 

 

 

 

 

ﬂ

T‘HLSﬁo

f)
MlCHlGAN STA E LIBRAR E

LIBRARY | llllllllllulllIlllllllllllllllll
Michigan State .. 31293017700554
‘ University ‘

 

i Q)‘\~——i‘ ’\

 

 

This is to certify that the

dissertation entitled

ACINAR MORPHOGENESIS IN NON-MALIGNANT
HUMAN PROSTATIC EPITHELIAL CELLS
AND ITS LOSS IN MALIGNANT CELLS

presented by

Diana Bello-DeOcampo

has been accepted towards fulfillment
of the requirements for

_Eh_._D_n___ degree in M

q 'Major professor
Datma ‘ l. l i IQ '
\l V

M5 U is an Afﬁrmative Action/Equal Opportunity Institution 0-12771

r

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINE: return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

WW I) '22 09

05‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/98 COW“

ACINAR MORPHOGENESIS IN NON-MALIGNANT HUMAN
PROSTATIC EPITHELIAL CELLS AND ITS LOSS IN
MALIGNANT CELLS

By

Diana Bello-DeOcampo

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Zoology

1999

ABSTRACT

ACINAR MORPHOGENESIS IN NON-MALIGNANT HUMAN PROSTATIC
EPITHELIAL CELLS AND ITS LOSS IN MALIGNANT CELLS

By

Diana Bello-DeOcampo

During the multi-step process of carcinogenesis in epithelial cells, multiple,
cumulative, genetic and epigenetic changes occur as cells progress from a normal to a
malignant, invasive and metastatic phenotype. During tumor progression, additional changes
lead to the development of autonomous, growth factor and hormone independent, highly
invasive metastatic cells. Little is known about the multiple steps in prostate carcinogenesis
and tumor progression. For the present study, it was hypothesized that during prostate
carcinogenesis and progression, the following changes occur: 1) that there is a progressive
loss of the ability of epithelial cells to undergo acinar morphogenesis, by polarization, to form
well deﬁned acini with lumens; 2) that adhesion of cells to laminin is altered and this occurs
as a result of changes in the expression of laminin integrin receptors which bind cells to
laminin; 3) that the loss of abnormal expression of integrins imparts to the cells increased
invasive ability; and 4) that cells show altered response to both stimulatory and inhibitory
growth factors, which occurs as a result of changes in the level of expression of growth
factors and their receptors.

Using a family of seven human prostate epithelial cell lines, which include a human
papillomavirus l 8 immortalized, non-tumorigenic cell line RWPE- l , and several progressively
more tumorigenic cell lines, derived from RWPE-l cells, a novel, 3-dimensional human
prostate cell culture model was developed to test the hypotheses. In this cell model, RWPE-l
cells mimic normal epithelial acinar morphogenesis as it occurs in vivo. Immunocytochemical
and confocal microscopy methods were employed. Results show that RWPE-l cells form
acini only when plated in a matrix of Matrigel or laminin, but not in collagen or ﬁbronectin.
When binding of cells to laminin is blocked by antibody to laminin, cells fail to form acini.

RWPE-lcells show strong expression of the laminin-speciﬁc 0L6Bl integrin receptor at their

basal end. When the 0L6 or [31 integrin subunits were blocked by their anti-functional antibody,
the ability of cells to form acini was reduced. This loss was more dramatic with the [31
integrin antibody. The six tumorigenic cells lines showed a progressive loss of acini forming
ability with increasing malignancy. While the least malignant WPE l -NA22 cells showed 0L6Bl
integrin expression similar to that seen in RWPE-l cells, a loss of a6 and diffuse expression
of B1 intergrin was observed in the most malignant, WPEl-NBZ6 cells. This abnormal
integrin expression showed a direct relationship to the loss of ability of cells to undergo acinar
morphogenesis and their increased invasive ability.

Exposure of RWPE-l cells to growth stimulatory exogenous EGF caused a dose-
dependent decrease in acinar formation. These results suggest that, because of increased cell
proliferation, malignant cells lose the ability to form acini and the degree of loss is related to
abnormal integrin expression and increased invasive ability. Further, the growth inhibitory,
exogenous TGF-Bl and B2 were more effective inhibitors of growth of RWPE-l and
WPEl-NA22 cells than that of WPE] l-NBZ6 cells. Both TGF-Bl and B2 enhanced the ability
of WPE] -NA22 cells to form acini and exposure of cells to anti-functional TGF-B antibody
resulted in a loss of polarization and acini forming ability. While RWPE-l cells expressed
endogenous TGF-B (TFG-BI>TGF-BZ>TGF-B3), and TGF-B R11 and 111, there was a
progressive loss of expression of both the TGF-B and its receptors with increasing
malignancy.

Taken together, these results demonstratethat during prostate carcinogenesis and I
tumor progression, there is a progressive loss of the normal expression of laminin integrin
receptors. Increased autocrine stimulation by growth factors, such as EGF, and the decrease
or loss of TGF-B and its receptors, the key negative regulators, leads to a loss of growth
homeostasis. Thus, changes in the expression of laminin integrins, growth factors and their
receptors, contribute to the inability of cancer cells to undergo acinar morphogenesis. This
is associated with increased invasive ability. These results facilitate the development of agents
which may be used to restore normal acinar morphogenesis and, thus, block carcinogenesis,

invasion and metastasis.

 

To my family and my mentor

iv

Acknowledgments

I would like to take this opportunity to recognize all the people who directly and
indirectly contributed to the work in this dissertation. I convey my gratitude to my mentor,
Dr. Mukta M. Webber, for her support, guidance and overall tenacity without which I
would have not been able to complete this work. But most importantly, I thank her for
believing in me. I thank my committee members, Dr. Stephen C. Bromley, Dr. Thomas K.
Huard, Dr. Richard Walshaw, Dr. David D. Wartinger and Dr. Daniel E. Williams and our
collaborators at NIH, Dr. Hynda K. Kleinman and Dr. Matthew P. Hoffman for their advice
and support. I especially wish to acknowledge and extend my appreciation to Dr. Shirley
A. Owens, Dr. Joanne H. Whallon and fellow graduate student, Timothy L. Murphy, at
the MSU laser Scanning Microscopy facility for their invaluable assistance, cooperation
and friendship. Additionally, I thank fellow lab rats Salmaan Quader, William W. Chu
Rebekah M. Sharp and Brandi Purcell for their support and friendship.

I extend my gratitude and appreciation to my mother and father, Siomara M. and
Rene J. Bello, for their continuing unconditional support and encouragement and my
sister, Stephanie Marie, for keeping them distracted. Finally and most importantly, I
extend my warmest appreciation to my husband, Nestor David DeOcampo, for just being

there when I needed him most and not complaining about doing housework.

TABLE OF CONTENTS

List of Tables ................................................................................... xi
List of Figures ................................................................................... xii
List of Abbreviations ........................................................................... xv
Introduction .................................................................................... I
Hypotheses ...................... I ........................................................ 3
Chapter One. The Human Prostate: background ....................................... 5
Anatomy ........................................................................ 6
Development of the Human Prostate ....................................... 7
Prostate Cancer ............................................................... 18
Chapter Two. Materials & Methods ....................................................... 29
Materials ................................................................................ 30
Media and Cell Culture Reagents ........................................... 30
Antibodies ...................................................................... 30
Growth Factors and Hormones ............................................. 31
Extracellular Matrix (ECM) ................................................ 31
Other Materials ............................................................... 31
Methods ................................................................................. 32
Cell Lines ...................................................................... 32
Cell Culture .................................................................... 33

vi

Cell Proliferation in Monolayer Cultures .................................. 34
EGF and mibolerone combined dose response assay ............ 34

Transforming growth factor (T GF)-Bl, B2 & B3
dose response assay .................................................. 35

Immunocytochemical analysis of using

monolayer cultures ................................................... 35
Laser scanning microscopy ......................................... 37
In Vitro 3-Dimensional Culture Assays ................................... 38
Acinar morphogenesis assay ...................................... 38

Immunocytochemical analysis of
Matrigel cultures ..................................................... 39

Induction of functional differentiation

by androgen in Matrigel cultures .................................. 41
Analysis of individual basement membrane components ....... 42
Induction of acinar formation assay ................................ 43
Inhibition of acinar formation assay ................................ 44

EGF effects on acinar morphogenesis .............................. 44
Invasion assay ................................................................ 46
Chapter Three. Extracellular Matrix and Integrins: background ................ 50
Epithelial cell polarity ........................................................ 51
Extracellular matrix (ECM) .................................................. 52
Integrins ........................................................................ 59
Prostate epithelial markers .................................................... 69

vii

Chapter Four. Acinar Differentiation by Non-malignant Immortalized Human

Prostatic Epithelial Cells and Its Loss by Malignant Cells ........ 76
Abstract ......................................................................... 77
Introduction .................................................................... 78
Materials and methods ......................................................... 81

Materials ............................................................... 81
Methods ................................................................ 81
Results .......................................................................... 84

Morphogenesis in 3-Dimensional cultures of RWPE-l,
non-tumorigenic cells and immunostaining for laminin,
0L6 and BI integrins, and PSA ....................................... 84

Comparison of the ability for structural morphogenesis
by the non-tumorigenic RWPE-l cells with that of the

tumorigenic RWPE-2 and DU-145 cells in 3-D cultures ........ 88

The invasive ability shows an inverse relationship

with the ability to undergo acinar morphogenesis ................ 88
Discussion ...................................................................... 90

Chapter Five: Extracellular Matrix and Integrins Modulate Acinar

Morphogenesis by RWPE-l Cells ............................................. 101
Abstract ....................................................................... 102
Introduction .................................................................... 103
Results .......................................................................... 106

laminin is required for acinar morphogenesis ................... 106
01 integrin is essential for acinar formation ...................... 111

viii

Discussion ..................................................................... 1 14

Chapter Six. Malignant Human Prostatic Epithelial Cells Lose the Ability to

Undergo Acinar Morphogenesis ........................................... 125
Abstract ................................................................................. 126
Introduction ............................................................................ 127
Results .................................................................................. 129

Loss of acinar morphogenesis in 3—D cultures

by malignant cell lines ....................................................... 129

Invasive ability shows an inverse relationship with the

ability to undergo acinar morphogenesis .................................. 133

Alterations in integrin or 601 expression in malignant cell lines ........ 134

Deregulation of growth by EGF inhibits acinar morphogenesis ........ 137
Discussion ............................................................................. 141

Chapter Seven. Regulation of Acinar Morphogenesis by TGF—B in Non-malignant

and Malignant Human Prostatic Epithelial Cells .................. 152
Abstract ........................................................................ 153
Introduction ................................................................... 155
Results ......................................................................... 157

Effects of TGF—B isoforms on the growth of RWPE-l cells ..157

RWPE-l cells express TGF-Bl, 2 and 3
and the TGF-B R11 and R111 receptors ............................ 159

TGF-Bl and TGF-BZ are necessary for acinar morphogenesis. 159

Effects of TGF-B isoforms on the growth of WPEl-NA22
and WPEl-NB26 cells ............................................... 162

ix

Alteration in TGF-Bl, 2 and 3 and TGF-B R11 and
TGF-B RIII receptor expression in malignant WPEl-NA22
and WPEl-NB26 cells .............................................. 164

Restoration of acinar-forming ability by
TGF-Bl and TGF-BZ ................................................ 168

Discussion ..................................................................... 169

List of Tables

Table 2-1. Antibodies and conditions for immunocytochemistry
on monolayer cultures ............................................................. 37

Table 2-2. Antibodies and conditions for immunocytochemistry
of Matrigel cultures ................................................................ 41

Table 2-3. Neutralizing antibodies and conditions for inhibition of

acinar formation assay ............................................................ 46
Table 3-1. Properties of basement membrane components ............................... 55
Table 3-2. Integrins and their ECM binding sites ........................................... 63

xi

List of Figures

Figure 1-1. Three-dimensional model of the three glandular zones of the prostate ...... 6
Figure 1-2. Differentiation of the urogenital sinus in the human male .................... 8
Figure 1-3. High grade PIN demonstrates 4 architectural patterns ...................... 21
Figure 1-4. Proposed model for prostatic carcinogenesis .................................. 22
Figure 3-1. Pseudo-stratiﬁed prostatic epithelial cell layer ................................ 54
Figure 3-2. A current model of basement membrane structure ............................ 55
Figure 3-3. Schematic model of Laminin-l ................................................... 57
Figure 3-4. A model of laminin-entactin/nidogen complex ................................. 58
Figure 3-5. A schematic structure of integrins ................................................. 61
Figure 3-6. The integrin family ................................................................ 61
Figure 3-7. Integrin receptor binding to basement membrane ............................ 62
Figure 3-8. Active sites for integrins .......................................................... 65
Figure 4-1. Appearance of RWPE-l cells on Matrigel ..................................... 85
Figure 4-2. 3-D cultures ........................................................................ 87
Figure 4-3. Cell organization on Matrigel and relationship with invasion ............... 89
Figure 5-1. Acinar morphogenesis by RWPE-l cells in Matrigel ...................... 107
Figure 5-2. Optical serial sections of RWPE—l acinus grown in Matrigel ............ 107

xii

Figure 5-3. Optical serial sections of RWPE-l acinus grown in laminin-1 ............ 109
Figure 5-4. RWPE-l cells on ECM matrices. ............................................. 109
Figure 5-5. Acinar formation on matrix components ..................................... 110

Figure 5-6. Inhibition of acinar formation by anti-laminin neutralizing antibody 110

Figure 5-7. a6 Integrin expression ........................................................... 112
Figure 5-8. IgG control ....................................................................... 112
Figure 5-9. 01 Integrin expression .......................................................... 112
Figure 5-10. a2 Integrin expression ......................................................... 113
Figure 5-11. [34 Integrin expression ......................................................... 113

Figure 5-12. Inhibition of acinar formation by anti-integrin
neutralizing antibodies ......................................................... 115

Figure 6-1. Acinar formation of immortalized RWPE-l cell line
and malignant prostate epithelial cell lines ................................... 131

Figure 6-2 Acinar formation of immortalized non—tumorigenic and
tumorigenic prostate epithelial cell lines grown in Matrigel .............. 132

Figure 6-3. Invasive ability of non-tumorigenic immortalized and

malignant human prostate epithelial cells ..................................... 135
Figure 6-4. Relationship between acini-forming ability and invasive ability ........... 135
Figure 6—5. (1601 laminin integrin expression ............................................. 136

Figure 6-6. Effects of EGF on the growth of RWPE-l cells and its modulation

by androgen mibolerone .......................................................... 138
Figure 6—7. Reduction of acinar formation by EGF ........................................ 140
Figure 6-8. Enhancement of acinar formation by EGF neutralizing antibody ............ 140

xiii

Figure 7-1. Effects of TGF-Bl, TGF-BZ and TGF-B3 on growth of
RWPE-l cells in monolayer cultures ........................................... 158

Figure 7-2. TGF-Bl, TGF-B2 and TGF-B3 expression in non-malignant
RWPE-l and malignant human prostatic epithelial cells .................... 160

Figure 7-3.TGF-B R11 and TGF-B RIII expression in non-malignant
RWPE-l and malignant human prostatic epithelial cells ..................... 161

Figure 7-4.Reduction of acinar formation by anti-TGF-B neutralizing antibodies ..... 163

Figure 7-5. Reduction of acinar formation by anti-TGF-B neutralizing
receptor antibodies ................................................................ 163

Figure 7-6. Effects of TGF-B l, TGF-BZ and TGF-B3 on growth of

WPEl-NA22 cells in monolayer cultures ..................................... 165
Figure 7-7. Effects of TGF-Bl, TGF-B2 and TGF-B3 on growth of

WPEl-NB26 cells in monolayer cultures ..................................... 166
Figure 7-8. Effects of TGF-Bl on the growth of RWPE-l, WPEl-NA22 and

WPEl—NBZ6 cells in monolayer cultures ..................................... 167
Figure 7-9. Enhancement of acinar formation by TGF-B ................................. 170

xiv

5 a-DHT

ABC

AR

BPE

BSA

CK-14

CK-18

D-PBS

DCS

EGF

HPV-18

IF

INT

K-SFM

LSM

MoAb

NMU

PIN

PSA

PSF

TGF-B

Abbreviations

........................................................... 5-0t-dihydrotestosterone
............................................................ Avidin—Biotin Complex
................................................................ Androgen Receptor
........................................................... Bovine Pituitary Extract
............................................................. Bovine serum albumin
...................................................................... Cytokeratin 14
...................................................................... Cytokeratin 18
.......................................... Dulbecco's Phosphate Buffered Saline
.................................................................. Donor Calf Serum
......................................................... Epidermal Growth Factor
....................................................... Human Papilloma Virus 18
............................................................ Intermediate Filaments
.................................................... P-Iodo-nitrotetrazolium violet
............................................... Keratinocyte-Serum Free Medium
...................................................... Laser Scanning Microscope
.............................................................. Monoclonal Antibody
........................................................... N—nitroso-N-methlyurea
................................................ Prostatic intraepithelial neoplasia
......................................................... Prostate Speciﬁc Antigen
.......................................... Penicillin, Streptomycin & Fungizone

.................................................. Transforming Growth Factor B

XV

Introduction

Adenocarcinoma of the prostate is the most common type of cancer, excluding skin
cancer, in adult men in the United States. It is estimated that over 184,500 men will be
diagnosed and 39,200 men will die of prostate cancer in 1998, thus, making prostate
cancer the second leading cause of cancer death in the US. (Landis et al., 1998).
Incidence of prostate cancer increases with age and about 80% of prostate cancers are
diagnosed in men over the age of 60 (Carter and Coffey, 1990). Autopsy studies show that
approximately 11 million men older than 45-50 years in the US. have latent carcinoma
of the prostate and it is predicted that one in ten men will develop prostate cancer in his
lifetime (Carter and Coffey, 1990, Foster, 1990). Although cases of precancerous lesions,
referred to as prostatic intraepithelial neoplasia (PIN), appear in men in their 20's and
30's, the majority of prostate cancer cases are diagnosed in men their 60's and 70's at an
advanced stage of the disease when invasion and metastasis have already occurred
(Bostwick, 1992). Presently, little is known about the causes of prostate cancer. Therefore,
a great deal of importance and attention is being placed on the need for information on the
mechanisms involved in carcinogenesis and on treatment strategies for the prevention of
the progression of precancerous, latent lesions to clinical cancer, as well as, effective
strategies for the treatment of the advanced disease. However, research in these areas has

been slow.

One of the major obstacles faced by researchers is the lack of suitable human cell
models and in vitro culture systems which would mimic the in viva micro—environment.
Currently, the use of human prostate epithelial cells for the study of prostate cancer is the
best available model system aside from human experimentation which is basically
unfeasible and unethical. Immortalized human prostatic cell lines provide uniform,
standardized and reproducible models for the study of the multi-step process of
carcinogenesis and the alterations that occur in malignant transformation. Such cells lines
have been developed and fully characterized in Dr. Webber's laboratory at Michigan State
University. Prostatic epithelial cells from a donor were immortalized with the human
papilloma virus-18 (HPV-18), thus establishing the RWPE-l cell line (Rhim et al., 1994
and Bello et al., 1997). Additional cell lines, which represent different stages of
carcinogenesis, were developed by transfection with the ras-oncogene or exposure of
RWPE-l cells to a carcinogen. RWPE-l cells were transformed by the addition of a Ki-ras
oncogene to produce the tumorigenic RWPE—2 cell line (Rhim et al., 1994 and Bello et a1. ,
1997). Other RWPE-l cells were treated with the chemical carcinogen, N—nitroso-N-
methlyurea (NMU), to establish cell lines of varying tumorigenicity: WPEl-NA22,
WPEl-NBll, WPEl-NB14, WPEl-NB26 and WPEl-NB27 (Webber, personal
communication). Collectively, these cells lines represent different stages of neoplastic
transformation from initiated non-invasive, non-tumorigenic to highly invasive,

tumorigenic cells.

Hypotheses:

The main focus of this study is the development of a 3-dimensional (3-D) culture
system, which mimics in vivo prostatic acinar morphogenesis, and its application to study
the factors involved in the progression of carcinogenesis utilizing the unique characteristics
of the newly developed cell lines. Speciﬁcally, the role of the individual extracellular
matrix components and their integrin receptors, as well as growth factors and their
receptors, in epithelial polarization and acinar formation and how they are altered in
malignant cells, was examined. The proposed hypotheses are that:

O RWPE-l cells undergo structural and functional differentiation when grown
on Matrigel, a reconstituted basement membrane.

0 Laminin is the key extracellular matrix component necessary for acinar
morphogenesis.
0 Cell attachment and epithelial polarization require the basal expression of

aGBI laminin integrin receptor.

6 TGF-B and EGF play an essential role in the epithelial organization and
formation of acini.

6 Neoplastic cells lose the ability to undergo acinar morphogenesis.

0 Loss of normal c1601 laminin integrin expression results in the loss of ability
to undergo acinar morphogenesis.

O Progression through the multi-step processes of carcinogenesis involves at least
two sequentially unique steps:

1. Deregulation of growth involving alterations in the expression and
responsiveness to growth factors.

2. Alterations in the 0L6BI laminin integrin receptor expression.

References

Bello D, Webber MM, Kleinman HK, Wartinger DD and Rhim JS: Androgen responsive
adult human prostatic epithelial cell lines immortalized by human papillomavirus
18. Carcinogenesis 18: 1212-1223, 1997.

Bostwick, D.G.: Prostatic intraepithelial neoplasia (PIN): current concepts. J. Cell.
Biochem. 164:10-19, 1992.

Carter, H.B. and Coffey, D.S.: The prostate: increasing medical problem. Prostate 16:39-
48,1990.

Foster, C.S.: Predictive factors in prostatic hyperplasia and neoplasia. Hum. Pathol.
21:575-577, 1990.

Landis SH, Murray T, Bolden S and Wingo PA: Cancer Stats, 1998. CA Cancer J. Clin.
48:1,1998.

Rhim JS, Webber MM, Bello D, Lee MS, Amstein P, Chen L and Jay G: Stepwise
immortalization and transformation of adult human prostate epithelial cells by a
combination of HPV-18 and v-Ki-ras. Proc. Nat. Acad. Sci. USA, 91:11874-

11878, 1994.

Chapter One

The Human Prostate:
Background

The Human Prostate

The prostate is an exocrine gland with an average weight of 20 grams. The adult
prostate gland lies caudal to the urinary bladder and cranial to the external urethral sphincter
(Rous et al., 1988). It is composed of glands, muscle and ﬁbrous tissue encapsulated by
connective tissue (McNeal, 1989, Rous et al., 1988). Although the prostate gland is not a
primary sex gland, it is considered an accessory sex gland because of its indirect involvement

in procreation. The secretion of the prostate gland constitutes a major component of semen.

The prostatic glands empty their secretion into prostatic ducts, that in turn, empty into the
prostatic urethra, where this secretion mixes with sperm from the testes that arrive via the
ejaculatory ducts. The prostatic secretion ﬁlnctions as a buffered vehicle for spermatozoa and
provides them nourishment (Rous, et al., 1988). This review will focus on the normal and the

malignant prostate.

Anatomy

 

The prostate is divided into the
ventral ﬁbromuscular portion and a
dorsal glandular portion. The glandular
prostate is composed of three zones: the

peripheral (75%), the central (15%) and

 

the transition (5%) (Figure 1-1). The

 

 

 

peripheral zone surrounds the Figure 1-1. Three-Dimensional model of the three
glandular zones of the prostate. Peripheral (P),

posterolateral peripheral aspect and transitional (T) and central (C ) zones of the prostate
(Lee et al., 1989).

extends ﬁ'om the apex to the base. It is

aspect and extends from the apex to the base. It is composed of small, simple acini lined
by columnar secretory epithelial cells. The central zone surrounds the ejaculatory ducts.
The acini in the central zone are large, irregularly contoured and are lined by low
columnar to cuboidal epithelium. The transition zone is formed by two bilaterally
symmetrical lobules lateral to the pre-prostatic sphincter. The transition zone is
histologically similar to the peripheral zone but differs in that its stroma is more dense and
compact (McNeal, 1989, Miller, 1993). The prostatic stroma is primarily composed of
smooth muscles cells and ﬁbroblasts which reside in the extracellular matrix (ECM). Due
to the lack of suitable in vitro human models and the obvious moral and ethical constraints
on the use of live human subjects for experimentation, the fetal and post-natal development
of the prostate has been studied extensively in the rodent model and the information has

been extrapolated to the human prostate.

Development of the Human Prostate

Embryonic development: The ﬁrst indication of prostatic development in humans
is seen at the twelfth week (3rd month) of fetal life as outpouchings from the urethral
epithelium in the distal portion of the urogenital sinus (Figure 1-2). These urethral
epithelial cells bud out as prostatic buds into the adjacent urogenital mesenchyme both
above and below the entrance of the mesonephric ducts and the openings of the ejaculatory
duct (McNeal, 1989, Rous, 1988, Smith, 1972). By the fourth month of fetal life the
prostate buds reach their final number. Five paired buds are formed in all. The top pairs

of buds are thought to be of mesodermal origin and form the inner zones of the prostate.

 

5 weeks 9 weeks

 

‘\

 

' ' Ureter
-i./V‘ ..
~ . _ .. was motlerian
‘ . M _ tubercle
12 weeks 6 months

Ureter

  
   

Tubular

' Me: n h '
urogenital dugtep m
sinus
Prostate
OiIIereniiating
mesenchyma prostatic
epitheliai
buds

 

 

 

 

Figure 1-2. Differentiation of the urogenital sinus in the human male (Smith, 1972).

The outer zone of the prostate is formed by the lower buds and is thought to be of
endodermal origin (Aumuller, 1979, Coffey, 1992). These buds branch and re-branch, to
form determinate glands that are directed toward the bladder, where they wind spirally
around the urethra. They will eventually comprise three distinct zones: the central,
transitional and peripheral zones. The ﬁbromuscular tissue that was originally continuous with
the bladder, becomes entangled with the branching buds and becomes the prostatic stroma
(Aumuller, 1979, Coffey, 1992, Smith, 1972).

The initial “budding” of prostatic ducts from the urethral epithelium is stimulated by
5a-dihydrotestosterone (Sa-DHT). Testosterone from the fetal testis is metabolically
converted to Sa-DHT by Sa-reductase. Although both the embryonic prostatic epithelial cells
and the surrounding mesenchymal cells are capable of converting testosterone to 50t-DHT,
the epithelium produces signiﬁcantly more Sa-DHT than the mesenchyme. Interestingly
enough, the ﬁnal target of the androgenic inﬂuence is the adjacent mesenchyme not the
epithelium. The epithelium converts testosterone to Sa-DHT, which then diffuses out to the
mesenchyme. Although both the epithelial and mesenchymal cells express androgen receptor,
the androgen receptor in the epithelial cells in the developing prostate does not bind Sa-DHT.
It has been shown that the androgen receptor mRNA found in the epithelial cells of the
prostatic buds is, at least, several hundred bases shorter than that of the androgen receptor
mRNA in the mesenchymal cells. Exactly why this so is not yet clear (Timme et al., 1994).
The fact that the mesenchyme, and not the epithelium, is the target of androgen action has
been demonstrated quite elegantly by Cuhna and associates (1987 and 1996) using epithelial-
mesenchymal recombinant experiments. In such experiments, androgen-sensitive urogenital
sinus mesenchyme and androgen-insensitive urogenital epithelium, with the addition of

9

androgens, were able to form prostatic acini, while the converse, androgen-insensitive
urogenital mesenchyme and androgen-sensitive urogenital epithelium failed to form acini.

In other experiments using pregnant females, anti-androgens were used at different
stages of pregnancy to determine the importance of androgens in eliciting prostatic budding.
It was found that when anti-androgens were administered on day 16.5 day postcoitum,
prostatic budding was completely abolished. But if the anti-androgens were administered on
day 18.5, prostatic budding occurred. The action of androgen in this case, is thought of to be
a turn on switch. Once you turn on the switch for morphogenesis you do not need to keep
your hand on it, i.e. ﬁrrther stimulation in viva by androgen is not needed beyond the initial
induction (Donjacour and Cuhna, 1988).

To date, the generally accepted mechanism for prostatic budding from the urogenital
sinus may be visualized as follows: testosterone from the embryonic testis is converted to
Sa-DHT by the “prostatic” epithelial cells, Sa-DHT then diffuses to the mesenchymal cells
where it binds to the androgen receptor and induces (turns on) the production of an unknown
growth factor(s), which, in turn, stimulates the urogenital epithelial cells to form prostatic
buds. Two questions arise from this explanation: what are the growth factors secreted by the
mesenchyme that induce the budding and what role do adhesion molecules, ECM bound
proteases and growth factors play in modifying the extracellular matrix (ECM) which the
prostatic buds are invading? Unfortunately, very little information about human fetal and
postnatal budding and branching is available due to the lack of ﬁmctional model systems, but
work in postnatal rat branching morphogenesis, has been done and is generally extrapolated
to the human condition to fill in the gaps. The 3-D human prostate epithelial cell model

developed and described in the present work should serve as an in vitra model for human

10

prostate morphogenesis.

Postnatal Development of the Prostate: In general, postnatal development of the

prostate may be divided into three phases: 1) postnatal involution (regression) phase (1 month
after birth to 2 months); 2) infantile resting phase (2 months to 10-13 years); 3) maturation
phase (14 to 18-21 years). During postnatal development no new lobe formation is observed
(Aumuller, 1979 and 1991).

At birth, acini are predominantly lined with epithelium showing squamous epithelial
metaplasia which is thought to be caused by maternal estrogen. There is some scattered
secretory activity but these secretions are strictly mucosal. The postnatal involution phase
following birth occurs in response to the cessation of maternal steroids. During this time the
metaplastic epithelium is replaced by cuboidal psuedostratiﬁed epithelium (Aumuller, 1979
&l991, Coffey, 1992).

During the first three months aﬁer birth there are large hormone surges that result in
hormone imprinting of the prostate. These hormonal surges involve transient increases in the
serum levels of estrogen, androgen and progesterone in the early postnatal life in both humans
and rats. The estrogen surge is represented by the high serum estradiol levels seen at birth.
Estrogen levels dramatically fall in the ﬁrst few days after birth which results in involution
(Coffey, 1992). In human males, a surge in testosterone is observed between the second and
third months. Testosterone levels increase up to 60 times higher than normal pre-pubertal
levels. In fact, at the time of the surge, testosterone levels almost match adult testosterone
levels. Progesterone levels are high at birth in both animals and are thought to be of placental
origin. In the human neonate, there is another surge of progesterone at two months (Coffey,

1992). From rat studies, we know that theses surges are critical in hormonally imprinting the

11

prostate. Such imprinting subsequently determines the gland’s future growth regulation by
steroid hormone. This growth regulation by imprinting is illustrated by experiments where
exogenous estrogen administration to the neonatal rat reduced the sensitivity of the adult
animal to androgens (Naslund and Coffey, 1986). On the other hand, exogenous
administration of androgen or progesterone to the neonatal rat resulted in an increase in
sensitivity to androgens in adulthood. It is proposed that the neonatal hormone surges are
altering the properties of the prostatic stem cells or that they are perhaps altering stem cell
numbers. Although the mechanism remains unclear, it does appear that the hormone surges
act directly on the prostate, and not through the hypothalamic-pituitary axis (Naslund and
Coffey, 1986). Since this phenomena of neonatal and pre-pubertal steroids imprinting of the
prostate has been established as a critical factor in rats, it is thought that the same process
occurs in the human male.

Following the post-involution/regression phase, only minimal growth, development
and histological differentiation occur in the infantile prostate. When boys reach 10 years of
age the prostate begins to develop its typical external shape from the spherical shape it had
at birth. With the onset of puberty at 11 to 13 years of age, the maturation of the glandular
epithelium begins. This maturation phase lasts from 1 l- l 3 to 18-22 years of age. During this
time, the glandular ducts begin to grow in diameter, and they branch (branching
morphogenesis) and develop new glands. The usually thin infantile epithelium starts to
transform into a multilayered columnar epithelium. In the mouse this phase, which occurs
within the ﬁrst 15 days of postnatal life, is very well studied. It had always been assumed that
branching morphogenesis was an androgen-dependant process, but that notion is being

challenged. When mice were castrated 24 hours after birth, a decrease in the number of tips

12

and branch points in both dorsal and ventral prostates was observed but both lobes still
underwent signiﬁcant branching (25-45%). In order to counter any residual androgen activity
or androgen contribution from the adrenal gland, anti-androgen was administered to the
castrated animals. This treatment did not cause any additional suppression of branching.
Therefore, it can be concluded that branching morphogenesis is sensitive to, but does not
require, chronic androgen stimulation. The idea is that unlike the fetal condition, where
androgen levels are necessary at critical periods for initial prostatic budding, postnatal
branching morphogenesis does not require androgen stimulation. Even if the preceding is true,
androgen is essential for functional differentiation of the prostate in the adult male (Aumuller,
1979 &l991, Coffey, 1992, Naslund and Coffey, 1986).

Branching morphogenesis stops at 15 days after birth in rats. What gives the stop
command? Cessation of branching is not likely due to the increase in androgen levels, because
branching stopped at 15 days in both the non-castrated and castrated animals alike. One
attractive proposed mechanism is that the decrease in the ratio of stroma to epithelium causes
branching to stop, but in the castrated neonates branching stops at 15 days even though there
is an abundance of stroma available. But it must be noted that the stroma of castrated animals
is abnormal, in that it has larger amounts of collagen and other ECM materials (Donjacour
and Cuhna, 1988). This is possibly due to improper ECM remodeling, a process which may
indirectly require androgens for proper functioning. Furthermore, although it is tempting to
extrapolate this lack of androgenic response to the human condition, it must be noted that,
unlike humans, rats have a very low sensitivity to testosterone, especially during this

branching period.

13

Because postnatal branching morphogenesis increases the number of glands, the
amount and degree of maturation of its secretory products also increases. At birth the few
secretory cells present secrete mucous, which is continued throughout infancy. Functional
secretory maturation usually occurs between ages 18-22, as acini with distinct lumens are
developed after postnatal branching morphogenesis. Secretions go from mucosal to prostate-
speciﬁc. As early as 14 years of age, most glands have a more or less duct-like appearance
and show some immunoreactivity for prostatic acid phosphate (PAP) and prostate speciﬁc
antigen (PSA) . There are no mucous producing cells in the adult prostate as there are in
infancy and in the early post-pubertal prostate (Aumuller, 1979 &l99l).

At the completion of maturation, the adult prostatic acini are said to generally contain
three types of cells: basal epithelial cell, luminal epithelial cells and neuroendocrine cells
(Verhagen et al., 1992). The basal cells are considered to be the stem cells from which
luminal cells arise. Luminal cells are the terminally differentiated secretory cells of the
prostate gland. A fourth type of cell population, called amplifying cells, is thought to exist
in the prostate gland. These amplifying cells are considered to be cells in transition from

the basal to the luminal type (Verhagen et al., 1992).

Growth factors involved in prostatic development: Recently, keratinocyte growth

factor (KGF) and transforming growth factor-0 (TGF-B) have been proposed to mediated
the mesenchymal-epithelial interaction in the developing prostate (Cuhna, 1996). KGF is
a member of the ﬁbroblast growth factor family and is primarily produced by ﬁbroblasts
and mesenchymal cells but has a high degree of speciﬁcity for epithelial cells (Cuhna,
1996). While, KGF is primarily produced by the mesenchymal cells, the KGF receptor is

principally expressed by epithelial cells, where it induces cell proliferation. In vitra organ

14

culture experiments, using the ventral prostate of new born rats, have been used to
investigate the role of KGF during postnatal branching. KGF transcripts were found in the
new born ventral prostate; additionally, transcripts for its receptor were found speciﬁcally
in the epithelium. When organ cultures were grown in the presence of testosterone, the
ventral prostate ducts showed extensive branching but when grown in the absence of
testosterone branching was modest and eventually stopped all together. The addition of
antibodies against KGF into the culture medium containing testosterone inhibited ductal
branching. The branching patterns of organ cultures grown in the presence of testosterone
and anti-KGF antibodies resembled the pattern seen when cultures were grown in the
absence of testosterone. Thus, the anti-KGF antibodies inhibited the trophic effects of
testosterone. Most striking was the observation that cultures grown in the absence of
testosterone, but in the presence of KGF, demonstrated ductal growth and branching
morphogenesis patterns similar to those patterns seen in cultures treated with testosterone
alone. That is to say, KGF almost completely mimicked the growth and branching
morphogenesis seen in response to androgens (Cuhna, 1996). Based on the above
evidence, it is proposed that one of the unknown growth factors produced by the
mesenchymal cells of the urogenital sinus in response to 50t~DHT induced prostatic
budding is KGF.

Transforming growth factor-,6 (TGF-ﬂ): Another growth factor that has been
associated with prostatic development is TGF-B. Because of its apparent important role in
branching and acinar morphogenesis in viva, I have investigated its effects on acinar

morphogenesis in the 3-D cell culture model. These results are reported in chapter seven.

15

We have previously reported the effects of TGF-[3l on prostatic epithelium in monolayer
cultures (Webber et al., 1996, Bello et al., 1997).

TGF-[35 are members of a superfamily of polypeptides that regulate cell cycle
progression, differentiation, chemotaxis and metabolism of many different types of cells
(Habib and Chem, 1994). Their biological effect on a given target cell is dependant on the
cell type, the growth conditions and the presence of other growth factors. Although a
bifunctional response to TGF-B by both ﬁbroblast and epithelial cells has been reported,
TGF-B generally stimulates the growth of ﬁbroblasts but inhibits the growth of epithelial
and endothelial cells (Franzen et al., 1993, Roberts and Spom, 1990, Webber et al.,
1996). In addition to inhibiting epithelial cell growth, TGF-B stimulates the production of
extracellular matrix components and protease inhibitors and causes a decreases in the net
activity of extracellular matrix degrading proteolytic enzymes (Roberts and Sporn, 1990;
Franzen et al., 1993). Presently three form of TGF 0 (TGF-B ,, TGF-B 2, and TGF-B 3)
have been identiﬁed in the prostate.

Although TGF-B is expressed by both the urogenital mesenchyme and the
urogenital epithelium of the fetal mouse, the highest levels are detected in the mesenchyme
and not the epithelium. Speciﬁcally, the highest levels are in the areas of active epithelial
duct formation. TGF-B expression actually follows the path the epithelial cells take as they
penetrate into the surrounding mesenchyme. For proteolytic remodeling of the
mesenchyme, serine proteases, such as tissue and urokinase plasminogen activator (tPA
& uPA), plasmin (from plasminogen by uPA) and some metalloproteinases are the best

candidates, since they have also been found to be increased in the urogenital mesenchyme

16

(T imme et al., 1994, Wilson et a1. 1992). Cathepsin D expression has been found in the
basal portion of the prostatic epithelial cells as they penetrate, but this expression is lost
with formation of lumens (Wilson, 1995). The reciprocal relationship between TGF-B and
proteases has been somewhat elucidated. TGF-l3 isoforms must be cleaved from a latent
form into active molecules. Therefore, one of the functions of these proteases is to activate
matrix bound growth factors, such as TGF-B (Aumuller, 1991 and Timme et al., 1994).
The other function is to degrade the ECM. TGF-B and proteases work together in prostatic
morphogenesis in the following manner: TGF-B induces prostatic epithelial cells to
produces matrix proteins and causes an up-regulation of their integrin receptor expression,
thus, providing invading epithelial cells attachment sites and the needed ECM receptors
which enable them to move. Meanwhile, proteases degrade the newly laid down ECM,
keep the epithelial cells from differentiating by removing the matrix and release growth
factors, such as bFGF, thus, maintaining their proliferative state.

In summary, there is little change in the actual size of the prostate gland from the
time between birth till puberty. At puberty, in response to the increased levels of
testosterone, the prostate again begins to grow. As discussed above, this growth is the
result of signiﬁcant hyperplasia of the ductal epithelium and the formation of bud sites.
This branching morphogenesis, which leads to the elaboration of the duct system, the
development of new acini and a condensation of the stroma, is accompanied by a slow and
continuous increase in the size and mass of the prostate. By age 30, when growth stops,
the prostate has grown to twice its pre-pubertal size primarily due to the development of

new acini and related ducts. The prostate remains largely unchanged in size until 45-55

17

years of age. By age 60, most men develop benign prostatic hyperplasia (BPH) and/or
prostate cancer (Aumuller, 1979, Rous, 1988, McNeal, 1989). One objective of the
present study is to examine changes in integrin expression and possible associated changes
in the ability of prostate cells to polarize and undergo acinar morphogenesis. Such changes
may also affect the invasive ability of prostate cancer cells. Some characteristics of
prostate cancer, invasion and the precancerous changes, characterized by prostatic
intraepithelial neoplasia (PIN) will now be described. It should be noted that the set of
cell lines used in the present study may represent a continuum of neoplastic changes from

normal epithelium to invasive cancer (see chapter 6).

Prostate Cancer

With the ever increasing life span in the United States, adenocarcinoma of the
prostate has become a major health concern. If prostate cancer is detected when it is
pathologically conﬁned to the prostate, that is, it has not yet metastasized, it can be
surgically removed or successfully treated. Therefore, it is the goal of modern medicine
to not only be able to accurately diagnose prostate cancer but to do so early, while it is still
conﬁned to the prostate. That is why the American Cancer Society continues to
recommend yearly prostate check-ups to men over 40 as a regular annual assessment for
prostate cancer (Cupp et al., 1993). Unfortunately, even with the advances in early
detection, 60% of prostate cancer cases are diagnosed in men in their 60's and 70's, when

invasion and metastasis have already occurred (Miller, 1993; Cupp et al., 1993).

18

Invasion: Invasion by cancer cells involves invasion of the stroma, of the bladder,
and/or of the seminal vesicles. Stromal invasion is noted by the lack of basal cells and the
basement membrane which separates the acini from the stroma. The acini may also appear
to be haphazardly dispersed in the stroma and the acini themselves frequently have
irregular shapes. Additionally, there might be signs of individual or groups of cells
growing out of the acini and invading into the stroma. Bladder and periurethral invasion
is characterized by the presence of malignant cells in the lamina propria that have no
relationship to the surface epithelium (Mostoﬁ et al., 1992). Seminal vesicle invasion is
usually associated with cancers that also have lymph node metastasis and are larger than
12 mm, however this association does not negate the probability of small cancers being
over-looked (Villers et al. 1991).

Pathology: Prostatic carcinoma has a great variety of architectural patterns that
have been arranged into a "linear scale of progressive dedifferentiation" (McNeal et al.,
1986). These patterns have been used to establish an architectural grading system. Some
of these unique patterns may also serve as criteria for the diagnosis of prostate cancer
(McNeal et a1. , 1986). The pathologist looks for the following architectural disturbances
(Mostoﬁ et a1. , 1992). In the normal prostate, glands are uniformly large, convoluted, and
are comprised of basal and luminal cells arranged in a pseudostratiﬁed pattern (Miller,
1993). Each gland is surrounded by a basement membrane and is distributed in a nodular
pattern (Mostoﬁ et al., 1992). In prostate cancer, the acini are small and the cells are
arranged in. a single layer devoid of basal cells. These small acini are packed closely
together or they may form large acini of a single layer of cells with no convolutions. A
pattern of large and small acini arranged side by side is also seen (Mostoﬁ et al., 1992).

19

Another distinct pattern is that of small or large single layered acini fused to each other
without any stroma between them. Other arrangements such as cribriform or gland in
gland patterns also denote prostate cancer (Mostoﬁ et a1. , 1992).

It is interesting to note that a number of benign lesions may be confused with or
perhaps even mis-diagnosed as prostate cancer. These benign lesions are basal cell
hyperplasia, atrophy, unusual variants of hyperplasia and cribriform hyperplasia.

Precursor changes- prostatic intraepithelial neoplasia (PIN): Progress has been
exceedingly slow in identifying the precursor to prostate cancer due to the inability to
accurately re-biopsy the prostate (Miller, 1993). Candidates for precursor lesions have
been proposed such as Roman bridging, adenosis, clear cell cribriform hyperplasia, and
intraductal dysplasia. All but interductal dysplasia have been discredited as possible
prostatic precancerous lesions (Miller, 1993).

Intraductal dysplasia was ﬁrst described by McNeal and Botswick in 1986 (McNeal
et al., 1986). They felt it was a precursor lesion to prostate cancer because of its
similarities in nuclear pleomorphorphism and nuclear prominence to cancer and its
association with concurrent prostate cancer in specimens (Miller, 1993). In 1989, the term
prostatic intraepithelial neoplasia (PIN) was proposed and replaced intraductal dysplasia
which had also been referred to as large acinar atypical hyperplasia, hyperplasia with
malignant change, marked atypia and duct-acinar dysplasia (Miller, 1993, Bostwick,
1995).

Prostatic intraepithelial neoplasia is generally divided into two grades: high and low
(Bostwick, 1995). Low grade PIN is characterized by glands that show epithelial cell
crowding with enlarged nuclei that vary in size. An undisrupted basal epithelial cell layer

20

is present. High grade PIN displays four architectural patterns: micro papillary, tufting,
cribriform and ﬂat (Figure 1—3). PIN spreads much like prostate carcinoma. Three patterns
of spread have been described for PIN. In the ﬁrst pattern, neoplastic PIN cells appear to
replace the luminal epithelial cells and the basal cells appear unaltered. In the second

pattern, the neoplastic luminal-like PIN cells appear to invade the acinar wall, disrupting

some basal cells (Bostwick, 1995). In the third pattern, which is rarely found, the PIN
cells have invaginated between the basal cell layer and columnar secretory cell layer. Some
high grade PIN show severe fragmention of the basal layer and invasion. In fact, it can not
be distinguish from carcinoma in situ (Bostwick, 1995).

The incidence of PIN, similar to prostate cancer, increases with age. However,
unlike prostate cancer, which shows a higher prevalence in blacks than in whites, the

prevalence of PIN is equal in both (Botswick, 1995). PIN has been found in men in their

 

 

 

 

 

Figure 1-3. High grade PIN demonstrates 4 architectural patterns : rnicropapillary,
tufting, cribriform and ﬂat (Miller, 1993).

21

20's (9%) and 30's (22%) and is thought to predate the onset of carcinoma by 5 to 10
years (Bostwick, 1995). As in prostate cancer, the peripheral zone of the prostate is the
most common site for PIN. Other similarities between PIN and prostate cancer are:
elevated levels of collagenase type IV, EGF receptor, and growth factors, such as EGF and
TGF-0t. Oncogene products such as erbB-2 and erbB-3 and bcl-2 have been shown to be
elevated in PIN. In contrast, differentiated markers such as cytokeratins and
neuroendocrine cells are lost with increasing grades of PIN. Although serum PSA levels
are increased in PIN (Kim et al., 1995), PSA staining of PIN cells is decreased or lost
(Bostwick, 1995). An elegant model for prostate carcinogenesis has been proposed, based
on the multistep theory of transformation, using the morphologic continuum of PIN
(Figure 1-4). In this model, low grade PIN corresponds to very mild and mild dysplasia
while high grade PIN encompasses the progression from mild to severe dysplasia including
in situ carcinoma. PIN ends and prostate cancer begins when the neoplastic cells invade

the stroma after the basal cell layer is completely disrupted (Botswick et al., 1987).

 

Dysplasia Carcinoma

1

Normal 1

T I V V l’ 1
[Very mild Mild Moderate Several In Situ Microinvasion

 

 

 

I V
Lumenal 90’ g . ‘
(secretory) I“ g I 1"900'93’0 1' 53513980: 0‘060
“$1.331" $55.5. I D c Q) W» F f«9 6‘
CIII uvcmdcjc io'.:'”-...-.’ -..‘ -0“ .. g1’0:
3:39:32: 1 LOW Gfade 1 High Grade I . ,Ji/
membrane

Prostatic lntraepithelial Neoplasia

 

 

 

Figure 14. Proposed model for prostatic carcinogenesis (Bostwick, 1996).

22

Clinically, PIN has become of very important because it has a strong predictive
value as a marker for prostate cancer (Botwick, 1995). In a study where 37 patients were
biopsied, 14 (38%) showed presence of PIN. When they were biopsied again within eight
years, they had all developed prostate cancer (Bostwick, 1995). Another study showed
similar results, where 39% of the 104 men that had been diagnosed with PIN, had
prostate cancer upon follow up a few years later (Park et a1. , 1989 see Bostwick, 1995).
In fact one study found that the probability of developing cancer was also depended on the
type of PIN diagnosed in the initial biopsy. Forty percent of men diagnosed with high
grade PIN had cancer when re-biopsied 18 months later, while only 13 % of those with low
grade PIN showed cancer (Bostwick, 1995). One could argue that the cancer was present
at the ﬁrst biopsy but was just not sampled until the second, but that is highly unlikely if
an adequate amount of locations are initially biopsied. Although it is very much implied,
no studies to date have deﬁnitively been able to determine whether or not PIN progresses
to cancer as in the carcinogenesis model. It is felt that this is just a matter of time
(Bostwick, 1995).

It should be noted that the set of cell lines used in the present study may represent
a continuum of neoplastic changes from normal epithelium to invasive cancer (see chapter

six).

23

References

Aumuller G: Prostate gland and seminal vesicles. Berlin: Springer-Verlag. pp. 12-88.

1979.

Aumuller G: Postnatal development of the prostate. Bullletin de l’association des

anatomistes, 742229, 39—42, 1991.

Bello D, Webber MM, Kleinman HK, Wartinger DD and Rhim J S: Androgen responsive
adult human prostatic epithelial cell lines immortalized by human papillomavirus

18. Carcinogenesis 18: 1212—1223, 1997.

Bostwick DG: Prostatic intrapeithelial neoplasia (PIN): Current concepts. J Cell. Molec. Biol.

suppl. 16H10-l9, 1992.

Bostwick DG: Prospective origins of prostate carcinoma. Prostatic intraepithelial neoplasia

and atypical adenomatous hyperplasia. Cancer. 78:330-6, 1995.

Bostwick DG: Progression of prostatic intraepithelial neoplasia to early invasive

adenocarcinoma. Eur. Urol. 301145-52, 1996.

Coffey DS: Physiology of the male reproduction. In Walsh, , P.C. , Retik, A.B. , Stamey,
T.A., Vaughan, E.D.(eds): "Campbell’s Urology"sixth edition. Pennsylvania:
W.B. Saunders Company, pp. 224-246, 1992.

24

Cunha G: Growth factors as mediators of androgen action during male urogenital

development. Prostate supplement. 6: 22-25, 1996.

Cunha G and Donjacour A: Stromal-epithelial interactions inn normal and abnormal
prostatic development. Current concepts and approaches to the study of prostate

cancer. 239: 251-272, 1987.

Cupp MR and Oesterling JE: Prostate-speciﬁc antigen, digital rectal examination, and
transrectal ultrasonography: their roles in diagnosing early prostate cancer. Mayo

Clin. Proc.68:297-306, 1993.

Donjacour A and Cunha GzThe effects of androgen depravation on branching

morphogenesis in the mouse prostate. Dev. Biol., 128: 1-14, 1988.

Franzen P, Ichijo H and Miyazono D: Different signals mediate transforming growth

factor-B l-induced growth inhibition and extracellular matrix production in prostate

carcinoma cells. Exp. Cell Res., 207: 1-7, 1993.

Habib FK and Chem C: Pathogenesis of benign prostatic hyperplasia. In: Handbook on

benign prostate hyperplasia. Chisholm, G.D. (ed.), New York, Raven Press, pp

19-31, 1994.

25

Henttu P and Vinkop P: Growth factor regulation of gene expression in the human

prostatic carcinoma cell line LNCaP. Cancer Res., 53: 1051-1058, 1993.
Kim ED, Smith ND and Grayhack JT: Prostate speciﬁc antigen in the expressed prostatic
fluid of men with benign prostatic hyperplasia and prostate carcinoma. J. Urol.

154:1802-5, 1995.

Lee F, Torp-Pedersen ST and Siders DB: The role of transrectal ultrasound in the early

detection of prostate cancer. CA-Cancer-J-Clin. 392337-60. 1989.

McNeal J E: Anatomy an embryology. In Fitzpatrick, J. M. and Krane, R. J. (eds): "The

Prostate" Edinburgh: Churchill Livingstone, pp 3-7, 1989.

McNeal JE, Bostwick DG, Kindrachuk RA, Redwine EA, Freiha FS and Stamey TA:

Patterns of progression in prostate cancer. Lancet. 1:60-3, 1986.

Miller GJ: Anatomical and pathological considerations. In Das, S. and Crawford, E. D.

(eds): "Cancer of the Prostate" New York: Marcel Dekker, Inc, pp 13-17, 1993.

Mostoﬁ FK, Davis CJ Jr, and Sesterhenn IA: Pathology of carcinoma of the prostate.

Cancer. 70(1 Suppl):23S-53, 1992.

26

Naslund MJ and Coffey DS: The differential effects of neonatal androgens, estrogen and

progesterone on adult rat prostate growth. J. Urol., 136: 1136-1140, 1986.

Roberts AB and Sporn MB: The transforming growth factor-BS. In: Peptide growth factors
and their receptors 1. Spom, MB. and Roberts, A.B. (eds), New York, Springer-

Verlag, pp. 419-472, 1990.

Rous SN: The Prostate Book. New York : W. W. Norton & Co, pp 19-25, 1988.

Smith DR: General Urology. Canada: Lange Medical Publications, p. 22, 1972.

Timme TL, Trouong LD, Merz VW, Krebs T, Kadmon D, Flanders KC, Park SH and
Thompson T: Mesenchymal-epithelial interaction and transforming growth factor-B

expression during mouse prostate morphogenesis. Endocrinology, 134: 1039-1045,

1994.

Verhagen APM, Ramaekers FCS, Aalders TW, Schaafsma HE, Debruyne MJ and Schlken
J A: Co—localizationn of basal and luminal cell-type cytokeratins in human prostate

cancer. Cancer Res., 52: 6182-6187, 1992.

Villers A, McNeal JE, Freiha FS and Stamey TA: Multiple cancers in the prostate.

Morphologic features of clinically recognized versus incidental tumors. Cancer. 70:

2313-8, 1992.

27

Webber MM, Bello D and Quader S: Immortalized and tumorigenic adult human prostatic
epithelial cell lines: characteristics and applications. Part 1. Cell markers and

immortilized nontumorigenic cell lines. Prostate 26:386-394, 1996.

Wilson JD: Recent studies of the mechanism of action of testosterone. N. England J.

Med., 287: 1284, 1972.

28

Chapter Two

Materials & Methods

29

Materials

Media & cell culture reagents

Keratinocyte-serum free medium (K-SFM) 17005-042, Antibiotic/antimycotic
mixture (PSF) 15240-013, Dulbecco' 5 modiﬁed Eagle's medium (DMEM) 21063-029 and
RPMI-1640 11875-119 (Gibco-BRL, Grand Island, NY); Dulbecco’s phosphate buffered
saline Ca2+/Mg2+-free (PBS) 28374 (Pierce, Rockford, IL); donor calf serum 1110-90

(Intergin, Purchase, NY).

Antibodies

Monoclonal antibody (mAb) to a6 integrin, MAB1972, mAb to 0:2 integrin,
AB1936 and mAb to B4 integrin, ABl922 (Chemcon, Temecula, CA); integrin mAb to
(16 integrin, clone GoH3 (AMAC Inc. , Westbrook, ME); mAb to 01 integrin (gift of Dr.
S.K.Akiyama, NIDR, NIH); mAb to laminin, C-827, mouse IgG, M8273, rat IgG, 14131,
goat IgG, I5256 and cytokeratin 18 , C-8541 (Sigma, St. Louis, MO); mAb to cytokeratin
14, 10003 (Cappel, Durham, NC); mAb to prostate speciﬁc antigen (PSA), M-0750
(Dako, Carpinteria, CA); polyclonal Ab to androgen receptor (AR), PA1—110, (Afﬁnity
BioReagents, Neshanic Station, NJ); Vectastain Elite ABC Peroxidase Kit, PK-6102, 3,3'-
diamino-bezidine (DAB) Substrate Kit, SK-4100, Vectashield mounting medium, H1000,
Vectashield mounting medium with P1, H1300, Anti-mouse Fluorescent Kit, FI-2100,
Anti-rat Fluorescein, FI-4000, Anti-goat Fluorescein, FI-5000, Mouse IgG, I-2000,
Fluorescein Avidin D, A-2001, biotinylated anti-goat IgG, BA5000 and biotinylated anti-

rabbit IgG, BA 1000 (Vector laboratories, Burlingame, CA); anti-TGF-Bl mAb, MAB240,

30

anti-TGF-BZ, AB—112-NA, TGF-B3, AB-244-NA, anti-EGF, MAB236, anti-TGF-Bl
mAb, MAB246, anti-TGF—BZ, MAB 612, TGF-B3, M643, anti-TGF-RII, AF-24l-NA and
TGF-RIII, AF-242-PB (R&D Systems, Minneapolis, MN); anti FAK, F15020

(Transduction Laboratories, Lexington, KY).

Growth factors and hormones

Epidermal growth factors (EGF), 40001, transforming growth factor (TGF)-Bl,
240-B, TGF-B2, 302-B2, TGF-B3, 243-B3 (R&D Systems, Minneapolis, MN);
mibolerone, W-300 (BIOMOL Biomolecular Research Laboratories, Plymouth Meeting,

PA).

Extracellular matrix (ECM)

Phenol red—free Matrigel, 4023C and growth factor reduced Matrigel, 40230,
(Collaborative Scientiﬁc, Lincoln Park, NJ); collagen type IV, 3410-01 and laminin,
3400-10 (Trevigen.; Gaithersburg, MD); human ﬁbronectin, 40008A and gelatin, G1890

(Sigma; St. Louis, MO).

Other materials

HEMA-3 Stain Set, 67-56-1 (Curtin-Matheson, Wood Dale, IL); 24 well plates
(Falcon), 08-7721; 12 mm circle coverslips, 12-545-80 and cell strainers (Falcon), 4011

pore size, 2340 (Fisher Scientiﬁc, Itasca, IL); Nuclepore membrane, 8;; pore size,

1550446, 96 well plates, 3596 (Costar, Cambridge, MA).

31

Methods

Cell lines

RWPE-l, RWPE-2, WPEl-NA22, WPEl-NBll, WPEl-NB14, WPEl-NB26,
WPEl-NB27 and DU-145 cells lines were used in this study. The RWPE-l cell line was
derived from non-neoplastic human prostatic epithelial cells by immortalization with
human papilloma virus 18. These cells mimic normal prostatic cell behavior in their
response to growth factors and androgen and in their expression of PSA and androgen
receptor in response to androgen stimulation. They do not form tumors in nude mice nor
are they invasive (Bello et al. , 1997). RWPE—l cells were transformed by the addition of
a Ki—ras oncogene to produce the tumorigenic RWPE-2 cell line. These cells show altered
response to growth factors and androgen as compared to the parent RWPE—l cell lines, but
do express PSA and androgen receptor in response to androgen stimulation. They form
tumors in nude mice and are anchorage independent and invasive (Rhim etal., 1994 and
Bello et al., 1997). WPEl-NA22, WPEl-NBll, WPEl-NB14, WPEl-NBZ6 and
WPEl-NB27 were derived in the following manner: RWPE-l cells were treated with the
chemical carcinogen, N-nitroso-N—methlyurea (NMU) and injected into nude mice
(Webber, personal communiation). Tumors were harvested, cells were isolated and grown
in culture. These cells were then grown in agar, colonies were isolated and placed back
into culture. After a second round through nude mice and agar, the WPEl-NA22,
WPEl-NBI 1, WPEl-NB14, WPEl-NBZ6 and WPEl-NBZ7 cell lines were established.
Collectively, these cell lines represent different stages of neoplastic transformation from

initiated non-invasive, non-tumorigenic to highly invasive, tumorigenic cells.

32

DU-145 cell line is a commercially available prostate cancer cell line derived from

a human brain metastasis. It was used as a negative control.

Cell Culture

RWPE—l, RWPE-2, WPEl-NA22, WPEl-NB11,WPE1-NB14, WPEl-NB26 and
WPEl-NB27 cells were maintained in basal Keratinocyte-Serum Free Medium (K-SFM)
supplemented with 50 ug/ml bovine pituitary extract (BPE), Sng/ml epidermal growth
factor (EGF), and 1% antibiotic/antimycotic mixture(PSF) which will be referred to as
complete K-SFM. Basal K-SFM medium refers to K—SFM medium which lacks BPE and
EGF. The DU-145 human prostate carcinoma cell line was maintained in RPMI-1640
medium containing 5% donor calf serum (DCS) and 1% PSF. Cells were passaged upon
conﬂuence and seeded at 2 x 106 cells/T—75 ﬂask. For passaging, cells were washed with
Ca+ +/Mg-free PBS, then incubated with 3 ml of trypsin: EDTA mixture (0.05% trypsin,
0.02 mM EDTA) diluted 1:1 with Ca”/Mg-free PBS for 7 min. Cells were dislodged by
tapping, neutralized by equal part addition of 2 % DCS in CaH/Mg-free PBS, and
recovered by centrifugation (2000 rpm) for 5 min. Cell counts were preformed with a

Coulter Counter.

33

Cell Proliferation in Monolayer Cultures

EGF and mibolerone combined dose response assay: The combined effect of EGF
and the synthetic androgen, mibolerone, on growth of RWPE-l cells, was investigated.
EGF was dissolved in sterile distilled water. EGF stocks were stored at -70°C until use.
Mibolerone was dissolved in ethanol and was protected from light and stocks were stored
at -70°C until used. The ﬁnal concentration of vehicle in the medium was 0.1% which was
not growth inhibitory. 5,000 cells were plated/well in triplicate, in 96 well plates in
complete K—SFM Forty-eight hours after plating, medium was changed to basal K-SFM
supplemented with BPE containing either 0 or 5 nM mibolerone and a range of EGF
concentrations (0, 1.25, 2.5 and 10 ng/ ml). Plates were ﬁxed at 5 days after treatment and
prepared for absorbance as follows: plates were washed twice with normal saline to
remove the dead cells and media, ﬁxed in 95% ethanol for l h and stained with 0.5 %
methylene blue for 45 min. During this time the dye is taken up by the ﬁxed cells. The
cells were then washed twice with distilled water and incubated at 37°C for 24 h in 1%
SDS to extract the dye. Absorbance was read at 620 nM using a Titertek microplate
reader. Becuase of the direct relationship between cell number and absorbance, the
absorbance data were plotted against time, analyzed and cell doubling times were
determined for low and high cell densities (Webber et al., 1996). All experiments were

conducted in triplicate.

34

T ransforming growth factor (TGH-ﬂl, ,62 & ,63 dose response assay: The effects
of growth factors TGF-13,, TGF—B2 and TGF-B3 on growth of RWPE-l , WPEl—NA22 and
WPEl-NB26 cells were investigated. All three TGF-B isoforms were dissolved in 4 mM
HCl containing 1 mg/ ml BSA. The ﬁnal concentration of vehicle in the medium was 0.1%
which was not growth inhibitory. 10,000 cells were plated/well in triplicate, in 96 well
plates in complete K-SFM. Forty-eight hours after plating, medium was changed to
complete K-SFM medium containing TGF-Bl, TGF—B2 or TGF-B3 at concentrations ranging
from 0 to 10 ng/ ml. Plates were ﬁxed at 5 days after treatment and prepared for

absorbance as described above. All experiments were conducted in duplicate.

Immunocytochemical analysis of using monolayer cultures: RWPE-l , WPEl-
NA22 and WPEl-NB26 cells were grown on sterile coverslips in 24 well plates at a
density of 20,000 cells/400 #1 in complete K-SFM medium/well. Upon sub-conﬂuence,
cells were rinsed twice with PBS and ﬁxed in a 1:1 solution of methanol and acetone for
2 min with gentle agitation. The coverslips were then either processed that day or stored
in a -20°C freezer. Immunocytochemistry was performed using an indirect immuno-
peroxidase method. Immunostaining experiments were done in duplicate following a
modiﬁed Vector protocol (Webber et al., 1996). All of the following steps were carried
out at room temperature unless otherwise indicated. Cells were blocked in normal horse
serum for 1 hr at room temperature and incubated with the appropriate monoclonal
antibody (mAb) diluted in normal horse serum. Table 2-1 shows speciﬁc antibody
dilutions, incubation times and temperature for each antibody used. Cells were stained

using either ﬂuoresceine-tagged anti-IgG secondary Ab (1:200/1 h) or biotinylated

Table 1. Antibody and conditions for immunocytochemistry

of monolayer cultures

 

 

 

 

 

 

 

 

 

 

Antibody Dilution Incubation time Temperature
TGF-[31 ‘3 1:100 2 h 21°C
TGF-[32 "3 1:100 2 h 21°C
TGF-[33 "3 1:100 2 h 21°C
TGF-B receptor II (Rll) 3-4 1:50 2 h 21°C
TGF-B receptor In (Rlll) 3-4 1:50 2 h 21°C

 

1, Mouse
2, Rabbit

36

 

anti-goat/rabbit IgG secondary Ab (1:200/ 1 h) followed by an avidin-ﬂuoresceine (1:200/
1h). Cells were washed twice in PBS over 10 min between all successive steps after
primary Ab application. Cells were washed with in PBS and mounted in Vectashield
mounting medium onto alcohol washed slides. Using a laser scanning confocal microscope
cells were determined to be either + or - for a given antigen as compared to controls that
received non-speciﬁc IgG as a control instead of primary antibody. All experiments were

conducted in duplicate.

Laser scanning microscopy

Immunocytochemical and in virra 3-D culture assays were analyzed using a Ziess
10 Laser Scanning Confocal Microscope housed in the laser scanning microscopy (LSM)
laboratory at Michigan State University. The LSM laboratory is a university wide facility
providing state—of-the-art light microscopy technology to the MSU research community.
Confocal microscopes allow for the examination with minimal distortion to the 3-D
architecture of the specimen. The outstanding feature of the confocal microscopy is its
ability to make "optical sections", images of thin, horizontal slices of ﬂuorescing
specimens. In conventional microscopy, light reaches the observer both from the narrow
plane, which is in focus (the focal plane image), and from the out-of-focus regions above
and below it. The out of focus light severely degrades the quality of the focal plane image
and limits the depth to which a specimen can be examined. Consequently, details such as
lumens, are not readily seen or easily discriminated. The classical method for overcoming
this problem is to parafﬁn-embed the specimen and slice it mechanically into thin serial

sections for conventional light microscope viewing. In contrast, with the confocal

37

microsc0pe, the specimen is scanned point-to—point with a ﬁnely focused laser beam. A
pinhole aperture is placed directly in front of a photomultiplier detector which results in
the exclusion of the out-of—focus light from the ﬁnal image. This produces an "optical
section", which is directly comparable to what would be seen if the specimen had been
mechanically cut at that exact position. However, in this case the specimen remains intact
with no disruption of its 3-D architecture and it is available for subsequent optical
sectioning in different planes. Additionally, light transmitted (non-confocal) images
collected, using a laser scanning microscope are sharper, clearer and crisper than images

captured with a conventional Tungsten lamp microscope.

In Vitra 3-Dimension Culture Assays

Acinar morphogenesis assay: Phenol red-free Matrigel, referred to as Matrigel,
was used. To the wells of 96-well plates, 50 pl of Matrigel (0.3 mg/50 m1) kept on ice,
were added and the plates were placed in a 37°C incubator for 1 h to allow the matrix to
gel. RWPE-l, RWPE-2, WPEl-NA22, WPEl-NBll, WPEl-NB14, WPEl-NB26,
WPEl-NB27 and Du-145 cells were plated at 10,000 cells/well in complete K-SFM with
6 wells/cell lines. Cells plated on plastic served as monolayer controls. Medium was
Changed every 48 h. At 4 days, cultures were washed gently with PBS, ﬁxed in 2.5%
buffered paraformaldehyde for 15 min, rinsed twice with PBS and then either processed
that day or stored in PBS at 4°C (Webber et al., 1997). Acinar formation was examined
by confocal microscopy of propidium iodide (1.5 ug/ ml) stained cultures using the criteria

of four or more spherically arranged polarized cells with lumens. For staining, cultures

38

were permeabilize using a 1:1 solution of methanol and acetone (3 min) and aqueously
mounted with Vectashield mounting medium with propidium iodide or stained with
acridine orange (15 min.) and aqueously mounted with Vectashield on alcohol washed
slides. Acinar counts were performed at low magniﬁcation using a laser scanning
microscope (LSM). Lumen formation was veriﬁed by randomly visualizing counted acini
using laser scanning confocal microscopy. The number of acini formed was expressed as
percent of control using RWPE-l acinar formation as 100%. All experiments were

conducted in triplicate.

Immunocytochemical analysis of Matrigel cultures: Immunocytochemistry was
performed using an indirect immuno-ﬂuorescence staining following a modiﬁed Vector
protocol. All of the following steps were carried out at room temperature unless otherwise
indicated. Cultures were blocked in 1.5% normal horse serum/PBS for 1 hr at room
temperature and incubated with the appropriate monoclonal antibody (MAb) diluted in
blocking solution. Table 2—2 shows speciﬁc antibody dilutions and incubation times for
each antibody used. Cultures were then stained using either ﬂuorescence tagged anti-
mouse/rat IgG secondary Ab (1:200/ 1 h) or biotinylated anti-goat or rabbit IgG secondary

Ab (1:200/ 1 h) followed by an avidin-ﬂuoresceine (1:200/ 1 h). Cultures were washed

39

Table 2. Antibody and conditions for immunocytochemistry

of Matrigel cultures

 

 

 

 

 

 

 

 

 

 

Antibody Dilution Incubation time Temperature
Anti-ca integrin "2 1:100 2 h 21°C
Anti-a6 integrin 17" 1:100 2 h 21°C
Anti-[31 integrin 3'4 1:100 2 h 21°C
Anti-[34 integrin 2'5 1:100 2 h 1 21°C
Anti-Laminin ‘1 1:200 2 h " 21°C
Prostate specific antigen (PSA) "2 1:20 24 h 4°C

 

 

1, Mouse
2, RWPE-1

3, RWPW-1, WPE1-7, WPE1-10, WPEl-NA22, WPE‘l-NBZ7 and WPEl-NBZG

4, Rat
5, Rabbit

40

 

twice in PBS over 10 min between all successive steps after primary Ab application.
Cultures were washed in PBS and mounted in Vectashield mounting medium onto alcohol
washed slides. Using a laser scanning confocal microscope, acini were determined to be
either + or - for a given antigen as compared to controls that received non-speciﬁc IgG

instead of the primary antibody. All experiments were done in duplicate.

Induction of functional differentiation by androgen in Matrigel cultures:
Experiments to examine the effects of androgens on morphogenesis and expression of
differentiated function employed mibolerone treatment using the acinar morphogenesis
assay described earlier. Twenty hours after plating RWPE-l cells on Matrigel, medium
was Changed to complete K-SFM medium containing 5 nM mibolerone. Medium was
subsequently changed every 48 h. After 6 days of mibolerone treatment, prostate speciﬁc
antigen (PSA) expression was detected by a modiﬁed avidin-biotin immunoperoxidase
Vector protocol, using monoclonal Ab. Matrigel cultures were washed 2X with PBS, ﬁxed
in 2% paraformaldehyde, parafﬁn embedded, sectioned at 51:, mounted on poly-L-lysine
coated slides and re-hydrated for immunocytochemical analysis. Control sections lacked
primary antibody. All sections were washed twice in 500 #1 PBS over 10 min between all
successive steps after primary Ab application. The following sequential steps were
conducted at room temperature unless otherwise noted: sections were blocked with 1.5 %
normal horse serum dissolved in PBS for 1 h, incubated with PSA Ab was diluted 1:20
in blocking solution for 24 h, followed by biotinylated secondary Ab (1:200) for 30 min,
treated with 3 % H202 for 3 min to quench endogenous peroxidase activity, incubated with

the avidin-biotin-peroxidase complex for 30 min, developed with diaminobenzidine—nickel

41

chloride (DAB) substrate for 4-6 min, dehydrated and mounted on acid/alcohol washed
slides. Using a light microscope, acini were determined to be either + or - for PSA
expression as compared to controls that received no primary antibody. All experiments

were done in duplicate.

Analysis of individual basement membrane components: To the wells of 96-well
plates, 50 #1 of laminin (1 mg/ml) kept on ice, were added and the plates were placed in
a 37°C incubator for 1 h to allow the matrix to gel. Due to the liquid state of ﬁbronectin
and Collagen type IV a gelatin base was need in order to form a gel. Forty microliters
(40 [11) of gelatin were added to each well and the plates were placed in a 37°C incubator
for 15 min to allow the gelatin to set. Twenty microliters (20 ul) of Collagen type IV
(0.5 mg/ml) or ﬁbronectin (5 mg/ ml) diluted 1:1 in gelatin was then added to the gelatin
base and placed in a 37°C incubator for 15 min to allow the mixture to gel. RWPE-l cells
were plated at 10,000 cells/well in complete K—SFM with 6 wells representing each test
matrix component. Cells plated on plastic and on Matri gel served as the controls. Medium
was changed every 48 h. At 4 days, cultures were washed gently with PBS, ﬁxed in 2.5 %
buffered paraformaldehyde for 15 min, rinsed with PBS and either processed that day or
stored in PBS at 4°C. Cultures were permeabilized using a 1:1 solution of methanol and
acetone (3 min) and mounted in Vectashield mounting medium containing propidium
iodide (1.5 ug/ml) on alcohol washed slides. Acinar formation was examined by laser
scanning confocal microscopy using the criteria of four or more spherically arranged
polarized cells with lumens. Acinar counts were performed at low magniﬁcation using the

LSM. The presence of a lumen was veriﬁed by randomly visualizing counted acini using

42

confocal microscopy. The number of acini formed was expressed as percent of control

using RWPE-l cells grown on Matrigel as 100%. All experiments were done in duplicate.

Induction of acinar formation assay: The ability of TGF-11,, TGF-B2 and TGF-B,
to induce acinar formation in tumorigenic WPEl-NA22 cell line was investigated.
Non-tumorigenic RWPE-l cells were used as controls. All three TGF-B isoforms were
dissolved in 4 mM HCI containing 1 mg/ml BSA. The ﬁnal concentration of vehicle in the
medium was 0.1% which was not growth inhibitory. To the wells of 96-well plates, 50 pl
of growth factor reduced Matrigel (0.3 mg/50 ul), kept on ice, were added and the plates
were placed in a 37°C incubator for 1 h to allow the matrix to gel. Cells were plated at
10,000 cells/well in complete K-SFM with 6 wells representing each concentration. Forty-
eight hours after plating, medium was changed to complete K-SFM medium containing
TGF-Bl, TGF-112 or TGF-B3 at concentrations ranging from 0 to 10 ng/ ml. Medium was
changed every 48 h. At 4 days of treatment, cultures were washed gently with PBS, ﬁxed
in 2.5% buffered paraformaldehyde for 15 min, rinsed twice with PBS and then either
processed that day or stored in PBS at 4°C. Cultures were permeabilize using a 1:1
solution of methanol and acetone (3 min) and mounted in Vectashield mounting medium,
containing propidium iodide (1.5 ug/ml), on alcohol washed slides. Acinar counts were
performed at low magniﬁcation using the LSM. The presence of a lumen was veriﬁed by
randomly visualizing counted acini using confocal microscopy. The number of acini
formed was expressed as percent of control using non-treated control wells as 100%. All

experiments were done in duplicate.

43

Inhibition of acinarformatian assay: The ability neutralizing anti-bodies (Ab)
against TGF—Bl, TGF-112, TGF-B3, TGF-B RII, TGF—BRIII, a6 integrin subunit, Bl integrin
subunit or laminin, respectively. to inhibit acinar formation in non-tumorigenic RWPE-l
cells was investigated. To the wells of 96-well plates, 50 pl of growth factor—reduced
Matrigel (0.3 mg/50 ml) kept on ice, were added and the plates were placed in a 37°C
incubator for l h to allow the matrix to gel. Cells were passaged and resuspended in media
containing the appropriate neutralizing Ab concentration (Table 2-3) and incubated at room
temperature for 45 min. Cells were plated at 10,000 cells/well in triplicate. Forty-eight
hours after plating, cultures were washed gently with PBS, ﬁxed in 2.5% buffered
paraformaldehyde for 15 min, rinsed twice with PBS and then either processed that day
or stored in PBS at 4°C. For staining, cultures were permeabilize using a 1:1 solution of
methanol and acetone (3 min) and mounted with Vectashield mounting medium containing
propidium iodide (1.5 pg/ ml). Acinar counts were performed at low magniﬁcation using
the LSM. The presence of a lumen was veriﬁed by randomly visualizing counted acini
using confocal microscopy. The number of acini formed was expressed as percent of

control using non-treated control wells as 100%. All experiments were done in duplicate.

EGF effects on acinar morphogenesis: The effects of EGF on acinar formation in
non-tumorigenic RWPE-l cell line was investigated. EGF was dissolved in sterile distilled
water. EGF stocks were stored at -70°C until use. The ﬁnal concentration of vehicle in the
medium was 0.1% which was not growth inhibitory. To the wells of 96-well plates, 50 pl
of growth factor reduced Matrigel (0.3 mg/50 m1) kept on ice, were added and the plates

were placed in a 37°C incubator for 1 h to allow the matrix to gel. Cells were plated at

44

Table 3. Neutralizing antibody and conditions for Inhibition of acinar and
ductal formation assay.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Antibody Dilution Incubation time Temperature
Anti-c5 integrin ‘ 1:5 45 min. 21°C
Anti-B1 integrin 1’ 1:33 45 min. 21°C
Anti-Laminin ‘ Full strength, 1:10 and 1:20 45 min. 21°C
TGF-B1 ‘ 1:42 45 min. 21°C
TGF-[32 ‘ 1:52 45 min. 21°C
TGF-B3 ‘ 1:15 45 min. 21°C
TGF-B receptor ll (R11) 3 1: 15 45 min. 21°C
TGF-B receptor In (Rlll) 3 1: 15 45 min. 21°C
EGF ‘ 1:42 and 1:62 45 min. 21°C
199 1.2.3 Full stength‘. 1:33 and 12000 45 min. 21°C

1, Mouse
2, Rat
3, Goal

45

 

10,000 cells/well in complete K—SFM with 6 wells representing each concentration. Forty-
eight hours after plating, medium was changed to basal K-SFM supplemented with BPE
medium containing EGF at concentrations ranging from 0 to 10 ng/ml. Medium was
changed every 48 h. At 4 days of treatment, cultures were washed gently with PBS, ﬁxed
in 2.5% buffered paraformaldehyde for 15 min, rinsed twice with PBS and then either
processed that day or stored in PBS at 4°C to be used later for analysis. Cultures were
permeabilize using a 1:1 solution of methanol and acetone (3 min),and mounted with
Vectashield mounting medium containing propidium iodide (1.5 pg/ ml) on alcohol washed
slides. Acinar counts were performed at low magniﬁcation using the LSM. The presence
of a lumen was veriﬁed by randomly visualizing counted acini using confocal microscopy.
The number of acini formed was expressed as percent of control using non-treated control

wells as 100%. All experiments were done in duplicate.

Invasion assay

The ability RWPE-l, RWPE-2, WPEl-NA22, WPEl-NBll, WPEl-NB14,
WPE] -NB26 and WPE] -NB27 cells to invade through Matrigel, a reconstituted basement
membrane, was examined using a Boyden chamber in vitra invasion assay modiﬁed in our
laboratory (Bello, 1997). DU-l45 cells, which show high invasive ability in this assay,
were used as the positive control. The modiﬁed assay was conducted as follows: on the
day prior to running the invasion assay, the following cells were plated in T-75 ﬂasks in
10 ml of the appropriate medium. RWPE-l, RWPE-2, WPEl-NA22, WPEl-NBll,
WPEl-NB14, WPEl-NB26 and WPEl-NB27 cells were plated in complete K-SFM at

3X10°/ ﬂask and DU—145 cells in RPMI-1640 with 5 % DCS at 2X106 cells/ ﬂask. Matrigel

46

was diluted 1:20 with cold (4°C), sterile distilled water on ice for a ﬁnal concentration of
500 pg/ml. Each Nuclepore ﬁlter was coated with 25 pg Matrigel in 50p] and left to dry
overnight at room temperature under sterile conditions. Cells were rinsed with D-PBS
24 h after plating and incubated with 3 ml of 1 mM EDTA for 8-10 min, dislodged by
tapping and recovered by centrifugation. Pellets were resuspended to obtain 2 million cells
per ml. RWPE-l, RWPE-2, WPEl-NA22, WPEl—NBI 1, WPEl-NB14, WPEl-NB26 or
WPEl-NBZ7 cells were re- suspended in 3 ml basal K-SFM containing 0.1% BSA.
DU-145 cells were resuspended in serum-free RPMI containing 0.1% BSA. The bottom
chamber contained 220 pl of NIH/3T3 cell conditioned medium which served as the
chemoattractant. For preparing conditioned medium, subconﬂuent NIH/3T3 cultures in
100 mm plates were fed with 7 ml of serum-free DMEM containing 50 pg/ ml ascorbic
acid. Conditioned medium was collected 24 h later, centrifuged to remove cell debri and
stored at -20°C. A cell suspension containing 400,000 cells in 200 pl medium was added
to the top chamber and allowed to remain undisturbed for 5 min before overlaying with
650 p1 of K-SFM for RWPE—l, RWPE-2, WPEl-NA22, WPEl-NBll, WPEl-NB14,
WPEl-NB26 or WPEl-NB27 cells or serum-free RPMI medium for DU-145 cells,
containing 0.1% BSA. Controls consisted of K-SFM medium containing 0.1% BSA or
serum-free RPMI medium. Cells were allowed to invade for 24 h at 37°C at which time
the ﬁlters were prepared for absorbance reading. The migrated cells were ﬁxed, stained
with HEMA-3 and allowed to hydrate in distilled water. Nuclear stain was extracted by
placing ﬁlters individually in wells of a 24-well plate containing 300 pl of 0.1 N HCl for
15 min and 200 pl from each well were then placed in a 96 well plate and the absorbance

measured at 620 nm using a Titertek microplate reader. In addition to evaluating the

47

number of migrated cells on the ﬁlter, cells that had migrated to the bottom chamber were
also counted. Medium in the bottom chambers was triturated ﬁve times and 100 pl were
taken from each of the three chambers, pooled into cuvettes containing 10 ml Isoton,
triturated ten times and counted. The DU-145 cells that invaded and were attached to the
underside of the ﬁlter and those counted in the bottom well were each taken as 100%. The
two percentages were added to obtain total invasion. Percent absorbance and cell counts
for RWPE-l, RWPE-2, WPEl-NA22, WPEl-NBll, WPEl-NB14, WPEl-NB26 and
WPEl-NB27 cells were calculated using DU-145 as 100%. Three replicate invasion

chambers were prepared per treatment. All experiments were done in triplicate.

48

References

Bello D, Webber MM, Kleinman HK, Wartinger DD and Rhim J S: Androgen responsive
adult human prostatic epithelial cell lines immortalized by human papillomavirus

18. Carcinogenesis 18: 1212-1223, 1997.

Rhim JS, Webber MM, Bello D, Lee MS, Amstein P, Chen L and Jay G: Stepwise
immortalization and transformation of adult human prostate epithelial cells by a
combination of HPV18 and v-Ki-ras. Proc. Nat. Acad. Sci. USA, 91:11874-

11878, 1994.

Webber MM, Bello D, Kleinman HK, Wartinger DD, Williams DE and Rhim JS:
Prostate speciﬁc antigen and androgen receptor induction and characterization of

an immortalized adult human prostatic epithelial cell line. Carcinogenesis 17: 1641-

1646, 1996.

Webber MM, Bello D, Kleinman HK and Hoffman MP: Acinar differentiation by non-

malignant immortalized human prostatic epithelial cells and its loss by malignant

cells. Carcinogenesis 18:1225-1231, 1997.

49

Chapter Three

Extracellular Matrix and Integrins

Background

50

Background

In prostatic development, the terms acinar and ductal morphogenesis refer
to the process by which inducers alter the shape and outgrowth (branching) of epithelial
cells into the mesenchyme for the ultimate establishment of acini and associated ducts
(Hay, 1981). These subsequently undergo functional differentiation, by inducer inﬂuence,
to secrete prostate speciﬁc proteins, a major component of semen. The creation
(morphogenesis) of acini and ducts from undifferentiated epithelial cells involves
coordinated interactions which all work toward the proper establishment of epithelial cell
polarization and acinar formation. In order to discuss this morphogenic process, one must
have a basic understanding of epithelial polarization, the extracellular matrix (ECM),
integrins, growth factors and hormones because all of these are involved in normal
glandular morphogenesis. Additionally, this review will brieﬂy discuss prostatic epithelial

cell makers such as intermediate ﬁlament cytokeratin proteins and prostate speciﬁc antigen.

Epithelial Cell Polarity

Epithelial cells are organized into sheets, glands and tubules (ducts) that form the
interface between tissue compartments in higher organisms. They function to vectorially
transport ions, sugars and amino acids against a transmembrane concentration gradient.
Thus, by their functions of secretion and/or absorption, they are able to alter the tissue
compartments which they separate (Alberts et al., 1994). In the prostate, epithelial cells
produce secretions which are a major component of semen. Prostatic epithelial cells form
glandular compartments in which epithelial cells apically secrete their products into a

lumen for transport through a ductal network. In this manner, the epithelial cells physically

51

separate the muscular stroma from the glandular lumen and associated ducts. In order to
accomplish this apical secretion and physical separation, the epithelial cells must be
polarized.

The induction and maintenance of epithelial cells polarity is a multistage process
which depends, in part, on cell to cell and cell to ECM interactions (Hay, 1981). These
interactions are made possible by surface adhesion molecules located in the plasma
membranes. The plasma membrane is divided into three distinct regions: apical, lateral and
basal (Alberts et al., 1994). Each region has a distinct protein and lipid composition. The
apical region, which borders the glandular and tubule lumen, is characterized by microvilli
and cilia involved in absorption and /or secretion. The lateral surface is involved in cell
to cell adhesion and cell communication. It is characterized by tight junctions that separate
it from the apical surface and by adhesion molecules, such as integrins and cadherins, that
seal them to neighboring cells. The basal surface is typically anchored by integrins to the
basement membrane. Epithelial cells polarize in response to contact with neighboring cells

and the basement membrane, which is a speciﬁc part of the ECM.

Extracellular Matrix (ECM)

The extracelluar matrix (ECM) is that structurally stable non-cellular material which lies under
the epithelial cells and surrounds connective tissue cells (Hay, I981). It may be viewed as the

ground substance in the tissues of multicellular organisms. The ECM was once

52

thought of as a cellular product which served as a mere scaffold onto which cells attached via
cellular receptors. Today we know the ECM to be an active and vital player in cell migration,
proliferation and differentiation, in addition to its role in establishing and maintaining cell
shape (polarization) (Hay, 1981). The macromolecules that constitute the ECM are products
of cells which reside in it. These cells take an active role in the patterning of the ECM in that
their cytoskeleton structure inﬂuences the ECM’s orientation (Alberts et al., 1994). During
embryonic development the ECM
plays an important role in
organogenesis and epithelial cell
polarization and differentiation. The
ECM is composed of three major
classes of macromolecules secreted
by cells: collagens, proteoglycans
and glycoproteins (Alberts et al.,

1994).

 

In the prostate, the basement

Figure 3—1. Psardo—stratiﬁed prostatic epithelial cell layer.
membrane separates the epithelial Basal cells and luminal cells are illustrated attached to the
basement membrane of the extracellular matrix.

cells of the glands and ducts from the

underlying stroma (ECM proper) (Figure 3-1). The basement membrane (basal lamina) is a
structurally distinct zone of the ECM composed of laminin, collagen type IV,
nidogen/entactin and heparan sulfate binding proteoglycans (Figure 3-2, Table 3-1). Basement
membranes are relatively thin, ranging from 20-300 nm in thickness. They either surround

cells as in muscle, fat and nerve axons or they separate cells form the underlying stroma as

53

   
  

Key:
g...
entacunlnldogen

proteoglycans

laminin
tlbronectln

type IV collagen

Figure 3-2. A cum-em model of the molecular structure of the basement membrane. The basement
membrane is formed by speciﬁc interactims betwem the protiens type IV collagen, laminin, entaetin
and proteoglylcans (Based on PD. Yurchenco and J.C. Schittny, 1990).

Table 3—1. Properties of basement membrane components with potential for interaction with
laminin. (Tirnpl et al., 1987)

 

 

 

Ultneenuiuael
dun
Component :3... M, (x l0“) Chain constituents Don-in structure
Collagen type IV 4.4 S 550 Two «MW and one 4120‘!) Four hummer-Ion;
chimencheboulmkna triplehelixlermimtimin
dobuhr donnin
Laminin 11.5 S 950 BI. Bl (each about 230 kDe). Cmsheped (us x 15 um)
and A chain (~60 kDa) with and
mdlike domino
Heparin sulfate 7.0 8 ISO Single-core protein chain (4!) Compact. rodlike core (to
LDproteodycnn kDe)wlththreeHSchains nm)HSchaim(90nm)el
(36 km one pole
HD Wye-n 4.5 S I30 Smell-core protein (S-IZ _ Sterile ttrncmre with ”-nm
km) with (om (3—5) HS HS chins
chains (29 kDI)
Nidogen 6J8 ISO Singleclhtn DInhbclH~7Jnnlnqurth
' Abbreviations: LD, low-density form; HD. high-density form; HS. Wm slime.

 

 

 

54

in epithelium and endothelium (Timpl and Aumailley, 1989). This section will brieﬂy
review ECM components: collagen type IV, laminin/entactin(nidogen), heparan sulfate
proteoglycans and ﬁbronectin.

Collagen Type IV: Collagen IV is the major structural component of the basement
membrane and functions to anchor other ECM component, most notably, the laminin2cell
complex (Timple and Aumailley, I989). Collagen IV may comprise between 30 to 80% of
the total basement membrane mass in different basement membranes. Individual collagen
IV chains (1,700 amino acids) assemble to form a 400 nm long triple helix. Although
collagen IV forms the scaffold onto which laminin and heparin sulfate proteoglycans bind
(Figure 2), it can directly bind to cells via speciﬁc collagen IV integrin receptors.

Laminin: Laminins are a family of large trimeric glycoproteins comprised of three
disulﬁde bonded chains. They are found, exclusively, in basement membranes (Malinda and
Kleinman, 1996). Developmentally, laminin is one of the ﬁrst synthesized ECM proteins.
In fact, the basement membrane of developing embryos consist mainly of laminin with little
or no collagen type IV. In 1979, laminin was ﬁrst isolated from a tumor and believed to be
a third chain of type four collagen. Subsequent, antibody localization and chemical analysis
found it to be present in the basement membrane of most tissues (Alberts et al, 1984,
Malinda and Kleinman, 1996).

Laminin is a large (”850 kDa) glycoprotein which exist as a cruciform-like structure
formed by three very long polypeptide chains (or, B, and y) which are held together by
disulﬁde bonds (Figure 3-3) (Alberts et al, 1984, Malinda and Kleinman, 1996). The or, B,
and y chains are structurally homologous but lack homology in their amino acid sequence.

There are various isoforms of each chain (ie al, a2, a3), which

55

 

 

 

assemble to form the different laminin
a chain
(‘00 kc)
isoform. At the amino terminus, the a
mo
chain has three globular domains m" /
chnln PM! s---Q--. ychain
separate by EGF-like repeats. At the piano k0) / (zoom)
. . . . Poona
carboxy terminus, it has a corled-corl
domain and a large globular domain.
m _
The B and y chains, being smaller than ' A"\t
. 3 k I. .
the or chain, possess only two globular o m“ I:
domains and two EGF repeats at the L I: .
tovoL'eTn

 

 

 

am'no te m‘n s and lack a lar e
l r 1 u g Figure 3-3. Schematic model of Laminin-1 (Malinda

globular domain at the carboxy and Klemman’1996)'

terminus. The three chains assemble at the carboxy end. This assembly is mediated by the
coiled-coil domain of the or chain, which is critical for chain-speciﬁc assembly of the
laminin molecule. The assembled cruciform laminin molecule possesses a number of
functional domains which bind type IV collagen, heparan sulfate entactin/nidogen and cell
surface receptors such as integrins.

Interestingly, entactin/nidogen, a small (150 kDa) sulfated polypeptide molecule,
binds through one of its globular domains to the center of the laminin cross (Figure 3-4)
and usually co-precipitates with laminin. Although its exact biological function is not
known, entactin/nidogen is thought to modulate the activity of a cell binding site in
laminin 1 fragment. Entactin/nidogen is able to independently bind to collagen IV, thus,
contributing to the intricate linkage pattern necessary for the stabilization of the basement

membrane (Malinda and Kleinman, 1996, Martin and Timpl, 1987).

56

 

As the key component of the basement
membrane, laminins are important both structurally

and biologically (Malinda and Kleinman, 1996).

 

Structurally, laminin is thought to initiate the

formation of the basement membrane because of its

 

 

 

 

M
. . . . m0 FF,“
abllity to bind collagen IV, entactrn and proteoglycans b M,

 

and as well as other laminin molecules. This claim is Figure 3.4_ A model of laminin-
entactin/nidogen complex

supported by the observation that it is the ﬁrst showing protease fragment
indentiﬁed by arabic numbers

extracellular macromolecule to be expressed by the and Phe lacation 0f some
biological acmvres (Timpl,

developing embryo. Biologically, laminins are 1989)'

involved in promoting cell growth, migration, epithelial cell polarity and differentiation,
nerve regeneration, wound repair and tumor growth and metastasis. The ﬁrst biological
activity elucidated for laminin was its binding to both normal and neoplastic epithelial
cells. Currently, speciﬁc adhesion-promoting peptide sequences in laminin have been
identiﬁed. laminins affect cellular function by binding to the cells adhesion receptors such
as integrins, thereby inducing both mechanical and chemical transduction pathways, which
alter gene expression and cell function.

Heparan sulfate proteoglycans: Heparan sulfate proteoglycans are divided into
high and low density molecules. The high density molecules (600-700 kDa) are small and
star-like in shape and have a low protein content. The low density molecules (130 kDa)
have a multi-domain core which is associated with heparan sulfate side chains attached at
one end (Figure 3—2). Heparan sulfate proteoglycans are considered to function as ﬁller

molecules and serve as selectively charged barriers which make the basement membrane

57

impermeable to some proteins. Although both types of proteoglycans can bind laminin
directly, high-density proteoglycans bind it weakly, while low density proteoglycans bind
it strongly.

Fibronectin: Although ﬁbronectin is not usually found in the basement membrane,
it is, nevertheless, an important ECM component (Figure 3-2). Fibronectin is a large
glycoprotein (210-260 kDa) which is found ubiquitously throughout the body tissues and
ﬂuids. Fibronectin functions as a bridge between cells and other ECM molecules, as an
attachment factor for cells or as a structural component of ECM. There are two major
forms of ﬁbronectin: plasma ﬁbronectin and cellular ﬁbronectin. Cellular ﬁbronectin has
large polypeptide chains which exist as monomers, disulﬁde-bonded dimers and multi-
meres (Akiyama and Yamada, 1987). The disulﬁde bonds are unique in that they are
usually conﬁned to the carboxy terminus. The odd placement of the disulﬁde bonds may
be to facilitate the ability of the molecule to stretch between distant points while ‘mediating
adhesive interactions’ (Yamada, 1983). These polypeptide chains are composed of discrete
globular functional domains. Fibronectin possesses a collagen binding domain, a heparan
binding domain, a hyaluronic acid-binding domain, an actin and DNA binding domain and
a plasma membrane binding domain (Akiyama and Yamada, 1983, Yamada, 1984). The
plasma membrane-binding domain contains an RGD (arginiine, glycine, aspartic acid)
sequence which is recognized by ﬁbronectin integrin receptors of cells (Garrat and
Humpheries, 1995, Sonnenberg 1993).

One aim of the present study is to examine the expression of laminin integrins in
human prostatic epithelial cells and their role in human prostatic epithelial polarization and

acinar morphogenesis. A second aim is to examine the changes in integrin expression in

58

human prostate cancer cells and the effects these changes may have on the ability of cancer

cells to polarize. Therefore, a brief review of integrin structure and function follows.

Integrins

Integrins are a family of cell adhesion receptors, located in the cell membrane, and
are involved in the coupling of cells to the ECM and neighboring cells. Integrins modulate
cell adhesion by mechanical transduction and function as signal transduction receptors.
Thus, integrins have been shown to play a critical role in embryogenesis, organogenesis,
wound repair, cytoskeletal organization, motility, cell growth and differentiation (Geiger
et a1. , 1995, Vinatier, 1995). In addition to their role in establishing cell polarity, integrins
are involved in the differentiation of epithelial cells into specialized secretory cells
(Sheppard, 1996). Integrins connect cells to their ligands which include laminin, collagens

and ﬁbronectin.

Structure

Integrins are heterodimeric transmembrane glycoproteins which consist of an a
(120 - 180 kDa) and a B (90-110 kDa) subunit (Figure 3-5). Each subunit has three
domains; a large extracelluar domain which binds to the ligand, a transmembrane domain

and a short cytoplasmic domain which is associated with the cytoskeleton (Sonnesberg,

59

 

1993, (Vinatier, 1995). At present

Connection
m cmktw" lntracytoplaamic
there are 22 known integrin receptors 11 portion
Cellular
mm..- l“ “v m" w «Tranamembrane
which are the products of different mm. portion

combinations of 16 different a-
subunits with 8 different B-subunits

(Figure 3—6). The B subunits are

 

promiscuous and can complex with

Ligand binding region

   
  

Extracellular
region

submit

 

 

Figure 3-5. A schematic structure of integrins

multiple (1 subunits, while the a

(Vinatier, 1995).

subunits are generally loyal to their [3 subunit partner with the exception of the av, (14 and

a6 subunits (Garret and Humphries, 1995,

Vinatier, 1995).

Classiﬁcation

Integrins are generally classiﬁed by their
mode of adhesion (V inatier, I995). The three
binding modes are: cell to ECM adhesion, cell to
cell adhesion, and platelet to ECM to platelet
adhesion. Integrins are best known for their role in
cell to ECM adhesion. Cell to ECM-binding
integrins may be loosely divided into two groups.
Group one consists of receptors for collagen and

laminin and group two consist of receptors for

 

 

 

 

 

 

 

Figure 3-6. The integrin family
(Sonnesberg, 1993).

ﬁbronectin and vitronectin. Generally, epithelial cells express integrins

60

from the collagen/laminin group,

 

 
    

while mesenchymal and endothelial Am“ ““0“th

cells express integrins from the MMCELLUW
ﬁbronectin/vitronectin group
(Vinatier, 1995) (Figure 3-7).

B Subunit
Fibronectin, laminin 0f mm

Ligand Speciﬁcity and Affinity uSubunit of integrin

 

 

 

Ligand speciﬁcity: The 'Figure 3-7. Integrin receptor binding to basement
membrane. (Gilbert, 1994).

different cr/B subunit combinations
specify the ligand binding ability of integrins. Although integrins are generally considered to
be promiscuous receptors because they may bind more than one type of ligand, there are
seven integrins (aSBl, avBl, orvB6, avBS, a6B1, a5 B4 and a7Bl) that demonstrate single
ligand binding speciﬁcity (Table 3—2) (Sonnenberg, 1993). Ligand speciﬁcity is also dependant
on the cell type, due to differences in post-translational modiﬁcation of individual integrin
subunits (Sonnenberg, 1993). An example of cell-type speciﬁcity is seen with the a2B1
integrin. When aZBl is expressed by platelet cells, it ﬁrnctions as a receptor for collagen, but
when it is expressed on endothelial cells, it acts as a receptor for collagen and laminin

(Sonnenberg, 1993; Vinatier, 1995).

61

Table 3-2. Integrins and their ECM binding sites (Sonnenberg, 1993).

 

[3. integrins
“int.
“an:

'35:

3401

“a”!

3e”:

Cznt

Cg".

evil:

I), integrins
at”:

Bin/’2

aaﬂl

ll, integrins
“are”:

'4‘:

[l7 integrins

am”?
“a”:

Gallo
«.55
av”.

av":

 

Receptor Alternative name

VLAol

VLA-Z. GPlaclla. ECMRII

VLA-3. VCA~2. ECMRI. Gaph 3
(hamster)

VLA-d, LPAM-Z (mouse)

VLA-s. FNR, GPlc-lla. ECMRVI

VLA-G. GPlc-lla

VLA-7

LFA-l
Mac-l. (ZR-3
p150. 95

GPllb-llla
VNR

M290 tat. (mouse)
LPAM-l (mouSe). am,

Other I! subunit containing integrins

GPIc-ICBP, TSP-180. A9 and EA-i
ﬂvﬂv Gvﬂao

Ligands

Co" (I. IV). Ln
Coll (l-V.Vl), Ln. Fn?
Ep. Ln, Nd/En. Fn. Collll)

Fn.alt, VCAM-l/lNCAMHO
Fn

Ln \

Ln ‘

7

Fn

ICAMJ, lCAM-Z. lCAM-3
C3bi. factor X. H). lCAM-l
Fb. C3bi?

Fb. Fn. vWF. Vn
Fb. Fn. vWF, Vn, Tsp. Use and Bspl

7
Fn.all, VCAM-l

Ln
Vn
Fn
?

 

Ligand afﬁnity: Cells can regulate the binding afﬁnity of integrins to their ligand
by either up or down regulating its expression or by direct modulation of the integrin
receptor (V inatier, 1995). Growth factors have been shown to modulate the synthesis and
expression of integrin subunits. The ﬁrst growth factor to be identiﬁed as a regulator of
integrin expression was transforming growth factor-B1 (T GF—Bl) (Garratt and Humphries,
1995). TGF-B, an important negative regulator of epithelial growth, is expressed during
embryogenesis. TGF-B1 regulates integrin expression at the level of transcription and by
affecting the rate of maturation of the B subunits (Garratt and Humphries, 1995).
Speciﬁcally, TGF-Bl up regulates the expression of al to a6, av, B1, B3, B5 and B6

integrin subunits. Other growth factors known to regulate integrin expression are nerve

62

 

growth factor, interferon, interleukine-IB and tumor necrosis factor-a (Vinatier, 1995).

Binding afﬁnity is also regulated by intracellular events via modiﬁcation of
integrins cytoplasmic domains as well as by extracellular factors (Garratt and Humphries,
1995). Modiﬁcation of the intracellular, cytoplasmic, domain alters the ability of the
extracellular domain to bind to its ligand. This is seen with the expression of a6Bl in
macrophages. a6B1 is a laminin receptor which can not bind laminin unless it is ﬁrst
phosphorylated. Phosphorylation occurs as a consequence of macrophage activation
(Garratt and Humphries, 1995; Vinatier, 1995). Extracelluarly, integrin afﬁnity is
regulated by lipids or by the type of cations present at the ligand binding site. For
example, the collagen binding aZBl integrin will bind collagen in the presence of

magnesium but not in the presence of calcium (Garratt and Humphries, 1995; Vinatier,

1995).

Ligand and Receptor Active Sites

Ligand active site: Integrins bind at speciﬁc peptide sequences within the ligand
called the ligand-receptor binding site. In 1984, the ﬁrst ligand-receptor binding site was
described within the central cell-binding domain of ﬁbronectin as an RGD (Arg-Gly—Asp)
tripeptide sequence (Garratt and Humphries, 1995 , Sonnenberg, 1993). Although the RGD
motif has been shown to be present in most matrix proteins, in some instances, it is
masked, thus, becoming unavailable for binding. This is observed in the case of laminin
and the avB3 integrin. The RGD motif in the laminin is not recognized by avB3 integrin
because it is hidden within the molecule. Upon proteolytic digestion of laminin, the RGD

sequence is exposed and binding by avB3 integrin occurs (Sonnenberg, 1993). This could

63

be one mechanism by which integrins modulate cell motility.

Because all matrix proteins do not contain RGD motifs, and even if they do, it may

be masked, this suggests that other receptor binding sites must also exist. At present not

all of the receptor binding sites have been discovered but all that have, seem to share one

common characteristic. They all have short peptide motifs which contain an acidic amino

acid such as aspartic acid and glutamine. Currently, the laminin receptor binding sites that

have been identiﬁed are fragment pl (cl-4) for alB 1 , fragment E8 for a6B1 and fragment

E1 for a3B1. Table 3-2 shows all the receptor binding sites that are presently known.

Receptor active site: The
site on the integrin receptor that
interacts with the ligand~receptor
site is composed of three distinct
ligand binding site, as shown in
figure 3-8 (Garratt and
Humphries, 1995). The ﬁrst two
sites are locates in the or subunit.

The ﬁrst site, called the aA

domain, is a 200 amino acid

polypeptide segment. The second

 

 

 

 

'U m l
oA-DOMAIN ,
[I H B]
/////
a lllllllllll—- l
I
‘EZTION' -e"'T‘GltN m m/—‘
sires
m
t! Izzyx-IIIIIAI
B) ‘
(198.209)
BA-DOMAIN

 

 

 

 

Figure 3-8. Active sites for integrins (Garratt and
Humphries, 1995).

site is unnamed and is thought to play a key role in maintaining structural integrity. The

third ligand-receptor site is located on the B subunit and contains more than one interacting

peptide sequence (Garratt and Humphries, 1995).

Interestingly, three cation binding sites co-localize with the three active binding
sites (Figure 3-8). The binding of divalent cations to integrins is required for receptor
activation. Generally, the binding of Mg2+ or Mn2+ to the cation binding sites, along with
ligand binding, is stimulatory, while Ca2+ binding in inhibitory. Physiologically, the role
of cation binding is to mediate ligand binding and regulate the activation state of the

integrins (Garratt and Humphries, 1995).

Intracellular Interaction

The intracellular domain of the integrin receptor plays an important role in
connecting the plasma membrane to the cytoskeleton either at adhesion junctions called
focal contacts, where they are indirectly linked to actin ﬁlaments, or at hemidesmosomes
which indirectly link them to cytokeratin intermediate ﬁlaments (Figure 3—7). Focal
contacts are cellzmatrix adhesion junctions that contain adhesion-associated proteins such
as talin, a-actinin, vinculin, tensin, paxillin, zyxin and cysteine-rich proteins molecules.
(Geiger et al., 1996, Sonnenberg, 1993). Prior to activation, the integrin a subunit
prevents the integrin receptor from localizing to focal contacts in the plasma membrane.
It is upon activation of an integrin receptor that receptors are recruited into a focal contact.
Here, the B subunits bind to the cytoskeletal components, while the a subunits are
primarily associated with the ligand (Geiger et al., 1996, Sonnenberg, 1993). Within the
focal contacts, the cytoplasmic portion of the B integrin subunit is connected to actin
ﬁlaments via talin or a-actinin. Talin and a-actinin are, in turn, connected to the actin
ﬁlaments of the cytoskeleton. Additionally, a-actinin along with a molecule, called
ﬁmbrin, connect neighboring actin ﬁlaments to one another. Tensin also binds different

65

actin ﬁlaments to each other and binds to vinculin. Zyxin and cysteine-rich proteins are
associated with the a-actinin molecules. Vinculin binds talin and possibly a-actinin, while
at the same time interacting with actin ﬁlaments. Finally, paxillin binds to vinculin, which
binds to talin that binds to both the B subunit of integrin and the actin ﬁlament of the
cytoskeleton. Paxillin is also associated with focal adhesion kinase (FAK), that is involved
in signal transduction (Geiger et al., 1996). All these connections play an important role
in how integrins exert their mechanical effect on the cytoskeleton which, in turn, alter

cellular function.

Intracellular signaling

Intracellular signaling by integrins may be accomplished by either mechanical
transduction or by signal transduction through secondary messenger pathways. It is
believed that the biochemical effects, observed as a result of integrin binding, are
predominantly controlled through signal transduction rather than a mechanical transduction
(Rosales et al., 1995).

Mechanical transduction: Integrins provide mechanical support by linking the
cytoskeleton to the ECM via the plasma membrane (Garratt and Humphries, 1995). Thus,
the ability of cells to sense position and shape is determined by the number and distribution
of its integrins and integrin-mediated adhesion. Integrins, by virtue of physically linking
the ECM to the cytoskeleton via focal contact proteins, create a cytoskeletal organization
and mechanical tension that directly inﬂuences events in the cytoplasm and the nucleus
(Garratt and Humphries, 1995, Rosales et a1. , 1995). In non-motile cells, such as epithelial
cells, the mechanical stress state is important for the maintenance of polarity and cell shape

66

and for cellular movement in cases of wound healing and inﬁamation (Garratt and
Humphries, 1995). Additionally, mechanical transduction works synergistically with
integrin-induced signal transduction events which modulate differentiation and speciﬁc
gene expression.

Signal transduction by integrins: In recent years, it has become clear that integrins
not only function to mediate cellular adhesion to the ECM, but also to trigger biological
responses in a manner similar to those transduced by growth factors and hormones
(Ruoslahti and Reed, 1994, Streuli et. al., 1995). There is a strong possibility that
integrin-initiated signal pathways may impinge pathways involving other type of receptor
systems. Although not much is known about the signal transduction pathways associated
with integrins some key factors have been identiﬁed. A 125 kDa FAK has been shown to
phosphorylate tyrosine and autophosphorylate in response to integrin activation and
clustering in focal contacts (Geiger et al., 1995, Rosales et al., 1995). The FAK kinase
has been associated with PI-3 kinase and GRB—2 which are involved in the tyrosine
pathway and G-protein/inositol-lipid pathway (Rosales et a1. , 1995). Thus, suggesting that
integrin FAK activation include proteins and cascades used in the other pathways such as
the receptor tyrosine and the G-protein pathway. Unfortunately, the mechanism by which

FAK works is unclear at present.

Integrin-Mediated Gene Expression

Integrin—mediated gene expression interacts with and is, thereby, regulated by

signals from other receptor system pathways such as hormone and growth factors. For

67

example, in monocytes, B1 integrin activation induces cytokine and metalloproteinase gene
expression, which are dependant on tyrosine phosphorylation (see Lin et a1. 1994 in
Garratt and Humphries, 1995, Yurochko et a1. 1992). In epithelial cells, control of gene
expression is best illustrated with the expression of B—casein in mammary epithelial cells
(Roskelley et a1. , 1994). B-casein is an abundant milk protein the expression of which is
regulated by prolactin and basement membrane via integrin binding. Basement membrane-
dependant expression of B-casein requires that mammary epithelial cells have functional
B1 integrin. B-casein gene expression is regulated transcriptionally by an ECM response
element located upstream of the B—casein gene (Roskelley et al., 1994).

Roskelley et a1. (1994) set out to determine if the B-casein expression, elicited by
integrin binding to the laminin rich ECM, was solely based on physical clustering or on
biochemical signals. Mammary cells were grown on non-adhesive substratum polyHEMA
to cause cell clustering. The cell clusters were then treated with prolactin and assayed for
B-casein expression but none was found. Thus, clustering was not sufﬁcient to induce
B-casein expression even in the presence of prolactin. When cells plated on the polyHEMA
coated plates were overlaid with ECM and treated with prolactin, B-casein expression was
seen within 8 hours. They also showed that within 30 minutes of overlay with laminin,
tyrosine phosphorylation increased. Therefore, it was concluded that Bl integrin regulates
B-casein expression by a biochemical component involving tyrosine phosphorylation
associated with integrin activation. Because the prolactin signal transduction pathway is
not clear, at present, it is not known how it interacts with the integrin stimulated signal

transduction pathway (Roskelley et al., 1994).

68

In summary, it can be stated that integrins have been shown to play a critical role
in embryogenesis, organogenesis, wound repair, cytoskeletal organization, motility, cell
growth and differentiation by their ability to trigger signal transduction as well as by
mechanical transduction (Geiger et al., 1995 , Vinatier, 1995).

In order to establish the validity of an in vitro human prostatic epithelial cell model,
it is necessary to demonstrate that the cells express epithelial cell marker protiens as well
as prostate-speciﬁc marker protien. We have previously demonstrated that cell lines used
in this study express the epithelial cell marker cytokeratins, as well as prostate epithelial
markers androgen receptor and prostate speciﬁc antigen (PSA) (Bello et al. , 1997). A brief

description of these markers follows.

Prostate Epithelial Markers

Prostate Specific Antigen

Prostatic secretory epithelial cells synthesize and secrete prostatic speciﬁc antigen
in vivo (Fong et a1. , 1991). Prostate speciﬁc antigen is a major component of the semen.
Its function is to liquefy the semen after ejaculation (Lija et al., 1987, Webber et al.,
1995). It is a serine protease with molecular weight of about 33,000 Da. It is composed
of a single 240 amino acid glycoprotein chain. The prostate speciﬁc antigen gene is
located on chromosome 19 ( Lundwall and Lima, 1987, Webber et al., 1995).

The effects of androgen on prostate speciﬁc antigen production are clearly
demonstrated by the prostatic cancer cell line LNCaP (Young et a1. , 1991, Webber et al,
1996). In LNCaP cells, prostate speciﬁc antigen is constitutively secreted, even in the

absence of an androgen stimulus, due to a mutation in its androgen receptor. Addition of

69

androgens into the medium causes an increase of prostate speciﬁc antigen expression (Fong
et al., 1992, Young et al., 1991). Additionally, growth factors, such as basic ﬁbroblast
growth factor and epidermal growth factor are known to slightly decrease the secretion of
prostate speciﬁc antigen. Epidermal growth factor down regulates prostate speciﬁc antigen
mRNA and interferes with androgen regulation of the gene (Henttu and Vinkop, 1993).
Clinically, serum prostate speciﬁc antigen levels, which are elevated in association with
prostate cancer, are used for the diagnosis of prostate cancer and for early detection (Garde
et al., 1993, Webber et al., 1995). Recently, prostate speciﬁc antigen has been shown to
be involved in invasion and metastasis of prostate cancer due to its ability to degrade
components of the extracellular matrix (Webber et al. , 1995).

As stated previously, it is the luminal secretory cells of the mature prostatic
epithelium that synthesize and secrete prostate speciﬁc antigen in response to androgens
(Fong et al., 1991, Webber et al., 1995, Webber et al., 1996). Since the expression of
prostate speciﬁc antigen is a differentiated function of normal prostate epithelial cells, it

is used as a marker for prostatic differentiation.

70

References

Alberts B, Bray D, Lewis J, Raff M, Roberts K and Wastson J D (eds.): Cell junctions,
cell adesion and the extracellular matrix. In: Molecular biology of the cells, third

edition. New York, Garland Plublishing Inc., pp. 949-1009, 1994.

Akiyama SK and Yamada KM: Fibronectin. Adv. Enzymol. 59:1-57, 1987.

Bello D, Webber MM, Kleinman HK, Wartinger DD and Rhim J S: Androgen responsive
adult human prostatic epithelial cell lines immortalized by human papillomavirus

l8. Carcinogenesis 18: 1212-1223, 1997.

Fong CJ, Sherwood ER, Braun EJ, Berg LA, Lee C and Kozlowski JM: Regulation of
prostatic carcinoma cell proliferation and secretory activity by extracellular matrix

and Stromal secretions. Prostate, 21: 121-131, 1992.

Garde SV, Sheth AR, Porter AT and Pienta K]: A comparative study on expression of
prostatic inhibin peptide, prostate acid phosphatase and prostate speciﬁc antigen in
androgen independent human and rat prostate carcinoma cell lines. Cancer Lett.,

70: 159-166, 1993.

Garratt AN and Humphries MJ: Recent insights into ligan binding, activation and

signalling by integrin adhesion receptors. Acta Anat. 154, 34-45, 1995.

71

Geiger B, Yehunda—Levenberg S and Bershadsky AD: Molecular interactions in the
submembrane plaque of cell-cell and cell-matrix adhesions. Acta Anat. 154, 46-62,

1995.

Gilbert SF: Developmental biology, fourth edition. Massachusetts, Sinauer Associates,

Inc. Publishers, pg. 111, 1994.

Hay E: Collagen and embryonic development. In: Cell biology of extracellular matrix.

Hay E (ed.), New York, Plenum Press, pp.379-405, 1981.

Henttu P and Vinkop P: Growth factor regulation of gene expression in the human

prostatic carcinoma cell line LNCaP. Cancer Res., 53: 1051-1058, 1993.

Lija, H. Oldbring, J., Rannevik, G. and Laurell, C. Seminal vesicle-secreted proteins.

J. Clin. Invest., 80: 281-285, 1987.

Lin TH, Yurochko A, Kornberg L, Morris J, Walker JJ, Haskill S and Juliano RL: The
role of protein tyrosine phosphorylation in integrin-mediated gene induction in

monocytes. J. Cell. Biol. 126, 1585-1593, 1994.

Lundwall, A. and Lima, H. Molecular cloning of the human prostate speciﬁc antigen

CDNA. FEBS Lett., 214: 317-322, 1987.

72

Malinda KM and Kleinman HK: Molecules in focus: the laminins. Int. J. Biochem. Cell

Biol. 28:957-959, 1996.

Martin GR: Laminin and other basement membrane components. Ann. Rev. Cell Biol.

3:57—85, 1987.

Rosales C, O’Brien V, Kornberg L and Juliano R: Signal transduction by cell adhesion

receptors. Biochim. Biophys. Acta 1242, 77-98, 1995.

Roskelley CD, Desprez PY and Bissel MJ : Extracellular matrix-dependent tissue speciﬁc
gene expression on mammary epithelial cells requires both physical and
biochemical signal transduciton. Proc. Natl. Acad. Sci. USA 91, 12378-12382,

1994.

Ruoslahti E and Reed J C: Anchorage dependence, integrins, and apoptosis. Cell 77:477-

478, 1994.

Sheppard D: Epithelial integrins. Bioessays 18:655-660, 1996.

Sonnenberg,A. (1993) Integrins and their ligands. Curr. Top. Microbiol. Immunol. 184,

7-35.

73

Streuli CH, Schmidhauser C, Bailey N, Yurchenco P, Skubitz AP, Roskelley C and
Bissell MJ: Laminin mediates tissue-speciﬁc gene expression in mammary

epithelia. J. Cell Biol. 129: 591-603, 1995.

Timpl R: Structure and bilogical activity of basement membrane protiens. Eur.J .Biochem.

180, 487-502, 1989.

Timpl R and Aumailley M: Biochemistry of basement mambranes. Adv. Nephrol. 18:59-

76, 1989.

Vinatier D: Integrins and reproduction. Euro. J. Obs. Gyn. Reprod. Bio. 59, 71-81, 1995.

Webber MM, Waghray A and Belle D: Prostate-speciﬁc antigen, a serine protease,

facilities human prostate cancer cell invasion. Clin. Cancer Res., 1: 1089-1094,

1995.

Webber, MM, Bello D, and Quader S: Immortalized and tumorigenic adult human

prostatic epithelial cell lines. Characteristics and applications. Part. 1. Cell markers

and immortalized, non-tumorigenic cell lines. Prostate 29:386- 394, 1996.

Yamada KM: Cell surface interactions with extracellular materials. Ann. Rev. Biochem.

522761-99, 1983.

74

Young CY, Montgomery BT, Andrews PE, Qui SD, Bilhartz DL and Tindall DJ:
Hormonal regulation of prostate-speciﬁc antigen messenger RNA in human

prostatic adenocarcinoma cell line LNCaP. Cancer. Res., 51: 3748-3752, 1991.

Yurochko KM, Liu DY, Eierman D and Hakill S: Integrins as a primary signal

transduction molecule regulating monocyte immediate-early gene induction. Proc.

Natl. Acad. Sci. USA 89, 9034-9038, 1992.

Yurchenco PD and Schittny J C: Molecular architecture of basement membranes. FASEB J.

4: 1577-90, 1990.

75

Chapter Four

Acinar Differentiation by Non-malignant
Immortalized Human Prostatic Epithelial
Cells and Its Loss by Malignant Cells

Work presented in Chapter 3 has been published:
Webber MM, Bello D, Kleinman HK and Hoffman MP: Acinar differentiation by
non-malignant immortalized human prostatic epithelial cells and its loss by
malignant cells. Carcinogenesis 18:1225-1231, 1997.
Other related work:
Bello D, Webber MM, Kleinman HK, Wartinger DD and Rhim JS: Androgen
responsive adult human prostatic epithelial cell lines immortalized by

human papillomavirus 18. Carcinogenesis 18: 1212-1223, 1997.

76

Abstract

Invasive prostatic carcinomas and prostatic intraepithelial neoplasia (PIN) are
characterized by a loss of normal cell organization, cell polarity, and cellzcell and
cell:basement membrane adhesion. The objective of this study was to establish in vitro three-
dimensional (3 -D) cell models which can be used to investigate mechanisms involved in acinar
morphogenesis and differentiation in normal prostatic epithelium and their abnormalities in
cancer cells. The process of acinar morphogenesis, including structural and functional
differentiation, was investigated by culture on basement membrane gels (Matrigel). The
human papillomavirus 18 immortalized, non-tumorigenic cell line RWPE-l, the v-Ki-ras
transformed, tumorigenic RWPE-Z cell line derived from RWPE-l cells and the human
prostatic carcinoma cell line DU-145 were used. When cultured on Matrigel, RWPE-Z cells
remain as single cells or form small aggregates and DU-145 cells form large amorphous cell
aggregates without any organization or lumen. In contrast, RWPE-l cells form acini of
polarized epithelium with a distinct lumen, show a distinct laminin basement membrane, and
express a6B1 integrins at their basal end. Exposure to conditioned medium from NIH 3T3
cultures accelerates glandular morphogenesis. Parallel cultures maintained as monolayers on
plastic remain as monolayers. In the presence of the synthetic androgen mibolerone, acinar
cells express prostate speciﬁc antigen (PSA) as determined by immunostaining. We conclude
that normal prostate cells can undergo acinar morphogenesis while tumorigenic cells have
lost this ability. The 3-D cultures provide physiologically relevant in vitro models for
elucidating regulation of growth, morphogenesis and differentiation in the normal human
prostate, for deﬁning heterotypic interactions in benign prostatic hyperplasia and for

establishing the basis for the loss of normal cell organization in early neoplastic lesions such

77

as PIN as well as during tumor progression in prostate cancer.

Introduction

Loss of normal cell organization, cell polarity, and cellzcell and cell:basement
membrane (BM) adhesion has been observed in many carcinomas including invasive prostatic
carcinomas and advanced prostatic intraepithelial neoplasia (PIN) (1,2). These abnormalities
are associated with a loss, decrease or altered expression of the basement membrane protein
laminin, cell adhesion molecules such as cadherins and integrin receptors for extracellular
matrix (ECM) proteins. Normal expression of these proteins is critical for normal
morphogenesis. The aim of this study was to establish three-dimensional (3-D) in vitro cell
models, derived from non-tumorigenic and tumorigenic human prostatic epithelial cells, which
could be used to investigate mechanisms involved in acinar morphogenesis and differentiation
in both normal and malignant prostatic epithelium.

The ultimate goal of an in vitro model, using normal prostatic epithelium, is to mimic
the in viva cell architecture and retain differentiated function. Monolayer cultures lack the 3-
dimensional cell organization of tissues in vivo and a rapid loss of cell ﬁrnction is often
observed. Three-dimensional cultures simulate the in vivo architecture more realistically than
monolayer cultures. Such a model for the human prostate has not previously been available.
A necessary prerequisite for such a model is to develop cell lines that retain morphology and
function similar to primary cells, and remain responsive to hormones and to other growth
regulators. We have developed and characterized several immortalized, non-tumorigenic and
tumorigenic human prostatic epithelial cell lines which express ﬁmctional characteristics of

78

normal cells (3-7). These cell lines now make it possible to study morphogenesis and
differentiation in both normal and malignant prostate cells using standardized and
reproducible cell types. It is recognized that immortalization may alter some cell
characteristics, however, we have demonstrated that the RWPE-l cell line, used in the present
study, has retained the Y chromosome and normal cytokeratin expression and the cells mimic
normal epithelium in their response to growth factors as well as androgens by the expression
of PSA and up-regulation of androgen receptor (3).

Although much work has been done on factors that stimulate the growth of prostatic
epithelial cells, little is known about ECM-mediated regulation of morphogenesis and
differentiation. ECM is necessary for epithelial cell proliferative response to growth factors
as well as differentiation which are both necessary for maintaining homeostasis (8). In the
normal adult prostate in homeostasis, there is little net growth because a balance between cell
loss and cell gain is maintained among and between different cell types. One objective of this
study was to develop a 3-D in vitro model for human prostate morphogenesis to mimic the
in vivo acini and to use this for comparison with malignant cell behavior. Freshly isolated
acini from benign prostatic hyperplasia (BPH) tissue embedded in collagen and then
maintained as xenograﬁs in nude mice have been used for short term cultures to examine
epithelial-stromal interactions but these cells only have a ﬁnite lifespan (9). Salivary and
mammary gland epithelial culture models using culture on extracellular matrix have been
described (10,1 1). However, a 3-D matrix culture model for human prostatic epithelium using
well characterized cell lines was not available prior to the present study. This study includes

results on polarization and organization of non-malignant prostatic epithelial RWPE-l cells

79

into acini when maintained in Matrigel, and androgen-induced functional differentiation,
resulting in PSA expression. The behavior of RWPE-l cells in 3-D cultures, the assembly of
laminin in the basement membrane, and the expression of integrins was examined. The ability
of the non-tumorigenic RWPE-l cells to undergo acinar morphogenesis was compared with
that of the tumorigenic RWPE-Z and DU-l45 cells along with their invasive ability.
Epithelial cell polarization requires both cell-cell and cell-matrix adhesion (12,13).
Laminin, an essential component of epithelial cell basement membrane (BM), is the key BM
glycoprotein in cell:ECM adhesion and polarization. Little is known about the role of laminin
and its integrin receptors in prostate morphogenesis and differentiation. Laminin is a large
~800 kDa glycoprotein consisting of three chains, or, B, and 7 arranged in a cruciform-like
structure (13). It is the ﬁrst ECM protein to be synthesized during embryogenesis, which
points to its crucial role in morphogenesis and organogenesis (8,13,14). Laminin is a major
(60%) component of Matrigel, a re-constituted BM (8), used in this study for the 3-D
cultures. Since integrin receptors mediate epithelial cell polarization, integrin expression was
examined. Integrins, a complex family of heterodimeric transmembrane receptors for ECM
proteins, consist of a and B subunits and provide the link between the ECM and the
cytoskeleton. Complexing of a and B subunits in different combinations results in their
binding speciﬁcity for different ECM proteins. The a6Bl receptor is considered to be the
speciﬁc receptor for the E8 cell-binding domain of the a laminin chain, and is essential for
cell polarization, for example, in developing kidney tubules (13-16). While cell: basement
membrane interactions lead to structural morphogenesis of normal epithelial cells, androgens

are necessary for functional morphogenesis of prostatic epithelium and for the maintenance

80

of a differentiated phenotype. Therefore, the expression of PSA was used in this study as a

marker of ﬁrnctional differentiation in response to androgen exposure.

Materials and methods
Materials

Keratinocyte-serum free medium (K-SFM) no. 17005-042, antibiotic/antimycotic
mixture (PSF) no. 15240-013 (Gibco-BRL, Grand Island, NY); Dulbecco's modiﬁed Eagle's
medium (DMEM) no. CC293, RPMl-l640 no.CC26l (Celox Laboratories, Hopkins, MN);
Dulbecco's phosphate buffered saline (D-PBS Ca2’/Mg2‘-free) no. 283 74 (Pierce, Rockford,
IL); donor calf serum (DCS) no. 1 I 10-90 (Intergen, Purchase, N.Y.); mibolerone no. W-3OO
(BIOMOL Biomolecular Research Laboratories, Plymouth Meeting, PA); monoclonal
antibody to 0:6 integrin, clone GoH3 (AMAC Inc., Westbrook, ME); monoclonal antibody
to B1 integrin clone ml3 (giﬁ of Dr. SK. Akiyama, NIDR, NIH); monoclonal antibody to
laminin no. c-827 1(Sigma;, St. Louis, MO); monoclonal antibody to prostate speciﬁc antigen
no. M-07SO (DAKO, Carpinteria, CA); Vectastain Elite ABC Peroxidase Kit no. PK-6102
and 3,3-diamino—benzidine (DAB) Substrate Kit no. SK-4100 (Vector Laboratories,
Burlingame, CA); 24 well plates (Falcon) no. 08-7721; 12 mm circle coverslips no. 12-545-80

(Fisher Scientiﬁc, Itasca, IL); Matrigel was prepared as described (17).

Methods
Cell culture: RWPE-l and RWPE-Z cells were maintained in the complete K-SFM

medium which contains 50 11ng of bovine pituitary extract (BPE) and 5 ng/ml EGF, plus

81

antibiotic/antimycotic mixture PSF (Penicillin, 100 U/ml, Streptomycin 100 ug/ml and
Fungizone, 25 lug/ml). DU-145 human prostate carcinoma cell line was maintained in RPM]-
1640 medium containing 5% donor calf serum (DC S) and 1% PSF. Cells were passaged upon
conﬂuence and seeded at 2 x 106 cells/ T-75 ﬂask.

Acinar morphogenesis; 3-D cultures on Matrigel: Phenol red-free Matrigel was
used. To the wells of 24-well plates, 200 pl of Matrigel kept on ice, were added and the plates
were placed in a 37°C incubator for 1 h to allow the matrix to gel. RWPE-l cells were plated
at 100,000 cells/ well in complete K-SFM. The wells were divided, 24 h later, into treatment
groups in triplicate and the medium was replaced with test medium. Cells plated on plastic
served as the monolayer controls. Medium was changed every 48 h. Acinar formation was
followed and recorded photographically. At the appropriate time intervals, cultures were
washed gently with PBS, ﬁxed in 2.5% buffered paraformaldehyde for 15 min, rinsed twice
with PBS and either stained as whole mounts with Hematoxylin and Eosin (H&E) for gross
morphology or paraffin embedded and sectioned for immunostaining.

Androgen treatment: Experiments to examine the effects of androgen on
morphogenesis and expression of differentiated function employed mibolerone, a non-
metabolizable, synthetic androgen to ensure continuous androgen presence during the
experimental period. Stock solutions and cultures were protected from light. Mibolerone was
dissolved in absolute ethanol and was added to test media at 5 nM concentration, on the basis
of previously observed effects on human prostate epithelium in vitro (3,5). Medium was
changed every 48 h. Androgen effects on differentiation were evaluated by the expression of

PSA using immunostaining.

82

Conditioned medium (CM) for effects on morphogenesis: NIH 3T3 cells were
maintained in DMEM with 5% DCS. When cultures in 100 mm plates reached ~80%
conﬂuence, they were rinsed twice with PBS and placed on 7 ml serum-free DMEM.
Conditioned medium was collected 24 h later, centriﬁrged to remove cell debris, aliquoted and
frozen at -20°C until needed. For CM experiments, K-SFM medium and CM were mixed in
a 1:1 ratio for RWPE-l and RWPE-2 cells or RPMI with 5% DCS and CM in a l :1 ratio for
DU-l45 cells.

Immunocytochemical analysis: Protein expression was detected by a modiﬁed
avidin-biotin immunoperoxidase Vector protocol, using monoclonal antibodies (3 ,5). Matrigel
cultures were washed 2X with PBS, ﬁxed in paraformaldehyde, parafﬁn embedded, sectioned
at 5,4, mounted on poly-L-lysine coated slides and re-hydrated for immunocytochemical
analysis. Immunocytochemistry was performed using an indirect immuno-peroxidase modiﬁed
Vector protocol (3,5). Control sections lacked primary antibody. Antibody dilutions were
made in normal horse serum. All sections were washed twice in 500 pl PBS over 10 min
between all successive steps aﬁer primary Ab application. The following sequential steps were
conducted at room temperature and cells were rinsed twice with PBS between steps after
application of the primary Ab: sections were blocked with normal horse serum for 1 h,
incubated with the appropriate speciﬁc antibody for the indicated time followed by
biotinylated secondary Ab (1:200) for 30 min, treated with 3% H202 for 3 min to quench
endogenous peroxidase activity, incubated with the avidin-biotin-peroxidase complex for 30
min, developed with diaminobenzidine-nickel chloride (DAB) substrate for 4-6 min,

dehydrated and mounted on acid/alcohol washed slides. For immunostaining of PSA, cultures

83

were treated in K-SFM medium containing 5 nM mibolerone, a non-metabolizable androgen,
beginning 48 h aﬁer plating. Cultures were processed as described above. The primary
antibody dilutions were: for laminin 1:200, (:6 integrin 1:100, Bl integrin 1:100, and PSA
1:20, and incubation time was 24 h at 40 C.

Invasion assay: The method for the invasion assay is described in detail in the
accompanying paper (3). Brieﬂy, cells were plated at a density of 400,000 cells/well, in
triplicate, on Matrigel-coated ﬁlters in semm-free medium containing 0.1% bovine serum

albumin. Cells were allowed to invade for 24 h.

Results

Morphogenesis in 3-Dimensional cultures of RWPE-l, non-tumorigenic cells and
immunostaining for laminin, a6 and B1 integrins, and PSA: The effects of a basement
membrane substratum (Matrigel) on RWPE-l cell morphology under different media
conditions were ﬁrst examined. Twenty-four hours aﬁer plating, wells were divided into
several groups and were maintained on: 1 ) complete K-SFM; 2) complete K-SFM containing
5 nM mibolerone; 3) NIH 3T3 conditioned medium (CM) in a l :1 ratio with K-SFM and 4)
CM in a 1:1 ratio with K-SFM containing 5 nM mibolerone. Within 6 h after plating, cells
attached and they began to migrate and organize by forming cytoplasmic connections with
neighboring cells and by 24 h, a web-like pattern was seen in all cultures (Figure 4-1). With
time, cells plated on Matrigel migrate into the gel. Morphogenesis by RWPE-l cells was a

little more rapid in CM treated-cultures and by day 6, distinct duct-like structures or tubes

84

 

 

Figure 41. Appearance of RWPE-l cells on Matrigel. RWPE-l cells (10’ cells)
when plated on Matrigel show a web-like organization 24 h after plating in
complete K-SFM medium. (X20).

85

and acini were seen. Extensive branching and budding from tubes into acini-like structures
was seen in greater numbers in the CM treated cells than the controls by day 12 (Figures 4-2a
and 4-2b). This budding is also apparent in hematoxylin and cosine stained sections (Figure
4-2c). Figure 4-2d shows the in viva branching morphogenesis in the developing rat prostate
(18) while Figure 4-2e shows the in vitro morphogenesis of RWPE-l cells in Matrigel at day
16. It is clear that the in vitro morphogenesis and the formation of acini is similar to that seen
in vivo. Whole acini in live cultures of RWPE-l cells appear as blebs (Figure 4-2f). In
sections, the acini show a well organized, polarized epithelium with a distinct lumen (Figure
4-2g), similar to that seen in vivo in a histological section of the human prostate (Figure 4-2h)
from which RWPE-l cells were originally derived.

The organization of laminin in the basement membrane and the expression of a6 and
B1 integrins, and PSA were examined by indirect immunoperoxidase staining. Cells organized
into acini show a distinct basement membrane consisting of laminin surrounding the acini
(Figure 4-2i). Strong positive staining for (16 and B1 integrins was observed at the basal end
of the acinar cells (Figures 4-2j, 4-2k) as compared to the control (4-2l). Most importantly,
cells in acini from mibolerone treated cultures show strong positive staining for PSA (Figure
2m) as compared to the control lacking primary antibody (Figure 4-2n). The secreted material
in the lumen of the acini also stained positive for PSA (Figure 4-2m). With time, the acini
became distended with secretion and contained dead cells as seen in Figures 4-2k and 4-2m.
Cells grown on plastic did not show acinar organization and only formed a monolayer (F ig.4-

20).

86

 

   

‘s

 

ﬁgure 4-2. 3-D cultures. Acinar formation in vitro and in vivo. Photographs
were taken between 12-17 days after growth on Matrigel. a. RWPE-l culture
on day 12 after plating on Matrigel. Extensive branching and budding from
tubes into acini-like structures was seen in cultures containing CM and 5nM
mibolerone (X50); b. budding acini at higher magniﬁcation.(Xl98); c. an H&E
stained section showing branching and budding, day 12 (X105); d. branching
morphogenesis in neonatal rat prostate (from Yamashita et a1. 1996, Ref. 18);
c. branching and acinar morphogenesis in a 3-D in vitro Matrigel culture of
RWPE-l cells, day 16 (X105) mimics that in vivo shown in Figure 2d; f.
surface architecture of an in vitro culture on Matrigel showing several blebs
which represent the acini, day 12 (X220); g. cross sections (H & E) of an
acinus in an in vitro Matrigel culture showing polarized organization of cells
around a central lumen, day 12 (X525); h. a gland in a histological section of
a normal human prostate, from which RWPE—l cells were derived, is shown
for comparison (X230); I-n. immunoperoxidase staining of acini I. Iarninin
shows a well deﬁned basement membrane, day 12 (X390); j & k. «6 and B1
integrin subunits shows strong staining at the basal end of cells, day 17 (X250);
1 . a negative control, day 17 (X250); m. positive staining for PSA, day 17
(X250); n. a serial section negative control for PSA (X250); 0. RWPE-l cells
in a monolayer culture on plastic, day 12 (X55).

87

Comparison of the ability for structural morphogenesis by the non-tumorigenic
RWPE-l cells with that of the tumorigenic RWPE-Z and DU-145 cells in 3-D cultures:
When plated on Matrigel, RWPE-l cells rapidly begin to organize and within 24 h form a
network of cells (Figure 4-3 a). However, RWPE—Z cells failed to show such organization and
remained as single cells or formed small clumps (Figure 4-3b), and DU-l45 cells remained
as a monolayer (Figure 4-3c). After culture for 12 days, the RWPE-l cells showed well
developed branching with branches terminating in acini (Figure 4-3 d), whereas, the RWPE-Z
cells remained either as single cells or formed small clumps without any organization with
evidence of cell death (Figure 3f), while DU-145 cells formed amorphous balls (Figure 4-3 h).
In cross sections, the RWPE-l acini show distinct lumens lined by a single layer of cells
(Figure 4-3 e). The RWPE-2 cells showed no such organization (Figure 4-3 g) and the DU-145
cells formed amorphous solid balls without lumens (Figure 4-3i). These results demonstrate
that RWPE-l cells retain the ability to organize into acini, similar to those seen in viva but
the malignant cells have lost this ability.

The invasive ability shows an inverse relationship with the ability to undergo
acinar morphogenesis: The invasive ability of RWPE-l cells through Matrigel was
compared with that of RWPE-2 and DU-145 cells, where DU-145 invasion was taken as
100%. In comparison to DU-145 cells, the RWPE-l cells were not invasive while the RWPE-

2 cells showed 49% invasion (Figure 4-3j).

88

 

Percent Invasion

 

RWPE-l RWP‘E-l DU-145

Figure 4-3.Cell organizaton on Matrigel and relationship with invasion. a, b, c. Gross
morphology of cells plated on Matrigel as seen after 24 h, (XSO). a. RWPE-l; b. RWPE-Z; c.
DU-145 cells. d,f,h. cells plated in Matrigel as seen after 16 days (X50). (1. non-tumorigenic
RWPE-I cells undergo branching and acinar morphogenesis; f. tumorigenic RWPE-2 cells
remain as single cells or form small clumps; h. DU-145 cells form amorphous balls. e,g,i. cross
sections (H & E) of cultures representing those shown in 2f, 2h and 2e. e. RWPE-l cells are
polarized and organized into an acinus containing a lumen (X525); 3. RWPE-Z cells do not show
any organization, (X250); I. DU-145 cells form solid balls without a lumen (X250); j. The
invasive ability of non-tumorigenic RWPE-I cells was compared with that of the tumorigenic
cell lines RWPE-Z and DU-l45 using a modiﬁed Boyden chamber assay (see Ref. #3). Cells
were plated on Matrigel-coated ﬁlters at 400,000 cells per chamber and were allowed to invade
for 24 h. Invasion by DU-l45 cells was taken as 100%.

89

Discussion

In prostatic tissue homeostasis, a balance is maintained between growth stimulation
and inhibition, where the latter involves a decrease in cell proliferation with terminal
differentiation into secretory cells which are eventually sloughed off. In this context then,
morphogenesis includes two steps: 1) structural differentiation involving acinar organization
and 2) ﬁrnctional differentiation involving expression of secretory products. Shifting the
balance in favor of ﬁmctional differentiation should result in a net decrease in growth, a
desirable effect for counteracting growth of cancer cells and re-initiation of growth in middle
age men which results in BPH. Our results demonstrate that RWPE- l , immortalized but non-
tumorigenic cells, retain the ability to undergo acinar differentiation as a result of their
interaction with ECM, whereas, the v-Ki-ras-transformed RWPE-2 tumorigenic cells and
DU-l45 human prostatic carcinoma cells have lost the ability to organize in the same manner.
Similar observations have been made when normal and malignant human breast epithelial cells
were grown on ECM (11,19).

Laminin plays a critical role in cell-ECM adhesion, polarization, morphogenesis and
differentiation and the response of cells to laminin in vitra resembles the differentiated
phenotype of cells in viva (8,13,14). Laminin, a major (60%) component of Matrigel, has
been used to study human endothelial and salivary gland morphogenesis (10,20). Our results
show that RWPE-l cells are able to assemble a well deﬁned, laminin-containing BM at the
periphery of the acini on the basal cell surface. Current evidence suggests that the a chain of
laminin is necessary for basement membrane assembly, morphogenesis, polarization and
differentiation in viva (2,16,21). While the induction of kidney tubulogenesis correlates with

the appearance of laminin in the mesenchyme, the expression of a chain correlates with

90

polarization. Inhibition of polarization in organ culture by antibody against a chain suggests
that a chain is needed for establishing epithelial cell polarity (2,16,21).

Because of the apparent role of a chain in cell organization, we examined the integrin
receptor which bind speciﬁcally to the E8 fragment of laminin or chain. Integrin receptors
mediate epithelial cells polarization. Amongst the known integrin receptors for laminin, a6B1
is considered to be the speciﬁc receptor for the E8 cell-binding domain of the a laminin chain
and is essential for cell polarization (13,15,16). Laminin a chain and a6 integrin are
coordinately expressed and are co-localized in regions where epithelial cell morphogenesis
occurs and polarization begins in kidney tubule formation (16). Antibodies to 0:6 integrin
subunit selectively block binding to the E8 domain and inhibit cell polarization in kidney
tubule formation (15,16). The Bl subunit is also an essential part of the laminin receptor and
antibody to B1 interferes with morphogenesis (10,16,22). A loss, decrease or diffuse
expression of integrin receptors is associated with loss of normal cell organization in cancer
cells (11). RWPE-l cells show strong expression of both a6B1 integrins at the basal end of
cells forming the acini and thus, this immortalized cell line maintains the ability to undergo
structural morphogenesis similar to normal prostatic epithelial cells.

Information about integrin receptors in the human prostate is very limited and is based
primarily on cancer cell lines. Little is known about integrin expression in normal human
prostatic epithelial cells. In one study, (16, B1 and B4 were all expressed at the basal end of
epithelial cells in the normal human prostate, whereas, (16 and B1 showed a diffused
distribution and B4 was not detectable in prostate cancer (23). Human prostate carcinoma cell
lines DU-l45 and PC-3 were shown to have an unstable, heterogeneous and diffuse

expression of a6Bl and (16134 integrins while in LNCaP cells a6B4 was absent (24,25). On

91

the basis of current knowledge from the prostate and other tissues as well as the present
study, a6Bl may emerge as an important laminin integrin receptor involved in prostate
morphogenesis.

Having established structural differentiation in our 3-D prostate epithelium model, the
second step in morphogenesis is to accomplish functional differentiation involving expression
of secretory products. Androgens regulate growth and functional differentiation of prostatic
epithelium (3,5,6,26). PSA is an important marker of the differentiated function of prostatic
epithelial cells and its expression is androgen-dependent. We have previously shown that,
depending on the duration of treatment and concentration, RWPE-l and RWPE-2 cells as
well as our PWR-lE cell line in monolayer cultures show heterogeneous expression of PSA,
upregulate AR and show a growth response when exposed to mibolerone (3,5). PWR-lE is
a human prostate epithelial cell line immortalized by an adenovirus- l 2/ simian virus-40 (Ad 12-
SV40) hybrid virus (5,6). DU-l45 cells are considered to be negative for PSA (7). An
increase in PSA secretion was observed when freshly isolated prostatic epithelial cells were
plated on matrix-coated plates as compared to those on plastic (27). In the acini formed by
RWPE-l cells in Matrigel, PSA expression is strong in acinar cells and the secretion
sequestered in the acini also shows strong positive staining for PSA. These results
demonstrate that the acini formed by RWPE-l cells in vitra are able not only to synthesize
but to also secrete PSA into the lumen when exposed to the synthetic androgen mibolerone
and that the process of secretion is polarized, similar to that in normal prostatic acini in viva.

Bissell and collaborators have developed an elegant model of hormone responsive,
secretory mouse mammary epithelium to study differentiation and morphogenesis in vitra

(11,28). In this model, laminin in combination with lactogenic hormones, regulates tissue-

92

speciﬁc gene expression resulting in cell differentiation and the expression of mammary
speciﬁc B-casein synthesis in mammary epithelial cells in vitra (28). Similarly, our 3-
dimensional in vitra model, employing immortalized prostatic epithelial cells in the presence
of androgens, mimics characteristics of the human prostate during puberty and gland
remodeling. The processes of branching morphogenesis, acinar formation and differentiation
with secretion of PSA occur in the adult prostate as well as in BPH.

The RWPE-l and RWPE-Z cells provide fully characterized, standardized, and
reproducible cell lines. When grown in 3-D cultures, physiologically relevant models are
obtained for studies on the relationship between growth, morphogenesis, and ﬁrnctional
differentiation in normal cells and abnormalities associated with malignant transformation. In
the normal adult prostate in homeostasis, there is little net growth because a balance between
cell loss and cell gain is maintained among and between different cell types. This balance can
be modulated not only by the level of a speciﬁc growth inducer or inhibitor but also by the
number of its receptors. Since the expression of growth factors and receptors can be mutually
regulated by androgens and growth factors, the picture becomes very complex. The paracrine
interactions with stromal cells add another level of complexity, making it difﬁcult to separate
the various interactions in viva. Such interactions have been studied in the developing rat
prostate (18). The in vitra acinar morphogenesis that we have observed, using human
prostatic epithelial cells, resembles that seen in viva in the rat prostate (see Figures 4-2d and
4-2e).

The 3-D in vitra models for human prostate epithelial cells make it possible to dissect
the complex actions and interactions between cells, ECM proteins, growth factors, androgens

and their receptors in normal and neoplastic prostate growth, differentiation and

93

morphogenesis. Since laminin deposition must precede cell organization even in the
commonly used collagen gel raft cultures, the 3-D cultures in and on Matrigel, provide a
better model for studying the mechanisms of growth regulation involving morphogenesis and
differentiation in normal and neoplastic cells. These models also provide the opportunity to
study heterotypic cell interactions and changes in homeostasis during initiation of BPH. The
3-D cultures are especially useﬁJl for examining changes in cell organization that occur in
early neoplastic lesions such as PIN. The comparison of the ability to undergo acinar
morphogenesis by the immortalized RWPE-l cells to the inability of the v-Ki-ras-
transformed RWPE-2 cells and the formation of amorphous cellular masses by DU-l45
prostatic carcinoma cells may suggest a stepwise progression from immortalized to less and
then to a more aggressive and invasive carcinoma cells phenotype. We have shown earlier that
the invasive ability of RWPE-Z and DU-145 cells is associated with a marked increase in the
secretion of urokinase and/or gelatinases (29). DU-145 cells, which are more invasive than
RWPE-2 cells, secrete higher levels of urokinase than RWPE-Z cells (29). A loss of the
ability to undergo morphogenesis appears to be related to increased ability to invade and the
more undifferentiated the cell type the more aggressive the cancer. Many aspects of this
aberrant behavior remain to be investigated and further studies to elucidate the basis for the

inability to undergo structural morphogenesis are in progress.

94

References

1. Bostwick, DC. (1992) Prostatic intrapeithelial neoplasia (PIN): Current concepts. J Cell.

Molec. Biol. sugppﬁl. l6H10-19.

2. Rodriguez-Boulan, E., and Nelson, W.J. (1989) Morphogenesis of the polarized

epithelial cell phenotype. Science 245:718-725.

3. Bello, D., Webber, M.M., Kleinman, H.K., Wartinger, DD. and Rhim, IS. (1997)
Androgen responsive adult human prostatic epithelial cell lines immortalized by human

papillomavirus 18. Carcinogenesis (accepted for publication).

4. Rhim, J.S., Webber, M.M., Bello, D., Lee, M.S., Amstein, P., Chen, Land Jay. G.
(1994) Stepwise immortalization and transformation of adult human prostate epithelial
cells by a combination of HPV-18 and v-Ki-ras. Proc. Ngtl. Acad. Sci. USA 91:1 1874-

11878.

5. Webber, M.M., Bello, D., Kleinman, HK,, Wartinger, D.D., Williams, DE. and Rhim,
IS. (1996) Prostate speciﬁc antigen and androgen receptor induction and
characterization of an immortalized adult human prostatic epithelial cell line.

Carcimenesis 17: 1641-1646.

95

10.

Webber, M.M., Bello, D. and Quader, S. (1996) Immortalized and tumorigenic adult
human prostatic epithelial cell lines. Characteristics and applications. Part 1. Cell markers

and immortalized, non-tumorigenic cell lines. Prostate 29:386-394.

 

Webber, M.M., Bello, D. and Quader, S. (1997) Immortalized and tumorigenic adult
human prostatic epithelial cell lines. Characteristics and applications. Part 2. Tumorigenic

cell lines. Prostate 30:58-64.

 

Kleinman, H.K., Graf, 1., Iwamoto, Y., Kitten, G.T., Ogle, R.C., Sasaki, M., Yamada,
Y., Martin, GR. and Luckenbill-Edds, L. (1987) Role of basement membranes in cell

differentiation. Ann. NY. Acad. Sci. 513:134-145.

Hayward, S.W., Del Buono, R.D., Deshpande, N. and Hall, PA. (1992) A functional
model of adult human prostate epithelium: The role of androgens and stroma in

architectural organization and the maintenance of differentiated secretory function. J. Cell

&1022361-372.

Hoffman, M.P., Kibbey, M.C., Letterio, J .J. and Kleinman, H.K. (1996) Role of laminin-

] and TGF-B3 in acinar differentiation of a human submandibular gland cell line (HSG).

J. Cell Sci. 109:2013-2021.

96

ll.

12.

13.

14.

15.

16.

Howlett, A.R., Bailey, N., Damsky, C ., Petersen, O.W. and Bissell, M.J. (1995) Cellular
growth and survival are mediated by B1 integrins in normal human breast epithelium but

not in breast carcinoma. J. Cell Sci. 108, 1945-1957.

Drubin, D.G. and Nelson, J.W. (1996) Origins of cell polarity. Cell 84:335-344.

Malinda, KM. and Kleinman, H.K. (1996) Molecules in focus: the laminins. Int. J.

Biochem. Cell Biol. 28:957-959.

Klein, G., Langegger, M., Timpl, R. and Ekblom, P. (1988) Role of laminin A chain in

the development of epithelial cell polarity. _C_e_ll 55:331-341.

Aumailley, M., Timpl, R. and Sonnenberg, A. (1990) Antibody to integrin a6 subunit

speciﬁcally inhibits cell-binding to laminin fragment 8. Ex; Cell. Res. 188255-60.

Sorokin, L., Sonnenberg, A., Aumailley, M., Timpl, R. and Ekblom, P. (1990)
Recognition of the laminin E8 cell-binding site by an integrin possessing the «6 subunit

is essential for epithelial polarization in developing kidney tubules. J. Cell Biol.

111:1265-1273.

17. Kleinman, H.K., McGarvey, M.L., Hassell, J.R., Star, V.L., Cannon, F.B., Laurie, G.W.

and Martin, GR. (1986) Basement membrane complexes with biological activity.

Biochemistry 25:312-318.

97

18.

19.

20.

21.

22.

Yamashita, A., Hayashi, N., Sugimura, Y., Cunha, GR. and Kawamura, J. (1996)
Inﬂuence of diethylstilbestrol, leuprolelin (a luteinizing hormone-releasing hormone
analog). ﬁnasteride (a Sa-reductase inhibitor), and castration on the lobar subdivisions

of the rat prostate. Prostate 29:1-14.

 

Petersen, O.W., Ronnov-Jessen, L., Howlett, AR. and Bissell, M.J. (1992) Interaction
with basement membrane serves to rapidly distinguish growth and differentiation pattern
of normal and malignant human breast epithelial cells. Proc. Natl. Acad. Sci. USA

89:9064-9068.

Kubota, Y., Kleinman, H.K., Martin, GR. and Lawley, T.J. (1988) Role of laminin and
basement membrane in the morphological differentiation of human endothelial cells into

capillary-like structures. 1. Cell Biol. 107:1589-1598.

Nomizu, M., Kim, W.H., Yamamura, K., Utani, A., Song, 8., Otaka, A., Roller, P.R.,
Kleinman, H.K. and Yamada, Y. (1995) Identiﬁcation of cell binding sites in the laminin

or] chain carboxyl-terrninal globular domain by systematic screening of synthetic

peptides. J. Biol. Chem. 270:20583-20590.

Kim, W.H., Jun, S.H., Kibbey, M.C., Thompson, E.W. and Kleinman, H.K. (1994-1995)

Expression of B1 integrin in laminin-adhesion-selected human colon cancer cell lines of

varying tumorigenicity. Invasion Metastasis 14:147-155.

98

23.

24.

25.

26.

27.

28.

Knox, J .D., Cress, A.E., Clark, V., Manriquez, L., Affinito, K., Dalkin, BL. and Nagle,
RB. (1994) Differential expression of extracellular matrix molecules and the a6-integrins

in the normal and neoplastic prostate. Am. J. Pgthol. 145:167-174.

Rabinovitz, I., Nagle, RB. and Cress, AF. (1995) Integrin a6 expression in human
prostate carcinoma cell is associated with a migratory and invasive phenotype in vitra

and in viva. Clin. Exp. Met. 13:481-491.

Rokhlin, O.W. and Cohen, ME. (1995) Expression of cellular adhesion molecules on

human prostate tumor cell lines. Prostate 26:205-212.

 

Tindall, D.J . (1996) Androgen regulation of gene expression in the prostate. International

Symposium on Biolggy of the Prostat_e, NIDDK, NIH, Washington DC. p. 17.

Fong, C., Sherwood, E.R., Sutkowski, D.M., Abu-Jawdeh, G.M., Yokoo, H., Bauer,
K.D., Kozlowski, J .M. and Lee, C. (1991) Reconstituted basement membrane promotes

morphological and ﬁrnctional differentiation of primary human prostatic epithelial cells.

Prostate 19:221-235.

 

Streuli, C.H., Bailey, N. and Bissell, M.J. (1991) Control of mammary epithelial
differentiation: Basement membrane induces tissue-speciﬁc gene expression in the

absence of cell-cell interaction and morphological polarity. J. Cell Biol. 115: 1383-1395.

99

29. Webber, M.M., Waghray, A., Bella, D. and Rhim, IS. (1996) Proteases and invasion in
human prostate epithelial cell lines. Implications in prostate cancer prevention and

intervention. Radiat. Oncol. Invest. 3:358-362.

100

Chapter Five

Extracellular Matrix and Integrins
Modulate Acinar Morphogenesis by
RWPE-l Cells

101

Abstract

Interactions between cells and the extracellular matrix (ECM) play a critical role
in acinar and branching morphogenesis during prostate development. Using HPV-18
immortalized cell line, RWPE-l , a 3-dimensional (3-D) cell culture model was developed
and used to identify the speciﬁc extracellular matrix protein and its integrin receptor,
required for acinar morphogenesis in vitro. When cells were plated on, laminin, type IV
collagen or ﬁbronectin, laminin was the only ECM protein that was able to induce cell
polarization and formation of Acini with a distinct lumen. Cells plated on collagen or
ﬁbronectin remained as a monolayer. The number of acini formed in laminin was similar
to that formed in Matrigel. Blocking the binding of cells to laminin by anti-laminin
antibody caused a decrease in acinar formation in a dose dependent manner. These results
show that laminin is required for cell polarization and acinar formation.

The expression of a6 and B1 integrin subunits was examined by confocal
immunoﬂuorescence microscopy. RWPE-l cells show strong expression of a6 and B1
integrin subunits, which are expressed primarily at the basal end of polarized cells.
Treatment of cells with anti-functional antibody to B1 reduced acinar formation to 10% of
the IgG-treated control which was taken as 100%. However, similar treatment with anti-
functional (16 antibody only reduced acinar formation to 75% of control. These results
demonstrate that while both (16 and B1 are important, acinar formation does not occur in
the absence of the B1 integrin. It can, therefore, be stated that laminin and B1 integrin are
the two key factors required for acinar morphogenesis to occur. This is the ﬁrst report of
acinar formation in vitra for human prostatic epithelial cells. This work provides the basic

understanding of cellzmatrix interactions in prostate acinar morphogenesis. The 3-D cell

102

culture model can not only be used to examine normal development of the human prostate,
but also for examining early changes in prostate carcinogenesis. This provides the
opportunity to identify agents which may be used to prevent the development and

progression of prostate cancer.

Introduction

Prostatic epithelial cells, in viva, exist in a 3-dimensional (3-D) environment in
which the extracellular matrix (ECM), androgens, growth factors, receptors and stroma
inﬂuence cellular growth and differentiation (Fang et al. 1991, Webber et al., 1996 Bello
et al. , 1997). In 2-dimensional cell cultures, human prostatic epithelial cells undergo a few
rounds of cell division, lose differentiated function and are then lost. This makes their in
vitro behavior, in many instances, of little relevance to their in viva condition. Therefore,
it is proposed that, to effectively study prostatic epithelial cell biology, a cell systems
which mimics the 3-D micro-environment and allows for the expression of normal
prostatic functional characteristics, is needed. The use of fully characterized immortalized
human prostatic epithelial cells in an in virra 3-D culture system, described in chapter 2,
has adequately fulﬁlled this need.

The ECM is essential for the morphogenesis of tissues. It is involved in cell
migration, proliferation, differentiation and in establishing and maintaining cell polarity
and cell architecture (Hay, 1981). During embryonic development, the ECM plays an
important role in organogenesis and epithelial cell differentiation. In the developing
prostate, both the budding urogenital epithelial cells and the mesenchymal cells secrete

ECM proteins which the epithelial cells then use to "move", as a train would use railroad

103

tracks, as they branch into the mesenchyme to form acini and ducts. The basement
membrane is a structurally distinct zone of the ECM. In the prostate, it separates the
epithelial cells of the acini and ducts from the underlying stroma.

Although work has been done in lung, kidney, breast and salivary gland
morphogenesis, not much is known about ECM-regulated acinar morphogenesis in the
prostate (Hoffman et al., 1996, Schuger L, 1990, Weaver et al., 1996, Wu and Santoro,
1996). This chapter, will focus on the use of the human prostatic 3-D culture model to
identify the key ECM components and integrins essential for epithelial polarization and
acinar formation. Because the progression from a normal to a malignant phenotype is
characterized by a loss of normal cell organization, cell polarity, cellzcell and cell:ECM
adhesion and appropriate integrin receptor expression, the data describe in this chapter
provide the necessary baseline for further study into how these factors are altered in the
process of carcinogenesis.

As the key component of the basement membrane, laminins are important in
cell:ECM adhesion and epithelial polarization. Laminins are -800 kDa, large glycoproteins
which are composed of three chains, or, B, and y, held together by disulﬁde bonds and
arranged in a cruciform-like structure (Malinda and Kleinman, 1996). Laminins are the
ﬁrst ECM proteins synthesized during embryonic development, and therefore, play a
crucial role in morphogenesis and organogenesis (Kleinman, 1987, Malinda and Kleinman,
1996). Type IV collagen is also a major structural component in the basement membrane
and functions to anchor other ECM components (Timple and Aumailley, 1989). Individual
type IV collagen chains (567 bp) assemble to form 400 nm long triple helices which

aggregate to form a loose but continuous network which serves as the ”scaffold" of the

104

basement membrane. Although type IV collagen may bind to cells directly, via cell matrix
receptors, epithelial cells attach to collagen indirectly through laminin. Fibronectin, a large
glycoprotein (210-260 kDa), found ubiquitously throughout the body tissues and ﬂuids,
functions as bridges between cells and other ECM molecules. Although ﬁbronectin is not
found in the basement membrane in adults, it is present in most embryonic basement
membranes (Kleinman et al. , 1982), which suggests a role in the initial stages of prostatic
acinar and ductal morphogenesis. It is, therefore, worthy of investigation.

ECM proteins bind to the cell surface via speciﬁc receptors. The ﬁrst family of
receptors identiﬁed to do so, was the integrin family. Integrins are known to play an
important role in embryogenesis, organogenesis, wound repair, cytoskeletal organization,
motility, cell growth and differentiation (Geiger etal., 1995, Vinatier, 1995). Because of
their critical role in epithelial cell polarization, the identiﬁcation of key integrins involved
in prostatic acinar morphogenesis was examined (Webber et al., 1997). Integrins are
heterodimeric transmembrane glycoproteins which consist of an a (120 - 180 kDa) and a
B (90-110 kDa) subunit both of which are transmembrane proteins. The B subunits are
promiscuous and can complex with multiple or subunits, while the ct subunits are generally
loyal to their B subunits partner (Garratt and Humphries, 1995, Vinatier, 1995).

The human prostatic epithelial cell line RWPE-l , which mimics normal prostatic
epithelial cells in viva, was used in this study. In order to identify the key basement
membrane proteins and integrin receptors, involved in prostatic epithelial cell polarization
and acinar morphogenesis, a 3-D cell culture model was used. The ability of single ECM
components, laminin, type IV collagen and ﬁbronectin, to induce RWPE-l cells to

undergo acinar morphogenesis, was examined. Further, the expression of integrin

105

receptors a6B1 and a6B4, which bind laminin exclusively, and a2131 , which binds laminin
as well as type IV collagen and ﬁbronectin, was examined to elucidate their role in cell
polarization. We have already established that RWPE-l cells undergo acinar
morphogenesis in Matrigel 3-D cultures (Webber et al 1997). Therefore, the Matrigel
cultures were used as a positive control while those in gelatin were used as a negative

control .
Results

Laminin is required for acinar morphogenesis

The ability of individual basement membrane glycoproteins, laminin (laminin-1 was
used in this study), type IV collagen or ﬁbronectin, to induce RWPE-l acinar
morphogenesis was ﬁrst examined to determine which component was essential for acinar
formation. RWPE-l cells were plated onto either laminin, type IV collagen or ﬁbronectin-
coated wells. Matrigel and gelatin-coated wells served as positive and negative controls,
respectively. Within 24 h after plating, cells plated on all matrices attached but only those
cells on Matrigel formed cytoplasmic connections with neighboring cells. At four days,
cells plated on Matrigel had formed distinct acini with associated ducts which gave a
"bunch of grapes" appearance (Figure 5-1 a). When the "grape bunch” structure was
stained with anti-laminin antibody (Ab) and optically sectioned by laser scanning confocal
microscopy, laminin expression was found to be most intense at areas immediately
surrounding individual acini and weaker at areas of duct—matrix contact (Figure 5-1 6).
Optical serial sections through a single propidium iodide-stained acinus demonstrated that
the acinus was a 3-D spherical gland consisting of polarized cells with a distinct central
lumen (Figure 5-2 a-f).

106

 

Figure 5-1. Acinar morphogenesis by RWPE-l cells in Matrigel. Localization of
laminin in four day 3-D Matrigel cultures using anti-laminin antibodies. a. LSM
transmitted image of acini and associated duct network. b. optical confocal section of
laminin expression. Note strong staining surrounding base of acini (Bar 25 u).

 

Figure 5-2. Optical serial sections of a propidium-iodide stained RWPE—l acinus grown on
Matrigel for 4 days. Sections are arrange in a top to bottom conﬁguration with (a) being the
top, (c)and (d) more middle and (f) the bottom of the acinus (Bar 25p).

107

Most importantly, cells plated on laminin formed distinct spherical acini consisting
of polarized cells with a distinct lumen (Figure 5-3 a-f). This clearly demonstrates the
ability of laminin-l to induce acinar formation. A comparison of acini formed on Matrigel
and laminin-1 (Figure 5-4 a and b, respectively), shows that both consist of well polarized
cells around a central lumen.

When cells were plated on ﬁbronectin and type IV collagen acinar formation did
not occur. Cells merely attached and spread (Figure 5-4 c and d, respectively). Cells
grown on gelatin did not form acini or spread but formed amorphous clumps (Figure 5-4
e). Cells grown on plastic did not show acinar organization and only formed a monolayer
(Figure 4-20, Chapter 4 page 87).

When the number of acini formed on the different matrices was compared, no
signiﬁcant difference between Matrigel (100%) and laminin-1 (98%) was observed
(Figure 5-5). No acini (0%) formed on either type IV collagen or ﬁbronectin.

If laminin is necessary for acinar formation, then blocking the binding of laminin
to the cells should result in the inhibition of acinar formation. This was tested using anti-
functional (neutralizing) antibody directed against laminin. The anti-laminin Ab binds to
the laminin in Matrigel and physically prevents the cells from binding to laminin. The
neutralizing laminin Ab was diluted 1 :20, 1:10 and 1:5 and added to Matrigel-coated wells
1 h before plating cells. Results showed that acinar formation was decreased to 23 % , 4%
and 2.5% when compared to IgG treated controls (100%) (Figure 5-6). These result

indicate that cell must bind to laminin in order for acinar formation to occur.

108

 

Figure 5-3. Optical serial sections of a propidium iodide-stained RWPE-l acinus
grown on laminin for 4 days. Sections are arrange in a top-bottom conﬁguration
with (a) being cranial, (c)and (d) medial and (f) posterior (Bar 25p).

 

Figure 5-4. RWPE-l cells on ECM Matrices. Optical sections of propidium iodide-
stained RWPE—l ceUs at 4 days of growth on (a) Matrigel (positive control), (b) laminin,
(c)ﬁbronectin, ((1) type IV collagen and (e) gelatin (negative control). Acinar formation
was only observed on (a) Matrigel and (b) laminin. Cells spread, but did not form acini,
on (c)ﬁbronectin and (d) type IV collagen. (e) Cells formed amorphous clumps on gelatin
(Bar 25 u).

109

3 so
a
C
.- 8
C
E .5 so
‘5
0
g 40
$
20
o ,7 .__ L 7 7 V—

 

 

 

 

 

 

 

 

 

Matrigel
Lamlnln
Type IV ‘
Collagen
Fibronectln
Gelatin

Matrix Component

Figure 5-5. Acinar Formation on Matrix Components. RWPE-l cells were grown on
Matrigel, laminin type IV collagen, ﬁbronectin or gelatin. Acini-forming ability was
determined by staining cultures with propidium iodide and counting individual acini using
the LSM transmitted light at low power. Counts are expressed as a percent of the control.
Matrigel control was set at 100% (Bar=+/-SE).

 

 

 

a 100 ——
80
E
o 60
E 2
2 2
§ 40
3
E
20 I D
0 .7 __ 7.“: ,_L_I:LL.
IgG 1:20 1:10 1:5
Control
Ab Dilution

Figure 5-6. Inhibition of acinar formation by anti-laminin neutralizing
antibody. RWPE—l cells were grown on Matrigel and pre-treated with laminin
anti-functional antibody. (a) Acini-t‘onning ability was determined by staining
cultures with propidium iodide and counting individual acini using the LSM
transmitted light at low power. Counts are expressed as a percent of the IgG
control which was set at 100% (Bar: +/-SE). (b) Optical confocal section of
a control acinus.(c) Optical confocal section of 1:20 antibody-treated cells
which form aggregates but lack acinar formation (Bar 25 u).

110

Bl integrin is essential for acinar formation

Epithelial cells bind laminin via integrin receptors. Therefore, a6B1, a6B4 and
a2B1 laminin integrin receptor expression was examined, using antibodies directed against
a2, (16, B1 or B4 integrin subunits, by indirect immunoﬂuorescence staining of intact
acini. Integrin expression was visualized by optical sectioning using a laser scanning
confocal microscope.

Strong positive staining for (16 (Figures 5-7) was observed at the basal end of the
acinar cells as compared to the IgG control (Figure 5-8). Strong staining was observed also
for the B1 integrin subunit (Figure 5-9). Bl integrin expression was located at the basal end
of the acinar cells in comparison to the IgG control. These results demonstrate that
RWPE-l cells show strong basal expression of the a6B1 laminin integrin. In comparison
to a6 or B1 staining, only weak staining for «2 was observed as compared to the IgG
control. Staining appeared focally at the basal end of the acinar cells (Figures 5-10).
Integrin B4 staining was also observed. Staining appeared weak but continuous at the basal
and lateral edge of acinar cells (Figures 5-11) as compared to the IgG control. Taken
together, the data show that RWPE-l cells weakly express a6B4 and a2B1 laminin
integrins, while the a6Bl laminin integrin expression is predominant.

In order to demonstrate that the (16 and B1 laminin integrin subunits are the key
integrins involved in epithelial polarization and acinar morphogenesis, anti-functional
antibodies against B1 or a6 integrins were employed. B1 or a6 integrin antibody, diluted
1:33 and 1:5, respectively, was added to the cell suspension and allowed to incubate for

1 h before cells were plated onto Matrigel-coated wells. Anti-B1 Ab decreased acinar

111

 

) .

Figure

 

g.

5-7. or

6 Integrin Expression.

   

    

  

RWPE-l cells, grown on Matrigel, were

ﬁxed at 4 days and stained for a6 integrin expression by immunoﬂuorescence
using anti-a6 antibody. (a) LSM transmitted image of an acinus. (b) Optical
confocal section of the same acinus showing a6 integrin expression. Note the
strong staining at the basal end of polarized cells forming the acinus (Bar 25 u).

 

Figure 5-8. IgG Control
for integrin expression.
RWPE-l cells, grown
on Matrigel, were ﬁxed
at 4 days and stained
using a non-speciﬁc IgG
instead of the primary
antibody . (a)LSM
transmitted image of an
acinus. (b) No non-
speciﬁc staining was
observed when lgG was
used (Bar 25 u).

 

Figure 5-9. B1 Integrin Expression. RWPE-l cells, grown on
Matrigel, were ﬁxed at 4 days and stained for BI integrin expression
by immunoﬂuorescence using anti-Bl antibody. (a) LSM transmitted
image of acini. (b, c and (1) Bl integrin expression as seen in optical
serial confocal sections. Note the strong staining at the basal end of
polarized cells forming acini and that different acini come into view
at different optical depth (Bar 25 u).

112

    

 

:il b

Figure 5—10. (12 Integrin Expression. RWPE-l cells, grown on Matrigel, were
ﬁxed at 4 days and stained for 0L2 integrin expression by immunoﬂuorescence
using anti-a2 antibody. (at) LSM transmitted image of acini. (b) Optical
confocal section of acini showing a2 expression. Note focal staining at the
basal end of polarized cells forming acini (Bar 25 u).

 

 

Figure 5-11. B4 Integrin Expression. RWPE-l cells. grown
on Matrigel, were ﬁxed at 4 days and stained for B4 integrin
expression by immunoﬂuorescence using ant-B4 antibody.
(a.b) LSM transmitted image of acini. (c,d) Optical confocal
section of acini showing B4 integrin expression. Note weak
staining at the basal and lateral surface of polarized cells
forming the acini (Bar 25 u).

113

formation to 10% of the IgG control which was set at 100% (Figure 5- 12). However, anti-
functional a6 integrin Ab only caused a decrease to 75 % of control in acinar formation
(Figure 5-12). These results show that binding of cells through the B1 integrins receptor

is essential for cell polarization and acinar morphogenesis.

Discussion

The process of prostatic acinar morphogenesis is poorly understood. Animal
models, although useful, intrinsically passes a plethora of unknown variables which can
not be fully controlled. Additionally, such in. vivo animal systems primarily use rodent
prostatic cell lines which may be injected at various sites into animals, producing data
which must then be extrapolated to the human condition. Monolayer 2-dimensional cell
cultures, utilizing primary human prostatic epithelial cells are ideal, but isolated cells
typically undergo a few rounds of cell division, lose differentiated function, undergo
senescence and are lost. In many instances, the information gathered from these 2-D cell
culture systems is of little relevance to the in viva condition. In order to overcome the
limitations of using primary cultures of human prostatic epithelial cells, our laboratory has
developed the RWPE—l cell line. The RWPE-l cell line was derived from non-neoplastic
human prostatic epithelial cells by immortalization with human papilloma virus 18. These
cells mimic normal prostatic cell behavior in their response to growth factors and androgen
and in their expression of PSA and androgen receptor in response to androgen stimulation.
They do not form tumors in nude mice nor are they invasive (Bello et al. , 1997). Because
these cells are immortalized but have retained their normal functional ability, they provide

a uniform, standardized and reproducible cell model for use in studying prostate

114

120 nr

100

80

60

Aclnl
(Percent of Control)

 

 

196 Control Integrin alpha 8 Integrin beta 1

Figure 5-12. Inhibition of acinar formation by anti-integrin neutralizing
antibodies. RWPE-l cells were grown on Matrigel, pre-treated with anti-
functional integrin 0.6 or B1 integrin antibody for l h. Acini-forming ability
was determined by staining cultures with propidium iodide and counting
individual acini using the LSM transmitted light at low power. Counts are
expressed as a percent of the IgG control which was set at 100%.

115

development, prostate cell biology and the neoplastic transformation which occurs in the
process of carcinogenesis. I have described a 3-D cell culture system, using RWPE-l
human prostatic epithelial cells, which mimics the in viva micro-environment and allows
for a reproducible model system. This model system can be used to study the factors
involved in prostatic acinar morphogenesis and how these factors are altered in the
progression through prostatic intraepithelial neoplasia (PIN) and carcinogenesis.

It should be stated at the outset that when performing 3-D cultures, it is important
to use strict criteria for evaluation of acini. The best deﬁnition for an acinus is: a spherical
arrangement of polarized cells which form a distinct central lumen. In order to conﬁrm
acinus formation under this deﬁnition, an investigator may examine cultures by either
2 um parafﬁn sections or by optically sectioning acini using a confocal microscope. Light
microscopy and phase contrast techniques alone are not sufﬁcient for evaluation of acini
formation. I have found that preperation of samples for confocal microscopy causes the
least distortion of the 3-D architecture, as compared to the embedding process, and the loss
of material due to processing is minimal. Recently, other investigators have reported
gland-like structures being formed when mouse prostatic cancer cells were grown on
Matrigel. Unfortunately, this study lacks clear illustration of polarized cells with distinct
lumen. The phase contrast pictures presented were not sufﬁcient for conﬁrmation of acini
formation (Jorcyk, et al., 1998). Results reported in the present study are the ﬁrst to show
acinar formation in vitra for human prostatic epithelial cells.

The focus of this chapter is to elucidate the role of ECM components and the
associated integrin receptors in acinar morphogenesis using this 3-D cell culture system.

In the developing prostate, urogenital epithelial cells bud out of the urogenital sinus to

116

invade the surrounding mesenchyme where they terminate and differentiate into acini. It
is known that the ECM, that is deposited by both the budding epithelial cells and the
ﬁbroblasts in the mesenchyme, plays an important role in acinar formation. It is, however,
not known as to which ECM component is essential for acinar formation. When grown on
the reconstituted basement membrane Matrigel, RWPE—l cells form distinct acini and ducts
(Figure 5-1 a and Figure 5—2). Therefore, the ability to form acini on laminin-1, type IV
collagen and ﬁbronectin was examined. Laminin-1 is the predominant component of the
basement membrane and Matrigel (Hoffman et al., 1996, Malinda and Kleinman, 1996).
Type IV collagen is another major component of the basement membrane and acts as the
scaffold onto which laminin and cells may bind (Hay, 1981). Although ﬁbronectin is not
found in adult basement membranes, it is present in embryonic basement membranes when
initial epithelial budding and branching is occurring (Kleinman et al., 1982). Results
presented here show that laminin-1 (Figure 5-4 b), and not type IV collagen or ﬁbronectin
(Figure 4-4 c and d), was able to induce acinar formation and the number of acini formed
in laminin was comparable to that observed in Matrigel (Figure 5-5). These results
demonstrate the ability of laminin-1 to induce acinar formation in the same manner and
frequency as Matrigel. The ability of RWPE-l prostate epithelial cells to undergo acinar
formation in laminin is supported by similar acinar organization seen by others who have
utilized mammary and submandibular epithelial cells (Hoffman et al. , 1996, Weaver et al. ,
1996).

If laminin truly plays an essential role in acinar formation as implicated above, then
blocking cell binding to laminin should disrupt acinar formation. Results show that the

addition of anti-functional Ab to laminin, which blocks cell-to-laminin binding, reduced

117

the ability of RWPE-l cells to form acini by 97.5 % at the lowest Ab dilution (1:5) and
77% at the highest dilution (1 :20) (Figure 5-6). Taken together, these results demonstrate
that the laminin component of the basement membrane is necessary for prostatic acinar
morphogenesis.

Cells attach to the ECM via integrin cell adhesion receptors. Integrins are critical
for establishing cell polarity and differentiation of epithelial cells into specialized secretory
cells (Sheppard, 1996, Webber et al., 1997). Speciﬁcally, the a6Bl is the key integrin
receptor for laminin—l and is essential for cell polarization (Sorokin, 1990, Webber et al. ,
1997). As shown earlier, a6 and B1 integrin subunits are strongly expressed by RWPE-l
cells, suggesting strong expression of the a6Bl integrin receptor (Figure 5-7 and 5-8).
RWPE-l acini also express «2 and B4 integrins. This suggests that a6B4 and a2Bl laminin
integrins may also be formed, but their expression is substantially weaker than a6Bl
expression, indicating that their role may be a supporting one (Figure 5-10 and 5-11). The
presented data indicate that the primary laminin integrin expressed by RWPE-l cells is the
a6B1 integrin.

If a6B1 integrin is the key integrin involved in cell polarization and acinar
organization, then the disruption of binding of the a6Bl integrin receptor to laminin should
result in inhibition of epithelial cells polarization, required for acinar formation. The
importance of a6Bl integrin in prostatic epithelial cells polarization and acinar
morphogenesis was further elucidated by blocking a6 and B1 function using neutralizing
Abs. Anti-functional Ab to B1 showed a dramatic 90% decrease in acinar forming ability
(Figure 5-12). While, inhibition of a6 function resulted in only a 25 % decrease. Although
studies on submandibular gland epithelial cells have shown that disruption of either (16 or

118

Bl integrin signiﬁcantly affects acinar formation (Hoffman et al., 1996), our ﬁndings
indicate that in the prostate, the B1 and not the a6 integrin subunit is critical for epithelial
polarization and acinar morphogenesis.

Normal prostatic epithelial cells and epithelial cells in prostatic intraepithelial
neoplasia (PIN) express a6B1, a6B4 and a2B1 laminin receptors, but the expression of a2
and B4 are lost in cancer cells (Cress et al., 1995). This is not surprising since cancer is
characterized by a loss of normal cell organization, cell polarity, cellzcell and cell:ECM
adhesion which would result from a loss of appropriate integrin receptor expression.
Although expression of a6B1 is present in prostate cancer cells, its expression is not
restricted to the basal aspects of the cell, as in normal cells, but is diffuse throughout the
cell membrane (Cress et al., 1995, Hao et al., 1996, Rabinovitz etal., 1995, Rokhlin and
Cohen, 1995). In RWPE—l cells, the expression of a6Bl integrin is restricted to the basal
cell surface. RWPE-l cells also express a2Bl and a6B4 integrins but their expression is
much weaker than a6Bl expression.

In summary, the data presented here show that laminin-l is essential for acinar
morphogenesis and that cell polarization and organization occurs in the presence of
functional a6Bl integrin binding. Absence of type IV collagen and ﬁbronectin does not
inhibit prostatic epithelial cells polarization and cell:matrix and cellzcell interactions.
RWPE-l prostatic epithelial cells undergo epithelial cell polarization and acinar formation
when grown on the reconstituted basement membrane, Matrigel, as a result of their
interaction with laminin via the a6B1 integrin. This work has provided the basic
understanding of cell:matrix interactions necessary for further study of changes in

cell:matrix interactions that occur during the process of carcinogenesis. It is hoped that

119

further studies, utilizing the 3-D cell model system, will shed light on the role of
laminin-l, a6B1 integrins and other factors in the progression from normal to malignant

prostatic epithelial cells.

120

References

Bello D, Webber MM, Kleinman HK, Wartinger DD and Rhim JS: Androgen responsive
adult human prostatic epithelial cell lines immortalized by human papillomavirus

18. Carcinogenesis 18: 1212-1223, 1997.

Cress AE, Rabinovitz I, Zhu W and Nagle RB: The alpha 6 beta 1 and alpha 6 beta 4
integrins in human prostate cancer progression. Cancer Metastasis Rev. 14:219-28,

1995.

Fong CJ, Sherwood ER, Sutkowski DM, Abu-Jawdeh GM, Yokoo H, Bauer KD, Kozlowski
JM and Lee C: Reconstituted basement membrane promotes morphological and

functional differentiation of primary human prostatic epithelial cells. Prostate 19:

221-35, 1991.

Garratt AN and Humphries Ml: Recent insights into ligan binding, activation and

signalling by integrin adhesion receptors. Acta Anat. 154, 34-45, 1995.

Geiger B, Yehunda-Levenberg S and Bershadsky AD: Molecular interactions in the
submembrane plaque of cell-cell and cell-matrix adhesions. Acta Anat. 154:46—62,
1995 .

Hao J, Yang,Y, McDaniel, KM, Dalkin BL, Cress AE and NagleRB: Differential expression
of laminin 5 (alpha 3 beta 3 gamma 2) by human malignant and normal prostate. Am.

J. Pathol.]49zl34l-9, 1996.

121

Hay E: Collagen and embryonic development. In: Cell biology of extracellular matrix.

Hay E (ed.), New York, Plenum Press, pp.379-405, 1981.

Hoffman MP, Kibbey MC, Letterio JJ and Kleinman HK: Role of laminin-1 and TGF-B3 in
acinar differentiation of a human submandibular gland cell line (HSG). J. Cell Sci.

109:2013-2021, 1996.

Jorcyk C, Liu-ML, Shibata MA, Maroulakou IG, Komschlies KL, McPhaul MJ, Resau
JH and GreenJE: Development and characterization of a mouse prostate
adenocarcinoma cell line: ductal formation determined by extracellular matrix.

Prostate 34:10-22, 1998.

Kleinman HK, Woodley DT, McGarvey ML, Robey PG, Hassell JR and Martin GR:
Interaction and assembly of basement membrane components. In extracellular

matrix eds. Hawkes S and Wang JL, Academic Press, 1982.

Malinda KM and Kleinman HK: Molecules in focus: the laminins. Int. J. Biochem. Cell Biol.

28:957-959, 1996.

Orian-Rousseau V, Aberdam D, Rousselle P, Messent A, Gavrilovic J, Meneguzzi G;
Kedinger M and Simon-Assmann P: Human colonic cancer cells synthesize and

adhere to laminin-5. J. Cell Sci. 30( Pt 14):1993-2004, 1998.

122

Rabinovitz l, Nagle RB and Cress AE: Integrin alpha 6 expression in human prostate

carcinoma cells is associated with a migratory and invasive phenotype in vitro and in

vivo. Clin. Exp. Metastasis. 13: 481-91, 1995.

Rokhlin OW and Cohen MB: Expression of cellular adhesion molecules on human prostate

tumor cell lines. Prostate. 26: 205-12, 1995.

Schuger L, O'Shea KS, Nelson BB and Varani J: Organotypic arrangement of mouse
embryonic lung cells on a basement membrane extract: involvement of laminin.

Development. 110: 1091-9, 1990.

Timpl R and Aumailley M: Biochemistry of basement mambranes. Adv. Nephrol. 18:59-

76, 1989.

Vinatier,D. Integrins and reproduction. Euro. J. Obs. Gyn. Reprod. Bio. 59, 71-81,1995.

Weaver VM, Fischer AH, Peterson OW and Bissell MI: The importance of the

microenvironment in breast cancer progression: recapitulation of mammary

tumorigenesis using a unique human mammary epithelial cell model and a

three-dimensional culture assay. Biochem. Cell Biol. 74:6,: 833-51, 1996.

123

Webber MM, Bello D, Kleinman HK and Hoffman MP: Acinar differentiation by non-
malignant immortalized human prostatic epithelial cells and its loss by malignant

cells. Carcinogenesis 18:1225-1231, 1997.

Wu J E and Santoro SA: Differential expression of integrin alpha subunits supports distinct

roles during lung branching morphogenesis. Dev. Dyn. 206:2, 169-81, 1996.

124

Chapter Six

Malignant Human Prostatic Epithelial Cells Lose
the Ability to Undergo Acinar Morphogenesis

125

Abstract

Acinar morphogenesis occurs as a result of complex multiple interactions between
cells and the surrounding ECM, via their integrin receptors, as well as growth factors
present in the micro-environment and growth factor receptors expressed by cells.
Alterations in these interactions may occur during carcinogenesis, as cells progress from
a normal to a malignant, invasive phenotype. A family of cell lines, which apparently
represent the multiple steps in carcinogenesis, similar to PIN, as well as subsequent tumor
progression, were used to determine whether the degree of loss of ability of cancer cells
to undergo acinar morphogenesis, in the 3-D cell culture model, has any relationship with
changes in integrin expression, their invasive ability, and their response to EGF. The cell
lines used include: RWPE-l, which represents the non-tumorigenic but immortalized state,
and the tumorigenic cell lines, arranged in order of their increasing malignant
characteristics, WPEl-NA22, WPEl-NB27, WPEl-NB14; WPEl-NBI 1, and
WPEl-NB26, all of which were derived from RWPE-l by transformation with NMU. The
DU-145, malignant and invasive prostate cancer cell line was used as a control. Results
show that an inverse relationship exists between the degree of ability to form acini and the
degree of invasive ability, where RWPE-l cells and WPEl-NA22 show high
acinar-forming ability but no or little invasive ability, while the more invasive cell lines
WPEl-NB26 and DU-145 cells show no acinar-forming ability. In addition, while
RWPE-l and WPEl-NAZZ cells show basal expression of (1631 integrin, which correlates
with their ability to polarize and form acini, a loss (16 and abnormal, diffused expression
of [31 integrin in WPEl-NB26 and DU-145 cells, was observed. Further, exposure of

RWPE-l cells to EGF causes a dose dependent decrease in acinar formation, at 10 ng/ml

126

EGF, to 56% of untreated control. Addition of EGF anti-functional antibody caused an
enhancement of acinar formation. These results suggest that malignant cells lose the ability
to undergo acinar morphogenesis and that the degree of loss is directly related to their
invasive ability and loss or abnormal expression of laminin-speciﬁc integrins. Another
contributing factor to loss of acinar morphogenesis appears to be increase in cell
proliferation. These results show that the in vitro 3-D cell culture model may mimic
carcinogenesis and tumor progression in vivo. This study provides basis for exploring the
potential for identifying agents which may normalize acinar formation, and thus, control

cancer development.

Introduction

Homeostasis in the prostate is maintained by an interplay of multiple factors which
directly or indirectly regulate cell growth and cell death. Factors involved in this regulation
of homeostasis include, laminin and its integrin receptors, hormones, and growth factors.
Prostate cancer arises when the balance between the cell proliferation and cell death is
altered in favor of proliferation (Sensibar, 1995). Prostatic intraepithelial neoplasia (PIN)
is considered to be the ﬁrst step in prostate carcinogenesis. Late PIN is considered to be
equivalent to carcinoma in situ (Webber et al., 1996). Established primary tumors are
heterogeneous and consist of invasive and non-invasive cell populations. The next step in
prostate cancer development is invasion of the basement membrane by cancer cells and
subsequent metastasis to secondary sites, which is the primary cause of death from prostate

cancer (Plug and Kopf-Maier, 1995, Webber et al., 1996).

127

Recently, it has been recognized that any scientiﬁc study of prostate cancer
necessitates a knowledge of morphogenesis ( Bonkhoff and Remberger, 1998, Webber et
a1, 1997). Therefore, the overall aim of this chapter is to describe a 3—D in vitro cell model
with which to study mechanisms involved in acinar morphogenesis and its loss in the
progression from normal epithelium, to PIN and then to invasive carcinoma. Such a model
system requires well characterized cells line that represent different stages in the multi-step
process of carcinogenesis. Such cells lines have been develOped and fully characterized in
Dr. Webber's laboratory. Cell lines, which represent different stages of carcinogenesis,
were developed by transfection or exposure of RWPE-l cells to a carcinogen. RWPE-l
cells were transformed by the addition of a Ki-ras oncogene to produce the tumorigenic
RWPE-Z cell line (Bello et al. , 1997, Rhim, 1994). Other RWPE-l cells were treated with
the chemical carcinogen, N-nitroso-N-methlyurea (NMU) , to establish cell lines of varying
tumorigenicity including: WPEl-NA22, WPEl-NB27, WPEl-NB14, WPEl-NBll and
WPEl-NBZ6 (Webber, personal communication). Collectively, these cells lines represent
different stages of neoplastic transformation from low tumorigenic to highly tumorigenic
cells.

Late stages of PIN and prostate cancer are characterized by a loss of normal cell
organization, cellular polarization and cellzcell and cell:basement membrane adhesion,
which may involve alterations in integrin receptors expression (Webber et al., 1997).
Proper expression of a6B1 integrin and the presence of laminin are necessary for normal
acinar morphogenesis (see Chapter 5). Alterations in integrin expression play a role in the
course of invasion where tumor cells must ﬁrst attach to the basement membrane before

degrading it and then detach and move into the ECM to intravasate into nearby blood

128

vessels to travel and metastasize to a secondary site.

Growth factors enable cells to maintain local homeostasis and adjust to their
biological environment. Epidermal growth factor (EGF) is involved in the regulation of
growth of the normal human prostatic epithelium (Bello et a1. , 1997, Webber at al. , 1996).
During malignant transformation, initiated cells may accomplish selective growth by
autocrine over-expression of EGF (J arrard et al. , 1994, Nakamoto, 1992), which has been
implicated in neoplastic transformation in numerous types of cancers such as gastric,
esophageal, breast and prostate cancers (Jarrard et al., 1994).

Studies reported in this chapter will examine the relationship between the degree
of loss of the ability to undergo acinar morphogenesis by cancer cells and their invasive
ability. Further, possible relationships with «6131 integrin expression as well as response

of various cells to exogenous EGF will be examined.

Results

Loss of acinar morphogenesis in 3-D cultures by malignant cell lines

WPEl-NA22, WPEl-NBZ7, WPEl-NBI4, RWPE-Z, WPEl-NBll and
WPEl-NBZ6 cell lines represent different stages of neoplastic transformation and tumor
progression from low—tumorigenic to highly tumorigenic cells. The RWPE—l cell line
served as the positive control, while, the DU-145 human prostate cancer cell line, which
originated from a brain metastasis, served as the negative control. All cell lines were
grown on the reconstituted basement membrane, Matrigel, to determine their ability to
undergo acinar morphogenesis. Within 24 h after plating, cells attached and all, except

DU-l45 cells, formed cytoplasmic connections with neighboring cells. Only DU-145 cells

129

formed a monolayer. At four days, cultures were ﬁxed, stained with propidium iodide and
counts of acini were performed at low magniﬁcation using laser scanning confocal
microscopy. Acini-forming ability of the different cell lines was determined by counting
individual acini using the LSM transmitted light at low power. Marked differences between
the cell lines were observed. The most notable difference was observed between the non-
tumorigenic RWPE-l cells and the tumorigenic WPEl-NB26 and DU-145 cells, where
the highly malignant WPEl-NB26 and DU- 145 cells did not organize at all into acini (0%)
(Figure 61), while the non-tumorigenic RWPE-l cells used as the positive control, formed
acini at a high frequency, and this number was set at 100%. The highest frequency of acini
formation amongst the tumorigenic cells lines was seen in WPEl -N A22 and WPEl-NB27
cells with 69% and 51% of control, respectively. Lower acini formation was seen in
WPEl-NBl4 at 21%, RWPE-2 at 10% and WPEl-NBll at 9% of the RWPE-l control.
These data demonstrates that as cells increase in tumorigenicity, their ability to polarize
and undergo acinar morphogenesis is decreased.

Acinar morphology was examined by optically sectioning propidium iodide-stained
acini using a laser scanning confocal microscope. RWPE-l cells formed acini consisting
of polarized cells forming a distinct central lumen (Figure 6-2 a z). WPEl-NA22 cells
formed acini morphologically comparable to those formed by RWPE-l (Figure 6-2 b).
WPEl-NA22 acini consisted of polarized epithelial cells forming a distinct central lumen
with an average area of ~4000 umz. WPEl-NB27 cells also formed small acini with well
polarized cells and a central lumen (Figure 6-2 c 1'). While mitotic cells were occasionally
observed in RWPE-l and WPEl-NA22 acini, WPEl-NB27 acinar cells tended to be

undergoing mitosis at a higher rate than cells in RWPE-l or WPEl-NA22 acini

130

§

 

A
8
1
_1

 

AL

Acini
(Percent of Control)

 

8

 

B

 

 

 

 

 

 

 

 

 

 

 

 

 

, s— ._ n

m1 WEI-A22 ”El-m7 WEI-14 m2 WEI-B11 WEI-m N145

Figure 6-1. Acinar Formation of immortalized RWPE-l cell line and malignant prostate epithelial
cell lines. Cells were grown on Matrigel for 4 days. Acini-forming ability was determined by
staining cultures with propidium iodide and counting individual acini using the LSM transmitted
light at low power. Counts are expressed as percent of the RWPE-l which was set at 100%
(Bars = +/-SE).

131

 

Figure 6-2. Acinar formation by immortalized non-tumorigenic and tumorigenic prostatic
epithelial cell lines grown in 3-D Matrigel cultures. Optical confocal sections of
propidium iodide-stained cells at 4 days of growth. (at) RWPE-l (positive control) and
(b) WPEl-NA22 cells form acini with well polarized cells and a distinct lumen. (31'!) Cell
which escape into the lumen undergo apoptosis. (c)WPE1-NBZ7 cells polarize and form
some small acini (i), but as the acini increase in size, the lumen become obstructed by
cells growing into it (ii and iii). (d) WPEl-NBl4 cells form half acini attached to large
masses of cells (di) and psuedo-acini structures that appear to have lumens but lack
polarized cells (d it). (e) WPEl-NBll cells form acini which get “plugged” by cells
migrating to the center of the lumen (er) or structures in a “ribbon” formation consisting
of sheets of polarized cells folded over themselves (eii). (t) RWPE—Z cells form small
amorphous balls who’s cells are undergoing apoptosis. (g) WPEl—NBZ6 cells do not form
acini but organize into large solid cylinder-like structures and amorphous balls. (h) Du-
l45 cells, which served as a negative control, form large amorphous balls (Bar 25p).

[32

(Figure 6-2 c ii). Additionally, WPEl-NB27 cells, that migrated into the lumen, did not
readily undergo apoptosis (Figure 6-2 c iii), however, those in RWPE-l acini did so
(Figure 6-2 a it). WPEl-NB14 cells attached and polarized but tended to form only half
acini which were attached to large masses of cells (Figure 6-2 di). Additionally,

WPEl—NBl4 cells tended to form psuedo-acini (Figure 6—2 d it), structures that appeared
to have lumens but lacked polarized cells (Figure 62 e) . WPEl-NBll cells did attach and
polarize but few acini with lumens were evident. Acini formed tended to get "plugged” by
cells migrating to the center of the lumen (Figure 6-2 ei). Although some migrated cells
were undergoing apoptosis, most were not. The predominant type of structures observed
in WPEl-NBll cultures were "ribbon" formation consisting of sheets of polarized cells
folded over themselves (Figure 62 e ii). Few RWPE—2 cells attached to the Matrigel and,
those that did, tended to undergo apoptosis immediately after polarizing and beginning to
form acini (Figure 6-2 t). WPE-NB26 cells did not form any acini but did organized into
large solid cylinder structures and solid balls (Figure 6-2 g). DU-145 cells also did not
form acini, they formed large solid balls with no apparent organization (Figure 6-2 h).
These results show that as cells increase in tumorigenicity, their ability to form

morphologically distinct, well polarized, acini deteriorates.

The invasive ability shows an inverse relationship with the ability to undergo acinar
morphogenesis

The invasive ability of RWPE-l cells through Matrigel, using the modiﬁed Boyden
chamber in vitro invasion assay, was compared with that of WPEl-NA22, WPEl-NB27,

WPE] ~NB14, RWPE-Z, WPE 1 -NBl 1 and WPE] -NB26 cells, where DU-145 invasion was

133

taken as 100%. In comparison to DU-145 cells, the RWPE-l cells were not invasive,
while the tumorigenic cells lines displayed varying degree of invasion (Figure 6—3). A
comparison of the acinar-forming and invasive ability shows a striking inverse relationship
(Figure 6-4), that is, as cells becomes more invasive and tumorigenic, they lose the ability

to undergo acinar morphogenesis.

Alterations in integrin (16131 expression in malignant cell lines

Cell lines representing varying degree of malignancy were selected for an
examination of (16131 integrin expression and include RWPE-l, WPEl-NA22,

WPEl -NB27, WPE] -NBZ6. RWPE—l cells represent a non-malignant stage. WPEl-NA22
and WPEl-NB27 cells represent low-tumorigenic and low-invasive stage, while
WPE-N826 cells represent a highly tumorigenic, highly invasive stage. a6 or [31 integrin
expression was visualized by optical sectioning using a laser scanning confocal
microscope.

RWPE-l acini show strong positive staining for 016 at the basal end of the acinar
cells as compared to the IgG control (Figures 6-5). WPEl—NA22 also showed strong basal
expression of a6 integrin (Figure 6—5). WPEl—NB27 acini and WPEl-NBZ6 ball structures
did not stain for a6 expression (Figure 6-5). Strong Bl integrin basal expression was
observed in RWPE-l acini (Figure 6-5). WPEl-NA22 also showed strong basal expression
of Bl integrin (Figure 6—5). However, WPEl-NB27 acini showed positive, but weaker,
staining for BI at the basal end of the acinar cells (Figure 6-5) as compared to the IgG
control (Figures 5-8). The ball structures formed by WPEl-NB26 cells showed strong
positive staining for [31 integrin but the expression was not basal but was diffuse

134

RW PE-1

 

WPE1-NA22 —
WPE1-NBZ1 m
WPE1-NB14 _—-c
WPE-NB" ——-
RwPE-z _
wm-neze ———.
ou-us —

0 20 4O 60 80 1 00 1 20

Percent Invasion

Figure 6-3. Invasive ability of non—tumorigenic immortalized and malignant human
prostatic epithelial cells was compared using a modiﬁed Boyden chamber assay. Cells
were plated on Matrigel-coated ﬁlters at 400,000 cells per chamber and were allowed
to invade for 24 hours. Invasion by DU—14S cells was taken as 100% (Bars: +/-SE).

% Invasive Ability

%Aclni Formation

 

Figure 6-4. (a) Relationship between acini-forming ability and invasive ability. As invasive ability
increases, the ability to undergo acinar morphogenesis decreases. (b) Optical confocal sections
of propidium iodide-stained cells grown on Matrigel for 4 days. Sections of acini correspond to
cell lines shown in the graph above (a).

135

   

WPE1-NBZ6

Figure 6-5. (1601 laminin integrin expression. Non-malignant and malignant
prostatic epithelial cells, grown on Matrigel, were ﬁxed at 4 days and stained and
examined by confocal microscopy. for (16 or Bl integrin expression by
immunoﬂuorescence using monoclonal antibody. Non-malignant RWPE-l cells show
strong basal expression of 0.6 and BI. WPEl-NA22 cells. representing an early stage
of malignance, show basal expression of a6 and BI. WPEl—NB27 cells, representing
a later stage of malignancy than WPEl-NA22. lack (x6 expression but show basal
expression of Bl. WPEl-NBZ6 cells, representing a late invasive stage of
malignancy, lack a6 expression and show a diffuse cell membrane expression of
Bl(Bar 25 u).
136

throughout the entire cell membrane (Figures 6-5). These results demonstrate that while
0L6 integrin expression is decreased or is lost in the malignant transformation in
WPEl-NB27 and WPEl-NB26 cells, the BI expression is altered so that it is no longer

conﬁned to the basal cell surface but is diffusely expressed throughout the cell membrane.

Deregulation of growth by EGF inhibits acinar morphogenesis

In the prostate EGF is a positive regulator of prostatic epithelial cell growth and
its expression is known to be elevated in prostate cancer cells. Because the prostate is a
hormone regulated organ, the effects of EGF and its modulation by androgen were
examined in RWPE-l cells. Cells were treated with or without 5 nM of a non-metabolizing
androgen, mibolerone, in the presence of EGF ranging from 0 to 10 ng/ml. EGF alone
caused a maximum growth stimulation at 10 ng/ml of 162% of untreated control which
was set at 100% (Figure 6-6). In combination with 5 nM mibolerone, 10ng/ml of EGF
caused a growth increase to 224% of control (Figure 6-6). These data demonstrate that
exogenous EGF causes an increase in cell growth and that this growth is further enhanced
in the presence of androgen.

To examine the effects of this growth deregulation by EGF in acinar
morphogenesis, RWPE-l cells were treated with EGF at .01, 0.1, 1 and 10 ng/ml
concentrations. Cultures were ﬁxed after four days of treatment. Acini-forming ability for
each treatment was determined by staining cultures with propidium iodide and counting
individual acini using the LSM transmitted light at low power. EGF caused a dose-
dependent decrease in the acinar-forming ability of RWPE—l cells when compared to the

non-treated vehicle control set at 100%. At 0.1 EGF ng/ ml, acini formation was decreased

137

240

      

—a—No Mibolerone

220
+5 nM Mibolerone
200
180

160

Cell Growth
(Percent of Control)

140 —+
120 i
100

so - a» “1 ~ +~ ..W __1,,_.,,_ .‘ _. z.
0 1.25 2.5 5 10
EGF (nglml)

Figure 6-6. Effects of EGF on the growth RWPE-l cells and its modulation by the synthetic
androgen Mibolerone. 10,000 cells/well were plated in 96 well plates. Cells were treated
with EGF in the absence and presence of 5 nM Mibolerone for 5 days. RWPE-l cells
showed a dose-dependent growth response to EGF which was accentuated in the presence
of Mibolerone (Bars: +/-SE).

138

to 67% of control and a maximum inhibition of 56% was seen at 10 ng/ ml of EGF (Figure
6-7). These results show that exogenous EGF reduces the ability of RWPE—l cells to
undergo acinar morphogenesis.

If the decrease in acinar formation is due to a deregulation, which favors cell
proliferation, then blocking EGF activity should induce acinar formation. This was tested
using anti-functional (neutralizing) antibody directed against EGF. The anti-EGF Ab binds
to the EGF, which is a component of the culture medium, thereby preventing its binding
to the cellular EGF receptors and, thus, inhibiting a growth response. The neutralizing
EGF antibody, diluted 1:63 and 1:14, was added to the cell suspension and allowed to
incubate for 1h before cells were plated onto Matrigel-coated wells. Anti-EGF Ab
increased acinar formation to 114% and 120%, respectively as compared to the IgG
control which was set at 100% (Figure 6-8). These results show that inhibition of cell

proliferation by neutralization of EGF activity enhances acinar morphogenesis.

Discussion

Any scientiﬁc study of prostate cancer necessitate a knowledge of morphogenesis.
The study of prostate cancer requires the analysis of the various factors which control
normal acinar morphogenesis and their alteration in the process of carcinogenesis.
Currently, little is known about the mechanisms that control normal acinar morphogenesis
or alterations which occur in the pathogenesis of prostate cancer (Bonkhoff and
Remberger, 1998, Webber et al, 1997). This is due, in part, to the lack of appropriate in
vitro cell model systems which mimic the process of prostatic carcinogenesis and still

retain the 3-D micro-environment found in vivo. To be useful as a consistently

139

120
100 4

 

 

 

 

 

T:

:2:

8 80':
E.— T
2 O 60

H

C .

Q +

0 40

a E

$ 20~

o e . +

 

 

 

 

 

 

 

 

 

 

 

0 0.01 0.1 1 10
EGF Concentration (nglml)

Figure 6-7. Reduction of acinar formation by EGF. RWPE-l cells were grown
on Matrigel and treated for 4 days with EGF. Acini-forming ability was
determined by staining cultures with propidium iodide and counting individual
acini using the LSM transmitted light at low power. Counts are expressed as a
percent of the non-treated control which was set at 100%(Bars = +/-SE).

125 ~
120 i
115
110
105
100 -i
95

90 . ‘ ‘ ; - h ;
0 8 36
EGF Ab Concentration (nglml)

 

 

*1 .__- _i.. _ ___ .1

 

Aclnl
(Percent of Control)

 

 

 

 

 

 

,_
1

 

Figure 6-8. Enhancement of acinar formation by EGF neutralizing antibody.
RWPE-l cells were grown on Matrigel and pretreated with EGF anti-
functional antibody. Acini-forming ability was determined by staining cultures
with propidium iodide and cormting individual acini using the LSM transmitted
light at low power. Counts are expressed as a percent of the IgG control which
was set at 100% .

140

reproducible cell system, such model systems must consist of cell lines which are
representative of different stages of malignancy and do not undergo senescence. Ideally,
these cells line would originate from a common non-neoplastic established parent cell line
which could also be used for comparison. Herein is described the use of such immortalized
cells lines, in a 3-D in vitro cell model, which together with their parent RWPE-l cell line,
represent different stages in the progression from normal epithelium to prostatic
intraepithelial neoplasia (PIN) and then to invasive carcinoma. RWPE—l cells were
transformed by the addition of a Ki-ras oncogene to produce the tumorigenic RWPE-2 cell
line (Rhim et a1. , 1994 and Bello et al. , 1997). Further, RWPE-l cells were treated with
the chemical carcinogen, N-nitroso-N-methlyurea (NMU) (Webber, personal
communication) and injected into nude mice. The WPEl-NA22, WPEl-NBZ7, WPEl-
NB14, WPEl-NBll and WPEl-NBZ6 cell lines were established form tumors in nude
mice which were subsequently cloned in agar. Unlike the parental RWPE-l cell line, they
show varying degree of tumorigenicity and anchorage independent growth. Additionally,
these cell lines show altered phenotypes, in their growth rates and response to growth
factors and androgen, in comparison to each other and to the parent cell line. They have
retained the ability to express PSA and androgen receptor in response to androgen
stimulation (Bello et al., 1997, Webber, personal communication). This chapter has
focused on the loss of ability of malignant cells to undergo acinar morphogenesis, and
changes in their integrin expression as well as the relationship of these characteristics to
their invasive ability. Additionally, the role of epidermal growth factor in acinar formation

was also examined.

14]

Homeostasis in the prostate is maintained through an interplay of multiple factors
which directly or indirectly regulate cell growth and cell death. Prostate cancer arises when
the balance between the cell proliferation and cell death is altered in favor of proliferation
(Sensibar, 1995). Carcinogenesis in the prostate involves the presence of PIN, carcinoma
in situ and invasive carcinoma. During progression to an invasive malignant state, cancer
cells are said to de-differentiate and lose normal cell organization, cell polarity, cell:cell
and cell:ECM adhesion. It is believed that this happens in a stepwise manner in which cells
progress from an immortalized to low tumorigenic to a highly tumorigenic, invasive
phenotype (Webber et al., 1997). This phenomenon was examined in prostate epithelial
cells using the RWPE—l, WPEl-NA22, WPEl-NB27, WPEl-NBl4, RWPE-Z, WPEl-
NBll and WPEl-NBZ6 cell lines.

Results show a loss in acinar-forming ability with increasing malignancy (Figure
6-1). Similar observations have been made in the breast when normal and malignant cells
were induced to form acini (Petersen et al. , 1992, Howlett et al. , 1995). When plated on
the reconstituted basement membrane, Matrigel, non-malignant RWPE-l prostatic
epithelial cells underwent acinar morphogenesis forming distinct acini with well polarized
cells and lumens (Figure 6—2 a). The low-tumorigenic WPEl-NA22 cells also formed
distinct acini with well polarized cells and lumens, which are comparable to those formed
by RWPE-l cells, but at a lower frequency of 69% as compared to the RWPE-l controls
(Figure 6—1 and 6-2 a and b, respectively). WPEl-NA22 seem to represent a stage in PIN,
in that they retain the ability to form acini, have a growth rate similar to RWPE-l, have
a very low invasive ability and form very small tumors. WPEl-NBZ7 cells, with a reduced

acini-forming ability of 51%, formed small acini with well polarized cells and a central

142

 

lumen (Figure 6-1 and 6-2 ci). Although mitotic cells were occasionally observed in
RWPE—l and WPEl-NA22 acini, WPEl-NB27 acinar cells tended to be undergoing
mitosis at higher rate suggesting a growth deregulation in favor of cell proliferation
(Figure 6-2 c ii). This is in agreement with WPEl-NB27 growth rate which show a
doubling time twice that of RWPE-l and WPEl-NA22 (Webber, personal
communication). Additionally, cells that migrated into the lumen did not readily undergo
apoptosis upon loss of attachment to the matrix as is seen in RWPE-l cells (Figure 6—1
c iii and a ii, respectively). Both WPEl-NB14 and WPEl-NBII cells appear to polarize
but are unable to properly regulated their cell proliferation, thus, forming fewer acini, at
21% and 10% of control, respectively. The acini that do form consist of polarized cells
with half lumens or lumens that tended to "get plugged" by cells migrating into it
(Figure 6—2 di and er). Only a few RWPE—2 cells attach to the Matrigel and, those that do,
tend to undergo apoptosis immediately after attachment (Figure 6-2 1‘). WPE-N326 cells
did not form acini but did organized into large solid cylinder structures and solid balls of
cells similar to those formed by DU-l45 (Figure 61 g and h, respectively). These results
suggest that interactions between prostate epithelial cells and the ECM microenvironment
are important for regulating normal acinar morphogenesis and that these interactions are
disturbed in the process of tumorigenesis.

The results from the in vitro invasion assay demonstrate that the invasive ability of
the malignant cell lines shows an inverse relationship with their acinar—forming ability
(Figure 6-7). This suggests that the factors which promote acinar differentiation inhibit
invasion and those which favor invasion deny the cells the ability to undergo

differentiation. Taken together, the preceding results, thus, allow the ranking of the cell

143

 

lines in order of their increasing malignancy: WPEl-NA22, WPEl-NB27, WPEl-NBI4,
RWPE-Z, WPEl-NBll to WPEl-NBZ6 cell lines (see order in Figure 6-4). The non-
tumorigenic, non-invasive immortalized RWPE—l cell line represents a non—malignant cell
type.

Extracellular matrixzcell interactions are mediated by integrin receptors. As
demonstrated in chapter 5, the laminin component of the basement membrane and its key
laminin integrin receptor, a601, mediate the ability of prostatic epithelial cells to polarize
and undergo acinar morphogenesis. Alterations in the expression of laminin integrins has
been observed in a majority of prostate and breast cancers (Howlett et al., 1995, Knox et
al., 1994, Rabinovitz et al., 1995, Wewer et al., 1997). Because of the importance of
laminin in epithelial polarization and acinar morphogenesis, alterations in Q6131 integrin
expression were examined. The non-malignant RWPE-l cell line formed acini that
demonstrated strong basal expression of the a6Bl integrin receptor (Figure 6-5). The
WPEl-NA22 cell line, which has low tumorigenecity and low invasive ability, also
showed basal expression of the (16131 integrin receptor (Figure 6—5). This is in agreement
with the observation that normal prostatic epithelial cells and PIN show basal expression
of (16131 laminin receptors in viva (Cress et al., 1995 , Knox et al., 1994, Rabinovitz et
al. , 1995). While WPE1-27 cells express Bl, they do not express the a6 integrin receptor.
This may suggest a weaker attachment of cells to the basement membrane. The highly
invasive and the most malignant WPE-26 cells lack (16 integrin expression. They show 131
expression but it is not restricted to the basal cell surface and is diffuse throughout the cell
membrane. This diffuse Bl staining pattern was also observed in the malignant and highly
invasive Du-145 (data not shown). This is in agreement with ﬁndings of other investigators

144

who have observed a diffuse Bl integrin expression pattern in invasive prostate and breast
cancer cells in vivo and in. vitro and have suggested that it may play a role in invasion
(Knox et al., 1994, Rabinovitz et al., 1995, Wewer et al., 1997).

However, the lack of a6 integrin expression in WPEl-NB26 and WPEl-NB27 is
in conﬂict with accounts where an increase in (16 expression in prostate cancer cells in vivo
and in established in vitro prostate cancer cell lines was observed (Knox et al., 1994,
Rabinovitz et al., 1995, Wewer et al., 1997). To explain this difference, I suggest that
integrins may be expressed in a transient manner in prostate carcinogenesis, similar to their
expression in the developing prostate. Perhaps WPEl-NBZ7 and WPEl-NB26 cells
represent two different stages in tumor progression in which the expression of a6 is turned
off. At the WPEl-NBZ7 stage, cells lack a6 expression but retain a normal basal
expression of B1, suggesting a weak attachment to the basement membrane. This would
ﬁt with their low acinar-forming ability and the slightly higher invasive ability as
compared to the 0(6B1 integrin expressing RWPE-l and WPEl-NAZZ cell lines.
WPEl-NB26 might represent a more advanced stage, when the cells are most motile and
the lack of a6 integrin expression maintains the cells detached from the ECM. This, in
combination with the diffuse expression of B1, would facilitate invasion. This is supported
by the observation that WPEl-NB26 cells are the most invasive amongst this family of
malignant cells lines. Additionally, matrix studies with normal and malignant breast cancer
cells show a diffuse staining for B1 and absence of a6 integrin staining in malignant cells,
as compared to normal breast cells (Howlett et al., 1995).

In analyzing the morphology of the acini formed by the malignant cell lines and

their integrin expression proﬁle, 1 observed that the loss of the ability to undergo acinar

145

formation could be divided into two phases. In the ﬁrst phase, cells lose the ability to
control their growth but retain the ability to attach and polarize, as is seen in WPEl-NA22
and WPEl—NB27. In the second phase, in addition to the loss of cell growth, cells also
lose proper integrin expression. In this phase, cells attach but do not polarize, instead, they
grow in masses which increase in size according to their growth rate as seen by RWPE-Z,
WPEl-NB26 and DU-145 cells.

Perhaps the decrease in acinar-forming ability seen in the WPEl-NA22 and
WPEl-NB27 cells lines, which retain some degree of acinar-forming ability, is due to the
deregulation of growth observed in malignant cells. The autocrine over-expression of
growth factors, such as EGF, has been implicated in neoplastic transformation (Jarrard et
al, 1994). In primary cultures of normal prostate cells exogenous EGF causes an increase
in cell proliferation (McKeehan et al. , 1984). In prostate cancer cell lines, such as PC-3
and DU-145, exogenous EGF causes little to no effect on cell proliferation. The lack of
growth response to EGF may be due to the autocrine production of EGF and, therefore,
the cells are not as sensitive to exogenous EGF. EGF dose response studies show that
WPEI -NA22, WPEl -NB27 are less sensitive the exogenous EGF treatment than RWPE-l
cells (Webber, personal communication). This, along with the fact that their doubling time
is twice that of RWPE-l , supports the idea that the decreased acinar-forming ability is due
to a deregulation of growth in favor of cell proliferation.

In order to determine if a deregulation of growth alone could alter acinar
formation, the role of exogenous EGF on RWPE-l cell growth and acinar-forming ability
was examined. Results show that exogenous EGF causes an increase in cell growth (Figure

6-6). Intentional deregulation of RWPE-l cell growth using exogenous EGF, signiﬁcantly

146

reduced the acini-forming ability of RWPE-l cells when plated on Matrigel (Figure 6-7).
Conversely, if EGF activity was blocked using neutralizing antibody, acinar formation was
signiﬁcantly enhanced. Furthermore, the increase in RWPE-l cell proliferation, in
response to EGF, is accentuated by the presence of androgen (Figure 6—6). Androgen is
known to causes an up-regulation of the EGF receptor (Carpenter and Wahl, 1990),
thereby increasing the sensitivity of cells to EGF . Since the development of prostate cancer
is dependant on the presence of androgen, the regulation of EGF responsiveness by
androgen presents one manner by which PIN cells are able to escape the conﬁnes of acini
and expand clonally.

In summary, the data presented in this chapter demonstrate the use of an in vitro
3-D cell culture system that mimics different stages in prostate carcinogenesis represented
by PIN in vivo. Using this system it was demonstrated that malignant cells lose the ability
to undergo acinar differentiation and that the degree of loss is correlated with their degree
of malignancy. The loss of the ability to form acini showed an inverse relationship to the
invasive ability of the cells. Additionally, data show that in the progression through the
multi-step process of carcinogenesis, alterations in laminin integrin 0(6B1 expression and
deregulation of growth, involving alteration in growth factors expression and
responsiveness, occur early on in the process. It is hoped that further studies, utilizing the
3-D cell culture model system, will shed light on the role of integrin receptors and other
factors in the progression from normal to malignant prostatic epithelial cells. This

information would facilitate the development of drugs for cancer prevention and treatment.

147

References

Bello D, Webber MM, Kleinman HK, Wartinger DD and Rhim J S: Androgen responsive
adult human prostatic epithelial cell lines immortalized by human papillomavirus

18. Carcinogenesis 18: 1212-1223, 1997.

Bonkhoff H and Remberger K: Morphogenetic concepts of normal and abnormal growth

in the human prostate. Virchows Arch. 433: 195-202, 1998.

Carpenter G and Wahl MI: The epidermal growth factor family. In: Peptide growth
factors and their receptors 1. Sporn, MB. and Roberts, A.B. (eds), New York,

Springer-Verlag, pp. 69-171, 1990.

Cress AE, Rabinovitz I, Zhu W and Nagle RB: The alpha 6 beta 1 and alpha 6 beta 4
integrins in human prostate cancer progression. Cancer Metastasis Rev: 14:219-28,

1995.

Flug M and Kopf-Maier P: The basement membrane and its involvement in carcinoma cell

invasion. Acta. Anat. Basel. 152: 69-84, 1995.

Howlett AR, Bailey N, Damsky C, Petersen OW and Bissell MJ: Cellular growth and

survival are mediated by beta 1 integrins in normal human breast epithelium but

not in breast carcinoma. J. Cell. Sci. 108: 1945-57, 1995.

148

Jarrard, D.F., Blitz, B.F., Smith, R.C., Patai, B. L. and Rukstalis. Effects of epidermal
growth factors on prostate cancer cell line PC3 growth and invasion.

Prostate, 24: 46—53, 1994.

Knox JD, Cress AE, Clark V, Manriquez L, Afﬁnito KS, Dalkin Bland Nagle RB:
Differential expression of extracellular matrix molecules and the alpha 6-integrins

in the normal and neoplastic prostate. Am. J. Pathol. 145:167-74, 1994.

McKeehan, W. L., Adams, PS. and Rosser, M.P. Direct mitogenic effects of insulin,
epidermal growth factor, glucocorticoid, cholera toxin, unknown pituitary factors
and possibly prolactin but not androgen, on normal rat prostate cells in serum-free,

primary cell cultures. Cancer Res., 44: 1998-2010, 1984.

Nakamoto, T., Chang, C.S., Li, A.K. and Chodak, G.W. Basic ﬁbroblast growth

factor in human prostate cancer cell lines. Cancer Res., 52: 571-577, 1992.

Petersen OW, Ronnov-Jessen L, Howlett AR and Bissell MJ: Interaction with basement
membrane serves to rapidly distinguish growth and differentiation pattern of
normal and malignant human breast epithelial cells [published erratum appears in

Proc. Natl. Acad. Sci. USA. 89: 9064-8, 1992.

149

Rabinovitz I, Nagle RB and Cress AE: Integrin alpha 6 expression in human prostate
carcinoma cells is associated with a migratory and invasive phenotype in vitro and

in vivo. Clin. Exp. Metastasis. 13: 481-91, 1995.

Rhim JS, Webber MM, Bello D, Lee MS, Amstein P, Chen L and Jay G: Stepwise
immortalization and transformation of adult human prostate epithelial cells by a
combination of HPV18 and v-Ki-ras. Proc. Nat. Acad. Sci. USA, 91:11874-

11878, 1994.

Sensibar JA: Analysis of cell death and cell proliferation in embryonic stages, normal
adult, and aging prostates in human and animals. Microsc. Res. Tech. 30: 342-50,

1995.

Webber MM, Bello D, Kleinman HK and Hoffman MP: Acinar differentiation by non-

malignant immortalized human prostatic epithelial cells and its loss by malignant

cells. Carcinogenesis 18:1225-1231, 1997.

Webber, MM, Bello D, and Quader S: Immortalized and tumorigenic adult human
prostatic epithelial cell lines. Characteristics and applications. Part. 1. Cell markers

and immortalized, non-tumorigenic cell lines. Prostate 29:386— 394, 1996.

150

Webber MM, Waghray A, Bello D and Rhim JS: Proteases and invasion in human
prostate epithelial cell lines: Implications in prostate cancer prevention and

intervention. Radiat. Oncol. Invest. 3:358-362, 1996.

Wewer UM, Shaw LM, Albrechtsen R and Mercurio AM: The integrin alpha 6 beta 1

promotes the survival of metastatic human breast carcinoma cells in mice. Am. J.

Pathol. 151: 1191-8, 1997.

151

Chapter Seven

Regulation of Acinar Morphogenesis by
TGF-B in Non-malignant and Malignant
Human Prostatic Epithelial Cells

152

Abstract

An important event in the transformation of normal cells to malignant cells is the
loss of normal growth regulation. This occurs as a result of a loss of balance between
stimulatory and inhibitory growth factors. Studies described in Chapter 6 showed that
excessive stimulation with EGF can cause a decrease in the ability of cells to undergo
acinar morphogenesis. This may be one factor responsible for the loss of ability of cancer
cells to polarize and forma acini. In the present study, the role of the inhibitory growth
factor TGF-B and its receptors in acinar morphogenesis in non-malignant RWPE-l and
progressively more malignant WPE] -NA22 and WPEl -N B26 cells was examined. Results
show that TGF-B2 and 3 were more effective in inhibiting growth of RWPE—l cells that
TGF-B3. An examination of the expression of the three TGF-B isoforms by
immunostaining showed that they were expressed in RWPE-l cells in the following order:
TGF-B1 > B2 > B3. A strong expression of TGF-B R11 and TGF-B RIII was also observed.
Exposure of cells to anti-functional TGF-B1 and 2 antibody caused inhibition of cell
polarization and reduced acini formation to 42% and 46% of control, respectively,
indicating that TGF-B1 and B2 are necessary for cell polarization but TGF-B3 is not
essential. The importance of TGF-B function was further demonstrated by treating cells
with anti-functional antibody to TGF-B R11 and R111 receptors. Anti-TGF-B R11 and R111
antibodies decreased acini formation to 26 and 35% of control, respectively, thus,
validating the importance of TGF-B binding in acinar morphogenesis.

Exposure of WPEl-NA22 cells to TGF-B1, 2 and 3, in monolayer cultures,

reduced growth to 54, 60 and 70% of control, respectively. Similar treatment of

153

WPEl-NB26 cells reduced growth to 74, 74 and 85 % of control, respectively. In
comparison to RWPE-l and WPEl-NA22, WPEl-NB26 cells show reduced sensitivity to
TGF-B1. Further, WPEl-NA22 cells showed strong expression of TGF-B1 but that of B2
and B3 was not detected. They do express TGF-B1 R11 and R111 receptors and the addition
of exogenous TGF-B1 or TGF-B2 induced an increase in acini formation by 134 and 159%
of control. However, the more malignant WPEl-NB26 did not show any TGF-B1
expression by immunostaining and only weakly expressed TGF—B1 RII. Taken together,
these results demonstrate that TGF-B1 and B2 play a critical role in acinar morphogenesis,
that a stepwise loss of TGF-B and/or receptor expression is associated with increasing
malignant characteristic of prostate cancer cells and that exogenous TGF-B can restore the

ability, at least of cells at the lower end of malignancy, to undergo acinar morphogenesis.

154

Introduction

During the multi-step process of carcinogenesis, an important event is the loss of
normal growth regulation. This deregulation of growth is achieved, in part, by alteration
in growth factor expression and in the ability of cells to respond to growth factors.
Transforming growth factor-B and epidermal growth factor (EGF) are two major growth
factors involved in the regulation of the normal prostatic epithelium (Webber at al. 1996).
In normal prostate epithelial cells, TGF-B acts predominantly as a negative growth
regulator, while EGF acts as a positive growth regulator. Changes in the level of these
growth factors have been observed in malignant prostate epithelial cells (Jarrard et al,
1994, Nakamoto et al., 1992). One characteristic of progression from a normal to a
malignant phenotype is reflected in the increased rate of cell proliferation caused by the
autocrine secretion of mitogenic growth factors and the down-regulation or cessation of
inhibitory growth factor function (Culig, 1994, Nakamoto et al. , 1992,). Data in chapter
6 demonstrated that a deregulation of growth by exogenous EGF inhibits acinar
morphogenesis and that the neutralization of EGF activity enhanced acinar-formation and
differentiation. This chapter will examine the role of TGF-B in acinar morphogenesis and
alterations in TGF-B expression and response in cancer cells, using the 3-D cell culture
model.

TGF-Bs are members of a superfamily of polypeptides that regulate cell cycle
progression, differentiation and chemotaxis of many different types of cells. Their
biological effect on a given target cell is dependant on the cell type, the growth conditions

and the presence of other growth factors. Although a bifunctional response to TGF-B by

155

both ﬁbroblast and epithelial cells has been reported, TGF-B generally stimulates the
growth of ﬁbroblasts but inhibits the growth of epithelial and endothelial cells (Franzen
et. al., 1993, Roberts and Sporn, 1990, Webber et. al., 1996). In addition to inhibiting
epithelial cell growth, TGF- B stimulates the production of extracellular matrix components
and protease inhibitors and causes a decrease in the net activity of extracellular matrix-
degrading proteolytic enzymes (Franzen et al., 1993, Roberts and Spom, 1990,).
Presently, three form of TGF B (TGF-B1, TGF-B2 and TGF-B3) have been identiﬁed in
the prostate (Dahiya et a1, 1996).

Like most growth factors, TGF-B exerts its biological activity through a speciﬁc
transmembrane receptor. Thus far, three classes of TGF-B receptor subunits have been
recognized: TGF-B type 1 (RI), TGF-B type 11 (R11) and TGF-B type 111 (R111) (Mckeehan,
1991). R1 (60- 70 kDa) and R11 (85-110 kDa) are serine-threonine kinase receptors, while
RIII (100-140 kDa), with a large extracellular domain and a short cytoplasmic tail,
possesses no obvious signaling motif. R11] is the predominantly expressed receptor in
comparison to the RI and R11 receptors, and has high afﬁnity for TGF-B1 and B2
(Mckeehan, 1991, Moustakas et al., 1993). A functional TGF-B receptor is formed as a
heterodimer by a single disulﬁde bond between different receptor subunits. The
heterodimer is formed upon exposure to the ligand. In the presence of TGF-B, RII/ R111
and R1/ R11 form heterodimers but RIII/RI heterodimers form rarely (Moustakas etal.,
1993). The RI/RII complex binds TGF-B1 and B3 well but binds TGF-B2 poorly, while
the RII/RIII complex binds TGF-B2 as well as TGF-B1 with high afﬁnity. The RII/RIII

complex also binds TGF-B3. The human prostate epithelial cells express all three TGF-B

156

receptors (Guo etal., 1997).

Although TGF-B is expressed by both the urogenital mesenchyme and the
urogenital epithelium of the fetal mouse, the highest levels are detected in the mesenchyme
and not the epithelium. Its exact role in prostatic morphogenesis has remained unclear.
Conﬂicting reports indicate that TGF-B expression follows the path the epithelial cells take
as they penetrate into the surrounding mesenchyme, while others ﬁnd focal loss of TGF-B
expression at bud-forming regions (Nemeth, et al. 1997, Silberstein et a1, 1992, Timme
et al. 1994).

Because of its apparent important role in branching and acinar morphogenesis in
vivo, the present chapter will focus on the role of a TGF-B 1, 2 and 3, and R11 and R111
receptors in normal acinar morphogenesis by RWPE-l cells and its alteration in the PIN-
like WPEl-NA22 cells line and the malignant WPEl-NB26 cells. Additionally, the ability
of TGF- B isoforms to restore the ability to undergo acinar morphogenesis will also be

examined.

Results

Effects of TGF-B isoforms on the growth of RWPE-l cells

In the human prostate TGF- B is a negative regulator of normal prostatic epithelial
cell growth. In order to elucidate the individual roles of TGF-B1, TGF-B2, and TGF-B3
on cell growth, RWPE-l cells were treated with either TGF-B1, TGF-B2 or TGF-B3 ranging
from 0 to 10 ng/ml in the presence of 5 ng/ml of EGF. At 10 ng/ml, TGF-B1 caused a
maximum growth inhibition to 66% of untreated control which was set at 100%

(Figure 7-1). TGF-B2 caused a maximum growth inhibition to 76% of untreated control

157

120 r
. +TGF-beta1

l +TGF-bet82
11° f—s—rcF-betaa

1
I

 

 

.5 100
h
g E
O 90 _
2 2 '
0 O
u
= s: so
0
o 3
h
0
9., 70
so -
so 1-7 ___- —..—_.—— —~~ 1%.”, _____-_ __4

o 2.5 5 10
TGF-B (nglml)

Figure 7-1. Effects of TGF-Bl, TGF-B2 and TGF-B3 on the Growth of RWPE—l cell in
monolayer cultures. 10, 000 cells/well were plated in complete K-SFM medium in 96 well
plates in triplieate. Cells were treated with TGF-B1, TGF-B2 or TGF-B3 for 5 days. RWPE-l
cells showed a dose-dependent growth inhibition for all three cells line. Note that TGF-Bl was
the most effective growth inhibitor. (Bars = +/-SE)

158

at 10 ng/ ml. TGF-B3 show similar response as TGF-B2 with a maximum growth inhibition
to 80% of untreated control at 10 ng/ ml (Figure 7-1). These data demonstrate that while

all three forms elicited growth inhibition, TGF-B1 was most affective.

RWPE-l cells express TGF-B1, 2 and 3 and the TGF-B RI] and R111 receptors

The expression of TGF-B1, TGF-B2 or TGF-B3 in RWPE-l cells was examined
by indirect immunoﬂuorescence using a laser scanning microscope (LSM). Because the
TGF-B RII/ RIII receptor complex binds both TGF-B1 and TGF-B2 with high afﬁnity, and
also binds TGF-B3, the expression of R11 and R111 was examined. RWPE-l cells show
strong positive staining for TGF-B1 as compared to the IgG control (Figures 7-2). TGF-B2
expression was detected but staining was weaker in comparison to TGF-B1 (Figure 7-2).
Very weak staining was observed for TGF-B3 as compared TGF-Bl or 2 (Figure 7-2).
RWPE-l showed strong expression for both the TGF-B R11 and TGF-B RIII as compared
to the IgG control (Figures 7-3). These results demonstrate that TGF-B1 is the predominant
type expressed by RWPE-l cells and that TGF-B2 and TGF-B3 are expressed in decreasing
order. Strong expression of TGF-B R11 and TGF-B RIII receptors was also observed in

these cells.

TGF-B1 and TGF-B2 are necessary for acinar morphogenesis
In order to determine if TGF-B 1 , TGF-B2 or TGF-B3, expressed by RWPE-l cells,
play a critical role in acinar morphogenesis, anti-functional antibodies against TGF-B1,

TGF-B2 or TGF-B3 were employed. TGF-B1, TGF-B2 or TGF-B3 antibody, diluted 1:42,

159

“K"

- RWPE-l WPE’l-NA22 WPEl-NBZE

 

Figure 7-2. TGF-Bl, TGF—B2 and TGF-B3 ermression in non-malignant RWPE—l and malignant
human prostatic epithelial cells, WPEl-NA22 and WPEl-NBZé. Cells were grown on glass
coverslips, ﬁxed and examined for TGF-Bl , TGF-B2 and TGF-B3 expression by
immunofluorescence using monoclonal antibody. Staining was visualized by LSM. Non-
malignant RWPE-l cells show strong expression of TGF—Bl . but weaker expression of TGF-B2
and TGF-B3. WPEl-NA22 cells, representing an early stage of malignancy, show positive
expression of TGF-Bl, but lack expression of TGF-B2 and TGF-B3. WPEl-NBZ6 cells,
representing a later invasive stage of malignancy than WPEl-NA22, lack expression of TGF-
Bl, TGF-B2 and TGF-B3 (Bar 25 u).

160

- RWPE-1 WPEl—NA22WPE1-NB26

  
  

   
 

TGF-B RII

TGF-BRIII

IgG contrOI - .-

Figure 7-3. TGF-B R11 and TGF-B R111 receptor expression in non-malignant RWPE—l and
malignant prostatic epithelial cells, WPEl-NA22 and WPEl—NBZé. Cells were grown on
Matrigel, ﬁxed at 4 days and examined for TGF-B R11 and TGF-B R111 receptor expression
by immunoﬂuorescence using monoclonal antibody. Staining was visualized by LSM. Non-
malignant RWPE-l cells show strong expression of TGF-B R11 and TGF-B RIII. WPEl-
NA22 cells, representing an early stage of malignancy, show positive expression of TGF-B
RH and TGF-B RIII. WPEl-NBZ6 cells, representing a later invasive stage of malignancy than
WPEl-NA22, show faint expression of TGF—B R11 and lack TGF-B R111. expression

(Bar 25 u).

161

1:62 and 1:16, respectively, was added to the cell suspension and allowed to incubate for
1 h before cells were plated onto Matrigel-coated wells. Anti-TGF—Bl neutralizing Ab
decreased acinar formation to 42 % of the IgG control which was set at 100% (Figure 7-4).
Anti-TGF-B2 neutralizing Ab also decreased acinar formation to 46% of the IgG control
(Figure 7-4). However, anti-functional TGF-B3 Ab only caused a decrease to 80% of
control in acinar formation (Figure 7-4). These results indicate that TGF-B1 and TGF-B2,
but not TGF-B3, are essential for cell polarization and acinar morphogenesis.

If TGF-B function is necessary for acinar formation, then blocking the binding of
TGF-B to the cells should result in the inhibition of acinar formation. This was tested using
anti-functional antibody directed against TGF-B RII or TGF-B R111. The neutralizing
TGF-B RII or TGF-B RIII Ab, diluted 1:16, was added to the cell suspension and allowed
to incubate for 1 h before cells were plated onto Matrigel-coated wells. Anti-TGF-B RII
neutralizing Ab decreased acinar formation to 26% of the IgG control which was set at
100% (Figure 7-5). Anti-TGF-B RIII neutralizing Ab decreased acinar formation to 35 %
of the IgG control (Figure 7-5). These results validate the importance of functional TGF-B

binding in acinar morphogenesis.

Effects of TGF-B isoforms on the growth of WPEl-NA22 and WPEl-NBZ6 cells
Because a loss of autocrine TGF-B activity and/or sensitivity to exogenous TGF-B

appears to provide cells with a growth advantage in malignant progression, effects of

TGF-B 1, 2 and 3, on the growth of WPEl-NA22 and WPEl-NB26 cells, were examined.

WPEI -NA22 and WPEl-NBZ6 cells were treated with either TGF-Bl , TGF-B2 or TGF-B3

162

Percent of 196 Control

 

IgG Control anti-TGF-beta 1 anti-TGF-beta 2 anti-TGF-beta 3

Figure 7—4. Inhibition of acinar formation by anti-TGF—B neutralizing antibodies.
RWPE-l cells were pretreated with TGF-B1, TGF-B2 or TGF-B anti-functional
antibody before plating on Matrigel. (a) Acini-forming ability was determined by
staining cultures with propidium iodide and counting‘individual acini using the LSM
transmitted light at low power. Counts are expressed as percent of the IgG control which
was set at 100% (Bars = +/— SE). (b) Optical confocal section of a control acinus. (c)
Optical confocal section of TGF-B antibody—treated cells which form aggregates but lack
polarization and acinar formation (Bar 25 u).

120 -

100-

O
O
.

Aclnl
(Percent of Control)
8 8

N
O

 

   

196 Control antl-TGFbeta RII antl-TGFbeta Hll

Figure 7-5. Inhibition of acinar formation by anti-TGF-B receptor
neutralizing antibodies. RWPE-l cells were and pre-treated with
TGF-B R1] or TGF-B RIII anti-functional antibody before plating on
Matrigel. Acini-forrning ability was determined by staining cultures
with propidium iodide and counting individual acini using the LSM
transmitted light at low power. Counts are expressed as a percent of
the IgG control which was set at 100% (Bars: +/- SE).
163

at concentrations ranging from 0 to 10 ng/ml. In WPEl-NA22 cells, TGF-B1 caused a
maximum growth inhibition to 54% of untreated control at 10 ng/ml (Figure 7-6).
Untreated control was set at 100%. TGF-B2 caused a maximum growth inhibition to 60%
of untreated control at 10 ng/ml (Figure 7-6). TGF-B3 was the least effective at eliciting
an inhibitory response, with a initial growth inhibition to 70% of untreated control at 2.5
ng/ml but at 10 ng/ml concentration growth was inhibited only to 82% (Figure 7-6).

In WPEl-NB26 cells, TGF-B1 caused a maximum growth inhibition to 74% of
untreated control at 10 ng/ ml (Figure 7-7). TGF-B2 caused a maximum growth inhibition
at 10 ng/ml to 74% of untreated control (Figure 7-7). TGF-B3 was the least effective at
eliciting an inhibitory response, with a initial growth inhibition to 74 % of untreated control
at 2.5 ng/ml but at 10 ng/ ml growth was inhibited only to 85% (Figure 7-7). These data
demonstrate that while all three forms elicited a growth inhibition, TGF-Bl and TGF-B2
were most effective. In comparison to RWPE-I and WPEl-NA22, WPEl-NB26 cells

show reduced sensitivity to TGF-B1 (Figure 7-8).

Alteration in TGF-B1, 2 and 3 and TGF-B R11 and TGF-B RIII receptor expression
in malignant WPEl-NA22 and WPEl-NBZ6 cells

The escape from the negative growth control imposed by TGF-B may be due to
either down-regulation of growth factor expression or a decrease or loss of receptor
expression. Therefore, the expression of TGF-B1, TGF-B2, and TGF-B3 and TGF-B R11
and R111 in WPEl-NA22 and WPEl-NB26 cells was examined by indirect

immunoﬂuorescence using the LSM. WPEl-NA22 cells show strong positive staining for

164

 

110 - —o—TGF-beta 1
: +TGF-beta 2
10° +TGF-beta 3

 

 

 

“73‘ 90 1
4: E
E 8 so
2 ..
0 2 I
= c 70
8 §
0
9. so .
i
so f
40 i——-~ ~ 1 -—--+~-- «w -~ ~—-—-~-~-—- -,
o 2.5 5 10

TGF-beta Concentration (nglml)

Figure 7-6. Effects of TGF-B1, TGF-B2 and TGF-B3 on the growth of

WPEl-NA22 cells. 10, 000 cells/well were plated in complete K-SFM medium in
96 well plates. Cells were treated with TGF-Bl, TGF-B2 or TGF-B3 for 5 days.
WPE 1 -NA22 cells showed a dose-dependent growth inhibition for all three
isoforms. Note that TGF-B1 and TGF-B2 were more inhibitory than TGF-B3

(Bars= +/- SE).

165

110

—o—TGF-beta 1
—-o—TGF-beta 3

3
..‘r.
s:
38
2...
50
-‘E
00
0:2
0
EL.

 

50 j

40 4— W - *-—--—+— —-—- - WW -- WW - 4
0 2.5 5 10

TGF-beta Concentration (nglml)

Figure 7-7. Effects of TGF-B1, TGF-B2 and TGF-B3 on the growth of WPEl-NBZ6
cells. 10, 000 cells/well were plated in complete K-SFM medium in 96 well plates.
Cells were treated with TGF-B1, TGF-B2 or TGF-B3 for 5 days. WPEl-NB26 cells
showed a growth inhibition for all three isoforms. Note that TGF-B2 was the most
inhibitory isoform (Bars = +/-SE).

166

110

 

 

—<>-RWPE-1
100 +VWE1-NA22
+WPE1-N326
g 90 f-
5 5
i 5 so ’
e 2
(9 g .r
% 5 70 ‘ .\
o e , z}
a r
V 60 ,
1
50 1
1
4o 1 ___-__ _ _ ____.__..___ ;_ ~__-a__..___a-_._w_1__.___ a“ ____ ___. , 1
0 2.5 5 10

TGF-B 1 Concentration (nglml)

Figure 7-8. Effects of TGF-B1 on the growth of RWPE-l, WPEl-NA22 cells and
WPEl-NB26. 10, 000 cells/well were plated in complete K-SFM medium in 96 well plates
for each cell line. Cells were treated with TGF-Bl for 5 days. WPEl-NA22 showed the

strongest inhibition by TGF-Bl. WPEl-NB26 cells showed the least inhibition when
compared to RWPE-l cells (Bars: +/-SE).

167

TGF-B1 as compared to the IgG control (Figures 7-2). TGF-B2 and TGF-B3 expression
was not detected by immunoﬂuorescence staining as compared to the IgG control (Figures
7-2). WPEl-NB26 cells show no positive staining for TGF-B1, TGF-B2 or TGF-B3
expression by immunoﬂuorescence staining, as compared to the IgG control (Figures 7-2).
WPEl-NA22 show positive expression for both the TGF-B R11 and TGF-B RIII receptors
as compared to the IgG control (Figures 7-3). WPEl-NB26 show weakly positive
expression for TGF-B RII receptor but do not show TGF-B R111 receptor expression as
compared to the IgG control (Figures 7-3). These results demonstrate that WPEl-NA22
cells express TGF-B1 but lack TGF-B2 and TGF-B3 expression and they do express TGF-B
R11 and R111 receptors. WPEl-NBZ6 cells show a lack TGF-B1, TGF-B2, TGF-B3.
WPEl-NBZ6 cells show only a weak expression of TGF-B R11 and lack TGF-B RIII

receptor expression.

Restoration of acinar-forming ability by TGF-B1 and TGF-B2

Because WPEl-NA22 cells show some acinar-forming ability (69%, see chapter
6 Figure 6-1), it was selected for use to determine if exogenous TGF-B1 and TGF-B2
could induce acinar formation. TGF-B1 and TGF-B2 were chosen because of their ability
to inhibit cell proliferation of WPE] -NA22. WPEI -N A22 cells were treated with TGF-B1
or TGF-B2 at concentration ranging from 0 to 10 ng/ml. Cultures were ﬁxed after four
days of treatment. Acini-forming ability for each treatment was determined by staining
cultures with propidium iodide and counting individual acini using the LSM transmitted

light at low power. TGF-B1 caused an increase in the acinar-forming ability of

168

WPEl-NA22 cells to 134% at 10 ng/ml as compared to the non-treated vehicle control
which was set at 100% (Figures 7-9). TGF—B2 caused a dose—dependent increase in the
acinar-forming ability of WPEl-NA22 cells to 159% at 10 ng/ml (Figures 7-9). These
results show that exogenous TGF-B1 and TGF- B2 enhance the ability of WPEl-NA22 cells

to undergo acinar morphogenesis.

Discussion

Loss of homeostasis of growth in the prostate is an important event in prostatic
carcinogenesis. It results from a shift in the balance between cell proliferation and cell
death in favor of proliferation (Sensibar, 1995). This deregulation of growth may result
from altered expression and sensitivity to growth factors. EGF and TGF-B are key growth
factors which regulate prostate epithelial growth (Webber at al. 1996). TGF-B acts
predominantly as a negative growth regulator of prostate epithelial cell growth, while EGF
acts as a positive growth regulator. As a part of this growth deregulation, exogenous EGF
was shown to inhibit acinar morphogenesis. Further, the neutralization of EGF activity
resulted in the enhancement of acinar-formation and differentiation (see chapter 6). The
focus of this chapter was to examine the role of the negative growth regulator, TGF-B, in
acinar morphogenesis in non-malignant and malignant prostatic epithelial cells.

The importance of TGF-B in prostate function is clearly illustrated by studies in
castrated rats. During castration—induced involution, the prostate responds to the ablation
of androgens with a decrease in EGF-receptor expression and a sharp increase in TGF-B

and TGF-B RII receptor expression. This increase in the TGF-B and receptor expression

169

200 I TGF-beta 1
13° 1:! TGF-beta 2

 

 

 

 

 

3‘ 160
E 140
E 1‘3 120 ~
g g 100
:3 ~
I-
G)
$ 40
20 7 _.
0 a , +-~ W17 - +-~ a
0 nglml 2.5 nglml 5 nglml 10 nglml
TGFbeta (nglml)

Figure 7-9. Induction of acinar formation by TGF-B. WPEl—NA22 cells were grown on
Matrigel and treated with TGF—B1 and TGF—B2 for 4 days. Acini-forming ability was
determined by staining cultures with propidium iodide and counting individual acini using
the LSM transmitted light at low power. Counts are expressed as a percent of the vehicle
treated-control which was set at 100 %(Bars= +/-SE).

170

causes the prostatic epithelial cells to undergo apoptosis (Byme et al. , 1996, Culig et al,
1996, Nishi et al., 1996). In the normal adult prostate, TGF-B functions as a negative
growth regulator and maintains growth homeostasis in prostatic epithelium. In fact, TGF-B
is the major antagonist of positive growth factors in the prostate (Culig et al, 1996).

Although TGF-B1, TGF-B2 and TGF-B3 are known to be expressed in the fetal
prostate, their roles in prostatic morphogenesis are not clear. In the mouse prostate model,
TGF-B2 and TGF-B3 expression, although present in the fetal urogenital sinus, is virtually
absent in the adult tissue, while TGF-B1 expression increases with time. Fetal expression
of TGF-B1 is localized to the urogenital mesenchyme surrounding the developing prostatic
buds (Timme et al., 1994). In contrast, work in breast and lung development has shown
a spatial and temporal pattern for TGF-B accumulation during acinar morphogenesis with
a notable loss in focal TGF-B1 expression at the bud forming regions (Daniel et al. , 1989,
Heine et al. 1990, Silberstein et al, 1992, Timme et al., 1994).

In order to elucidate the roles of TGF-B 1 , TGF-B2, and TGF-B3 in human prostatic
acinar morphogenesis, their effect on non-malignant immortalized human prostatic
RWPE-I cells was examined. TGF-B1, TGF-B2 and TGF-B3, respectively, inhibits EGF
induced RWPE—l cell growth. This EGF, at 5 ng/ml, is a component of the culture
medium. RWPE-l cells demonstrate a differential sensitivity to TGF-B isoforms, as
TGF-B1 >TGF-B2>TGF-B3, with cells being most sensitive to TGF-B1 (Figure 7-1).
RWPE-l cells show strong endogenous expression of TGF-B1 and weak expression of
TGF-B2 and TGF-B3 (Figures 7-2). RWPE-l cells express TGF-B R11 and TGF-B RIH

receptors suggesting the formation of RII/RIII TGF-B receptor heterodimer (Figures 7-3).

171

The RII/RIII receptors binds both TGF-B1 and TGF-B2 with high afﬁnity as compared to
their binding of TGF-B3. These data suggest an autocrine growth regulation of RWPE-l
cells by TGF-B.

TGF-B1 and TGF—B2 expression plays a critical role in acinar morphogenesis of
RWPE-l cells. Inhibition of TGF-Bl or TGF-B2 activity by anti-functional antibodies
reduced acini-formation by 48 % and 44 % , respectively. Inhibition of TGF-B3 activity only
reduced acinar-formation by 20%. The data clearly indicate that TGF-B1 and TGF-B2 are
essential for cell polarization and acinar morphogenesis. The importance of TGF-B in
prostate acinar morphogenesis is supported by similar ﬁndings in the submandibular gland
(Hoffman, 1996).

If TGF-B function is truly necessary for acinar formation , then blocking the binding
of TGF-B to the cells should result in the inhibition of acinar formation. Results show that
the addition of anti-functional Ab to TGF-B RII or TGF-B RIII, which block the binding
of TGF-B, reduced the ability of RWPE-l cells to form acini by 74% and 65%,
respectively. Taken together, these data show that functional TGF-B binding is necessary
for acinar morphogenesis and suggest that disruption of either TGF-B or TGF-B receptor
function inhibits acinar-forming ability.

As stated earlier, a deregulation of epithelial cell growth is a step in prostate
carcinogenesis. Tumor cells at different stages in progression from slow-growing, low-
tumorigenic, to fast-growing, invasive and highly tumorigenetic, exhibit distinct
alterations in the expression of speciﬁc growth factors and their receptors (McKeehan,

1993). Additionally, a decrease in sensitivity to exogenous TGF-B may provide cells with

172

a growth advantage in malignant progression. Therefore, the effects of TGF-B 1, TGF-B2
and TGF-B3 on PIN-like WPEl-NA22 and malignant WPEl-N B26 cells were examined.
TGF-B1, TGF-B2 and TGF-B3, respectively, differentially inhibited EGF induced
WPEl-NA22 and WPEl-NB26 cell growth. WPEl-NA22 cells show relatively equal
sensitivity to TGF-Bl and TGF-B2 (Figure 7-6). WPEl-NB26 cells demonstrate greater
sensitivity to TGF-B2 than to TGF-B1 (Figure 7-6). In comparison to RWPE-l and
WPEl-NA22, WPEl-NB26 cells show reduced sensitivity to TGF-Bl (Figure 7-8). Thus,
inhibition of cell growth in response to TGF-B decreases as cells progress toward a more
malignant and invasive state. This is in agreement with studies which observed that the
inhibitory effects of TGF-B on prostate cancer diminish as cells progress toward a less
differentiated and more malignant state (Wilding, 1991). In both WPEl-NA22 and
WPEl-NB26, TGF-B3 at 2.5 ng/ml initially caused an a 30% and 26%inhibition in
growth, respectively, but as the concentration of TGF-B3 increased to 10 ng/ml, cells
became less sensitive showing only a 18 % and 15 % reduction in growth (Figure 7-6). This
may be due to a possible biphasic response to TGF-B or simply due to a down regulation
of the receptor by the higher level of the growth factor. Additionally, TGF-B1 and
TGF-B2 have similar in viva biological effects (Hoffman et al. , 1996).

The escape from the negative growth control imposed by autocrine TGF-B
expression may be due to either down regulation of growth factor expression and/or a loss
of receptor expression (Jakowlew et al, 1997). WPEl-NA22 cells, which may represent
an early stage in carcinogenesis, show strong expression of TGF-B1 but TGF-B2 and TGF-

B3 expression was not detected (Figures 7-2). On the other hand, WPEl-NB26 cells which

173

are highly invasive and may represent a late stage in tumor progression, show no TGF-B1,
TGF-B2 or TGF-B3 expression (Figures 7—2). There are conﬂicting reports in the literature
as to the expression of TGF-B in cancer (Kostenuik et al, 1997, Merz et al, 1994, Sehgal
et al, 1996). Some reports note an increased in TGF-B expression associated with prostate
cancer, while others ﬁnd no difference in TGF-B expression between the normal prostate
and prostate cancer (Glynne-Jones et al. 1994). It must be noted that in the latter studies,
TGF-B expression was localized in the stroma ( Kostenuik et al, 1997, Merz et al, 1994,
Sehgal et al., 1996, Thompson et al., 1992,). Another study suggests that autocrine
expression of TGF-B functions as a tumor suppressor (Sun and Chen 1997), in which a
decrease in TGF-B autocrine expression ultimately results in a deregulation of growth,
favoring cell proliferation and carcinogenesis. This is in agreement with my ﬁnding that
non-malignant RWPE-l express TGF-B1, 2 and 3, while WPEl-NA22 and WPE-26 cells
demonstrates a stepwise loss of TGF-B isoform expression. WPEl-NA22 cells express
TGF-B1 but lack TGF-B2 and TGF-B3, while WPEl-NB26 show no expression of any
TGF- B isoforms as examined by immunostaining (Figures 7-2). Additionally, the observed
loss of TGF-B3 is in agreement with its decrease observed in prostate cancer in viva (Merz
et al, 1994)

Another mechanism by which cells may become insensitive to TGF-B regulation
is by the loss of TGF-B receptor expression. Result show that WPEl-NA22 express both
the TGF-B R11 and TGF-B RIII receptors, similar to the parent non-tumorigenic cell line
RWPE-l , while WPEl-NB26 only weakly express TGF-B RII receptor and lack TGF-B

RIII receptor expression (Figures 7-3). These results suggest that a decrease in

174

responsiveness of the malignant WPEl-NB26 cell line may be due to a decrease or loss of
TGF-B receptor expression. This is agreement with other reports where a decrease or loss
of R11 receptors in prostate cancer cells has been observed (Gerdes et a1, 1998, Guo et al,
1997). In fact, the loss of R11 has been correlated with Gleason grade and is associated
with poor prognosis (Glynne-Jones et al. , 1994). WPEl-NBZ6 cells do not express TGF-B
and show very weak expression of only one receptor but it does exhibit some growth
inhibition in response to exogenous TGF-B. This maybe due to the fact that the TGF-B
receptor is upregulated in the presence of the growth factor. Additionally, interruption of
TGF-B in knock out mice does not cause cancer but makes cells very susceptible to cancer
by enhancing cell proliferation during promotion (Schirmacher et al., 1998).

The restoration of acinar-forming ability by reestablishing TGF-B negative growth
regulation was examined using WPEl-NA22 cells. Because WPEl-NA22 represent a PIN-
like stage in carcinogenesis, it shows some acinar-forming ability (69%, see chapter 6
Figure 6-1). Results show that TGF-B1 and TGF-B2 caused an increase in the acinar-
forming ability of WPEl-NA22 cells by 34% and 59%, respectively (Figures 7-9). These
results show that exogenous TGF-B1 and TGF-B2 can enhance the ability of WPEl-NA22
cells to undergo acinar morphogenesis.

In summary, results demonstrate the critical role of autocrine TGF-B1 and TGF-B2
expression in human prostatic acinar morphogenesis. The loss of sensitivity apparently
results from a decrease in TGF-B growth factor and receptor function and appears to be
associated with a loss in the ability to undergo acinar morphogenesis. A stepwise loss in
autocrine TGF-B expression and TGF-B receptor expression was seen in malignant WPEl-

NA22 and WPEl-NB26 cells. These decreases caused a decrease in sensitivity that was

175

correlated to their malignant characteristics and their loss of ability to undergo acinar
morphogenesis. Additionally, restoration of acini-forming ability was stimulated by
exogenous TGF-B treatment which re-established negative growth control. The data in this
chapter have enhanced our understanding of the role of TGF-B in prostate acinar

morphogenesis and its alteration in the process of carcinogenesis and tumor progression.

176

 

References

Byme RL, Leung H and Neal DE: Peptide growth factors in the prostate as mediators of

stromal epithelial interaction. Br. J. Urol. 77: 627-33, 1996.

Culig, A. Androgen receptor activation in prostate tumor cell lines by insulin-like growth
factor-I, keratinocyte growth factor, and epidermal growth factor. Cancer Res. , 54:

5474-5478, 1994.

Daniel CW, Silberstein GB, Van-Horn K, Strickland P and Robinson S: TGF-beta
l-induced inhibition of mouse mammary ductal growth: developmental speciﬁcity

and characterization.Dev-Biol. 135:20-30, 1989.

Dahiya R, Lee C, Haughney PC, Chui R, Ho R and Deng G: Differential gene expression
of transforming growth factors alpha and beta, epidermal growth factor,
keratinocyte growth factor, and their receptors in fetal and adult human prostatic

tissues and cancer cell lines Urology. 48: 963-70, 1996.

Franzen, P., Ichijo, H. and Miyazono. Different signals mediate transforming growth
factor-b 1-induced growth inhibition and extracellular matrix production in prostate

carcinoma cells. Exp. Cell Res., 207: 1-7, 1993.

177

Gerde MJ, Larsen M, McBride L, Dang TD, Lu B and Rowley DR: Localization of
transforming growth factor-betal and type II receptor in developing normal human

prostate and carcinoma tissues .J. Histochem. Cytochem. 46:379-88, 1998.

Glynne-Jones E, Harper ME, Goddard L, Eaton CL, Matthews PN and Grifﬁths K:
Transforming growth factor beta 1 expression in benign and malignant prostatic

tumors. Prostate. 25: 210-8, 1994.

Guo Y, Jacobs SC and Kyprianou N: Down-regulation of protein and mRNA expression
for transforming growth factor-beta (TGF-betal) type I and type 11 receptors in

human prostate cancer. Int. J. Cancer. 71:573-9, 1997.

Heine U1, Munoz EF, Flanders KC, Roberts AB and Sporn MB: Colocalization of
TGF-beta 1 and collagen I and III, ﬁbronectin and glycosaminoglycans during lung

branching morphogenesis. Development. 109:29-36, 1990.

Hoffman MP, Kibbey MC, Letterio JJ and Kleinman HK: Role of laminin-l and TGF-
B3 in acinar differentiation of a human submandibular gland cell line (HSG). J.

Cell Sci, 109:2013-2021, 1996.

Jakowlew SB, Moody TW, Mariano JM: Transforming growth factor-beta receptors in
human cancer cell lines: analysis of transcript, protein and proliferation. Anticancer

Res. 17:1849-60, 1997.

178

Jarrard DF, Blitz BF, Smith RC, Patai BL and Rukstalis DB: Effect of epidermal growth
factor on prostate cancer cell line PC3 growth and invasion. Prostate. 24:46-53,

1994.

Kostenuik PJ, Singh G and Orr FW: Transforming growth factor beta upregulates the
integrin-mediated adhesion of human prostatic carcinoma cells to type I collagen.

Clin. Exp. Metastasis. 15:41-52, 1997.

McKeehan WL: Growth factor receptors and prostate cell growth. Cancer Surv. 1 1 : 165-75 ,

1991.

McKeehan WL, Hou J, Adams P, Wang F, Yan GC and Kan M: Heparin-binding
ﬁbroblast growth factors and prostate cancer. Adv. Exp. Med. Biol. 330:203-13,

1993.

Merz VW, Arnold AM and Studer UE: Differential expression of transforming growth
factor-beta 1 and beta 3 as well as c-fos mRNA in normal human prostate, benign

prostatic hyperplasia and prostatic cancer. World J. Urol. 12:96-8, 1994.

Moustakas A, Lin HY, Henis YI, Plamondon J , O'Connor-McCourt MD and Lodish HF:
The transforming growth factor beta receptors types I, II, and ID form

hetero-oligomeric complexes in the presence of ligand.J-Biol-Chem. 1993 Oct 25;

268(30): 22215-8, 1993.

179

Nakamoto, T., Chang, C.S., Li, A.K. and Chodak, G.W. Basic ﬁbroblast growth factor

in human prostate cancer cell lines. Cancer Res., 52: 571-577, 1992.

Nemeth J A, Sensibar J A, White RR, Zelner DJ, Kim IY, Lee C: Prostatic ductal system
in rats: tissue-speciﬁc expression and regional variation in stromal distribution of

transforming growth factor-beta 1. Prostate. 33: 64-71, 1997.

Nishi N, Oya H, Matsumoto K, Nakamura T, Miyanaka H and Wada F: Changes in gene
expression of growth factors and their receptors during castration-induced

involution and androgen-induced regrowth of rat prostates. Prostate. 28: 139-52,

1996.

Roberts, AB. and Sporn, MB. The transforming growth factor-BS. In: Peptide growth
factors and their receptors 1. Sporn, MB. and Roberts, A.B. (eds.), New York,

Springer-Verlag, pp. 419-472, 1990.

Schirmacher P, Amendt C , Weber H and Blessing M: Expression of a dominant negative
type II TGF-beta receptor in mouse skin results in an increase in carcinoma
incidence and an acceleration of carcinoma development. Oncogene. 17: 25-34,

1998.

180

 

Sehgal I, Baley PA and Thompson TC: Transforming growth factor betal stimulates
contrasting responses in metastatic versus primary mouse prostate cancer—derived

cell lines in vitro.Cancer-Res. 56:3359-65, 1996.

Sensibar JA: Analysis of cell death and cell proliferation in embryonic stages, normal
adult, and aging prostates in human and animals. Microsc. Res. Tech. 30: 342-50,

1995.

Silberstein GB, Flanders KC, Roberts AB and Daniel CW: Regulation of mammary
morphogenesis: evidence for extracellular matrix-mediated inhibition of ductal

budding by transforming growth factor-beta 1. Dev-Biol. 152: 354-62, 1992.

Sun L and Chen C: Expression of transforming growth factor beta type III receptor
suppresses tumorigenicity of human breast cancer MDA-MB-23l

cells.J-Biol-Chem. 272:25367-72, 1997.
Thompson TC, Truong LD, Timme TL, Kadmon D, McCune BK, Flanders K, Scardino

PT, Park SH: Transforming growth factor beta 1 as a biomarker for prostate

cancer. J. Cell Biochem Suppl. 16Hz54-61, 1992.

181

Timme TL, Truong LD, Merz VW, Krebs T, Kadmon D, Flanders KC, Park SH and
Thompson TC: Mesenchymal-epithelial interactions and transforming growth
factor-beta expression during mouse prostate morphogenesis. Endocrinology. 134:

1039-45 , 1994.

Webber, MM, Bello D, and Quader S: Immortalized and tumorigenic adult human

prostatic epithelial cell lines. Characteristics and applications. Part. 1. Cell markers

and immortalized, non-tumorigenic cell lines. Prostate 29:386— 394, 1996.

Wilding G: Response of prostate cancer cells to peptide growth factors: transforming

growth factor-beta. Cancer Surv. 112147-63, 1991.

182

 

'"I111111qu