AN 1IV HVO MODEL FOR PSA PRODUCTION BY BREAST CANCER CELLLINES

GROWING AS XENûGRAF-ïS IN SCID MICE

Ilana Kogan

A thesis submitted in conformity with the requirements

for the degree of Master of Science,

Graduate Department of Medical Biophysics,

University of Toronto

O Copyright by Ilana Kogan (1997)

Acquisitions and Acquisitions et Bibliographie Services services bibliographiques

395 Wellington Street 395, rue Wellington Ottawa ON K I A ON4 Ottawa ON K1A ON4 Canada Canada

Your lile Votre référence

Our file Norre rtiiérence

The author has granted a non- L'auteur a accordé une licence non exclusive licence allowing the exclusive permettant à la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or sel1 reproduire, prêter, distribuer ou copies of this thesis in microform, vendre des copies de cette thèse sous paper or electronic formats. la forme de rnicrofiche/film, de

reproduction sur papier ou sur format électronique.

The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantid extracts fiom it Ni la thèse ni des extraits substantiels may be printed or othenvise de celle-ci ne doivent ê e imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation.

AN IN ?!IV0 MODEL FOR PSA PRODUCïiON BY BREAST CANCER (7ELGLINES

GROWING AS XENOGRAliTS IN SCID MICE

Master of Science, 1997

Ilana Kogan

Department of Medical Biophysics, University of Toronto

ABSTRACI'

The presence of prostate-specific antigen (PSA) in human breast tumours has been

associated with a good prognosis. An in vivo PSA-induction model was established, using the

steroid-hormone receptor-positive hurnan breast cancer cell-lines T47Q MCF7, or BT474,

growing as xenografts in SCID mice. PSA production was initiated by stimulahg the mice with

norgestrel for 5-7 days. Mice injected with the prostate cancer celhline LNCaP or the steroid-

hormone receptor-negative human breast cancer cell-line BT20 fùnctioned as positive or negative

controls for PSA production, respectively. Norgestrel induced low levels of PSA in the receptor-

positive breast tumour xenografts. In vitro data showed that norgestrel can stimulate

exponentially growing T47D or BT474 cells to produce PSA and that low PSA concentrations

do not affect their proliferation rate. This in vivo model is usehl for studying îhe role of PSA

as a prognostic rnarker for breast cancer.

ACKNOwLEDGEMErn

1 would like to express my gratitude and appreciation to Dr. Mike Rauth for his helpful

guidance, support, and encouragement throughout this project. Special thanks to Drs. Jim

Ballinger, David Hedley, and E. P. Diamandis for their advice and helpful discussions. 1 also

thank Mr. Robert Kuba for the technical assistance. 1 would also like to acknowledge the

generous contributions of Dr. J. Mullen for the irnrnunohistochemistry, Dr. S. Minkin for the

help with the statistical analyses, and Dr. D. J. A. Sutherland for the receptor analyses. 1 also

thank Tricia Melo, Veet Misra, Christine Brezden, Leora Hom, Hugo Zhang and Jonathan

Tunggal for the helpful discussions and the moral support throughout this project. Special thanks

to my family for their continuous encouragements and the unfailhg belièf in my future. Finally,

1 also thank Zoltan Gombos for the great motivation, his advice, and faithtùl support.

TABLE OF C0h"TENTS

Abstract ......................................................................................................................

Acknowlsdgements

Table of Contents

List of Tables

List of Figures

Chapter 1:

1.1.

1.2.

1.3.

1.4.

Introduction .............................................................................................. .............................................................................................. ûverview

Introduction ..............................................................................................

The History of Prostate-Specific Antigen ..........................................

PSA Gene and Protein ......................................................................

...................... PSA 1s a Member of the Kallikreim Gene Family

The PSA Protein ...................................................................

......................................... PSA Processing and Secretion

Molecular F o m of PSA in the Serninal Plasma and the S e m ..................................................................................

PSA fkom Non-Prostatic Chigin ..............................................

General Features of Steroid Hormones and Steroid-Hormone Receptors ............................. .. ................................................

PW . . 11

. . 11

i\

vii

in

1.6. Factors Affecting PSA Synthesis and Expression ..................................

.......... 1.6.1. Steroid Hormones and Steroid Hormone Recepto~s

1.6.2. Growth Factors and Other Molecules ..................................

1.7. PSA Function ..................................................................................

...................................................................... 1 .8 . PSA and Breast Cancer

1.9. Breast Cancer ..................................................................................

1.9.1. Epidemiology ......................................................................

.................................. 1.9.2. Mammary Gland and Mamrnary Cancers

1.9.3. Prognostic Markers in Breast Cancer ..................................

.................................................................................. 1.10. Thesis Outline

............................................................................................ 1.1 1 . References

Chapter 2: ProstatôSpecific Antigen Induction by a Steroid Hormone in T47D Cells Growing in SClD Mice ..................................

.............................................................................................. Summary

.............................................................................................. Introduction

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

.................................................................................. Ce11 Lines

Animals ..................................................................................

.......... Genemtion of SCID Mice Xenografk and PSA Induction

Preparation of Tissue Extracts ..............................................

.................................................................................. PSA Assay

2.3.5.1. Assay Description and Instrumentation ......................

.............................................. 2.3.5.2. Reagents and Solutions

2.3.5.3. Antibodies ......................................................................

.................................. 2.3.5.4. Calibrators (PSA Standards)

........................................................ 2.3.5 .5 . Assay Procedures

.......... Total Protein Determination and PSA Imrnunoreactivity

Imrnunohistochernistry ..........................................................

...................... Progesterone- and Estrogen-Receptor Analyses

...................................................................... Statistical Analysis

.............................................................................................. Discussion

References ......................... ... ............................................................

Chapter 3: Summary and Conclusions. PreIiminary Results. and Future Directions ..............................................................................................

................................................................ 3.1. Surnrnary and Conclusions

3.2. Prelirninary Results ..................................................................................

3.2.1. Stimulation of T47D Cells with the Progestin Norgestrel ..........

3.2.2. Effect of PSA on the Growth Rate of T47D and BT474 ................................................................................. Cells

3.2.3. Prostate-Specific Antigen Induction in SCID Mice Be-g BT474, MCF7. and BT20 Cells ......................

3.3. Future Directions ..................................................................................

3.3.1. Effect of PSA Production on the Sensitivity of Human Breast-Cancer Cell-Lines to Chemotherapeutic Dmgs and Ionizing-Radiation . In Vifro Studies ......................

3.3.2. Effect of Chemotherapeutic Dru@ and Ionizing Radiation on PSA-Positive Tumeurs vs . PSA-Negat ive Turnows,

Growing as Xenografts in SCID Mice .................................

3.4. References ..............................................................................................

vii

LIST OF TrnLES

Chapter 2

Table 2-1. Means and SDs of PSA concentration in turnours and tissues of control, 5-day norgestrel-stimulated, and 7-day norgestrel- stimulateci T47D mice ....................................................... 64

Table 2-2. Means and SDs of PSA concentration in tumours and tissues of LNCaP mice ................................................................................. 66

Table 2-3. Levels of Prostate-Specific Antigen, progesterone receptor, and estrogen receptor in turnours fiom control, 5-day norgestrel- stimulated, and 7-day norgestrel-stimulated T47D rnice .......... 73

... Vlll

LIST OF FZGUmS

Chapter 1

Figure 1-1.

Figure 1-2.

Chapter 2

Figure 2-1.

Figure 2-2.

Figure 2-3.

Figure 2-4.

Figwe 2-5.

Figure 2-6.

Figure 2-7.

Schematic representation of the PSA gene ...,..................

The mechanism of action of steroid hormones ......................

PSA Assay ..................................................................................

Means and SDs of PSA concentration in tumours from control, 5-day norgestrel-stimulated, and 7-&y norgestrel-stimulated T47D mice and LNCaP

Intra- and inter-tumour variability in IpSA] in T47D rnice . . . . . . . .. . . . . . . . . . .. . . .. . . . . . . . . . . . . . . . . . . . . . . . .. . , . . . . . . . . . .. . , . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . Intra- and inter-tumour variability in [PSA] in LNCaP mice ...................................................... #............................. -

Immunohistochernical staining of tumour sections fmm LNCaP mice ushg a polyclonal anti-PSA antibody ..,.......

The level of progesterone receptors in tumours of control, 5-day norgestrel-stimulated, and 7-day norgestrel-stimulated T47D mice ...........................................,.........,...,..,,,.+,~,.,....,.......

The level of estrogen receptors in turnours of control, 5-day norgestrel-stimulated, and 7-day norgestrel-stimulated T47D mice ...... ............................... ... ...., ,................... ............... ..

Chapter 3

Figure 3-1.

Figure 3-2.

Figure 3-3.

Figure 3-4.

Figure 3-5.

Figure 3-6.

Figure 3-7.

Figure 3-8.

Figure 3-9.

PSA production by T47D cells post norgestrel stimulation . . ... . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . .

Chernical structure of M I T which by reduction is converted to a formazan product .........................................................,

Effect of varying PSA concentrations on the proliferation of T47D cells measured by the MTT assay ......................

Effect of varying PSA concentrations on the proliferation of BT474 cells measured by the MTï assay ......................

Means and SDs of PSA concentration in turnours of control BT474 NODISCID mice and in tumours and tissues of 7-day norgestrel-stimulated BT474 NODISCID mice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . , m. . . . . . . . . .

Means and SDs of PSA concentration in tumours of control MCF7 rnice and in tumours and tissues of 7-day norgestrel-stimulated MCF7 mice ...................... .. ......

Mean and SD of [PSA] in turnours of rnice fiom different treatment groups ................................................,........

Mean and SD of the level of estrogen receptors in T47D, BT474, MCF7, and BT20 mice ..................................

Mean and SD of the level of progesterone receptors in T447D, BT474, MCF7, and Bï20 mice ..................... , ,,,,........

1.1. OVERVIEW

Cancer is a multistage disease caused by an accumulation of mutations in different type:

of genes, such as proto-oncogenes, turnow suppressor genes, and genes involved in DNA repair

Mutations in these susceptibility genes, rnany of which controi ceil growth, can lead to biologica:

changes that alter ce11 function, thus leadiig to uncontrolled proliferation (Helzlsouer, 1993)

Cancer has been a major problem throughout human history. Cmently, it is estimated thai

approximately 40% of Arnericans will eventually have the disease and more than one in five wili

die of it. Mortality fiom some types of cancer, such as Hodgkin's disease, Burkitt's lymphoma

and t~ticular cancers, has declined drarnatically in developed corntries in the last few years dur

to more effective detection and treatrnent procedures (Rennie and Rustiig, 1996). Nevertheless

the incidence of the most significant forms of cancer has remained constant or increased, partl~

due to an increase in the aging population. Furthmore, cancers of the lung, breast, prostate

and colon have al1 become more fiequent in countries where environmental risk factors, such a(

cigarette smoking, unhealthy dietary habits, and exposure to carcinogens, are now more cornmor

(Trichopoulos et al., 1996).

Breast cancer is the most common cancer among North American women. Preventior

of breast cancer is dificult since the initiation of the disease involves many endogenous factor!

that are diflïcult to control. Currently, early diagnosis and administration of effective treatrnent:

are the major means to reduce the mortality rate. Due to the heterogeneity of breast cancer,

physicians use prognostic markers to identi& patients who are at high risk for disease recwrenct

and, therefore, might benefit fiom adjuvant therapy (McGuire and Clark, 1992). Although s

variety of prognostic markers are available, these markers are not very sensitive and specific!

3

rnaking it dificult to identify high risk patients. It is important to characterize new prognostic

parameters to better understand the biology of breast cancer and design new treatrnents for some

subgroups of breast cancer patients. The recent observation that some breast hunours express

prostate-specific antigen (PSA; Yu et al., 1994a) and that it is a positive prognostic factor (Yu

et al., 1995) has focused attention on PSA's role in breast cancer.

1.2. INTRODUrnON

PSA is a serine protease that, until recently, was thought to be secreted exclusively by

epithelial cells of the prostate gland. In the past several years, however, new reports have show

that many tissues and tumom can produce PSA, although in much lower concentrations. For

example, PSA was detected in the endomeûiurn (Oesterling, 1991; Clements and Mukhtar, 1994),

amniotic fluid (Yu and Diamandis, 1995a), normal breast tissue, and in tissue of benign breast

disease (Yu et al., 1996a). PSA-irnmunoreactivity was also exarnined in more than twelve

hundred breast cancer patients. Results showed that immunoreactive-PSA was present in >30%

of the breast turnour cytosols (Diamandis et al., 1994; Yu et al., 1994a). Monne et al. (1994)

later c o n h e d that the immunoreactive PSA detected in the breast turnour cytosols was actually

PSA. Multivariate analysis indicated that breast cancer patients with PSA-positive tumours had

a reduced risk for relapse, suggesting PSA to be an independent favourable prognostic marker

for some breast cancer subgroups (Yu et al., 1995).

In the following sections, a brief review is given of the discovery of PSA and some

aspects of the PSA gene and protein. Since the PSA gene is upregulated by steroid hormones,

the mode of action of steroid hormones and their receptors is discussed. The role of PSA in the

4

semen, and the possible hctions of PSA in prostate cancer and breast cancer are then

summdzed. To help unde~tand the possible role of PSA in the etiology of breast cancer, the

role of exogenous and endogenous factors affecthg the development of the rnarnrnary gland are

presented. This chapter provides the background for understanding the rationale behind

developing an in vivo mode1 for PSA production by breast tumom, which is presented in

Chapter 2.

lb3. THE HISTORY OF PROSTAïX-SPECIFIC ANTIGEN

The identification of the protein subsequently named PSA was independently reported by

several groups during the 1970s, who initially named the protein based on its physical or

biochernical characteristics. In 1971, a group of Japanese investigators identified the protein in

the serninal plasma by imrnunoelectrophoresis and named it y-sehoprotein (Hm et al., 1 97 1).

In 1973, a similar protein was isolated fiom the semen and named protein-E, based on its

electrophoretic mobility (Li and Beling, 1973). Using sodium dodecyl sulphate polyacrylamide

gel electrophoresis (SDS-PAGE) and irnmunological methods, Sensabaugh characterized the

protein fiom human seminal plasma biochemically, determinhg its molecular weight as 30 kDa.

nie protein was, therefore, named p30 (Sensabaugh, 1978).

In the prostate, this protein was fmt identified in 1977 by Wang and colleagues.

However, it was not purified fiorn this tissue until 1979 (Wang et al., 1979). Since this group

could detect the protein only in prostatic tissue they named it prostate-specific antigen. In 1980,

Papsidero and colleagues detected PSA in the serum of men with advanced prostate cancer by

an imrnunoelectrophoresis technique. Two years later, Wang and CO-workers (1 982) showed that

5

PSA was present in the seminal plasma and that it was sirnilar to the PSA they detected in the

prostate. Although some investigators were sceptical at that time that y-seminoprotein, protein-

E,, p30, and PSA were the sarne protein, it is now accepted that al1 these names refer to PSA

(McCormack et al., 1995).

1.4. PSA GENE AND PRû'ïEIN

1.4.1. PSA 1s a Member of the Kallikrein Gene Family

PSA is a mernber of the human kallikrein gene farnily and is local id to the long a m

of chromosome 19, in the region q13.3 (Riegrnan et al., 1989a; Schonk et al., 1990). Kallikreins

are encoded by a sequentid multi-gene farnily in many species. Although genes related to the

human PSA gene were detected in several primate species, they were not found in other

mammalian species, including rat or mouse (Karr et al., 1995). In humans, the kallikrein gene

locus is -60-70 kilobases (kb) long, and includes three members: tissue (kidney/pancreas)

kallikrein (KLKl), glanduiar kallikrein (KLK2 or hGK-l), and PSA (KLK3; Drinkwater et al.,

1988; MacDonald et al., 1988). The PSA gene has a high degree of homology (82%) to KLK2

(Henttu and Vihko, 1989) and extensive homology (68%) to KLKl (Watt et al., 1986).

Moreover, the promoter ngions of PSA and KLK2 are 91% identical (HenMi and Vihko, 1989;

Henttu and Vihko, 1994). The genes for PSA and KLK2 are express& in normal, benign, and

rnalignant prostatic epithelium (Watt et al., 1986; Chapdelaine et al., 1988). HenMi and

colleagues (1990) reported that there are approximately five times higher levels of PSA mRNA

compared to KLK2 rnRNA in benign and malignant prostatic tissue. Other studies suggested that

the levels of KLK2 are half that of PSA mRNA in normal prostate or benign prostate cancer

6

(Young et al., 1992).

The PSA gene was characterized by several research groups (Lmdwall and Lilja, 1987;

Henttu and V i o , 1989; Riegrnan et al., 1989a; Riegman et al., 1989b). Riegrnan and

colleagues (1 989b) sequenced the gene encoding PSA and showed that it has five exons and four

introns over the length of approximately 6 kb. Figure 1-1 provides a schematic representation

of the PSA gene. The promoter of the PSA gene contains a hctional androgen-responsive

element (ARE) at position -170 to -156 (AGAACAgcaAGTGCT), showing that the gene is under

androgen regulation. This ARE closely resembles the ARE consensus sequence, which is

necessary for the bindiig of staoid (glucocorticoid, progestin, and androgen) hormone mptors

( Ham et al., 1988; Riegman et al., 1989a; Riegman et al., 1991).

A 35 bp segment starting at -400, the androgen-response region (ARR), cooperates with

the ARE in androgen induction of the PSA promoter. Both segments are required for hi&

androgen induced activity of the PSA promoter (Cleutjens et al., 1996). Schuur and CO-workers

(1996) identified a prostate-specific enhancer, located between -5327 and -3737, which is

androgen responsive and requires a promoter for activity. These results suggest that biidiig of

the androgen receptor to the ARE may not be sufficient for prostate-specific gene expression.

Furthermore, gene expression may require interaction of prostatespecific proteins with the

enhancer. Other well characterized regulatoy elements on the PSA gene, which are important

for the regulation of transcription, include: a variant TATA box (ïTïATA) at position -28 to -23,

a GC-box at -53 to -48, and a CACCC-box at -129 to -125. Two transcription initiation sites

were also identified at positions +l and +7. The open readiig h e starts at +42 in the first

exon and ends with a STOP codon at 1-5204 Wegrnan et al., 1989b).

(START) (STOP) 5' 3'

"-m.

-5327 -3737 -4M -3a-170 -156-129 -125 -53 -48 -28 -23 5204 5880 \ I +I 7 42 \ I \ I \ 1 / EXON 1 EXON2 EXON3 EXON4 EXON5 \ I \ I \ I \Yt

Promoter Region

1 Prostate-Specific Enhancer 4 CACCC-Box 2 ARR (Androgen-Responsive pegion) 5 G c - 6 0 ~

3 ARE (Androgen-Response Element) 6 Variant TATA Box

+1, +7 two transcription initiation sites

Figure 1-1. Schernatic representation of the PSA gene, showing cis-acting transcriptional elements (1-6; open boxes), two transcription initiation sites at +1 and +7, START codon at 4-42, and STOP codon at +5204. Closed boxes represent exons. Diagram is not to scale.

8

1.4.2. The PSA Protein

PSA is a single-chain glycoprotein (Wang el al,, 1982). B a d on the sequence of its

arnino acids, the mature, catalytically active PSA molecule is thought to consist of 237 arnino

acid residues, with a calculated molecular weight of approximately 26 D a . However, due to the

glycosylation of the protein and the heterogeneity of the sugar chains, the molecular weight of

the protein is approximately 33 kDa as detected by gel filtration and gel electrophoresis (Wang

et al., 1981; Watt et al., 1986; Lundwall and Lilja, 1987; Schaller et al., 1987).

Watt and coIIeagues (1986) suggested that there are four sugar chains that are linked to

the PSA polypeptide chaii: an N-linked carbohydrate side chain at Asn-45, two Olinked

carbohydrate side chaii at Sa-69 and Ser-71, and one Olinked carbohydrate at Thr-70.

However, others believe that there is only one Asn-linked carbohydrate moiety at position 45

(van Halbeek et al., 1985; Lundwall and Lilja, 1987; Schaller et al., 1987; Belanger et al., 1995;

McCormack et al., 1995). The rnicroheterogeneity in the sugar components of the PSA gene

(hexoses, hexosamines, and sialic acid) gives rise to several PSA isoforrns with diffaent

isoelectric points (p1) that range fiom 6.8 to 7.5. The most abundant PSA isoform exhibits a pI

of 6.9 (Wang et al., 1979; Wang et al., 1982).

Tissue-kallikrein and kallikrein-like proteins are a subgroup of serine proteases that can

cleave specific precursor proteins at selected sites, generating mature and biologicalIy active

proteins. Amino acid sequence analysis showed that, like other kallikreins, PSA is also

structurally homoIogous to members of the serine protease family. There are four major classes

of proteases that are disthguished by the principal functional group in their active site: serine,

thiol, carboxyl, and metalIo (Kraut, 1977; Watt et al., 1986). The active site residues of the PSA

9

molecule @s-41, Asp-96, and Ser-189) compond to the three residues invariably found in other

serine proteases (Neurath and Walsh, 1976; Schaller et al., 1987).

Serine proteases play an important role in post-translational processing of polypeptides

@rinkwater et al., 1988; MacDonald et al., 1988). By virtue of this processing role, kallikreins

are vital regulators of diverse aspects of local functions in many tissues, such as blood

flow/pressure, inflammation, proliferation, and differentiation (Clements, 1994). The substrate

specificity of PSA is similar to that of chymotrypsin, since both enzymes cleave peptide bonds

at the carboxy-teminal of certain Leu and Tyr residues. However, some synthetic substrates

cleaved at high rates with chymoûypsin are poor substrates for PSA. ïherefore, PSA is

considerd to have a restricted chymotrypsh-like activity (Lilja, 1985; Akiyama et al., 1987;

Christensson et al., 1990).

There are ten Cys residues on the PSA polypeptide at positions that are homologous to

those of other serine proteases, suggesting that five disulphide bonds are present on the PSA

molecule (Watt et al., 1986). Most of PSA isolated fiom the serninal fluid is enzymatically

active (Lilja, 1985; Christensson et al., 1990). However, studies showed that 20-30% of PSA

in these preparations is clipped between residues 85-86, 145-146, or 182-183. It is known that

the clavage between arnino acid residues 145 and 146 inactivates the rnolecule (Christensson

et al., 1990). The clipped protein remains comectd by the interna1 disulphide bonds and co-

migrates with the intact PSA during non-reducing SDS-PAGE analysis. However, due to the

lower molecular weight of the nicked form of PSA under denaturing conditions, it is possible to

separate this molecule fiom the active forrn (Watt et al., 1986; Schaller et al., 1987). Both the

inactive and the active 33 kDa proteins are described as fie PSA.

1 O

1.4.3. PSA Processing and Secretion

Knowledge of the sequence of the mature PSA molecule and its precursor (Lilja, 1985;

Lundwall and Lilja, 1987; Schaller et al., l987), experiments with other serine proteases (Neurath

and Walsh, 1976; Neurath, 1986; MacDonald et al., 1988), and data on the intracellular

localization of PSA (Sinha et al., 1987) a11 showed that PSA is processed in a way similar to

otha extracellular serine proteases. Althougti there is no direct experimental evidence, it is

believed that PSA is translated as an inactive 261 amino acid PSA precmor. The PSA precursor

contains a 24 amino acid extension at the N-terminai of the mature molecule. This extension

d e s up both the signal peptide (17 residues) and the propeptide (7 residues; Lilja, 1985).

The signal peptide of PSA guides the protein to and across the endoplasmic reticulurn

membrane, thus directing PSA to enter the secretory pathway. In the endoplasmic reticulurn

membrane, the signal peptide is cleaved and the resulting molecule, the proPSA, is glycosylated

and processed in the Golgi apparatus. men, it is transported within vesicles to the plasma

membrane. In the prostate, proPSA is secreted fiom the plasma membrane into the lumina of

the prostate ducts. ProPSA is the inactive precursor of PSA that contains seven additional

residues at its N-terminal. It is thought that cleavage of these additional residues results in the

mature, catalytically active PSA mo1ecule that contains 237 amino acids. However, the

protease(s) responsible for cleaving the proPSA to the mature PSA has not been identified (Lilja,

1985).

11

1.4.4. Molecular Forms of PSA in the Seminal Plasma and the Semm

In the seminal plasma, more than 60% of PSA shows enzymatic activity. Nevertheles,

approximately one third of PSA purified from seminal plasma is lacking enzymatic activity due

to the cleavage at the carboxy-terminal of Lys-145 (Lilja, 1985; Christensson et al., 1990). A

rninor fiaction, less than 5% of seminal PSA, is inactivated by binding to the protein C inhibitor

(PCI) and f o m a high molecular mass complex of approximately 90 ma. PCI is a glycoprotein

of about 57 kDa that belongs to the family of extracellular serine protease inhibitors (serpins) and

is rnainly produced by the seminal vesicle (Lilja, 1985; Christensson et al., 1990; Christensson

and Lilja, 1994). ïhe PSA-PCI complex has not been detected in serum (Espaîia et al., 1993).

Besides PCI, the serpin farnily includes a-1-anti-trypsin and a-1-anti-chymotrypsin (ACT;

Christensson et al., 1990).

h w PSA concentrations (< 4 y&) are usually found in the serum of normal males.

These levels correspond to approximately IO6 of the physiologic PSA concentrations present in

the seminal fluid, which range fiom 0.5 to 3.0 gL (Lilja, 1985). Papsidero and colleagues (1980)

observed that immunoreactive PSA in the s e m has a molecular mass different fiom that found

in the seminal plasma. Since PSA is a serine protease, they hypothesized that PSA in the serurn

rnight fonn complexes with other proteins, thus becorning enzymatically inactive. It was later

discovered that in the senim, enzymatically active PSA (f-PSA) forms stable complexes with two

protease inhibitors: ACT and a-2-macroglobulin ( a m . The a,M analogue - the pregnancy zone

protein - is the third serine protease inhibitor that was shown to forrn complexes with PSA

(Christensson et al., 1990).

12

The dominant form of PSA in the serum, as determined by conventional irnmunoassays,

is bowd to ACT and constitutes approximately 90% of al1 PSA. Reaction of PSA with ACT

causes inactivation of PSA by the formation of an equimolar (1: 1) ratio complex (PSA-ACT) of

about 90-100 kDa (Lilja et al., 1991). The s e m levels of ACT (0.5 mgfml) and a,M (3 mgM)

are 1O5-106 times greater than the normal PSA levels (4 pa). Therefore, al1 enzymatically

active PSA in serum is expected to form complexes with protease inhibitors (McCorrnack et al.,

1995). When PSA is bound to ACT, a few antigenic determinants are still accessible for

irnrnunologic reactions. However, there are no detectable PSA epitopes when the molecule reacts

with a,M due to the total encapsdation of the PSA molecule (Lilja et al., 1991). f i s complex

c m be detected by Western blotting &a SDS-PAGE (Zhou et al., 1993).

In the serum, fiee PSA rnakes up only a minor hction of the immunoreactive PSA and

most probably represents the proPSA or an internally cleaved form (Liija et al., 1991). Free-PSA

is enzymatically inactive and is incapable of forming complexes with protease inhibitors,

therefore, circulating in the blood as a &, uncomplexed form (Lilja et al., 1991; Christensson

et al., 1993; Abraharnsson and Lilja, 1994). This form of PSA has a half-life of l a s than 2 hrs

and is rapidly elirninated fkom the circulation by renal clearance (Lilja et al., 1991; Stemm et

al., 1991). Nevertheless, the PSA-ACT complex has a prolonged serum half-life of 2-3 days

since it is too large to allow elimination fiom the blood circulation by glomemlar filtration

(Stamey et al., 1987; Oesterling et al., 1988). One hypothesis suggested that the Iiver may aid

in the clearance of the PSA-ACT complex, by binding of the complex to serine proteinasdserpin-

receptors on hepatocytes (Perlmutter et al., 1990; Mast et al., 1991).

1.4.5. PSA from Non-Prostatic Origin

Until recently, it was believed that only qithelial cells of the prostate produced PSA.

This conclusion was based on imrnunoassays and immunohistochernical studies that did not detect

PSA in non-prostatic tissues (ûesterling, 1991). Howeva, as more sensitive assays becarne

available over the last few years (detection limit of -100 ng/L), some reports have challenged the

tissue specificity of PSA. Several groups, using immunohistochemical studies, have proved that

male anal glands and both male and female periurethral glands stained positively for PSA (Pollen

and Dreilinger, 1984; Tepper et al., 1984; Nowels et al., 1988; Karnoshida and Tsutsumi, 1990).

In 1993, Iwakiri and colleagues investigatd the presence of PSA in male urine before and afler

radical prostatectomy. ïhey confimieci that the periurethral glands can produce and secrete PSA

into male urine. Breul and colleagues (1997) also showed the presence of PSA in the urine of

females, receiving continuous testosterone.

Recent studies, rnaking use of ultrasensitive detection-techniques for PSA (detection limit

of 4 0 n a ) , have shown that other human tissues and turnouis can produce PSA at low levels.

For example, PSA was detected in the following tissues: normal endometriurn (Clements and

Mukhtar, 1994), normal breast, tissue of benign breast diseases, and in breast cancers. Highest

expression was seen in benign breast disease and lowest expression in an advanced stage

cancerous tissue (Yu et al., 1996a). It was also observed that post-pregnancy, the normal breast

produces PSA and secretes it into the rnilk of lactating women (Yu and Diamandis, 1995b). The

same group identified PSA also in amniotic fluid (Yu and Diamandis, 1995a) and in colon,

ovarian, liver, kidney, parotid, breast, and adrenal tumours (Diamandis and Yu, 1994; Diamandis

and Yy 1995; Levesque et al., 1995). 'Iherefore, al1 these data show that PSA production is not

14

specifrc to the male prostate but that PSA is a biochemical marker produced by several normal

tissues, turnows, and during pregnancy. Although many tissues produce PSA, it is important to

emphasize that the PSA concentrations detected in these tissues and tumours are much lower than

those produced by normal prostatic cells, by a factor of at least 104 (Diamandis and Yu, 1995).

1.5. GENERAL FEATURES OF STEROID HORMONES AND STERODHORMOIW RECEPTORS

It is now well documented that androgens are the most important factors that affect PSA

production (Jackson et al., 1989; Weber et al., 1989; Bilhartz et ad., 1990; Sherwood et al.,

1990). Consequently, to better understand the regulation of the PSA gene, it is essential ta

recognize the mechanisrns by which steroid hormones and their recepton function. Steroid

hormones are potent chernical messengers that have important effects on ce11 differentiation,

homeostasis, and morphogenesis. Steroid hormones are derived fiom cholesterol and share a

comrnon chernical structure. However, each class of steroids, Le., estrogens, androgens,

progestins, mîneralocorticoids, and glucocorticoids has additional chemical groups bound to the

steroid core. These groups provide specificity to the different steroid classes (Clark and

Markaverich, 1988; Sutherland and Mobbs, 1992). Steroid hormones circulate in the blood and

fieely d i h e through the ce11 membrane (Figure 1-2). However, their biologic effects are

rnanifested only in cells containing intracellulm receptors that recognize specific ligands with

high a£fiity (Strahle et al., 1989; McDonnell et al., 1993). Although the recepton are a

prerequisite for hormone action, they are not sufficient. Ubiquitous and specific factors, which

are present in each particular cell, mediate the fùnction of receptors. Furthemore, the action of

steroid-hormone receptors is limited by the accessibility of genes in chromatin (Tms and Beato,

Nuclear Membrane

// .. Activated \\

Inactive eceptor

@ Steroid Hormone

lnactive Steroid Hormone Receptor

Activated Steroid Hormone Receptor After Steroid Hormone Binding

HRE Hormone Response Element

Figure 1-2. The rnechanisrn of action of steroid homiones. After a passive diffusion thmu& the ce11 membrane, steroid hormones bind to specific intracellular receptors Iocated either in th€ cytoplasm or nucleus. men, activated receptors bind to regulatory elements on the DNA @RE) thereby activating transcription.

16

1993). Many different factors are, therefore, necessary for the complex activation of genes by

steroid hormones.

ïhe steroidhhyroid hormone receptor farnily includes receptors for steroid hormones,

vitamin D, thyroid hormone, and retinoic acid (Yamamoto, 1985). Al1 steroid receptors have a

similar structure. The N-terminal region is essential for the tram-activating function but it is not

very well conserved among the various rnembers of the receptor farnily. The centrril DNA

binding dornain is highly conserved and contains eight conserved Cys residues, which interact

with two zinc ions, and is Arg- and Lys-rich. It is possible to divide the steroid hormone

receptors into two groups, based on the similarity in the DNA binding dornain: one including the

glucocorticoid, progesterone, androgen, and mineralocorticoid receptors, and the other including

the estrogen, thyroid hormone, retinoic acid, and vitamin 4 receptors. Members of the same

group might recognize and regulate transcription through similar target sequences - the hormone

response elements (HRE) - in hormone-responsive promoters (Scheidereit et al., 1983; Umesono

et al., 1986; Beato, 1989; Baniahmad and Tsai, 1993).

The C-terminal region of the receptor contains the ligand binding domain. It also contains

the information for nuclear translocation, receptor dimerization (McDonnell et al., 1993),

interaction with heat shock proteins (Housley et al., 1990), and at least in part of the receptors,

a second tram-activating domain (Dobson et al., 1989). DNA binding and km-activation are

both controlled by ligand binding (Ham et al., 1988; Beato, 1989; McDonnell et al., 1989; Fawell

et al., 1990; Riegrnan et al., 1991).

The cfassical mode1 for steroid hormone action is shown in Figure 1-2. The steroid

hormone passively diffuses into a target ceIl where it binds to a specific int~acellular receptor -

17

located either in the cytoplasm or the nucleus - to fom a complex. Then, the receptor-ligand

complex undergoes a conformational change, the "activation" step. Activation of the receptor

leads to its intracellular translocation, when the receptor is cytoplasmic, and binding to specific

regulatory sequences, the HRE, at the regulatory regions of target genes (McDonnell et al., 1993).

Tsai and colleagues (1988) showed that active steroid receptors bind to the response elements as

dimers. HREs are cis-acting DNA sequences that have the properties of transcriptional enhancers

(Baniahmad and Tsai, 1993). It is important to notice that whle the DNA response element is

necessary for receptor-specific activation of transcription, it is not sufficient on its own. Both

the promoter and the cellular context are important fhctors required for transcriptional activation

(Berry et al., 1990).

1.6. FACTORS AFFEClïNG PSA SYNTHESIS AND EXPRESSION

1.6.1. Steroid Hormones and Steroid Hormone Receptors

Androgens are the rnost important factors that affkct PSA production (Jackson et al.,

1989; Weber et al., 1989; Bilhartz et al., 1990; Sherwood et al,, 1990). Gleave and CO-workers

showed that androgens increased PSA levels in athyrnic mice, bearing LNCaP (androgen-sensitive

hurnan prostate cancer cell-line) xenografh. They observed that serum PSA levels decreased up

to 8-fold following castration of the animals and increased up to 20-fold d e r androgen

supplementation. Furthemore, using Northem blot analysis, they illustrated that the effects of

androgens on PSA production occur at the genetic level, influencing the synthesis of specific

rnRNAs (Gleave et al., 1992).

18

Effects of steroids on target cells are mediated by their respective receptors. Androgen

receptors ( A h ) are structurally and hctionally organized into domains that mediate hormone

binding, nuclear translocation, dimerization, DNA binding, and transcriptional activation (Beato,

1989; Ham and Parka, 1989). Androgen and its agonists upregulate the PSA gene while

antagonists down regulate the gene (Ham et al., 1988; Riegman et al., 1991). Studies showed

that PSA expression is correlated to the ability of ARS to bind to HRE in the PSA promoter.

Inhibition of complex formation between HREs and ARS plays a key role in the down-regulation

of the PSA gene expression in LNCaP cells (Wang et al., 1997). In men, PSA gene regulation

is under the control of testicular androgens through the ARS. The epithelial cells of the prostate

are rich in ARs @vine, 1995). In the prostate, therefore, androgenic steroids can bind to ARS

and regulate transcription of the PSA gene (Henmi et al,, 1992; Young et al., 1992).

Clinical data suggested PSA to be upregulated in sorne bmst turnours and not others.

Furthemore, PSA presence was associated with the presence of both estrogen and progesterone

receptors, however, this association was strongest between PSA and progesterone receptors.

Therefore, it was hypothesized that the presence of PSA in breast cancer might be associated with

the progesterone receptor action (Yu et al., 1994a). To investigate the mechanism of PSA gene

regulation in the breast, Yu et al. (1994b) developed a tissue culture system that reproduced, in

vitro, PSA production by breast cells. In these experiments they grew different cell-Iines in

multi-well plates to near confluence. Then, different steroids were added to the wells to stimulate

PSA production, and PSA concentration was assesseci in the media at different intervals using

an irnrnunofluorometric procedure. Results showed that the stemid-hormone receptor-positive

human breast carcinoma cell-lines T47D and MCF7, did not produce detectable PSA when

19

cultured in media lacking steroid hormones. However, upon stimulation by progestins,

androgens, mineralocorticoids, or glucocorticoids, both cell-lines produced PSA in a dose

dependant manner. Estrogens failed to induce such stimulation in both cell-lines and could

antagonize the induction by androgens in T47D cells. These data show that the estrogen receptor

is not diiectly involved in PSA gene upregulation in this cell-line. It is possible, however, that

a delicate balance between androgens, progestins, and estrogens is important for PSA gene

regulation. The steroid-hormone receptor-negative breast carcinoma cell-line, Bï20, failed to

produce detectable PSA upon stimulation (Yu et al., 1994b).

Recent studies revealed that steroid hormones and their receptors, though necessary, may

not be sacient for PSA production. When SAOS (osteosarcorna) and BG-1 (ovarian

carcinoma) cell-lines, which are positive for estrogen and progesterone reeptors, wae stimulated

with different steroids, PSA production was not detected in the culture media (Zarghami et al.,

1997). It is important to investigate whether the recepton in these cell-lines are functional to

exclude the possibility that lack of PSA production was due to problerns with inactive steroid

receptors.

It is accepted that AR cm mediate PSA production. However, recent in vitro studies

using T47D hurnan breast tumour cells indicate that progesterone receptor can regulate the PSA

gene independently of the AR l i s conclusion is based on the following observations. It was

found that the progesti norgestirnate and triarncinolone acetonide - which cannot bind to the

AR but can bind to the PR - both upregulated the PSA gene at concentrations as low as 1Ol0M.

The antiandrogen cyproterone acetate - which blocks the AR but has progesterone activity -

upregulated the PSA gene at concentrations as low as 10-lOM. Also, the antiprogestin

20

mifepristone completely blocked the stimulation of the specific progestin norgestimate (Zarghami

et al., 1997). In women, PSA is produced in organs rich in steroid hormone receptors, such as

the breast (Yu et al., 1994a) and the endometriurn (Clements and Mukhtar, 1994). Based on the

in vitro stimulation experiments of the T47D cell-line, Zarghami and colleagues (1997) suggested

that in women, progestins and androgens rnight mediate PSA production îhrough the independent

action of progesterone- and androgen-receptors.

1.6.2. Growth Factors and Other Molecules

Some observations suggest that the PSA gene can be activated in a hormone-independent

way. Culig and CO-workers (1994) found that several growth factors, such as insulin-like growth

factor-1 (IGF-1), keratinocyte growth factor, and epidermal growth factor (EGF), can activate

androgen receptors without îhe presence of androgens. The molecular mechanisms involved in

the interaction between growth factors and androgen receptors are not clearly understood.

However, it is possible that phosphorylation of the receptors by the growth factors rnight play

a role in this regdation (Ikonen et al., 1994; Kuiper and Brinkrnann, 1995).

Autocrine and paracrine growth factors have been implicated as possible mediators of

androgen-induced growth pathways and prostate cancer progression (Mydlo et al., 1988;

Schuurmans et d, 1989; Wilding et al., 1989a; Wilding et al., 1989b; Henttu and Vihko, 1993).

PSA expression and synthesis are typical indicators of AR activation in prostatic cell-Iines. As

a result, PSA production can be usefiil to assess possible interactions of growth factors with the

androgen signal transduction cascade. Studies showed that PSA mRNA expression in LNCaP

cells was increased by transforming growth factor (TGF) fl and decreased by basic fibroblast

2 1

growth factor (bFGF). However, changes in PSA expression did not parallel changes in ceIl

growth rate produced by these growth factors. Therefore, it was suggested that changes in PSA

production do not simply reflect changes in the growth rate (Gleave et al., 1992). Ford and

colleagues (1985) showed that there are no changes in the expression of TGF-B, TGF-4 and

bFGF in response to in vitro androgen stimulation. Therefore, these results indicate that PSA

production and LNCaP ce11 growth rate are independent processes (Gleave et al., 1992) and

growth factors are not induced in cells undagoing androgen-induced changes in PSA expression

(Ford et al., 1985).

Henttu and V i o (1993) also found that growth factors not only affect the rate of ce11

division of LNCaP cells but also regulate the biosynthesis and secretion of prostatic proteins,

such as PSA. EGF and TGF-a decreased PSA mRNA synthesis, thus, reducing the secretion of

PSA by LNCaP cells. Furthermore, their study showed that EGF and TGF-a altered the

androgen regulation of the PSA gene by altering levels of the androgen receptor. In contrast to

previous conclusions (Ford et al., 1985), these data indicate that the androgen-regulatory system

may interact with the growth factor regdatory systerns at multiple levels in prostatic cells,

emphasizing the complexity of PSA biosynthesis.

Killian and colleagues (1993) revealed an interesting relationship between TGF-P and

PSA. TGF-P is a multi-fùnctional regulator of cellular growth, differentiation, and development.

It is usually described as a negative growth regulator, which restricts turnour progression (Sporn

and Roberts, 1985). However, there is some evidence to suggest that turnom secrete high levds

of TGF-P @erynck et al., 1987). Furthermore, TGF-P has a strong rnitogenic effect on

osteoblasts and fibroblasts in vitro. This growth factor is produced in an inactive form by several

22

tissues, including the prostate, and must be activated to carry out its biological hction. Recent

studies have shown that PSA can proteolytically cleave and activate TGF-P, which then can

stimulate production of PSA (Killian et al., 1993).

In prostatic cells, other molecules such as retinoic acid (RA) also regulate PSA

production. RA is the rnost biologically active and physiologically abundant metabolite of

vitarnin A (Sporn and Roberts, 1983), and its receptor is a member of the steroidlthyroid

hormone receptor family (Petkovich et al., 1987). In contrast to the promoting effect of

androgens on tumour development (Schuurmans et al., 1989), it has been suggested that retinoids

have the potential to inhibit cancer growth and promote cellular differentiation. RA induced a

dose-dependent increase in PSA secretion by LNCaP cells at concentrations of 0.1 to 1 PM.

Furthemore, RA also suppressed the proliferation of LNCaP cells at the same concentrations

pong et al., 1993). Taken together, these data suggest that PSA might be important for the

inhibition of ce11 proliferation. However, Young and colleagues (1994) found that highe~

concentrations of RA (10 FM) can inhibit androgen-regulated genes such as PSA by reducing the

level of A&. Their fmdings propose that the suppression of proliferation and function ol

prostatic cells by RA may be via modulatory effects on the AR, suggesting that RA does no1

directly affect PSA expression. ï he mechanism by which RA regulates the AR gene, is no1

known and remains to be determineci.

Substances related to tumour-ceIl microenvironment or to the extracellular matrix may

also influence the production of PSA (Sherwood et al., 1990). Guo and colleagues (1994)

showed that extracellular matrix components, present in Matrigel" induce a transient decrease

in the levels of PSA mRNA in LNCaP cells. Al1 these different studies show that the regulation

23

of PSA biosynthesis is a complex process and rnight involve the coordinated and cooperative

action of many different molecules and growth factors.

1.7. PSA FUNCïiON

Both the molecular and the biochemical properties of PSA have been well characterized.

Nevertheless, ow knowledge of PSA's physiological properties is still lirnited and confined to

its role in the seminal plasma. As described previously, PSA is a serine protease with a

chymotrypsin-like activity. ïhe main physiological substrates for PSA identified in the seminal

plasma are the gel-forming proteins produced and secreted h m the serninal vesicles (Wang et

al., 1982; Lilja et al., 1987; McGee and Hm, 1988). Semenogelin 1 and II are the two major

gel-forming proteins in the seminal plasma, whereas fibronectin, which is also a component of

the extracellular rnatrix, constitutes a srnaller part. At ejaculation, the gel-fonning proteins are

degraded into smaller, soluble hgrnents due to the proteolytic activity of PSA. As a result, the

serninal coagulum formed at ejaculation is liquified rapidly, followed by the release of

progressively rnotile sperrnatozoa (Malm and Lilja, 1995). Therefore, PSA plays an important

role in male fertility (Peehl et al., 1995).

Espaiia and colleagues (1991) showed that following ejaculation, PSA binds and

inactivates PCI. Serninal plasma PCI is secreted as an active protein. The inactivation of PCI

occurs in parallel with the appearance of PCI-PSA complexes and with the dissolution of the gel

stnicture of semen. Results suggested that only 5% of PSA in semen f o m a complex with 40%

of PCI. Therefore, it is believed that PSA rnight modulate the fùnction of PCI by binding to it

(Malm and Lilja, 1995).

24

Except its hc t i on in the semen, little has been reported on the physiological role of PSA

in the prostate and other tissues. Recently, in cuntrast to the conclusions previously reported

(Gleave et al., 1992), Wang and colleagues (1996) found that PSA plays a role in the growth

stimulation of the androgen-responsive prostate cancer cell-line, LNCaP. Other studies, discussed

above, showed that PSA can cleave and activate the latent f om of TGF-P. Active TGF-P is

present in diverse tissues and is vital for nomal cellular growth, differentiation, and development

(Fynan and Reiss, 1993). Furthmore, Killian and CO-workm (1993) revealed that TGF-b has

a strong mitogenic effect on osteoblasts and fibroblasts in vitro. Based on these data, it is

possible that PSA might play an important role in carcinogenesis.

Further clues to the relationship between growth factors and PSA have been found around

1993, when scientists investigated the possible significance of the proteolysis of insulin-like

growth factor binding proteins (IGFBPs) by PSA (Kanety et al., 1993; Cohen et al., 1994).

Insulin-like growth factors (IGFs) stimulate proliferation of many tissues and ce11 types, both in

vivo and N1 vitro. Furthmore, they may act as endocrine, paracrine, and autocrine modulators

of normal and malignant proli ferat ion (Daughaday and Rotwein, 1 989; Daughaday, 1 990). It was

shown that IGFs could also act as rnitogens and promote the growth of breast tumour cells (Myal

et al., 1984; Arteaga, 1992). In serum, IGFBPs bhd IGFs with high aifiinity, thus modulating

their proliferative and rnitogenic effects (HXdouin et al., 1989; Rosenfeld et al., 1990; Cohen et

al., 1991; LeRoith et al., 19%). The molecular mechanisrns, involved in the interaction of the

IGFBPs with the IGFs and their receptors, are still unclear. However, it seerns that IGFBPs

regulate the availability of fi-ee IGFs for interaction with IGF receptors (Rosenfeld et al., 1990;

LeRoith et al., 1992). Thus, P S A may upregulate ce11 growth by cleaving IGFBPs.

25

In plasma, IGFBP-3 is the major IGFBP (Cohen et al., 1994). Kanety and colleagues

(1993) observed a significant negative correlation between serum IGFBP-3 and PSA in patients

with metastatic prostate cancer, i.e., decreased levels of IGFBP-3 in these patients were correlated

with higher PSA levels. Cohen and co-workers (1994) found that PSA could cleave IGFBP-3

to form a lower af5nity fiagment. ïhis cleavage of IGFBP-3 abolished its inhibitory effect on

IGF-1-stirndated growh of prostate cells in vitro. Yu and colleagues (1996a) proposed that fiee

IGF-1 couid then bind ARS to induce PSA production and to exert other physiological actions.

This hypothesis offers a regulatory loop between PSA, growth factors, and growth factor binding

proteins in prostate cancer. It is possible that PSA might be important for the growth regulation

and proliferation of prostate cancer cells. In breast tumows, however, no correlation or

association between PSA and IGFdIGFBPs was detected (Yu et al., 1996b). Thus, this

hypothesis offers no explanation for the role of PSA in breast tumows.

Yoshida and CO-workers (1995) suggested PSA to be an initiator of the protease cascade

involved in prostate cancer invasion and metastasis. Tumour ce11 invasion into the adjacent tissue

requires pericellular proteolysis of the extracellular matrix. Urokinase-type plasminogen activator

(LPA) activates plasminogen to plasrnin, which degrades different components of the extracellular

ma@ such as, fibronectin, laminin, proteoglycans, and collagen ( D m et al., 1985). Plasrnin

also activates latent collagenase (Testa and Quigley, 1990) and growth factors (Lyons et al.,

1988). Tumour cells or interstitial cells produce and secrete uPA as a single-chah zymogen form

(scuPA), which is bound to the uPA receptor on the tumour ce11 surface. Expression of uPA and

the uPA receptor by turnour cells results in a high invasive capability (Ellis et al., 1992; Kwaan,

1992). It has k e n reported that PSA might be responsible for the initial generation of active

26

uPA in prostate cancer. niese results suggest an important role for PSA in the invasion and

metastatic processes in prostate cancer (Yoshida et al., 1995). Unfortunately, this data cannot

provide clues to the importance of PSA in breast cancer since PSA presence in breast tumours

was associated with a good prognosis (Yu et al., 1995). It is possible that diEerent

concentrations of PSA, such as those found in breast and prostate cancer, may result in distinct

effects on ce11 fimctions. Therefore, to better understand PSA's role, it is important to investigate

what effects different PSA levels have on different ce11 types.

PSA is present in about 30% of breast tumeurs (Yu et al., 1994a), and some breast cancer

cell-lines can produce PSA in vitro upon stimulation with steroids (Yu et al., 1994b). In an

attempt to understand the possible role of PSA in the breast, Lai and culleagues (1996) studied

the effect of PSA on growth of sorne human breast mm cell-lines. They found that low

concentrations of PSA (0.001 pg/L to 0.1 y&) inhibited MCF7 (steroid hormone receptor

positive cell-line) ce11 growth, while higher PSA concentrations (1 p g L to 100 pgk) did not.

PSA did not affect the steroid hormone receptor negative cell-line, MDA-MB-231. îlese data

suggest PSA rnay act as a negative growîh regulator in hormone-dependant breast cancer over

a limited concentration range (Lai et al., 1996).

1.8. PSA AND BREAST CANCER

PSA irnrnunoreactiviiy was examined in a cohort of more than 1,200 breast cancer

patients, with a sensitive time-resolved irnrnunofluorometric assay. At the cut-off leveI of 30

pglmg of total protein, immunoreactive-PSA was detected in greater than 30% of fernale breast

turnour cytosolic extracts. Some of these breast tumours contained PSA levels of 50 @mg of

27

total protein or even higher (Diamandis et al., 1994; Yu et al., 1994a). 'The molecular weight

of PSA in female breast tumours, as detected by hi& performance liquid chromatogaphy

(HPLC) and Western blot analysis, was identical to that of seminal PSA and fiee senun PSA (-33

kDa). These data suggest PSA in the fernale samples to be enzyrnatically active (Diamandis et

al., 1994). Other studies, however, showed that both fiee PSA and the PSA-ACT complex were

present in breast turnour cytosols (Wu et al., 1995).

Yu and colleagues (1995) measured immunoreactive PSA in tumour cytosols of 174 breast

cancer patients, and classifiecl the breast cancers as PSA-positive or PSA-negative, based on a

cutoff level of 30 pg/mg of total protein. To fmd the precise contribution of PSA to the patients'

risk for survival, it was necessaty to m i d e r the information provided by other parameters in

the study, such as age, clinid stage, tumour size, histological grade, nodal status, and ER- and

PR-status. Afta adjusting for most of these variables, a rnultivariate analysis showed that breast

cancer patients with PSA-positive turnours had a reduced risk for relapse. k e data suggest

PSA to be an independent favourable prognostic rnarker for some breast cancer subgroups. Later

studies by the same group showed that the presence of PSA in tumour samples was related to

other variables associated with a good prognosis. For exarnple, the proliferative activity,

represented as the fiaction of cells in S-phase, was significantly lower in PSA-positive tumows

than in PSA-negative tumours. Furthemore, PSA was more fiequently found in diploid turnours

(Levesque et al., 1995). These studies reveal that the presence of PSA in breast turnours may

be usefiil as a new biological rnarker to distinguish a subgroup of breast cancer patients who have

a good prognosis.

1.9. BREAST CANCER

1.9.1. Epiderniology

In North America, breast cancer is the most fiequently diagnosed tumour in women.

Although the rates of breast cancer are much lower in Asian countries, they are increasing rapidly

in that region and rising slowiy in other developed countries (Parkin et al., 1993). In the United

States, about 44,000 women die of breast cancer every year, and each of them typically loses

nearly two decades of life. The loss in Europe is sirnilar (Parkin et al., 1992; Coleman et al.,

1993; Wingo et al., 1995). It is estimated that one of every eight women in the United States

will develop breast cancer at some time during ha life (Feuer et al., 1993).

Despite advances in the early detection and treatment of breast cancer, the mortality rate

fiom this disease has not changed appmiably in the last 50 years (Harris et al., 1992).

Epidemiological studies have identified various risk factors contributing to the development of

breast cancer. Cumulative exposure to estrogens underlies most of the known risk factors.

Therefore, past epidemiological research on brmt cancer has focused to a large extent on cultural

and lifestyle differences that may explain international variations in the incidence of breast cancer

in t e m of reproductive hormone levels. It is acceptai that the risk of breast cancer is related

to the hormonal exposwe. ïherefore, the earlier in life menses begins, the longer monthly cycIes

extend, and the earlier and longer birth control pills and estrogen replacement are in use, the

greater a wornan's risk of breast cancer (Kelsey and Bernstein, 1996).

A farnily history of breast cancer increases the likelihood of breast cancer development.

This may be due to the genetic and environrnental similarities among farnily members (Hulka and

Stark, 1995). Early age at onset is the strongest indicator of genetic susceptibility (Slattery and

29

Kerber, 1993). Overall, only 10-15% of breast cancer is attributed to f&ly history, and about

half of this is due to dominantly inherited susceptibility genes, such as BRCAl and BRCA;!

(Weber and Garber, 1993).

Other factors that affect the susceptibility for breast cancer are often related to diet and

lifestyle. For instance, high fat diet, charbroiled food, vitamin D deficiency, smoking, and

alcohol may increase the risk for breast cancer. Physical activity, however, reduces the risk for

developiig the disease (Kelsey and Bernstein, 1996). T h e of life, when exposwes to chernical

carcinogens such as cigarettes and pesticides take place, is also an important factor in breast

cancer. î l e period between age at menarche and age at fmt fùll-term pregnancy has been

identified as a tirne when breast tissue rnay be more susceptible to damage (Russo et al., 1990).

Dietary and pharmacological trials are in progress to assess the potential means for reducing the

incidence of breast cancer. Currently, prevention of breast cancer is dificult since many nsk

factors are correlateci and diEcult to control. To fmd the appropriate preventive measures, it is

important to understand the etiology of this disease and the mechanisms involved in its initiation.

Studies of the mechanisms by which endogenous hormones have their effects will lead to

increasd understanding of risk factors and preventive maures.

1.9.2. Mammary Gland and Mammary Cancers

The rnammary gland consists of a secretory epithelium (parenchyma) embedded in a

stroma, which is made up of connative tissue, adipose tissue, and a vascular network. Two

types of epithelial cells make up the parenchyrnal compartment: luminal and myoepithelial.

Luminal rnammary epithelial cells (LMECs) make up the inner lining of ducts and alveoli, which

30

are important for the synthesis and passage of milk via the nipple. The myoepithelial cells are

the contractile elements needed for expulsion of niilk. Both the stroma and the myoepithelial

cells produce extracellular matrix @CM) components. ECM, hormones, and growth factors are

essential for the growth, fiinction, and morphogenesis of the marnrnary gland (Nandi et al., 1995).

Studies have shown that an appropriate homona1 milieu is necessary for proliferation 01

mammary epithelial cells and is aiso a prerequisite for breast carcinogenesis (Reid et al., 1996).

The capacity of normal breast tissue to develop depends on age, composition of breast tissue, a n d

hormonal environment. Most proliferation of the lobules occm during the teenage period aiter

puberiy (Hughes et al., 1989). Additionally, during each monthly cycIe between menarche ano

menopause the ductal epithelium is renewed continuously due to the cyclical menstnial hormone

(Furnival, 1986). Estrogen, and to a lesser extent, progesterone, are mitogenic to breast tissue,

Changes in the levels of these steroid homones during the menstnial cycle profoundly influence

both the stroma and the epithelium (Vogel et al., 1981). Childbirth, with its post-lactational

involution, causes the lobular structure to become more differentiated. As a result, there is i

decrease in lobules without any significant changes in the ducts and the connective tissue stroms

(Russo and Russo, 1995). During menopause, the ductal and lobular epithelium, as well as th€

adjacent fibrous connective tissue stroma, regress and each is gradually replaced by adipose tissu

(Hughes et al., 1989).

Over 85% of the spontaneous mammary cancers originate in LMECs. Breast cancers art

classifieci as either homone-dependent tumours (HDTs) or hormone-independent tumours (HITs),

HDTs require the presence of hormones - which are produced by the ovaries and the pituitary -

for their proliferation, whereas HITs do not. Both types of tumours can undergo furSie1

3 1

progression, giving rise to turnours of a heterogeneous phenotype (Nandi et al., 1995). Breast

cancer is a heterogeneous disease, causing difficulties in selecting the appropriate therapy for

many patients. Therefore, the use of prognostic markers can aid clhicians in identifying

subgroups of patients, providing them with better treatment.

1.9.3. Prognostic Markers in Breast Cancer

The ultimate goal of clinical prognostic rnarkers is to provide information that will allow

physicians to choose a specific and beneficial therapy for each patient, thereby eliminatimg

micrometastases and increasing the disease-fke survival. Fifty percent of women with primary

operable breast cancer are node-negative and have a favomble prognosis. lhey either sustain

a long term remission or are cured of the disease by surgery alone or in combination with

radiation. It is important to differentiate this group of patients fiom the node-positive patients

who have an unfavourable prognosis. ïhis classification allows physicians to identifj the patients

who might benefit fiom adjuvant therapy and spare others the side effects (Fisher et al., 1985).

Adjuvant therapy is given to patients afler surgery to prevent or minimize the growth of

micrometastases that might grow into a recurrent tumow. Some common drug combinations that

are aven to premenopausal breast cancer patients at high risk for residual rnicroscopic metastases

are: cyclophosphamide plus methotrexate plus 5-fluoroumcil (CMF), doxombicin plus

cyc1ophosphamide (AC), and cyclophosphamide plus doxorubicin plus 5-fluorouracil (CAF).

Tamoxifen is often given to postmenopausal women who are at high risk for disease recurrence

(Dollinger et al., 1995).

32

Although it is possible to anticipate whether patients would respond to endocrine therapy,

it is still dificult to predict whether patients would benefit h m adjuvant therapy (Porter-Jordan

and Lipprnan, 1994). It would be usefùl, both medically and economically, if we could identiS,

node-negative women who are at a very low nsk of recurrence, and thus, do not need adjuvant

therapy. Neverthelas, a relapse rate of 25-3O% has been reported in node-negative patients, and

thus it is important to identifj these patients for additional treatnient (Copper, 1991). Moreover,

it would be important to identia through a combination of prognostic markers, which node-

positive tumom are very aggressive, allowing patients to consider more intensive or

investigational therapies (Elledge et al., 1992; Reynolds, 1994).

Current prognostic markers can be divided into two groups: host factors (e.g., age,

menopausal status, infiammatory response) and tumour factors (e.g., size, grade, vascular

invasion, nodal involvement, estrogen and progesterone receptor level, ploidy and S-phase

determination). Tumour six, histological grading, nodal status, and ER status, are the most

widely accepted and used indicators in breast cancer. Of al1 prognostic factors, tumour s i x is

the simplest to measure, cheapest, and is a good indicator of the disease stage. Increasing tumour

size is related to an increasing probability of positive nodes and higher risk of recurrence and

death (Fisher et al., 1985). ER status is a predictive- rather than a prognostic-factor since it

predicts response to hormone therapy and does not indicate the general outcorne (Reynolds,

1994).

One major problem in the breast cancer prognostic field is that rnost of the prognostic

markers are highly correlated. As a result, the information they provide is largely redwidant.

The challenge for researchers, therefore, is to discover the best independent predic tors, b y using

33

multivariate analyses (Reynolds, 1994). Then, a combination of thme independent factors will

allow a more accurate identification of additional subsets of patients that do worse or better than

expected soleiy based on nodal status (Elledge et al., 1992). Since PSA was found to be an

independent favourable prognostic marker for some breast cancer subgroups, it is of great interest

to understand what role PSA plays in the female breast. In the future, these studies might aid

scientists and clhicians to design new therapies based on the PSA concentrations in the breast

tumours.

1.10. THESIS OUTLINE

PSA is a valuable tumour rnarka that is used for the diagnosis and monitoring of patients

with prostate cancer @iarnandis and Yw, 1995). In the past few years, several PSA-produchg

prostate cancer models have been developed to fùrther understand the biology of prostate cancer

(Papsidero et al., 1981; Horoszewicz et al., 1983; Gleave et al., 1991; Gleave et al., 1992). PSA

was also detected in more than 30% of hurnan breast turnom, although in much lower

concentrations (Diamandis et al., 1994; Yu at al., 1994a), and was suggested to be a good

prognostic marker for some breast cancer subgroups (Yu et al., 1995). Furthemore, in vitro

results showed that the steroid hormone receptor positive hurnan breast cancer cell-lines, T47D

and MCF7, could produce PSA upon induction with steroids (Yu et al., 1994b). However, so

far, no in vivo model for PSA production by breast tumours has b e n estabiished.

The purpose of the present study, therefore, was to develop a PSA-producing breast cancer

model, using human breast cancer cell-lines growing in severe combined imrnmodeficient (SCID)

mice. The availability of ultrasensitive irnmunofluorometric assays aIlowed the detection and

3'

quantification of low PSA leveIs present in the breast tumour xenografts. This mouse model foi

PSA production allows in vivo evaluation of the replation of PSA synthesis and expression. Th(

development and documentation of the in vivo model for PSA induction is discussed in Chaptei

2. Chapter 3 summarizes the conclusions fiom the in vivo study and provides some preliminaq

data regarding the production of PSA in vitro, as a fiinction of ce11 growth, as well as t h

possible role of exogenous PSA on the growh rate of T47D cells. Chapter 3 also introduce

three other in vivo mouse models for PSA induction using different breast cancer ce11 lines

Proposed experiments for the future complete Chapter 3.

Abrahamsson, P. A. and Lilja, H (1994). Free and complexed f o m of prostate-specific antigen in serum. Tmour M a r k W t e 6,l-4.

Akiyama, K., Nakamura, T., Iwanaga, S., and Hara, M. (1987). ïhe chymotrypsin-like activity of human prostate-specific antigen, y-seminoprotein. FEBS Lett. 225, 168-172.

Arteaga, C. L. (1992). Interference of the IGF system as a strategy to inhibit breast cancer growth. Breast Cancer Res. Tr. 22, 10 1-1 06.

Baniahrnad, A. and Tsai, M. (1993). Mechanisrns of transcriptio~ial activation by steroid-hormone receptors. J: Cell. Biochem. 51, 151-156.

Beato, h4. (1989). Gene regulation by steroid hormones. Cell56,335-344.

Belanger, A., van Halbeek, H., Graves, H. C., Brandbois, K., Starney, T. A., Huang, L., Poppe, I., and Labrie, F. (1995). Molecular mass and carbohydrate structure of prostate specific antigen: Studia for establishment of an international PSA standard. Prostate 24, 187-197.

Berry, M, Metzger, D., and Chambon, P. (1990). Role of the two activating domains of the oestrogen receptor in the cell-type and promoter-context dependent agonist activity of the anti- oestrogen 4-hydroxytamoxifen. &?BO J. 9, 281 1-281 8.

Bilhartq K. L., Young, C. Y. F., Flanagan, W. F., He, W., and Tindall, D. J. (1990). Androgen regulation of prostate specific antigen (PSA) in human prostate carcinoma. J Urol. 143,203A.

Breul, J., Pickl, U., and Jürgen, S. (1997). Extraprostatic production of prostate specific antigen is under hormonal control. J Uiology 157, 212-213.

Chapdelaine, P., Paradis, G., Tremblay, R R, and Dubé, J. Y. (1988). High level of expression in the prostate of a human glandular kallikrein mRNA related to prostate-specific antigen. FEBS Lett. 236, 205-208.

Christensson, A., Laurell, C. B., and Lilja, H. (1990). Enzymatic activity of the prostate-specific antigen and its reactions with extracellular serine protehase inhibitors. Eur. J Biochem. 194,755- 763.

Christensson, A., Bjork, T., Nilsson, O., Dahlen, U., Matikainen, T., Cockett, A. T. K., Abrahamsson, P. A., and Lilja, H. (1993). S e m prostate-specific antigen complexed to cxl- antichymotrypsin as an indicator of prostate cancer. J Urology 150, 100-1 05.

Christensson, A. and Liija, H. (1994). Complex formation between protein C inhibitor and prostate-specifk antigen in vitro and in human semen. Ew. J Biochem. 220,45-53.

Clark, J. H. and Markaverich, B. M. (1988). Actions of ovarian steroid hormones. In: f ie Physiology of Reprodudion. (eds. Knobil, E. and Neill, S.). New York, NY: Raven Press, pp. 675-724.

Clements, J. A. (1994). n e human kallikrein gene family: a diversity of expression and fùnction. Mol. Cell. Ena'ocrinol. 99, Cl-C6.

Clements, J. and Mukhtar, A. (1994). Glandular kallikreins and prostate-specific antigen are expressed in the hurnan endometrim. J Clin. Endocr. Metab. 78, 1536-1539.

Cleutjens, K. J. M, van Eekelen, C. C. E. M., van d a Korput, H. A. G. M., Brinkmann, A. O., and Traprnan, J. (1996). Two androgen response regions mperate in steroid hormone regulated activity of the prostate-specific antigen promoter. J. Biol. C h m 271,6379-6388.

Cohen, P., Fielder, P. J., Hasegawa, Y., Frisch, H., Giudice, L. C., and Rosenfeld, R G. (1991). Clinical aspects of insulin-like growth factor binding proteins. Acta Enddnol. 124,74-85.

Cohen, P., Psehl, D. M., Graves, H. C. B., and Rosenfeld, R G. (1994). Biologicd effécts of prostate-specific antigen as an insulin-like growth factor binding protein-3 protease, J Endocriml. 142,407-41 5.

Coleman, M. P., Estèva, J., Damiecki, P., ArsIan, A., and Renard, H. (1993). T r e d in Cancer Incidence and Mortality. Lyon, France: M C Scientific Publications.

Copper, M. R (1991). The role of chemotherapy for node-negative breast cancer. Cancer 67, 1744- 1747.

Culig, Z., Hobisch, A., Cronauer, M. V., Radrnayer, C., Traprnan, J., Hittmair, A., Bartsch, G., and Klocker, H. (1994). Androgen receptor activation in prostatic turnor ce11 lines by insulin-like growth factor-1, keratinocyte 6ow-h factor, and epidernd growth factor. cancer es. 54,5474- 5478.

D m , K, Andersen, A, Grmdahl-Hansen, J., Kristensen, P., Nielsen, L. S., and Slu-iver, L. (1985). Plasrninogen activators, tissue degradation and cancer. A h . Cancer Res. 44, 139-266.

Daughaday, W. H. and Rotwein, P. (1989). Insulin-like growth factors 1 and 2. Peptide messenger ribonucleic acid, gene structures and serurn and tissue concentrations. Endocr, Rm 10, 68-91.

Daughaday, W. H. (1990). Editorial: the possible autocrinelparacrine and endocrine roles of insulin-like growth factors of human tumours. Endocrinology 127, 1-4.

Derynck, R, Goeddel, D. V., Ullrich, A., Guttennan, J. U., Williams, R D., Bringman, T. S., and Berger, W. H. (1987). Synthesis of messenger RNAs for transforrning growth factors alpha and beta and the epidermal growth factor receptor by h u m turnours. Cancer Res. 47, 707-712.

Diamandis, E. P., Yu, H., and Sutherland, D. J. A. (1 994). Detection of prostate specific antigen irnmunoreactivity in breast tumors. Bremt Cuncer Res. Tr. 32,301-3 10.

Diamandis, E. P. and Yu, H. (1994). Prostate specific antigen and lack of specificity for prostate cells. k e t 345, 1 186.

Diamandis, E. P. and Yu, H. (1995). New biological functions of prostate specific antigen? J C h Endom Metab. 80, 151 5-1517.

Dollhger, M., Rosenbaum, E. H., and Cable, G. (1995). Everyone's Guide to Cancer Therapy: How Cancer is Diagnosed, Treated, and Managed Day to Day. 2nd ed. (ed. Hasselback, R). Toronto, ON: Somarville House.

Dubson, A. D. W., Conneely, O. M., Beattie, W., Maxwell, B. L., Mak, P., Tsai, M. J., Schrader, W. T., and O'Malley, B. W. (1989). Mutational analysis of the chicken progesterone receptor. J Biol. Chem. 264,4207-42 1 1.

Dnnk'livater, C. C., Evans, B. A., and Richards, R I. (1988). Kallikreins, kinins, and growth factor biosynthesis. T r e d Biochem Sci. 13, 169-1 72.

Elledge, R M., McGuire, W. L., and Osborne, C. K. (1992). Prognostic factors in breast cancer. Semin. ûncol. 19,244-253.

Ellis, V., Pyke, C., Eriksin, J., Solberg, H., and Ilam, K. (1992). The urokinase receptor: involvement in cd-surface proteolysis and cancer invasion. Ann. AT Acad Sei. 667, 13-3 1.

Espafia, F., Gilabert, J., Estellés, A., Romeu, A., h a r , J., and Cabo, A. (1991). Functionally active protein C inhibitorlplasminogen activator inhibitor-3 (PCYPAI-3) is secreted in seminal vesicles, occurs at high concentrations in human seminal plasma and complexes with prostate- specific antigen. ïhromb. Res. 64, 309-320.

Espaiïa, F., Sanchez-Cuenca, J., Vera, C. D., Estelles, A, and Gilabert, J. (1993). A quantitative ELISA for the measurement of complexes of prostate-specific antigen with protein C inhibitor when using a purified standard. J Lab. C h Med 122,711-719.

Fawell, S. E., Lees, J. A,: White, R, and Parka, M. 0. (1990). Characterimtion and colocalization of steroid bindmg and dirnerization activities in the mouse estrogen receptor. Ce11 60,953-962.

Feuer, E. J., Wun, L. M., Boring, C. C., Flanders, W. D., Tirnrnel, M. J., and Tong, T. (1993). The life-time risk of developiig breast cancer. J Natl. Cancer 1. 85, 892-897.

Fisher, B., Bauer, M., Margolese, R , Poisson, R , Pilch, Y., Redmond, C., Fisher, E., Wolmark, N., Deutsch, M., Montahue, E., Saffer, E., Wickerham, L., Lerner, H., Glass, A., Shibata, H., Deckers, P., Ketcham, A., Oishi, R, and Russell, 1. (1985). Five-year results of a randomized clinical trial cornparhg total mastectomy and segmental rnastectomy with and without radiation in the treatment of breast cancer. N w Engl. J Med 312, 665-672.

Fong, CC., Sutkowski, D. M., Braun, E. J., Bauer, K. D., Shenvood, E. R, Lee, C., and Kozlowski, J. M. (1993). Effect of retinoic acid on the proliferation and secretory activity of androgen-responsive prostatic carcinoma cells. J Urology 149, 190-1 194.

Ford, T. F., Butcher, D. N., Masters, J. R W., and Parkinsson, M. C. (1985). Immunocytochemical localization of prostate-specific antigen: specificity and application to clinical practice. Brit. J Urol. 57, 50-55.

Furnival, C. M. (1986). Function of the normal breast: Physiology and endocrine control. In Clinical Surgery (Cl'ol.lO): Breast Disease (ed. Forbes, J. F.). Ediibmgh, Sçotland: Churchill Livingstone, pp. 1-19.

Fynan, T. M. and Reiss, M. (1993). Resistance to inhibition of ce11 growth by transforming growth fator-p and its role in onçogenesis. Crit. Rev. Oncogenesis 4,493-540.

Gleave, M., Hsieh, J. T., Gao, C., von Eschenbach, A. C., and Chung, L. W. K. (1991). Acceleration of human prostate cancer growth in vivo by factors produced by prostate and bone fibroblasts. Cancer Res. 51, 3753-3761.

Gleave, M. E., Hsieh, J-T., Wu, H-C., von Eschenbach, A. C., and Chung, L. W. K. (1992). Serum prostate specific antigen levels in mice bearing human prostate LNCaP tumors are determuid by tumor volume and endocrine and growth factors. Cancer Res, 52, 1598-1605,

Guo, Y., Pili, R, and Passatiti, A. (1 994). Regulation of prostate-specific antigen gene expression in LNCaP human prostatic carcinoma cells by growth, dihydrotestosterone, and extracellular matrix. Prostate 24, 1-10.

van Halbeek, H., ûerwig, G. J., Vleigenthart, J. F., Tsuda, R, Ham, M., Akiyarna, K., and Schrnid, K. (1985). Occurrence of the Y determinant on the N-glycosidic carbohydrate units of human gamma-seminoprotein. Biochem Bioph Res. Co. 131, 507-5 14.

Ham, J., ïhomson, A, Beedham, M., Webb, P., and Parker, M. (1988). Characterization of response elernents for androgen, glucocorticoids, and progestins in mouse marnmary tumour virus. Nucleic Acidr Res. 16, 5263-5275.

Jih, J. and Parker, M. G. (1989). Regdation of gene expression by nuclear honnone receptors. Cm. @in Ce11 Biol. 1, 503-5 1 1.

Hara, M, Koyanagi, Y., Inoue, T., and Fukuyama, Y. (1971). Some physico-chernical characteristics of "y-seminoprotein", an antigenic cornponent specific for hurnan seminal plasma. Forensic immunological study of body fluids and secretions VII. Jpn. J k g a l Med 25,322-324.

Hardouin, S., Gounnelen, M., Noguiez, P., Seurin, D., Roghani, M., Le Bouc, Y., Povoa, G., Merimee, T. J., Hossenlopp, M. P., and Binoux, (1989). Molecular forms of s e m insulin-like growth façtor (1GF)-binding proteins in man: relationships with growth homone and IGFs and physiological significance. J Clin. Endocr. Metab. 69, 129 1-1301.

Harris, J. R, Lippman, M E., Veronesi, U., and Willett, W. (1992). Breast cancer. Nau Engl. J Med 327, 319-328.

Helzlsouer, R 1. (1993). Epiderniology, early detection, and prevention of breast cancer. Cwr. @in Oncol. 5, 955-959.

Henttu, P. and Vihko, P. (1989). cDNA coding for the entire human prostate sgecific antigen shows high homologies to the hurnan tissue kallikrein genes. Biochem Bioph. Res. Co. 160,903- 910.

Henttu, P., Lukkarinen, O., and V i o , P. (1990). Expression of the gene coding for human prostate-specific antigen and related hGK-1 in benign and rnalignant tumors of the human prostate. I . . J Cancer 45, 654-660.

Henttu, P., Liao, S., and V i o , P. (1992). Androgens upregulate the human prostate-specific antigen messenger ribonucleic acid (rnRNA), but down-regulate the prostatic acid phosphatase mRNA in the LNCaP ce11 line. Endocrirwlogy 130, 766-772.

Henttu, P. and Vihko, P. (1993). Growth factor regulation of gene expression in the human prostatic carcinoma ce11 line LNCaP. C a m Res. 53, 1051-1058.

Henttu, P. and Vihko, P. (1994). Prostate-specific antigen and hurnan glandular kallikrein: two kallikreins of the hurnan prostate. Ann Med 26, 157-164.

Horoszewicz, J. S., Leong, S. S., KaWinski, E., Karr, J. P., Rosenthal, H., Chu, T. hl., Mirand, E. A., and Murphy, G. P. (1983). LNCaP mode1 of hurnan prostatic carcinoma. Cancer Res. 43, 1809-1818.

HousIey, P. R, Sanchez, E. R, Danielsen, M., Rîngold, G. M, and Pratt, W. B. (1990). Evidence that the conserved region in the steroid bindig dornain of the glucocorticoid receptor is required for both optimal binding of hsp 90 and protection fiom proteolytic cleavage. J Biol. Chem. 265, 12778- 12781.

Hughes, L. E., Mansel, R E., and Webster, D. J. T. (1989). Benign Disorders and Diseases of the Breast. London, England: Bailliere Tindall.

Hulka, B. S. and Stark, A. T. (1995). Breast cancer: cause and prevention. Lumet 346, 883-887.

Ikonen, T., Palvirno, J. J., Kallio, P. A., Reinikainen, P., and J b e , O. A. (1994). Stimulation of androgen-regulated transactivation by modulators of protein phosphorylation. Endocrinology 135, 1359-1366.

Iwakiri, J., Grandbois, K, Wehner, N., Graves, H. C. B., and Starney, T. A. (1993). An analysis of urinary prostate specific antigen before and after radical prostatectomy: evidence for secretion of prostate specific antigen by the periwethral glands. J Urology 149,783-786.

Jackson, J. A., Waxman, J., and Spiekerman, A. M. (1989). Prostatic complications of testosterone replacement therapy. Arch Intem. Med 149,2365-2366.

Karnoshida, S. and Tsutsumi, Y. (1990). Extraprostatic localization of prostatic acid phosphatase and prostate-specific antigen: distribution in cloacogenic glandular epithelium and sex-dependent expression in human ana1 gland. H m PafhoI. 21, 1 108-1 1 1 1.

Kanety, H., Madjar, Y., Dagan, Y., Levi, J., Papa, M. Z., Pariente, C., Goldwasser, B., and Karasik, A. (1993). S e m insulin-like growth factor-binding protein-2 (IGFBP-2) is increased and IGFBP-3 is increased in patients with prostate cancer: correlation with serum prostate- specific antigen. J Clin Endocr. Metab. 77,229-233.

Karr, J. F., Kantor, J. A., Hancl, P. H., Eggensperger, D. L., and Schlom, J. (1995). The presence of prostate-specific antigen-retated genes in primates and the expression of recombinant human prostate-specific antigen in a transfected mutine cell-line. Cancer Res. 55, 2455-2462.

Kelsey, J. L. and Bernstein, L. (1996). Epidemiology and prevention of breast cancer. Annu Rev. Publ. Health 17,47-67.

Killian, C. S., Corral, D. A., KaWinski, E., and Constantine, R 1. (1993). Mitogenic response of osteoblast cells to prostate-specific antigen suggests an activation of latent TGF-P and a proteolytic modulation of ce11 adhesion receptors. Biochem Bioph Res. Co. 192, 940-947.

Kraut, J. (1977). Serine proteases: structure and mechanism of catalysis. Annu Ra! Biochem. 46, 33 1-358.

Kuiper, G. G. J. M. and B r i i A. 0. (1995). Phosphotryptic peptide analysis of the human androgen receptor: detection of a hormone-induced phosphopeptide. Biochemisby 34,185 1-1 857.

Kwaan, H. C. (1992). ïhe plasminogen-plasrnin systern in malignancy. Camer Metast. Ra! 11, 291-311.

Lai, L. C., Erbas, H., Lennard, T. W. J., and Peaston, R T. (1996). Prostate-specific mtigen in breast cyst fluid: possible role of prostate-specific antigen in hormone-dependent breast cancer. Int. J Cancer 66, 743-746.

LeRoith, D., Werner, H., Burgueta, B., Roberts, C. T., Mulroney, S., and Haramati, A. (1992). The insulin-like growth factor farnily of peptides, binding proteins and receptors: their potential role in tissue regeneration. A& E q . Med Biol. 321, 21-28.

Levesque, M, Yu, H., D'Costa, M., and Diamandis, E. P. (1995). Prostate specific antigen expression by various twnors. J Clin. Lab. And. 9, 123-128.

Levine, A. C. (1995). Pathogenesis and medical management of benign prostatic hyperplasia. Trends Endocrinol. Metab. 6, 128- 132.

Li, T. S., and &lin& C. G. (1973). Isolation and characterization of two specific antigens of hurnan seminal plasma. Fertil. Steril. 24, 134-143.

Lilja, H. (1985). A kallikrein-like serine protease in prostatic fluid cleaves the predominant seminal vesicle protein. J Clin Imest. 76, 1899-1903.

Lilja, H., Oldbring, J., Rannevik, G., and Laurell, C. B. (1987). Seminal vesicle-secreted proteins and their reactions during gelation and liquefaction of hurnan semen. J Clin, Invest. 80,281-285.

Lilja, H., Christensson, A., Dahlen, U., Matikaiien, M. T., Nilsson, O., Pettersson, K., and Lovgren, T. (1991). Prostate-specific antigen in hurnan senun occm predominately in complex with alpha-1 -antichyrnoüypsin. Clin. Chem 37, 16 18-1 625.

Lundwall, and Li'ja, H. (1987). Molecular cloning of human prostate specific antigen cDNA. FEBS k t t . 214, 3 17-322.

Lyons, R M, Genrty, L. E., Pwchio, A. F., and Moses, H. L. (1988). Mechanism of activation of latent recombinant transforrning growth factor pl by plasmin. J Ce11 Bol. 110, 1361-1 367.

MacDonald, R J., Margolius, H. S., and Erdos, E. G. (1988). Molecular biology of tissue kallikrein. Biochem J: 253, 3 13-321,

Malm, J. and Lilja, H. (1995). Biochernistry of prostate specific antigen, PSA. Scund J Clin. Lab Iw. 221 (55 S~ppl.) , 15-22.

Mast, A. E., Enghld, J. J., P i m , S. V., and Salvesen, G. (1991). Analysis of the plasrna elimination kinetics and conf'orrnationaI stabilities of native, proteinase-complexed, and reactive site cleaved serpins: cornparison of q-proteinase inhibitor, q-antichyrnoîrypsin, antithrombin III, a,-antiplasrnin, angiotensinogen, and ovalbumin. Biochemishy 30, 1723- 1730.

McCormack, R T., Wang, T. J., Rittenhouse, H G., Wolfett, R L., Finlay, J. A., Lilja, H., Sokoloff, R L., and Oestterling, J. E. (1995). MoIecuIar forrns of prostate-specific antigen and the hurnan kallikrein gene farnily: a new era. Uvology 45, 729-744.

McDonnell, D. P., Scott, R A., Kerner, S. A, O'Malley, B. W., and Pike, J. W. (1989). Functional domains of the hurnan vibmin 4 receptor regulate osteocalch gene expression. Mol. Endocrinol. 3,635-644.

McDonnell, D. P., Clevenger, B., Dana, S., Santiso-Mere, D., Tzukerman, M. T., and Gleeson, M. A. G. (1993). The rnechanism of action of steroid hormones: a new twist to an old tale. J Clin Pharmacol. 33, 1 165-1 172.

McGee, R S. and Herr, J. C. (1988). Human seminal vesicle-specific antigen is a substrate for prostate-specific antigen (or P-30). Biol. Reprod. 39,499-510.

McGuire, W. L. and Clark, G. M. (1992). Prognostic factors and treatment decisions in axillary- node-negative breast cancer. New EngI, J Md 326, 1756- 1761.

Monne, M., Croce, C. M., Yu, H., and Diamandis, E. P. (1994). Molecular characterization of prostate-specific antigen messenger RNA expressed in breast tumors. Cancer Res. 54,6344-6347.

Myal, Y., Shiu, R P. C., Bhaumick, B., and Bah, M. (1984). Receptor bindhg and growth- promoting activity of insulin-like growth factors in human breast cancer cells (T47D) in culture. Cancer Res. 44, 5486-5490.

Mydlo, J. H., Michaeli, J., Heston, W. D. W., and Fair, W. R (1988). Expression of basic fibroblast growth factor mRNA in benign prostatic hyperplasia and prostatic carcinoma. Prostate 13,241-247.

Nandi, S., Guzman, R C., and Yang, J. (1995). Hormones and rnammary carcinogenesis in mice, rais, and hurnans: A unifying hypothesis. P m . N d . Acad Sci. USA 92,3650-3657.

Neurath, H. and Walsh, K A. (1976). Role of proteolytic enzymes in biological regdation. Proc. Natl. Acad Sci. USA 73, 3825-3832.

Neurath, H. (1986). The versatility of proteolytic enzymes. J Cell. Biochem. 32,35-49.

Nowels, K , Kent, E., Rinsho, K., and Oyasu, R (1988). Prostate specific antigen and acid phosphatase-reactive cells in cystitis cystica and glandularis. Arch. Puthol. Lab. Med 112, 734- 737.

Oesterling, J. E., Chan, D. W., Epstein, J. L, Kimball, A W., Bnizek, D. J., Rock, R C., Brendler, C. B., and Walsh, P. C. (1988). Prostate specific antigen in the preoperative and postoperative evaluation of localized prostatic cancer treated with radical prostatectomy. J Urol. 139,766-772.

Oesterling, J. E. (1991). Prostate specific antigen: a critical assessrnent of the most usefbl tumor marker for adenocarcinorna of the prostate. J Urology 145, 907-923.

Papsidero, L. D., Wang, M. C., Valenniela, L. A., Murphy, G. P., and Chu, T. M. (1980). A prostate antigen in sera of prostatic cancer patients. Cancer Res. 40, 2428-2432.

Papsidero, L. D., Kuriyarna, M., Wang, M C., Horoszewicz, Jey Leong, S. S., Valemela, L., Murphy, G. P., and Chu, T. M. (1981). Prostate antigen: a marker for hurnan prostate epithelial cells. 1 NatL Cmer 1. 66, 37-4 1.

Parkin, D. M., Muir, C. S., Whelan, S. L., Gao, Y. T., Ferlay, J., and Powell, J. (1992). Cancer Incidence in Five Continents. Lyon, France: IARC Scientific Publication.

Parkin, D. My Pisani, P., and Ferlay, J. (1993). Estirnates of the worldwide incidence of eighteen major cancers in 1985. Int. 1 Cancer 54, 594-606.

Peehl, D. M., Cohen, P., and Rosdeld, R G. (1995). The insulin-like growth factor systern in the prostate. World l Urol. 13, 306-3 1 1.

Perlmutter, D. H., Glover, G. I., Rivetna, M., Schasteen, C. S., and Fallon, R J. (1990). Identification of a serpin-enzyme complex receptor on human hepatoma cells and hurnan monocytes. Proc. Natl. Acad Sci. USA 87, 3753-3757.

Petkovich, M., Brand, B. J., Knist, A., and Chambon, P. (1987). A hurnan retinoic acid receptor which belongs to the family of nuclear receptors. Nature 330,44440.

Pollen, J. J. and Dreilinger, A. (1 984). Imrnunohistochernical identification of prostatic acid phosphatase and prostate specific antigen in fernale periwethra1 glands. Urology 23, 303-304.

Porter-Jordan, K. and Lippman, M. E. (1994). Overview of the biologic markers of breast cancer. HemafoZ. ~ d . Clira N 8,73- 100.

Reid, S. E., Murthy, M. S., Kaufinan, M., and Scanlon, E. F. (1996). Endocrine and paracrine hormones in the promotion, progression, and recurrence of breast cancer. Brjt. J S q 83, 1037- 1046.

Rennie, J. and Rusting, R (1 996). Making headway against cancer. Sci Am September, 56-59.

Reynolds, T. (1994). Breast cancer prognostic factors - the search goes on. J Natl. Chmer 1. 86, 480-483.

Riegman, P. H., Vlietstra, R J., Klaassen, P., van der Korput, J., Gewts van Kessel, A., Rornijn, J., and Traprnan, J. (1989a). ïke prostate-specific antigen gene and the human glandular kallikrein-1 gene are tmdernly located on chromosome 19. FMS Lett. 247, 123-126.

Riegrnan, P. H., Vlietstra, R J., van der Korput, J. A. G. M., Rornijn, J. C., and Traprnan, J. (1 Wb). Characterization of the prostate-specific antigen gene: a novel human ka1 likrein-Ii ke gene. Biochem. Bioph. Res. Co, 159,95- 102.

Riegrnan, P. H., Vlietstra, R J., van der Korput, J., Brinkrnann, A. O., and Trapman, J. (1991). The promoter of the prostate-specific antigen gene contains a fùnctional androgen responsive element. Mol. Endocrinol. 5, 192 1- 1930.

Rosedelci, R G., Lamson, G., Pham, H., Oh, Y., Conover, C., DeLeon, D. D., Donovan, S. M., Ocrant, I., and Giudice, L. C. (1990). Insulin-like growth factor bindimg proteins. Rec. Prog. Hormone Res. 46,99- 163.

Russo, J., Gusterson, B. A, Rogers, A. E., Russo, 1. H., Wellings, S. R , and van Zwieten, M. J. (1990). Biology of disease. Comparative study of hurnan and rat mammary hunorigenesis. Lab. Invest. 62,244-278.

Russo, J. and Ruso, 1. (1995). Hormonally induced differentiation: a novel approach to breast cancer prevention. .l Cell. Biochem 22, 58-64.

Schaller, J., Akiyama, K, Tsuda, R, Hara, M., Marti, T., and Rickli, E. E. (1987). Isolation, characterimtion and amho acid sequence of y-seminoprotein, a glycoprotein fiom human seminal plasma. Eur. J Biochem 170, 11 1-1 18.

Scheidereit, C., Geisse, S., Westphal, H. M., and Beato, M. (1983). The glucocorticoid receptor binds to defmed nucleotide sequences near the promoter of mouse rnarnrnary tumow virus. Nature 304,749-752.

Schonk, D., van Dijk, P., Riegman, P., Traprnan, J., Holm, C., Wilcocks, T. C., Siilekens, P., van Venrooij, W., Wimmer, E. M., Geurts van Kessel, A., Ropers, H. H., and Wieringa, B. (1990). Assignment of seven genes to distinct intervals on the rnidportion of human chromosome 19q surrounding the motonic dystrophy region. Cytogenet. Cell Genet. 54, 15- 19.

Schuur, B. R, Henderson, G. A., Kmetec, L. A., Miller, J. D., Lamparski, H. G., and Henderson, D. R (1996). Prostate-specific antigen expression is regulated by an upstream enhancer. J Bol. Chem. 271, 7043-705 1.

Schuurmans, A. L. G,, Bolt, J., and Mulder, E. (1989). Androgen receptor-mediated growth and epidermal growth factor receptor induction in hurnan prostate ce11 iine LNCaP. Urol. ]nt. 44,7 1- 76.

Sensabaugh, G. F. (1978). Isolation and characterization of a semen-specific protein fiom hurnan seminal plasma: a potentiai new marker for semen identification. J Forensic Sci. 23, 106-1 15.

Sherwood, E., Kozlowski, J., and Lee, C. (1990). Effect of androgens and prostatic stroma on growth and PSA/PAP secretion by the human prostate carcinoma line LNCaP. J Uvology 143, 203A.

Sinha, A. A., Wilson, M. J., and Gleason, D. F. (1987). Irnrnunoelectron microscopie localization of prostatic-specific antigen in human prostate by the protein A-goId complex. Cmer 60, 1288- 1293.

Slattery, M. L. and Kerba, R A. (1993). A comprehensive evaluation of family history and breast cancer risk, the Utah population database. J M 270, 1563-1 568.

Spom, M. B. and Roberts, A. B. (1983). Role of retinoids in differentiation and carcinogenesis. Cmer Res. 43,30343040.

Spom, M. B. and Roberts, A. B. (1985). Autocrine growth factors and cancer. Nature 313,745- 747.

Stamey, T. A, Yang, N., Hay, A. R, McNeal, J. E., Freiha, F. S., and Redwine, E. (1987). Prostate specific antigen is a s e m marker for adenocarcinorna of the prostate. New Engl. J Med 317,909-918.

Stenrnan, U., Leiionen, J., Alfian, H., Rannikko, S. Tuhkanen, K., and Afthan, 0. (1991). A complex between prostate-specific antigen and g -antichymotrypsin is the major form of prostate- specific antigen in senun of patients with prostatic cancer: assay of the complex improves sensitivity for cancer. Cancer Res. 51, 22-226.

Strahle, U., Boshart, M, Klock, G., Stewart, F., and Schüîq G. (1989). Glucocorticoid- and progesterone-specific effects are determined by diffaential expression of the respective hormone receptors. Nature 339,629-632.

Sutherland, D. J. suid Mobbs, B. G. (1992). Hormones and Cancer. In Basic Science of Oncoloay, Second Edition. (ed. Tannock, 1. F. and Hill, R P.). New York, IW McGraw Hill miblishing, pp. 207-232.

Tepper, S. L., fagirdar, J., Heath, D., and Geller, S. A. (1984). Homology between the female paraurethral (Skene's) glands and the prostate. Arch Pathol. Lab. Med. 108,423-425.

Testa, J. E. and Quigley, J. P. (1990). The role of urokuiase-type plasminogen activator in aggressive tumur-ce11 behaviour. Cancer Metast. Rev. 9,353-367.

Trichopoulos, D., Li, F. P., and Hunter, D. J. (1996). What causes cancer? Sci. Am. Sept., 80-87.

Truss, M. and Beato, M. (1993). Steroid hormone receptors: Interaction with deoxyribonucleic acid transcription factors. Endocr. Rev. 14,459-479.

Tsai, S. Y., Carlstedt-Duke, J., Weigel, N. L., Dahlman, K., Gustafsson, J. A, Tsai, M. J., and O'Malley, B. W. (1988). Molecular interactions of steroid hormone receptor with its enhancer element: evidence for receptor d i a formation. Ce11 55,361-369.

Urnesono, K, Giguere, V., Glass, C. K, Rosenfeld, M. G., and Evans, R M. (1986). Retinoic acid and thyroid hormone induce gene expression through a cornmon responsive element. Nature 336, 262-265.

VogeI, P. M, Georgiade, N. G., Fetter, B. F., Vogel, F. S., and McCarty, K S. (1981). The comlation of histologie changes in the hurnan breast with the menstrual cycle. Am. J Pathol. 104923-34.

Wang, M. C., Valenniela, L., Murphy, G. P., and Chu, T. M. (1977). Tissue specific and tumour specific antigens in human prostate. Fed Proc. 36, 1254.

Wang, M. C., Valemela, L. A., Murphy, G. P., and Chu, T. M. (1979). Purification of hurnan prostate-specific antigen. Imst. Urol. 17, 159-163.

Wang, M. C., Papsidero, L. D., Kuriyama, M., Valenarela, L. A., Murphy, G. P., and Chu, T. M. (1981). Prostate antigen: a new potential rnarker for prostatic cancer. Prostate 2, 89-96.

Wang, M. C., Kuriyarna, M., Papsidero, L. D., Loor, R M, Valemela, L. A., Murphy, G. P., and Chu, T. M. (1982). Prostate antigen of human cancer patients. In Methods in Cancer Research XX (ed. Busch, H. and Yeoman, L.). New York, NY: Academic Press, pp. 179- 197.

Wang, L. G., Liu, X M., Kreis, W., and Budman, D. R (1996). Involvement of prostate specific antigen in the stimulation of LNCaP ce11 growth. Oncol. Rep. 3, 91 1-917.

Wang, L. G., Liu, X M., Kreis, W., and Budman, D. R (1997). Down-regulation of prostate- specific antigen expression by finasteride through inhibition of complex formation between androgen receptor and steroid receptor-binding consensus in the promoter of the PSA gene in LNCaP cells. Cancer Res. 57, 714-719.

Watt, K W. K., Lee, P. J. L., M'Tirnkuly T., Chan, W. P., and Loor, R (1986). Human prostate- specific antigen: Structural and functional sirnilarity with serine proteases. Proc. Natl. Acad Sci. USA 83, 3 166-3 170.

Weber, J. P., Oesterling, J. E., Peters, C. A., Partin, k W., Chan, D. W., and Walsh, P. C. (1 989). The influence of reversible androgen deprivation on s e m prostate-specific antigen levels in men with benign prostatic hyperplasia. J. Urology 141,987-992.

Weber, B. L. and Garber, J. E. (1993). Family history and breast cancer, probabilities and possibilities. J M 270, 1602-1603.

Wilding, G., Valverius, E., Knabbe, C., and Gelrnann, E. P. (1989a). Role of transforming growth factor-a in human prostate cancer ce11 growth. Prostate 15, 1-12.

Wildiig, G., Zugmeier, G., Knabbe, C., Flanders, K., and Gelrnann, E. (1989b). Differential effects of transforming growth factor P on human prostate cancer mils in vitro. Mol. Cell. Endocrinol. 62, 79-87.

Wingo, P. A., Tong, T., and Bolden, S. (1995). Cancer statistics. CA-Cancer J. Clin 45, 8-30.

Wu, J. T., Zhang, P., Astill, M. E., Wilson, L. W., Lyom, B. W., Wu, L. L., and Stephenson, R (1995). PSA immunoreactivity detected in LNCaP ce11 medium, breast tumour cytosol, and fernale serum. 3: Clin Lab. Anal. 9,243-25 1.

Yamamoto, K. R (1985). Steroid receptor regdatecl transcription of specific genes and gene networks. Annu RD. Genet. 19,209-252.

Yoshida, E., Ohmura, S., Su&, M., Mamyarna, M, and Mihara, H. (1995). Prostate-specific antigen activates single-chain urokinase-type plasminogen activator. Int. J. Cancer 63,863-865.

Young, C. Y. F., Andrews, P. E., Montgomery, B. T., and Tindall, D. J. (1992). Tissue-specific and hormonal regdation of human prostate-specific glanduiar kallikrein. Biochemishy 31,818- 824.

Young, C . Y. F., Murtha, P. E., Andrews, P. E., L'mcky, J. IL, and Tindall, D. J. (1994). Antagonism of androgen action in prostate tumour celis by retinoic acid. Prostate 25, 39-45.

Yu, H., Diamandis, E. P., and Sutherland, D. J. A. (1994a). Immunoreactive prostate-specific antigen levels in fernale and male breast tumors and its association with steroid homone receptors and patient age. Clin. Biochem. 27, 75-79.

Yu, H., Diamandis, E. P., Zarghami, N., and Grass, L. (1994b). Induction of prostate specific antigen production by steroids and tarnoxifen in breast cancer ce11 lines. Breast Cancer Res. Tr. 32, 291-300.

Yu, H. and Diamandis, E. P. (1995a). Prostate specific antigen imrnunoreactivity in amniotic fluid. Clin Chem. 41, 204-210.

Yu, H. and Diarnandis, E. P. (1995b). Prostate specific antigen in milk of lactating women. Clin. Chem. 41, 54-58.

Yu, H., Giai, M., Diamandis, E. P., Katsaros, D., Sutherland, D. J. A., Levesque, M. A., Roagna, R, Ponzone, R , and Sismondi, P. (1995). Prostate-specific antigen is a new favorable prognostic indicator for women with breast cancer. a m e r Res. 55,2104-21 IO.

Yu, H., Diamandis, E. P., Levesque, M., Giai, M, Roagna, R, Ponzone, R , Sismondi, P., Monne, M., and Croce, C. (1996a). Prostate specific antigen in breast cancer, benign breast disease and normal breast tissue. Bremt Crmcer Res. Tr. 40, 171-178.

Yu, H., Levesque, M. A., Khosravi, M. J., Papanastasiou-Diamandi, A., Clark, G. M., and Diamandis, E. P. (1996b). Association between insulin-like growth factors and their proteins and other prognostic indicators in breast cancer. Brit. J Cancer 74, 1242- 1247.

Zargharni, N., Grass, L., and Diarnandis, E. P. (1997). Steroid hormone regulation of prostate- specific antigen gene expression in breast cancer. Brit. J a m e r 75, 579-588.

Zhou, A., Tewari, P. C., Bluesteiq B. L., Caldwell, G. W., and Larsen, F. L. (1993). Multiple forms of prostate-specific antigen in Sem: differences in irnrnuno-recognition by monoclonal and polyclonal assays. Clin. Chem 39, 2483-2491.

PROSTATESPECIFIC ANTIGEN INDUCTION BY A STEROID HORMONE IN T47D CEWS GROWDJG IN SCID MICE

Ilana Kogan, James R Ballinger, Russell Redshaw, EleAherios P. Diamandis, Dimitrios N. Melegos, Robert M. Kuba, and A. Michael Rauth

Subrnitted for publication in: Breast Cancer Research and Treatment

50

2.1. SUMMARY

Previous studies revealed that prostate-specific antigen (PSA) is present in more than 30%

of human breast turnour cytosols. SuMval analysis showed that patients with PSA-producing

tumeurs have a reduced risk for relapse, suggesting PSA to be an independent favourable

prognostic rnarker for a large subset of breast cancer patients. The present investigation

establishes an in vivo model for the induction of PSA in human breast cancer tumours growing

as xenografts in severe combined irnmunodeficient (SCID) mice. î l e human mammary cancer

cell-line T47D was grown i.m. in female rnice. When the tumow and leg diameter reached 10

mm, the mice were stimulated daily with norgestrel for either 5 or 7 days to produce PSA, and

sacrificed on day 8. The prostate cancer cell-line LNCaP was gown in male mice and

functioned as a positive control for PSA production. M e r T47D and LNCaP mice were

sacrificed, a highly sensitive irnmunofluorometric assay was used to analyze the PSA

concentration in the turnour, muscle, liver, and kidney cytosols. Norgestrel-stimulated T47C

mice showed significantly greater PSA levels in the tumours compared to tumours of the control

rnice. However, PSA levels in tumours of the stimulated mice were significantly lower than

those in the LNCaP xenografts. No PSA levels above background were present in the blood anc

normal tissue of the norgestrel-stimulated and controI T47D xenogmfk. This mouse model will

be a valuable tool for investigating and screening new therapies for a subgroup of breast cancer

patients who have significant PSA concentrations in their tumours.

5 1

2.2. rnODU(7IlrON

Prostate-specific antigen (PSA) is a 33 kDa single chah glycoprotein (Oesterling, 1991).

The human gene encodiig PSA was cloned, sequenced, and localized to the long arm of

chromosome 19 (Riegman et al., 1989). PSA is a kallikrein-like serine protease with a

chymotrypsin-like activity. It is present in semen at concentrations of 0.5-3.0 g/L, and

participates in semen liquefaction after ejaculation (Malm and Lilja, 1995). Serum of normal

men contains PSA levels below 4 pgL. However, there are higher PSA leveIs in the serurn of

patients with benign diseases of the prostate and prostatic carcinomas. The on cent ration of PSA

in serurn is a valuable biological marker for screening, diagnmis, and monitoring patients with

prostate cancer (Oestaling, 1991).

Initially, it was believed that only epithelial cells of the prostate prduced PSA. Receni

studies, using ultrasensitive detection-techniques for PSA, have show that other hurnan tissua

and tumours can produce PSA at low levels. For exarnple, colon, ovarian, liver, kidney, and

adrenal tumours can produce PSA (Diamandis and Yu, 1994; Diamandis and Yu, 1995; Levesque

et al., 1995). PSA immunoreactivity was also examined in a cohort of more than 1,200 breasl

cancer patients, with a sensitive time-resolved irnmunofluorometnc assay (Diamandis et al., 1994;

Yu et al., 1994a). Results showed that immunoreactive PSA was present in more than 30% ol

the brmt tumour cytosols, at concentrations greater than 30 pglmg of total protein. Studies

confkmed that the immunoreactive PSA, detected in the breast turnow cytosols, was identical tc

the PSA produced by the prostate gland (Monne et al., 1994).

Breast cancer patients with PSA-positive tumours are a subset among the patients with

steroid-hormone receptor-positive twnours. A multivariate analysis showed that patients with

52

PSA-produchg tumours have a reduced risk for relapse. Thus, PSA is an independent favourable

prognostic marker for a large subset of breast cancer patients (Yu et al., 1995).

Breast cancer is a heterogeneous disease. Patients are usually divided into subgroups

based on diffaent prognostic and predictive markers. These markers are used to classi@ patients

into hi& or low risk groups when rnaking treatment decisions afler local surgery (McGuire and

Clark, 1992). Although a variety of markers are available, physicians cannot readily identiQ

patients who may benefit fiom adjuvant treatrnent. The identification process has been difficult

because the available markers are not specific enough (Yu et al., 1995). To better understand

the pathology of breast cancer and defme the therapeutic options for the different subgroups, it

is necessary to identify and characterize new prognostic marken such as PSA.

Accordhg to Young and colleagues (1992), androgenic steroids bind to androgen

receptors and regulate transcription of the PSA gene. However, steroid hormones and their

receptors share extensive structural sirnilarities. The progesterone, androgen, glucocorticoid, and

mineraloçorticoid receptors recognize highly homologous hormone response elements (HIE) on

DNA, and could al1 regulate genes whose promoters contain such response elements (Beato,

1989). Clinical data suggested PSA to be upregulated in some tumom and not others (Yu et al.,

1994a). To investigate the mechanism of PSA gene regulation in the breast, Yu and CO-workers

(1994b) developed an in vitro system that reproduced this PSA production by breast tumour cells.

'Ihey found that steroid hormones such as progestins can induce steroid-hormone receptor-

positive breast cancer cell-lines to produce PSA in vitro. Furthemore, estrogens failed to induce

such stimulation and could block PSA induction by androgens in T47D cells. Norgestrel was

shown to be a strong PSA-induchg progestin.

53

The present investigation establishes an in vivo rnodel for PSA gene induction as an

extension of the in vitro rnodel of Yu and colleagues (1994b). The human breast cancer dl-line

T47D and the prostate cancer cell-line LNCaP were grown in SCID mice, and the ability of

norgestrel to induce PSA production in the breast tumours was investigated. Our results show

that it is possible to induce PSA production in vivo using the T47D SCID rnouse xenograft

system.

2.3. MATERLQLS AND METHODS

2.3.1. Cell Lines

The steroid-hormone receptor-positive breast cancer cell-line T47D and the prostate cancer

cell-line LNCaP were obtained fiom ATCC (Rockville, MD). Cells were grown initially in fiasks

at 37°C and 5% CO2 in Alpha medium supplemented with antibiotics and 10% fetal calf serurn

(growth medium). Typical ce11 doubling tirnes were 20-30 hrs. The cells were then trypsinized,

counted with a particle counter and haemocytometer (to assess clurnping), washed, and

resuspendd in growth medium for injection into anirnals.

2.3.2. Animals

Fernale and male SCID mice with a Balbic genetic background (8-10 weeks old) were

obtained from the breeding colony of Ontario Cancer Institute (Toronto, Canada). Animals were

kept in micro-isolator cages that were changed twice weekly. Mice received a supply of sterile

water and y-irradiated rodent food (Teklad, WI; 5% rodent diet #7012) ad libitum. To prevent

infections, the animals were handled under a iaminar flow hood.

54

2.3.3, Generation of SCID Mice Xenografts and PSA Induction

In vivo growth was initiated by i.m. injection of cells into the left hind leg of rnice. T47D

cells (3x106) were injected into female mice (T47D mice), and 5x106 LNCaP cells were injected

into male mice (LNCaP mice). Ail cells were injected in a volume of 0.05 ml growth medium.

To stimulate tumour growth, T47D mice were injected S.C. in the scapular region with 0.1 ml P estradiol 17-valerate (5 mglml in sesame oil; Sigma Chemical Co., St. Louis, MO) once every

2 weeks, beginning 1 week &er ce11 injection (Leung and Shiy 1981). When tumours grew to

a leg diameter (leg plus tumour) of 10 mm (4.5 g of tumour), fkstradiol 17-valaate injections

were stoppeù, and T47D rnice were randomly divided into 3 groups. One group was stimulated

daily with 0.05 ml of norgestrel(1 mghi in sesarne oil; Sigma Chemical Co., St. Louis, MO)

for 5 days. The second group was stimulated daily with 0.05 ml norgestrel(1 mgml in sesame

oil) for 7 days. The third group was stimulated daily witfi 0.05 ml of the vehicle (sesarne oil;

Sigma Chernical Co., St. Louis, MO) for 7 days. In al1 stimulation experiments, mice were

injected S.C. in the scapular region. Al1 rnice were sacrificed by cervical dislocation 8 days after

the fust norgestreVvehicle injection. Mice injected with the LNCaP dl-line received no M e r

treatment and were sacrificed when tumour and leg size reached 10 mm in diameter. The

following tissues were obtained fiom each mouse: tumour, muscle (fiom the opposite leg), liver,

and kidneys. Al1 tissues were immediately fiozen in liquid nitrogen and stored at -70°C until

tissue extraction.

55

2.3.4. Preparation of Tissue Extracts

Three samples (-10 mg each) were obtained fiom different regions of each fiozen tumour

or normal tissue. Each tissue sample was pulverized to a fine powder at -70°C. The tissues

were then lysed for 30 min on ice, using 0.5 ml of a lysis buffer [50 mM Tris (pH=8.0), 150 rnM

NaCl, 5 mM EDTA, 1% (wh) NP40 surfactant, and 1 rnM of PMSF]. Lysates were centrifuged

at 15,000 g at 4°C for 30 min, and the supematants were collected for PSA and total protein

analyses.

2.3.5. PSA Assay

2.3.5.1. Assay Description and Instrumentation

Atirne-resolved irnmunofluorometric assay was used for the quantification of PSA in each

sample (Figwe 2-1). The method is a sandwich-type irnrnunoassay which incorporates 2 mwine

monoclonal anti-PSA antibodies (Diagnostic Systerns Laboratones, Webster, TX). The first

antibody is useù for coating the plate and the second is a detection antibody that is conjugated

with biotin. Streptavidin conjugated with alkaline phosphatase (SA-ALP) is used as a Iabel to

bind to biotin. It is possible to detect the enzyrnatic activity of alkaline phosphatase (ALP)

through the hydrolysis of a substrate, 5-fluorosalicyl phosphate (FSAP). The dephosphorylated

form of FSAP (5-fluorosalicylic acid; FSA) fùrther reacts with Tb3+-EDTA chelates, fomiing a

ternary fluorescent complex. ïhe fluorescence of this complex is measured (61 5 nm) with a tirne

resolved fluororneter, the 615TM Immunolyzer (CyberFluor, Toronto, ON), following laser

excitation (337.1 nm).

FSAP FSA+T~*-EDTA -+ FS A - T ~ ~ E D T A + ResoIved

u Fluororneter

1. monoclonal anti-PSA antibody (1 st) attached to plate surface

2. monoclonal anti-PSA antibody (2nd) with a biotin molecule (filled circle)

3. SA-ALP= sterptavidin alkaline phosphatase complex

4. FSAP= 5'-fluorosalicyl phosphate

5. FSA-~~~+-EDTA = ternary cornplex of fluorosalicylic acid, terbium and EDTA

Figure 21. PSA assay, illustrahng steps involveci in the imrnunofluorometric analysis.

57

Time-resolved fluorometers have al1 the usual components of a conventional fluorometer

plus a system for time-gated measwements of ody a part of the total emission cycle. During

the rest of the emission cycle, the photomultiplier is inactive and unwanted events (eqg, short-

lived fluorescence) are undetected. Some main advantages of the fluorescent Tb3' chelates

include separation of excitation and ernission wavelengths (i.e., large Stoke shih), narrow

emission bands, and a long fluorescence Iifetime. This long lifetime makes Tb3' chelates

especially suitable as labels for pi time-resolved fluorescence irnmunoassay. By delaying

measurement of fluorescence after a flash excitation of the sample, background short-lived

fluorescence due to senim, solvents, cuvettes, and reagents is excluded. The time-gated settings

of this instrument are as follows: laser ernission wavelength 337.1 nm; laser pulse duration 3-4

ns; repetition rate 20 pulseds; delay tirne 200 p ; measurement time 400 ps; recovery tirne afier

completion of measurement to next flash 49.4 ms; cycle tirne 50 ms; measurement timdwell 1

s (20 flashes). The wavelength of the interference filter used in the ernission pathway is 615 nrn

(Diamandis, 1988; Christopoulos and Diamandis, 1992; Papanastasiou-Diamandi et al., 1992).

In this assay, exceptional sensitivity is achieved because of the enzyrnatic amplification

introduced by ALP and the quantification by the tirne-resolved fluorometer (Christopoulos and

Diamandis, 1992). The detection limit of the assay is 1-2 ng PSA/L. Using this assay, the PSA

content of al1 tissue extracts was measured in duplicate and comparecl to a standard curve, which

was run daily for each assay.

58

2.3.5.2. Reagents and Solutions

Reagents used in this assay, unless otherwise stated, were purchased from Sigma Chernical

Co., St. Louis, MO. The assay buffer was 6% (wlv) bovine serum albumin (BSA), 0.5 M KCl,

0.05% (w/v) N3Na, 5% (vlv) normal mouse serwn, 0.05% (wh) Triton X-100, and 50 rnM Tris

(pH 7.8). The wash solution was 0.15 M NaCl and 0.05% (wh) Tween 20 in 5 mM Tris (pH

7.8). The SA-ALP (Jackson Imrnunoresearch, West Grove, PA) diluent contained 6% (wlv) BSA

in 50 mM Tns (pII 7.8). ïhe substrate buffier was a 0.15 M NaCl, 1 mM MgCl,, 0.05% N3Na

and 0.1 M Tris (pH 9.1). ï he substrate stock solution was a 10 rnM FSAP (CyberFluor, Toronto,

ON) in 0.1 M NaOH. The development solution contained 0.4 M NaOH, 2 rnM TbCl,, 3 rnM

EDTA, and 1 M Tris (Ferguson et al., 1996).

2.3.5.3. Antibodies

The murine monoclonal anti-PSA antibodies were obtained h m Diagnostic Systems

Laboratories, Webster, TX Antibody DSL-O1 was present on pre-coated plates and DSL-11 was

obtained as 0.1% solution and used for detection. ïhe monoclonal DSLll was prepared for

biotinylation by overnight dialysis against 0.1 M sodium bicarbonate. An equal volume of 0.5

M carbonate bufEer (pH 9.1) was added to yield a 0.05% fmal protein solution. The N-

hydroxysuccinirnide ester of biotin (NHS-LC-Biotin) was solubilizd in DMSO (1 mg in 50 PL)

before incubation with the antibody (2 hrs, 25°C). One mg of NHS-LC-Biotin was used per mg

of antibody (Ferguson et al., 1996).

59

2.3.5.4. Calibrators (PSA Standards)

Purified seminal plasma PSA calibrators were prepared by diluting human serninal PSA

in 50 mM Tris (pH 7.8) containing 6% (wh) BSA. The concentrations of the preparations used

for the calibration were O, 5, 10, 25, 100, 500, 2000, and 10000 ngL (Ferguson et al., 1996).

2.3.5.5. Assay Procedures

Calibrators or turnour cytosol extracts (50 pL) were added in duplicate to coated

microtiter wells of 96-well plates. Into each of these wells, 50 pL of assay bufk was added,

containing diluted biotinylated monoclonal antibody (0.5 m@). Wells, containing assay buffer,

detection antibody, and either calibrator or sarnple, were incubated for 1 hr at 25°C with shaking.

At the end of the incubation p&od, the plates were washed 6 tirnes with wash solution by using

an automated rnicrotiter plate washer. Then, 100 PL of SA-ALP conjugated stock solution,

diluted 1 :20,000 with SA-ALP diluent, was added to each well. Mer an incubation period of

15 min at room temperature with shaking, the plates were washed 6 times with wash solution.

The diluted substrate (100 yUwell; 1 : 10 dilution of stock FSAP solution in substrate buffer

before use) was added to each well, and the plates were shaken for 10 min at r o m temperature.

At the end, 100 pL of developing solution was added to each well. After incubating the plate

for 1 min at room temperature, the Tb3+-specific fluorescence was read with CyberFluor 615

Immoanalyzer. Data reduction was done automatically by the machine (Ferguson et el., 1996).

Al1 samples were run in duplicate and intra-sample agreement was approximately 95%.

60

2.3.6. Total Protein Determination and PSA Irnmunoreactivity

The concentrations of total protein in the cytosolic extracts were detemiined by the

bicinchoninic acid (BCA) method, using a commercial kit (Pierce Chemical Co., Rockford, IL).

In this system, proteins react with Cu2' in an alkaline medium, yielding Cu'. Two molecules of

BCA - a highly sensitive detection ragent for Cu' - bind to 1 Cu', forrning a pwple reaction

product. ïhis product exhibits a strong absorbante at 562 nm, allowing the spectrophotometric

quantitation of proteins in aqueous solution. Total protein in each sample was measured in

duplicate and compared to a standard curve that was run daily for each assay. The relative

amount of immunoreactive PSA in each tissue was calculated by dividing each PSA value by the

total protein in each sample to yield pg PSAfmg protein (PSA level).

2.3.7. Imrnunohistochemistry

LNCaP turnours were fixed in formaldehyde, embedded in paraffin, sectioned and stained

for PSA using an indirect labelling method that detects antigen sites in the specimen. FUst,

tumour sections were incubated with a prirnary polyclonal rabbit anti-human PSA antibody

@ A m Diagnostic Canada Inc., Mississauga, ON) which was diluted 1:800. Then, a biotinylated

secondq antibody (Signet Laboratories, Inc., Dedham, MA) was diluted 15 and incubated with

the samples. The sections were then incubated with a peroxidase-conjugated streptavidii-biotin

complex (Signet Laboratories, Inc., Dedharn, MA), followed by the addition of a solution

containhg the chromogenic substrate 3,3'-diarninobenzidine tetrahydrochloride @AB; Sigma

Chemical Co., St. huis, MO) and H202. This produced a localized coloured precipitate at the

antigen sites. Haematoxylin was used as a counterstain to visuaiize al1 cells. The

61

irnmunohistochernical analysis was done at Mount Sinai Hospital, Department of Pathology,

courtesy of Dr. J. Mullen.

2.3.8. Progesterone- and Estrogen-Receptor Analyses

Concentrations of progesterone- and estrogen-receptors in the turnour samples (finoVmg

of total protein) were measured with the Abbott PgR/ER enzyme immunoassay kits (Abbott

Laboratories, Abbott Park, IL) at the Sunnybrook Health Science Centre, courtesy of Dr. D. J.

A. Sutherland. Briefly, beads coated with anti-PgEUER (rat, monoclonal) are incubated with

tissue cytosols or standards. During this incubation, PgR/'ER present in the sample or standards

binds to the anti-PgR/anti-ER on the beads. Unbound materials present in the sample a r e

removed by washing the beads. Then, a second anti-PgRIER (rat, monoclonal) conjugated with

horseradish peroxidase (anti-PgR~EEI-lRPO) is incubated with the beads, followed by binding

of anti-PgEVERHRPO to the PgRlER bound to the beads. Unbund conjugate is removed bq

washing the beads. The beads are next incubated with O-phenylenediamine (OPD) substrate

solution which contains H202. Reaction of OPD with HRPO yields a yellow-orange colow. Th€

intensity of the colour formed by the enzyme reaction is proportional to the concentration ol

PgR/ER in the sarnple. The enzyme reaction is stopped by the addition of 1 M su l f i c acid anc

the intensity of the colour developed is r a d using a spectrophotometer set at 492 nm.

62

2.3.9. Statistical Analysis

PSA levels in T47D mice did not have a normal distribution and the variances were

heterogeneous. Therefore, a correction for the heterogeneity and non-normality was carrieci out

by transforming the data fiom their original forrn to a logarithmic form. One way ANOVA was

used to analyze the difference in PSA level between the tuinours of norgestrel-stimulated and non

stimulated T47D rnice. The test was also used to compare the PSA levels in the different tissues

of T47D mice.

2.4. RESULTS

Turnours developed in the T47D and LNCaP mice 3-4 months de r ceIl inoculation.

Figure 2-2 shows the mean and SD for PSA levels in tumow dissected iiom control T47D

bearing mice, 5- or 7-day norgestrel-stirnulated T47D bearing rnice, and LNCaP bearing mice.

Very low PSA levels were present in the T47D bearing rnice that were injected with sesame ail

(control), with a mean and SD of 2.4k1.5 pg PSA/mg protein. This level was not significantly

different fiom background levels. In contrast, elevated PSA levels were found in 5- and 7-day

norgestrel-stimulated T47D mice, with the means and SDs of 47Ml and 14M123 pg PSA/mg

protein, respectively. Clearly, norgestrel induced PSA production in T47D mice, when

adrninistered for either 5 or 7 days. The mean and SD for PSA Ievels in LNCaP mice were

443W7lO pg PSNmg protein, a concentration that is -30-1 00 times greater than the PSA levels

in the stimulated T47D mice (Figure 2-2). Background PSA levels were detected in the muscle,

liver, and kidneys &om both control and norgestrel-stimulated T47D mice (Table 2-1). However,

the normal tissues fiom LNCaP mice had PSA levels that were -20% of those detected in the

Control 5 Days 7 Days LNCaP

Treatments

Figure 2-2. Means and SDs of PSA concentration (pg PSA/mg protein) in tumour fiom control(4 tumours), 5-day norgestrel-stimulated (4 tumours), and 7-day norgestrel-stimulated (5 tumours) T47D mice, and LNCaP mice (12 turnours).

Table 2-1. Means and SDs of PSA concentration @g/mg total protein) in tumours and tissues of control, 5-day norgestrel-stimula and 7-day norgestrel- stimuiated T47D rnice.

Tissue

Tumour

Muscle

Kidney

Liver

Control SDay Stimulation 7-Day Stimulation (n=4) (n4) ( n 3

2.4&1.50 47a1 14&123

65

tumours (Table 2-2).

When analyzing the turnow samples fiom the 4 different mouse groups, a large SD was

observed in al1 treatment groups (Figure 2-2). f i e large SD might be due to both intra- and

inter-tu11iow variability in PSA levels. ïhus, both sources of variability were examined in each

treatment group. A large variability in PSA level was observed arnong tumours fiom individual

T47D mice in the sme treatrnent group (Figure 2-3). The mean PSA level in the 5-day

norgestrel-stimulated mice ranged from 9.3 to 102 pg PSA/mg protein. In the 7-day norgestrel-

stirnulated mice, the PSA mean level ranged fiom 37 to 353 pg PSAlmg protein. Furthermore,

when analyzing different sarnples from a single twnow, we found intra-tumour heterogeneity in

the different samples, as seen by the large SD for the individual mice (Figure 2-3). Inter- and

intra-tumour heterogeneity both contribute to the large SD (Figure 2-2).

Turnom fiom different rnice inoculated with LNCaP cells also had variable levels of

PSA, ranging from -1600 to 9000 pg PSNmg protein. M e n analyzing 3 different samples fiom

the same tumour, intra-tumour variability in PSA levels was observed in these mice, as seen by

the large SD for most tumours (Figure 2-4). Furthermore, this intra-tumour heterogeneity was

observed in immunohistochernically stained sections fiom LNCaP mice (Figure 2-5). Thus, the

inter- and intra-tumeur heterogeneity seen in norgestrel-stimulated rnice was also observed in

prostate tumours.

Previous data has suggested that PSA presence in tumours is associated with the presence

of progesterone receptors (Yu et al., 1994a). Figure 2-6 shows the progesterone receptor levels

in T47D mice Tom the different treatment groups. The means and SDs of progesterom receptor

concentrations in the control and 5-day stirnulated mice were 61&80 and 60N101 hoVmg total

Table 22. Means and SDs of PSA cuncentration (pg/mg total proteii) in turnours and tissues of LNCaP rnice.

II Tissue Qpe 1 LNCaP (n=12) 11 m

II Muscle 1 387k280 11 Kidney 722337

Liver 53643 12

Control 5 Days

Treatments

7 Days

Figure 2-3. Intra- and inter-tumour variability in [PSA] in T47D mice. Each bar represents a single mouse. The mean and SD for [PSA] for each tumour were obtain by analyzing 3 different sarnples from a single tumour.

LNCaP Mice

Figure 2-4. Intra- and inter-turnour variability in [PSA] in LNCaP mice. Each bar represents a single tumour. The mean and SD for [PSA] for each tumour were obtainel by analyzing 3 different samples fiom a single tumour.

Figure 2-5. Irnrnunohistochernical staining of tumour sections fiom LNCaP mice using polyclonal anti-PSA antibody. ïhe brown areas represent cluters of cells that are positive f PSA while the lighter areas represent cells that do not contain PSA. The rnagnification is 25C

Control 5 Days 7 Days

Figure 2-6. The level of progesterone receptors in tumours of control, 5-day norgestrel-stimulated, and 7-day norgestrel-stimulated T47D rnice. The number of tumours assayed is given by n.

7 1

protein, respectively. However, the 7-&y norgestrel-stimulated T47D mice had lower

progesterone-receptor levels with the mean and SD of 265&8 fmol/mg total protein.

Progesterone receptor levels in LNCaP mice were not detectable (data not shown). The levels

of estrogen receptors in mice fiom different T47D treatrnent groups are show in Figure 2-7.

Control T47D mice, 5-day stimulated T47D mice, and 7-day stimulated T47D rnice had low

estrogen receptor concentrations with the means and SDs of 1 W, 116 , and 1% fmoVmg total

protein, respectively. Estrogen receptor levels in LNCaP mice were not detectable (data not

shown). Table 2-3 surnmarizes the PSA concentrations as well as the progesterone and estrogen

receptor levels in the individual T47D mouse tumours studied. There was no obvious correlation

between receptor status and PSA production perhaps due to the small nurnber of animals used

and the low range of receptor levels.

- Control 5 Days 7 Days

Figure 2-7. The level of estrogen receptors in tumours of control, 5-day norgestrel. stimulated, and 7-day norgestrel-stimulated T47D mice. The number of tumours assayed is given by n.

Table 2-3. Levels of prostate-specific antigen (PSA), progesterone receptor (PgR), and estrogen receptor (ER) in turnours fiom control, 5-day norgestel-stimulated, and 7-&y norgestel-stimulated T47D mice.

11 Mouse Nurnber and Type

11 483; T47D Control

11 486; T47D Contrd

tPSAhotal protein (pglmg). 'PgRhotal protein (finoYmg). QWtotal protein (IÏnoVmg). p on cent rations of 0-4 £inoumg are negaiive, 5-14 £inoumg are equivocal, and 15 houmg and higher are positive.

1 488; T47D Contrd

2.5. DISCUSSION

In vitro experiments showed that steroid hormones, including progestins and androgens,

can induce the hurnan breast cancer cell-line T47D to produce PSA (Yu et al., 1994b). In the

present investigation, an in vivo mode1 for breast cancer PSA induction was estabiished using

SCID mouse xenografts. Elevated PSA levels were observed in T47D tumours in rnice that were

stimulated with norgestrel for either 5 or 7 days. However, T47D tumours in control mice that

were injected with the vehicle only, showed low PSA concentrations that were not significantly

different fiom background levels (Figure 2-2). Tmours fiom the stimulated T47D mice had a

statistically significant difference in PSA levels compared to the tumom from the control mice

@<0.00 1). Furthermore, PSA production in the norgestrel-stimulated T47D rnice was

tumour-specific, since elevated PSA levels were only found in the tumours. In contrast,

background PSA levels were present in the muscle, liver, and kidneys of these T47D mice as

well as in al1 tissues of control mice (data not shown).

Tumours of mice, which were injected with the prostate cancer cell-line LNCaP, had PSA

concentrations that were about 30-100 times greater than the PSA concentrations in the

norgestrel-stirnulated T47D turnours. These results parallel observations of other investigators,

who showed that the PSA concentrations in most hurnan breast himours are much lower than the

PSA concentrations found in hurnan prostate cancer (Oesterling, 1991; Yu et al., 1994a). The

normal tissue fiom LNCaP mice had PSA levels that were 10-20% of those detected in the

tumours (Table 2-2). It appears that high PSA concentrations in LNCaP tumours caused elevated

levels in the blood, contributing to elevated PSA levels in the normal tissues.

75

Yu and colleagues (1994a) analyzed 1275 breast cancer cytosolic extracts and observed

a large variability in the PSA concentration in the tumours. The PSA levels in the'r study varîed

fiom 10 to 100,000 pg PSAhg protein and the PSA levels in the population were not normally

distributed. In the present study, a large inter-tumour variability in PSA concentration was found

in the norgestrel-stimulated T47D mice. ïhe PSA concentrations in these rnice ranged fiom 9

to 353 pg PSAhg protein (Figure 2-3). Moreover, the concentrations of PSA in these rnice were

also not normally distributed. Therefore, there are similatities in PSA levels in the 2 populations,

suggesting that SCID mouse xenografts may be usefiil in investigating PSA production by breast

When analyzing different cross-sections fiom a single tumour, intra-tumour variability in

PSA concentration was found in both T47D and LNCaP hunours, as observed by the large SD

(Figures 2-3 and 2-4), and by the irnmunohistochernical staining of LNCaP tumours (Figure 2-5).

Zarghami and Diamandis (1996) found that the PSA irnrnunoreactivity in human breast tumows

is focal and restricted to clusters of cells. They explained these results by suggesting that not

al1 tumour cells in PSA-positive tumours produce PSA, resulting in an uneven distribution of

PSA in the turnours. However, it is also possible that distinct turnour sarnples fiom the LNCaP

and norgestrel-stimulated T47D mice have different ratios of PSA-producing tumour cells to

normal cells (e.g., stroma1 cells and iymphocytes) which do not produce PSA. The distribution

of PSA in LNCaP twnours was also focaI and restricted to clusters of cells, prîm~ily in the

cytoplasm of these cells (Figure 2-5). Therefore, one could expect different samples fiom the

same turnour to display different PSA levels, resulting in turnour heterogeneity.

76

The heterogeneity in PSA concentration in breast tumom c m be attributed to factors such

as androgen levels (Gleave et al., 1992) and growth factors (Cohen et al., 1992; Henttu and

V i o , 1993; Kanety et al., 1993; Culig et al., 1994), which affect PSA synthesis and expression.

However, the SCID mouse population was "homogeneous" due to homozygosity and the identical

nurnber of turnour cells that were inoculated. Thus, one would not expect varying androgen and

growth factor levels in the different mice. The cause of the heterogeneity in PSA expression in

the different T47D tumours remains unexplained.

Yu and colleagues (1994a) reportai a close association between PSA presence in breast

turnow and the presence of both progesterone- and estrogen-receptors. There was a stronger

association between PSA and progesterone receptors. Therefore, in the current study, the

relationship between PSA, progesterone-, and estrogen-receptors was investigated. Limited data

showed some qualitative association between PSA levels and progesterone receptor levels in

control, 5-&y norgestrel-stimulated, and 7-day norgestrel-stimulated T47D mice. However, the

narrow range of receptor levels measured and low number of turnours available prduded

quantitative conclusions (Table 2-3). The high progesterone receptor levels relative to estrogen

receptors suggests that the former rnay be more important than the latter in affecthg PSA

synthesis.

It is well known that progestins cm down-regulate progesterone receptors (Alexander et

al., 1989; May et al., 1989). In the present study, the levels of progesterone receptors in tumours

kom 7-day stimulated T47D mice were approximately 2.5 times lower than those found in both

the control and 5-day stimulated T47D mice (Figure 2-5). nese results show that continuous

stimulation with norgestrel (i.e., 7 days) causes down-regulation of the progesterone receptor,

77

while shorter stimulation (ie., 5 days) followed by 2 days without norgestrel stimulation, does

not affect the receptor levels. The levels of estrogen receptors were not affected by the norgestrel

stimulation (Figure 2-6). Al1 LNCaP tumours showed undetectable levels of both estrogen and

progesterone receptors (data not show).

Using an ultrasensitive time-resolved irnrnunofluorometric assay, Diamandis and

colleagues (1994) found that more than 30% of breast tumour cytosols contain immunoreactive

PSA. Their results suggested PSA to be an independent favourable prognostic rnarker for some

breast cancer subgroups (Yu et al., 1995). The goai of clinical prognostic markers is to provide

ht?onnation that will allow physicians to choose a specific and beneficial therapy for each patient

(Porter-Jordan and Lippman, 1994). Currently, prevention of breast cancer is not feasible because

many associated factors are endogenous and thus dificult to manipulate. Therefore, the only way

to reduce mortality is through early diagnosis and admlliistration of effective treatrnent.

Due to the heterogeneity of breast cancer, physicians use prognostic markers to identifj

patients who are at high risk for disease recurrence (McGuire and Clark, 1992). Although a

variety of prognostic markers are available, these markers are neither very sensitive nor specific,

rnaking the classification of patients into subgroups inaccurate. Thus, it is quite dificult to

identiQ patients who may benefit ii-om adjuvant treatment, and selection of patients for

appropriate therapy remains difficult and confùsing (Gasparnini et al., 1993; Porter-Jordan and

Lippmn, 1994). nie present mouse xenografi mode1 may be a vaiuable tool for studying the

response of breast tumours to adjuvant therapy based on the PSA concentration. Furthemore,

this mode1 can contribute to the development and screening of noveI therapies for a large

subgroup of breast cancer patients.

2.6. REFERENCES

Alexander, 1. E., Clarke, C. L., Shine, J. and Sutherland, R L. (1989). Progestin inhibition of progesterone receptor gene expression in human breast cancer cells. Mol. EEndocriml. 3, 1377- 1386.

Beato, M. (1989). Gene regulation by steroid hormones. Ce11 56, 335-344.

Christopoulos, T. K. and Diarnandis, E.P. (1992). Enzymatically amplifid tirne-resolved fluorescence imrnunoassay with terbium chelates. Anal. C h . 64, 342-346.

Cohen, P., Graves, H. C. B., Peehl, D. M., Kamarei, M., Giuduce, L. C. and Rosenfeld, R G. (1 992). Prostate-specific antigen (PSA) is an insulin-like growth factor binding protein-3 protease found in seminal plasma. J Clin Endocr. Metab. 75, 1046-1053.

Culig, Z., Hobisch, A., Cronauer, M. V., Radmaya, C., Traprnan, J., Hittmair, A., Bartsch, G. and Klocker, H. (1 994). Androgen receptor activation in prostatic tumour ce11 lies by insulin-like growth factor-1, keratinocyte growth factor, and epidennal growth factor. Chmer Res. 54,5474- 5478.

Diamandis, E. P. (1988). Immunoassays with timeresolved fluorescence spectroscopy: principles and applications. Ch. Biochem. 21, 139- 150.

Diamandis, E. P. and Yu, H. (1994). New biological fùnctions of prostate specific antigen? J Clin Endocr. Metub. 84 1515-1517.

Diarnandis, E. P., Yu, H. and Sutherland, D. J. A. (1994). Detection of prostate specific antigen irnmunoreactivity in breast tumours. Bremt Cancer Res. Tr. 32, 301-310.

Diarnandis, E. P. and Yu, H. (1995). Prostate specific antigen and lack of specificity for prostate cells. h e t 345, 1 186.

Ferguson, R A, Yu, H., Kalyvas, M., Zarnmit, S. and Diamandis, E. P. (1996). Ultrasensitive detection of prostate-specific antigen by a time-resolved Immunofluorometric csay and the Irnmulite@ immunachemiluMnescent third-generation assay: potential applications in prostate and breast cancers. C h Chem. 42, 675-684.

Gaspamini, G., Pozza, F. and Himis, A. L. (1993). Evaluating the potential usefulness of new prognostic and predictive indicators in node-negative breast cancer patients. J Natl. Cancer 1. 85, 1206-1219.

Gleave, E., Hsieh, J.-T., Wu, H.-C., von Eschenbach, A. C. and Chung, L, W K. (1992). S e m prostate specific antigen levels in rnice bearing human prostate LNCaP tumours are determined by tumour volume and endocrine growth factors. h c e r Res. 52, 1598-1605.

Henttu, P. and Vihko, P. (1993). Growth factor regulation of gene expression in the hurnan prostatic carcinoma ce11 line LNCaP. Cancer Res. 53, 1051-1058.

Kanety, H., Macljar, Y., Dagan, Y., Levi, J., Papa, M. Z., Pariente, C., Goldwasser, B. and Karasik, A. (1993). S e m insulin-like growth factor-bindiig protein-2 (IGFBP-2) is increased and IGFBP-3 is increased in patients with prostate cancer: correlation with serurn prostate- specific antigen. J Clin Endocr. Metab. 77, 229-233.

Leung, C. K H. and Shiu, R P. C. (1981). Required presence of both estrogen and pituitary factors for the growth of human breast cancer cells in athymic nude rnice. Cancer Res. 41, 546- 551.

Levesque, M., Yu, H., D'Costa, M. and Diamandis, E. P. (1995). Prostate specific antigen expression by various turnours. J Clin Lab. Anal. 9, 123-128.

Ivlalm, J. and Lilja, H. (1995). Biochemistry of prostate specific antigen, PSA. Scand J Clin. Lab. Inv. 55 (Suppl. 22f), 15-22.

May, F. E. B., Johnson, M. D., Wiseman, L. R, Wakeling, A. E., Kastner, P. and Westley, B. R (1989). Regulation of progesterone receptor gene expression in hurnan breast cancer cells. J Steroid Biochem. 33, 1035- 104 1.

McGuire, W. L. and Clark, G. M. (1992). Prognostic factors and treatment decisions in axillary- node-negative breast cancer. New Engl. J Med. 326, 1756-1 761.

Monne, M., Croce, C. M., Yu, H. and Diamandis, E. P. (1994). Molecular characterization of prostate-specific antigen messenger RNA expressed in breast turnours. Cancer Res. 54, 6344- 6347.

Oesterling J. E. (1991). Prostate specific antigen: a critical assessrnent ofthe most useful tumour mker for adenocarcinorna of the prostate. J Urology 145, 907-923.

Papanastasiou-Diamandi, A., Christopoulos, T. K. and Diamandis, E. P. (1992). Ultrasensitive thyrotropin imunoassay based on enzymatically arnplified time-resolved fluorescence with terbium cheIate, C h Chem. 38, 545-548.

Porter-Jordan, K, Lipprnan, M. E. (1994). ûverview of the biologic markers of breast cancer. Hemat. Oncol. C h N 8,73-100.

Riegman, P. H. J., Vlietstrri, R J., Klaassen, P., van der Korput, J. A. G. M., Romijn, J. C. and Traprnan, J. (1989). The prostate-specific antigen gene and the human glandular kallikrein-1 gene are tandernly located on chromosome 19. FEBS Lett. 247, 123-126.

Young, C. Y. F., Andrews, P. E., Montgomery, B. T. and Tindall, D. J. (1992). Tissue-specific and hornml regulation of hurnan prostate-specific glandular kaIIikrein. Biochernistry 31, 81 8- 824.

Yu, H., Diarnandis, E. P. and Sutherland, D. J. A. (1994a). Immunoreactive prostate-specific antigen levels in fernale and rnaie breast turnours and its association with steroid hormone receptors and patient age. Clin Biochem. 27,75-79.

Yu, H., Diamandis, E. P., Zarghami, N. and Grass, L. (1994b). Induction of prostate specific antigen production by steroids and tamoxifen in breast cancer ce11 lines. Breast Cancer Res. Tr. 32,291-300.

Yu, H., Giai, M, Diarnandis, E. P., Katsaros, D., Sutherland, D. J. A., Levesque, M. A., Roagna, R, Ponzone, R and Sismondi, P. (1995). Prostate-specific antigen is a new favourable prognostic indicator for women with breast cancer. Cancer Res. 55,2104-21 10.

Zarghami, N. and Diamandis, E. P. (1996). Detection of prostate-specific antigen rnRNA and protein in breast tumom. Clin Chem. 42,361-366.

SUMMARY AND CONCLUSIONS, PRELIMINARY RESULTS, AND FUTURE DIRECTIONS

82

3.1. SUMMATiY AM) CONCLUSIONS

Chapter 2 reported the establishment of an in vivo mode1 for PSA production by the

hurnan breast cancer cell-line T47D, growing as a xenograil in SCID rnice. The prostate cancer

ce11 line LNCaP was injected into SCID mice and fùnctioned as a positive control for PSA

production. Results showed that both T47D and L N 0 cells can grow in SCID mice and

develop tumom within 3 4 months. Stimulation of the T47D rnice with the norgestrel for 5 or

7 days resulted in PSA production by the hunow. Furthmore, PSA production was tumour-

specific since elevated PSA levels were only detected in the tumouts of the norgestrel-stirnulated

T47D mice. PSA concentrations were at background in the normal tissues of the stirnulated

T47D rnice or in the turnours of T47D control mice, which were injected with the vehicle. PSA

production by the T47D xenografis paralleled the in vitro data reported by Yu and colleagues

(1994a), who showed that norgestrel c m induce PSA production in steroid-honnone receptor-

positive breast cancer cell-lines. Turnours of mice bearing the LNCaP cells produced high PSA

levels without stimulation.

Turnours fiom T47D and LNCaP xenografb exhibited both inter- and intra-tumou

heterogeneity in PSA concentrations, poulting to the complexity of PSA gene regdation. Both

the PSA levels detected in the T4ïD xenografts and the heterogeneity in PSA concentrations were

comparable to those detected in tumeurs of breast cancer patients (Yu et al., 1994b; Zargharm

and Diamandis, 1996).

Steroid hormones and their receptors were shown to be involved in PSA production

(Young et al., 1992; Yu et al., 1994a; Zarghami et al., 1997). To investigate whether there is

an association between PSA production and receptor levels in the mouse modeI, the progesterone-

83

and estrogen-receptor concentrations were assayed in the tumom of the control, Sday, and 7-day

norgestrel-stimulated T47D mice. Lirnited data showed some qualitative association between

progesterone receptor levels and PSA levels. However, there was no association between the

levels of estrogen receptors and PSA production. These data are consistent with previous in vitro

reports (Yu et al., 1994a; Yu et al., 1994b; Zarghami et al., 1997), indicating that progesterone-

receptors rnight play a role in PSA synthesis.

Based on these similarities, the present in vivo mouse model for PSA production by breast

turnours will be a usefiil system for investigating the possible role(s) of PSA in breast cancer.

Moreover, this model could aid in understandimg why patients with PSA-positive turnours have

a better prognosis than patients with PSA-negative tumours. In the next section, prelirninary

results pertaining to various aspects of the present breast cancer model are presented. Future

experiments to examuie the model are discussed at the end of this chapter.

3.2. PRELIMINARY RIESULTS

3.2.1. Stimulation of T47D Cells W h the Progestin Norgestrel

Yu et al. (1994a) found that upon stimulation with steroids, the steroid-hormone receptor-

positive breast cancer cell-lines T47D and MCF7 produced PSA, which was detected both in the

supernatant and in the ce11 lysates. Neither cell-line produced rneasurable arnounts of PSA when

grown in culture. These experiments were done with cells approaching confluence. It was of

interest whether norgestrel could also stimulate exponentially growing cells to produce PSA, and

to what degree norgestrel rnight &eçt ceIl growth. It was hypothesized that as the cells divide

in the presence of norgestrel, PSA would accumulate in the media proportional to ce11 nurnber,

84

According to this hypothesis, the rate of PSA production should be exponential and relate to the

rate of ce11 doubling time. However, this mode1 relies on the following assumptions: PSA

induction is rapid, PSA production occurs throughout the ce11 growth phase, and neither

norgestrel nor PSA affects ce11 growth.

To test this hypothesis, 5x104 T47D cells were seeded in each well of a 24 multi-well

plate (-5- 10% confluent). The next day, "day O," medium (see Chapter 2) fiom each well was

removed and fiesh medium with norgestrel at a fmal concentration of either 1 or 10 pM was

added. From "day 1" to "day 1 lu, supmatants were removed fiom 2 wells each day for PSA

analyses (see Chapter 2). Then, fiesh media with no norgestrel was added to these wells to allow

cells to grow M e r until the end of the experirnent ("day 11"). Due to the different stimulation

periods of cells with norgestrel, ie. , fiom 1 to 11 days, cells in the different wells were exposed

to different PSA concentrations. Therefore, it was possible to determine whether different PSA

levels or times of norgestrel exposure affect the rate of ce11 proliferation. The plates were staùled

with trypan blue on "day 1 1", to assess the fmal state of ce11 growth in the different wells.

PSA levels in the medium, afler stimulation with either 1 or 10 pM norgestrel, are shown

as a function of tirne in Figure 3-1. As seen in the figure, there are 3 phases for the appearance

of PSA in the medium. The first phase, fiom "day O" to "day 3", is the initial induction period

which has been previously shown to take approximately 2 days in vitro (Yu et al., 1994a).

Detectable levels of PSA were fmt seen on "day 2" and the stabilization of PSA production

ocçurred by "day 3 " when ce11 growth was well established. The second phase, from "day 3" to

"day 7", showed an exponential increase in PSA levels with a doubling time of about 24 hrs.

This increase reflects the reported 24-30 hrs doubling time of T47D cells (Sutherland et al.,

O 2 4 6 8 ?O

Days After Norgestrel Stimulation

Figure 3-1. PSA production by T47D cells post norgestrel stimulation. Cells were exposed to either 1 or 10 pM norgestrel on "day O". Samples from the culture medium were removed from "day 1 " to "day 1 1" for PSA analysis.

86

1988) the accumulation of PSA in the medium with time.

During the final phase, fiom "day 7" to "day 1 l", PSA levels increased more slowly with

a doubling tirne of about 48 hrs. This reduced rate of PSA production reflects the decrease in

the rate of ceII proliferation fi-om "day 8" to "day 1 l", the confluency stage of the ce11 growtl-

curve. Confluency on "day 8" was visible in al1 wells, regardless of the stimulation period witk

norgestrel. Due to ce11 clurnping and loss on trypsinization, it was difficult to count cells direct11

wiîh a Coulter Counter. Staining on "day 11" showed that the degree of confluency reached bq

al1 cultures was the same. ïhus, the hypothesis was ~onfirmed, showing that during exponential

ce11 growth, fiom "&y 3" to "day 7", PSA increased exponentially, consistent with continuour

PSA production per ce11 at a constant rate. Attempts to model the behaviour quantitative11

showed that the theoretical results were vay dependent on the time and initial level of PSP

chosen. However, the data could be fit well with ce11 doubling times of about 24 hrs (data no1

show).

PSA presence in breast tumours is a good prognostic rnarker for some breast cancei

subgroups (Yu et al., 1995). One possibility is that the role of PSA in breast cancer is to inhibi

ce11 growth, thus, causing the tumours to be less aggressive. In these experiments, however, PSP

production after norgestrel stimulation did not inhibit ce11 growth. Therefore, these data do no

support a model in which PSA producing cells grow more slowly, thus, causing patients witl

PSA-positive turnours to have a better prognosis. Nevertheless, these studies show, that PSP

production upon stimulation with steroids is not confmed to confluent cells gowing in vitro, bu

that exponentially growing cells c m also produce PSA. A more sensitive assay for studying thi

effect of PSA on ce11 growth, as discussed in the next section, was carried out to M e r test thi:

conclusion.

3.2.2, Eff'ect of PSA on the Growth Rate of T47D and ET474 Cells

To more diiectly investigate the effect of PSA on the ce11 growth of breast cancer ce11

lines, such as T47D and BT474, the commercially availabie M I T assay (CeIlTiter 96TM Non-

Radioactive Proliferation Assay; Fisher Scientific, Nepean, ON) was used. This method is basec

on the cellular conversion of a tetrazolium salt 04Tï) into a formazan product by viable ce115

(Figure 3-2). The formazan product can be easily detected using an enzyme-linkec

immunosorbent assay (ELISA) plate-reader.

Two human breast cancer ce11 lines, T47D and BT474, were used for these experiments

The BT474 ce11 line is a staoid-hormone receptor-positive ce11 line that was shown to produa

five-times higher PSA levels in vitro after stimulation with staoids, as compared to T47D celh

@lamandis et al., in preparation). In the present expaiments, 2x104 cel!s/well (either T47D oi

BT474) were seeded in 96-well plates. Then, diluted PSA (purifieci from human seminal plasma

a gifi fiom Dr. Diamandis) was added to each well to yietd final PSA concentrations of 0.001

0.005, 0.01, or 0.1pgL. PSA concentrations were chosen based on the in vitro experiments O:

Lai and colleagues (1996), who incubated MCF7 cells with PSA and showed that lovi

concentrations of PSA (0.001-O.lp&) caused small but signifiant inhibition of ce11 growth

Ce11 numbers were estimated by lysing the cells and counting their nuclei using a Coultei

Counter. They found that the degree of inhibition decreased with increasing concentrations O:

PSA. In the present study, d e r PSA addition, the T47D/BT474 cells were incubated for 4 day:

at 37°C in a 5% CO, incubator. At the end of the incubation period, a dye solution containini

Figure 3-2. Chernical structure of MTï (colourless) which by reduction is convertecl to a forrnazan product (coloured).

89

the tetrazolium salt was added to each well and the plate was incubated for 4 hrs. Then, the

solubilization/stop solution was added to the wells and the absorbance was recorde. at 570 m.

The absorbance is directly proportional to the number of viable cells (manufacturer

specifications).

Results showed that PSA concentrations between 0.001 and 0.1pglL did not affect the

proliferation of T47D and BT474 cells (Figures 3-3 and 3-4). Although at the PSA concentration

of 0.001pgL there was a slight inhibition on the growth of both T47D and BT474 ceils, the

differences in absorbance were not statistically significant. In contrast to the results of Lai and

colleagues (1996), the present data show that PSA levels between 0.001 and O.lpg/L do not

affect the growth of eiîher T47D or BT474 ce11 lines. Possible explmations for this difference

include: differences in assays use4 ceIl lines, sample sizes, and statistical analyses.

3.2.3. Prostate-Specific Antigen Induction in SCID Mice Bearing BT474, M m , or BT20 Cells

To better understand PSA production by breast tumours, it is important to develop an in

vivo mouse model that produces PSA concentrations that are comparable to those observed in

breast cancer patients. The availability of different human breast cancer cell-lines that have

different characteristics, alIows the testing of these ce11 lines to find which can grow best in SCID

rnice and produce the highest PSA levels. As discussed in Chapter 2, it was faund that mice

injected with the T47D cell-line developed tumours within 3-4 months and could produce PSA

after stimulation with norgestrel. Two other steroid-hormone receptar-positive human breast

cancer cell-lines, MCF7 and BT474, were tested in the current work to develop an in vivo model

for PSA induction. The MCF7 cd-Iine was used extensively in the initial in vitro stimulation

PSA Concentration (pglL)

Figure 3-3. Effect of varying PSA concentrations on the proliferation of T47D cells, measured by the MTT assay, The number of wells (n) assessed at each concentration are shown. Error bars represent the standard deviation of the mean.

PSA Concentration (pglL)

Figure 3-4. Effect of varying PSA concentrations on the proliferation of BT474 ce1 measured by the MTT assay. The number of wells (n) assessed at each concentratio are shown. The error bars represent the standard deviation of the mean.

92

experirnents of Yu and colleagues (1994a). In vitro data showed that the BT474 cell-line

produced five-times higher b e l s of PSA upon stimulation with steroids than the T47D and

MCF7 ce11 lines (Diamandis et al., in preparation). Therefore, these cell-lines were tested in the

in vivo mouse model.

To test the MCF7 and BT474 tumow cell-lines, the sarne materials and methods were

applied as described in Chapter 2. One dificulty with developing the T47D mouse model was

the long tirne it took for tumours to grow in the SCID rnice (i.e., 3-4 months). By using the non-

obese diabetic (NOD/SCID) mouse strain, we attempted to speed up tumow growth. NOD/SCID

mice lack T and B cells, like SCID mice (Hendrickson, 1993). Furthmore, they have reduced

activity of natural-kilIer (K) cells, and have defects in complement activity and macrophage

bction (Shultz et al., 1995). ïhme mice were shown to be betta recipients for human bone

rnarrow transplants than SCID rnice (Larochelle et al., 1995) and therefore, served as subjects

in developing BT474 xenografts. However, our results showed that only about 20% of rnice

injected with the BT474 cells developed twnom, and the tumour growth in these mice required

the same time as the SCID mouse strain (data not shown). Therefore, these mice were not used

for M e r experiments and SCID rnice were used to establish the MCF7 and BT20 xenografts.

Steroid hormone receptors are important for PSA production by breast cancer celIs. In

vitro data showed that the steroid-hormone receptor-negative human breast cancer cell-line, BT20,

did not produce PSA upon stimulation with steroids (Yu et al., 1994a). We were interesteci to

conf i i that there would be no PSA induction in turnours of rnice bearing the BT20 cd-line, as

compared to their normal tissues. Thus, an in vivo model with mice bearing BT20 tumours was

also established. The relative concentrarions of PSA, progesterone-, and estrogen-receptors in

93

turnours and tissues Fom BT474, BT20, and MCF7 bearing mice were analyzed according to the

methods discussed in Chapter 2.

Tumours of NODlSCID mice bearing the BT474 cell-line produced tiigh levels of PSA

after stirndating the rnice with norgestrel for 7 days, with a mean of 303 pg PSNmg total protein

(Figure 3-5). ïhe PSA levels in the tumours of these mice were approximateIy three-times

higher than in the tumom of 7-day stimulated T47D rnice (Chapter 2). Production of higher

PSA levels by these tumours was expected since this ceIl-line produced five-times higher PSA

levels than the T47D ce11 line upon stimulation with norgestrel (Diamandis et al., in preparation).

Normal tissues, muscle, kidney, and liver were not signifimtly different fiom the tumours ol

the BT474 control rnice (stimulated with sesame oil). These background levels were

approximately ten-fold higher than the T47D controls (Chapter 2) and this difference is believed

to be due to technical problems with reagents used in this assay, contribuhg to highe~

background levels.

Unexpected results were obtained with mice injected with the MCF7 ce11 line. In v i h

data showed that this cell-line c m be stimulated with steroids to produce PSA (Yu et al., 1994a),

However, we did not detect elevated PSA concentrations in the tumours of MCF7 rnice

stimulated with norgestrel for 7 days or in normal tissues (Figure 3-6). Moreover, the MCFi

rnice thai were stimulated with the sesarne oil (negative control) had higher PSA levels than the

7-day stimulated mice, although the difference was not statistically significant. Previous in vivc

data with the MCF7 cell-line showed some PSA production upon induction with norgestrel

(Rauth and Ballinger, unpublished data). However, times of tumour growth were about 1-2

months, probably due to the higher number of cells injected (2x107 vs. 3x106 in the present

Tumour-Control Tumour Muscle

Treatments

Kidney Liver

Figure 3-5. Means and SDs of PSA concentration in tumours of control BT474 NOD/SCID rnice and in tumours and tissues of 7-day norgestrel-stimulated BT474 NOD/SCID rnice. The number of mice (n) per group are shown.

Tumour-Control Tumour Muscle

Treatments

Kidney

n=6

1

Liver

Figure 3-6. Means and SDs of PSA concentration in turnours of control MCF7 mice and in tumours and tissues of 7-day norgestrel-stirnulated MCF7 rnice. The number r: mice (n) in each group are shown.

96

experiments). More recent in vitro experiments with the MCF7 ce11 line showed inconsistency

in PSA induction (Diamandis et al., in preparation). Therefore, one possibility is that extensive

subculturing of these cells may alter some of their control mechanisms, such as those involved

in PSA synthesis and expression. For exarnple, it is possible that after certain nurnber of ceIl

divisions the progesterone-receptor levels are down-regulated, resulting in very low PSA

production.

Results showed that norgestrel did not induce PSA production in twmom of BT20 mice

as compared to the normal tissues. These observations were expected fiom the in vitro data.

Since the BT20 cell-line has very low levels of staoid-hormone receptors, it cannot be stirnulated

with steroids to produce PSA (Yu et al., 1994a). Figure 3-7 summarim the PSA ievels in

turnours of BT474, BT20, or MCF7 mice.

The estrogen- and progesterone-receptor levels in sorne twnours use. in these studies were

assayed - Figures 3-8 and 3-9, respectively - and compared to the receptor levels in T47D mice

(Chapter 2). Estrogen-receptor levels were similar in T47D, BT474, and BT20 mice. The

estrogen-receptor levels in MCF7 turnours were higher, but showed large variations (Figure 3-8).

Clearly, there was no correlation between estrogen-receptor levels and PSA production.

Progesterone-receptor levels in the tumours of T47D mice were high as compared to BT474,

MCF7, or BT20 tumours (Figure 3-9). Al1 tumours showed dom-regulation of progesterone-

receptor levels d e r 7-day norgestrel-stimulation. The low progesterone-receptor leveIs of MCF7

and BT2O tumours, d e r 7-&y stimulation, are consistent with their low rate of PSA production.

However, the Iow levels of progesterone-receptors in BT474 tumours are not consistent with the

high PSA levels detected in these tumours.

Control BT474 7 Days

- - ~

7 Days

Treatments

Control 7 Days

Figure 3-7. Mean and SD of [PSA] in tumours of mice fiom different treatment groups. The mice were either injected with sesame oil and functioned as a control or were stimulated with norgestrel for 7-days. Number of mice/group (n) are shown.

control 5 days 7 days control 7 days control 7 days 7 days

T47D BT474 MCF7 BT20

Figure 3-9. Mean and SD of the level of progesterone receptors in T47D, BT474, MCF7, and BT20 mice. The nurnbcr of mice per treatment group (n) are shown.

1 O0

The receptor experiments show that PSA synthesis and expression is a complicated

process that may not only involve the action of progesterone-receptors. It is also possible that

the progesterone-receptors in the T47D cells/tumours, although in high concentration, rnight be

mutated and thus non-functional (McGuire et al., 1977; Yamamoto, 1985). The irnrnunoassays

used in these studies for the estrogen- and progesterone-receptor analyses (Chapter 2) measured

the levels of the receptors and not their functionality. To better understand the role of the

receptors in PSA production, in future studies, it is important to determine whether these

receptors are active.

From fiese experiments it is possible to conclude that the MCF7 ce11 line is not usehl

for studying PSA production by breast tumeurs since it appears to have unstable characteristics

in vitro and in vivo. The BT474 cell-line is more usefiil for PSA induction since BT474 tumom

produce high PSA levels. Nevertheless, the use of this cell-line is not recommended since

tumours developed in only about 20% of mice injected, when either SCID or NOD/SCID mouse

strains were used (data not shown). Based on the data presented in Chapter 2 and in this section,

PSA induction can be ba t achieved in T47D cells growing as xenografk in SCID mice.

3.3. FUTURE DIRECTIONS

3.3.1. Effect of PSA Production on the Sensitivity of Human Breast Cancer Cell-Lines to Chemotherapeutic Drugs and Ionizing Radiation - In Ktro Studies

It has been shown that breast cancer patients with PSA-positive himours have a better

prognosis than patients with PSA-negaiive tumours. However, the function of PSA in the breast,

if any, is yet to be fowid. As show in previous sections, PSA did not affect the growth rate of

breast cancer 41-lines, inconsistent wiîh the hyjmthesis that tumours which produce PSA

101

proliferate more siowly than tumours that do not produce PSA. Another possibility is that PSA-

positive cells are more sensitive to chemotherapeutic drugs a d o r radiation, thus, causing patients

with PSA-positive turnows to respond better to treatments and have a better prognosis. To test

this alternative hypothesis, the T47D and BT474 ce11 lines will be used for the following

experiments. The cells will be first exposed to doxombicin, 5-fluorouracil, or methotrexate at

diffkrent concentrations to determine the toxicity range. These cimg are routinely used alone

or in combinations in the treatment of breast cancer patients with adjuvant therapy (Clavel and

Catimel, 1993; Mouridsen, 1993). Mer choosing the appropriate drug concentrations, cells will

be stimulated with norgestrel to produce PSA, as described in Section 3.2.1. Control samples

will not be stimulated and therefore should not produce PSA. The BT20 cell-line will hction

as a second negative control and will be stimulated with norgestrel.

Stimulation of cells with norgestrel will start on "day O" and continue until "day 3", to

allow cells to reach exponential ce11 growth and PSA production (Section 3.2.1). On "day 4":

doxombicin, 5-fluorouracil, or methotrexate (at the appropriate concentrations) will be added tc

the cells for 3 days. At the end of the incubation period, ce11 nurnber will be analyzed using the

MïT assay, as described in Section 3.2.2. If a difference in ce11 number is fowid between the

norgestrel-stimulated and non-stimulated treatrnent groups, the colony assay will be used to more

specifically assess the sensitivity of PSA producing cells to these drugs. In the future, these

experiments could aid clinicians in selecting effective therapies for breast cancer patients with

PSA-positive tumours.

Radiotherapy is ofien used as part of standard treatment for breast cancer after the primary

hunow has been removed. Due to the risk of undetectable microscopic cancer cells remaining

102

in the breast after the surgery, radiation therapy plays an important role in preventing the growth

of these tumour cells into a recurrent tumour (Dollinger et al., 1995). Therefore, to understand

the role of PSA in breast cancer, it is important to investigate whether there is a clifference in the

sensitivity of PSA-positive vs. PSA-negative cells to ionizing radiation. Based on clinical data

(Yu et al., 1995), it is hypothesized that PSA-producing cells will be more sensitive to radiation

than çells that do not produce PSA. To test this hypothesis, cells will be fmt irradiated with

different doses of ionizing radiation to detennine the surviving fi-action of non-stimulated

T47D/l3T474 cells. men, cells will be eitfia stimulated with norgestrel or with medium (control

group) fiom "day O" to "&y 6". On "day 7", cells will be exposed to the proper dose of

radiation, washed, trypsinized, and plated for survival analysis. Clonogenic assays will show

whether PSA production affects the sensitivity of cells to ioniziig radiation.

3.3.2. Effect of Chemotherapeutic h g s and Ionizing Radiation on PSA-Positive Tumours vs. PSA-Negative Tumours, Growing as Xenografts in SClD Mice

If the in viho results confirm the sensitivity of PSA-positive cells to either

chemotherapeutic dmgs or ionizing radiation, a mouse mode1 will be developed for investigating

this phenomenon in vivo. In the absence of positive in vitro results, lirnited studies will still be

carried out to assess possible micro-environmental effects that may mur in PSA-positive and

PSA-negative turnours. Initially, SCID rnice bearing T47D tumom will be stimulated with

norgestrel to produce PSA (as described in Section 2.3.3) or with sesame oil (negative control).

Then, mice will be divided into 2 groups, both groups containing norgestrel-stimulated and non-

stimulated animals. The fmt group of mice will be injected with chemotherapeutic drugs, to

which PSA-positive cells were found sensitive in vitro. The response of tumours to these dnigs

101

will be deterrnined by measuring leg diameter. Tumou regrasion will show whether the

treatment affects PSA-positive turnours relative to PSA-negative twnours.

The second group of mice will be treated with a single dose of ionizing radiation locallj

to the tumour. Turnorn of both norgestrel-stimulated and non-stimulated mice will be measurec

to evaluate the effect of radiation on tumour six. If PSA-positive tumours are found to be mort

sensitive to either chemotherapy or ionizing radiation as compareci to PSA-negative tumours

these results could have major implications to the treatment selection for some breast cancei

patients. Furthemore, these data could contribute to unravelling the hct ion of PSA in th(

female brwt.

Clavel, M. and Catirnel, G. (1993). Breast cancer: chernotherapy in the treatment of advanced disease. Eur. J Gncer 29A, 598-604.

Dollinger, M., Rosenbaurn, E. H., and Cable, G. (1995). Everyone's Guide to Cancer Therapy: How Cancer 1s Diagnosed, Treated, and Managed Day to Day. 2nd ed. (ed. Hasselback, R). Toronto, ON: Somarville House.

Hendrickson, E. A. (1993). The SCID mouse: relevance as an animal mode1 system for studying human disease. Am. J Pathul. 143, 15 1 1- 1522.

Lai, L. C,, Erbas, H., Lennard, T. W. J., and Peaston, R T. (1996). Prostate-specific antigen in breast cyst fluid: possible rde of prostate-specific antigen in hormone-dependent breast cancer. Int. J Cancer 66,743-74.

Larochelle, A., Vormwr, J., Lapidot, T., Sha, G., Furukawa, T., Li, Q., Shultz, L. D., Olivieri, N. F., Stamatoyannopoulos, G., and Dick, J. E. (1995). Engraftment of immune-deficient mice with primitive hematopoietic cells £hm Pthalassernia and sickle ce1 anernia patients: Implications for evaluiition human gene therapy protocols. H m . Mol. Genet. 4, 163-172.

McGuire, W. L., HoMitz, K. B., Pearson, O. H., and Segaloff, A. (1977). Current status of estrogen and progesterone receptors in breast cancer. Cancer 39, 2934-2947.

Mouridsen, H. T. (1993). Adjuvant systernic therapy in breast cancer. Eur. J Cancer 294 595- 598.

Shultz, L. D., Schweitzer, P. A., Christianson, S. W., Gott, B., Schweitzer, 1. B., Tennent, B., McKenna, S., Mobraaten, L., Rajan, T. V., Greiner, D. L., and Leiter, E. H. (1995). Multiple defects in innate and adaptive irnmunologic function in NODILtSz-scid rnice. J. Immunol. 154, 180-191.

Sutherland, R L., Hall, R E., Pang, G. Y. N., Musgrove, E. A., and Clarke, C. L. (1988). Effect of medroxyprogesterone acetate on proliferation and ce11 cycle kinetics of human mammary carcinoma cells. Cancer Res. 48, 50845091.

Yamamoto, E. R (1985). Steroid receptor regulated transcription of specific genes and gene networks. Annu. Rev. Genet. 19, 209-252.

Young, C. Y. F., Andrews, P. E., Montgomery, B. T,, and Tindall, D. J. (1992). Tissue-specific and hormonal regulation of human prostate-specific glandular kallikrein. Biochemist~ 31,818- 824.

Yu, H., Diamandis, E. P., Zarghami, N., and Grass, L. (1994a). Induction of prostate specific antigen production by steroids and tamoxifen in breast cancer ce11 lines. B m t Cancer Res. Tr. 32,291-300.

Yu, H., Diamandis, E. P., and Sutherland, D. J. A. (1994b). Immunoreactive prostate-specific antigen levels in female and male breast tumors and its association with steroid hormone receptors and patient age. Clin. Biochem. 27, 75-79.

Yu, H., Giai, M., Diamandis, E. P., Katsaros, D., Sutherland, D. J. A., Levesque, M. A., Roagna, R, Ponzone, R, and Sismondi, P. (1995). Prostate-specific antigen is a new favorable prognostic indicator for women with breast cancer. Cancer Res. 55,2 104-21 10.

Zarghami, N. and Diamandis, E. P. (1996). Detection of prostate-specific antigen mRNA and protein in breast tumors. Clin C h 42, 361-366.

Zargham, N., Grass, L., and Diamandis, E. P. (1997). Steroid homme regdation of prostate- specific antigen gene expression in breast cancer. Brit. J &mer 75,579-588.

APPLIED A IMAGE. lnc 1653 East Main Street - -. , Rochester, NY 14609 USA -- --= Phone: 71 61482-0300 -- -- - - Fax: 71 61288-5989

0 1993. Applied Image. Inc.. All fiights Reserved