Tumor-Associated Trypsin Participates in Cancer Cell-Mediated Degradation of Extracellular Matrix1

(CANCER RESEARCH 51. 2107-2112, April 15. 1991] Tumor-associated Trypsin Participates in Cancer Cell-mediated Degradation of Extracellular Matrix1

Erkki Koivunen,2 Ari RistimÃ¤ki,Outi Itkonen, Sirpa Osman, Matti Vuento, and Ulf-HÃ¡kan Stenman Department of Obstetrics and Gynecology, Helsinki L'nirersity Central Hospital, Haartmaninkatu 2, SF-00290 Helsinki [E. K., A, Râ€žO. I., S. O,, U-H. SJ, and Department of Biochemistry, L'nirersity of Helsinki, L'nioninkatu 35, SF-OOI70 Helsinki [M. V.I.Finland

ABSTRACT (11), and cathepsin B (4) have been found to correlate with invasiveness and the metastatic ability of tumor cell lines. In We have recently demonstrated that many cancer cell lines produce a extracts of tumor tissues the levels of urokinase-type plasmin novel trypsinogen isoenzyme called tumor-associated trypsinogen 2 (TAT-2). It was found during a search of the target protease for tumor- ogen activator (u-PA) generally correlate with the degree of associated trypsin inhibitor (TATI). We now show that degradation of malignancy of the tumors (12). It is interesting that TAT-2, the subendothelial cell extracellular matrix (ECM) by four different cell lines major TAT isoenzyme, also shows close association with ma (COLO 205 colon carcinoma, K-562 erythroleukemia, CAPAN-1 pan lignancy; the levels are higher in malignant and borderline creatic carcinoma, and HT I080 fibrosarcoma) can be partially inhibited ovarian tumor cyst fluids than in benign ones (13). In addition, by TATI or neutralizing trypsin antibodies. When cells were cultured in purified TAT-1 and TAT-2 activate pro-u-PA in vitro (8). These serum-free medium on ECM, TATI and trypsin antibodies inhibited the results suggest that TAT-2 may promote the dissemination of release of immunoreactive fibronectin fragments from ECM by 47-54 and 40%, respectively. Degradation of isotopically labeled ([-'Miserine, tumor cells. TAT-2 zymogen is produced by many cancer cell lines, and |3H)proline, and (35S|sulfate) ECM was also significantly prevented by it has been purified from the culture medium of COLO 205 TATI. At its maximum, it exerted a 57% inhibition on the degradation of [3H|serine-labeled ECM. Plasminogen added exogenously to the cul colon adenocarcinoma and HT 1080 fibrosarcoma cell lines ture medium further potentiated the proteolysis of ECM. Interestingly, (14). In the present study we have examined the role of TAT-2 addition of enteropeptidase, an activator of TAT-2, also enhanced cell- in cell-mediated degradation of extracellular matrix. We devel mediated proteolysis as assessed by degradation of purified fibronectin oped a method for determining degradation of fibronectin based coated onto the surface of wells. Immunoblot analysis showed that on analysis of immunoreactive fibronectin fragments liberated enteropeptidase-mediated proteolysis generated a pattern of fibronectin by tumor cells growing on protein-coated surfaces. We also fragments similar to that obtained by digestion of purified fibronectin by analyzed the degradation of isotopically labeled ECM. The TAT-2. These results demonstrate the existence of a proteolytic system results show that tumor cell-mediated proteolysis can be signif in tumor cells which is dependent on the activation of TAT-2. We suggest icantly abrogated by inhibiting TAT-2. that TAT-2 is involved in a protease cascade-stimulating tumor cell invasion and degradation of extracellular matrix. MATERIALS AND METHODS INTRODUCTION Chemicals. Reagents were obtained as follows: aprotinin from Sigma Chemical Co. (St. Louis, MO); EACA from Fluka AG (Buchs, Swit Production of proteolytic enzymes by malignant tumor cells zerland); RPMI 1640, Medium 199, PBS (without magnesium and is believed to be essential to the ability of the tumor to invade calcium), and fetal bovine serum from Flow Laboratories (Irvine, Scot and degrade extracellular matrix. Of the proteases secreted by land); pooled human serum from the Finnish Red Cross Blood Trans tumor cells, collagenases (1), plasminogen activators (2), tran- fusion Service (Helsinki, Finland); L-glutamine and antibiotics from sin/stromelysin (3), and cathepsins (4-6) have been the most GIBCO Laboratories (Grand Island, NY); and gelatin from Merck thoroughly studied. We have recently characterized a novel (Darmstadt, Germany). tumor-associated protease, TAT,3 (7, 8). It was initially identi Proteinases. Human plasminogen and plasmin were obtained from fied in the cyst fluid of mucinous ovarian tumors during a Kabi Vitrum (Stockholm, Sweden). Porcine enteropeptidase and bovine trypsin were obtained from Sigma. Enteropeptidase was dissolved in search of the target protease for TATI, which is a marker for PBS and passed through an affinity column of TATI-Sepharose 4B to mucinous ovarian cancer (9). TAT occurs as two isoenzymes, remove contaminating trypsin-like enzymes. TAT-2 was purified by a TAT-1 and TAT-2. They are identical to the corresponding monoclonal antibody affinity column from serum-free culture medium pancreatic trypsins with respect to amino-terminal amino acid of COLO 205 cells as described (14). TAT-2 was obtained in zymogen sequence, molecular weight, and immunoreactivity, but they form and autoactivated by neutralizing the pH. differ with respect to substrate specificity, enzyme stability, and Purification of TATI. TATI was purified by a novel affinity chroma elution pattern in ion-exchange chromatography. At present it tography method from the urine of cancer patients. The pH of the urine is not known whether these differences are explained by tissue- was adjusted to 7.4, and it was dialyzed against water with a hollow specific modification of trypsin or whether distinct genes code fiber dialyzer (15) until the conductivity corresponded to that of 50 mM Tris, pH 7.4. Dialyzed urine was then applied at a flow rate of 60 ml/ for trypsinogens derived from tumors and pancreas. h at 4Â°Cto a Q Sepharose anion-exchange column (Pharmacia, Upps The levels of type IV collagenase (10), plasminogen activators ala, Sweden) and an affinity column of bovine trypsin-Sepharose 4B Received 7/31/90; accepted 2/4/91. previously equilibrated with Tris buffer. Most urine proteins were The costs of publication of this article were defrayed in part by the payment removed by the first anion-exchange column. After the trypsin-Sepha of page charges. This article must therefore be hereby marked advertisement in rose column was washed with Tris buffer containing l M NaCI, 0.1% accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1This work was supported by grants from the Academy of Finland, the Finnish Brij 35, and 1% 2-propanol, TATI was eluted with 0.1 % trifluoroacetic Cancer Society, the Finnish Social Security Institution, the Jenny and Antti acid. TATI was finally purified by reversed-phase chromatography on Wihuri Foundation, and the Ida Montin Foundation. a C,Â«column with a linear acetonitrile gradient (from 10 to 60% in 20 2To whom requests for reprints should be addressed. 3The abbreviations used are: TAT, tumor-associated trypsin(ogen); TATI, min) in 0.1% trifluoroacetic acid. By this method, 2 mg of TATI were tumor-associated trypsin inhibitor; u-PA, urokinase-type plasminogen activator; obtained from 1 liter of urine. ECM, extracellular matrix: EACA. Â«-aminocaproicacid; PBS. phosphate-buffered Anticatalytic Antibodies. The IgG fraction was prepared from a rabbit saline. preimmune serum and antiserum against human pancreatic trypsin 1 2107

(8) by sequential affinity Chromatograph) on Protein G-Sepharose and RESULTS protein A-Sepharose (Pharmacia). Anti-u-PA IgG was prepared from goat antiserum against human low-molecular-weight u-PA (Biopool The adherent tumor cell lines used in this study can easily be AB, UmeÃ ,Sweden) by using protein A-Sepharose. Anti-trypsin IgG detached from culture dishes by removing the serum and incu and anti-u-PA IgG were anticatalytic. because they inhibited the ami- bating the cells in PBS. When COLO 205 cells collected this dolytic activity of TAT-2 and u-PA, respectively. way were plated in serum-free medium in fibronectin-coated Cell Culture. Human cancer cell lines COLO 205 colon adenocarci- wells, immunoreactive fibronectin was released into the serum- noma (ATCC CCL 222). HT 1080 fibrosarcoma (ATCC CCL 121), K- 562 erythroleukemia (ATCC CCL 243), and CAPAN-1 pancreatic free medium as detected by fibronectin immunoassay (Fig. 1). adenocarcinoma (ATCC HTB 79) were obtained from the American Control experiments showed that fibronectin immunoreactivity Type Culture Collection (Rockville. MD). Cells were cultured in 800- was not released into the medium in the absence of cells. In ml Nunclon flasks (Roskilde, Denmark) and maintained in RPMI 1640 addition, COLO 205 cells themselves were not found to secrete supplemented with 10% heat-inactivated fetal bovine serum, 2 mM L- fibronectin. glutamine. 100 units/ml penicillin, and 100 ^g/ml streptomycin. Treat The cell-mediated release of coated fibronectin could be ment with trypsin was not required for detachment of the adherent significantly increased by adding plasminogen zymogen or ac cells, possibly because the cells produce their own trypsin. After serum tive enteropeptidase to the culture medium (Fig. 1). Fibronectin was removed, and the flasks were washed twice with PBS, cells were immunoreactivity was readily detectable after a 4-h culture and incubated for 10-15 min in PBS at room temperature. The flasks were increased further over 65 h. Control experiments indicated that vigorously shaken until all cells detached. Preparation of Wells Coated with ECM and Kibronectin. Endothelial plasminogen and enteropeptidase alone at the low concentra cells were isolated from human umbilical cord veins by the method tions used (6 and 1 Mg/ml, respectively) did not cause degrada originally described by Jaffe et al. (16), and ECM was prepared essen tion of coated fibronectin during a 24-h incubation. However, tially as described (17. 18). Endothelial cells were seeded in Nunclon fibronectin appears to be sensitive to higher concentrations of 24-well multidishes, which were previously coated with gelatin, and enteropeptidase (20 Â¿ig/ml),as demonstrated by release of fi cultured in Medium 199 containing 20% human serum. Labeling bronectin immunoreactivity from the wells (not shown). medium containing 5 Â¿iCi/inlpHlserine, 5 ^Ci/ini ['Hjproline, or 20 fiCi/ml [<5S]sulfate was supplemented with 20 Â¿ig/mlfresh ascorbate The fibronectin immunoreactivity released into the culture medium during 24-h culture was analyzed by gel filtration and added daily. Labeling medium was replaced after 3 days of culture. immunoblotting. Gel filtration experiments indicated that pro- After culturing for 5 to 7 days, cells were lysed by addition of 30 mM NHjOH for 30 min. The wells were washed with PBS, H2O. and 70% teolysis mediated by both untreated cells and cells treated with ethanol. To prepare wells coated with fibronectin, Nunclon 24-well plasminogen or enteropeptidase yielded a major immunoreac multidishes were incubated with fibronectin (2 Mg/ml in PBS, 1 ml/ tive fragment with a molecular mass of about 200 kDa (not well) overnight at 37Â°C.Before use, the wells were washed twice with shown). In addition, several minor fragments of smaller molec PBS. Fibronectin was purified from human plasma as described (19). ular weight were detected. Immunoblot analysis showed differ Degradation of Fibronectin and Labeled ECM. Cells were plated at ences in the patterns of fibronectin fragments generated by varying densities in coated wells as described in the text and cultured treatment with enteropeptidase and plasminogen, respectively in serum-free RPMI. In some experiments plasminogen and enteropep- (Fig. 2). Plasminogen-mediated proteolysis generated larger tidase were added at final concentrations of 6 and I Â¿ig/ml,respectively. Immune and preimmune IgG were used at concentrations of 10-25 /jg/ fibronectin fragments (>100 kDa) than those mediated by en ml. The concentrations of TATI and aprotinin were 1-10 ^g/ml, and teropeptidase (Fig. 2, Lanes 4 and 5). A similar difference in that of EACA was 330 Â¿ig/ml.Aliquots of 50 and 200 ÃŸ\werecollected fragment pattern was found when purified fibronectin was at various time points, centrifuged at 9000 rpm for 5 min, and analyzed digested by plasmin and TAT-2, respectively (Fig. 2, Lanes 1 by fibronectin immunoassay and ^-scintillation counting, respectively. and 2). In addition, enteropeptidase treatment and TAT-2 Detection of Fibronectin Immunoreactivity. Polyclonal antibodies digestion yielded a characteristic fragment migrating at 85 kDa. were prepared from rabbit antiserum against fibronectin (20) by Na2SO4 Proteolysis mediated by untreated cells predominantly gener precipitation and used to develop a time-resolved immunofluorometric ated large (>100 kDa) fibronectin fragments (Fig. 2, Lane 6). assay according to methods previously described (21). In a typical assay of cell culture media, the samples were incubated for 60 min at 25Â°C The plasminogen-mediated proteolysis of fibronectin was in microtiter wells coated with the fibronectin antibodies. After wash ing, the wells were incubated for 30 min with fibronectin antibodies 200 that had been labeled with a europium chelate (Wallac Oy, Turku, no additions Finland) (22). After washing of the wells and addition of an "enhance -f enteropeptidase ment solution" (22), the bound europium was measured with an Arcus o a 150 + plasminogen 1230 time-resolved fluorometer (Wallac). Using a 50-^1 sample volume, 4) w O the linear measuring range of the assay was 0.5-250 ng/ml of fibronec C 3 tin, and the coefficient of variation was below 10% in this concentration E range. E 100 Characterization of Fibronectin Fragments. The culture media were dialyzed against 5 mM NH4HCO, for 2 h. lyophilized, and analyzed by gel filtration and immunoblotting. Gel filtration was performed on a SuperÃ³se 12 column (Pharmacia) at a flow rate of 30 ml/h. using 50 mM Tris-HCI buffer (pH 7.4) containing 0.1 M NaCI and 0.02% NaN3. Fractions of 400 n\ were collected and analyzed by fibronectin immu noassay. Sodium dodecyl sulfate gel electrophoresis was performed on 20 40 60 7.5% polyacrylamide gels under nonreducing conditions (23). Proteins Time(h) were electrotransferred to nitrocellulose membranes (24), and fibronec Fig. I. Time course of COLO 205 cell-mediated release of fibronectin frag ments from fibronectin-coated wells. Cells were plated at a density of 2 X IO*/ tin fragments were detected by incubating with polyclonal fibronectin antibodies followed by secondary peroxidase-conjugated swine anti- well in 1 ml serum-free medium in the absence (X) or presence of plasminogen (O) (6 fjg/ml) and enteropeptidase (D) (l Â¿ig/ml).Aliquots of culture media were rabbit antibodies (Dakopatts, Glostrup, Denmark). Peroxidase staining collected for analysis by fibronectin immunoassay at the time points indicated. was performed with 3.3'-diaminobenzine tetrahydrochloride (Fluka). Points, means for quadruplicate wells; bars. SD. 2108

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 1991 American Association for Cancer Research. TUMOR CELL TRYPSIN AND MATRIX DEGRADATION inhibited by aprotinin, EACA, and anticatalytic polyclonal u- 500 PA IgG, but not by TATI or anticatalytic polyclonal trypsin IgG (Table 1). By contrast, TATI and anti-trypsin IgG inhibited both the enteropeptidase-mediated proteolysis and the proteolysis mediated by untreated cells (Table 1). TATI and anti- trypsin IgG caused 81 and 92% inhibition on the enteropepti dase-mediated proteolysis, respectively, and 75 and 32% inhi bition on the proteolysis mediated by untreated cells, respec tively. Preimmune IgG had no effect. When COLO 205 cells were plated on ECM deposited by human umbilical vein endothelial cells, proteolysis of fibronectin was very rapid (Fig. 3). Fibronectin immunoreactivity was 246 detected in the medium after a 1-h culture and further increased Time (h) for 24 h. After a 6-h culture, large (>200 kDa) fibronectin Fig. 3. Inhibition of COLO 205 cell-mediated degradation of ECM fibronectin by trypsin antibodies. Cells were plated on ECM-coated wells at a density of 2 x fragments predominated in the culture medium, as detected by lO'/well and incubated in 1 ml serum-free medium in the absence (O) or presence of anti-trypsin IgG (â€¢)and preimmune IgG (Ãœ).Control wells were incubated without cells (X). The concentrations of immune and preimmune IgG were 25 5 6 Mg/ml. Aliquots of culture media were collected for analysis by fibronectin immunoassay at 2-h intervals. Points, means for quadruplicate measurements; ears. SD. i 200 800

o 600-

E 400

_ 200

COLO 205 K-562 CAPAN-1 Cancer cell line 30 Fig. 4. Prevention of cancer cell-mediated degradation of ECM fibronectin by TATI. COLO 205. K-562. and CAPAN-I cells were seeded at densities of 2 x 10'. 5 x IO5,and 2.5 x 105/well. respectively, in 0.5 ml serum-free medium. Cells were cultured in the absence (â€¢)orpresence (EH)ofTATI, and the culture media were analyzed for fibronectin immunoreactivity. COLO 205 cells were incubated Fig. 2. Immunoblot analysis of fibronectin fragments released by cells from for 4 h, and K-562 and CAPAN-1 cells were incubated overnight. The concentra coated wells or generated by digestion of purified fibronectin with TAT-2 and tion of TATI was 10 ^g/ml. Columns, means from triplicate wells; bars, SD. plasmin. Lane I, fibronectin digested for l h with plasmin (0.025 casein unit); Untreated control wells did not release fibronectin immunoreactivity. Lane 2, fibronectin digested for l h with TAT-2 (IO ng); Lane 3, untreated fibronectin; Lane 4, culture medium of plasminogen-treated cells; Lane 5, culture medium of enteropeptidase-treated cells; Lane 6. culture medium of untreated cells. Asterisks, 85-kDa fragment characteristic of enteropeptidase- and TAT-2- immunoblotting (not shown). TATI and anti-trypsin IgG inhib mediated proteolysis. ited degradation by 51 and 40%, respectively (Figs. 3 and 4). The proteolysis caused by two other cell lines which produce TAT-2, K-562 (14) and CAPAN-1 was similarly inhibited by Table 1 Effect of protease inhibitors and anticatalytic protease antibodies on COLO 205 cell-mediated release of fibronectin fragments from fibronectin-coated TATI (Fig. 4). TATI prevented K-562 and CAPAN-1 cell- wells mediated proteolysis by 47 and 54%, respectively. ci of inhibition with following additions to TATI was also effective in inhibiting the destruction of iso- serum-free culture medium" topically labeled ECM, when cells were cultured under serum- Inhibitor" None Enteropeptidase Plasminogen free conditions in the absence of plasminogen. TATI added at a concentration of 10 Mg/ml reduced the release of ['H]proline TATIAprotininEACAAnti-trypsin Â±3.589 Â±6.968 label from ECM by all three cell lines studied. After a 24-h Â±0.40Â±0.8ND49 IgGPreimmune incubation, TATI prevented the proteolysis mediated by COLO IgGAnti-u-PA 205, CAPAN-1, and HT 1080 by 40, 48, and 35%, respectively IgG7580N32306.2'8.4D"5.91.28.08182N92041.22.4D5.56.23.92Â±1.5 (Fig. SB). The degradation of ECM labeled with ['Hjserine Â°TATI and aprotinin Â»ereused at a concentration of 1 ^g/ml. EACA at a could be inhibited by TATI in COLO 205 and in CAPAN-1 concentration of 330 fig/ml, and prcimmune and immune IgG at a concentration of 10 fig/ml. cells (Fig. 5A). The percentage of inhibition was 57 and 51%, * Fibronectin immunoreactivity released into the culture medium was deter respectively. However, TATI did not prevent HT 1080 cell- mined by the fibronectin immunoassay after culturing for 16-24 h. The inhibition mediated degradation of ['Hjserine-labeled ECM (data not is given as the decrease of the immunoreactivity compared with that observed in the absence of inhibitors. Cells (2 x 10*) were plated on fibronectin-coated wells shown). It is notable that of the three cell lines studied, HT in the presence or absence of plasminogen and enteropeptidase in serum-free 1080 hydrolyzed the matrix most weakly, and it secreted the medium. c Mean Â±SD for at least 2 experiments with 2 wells/experiment. lowest amount of TAT-2 as analyzed by TAT-2 immunoassay d ND. not determined. (not shown). The degradation of [35S]sulfate-labeled ECM could 2109

2000 mediated degradation of ECM. To demonstrate TAT-2 activity in vitro, it was essential to culture cells in serum-free medium to avoid the presence of serum trypsin inhibitors and plasmin 1500- ogen. Under these conditions, the specific inhibitor TATI pre vented the proteolysis of fibronectin or isotopically labeled ECM by tumor cell lines that produce TAT-2. Further evidence 1000 for involvement of TAT-2 in proteolysis was the finding that anticatalytic trypsin antibodies were inhibitory but preimmune antibodies had no effect. In addition, the cell-mediated degra 500 dation of fibronectin was augmented by adding exogenous enteropeptidase, an activator of trypsinogen, to the culture medium. Analysis of fibronectin fragments generated indicated COLO205 CAPAN-1 COLO 205 CAPAN 1 HT 1080 COLO205 that enteropeptidase-induced proteolysis yielded fragments Cancer cell line similar to those formed by digestion of purified fibronectin with Fig. 5. Inhibition of cancer cell-mediated destruction of isotopically labeled TAT-2. matrices by TATI. A, Â¡'HJserine-labeled ECM: B. [3H|proline-labeIed ECM: C. |35S]sulfate-labeled ECM. COLO 205 and HT 1080 cells were seeded at a density Previous studies have indicated that cell-mediated degrada of 5 x lO'/well, and CAPAN-1 cells at a density of 2.5 x lO'/well in 0.5 ml tion of ECM or protein substrates is enhanced by addition of serum-free RPMI medium. Cells were cultured for 24-48 h in the absence (â€¢)or plasminogen (17,18, 25-28). This was also found in the present presence (D) of TATI, and the isotope label released into the medium was measured by Â¿-scintillation counting. The concentration of TATI was 10 ng/ml. study. Addition of plasminogen to serum-free medium in Columns, means from triplicate wells; bars. SD. Untreated control wells released creased up to 10-fold the degradation of labeled ECM and 229 Â±10, 123 Â±21, and 73 Â±4 cpm |'H]serine, |'H]proline, and [35S]sulfate fibronectin by COLO 205 cells. Plasminogen-dependent tumor label, respectively. cell-mediated proteolysis can be inhibited by protease nexin I (29), plasminogen activator inhibitor 1 (30), or plasminogen 4000 activator inhibitor 2 (18), indicating an important role for u- Control + plasminogen PA in the protease cascade. However, several investigators have found that there is also a basal cell-mediated proteolytic activity IH plasminogen + TATI in the absence of plasminogen (25-28). Our results suggest that in the cell lines studied a major portion of this is due to the activity of TAT-2. In addition to degrading ECM by itself, TAT-2 can also participate in the activation of pro-u-PA (8). This could be an explanation for the rapid activation of plas minogen exogenously added to cell culture. Perhaps surprisingly, TATI prevented the degradation of [3H]proline-labeled ECM by all three tumor cell lines studied: serine sulfate COLO 205; CAPAN-1; and HT 1080. As the [3H]proline label Labeled ECM is assumed to be biosynthetically incorporated primarily into Fig. 6. Effect of plasminogen on degradation of isotopically labeled ECM by COLO 205 cells. Cells (5 x 10') were incubated without plasminogen (â€¢).with the collagen component of newly synthesized subendothelial matrix (26), our results suggest that TAT-2 could mediate the plasminogen (6 Mg/ml) (D), or with plasminogen (6 /jg/ml) and TATI (10 ng/ml) (D) in 0.5 ml serum-free medium. After culturing for 16 h, the released radioac degradation of collagenous material. This is also supported by tivity was measured by /Â¿-scintillationcounting. Columns, means from triplicate the finding that purified TAT-2 released ['H]proline label from wells; bars, SD. Control wells treated with plasminogen alone released 386 Â±29, 177 Â±49, and 108 Â±9 cpm [3H]serine, [3H|proline, and ("S]sulfate label, the matrix. It remains to be shown whether TAT-2 can degrade respectively. native collagens in ECM or whether it degrades other trypsin- sensitive proline-containing matrix components. It is also pos sible that TAT-2 can activate latent collagenase, because trypsin be prevented by TATI in COLO 205 cells (Fig. 5C), but not in CAPAN-1 or HT 1080 cells (not shown). The inhibition of is a known activator of procollagenase (1). TATI inhibited the degradation of ['Hjserine-labeled ECM TATI in COLO 205 cells was 63%. by COLO 205 and CAPAN-1 cells and the degradation of As in the case of fibronectin proteolysis, addition of plasmin ["Sjsulfate-labeled ECM by COLO 205 cells. The ['Hjserine ogen to the culture medium substantiated the degradation of all three isotopically labeled matrices by COLO 205 cells. In label would be biosynthetically incorporated into most matrix proteins (18), and [35S]sulfate label into sulfated proteoglycans contrast to the proteolysis of fibronectin, TATI weakly but significantly prevented proteolysis in the presence of plasmin and proteins (25). These results suggest that depending on ogen (Fig. 6). The TATI-mediated inhibition (expressed in cpm) tumor cell line, TAT-2 contributes to the degradation of matrix corresponded to the radioactivity released in the plasminogen- (glyco)proteins and proteoglycans. Because TATI did not in independent proteolysis. Therefore, TATI may have inhibited hibit the destruction of all labeled matrices by CAPAN-1 and the portion of proteolysis that was exerted by TAT-2 but not HT 1080 cells, other proteinases not characterized in this study by plasmin. Active TAT-2 purified from COLO 205 culture are also involved in the proteolysis. When proteolysis by COLO medium (100 ng) hydrolyzed all three labeled matrices, as 205 cells was studied in plasminogen-supplemented medium, demonstrated by release of label into the medium (not shown). TATI caused an inhibition similar to that observed in the absence of plasminogen, apparently that portion that was due to TAT-2. These results suggest that ECM degradation is DISCUSSION accomplished by the concerted action of plasmin, plasminogen These studies demonstrate that the novel trypsinogen isoen- activators, TAT-2, and other proteases. In COLO 205 cells zyme TAT-2 produced by cancer cells contributes to the cell- plasminogen was activated by u-PA, inasmuch as the proteolysis 2110

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 1991 American Association for Cancer Research. TUMOR CELL TRYPSIN AND MATRIX DEGRADATION of fibronectin could be partially inhibited by anticatalytic uro- correlation with metastatic potential. Science (Washington DC), 212: 1151- 1153, 1981. kinase antibodies. Recently, COLO 205 cells have been shown 5 Denhardt, D. T., Greenberg, A. H., Egan, S. E., Hamilton, R. T., and Wright, to express u-PA and its mRNA (18, 31). J. A. Cysteine protease cathepsin L expression correlates closely with the Because TAT-2 is secreted in trypsinogen form from the cells metastatic potential of H-ras-transformed murine fibroblasts. Oncogene, 2: 55-59, 1987. (14), an obvious question is how it is activated. Part of TAT-2 6. Spyratos, F., Brouillet, J-P., Defrenne, A., Macene, K., RouÃ©sse,J.,Maude- is in active form in cell culture as indicated by inhibition of londe. T., Brunei, M.. Andrieu, C., Desplaces, A., and Rochefort, H. Cathep sin D: an independent prognostic factor for metastasis of breast cancer. proteolysis by TATI and trypsin antibodies. Enteropeptidase Lancet, 2: 1115-1118, 1989. (32) and cathepsin B (33) are activators of pancreatic trypsin- 7. Stenman, U-H., Koivunen, E., and Vuento, M. Characterization of a tumor- ogen. Purified TAT-2 is activated by enteropeptidase (14), and associated serine protease. Biol. Chem. Hoppe-Seyler, 369: 9-14, 1988. Koivunen, E., Huhtala, M-L., and Stenman, U-H. Human ovarian tumor- we show here that it potentiates COLO 205 cell-mediated associated trypsin. Its purification and characterization from mucinous cyst degradation of fibronectin. The potentiation is due to activation fluid and identification as an activator of pro-urokinase. J. Biol. Chem., 264: of TAT-2, as the proteolysis can be inhibited by TATI and 9 14095-14099, 1989. Halila. H., Lehtovirta, P., and Stenman, U-H. Tumour-associated trypsin trypsin antibodies. It remains to be shown whether an entero- inhibitor (TATI) in ovarian cancer. Br. J. Cancer, 57: 304-307, 1988. peptidase-like protease-activating TAT-2 exists in tumor cells. 10 Liotta, L. A., Tryggvason, K., Garbisa, S., Hart, I., Foltz, C. M., and Shafie, S. Metastatic potential correlates with enzymatic degradation of basement This study demonstrates that TATI, also known previously membrane collagen. Nature (Lond.), 284:67-68, 1980. as pancreatic secretory trypsin inhibitor (34), is a potential 11. Carlsen, S. A., Ramshaw, I. A., and Warrington, R. C. Involvement of inhibitor of tumor cell-mediated degradation of ECM. Even plasminogen activator production with tumor metastasis in a rat model. Cancer Res., 44: 3012-3016, 1984. though the plasminogen-dependent proteolysis was poorly or 12. Hasui, Y., Suzumiya, J., Marutsuka, K., Sumiyoshi, A., Hashida, S., and not at all inhibited by TATI, apparently because it is only a Ishikawa, E. Comparative study of plasminogen activators in cancers and normal mucosae of human urinary bladder. Cancer Res., 49: 1067-1070, weak inhibitor of plasmin with a KÂ¡of9.7 nivi (35), TATI may 1989. play an important matrix-stabilizing role by controlling the 13. Koivunen, E., Itkonen. O., Halila, H., and Stenman, U-H. Cyst fluid of activation of TAT-2 zymogen. Coexpression of TATI with ovarian cancer patients contains high concentrations of trypsinogen-2. Can cer Res., 50: 2375-2378, 1990. TAT-2 in the cyst fluids of ovarian tumors (13) and cancer cell 14. Koivunen. E., Saksela, O.. Itkonen. O., Osman, S.. Huhtala, M-L., and lines COLO 205 (14) and CAPAN-1 (36) further suggests a Stenman, U-H. Human colon carcinoma, fibrosarcoma and leukemia cell lines produce tumor-associated trypsinogen. Int. J. Cancer, in press, 1991. specific role for TATI in controlling TAT-2. TATI was initially ,5 Stenman, U-H., Pesonen, K., and Huhtala, M-L. Rapid concentration of identified as a tumor marker for ovarian cancer (37, 38), but urinary peptides and proteins. Anal. Biochem., Â¡23:291-294, 1982. the levels in serum and urine are also elevated in patients with 16-Jaffe, E. A., Nachman, R. L., Becker, C. G.. and Minick. C. R. Culture of human endothelial cells derived from umbilical veins. Identification by mor other types of advanced cancer (9, 37, 39). We suggest that the phologic and immunologie criteria. J. Clin. Invest., 52: 2745-2756, 1973. elevation of TATI is a reaction to TAT-2 expression with the n. Jones, P. A., and DeClerck, Y. A. Destruction of extracellular matrices containing glycoprotein. elastin, and collagen by metastatic human tumor aim of controlling TAT-2 activity and ECM degradation. How cells. Cancer Res., 40: 3222-3227, 1980. ever, TATI may also have other functions. TATI (or pancreatic is. Baker, M. S., Bleakley, P., Woodrow, G. C., and Doe, W. F. Inhibition of secretory trypsin inhibitor) at concentrations present in serum cancer cell urokinase plasminogen activator by its specific inhibitor PAI-2 and subsequent effects on extracellular matrix degradation. Cancer Res., 50: is mitogenic on human fibroblasts (40, 41) and endothelial cells 4676-4684, 1990. (42). Furthermore, it has been demonstrated that TATI binds 19. Vuento. M., and Vaheri. A. Purification of fibronectin from human plasma specifically to various cultured cells. A 140-kDa cell surface by affinity chromatography under non-denaturing conditions. Biochem. J., 183: 331-337, 1979. receptor protein mediating the binding could be characterized 20. Vuento, M., Salonen. E.. Salminen, K., Pasanen. M., and Stenman, U-H. by chemical cross-linking (43). It is not known whether cell Immunochemical characterization of human plasma fibronectin. Biochem. J., 191:119-727. 1980. surface proteases are involved in the cell binding and mitogenic 2i Stenman, U-H.. Alfthan, H.. Ranta, T., Vartianinen, E., Jalkanen, J., and activities of TATI. SeppalÃ¢,M. Serum levels of human chorionic gonadotropin in nonpregnant women and men are modulated by gonadotropin-releasing hormone and sex Degradation of ECM and basement membrane is an essential steroids. J. Clin. Endocrino!. Metab., 46: 730-736, 1987. step at several stages of cancer invasion (44). The proteolysis 22. HemmilÃ¤, I., Dakubu, S., Mukkala. V-M., Siitari. H., and LÃ¶vgren, T. assay used in the present study can be envisaged as a simple Europium as a label in time-resolved immunofluorometric assays. Anal. Biochem., 137: 335-343. 1984. model of endothelial basement membrane degradation occur- 23 I.animili. U-K. Cleavage of structural proteins during the assembly of the ring during intra- and extravasation of tumor cells. Our studies head of bacteriophage T4. Nature (Lond.). 227: 680-685, 1970. suggest that TAT-2 may have a significant role in the degra- 24- Towbin, H., Staehelin, T., and Gordon, J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some appli dation of basement membranes by tumor cells. TAT-2 may be cations. Proc. Nati. Acad. Sci. USA, 76: 4350-4354, 1979. one component of the protease cascade that cells utilize for 25. Kramer, R. H., Vogel, K. G., and Nicolson, G. L. Solubilization and degra dation of subendothelial matrix glycoproteins and proteoglycans by met migration through tissue barriers. astatic tumor cells. J. Biol. Chem., 257: 2678-2686, 1982. 26. Laug, W. E., DeClerk, Y. A., and Jones, P. A. Degradation of the subendo thelial matrix by tumor cells. Cancer Res., 43: 1827-1834, 1983. ACKNOWLEDGMENTS 27. Fairbairn, S., Gilbert. R., Ojakian, G., Schwimmer, R., and Quigley, J. P. The extracellular matrix of normal chick embryo fibroblasts: its effect on We thank Dr. Olii Saksela for helpful discussions and Lusa Airas transformed chick fibroblasts and its proteolytic degradation by the trans for technical assistance. formants. J. Cell Biol.. 101: 1790-1798, 1985. 28. Rezaee, M., Chen, L., and Kramer, R. H. Measurement of plasminogen activator activity from human fibrosarcoma cells by a new microassay. Int. J. Cancer, 40: 823-829. 1987. REFERENCES 29. Bergman, B. L., Scott, R. W., Bajpai, A., Watts, S., and Baker, J. B. Inhibition of tumor-cell-mediated extracellular matrix destruction by a fibroblast pro- 1. Tryggvason, K., HÃ¶yhtyÃ¤,M.,and Salo. T. Proteolytic degradation of extra teinase inhibitor, protease nexin I. Proc. Nati. Acad. Sci. USA, 83: 996- cellular matrix in tumor invasion. Biochim. Biophys. Acta, 907: 191-217. 1000, 1986. 1987. 30. Cajot. J. F.. Bamat. J., Bergonzelli. G. E., Kruithof. E. K. O., Medcalf, R. 2. Saksela. O., and Rifkin, D. B. Cell-associated plasminogen activation: regu L., Testuz, J.. and Sordat, B. Plasminogen activator type I is a potential lation and physiological functions. Annu. Rev. Cell Biol.. 4: 93-126. 1988. natural inhibitor of extracellular matrix degradation by fibrosarcoma and 3. Matrisian. L. Y., Bowden. G. T., Krieg. P.. FÃ¼rstenberger,G., Bilami. J. P., colon carcinoma cells. Proc. Nati. Acad. Sci. USA, 87:6939-6943, 1990. Leroy, P., and Breathnach. R. The mRNA coding for the secreted protease 31. Quax, P. H. A., van Leeuwen, R. J. T., Verspaget, H. W., and Verheijen, J. transin is expressed more abundantly in malignant than in benign tumors. H. Protein and messenger RNA levels of plasminogen activators and inhibi Proc. Nati. Acad. Sci. USA, 83: 9413-9417, 1986. tors analyzed in 22 human tumor cell lines. Cancer Res.. 50: 1488-1494, 4. Sloane, B. F., Dunn, J. R., and Honn. K. V. Lysosomal cathepsin B: 1990. 2111

32. Light, A., and Janska, H. Enterokinase (enteropeptidase): comparative as- Excretion of a tumor-associated trypsin inhibitor (TATI) in urine of patients pects. Trends Biochem. Sci., 14: 110-112, 1989. with gynecological malignancy. Int. J. Cancer, 31: 71 1-714, 1983. 33 Greenbaum, L. M., Hirshkowitz, A., and Shoichet, 1. The activation of 40. Ogawa, M., Tsushima, T., Ohbu, Y., Ogawa, N., Tanaka, S., Ishida, M., and trypsinogen bv cathepsin B. J. Biol. Chem., 234: 2885-2890, 1959. Mori, T. Stimulation of DNA synthesis in human fibroblasts by human pancreatic secretory trypsin inhibitor. Res. Commun. Chem. Pathol. Phar- 34. Greene, L. J.. Pubols, M. H., and Bartelt. D. C. Human pancreatic secretory trypsin inhibitor. Methods Enzymol., 45: 8 13-856, 1976. macol 50: 155-158 1986. 35. Turpeinen. U., Koivunen, E., and Stenman. U-H. Reaction of a tumor- 41. Hamilton I., Reynolds. G., W., Scott, G. K., Sharfe, N, and Tse, C. A. r . , ...... â€¢ â€¢.j -.K i associated trypsm inhibitor with serine protemases associated with coagula- Effects of human and ovine pancreatic secretory trypsin inhibitors on the proliferation of norma| numaâ€žfibroblasts. Biol. Chem. Hoppe-Seyler, 37 J: lion and tumor invasion. Biochem. J.. 254:9] 1-914, 1988. 79-83 1990 36. Ogata. N. Demonstration of pancreatic secretory trypsin inhibitor in serum- ^ McKeehan, W. L., Sakagami, Y., Hoshi. H., and McKeehan, K. Two appar- free culture medium conditioned by the human pancreatic carcinoma cell line en( human endotheiia| ceMgrowth factors from human hepatoma cells are CAPAN-l.J. Biol. Chem.. 263: 13427-13431, 1988. tumor-associated proteinase inhibitors. J. Biol. Chem., 261: 5378-5383, 37. Stenman. U-H., Huhtala, M-L., Koistinen. R.. and SeppÃ¤lÃ¤,M.Immuno- \986. chemical demonstration of an ovarian cancer-associated urinary peptide. Int. 43 Niinobu, T., Ogawa, M., Murata, A., Nishijima, J-I., and Mori, T. Identifi- J. Cancer. 30: 53-57, 1982. cation and characterization of receptors specific for human pancreatic secre- 38. Huhtala. M-L., Pesonen. K., Kalkkinen. N., and Stenman. U-H. Purification tory trypsin inhibitor. J. Exp. Med., 172: 1133-1 142, 1990. and characterization of a tumor-associated trypsin inhibitor from the urine 44. Liotta, L. A., Rao, N.. and Wewer, U. M. Biochemical interactions of tumor of a patient with ovarian cancer. J. Biol. Chem., 257: 13713-13716, 1982. cells with the basement membrane. Annu. Rev. Biochem., 55: 1037-1057, 39. Huhtala, M-L., KahanpÃ¤Ã¤,K..SeppÃ¤lÃ¤.M.,Halila, H., and Stenman, U. H. 1986.

2112

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 1991 American Association for Cancer Research. Tumor-associated Trypsin Participates in Cancer Cell-mediated Degradation of Extracellular Matrix

Erkki Koivunen, Ari Ristimäki, Outi Itkonen, et al.

Cancer Res 1991;51:2107-2112.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/51/8/2107

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/51/8/2107. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.