Srinivasan et al., 1989). One of these domains imparts the immortalizing
function of the protein. This domain contains the binding site for the cellular
protein p53 and forms a stable complex with this protein in cells infected and
transformed with the virus (Lane and Crawford, 1979). The p53 protein has
been associated with the process of neoplastic tranSformation in a number of
cell types (Dippel et al.,1981; Oren et al.,1981). Recent studies by Finlay et al.
(1989) support the idea that the p53 protein may act as a suppressor to block
cell transformation. The transforming activity of the large T protein may in
part be due to an inactivation of the suppressing activity of p53 through the
formation of this protein complex.

The second domain of the large T protein forms a complex with
another known tumor suppressor gene, the retinoblastoma protein (p105-RB)
(DeCaprio et al.,1988). Inactivation of this protein by the formation of the
large T/ p105-RB complex is also thought to play an important role in the

transforming activity of the SV40 virus.

A similar complex is formed with the human papilloma virus -16
transforming protein E7 and the RB protein (Dyson et al.,1989). The
formation of this complex is not surprising based on the structural similarity

between the E7 protein and the SV40 large T antigen.

b. Oanenic Adenoviruses
Oncogenic members of the Adenovirus family include types 2, 5 and 12

(Delbecco and Ginsberg, 1980). None of these viruses cause tumors in
humans but they can induce the development of sarcomas in the
nonpermissive hamster or rat. The genes required for transformation are
localized in the 6-7% extreme left region of the viral genome involved in
early gene expression. As seen with the papova viruses, transformation
occurs as a result of expression in the early genes. Two transforming genes,
Ela and Elb, have been localized within this early region. Transformation of
rodent cells in culture can be achieved if they are transfected with
recombinant plasmids that contain only these early genes (Iochemsen et
al.,1982). 4

The role of the Ela and Elb proteins in cellular transformation has
been extensively examined. Ela has been shown to immortalize cells in
culture when transfected with this gene (Houweling et al., 1980). In addition,
it has been demonstrated that Ela can cooperate with the ras cellular
oncogene to induce complete neoplastic transformation of primary rat

embryo fibroblasts in culture (Ruley, 1983).

41

Both the Ela and Elb proteins form complexes with specific cellular
proteins which have been shown to be involved in oncogenic
transformation. Ela, like the large T antigen of SV40 and the E7 protein of
HPV-16, forms a complex with the RB protein (Whyte et al., 1988). The EIB
protein forms a complex with the p53 suppressor protein (Sarnow et al.,1982).
It is believed that the formation of these protein complexes is an important

function in the oncogenic activity of these proteins.

c. Oncogenic Herpesviruses

Epidemiological data have demonstrated a strong association with the
infection of Epstein-Barr virus (EBV) and the development of several types of
human lymphomas (e.g. Burkitt's lymphoma) and human nasopharyngeal
carcinoma (Cotran et al., 1989). Burkitt's lymphoma is endemic in certain
parts of Africa and New Guinea, whereas nasopharyngeal carcinoma is
endemic to parts of Africa, Southern China and in Arctic Eskimos. Nearly
100% of patients in those areas of the world with these diseases carry parts of
the EBV genome in their cells and have elevated antibody titers to the capsid
proteins.

Recent work in several laboratories suggests that the Epstein-Barr virus
carries transforming genes. Elliott Kieff's group has shown that the viral
latent membrane protein (LMP) gene has transforming activity (cited by Man,
1989). The mechanism of action for this gene is presently unknown, but it is
thought to code for a growth factor or ion channel receptor. Another gene,
EBNA-Z (Epstein-Barr virus nuclear antigen-2), was shown by William

42

Sugden to 'induce immortalization of lymphocytes (cited by Marx, 1989). This
may be accomplished in part by stimulating the expression of other viral and
cellular genes including the LMP gene.

Additional genes or factors must be involved in the development of
these cancers besides EBV infection. Most adults throughout the world are
infected by EBV but outside of these particular geographic areas they are
asymptomatic for these cancers (Cotran et al., 1989). One such factor in the
case of Burkitt's lymphoma may be the requirement for specific genetic
rearrangements. All of the patients with this disease have a chromosomal
translocation of the distal part of chromosome 8 with either chromosomes 14,
22, or 2 (Klein, 1983). As will be discussed later, these translocations activate

the expression of the cellular oncogene c-myc.

d. DNA tumor viruses without oncogenes

There are several oncogenic viruses which do not have any known
transforming genes. One of these, the hepatitis B virus (HBV), has been
closely associated with the occurrence of human liver cancer (Cotran et al.,
1989). Like the Epstein Barr virus, HBV is endemic in many third world
countries particularly those in Africa and in the Far East. Correspondingly
these areas have very high incidences of hepatocellular carcinomas. A 10 year
epidemiological study conducted in Taiwan has shown that carriers of the
virus have about a 100 times higher probability of developing liver cancer
than do uninfected individuals (United States Public Health Service
Document, 1989). This is a significantly high risk when it is compared to

43

other cancer risk factors. For example, the epidemiological data are very
conclusive in demonstrating that cigarette smoking is an important etiologic
factor contributing to the development of lung cancer (Doll and Hill, 1950).
Even as strong as this association is, the risk of developing lung cancer in
smokers is still 5 times lower than the risk of HBV infected individuals to
develop liver cancer. This level of risk leaves very little doubt that HBV is a
major cause of this disease.

Although transforming genes have not been detected in this virus
several mechanisms have been proposed to explain its' oncogenicity. Work
by Pierre Tiollais has found that sometimes when the virus inserts itself into
the host genome it does so near a cellular oncogene. As a result this gene is
inadvertently activated leading to its' unscheduled expression (Marx, 1989).

Chronic liver damage, as a result of persistent HBV infection, also
appears to play a important role in the tumorigenicity of this virus (Chisari et
al., 1989). Transgenic mice were developed carrying the gene for the viral
surface antigen. Expression of this gene in these animals resulted in large
deposits of the antigen protein within the liver causing necrosis. This
damage was accompanied by an increase in cell division to regenerate the
damaged tissue. After months of tissue damage and regeneration, cancerous
nodules began to develop in the liver of these animals. Although a build-up
of this antigen does not occur in humans, chronic infection does cause liver
injury as a result of immune attack on infected cells. The mechanism by

which this virally induced repeated cellular division causes liver tumor

development may be very similar to that discused above with certain

nongenotoxic chemical carcinogens.

'3.ClllQ 'C' .

The identification of transforming genes in animal tumor viruses and
the detection of sequences homologous to the retroviral oncogenes in normal
mammalian cells suggested that neoplastically transformed cells may contain
activated oncogenes. Experiments conducted in the late 1970's by Dr. Robert
Weinberg's laboratory provided the first evidence for the existence of these
cellular transforming genes (Shih et al., 1979). DNA was isolated from
several chemically transformed rodent cell lines and when transfected into
NIH 3T3 cells caused the morphological transformation of these cells. These
experiments established the presence of oncogenic sequences within the DNA
of the chemically transformed donor cell lines.

Since these pioneering experiments, numerous studies have
demonstrated the presence of transforming sequences in a variety of animal
and human tumors (Bishop, 1987). Many of these transforming sequences
have been identified to be homologous to the retroviral transforming genes.
Other transforming genes found within these tumors have no apparent viral
counterpart and therefore represent unique cellular oncogenes (Table 1).

DNA sequences homologous to the cellular oncogenes have been

found in normal mammalian cells and in organisms as evolutionarily

45

divergent as drosophila and yeast (Shilo and Weinberg, 1981; Powers et al.,
1984; Chang et al., 1982). This suggests that the transforming oncogenes are
derived from normal cellular genes (proto-oncogenes). The high degree of
evolutionary conservation also indicates that the proto—oncogenes play a basic
role in the cells normal biochemistry. These concepts raise several very
irriportant and fundamental questions concerning the role and function of
cellular oncogenes in the development of cancer: 1) how are the proto-
oncogenes converted into their oncogenic counterparts and 2) what is the
normal function of the proto-oncogene protein products and how is this
altered by conversion to a transforming oncogene? Three basic mechanisms
have been identified describing how proto-oncogenes are converted into
transforming oncogenes: (1) point mutations (2) gene rearrangements (3) gene

amplification.

a. Activation by mint mutation

The ras oncogenes represent the first and most extensively studied
example of a family of proto-oncogenes activated by a point mutation. Point
mutations activating this family of oncogenes (H, K and N) have been
detected in three distinct functional regions within the gene. Mutations
affecting codons 12 and 13, 61 and 117 have been detected in various human
tumors and tumor cell lines as well as in experimentally induced animal
tumors (Barbacid, 1987; Reynolds et al., 1987). Experimentally induced
mutations in codons 59, 63, 116 and 119 have also been demonstrated to cause

the activation of the ras gene (Fosano et al., 1984; Walter et al., 1986; Sigel et

al., 1986). A single amino acid change at one of these codons is sufficient to
convert the physiologically normal 21,000 dalton ras protein into its
oncogenic form.

The neu oncogene is a recent example of another oncogene which is
activated by a point mutation. This gene was originally isolated from
transformed NIH 3T3 cells that were transfected with DNA from cell lines
derived from ethylnitrosourea induced rat neuro/glioblastomas (Schechter et
al., 1984). The neu oncogene is homologous to, but distinct from, the gene
coding for the epidermal growth factor (EGF) receptor. DNA sequence
analysis of four independently derived NIH 3T3 transformed clones has
revealed a single point mutation occuring at the same codon position within
this gene (Bargman et al., 1986). This mutation results in the substitution of
the normal valine with a glutamic acid within the transmembrane domain of
the protein. Unlike the ras gene where an activating point mutation can
occur at several different regions, the activation of the neu proto-oncogene
apparently involves just a single site. The mechanism by which this
mutation activates the neu gene is not understood at this time. Significant
progress has, however, been made in understanding the transforming activity

of the ras gene. This will be described later in more detail.

b. Activation by gene rearrangement

Specific chromosomal translocations occur at high frequencies in

certain types of tumors, suggesting that these chromosomal rearrangements

47

may have an etiologic role in the development of these tumors (Brodeur,
1986a). This concept has gained greater acceptance with the finding that in
several of these cancers the translocations involve the rearrangement of
specific oncogenes.

These rearrangements can result in the abnormal expression of the
translocated gene or lead to the genesis of a hybrid gene whose protein
product has transforming activity.

One of the best examples involving this type of oncogene activation
involves Burkitt's lymphoma. One hundred percent of the patients with this
type of cancer have one of three translocations involving chromosome 8
(Cotran et al., 1989). The most common rearrangement involves a reciprocal
translocation between chromosome 8 and chromosome 14. The c-myc
oncogene has been located on the distal end of chromosome 8 where in
normal cells expression is regulated during the cell cycle (Campisi et al., 1984;
Dalla-Favera et al., 1982). This translocation results in the placement of the c-
myc oncogene near the heavy chain irnmunoglobin gene (Taub et al., 1982).
At this new chromosomal location the expression of the c-myc gene is under
the control of the heavy chain gene. As a result high levels of c-myc
expression are detected in these cells (Erickson et al., 1983a).

Two alternative translocations involving chromosome 8 are observed
in variants of Burkitt's lymphoma. In these variant translocations the c-myc
gene remains on chromosome 8 and the immunoglobin kappa light chain
gene on chromosome 2 or the lambda light chain gene on chromosome 22 are

translocated and positioned near the c-myc gene (Erikson et al., 1983b; Hollis

et al., 1984). In both cases the c-myc gene comes under control of the
irnmunoglobin genes where transcription is greatly increased.

A similar translocation is observed in acute T-cell or chronic
lyniphocytic leukemia. Deregulation of the c-myc gene occurs when the T cell
receptor on chromosome 14 is translocated near the c-myc gene on
chromosome 8 (Erikson et al., 1986).

A somewhat different type of oncogene activation involving a
chromosomal rearrangement is found in chronic myelogenous leukemia.
Greater than 95% of these cancers involve a reciprocal translocation between
chromosome 9 and chromosome 22 (Nishimura and Sekiya, 1987). This
translocation results in an abbreviated form of chromosome 22 which can be
used as a prominent cytological marker (referred to as the Philadelphia
chromosome) for the clinical diagnosis of the disease.

The c-abl gene has been located on chromosome 9. This particular
translocation joins the c-abl gene with the bar (breakpoint cluster region) gene
on chromosome 22 to form a new hybrid gene (Heisterkamp et al., 1983). The
normal c-abl gene shares sequence homology with the tyrosine kinase family
of oncogenes but the protein product of this gene does not have any kinase
activity (Ponticelli et al., 1982). The c-abI-bcr hybrid produces a chimeric
protein with kinase activity. As a result of this specific translocation a new
gene has been generated whose protein product has the biochemical activity

of several other known oncogenes.

c. Activation through gene amplification

49

The above discussion provided several examples of how increased
oncogene expression, as a result of chromosomal translocations, could be
associated with neoplastic development. Gene amplification is another
mechanism in which high levels of oncogene transcriptional activity can be
achieVed. This form of oncogene activation appears to be an important step
in the genesis of certain types of malignancies. ‘

- For example, in approximately 40% of human neuroblastomas, the c-
myc gene is amplified between 3 and 300 times (Brodeur, 1986b). There is a
strong correlation between the degree of neoplastic progression and clinical
prognosis with the extent of c-myc expression. The amplification of the c-myc
gene has also been observed in neuroendocrine cells isolated from human
colon carcinomas (Alitalo et al., 1983).

Amplification of the neu oncogene has been detected in approximately
30% of human breast cancers (Slamon et al., 1987). As was observed with the
c-myc gene, overall survival as well as the probability of remission is
correlated to the degree of amplification. One possible explanation as to how
amplification of this gene is related to these clinical parameters is provide by
studies which have demonstrated that breast carcinoma cells expressing the
epidermal growth factor receptor have the same poor prognosis (Sainsbury et
al., 1987). Presumably the same ligand that binds to the EGF receptor also
binds to the product of the neu oncogene (which is similar to the EGF
receptor). Many breast carcinoma cells also produce transforming growth
factor a (TGFva) which can bind to the EGF and neu receptor to stimulate

growth. Synthesis of TGF-a by the breast carcinoma cells and its binding to

50

the neu receptor protein may set up an autocrine growth stimulatory process
that contributes to the negative clinical prognosis associated with the
amplification of the neu gene.

Amplification of these genes can occur within the confines of the
chromosome or externally to it (Schwab et al., 1983). The internal sites of
amplification are often located in areas represented as homogeneously
staining regions (HSR's). External amplification is found in extra-
chromosomal fragments referred to as double minute chromosomes. Both
types of amplification can visibly be observed through the cytological

examination of cells isolated from the tumors.
4. W

a. The retinoblastoma paradigm

The above discussion has focused on a group of dominant acting
oncogenes that cause neoplastic transformation, in part, by promoting
excessive or inappropriate cell proliferation. These genes are considered
dominant because they can induce neoplastic transformation when only one
of the two normal proto-oncogene alleles is converted to its oncogenic form.
If neoplastic transformation can be induced by the activation of a set of
growth-promoting genes then, alternatively, cancers could arise by the
inactivation of genes that normally function to suppress cell proliferation.

Evidence to support this idea has been provided by a number of cell
fusion experiments (Klinger, 1982; Klein, 1987). In these experiments cancer

51

cells are fused with normal cells and frequently the tumorigenic phenotype is
lost in the hybrid cell. These results suggest that at least one of the properties
associated with the neoplastic transformation of a cell involves the loss of
growth-regulatory genes which are restored by the fusion to a normal cell.

It is inferred from these experiments that normal cells contain specific
genes that function to suppress properties associated with neoplastic growth.
A single normal allele from one of these genes is all that would be necessary
to exert this suppressive activity (unless there is a dosage effect).
Consequently, both alleles must be inactivated if it is to participate in the
neoplastic conversion of the cell. Because of this, these genes are sometimes
referred to as recessive cancer genes (Hansen and Cavenee, 1988).

Convincing evidence supporting the existence of cancer suppressor
genes has been provided by the elegant studies determining the mechanisms
involved in the causation of retinoblastoma. This particular type of cancer is
relatively rare and affects approximately 1 out of every 20,000 infants and
children (Cotran et al., 1989). The disease occurs sporadically in about 60% of
the cases and in the remaining 40% there is a hereditary predisposition
toward the development of the disease.

Nearly 20 years ago, Knudson hypothesized that the development of
this disease required two genetic events (Knudson, 1971). He suggested that
in the hereditary cases, the first event is acquired from one of the parents.
Careful cytological analysis of cells from individuals with the hereditary form
of the disease occasionally revealed a small deletion in one of the number 13

chromosomes (Yunis and Ramsay, 1979). It was hypothesized that a gene

52

(referred to as the Rb gene) was located in the area of this deletion and that
the inactivation of the Rb gene was involved in the development of
retinoblastoma. The inactivation of the single Rb allele is however not
sufficient to cause the disease. Only focal lesions are observed within the
retina of individuals with the hereditary form even though all of the cells
contain this initial defect. This suggests, as Knudson had hypothesized, that a
second genetic event is necessary. This observation also supported his
postulate that one of these events is genetically transmitted in the familial
form of the disease and that in the sporadic cases two somatic events must
occur within the same cell.

Evidence supporting the existence of the second event was obtained by
detecting DNA restriction fragment length polymorphisms between normal
and tumor tissue at the Rb locus of individuals with the hereditary form of
the disease (Sparkes et al., 1983). This analysis revealed that in the normal
tissue of the retinoblastoma patient the Rb locus was heterozygous but in the
tumor cells this locus existed in the homozygous condition. This established
the fact that not only were two genetic events required but that these events
involved the inactivation of both alleles from the same gene. Because both
alleles of the Rb gene need to be inactivated to cause retinoblastoma, it was
concluded that the Rb gene normally functions to suppress the tumor

phenotype. These studies provided convincing evidence for the existence of

tumor suppressor genes.

b. Other possible tumor suppressor genes

53

A specific chromosomal deletion is also associated with Wilm's
tumors. In this disease, the deletion occurs in the p13 band of chromosome
11 (Cotran et al., 1989). An experimental approach similar to the one
described in the retinoblastoma studies has indicated that loci homozygocity
also exists at the p13 region of chromosome 11 in Wilm's tumor cells (van
Heyningen et al., 1985). Although the specific gene responsible for this
disease has not been isolated, a recessive suppressor gene is suspected because
of the similarity in the observations to the retinoblastoma example.

A recent report by Levine and his colleagues provides data implicating
the p53 gene as a possible suppressor gene (Findlay et al., 1989). In this study
they generated transformed primary rat embryo fibroblasts by cotransfecting
cultures with various combinations of oncogenes. For example, transformed
foci were generated by transfection with the adenovirus EM and the activated
ras oncogenes. If these primary cells were transfected with EIa, ras and with
the wild type p53 gene, a significant reduction (67%-80%) in the number of
transformed foci was observed. This result suggests that the p53 gene
suppressed the transforming potential of the E1 a and ms genes. The
incomplete suppression of foci development by the p53 gene was explained in
the subsequent analysis of the few transformed foci that did develop. This
analysis revealed that the p53 gene had been integrated into the genome of
these transfectants but that it either was either not being expressed or was
expressed in a mutant form. Because the p53 gene was basically inactive it

could not exert its suppressive activity in those cells.

Additional evidence supporting the idea that the p53 gene is a

- suppressor gene is provided by the studies examining the interaction of this
gene with the products several DNA tumor viral oncogenes. As was
discussed earlier, these viral transforming proteins form stable complexes
with the p53 protein. It is thought that the transforming activity of these
viral proteins is due in part to the inactivation of the p53 protein through the
formation of this complex. Such a mechanism would provide a very
plausible explanation for the transforming activity of these viruses if the p53
gene has a suppressor function. This data, in conjunction with the
information presented above, strongly suggests that the p53 gene may act as a

suppressor of neoplastic transformation.

ummmmmwm

The regulation of normal cellular proliferation involves a complex
interaction between extracellular and intracellular proteins. Information
present in specific extracellular proteins is transmitted from outside the cell
through a series of biochemical reactions into the nucleus which in turn
instructs the cell to divide. Many of the viral and cellular oncogenes appear
to play an important role in this information transfer. The protein products
of these genes have altered biochemical activity which can interfere with the
regulated transfer of this information resulting in the abnormal proliferation

commonly associated with the ne0plastic cell. These oncogenic proteins can

55

be grouped according to their subcellular localization and biochemical activity
(Table 1).

1. Wins

An important step linking the protein products of oncogenes with
aberrations in growth regulation was the finding that the simian sarcoma
virus oncogene, v-sis, coded for a protein that was similar to the [3 chain of
human platelet-derived growth factor (PDGF) (Waterfield et al., 1983). PDGF
is a potent serum rnitogen that can induce DNA synthesis and cell division in
many cells in the absence of other growth factors (Goustin et al., 1986).
Several studies have shown that most transformed mesenchymal cells
produce PDGF (Bowen-Pope et al., 1984). The biologically active form of
PDGF consists of a dimer composed of a 14-18 kDa a chain and a 16 kDa [3
chain (Iohnsson et al., 1982). Stimulation of cells with PDGF results in the
autophosphorylation of the 185 kDa PDGF receptor protein (Ek and Heldin,
1982). This presumably initiates a cascade of chemical reactions transmitting
the proliferative information of PDGF to the cell nucleus.

The transforming activity of the simian sarcoma virus is thought to be
caused by the high level of. v-sis gene expression. This high expression of the
v-sis gene results in the overproduction of PDGF which is believed to
stimulate the tumor cell to grow by interaction with its own receptor through
an autocrine mechanism (Sporn and Todaro, 1980). Evidence contradicting

this autocrine hypothesis has been provided by studies where high

concentrations of anti-PDGF antibodies added to the culture medium did not
reverse the transformed phenotype of SSV-transformed cells (Huang et al.,
1984). Furthermore, continuous addition of PDGF to the cultures of normal
NRK cells does not induce cellular transformation.

This contradiction may in part be explained by recent evidence where a
mutant form of the v-sis gene was constructed which causes the v-sis protein
product to be retained within the cell (Bejcek et al., 1989). PDGF produced by
this gene is sequestered within the endoplasmic reticulum and the Golgi.
Cells expressing this gene were transformed to the same extent as were cells
transformed with the wild type v-sis gene. This data supports a mechanism
whereby the v-sis and possibly the c-sis gene products may internally
stimulate the PDGF receptor.

2. WWW

There are several oncogenes, both of viral and of cellular origin, that
code for growth factor receptor proteins located within the cytoplasmic
membrane. These receptor proteins can be divided into three functional
domains; an external domain which binds the growth factor protein, a
transmembrane domain and a cytoplasmic domain having protein kinase
activity.

The first oncogene identified to code for a growth factor receptor was
the transforming gene of the avian erythroblastosis virus (AEV), v-erb-B.

Using a computer protein sequence data base, Downward et al. (1984) found

that the the v-erb-B protein has extensive homology to the human epidermal

57

growth factor (EGF) receptor. The protein synthesized from the v-erb -B gene
was however a truncated form of its cellular counterpart. This protein lacked
a complete external domain which binds the EGF protein. The authors
hypothesized that the transforming activity of the v- erb -B gene may be
caused by a continuous proliferative signal elicited by the truncated receptor
even in the absence of EGF. This hypothesis is partially substantiated by
studies showing that AEV transformed erythroblasts fail to bind EGF

(W aterfield, 1986).

The myloid cell growth factor, CSF-l, is another example of a growth
factor that is coded for by a potential cellular oncogene. This protein is
synthesized by cellular proto-oncogene fms (Sher et al., 1985). A truncated
version of this receptor protein is coded by the feline sarcoma virus oncogene
v-fms. It seems that during the viral transduction of the cellular fms gene
part of the 3' end of the gene was deleted (Waterfield, 1986). As a result the
carboxyl terminal end of the v-fms protein is missing. This region of the
protein represents the internal cytoplasmic domain of the molecule. The
protein kinase activity of the protein is located in this portion of the receptor.
It is thought that the transforming activity of this oncogene product is related
to a possible loss of kinase activity.

The neu, cellular oncogene codes for a protein which may represent a
new growth factor receptor protein. The amino acid sequence of this protein
is similar to but distinct from the EGF receptor. Unlike the erb-B gene, the
neu oncogene has no viral counterpart. The activation of this gene also

differs from that of the v-erb-B gene. Activation involves a single point

58

mutation in the transmembrane domain of the protein (Bargman et al., 1986).
How this single mutation activates this gene is presently unknown.

Many human tumors, particularly breast tumors, contain amplified
copies of this proto—oncogene. In addition to the point mutation, an increase
in gene expression as a result of amplification of this gene may play a role in

its tumorigenicity (Slamon et al., 1987).

3. W

The majority of the viral and cellular oncogene proteins reside within
the cytoplasm however many are closely associated with the cytoplasmic
membrane (Table 1). In general, the biochemical activity and role that these
proteins play in cellular transformation are poorly understood. Many of
these proteins have phosphokinase activity (Teich, 1986; Cotran, 1989). The
majority of these proteins exclusively phosphorylate tyrosine although there
are several that can phosphorylate serine and threonine (ie. mos, raf). These
proteins also have the capability of phosphorylating themselves. This process
of autophosphorylation may serve an important regulatory function in
controlling their biochemical activity.

One of the best studied oncogenic proteins within this group is the src
protein. The gene coding for the src protein was originally identified as the
transforming gene of the Rous avian sarcoma virus (Duesberg and Vogt,
1970). A homologous sequence, c-src, was subsequently identified and
isolated in mammalian cells (Bishop, 1983). Activation of the c-src gene

during viral transduction involves both genetic substitutions and point

59

mutations (Takeya and Hanafusa, 1983). The v-src gene codes for a protein
with a molecular weight of 60,000 daltons and is usually referred to as pp605'c.
It is believed that the transforming ability of this protein involves aberrations
in the specific tyrosine kinase activity (Piwnica-Worms et al., 1987).

A number of cellular proteins with phosphorylated tyrosine residues
have been identified in cells transformed with v-src. One that is of particular
interest is vinculin, a cytoskeletal protein (Cotran et al., 1989). This protein is
a component of the cytoskeletal network of mesenchymal cells and is located
at the points of contact between cells known as focal adhesion plaques.
Changes in the phoshorylation of vinculin may be associated with some of
the morphological changes detected in transformed cells such as rounding up
and decreased cohesiveness.

Some evidence also exists suggesting that pp605l‘c may also be capable of
phosphorylating certain lipids (Cotran et al., 1989). Increased cellular
proliferation by an activated src protein may involve a stimulation of the
phosphotidylinositol signal transduction pathway.

Another important group of cytoplasmic proteins, which lack
phosphokinase activity, are the proteins coded for by the ras family of
oncogenes. In light of the particular relevance of these proteins to this thesis,
this class of oncogenic proteins will be examined in a separate section

following the present general discussion of oncogene proteins.

60

4.:2 I'lll’ll'll 1

Products of several viral and cellular oncogenes are located within the
nucleus. These genes include myc, fos, jun, myb, ets and ski (Table 1). Very
little information is known about the ets and ski oncogenes. On the other
hand the myc, fos, fun, and myb oncogenes have been studied more
extensively. In general, the protein products of these oncogenes bind to
DNA. Because of this, it is believed that they are involved in the regulation
of gene expression either by activating or inactivating a particular gene or set
of genes that participate in cellular proliferation or differentiation.
Aberrations in this gene regulatory activity are thought to contribute to the
transforming potential of these genes.

This is exemplified by the c-myc gene protein. In humans this protein
has 439 amino acids and can bind to single or double stranded DNA (Persson
and Leder, 1984; Stewart et al., 1984). It is generally expressed at low levels in
quiescent cells but expression is significantly increased after stimulation by
certain growth factors or mitogens (Kelly et al., 1983). Regulated expression is
absent in several chemically transformed cells where c—myc transcription is
observed (Campisi et al., 1984).

A viral homolog to the c-myc gene has been isolated from 4 different
avian retroviruses that cause various sarcomas, carcinomas and leukemias
(Klein, 1982). As discussed previously, chromosomal translocations involving
this gene have also been associated with the causation of human Burkitt's

lymphoma. In addition to the cellular myc gene, several related genes, N-myc

61

and L-myc have been isolated from human neuroblastomas and small cell
lung carcinomas respectively (Klein, 1982).

Interestingly, the v-myc or the c-myc gene when linked to a
transcriptional promoter cannot neoplastically transform cells in culture.
Instead transfection with these genes leads to cellular immortality (Mougneau
et al., 1983). Transfection of quiescent cells with c-myc does not cause
proliferation but instead prepares the cell for other growth signals.
Consequently the myc gene is referred to as a competence gene which requires
the action of other progression genes to complete the transformation process.
This concept is illustrated in experiments conducted by Land et al. (1983) who
showed that both the activated myc and ras genes were required to transform
primary rat embryo fibroblasts in culture. Although considerable efforts have
been extended toward understanding the role the myc genes play in cellular
transformation, other than its DNA binding properties, the biochemical
function of the myc protein still remains obscure.

An interesting relationship between the protein products of the fos and
jun oncogenes has been revealed during the past several years. The fos and
jun genes were originally identified as the transforming genes of the FBI
murine sarcoma (FBI-MuSV) virus and the avian sarcoma virus-17
respectively (Curran and Teich, 1982; Maki et al., 1987). Cellular homologues
to both viral genes have also been identified (Van Beveren et al., 1983;
Bohmann et al., 1987).

Both v—fos and c-fos genes code for a 55,000 dalton protein which is
localized in the nucleus (Curran et al., 1984). The pSSC' 5 protein differs from

62

the p55V'f05 protein by 49 amino acids at the carboxy terminus and it is also
extensively post-transcriptionally modified. The fos protein, by itself, does
not appear to have any DNA binding activity in vitro. However, antibodies
to c-fos precipitate a protein complex which can bind to specific gene
transcriptional control elements (Distal et al., 1987). One protein identified
within this complex is a 39,000 dalton (p39) protein which has been identified
to be the protein product of the c-jun oncogene (SassoneCorsi et al., 1988)

Amino acid sequence analysis of the c—jun protein has revealed
significant homology with the yeast transcription factor GCN4 (Vogt et al.,
1987). This sequence homology reflects a functional similarity as both
proteins bind to the same DNA sequence, TGACTCA (Hope and Struhl, 1985;
Hill et al., 1986). The mammalian transcription factor AP-l also recognizes
this sequence suggesting that the jun protein may be associated with AP-l
(Angel et al., 1987; Lee et al., 1987). Using antibody cross reactivity reactions
together with partial peptide sequence analysis it has been established that the
c-jun protein is a major polypeptide of the AP-1 transcription complex
(Rauscher et al., 1988; Sassone-Corsi et al., 1988).

Separate in vitro translation studies with c-jun and c-fos have
demonstrated that the c-jun protein forms a homodimer which binds to the
AP-l DNA site while the c-fos protein fails to dimerize and does not bind to
DNA (Halazonetis et al., 1988). However if both genes are cotranslated, a
heterodirner is formed that binds approximately 25 times more efficiently to
the AP-l site" than does the c-jun homodimer. These experiments suggest

that the transcriptional activation of genes by AP-1 requires the coordinated

63

synthesis and interaction of the c—jun and c-fos proteins. If there was an over
expression of the c-jun gene in the absence of c-fos expression then a
predominant formation of a c-jun homodimer may occur. Because this
homodimer binds less efficiently to the AP-l binding region compared to the
jun lfos heterodirner it could cause the deregulation in the control of gene
transcription. This may be one mechanism explaining how these genes are

involved in neoplastic transformation.

BMW

1. Weaning:

The mammalian ras genes are members of a multigene family (H-ras,
K-ras, and N-ras) which have been implicated in playing an etiologic role in
most human tumors as well as in a variety of spontaneous and
experimentally induced animal tumors (Barbacid, 1987; B05, 1988). Two ras
genes, H and K, are homologous to the retroviral transforming genes of the
Harvey and Kirstin murine sarcoma viruses. The N-ras gene was isolated
from a human neuroblastoma cell line and does not have a viral counterpart
(Shimizu et al., 1983).

Sequences homologous to the mammalian ras genes have been
detected in yeast and drosophila (Shilo and Weinberg, 1981; Chang et al.,
1982). The conservation of these genes throughout eucaryotic evolution
suggests that they play an intrinsic role in basic cellular physiology.

All 3 ras genes contain a 5' non-coding exon followed by 4 coding exons

which code for a very similar 21 kDalton protein (Pulciani et al., 1982;

McGrath et al., 1983; Cichutek and Duesberg, 1986). The K—ras gene can
produce several variations of the ras protein which differ primarily at the
carboxyl terminus. These variations arise as a result of alternative mRN A
spicing patterns transcribed from two different number four exons (McGrath
et al., 1983; Capon et al., 1983, Fasano et al., 1983).

The ras proto—oncogenes can be activated by the induction of specific
point mutations within the coding exons of the gene. All of these mutations
are clustered within specific codons in exons 1, 2 and 3. Mutations within
codons 12, 13, 61 and 117 have been detected in naturally occuring human and
in chemically induced animal tumors (B05, 1989; Reynolds et al., 1987).
Earlier studies have indicated that only 10-20% of human tumors contained
activated ras genes (Pulciani et al., 1982). With the development of more
sensitive and rapid techniques of analysis, this frequency has increased
substantially for some types of tumors. For example, more than 90% of
human pancreatic tumors have been shown to contain an activated K-ras
gene (Smit et al.,1988).

In vitro mutagenesis of the ras genes has revealed several other
activating codon positions. Treatment of the H-ras gene with bisulfite
induces mutations in codons 59 and 63 that can transform transfected NIH
3T3 cells in culture (Fasano et al., 1984). Site directed mutagenesis
experiments have also indicated that changes in the coding sequence at
codons 116 and 119 have similar transforming capabilities (Walter et al., 1986;
Sigel et al., 1986). Most of these natural and synthetic mutations involve

single base pair substitutions within the specific activating codon. A double

base pair substitution has been detected in the 12th codon of the H-ras gene
form Balb-C murine sarcoma virus (Reddy et al., 1985). The normal mouse
nucleotide sequence at this position is a GGA coding for a glycine and in the
virus this codon has been mutated to an AAA which codes for a lysine. Site
directed mutagenesis experiments by Seeburg (1984) have indicated that
mutations substituting any amino acid except proline at position 12 will
activate the H-ras gene. Some mutations were observed to be more potent
than others as determined by in vitro transfection analysis into rat-1 cells.

Activation of the ras genes can also occur as a result of increased gene
expression. Multiple copies of the H-ras gene transfected into NIH 3T3 mouse
fibroblasts‘cause the morphological transformation of these cells (Pulciani et
al., 1985). Fusion of strong viral promoters to the normal H-ras gene leads to
the enhanced expression of this gene which can transform cells in culture
(Chang et al., 1982). In addition, elevated expression of the human ras genes
have been frequently detected in premalignant and malignant tumors of the
colon and to a lesser extent in the lung (Spandidos and Kerr, 1984; Pulciani et
al., 1985).

There have been conflicting reports addressing the issue as to whether
the ras genes alone are sufficient to cause neoplastic transformation. Studies
using primary rodent fibroblasts in culture have demonstrated that at least
one other oncogene is required for tumorigenic conversion by the ras genes.
Cellular immortalization appears to be an important prerequisite for this
transformation. Prior immortalization by exposure to mutagenic chemicals

or transfection with either the myc or the adenovirus EIa gene was necessary

for successful transformation of primary rodent cells by ras (Land et al., 1983;
Newbold and Overell, 1983; Ruley, 1983).

' Multiple gene activation also appears to be a requirement for the
complete transformation of diploid human fibroblasts by the ras oncogenes.
Studies by Hurlin et al. (1987) and Wilson et al. (1987) have demonstrated that
the transfection of normal human fibroblasts by the H-ras and N-ras genes
resulted in these cells acquiring several characteristics of transformed cells (ie.
focus formation and anchorage independence) but failed to exhibit an infinite
lifespan in culture and were not tumorigenic. If however an immortalized
human cell line was used in these experiments then complete neoplastic
transformation was achieved by transfection with the H-ras gene (Hurlin et
al., 1989; Wilson et al., 1989).

In vivo evidence in support of the requirement for multiple gene
activation for neoplastic transformation is provided by studies where
transgenic mice have been created carrying the v-H-ras gene linked to the
mouse mammary tumor virus promoter (Sinn et al., 1987). These mice
developed tumors in the mammary, salivary and lymphoid tissues. The
tumors that did develop in these tissues were however isolated focal lesions
of monoclonal origin indicating that additional somatic events were required
for full malignant transformation.

Contrary to the results of these studies, Spandidos and Wilkie (1984)
have reported the complete neoplastic transformation of early passage
Chinese or Syrian hamster cells by the T24 H-ras oncogene linked to

transcriptional enhancers. Transformation was dependent on both the

67

mutational activation and enhanced expression of the ras gene. Transfection
of the H-ras gene without the enhancer elements resulted in immortalization
only. These results were confirmed in a separate study using rat embryo cells
transfected with the H-ras gene (Pozzati et al., 1986).

One problem with the interpretation of these studies is that it has not
been determined whether additional oncogenes were spontaneously activated
in these cells. Consequently, it cannot definitively be established that the
activation of a single oncogene is sufficient to induce complete neoplastic
transformation. Indeed, the bulk of experimental evidence supports the
concept that carcinogenesis is a multi-step process of which one may be the

activation of the ras gene.

2. W

The ras protein can be divided into four functional domains based on
their biochemical activity and evolutionary conservation (see Figure 2). The
first domain is represented by the first 80 amino acids where almost complete
evolutionary conservation has occurred among the mammalian ras genes
(Santos and Nebreda, 1989). Located at codons 12 and 61 within this domain
are two areas where the majority of the activating mutations take place.
These two areas also encompass two of the four GTP binding sites which are
essential for oncogenic transformation by the protein (DeVos et al., 1988).
Another essential region found within this domain occurs between amino

acids 32 and 40. This region is not involved with GTP binding or hydrolysis

hOwever any amino acid changes occurring at this site eliminate the

transforming activity of the protein.

69

89 .3282 new 3.5% E0... potato Am

865m $8282 I 52.8 29885 _H_

52.8 .285 a . 58.8 .9533 I
NVFAVVF $2.-wa mmtmm . owtww .
< < e < 4 4 4 4

 

1000
2: 8e 9; 8e 2: 8 8 9. ow .
mEmEou .2285“. .n
:99:
225283 a\e 8-2 o\o 3A
:08 _ _ _ _ _ _ _ ~12
8. of 9; our 2: 8 8 9. 8 e

cozmzmmcoo 9325.05 8

ocom ow: :EEEEmE 05 do mEmEou .mcozoca can 55625 .N 9.39....

70

It has been suggested that this region may be involved in the interaction of
the ras protein with its putative cellular target and consequently it has been
named the effector domain (DeVos et al., 1988).

The second domain encompasses amino acids 81-160 where
approximately 80-90% amino acid sequence homology exists among the
various mammalian ras proteins. Within this domain there are three
essential regions ( a.a. 77-92, 109-123 and 139-165) that are absolutely necessary
for cellular transformation. The other two GTP binding regions are located
within two of these essential regions occuring between amino acids 116-119
and 143 -147.

The remaining amino acids, excluding the last 4, constitute the third
domain and are generally referred to as the hyper-varible portion of the
protein. The amino acid sequence within this domain differs for each of the
ras proteins. Most amino acid changes within this region do not appear to
effect the transforming activity of the protein.

The fourth domain is represented by the carboxy terminal four amino
acids and is designated as the CAAX box where the C is a cysteine residue, the
A can be any aliphatic amino acid and the X is any amino acid. This amino
acid sequence is highly conserved in all of the ras proteins and is present on
several other unrelated G proteins, nuclear lamins and the yeast alfa mating
factor (Clarke et al., 1988). This motif plays an important role in the

posttranslational modification of the ras proteins.

71

3.2 II I I' 1 IT I' [I] I .
The newly synthesized ras proteins are located within the cytoplasm

and are biochemically inactive. Functional activation requires the
localization of the ras proteins at the plasma membrane. Transport to the
membrane involves a series of complex posttranslational modifications.
These modifications involve proteolytic cleavage, acylation,
polyisoprenylation and palmitoylation of the carboxyl terminal portion of the
protein. The terminal CAAX motif plays a pivotal role in this process.

Early studies indicated that palmitoylation of the ras proteins was an
important process for membrane localization (Chen et al., 1985; Willumsen et
al., 1984). In these studies it was demonstrated that the replacement of the
terminal Cys136 with serine resulted in the blocking of palmitoylation and
membrane localization. These results suggested that the Cy5136 residue was
the site of protein palmitoylation and that the CAAX box was the
determinant for processing and localization of p21 ras.

Very recent studies have shown however that Cys186 is not the site for
palmitoylation and that the posttranslational processing of the ras proteins
involves a two step process. Gutierrez et al. (1989) have demonstrated that
during step 1 the initial translation product, pro-p21, is synthesized within
the cytoplasm and rapidly converted to an intermediate form, c-p21. This
conversion involves the proteolytic cleavage of the terminal AXX amino
acids and the methylation of the a-carboxyl group on Cysl36. These
modifications increase the hydrophobicity of the c-p21 compared to the pro-

p21 but the protein is still located in the cytoplasm and is not palmitoylated.

Subsequent studies by Hancock et al. (1989) have shown that the cytoplasmic
ras proteins are also polyisoprenylated during this first step and that the site

of this isoprenylation is the terminal Cys186 residue.

During the second step the palmityl groups are added and the protein is
localized at the membrane (Hancock et al., 1989). Palmitoylation of the
cytoplasmic ras proteins occurs via thioesterification of palmitic acid with
cysteine residues that are upstream of Cysl36. This takes place at Cys181 and
Cys184 on H-ras, Cys181 on N—ras and presumably at Cys180 on the viral K-ras
proteins. The p21 K(B)-ras mammalian protein is not palmitoylated because
it lacks cysteine residues upstream of Cys‘86. Since the K-ras protein is not
palmitoylated it must not be an essential modification for protein activity
because this protein still has transforming activity. All ras proteins are
therefore isoprenylated but not necessarily palmitoylated. Hancock et al.
(1989) did show that palmitylation increases the avidity of membrane binding
as well as the transforming activity of the ras proteins.

The importance of polyisoprenylation for ras function is demonstrated
in an experiment in which polyisoprenoid biosynthesis is inhibited by
treatment of cells with mevinolin (Hancock et al., 1989). This blockage of
polyisoprenylation abolished p21 membrane binding. Because membrane
binding is essential for the biological activity of the ras proteins,

polyisoprenylation is essential for cellular transformation.

73

4.111.] 'll"l [I] I'

Extensive analysis of the protein products of the proto-oncogenic and
oncogenic ras genes has yielded very little functional information in regards
to their role in normal or neoplastic cellular processes. In spite of this lack of
functional information, a considerable amount of data has been generated
concerning the structure and biochemistry of these proteins.

The ras proteins share many biochemical properties with a class of
proteins, commonly referred to as the G-proteins, which are involved in
membrane receptor mediated signal transduction pathways. These
similarities suggest that the ras proteins may also function in this capacity and
that aberrations in this process may explain their transforming activity.

Like the G-proteins, the ras proteins have the ability to bind, exchange
and hydrolyze guanine nucleotides (Scolnick et al., 1979; Hurley et al., 1984).
Regulation of the biochemical activity of the G—proteins involves the cycling
between the inactive GDP-bound form and the active GTP-bound form
(Gilrnan, 1984). A similar mode of activation and inactivation is thought to
occur with the ras proteins. It has been hypothesized that the transforming
activity of the mutant ras proteins may be the result of abnormally high
levels of p21 in the active GTP-bound state (Gibb et al., 1984). These mutations
could induce high levels of GTP-bound p21 by either altering the GTP binding
affinity of the molecule or by inhibiting the proteins intrinsic GTPase activity.

Structural studies coupled with mutational analysis have determined
that the binding sites for GTP are centered around the transformation

activating ,codons 12 and 61 (DeVos et al., 1988). Mutations at these codon

74

positions however do not appear to manifest their transforming potential by
effecting GTP binding because both the normal and transforming ras proteins
bind GTP with similar affinities (Finkel et al., 1984).

Comparative biochemical analysis between normal and the activated
p21 protein has revealed significant decreases in the GTPase activity of
various mutant ras proteins (Gibbs et al., 1984; McGrath et al., 1984; Manne et
al., 1985). A correlation between the decrease in GTPase activity of these
proteins with in vitra and in viva transforming potential however cannot be
established (Lacal et al., 1986; Trahey et al., 1987a).

The intrinsic GTPase activity of the normal p21 protein is very weak
under normal physiologic conditions and is insufficient to maintain the ras
protein in the inactive GDP state (Trahey and McCormick, 1987b). Recently a
protein has been isolated which associates with p21 and stimulates its GTPase
activity (Trahey and McCormick, 1987b). This GTPase activating protein
(GAP) catalyzes the conversion of p21-GTP to p21-GDP at a rate which is
several hundred fold greater than the intrinsic rate. It is of particular
significance that this protein does not have a stimulatory effect on the GTPase
activity of ras proteins with mutations in codons 12, 59 or 61 (Adari et al.,
1988). Because many of these mutant proteins already have reduced GTPase
activity, their in viva GTPase activity when coupled with GAP is several
orders of magnitude lower than the wild type p21.

Functional mutation studies have identified a region of the p21 protein
located between amino acids 32 and 40 as a possible interaction site between

the ras protein and its putative cellular target (DeVos et al., 1988). Mutations

75

within this region do not effect GTP binding, GTPase activity or membrane
localization but is essential for ras transforming activity. This area has been
termed the ras effector region.

Recent studies have tested the ability of various mutant ras proteins to
interact with GAP (Adari et al., 1988). Proteins with mutations in non-
essential regions were stimulated by GAP but proteins with mutations in GTP
binding regions or in the effector region lost their ability to respond to GAP.
It was concluded that the ras effector region is the site for GAP binding and it
has been suggested that GAP is the ras effector protein (McCormick, 1989).

A model has been proposed by McCormick (1989) which postulates
GAP as being both an effector and regulator of ras protein activity. In this
model membrane bound p21-GDP is converted to p21-GTP as the result of
some upstream stimulatory signal (for example it could be initiated by a
growth factor). This nucleotide exchange results in the formation of a p21-
GTP/ GAP complex. This complex may serve to position GAP at a specific site
in the membrane where GAP can further interact with some other cellular
factor. Thus, the input signal is propagated through a ras-GAP complex
where GAP serves as the effector transmitting the stimulatory signal to the
next component in the signal cascade. The duration of the signal is limited
because p21-GTP is quickly converted to p21-GDP by GAP resulting in the
disassociation of GAP from the ras protein. GAP therefore participates in the
signal transmission (as an effector) and at the same time (as a regulator) down
regulates the input signal . Transforming ras proteins still associate with GAP

but the output signal is no longer down regulated because the ras GTPase is

76

not stimulated by GAP. GAP remains associated with ras and is properly
positioned to participate in the signal cascade. Consequently, an unregulated,
constitutive output signal is propagated in cells carrying mutant ras genes
contributing to the transformed phenotype.

EC] '1: . '3']! I

1. Methndsmedinriskassessmcnt

The extensive use of synthetic chemicals by industrialized societies in
conjunction with the possible carcinogenic hazard associated with exposure to
these chemicals has created the need to predict and evaluate their
carcinogenic potential. This process currently relies on data generated from a
a number of experimental methodologies because no single test can
equivocally determine a chemical's carcinogenic potential. This usually
includes the use of human epidemiological studies in conjunction with
various in vitra and in viva test systems. Data generated from all of these
sources are assimilated to establish the carcinogenic potential of a particular
chemical. It is important to recognize that even with the use of this multi-
faceted approach the evaluation of a chemical's carcinogenic potential is still
an approximation. This section will discuss the approach generally used to
make this assessment and why the interpretation of results from these assays
provides only an approximation rather than an absolute determination of

carcinogenic risk.

a.) Epidemiological analysis in risk assessment
A fundamental objective of the cancer epidemiologist is to study the

patterns of human cancers in an attempt to elucidate the various factors
associated with its cause. This is basically achieved by describing the
occurrence of cancers in relation to differences in a number of factors such as
sex, age, race, geographical location, occupation and socio—economic class to
determine a common parameter associated with the disease outcome. The
utility of this approach has been validated by a number of studies correlating
specific human cancers with exposure to a specific chemical agent. These
studies have provided great insight into the various causes of cancer which
has been beneficial in identifying and eliminating exposure to a number of
hazardous materials.

Perhaps one of the most successful examples of how epidemiology can
contribute to the identification of a human carcinogen is provided by the
studies which have identified cigarette smoking as an important etiologic
agent in the development of lung cancer (Doll and Hill, 1950). An
examination of cancer death statistics revealed an enormous increase in the
lung cancer incidence in males, but not in females, between 1920 and 1944.
This suggested that some exclusive activity of males was the cause of this
increase. After careful analysis of a large volume of data the common factor
of cigarette smoking was identified. One of the difficulties in identifying
cigarette smoking as the etiologic factor for the increase in male lung cancers
was the long latency period between when the individual began smoking and

the onset of disease. This latent period averaged around 20 to 30 years from

78

the time of initial exposure to the appearance of a clinical tumor. The use of
epidemiological analysis not only identified the causative agent for the
increase in this particular cancer but it also provided an important insight
into the mechanism of chemical carcinogenesis.

Once the causative agent was determined, it was possible to initiate
studies to identify the specific chemicals within tobacco smoke that were
responsible for inducing the cancer. Many of these chemicals have been
isolated and identified and include a number of polycyclic aromatic
hydrocarbons such as benzo(a)pyrene (IARC Monograph, 1982).

Using an approach similar to that used in the smoking and lung cancer
studies, epidemiologists have identified a number of other chemicals as
human carcinogens. This list includes such industrial chemicals as benzene,
4-aminobiphenyl, asbestos, benzidine, cyclophosPhamide and vinyl chloride
in addition to several pharmaceutical agents such as diethylstilbestrol and
other conjugated oestrogens (IARC Monograph, 1982).

In spite of the utility of using epidemiological studies to evaluate
human chemical carcinogenicity there are several important limitations to
this approach (MacMahon, 1979; Muir, 1979). One of the most serious
limitations is that epidemiology is primarily an observational science.
Conclusions depend on collecting and comparing a large number of
observations and then finding a common parameter associated with a
particular disease state. It is often a very difficult task to accomplish this with

a high degree of statistical and biological confidence.

79

These analyses are dependent on the availability of accurate personal
records (ie medical, occupational, family histories etc.) to determine cause and
effect relationships. Consequently inadequate records can severely hinder the
interpretive conclusions by the investigator.

Achieving a high degree of statistical confidence usually necessitates
the analysis of fairly large population sizes. In many cases the exposed
populations are very small and it is often very difficult to ascribe any
statistically significant increase in tumor development with exposure. Also it
often requires taking multiple measurements which is rarely possible.

A greater degree of confidence is obtained if the disease is relatively-
uncommon and the exposure causing the disease is unusual. This is
illustrated by the epidemiological study associating the development of
angiosarcoma of the liver with the exposure to vinyl chloride (Doll, 1988).
Because the incidence of this disease is quite rare and the occurrence was
confined to a specific occupational exposure group, a great deal of confidence
was obtained in assigning the increase in this cancer with exposure to vinyl
chloride. It is more difficult to correlate the increase in a disease with
exposure when the the disease is fairly common and the natural incidence
fluctuates with factors such as time, geographic location and socio-economic
status (Muir, 1979).

All of these factors limit the ability of the epidemiologist to evaluate
the carcinogenic potential of any chemical agent. This is reflected in the fact
that of the several hundred chemicals which have been assigned animal

carcinogens less than 25 have been confirmed as human carcinogens by

epidemiological analysis (IARC Monograph, 1982). Consequently our risk
evaluation process has evolved to include an experimental approach to assist

in the determination of human chemical carcinogenic potential.

b.) Shag-m tgsts fgr thg gvaluatign 9f chemical sarg’ngggnig
mtgntial

The retrospective nature of epidemiological analysis severely limits the
utility of these studies for predictive human carcinogenic risk evaluation.
The animal bioassay was developed to provide a prospective experimental
method to assess carcinogenic risk. These assays are, however, very time
consuming and expensive to conduct. This has necessitated the development
of alternative methods to screen for the thousands of chemicals with
potential carcinogenic activity. To assist in this task a number of short-term
tests for mutagenicity have been developed. The rationale for the use of a
mutagenicity test as a predictor of carcinogenicity is based on the
preponderance of data implicating mutagenicity with carcinogenic activity. In
addition, the speed, relatively low cost and high predictive value has added to
the appeal of these test systems and as a result they have become an
important component in carcinogenicity evaluation.

Because of specific interpretive limitations with any one assay, the
evaluation process has evolved into the use of a battery of test systems. These

testing protocols ideally would include assays to measure mutagenicity in

81

bacteria and mammalian cells in vitra and gross chromosomal damage both
in. vitra and in viva (Wiliams et al., 1980).

Of the over 100 short-term tests that exist the Ames Test is the most
well known and widely used. This test assesses chemical mutagenicity using
several specially developed strains of Salmonella bacteria (Ames et al., 1973).
Each strain has been constructed to detect a chemical's potential to induce
either base pair substitutions or frameshift mutations. Because many
chemicals require bioactivation to their active mutagenic form, the test is
conducted with and without a rat liver preparation that is rich in
biometabolizing enzymes (59 mix).

The bacteria themselves are auxotrophic mutants which cannot
synthesize histidine. The basis of the test is to determine whether a chemical
can induce mutations causing a reversion of the mutant bacteria to a wild
type phenotype that can then be selected for on minimal media plates.

This test detects most of the known organic human chemical
carcinogens with the exception of benzene and diethylstilbesterol (IARC,
1980). An extensive analysis has been conducted to evaluate the predictive
value of Ames tests results with animal bioassay carcinogenicity data
(McCann et al., 1975 ). Of 254 positive carcinogens tested there was a 90%
correlation between mutagenicity and carcinogenicity. In general, the metal
carcinogens, hormones and nongenotoxic carcinogens were not detected in
the test system. In the evaluation of noncarcinogens, no mutagenic activity

was detected in 46 of 46 tested chemicals.

82

Because mammalian cell DNA is more complex than bacterial DNA, in
vitra mammalian mutagenicity assays are also conducted. The basic design of
these test systems is in principle very similar to the Ames test. The objective
of these tests is to determine the ability of the chemical to induce mutations
in a specific genetic marker in mammalian cells in culture. These markers
typically utilize a gene involved in an easily selected for biochemical
metabolic pathway . Two of the more commonly used systems are designed
to detect mutations in either the hypoxanthine guanine phosphoribosyl
transferase (HGPRT) gene or the thymidine kinase (TK) gene (Williams et al.,
1980). .

Some chemicals damage genetic material at the level of the
chromosome and consequently would not be detected as mutagens using the
assays described above. Chromosomal damage can be classified in two general
categories (IARC, 1980). One involves the whole chromosome and the other
involves the individual chromatids. Whole chromosomal aberrations occur
during the G1 phase of the cell cycle prior to DNA replication. This damage
involves breakages that can lead to gaps, terminal and interstitial deletions,
inversions, centric and acentric ring formation, and asymmetric interchanges
that can produce reciprocal translocations.

Chromatid aberrations involve damage to the single chromatid which
can occur at any time during or after the replication of the chromosome. A
wider variety of damage can be observed with this type of aberration

compared to the whole chromosome. Both in viva and in vitra tests have

83

been developed to evaluate a chemical's potential to induce either type of
damage.

The in vitra systems typically use peripheral blood lymphocytes or
fibroblasts for analysis. Some of the advantages of the in vitra systems are
that they can be conducted with human cells and that the concentrations of
test material can be much higher than would be possible with in viva studies.
A major disadvantage is that the cells in culture may be unable to metabolize
the chemical to its mutagenic form. This is circumvented by the addition of a
metabolizing system similar to what is used in the Ames test.

The use of in viva systems is very desirable because the whole animal
is exposed to the test material which allows the influence of bioavailability
and metabolizing functions on mutagenicity to be evaluated. This makes it
possible to determine quantitative measurements of dose response which is
very important for risk assessment analysis (Williams, 1980). In viva analysis
also allows for the analysis of a wider variety of cell types which can include
peripheral lymphocytes and cells of the bone marrow, liver, gonads etc.
(IARC, 1980).

Genetic endpoints for clastogenic in viva and in vitra test systems
include the cytological chromosomal abnormalities discussed above, the
presence of micronulcei, and the determination of sister chromatid
exchanges. Endpoints restricted to the in viva systems would include
dominant lethal mutations and heritable translocations within the germ cells

(IARC, 1980). One of the major disadvantage of these last two endpoints is

that they are time consuming and labor intensive and as a result studies to
evaluate them are very expensive to conduct.

'An obvious major limitation for the evaluation of chemical
carcinogenicity using the short-term tests described above is that these assays
are determining mutagenic potential rather than the ability to cause cancer.
Even though there is a strong relationship between mutagenic potential and
carcinogenic potential, the correlation is not 100%. In spite of these
limitations the results from short-term assays continue to play an important

role in the evaluation of a chemicals carcinogenic potential.

c.) Thg animal bigassay

Data generated from epidemiological and mutagenicity studies are
important contributions in the evaluation of a chemical's carcinogenic
potential. The results from the rodent bioassay for carcinogenicity however
continues to have a major influence in this decision making process. The
general premise of the long-term bioassay is to expose animals to a specific
chemical for the majority of their lifespan and determine whether this
exposure results in the increased incidence of neoplastic lesions. These
studies are a time-consuming and very costly endeavor and it is important
that a degree of uniformity be achieved in their conduct so results from
different laboratories will have comparable interpretations. To achieve this
objective various national and international committees have convened to
adopt protocol guidelines for the conduct of these studies (WHO, 1969, 1978;
FDA, 1971).

i.) Bssic condugt

The rodent bioassay, as it is currently conducted by the National
Toxicology Program under the auspices of the National Institute of
Environmental Health Sciences, was adopted during the early 1970's and has
been used to evaluate over 300 chemicals for potential carcinogenicity. These
studies are presently concluded according to international guidelines and
conform to the principles of Good Laboratory Practices (Federal Register,
1978). These guidelines specify that males and females from two rodent
species will be exposed to the test material for a 24 month period. The route
of administration of the test material will be chosen to reflect the probable
route of human exposure. Experimental groups will consist of a control
group receiving no test material, a high dose group in which the exposure
concentration is set at the maximum tolerated dose (MTD) and a low dose
group where the exposure concentration is 1 / 2 the MTD. Each group is
composed of 50 animals and the species primarily used are the B6C3F1 mouse
and the F344 rat. At the termination of the study a chemical will be
considered positive for carcinogenicity if a statistically significant increase is
observed in the overall tumor incidence or in a specific tumor type in either

sex in at least one of the two test species.

ii) Problems with interpretation

The development of animal tumors in the bioassay does not
definitively establish a chemical as a human carcinogen. A number of factors

complicate the interpretation and extrapolation of bioassay results. The

specific influence that many of these factors have on the eventual outcome of
bioassay results is not completely understood at this time. This makes the
evaluation of human carcinogenic risk a difficult and uncertain task.

I For example, in many bioassays there is a difference in the
susceptibility to develop tumors between the 2 rodent species used. This is
illustrated in an examination of 226 chemicals that were positive for tumor
induction in the bioassay. Of these 226 bioassays, 42% (96 of 226) produced
tumors in the mouse and not in the rat or vice versa (Bernstein et al., 1985).
It is difficult to assign a carcinogenic risk in humans when such a disparity
exists in the tumor response of two so closely related species.

Metabolic and pharmacokinetic differences between rodents and higher
mammals can also complicate the extrapolation of bioassay data. This is
particularly important for the many compounds that must be biotransformed
to a reactive carcinogenic intermediate. Differences in uptake and
elimination rates can result in the saturation of primary detoxifying
enzymatic pathways in one species and not in another. This can lead to the
activation of secondary routes of metabolism which can produce reactive
intermediates. These secondary metabolic intermediates may then interact
with important macromolecules, such as DNA, and initiate neoplastic
development.

This problem is illustrated in rodents which tend to use glutathione
rather than epoxide hydrase for detoxification of xenobiotics. Man and
primates utilize glutathione only as a reserve after epoxide hydrase activity

has been saturated (Pacifici et al., 1981). Consequently at high exposure

87

concentrations intracellular glutathione is depleted early in the detoxification
process making rodents potentially more vulnerable to the generation of
various oxidants and free radicals. These differences in metabolic routes of
detoxification can influence the sensitivity of a species to develop tumors
after exposure to a particular chemical. Such species differences must be
considered when extrapolating bioassay results to assess human carcinogenic
risk.

The experimental design of the bioassay is also a contributing factor
effecting the interpretation of results. As indicated previously, the protocol
for the rodent bioassay calls for the exposure of a relatively small number of
animals to very high doses of the test material. The use of a small number of
animals is based on logistic and economic reasons and the use of high doses is
to maximize the probability of inducing a biologic effect. As was previously
discussed, high exposure concentrations may alter the metabolic fate of the
test chemical and produce a response that would not be observed at the much
lower exposure levels to which humans typically would be exposed.

Many of the inbred strains of rodents have high spontaneous tumor
incidences. For example nearly 100% of F344 male rats will develop
spontaneous testicular tumors within two years (WHO, 1980). The Sprague-
Dawley rat and the A / I mouse have high spontaneous incidences of
mammary and lung tumors respectively. One of the more controversial
factors complicating the evaluation of bioassay results involves the relevance
of an increase in a normally high spontaneous tumor incidence to the

potential human risk. This problem is particularly relevant with the B5C3F1

mouse because of its extensive use in many bioassays. Males of this strain
have a characteristically high spontaneous liver tumor incidence which
averages around 30% in control p0pulations but this incidence can vary from
10%-50% (Chan, 1984).

In manylbioassays involving the 86C3F1 mouse an increase in this
Mar incidence is often observed following treatment with the test
compound. This is illustrated in an examination of 85 positive bioassays
conducted by the National Cancer Institute between 1977 and 1980 involving
the B5C3F1 mouse and either the F344 or Osborne Mendal rat (Hamm, 1984).
Of the 85 chemicals that were carcinogenic in at least one sex of one species,
59% (50/85) caused liver tumors in the B6C3F1 mouse whereas only 22%

(19/ 85) produced liver tumors in the rat. In addition, 24% (20/ 85) of the
chemicals caused only liver tumors in the mouse and no tumors of any type
were observed in either rat strain. The significance of the increase in only
mouse liver tumors to a potential human carcinogenic risk from exposure to
these chemicals is presently a topic of considerable debate. Numerous task
forces have been assembled in an attempt to resolve this issue with little
success.

Because of the preponderance of data associating mutations with
neoplastic development there is less controversy interpreting the significance
of B5C3F1 mouse liver tumors induced with genotoxic chemicals. Since these
agents are also likely to produce mutations in humans there is greater
confidence in classifying them as potential human carcinogens. However, an

increase incidence in the B6C3F1 mouse liver tumors has also been observed

89

for a number of chemicals that appear to have little if any mutagenic activity.
For example, the chlorinated hydrocarbons tetrachloroethylene,
trichloroethylene, tetrachloroethane and chloroform are generally negative
in various short-term tests for genotoxicity yet all cause a significant increase
in B6C3F1 mouse liver tumors (Environmental Health Criteria, 1985;
Reichert, 1983). The debate concerning the significance of B6C3F1 mouse liver
tumors to potential human carcinogenic risk is currently centered around
bioassays involving compounds of this type. To accurately evaluate human
carcinogenic risk from nongenotoxic chemicals it is important to
understanding the mechanism by which these spontaneous tumors arise and

how chemical agents enhance this tumor response.

90

GO ll'l' 'BCEM I' I

1.) MW

The detection of activated cellular oncogenes in a variety of human
tumors suggested that cellular oncogenes might also be associated with the
induction of spontaneous liver tumors in the B5C3F1 mouse. The presence of
dominant acting oncogenes can be detected using cell mediated DNA transfer
techniques (Shih, 1979). To determine whether spontaneous 86C3F1 mouse
liver tumors contained dominant cellular oncogenes several investigators
transfected DNA isolated from these tumors into NIH 3T3 cells (Fox and
Watanabe, 1985; Reynolds et al., 1986). These transfection experiments
resulted in the development of morphologically transformed foci indicating
the presence of a dominant oncogene(s) (Table 3). The percentage of animals
whose tumor DNA induced transformed foci ranged from 56% to 82%
suggesting that the activation of cellular oncogenes was an important event
in the development of these spontaneous tumors. The B6C3F1 mouse is a
hybrid strain resulting from the cross between a C3H/He] male and a
C57BL/6I female. The C3H/ He] male has an even higher spontaneous liver
tumor incidence than does the B6C3F1 male mouse (Graso and Hardy, 1975).
This suggests the B6C3F1 mouse may have a genetic predisposition to the
development of a high spontaneous liver tumor incidence. This genetic

factor could be an activated cellular oncogene.

91

Table 3. Transforming genes in spontaneous hepatocellular tumors
of the B5C3F1 mouse.

DNA source Frequency of Transformation
positive efficiency,
transforming tumors foci per ug of DNA

Spontaneous liver tumors 3 9 of 11 (82%) 0.005 - 0.020
(24 months)
Spontaneous liver tumors b 13 of 23 (57%) 0.012 - 0.037

(24 months)

a Fox and Watanabe. 1985

b Reynolds et al., 1986

92

However, when DNA isolated from the normal tissue surrounding the
liver tumors was transfected into the NIH 3T3 cells no foci were detected (Fox
and Watanabe, 1985). This indicates that the genetic predisposition did not
involve a dominant acting oncogene. It also indicates that the oncogene
detected in the spontaneous tumors was the result of a somatic event
occurring within the individual liver cells.

An activated H-ras oncogene was identified in the majority of the
tumors that were positive in the NIH 3T3 transfection assay. The frequency
of H-ras gene activation within these tumors was determined to be 66% (4 of
6) (Fox et al., 1987) and 88% (15 of 17) (Reynolds et al., 1987) in two
independent studies (Table 4). This data indicates that the H-ras oncogene is
frequently activated in 86C3F1 mice spontaneous liver tumors.

Other dominant oncogenes apparently were also activated in the
spontaneous liver tumors because not all of the NIH 3T3 cell transformants
contained an activated H-ras gene. Studies by Reynolds et al. (1987) found
that the raf oncogene was activated in at least one tumor and an unidentified
oncogene(s) appeared to be activated in three other tumors. In addition, DNA
from several spontaneous liver tumors did not cause the transformation of
NIH 3T3 cells suggesting that they did not contain an activated oncogene.
There are however many oncogenes which cannot be detected using the NIH
3T3 transformation system and therefore the possibility exists that one of

these oncogenes may be activated in those tumors.

93

Table 4. Oncogene activation in B5C3F1 mouse liver tumors

 

Treatment Number of Oncogene activated
used to induce positive
tumors tumors in NIH
3T3 assay
H-ras K-ras raf unknown
None a,d 23 19 0 1 3
(spontaneous)
N-HO-AAF b 7 7 0 0 0
V--CARb 7 7 0 0 0
HO-DHE b 12 10 2 0 0
DMN C 14 14 0 0 0
Furfural ‘1 13 10 2 1 0
Furan d 13 9 l 0 3
a) Fox et al., 1987
b) Wiseman et al., 1986
c) Stowers et al., 1988

d) Reynolds et al., 1987

94

2.)!I'I' [ll] .1 .11.: 132E
museJixemenrs

To determine if and to what extent the H-ras gene is activated in
chemically induced B6C3F1 mouse liver tumors, Wiseman et a1 (1986) treated
12 day old .mice with a single i.p. injection of either N—hydroxy-2-
acetylaminofluorene (N -HO-AAF), vinyl carbamate (VC) or 1'-hydroxy-2',3'-
dehydroestragole (HO-DHE). Multiple hepatomas developed within 7-9
months without any additional treatment.

A high frequency of 3T3 cell transformation was exhibited by DNA
isolated from these tumors (7 of 7 N-OH-AAF, 7of 7 VC, 11 of 11 HO-DHE)
indicating the presence of an activated oncogene. The H-ras gene was
identified as the transforming oncogene in the DNA from 24 of the 26
hepatomas (Table 4). Selective oligonucleotide hybridization analysis
revealed that the activation of the H-ras gene in these tumors was the result
of a single base pair substitution within codon 61 (Table 5).

These three chemicals are potent genotoxic carcinogens whose
electrophilic metabolites react with specific sites on DNA (Miller et al., 1985;
Lai et al., 1985; Fennell et al., 1985). Analysis of the particular base pair
substitutions occuring at codon 61 has indicated that each chemical produced
a distinct pattern of H-ras mutations (Table 5).

Studies in E. cali suggest that the primary DNA adduct produced by N-
HO-AAF exposure is the N-(deoxyguanosin-8-yl)-aminofluorene (Lai et al.,
1985).

95

Table 5. Mutation spectrum in codon 61 (CAA) of the H~ras gene in
B5C3F1 mouse liver tumors.

 

Treatment used to Number of Codon 61 mutation spectrum
induce tumors tumors with
codon 61
mutations
AAA CGA CTA
None (spontaneous) a 15 9 3 3
N-HO-AAF b 7 7 0 0
V-CAR b 7 0 1 6
HO-DHE b 10 0 5 5
DEN C 14 7 3 4
Frufural a 6 5 1 0
Furan a 5 4 1 0
a Reynolds et al., 1987

b Wiseman et al., 1986
C Stowers et al., 1988

96

This adduct should give rise to CG => A-T transversions which was the only
base pair substitution observed at codon 61 in the liver tumors induced with
this compound.

The base substitutions detected with HO-DHE and VC were A-T => T-A
transversions and AT => GC transitions at the second nucleotide position
within codon 61 (Table 5). These particular nucleotide substitutions were
somewhat surprising considering the types of adducts observed in other
studies with these compounds. Analysis of DNA base adducts in mouse liver
after a carcinogenic dose of HO-DHE revealed that greater than 90% were at
the N2 position of deoxyguanosine (Fennell et al.,1985). Data for vinyl
carbamate adduct formation is not available but studies with vinyl chloride, a
closely related carcinogen, indicates that binding occurs primarily at the N7
position of deoxyguanosine residues (Barbin et al.,1984). Interpretation of the
mutagenic specificity of these compounds is complicated by the fact that
analysis is restricted to a small number of codons (12, 13, 61, 117). In addition,
not all of these mutations are equally selected for since they do not have the
same efficiency of inducing morphological transformation (Reynolds et al.,
1987). Therefore other mutations may be present at additional sites but they
are not available for analysis since they cannot be selected.

Even though the specific base substitutions in the tumors were
different than what would have been predicted, each of the three chemicals
exhibited a unique mutational spectrum (Table 5). This suggests that these

agents activated the H-ras gene as the result of a genotoxic interaction.

97

This is even more evident when these mutational spectra are compared to
the mutational spectra observed in spontaneous tumors (Table 5). As can be
seen in the table, the frequency distribution of codon 61 mutations of the
spontaneous tumors is distinct from those of the chemically induced tumors.

In a similar study, Reynolds et al. (1987) induced tumors in the B6C3F1
mouse liver by chronic exposure to either furan or furfural. As was observed
by Wiseman et al. (1986) the majority of the tumors contained an activated H—
ras gene (Table 4). Their results differed from that of Wiseman et al. (1986) in
that not all of the activating mutations were confined to codon 61. Some of
the tumors contained mutations at codons 13 and 117 (Table 6). These
compounds were determined to be nonmutagenic based on the results of the
Ames bacterial mutagenicity test (Reynolds et al., 1987). However, because
mutations were detected at codon positions not observed in spontaneous
tumors Reynolds suggests that these agents are genotoxic under in viva
conditions.

One of the problems with evaluating potential human risk from
nongenotoxic chemicals is that they can induce animal tumors without any
apparent in vitra mutagenic activity. It is important for the risk assessment
process to determine with a high degree of confidence whether these agents
are truly nongenotoxic in the animal. Analysis of the mutational spectrum
within the H—ras gene of liver tumors induced with a particular chemical may
provide a sensitive method to evaluate the presence or absence of in viva

genotoxicity.

98

Table 6. Distribution of activating codon mutations within the H-ras
gene from B5C3F1 mouse liver tumors.

 

Treatment used Number of tumors Codons activated
to induce analyzed
tumors
61 13 117
None a ' 15 15 0 0
(spontaneous)
N-HO-AAF b 7 7 0
VCAR b 7 7 0 0
HO-DHEb 10 10 0 0
DMN C_ 14 14 O O
Furfural a 9 5 0 4
Furan a 9 6 2 1
3‘ Reynolds et al., 1987

b Wiseman et al., 1986
C Stowers et al., 1988

This would eliminate some of the uncertainty concerning the interpretation
of these bioassays in regard to assignment of potential human risk.

The basic principle underlying this type of analysis is the comparison of
the mutational spectrum within the H-ras oncogene from spontaneous
tumors with that of the chemically induced tumors. This comparison
involves the determination in the frequency distribution of base pair
substitutions within codon 61 (intracodon analysis) and also identifying
which codons within the H-ras gene are activated (intercodon analysis).
Statistically significant deviations, in either one of these parameters, from
that observed in spontaneous tumors would be an indication of possible
genotoxic activity. However, as can be seen in Table 5, diethylnitrosamine a
potent genotoxic carcinogen, has a mutational spectrum in codon 61 that is
very similar to that observed in spontaneous tumors. Consequently not all
genotoxic agents elicit a unique mutational spectrum.

If the spontaneous tumors are to be used as a benchmark, then the
mutational spectrum of these tumors must be determined with a high degree
of confidence. Previously only 15 spontaneous tumors have been analyzed in
this manner (Reynolds et al., 1987). Analysis of additional tumors will be
required to achieve the statistical confidence necessary to make the
appropriate comparisons. In addition, to determine whether nongenotoxic
carcinogens have cryptic mutational activity, mutational analysis of the H-ras
gene in tumors induced with these types of chemicals needs to be conducted.

These studies should increase the understanding of how genotoxic and

100

nongenotoxic agents activated the H—ras gene in these tumors which should

aid the interpretation of bioassay data using this strain of mouse.

LIST OF REFERENCES

101

References

Adari, H., Lowy, D. R., Willumsen, B. M., Der, C. 1., McCormick, F., (1988),
Guanosine triphosphatase activating protein (GAP) interacts with the p21 ras
effector binding domain, Science, 240: 518-521.

Alitalo, K., Schwab, M., Lin, C. S., Varmus, H. E., Bishop, 1. W., (1983),
Homogeneously staining chromosomal regions contain amplified copies of
an abundantly expressed cellular oncogene (c-myc) in malignant
neuroendocrine cells from a human colon carcinoma, Proc. Natl. Acad.Sci.

U.S.A., 80: 1707-1711.

Andervant, H. B., (1950), Studies on the occurrence of spontaneous
hepatomas in mice strains C3A and CBA,1., Natl. Cancer Inst., 11: 581-592.

Angel, P., Imagawa, M., Chiu, K, Stein, B., Irnbra, R. 1., Rahmsdorf, H. 1.,
Ionat, C., Herrlich, P., Karin, M., (1987), Phorbol ester-inducible genes contain
a common cis element recognized by a TPA-modulated trans-acting factor,
Cell, 49: 729-739.

Angel, P., Poting, A., Mallick, U., Rahmsdorf, H. 1., Scharpp, M. and Herrlich,
P., (1986), Induction of metallothionein and other mRN A species by
carcinogens on tumor promotors in primary human skin fibroblasts. Mol.
Cell Biol., 6: 1760-1766.

Barbicid, M., (1987), ras Genes, Annul. Rev., Biochem., 56: 779.

Barbin, A., Besson, F., Perrard, M-H., Bereziat,1-C., Kaldor,1., Michel, G.,
Bartsch, H., (1985), Induction of specific base-pair substitutions in E. cali type
A mutants by chloroethylene oxide, a carcinogenic vinyl chloride metabolite,
Mutation Research, 152: 147-156.

Bargmann, C. 1., Hung, M.-C., and Weinberg, R.A., (1986), Multiple
independent activation of the neu oncogene by a point mutation altering the
transmembrane domain of p185, Cell, 45:649-657.

Becker, F. F., (1982), Morphological classification of mouse liver tumors based
on biological characteristics, Cancer Research, 42: 3918-3923.

Berenblum, I., (1929), Tumor formation following freezing with carbon
dioxide snow, Brit. 1. Exper. Pathol., 10: 179-184.

102

Bernstein, T. Gold, L. S., Ames, B. N., Pike, M. 5., Hoel, D. G., (1985), Some
tautologous aspects of the comparison of carcinogenic potency in rats and
mice, Fundam. Appl. Toxicol., 5: 79-86.

Bishop, 1. M., (1983), Annu. Rev. Biochem., 52: 301-354.
Bishop, 1. M., The Molecular Genetics of Cancer, Science 235: 306-311.

Bishop, 1.M., (1983), Cellular oncogenes and retroviruses, Annu. Rev.
Biochem., 52: 301.

Bister, K. and Deusberg, RH, (1982), in Advances in Viral Oncology (ed.)
Klein, G. Raven Press, New York, 3-42.

Bohmann, D., Bos, T. 1., Admon, A., Nishimura, T., Vogt, P. K. and T. Jan,
(1987), Human proto-oncogene c-jun encodes a DNA binding protein with
structural and functional properties of transcription factor AP-l, Science, 238:
1386-1392.

Bonfiglio, T. A. and Stoler, M. H., (1988), Human papillomavirus and cancer
of uterine cervix, Hum. Pathol. 19: 621.

Bos, 1. L., (1989), ms Oncogenes in human cancer, a review, Cancer Research,
49: 4682-4689.

Bos,1., (1988), The ras gene family and human carcinogenesis, Mutation
Research, 195(3): 255-271.

Boutwell R.K., (1964), Some biological aspects of skin carcinogenesis, Prog.
Exp. Tumor Res., 4: 207.

Bowen-Pope, O. F., Vogel, A., and Ross, R., (1984), Production of platelet-
derived growth factor-like moleculas and reduced expression of platelet-
derived growth factor receptors accompanying transformation by a wide
variety of agents, Proc. Natl. Acad. Sci. U.S.A., 81: 2396-2400.

Brodeur, G. M., (1986), Clinical implications of oncogene activation in human
neuroblastomas, Cancer, 58: 541.

Brodeur, G. M., (1986), Molecular correlates of cytogenetic abnormalities in
human cancer cells: implications for oncogene activation, Prog. Hematol., 16:
229.

103

Butterworth, BE. and Slaga, T.1., (1987) Banbury Report 25: Nongenotoxic
Mechanisms in Carcingenesis, Cold Spring Harbor.

Campisi,1., Gray, H. E., Pardee, A. B., (1984), Cell-cycle control of c-myc but no
c-r_as expression is lost following chemical transformation, Cell, 36: 241 -247.

Campisi,1., Gray, H. E., Pardee, LA. B., Dean, M., Sonenshein, G. E., (1984), Cell
cycle control of c-myc but not c-ras expression is lost following chemical
transformation, Cell, 36: 241-247.

Capon, D. 1. Chen, E. Y., Levinson, A. D., Seeburg, P. H., Goedel, D. V., (1983),
Complete nucleotide sequences of T24 human bladder carcinoma oncogene
and its normal homologue, Nature, 302: 33-37.

Chan, C., (1984), NTP Technical Report, Cas N o. 597-25-1.

Chang, E. H., Furth, M. E., Scolnick, E. M., Lowry, D. R.,(1982), Tumorigenic
transformation of mammalian cells induced by a normal human gene
homologous to the oncogene of Harvey murine sarcoma virus, Nature, 297:
479-483.

Chang, E. H., Gonda, M. A., Ellis, R. W., Scolnich, E. M. Lowry, D. R, (1982),
Human genome contains four genes homogeneous to transforming genes of

Harvey and Kirsten murine sarcoma virus, Proc. Natl. Acad. Sci. U.S.A., 79:
4848-48952

Chen, Z. Q, Ulsh, L S., DuBois, G. and Shih, T., (1985), Posttranslational
processing of p21 ras proteins involves palmitoylation of the c-terminal
tetrapeptide containing cys 186,1. Viral, 56: 607-612.

Chisari,F.V., (1989), Analysis of hepatitis B virus gene expression in
transgenic mice, In Concepts in Viral Pathogenesis Vol 3, (ed.) Notkins and
Oldstone, 164-169.

Cichutek, K. and Duesberg, P. H., (1986), Harvey ras genes transform without
codon mutations, apparently activated by truncation of a 5' exon (exon-1),
Proc. Natl. Acad. Sci., 83: 2340-2344.

Cifone, M. and Fidler,1., (1981), Increasing metastatic potential is associated
with increasing genetic instability of clones isolated from murine neoplasma.
Proc. Natl. Acad. Sci. U.S.A. 78: 6249-6252.

104

Clarke, S., Vogel, I. P., Deschenes, R. 1., Stock,1., (1988), Posttranslational
modification of the H-ras oncogene protein: evidence for a third class of
protein carboxyl methyltransferase, Proc. Natl. Acad. Sci. U.S.A., 85: 4643-4647.

Cotran, R. 5., Kumar, V., Robbins, S. L., (1989), Neoplasia, in Pathologic Basis
of Disease, W. B. Saunders Co., 239-306.

Courtneidge, S. A. and Smith, A. F., (1983), Polyoma virus transforming
protein associates with the product of the c-src cellular gene, Nature 303: 435-
436.

Couturier, A., Bazzar, 5., and Castagna, M., (1984), Further characterization of
tumor promoter mediated activation of protein kinase C, Biochem. Biophys.
Res., Commun., 121: 448-455.

Croce, C. M. and Nowell, P. C., (1985), Molecular basis of human B cell
neoplasia. Blood 65: 1-7.

Curran, T. and Franza, BR, Pas and jun: the AP-1 connection, Cell, 55: 395-
397.

Curran, T., Miller, A. D., Zokas, L., Verma, I. M., (1984), Viral and cellular fos
proteins: a comparative analysis, Cell, 36: 259-268.

Dalla-Favera, R., Bregni, M.,. Erikson,1., Patterson, D., Gallo, R. C. and Croce,
C. M., (1982), Assignment of the human c-myc oncogene to the region of
chromosome 8 which is translocated in Burkett lymphoma cells, Proc. Natl.
Acad. Sci. U.S.A. 79: 7824-7827.

Darnell, 1., Lodish, H., and Baltimore, D., (1986), Molecular Cell Biology, W.
H. Freeman and Company, N.Y., N.Y.

DeCapiro, 1. A., Ludlow, 1. W., Figge,1., Shew,1-Y., Huang, C-H., Lee, W-H.,
Marsilio, E., Poucha, E. and Livingston, D. M., (1988), SV-40 large tumor
antigen forms a specific complex with the product of the retinoblastoma
susceptibility gene, Cell, 54: 275-283.

Delbecco, R. and Ginsberg, H S., (1980), Virology, Harper and Row, Publisher,
Philadelphia, PA.

105

Delclos, K. B., Tarpley, W., G., Miller, E. C., Miller, 1. A., (1984), 4-
Aminoazobenzene and N ,N -dimethyl-4-aminoazobenzene as equipotent
hepatic carcinogens in male C57BL/ 61 X C3H/He1 F1 mice and
characterization of N-(deoxyguanosin-8-yl)-4-aminobenzene as the major
presistent hepatic DNA-bound dye in these mice, Cancer Res. , 44: 2540-2550.

DeVos, A. M., Tong, L., Milburn, M. V. Matias, P. M., 1ancarik, 1., Noguchi, S.,
Nishimura, S., Miura, K.,. Ohtsura, E., Kim, S-H., (1988), Three-dimensional
structure of an oncogene protein: catalytic domain of human c-H-ras p21,
Science, 239: 888-893.

Dippold, W. G.,1ay, G., DeLev, A. B., Kharry, G., and Old, L. 1., (1981), p53
transformation-related protein detection by monoclonal antibody in mouse
and human cells, Proc. Natl. Acad. Sci. U.S.A., 78: 1695-1699.

Distel, R. 1., R0, H-S, Rosen, B. S., Groves, D., L., Spielgelman, B. M., (1987),
Nucleoprotein complexes that regulate gene expression in adipocyte
differentiation: direct participation of c-fas, Cell, 49: 35-844.

Doll, R. and Hill, A. B.,(1950), Smoking and carcinoma of the lung, British
Medical Journal, 2: 739-48.

Doll, R., (1988), Effects of exposure to vinyl chloride, Scand. 1. Work Environ
Health, 14: 61-78.

Doll, R., Peto, R., (1981), The causes of cancer: quantitative estimates of
avoidable risks of cancer in the United States today, 1. Natl. Cancer Inst, 66(6):
1192-1303.

Downward, 1., Yarden, Y., Mayes, E., Scarce, G., Totty, N. Stockwell, P., Ulrich,
A., Schlessinger,1., Waterfield, M. D., (1984), Close similarity of epidermal
growth factor receptor and the v-erb B oncogene protein sequences, Nature,
307: 521-527.

Drake, 1. W., Allen, E.F., Forsberg, S.A., Preparota, R.M., and Greenberg, E.O.,
(1969), Spontaneous mutations, Nature (London), 221: 1128-1132.

Drinkwater, N. R. and Ginsler, 1. 1., (1986), Genetic control of
hepatocarcinogenesis in C57BL/61 and C3H/ e1 inbred mice, Carcinogenesis, 7:
1701-17607.

106

Druckey, H., (1967), In Potential Carcinogenic Hazards from Drugs, Evaluation
of Risk. Terhaut R (ed.): UTCC Monograph Series, Vol. 7. Berlin:Springer-
Verley, 60-77.

Duesberg, P. H. and Vogt, P. K., (1970), Differences between ribonucleic acids of
transforming and nontransforming avian tumor viruses, Proc. Natl. Acad.
Sci. U.S.A., 67: 1673-1680.

Dutton, DR. and Bowden, G. T., (1985), Indirect induction of a clastogenic
effect in epidermal cells by a tumor promoter, Carcinogenesis 6: 1279-1284.

Dyson, N., Howley, P. M., Munger, K., Harlow, E., (1989), The human
papilloma virus-16 E7 oncogene is able to bind to the retinoblastoma gene
product, Science, 243: 934-937.

Ehrlich, G. D. and Poiesy, B. 1., (1988), Clinical and molecular parameters of
HTLV-l infection, Clin. Lab Med. 8: 65.

Ek, B and Heldin, C-H., (1982), Characterization of a tyrosine-specific kinase
activity in human fibroblast membranes stimulated by platelet-derived
growth factor, 1. Biol. Chem., 257: 10486-10492.

Elcomb, C. K, Rose, M. S., Pratt, I. S., (1985), Biochemical and histological
changes in rat and mouse liver following the administration of
tricthroethylene: possible relevance to species differences in
hepatocarcinogenicity, Toxicol. and Applied Pharmacol. 79: 365-76.

Environmental Health Criteria, (1975), Trichloroethylene, WHO, Geneva, 50.

Erikson, 1. Nishikura, K., ar-Rushdi, A., Finan, 1., Emanuel, B., Lenoir, G.,
Nowell, P. and Croce, C., (1983b), Translocation of an immunoglobin kappa
locus to a region 3' of an unrearranged c-myc oncogene enhances c-myc
expression, Proc. Natl. Acad. Sci. USA, 80: 7581-7585.

Erikson,1., Finger, 1., Sun, L., ar-Rushdi, A., Nishikura, K., Minowada, 1.,
Finan,1., Emanuel, B. S., Nowell, P. C., Croce, C. M., (1986), Deregulation of c-
myc by translocation of the alpha-locus of the T-cell recptor in T- cell
leukemias, Science, 232: 8894-886.

Erikson,1., Rushdi, A., Drwinga, H. L., Nowell, P. C., Croce, C. M., (1983a),
Transcriptional activation of the translocated c-myr oncogene in Burkitt's
lymphoma, Proc. Natl. Acad. Sci., 80: 820-824.

107

Fasano, 1., Taparowsky, E., Fiddles, 1., Wigler, M., Goldfarb, M., (1983),
Sequence and structure of the coding region of the human H-ras-l gene from
T24 bladder carcinoma cells, 1. Male & Applied Gen., 2(2): 173-180.

Fasano, O., Aldrich, T., Tamanoi, F., Taparowsky, E. Furth, M. and Wigler,
M., (1984), Analysis of the transforming potential of the human H-ras gene by
random mutagenesis, Proc. Natl.Acad. Sci., (81): 4008-4012.

Fennell, T. R., Wiseman, R. W., Miller, 1. A., Miller, E., C., (1985), Major role
of hepatic sulfotransferase activity in the metabolic activation, DNA adduct
formation, and carcinogenicity of 1'-hydroxy-2',3'-dehydroestragole in infant
male C57BL/61x C3H/ He] F1 mice, Cancer Res, 45: 5310-5320.

Finlay, C., A., Hinds, P. W., Levine, A. 1., (1989), The p53 proto-oncogene can
act as a suppressor of transformation, Cell, 57: 1083-1093.

Food and Drug Administration, Advisory Committee on Protocols for Safety
Evaluation: Panel on Carcinogenesis A Report on Cancer Testing in the
Safety Evaluation of Food Additives and Pesticides, (1971), Toxicol., Appl.
Pharmacol., 20: 419-438.

Ford, R Maizel, A. ,(ed) (1986), Mediators in Cell Growth and Differentiation.
New York: Raven Press.

Fox, T. R. and Watanabe, P. G.,(1985), Detection of a cellular oncogene in
spontaneous liver tumors of B5C3F1 mice, Science, 228: 596-597.

Fax, T. R., Schumann, A. M., Watanabe, P., G., (1987), Activation of a cellular
proto-oncogene in spontaneous liver tumor tissue of the B5C3F1 mouse,
Arch. Toxicol., Suppl 10: 217-227.

Franks, L. M., (1986), What is Cancer? In Cellular and Molecular Biology of
Cancer Ed. L. M. Frank and N. Teich, Oxford Scientific Pub. Oxford.

Frenkel, K., and Chrzan,K., (1987), DNA base modifiction by tumor promoter-
activated polymorphonuclear leukoctyes, Carcinogenesis 8: 455-460.

Gibbs, 1., B., Sigal, I. S., Poe, M., Scolnick, E. M., (1984), Intrinsic GTPase activity
distinguishes normal and oncogenic ras p21 molecules, Proc. Natl. Acad. Sci.
U.S.A., 81: 5704-5708.

108

Goldstein, B.D., Witz, G., Amoruso, M., Stone, D.S., and Troll, W., (1981),
Stimulation of human polymorphonuclear leukocyte superoxide anion
radical production by tumor promoters, Cancer Lett. 11: 257-262.

Goustin, A. 5., Leaf, E. B., Shipley, G. D., Moses, H. L., (1986), Growth factor
and cancer, Cancer Research, 46: 1015-1029.

Graso, P. and Hardy, 1.,(1975), Strain differences in natural incidence and
response to carcinogens, In Butler, W. H. and Newberne, P. M., (ed), Mouse
Hepatic Neoplasia, Elsevier, New York, p.111-131.

Grasso, P. and Hardy, 1., (1975), Strain differences in natural incidences and
response to carcinogens, In Butler, W. H. and Newberne, B. M. (eds) , Mouse
Hepatic Neoplasia, Elsevier, New York, pp.111-131.

Groffen, 1., Stephenson, 1. R. Heisterkamp, N., deKlein, A., Bartram, C. and
Grosveld, G., (1984), Philadelphia chromosome breakpoints are clustered
within a limited region, bcr, on chromosome 22, Cell 36: 93-99.

Gross, L., (1983), Oncogenic viruses, 3rd ed., Pergamon Press, NY., NY.

Gutierres, L. Magee, A. 1., Marshall, C. 1., and Hancock, 1. F., (1989), Post-
translational processing of p21 ras is a two-step and involves carboxyl-
methylation and carboxy-terminal proteolysis, EMBO], 8: 1093-1088.

Hamm, T. E., (1984), Occurrence of Mouse Liver Neoplasma in the National
Cancer Institute Bioassays ed. Popps 1. A. , Hemisphere Publishing, Mouse
Liver Neoplasma, Current Perspective, p.27-38.

Hancock, 1. F., Magee, A. 1., Childs, 1. E. and Marshall, C. 1., (1989), All ras
proteins are polyisoprenylated but only some are palmitolyated, Cell, 57: 1167-
1172.

Hanigan, M., H., Kemp, C. 1., Ginsler,1., 1. and Drinkwater, N. R., (1988), Rapid
growth of preneoplastic lesions in hepatocarcinogen-sensitive C3H/He1 male
mice relative to C57BL/61 male mice, Carcinogenesis, 9(6): 885-890.

Hansen, M., and Cavenee, W., (1988), Tumor suppressors: Recessive
mutations that lead to cancer, Cell, 53: 172-173.

109

Harris, 1. F., Chambers, A. F., Hill, R. P. and Ling, V., (1981), Metastatic
varients are generated spontaneously at a high rate in mouse KI-IT tumor,
Proc. Natl. Acad. Sci. U.S.A. 79: 5547-5551.

Hartman, P. E., (1980), Bacterial mutagenesis - review of new insights,
Environ. Mut., 2: 3-16.

Heisterkamp, N., Stephenson, 1. R., Goffen, 1., Hansen, P. F., deKlein, A.,
Bartram, C. R., Grosveld, G., ( ), Localization of the c-abl oncogene adjacent to
a translocation breakpoint in chronic myelocytic leukemia, Nature, 306: 239-
242.

Heppner, G. H., (1984), Tumor heterogeneity, Cancer Research, 44: 2259-2265.

Hill, D. E., Hope, I.A., Macke, 1. P., Struhl, K., (1986), Saturation mutagenesis
of the yeast his3 regulatory site: requirements for transcriptional induction
and for binding by GCN 4 activator protein, Science, 234: 451-457.

Hollis, G. F., Mitlchell, K. F., Battey,1., Potter, H., Taub, R. A., Lenoir, G. Ledey,
P., (1984), A varient translocation places the lambda immunoglobin genes 3'
to the c-myg oncogene in Burkett lymphoma, Nature, 307:752-755.

Hope, I. A., Struhl, F.,(1985), GCN4 protein, synthesized in vitro, binds HIS3
regulatory sequences: implications for general control of amino acid
biosynthetic gene in yeast, Cell, 43: 177-188.

Houwelling, A., Van der Elsen, P. 1. and Van der Eb, A. 1., (1980), Partial
transformation of primary rat cells by the leftmost 4.5% fragment of
adenovirus 5 DNA, Virology, 105: 537-550.

Huang, 1. 5., Huang, 5. S. and Deuel, T. F., (1984), Transforming protein of
simian sarcoma virus stimulates autocrine grthh of SSV - transformed cells
through PDGF cell - surface receptors, Cell, 39: 79-87.

Huebner, RJ. and Todaro, G.1., (1969), Oncogenes of RNA tumor viruses as
determinants of cancer, Proc. Natl. Acad. Sci. U.S.A., 64: 1087.

Hunter, T., (1984), Oncogenes and proto-oncogenes: How do they differ, 1NCI,
73(4): 773-786.

110

Hurley, 1. B., Simon, M. I., Teplow, D. B. Robishaw, J. D., Gilman, A. G.,
(1984). Homologies between signal transducing G proteins and ras gene
products, Science, 226: 860-862.

Hurlin, P. 1., Fry, D. G., Maher, V. M., McCormick, J. J., (1987), Morphological
transformation, focus formation and anchorage independence induced in
diploid human fibroblasts by expression of a transfected H-ras, Cancer Res., 47:
5752-5757.

Hurlin, P. 1., Maher, V. M., McCormick, J. 1., (1989), Malignant transformation
of human fibroblasts cause by expression of a transfected T24 H-ras oncogene,
P.N.A.S., 86: 187-191.

IARC Monograph, (1980), Long-term and short-term assays for screening
carcinogens: a critical appraisal, Supplement #2,

Interdisciplinary Panel on Carcinogenicity, (1984), Criteria for evidence of
chemical carcinogenicity, Science, 225: 682.

Inversen, U. M. and Inversen, O. H., (1979), The carcinogenic effect of TPA
when applied to the skin of hairless mice, Virchows Arch. 30:33-42.

Jensen, R. K., Sleight, S. D., Goodman, J. I., Aust, S. D., and Trosko, J. E. (1982),
Polybrominated biphenyls as promoters in experimental hepatocarcinogensis
in rats, Carcinogenesis, 3: 1183-1186.

Jochemsen, H., Daniels, G.S.G., Hertoghs, P.1., Schrier, van den Elsen, P.1., and
van der Eb, A.1., 1982, Identification of adenovirus type 12 gene products
involved in transformation and oncogenesis, Virology, 122: 15

Johnson, A., Heldin, C-H, Westermark, G., Wasteson, A., (1982), Platelet-
derived growth factor: identification of constituent polypeptide chains,
Biochem. Biophys. Res. Comm., 104: 66-74.

Kakunaga, T. and Crow, 1. D., (1980), Cell varients showing differential
susceptibility to ultraviolet light-induced transformation, Science 209: 505-507.

Kaminski, M. and Auerbach, R., (1988), Tumor cells are protected from NK
cell-medicated lysis by adhesion to endothelial cells, Int. J. Cancer 41: 847.

Klein, G.‘, (1983), Specific chromosomal translocation and the genesis of B-cell-
derived tumors in mice and men, Cell, 32: 311-315.

111

Klein, G., (1987), The approaching era of the tumor suppressor genes, Nature,
238: 1539-1545.

Klinger, H. P., (1982), Suppression of tumorigenicity, Cytogenet. and Cell.
Genet. , 32: 68-84.‘

Kluwe, W. M., Huff, J. E., Matthews, H. B., Irwin, R., Haseman, J. K., (1985),
Comparative chronic toxicities and carcinogenic potentials of 2-ethylhexyl-
containing compounds in rats and mice, Carcinogenesis 6: 1577-83.

Knudson, A. G., (1971), Mutation and cancer: statistical study of
retinoblastoma, Proc. Natl. Acad. Sci. U.S.A, 68: 820-823.

Krieg, P., Finch, J., Fustenberger, G., Melber, K., Matrisian, L. M., Bowden, G.
T., (1988), Tumor promoters induce a transient expression of tumor-
associated genes in both basel and differentiated cells of the mouse epidermis,
Carcinogenesis, 9 (1): 95-100.

Lacal,1. C., Srivastava, S. K., Anderson, P. 5., Aaronson, S. A., (1986), Ras p21
proteins with high or low GTPase activity can efficiently transform NIH/3T3
cells, Cell, 44: 609-617.

Lai, C-C, Miller, J. A., Miller, E. C., Liem, A., (1985), N sulfoxy-2-
animofluorene is the major ultimate electrophilic and carcinogenic
metabolite of N hydroxy-2—acetylaminofluorene in the livers of infant male
C57Bl / 61 x C3H/ HeJ F1 B6C3F1 mice, Carcinogenesis 6 (7): 1037-1046.

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.

Land, H., Parada, L., and Weinberg, R. A., (1983), Cellular oncogenes and
multistep carcinogenesis, Science, 222: 771 -778.

Lane, D. P. and Crawford, L. V., (1979), T Antigen is bound to a host protein in
SV40 transformed cells, Nature, 278: 261-263.

Lee, W. Mitchell, D., Tjian, R., (1987), Purified transcription factor AP—l
interacts with TPA-inducible enhancer elements, Cell, 49: 741-752.

112

Ling, V., Chambers, A. F., Harris, J. F. and Hill, R. P., (1985.), Quantitative
genetic analysis of tumor progression, Cancer Metastasis Rev., 4: 173-194.

Liotta, L. A. et al., (1988), Biochemical mechanisms of tumor invasion and
metastases, Prog. Clin. Biol. Res. 256: 3.

Lippman, M. F., dickson,R.B., Gelmann, E.P., Rosen, N., Kaufman, D.,
Knabbe, C., Bates, S., Kasid, A., Salomon, D., (1987), Growth regulation of
human breast carcinoma occurs through regulated growth factor secretion, 1.
Cell Biochem. 35 (1): 1-16.

Loeb, L. A., Weymouth, L. A., Kunkel, T. A., Gopinathan, K. P., Beckman, R.
A., Dube, D. F., (1978), Cold Spring Harbor Symposium Quant., Biol. 43: 921-
927.

Loury, DJ. Goldsworthy, T.L., and Butterworth, BE, (1987), The role of
measuring cell replication as a predictive index of tissue-specific tumorigenic

potential, Banbury Report 25: Nongenotoxic Mechanisms in Carcingenesis,
Cold Spring Harbor.

MacMahon, B., (1979), Epidemiology's strengths, weaknesses in
environmental health issues, The National Research Council in 1979:
Current Issues and Studies, p340.

Maher,1. M.,, Dorney, D. 1., Mendrala, A. L, Konge-Thomas, B., McCormick, J.
1., (1979), DNA excission-repair processes in human cells can eliminate the

cytotoxic and mutagenic consequences of ultraviolet irradation, Mutation
Research, 62: 311-323.

Maki, Y., Bos, T. 1., Davis, C., Starbuck, M. land Vogt, P. K., (1987), Avain
sarcoma virus 17 carries the jun oncogene, Proc. Natl. Acad. Sci., 84: 284-2852.

Manolova, G. and Manolova, Y., (1972), Marker band in one choromosome 14
from Burkitt lymphoma, Nature 237: 33-34.

Manne, V., Bekesi, E. and Kung, H., (1985), H-ras proteins exhibit GTPase
activity: point mutations that activate Ha-ras gene products result in
decreased GTPase activity, Proc. Natl. Acad.Sci U.S.A., 82: 376-380.

Marx, J. L. , How DNA viruses may cause cancer. (1989), Science, 243: 1012-
1013.

113

McCann, H. 1.., Choi, E. Yamasaki, E., Ames, B. N., (1975), Detection of
carcinogens as mutagens in the Salmonella / microsomal test: assay of 300
chemicals, Proc. Natl. Acad. Sci. U.S.A., 72(12): 5135-5139.

McCann, J. Choi, E., Yamasaki, Ames B., (1975), Detection of carcinogens as
mutagens in the Salmonella microsomal test: Assay of 300 chemicals, Proc.
Natl. Acad. Sci. U.S.A., Vol. 72: 135-5139.

McCormick, F.,(1989), ras GTPase activating protein: Signal transmitter and
signal terminator, Cell, 56: 5-8.

McGrath ,1. P., Capon, D. J., Goedel, D. V., Levinson, A. D.,(1984),
Comparative biochemical properties of normal and activated human ras p21
protein, Nature, 310: 644-649.

McGrath, J. F., Capon, D. J., Smith, D. H., Chen, E. Y., Seeburg, P. H., Goedel, D.
V. and Levinson, A. D., (1983), Structure and organization of the human K-
ras proto-oncogene and a related processed pseudogene, Nature, 304: 501-506.

Miller, E. C. and Miller, J. A., (1981), Mechanisms of chemical carcinogenesis,
Cancer 47:1055—1064.

Miller, E. C., Miller, J. A., Boberg, E. W., Delclos, K. B., Lai, C-C, Fennell, T. R.,
Wiseman, R. W., Liem, A., (1985), Sulfuric acid esters as ultimate electrophilic
and carcinogenic metabolites of some alkenylbenzenes and aromatic amines
in mouse liver, Carcinogenesis, 10: 93-107.

Mirsalis,1. C., Tyson, C. K., Loh, E. N., Steinmetz, K. L., Bakke, J. P., Hamilton,
C. M., Spak, D. K. and Spalding , J. W., (1985), Induction of hepatic cell
proliferation and unscheduled DNA synthesis in mouse hepatocytes
folloWingin viva treatment, Carcinogenesis, 6: 1521-1524.

Mougneau, E., Lemieux, L., Russoulzadegan, M., Cuzin, F., (1984), Biological
activities of v-myc and rearranged c-myc oncogenes in rat fibroblast cells in
culture, Proc. Natl. Acad. Sci., 81: 5758-5762.

Muir, C. S.,(1979), Limitations and advantages of epidemiological
investigations in environmental carcinogenesis (1979), Annals New York
Academy of Sciences, 153-165.

114

Nerenberg, M., Hinrechs, 5. H., Reynolds, R.K., Khoury, G., Jay, G., (1987), The
fat gene of human T-lymphototropic virus type 1 induces mesenchymal
tumors in transgenic mice, Science 237: 1324-1329.

Nicolson, G. and Milos, L., (ed). (1984), Cancer Invasion and Metastasis:
Biologic and Therapeutic Aspects, New York, Raven Press.

Nishimura, 5., and Sekiya, T., (1987), Human cancer and cellular oncogenes, J.
of Biochem., 243: 313-327.

Nowell, P. C., (1986), Mechanism of Tumor Progression, Cancer Research 46:
2203-2207.

Nutrition Foundation, (1983), The relevance of mouse liver hepatomas to
humand carcinogenic risk, Report to the international expert advisory
committee to the Nutrition Foundation, The Nutrition Foundation, Inc,
Washinton, D.C.

O'Connell, J. F., Rotstein, J.B., Slaga, T.1., (1987), Tumor promotion and
progression in the mouse skin model, In Banbury Report 25: Nongenotoxic
Mechanisms in Carcingenesis, Cold Spring Harbor, 11-24.

Oksala, T. and Therman, E., (1979), Mitotic abnormalities in cancer.
Chromosomes and Cancer. 1. German (ed) pp. 239-263, New York. John
Wiley 8: Sons .

Omar, R. A. and Lanks, K. W., (1983), Origin of phenotypic variation in clones
of smian virus 40-transformed mouse embryo cells, Cancer Res., 43: 1835-1841.

Oren, M., Maltgman, W., and Levine, A. 1., (1981), Post-translational
regulation of 54 K cellular tumor antigen in normal and transformed cells,
Mol. Cell. Biol., 1: 101-110.

'Pacifici, G. M., Boobis, A. R., Brodie, M. 1., McManus, M. E., Davis, D. 5., (1981),
Tissue and species differences in enzymes of epoxide metabolism,
Xenobiotica, 11: 73-79.

Persson, H. and Iede, D., (1984), Nuclear localization and DNA binding
properties expressed by human c-myc oncogene, Science, 225: 718-721.

115

Peterson, 1. A., Ceriani, R. L., Blank, E. W. and Osvaldo, L., (1983), Comparison
of rates of phenotypic variability in surface antigen expression in normal and
cancerous human breast epithelial cells, Cancer Research 43: 4291-4296.

Pitot, H. C. Goldsworthy, T., Cambell, H. A., Poland, A., ( 1980 b), Quantitative
evaluation of the promotion by 2, 3, 7 8-tetrachlorodibenzo-p-dioxin of
hpatocarcinogenesis for diethylnitrosamine, Cancer Research 40 (10): 3616.

Pitot, H. C., and Sirica, A.E., (1980 a), The stages of initiation and promotion in
hepatocarogenesis, Biochem. Biophys. Acta, 605: 191-215. .
Pitot, H. C., Goldsworthy, T., Campbell, H. A. and Poland, A., (1980),
Quantitative evaluation of the promotion by 2,3,7,8-tetrachlorodibenzo-p-
dioxin of hepatocarcinogenesis from diethylnitrosamine, Cancer Res, 40:
3616-3620.

Pitot, H. C., Goldworthy, T. and Moran, S., (1982), The natural history of
carcinogenesis: Implications of experimental carcinogenesis in the genesis of
human cancer, Mechanism of Chemical Carcinogenesis. Ed. Harris, C. C.,
Cerutti, P. A. Pub. Alan R Liss Inc., NY, NY.

Pitot, H.C., (1986) Fundamentals of Oncology, 3rd edition, Marcel Decker, New
York. '

Piwnica-Worms, H., Saunders, K. B., Roberts, T. M., Smith, A. E., Cheng, S.
H., (1987), Tyrosine phosphorylation regulates the biochemical and biological
properties of pp60 c-src, Cell, 49: 75-82.

Poland, A. and Knutson, J. C., (1982), 2,3,7,8-Tetrachlorodibenzo-p-dioxin and
related aromatic hydrocarbons: examination of the mechanism of toxicity,
Ann. Rev. Pharmacol. 22: 217-54.

Poland, A., Glover, E., (1975), Genetic expression of aryl hydrocarbon
hydroxylase by 2, 3, 7, 8-tetrachlorodibenzo-p-dioxin evidence for a receptor
mutation in genetically nonresponsive mice, Mol. Pharmacol. 11(4): 389-398.

Ponticelli, A. 5., Whitlock, C. A., Rosenberg, N. and Witte, ON, (1982), In
viva tyrosine phosphorylation of the Abelson virus transforming protein are
absent in its normal cell homolog, Cell, 29: 953-960.

116

Powers, 5. Katovka, T., Fasano, O., Goldfarb, M., Stratherm,1., Broach, 1. and
Wigler, M., (1984), Genes in S. cerevisiae encoding proteins with domains
homologous to the mammalian ras protein, Cell, 36: 607-612.

Pozzati, R, Muschel, R, Williams, 1., Padmanabhan, R., Howard, B., Liotta, L.,
Khoury, G., (1986), Primary rat embryo transformed by one or two oncogenes
show different, metastatic potentials, Science, 232: 223-227.

Pulciani, 5., Santos, E., Lauver, A. V., Long, L. K., Aaronson, S. A., Barbacid,
M.,(1982), Oncogenes in solid human tumors, Nature, 300: 539-542.

Rassoulzadegan, M., Cowie, A., Carr, A., Glaichenhous, N., Kamen, R., Cuzin,
F., (1982),The roles of individual polyoma virus early proteins in oncogenic
transformation, Nature, 300: 713-718.

Reddy, E., P., Lipman, D., Andersen, P. R, Tronick, 5. R., Aaronson, S. A.,
(1985), Nucleotide sequence analysis of the balb/ c murine sarcoma virus
transforming gene, J. Virol. 53: 984-987.

Reddy, 1. K. and Lalwani, N. D., (1983), Carcinogenesis by hepatic peroxisomal
proliferations: evaluation of the risk of hypolipidemic drugs and industrial
plasticizers to humans, CRC Crit. Rev. Toxicol. 12: 1-58.

Reddy, J.K., and Lalwani, ND, (1983), Carcinogenesis by hepatic peroxisome
porliferatiors: Evaluation of the fisk of hypolipidemic drugs and industrial
plasticizers to humans, CRC Crit. Rev. Toxicol. 12: 1-58.

Reichert, D., (1983), Biolgical actions and interactions of tetrachloroethylene,
Mutation Res, 123: 411-429.

Reynolds, 5. H., Stowers, S. 1., Maronpot, R. R., Anderson, I. M.. W.,
Aaronson, S. A.,(1986), Detection and identification of activated oncogenes in
spontaneously occurring benign and malignant hepatocellular tumors of the
B6C3F1 mouse, Proc. Natl. Acad. Sci. U.S.A., 83: 33-37.

Reynolds, 5. H., Stowers, 5. 1., Patterson, R M., Maronpot, R R, Aaronson, S.
A. and Anderson, M. W., (1987) Activated oncogenes in B5C3F1 mouse liver
tumors: Implications for risk assessment. Science, 237: 1309-1316

Reynolds, 5. H., Stowers, S. 1., Patterson, R. M., Maronpot, R. R, Aaronson, S.
A., Anderson, M. W.,(1987), Activated oncogenes in B6C3F1 mouse liver
tumors; Implications for risk assessment, Science 237: 1309-1316.

117

Ruby, H. E., (1983), Adenovirus early region 1A enables viral and cellular
transforming genes to transform primary cells in culture, Nature, 304: 602-
606.

Ruley, H. E.,(1983), Adenovirus early region 1A enables viral and cellular
transforming genes to transform primary cells in culture, Nature, 304: 602-
606. -

Sainsbury, J.RC., (1988), Epidermal growth factor receptor status as a predictor
of early recurrence of and death from breast cancer, Lancet, 1: 1398.

Santos, E. and Nebreda, A. R, (1989), Structural and functional properties of
ras proteins, The FASEB Journal, 3(10): 2151-2163.

Sarnow, P., Ho, Y. S. , William, J., and Levine, A. J., (1982), Adenovirus E1b-58
kd tumor antigen and SV-40 large tumor antigen are physically associated
with the same 54 kd cellular protein in transformed cells, Cell, 28: 387-394.

Sassone-Corsi, P., Lamph, W. W., Kamps, M., Verma, I. M. (1988), fas-
Associated cellular p39 is related to nuclear transcription factor AP-l, Cell, 54:
553-560.

Schecter, A. L., Hung, M. C., Vardyanathan, L., Decker, S. 1., Drebin, J. A.,
Greene, M. I. and Weinberg, R. A., (1984), The neu oncogene: an erb b-related
gene encoding a 185,000 Mr tumor antigen, Nature, 312: 513-516.

Schulte-Hermann, R, Schuppler, 1., Ohde, G., and Timmermann-Trosiemer,
I., (1982), Effect of tumor promoters on proliferation of putative preneoplastic
cells in rat liver, E. Hecker, N. Fusenig, W. K. Lung, F. Marks and H.
Thielman (eds),‘In Carcinogenesis, Vol. 7, Raven Press, New York, pp99-104
(1982).

Schulte-Hermann, R., Timmermann-Trosiener, I. and Schuppler, J., (1983),
Promotion of spontaneous preneoplastic cells in rat liver as a possible
explaination of tumor production by nonmutagenic compounds, Cancer Res,
43:839-844.

Schwab, M., Alitalo, K., Varmus, H. E., Bishop, 1. M., (1983), A cellular
oncogene (c-Ki-ras) is ampliphied, overexpressed, and located within
karyotypic abnormalities in mouse adrenocortical tumour cells, Nature, 303:
497-501.

118

Scolnick, F. M., Papogeorge, A. G., Shih, T. Y., (1979), Guanine nucleotide-
binding activity as an assay for src protein of rat-derived murine sarcoma
virus, Proc. Natl. Acad. Sci., 76: 5355-5359.

Seeburg, P. H. Colby, W. W., Capon, D. 1., Goedel, D. V. and Levinson, A. D.,
(1984), Biological properties of human c-Ha-ras 1 genes mutated at codon 12,
Nature, 312: 71-75.

Sher, C. 1., Rettenmier, C. W., Socca, R, Roussel, M. F., Look, A. T. Stanley,
E., R, (1985), The c-fms proto—oncogene product is related to the receptor for
the mononuclear phagocyte growth factor CSF-l, Cell, 41: 665-676.

Shilo, B-Z., Weinberg, R A., (1981), DNA sequences homologous to
vertebrate oncogenes are conserved in Qrgsophila melanogasler, Proc. Natl.
Acad. Sci. U.S.A., 78: 6789-6792.

Shimizu, K., Goldfarb, M., Suard, Y., Perucha, M., Yen, L., Kamota, T.,
Feramisco,1., Stavnezer, E., Fogh, 1., Wigler, M. H., (1983), Three human
transforming genes are related to the viral ras oncogenes, Proc. Natl .Acad.
Sci., 80: 2112-2116.

Slamon, D. 1., (1987), Human breast cancer: Correlation of relapse and
survival with amplification of the HER-Z/neu oncogene, Science, 235: 177.

Smets, L. A., (1980), Cell transformation as a model for tumor induction and
neoplastic growth, Biochem. Biophys. Acta, 605: 93-111.

Srnit, V. T., Boot, A. J. M, Smits, A. M.M, Fleuren, G. 1., Comelisse, C. 1., and
Bos, J. L., (1988), K-ras codon 12 mutations occur very frequently in pancreatic
adenocarcinoma, Nuclei Acid, Res, 16: 7773-7782.

Solt, D., Farber, E., (1975), New principle for the analysis of chemical
carcinogenesis, Nature, 263(5579): 701-703.

Spandidos, D. A. and Wilkie, N. M., (1984), Malignant transformation of early
passage rodent cells by a single mutated human oncogene, Nature, 310: 469-
475.

SpandidOs, D. A., and Kerr, 1. B., (1984), Elevated expression of the human ras
oncogene family in premalignant and malignant tumour of the colorectum,
Br. J. Cancer, 49: 681-688.

119

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

Srinivasan, A., Peden, K. W. C. and Pipas, 1. M., (1989), The large tumor
antigen of simian virus 40 encodes at least two distinct transforming
functions, 1. of Virology, 63 (12): 5459-5463.

Stewart, T. A., Pattengale, P. K, Ieder, P., (1984), Spontaneous mammary
adenocarcinoma in transgenic mice that carry and express MTV / myc fusion
genes, Cell, 38: 627-637.

Stowers, S. 1., Wiseman, R. W., Ward, J. M., Miller, E. C. Miller, J. A.,
Anderson M. W., Eva, A., (1988), Detection of activated proto-oncogenes in N-
nitrosodiethylamine induced liver tumors: a comparison between B6C3F1
mice and Fischer 344 rats, Carcinogenesis, 9 (2): 271-276.

Takeya, T., Hanafusa, H., (1983), Structure and sequence of the cellular gene
homologous to the RSV src gene and the mechanism for generating the
transforming virus, Cell, 32: 881-890.

Task Force of Past Presidents - Society of Toxicology, (1982), Fundamental and
App] Tox., 2: 101. '

Taub, R, Kirsch, I., Morton, C., Lenoir, G., Swan, D., Tronick, 5., Aaronson, 5.,
and Leder, P., (1982), Translocation of the c-myc gene into the immunoglobin

heavy chain locus in human Burkett's lymphoma and murine plasmacytoma
cells, Proc. Natl. Acad. Sci. U.S.A., 79: 7837-784.

Teich, M. N ., (1986) In Cellular and Molecular Biology of Cancer, Ed. Frank,
L.M. and Teich, N., Oxford Science Pub.

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

Trahey, MN ., Milley, R. 1., Cole, G. E., Innis, M., Paterson, H., Marshall, C. 1.,
Hall, McCormick, F., (1987a), Biochemical and biological properties of the
human N-ras p21 protein, Mol. Cell Biol., 7(1): 541-544.

120

United States Public Health Service, (1989), The health consequences of
smoking - Cancer: A report of the surgeon general, Rockville, MD., United
States Dept. of Health and Human Services, Office on Smoking and Health.

Van Beveren, C., van Stroaten, F., Curran, T., Muller, R, Verma, I. M., (1983),
Analysis of FBT-MuSV provirus and c- fos (mouse) gene reveals that viral
and cellular fos gene products have different carboxy termini, Cell, 32: 1241-
1255.

van dem Berhe, H., Parloir, C. Gosseyl, 5., Eglebienne,1., Comu, G. and Sokal,
G., (1976), Varient translocations in Burkett lymphoma, Cancer Genet.
Cytogenet., 1: 9-14.

Verma, A. K.,. Boutwell, R K., (1980), Effects of dose and duration of
treatment with the tumor promoting agent 12-O- tetra decanoyl phorbol 13
acetate on mouse skin carcinogenesis, Carcinogenesis, 1: 271-276.

Vogt, P. F., Bos, T. J. and Doolittle, R. F., (1987), Homology between the DNA
binding domain and the GCN 4 regulatory protein of yeast and the carboxyl-
terrninal region of a protein coded for by the oncogene jun, Proc. Natl. Acad.
Sci. U.S.A, 84: 3316-3319.

Wanghorne, C., Thomas, M., Lagarde, A., Kerbel, R. 5., Breitrnan, M. L.,
(1988), Genetic evidence for progressive selection and overgrowth of primary
tumors by metatstatic cell subpopulations, Cancer Res, 48 (21): 6109-6114.

Waterfield, M. D., (1986), in Cellular Molecular Biology of Cancer, Ed. Frank L.
M. and Teich, N., Oxford Science Pub. ‘

Waterfield, M. D., Scarce, G. T., Whittle, N. Stroobant, P., Johnson, A.,
Wasteson, A. Westermark, B., Heldin, C. H., Huang, 1. 5., Deuel, T. F., (1983),
Platelet-derived growth factor is structurally related to the potential
transforming protein p28 sis of simian sarcoma virus, Nature, 304: 35-39.

WHO (192), IARC, (1982), Monograph on the evaluation of the carcinogenic
risk of chemicals to humans, Supplement 4.

WHO, (1969), Principles for the Testing and Evaluation of Drugs for
Carcinogenicity , WHO Tech. Report N o. 426, Geneva, World Health
Organization

121

WHO, (1978), Carcinogenicity and mutagenicity, In Principles and Methods
for Evaluating the Toxicity of Chemicals Part 1 Geneva, World Health
Organization p. 236-272.

Whyte, P., Buchkovich, K. 1., Horowitz, J. M., Friend, S.H., Raybuck, M.,
Weinberg, R A., Harlow, E., (1988), Association between an oncogene and an
anti-oncogene: the adenovirus E1A proteins bind to the retinoblastoma gene
product, Nature, 334: 124-129.

Wigley, C., (1986), Chemical Carcinogenesis, In Cellular and Molecular
Biology of Cancer. ed. L. M. Franks and N. Teich, Oxford Scientific Pub.,
Oxford '

Williams G. and 'Weisburger, (1986), Chemical Carcinogens in Toxicology,
The Basic Science of Poisons, ed. Casarett and Doull, Macmillan Publishing
Co., NY.

Williams, G. M., Kroes, R, Waayers, H. W., van de Poll, K. W., (1980), The
predictive 'value of short-term screening tests in carcinogenicity evaluation,
Applied Methods in Oncology, Vol. 3.

Willumsen, B. M., Norris, K., Papageorge, A. G., Hubbert, N. L. and Lowry, D.,
R, (1984), Harvey murine sarcoma virus p21 ras protein: biological and

biochemical significance of the cysteine nearest the carboxy terminus, J.
EMBO, 3: 2581-2585.

Wilson, D. M., Fry, D.G., Maher, V. M., McCormick, J. 1., (1989),
Transformation of diploid human fibroblasts by transfection of N-ras
oncogenes, Carcinogenesis, 10(4): 635—640.

Wiseman, R. W., Stowers, S. 1., Miller, E. C., Anderson, M. W. and Miller,J.A.,
Activating mutations of the C-Ha-ras proto-oncogene in chemically induced
hepatomas of the male B6C3F1 mouse. Eggs. NsLl. Acad. &- USA, 83: 5825-
5829, 1986.

Yamagiwa, K. and Ichikawa, K., (1918), Experimental study of the
pathogenesis of carcinoma, 1. Cancer Res. 3: 1-29.

Yunes, J. J. and Ramsey, N., (1979), Retinoblastoma and subband deletion of
chromosome 13, Am. J. Dis. Child, 132: 161-163.

122

CHAPTER II. MUTATIONAL ANALYSIS or THE H-ras ONCOCENE lN
' SPONTANEOUS 35cm MOUSE LIVER TUMORS AND
TUMORS INDUCED WITH A GENOTOXIC AND
NONGENOTOXIC HEPATOCARCINOGENS

Tony R Foxarb, Alan M. Schumanna, Philip G. Watanabea,
Barry L. Yanoa, Veronica M. Maherb, 1. Justin McCormickb

aThe Toxicology Research Laboratory, The Dow Chemical Company,
Midland, Michigan 48674.

bCarcinogenesis Laboratory, Fee Hall, Michigan State University, East Lansing,
NH 48824-1316.

123

MACE

The frequency and mutational profile of. H-ras gene activation was
determined in spontaneous liver tumors of male B6C3F1 mice and in tumors
induced with the genotoxic hepatocarcinogen benzidine-2H0 or the
nongenotoxic hepatocarcinogens phenobarbital, chloroform or ciprofibrate.
DNA sequence analysis of the H-ras gene from representative tumors
revealed that 32 of 50 (64%) spontaneous tumors and 13 of 22 (59%)
benzidine-ZHCl induced tumors contained a point mutation in codon 61.
Tumors induced with the nongenotoxic agents had a much lower frequency
of codon 61 mutations, i.e., phenobarbital 1 of 15 (7%), chloroform 5 of 24
(21%) and ciprofibrate 8 of 39 (21 %). No mutations were observed at codons
12, 13 or 117 in tumors from any of the groups. Only three base pair
substitutions within codon 61 were found. The one most frequently detected
in all of the groups was a CC to AT transversion at the lst nucleotide
position, occurring at a 59%, 85%, 100%, 80%, and 88%, frequency in the
spontaneous tumors and in the tumors induced with benzidine 2-HC1,
phenobarbital, chloroform and ciprofibrate respectively. In these same groups
a ArT to CC transition or a ArT to T-A transversion at the second nucleotide
position occurred at a frequency of 34%, 8%, 0%, 0%, 12% and 6%, 8%, 0%,
20%, 0% respectively. The number of tumors carrying an activated H-ras gene
in the nongenotoxic treatment groups is within the range that would be
expected if those animals had not received any treatment. This indicates that

the activation of the H-ras gene in those tumors is probably the result of a

124

spontaneous event. The data suggest that these toxicologically and
pharmacologically diverse nongenotoxic hepatocarcinogens increase the
frequency of liver tumors but do not induce mutations in the H-ras gene.
Instead these agents appear to interact with a population of cells that do not
contain an activated H-ras gene. This suggests that the mechanisms of tumor
development by these nongenotoxic carcinogens differs at least partially from
the mechanisms responsible for the development of spontaneous tumors or

those induced by a typical genotoxic agent.

W

The results of the rodent bioassay play an important role in the
evaluation of whether a chemical is considered potentially carcinogenic for
humans. However, the use of such test results to predict human risk is often
complicated by biological factors peculiar to the strain of animal used. For
example, the B6C3F1 mouse is frequently used to assay various potential
carcinogens. Male B5C3F1 mice have an average spontaneous liver tumor
incidence of 20-30% (1). These mice are apparently hypersensitive to the
development of such tumors which calls into question the relevance of a
chemically induced increase in tumor frequency as a predictor of human
cancer risk. This concern is underscored by the fact that many genotoxic and
nongenotoxic agents significantly increase the frequency of liver tumors in

this animal but may not be tumorigenic at other sites or in other species (2).

125

For example, Ashby and Tennant surveyed 222 carcinogen bioassays
conducted by the NCI/NT'P (3). Of the 115 chemicals found to be tumorigenic,
nearly 40% were hepatocarcinogenic in the B6C 3F 1 mouse . For 20% of the 1 15
carcinogens the only tumorigenic response was an increase in mouse liver
tumors. Based on in m mutagenicity tests many of these tumorigenic
chemicals have been shown to have genotoxic activity (ie. damage DNA).
The mutagenic activity of these agents can be initiated from the parent
molecule or from electrophilic metabolic intermediates. It is believed that
damage to DNA resulting in somatic mutations plays an important role in
the tumorigenicity of this class of chemical carcinogens (4).

Not all carcinogenic chemicals have mutagenic activity as determined
by in vitra mutagenicity assays. An important class of chemical carcinogens,
generally referred to as nongenotoxic, do not cause mutations but still can
cause an increase in animal tumors. The mechanism by which these agents
increase mouse liver tumors is presently unclear. They may exert their
carcinogenic activity by promoting previously inititated cells (5).
Alternatively, these agents may induce somatic mutations by indirectly
damaging DNA as a consequence to some abnormally induced physiologic
response. For example, this secondary damage to DNA may occur as a result
of the interaction of reactive oxygen radicals with DNA generated by the
infiltration of polymorphonuclear leukocytes as a consequence to chemically
induced cellular toxicity (6).

Much of the controversy surrounding the interpretation of bioassay

results concerns the relevance to human risk from exposure to nongenotoxic

126

agents that cause an increase in mouse liver tumors. In an attempt to gain a
better understanding of how liver tumors arise in the B6C3F1 mouse and how
this tumor incidence is increased by chemical exposure, several groups have
examined the role of cellular oncogenes in the hepatocarcinogenic process of
the B6C3F1 mouse. It was demonstrated. that the H-ras oncogene is activated
in a high percentage of both spontaneous tumors and tumors induced with
several genotoxic hepatocarcinogens (7-10). In addition Wiseman et al (9) and
Reynolds et al (10) showed that certain genotoxic carcinogens cause chemical
specific mutational patterns within the H-ras gene.

To investigate the mechanisms by which nongenotoxic
hepatocarcinogens enhance liver tumor development in the B6C3F1 mouse,
we determined the frequency and mutational profile of H-ras gene activation
in tumors that arose spontaneously and in tumors induced by exposure to a
genotoxic or several nongenotoxic carcinogens. This was carried out by
amplifying regions of exons 1, 2, and 3 of the H-ras gene which include the
activating codons 12, 13, 61 and 117 from tumor DNA using the polymerase
chain reaction technique (PCR). The amplification products were subjected to
direct DNA sequence analysis to identify point mutation within these
activating codons. The results showed that tumors induced with the
nongenotoxic hepatocarcinogens phenobarbital, chloroform or ciprofibrate
have a much lower frequency of H-ras gene activation than do those that
arose spontaneously or those induced with the genotoxic carcinogen

benzidine-2 HCl.

127

MATERIALS AND METHODS

Tumor Induction. Spontaneous liver tumors were obtained from control
male B6C3F1 mice collected at the terminal necropsy of 2-year bioassays for
carcinogenicity conducted at the Environmental and Toxicology Research
Laboratory of The Dow Chemical Company (Midland, MI). Chemically
induced liver tumors were obtained from independently treated groups of
male B6C3F1 mice administered either benzidine-2 HQ (120 ppm, drinking
H20, 1 year), phenobarbital (0.05%, drinking H20, 1 yr), chloroform (200
mg/ kg corn oil gavage, 2X weekly for 1 yr) or ciprofibrate (0.0125% diet, 2 yr).
Animals were put on test at 6 weeks of age and necropsied after 18-24 months.
At necropsy tumors were excised and a section of each tumor was fixed in
neutral phosphate buffered 10% formalin. Tissues were processed by
conventional techniques, embedded in paraffin, cut at approximately 6pm,
stained with hematoxylin and eosin and examined using a light microsc0pe.
The remainder of the tumor was frozen on dry ice and stored at -70°C until

DNA was isolated.

Oligonucleotide Primer Preparation. Oligonucleotide primers used in PCR
amplification and DNA sequencing were synthesized using the
phosphoramidite method on an Applied Biosystems Model 380A DNA
synthesizer. Detritylated primers were purified by HPLC using a 101i RP-300
reverse phase column (Brownlee Laboratories). Elution of primers from the

column was achieved with a linear 7-17% acetylnitrile gradient (50 min., 1

128

ml/ min.) in 0.1 M tetramethylammonium acetate (pH 7.0). HPLC fractions
containing the purified primer were pooled, evaporated to dryness,
reconstituted with distilled H20 and purified by ultrafiltration using a
Centricon-3 (Amicon, W. R. Grace, Danvers, Ma.).

PCR Amplification. Total genomic DNA was isolated from liver tumor
tissue as previously described (5). Prior to PCR amplification approximately 2
ug of tumor DNA was digested with 2 units of Bam HI restriction
endonuclease (37°C, 30 min) as recommended by the manufacturer (Bethesda
Research Laboratories). Restriction endonuclease digestion of the genomic
DNA resulted in an increase amplification yield of the H-ras gene sequences
presumably by increasing the template accessibility to the PCR polymerizing
enzyme (unpublished studies, T. Fox).

PCR amplification of the digested tumor DNA was conducted in a total
volume of 50p] containing 50 mM KCl, 10 mM Tris-HCl (pH 8.3), 15 mM
MgC12, 400 M dATP, dT'I'P, dGTP, dCTP, approximately 1 ug each of 3' and 5'
oligonucletotide primers and 2.5 units of Taq polymerase (Perkin-Elmer
Cetus)(10). The reaction mixture was incubated for 5 min at 94°C to ensure
complete denaturation of genomic DNA. Primers were than allowed to
hybridize to the template DNA at 65°C for 1 min. Amplification was initiated
with the addition of Taq enzyme followed by 30 cycles of successive
incubations for 1 min each at 65°C (annealing) 70°C (polymerization) and

94°C (denaturing). Confirmation of amplification was assessed by agarose gel

129

electrophoresis. 10111 of the PCR reaction mixture was loaded onto a 3%
Nusieve (FMC) 1% agarose gel and DNA bands were visualized by ethidium

bromide staining.

Sequencing. The protocol for DNA sequence analysis was adopted from that
of Yang e_t s1. (11). Following the termination of amplification the PCR
reaction mixture was purified by centrifugation through a Centricon-30 to
remove unincorporated nucleotides and excess oligonucleotide PCR primers.
The Centricon-30 retentate was sequenced directly using a modification of the
Sanger dideoxy chain termination protocol (12 ). Five microliters of the PCR
amplified sample was added to 15111 of a solution containing 40 mM Tris-HCl
(pH 7.5), 20 mM MgC12, 50 mM NaCl and approximately 5 pmoles of 32P end-
labeled sequencing primer.

Sequencing primers were end-labeled with (7-32P) ATP by incubation in 60
mM Tris-HCl (pH 7.8) 10 mM MgC12, 330 mM ATP, 15 mM 2-
mercaptoethanol and 15 units of polynucleotide kinase at 37°C for 30 min.
The reaction mixture was centrifuged through a N u-Clean D25 disposable
spin column aspdescribed by the manufacturer (IBI). Specific activities of the

sequencing primers routinely ranged from 1-5 X 108 CPM/ ug.

The mixture containing the amplified DNA and sequencing primer was
heated for 5 min. at 96°C, cooled on ice, and 2 units of sequenase (United

States Biochemical) was added. The sequencing reaction was initiated by

130

adding 3.5 ul of this mixture to 1 of 4 tubes containing 2.5 ul of the '
appropriate combination of deoxynucleotides/ dideoxynucleotide at a 10:1
ratio. Deoxynucleotide concentrations were 33 Md and dideoxynucleotides
were 3.3 M. The tubes were incubated at 37°C for 15 min. and the reaction
was terminated by the addition of 4 pl of 95% formamide containing 0.05%
bromophenol blue and xylene cyanol FF tracking dyes. The reaction tubes
were heated at 96°C for 3 min, cooled on ice and 2.5 ul was loaded onto a 8%
polyacrylamide/BM urea gel and electrophoresed at 40 m Amps. The gel was
transferred to Whatrnan 3 mm paper, dried under vacuum and exposed to
Kodak XAR-S film at -70°C. for 24-48 hrs.

RESULTS

Tumor Induction. Tissue used for analysis in this study was derived from
male B6C3F1 mice liver tumors occurring spontaneously or induced by
treatment with the genotoxic hepatocarcinogen benzidine-2 HCl or the
nongenotoxic hepatocarcinogens phenobarbital, chloroform or ciprofibrate.
Mice bearing spontaneous tumors, obtained from the control groups of 6
different bioassays, had an average tumor incidence of 42% with 22% of the
tumor bearing animals having multiple liver tumors (Table 1). A treatment
related effect was observed for all compounds within the treatment groups as
indicated by the increase in percent tumor incidence and tumor multiplicity
compared to the control animals (Table ’1). Tumors ranged in size from

approximately 1 mm to 2.5 cm in diameter.

131

H-ras Gene Activation. The activation of the H-ras gene in a variety of
human and rodent tumors has been shown to involve point mutations at
codons 12, 13, 61, or 117 (10,14). The extent of H-ras gene activation in the
B6C3F1 mouse liver tumors of the various treatment groups was assessed by
determining the presence or absence of point mutations within these
activating codons. Codons 12 and 13, 61 and 117 reside in exons 1, 2 and 3,
respectively, of the H-ras gene (15). Point mutations in these codons were
identified by direct sequence analysis of amplified regions within the
corresponding exons from genomic tumor DNA. These exon regions were
amplified using the polymerase chain reaction (PCR) technique. The
amplified regions of the H-ras gene with the corresponding PCR and
sequencing primers are illustrated in Figure 1.

Sequence analysis of the H-ras gene was carried out on a total of 151
tumors from the combined groups. Mutations were detected only in codon 61
and not at codons 12, 13 or 117 in any of the tumors. As a result the
percentage of tumors with a mutation in codon 61 is equivalent to the
percentage of tumors with an activated H-ras gene.

As can be seen in Table 2, the percentage of tumors with an activated
H-ras gene in the spontaneous and benzidine-2 HCl-induced tumors is
considerably higher than in the tumors induced by treatment with
phenobarbital, chloroform or ciprofibrate. Analysis of spontaneous tumors
revealed that 64% contained a point mutation in codon 61. A similar

response was observed in the benzidine-2 HCl-induced tumors, where 59%

132

tumors had codon 61 mutations. This is in contrast to the tumors induced by
the nongenotoxic agents where activating H-ras gene mutations were detected
at a much lower frequency, i.e., phenobarbital 7% , chloroform 21 % and
ciprofibrate 21 %..

Histopathological examinations were conducted on all tumors except
those that were too small to forfeit tissue for analysis. A correlation between
the incidence of H-ras gene activation and the development of either a
hepatocellular adenoma or hepatocellular carcinoma was made and the data
summarized in Table 3. There was no statistically significant difference
between the frequency of H-ras gene activation in the hepatocellular
adenomas and carcinomas (statistical analysis determined using a chi square
contingency table test). Histopathological examination of the spontaneous
tumors and the tumors induced with benzidine-ZHCI, phenobarbital, and
chloroform did not reveal any significant differences in morphology or
staining characteristics. The ciprofibrate induced tumors were more
eosinophilic as were the surrounding normal hepatocytes. This observation
would be expected from a treatment related increase in cytoplasmic
peroxisomes.

The mutational spectrum of activating codon 61 mutations within the
H-ras gene was determined and the data are summarized in Table 4. The
normal nucleotide sequence for codon 61 in the mouse H-ras gene is CAA.
The activating mutations observed in all tumor groups involved a single base

pair substitution changing this sequence to AAA, CGA or CTA.

133

Representative examples of sequence gels exhibiting these three mutations
are shown in Figure 2.

The most frequently detected mutation was a C-G to AT transversion
at the lst nucleotide position. The frequency of this base pair substitution in
the spontaneous tumors and in the tumors induced with benzidine 2-HCl,
phenobarbital, chloroform and ciprofibrate was 59%, 85%, 100%, 80% and 88%
respectively. An A-T to CC transition or a A-T to T-A transversion at the
second nucleotide position occurred in these same groups at a frequency of
34%, 8%, 0%, 0%, 12%, and 6%, 8%, 0%, 20%, 0% respectively.

Many of the animals in the chemically induced tumor groups had
multiple tumors as is indicated in Table 1. A comparison of the H-ras codon
61 nucleotide sequence of the individual tumors from animals with multiple
tumors revealed that in approximately one-half of those animals examined a

different codon 61 sequence was observed.
DISCUSSION

Mutational analysis of tumor DNA from spontaneous tumors and
tumors induced with benzidine-2 HCl, phenobarbital, chloroform, or
ciprofibrate revealed activating mutations exclusively in codon 61. Only 3
activating point mutations within this codon were detected; a CG to A-T
transversion at the lst nucleotide position, an A-T to G-C transition or an A-T

to T-A transversion at the second nucleotide position. These point mutations

134

resulted in the amino acid substitution of glycine in the H-ras P21 protein
with lysine, arginine or leucine respectively.

The predominant mutation in all treatment groups was the point
mutation converting the normal codon 61 sequence of CAA to AAA. It has
been demonstrated by Strauss et a1 (17) that during DNA synthesis on a
noninstructional template the most frequent polymerase error involves the
preferential insertion of an adenine, the"A rule". In the spontaneous
tumors the high frequency of CC to AT transversions observed at this
nucleotide position may be attributed to this type of polymerase error.

The frequency of AAA codon 61 mutations in DNA from tumors taken
from benzidine2 HCl treated animals was slightly higher than that observed
in the DNA from spontaneous tumors. Chi square analysis showed that this
increase was not statistically significant. Some of the tumors carrying this
Specific mutation could be of spontaneous origin considering the high
frequency of this mutation observed in spontaneous tumors. However, this
increase could be due to the mutagenic activity of benzidine- 2HC1.
Benzidine ~2HCl is a potent genotoxic agent which has been shown to cause
primarily C8 adducts on guanine (18). The formation of such an adduct on
the guanine opposite cytosine at the first position in codon 61 could be
expected to cause a CG to AT transversion. This transversion could occur
either by the preferential insertion of an adenine across from a
noninstructive base or as a result of the adduct allowing stable base pairing
between the modified guanine with an adenine as was reported recently for

guanine carrying an AP residue at the C8 position (19).

135

Several investigators have also determined the mutational spectra of
codon 61 within the H-ras gene from B6C3F1 mouse liver tumors induced
with various genotoxic carcinogens. Wiseman et al. (9) induced B6C3F1 liver
tumors with N-hydroxy-Z-acetylaminofluorene, vinyl carbamate, or 1'-
hydroxy-2',3', dehydroestragole. These agents produce codon 61 mutational
spectra that are distinct from the mutational spectra observed in spontaneous
tumors. A similar analysis by Reynolds et al. (8) of tumors induced with
furan and furfural observed mutations not only in codon 61 but at codons 13
and 117, two activating codons that have not been detected previously in
tumor DNA. The carcinogenic agents used in these studies therefore produce
unique chemical mutation spectra within the H-ras gene which is consistent
with their apparent genotoxic activity. It has been suggested that the
evaluation of the mutational spectra within the H-ras gene of B6C3F1 mouse
liver tumors may be useful in determining the in viva genotoxic potential of
an agent which is positive for tumorigenicity in the rodent bioassay (10). This
type of analysis may prove valuable in selected circumstances but there are
limitations that must be realized; This is illustrated by the analysis of liver
tumors generated in the B6C3F1 mouse with the potent mutagenic carcinogen
N-diethylnitrosamine. The codon 61 mutation spectrum observed in these
tumors was similar to the mutation spectrum observed in spontaneous
tumors (20). Consequently, the presence or absence of genotoxic activity for
this particular chemical cannot be concluded using this procedure. This type
of analysis is useful only if the mutational spectra of the chemically induced

tumors deviates significantly from the spontaneous tumor spectrum.

136

Because the spontaneous tumor spectrum is used as a benchmark for this
analysis, a high degree of confidence in this spectrum is an obvious
prerequisite.

One of the objectives of the present study was to examine a large number
of spontaneous tumors to increase the statistical confidence in this spectrum.
Our results have indicated that the mutational spectrum observed in the
spontaneous liver tumors is similar to the mutational spectrum previously
reported for 15 spontaneous tumors (10). The inclusion of this data to the
preexisting data base should increase the degree of confidence in this spectrum.

Analysis of the codon 61 mutational spectra of the tumors induced with
the nongenotoxic agents used in this study was not meaningful because so few of
the tumors contained codon 61 mutations. However, it was noted that the most
frequently observed mutation in these tumors, a CC to A-T base transition at the
first nucleotide position, was the most common mutation detected in the
spontaneous tumors.

Analysis in the frequency of H-ras gene activation did reveal significant
differences between the tumors induced with the nongenotoxic carcinogens
compared to the spontaneous tumors or tumors induced with the genotoxic
agent benzidine - 2HCL. The H-ras gene was activated to a similar frequency in
spontaneous and benzidine -2 HCl induced tumors (64% and 59% respectively).
The frequency of H-ras gene activation was significantly lower in tumors
induced with phenobarbital, chloroform, and ciprofibrate (7%, 21 %, 21%
respectively). If one determines the number of tumors within the nongenotoxic

induced groups that would be of spontaneous origin (42%), and then calculates

137

the subgroup of these that can be expected to exhibit an H-ras activation, i.e. 27%
(64% of 42%), this value approximates the number of tumors observed within
these groups having an activated H-ras gene. In other words, the number of
tumors in these groups that contained an activated H-ras gene is the number that
would be expected if those animals had received no treatment. This suggests that
the tumors induced with the nongenotoxic agents having an activated H-ras
gene are probably of spontaneous origin and, therefore, the increase in tumor
frequency above background caused by exposure to these nongenotoxic agents
results from their interaction with a population of non-H-ras activated cells. We
concluded that these nongenotoxic agents do not cause mutations within the H-
ras gene, and that they increase the frequency of mouse liver tumor
development by a mechanism that differs from the process occurring
spontaneously or after treatment with a genotoxic carcinogen.

Activation of this gene was detected in hepatocellular adenomas as small
as 1 mm in diameter. Balmain et al (21) have detected an activated H-ras gene in
chemically induced benign mouse skin papillomas. The detection of an
activated H-ras gene in very small hepatocellular adenomas as well as in benign
skin papillomas suggests that activation of this gene may be an early initiating
event in tumorigenesis. The high frequency of H-ras gene activation in
spontaneous tumors also indicates that this gene plays an important role in
B6C3F1 mouse liver tumor development. However, in approximately 35% to
40% of the spontaneous and benzidine.2HCl induced tumors and in
approximately 80% to 90% of the tumors induced with the nongenotoxic agents

did not have a detectable activated H-ras gene. This indicates that alternative

138

pathways independent of H-ras gene activation can also lead to hepatic tumor
development. The data suggest that there are at least two populations of cells
which can give rise to tumors within the liver, one consisting of cells having an
activated H-ras oncogene and another which does not. The data do not preclude
the possibility that cells giving rise to tumors that do not have an activated an H-
ras gene may have other activated ras genes. (i.e. K-ras, N-ras). Reynolds st 31.,
did detect an activated K-ras gene in 7% and 15% of liver tumors in the B6C3F1
mouse induced with furfural and furan respectively (10).

The three nongenotoxic agents used in this study have diverse
pharmacologic and toxicologic activities. Phenobarbital has general hepato-
stimulatory properties resulting in increased gene expression with an
associated induction of cellular proliferation (22). Animal tumors induced
with chloroform are accompanied by cell necrosis, death, and subsequent
regeneration (23). Although cellular proliferation alone is probably not
sufficient to cause tumors it has been shown to be a necessary prerequisite for
the fixation of the initiating event in cell culture and it is believed to be
important for the clonal expansion of the initiated cell population in viva
(24, 25). Peroxisome proliferators, such as ciprofibrate, have been
hypothesized to be responsible for tumor induction by various mechanisms.
One of these suggests that they induce indirect secondary damage to DNA as a
consequence of an increased production of oxygen free radicals, however,
ciprofibrate has not been shown to cause mutations in vitra (26).

Alternatively it has been proposed that such agents stimulate the growth of

preneoplastic foci suggesting that they are acting as tumor promoters (27).

139

Even though the biological activity of these nongenotoxic animal tumorigens
is diverse, their tumorigenic activity is only observed at doses that cause
perturbations in the cells' normal biochemistry and/ or physiology. This
implies that at lower doses they will not cause tumors.

In summary, results from the present study indicate that the
nongenotoxic agents phenobarbital, chloroform and ciprofibrate induce liver
tumors in the B 6C3F1 mouse by a process that does not require the activation
of the H-ras gene. This data suggests that at least two different populations of
cells, one requiring H-ras activation and the other independent of this event,
can give rise to liver tumors in these animals. The nongenotoxic agents used
in this study apparently increase B6C3F1 mouse liver tumors by interacting
with a population of cells that do not contain an activated H-ras gene.
Additional studies comparing tumors with and without an H-ras gene
mutation will be required to elucidate the molecular and cellular differences
between these cell populations. Understanding these differences should
provide insight into how various classes of carcinogens interact with cells to
give rise to liver tumors in the B6C3F1 mouse. Such results will be useful in
extrapolating bioassay results using this strain of mouse to evaluate potential

chemical carcinogenic risk for humans.

140

ACKNOWLEDGEMENTS

The authors would like to thank our colleague Dr. Jia-ling Yang, presently at

the Institute of Biomedical Sciences, Academia Sinica, Taipei, Taiwan for here
advice on the DNA sequence analysis. We also thank Dr. Peggy Schott for her
technical advice concerning the DNA oligonucleotide synthesis, Sherry Pagels

for technical assistance and Jean Shuler for typing the manuscript.

141

REFERENCES

Chen, C., NTP technical report on the toxicology and characterization
studies of dimethylmorphoenophosphoramidate in F344 / N rats and
B5C3F1 mice. Cas No. 597-25-1, 1984.

Ward, J. M., Goodman, D. G., Squire, R A., Chu, K. C., and Linhart, M.S.,
(1979), Neoplastic and nonneoplastic lesions in aging (57BL/ 6N x
C3H/HeN) F1 (B6C3F1), J. Natl. Cancer Inst, 63(3): 849-854.

Ashby, J. and Tennant, R. W. Chemical structure, Salmonella
mutagenicity and extent of carcinogenicity as indicators of genotoxic
carcinogenesis among 222 chemicals tested in rodents by the US.

NCI/NTP. Mutation Bis, 204:17-115, 1988.

Cellufilr and Molecul_ar Biology of Cancer, Editors Franks, L. M., and
Feich, N., Oxford University Press, New York, NY, 1986.

Schulte-Hermann, R, Timmermann-Trosiemer, I., Schuppler, 1., (1983),
Promotion of spontaneous preneoplastic cells in rat liver as a possible
explanation of tumor production by nonmutagenic compounds, Cancer

Res, 43: 839-844.

10.

142

Frenkel, K., and Chrzan, K., (1987), DNA base modification by tumor
promoting activated polymorphonuclear leukocytes, Carcinogenesis 8:
455-460.

Fox, T. R, Schumann, A. M., Watanabe, P. G. Activation of a cellular
proto-oncogene in spontaneous liver tumor tissue of the B6C3F1 mouse.

Arch Toxicol. Suppl., 10:217-227, 1987.

Reynolds, 5. H., Stowers, S. 1., Maronpot, R R, Anderson, M. W. and
Aaronson, S. A. Detection and identification of activated oncogenes in
spontaneously occurring benign and malignant hepatocellular tumors of

the 86C3F1 mouse. Proc. Natl. Acad. _S£i_. USA, 86:33-37, 1986.

Wiseman, R W., Stowers, S. 1., Miller, E. C., Anderson, M. W. and
Miller, 1. A. Activating mutations of the C-Ha-ras proto-oncogene in
chemically induced hepatomas of the male B6C3F1 mouse. PLog. Na_tl.
_A_csc_l. S_c_i. USA, 83:5825-5829, 1986.

Reynolds, 5. H., Stowers, 5.1., Patterson, R M., Maronpot, R R,
Aaronson, S. A. and Anderson, M. W. Activated oncogenes in B6C3F1

mouse liver tumors: implications for risk assessment. Science, 237:1309-

1316, 1987.

11.

12.

13.

14.

15.

16.

143

Mullis, K. and Faloona, F. Specific synthesis of DNA in vitra via a
polymerase catalyzed chain reaction. Methods Enzmol., 155:335-350,
1987.

Yang, 1-L., Maher, V. M., McCormick, J. 1. Amplification and direct
mucleotide sequencing of cDN A from the lysate of low numbers of

diploid human cells. Gene (in press).

Sanger, F., Nicklen, S. and Coulson, A. R DNA sequencing with chain
terminating inhibitors. Proc. Natl. Acad. E. USA, 74:5463-5467, 1977.

Barbacid, M. Ras genes, Annual Rev. Biochem., 56:779-827, 1987.

Fasano, O., Taparowsky, E., Fiddes, 1., Wigler, M. and Goldfarb, M.
Sequence and structure of the coding region of the human H-ras-l gene
from T24 bladder carcinoma cells. _1. Molecular and Applied §e_nt. 2:173-
180, 1983.

Reddy, E. P., Lysman, D., Andersen, P. R, Tronick, S. R. and Aamson, S.
A. Nucleotide sequence analysis of the Balb/c murine sarcoma virus

transforming gene. 1. Virology, 53:984-987, 1985.

17.

18.

19.

20.

21.

144

Strauss, B., Rabkin, F., Sagher, D., Moore, P. The role of DNA
polymerase in base substitution mutagenesis on noninstructional

templates. Biochimie 674:829-838, 1982.

 

Josephy, P. D., Benzidine: mechanism of oxidative activation and

mutagenesis. m. Proc., 45(10):2465-2470, 1986.

Norman, D., Abuaf, P., Hingerty, B.F., Lide, D., Grunberger, D. ,Broyde. S.
and Patel, D. J. N MR and Conformational characterization of the N-
(Deoxyguanosine 8-yl)aminofluorine adduct-[(AF)G] opposite adenosine
in DNA; (AFG[syn]- A[anti] pair formation and its' pH dependence.
Biochemist_ry, 28: 1989 (in press)

Stowers, S. J., Wiseman, RW., Ward, 1.M., Miller, E.C., Miller, I.A.,
Anderson, M.W., Eva, A., (1988), Dection of activated proto-oncogenes in
N-nitrosodiethnylamine induced liver tumors: a comparison between

B6C3F1 mice and Fischer 344 rats, Carcinogenesis, 9 (2): 271-276.

Balmain, A., Ramsden, M., Bowden, G. T, and Smith, 1. Activation of
the mouse cellular Harvey-ras gene in chemically induced benign skin

papillomas. Nature 307:658-660, 1984.

Peraino, C., Fay, R. J. M., Staffeldt, E. and Christopher, J. P. Comparative

enhancing effects of phenobarbital, amobarbital, diphenylhydantoin and

24.

26.

27.

145

dichlorodiphenyl-trichloroethyane on 2-acetylaminofluorene induced
hepatic tumorigenesis in the rat. Cancer Research, 35:2884-2890 (1975).

Reitz, R H., Fox, T. R and Quast, J. F. Mechanistic considerations for
carcinogenic risk estimation: chloroform. Envirn. Health Perspectives,

46:163-168 (1982).

Kakunaga, T. The role of cell division in the malignant transformation
of mouse cells treated with 3-methylcholanthrene. Cancer Research

35:1637-1642 (1975).

Mironescu, 5., Love, R. DNA synthesis and transformation induced in
density-inhibited cultures of hamster embryo cells by the carcinogen
benzo(a)pyrene. Cancer Research 34:2562-2570 (1974).

Reddy, 1. K and Lalwani, N. D. Carcinogenesis by hepatic peroxisome
proliferators: Evaluation of the risk of hypolipidemic drugs and
industrial plasticizers to humans. CRC Crit. Rev. Toxicol. 12:1-58 (1983).

Cattley, R. C. and Popp, J. A. Differences between the promoting
activities of the peroxisome proliferator WY-14,643 and phenobarbital in

rat liver. Cancer Research, 49:3246-3251 (1989)

146

Figure Legend
Figure 1. . Nucleotide sequence of the coding strand for the H-ras gene

Figure 2.

which was PCR amplified and subsequently sequenced.
Fragment I represents the region of exon 1 containing codons 12
and 13. Fragment II is region of exon 2 containing codon 61.
Fragment III is the entire sequence of exon 3 containing codon
117. The darkly shaded areas indicate position of the 5' and 3'
PCR primers. Lightly shaded areas represent the sequencing

primers.

Representative DNA sequence of the H-ras gene codon 61 region
derived from PCR amplified liver tumor DNA.
(a) Normal codon 61 sequence (CAA) derived from a

ciprofibrate induced tumor.

(b) CG —> A-T transversion mutation at codon 61 (AAA)

from ciprofibrate induced liver tumor.

‘ (c) A-T _> G-C transition (CGA) from benzidine-ZHCl

induced liver tumor.

(d) A-T —> T-A transversion (CTA) from spontaneous liver

tumor.

147

EmEca 0:65:08 3
295a mom . ..

m8_> mEoocmw 2:56 928 E0: 32.8 mm ..

 

 

 

 

 

  

 

I I I a a c a i
.
n t e t. o a e a s . es.

148

Figure 2.

ng

 

149

Table 1: Incidence of spontaneous and chemcially induced

liver tumors in male B5C3F1 mice.

 

% of
tumor
bearing

% of animals
Treatment Animals animals with
used to induce Number of with with . multiple
tumors animals tumorsa tumors tumors
None 293 124 42 22
(Spontaneous)
Benzidine - 27 23 85 39
2HC1
Phenobarbital 25 17 68 65
Chloroform ‘ 17 15 80 80
Ciprofibrate 31 29 94 79

a) All animals were necropsied after 18-24 months of age and liver tumors
were removed for further analysis.

150

Table 2: Frequency of activation of the H-ras gene at codon 61 in
male B5C3Fl mouse liver tumors.

 

Treatment used to Number of tumors Tumors with an
induce tumors analyzed“ activaed H-rasa
None 50 32 (64%)
(Spontaneous)
Benzidine ~2HCl 22 ' 13 (59%)
Phenobarbital 15 1 (7%)
Chloroform 24 5 (21%)
Ciprofibrate 39 8 (21%)

a) Total genomic DNA was isolated from liver tumor tissue. Defined
regions of exon 1, 2, and 3 of the H-ras gene incorporating codons 12, 13, 61,
and 117 were PCR amplified and sequenced directly. The sequence was
determined for 97 of the total 190 codons of the H-ras gene.

151

Table 3: Frequency of H-ras gene activation in hepatocellular
adenomas and carcinomas in male B5C3F1 mice.

% of Adenomas with % of Carcinomas with

 

Treatment used to activated activated
induce tumors H-rasa H-rasa
None

(Spontaneous) 61% (11/18) 64% (9/ 14)
Benzidine -2HCl 53% (8/ 15) 67% (4/6)
Phenobarbital 1 1 % (1 / 9) 0% (0/ 5)
Chloroform 15% (2/ 13) 33% (3/ 9)
Ciprofibrate 21% (6/ 29) 22% (2/ 9)

a) A portion of each tumor was excised, fixed at necropsy in neutral phosphate
buffered 10% formalin, processed by standard techniques, and embedded in
prarfin. Sections were cut (6 pm) , stained, and evaluated by light microscopy.

152

Table 4: Mutation spectrum in codon 61 (CAA) of the H-ras gene in
male B5C3F1 mouse liver tumors.

Number of tumors with specific

 

Treatment used . Tumors with an codon 61 mutations (%)a
to induce activated H-ras
tumors gene AAA CGA CTA
None 32 19 (59) 11 (34) 2 (6)
(spontaneous)

Benzidine . 2HCl 13 11 (85) 1 (8) 1 (8)
Phenobarbital 1 1 (100) 0 0
Chloroform 5 4 (80) 0 l (20)
Ciprofibrate 8 7 (88) 1 (12) 0

a) Codon 61 sequence was determined as described in Table 2.

153

CHAPTERHLAMQDELEQW
nu rant! 0' tie 01L .t - MNT

Determining the mechanisms involved in the development of B6C3F1
mouse liver tumors is not only important for interpreting cancer bioassay
data, but may also provide a useful model system for understanding
carcinogenesis in general. Even though the specific events of neoplastic
development in the mouse liver have not been completely elucidated, results
from a number of recent studies have provided important insights into the
process of spontaneous and chemically induced hepatocarcinogenesis in this
strain of mouse.

The B6C3F1 mouse is a hybrid obtained through the genetic cross
between a male C3H/He1 and a female C57BL/61. These two parental strains
have widely differing susceptibilities to the development of both spontaneous
and chemically induced liver tumors. Within 15 months of age nearly 50% of
the C3H/HeJ mice will develop spontaneous liver tumors (Andervant, 1950).
In contrast less that 5% of the C57BL/ 61 mice will have liver tumors by 2 years
of age (Grasso and Hardy, 1975). This difference in spontaneous tumor
susceptibility is also exhibited in the chemical induction of liver tumors. A 15
fold greater multiplicity of tumor formation is observed in the C3H/ He]
mouse compared to the C57BL/ 6] mouse after treatment with 4-
aminoazobenzene or N,N-dimethyl-4-aminoazobenzene (Declos et al., 1984).

The C3H/HeJ mouse is also 20-50 fold more susceptible than the C57BL/6J

154

mouse to the induction of liver tumors by N-ethyl-N-nitrosourea (ENU) or
N,N-diethylnitrosamine (DEN) (Drinkwater and Ginsler, 1986).

The susceptibility of the C3H/ He] mouse to develop both spontaneous
and chemically induced liver tumors is presumably due to a genetically
transmitted predisposition gene(s). A study to delineate the genetic basis for
this susceptibility was recently reported by Drinkwater (Drinkwater and
Ginsler, 1986). DEN induced tumor multiplicity was used as a genetic marker
in segregating cross analysis between various susceptible and resistant strains.
In these studies it was determined that tumor susceptibility was due to allelic
differences for at least 2 independent loci (Drinkwater and Ginsler, 1986). A
single locus, however, appears to be responsible for approximately 85% of this
sensitivity. This locus has been denoted the Hcs gene (Hepato-carcinogen
sensitivity).

The dominant and recessive character of this locus was evaluated by
exposing a series of recombinant inbred mice strains to ENU and determining
the relative susceptibilities to develop liver tumors. Three different
susceptibility phenotypes were observed. One was very sensitive, one was
resistant and one was intermediate to the others. This data indicates that the
alleles of the Hcs locus in the C57BL/ 61 and C3H/HeJ mice are semi-
dominant.

The susceptibility of the B6C3F1 mouse to develop spontaneous and
chemically induced tumors falls between the two parents (Becker, 1982,
Drinkwater and Ginsler, 1986). The intermediate tumor sensitivity of this

mouse is a logical outcome given the semi-dominant nature of the Has gene.

155

At what stage in the carcinogenic process does this gene exert its
influence? To address this issue, Drinkwater and his collaborators conducted
several studies ‘to determine whether the H as locus affects initiation or post-
initiation events in the liver. Newborn C5le 6] and C3H/ He] mice were
injected with [C 14] DEN and the relative levels of DNA alkylation was
determined. The results demonstrated that there was no significant
difference in the extent of ethylation of hepatic DNA, or in the relative levels
of N-7-ethylguanine or Oé-ethylguanine adducts between the two strains
(Drinkwater and Ginsler, 1986). The genetic influence of this locus does not,
therefore, appear to be at the level of DNA modification by hepatic
carcinogens. This suggests that the Has gene may be acting at the promotion
stage of liver tumor development.

To investigate this possibility, the comparative development of
putative preneoplastic lesions induced with ENU was made between each
strain (Hanigan et al., 1988). In both strains an increase in the number and
size of these lesions was observed. At 12 weeks of age there was
approximately a 25 fold greater number of foci in the C3H/ He] mice compared
to the C57BL/6J mice . This data could be interpreted to mean that the Hcs
gene is influencing the initiation stage of tumor development. The
difference however had decreased to only 5 fold by 20 weeks of age. On the
other hand, the ratio in the mean diameter of these lesions between the
C3H/ He] and C57BL/6J mice continued to increase with time, indicating that
the foci in the C3H/ He] mice were growing faster than the foci in the

C57BL/6J mice. This rate of growth was determined to be about 70% greater

156

in the C3H/ He] mice compared to the C57BL/6J mice. Taking into
consideration both the divergence over time in the size of the foci with the
convergence in the total number of foci, the data suggests that the Hcs gene is
probably acting at the level of promotion. From these data it was concluded
that the greater susceptibility of the C3H/ He] and B6C3F1 mice to develop
liver tumors compared to the C57Bl/6] mouse is the result of a strain
dependent difference in the rate of grth in preneoplastic lesions caused in
part by the Has gene. .

It is now evident that neoplastic development within the liver of the
B6C3F1 mouse involves the participation of more than the Hcs gene. As
indicated in the discussion from previous sections of this thesis, the H-ras
oncogene is found in a majority of spontaneous tumors and tumors induced
by exposure to various genotoxic carcinogens. This oncogene appears to play
an important role in the development of these tumors and most likely at the
initiation stage of tumor development. It does not however appear to be
essential for hepatocarcinogenesis because not all of these tumors contain an
activated H-ras gene. This is even more evident in tumors induced with
certain nongenotoxic carcinogens where the vast majority do not contain this
activated oncogene. The transforming function exerted by the H-ras gene
apparently can be functionally substituted for by the activation of some other
gene(s).

Given this information, what type of model can be constructed to
explain the sensitivity of the B6C3F1 mouse to develop liver tumors after

exposure to various nongenotoxic agents? Many nongenotoxic carcinogens

157

exhibit their tumorigenic action only at doses which are high enough to
induce cellular toxicity. As was discussed in the section on nongenotoxic
carcinogens, an enhanced level of cell division can occur as a consequence of
a toxicity induced degenerative / regenerative response in the target organ.
During periods of rapid cellular proliferation additional mutations in the
DNA can occur because less time is available for premutagenic lesions to be
repaired (Maher et al., 1979). Consequently a greater number of these
premutagenic lesions will become fixed as mutations. This higher incidence
of spontaneous mutations may create a greater number of initiated cells.
With an expanded population of initiated cells the probability increases that
one of these cells will give rise to a tumor.

Alternatively, if these mutations occur in previously initiated cells then they
may be advanced to a higher state of neoplastic transformation and therefore
more prone to develop into a tumor.

Alternatively, the nongenotoxic carcinogens may be exerting their
tumorigenic action by acting as hepatic tumor promoters. This promotional
activity may be exerted in several different ways. It has been suggested that
these agents may provide a promotional stimulus by increasing hepatic
cellular proliferation for example through a degenerative/ regenerative
hyperplasia (Schulte-Hermann et al., 1982; Mirsalis et al., 1985). An increase
in the level of liver cell proliferation could result in an expansion of the
previously initiated cell population. This expansion in the number of
initiated cells increases the probability that one of these cells may progress

further towards neoplastic development and eventually give rise to a tumor.

158

Some nongenotoxic agents can stimulate cell division by mechanisms
other than by inducing cellular toxicity. Many of the classic tumor promoting
phorbol esters have have inherent mitogenic activity that can stimulate cell
proliferation. These agents have been demonstrated to be potent activators of
protein kinase C (Couturier et al., 1984). The activation of this enzyme can
have many physiological consequences including the induction of cell
division.

Some nongenotoxic agents may also exert their promotional activity by
altering the expression of gene transcription within the initiated cell. For
example, the polychlorinated biphenyl 2,3,7,8-tetrachlorodibenzodioxin
(TCDD), along with other halogenated aromatic hydrocarbons, has been
shown to be a potent liver tumor promoter in rats (Pitot et al., 1980; Jenson et
al., 1982). These materials bind to high affinity cyt0plasmic receptors which
then can interact with specific promoter regions within the DNA causing
alterations in gene transcription. The consequences of this altered gene
activity on neoplastic transformation are not well understood but changes in
gene expression are considered to be an important aspect of tumor
promotion.

Recent studies by Drinkwater et al. (in press) examined the relationship
between the promoting activity in one of these polyhalogenated biphenyl
tumor promoters and the Has gene. They compared the susceptibility to
develop tumors in ENU initiated C57BL/ 61 and C3H/ He] mice with the
promotional activity of the TCDD congener 3,4,5,3',4',5'-hexabromobiphenyl
(HBB). Their results demonstrated that the Hcs gene can act synergistically

159

and independently with this tumor promoter to enhance the induction of
liver tumors in the sensitive C3H/ He] mouse.

The increase in B6C3F1 mouse liver tumors as a result of chronic
exposure to nongenotoxic agents may be the result of an increase in the
number of initiated cells and/ or the promotion of existing spontaneously
initiated cells. Both of these actions could be the consequence of an increase
in cellular proliferation or through alterations in gene expression. These
activities when coupled to the synergistic action of the Has gene have the
apparent enhancing effect of increasing the tumor response in these animals.

One perplexing question still remains to be addressed. Why do the
tumors induced with these agents have a much lower frequency of the
activated H-ras gene compared to spontaneous tumors or tumors induced
with genotoxic carcinogens? This is a difficult question to answer. It was
originally hypothesized that this class of chemicals caused an increase in liver
tumors by simply enhancing the spontaneous process of tumor development.
As a result treatment by these agents would increase the tumor incidence, but
the distribution of tumors with or without an activated H-ras gene would be
the same as is in the control animals.

The few tumors observed within the nongenotoxic treatment groups
that contain an activated H-ras gene are approximately the number that
would be expected if those animals had received no treatment. This suggests
that the increase in tumor frequency was due to an enhancement in the
tumorigenic conversion of those cells which did not contain an activated H-

ras gene. This implies that the increase in liver tumors within the

160

nongenotoxic treatment groups was the result of a preferential stimulation of
the non H-ras containing cells by these agents.

How might this occur? To address this question we must consider the
process of initiation. It is not known whether there is more than one type of
initiating. event. The present analysis supports the concept that there may be
at least two types of initiation within the liver; activation of the H-ras gene
and another not involving the activation of the H-ras gene. It is also not
known whether these different initiation events are of equal potency. One
type of initiated cell may have a greater potential to progress towards
neoplastic transformation that another.

Because mare of the spontaneous tumors contain an activated H-ras
gene than do not, it could be argued that the activation of the H-ras gene is a
more potent initiation event (assuming all initiating events occur with equal
frequency). The cells containing an activated H-ras gene exist at a more
advanced state of neoplastic progression than do non-H-ras initiated cells.
Consequently they require fewer additional changes to arrive at a state of
complete neoplastic conversion. This explains why a higher percentage of
spontaneous tumors contain an activated H-ras gene.

Suppose that (1) the nongenotoxic agents influence tumor
development only at a certain step in the tumorigenic process and (2) only the
cells containing an activated H-ras have progressed beyond the point of
influence by the nongenotoxic agents. If this is true, then the progression of
the H-ras containing cells to a more advanced level of neoplastic

transformation is not influenced by the exposure to these agents.

161

On the other hand, the ne0plastic progression of those cells without an
activated H-ras gene is influenced by exposure to these chemicals. Perhaps
exposure to these agents alters the expression of specific genes which elevates
those cells to a level of neoplastic conversion that is comparable to the H-ras
containing cells. Once they have reached that point both cell types now have
an equal probability of further neoplastic progression. However, the exposure
to these agents has now increased the number of cells containing an activated
H-ras gene to this elevated level of neoplastic transformation. This increases
the probability that one of these cells will develop into a tumor and as a result
an increase in the number of tumors without an activated H-ras gene occurs.

These concepts are based on a number of unproven assumptions.
Additional work comparing the molecular and biochemical differences
between these different populations of cells will be required to validate or
refute this hypothesis. Insight into how these agents enhance the
development of tumors in this strain should increase our understanding of
chemical carcinogenesis in addition to aiding the interpretation of animal

bioassay data.

APPENDD(

162

APPENDIX

To determine the frequency and mutational profile of H-ras gene
activation in spontaneous and chemically induced liver tumors of the B6C3F1

mouse the following experimental protocol was developed.
1. Induce liver tumors in male B6C3F1 mice.

2. Isolate DNA from liver tumor tissue.

3. Amplify, using the polymerase chain reaction technique (PCR),
designated regions of the H-ras gene which contain the activating

codons 12, 13, 61 and 117.

4. Directly sequence the amplified H-ras gene products.

The sections below provide a more detailed description for each of the

four procedural steps listed above.

1. Tumor Induction:

Male B6C3F1 mice were obtained at the age of 5 weeks from the Charles
River Breeding Laboratories (Portage, MI.). Mice were acclimated for 1 week
prior to being placed on study. Mice in the benzidine-2H0, phenobarbital and
cthroform groups were housed 5 per cage in plastic tubes. Control mice and
ciprofibrate treated mice were housed individually in stainless steel cages.
Certified rodent chow and municipal tap water were provided ad libitum.

Thirty animals were initially placed on study for each of the chemical
treatment groups. The dosing regimen for these group is illustrated below.

 

Chemical Dose Route Duration
Benzidine-2HC1 120 ppm Drinking water 1 year
Phenobarbital 0.05 % Drinking water 1 year
Chloroform 200 mg/ kg Corn oil gavage 2X /wk., 1 year
Ciprofibrate 0.0125% Diet 2 year

2. DNA Isolation from Liver Tumor Tissue:

163

Genomic DNA was isolated from the frozen tumor tissue as described below.

1.

10.

11.

The frozen liver tumor tissue is homogenized in approximately 5 ml of a
10 mM Tris-HCl (pH 7.8), 150 mM NaCl, 2 mM EDTA solution.

The appropriate volume of a 10 % SDS solution is added to the tissue
homogenate to give a final concentration of 0.5 %.

0.2 volumes of 5.0 M NaClO4 is added, mixed and heated for 10 min. at
60° C.

The lysate is cooled to room temperature and extracted with an equal
volume of chloroform/isoamylalcohol (25:1).

The extract is centrifuge at at 6000 G for 10 min.after which the upper
aqueous layer is removed.

The DNA is precipitated by the addition of 2.5 volumes of cold ethanol.

The DNA is dried, resuspended in 21111 of homogenization buffer and
incubated with ribonuclease A (50 ug/ ml) for 1 hr. at 37° C.

The above incubation is then followed by an additional incubation with
proteinase-k (200 ug/ ml) for 1-2 hr. at 37° C.

The incubation mixture is then extracted by successive extractions with an
equal volume of phenol; 50:50 mixture of phenol : chloroform/
isoamylalcohol; and chloroform / isoamylalcohol (24:1).

Precipitation of the DNA is performed by the addition of 2.5 volumes of
cold ethanol.

The DNA is dried, resuspended in 10 mM Tris-HCl (pH 7.5), 1 mM EDTA,
and quantitated by measuring u.v. absorption at 260 nM.

164

3. Polymerase Chain Reaction Amplification of the H-ras Gene From Liver
Tumor DNA:

a. Oligonucleotide Primer Preparation.

DNA oligonucleotides used as primers for the PCR reaction and DNA
sequencing were synthesized using the phosphoramidite method on an
Applied Biosystems Model 380A DNA synthesizer. Detritylated primers were
purified by HPLC using a 10 u RP-300 reverse phase column. Elution of the
primers from the column was achieved with a linear 7-17% acetylnitrile
gradient (50 min.,1 ml/ min.) in 0.1 M tetramethylammonium acteate
(pH 7.0). A representation of a typical elution profile is provided in Figure 1.
HPLC fractions containing the purified primer were pooled, evaporated to
dryness, reconstituted with distilled water and purified by ultrafiltration
using a Centricon-3.

b. Additional Information Concerning PCR Amplification.

The PCR reaction was conducted as described in the Materials and
Methods section discussed in Chapter 2 of this thesis. Additional information
concerning the development of the PCR amplification protocol used in this
thesis to amplify genomic tumor DNA is provided below:

Initial experiments to amplify the codon 61 region of the H-ras gene
from liver tumor DNA resulted in irregular amplification or when
amplification did occur low yields were often obtained. These results were
confusing because amplification of the liver DNA was concluded under the
same conditions which were successful in amplifying a 500 bp region of the
lambda genome. After thoughtful analysis two factors were identified as
probable sources of the problem.

The first problem involved a possible change in the pH and salt
concentration of the PCR reaction mixture. Initial development of the PCR
oligonucleotide primmer purification involved the pooling and subsequent
concentration of HPLC peak fractions containing the purified primers. The
eluent used for the HPLC analysis was a 0.1 M tetramethylammonium acetate
solution. Concentration of the pooled HPLC fractions resulted in an
approximate 50 to 100 fold increase in the salt concentration of the primer
solution. Addition of the primer solution to the PCR reaction volume
would increase the salt concentration by 5 to 10 fold. Also, the addition of the
concentrated acetate solution may have changed the pH of the PCR reaction.
Both the increase in the salt concentration and change in pH may have

165

Figure 1: Representative chromatogram of HPLC purification for
DNA oligonucleotides primers used in PCR amplification and

DNA sequencing. Parameters used for the analysis are
discribed in the Materials and Methods section within Chapter 2.

166

Figure 1:

18 met

 

Absorption 254 nm

 

 

 

Distance [cm]

167

adversely inhibited the Taq polymerase. To eliminate this problem the high
salt concentration of the primmer solutions was removed, prior to use in the
PCR reaction, by ultrafiltration.

The second problem involved the accessibility of the Taq enzyme to the
H-ras gene sequence. Purification of genomic DNA at high concentrations
results in a very viscous suspension. The high viscosity of this suspension
leads to poor solubility when a small aliquot is added to to the PCR reaction
volume. Because of this, it was rationalized that the the H-ras gene may be
poorly accessible to the Taq enzyme. To alleviate this potential problem an
experiment was conducted to determine whether preincubation of the
genomic DNA with the endonuclease Bam H1 would increase the efficiency
of the PCR reaction. It was rationalized that the digestion of the genomic
DNA with the endonuclease would break down the long genomic fragments
facilitating the dispersion of the DNA in the reaction solution and thereby
increase the template accessibility to the Taq enzyme. The effectiveness of the
Barn H1 preincubation on PCR amplification effeciency is illustrated in Figure
2.

168

 

Figure 2: Agarose electrophoretic gel analysis demonstrating the utility of
Bam H1 digestion on increasing PCR amplification efficiency. PCR
amplifications were conducted as described in the Materials and
Methods section of Chapter 2.

Lane 1: Hind III digest of lambda DNA

Lane 2: 1.75 pg of normal liver DNA (undigested) was subjected to
PCR amplification of an 80 bp region within exon 2 of the
H-ras gene.

Lane 3: 5.25 ug of normal liver DNA (undigested) was subjected to
PCR amplification of an 80 bp region within exon 2 of the
H-ras gene.

Lane 4: 1.75 pg of normal liver DNA was digested with Barn H1
endonuclease prior to PCR amplification of an 80 bp region
within exon 2 of the H-ras gene.

169

 

Figure 3: Agarose electrophoretic gel analysis from a representative PCR
amplification of a 110 bp region within the H-ras gene from liver
tumor tissue. Total genomic DNA, isolated from independent liver
tumors induced with ciprofibrate, was subjected to PCR
amplification as described in the Materials and Methods within
Chapter 2. The sequence of the amplified 110 bp region is illustrated
in Figure 1 of Chapter 2. Lane 1 is a Hind III digest of lambda DNA;
Lanes 2-14 are the ciprofibrate samples.

170

(a)

117-A

 

Figure 4: Representaive DNA sequence analysis of the H-ras gene codon 12,
13 and 117 regions. PCR amplified regions of exon 1 and 3 within
the H-ras gene (illustrated in Figure 1 of Chapter 2) were sequenced
as described in the Materials and Methods section of Chapter 2.

(a) representative sequence of the H-ras gene encompassing codons
12 and 13. The DNA derived for this analysis was from a
spontaneous liver tumor. (b) representative sequence of the H-ras
gene encompassing codon 117. The DNA derived for this analysis
was also derived from a spontaneous liver tumor.

"Illilllllllllllllllll