Download Primary Tumor and Metastasis in Ovarian Cancer Differ in Their

[C'ANC'ER K|-:S|-:AKC|| SS. .«SU-MM. September 15, l")95|
Advances in Brief
Primary Tumor and Metastasis in Ovarian Cancer Differ in Their Content of
Urokinase-type Plasminogen Activator, Its Receptor, and
Inhibitors Types 1 and 21
B. Schmalfeldt,-
W. Kühn,U. Reuning, L. Pache, P. Dettmar, M. Se-limiti, F. Jänicke, H. Höfler,and H. Graeff
Frauenklinik der Technischen Universität München. Klinikum rechts tier Isar. Ismaninger Strasse 22. D-KI675. Munich ¡B.S.. W. K., U. R., L P.. M. S., F. J., H. G.I. and
Institut fürAllgemeine Pathologie und Pathologische Anatomie, Technische Universität München¡P.D., H. H. I, Munich, Germany
Abstract
The relevance of urokinase-type
plasminoceli activator (uPA) and plas-
minogen activator inhibitor (PAI) type I in predicting the survival prob
ability of patients with advanced ovarian cancer after radical surgery and
adjuvant chemotherapy by assessing the patients' primary tumors has
recently been shown by us (W. Kühnel al., Gynecol. Oncol., 55: 401-409,
1994). In the present study, we determined uPA, uPA receptor, PAI-1, and
PAI-2 concentrations in primary tumors and tumor-infiltrated omentum
metastasis is controlled by a series of successive events. After local
proteolysis of the surrounding extracellular matrix, detachment and
spread of tumor cells is facilitated followed by the invasion of the
adjacent tissue, crossing of tissue boundaries and the basement mem
brane. Intravasation and extravasation are followed by reimplantation
of tumor cells into the respective target organ, remodeling of a new
tumor stroma, and angiogenesis in order to consolidate a secondary
tumor at a distant locus. For these events a fine-tuned balance between
and rctroperitoneal lymph nodes of ovarian cancer patients. The group
proteolytic factors, their inhibitors, adhesive proteins, angiogenic fac
consisted of 39 patients with advanced ovarian carcinoma stages Fédéra tors, oncoproteins, and growth factors has to be adjusted (2, 3).
tion Internationale de Gynécologieet d'Obstétriquei l-ÕGOiHic or IV; for
comparison 7 patients with early carcinoma stage FIGO I were also
included. In métastasesof the omentum from ovarian cancer stage FIGO
lile or IV patients, we noted a 4-fold elevated uPA content, a 2-fold
increase in PAI-I, and also a significant increase in uPA receptor and
PAI-2 over primary tumors. In métastasesof the lymph nodes the levels of
the respective antigens were also increased when compared to primary
tumors. These data may indicate that elevated levels of components of the
fihrinolytic system at sites of métastasesmay contribute to the aggressive
potential of cancer cells by favoring their rcimplantation and/or the
consolidation of a new tumor stroma.
Introduction
Ovarian carcinoma is one of the most severe gynecological malig
nancies and because of a long asymptomatic course of the disease
displays already advanced stages at diagnosis with extensive cancer
spread. This is reflected by the high incidence (70% of the cases) of
mortality within 5 years, a fact that stresses the necessity of an early
diagnosis for risk selection and the prediction of the course of the
disease. Several investigators have demonstrated that the prognosis of
patients with advanced ovarian cancer could be improved by radical
surgery followed by adjuvant chemotherapy. Although the absence of
residual tumor after cytoreductive surgery proved to be a strong
prognostic factor for survival, the complete resection of the tumor
mass was not sufficient to protect all patients from the recurrence of
the disease and subsequent death. Several attempts have been under
taken to improve the prognosis of cancer patients and to establish
predictive factors that are related to tumor biology. Numerous factors
have been screened, among them the receptors for estrogen, proges
terone, and epidermal growth factors, DNA ploidy, S-phase, and
oncogenes ( 1).
In recent years special attention has been paid to tumor cell surfaceassociated proteolytic enzyme systems regarding their impact on
cancer progression and metastasis. The process of tumor invasion and
Received ft/15/95; accepted 8/4/95.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore he hereby marked advertisement in accordance with
18 U.S.C'. Section 1734 solely to indicate this fact.
1This work was supported by the Deutsche Forschungsgemeinschaft (Klinische For
schergruppe GR 280/4-1) and the BIOMF.D I program (BMH1CTO1346).
: To whom requests for reprints should be addressed.
Basic and clinical research in tumor invasion and metastasis have
focused on the role of tumor-associated proteases such as metalloproteinases and the serine proteases of the plasmin/plasminogen activator
system (2). Metalloproteinases seem to be correlated with tumor cell
invasion and metastasis in animal models; however, conclusive data
on their clinical significance in humans are still lacking. In contrast,
for several different solid cancers a strong prognostic impact of the
scrine protease uPA1 and/or its inhibitor PAI-1 as predictors for the
course and the outcome of the cancer has been reported. Besides their
statistically independent prognostic impact in cancer of the breast (4,
5), lung (6), colon (7), and the gastrointestinal tract (8), significantly
elevated levels of both uPA and PAI-1 have been demonstrated in
ovarian cancer tissues compared to those in normal ovarian tissue
specimens. Patients with advanced ovarian cancer can be subgrouped
on the basis of computer-optimized "cutoff values for uPA and
PAI-1: patients with low uPA and PAI-1 (uPA < 0.9 ng/mg protein;
PAI-1 < 13.5 ng/mg protein) had a statistically better prognosis than
patients with high uPA and PAI-1 antigen levels (9). First indications
for a role of the uPA-mediated proteolytic system in tumor invasion
and metastasis came in 1976 from Astedt and Holmberg (10), who
demonstrated for the first time high uPA concentrations in cultured
ovarian carcinoma tissue. In the following years increased uPA and
PAI-1 levels were also found in tumor tissue extracts of patients with
ovarian carcinoma (11, 12) and in malignant ascites (13).
uPA is synthesized and secreted by normal and tumor cells. Both
the inactive proenzyme form of uPA, pro-uPA, and protcolytically
active uPA bind to a specific glycanlipid-anchored receptor (uPA-R,
CD 87) on the tumor cell surface. Upon binding of pro-uPA, its
activation by other proteases (e.g., plasmin) is facilitated. uPA acti
vates the proenzyme plasminogcn into the broad spectrum serine
protease plasmin, which may in turn either directly degrade extracel
lular matrix proteins such as fibrin, fibronectin, laminin, and proteoglycans or activates certain matrix-degrading enzymes such as procollagenase. The activation of uPA is controlled by its specificinhibitors PAI-1 and PAI-2. Binding of PAI-1 or PAI-2 to uPA-Rbound uPA results in the subsequent internalization of the ternary
1The abbreviations used are: uPA. urokinase-type
minogen activator inhibitor; uPA-R. uPA receptor.
plasminogcn activator; PAI, plas-
3958
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1995 American Association for Cancer Research.
iil'A. iil'A-R.
AND PAI-1/2 IN PRIMARY TUMOR AND MUTASI ASÕS
complex, thereby regulating cell surface plasmin generation. uPA-R
Table 2 Owi/ximwi ofiiPA. uPA-R. PAI-1.anil PAI-2 utilicen /nr/.v//( primary
minorversus
IVIAntigen(ng/mg
tuimir-injiliratal iimenlum nuijus in iul\tinced tmirian cunïmmii Ã-r'lÃ-iOlile ¡ir
in combination with specific plasmin(ogen) receptors focalizes the
plasmin/plasminogen activator system to the tumor cell surface, thus
constituting an effective proteolytic enzyme system (2).
Tumor-associated proteolytic enzyme systems are not only impor
tant for tumor cell spread from the primary tumor but seem to be
crucial parameters for the events at the site of métastases.In this
report we investigated the levels of uPA, uPA-R, PAI-1, and PAI-2 in
tissue extracts of primary tumors, tumor-infiltrated omentum, and
tumor-infiltrated lymph nodes of patients suffering from advanced
ovarian carcinoma.
protein)uPA(median
FIGOIlle/IV
tumor
metastasis(«
39)"(UM(0.05-10.16)2.87(0.55-8.66)16.99(0.14-105.1)0.25'(0.01-11.34)Omentum
(11=
33)"3.65*(0.66-16.58)4.3o(1.15-8.71)34.14(4.60-125.3)1.72
=
(P)II.IKIOI0.050.00090.1104
range)uPA-R(median
range)PAI-1(median
range)PAI-2(median
range)Primary
Patients and Methods
" il. number of patients.
'' n = 34.
' n = 38.
Forty-six patients with early (FIGO I) or advanced ovarian cancer stages
(FIGO lile or IV) were enrolled as par! of a prospective study on ovarian
cancer patients undertaken at the Frauenklinik der Technischen Universität
München(Table 1). Patients with stages FIGO II, Ilia, and Illh were excluded
from the present analysis because of a too small number of eases. Patients with
advanced ovarian carcinoma received the following radical surgical treatment:
longitudinal laparotomy with hysterectomy, bilateral adenectomy. appendec
tomy, infragastric omentectomy, and bilateral pelvic and paraaortic lymphadenectomy. Among the group of patients, seven patients received limited
surgical treatment because of their had health condition. In younger patients
(<35 years) with tumor stage FIGO I. less radical surgery was performed in
order to preserve the fertility of the patients. Histológica! examinations of all
of the tumor tissue sections were performed at the Institut fürAllgemeine
Pathologie und Pathologische Anatomie, Technische Universität München
(Munich, Germany). Benign ovarian tumors (n = 21), e.g., benign ovarian
cysts, serous cystadenoma, or cystadenof'ibroma served as controls.
Tissue Collection and Extraction. Ovarian cancer tissue specimens from
primary tumors, omentum majus, and lymph nodes were collected during
surgery, classified by the pathologist, and stored in liquid nitrogen until used.
Dcep-fro/.en specimens of 200-500 mg wet weight were pulverized (30 s)
using the Micro-Dismembrator (Braun, Melsungen, Germany) set to maximum
power. The resulting powder was immediately suspended in 2 ml TBS (0.02 M
Tris-HCI, 0.125 M NaCI, pH 8.5), 1% (v/v) Triton X-IOO (Sigma. Munich,
Germany), and extracted at 4°Cfor 12 h followed by an ultracentrifugation
step (1()0,(XK) x K, 45 min, 4°C)to separate cell debris.
assisted analysis (EIA program; ICN-Flow Laboratories,
Meckenheim.
Ger
many). Protein concentrations were determined using the BCA Protein Assay
Reagent kit (Pierce, Rockford, IL). Values for uPA, uPA-R, PAI-1, and PAI-2
concentrations are expressed as ng/mg protein of the tissue extracts. In our
laboratory the interassay variation coefficient for u PA was 5.8%, for uPA-R
5.6%, for PAI-1 4.0%, and for PAI-2 1.4%, thus fulfilling the criteria for a
reliable test. Immunohistochemical staining for uPA, its receptor (uPA-R), and
the inhibitor PAI-1 was performed as described in detail by Janicke el al. (4).
The following antibodies supplied by American Diagnostica were used at 10
/ng/ml: uPA (mAb 3689), uPA-R (mAb 3639), and PAI-1 (mAb 3785).
Statistical Analyses. All results are given as median values. Antigen
determinations were performed in duplicate. The significance level of differ
ences between the antigen levels in cancer tissue extracts versus controls was
evaluated using the Mann-Whitney (/ test. Differences in antigen concentra
tions between specimens from the same patient were tested for significance
using the Wilcoxon test for combined samples (5). Differences were consid
ered to be significant when P < 0.05.
Results
Due to our recent findings that both uPA and PAI-1 may serve as
predictors for survival in patients with advanced ovarian carcinoma
(9), we determined in the present study the content of uPA, uPA-R,
PAI-1, and PAI-2 in primary tumors and typical sites of metastasis in
bind 821, 822, 823, and 893, respectively (American Diagnostica, Greenwich.
CT; Refs. 1, 10, 12, and 14). The data obtained were subjected to computcrovarian cancer patients, e.g., omentum majus and, in addition, in
retropcritoneal lymph nodes. The respective tissues were extracted in
the presence of the nonionic detergent Triton X-100, and the antigen
Tahlc 1 Patient dilla tm a^e. Iunior stai;?, Ã-iisl(ilt>u\.grading, nodal status, unti
l\ini>liadi'ni'ctomy
content of uPA, uPA-R, PAI-1, and PAI-2 was determined using the
ELISA. The important new findings in this study can be summarized
VariableMedian
as
follows.
yr
(20-85)72415284595II23713267322597
(range)StageFIGO
age
uPA. The median concentration of uPA in primary tumor tissue of
IFIGO
patients with tumor stage FIGO lile and IV was 0.91 ng/mg protein.
lileFIGO
In extracts of tumor-infiltrated omentum majus, the uPA content was
IVHistologySerousMucinousrindomctroidUnditïerenliatcdGradingGlG2G3G4Nodal
approximately 4-fold higher than in primary tumors exhibiting statis
tical significance (P = 0.0001, Table 2). The corresponding values for
Laboratory Assay Systems. uPA, uPA-R, PAI-1, and PAI-2 concentra
tions were determined by using the commercially available ELISA kits Imu-
statusNilNlNXLvmphadenectomyPelvicParaaorticPelvic
paraaorticSamplingNone60
and
the uPA content in primary tumor tissue and omentum metastasis of
patients with ovarian carcinoma FIGO lile or IV are depicted as
paired samples in Fig. 1. In tumor-infiltrated retroperitoneal lymph
nodes, an approximately 2-fold increase in the median value of uPA
was noted compared to the uPA content in tissue extracts from
primary tumors; however, statistical significance was lacking
(P = 0.68).
uPA-R. The median value for the receptor of uPA, uPA-R. in
extracts of primary tumor tissues of the same patients was 2.87 ng/mg
protein compared to a median value of 4.36 ng uPA-R/mg protein in
tumor-infiltrated omentum majus. This difference was statistically
significant (P = 0.05, Table 2). Values for corresponding uPA-R
concentrations in primary tumor and omentum metastasis from each
3959
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1995 American Association for Cancer Research.
uPA. uPA-R, AND PA1-1/2 IN PRIMARY TUMOR AND METÕSTASIS
or IV were significantly higher than in extracts prepared from benign
ovarian tumors. PAI-2 levels were almost comparable in extracts of
benign and malignant tumors (Table 3). uPA, uPA-R. and PAI-1
levels in extracts of tumor-infiltrated lymph nodes and omentum
100
majus of patients with advanced ovarian carcinoma stages FIGO lile
or IV were significantly higher than in extracts of tumor-free lymph
nodes and omentum majus of patients with early ovarian cancer stage
FIGO I. However, PAI-2 levels were not significantly increased in
extracts of tumor-infiltrated lymph nodes and omentum majus com
pared to their tumor-free counterparts of FIGO I patients.
To demonstrate the cellular origin of uPA in advanced ovarian
cancer, immunohistochemical stainings were performed in compari
son to benign cystadenoma of the ovary (Fig. 5, B and C). All primary
tumor specimens (FIGO IIIc/IV) examined in the present study by an
ELISA displayed a positive reaction of tumor and stromal cells with
mAb 3689 (American Diagnostica) directed to uPA when assessed by
immunohistochemistry. A typical example is shown in Fig. 5C. In
contrast, although stromal cells of benign ovarian tumors were reac
tive with the uPA antibody, epithelial cells failed to react (Fig. 5B).
To demonstrate the specificity of mAb 3689, uPA-rich kidney tissue
specimens were also stained with mAb 3689 (Fig. 5A). As expected,
mAb 3689 intensively stained the renal tubule cells, whereas the
glomeruli were negative. Tumor cells of primary advanced ovarian
carcinomas were also found to be positive for uPA-R (mAb 3639;
American Diagnostica) and PAI-1 (mAb 3785; American Diagnos
tica; data not shown). As for uPA, stromal cells of different origin
were found to react with mAbs to uPA-R or PAI-1.
0.1
0,01
Primary
Tumor
Omentum
Metástasis
Fig. 1. uPA antigen in tissue extracts of primary tumor and omentum metastasis. Paired
samples are displayed for each of the 34 patients with ovarian cancer stage FIGO lile or
IV. Note that there is an increase for the uPA antigen concentration in omentum metastasis
over primary tumor in 30 of 34 patients (P —0.0001).
Discussion
Thus far only little information is available concerning differences
in the content of uPA, uPA-R, and the inhibitors PAI-1 and PAI-2 in
of the ovarian cancer patients are depicted as paired samples in Fig. 2.
As for uPA. the median uPA-R content in extracts of tumor-infiltrated
lymph nodes was increased compared to primary tumors, but lacked
statistical significance (P = 0.53).
PAI-1. The median value of PAI-1 in extracts obtained from pri
mary tumor tissues was 16.99 ng/mg protein; in the corresponding
sites of métastases(tumor-infiltrated lymph nodes, 35.42 ng/mg pro
tein; métastasesof the omentum majus, 34.14 ng/mg protein), an
approximately 2-fold increase over extracts of primary tumors was
noted. The difference in the PAI-1 concentrations in extracts of
primary tumors and omentum majus was statistically significant
(P = 0.0009, Table 2), in contrast to only an insignificant increase in
extracts of tumor-infiltrated lymph nodes (P = 0.11). The PAI-1
concentration in primary tumor tissue and the corresponding one in
omentum metastasis tissue are depicted as paired samples for each of
the patients with ovarian cancer FIGO lile or IV in Fig. 3.
PAI-2. In extracts obtained from primary tumor tissues of patients
with advanced ovarian carcinoma, a low PAI-2 antigen content was
determined and reflected by a median value of 0.91 ng/mg protein. In
extracts of tumor-infiltrated omentum majus, the median PAI-2 con
centration was approximately 7-fold higher than in extracts of primary
tumors. This difference was statistically significant (P = 0.004, Table
2). Corresponding values for PAI-2 in primary tumor and omentum
metastasis from each of the 33 ovarian cancer patients (FIGO lile or
IV) are depicted as paired samples in Fig. 4. PAI-2 antigen measure
ments in extracts of tumor-infiltrated retroperitoneal lymph nodes and
primary tumors did not reveal significant differences (P = 0.82). No
statistical correlation was noticed among uPA, uPA-R, PAI-1, and
PAI-2 content, neither in the primary tumor nor in the métastases.
The concentrations of uPA, uPA-R, and PAI-1 extracts from pri
mary tumor tissue of patients with advanced ovarian cancer FIGO lile
i
Q.
CO
I
Primary
Tumor
Omentum
Metastasis
Fig. 2. uPA-R antigen in tissue extracts of primary tumor and omentum metastasis.
Paired samples are diplayed from each of the 33 patients with ovarian cancer stage FIGO
lile or IV. Note that there is an increase for the uPA-R antigen concentration in omentum
metastasis over primary tumor in 20 of 33 patients (P = 0.05).
3960
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1995 American Association for Cancer Research.
uPA. uPA-R, AND PAI-1/2 IN PRIMARY TUMOR AND METASTASIS
1000-1
fine-tuned balance among uPA, uPA-R, and PAI-1/2 might be impli
cated. With respect to PAI-2, a significant increase was observed in
cystic fluids of patients with malignant ovarian tumors compared to
benign tumors (17). In ascites of patients with ovarian cancer, PAI-2
was found to be an independent factor indicating poor prognosis;
however, measurements in tumor tissues were not performed (18).
PAI-2 levels are also elevated in tissue extracts from breast carcino
mas compared to benign breast tumors, although this increase is of no
clinical relevance (19). We observed only low expression of PAI-2 in
100-
extracts of primary tumors of ovarian carcinoma and of benign tissue
with almost comparable concentrations. These data are in good ac
cordance with data from colon and breast carcinoma (18-20).
1CH
100
10
0,1
Primary
Tumor
Omentum
Metastasis
Fig. 3. PAI-I antigen in tissue extracts of primary tumor and omentum metastasis.
Paired samples are displayed from each of the 33 patients with ovarian cancer stage FIGO
IIIc or IV. Note that there is an increase for the PAI-1 antigen concentration in omentum
metastasis over primary tumor in 26 of 33 patients (P = 0.0009).
i
^
ÛL
primary versus secondary ovarian carcinomas. Comparing the plasminogen activator activity in tissue extracts from primary tumors and
metastatic ovarian carcinomas as early as 1987, it has been shown that
this activity was mainly due to the action of uPA; however, a repre
sentative comparison of uPA activity in primary tumor versus métas
tases has not been reported (14). Information on PAI-I and PAI-2
antigen levels in primary tumor versus metastasis is lacking in ovarian
cancer. In breast carcinoma, significantly elevated PAI-1 levels in
métastaseshave been reported (4). These findings are consistent with
our data on ovarian carcinoma. However, the uPA concentration in
lymph node métastasesof breast tumors was not significantly altered
compared to its content in primary tumors. Similar data have been
reported for colon cancer showing no increase in uPA activity in
contrast to increased PAI-1 levels in liver metastasis (7). Increased
levels of PAI-1 and also PAI-2 at the site of métastasesmight locally
down-regulate uPA activity, thus favoring the process of tumor cell
reimplantation and the formation of a new tumor stroma. In the
present study on ovarian carcinoma, we found both uPA and PAI-1/2
levels elevated at the sites of metastasis. A possible explanation for
the simultaneous increase of the protease uPA as well as its inhibitors
in métastasesof ovarian versus breast and colon carcinomas is that in
ovarian cancer a different type of tumor growth and metastasis is
displayed compared to other cancers. In breast and colon carcinomas,
tumor cells mainly spread hematogenously and/or lymphogenously, a
process in which the formation of tumor cell-loaded microthrombi in
small vessels and the consecutive generation of a new tumor stroma
arising from the tumor cells within the thrombus may play a role.
According to this so-called "microthrombus theory," an increase in
fibrinolytic activity via the action of uPA in the tumor cell vicinity
might prevent metastasis (16). In contrast, in ovarian cancer, tumor
cells spread on and invade the peritoneum, a process in which a
0,1
0,01
Primary
Tumor
Omentum
Metastasis
Fig. 4. PAI-2 antigen in tissue extracts of primary tumors and omentum metastasis.
Paired samples are displayed from each of the 33 patients with ovarian cancer stage FIGO
Mie or IV. Note that there is an increase for the PAI-2 antigen concentration in omentum
metastasis over primary tumor in 24 of 33 patients (P = 0.004).
Table 3 Comparison of am.of
tif'A. uPA-R. PAI-1.
malignantAntigen(ng/mg benign and
protein)uPA(median
levelsovarian
PAI-2 antigen
tumorsOvarian
extractsStatisticalsignificance(P)0.0
tissue
cancer(Primary
ovariantumors
tumor;,1
21)"0.31(0.01-2.54)1.63(0.41-2.83)4.50(1.69-15.30)0.18(0.01-2.47)/
(n =
.39)"0.91(0.05-10.16)2.87(0.55-8.66)16.99(0.14-1(15.1)0.2
=
range)uPA-R(median
range)PAI-1(median
range)PAI-2(median
range)"
patients.*
;i, number of
38.'
n =
n.s., not significant.Benign
3961
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1995 American Association for Cancer Research.
ul'A. iil'A-R. ANI) PAl-l/a
IN PRIMARY TUMOR AND Ml-TASTASIS
In the present study an increase in the uPA-R concentration in
tumor-infiltrated omcntum and lymph nodes of advanced ovarian
tumor stages has been observed. In primary colonie carcinomas and
their métastasesin the liver, an increased presence of uPA-R has been
found, too (21). For colon carcinomas it has been demonstrated that a
high content of uPA-R in tissue extracts correlated with a poor
prognosis for the patients (22). For ovarian carcinomas it remains to
be established whether the amount of uPA-R correlates with Ihe
progression of tumor growth and malignancy. Interestingly, in ascitic
fluid, but also in tumor tissue and blood, a soluble, ligand-free form
of uPA-R has been detected in invasive carcinomas. uPA-R may
derive from the tumor via a phospholipase and/or proteascs-mediated
shedding of cell surface-attached uPA-R (23).
High concentrations of uPA, uPA-R, and PAI-1 in malignant ovar
ian tumor tissue reflect an increased expression of these factors in
tumor cells or nonmalignant cells of the tumor stroma. In the present
investigation on protcolytic factors in advanced ovarian cancer, by
using the same mAbs (3689 and 3639) to uPA and uPA-R, respec
tively, we could confirm the results by Young et al. (24) that both
tumor cells and stromal cells in fact react with the respective anti
bodies. Moreover, our immunohistochemical results support the as
sumption that ovarian cancer cells not only contain uPA and uPA-R
but also the inhibitor PAI-1.
However, it is still an open question whether the process of metas
tasis might depend on the selection of tumor cell clones with increased
expression of plasminogen activators, or whether the synthesis of uPA
and PAI-1/2 is due to regulatory mechanisms or malignant transfor
mation. In several metastasis models, it has been shown that primary
tumors display a heterogeneity with respect to their metastatic poten
tial (25). The occurrence of several types of ovarian carcinomas
eliciting different aggressive potential has been reported: in patients
with tumor stage FIGO IV accompanied by extensive métastasesa
smaller intraabdominal tumor mass had frequently been observed
compared to patients with tumor stage FIGO III, although they had a
comparable rctropcritoneal lymph node involvement. These observa
tions might be related to differences in the expression of plasminogen
activators and their inhibitors. Thus far only little information is
available about the different patterns of protein expression in primary
tumors versus métastasesduring the progression of cancer. In a recent
publication on breast cancer, it has been reported that the CD44
variant expression was increased by approximately 15-35% in lymph
node métastasescompared to primary tumors of breast cancer patients
(26). Likewise, elevated levels of uPA, its inhibitors, and/or uPA-R at
sites of metastasis may contribute to the spread and invasion of
ovarian cancer cells and their capability to reimplant in the perito
neum. The elucidation of the concerted action of the different com
ponents of the plasmin/plasminogen activator system and their inter
dependence involved in tumor metastasis will help to further the
understanding of tumor progression and the promotion of malignancy
in ovarian carcinoma.
Fig. 5. A. immunohistochemical staining of uPA in formalin-fixed, paraffin-embedded
human kidney. The region showed displays the glomerulus and Ihe tubules oÃ-the cortical
area. The glomerulus is devoid of uPA staining, in contrast, the cytoplasm of the epithelial
cells of the renal tubules shows a strong positive reaction with mAh 3689. uPA staining
in red (APAAP); nuclei (light blue) are stained with hemaloxylin. lì,immunohistochem
ical staining of uPA in formalin-fixed, paraffin-embedded benign cystadenoma of the
mucinous type. The epithelial layer (Ã-Ã-rrmv/jtW.v)
does not react with mAb 36S9, whereas
stronial cells are found to be positive. uPA staining in red (APAAP); nuclei (light blue)
arc stained with hematoxylin. C. immunohistochemical staining of uPA in formalin-fixed,
paraffin-embedded ovarian carcinoma (primary tumor, FIGO Mie) of the mucinous type.
Both tumor cells (intense red cytoplasmic slain) and occasionally stroma cells (light red)
react with mAb 36X9. uPA staining in red (APAAP); nuclei (light blue) are stained with
hematoxylin. A-C, X 2(K).
Acknowledgments
We thank Dr. Richard Hart (American Diagnostica, Greenwich, CT) for his
generous support. The expert technical assistance of Erika Scdlac/ck. Brigitte
Jaud-Münch. and Hildegard Seihold is acknowledged.
References
1. Friedlander, M. L., and Dcmbo, A. J. Prognostic factors in ovarian cancer. Semin.
Oncol., IK: 205-212, 19" I.
2. Schmid. M., Wilhelm, O.. Jänicke.F.. Magdolen. V.. Reuning. U., Ohi. II.. Moniwa,
N.. Kobayashi, II.. Weidle, U., and Graeff, H. Urokinase-type plasminogen activator
(uPA) and its receptor (C'DK7). A new target in tumor invasion and metastasis. J.
Obstet. Gynaecol. (Basingslokc), 21: 151-165, 1995.
3. Graeff, H. Tumorbiologie-orientierte
Bchandlungskon/epte
in der gynäkologischen
3962
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1995 American Association for Cancer Research.
ul'A. iiPA-R. ANI) PAI-]/:
IN PRIMARY TUMOR AND MlilASIASIS
Onkologie, Gcburlsh. Fraucnhcilkd., .5.5;MI7-M19. 1995.
Jánicke. F., Schmilt. M.. Hafter. R.. Hollrieder. A., Bahic. R., Ulm. K.. Gossner. W..
and Ciracff. M. Urokinase-typc plasminogcn activator (uPA) antigen is a predictor of
early relapse in breast cancer. Fibrinolvsis. 4: 69-78, WO.
Jänicke.F., Schmill. M.. Pache, L., Ulm. K., llarhcck. N.. Höfler,H., and Gracff. H.
Urokinase (uPA) and its inhibitor PAI-1 are strong and independent prognostic factors
in node-negative breast cancer. Breasl Cancer Res. Treat., 24: 195-208, 1993.
Pederscn. H.. (irondahl-Hansen. J.. Francis. I).. Osterlind. K- Hansen, H. H., Daño,
K., and Brünner.N. Urokinase and plasminogen activator inhibitor type I in pulmo
nary adenocarcinoma. Cancer Res., 54: 120-123, 1994.
Sier, C. F. M., Vloedgraven. J. M., Gancsh, S., Griffioen, G.. Ouax. P. H. A.,
Verheijen. J. J., Dooijewaard, (1.. Welvaart, K.. Van De Velde, C. J. H.. Lamcrs.
C".B. II.. and Verspaget. H. W. Inactive Urokinase and increased levels of its inhibitor
type 1 in coloréela)cancer liver metastasis. Gastroenterology, 107: 1449-1456. 1994.
Nekarda. H.. Siewert. J. R.. Schmilt. M., and Ulm, K. Tumor-associated proteolytic
factors uPA and PAI-1 and survival in totally resected gastric cancer. Lancet. 343:
117, 1994.
Kühn.W.. Pache. L... Schmalleldl. B.. Dcttmar. P.. Schmilt. M.. Jänicke, F., and
Grácil. H. Urokinasc (uPA) und PAI-1 predict survival in advanced ovarian cancer
palients (FIGO III) after radical surgery anil platinum-based chemotherapy. Gynccol.
Oncol., 55.- 401-409, 1994.
10. Astedt. B.. anil Holmberg. L. Ininnmological
identity ol urokinase and ovarian
activator released in tissue culture. Nature (I.ond.). 261:
carcinoma plasminogcn
595-547 m7(,
Pujade-Lauraine. E., Lu, IL. Mirshahi. S., Soria, J., Soria, C., Bernadou, A..
Kruilholf. F. K. O.. I.ijnen. 11. R.. and Burtin. P. The plasminogen-activation system
in ovarian tumors. Int. J. Cancer, 55: 27-31. 1993.
Berns. F.. M. J. J.. Klijn, J. Ci. M., Schmitt. M.. Hen/en-Utgmans, S. C., van der Burg.
M. E. L., Jänicke. F., and Fockens J. A. Occurrence and interrelationships of
urokinase-type plasminogcn activator, plasminogen activator inhibitor l and cathepsin
D in human ovarian tissues. Proc. Am. Assoc. Cancer Res.. 25: (XXI, 1994.
Oh, D„Hwang. S. G.. Hong. S. Y.. Suh, C. L. Lee, H. B., Kye, K. C., Park. S., and
Kim. N. K. Fibrinolytic changes in malignant ascites. Fibrinolysis, <V:156-161. 1994.
Camiolo. S. M.. Markus. G., and Pivers. M. S. Plasminogen activator content of
gynecological tumors and their metaslasis. Gynccol. Oncol., 26: 364-373, I9S7.
Sicr. C'. F. M., Vcrspagel. H. W.. Griffioen. G.. Verheijen. J. J., Ouax. P. H. A..
Dooijewaard. G.. De Bruin Paf. C'. J. H.. and Lamers, C. B. H. Imbalance of
plasminogen activators and their inhibitors in human colorectal neoplasia. Gaslrocn-
terology. 101: 1522-1528. 1991.
16. Markus. G.. Camiolo. M.. Kohga. S.. Madeja. J. M., and Mittclman. A. Plasminogen
activator secretion of human tumors in short-term organ culture, including a com
parison of primary and metastatic colon tumors. Cancer Res., 43: 5517-5525. 1983.
17. Cassieri. B.. Bossmar. T.. Lecander. !.. and Asledt. B. Plasminogen activators and
plasminogen activator inhibitors in Nun,I and tumor fluids of patients with ovarian
cancer. Eur. J. Cancer. 3ÃœA:1302-1311, 1994.
Chambers, S. K., Gertz, R. E., Ivins. C'. M.. and Kacinski. B. M. The significance of
18.
urokinase-type plasminogen activator, its inhibitors and its receptor in ascites of
patients with epithelial ovarian cancer. Cancer (Phila.). 75: 1627-1633, 1995.
19. Foekens, J. A.. Buessecker, F., Peters, H. A.. Krainick. U.. van Putten. W. L. J., Look,
M. P.. Klijn, J. (i. M., and Kramer, M. I). Plasminogen activator inhibilor-2:
prognostic relevance in 1012 patients with primary breast cancer. Cancer Res., 55:
1423-1427, 1995.
20. Bouchet. C.. Spyratos, F., Martin. P. M.. Ilacene. K., Gentile. A., and Oglobine. J.
Prognostic value of urokinase-type plasminogen activator (uPA) and plasminogen
activator inhibitors PAI-1 and PAI-2 in breast carcinomas. Br. J. Cancer. 69: 398405. 1994.
21. Sier, C. F. M.. Ouax. P. H. A.. Vloedgraven. H. J. M.. Verheijen, J. J., Griffioen, G.,
Ganesh, S.. Lamers C'.. B. H. W.. and Verspaget. H. W. Increased urokinase receptor
levels in human gastrointestinal neoplasia related liver métastases.Invasion Metas
tasis, 13: 277-288. 1993.
Gancsh. S., Sier, C. F.. Heerding. M. M.. Griffioen. G., (.amers. C".B.. and Verspaget,
H. W. Urnkinase receptor and colorectal cancer survival. Lancet, 344: 401-402,
1994.
23. Pedersen. N.. Schmilt. M.. Ronne. E.. Nicolelli. M. L, Hoyer-Hansen. G.. Conese. M.,
Giavazzi. R.. Dano. K.. Kühn.W.. Jiinickc. F., and Blasi. F. A ligand-ircc. soluble
urokinase receptor is present in the ascitic tluid from patients with ovarian cancer. J.
Clin. Invest., 92: 2160-2167, 1993.
24. Young, T. N.. Rodriguez, G. C., Moscr, T. L., Basi, R. C.. Jr., Pizzo. S. V., and Stack,
M. S. Coordinate expression of urinary-type plasminogen activator and its receptor
accompanies malignant transformation of the ovarian surface epithelium. Am. J.
Obstet. Gynccol., 170: 1285-1296, 1994.
Fidler. 1. J. Critical factors in the biology of human cancer metastasis: Twenty-eighth
G. 11. A. Clowes Memorial Award Lecture. Cancer Res., 50: 6130-6138. 199(1.
Kaulmann. M., lleider. K. H.. Sinn. H. P.. v. Minckwilz. Ci.. PonÃ-a.IL. and Herrlich.
P. CD44 variant cxon epilopes in primary breast cancer and length of survival,
l-ancel. .«5: 615-619. 1995.
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1995 American Association for Cancer Research.
Primary Tumor and Metastasis in Ovarian Cancer Differ in Their
Content of Urokinase-type Plasminogen Activator, Its Receptor,
and Inhibitors Types 1 and 2
B. Schmalfeldt, W. Kuhn, U. Reuning, et al.
Cancer Res 1995;55:3958-3963.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/55/18/3958
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1995 American Association for Cancer Research.