Vol. 1, 1089-1094, October 1995 Clinical Cancer Research 1089
Advances in Brief
Prostate-specific Antigen, a Serine Protease, Facilitates Human
Prostate Cancer Cell Invasion1
Mukta M. Webber,2 Anuradha Waghray,
and Diana Bello
Departments of Medicine and Zoology, Michigan State University,
East Lansing, Michigan 48824-i3i2
Abstract
Human prostatic epithelial cells constitutively secrete
prostate-specific antigen (PSA), a kallikrein-like serine pro-
tease, which is a normal component of the seminal plasma.
PSA is currently used as a specific diagnostic marker for the
early detection of prostate cancer. We demonstrate that PSA
degrades extracellular matrix glycoproteins fibronectin and
laminin and, thus, may facilitate invasion by prostate cancer
cells. Blocking PSA proteolytic activity with PSA-specific
mAb results in a dose-dependent decrease in vitro in the
invasion of the reconstituted basement membrane Matrigel
by LNCaP human prostate carcinoma cells which secrete
high levels of PSA. A novel PSA-SDS-PAGE zymography
method for the detection of matrix degrading ability of PSA
is also described. We propose that: (a) because of the dys-
plastic cellular disorganization in early neoplastic lesions
called prostatic intraepithelial neoplasia (PIN), PSA may be
secreted not only at the luminal end but also, abnormally, at
the cell-basement membrane interface, causing matrix deg-
radation and facilitating invasion; and (b) PSA, along with
urokinase, another serine protease secreted by prostatic
epithelium, may be involved in the proteolytic cascade dur-
ing prostate cancer invasion and metastasis. The discovery
of the extracellular matrix degrading ability of PSA not only
makes it a marker for early detection but also a target for
prevention and intervention in prostate cancer.
Introduction
Prostate cancer is the most common cancer in adult men in
the United States. An estimated 244,000 new cases and 40,400
deaths from prostate cancer will occur in the United States in
1995 (1). The incidence increases with age, and about 80% of
prostate cancers are diagnosed after the age of 65 years (2).
Because of the increasing life span and an aging American male
population, prostate cancer is a major health concern. The most
threatening and primary cause of death from prostate cancer is
invasion and metastasis. One of the first events in progression to
malignancy is the degradation of the BM3 and ECM, followed
by invasion, a critical early step in the metastatic cascade.
Proteases intervene at the transition from in situ to invasive
carcinoma where local dissolution of the BM occurs. A corre-
lation between an increase in the secretion of matrix-degrading
serine proteases and metalloproteases, and invasion, metastasis,
and aggressiveness of cancer has been demonstrated (3). We
have shown a direct relationship between the level of secreted
urokinase (u-PA) and the ability of the DU145 human prostate
carcinoma cell line, not producing PSA, to degrade and invade
ECM (4, 5). We now show that PSA, which is abundantly
secreted by prostatic epithelial cells, may also play a role in
prostate cancer invasion.
PSA is a single-chain, 240-amino acid glycoprotein with
Mr �33,000 and a primary structure showing considerable ho-
mology to kallikrein (6). The human PSA gene on chromosome
19 has been cloned (7). PSA has the His-Asp-Ser triad in its
catalytic domain, a characteristic of serine proteases. It has
chymotrypsin-like activity, does not hydrolyze synthetic sub-
strates for plasmin, and displays a weak interaction with apro-
tinin, a plasmin inhibitor (8). This suggests that PSA primarily
acts independently as a protease in protein degradation, and not
via plasmin, as does u-PA. PSA is organ-specific, is character-
istically expressed in prostatic epithelial cells, and its expression
is regulated by androgens (9). The androgen-responsive LNCaP
human prostatic carcinoma cells (10) respond by increased PSA
expression (1 1). Following surgical or hormonal castration for
prostate cancer, serum PSA levels in patients decline. A subse-
quent increase in serum PSA level indicates cancer recurrence
(12). Although elevated levels of PSA in the serum of prostate
cancer patients were observed over 25 years ago, PSA measure-
ment has come into wide use only recently for early detection of
prostate cancer, for monitoring patients following radical pros-
tatectomy, and for identifying metastatic tumors of unknown
origin (13). Serum PSA measurement in combination with dig-
ital rectal examination and transrectal ultrasonography has
greatly increased the ability to detect prostate-confined cancer
(12, 13).
PSA is an important component and one of the most
abundant serine proteases in the seminal plasma, where it is
found at an average concentration of about 1.0 mg/ml (9).
Immediately after ejaculation, the seminal plasma coagulates
into a viscous gel which liquefies within about 20 mm. PSA
mediates this liquefaction (14). An understanding of the lique-
faction process provides clues to the mechanism of matrix
degradation by PSA in prostate cancer invasion. The seminal
coagulum is formed by the two most abundant, large molecular
Received 3/23/95; revised 5/16/95; accepted 5/30/95.
I This work was supported by the Biotechnology Research Center and
Research Initiation Grant (Michigan State University).
2 To whom requests for reprints should be addressed, at 5-350 Plant
Biology Building, Michigan State University, East Lansing, MI 48824-
1312.
3 The abbreviations used are: BM, basement membrane; ECM, extra-cellular matrix; PIN, prostatic intraepithelial neoplasia; PSA, prostate-
specific antigen; PSA-Ab, antibody to PSA; u-PA, urokinase-type plas-
minogen activator; mAb, monoclonal antibody.
Research. on June 15, 2020. © 1995 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
1090 PSA Facilitates Human Prostate Cancer Cell Invasion
weight proteins in semen, semenogelin and fibronectin, both of
which are contributed by the seminal vesicles. During liquefac-
tion, PSA-mediated degradation of semenogelin and fibronectin
runs parallel with the dissolution of the seminal gel (14). Lami-
nm, type IV collagen, and fibronectin are major components of
the BM and ECM, respectively, and their degradation is the
initial step in the metastatic cascade. Our objective was to
determine whether PSA was involved in ECM degradation and
invasion.
Materials and Methods
Materials. PSA, Calbiochem 539832, mAb to PSA,
DAKO M750; polyclonal, anticatalytic antibody to u-PA,
American Diagnostica 389; #{128}-aminocaproic acid, A7824, mono-
clonal mouse antihuman antibody against fibronectin, F-7387,
monoclonal mouse antihuman antibody against laminin,
L-827!, human plasminogen 5661, and BSA, A-2153 (Sigma);
gelatin, Bio-Rad 170-6537; fetal bovine serum, Intergen 1020-
75, fibronectin, 40008 laminin, 40232 (Collaborative Research);
Immobilon-P 1PVH 304FO, Millipore; Vector ABC kit, AK-
5001 for Western blots; Vector peroxidase Elite ABC kit PK-
6102, and 3,3’-diaminobenzidine (DAB) nickel kit SK-4100 for
immunostaining (Vector Laboratories); RPM! 1640 medium,
GIBCO 320-1875AJ; Nuclepore filters, 8-p.m pore size, Costar
150446, and HEMA-3, Curtin Matheson 122-911 were used.
SDS-PAGE Zymography for Detection of Urokinase
Activity. PSA samples were analyzed using SDS-PAGE zy-
mography (4) to confirm the absence of u-PA and plasmin
activity. Substrates for u-PA and plasmin were incorporated at a
concentration of 12 �i.g of a solid plasminogen preparation
(activity, 0.44 units/mg solid) and 0.9 mg gelatin, respectively,
per ml of acrylamide in a 10% separating gel. Six p.g PSMane
were run in minigels (200 V, 4#{176}C,45 mm). The gels were further
processed to renature the enzymes and stained with Coomassie
blue. The presence of enzyme activity is indicated by bands of lysis
against a dark background (4). This u-PA activity could be
blocked by antibody to u-PA. Six p.g PSA were mixed with 20
�i.g u-PA antibody and 50 p.g #{128}-aminocaproic acid and incubated
for 2 h at 37#{176}Cbefore SDS-PAGE zymography.
PSA-SDS-PAGE Zymography for Detection of PSA
Proteolytic Activity and the Anticatalytic Ability of Anti-
body to PSA. To determine whether PSA alone has protease
activity and can degrade fibronectin, PSA samples were ana-
lyzed in our new and novel SDS-PAGE zymography method
where fibronectin was incorporated into a 12% acrylamide gel
to serve as a substrate for PSA. Gels were run as described
above, but stained using a silver stain (15). 1� determine
whether the antibody to PSA (PSA-Ab) has anticatalytic ability,
10-pig PSA samples were preincubated with 0.75 pg PSA-Ab
(IgG) for 17 h at 37#{176}Cprior to SDS-PAGE zymography. It
should be noted that not only does the anticatalytic ability of
various PSA antibodies but also the protease activity of different
PSA samples needs to be established because PSA can lose this
activity during the purification process.
SDS-PAGE and Western Blot Analysis. Affinity-pun-
fied PSA derived from human seminal plasma was used for
Western blots and for experiments to further examine fibronec-
tin and laminin degradation by PSA. To determine whether PSA
can degrade fibronectin and laminin, it was necessary to first
establish whether the PSA sample had any contaminating u-PA
activity. Therefore, PSA samples were subjected to Western blot
analysis as well as zymography. Twelve �i.g PSA/lane were
analyzed using SDS-PAGE on 4-15% gradient minigels (200
V, 4#{176}C,45 mm) and transferred to Immobilon membrane. Blots
for PSA were stained with mAb to PSA (1 : 100), and for uroki-
nase, with polyclonal antibody to urokinase (1:500), followed
by avidin-biotin alkaline phosphatase using a Vector ABC kit.
Fibronectin and Laminin Degradation. Degradation of
fibronectin and laminin by PSA is demonstrated by immunoblot
analysis. Twelve pg PSA were incubated with 20 pg antibody
to u-PA and 40 pg E-aminocaproic acid for 2 h at 37#{176}Cto block
any trace of u-PA and plasmin activity. A 2.5-p.g sample of
fibronectin only was run as a control. For fibronectin degrada-
tion, 2.5 pig fibronectin were incubated (17 h, 37#{176}C)with 12 jig
PSA prepared as above, run on a 4-15% gradient gel, trans-
ferred to Immobilon, and stained with mouse antihuman primary
antibody against fibronectin (1 :500), biotinylated goat anti-
mouse secondary antibody (1: 10,000), and avidin-biotin alka-
line phosphatase. For a laminin control, 2.5 jig of reduced
laminin only, boiled with �3-mercaptoethanol for 5 mm, were
run. For laminin degradation, 2.5 jig laminin were incubated
with 12 jig PSA prepared as described above, resolved by
SDS-PAGE, and stained with mouse antihuman primary anti-
body against laminin (1:500). Laminin sample was reduced
before loading into the gel. These experiments were repeated
four times.
Immunocytochemical Staining for PSA in LNCaP
Cells. LNCaP cells were plated on glass coverslips in RPMI
1640 medium and 15% fetal bovine serum at a density of
20,000/400 p.1 medium/well in 24-well culture plates. The fol-
lowing steps were performed at room temperature. Cells were
rinsed with PBS (10 mm) between all successive steps after
primary antibody application. Cells were fixed in 50:50 meth-
anol:acetone, blocked with horse serum (Vector kit) for 1 h, and
incubated for 24 h with mAb to PSA diluted 1:20 in horse
serum. Coverslips lacking primary antibody served as controls.
Prior to staining, cells were treated with 3% hydrogen peroxide
for 3 mm to quench endogenous peroxidase activity, then
stained using a Vectorstain Elite ABC avidin-biotin peroxidase
complex for 30 mm and developed with diaminobenzidine-
nickel chloride substrate for 5 mm.
Invasion Assay. Nuclepore filters were coated with 500
jig/ml Matnigel (16). LNCaP cells were released from mother
flasks using 1 mM EDTA and suspended in RPM! 1640 medium
containing 0.1% BSA. Two hundred thousand or 400,000 cells/
200 jil medium were plated on the coated filter in Boyden
chambers containing 650 p1 medium. The lower chamber con-
tamed 220 p1 of the chemoattractant conditioned medium from
NIH/3T3 fibroblasts grown for 24 h in serum-free medium
containing 50 jig/ml ascorbic acid. Cells were allowed to mi-
grate for 6 to 48 h, fixed on the filter, and stained with HEMA-3.
Nuclear stain was extracted for 15 mm with 0.1 N HC1, and
absorbance was measured at 620 nm using a Titertek microplate
reader (17). Three replicate cultures were prepared per treat-
ment, and the mean values were plotted as percentage of control,
taking untreated invaded cells as 100%. In experiments to de-
termine the role of PSA in invasion, PSA activity was blocked
Research. on June 15, 2020. © 1995 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
1
A
1 2
3lkDa-54kDa-
C
A12 1 2
Clinical Cancer Research 1091
B
D1 2
440 -
220
2 1 0
B
400-. _
220
-31kDa-i� .�
.� 8��#{149}”�
Fig. 1 A, an immunoblot of affinity-purified PSA derived from human
seminal plasma. Lanes I and 2, a PSA sample. Lane 1, stained with mAb
to PSA (1:100). A series of bands at Mr �33,0(X) represent PSA. Lane
2, stained with polyclonal antibody to urokinase (1:500). B, an SDS-
PAGE zymogram (10% gel) for the detection of trace urokinase activityin the PSA sample. Lanes I and 2, received a sample of PSA. Lane 1,
faint zone of lysis at Mr 54,(K)O represents urokinase. Lane 2, PSAsample was blocked with antibody to u-PA. C, SDS-PAGE zymogram(12% gel) to determine whether PSA has protease activity and candegrade fibronectin. Three mg fibronectin were incorporated into the gel
as a substrate for 6 jig PSA loaded into the lane. Lysis caused by PSA
is indicated (arrowhead) at a level lower than the Mr 31,00() marker. D,
SDS-PAGE zymogram (12% gel) to determine whether the antibody toPSA has anticatalytic ability. The gel contained 5 mg fibronectin sub-
strate for PSA. Lane I, 10 jig PSA. Two zones of lysis at Mr �31,000
and a lower molecular weight (arrowhead) are seen. Lane 2, 10 p.g PSAwere incubated with 0.75 jig antibody (lgG) to PSA for 17 h at 37#{176}C
prior to SDS-PAGE.
with PSA antibody (IgG) at 6.25, 12.5, 25, 50, and 100 ngjml in
the final cell suspension. A suspension of one or two million
LNCaP cells/ml of serum-free RPMI 1640 medium containing
0.1% BSA was incubated with the antibody for 2 h before
performing the assay. A nonspecific immunoglobulin (IgG) was
used as a control. Nine such separate experiments were con-
ducted.
Results and Discussion
To attribute degradation of fibronectin and laminin to PSA,
it was necessary to first ascertain whether any u-PA activity was
present in the PSA sample to be used to assess its potential for
matrix degradation. Both PSA and urokinase are normally se-
creted by the prostate and constitute the prostate’s contribution
to the seminal plasma, from which PSA was purified. As stated
earlier, urokinase is also involved in matrix degradation (4, 5).
In the immunoblot in Fig. 1A, both lanes received a PSA
sample. Lane 1 was stained with human PSA-specific and Lane
2 with u-PA-specific antibody. In Lane 1, a series of bands at Mr
25,000-33,000 represent PSA. Variations in prostatic PSA mo-
lecular weight (20,000-35,000) due to glycosylation and alter-
native splicing have been reported (6, 18). Absence of staining
in Lane 2 indicates that u-PA was not detected in the PSA
Fig. 2 Degradation of fibronectin and laminin by PSA is demonstrated
by Western blot analysis. Samples were resolved on 4-15% gradientgels using SDS-PAGE and transferred, and blots were stained with the
respective antibodies. A, immunoblot to show fibronectin degradation
by PSA. Lane 1, 2.5-jig sample offibronectin only (control); Lane 2, 2.5
�i.g fibronectin were incubated with PSA prepared as described in
‘ ‘ Materials and Methods. ‘ ‘ fibronectin degradation products are shown.
Blot was stained with mouse antihuman primary antibody against fi-
bronectin (1:5(X)). B, immunoblot for laminin degradation by PSA. Lane1, 2.5 jig of reduced laminin only; Lane 2, 2.5 jig laminin were
incubated with PSA prepared as described above. Laminin sample was
reduced before loading into the gel. Blot was stained with mouse anti-human primary antibody against laminin (1:5(X)).
sample using immunoblot analysis. A more sensitive zymo-
graphic assay, which can detect as little as 10- “ mol u-PA (19),
was then used to detect trace u-PA activity in the PSA sample by
SDS-PAGE in which the gel also contained plasminogen, a
substrate for urokinase (4). The faint zone of lysis at Mr
-54,000 (Fig. 1B, Lane 1) represents the high molecular weight
u-PA, indicating the presence of trace u-PA activity in the PSA
sample. This activity could be completely blocked by urokinase-
specific antibody (Fig. 1B, Lane 2). Even traces of urokinase are
sufficient to generate plasmin in the presence of plasminogen.
Plasmin, also a serine protease, is formed by urokinase-medi-
ated catalysis of plasminogen to plasmin, and it can degrade
both fibronectin and laminin. The zone of lysis in the PSA
zymogram in Fig. 1C (6 jig PSA), which used fibronectin as a
substrate, shows that PSA has protease activity and it can
degrade fibronectin. In Fig. 11) (Lane 1), when a larger amount
(10 jig) of PSA was run in a similar gel, two zones of lysis, one
at Mr �3l,000 and the other at a lower molecular weight
(-25,000) were observed. This lysis could be blocked by pre-
incubation of PSA with PSA-Ab (Fig. 1D, Lane 2), indicating
that the antibody to PSA used has anticatalytic ability.
To further determine whether PSA alone can degrade fi-
bronectin and laminin, samples of PSA were incubated with
antibody to urokinase to block u-PA activity. Although the
absence of other lysis bands in the zymogram (Fig. !B, Lane 1)
excluded the presence of plasmin, #{128}-aminocaproic acid, a plas-
mm inhibitor, was also incubated with PSA concurrently to
block any trace plasmin activity. PSA samples prepared in this
Research. on June 15, 2020. © 1995 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
PSA Ab
DIgG
C0(I)
>C
LNCaP 6.25 12.5 25 50 100
Antibody Concentration (ng/mI)
Fig. 4 Effects of antibody to PSA on the ability of LNCaP humanprostate carcinoma cells to invade the reconstituted BM Matrigel, using
the Boyden chamber assay. A nonspecific immunoglobulin (IgG) wasused as a control. Bars, SD.
1092 PSA Facilitates Human Prostate Cancer Cell Invasion
4 Unpublished data.
b
Fig. 3 Indirect avidin-biotin immunoperoxidase staining of LNCaP cells using mAb to PSA. a, cells stained with PSA antibody; b, control. Bar, 20
p.m. X 532.
manner were then used to examine the ability of PSA to degrade
the matrix proteins. Samples of fibronectin or laminin were
incubated with PSA, and the degradation products were sepa-
rated by SDS-PAGE and detected using immunoblot analysis.
Fig. 14 (Lane 1) shows the undegraded, control fibronectin
bands at Mr �440,000 dimer and at Mr 220,000 and 210,000
monomeric forms. After incubation of fibronectin with PSA,
several low molecular weight fibronectin degradation bands
were observed, indicating that PSA can degrade fibronectin
(Fig. 2A, Lane 2). Similarly, in Fig. 2B, Lane 1 shows the
undegraded, control laminin bands at MrS �400,000 and
220,000, and in Fig. 2B, Lane 2 shows loss of the two high
molecular weight laminin bands and the appearance of several
low molecular weight degradation products, indicating that PSA
degrades laminin.
To demonstrate that PSA may play a role in invasion by
prostate cancer cells, the relationship between free, secreted
PSA and the invasive ability of LNCaP cells was examined.
LNCaP cells are a useful model for this study because they
produce only a small amount of u-PA as compared to a more
invasive DU145 human prostate carcinoma cell line, and they
also secrete very low levels of gelatinases.4 Nevertheless, they
are invasive, although much less than DU145 cells in an in vitro
invasion assay.4 LNCaP cells were derived by Horoszewicz et
a!. (10) from a lymph node metastasis of prostate carcinoma,
and they formed invasive tumors at the site of injection in nude
mice 8 weeks after injection, but distal metastases were not
apparent at that time. LNCaP cells express (Fig. 3) and secrete
high levels of PSA. The possible role of PSA in invasion is
further suggested by results which show (Fig. 4) that in the
presence of PSA-specific antibody, the ability of LNCaP cells to
invade a reconstituted BM Matrigel is markedly reduced in a
dose-dependent manner. However, a nonspecific immunoglob-
ulin did not inhibit invasion. PSA zymography results (Fig. 1D)
already showed that the PSA-Ab used has anticatalytic ability.
Therefore, a dose-dependent inhibition of invasion by this an-
tibody in the invasion assay suggests that PSA may be associ-
ated with invasion by LNCaP cells.
On the basis of these results, we propose that PSA, along
with u-PA, may be involved in the proteolytic cascade in pros-
tate cancer. Evidence suggests that a proteolytic cascade inde-
pendent of metalloproteases may exist and that degradation of
type IV collagen, an important component of the BM, can occur
via a plasmin-dependent but metalloprotease-independent path-
way (20). Furthermore, kallikreins such as PSA can activate
prourokinase to its active form and subsequently, both u-PA and
plasmin can activate procollagenases (5). Our results show that
ECM components can serve as substrates for PSA, and, thus,
PSA may be involved in localized proteolysis in tumor cell
invasion in prostate cancer.
Research. on June 15, 2020. © 1995 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
Basementmembrane
Invading Capillarycells
Clinical Cancer Research 1093
If PSA plays a role in the proteolytic cascade in prostate
cancer, as suggested here, why are LNCaP cells, which secrete
PSA, less invasive than DU145 human prostatic carcinoma
cells, which do not secrete PSA? The answer may lie in the fact
that DU145 cells secrete high levels of urokinase (4). In the
presence of plasminogen, which is abundantly available in vivo
in blood and extracellular fluids, and in vitro, in serum and
pituitary extract, plasmin is generated by the activation of plas-
minogen by u-PA. Plasmin, a potent protease, has a wide sub-
strate specificity and is able to degrade not only fibrin, fibronec-
tin, and laminin but also type IV collagen (20). Therefore,
plasmin-mediated proteolysis is much more efficient than that
mediated by PSA. As a result, DU145 cells, which secrete high
levels of u-PA, are more invasive than LNCaP cells. If prostate
cancer cells are secreting urokinase, PSA would not be required
for invasion. However, PSA may contribute to the invasive
ability of prostate cancer cells. Our results suggest that in the
absence of high urokinase secretion, PSA could mediate inva-
sion, as in LNCaP cells. These proposed mechanisms do not,
however, exclude the involvement of other proteases in the
proteolytic cascade.
The presence of low grade PINs has been observed in men
in their 20’s and 30’s, the frequency of preinvasive high-grade
PIN increases with age and PINs may precede cancer by 10 or
more years (21, 22). Focal proliferation, cellular disorganization
and heterogeneity, and disruption of the basal epithelial cell
layer, followed by disruption of the BM in high-grade PIN have
been observed (22, 23). In the normal prostate gland in viva,
polarized secretion of prostatic proteases, such as PSA and
u-PA, takes place at the apical, luminal end of glandular epi-
thelium. Disorganization of prostatic epithelial cell monolayers
in vitro can lead to a loss of polarized secretion so that cells
begin to secrete proteases at their apical as well as at their basal
end (24). Accepting the cellular disorganization as seen in
dysplasia in vivo, changes in the polarity of secretion by epi-
thelial cells could take place, leading to similar loss of polarized
secretion. This would result in the secretion of PSA and u-PA
not only into the gland lumen but also at the cell-BM interface,
causing localized proteolysis of the BM and ECM which cul-
minates in invasion and metastasis. Such protease secretion
would lead to their leakage and entry into capillaries (Fig. 5),
resulting in increased serum PSA levels in prostate cancer
patients. This would also explain why increased serum PSA
levels have been observed as early as 6 years before prostate
cancer is discovered by rectal examination (25). For example,
the mean serum PSA level in normal, high-grade PIN, and
carcinoma patients is 4, 7, and 17.9 ng/ml, respectively (23).
The increased serum PSA levels need not necessarily result
from an increased production by carcinoma cells but may be due
to its leakage into the blood vessels. Detection of PSA mRNA
by PCR in cells from peripheral blood of prostate cancer pa-
tients also suggests that dissemination of prostate cancer may be
a relatively early event (26).
We postulate that since prostatic epithelium secretes both
urokinase and PSA, cells in preinvasive prostatic lesions, such
as high-grade PIN, may have an inherent ability and predilection
to invade and metastasize. This may explain why a majority of
prostate cancer patients already have disseminated disease be-
yond the prostate at initial diagnosis by rectal examination. In
Fig. 5 Diagrammatic representation of a prostatic gland with early
invasive carcinoma, which suggests the role of PSA in invasion. Normal
pseudostratified epithelium rests on a continuous BM, whereas in the
area of invasion, it is interrupted. The focus of malignant cells shows
dysplasia so that now the cells are secreting PSA into the gland lumen
as well as at the basal end, causing matrix degradation. The diagram
depicts some of the secreted PSA entering the blood circulation via
capillaries, which would result in a rise in serum PSA levels.
invasion, the net extracellular proteolysis at the cell-BM inter-
face will depend on the ratio between the levels of extracellular
proteases and their natural inhibitors. The balance between the
protease and its inhibitors must be in favor of the active protease
in order for proteolysis to occur. Serine protease inhibitors
(serpins; Ref. 27) and their modulators may serve as useful
agents for blocking the progression of preinvasive to invasive
prostatic carcinoma. Work on the association between the se-
cretion of different proteases and the invasive ability of several
different immortalized and malignant human prostate cell lines
(28) is in progress. The discovery of the ECM-degrading ability
of PSA not only makes it a marker for early detection but also
a target for prevention and intervention in prostate cancer.
Acknowledgments
We thank Dr. Nikolay Dimitrov for helpful comments, Dr. Hynda
K. Kleinman for providing the Matrigel, Aaron I. Steele for preparing
the manuscript, and Kay Steele and Chris Fetters for graphics.
References
1. Wingo, P. A., Tong, I., and Bolden, S. Cancer statistics, 1995. CA
Cancer J. Clin., 45: 8-30, 1995.
2. Carter, B. S., Carter, H. B., and Isaacs, J. I. Epidemiologic evidence
regarding predisposing factors to prostate cancer. Prostate, 16: 187-197,
1990.
3. Liotta, L. A., and Stetler-Stevenson, W. G. Tumor invasion and
metastasis: an imbalance of positive and negative regulation. Cancer
Res., 51 (Suppl.): 5054s-5059s, 1991.
4. Waghray, A., and Webber, M. M. Retinoic acid modulates extracel-
lular urokinase-type plasminogen activator activity in DU-!45 human
prostatic carcinoma cells. Clin. Cancer Res., 1: 747-753, 1995.
Research. on June 15, 2020. © 1995 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
1094 PSA Facilitates Human Prostate Cancer Cell Invasion
5. Webber, M. M., and Waghray, A. Urokinase-mediated extracellular
matrix degradation by human prostatic carcinoma cells and its inhibition
by retinoic acid. Clin. Cancer Res., 1: 755-761, 1995.
6. Watt, K. W. K., Lee, P., M’Timkulu, I., Chan, W., and Loor, R.Human prostate-specific antigen: structural and functional similaritywith serine proteases. Proc. Natl. Acad. Sci. USA, 83: 3166-3170,
1986.
7. Lundwall, A., and Lilja, H. Molecular cloning of human prostate
specific antigen eDNA. FEBS Lett., 214: 317-322, 1987.
8. Christensson, A., Laurell, C., and Lilja, H. Enzymatic activity of
prostate-specific antigen and its reactions with extracellular serine pro-
teinase inhibitors. Eur. J. Biochem., 194: 755-763, 1990.
9. Christensson, A., and Lilja, H. Complex formation between proteinC inhibitor and prostate-specific antigen in vitro and in human semen.
Eur. J. Biochem., 220: 45-53, 1994.
10. Horoszewicz, J. S., Leong, S. S., Kawinski, E., Karr, J. P.,
Rosenthal, H., Chu, I. M., Mirand, E. A., and Murphy, G. P. LNCaPmodel of human prostatic carcinoma. Cancer Res., 43: 1809-1818, 1983.
1 1 . Wolf, D. A., Schulz, P., and Fittler, F. Transcriptional regulation ofprostate kallikrein-like genes by androgen. Mol. Endocrinol., 6: 753-
762, 1992.
12. Oesterling, J. E. Prostate-specific antigen: a valuable clinical tool.
Oncology, 5: 107-128, 1991.
13. Catalona, W. J., Richie, J. P., Ahmann, F. R., Hudson, M. A.,
Scardino, P. 1., Flanigan, R. C., deKernion, J. B., Ratliff, T. L.,
Kavoussi, L. R., Dalkin, B. L., Waters, W. B., MacFarlane, M. I., and
Southwick, P. C. Comparison of digital rectal examination and serum
prostate specific antigen in the early detection of prostate cancer: results ofa multicenter clinical trial of 6,630 men. J. Urol., 151: 1283-1290, 1994.
14. Lilja, H., Oldbring, J., Rannevik, G., and Laurell, C. Seminal
vesicle-secreted proteins and their reactions during gelation and lique-
faction of human semen. J. Clin. Invest., 80: 281-285, 1987.
15. Blum, H., Beier, H., and Gross, H. J., Improved silver staining ofplant proteins, RNA and DNA in polycrylamide gels. Electrophoresis,
8: 93-99, 1987.
16. Albini, A., Iwamoto, Y., Kleinman, H. K., Martin, G. R., Aaronson,S. A., Kozlowski, J. M., and McEwan, R. N. A rapid in vitro assay forquantitating the invasive potential of tumor cells. Cancer Res., 47:
3239-3245, 1987.
17. Grotendorst, G. R. Spectrophotometric assay for the quantitation of
cell migration in the Boyden chamber chemotaxis assay. MethodsEnzymol., 147: 144-152, 1987.
18. Riegman, P. H. J., Klaassen, P., van der Korput, J. A. G. M.,
Romijn, J. C., and Trapman, J. Molecular cloning and characterizationof novel prostate antigen eDNA’s. Biochem. Biophys. Res. Commun.,
155: 181-188, 1988.
19. Dan#{248},K., Andreasen, P. A., Gr#{248}ndahl-Hansen, J., Kristensen, P.,Nielsen, L. S., and Skriver, L. Plasminogen activators, tissue degrada-
tion and cancer. Adv. Cancer Res., 44: 139-226, 1985.
20. Mackay, A. R., Corbitt, R. H., Hartzler, J. L., and Thorgeirsson,U. P. Basement membrane type IV collagen degradation: evidence for
the involvement of a proteolytic cascade independent of metalloprotein-
ases. Cancer Res., 50: 5997-6001, 1990.
21. Sakr, W. A., Haas, G. P., Cassin, B. F., Pontes, J. E., and Crissman,
J. D. The frequency of carcinoma and intraepithelial neoplasia of theprostate in young male patients. J. Urol., 150: 379-385, 1993.
22. Bostwick, D. G. Prostatic intraepithelial neoplasia (PIN): current
concepts. J. Cell. Biochem., 16H (Suppl.): 10-19, 1992.
23. Brawer, M. K. Prostate specific antigen. Acta Oncol., 30: 161-168,
1991.
24. Djakiew, D., Pflug, B., Delsite, R., Lynch, J. H., and Onoda, M.Density dependent polarized secretion of a prostatic epithelial cell line.
Prostate, 20: 15-27, 1992.
25. Helzlsouer, K. J., Newby, J., and Comstock, G. W. Prostate-specificantigen levels and subsequent prostate cancer: potential for screening.
Cancer Epidemiol. Biomarkers Preven., 1: 537-540, 1992.
26. Moreno, J. G., Croce, C. M., Fischer, R., Monne, M., Vihko, P.,
Mulholland, S. G., and Gomella, L. G. Detection of hematogenousmicrometastasis in patients with prostate cancer. Cancer Res., 52:
6110-6112, 1992.
27. Zou, Z., Anisowicz, A., Hendrix, M. J. C., Thor, A., Neveu, M.,
Sheng, S., Rafidi, K., Seftor, E., and Sager, R. Maspin, a serpin with
tumor-suppressing activity in human mammary epithelial cells. Science
(Washington DC), 263: 526-529, 1994.
28. Rhim, J. S., Webber, M. M., Bello, D., Lee, M. S., Arnstein, P.,Chen, L., and Jay, G. Stepwise immortalization and transformation ofadult human prostate epithelial cells by a combination of HPV-18 and
v-Ki-ras. Proc. Natl. Acad. Sci. USA, 91: 1 1874-11878, 1994.
Research. on June 15, 2020. © 1995 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
1995;1:1089-1094. Clin Cancer Res M M Webber, A Waghray and D Bello prostate cancer cell invasion.Prostate-specific antigen, a serine protease, facilitates human
Updated version
http://clincancerres.aacrjournals.org/content/1/10/1089
Access the most recent version of this article at:
E-mail alerts related to this article or journal.Sign up to receive free email-alerts
Subscriptions
Reprints and
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Permissions
Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)
.http://clincancerres.aacrjournals.org/content/1/10/1089To request permission to re-use all or part of this article, use this link
Research. on June 15, 2020. © 1995 American Association for Cancerclincancerres.aacrjournals.org Downloaded from