MECHANISTIC AND CLINICAL ASPECTS OF PROSTATE SPECIFIC ANTIGEN EXPRESSION IN

NON-PROSTATIC TISSUES

Nosratollah Zarghami

A Thesis subrnitted in conformity with the requirements for the Degree of Doctor of Philosophy

Graduate Department of Clinical Biochemistry University of Toronto

OCopyright by Nosratollah Zarghami 1997

National Libraiy u*m of Canada Bibliothèque nationale du Canada

Acquisitions and Acquisitions et Bibliographie Services services bibliographiques

395 Wellington Street 395. rue Wellington Ottawa ON K1A ON4 OnawaON KlAON4 Canada Canada

The author has granted a non- exclusive licence allowing the National Library of Canada to reproduce, loan, distribute or seIl copies of this thesis in microfom, paper or electronic formats.

The author retains ownership of the copyright in ths thesis. Neither the thesis nor substantial extracts fiom it may be printed or otherwise reproduced without the author's permission.

L'auteur a accordé une licence non exclusive permettant à la Bibliothèque nationale du Canada de reproduire, prêter, distribuer ou vendre des copies de cette thèse sous la fome de rnicrofiche/fïlm, de reproduction sur papier ou sur format électronique.

L'auteur conserve la propriété du droit d'auteur qui protège cette thèse. Ni la thèse ni des extraits substantiels de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation.

MECHANISTIC AND CLINICAL ASPECTS OF PROSTATE SPECIFIC ANTIGEN EXPRESSION IN NON-PROSTATIC

TISSUES

Nosratollah Zarghami Ph.D. 1997

Department of Clinical Biochemistry University of Toronto

Abstract

Prostate specific antigen (PSA) is a 30 KDa serine protease which shares significant protein

and gene sequence homology with pancreatic-rend kallikrein and glandular kallikrein. Prostate

specific antigen, as its name denotes, was initially discovered in seminal plasma and extracts from

the prostate gland. Since the breast is a steroid hormone regulated organ and the knowledge that

PSA is regulated by androgens in the prostate, 1 hypothesize that the expression of PSA in breast

tissue is under the regulation of the steroid hormone receptor system. In this thesis, I have devised

appropriate experiments and tissue culture systems to study the reguiation of PSA in breast tissues

and ce11 lines.

The results from our studies on the PSA regulation in non-prostatic tissues yielded the

following :

1. We have developed reverse transcription-polymerase chah reaction (RT-PCR) methods for

detecting prostate-specific antigen (PSA) mRNA expression in tumors. Our data

. . 11

demonstrate that there is good agreement between presence of PSA protein and PSA mRNA

in breast tumors and that these tests give complementary information.

2. Based on clinical observations, we have developed a tissue culture system to study the

regulation of PSA gene in breast cancer. We have show that the breast cancer ce11 lines T-

47D and BT-474 produce PSA when stimulated by androgens, progestins and

glucocorticoids/ mineralocorticoids but not estrogens.

3 . Cyclical changes of PSA concentration during the menstmal cycle of normal women were

monitored and we found that PSA levels in serum are highest dunng the mid-late follicular

phase, drop continuously with a half-life of 5-6 days between the late follicular phase and

mid-cycle and reach a minimum during the mid luteal phase.

4. Our studies indicate that the PSA gene is expressed in prirnary lung tumors. The identity

of the PSA gene expression was venfied using PSA immunoassay, imrnunohistochernistry,

RT-PCR, Southem blot hybridization with a specific PSA RNA probe and DNA

sequencing. By cloning and sequencing of aberrant PCR products, we were able to isolate

a 450 bp unique sequence not previously deposited in GenBank.

Acknowledgements

1 would like to express rny gratitude and appreciation to Drs. Eleftherios Diamandis,

Reinhold Vieth, and Cliff Lingwood for their guidance, advice, encouragement, and support, and

Dr. Elefiherios Diamandis for his supervision of this research project and detailed review of the

dissertation. I would also like to th& Drs. Donald Sutherland, Dionyssios Katsaros, Mario

D'Costa, Edward Sauter, Linda Giudice, and The Canadian Red Cross Blood Transfùsion Service

for providing human specimens, usehl information, helpfùl discussion, advice, technical support

and help. Finally, I would like to specially thank Linda Grass, Dr. He Yu, Michael Levesque,

Katerina Angelopoulou and Lisene Santos for their generous help and contributions to this research

project.

Table of Contents

CHAPTER PAGE

CHAPTER 1. Introduction

CELAPTER 2. Literature Review 2- 1. A histoncal Perspective of PSA 2-2. Structural and Molecular Features of PSA 2-3. The PSA Gene 2-4. Regulation of PSA Production

2-4- 1. Positive Control Mechanisms 2-4-2. Negative Control Mechanisms

2-5. Effects of Growth Factors on PSA Gene 2-6. Biological Function of PSA 2-7. Moiecular Foms of PSA 2-8. PSA in Non-Prostatic Tissues

2-8- I . Penurethral Glands 2-8-2. Breast

2-8-2- 1 Breast Cancer Tissue 2-8-2-2 Breast Cancer Cells 2-8-2-3 Other Bremi Tissues 2 -8 -24 Breast Fhr ids 2-8-2-5 PSA Regdation and Fzinction in Breast Tissue 2-8-2- 6 Chical Applications

2-8-3. Arnniotic Fluid 2-8-4. Female Semm 2-8-5. Ovary, Endometrium and Other Tissues 2-8-6. Other Cancers

2-9. S teroid Hormone Receptors 2-9- 1. Mechanisms of Gene Regulation 2-9-2. Clinical Application of Estrogen & Progesterone Receptors

in Breast Cancer 2-9-3. Steroid Hormone Receptors and Cancer

CEMPTER 3. Hypothesis

CHAPTER 4. Objectives

CHAPTER 5. Materials and Methods 5- 1. Preparation of PSA cDNA

5- 1- 1. Transformation Procedure 5-2. Positive and Negative Control Ce11 Lines 5-3. Preparation of Tumor Cytosol Extracts

5-4. PSA Measurements 50 5-5. High Performance Liquid Chromatography (HPLC) 5 1 5-6. Total Protein Determination 52 5-7. Isolation of Total RNA 52 5-8. Reverse Transcriptase 53 5-9. Oligonucleotide Prirners 54 5- 10. PCR Protocol 54 5- 1 1. Nested Primer PCR 55 5- 12. Labelling of PSA cDNA Probe 5 5 5- 13. RNA Labelling by In-Vitro Transcription 56 5- 14. Direct Incorporation of DIG-dUTP Dunng PCR 56 5- 1 5. Detection Limit of PCR 57 5- 1 6. Gel Electrophoresis, Southem Blotting and Hybridization 57 5- 1 7. Detection Protocol 58 5- 18. Cloning of PCR Products 59 5 - 1 9. DNA Sequencing 59 5-20. Tissue Culture Systern 59

5-2 O- 1. Compoundî 59 5-20-2. Cell LNies 60 5-20-3. CeII Culture Procedure 65 5-20- 4. Stimulation wzth Steroid Compoirndr 65 5-20-5. Stimzilution with Semm Sampies and Goriadotropin Hormones 66 5-20-6. Blochg Experiments 66 5-20- 7. Kineîic Ewperiments 67

5-2 1 . Lysis Procedure 67 5-22. Measurement of Steroid Hormone Receptors 67 5-23. Measurement of Hormones in Serurn 68 5-24. Irnmunohistochemistiy 68 5-25. Subject Selection 69

3-25 1. Breast Tissue Specimens 69 5-25-2. Serum Specimens Ob tained Dtiring Menstmal Cycle 69 5-25-3. L ung Cancer Patients 70

5-25-3- 1. Case Study 70 5-25-3-2. Lzrng Ttimor Tissues 71

5-26. Ethical issues 7 1 5-27. Statistical Analysis 73

CHAPTER 6. Results 6- 1. Optimization of PCR Protocol 6-2. Determinhg Limits of Detection Methods

6 - 2 4 Ethidium Bromide Stuining 6-2-2. Nested PCR 6-2-3. Direct DIG-dUTP Incorporation 6 - 2 4 Souîhern BIot Hybridization

6-3. PSA Gene Expression in Breast Tumors 6-4. Steroid Hormone Regulation of PSA Gene Expression in Breast Cancer

6-4- 1. Kinelics of PSA Ekpresszon 6 - 4 2 Induction of PSA Produciion by Steroid Homones 6-4-3. Detemination of Dose-Response Stimulatory Effecls of

Steroid Hormones 6- 4-4. Blocking Steroid Hormone Receptors

6-5. Variation of PSA Protein During the Menstmal Cycle 6-6. Expression of PSA Gene in Lung Tissue 6-7. Frequency of PSA mRNA Expression in Lung Tumors

CEMPTER 7. Discussion 7- 1. Detection of PSA Protein and rnRNA in Breast Turnors 7-2. Steroid Hormone Regulation of PSA Gene in Breast Cancer 7-3. PS A Production Dunng the Menstrual Cycle 7-4. Expression of PSA Gene in Lung Tissue 7-5. Frequency of PSA mRNA Expression in Lung Tumors 7-6. Conclusions

CHAPTER 8. Concluding Remarks and Future Directions

CEIAPTER 9. References

vii

List of Figures

Page Figure

Figure 1- 1

Figure 1-2

Figure 2-2- 1

Figure 2-9- 1

Figure 6-2- 1 - 1

Figure 6-2-2- 1

Figure 6-2-3 - 1

Figure 6-3 - 1.

Figure 6-4- 1 - 1.

Figure 6-4- 1-2.

Figure 6-4- 1-3.

Figure 6-4- 1-4

Figure 6-4-2- 1.

Figure 6-4-3- 1.

Figure 6-5- 1

Figure 6-5-2

Figure 6-5-3

Mechanism of action of steroid hormones 2

Structural requirements for androgen regulation of gene transcription 3

Homology of the human tissue kallikrein gene farnily 11

Generai stmctural and fùnctional organization of nuclear receptors and concsensus responsive elements 35

RT-PCR of PSA niRNA and P -actin rnRNA 76

Nested primer PCR 78

Direct incorporation of DIG-dUTP in PCR reaction 79

Detection of PSA mRNA by hybridization 83

RT-PCR of PSA rnRNA extracted fiom T47D cells 86

Appearance of PSA rnRNA and PSA protein inside and outside 87

Representative chromatograrz of the PSA cDNA sequence 89

Complete sequence of the 6 16 nucleotide region of the PSA cDN A 90

PSA concentration in T-47D ce11 line tissue culture supernatant after stimulation with various compounds 92

Dose-response expenment of six representative steroids 94

Senim PSA and progesterone dunng menstrual cycle of M.T 99

Semm PSA and progesterone during menstmal cycle of C.J and L.S 1 O0

Semm progesterone levels during the menstrual cycle and its stimulation effect 102

Figure 6-5-4

Figure 6-6- 1

Figure 6-6-2

Figure 6-6-3

Figure 6-6-4

Figure 6-6-5

Figure 6-6-6

Figure 6-7- 1

Figure 6-7-2

Figure 6-7-3

RT-PCR of PSA mRNA and actin obtained fiom T-47D stimulate with semm of voIunteer M.M 1 03

Molecular weight identitication of PSA in the lung tissue extract 105

RT-PCR of RNA extracted fiom various tissues 1 06

Representative chromatograrn of the PSA cDNA sequence of the lung tissue 1 07

Complete sequence of the PSA cDNA sequence isolated fiom lung tissue 108

Imrnunohistochemical localization of PSA in lung tissue 110

Production of PSA by BT-474 cells d e r stimulation with becIomet hasone 112

RT-PCR for PSA mRNA extracted fiom lung tumors 115

RT-PCR for PSA mRNA extracted from lung tumors with aberrant bands 116

Sequence of unique 450 bp PCR product 118

List of Tables

Tabte Page

Table 2-2-1

Table 2-7- 1

TabIe 5-20- 1 - 1

Table 5-20-2- 1

Table 5-35-3-2- 1

Tab te 6-2-4- 1

Table 6-3- 1

Table 6-4- 1

Table 6-4-2- 1

Table 6-4-4- 1

Table 6-6- 1

Human tissue kallikrein gene family

Nomenclature of hK proteins

List of steroid/cornpounds tested for PSA production

Ce11 lines tested for PSA production in tissue culture system

Clinicopathologicai findings in primary lung tumors

Detection of PSA mRNA with various RT-PCR protocols

Detection of PSA protein and PSA mRNA in pnmary breast tumors

Steroid hormone receptor levels in different cell lines

Regulation of PSA gene expression by various compounds in the breast carcinoma ce11 Iine T-47D

Blocking of PSA production by vanous compounds

Cornparison of DNA sequence of PCR product found in lung tissue with sequence deposited in GenBank

Table 6-7- 1 Distribution of PSA protein in lung tumor cytosols 113

List of Abbreviations

AR

ARE

ApoD

ALP

BPH

dNTP

DIG-dUTP

DMSO

ER

EGF

EDTA

FGF

FSH

GR

hGK- 1 (hKLK2)

hKLK 1

hKLK3

IGF-1

IGF-II

IGFBPs

IGFBP-3

androgen receptor

androgen response element

apolipoprotein D

alkaline phosphatase

benign prostatic hyperplasia

deoxynucleoside triphosphate

digoxigenin-labelled deoxyuridine triphosphate

dirnethylsulfoxide

estrogen receptor

epidemal growth factor

ethylenediaminetetraacetic acid

fibroblast growth factor

follicle stimulating hormone

glucocorticoid hormone receptor

human glandular kallikrein 1

tissue kallikrein

human kallikreid (PSA)

insulin-like growth factor-1

insulin-like growth factor-II

insulin-like growth factor binding proteins

insulin-like growth factor binding protein-3

LB

LH

MR

PMSF

PR

PI

PBS

PKC

RA

RAR

RT-PCR

SSC

SDS

TR-FIA

TPA

TGF-a

TGF- P

TR

VDR

Luria-Bertani

Iuteinizing hormone

rnineralocorticoid hormone receptor

phenylrnethylsulphonyl fluoride

progesterone receptor

isoelectric point

p hosp hate-buffered saline

protein kinase C

retinoic acid

retinoic acid receptor

reverse transcription-polymerase chain reaction

saline sodium citrate

sodium dodecyl sulfate

time-resoived imunofluorometric assay

tumor-promoting phorbol ester

transforming growth factor alpha

transforming growth factor beta

thyroid hormone receptor

vitamin D receptor

1

CHAPTER 1. Introduction

Steroid hormone receptors belong to a superfamily of ligand-inducible transcription factors

that regulate hormone-responsive genes. Steroid hormone responsive tissues can be found in the

breast, prostate, endometriun, uterus, ovaries, and other diverse tissues. These tissues have steroid

hormone receptors that are present as inactive proteins, but once they associate with their cognate

hormone, they become activated, bind to specific regions of the DNA, and regulate hormone-

responsive genes. The current mode1 for steroid hormone action (Figure 1-1) starts with the passive

diffùsion of steroids into the cell. The steroid then binds noncovalently and with high affinity to the

steroid hormone receptor in the cytoplasm. Ligand binding causes the receptor to undergo a process

called activation or transformation in which a conformational change in the receptor is thought to

occur. The receptor dissociates frorn the nonsteroid-binding subunits with which it is normally

associated, unmasking the DNA binding domain of the receptor protein. The steroid-receptor

complex then translocates to the nucleus where it binds to specific DNA sequences called steroid

hormone-regdatory elements (HREs) such as androgen-regulatory elements (ARES). M e r binding

to HRE, the steroid-receptor complex alters the transcription rate of specific genes near or in which

the ARE is located. A typical androgen-regulatory responsive gene is shown in Figure 1-2. The

.4RE, shown with the consensus DNA sequence, is located in the 5' regulatory region of the gene

where an adrogen receptor bound to the ARE can interact with transcription factors that bind to

other regulatory elements, such as the TATA and CAAT boxes, which are also present in this

region. In this fashion, androgens increase or decrease the amount of messenger RNA and

ultimately the level of protein that is synthesized fiom these genes, and thus alter cellular functions.

The presence of such receptors in tumors is considered a good prognostic indicator since the tumor

Figu 1-1. R d t of hormone in s i~oid ceceptoi u t i ~ a t i a i . Progestcron i ~ c p t œ trisu in a tranuripion>lly inad* 8-102 cnnplct ia association wi'th k a t hdc pmteim hsp90, hg70, rd w56 in the Jwm of homsrr. Addition d hamne (Hl chnaes tk anformation d che tuepta (dbwcd by dissœiation of M. -. r fm of hsps, hich c m be *id rher u l t nd AlP u u m t , still rcquircr hamonc br mnscriptiavl xtinücn. hamnt bdcd recepa & d r to its mponw t * m n t (HU). kcom p h p h o y l a t d (m. a*ta 6- tmmaiptiorr (rcpceimted by an a m W .

CONSENSUS ELEMENTS

ENHANCER PROMOTER ELEMENTS ELEMENTS

,

GGTAC AXWTTTCT

Fig e 1-2. Structurai requirements for androgen regulation of gene transcription.

i

ARE ARE CAAT TATA -1 . BOX BOX T T

TRAKSCWPTION TERMTNATlON START SITE S I E

GENE - 3- r -

h - -

4

will iikely respond to hormonal treatrnent with hormone antagonias which block the receptors and

ultimate turnor proliferation. However, not al1 receptor positive tumors will respond to endocrine

treatrnent and this finding was proposed as an indication that in sorne tumors the receptors are

inactive or defective, or post-receptor defects may be present. In the prostate, for example, cells

have androgen receptors (AR) that respond to the hormone dihydrotestosterone. This mechanisrn

" up-regdates" the gene that produces prostate-specific antigen (PS A). In the breast, estrogen

regulated genes include pS2, cathepsin-D, and the progesterone receptor. Apo D is another recent

known biochernical marker that is up-regulated by androgens in human breast cancer ce11 lines.

Apo D protein is present in seminai fluid and gene expression is also under the regdatory function

of androgen receptor. Expression of pepsinogen C, an aspartyl proteinase enzyme, is also under

the control of steroid hormones in the breast. Studies on the hormonal regdation of pepsinogen C

in breast cancer cells have revealed that androgens, glucocorticoids and to a lesser extent

progesterone, are able to induce the expression of this gene in breast cancer ce11 lines. The

measurement of steroid hormone receptor-inducible proteins is one of the ways to test for the

presence of active receptor molecules in the tumor cells. The inducible proteins may have better

prognostic significance than the receptors thernselves. In this project we examined the mechanisrn

of production and clinical aspects of prostate specific antigen (PSA) in breast and lung tissues.

Prostate specific antigen (PSA) has become an important rnarker in clinical medicine over

the last decade. Serum PSA is a routine marker of prostatic disease, especially for prostate

adenocarcinorna, since the first methods becarne available for in-vitro testing in the rnid- 1980s.

Over the past few years, our understanding of the molecular structure, physiology, and

pathophysiology ofthis protein has grown irnmensely: PSA is no longer simply an "antigenNdefined

5

by a specific antibody but rather, a protein with a biological role, evolutionary history and clinical

significance.

PSA is a 33 kDa single chain glycoprotein and a senne protease with chymotrypsin-like

enzymatic activity. Until recently, PSA was thought to be produced exclusively by the prostate

epithelial cells lining the acini and ducts of the prostate gland and to be absent fiom al1 female

tissues and fluids. PSA is present at high concentrations in seminal plasma and used as a marker

for screening, diagnosis, and monitoring of prostate cancer (1 -3). Immunohistochernical detection

of PSA in tumors of non-prostatic ongin has been reported as a rare event (4). Since the recent

description by Our group of the production of PSA protein by breast tumors (5 ), we have

demonstrated that PSA is present in 30-40% of femaie breast tumors (6) , more rarely in other

tumors (7), in normal breast tissue (8), and in biological fluids including breast rnilk (9), breast cyst

fluid (lO), breast discharge fluid and amniotic fluid (1 1). Some female breast tumors produce very

high levels of PS A while othen do not express any PSA.

Association analysis between PSA levels and levels of progesterone (PR) and estrogen

receptors (ERS) in female breast tumors, clearly demonstrated that PSA was associated with the

presence of the PR, but not the ER. Survival analysis reveaied that patients with breast tumors

producing PSA live longer and relapse less fkquently than patients with tumors that do not produce

PSA. Thus, PSA is a new favourable prognostic indicator in female breast cancer (12).

The regdation of this glycoprotein in-vivo has important clinical application. There is

increasing evidence that transcription of the prostate specific antigen (PSA) gene is androgen

regulated. The PSA prornoter contains a sequence which is closely related to the ARE (androgen

response element) consensus sequence. This sequence is a high af£inity androgen binding site and

6

acts as a fùnctional ARE in transfected LNCaP cells (13). PSA gene regulation has been studied

extensively in the LNCaP prostate cancer ceU line and has been found to be positively regulated

by a number of steroid hormones, such as R5020, estradiol, and epitestosterone ( 1 3). However, the

androgen receptor in this ce11 line has been found to have a point mutation which leads to a relaxed

atflluty for steroid hormones and, as such, the LNCaP ce11 line may not be the ideal cell mode1 with

which to study the regulation of this gene.

The growth and differentiation of the mammary gland is a result of complex interaction of

a number of endocrine hormones which act alone and in combination. The development and

progression of breast cancer is also strongly influenced by hormonal status, yet the molecular

mechanisms underlying the complex pattern of multihomone regulation of marnrnary epithelium

are not well understood. Given the fact that PSA is a steroid hormone-regulated protein in the

prostate, we thought that PSA may indeed be expressed in non-prostatic tissues. Particularly, its

presence in breast cancer patients may reflect the existence of a complete hormone receptor

pathway and that understanding of its fùnction and regulation in this tissue may have importance

in breast cancer research both in basic and clinical applications. For these considerations, we were

prompted to examine the possibility that production of PSA protein and PSA mRNA by breast

cancer cell lines could be mediated by homonal factors and their related receptors.

CHAPTER 2. Literature Review

2-1. A Historieal Perspective of PSA

PSA is a normal constituent of prostate secretions. Very few other tissues or organs have

been demonstrated to produce this protein. Its history and biochemistry have been extensively

reviewed (1 4). In brief, PSA was initially discovered in seminal plasma and was named gamma-

seminoprotein by a group of Iapanese scientists who were searching for a sernen marker in the late

1960s (1 5). Over the past few years another group of researchers isolated a similar protein from

semen and called it protein E, based on its electrophoretic mobility (16). In 1978, Sensabaugh

characterized this protein biochernically with the use of an immunoelectrophoretic method and

named it as p30 according to its molecular weight (17). Because of its NI-vitro stability,

immunogenecity, and tissue specificity, p30 was considered to be a usefûi marker in criminai and

forensic investigations ( 1 8).

PSA was found in prostate by Wang et al in 1977 (19) and was later isolated from this tissue

by the sarne research group (20). Since this protein was identified only in prostatic tissue at that

time, it was named prostate-specific antigen. Wang et al also demonstrated that PSA was present

in seminai plasma and that it was similar to the PSA they found in the prostate (2 1). Now, although

it has not been universally accepted that PSA is the only protein under discussion, consensus has

been reached that gamma-seminoprotein, protein E,, p30, and PSA are the same protein (22). PSA

was later identified in male serum by the same group who found PSA in the prostate (23).

Over the last few years, a nurnber of reports have challenged the absolute tissue specificity

of PSA. However, many groups have reported immunoreactivity for PSA in a variety of normal and

neoplastic tissue types. These include rare apocnne sweat gland carcinoma (24), apocnne breast

8

carcinorna (24), salivary gland neoplasm (4), colon carcinomas (25)- pancreatic acinar ce11

carcinoma (26), pnmary ovmian carcinoma (27), adenocarcinorna of paraurethrai gland (28,29),

biliary tract carcinomas (25), and bladder carcinoma (30). Moreover, PSA has been

immunohistochernically demonstrated in normal axillary and perineal apocrine sweat glands (24),

in addition to the paraurethral (3 1) and perianal glands (32). In most cases, PSA positivity was

obtained when polyclonal antibodies were used in imrnunohistochernistry. However, none of these

tumors stained positive when monoclonal PSA antibodies were used, and it was concluded that the

positive staining was due to cross-reactivity of the polyclonal antibodies.

The isolated reports that challenged the absolute tissue specificity of PSA did not attract

much interest, because of their rarity. However, with newly developed ultrasensitive imrnunological

assays for PSA, our group recently reported that PSA is frequently present in extracts of normal,

hyperplastic and cancerous breast tissues (5-8), in female discharge fluid and in breast rnilk (9).

PSA was also found in some ovarian tumor extracts (28), in endometrium (33)- and in many other

tumor types (7).

2-2. Structural and Molecular Features of PSA

PSA is a single-chah glycoprotein of approximately 33 kDa. It has been initiaily

demonstrated that PSA has 240 amino acid residues (34), howwer, based on sequence of its mature

protein is believed to consist of 237 amino acids (35,36). The PSA polypeptide chah has five

interna1 disulphide bonds and these bonds allow saminai PSA to have two different rnolecular

foms. One has a complete and intact structure known as native form; the other has a cleavage at

lysine-lysine bond located between amino acid residues 145 and 146 and is called "nicked".

9

Because of the five intemal disulphide bonds in the mature protein, the nicked molecular form

could CO-migrate with the native form on a sodium dodecyl sulphate polyacrylamide gel

electrophoresis (SDS-PAGE) under no-reducing conditions, but under denatunng/reducing

conditions, this form of PSA separates fiom the native molecular form becoming PSA fragments

with smaller molecular masses (34-36).

Human PSA is also known to be composed of 7 to 12 % carbohydrates (37). Based on

different analytical techniques, the rnolecular mass of this single polypeptide chah could be

different. The molecular weight of the complete protein is around 33 kilodalton (kDa) determined

by gel filtration and gel electrophoresis (38). Based on the component of the PSA protein, i.e. 240

or 237 amino acid residues, the estimated molecular weight is 26.5 or 26.1 ma, respectively (34-

3 6).

Recently, carbohydrate structure analysis of PSA protein revealed that there are 4

carbohydrate side chains conjugated to the PSA polypeptide chain. Of the 4 sugar chains, one is N-

linked to an asparagine at position 45, two O-linked to a senne at positions 69 and 7 1, and one 0-

linked to a threonine at position 70 of the arnino acid sequence (34). However, other studies

reported that there is oniy 1 carbohydrate side chah N-linked to PSA at asparagine. at position 45

(22). The carbohydrate chah consists of hexose, hexosarnine, and sialic acid (20). Approximately

1% of the 7% carbohydrate residues in the PSA molecule is neuraminic acid giving rise to

microheterogeneity when the protein is subjected to isoelectrophoresis. This could explain the

presence of several PS A isofoms with different isoelectnc points (PI) ranging From 6.8 to 7.5. The

major isoform has PI of 6.9 (39). Some features of the human tissue kallikrein gene family are

summarized in Table 2-2- 1 and Figure 2-2- 1 .

TABLE 2-2-1 Numan tissue kollihrein gene fornily

Old

% muence NO. 01 Amino Acids For Mature Proteln:

New lden t i ty in Reglon: M W (kOa) Pr Cene Protein Designation Designation ~ o Q S A pre pro mature obsewed cakulatedb oôsennd calculated hKLK I Pancreaticlrenal kalltkretn hPRK hK I -62 17 7 238 ND 26.4 ND 4.4 hKLK2 Hurnan glandular kallikrein- l hGK- 1 hK2 - 80 17 7 237 ND 26.2 ND 6.9 hKLK3 Prostate-specific antigen PS A hK3 1 O0 17 7 237 28.5 26.1 6.8- 7.2' 7.4

2.3. The PSA Gene

A gene encoding PSA was identified and cloned by several groups in the 1980s (40-42).

The PSA gene is located on the long arm of chromosome 19 and spans a DNA region of DNA about

6 kilobases (kb) in length. It contains 5 exons, 4 introns, 2 promoter sites, and an untranslated

region at 3' end. PSA gene (KLK3 or hKLK3) belongs to a human kallikrein gene family which

contains two other members, the human glandular kallikrein 1 gene (hGK- 1, KLK2 or IiKLKî) and

the tissue (rendpancreatic) kallikrein gene (KLK1 or W U ) . The PSA gene is approximately

80% homologous with human glandular kallikrein (hGK-1) gene and 68% homologous with tissue

kallikrein (hKLK 1) gene (40,43,44). In addition, recent studies of hGK- 1 mRNA suggest that PSA

and hGK-1 genes are very sirnilar as they both seem to be present only in the prostate and are

regulated by androgen via androgen receptors (45.46). In the prostate, the amount of hGK- 1 mRNA

is only 10% to 50% of the arnount of PSA rnRNA (47,48). These two genes, based on their DNA

sequences, are supposed to encode two similar proteins in terms of their polypeptide structure and

enzyrnatic activity (49). Since the protein encoded by hGK-1 gene has not been identified yet,

Peehl suspects that seminal PSA is actually a mixture of two proteins which are separately encoded

by PSA and hGK-1 genes (50). However, no one has been able to prove this hypothesis. Based on

the nucleotide sequence of the hGK-1 gene, the protein is predicted to have trypsin-Iike substrate

activity, in contrast to the chymotrypsin-like activity of PSA (50).

In spite of many similarities between PSA and kallikrein genes, it is stiil believed by other

investigaton that these genes and their related proteins are distinct from each other in many aspects

(45). Fust, these genes are located far apart on the chromosome 19; PSA gene resides 12 kb away

13

from the hGK- 1 gene and at least 15 kb away fkom the tissue kallikrein gene (5 1,52). Secondly, the

promo ter site and other untranslated regions are different between these genes (40,4 1,43.50).

Finally, a crucial difference is found in some arnino acids that determine the binding specificity of

the enzymes to their substrates (35.4 1). PSA is believed to have chyrnotrypsin-like activity since

a serine is present at one of the substrate binding sites (arnino acid 184). whereas hGK-1 is

supposed to have trypsin-like activity as the position is occupied by an aspartic acid.

Several RNA species have been reponed to transcribe frorn the PSA gene (48). A 1.6 kb

transcript is considered to be the major species being translated into the PSA protein. The other two

transcnpts with 1.9 and 0.9 kb in length are thought to be the products of alternative splicing.

Different PSA M A S have been seen in benign prostatic hyperplasia (BPH) and prostate cancer

(48.53). However, no one has been able to identify any different forms of PSA protein translated

from these transcripts as being clinically distinct as well as useful.

2-4. Regulation of PSA Production

2-4- 1. Positive Control Mechanisms

Evidence for androgen regulation of the PSA has been denved pnrnarily from clinical

studies. Goldfarb et a1 (54) studied the age-related changes in PSA expression in human prostate

utilizing irnmunohistochemical techniques. PSA levels in prostatic tissue are altered significantly

d e r birth throughout puberty and these aiterations match well to the changes of androgen level in

serum (5 5). Weber and associates (5 6 ) demonstrated a significant correlation between circulating

androgen levels and semm PSA. At birth, serurn PSA detectable in both genders and no difference

exias between males and femaies. After birth, the levels start to drop to undetectable levels within

14

several months and they begin to rise again in male teenagers at puberty. Considerable evidence

exists in support of clinid studies that therapeutically increasing or decreasing testosterone levels

in the male body are always followed by changes in semm PSA ievels (52,5 8).

Based on these clinicd observations, many investigators suggested a possible role of

androgen in the regulation of PSA gene expression. Since LNCaP cells were used to study the

androgenic effkts on PSA mRNA expression, convincing evidence for androgen regulation of PSA

gene was provided. For example, Northern blot analysis showed that PSA mRNA increases rapidly

(within 2-6 hours) following androgen treatment of LNCaP cells (59). DNA sequence analysis of

the promoter region of the PSA gene revealed a consensus androgen-response element (ARE)

located approximately 160 base pairs upstream of the transcnpt start site (59). Young et al have

demonstrated that in their ce11 culture experiments with the use of the prostate cancer ce11 line

LNCaP, the production of PSA was up-regulated by androgens through the androgen receptor and

that the regulation was accomplished at the level of PSA gene transcription (59). Recently. it has

been demonstrated that vitamin D, as well as severai paracnne factors secreted fiom prostatic

stromal cells can up-regulate PSA protein expression in the absence of androgen (60). An autocrine

factor produced by a subline of LNCaP was found to up-regulate PSA mRNA (6 1 ) .

2 - 4 2 Negatzve Control Mechanisms

There are at least two signalling pathways that exhibit a negative effect on androgen

induction of PSA expression. One is the phorbol ester-activated protein kinase C (PKC) pathway.

In further experiments, a tumor-promoting phorbol ester (TPA) that activates PKC, had profound

effects on blocking androgen induction of PSA rnRNA and protein, whereas forskolin, a protein

15

kinase A activator, had linle or no effect on the expression of PSA (62,63). Therefore, it is

suggested that the PKC activity rnay function to counteract androgen action.

A second pathway that negatively regulates androgen action is mediated by retinoic acid

(RA). Young et al (63) has shown that RA can inhibit ce11 growth as well as repress androgen

induction of PSA &A in LNCaP cells. Kinetic studies have shown that at least 16 hours of RA

treatment were required to effect the repression of PSA gene. This is in marked contrast to the TPA

effect, which occurred fairly rapidly (4-6 hours). Moreover, AR protein levels were also reduced

afier 24 hours of RA treatment. Therefore, it has been postulated that the antagonistic effects of RA

on the proueration and function of prostatic cells may be at least in part via modulatory effects on

the AR, although the acnial mechanism remains to be determined.

2-5. Effects of Growth Factors on the PSA Gene

In addition to steroid hormones, growth factors are thought to also participate in the

regulation of PSA gene expression. Results From ce11 culture studies indicated that epidermal

growth factor (EGF) and transforming growth factor alpha (TGF-a) could down-regulate the level

of PSA mRNA in prostate cancer cells and that TGF-P and bFGF seemed to have little influence

on the regulation of PSA expression (64,13). It was suggested in another study that some

unidentified autocrine regulators may exist modulating the expression of PSA gene in prostate

cancer cells which grew independently fi-om androgen (61). Extracellular matrix also seemed to

affect the expression of PSA gene through its action on the growth of prostate cancer cells (65).

16

2-6. Biologieal Function of PSA

In recent years PSA has been studied extensively and characterized well regarding its

rnolecular and biochemical aspects in biological fluids. Although the physiologicai role and

importance of PSA protein in non-prostatic tissue still needs to be elucidated, our knowledge of its

physiological functions is mostly limited to the findings in the serninal fluid. PSA Functions as a

serine protease and its catalytic triad consists of three amino acids including histidine 4 1, aspartate

96, and serine 189 (34-36). It has been shown that PSA is able to cleave other proteins at certain

leucine and tyrosine residues fiom the protein's carboxy-terminus (66-68). Arginine and methionine

residues at the carboxyl-terminal end have also been mentioned to be the possible cleavage sites

of PSA on protein C inhibitor (PCI) or secretory protease inhibitor, both of which are present in

serninal plasma (69,7O).

PSA is a serine protease, and in al1 tissues and fluids, the predorninant form is the non

complexed, 33 kDa free PSA monomer that is the enzymaticaily active form of this enzyme.

However, complexes with proteinase inhibitors also exist at much lower concentrations. The

sequence of PSA shows extensive hornology with y-nerve growth factor ( 56%), epidermai growth

factor-binding protein (53%), and a-nerve growth factor (5 1%) (35). In spite of having sequence

hornology with these molecules, it is not known that PSA has functions sirnilar to these factors. It

is proposed that PSA may enzymatically act upon one or more substrates and modiQ their action

in a fashion similar to the function of the other proteinases of the kallikrein family. Earlier studies

suggested that PSA had both chymotrypsin-like and trypsin-like enrymatic activities (68,71).

However, it is now believed that PSA has only chymotrypsin-like activity (50.72). The finding of

trypsin-like activity in PSA is suspected to be due to contamination of purified PSA with hGK- 1

17

protein which is supposed to have trypsin-like activity based on the amino acid sequence of kGK- 1

derived from its mRNA (72). However, since no one has been able to isolate the hGK-1 protein,

this view of PSA enzymatic activity is still hypothetical.

Findings by an in-vitro study showed that PSA can hydrolyse a number of substrates

including gelatin, interleukin 2, myoglobin, ovdbumin, fibnnogen, and insulin at both a and P

chains (34). In fact, major physiological function of PSA was initially demonstrated in seminal

plasma. PSA in seminal plasma was found to be able to digest two large seminal vesicle proteins

called seminogelin 1 and II, to smaller molecular weight fragments (66,67). Cleavage by PSA

occurs at carboxy-termini of certain tyrosine and leucine residues in seminogelin (73). Proteins

narned serninal-vesicle-specific antigens were found by another group to be the substrates of PSA

in semen (74) are thought to be identical to the seminogelins (72). The extracelluar matnx

component fibronectin is also a target of PSA's enzymatic activity (73). Fibronectin, iike

seminogelin, is a component of the seminal fluid and was also discovered to be cleaved by PSA in

semen (75,73). Liquefaction of seminal coagulum formed at ejaculation is known to be the

consequence of the cleavage of these proteins by PSA. This PSA function has been linked to

human fertility because, presumably, sperm motility is enhanced after the semen clot is liquefied

(50).

In spite of the PSA function in seminal plasma, the physiological role and importance of

PSA in non-prostatic tissues needs to be elucidated. Recent data on prostatic tissue could be

extrapolated to support the view that PSA may be involved in the growth regulation of mammary

and other tissues. Evidence indicates that PSA may play an imponant role in the regulation of

growth factors. It has been reponed that PSA levels in semm of prostate cancer patients are

18

associated with the semm levels of insulin-Iike growth factor binding proteins (IGFBPs) (76). The

IGFBPs form a large farnily of related proteins with at least six members (77). In one study, PSA

was shown to be able to digest one of the six members, IGFBP-3. This activity is thought to

regulate the insulin-like growth factor-1 (IGF-1) concentration, because digestion of IGFBP-3 by

PSA releases biologically active IGF-1 (78). insulin-like growth factors (IGFs) are known to be

important autocrine, paracrine, and endocrine peptide hormones having a variety of effects on the

proliferation and differentiation of various normal cells as well as on the transformation and

apoptosis of cancer cells (79,80). IGFs must bind to specific cellular membrane receptors to exert

their hnctions. The binding of IGFs to their receptors is modulated by another group of proteins

called IGFBPs (81). It is believed that IGFBPs carry IGFs to their target cells and regulate the

function of IGFs through blocking the binding of IGFs to their receptors. Thus, the digestion of

IGBP-3 by PSA results in the release of IGFs to which IGFBP-3 binds (82). In addition to its

binding to IGFs, IGFBP-3 may have other hnctions independent from IGFs because IGFBP-3

receptor has been found in the membrane of certain type of cells and IGFBP-3 could bind to those

receptors exerting its own hnctions (83). In this case, PSA may regulate the fùnction of IGFBP-3

per se. Ln addition, It has been found that PSA has rnitogenic activity, presumably due to activation

by PSA of latent transforming growth factor+ and through modulation of ceIl adhesion (84).

Al1 these observations have suggested that PSA rnight have a biologic role. This role is

dependent on the enzyrnatic activity of PSA. Proteolytic activity in tissues is usually tightly

regulated. Because of its proteolytic activity, PSA could bind to sorne serine protease inhibitors.

PSA occurs predominantly in complex with a,-antichymotrypsin (PSA-ACT) in semm and several

other serine protease inhibitors (67). Others have shown that in seminal plasma, PSA binds and

19

inactivates another serine protease inhibitor cailed protein-C inhibitor (69.70).

Except serine protease inhibitors, other mechanisms may be involved in the regulatioa of

PSA function. One possible mechanism is that PSA rnay be secreted as a prozymogen and needs

to be processed enzymatically as some other proteolytic enzymes. It has been suggested that the

initial PSA polypeptide chain has 261 arnino acids and it undergoes several proteolytic changes to

become a functionally active mature molecule with 237 amino acids. Three cleavage sites have

been found to exist and cleavage at 2 of the sites may result in release of 2 small peptide chains

fiom the large molecule. One has 17 molecules and the other contains 7 amino acids. These changes

rnay facilitate the activation of PSA fiom zymogen to its active form. This activation process is

thought to take place inside the prostate gland, but the activating enzyme has not been identified.

One study suggested that hGK-1 protein may be the candidate enzyme because the proteolysis is

trypsin-like cleavage and hGK- 1 is supposed to have this activity.

There is a third cleavage site at the lysine-lysine bond between amino acid residues 145 and

146 of PSA aminoacid sequence. Cleavage at this intemal site may produce a two-chain form of

PSA (nicked form of PSA) which is enzymatically inactive. The protease enzyme respoiisible for

this cleavage has not been reponed. The free form of PSA in semm may have this structure, as it

is enzymatically inactive, and is unable to bind to a,-antichymotrypsin. Formation of free PSA is

suspected to occur before it enters the blood circulation because it would be highly unlikely for the

srnall amount of free and active PSA not to react with a,-antichymotrypsin while it is present in

serurn in a large excess arnount over the free PSA. Other cleavage sites on the PSA molecule may

also exist, but the consequence of the cleavages is not clear (66,67). Physiologically, inactivation

of PSA function can also be achieved through the binding of PSA to other proteins. It is believed

20

that once PSA gets into the blood circulation the native form binds quickiy to several senne

protease inhibitors, rnainly a ,-antichymotrypsin, and becomes enzymatically inactive.

2-7. Molecutar Forms of PSA

PSA in semen has two major molecular forms. one of which is the intact polypeptide

chain called the native forrn. n i e other cailed Mcked form. has a cleavage at aminoacids Lys 145

and Lys 146 (85). As mentioned earlier. a small amount of PSA in semen also binds to PCI.

Several molecular forms of PSA exkt in serurn. including free PSA. a 1 -antichymotrypsin-bound

PSA (PSA-ACT complex). and a,-macroglobulin bound PSA (70.85). The PSA-ACT complex

is the major molecular form of PSA in serurn. Al1 these forms are believed to be enzymatically

inactive. Most current PSA assays c m detect free PSA and PSA-ACT. PSA assays which only

detect fiee PSA have been developed and the ratio of free versus total PSA appears to provide

a slightly better sensitivity and specificity for diagnosis of prostate cancer (86). The new

nomenclature of PSA molecular forms and kalliheins is shown in Table 2-7- 1.

2-8. PSA In Non-Prostatic Tissues

2-8- 1. Pen'urethral Glands

The periurethral gland is the first fernale tissue which was suggested to be able to

produce PSA. PSA production by this gland is based on the findings of immunohistochemical

studies which demonstrate positive stainings for PSA in both male and female periurethral glands

(87.88.89.90) and of PSA in male urine of prostate cancer patients who have had their prostate

removed (87). Further histological studies indicate that the tissue structure of female periurethral

Table 2-7-1 Nomencluture of hK proteins Formal Name Common Name Description hK1 PRK Pancreaticl$enal kallikrein h K2 hC K. I hK3 . PSA

Total PSA t-PSA

Free PSA f- PSA

PSA complexes PSA-ACT

PSA-AT

PSA-IT

Clandular kalfikrein Prostate-specific antigen All imrnunodetectable forms in serum. primarily f-PSA and PSA-ACT Noncomplexed B A ; may be proteoiyticaliy active or inactive in seminal muid. oniy inactive in senrm PSA covatently bound to a, -antichymouypsin inhibitor; synonymous with PSA ~0mplex; major immunodetectable form in serum PSA covaiently linked and encapsulated by a,- macroglobulin; not detected in irnrnunoassays; synonymous with occult PSA PSA covaiently bound to protein-C inhibitor: minor component in serninal fluid; not detected in serum PSA covaiently bound to a, .antitrypsin; trace component in serum PSA covalently bound to inter-alpha trypsin inhibitor: trace cornponent in serum

22

gland is similar to that of male prostate before puberty. but the gland remains underdeveloped

throughout the whole life due to the lack of androgenic stimulation (91 .) Therefore. the female

periurethral gland c m be considered the 'fernale prostate' (14,92). An intcresting animal

rxperirnent suggests that endoderm-derived urethral epithelium from newbom female mice would

develop into a prostate-like tissue producing a PSA-like protein if the cells were allowed to grow

in the body of male mice (93).

2-8-2. Breast

2-8-2- 1. Breasr Cancer Tissue

In 1993. PSA imrnunoreactivity in breast cancer cytosolic extracts. prepared for

measurement of steroid hormone receptors. was accidentally discovered in our laboratory while

we were evaluating the analytical specificity of a new ultrasensitive PSA assay developed for

monitoring prostate cancer patients afier radical prostatectomy. PSA irnmunoreactivity in non-

prostatic tissues was thought to be due to the cross reaction of polyclonal antibodies to non-PSA

proteins. However. this possibility was excluded when similar irnmunoreactivity was seen with

double-monoclonal PSA assays (94). Moreover. the possible existence of PSA in breast cancer

was suggested by the molecular weight of PSA immunoreactive species in the cytosols. The

molecular weight determined by gel electrophoresis and gel filtration was identical to that of

PSA in seminal plasma or to the free fom of PSA in male serum. Finally. PSA mRNA was

detected in breast cancer tissues which were positive for PSA protein measured by imrnunoassay

(95). In contrast. tissues negative for PSA protein were also negative for PSA mRNA.

Since PSA concentrations in breast cancer cytosols are relatively low compared to the

levels in male prostate, the number of sarnples which have detectable PSA may Vary depending

upon the sensitivity of the PSA assay used. Using a PSA assay which has a detection limit of

10 n_d . we found about 30% of breast cancer cytosol samples having PSA levels at or higher

than 0.03 ng of PSA per mg of total protein.

2-8-2-2. Breast Cancer Cells

Several breast cancer ce11 lines which do not produce PSA in the absence of steroid

hormones (baseline) are able to produce PSA afier the cells are incubated with androgens.

progestins. glucocorticoids. and mineralocorticoids (96). The stimulation of PSA production by

these steroid hormones is tirne- and dose-dependent. The molecular weight of PSA produced by

these breast cancer cells is identicai to seminal PSA or to PSA from the prostate cancer ce11 line

LNCaP. PSA mRNA is readily identified in these cells lines d e r the stimulation by steroids

(95) . Estrogens do not have the ability to stimulate the cells to produce PSA. but can impair

the stimulating effect of androgens (96). The production of PSA is observed only in breast

cancer ce11 lines which have steroid hormone receptors (MCF-7 or T47-D). but for the cells

which do not possess the receptors (BT-20) no PSA is produced afier similar stimulation (96).

However, for as yet unknown reasons. many ce11 lines which are positive for receptors are not

able to produce PSA after steroid hormone stimulation.

2-8-2-3. Other Breast Tissues

PSA has also k e n found in both normal breast tissues (specimens fiom cosrnetic breast

reduction surgery) and breast tissues from benign breast disease (BBD). The percentage of

24

women who have detectable levels of PSA in their breast tissues is siçnificantly higher in

patients with BBD (60%) than in normal women (30%). but the positivity rates are not different

between breast cancer patients and normal women (97). The molecular weight of PSA in these

specimens is similar to the one found in seminal plasma (33 KDa).

2-8-2-4. Breasr Fluids

PSA is detectable in dl tested breast fluids. including rnilk (9). cyst fluid (98). and nipple

aspirate fluid (99). PSA concentrations in these fluids Vary widely from undetectable to 5.000

rn@. PSA levels in mik decrease post-delivery with t h e but are not affected by mother's age

or baby's gender. Free PSA is the major molecular fonn of PSA in milk. PSA-ACT complex

is also present in milk. but accounts for a5%. In cyst fluid. PSA levels vary between <0.01

and 82 pg/L. and free PSA and PSA-ACT complex are present in roughly equal proportions

(98). Even wider variations of PSA concentrations are seen in nipple aspirate fluid (0-5,000

m a ) : the rnolecular forms of PSA in this fluid are free PSA and PSA-ACT. in roughly equal

proportions.

2-8-2-5. PSA Regulation and Funcrion in Breasr Tissue

Based on the resuits of ce11 culture experiments. we concluded that the PSA gene

expression in breast cancer cells is up-regulated by androgens. glucocorticoids and progestins.

The regulation requires androgen, glucocorticoid and progesterone receptors. respectively

(unpublished data). Estrogen not only does not up-regulate but can interfere with the effect of

androgen. The exact mechanism is unknown. Since al1 these observations are based on in vitro

25

experiments using breast cancer ceil lines, it remains unknown as to which steroid hormones are

involved in the PSA gene expression and PSA protein production in normal breast cells and

during pregnancy. However. there is indirect evidence for PSA production in normal breast tissue

induced by progesterone. We have andyzed normal breast tissues for PSA from I l healthy

women who underwent cosmetic breast reduction surgery. Of these wornen. none but one had

very high levels of PSA in the tissue: that woman was also the only one who was on oral

contraceptives which contained high levels of progestin (8). Clinical studies have s h o w that

the presence of PSA in breast cancer cytosols is associated with the presence of estrogen and

progesterone receptors. Although we did not examine the relationship between androgen

receptors and PSA status in the clinical study. it is known that androgen receptors are present

in breast cancer cytosols (99) and are positively correlated with estrogen and progesterone

receptors ( 100).

Interestingly. PSA is only associated ~4th ER or PR by their dichotomous status. Le.. the

presence or absence of these proteins. No correlation could be seen between the levels of PSA

and ER or PR. This phenornenon may suggest that the associations between PSA and ER or PR

are indirect and the key luikage between PSA and the receptors is the ligands which bind to the

receptors and activate PSA gene transcription. Therefore. the amounts of receptors present in the

cells at and above certain levels are not the key factors for PSA gene transcription. Another

interesting observation is that ER and PR positivity in breast cancer increase with age of the

patients. but PSA positivity decreases slowly with age (6). The opposite direction of correlation

between age and PSA or receptors may indicate that PSA and the receptors are both associated

with the steroid hormones released from the ovary because the production of ovarïan hormones

26

is an age-dependent phenornenon.

The enzyrnatic activity and physiological role of PSA in breast tissue remains

undetermined. In seminal plasma PSA is found to be able to proteolyze IGFBP-3 (78). An

inverse correlation between PSA and IGFBP-3 is also seen in the senun of prostate cancer

patients (76). Based on these information. we have examined PSA levels in breast cancer

cytosols in association with IGFBP-3. However. no correlation or association was seen between

PSA and IGFBP-3 in our preliminary study (101). The relationship between PSA with other

members of IGF farnily. IGF-1. IGF-II. and IGFBP-1. was also analyzed. but no association was

found.

The fact that PSA is present in milk and amniotic fluid (mentioned later) and that PSA

levels in amniotic fluid change with gestational age suggests that PSA may play a role in fetal

andor newbom development.

The relationship between PSA and IGFBP-3 indicates a possible regulatory pathway for

IGFs. In an in vitro study. it was found that IGF-1 could activate androgen receptor (AR)

resulting in AR-mediated gene transcription (102). PSA gene transcription is one of those genes

up-regulated by androgen via AR (59). Taken together. these data suggest that it may be likely

that there is a regulatory loop on IGF-I through PSA and IGF binding proteins. IGF-1 increases

the production of PSA through androgen receptor. and PSA degrades IGFBP-3 causing release

of IGF-1 from its binding protein. It is believed that IGF binding proteins control the

bioavailability of IGFs through regulation of their transportation and binding to IGF receptors.

Whether the regulation results in enhanced or suppressed action of IGF depends on the type of

tissues involved (79). Because of this tissue-specific regulation, it is difficult to speculate

whether the loop is up- or dom-regulating

Other interesting findings about IGF

the action of IGFs.

include estrogen regulated expression of IGFs and

their binding proteins in breast tissue (103) and involvement of the IGF farnily in ce11

transformation and apoptosis (86.1 04). PS A. through IGFBP-3 proteolysis could be indirectly

involved in such phenomena.

2-8-2-6. Clinical Applicatiom

The presence of PSA in breast cancer cells and its association with steroid hormone

recepton indicated potentiai clinical utilities of this protein in breast cancer. Preliminary clinical

studies did suggest the possibility of PSA king a favorable indicator of prognosis. It was found

that PSA presence was associated with early stage of the disease and small size of the tumor.

and that patients with PSA positive breast cancer could have longer survival compared to those

with PSA negative cancer. Furthemore. the reduced risk for relapse by PSA was sustained afier

other major clinical and pathological factors were controlled in the analysis (12). Survival

anaiysis in subgroups of patients indicated that PSA status may help to further identify patients

with favourable prognosis among those who were node positive or estrogen receptor negative.

One recent study on breast nipple aspirate fluid found that PSA levels in this fluid were

significantly lower in women with high risk of breast cancer compared to women with low risk

of the disease (99). However. dl these observations are based on small number of patients.

Large scale and well-designed studies are needed to confirm these findings.

It is still not clear what is the biological rnechanism of PSA being a good prognostic

indicator for breast cancer. One possible interpretation is that PSA can only be produced by well

28

differentiated cancer cells. Being a product of steroid hormone regdation through their receptors.

the presence of PSA indicates the presence of the regulatory system. which is a sign of good

differentiation of cancer cells. Moreover, PSA in breast cancer cells indicates not only the

existence but d so the fûnctional status of the system. Based on the observation of 30% ER

positive patients who do not respond to endocrine treatrnent (105). it is thought that these

receptors are mutant products which are not functional (106). Therefore. the physical existence

of these receptors does not necessarily indicate their functional status. From this point of view.

PSA could be a more direct and useful marker for this system since it is placed downstrearn

from the fünction of the ER and PR.

The other possibility is the regulatory balance between androgens and estrogens in the

target cell. Since PSA is an androgen up-replated protein in the prostate. the impact of androgen

on breast cancer needs to be considered. The androgen receptor is present in breast cancer and

is correlated with ER and PR. Ce11 culture experiments shows that androgen could inhibit the

proliferation of breast cancer cells (107) and counteract the effect of estrogen ( 108). In addition.

androgen has been used effectively in the treatrnent of some breast cancer patients (109). Our

ceII culture study demonstrates that PSA production in breast cancer cells c m be induced by

androgens (96). Based on this evidence. the presence of PSA in breast ceils may suggest an

androgenic suppression of estrogen effect which is believed to play an essential role in the

development and progression of breast cancer ( 1 10.1 1 1.1 12).

2-8-3. Amniotic Fluid

In addition to milk and other breast fluids. amniotic fluid is another female fluid

29

containing detectable levels of PSA ( 1 1). Both free PSA and PSA-ACT complex are found in

the fluid but the former is the major component. The median concentration of PSA in amniotic

fluid increases between gestationai weeks 14 and 21 and then seerns to level off or decrease until

parturition. The presence and change of PSA concentration in arnniotic fluid seem to have an

impact on senun PSA in pregnant women. S e m PSA is significantly higher in pregnant women

than in non-pregnant women. and changes in serum PSA of pregnant wornen match well with

the change of PSA in amniotic fluid (1 13). Extrernely low or high levels of PSA in arnniotic

fluid at a specific gestational age rnay be linked to fetal abnormalities such as trisomy 21 ( 1 l 3 ) .

ïhe source and physiological role of PSA in amniotic fluid remain unknown. but the levels are

not related to fetai gender. The relationship between PSA and IGFBP-3 or other members of IGF

fmily in amniotic fluid has not been examined.

2-8-4. Fernale Semm

Using conventional PSA assays with a detection limit of 0.1 pg/L or higher, it is possible

to find detectable levels of PSA in less than 10% of fernale sera ( 1 14- 1 18). Because of the lack

of interest. the real identity of PSA irnmunoreactivity in female senun has never been studied.

It was assumed that this irnrnunoreactivity was an artifact which was caused either by cross

reaction of polyclonal anti-PSA antibodies to non-specific proteins or by the presence of

heterophilic antibodies or anti-PSA antibodies.

Based on our studies, the percentage of female sera which are positive for PSA increased

when we use an increasingly sensitive PSA method (detection lirnits between 0.01 and 0.001

pg/L). This suggests that trace amounts of PSA are present in the majority of female sera (1 19).

30

The molecular form of PSA in normal fernale senim as determined by gel chromatography is

similar to the one in male s e m i.e. rnainly PSA-ACT complex (1 19).

In the prostate, androgen up-regulates the expression of PSA and the regulation of PSA

production by androgen is also observed in the cultured breast cancer cells. Based on this. it is

interesting to know if PSA is elevated in the s e m of women who have high levels of androgen.

Our recent study shows that serum PSA is increased significantly in hirnite women and that PSA

levels are positively correlated with the levels of testosterone and 3 a -androstanediol glucuronide

(unpublished data). A potential clinical utility of serum PSA in hirsutism is suggested.

The source of PSA in female s e m remains undeterrnined. but several possibilities exist

based on preliminary observations. Serurn PSA in pregnant women could come from amniotic

fluid since PSA levels increases during prepancy and change with the gestational age. Amniotic

fluid has 20-40 times higher PSA than serum of pregnant women. Obviously. this source should

be taken into account only when women are pregnant. It is reasonable to speculate that breast

tissue is an important source of serum PSA in women. However. when we measured PSA in

matched pre- and pst-surgical sera of breast cancer patients. we found no change in s e m PSA

atier surgery and no correlation in PSA levels between breast tissue and serum ( 120). However.

these observations may have the following limitations. First. the PSA assay we used in the study

was still not sensitive enough to detect minute changes of PSA in female senim. Secondly. serum

PSA can come from both normal and cancerous breast tissues. Thus. the presence and removal

of the tumor tissue may not necessarily affect the levels of PSA in s e m . However. our most

recent (unpublished) data does suggest that removal of breast cancer could result in decrease of

semm PSA when a more sensitive PSA assay is used.

3 1

It is possible that PSA in female semm arises from other sources such as ovary.

endometri-. or lung. but the frequency of these tissues h a k g detectable PSA is relatively low.

Although PSA is also produced by the female periurethral gland. the arnount that could enter

blood circulation is thought to be minimal. Thus. PSA from this gland is not expected to be a

major source of PSA in female s e m .

Recently, we observed changes in the molecular t o m of PSA in female serum in

association with disease statu. For hedthy women and post-operative patients with breast cancer.

the major molecular form of PSA in senun is PSA-ACT cornplex. whereas for the pre-operative

women with breast cancer. the major form is free PSA (12 1). However. this observation is

preliminary . The explanation for this phenornenon and its potential clinical utility remain

unknown.

2-8-5. Ovary, Endometrium and Other Tissues

PSA is detectable in some ovarian cancer cytosols. but the frequency of detection in

primary ovarian cancer is much lower (3%) compared to the frequency of detection in breast

cancer (30%) (122). Ovarian cancer metastatic to breast also has detectable PSA provided that

the primary tumor produces it. We have measured such specimens from 3 women. and 2 of them

were found to have detectable PSA. Of the 3 patients. clinical information was available for one

PSA negative patient and one PSA positive patient. Interestingly. the patient with PSA positive

cancer. which was negative for steroid hormone receptors. responded well to the endocrine

treatment and survived many years afier distant metastatsis. However. the patients with PSA

nrgative cancer was positive for the receptors and she died shortly afier the relapse (122).

32

Although this observation is only based on two patients. it is consistent with our îïnding of PSA

being a favourable prognostic indicator in primary breast cancer ( 1 2).

Clements and Mukhtar reported that PSA mRNA is detectable in endometrial tissues

dong with two other members of the Mlikrein gene family (33). Other tissues which have been

reported to have PSA immunoreactivity by imrnunohistochemistry include kidne y. pituitary (44).

normal axillary and perennial apocrine sweat glands. and some apocrine foci in fibrocystic breast

tissues (24). However. most of the immunohistochernical findings are believed to be due to the

cross-reaction of polyclonal antibody to non-specific antigens because the results were not

reproducible when a monoclonal antibody was used (24).

2-8-6. Other Cancers

Some prirnary lung cancer tissues fiom both males and females contain low but detectable

amounts of PSA (123.7). Other cancer tissues which are found to have trace amounts of

detectable PSA in our study include kidney, adrenai. and colon (7). It has also been reported that

serurn PSA levels were increased in a woman with renat ceil carcinoma (124). Van Krieken

found PSA in salivary gland carcinomas (4) and Papoai et al. reported carcinomas in sweat

glands and breast apocrine glands having false-positive staining for PSA (24).

2-9. Steroid Hormone Receptors

2-9- 1. Mechanisms of Gene Replation

It has been known for several years that steroid hormones function ihrough the action of

specific receptor proteins. In target tissues, receptors are activated by hormonal ligands and

33

thereafler modulate the expression of a network of specific target genes. Receptors for steroids,

glucocorticoid. androgen, estrogen, progesterone, and mineraiocorticoid, and the sterol, 1,25-

dihydroxyvitamin D, have been identified and characterized by biochemical and molecular

biologicai methods (125). Other non-steroid, seemingly unrelated molecules such as Thyroxine (T,)

and retinoic acid (RA) have been found to have receptor hormones and appear to act through sirnilar

mechanisms as steroid receptors ( 126).

Even before the steroid receptors were cloned, it was clear that they were organized in

funcrional domains. Proteolytic cleavage analysis first revealed receptor fragments in which DNA-

binding activity was separateci &om steroid binding. Once the receptors were cloned. these domains

were better dehed. Functionaliy, the steroid receptors interact with inhibitory proteins such as heat

shock proteins ( Figure 1-I), bind to ligand, dimenze. bind to specific sequence of DNA with high

affinity, and activate transcription.

A number of structural features are sirnilar in al1 members of the steroid receptor gene

family, suggesting that they are considered rightly as members of a class of regdatory proteins.

These features include a structurally separate hormone binding site that is a fraction of the total

receptor polypeptide chain, the presence of a high-afinity DNA binding site distinct fiom the

hormone site, a tendency to aggregate at low ionic strength to form either dimers or tetramers of

the subunits, and enhanced affinity for the ceil nucleus in the presence of bound hormone. These

four characteristics can be observed in both cmde extracts and punfied preparations and hence are

probably characteristic of the proteins in situ.

On the basis of cDNA cloning and the cornparison of the deduced amino acid sequence of

different hormone receptors, al1 receptors analyzed thus far are structured in a sirnilar way: they

34

exhibit a variable N-terminal region, a short and well conserved cysteine-rich centre domain, and

a relatively well conserved C-terminal half ( Figure 2-9-1). These three main domains are separated

by less well-defined regions, and are probably subdivided into shorter structural motifs. For

instance, the central domain exhibits an array of cysteine residues compatible with the formation

of two- so-called zinc fingers (127), each of them with a zinc atom tetrahedrically coordinated to

four cysteines. This type of structure, as well as in vitro binding studies and functional evidence,

make it highly probable that the central domain is responsible for the DNA-binding activity of the

receptors. The cysteine residues implicated in metal binding have recently been established by

mutational analysis. Interestingly , each one of the hypothetical zinc figures is encoded by a

separate exon of the receptor gene, but one figure alone appears to be unable to bind to DNA or to

activate transcription ( 1 27,128).

Irnmediately adjacent to the C-terminal zinc figure is a short sequence similar to those found

in SV40 T antigen that appears to be in part responsible for intranuclear locaiization of the

glucocorticoid receptors. Another hormone dependent nuclear translocation signai rnay be located

in the carboxy terminal haif overlapping the steroid binding domain. In this region also reside the

amino acid residues thought to be responsible for the interaction with the 90 kDa heat shock

protein. Since binding of the hormone appears to promote dissociation fiom hsp90, the effect of

hormone binding to nuclear translocation rnay be indirect.

The C-terminal; seems to be complex structurally and functionally, since in addition to its

role in hormone binding and nuclear translocation, it may also contain tram-activation and

dimerization tùnctions. Some of the amino acid residues involved in binding of natural or synthetic

ligands have been identified by mutational and chernicd analysis (1 29- 130).

N - Modulator IONAB Ligand ] - c FUNCTldN

- - - iransacirvaiion

CONSENSUS RESPONSlVE ELEMENTS FOR NUCLEAR RECEPTORS

11 13 15 1 2 3 4 5 6 10 1 2 1 4

1. GRE (+ ) GGTACAnnnTGmCT 2. PR€ II

3. ARE 8 @

4. MRE a *

5. ER€ AGGTCAnnnTGACCT 6. EcRE AGGGTTnnnTGCACT

Figure 2-94. General stnictual and Functional organization of nuclear receptors and consensus responsive elements.

36

The N-terminal domain is less well charactenzed and may have a modulatory effect on

tram-activation. Most of the receptor antibodies descnbed thus far are directed against this region

of the protein. in the case of glucocorticoid receptor, the junction between this and the central

domain appears to be very sensitive to proteolytic attack, as a 40 kDa form of the receptor

containing only the central and C-terminal; domains is ofien found in liver cposol as a product of

receptor deregulation (129-13 1). This finding underlines the structural array of the receptor

molecules, with protease-sensitive regions comecting more resistant domains.

The function of the steroid hormone receptors concem the regulation of the expression of

particular categories of genes. this regulation is exerted by interaction with specific DNA sequences

and with other proteins involved in gene expression control. The DNA sequence required to the

specific binding of nuciear receptors (hormone response element; HREs) are generally situated

upstream of the transcription site and present two short (5 to 6 base pairs) segments of a very

similar or identical sequence (in opposite direction for most nuclear receptors), interspersed or not

by a short non-conserved segment. A comparison of the available sequence information yields a

15-mer consensus sequence for the glucocorticoid response element (GRE) and a 13-mer consensus

sequence for the estrogen response element (ERE) (Figure 2-9-1). WhiIe no clear consensus

sequence has yet been defineci for the progesterone and androgen response elements, a GRE 15-mer

is able to mediate induction Dy progesterone, androgen, and mineralocorticoids (1 32).

The specificity of the natural elements, such as that found in the MMTV-LTR. for a

particular hormone receptor is not tightly restricted, and these elements are able to mediate

induction by several hormones, including glucoconicoids, progestins, androgens, and

mineralocorticoids. Detailed studies of the contacts of the glucocorticoid and the progesterone

37

receptors with a particular HRE show very sirnilar patterns for the recepton. These differences are

probably functionally relevant, as mutation of individuai sequences has a differentiai infiuence on

the response to glucocorticoids and progestins. Thus, it seems that although identical sequences can

mediate response to different hormones, the requirements for optimal interaction are not identical

for the different hormone receptors (132).

A consensus sequence has also been postulated for the thyroid hormone and retinoic acid

responsive elements containing half palindromes sirnilar to those of the ERE but with different

spacing (TRE,RRE, Figure 2-9-1). It has been s h o w that the thyroid hormone receptor cm bind

to the ERE, but that this binding does not result in transcriptional activation (126,132).

The ultimate hnction of receptors is to modulate specific effects at the transcriptional level.

This fùnction has been analyzed extensively in vivo using receptor-deficient ce11 lines transfected

with expression vectors contain cDNAs encoding receptors and reporter vector containing specific

response elements linked to a gene whose product is readily assayable such as chiorarnphenicol

acetyltransferase (CAT) or luciferase. Using these analyses, the regions of the receptor responsible

or gene activation can be defined. for exarnple, mutations that affect conserved cysteine in the DNA

binding domain result in loss of glucocorticoid receptor trans-activation fùnction. In the rat

glucocorticoid receptor an 86-amino acid region including the DNA binding region is sufficient

to mediate gene activation in stably and transiently transfected cells. However in some cases the

activation capacity of mutant receptors was low compared with wild-type receptors. A mutant of

the human estrogen receptor that lacked the hormone binding domain exhibited only 5% of activity.

For several receptors, deletion of amino acids in the N-terminus results in reduced activation

capacity. Thus multiple regions of receptors, both N-terminal and C-teminal, are involved in the

38

transcriptional activation process (129).

A series of recent reports suggested that transcription efficiency of steroid hormone

receptors, once positioned on the right promoter, require functional interaction among the receptor

molecules as well as interactions with other essential transcriptional factors. Functional

cooperativity between two binding sites for the glucocorticoid receptor was first demonstrated in

the regdatory region of the tyrosine amino transferase gene. However, in the case of h4MTV.

binding of glucocorticoid and progesterone receptors to linear DNA segments containing severai

receptor binding sites, does not occur in a cooperative fashion. Additional experiments with

artificial combinations of GREs and binding sites for transcriptional factors showed that not only

NF- 1, but a whole battery of other well characterized factors, are able to act in combination with

the hormone receptors for the activation of an adjacent promoter. When the binding site for receptor

is not very strong, the influence of binding sites to other transcription factors is synergetic, whereas

with a strong GRE the effect are additive, in fact another GRE is as effective as a binding site for

NF- l or SP- 1, a CACCC sequence, or an octanucleotide motif (129,134).

Beside transcription factors, a participation of chromatin structural features in the

mechanism of hormone gene regulation has been repeatedly suggested. In particular, the

observation that following hormonal treatment DNAase I hypersensitive regions appear in the

vicinity of the regulated promoters has suggested that the hormones may act by altering the

chromatin structure of target genes. Recently, it has been postulated that phosphorylation of steroid

hormone receptor is another potential pathway for control of hormone action, particularly if

phosp horylation is hormone dependent. For example, hormone treatments increase the

phosphorylation of the progesterone and glucocorticoid receptors and it rnay regulate the DNA-

39

binding and/or transcriptionai activation fùnction of steroid receptors. However, the fùnctional

significance of steroid receptor phosphorylation has yet to be determined (1 3 5).

2-9-2. Clinical Applicafion of Estrogen and Progesterune Recep fors in Breast C'amer

ER and PR belong to a superfamily of nuclear hormone receptors that hnction as

transcription factors when they are bound to their respective ligand. Estrogens and progestins are

considered the pnmary regulators of growth and differentiation of normal mammary tissue.

Moreover, estrogens are thought to play an important role in the development and progression of

breast cancer. This concept is based on severai lines of evidence. First , epidemiologic evidence

indicates a protective effect concomitant with ovariectomy and an increased risk of developing

breast cancer arnong women treated with the estrogen diethylstilbestrol. In addition, the mitogenic

effects of estrogens on various breast cancer ce11 lines grown in culture, and in vivo in

ovariectomized nude mice have been well documented in the literature. Reports have also

demonstrated that the use of oral contraceptives starting at a younger age significantly increase the

risk of breast cancer. This cohort of wornen had a significant relationship between survival and

exposure to oral contraceptives at a young age. These results suggest that estrogen does indeed

participate in tumor growth promotion and, later, in progression of the disease ( 13 6- 1 3 8).

Estrogens and progestins exert their cellular effects through the binding and activation of

specific nuclear receptors, the ER and the progesterone receptor (PR). When evaluating the role of

these receptors in breast cancer, two issues are important. First, the presence of ER and PR predicts

improved disease-free and overall suMval of breast cancer patients. Second, the presence of these

receptors in tumors predicts the likelihood of benefit from endocrine therapy. Since the original

40

publication suggesting the value of ER as a prognostic biomarker of delayed recurrence in primary

breast cancer, ER and PR assays have become standard practice in the management of breast

cancer. As a single factor, PR has been show to be a more powerful independent predictor of

tmor recurrence that ER in certain subset of patients, such as those with stage II breast cancer. It

is now known that the correlation between ER and PR status and patient outcome is a general

reflection of the intrinsic biological behaviour of receptor-positive tumors. ER-and PR-positive

tumors are more likely to be highly differentiated, to be diploid, and to exhibit lower proliferation

rates than are receptor-negative tumors. The value of steroid receptors as single, independent

prognostic biomarkers is limited, however, and they are most effectively used in combination with

their prognostic biomarkers ( 1 39,140).

Potentially of great importance is the ability of ER and PR to predict response to endocrine

therapy. The ER assay is most usefùl if the tumor is ER negative; these patients seldom respond to

endocrine therapy. The knowledge or ER and PR status together improves the ability to predict

endocrine responsiveness. This findings is based on a hypothesis first published by Horwitz and

associates, suggesting that PR rnight be a better marker than ER for an intact endocrine response

pathway because PR is a product of estrogen action. Indeed, this appears to be the case. Response

rates higher than 70% are seen in metastatic ER positive. PR positive tumors. In addition a

quantitative correlation appears to exist between receptor content and response. with some studies

finding a quantitative correlation only with ER levels and other finding correlations with both ER

and PR levels (14 1).

41

2-9-3. Steroid Hormone Receptors and Cancer

The biological effécts of sex steroid hormones on growth and development of reproductive

tissues and on the growth and progression of endocrine-dependent neoplasia ( 142- 149) are rnediated

by intracellular receptors that are members of a gene family of ligand-dependent transcription

factors (149- 152). In addition to sex steroid hormones and glucocorticoids, the gene family also

includes receptors for thyroid hormones (TR), vitamin D, (VDR), retinoic acid (FUR), and a

number of orphan receptors for which a ligand has not yet been identified. Intracellular receptors

are modular proteins composed of separate domains for ligand binding, DNA binding and

transcriptional enhancement. For the steroid hormone group of receptors, the C-terminal ligand

binding domain (LBD) also harbours sequences for a second ligand-dependent transcriptional

activation domain, dimerization, nuclear localization and binding of heat shock proteins (1 49- 142)

There are similarities in the generai rnechanism by which ligands activate steroid hormone

receptors. In the absence of hormone, receptors from an inactive oligometric complex with heat

shock protein 90 (hsp 90), hsp70, immunophilins and possibly other proteins of unknown identity

(1 53- 155). In response to binding hormone, receptors dissociate from the oligornetric complex and

acquire the ability to dimerize (1 56- 159), and to bind to hormone response elernents (HRE) that are

usually located in the regdatory region of steroid responsive genes ( 160). Receptor association with

HREs leads to an increase or decrease in transcnption (161) by mechanisms that are not fully

understood. A central event in the activation process appears to be a ligand induced conformational

change in receptor structure (162).

It has been demonstrated that steroid hormone receptor presence is a necessary condition

for hormone responsiveness in a target tissue. This crucial role has led numerous investigators to

42

examine malignant tissues that are known to respond to hormonal manipulations and to correlate

the clinical responsiveness to the presence and absence of a steroid receptor. Their presence has

been well documented in some human breast cancers (163) and acute leukaernia (164) and

information is accumulating in other hormone dependent cancers, such as prostate (165) and

endometrial cancer ( 166). In addition, steroid hormone receptors have recently been identified in

some rnalignant melanoma sarnples, lung tumors and colon cancer (1 6% 169). The lack of receptor

almost always predicts a lack of response to hormone therapy, while many but not dl tumors that

contain measuriible receptors are responsive to hormonal manipulations at least in the case of breast

cancer and leukaernia.

Function of steroid hormone receptors could be assessed by different biochemical and

molecular biological methodologies. One possible approach is to look at the expression of steroid

hormone-inducible proteins in target tissues using receptor inhibitors. Several synthetic ligands

(both steroid and non-steroidal) for estrogen (ER), progesterone (PR), androgen (AR),

glucocorticoid (GR), rnineralocorticoid (MR), and retinoic acid (RAR) have been developed which

compete for binding with natural hormones and are capable of inhibiting receptor activity The

antagonist forms a complex with the receptor and then interferes with the normal function of a

ligand-bound receptor by not translocating to the nucleus, not binding to the appropriate DNA

sequences with high affinity, or not aflfecting transcription rates. These components have also been

proven to be valuable for dissecting normal receptor activation mechanisms ( 170- 174).

Breast cancer is the second most fiequent cause of cancer death arnong American women,

accounting for 18% of al1 cancer deaths among women and trailing only lung cancer (1 75). AS

mentioned above, PSA is a good prognostic indicator in breast cancer patients and associate with

43

the presence of steroid hormone receptors ( 6 ) and this may be used as a new approach to design

therapeutic agents for treatrnent of endocrine dependent breast and other hormone dependent

cancers. Approximately 30% of aü breast cancer patients respond to endocrine therapy. The overail

response rate to antiestrogens increases to 70% or more in patients whose tumor express receptors

for both estrogen and progesterone (1 75-1 78). To define the mechanism of action of endocrine

therapies and to develop and screen new agents and therapies we require models that exhibit an

endocrine response profile comparable to that found in breast cancers.

In this regard, breast cancer ce11 lines that grow in-vitro represent one of the most widely

used experimental models of breast cancer. For many studies, these models provide the only means

to address a specific hypothesis. Therefore, studying the steroid hormone regdatory pathway of

PSA expression in non-prostatic tumors, particularly breast cancer, should open a new window of

predicting the prognosis of breast cancer patients and possibly the response of breast cancer to

endocrine treatment.

CHAPTER 3. Hypothesis

Prostate specific antigen (PSA) is a protein selectively expressed by prostatic tissue. The

PSA gene contains in its prornoter region an androgen response eiement. PSA gene expression in

the prostate gland as well as in prostatic carcinoma ce11 lines is regulated by the androgen receptor

and androgenic steroids produced by the testes. The epitheiial cells of the prostate, which produce

high arnounts of prostate specific antigen, are rich in androgen receptors. It is now well established

that the PSA gene in prostatic tissue is regulated by the steroid hormone receptor system. PSA was

recently discovered in breast tissue extracts, in breast tumors as well as in al1 breast secretions.

However, the physiological role of PSA in non-prostatic tissues and its mode of regulation are

unknown and not as yet been studied.

In this thesis, we have hypothesized that the expression of PSA in breast tissue is under the

regulation of the steroid hormone receptor system.

We also hypothesize that if the PSA gene is regulated by steroid hormones and their

receptors in non-prostatic tissues, then, PSA would be present and its concentration should be

changing in physiological situations where steroid hormone levels undergo cyclic changes, as it

happens during the menstmal cycle. We have further postulated that apart frorn the breast, which

is known to possess steroid hormone receptors, other tissues which are now known to contain

steroid hormone receptors e.g. lung tissue, may also have the ability to produce PSA.

Consistent with the Hypothesis of this thesis, we have formulated the following research

questions:

1. Could we design highly sensitive reverse transcription-polymerase chah reaction methods

45

to rnonitor the PSA mRNA with extremely high sensitivity and specificity?

2. Is PSA expression in non-prostatic tissues regulated by steroid hormones? 1s this regulation

dose dependent? Are there major differences in the ability of various classes of steroid

hormones to regulate the PSA gene? Are there steroid hormones which act to uprepulate

andor downregulate the PSA gene?

3 . If the PSA gene is indeed regulated by steroid hormones and their receptors in tissue culture

systems, is the same regulation occumng in-vivo? Could we establish cyclic changes of

PSA concentration in semm of women during the menstrual cycle? Couid we irnpiicate

cyclically changing steroid hormones during the menstrud cycle in the regulation of the

PSA gene?

4. 1s breast tissue the only female tissue that produces PSA or are there other tissue that

possess steroid hormone receptors that could also upregulate this gene? For example, is the

PSA gene upregulated in lung tissue, which is known to contain steroid hormone receptors?

Would this gene be upregulated in patients who receive exogenously administered steroid

hormones?

CHAPTER 4. Objectives

The objectives of my investigation are listed below.

1. In order to examine in detail and with high sensitivity, the expression of prostate specific

antigen in non-prostatic tissues under non-stimulatory and stimulatory conditions by various

agonists, I would need to have appropriate tools. A highly sensitive, irnrnuno£luoromet~c

procedure for measuring PSA protein has already been developed in Our laboratory. This

method has a detection limit of 1 ngL and is the rnost sensitive assay presently available

for PSA My first objective was to develop a highly reliable and extremely sensitive reverse

transcription polymerase cham reaction method to monitor PSA mRNA levels. This method

would be invaluable in monitoring PSA expression at the mRNA level in ce11 lines as well

as in clinical material to be studied.

7 -- Among an extensive collection of ce11 lines of which some possess steroid hormone

receptors while others are receptor negative, my objective is to isolate ce11 lines in which

the PSA gene could be reguiated by exogenously administered hormones. This particular

system will be invaluable to study the mechanisms of PSA regulation. Once the ce11 lines

are identified, rny objective is to examine if they could regulate the PSA gene under

stimulation by androgenic, progesterogenic, estrogenic, glucocorticoid and

mineralocorticoid agonists. In parallel experirnents, another objective is to identie which

of the agonist actions could be blocked by molecules that antagonize these particular

steroids. The results fiom these studies will ailow me to delineate which receptors are

involved in PSA gene regdation.

3. Another objective is to examine if the results of the tissue culture system would be

applicable to clinical situations. 1 have thus exarnined if the PSA levels in serum of women

change during the menstrual cycle. Since many aeroid hormones are regulated through the

menstrual cycle and their concentration changes drarnatically during the luteal or follicular

phases. 1 was interested to establish if there is any cyclic change of PSA concentration in

serum of these women.

4. Since tissues other than the prostate and breast possess steroid hormone receptors, another

objective was to establish if such tissues (e-g. lung tissue) are also able to produce PSA

especially after stimulation by exogenously administered hormones.

CHAPTER 5. Materiais and Methods

5-1. Preparation of PSA cDNA

5- f - 1. Transformation Procedure

DHSa E. Coli competent cells were transfected with a recombinant pA75 plasmid (a gifl

from Dr. J. Trapman, University of Texas, M.D. Anderson Cancer, Houston, TX) using the heat-

shoc k transformation method. This plasmid contains the complete sequence of PS A cDNA

(179,180). In brief. DHSa competent cells were thawed on ice and mixed gently. 50pL of

competent cells were aliqoted into chilled microcentrifuge tubes. 1-3 PL plasmid pA75 was added

directiy to a tube containing competent cells, moving the pipette through the cells while dispensing.

Cells were incubated on ice for 30 minutes and then heat-shocked 20 seconds at 37 OC. Following

that, cells were placed on ice for 2 minutes. 0.95 rnL of LB media was added and shaken at 225 rpm

for 1 hour at 37 OC for expression. In pardel, transformation efficiency was controlled by adding

5 p L (0.5 ng) control pUC19 vector to one tube containing 50 pL of competent cells. After

tramfaion, the DH5 a cells were grown on LB-ampicillin media (10 g bacto-tryptone, 5 g bacto-

yeast extract, 10 g NaCl, pH 7.4, per liter, ampicillin 50 mg/mL). Cultures were incubated for 4

hours at 3 7 OC with shaking and then used to inoculate larger volumes of kesh media which were

incubated at 37 OC for 24 hours. Isolation and purification of the pA75 plasrnid was performed using

the Prornega Magic Maxiprep kit (Promega, Madison, WI, 53 1 1-5399). The procedure is described

briefly as follows. About 100-500 mL the cells were pelleted by centrifugation at 14,000 x g for 15

minutes. The ceils pellet was resuspended in 15 mL of ce11 resuspension solution (50 mM Tris-HCI,

pH 7.5, 10 mM EDTq 100 pg/mL RNase A) to disrupt the cells. 15 mL of ce11 lysis solution ( 0.2

49

M N a 0 6 1% SDS) were added and k e d gentiy, but thoroughly, by stimng or inverting. The ce11

suspension should becorne clear and viscous. Following few time stimng or inverting, 15 mL of

neutralization solution (2.55 M potassium acetate, pH 4.8) were added and immediately mixed by

inverting the centrifuge boale several times. Centrifiigation at 14,000 x g for 15 minutes at 4 OC

separates the cleared supernatant from the white precipitate which was removed carefblly. To

extract DNA, 0.6 volumes of isopropanol were used and mixed by inversion. Centrifugation at

14,000 x g for 15 minutes separates the DNA pellet which is resuspended in 2 mL of TE buffer (

50 mh4 Tris-HCI, pH 7.5, 1 rnM EDTA).

Plasmid purification was carried out by adding 10 mL of Magic Maxipreps DNA

purification resin to the DNA solution. The resin/DNA mix was transferred into the Maxicolumn

and washed with 13 rnL of column wash solution (200 rnM NaCL, 20 rnM Tris-HCI, pH 7.5, 5 mM

EDTA diuted 1: 1 with 95% ethanol ). Another 12 mL of column wash solution were added to the

Maxicolurnn and applied a vacuum to draw the solution through the Maxicolumn. To rinse the

resin, 5 mL of 80% ethanol were added to the Maxicolurnn by continuing to draw a vacuum for an

additional 10 minutes. The maxicolurnn was removed from the vacuum manifold and placed in the

provided reservoir (50 rnL screw cap tube). 1.5 mL of preheated (65-70 OC) water was applied to

the Maxicolumn and waited 1 minute. DNA was eluted by centrifugation the

~axko~urnn/Reservoir at 2,500 rpm (1,300 x g ) for 5 minutes. The concentration and purity of the

plasmid DNA was determined spectrophotometrically.

A plasmid containing the complete p-aain cDNA was obtained from Clontech Laboratones

Inc, Palo Alto, CA, 943034407.

50

5-2. Positive and Negative Control Ce11 Lines

The PSA-expressing human prostatic carcinoma ce11 lhe LNCaP and the non-PSA

expressing human breast carcinoma ce11 line BT-20 were used in control experiments. These ce11

lines were obtained fiom the American Type Culture CoIIection, Rockville, MD. Culture of ceII

lines was performed in flasks at 37 'C and 5% COz in RPMI 1640 medium supplemented with

bovine insulin (200 uni t f i ) glutamine (29 g/L) and 10% fetal calf serum. When the cells were

grown to near confiuency, they were washed with isotonic saline, detached by trypsin-EDTA

treatment and counted. 10' cells were kept frozen at -70 OC until RNA extraction was performed.

5-3. Preparation of Tumor Cytosol Extracts

Approximately 0.2 g of tissue £Yom each tumor was quick frozen and pulverized manually

with a hamrner to a fine powder at -80 OC. The cells were lysed for 30 minutes on ice with 1 mL of

lysis buffer (50 mrnoi/L Tris buffer, pH 8.0, containing 150 mmom NaCI, 5 rnmol/L EDTA, IO@

Nonidet NP-40 surfactant and I mmoVL phenylmethylsulfonyl Ruoride). The lysates were

centrifuged at 15,000 g at 4 OC for 30 minutes and the supematants (cytosolic fractions) were

assayed for PSA and total protein.

5-4. PSA Measurements

PSA in the cytosolic extracts andor tissue culture supematants were measured with a highly

sensitive and specific time-resolved immunofluorometric technique previously established and

described in detail elsewhere (1 8 1 ). In bnef, the PS A assay uses a mouse monoclonal anti-PS A

capture antibody coated to polystyrene microtiter wells, a biotinyiated monoclonal anti-PSA

5 1

detection antibody and alkaline phosphatase-labelled streptavidin (SA-ALP). In this imrnunoassay,

100 pL of sample is incubated with the capturing antibody in the presence of 50 pL of assay buffer

containhg the monoclonal biotinylated anti-PSA detection antibody. M e r 1 h incubation followed

by washing x6, the SA-ALP conjugate is added for 15 min followed by another washing x6. The

activity of ALP is then measured by adding the substrate 5'-Buorosalicyl phosphate, incubating for

10 min and then by adding a Tb3* and EDTA-containing developing solution. After 2 min. the

fluorescence is measured in the time-resolved fluorornetric mode with the CyberFluor-615

imrnunoanalyzer (CyberFluor Inc., Toronto, Ontario). This assay has a detection limit of 1 n g L of

PSA and can measure PSA at leveis of I ng/L or higher (up to 10,000 ng/L) with imprecision of

40%. Al1 assays were performed in duplicate. Tissue culture supenatants and/or tumor cytosolic

extracts were measured undiluted using 100 pL aliquots per assay, udess othenvise stated.

5-5. High-Performance Liquid Chromatography (HPLC)

HPLC analysis for molecular weight venfication of immunoreactive substances was

performed with a Hewlett Packard 1050 system. The mobile phase for the gel filtration

chromatography was a 0.1 moVL sodium sulfate and 0.1 moVL sodium dihydrogen phosphate

solution, pH 6.80. The ske exclusion column used was a TSK-GEL G3000SW. 60 cm x 7.5 mm

in combination with a guard column (TosoHaas, Montgomeryville, PA) with exclusion limit of 600

kDa and was calibrated with a molecular mass standard solution fiom Bio-Rad Laboratories,

Hercules, CA, which contains thyroglobulin (670 D a ) , IgG ( 158 kDa), ovalburnin (44 kDa).

myoglobin (1 7 kDa), and cyanocobalamin (1.35 kDa). The fIow rate was 0.5 W m i n and HPLC

was mn isocratically. After injection of 500 pL of centrifùged sample, fractions of 0.5 mL were

collected with a fraction collecter (Mode1 FRAC-100; Pharmacia, Upsala, Sweden).

5-6. Total Protein Determination

The concentration of total protein in ce11 lysates, tissue cytosolic extracts, and biological

fluids were measured in duplicate using a cornrnercially available kit (Pierce Chernical Co.

Rockford IL 61 105) based on the use of the bicinchoninic acid @CA) detection reagent, as

directed by the product literature for perfodng the "microtiter plate protocol". Assayed in parallel

were protein standard solutions, ranging in concentration from 0.05 to 2 g/L, which were made by

serial dilution of a 2 g/L. albumin standard with a buffer containing 50 mmol/L Tris, pH 7.4, and

7.5 mrnol/L NaN,. The same buffer was used to dilute turnor extracts whose initial total protein

concentration exceeded 2 mg/rnL. Ninety-six well, transparent polystyrene microtiter plates

(Dynatech Laboratories Inc., Chantilly VA) served as the reaction vessels. Absorbance at 545 nm

(close to the absorbance maximum of the reaction product, 562 nm) was measured on the ELISA

plate reader (Bio-Tek Instruments Inc., Winooski VT). Calculation of protein concentrations in the

turnor extracts, ce11 lysate, and biologicd fluids by interpolation From a linear calibration curve, was

performed using Deltasofl ELSA Analysis software (BioMetallics Inc., Princeton NJ). PSA

concentration in al1 cytosols is expressed as ng of PSA per g of total protein.

5-7. Isolation of Total RNA

Total RNA isolation from tissues and ce11 lines was performed using the TRIzol method

(GIBCO BRL, Gaithesburg, MD, TRIzol reagent) following the instructions of the manufacturer.

The m o l method is descnbed briefly as follows. About 200-500 mg of fiozen tumor tissue were

53

first pulverized to a fine powder at -80 OC, and homogenized in 1 rnL of TRIzoI reagent per 50 to

100 mg tissue. 10' cells were pelleted and used for total RNA extraction. M e r incubating the

samples and/or ce11 pellets for 5 minutes at room temperature to permit the complete dissociation

of nucleoprotein complexes, 200 pL chloroform per 1 mL of TRIzol reagent were added into the

solution. Following vigorous mixing, the solution was centrifùged at 12,000 x g for 15 minutes at

4 OC. Centrifugation separates the biphase mixture into the lower red phenol-chforoform phase and

upper colorless aqueous phase which was removed carefdly. The RNA was precipitated from the

aqueous phase by rnixing with 508 pL of isopropanol per i mL of initial TRIzol reagent. The

samples were incubated at room temperature for 10 minutes and centrifuged at 12, 000 x g for 10

minutes at 4 OC. The supernatant was removed and the RNA pellet was washed once with 75%

ethanol. M e r centrifugation and removal of the alcohoi, the pellet was air dned and dissolved in

diethyl pyrocarbonate (DEPC) treated water. The integnty of the RNA was checked

electrophoretically on 2% agarose-formaldehyde gels, and the amount and punty by

spectrophotometry at A?, and

5-8. Reverse Transcription

The synthesis of cDNA from the isolated total RNA was carried out with a first-strand

cDNA synthesis kit using Superscript II reverse transcriptase (GIBCO BRL, Gaithersburg, MD).

Briefly, 1-5 pg of RNA and oligo (dT),2.,, primers (500 ng), were first denatured for 10 min at 70

OC, chilleci on ice for 1 min and then incubated for 5 min at 42 OC in a 19 pL reaction mixture which

includes 10 x PCR buffer ( containing 200 mmoi/L of Tris-HCI and 500 rnmoVL of KCl, pH 8.4)

(Boehringer Mannheim), 10 mmoVL of deoxynucleotide triphosphate rnix (Boehringer Mannheim),

54

10 mmoVL of dithiothreitol @TT), and 25 mmoiL of MgCl? Then, 200 units (1 PL) of Superscript

II reverse transcriptase were added to the reaction mixture, incubated for 50 min at 42 OC,

terminated at 70 OC for 15 min and chilled on ice. The mixture was then treated with 1 pL of RNase

H for 20 min at 37 OC before proceeding to amplification of the target cDNA. Negative control

reactions for RT-PCR were performed using al1 reagents but without added Superscript II.

5-9. Oiigonucleotide Primers

Two oligonucleotide primers were used to ampli@ a 754 base pair region of PSA cDNA.

These, originally desctibed by Deguchi et al. (182) have the foliowing sequence: PSA Al : 5'-

TGCGCAAGTTCACCCTC A-3', PS A B I : 5'-CCCTCTCCTTACTTCATCC-3'. We have fiirther

developed another two primers for nested primer PCR as follows: PSA NI: 5'-

CTGTGTGCTGGACGCTGG-3', PS A N3 : 5'-ACCTCAC ACCTAAGGAC A-3'. For actin cDNA

amplification, we used the following primers, previously published (183). ACT 1: 5'-

AC AATGAGCTGCGTGTGGCT-3', ACT 2 : 5'-TCTCCTTAATGTC ACGC ACGA-3'.

PCR with pnmers AIE3 1 yields a 754 bp fragment fiom exons 4, 5 and the 3' untranslated

region, with primers NIN3 for nested primer PCR a 303 bp fragment and with ACTUACTZ a 372

bp fragment.

5-10. PCR Protocol

PCR was performed in 0.2 mL thin-walled MicroArnp reaction tubes on a Perkin-Elmer

Gene Arnp 2400 system. Total volume was 50 PL. The reaction mixture contained PCR buffer (50

mrnol/L KCI, 10 rnrnolR. Tris buffer, pH 8.3, 1.5 mrnoVL MgCl,, I OmgIL gelatin), 200 pmoVL of

55

deoxynucleoside triphosphates, (dNTPs), 1 pmol/L of PCR pnmers, 2.0 units of Taq DNA

polymerase (Boehringer Mannheim) and 5 pL of cDNA target (added last). The PCR was

performed with one cycle at 94 "C for 5 min, 30 cycles with denaturation at 94 "C for 30 S.

annealing at 60 OC for 30 s and extension at 72 "C for 30 s and one cycle at 72 OC for 7 min. Twenty

pL of PCR reactions were electophoresed on 2% agarose gels and visualized with ethidium bromide

staining. Negative controls did not contain template DNA.

5- 1 1. Nested Primer PCR

The procedure was the sarne as for the above PCR protocol except that 25 cycles were used

in the first PCR, pnmers Nl/N3 were used in the nested primer PCR for 20 cycles and 1 pL of the

first PCR product was added as target.

5-12. Labeling of PSA cDNA Probe

PSA cDNA plasmid was linearized with Hind III enzyme digestion and labeled with the

random primer method by incorporation of digoxigenin-labeled deoxyuridine triphosphate (DIG-

dUTP). We used the DIG-DNA labeling kit fiom Boehnnger Mannheim and the protocol

recornmended by the manufacturer. PSA cDNA plasmid was first incubated for 1 hour at 3 7 OC with

150 units of Hind-III restriction enzyme (Boehringer Mannheim) in a 10 rnmol/L of Tris-HCI

buffer, pH 8.0, containing 0.1 mol of NaCI, 5 mrnollL of MgCL, and 1 mmoK of P-2-

rnercaptoethanol per liter. Then, the PSA cDNA plasmid was denatured by heating in a boiling

water bath for 10 minutes and chilled on ice. Thirty pL of, 10 x concentrated dNTP mix, including

I rnmolL of dATP, 1 mmol/L of dCTP, 1 mrnoi/L of dGTP, 0.65 m o V L of dTTP, and 0.35

mrnol/L of DIG-dUTP. The whole mixture was incubated at least 60 minutes at 37 OC.

5- 13. RNA Labeling by In- Vitro Transcription

The PSA PCR product was cloned into the polylinker site of the pCR 2.1 transcription

vector. The recombinant PCR 2.1 transcription vector contains promoters for SP6 and T7 RNA

polymerases. After linearization of the vector with Hind III restriction enzyme, the T7 RNA

polymerase was used to create "mn-off' transcripts using the DIG RNA Labeling Kit (Boehringer

Mannheim, Germany) and the protocol recommended by the manufacturer. DIG-UTP was used as

a substrate and was incorporated into the transcript. The DIG-labelled cRNA was used as a non-

radioactive probe in Southem blots. In brief description, the following reagents were added to a

microfuge tube on ice: 20 pL of, 10 x transcription buffer, 20 PL of, 10 x concentrated NTP

labeling d u r e containing 10 mM ATP, 10 mM CTP, 10 &I GTP, 6.5 mM UTP, 3.5 rnM DIG-

UTP, pH 7.5, 10 pg o f linearized DNA, 400 units of T7 RNA polymerase, and 10 pL of RNase

inhibitor and centrifuged briefly. The reaction mixture was incubated for 2 hours at 37 OC. Longer

incubation does not increase the yield of labelled CRNA. DIG-labelled RNA probes have the

foliowing advantages: 1. they are of defined length; 2. they are single strand specific and therefore

al1 labelled RNA is available for hybridization and does not renature as in the case of DNA. cRNA

which is DIG-labelled according to the above protocol allows the detection of O. 1 pg homologous

DNA or RNA in a dot blot.

5-14. Direct Incorporation of DIGdUTP During PCR

We have used the same PCR protocol to ampli& PSA cDNA but we included a dNTP

57

mixture that contains, in addition :O the four nucleoside triphosphates, DIG-1 1-dUTP. The final

concentration of d N T P s and DIG- 1 1 -dUTP in the PCR reaction were 200 pmoI/L and 0.7 pmoVL,

respectively. PCR products were electrophoresed on 2% agarose gels, transferred and fixed as

descnbed. The DIG- 1 1-dUTP incorporated was detected using cherniluminescence as descnbed

in the previous section.

5-15. Detection Limit of PCR

We have used totai RNA extracteci from 10' LNCaP prostate carcinoma cells as a reference

preparation for determining lirnits of detection. This total RNA was reverse transcribed and then

diluted successively to give cDNA corresponding to a certain number of cells. We have also used

punfied plasrnid pA75 containing full-length of PSA cDNAs to calculate the detection lirnit of the

PCR methods. Plasmid and LNCaP cDNAs were diluted in a 10 rnrnol/L Tris buffer, pH 7.80

containing 1 mg/ml of salmon sperm DNA to avoid losses due to adsorption to tubes. Negative

controls were included as necessary with the diluent used as target.

5-16. Gel Electrophoresis, Southem Blotting and Hybridization

Aliquots of PCR produas (20 pL) were electrophoresed at 100 V for 45 min on 2% agarose

minigels containing ethdium brornide. The gels were then Southem blotted ont0 positiveiy charged

nylon membranes (Boehnnger Mannheim) by overnight alkali capillary blotting with use of 0.4

mol/L NaOH. The membranes were then baked for 15-30 min at 120°C. For hybridization, the

membranes were placed in a roller bottle with 20 rnL of hybndization buffer per 100 cm' of

membrane. We used a commercial hybridization solution @IG Easy Hyb, Boehnnger Mannheim).

58

Prehybridization was at 42 OC for 1 h. The solution was then replaced with 5 mL per 100 cm'

membrane of hybndization buffer containing 50 ng/rnL of fieshly denatured labeled PSA cDNA

andlor 200 ng/rnL DIG-labeled cRNA probe. Hybridization was carried out for 12- 16 h at 42 OC.

Filters were subsequently washed twice with 2 x SSC containing 0.1% SDS at room temperature

(5 min per wash) and twice with O. 1 x SSC containing 0.1% SDS at 68 OC (1 5 min per wash).

5-1 7. Detection Protocol

The detection was performed as follows: The membranes were first washed briefly in a

washing b&er containing 100 mM maleic acid, 150 rnM NaCl, pH 7.5 and 0.3% Tween 20 for 1-5

minutes at room temperature and then were incubated in blocking solution (Boehringer Mannheim)

for 30 minutes. Alkaiine phosphatase-conjugated anti-digoxigenin antibody (Fab Fragment) was

diluted to a final concentration of 35 mU/mL ( 5000-fold) in blocking solution. The membranes

were incubated in diluteci antibody conjugate solution for 30 minutes and then washed twice for 15

minutes each at room temperature in washing buffer. Finally, the membranes were equilibrated in

a buffer containing 0.1 moVL of Tris-HCl , 0.1 moVL of NaCl and 50 rnmoi/L of MgCl,, pH 9.5,

for 5 minutes at room temperature and incubated with 100-fold diluted cherniluminescent substrate

in the bfler stated above, AMPPD (1 0 rng/rnL, Boehringer Mannheim), andior CDP-star (25 mM,

Tropix, Bedford, MA). While CDP-star equilibrating buffer was 0.1 M diethanolamine, 0.1 rnM

MgC12, pH 9.5. The cherniluminescent signal was captured on X-ray film with 15-20 minutes

exposure for AMPPD and 10-60s exposure for CDP-star at room temperature.

Other methods for Southem blotting, probe radiolabeling, hybridization, and

autoradiography were aiso perforrned according to standard techniques (1 84).

59

5-18. Cloning of PCR Products

We have cloned the 754 bp PCR product of PSA cDNA and another five PCR products with

rnolecular weights between 250-500 bp using the TA cloning kit (Invitrogen Corporation, San

Diego, CA). We followed the protocol recomrnended by the manufacturer. Large quantities of the

recombinant plasrnids were prepared by culturing transformed E. coli cells in LB-ampicillin media

and extracting the plasmids with the Qiagen Midi plasmid purification kit (Qiagen Inc, Chatsworth,

CA).

5-19. DNA Sequencing

PCR products were directly sequenced using the Thermo Sequenase kit (Amersham

Internat ional, Buckinghamshire, England). The protocol recomrnended by the manufacturer was

used throughout. Sequencing primers, labelled at the S'-end with Cy5 fluorescent dye? were as

follows: PSA-S 1 5'-AAGGTGACCAAGTTCATG-3' (binds 19 nucleotides intemally from PCR

primer PSA-A 1). PSA-S2: 5'-CCATCCCATGCCAAAGGA-3' (binds 19 nucleotides intemally

6om PCR pimer PSA-A2). Al1 sequencing reactions were loaded on the ALF Express Automatic

Sequencer (Pharmacia Biotech, Uppsala, Sweden). Sequence cornparisons were performed with

BLAST and DNASIS software.

5-20. Tissue Culture System

5-20- l. Conipo~ndr

Al1 steroidal and non-steroidal compounds used in this study were obtained from Sigma

Chernical Co., St. Louis, MO 63 l78., except for the following: ICI 102,780 and Casodex (ICI

60

1 76,334) (Zeneca Pharma Inc., Mississauga, ON, Canada); RU58,668, RU54,876, RU56,187,

Nilutamide (Anandron, RU23,908), and Mifepristone (RU486, RU38,486) (Russel-UCLAF,

Romainville, France), Beclomethasone Dipropionate (Glaxo Wellcome Inc., Mississauga, ON,

Canada), vitamin D analogs ( ' O 8-8892, Ro 2 1-553 5, Ro 23-7553, and Ro 24-553 1) (Hofian-La

Roche Inc, Nutley, New Jersey, USA), EB- 1 O89 (Leo Pharmaceuticai Products Ltd, Allerup.

Denmark), LG100 153, LGlOO272, and Degnelin (LIGAND Pharmaceuticals Inc, San Diego, CA),

Recombinant Human FGF acidic, EGF, IGF-1, and IGF-II (R&D Systems, Minneapolis, MN).

Hydroxyflutarnide was a gifi from Dr. Doma Peehl. Stanford University. Stock solutions ( 10'~ or

lo5 M) were prepared in absolute ethanol. More dilute solutions were also prepared in the sarne

solvent. A total number of 64 substances were tested in tissue culture expenments, which included

androgens, progestins, estrogens, glucocorticoids, mineralocorticoids, vitamin D, retinoic acid,

growth factors, gonadotropin hormones, cyclohexirnide, and steroid hormone antagonists. The list

of al1 compounds along with their biological function are s h o m in Table 5-20-1-1.

5-20-2. Cell lines

In tissue culture expenments various tumor ceil lines were tested for PSA expression. Al1

the ce11 lines used were obtained from the Amencan Type Culture Collection (ATTC), Rockville,

MD 20852., except for the following: SAOS (osteosarcorna; provided by Dr. M. Grynpas, Mount

Sinai Hospital, Toronto), BG-1 (ovanan carcinoma; provided by Dr. H. Rocheford, INSERM,

Montpelier, France), and MFM-223 and MFE-296 (breast and endometrial cancer ce11 lines

provided by DR. R. Hackenberg, Klinikum der Phillips-Universitat, Marburg, Germany). The list

of the ce11 lines tested along with their tissue origin are shown in Table 5-20-2-1.

Table 5-20-1-1. List of Steroid/Cornpounds Tested for PSA Production in Tissue Culture Systern

Biological Action

1 1 1 Testosterone 1 Androgen

--

~ ~ d r o a n d r o s t e r o n e 7 Androgen

2

1 4 1 RI 88 1 (methyltrienolone) 1 Androgen

Androsterone Androgen

6

1 9 1 Estriol 1 Estrogen metabolite

7

8

Dihydrotestosterone Androgen

Dihydroisoandrosteronee Sulfate (DEA-SO,)

Estrone

10

1 13 1 Hydrocortisone 1 Glucocorticoid

Androgen Met abolite

Estrogen

1 1

12

1 7a-Ethynylestradiol Estrogen

P-Estradiol

Corticosterone

14

1 16 1 Prednisone 1 Glucocorticoid

Estrogen

Glucocorticoid

15

Betarnethasone 1 7-valerate Glucocorticoid

Dexamethasone

1 7

Glucocorticoid

1 8

1 21 1 Norethidrone 1 Progestin

Beclomethasone dipropionate

19

20

Glucocorticoid

1 7 a-Hydroxyprogesterone

1 23 1 Depo-Provera 1 Progestin

Progesterone precursor

Progesterone

Norethynodrel

22

Progestin

Progestin

Norgestrel Progestin

Steroid/ Compounds Biological Action

1 25 1 Triarncinolone Acetonide 1 Progestin/ Glucocorticoid 1 24

1 26 1 Aidosterone 1 Mineralocorticoid 1

Norgestimate

( 29 1 Ro23-7553 1 Vitamin D, analog

Progestin

27

28

1 32 1 LG100275 (TTNPB) 1 RAR Agonist

EB1089

Ro 23-553

30

3 1

1 34 1 13-cis Retionoic Acid 1 Retinoic Acic

Vitarnin D, analog

Vitamin D, analog

1 35 1 Al1 Tram Retinoic Acid 1 Retionoic acid

Ro-24-5531

Degnelin

Vitamin D, analog

Rentioc Acid Agonist

1 38 1 ICI 182, 780

36

37

Antiestrogen

Antiestrogen

LG 100153

Tarnoxifen

RXR Agonist

Antiestrogen

1 42 1 Cyproterone Acetate 1 Antiandroged Progestin

40

4 1

1 43 1 Casodex (ICI 176,334) 1 Antiandrogen

1 44 1 RU%, 187 Antiandrogen

RU 54, 876

HydroxJutamide

- .- - - - . - -

I 1 45 1 Nilutamide (Anandron) Antiandrogen

Antiestrogen

Antiandrogen

1 46 1 Mifepristone (RU 4861 RU 38,486) 1 Antiprogestin

1 47 1 ~ortexolone 1 Antiglucocorticoid

1 48 1 Spironolactone 1 htirnineralocorticoid

I i ! Biological Action

51 1 GH 1 Growth Hormone

49

50

Lactogenic

53 1 LH 1 Gonadotropin

Cholesteroi

Cycloheximide

54 1 FSH 1 Gonadotropin

Steroid hormone precusor

Protein Synthsis inhibitor

55 1 P-HCG 1 Chriogonadotropin

56 1 IGF-I 1 Growth Factor

57 1 IGF-II 1 Growth Factor

58 1 a-FGF 1 Growth Factor

59 1 EGF 1 Growth factor

Oral Contraceptive

Oral Contraceptive

Control

64 1 Nothing 1 Control

Table 5-20-2-1. Cell lines Tested for PSA Production in Tissue Culture System

Tissue of Origin 1 Steroid Hormone Receptor Status

1

2

BT-474 1 Brdast 1 ER (+), PR (+), AR (+) '.'

3

4

T-47D

MCF-7

Breast 1 ER (-), PR (-), AR (-)'.'

ZR-75- 1

MDA-M.-453

6

7

9 1 BG- 1

Breast

B reast

10 1 HBT-75 1 0varia.n ( Not Reported

ER (+), PR (+), AR (+)l.'

ER (+), PR (+), AR (+)'.'

Breast

Breast

MFM-23 3

HBL- 1 O0

ER (+), PR (+), AR (+)'y2

ER (-), PR (+), AR (+)L*2

13 ( A-427 1 Lung 1 Not Reported

Breast

B reast

1 1

12

Not Reported

ER (-), PR (-), AR (+)'*'

ER (-), PR (-), AR (-)'s2

NCbH460 1 Lung 1 Not Reported

HBT- 16 1

HBT- 173

16 1 SOAS 1 Bone

- --

Not Reported

Ovarian

Lung

19 1 Co10 320 1 Colon 1 Not Reported

AR (+), ER (+), PR (?)"

Not Reported

prostate 1 ER (-), PR (-), AR (+) '+'

1. J Steroid Biochem Mol Biol. 56: 1 13-1 17, 1996 2. Clarke R Leonessa F, Bmmer N, Thompson EW. ( 1996). Diseases of the breast : In vitro

models of breast cancer.Harris JR, Lippman ME, Morrow M, Hellman S (eds). Lippman- Raven Press: New York, pp.245-26 1.

3. J Clin Endocrino1 Metab 76: 1497- 1502, 1992 4. Cancer Res 4452864290, 1984 5. Biochem Biophys Res Commun, 192:940-947, 1993

5-2 0-3. Ce f f Culture Procedure

fhe ce1 lines were initially grown in plastic culture flasks at 37" with 5% of COZ in RPMI

1640 culture medium (GIBCO BRL, Gaithersburg, MD) supplemented with glutamine (200

mrnollL), bovine insulin ( 10 mg/L), fetai bovine semm ( 1 O%), antibiotics (penicillin, streptomycin)

and antirnycotics (amphitericin B) unless otherwise recornrnended by the providers. The cells were

cultured to near conûuency and then transferred to phenol red-free media containhg 10% charcoal-

stripped fetal bovine serum (Gibco BRL, Gaithesburg, MD)with antibiotics/antimycotics. Phenol

red-free media were used since phenol red was found previously to have weak estrogenic activity

(185) and charcoal-stripped fetal bovine serum is devoid of steroid hormones. The cells were then

detached by trypsin-EDTA treatment, washed with PBS buffer and subcultured as above in 24-well

tissue culture plates (Corning # 25820) until confluence with change in media at 3 days. Each well

contained 2 mL of phenol red-fiee media with 10% charcoai-stripped fetai calf serum and

antibiotics/antimycoticç.

5-20-4. S~imulution w i h Steroid Compozrnak

Stimulation was initiated by adding 2 pL of each steroid and steroid compounds (IO-<- 1

M) dissolved in 100% ethanol and incubating for a cenain penod of time (usually up to 8 days).

Tissue culture supematants (-150 pL ) were removed for PSA anaiysis at days 3,5, and 8 . Slight

modifications of this protocol were introduced as necessaq. Appropriate multiple positive and

negative controls (oniy alcohol added) were included in each expenment. Wells with microbial

contamination were excluded fiom the data anaiysis.

66

5-20-5 Stimulation with S e m Smp[es and Gonadotropin Hormones

Stimulations were camied out with confluent T-47D and BT-474 breast carcinoma ce11 lines

containhg 1 ml of phenol red-ffee media and 10% charcoal-stripped fetal calf serum and

antibiotics/antimycotics. Stimulation was initiated by adding 1 mi of filter-steralized senim sarnple

and incubating for a certain time (24 h). Tissue culture supernatant (- 150 PL) was removed for PSA

protein analysis after 24 h . Slight modifications of this protocol were introduced as necessary.

Appropriate multiple negative controls (no semm added) were included in each experiment .

Positive controls consisted of T-47D cells stimulated with the synthetic progestin Norgestrel which

was found previously (96) to induce PSA mRNA and PSA protein production by T-47D cells.

Additional experiments were performed by using progesterone at final concentrations of 1 O-'- 1 0-l1

M. prolactin (5 ng/mL-5 pg/rnL), growth hormone (10 ng/mL-10 pg/mL), LH (10 ng/mL-10

pg/mL), FSH (1 0 ng/mL- 1 O pg/rnL) and choriogonadotropin (5 IU/mL-0.005 IU/rnL). Al1

concentrations are final, in the microtiter weIls.

5-20-6. Blockuzg Experirnents

Blocking experiments were performed by simultaneously checking for the following

possibilities: (a) Stimulation by the blocker alone at a final concentration of 1 0-8 M. (b) Stimulation

by the stimulating steroid alone at a final concentration of IO-^ MM. (c) Adding the blocker to the

cells at a final concentration of IO-' M, incubating for 1 h and then adding the stimulant at a

concentration of 10 '~ M. (d) Including controls with ethanol only (negative controls). This protocol

allows for a direct cornparison of the stimulating activity of either the blocker or the stimulant and

the effect of the blocker on the ability of the stimulant to induce PSA expression when the blocker

67

is allowed to bind to the receptors at 10-fold higher concentrations for I h before the addition of

the stimulant.

5-20- 7. Kineizc Experiments

The kinetics of PSA production by T-47D and BT-474 cells were studied as follows.

Confluent T-47D and BT-474 cells were stimulated with Norgestrel at a final concentration of 10"

M and the cells were harvested dong with tissue culture supematants at 1, 2, 4, 8, 24 and 48 h.

Control cells were harvested at 48 h without any stimulation (ethanol added only). The tissue

culture supematants and portion of the cells were used for PSA protein analysis; another portion

of the cells was used to extract total RNA for PCR analysis of PSA mRNA.

5-21. Lysis Procedure

AU cell pellets were lysed for 30 minutes on ice with 1 mL of lysis buffer. Lysis buffer was

50 mmol/L Tris, pH 8.0 containing 150 mrnoi/L sodium chioride, 5 mmol/L

ethylenediaminetetraacetic acid (EDTA), 10 g/L Nonidet NP-40 surfactant, 1 mrnoVL

phenylmethylsulphonyi fluoride and 1 m g L each of aprotinin and leupeptin as proteinase

inhibitors. The lysate was centrifugecf at 15,000 g at 4 "C for 30 minutes, the supernatant was

collected and immediately assayed for PSA and total protein.

5-22. Measurement of Steroid Hormone Receptors

The concentration (fhoVrng) of estrogen receptors (ER) and progesterone receptors (PR)

in al1 tumor ceIl lines and breast tumor specirnens were measured with the Abbott enzyme

68

immunoassay kits (Abboa Laboratories, North Chicago, IL) at the Sumybroo k Heaith Science

Centre.

5-23. Measurement of Hormones in Serum

AU serum samples collected dunng the course of the menstrual cycle were analyzed for the

following hormone levels: FSH, LH, estradiol and progesterone. FSH and LH were assayed with

the ~ccess" Immunoassay Analyzer (Sanofi Diagnostics Pasteur, Montreal, Quebec, Canada). The

Iowest detectable level of FSH and LH distinguishable fiom zero with 95% confidence is 0.2 U L

and both assays exhibit total imprecision under 10% across the assay range. Estradiol was measured

by a solid-phase, chemiluminescence enzyme immunoassay system (Immulites, Diagnostic

Products Corporation, Los Angels, CA). Detection limit and coefficient of variation are 0.044

nrn0i.L and less than 10%, respectively. Progesterone was measured with the Ciba-Corning ACS*

progesterone cherniluminescence irnmunoassay (Ciba-Corning Canada, Markharn, ON, Canada).

The ACS" progesterone assay has a minimum detectable concentration of 0.35 nmoVL and

coefficients of variation 4 0 % within the measuring range. The detection limit of the assay is

approximately 0.1 5 pg/rnL and precision < 10%.

5-24. Immunohistochemistry

Immunohistochemical stains for PSA protein were performed on Cpm-thick paraffin

sections of selected blocks of the pnmary lung adenocarcinorna using a streptavidin-biotin

technique and a polyclonal antiserum directed against human seminal plasma PSA protein

(Biomeda, Foster City, CA). Sections were pretreated with microwave heating to facilitate antigen

69

retrieval. Human prostate was used as a positive control while non-immune rabbit serum served

as a negative control for assessrnent of non-specific staining.

5-25. Su bject Selection

5-25 1. Bremi Tisnre Specimens

Thirty breast cancer specimens were obtained from women undergoing surgery for primary

breast cancer. The breast tumor tissues were stored in liquid nitrogen irnrnediately &er surgical

resection, transported to the laboratory and subsequently stored at -70 OC until protein and RNA

extractions were performed. Al1 tumors were primary lesions except for one ovarian carcinoma,

which was metastatic from a primary breast cancer. We also used one lung carcinoma which was

metastatic from a primary prostate cancer, one prirnary liver cancer and one primary prostate

cancer, al1 as control tissues. Al1 tissues were lefkovers from routine pathological exarninations.

Steroid hormone receptor analysis (ER and PR) was also performed on each tumor specimens using

the Abbon enzyme immunoassay kits.

5-25-2. Semm Specimens Obtahed During Menstnial Cycle

Fourteen healthy, reproductive age women with regular owlatory menstrual cycles, as

determined by history and either basal body temperature charting or luteal phase progesterone

measurements, were studied dunng the period of one whole menstrual cycle. Blood was collected

over the one month cycle. Criteria for inclusion in this study were: female between 20-46 years of

age, normal menstrual cycle and not on birth control pills or other medications. Samples were

collected eight times (every 3-4 days) over the course of the cycle, usually three in the phase after

70

menstmation (follicular phase), two near ovulation and three in the luteal phase. Al1 women had

a normal screening medical history; one reported a past history of breast fibroadenoma which was

not treated. Six of the volunteers were nulliparous (no pregnancy) and one had two normal

pregnancies. Race of the volunteers was two Blacks, two Oriental, and ten Caucasians. Menstrual

cycle regularity and ages of the women were between 26-39 days and 29-46 years, respectively.

5-53. Lztng Cancer Patients

5-25-3- I. Case stucs

Patient: The patient was an 87 year old femaie who was initially admitted to the Ernergency

Department of St. Joseph's Health Centre, Toronto, with acute cholecystitis. A chest x-ray at the

time of admission revealed that the patient had fairly severe lung disease with chronic emphysema

but also a nodule in the right lung suspicious for carcinoma. The patient undenvent wedge

resection. On pathological examination, the lung lesion proved to be a moderately differentiated

adenocarcinorna with a maximum dimension of 1.5 cm. The cancer arose in an area of a scar with

extensive deposition of anthracotic pigment. The tumor consisted of sheets and glands of large cells

with eosinophilic cytoplasm and large hyperchromatic nuclei with central nucleoii. The pleural

surface and resection margin were fiee of tumor.

Post-operatively, the patient did well initially, however, she soon developed abdominal pain

and was totally unable to eat. With a diagnosis of cholecystitis, the patient undenvent a video-

assisted cholecystectomy. Post-operatively, she developed complications with abdominal bleeding

and died 32 days d e r admission.

During the course of her disease, she was treated with the following dmgs: Zantac

71

(ranitidine hydrochloride; H, receptor antagonist); Dilitiazem (antianginal agent; calcium channel

blocker); Ventolin (salbutamol; bronchodilator, beta-2 adrenergic stimulant); Lasix (furocexnide,

diuretic), Theo-Dur (theophylline, bronchodilator); Demerol (meperidine, narcotic, analgesic);

Morphine (narcotic, analgesic); and Becloforte (beclomethasone dipropionate, antinflamato~y

corticosteroid).

Tumor tissue was snap-fiozen in liquid nitrogen and subsequently stored at -80 O C . In

parallel, tumor tissue and adjacent normal tissue excised during surgery was fixed in formalin and

embedded in paraffin. Fresh fiozen tissue was used to prepare cytosolic extracts and for RNA

extraction; the parafin-embedded tissue was used for histopathology and immunohistochernistry.

5-25-3-2. Lung Ttmor Tissues

Patients. 52 turnor tissue specimens were obtained h m patients undergoing surgery for primary

lung cancer. Seven specirnens of histologically normal lung tissue were recovered fiom adjacent

tissues of the lung tumors. Diagnosis was histologically confirmed and tumors were staged using

the TNM (T-rimary tumor; N=regional lyrnph nodes; M=distant met astasis) classification system.

Tumor tissues and adjacent normal tissues, excised during surgery, were snap-fiozen in liquid

nitrogen and subsequently stored at -80°C until use.

Detailed clinicopathological features of Our patient population are summarized in Table 5-

25-3-2- 1.

5-26. Ethical Issues

Al1 research projects involving human biological specimens and clinical information were

Table 5-25-3-2-1. Clinicopathological Findings in 52 patients with primary Lung Tumors

Variable* Number of Patients ?40

Sex

Male

Fernale

Stage

Histological Type

Squarnous

Adenocarcinorna

Smail Cell

Carcinoid Tumor

Large Ce11

Nodal Status

Node Negative 38 73

Node Positive 14 27 *Age was between 42 -89 years (median 68). "Grade unknown for 5 patients.

73

approved by the local ethics authorities. Informed consent was signed at the local institution by the

patients who were d i n g to provide their biological specirnens and clinical information for medical

research. Al1 information released or published from the studies referred to code numbers known

only to the principal investigators. Extraordinary care was taken to preserve the confldentiality of

participating patients.

5-27. Statistical Analysis

The distribution of demographic and clinicopathologic variables, including age, stage,

histological grade and type, and nodal status were compared between PSA-positive and PSA-

negative patients with the use of contingency tables analyzed by the chi-square test or Fisher's Exact

test where appropnate. AU analyses were performed using SAS software (SAS Institute, Cary, NC,

USA).

CHAPTER 6, Results

6-1. Optimization of PCR Protocol

The polymerase chain reaction (PCR) is a recently developed procedure for the in-vitro

arnpilification of DNA sequences that has gained widespread acceptance in many areas of

rnolecular biology particularly in tumor biology, clinical medicine and more recently in routine

application for diagnosis. Although the basic principles are simple and straightforward, the reaction

itself involves complex kinetic interactions between template, product, pnmer, nucleotide

triphosphates, and enzymes that change throughoiit the course of the reaction. Despite this

complexity, PCR amplification is remarkably improved with the new thermostable Taq DNA

polymerase. There are, however, adjustments that can be made to some of the reaction parameters

that in some cases will drarnaticaily improve specificity and yield. These adjustments include

alterations of the reaction buffer, particularly MgC12, pnmer concentration, dNTP concentration,

enzyme conccentration, annealing time and temperature, and extension time and temperature. The

efficacy of PCR is measured by its specificity, efficiency (i.e. yield), and fidelity. A highly specific

PCR will generate only one amplification product that is the intended target sequence. More

efficient amplification will generate more product with fewer cycles. A highly accurate (Le. high

fidelity) PCR will contain a negligible amount of DNA polymerase induced errors in its product.

An ideal PCR would be the one with high specificity, yield and fidelity. Therefore, dunng

the initial attempts of developinç a PCR method for PSA gene amplification, we designed specific

primers for PSA and optimized al1 conditions of ampilification including reaction buffer, primer

concentration, annealing tirne and temperature, dNTP concentration, number of cycles, DNA Taq

75

polymerase concentration, and arnount of target used. As mentioned, PCR is a very sensitive

method. It has been s h o w that the technique can arnplfy successfùlly DNA targets that are

presents in complex mixtures as single molecules (186). Such studies are bea done with senal

dilutions of plasmid vector.

6-2. Determining Limits of Detection Methods

6-2- 1. E~hidium Bromide Staining

The detection limit of Our PCR method was first determined by amplification of plasmid

pA75 containing the full-length PSA cDNA. Ethidium bromide staining of the PCR product on a

2% agarose gel revealed a detection limit of - 1,000 copies of PSA cDNA (Figure 6-2-1-1). A

plasmid containing p-actin cDNA was also amplified using p-actin specific primers. The detection

limit of actin cDNA with ethidium bromide staining was -200 copies. In both PCRs for PSA

cDNA and actin cDNA, the PCR products were of the expected length i.e. 754 and 372 bp,

respectively .

Total RNA was extracted from the human prostatic carcinoma cell line LNCaP which is

known to express PSA and from the human breast carcinoma ce11 iine BT-20 which is steroid

hormone receptor negative and does not express PSA protein. The total RNA from both ce11 lines

was reverse transcnbed and the cDNA diluted serially and amplified by PCR. The BT-20 cDNA

did not yield any amplification products at any dilution confirming absence of PSA mRNA in this

ce11 line. The LNCaP cDNA yielded the expected 754 bp amplification product with cDNA

equivalent to 2 LNCaP cells or higher. Actin cDNA was successfully amplified fiom both ce11

lines. These data are presented in Figure 6-2-1-1B.

Figure 6-2-1-2. RT-PCR of PSA &A (A upper panel) and p-actin mRNA (B, lower panel). The PCR product was detected on agarose gels by ethidium bromide staining in both cases. Lane 1. Molecular weight markers in base pain. Lanes 2 - 8. mRNA extracted fkom LNCaP cells corresponding to 20,000 (2). 2,000 (3), 200 (4). 20 (5). 2 (6) , 0.2 (7) and 0.02 ( 8 ) cells. Lane 9. Control PCR with 9,000 molecules of plasmid pA75, containing full-length PSA cDNA as target. Lanes 10-1 1 . mRNA extracted fiom 20,000 (10) or 2,000 (1 1 ) BT-20 cells which are negative for PSA mRNA but positive for p-actin rnRNA. The detection lirnit is PSA mRNA equivalent to 2 LNCaP cells.

6-2-2. Nested PCR

Similar expenments were performed using the nested-primer protocol. With this method,

we could detect less than 10 molecules of plasmid pA75 and cDNA equivdent to 0.02 LNCaP cells

(Figure 6-2-2-1 ). This data suggests that the nested primer PCR method is about 100 times more

sensitive than the regular PCR protocol when using ethidiurn bromide stained gels for detection of

PCR products.

6-2-3. Direct DIGdUTP Incorporation

We have further developed a third highly sensitive PCR procudure for PSA cDNA using

one PCR reaction, in which digoxigenin-labeled Il-dUTP is direaly incorporated into the PCR

product. DIG- 1 1 -dUTP is added in minute amounts in the PCR mix dong with the other dNTPs.

The DIG-1 1-dUTP labeled PCR product is first run on agarose gels, it is then transferred by

Southem blotting to nylon membranes and subsequently detected using anti-digoxigenin antibodies

labeled with aikaiine phosphatase and cherniluminescence. This method gave detection lirnits of

about 10 molecules of pA75 plasrnid and 0.2 LNCaP cells (Figure 6-2-3-1 ).

6 - 2 4 Souihem B Iot Hybridtzation

A fourth detection method of PSA cDNA was based on a single PCR reaction, agarose gel

electrophoresis of the product, Southem transfer to nylon membranes, hybndization with a

digoxigenin-labeled PSA cDNA probe and detection by cherniluminescence. This method gave

detectabilities of about 10 molecules of pA75 plasmid and 0.2 cells of LNCaP. A summary of the

Figure 6-2-2-1. Nested primer PCR protocol for PSA mRNA. A. The target is pA75 plasmid. Number of pA75 m o l d e s per lane (in brackets) as follows: 0(2), 1 (3), 10 (4), 100 ( S ) , 1,000 (6), 10,000 (7). Lane 1 contains molecular wieght markers in base pairs and Lane 8 is negative control. B. The target is LNCaP mRNA equivalent to the following number o f cells: 0.002 (2), 0.02 (3), 0.2 (4), 2 (5), 20 (6). 200 (7). The detection limit in A is 510 pA75 molecuies and in B is 0.02 LNCaP celIs.

Figure 6-2-34. RT-PCR protocol for PSA mRNA with direct incorporation of DIG- 1 1-dUTP and cherniluminescence detection. A The target is pA7S plasmid. Number of pA75 molecules per lane (in brackets) as follows: O ( 2 , 1 ( 3 , O ( 4 100 (5). 1,000 (6), 10,000 (7). Lane 1 contains biotinylated rnolecular weight markers. B. The target is LNCaP rnRNA equivalent to the following number of cefls: O (2), 0.02 (3), 0.2 (4), 2 (9, 20 (6) , 200 (7). The detection lirnit in A is 10 pA75 molecules and in B is 0.2 LNCaP cells.

Table 6-2-4-1. Detection of PSA mRNA with various RT-PCR protocols

PROTOCOL DETECTION METHOD DETECTION LIMIT -

pA75 Plumid LNCaP (No. of copies) (No. of ctlb)

-- - -

1 . PCR with one set of Agarose gel elcctrophorcsis 1000 2 prîmen (AGE); ethidium bromide

staiaing (EBS) - - - -

2. Nestcd primer PCR AGE, EBS 510 0.02 - - - - -- -

3. PCR with one set of Incorporatin of DIG- 1 LdUTP; 1 O 0.2 p rim ers AGE; Southent blotting (SB);

anti-DIG antibodics labeled with alkaline phosphatase (ALP); cherniluminescence. - - -- --

4. PCR with one set of AGE; SB; hybridization with - < I O 0.2 prirners dig-labeled PSA cDNA probe;

anci-DIG antibodies labeled with ALP; cherniluminescence

developed techniques and their detectability is presented in Table 6-2-4-1.

6-3. PSA Gene Expression in Breast Tumors

In order to study PSA mRNA expression in breast tumors, we have initially screened 240

breast tumor cytosolic extracts for PSA protein using a highly sensitive immunological assay (1 74).

We then classified these tumors into three categones based on the arnount of PSA expressed Le.

highly positive tumors (PSA > 100 ng/g of total protein), weakly positive tumors ( 15 ng/g CPSA

< 100 ng/g) and negative tumors (PSA < 1 5 ng/g). We then selected 30 tumors for fùrther study at

the mRNA level using PCR technology. The results are shown in Table 6-11. The nested primer

method was not used for tumor RNA analysis since we found that this assay could not be

completely controlled for PCR contamination.

From the 10 tumors that are highiy positive for PSA protein, 5 were positive with PCR

method 1 (one PCR, ethidium bromide staining), 10 were positive by method 3 (DIG-dUTP

incorporation) and 9 were positive by method 4 (probe hybridization). One sarnple was not

evaluated by method 4. From the 10 tumors which are weakly positive for PSA protein, none was

positive with method 1, 6 were positive by method 3 and 6 were positive by method 4. From the

6 positives there were two discrepant sarnples; sarnples 12 and 1 7. From the 10 tumors which are

negative for PSA protein, none of the methods detected PSA mRNA except for sarnple 24 which

was found weakly positive only by method 4 (Table 6-3-1).

Resuits of PCR analysis of some breast tumors are shown in Figure 6-3-1. In addition to

these tumors, we also tested control tissue fiom prostate (positive with al1 methods), a lung cancer

which was metastatic fiom a prostate primary lesion (positive with al1 rnethods) and one primary

Iiver cancer (negative by dl methods).

Table 11-34 Detection

Spechen #

1. 38101

2. 38 10-5

3. 38 10-7

4. 3810-10

5. 38 14-6

6. 3816-1 1

7. 3828-3

8. 383 1-7

9. 3832-5

10. 3818-1

I I . 3827-5

12. 3833-6

13. 383 1-6

14. 38 13-8

15. 38 14-5

16. 3822-3

17. 3822-1

18. 3818-12

19. 3828-I

20. 3832-7

21. 3827-3

3 3 -- . 3832-2

23. 3832-3

24. 38324

25. 3833-1 1

26. 38 14-3

27. 3822-1 1

28. 3828-10 --

29. 38284

30. 3833-1

1 .Methoci 1 is one PCR, digo'cigenin- 1 I -dUTP

of PSA Protein PSA

( ng/g of total protein)

269

1102

828 1

1 106

3417

2387

IO4

14 15

724 1

18 1

64

8

23

24

12

79

5 2

57

20

86

O

O

O

O

O

O

O

O --

O

O

agarose gel eIectrophoresis incorporation. 3 .Method 4 is

and PSA mRNA in 30 ER/PR

(- of

18 1.'80

2'4

45 1.'661

235,'s 15

71.;16

8i4

143A2 1

16i253

341 19

1901368

0: 5

4: 1

36/73

7321296

3901307

595

21'32

32379

701'1 18

2257

5853

266i5

99; 128

971'85

O/ 0

2781582

01 1

1!0

1891245

6312

and ethidium bromide based on Sûuthern bfot and

Method 4@'

- +

NDW

- +

-

-

.-

based on

Primary Breast

Methd 1'')

-

- -

+

-

stahing. hybridization.

Tumors PSA mRNA

Method 3a)

- - +

- - -

6

- -

A

- -

-

- 2.Method 3 1s 4.ND~iotdone

Figure 11-31. Detection of PSA mRNA by RT-PCR and hybridization (Method 4) in breast tumors weakly positive for PSA protein. Lane 1 . Biotinylated molecular weight markers. Lanes 2- 1 1. Samples 1 1 to 20 of Table 2. Lane 12. Negative control. Actin PCR was positive in al1 samples (data not shown).

6-4. Steroid Hormone Regulation of PSA Gene Expression in Breast Cancer

AI1 ce11 lines used were first measured for the presence of the steroid hormone receptors,

ER and PR. The obtained results were summarized in Table 6-41.

In order to investigate the mechanism of PSA gene regulation in the breast, al1 the breast

cancer ce11 lines tested (Table 6-20-2-1) for PSA production only two steroid hormone receptor

positive ce11 lines, T-47D and BT-474, were able to expressed PSA at the protein and the rnRNA

level in response to stimulatory eRects of steroid hormones. Therefore, we were able to establish

a highly sensitive tissue culture system which reproduces in-vitro the phenornenon of PSA

production by breast cells. These MO steroid hormone receptor-positive breast carcinoma ceil lines,

T-47D and BT-474, do not produce detectable PSA when cultured in media devoid of steroid

hormones. Upon stimulation by steroid hormones , these ce11 lines produce PSA in a dose-response

manner. We have used these system to study the kinetics of PSA production, the dose-response of

PSA production by various steroid hormones and the blocking effect of antihormones.

6-4- 1. Kinetics of PSA Expression

The breast carcinoma ce11 lines, T-47D and BT-474, were cultured in the absence of any

stimulating steroid and in the presence of the stimulating steroid Norgestrel at a concentration of

10' M. The appearance of PSA mRNA was monitored with reverse transcription-polymerase chah

reaction (RT-PCR). The appearance of PSA protein inside and outside of the ce11 (secreted protein)

was monitored by the irnmunoassay procedure. The results are summarized in Figures 6-4-14 and

6-40 1-2.

Table 6-41. Steroid Hormone Receptor Levels in Cell lines Tested for PSA Production

Progesterone Receptor

(PR) fmoümg*

Tissue of Origin

--

Breast 1 1 1 1 1 023

Estrogen Receptor (ER) fmoWmg*

Breast 1 14 1 38

Breast 1 112 1 482

Breast 1 O 1 O

Breast 1 1 I 1

Breast 1 O 1 O

Breast 1 O 1 O

19 1 BG- 1

HBT- 173

Lung I 0 l 1

SOAS

Colo 320 01 Bone 78

Endornetrium 1

Coion I 1 I O

3

1

Prostate I O I O

* hoVrng of total protein

Figure 6-4-1-1. RT-PCR of PSA mRNA extracted from T-47D cells. A. Ethidium bromide-stained agarose gel. B. The gel was Southern-transferred and the PCR product, in which digoxigenin- dUTP was incorporated during PC& was detected with anti-digoxigenin antibodies and chernilurninescence. Lanes 1 and 7. T-47D cells were stimulated with absolute alcohol (solvent) and mRNA extracted at the beginning (O time) or at the end (48 h) of the experiment, respectively. Lanes 2 ,3 ,4,5,6 and 8. The T-47D ceUs were stimulated with IO-' M Norgestrel once and mRNA extracted f i e r 1 h (lane 2), 2 h (lane 3), 4 h (lane 4), 8 h (lane 5), 24 h (lane 6) and 48 h (lane 8). M. W., molecular weight markers. The PCR product is 754 bp in size. Actin RT-PCR was performed in al1 cDNAs Frorn lanes 1-8 and it was positive in al1 cases.

O Hours 1 2 4 8 24 4 8

Figure 6-4-1-2. Appearance of PSA mRNA and PSA protein intracellularly (O) or in the tissue culture supernatant (4) foilowing stimulation of T47D ceils with lod M Norgestrel. The post- stimulation time of first appearance of PSA mRNA is 2 h, of intracellular PSA protein is 4 h and of PSA protein in tissue culture supernatant is 8 h. Cells not stimulated at al1 (data not shown) or stimulated with ethanol did not produce PSA mRNA or protein during the 48 h duration of this experirnent. No PSA mRNA or protein was seen in Norgestrel- stimulated or unstimulated BT-20 breast carcinoma ce11 lines.

88

PSA mRNA is undetectable in either unstimulated cells, cells stimulated with ethanol for

up to 48 h or cells stimulated with Norgestrel after 1 h post-stimulation. PSA mRNA becomes

detectable in the Norgestrel-stimulated cells at 2 h, its concentration increases at 4 h and it persists

for at least 48 h (Figure 6-4-1-1). However, this test is semiquantitative. Quantitative information

was provided by protein data. PSA protein is first detected in the ce11 cytoplasm 4 h post-

stimulation by Norgestrel and accumulates over the 48 h study period. PSA secreted into the

culture medium is first detected at 8 h and its concentration increases rapidly with time (Figure 6-40

1-2).

An identical experiment was perfomed using the steroid hormone receptor-negative and

positive breast carcinoma ce11 lines BT-20, MDA-MB-453, ZR-75- 1, and A-427 lung carcinoma

ce11 line. These ce11 lines did not produce any detectable PSA mRNA or protein after stimulation

with Norgestrel at the indicated time periods.

The identity of the PCR product was verïfied by complete sequencing of both strands.

Partial sequencing data are shown in Figure 6-4-1-3. The entire sequence is shown in Figure 6-4-

1-4. The sequence, spanning 6 16 nucleotides from exons 4, 5 and the 3'-untransiated region is

>99% homologous to the published sequence of PSA cDNA or genomic DNA. We found 1000/0

homology with the sequence published by Lundwall(187), Digby et al (5 1) and Klobeck et al (188).

There is one base difference (A to G) at position 439 (3'-untranslated region) of our

sequence and the sequence published by Lundwall and Lilja (39 , Shultz et al (1 79), Henmi and

Vihko (41), and Riegman et al (180). Also, at position 419 (3'-untranslated region) we and others

have identified G but a few other investigators have identified T. These differences are likely

polymorphisrns.

GCTCGGGTGA

ATCACGTCAT

GTACACCAAG

CWCCCCTG

7TGGAAATGA

GTCCTTAGGT

AGGTGTAGAC

TCCTGGGGAA

GGACACAGAT

AAGAGGGGTG

ACTGTCCATG

TCACAGCAAG

TGGWGCCT

ITCTGGGGGC

GGGGCAGTGA

GTGGTGCATT

AGCACCCCTA

CCAGGCCAAG

GTGAGGTCCA

CAGAGTGrrr

TACTGGCCAT

AGGAT GGGGT

GGATCCACAC

MGCACTGAG

GATGGAGCT G

AGAGAA

CCACftGTCT

ACCAT GTGCC

ACCGGAAGTG

TCAACTCCCT

ACTCAGGCCT

GGGITGCTAG

CTTAMTGGT

GCCTGGAGAC

GTCTGTGTTA

TGAGAGAGTG

CAGAAGCTGG

AAAAC ATAAC

GTAATGGT GT

CTGCCCGAAA

GATCAAGGAC

ARGTAGtAA

CCCCAGiTCT

GMMGAAAT

GTAArnGT

ATATCACT CA

r t f G f GGGGT

GAGAGTGACA

AGGCACAACG

CCACf CTGTC

GGCCTKCCT

ACCATCGTGG

ACITGGAACC

ACTGACCn-r

CAGCAGACAC

CCTCTCTGTG

AmCTCTGA

ACAGAGATGA

TGTGCTGGAC

CACCAGACAC

CTGGAGGCAC

Figure 6-4- 1-4. Complete sequence of the 6 16 nucleotide region of die PSA cDNA sequence from Norgestrel-stimulated T47D cells. For discussion see text. Bold and underhe indicate end of exon 4 and exon 5 , respectively.

9 1

Cornparison of Our 616 nucleotide sequence with the sequence of human glandular

kailikrein gene revealed only about 80% homology. These data confirm that the mRNA isolated

fiom Norgestrel stimulated T-47D and BT-474 cells is indeed the PSA rnRNA.

6-4-2. Induction of PSA Production by Steroid Hormones

In order to establish which of the compounds shown in Table 5-20-1-1 act as PSA gene

regulators, we stimulated T47D and BT-474 breast cancer cells and measured PSA in the tissue

culture supernatant at 3, 5 and 8 days post-single stimulation at a compound concentration of IO-' M.

The compounds were then qualitatively categorized into t hree groups; non-stimulators, weak

stimulators and strong stimulators. For this classification, we arbitrarily selected a scpematant PSA

concentration at 8 days of >ZOO ng/L, 10-200 ng/L or <IOng/L for strong, weak and non-

stimulators, respectively. Some stimulation data are shown graphicaily in Figure 6-4-24. Detailed

data are summarized in Table 6-4-2- 1.

6-4-3. Delermina~ion of Dose- Response Stimulatory Effects of Steroid Hormones

Dose-response experiments were designed to detemine the lowest concentration of

stimulating steroids which could still induce production of PSA. Representative dose-response

experiments are shown in Figure 6-4-3-1. The lowest concentration of stimulating steroids which

could induce PSA production is shown in Table 6-4-2-1. Among the steroids tested, the most

potent were three androgens (dihydroandrosterone, R1881, and dihydrotestosterone) and four

synthetic progestins (Norgestrel, Norethynodrel, R2050, and Depo-provera). These four steroids

couid induce PSA production at final concentrations down to IO-" M. Two other androgens

(testosterone and androsterone) and a synthetic progestin (Norgestimate) were potent at

Stimulation Compound

Figure 64-24. PSA concentration in T-47D ceil line tissue culture supematants d e r a single IO-' M stimulation with various compounds and sarnpling of the supematants at 3(.), 5(.) and 8(A) days post-stimulation. The compounds tested were: 1, testosterone; 2, estrone; 3, vitamin D; 4, corticosterone; 5 , dihydroandrosterone; 6, estriol; 7, 17a-hydroxyprogesterone; 8, androsterone; 9, hydrocortisone; 10, 1 7a-ethynylestradiol; 1 1, norethynodrel; 12, tamoxifen; 13, P-estradiol; 14, betarnethasone- 17-valerate; 15, norethidrone; 16, norgesuel; 1 7, aldosterone; 1 8, dexamethasone; 1 9, cholesterol; 20, Ri88 1; 21, R5020; 22, 23, no stimulation; 24, aicohol. in ali cases the most drarnatic change in PSA concentration occurs in the interval between 3-6 days. The strongest stimulators are androgens (1,5,8,20,21) and progestlis (1 l,l5,l6). More data are given in Table 642-2 .

93 Table 6-4-2-1. PSA concentration in T-47D ce11 Iine tissue culture supernatant after stimulation with various compounds

1

VaHydraxvprogestm Progesteme peairsor Nothing

f+ '%Wtae Rogesîm Weak

8e tamethasone 1 ?-valerat e

Dexarnetha~orie

Prednisone

G l u c a u k n d

Glucocwticoid

Glucocorticoid

- - . - - -

NOrgestrel

DepuPrwera

Norgesh'mate

Aldosteme

RU 58,668 1 Antjeserogen I Nothing I

- - -

s m g

S W

Ming

- -

Progestin Progestin

Triamiandone acetonde

TamOufen

- p. - - -

ND

10.'

Progestin I Strong Mineraloconicoid Weak

1 Antirnineralocorticoid 1 Weak 1 IO-'

- - - -- -

SLrorig

s m l IOW

10"

Progestin/giUOOCOrtjcoid

Anliesîrogen

Casodex (Cl 176.334)

RUS6.187

Nitutmide (Anandron)

Mdepistone (Rü486iRU38.486)

Cuîexdane

1. For definitions. see lexi. 2. ND; not done.

r 10"

rl(rl'

Str0ng

Noihing

Antiandrogen

An fiandmgen

Antiandrogen

Antiprogestin

A n t i g l u m ï d

IDw

Noîhing

Weak

Nothing

Weak

Weak

lo4

l@

10'

Figure 64-11. Dose-response experiments of six representative steroids. For more details and discussion see text.

95

concentrations down to lO-'OM. 17a-Ethinyle~radiol, unlike to al1 other estrogens tested, was able

to positively regulate PS A production but only at relatively high concentrations ( 1 0"- 1 0' M).

Arnong the glucocorticoids, corticosterone was a weak stimulator while betarnethasone and

dexarnethasone were strong stimulators but only at concentrations >IO-' M. Aldosterone was a

weak stimulator but the synthetic compound triamicinolone acetonide (TA) was a strong

stimulator, acting at low concentrations (lO*"' M). None of the antiestrogens exhibited stimulatory

activity. Among the antiandrogens, some did not induce any PSA production, one was a weak

stimulator (RU 56,187) and one (cyproterone acetate) was a strong stimulator, acting at

concentrations as low as 10-'O M. The antiprogestin Mifepristone (RU 38,486; RU 486) as well as

the antiglucocorticoid cortexolone and the antimineralocorticoid spironolactone were weak

stimulators, acting at concentrations 1 O"- 1 0-'M. The androgen rnetabolite dihydroisoandrosterone

sulfate and the progesterone precursor 17a-hydroxyprogesterone were inactive. Dose-response

experiments with the estrogens estrone and estradiol have s h o w that these compounds remain

inactive with respect to PSA gene up-regdation at any concentration between 1 O-'- 1 O-'' M.

Stimulation expenments were funher conducted using the steroid hormone receptor-

negative breast carcinoma ce11 lines BT-20, HBL- 100, and the steroid hormone receptor-positive

breast carcinoma cell iines MCF-7,ZR-75- 1, MDA-MB-453, and MFM-23 3 . The sarne experirnent

was carried out using the steroid hormone receptor positive cell lines SAOS (osteosarcorna), MFE-

296 ( endometnum) and BG-1 (ovarian carcinoma) and the unknown receptor status of ce11 lines

HBT- 16 1 (ovaian), HBT- 16 1 (ovarian), HBT- 173 (ovarian), A-427 (lung), SK-MES- 1 (lung),

NCLH460 (lung), MIA PaCa (pancreas), and Co10 320 (colon). None of the cornpounds listed in

Table 5-20-2-1 was able to induce detectable PSA protein production.

96

6 - 4 4 Blochg Steroid Hormone Receptors

In order to fbnher elucidate the mechanism of regdation of PSA production by steroid

hormones, we have conducted blocking experiments. In these studies, we have first treated the T-

47D and BT-474 cells with a steroid hormone receptor blocker for 1 h followed by the addition of

a 10-fold lower concentration of a stimulating steroid. The data are presented in Table 6-4-44.

Estradiol, as well as the antiandrogens Nilutamide, RU56,187 and hydroxyflutamide, were

able to significantly block the stimulating action of dihydrotestosterone in T-47D but not as strong

as in BT-474 celi line. Mifepristone was also a potent blocker in T-47D and a weak blocker in BT-

474 ; no blocking activity was seen arnong the antiestrogen ICI 182,780 the antimineralocorticoid

spironolactone and the antiglucocorticoid cortexolone in T-47D but only ICI 182, 780 had a weak

blocking effect on dihydrotestosterone in BT-474 ce11 line.

The stimulatory activity of Norgestrel was blocked minimally by estradiol, Nilutamide,

hydroxyflutamide and ICI 182,780 and to a higher degree by RU 56, 187 in T-47D. However,

arnong those blocker only RU 56.187 was able to block a fesser degree in BT-474. The most potent

blocker of Norgestrel's action was the antiprogestin Mifepristone (blocking 90- 100%) in both cell

lines. The stimulatory activity of the highly selective progestin agonist Norgestimate was only

blocked by the antiprogestin Mifepnstone (blocking 100%).

The stimulatory activity of dextarnethasone and aldo sterone were not bloc ked by estradiol

as well as Nilutamide, hydroxyflutarnide, RU56,187, Mifeprktone, ICI 182, 780, spironolactone,

and cortexolone in BT-474 breast carcinoma ceIl Iine.

Table 6 - 4 4 2 . Blocking of PSA Production in T-47D ceiis by Various Compounddl)

Stimularing Compound (1 0%)

Blochg Cornpound ( 10SM)

Estradio1

Nilutamide

RU 56,187

Hyiroyflutamrde

ICI 182,780

Mifepnstone

Spironolactone

Cortexolone

1. For detaiied protocol refer to Methods section.

7 . Range of 3 dfierent experiments.

98

6-5. Variation of PSA Protein Dunng the Menstmai Cycle

We collected approximately 8-10 serum sarnples from each of the 14 normal fernale

volunteers during the foilicular, mid cycle and luteal phase of the menstrual cycle. PSA protein was

measured in these semm sarnpies using a highly sensitive tirne-resolved immunofluorometnc

technique (1 8 1). This assay has within-run coefficients of variation of < 10% at PSA levels around

1 ng/L or higher. AI1 samples from the same patient were measured in duplicate or tnplicate in the

same assay mn to minimize variability. The PSA concentration patterns were verified in at least

3 different assay runs per patient and the results were consistent in al1 cases.

Anaiysis for PSA reveaied that in 11 patients the PSA levels were below 4 ng/L in dl

samples coliected. Since the detection lirnit of the assay used is 1-2 ng/L, we did not use these data

for further analysis. Among the three volunteers whose senim PSA levels were >4 n g L , we

collected additional sera from two. These new sera spanned two consequtive menstruai cycles and

were collected at least two months after the sera f?om initial collection. Thus, in total, we studied

7 informative menstrual cycles From three different patients. The days of the menstrual cycle in

each case was verified by analysis of progesterone, estradiol, LH, and FSH.

In Figure 6-51, we present the semm PSA changes during the menstrual cycle of patient

M.T. (3 cycles) and in Figure 6-5-2 the data for patient C.J. (3 cycles) and L.S. (1 cycle). In the

sarne graph, we present the progesterone values as well. When we plotted the PSA data along with

data for LH, FSH, or estradiol, no recognizable pattern or relationship was seen. The data of

Figures 6-54 and 6-5-2 show a consistent pattern with PSA peaking during the rnid-late follicular

phase and reaching a minimum during the rnid-Iate luteal phase. The difference between PSA and

progesterone peaks is about 10-20 days.

O 1 O 20 30

Days of Menstrual Cycle

O 10 20 30 40 SO 6 0

Days of Menstrual Cycle

Figure 6-51. S e m prostate specific antigen (m) and progesterone (+) Ievels during the menstnial cycle ofvolunteer M.T. A One rnaistnial cyde. B. Sera fiom two additionai consecutive menstrual cycles coilected at least two months after the sera of panel A. Notice the peak of PSA which foiiows the progesterone peak with a delay of 10-20 days.

O 1 O 20 10

Dayr of Mrnstrual Cycle

O t

O 1 O 20 3 O O

Days of Menstrual Cycle

Oays of Menstrual Cycfe

Figure 6-5-2. S e m prostate specific antigen (@) and progesterone (*) levels during the menstrual cycle of volunteers C. J. and L. S. A. One rnenstrual cycle of volunteer C. J. B. One menstruai cycle of volunteer L. S. C .Sera f?om h ~ o additional consecutive menstmal cycles of volunteer C.J. coilected at least two months &er the sera o f panel A Other cornrnents as in Figure 6-54.

101

We have further exarnined if semm obtained during the menstrual cycle has the ability to

stimulate PSA production in the breast carcinoma ce11 line T-47D. This expenment was performed

by measuring either PSA rnRNA or protein. We found that non stimulated T-47D cells do not

produce PSA protein and do not express PSA mRNA. When the cells are stimulated with serum

obtained during the menstnial cycle, the results of Figure 6-5-3 are obtained. The ability of the

semm to induce PSA production parallels the levels of progesterone. Maximal stimulation is

achieved with the semm that contains the maximal progesterone concentration (day 24 of the

menstnial cycle).

PSA mRNA in T-47D cells can be detected in trace amounts if the serum used for

stimulation is collected during the follicular phase of the cycle. When the serum is collected during

the iuteal phase, PSA mRNA expression is dramatically increased (Figure 6-5-4). These changes

in PSA mRNA levels parallel the changes in progesterone levels. PSA mRNA is undetectable in

T-47D cells which are not stimulated by serum.

We have further stirnulated T-47D cells with the glycoprotein hormones LH, FSH, hCG and

with prolactin and growth hormone. Despite the wide range of concentrations used, none was able

to induce PSA production. Progesterone was able to induce T-47D cells for PSA protein production

and PSA mRNA expression at concentrations between 1 O-'- 1 0-l0 M-

6-6. Expression of the PSA Gene in Lung Tissue

The tumor extract from the primary lung adenocarcinorna was measured for PSA protein;

the PSA concentration obtained was 958 ngL and was considered highly positive for PSA

immunoreactivity . In contrast, another 1 0 lung adenocarcinornas from women, when extracted,

O 1 O 20 3 0

Days

Figure 6-5-3. Serum progesterone levels during the rnenstnial cycle of a nomaliy cychg volunteer (m). These sera were used to stimulate PSA production in the breast carcinoma ce11 line T-47D as dûcnbed under methods. The PSA concentration in the asnie culture supematants is shown (+). Notice that sera which contain high levels of progesterone are able to induce PSA production in the c d h e T-47D. Data on PSA rnRNA expression are presented in Figure 6-5-4.

Figure 6-5-4. Reverse transcription polymerase chah reaction of PSA mRNA (A,B) and actin (C) obtahed from T47D celis stimdated with semm of volunteer MM. The serum was drawn at days 5 (lane 2). 7 (lane 3), 10 (lane 4), 14 (lane 9, 16 (lane 6), 21 (lane 7), 24 (lane 8) and 28 (lane 9) of the mensaual cycle. Lane 10 represents a positive PCR control (plasmid containhg full-length cDNA for PSA) and lane 1 1 is the negtive PCR control (no target). Notice the intense PCR bands produced when s e m Eorn the luteal phase of the cycle is used to stimulate T-47D ceils. In panel A the PCR product was detected with cherniluminescence; In panels B and C with ethidium bromide staining of agarose gels.

1 O4

exhibited PSA concentrations <20 n@. Contamination of the tumor extract by blood PSA is

unlikeiy since the PSA concentration in the serum of this woman was 4 ng/L preoperatively and 5

ng/L postoperatively.

In order to estabiish the molecular weight of the immunoreactive species in the tumor

extract, we performed high performance iiquid chromatography with a gel filtration colurnn (Figure

6-44). Virtuaily 100% of the immunoreactive species elutes at fraction 40, corresponding to a

molecular weight of approximately 33 KDa (free PSA). No complex between PSA and a,-

antichymotrypsin (molecular weight - 100 KDa) was seen.

Total RNA exrracted from the lung tissue was reverse-transcnbed to cDNA and then

arnplified by PCR using pnmers specific for the PSA cDNA sequence. Southern blot hybndization

of the PCR product with a PSA DIG-labelled cRNA probe detected the PSA band of the expected

size (754-bp) in the sarnple but not in the negative control (Figure 6-6-2). This band was

reamplified (to obtain enough material) and sequenced. Both strands were sequenced and the data

were compared with the published PSA and glandular kallikrein cDNA sequences.

Partial sequencing data are shown in Figure 6-6-3. The entire sequence obtained is

presented in Figure 6-6-4. The sequence, spanning 549 nucleotides from exons 4, 5 and the 3'-

untramlateci region has >99% sequence similarity to the published sequence of PSA cDNA or genomic

DNA (Table 6-64). However, there is only about 80% similarity with the human glandular kallikrein

gene in accordance with literature reports (22).

We found four positions where there is variation between Our sequence and those reported in

the literahire. These ciifferences, alI present in the 3' untranslated region. may represent polymorphisms.

At two positions (position 427 and 447) the variation appears in the same spot (underlined) of the

Free PSA

2 5 SS 4 S

Fraction Number

Figure 6-64 Molecular weight identification of PSA irnmunoreactivity in the lung tissue exuaa ushg HPLC with a gel filtration column. Mer injecting 500 PL of tumor extract, HPLC &actions (0.5 mL) were analyzed with an assay that recognizes both free PSA (33 KDa) and PSA-ACT complex (100 KDa). Al1 PSA immunoreadvity in the lung tumor extract elutes at a molecular weight of 33 KDa (fraction 40 + 2) and represents f i e PSA PSA-ACT complexes (eluthg at fiaction 3222) were not found.

Figure 6-6-2. Reverse transcription PCR of RNA extracted fiom various tissues. Panel A. The PCR products were run on an agarose gel, stained with ethidium brornide. Lane 1. Molecular weight markers with length in base pairs shown on the left. Lane 2. Digoxigenin-labeled DNA markers (not stained). Lanes 3 and 4. RT-PCR product for the patient described here; no bands are visible with this method. Lane 5. RT-PCR produa for a patient with a PSA-positive breast turnor (positive control). Lane 6 . RT-PCR product for a patient with a PSA-negative breast turnor (negative control). Lane 7. Amplification of 9,000 m o l d e s of a plasrnid containing the f ù U sequence of PSA cDNA (positive control). Lane 8. PCR negative control (no target added). Lane 9. PCR for actin cDNA for the patient described here, venfjmg the success of the reverse transcription. In lanes 5 and 7, the PCR produa is 754 bp and in iane 9 is 372 bp. Panel B. PCR products were run on an agarose gel, Southem-blotted and hybridized with a DIG-labeled RNA probe. Detection of hybrids was achieved with cherniluminescence. Lane 1. DIG labeled molecular weight markers with length in base pairs show on the lefi. Lane 2. RT-PCR produa for the patient described in this paper. The 754 bp PCR product is clearly identified dong with a higher rnolecular weight produa (genomic DNA). Lane 3. PCR negative control.

AAGCACCTGC

TTCAAGGTAT

CCTCCCTGT

CATCGTGGCC

lTGGAACCIT

TGACCllTGT

GCAGACACAG

TCTCTGTGTC

TTCTCTGAGG

AGAGATGAAA

TGCTGGACAC

TCGGGTGATT

CACGTCATGG

ACACCAAGGT

AACCCCTGqG

GGAAATGACC

CCTTAGGTGT

GTGTAGACCA

CTGGGGAATA

ACACAGATAG

GAGGGGTGGG

TGTCCATGAA

CTGGGGGCCC

GGCAGT GAAC

GGTGCAlTAC

CACCCCTATC

AGGCCAAGAC

GAGGTCCAGG

GAGTGrrrCT

CTGGCCATGC

GATGGGGTGT - ATCCACACTG

GCACTGAGCA

ACTTGTCTGT

CATGTGCCCT

CGGAAGTGGA

AACICCCTAT

T CAGGCCTCC

GUGCTAGGA

TAAATGGTGT

CTGGAGACAT

CTGTGlTATT

AGAGAGTGGA

GAAGCTGGAG

AATGGTGTGC

GCCCGAAAGG

TCAAGGACAC

TGTAGTAAAC

CCAGlTCTAC

AAAGAAATCA

A A r n G T C C

ATCACTCAAT

TGTGGGGTAC

GAGTGACATG

GCACAACGC

Figure 6-6-4. Complete sequence of the 549 nucleotide region of the PSA cDNA sequence from the mRNA isolated from the lung tissue. Bold and underline indicate end of exon 4 and exon 5, respectively. Underiined sequences represent a six base repeat in which one nucleotide (double underiined) is fiequently polymorphic (Table 1). Sequence variation was also seen at positions 234 and 184 (underhed).

Table 6-61. C o m p ~ s o n of DNA sequence of PCR product found in this mdy with sequence deposited in GenBank

GenBank Accession Positiodl)

DNA 427 447 234 184

This Study (Lung Tissue) cDNA

cDNA

cDNA

Genomic

cDNA

cDNA

Genomic

cDNA

Genomic

1. Refcn to nucleotide nurnbcr shown in Figure 6-64. Yucleotides not shown are identical in al1 sequencts

Figure 6-64. Immunohistochemical localization of PSA in lung tissue with polyclod antibody. A: The lung tumor does not contain PSA imrnunoreactivity in epithelial neoplastic ceiis or in the stroma1 component. B: A trapped non-neoplastic bronchiole at the edge of the m o r contains PSA imrnunoreactivity in epitheliai lining ceus. The cytoplasmic positivity is arong but restricted to scattered ceus (arrowheads). Tumor ceiis (T) around the bronchiole are negative. (Streptavidin-biotin technique; ~ 2 5 0 ) .

1 1 1

sequence repeat T G G E T (Figure M).

We localized the PSA immunoreactivity with immunohistochemistry (Figure 6-6-5). While

the lung tumor was entirely negative for P S 4 trappeci non-tumorous bronchioles at the edge of the

tumor contained focal PSA immunoreactivity in epithelial lining ceiis.

Since the PSA gene is regulated by steroid hormones (96), we examineci if the only steroid

administered to the patient durùig hospitalization (beclomethasone dipropionate) was able to stimulate

PSA production in a tissue culture system. The breast carcinoma ceii lime BT-474 does not produce

PSA when cultureci in the absence of steroid hormones but it does so when stimuiated by androgens,

progestins and glucocorticoids. Beclomethasone dipropionate was able to stimulate PSA production

in this ce11 line in a dose-dependent manner at levels 10"- 10" moi/L (Figure 6-6-6).

6-7. Frequency of PSA mRNA Expression in Lung Tumon

Due to the marked differences between semm PSA concentrations in males and females

(males have approximately 500-fold higher levels), a11 data were analyzed separately between males

(N=3 5) and females (N=17). PSA levels in tumor cytosols were expressed as ng of PSA per g of

total protein to normaiize for the arnount of tissue extracted. The average total protein

concentration of these extracts was -1 g/L. Any level of PSA 4 1 ng/g or < 1 ng/L was considered

as non-detectable since it is below the detection limit of the PSA assay used (50).

The distribution of lung tumor PSA content between males and females is shown in Table

6-7- 1. In the sarne Table, we summarize data of serum PSA in these male and female patients.

Both pre-operative (-5 days before surgery) and post-operative (-5 days post-surgery), PSA levels

were measured. The following comments apply: (a) Levels of PSA in tissue extracts fkom males

Figure 6-66. Production of PSA by the breast carcinoma ceIl line BT-474 der stimulation by beclomethasone dipropionate and other steroids. PSA was measured in the tissue culture supernatant 5 days post-stimulation of confluent celIs with a simgle dose. 1 . Norgestrel (progestin). 2. Beta- methasone (glucocorticoid). 3. Dexamethasone (glucocorticoid). 4. BecIomethasone (glucocorticoid). 5 . Prednisone (glucocorticoid). 6. Estradio1 (estrogen). Dosedependent PSA production is seen at beclomethasone levels between 1 O-'- 1 0.7 moVL.

Table 6 - 7 4 Distribution of PSA in Lung Tumor Cytosols and in preoprative and Postoperative Sera fiom Male (M) and Fernale (F) patients.

Percentile

Tumor Cytosols (ng/g)

Preoperative Semm (ng/L)

Postoperative Senim (ng/L)

114

are higher than those in females. (b) We found no consistent changes between pre-operative and

post-operative semm PSA in either males or females. (c) Analysis of normal lung tissue extracts

from 3 males and 4 females has revealed that PSA was also detectable at levels 1.9 ng/g (median)

for males and 2.9 ng/g (median) for femaies.

We have examined if the PSA levels in the tumor cytosoIs were associated with the semm

levels in both males and femaies. The Spearman correlation coefficient between tumor cytosol PSA

and preoperative serum PSA in males was 0.80 (p=0.000 1) and in femaies was 0.53 (p=0.05). The

Spearman correlation coefficient between tumor cytosol PSA and postoperative semm PSA in

males was 0.5 1 (p=0.0 1) and in femaies was 0.37 (p=O. 18). These data suggest that there is a

significant correlation between tumor levels of PSA and serum P S 4 at least in male patients.

In Our series, only four patients died during the follow-up of 6-1 2 months. For this reason,

and due to the smdl number of patients, no survival anaiysis in relation to PSA levels in the tumors

was performed.

PSA mRNA was detected in lung tumors with RT-PCR and Southern blot hybridization.

From the 35 lung tumors from males, 24 (69%) were positive for PSA rnRNA. From the 17 lung

tumors from females, 9 (53%) were positive for PSA mRNA Of the 7 adjacent normal lung tissues,

five (71%) were dso positive for PSA mRNA. Examples of PSA mRNA detection in lung tumors

are shown in Figures 6-7-1 and 11-7-2. In addition to the expected 754 bp PCR product, we

occasionally detected bands of molecular weights between 250-500 bp (Figure 6-70 2). The

identity of these bands is discussed below. Sequencing of the 754 bp PCR product from one female

patient with iung cancer confirmed the 100% identity with the published sequence of PSA cDNA

(187).

Figure 6-7-1. A: RT-PCR for PSA mRNA extracted from lung tumors. B: RT-PCR for actin mRNA extracted fiom lung tumors. Lane 1. Molecular weight markers. Sizes in base pairs (bp) are shown in lefi panel. Lanes 2-1 1 . Lung tumors. Lane 12. PCR negative control (no template added). In A, the PCR products were detected by Southem hybridization and cherniluminescence. In B, the PCR products were detected by ethidium bromide staining. The 754 bp PCR product between molecuiar weight markers 900 and 692 bp corresponds to PSA. The 372 bp PCR product (in panel B) corresponds to actin. Tumors in lanes 2, 4, 5, 7, 8, 9 and 10 are positive for PSA rnRNA.

Figure 6-7-2. A: RT-PCR for PSA mRNA extracted from lung tumors. B: RT-PCR for actin mRNA extracted fiom lung tumors. Lane 1 . Molecular weight markers. Sizes in bp are s h o w in Iefl panel. Lanes 2-5. Lung tumors. Other cornments as in Figure 14-7- 1. Aberrant PCR bands appear in lanes 4 and 5 o f panel A Other aberrant bands appeared at molecular weights of 250-500 bp (not shown).

117

The association between PSA protein and mRNA expression in lung tumors and other

clinicopathologicai variables was examined by contingency tables and chi-square statistical

analysis. In males, using as cutoff the median PSA level in lung tumors (expressed as ng/g of total

protein), we found no association between PSA protein and presence or absence of mRNA

(p4 .33) . patient age @=0.88), stage (p=0.24), grade (p=0.36), histological type (p=0.67) or nodal

status (p=0.23). The only statistically significant association was with preoperative (p-0.003) and

postoperative (p4.006) serum PSA, an association also revealed by Spearman correlation analysis

(already described). In females, no significant association was found between tumor PSA levels

and any of the above clinicopathological variables.

We have fùrther examined if the presence of PSA mRNA in lung tumors is associated with

any of the clinicopathologicai variables. In males or females, the presence of PSA M A was not

associated with patient age, stage, grade, pre-surgical or post-surgical serum PSA or nodal status.

We found some evidence that the presence of mRNA in males is associated with squamous ce11

carcinomas. From the I I tumors which were PSA mRNA-negative, four (36%) were squarnous

cell, six (55%) were adenocarcinomas and one (9%) was of other histologic type. From the 24 PS A

mRNA positive tumors, seventeen (7 1 %) were squamous cell, four (1 7%) were adenocarcinomas

and three (13%) were of other histologic type (p=0.07).

In Figure 6-7-2, we have shown that we occasionally detected aberrant PCR products of

rnolecular weights between 250-500 bp. We have isolated and cloned six of these products and

subsequently sequenced al1 of them using vector-specific sequencing primers. Three of these

products proved to be primer-dimer artifacts which hybndized to the probe due to sequence

complementarity between the PCR primers and our riboprobe which contains the primers in its

TGCCAATGGG

T C i G A C A

AACI'TAAAAA

TGGATCTTCA

GGGAGAACTG

CAAGACATGG

T GITCTGCAA

ACTGCAACTA

GGrrrAAGAA

TiATCCAGCA

TGAATGTAGT

TAC CTACAAA

GAATACTGAA

ITCTAAGGTG

AAAAAAACAT

TGCTAGGTGT

GAAITGT GGT

CTAAGCAGTA

AGAAGCAGAT

TGAAATCCAT

ATCCAGGGAG

TCCCAAAm

ACCCAGGAAT

TAAAATGGAT

GCCClTGTAG

CCCTGGTCAT

TGTGTAAGCT

CAGAATATAG GMATATCAC

m A A G T T A GTTAAGGACT

ACTGAClTCA GTAlTïGTCA

CAATACTCGC TTCTCAAACA

GCCAGCCATA CAAATATGCT

TTGCTGCAAC TGATAITCTG

TCCAAGAACT AACACTGCAG

GTCACTTCCA AGAAGTATCA

GAAAAGGGAT TAAGAGTGAG

Figure 6-7-3. Sequence of a 450 bp PCR product detected in addition to the 754 bp PSA PCR product in some lung tumors. This sequence has no homology to any other sequence deposited in GenBank.

118

sequence. The other three aberrant PCR products represented the same sequence consisting of 450

bp (Figure 6-7-3). This PCR produa was created only by primer PSA-A2 since the complete

sequence of this PCR product had in one end the primer PSA-A2 and at the other end of the same

strand a sequence complementary to PSA-A2. The 450 bp sequence, venfied by sequencing of both

strands, had no sequence hornology to PSA cDN4 kallikrein cDNA or any other sequence

deposited in GenBank as revealed by BLAST or DNASIS software analysis.

CHAPTER 7. Discussion

7-1. Detection of PSA Protein and mRNA in Breast Tumors

Prostate specific antiçen (PSA) is a senne protease found at very high concentrations in

seminal plasma. It has been suggested that PSA is involved in semen liquefaction post-ejaculation.

PSA is considered a highiy specific biochemical marker of the prostate gland and is currently used

for diagnosis, prognosis, and management of patients with prostate cancer (1-3).

Our group has recently dernonstrated that approximately 30% of fernale breast tumors

produce a 33 D a protein with stnking sirnilarities to seminal plasma PSA (4-6). This

immunoreactive PSA was shown to be associated with the presence of steroid hormone receptors,

earlier disease stage, and younger patient age (6). Preliminary data suggest that PSA may be a new

favorable pronostic indicator in breast cancer (12).Irnrnunoreactive PSA has been detected in

various tumors including ovarian, colon, h g , and parotid tumors. However, PSA expression in

these tumors is infiequent and occurs at much lower levels than in breast tumors (7). Recently, it

has been dernonstrated that PSA mRNA isolated from a few PSA positive breast tumors had an

identical sequence to the PSA cDNA derived fiorn prostate tissue (95).

Polyrnerase chah reactraction (PCR) is a very sensitive procedure for in-vitro amplification

of DNA sequences. PCR has gained widespread acceptance in many areas of molecular biology

particularly in tumor biology, clinical medicine, and more recently in routine diagnostic

applications. RT-PCR-based methods for analyzing rnRNA expression specific for neoplastic ceils

have been applied for detecting micrometastasis or minimal residual disease (1 82,189).

120

In this study, 1 have developed sensitive RT-PCR methods for detecting PSA rnRNA.

Among the variations tested, the nested primer assay was the most sensitive but this method is

difficult to control for contamination and was not used to assay the breast tumors. The simple PCR

assay in which PCR products are detected by ethidium bromide staining was also found inadequate

since many PSA protein-positive tumors tested negative for PSA mRNA with this method. The

other two modifications in which the PCR products are detected with cherniluminescence with use

of digoxigenin as label were found to be satisfactory. Data generated with these two methods were

in good agreement with data fiom the analysis o f PSA protein with a highly sensitive immunoassay.

The observed few discrepancies are likely due to tumor heterogeneity since in PSA-positive tumors,

not ail tumor cells produce PSA; the immunoreactivity is focal and restricted to clusters of cells as

detected by imrnunohistochernical localization. Tumor heterogeneity is likely the reason why we

did not observe a consistent correlation between levels of PSA protein and intensity of the PCR

bands; two different tumor pieces were used for cytosol extraction and PSA protein analysis and

for rnRNA extraction and PCR.

In the last few years, rnany other groups have developed sensitive PCR procedures for PSA

mRNA. Deguchi et al. (1 82) developed a PCR assay with sensitivity between O. 1- 1 LNCaP ceIl

and used it to detect prostatic carcinoma turnor cells invading the regional lymph nodes. Moreno

et al. (190) developed a PCR assay of unspecified sensitivity and used it to detect prostate cancer

cells in the circulation of patients with advanced cancer. Katz et al. (19 1) developed a PCR assay

for PSA mRNA which incorporates DIG- 1 1 - d W in the reaction products and demonstrated that

they could detect 1 LNCaP ce1 admixed with 100,000 human blood lymphocytes. When they used

their method to assay human blood, they found that 78% of prostate cancer patients with metastatic

121

disease and 39% with localized disease had cancer cells in the circulation. More recently, Jaakkola

et al. (192) have developed a nested primer PCR assay with the ability to detect -2 LNCaP cells

admixed with 106 leukocytes. These authors found that PSA mRNA could not be detected in the

blood of patients with benign prostatic hyperplasia, non-prostatic types of cancer or in patients with

localized prostate cancer; 50% of patients with metastatic disease tested positive.

A universal finding arnong these midies (190- 192) is that none of the numerous bloods From

normal males or females tested positive for PSA mRNA. However, Smith et al. (1 93) reported

100% positivity arnong 7 normal males and 6 normal female blood samples, using a highly sensitive

nested primer protocol. Since nested-primer procedures are prone to contamination, we recommend

caution before concluding that blood samples from any male or female contain traces of PSA

mRNA until the data by Smith et al. are reproduced by others.

The data presented here support the view that assessrnent of PSA production by breast

tumors c m be studied by either highly sensitive immunoassays for PSA protein (1 82) or by highly

sensitive RT-PCR assays, like those described here for PSA mRNA. We anticipate that the RT-

PCR-based assays will be usehl to assess PSA mRNA levels in tumors other than those of the

breast and in non-prostatic tumors and tissues from males. Although the PSA protein assay is

extremely sensitive, it suffers from the major limitation that male blood contains significant

amounts of PSA which may contaminate tumor or tissue extracts from males. Detection of PSA

mRNA would effectively circumvent this potential problem.

7-2. Steroid Hormone Regulation of PSA Gene in Breast Cancer

The PSA gene is known to be regulated by androgens in the male prostate (194,195). The

122

epitheliai cells of the prostate gland are rich in AR, some stromai cells aiso contain AR, as well as

the enzyme 5 a-reductase which reduces testosterone to dihydrotestosterone ( 1 94). The PS A gene

has an HRE to which the activated AR can bind (45,188,195-197). The PSA gene is up-regulated

by androgens and androgen agonists and is down-regulated by antiandrogens.

We have recently shown that female breast tissue and breast secretions contain high levels

of PSA (198). Although the physiological role of this protein in the female breast is still unknown.

we have demonstrated that the presence of PSA is strongly associated with presence of steroid

hormone receptors (6) . I have thus postulated that the PSA gene in the fernale breast is regulated

by steroid hormones. In this study, I have developed a tissue culture syaem to fùrther examine this

regulation and study the involvement of the various steroid hormone receptors.

We have first shown that the steroid hormone receptor-positive breast carcinoma ce11 Iine

T-47D is capable of producing PSA under appropriate stimulation by steroid hormones. T-47D

cells as well as MCF-7 cells do not produce PSA in the absence of steroid hormones (96,193). The

PSA mRNA produced by T47D and BT474 cells is identical to the sequence of PSA mRNA from

prostate celis. In contrast, the breast carcinoma ceIl line BT-20, which is devoid of steroid hormone

receptors did not produce PSA after stimulation by any of the compounds listed in Table 5-20-1-1.

1 have thus postulated that PSA production by breast cells is dependent on the steroid

honnone/steroid hormone receptor system. I fùrther demonstrated that the receptors and hormones

are necessary but not sufficient for PSA production. When I stirnuiated the steroid hormone

receptor-positive ceil Lines SAOS (osteosarcoma) and BG- 1 (ovarian carcinoma), ZR-75- 1, MCF-7,

(breast carcinoma cell lines) with the compounds show in Table 5-20-1-1, none was able to induce

PSA production. The presence of estrogen and progesterone receptors in these ceIl lines was

123

confirmed by aiiysis with estabtished enzyme immunoassay kits (Table 6-4-1). Apparently, either

post-receptor defects are present in these ce11 lines, the receptors are defective or the promoter of

the PSA gene in these ce11 lines is tissue specific. These possibilities were not studied further.

Among al1 androgenic compounds tested, only dihydroisoandrosterone sulfate, an inactive

metabolite, was not able to stimulate PSA production. Al1 other androgens were strong stimulators

(Table 6-4-2-1). The physiologicai androgens testosterone and androsterone and their reduced

forms dihydrotestosterone and dihydroandrosterone were able to induce PSA production at levels

as low as 10-'O and IO-'' M. respectively. The lower active concentration of dihydrotestosterone

(and dihydroandrosterone) is in accord with its higher affinity for the androgen receptor than

testosterone ( 194). Strong stimulation was also observed with the synthetic compounds R 1 88 1 and

RS020. In al1 cases tested, we observed a dose-response relationship. Although there is always a

degree of cross-reaaivity of steroid hormones with receptors other than the cognate receptors. the

activity of androgens at levels around 1 O-'' M (a concentration 10- 100-fold lower than the affinity

constant of the testosterone-AR complex) strongiy suggests that the effect is mediated through high

affinity binding to the androgen receptor and not through low affinity binding to cross-reacting

receptors. Among the four estrogens tested, the three naîural estrogens estradiol, estrone and estnol

did not mediate any PSA production. These data suggest that the estrogen receptor is not involved

in PSA gene up-regdation in the breast carcinoma cell line T-47D and BT-474. 17a-

Ethinylestradiol, a synthetic estrogen, was a weak but consistent stimulator at concentrations 21 0'

M. suggesting that this steroid is not a pure estrogen. Our data suggest that this steroid interacts

with the androgen and/or the progesterone receptor leading to active complexes capable of weakly

up-regulating the PS A gene.

124

Arnong the group of glucocorticoids tested, the physiological glucocorticoid hydrocortisone

and the synthetic glucocorticoid prednisone had no effect. The strong induction of betarnethasone

and dexamethasone and the weak induction by corticosterone, (at concentrations ~ 1 0 . ~ M) dl of

which have higher affinities for the glucocorticoid receptor than hydrocortisone and cortisone and

do not bind to either AR or PR (199,) suggest that the glucocorticoid receptor is capable of

regulating the PSA gene as well. The PSA stimulation at high glucocorticoid concentrations ody,

may reflect the low concentration of this receptor in T47D cells (200,20 1).

Wit h the exception of the inactive progesterone precursor 1 7 a -hydroxyprogesterone, al1

other progesterone agonists tested, were strong stimulators of PSA production. In dose-response

experiments, we have shown that the three tested progestins were active at levels 1 O-"- 1 O-" M. In

particular, Norgestimate, which exhibits highly specific high affinity binding to the progesterone

receptor than other progestin agonists (202-204) and binds to androgen receptor very poorly, was

active at levels down to 10 - '~ M. The data presented for the progestin agonists, in combination with

data of blocking experiments (discussed below) strongly suggest that the progesterone receptor,

activated by progestin, is capable of directly up-regulating the PSA gene.

In males, PSA gene regulation is under the control of testicular androgens through the

androgen receptor. 1 speculate that in fernales, PSA gene regulation in organs like the breast and

the endometnum (33) is mediated by progestins and androgens through the independent action of

the progesterone and the androgen receptor.

Aidosterone, a natural mineralocorticoid, was capable of PSA regulation only weakly and

at concentrations > 1 om8 M.

Triarnicinolone acetonide, a compound known to interact with the PR and GR but not the

125

androgen receptor ( 199,205) was found to strongly stimulate PSA production at concenaations as

low as IO-'' M. This finding M e r strengthens our suggestion that the progesterone and

glucocorticoid receptor c m mediate PS A production without involvement of the AR.

Among the four antiestrogens, none was able to mediate PSA production consistent with

the suggestion that the estrogen receptor is not involved in PSA gene up-regdation. Among the

group of antiandrogens, 1 observed some interesting phenornena. Al1 these compounds bind to the

androgen receptor leading to either inactive complexes (pure antiandrogens) or to complexes with

some biologicai stimulatory activity (antiandrogens with weak agonist activity). In our system,

hydroxytlutamide, Casodex and Nilutamide (Anandron) which are known to bind to the androgen

receptor with low atfinity ( ~ 2 % of testosterone affinity) (206) did not mediate any PSA production

suggesting formation of weak and inactive complexes with the androgen receptor. Cyproterone

acetate, which binds to the androgen receptor with affinity -10% of that of testosterone (206), was

found to be a strong stimulator of PSA production, with activity even at concentrations - 10"' M.

This data suggests that cyproterone acetate, a known antiandrogen which also interacts with the

progesterone receptor (194,206) and has biological progestational activity ( 1 94) exerts its action

on PSA replation through binding to the progesterone receptor. Our finding that cyproterone

acetate can up-regulate the PSA gene through the progesterone receptor in parallel to its expected

d o m regulation of the PSA gene through androgen receptor blockade, requires fùrther investigation

since monitoring PSA levels during cyproterone acetate treatment of prostate cancer may not be a

reliable index of clinical response. Up-regulation of the PSA gene by cyproterone acetate through

its progestational activity has not, to Our knowledge, as yet been reported. It is currently unknown

if this indeed happens in prostate cells in addition to breast cancer ce11 lines.

126

The newer antiandrogen RU56,187 has f i n i t y for the androgen receptor similar to

testosterone but no detectable afftnity for progesterone, glucocorticoid, mineralocorticoid or

estrogen receptors (206). In our system, 1 detected weak PSA gene up-regdation at RU56,187

concentration > 1 M. This up-regdation strongly suggests that RU56.187 has weak androgen

agonist activity. In this respect, Our tissue culture system appears to be more sensitive than the in-

vitro systems used by Teutsch et al to evaluate RU56,187, concluding that this compound is totally

devoid of binding to other steroid receptors and of any agonist effect (206). It remains to be

determined if the weak agonist activity of RU56,l87 has any biological significance.

Mifepristone (RU486, RU3 8,486) is a new antiprogestational agent with antiglucocorticoid

and antiandrogenic activity . RU486 has been commercialized as an antiprogestin for first trimester

pregnancy interruption. In Our system, 1 found that Mifepristone has weak agonist activity,

mediating PSA gene up-regulation at concentrations >IO-' M. This agonist activity was not

observed by Philibert et al (207), fbrther suggesting that their biological tests are not as sensitive

as our tissue culture system in detecting such an effect. In support of our data are reports by others

showing weak agonist activity of RU486 in various systems (208,209,99). It remains to be seen if

the weak agonist activity of Mifepristone is mediated by the androgen, glucocorticoid or the

progesterone receptor. Very weak agonist activity was dso observed for the antiglucocorticoid

cortexolone and the antimineralocorticoid spironolactone at concentrations 21 0.' M.

While estradiol has no positive effect on PSA gene regulation, blocking experiments have

revealed that estradiol could block the action of dihydrotestosterone and to a much lesser degree

Norgestrel but not Norgestimate on PSA gene regulation (Table 6-4-2-1). There are two possible

explanations for this phenornenon. First, high doses of estradiol could cause its binding to the

127

androgen receptor thus blocking the action of dihydrotestosterone (1 99,99). Second, estradiol

would bind to the estrogen receptor in T-47D cells and the active cornplex further inhibit the action

of active AR complexes. Active ER, AR and PR complexes compete for the same transcription

factors including c-jun and c-fos as suggested previously (210). The fact that estradiol blocks the

stimulation by dihydrotestosterone but not the stimulation by Norgestimate suggests that estradiol

blockade targets the AR but not the PR. The finding of positive regulation of the PSA gene by

androgen and progestin and the negative regulation by estrogen suggests that PSA is regulated by

a delicate balance between androgens, progestins and estrogens.

Nilutamide and hydroxyflutamide, two antiandrogens that bind with low affinity to the

androgen receptor, had moderate but not complete blocking activity (-30-40% on average) on

dihydrotestosterone and an even lower blocking activity ( 9 0 % ) on Norgestrel and no blocking

activity on Norgestirnate. This data are expected since the stimulating steroids (eg.

dihydrotestosterone), having higlier &nity for the androgen receptor, would displace a fraction

of the blocker after they are added into the tissue culture system. On the other hand, RU56.187

which has an affinity for the androgen receptor similar to testosterone, was able to block 85-9 1%

of the activity of dihydrotestosterone. The lower blockade on Norgestrel action (1 7-40% on

average) and the absence of blockade on Norgestimate action further suggests that a significant

portion of Norgestrel's and 100% of Norgestirnate's stimulation is mediated through the

progesterone receptor to which RU56,187 does not bind and could not block.

The antiestrogen ICI 182,780 had little or no blocking effect on the stimulation of PSA

production by dihydrotestosterone, Norgestrel or Norgestimate, in accordance with our view that

the estrogen receptor does not positively mediate PSA production in Our system.

128

Mifepnstone was an effective blocker of PSA production by dihydrotestosterone (70-80%)

and an almost complete blocker of Norgestrel and Norgestimate (90-1 00%). This is in accord with

Our view that PSA production is mediated independently by the AR and the PR since Mifepnstone

is known to block effectively the progesterone receptor and to a lesser but significant degree the

androgen receptor (207).

As expected. the antiglucocorticoid cortexoione and the antimineralocorticoid

spironolactone had no effect on either dihydrotestosterone, Norgestrel or Norgestimate action since

these two antihormones bind pnmarily to GR and MR and only with Iow afinity to other receptors

which are invoived in PSA production.

Stimulation experiments were also conducted using growth factors, vitamin D, analogs,

dl-tram and cis-retinoids (Table 6-20-1-1) in T-47D and BT-474 breast carcinoma ce11 lines, but

none of thern was able to induce PSA production. However, recently it has been reported that in

addition of steroid hormones, growth factors, IGF-1, KGF, EGF, were thought to participate in the

regulation of PSA gene expression through the androgen receptor activation in prostatic tumour ce11

line (102). In addition to traditional steroid hormones, four growth factors IGF-1, FGF, TGF, and

EGF affect breast cancer and treatment with trans-retinoids (analog of vitamin A) slows tumor

growth and reduces circulating IGF-1, particularly in women under 50. The mechanism by which

retinoids affect IGF-induced growth of breast cancer cells involves modification of local IGFBP

production (2 1 1).

Taken together, our data suggest the following: The breast carcinoma ce11 lines T-47D and

BT-474 have the necessary receptors and other transcriptional machinery to produce PSA. Once

stimulated by a steroid hormone, T-47D and BT-474 cells produce PSA rnRNA within 1-2h,

129

synthesize detectable intracellulas protein within 4 h and secrete detectable protein within 8 h. PSA

gene regulation is under the control of andorgens and progestins through the independent action of

the androgen and progesterone receptors (positive regulation). Weak positive regulation may also

be effected by high concentrations of glucocorticoids and mineralocorticoid. Estrogens do not

positively regulate the PSA gene but they act as blockers of androgen action. The most effective

blockers of PSA gene regulation were found to be the antiandrogen RU56.187 and the

antiprogestin Mifepristone. Our data, showing multihormone regulation of the PSA gene, are

in accord with those of GIover and Darbe (200) who concluded the same using T-47D celis

transfected with the mamrnary turnor virus long terminal repeat sequences.

Our tissue culture system not only reproduces the phenornenon of PSA production by breast

cancers but it also offers a means of testing the biological activity of candidate new hormonal and

antihormonal agents. With this systern, 1 have shown that two new antihormones, the antiandrogen

RU56,187 and the antiprogestin Mifepristone, which were found to be completely devoid of agonist

activity by traditional in-vivo and in-vitro techniques, have Iow but detectable androgen and/or

progestin agonist activity demonstrated by their ability to up-regulate the PSA gene.

Recently, we have obtained evidence that the progestin-mediated up-regulation of the PSA-

gene occurs in-vivo as well. We have reported PSA production by normal breast tissue in a female

patient who was receiving an oral contraceptive containing Norethidrone (8). PSA regulation by

progestins in the prostate has not been reported but it is known that the prostate cells, in addition

to AR, also contain PR (2 12). In another report, we described a patient who was receiving high

doses of glucocorticoids and had a . ovarian tumor producing PSA (27).

This data fùrther supports the view that PSA may be regulated in diverse tissues containing

130

PR. Tissues which contain PR include, the breast (6), the endometrium (33), brain meningiomas

(2 13), blood vesse1 walls (2 14), unnary (2 15) and osteoclasts (2 16). In view of these and other

hdings, 1 believe that it is tirne to snidy in detail the biological role of PSA in non-prostatic tissues.

The presence of a significant prognostic value of this molecule in breast cancer ( 12) suggests that

this eiegantly regulated enzyme rnay play some role in breast cancer initiation and progression.

7-3. PSA Production During the Menstrual Cycle

There is now extensive evidence that the PSA gene is up-regulated by androgens and

progestins but not estrogens. The PSA concentration in female semm is very low and usually not

measurable by commercial PSA assays. Using highiy sensitive PSA procedures that we developed

(1 8 l), we demonstrated that many female sera had measurable PSA concentrations (1 19). We have

further demonstrated that much higher PSA levels can be found in female breast secretions. For

comparative purposes, 1 here report approximate levels of PSA in various human fluids; serninal

plasma, 1,000,000,000 ng/L; male serum, 1,000-2,000 ng/L; normal breast discharge fluid,

5,000,000 ngL; rnilk of lactating women, 100,000 ng/L; female senim, 2-4 n a . There is

approximately a lo4-fold difference between PSA levels in the physiological secretion (e.g. seminal

plasma or breast discharge fluid) and the corresponding serum. The physiological role of PSA in

seminal plasma seerns to be established (66). The role of PSA in the breast and its secretions is still

obscure. We found that in breast cancer, patients with PSA-producing tumors have better prognosis

than patients with tumors which do not express PSA ( 12).

Since PSA expression is under the control of steroid hormones and their receptors, I

speculated that its concentration may change during the menstrual cycle. This was demonstrated

13 1

by assessing both the levels of semm PSA during the menstrual cycle of normal women and by

examining the ability of serum from these women to stimulate PSA production and PSA mRNA

expression in a tissue culture system. 1 demonstrated that semm PSA levels change significantly

during the menstruai cycle, following a specific pattern. This pattern was similar in three patients

and was reproducible on repeated cycles from the same patients (Figure 6-5-1 and 6-5-2). PSA

concentration increase follows the progesterone concentration increase but with a 10-20 day lag

period, suggesting that PSA is kely a product of progesterone action in the target tissue. Based on

many recent indirect findings, we suggest that the target tissue is the female breast but the

contribution of other steroid hormone responsive tissues, like the endornetrium cannot be excluded

(33). Recently, Clements et al. measured PSA mRNA in endometrial tissue of normal cycling

women and found that highest levels occur during the follicular phase and lowest levels during the

late luteal phase in general accordance with the data reported here (2 17). No association was seen

between the changes of PSA and changes in LH, FSH, Free testosterone or estradiol. Our tissue

culture system confirmed the following: (a) That progesterone cm up-regulate PSA production. (b)

that estradiol LH, FSH, prolactin, growth hormone and choriogonadotropin do not rnediate PSA

production in the T-47D and BT-474 ceil lines. (c) That sera collected at the luteai phase of the

cycle (which are high in progesterone) but not sera collected at the follicular phase or rnid-cycle)

have the ability to up-regulate PSA protein production and increase PSA mRNA expression in the

breast carcinoma ce11 line T-47D.

The data presented allow us to speculate that corpus luteum steroids, one of which is

progesterone, stimulate target tissues capable of producing PSA (one of which is the breast) for

PSA production and release into the mamrnary ducts. A minor Fraction of this PSA difises into the

132

general circulation and can be measured in the serum. Peak levels of PSA appear 10-20 days post-

progesterone peak. Once the corpus luteum regesses, PSA concentration decreases with an

apparent half-life of approximately 5-6 days. Based on these findings, we suggest that PSA is a

protein regulated by the corpus luteum and it is a marker of progesterone action in target tissues.

Since this protein has senne protease activity, it will be interesting to find its biological role in the

breast and its secretions and its possible physiological substrates. This issue has recently been

discussed ( 198).

Yu et al. (8) have demonstrated that prostate specific antigen gene can be upregulated by

progestin containing oral contraceptives. Recently, the role of oral contraceptives in the

development of breast cancer has been reviewed in over 150,000 patients. The conclusion fiom this

study is that women who are currently using combined oral contraceptives or have used them in the

past 10 years are at a slightly increased risk of having breast cancer diagnosed, although the

additional cancers diagnosed tend to be localized to the breast. There is no evidence of an increase

in the risk of having breast cancer diagnosed 10 or more years after cessation of use, and the

cancers diagnosed then are less advanced clincally than the cancers diagnosed in never-users (2 18).

7-4. Expression of PSA Gene in Lung Tissue

More recently, others have proposed that PSA may cleave biologically important substrates

like insulin-like growth factor binding protein 3 (IGF'BP-3) (78), parathyroid hormone related

peptide (PTHrP) (219) or unknown substrates which release peptides with smooth muscle

contraction activity (220). The search for new biological functions of PSA has recently been

intensified d e r the realization that PSA is expressed in diverse tissues (198). Striking sirnilarities

133

between PS A expression and expression of the breast cancer susceptibility gene BRC A- l have been

noted (22 1 ). For example, BRC A- 1 and PS A were found to be upregulated dunng pregnancy, are

both secreted into the milk of lactating women and are both regulated by steroid hormones. If

BRCA- 1 is a granin then PSA and BRCA- 1 may be secreted by the breast epithelial cells. Others

have found that PSA may act as a negative growth regulator in hormone-dependent breast cancers

by stimulating the conversion of estradiol to estrone (222). The highest concentration of PSA,

except for seminal plasma, was found in breast discharge fluid (up to 5 mg/L) and in milk of

lactating women (up to 0.3 m a ) (99).

It has been known for years that, in the prostate, the PSA gene is regulated by androgens

through the androgen receptor (22). Using a tissue culture system with breast carcinoma ce11 lines,

we have demonstrated that androgens, progestins and glucocorticoids can independently up-regulate

the PSA gene (96). We have also provided evidence that this regdation occurs irz-vivo by

describing a woman whose breast tissue PSA was very high due to stimulation by a progestin

containing oral contraceptive (8) and another woman whose ovarian tumor was producing PSA,

likely due to stimulation by glucocorticoids (27).

Although two previous studies have suggested that PSA may be produced by lung tissue and

tumon (223,123), their data were based only on protein measurements. Since PSA has extensive

sequence homology with the hGK-1 gene and protein and PSA antibodies cross react with hGK- 1

(224,225), the only certain way to verify the presence of PSA in lung is by either sequencing the

purified protein (which is irnpractical due to the very low levels present) or by molecular analysis

of the expressed PSA rnRNA. 1 describe here a patient whose lung tissue extract was highly positive

for PSA protein in its 33 KDa form. In this patient, 1 was able to arnplify PSA mRNA and sequence

134

PSA cDNA to ven@ its identity as PSA cDNA and not hGK-1 cDNA. Immunohistochernical

localization has shown that the PSA-positive tissue is the normal lung epithelium adjacent to the

tumor. This is in accord with data fiom breast tumors in which we noted more fiequent staùiing

for PSA of normal and hyperplastic breast epithelium and well-differentiated tumors. Poorly

differentiated and steroid hormone receptor-negative breast tumors are usually negative for PSA.

This patient was unique since her lung tissue contained more than 50 times the PSA

immunoreactivity of another 10 lung tumor extracts from women. 1 postulate that in this patient,

the PSA was up-regulated by the steroid beclomethasone dipropionate since this drug was found

to be able to up-regulate PSA production in-vitro, in a dose-dependent manner (Figure 6-6-6).

These data parallei observations in a patient with ovarian carcinoma (27) and a patient receiving

oral contraceptives (8). Recently, in support of our hypothesis, Beattie et al. reported presence of

high affinity receptors for androgen, estrogen and glucocorticoids in lung tissues and carcinomas

(169). Beer and Malkinson reported similar data in mouse lung tumors (226). I conclude that PSA

can be expressed in a variety of tissues bearing steroid homone receptors and that its expression

can be manipulated by exogenously administered hormones. The biological role of PSA in these

tissues is still unknown.

7-5. Frequency of PSA mRNA Expression in Lung tumon

PSA in lung tissue was previously found using irnrnunological assays that measure PSA

protein (123). However, since PSA is also present in the blood, there is always the possibility of

contamination of tissue extracts by residual blood in the tissue. Also, the close sequence homology

berneen PSA and hGK-1 may cause problems of cross-reactivity (22). In this study, 1 have s h o w

135

that contamination of the tumor extracts by blood PSA is likely, since 1 found a signdlcant

correlation between tumor extract PSA and serum P S 4 especially in male patients.

Since PSA mRNA-containing celis do not seem to exist in the circulation of patients without

prostate cancer (208,220,221), 1 have attempted to show evidence for PSA expression in lung tissue

by searching for PSA mRNA presence in lung tissue extracts. 1 found PSA &A in 69% of lung

tumors from males, 53% of lung tumors fiom females and in 71% of normal lung tissues. These

dzta support the view that low levels of PSA mRNA are found in the majority of lung tissue

extracts. These levels are much lower than those found in femde breast since in the latter tissue,

PSA mRNA could be detected without the need of Southem blot hybndization (227) and PSA

protein levels in breast tumor extracts are much higher then those reported in Table 6-7-1 (8). As

in the case of breast, normal lung tissue contains PSA mRNA as well.

1 found no association between PSA mRNA presence and clinicopathoiogical variables

including patient age, stage, grade, histological type or nodal status. Consequently, the prognostic

value of this marker in lung cancer is questionable at present. The regdation of PSA rnRNA

expression in lung tissue and its possible biological role are unknown.

By cloning and sequencing of aberrant PCR products, 1 was able to isolate a 450 bp unique

sequence not previously deposited in GenBank (Figure 6-7-3). Additional work will reveal if this

fragment belongs to a new gene that is overexpressed in lung tumors. Work dong these lines is

currently in progress.

This data support the view that PSA is fiequently expressed at low levels in cancerous and

normal lung tissue. The role of this molecule in lung physiology and pathobiology is currently

unknown.

136

7-6. Conclusion

For studying of PSA gene expression in non-prostatic tissues 1 have first developed reverse

transcription-polymerase chah reaction (RT-PCR) methods for detecting prostate-specific antigen

(PSA) mRNA Using this methods, and a highly sensitive irnmunofluorometnc assay for measuring

PSA proteinJ found good agreement between presence of PSA protein and PSA rnRNA in breast

tumors. 1 thus proposed that in women, detection of PSA protein or PSA mRNA in tissues and

turnors offers equivalent information. Because PSA protein is present in male blood and thus could

contaminate extracts fiom tumors and tissues from men, I propose that the RT-PCR methods 1

describe be used to assess non-prostatic expression of the PSA gene in men.

For investigating the mechanism of PSA gene regulation in the breast, I have developed a

tissue culture system which reproduces in-vitro the phenornenon of PSA production by breast cells.

The tissue culture expenments showed that steroid hormone receptor positive breast cancer cells,

T-47D and BT-474, are capable of producing PSA when stimulated by androgens, progestins and

çlucocorticoidslrnineralocorticoid but not estrogens. However, Our studies demonstrated that those

ce11 lines which were devoid of steroid hormone receptors, PSA production was not observed after

stimulation with steroid hormones. The steroid hormone induced PSA production in breast cancer

cells was time- and dose-dependent. Kinetic studies showed that PSA mRNA expression appears

-2 hours post stimulation; PSA protein appears after 4-8 h. Among sixty four compound tested,

only androgens and progestins were able to stimulate PSA production at concentrations beiow 10"

M. Evidence that the progesterone and androgen receptors cm regulate the PSA gene

independentiy was provided as follows: (a) The progestin Norgestimate which does not bind to the

androgen receptor up-regdates the PSA gene at concentrations as low as 1 O-'' M. (b) Tnamcinolone

137

acetonide which does not bind to the AR but binds to the PR acts sirnilarly to Norgestimate (c) The

antiandrogen cyproterone acetate which blocks the androgen receptor but has progesterone activity,

up-regulates the PSA gene at concentrations as low as 10'" M. (d) the antiprogestin Mifepristone

blocks completely the stimulation of the specific progestin Norgestimate. Our tissue culture system

identified androgedprogestin agonist activities of 17a-ethinylestradiol, the antiandrogen RU56,187

and the antiprogestin Mifepristone. These data suggested that the expression of the PS A gene in the

fernale breast is under the control of androgens and progestins. Our tissue culture system is a highly

sensitive in-vitro method for evaluating the biological activity of candidate compounds having

agonist and antagonist steroid hormone activity.

The regulation of the PSA gene by steroid hormones in the breast led us to speculate that

this protein may be differentially expressed during the menstrual cycle. By examining the PSA

concentration in serum during the menstrual cycle of normal women, 1 found that PSA levels in

serum are highest dunng the mid-late follicular phase, drop continuously with a half-life of 5-6 days

between the late follicular phase and rnid-cycle and reach a minimum dunng the mid-luteal phase.

PSA changes do not correlate with changes in LH, FSH, or estradiol levels. However, PSA peaks

seem to follow the progesterone concentration peaks with a delay of 10-20 days. Using a tissue

culture system based on T-47D breast carcinoma ce11 line, only sera obtained during the rnid-late

luteai phase were able to up-regulate the PSA mRNA and protein. In stimulation experiments in-

vitro, progesterone, but not LEI, FSH, estradiol, hCG, prolactin or growth hormone, was able to up-

regulate PSA mRNA and protein in the T-47D ce11 line. These data suggested that progesterone,

and possibly other corpus luteum steroids, stimulate target tissues for ?SA production in a cyclical

manner during the menstmal cycle. However, the relatively long tirne difference between

138

progesterone and PSA peaks does not support a direct induction-secretion mechanism. This issue

needs clarification with further expenmentation.

Further, the ability of non-prostatic tissues to produce PSA was further reported in a female

patient with lung adenocarcinorna whose tumor extracts was highly positive for PSA

immunoreactivity. The identity of PSA in lung tissue was verified by using RT-PCR, HPLC,

immunohistochemistry, and sequencing of the PCR products. In addition, tissue culture experiments

suggested that beclomethasone, a glucocorticoid used to treat the patient, was able to up-regulate

PSA gene expression. This was the first report that unequivocally demonstrated PSA expression

in lung tissue.

In addition, further studies demonstrated that PSA protein and mRNA was present in

primary lung tumor tissues in both males and fernales. PSA protein was detected more fiequently

and at higher levels in lung tumor extracts fiorn male patients, suggesting a possible contamination

of the tumor extracts with PSA fiom residual blood in the tumor vasculature. Clinicai studies

indicated that the expression of PSA protein did not associate with patient age, stage, grade,

histological type or nodal status. Similarly, PSA mRNA was not associated with any

clinicopathological variables but squarnous ceil carcinomas, especially in males, were more

fiequently positive. Consequently, the prognostic value of this marker in lung cancer is questionable

at present. Our data support the view that PSA is fiequently expressed at low levels in cancerous

and normal lung tissue. The role of this rnolecule in lung physiology and pathobiology is currently

unknown.

More interestingly, by cloning and sequencing of aberrant PCR products, 1 was able to

isolate a 450 bp unique sequence not previously deposited in the GenBank. Additionai work will

139

reveal if this fragment belongs to a new gene that is overexpressed in lung tumors. Work dong

these lines is currently in progress.

CHAPTER 8. Concluding Remarks and Future Directions

In this thesis, 1 have generated the following new information.

1 . 1 have devised an ultrasensitive, highly specific rnethod for monitoring prostate specific

antigen mRNA levels. This method was s h o w to produce results in excellent concordance

with an ultrasensitive imrnunofluorometric procedure for measuring PSA protein.

2. 1 have discovered two breast carcinoma ce11 lines which can upregulate the prostate specific

antigen gene d e r exogenous stimulation by steroid hormones. By using this system, 1 have

s h o w that the PSA gene can be upregulated at steroid hormone concentrations below 10''~

M by progestins and androgens. In addition, higher concentrations of glucocorticoids and

mineralocorticoids can also upregulate the PSA gene. 1 have s h o w that estrogens not only

do not upregulate the PSA gene but they block PSA gene upregulation by progestins and

androgens. 1 have also studied the kinetics of PSA mRNA protein regulation in these ce11

lines. The data generated have shown that the PSA gene, in these ce11 lines is under the

control of steroid hormones.

3 . 1 have shown that the data generated with the VI-vitro system are in accord with the data that

1 have produced Nt-vivo. by studying the PSA gene expression during the menstrual cycle.

1 have shown for the first time that the PSA gene is regulated differentially in a cyclical

manner during the rnenstrual cycle. 1 have also provided evidence that during the menstrual

cycle, the major regulating steroid is progesterone.

By using a combination of protein measurements, high performance liquid chromatography,

RT PCR, Southem blot hybridization, immunohistochernistry and DNA sequencing

141

techniques, 1 have shown unequivocally that the prostate specific antigen gene is expressed

by lung tissue and lung tumors. 1 provided evidence that at least in one patient that the PSA

gene in lung tissue is upregulated by exogenously administered steroids. 1 found no

association between PSA gene expression and lung cancer and clinicopathological variables

of the patients.

It is now very clear that the PSA gene is not expressed specifically in prostatic tissue as it

was believed for many years. However, there are still many questions which have originated fiom

my investigations and those of other members of our research laboratory. For example, 1 tabulate

below a number of unanswered questions.

The physiological role of PSA in breast and other non-prostatic tissues remains obscure.

PSA is a senne protease with chymotrypsin like enzymatic activity and it is likely exerting

its physiological role by cleaving other substrates. It has recently been postulated that PSA

may be a protease for the BRCA-1 gene products (E.P. Diamandis. BRCAl protein

products: Antibody specificity. functional motifs and secreted tumour suppressors. Nature

Genetics, 1 996; 1 3 : 268).

1 postulated that the PSA molecule may be acting through surface receptors in regulating

ce11 proliferation. However, extensive experimentations to identi@ such a receptor have

been unsuccessful. 1 found no evidence of presence of a receptor in these ce11 lines.

It will be interesting to examine and discover why PSA is a favourable prognostic indicator

in breast cancer and a marker of reduced risk for developing breast cancer. It appears that

PSA is a molecule with anticarcinogenic properties in the breast.

I postulated that the tissue culture systems that 1 developed may have potential for screening

142

various candidate compounds with steroid hormone agonist and steroid hormone antagonist

activities. Preliminary results by other members of our research laboratory suppon this

hypothesis.

5 . It will be interesting to examine if the functions of prostate specific antigen are rnediated

through growth factor regulation. It has recently been proposed that PSA is a protease for

IGF binding proteins and this rnay constitute a regulatory loop for growth factor regulation.

6 During my investigations, 1 found two breast carcinoma ce11 lines, which are steroid

hormone receptor positive and can upregulate the PSA gene d e r induction by progestins

and androgens. However, 1 have also identified other ceif lines which contain steroid

hormone receptors (e.g. ZR-75, BG- 1) which do not upregulate the PSA gene after steroid

hormone induction. I have postulated that these ce11 lines have either defective receptors

or are lacking other critical transcription factors that are involved in PSA gene regulation.

CHAPTER 13. References

Oesterling JE. Prostate specific antigen : a cntical assessrnent of the most useful tumor

marker for adenocarcinoma of the prostate. I Urol, 145 :9O7-923, 1 99 1.

Papsidero LD, Kuriyarna M. Wang MC, Horozewicz I, Leong SS, Valemela L, et al.

Prostate specific antigen: a marker for human prostatic epithelial cells. J Nat1 Cancer Inst,

66:37-42, 1981.

Stamey TA Yang N, Hay AR, McNeai JE, Freiha FS, Redwine E. Prostate-specific antigen

as a serum marker for adenocarcinoma of the prostate. N Engi J Med, 3 17:909-9 16, 1987.

Van Krieken IH. Prostate marker imrnunoreactivity in saiivary gland neoplasms. A rare

pitfall in immunohistochernistry. Am J Surg Pathoi, 17:4 10-4 14, 1993.

Diamandis EP, Yu H, Sutherland DIA. Detection of prostate specific antigen

immunoreactivity in breast tumors. Breast Cancer Res Treat, 32:30 1 -3 10, 1 994.

Yu H, Diamandis EP, Sutherland DJA. Immunoreactive prostate specific antigen levels in

fernale and male breast tumors and its association with steroid hormone receptors and

patient age. Clin Biochem, 27:75-79, 1994.

Levesque M, Yu H, D'Costa M, Diarnandis EP. Prostate specific antigen expression by

various tumors. J Clin Lab Analysis, 9: 125- 128, 1995.

Yu H, Diamandis EP, Morne M, Croce CM. Oral contraceptive-induced expression of

prostate specific antigen in the female breast. J Bi01 Chem, 270:66 15-66 18, 1995.

Yu H, Diamandis EP. Prostate specific antigen in milk of lactating women. Clin Chem,

4 1 : 54-58, 1995.

Diamandis EP, Yu H, Lopez-0th C. Prostate specific antigen - A new constituent of breast

144

cyst fluid. Breast Cancer Res Treat, 38259-264. 1996.

Yu H, Diamandis EP. Prostate specific antigen immunoreactivity in amniotic fluid. Clin

Chem, 411204-210, 1995.

Yu H, Giai M., Diamandis EP, Katasaros D, Sutherland DJA, Levesque Mq Roagna R

Pomone R Sismondi P. Prostate specific antigen is a new favourable prognostic indicator

for women with breast cancer. Cancer Res, 5 5 :2 104-2 1 10, 1995.

Montgomery BT, Young CY-F, Bilhartz DL, Andrews PE, Prescott JL, Thompson NF

and Tindall DI. Hormonal regdation of prostate specific antigen (PSA) glycoprotein in the

human prostatic adenocarcinorna ce11 line, LNCaP. Prostate, 2 1 :63-73, 1992.

Ambruster DA. Prostate-specific antigen: Biochemistry, analytical methods, and clinical

application. Clin Chem, 39: 18 1 - 195, 1993.

Hara M, Koyanagi Y, houe T, Fukuyama T. Some physio-chernical charactenstics of

gamma serninoprotein; an antigenic component specific for human seminal plasma. Jpn J

Legal Med, 25 :3 22-3 24, 1 97 1.

Li TS, Beling CG. Isolation and characterization of two specific antigens of human seminal

plasma. Fertil Stenl, 24: 134- 143, 1973.

Sensabaugh GF. Isolation and characterization of a semen-specific protein from human

seminal plasma: a potential new rnarker for semen identification. J Forensic Sci, 23: 106-

115, 1978.

Graves HCB, Sensabaugh GF, Blake RT. Postcoital detection of male-specific protein.

Application to the investigation of rape. N Engl J Med, 3 12:338-343, 1985.

Wang MC, Valenzuela LA, Murphy GP, Chu TM. Tissue specific and tumor specific

145

antigens in human prostate. Fed Proc, 36: 1254.1977.

Wang MC, Valenzuela LA, Murphy GP, Chu TM. Purification of a human Prostate specific

antigen. Invest Urol, 17: 159- 163, 1979.

Wang MC, Valenniela L 4 Murphy GP, Chu TM. A simplified purification procedure for

human prostate antigen. Oncology, 39: 1-5, 1982.

McCormack RT, Rittenhouse HG, Finlay J 4 Sokoloff RL, Wang TJ, Wolfert EU, Lilja H,

Oesterling JE. Molecular forms of prostate-specific antigen and the human kallikrein gene

farnily: a new era. Urology, 45 :729-744, 1995.

Papsidero LD, Wang MC Valenzuela LA, Murphy GP, Chu TM. A prostate antigen in sera

of prostatic cancer patients. Cancer Res, 40:2428-2432, 1980.

Papotti M, Paties C, Peveri V, Moscuzza L, Bussolati G. Irnrnunocytochemical detection

of prostate-specific antigen (PSA) in skin adnexai and breast tissues and tumors. Bas Appl

Histochem, 33 :25-29, 1989.

Wilbur DC, Krenzer K, Bonfiglio TA. Prostate specific antigen (PSA) staining in

carcinomas of non-prostatic ongin. Am J Clin Pathol, 88 (Abstr): 530, 1987.

Kuopio T, Ekfors TO, Nikkanen V, Nevalainen TJ. Acinar ce11 carcinoma of the pancreas:

Report of three cases. APMIS, 103 :69-78, 1995.

Yu H, Diamandis EP, Levesque M, et al. Expression of the prostate-specific antigen gene

by a primary ovarian carcinoma. Cancer Res, 5 5 : 1603- 1606, 1995.

Zaviacic M, Sidlo J, Borovsky M. Prostate specific antigen and prostate specific acid

phosphatase in adenocarcinorna of Skene's paraurethral glands and ducts. Virchows Arch,

423 :503-505, 1993.

146

Dodson MK, CIiby W k Keeney GL, Peterson MF, Podratz KC. Skene's gland

adenocarcinorna with increased serum level of prostate-specific antigen. Gynecol Oncol,

55:304-307, 1994.

Grignon DL, Ro Pr', Ayala AG, Johnson DE, Ordonez NG. Primas, adenocarcinorna of the

urinary bladder: A clinicopathological analysis of 72 cases. Cancer, 67:2 165-2 172, 199 1.

Frazier E-W, Humphrey P 4 Burchette JL, Paulson DF, Immunoreactive prostatic specific

antigen in male periurethral glands. I Urol, 147:246-248, 1992.

Karnoshida S, Tsutsumi Y. Earaprostatic localization of prostate acid phosphatase and

prostate-specific antigen: Distribution in cholacogenic glandular epithelium and sex-

dependent expression in human anal gland. Hum Pathol, 2 1 : 1 108- 1 1 1 1, 1990.

Clements J 4 Mukhtar A. Glandular kallikriens and prostate specific antigen are expressed

in the human endometrium. .J Clin Endocrinol Metab, 78: 1 536- 1 539, 1994.

Watt KWK, Lee PJ, MTimkulu T, Chan WP, Loor R. Human prostate-specific antigen:

structural and functional similarity with serine proteases. Proc Nat1 Acad Sci USA,

83 :3 166-3 170, 1986.

Lundwall A, Lilja H. Molecular cloning of human prostate specific antigen cDNA. FEBS

Lett, 2 l4:3 17-322, 1987.

Schaller J, Akiyama K, Tsuda R, Hara M, Marti T, Rickii EE. Isolation, charactenzation

and amino-acid sequences of gamma-seminoprotein, a glycoprotein fiom human seminal

plasma. Eur J Biochem, 170: 1 1 1- 120, 1987.

Wu, JT, Zhang P, Wang T, Wilson L, Asti11 M. Evaluation of fiee PSA isoforms, PSA

complex formation and specificity of anti-PSA antibodies by HPLC and PAGE-

147

imrnunoblotting techniques. J Clin Lab Analysis, 9: 1 - 14, 1995.

Wang MC, Papsidero LD, Kuriyama M, Valenzuela LA, Murphy GP, Chu TM. Prostate

antigen: a new potential marker for prostatic cancer. Prostate, 2:89-96, 198 1 .

Wang MC, Kuriyarna M, Papsidero LD, et al. Prostate antigen of human cancer patients.

Methods Cancer Res, 19: 179- 197, 1982.

Riegman PHJ, Vlietstra RI, van der Korput JAGM, Romijn JC, Trapman J. Characterization

of the prostate-specific antigen gene: a novel human kallikrein-like and gene. Biochem

Biophys Res Commun, 159:95-102, 1989.

Henttu P, V i o P. cDNA coding for the entire human prostate specific antigen shows high

homologies to the human tissue kallikrein genes. Biochem Biophys Res Commun, l60:903-

910, 1989.

Henmi P. Vihko P. Prostate-specific antigen and human glandular kallikrein: two kallikreins

of the human prostate. Ann Med, 26: 157- 164, 1994.

Schedlich LJ, Bemetts BH, Moms BI. Primary structure of a human glandular kallikrein

cene. DNA, 6429437, 1987. C

Clements JA. the glandular kallikrein family of enzymes: Tissue-specific expression and

hormonal regulation. Endocrine Rev, 10:393-4 19, 1989.

Murtha P, Tindall DI, Young CY. Androgen induction of a human prostate-specific

kallikrein, hKLK2: characterization of an androgen response element in the 5' promoter

region of the gene. Biochemistry, 32:6459-6464, 1993.

Young CY, Andrews PE, Montgomery BT, Tindall DJ. Tissue-specific and hormonal

regulation of human prostate-specific glandular kallikrein. Biochemistry, 3 1 : 8 18-824, 1992.

148

Chapdelaine P, Paradis G, Tremblay RR, Dube JY. High level of expression in the prostate

of a human glandular kallikrein mRNA related to prostate-specific antigen. FEBS Lett.

236:205-208, 1988.

Henttu P. Lukarinen O, Vihko P. Expression of the gene coding for human prostate-specific

antigen and related hGK- 1 in benign and malignant tumors of the human prostate. Int J

Cancer, 45654-660, 1990.

Clements JA The human kaliikrein gene family: a diversity of expression and function. Mol

Ce11 Endocrin, 99:C 1 -6, 1994.

Peehl DM. Prostate specific antigen role and function. Cancer , 75 202 1-2026, 1995.

Digby M, Zhang XY, Richards RI. Human prostate specific antigen (PSA) gene: structure

and linkage to the kallikrein-like gene, hGK- 1. Nucleic Acids Res, 17:2 13 7, 1989.

Riegman PHJ, Vlietstra RJ, Klaassen Pl van der Korput JAGM, Guerts van Kessel A,

Romijin JC, Trapman J. The prostate-specific antigen gene and the human glandular

kallikrein- 1 gene are tandemly located on chromosome 1 9. FEBS Lett, 247: 123- 1 26, 1989.

Morris BJ. hGK-1: a kallikrein gene expressed in human prostate. Clin Exp P h a m Physiol,

l6:345-35 1, 1989.

Goldfarb D4 Stein BS, Shamszadeh M, Petersen RO: Age-related changes in tissue levels

of prostatic acid phosphatase and prostate-specific antigen. I Urol, 136: 1266- 1269, 1986

Vieira JGH, Nishida SK, Pereira AB, Arraes RF, Verreschi ITN. Semm levels of prostate-

specific antigen in normal boys throughout puberty. J Clin Endocnnol Metab, 78: 1185-

1187, 1994.

Weber JP, Oesterling JE, Peter CA, Partin AW, Chan DW, Walsh P: The infiuence of

149

reversible androgen depnvation on semm prostate-specific antigen levels in men with

benign prostatic hyperplasia. J Urol, 14 1 :987-992, 1989.

Hanash Kq Mostofi W. Androgen effect on prostate specific antigen secretion. J Surg

Oncol, 49: 202-204, 1 992.

Stamey TA, Kabah IN, Ferrari M, Yang N. Prostate specific antigen and the diagnosis and

treatment of adenocarcinoma of the prostate. IV. Anti-androgen treated patients. .i Urol,

141: 1088-1090, 1989.

Young CYF, Montgomery BT, Andrews PE, Qiu S, Bilhartz DL, Tindall DJ. Hormonal

regulation of prostate-specific antigen messenger RNA in human prostatic adenocarcinoma

ce11 line LNCaP. Cancer Res, 5 1 : 3 748-3 752, 199 1.

Skowronski RI, Peehl DM, Feldman D. Vitamin D and prostate cancer: 1,25

dihydroxyvitamin D, receptors and actions in human prostate cancer ce11 lines.

Endocnnology, 132: 1952- 1960, 1993.

Hsieh J-T, Wu H-C, Gleave ME, von Eschenbach AC, Chung LW. Autocnne regulation of

prostate-specific antigen gene expression in a human prostate cancer (LNCaP) subline.

Cancer Res, 53 :2852-2857, 1993.

Andrews PE, Young CYF, Montgomery BT, Tindall DI. Tumor-promoting phorbol ester

down-regulates the androgen induction of prostate-specific antigen in a human prostatic

adenocarcinoma ceIl line. Cancer Res, 52: 1525- 1529, 1992.

Young CYF, Murtha PE, Andrews PE, Lindzay K. Tindall DJ. Antagonism of androgen

action in prostate tumor cells by retinoid acid. Prostate, 25:39-45, 1994.

Henttu P, Vihko P. Growth factor regulation of gene expression in the human prostatic

carcinoma ceIl Iine LNCaP. Cancer Res, 53 : 105 1 - 1058, 1993.

65. Guo Y, Pili Passaniti A. Regulation of prostate-specific antigen gene expression in

LNCaP human prostatic carcinoma cells by growth, dihydrotestosterone, and Extra celluar

matrix. Prostate, 24: 1 - 10, 1994.

66. Lilja H. A kallikrein-like senne protease in prostatic fluid cleaves the predominant seminal

vesicle protein. .J Clin Invest, 76: 1899- 1 903, 1 985.

6 7 Akiyama K, Nakamura T, Iwanaga S, Hara M. The chyrnotrypsin-like activity of human

prostate-specific antigen, r-seminoprotein. FEBS Lett, 225: 168- 1 72, 1987.

68. Christewson 4 Laureli CB, Lilja H. Enzymatic activity of the prostate-specific antigen and

its reactions with Extracelluar senne proteinase inhibitors. Eur J Biochem, 194:755-763,

1990.

69. Lilja H, Abrahamsson PA, Lundwall A. Sernenogelin, the predominant protein in human

semen: primary structure and identification of closely related proteins in the male accessory

sex glands and on the spermatozoa. J Bi01 Chem, 264: 18%- 1900, 1989.

70. Christensson A, Lilja H. Complex formation between protein C inhibitor and prostate-

specific antigen Nt-vitro and in human semen. Eur J Biochern, 220:45-53, 1990.

71. Ban Y, Wang MC, Watt KWK, Loor R, Chu M. The proteolytic activity of human prostate-

specific antigen. Biochem Biophys Res C o w 123 :482-488, 1984.

72. Lilja H. Structure, function, and regulation of the enzyme activity of prostate-specific

antigen. World J Urol, 1 1 : 188- 19 1, 1993.

73. Lilja H, Cockett ATK, Abrahamsson P-A. Prostate specific antigen predominantly forms

a complex with alpha, -antichymo-trypsin in blood. Cancer, 70(Suppl):230-234, 1992.

15 1

McGee RS, Herr JC. Human serninal vesicle-specific antigen is a substrate for prostate-

specific antigen (or P-30). Bi01 Reprod, 39:499-5 10, 1988.

Lilja H, Oldbring J, Rannevik G, Laurel1 CB. Seminal vesicle-secreted proteins and their

reactions durhg gelation and liquefaction of human semen. J Clin Invest, 80: 28 1-285, 1 987.

Kanety H, Madjar Y, Dagan Y, Levi J, Papa MZ, Pariente C, Goldwasser B. Karasik A.

Semm insulin-like growth factor-binding protein-2 (IGFBP-2) is increased and IGFBP-3

is decreased in patients with prostate cancer: correlation with serurn prostate-specific

antigen. J Clin Endocrinol Metab, 77229-233, 1993.

Cohen Pl Fielder PI, Hasegawa Y, Frisch H, Giudice LC, Rosenfeld RG. Clinical aspects

of insulin-like growth factor binding proteins. Acta Endocrinol (Copenh), 124 (Suppl 2):74-

85, 1991.

Cohen P, Graves HCB, Peehl DM, Karnarei M, Giudice LC, Rosenfeld RG. Prostate-

specific antigen (PSA) is an insulin-like growth factor binding protein-3 protease found in

serninal plasma. J Clin Endocnnol Metab, 75 : 1046- 1053, 1992.

K W c h i RF, Yee D. The insulin-like growth factor family of ligands, receptors, and

binding proteins. Breast Cancer Res Treat, 22: 7- 19, 1992.

Baserga R. The insulin-like growth factor I receptor: a key to tumor growth. Cancer Res,

55 :249-252, 1995.

Figueroa JA, Yee D. The insulin-like growth factor binding proteins (IGFBPs) in human

breast cancer. Breast Cancer Res Treat, 22:8 1-90, 1992.

Cohen P. Peel DM. Graves HCB, Rosenfeld RG. Biological effects of prostate specific

antigen as an insulin-like growth factor-binding-3 protease. J Endocnnol, 142:407-415,

1994.

Oh Y, Muller HL, Larnson G, Rosenfeld RG. Insulin-like growth factor (1GF)-independent

action of IGF-binding protein-3 in Hs578T human breast cancer cells. J Bi01 Chem 268,

14964-14971, 1993.

Killian CS, Corral DA, Kawinski E, Constantine RI. Mitogenic response of osteoblast ceils

to prostate specific antigen suggests an activation of latent TGF-B and a proteolytic

modulation of cell adhesion receptors. Biochem Biophys Res Commun, 192:940-947, 1993

Lilja H. Christensson A, Dahlen U. et al: Prostate-specific antigen in serum occurs

predominantly in cornplex with al -antichymotrypsin. Clin Chem 37: 16 18-1625. 199 1 .

Partin AW. Oesterling JE: The clinical usefulness of prostate specific antigen: Update

1994. J Ur01 1521358-1368. 1994.

I w a k i J. Grandbois K. Wehner N. et al. An anaiysis of Lirinary prostate specific antigen

before and after radical prostatectomy : evidence for secretion of prostate speci fic antigen

by the periurethral glands. J Ur01 149:783-786. 1993.

Nowels K. Kent E. Rinsho K. et al: Prostate specific antigen and acid phosphatase-

reactive cells in cystitis cystica and glandularis. Arch Pathol Lab Med 1 12734737. 1988.

Pollen JJ. Dreilinger A: Imrnunohistochemical identification of prostatic acid phosphatase

and prostate specific antigen in female periurethral glands. Urology 23:303-304. 1984.

Tepper SL, Jagirdar J, Heath D, et al: Homology between the female paraurethral

(Skene's) glands and the prostate. Arch Pathol Lab Med 108:423-425. 1984.

Wernert N. Albrech M. Sesterhenn 1. et al: The 'female prostate': location. morphoiogy.

153

immunohistochemical characteristics and significance. Eur Ur01 2264-69, 1992.

Ablin RJ: Prostate-specific antigen and the female prostate. Clin Chem 35507-508. 1989.

Boutin EL. et al: The response of female urogenital tract epithelia to mesenchymal

inductors is restricted by the germ layer origin of the epithelium: prostatic inductions.

Differentiation. 48:99- 105, 199 1.

Diamandis EP. Yu H, Sutherland DJA: Detection of prostate-specific antigen

irnmunoreactivity in breast tumors. Breast Cancer Res Treat 3230 1-3 10, 1994.

Morne M, Croce CM, Yu H, Diamandis EP. Molecular charactenzation of prostate specific

antigen mRNA expressed in breast tumors. Cancer Res, 54:6344-6347, 1994.

Yu H, Diamandis EP, Zarghami N, Grass L. Induction of prostate specific antigen

production by steroids and tarnoxifen in breast cancer ceil lines. Breast Cancer Res Treat,

32:291-300, 1994.

Diamandis EP. Yu H. Lopez-Otin C: Prostate specific antigen - a new constituent of

breast cyst tluid. Breast Cancer Res Treat 38259-264. 1996.

Kuenen-Boumeester V. et al. Irnmunohistochemical determination of androgen receptors

in relation to oestrogen and progesterone receptors in female breast cancer. Int J Cancer

5258 1-584. 1992.

Sauter ER. Daly M. Lenahan K. et al: Prostate specific antigen in nipple aspirate fluid

correlate with breast cancer risk. Proc AACR (abstract) 37213. 1996.

McCormack RT, Rittenhouse HG, Finlay JA, SokolofFRL, Wang TJ, Wolfert RL, Lilja H,

Oesterling .JE. Molecular forms of prostate specific antigen an dthe human kallikrein gene

family: a new era. Ur01 45 : 729744, 1 995.

Yu H. Levesque MA, Mohammad K. et al: Associations between insulin-like growth

factors and their binding proteins and other prognostic indicaton in breast cancer. Br J

Cancer (in press). 1996.

Culig Z, Hobisch A, Cronauer MV, Radmay C, Trapman J, Hittmair A, Bartsch G, and

Klocker K. Androgen receptor activation in prostatic tumor ce11 lines by insulin-like growth

factor-1, keratinocyte growth factor, and epidermal growth factor. Cancer Res, 545474-

5478, 1994.

Manni A. Badger B. Wei L. et al: Hormonal regulation of insulin-like growth factor II

and insulin-like growth factor binding protein expression by breast cancer ce11 in vivo:

evidence for stroma1 epithelial interactions. Cancer Res 5429342942. 1994.

Westley BR. May FEB: Insulin-like growth factors: the unrecognised oncogenes. Br J

Cancer 72: 1065- 1066. 1995.

Desomber ER: Steroid receptors in breast cancer. In McBivitt RW. et al. (eds): The

Breast. Baltimore. Williams & Wilkins. 1 984. p 149- 1 74.

Wei LL: Transcriptional activation of the estrogen receptor. Clin Chem 3934 1-345.

1993.

de Launoit Y. Dauvois S. Dufour M. et al: Inhibition of ce11 cycle kinetics and

proliferation by the androgen sa-dihydrotestosterone and antiestrogen N, n-butyl-N-

methyl- 1 1 -[16'a-chloro-3'. 1 7B-dihydroxy-estra- 1 '. 3'. Y-( 1 07)triene-7'a-yl]

undecanamide in human breast cancer ZR-75- 1 cells. Cancer Res 5 1 :2797-2802, 199 1.

155

Dauvoid S. Geng SC. Levesque C. et al: Additive inhibitory effects of an androgen and

the antiestrogen EM- 1 70 on estradiol-stirnulated growth of human ZR-75- 1 breast tumors

in athymic mice. Cancer Res 5 1 :3 13 1-3 135, 199 1.

Pritchard KJ. The use of endocrine therapy. Hematol Oncol Clin N Am 3:765-805. 1989.

Harris JR, Lippman ME. Veronesi U. et al. Breast cancer. N Engl J Med 3275 19-328.

1992.

Hendenon BE. Ross RK. Pike MC: Hormonal chemoprevention of cancer in women.

Science 259533-638. 1993.

Jordan VC. Murphy CS: Endocrine pharmacology of antiestrogens as antitumor agents.

Endocr Rev 1 1578-6 10. 1990.

Melegos DN, Yu H. Allen LC. et al: Prostate specific antigen in amniotic fluid of normal

and abnormal pregnancies. Clin Biochem (in press). 1 996.

Graves HCB. Wehner N. Starney TA: Ultrasensitive radioirnmunoassay of prostate-

specific antigen. Clin Chem 38:735-742. 1992.

Klee GG, Dodge LA. Zincke H. et al: Measurement of serum prostate specific antigen

using [Mx prostate specific antigen assay. J Urol 15 1 :94-98. 1994.

Rock RC, Chan DW. Bruzek D. et al: Evaluation of a monoclonal irnmunoradiometric

assay for prostate-specific antigen. Clin Chem 3 3 : 2 X 7-226 1. 1 987.

Vessella RL. Noteboom J. Lange PH: Evaluation of the Abbott IMx automated

immunoassay of' prostate-specific antigen. Clin Chem 3 8:2044-2054, 1992.

Vihko P. Kurkela R, Ramber~ J, et al: Time-resolved immunofluorometric assav of

1%

human prostate-specific antigen. Clin Chem 36:92-95. 1990.

Yu H. Diarnandis EP: Measurement of prostate specific antigen level in the serum of

fernales and prostatectornized males with an ultrasensitive h u n o a s s a y technique. J Ur01

153:1004-1008. 1995.

Giai M. Yu H. Katsaros D. et al: Comparison of prostate specific antigen in the serurn

of normal women and wornen with breast cancer before or afier surgery. Br J Cancer

72:728-73 1. 1995.

Melegos D. Diarnandis EP: Diagnostic value of molecular forms of prostate specific

antigen for female breast cancer. Clin Biochem 29: 193-200. 1996.

Yu H. Diarnandis EP. Levesque MA. et al: Ectopic production of prostate specific

antigen by a breast turnor metastatic to the ovary. J Clin Lab Analysis 8251-253. 1994.

Levesque M, Yu H, D'Costa M, Tadross L, and Diarnandis EP. Immunoreactive prostate

specific antigen in lung tumors. I Clin Lab And 9 375-379, 1995.

Pummer K. Wimsberger G. Purstner P. et al:. False positive prostate specific antigen

values in the sera of women with renal ce11 carcinoma. J Ur01 148:2 1-23. 1992.

Walters ;MR. Steroid hormone receptors and the nucleus. Endocrine Rev 6 : s 12- . 1985.

Evans RM. The steroid and thyroid hormone receptor superfamily. Science 240389-895,

1988.

Miller J, McLachlan AD, Klug A. Repetitive zinc-binding domains in the protein

transcription factor IIIA fiom Xenopus oocytes. EMBO J 4: 1609- 16 14, 1 985.

Hollenbern SM- Evance RM. Multide and coo~erative trans-activatinrr domains of the

huma. glucocorticoid receptors. Ce11 55 : 899-906, 1988.

Carson-Junca MA, Schrader WT, O'Malley BW. Steroid receptor family: structure

and fùnctions. Endocrine Rev 1 1 :20 1, 1990.

Smith LI, Bodwell JE, Mendel DEI, Ciradelli T, North WG, Munck A. Identification of

cystein-644 as the covalent site of attachent of dexamethasone 2 1 -mesylate to munne

glucocorticoid receptors in WEH 1-7 cells. Biochemistry 27: 3 747-3 753, 1988.

Westphal HM, Beato M. The activated glucocorticoid receptor of rat liver. Purification

and physical characterization. Eur J Biochem. 106:395-403, 1980.

Strahle U, Klock G, Schutz G. A DNA sequence of 15 base pairs is sufficient to mediate

both glucocorticoid and progesterone induction of gene expression. Proc Nat1 Acad Sci

USA 84:7871-7875, 1987.

Petrangeli El Lubrano C, Ortolani F, Ravenna L, Vacca A, Sciacchitano L, Frati A, Gulino

A. Estrogen receptors: New perspectivein breast cancer management. .i Steroid Biochem

Molec Bio1.49:327-33 1, 1994.

Schule R Muller M, Kaltschmidt C, Renkawitz R. Many transcription factors interact

synergistically with steroid receptors. Science 242: 14 1 8- 1420, 1988.

Hoeck W, Rusconi S, Groner B. Down-regulation and phosphory!ation of glucocorticoid

receptor in cultured cells. Investigations with a monospecific antiserum against a

bactenally expressed receptor fragment. J Bi01 Chem 264: 14296- 14402, 1989.

Vessey MP. The involvement of oestrogen in the development and progression of breast

disease: epidemiological evidence. Proc R Soc Edinb 95B :3 5, 1989.

Katzenellenbogen BS, Kendra KL, Norman MJ, et al. Proliferation, hormone

158

responsiveness and estrogen receptor content of MCF-7 human breast cancer cells grown

in the short-term and long-term absence of estrogens. Cancer Res 47:4355, 1987.

Osborne CK, Hobbs K, Clark GM. Effect of estrogens and antiestrogens o n growth of

human breast cancer cells in athymic rnice. Cancer Res 45584, 1985.

Clark GM, McGuire WL, Hubay CA, et ai. Progesterone receptors as a prognostic factor

in stage II breast cancer. N Engl I Med 309: 1343, 1983.

Honvitz KB, McGuire WL, Pearson OH, et al. Predicting response to endocrine therapy

human breast cancer: a hypothesis. Science 189:726,1975.

Ravdin PM, Green S, Dorr TM, et ai. Prognostic significance of progesterone receptor

levels in estrogen receptor-positive patients with metastatic brest cancer treated with

tamoxifen: results of a prospective Southwest Oncology Group study. J Clin Oncol 10: 1284,

1992.

Santen Manni A, Harvey H, Redmond C. Endocrine treatment of breast cancer in women.

Endocrine Rev 1 1 :22 1-265, 1990.

Dreicer R, and Wilding G. Steroid hormone agonists and antagonists in the treatment of

cancer. Cancer Invest 10 27-4 1, 1992.

Read LD, and Katzenellenbogen BS. Characterization and regulation of estrogen and

progesterone receptors in breast cancer. In Genes, Oncogenes and Hormones: Advances in

Cellular and Molecular Biology of Breast Cancer (Edited by R. B. Dickson and M. E.

Lippman). Kluwer-Nij hoff, Dordrecht, Net herlands pp. 277-299,1992.

Dickson RB, Lippman ME. Estrogenic regulation of growth and polypeptide growth factor

secretion in human breast carcinoma. Endocrine Rev 8 : 29-43, 1 987.

159

Clarke C L, Sutherland RL. Progestin regulation of cellular proliferation. Endocrine Rev

11:266-301, 1990.

Horwitz KB. The molecular biology of RU486. 1s there a role for antiprogestins in the

treatment of breast cancer? Endocrine Rev 13 : 149- 163, 1992.

Osborne CK, Yochowitz MG, Knight WA III, and McGuire W L. The value of estrogen

and progesterone receptors in the treatment of breast cancer. Cancer, 462884-2888, 1980.

Clark GM and McGuire WL. Steroid receptors and other prognostic factors in primary

breast cancer. Semin Oncol 15 :20-25, 1988.

Evans RM. The steroid and thyroid hormone receptor superfarnily. Science, 240:889-895,

1988

Carson-Jurica Mq Schrader WT, and O'Malley B W S teroid receptor family : structure and

hnctions. Endocrine Rev 1 1 :20 1-220, 1990.

McDonnelI DP, Vegeto E, Gleeson MA : Nuciear hormone receptors as targets for new dmg

discovery. Biotechnology 1 1 : 1259- 126 1, 1993.

Onate SA, Estes P 4 Welch WJ, Nordeen SK, and Edwards DP. Evidence that heat shock

protein-70 associated with progesterone recepton is not involved in receptor-DNA binding.

Molec Endocrin 5 : 1993-2004, 199 1.

Pratt WB. The role of heat shock proteins in regulating the function, folding and traffcking

of the glucocorticoid receptor. J Bi01 Chem 268:2 1,455-458, 1993.

Smith DF, Toft DO. Steroid receptors and their associated proteins. MoIec Endocrinol 714-

11, 1993.

FawelI SE, Lees JA., White R, and Parker MG. Characterization and localization of steroid

160

binding and dimerization activities in the mouse estrogen receptor. Ce11 60:953-962, 1990

Guiochon-Mante1 A, Loosfelt H, Lesxop P, Sar S. Atger M. Perrot-Applanat M, and

Milgrom E. Mechanisms of nuclear localization of the progesterone receptor: evidence for

interaction between monomers. Ce11 57: 1 147- 1 1 54, 1 989.

Kumar V, and Chambon P. The estrogen receptor binds tightly to its responsive element as

a ligand-induced homodimer. Ce11 55 145- 156, 1987.

DeMarzo AM, Beck CA., Onate SA, Edwards DP. Dimenzation of mammaiian

progesterone receptors occurs in the absence of DNA and is related to the release of the 90-

kDa heat shock protein. Proc Natl Acad Sci USA 88: 72-76.199 1 .

Tmss MK, Beato M. Steroid hormone receptors and interaction with DNA a ~ ~ d transcription

factors. Endocrine Rev 14:459-479, 1993.

Bagchi MK, Tsai M-J, OIMalley BW, and Tsai SY. Analysis of the mechanism of steroid

hormone receptor-dependent gene activation in cell-fiee systems. Endocrine Rev 14525-

535, 1992.

Man GF, Leng X, Tsai S, Weigel NL, Edwards DP, Tsai M-J, and OIMalley BW. Hormone

and antihormone induce distinct conformational changes which are central to steroid

receptor activation. J Bi01 Chem 267 19: 5 13- 520, 1992.

McGuire WL, Carbone PP, and Vollmer EP. Estrogen receptors in human breast cancer.

New York, Raven Press 1 975.

Lippman ME, Haiterman RH, Leventhal BG, Perry S, and Thompson EB. Glucocorticoid

binding proteins in human acute lymphoblastic leukaemia blast cells. J Clin Invest 52: 171 5-

1725, 1973.

16 1

Hsu RS. Androgen receptors in human prostates. In: Normal and Abnormal Growth of the

Prostate. Springfield, Illinois, Charles C Thomas, 1975; pp. 663-675.

Poilow K, Lubbert FI, Boguoi E, Kreuzer G, and Pollow B. Characterization and comparison

of receptors for 17P-estradiol and progesterone in human proliferative endometnurn and

endometrial carcinoma. Endocnnology, 96: 3 19-328, 1975.

Fisher RI, Neifield ME, and Lippman ME. Estrogen receptors in human rnalignant

melanoma. Lancet, 2: 3 3 7-339, 1976.

Carnadall Mord T, Do KM., Geeihoed GW, Tsangaris T, Lipprnan ME. Steroid hormone

receptors in human colon cancers. cancer, 43 :980-984, 1979

Beattie CW, Hansen NW, Thomas PA. Steroid receptors in human lung cancer. Cancer

Res, 49420642 14, 1985.

Gagne D, Pons M, PhiIbert D. RU 38486: A potent antiglucocorticoid in-vitro and NI-vzvo.

Steroid Biochem, 23 :247-25 1, 1985.

Waeber B, Nussberger J, and Brunner HR. Clinical applications of antimineralocorticoids.

Steroid Biochem, 3 1 :739-743, 1988.

Jordan VC. Biochernical pharmacology of antiestrogen action. Pharm Rev, 36245-276,

1984

Baulieu EE. Contragestion and other clinicd applications of RU 486, an antiprogesterone

at the receptor. Science, 245: 13 5 1 - 13 57, 1989.

Neumann F, von Berswordt-Wallrabe Elger W, Steinbeck H, Hahn ID, and Krarner M.

Aspects of androgen-dependent events as studied by antiandrogens. Rec Prog Hom Res,

26:337-341, 1970.

162

Boring CC, Squires TS, Tong T, et al. Cancer Statistics, . CA, Cancer J Clin, 44:7, 1994

Magdelenat H, Pouillart P, Steroid hormone receptors in breast cancer. In: Sheridan PJ,

Blum K, Trachtenberg MC, eds. Steroid receptors and disease: cancer autoimmune, bone

and circulatory disorders. New York, Marcel Dekker, 43 5, 1988.

McGuire WL, and Char1 GM. The prognostic role of progesterone receptor levels in breast

cancer. Sernin Oncol, 10:2-6, 1983.

Radvin PM, Green S, Don TM, et al. Prognostic significance of progesterone receptor levels

in estrogen-positive patients with metastatic breast cancer treated with tamoxifen: results

of a prospective Southwest Oncology Group study. J Clin Oncol, 10: 1284- 129 1, 1992.

Reigrnan PHJ, Kiaassen P, Van der Korpat, JAGM, Romijin JC, Trapman J. Molecular

cloning and characterization of novel prostate-antigen gene cDNA's. Biochem Biophys Res

Comun, 1%: 18 1-8, 1988.

Schuiz P. Stucka R, Feldmann H, Combnato G, Klobeck HG, Fittler F. Sequence of a cDNA

clone encompassing the complete mature prostate specific antigen (PSA) and an unsplised

leader sequence. Nucleic Acid Res, l6:6226, 1 988.

Ferguson, EU, Yu H, Kalyvas M, Zammit S, and Diamandis EP. Ultrasensitive detection

prostate specific antigen by time resolved imrnunofluorometric assay and the Imrnulite"

Immunochemiluminescent Third Generation Assay: potential applications in prostate and

breast cancers. Clin Chem, 42375-684, 1996.

Deguchi T, Doi T, Ehara H, Ito S, Takahashi Y, Nishino Y, Fujihiro S, Kawamura T,

Komeda H, Horie M, Kaji H, Shimokawa K. Tanaka T, Kawada Y. Detection of

micromet astatic prostate cancer cells in lymph nodes by reverse transcriptase-polymerase

163

chah reaction. Cancer Res, 53.33 50-53 54, 1993.

Okazaki T. Detection of estrogen receptor (ER) mRNA by use of reverse transcriptase-

polymerase chah reaction (RT-PCR) assay; comparison with dextran coated charcol @CC)

assay and imunocytochemicai assay. Jpn J Cancer Chemother, 19:36 1-366, 1992

Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Laboratory Manual. Cold Spnng

Harbor, NY: Cold Spring Harbor Laboratory, 1989.

Bert hois Y, Katzenellenbogen JA, Katzenellenbogen B S. Phenol red in tissue culture

medium is a weak estrogen: implications conceming the study of estrogen responsive cells

in culture. Proc Nat1 Acad Sci USA, 83:2496-2500, 1986.

Hart C, Spira T, More I, Schochetman G, Lifson, 4 Galphin J, Ou C-Y. Direct detection

HIV RNA expression in seropositive subjects. Lancet, ii:596-599, 1988.

Lundwail A. Characterization of the gene for prostate specific antigen, a huma. glandular

kallikrein. Biochem Biophys Res Commun, 16 1 : 1 1 5 1 - 1 1 59, 1989.

Klobeck HG, Combriato G, Schulz P, Arbuson V and Fittler F. Genomic sequence of

human prostate specific antigen (PSA). Nucleic Acids Res, 17398 1, 1989.

Diamandis EP, Yu H. Prostate cancer, prostate specific antigen and the polymerase chah

reaction. CIin C hem, 4 1 : 1 77- 1 79, 1995.

Moreno JG, Croce CM, Fischer R, Monne M, Vihko P, Mulholland SG, Gomella LG.

Detection of hematogenous micrometastasis in patients with prostate cancer. Cancer Res,

52:6I 10-61 12, 1992.

Katz AE, Olsson CA, Raf5o AJ, Carna C, Perlman H. Molecular staging of prostate cancer

with the use of an enlianced reverse transcriptase-PCR assay. Urology, 43:765-775, 1994.

164

Jaakkola S, Vomanen T, Leinonen J, Rannikko S, Stenman W. Detection of prostatic cells

in peripheral blood: Correlation with semm concentrations of prostate specific antigen. Clin

Chem, 41: 182-186, 1995.

Smith Biggar S. Hussain M. Prostate specific antigen messenger RNA is expressed in

non-prostate cells: Implications for detection of micrometastases. Cancer Res, 552640-

2644, 1995.

Levine AC. Pathogenesis and medical management of benign prostatic hyperplasia. Trends

Endocrinol Metab, 6: 128- 132, 1995.

Henttu P, Liao S. and Vihko P. Androgens up-reguiate the human prostate-specific antigen

messenger ribonucleic acid (mRNA), but dom-regulate the prostatic acid phosphatase

mRNA in the LNCaP cell Iine. Endocnnology, 130:766-772, 1992.

Riegman PH, Vliestra RI, Van der korput IAGM, Bnnkman AO, and Trapman J. The

promoter of the prostate specific antigen çene contains a functional androgen responsive

element. Mol Endocrinol, 5 : 192 1- 1930, 199 1.

Luke MC, and Coffey DS. Human androgen receptor binding to the androgen response

element of prostate specific antigen. I Androl, 15 :4 1-5 1, 1994.

Diamandis EP, Yu H. New biological functions of prostate specific antigen? J Clin

Endocrino1 Metab, 80: 15 15- 15 17, 1995.

Ojasso T, Dore JC, Gilbert J, and Raynaud P. Binding of steroids to the progestin and

glucocorticoid receptors analyzed by correspondence analysis. J Med Chem,3 1 : 1 160- 1 169,

1988.

Glover JF, and Darber FD. Multihormone regulation of MMTV-LTR in transfected T-47D

human breast cancers. J Steroid Biochem, 32:3 57-363, 1989.

20 1. Nordeen SK, Kuhnel B, Lawler-Heavner J, and Barber DA, Edwards DP. A quantitative

cornparison of dual control of hormone response element by progestins and glucocorticoids

in the same ce11 line. Mol Endocrinol; 3 : 1270- 1278, 1989.

202. Kafissen ME. A norgestimate-containing oral contraceptive: Review of clinical studies.

Am J Obstet Gynecol, 167: 1 196- 1202, 1992 .

203. Phillips 4 Hahn DW, and McGuire L. Preclinical evaluation of norgestimate, a progestin

with minimal androgenic activity. Am J Obstet Gynecol, 167: 1 19 1 - 1 196, 1992.

204. Phillips A. The selectivity of a new progestin. Acta Obstet Gynecol Scand Suppl, l52:Z 1 -

24, 1990.

205. Zava DT, Landrum B, Horwitz KB, and McGuire WL. hdrogen receptor assay with

['~]meth~ltrienolone (RI 88 1) in the presence of progesterone receptors. Endocrinology,

104: 1007-1012, 1979.

206. Teutsch G, Goubet F, Battmann T, Bonfils A, Bouchoux F, Cerede E, Gofflo D, Gaillard-

Kelly M, and Philbert D. Non-steroidal antiandrogens: synthesis and biological profile of

high-fity Ligands for the androgen receptor. J Steroid Biochem Molec Biol, 48: 1 1 1 - 1 19,

1994.

207. Philibert D, Moguilewsky M, Mary 1, Lecaque D, Tournemine C, Secchi J and Deraedt R.

Pharmacological profile of RU 486 in animals. In: The antiprogestin steroid RU 486 and

human fertility control. Beaulieu EE, Segal SJ (editors): pp 49-68. Plenum Publishing

Corporation, 1985.

208. Gravanis A Schaison G, George M, Debrux J, Satyaswaroop PG, Baulieu EE, and Robe1

P. Endometrial and pituitary responses to the steroidai antiprogestin RU 486 in

postmenopausal women. J Clin Endocrinol Metab, 60: 156- 163, 1985.

Gronemeyer H, Meyer M-E, Bocquel M-T, Kastner P, Turcotte B and Chambon P. .

Progestin receptors: Isofoms and antihormone action. J Steroid Biochem Molec Biol,

4O:27 1-278, 199 1

Pearce, and Yamamoto KR. Mineralocorticoid and glucocorticoid receptor activities

distinguished by non-receptor factors at composite response element . Science, 259: 1 16 1 -

1165, 1993.

LeRoith D. Insulin-like growth factors and cancer. Ann Int Med, 122 (1): 54-59, 1995.

Mobbs BG and Liu Y. Immunohistochernicai localkation of progesterone receptor in

benign and malignant human prostate. Prostate, l6:245-25 1, 1 990.

Carroll RS, Glowacka D, Dashner K, and Black PM. Progesterone receptors in

meningiomas. Cancer Res, 53 : 13 12- 1 3 16, 1993.

Bergqvist A, Bergqvist D, and Femo M. Estrogen and progesterone receptors in vesse1

walls: biochemical and immunological assays. Acta Obstet Gynecol Scand, 72: 10- 1 6, 1 993

Pacchioni D, Revelli A, Casetta G, et ai. Immunochemical detection of estrogen and

progesterone receptors in the normal urinas, bladder and in pseudomembranous trigonitis.

J Endocrinol Invest, 1 5 : 7 19-725, 1992.

Boivin G, Terrier-Anthoine C, and More1 G. Ultrastructural locaiization of endogenous

hormones and receptors in bone tissue: an irnmunocytological approach in fiozen samples.

Micron, 25 15-27, 1994.

Clements J, Yap, B. Mukhtar 4 Ehrlich 4 Salamonsen L. Prostate specific antigen and

167

glanduiar kailikrein expression in the human endometrium across the rnenstniai cycle. Proc

Int Congress Endocrinol, pg 840, No p3-344, 1996.

218. Collaborative group on hormonal factors in breast cancer. Breast cancer and hormonal

contraceptives: Collaborative reanalyses of individual data on 53,297 women with breast

cancer and 100,239 women without breast cancer from 54 epidemiological studies. Lancet

l996;347: 17 13- 1727.

2 19. Cramer SD, Zuxiong C, and Peehl DM. Prostate specific antigen (PSA) cleaves parathyroid

hormone related protein in the PTH-like domain: inactivation of PTHrP-stimulated CAMP

accumuIation in mouse osteoblast. J Urol, l56:526-53 1, 1996.

220. Fichtner J, Graves HCB, Thatcher K, Yemoto C, and Shortliffe LMD. Prostate specific

antigen releases a kinin-like substance on proteolysis of seminal vesicle fluid that stimulates

smooth muscle contraction. J Urol, 1 5 5 : 73 8-742, 1996.

22 1 . Diamandis EP. BRCAl protein products: Antibody specificity, functional motifs and

secreted tumour suppressor. Nature Genetics, 13:268, 1996.

222. Lai LC, Erbas H, Lennard TWJ, Peaston RT. Prostate specific antigen in breast cyst fluid:

possible role of prostate specific antigen in honnone dependent breast cancer. Int J Cancer,

66:743-746, 1996.

223. Bilgrami, S, Singh NT, Shafi N, and Ciesielski T. Raised prostate specific antigen in

adenocarcinorna of lung. Lancet, 344: 137 1- 1372, 1994.

224. Saedi MS, Cass MUI, Goel AS, Grauer L, Hogen KL, Okaney T, Griffin BY, Klee GG,

Young CY-F, and Tindal DJ. Overexpression of a human prostate specific glandular

kallikrein, hK2, in E. coli and generation of antibodies. Mol Ce11 Endocrïnol, 1 O9:Z 7-24 1,

1995.

225. Lovgren .J, Pironen Tl Overmo C, Dowell B, Karp M, Pettersson K, Lilja H, and Lundwall

A. Production of recombinant PSA and EX2 ande analysis of their immunologie cross-

reactivity. Biochem Biophys Res Communs, 2 13 : 888-895.1 995.

226. Beer DG, and Malkinson AM. Localization of specific ['H]dexamethasone binding in

urethan-induced mouse lung tumors. Cancer Res, 44:3 546-3 5 53, 1984.

227. Zargharni N, Diamandis EP. Detection of prostate specific antigen mRNA and protein in

breast tumors. Clin. Chem, 42:X 1-366, 1996.

i i v i n u L ~ v n L U ~ I I U I Y

TEST TARGET (QA-3)

APPLIED & IMAGE. lnc - = 1653 East Main Street - -. - Rochester. NY 14609 USA -- -- Phone: 716/482-0300 -- -- - - F a : 71 61288-5989

O 1993. A p p i d image. lm.. AI1 Rightç Reserved