Inhibition of in Vitro Ovarian Cancer Cell Invasion by Modulation of Urokinase-
type Plasminogen Activator and Cathepsin BHiroshi Kobayashi,1 Hidekazu Ohi, Motoi Sugimura, Hiromitsu Shinohara, Toshiro Fuji!, and Toshihiko Terao
Department of Obstetrics and Gynecology, Hamamatsu University School of Medicine, Handacho 3600, Hamamatsu, Shizuoka, 431-31, Japan
ABSTRACT
HOC-I ovarian cancer cells express the single-chain form of theurokinase-type plasminogen activator (uPA) and cathepsin B (calli B) ontheir cell surface. The significance of the expression of cell surface uPA/cath B activity to the invasive potential was examined by preincubatingwith uPA/cath B-modulating agents in in vitro invasion assay. The anti-uPA monoclonal antibody 394 effectively inhibited invasion in a dose-dependent manner. On the contrary, anti-cath B antibody did not affectthe invasive potential of the cells. 1-6-1. a specific inhibitor for cysteineproteases, blocked invasion as effectively as monoclonal antibody 394.The data reveal that the uPA and cysteine proteases contribute significantly to the invasive capacity of the cells. We suggest that the cysteineproteases facilitate the action of uPA, possibly by activating proenzymeuPA produced by cancer cells. Evidence for the role of a cathepsin-uPAactivation cascade in HOC-I cell invasion is provided.
INTRODUCTION
Proteolytic enzymes such as collagenase, cathepsins, plasmin,or uPA: have been implicated in tumor invasion and metastasis
formation (1,2). It has been shown in a number of experimentalmodels that tumor cells elaborate enzymes that can degradecomponents of the extracellular matrices and basement membranes (1). Cath B is considered to be one such enzyme involvedin tumor invasion and metastasis (3-5). Cath B, which iselevated in tumors compared with benign tissues, plays a regulatory role in type IV collagen degradation, since it can convertinactive procollagenase type IV to its active form (6, 7). Severalstudies have demonstrated that cath B can degrade extracellularmatrix and has been correlated with cellular invasion (3-5).Recently we demonstrated that cath B can efficiently convertthe soluble or tumor cell receptor-bound form of the pro-uPAto enzymatically active two-chain uPA (7). Thus, the cellularproteinase cath B may substitute for the plasma proteinaseplasmin in the activation of pro-uPA released by cancer cells.In addition, a role for uPA in regulating tumor cell invasivenesshas been proposed on the basis of the generally increased uPAactivity in several metastatic tumors. Thus, uPA may also beinvolved in tumor invasion and metastasis (1). Various testsystems have been devised to study the mechanism of invasionand metastasis as well as inhibition of experimental metastasisin vitro, suggesting that specific antibodies and/or proteinaseinhibitors prevented cancer cell invasion to chick chorioallan-toic membrane (8-10), human amnion (11), and reconstitutedbasement membranes (12-17).
In the present study, we demonstrate the presence of cellsurface-localized uPA and cath B on HOC-I ovarian cancercells and inhibition of invasive potential of the cells by anti-
Received2/7/92;accepted4/24/92.The costs of publication of this article were defrayed in part by the payment
of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.
' To whom requests for reprints should be addressed.2The abbreviations used are: uPA, urokinase-type plasminogen activator; cath
uPA/cath B antibodies and specific inhibitors for cysteine pro-teinases and serine proteinases, having examined their role inan in vitro Matrigel assay. The results of this study suggest that(a) proteolysis of extracellular matrix components by cell surface-localized uPA is a critical step during the process of tumorcell invasion through basement membranes and (b) cath B mayplay a role indirectly in cancer invasion, not through tissuematrix degradation. Accordingly, the results support our hypothesis that membrane-associated cath B may be importanton the activation of pro-uPA produced by cancer cells.
MATERIALS AND METHODS
Cells and Culture. An ovarian cancer cell line, HOC-I, was establishedfrom a recurrent region of ovarian endometrioid carcinoma (18). Thecells were maintained under an atmosphere of 5% CO2 in RPMI 1640supplemented with 10% fetal calf serum.
Antibodies and Protease Inhibitor. Antibodies used include the uPA-specific moAB 377, reacting with the A-chain of pro-uPA/uPA, andmoAB 394, reacting with the B-chain of pro-uPA/uPA (AmericanDiagnostica Co.). The increasing concentrations of moAB 394 progressively inhibited uPA enzymatic activity. In contrast, moAB 377 did notsignificantly affect the uPA activity (data not shown). Sheep polyclonalantibody and peroxidase-conjugated antibody raised against humancath B were purchased from The Binding Site, Ltd. (Birmingham,England) and Chemicon International, Inc. (Temecula, CA), respectively. Anti-cath B antibody inhibits cath B activity (data not shown).Preimmune murine or sheep serum was used as a control in an in vitroinvasion assay. E-64 and E-ACA were used as specific inhibitors forcath B and for serine proteinase, respectively (Sigma). E-64, which isineffective against metallo- and serine proteases, is an irreversibleinhibitor of cysteine proteases such as cath B, cath H, and cath L andpapain. Trasylol was a gift from Bayer Werk, Germany. All otherreagents were of analytical grade.
Enzyme-linked Immunosorbent Assay. Cells were grown to semicon-fluence in 96-well microtiter plates (Costar, Cambridge, MA), washedthree times with PBS containing 2% BSA, and then reacted with uPA-and cath B-specific antibodies (Table 1). Nonspecific binding wasascertained using reactions with murine or sheep immunoglobulins asa control. Specifically bound uPA antibodies were detected with biotin-conjugated second antibodies (1.5 ¿ig/ml;DAKO; 1 h, 23°C),followedby the addition of avidin-peroxidase (0.4 Mg/ml, DAKO; 1 h, 23°C).
Peroxidase-conjugated anti-cath B polyclonal antibody was used todetermine cath B molecules immunologically (3 h, 23°C).The reactionswith tetramethylbenzidine were terminated after 10 min at 23°C,and
the AW of each well was measured in an enzyme immunoassay reader(Model 2550; Bio-Rad, Richmond, CA) (7).
Enzyme Assay. Chromogenic assays for uPA activity and cath Bactivity were carried out as detailed. Briefly, 150 n\ of the cell suspension(1 x IO6cells/ml complete medium) are added to each well of microtiterplates. A 5-day incubation period resulted in over 90% adherent cells.After three washings with PBS-0.1% BSA, each well was incubatedwith the uPA-specific chromogenic synthetic substrate S-2444(KabiVitrum) or the cath B-specific substrate Z-Arg-Arg-2-naphthy-lamide (Peptide Research Foundation, Osaka, Japan) (19) for differentlengths of time. To demonstrate the specificity of cysteine proteaseactivity, cells were preincubated with the inhibitor E-64, and thensubstrate was added (20).
Indirect Immunofluorescence Assay for uPA by Flow Cytometry. Forthe assay, the cells were washed with PBS-0.1 % BSA, then 1 miviEDTA
Table 1 Immunological determination ofuPA and cath B on the surface ofHOC-1 ovarian cancer cells by cellELISAuPA-specific
antibodyTreatmentsControl"
After saturation'Acid treatment'
Acid treated + saturationmoAB
3770.11±0.04"
0.29 ±0.07<0.01
0.30 ±0.05moAB
3940.16
±0.030.41 ±0.09
<0.010.40 ±0.08Cath
Bantibody0.20
±0.05NT'
NTNT
" HOC-1 cells were washed with PBS-0.1% BSA.* Mean ±SD of Atv>of assays carried out in triplicate.'The cells were incubated with purified, unlabeled HMW-uPA (5 n\t) under
conditions that achieve saturable binding.äNT, not tested.' The cells were treated with mild acidic buffer under conditions that have been
shown to dissociate polypeptide ligands from their receptors.
was added (trypsin is not used because of degradation of uPA receptors),and the cells were detached with Vortex mixer apparatus. The cellswere then collected by centrifugation (1200 rpm, 10 min) and washedtwice. The cells were dispersed by pipeting until a single cell suspensionwas obtained. An aliquot of sample was taken, and cell numbers andviability by trypan blue exclusion were determined. The cells weresuspended in PBS-0.1% BSA at a concentration of 1 x IO6cells/ml inthe absence or presence of 5.0 ni\i HMW-uPA (30 min, 4'C). A 2.0 n\i
concentration of HMW-uPA is enough to saturate cell surface uPAreceptors (data not shown). Also, cells were treated with 50 m.Mglycine-HCI-0.5 M NaCl, pH 3.0, first to dissociate receptor-bound uPA andthen were neutralized with 4-(2-hydroxyethyl)-l-piperazineethanesul-fonic acid buffer (21). The cell pellet was resuspended in PBS-0.1%BSA containing 5.0 nM HMW-uPA. After two washings with PBS-0.1% BSA, uPA-specific antibodies (moABs 377 and 394; 5.0 Mg/ml)were applied to cell suspension (1 h, 4°C),followed by the addition of
fluorescein isothiocyanate-conjugated anti-mouse secondary antibody(1 h, 4°C).Cell-associated fluorescence was measured with the EPICS
PROFILE flow cytometer (Coulter).Immunoadsorbent-amidolytic Assay of Pro-uPA and HMW-uPA in
either with 150 M!of 5.0 Mg/ml plasmin solution or with buffer alone.After incubation (1 h, 37'C) the plasmin-catalyzed reaction was blocked
by adding 25 /j'/well of a 0.35-Mg/ml aprotinin solution. Then 25 M'/well of chromogenic synthetic substrate S-2444 (1.5 mg/ml) wereadded. The color was allowed to develop (0-60 min, 37"C) and read at
405 nm in an EIA reader. The total uPAs and HMW-uPA concentrations were obtained by interpolating the absorbance of wells treatedand untreated with plasmin on the dose-response curve of standards.The concentration of pro-uPA was obtained by subtracting HMW-uPAfrom the total uPAs (22).
Cell Attachment Assay. The fibronectin solution (10 Mg/ml PBS) wasincubated in 96-well microtiter-plates (2 h, 23°C).The unattached
protein was removed by three washings with PBS (23). To ensure thecells in the microtiter well, 100 M' of RPMI 1640 in the absence orpresence of uPA/cath B-modifying agents (antibodies and proteaseinhibitors; see Figs. 3 and 4) were added to each well (1 h, 23°C).Thisfollowed by the addition of 100 M!of the cell suspension (1 x 10*cells/ml). After incubation (2 h, 37°C),the unattached cells were removed
simply by discarding the medium and washing. The plate is floodedwith PBS, and the washing solution is discarded. Under optimal conditions, about 70% of these cells remain attached to the well afterwashing. The cells were then fixed with 3% paraformaldehyde in PBSand stained with 1% toluidine blue-3% formaldehyde in PBS, andattached cells were counted with an Olympus light microscope (x 200).The data are expressed as the number of attached cells at a randomlychosen area (each sample was assayed using triplicate wells and eachwell was counted at three areas).
Cell Invasion Assay. Tumor cell invasiveness was determined usinga modification of the membrane invasion culture system (12-17).Briefly, polycarbonate filters, 8-Minpore size (Costar), were coated with
an extract of basement membrane components (Matrigel; 50 Mg/fi'ter,1.2 Mg/mnr; Collaborative Research Co.,) and dried (2 h, 37°C).A
blind well Boyden chamber were filled with 600 M!RPMI 1640-0.1%BSA in the lower compartment and coated filters were mounted in thechamber. The cells to be studied were collected by short exposure toEDTA (1 HIM),resuspended in RPMI 1640-0.1% BSA and shaken asdescribed above. Wells were pretreated with uPA/cath B-modifyingagents (at the concentrations noted) as described in Figs. 3 and 4,typically for 60 min at 23°Cin serum-free medium, with occasionalmixing, and then 200 M!cell suspension (1 x IO*1cells/ml) were placed
in the upper compartment of the chamber (22). The plates were incubated at 37°Cin a 5% CO; atmosphere saturated with H:O for 24 h.
Human fibroblast-conditioned media and fibronectin (25 Mg/ml) wereused as chemoattractants in the lower compartment of the chamber.Comparable results were obtained using either source of chemoattractants. The different behaviors with the antibodies and inhibitors of theserine and cysteine proteinases were not obtained. In the present studies,fibroblast-conditioned media were placed in the lower compartment asa source of chemoattractants. Under this condition, few cells died within24 h as measured by trypan blue exclusion in preliminary experiments.
We assessed the penetration of HOC-I cells as a function of time.Few cells were found to penetrate through the Matrigel during the first6 h of the assay. After 12 h, many more cells were observed to penetrate(data not shown). At the end of the incubation, the cells on the uppersurface of the filter were completely removed by wiping with a cottonswab, as monitored visually under a light microscope (x 200). Cellsfrom various areas of the lower surface were counted and each assaywas carried out in triplicate.
Chemotactic Assay. The chemotactic assay was conducted in a similarfashion without coating of Matrigel (15, 16). After incubation (24 h,37"C) the filters were then removed, fixed, and stained. All materials
from the upper surface of the filter was carefully removed by scrapingwith a cotton tip, and invasive cells adhering to the lower surface of thefilter were quantitated with a light microscope (x 200).
The statistical significance of the results was calculated using thenonparametric Wilcoxon test.
RESULTS
To argue the evidence for the involvement of uPA and cathB in invasive phenomena, we have focused our attention oncancer cells that produce uPA and cath B and investigatedwhether inhibition of uPA or cath B diminishes the invasivebehavior.
Immunological and Enzymological Determination of uPA andCath B on the Surface of HOC-I Ovarian Cancer Cells. ELISAusing different uPA-specific monoclonal antibodies detects surface membrane-localized (receptor-bound) uPA, since the cellsadhered on the microtiter plate without being permeabilized.uPA on the surface of HOC-I cells was produced only in thelogarithmic phase of cell growth, whereas once the cells reachedconfluence, significantly fewer cells expressed and produceduPA. Also, most of the cell-associated uPA is on the cell surface,as shown by indirect immunofluorescence (data not shown).Our studies were performed in subconfluence.
HOC-I cells showed detectable, reproducible levels of expression of surface uPA by cell ELISA (Table 1). After the cellswere incubated with purified, unlabeled HMW-uPA under conditions that achieve saturable binding, the cell surface-associated uPA activity was measured immunologically. HOC-I cellssaturated with HMW-uPA showed more specific binding. Wetreated HOC-I cells with mild acidic buffer under conditionsthat have been shown to dissociate polypeptide ligands fromtheir receptors (21). Specific immunoreactivity is drastically
Table 2 Immunological determination ofuPA and cath B on the surface ofHOC-1 ovarian cancer cells by flow cytometry
uPA-specificantibodyTreatmentsControl"
After saturation'Acid treatment-^Acid treated + saturationmoAB
37720.3±5.6*
46.3 ±10.0''
9.4 ±3.947.2 ±14.1moAB
39447.9
±9.171.9 ±I2.(,d
10.6 ±4.880.3 ±11.7Cath
Bantibody24.7
±6.9NT'
NTNT
°HOC-I cells were washed with PBS-0.1% BSA.h Mean ±SD of fluorescence mean channel of assays carried out in triplicate.' The cells were incubated with purified, unlabeled HMW-uPA (5 HM)under
conditions that achieve saturable binding.'The binding of moAB 377 to pro-uPA/uPA is impeded by a steric hindrance
to some extent.' NT, not tested.'The cells were treated with mild acidic buffer under conditions that have been
shown to dissociate polypeptide ligands for their receptors.
decreased following acid treatment. HMW-uPA was able tobind to the uPA receptor uncovered by the acid treatment. Acid-treated cells acquired the property of binding uPA. Also, HOC-I cells showed significantly reproducible levels of expression ofcath B on their cell surface. Furthermore, flow cytometricanalysis also indicated that uPA is not an integral membraneprotein but is bound to a specific surface receptor that is notcompletely saturated (Table 2). A mild acid treatment uncoversthe surface receptors by dissociating uPA.
For a quantitative assay of uPA activity on the surface ofHOC-I cells, we have applied a chromogenic synthetic substrate, S-2444, to the cells adhering to the microtiter plates(Fig. 1). In the absence of cells, the negative control showed noenzymatic activity during the course of the assay. HOC-I cellsalone were capable of degrading S-2444. When HOC-I cellswere preincubated with plasminogen, considerably elevated levels of amidolytic activity was observed. Commercially availableplasminogen is thought to contaminate a trace amount ofplasmin (24). In the presence of HOC-I cells and plasmin, therewas a significant increase in amidolytic activity. The inhibitoryanti-uPA moAB 394 reduced the amidolytic activity by about70%, when the cells were preincubated with moAB 394. mo AB377 has essentially no inhibitory effect.
Pro-uPA and HMW-uPA levels in HOC-I cell supernatantswere measured by immunoadsorbent-amidolytic assay (22) (Fig.2). The significant increase in the amidolytic activity afteractivation of uPAs by plasmin suggests that more than one-half(~60%) of the total immunoreactive uPAs in HOC-I cell supernatants is enzymatically inactive pro-uPA.
Effect of Selective Antibodies and Protease Inhibitors on Cancer Cell Invasion. If uPA and cath B are essential to the invasionprocess of cancer cells, an inhibition of invasion would beobserved by selective antibodies or proteinase inhibitors. Theinvasiveness of the cells and its sensitivity to anti-uPA/cath Bantibodies and cysteine proteinase inhibitor as well as serineproteinase inhibitors were assessed using the in vitro invasionassay. Experiments with moAB 394 show a statistically significant inhibition of invasion, while moAB 377 has essentially noeffect. Of the two anti-uPA antibodies, only anti-uPA antibodyinhibiting enzymatic activity of uPA was effective. Nonimmuneserum did not inhibit basement membrane invasion. An attemptwith polyclonal anti-cath B antibody shows no significant effects. In comparison, the cysteine proteinase inhibitor (E-64;50 Mg/ml) and serine proteinase inhibitors [trasylol (500 units/ml) and E-ACA (500 Mg/ml)] reduced the invasion by about70-80%, showing inhibition of the cell invasion as effectivelyas mo AB 394 (100 Mg/ml). The concentrations used are notcytotoxic (11). The inhibitory effect of the anti-uPA antibody
(moAB 394) and of the serine proteinase inhibitors was dosedependent.
To examine the accessibility of antiserum and inhibitor tothe cellular enzymes involved in tumor invasion, we addeduPA/cath B-modifying agents under another condition described by Mignatti et al. (11). The antiserum and inhibitorwere added to the cell suspension before the cells were seededinto the culture wells. The results obtained showed that moAB394, E-64, Trasylol, and E-ACA effectively inhibited invasion(data not shown). The above mentioned agents appeared capable of blocking invasion when they were added with the cellsand when they were preadsorbed to the basement membrane(Fig. 3).
The chemotactic response of the cells was also tested todetermine if the inhibition of the invasive properties of the cellswas due to their inhibition of this response by antibodies orproteinase inhibitors (Fig. 4). HOC-I cells tested here showeda good chemotactic response in the presence of modulatingreagents. These results indicate that these modulating reagentshave no effects on cancer cell chemotaxis, because the cancercells are able to migrate to chemoattractants even in the presence of antibodies or proteinase inhibitors.
0.3-
0.2-
0.1-
15 30Time (min)
60
Fig. 1. Quantitative detection of uPA activity in HOC-I cells. The effects ofmodulation of uPA activity were investigated using chromogenic synthetic substrate S-2444. HOC-I cells were preincubated with agents noted below (1 h, 23°C),
and the cells were washed twice. Assays were routinely carried out in triplicate.O, buffer alone; D, HOC-I cells alone; •¿�,HOC-I cells plus plasmin (0.5 jig/ml);•¿�HOC-I cells plus plasminogen (15 ¿ig/ml);A, HOC-I cells plus plasmin plusanti-uPA IgG (moAB 394, 100 i<g/ml); A, HOC-I cells plus plasmin plus anti-uPA IgG (moAB 377, 100 ^g/ml). Bars, SD.
0.5-1
0.4-
0.3-
0.2-
0.1
15 30Time (min)
60
Fig. 2. Differential detection of pro-uPA and HMW-uPA in HOC-I cellsupernatants. The effects of plasmin treatment of uPA, which was contained inHOC-I cell supernatants, bound to moAB 377(15 fig/ml)-coated microtiter plateswere examined. Amidolytic activity of HOC-I cell supernatants (150 nl) treatedwith 5 *ig/ml (•)or 0 /-u nil (A) plasmin solution; O, amidolytic activity ofHMW-uPA (10 nM) treated with plasmin (5 /ig/ml). Bars, SD.
Fig. 3. Effect of selective antibodies and inhibitors on in vitro basement membrane invasionby HOC-I cells. E-64 (0.5, 5. and 50 ^g/ml). ahighly specific inhibitor for cysteine proteases,was added to the invasion wells 60 min beforethe cells were plated on the Matrigel. EachmoAB (1. 10. and 100 vg/ml) was treated asdescribed in "Materials and Methods." For each
experiment three replicates were performed. Acontrol experiment was carried out in the absence of any antibodies and inhibitor. Columns,mean of three experiments; ears. SD. Similarresults were obtained when these agents wereadded to the cell suspension before the cells wereseeded into the culture wells. Trasy.. Trasylol.
Fig. 4. Effect of selective antibodies and inhibitors on the chemotactic response and cellattachment of HOC-I cells. D, chemotactic response (cells/field; °t);•¿�.cell attachment (at
tached cells/field; %); C, control medium: P.preimmune serum (100 »jg/ml).Columns, meanof three experiments; bars, SD.
anti-uPA moABanti-cith B poAB E-64 Trasy E-ACA
Furthermore, the effects of modulating reagents used in theexperiments on cell attachment had been studied. Little or noinhibition of attachment to fibronectin coated on wells was seenwith modulating reagents. The results for the cells and antibodies or proteinase inhibitors are shown in Fig. 4.
DISCUSSION
In this paper we have shown that HOC-I ovarian cancer cellshave a surface-associated uPA activity and cath B activity andthat inhibition of these activities by specific antibodies andprotease inhibitors leads to a decrease in the invasive potentialin an in vitro invasion assay. The reconstituted basement membranes provide an excellent model system for the study ofinvasion (12-17, 25). To detect and to determine the characteristics of cell-associated uPA and cath B, several means includingcell ELISA, flow cytometric analysis, and immunoadsorbent-amidolytic assay were carried out. In fact, our studies indicatethat (a) HOC-I cells have uPA molecules on their cell surfaceand most of the cell-associated uPA on the cell surface isenzymatically inactive pro-uPA, (b) pro-uPA is bound to aspecific surface receptor that is not completely saturated, (c) itsactivity can be inhibited by anti-uPA moAB 394, and (d) thecells have enzymatically active cath B on their cell surface.
The ability of antibodies against uPA and serine proteinaseinhibitors to inhibit invasion of the basement membranes byHOC-I cells implicates uPA in the invasive process of the cells.The fact that moAB 377, which does not inhibit uPA activity,did not abrogate invasion to a greater extent than moAB 394,effectively blocking uPA activity, indicates that the inhibitionof uPA activity is contributing to the invasive process.
The cysteine proteinase inhibitor E-64 used in this studyreduced the invasiveness of the cells tested by some 70%. The
fact that anti-cath B polyclonal antibody manifesting inhibitionof cath B activity did not abrogate invasion to a greater extentthan E-64 argues that cath B is not contributing more to theinvasive phenotype than other cathepsins such as cath L. Theability of E-64, which is thought to be a strong inhibitor of cathB and cath L (20), to inhibit the invasion strongly suggests amajor role for cath L in facilitating membrane penetration (26-28). Furthermore, the incompleteness with which the inhibitinguPA antibody and E-64 block invasion indicates that uPA andcysteine proteinases are not sufficient in invasion mechanisms,suggesting that HOC-I cells have a significant plasmin independent protease activity. We speculate that in HOC-I cells,cath B may play a role indirectly in cancer cell invasion, notthrough tissue matrix degradation, supporting our hypothesisthat membrane-associated cath B may be important in theconversion of pro-uPA produced by cancer cells to enzymatically active HMW-uPA (7).
A possible correlation between metastatic potential and theextent of enzymatic degradation of collagen type IV has beenreported for variants of the B16 melanoma cell lines (29) andfor rat mammary adenocarcinoma cell lines (26, 30). Also,inhibitors of metalloproteinases and serine proteinases reportedly significantly reduce the penetration of the amnionic membrane by invasive cells (11,31). Although a positive correlationis observed between type IV collagenase activity and tumor cellinvasion, the secretion and activation of metalloproteinases arenot enough to degrade the tumor stromas and extracellularmatrix as well as basement membranes (32, 33). The balancebetween activated type IV collagenase and endogenous inhibitor(tissue inhibitor of metalloproteinases 2) is thought to be acritical determinant of tumor cell invasiveness. Albini et al. (33)reported that addition of free tissue inhibitor of metalloproteinases 2 or specific antiserum against the M, 72,000 type IV
collagenase inhibited HT-1080 cell invasion, suggesting thatmetalloproteinases, in particular the M, 72,000 type IV collagenase, are critical for tumor cell invasion of the reconstitutedbasement membranes.
Our studies have not addressed the expression and the production of metalloproteinases in cultured ovarian cancer cells.In addition, we did not examine additional evidence for theparticipation of metalloproteinases. Caution should be exercised in drawing conclusions as to the invasive behavior ofovarian cancer cells. Notwithstanding these limitations, thepresent study strongly argues for a role of cell-associated uPAand cysteine proteinases in facilitating the in vitro invasion ofovarian cancer cells.
ACKNOWLEDGMENTS
We thank Dr. Lothar Goretzki (Frauenklinik der Technischen Universität,Klinikum rechts der Isar, Munich, Germany) for helpful discussions on cath B and L and Dr. Kazuhiro Sumimoto for insight intodata interpretation.
REFERENCES
1. Daño,K., Andreasen, P. A., Grondahl-Hansen, J., Kristensen, P. I., Nielsen,L. S., and Skriver, L. Plasminogen activators, tissue degradation, and cancer.Adv. Cancer Res., 44: 139-266, 1985.
2. ÑutÃa.L. A., Rao, C. N., and Barsky, S. H. Tumor invasion and theextracellular matrix. Lab. Invest., 49: 636-649, 1983.
3. Sloane, B. F., Honn, K. V., Sadler, J. G., Turner, W. A., Kimpson, J. J., andTaylor. J. D. Cathepsin B activity in B16 melanoma cells: a possible markerfor metastatic potential. Cancer Res., 42: 980-986, 1982.
4. Rozhin, J., Robinson, D., Stevens, M. A., Lah, T. T., Honn, K. V., Ryan, R.I .. and Sloane, B. F. Properties of a plasma membrane-associated cathepsinB-like cysteine proteinase in metastatic B16 melanoma. Cancer Res., 47:6620-6628, 1987.
5. Keren, Z., and LeGrue. S. J. Identification of cell surface cathepsin B-likeactivity on murine melanomas and fibrosarcomas: modulation by butano!extraction. Cancer Res., 48: 1416-1421, 1988.
6. Eckhout, Y., and Vaes, G. Further studies on the activation of procollagenase,the latent precursor of bone collagenase: effect of lysosomal cathepsin B,plasmin, and kallikrein, and spontaneous activation. Biochem. J., 166: 21-31, 1977.
7. Kobayashi, H., Schmitt, M., Goretzki, L., Chucholowski, N., Calvete, J.,Kramer, M., Gunzler, W. A., Janicke, F.. and Graeff, H. Cathepsin Befficiently activates the soluble and the tumor cell receptor-bound form ofthe proenzyme urokinase-type plasminogen activator (pro-uPA). J. Biol.Chem., 266: 5147-5152, 1991.
8. Ossowski, L. In vivo invasion of modified chorioallantoic membrane by tumorcells: the role of cell surface-bound urokinase. J. Cell Biol., 707: 2437-2445,1988.
9. Ossowski, L., and Reich, E. Antibodies to plasminogen activator inhibithuman tumor metastasis. Cell, 35: 611-619, 1983.
10. Ossowski, L., and Reich, E. Experimental model for quantitative study ofmetastasis. Cancer Res., 40: 2300-2309, 1980.
11. Magnatti, P., Robbins, E., and Rifkin, D. B. Tumor invasion through thehuman amniotic membrane: requirement for a proteinase cascade. Cell, 47:489-498, 1986.
12. Gehlsen, F. R., Wagner, H. N., and Hendrix, M. J. C. Membrane Invasion
Culture System (MICS). Med. Instrum., ¡S:268-271, 1984.13. Kleinman, H. K., McGarvey, M. L., Hassell, J. R., Star, V. L., Cannon, F.
B., Laurie, G. W., and Martin, G. R. Basement membrane complexes withbiological activity. Biochemistry, 25: 312-318, 1986.
14. Terranova, V. P., Hujanen, E. S., and Martin, G. R. Basement membraneand the invasive activity of metastatic tumor cells. J. Nati. Cancer Inst., 77:311-316,1986.
15. 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 for quantitatingthe invasive potential of tumor cells. Cancer Res., 47: 3239-3245, 1987.
16. Reich. R., Thompson, E. W., Iwamoto, Y., Martin, G. R., Deason, J. R.,Fuller, G. C., and Miskin, R. Effects of inhibitors of plasminogen activator,serine proteases, and collagenase IV on the invasion of basement membranesby metastatic cells. Cancer Res., 48: 3307-3312, 1988.
17. Gehlsen, K. R., Argraves, W. S., Pierschbacher, M. D., and Ruoslahti, E.Inhibition of in vitro cell invasion by Arg-Gly-Asp-containing syntheticpeptides. J. Cell Biol., 106: 925-930, 1988.
18. Fujii, T. Establishment and characterization of human ovarian endometrioidcarcinoma cell line. Acta Obstet. Gynaecol. Jpn., 41: 161-166, 1989.
19. Barrett, A. J., and Kirschke, H. Cathepsin B, cathepsin H, and cathepsin L.Methods Enzymol., 80: 535-561, 1981.
20. Barrett, A. J., Kembhavi, A. A., Brown, M. A., Kirschke, H., Knight, C. G.,Tamai, M., and Honada, K. l-Iranj-Epoxysuccmylleucylamido(4-guani-dino)butane (E-64) and its analogues as inhibitors of cysteine proteinasesincluding cathepsin B, H and L. Biochem. J., 201: 189-198, 1982.
21. Stoppelli, M. P., Tacchetti, C., Cubellis, M. V., Corti, A., Hearing, V. J.,Cassani, G., Appella, E., and Blasi, F. Autocrine saturation of pro-urokinasereceptors on human A431 cells. Cell, 45:675-684, 1986.
22. Corti, A., Nul li, M. L., and Cassani, G. Differential detection of single-chainand two-chain urokinase-type plasminogen activator by a new immunoadsor-bent-amidolytic assay (IAA). Thromb. Haemost., 56:407-410, 1986.
23. Ruoslahti, E., Hayman, E. G., Pierschbacher, M., and Engvall, E. Fibronec-tin: purification, immunochemical properties, and biological activities. Methods Enzymol., 82: 828-831, 1982.
24. Hearing, V. J., Law, L. W., Corti, A., Appella, E., and Blasi, F. Modulationof metastatic potential by cell surface urokinase of murine melanoma cells.Cancer Res., 48: 1270-1278, 1988.
25. Cajot, J. F., Schleuning, W. D., Medcalf. R. L., Bamat, J., Testuz, J.,Liebermann, L., and Sordat. B. Mouse L cells expressing human prouroki-nase-type plasminogen activator: effects on extracellular matrix degradationand invasion. J. Cell Biol., 709: 915-925, 1989.
26. Yagel, S., Warner, A. H., Nellans, H. N., Lala, P. K., Waghorne, C., andDenhardt, D. T. Suppresion by cathepsin L inhibitors of the invasion ofamnion membranes by murine cancer cells. Cancer Res., 49: 3553-3557,1989.
27. Rozhin, J., Wade, R. L., Honn, K. V., and Sloane, B. F. Membrane-associatedcathepsin L: a role in metastasis in melanomas. Biochem. Biophys. Res.Commun., 164: 556-561, 1989.
28. Yamaguchi, N., Chung, S. M., Shiroeda, O., Koyama, K., and Imanishi, J.Characterization of a cathepsin L-like enzyme secreted from human pancreatic cancer cell line HPC-YP. Cancer Res., 50: 658-663, 1990.
29. Liotta, L. A., Tryggvason, K., Garbisa, S., Hart, I., Foltz, C. M., and Shafie,S. Metastatic potential correlates with enzymatic degradation of basementmembrane collagen. Nature (Lond.), 284:67-68, 1980.
30. Nakajima, M., Welch, D. R., Belloni, P. N., and Nicolson, G. L. Degradationof basement membrane type IV collagen and lung subendothelial matrix byrat mammary adenocarcinoma cell clones of differing metastatic potentials.Cancer Res., 47: 4846-4876, 1987.
31. Thorgeirsson, U. P., Liotta, L. A., Kalebic, T., Margulies, I. M., Thomas,K., Rios-Candelore, M., and Russo, R. G. Effect of natural protease inhibitorsand a chemoattractant on tumor cell invasion in vitro. J. Nati. Cancer Inst.,69: 1049-1054, 1982.
32. Liotta, L. A., and Stetler-Stevenson, W. G. Tumor invasion and metastasis:an imbalance of positive and negative regulation. Cancer Res. (Suppl.), 51:5054s-5059s, 1991.
33. Albini, A., Melchiori, A., Santi. L., Liotta, L. A., Brown, P. D., and Stetler-Stevenson, W. G. Tumor cell invasion inhibited by TIMP-2. J. Nati. CancerInst., «5:775-779, 1991.
