[CANCER RESEARCH 52, 6754-6760, December 15, 1992]
Invasion of Connective Tissue by Human Carcinoma Cell Lines: Requirement for
Urokinase, Urokinase Receptor, and Interstitial Collagenase 1
Liliana Ossowski
Division of Medical Oncology, The Mount Sinai School of Medicine, New York, New York 10029
ABSTRACT
We have screened six human squamous carcinoma cell lines for their
ability to invade connective tissue by using the experimentally modified
chorioallantoic membrane of a chick embryo as an in vivo model of
invasion. In confirmation of our earlier studies, all the invasive cell lines
expressed high levels of surface-bound urokinase type plasminogen activator (uPA). However, some cell lines expressing this activity were not
invasive, suggesting that surface uPA, although necessary, was not sufficient. Since in addition to fibronectin, that can be degraded by uPA or
plasmin, chorioallantoic membrane connective tissue contains collagen,
we examined the profile of collagenases secreted by the various cell lines
in search for an activity that would coincide with the invasive phenotype.
We found, using gelatin substrate gels, that type IV gelatinase was
produced by all six cell types tested, three cell types produced the Mr
92,000 gelatinase, and three a lower-molecular-weight activity, which
we identified by immunoprecipitation with specific antibodies, and by a
direct assay of activity, as interstitial collagenase. Only the latter cells
were found to be highly invasive.
We showed previously that continuous culture in vitro of one of the
carcinoma cell lines, HEp3, led to a gradual extinction of their malignant phenotype. To confirm the correlation between invasion and the
production of interstitial collagenase, we examined these two functions
in cells freshly isolated from a HEp3 tumor and intermittently during
passage in vitro. We found that, although the surface uPA activity was
slightly diminished in the in vitro grown cultures, it was still within the
range of values found in highly malignant cells, suggesting that it is not
the reason for the decrease in invasiveness. In contrast, the reduction in
interstitial collagenase closely followed the loss of the invasive phenotype; after 30 in vitro passages the cells were almost completely devoid
of interstitial collagenase and unable to invade. The decrease in collagenase activity was not the result of an increased tissue inhibitor of
metalloproteinases production.
INTRODUCTION
The complexity of the metastatic cascade requires that individual steps be studied separately and the overall process reconstructed from the results of such studies. An important intermediate step in the formation of metastasis is the local invasion
of connective tissue by a growing primary tumor. To study
invasion in vivo, we developed a model consisting of an altered
CAM 2 of a chick embryo (1). The experimental alteration
(wounding followed by a period of "resealing") produces a tissue which is enriched in fibroblasts and extracellular matrix and
contains blood vessels of various diameters. Unlike the intact
CAM, it is devoid of the apical epithelium and the underlying
basement membrane (1). Such structure provides a versatile
model for the study of invasion since it accommodates tumor
cells of varying invasive potential and allows for an accurate
quantitation of the invading cells.
It is widely accepted that several proteases, including PA, and
especially uPA, are intimately involved in the metastatic spread
of tumor cells (2-10). Elevated levels of uPA have been shown
to be present in most malignant tumors (7, 10-13) and their
presence was shown to be associated with a more aggressive
disease in several human malignancies (14, 15). In experimental
models, following i.v. injection of tumor cells, modulation of
tumor uPA was shown to correlate with lung colonization
(4-6). Spontaneous metastasis was also blocked by inhibitory
anti-uPA antibodies (2). The unique characteristic of uPA is
that it interacts with a specific receptor on the surface of most
mammalian cell types (16, 17). The invasive ability of several
colon carcinoma cell lines in vitro was shown to be correlated
with the number of surface uPA receptors (8, 9). Using the in
vivo CAM model we showed that cells which simultaneously
expressed both the surface receptor for uPA and elevated levels
of the enzyme invaded more efficiently than cells which expressed either one of the components (1, 8, 9, 18). To assess
whether these properties alone were indicative of an invasive
phenotype, we compared several uPA-producing and receptorexpressing human carcinoma cell lines and found that their
invasiveness varied by 10- to 12-fold, suggesting the presence of
other rate-limiting factors. Since most of the cells under study
originated from squamous cell carcinomas, which are known to
produce collagenase(s) (19-21), and since collagenases are
strongly implicated in invasion (especially in in vitro assays)
(22-25), we tested whether they were a part of the invasive
phenotype in the CAM model. This paper describes the results
of experiments which correlate local invasiveness with the production of interstitial type collagenase. Thus, we found that to
be effective in local invasion, a tumor cell has to be able to
express at least uPA, its receptor and interstitial collagenase,
and possibly, type IV gelatinase.
MATERIALS AND METHODS
Materials. Dulbecco's modified Eagles medium and RPMI 1640
were obtained from Grand Island Biological Co. (Grand Island, NY) in
powdered form and prepared in the laboratory; fetal bovine serum was
from JRH Biosciences (Lenexa, KS); complement fixation avian leukosis (COFAL)-negative embryonated hens' eggs were from SPAFAS
(Norwich, CT); Triton X-100, collagenase type IA, urokinase from
human kidney cultures (used as standard), porcine skin gelatin type II,
1,10-phenanthroline, and collagen from rat tail (type VII) were from
Sigma Chemical Co. (St. Louis, MO); proteinA-Sepharose was from
Pharmacia (Piscataway, N J); chromogenic substrate for plasmin (Spectrozyme PL) was from American Diagnostica (Greenwich, CT). Native,
single chain urokinase (American Diagnostica) was iodinated and used
for receptor-binding experiments, and Abbokinase (low-molecularweight urokinase) from Abbott Laboratories (North Chicago, IL) was
Received 2/20/92; accepted 10/2/92.
iodinated and used for PA inhibitor assay. Human plasminogen was
The costs of publication of this article were defrayedin part by the payment of purified from fresh frozen human plasma, as described previously (26).
page charges.This article must thereforebe herebymarkedadvertisementin accord[~25-I]iododeoxyuridine (specific activity, 2200 Ci/mmol), 125-1, and soance with 18 U.S.C. Section 1734 solely to indicate this fact.
1This work was supported by USPHS Research Grant CA-40758 and the Sam- dium [3H]borohydride (100 mCi/mmol) were obtained from DuPont
uel Waxman Cancer Research Foundation.
New England Nuclear (Boston, MA).
2 The abbreviations used are: CAM, chorioallantoic membrane; PA, plasminoCell Lines. HEp3 [human epidermoid carcinoma (27)] cells were
gen activator; uPA, urokinase-type plasminogen activator; TIMP, tissue inhibitor
of metalloproteinases; PAl 2, plasminogen activator inhibitor 2; PAGE,polyacryl- obtained by dissociation of HEp3 tumors serially passaged on CAMs of
amide gel electrophoresis; PMSF, phenylmethylsuifonylfluoride.
10-day-old chick embryos (28), IC8 and IH4 were isolated from the
6754
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1992 American Association for Cancer
Research.
SURFACE PLASMINOGEN ACTIVATOR. COLLAGENASE, AND LOCAL INVASION
radioactivity present in each well. Activation of samples with 4-amiHEp3 tumor and were previously described (29), Colo 16 (30), Colo
227 (31), and A431 are human squamous cells carcinomas; Colo 227, a nophenyl mercuric acetate (1 mM; 1 h at 37~ increased the activity of
the samples only by 15 to 25% (not shown).
gift from Dr. G. B. Moore, originated in an old burn scar.
Identification of Collagenase by Specific Immunoprecipitation. ConQuantitation of CAM Invasion. This was done essentially as previously described (1). Briefly, cells grown in culture for 2 to 3 days were ditioned medium of HEp3 cells (passage 0) was incubated with protein
labeled with [~25I]iododeoxyuridine (0.2 uCi/ml of medium, for 24 h in A-Sepharose preadsorbed with a mixture of 5 monoclonal antibodies
Dulbecco's medium with 5% fetal bovine serum). The cells were de- against rabbit interstitial collagenase (a generous gift from Dr. Z. Werb,
tached, washed extensively, and resuspended at 3 x 105 cells/50 ul of University of California, San Francisco, CA). Four of these antibodies
phosphate-buffered saline and inoculated onto wounded CAMs of chick recognize the human enzyme. The mixture of antibodies adsorbed to
embryos which were incubated at 37~ for 18 to 21 h after the wound- protein A-Sepharose in 0.1 M Tris buffer, pH 8.0 (150/~l with 100 t~l of
ing (and prior to inoculation with cells). The most invasive cell lines undiluted antibody) was preincubated for 30 min at room temperature,
were tested on CAMs incubated for 30 h after wounding. Invasion was washed extensively, and incubated with 80 #l of HEp3-conditioned
quantitated by determining the number of cells which penetrated the medium for 2 h at 4~ The supernatant of this reaction was incubated
CAM mesenchyme 24 h after inoculation and thus became insensitive again with the antibodies and the two pellets were washed and eluted
with I x sample buffer. The supernatant and the eluates of the 2 imto the release by trypsin (1).
Quantitation of Surface uPA Receptors. Native single chain uroki- munoprecipitates were electrophoresed on a PAGE-gelatin gel and innase (American Diagnostica) was iodinated by using immobilized cubated as described above for 40 h. Mouse monoclonal anti-human
Iodo-Gen. In each case the specific activity was 2 to 5 x 107 cpm/~g uPA antibody (50 ug/ml) served as a negative control.
PA Inhibitor Assay. PA inhibitors secreted by HEp3 passage 0 and
protein. Under reducing (10%/3-mercaptoethanol) or nonreducing conditions in 9% PAGE, followed by autoradiography, more than 95% of passage 41 were assayed as described (38). A constant amount of iodihated low-molecular-weighturokinase was mixed and incubated for I h
the uPA appeared as a single, Mr 55,000 band. The uPA-binding assays
were performed exactly as described (1), except that 2 • 105 cells were at 4~ with undiluted, or 1:5 diluted, conditioned media. The samples
were then subjected to PAGE and autoradiography. RAI 2 formed a
plated per well and the incubation with the ligand was 90 rain at 4~
The number of free receptors, as well as the total number of receptors complex with the low-molecular weight urokinase which ap(obtained after stripping of the endogenous uPA (32), was determined peared as a high-molecular-weightband (approximately Mr 75,000) on
the gel.
for each cell line.
Plasminogen Activator Assay. Each cell type (except HEp3 passage
0) was plated in medium with 10% fetal bovine serum at 3 x l0 s R E S U L T S
cells/60-mm dish. (Hep3, IC8, and IH4, and A431 cells were plated in
IS High Surface uPA Expression a Sufficient Determinant
Dulbecco's medium, Colo 16 and 227 were plated in RPMI.) HEp3
passage 0 cells were plated at 7 x 105 cells/dish. After a 24-h incubation of Local Tumor Invasiveness? Our earlier work, using a h u m a n
the cultures were washed 2 times and incubated for an additional 24 h
squamous cell carcinoma (HEp3), provided evidence for a funcin medium without serum. Conditioned media and cells (lysed in 0.5%
tional link between the expression of u P A and its receptor, and
Triton X-100) were collected and frozen at -20~ until assayed for PA.
the invasive phenotype (1). This was shown both in an experiPA assays were performed exactly as described (33), using human pu- mentally modified chorioallantoic m e m b r a n e of a chick embryo
rified plasminogen (200 ug/ml) and chromogenic substrate (Specand in s.c. injected nude mice (18). To d e t e r m i n e w h e t h e r other
trozyme PL) for plasmin.
Gelatinase Assay. The assay was carried out exactly as described squamous cell carcinoma cell lines displayed similar functional
(34). Briefly, conditioned media from different cell lines were prepared relationship, we inoculated experimentally modified C A M s
as described above, and used fresh or after freezing. They were electro- with 6 different cell lines and quantitated their invasion (Table
1). These cell lines included 4 individual squamous cell carciphoresed on 8% gels copolymerized with 1 mg/ml of porcine gelatin.
The gels were washed in 2.5% Triton X-100 for 30 min, incubated for n o m a s (Colo 16, Colo 227, HEp3, and A431) and 2 clones
40 h at 37~ in buffer containing 50 mM Tris-HCl, pH 7.7, 5 mM CaCI2, isolated from the H E p 3 t u m o r (IC8 and IH4) (29). W e found
and 0.02% NaN3, ("gelatinase buffer"), stained for 10 min in Coo- that the 6 cell lines differed vastly in their ability to invade the
massie blue, and destained to identify the zones of lysis. To characterize modified CAM. Colo 16 and H E p 3 showed the highest invathe proteases, following PAGE and Triton X-100 wash, some gels were siveness; Colo 227 was s o m e w h a t less invasive (Table 1). To
incubated in TIMP (14 ~tg/ml), PMSF (5 mM), or 1,10-phenanthroline accurately quantitate the invasive potential of the 3 additional
(5 mM). In one experiment samples were activated by incubation with 1
cell types (A431, IC8, and IH4) we lowered the stringency of
mM 4-aminophenyl mercuric acetate or 10 #g/ml of trypsin for 1 h at
the assay (1) by shortening the time between the w o u n d i n g of
37~ Samples of conditioned media of each cell line were also electrophoresed on an 8% gel copolymerized with 1 mg/ml of casein (CarnaTable 1 Comparison of invasive ability of squamous cell carcinomas on CAMs
tion powdered milk, Carnation Co., Los Angeles, CA). TIMP activity
Tumor cell lines were serially propagated in culture (with the exception of
was analyzed by using a published procedure (35). Briefly, 50-fold concentrated conditioned media were electrophoresed on PAGE-gelatin, HEp3 p0, which was obtainedfrom CAM tumors 2 days beforethe experiment,
see "Materials and Methods"). All cells were labeled with [~25-1]iododeoxyuriand the gels were washed for 2 h in 2.5% Tris-Triton X-100 and incu- dine (0.2 #C/ml) for 24 h, washed extensively,detached with 1 mM EDTA in
bated in gelatinase buffer containing 20% conditioned medium from phosphate-bufferedsaline and resuspendedin phosphate-bufferedsaline (3 x 105
U937 cells stimulated for 40 h with 50 nM phorbol myristate acetate as cells/50 #l), inoculatedonto CAMs that had been woundedand allowedto reseal
a source of gelatinase. After 4 h of incubation the gels were transferred (see "Materials and Methods") for 20 to 30 h. Each type of cell was tested in 3
individual experiments in a total of 18 to 32 embryos. Invasion was quantito gelatinase buffer and incubated for an additional 15 h, stained, and
tated as described(1). Analysis of variance (Systat Statistics) gave a P value of
destained to reveal the areas which were protected from gelatinase <0.0001. The P values shown in the table were obtained by contrasting the
digestion and appeared as darker bands. To measure interstitial colla- invasion values of each cell line with that of Colo 16, usingpost hor tests.
genase we used the quantitative collagen film assay (36) in which type
Invasion (%)
I collagen was labeled (specific activity, 6 x 105 cpm/mg) with tritium,
Cell type
Mean + SD
Median (range)
P
using pyridoxal phosphate and reduction with sodium [3H]borohydride
Colo
16
48.5-+
23
50.5
(9.8-83.5)
(37). Twenty #1 of labeled collagen (1 mg/ml, 3000 cpm/20 ~1) were
HEp3 p0
44.1 + 29
42.6 (5.3-95.9)
0.547
distributed into 96-well trays, gelled, and dried exactly as described
Colo 227
26.6 -+ 19
16.5 (10.1-47.7)
0.038
(36). The samples (30 t~l), concentrated 50x or undiluted, were incuA431
20.9-+ 23
ll.0 (1.8-78.4)
0.0001
IH4
6.6-+ 6
4.8 (1.5-15.2)
0.0001
bated (in duplicate) for 20 h at 28~ and the radioactivity released into
IC8
4.4 _+3
3.4 (0.5-11.2)
0.0001
the medium was measured and expressed as the percentage of total
6755
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1992 American Association for Cancer
Research.
SURFACE PLASMINOGEN ACTIVATOR. COLLAGENASE, AND LOCAL INVASION
Table 2 Quantitation of uPA production and uPA receptors by human
carcinoma cells
For uPA determination, cells were grown to confluency (usually 48 h after
seeding), the cultures were washed and incubated for 24 h in medium without
serum. Conditioned medium and cell lysateswere used to determine uPA activity (see "Materials and Methods"). The results shown are the mean of biological duplicates which varied by no more than 20%. The number of receptors
per cell and the number of occupied receptors were calculated as described in
"Materials and Methods." The Ka for uPA binding to each of the cell lines was
2 to 7 x -~o M.
Cell type
Colo 16
HEp3
Colo 227
A431
IH4
|C8
uPA
uPA
receptors Occupied
uPA released
cell-associated
per cell receptors
(units/106cells) (units/mgprotein) (• 10 -5) (% of total)
0.4
0.9
1.8
2.6
1.9
0.3
0.1
0.14
0.35
0.4
0.45
0.08
0.68
1.7
ND a
0.3
1.0
2.3
81
92
ND
84
50
73
a ND, not determined.
the CAM and cell inoculation from 30 to 20 h. Even under
these conditions these 3 cell types invaded inefficiently (Table
1).
Were the elevated surface-bound uPA the only requirement
for an efficient invasion, the two properties would be expected
to cosegregate. The results in Table 2 indicate that this is not
the case; although all cell types produced considerable amount
of surface-bound uPA (Table 2), the most highly invasive cells
(Colo 16) were one of the lesser uPA producers, while a poorly
invasive |H4 produced as much as 2.3 units of uPA. We and
others have shown that enhanced uPA production alone is
not a sufficient determinant of an invasive phenotype (1, 8, 9),
and that the cells had to concentrate the enzyme on their surface via specific receptors. All the cell lines tested, however,
displayed a variable but relatively high level of surface uPA
receptors (Table 2) which were at least 50% occupied by endogenously produced uPA (Table 2). Several other cell lines, which
produced tissue type PA, or did not produce any PA, and thus
did not have surface-bound activity (1) 3 invaded poorly in the
CAM assay. This suggests that proteolysis mediated via surface-bound uPA is a necessary but not a sufficient determinant
of invasiveness.
Correlation between Production of Interstitial Collagenase
and the Invasive Phenotype of Squamous Cell Carcinomas.
The stroma of the CAM is rich in collagen and fibronectin
(39, 40). Since both uPA and plasmin cleave fibronectin at
specific and not overlapping sites (41), but they do not cleave
native collagen, we hypothesized that collagen-degrading enzymes were also needed for efficient CAM invasion. A direct
assay of interstitial collagenase activity, using sodium [3H]borohydride-labeled collagen type I as described (36, 37), revealed that the highly invasive cells produced 17 to 600 times
more of the enzyme than noninvasive cells (Table 3). A similar
result was obtained in an additional test in which conditioned
media were analyzed by electrophoresis in a polyacrylamide gel
copolymerized with 1 mg/ml of gelatin, followed by zymography. Three major bands of lysis were revealed (Fig. 1); 2 of the
bands often appeared as doublets. The activities were much
more pronounced (especially the Mr 48,000/43,000 band) if the
conditioned media were electrophoresed immediately after collection and not frozen (compare Fig. 1A with Fig. 1C). Also, the
fresh samples contained bands of lysis migrating with apparent
molecular weights of 39,000 and less, (Fig. 1A), which were not
3 Unpublished results.
present in the frozen samples (Fig. 1C). The calculated apparent molecular weights of the lysis zones obtained by comparison
with reduced molecular weight standards were 83,000, 63,000,
and 48,000/43,000. The 2 slower moving bands most likely
represent the Mr 92,000 and the Mr 70,000/62,000 type IV
collagenases/gelatinases (Fig. 1, A and C). Neither of these
activities correlated with the invasive phenotype; all cells tested,
regardless of their ability to invade the CAM, produced
strong lysis zones corresponding to the type IV gelatinase
(Fig. 1, A and C, Lanes 1, 2, 3, 4, and 5), and the highest levels
of Mr 92,000 gelatinase were found in one of the most invasive
cells (HEp3 passage 0) and in the least invasive cell type (IC8).
The activity that correlated with the invasive phenotype was
the Mr 48,000/43,000 doublet; it was found only in the 3 highly
invasive cells (Colo 16, HEp3, and Colo 227) (Fig. 1A, Lanes 2,
3, and 4, and Fig. 1C, Lanes 2, 3, 4, and 5). Noninvasive IC8
(Fig. 1, A and C, Lanes 1), and IH4 cells (not shown) were
completely devoid of this activity.
None of the lysis zones developed if the incubation buffer
contained the collagenase inhibitor, TIMP (Fig. 1C, Lanes 4T
and 573 or 5 mM concentration of the zinc chelator, 1,10phenanothroline (results not shown). The activity was not inhibited by 5 mM PMSF (results not shown).
Since interstitial collagenase and stromelysin have similar
molecular weights, to identify the Mr 48,000/43,000 lysis
zones, we compared their pattern and intensity on gelatin and
casein gels. Stromelysins produce stronger zones of lysis on
casein versus gelatin substrate gels (34). As shown in Fig. 1B,
the lysis zones produced on casein were weaker than on gelatin
and, both the low-molecular weight activity and the presumed
type IV gelatinase were expressed weakly, suggesting that the
lysis zones were produced by a collagenase. To confirm this
conclusion, we immunoprecipitated conditioned medium from
HEp3 cells with an "oligoclonal" mixture of mouse antibody to
interstitial collagenase. The precipitate and the supernatant
were analyzed by zymography. Fig. 2 shows that the anti-collagenase antibodies removed all of the Mr 48,000/43,000 activity from the conditioned medium and that the activity was
recovered by solubilization from the immobilized antibodies
(Fig. 2, Lanes 2 and 3). A control antibody did not precipitate
the low-molecular-weight activity (Fig. 2, Lanes 4, 5, 6).
A431 cells showed a distinct pattern consisting of 2 lysis zones
Table 3 Interstitial collagenase activity
Cells were grown to confluency in medium with 10% serum, washed 2 times,
and incubated for 24 h in medium without serum. Interstitial collagenase activity
was measured in conditioned medium directly or after concentration 5- or 50fold, on type I, biochemically labeled collagen as substrate (see "Materials and
Methods"). The reaction was carried out at 28~ for 20 h. The results are the
mean of duplicate determinations which differed at most by 11%. The values
were: background (medium alone) 14.6%; trypsin (50 ~g/ml) 1.2% above background; 2 #g/ml of bacterial collagenase 9.5% above background. Soybean
trypsin inhibitor (100 ug/mi) did not inhibit, and 0.1 r~ EDTA completely inhibited the degradation by the 50-fold concentrated HEp3-conditioned medium.
Conditioned medium
(fold concentration)
Digested collagen
(% of total)
HEp3 passage 0
1
5
50
Below detection
Below detection
23.3
HEp3 passage 40
1
50
Below detection
3.4
1
Below detection
1.4
Cell type
IC8
50
Colo 16
Colo 227
1
1
6756
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1992 American Association for Cancer
Research.
4.1
17.0
SURFACE PLASMINOGEN ACTIVATOR,COLLAGENASE,AND LOCAL INVASION
A,
B.
1
KDa
2
3
4
5
1
2
3
4
5
KDa
.---107
83-.Fig. 1. Collagenases(gelatinases)produced
by human carcinomas. Different cell types
were grown to confluency,washed twice, and
incubated in serum-free medium for 24 h.
When indicated, cells were lysed in 0.5% Triton X-100 and 0.1 MTris. pH 8. I, Conditioned
media or cell lysates (15 ~1) were mixed with
sample buffer and eleetrophoresed on an 8%
PAGE copolymerizedwith 1 mg/ml of gelatin
(,4, C) or I mg/ml casein (B) and dc~elopedas
described in "Materials and Methods." .4,
Lane 1, IC8: Lane 2, HEp3 p0; Lane 3, Colo
16: Lane 4, Colo 227: Lane 5, HeLa. B, Lane
1. HEp3 p0: Lane 2, HEp3 p6: Lane 3, IC8:
Lane 4, Colo 16: Lane 5, Colo 227. C, Lane 1,
IC8: Lane 2, Colo 16: Lane 3, Colo 227: Lane
4. HEp3 p0; Lane 5, HEp3 pl: Lane 4T, HEp3
p0: Lane 5T, HEp3 pl (4T and 5Twere developed in presence of 14 ~g/ml of TIMP).
Samples shown in Gel A were prepared and
analyzed without freezing. K/)a, molecular
weights in thousands.
-'--76
63----
.,,..52
48.,=
43.- ,~
36--,-
~49
Co
KDa 1
2
3
4
5
4T
5T
83~
63.-~
48-..,=
43 .~
els and Invasive Potential of H E p 3 Cells Maintained in Culture.
Our previous work has shown that highly malignant H E p 3
KDa
cells, obtained directly by dissociation from a tumor, lose their
malignant potential during in vitro culture (29). Temporally,
the loss of ability to metastasize preceded the loss of tumorigenicity by weeks. We examined whether consecutively in vitro
passaged HEp3 cells show a diminished invasive ability in the
"*-66
C A M assay and if so, what biochemical parameters accompany
this change. We found that the invasive ability was seriously
"-43
curtailed between in vitro passages 12 and 18 (Table 4) and was
reduced to the level of noninvasive clones (Table 1) after 30 in
vitro passages. We quantitated uPA production, uPA receptors,
the affinity of uPA binding, and the production of (PAI-2)
which has been shown to be the major inhibitor produced by
""31
these cells (38). Comparison of highly malignant passage 0
HEp3 cells and cells kept in vitro for 33-41 passages revealed
that the uPA and the receptor production in the latter, although
slightly reduced, was still well within the range found in highly
Fig. 2. Identificationof the Mr 48,000/43,000 bands as interstit;al collagenase. invasive cells (Table 5). I n vivo, the specific PA activity
lmmunoprecipitationwith antibodies specificfor interstitial collagenaseLanes 1, (units/mg of protein) of the tumors produced by noninvasive
2, 3 (for details see "Materials and Methods") or irrelevantantibodies,Lanes 4, 5,
6. Lanes 1 and 4, supernatants, Lanes 2 and 5 first, and Lanes 3 and 6 second cells was much lower (Table 5), mainly due to the relatively
immunoprecipitate (see "Materials and Methods"). KDa, molecular weight in greater contribution of the normal, non-uPA-producing C A M
thousands.
tissue. The concentration of free PAI 2 was greater in the conditioned medium from the highly malignant cells than in the in
vitro grown cells (Table 5).
of apparent molecular weight of 75,000 and 55,000 (results not
The examination of collagenase (gelatinase) production reshown). Incubation of the conditioned medium from these cells
vealed that lysis zones, previously identified as interstitial
with the anti-interstitial collagenase antibodies did not remove
collagenase (Fig. 2), were most p r o m i n e n t in conditioned media
any activity, suggesting that either these cells were devoid of
of highly malignant H E p 3 cells (passage 0 and 2), less prointerstitial collagenase or that their enzyme was antigenically
nounced in medium from passage 18, and completely absent
different.
from the conditioned medium and the cell lysates of passage
Temporal Correlation between Interstitial Collagenase Lev30 (Fig. 3). Similarly, a direct assay of interstitial collagenase
6757
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1992 American Association for Cancer
Research.
1
2 3 4
5
6
SURFACE PLASMINOGEN ACTIVATOR, COLLAGENASE, AND LOCAL INVASION
Table 4 Correlation of CAM invasion with the number of HEp3 passages
in culture
HEp3 cells were isolated from tumors growing on the CAMs of chick embryos
and passaged in culture for the indicated number of times. They were labeled
with [J251]iododeoxyuridinefor 24 h prior to inoculation. Each CAM, which was
wounded and allowed to reseal for 18 h (passage 12, 18, and 30) and 28 h
(passage 0 and 2) was inoculated with 3 x 105 cells. Invasion was determined
24 h later. (For detail see "Materials and Methods"). Invasion is the number of
cells expressed as the percentage of total cells. The results shown are the mean
of determinations made in 14 eggs/passage. The comparison between groups was
done by using analysis of variance (P < 0.0001). The P values shown above were
obtained from post hoc tests in which the invasion of each passage was compared
to that of HEp3 p0.
DISCUSSION
T h i s study was a i m e d at i d e n t i f y i n g t h o s e p a r a m e t e r s o f the
m a l i g n a n t p h e n o t y p e w h i c h enable a t u m o r cell to display locally invasive behavior. W e f o u n d that, in a d d i t i o n to surfaceb o u n d u P A , s q u a m o u s cell c a r c i n o m a cells m a y r e q u i r e interstitial c o l l a g e n a s e for invasion. T h e n e e d for c o l l a g e n o l y t i c
activity in m a l i g n a n c y has long been s u s p e c t e d (19, 20, 43); type
IV gelatinase was s h o w n to be associated with m o r e m a l i g n a n t
cells ( 4 4 - 4 6 ) , a n d c o l l a g e n a s e s w e r e s h o w n to be r e q u i r e d for
invasion o f artificial a n d n a t u r a l b a s e m e n t m e m b r a n e s ( 2 2 - 2 5 ) .
M o r e recently, an a s s o c i a t i o n b e t w e e n the m a l i g n a n t p h e n o t y p e
a n d the p r e s e n c e o f M r 9 2 , 0 0 0 g e l a t i n a s e has been p r o p o s e d
(47, 48). T o the best o f o u r k n o w l e d g e , h o w e v e r , this r e p o r t is
the first to s h o w a close c o r r e l a t i o n b e t w e e n the p r o d u c t i o n o f
interstitial c o l l a g e n a s e a n d the ability to i n v a d e c o n n e c t i v e tissue in vivo.
T h i s c o r r e l a t i o n was identified by two a p p r o a c h e s : in the first
we d e m o n s t r a t e d that the d i s t r i b u t i o n o f interstitial c o l l a g e n a s e
in a series o f s q u a m o u s c a r c i n o m a cell lines p a r a l l e l e d t h e i r
ability to i n v a d e c o n n e c t i v e tissue (Table 1; Fig. 1). In the
s e c o n d a p p r o a c h we s h o w e d that in H E p 3 cells the loss o f the
Invasion (%)
Passage no.
0
2
12
18
30
Mean + SD
28.6
25.0
18.6
10.2
6.1
_+ 20.1
_+ 18.5
_+ 7.1
_+ 9.9
_+ 5.6
Range
P
3.5-69.0
6.8-61.0
7.7-28.6
0.7-37.8
0.4-18.1
0.600
0.143
<0.0001
<0.0001
Table 5 Properties of malignant (in vivo grown) and nonmalignant
(in vitro grown) HEp3 cells
The uPA and uPA receptor number measurements were as in the legend to
Table 2. PAI 2 was determined as previously described (38) and as detailed in
"Materials and Methods." Gelatinase were quantitated by gelatin zymography
and interstitial collagenase by zymography and direct assay (36, 37). Tumorigenicity was determined by inoculating wounded CAMs of 10-day-old embryos
with 2 • 10-5 cells/CAM and incubating the embryos for 7 days, at which time
the tumors were excised, weighed and homogenized in 0.5% Triton X-100. The
lysates were used for uPA determination (see "Materials and Methods"). Invasion was determined as described in the legend to Table 1. Metastasis was
quantitated as previously described (2).
KDa
Properties
0
1.12
+++
1.7 • 105
91.6
3.4 X 10-1o
1
2
3
4
5
6
7
116
97
HEp3 cells (passage in vitro)
uPA in conditioned medium
(units/ml/10 -6 cells)
Free uPA inhibitor (PAl 2)
uPA receptors
Total sites/cell
% of occupancy
KD (M)
Collagenases, (gelatinases)
Mr 92,000, type 1V
Mr 72,000, type IV
Interstitial
Tumorigenicity on CAM
Tumor wt (g)
Tumor uPA (units/mg)
Invasion of CAM
Metastasis to lungs
Std
200
66
33-41
0.74
+
0.9 • 105
83.0
2.0 • 10-l~
+
+++/+
+/+ +
+/+++/++
-/-
0.193 + 0.04
1.980 + 0.87
+++
+++
0.024 +_0.005
0.449 + 0.066
+
-
Fig. 3. The change in the profile of collagenases in HEp3 cells with passage in
vitro. Conditioned media or cell lysates were prepared as described in legend to
Fig. 1 and analyzed on 8% PAGE-gelatin. Std, high-molecular-weight standards;
Lanes 1, HEp3 p30 cell lysates; Lane 2, conditioned medium; Lane 3, HEp3 p18;
Lane 4, HEp3 p2; Lane 5, HEp3 p0; Lane 6, IC8; Lane 7, Colo 227. KDa,
molecular weight in thousands.
BIB
AI
1
revealed greatly r e d u c e d e n z y m e activity in H E p 3 cells passaged in vitro (Table 3). T h u s , the r e d u c t i o n in the interstitial
c o l l a g e n a s e c o r r e l a t e d well with the decline in invasiveness (Table 4). T h e activity o f type IV g e l a t i n a s e was slightly i n c r e a s e d
in the n o n i n v a s i v e cells a n d the M r 9 2 , 0 0 0 gelatinase was
p r e s e n t in passages 0 to 18 a n d was d i m i n i s h e d in passage 30.
T h e o b s e r v e d decline in c o l l a g e n a s e activity was n o t the result
o f i n c r e a s e d p r o d u c t i o n o f T I M P since the invasive H E p 3 passage 0 cells, w h i c h p r o d u c e relatively high levels o f interstitial
c o l l a g e n a s e , were the only cells f o u n d to p r o d u c e also high
levels o f T I M P (Fig. 4, L a n e 4). All 3 cell lines tested p r o d u c e d
low levels o f T I M P 2. As a c o m p a r i s o n , u n d i l u t e d c o n d i t i o n e d
medium from PMA-stimulated U937 contained much more
T I M P 2 t h a n 50-fold c o n c e n t r a t e d m e d i a f r o m the tested cells
(Fig. 4). Several o t h e r areas o f g e l a t i n a s e i n h i b i t i o n (possibly
o t h e r f o r m s o f T I M P ) o f h i g h e r m o l e c u l a r weights, were det e c t e d in the reverse z y m o g r a p h y , but w e r e n o t investigated
further.
23456
1
2
3
4
5
lip
I
GID
Fig. 4. Analysis of production of functional TIMPs. Conditioned media of
HEp3 passage 0, 30, or IC8 were concentrated 50-fold; conditioned medium from
phorboi myristate acetate-stimulated U937 cells was used without concentration.
A, samples electrophoresed on PAGE and stained with Coomassie blue. Lane 1,
molecular weight standards; Lane 2, IC8; Lane 3, HEp3 passage 30; Lane 4,
HEp3 passage 0; Lane 5, phorbol myristate acetate-stimulated U937 cells; Lane
6, recombinant TIMP. B, the samples were electrophoresed on PAGE-gelatin,
and processed as described in "Materials and Methods" for reversed zymography.
Dark bands indicate areas in which gelatin was protected from digestion by
gelatinase. Upper arrow, position of TIMP; lower arrow, position of TIMP 2.
Lanes 1-5, as in A.
6758
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1992 American Association for Cancer
Research.
SURFACE PLASMINOGEN ACTIVATOR. COLLAGENASE, AND LOCAL INVASION
ability to invade connective tissue upon repeated in vitro pas- ered several possibilities. For instance, we tested whether uPA
sage was temporally linked with the loss of interstitial collagen- production in the highly invasive ceils is up-regulated by interase (Tables 3, 4, and 5, Figs. 2, 3, and 4).
actions with extracellular matrix components. This was done by
We used gelatin zymography and a direct assay for interstitial comparing the specific uPA activity of invasive and noninvasive
collagenase to identify the proteases produced by the various cells in vivo in the CAM, and in vitro, using fibronectin as a
squamous carcinoma cell lines. The zymography revealed the matrix (results not shown). We found the basic uPA production
presence of several bands of gelatinolytic activity with the ap- to be unaltered by these interactions. We have also examined
parent molecular weights of 83,000, 63,000, and 48,000/ the possibility that invasion is associated with the triggering of
43,000. All 3 lysis zones were judged to belong to the metallo- the host inflammatory response. The HEp3 tumor has been
proteinase family since their activity was inhibited by incu- shown to induce increased neutrophil production (18). Zymogbation of the gels in the presence of TIMP (Fig. 1C), a zinc raphy of conditioned media obtained from HEp3 tumor cells
chelator (1,10-phenanthroline), or EDTA (results not shown), cultured in vitro for only 2 days (Fig. 1A, Lane 2, and Fig. 3,
while a serine protease inhibitor (PMSF, 5 mM) was without Lane 5) showed additional activities (of slow mobilities) which
effect (results not shown). We presumed that the 3 observed disappeared with time in culture. We have not identified the
lysis zones correspond to the previously identified Mr 92,000, source of these additional lysis zones but, since conditioned
70,000/62,000, and 57,000/52,000 species of collagenase media of control CAMs of the equivalent age cultured in vitro
(45, 47). The discrepancy between the published apparent mo- for 2 days, did not show similar activities (results not shown),
lecular weights and the molecular weights seen in our zymog- we concluded that inflammatory chick cells stimulated by the
raphy may be technical in nature; we used reduced protein as tumor may be contributing these activities. It should be kept in
molecular weight standards. Also, most publications do not mind, however, that the most invasive cell line, Colo 16, did not
show the actual position of standards but in one (46), in which produce an inflammatory response, suggesting that it was not
the standard protein bands were shown, the relative mobility of necessary for this stage of metastasis. Another possibility was
the Mr 70,000/62,000 gelatinase was similar to ours. Thus, we that some cell lines may have the ability to activate chicken
concluded that squamous carcinoma cell lines, in addition to plasminogen more efficiently, thus generating more proteolytic
producing an activity (Mr 48,000/43,000) which we identified activity. We found, however, that the rate of activation was
with the aid of specific antibodies (Fig. 2) and by a direct assay directly proportional to the amount of uPA produced. This was
(Table 3) as interstitial collagenase, also produce a Mr 92,000 true for the soluble as well as for the surface-bound uPA (results
and a 70,000/62,000 collagenase. The Mr 48,000 and 43,000 not shown). Another property, the ability to survive on the
are most likely glycosylation forms of the same enzyme (49) or CAM, was similar for all the cell lines because when inoculated
possibly nonactivated and activated forms. The type IV activity onto CAMs immediately after wounding, all cell types "invadappeared in some cell types as two lysis zones; the lower-mo- ed" equally well (1). 3 Finally, although the level of TIMP prolecular-weight species was slightly more pronounced in fresh duction appears to play a crucial role in some models of invaconditioned media (Fig. 1, A and C). The very low-molecular- sion and metastasis (35, 52, 53), it does not seem to be
weight (<40,000) lysis zones were present only in fresh condi- responsible for the decline in invasiveness of HEp3 cells pastioned medium (Fig. 1A) or following activation with 1 mM saged in vitro (Fig. 4); the most highly invasive cells, which
4-aminophenyl mercuric acetate or trypsin (results not shown). produced high levels of interstitial collagenase, also produced
These were immunoprecipitated by the anti-collagenase anti- high levels of TIMP (Fig. 4). Thus, in addition to a sufficient
bodies (results not shown). These results confirmed published level of surface uPA activity, interstitial collagenase, and posobservations (50, 51) showing that the Mr 70,000/62,000 and sibly type IV gelatinase, are at present the only identifiable
the 48,000/43,000 collagenases may be at least partially acti- factors responsible for the locally invasive behavior of squamous cell carcinomas. The need for at least two, and possibly
vated in culture medium.
We found that the cell lines examined in this report (and three, proteolytic enzymes suggests the existence of a proother human carcinoma cell lines not included here), regardless teolytic cascade in which surface uPA generates plasmin which
of their invasive ability, all produced type IV gelatinase. Thus, in turn has a dual function as a fibronectin-degrading enzyme
we cannot conclude whether the enzyme is essential, or not and as an activator of interstitial collagenase. Such interactions
required at all in local invasion. The fact that the least invasive have been shown to be present in cultures of keratinocytes and
cells (IC8) were the best producers of type IV activity (Fig. 1A fibroblasts (51). The final direct proof of the causal role of
and 16, Lane I) supports the latter possibility. It also appears interstitial collagenase in local invasion awaits transfection with
that local invasion of connective tissue may proceed in the antisense mRNA.
absence of the Mr 92,000 gelatinase, since two of the most
invasive cells (Colo 16 and Colo 227) did not produce this ACKNOWLEDGMENTS
activity (Fig. 1C, Lanes 2 and 3). These results support the
widely accepted hypothesis that a specific set of properties is
I thank Dr. Zena Werb from the University of California, San Francisco, for the generous gift of the anti-interstitial collagenase antibodrequired for each step of metastasis.
In assessing the link between the interstitial collagenase and ies, Jin Yi for technical assistance, T. Ossowski for help with the prepthe process of invasion, we considered the following facts. First, aration of the manuscript, Dr. Rafael Mira y Lopez for critical reading
since we have shown that a high level of surface uPA activity is of the manuscript, and Dr. Samuel Waxman for stimulating discussions
required for invasion of the CAM (1), we asked whether all and support.
efficiently invading cells possess this property. We found that
all the squamous carcinoma cell lines tested, regardless of their REFERENCES
1. Ossowski, L. In vivo invasion of modified membrane by tumor cells: the role
invasive ability, were equipped with this activity (Table 2), sugof cell surface-bound urokinase. J. Cell Biol., 107- 2437-2445, 1988.
gesting that in addition to surface proteolysis mediated via
2. Ossowski, L., and Reich, E. Antibodies to plasminogen activator inhibit
uPA, additional factors were required for invasion. We considhuman tumor metastasis. Cell, 35:611-619, 1983.
6759
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1992 American Association for Cancer
Research.
SURFACE PLASMINOGEN ACTIVATOR. COLLAGENASE, AND LOCAL INVASION
3. Ossowski, L. Plasminogen activator dependent pathways in the dissemination of human tumor cells in the chick embryo. Cell, 52: 321-328, 1988.
4. Hearing, V. J., Law, L. W., Corti, A., Appella, E., and Blasi, F. Modulation
of metastatic potential by cell surface urokinase of murine melanoma cells.
Cancer Res., 48: 1270-1278, 1988.
5. Axelrod, J. H., Reich, R., and Miskin, R. Expression of human recombinant
plasminogen activators enhances invasion and experimental metastasis of
H-ras-transformed NIH 3T3 cells. Mol. Cell. Biol., 9: 2133-2141, 1989.
6. Yu, H., and Schultz, R. M. Relationship between secreted urokinase plasminogen activator activity and metastatic potential in murine B 16 cells transfected with human urokinase sense and antisense genes. Cancer Res., 50:
7623-7633, 1990.
7. Markus, G. The relevance of plasminogen activators to neoplastic growth.
Enzyme (Basel), 40: 158-172, 1988.
8. Schlechte, W., Murano, G., and Boyd, D. Examination of the role of urokinase receptors in human colon cancer mediated laminin degradation. Cancer
Res., 49: 6064-6069, 1989.
9. Hollas, W., Blasi, F., and Boyd, D. Role of the urokinase receptors in facilitating matrix invasion by cultured colon cancer. Cancer Res., 50: 36903695, 1991.
10. Testa, J. E., and Quigly, J. P. The role of plasminogen activator in aggressive
tumor behavior. Cancer Metastasis Rev., 9: 353-367, 1990.
11. Dano, K., Andreasen, P. A., Grondahl-Hansen, J., Kristensen, P., Nielsen, L.
S., and Skriver, L. Plasminogen activators, tissue degradation and cancer.
Adv. Cancer Res., 44: 140-239, 1985.
12. DeBruin, P. A., Griffioen, G., Verspaget, H. V., and Verheijen, J. H. Plasminogen activators and tumor development in the human colon. Activity
levels in normal mucosa, adenomateous tissue, polyps and adenocarcinomas.
Cancer Res., 47: 4654-4657, 1987.
13. Sappino, A. P., Busso, N., Belin, D., and Vassalli, J-D. Increase of urokinasetype plasminogen activator gene expression in human lung and breast carcinomas. Cancer Res., 47: 4043-4046, 1987.
14. Janicke, F., Schmidt, M., Hollrieder, A., Babic, R., Uim, K., Gossner, W.,
and Graeff, H. Urokinase-type plasminogen activator (uPA) is a predictor of
early relapse in breast cancer. Fibrinolysis, 4: 69-78, 1990.
15. Sim, P-S., Stephens, R. W., Fayle, D. R. H., and Doe, W. F. Urokinase-type
plasminogen activator in colorectal carcinomas and adenomatous polyps:
quantitative expression of active and proenzyme. Int. J. Cancer, 42: 483-488,
1988.
16. Vassalli, J-D., Baccino, D., and Belin, D. A cellular binding site for the Mr
55,000 form of the human plasminogen activator, urokinase. J. Cell Biol.,
100: 86-92, 1985.
17. Stoppelli, M. P., Corti, A., Soffientini, A., Cassani, G., Blasi, F., and Assoian, R. K. Differentiation-enhanced binding of the amino-terminal fragment of human urokinase plasminogen activator to a specific receptor on
U937 monocytes. Proc. Natl. Acad. Sci. USA, 82: 4939-4943, 1985.
18. Ossowski, L., Russo-Payne, H., and Wilson, E. L. Inhibition of urokinasetype plasminogen activator by antibodies: the effect on dissemination of a
human tumor in the nude mouse. Cancer Res., 51: 274-281, 1991.
19. Burman, J. F., and Carter, R. L. Lysis of type-I collagen by squamous
carcinomas of the head and neck. Int. J. Cancer, 36:109-116, 1985.
20. Tsuboi, R., Yamaguchi, T., Kurita, Y., Nakao, H., Ogawa, H., and Ishihara,
K. Comparison of proteinase activities in squamous cell carcinoma, basal cell
epithelioma and seborrheic keratosis. J. Invest. Dermatol., 90: 869-872,
1988.
21. Sappino, A-P., Belin, D., Huarte, J., Hirschel-Scholtz, S., Saurat, J-H., and
Vassalli, J-D. Differential protease expression by cutaneous squamous and
basal cell carcinomas. J. Clin. Invest., 88: 1073-1079, 1991.
22. Mignatti, P., Rohbins, E., and Rifkin, D. B. Tumor invasion through the
human amniotic membrane: requirement for a proteinase cascade. Cell, 47:
487-498, 1986.
23. Brown, P. D., Levy, A. T., Margulies, I. M. K., Liotta, L. A., and StetlerStevenson, W. G. Independent expression and cellular processing of Mr
72,000 type IV coilagenase and interstitial collagenase in human tumorigenic
cell lines. Cancer Res., 50: 6184-6191, 1990.
24. Cajot, J-F., Schleuning, W-D., Medcalf, R. L., Bamat, R. J., Testa, J., Liebermann, L., and Sordat, B. Mouse L-cells expressing human prourokinase
type PA: effects on extracellular matrix degradation and invasion. J. Cell
Biol., 109: 915-929, 1989.
25. Reich, R., Thompson, E. W., Iwamoto, Y., Martin, G. R., Deason, J. R.,
Fuller, G. C., and Miskin, R. Effect of inhibitors of plasminogen activator,
serine proteinases and collagenase IV on the invasion of basement membranes by metastatic cells. Cancer Res., 48: 3307-3312, 1988.
26. Deutsch, D. G., and Mertz, E. T. Plasminogen purification from human
plasma by affinity chromatography. Science (Washington DC), 170: 10951096, 1970.
27. Toolan, H. W. Transplantable human neoplasms maintained in cortisonetreated laboratory animals: HS# 1, H.Ep.#1, H.Ep.#2, H.Ep.#3 and H.Emb.
#1. Cancer Res., 14: 660-666, 1954.
28. Ossowski, L., and Reich, E. Experimental model for quantitative study of
metastasis. Cancer Res., 40: 2300-2309, 1980.
29. Ossowski, L., and Reich, E. Changes in malignant phenotype of a human
carcinoma conditioned by growth environment. Cell, 33: 323-333, 1983.
30. Moore, G. B., Merick, S. B., Woods, L. K., and Arabasz, N. M. A human
squamous cell carcinoma cell line. Cancer Res., 35: 2684-2688, 1975.
31. Giard, D. J., Aaronson, S. A., Todaro, G. J., Arnstein, P. J., Kersey, J. H.,
Dosik, H., and Parks, W. P. In vitro cultivation of human tumors: establishment of cell lines derived from a series of solid tumors. J. Natl. Cancer Inst.,
51: 1417-1423, 1973.
32. Stoppelli, M. P., Tacchetti, C., Cubellis, M. V., Corti, A., Hearing, V. J.,
Cassani, G., Apella, E., and Blasi, F. Autocrine saturation of pro-urokinase
receptors on human A431 cells. Cell, 45: 675-684, 1986.
33. Mira y Lopez, R., and Ossowski, L. Hormonal modulation of plasminogen
activator: an approach to prediction of breast tumor responsiveness. Cancer
Res., 47: 3558-3564, 1987.
34. Herron, G. S., Banda, M. J., Clark, E. J., Gavrilovic, J., and Werb, Z.
Secretion of metalioproteinases by stimulated capillary endothelial cells. II.
Expression of collagenase and stromelysin activities is regulated by endogenous inhibitors. J. Cell Biol., 261: 2814-2819, 1986.
35. DeClerck, Y. A., Perez, N., Shimada, H., Boone, T. C., Langley, K. E., and
Taylor, S. M. Inhibition of invasion and metastasis in cells transfected with
an inhibitor of metalloproteinases. Cancer Res., 52: 701-708, 1992.
36. Johnson-Wint, B. A quantitative collagen film collagenase assay for large
number of samples. Anal. Biochem., 104: 175-181, 1980.
37. Birkedal-Hansen, H., and Dano, K. A sensitive collagenase assay using [all]collagen labeled by reaction with pyridoxal phosphate and [3H]borohydride.
Anal. Biochem., 115: 18-26, 1981.
38. Ossowski, L., and Belin, D. Effect ofdimethyl sulfoxide on human carcinoma
cells, inhibition of plasminogen activator synthesis, change in cell morphology, and alteration of response to cholera toxin. Mol. Cell. Biol., 5: 355233559, 1985.
39. Fairbarn, S., Gilbert, R., Ojakian, G., Schwimmer, R., and Quigley, J. P. The
extracellular matrix of chick embryo fibroblasts and its proteolytic degradation by the transformants. J. Cell Biol., 101: 1790-1798, 1985.
40. lngber, D. E., Madri, J. A., and Folkman, J. A possible mechanism for
inhibition of angiogenesis by angiostatic steroids: induction of capillary basement membrane dissolution. Endocrinology, 119:1768-1775, 1986.
41. Quigley, J. P., Gold, L. I., Schwimmer, R., and Sullivan, L. M. Limited
cleavage of cellular fibronectin by plasminogen activator purified from transformed cells. Proc. Natl. Acad. Sci. USA, 84: 2776-2780, 1987.
42. Kruithof, E. K. O. Plasminogen activator inhibitors--a review. Enzyme
(Basel), 40: 113-121, 1988.
43. Barsky, S. H., Togo, S., Garbisa, S., and Liotta, L. A. Type IV collagenase
immunoreactivity in invasive breast carcinoma. Lancet, 8319: 296-297,
1983.
44. Liotta, L. A., Tryggvason, K., Garbisa, S., Hart, I., Folty, S. M., and Shafie,
S. Metastatic potential correlates with enzymatic degradation of basement
membrane collagen. Nature (Lond.), 284: 67-68, 1980.
45. Garbisa, S., Pozzatti, R., Muschel, R. J., Saffiotti, U., Ballin, M., Goldfarb,
R. H., Khoury, G., and Liotta, L. A. Secretion of type IV collagenolytic
protease and metastatic phenotype: induction by transfection with c-Ha-ras
but not with c-Ha-ras plus Ad2-Ela. Cancer Res., 47: 1523-1528, 1987.
46. Spinucci, C., Zucker, S., Wiemar, J. M., Lysik, R. M., Imhof, B., Ramamurthy, M., Liotta, L. A., and Nagasi, H. Purifcation of a gelatin degrading type
IV collagenase secreted by ras oncogene-transformed fibroblasts. J. Natl.
Cancer Inst., 80: 1416-1420, 1988.
47. Bernhard, E. J., Muschel, R. J., and Hughes, E. N. Mr 92,000 gelatinase
release correlates with the metastatic phenotype in transformed rat embryo
cells. Cancer Res., 50: 3872-3877, 1990.
48. Yamagata, S., Tanaka, R., Ito, Y., and Shimizu, S. Gelatinase of murine
metastatic tumor cells. Biochem. Biophys. Res. Commun., 158: 228-234,
1989.
49. Wilhelm, S. M., Eisen, A. Z., Teter, M., Clark, S. D., Kronberger, A., and
Goldberg, G. I. Human fibroblast coilagenase: glycosylation and tissue-specific levels of enzyme synthesis. Proc. Natl. Acad. Sci. USA, 83: 3756-3760,
1986.
50. Quigley, J. P., Berkenpas, M. B., Aimes, R. T., and Chen, J. M. Serine
protease and metalloprotease cascade system involved in pericellular proteolysis. Cell Differ. Dev., 32: 263-276, 1990.
51. He, C., Wilhelm, S. M., Pentland, A. P., Marmer, B. L., Grant, G. A., Eisen,
A. Z., and Goldberg, G. I. Tissue cooperation in a proteolytic cascade activating human interstitial collagenase. Proc. Natl. Acad. Sci. USA, 86: 26322636, 1989.
52. Khokha, R., Waterhouse, P., Yagel, S., Lala, P. K., Overall, C. M., Norton,
G., and Denhardt, D. T. Antisense RNA induced reduction in murine TIMP
levels confers oncogenicity on Swiss 3T3 cells. Science (Washington DC),
243: 947-950, 1989.
53. Ponton, A., Coulombe, B., and Skup, D. Decreased expression of tissue
inhibitor of metalloproteinases in metastatic tumor cells leading to increased
levels of collagenase activity. Cancer Res., 51:2138-2143, 1991.
6760
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1992 American Association for Cancer
Research.
Invasion of Connective Tissue by Human Carcinoma Cell
Lines: Requirement for Urokinase, Urokinase Receptor, and
Interstitial Collagenase
Liliana Ossowski
Cancer Res 1992;52:6754-6760.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/52/24/6754
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 15, 2017. © 1992 American Association for Cancer
Research.