THE-as

”L

MICHIGAN SYATEU

l I! (NH! ll/llllllllll'ill/ll!!!

3 129 942739

ll

 

 

 

 

 

 

 

 

ll

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled
Cloning and characterization of genes associated with the

neoplastic transformation of human fibroblastic cells
presented by
Jing .Qing

has been accepted towards fulfillment
of the requirements for

degree in Microbiology

EM at

Major professor

Ph.D.

nateAfilzr r117

0-12771

MS U is an Affirmative Action/Equal Opportunity Institution

 

 

LIBRARY

Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove thie checkout from your record.
To AVOID FINES Mom on or before dete due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU le An Affirmative Action/Equal Opponunlty lnetituion
Willis-9.1

 

CLONING AND CHARACTERIZATION OF GENES ASSOCIATED WITH THE
NEOPLASTIC TRANSFORMATION OF HUMAN FIBROBLASTIC CELLS

By

Jing Qing

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

DOCTOR OF PHILOSOPHY

Department of Microbiology

1 997

ABSTRACT

CLONING AND CHARACTERIZATION OF GENES ASSOCIATED WITH
THE NEOPLASTIC TRANSFORMATION OF HUMAN FIBROBLASTIC CELLS

By

Jing Qing

To understand the genetic changes involved in the malignant transformation of human
fibroblast cells, I carried out differential mRNA display analysis, comparing the infinite
life span, karyotypically stable, non-tumorigenic human fibroblast cell strain MSU-1.1
and one tumor-derived cell line, designated 6AISB1, which had been malignantly
transformed by carcinogen treatment of MSU-1.1 cells. Five of the nine differentially
expressed cDNA fragments identified by differential display were confirmed to be
markedly downregulated in 6AISB1 cells by Northern blot analysis. DNA sequencing
followed by a computer search of databases indicated that two of these were unique.
One of the cDNAs corresponds to fibulin-1D, and the other one corresponds to a novel
gene, designated 8T7. Northern and Western blotting analysis of 16 cell lines
established from tumors formed in athymic mice by MSU-1.1-derived cell strains
independently transformed in culture showed that 44% exhibited low level or lack of
expression of fibulin-1D mRNA and protein. In a similar analysis of 15 malignant cell
lines derived from patients, 80% showed low level or no expression of fibulin-1 D. To
study the role of fibulin-1D in transformation, l transfected 6AISB1 cells and a human
fibrosarcoma-derived cell line (SHAC) with a fibulin-1D cDNA expression construct.
Transfectants displaying high levels of fibulin-1D were isolated and characterized.
Elevated expression of fibulin-1D led to reduced ability to form colonies in soft agar and

reduced invasive potential as tested in a matrigel in vitro invasion assay. Furthermore,

expression of fibulin-1D resulted in a markedly extended latency in tumor formation in
athymic mice.

Using Northern analysis, I found that 10 out of 15 tumor cell lines derived from
patients and 6 out of 16 cell lines established from tumors formed in athymic mice by
MSU-1.1 cells transformed in culture have low or undetectable levels of $17 mRNA.
Molecular cloning of a near full-length cDNA revealed that the novel gene encodes a
putative transmembrane protein composed of 859 amino acids: the N-terminal domain
consisting of 492 amino acids including a 5-fold cysteine-rich repeat of 40 amino acids
homologous to the ligand binding repeat of the known low density lipoprotein receptor,
a 24 hydrophobic amino acid stretch spanning the plasma membrane, and a C-
terminal domain of 343 residues. The $17 gene is widely expressed in normal human
tissues and is particularly abundant in human heart and skeletal muscle. Western
blotting analysis using a specific anti-ST7 peptide antibody demonstrated that the
levels of ST7 protein are high in normal fibroblasts and low in 12 sarcoma-derived cell
lines tested. Altered expression of ST7 appears to be controlled at the transcriptional

and posttranscriptional level.

DEDICATION

To my daughter, Julia Yaxin Wei

for the magical, purest smile you give me everyday

To my husband, Dong Wei

without whom the research wouldn’t be so joyful

To my parents, Shuhua Qing and Hanguo Li

for their endless love and faith in me

ACKNOWLEDGMENTS

First and foremost, I would like to thank my major professor, Dr. J. Justin McCormick,
for giving me the benifit of his wisdom, insight, encouragement, and never failing help
during my years in the Carcinogenesis Laboratory. In addition, I wish I could use words
to express my sincerest and deepest gratitude to Dr. Veronica M. Maher, for her advice,
encouragement and lively discussions on science and other matters. Furthermore, I
would like to give special thanks to the other members of my graduate guidance
committee, Drs. Paul Coussens, Patrick Venta, and Richard Schwartz for their advice
and invaluable time.

Also I owe thanks to Terry McManus for his help in English proficiency in part of
Chapter I. A million thanks go to the past and present members of the Carcinogenesis
Liboratory for their generous friendship, assistance and stimulating discussions on
science. They include Scott Boley, Clarissa Dallas, Philip Doroh, April Hoffman, Shixia
Huang, Suzanne K. Kohler, W.Glenn McGregor, Lonnie Milam, Rebecca Odewaller,
Sandra O’Reilly, Joseph Prezybewski, Beatrice Tung, Qingping Wang, Dong Wei, and

Hong Zhang.

TABLE OF CONTENTS

LIST OF TABLES .................................................................................................. ix
LIST OF FIGURES ................................................................................................ x
INTRODUCTION ................................................................................................... 1
References ..................................................................................................... 5
CHAPTER I
LITERATURE REVIEW
A. Evidence that supports cancer is a genetic disease
1. Chromosomal abnormalities in cancer cells .......................................... 8
2. Familial cancers .................................................................................... 9
3. Defect in DNA repair and predisposition to cancer ................................ 10
4. Carcinogens and mutagens .................................................................. 11

B. The genetic elements governing cancer: Proto-oncogenes

1. Oncogene hypothesis ............................................................................ 12
2. Identification and isolation of oncogenes
2.1. Identification by homology with retroviral oncogenes ...................... 14
2.2. Identification by gene transfer ........................................................ 15
2.3. Identification by chromosome translocations .................................. 16
2.4. Identification by other routes
2.4.1. Identification by gene amplification ....................................... 18
2.4.2. Identification by structural homology with
other oncogenes .................................................................. 19
3. Functions of proto—oncogenes
3.1. Growth factors as oncogenes ......................................................... 20

3.2. Protein tyrosine kinases as oncogenes
3.2.1. Receptor protein tyrosine kinases: EGF receptor

as the prototype ................................................................... 22
3.2.2. Nonreceptor protein tyrosine kinases: Src

as the prototype ................................................................... 23
3.3. Membrane associated G-proteins as oncogenes ............................ 24
3.4. SerlT hr kinases as oncogenes: Raf-1 as the prototype .................. 25
3.5. Nuclear transcription factors as oncogenes .................................... 26
3.6. Cell cycle kinase activators as oncogenes ..................................... 27
3.7. Inhibitors of apoptosis as oncogenes ............................................. 29

vi

C. The genetic elements governing cancer: Tumor suppressor genes
1. Tumor suppressor gene hypothesis

1.1.The suppression of malignancy in cell hybrids ................................ 31
1.2. Familial cancer syndromes: Knudson's hypothesis ........................ 32
2. Identification and isolation of tumor suppressor genes
2.1. Identification by gene transfer ........................................................ 33
2.2. Identification by subtractive hybridization, differential hybridization
and differential mRNA display ........................................................ 35

2.3. Identification by positional cloning
2.3.1. Localization of putative tumor suppressor genes

2.3.1.1. Localization by cytogenetics .................................... 37

2.3.1.2. Localization by molecular cytogenetics:
Loss of heterozygosity ............................................ 38

2.3.1.3. Localization by molecular cytogenetics:
Linkage analysis ..................................................... 38

2.3.1.4. Localization by molecular cytogenetics:
Comparative genomic hybridization ......................... 39
2.3.2. Isolation of tumor suppressor genes .................................... 40

3. Function of tumor suppressor genes

3.1. Function of the p53 tumor suppressor gene ................................... 42
3.2. Function of the Rb tumor suppressor gene ..................................... 44
3.3. Function of the p16 tumor suppressor gene ................................... 45

D. Carcinogenesis is a multistep process: The role of oncogenes and tumor

suppressor genes
1. Epidemiological studies of cancer .......................................................... 47
2. Experimental animal studies: Mouse skin carcinogenesis ...................... 47
3. Human cancer studies: Colorectal carcinogenesis ................................. 50
4. Transformation of human cells in culture
4.1. Immortalization of human cells in culture ........................................ 52
4.2. Neoplastic transformation of immortal human cells ........................ 55
LIST OF REFERENCES ................................................................................ 56
CHAPTER II

SUPPRESSION OF ANCHORAGE-INDEPENDENT GROWTH AND
MATRIGEL INVASION AND DELAYED TUMOR FORMATION BY
ELEVATED EXPRESSION OF FIBULlN-1D IN HUMAN
FlBROSARCOMA—DERIVED CELL LINES

Abstract .......................................................................................................... 82
Introduction .................................................................................................... 83
Results
Identification of fibulin-1 D as a differentially expressed mRNA ................. 85
Expression analysis of fibuIin-1D in multiple human tumor
ceflfines ............................................................................................... 86

Establishment of SHAC and 6AISB1 clones stably expressing

vii

high levels of fibulin-1D protein ............................................................ 87
Anchorage-independence growth of the clones

that express fibulin-1 D ......................................................................... 90
In vitro invasiveness offibulin-1Dtransfectants .......................................... 94
In vivo tumorigenicity of the fibulin-1 D expressing clones .......................... 94
In vitro proliferation offibulin-1Dtransfectants ........................................... 94
Discussion ...................................................................................................... 103
Materials and methods
Cells and cell culture .................................................................................. 112
Growth rate ................................................................................................ 112
Differential mRNA display .......................................................................... 1 12
Northern blot analysis ................................................................................ 1 13
Preparation of cell lysates and conditioned medium .................................. 114
Western blot analysis of fibulin-1 ............................................................... 115
Plasmid construction and transfection ....................................................... 116
Anchorage independence assay ................................................................ 1 16
In vitro invasion assay ............................................................................... 116
Tumorigenicity assay ................................................................................. 1 16
Acknowledgment ............................................................................................ 1 18
References ..................................................................................................... 1 19
CHAPTER III

CLONING AND CHARACTERIZATION OF A NOVEL GENE ENCODING
A PUTATIVE TRANSMEMBRANE PROTEIN WITH ALTERED
EXPRESSION IN HUMAN TUMORS

Abstract .......................................................................................................... 123
Introduction .................................................................................................... 124
Results
Identification of ST7 as a gene differentially expressed
between preneoplastic and malignant human cells .............................. 126
Expression of ST7 in multiple human tumor cell lines ................................ 126
Molecular cloning of the full-length human ST7 cDNA ............................... 127
Primary structure suggests a novel transmembrane protein ...................... 130
Pattern of expression of ST7 mRNA in various human tissues ................. 136
Western blotting analysis of ST7 protein in human
tumor cell lines .................................................................................... 136
Discussion ...................................................................................................... 146
Materials and methods
Cells and cell culture .................................................................................. 149
Differential display ..................................................................................... 149
Northern blot analysis ................................................................................ 149
Cloning of human ST7 cDNA ..................................................................... 150
DNA sequencing and sequence analysis ................................................... 151
Production of anti-ST7 antibody ................................................................. 151
Western blotting analysis ........................................................................... 152
Acknowledgment ............................................................................................ 153
References ..................................................................................................... 154

viii

LIST OF TABLES

Chapter II
Table 1. Human tumor-derived cell lines tested for fibulin-1 expression ................ 91
Table 2. Growth properties of parental cells, cells expressing a

transfected-fibulin-1 D gene, and vector-transfected control cells ............ 102

LIST OF FIGURES

Figure
Chapter II
1. (A) Differential display comparing mRNAs from parental MSU-1.1
cell strain and its malignant derivatives cell line 6AISB1 ........................... 89
(B) Northern blot analysis confirming differential gene expression
for fragment SG7 using the cloned cDNA fragment as a probe ................. 89
(C) Northern blot analysis using fibulin-1 D cDNA as a probe .......................... 89
(A) Northern blot analysis of fibulin-1 expression in multiple
human tumor-derived cell lines .................................................................. 93
(B) RNA loading evaluated by a GAPDH cDNA probe ..................................... 93

. Western blot analysis of the expression (A) and secretion (B) of

fibulin-1 D protein by normal cells and 10 tumor-derived
cell lines under non-reducing conditions .......................................................... 96

. Western blot analysis of the expression and secretion

of fibulin-1 D protein by Fibulin-1 D transfectants .............................................. 99

Colony formation by fibulin-1Dtransfected cell
strains in 0.33% agarose ................................................................................ 101

In vitro invasiveness of fibulin-1 D transfected cell
strains in matrigel assays ................................................................................ 105

. Tumorigenicity offibulin-1Dtransfected cell

strains in athymic mice .................................................................................... 108
Chapter III
(A) Differential display comparing mRNAs from parental MSU-1.1
cell strain and its malignant derivatives cell line 6AISB1 ........................... 129
(B) Northern blot analysis confirming differential gene expression
for fragment ST7 using the amplified DNA fragment as a probe ................ 129

2. Northern blot analysis of ST7 expression in multiple

human tumor-derived cell lines ................................................................ 132
x

3. Characterization of $17 cDNA clones
(A) A diagram showing the four overlapping ST7

cDNA clones ............................................................................................ 134
(B) A near full-length cDNA sequence of ST7 gene and
the deduced amino acid sequence ........................................................... 135

4. Alignment of the cysteine-rich repeat region of ST7

with that of the human LDL receptor ............................................................... 138
5. Expression of ST7 gene in normal human tissues ........................................... 141
6. Analysis of the specificity of the antibody B250 ............................................... 143

7. Western blot analysis of the expression of ST7 protein in normal
human fibroblast cell lines and 12 tumor-derived cell lines .............................. 145

xi

INTRODUCTION

It is generally recognized that cancer development in humans is a multistep process,
in which a series of genetic and epigenetic events lead to the emergence of a clone of
cells that have escaped normal growth control (Peto, 1977; Father, 1984; Klein, 1987;
Weinberg, 1989; Bishop, 1991; Hunter, 1991). Altered expression of proto-oncogenes
and tumor suppressor genes as a result of mutations or other changes plays a pivotal
role in the development and progression of tumors. For example, at least four or five
genetic changes, including loss of function of tumor suppressor genes and activation of
proto-oncogenes, are necessary for the development of colon carcinomas (Fearon and
Vogelstein, 1990; Cunningham and Dunlop, 1996). In human breast cancers, mutations
in a number of oncogenes, tumor suppressor genes and many other genetic loci are
required before a cell becomes malignant (Jones et al., 1995).

Despite the vast increase in our knowledge of oncogenes and tumor suppressor genes
associated with human neoplasia over the past 15 years, the molecular events leading
to the formation of most types of human tumors are still not well understood. A variety of
genetic lesions are found in tumors, including point mutation, deletion, DNA amplification
and chromosome rearrangement (Salomon et al., 1991; Lasko et al., 1991). Genes
involved in cancer affect the normal functions of many cellular processes, including cell
proliferation, senescence, apoptosis, cell-cell and cell-matrix interactions, DNA repair,
invasion and motility, angiogenesis, and others (Sager, 1997). Yet very few cancer-

related genes affecting these processes have been identified in human cancers.

1

2
Therefore, to understand the multistep nature of tumor formation, it is necessary to

identify the genes that are involved in the transition of non-tumorigenic cells to malignant
cells and establish any correlation between the disruption their functions and the tumor
phenotype.

The recent introduction of several in vitro transformation systems utilizing human
cells in culture (Stoner et al., 1991; Reznikoff et al., 1993; McCormick and Maher,
1994; Rhim et al., 1994; Park et al., 1995) has facilitated the investigation of the
cellular and molecular mechanisms involved in the multistep carcinogenic process.
One of the advantages of these model systems is that they provide a means of
dissecting the carcinogenic process under well-defined conditions. In our laboratory,
transfection of the v-myc oncogene into a human neonatal foreskin-derived fibroblast
cell line LG1 led to the establishment of a near-diploid, karyotypically stable, infinite life
span human fibroblast cell strain MSU-1.1 (Morgan et al., 1991). MSU-1.1 cells are
phenotypically normal and do not form tumors in athymic mice. Transfection of MSU-
1.1 cells with an activated ras oncogene expressed at high levels (Hurlin et al., 1989;
Wilson et al., 1990) or treatment of MSU-1.1 cells with chemical carcinogens such as
(_+_)-7B,8a-dihydroxy-9a,10a-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (BPDE) (Yang
et al., 1992) or gamma irradiation (Reinhold et al., 1996), followed by selection of
focus-fanning cells, results in cells capable of forming tumors in athymic mice.
However, the cellular genes responsible for the neoplastic transformation induced by
these carcinogens remain poorly understood.

One way to identify critical genes associated with malignant transformation is to
analyze the gene expression profiles in normal and cancer cells with the same genetic
background. Conventionally this is achieved using subtractive hybridization (Wieland

et al., 1990; Lee et al., 1991) or differential hybridization (Steeg et al., 1988). The

3
recently developed differential mRNA display method (Liang and Pardee, 1992; Liang

et al., 1993) involves the reverse transcription of the mRNAs followed by PCR
amplification. The amplified cDNA fragments corresponding to the 3’ termini of
mRNAs are separated on a DNA sequencing gel. The advantage of this method is that
it allows one to display rapidly and simultaneously the expression of mRNAs from
various phenotypically distinct variants. Therefore, genes that are overexpressed or
downregulated can be identified at the same time.

The objectives of the present study were (1) to identify genes that are differentially
expressed between MSU-1.1 cells and one fibrosarcoma-derived cell line, designated
6AISB1, which had been malignantly transformed by carcinogen treatment of MSU-1.1
cells; (2) to determine whether the altered expression of identified genes is relevant in
human tumor development in vivo by surveying the expression of the identified genes
in tumor-derived cell lines from patients; (3) to clone the full-length cDNAs that
correspond to the differentially expressed genes; (4) to define the biological function of
the cloned gene(s) in tumor development by introducing the gene(s) into a proper
recipient cell line and testing transformation-related phenotype. By identification,
cloning and characterization of differentially expressed genes involved in the malignant
transformation of MSU-1.1 cells, we should gain insight into the nature of the genetic
changes underlying the neoplastic transformation of human fibroblasts in culture. We
hope this information will provide useful diagnosis markers. It may also provide new
approaches to therapy.

Chapter I of the thesis reviews the literature that supports the mutation theory of the
origin of human cancer. The various methods utilized to identify and isolate cellular
oncogenes and tumor suppressor genes and the biochemical functions of these genes

are discussed. Also discussed in Chapter I is the evidence that supports the multistep

4
hypothesis of carcinogenesis. Chapter II consists of a manuscript that will be

published in the journal Oncogene. It describes the research I carried out which shows
that fibulin-1 D, an extracellular matrix protein, is downregulated in 6AISB1 cells and a
large fraction of fibrosarcoma—derived cell lines. Elevated expression of fibulin-1D in
6AISB1 cells and a human fibrosarcoma-derived cell line, SHAC, led to reduced ability
to form colonies in soft agar and reduced invasive potential as tested in a matrigel in
vitro invasion assay. Furthermore, expression of fibulin-1D resulted in a markedly
extended latency in tumor formation in athymic mice. Chapter III consists of a
manuscript that has been submitted to the journal Oncogene. It describes my
research to identify and clone a novel gene, designated $17, which is downregulated
in 6AISB1 and many other fibrosarcoma-derived cell lines. The deduced amino acid
sequence suggests that ST7 encodes a putative transmembrane protein. Using a
specific anti-ST7 peptide antibody, l demonstrated that the protein level of ST7 is high

in normal human fibroblasts and low in sarcoma-derived cell lines.

LIST OF REFERENCES

LIST OF REFERENCES

Bishop, J.M. (1991). Molecular themes in oncogenesis. Cell, 64:235-248.

Cunningham, C., and Dunlop, MG. (1996). Molecular genetic basis of colorectal
cancer susceptibility. Br. J. Surg., 83:321-329.

Farber, E. (1984). Cellular biochemistry of the stepwise development of cancer with
chemicals: G. H. A. Clowes memorial lecture. Cancer Res., 44: 4217- 4223.

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

Hunter, T. (1991). Cooperations between oncogenes. Cell, 64:249-270.

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

Jones, K. A., Brown, M.A., Solomon, E. (1995). Molecular genetics of sporadic and
familial breast cancer. Cancer Surv., 25: 315-334.

Klein, G. (1987). The approaching era of the tumor suppressor genes. Science, 238:
1539-45.

Lasko, D., Cavenee, W., and Nordenskjold, M. (1991). Loss of constitutional
heterozygosity in human cancer. Ann. Rev. Genet, 25: 281-314.

Lee, S.W., Tomasetto, C., and Sager, R. (1991). Positive selection of candidate tumor
suppressor genes by subtractive hybridization. Proc. Natl. Acad. Sci. USA., 88:2825-
2829.

Liang, P., and Pardee, A. B. (1992). Differential display of eukaryotic messenger RNA
by means of the polymerase chain reaction. Science, 257: 967-969.

Liang, P., Averboukh, L., and Pardee, A. B. (1993). Distribution and cloning of
eukaryotic mRNA by means of differential display: refinements and optimization.
Nucleic Acids Res., 21:3269-3275.

McCormick, J.J., and Maher. V.M. (1994). Analysis of the multistep process of
carcinogenesis using human fibroblasts. Risk Anal., 14:257-263.

5

6

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

Park, N.H., Gujuluva, C.N., Baek, J.H., Cherrick, H.M., Shin, K.H., and Min, BM.
(1995). Combined oral carcinogenicity of HPV-16 and benzo(a)pyrene: an in vitro
multistep carcinogenesis model. Oncogene, 10: 2145-2153.

Peto, R. (1977). Epideology, multistage models, and short term mutagenesis tests. In
Origins of Human Cancer. Cold Spring Harbor Laboratory, Cold Spring Harbor,
NY,1403-1428.

Reinhold, D.S., Walicka, M., Elkassaby, M., Milam, L.D., Kohler, S.K., Dunstan, R.W.
and McCormick, J.J. (1996). Malignant transformation of human fibroblasts by ionizing
radiation. Int. J. Radiat. Biol., 69: 707-715.

Reznikoff, C.A., Kav, C., Messing, E.M., Newton, M., and Swaminathan, S. (1993). A
molecular genetic model of human bladder carcinogenesis. Semin. Cancer Biol., 4:
143-152.

Rhim, J.S., Webber, M.M., Belle, 0., Lee, M.S., Amstein, P., Chen, LS, and Jay, G.
(1994). Stepwise immortalization and transformation of adult human prostate epithelial
cells by a combination of HPV-18 and v-Ki-ras. Proc. Natl. Acad. Sci. USA., 92: 11874-
11878.

Sager, R. (1997). Expression genetics in cancer: Shifting the focus from DNA to RNA.
Proc. Natl. Acad. Sci. USA, 94: 952-955.

Salomon, E., Bozzow, J., and Goddard, AD. (1991). Chromosome aberrations and
cancer. Science, 254: 1 153-1160.

Steeg, P.S., Bevilacqua, G., Kopper, L., Thorgeirsson, U.P., Talmadge, J.E., Liotta,
LA, and Sobel, ME. (1988). Evidence for a novel gene associated with low tumor
metastatic potential. J. Natl. Cancer Inst, 80:200-204.

Stoner, G.D., Kaighn, M.E., Reddel, R.R., Resan, J.H., Bowman, 0., Naio, 2.,
Matsukura, M., You, M., Galati, A.J. and Harris, CC. (1991). Establishment and
characterization of SV40 T—antigen immortalized human esophageal epithelial cells.
Cancer Res., 51: 365-371.

Weinberg, RA (1989). Oncogenes, antioncogenes, and the molecular bases of
multistep carcinogenesis. Cancer Res., 49: 3713-3721.

VWeland, l., Bolger, G., Asouline, G., and nger, M. (1990). A method for difference
cloning: gene amplification following subtractive hybridization. Proc. Natl. Acad. Sci.
USA, 87: 2720-2724.

Wilson, D.M., Yang, D.J., Dillberger, J.E., Dietrich, S.E., Maher,V.M. and McCormick,

7

J.J. (1990). Malignant transformation of human fibroblasts by a transfected N-ras
oncogene. Cancer Res., 50: 5587-5593.

Yang, D, Louden, C, Reinhold, D.S., Kohler, S.K., Maher, V.M. and McCormick, J. J.
(1992). Malignant transformation of human fibroblast cell strain MSU-1.1 by (1)-78,8a-
dihydroxy-9a,10a-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene. Proc. Natl. Acad. Sci.
USA, 89: 2237-2241.

CHAPTER 1

LITERATURE REVIEW

A. Evidence that supports cancer is a genetic disease

It is generally recognized that cancer is fundamentally a genetic disease, arising from
changes in the DNA. Several lines of evidence support this theme: the detection of
damaged chromosomes in cancer cells, the recognition of hereditary predisposition to
cancer, the connection between cancer susceptibility and impaired ability of cells to
repair damaged DNA, the evidence that relates the mutagenic potential of substances to
their carcinogenicity, and the persuasive power of identified alterations in cancer-related
genes. Here I describe some of the evidence in favor of the mutation theory of the
origin of human cancer.
1. Chromosomal abnormalities in cancer cells

In 1914, Boveri first formulated the somatic mutation hypothesis on the origin of

human cancer (translation published in 1929). He proposed that the origin of cancer
cells was due to the “wrongly combined chromosome complex occurring in a somatic
cell and that this caused abnormal cell proliferation. The unlimited tendency to rapid
proliferation in malignant tumor cells [could result] from a permanent predominance of
the chromosomes that promote division...Another possibility is the presence of definite
chromosomes which inhibit division...Cells of tumors with unlimited growth would arise if
those inhibiting chromosomes were eliminated.” (Boveri, 1929). He proposed further that

this defect was passed on to all cellular descendants of the original cancer cell.

8

9
The first chromosomal abnormality definitively associated with human cancer is the

Philadelphia chromosome found in the leukemia cells of more than 90% of patients with
chronic myelocytic leukemia. The Philadelphia chromosome results from a reciprocal
chromosomal translocation between chromosome 9 and 22. So far, more than 100
commonly occurring translocations have been observed in Ieukemias, lymphomas and
solid tumors (Solomon et al., 1991; Rabbitts, 1994). The consistent association of
specific translocations with particular disease types, particularly when present as the
sole chromosome abnormality, has led to the realization that these rearrangements
identify critical genetic loci in the oncogenic process.

In addition to chromosomal translocation, other chromosomal abnormalities found in
cancer cells include chromosomal inversions, insertions, deletions, amplification and
aneuploidy (Ruddon, 1995). Some of these aberrations occur consistently in certain
specific cancer types. For example, deletions of 11p13 are found in Wilms’ tumors
(Riccardi et al., 1990), and chromosome 9 monosomy is found in bladder
adenocarcinoma (Solomon et al., 1991). These lead to the hypothesis that a specific
locus is involved in a key way in generating these malignancies. It has also been found
that certain chromosomal aberrations are common to tumors of different cellular origin.
For example, the 3p13-23 region is commonly affected by deletions in small cell
carcinomas and adenocarcinomas of the lung, renal cell carcinomas and ovarian
adenocarcinomas (Dal-Cin and Sandberg, 1989). Deletions of the 1p11-22 region are
found in melanoma, breast adenocarcinomas, and malignant fibrous histiocytomas
(Solomon et al., 1991). These data suggest that a common locus must be inactivated for
these tumors to develop.

2. Famllial cancers

Inherited cancers represent a small fraction, perhaps 1% to 3% of total human

10
cancers (Knudsen, 1977). About 50 forms of hereditary cancers have been reported (Li,

1988; Birch, 1994). For example, dominant genetic inheritance accounts for about 40%
of retinoblastoma and 20% to 40% of Wilms’ tumor. Familial adenomatous polyposis
(FAP) of the colon is another example of a cancer transmitted as a Mendelian-dominant
trait, with about 80% penetrance. Colorectal cancer will eventually develop in nearly
100% of untreated patients with FAP (Fearon and Johns, 1992). The Li-Fraumeni
syndrome represents another hereditary neoplasia. Members of these affected families
may develop a variety of cancers, including sarcomas, breast cancers, brain cancers,
and leukemia at an early age. The inheritance pattern is that of an autosomal dominant
trait with high penetrance (Garber et al., 1991). Collectively, these diseases suggest
that a number of inherited predisposing mutations contribute to the causation of these
cancer. The defective genes responsible for some of these diseases have been
identified and cloned. For example, the Rb gene is responsible for retinoblastomas
(Friend et al., 1986; Lee et al., 1987), the p53 gene for Li-Fraumeni syndrome
(Srivastava et al., 1990), and the APC gene for FAP (Fearon and Johns, 1992).

3. Defect in DNA repair and predisposition to cancer

One of the most convincing piece of evidence for a causal relationship between genetic
mutations and cancer in humans comes from the fact that the incidence of cancer in
patients with DNA repair deficiencies is greatly increased (Lindahl, 1994). In patients
with certain recessively inherited disorders, such as xeroderrna pigmentosum (XP),
ataxis telangiectasia, Fanconi’s anemia, Bloom’s syndrome, and hereditary
nonpolyposis colorectal cancer (HNPCC), the frequency of specific tumors is
significantly higher than in the general populations. Each of these disorders is
characterized by the inability to repair specific kinds of physical or chemical damage to

DNA.

11
XP is the most widely studied of the repair-deficient human diseases. Seven

complementation groups of XP and one group that does not complement (XP variant)
have been identified so far (Cleaver, 1990). These patients are characterized by
extreme sensitivity of the skin to sunlight and usually suffer an age-specific incidence of
skin cancer that is several thousand times higher than normal individuals. They typically
develop multiple skin tumors that lead to death from metastatic squamous or basal cell
carcinoma or malignant melanoma (Cleaver and Kraemer, 1989). All XP patients except
XP variants are defective in nucleotide excision repair. Due to this defect, the cells of
XP patients are less efficient in removing DNA lesions induced by DNA-damaging
agents, for example, lesions caused by the ultraviolet light from the sun, and therefore
are more prone to acquire genetic mutations than normal cells.

Another well studied disease is HNPCC. Recently it has been found that the majority
of HNPCC cases can be attributed to a defect in any one of the multiple loci responsible
for DNA mismatch repair, for example, the hMSH2 (Fishel et al., 1993), hMLH1
(Papadcpoulos et al., 1994; Bronner et al., 1994), hPMS1 and hPMSZ (Nicolaides et al.,
1994) genes. It has been proposed that the initial event in the development of tumors in
HNPCC patients is the functional loss of a critical mismatch repair activity (Modrich,
1994), resulting in expansion or contraction of short nucleotide repeat sequences, which
ultimately lead to mutations in cancer-related genes. For instance, the type II TGFB
receptor gene, which is involved in negatively regulating epithelial cell growth, is
mutated in many HNPCC tumors with the mismatch repair deficiency (Markowitz et al.,
1995). The mutations found were either insertion or deletions in simple repeated
sequences, which is consistent with defective mismatch repair.

4. Carcinogens and mutagens

The evidence that chemicals can induce cancer in humans has been accumulating for

12
more than two centuries (Miller, 1978). Since most chemical carcinogens or their

activated metabolites can react with DNA and cause mutations (Maher et al., 1968;
Miller and Miller, 1981; Conney, 1982), the most plausible theory of carcinogenesis is
that cancer is caused by genetic mutations. The mutagenicity of an agent can be
assayed by various methods. For example, the Ames test (Ames et al., 1973; Maron
and Ames, 1983) uses several strains of bacteria, Salmonella typhimurium, which are
histidine auxotrophs and have poor nucleotide excision repair mechanism and an
increased permeability to exogenously added chemicals. Using this system, Ames and
his colleagues have estimated that about 90% of all carcinogens tested are also
mutagens. Moreover, few non-carcinogenic agents show significant mutagenicity in this
test system (McCann et al., 1975). Consistent with this mutation theory of
carcinogenesis, chemical agents that cause genetic changes are frequently
carcinogenic (Lawley, 1989). The extent of formation of some specific DNA adducts, for
example, alkyl-OB-guanine, has been shown to quantitatively correlate with mutagenicity
and carcinogenicity of nitrosamines and similar alkylating agents (Frei et al., 1978;
Swenson et al., 1986). Polycyclic hydrocarbons, such as benzo(a)pyrene and 3-
methylcholanthrene cause mutations in cultured Chinese hamster V79 cells, as
measured by induction of resistance to 8-azaguanine and ouabain, and the degree of
mutagenicity was related to the degree of carcinogenicity of the chemicals in vivo
(Huberrnan and Sach, 1974; 1976).

B. The genetic elements governing cancers: Oncogenes

1. Oncogene hypothesis

Much of our understanding of the molecular mechanisms involved in cellular
transformation comes from studies of tumor viruses. In 1910, Payton Rous first isolated

a transforming agent from a spontaneous sarcoma formed in a Plymouth Rock chicken.

13
He demonstrated that tumor extracts still had the transforming ability after being passed

through filters designed to exclude bacteria (Rous, 1911). Very little progress occurred
on the identification of the agent until the late 1950s, when electron microscopic
techniques enabled investigators to identify the infectious agent as a virus (Bernhard,
1960). In the 1960s and 1970s, the development of techniques of molecular virology led
to the isolation and cloning of many animal tumor viruses, including those from mouse,
chicken, cat and monkey (T ooze, 1980 and Weiss et al., 1982). Molecular
characterization of these viruses revealed that they could be classified into two groups:
one group having a genome composed of RNA and the other group having a genome
composed of DNA (T ooze, 1980; Weiss et al., 1982). The RNA tumor viruses, named
retroviruses, have a small RNA genome. Genetic analysis showed that, for many of
these retroviruses, a viral gene(s) conferred transformed phenotype, i.e., increased
refractivity or rounding of cells, or both, to host cells in culture. For example, several
groups isolated temperature sensitive mutants of Rous sarcoma virus (RSV) that could
induce morphological transformation of chick embryo fibroblasts at 36°C but not at the
non-permissive temperature (Martin,1970; Friis et al., 1971; Kawai and Hanafusa,
1971). Golde (1970) and Vogt and colleagues (1970) showed that irradiation of certain
strains of avian sarcoma virus resulted in the formation of viruses that retained the
ability to grow but were no longer able to transform. Transforming genes were also
found in DNA tumor viruses such as simian virus 40 and polyoma virus (Abrahams et
al., 1975; Treisman et al., 1981; Bishop, 1985). These genes capable of inducing
transformation were termed oncogenes (Bishop and Varmus, 1983). It was
hypothesized that fragments of endogenous retrovirus genomes lay scattered
throughout the mammalian genome as residues of ancient gerrnline infections by

retroviruses. These fragmented viral genes might be able to transform the cell, once

14
activated by specific stimuli like those known to turn on latent proviruses (Huberrnan and

Todaro, 1969). Even though this hypothesis is wrong, it led to our current insights into
the molecular mechanisms of cancer development.
2. Identification and isolation of oncogenes

By differential hybridization between a transformation competent RSV strain and a
transformation defective RSV strain, Bishop and Vannus and their colleagues purified
complementary DNAs corresponding to the transforming gene of RSV (c-src) (Stehelin
et al., 1976a). Using a src cDNA as a probe, they carried out hybridization experiments
with the genomic DNA of transformed chicken cells as well as normal cells and found
that sequences homologous to this cDNA were present in both types of cells (Stehelin et
al., 1976b). Later it was found that in normal salmon, mouse, calf and human DNA,
there exist DNA sequences homologous to src (Spector et al., 1978a). RNA sequences
corresponding to src were found in the cellular RNA of normal and neoplastic chicken
cells, indicating that the gene is not only present but is also transcribed in both normal
and transformed cells (Spector et al., 1978b). Further work extended this finding to
many other retroviral oncogenes (Varmus, 1989).

Searching for the cellular homologue of viral transforming genes was a fruitful way to
uncover cellular oncogenes. Using the retroviral oncogenes as probes, more than
twenty cellular oncogenes have been identified and cloned. A comparison of the cDNA
sequence of the retroviral genes and intron-exon structure of the vertebrate genes
makes it clear that these oncogenes were normal components of the genome, and that
the retrovirus had transduced them (Bishop and Varmus, 1982). Analysis of multiple
viral oncogenes and their homologues in animal genomes by DNA sequencing revealed

that the viral genes often were mutated or altered substantially (Varmus, 1984; Bishop,

15
1985; 1987). The genes responsible for transformation in both naturally occurring and

virus-induced tumors are designated oncogene, and their normal, unaltered cellular
forms are called proto-oncogenes (Varmus, 1989; Bishop, 1991).
2mm
In 1972, Hill and Hillavo showed that the DNA of cells transformed by RSV-infection
was able to induce transformation of chicken embryo fibroblasts in culture as a result of
transfer of the RSV genome. This finding led to the hypothesis that the DNA from cells
transformed by chemical carcinogens or from cancer cell lines might also be able to
transform normal cells. Weinberg and colleagues were the first to demonstrate that
DNA from chemically transformed mouse fibroblasts were able to induce morphological
transformation (focus formation) of NIH3T3 cells, an established non-neoplastic mouse
fibroblast cell line (Shih et al., 1979). In 1981, several research groups reported that
DNA from certain human tumor cell lines can also transform NIH3T3 cells (Shih et al.,
1981; Krontiris and Cooper, 1981; Perucho et al., 1981). Cloning strategies were
designed to rescue the dominant-acting human transforming genes from the
transformed NIH3T3 cells. One method took advantage of the different repetitive
sequences found in human and mouse cells. Alu sequences are highly repeated
sequences (10 X 105 copies per genome) interspersed in the human genome and are
present on average about once every 5 kb. During DNA transfection, approximately
10-30 kb of genomic DNA from the human tumor cells is transferred and integrated into
the genome of the mouse cells. The detection of Alu sequences in the transformed
NIH3T3 cells indicates that a human sequence has been taken up by the cells and may
be contributing to the transformed phenotype. To eliminate human DNA that has
nothing to do with the transformed phenotype, genomic DNA from NIH3T3

transforrnants was again transfected into normal NIH3T3 cells. Multiple rounds of

16
transfection and isolation of transformed cells lead to the enrichment of human

sequences required for transformation of NIH3T3 cells. After enrichment, a genomic
library was constructed using the DNA from NIH3T3 transforrnants. The DNA insert of
human origin was identified by hybridization with Alu sequences and tested for
transforming ability. This strategy led to the cloning of the res oncogene from the EJ
human bladder carcinoma cell line (Shih and Weinberg, 1982). At the same time, the
res oncogene was similarly cloned from several EJ-related bladder carcinoma cell lines
(Santos et al., 1982; Der et al., 1982; Goldfarb et al., 1982).

Many modifications of the transfection focus assay have been developed. One
approach was to co-transfect tumor DNA with an antibiotic resistance marker gene. The
drug-resistant clones were then selected in vitro, pooled and injected into athymic mice
where tumor formation was monitored. This technique has revealed at least one new
oncogene, mas (Young et al., 1986). Another approach makes use of an expression
cDNA library instead of the fragmented genomic DNA. The cDNA library is constructed
with mRNA from tumor cells and introduced into NIH3T3 cells to select for focus.
Several potential oncogenes have been identified with this method (Miki et al., 1991;
Chan et al., 1993; 1994).

Such cloning experiments carried out with a large number of human tumor cell lines
have revealed a diverse assortment of oncogenes (Varmus, 1984; Bishop, 1987; Miki
and Aaronson, 1995). This method, however, suffers from several limitations. Among
the most serious is its reliance on a single mesenchymal-derived recipient cell line which
may not be susceptible to transformation by oncogenes from other cell types or to
oncogenes which confer a property already acquired by NIH-3T3 cells (Varmus, 1984;

Hunter, 1991).

23lll'fil'll | I I'

17
Specific chromosome translocations are often associated with distinct types of

neoplasms (Heim and Mitelman, 1989; Soloman et al., 1991; Rabbitts, 1994). This
observation implies that genes affected by the translocations might be instrumental in
the genesis of the associated tumors. Based on chromosome translocations, a number
of oncogenes have been identified. A good example is the cloning of the bcI-2
oncogene from B cell lymphoma cells.

A common chromosomal abnormality that occurs in at least 90% of human follicular
lymphomas is the reciprocal t (14;13) (q32 ; q21) translocation (Yunis et al., 1982). Since
it was known that the immunoglobulin heavy chain gene (lgH) resides in chromosome
band 14q32, several groups independently carried out Southern blot analysis using the
lgH gene as a probe to test whether any chromosome rearrangements occurred in the
lgH locus. Once they confirmed that this locus was mediating the translocation,
genomic libraries were constructed from the DNA of follicular lymphoma cells carrying
the t (14;18) chromosomal translocation, and screened with an lgH specific probe to
isolate DNA fragments containing the breakpoint (T sujimoto et al., 1984; Cleary and
Sklar, 1985). Subsequent chromosome walking experiments with chromosome 18 DNA
fragments flanking the breakpoint ultimately led to the isolation of the novel oncogene,
bcI-2 (T sujimoto et al., 1985; Cleary et al., 1986). This translocation resulted in the
juxtaposition of the bcI-2 gene with the lgH gene, placing the bcI-2 upstream of the
enhancer of the immunoglobulin gene and resulting in overexpression of the bcl-2 gene
product (Graninger et al., 1987).

Indeed several of the known cellular oncogenes have been found adjacent to
interchromosomal breaks. For example, the human c-myc cellular oncogene, normally
located at chromosome 8q“, is juxtaposed with the heavy, x, or k immunoglobulin (lg)

genes from chromosome segments 14q32, 2p", or 22q11 in the well characterized

18
Burkitt’s lymphoma translocations (Taub et al., 1982; Dalla-Favera et al., 1983; Leder et

al., 1983). These interchromosomal recombinations place the c-myc gene next to an lg
locus and result in enhanced expression of the c-myc gene (Erikson et al., 1983; Taub
et al., 1984). Another example is the 9:22 Philadelphia translocation present in almost
all cases of chronic myelogenous leukemia (deKlein and Hagemeijer, 1984). The
translocated gene on chromosome 9 is the c-abl oncogene (deKlein et al., 1982). The
translocation results in the deletion of the amino-terminal portion of the abl gene product
and the replacement of it with the amino-terminal of the bcr (for breakpoint cluster
region) gene (Heisterkamp et al., 1985). This fusion protein exhibits elevated tyrosine
kinase activity (Konopka et al., 1984).

MW

Many human tumors have amplified DNA sequences. The presence of amplified DNA
sequences is often, although not always, manifested by the presence of chromosomes
containing homogeneously staining regions (HSRs), or paired acentric chromatin bodies
termed double minutes (DMs). Analysis of tumor cells and cell lines that exhibit DMs or
HSRs has revealed the amplification of oncogenes. One such example is the isolation of
the mdm-2 oncogene.

A spontaneously transformed derivative of a mouse 3T3 cell line (designated 3T3-DM)
contains an average of 25-30 DMs per cell. To isolate cDNA clones that represent
genes amplified and overexpressed in the 3T3-DM cells, Cahilly-Snyder et al. (1987)
constructed a cDNA library using mRNA from these cells. By differential hybridization
with RNA from mouse cells lacking DMs or HSRs, they isolated two cDNAs
representing amplified genes that were associated with the DMs, designated mdm-1

and mdm-2 (for mouse double minute). Later they demonstrated that enhanced

19
expression of the mdm-2 gene could induce tumorigenicity in NIH3T3 and Rat 2 cells

(Fakharzadeh et al., 1991).
WWW!

Several of the oncogenes identified as described earlier belong to multigene families. It
is attractive to assume that genes homologous to known oncogenes may also be
candidates for oncogenes. For example, using probes derived from the c-myc
oncogene, three other related genes have been identified, i.e., N-myc, L-myc and R-myc
(Depinho et al., 1987). Furthermore, it has been shown that all four myc genes have
transforming activity (Schwab et al., 1985; Depinho et al., 1987). Another example is
the cloning of ras-related genes. Ha-, Ki-, and N-ras genes belong to a closely
conserved family. Using ras probes, at least five more distantly related genes have
been identified. One of these, R-ras (Lowe et al., 1987), has been reported to be
mutated in human ovarian carcinoma cell lines and the mutated gene, when introduced
into NIH3T3 cells, induced a fully malignant phenotype (Saez et al., 1994).

3. Function of oncogene products

To date, more than 100 oncogenes have been identified. Extensive studies over the
past two decades have revealed that proto-oncogene products are involved in the
regulation of normal cellular growth and differentiation. Based on their biochemical
functions, the protein products encoded by cellular oncogenes can be classified into
seven groups. I will review examples in each category.

3.1. Growth factors as oncogenes

In 1983, two groups of investigators reported that the sequence of the v-sis
transforming gene, the acutely transforming oncogene of simian sarcoma virus, was
homologous to the gene encoding the B chain of the platelet-derived growth factor

(PDGFB) (Doolitle et al., 1983; Waterfield et al., 1983). Since then, several oncogene

20
products have been found to be growth factors. For example, oncogenes hst (Delli et

al., 1987) and int-2 (Moore et al., 1986) encode fibroblast growth factor-related proteins.
In general, growth factor stimuli are transmitted into the cell via specific transmembrane
receptors that modify key regulatory proteins in the cytoplasm. These in turn affect the
decisions controlling cell proliferation and differentiation, including changes in gene
expression and reactivity to other factors. The relationship between growth factors and
malignant transformation was first suggested by Spam and Todaro (1980) who
proposed the autocrine model. This model proposes that growth factors and their
cognate receptors are co-expressed by malignant cell populations, and that ligand-
receptor interaction results in the autonomous proliferation of tumor cells. Several lines
of evidence support this hypothesis. Expression of PDGF and its receptors has been
reported in a high fraction of sarcomas as well as glioblastomas (Heldin and
Westermark, 1989; Maxwell et al., 1990). Similarly, at least 70% of small cell lung
cancer tumors and tumor-derived cell lines co-express the genes for stem cell factor and
its receptor, the c-kit proto-oncogene (Krystal et al., 1996). Human colorectal
carcinomas have been demonstrated to express a variety of growth factors, such as
epidermal growth factor (EGF), transforming growth factor a (TGFa) and their cognate
receptors, forming multi-autocrine loops (Shirai et al., 1995). These data suggest that
aberrant production of secreted growth factors can play decisive roles in tumorigenesis
by increasing the proliferation rate and degree of cellular autonomy (Aaronson, 1991;
Cross and Dexter, 1991).
3.2. Protein tyrosine kinases as oncogenes
The association of tyrosine phosphorylation with malignant transformation was first
suggested by the discovery that many of the viral oncoproteins catalyzed the

phosphorylation of proteins on tyrosine residues, and the level of phosphotyrosine in

21
proteins is elevated as much as 10—20 fold in cells transformed by these oncogenes

(Hunter, 1987). To date, more than 40 protein tyrosine kinases have been identified.
They can be classified into two groups: (1) Receptor protein tyrosine kinases, which
span the plasma membrane with large extracellular and cytoplasmic domains; (2)
cytoplasmic non-receptor protein tyrosine kinases, many of which are attached to the
plasma membranes. 1 will describe an example for each category.
:-.----o-"'uion-.n—.-u- -.-o-._ roe-to:

Because growth factor receptor tyrosine kinases have the ability to generate a
mitogenic signal, these molecules possess a latent oncogenic potential which, when
activated, results in unregulated cell growth. In fact, several oncogenes are
homologues of receptor tyrosine kinase. For example, the v-erbB oncogene is derived
from the chicken EGF receptor gene (Yarden and Ullrich, 1988). The v-fms oncogene is
derived from the gene for the receptor for the colony-stimulating factor 1 (Sherr, 1990).

The EGF receptor (EGFR) is a 170 kDa transmembrane tyrosine kinase that is
expressed on a wide variety of cell types. Other members in the EGFR family include
p185neu tyrosine kinase (also known as HER2 or c-erbB-2), erbB-3 and erbB-4
(Schlessinger and Ullrich, 1992; Plowman et al., 1993). All four receptors possess a
glycosylated, cysteine-rich, extracellular ligand binding domain, a single hydrophobic
transmembrane region, and a cytoplasmic domain that contains a tyrosine kinase
catalytic domain (Ullrich and Schlessinger, 1990).

The interaction of a ligand with the EGFR or the p185neu/HER2 results in receptor
dimerization, autophosphorylation (Ullrich and Schlessinger, 1990), and enhanced
tyrosine kinase activity toward other substrates which elicit a mitogenic response in
EGF-sensitive cells. Certain signaling molecules become physically associated and/or

phosphorylated by the activated EGFR kinase. Those identified include the

22
phosphatidylinositol (PD-specific phospholipase C-y (PLC-y) (Rotin et al., 1992; Vega et

al., 1992), the Pl-3'- kinase (PI3K), the PI4 kinase and the P15 kinase (Cochet et al.,
1991), the GTPase activating protein (GAP) (Margolis et al., 1990), and the adaptor
protein Shc (Pelicci et al., 1992). PLC4y is involved in the generation of two important
messengers, inositol triphosphate and diacyl glycerol (Berridge and Irvine, 1989;
Kikkawa et al., 1989). The former causes release of stored intracellular calcium and the
latter activates protein kinase C (PKC). These secondary messengers appear rapidly in
cells following stimulation with growth factors such as EGF. PKC belongs to a multigene
family that encodes at least ten distinct isoforms (Nishizuka, 1992). Overexpression or
gene alteration of certain isoforms of PKC has been reported to increase cell
proliferation in culture (Housey et al., 1988; Hsiao et al., 1989) and induce neoplastic
transformation (Cacace et al., 1993). The actions of a number of tumor promoters are
thought to be mediated by PKC (Berridge and Irvine, 1989; Kikkawa et al., 1989). The
importance of the PI3 kinase in transformation has been underscored by the recent
identification of a new avian sarcoma virus, ASV16. The oncogene carried by ASV 16
encodes a Gag-PI3K p110 catalytic subunit fusion protein, which has PI3K activity
(Hunter, 1997; Chang et al.,1997). The adaptor protein Shc has been shown to be a
potent transforming gene when overexpressed (Pelicci et al., 1992).

Constitutive activation of receptor protein tyrosine kinases can be achieved in a
number of ways. In the case of the v-erbB oncogene, deletion of the extracellular ligand
binding domain eliminates the negative control that this structure normally exerts on the
cytoplasmic domain. A single point mutation in the p185neu transmembrane domain
results in constitutive receptor oligomerization and activation (Weiner et al., 1989).
Receptor-derived oncogenes possess other structural lesions such as point mutations

and deletions in the cytoplasmic region, and carboxyl-terminal truncations that appear to

23
enhance and modulate the transforming signal (Woolford et al., 1988; Khazaie et al.,

1989). However, these mutations seem to play a minor role in human cancer, since the
most common cellular lesion found in human cancers involves autocrine activation (see
831) in association with receptor overexpression. Members of the EGFR family have
been shown to be amplified or overexpressed in human tumors of several types,
including lung cancer, breast cancer, ovarian cancer (Veale et al., 1987; Slamon et al.,
1989), head and neck cancers (Grandis and Tweardy, 1993), glioblastomas (Nishikawa
et al., 1994) and bladder cancer (Lipponen and Eskelinen, 1994).

keg-[eeee‘l .".-,.._- ., .‘eeee-

As noted above, Src was the first transforming oncogene discovered. At least nine
members of the Src gene family have been discovered (Bolden et al., 1992; Brown and
Cooper, 1996). Several were initially discovered as the normal counterparts of viral
oncogenes. These proteins share structural similarities. They all have an extreme N-
terrninal myristoylation signal presumably required for their attachment to the plasma
membranes, the Src homology (SH) 3 and SH2 regions, a kinase domain, and a C-
terrninal regulatory tail.

The normal cellular functions of Src are still not clear. It has been shown that Src
kinase is activated in PDGF-, CSF1-, EGF-, and FGF-stimulated fibroblasts (Brown and
Cooper, 1996), and associates with the PDGF and CSF-1 receptors through its SH2
domain (Alonso et al., 1995). Activated Src protein has been shown to phosphorylate
proteins that are implicated in mitogenic signaling pathways such as Pl-3 kinase
(Gutkind et al., 1990; Haefner et al., 1995), GAP protein (Brott et al., 1991), and the
adaptor protein, Shc (McGlade et al., 1992; Egan et al., 1993). These findings provide a
biochemical link between the Src-family kinases and pathway utilized by activated

growth factor receptors and suggest a role for Src in growth control. Therefore, v-Src

24
and activated c-Src may transform cells by upregulating mitogenic signaling pathways

normally stimulated by growth factors.

Another class of Src kinase substrates that have been identified includes proteins
involved in cytoskeletal organization (e.g., vinculin), cell-substrate adhesion (e.g.,
fibronectin receptor), and cell-cell adhesion (e.g., focal adhesion kinase) (Reviewed by
Brown and Cooper, 1996). The functional significance of phosphorylation of these
proteins has yet to be established.

3.3. Membrane associated G-proteins as oncogenes
Two forms of membrane associated G-proteins have been implicated in the regulation
of cellular proliferation (Boume et al., 1990): the heterotrimeric G proteins and the
monomeric products of the ras and related genes, including Ha-, Ki-, and N-ras.
Recently, two more ras-like genes, T021 and R-ras have become potential targets for
mutational activation in human cancers (Saez et al., 1994; Hunter, 1997).

The normal function of ras family proteins is to serve as an ubiquitous signal
transducer for multiple growth factor receptor tyrosine kinases (Crews and Erikson,
1993; Egan and Weinberg, 1993; McCormick, 1993; Marshall, 1995; Seger and Krebs,
1995). When a ligand such as EGF or PDGF binds to its receptor, the receptor is
activated, triggering an autophosphorylation event. This leads, through intermediates
such as adaptor proteins Shc and Grb2 (for growth factor receptor binding) to the
guanine nucleotide exchanging factor 80$ (for son of sevenless), resulting in recruitment
of Sos to the plasma membrane where ras is bound. The interaction between the
guanine nucleotide releasing activity of Sos and ras leads to the release of GDP and
binding of GTP to ras (Feig, 1993). This activates ras, which in turn leads to activation
of a kinase cascade including Raf, Mek (also known as mitogen-activated protein kinase

kinase or MAPKK), and MAP kinase (MAPK). The substrates of MAPK include various

25
molecules associated with cell growth control, including transcription factors, enzymes,

and regulators of protein synthesis (Marshall, 1995).

This signal transduction pathway is a crucial pathway regulating growth and
differentiation in many cell types. In fact, mutations in the ras family of oncogenes are
the most frequently detected alterations in oncogenes in both animal tumor model
systems and in human cancers (Bos, 1989). Oncogenic forms of ras genes differ
fromtheir normal counterparts by having a single mutation in codon 12,13, 59 or 61,
which reduces their intrinsic GTPase activity and their ability to interact with GAP, thus
keeping p21ras in the GTP-bound, activated mode (Tong et al., 1989; Krengel et al.,
1990).

Besides this well characterized ras-Raf-MAPK pathway, it has recently been shown
that ras activation of Raf-MAPK independent pathways is sufficient to cause tumorigenic
transformation (Khosravi-Far et al., 1996). Among the targets for ras required for
transformation is Rac, a member of the Rho family of small G proteins. Mutant Rac has
modest transforming activity, and a dominant negative mutant Rac blocks ras induced
transformation (Khosravi-Far et al., 1995; Qiu et al., 1995).

3.4. SerlT hr kinases as oncogenes: Raf-1 as the prototype

Raf-1 was first identified as the cellular homologue of the viral oncogene v-raf. The
Raf family of oncogenes encodes serine/threonine-specific protein kinases. It comprises
three members: Raf-1, A-Raf and B-Raf (Storm et al., 1990; Rapp, 1991). The three
genes are dispersed over different chromosomes and have been mapped to sites that
are frequently altered in human tumors (Storm et al., 1990). All three genes can be
converted into oncogenes by N-terminal fusion, deletion or site-specific mutations
(Rapp, 1991; Storm and Rapp, 1993). Raf proteins are believed to be key components

of the growth factor receptor-ras-MAPK mitogenic signal transduction pathway (See

26
3.3.3.). Raf proteins can also be activated by phorbol ester through protein kinase C

(Rapp, 1991; Stephens et al., 1992) or by G-protein coupled receptors (Cook et al.,
1993; Howe and Maeshall, 1993).

A number of studies also suggest that Raf-1 can elicit its oncogenic effect independent
of MAPK activation. In Rat-1 cells, expression of a transfoming, activated Raf-1 does
not stimulate MAPK activity and yet is sufficient to induce transformation (Gallego et al.,
1992; Samuels et al., 1993). The mechanisms underlying the transformation are not
clear.

Raf-1 may also positively regulate cell growth by promoting progression of the cell
cycle. It has been shown that Raf-1 can phosphorylate and activate the cdc25
phosphatases (Galaktionov et al., 1995).

3.5. Nuclear transcription factors as oncogenes

Modulation of gene expression is frequently an important ramification of intracellular
signaling and plays a vital role in the control of cell proliferation and differentiation. A
number of nuclear transcription factors and transcription regulators have been
implicated in different types of cancer (Lewin, 1991; Forrest and Curran, 1992). One
example is the activator protein 1 (AP-1).

AP-1 serves as the nuclear target of many oncogenic signal transduction pathways
(Angel and Karin, 1991; Treisman, 1994). This multigene group includes members from
the Fos (c-fos, fosB, fra-1, and fra-2) and Jun (c-jun, jun B, and jun D) families.
Members of the Fos family proteins form heterodimers with Jun proteins via their bZip
(basic region plus Ieucine zipper) regions. The complex binds to specific DNA
sequences and regulates the transcription of genes containing the cis-acting element in
their promoters.

Both Fos and Jun were initially identified as retroviral oncogenes (Curran et al., 1982;

27
Maki et al., 1987). In keeping with their role in the control of cell proliferation,

expression and activity of these genes are transiently stimulated by mitogenic signals.
Rapidly dividing cells show increased levels of AP-1 transcriptional activation (Vogt,
1994). Studies using loss-of-function approaches, i.e., the expression of antisense RNA
or the microinjection of antibodies showed that AP-1 complexes are required for
theinitiation of DNA synthesis during S phase (Holt et al., 1986; Carter et al., 1991;
Kovary and Bravo, 1991; Smith and Prochownik, 1992). This function may be mediated
by regulating the transcription of cyclin D1 (Albanese et al., 1995; Phuchareon and
Tokuhisa, 1995). Studies using transgenic mice showed that stable expression of c-fos
led to dysregulation of bone growth, eventually resulting in osteosarcomas and
chondrosarcomas (Ruther et al., 1987; 1989; Wang et al., 1991). Transgenic mice
expressing an oncogenic form of jun developed fibrosarcomas at sites of wound healing
(Schuh et al., 1990). A recent study with c-fos-null mice (Saez et al., 1995)
demonstrated that c-fos is necessary for progression from benign papillomas to
malignant squamous cell carcinomas and spindle cell tumors. Furthermore, activation
of AP-1 activity has been documented in several types of human cancer, including
human squamous cell carcinoma of the lung (Volm et al., 1992; Wodrich and Volm,
1993) and breast cancer (Walker and Cowl, 1991). Collectively, these studies
demonstrate that altered activity of AP-1 complex could affect mitogenic control and
promote neoplastic transformation of cells in specific tissues.

However, AP-1 is not connected exclusively to positive growth signals (Sassone-Corsi,
1992). In a number of systems increased AP-1 activity is correlated with cessation of
cell growth and differentiation (Vogt, 1994). The role of AP-1 as a growth stimulator or a
growth attenuator is determined by additional, largely unknown factors.

3.6. Cell cycle kinase activators as oncogenes

28
In eukaryotic cells, progression through the cell cycle is strictly controlled by a family

of cyclin proteins, the cyclin-dependent kinases (cdks), and kinase inhibitor proteins
(Hartwell and Kastan, 1994; Hunter and Pines, 1994; Nurse, 1994). Unique
combinations of cyclins and cdks assemble during each phase of the cell cycle.
Theassociation of cyclins and cdks allows the subsequent activation of the complex and
drives cell proliferation forward by phosphorylating specific substrates.

The control of mammalian cell proliferation by extracellular signals occurs largely
during the G1 phase of the cell cycle (Hunter and Pines, 1994; Hall and Peters, 1996).
During the G1 phase, cells respond to extracellular signals by either advancing toward
another division or withdrawing from the cycle into a resting state (Go) (Pardee, 1989;
Sherr, 1994). Unlike transit through the S, 62, and M phases, G1 progression normally
relies on stimulation by mitogens and can be blocked by antiproliferative agents. The D
type cyclins (D1, DZ, and 03) and cyclin E are the primary G1 phase cyclins in
mammalian cells (Sherr, 1993). Expression of the D type cyclins is rapidly induced after
exposure of cells to mitogens; this expression declines when mitogens are withdrawn or
antiproliferative agents are added (Sherr, 1994). Thus one might expect that the
deregulation of cyclin D synthesis would make cell cycle progression less dependent on
growth factors and contribute to oncogenesis.

Recent studies have shown that, as a group, mutations in or deregulated expression
of cell cycle genes constitute the most common genetic changes in tumor cells (Sherr,
1996), especially in one or another of the genes involved in controlling progression
through the G1 phase of the cell cycle. The best characterized one is cyclin D1.

Cyclin D1 is encoded by the CCND1 gene on chromosome 11q13 (Xiong et al., 1992).
It was originally cloned as an oncogene, termed PRAD1, which was found to be

activated by inversion of chromosome 11 in the genome of parathyroid adenoma cells

29
(Motokura et al., 1991; Rosenberg et al., 1991a). This inversion places CCND1/PRAD1

gene under the control of the parathyroid hormone gene promoter, resulting in
overexpression of cyclin D1. In centrocytic lymphoma and multiplemyeloma,
overexpression of cyclin D1 is achieved by a reciprocal translocation of chromosome
11q13 and the lgH locus on 14q34, bringing cyclin D1 gene under the influence of the
lgH enhancer (Seto et al., 1992; Rosenberg et al., 1991b). Amplification and
overexpression of cyclin D1 have also been documented in a broad spectrum of
common human cancers, including breast cancers, squamous cell carcinomas of the
head and neck, bladder cancers, small cell lung cancers and esophageal carcinomas
(Hall and Peters, 1996). Consistent with the oncogenic role of cyclin D1 are
observations that transgenic mice engineered to overexpress this cyclin in their breast
tissues are prone to mammary adenocarcinomas (Wang at al., 1994), and that co-
expression of cyclin D1 and myc genes in the lymphoid tissues of transgenic mice leads
to rapid development of lymphomas (Bodrug et al., 1994; Lovec et al., 1994).
3.7. Inhibitors of apoptosis as oncogenes
In the past few years, accumulating evidence suggests that an imbalance of

homeostasis between cell growth and death plays a role in cancer development. Cancer
cells usually acquire the ability to escape from programmed cell death (apoptosis) in
response to DNA damage or other suboptimal conditions that would induce a normal cell
to die. Bcl-2 family proteins are key regulators of apoptosis and are the best candidates
implicated in carcinogenesis.

As discussed earlier (see B.2.3), the bcI-2 gene was first identified as an oncogene
involved in the development of human follicular lymphoma (T sujimoto et al., 1984). The
t(14;18) translocation brings the bcl-2 gene under the control of the lgH enhancer,

resulting in constitutively high expression of bcl-2 protein (Graninger et al., 1987).

30
The frequent association of a translocation involving the bcl-2 locus with cancer

suggests that the bcl-2 gene may be an oncogene. Introduction of bcl-2 into
aninterleukin 3-(IL-3)-dependent cell line led, not to factor-independent growth, but to
factor-independent survival, and the cells were not tumorigenic (Vaux et al., 1988).
Reed et al. (1988) reported that 3T3 cell constitutively expressing bcl-2 proteins only
rarely became transformed. These studies were interpreted as indications that bcl-2
can contribute to tumorigenesis by allowing cells to survive, but malignancy requires
additional genetic changes.

Consistent with this view, the generation of bcl-2 transgenic mice using the lgH
enhancer demonstrated that in the B cell lineage, deregulated cell survival resulted in
dramatic polyclonal expansion of mature B lymphocytes and the development of diffuse
immunoblastic B cell lymphomas. A high proportion of these tumors had
rearrangements of the c-myc gene as a second genetic alteration (Strasser et al., 1990;
McDonnell and Korsmeyer, 1991; Katsumata et al., 1992). Targeted overexpression of
the bcl-2 gene in thymic lymphocytes, mammary and intestinal epithelium, and myeloid
populations has not produced tumors (Hockenbery, 1994).

One model for the function of the bcI-2 gene in carcinogenesis proposes that the bcI-2
gene acts synergistically with specific oncogenes to inhibit cell death under adverse
growth conditions. This concept follows the recognition that several oncogenes have
dual effects on cell proliferation and cell death. For example, expression of the c-myc
gene promotes cell proliferation when growth conditions are favorable. However, when
cells enter a growth arrest phase due to absence of growth factors or the action of
cytostatic drugs, expression of c-myc induces apoptosis (Packham and Cleveland,
1995). C-myc-induced apoptosis has been shown to be blocked by overexpression of

bcl-2 (Bissonnette et al., 1992; Evan et al., 1992; Fanidi et al., 1992). Additional cellular

31
oncogenes that can predispose cells to undergo apoptosis have been found, including c-

fos and c-rel (Smeyne et al., 1993; Abbadie et al., 1993).

Deregulation of the bcl-2 gene may also function as a survival mechanism in cancer
cells predisposed to cell death by environmental factors. BcI-2 acts as a broad anti-
apoptotic factor and opposes cell death following treatment with ionizing radiation and
cancer drugs as well as hormone manipulations (Sentman et al., 1991; Miyashita and
Reed, 1993). Several non-lymphoid neoplasms, including lung, prostate, colon and
breast cancers, express high levels of bcI-2 protein (McDonnell et al., 1992; Pezzella et
al., 1993; Bronner et al., 1994).

C. The genetic elements governing cancers: Tumor suppressor genes
1. Tumor suppressor gene hypothesis

The existence of tumor suppressor genes that can suppress or inhibit the growth of
tumor cells is supported by three lines of evidence: (1) studies on somatic cell hybrids
between normal and tumor cells; (2) familial cancer syndromes; and (3) the loss of
specific chromosomes in cancers. 1 will focus on studies on somatic cell hybrids and
familial cancer syndromes.

1.1. The suppression of malignancy in cell hybrids

Somatic cell hybrid experiments provided the first evidence that genetic changes
underlying the neoplastic transformation might result from the loss of function of specific
genes. In these experiments, tumorigenic cells were fused with nontumorigenic cells,
and the resulted hybrid cells were nontumorigenic (Harris et al., 1969; Harris, 1988).
However, hybrids passaged in culture for extended periods were sometimes
tumorigenic. Careful studies demonstrated that the tumorigenic segregants had lost one
or more chromosomes from the initial hybrids (Harris and Klein, 1969; Stanbridge et al.,

1981; Evans et al., 1982), and that suppression of malignancy in the hybrid cells

32
required the retention of specific chromosomes derived from the normal cells

(Stanbridge et al., 1981; Klinger, 1982; Harris, 1988). This phenomenon of suppression
holds true for a wide variety of cell or tissue types (Sager, 1985; 1989; Levine, 1993).
Taken together, these experiments suggest that recessive genetic changes can be
responsible for the tumorigenic phenotype and that gene products of normal alleles from
normal cells are able to suppress tumorigenicity.

By analyzing chromosomal loss in the tumorigenic variants of suppressed hybrids and
using a microcell transfer technique, in which only one single chromosome derived from
the normal cells is transferred to tumorigenic cells, experiments have identified and
mapped the specific chromosomal location of some of these suppressor genes
(Marshall, 1991; Levine, 1993).

1.2. Familial cancer syndromes: Knudson’s hypothesis

A small proportion of most types of human tumors occur in familial form (Knudson,
1977; Li, 1988). Several well-defined tumor syndromes are inherited as simple
autosomal dominant traits with high penetrance (Birch, 1994). Retinoblastoma, a rare
pediatric eye tumor, is the prototype of this group.

From the epidemiological studies of retinoblastoma, Knudson (1971) noted that 40% of
retinoblastoma cases occurred in young children with a mean age of 14 months, and
these tumors were often bilateral (originating in both eyes) with the mean number of
tumors being three. In some cases there was a family history of this disease. If these
patients were cured through surgical intervention, they often had a substantially
increased risk of developing other malignancies, especially osteosarcoma. These
observations suggested an inherited component for these cancers. Howevr, about 60%
of the retinoblastoma cases didn’t fit this pattern. In these cases, no familyhistory of

cancers was noted. These tumors were typically first detected at a later age, and were

33
always unilateral (affecting only one eye) with one cancer per patient. This class of

retinoblastomas was quite rare, occurring in about one in 30,000 people.

Following the earlier suggestions by De Mars (1970), Knudson (1971) proposed a
hypothesis to explain the origins of these two categories of retinoblastoma. He
suggested that children who develop retinoblastomas at an early age inherited a mutant
allele of a gene, later called the retinoblastoma susceptibility gene or Rb, and that a
second somatic mutation in this gene in a retinoblast would then give rise to this cancer.
This was a common event, which was consistent with an average of three independent
cancers per patient and occurred in 90% of such patients. In contrast, children who
develop retinoblastoma at a later age did not inherit a mutant Rb allele, and two
independent mutations of this gene in a single retinoblast must have occurred to give
rise to a tumor since the Rb gene is recessive at the cellular level. This was consistent
with the low frequency (one in 30,000 people). Predisposition to retinoblastoma was
autosomally dominant because at typical mutation rates almost everyone who has
inherited one defective Rb gene will sustain a second mutation in the Rb gene in at least
one retinoblast. This unifying hypothesis leads to the concept of genes whose product
normally prevents cancerous growth and both alleles must be lost via mutations for it to
play a causal role in cancer development.

2. Identification and isolation of tumor suppressor genes
2.1. Identification by gene transfer

Unlike the dominant oncogenes, which can be identified with relative ease, assaying
directly for tumor suppressor gene function is difficult. It is easy to identify a small
number of transformed cells admist a background of normal cells because
thetransformed cells have a growth advantage. In contrast, identifying a small number

of normal (revertant) cells which grow like normal cells in a background of transformed

34
cells is not a trivial task. Based on observations that alteration in certain phenotypes in

vitro is associated with loss of malignancy in vivo, several groups devised assays to
clone tumor suppressor genes directly using a gene transfer technique (Noda, 1990).

In the course of characterizing the in vitro phenotypes of spontaneously
nontumorigenic revertants from the v-Ki-ras-transformed NIH3T3 cells, Noda and
colleagues found that flat revertants exhibited greatly reduced malignancy in vivo (Noda
et al., 1983; Basin and Noda, 1987). They used “flatness” as an in vitro marker to clone
tumor suppressor genes. They transfected the tumorigenic DT cell line (a subline of v-
Ki-ras-transformed NIH3T3 cell) with a normal human foreskin fibroblast expression
library, isolated flat revertants, and recovered the transfected cDNA, Krev-1, from the
genome of the revertant cells. When Krev-1 was introduced into DT cells and
expressed at high levels, some of the transformed phenotypes, e.g., the efficiency of
colony formation and the size of the colonies in soft agar, were reduced. The Krev-1
gene encodes a protein homologous to the p21ras protein (Kitayama et al., 1989).

Since H-ras transformed rat FE-8 cells showed an increased sensitivity toward ouabain
when compared to their normal counterparts (Noda et al., 1983), Schaefer et al. (1988)
established a functional assay to identify and clone DNA sequence capable of
suppressing neoplastic transformation. Genomic DNA from normal human placenta was
introduced into FE-8 cells by co-transfection with a plasmid conferring drug resistance.
Drug-resistant cells were subjected to treatment with ouabain. The surviving clones lost
the morphology of transformed cells, acquired the ability to grow in an anchorage-
dependent manner only, and showed reduced ability to form tumors in athymic
mice.Using the human Alu repetitive sequences as probes, the suppressor gene in a
secondary revertant clone was isolated.

These cloning experiments are largely based on alterations in certain phenotypes in

35
vitro that correspond well to loss of malignancy in vivo. The problem is lack of an

efficient selection procedure. Also, the spontaneous revertants will give rise to false
positives. Because of these limitations, gene transfer is not an efficient method to clone
tumor suppressor genes.
2.2. Identification by subtractive hybridization, differential hybridization and
differential mRNA display

Since malignant transformation is caused by accumulation of diverse genetic changes,
including mutations and/or altered expression of critical genes (Sager, 1997), gene
transfer experiments may not detect genes that are required but are not sufficient to
suppress the malignant phenotype. Therefore, identification and cloning of genes
differentially expressed between tumor cells and well matched normal counterparts,
especially those genes that are downregulated in the tumor cells, may uncover new
tumor suppressor genes. Currently, several methods are available for this purpose,
namely, subtractive hybridization, differential colony hybridization and differential mRNA
display.

Subtractive hybridization is designed to select for genes expressed uniquely or
preferentially in one of a pair of closely related cell populations (Sargent, 1987). cDNA
synthesized from the mRNA of one cell population is hybridized with excess amount of
mRNA from the other cell population. The cDNA-RNA hybrids representing mRNAs that
are equally expressed in both cell populations are efficiently separated from the
unpaired cDNAs by hydroxyapatite chromatography. The unpaired cDNAs from the first
round of hybridization are hybridized again with excess amount of mRNA from the other
cell population. Usually the unpaired cDNAs from the second round of hybridization
represent mRNAs that are uniquely expressed in the cell of interest. These cDNAs are

inserted to a vector to construct a subtractive library or are used as probes to screen a

36
library. Using such a strategy, Sager and colleagues isolated several putative tumor

suppressor genes involved in the development of breast cancer (Lee et al., 1991 ).

The technique of differential colony hybridization has been widely used to identify
mRNAs regulated by cell differentiation, growth and cytokine treatment (Leonard et al.,
1987; Sarrna et al., 1992; Femandez-Pol et al., 1993; Green et al., 1995). To identify
genes uniquely expressed in metastatic cells, Steeg et al. (1988) utilized a series of
related murine melanoma cell lines of varying metastatic potential. These cell lines are
derived from K-1735 cell line. A cDNA library was constructed using mRNA from K-
1735 cells. Duplicate filters of the library were prepared and one was hybridized to
labeled cDNA from a low-metastatic cell line and one to cDNA from a high-metastatic
cell line. Using this approach, they isolated a novel gene, termed nm23, whose
expression level was inversely correlated to tumor metastatic potential (Sobel, 1990).
The major drawbacks of this techniques are poor sensitivity and difficulty in isolating the
relatively few positive clones from the high background hybridization (Cochran et al.,
1987).

The differential mRNA display procedure developed by Liang and Pardee compares
the profile of gene expression in closely related eukaryotic cells by systematically
amplifying the 3' end of mRNAs using reverse transcription and PCR ( Liang and
Pardee, 1992; Liang et al., 1993). It consists of two steps. First, a subset of the mRNA
populations are reversely transcribed using an anchored oligo(dT) primer; secondly,
the3' end of mRNAs are amplified by PCR using an arbitrary 10mer primer and an
anchored oligo(dT) primer. The PCR products are separated on a DNA sequencing gel.
By comparing the band patterns from two cell populations, mRNAs that are upregulated
or downregulated can be identified simultaneously. Compared with the conventional

subtractive or differential hybridization method, differential mRNA display provides a

37
rapid and efficient approach to detect differentially expressed mRNAs from pairs of

phenotypically distinct cells of the same genetic background. Using this approach,
Sager and colleagues have isolated more than 50 novel genes which are no longer
expressed or are strongly downregulated in breast carcinoma cell lines (Sager, 1997).
Genes that are downregulated in other types of cancer have also been identified with
this method (Sun et al., 1994; Simon et al., 1996).
2.3. Identification by positional cloning

A powerful way to identify and clone tumor suppressor genes is positional cloning,
whereby a gene associated with certain types of cancer is isolated on the basis of its
approximate chromosomal position (Collins, 1991; Wrcking and lMlliamson, 1991). The
first step in this approach is the mapping of the gene to a particular human
chromosome. Then various strategies can be used to identify and characterize the
specific gene in that genomic region. I will describe the various methods available for

localization and cloning of such genes.

2311 I' I' [III
23111 III | l'

Cytogenetic studies in solid tumors have generally yielded a plethora of inconclusive
data but there have been some consistent findings that have led to the isolation of new
tumor suppressor genes. For example, about 3% to 5% of hereditary
retinoblastomacases were found to be associated with gross abnormalities in
chromosome 13 (Gallie et al., 1990). Other studies found patients who had small
interstitial deletions of 13q (Franke, 1978; Yunis and Ransay, 1978). In all cases
studied, the deletion involved band 13q14.1. This regional deletion was also
demonstrated in some sporadic retinoblastomas (Balaban et al., 1982). These studies

suggested that the putative Rb tumor suppressor gene resided within 13q14.1.

38
Cytogenetic studies also helped to localize the tumor suppressor gene of Wilms’ tumor

on chromosome 11p13 (Riccard et al., 1990). However, most of the time, the regional
chromosome loss is too small to be detected at the visible chromosome level.
Molecular cytogenetics provides a more sensitive method to detect chromosome
abnormality by analysis of extracted DNA.

Loss of heterozygosity (LOH) involves comparing normal and tumor tissues from the
same patient to detect allele imbalance at polymorphic loci. The typical situation is that
the two alleles from the non-tumorigenic cells of the body contain a restriction fragment
length polymorphism (RFLP), while the cells from the tumor tissue have lost one of
these two alleles. By using panels of RFLPs that map to different regions of a given
chromosome, and studying LOH in enough tumors, a common region of deletion can be
identified. LOH could arise from a loss of a chromosome, a small deletion of the genetic
locus, or a gene conversion by homologous recombination. Typically a mutation arises
in a tumor suppressor gene and this is followed by LOH at that locus, thus eliminating
both wild-type alleles.

Besides RFLPs, microsatellite sequences are particularly informative for detection of
LOH. They are extremely polymorphic so that most individuals are constitutionally
heterozygous (Weber, 1990). Furthermore, they can be detected by PCR and are not
dependent on the presence or absence of a restriction enzyme site.

For those genes associated with hereditary predisposition to some form of cancer,
genetic linkage analysis in affected families provides an additional route to localization.
This technique depends on the crossover between homologous chromosomes at

meiosis which means that DNA sequences, even on the same chromosome arm, will

39
not be transmitted together over several generations unless they are physically very

close together. If homologous sequences on the two parental chromosomes are
polymorphic, meiotic reassortment can be detected. The purpose of linkage analysis is
to identify the close physical neighbors of the gene of interest in the genome (Yates and
Connor, 1986). This involves finding a large number of families in which multiple
individuals are affected with a specific cancer, and the testing of the individuals in these
families with a large panel of DNA markers that show polymorphism. When a marker
that tends to be co-inherited with the cancer gene in all (or almost all) affected members
of those families is found, it indicates that the cancer gene is located near to that marker
and thus allows mapping of the gene to that chromosome. Linkage analysis then
proceeds by testing additional markers from the same chromosome arm and measuring
the frequency of meiotic crossover between each of them and the cancer trait. In this
way, the relative positions of a number of DNA sequences (including the putative cancer
gene) can be established to facilitate gene isolation. The major limitation of linkage
analysis in identifying tumor suppressor sequences is that it can only be used to study
familial cancers and requires relatively large, well-defined pedigrees.
Il" _—_ 1"‘I‘. 'll'z-‘Q‘l'll I 'lQ-‘l

In comparative genomic hybridization (CGH), a single hybridization allows DNA copy
number changes in the entire genome of a tumor to be assessed in comparison with
normal tissue DNA (Kallioniemi et al., 1992). The principle of CGH is simple.
Representative genomic DNA is prepared separately from tumor and normal tissue of
the same patient. The DNAs are labeled with two different fluorochromes, e.g., one with
a green fluorochrome, and the other with a red. Equal amounts of these labeled DNAs
are mixed and simultaneously hybridized to normal human metaphase spreads. The

relative amounts of tumor and normal DNA probes bound at a given chromosomal locus

40
depend on the relative abundance of these sequences in the two DNA samples and can

be quantitated by measurement of the ratio of green to red fluorescence along the
length of each chromosome. The normal DNA probes serve as a control for local
variations in the ability to hybridize to the target chromosome. Therefore, where there
has been either a deletion or an amplification of DNA in the tumor, the green-to-red
ration will deviate from the norm. This method has led to the localization of several
susceptibility loci for human cancers (Hemminki et al., 1997; Kallioniomi, 1997).
MW

Once the approximate location of a putative tumor suppressor gene is mapped within
two flanking markers, several methods can be used to clone the gene.

The first method is the candidate gene method. If there are one or several known
genes in the region of interest, these then become “candidates” for the gene causing the
cancer in question and are examined, for any mutations that segregate with the disease
in affected families and for a putative function that may plausibly account for the
observed phenotype. This procedure led to the identification of the p53 tumor
suppressor gene.

By using multiple probes on the short arm of chromosome 17, Baber et al. (1989)
defined a region that was commonly lost in colon carcinomas. This region contains the
site of the p53 gene. The p53 protein was implicated in tumorigenesis because of its
interaction with the transforming large T antigen of simian virus 40. Since no alteration in
the remaining p53 locus could be detected in tumors, the p53 cDNA from tumors that
had lost one p53 allele was sequenced. This analysis with two tumor samples found the
presence of single point mutations occurring in evolutionarily conserved regions (Finlay
et al., 1988). Additional cDNA sequencing showed that many different human tumors

commonly contained mutant p53 proteins (Hollstein et al., 1994; Lane, 1994).

41
The second method is chromosome walking. This technique was used by Dryja et al.

(1986) to clone the Rb tumor suppressor gene. They showed that a chromosome 13
sequence that mapped to q14 detected a region that was homozygously deleted in the
tumor DNA from two unrelated Rb patients. This sequence was then used in a
chromosome walking to identify a DNA fragment conserved in mouse and human, which
suggested that it contained a coding sequence. This DNA fragment identified a 4.7 kb
mRNA, which was expressed in normal retinal cells, but was altered or not expressed in
retinoblastomas (Friend et al., 1986; Lee et al., 1987)

Another approach involves a direct search for transcribable sequences. Currently two
methods are frequently used: hybridization selection and exon trapping (Duyk et al.,
1990; Nehls et al., 1994; Datson et al., 1996). With hybridization selection, one first
screens a genomic library with the marker sequences believed to lie close to the gene of
interest. Then one screens a cDNA library from an appropriate tissue with the labeled
genomic clone to isolate the cDNA. With exon trapping, genomic DNA of interest is
partially digested with a restriction enzyme and cloned into a vector containing a splice
donor and an acceptor sites positioned flanking the inserted genomic DNA. RNA
transcripts derived from such vectors are processed in vivo and exons contained within
the inserted genomic fragments become flanked by known sequences in the resulting
mRNAs. Reverse transcription-coupled PCR can then be used for subsequent cloning
and sequence analysis of trapped exons. Identification of cDNA by either method
results in a candidate tumor suppressor gene. Confirmation that it is the tumor
suppressor gene usually requires that mutations that segregate with cancer patients can
be identified in that gene or that its biological function can account for the phenotype.

3. Function of tumor suppressor genes

So far, about a dozen confirmed tumor suppressor genes have been cloned (Sager,

42
1997). Their biological functions range from transcriptional regulation to cell cycle check

point control. I will describe the functions of three tumor suppressor genes that have
been extensively studied, namely p53, Rb and p16.
3.1 . Function of the p53 tumor suppressor gene

More than 50% of human tumors contain mutations in the p53 gene (Hollstein et al.,
1994). The p53 tumor suppressor gene encodes a 393 amino acids nuclear
phosphoprotein. The protein has three structural and functional domains. The N-
terrninal region of p53 protein is highly charged and acts as a transcription activation
domain. It interacts with the basal transcription machinery to regulate gene expression
(Lu and Levine, 1995). p53 transcription activation is negatively regulated by MDM2
protein which can act as an oncoprotein (Lin et al., 1995). The core or central region of
p53 is highly conserved in evolution and has sequence-specific DNA binding activity.
More than 90% of p53 missense mutations reside in this region. The C-tenninal domain
of p53 protein is responsible for its oligomerization. p53 proteins assemble through this
domain to form stable tetramers. The last 28 amino acids of p53 act as a site of
negative regulation for the sequence-specific DNA binding function of the central domain
(Hupp and Lane, 1994) and may also encode a non-specific DNA binding site (Levine,
1997)

Normally the amount of p53 protein in a cell at steady state is low because of its
relative short half life (about 20 minutes). Different types of DNA damage can activate
p53 protein, including double-strand breaks induced by y-irradiation, ultraviolet
irradiation, and chemical damage to DNA. This results in a rapid increase in the level of
p53 protein in cells and activation of p53 as a transcription factor (Cox and Lane, 1995).
One of the downstream genes of p53 is the p21 protein, an universal inhibitor of cyclin-

dependent kinases (El-Deiry et al., 1993). The p21 protein binds to a number of cyclin-

43
cdk complexes, inhibiting cdk kinase activity and blocking cell cycle progression (Xiong

et al., 1993). The p21 protein also binds to proliferating cell nuclear antigen (PCNA),
blocking its role as a DNA polymerase processitivity factor in DNA replication (Waga et
al., 1994). Therefore, p21 can act on cyclin-cdk complexes and PCNA to stop DNA
replication, and arrest cells in the G1 phase. This p53-mediated G1 arrest is believed to
give cells sufficient time to repair DNA damages before the cells enter into the S phase.
The p53 protein has also been shown to mediate GZIM phase arrest. When mitotic
spindle inhibitors, such as nocodazole, are added to cells with wild type p53, the cells
are arrested in 62. In the absence of p53, these cells will reinitiate DNA synthesis,
increasing the ploidy of the cells (Cross et al., 1995). In addition, p53 appears to be
involved in regulating the number of centrosomes in a cell. Mouse embryo fibroblasts
(MEFs) from p53 null mice produce abnormal numbers of centrosomes after a few
doublings in cell culture and initiate spindles with three or four poles, whereas MEFs
from normal mice in culture at the same passage level do not exhibit this phenotype
(Fukasawa et al., 1996). This p53-mediated GZIM check point may account for the
phenotype of genomic instability that is commonly associated with a p53 mutation.
Another important function of the p53 protein is to regulate apoptosis. Normal
thymocytes will undergo apoptosis in response to DNA damage, whereas thymocytes
from p53 null mice do not (Lowe et al., 1993). p53 can also initiate apoptosis in
response to the expression of a viral or cellular oncogene. The expression of the
adenovirus E1A protein in rat fibroblasts stabilizes and activates p53 protein. The
resultant cells die of apoptosis (Debbas and White, 1993). Cells overexpressing E2F-1
and a temperature sensitive mutant p53 protein undergo apoptosis at 32°C but not at
37-39°C, where p53 is inactive (Wu and Levine, 1994). Similarly, cells overexpressing

myc and a temperature-sensitive p53 have a temperature-sensitive apoptotic response

44
(Wagner et al., 1994).

Besides these factors, hypoxia is able to stimulate p53 levels and induce apoptosis
(Graeber et al., 1996). It has been suggested that this process represents another way
that p53 may act as a tumor suppressor (Kinzler and Vogelstein, 1996). Most tumors
are initiated by genes other than p53 and go through many rounds of clonal expansion
to reach a critical size when the blood supply becomes rate-limiting. Angiogenesis is
required for clinically apparent growth of a tumor. The resultant hypoxia induces a p53-
mediated apoptosis to limit the progression of tumors.

In summary, the current knowledge of p53 function suggests that the p53 protein acts
to protect the organism from genetic damage. In response to damage or potential
damage to the DNA, p53 protein initiates a protective cell cycle arrest or apoptotic cell
death. A number of factors affect the decision of a cell to enter a p53-mediated cell
cycle arrest or apoptotic pathway. Under conditions in which the DNA is damaged,
growth factors are deprived, or an activated oncogene is forcing the cell into
proliferation, p53-mediated apoptosis prevails. In this way, cells with damaged DNA or
cells in a stressed environment are eliminated in a p53-dependent apoptotic event.
Loss of p53 gene function allows the propagation of cells with genetic damage and this
may be a key step in the development of neoplasia (Lane, 1994; Levine, 1997).

3.2. Function of the Rb tumor suppressor gene

Rb, the product of the retinoblastoma tumor suppressor gene, exerts its cell growth
inhibitory function mainly by inhibiting the activity of transcription factors E2Fs. E2Fs
are a family of heterodimeric transcription factors that can transactivate genes whose
products are important for S phase entry, including most notably c-myc, B-myb, cdc2,
thymidine kinase and dihydrofolate reductase (Nevins, 1992; Lathangue, 1994).

Rb exerts most of its effect in a defined window of time in the first two thirds of the G1

45
phase of the cell cycle. Cells entering G1 from mitosis require exposure to serum

mitogens continuously until several hours before the onset of S phase; thereafter they
become relatively serum-independent. This transition from a serum-dependent to
serum-independent state is demarcated by a discrete point in time, termed R
(restriction) point (Pardee, 1989). Before a cell reaches the R point, the Rb protein is
found in an underphosphorylated form. The hypophosphorylated Rb binds to a subset of
E2F complexes, converting them to repressors that constrain expression of E2F target
genes (Weinberg, 1995; Sherr, 1997). During the last several hours of G1, the Rb
protein is initially phosphorylated by cyclin D-dependent kinases and then by cyclin E-
CDK2 complex. Phosphorylated Rb protein dissociates from E2Fs, enabling them to
transactivate the genes important for S phase entry.

A diverse body of evidence indicates that the Rb protein is involved in the
pathogenesis of a variety of human tumors. In retinoblastomas, in small cell lung
carcinomas, and in many sarcomas and bladder carcinomas, Rb function is lost through
mutations of the Rb gene (Horowitz et al., 1990). In the great majority of cervical
carcinomas, the cells have been previously infected with one of the oncogenic forms of
the papilloma virus and the inactivation of Rb is achieved by the binding of the E7 viral
protein (zur Hausen, 1991). As discussed earlier, many tumor cells constitutively
overexpress cyclin D1 (see B36), and this in turn activates cyclin D-dependent kinases
and phosphorylates Rb protein. Taken together, as a direct consequence of Rb
inactivation (either by genetic mutation or by functional inactivation), E2Fs are liberated
from Rb inhibition and the progression of cells into late G1 and S phase become
uncontrolled.

3.3. Function of the p16 tumor suppressor gene

The cell cycle is controlled by the sequential activation and inactivation of cyclin-

46
dependent kinases. As discussed earlier, among the members of the cyclin family, D1

can be an oncogene and functions as an effector of mitogen—induced proliferation acting
during the G1 phase of the cell cycle. The D1-dependent kinases are subject to
additional levels of regulation, including the association with inhibitory subunits (Morgan,
1995; Sherr and Roberts, 1995). One such inhibitor, p16, specifically binds and inhibits
the cyclin D-dependent kinases by competing with D cyclins. Since cyclin D-dependent
kinase activity is required to phosphorylate the Rb protein during the middle to late G1
phase as described above, p16 can negatively regulate cell proliferation by suppressing
hyperphosphorylation and functional inactivation of the Rb protein.

The involvement of p16 in the development of human cancers was implied by the
observation that the p16 gene is mutated in many tumor-derived cell lines and maps to
chromosome 9p21, a region frequently altered in human malignancies (Kamb et al.,
1994; Nobori et al., 1994; Okamoto et al., 1994). Several mechanisms of p16
inactivation have been characterized. Point mutation and small deletions are common
in pancreatic adenocarcinomas, esophageal carcinomas, and in families with hereditary
susceptibility to melanoma (Hirama and Koeffler, 1995; Pollock et al., 1996).
Homozygous deletions of the p16 locus occur commonly in non-small cell lung
carcinomas, head and neck tumors, prostate tumors and bladder carcinomas (Hirama
and Koeffler, 1995). Finally, methylation of the p16 gene locus is common in breast and
colon cancers and results in silencing of the p16 promoter (Herman et al., 1995).
Consistent with its tumor suppressor function, p16 null mice spontaneously develop a
spectrum of different tumors by 6 months of age, with the rate of tumor formation
accelerating in response to carcinogen treatment (Serrano et al., 1996). Cultured p16”'
embryo fibroblasts do not enter senescence, and unlike their wild type counterparts,

they can be transformed by oncogenic ras alone (Serrano et al., 1996). Presumably,

47
p16 loss might mimic the overexpression of cyclin D1, leading to the

hyperphosphorylation and inactivation of Rb protein.

D. Carcinogenesis is a multistep process: The role of oncogene and tumor
suppressor gene
1. Epidemiological studies of cancer

Cancer is predominantly a disease of the elderly, with the risk of acquiring this disease
increasing with age. The first clue to the multistep nature of carcinogenesis comes from
epidemiological studies. Earlier epidemiological studies of cancer incidence as a
function of age using mathematical modeling have shown that for adult human tumors
four to six genetic changes were required for genesis of a tumor (Arrnitage and Doll,
1954; Peto, 1977; Dix, 1989). Recently Renan (1993) used the same method but a
better defined data set to address the question of the number of mutational changes
required for 28 different human malignancies. By plotting the log of the age-specific
mortality rate against the log of the age in years of the person affected, and determining
the best-fit linear regression coefficients for each tumor type, he concluded that the
common adult human cancers of the lip, stomach, liver, pancreas, kidney, skin and
bladder required 7 to 8 mutational changes, and tumors with very late onset, such as
prostate cancer, required 12 changes. Though this type of studies provides us a general
idea about the multistep nature, it can not tell what genetic changes are involved.
2. Experimental animal studies: Mouse skin carcinogenesis

A classical model of multistep carcinogenesis is the pathogenesis of mouse skin cancer
(Mottram, 1944; Boutwell, 1964; 1974; Hennings et al., 1990; Yuspa, 1994). Sequential
application of chemical agents to mouse skin can induce tumors, and tumor
development can be divided into three stages: initiation, promotion and progression.

Typically, tumor initiation is brought about by the single application of a mutagen, such

48
7,12-dimethylbenz(a)anthracene (DMBA); promotion is carried out by repeated

application of a phorbol ester, such as 12-O-tetradecanoylphorboI-13-acetate (T PA) or
by a natural promoting stimulus such as wounding. Papillomas begin to appear at 12 to
20 weeks after the promotion has begun and by about 1 year, about 40% to 60% of the
animals have some papillomas that become squamous cell carcinomas. If the
promoting agents are given alone or before the initiating agent, usually no malignant
tumors occur.

Detailed studies of the molecular basis of each stage of the skin tumor development
indicate that initiation involves a permanent, heritable changes in the gene expression of
the initiated cells which produces a subtle change in the keratinocyte phenotype.
Genetic analyses revealed that c-Ha-ras alterations are associated with the initiated
phenotype. It has been shown that c-Ha-ras mutations are usually heterozygous in
papillomas and can be detected in initiated skin prior to the emergence of tumors
(Nelson et al., 1992). Furthermore, the initiating agent used determines the existence,
nature, and site of the c-Ha-ras mutation (Quintane et al., 1986; Brown et al., 1990).
For example, when DMBA is metabolized, the diol epoxide produced primarily binds to
adenine residues in DNA (Cheng et al., 1988). When H-ras mutations were analyzed in
skin tumors of DMBA-treated mice, the mutations were A181 to T transversions in 45 of
50 tumors (Brown et al., 1990).

When a subpopulation of keratinocytes isolated from carcinogen-initiated skin are
cultured in vitro, they resist the Ca2+ signal for terminal differentiation and evolve as foci
which continue to grow in medium containing >0.1 mM Ca”. Biochemical analyses
indicate that overexpression of the TGFa protein and an alteration of PKC are essential
for mediating this phenotype. Differential modification of PKC isoforms, particularly

activation of PKCa through increased levels of cellular diacylglycerol and functional

49
inhibition of PKC8 by tyrosine kinase, produces keratinocytes with enhanced

proliferative capacity and reduced sensitivity to signals for terminal differentiation
(Yuspa, 1994).

Application of tumor promoters to initiated epidermis causes the selective clonal
outgrth of initiated cells to produce multiple benign squamous cell papillomas, each
representing an expanded clone of initiated cells (Deamant et al., 1987; lannaccone et
al., 1987). The mechanisms of exogenous promotions are likely to be epigenetic in most
cases because (1) a single genetic change in normal keratinocytes is sufficient to
produce a papilloma phenotype (Greenhalgh et al., 1993) and (2) most promoting
agents are not mutagens (Yuspa and Dlugosz, 1991).

Progression of a papilloma to a carcinoma in mouse skin is usually a spontaneous
process. Progression and malignant conversion can be enhanced and accelerated by
exposing animals beefing papillomas to a mutagen, supporting a genetic basis for
progression (O’Connell et al., 1986; Hennings et al., 1990). Genetic studies indicate that
non-random, sequential chromosomal aberrations are associated with progression of
mouse skin papillomas; particularly prominent are trisomies of chromosomes 6 and 7
(Yuspa, 1994). Changes in two cellular genes, c-Ha-ras and p53, have been closely
identified with malignant conversion of skin tumors. The mutated c-Ha-ras gene, which
is heterozygous in papillomas, is frequently homozygous in carcinomas (Bianchi et al.,
1990). Mutations in the p53 tumor suppressor gene are rarely found in chemically
induced papillomas but are frequently detected in squamous carcinomas, particularly
those induced by benzo(a)pyrene (Ruggeri et al., 1991; 1993; Kress et al., 1992).
Recent studies indicate that 90% of squamous carcinomas are devoid of TGFB1 and
TGFBZ. Direct evidence linking TGF81 loss and accelerated malignant tumor

progression comes from studies using keratinocytes cultured from TGFB1 null mice.

50
Introduction of v-Ha-ras oncogene into cultured keratinocytes of TGFB1 null mice or

wild-type and heterozygous Iittermates results in papillomas when these cells are
grafted to nude mice. However, carcinomas only develop in the TGFB1 null papillomas
but not in those of other genotypes (Yuspa, 1994). These results indicate that the TGFB
family of growth inhibitors can serve as suppressor of malignant progression.

The mouse skin carcinogenesis model clearly demonstrates that the genesis of skin
tumors requires non-genetic changes as well as a series of genetic changes. Mutations
in oncogenes, e.g., c-Ha-ras, tumor suppressor genes, e.g., p53, altered expression of
TGFj31, altered activity of PKC and characteristic chromosomal abnormalities all
contribute to the development of the squamous cell carcinomas.

3. Human cancer studies: Colorectal carcinogenesis

Colorectal carcinomas have proven to be an excellent model system in which to study
the genetic changes involved in the initiation and progression of human solid tumors.
There is a well-defined progression from benign to malignant tumors, and the cancers
that arise are clonal in origin. Tumors of various stages of dysplasia and malignancy,
ranging from benign adenomas to invasive carcinomas, can be obtained surgically. In
addition, colorectal cancer exists in both sporadic and inherited form. For example, in
patients with familial adenoma polyposis (FAP), hundreds to thousands of adenomatous
polyps will develop in their colons and rectums, and a small percentage of these go on
to become malignant. The high frequency of polyps at various stages of progression
allows one to study the stage-wise pattern of colon cancer. Studies of this inherited
cancer syndrome and large number of sporadic tumors allowed Vogelstein and
colleagues to identify the critical genetic events driving colorectal carcinogenesis
(Vogelstein et al., 1988; Fearon and Vogelstein, 1990).

An early change that occurs in the genome of cells of small, early adenomas from FAP

51
patients and patients without familial predisposition is chromosomal allelic loss of 5q.

Linkage analysis and molecular cloning have identified the affected gene on
chromosome 5q21 in FAP and the gene is called APC ( for adenomatous polyposis coli)
(Groden et al., 1991; Kinzler et al., 1991). The APC gene is mutated in the gennline of
FAP patients (Mandi et al., 1994) while truncating APC mutations have been identified in
60% of sporadic colorectal cancers and adenomas (Powell et al., 1992; lchii et al.,
1993). Both copies of APC are inactivated either by mutations on each allele, or by a
mutation on one allele and a structural deletion in the other. The data indicate that APC
is a tumor suppressor gene, and one or both alleles of the APC gene are lost
orinactivated at an early stage in the development of colorectal cancer.

Mutations in the K-ras oncogene have been identified in about 10% of small
adenomas, 50% of larger adenomas and in 50% of carcinomas. Such ras mutations,
usually in codons 12 or 13, tend to be observed in more dysplastic adenomas. Ras gene
mutations appear to occur in one cell of a preexisting adenoma and confers it a growth
advantage. Through clonal expansion of the cells with the mutation, a small adenoma is
converted into a larger and more dysplastic one.

Loss of heterozygosity of additional tumor suppressor genes appears to be critical in
later stages of colorectal tumorigenesis. The chromosomes most frequently deleted
include chromosome 18q and 17p. 18q is deleted in 50% of late adenomas and more
than 70% of carcinomas (Vogelstein et al., 1988). A candidate tumor suppressor gene
from this region has been identified. This gene, termed DCC (for deleted in colorectal
cancer), encodes a protein with putative cell adhesion properties. Its expression is
absent or reduced in colorectal carcinomas, suggesting that its involvement in normal
cell-cell or cell-matrix interactions is required to maintain a normal state of

differentiation. The loss of a large portion of one copy of chromosome 17p has been

52
observed in more than 75% of carcinomas. This loss is rarely seen in adenomas at any

stage. The common region lost on 17p contains the p53 tumor suppressor gene.
Nucleotide sequencing analysis of the p53 cDNA derived from colorectal cancers has
shown that in 70% to 80% of the cases there is a missense mutation in the remaining
p53 allele in the cancer cells. Additional chromosome losses, including 1q, 4p, 6q, 8p,
9g and 22q, has been observed in colorectal cancers. On average, colorectal
carcinomas contain four or five allelic losses. Patients with more than the median
number of losses in their tumors have a poorer prognosis.

In summary, colorectal tumors appear to arise as a result of the mutational activation
of specific oncogenes and inactivation of specific tumor suppressor genes. Mutations in
at least four to five genes are required for the formation of a malignant tumor. Fewer
changes are sufficient for benign tumorigenesis. These studies clearly demonstrate the
complex multistep nature of human tumor development.

4. Transformation of human cells in culture

One of the advantages in using human cells in culture to study neoplastic
transformation is that it provides a means to dissect the carcinogenesis process under
well-defined conditions. However, normal human cells in culture have never been found
to undergo spontaneous malignant transformation and have proven to be extremely
difficult to be malignantly transformed by any means (McCormick and Maher, 1988;
Rhim, 1993). Recently neoplastic transformation of human cells in culture has been
achieved in a stepwise fashion-immortalization and malignant conversion of the
immortalized cells. These studies support a multistep process for neoplastic
transformation and provide insights into the molecular mechanisms underlying the
process.

4.1. Immortalization of human cells in culture

53
Normal human cells in culture have a limited life span, beyond which the cells enter

the terminally nondividing state referred to as senescence (Hayflick and Moorhead,
1961). Finite life span normal human cells can only undergo two successive clonal
selections before they enter crisis and senesce (McCormick and Maher, 1988). Earlier
studies trying to transform finite life-span human cells in culture were not successful,
suggesting that acquiring immortality is a prerequisite if a cell is to acquire sequentially
all the genetic changes needed to become malignant (Kuroki and Huh, 1993;
McCormick and Maher, 1994). Whether this is the case for cells in the human body is
not known for certain. What is known is that many malignant tumor-derived cells can be
grown indefinitely in culture, whereas cells from normal tissues are never able to grow
indefinitely in culture.

Human cells can be immortalized by repeated treatment with chemical and physical
carcinogens, and by infection or transfection with certain viral genes. However it occurs
at a very low frequency (McCormick and Maher, 1988; Shay et al., 1991; Bai et al.,
1993; Kuroki and Huh, 1993). In order to immortalize human fibroblasts, for example,
Namba et al. (1988) had to expose them to more than 10 treatments with either
4-nitroquinoline 1-oxide or gamma-irradiation and has been successful only a few times.
Nevertheless, immortalization of human epithelial cells, for example, epidermal
keratinocytes, bronchial epithelial cells, mammary epithelial cells, and prostate epithelial
cells, have been achieved by infection with the AD12-SV40 hybrid virus or human
papilloma virus (Kuroki and Huh, 1993; Rhim, 1993).

Immortalization of diploid human foreskin fibroblast has been achieved in this laboratory
by transfection with a v-myc oncogene. An early passage, foreskin-derived normal
human fibroblast cell line was transfected with a plasmid carrying the neo gene and a v-

myc gene. The transfectants were selected for geneticin resistance and clonally—derived

54
cell strains expressing the v-myc protein were isolated and propagated for many

generations. Eventually all cell strains senesced, but among the senescing progeny
viable cells could be seen. These eventually gave rise to an infinite life span cell strain
designated MSU-1.1 (Morgan et al., 1991). This experiment demonstrates that
expression of the v-myc oncoprotein alone is not sufficient to cause human fibroblasts to
acquire an infinite life span, but we do not yet know what additional genetic change(s) is
required.

At present the mechanisms underlying immortalization are poorly understood. Careful
analyses of the immortalization process suggest that immortalization is caused by
multiple genetic changes. For example, following infection by SV40 virus or stable
transfection with the SV40 T-antigen, normal proliferating human fibroblasts replicate for
about 20-30 more population doublings beyond their normal senescent point. The
population then enters a state of crisis, during which cell number remains constant or
declines due to an increase in cell death. From this crisis population arise the rare,
immortal clones, presumably due to additional genetic events (Wright and Shay, 1992).
It has been reported that (Shay and Wright, 1989) for a population of SV40 T-antigen
transfected human lung fibroblasts, the frequency of immortalization is about 3 X 10‘7.
Somatic cell hybridization between an immortal cell line and a finite life span cell line
results in cells with limited life span, indicating that immortalization is caused by
recessive gene mutations (Smith and Pereira-Smith, 1996). Recently, a number of
studies indicate that the p53 and Rb tumor suppressor genes may play a causal role in
immortalization of certain cell types (Shay et al., 1991; Vojta and Barrett, 1995). For
example, fibroblasts from seven out of eight Li-Fraumeni syndrome patients escaped
senescence and were immortalized spontaneously in culture (Bischoff et al., 1990).

Hara et al. (1991) reported that targeted functional knock-out of the Rb and p53 with

55
antisense oligomers resulted in an extended life span of human fibroblasts.

4.2. Neoplastic transformation of Immortal human cells

While immortality is not sufficient for transformation, most immortalized cells have an
increased sensitivity for spontaneous, carcinogen- or oncogene-induced neoplastic
progression. For example, ADIZ-SV40 hybrid virus immortalized human epidermal
keratinocytes can be malignantly transformed by retroviral oncogenes such as H-ras,
fms, erbB and src; they can also be transformed by chemical carcinogens or by x-ray
irradiation treatment (Reviewed by Rhim, 1993). This laboratory has successfully
converted MSU-1.1 human fibroblasts into malignant cells by transfection of an
activated H-ras or an N-ras oncogene in a vector engineered for overexpression of the
oncogene. Expression of the same ras oncogenes at the level found for the endogenous
H-ras or N-ras proto-oncogene did not cause malignant transformation. However, if
MSU-1.1 cells that expressed a transfected H-ras or v-sis oncogene at low levels were
subsequently transformed with a v-fes oncogene, then the cells became malignant (Lin
et al., 1995). These studies indicate that more than one genetic changes are needed for
the malignant transformation of MSU-1.1 cells. They also demonstrate the
complementary role between oncogenes.

In summary, normal human cells in culture can be immortalized by a variety of means
(viruses, chemical carcinogens, irradiation and oncogene transfection). Additional
exposure of these immortal cells to a carcinogenic agent and appropriate selection can
result in malignantly transformed cells. Thus these studies demonstrate that similar to
tumor development in vivo, neoplastic transformation of human cells in culture is indeed

a multistep process.

LIST OF REFERENCES

LIST OF REFERENCES

Aaronson, SA. (1991). Growth factors and cancer. Science, 254:1146-1153.

Abbadie, C., Kabrun, N., Bouali, F et al. (1993). High levels of c-rel expression are
associated with programmed cell death in the developing avian embryo and in bone
marrow cells in vitro. Cell, 75:899-912.

Abrahams, P.J., Mulder, C., van der voode A, et al.. (1975) Transformation of primary
rat kidney cells by fragments of simian virus 40. J. Virol., 16:818-823.

Albanese, C., Johnson, J., Watanabe, G., Eklund, N., Vu, 0., Arnold, A., and Pestell,
R. G. (1995). Transforming p21ras mutants and c-Ets-2 activate the cyclin D1 promoter
through distinguishable regions. J. Biol. Chem, 270: 23589-23597.

Alonso, G., Koegl, M., Mazurenko, N., and Courtneidge, SA. (1995). Sequence
requirements for binding of Src family tyrosine kinases to activated growth factor
receptors. J. Biol. Chem., 279:9840-9848.

Ames, B.N., Durston, W.E., Yamasaki, E., and Lee, FD. (1973). Carcinogens are
mutagens: simple test system combining liver bacteria for detection. Proc. Natl. Acad.
Sci. USA., 70:2281-2285.

Angel, P., and Karin, M. (1991). The role of Jun, Fos and AP-1 complex in cell-
proliferation and transformation. Biochim. Biophys. Acta, 1072:129-157.

Annitage, P., and Doll, R. (1954). The age distribution of cancer and a multistage
theory of carcinogenesis. Brit. J. Cancer Res., 8:1-12.

Bai, L., Mihara, K., Kondo, Y., Honma, M., and Numba, M. (1993). Immortalization of
normal human fibroblasts by treatment with 4-nitroquinoline 1-oxide. Int. J. Cancer,
53:451-456. -

Balaban, G., Gilbert, F., Nichols, W., et al.., (1982). Abnormalities of chromosome 13 in
retinoblastomas from individuals with normal constitutional karyotypes. Cancer Genet.
Cytogenet, 6:213-221.

Bassin, RH, and Noda, M. (1987). Oncogene inhibition by cellular genes. Adv. Viral.

Oncol. 6, 103-127.

56

57

Bernhard, W. (1960). The detection and study of tumor viruses with the electron
microscope. Cancer Res., 20:712-727.

Berridge, M.J., and Irvine, RF. (1989). lnositol phosphates and cell signaling. Nature,
341:197-205.

Bianchi, A.B., Aldaz, CM, and Conti, C.J. (1990). Nonrandom duplication of the
chromosome bearing a mutated Ha-ras-1 allele in mouse skin tumors. Proc. Natl. Acad.
Sci. USA., 87:6902-6906.

Birch, J.M. (1994). Familiar cancer syndromes and clusters. Brit. Med. Bullet, 50:624-
639.

Bischoff, F.Z., Yim, S.O., Pathak, S., Grant, 6., Siciliano, M. J., Giovanella, B. C.,
Strong, L. C., Tainsky, M. A. (1990). Spontaneous abnormalities in normal fibroblasts

from patients with Li-Fraumeni cancer syndrome: aneuploidy and immortalization.
Cancer Res. 50: 7979-7984.

Bishop, J.M. (1991) Molecular themes in oncogenesis. Cell, 64:235-248.

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

Bishop, J.M. (1985). Wral oncogenes. Cell, 42:23-38.

Bishop, J.M., and Varmus, H. Functions and origins of retroviral transforming genes. In
Weiss, W., Teich, N., Varmus, H., and Coffin, J. (Eds): RNA tumor viruses. Cold Spring
Harbor, NY, Cold Spring Harbor Laboratory Press, 1982, pp.999-1108.

Bissonnette, R.P., Echeveri, F., Mahboubi, A., and Green, D. (1992). Apoptotic cell
death induced by c-myc is inhibited by bcl-2. Nature, 359:552-554.

Bodrug, S.E., Warner, B.J., Bath, M.L., Lindeman, G.J., Harris, A.W., and Adams, J.M.
(1994). Cyclin D1 transgene impedes lymphocyte maturation and collaborates in
lymphomagenesis with the myc gene. EMBO J., 13:2124-2130.

Bolden, J.B., Rowley, R.B., Spana, C., and Tsygankov, A.Y. (1992). The Src family of
tyrosine protein kinases in hemopoietic signal transduction. FASEB J. 6:3403-3409.

Bos, J.L. (1989). ras oncogene in human cancer: A review. Cancer Res. 49:4682-
4689.

Boume, H.R., Sanders, DA, and McCormick, F. (1990). The GTPase superfamily: l. A
conserved switch for diverse cell functions. Nature, 348:125-131.

Boutwell, R.K. (1974). The function and mechanism of promoters of carcinogenesis.
CRC Crit. Rev. Toxicol., 2(4): 419-43.

Boutwell, R.K. (1964). Some biological aspects of skin carcinogenesis. Prog. Exp.
Tumor Res., 4: 207-250.

S8

Boveri, T. (1929). The origin of malignant tumors. (Baltimore: The Vlfllliams and Wilkins,
00.), pp.26—27.

Bronner, M.P., Culin, C., Reed, J., and Furth, E. (1995). BcI-2 proto-oncogene and the
gastrointestinal mucosal epithelial tumor progression model. Am. J. Pathol., 146:20-26.

Bronner, C.E., Baker, S.M., Morrison, P.T., Warren, G., Smith, L.G., Lescoe, M.K.,
Kane, M., Earabino, C., Lipford, J., Lindblom, A, et al.. (1994). Mutation in the DNA
mismatch repair gene homologue hMLH1 is associated with hereditary non-polyposis
colon cancer. Nature, 368: 258-261.

Brott, B.K., Decher, S., Shafer, J. et al., (1991). GTPase-activating protein (GAP)
interacts with viral and cellular src kinase. Proc. Natl. Acad. Sci. USA, 88:755-759.

Brown, MT, and Cooper, J.A. (1996). Regulation, substrates and regulation of src.
Biochim. Biophys. Acta, 1287:121-149.

Brown, K., Buchmann, A, and Balmain, A. (1990). Carcinogen-induced mutations in
the mouse c-Ha-ras gene provide evidence of multiple pathways for tumor progression.
Proc. Natl. Acad. Sci. USA, 87: 538-542.

Cabilly-Snyder, L., Yang-Feng, T., Francke, U., and George, D.L. (1987). Molecular
analysis and chromosomal mapping of amplified gene isolated from a transformed
mouse 3T3 cell lines. Somat. Cell Mol. Genet, 13:235-244.

Cacace, A.M., Guadagno-Nichols, S., Krauss, R., et al., (1993). The epsilon isoform of
PKC is an oncogene when overexpressed in rat fibroblasts. Oncogene, 8:2095-2104.

Carter, R., Cosenza, S.C., Pena, A, Lipson, K., Soprano, DR, and Soprano, K. J.
(1991). A potential role for c-jun in cell cycle progression through late G1 and S.
Oncogene, 6: 229-235.

Chan, A.M., Chedid, M., McGovern, E.S., Popescu, N.C., Miki, T., and Aaronson, SA
(1993). Expression cDNA cloning of a serine kinase transforming gene. Oncogene, 8:
1329-33.

Chan, A.M., Miki, T., Meyers, K.A., and Aaronson, SA. (1994). A human oncogene of
the RAS superfamily unmasked by expression cDNA cloning. Proc. Natl. Acad. Sci.
USA, 91: 7558-62.

Chang, H.W., Aoki, M., Fruman, D., Auger, K.R., Bellacosa, A, Tsichlis, P.N., Cantley,
L.C., Roberts, TM, and Vogt, PK. (1997). Transformation of chicken cells by the gene
encoding the catalytic subunit of PI 3-kinase. Science, 276:1848-1850.

Cheng, S.C., Prakash, A. S., Pigott, M. A, Hilton, B. D., Lee, H., Harvey, R. G., and
Dipple, A. (1988). A metabolite of the carcinogen 7,12-dimethylbenz[a]anthracene that
reacts predominantly with adenine residues in DNA. Carcinogenesis, 9: 1721-1723.

Cleary, ML. and Sklar, J. (1985). Nucleotide sequence of a t(14;18) chromosome

59

breakpoint in follicular lymphoma and demonstration of a breakpoint-cluster region near
a transcriptionally active locus on chromosome 18. Proc. Natl. Acad. Sci. USA,
82:7439-7443.

Cleary, M.L., Smith, SD, and Sklar, J. (1986). Cloning and structural analysis of cDNA
for bcI-2 and a hybrid bcI-2/immunoglobulin transcript resulting from the t(14;18)
translocation. Cell, 47:19-28.

Cleaver, J.E., and Kraemer, K.H. (1989). In The Metabolic Basis of Inherited Disease.
6th Edition, Scriver,C.A, Beaudet, A.L., Sly, W.S., and Vale, D. Eds. New York,
McGraw-Hill, pp. 2949-2971.

Cleaver, J.E. (1990). Do we know the cause of xeroderma pigmentosum?
Carcinogenesis, 1 1 :875-882.

Cochet, C.C., Filhol, O., Payrastre, B., Hunter, T., and Gill, GM. (1991). Interaction
between the epidermal growth factor receptor and phosphoinositide kinases. J. Biol.
Chem., 266:637-644.

Cochran, B.H., Zumstein, P., Zullo, J., Rollins, B., Mercola, M., and Stiles, CD. (1987).
Differential colony hybridization: molecular cloning from a zero data base. Methods
Enzymol., 147: 64-85.

Collins, PS. (1991). Of needles and haystacks: Finding human disease genes by
positional cloning. Clin. Res., 39:615-623.

Conney, AH. (1982). Induction of microsomal enzymes by foreign chemicals and
carcinogenesis by polycyclic aromatic hydrocarbons. Cancer Res., 4224875.

Cook, S.J., Rubinfeld, B., Albert, 1., and McCormick, F. (1993). RapV12 antagonizes
ras-dependent activation of ERK1 and ERK2 by LPA and EGF in Rat-1 fibroblasts.
EMBO J.,12: 3475-3485.

Cox, LS, and Lane, DP. (1995). Tumour suppressors, kinases and clamps: how p53
regulates the cell cycle in response to DNA damage. Bioessays, 17: 501-508.

Crews, CM, and Erikson, R.L. (1993). Extracellular signals and reversible protein
phosphorylation: What to Mek of it all. Cell, 74:215-217.

Cross, M., and Dexter, TM. (1991). Growth factors in development, transformation, and
tumorigenesis. Cell, 64:271-280.

Cross, S.M., Sanchez, C.A., Morgan, C.A., Schimke, M.K., Ramel, S., ldzerda, R.L.,
Rasking, W.H., and Reid, B.J. (1995). A p53-dependent mouse spindle check point.
Science, 267: 1 353-1 356.

Curran, T., Peters, G., Van, B.C., Teich, NM, and Verma, 1M. (1982). FBJ murine
osteosarcoma virus: identification and molecular cloning of biologically active proviral
DNA. J. \firol.,44:674-682.

60

Dal-Cin, P., and Sandberg, AA (1989). Chromosomal aspects of human oncogenesis.
Crit. Rev. Oncog., 1: 113-26.

Dalla-Favera, R., Martinotti, S., Gallo, R., Erikson, J., and Croce.C.M. (1983).
Translocation and rearrangement of the c-myc oncogene locus in human
undifferentiated B-cell lymphomas. Science, 219:963-967.

Datson, N.A, van-de Vosse, E., Dauwerse, H.G., Bout, M., van-Ommen, G.J., and
den-Dunnen, J.T. (1996). Scanning for genes in large genomic regions: cosmid-based
exon trapping of multiple exons in a single product. Nucleic Acids Res., 24: 1105-1111.

De Mars, R. 23rd Annual Symposium, Fundamental Cancer Research 1969. Baltimore,
Williams and Wilkins 1970; pp. 105-106.

De Klein, A, and Hagemeijier, A. (1984). Cytogenetics and molecular analysis of the Ph’
translocation in chronic myeloid leukemia. Cancer Surv., 3:515.

Deamant, F.D., and lannaccone, PM. (1987). Clonal origin of chemically induced
papillomas: separate analysis of epidermal and dermal components. J. Cell Biol.,
88:305-312.

Debbas, M., and White, E. (1993). Wild-type p53 mediates apoptosis by E1A which is
inhibited by E1B. Genes Dev., 7:546-554.

DeKlein, A, van Kessel, A.G., Grosveld, G., Bartram, C.R., Hagemeijer, A, Bootsma,
D., Spurr, N.K., Heisterkamp, N., Grotten, J., and Stephenson, JR (1982). A cellular
oncogene is translocated to the Philadelphia chromosome in chronic myelocytic
leukemia. Nature, 300:765-767.

DeIIi-Bovi, P., Curatola, AM., Kern, F.G., Greco, A, lttmann, M., and Basilico, C.
(1987). An oncogene isolated by transfection of Kaposi's sarcoma DNA encodes a
growth factor that is a member of the FGF family. Cell, 50: 729-737.

DePinho, R., Mitsock, L., Hatton, K., Ferrier, P., Zimmerman, K., Legouy, E., Tesfaye,
A, Collum, R., Yancopoulos, G., Nisen, P., et al. (1987). Myc family of cellular
oncogenes. J. Cell Biochem., 33: 257-66.

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

Dix, D. (1989). The role of aging in cancer incidence: an epidemiological study. J.
Gerontol., 44:10-18.

Doolittle, R.F., Hunkapiller, M.W., Hood, L.E., Devare, S.G., Robbins, K.C., Aaronson,
SA and Antoniades, HM (1983). Simian sarcoma virus onc gene, v-sis, is derived
from the gene(s) encoding a platelet-derived growth factor. Science, 221:275—277.

Dryja, T., Rapaport, J., Joyce, J., et al.. (1986). Molecular detection of deletions

61

involving band q14 of chromosome 13 in retinoblastomas. Proc. Natl. Acad. Sci. USA,
83:7391-7394.

Duyk, G.M., Kim, S.W., Myers, RM, and Cox, DR. (1990). Exon trapping: a genetic
screen to identify candidate transcribed sequences in cloned mammalian genomic DNA.
Proc. Natl. Acad. Sci. USA, 87: 8995-8999.

Egan, SE, and Weinberg, RA (1993). The pathway to signal achievement. Nature,
365:781-783.

Egan, S.E., Giddings, B.W., Brooks, M.W., Buday, L., Sizeland, AM., and Weinberg,
RA (1993). Association of $05 ras exchange protein with Grb2 is implicated in tyrosine
kinase signal transduction and transformation. Nature, 363245-51.

EI-Deiry, W.S., Tokino, T., Velculescu, V.E., Levy, D.B., Parsons, R., Trent, J.M., Lin,
0., Mercer, W.E., Kinzler, K.W., and Vogelstein, B. (1993). WAF1, a potential mediator
of p53 tumor suppression. Cell, 75:817-825.

Erikson, J., ar-Rushdi, A, Drwinga, H.L., Nowell, PC, and Croce, CM. (1983).
Transcriptional activation of the c-myc oncogene in Burkitt's lymphoma. Proc. Natl.
Acad. Sci. USA, 80:820-824.

Evan, G.l., Wyllie, AH., Gilbert, C.S. et al.. (1992). Induction of apoptosis in fibroblasts
by c-myc protein. Cell, 69:119-128.

Evans, E.P., Burtenshaw, M.D., Brown, B.B., Hennion, R., and Harris, H. (1982). The
analysis of malignancy by cell fusion. IX. Reexamination of clarification of the
cytogenetic problem. J. Cell Sci., 56:113-130.

Fakhaizadeh, S.S., Trusko, SP, and George, D.L. (1991). Tumorigenic potential
associated with enhanced expression of a gene that is amplified in a mouse tumor cell
line. EMBO J., 10:1565-1569.

Fanidi, A, Harrington, EA, and Even. G. (1992). Cooperative interaction between c-
myc and bcl-2 protooncogenes. Nature, 359:554-556.

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

Fearon, ER, and Jones, PA (1992). Progressing toward a molecular description of
colorectal cancer development. FASEB J., 6:2783-2790.

Feig, LA (1993). Science, 260:767-768.
Femandez-Pol, J.A., Klos, DJ, and Hamilton, PD. (1993). A growth factor-inducible
gene encodes a novel nuclear protein with zinc finger structure. J. Biol. Chem, 268:

21198-21204.

Fishel, R., Lescoe, M.C., Rao, M.R., Copeland, N.G., Jenkins, N.A., Garber, J., Kane,

62

M., et al., (1993). The mutation mutation gene homologue MSH2 and its association
with hereditary nonpolyposis colon cancer. Cell, 75:1027-1037.

Forrest, D., and Curran, T. (1992). Crossed signals: oncogenic transcription factors.
Curr. Opin. Genet. Dev., 2: 19-27.

Francke, U. (1978). Retinoblastoma and chromosome 13. Birth Defects, 12:131-137.

Frei, J. V., Swenson, D. H., Warren, W., and Lawley, P. D. (1978). Alkylation of
deoxyribonucleic acid in vivo in various organs of CS7BL mice by the carcinogens
N-methyl-N-nitrosourea, N-ethyl-N-nitrosourea and ethyl methanesulphonate in relation
to induction of thymic lymphoma. Some applications of high-pressure liquid
chromatography. Biochem. J., 174: 1031-1044.

Friend, S.H., Bemards, R., Rogelj, S., Weinberg, R.A, Rapaport, J.M., Albert, OM, and
Dryja, TR (1986). A human DNA segment with properties of the gene that predisposes
to retinoblastomas and osteosarcoma. Nature, 323:643-646.

Friis, R. R., Toyoshima, K., and Vogt, P. K. (1971).Conditional lethal mutants of avian
sarcoma viruses. 1. Physiology of ts 75 and ts 149. Virology. 43: 375-389.

Fukasawa, K., Choi, T., Kuriyama, R., Rulong, S., and Vande Woude, GE (1996).
Abnormal centrosome amplification in the absence of p53. Science, 271:1744-1747.

Galaktionov, K., Jessus, C., and Beach, D. (1995). Raf1 interaction with Cdc25
phosphatase ties mitogenic signal transduction to cell cycle activation. Genes Dev.,
9: 1046-1058.

Gallego, C., Gupta, S.K., Heasley, L. E., Qian, N. X., and Johnson, G.L. (1992).
Mitogen-activated protein kinase activation resulting from selective oncogene
expression in NIH 3T3 and rat 1a cells. Proc. Natl. Acad. Sci. USA, 89: 7355-7359.

Gallie, B.L., Squire, J.A., Goddard, A, et al.., (1990). Biology of disease: Mechanisms of
oncogenesis in retinoblastomas. Lab Invest, 62:394-408.

Garber, J.E., Goldstein, AM., Kantor, A.F., Dreyfus, M.G., Fraumeni, J.F., and Li, F-P.
(1991). Follow-up study of twenty-four families with Li-Fraumeni syndrome. Cancer
Res., 5126094.

Golde, A. (1970). Radio-induced mutants of the Schmidt-Ruppin strain of rous sarcoma
virus. Virology, 40: 1022-1029.

Goldfarb, M., Shimizu, K., Prucho, M., and Wigner, M. (1982). Isolation and preliminary
characterization of a human transforming gene from T24 bladder carcinoma cells.
Nature, 296:404-409.

Graeber, T.G., Osmanian, C., Jacks, T., Housman, D.E., Koch, C.J., Lowe, S.W., and
Giaccia, A.J. (1996). Hypoxia-mediated selection of cells with diminished apoptotic
potential in solid tumors. Nature, 379:88-91.

63

Grandis, JR, and Tweardy, DJ. (1993). Elevated levels of transforming growth factor
alpha and epidermal growth factor receptor messenger RNA are early markers of
carcinogenesis in head and neck cancer. Cancer Res., 53:3579—3584.

Graninger, W.B., Seto, M., Boutain, B., Goldman, P., and Korsmeyer, SJ. (1987).
Expression of Bcl-2 and Bcl-2-lg fusion transcripts in normal and neoplastic cells. J.Clin.
Invest. 8021512.

Green, R.S., Lieb, M.E., Weintraub, AS., Gacheru, S. N., Rosenfield, C.L., Shah, S.,
Kagan, HM, and Taubman, MB. (1995). Identification of lysyl oxidase and other
platelet-derived growth factor-inducible genes in vascular smooth muscle cells by
differential screening. Lab Invest, 73: 476-482.

Greenhalgh, D.A., Rothnagel, J.A., Quintanilla, M.I., Orengo, C.C., Gagne, T.A,
Bundman, D.S., Longley, MA, and R00p, DR. (1993). Induction of epidermal
hyperplasia, hyperkeratosis, and papillomas in transgenic mice by a targeted v-Ha-ras
oncogene. Moi. Carcinogenesis, 7: 99-1 10.

Groden, J., Thliveris, A, Samowitz, W., Carlson, M., Gelbert, L., Albertsen, H., Joslyn,
6., Stevens, J., Spirio, L., Robertson, M., et al. (1991). Identification and
characterization of the familial adenomatous polyposis coll gene. Cell, 66: 589-600.

Gutkind, S.J., Lacal, RM, and Robbim, KC. (1990). Thrombin-dependent association
of phosphatidylinositol-3 kinase with p60c-src and p59fyn in human platelets. Mol. Cell
Biol., 10:3806-3809.

Haefner, B., Baxter, R., Fincham, V.J., Downs, GP, and Frame, MC. (1995). J. Biol.
Chem., 270:7937-7943.

Hall, M., and Peters, G. (1996). Genetic alterations of cyclins, cyclin-dependent
kinases, and Cdk inhibitors in human cancer. Adv. Cancer Res., 68: 67-108.

Hansen, M., and Cavenee, W. (1987). Genetics of cancer disposition. Cancer Res.,
47:5518-5527.

Hara, E., Tsurui, H., Shinozaki, A, Nakada, S., and Oda, K. (1991). Cooperative effect
of antisense-Rb and antisense-p53 oligomers on the extension of life span in human
diploid fibroblasts, TlG-1. Biochem. Biophys. Res. Commun., 179: 528-534.

Harris, H. (1988). The analysis of malignancy by cell fusion: The position in 1988.
Cancer Res., 48:3302-3306.

Harris, H., Miller, O.J., Klein, G., Worst, P., and Tachibanaq, T. (1969). Suppression of
malignancy by cell fusion. Nature, 223:363-368.

Harris, H., and Klein, G. (1969). Malignancy of somatic cell hybrids. Nature, 224:1314-
1316.

Hartwell, L.H., and Kastan, MB. (1994). Cell cycle control and cancer. Science, 266:

64
1821 -8.

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

Helm, S., and Mitelman, F. (1989). Primary chromosome abnormalities in human
neoplasia. Adv. Cancer Res., 5221-43.

Heisterkamp, N., Stam, K., Grotten, J., deKlein, A. And Grosveld, G. (1985). Structural
organization of the bcr gene and its role in Ph' translocation. Nature, 315:758-761.

Heldin, CH, and Westermark, B. (1989). Platelet-derived growth factors: a family of
isoforms that bind to two distinct receptors. Brit. Med. Bull., 452453-464.

Hemminki, A, Tomlinson, l., Markie, D., Jarvinen, H., Sistonen, P., Bjorkqvist, A-M.,
Knuutila, S., Salovaara, R., Bodmer, W., Shibata, D., Chapelle, A, and Aaltonen, LA
(1997). Localization of a susceptibility locus for Peutz-Jeghers syndrome to 19p using
comparative genomic hybridization and targeted linkage analysis. Nature Genet,
15:87-90.

Hennings, H., Glick, A.B., Greenhalgh, D.A., Morgan, D.L., Strickland, J.E., and Yuspa,
SH. (1993). Critical aspects of initiation, promotion, and progression in multistage
epidermal carcinogenesis. Proc. Soc. Exp. Biol. Med, 20221-8.

Hennings, H., Shores, R.A., Poitier, M.C., Reed, E., Tarone, RE, and Yuspa, SH.
(1990). Enhanced malignant conversion of benign mouse skin tumors by cisplatin. J.
Natl. Cancer Inst, 82:832-840.

Herman, J.G., Merlo, A, Mac, L., Lapidus, R.G., lssa, J-P., Davidson, N.E., Sidransky,
D., and Baylin, SB. (1995). Inactivation of the CDKN2/p16IMTS1 gene is frequently
associated with aberrant DNA methylation in all common human cancers. Cancer Res.,
55:4525-4530.

Hill, M., and Hillova, J. (1972). Virus recovery in chicken cells tested Rous sarcoma cell
DNA. Nature New Biol., 237235-39.

Hirama, T., and Koeffler, HP. (1995). Role of the cyclin-dependent kinase inhibitors in
the development of cancer. Blood, 86:841-854.

Hockenbery, OM. (1994). Bel-2 in cancer, development and apoptosis. J. Cell Sci.,
18:51-55.

Housey, G.M., Johnson, MD, Hsiao, W.L. et al., (1988). Overproduction of protein
kinase C causes disordered growth control in rat fibroblasts. Cell, 52:343-354.

Hollstein, M., Rice, K., Greenblatt, M.S., Soussi, T., Fuchs, R., Sorlie, T., Hovig, E.,
Smith-Sorensen, B., Montesano, R., and Harris, CC. (1994). Database of p53 gene
somatic mutations in human tumors and cell lines. Nucleic Acids Res., 22: 3551-3555.

65

Holt, J.T., Gopal, T.V., Moulton, AD., and Nienhuis, AW. (1986). Inducible production
of c-fos antisense RNA inhibits 3T3 cell proliferation. Proc. Natl. Acad. Sci. USA, 83:
4794-4798.

Horowitz, J.M., Park, S.H., Bogenmann, E., Cheng, J.C., Yandell, D.W., kaye, F.J.,
Minna, J.D., Dryja, T.P., and Weinberg, RA (1990). Frequent inactivation of the
retinoblastoma antioncogene is restricted to a subset of human tumor cells. Proc. Natl.
Acad. Sci. USA, 87:2775-2779.

Howe, LR, and Marshall, C.J. (1993). Lysophosphatidic acid stimulates
mitogen-activated protein kinase activation via a G-protein-coupled pathway requiring
p21ras and p74raf-1. J. Biol. Chem., 268: 20717-20720.

Hsiao, W.L., Housey, G.M., Johnson, MD. et al.. (1989). Cells that overproduce protein
kinase C are more susceptible to transformation by an activated H-ras oncogene. Mol.
Cell Biol., 922641-2647.

Huberman, E., and Sachs, L. (1974). Cell-mediated mutagenesis of mammalian cells
with chemical carcinogens. Int. J. Cancer, 132326-333.

Huberman, E., and Sachs, L. (1976). Mutability of different genetic loci in mammalian
cells by metabolically activated carcinogenic polycyclic hydrocarbons. Proc. Natl. Acad.
Sci. USA, 73:188-192.

Huebner, R.J., and Todaro, G.J. (1969) Oncogenes of RNA tumor viruses as
determinants of cancer. Proc. Natl. Acad. Sci. USA, 6421087-1094.

Hunter, T., and Pines, J. (1994). Cyclins and cancer. ll: Cyclin D and CDK inhibitors
come of age. Cell, 79: 573-582.

Hunter, T. (1997). Oncoprotein networks. Cell, 882333-346.
Hunter, T. (1991). Cooperations between oncogenes. Cell, 642249-270.
Hunter, T. (1987). A thousand and one protein kinases. Cell, 502 823-829.

HUpp, TR, and Lane, DP. (1994). Allosteric activation of latent p53 tetramers. Curr.
Biol., 42865-875.

lannaccone, P.M., Weinberg, WC, and Deamant, FD. (1987). On the clonal origin of
tumors: a review of experimental models. Int. J. Cancer, 39:778-784.

lchii, S., Takeda, S., Horii, A, Nakatsuru, S., Miyoshi, Y., Emi, M., Fujiwara, Y.,
Koyama, K., Furuyama, J., Utsunomiya, J., et al. (1993). Detailed analysis of genetic
alterations in colorectal tumors from patients with and without familial adenomatous
polyposis (FAP). Oncogene, 82 2399-405.

Kallioniemi, A, Kallioniemi, O.P., Sudar, D., Rutovitz, D., Gray, J.W., Waldman, F., and
Pinkel, D. (1992). Comparative genomic hybridization for molecular cytogenetic

66
analysis of solid tumors. Science, 258:818-821.

Kallioniemi, OF. (1997). Linking chromosomal clues. Nature Genet, 1525-6.

Kamb, A, Gruis, N.A, Weaver-Feldhaus, J., Liu, 0., Harshman, K., tavigian, S.V.,
Stochert, E., Day, R.S.Ill, Johnson, BE, and Skolnick, M.H. (1994). A cell cycle
regulator potentially involved in genesis of many tumor types. Science, 264:436-440.

Katsumata, M., Siegel, R.M., Louie, D.C. et al.. (1992). Differential effects of Bcl-2 on T
and B cells in transgenic mice. Proc. Natl. Acad. Sci. USA, 89:11376-11380.

Kawai, S., and Hanafusa, H. (1971). The effects of reciprocal changes in temperature
on the transformed state of cells infected with a rous sarcoma virus mutant. Virology,
46: 470-479.

Khazaie, K., Dull, T.J., Graf, T., Schlessinger, J., Ullrich, A, Beug, H., and Vennstomi,
B. (1988). Truncation of the human EGF receptor leads to different transforming
potentials in primary avian fibroblasts and erythroblasts. EMBO J., 723061-3071.

Khosravi-Far, R., Solski, P.A, Clark, G.J., Kinch, MS, and Der, C.J. (1995). Activation
of Rac1, RhoA, and mitogen-activated protein kinases is required for ras transformation.
Mol. Cell Biol., 152 6443-6453.

Khosravi-Far, R., White, M. A, Westwick, J.K., Solski, P.A, Chrzanowska-Wodnicka,
M., Van-Aelst, L., Wigler, M.H., and Der, C.J. (1996). Oncogenic ras activation of
Raf/mitogen-activated protein kinase-independent pathways is sufficient to cause
tumorigenic transformation. Mol. Cell Biol., 16: 3923-3933.

Kikkawa, U., Kishimoto, A, and Nishizuka, Y. (1989). The protein kinase C family:
Heterogeneity and its implications. Annu. Rev. Biochem., 58:31-44.

Kinzler, K.W., Nilbert, M.C., Su, L.K., Vogelstein, B., Bryan, T.M., Levy, D.B., Smith,
K.J., Preislnger, AC., Hedge, P., McKechnie, D., et al. (1991). Identification of FAP
locus genes from chromosome 5q21. Science, 253: 661-665.

Kinzler, K.W., and Vogelstein, B. (1996). Life and death in a malignant tumor. Nature,
379219-20.

Kitayama, H., Sugimoto, Y., Matsuzaki, T., lkawa, Y., and Noda, M. (1989). A ras-
related gene with transformation suppressor activity. Cell, 56277-84.

Klein, G. (1981). The role of gene dosage and genetic transposition in carcinogenesis.
Nature, 294:313-318.

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

Klinger, HP. (1982). Suppression of tumorigenicity. Cytogenet. Cell Genet, 32:68-84.

67

Knudson, AG. (1971). Mutation and cancer: statistical study of retinoblastoma. Proc.
NatI.Acad. Sci. USA. 68:820-823.

Knudson, AG. (1986). Genetics of human cancer. Ann. Rev. Genet, 20:231-251.
Knudson, AG. (1993). All in the cancer family. Nature Genet, 52103-104.

Knudson, A. G. (1989). Hereditary cancers: Clues to mechanisms of carcinogenesis.
Brit. J. Cancer, 59:661-666.

Knudson, AG. Genetic predisposition to cancer, in Hiatt, H.H., Watson, JR, and
Winsten, J.A(eds). Origin of Human Cancer, Cold Spring Harbor, NY: Cold Spring
Harbor Laboratory, 1977, pp.45-52.

Konopka, J.B., Watanabe, S.M., Witte, ON. (1984). An alteration of the human c-abl
protein in K562 leukemia cells unmasks associated protein tyrosine kinase activity. Cell,
3721035.

Kovary, K., and Bravo, R. (1991). The jun and fos protein families are both required for
cell cycle progression in fibroblasts. Mol. Cell Biol., 11: 4466-4472.

Krengel, U., Schlichting, L., Scherer, A, Schumann, R., Frech, M., John, J., Sabsch, W.,
Pai, BF, and Wittinghofer, A. (1990). Three-dimensional structures of H-ras p21
mutants: Molecular basis for their inability to function as signal switch molecule. Cell,
62:539-548.

Kress, S., Sutter, C., Strickland, P.T., Mukhtar, H., Schweizer, J., and Schwartz, M.
(1992). Carcinogen-specific mutational pattern in the p53 gene in ultraviolet B radiation-
induced squamous cell carcinomas of mouse skin. Cancer Res., 52:6400-6403.

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

Krystal, G.W., Hines, S.J., and Organ, CF. (1996). Autocrine growth of small cell lung
cancer mediated by coexpression of c-kit and stem cell factor. Cancer Res., 56:370-
376.

Kuroki, T., and Huh, N-h. (1993). Why are human cells resistant to malignant cell
transformation in vitro. Jpn. J. Cancer Res., 84:1091-1100.

Lane, DP. (1994). p53 and human cancers. Brit. Med. Bulletin., 50:582-599.

Lathangue, MB. (1994). DRTF1IE2F2 an expanding family of heterodimeric transcription
factors implicated in cell-cycle control. Trends Biochem. Sci., 192108-114.

Lawley, PD. (1989). Mutagens as carcinogens: development of current concepts.
Mutation Res., 21323-25.

Leder, P., Battey, J., Lenoir, G., Moulding, G, Murphy, W., Potter, H., Stewart, T., and

68

Taub, R. (1983). Translocation among antibody genes in human cancer, Science,
222:765-771.

Lee, W-H., Boohstein, R., Hong, P., Young, L-J., Shew, J-Y., and Lee, E.Y-H.P. (1987).
Human retinoblastoma susceptibility gene: cloning, identification, and sequence.
Science, 235:1394-1399.

Lee, S.W., Tomasetto, C., and Sager, R. (1991). Positive selection of candidate tumor
suppressor genes by subtractive hybridization. Proc. Natl. Acad. Sci. USA, 8822825-
2829.

Leonard, D.G.B., Ziff, EB, and Greene, L.A(1987). Identification and characterization
of mRNAs regulated by nerve growth factors in PC12 cells. Mol. Cell Biol., 723156-
3167.

Levine, AJ. (1993). The tumor suppressor genes. Annu. Rev. Biochem., 62:623-651.

Levine, A.J. (1997). p53, the cellular gatekeeper for growth and division. Cell, 88:323-
331.

Lewin, B. (1991). Oncogenic conversion by regulatory changes in transcription factors.
Cell, 64:303-312.

Li, PP (1988). Cancer families: Human models of susceptibility to neoplasia-the
Richard and Hinda Rosenthal Foundation Award Lecture. Cancer Res., 4825381-5386.

Liang, P., and Pardee, A. B. (1992). Differential display of eukaryotic messenger RNA
by means of the polymerase chain reaction. Science, 257: 967-969.

Liang, P., Averboukh, L., and Pardee, A. B. (1993). Distribution and cloning of
eukaryotic mRNA by means of differential display: refinements and optimization.
Nucleic Acids Res., 21 :3269-3275.

Lin, C., Maher, V.M., and McCormick, J.J. (1995) Malignant transformation of human
fibroblast strain MSU-1.1 by v-fes requires an additional genetic change. Int. J.Cancer.,
63: 140-7

Lin, J., Wu, X., Chen, J., Chang, A, and Levine, A.J. (1995). Functions of the p53
protein in growth regulation and tumor suppression. Cold Spring Harbor Symposia on
Quantitative Biology LIX, 215-223.

Lindahl,T. (1994). DNA surveillance defect in cancer cells. Current Bi0109Y. 4:249.
Lipponen, P., and Eskelinen, M. (1994). Expression of epidermal growth factor receptor
in bladder cancer as related to established prognostic factors, oncoprotein (c-erbB-2,

p53) expression and long-term prognosis. Brit. J. Cancer, 6921120-1125.

Lovec, H., Grzeschiczek, A, Kowalski, MB, and Moroy, T. (1994). Cyclin D1/bcl-1
cooperates with myc genes in the generation of B-cell lymphoma in transgenic mice.

69
EMBO J., 13:3487-3495.

Lowe, D.G., Capon, D.J., Delwart, E., Sakaguchi, A.Y., Naylor, S.L., and Goeddel,
D.V. (1987). Structure of the human and murine R-ras genes, novel genes closely
related to ras proto-oncogenes. Cell, 48: 137-146.

Lowe, S.W., Schmitt, E.M., Smith, S.W., Osborne, BA, and Jacks, T. (1993). P53 is
required for radiation induced apoptosis in mouse thymocytes. Nature, 362:847-849.

Lu,H., and Levine, A.J. (1995). Human TAF-31 is a transcriptional coactivator of the
p53 protein. Proc. Natl. Acad. Sci. USA, 92:5154-5158.

Maher, V.M., Miller, E.C., Miller, J.A., and Szybalski, W. (1968). Mutations and
decreases in density of transforming DNA produced by derivatives of the carcinogens
2-acetyl-aminofluorene and N-methyl-4-aminoazobenzene. Mol. Pharmacol., 42
411-426.

Maki, Y., Bos, T.J., Davis, 0., Starbuck, M., and Vogt, PK. (1987). Avian sarcoma virus
17 carries the jun oncogene. Proc. Natl. Acad. Sci. USA, 84:2848-2852.

Mandi, M., Paffenholz, R., Friedl, W., Caspari, R., Sengteller, M., and Propping, P.
(1994). Frequency of common and novel inactivating APC mutations in 202 families with
familial adenomatous polyposis. Hum. Mol. Genet, 3: 181-184.

Margolis, B., Li, M, Koch, A, et al.. (1990). The tyrosine phosphorylated carboxyl
terminus of the EGF receptor is a binding site for GAP and PLC-y. EMBO J., 924375-
4380.

Markowitz, S., Wang, J., Myeroff, L., Parsons, R., Sun, L., Lutterbaugh, J., Fan, R. S.,
Zborowska, E., Kinzler, K.W., Vogelstein, B., et al.. (1995). Inactivation of the type II
TGF-beta receptor in colon cancer cells with microsatellite instability. Science, 268:
1336-1338.

Maron, D.M., and Ames, B.N. (1983). Revised methods for the Salmonella mutagenicity
test. Mutat. Res., 1132173-215.

Marshall, C.J. (1991). Tumor suppressor genes. Cell, 64:313-326.
Marshall, MS. (1995). ras target protein in eukaryotic cells. FASEB J., 9:1311-1318.

Martin, GS. (1970). Rous sarcoma virus: A function required for the maintenance of
the transformed state. Nature, 22721021-1023.

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

McCann, J., Choi, E., Yamasaki, E., and Ames, B.N. (1975). Detections of carcinogens

70

as mutagens in the Salmonella/microsome test: Assay of 300 chemicals. Proc. Natl.
Acad. Sci. USA, 7225135.

McCormick, J.J., and Maher. V.M. (1994). Analysis of the multistep process of
carcinogenesis using human fibroblasts. Risk Anal., 14:257-263.

McCormick, F. (1993). How receptors turn ras on. Nature, 363215-16.

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

McDonnell, T.J., Troncoso, P., Brisbay, SM, et al.. (1992). Expression of the proto-
oncogene bcI-2 in the prostate and its association with emergence of androgen-
independent prostate cancer. Cancer Res., 52: 6940-6944.

McDonnell, T.J., and Korsmeyer, SJ. (1991). Progression from lymphoid hyperplasia to
high-grade malignant lymphoma in mice transgenic for the t(14;18). Nature, 3492254-
256.

McGIade, J., Cheng, A, Pelicci, G., Pelicci, PG, and Pawson, T. (1992). She proteins
are phosphorylated and regulated by the v-Src and v-Fps protein-tyrosine kinases. Proc.
Natl. Acad. Sci. USA, 89:8869-8873.

Miki, T., Fleming, T.P., Crescenzi, M., Molloy, C.J., Blam, S.B., Reynolds, SH, and
Aaronson, SA (1991). Development of a highly efficient expression cDNA cloning
system: application to oncogene isolation. Proc. Natl. Acad. Sci. USA, 882 5167-71.

Miki, T., and Aaronson, SA (1995). Isolation of oncogenes by expression cDNA
cloning. Methods Enzymol., 254: 196-206.

Miller, EC. (1978). Some current perspectives on chemical carcinogenesis in human
and experimental animals: Presidential Address. Cancer Res., 3821479.

Miller, EC, and Miller, J.A. (1981). Searches for ultimate chemical carcinogens and
their reaction with cellular macromolecules. Cancer, 4722327.

Miyashita, T., and Reed, JC. (1993). Bcl-2 oncoprotein blocks chemotherapy-induced
apoptosis in a human leukemia cell line. Blood, 812151-157.

Modrich, P. (1994). Mismatch repair, genetic instability, and cancer. Science, 26621059-
1060.

Moore, R., Casey, G., Brookes, 8., Dixon, M., Peters, G., and Dickson, C. (1986).
Sequence, topography and protein coding potential of mouse int-2: a putative oncogene
activated by mouse mammary tumor virus. EMBO J., 5: 919-24.

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

71
oncogene. Exp. Cell Res.,197: 125-136.

Morgan, DO. (1995). Principle of CDK regulation. Nature, 374:131-134.

Motokura, T., Bloom, T., Kim, H.G., Juppner, H., Rudennan, J.V., Kronberg, HM, and
Arnold, A. (1991). A novel cyclin encoded by a bcI-1 linked candidate oncogene.
Nature, 3502512-515.

Mottram, J. C. (1944). A developing factor in experimental blastogenesis. J. Pathol.
Bacterial, 562181-187.

Namba,M., Nishitani,K., Fukushima, F., Kimoto, T. (1988). Multistep carcinogenesis of
normal human fibroblasts. Human fibroblasts immortalized by repeated treatment with
Co-60 gamma rays were transformed into tumorigenic cells with Ha-ras oncogenes.
Anticancer Res., 82 947-958

Nehls, M., Pfeifer, D., and Boehm, T. (1994). Exon amplification from complete libraries
of genomic DNA using a novel phage vector with automatic plasmid excision facility:
application to the mouse neurofibromatosis-1 locus. Oncogene, 92 2169-2175.

Nelson, M.A, Futscher, B.W., Kinsella, T., Wymer, J., and Bowden, GT (1992).
Detection of mutant Ha-ras genes in chemically initiated mouse skin epidermis before
the development of benign tumors. Proc. Natl. Acad. Sci. USA, 89:6398-6402.

Nevins, JR. (1992). E2F: a link between the Rb tumor suppressor protein and viral
oncoproteins. Science, 2582424-429.

Nicolaides, N.C., Papadopoulos, M, Liu, B., Wei, Y.F., Carter, K.C., Ruben, S.M.,
Rosen, C.A., Haseltine, W.A, Fleischmann, R.D., Fraser, C.M., et-al. (1994). Mutations
of two PMS homologues in hereditary nonpolyposis colon cancer. Nature, 371: 75-80.

Nishikawa, R., Ji, XD., Harmon, R.C., lazar, C.S., Gill, G.N., Cavenee, W.K., and
Huang, H.J. (1994). Proc. Natl. Acad. Sci. USA, 9127727-7731.

Nishizuka, Y. (1992). Intracellular signaling by the hydrolysis of phospholipids and
activation of protein kinase C. Science, 2582607.

Nobori, T., Miura, K., Wu, D.J., Lois, A, Takabayashi, K., and Carson, DA (1994).
Deletion of the cyclin-dependent kinase 4 inhibitor gene in multiple human cancers.
Nature, 368:753-756.

Noda, M. (1990). Expression cloning of tumor suppressor genes: A guide for optimists.
Moi. Carcinogenesis, 32251-253.

Noda, M., Selinger, Z., Scolnick, EM, and Bassin, RH. (1983). Flat revertants isolated
from kirsten sarcoma virus-transformed cells are resistant to the action specific
oncogenes. Proc. Natl. Acad. Sci. USA, 8025602-5606.

Okamoto, A, Demetrick, D.J., Spillare, E.A., Hagiwara, K., Hussein, S.P., Bennett,

72

WP, Forrester, K., Gerwin, B., Serrano, M., Beach, DH, and Harris, CC. (1994).
Mutations and altered expression of p161NK4 in human cancer. Proc. Natl. Acad. Sci.
USA, 91:11045-11049.

O’Connell, J.F., Klein-Szanto, A.J.P., Digiovanni, D.M., Fries, J.W., and Slaga, T.J.
(1986). Malignant progression of mouse skin papillomas treated with ethylnitrosourea,

N-methyI-N’-nitro-N-nitrosoguanidine, or 12-O-tetradecanoylphorbol-13-acetate. Cancer
Letter, 302269-274.

Packham, G., and Cleveland, J.L. (1995). C-Myc and apoptosis. Biochim. Biophy.
Acta, 1242:11-28.

Papadopoulos, N., Nicolaides, N.C., Wei, Y.F., Ruben, S.M., Carter, K.C., Rosen, C.A,
Haseltine, W.A., Fleischmann, R.D., Fraser, C.M., Adams, MD, et al. (1994). Mutation
of a mutL homolog in hereditary colon cancer. Science, 2632 1625-1629.

Pardee, AB. (1989). G1 events and regulation of cell proliferation. Science, 2462603-
608.

Pelicci, G., Lanfrancone, L., Grignani, F., McGIade, J., Cavallo, F., et al., (1992). A
novel transforming protein (Shc) with an SH2 domain is implicated in mitogenic signal
transduction. Cell, 70293-104.

Perucho, M., Goldfarb, M., Shimizu, K., Lama, C., Fogh, J., and Wigler, M. (1981).
Human-tumor-derived cell lines contain common and different transforming genes. Cell,
27:467-476.

Peto, R. (1977). Epideology, multistage models, and short term mutagenesis tests. In
Origins of Human Cancer. Cold Spring Harbor Laboratory, Cold Spring Harbor,
NY,1403-1428.

Pezzella, F., Turley, H., Kuzu, I., et al.. (1993). Bel-2 protein in non-small cell lung
carcinoma. New Engl. J. Med., 329:690-694.

Phuchareon, J., and Tokuhisa, T. (1995). Deregulated c-FosIAP-1 modulates
expression of the cyclin and the cdk gene in splenic B cells stimulated with
Iipopolysaccharide. Cancer Lett, 92: 203-208.

Plowman, G.D., Green, J.M., Culouscou, J.M., Carlton, G. W., Rothwell, V.M., and
Buckley, S. (1993). Heregulin induces tyrosine phosphorylation of HER4Ip180erbB4.
Nature, 3662 473-475.

Pollock, P.M., Pearson, JV, and Hayward, MK. (1996). Compilation of somatic
mutations of the CDKN2 gene in human cancers: non-random distribution of base
substitutions. Genes Chromosome Cancer, 15277-88.

Powell, S.M., Petersen, G.M., Krush, A.J., Booker, S, Jen, J., Giardiello, F.M.,
Hamilton, S.R., Vogelstein, B., and Kinzler, K.W. (1992). Molecular diagnosis of familial
adenomatous polyposis. N. Engl. J. Med., 329: 1982-987.

73

Qiu, R.G., Chen, J., Kim, D., McCormick, F., and Symons, M. (1995). An essential role
for Rac in ras transformation. Nature, 3742 457-459.

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

Rabbits, TH. (1994). Chromosomal translocations in human cancer. Nature, 372:143-
149.

Rapp, U.R. (1990). Role of Raf-1 serine/threonine protein kinase in growth factor signal
transduction. Oncogene, 62495-500.

Reed, J.C., Cuddy, M., Slabiak, T., Croce, OM, and Nowell, PC. (1988). Oncogenic
potential of bcl-2 demonstrated by gene transfer. Nature, 336: 259-61.

Renan, M.J. (1993). How many mutations are required for tumorigenesis? Implications
from human cancer data. Mol. Carcinogenesis, 72139-146.

Rhim, J.S. (1993). Neoplastic transformation of human cells in vitro. Crit. Rev. Oncol.,
42313-335.

Riccardi, V.M., Hittner, H.M., Francke, U., et al.. (1990). The aniridia-Wilms’ tumor
association: The critical role of chromosome band 11p13. Cancer Genet. Cytogenet.,
22131-137.

Rosenberg, C.L., Wong, E., Petty, E.M., Bale, AE., Tsujimoto, Y., Harris, ML, and
Arnold, A. (1991b). PRAD1, a candidate BCL1 oncogene: mapping and expression in
centrocytic lymphoma. Proc. Natl. Acad. Sci. USA, 88:9638-9642.

Rosenberg, C.L., Kim, H.G., Shows, T.B., Kronberg, HM, and Arnold, A. (1991a).
Rearrangement and overexpression of D118287E, a candidate oncogene on
chromosome 11q13 in benign parathyroid tumors. Oncogene, 62449-453.

Rotin, D., Margolis, B., Mohammadi, M., Daly, R.J., Daum, G., Li, M, Fischer, E. H.,
Burgess, W.H., Ullrich, A, and Schlessinger, J. (1992). SH2 domains prevent tyrosine
dephosphorylation of the EGF receptor identification of Tyr992 as the high-affinity
binding site for SH2 domains of phospholipase C gamma. EMBO J., 11:559-567.

Rous, P. (1911) A sarcoma of the fowl transmissible by an agent separable from the
tumor cells. J. Exp. Med, 13:397-411.

Rous, P. (1910) A transmissible avian neoplasm: sarcoma of the common fowl. J. Exp.
Med., 12:696.

Ruddon, R.W. (1995). In Cancer Biology. Oxford University Press, pp.62-69.

Ruggeri, B., Caamano, J., Goodrow, T., Dirado, M., Bianchi, A, Trono, D., Conti, C.J.,
and KIein-Szanto, A.J. (1991). Alterations of the p53 tumor suppressor gene during

74
mouse skin tumor progression. Cancer Res., 51 :6615-6621.

Ruggeri, B., Dirado, M., Zhang, S.Y., Bauer, B., Goodrow, T., and Klein-Szanto, A.J.
(1993). Benzo(a)pyrene-induced murine skin tumors exhibit frequent and characteristic
G to T mutations in the p53 gene. Proc. Natl. Acad. Sci. USA, 9021013-1017.

Ruther, U., Komitowski, D., Schubert, ER, and Wagner, E.F. (1989). C-fos expression
induced bone tumors in transgenic mice. Oncogene, 42861-865.

Ruther, U., Garber, C., Komitowski, D., Muller, R., and Wagner, E.F. (1987).
Deregulated c-fos expression interferes with normal bone development in transgenic
mice. Nature, 325:412-416.

Saez, R., Chan, AM., Miki, T., and Aaronson, SA (1994). Oncogenic activation of
human R-ras by point mutations analogous to those of prototype H-ras oncogenes.
Oncogene, 92 2977-2982.

Saez, E., Rutberg, S., Mueller, E., Oppenheim, H., Smoluk, J., Yuspa, H., and
Spiegelman, BM. (1995). C-fos is required for malignant progression of skin tumors.
Cell, 822721-732.

Sager, R. (1985). Genetic suppression of tumor formation. Adv. Cancer Res., 44:43-
68.

Sager, R. (1989). Tumor suppressor genes: The puzzle and promise. Science,
246:1406-1412.

Sager, R. (1997). Expression genetics in cancer: Shifting the focus from DNA to RNA.
Proc. Natl. Acad. Sci. USA, 94:952-955.

Samuels, M.L., Weber, M. J., Bishop, J. M., and McMahon, M. (1993). Conditional
transformation of cells and rapid activation of the mitogen-activated protein kinase
cascade by an estradiol-dependent human raf-1 protein kinase. Mol. Cell Biol., 13:
6241-6252.

Santos, E., Tronick, S.R., Aaronson, S.A, Pulciani, S., and Barbacid, M. (1982). T24
human bladder carcinoma oncogenes is an activated form of the normal human
homologue of BALB and Harvey-MSV transforming genes. Nature, 2982343.

Sargent, TD. (1987). Meth. Enzymol., 152:423-431.

Sarma, V., Wolf, F.W., Marks, R.M., Shows, TB, and Dixit, V.M. (1992).Cloning of a
novel tumor necrosis factor-alpha-inducible primary response gene that is differentially
expressed in development and capillary tube-like formation in vitro. J lmmunol., 148:
3302-3312.

Sassone-Corsi, P. (1992). Fatal attraction: The fos-jun affair with nuclear receptors. lntl.
J.Oncol., 1:5.

75

Schaefer, R., lyer, J., lten, E., and Nirkko, A. (1988). Partial reversion of the
transformed phenotype in HRAS-transfected tumorigenic cells by transfer of a human
gene. Proc. Natl. Acad. Sci. USA, 85:1590-1594.

Schlessinger, J., and Ullrich, A. (1992).Growth factor signaling by receptor tyrosine
kinases. Neuron, 9: 383-91.

Schmid, M., Haaf, T., and Grunert, D. (1984). 5-Azacytidine-induced
undercondensations in human chromosomes. Hum. Genet, 67:257-263.

Schuh, A.C., Keating, S.J., Monteclaro, F.S., Vogt, PK, and Breitman, ML. (1990).
Obligatory wounding requirement for tumorigenesis in v-jun transgenic mice. Nature,
3462756-760.

Schwab, M., Varmus, HE, and Bishop, J.M. (1985). Human N-myc gene contributes to
neoplastic transformation of mammalian cells in culture. Nature, 316: 160-162.

Seger, R., and Krebs, EC. (1995). The MAPK signaling cascade. FASEB J., 92726-
735.

Sentman, C.L., Shutter, J.R., Hockenbery, D., Kanegawa, O., and Korsmeyer, SJ.
(1991). Bcl-2 inhibits multiple forms of apoptosis but not negative selection in
thymocytes. Cell, 67:879-888.

Serrano, M., Lee, H-W., Chin, L., Cordon-Cardo, C., Beach, 0., and DePinho, RA
(1996). Role of the INK4a locus in tumor suppression and cell mortality. Cell, 85:27-37.

Seto, M., Yamamoto, K., lida, S., Akao, Y., Utsumi, K.R., Kubonishi, l., Miyoshi, l.,
Ohtsuki, T., yawata, Y., Namba, M., Motokura, T., Arnold, A, Takahashi, T. And Ueda,
R. (1992). Gene rearrangement and overexpression of PRAD1 in lymphoid malignancy
with t(1 1 ;14) (q13;q32) translocation. Oncogene, 721401-1406.

Shay, J.W., Wright,W.E., Werbin, H. (1991). Defining the molecular mechanisms of
human cell immortalization. Biochim. Biophys. Acta, 1072: 1-7.

Shay, J.W., and Wright, W.E. (1989). Quantitation of the frequency of immortalization of
normal human diploid fibroblasts by SV40 large T-antigen. Exp. Cell Res., 1842109.

Sherr, C.J. (1990). Colony-stimulating factor-1 receptor. Blood, 75:1-12.
Sherr, C.J. (1993). Mammalian G1 cyclins. Cell, 73:1059-1065.
Sherr, C. J. (1994). G1 phase progression: cycling on cue. Cell, 79: 551-555.

Sherr, C.J., and Roberts, J.M. (1995). Inhibitors of mammalian G1 cyclin-dependent
kinases. Genes Dev., 921149-1163.

Sherr, C.J. (1996). Cancer cell cycle. Science, 274:1672-1677.

76

Shih, C., Padhy, L.C., Murray, M., and Weiberg, RA (1981). Transforming genes of
carcinomas and neuroblastomas introduced into mouse fibroblasts. Nature, 2902261-
264.

Shih, C., Shilo, B-Z., Goldfarb, M.P., Dannenberg, A, and Weinberg, RA (1979).
Passage of phenotypes of chemically transformed cells via transfection of DNA and
chromatin. Proc. Natl. Acad. Sci. USA, 76:5714-5718.

Shih, C., and Weiberg, RA (1982). Isolation of a transforming sequence from a human
bladder carcinoma cell line. Cell, 29:161-169.

Shirai, H., Ueno, E., Osaki, M., Tatebe, S., Ito, H., Kaibara, N. (1995). Expression of
growth factors and their receptors in human early colorectal carcinomas:
immunohistochemical Study. Anticancer Res., 15:2889-2894.

Simon, H-G., Risse, B., Jost, M., Oppenheimer, 8., Karl, C., and Rodeck, U. (1996).
Identification of differentially expressed messenger RNAs in human melanocytes and
melanoma cells. Cancer Res., 56:3112-3117.

Slamon, D.J., Godophin, W., Jones, L.A, Holt, J.A., Wong, S.G., Keith, D.E., Levin,
W.J., Stuart, S.G., Udove, J., Ullrich, A, and Press, MP. (1989). Studies of the
Her2/neu proto-oncogene in human breast and ovarian cancer. Science, 244:707-712.

Smeyne, R.J., Vendrell, M., I-layward, M., et al., (1993). Continuous c-fos expression
precedes programmed cell death in vivo. Nature, 363:166-169.

Smith, JR, and Pereira-Smith, OM. (1996). Replicative senescence: implication for in
vivo aging and tumor suppression. Science, 2732 63-67.

Smith, M.J., and Prochownik, E. V. (1992). Inhibition of c-jun causes reversible
proliferative arrest and withdrawal from the cell cycle. Blood, 79: 2107-2115.

Sobel, ME. (1990). Metastasis suppressor genes. J. Natl. Cancer Inst, 82: 267-276.

Solomon. E., Borrow, J., Goddard, AD. (1991). Chromosome aberrations and cancer.
Science, 25421 153-1 160.

Spector, D.H., Varmus, HE, and Bishop, J.M. (1978a). Nucleotide sequences related to
the transforming gene of avian sarcoma virus are present in DNA of uninfected
vertebrates. Proc. Natl. Acad. Sci.USA, 75: 4102-4106.

Spector, D.H., Smith, K., Padgett, T., McCombe, P., Roulland-Dussoix, D., Moscovic,
C., Varmus, HE, and Bishop, J.M. (1978b). Uninfected avian cells contain RNA
related to the transforming gene of avian sarcoma viruses. Cell, 13:371-379.

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

Srivastava, S., Zou, Z., Pirollo, K., Blattner, W., and Chang, EH. (1990). Gemi-line

77

transmission of a mutated p53 gene in a cancer-prone family with Li-Fraumeni
syndrome. Nature, 348:747-749.

Stanbridge, E.J., Flanderrneyer, R.R., Daniels, 0., and Nelson-Rees, WA (1981).
Specific chromosome loss associated with the expression of tumorigenicity in human
cell hybrids. Somatic Cell Genet, 7:699-712.

Stehelin, D., Guntaka, R.V., Varmus, HE, and Bishop, J.M. (1976a). Purification of
DNA complementary to nucleotide sequences required for neoplastic transformation of
fibroblasts by avian sarcoma viruses. J. Mol. Biol., 1012 349-365.

Stehelin, D., Varmus, H., Bishop, J.M., and Vogt, P.K. (1976b) DNA related to the
transforming genes of avian sarcoma viruses is present in normal avian DNA. Nature,
2602170-173.

Stephens, R.M., Sithanandam, G., Copeland, T.D., Kaplan, D.R., Rapp. U.R., and
Morrison, D.K. (1992). 95-kilodalton B-Raf serine/threonine kinase: identification of the
protein and its major autophosphorylation site. Mol. Cell Biol., 12: 3733-3742.

Storm, SM, and Rapp, U.R. (1993). Oncogene activation: c-raf-1 gene mutations in
experimental and naturally occurring tumors. Toxicol. Lett, 67: 201-210.

Storm, S.M., Brennscheidt, U., Sithanandam, G., and Rapp, U.R. (1990). raf oncogenes
in carcinogenesis. Crit. Rev. Oncog., 2: 1-8.

Strasser, A, Harris, A.W., Bath, M.L., and Cory, S. (1990). Novel primitive lymphoid
tumors induced in transgenic mice by cooperation between myc and bcl-2. Nature,
3482331-333.

Strasser, A, Harris, A.W., and Cory, S. (1991). Bcl-2 transgene inhibits T cell death and
perturbs thymic self-censorship. Cell, 67:889-899.

Sun, Y., Hegamyer, G., and Colbum, N. H.(1994). Molecular cloning of five messenger
RNAs differentially expressed in preneoplastic or neoplastic JB6 mouse epidermal cells:
One is homologous to human tissue inhibitor of metalloproteinases-3. Cancer Res.,
5421 139-1 144.

Swenson, D.H., Petzold, G.L., and Harbach, PR. (1986). The binding of
1-(2-hydroxyethyl)-1-nitrosourea to DNA in vitro and to DNA of thymus and marrow in
C57BL mice in vivo. Cancer Lett, 33: 75-81.

Taub, R., Moulding, C., Battey, J., Murphy, W., Vasicek, T., Lenoir, GM, and Leder, P.
(1984). Activation and somatic mutation of the translocated c-myc gene in Burkitt
lymphoma cells. Cell, 362339-348.

Taub, R., Kirsch, l., Monton, C., Lenoir, 6., Swan, 0., Tronick, S., Aaronson, S., and
Leder, P. (1982). Translocation of the c-myc gene into the immunoglobulin heavy chain
locus in human Burkitt’s lymphoma and murine plasmacytoma cells. Proc. Natl. Acad.
Sci. USA, 79:7837-7841.

78

Tong, L. A, DeVos, AM., Milbum, M.V., Jancarik, J., Noguchi, S., Nishimora, S., Miura,
K., Ohtsuka, E., and Kim, S-H. (1989). Structural differences between a ras oncogene
protein and the normal protein. Nature, 337290-93.

Tooze, J. (Ed): DNA tumor viruses. Cold Spring Harbor, NY, Cold Spring Harbor
Laboratory Press, 1980.

Toyoshima, K., Friis, R. R., and Vogt, P. K. (1970). The reproductive and
cell-transforming capacities of avian sarcoma virus B77: inactivation with UV light.
Virology, 42: 163-170.

Treisman, R.H., Novak, U., Favaloro, J., and Kamen, R. (1981) Transformation of rat
cells by an altered polyoma virus genome expressing only the middle T protein. Nature,
292:595-600.

Treisman, R. (1994). Ternary complex factors: growth factor regulated transcriptional
activators. Curr. Opin. Genet. Dev., 4: 96-101

Tsujimoto, Y. Finger, L.R., Yunis, J., Nowell, PC, and Croce, CM. (1984). Cloning of
the chromosome breakpoint of neoplastic B cells with the t(14;18) chromosome
translocation. Science, 226:1097-1099.

Ullrich, A, and Schlessinger, J. (1990). Signal transduction by receptors with tyrosine
kinase activity. Cell, 61:203-212.

Varmus, H. (1989). An historical overview of oncogene. In Oncogenes and Molecular
Origins of Cancer. R. A. Weinberg, ed. (New York: Cold Spring Harbor Laboratory
Press). pp.3-44.

Varmus, HE. (1984). The molecular genetics of cellular oncogenes. Annual Rev.
Genet, 18:553-612.

Vaux, D.L., Cory, S., and Adams, J.M. (1988). Bel-2 gene promotes haemopoietic cell
survival and cooperates with c-myc to immortalize pre-B cells. Nature, 335:440-442.

Veale, D., Ashcroft, T., Marsh, 0., Gibson, G.J., and Harris, AL. (1987). Epidermal
growth factor receptors in non-small cell lung cancer. Brit. J. Cancer, 55:513-516.

Vega, Q.C., Cochet, C.C., Filhol, 0., Chang, C-P., Rhee, S.G., and Gill, G.N. (1992). A
site of tyrosine phosphorylation in the C terminus of the epidermal growth factor receptor
is required to activate phospholipase C. Mol. Cell Biol., 122128-135.

Vogelstein, B., Fearon, E.R., Hamilton, 8., Kern, E., Preislnger, AC., Leppert, M.,
Nakamura, Y., White, R., Smits, AM.M., and Bos, J.L. (1988). Genetic alterations
during colorectal tumor development. N. Engl. J. Med., 319:525-532.

Vogt, PK (1994). Oncogenic transformation by Jun. In The Fos and Jun families of
transcription factors. Angel, PE, and Hrrrlich, P.A, Eds., (CRC Press, Inc., Florida),
pp.203-219.

79

Vojta, P.J., and Barrett, JC. (1995). Genetic analysis of cellular senescence. Biochim.
Biophy. Acta, 1242229-41.

Volm, M., Efferth, T., Mattem, J., and Wodrich, W. (1992). Overexpression of c-fos and
c-erbB-1 encoded proteins in squamous cell carcinomas of the lung of smokers. Int. J.
Oncol., 1269-71.

Waga, S., Hannon, G.J., Beach, 0., and Stillman, B. (1994). The p21 cyclin-dependent
kinase inhibitor directly controls DNA replication via interaction with PCNA. Nature,
369:574-578.

Wagner, A.J., Kokontis, J.M., and Hey, N. (1994). Myc-mediated apoptosis requires
wild type p53 in a manner independent of cell cycle arrest and the abilit6y of p53 to
induce p21‘“"’°'°‘- Genes Dev., 822817-2830.

Walker, RA, and Cowl, J. (1991). The expression of c-fos protein in human breast. J.
Pathol., 1632323-327.

Wang, T.C., Cardiff, R.D., Zukerberg, L., Lees, E., Arnold, A, and Schmidt, E.V. (1994).
Mammary hyperplasia and carcinoma in MMTV-cyclin D1 transgenic mice. Nature,
369:669-671.

Wang, Z.Q., Grigoriadis, AE., Mohle-Steinlein, U., and Wagner, E.F. (1991). A novel
target cell for c-fos-induced oncogenesis: development of chondrogenic tumors in
embryonic stem cell chimeras. EMBO. J., 10: 2437-2450.

Waterfield, M.D., Scrace, G.T., Whittle, N., Stoobant, P., Johnson, A, Wasteson, A,
Westermark, B., Heldin, C-H., Huang, J.S., and Deuel, TE (1983) Platelet-derived
growth factor is structurally related to the putative transforming protein p28sis of simian
sarcoma virus. Nature, 304235-39.

Weber, J.L. (1990). Human DNA polymorphisms based on length variations in simple-
sequence tandem repeats. Genome Anal., 12159-181.

Weinberg, RA (1995) The retinoblastoma protein and cell cycle control. Cell, 81:323-
330.

Weiner, D.B., Liu, J., Cohen, J.A., Vifilliams, W.V., and Greene, M. (1989). A point
mutation in the neu oncogene mimics ligand induction of receptor aggregation. Nature,
3392230231.

Weiss, W., Teich, N., Varmus, H., and Coffin, J. (Eds): RNA tumor viruses. Cold Spring
Harbor, NY, Cold Spring Harbor Laboratory Press, 1982.

Wicking, C., and Williams, B. (1991). From linked marker to gene. Trends Genet,
72288-293.

Wodrich, W., and Volm, M. (1993). Overexpression of oncoproteins in non-small cell
lung carcinomas of smokers. Carcinogenesis, 14:1121-1124.

80

Woolford, J.W., McAuliffe, A. and Rohrschneider, LR. (1988). Activation of the feline c-
fms proto-oncogene: multiple alterations are required to generate a fully transformed
phenotype. Cell, 55:965-977.

Wright, W.E., and Shay, J.W. (1992). The two-stage mechanism controlling cellular
senescence and immortalization. Exp. Gerontol., 27:383-389.

Wu, X., and Levine, A.J. (1994). p53 and E2F-1 cooperate to mediate apoptosis. Proc.
Natl. Acad. Sci. USA, 91:3602-3606.

Xiong, Y., Menninger, J., Beach, D. And Ward, DC. (1992). Molecular cloning and
chromosomal mapping of CCND genes encoding human D-type cyclins. Genomics,
13:575-584.

Xiong, Y., Hannon, G.J., Zhang, H., Casso, D., Kobayashi, R., and Beach, D. (1993).
p21 is a universal inhibitor of cyclin kinases. Nature, 366:701-704.

Young, D., Waitches, G., Birchmeier, C., Fasano, O., and \Mgler, M. (1986). Isolation
and characterization of a new cellular oncogene encoding a protein with multiple
potential transmembrane domains. Cell, 452 711-719.

Yaraden, Y., and Ullrich, A. (1988). Growth factor receptor tyrosine kinase. Annu. Rev.
Biochem., 57:443-478.

Yates, J.W., and Connor, J.M. (1986). Genetic Linkage. Brit. J. Hosp. Med, 133-136.

Yunis, J.J., and Ransay, N. (1978). Retinoblastoma and subband deletion of
chromosome 13. Am. J. Dis. Child, 132:161-163.

Yunis, J.J., Oken, N., Kaplan, M.E., Ensrud, K.M., Howe, K.R., and Theoligides, A.
(1982). Distinctive chromosomal abnormalities in histologic subtypes of non-Hodgkin’s
lymphomas. N. Engl. J. Med, 307:1231-1236.

Yuspa, SH, and Dlugosz, A A (1991). Cutaneous carcinogenesis: natural and
experimental. In: LA. Goldsmith (ed), Physiology biochemistry and molecular biology of
the skin., pp. 1365-1402. New York: Oxford University Press.

Yuspa, SH. (1994). The pathogenesis of squamous cell cancer: Lessons learned from
studies of skin carcinogenesis-Thirty-third G.H.A Clowes memorial award lecture.
Cancer Res., 54:1178-1189.

Zur Hausen, H. (1991). Viruses in human cancer. Science, 254:1167-1173.

CHAPTER II
Suppression of Anchorage-independent Growth and Matrigel Invasion and
Delayed Tumor Formation by Elevated Expression of Fibulin-1D

in Human Fibrosarcoma-derived Cell Lines

Jing Qing, Veronica M. Maher, Huan Tran,

W. Scott Argraves, Robert W. Dunstan, J. Justin McCormick

Carcinogenesis Laboratory-Fee Hall
Department of Biochemistry and Department of Microbiology
The Cancer Center

Michigan State University, East Lansing, MI 48824

81

82
ABSTRACT

Using differential display, we identified an mRNA that is markedly down-regulated in
cell line 6AISB1, derived from a fibrosarcoma formed in an athymic mouse following
injection of carcinogen-transformed MSU-1.1 cells. The nontumorigenic parental cell
strain, MSU-1.1, expresses high levels of this mRNA. Sequencing of the corresponding
cDNA fragment revealed that it corresponded to an expressed sequence tag, which
ultimately led to its identification as the fibula-JD gene. Fibulin-1 is a cysteine-rich,
calcium-binding extracellular matrix and plasma protein, which has four isoforms A-D
derived from alternative splicing. Northern and Western blotting analysis of 16 cell lines
established from tumors formed in athymic mice by MSU-1.1-derived cell strains
independently transformed in culture showed that 44% exhibited low level or lack of
expression of fibulin-1D mRNA and protein. In a similar analysis of 15 malignant cell
lines derived from patients, 80% showed low level or no expression. To study the role of
fibulin-1D in transformation, we transfected 6AISB1 cells and a human fibrosarcoma-
derived cell line (SHAC) with a fibulin-1D cDNA expression construct. Transfectants
displaying high levels of fibulin-1D were isolated and characterized. Elevated expression
of fibulin-1D led to reduced ability to form colonies in soft agar and reduced invasive
potential as tested in a matrigel in vitro invasion assay. Furthermore, expression of
fibulin-1D resulted in a markedly extended latency in tumor formation in athymic mice.
These results indicate that low expression of fibulin-1D plays a role in tumor formation

and invasion.

83
INTRODUCTION

Extensive research over the past two decades has established that cancer in humans
arises from a cell that has acquired multiple genetic alterations in oncogenes and/or
tumor suppressor genes (Weinberg, 1989; Fearon and Vogelstein, 1990; Hunter,
1991). Mutations and/or altered expression of these critical genes and their
downstream effector genes allow a cell to escape from normal growth control and
become malignant. A direct comparison of gene expression between non-tumorigenic
and tumorigenic cells of the same genetic background should provide information on
genetic changes involved in the transformation process. Subsequent cloning and
functional analysis of these differentially expressed genes may enrich our understanding
of cancer development.

To study the molecular mechanisms underlying neoplastic transformation of human
cells, we used the near-diploid, karyotypically stable, infinite life span human fibroblast
cell strain MSU-1.1 as a model system. Cell strain MSU-1.1 was established in this
laboratory from a normal diploid human neonatal foreskin fibroblast cell line (Morgan et
al., 1991). The cells are phenotypically normal and do not form tumors in athymic mice.
Transfection of MSU-1.1 cells with an activated ras oncogene expressed at high levels
(Hurlin et al., 1989; Vifllson et al., 1990) or treatment of the cells with chemical
carcinogens, such as (:)-7[3,8a-dihydroxy-9a,10a-epoxy-7,8,9,10-
tetrahydrobenzo[a]pyrene (BPDE) followed by selection for focus-forming cells (Yang at
al., 1992) results in transformants capable of tumor formation in athymic mice.
However, the molecular basis of the neoplastic conversion in MSU-1.1 cells by
treatment with chemical carcinogens is not understood. To identify putative
oncogene(s) and tumor suppressor gene(s), as well as other types of genes associated

with malignant transformation, we used differential mRNA display (Liang and Pardee,

84
1992; Liang at al., 1993) to compare the profile of gene expression in parental cell strain

MSU-1.1 and cell line 6AISB1. The latter was derived from a fibrosarcoma formed in an
athymic mouse by a BPDE-transformed MSU-1.1 cell strain. One of the differentially
expressed genes identified corresponds to fibulin-1 D, an extracellular matrix (ECM) and
plasma protein (Argraves et al., 1990, Tran e1 al., 1997). It is expressed at high levels in
MSU-1.1 cells and is markedly down-regulated in 6AISB1 cells. Analysis of 16 cell lines
established from tumors formed in athymic mice by MSU-1.1-derived cell strains
transformed in culture by various agents showed that seven (44%) had a low level or
complete lack of expression of fibulin-1D mRNA and protein. Similarly, analysis of 15
cell lines derived from tumors taken from patients showed that 12 (80%) had a low or
complete lack of expression of this mRNA and protein.

To determine if low expression of fibulin-1D were causally involved in the malignant
transformation of these cell lines, we studied the effect of stably expressing fibulin-1 D in
human fibrosarcoma-derived cell line SHAC and in 6AISB1ceIIs, which lack the
expression of this protein. The results showed that elevated expression of fibulin-1D
significantly decreased the anchorage-independent growth of both SHAC and 6AISB1
cell derivatives, as well as their invasiveness when tested in a matrigel in vitro invasion
assay, and significantly delayed the onset of tumor formation in athymic mice by these

cell lines.

85
RESULTS

To identify transformation-related genes, we carried out differential mRNA display
analysis, comparing the non-tumorigenic parental MSU-1.1 cell strain and a tumorigenic
derivative cell line, 6AISB1. RT-PCR reactions were carried out using 80 different
combinations of primer sets, composed of four degenerate anchored oligo(dT) primers,
T12MN (M was dG,dC or dA; N was dG, dC, dA or dT) and 20 arbitrary 10-mers (OPA 1-
20, Operon Technologies, Alameda, CA). About 8,000 bands were observed. The band
pattern for cell line 6AISB1 differed from that of parental MSU-1.1 in nine bands, each of
which was down-regulated in 6AISB1 cells. These cDNAs were excised from the
sequencing gel, reamplified by PCR, and used as probes for Northern blot analysis.
Such analysis showed that only five of these nine differentially-displayed cDNAs
represented mRNAs that actually are expressed at significantly lower levels in the
6AISB1 cells as compared with the MSU-1.1 cells. One of the five unique bands,
designated SG7, is characterized in this report. As shown in Figure 1B, this cDNA
hybridized to a 2.7-kb transcript showing a 14-fold lower level of expression in 6AISB1
cells than in MSU-1.1 cells. Cloning and sequencing of this partial cDNA indicated that
it consisted of 311 base pairs and was 98.7% identical to an expressed sequence tag
(Genbank accession No. T19384).

To get the full-length cDNA corresponding to SG7, EST clone T19384 was obtained
from the investigator who reported it, Dr. Georges Guellaen (Hopital Henri Mondor,
Creteil, France), and the 5' and the 3' regions of the EST clone were sequenced. A
database search revealed that the EST corresponded to the human fibula-10 gene,
with 97.4% identity in a 312-bp overlap in the 5' region (bases 1-312 of fibulin-1 D) and

98.9% identity in a 188-bp overlap in the 3' region (bases 2172-2359 of fibulin-1D). The

86
SG7 cDNA that we isolated by differential display contained a sequence that extended

beyond the 3' sequence of the fibulin-1 D cDNA sequence deposited in the database.

To confirm the identity of SG7 as human flbulin-1D, we carried out Northern blot
analysis using the 2.3-kb human fibulin-1D cDNA insert from plasmid
pBluescriptSKfibulin1 D (Tran et al., 1997) to probe the same membrane previously used
to detect SG7 expression. As shown in Figure 10, the fibulin-1 D cDNA probe detected
the 2.7-kb mRNA transcript that had been detected by SG7 (Fig.1B), but it also detected
a 2.3-kb mRNA transcript which corresponds to fibulin-1 C. This is consistent with the
fact that these two mRNAs have identical 5' regions (bases 1-1707). Both transcripts
exhibited coordinately decreased expression in 6AISB1 cells compared with the parental
MSU-1.1 cells.

A. _ .H .4.. .. pf.- ' . “h . _ ”h ,... .- .I_

To test whether the decreased expression of fibulin-1 D transcript is a common feature
of human tumor cells, we carried out Northern analysis using fibulin-1D cDNA as a
probe with RNA from 15 cell lines derived from malignant tumors taken from patients
and 16 cell lines established from tumors produced in athymic mice by MSU-1.1-derived
cell strains transformed in culture by oncogene transfection or carcinogen treatment.
The origin of these cell lines is listed in Table 1. Examples of results are shown in
Figure 2. Very low levels of expression of fibulin-1 were found in five out of five
fibrosarcoma-derived cell lines (lanes 15-19), in two out of two osteosarcoma-derived
cells (lanes 8 and 9), in two out of three neurofibrosarcoma-derived cells (lanes 12 and
20), in one out of two human bladder carcinoma-derived cells (lane 11), and in the
rhabdomyosarcoma- and cervical carcinoma-derived cell lines (lanes 7 and 10) tested in
this study. Downregulation of fibulin-1 mRNA expression was also found in 6 out of 16

MSU-1.1 lineage-derived tumor cells (lanes 3-6, and 22-23), in addition to cell strain

87
6A/SBl, in which the downregulation was first identified (Fig. 1). Taken together, 19 out

of 31 tumor-derived cell lines exhibited low or no expression of fibulin-1 (C and D
transcripts).

To determine whether mRNA levels correlate with fibulin-1 protein expression, we
performed Western blot analysis of cell extracts and their serum-free conditioned culture
media. As a control, purified fibulin-1 protein from human placenta was included in
each assay. As shown in Figure 3, lane 1, under non-reducing conditions, mouse-anti-
human fibulin-1 monoclonal antibody 3A11 specifically recognized the fibulin-1 protein
(apparent molecular weight, 80 kDa). We found that, consistent with results from
Northern analysis, the expression of fibulin-1 protein was very low or undetectable in a
panel of tumor cell lines, not only in the cell extracts (Fig.3A) but also in their
conditioned culture media (Fig.3B), whereas a neonatal foreskin-derived human
fibroblast line, LG1, and MSU-1.1 cells expressed high levels of fibulin-1 proteins. The
presence of multiple immunoreactive polypeptides that migrated faster than the mature
fibulin-1 polypeptide on SOS-PAGE (Fig.3A) may correspond to proteolytic fragments of
fibulin-1. Similar fragments have been observed in purified fibulin-1 preparations
derived from placenta and lung and have been generated in vitro by matrilysin and
leucocyte elastase digestion (Sasaki et a], 1996).

._.'...-. . .; .,.._.A : ..- ._. -,.- H...” - - . .._'.e
DLQIQ'IIJ

The levels of flbulin-1 protein in either cell extracts or conditioned media (Fig.3) from
fibrosarcoma-derived cell lines SHAC and 6AISB1 is almost undetectable. To examine
the effect of fibulin-1D expression on malignant transformation, we constructed a human
fibulin-1D cDNA expression vector, designated pPuro-Fibulin-1D, and transfected it into

SHAC and 6AISB1 cells. Puromycin resistant clones were isolated from cells that were

88

Fig.1. (A) Differential display comparing mRNAs from the parental MSU-1.1 cell strain
(lanes 1 and 2) and its malignant derivative cell line 6AISB1 (lane 3 and 4). For each
cell line, two independent preparations of total RNA were extracted, reverse
transcribed, and amplified by PCR in the presence of [oz-35$]dATP. The PCR products
were separated on a 6% polyacrylamide gel and autoradiographed. The signal
demonstrating altered expression is marked by an arrow. Primers used were T12MG
and OPA7. (B) Northern blot analysis confirming differential gene expression for
fragment SG7 using the cloned cDNA fragment as a probe. Total RNA (15 pg) obtained
from MSU-1.1 cells (lane 1) and 6AISB1 cells (lane 2) was subjected to Northern
analysis as detailed in “Materials and Methods”. The blot was stripped and reprobed
with a GAPDH cDNA (lower panel) as the loading control. (C) Northern blot analysis
using fibulin-1 D cDNA as a probe. The same blot of (B) was stripped and hybridized with
a radiolabeled 2.3-kb human fibulin-ID cDNA insert from the pBluescriptSKfibulin-1D

plasmid.

 

 

89

+

4. . _ .2 2 mi;
3 - _.._ .
23.no- Boga
1.;3 8.;

A

2 u

.

Fig.1

90
the result of independent transfection events, i.e., from independent dishes.

lmmunoblot analysis of whole cell extracts was used to identify clones that expressed
high levels of fibulin-1D protein. For SHAC cells, a total of 24 clonally-derived, drug-
resistant populations were tested. Five independent clones exhibited high levels of
fibulin-1D protein. Two clones of SHAC cells that had been transfected with the control
vector pBABE-Puro and displayed no detectable fibulin-1D protein were used as
negative controls. The same procedure was carried out with the 6AISB1 cells. Of 45
puromycin-resistant clones assayed, 10 showed high expression of fibulin-1D. Six
independent positive clones and three vector-transfected negative control clones were
selected at random for further characterization.

We also evaluated the level of flbulin-1 secreted into the conditioned culture medium
by these cell strains. As expected, neither the parental cells nor the vector-transfected
control clones secreted fibulin-1. In contrast, the transfected clones that expressed high
levels of cellular fibulin-1D secreted relatively high levels of mature proteins into the
medium (Fig.4).

;. ...,._--'.._-.-..,-. eeule .- ..- .- -,.- pf..-

Since most malignantly transformed cells, including SHAC and 6AISB1 cells exhibit
anchorage-independent growth, i.e., do not require attachment to a substratum for
proliferation, whereas normal human fibroblasts and MSU-1.1 cells do, we tested
whether fibulin-1D expression affects the ability of SHAC and 6AISB1 cells to form
colonies in agarose. As shown in Figure 5 and Table 2, all the transfectant clones
expressing high levels of fibulin-1D exhibited a reduced ability to grow in agarose
compared to their parental cells and their vector-transfected control cell strains. The
SHAC cells were counted after 21 days; the 6AISB1 cells after 14 days. A substantial

decrease in the size of colonies was observed for SHAC cell transfectants. A moderate

91
Table 1 Human tumor-derived cell lines tested for fibulin-1 expression

 

Cell lines3

 

Origin of the cell lines

 

L451lB5T, L46l/7T, L551l3T, 2FTfl'1
ZC1IST1, 6AISB1, 11C/SB1
MSU-1.1-y1- 2A1/T, MSU-1.1- y4-2A/SF1"
MSU-1.2-y1/SF1, msu-1.2-yz/sr1

2MT, 3MT, 178-2DT, DW5T,1.1-sisB/ST2
wsu1rr1, WSU17/T1

NCI, VIPzFT, SHAC, HT1080, 8387°
143 BTK, TE85

RD

Hela

GM03808“

T24, A1698“

NF1.90.8/FT1'

Derived from

spontaneously transformed MSU-1.1cells
Derived from

BPDE-transformed MSU-1.1 cells
Derived from

6°Co-transformed MSU-1.1 cells
Derived from

”Co-transformed MSU-1.2 cells
Derived from
oncogene-transfonned MSU-1.1 cells
Derived from

human neurofibrosarcomas
Derived from

human fibrosarcomas

Derived from

human osteosarcomas

Derived from

a human rhabdomyosarcoma
Derived from

a human cervical carcinoma
Derived from

a Wilms' tumor

Derived from

human bladder carcinomas
Derived from

a human neurofibrosarcoma

 

Cell lines in the first five lines of the table were established in this laboratory from

sarcomas generated in athymic mice by injection of transformed MSU-1.1 cells; those on
line six were established in this laboratory from human neurofibrosarcomas. The rest,
with the exception of 8387, GM03803, A1698, and NF1.90.8/FT1, were obtained from
American Type Culture Collection, Rockville, MD.

Referred to as MSU-1.1-y1 in Figure 2.

‘OQOU'

From Dr. Stuart A. Aaronson, National Cancer Institute, Bethesda, MD.
From Coriell Institute for Medical Research, Camden, NJ.

From Dr. 0. M. Pereira-Smith, Baylor College of Medicine, Houston, TX.
From Dr. Thomas Glover, University of Michigan, Ann Arbor, MI.

92

Fig. 2. Northern blot analysis of fibulin-I expression in multiple human tumor cell lines.
Total RNA (15 pg), obtained from a series of cell lines, was hybridized with fibulin-1D
cDNA as in Figure 1C. These were: fibroblast cell lines from normal donors (lanes 1
and 2); malignant cell lines derived from tumors formed in athymic mice after injection of
MSU-1.1 cells transformed by a transfected oncogene (lanes 3-5) or spontaneously
transformed (lane 6); malignant cell line derived from a patient’s rhabdosarcoma (lane
7), osteosarcoma (lanes 8 and 9), cervical carcinoma (lane 10), bladder carcinoma (lane
11), neurofibrosarcoma (lane 12); the normal cell line from which MSU-1.1 cells are
derived (lane 13); parental MSU-1.1 cells (lane 14); malignant cell lines derived from a
patient's fibrosarcoma (lanes 15-19), or neurofibrosarcoma (lane 20); and malignant cell
lines derived from athymic mice tumors formed after injection of cells transformed by
cobalt 60, viz., MSU-1.1 cells (lane 21), MSU-1.2 cells (lanes 22 and 23). The KD1 cell
line was originally obtained from the late Dr. Takeo Kakunaga. The other cell lines are

described in Table 1. (B) RNA loading evaluated by a GAPDH cDNA probe.

 

«or.

mNNNPNONGmehwoFmFQFMF NFFFOF m m h o m w m NF

gifieeommeJ-Jmcte-eeo-eno-

 

 

9 Essen/i I I s. .-.
Dr .. : .- r e.! i _

Bessel

H A W1Mi. I18 us EX

swswummmmsmSflwmflammmmmm
.nnmflnwzomi... .n m 19 Mm...
till .I M 1
seems .. m. ._
uuui

94
reduction in the size of colonies was seen for 6AISB1 cell transfectants. These results

indicate that elevated expression of fibulin-1D can partially suppress anchorage
independence in the flbrosarcoma-derived cell lines tested.

To analyze the invasive potential of the fibulin-1D transfectants, we used a modified
Boyden Chamber Matrigel assay. As shown in Figure 6, the ability of the fibulin-1D
expressing transfectants to reach the bottom of the filters through the matrigel was
reduced by 50-60% compared to that of parental cells and the vector-transfected control
cell strains of both SHAC and 6AISB1 cells.

I . | ...| [llllll'-ID- . I

To determine whether restoration of the expression of the fibulin-1D gene would also
suppress in vivo tumor growth, we inoculated the fibulin-1D-expressing transfectant cell
strains, their nontransfected parental cell lines, and vector-transfected controls into
athymic mice and monitored the animals for growth of tumors. As shown in Figure 7,
the latency of tumor formation for the fibulin-1D-expressing clones was longer than that
of the parental cells and the vector-transfected control cell strains. For SHAC cells and
their vector-transfected control cell strains, the tumor volume reached 500 mm3 after an
average of 25 days postinjection; whereas for fibulin-1D-transfected cells, the average
time was 55 days. For 6AISB1 cells and their vector-transfected control strains, with
the exception of one, the tumor volume reached 500 mm3 after an average of 20 days
postinjection; whereas the average time for four fibulin-1D-expressing clones was 48
days, and the other two fibulin-1D-expressing clones failed to form tumors in three

months.

i 'I I'll |' [HI l'-Il'.2| IE I l

95

Fig. 3. Western blot analysis of the expression and secretion of fibulin-1 by normal cells
(LG1 and MSU-1.1) and 10 tumor-derived cell lines under non-reducing conditions.
Cellular protein (20 pg) (A) and 50 pl of the medium conditioned by the same cell lines
(B) were analyzed by electrophoresis in 10% polyacrylamide gel and probed with a
monoclonal antihuman fibulin-1 antibody 3A11. The positive control (lane 1) was fibulin—
1 protein isolated and purified from human placenta; lanes 4-8 are from malignant cell
lines derived from human fibrosarcomas from patients; lane 9 from a
rhabdomyosarcoma; lane 10 from an osteosarcoma; lanes 11-13 are from malignant cell
lines derived from tumors in athymic mice from carcinogen-transformed MSU-1.1 cell

strains.

 

96

—80 KD

I-:ISII»l 'Z'I
1W8
I-BSN9
9831.

G H

080 LLH
L888
OVI-IS
szlA
ION
l'I-‘nSW
I91

I- uII"‘I!:I

 

- “'80 KD

Fig.3

97
If the transfected cells expressing fibulin-1 D grew more slowly than the parental cells,

this could explain the biological data described above. To determine the doubling time
of each of the fibrosarcoma—derived cell lines, the parental cells, two vector control
transfectants, and three fibulin-1D-expressing clones were allowed to grow exponentially
and the number of cells was determined at several time points (Table 2). For SHAC cell
line and its clonal derivatives, there was no difference in the growth rate of the vector-
transfected control cell strains and that of the fibulin-1D-expressing cell strains. For cell
line 6AISB1 and its clonal derivatives, introduction of fibulin-1D had no consistent effect
on the growth rate of these cells. The three fibulin-1D expressing clones grew
somewhat more slowly (23.8-25.8 h per doubling) than the parental cells (20.2 h). One
of the vector control clones, CON3, also grew at this slower rate (25.7 h), but did not
exhibit a longer than normal tumor latency period. Therefore, the lengthened latency
period seen in Figure 7 cannot be accounted for by a slower growth rate of the fibulin-

1D-transfectant clones.

98

Fig. 4. Western blot analysis of the expression and secretion of flbulin-1D protein by
Fibulin-1D transfectants. Whole cell lysates and conditioned medium from (A) LG1ceIIs
and 6AISB1 cells, three vector-transfected controls (CON1, CON2 and CON3) and six
fibulin-1D-expressing transfectants of 6AISB1 cells and (B) SHAC cells, two vector-
transfected controls (8a1 and 8b2) and five fibulin-1D-expressing transfectants of SHAC
cells were assayed. Cellular protein (20 pg) (upper panel) and 50 pl of the media
conditioned by these cell lines (lower panel) were subjected to immunoblotting analysis
as described in Figure 3. An arrow indicates the dimer of fibulin-1D, and an arrowhead

indicates the monomer.

 

 

99

 

as
“E
2P9
909
£39
398
um
ONHS

4.9“.

v 8...-

all
QOI

GB

99
Oz
92
SNOO
ZNOO
INOO
ISSNB

l91

iloxpm

.onom

.leoww

.IDXmBN

100

Fig. 5. Colony formation by fibulin-ID transfected cell strains in 0.33% agarose. Cell
strains used were: 6AISB1 cells (A), its vector-transfected control clone CON3 (B) and
its fibulin-1D-expressing clone 2b (C) and clone 11a (D), SHAC cells (E), its vector-
transfected control clone 8a1(F) and its flbulin-1D-expressing clone 5a3 (G) and clone
5C4 (H). For each cell strain 2 x 104 cells were plated in agarose at 5,000 cells per 60
mm-diameter dish. After 2 weeks (AD) or 3 weeks (E-G) of incubation, a
photomicrograph of a representative field was taken from each dish. Representative

examples are shown.

IOI

 

Fig.5

102

Table 2 Growth properties of parental cells, cells expressing a transfected-fibulin-1 D

gene, and vector-transfected control cells

 

 

 

Cell Number of colonies of Population Cell No- Of COIONIBS Population
Strains a given size per 5,000 doubling strains with a diameter doubling
or lines cells assayed in agarose time (h) or lines 2 100pm per time (h)

mean 1 SD 5,000 cells mean 1 SD
assayed
Diameter Diameter
2 150 pm _>_ 200 pm
6AISB1 1250 800 16.7 1 2.1 SHAC 1000 18.4 1 0.5
CON1 1440 730 16.6 1 1.6 8a1 1280 23.2 1 2.1
CON2 1300 700 ND‘ 8b2 1920 22.9 1 2.2
CON3 930 740 23.5 1 4.7 5a3 30 22.1 1 1.1
2b 130 0 27.3 1 4.6 504 0 19.5 1 1.8
2c 100 0 ND 5d2 30 22.6 1 2.0
5c 60 0 ND 5e1 60 ND
8b 990 290 24.1 1 2.4 5f3 0 ND
10b 210 0 ND
11a 30 0 23.2 1 2.2

 

 

‘ Not determined.

103
DISCUSSION

In the present study, using mRNA differentially display, we found that the low
expression of an extracellular matrix and glycoplasma protein, fibulin-1 D, is associated
with malignant transformation. High levels of fibulin-1 expression have been reported in
a wide spectrum of normal human tissues and organs (Zhang et al., 1994). In the
experiments described here, we assayed 15 tumor cell lines derived from human cancer
patients and found that 12 cell lines showed very low levels of fibulin-1 expression.
Since the normal cells from which these tumor cells were derived are not available, we
cannot be certain that the low expression is the result of downregulation. However, we
observed that seven out of 16 MSU-1.1-derived tumor cell lines exhibited markedly
reduced levels of fibulin-1 expression compared with the parental MSU-1.1 cells,
indicating that down regulation of fibulin-1 expression occurs relatively frequently in the
transformation of human fibroblasts in culture. Our data also showed that fibulin-1D
downregulation is not specifically correlated with the transformation induced by
particular oncogenes or by specific carcinogens. Southern blotting analysis revealed
that the downregulation of fibulin-1D in the MSU-1.1 cell strain derivatives cannot be
attributed to gross genetic deletions or rearrangements (data not shown).

Fibulin-1, together with fibulin-2, belongs to a family of extracellular matrix proteins
(Pan 9} al., 1993a; Tran e1 al., 1997). Alternative splicing of fibulin-1 precursor RNA
results in four transcripts that encode polypeptides differing from each other only at the
carboxyl terminal regions. The four isoforms are designated AD. The dominant forms
expressed in most human tissues and cell lines in culture are fibulin-1C and fibulin-1D
(Zhang etaL, 1994; Tran et al., 1997). Fibulin-1 proteins are multimodular proteins
containing three cysteine-rich anaphylatoxin-related segments, nine epidermal growth

factor-like repeats with a consensus motif for calcium-binding, and a variable carboxyl-

104

Fig. 6. In vitro invasiveness of fibulin-1D transfected cell strains in matrigel assays.
The invasive index for each cell strain was determined from the number of cells that
passed through the matrigel inserts as a fraction of the number of cells that passed
through the control uncoated filters. The invasion index of SHAC and 6AISB1 cells was

normalized to 100.

105

 

 

 

 

:o_mo>:_ Egan:

103 11A 23

SC

BA CON1 CON2 CON3 2C

 

 

 

 

:o.»u>:_ .3322

513

362 5.3 504 5d2 5.1

SHAC 8'1

6

Fig

106
terminal domain (Argraves et al., 1990; Pan et al., 1993b). Biochemical studies have

shown that fibulin-1 interacts with itself and binds other extracellular matrix proteins,
including fibronectin (Balbona et al., 1992), laminin, and nidogen (Pan e1 al., 1993b).
lmmunohistochemical studies have shown that fibulin-1 proteins are widely expressed
intercellular components of connective tissues which are present in matrix fibers and
basement membranes (Roark at al., 1995). However, until now little was known about
fibulin-1 playing a role in the process of transformation.

Since we found that the expression of fibulin-1 was low in many tumor-derived cell
lines, we sought to determine whether this low expression of fibulin-l is merely a
consequence of malignant transformation, or whether it plays a role in the malignant
transformation process. Our results with a transfected fibulin-1 D gene demonstrate that
in each case (11 out of 11), elevated expression of fibulin-1 can suppress the ability of
fibrosarcoma-derived cells to proliferate in soft agar and to delay the tumorigenicity in
vivo. The mechanisms underlying the observed effects have not yet been elucidated.
However, a number of studies regarding the role of other ECM molecules in
transformation indicate that there is a good correlation between the loss of specific ECM
molecules and the acquisition of transformed phenotypes in vitro and tumorigenicity in
vivo. For instance, hybrid cells formed by fusion of mouse melanoma cells and normal
mouse fibroblasts are not malignant. However, inhibition of fibronectin synthesis in such
hybrids by an antisense strategy causes the cells to become tumorigenic (Steel and
Harris, 1989). Conversely, addition of purified fibronectin to Herpes simplex virus-
transformed hamster fibroblasts can restore normal morphology and adhesive
properties characteristic of normal cells (All at al., 1977). Overexpression of fibronectin
in human fibrosarcoma-derived HT1080 cells with increased deposition of fibronectin on

the cell surface leads to reduced anchorage-independent growth and tumorigenicity

107

Fig. 7. Tumorigenecity of fibulin-1D transfected cell strains in athymic mice. The cell
strains used were: SHAC cells (A), 6AISB1 cells (B), their control vector or fibulin-1D
transfectants. For each cell line, two mice (total four sites) were injected with 1 x 10°
cells at each site. Each symbol represents the volume averaged from four tumors from
two mice: parental cell lines (-), control vector transfectants (other closed symbols),

and fibulin-1Dtransfectants (open symbols).

 

Average volume of tumors (mma)

Average volume of tumors (mm’)

1 200

108

 

1000 a

800-

600-

400-

200-

 

0

 

    

 

 

' ‘ " T r r

404448

Days after injection (day)

 

 

 

O
V
V
O
, . , e
/ l O ‘ A
I O
é!”t’: _A A .

121620242832364044

Days after Injection (day)

Fig. 7

109
(Akamatsu e1 al., 1996). A recent study shows that expression of SPARC, an ECM

protein, in ovarian cancer cells results in a significant decrease in cancer cell growth and
tumorigenic potential (Mok et al., 1996). Collectively, these data demonstrate that
certain ECM molecules play an important role in the regulation of malignant
transformation. Perturbation in expression of these molecules could lead to global
effects by interfering with the formation and stabilization of extracellular matrix
structures, or as suggested by Juliano and Haskill (1993), with normal cell growth
signaling pathways that act through extracellular matrix.

An important conclusion that can be drawn from our results is that loss of fibulin-1D
plays a role in invasion and perhaps metastasis. One possible model is that fibulin-1D
negatively regulates the ability of cells to interact with migration-promoting ECM
components, thereby inhibiting cellular migration. The ability of fibulin-1 to interact with
several adhesive and migration-promoting proteins such as fibronectin and laminin has
been established (Balbona et al., 1992; Pan et al., 1993b). It is possible that fibulin-1
binding to these proteins interferes with the ability of cells to derive migratory stimuli.
We know that the fibulin-1 binding site in fibronectin is located within type III repeats 13
and 14 (Balbona e1 al., 1992). This region lies adjacent to the Arg-Gly-Asp (RGD)-
containing type lllm adhesive domain and the CSIII adhesive domain (Argraves and
Gehlsen, 1991). These domains mediate cellular interaction with fibronectin via a wide
array of integrins, including 0:581 binding to the RGD site and (1481 binding to the CSIII
site. It is possible that the binding of fibulin-1 to fibronectin interferes with integrin
binding to either or both of these sites. The fibulin-1 binding site with Engelbreth-Holm-
Swarm tumor laminin has been mapped to a site contained within the E3 fragment (Pan
et al., 1993b; Brown et al., 1994), a fragment derived from the carboxy terminus of the a

chain. This domain has been shown to mediate integrin 0:381 binding (Gehlsen et al.,

110
1989; Gehlsen et al., 1992) and lies in the vicinity of binding sites for the integrins (1681

(Aumailley et al., 1990; Hall et al., 1990; Sonnenberg et al., 1991) and 0:781 (Kramer et
al., 1991; von der Mark e1 al., 1991). In addition, the laminin receptor dystroglycan
binds to the E3 fragment (Gee et al., 1993). Studies are under way to determine
whether fibulin-1 modulates any of these receptor-adhesive protein interactions.

Interestingly, in contrast to our findings, Clinton at a]. (1996) recently reported that
estrogens increase the expression of fibulin-1 in human ovarian cancer cells. They
suggest that the augmented fibulin-I expression facilitates ovarian tumor cell invasion.
However, they have not yet carried out studies to test this hypothesis. It may also be
that estrogen-augmented fibulin-1 expression acts to inhibit tumor cell migration.
Clearly, additional experiments are required to determine the role that fibulin-1 plays in
the movement and growth of various types of tumor cells.

In summary, by using differential mRNA display we have identified fibulin-1 as a gene
which is downregulated in a human fibroblastic cell strain transformed in culture. We
have found that the fibulin-1D protein is expressed at a low or completely undetectable
level in a number of human fibroblastic cells transformed in culture as well as in a
number of human tumor-derived cell lines. We have demonstrated that increased
expression of fibulin-1D from a transfected gene in human fibrosarcoma-derived cell
lines reduced anchorage-independent growth and delayed tumor growth in athymic
mice. Furthermore, the invasive ability of these cells was greatly suppressed. These
findings indicate that the loss of fibulin-1D expression contributes to the transformation
and progression progress of human fibrosarcoma. This study points to the importance
of recent studies that have shown that ECM molecules can directly regulate critical
cellular process such as growth, differentiation and apoptosis (Jones et al., 1993; Lin

and Bissell, 1993; Meredith et al., 1993; Frisch and Francis, 1994). Further research in

111
various types of cells will be required to address the precise mechanism by which ECM

contributes to the transformation process.

1 12
MATERIALS AND METHODS

Cell§_and_c§ch_ulture

The infinite life span human fibroblast cell strain MSU-1.1 and its derivative cell lines
were routinely cultured in Eagle's minimum essential medium, modified by addition of L-
aspartic acid (0.2 mM), L-serine (0.2 mM) and pyruvate (1mM) and supplemented with
10% supplemented calf serum (803) (Hyclone Laboratory, Logan, UT), penicillin (100
units/ml), streptomycin (100 pg/ml) and hydrocortisone (1 pglml) (complete medium) at
37°C in a humidified incubator containing 5% 002 in air.

Qumtluate.

Growth curves were obtained by plating 2 x 104 cells per 60 mm diameter dish and
harvesting triplicate dishes on day 1, 3, 4, 5 and 7. The total number of cells per dish
was determined using an Elzone Counter. The exponential growth phase of the growth
curve was used to calculate the population doubling time.

Non-tumorigenic human fibroblast cell strain MSU-1.1 and its malignant derivative cell
line 6AISB1 were used as sources of RNA. Total RNA from cells in exponential growth
was extracted using RNAzolB (Tel-Test, Friendswook, TX). Differential mRNA display
was performed as described (Liang and Pardee, 1992; Liang 9.1 al., 1993) with slight
modifications.

Briefly, total RNA (0.2 pg) from each cell line was reverse transcribed with each of the
four degenerate oligo (dT) primers T12MA, T12MT, T12MG and T12MC (where M may be
dG, dC, or dA), followed by PCR amplification of the cDNA in the presence of [Oi-35$]
dATP (Dupont, Wilmington, DE) using the corresponding T12MN as the 3' primer and
one of the arbitrary 10-mers (OPA1-20, Operon Technologies, Alameda, CA) as the 5'

primer. The PCR thermal cycle parameters were as follows: 94°C for 30 s, 40°C for 2

113
min, and 72°C for 30 s for 40 cycles, followed by extension at 72°C for 7 min. PCR

products were separated in a 6% denaturing polyacrylamide gel. Gels were dried
without fixation and exposed to Kodak XAR film. To confirm the results, reactions
showing differentially expressed mRNAs in MSU-1.1 and 6AISB1 cells were repeated in
duplicate using two independent RNA preparations. The cDNA fragments representing
uniquely expressed mRNAs were excised from the dried gels and reamplified by PCR
using the same set of primers originally used. The PCR products were run on a 1.5%
agarose gel, and the bands of the appropriate size were cut from the gel and purified
using QlAquick gel extraction kit (Qiagen, Chatsworth, CA).

S | | . | El I? .

The purified DNA was used as probes for Northern blot analysis or subcloned into the
pCRll vector by the TA cloning method (lnvitrogen, San Diego, CA) according to
manufacturer's instructions. Subsequent identification of insert-containing clones was
carried out by the dot-blot DNA hybridization procedure described by Callard e1
al.(1994). The subcloned cDNA inserts were again used as probes for Northern
analysis to confirm the differential expression. DNA sequencing was performed directly
from the TA cloning vector with the SP6 and the T7 primer using the Fidelity DNA
sequencing system (Oncor, Gaithersburg, MD). DNA database searches were
performed using the GCG FASTA program (Genetics Computer Group, Madison, WI) or
the BLAST program from National Center for Biotechnology lnforrnation.
Nommhlotjnallsls

Total RNA (15 pg) was electrophoresed on a denaturing formaldehyde agarose gel
(1.2%), transferred to Hybond-N membrane (Amersham, Arlington Heights, IL) by the
downward capillary transfer technique (Chemczynski, 1992) using a 20 fold

concentration of standard saline citrate buffer, and fixed by UV crosslinking (UV

114
Stratalinker 2400, Stratagene, La Jolla, CA). DNA probes were radiolabeled by the

random primed labeling method (Feinberg and Vogelstein, 1983). Northern
hybridization was performed at 42°C overnight in 50% formamide containing 5 x SSPE,
5 x Denhardt’s solution, 0.1% sodium dodecyl sulphate (SDS) and 0.1 mglml salmon
sperm DNA. The blots were then washed twice in 1 x SSPE, 0.1% SDS at 42°C, and a
third wash was carried out in 0.25 x SSPE with 0.1 % SDS at 55°C. To strip the
membrane for reprobing, the blots were treated with boiling 0.1% SDS solution and
allowed to cool to room temperature. Variation in RNA loading per lane was evaluated
by probing with glyceraldehyde 3-phosphate dehydrogenase (GAPDH) cDNA as the
control.

E l' E || | | | IT I I.

Cells were grown in 100 mm-diameter dishes in Eagles’s medium containing 10%
SCS until subconfiuent. To harvest the cells, the cell monolayer was quickly washed
four times with cold phosphate buffered saline (PBS), and the cells were lysed for 20
min on ice in 0.5 ml of RIPA buffer composed of 50 mM Tris-HCI, pH 7.2, 150 mM NaCl,
1% Triton X-100, 0.1% SDS, 0.5% deoxycholic acid, 2 mM phenylmethylsulfonyl fluoride
(PMSF), 1 mM EDTA and 0.15 units/ml aprotinin. The cell lysates were collected with a
rubber scraper, centrifuged at 14,000 x g for 15 min at 4°C, and the protein
concentration of the supernatant was determined using a bicinchoninic acid protein
assay kit (Pierce Chemical, Rockford, IL) with bovine serum albumin as the protein
standard. The cell lysates were used immediately or stored at -80°C until use. To
collect conditioned medium, nearly confluent cells cultured in 60 mm-diameter dishes
(i.e., ~ 2 x 106 cells) were washed four times with Eagle’s medium, and incubated for
another 48 h in 4 ml of complete medium without serum. The conditioned medium was

collected, and PMSF (final concentration, 1 mM) and aprotinin (0.15 units/ml) were

115
added, and the medium was centrifuged at 14,000 x g for 5 min to remove cell debris. A

second duplicate dish for each sample was used to prepare cell lysate as described
above, and the protein concentration of the cell extract was used to normalize the
volume of conditioned medium used for Western blot analysis.

Aliquots of cell lysates containing 20 pg of protein or 50 pl of the conditioned medium
were mixed with the sample buffer (0.05M Tris-HCl, pH6.9, 9% glycerol, 2.3% SDS and
0.1% bromophenol blue) without a reducing agent, and the proteins were separated on
a 10% SDS/polyacrylamide gel. The proteins were then electroblotted onto an
Immobilion-P membrane (Millipore, Bedford, MA). The blots were blocked for 2 h at
room temperature in Tris-buffered saline (20 mM Tris-HCI, pH 7.6, 137 mM NaCI)
containing 0.1% (v/v) Tween 20 and 5% (w/v) non-fat dry milk (blocking solution), and
then incubated for 2 h at room temperature with the mouse monoclonal anti-fibulin-1 lgG
3A11 diluted 1:20,000 in the same solution. This antibody was generated as part of an
earlier study (Argraves at al., 1990). The blots were washed several times and then
incubated with horseradish peroxidase-conjugated goat-anti-mouse lgG (Boehringer
Mannheim, Indianapolis, IN) that had been diluted 1:3000 with blocking solution.
Enhanced chemiluminescence (Amersham, Arlington Heights, IL) was used according to
the manufacturer’s recommendations to detect the signal.

A 2.3-kb EcoRl fragment of pBluescript-Fibulin-1 D, containing the entire coding
sequence of human Fibulin-1 D, was inserted into the unique EcoRl cloning site of the
mammalian expression vector pBABE-Puro (Morgenstem and Land, 1990) under the
control of Moloney murine leukemia virus long terminal repeat, generating the plasmid

pPuro-Fibulin1D. The sense orientation of the cloned fragment was confirmed by

116
restriction mapping. To produce stably transfected cell lines, cells in exponential growth

were plated in 100-mm dishes (2 x 105 cells/dish). After 18 to 24 h, 2 pg of plasmid
DNA was transfected into the cells using lipofectamine (GIBCO, Gaithersburg, MD)
according to the manufacturer’s instructions. Transfected cells were selectively grown in
culture medium containing 0.5 pg/ml puromycin (Sigma, St. Louis, M0) for 2-3 weeks.
Drug-resistant clones were randomly selected and subcloned individually for further
study.

WW!

For each cell line or strain, 2 x 104 cells were plated at 5,000 cells per 60 mm-
diameter dish in a top layer of 0.33% agarose in McM medium containing 2% fetal calf
serum, essentially as described by Hurlin at a]. (1989). The agarose plates were
maintained at 37°C in a humidified incubator with 3% C02 in air. Each week the
medium over the top layer of agarose was removed and fresh McM medium with 2%
fetal calf serum was added. At the end of 14 days, 6AISB1 cells and their transfectants
or 21 days for SHAC cells and their transfectants, the cells were fixed with 2.5%
glutaraldehyde. Photomicrographs covering 0.7 cm2 were made of representative fields
from each dish. All colonies greater than 100 lm in diameter were scored and the size
of each colony was recorded. The frequency of colonies of a given size in this area
from each dish was averaged, and the data were expressed per 5,000 cells assayed.

The invasiveness of SHAC cells, 6AISB1 cells and their transfectants was assayed by
a modified Boyden Chamber Matrigel method (Albini at al., 1987) according to the
manufacturer’s instructions (Collaborative Biomedical Product-Becton Dickson, Bedford,
MA). The cells to be studied were washed with PBS three times and harvested by short

exposure to 5 mM EDTA. Cells were washed once with Eagle’s medium, and 7 x 105

117
cells in Eagle’s medium containing 0.1 % BSA were seeded onto the control uncoated

filters (8 pm pore size) or matrigel-coated filters in Biocoat Matrigel invasion chambers.
Eagle’s medium (2.5 ml) containing 0.1% BSA and 10 nglml platelet-derived growth
factor (Sigma) was added to the lower compartment. After incubation at 37°C for 48 h
(for 6AISB1 cells and its transfectants) or 72 h (for SHAC cells and its transfectants),
the filters were fixed in ethanol and stained with 1% crystal violet. Cells that invaded the
lower surface of the filters were counted under a light microscope. Each assay was
done in triplicate.

BALB/c athymic mice 6 weeks of age were injected with 1 X 10° cells in 0.2 ml of
serum-free Eagle’s medium subcutaneously in the left front and right hind flank regions.
Tumor dimensions were measured twice weekly using a vemier caliper, and the size of
tumors was calculated using the formula for the volume of a hemiellipsoid, the geometric
figure most nearly approximating the shape of the tumor: Volume=0.5236 X length X

width X height (Rockwell at al., 1972).

l 18
ACKNOWLEDGMENTS

We thank Dr.Georges Guellaen at Hopital Henri Mondor (Creteil, France) for kindly
providing EST clone T19384 and Dr. Richard Schwartz of Michigan State University for
kindly providing a plasmid containing the 1.3-kb GAPDH cDNA and the expression
vector pBABE-Puro. We thank our colleague Dr. Dong Wei for help with DNA
sequencing and helpful discussions during the course of this study. This research was
supported by DHHS grant CA60907 from the National Cancer Institute, AG11026 from
the National Institute of Aging (J. J. M) and GM42912 from the National Institute of

General Medicine (W. S. A).

119
REFERENCES

Akamatsu, H, lchihara-Tanaka, K, Ozono, K, Kamiike, W, Matsuda, H and Sekiguchi, K.
(1996). mm, 56, 4541-4546.

Albini, A, lwamoto, Y, Kleinman, HK, Martin, GR, Aaronson, JM, Kozlowski, JM and
McEwan, RN. (1987). W., 47, 3239-3245.

Ali, lU, Mautner, V, Lanza, R, and Hynes, R0. (1977). Cell, 11, 115-126.

Argraves, WS and Gehlsen, KR. (1991). (alive, 5, 489-492.

Argraves, WS, Tran, H, Burgess, WH and Dickerson, K. (1990). .L_Qell_B[QL, 111,
3155-3164.

Aumailley, M, Timpl, R and Sonnenberg, A. (1990). Exam, 188, 55-60.

Balbona, K, Tran, H, Godyna, S, lngham, KC, Strickland, DK and Argraves, WS.
(1992). .1._B£QL_QtLem., 267, 20120-20125.

Brown, JC, Wiedemann, H and Timpl, R. (1994). J,_C_e[L§_ci.,107, 329-338.

Callard, D, Lescure, B and Mazzolini, L. (1994). BLOIQQLIHL'QUQS, 16, 1096-1103.

Chomczynski, P. (1992). mm, 201, 134-139.

Clinton, GM, Rougeot, C, Derancourt, J, Roger, P, Defrenne, A, Godyna, S, Argraves,
WS and Rochefort, H. (1996). W, 93, 316-320.

Fearon, ER and Vogelstein, B. (1990). Cell, 61, 759-767.

Feinberg, AP and Vogelstein, B. (1983). Andi/322mm, 132, 6-13.

Frisch, S and Francis, H. (1994). J._QQILBLQL. 124, 619-626.

Gee, SH, Blacher, RW, Douville, PJ, Provost, PR, Yurchenco, PD and Carbonetto, S.
(1993). .L_BL'QL_Qh_em., 268, 14972-80.

Gehlsen, KR, Dickerson, K, Argraves, WS, Engvall, E and Ruoslahti, E. (1989). _,L_B£QL

£11601, 264, 19034-38.

120
Gehlsen, KR, Sriramarao, P, Furcht, LT and Skubitz, AP. (1992). 142215111. 117, 449-

459.

Hall, DE, Reichardt, LF, Crowley, E, Holley, B, Moezzi, H, Sonnenberg, A and Damsky,
CH. (1990). .LQQILBLQL 110, 2175-84.

Hunter, T. (1991). Qell, 64, 249-270.

Hurlin, PJ, Maher, VM and McCormick, JJ. (1989). MAM, 86, 187-
191.

Jones, PC, Schmidhauser, C and Bissell, MJ. (1993). W., 3,
137-154.

Juliano, RL and Haskill, S. (1993). LQQILBL'QL120, 577-585.

Kramer, RH, Vu, MP, Cheng, YF, Ramos, DM, Timpl, R and Waleh, N. (1991). Cell
899.111.. 2, 805-817.

Liang, P, Averboukh, L and Pardee, AB. (1993). ALL/gleam, 21, 3269-3275.

Liang, P and Pardee, AB. (1992). Science, 257, 967-969.

Lin, CG and Bissell, MJ. (1993). EASEB_J., 7, 737-743.

Meredith, JE, Fazeli, B and Schwartz, MA. (1993). MQL_B£QL_C_QL4, 953-961.

Mok, SC, Chan, WY, Wong, KK, Muto, MG and Berkowitz, RS. (1996). Qacogene, 12,
1895-1901.

Morgan, TL, Yang, D, Fry, DG, Hurlin, PJ, Kohler, SK, Maher, VM and McCormick, JJ.
(1991). mm 197, 125-136.

Morgenstem, JP and Land, H. (1990). MAM. 18, 3587-3596.

Pan, TC, Kluge, M, Zhang, R-Z, Mayer, U, Timpl, R and Chu, M-L. (1993b). EuL_,.L
Biochem., 215, 733-740.

Pan, TC, Sasaki, T, Zhang, R-Z, Fassler, R and Timpl, R. (1993a). J._QelLBi_QL, 123,

1 269-1 277.

121
Roark, EF, Keene, DR, Haudenschild, CC, Godyna, S, Little, CD and Argraves, WS.

(1995). J.W. 43, 401-411.

Rockwell, SC, Kallaman, RF and Fajardo, L. (1972). W., 49, 735-747.

Sasaki, T, Mann, K, Murphy, G, Chu, ML and Timpl, R. (1996). W, 240,
427-434.

Sonnenberg, A, Gehlsen, KR, Aumailley, M and Timpl, R. (1991). W, 197,
234-244.

Steel, DM and Harris, H. (1989). W, 93, 515-524.

Tran, H, Mattei, M, Godyna, S and Argraves, WS. (1997). Matulfliology, in press.

Von der mark, H, Durr, J, Sonnenberg, A, von der Mark, K, Deutzmann, R and
Goodman, S. L. (1991). W., 266, 23593-601.

Weinberg, RA. (1989). QanceLBes“ 49, 3713-3721.

\NilSOfI, DM, Yang, DJ, Dillberger, JE, Dietrich, SE, Maher,VM and McCormick, J J.
(1990). W, 50, 5587-5593.

Yang, D, Louden, C, Reinhold, DS, Kohler, SK, Maher, VM and McCormick, JJ. (1992).
MW 89, 2237-2241.

Zhang, RZ, Pan, TC, Zhang, ZY, Mattel, MG, Timpl, R and Chu, ML. (1994).

Genomics, 22, 425-430.

CHAPTER III
Cloning and Characterization of a Novel Gene Encoding a Putative

Transmembrane Protein with Altered Expression in Human Tumors

Jing Qing, Dong Wei, Veronica M. Maher, J. Justin McCormick

Carcinogenesis Laboratory, Fee Hall
Department of Microbiology and Department of Biochemistry
The Cancer Center

Michigan State University, East Lansing, MI 48824-1316

122

123
ABSTRACT

Identification and characterization of genes expressed in normal cells and decreased
in their malignant counterparts is an important method for detecting candidate tumor
suppressors. Using differential display of mRNAs from nontumorigenic infinite life span
human fibroblast cell strain MSU-1.1 and an isogenic fibrosarcoma-derived cell line,
6AISB1, which was derived from chemical carcinogen transformed MSU-1.1 cells, we
identified a novel gene, ST7, showing a 6-fold lower expression in 6AISB1 cells
compared with parental MSU-1.1 cells. Molecular cloning of a near full-length cDNA
revealed that the novel gene encodes a putative transmembrane protein composed of
859 amino acids: the 492 N-terminal amino acids including a 5-fold cysteine-rich repeat
of 40 amino acids homologous to the ligand binding repeat of the known low density
lipoprotein receptor, a 24 hydrophobic amino acid stretch spanning the plasma
membrane, and a C-terminal domain of 343 residues. The ST7 gene is widely
expressed in normal human tissues and is particularly abundant in human heart and
skeletal muscle. Northern analysis showed that 10 out of 15 tumor cell lines derived
from patients and 6 out of 16 cell lines established from tumors formed in athymic mice
by MSU-1.1 cells transformed in culture have low or undetectable levels of ST7 mRNA.
Furthermore, Western blotting analysis using a specific anti-peptide antibody
demonstrated that the levels of ST7 protein are high in normal fibroblasts and low in 12
sarcoma-derived cell lines tested. Altered expression of ST7 appears to occur at both
the transcriptional and posttranscriptional level. These studies make it possible to
characterize a novel putative receptor protein, whose expression is downregulated in
many malignantly transformed fibroblastic cells, and which may play an important role in

the transformation of these cells.

124
INTRODUCTION

Cancer results from the accumulation of a series of genetic and biochemical changes
in normal cells (Peto, 1977; Farber, 1984; Klein and Klein 1985; Fearon and Vogelstein,
1990). Despite the vast increase in our knowledge of oncogenes and tumor suppressor
genes associated with human neoplasia over the past 15 years, the molecular events
leading to the formation of most types of human tumors are still not well understood.
Therefore, identification of genes that are involved in the transition of non-tumorigenic
cells to malignant cells can be expected to help us understand the molecular

mechanisms underlying tumor formation.

The recent introduction of several in vitro transformation systems utilizing human cells
in culture (Stoner at al., 1991; Reznikoff e1 al., 1993; McCormick and Maher, 1994; Rhim
et al., 1994; Park et al., 1995) has facilitated the investigation of the cellular and
molecular mechanisms involved in the multistep carcinogenic process. In our
laboratory, transfection of the v-myc oncogene into a human neonatal foreskin-derived
fibroblast cell line LG1 led to the establishment of a near-diploid, karyotypically stable,
infinite life span human fibroblast cell strain MSU-1.1 (Morgan e_t al., 1991). MSU-1.1
cells are phenotypically normal and do not form tumors in athymic mice. Treatment of
MSU-1.1 cells with chemical carcinogens such as benzo(a)pyrene diol epoxide (BPDE)
(Yang et al., 1992) or gamma irradiation (Reinhold e_t al., 1996), followed by selection of
focus-forming cells, results in cells capable of forming tumors in athymic mice. Since
the focus-derived, tumorigenic cells have various properties not possessed by the
parental cells, it is presumed that additional genetic alterations induced by carcinogens
are required in the transformation process. However, the cellular genes responsible for

the neoplastic transformation induced by these carcinogens remain poorly understood.

125
The present study was carried out seeking to isolate the oncogene(s), or tumor

suppressor gene(s) that are involved in this neoplastic conversion.

The recently developed differential mRNA display method (Liang and Pardee, 1992;
Liang at al., 1993) allows one to identify genes that are differentially expressed between
closely related eukaryotic cells. In our study, we applied this method to compare the
mRNA profile between the nontumorigenic parental MSU-1.1 cells and one of its
malignant derivative cell lines, 6AISB1, which was established from a fibrosarcoma
formed in an athymic mouse by BPDE-transformed MSU-1.1 cells. We have identified a
novel gene, designated ST7, whose mRNA is markedly downregulated in 6AISB1 cells
compared with MSU-1.1 cells. A near full-length cDNA was cloned and the deduced
amino acids revealed that this novel gene encoded a putative transmembrane protein of
859 amino acids. ST7 mRNA and protein levels were low in a large fraction of the tumor
derived cell lines tested. Further characterization of ST7 should lead to better

understanding of the carcinogenesis process in human cells.

126
RESULTS

ldentlfleatien Qf $11 as a Gene Differentially Examssed BeMeen W and
Malignant Human Calls. When we compared MSU-1.1 cell strain and its tumorigenic
derivative cell line 6AISB1 using mRNA differential display, we saw 8,000-10,000
displayed cDNA fragments and found in two independent experiments that nine DNA
fragments reproducibly showed differential intensities between MSU-1.1 cells and the
tumor-derived cells (Figure 1A and data not shown). cDNA fragments were isolated
from the sequencing gel and used as probes in Northern analysis to test for differential
mRNA expression between MSU-1.1 and 6AISB1 cells. Differential expression of
mRNA was confirmed for five of the nine cDNAs. Subsequent analysis showed that four
of these five corresponded to the same gene, i.e., fibulin-1D (Qing at al., 1997). The
fifth, designated ST7, is shown in Figure 1B. This cDNA fragment hybridized with a 3.7
kb transcript showing a 6-fold lower expression in 6AISB1 cells than in MSU-1.1 cells.
The ST7 DNA fragment was subcloned, and the insert DNA from several individual
plasmids was also found to hybridize to the same sized RNA and the transcript exhibited
differential expression in MSU-1.1 cells compared with the tumor cell line. Subsequent
sequencing analysis of this partial cDNA revealed that it contained 485 hp with no

significant homology with any known genes in the nucleotide sequence databases.

Expression $21 $11 in Multiple Human Tumor Cell Lines. To test whether the
expression of ST7 mRNA is also downregulated in other tumor-derived MSU-1.1 cell
lines malignantly transformed by various methods, RNA from 15 additional such cell
lines was assayed by Northern blotting analysis. We also examined 15 tumor-derived
cell lines from patients. The origin of these cell lines has been described previously

(Qing at al., 1997). Representative data are shown in Figure 2. We found

127
downregulation of ST7 in 5 out of 15 MSU-1.1 derivatives (Lanes 4, 5, 8, 21 and 22).

We also found that in 10 out of 15 tumor-derived cell lines from patients, the expression
of ST7 was very low or undetectable, i.e., in cells from three out of five fibrosarcomas
(Lanes 10, 12-13), in cells from two out of two osteosarcomas (Lanes 6 and 16), in cells
from the three neurofibrosarcomas (Lanes 18-20), and in cells from a cervical carcinoma
(Lane 7) and a bladder carcinoma (Lane 3). Collectively, of the 31 tumor-derived cell

lines assayed, 16 exhibited either low or no expression of ST7.

Meleeular filming of the Eullzlengtn Human SIZ cDNA. To obtain the full-length
cDNA of ST7, we screened a human fibroblast cDNA library (library 9 of Legerski) with
an ST7 gene-specific primer and a vector-specific primer, using the High Fidelity Expand
polymerase chain reaction method (Boehringer Mannheim). Several specific PCR
products were obtained. We sequenced the longest one (clone A), which was about 2.6
kb (Figure 3A). To recover the 5' end of the cDNA, we screened a human skeletal
muscle cDNA library with another ST7-specific primer and a vector primer using the
PCR method described above. This produced a fragment we called clone B. To obtain
additional 5' sequence, we carried out the 5'-rapid amplification of cDNA ends (RACE)
reaction using a human heart Marathon-ready cDNA (Clontech, Palo Alto, CA) and
designated the generated fragment clone C. These three clones overlapped with each
other (Figure 3A). Two forms of the cDNA differing from each other at the 5' end were
isolated. The longer form has 57 more nucleotides than the shorter one. The assembled
nucleotide sequence of the longer form (total 3078 bp) and the deduced amino acid
sequence are shown in Figure 3B. An open reading frame extends from nucleotide 43
to 2619 and encodes a protein of 859 amino acids with an estimated molecular weight

of 92.8 kDa. The first ATG codon (nucleotides 43-45) lies in the context of the Kozak

128

Fig.1. (A) Differential display comparing mRNAs from the parental MSU-1.1cell strain
(lanes 1 and 2) and its tumorigenic derivative cell line 6AISB1 (lanes 3 and 4). For each
cell line, two independent preparations of total RNA were extracted, reverse
transcribed, and amplified by PCR in the presence of [or-3°SldATP. The PCR products
were separated on a 6% polyacrylamide gel and autoradiographed. The signal
demonstrating altered expression is marked by an arrow. The primers used were T12MT
and OPA7 as described by Qing at a]. (1997). (B) Northern blot analysis confirming
differential gene expression for fragment ST7 using the amplified DNA fragment as a
probe. Total RNA (15 pg) obtained from MSU-1.1 cells (lane 1) and 6AISB1 cells (lane
2) was subjected to Northern analysis as detailed in “Materials and Methods”. The blot
was stripped and reprobed with a glyceraldehyde 3-phosphate dehydrogenase

(GAPDH) cDNA (lower panel) as the loading control.

 

129

.0
.‘h

it
I

: «ST7

 

N If!“ |l

 

. 1'ST7

..<-GAPDH

130
consensus initiation site of eukaryotic mRNA translation (Kozak, 1991). The $17 cDNA

contains a putative 3' untranslated region of 450 bp followed by a poly A tail and three
consensus polyadenylation signals (Wahle and Keller, 1992) at nucleotide positions
2846, 2976 and 3034. Comparison of $17 cDNA sequence to the Genbank and EMBO
database using the FASTA and the BLAST program revealed that there is no significant

homology with any known genes.

Enmanr StmfluLe Suggest: a Nolel Itanemembtane ELotein. Comparison of the
deduced amino acid sequence of ST7 with a nonredundant protein sequence database
revealed that overall there was no significant homology with other proteins. A striking
feature of this protein is that the amino-tenninus is composed of five imperfect repeats
of a 40-amino acid cysteine-rich repetitive sequence homologous to that found in the
human low density lipoprotein (LDL) receptor (Shdhof at al., 1985) (Figure 4) and
complement protein C9 (Stanley e1 al.. 1985). The similarity also includes the highly
conserved, negatively charged Ser-Asp-Glu triad, which occurs near the COOH-terminal
end of each repeat. Hydropathy analysis (Kyte and Doolittle, 1982) revealed that the
coding sequence contained a stretch of 24 hydrophobic amino acids extending from
residues 493 to 516 (underlined in Figure 3B), which is flanked at both ends by
positively charged residues. This hydrophobic sequence resembles the membrane-
spanning region of other transmembrane proteins (Sabatini at al., 1982). The predicted
protein also contains other putative functional domains, including nine putative N-linked
glycosylation sites, a number of potential phosphorylation sites for protein kinase C and
casein kinase II, and several N-myristoylation sites. The importance of these putative

domains of ST7 protein under physiological conditions remains to be determined.

131

Fig. 2. Northern blot analysis of ST7 expression in multiple human tumor cell lines. (A)
Total RNA (15 pg), obtained from a series of cell lines, was hybridized with ST7 cDNA
using a cloned ST7 DNA fragment as a probe. These were: fibroblast cell line from a
normal donor (lane 1); parental MSU-1.1 cells (lanes 2 and 9); malignant cell lines
derived from a patient’s cervical carcinoma (lane 3); malignant cell lines derived from
tumors formed in athymic mice alter injection of MSU-1.1 cells transformed by a
transfected oncogene (lane 4) or by carcinogen treatment (lane 5); malignant cell lines
derived from a patient’s osteosarcoma (lanes 6 and 16), bladder carcinoma (lane 7);
malignant cell lines derived from tumors formed in athymic mice after injection of MSU-
1.1 cells spontaneously transformed (lane 8); malignant cell lines derived from a
patient’s fibrosarcoma (lanes 10-14), rhabdomyosarcoma (lane 15), Wilms' tumor (lane
17), and neurofibrosarcoma (lanes 18-20); and malignant cell lines derived from tumors
formed in athymic mice after injection of MSU-1.2 cells transformed by cobalt 60 (lanes
21 and 22). (B) RNA loading, evaluated by rehybridizing the same blots with a GAPDH

cDNA probe.

 

ZJ-Z'l'nSlN
IJ-Z'l-DSW
Luunsm
LIJIfISM
lull-:IN
eoeewo
9931

a a
oeour-r
L888
OVHS
.l.:I=dIA

ION
l'I-I‘ISIN

JJ./IQV'I

173 .I.
M188“-

183/0”

10 2'82 1-
Iflat-I

l' L'DSW
6818

<

132

b. on: “

 

910111213141516171819 202122

2345678

1

133

Fig. 3. Characterization of ST7 cDNA clones. (A) A diagram showing the four
overlapping ST7 cDNA clones. They were isolated from MSU-1.1 cells, a human
fibroblast library, a skeletal muscle cDNA library, and a human heart Marathon-ready
cDNA, as described in Materials and methods. (B) A near full-length cDNA sequence of
ST7 gene and the deduced amino acid sequence. The amino acid sequence is
numbered beginning with the first methionine of the longest open reading frame. The
one-letter code is used for amino acids. The 19 amino acids existing only in the longer
isoform is indicated by double underlining. The putative transmembrane domain is
underlined. The peptide used for antibody generation is in bold font. The putative

polyadenylation signals are indicated with sets of asterisks.

 

134

 

 

A
MSU-1.1 [:1 oopcn (0.5 kb)
Library 9 l 1 Clone A (2.8 kb) '
Skeletal Muscle [:3 Clone B (0.7 kb)
Normal Heart :3 Clone C (0.7 Kb)

135

CTCCTCCTCCGTCTCCTCCTCTCTCTCTCATCTGCTGTGGTTATGGCCTGTCGCTGGAGCACAAAAGAGTCTCCGCGGTGGAGGTCTGCGTTGCTCTTGCTUTCCTCGCTGGGGTGTAC
HACRWSTKESPRNRSALLLLFLAGVY

(ISMATGGTGCTCTTGCAGAAMTTCTGAAAATGTGCATAUTCAGGAGTGTWCTGCTTGTGGAGAGACTCCAGAGCAAATACGAGOACCAAGTGGCATAATCACAAGCCCAGGCTGG
GNQALAEHJENVHJSGVSTACGETPEOIRAPSGIITSPGH

CCTT CT GAATATCCTGCAAAAATCAACTGTAGCTGGTT CATAAGGGCAAACCCAGGCGAAATCATTACTATAAGWTTCAGGATTTTGATATTCAAGGATCCAGAAGGTGCAATITGGAC
PSEYPAKINCSNFIRANPGEIITISFQDFDIOGSRRCNLD

TGGTTGACAATAGAAACATACAAGAATATTGAAAGTTACAGAGCTTGTGGTTCCAOAATTCCACCTCCGTATATCTCTTCACAAGACOACATCTGGATTAGGTTTCATTCGGATGACAAC
HLTIETYKNIESYRACGSTIPPPYISSQDHIHIRFHSDDN

ATCTCTAGAAAGGGTI'TMGACTGGCATA | l | l I CAGGGAAATCTGAGGAACCAAATTGTGCTTGTGATCAGTFTCGTT GT GGTAATGGAAAGT GT ATACCAGAAGCCTGGAAATGCAAT
ISRKGFRLAYFSGKSEEPNCACDOFRCGNGKCIPEAHKCN

AACATGGATGAATGTGGAGATAGTTCCGATGAAGAGATCTGTGCOAAAGAAGCAAATCCTCCAACTGCTGCTGCTTTTCAACCCTGTGCTTACAACCAGTTCCAGTGTTTATCCCGTTTT
NHDECGDSSDEEICAKEANPPTAAAFQPCAYNQFQCLSRF

ACCAAAGI'ITACACTTGCCTCCCCGAATCTTTAAAATGTGATGGGAAOATTGACTGCCTTGACCTAGGAGATGAGATAGACTGTGATGTGCCAACATGTGGGCAATGGCTAAAATATTIT
TKVYTCLPESLKCDGNIDCLDLGDEIDCDVPTCGQHLKYF

TATGGTACTTTTAATTCTCCCAATTATCCAGACTTTTATCCTCCTGGMGCAATTGCACCTGGTTAATAGACACTGGTGATCACCGTAAAGTCAWTTACGCTTCACTGACTWAAACTT
YGTFNSPNYPDFYPPGSNCTHLIDTGDHRKVILRFTDFKL

GATGGTACTGGTTATGGTGATTATGTCAAAATATATGATGGATTAGAGGAGAATC CAMCAAGCTWTGCGTGTGTTGACAGCTTTTGATTCTCATGCACCTCTTACAGTTGTTTCTTCT
DGTGYGDYVKIYDGLEENPHKLLRVLTAFDSHAPLTVVSS

TCTGGAOAGATAAGGGTAUI l l l | GT GCTGATAAAGTGAATGCTGOAGGGGATTTAATGCTACTTACCAAGTAGATGGGTT CT GTTT GCCATGGGAAATACCCT GTGGAGGTAACTGG
SGQIRVHFCADKVNAARGFNATYOVDGFCLPHEIPCGGNH

GGGTGTT ATACTGAGCAGCAGCGTT GT GATGGGT ATTGGCATTGCCCAAATGGAAGGGATGAAACCAATTGTACCATGT GCCAGAAGMAGAATTTCCATGTT CCCGAAATGGTGT CTGT
GCYTEQQRCDGYHHCPNGRDETNCTNCQKEEFPCSRNGVC

TATCCTCGTTCTGATCGCTGCAACTACCAGAATCATTGCCCAAATGGCTCAGATGAAAAAAACTGC | | | l | l | GCCAACCAGGAAATTTCCATTGTAAAAACAATCGTTGTGTGTTTGM
YPRSDRCNYQNHCPNGSDEKNCFFCQPGNFHCKNNRCVFE

AGTT GGGTGT GT GATTCT CAAGATGACTGTGGT GATGGQGCGATGAAGAAAATTGCCCAGT AATCGTGCCTACAAGAGTCATCACTGCTGCCGT CATAGGGAGCCT CATCT GT GGCCTG
SHVCDSQDDCGDGSDEENCPVIVPTRVLTAAVIGSLICEL

TTACTCGTCATAGCATTGGGATGTACW GTAAGCTTTATTCTCTGAGAATGTTT GAAAGAAGATCATTTGAAAOACAGTT GT CAAGAGT GGAAGOAGAATTGTTAAGMGAGAAGCTCCT
L_LV1ALGCTCKLYSLRMFERRSFETOLSRVEAELLRREAP

CCCT CGT ATGGACAATTGATTGCT CAGGGTTT AATTCCACCAGTTGAAGATTTTCCT GmGTTCACCT AATflGGCTTCT G I | l IGGAAAATCTGACISCTAGCGGTACGATCTCAGCTT
PSYGQLIAOGLIPPVEDFPVCSPNOASVLENLRLAVRSQL

GGATTTACTTCAGTCAGGCTTCCTATGGCAGGCAGATCAAGCAACATTTGGAACCGTAI | l | lAATHTGCAAGATGACGTCATTCTGGGTCATTGGCTTTGGTCTCAGCAGATGGAGAT
G FTSVRLPIIAGRSSNIHNRIFNFARSRHSGSLALVSADGD

GAGGTTGTCCCTAGTCAGAGTACCAGTAGAGAACCTGAGAGAAATCATACTCACAGAAGTTTGTTTTCCGTGGAGTCTGATGATACAGACACAGAAAATGAGAGAAGAGATATGGCAGGA
E VVPSQSTSREPERNHTHRSLFSVESDDTDTENERRDHAG

GCATCTGGTGGGGTTGCAGCTCCTTTGCCTCAAAAAGTCCCTCCCACAACGGCAGTGGAAGCGACAGTAGGAGCATGTGCAAGTTCCTCAACTCAGAGTACCCGAGGTGGTCATGCAGAT
A SGGVAAPLPQKVPPTTAVEATVGACASSSTQSTRGGHAD

MTGGAAGGGATGTGACAAGT GT GGAACCCCCAAGTGT GAGTCCAGCACGTCACCAGCTTACAAGTGCACTCAGT CGTATGACT CAGGGGCTACGCTGGGTACGWTTACATTAGGACGA
N GRDVTSVEPPSVSPARHQLTSALSRHTQGLRIIVRFTLGR

TCAAGTTCCCTAAGTCAGAACCAGAGTCCTTTGAGAMACTTGATAATGGGGTAAGTGGAAGAGAAGATGATGATGATGTTGAAATGCTAATTCCAATITCTGATGGATCTTCAGACTTT
S SSLSQNQSPLRQLDNGVSGREDDDDVEI‘ILIPISDGSSDF

GATGTGAATGACTGCTCCAGACCTCTTCTTGATCTTGCCTCAGATWGGAWGGGCWAGAWCCATATAATGCMCAAATCCTGGAGTAAGGCCAAGTAATCGAGATGGCCCCTGT
D VNDCSRPLLDLASDQGQGLRQPYNATNPGVRPSNRDGPC

GAGCGCTGTGGTATTGTCCACACTGCCCAGATAC CAGACACTTGCTTAGAAGTAACACTGAAAAACGAAACGAGTGATGATGAGGCTFTGTTACTTTGTTAGGTACGMTCACATAAGGG
E RCGIVHTAQIPDTCLEVTLKNETSDDEALLLC

AGATTGTATACAAGTTGGAGCAATATCCATTTATTAUTTGTAACTTTACAGTTMACTAGTTTTAGWTAAAAAGAAAAAATGCAGGGTGATTTCTTATTATTATATGTTAGCCTGCAT

GGTTAAATTCGACAACTTGTAACTCTATGAACTTAGAGTTTACTATTTTAGCAGCTAAAAATGCATCACATATTGCATATTGTTCAATAATGGTCCTTTCATTTGTTTCTGATTG| l 1 IC

******

ATCCTGATACTGTAGTTCACTGTAGAAATGTGGCTGCTGAAACTCATTTGATTGTCAl l l I IATCTATCCTATGTTAAATGGI l IGI l | l IACAAAATAATACCTTATTTTAATTGAAAC

******

GTTI'ATGCI | l IGCCAAGCACATCTTGTMCTTAATATAGCTAGATGTTAAGGTTGTTAATGTACCAAAAAMAMAA

*iiiit

Figure 38

120
26

240
360
106

480
146

600
186

720
226

840
266

960
306

1080
346

1200
386

1320
426

1440
466

1660
506

1680
546

1800
586

1920
626

2040
2160
706

2280
746

2400
786

2520
826

2640

859

2760

2880

3000

3078

136
Banem 9.1 Emmaion 91 SIZ mRNA in Manage Human Iiesues. Northern analysis

using a human normal tissue blot (Clontech) revealed that ST7 cDNA hybridized to a
single transcript of 3.7 kb. The ST7 mRNA was found to be most abundant in heart and
skeletal muscle, expressed at moderate levels in brain, lung, placenta and pancreas,

and barely detectable in tissues containing large number of epithelial cells such as liver

and kidney (Figure 5).

mmmmmflmjflnmflmanlumarmm. Arabbitanti-
ST7 polyclonal antibody (designated B250) was raised against a synthetic peptide as
described in Materials and Methods. This antibody recognized a protein with an
apparent molecular mass of 85 kDa, which is lower than the estimated molecular mass.
The reason for this anomalous migration of ST7 protein in SDS-polyacrylamide gel is not
clear- The specificity of this antibody for ST7 protein was tested by determining the
extent of inhibition with the antigen peptide. After preincubation of the antibody with the
antigen peptide (60 uM), the signal for the ST7 protein band in immunoblots was
dramatically reduced (Figure 6). Using this antibody, the expression level of ST7 protein
in normal cells and tumor-derived cell lines was analyzed by Western blotting analysis.
As shown in Figure 7, MSU-1.1 cells and normal human fibroblast cell line SL89
exhibited high levels of $17 protein, whereas the six malignant MSU-1.1 derivatives
(lanes 3-8) and three cell lines derived from human fibrosarcomas (lanes 9-11) showed
low levels of ST7 protein, which was consistent with the levels of ST7 mRNA transcripts
in these cells (Figure 2). Interestingly, human fibrosarcoma-derived cell lines HT1080
and VIP:FT and rhabdomyosarcoma-derived cell line RD showed an mRNA level of ST7
comparable to that in MSU-1.1 cells (Figure 2, compare lane 9 and 11,14 and 15),

whereas the protein was undetectable (Figure 7, compare lane 1 and 12-14). This result

137

Fig. 4. Alignment of the cysteine-rich repeat region of ST7 with that of the human LDL
receptor. Amino acids are numbered on the left. Conserved amino acids are boxed.
The data of human LDL receptor are taken with permission from “The LDL receptor

gene: A mosaic of exons shared with different proteins” by Sudhof at al., Science, 1985.

 

 

138

U]

L)

U C)

 

m

0

CT (4

 

[L5

 

 

C) d: 3 Id

m

 

 

 

 

23 3
O

X:
C)

 

 

DJ

(J

L) L)

 

[U

 

 

 

 

 

 

 

 

 

m C) m (D

C)

 

r0

 

 

 

(J L) U

 

 

>flz

x“.

a:
I

01

 

 

r0

0|

C) M

 

[5)

I:

C)

 

’Eritu

7h

ha

ha

 

 

A (3 C)

d

U]

L)

 

L9

L) o (J o

 

L5)

 

 

m m m

mmvlmvv

mevlaav

OHvIOBM

vaIoHN

oomrama

.oz osoflmom

£995 Cw .m

MDMCUWCOU
> z > mmNIomN
fl > 4 mvNIHHN
m m o oamlmha
A > m meIMNH
x m m NHHrvm
> m A mmIMv
a o > ~e-~

.02 ODUfimOm

5389-43 .<

139
suggests that downregulation of ST7 gene product occurs at both transcriptional and

posttranscriptional levels.

140

Fig. 5. Expression of ST7 gene in normal human tissues. (A) Human multiple tissue
blot (Clontech) containing poly (A)+ RNA (2 jig/lane) from the indicated tissues probed

with labeled 2.6 kb ST7 cDNA fragment. (B) The blot, stripped and rehybridized with a

probe for B-actin as a loading control.

 

141

mamaocmn.
>962

eEms—2 68.3w
..o>_._

was;

Scoomi

59m

two...

(Kb)

 

142

Fig. 6. Analysis of the specificity of the antibody B250. lmmunoblots of cell lysate (25
pg protein/lane) from MSU-1.1 cells were incubated with the rabbit anti-ST7 peptide
antibody B250 without (lane 1) or with (lane 2) prior incubation with the antigen peptide.

The position of the ST7 signal is indicated by an arrow.

 

143

c
It“:

a. 1 ¢ 85 kDa

Fig.6

144

Fig. 7. Western blot analysis of the expression of ST7 in normal human fibroblast cell
lines (MSU-1.1 and SL89) and 12 tumor-derived cell lines. Cellular protein (25 pg) was
analyzed by electrophoresis in 10% polyacrylamide gel and probed with antibody B250.
The origin of the tumor-derived cell lines are: malignant cell lines derived from tumors
formed in athymic mice after injection of MSU-1.1 cells transformed by various methods
(lanes 3-8); malignant cell lines derived from a patient’s fibrosarcoma (lanes 9-13) and

rhabdomyosarcoma (lane 14).

145

n— m
tun—5
one 2....
.02

swam

U<Im
Ewe.—
PDNBNP
Nara. Tam—2
_.._-N. Tam—2
wmmfi: _.
9mm);

mar—m
_.. Tam—2

5 6 7 8 91011121314

4

3

146
DISCUSSION

We have cloned and identified a novel gene, ST7, that encodes a 3.7 kb mRNA. Our
analysis of ST7 expression in multiple human tumor-derived cell lines revealed that 10
out of 15 tumor-derived cell lines from patients exhibit low or no expression of ST7
mRNA. Without the normal cells from which these tumor cell lines were derived, we
cannot be certain that the low expression resulted from downregulation. However,
because we found that ST7 is widely expressed in normal human tissues, including
heart, skeletal muscle, brain, lung, pancreas and placenta, but not in tissues consisting
of a large number of epithelial cells, such as liver and kidney, and that a series of normal
human fibroblast cell lines in culture expressed relatively constant levels of ST7 mRNA
and protein (data not shown), we conclude that $17 is ordinarily expressed at least in
most cells of mesenchymal origin. Because six out of 16 tumor-derived cell lines from
MSU-1.1 cells malignantly transformed by various methods clearly exhibited
downregulation of ST7 mRNA and protein, we conclude that at least in mesenchymal-
derived tumor cell lines, the downregulation of ST7 is frequently associated with

neoplastic transformation.

Our data also suggest that the regulation of ST7 expression occurs at both the
transcriptional and posttranscriptional levels. In most cases, the steady state protein
level of ST7 correlates well with its mRNA level. However, in three malignant cell lines
tested in this study, i.e., HT1080, VIP:FT and RD, which were derived from sarcomas
taken from patients, the steady state mRNA level of ST7 was comparable to that seen in
normal human fibroblasts, whereas the ST7 protein was barely detectable. This may

reflect failure of translation initiation or a defect in the stability of the protein product.

147
According to the apparent molecular weight of the protein on SDS-PAGE gel, the

cDNA we have isolated can account for the entire coding region of the novel protein. Of
the two forms of the ST7 mRNA that we isolated, the shorter lacks 57 nucleotides at the
5' end and encodes a protein of 840 amino acids that lacks residues 27 to 45 found at
the amino terminal end of the longer ST7 protein (indicated by double underline in
Figure 3B). This omission does not affect the open reading frame. The two isoforms
presumably result from alternative splicing. The significance of the 5' end heterogeneity

awaits more detailed functional analyses.

The amino-terminal half of the predicted ST7 protein product contains five imperfect
40-amino acid repeats. This set of repeats is homologous to the cysteine-rich, ligand-
binding domain in the human LDL receptor. Many cell surface and secreted proteins
contain repeated cysteine-rich motifs that are each 40-50 amino acids in length and
contain six cysteine residues linked in three disulfide bridges (Doolittle et al., 1984;
Appella et al., 1988; Daly et al., 1995). The common functional characteristic of these
repeats is their ability to mediate protein-protein interactions. Therefore it is possible
that this conserved sequence in ST7 is involved in ligand binding. However, it cannot be
ruled out that this cysteine-rich repeat sequence codes for a structural motif common to
a number of extracytoplasmic protein domains. Since the ST7 protein shares no other
significant homology with any known protein, we can only speculate as to its function.
One possibility is that the $17 protein is a component of a signal transduction pathway
that negatively regulates cell growth. The predicted membrane-spanning structure of the
ST7 protein suggests that it acts as a cellular receptor or co-receptor, functioning as a
ligand-regulated suppressor of a signaling unit like the recently cloned tumor suppressor

gene patehed, which encodes a receptor for Sonic hedgehog (Stone et al., 1996).

148
Another possibility is that $17 protein participates in cellular adhesion in a fashion

similar to the candidate tumor suppressor gene DCC (for deleted in colorectal cancer)
(Cho and Fearon, 1995; Fearon, 1996), or $17 protein undergoes proteolysis on the cell
surface, releasing a locally acting chemical signal. Any of these mechanisms could play
a role in tumorigenesis. The cloning of the ST7 cDNA and the availability of the ST7-
specific antibody should facilitate further experiments designed to better understand the

function of the novel gene and its role in tumorigenesis.

149
MATERIALS AND METHODS

Cells and Cell Culture. The infinite life span human fibroblast cell strain MSU-1.1 and
its derivative cell lines were routinely cultured in Eagle’s minimum essential medium
modified by addition of L-aspartic acid ( 0.2 mM), L-serine (0.2 mM) and pyruvate (1mM)
and supplemented with 10% SCS (Hyclone Laboratory, Logan, UT), penicillin (100
units/ml), streptomycin (100 pglml) and hydrocortisone (1 pglml) (complete medium) at

37°C in a humidified incubator containing 5% 002 in air.

QitfeLential mRNA Display, Non-tumorigenic infinite life span human fibroblast cell
strain MSU-1.1 and one of its tumorigenic derivative cell line, designated 6AISB1, were
used as sources of RNA for this study. Differential mRNA display and TA cloning were

carried out essentially as described (Qing et al., 1997).

Nenhem Blet Analyeie. Total RNA from cells in exponential growth was extracted
using RNAzolB (T el-Test, Friendswook,TX) according to manufacturer’s instructions.
For Northern blot analysis, RNA (15 pg) from each cell line was electrophoresed on
1.2% agarose/2.2 M formaldehyde gels, and then was transferred to hybond-N
membrane and immobilized by UV crosslinking (UV Stratalinker 2400, Stratagene, La
Jolla, CA). The cDNA probe was radiolabeled using a random primed labeling method
(Feinberg and Vogelstein, 1983). The blots were hybridized as previously described
(Qing at al., 1997). For analysis of ST7 expression in various normal human tissues,
the Multiple Tissue Northern Blot was purchased from Clontech (Clontech, Palo Alto,
CA). Variation in RNA loading per lane was evaluated by probing with the GAPDH

cDNA or n-actin as the controls.

150
Claning 91 Human 6]] cDNA. The directional human fibroblast cDNA library ( a

generous gift from Dr. Legerski, The University of Texas, MD. Anderson Cancer Center,
Houston, that is referred to as Library 9) was used to obtain the full-length cDNA
corresponding to the ST7 gene. We screened the library by the High Fidelity Expand
PCR method (Boehringer Mannheim, Indianapolis, IN) with a vector-specific primer (5'-
CCGGAAGCTTCTAGAGATCCCTCGA) and a ST7 gene specific primer based on the
partial ST7 sequence obtained from differential display (5'-
GCTCCAACTTGTATACAATCTCCC). Plasmid DNA derived from 10 x 10° independent
clones was used as the template. The 50 pl PCR mixture contained 1.75 mM MgCl2,
0.2 mM dNTP, 15 pmol of each primer, 100 ng of library plasmid DNA, and 2.5 units of
the mixture Taq and Pwo DNA polymerase. The PCR cycling consisted of initial
denaturation at 94°C for 2 min, followed by 10 cycles of 94°C for 30 s, 63°C for 30 s,
and 68°C for 3 min, followed by another 20 cycles with the same parameters, except
that the elongation time was extended for 15 s for each new cycle, followed by final
elongation at 68°C for 10 min. The PCR product was separated by electrophoresis in
1% agarose gel, and the major bands were purified using Qiaquick Gel Extraction kit
(Qiagen, Chatsworth, CA). The purified DNA (about 2.6 kbp, noted as clone A) was
used as probes for Northern analysis and cloned into the pCRII vector using the TA

cloning method (Invitrogen, San Diego, CA).

To obtain additional 5' sequence of this gene, we screened a human skeletal muscle
cDNA library (a generous gift from Dr. Ki-Han Kim, Purdue University, West Lafayette)
by PCR using one primer from the 5' end of clone A, designated JM131 (5'-
GGGTI'GAAAAGCAGCAGGAGTTGGAGG) and another vector-specific primer from the

region of the cloning site of kgt11 (5' GATTGGTGGCGACGACTCCTGGAGC). The

151
fragment generated was subcloned and designated clone B.

Clone C, which contains the first translation initiation codon, was isolated by the 5'
rapid amplification of cDNA ends (5'-RACE) method using a human heart Marathon-
ready cDNA (Clontech) with the ST7 gene-specific primer JM131. The PCR products

were subcloned into the pCRll vector as above and sequenced.

DNA Segueneing and Sequence Analusjs. Both strands of the cDNA inserts in the

pCRII vector were sequenced manually by the dideoxy chain termination method with
the SP6 and T7 primers using a Fidelity DNA Sequencing kit (Oncor, Gaithersburg,
MD). For long cDNA inserts, synthetic oligonucleotides were used as primers to
complete the sequence. Resolution was improved in some regions by replacing dGTP
with deaza-leP in the nucleotide mixture. The cDNA sequence and deduced protein
sequences were analyzed by FASTA and BLAST program with the DNA and protein
databases at the National Center for Biotechnology Information (NCBI). Secondary
structure predictions and the properties of the putative protein were calculated using the

GCG program (Genetic Computer Group, Madison, WI).

Eredugflen Qt Antjfi‘fl Antibody, The peptide corresponding to the C-terminus of the
ST7 protein (CLEVTLKNESTDDEA in the single-letter amino acid code; corresponding
to amino acids 841 to 855 of ST7) was synthesized by the Macrostructural Facility of the
Department of Biochemistry, Michigan State University. The synthetic peptide was
coupled to keyhole limpet hemocyanin (KLH) with the chemical crosslinker
glutaraldehyde. To obtain anti-ST7 antibody, 200 pg of KLH-conjugated peptide was

emulsified with an equal volume of TiterMax (Cytrx, Norcross, GA) in a total volume of 1

152
ml, and 0.1 ml was injected subcutaneously into each of four sites on each of two

female New Zealand White rabbits. The rabbits were administered booster shots after
four weeks. They were bled on day 42 and 56 and serum was prepared according to

standard protocol (Sambrook et al., 1989) and was designated B250.

Western BJetting Analysis, Cell lysates were prepared with RIPA buffer composed of
50 mM Tris-HCI, pH 7.2, 150 mM NaCl, 1% Triton X-100, 0.1% SDS, 0.5% deoxycholic
acid, 2 mM phenylmethylsulfonyl fluoride, 1 mM EDTA and 0.15 units/ml aprotinin as
described (Qing et al., 1997). Aliquots of cell lysates containing 25 pg of protein were
mixed with the sample buffer (0.05M Tris-HCI, pH6.9, 9% glycerol, 2.3% SDS, 0.1%
bromophenol blue and 5% B—Mercaptoethanol), separated on a SDS/polyacrylamide gel
(10%), and electroblotted onto an lmmobilon-P membrane (Millipore, Bedford, MA). The
blots were incubated for 2 h at room temperature in Tris-buffered saline (20 mM Tris-
HCl, pH 7.6, 137 mM NaCl) containing 0.1% (v/v) Tween 20 and 5% (w/v) non-fat dry
milk (blocking solution), and then incubated for 2 h at room temperature with B250, the
rabbit anti-ST7 antibody, diluted 1:500 in the same solution. The blots were washed
several times and then incubated with horseradish peroxidase-conjugated goat-anti-
rabbit IgG (Sigma, St. Louis, MO) that had been diluted 1:5000 with blocking solution.
Enhanced Chemiluminescence (Amersham, Arlington Heights, IL) was used according

to the manufacturer’s recommendations to detect the signal.

153
ACKNOWLEDGMENTS

We thank Dr. Legerski at the University of Texas, MD. Anderson Cancer Center,
Houston and Dr.Ki-l-Ian Kim at Purdue University, West Lafayette for kindly providing us
the cDNA libraries. The expert assistance on computer work from Dr. David Dewitt at
Michigan State University is gratefully acknowledged. This research was supported by
DHHS grants CA60907 from the National Cancer Institute, and AG11026 from the

National Institute of Aging.

 

154
REFERENCE

Appella, E, Weber, IT and Blasi, F. (1988). E85 Lett., 231, 1-4.

Cho, KR, and Fearon, ER. (1995). Cum Quin, Genet, Dex" 24, 3-17.

Daly, NL, Scanlon, MJ, Djordjevic, JT, Kroon, PA and Smith, R. (1995). Ema. NatL
Acad. Sci. USA, 92, 6334-6338.

Doolittle, RF, Feng, DF and Johnson, MS. (1984). NatuLe, 307, 558-566.

Farber, E. (1984). Caneex Res., 44, 4217- 4223.

Fearon, ER. (1996). Biecnjm. Biephys. Acta, 1288, 17-23.

Fearon, ER and Vogelstein, B. (1990). Cat], 61, 759-767.

Feinberg, AP and Vogelstein, B. (1983). AnaL Bieebem, 132, 6-13.

Klein, G and Klein, E. (1985). NatuLe, 315, 190-195.

Kozak, M. (1991). J. BieL Chem. 266, 19867-19870.

Kyte, J and Doolittle, RF. (1982). J. MeL Biel. 157, 105-132.

Liang, P, Averboukh, L and Pardee, AB. (1993). Nueleie Aeide Best, 21, 3269-3275.

Liang, P and Pardee, AB. (1992). Seienee, 257, 967-969.

McCormick, JJ and Maher, VM. (1994). Riel: Anal., 14, 257-263.

Morgan, TL, Yang, D, Fry, DG, Hurlin, PJ, Kohler, SK, Maher, VM and McCormick, JJ.
(1991). Exp. ecu Bee..197, 125-136.

Park, NH, Gujuluva, CN, Baek, JH, Cherrick, HM, Shin, KH and Min, BM. (1995).
anegene, 10, 2145—2153.

Peto, R. (1977). anflginsthlumanQangeL Cold Spring Harbor Laboratory, Cold

Spring Harbor, NY, 1403-1428.
Qing, J, Maher, VM, Tran, H, Argraves, WS, Dunstan, RW and McCormick, JJ. (1997).

Qneegene. In press.
Reinhold, DS, Walicka, M, Elkassaby, M, Milam, LD, Kohler, SK, Dunstan, RW and

155

McCormick, JJ. (1996). Int. J. Radiat. BieL, 69, 707-715.

Reznikoff, CA, Kav, C, Messing, EM, Newton, M and Swaminathan, S. (1993). Semin.
Cancer Biel., 4, 143-152.

Rhim, JS, Webber, MM, Bello, D, Lee, MS, Amstein, P, Chen, LS and Jay, G. (1994).
Bree. Natl. Aead. Set. USA, 92, 11874-11878.

Sabatini, DD, Kriebich, G, Morimoto, T and Adesnik, M. (1982)._J. CetlBiel. 92, 1-21.

Sambrook, J, Fritsch, EF and Maniatis, T. (1989). Melegutar Clentng: A Laberatery
ManuaL Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.

Stanley, KK, Kocher, HP, Luzio, JP, Jackson, P and Tschopp, J. (1985). EMB_Q ,L, 4,
375-382.

Stone, DM, Hynes, M, Armanini, M, Swanson, TA, Cu, 0, Johnson, RL, Scott, MP,
Pennica, D, Goddard, A, Phillips, D, Noll, M, Hooper, JE, de Sauvage, F and
Rosenthal, A. (1996). Nature, 384, 129-134.

Stoner, GD, Kaighn, ME, Reddel, RR, Resan, JH, Bowman, D, Naio, Z, Matsukura, M,
You, M, Galati, AJ and Harris, CC. (1991). Cancer Res., 51, 365—371.

Sudhof, TC, Goldstein, JL, Brown, MS and Russell, DW. (1985). Scienee, 228, 815-
822.

Wahle, E and Kellert, W. (1992). Ann. BBL Biechem. 61, 419-440.

Yamamoto, T, Bishop, RW, Brown, MS, Goldstein, JL and Russell, DW. (1986).
Seienee, 232, 1230-1237.

Yang, D, Louden, C, Reinhold, DS, Kohler, SK, Maher, VM and McCormick, J. J.

(1992). ELQQ. flat]. Aged. Sci. USA” 89, 2237-2241.

”‘lllllllllllllllllllli