LIBRARY
Michigan State
University

76/74/55 *-

This is to certify that the
thesis entitled

EVALUATION OF THE TUMOR SUPPRESSOR GENES
14-3-3 SIGMA AND P53 IN FELINE MAMMARY CARCINOMA

presented by

MEGAN STROHMEYER ALBERTELLI

has been accepted towards fulfillment
of the requirements for the

MS. degree in Small Animal Clinical Sciences

 

 

”C... r
go"

/

%/7///// 9

Major Profes’sor’s Signature '

m. M/ MES”
a r

Date

 

MSU is: an Affirmative Action/Equal Opportunity Institution 3

 

 

 

PLACE IN RETURN Box to remove this ched<out from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 c:/CI_FICIDetoDchndd-ms

 

EVALUATION OF THE TUMOR SUPPRESSOR GENES 14-3-3 SIGMA AND P53 IN
F ELINE MAMMARY CARCINOMA

By

Megan Strohmeyer Albertelli

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
' for the degree of
MASTER OF SCIENCE
Department of Small Animal Clinical Sciences

2005

ABSTRACT

EVALUATION OF THE TUMOR SUPPRESSOR GENES 14-3-3 SIGMA AND P53 IN
F ELINE MAMMARY CARCINOMA

By
Megan Strohmeyer Albertelli
Few studies have examined the genetic changes that occur in feline mammary
tumorigenesis. Therefore, we examined feline mammary carcinoma surgical biopsy
samples for hypermethylation of 14-3-3 o (o) and loss of heterozygosity (LOH) of P53,
two common molecular changes found in human primary breast tumors. Feline a was
amplified using PCR primers designed for conserved regions in human and mouse a and
sequenced. DNA isolated from normal and tumor tissue was modified with bisulfite,
amplified by PCR, and sequenced in order to differentiate methylated and unmethylated
sites. The number of methylated sites in 5 paired normal and tumor DNA was compared
and found to not be significantly different in each case. The number of methylated sites in
8 unpaired samples was compared and no significant difference was found between
normal and tumor DNA groups. Therefore, unlike human breast cancer,
hypermethylation of a does not play a significant role in feline mammary carcinoma
tumorigenesis. P53 LOH was examined by genotyping normal tissue DNA by restriction
digest at three single nucleotide polymorphism (SNP) markers in PS3. Tumor DNA was
then genotyped for cases in which the normal DNA was heterozygous at one or more
markers. Of the 20 informative cases studied, 3 cases (15%) showed LOH (complete loss
of one allele), 2 cases (10%) showed allelic imbalance (incomplete loss of one allele), 1
case (5%) showed LOH at only one of two informative markers, and 14 cases (70%)

showed no LOH. Therefore, the P53 LOH rate is similar between felines and humans.

ACKNOWLEDGEMENTS

I would like to thank my mentor, Vilma Yuzbasiyan-Gurkan, for all her support
and guidance during this project. I would also like to thank the members of the
Comparative Genetics lab and my committee for their help and support. In addition I
would like to thank all the students and staff of the College of Veterinary Medicine (and
their cats) who enthusiastically responded to my request for feline buccal swabs. And, of

course, I would like to thank Guy for all his patience and support during this project.

iii

TABLE OF CONTENTS

List of Tables ................................................................................................................. vi

List of Figures ............................................................................................................... vii
Chapter 1

Introduction. .................................................................................................................... l

1.1 Feline Mammary Neoplasia ........................................................................ 1

1.2 Cancer Genetics .......................................................................................... 3

1.3 Cats as Animal Models for Human Breast Cancer ....................................... 8

References ........................................................................................................... 9
Chapter 2

Summary of Cases ........................................................................................................ 11

2.1 Description of Sample Population ............................................................. 11

2.2 Tumor Types ............................................................................................. 11

2.3 DNA Isolation ........................................................................................... 12
Chapter 3

14-3-3 o ........................................................................................................................ 14

3.1 Background .......................... . .................................................................... 14

3.2 Materials and Methods .............................................................................. 17

3.2.1 Sequencing Feline 14-3-3 0 ........................................................ 17

3.2.2 Methylation Status of Feline Mammary Adenocarcinoma Cases. 18

3.2.3 Statistical Analysis ..................................................................... 22

3.3 Results ...................................................................................................... 22

3.3.1 Feline l4-3-3 0 Sequence ........................................................... 22

3.3.2 Feline Mammary Carcinoma Methylation Status ........................ 24

3.4 Discussion ................................................................................................... 24

References ......................................................................................................... 28
Chapter 4

P53 ................................................................................................................................ 30

4. 1 Background ............................................................................................... 3O

4. 2 Materials and Methods .............................................................................. 34

4.2.1 Development of SNP Genotyping Tests ..................................... 34

4.2.2 Allele Frequencies ...................................................................... 35

4.2.3 Genotyping of Feline Mammary Carcinoma Samples ................. 38

4.3 Results ...................................................................................................... 40

4.3.1 SNP Genotyping Tests ................................................................ 40

4.3.2 Allele Frequencies ................................................... . ................. 41

4.3.3 Loss of Heterozygosity of P53 ................................................... 41

4.4 Discussion ................................................................................................ 42

References ......................................................................................................... 45

iv

Chapter 5

Conclusion .................................................................................................................... 47
5.1 Potential Gene Targets in Future Studies of Feline Mammary
Adenocarcinoma ................................................................................................ 47
5.2 Methylation in Feline Mammary Adenocarcinoma .................................... 51
5.3 Cats as Animal Models for Human Breast Cancer ..................................... 52
References ......................................................................................................... 54

LIST OF TABLES

Table 2.1: Tumor Type Summary .................................................................................. 13
Table 3.1: Cases Examined for o Hypermethylation ...................................................... 25
Table 3.2: Number of Methylated Cst in Paired Feline Mammary Carcinoma Samples2 6
Table 3.3: Number of Methylated Cst in Unpaired Feline Mammary Carcinoma

Samples ......................................................................................................................... 26
Table 4.1: Reported Polymorphisms in Feline P53 ........................................................ 34
Table 4.2: SNP Genotyping Tests: PCR Conditions ....................................................... 38
Table 4.3: SNP Genotyping Tests: Restriction Digests .................................................. 39
Table 4.4: SNP Allele Frequencies ................................................................................ 41
Table 4.5: PS3 LOH Results .......................................................................................... 42
Table 4.6: Cases Studied for P53 LOH .......................................................................... 43

vi

LIST OF FIGURES

Figure 1.1: The Cell Cycle .............................................................................................. 5
Figure 3.1: Human and Mouse 0 ................................................................................... 19
Figure 3.2: Primers Used in the Heminesting PCR of Bisulfite Modified or ................... 21
Figure 3.3: Human and Feline 0 Exon 1 ....................................................................... 23
Figure 4.1: Feline P53 Genomic Sequence: Exon 5 - Exon 8 ......................................... 32
Figure 4.2: Schematic of Restriction Digest Genotyping Tests ...................................... 36
Figure 4.3: Complete LOH versus Allelic Imbalance ..................................................... 40

vii

CHAPTER 1

INTRODUCTION

1.1 Feline Mammary Neoplasia

Mammary tumors are the third most frequent tumor occurring in the cat

(following hematopoietic and skin neoplasia).l Mammary tumors make up 12% of all

feline malignant tumors and 17% of all tumors in female cats, with an incidence of 12.8

per 100,000 total cats and 25.4 per 100,000 female cats.1 Intact female cats are at the

greatest risk, with ovariectomized females having 0.6 of the relative risk of developing

mammary carcinoma and males very rarely affected (less than 1% of total cases).2 The

mean age of occurrence is 10 to 12 years, although cases ranging from 9 months to 23
years have been reported.3 Siamese cats develop mammary tumors at a younger age (7-9

years) but are not at increased risk after this time period.2

Feline mammary tumors are malignant in 80 to 85% of cases, growing rapidly and

invasively and metastasizing most commonly to the regional lymph nodes and lungs.3’4

80% of these tumors are adenocarcinomas, with tubular adenocarcinomas, papillary

adenocarcinomas, and solid carcinomas being the most common types.3 Feline mammary
carcinomas rarely contain estrogen receptors but frequently contain progesterone

receptors, and studies have shown that exogenous progestogen administration increases

the risk of both malignant and benign mammary tumors.5
Surgery is the most common therapy for feline mammary carcinoma. Radical
mastectomy (the removal of all mammary glands on an affected side) has been shown to

significantly increase the disease-free interval and non-significantly increase survival

time when compared to conservative surgery.6 Recurrence of the tumor at the tumor site

occurs in 66% of cases treated with conservative surgery.3 Combination chemotherapy
with doxorubicin and cyclophosphamide has been shown to induce short-term partial

responses (250% reduction in tumor mass) in 50% of cats with metastasis or

nonresectable tumors, increasing average survival time from 2.5 to 5 months.7
Adriamycin, mitoxantrone, vincristine, and cisplatin but not recombinant human tumor

necrosis factor alpha or recombinant feline interferon gamma have been shown to be

effective in vitro.8’9

Feline mammary carcinomas are generally given a guarded to poor prognosis due
to their invasiveness and likelihood of metastasis.3 The most significant prognostic factor

is tumor size, with tumors 1 cm3 to 8 cm3 associated with the lon est disease-free
g

interval and survival time, while age at diagnosis and breed do not have any prognostic

value.6 There is no standardized histologic grading system that can be related to
prognosis for feline mammary carcinomas. The Elston and Ellis method, which grades
tumors through assessment of degree of tubule formation, degree of nuclear and cellular
pleomorphism, and mitotic count, shows survival time predictive value for well-

differentiated carcinomas (grade I) and poorly differentiated carcinomas (grade 111) but
not for moderately differentiated carcinomas (grade II).10 The average survival time afier

detection of a mammary tumor in both treated and untreated cats is 10-12 months.3
However, cats with tumors of less than 2 cm in diameter have a average survival time of

over 3 years, emphasizing the importance of early detection and treatment in this

disease.3 Unfortunately, cats are presented to the veterinarian an average of five months

after the owner initially notes the tumor, and therefore most cats are in an advanced state

of disease when treated clinically.3

1.2 Cancer Genetics.

The etiology of mammary adenocarcinoma is still not fully understood, although
environmental toxicants, viruses, and inherited traits have all been suggested as possible
causes. These agents may seem diverse, but the common thread between them is that they
can damage or propagate changes in normal DNA. All cancers ultimately are caused by
the modification or inappropriate activation or inactivation of genes involved in the
regulation of cell growth and differentiation, resulting in uncontrolled cellular
proliferation. Thus, cancer is a genetic disease.

The process of cellular reproduction, the cell cycle, is normally a tightly
controlled process involving many genes that either promote or halt cellular proliferation.
The cell cycle consists of four phases: G1 (or gap 1), S (synthesis of DNA), GZ (or gap

2), and M (mitosis, the division of cellular components from the original cell into two

daughter cells).11 While production of cellular proteins in order to double cell size occurs
throughout the cell cycle, DNA replication only occurs at a specific point in the process.
In order to commit to DNA replication the cell must pass the G1 checkpoint, in which a

feedback system assesses cellular and environmental signals for the appropriateness of

cell division and triggers the events necessary for synthesis of DNA. 12 After DNA
synthesis, the cell must then commit to cell division by passing the G2/M checkpoint,
which in a similar fashion to the G1 checkpoint uses a feedback system to assess the state

of DNA replication and, when replication is complete, triggers the events necessary for

mitosis.13 This system ensures that the DNA is undamaged before replication and that

. DNA replication is completed before mitosis. The feedback system responsible for
controlling the cell cycle is based on two families of proteins called cyclins and cyclin-
dependent protein kinases (or Cdks). Cdks activate proteins involved in the cell cycle by
phosphorylating their serine and threonine residues. Cdks are not active unless bound to
cyclins, which are synthesized and degraded in each cell cycle round. Cyclins and Cdks
are conserved throughout eukaryotes, although the process is best understood in yeast,

where it was shown that a specific Cdk named cdc2 was necessary for advancement

through both the GI and GZ/M checkpoints.1 1'] 3 Cdc2 was shown to associate with
different cyclins (G1 or mitotic cyclins) in different stages of the cell cycle, which
conferred specificity to cdc2 and allowed it to phosphorylate different target proteins at
each stage. In mammals, several Cdks and cyclins are involved in passing the GI and

G2/M checkpoints. In order to pass the G1 checkpoint, a mammalian cell requires
cdk2,3,4,6 and the cyclin D and E groups.12 In order to pass the G2/M checkpoint, a

mammalian cell requires cdc2 and cyclin groups A and B.13 See Figure 1.1 for a
summary of the cell cycle.

Genetic alterations that predispose an individual to cancer may be inherited or
somatic. Inherited, or germline, mutations are passed on from the parents' gametes and
are thus found in every cell of the offspring's body. Somatic mutations, on the other hand,

occur in individual cells of the body and are not widespread. ‘4

These genes are divided into oncogenes and tumor suppressor genes”:1 5

Oncogenes are altered forms of normal cellular genes called proto-oncogenes. Proto-

oncogenes encode a number of proteins such as growth factors, growth factor receptors,

  
 
 
  
  
 
  
 
 
  

P53
fl 1
14-3-3 0‘

 

Cydm A cyclin B/A
DNA + CDK? + CD02
damage G_2/M
1 G1 M
P53 —' ch;c_k2t
cyclin E
+ CDK2

cyclin D
+ CDK4 &. 6

Figure 1.1: The Cell Cycle. Key cyclins and Cdks involved in each phase are
shown in italics, checkpoints are shown underlined, and genes studied in this paper are
shown in bold. See text for details.

GTPases, and nuclear transcription factors that are involved in tightly controlled
signal cascades promoting cellular growth. When these proto-oncogenes escape
regulation, cellular growth is continuously supported and cells proliferate without control.
Proto-oncogenes can be transformed into oncogenes by a variety of mechanisms.
Mutations in the gene, whether deletions, insertions, or base-pair substitutions, can alter
portions of the protein product that are important in regulation, resulting in constitutive
activity. If the proto-oncogene is near a chromosomal break point, chromosome

translocations may place it under the control of a different, more active promoter or

create a constitutively active fusion protein. A prom-oncogene may also undergo
amplification, in which a not fully understood mechanism creates multiple tandem copies
of the same gene, resulting in increased gene product. RNA viruses may insert an

oncogene into a new cell or disrupt the regulation of an existing proto-oncogene. (See

14-16

recent reviews regarding oncogenes.) Alterations in many proto—oncogenes, such as

HER-Z/neu, ms, and myc, have been observed in human breast adenocarcinomas.17

Tumor suppressor genes also regulate the cell cycle but have an antagonistic role
to the growth-promoting prom-oncogenes. Since tumor suppressor genes inhibit cellular

proliferation, it is the loss of their function that leads to uncontrolled cell growth and

neoplasia.14’15’l 8 A cell has two copies of each gene and can still function normally
providing it has at least one functional copy of a tumor suppressor gene; however, if both
copies of a tumor suppressor gene are inactivated then the cell may become tumorigenic.
This idea was first developed by Knudson in 1971 and is thus called “Knudson's Two-Hit
Hypothesis.” Knudson was studying retinoblastoma and noted that familial cases were
much more likely to be bilateral and developed at an earlier age when compared to
sporadic cases. He proposed that two mutagenic events, or “hits,” were necessary to
cause retinoblastoma. Sporadic cases were rare as it was very unlikely that two
mutational events would occur in the same cell. However, patients with the familial form
of the disease had inherited a mutation from a parent and thus had the first “hit” present

in every cell of the body. Therefore, the patient would then only need one mutational

event to occur in any cell to develop the disease.19

Normal tumor suppressor gene function can be lost in a variety of ways.

Deletions, insertions, or base-pair substitutions can alter or halt protein expression. The

whole gene may be lost when a portion of a chromosome is deleted or an entire
chromosome is lost in non-disj unction during mitosis. An epigenetic process such as
hypermethylation, described in more detail in chapter 3, may silence the expression of an
otherwise normal gene. As mentioned earlier, a cell can function normally with one good
copy of a tumor suppressor gene. However, if the one good gene is lost through any of
the methods described above, or replaced with an additional copy of the non-functional
gene through such mechanisms as chromosomal non-disj unction and reduplication of the
remaining chromosome or mitotic recombination, then the cell may become tumorigenic.
One hallmark of a tumor suppressor gene is loss of heterozygosity (LOH), in which
normal cells are heterozygous for a genetic marker within or near the tumor suppressor
gene of interest. However, tumor cells only contain one allele of the marker, suggesting
that the individual harbored one functional copy and one nonfunctional copy of the tumor

suppressor gene, and that the tumor cells have lost the functional copy. (See recent

1435:” regarding tumor suppressor genes.)

reviews

Although the classic tumor suppressor gene model suggests that both copies of the
gene need to be inactivated in order for tumorigenesis to occur, recent evidence suggests
that haploinsufficiency, or loss of only one functional allele, may contribute to

tumorigenesis. This effect may be due to dosage sensitivity of a gene product, or may

affect the cell when combined with mutations in other oncogenes or tumor suppressor

geneszo’21 Few examples of haploinsufiiciency and tumorigenesis are currently known,
but future work in this field may discover new tumor suppressor genes despite the

absence of LOH.

1.3 Cats as Animal Models for Human Breast Cancer

Breast cancer is the most common malignancy of women, with one of every eight

women in the US. likely to develop the disease.17 There are many similarities between
feline mammary tumors and human breast cancers that make cats suitable animal models

for therapeutic trials and the study of mammary tumorigenesis. Adenocarcinoma is the

predominant type of mammary tumor in both cats and humans.2 Both species exhibit a

high rate of metastasis with the most common sites of metastasis being the regional
lymph nodes and the lungs.8 Women and cats most often develop mammary cancer

during middle age8 and tumor size is an important prognostic indicator in both.22 Unlike

canine mammary tumors, feline mammary tumors are responsive to chemotherapeutic

agents used to treat human tumors, particularly doxorubicin.23 Mammary carcinoma in
Siamese cats suggests a genetic component to the disease as their younger mean age of

tumor development when compared to other breeds is similar to the earlier onset of breast

cancer in women from families with a history of breast cancer.2 The main difference
between feline and human breast cancers is hormone receptor status: 70% of human

breast adenocarcinomas are estrogen receptor (ER) positive as compared to 10% of feline

2,22

mammary adenocarcinomas. Therefore, feline mammary tumors may provide an

excellent model for ER negative human breast tumors.

REFERENCES

1. Dom CR, Taylor DO, Schneider R, et a1. Survey of animal neoplasms in Alameda
and Contra Costa Counties, California. 11. Cancer morbidity in dogs and cats from
Alameda County. J Natl Cancer Inst 1968;40:307-318.

2. Hayes HM, Jr., Milne KL, Mandel] CP. Epidemiological features of feline
mammary carcinoma. Vet Rec 1981;108:476-479.

3. Withrow SJ, MacEwen EG. Tumors of the mammary gland. Small animal clinical
oncology. 2nd ed. Philadelphia: W.B. Saunders, 1996;356-372.

4. Hayes AA, Mooney S. Feline mammary tumors. Vet Clin North Am Small Anim
Pract 1985;15:513-520.

5. Misdorp W, Romijn A, Hart AA. Feline mammary tumors: a case-control study of
hormonal factors. Anticancer Res 1991;11:1793-1797.

6. MacEwen EG, Hayes AA, Harvey HJ, et a1. Prognostic factors for feline
mammary tumors. J Am Vet Med Assoc 1984;185:201-204.

7. Jeglum KA, deGuzman E, Young KM. Chemotherapy of advanced mammary
adenocarcinoma in 14 cats. J Am Vet Med Assoc 1985;] 87: 1 57-160.

8. Weijer K, Head KW, Misdorp W, et a1. Feline malignant mammary tumors. 1.
Morphology and biology: some comparisons with human and canine mammary
carcinomas. J Natl Cancer Inst 1972;49:1697-1704.

9. Muleya J S, Nakaichi M, Taura Y, et al. In-vitro anti-proliferative effects of some
anti-tumour drugs on feline mammary tumour cell lines. Res Vet Sci 1999;66:169-174.

10. Castagnaro M, Casalone C, Bozzetta E, et a1. Tumour grading and the one-year
post-surgical prognosis in feline mammary carcinomas. J Comp Pathol 1998;119:263-
275.

11. Vermeulen K, Van Bockstaele DR, Bememan ZN. The cell cycle: a review of
regulation, deregulation and therapeutic targets in cancer. Cell Prolif2003;36: 13 1-149.

12. Massague J. G1 cell-cycle control and cancer. Nature 2004;432:298-306.

13. Taylor WR, Stark GR. Regulation of the G2/M transition by p53. Oncogene
2001;20:1803-1815.

14. Knudson AG. Cancer genetics. Am J Med Genet 2002;111:96-102.

15. Vogelstein B, Kinzler KW. Cancer genes and the pathways they control. Nat Med
2004;10:789-799.

16. Park M. Oncogenes. The Geneticlbasis of human cancer. New York: McGraw-
Hill Health Professions Division, 1998;537-563.

17. Couch F J, Weber BL. Breast cancer. The Genetic basis of human cancer. New
York: McGraw-Hill Health Professions Division, 1998;537-563.

18. F earon ER. Tumor Suppresor Genes. The Genetic basis of human cancer. New
York: McGraw-l-Iill Health Professions Division, 1998;537-563.

l9. Knudson AG. Chasing the cancer demon. Annu Rev Genet 2000;34:1-19.

20. Santarosa M, Ashworth A. Haploinsufficiency for tumour suppressor genes: when
you don't need to go all the way. Biochim Biophys Acta 2004;1654:105-122.

21. Sherr CJ. Principles of tumor suppression. Cell 2004;116:235-246.

22. Hahn KA, Bravo L, Avenell J S. Feline breast carcinoma as a pathologic and
therapeutic model for human breast cancer. In Vivo 1994;82825-828.

23. MacEwen EG. Spontaneous tumors in dogs and cats: models for the study of
cancer biology and treatment. Cancer Metastasis Rev 1990;9:125-136.

10

CHAPTER 2

SUMMARY OF CASES

2.1 Description of Sample Population

Feline mammary adenocarcinoma cases for study were randomly selected from

surgical biopsy samples submitted to the Animal Health Diagnostic Laboratory (AHDL)l
from January 1, 1997 to April 30, 2000. 211 total cases were represented. The majority of
cats were Domestic Shorthair (1 77/83.9%) followed by Siamese (18/8.5%), Persian

(3/ 1.4%), Himalayan (1/0.5%), Maine Coon (1/0.5%), Ragdoll (1/0.5%), and Tonkanese
(1/0.5%). The breed was not reported for nine cases (4.3%). The median age of the
population was 11313.6 years. The majority of cases were female (196/92.4%) with 8
cases (3.8%) being male and 8 cases (3.8%) being unreported. Of the females, 138 of the
196 cases (70%) were ovariohysterectomized and 58 cases (30%) were intact. All males

were neutered.

2.2 Tumor Types

Generally, feline mammary adenocarcinomas can be classified based on three
patterns of proliferation. Papillary adenocarcinomas arise from the epithelium of
. mammary ducts and appear as papillary projections. Tubular adenocarcinomas also arise
from mammary duct epithelium but form tubules rather than papillary projections.
Lobular adenocarcinomas, also called acinar or alveolar adenocarciomas, form distinct
acini that may be divided by connective tissue septa into lobules. Often a tumor may

contain more than one proliferation pattern. Mammary carcinomas may also be classified

 

1 Currently known as the Diagnostic Center for Population and Animal Health

11

as solid carcinomas, which contain solid nodules or sheets of neoplastic epithelial cells
not arranged in one of the patterns described above. Mixed mammary tumors contain
neoplastic epithelial cells as well as neoplastic myoepithelial cells with differentiation
into cartilage and bone. A mammary carcinoma may be described as scirrhous if it is
accompanied by collagenous connective tissue proliferation.

Of the 212 cases, lobular adenocarciomas were the most common (82 or
38.7%) and papillary adenocarcinomas were the second most common (61 or
28.7%). There were no cases classified as strictly tubular adenocarcinomas although three
tumors (1.4%) were classified as having both tubular and papillary patterns. A summary

of tumor types for all 212 cases can be found in table 2.1.

2.3 DNA Isolation

Feline mammary adenocarcinoma samples were obtained as formalin-fixed
paraff'm-embedded samples. A veterinary pathologist (Dr. Yamini) examined
a section of each sample and outlined the tumor with permanent marker on the
corresponding paraffin block, thus demarcating mammary adenocarcinoma from normal
mammary tissue. Tissue was then excised from the middle of the indicated tumor region
as well as outside the indicated tumor region in order to obtain samples of
adenocarcinoma cells and normal cells from each sample.

A small section (3 mm in diameter and 0.5 mm thick) of either tumor or normal
tissue was obtained with a scalpel blade and placed in 400 pl of digestion buffer (50 mM
Tris pH 8.5, 1 mM EDTA, 0.5% Tween®). Each sample was then heated at 95°C for 10

minutes, pulsed in microwave twice for 30 seconds at high power, and cooled to room

12

temperature. Each sample was then digested by adding proteinase K and incubating at

42°C overnight. Each sample was then heated at 95°C for 10 minutes to inactivate the

proteinase K and centrifuged at 12,000 rpm for 10 minutes. An aliquot of this digested

lysate was used as a template for PCR or then modified with bisulfite treatment to study

methylation.

Table 2.1: Tumor Type Summary

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Tumor Type # Cases Percenta e
Lobular Adenocarcinoma 82 38.9
PapillaryAdenocarcinoma 61 28.9
Unspecified 15 7. 1
Adenocarcinoma
Lobular Adenocarcinoma + 12 5.7
Solid Carcinoma
Papillary + Lobular 11 5.2
Adenocarcinoma
Solid Carcinoma 9 4.3
Lobular + Scirrhous 5 2.4
Adenocarcinoma
Scirrhous Adenocarcinoma 4 1.9
Tubular + Papillary 3 1.4
Adenocarcinoma
Papillary Adenocarcinoma 3 1.4
+ Solid Carcinoma
Mixed Tumor 2 0.9
Papillary + Scirrhous 2 0.9
Adenocarcinoma
Scirrhous Adenocarcinoma 1 0.5
+ Solid Carcinoma
Adenocarcinoma w/ 1 0.5
Squamous Differentiation
Totals 211 100

 

 

13

CHAPTER 3

14-3-3 o

3.1 Background
14-3-3 a, hereafier called 0, is a member of the highly conserved 14-3-3 gene

family, which in mammals consists of at least seven isoforms with the general function of
facilitating protein-protein interactions.1 Human o is a 9876 bp gene (Genbank accession
AF 029081) consisting of one 1245 bp exon which produces a 25 kDa protein.2 The gene
is located on human chromosome 1p35.3 0 was first identified in 1992 and called human
mammary epithelial marker 1 (HMEI)2 while another group independently identified the
gene in 1993 and called it stratifin.4

0 protein is expressed only in epithelial tissues, especially those enriched in
stratified squamous keratinizing epithelium.4 Expression of 0 results in arrest of the cell

cycle at the G2/M checkpoint.3 The a promoter contains a p53 binding site and 0
expression is induced by p53, a transcription factor expressed in response to DNA
damage.3 Once expressed, a protein binds CDK2 and CDC2 in the nucleus, nansports
them out of the nucleus by means of a nuclear exporting signal and sequesters them in the

cytoplasm. CDK2 and CDC2, key proteins in the GZ/M checkpoint of the cell cycle, are

unable to phosphorylate nuclear proteins fiom the cytoplasm and thus the cell does not

proceed to mitosis.5 Human colorectal cancer cells deficient in (I initially arrest at the

G2/M checkpoint but are unable to maintain this state, undergoing cell death in "mitotic

catastrophe" as they enter mitosis.6 o deficient cells also show more frequent

l4

chromosomal aberrations (such as chromosomal breaks, end-to-end fusions, and

unbalanced translocations) than cells containing 0.7 Without a to maintain G2/M arrest,
these chromosomal changes go unrepaired, leading to genetic instability and increasing
the likelihood that these cells undergo carcinogenesis.

0 was first associated with human cancer in 1992 when the first report describing
the gene noted that while normal human mammary epithelial cells expressed 0 mRNA,

two cell lines fi'om spontaneous human mammary carcinomas expressed greatly reduced

amounts of 0 mRNA.2 In 1999, 0 was shown to be overexpressed in a mitoxantrone
resistant pancreatic adenocarcinoma cell line, leading the authors to suggest that

0 modulated protein kinase C which in turn upregulated proteins involved in drug

resistances Ferguson et al then reported in 2000 that 0 mRNA expression was
undetectable in 94% of primary human breast tumors examined, although loss of .
heterozygosity or mutations of the gene were extremely rare. However, 91% of primary
human breast tumors and breast tumor cell lines examined contained hypermethylated

CpG islands in the 5' portion of the 0 gene, silencing 0 expression (as further explained

in the following paragraph).9 Another study performed proteomic profiling of normal and
tumor human breast tissue and found that all primary breast tumors examined contained
an average of lO-fold less 0 protein than normal breast tissue. 10 Hypermethylation of

0 was also identified in human gastric cancer, colorectal cancer, and hepatocellular

cancer cell lines and was observed in 43% of primary gastric adenocarcinomas

examined.l ‘

Many genetic changes have been noted in carcinogenesis but it is a relatively new

15

observation that epigenetic changes such as hypermethylation also play a role. An
epigenetic phenomenon creates heritable states without altering the DNA nucleotide
sequence itself. Methylation is an epigenetic event that is part of the regulation of gene
expression in normal cells. Methylation occurs at cytosine nucleotides that are 5' to
guanine nucleotides in the genome; these sites of potential methylation are called Cst.
Cst are methylated by DNA methyltransferases (DNMTs). DNMTI , responsible for
maintenance of methylation, recognizes hemimethylated Cst and methylates the

unmethylated site. DNMTs responsible for de novo methylation have not been identified

although DNMT3a and B are strong candidates.12 Cst occur at a less then expected
frequency in the genome, but some stretches of DNA, called CpG islands, contain the
expected or higher than expected frequency of the dinucleotide. CpG islands are often
located in the promoter and 5' coding regions of genes and play in a role in regulation of
gene expression. Genes with methylated CpG islands are not transcribed; examples are
imprinted genes and genes on the inactive X chromosome in females. Genes with
unmethylated CpG islands are transcribed; these are often housekeeping genes. The
mechanisms of this transcriptional control are not completely understood, but it is known
that the chromatin surrounding methylated CpG islands is in a "closed", transcriptionally
inactive state characterized by de-acetylated histones. Chromatin surrounding
unmethylated CpG islands, on the other hand, is in a transcriptionally favorable state
characterized by acetylated histones. Levels of methylation within each CpG island are
variable, but even a small amount of methylation results in a significant decrease in gene
expression. In one experimental system, methylation of 7% of Cst within a CpG island

resulted in 67-90% reduction in gene expression while higher levels of methylation

16

silenced the gene completely.l3 For general reviews of methylation, see Momparler and

' Boveni, 2000;l4 Baylin and Herman, 2000;15 and Robertson and Jones, 2000.16
Methylation can contribute to carcinogenesis in two ways. The first is mutation;
methylated Cst are hotspots for mutation as the 5-methylcytosine may deaminate to
form thymine, resulting in a C to T transition mutation. The second is hypermethylation,
in which normally unmethylated CpG islands are methylated by an unknown mechanism,
thus silencing expression of a normally expressed gene. If this silenced gene is a tumor
suppressor, the cell now lacks this protein and may begin to uncontrollably proliferate,
resulting in a tumor. Hypermethylation may occur in both copies of a tumor suppressor
gene, or may silence one copy while the other is inactivated through mutation or deletion.
Hypermethylation of several tumor suppressor genes has been noted in several types of
human cancers and demethylating chemotherapeutic agents such as 5-azacytidine and
decitabine are currently in clinical trials (see http://www.cancernet.nci.nih.gov).
Hypermethylation has not been studied in the cat. Further characterization of this
process is the first step in determining if demethylating chemotherapeutic agents will also
be effective in the cat. As hypermethylation of o is a common event in human breast
cancers and feline and human breast cancers share many similarities, I hypothesize that
hypermethylation of o is a frequent event in feline mammary cancers and may represent

an important target for intervention.

3.2 Materials and Methods

3.2.1 Sequencing Feline 14-3-3 o

0 has been sequenced in the human, mouse, and sheep and is highly conserved

17

between these species. Figure 3.1 demonstrates the homology between the human and
mouse 0 sequences, with each vertical line representing a conserved nucleotide.

. As feline a has not been previously sequenced, the most highly conserved regions

between the sequenced species were used to design several primer sets with the intention

of amplifying o exon 1 from feline DNA by polymerase chain reaction (PCR). During

primer design, care was taken to not place the 3' end of the primer at the 3rd codon
position, as this is most likely to be variable, decreasing the likelihood of amplification.
These primer sets were then used to amplify both human and feline genomic DNA
isolated from white blood cells in order to optimize PCR conditions. Primers used most
successfully to amplify a from the human and the cat are shown in figure 3.1 and were
used in a 25 pl PCR reaction containing 2.5 mM MgC12, 2.5 U Taq, and 0.4 uM each
primer. The PCR conditions consisted of 40 cycles of 94°C for 1 minute, 60°C for 2
minutes, and 72°C for 3 minutes. PCR products were visualized on 1% agarose gels and
appropriately sized bands were excised with a clean scalpel blade. DNA was purified
using the Qiaex II bead kit (Qiagen) according to the manufacturer's directions. DNA was
sequenced on an ABI 377 automated sequencer (Applied Biosystems) using Big Dye
terminators (Applied Biosystems). The feline sequence was then examined for Cst in
order to assess the potential of feline o for methylation.
3.2.2 Methylation Status of Feline Mammary Adenocarcinoma Cases

In order to assess methylation of normal and tumor tissue samples from feline
mammary adenocarcinoma cases (as described in Chapter 2), DNA was treated with
bisulfite and then sequenced. Bisulfite treatment chemically modifies cytosine

nucleotides of DNA to uracil, but does not modify methylated cytosines. During

18

Figure 3.1: Human and Mouse 0. Vertical lines represent conserved nucleotides.

Primers designed for conserved regions are shown. Genbank accession numbers:
AF 029081 (human), AF 058798 (mouse).

Human

Mouse:

Human:

Mouse:

Human:

Mouse:

Human:

Mouse:

Human:

Mouse:

Human:

Mouse:

Human:

Mouse:

Human:

Mouse:

Human:

Mouse:

Human:

Mouse:

Human:

Mouse:

Human:

Mouse

:8637

163

8697

223

8757

283

8817

343

8877

403

8937

463

8997

523

9057

583

9117

643

9177

703

9237

763

9297

:823

gagagagccagtctgatccagaa

a
catggagagagccagtctgatccagaaggccaagctggcagagcaggccgaacgctatga
llllllllllllllllIlllllllllllllllll llll II Illllllllll Illll
catggagagagccagtctgatccagaaggccaagttggctgaacaggccgaacggtatga

 

ggacatggcagccttcatgaaaggcgccgtggagaagggcgaggagctctcctgcgaaga
lllllllllll IIIIIIII IIIIIIIIII IIlIIlIIllIIlllllllllll ll
agacatggcagctttcatgaagagcgccgtggaaaagggcgaggagctctcctgcgagga

gcgaaacctgctctcagtagcctataagaacgtggtgggcggccagagggctgcctggag

|||||||||l|| || llllllll lllllllllllllllllllllll ll ||||||||
gcgaaacctgctttccgtagcctacaagaacgtggtgggcggccagagagcggcctggag

ggtgctgtccagtattgagcagaaaagcaacgaggagggctcggaggagaaggggcccga
lII llllllll ll llllllll llllllllllllll II II llllllll lllll
ggtcctgtccagcatcgagcagaagagcaacgaggaggggtcagaagagaagggccccga

ggtgcgtgagtaccgggagaaggtggagactgagctccagggcgtgtgcgacaccgtgct
IIII lllllllllllllllll lllll IIIIII II llllllllllllll II
ggtgaaagagtaccgggagaaggtagagaccgagctcagaggtgtgtgcgacaccgtact

gggcctgctggacagccacctcatcaaggaggccggggacgccgagagccgggtcttcta

lIlllIIllIll lllllllllll I III II II II llllllll Illlllll
cggcctgctggactcgcacctcatcaaaggggctggagatgcagagagccgcgtcttcta

cctgaagatgaagggtgactactaccgctacctggccgaggtggccaccggtgacgacaa

lllllllllllllllllllllllllllllllll Illlllllllllll II II Illll
cctgaagatgaagggtgactactaccgctacctagccgaggtggccactggcgatgacaa

gaagcgcatcattgactcagcccggtcagcctaccaggaggccatggacatcagcaagaa

llllllllllll II II llllllllllllllIIlllllllllllllllllllllllll
gaagcgcatcatcgattctgcccggtcagcctaccaggaggccatggacatcagcaagaa

ggagatgccgcccaccaaccccatccgcctgggcctggccctgaacttttccgtcttcca

Illlllllllll llllllllllllllllllllllllllllllllllllll IIIIIIII
ggagatgccgcctaccaaccccatccgcctgggcctggccctgaacttttcagtcttcca

ctacgagatcgccaacagccccgaggaggccatctctctggccaagaccactttcgacga

lllllllll IllllllllllllllllllllllIll Illlllllllllll llllllll
ctacgagatagccaacagccccgaggaggccatctcgctggccaagaccaccttcgacga

ggccatggctgatctgcacaccctcagcgaggactcctacaaagacagcaccctcatcat

lllllllll ll llllllllllllll llllllllllllll lllllllllllllllll
ggccatggccgacctgcacaccctcagtgaggactcctacaaggacagcaccctcatcat

gcagctgctgcgagacaacctgacactgtggacggccgacaacgccggggaagagggggg
Illlll III IIIIIIIIIIIII IIIIIIII lllllll ll lllllllllll II
gcagctcctgagagacascctgacgctgtggacagccgacagtgctggggaagagggtgg

 

tggactgtgacacctgccg

19

subsequent PCR the uracils are converted to thymine. Samples are then sequenced and
the methylation status can be determined by examining each CpG (the locations of which
were determined in previous sequencing of an unmodified sample) for the presence of a

cytosine or a thymine. Formalin-fixed parafi'm—embedded tissue has been shown to be a

useful source of DNA for this technique as the fixation does not affect methylation.l7

In order to perform bisulfite treatment, DNA was isolated from formalin-fixed
paraffm-embedded tissue samples as described in chapter 2. Approximately 1 pg of DNA
and 20 pg glycogen were incubated in 0.2 M NaOH (50 pl total volume) for 10 minutes
at 37°C. 30 pl of freshly prepared lOmM hydroquinone and 520 pl of freshly prepared
3.5 M sodium bisulfite were added to each sample and incubated under mineral oil for 16
hours at 50°C. Samples were purified with Wizard PCR Preps (Promega) and eluted with
50 pl water. Samples were then treated with NaOH (0.3 M final concentration), incubated
for 5 minutes at room temperature, and ethanol precipitated.

Once modified, the DNA was amplified with primers designed for the bisulfite-
modified sequence. Primers were designed for regions without Cst in order to eliminate
potential sequence variations between methylated and unmethylated DNA that may have
interfered with amplification. A heminesting PCR strategy was used to amplify the
bisulfite modified DNA as shown in figure 3.2. In the first PCR, 2.5 pl — 5 pl of the

modified DNA sample was added to a 25 pl reaction containing 2.5 mM MgClz, 0.4 pM
. primer 01, and 0.4 pM primer 02. The reaction was boiled for 3 minutes, 0.25U Taq was
added, and the reaction was cycled 35 times at 94°C for 1 minute, 54°C for 2 minutes,
and 72°C for 3 minutes. 1 pl of this completed PCR was then added to a second 25 pl

reaction containing primers 01 and 03. All conditions were the same as the first PCR

20

except the annealing temperature of 56°C. The final amplified product was 325 bp and
contained 17 CpG sites.

Figure 3.2: Primers Used in the Heminesting PCR of Bisulfite-Modified 0. Primers
are shown in bold and modified Cst are shown underlined.

Primer a]

 

$

eacacacrracrrrcaarracaacGTTAAGTTGGTAGAGTAGGTIEAAEQTTAEQAGGAT
CTCTCTCAATCAAACTAAATCTTCCAATTCAACCATCTCATCCAACTTACAATACTCCTA

ATGGTAGTTTTTATGAAGAGTGTTGTGGAAAAGGGTGAGGAGTTATTTTGTGAAGAGTGA
TACCATCAAAAATACTTCTCACAACACCTTTTCCCACTCCTCAATAAAACACTTCTCACT

AATTTGTTTTTAGTGGTTTATAAGAATGTGGTGGGTEGTTAGAGGGTTGTTTGGAGGGTT
TTAAACAAAAATCACCAAATATTCTTACACCACCCACCAATCTCCCAACAAACCTCCCAA

TTGTTTAGTATTGAGTAGAAAGGTAATGAGGAGAGTTTGGAAGAGAAGGGTTTGGAGGTG
AACAAATCATAACTCATCTTTCCATTACTCCTCTCAAACCTTCTCTTCCCAAACCTCCAC

TGAGAGTATTGGGAGAAGGTGGAGATTGAGTTTTGGGGTGTGTGTGATATGGTGTTGGGT
ACTCTCATAACCCTCTTCCACCTCTAACTCAAAACCCCACACACACTATACCACAACCCA

TTGTTGGATATTTATTTTATTAAGGAGGTEQGTGAZETEEAGAGTIE // 286 bp //
AAcaaccramaaamaaaaraarrccrCCAACCACTACAACTCTCAAC

‘

 

Primer a3

TTATAAAGATAGTATTTTTATTATGTAGTTGTTGTGAGATAATTTGATATTGTGGATGGT
AATATTTCTATCATAAAAATAATACATCAACAACACTCTATTAAACTAIAACACCTACCA

4

 

Primer 02

Once amplified, the PCR product was visualized on a 1% agarose gel and
appropriately sized bands were excised with a clean scalpel blade. The excised DNA was
purified with the Qiaex II bead kit (Qiagen) according to the manufacturer's directions.
DNA was then sequenced on an ABI 377 automated sequencer (Applied Biosystems)
using Big Dye terminators (Applied Biosystems). The status of each amplified CpG was
examined for the presence of a cytosine or a thymine in order to determine the

methylation status, with a cytosine indicating complete methylation, a thymine indicating

21

no methylation, and the presence of both nucleotides indicating partial methylation.
3.2.3 Statistical Analysis

Methylation of each tumor and corresponding normal tissue sample was
quantified as the number of CpG sites observed to be methylated. The number of CpG
sites was the same for each sample. The number of methylated CpG sites was then

compared between a tumor sample and the corresponding normal tissue sample using the

paired t-test.18

3.3 Results
3.3.1 Feline 14-3-3 0 Sequence

575 base pairs of the first exon of feline a were sequenced. Figure 3.3 shows this
~ sequence as compared to the homologous human 0 sequence (Genbank accession
AF 029081). The entire human 0 first exon is 744 bp, as represented by nucleotides 8638
- 9381 . Overall, human and feline 0 are highly conserved, with 94% nucleotide identity
and 92% amino acid identity.

Feline or contains 37 CpG sites in the sequenced region as compared to 34 in
human 0. 27 Cst are conserved between the cat and the human (73% of the total cat
Cst). There are 10 new CpG sites in the cat, representing 27% of the total cat Cst. 7
CpG sites appear in the human but not the cat, representing 21% of the total human

Cst. Figure 3.3 illustrates these CpG sites.

22

Figure 3.3: Human and Feline a Exon 1. Vertical lines represent conserved
nucleotides. Conserved CpG sites are shown in bold and unconserved sites are
underlined.

Human:

Cat:

Human:

Cat:

Human:

Cat:

Human:

Cat:

Human:

Cat:

Human:

Cat:

Human:

Cat:

Human:

Cat:

Human:

Cat:

Human:

Cat:

8674

1

8734

61

8794

121

8854

181

8914

241

8974

301

9034

361

9094

421

9154

481

9214

541

GCAGAGCAGGcggAAggcrATGAGGACATGGCAGCCTTCATGAAAGGggcggTGGAGAAG

llllllllllllllllllll lllllllllllllllllllllll ||||l||||| Ill
GCAGAGCAGGCQQAAQQCTAEEAGGACATGGCAGCCTTCATGAAGAGggcggTGGAAAAG

GGEEAGGAGCTCTCCTGggAAGAGggAAACCTGCTCTCAGTAGCCTATAAGAAEETGGTG

ll |||||||| IlllllllllllIllllllllllllllll ||||| lllll ||||||
GGTGAGGAGCTATCCTGggAAGAGggAAACCTGCTCTCAGTGGCCTACAAGAATGTGGTG

GGQQGCCAGAGGGCTGCCTGGAGGGTGCTGTCCAGTATTGAGCAGAAAAGCAAQQAGGAG

||||||||||||l||||||||||||| ||||||l|||| |l||||||| IIIITTIIIII
GGggGCCAGAGGGCTGCCTGGAGGGTCCTGTCCAGTATEEAGCAGAAAGGCAAQQAGGAG

GGCngGAGGAGAAGGGGCCCGAGGTGggTGAGTAcggGGAGAAGGTGGAGACTGAGCTC

lllllll ||||l||l ll llllllll Illlllllllllllllllllllllllllll
AGCngGAAGAGAAGGGCCQQGAGGTcggaGAGTAcggGGAGAAGGTGGAGACTGAGCTC

CAGGGggTGTGCGACACQQTGCTGGGCCTGCTGGACAGCCACCTCATCAAGGAGGcggGG

l ||ll|l|l|_T|||| lllllllllllllllllll lllllllllllllllllllll

EGGGGggTGTGTGACAEEGTGCTGGGCCTGCTGGACACCCACCTCATCAAGGAGGCQQGT

GAggcggAGAGcggGGTCTTCTACCTGAAGATGAAGGGTGACTACTACQQCTACCTGGCC

||||||l|||| lllllllllllllllll Illlllll llllllllllllllllllll
GAQQCQQAGAGTQQGGTCTTCTACCTGAAAATGAAGGGEEACTACTACQQCTACCTGGCT

GAGGTGGCCACEGGTGAggACAAGAAGQQCATCATTGACTCAGCCQQGTCAGCCTACCAG

lllllllllll llllllllllllllllllll|||||l||| llllllll |||||||||
GAGGTGGCCACTGGTGAggACAAGAAGQQCATCATTGACngcccggGTCEECCTACCAG

GAGGCCATGGACATCAGCAAGAAGGAGATGCQgCCCACCAACCCCATCQQCCTGGGCCTG

||||l||||llll|llIllll|||||IIIlllllllll|l||||||||||||||||||||
GAGGCCATGGACATCAGCAAGAAGGAGATGCQQCCCACCAACCCCATCQQCCTGGGCCTG

GCCCTGAACTTTTCEQTCTTCCACTAggAGATQQCCAACAGCCcggAGGAGGCCATCTCT

|| Illllllllll lllllllllllllllllllllllllllllllllllllllll|||
cggcrGAACTTTTCAGTCTTCCACTAggAGATQQCCAACAGcccggAGGAGGCCATCng

CTGGCCAAGACCACTTngAggAGGCCATGGCTGA

llllllllllllll llllllllllllllllllll
CTGGCCAAGACCACCTngAggAGGCCATGGCTGA

23

3.3.2 Feline Mammary Carcinoma Methylation Status

Paired normal and tumor DNA from 5 cases was examined for methylation status
of the o CpG island. Eight unpaired samples (three normal and five tumor DNA) were
also examined. Details of each case examined can be found in table 3.1. For each sample,
each CpG site was determined to be methylated, partially methylated, or unmethylated.
For each statistical calculation, partially methylated sites were counted as 0.5 completely
methylated sites counted as 1. Results are summarized in Tables 3.2 (paired samples) and
3.3 (unpaired samples).

For paired samples, the number of methylated Cst in the normal sample was
subtracted from the number of methylated Cst in the tumor sample in order to calculate
the difference between the two. The differences were then analyzed with a t-test using
SAS software. The two groups were found not to be significantly different (P=0.1671).
All samples were then divided into normal DNA and tumor DNA groups and the mean
number of methylated Cst for each group was compared with a t-test using SAS
software. Again, the two groups were not found to be significantly different (P=0.3565).
In conclusion, there is no significant difference in methylation of 0 between normal

feline mammary tissue and feline mammary carcinomas.

3.4 Discussion
In this study it was found that 14-3-3 a is not hypennethylated in feline

mammary carcinomas. This is in contrast to human breast tumors, which have been found

to have 0 hypermethylation in >90% of examined cases.9 In human breast tumors with

hypermethylation, all or almost all CpG sites were found to be completely methylated,

24

Table 3.1: Cases Examined for a Hypermethylation. F = female, F /S = female/spayed,

NR = not reported.

 

Paired Samples

' Case Breed Age Tumor Type Malignancy Prognosis Comments
# (years)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10 Mixed 12 F NR High Guarded High mitotic index,
cellular atypia

13 NR NR NR Acinar High Poor Recurring tumor,
high mitotic index

121 Mixed 15 F/S Acinar/Papillary Low Guarded

143 Mixed 16 NS Acinar/Solid High Poor Recurring tumor,
locally invasive,
vascular invasion

158 Mixed 11 F / S Papillary Moderate Guarded

Unpaired Samples: Normal DNA Examined

12 Mixed 10 F/S Acinar High Guarded Locally invasive

16 Persian 12 HS NR NR Guarded High mitotic index,
cellular atypia

151 Mixed 10 F/S Papillary High Guarded Locally invasive

Unpaired Samples: Tumor DNA Examined

130 Mixed 12 F/S Acinar/Papillary NR Guarded Multinodular,
multicystic

133 Mixed 8 F/S Papillary High Guarded Locally invasive

139 Siamese 7 F/S Acinar/ High Poor Vascular invasion

Scirrhous

149 Mixed 10 F Solid High Poor Vascular invasion

166 Mixed 8 F/S Papillary/Solid High Poor Possibly recurring
trunor, vascular
invasion

 

25

 

Table 3.2: Number of Methylated Cst in Paired Feline Mammary Carcinoma
Samples. P = partially methylated, C = completely methylated.

 

 

 

 

W N...
Sam-1e

10 0 4P, 1C 4P, 1C

13 4P 4P, 2C 2C

121 0 0 0

143 0 3P 3P

158 3P 2P -1P

 

 

 

 

 

 

Table 3.3: Number of Methylated Cst in Unpaired Feline Mammary Carcinoma
Samples. P = partially methylated, C = completely methylated

12
16
151

130
133
139
149
166

# Sites

Normal
0
6P 1C
0
Tumor
0
IP
IP
3P 2C
4P

 

lated

while normal breast tissue exhibited at most one methylated site.9 In the feline cases

examined in this study, a maximum of 2 CpG sites were found to be completely

methylated and a maximum of 6 sites were found to be partially methylated out of the 14

total CpG sites analyzed. Unlike human cases, 0 methylation in the cat did not correlate

with tumor samples: some tumor samples exhibited no methylated Cst, and the sample

with the most methylated sites (six partial and one complete) was from normal tissue.

The process of bisulfite modification and amplification of DNA is technically

challenging and therefore the number of samples successfully analyzed for o methylation

26

in this study is small. However, the sample population is representative of the larger
population of feline mammary tumor biopsies submitted to AHDL, covering a variety of
tumor types and characteristics. Several of the samples were highly malignant, increasing
the likelihood of identifying 0 hypermethylation even if it were a late-occurring event in
feline mammary tumorigenesis.

Several questions concerning feline o and feline methylation are unanswered in
this study. It is unknown whether 0 is normally expressed in the cat, and the finding that

0 is expressed in rat but not mouse mammary tissue indicates there is variation between

species.2 If 0 expression does occur in the eat, this study does not completely rule out
0 involvement in feline mammary tumorigenesis and further studies need to be done to
characterize the role of this gene in the cat. This is the first reported study of
hypermethylation and feline carcinogenesis, and more work needs to be done in the cat

on this epigenetic phenomenon that plays a role in several human cancers.

27

REFERENCES

1. Wang W, Shakes DC. Molecular evolution of the 14-3-3 protein family. J Mol
Evol 1996;43:384-398.

2. Prasad GL, Valverius EM, McDuffie E, et a1. Complementary DNA cloning of a
novel epithelial cell marker protein, HME 1 , that may be down-regulated in neoplastic
mammary cells. Cell Growth Difi'er 1992;3z507-513.

3. Henneking H, Lengauer C, Polyak K, et al. 14-3-3 sigma is a p53-regulated
inhibitor of G2/M progression. Mol Cell 1997;] :3-1 1.

4. Leffers H, Madsen P, Rasmussen HH, et a]. Molecular cloning and expression of
the transformation sensitive epithelial marker stratifm. A member of a protein family that
has been involved in the protein kinase C signalling pathway. J Mol Biol 1993;231:982-
998.

5. Laronga C, Yang HY, Neal C, et a]. Association of the cyclin-dependent kinases
and 14-3-3 sigma negatively regulates cell cycle progression. J Biol Chem
2000;275 :23 106-231 12.

6. Chan TA, Henneking H, Lengauer C, et al. 14-3-3Sigma is required to prevent
mitotic catastrophe after DNA damage. Nature 1999;401:616-620.

7. Dhar S, Squire JA, Hande MP, et al. Inactivation of 14-3-3sigma influences

telomere behavior and ionizing radiation-induced chromosomal instability. Mol Cell Biol
2000;20:7764-7772.

8. Sinha P, Hutter G, Kottgen E, et al. Increased expression of epidermal fatty acid
binding protein, cofilin, and 14-3-3-sigma (stratifm) detected by two-dimensional gel
electrophoresis, mass spectrometry and microsequencing of drug-resistant human
adenocarcinoma of the pancreas. Electrophoresis 1999;20:2952-2960.

9. Ferguson AT, Evron E, Umbricht CB, et al. High frequency .of hypermethylation
at the 14-3-3 sigma locus leads to gene silencing in breast cancer. Proc Natl Acad Sci U S
A 2000;97:6049-6054.

10. Vercoutter-Edouart AS, Lemoine J, Le Bourhis X, et a]. Proteomic analysis

reveals that 14-3-3sigma is down-regulated in human breast cancer cells. Cancer Res
2001;61:76-80.

11. Suzuki H, Itoh F, Toyota M, et al. Inactivation of the 14-3-3 sigma gene is

associated with 5' CpG island hypermethylation in human cancers. Cancer Res
2000;60:4353-4357.

28

12. Hendrich B, Bird A. Mammalian methyltransferases and methyl-CpG-binding
domains: proteins involved in DNA methylation. Curr T op Microbiol Immunol
2000;249:55-74.

13. Hsieh CL. Dependence of transcriptional repression on CpG methylation density.
Mol Cell Biol 1994;14:5487-5494.

14. Momparler RL, Bovenzi V. DNA methylation and cancer. J Cell Physiol
2000;183:145-154.

15. Baylin SB, Herman JG. DNA hypermethylation in tumorigenesis: epigenetics
joins genetics. Trends Genet 2000;16:168-174.

16. Robertson KD, Jones PA. DNA methylation: past, present and future directions.
Carcinogenesis 2000;21:461-467.

17. Kitazawa S, Kitazawa R, Maeda S. Identification of methylated cytosine from
archival formalin-fixed paraffm-embedded specimens. Lab Invest 2000;80:275-276.

18. Rao PV. Statistical research methods in the lifiz sciences. Pacific Grove, CA:
Duxbury Press, 1998.

29

CHAPTER 4
P53
4.1 Background
P53 is an extensively studied tumor suppressor gene. In the human, this

approximately 20 kb gene located on chromosome l7pl3 consists of 11 exons and
encodes a 393 amino acid nuclear phosphoprotein.l The 53 kDa protein was first reported

in 1979 as a component of cells transformed by simian virus 40 (SV40).2 In 1990,
germline P5 3 mutations were discovered to be the cause of Li-Fraumeni syndrome, an

autosomal dominantly inherited disorder in which members of affected families develop

one or more types of cancer at an early age of onset.I By the 19903, tumor suppressor
genes in general and P53 in particular were enthusiastically studied by many in order to

better understand the development and treatment of cancer, thus earning P53 the title

"Molecule of the Year" from Science in 1993.3 In the year 2003 alone there were 3399
Medline citations found with the keyword search "P53".

P53 normally functions as part of the Gl/S checkpoint in the cell cycle, causing
cell cycle arrest and cellular apoptosis in the presence of DNA damage. DNA damage

induces the expression of P53, which then acts as a transcription factor in the nucleus and

affects expression of several other genes including p21, MDM2, and Bax,4 which in turn
regulate the cell cycle and apoptosis. P53 also induces genes involved in the G2/M

checkpoint such as 14-3-3 o: The normal cell only contains a small amount of P53 as the

protein has a short half life and is targeted for degradation by the protein MDM24. The
accumulation of P53 in a cell, as detected by immunocytochemical staining, is indicative

of a missense mutation which stabilizes the protein and reduces its ability to induce

30

MDM2.5

Alterations in P5 3 have been found in 50-55% of human cancers.6 These

alterations have been found in a variety of cancers, including 75-80% of colorectal

tumors2 and 30-40% of breast tumors.7 P53 may be altered through deletions, insertions,
base pair substitutions, chromosomal loss, or other mechanisms affecting tumor
suppressor gene expression as discussed in Chapter 1. Many point mutations have been
reported, with most being missense mutations resulting in an altered protein (rather than
nonsense mutations resulting in a truncated protein). These mutations are clustered

between amino acids 130-290, with most occurring within four domains that are highly

conserved between many species.2 Amino acid residues 175, 248, and 273 frequently

contain point mutations and have been dubbed mutational "hot spots".2

Changes in P53 have also been extensively studied in spontaneous breast cancer
cases in order to determine if the presence of P53 mutation is useful as a prognostic

factor or if it can be used to predict a response to various therapies. Several studies have

found that human patients with P53 overexpression have a poorer prognosis.8 There have
been variable results in studies looking at P53 status and therapy efficacy. Some studies
have found that P53 alteration is predictive for resistance against tamoxifen, doxorubicin,

and radiotherapy, while other studies have not found predictive value for these

treatments.9 More studies specifically assessing P53 mutations need to be done in order

to obtain a true picture on the usefulness of P53 status as a factor in therapy selection.
The involvement of P53 in feline cancers has just begun to be studied. Feline P53

mRNA has been sequenced (Genbank accession D26608) and shows 86% nucleotide

conservation with human P53 mRN A (Genbank accession NM000546). Feline P53

31

introns 5, 6, and 7 have also been sequenced (Mayr et a]. 10, Genbank accessions U81292,

and U81298, respectively. See figure 4.] for sequences.) Using feline x rodent somatic

cell hybrids, feline P53 has been mapped to feline chromosome E

I.“

Figure 4.1: Feline P53 Genomic Sequence: Exon 5 - Exon 8. Uppercase = exons,
lowercase = introns. SNP locations are underlined.

1

51
101
151
201
251
301
351
401
451
501
551
601
651
701
751
801
851
901
951
1001
1051
1101

TACTCCCCTC
CGTGCAGCTG
CCATGGCCAT
TGTCCCCACC
gggctacaga
cctccccgat
GGAAGGAAAC
ATAGCGTCGT

ggtctctggg
gagatggggg
cggtgtgcac
gcctacacac
tCttthtCC
TGTGTAACAG
ATCATCACCC
aggccactct
tgtggaatct
ctctggcttt
tttaggctcc
tccctcactg
CTGGGACGGA
CCGGCGCACC
AGCCGCCC

CCCTCAACAA
TGGGTCCGAT
TTAQAAGAAG
ACGAGCGCTG
tggggcaggg
tgctctcagG
TTGCATGCCA
GGTGCCCTAC
aggaggtggq
gggctttctc
agccagccgg
tgcaggcctg
cagGTCGGCT
TTCCTGCATG
TGGAAGACTC
ctcccgtgct
cctctgctgt
gggaccttct
acataggatg
cctccagcgt
ACAGCTTCGA
GAGGAGGAAA

GCTGTTTTGC
CGCCGCCCCC
TCAGAGTTCA
CCCTGACAGT
cctgctgcta
TCTGGCGCCT
AGTACCTGGA
GAGCCGCCCG
ggaggggttt
cttcttatgc
gtggtcccca
cccggcgctg
CTGACTGTAC
GGGGGCATGA
CAAgtaggga
accgcccatc
cccccaccct
cttacccggc
aaggaggtgg
ctgtcttctt
GGTACGAGTT
ATTTCCGCAA

CAGCTGGCGA
ACCGGGAACC
TGACAGAGGT
AGCGATGggt
gggtcccccg
CCCCAGCATC
CGACAGAAAC
Athctgctt
gtcagcggcc
aacctcccca
gtgcacggtt
ggtggcctca
CACCATCCAC
ACCGGAGGCC
cccgcaggcc
ccgcctgtgg
ccgcctccaa
ttctcgatac

ggagtaaggg
ac gtgggtag
TGTGCCTGTC
GAAGGGGGAG

AGACCTGCCC
TGTGTCCGCG
CGTGAGGCGC
gagccgtcgg
gcccctgatt
TCATCCGAGT
ACTTTCCGAC
tggcatctgg

gtccaggtgg
cggcggcgtg
gaggaaacca
ctcggccgga
TACAATTTCA
CATCATCACC
accctgcccc
aatccccgcc
gttttctttt
tccttaggct
gggccccatc
TGGGAAGCTG
CTGGGAGAGA
CCTTGCCCTG

Several different types of feline neoplasms have been surveyed to determine if

P53 plays a role in feline cancer. Seventy-seven feline tumors of seven different types

were examined with immunocytochemical analysis and 20 were positive for P53 staining,

including 3 of 9 mammary carcinomas.5 P53 involvement in vaccine-associated feline

sarcomas has been of interest, with one study finding 7 of 18 informative cases (39%)

showing loss of heterozygosity12 and another study reporting 8 of 21 cases (38%)

32

showing dark P53 immunostaining. '3 One group in particular at the Veterinary
University of Vienna, Austria has examined many different feline tumors for P5 3

mutations through PCR and sequencing. Of the 29 mammary carcinomas this group has

10,14-

reported studying, 3 had mutations. '7 One had a missense mutation resulting in an

arginine to tryptophan amino acid transition at codon 282,14 one had a missense mutation
resulting in an arginine to cysteine amino acid transition at codon 158,17 and one had a 9

bp deletion affecting codons 251-256. 1°

Loss of heterozygosity (LOH) is a hallmark of tumor suppressor gene
involvement in tumorigenesis. A cell needs at least one functioning copy of a tumor
suppressor gene to control growth. If that one functioning copy of the gene is lost, then
the cell may undergo uncontrolled cell growth and tumorigenesis. In LOH, a
heterozygous marker is identified in or near the gene of interest. Marker alleles are
compared between normal tissue and tumor tissue. If a tumor tissue marker is .
homozygous as compared to the heterozygous normal tissue, than LOH has occurred. If
one marker allele has disappeared, it is an indication that the functioning tumor
suppressor gene allele has been lost, leaving behind a gene copy that is not functioning
due to changes such as mutation or hypermethylation.

In order to assess loss of heterozygosity, one must identify a heterozygous marker
in or near the tumor suppressor gene of interest. Single nucleotide polymorphisms, or
SNPs, have become a very useful marker for this purpose. A SNP is a locus in genomic
DNA in which different alleles exist for a single base pair in normal individuals of a
population. SNPs are plentiful, with more than 1,400,000 SNPs reported in the NCBI

database for the human genome and more being published all the time. Ofien SNPs are

33

located in introns or other non-coding regions, thus creating silent changes that can be
genotyped through methods such as sequencing or assessing restriction site changes.

A study concentrating on the role of P5 3 alterations in feline mammary carcinoma
has not been published. As feline mammary carcinomas are similar to human breast

tumors, I propose that the two cancers have a similar rate of P5 3 loss of heterozygosity.

4.2 Materials and Methods
4.2.1 Development of SNP Genotyping Tests

Genbank and the literature were searched for polymorphisms within feline P5 3 ,
listed in table 4.1. The reported minor allele frequencies were calculated from small
European sample populations. Polymorphism base pair positions are given according to

the nucleotide numbering in figure 4.].

Table 4.1: Reported Polymorphisms in Feline P53. NR = Not Reported.

Polymorphism Location Minor Allele
Fre . uenc

    
 

 

Position

   

 

 

 

 

 

 

C/TSNP Exon5 114 T=NR Mayr, 199510
C / T SNP Intron 6 495 T = 0.2 Mayr, 199816
C / T SNP Intron 7 737 T = 0.5 Mayr, 199316
T / C SNP Intron 7 969 C = NR Kanjilal, 199912
T insertion Intron 7 970 T insertion = NR Genbank
AF 175762
C / T / G SNP Intron 7 982 T = 0.5, G = NR Mayr, 199316;
Kanjilal, 199912

 

 

 

 

 

 

Restriction enzymes recognize specific short sequences of DNA and cleave the

34

DNA at this restriction site. The polymorphisms were examined to determine if they were
part of a restriction site. Polymorphisms 495, 73 7, and 970 had one allele that created a
restriction site while the other allele did not. The details of the genotyping test for each
polymorphism are given in section 4.3.1, but the general design plan is as follows.
Primers were designed to amplify a region of DNA by PCR containing the polymorphic
restriction site. In order to provide a control for the restriction enzyme digestion, the
amplified fragment also contained a restriction site that does not contain a polymorphism
and is thus always cleaved. The amplified DNA was then incubated with the proper
restriction enzyme overnight and the size of the resulting fragments were examined by
agarose gel electrophoresis. Fragment sizes differ depending on whether or not the
restriction enzyme was able to cleave the polymorphic site, thus providing a rapid,
reliable, and inexpensive method of genotyping polymorphisms. Figure 4.2 illustrates the
design of the restriction digest genotyping tests.
4.2.2 Allele Frequencies

The polymorphisms selected for study either did not have allele frequencies
reported or had allele frequencies reported for a small number of samples from a
European cat population. In order to determine allele frequencies for a North American
cat population, buccal cell sampling using cytology brushes was performed on cats
belonging to Michigan State University College of Veterinary Medicine staff and
students. Four buccal samples were performed by the owner of each cat. DNA was
isolated from one swab per cat and the other three stored for archival purposes. In order

to isolate genomic DNA from each swab, the swab was placed in a 1.5 ml eppendorf tube

35

Figure 4.2: Schematic of Restriction Digest Genotyping Tests.

 

 

 

 

 

 

 

 

 

 

 

Allele l
31 bp 160 hp 344 bp 97 bp
1 l l
Polymorphic Ease of Separation Digestion control
Tsel site MaeIII site Tsel site
1 l
191 bp 344 bp 97 bp
Allele 2
S N P 7 37
Allele 2
42 bp 421 bp 355 bp
1 l
Digestion control Polymorphic
SphI site SphI site
1
42 bp 776 bp
Allele 1

36

Figure 4.2 continued: Schematic of Restriction Digest Genotyping Tests.

 

 

 

 

 

 

 

 

Allele l
202 bp 307 bp ' 123 bp
l l
Digestion control Polymorphic
AleI site AleI site
1 w
202 bp 430 bp
Allele 2
SN P 495 + 970
Allele 1: 495 3 Allele 1: 970
31 bp 170 bp 307 hp 26 bp 97 bp
1 l l l
Polymorphic Digestion control Polymorphic Digestion control
Tsel site AleI site Alel site Tsel site
(SNP 495) 1 (SNP 970) l
201 bp 333 bp 97 bp
Allele 2: 495 Allele 2: 970

'37

 

    

and immersed in 600 pl of 50 mM NaOH. The tube was vortexed and heated to 95 °C for
5 minutes and vortexed again with brush still in the tube. 60 p] of 1 M Tris (pH 8.0) was
added to neutralize the solution. The tube was vortexed again and stored at 4 °C with the
brush still in the tube. 4-10 p] of this solution was used as template in a 25 pl PCR
reaction.

These samples were genotyped for the three selected P53 polymorphisms
described above. Each sample was amplified by PCR and digested with a restriction

enzyme as detailed in tables 4.2 and 4.3.

Table 4.2: SNP Genotyping Tests: PCR Conditions

  
 

   

     

  

 

  
   
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SNP Primers (Forward/Reverse) Temp- Primer 2 Taq Annea]. Cycles PCR
MgCl
late (pM (mM) (U) Temp.
(Ill) each) (°C) '
5' GGCT’I’I‘CT CC l'l'CIT ATGCAACCT 3’ 66 40
5' AAGGCTCCCCC’I'I‘CTTGCGG 3'
737 5' CGCCTCCCCAGCATCTCATC 3' 10 0.6 2 2.5 70 35 818
5' AAGGCTCCCCC'ITCTTGCGG 3'
970 5' GGC’ITI‘ CT CCTT C'I'l‘ AT GCAACCI‘ 3’ 4 0.4 1.5 2.5 66 40 632
5' AAGGCTCCCCCTI‘CTTGCGG 3'
495 S'GGC'ITTCTCC'ITCTTATGCAACCT 3' 4] 0.4 1.5 2.5 66 40 632
+ 5' AAGGCTCCCCCTTCT'I‘GCGG 3'
970

 

4.2.3 Genotyping of Feline Mammary Carcinoma Samples

Normal mammary tissue DNA, isolated from feline mammary carcinoma samples

as described in Chapter 2, was genotyped for one or more selected polymorphisms as

described in sections 4.2.] and 4.3.1. Samples were considered informative if the normal

tissue was heterozygous at one or more polymorphic loci. Tumor DNA from informative

samples was genotyped and the alleles present compared to those of normal tissue DNA.

Samples were then categorized as showing no LOH, partial LOH, or complete LOH.

38

 

Samples with no LOH were heterozygous in both the normal and tumor DNA. Samples

with LOH were heterozygous for a given polymorphism in the normal DNA but

Table 4.3: SNP Genotyping Tests: Restriction Digests

     
 

  
     

          
   

 

 

     

 

 

 

 

Restriction Additional Incubation Band Sizes (bp)
Enzyme 50 mM Temp Gel %
(pl) MgClz Allele 1 Allele 2 Heterozygote
495 Tsel = 0.5, 2.2 55 °C 3 % 31, 97, 97, 191, 31, 97, 160,
MaeIII = 160, 344 344 191, 344
1.0
737 SphI = 1.0 2.0 37 °C 2 % 42, 776 42, 355, 42, 421, 355,
421 776
970 AleI = 2.0 37 °C 1.5 % 123, 202, 202, 430 123, 202, 307,
0.5 307 430
495 AleI = 2.0 37 °C 3 % For 495: 97, 201 31, 97, 170,
+ 0.5, Tsel = 31, 170 201
970 0.5
For 970: 97, 333 26, 97, 307,
26, 97, 307 333

 

 

 

 

 

 

 

 

 

 

homozygous in the tumor DNA. Samples with partial LOH were heterozygous in both the
normal and tumor DNA, although the tumor DNA showed allelic imbalance. When
visualizing bands formed by differently sized DNA fragments on an agarose gel, bands
representing both alleles in a heterozygous sample normally appear as approximately the
same brightness, indicating approximately the same amount of DNA in each band.
However, the bands are of different brightness in a sample with allelic imbalance,
indicating different amounts of DNA in each band. Therefore, a sample with allelic
imbalance contained a mixed population of cells: some cells have undergone LOH while

others have not. Figure 4.3 illustrates allelic imbalance.

39

Figure 4.3: Complete LOH versus Allelic Imbalance. M = DNA marker, N = normal
tissue, T = tumor tissue

A) Complete LOH B) Allelic Imbalance

MNT MNT

\

    

=
,.__

 

4.3 Results

4.3.1 SNP Genotyping Tests

A total of four restriction digest genotyping tests were designed. Three tests
genotype a single polymorphism (495, 737, and 970) while one test genotypes two
polymorphisms at once (495 and 970). The general plan for these tests was described
previously in section 4.2.], while specific details for each test can be found in table 4.2.
A typical PCR reaction contained 4-10 pl DNA template, 1.5-2 mM MgC12, 0.12 mM
dNTPs, 0.4-0.6 pM each forward and reverse primer, 1x buffer, and 2.5 U Taq
polymerase. Each reaction was denatured at 94°C for 4 minutes; cycled through 94°C for
1 minutes, 66°C or 70°C for 2 minutes, and 72°C for 3 minutes for 35-40 cycles; and

incubated at 72°C for 8 minutes in a thermocycler. 10 pl of the PCR product was then

40

combined with additional MgClz and restriction enzyme, incubated overnight, and
examined by agarose electrophoresis (see Table 4.3 for details for genotyping each SNP).
4.3.2 Allele Frequencies

Allele frequencies for each of the three SNPs genotyped are reported in table 4.4.
Allele frequencies for both the CVM reference cat population and the feline mammary
carcinoma samples are reported. When the allele frequencies of the two groups were

compared with chi square analysis there were no significant differences between them.

Table 4.4: SNP Allele Frequencies

Calc.

 

 

 

 

 

 

 

495 Ref 0. 62 0. 38 0. 50 0. 47
Test 30 0.62 0.38 0.43 0.47
737 Ref 20 0.62 0.38 0.25 0.47
Test 29 0.71 0.29 0.17 0.41
970 Ref 84 0.91 0.09 0.16 0.17
Test 29 0.86 0.14 0.28 0.24

 

 

 

 

 

 

 

 

4.3.3 Loss of Heterozygosity of P53

28 normal samples were genotyped at one or more SNPs in order to obtain 20
informative cases, for an informative rate of 71%. A summary of results is found in table
4.5. A total of six samples showed complete or partial LOH, a rate of 30%. A summary

of cases studied is shown in table 4.6.

41

Table 4.5: PS3 LOH Results

Cate 0 W Percen --e

 

 

 

 

No LOH 14 70
LOH 3 15
Allelic Imbalance 2 10
LOH at 1 of 2 Informative 1 5
Loci

Totals 20 100

 

 

 

 

 

4.4 Discussion

The SNP genotyping tests presented are a rapid, inexpensive, and reliable way to
examine feline P5 3 LOH. They are used here to study feline mammary carcinomas but
will be useful in the investigation of P53 LOH in other feline cancers. Genotyping tests
for three different SNPs increases the number of informative cases available for study.
The combined 495/970 SNP test is useful in genotyping two SNPs at once, but if the
results from this test are unclear, then the presented alternative tests for each single SNP
can be used.

The LOH rate of 30% is similar to the 30-40% P53 LOH rate in human breast

tumors.7 This similarity is not unexpected as P5 3 is highly conserved in structure and
function between species, with P53 playing a critical role in the control of the cell cycle.
The results of this initial study encourage further investigation into the role of P53 in
feline mammary carcinoma. If more similarities between the role of P53 in feline and
human mammary cancers are found, treatments developed for P53 deficient human
tumors may be applied to feline cancer patients.

This study did not contain a large enough sample size to demonstrate a

statistically significant correlation between P53 LOH and tumor invasiveness, metastasis,

,42

Table 4.6: Cases Studied for PS3 LOH. NR. = Not reported.

 

Status Breed Age Sex Tumor T .-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Persian N.R.
ILOH [Mixed NR. N.R. Solid/acinar High Poor Vascular invasion I
ILOH [Mixed 8y F/S Solid/papillary 13h Poor Vascular invasion I
artial IMixed 12y F/ S Acinar High Guarded ocally invasive
OH
lelic 'xed 9y F/S Acinar High Guarded Locally invasive I
mbal.
llelicl 'xed 11y F/S I'apillary 'gh Poor L.N. metz., recurring
Imbal. tumor
I o 'xed 11y F/S I’apillary I od. Guarded
I OH
I 0 'Mixed 10y F/S Solid Low Guarded
I OH
E10 lMixed 18y /s Papillary Low Guarded
OH
E10 I[Mixed 16y F/S Solid/acinar High Poor Locally invasive, vascular
OH invasion, recurring tumor
E10 Siamese 7y F/S Acinar/Scirrhous High Poor Vascular invasion ]
OH
0 aine 9y F/S Acinar N.R. Guarded J
OH oon
o Pylixed 8y F/S Papillary igh Guarded Locally invasive I
OH
0 IMixed 13y IF Acinar High Poor Vascular invasion I
OH
E10 [Mixed 15y F/S Acinar/papillary Low Guarded I
OH
E10 jlvfixed 18y F Acinar/mixed High Poor Vascular invasion, metz on]
OH radiographs
0 [Mixed 7y F apillary Low Guarded
OH
0 [Mixed 13y NR Papillary N.R. Guarded
OH
E10 [Mixed 15y F Acinar N.R. Guarded
OH
0 'xed 10y I: Papillary IMod. Guarded
0H
_ _

 

 

 

 

 

 

 

43

or prognosis. However, all cases found to exhibit P53 LOH were described as highly
malignant while all cases of low or moderate malignancy did not exhibit P53 LOH.
Therefore, with further study, P53 LOH may prove to he a useful prognostic indicator in
feline mammary carcinomas. As there is no currently accepted histological grading
system for these tumors that correlates with prognosis, especially for moderately differen-
tiated carcinomas, a molecular indicator of prognosis would be a useful tool for

clinicians.

44

REFERENCES

1. Malkin D. The Li-Fraurneni Syndrome. The Genetic basis of human cancer. New
York: McGraw-Hill Health Professions Division, 1998;393-407.

2. Levine AJ, Momand J, Finlay CA. The p53 tumour suppressor gene. Nature
1991 ;351 :453-456.

3. Koshland DE, Jr. Molecule of the year. Science 1993;262:1953.

4. Levine AJ. p53, the cellular gatekeeper for growth and division. Cell
1997;88:323-331.

5. Nasir L, Krasner H, Argyle DJ, et al. Immunocytochemical analysis of the tumour
suppressor protein (p53) in feline neoplasia. Cancer Lett 2000;155:1-7.

6. Hollstein M, Rice K, Greenblatt MS, et al. Database of p53 gene somatic
mutations in human tumors and cell lines. Nucleic Acids Res 1994;22:3551-3555.

7. Couch FJ, Weber BL. Breast cancer. The Genetic basis of human cancer. New
York: McGraw-Hill Health Professions Division, 1998;537-563.

8. Liu MC, Gelmann EP. P53 gene mutations: case study of a clinical marker for
solid tumors. Semin Oncol 2002;29:246-257.

9. Hamilton A, Piccart M. The contribution of molecular markers to the prediction
of response in the treatment of breast cancer: a review of the literature on HER-2, p53
and BCL-2. Ann Oncol 2000;11:647-663.

10. Mayr B, Schaffner G, Kurzbauer R, et al. Mutations in tumour suppressor gene
p53 in two feline fibrosarcomas. Br VetJ 1995;151:707-713.

1]. Okuda M, Umeda A, Matsumoto Y, et al. Molecular cloning and chromosomal
mapping of feline p53 tumor suppressor gene. J Vet Med Sci 1993;55:801-805.

12. Kanjilal S, Banerji N, Fifer A, et a]. p53 tumor suppressor gene alterations in
vaccine-associated feline sarcoma. 19th Annual Veterinary Cancer Society Conference
1999.

13. Nambiar PR, Haines DM, Ellis JA, et a]. Mutational analysis of tumor suppressor
gene p53 in feline vaccine site-associated sarcomas. Am J Vet Res 2000;61:1277-1281.

14. Mayr B, Schaffner G, Kurzbauer R, et a]. Sequence of an exon of tumour

suppressor p53 gene--a comparative study in domestic animals: mutation in a feline solid
mammary carcinoma. Br Vet J 1995;151:325-329.

45

15. Mayr B, Reifinger M, Alton K, et al. Novel p53 tumour suppressor mutations in
cases of spindle cell sarcoma, pleomorphic sarcoma and fibrosarcoma in cats. Vet Res

Commun 1998;22:249-255.

16. Mayr B, Reifinger M, Loupal G. Polymorphisms in feline tumour suppressor gene
p53. Mutations in an osteosarcoma and a mammary carcinoma. Vet J 1998;155:103-106.

17. Mayr B, Blauensteiner J, Edlinger A, et a1. Presence of p53 mutations in feline
neoplasms. Res Vet Sci 2000;68:63-70.

46

CHAPTER 5

CONCLUSION

5.1 Potential Gene Targets in Future Studies of Feline Mammary Adenocarcinoma

Understanding the molecular genetic changes that occur during tumorigenesis has
become increasingly important in human oncology; molecular markers are now being
used as prognostic indicators and treatment targets. However, molecular changes in
veterinary cancer patients have not received the same attention. To begin studying
molecular markers of feline mammary carcinomas, two genes that are associated with
human breast tumors were selected. P53 was selected as it has been extensively studied
and is thought to play a role in tumorigenesis in many types of human cancers, including
breast cancer. On the other hand, 0 has only recently been implicated in human breast
cancer. However, changes in this gene were found in a high percentage of tumors
examined and were being investigated as a target of therapy, making it an interesting
gene to examine in another species such as the cat.

P5 3 and 0 represent only a fraction of the genes that have been studied in human
breast cancer, leaving many more to investigate in feline mammary carcinomas. One such

gene is HER2, also called ErbBZ or neu, a tyrosine kinase receptor that has been found to

be overexpressed in 20-40% of human breast cancers.1 The most common mechanism of
overexpression of HER2 is gene amplification, in which several copies of a gene or

chromosomal region are present. In humans, HER2 amplification is associated with

aggresive tumor behavior, shorter suvival time, and overall poor prognosis.2 HER2
overexpression is also predictive of response to some types of therapy. HER2+ tumors are

resistant to hormonal therapies such as tamoxifen but have increased sensitivity to the

47

chemotherapeutic agent anthracycline.3 Recently, HER2 itself has become a therapeutic

target: trastuzurnab, a humanized anti-HER2 monoclonal antibody, has been approved by

the FDA for treatment of women with HER2 + breast tumors.3

The first step in investigating HER2 in the cat would be to determine the feline
nucleotide sequence for this gene. The same process used to determine the feline
a sequence could be used to determine the feline HER2 sequence. Currently, the HER2
genomic sequence has been determined for the human (Gen-bank accession NM__004448)
and the mouse (Genbank accession NT_031413.2) while the mRN A sequence has been
determined for the rat (Genbank accession X03362). This existing sequence information
could be used to design oligonucleotide primers that would bind to highly conserved
regions of HER2, which could then be used to amplify and sequence feline HER2. Once
feline HER2 is sequenced, current methods used to evaluate human HER2
amplfication/overexpression could be assessed for their usefulness in the cat. One method
commonly used to assess HER2 status in human breast cancer patients is
immunohistochemistry (IHC), in which mammary tissue slides are stained with an
antibody which binds to HER2 protein. In order to visualize the antibody binding sites,
the antibody itself may carry a marker such as fluorescein or horseradish peroxidase, or a
secondary antibody carrying a marker which then binds to the primary antibody may be
used. The slide is then visually examined for presence of the marker. There are currently

more than 30 anti-human HER2 antibodies as well as a commercial kit (HercepTest,

DAKO) which are available for IHC.4 If feline and human HER2 are sufficiently similar,
then these antibodies may be used to perform IHC on feline mammary tissue and tumor

samples and assess HER2 overexpression. However, if anti-human HER2 antibodies do

48

not bind to feline HER2, the prohibitive time and expense involved in generating anti-
feline HER2 antibodies makes IHC a less practical techinique for examining feline HER2
overexpression. Fluorescence in situ hybridization (FISH) is another technique that is
currently being used to detect human HER2 amplification. In FISH, a slide-mounted
tissue section is stained with a fluorescently labeled oligonucleotide probe which binds to

HER2 in the cells' chromosomes. The slide is then visually examined for the- intensity of

the fluroescent signal in order to determine HER2 copy number.5 Although feline-
specific oligonucleotide probles are more easily manufactured than feline-specific
antibodies, the specialized equipment necessary to perform FISH may limit the use of this
technique in some laboratories. Although less sensitive than FISH, techniques such as
differential PCR or Southern blots may also be used to determine if HER2 is amplified in
feline mammary tumor samples. In differential PCR, the target gene and a reference gene
are co-amplified by PCR and the product amounts measured by densitometry. The ratio

of the target to reference gene product represents the amount of DNA originally in the

sample.6 In Southern blot, DNA is isolated fi'om cells, cleaved into fragments by
restriction enzymes, and the fragments are seperated through gel elecrophoresis. The
DNA is then transferred to a nitrocellulose membrane and washed with a solution
containing a radiolabled oligonucleotide probe which will hybridze to the gene of
interest. Radiograph film is then exposed to the nitrocellulose membrane, allowing the
fragment size and amount of DNA present to be measured. Most of the above techniques
have been shown to be effective with formalin-fixed paraffin embedded samples,
allowing examination of the feline mammary carcinoma samples we have already

collected.

49

BRCA1 and BRCA2 are two other genes that should be investigated in feline
mammary cancers. These genes have been associated with familial breast and ovarian
cancers, in which a germline mutation in either BRCAI or BRCA2 is passed on from
generation to generation, increasing the risk of mammary tumor development. Although
somatic mutation is very rare in sporadic human breast cancer cases, loss of

heterozygosity is common: 50-70% of sporadic ovarian and breast tumors have LOH of

BRCA1 and 30-50% have LOH of arc/12.7 The flmctions of BRCA1 and BRCA2

proteins are closely related, both being involved in control of homologous recombination

and DNA double-strand break repair.8

Although familial mammary cancer has not been reported in the cat, the role of
BRCA1 and 2 in sporadic feline mammary cancer should be investigated. LOH of BRCA1
and 2 can be studied in much the same way as P53 LOH was studied in this project. The
first step would be to obtain the nucleotide sequence of the feline BRCA1 and 2 genes,
which could be done in much the same was as the 0 sequence was obtained. A 2845 bp
partial coding sequence for feline BRCAI has been determined (Genbank accession
AF 2840] 8) which provides a good starting point for oligonucleotide primer design for
PCR. Feline BRCA2 has not been sequenced, but published BRCA2 sequences for several
other species (including human, mouse, and dog, Genbank accessions NM_000059,
NM_009765, and AB043 895, respectively) provide information to find conserved
sequences on which to base PCR primer design. As there are no published SNPs or other
genetic markers for feline BRCA1 and 2, the next step would be to identify these genetic
markers in the genes of interest. In order to identify SNPs, a pool-and-sequence method

can be used. In this method, DNA from several individual cats is combined, the DNA

50

region of interest is sequenced, and SNPs are identified through the presence of multiple

nucleotides at a single locus.9 Once SNPs are identified, restriction enzyme tests sirniliar
to those used in the P53 portion of this study can be developed in order to distinguish
between SNP alleles. The same feline mammary tissue DNA samples that were used in
this study can then be examined for SNP genotype. Once normal mammary tissue
samples that are heterozygous at a SNP locus are identified, the corresponding tumor

DNA sample can be genotyped in order to determine if it has undergone LOH.

5.2 Methylation in Feline Mammary Adenocarcinoma

In this study, we did not find hypermethylation of 14-3-3 o in the cat. One
possibility is that hypermethylation does not occur in the cat at all, as there have been no
other published studies documenting any hypermethylation in this species. However, this
lack of reporting is probably due to the scarcity of studies involving genetic and
epigenetic phenomena in the cat rather than a lack of feline hypermethylation. As
methylation is a highly conserved process and hypermethylation has been documented in
other species such as the rat and the mouse, it is probable that hypermethylation does
occur at other loci in the cat

Hypermethylation of several genes other than 0 has been linked to human breast
cancer. E-cadherin (or E-cad) is a cell adhesion molecule, downregulation of which has
been linked to invasion and metastasis in various human carcinomas. E-cad contains a
CpG island located in the promoter through intron 1 of this gene. Hypermethylation of

this CpG island has been found in greater than 50% of human primary breast tumors

examined. 10 Hypermethylation of the BRCA I promoter has also been identified One

51

study found that 13% of human primary breast tumors had hypermethylation of

BRCA 1. 11 This same study also examined the relationship of BRCA1 LOH and
hypermethylation; 20% of tumors exhibiting LOH were hypermethylated while only 5%
of tumors not exhibiting LOH were hypermethylated.

Both E-cad and BRCA1 would be interesting loci to study hypermethylation in
feline mammary carcinoma. These genes could be examined in much the same way a
was examined for hypermethylation, using the same DNA samples already isolated from
feline mammary tissue and tumors. In order to design PCR primers, conserved regions
from the promoter sequences of the human (Genbank accession L34545 (E-cad), L78833
(BRCA 1)), mouse (Genbank accession M81449 (E-cad)), rat (Genbank accession
AF 080590 (BRCA1)), and dog (Genbank accession AF 330163 (E-cad)) can be used.
DNA samples can be treated with sodium bisulfite and sequenced as described in chapter
3 in order to determine if the promoters of these genes are hypermethylated in feline

mammary carcinoma cases.

5.3 Cats as Animal Models for Human Breast Cancer

Cats have been proposed as a good animal model for human breast cancer as the
pathology, behavior, and drug response of mammary tumors is very similar between
these species. Molecular genetic changes in breast tumors have been extensively studied
in the human but not the cat, making this an unknown factor when evaluating the cat as
an appropriate animal model. In this study, we examined two molecular genetic changes
that have been documented in human breast tumors but had not previously been studied

in the cat. In the case of 14-3-3 0, cats and humans were very dissimilar- greater than

52

90% of human primary tumors exhibited hypermethylation while none of the feline
tumors exhibited hypermethylation. In the case of P5 3 , cats and humans were very
similar-mammary tumors of both species had an approximately 30% LOH rate. These
disparate results illustrate why careful consideration is necessary when selecting an
animal model if the goal of the study is obtain results applicable to human disease.
Caution is also needed when applying therapies developed for human disease to
veterinary species, as dissirnilarities at the molecular level may result in treatment failure.
More information is needed concerning molecular genetic events of tumorigenesis in
animals in order to develop better animal models for human disease as well as to allow

veterinary medicine to take advantage of advances in human disease treatment.

53

REFERENCES

l. Couch FJ, Weber BL. Breast cancer. The Genetic basis of human cancer. New
York: McGraw-Hill Health Professions Division, 1998;537-563.

2. Yarden Y. Biology of HER2 and its importance in breast cancer. Oncology
2001;61 Supp12:1-13.

3. Lohrisch C, Piccart M. An overview of HER2. Semin Oncol 2001;28:3-11.
4. Hanna W. Testing for HER2 status. Oncology 2001;61 Suppl 2:22-30.

5. Pauletti G, Godolphin W, Press MF, et a]. Detection and quantitation of HER-
2/neu gene amplification in human breast cancer archival material using fluorescence in
situ hybridization. Oncogene 1996;] 3:63-72.

6. Deng G, Kim YS. Quantitation of erbB-2 gene copy number in breast cancer by
an improved polymerase chain reaction (PCR) technique, competitively differential PCR.
Breast Cancer Res Treat 1999;58:213-217.

7. Welcsh PL, King MC. BRCA1 and BRCA2 and the genetics of breast and ovarian
cancer. Hum Mol Genet 2001;10:705-713.

8. Venkitararnan AR. Cancer susceptibility and the functions of BRCA1 and
BRCA2. Cell 2002;108:171-182.

9. Brouillette JA, Andrew JR, Venta PJ. Estimate of nucleotide diversity in dogs
with a pool-and-sequence method. Mamm Genome 2000;11:1079-1086.

10. Graff JR, Herman JG, Lapidus RG, et al. E-cadherin expression is silenced by
DNA hypermethylation in human breast and prostate carcinomas. Cancer Res
1995;55:5195-5199.

1 1. Esteller M, Silva JM, Dominguez G, et a]. Promoter hypermethylation and
BRCA1 inactivation in sporadic breast and ovarian tumors. J Natl Cancer Inst
2000;92:564-569.

54

 

Ilfllj‘ljflllfllfljljl[@1le