Cysteine Cathepsins Are Central Contributors of Invasion by Cultured Adenosylmethionine Decarboxylase-Transformed Rodent Fibroblasts

[CANCER RESEARCH 64, 8831–8838, December 15, 2004] Cysteine Cathepsins Are Central Contributors of Invasion by Cultured Adenosylmethionine Decarboxylase-Transformed Rodent Fibroblasts

Kirsi Ravanko,1 Kristiina Ja¨rvinen,1 Jari Helin,2 Nisse Kalkkinen,2 and Erkki Ho¨ltta¨1 1Department of Pathology, Haartman Institute, University of Helsinki and Helsinki University Central Hospital, Helsinki, Finland; and 2Institute of Biotechnology, University of Helsinki, Helsinki, Finland

ABSTRACT including the capability to initiate a cascade leading to MMP activation (4). In addition, cancer cells have been described as overexpress- Adenosylmethionine decarboxylase (AdoMetDC), a key enzyme in the ing several other invasion-related proteases encompassing members biosynthesis of polyamines, is often up-regulated in cancers. We have of the serine, cysteine, and aspartic proteinases (4–10). demonstrated previously that overexpression of AdoMetDC alone is suf- ficient to transform NIH 3T3 cells and induce highly invasive tumors in Previously, we have shown in rodent fibroblasts that overexpres- nude mice. Here, we studied the transformation-specific alterations in sion of S-adenosylmethionine decarboxylase (AdoMetDC), a key gene expression induced by AdoMetDC by using cDNA microarray and regulatory enzyme of polyamine biosynthesis, induces cell transfor- two-dimensional electrophoresis technologies. We specifically tried to mation and highly invasive tumors in nude mice (11), thus providing identify the secreted proteins contributing to the high invasive activity of a good model for studying the molecular mechanisms involved in cell the AdoMetDC-transformed cells. We found a significant increase in the invasion. Significantly, overexpressed AdoMetDC levels have also expression and secretion of procathepsin L, which was cleaved and acti- been detectable in various human cancers, and inhibition of tumor cell vated in the presence of glycosaminoglycans (heparin), and a smaller AdoMetDC activity prevents tumor formation and metastasis (12–14). increase in cathepsin B. Inhibition of the cathepsin L and B activity by Our intention was to identify fibroblast cell invasion-related factors by specific peptide inhibitors abrogated the invasive capacity of the use of cDNA microarray and proteome analyses. The results of the AdoMetDC transformants in Matrigel. The transformed cells also showed a small increase in the activity of gelatin-degrading matrix metallopro- studies indicate that the major proteases involved in the invasion of teinases (MMPs) and urokinase-type plasminogen activator activities, transformed rodent fibroblasts are cysteine cathepsins and, surpris- neither of which was sensitive to the inhibitors of cathepsin L and B. ingly, not MMPs or uPA. Furthermore, the invasive potency of the transformed cells remained unaffected by specific inhibitors of MMPs. The results suggest that cys- MATERIALS AND METHODS teine cathepsins are the main proteases contributing to the high invasiveness of the AdoMetDC-transformed cells and that the invasion potential is Cell Culture. NIH 3T3 cells (ATCC CRL 1658; American Type Culture largely independent of activation of the MMPs. Collection, Manassas, VA) stably transfected with the neomycin resistance gene (4N) or cotransfected with human AdoMetDC cDNA (Amdc-s) and Rat-1 cells stably transfected with the AdoMetDC cDNA have been described INTRODUCTION previously (11). Both pools of transfectants and Ͼ10 individual Amdc-s clones were used for the experiments. A normal cell undergoes several changes during the course of its The cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) transformation into a cancer cell. A fully transformed cell has usually containing penicillin, streptomycin, gentamicin, G418, and 5% (v/v) fetal or gained six new abilities: it can generate its own growth signals, newborn calf serum (Life Technologies, Inc., Paisley, Scotland, United King- acquire resistance to antigrowth signals, evade apoptosis, replicate dom), at 37°C in a humidified 5% CO2 atmosphere. with limitless potential, and induce neoangiogenic and invasive Metabolic Labeling and Concentration of Conditioned Medium. Cells growth (1). To escape the primary tumor and invade the adjacent were grown for 1 day to 80% confluence, and cultures were rinsed twice with tissue, the cancer cell has to degrade the surrounding basement mem- modified Eagle’s medium (MEM). The cells were incubated with methionine- 35 brane and extracellular matrix (ECM). This is a complex process that and cysteine-free MEM supplemented with 0.1 to 0.2 mCi/mL [ S]methi- likely involves multiple extracellular proteases and requires strictly onine/cysteine (Perkin-Elmer Life Sciences, Inc. Boston, MA) and one-twen- tieth volume of DMEM (to avoid depletion of methionine and cysteine) for 16 organized interaction of the cancer cell and the tumor microenviron- hours, after which the conditioned medium (CM) was collected. CM was ment. The most important regulators of this degradation process are concentrated to Յ one-fourteenth volume with the Ultrafree-15 concentration generally believed to be the matrix metalloproteinases (MMPs). These device MWCO 5000 (Millipore, Bedford, MA). Physiologic salts in the sam- are a family of zinc-containing endopeptidases comprising more than ples were removed by adding 10 mmol/L Tris-HCl (pH 7.4) to the initial 25 members capable of degrading various ECM components. Indeed, volumes and reconcentrating the samples. The protein concentrations of the several studies have shown a close correlation between stage of tumor samples were measured with the Protein Assay Kit (Bio-Rad Laboratories, progression and MMP expression (2, 3). Another extracellular prote- Hercules, CA). ase often associated with invasion is the urokinase-type plasminogen Two-Dimensional Electrophoresis. Equal amounts of proteins (50–100 activator (uPA). uPA is a serine protease with multiple actions, ␮g) from the CM were rehydrated on 18-cm NL DryStrips (pH 3–10; Amer- sham Pharmacia Biotech, San Francisco, CA) with a buffer containing 7 mol/L urea, 2 mol/L thiourea, 2 mmol/L tributylphosphine, 3.5% 3-[(3-cholamido- Received 9/25/03; revised 10/22/04; accepted 10/27/04. propyl)dimethylammonio]-1-propanesulfonic acid, 0.5% IPG buffer (pH 3–10) Grant support: The University of Helsinki, the Finnish Cancer Organizations, the Sigrid Juselius Foundation, the Academy of Finland, Helsinki University Central Hospital NL, and bromphenol blue for 12 hours at 20°C. Isoelectric focusing (IEF) was Research Funds, and the Ida Montin Foundation. K. Ravanko is a predoctoral fellow of the performed using Amersham Pharmacia Biotech IPGPhor at 500 V for 1 hour, Helsinki Graduate School in Biotechnology and Molecular Biology. 1,000 V for 2 hours, 4,000 V for 2 hours, and 8,000 V for 4 hours, with Imax The costs of publication of this article were defrayed in part by the payment of page per strip of 50 ␮A. In studying cathepsin L processing, 7-cm NL DryStrips (pH charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 3–10) were used, and IEF was performed at 500 V for 30 minutes, 1,000 V for Note: K. Ravanko and K. Ja¨rvinen contributed equally to this work. 1 hour, 4,000 V for 1 hour, and 8,000 V for 2 hours. After IEF, the strips were Requests for reprints: Erkki Ho¨ltta¨, Department of Pathology, Haartman Institute, incubated in equilibration buffer [6 mol/L urea, 50 mmol/L Tris-HCl (pH 6.8), University of Helsinki and Helsinki University Central Hospital, P. O. Box 21, FIN-00014 2% SDS, 30% glycerol, 5 mmol/L tributylphosphine, and bromphenol blue] for Helsinki, Finland. Phone: 358-0-1912-6516; Fax: 358-0-1912-7765; E-mail: erkki.holtta@ helsinki.fi. 25 minutes at room temperature and run in 12% SDS-PAGE. The gels were ©2004 American Association for Cancer Research. then stained with silver (15) or blotted onto nitrocellulose membranes. For 8831

autoradiographic detection, the stained gels were dried and exposed to Fuji the cell lysates or CM were added up to 100-␮L final volume of lysis buffer BAS-2500 phosphorimager plates. supplemented or not supplemented with 200 ␮g/mL heparin. The different In-Gel Tryptic Digestion. Silver-stained spots of proteins were cut from cathepsin L inhibitors [cathepsin L inhibitor I (Calbiochem, Darmstadt, Germany), the gel and destained by incubating these gel pieces for 30 minutes at 37°C Z-Phe-Tyr(tBu)-dimethylketone, or Ac-Leu-Leu-Nle-aldehyde (Bachem, Buben- ␮ with 0.1 mol/L NH4HCO3 in 50% acetonitrile (ACN). The gel pieces were dorf, Switzerland)] at 0, 0.5, 15, 25, or 50 mol/L concentrations were incubated shrunk with 100% ACN, after which ACN was removed. Proteins were with the samples for 1 hour at 37°C. Thereafter, 100 ␮L of reaction buffer [0.4

reduced with 20 mmol/L dithiothreitol in 0.1 mol/L NH4HCO3 at 56°C for 30 mol/L sodium acetate (pH 5.5), 4 mmol/L EDTA, and 8 mmol/L dithiothreitol] minutes and shrunk again with ACN. To block disulfide formation, the and 150 ␮mol/L cathepsin L substrate (Z-Phe-Arg-para-nitroanilide; Bachem)

sulfhydryl groups were alkylated by incubating the gel pieces with 55 mmol/L were added, the samples were incubated at 37°C, and A405 nm was measured as a iodoacetamide in 0.1 mol/L NH4HCO3 for 15 minutes at room temperature in function of time. the dark. The pieces were washed twice for 10 minutes with 0.1 mol/L Highly purified cathepsin L (specific activity, 6,036 milliunits/mg protein)

NH4HCO3, shrunk with ACN, and dried. For tryptic digestion of the proteins, from human liver (Calbiochem) was used as a positive control in the assays. the pieces were rehydrated with 10 ␮L of 0.05 mg/mL modified trypsin The activity of the purified cathepsin L was also measured at physiologic pH

(Promega, Madison, WI) in 0.1 mol/L NH4HCO3 and 10% ACN for 10 7.0 (HEPES buffer) and found to be only 16% of the activity at pH 5.5 (sodium

minutes, after which 0.1 mol/L NH4HCO3 with 10% ACN was added to cover acetate buffer). The specificity of cathepsin L inhibitor I toward the highly the pieces. Digestions were continued overnight at 37°C. The peptides formed purified cathepsin L and cathepsin B was tested [purity by SDS-PAGE, 95%;

were extracted from the gels once with 25 mmol/L NH4HCO3 for 10 minutes specific activity, 32.7 units/mg protein (Calbiochem)]. and twice with 5% HCOOH (15 minutes each). Assay of Secreted Urokinase-Type Plasminogen Activator Activity. Matrix-Assisted Laser Desorption Ionization Time-of-Flight Mass Cells were grown for 1 day, after which they were rinsed twice with DMEM Spectrometry. Peptides were desalted by reversed phase material (OligoR3; and incubated for 16 hours with serum-free DMEM. The CM was centrifuged PerSeptive Biosystems, Farmingham, MA) packed into an Eppendorf gel- to remove cell debris and concentrated as described above. The hydrolysis of loading pipette tip. Peptides were eluted from the column directly onto the H-D-Nle:HHT-Lys-para-nitroanilide in 11 mmol/L Tris-HCl (pH 7.4) buffer matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) target supplemented with 0.0024% Tween 20, 155 ␮g/mL bovine plasminogen, 1 plate with the matrix (␣-cyano-4-hydroxycinnamic acid, a saturated solution in mmol/L Spectrozyme PL (American Diagnostica, Greenwich, CT), 1.36 0.1% trifluoroacetic acid; 60% ACN) together with internal calibration stand- mmol/L 6-aminohexanoic acid, and the sample containing 10 ␮g of protein ards (angiotensin II and adrenocorticotropic hormone-clip 18–39; Sigma- was monitored by following the absorbance at 405 nm at 37°C (16). Aldrich Chemie GmbH, Steinhein, Germany). MALDI-TOF mass spectrom- Gelatin Zymography. Cells were cultured for 1 or 2 days, after which they etry (MS) was performed on a BiflexTM time-of-flight instrument (Bruker were rinsed with DMEM and incubated overnight in DMEM with or without Daltonik, Bremen, Germany) equipped with a nitrogen laser operating at 337 200 ␮g/mL heparin, and in both cases with or without 25 ␮mol/L cathepsin L nm. For protein identification, the mass spectrometric data generated were inhibitor I (Calbiochem). CM was concentrated as described above. Equal analyzed by the public domain search program ProFound3 with the NCBInr amounts of proteins (1 ␮g) from CM were suspended in nonreducing Laemmli databases. sample buffer and incubated at 37°C for 1 hour. Proteins were separated on 8% Complementary DNA Microarray Analysis. Polyadenylated mRNA was SDS-PAGE containing 2 mg/mL gelatin (Sigma-Aldrich Chemie GmbH). To isolated by oligo(dT) cellulose chromatography. The mRNA expression levels activate gelatin-degrading MMPs, the gel was incubated in buffer 1 [50 ␮ were examined with the Atlas cDNA expression array, mouse 1.2 (Clontech mmol/L Tris-HCl (pH 7.6), 5 mmol/L CaCl2,1 mol/L ZnCl2, and 2.5% ϩ Laboratories, Palo Alto, CA).4 The poly(A) RNA samples were treated with Triton X-100] at room temperature twice for 15 minutes, in buffer 2 [50 ␮ DNase I as described in the Clontech expression array user manual before mmol/L Tris-HCl (pH 7.6), 5 mmol/L CaCl2, and 1 mol/L ZnCl2] at room probe synthesis by reverse transcription. Complementary DNA probes were temperature for 5 minutes, and in buffer 3 [50 mmol/L Tris-HCl (pH 7.6), 5 ␣ 32 ␮ synthesized with [ - P]dATP (Amersham Pharmacia Biotech). The mem- mmol/L CaCl2,1 mol/L ZnCl2, 1% Triton X-100, and 0.02% NaN3] at 37°C branes were prehybridized, hybridized, and washed in a hybridization oven, overnight. Finally, the gel was stained with Coomassie Blue (Novex, San according to the manufacturer’s instructions. The imaging plates were scanned Diego, CA). with a Fuji BAS-2500 phosphorimager using MacBas 2.5 software. AtlasIm- Invasion Assay. Twenty four-well plates were overlaid with 300 ␮Lof age 2.0 software, in the global normalization mode, was used to normalize and growth factor-reduced Matrigel (Becton Dickinson Biosciences, Bedford, MA) compare signal intensities on the arrays. diluted 1:3 in DMEM. Matrigel was allowed to polymerize for 30 minutes at Immunoblotting. CM proteins were separated on 12% SDS-PAGE and 37°C. Thereafter, 10,000 cells were plated on top of the Matrigel in 100 ␮Lof transferred to nitrocellulose filters (Bio-Rad) using the trans-Blot cell (Bio- DMEM and allowed to adhere for 1 hour at 37°C. Excess DMEM was Rad). The filters were incubated in blocking buffer [25 mmol/L Tris (pH 8.0), removed, and a second 250-␮L Matrigel layer was cast above the cells. Finally, 125 mmol/L NaCl, 0.1% Tween 20, 2% bovine serum albumin, and 0.1% 500 ␮L of DMEM containing 10% fetal calf serum were added on top of the

NaN3] overnight at room temperature and then incubated with a specific Matrigel matrix. The experiment was performed without or with 0.5, 5, or 25 antibody to cathepsin L (D-20; Santa Cruz Biotechnology, Santa Cruz, CA), ␮mol/L cathepsin L inhibitor I (Calbiochem) or MMP inhibitors (Ilomastat diluted in the blocking buffer, for 2 hours at room temperature. Filters were GM6001 or its negative control and Marimastat BB2516; Calbiochem) in- rinsed five times in washing buffer [10 mmol/L Tris (pH 8.0), 150 mmol/L cluded in the Matrigel and growth medium. The cells were grown in three- NaCl, 0.05% Nonidet P-40, and 0.05% Tween 20] and incubated with horse- dimensional Matrigel for the indicated times, and the growth medium was radish peroxidase-conjugated rabbit antigoat IgGs (DAKO, Carpinteria, CA) replenished every third day. The growth patterns and morphology of the cells for 30 minutes at room temperature. Finally, the filters were rinsed five times were examined daily by phase-contrast microscopy and photographed. To also in the washing buffer, rinsed for 15 minutes in high-salt buffer [10 mmol/L count the cells, the Matrigel was allowed to liquify on ice overnight, and the Tris (pH 8.0) and 300 mmol/L NaCl], and rinsed three times in Tris-buffered number of the cells was determined with a Coulter particle counter (particles saline [10 mmol/L Tris (pH 8.0) and 150 mmol/L NaCl]. The proteins were ␾ Ն 6 ␮m counted). MDA-MB-231 breast cancer and HT-1080 fibrosarcoma visualized by enhanced chemiluminescence (Pierce, Rockford, IL) exposing cell lines served as a reference. filters to the Fuji RX film. Bovine Capillary Endothelial Cell Migration Assay. The endothelial cell Determination of Cathepsin L Activity. Cells were grown for 2 days, migration assay was performed essentially as described previously (17). Six- harvested by centrifugation, washed twice with PBS, and suspended in lysis well plates were coated with 800 ␮L of 1.9 mg/mL collagen I (Upstate buffer [0.1 mol/L sodium acetate (pH 5.0), 1 mmol/L EDTA, and 0.01% Triton Biotechnology, Charlottesville, VA) made in MEM containing 5% newborn X-100]. Lysis was completed by freezing and thawing the samples three times, calf serum (Life Technologies, Inc.). A 1-cm–diameter cylinder was placed in after which the lysates were clarified by centrifugation at maximal speed in an the middle of the well, and 200,000 bovine capillary endothelial cells were Eppendorf microcentrifuge at 4°C for 10 minutes. Next, 10 ␮g of proteins from allowed to adhere to the area inside the cylinder. The cylinder was removed, and the cells were rinsed with MEM. Then another collagen layer was spread 3 http://prowl.rockefeller.edu. with a concomitant casting of a 2-mm–diameter sample well at an approxi- 4 www.atlas.clontech.com. mately 3-mm distance from the endothelial cells. Finally, a 3-mL layer of 1.5% 8832

agar in MEM with 5% newborn calf serum was spread on top of the collagen. were synthesized by the cells and did not represent serum proteins The chemoattractant samples were prepared as follows: 18 ␮g of proteins from differentially bound to the cell surfaces of the normal and transformed the CM of Amdc-s or 4N cells with 0 or 25 ␮mol/L cathepsin L inhibitor I or cells. 20 ng of specific antibody to cathepsin L (D-20; Santa Cruz Biotechnology) The AdoMetDC-transformed cells (pools of transfectants and indi- ␮ were incubated at 37°C for 10 minutes, 200 g/mL heparin was added, and the vidual clones) consistently showed a markedly increased expression incubation was continued for 1 hour to activate the samples. Ten micrograms of a set of five proteins at about 39 kDa (pI ranging from 6.5 to 8) as of basic fibroblast growth factor was used as a positive control, and mere DMEM was used as a negative control. The cell cultures were grown at 37°C compared with expression in normal cells (Fig. 1). The five protein spots were individually excised from the gels, digested “in-gel” with and 5% CO2, and fresh chemoattractants were added to the sample wells every second day. Migration of the bovine capillary endothelial cells toward the trypsin, and analyzed by MS. The MALDI-TOF mass fingerprint sample wells was monitored daily and visualized by photography after 7 to 10 spectra obtained from all of the spots were highly similar; one of them days. is shown in Fig. 2. The peptide mass data were then used to identify Immunohistochemical Staining of Cathepsin L. Paraffin-embedded tis- the proteins from the NCBInr databases. All of the peptide finger- ␮ sue sections of AdoMetDC-induced tumors in nude mice were cut at 5 m, prints from the five proteins individually gave the cathepsin L pre- deparaffinized, and heated in microwave oven at 900 W for 2 minutes and 400 cursor their top score. W for 8 minutes in 10 mmol/L citric acid buffer (pH 6). Endogenous perox- Immunoblotting was then additionally used to confirm that the five idase activity was blocked with 0.3% H2O2 in methanol for 30 minutes. The sections were incubated with the primary antibody to cathepsin L (D-20; Santa spots identified by MALDI-TOF MS analysis represented different Cruz Biotechnology) at 4°C overnight followed by processing using the forms of cathepsin L. Equal amounts of secreted proteins from the StreptABComplex/horseradish peroxidase duet, mouse/rabbit kit (DAKO) ac- normal (4N) and AdoMetDC-transformed NIH 3T3 and Rat-1 cells cording to the manufacturer’s protocol. The chromogen used was 3-amino-9- were separated by two-dimensional electrophoresis, transferred to ethylcarbazole (AEC) (Sigma-Aldrich Inc.). The sections were gently coun- nitrocellulose membrane, and immunoblotted with a specific antibody terstained with Harris hematoxylin (Polysciences, Inc.). Sections incubated to cathepsin L. By standard enhanced chemiluminescence detection, without the primary antibody or with normal goat IgGs were used as negative signals were detected only in the immunoblots of the transformed controls, and they gave no reactions. All of the stainings were repeated at least cells, and they were at three distinct locations: at 39 kDa, with three times. Images were obtained using a Nikon Eclipse E800M microscope isoelectric point (pI) ranging from about pH 6.5 to 8; at 25 to 30 kDa, equipped with a digital camera system, and the images were captured using the Յ Nikon Capture software package. with pI around 4 to 5; and at 10 kDa, with pI about 6.5 (Fig. 3; data not shown). The five 39-kDa signals thus represent differentially charged procathepsin L forms. The immunologically detected 25- to RESULTS 30-kDa molecular mass spots were then subjected to reciprocal Cathepsin L Was the Major Secreted Protein of AdoMetDC- MALDI-TOF MS analyses (spectra not shown) and identified as Transformed Cells. As a first attempt to identify the proteins re- cathepsin L, the mature forms. sponsible for the high invasiveness of the AdoMetDC-transformed In parallel with the proteomic analyses, we examined the gene fibroblasts, we performed a comparative analysis of the secreted expression patterns of the normal and AdoMetDC-transformed NIH proteins between the normal and transformed cells. The normal and 3T3 cells by Atlas cDNA expression array (Clontech). The overall AdoMetDC-transformed NIH 3T3 (Fig. 1A) and Rat-1 (Fig. 1B) cells hybridization patterns of the microarrays were remarkably similar, were metabolically labeled with [35S]methionine and [35S]cysteine, with only about 1% of the genes showing a Ն2-fold change (data not and the proteins secreted into the media were analyzed by two- shown). We found, however, that the mRNA level of cathepsin L was dimensional electrophoresis. For visualization of the proteins, the gels highly up-regulated (approximately 10-fold) in the AdoMetDC-trans- were stained with silver and exposed to phosphorimager plates. The formed cells. Moreover, the mRNA levels of cathepsins B and C in metabolic labeling of the proteins allowed more accurate determina- the transformed cells were slightly elevated, whereas the cathepsins tion of the differences in protein expression than did silver staining, A, H, W, and D showed no transformation-specific change (Fig. 4; which is limited by the narrow linear range in which silver quantita- data not shown). The increase in cathepsin L mRNA was also one of tively binds to proteins. The labeling also confirmed that the proteins the greatest changes in mRNA expressions when Ϸ9,500 genes were

Fig. 1. Analysis of secreted proteins from the normal and AdoMetDC-transformed NIH 3T3 (A) and Rat-1 (B) cells. Equal amounts of proteins (100 ␮g) from conditioned media of the metabolically labeled normal NIH 3T3 (4N) and Rat-1 fibroblasts and their human AdoMetDC cDNA transfectants (Amdc-s) were separated by two-dimensional electrophoresis using 18-cm IPG strips. Gels were stained with silver and exposed to phosphorimager plates to analyze the total and radioactively labeled proteins, respectively. Spots chosen for further in- vestigations are circled. Molecular mass standards (in kDa) are indicated on the left.

8833

Fig. 2. MS spectrum of peptides obtained by digestion of one of the circled proteins (39 kDa, pI 6.5–8) in Fig. 1 by trypsin. Protein spots were cut from the silver-stained gels, digested with trypsin, and analyzed by MALDI-TOF MS.

analyzed by Incyte MouseGEM arrays.5 It is also noteworthy that our 85%, respectively. During the cell culturing, part of the inhibitor was global gene expression analyses with the Affymetrix Mouse Expres- found to be precipitated, resulting in a decrease in the active inhibitor sion Set 430 arrays (containing over 34,000 genes) have not revealed concentration in the medium. Secreted cathepsin L activity was also any additional cathepsins, except cathepsin K (low expression), to be inhibited with the other inhibitors tested [Z-Phe-Tyr(tBu)-dimethylk- significantly increased in the AdoMetDC-transformed cells.6 Because protease expression in cultured cells may differ from that in vivo, we wanted to verify that cathepsin L is also expressed in AdoMetDC-induced tumors in nude mice (11). Significantly, immunohistochemical analyses showed high levels of cathepsin L expression in the tumor tissue (Fig. 5). Histologically, the tumors were composed of a fairly homogenous population of typical fibrosarcoma cells. The specimens were also stained with anti-macrophage L1 protein (ab 8740; Abcam); no macrophages were detected inside the tumors (data not shown). Cathepsin L Activity Was Elevated in the Growth Medium but not Intracellularly in the AdoMetDC-Transformed Cells. We then measured cathepsin L activity in cell lysates and in CM of the normal and AdoMetDC-transformed NIH 3T3 cells. The amount and activity of the intracellular cathepsin L remained unchanged (data not shown), whereas the activity of cathepsin L in the medium of the AdoMetDC- transformed cells was higher than that in the normal cells (Fig. 6A). Direct analysis of the culture medium, however, showed the enzymatic activity toward the synthetic substrate to be very low, but preincubation of the CM with glycosaminoglycans such as heparin at acidic pH resulted in recovery of high activity in CM of the transformed cells. Instead, the activity of intracellular cathepsin L was not enhanced by the addition of heparin. The activity of secreted cathepsin L could be blocked by addition to the culture medium of cathepsin L inhibitor I (Z-FF-FMK), with Ն15 ␮mol/L Z-FF-FMK causing a complete inhibition (Fig. 6A). Analysis of the inhibitor specificity toward highly purified enzymes from human liver in a test tube showed that Z-FF-FMK inhibited not only cathepsin L but also cathepsin B. At inhibitor concentrations of 0.5, 5, and 25 ␮mol/L, the activity of human liver cathepsin L was inhibited by 96%, 98%, and 99% and the activity of cathepsin B was inhibited by 50%, 79%, and Fig. 3. Secreted cathepsin L is found in different forms in AdoMetDC-transformed NIH 3T3 and Rat-1 cells. CM of the normal fibroblasts (4N) and AdoMetDC transformants was resolved on two-dimensional electrophoresis, transferred to nitrocellulose, and 5 P. Nummela and E. Ho¨ltta¨, unpublished data. immunoblotted with a specific antibody to cathepsin L. Signals were visualized by 6 M. Yin and E. Ho¨ltta¨, unpublished data. enhanced chemiluminescence. Molecular mass markers are shown on the left. 8834

ically, MMP-2) was activated in the AdoMetDC-transformed cells, we also tested the effects of various MMP inhibitors, including Ilomastat GM6001, Marimastat BB2516 (Calbiochem), and BB3302 (British Biotech Pharmaceuticals) on this cellular invasion. Inhibition of MMPs somewhat reduced the invasive potential of AdoMetDC- transformed cells at the higher inhibitor concentrations tested, but the same effect was also observable with a control peptide (GM 6001, negative control; Calbiochem), suggesting that the effect was nonspe- cific (Fig. 8B; data not shown). As a reference, we also cultured MDA-MB-231 breast cancer and HT-1080 fibrosarcoma cell lines, which were previously shown to be highly invasive, in the Matrigel matrix. Parallel analyses showed that Fig. 4. Complementary DNA microarray analysis of cathepsin gene expression in normal and AdoMetDC-transformed NIH 3T3 cells. Representative sections of the cDNA the invasive capacity of the AdoMetDC-transformed cells is much expression arrays shown were scanned with a phosphorimager. Expression levels were higher than that of the MDA-MB-231 or HT-1080 cells. Invasion by normalized and quantified densitometrically. The Atlas arrays include single spots of each the MDA-MB-231 and HT-1080 cells was also inhibited by the cDNA. ␤-Actin is included as a control. cathepsin L inhibitor I (data not shown). Cathepsin L Activity Did Not Induce Migration of Capillary etone or Ac-Leu-Leu-Nle-aldehyde (data not shown)]. In the presence Endothelial Cells. Recent studies indicate that cathepsin L may also of heparin and acidic pH, the 39-kDa procathepsin L secreted was play a role in the regulation of angiogenesis (19). Our group has processed to 25- to 30-kDa forms with a more acidic pI, and the noticed that proteins/factors secreted by the AdoMetDC-transformed cathepsin L inhibitor Z-FF-FMK inhibited this processing, at least in NIH 3T3 cells induced migration of bovine capillary endothelial cells part, as shown by the one- and two-dimensional electrophoresis and in a three-dimensional collagen gel.7 To test whether secreted cathep- immunoblotting analyses (Fig. 6B and C). There are studies indicating sin L contributes to this phenomenon, we studied the capability of CM that cathepsin L can promote DNA synthesis and cell growth (see ref. of AdoMetDC-transformed NIH 3T3 cells, treated or not treated with 18). However, growth of the AdoMetDC-transformed cells remained cathepsin L inhibitor or specific antibody to cathepsin L, to induce the undisturbed or was only slightly inhibited (Յ15%) in the presence of migration of bovine capillary endothelial cells. Elimination of cathep- the cathepsin L inhibitor I during a culture for 3 days (data not sin L activity had no effect on this migration (data not shown). shown). Cathepsins L and B Did Not Promote the Activation of Gela- DISCUSSION tinases or uPA in AdoMetDC Transformants. A recent suggestion is that cathepsins can initiate a proteolytic cascade, resulting in the In this study, our comparative cDNA microarray and two-dimen- activation of uPA and MMs (for a review, see ref. 7). To study the sional electrophoresis analyses of normal and AdoMetDC-trans- possible role of cathepsins L and B in MMP activation, we performed formed fibroblast models revealed only a limited number of proteins zymographic analysis of gelatinase activities in the CM of normal and whose expression was consistently altered to a significant degree. One AdoMetDC-transformed NIH 3T3 cells cultured with and without of the most marked changes in the AdoMetDC-transformed cells was cathepsin L inhibitor I. Results showed that the AdoMetDC-trans- increased expression and secretion of procathepsin L, which, most formed cells displayed elevated MMP-2 content and activity and that likely, is one important factor contributing to the high invasiveness of inhibition of the cathepsin L and B activity by Z-FF-FMK did not these transformants. affect the MMP-2 activity (Fig. 7A; data not shown). Furthermore, we Cathepsin L is a lysosomal cysteine protease, which is ubiquitously analyzed the activity of uPA in the CM of the normal and transformed expressed. It is translated as preprocathepsin L, which is then pro- cells, and the effect of the cathepsin L inhibitor I on uPA, which cessed to procathepsin L with a molecular mass of 39 kDa and further activates plasminogen to plasmin, a broad spectrum serine protease processed to the mature cathepsin L, which exists in either single- able to also activate MMPs. The uPA activity appeared to be only chain (30 kDa) or two-chain (25 and 5 kDa) forms (9, 20–22). slightly higher in the AdoMetDC transformants as compared with that Cathepsin L is suggested to participate in turnover of intracellular or in the normal cells (Fig. 7B), and it was not affected by the cathepsin endocytosed proteins (8), degradation of ECM proteins, antigen proc- L inhibitor I. essing and presentation (23), bone resorption (24, 25), and various Invasive Capacity of AdoMetDC-Transformed NIH 3T3 and other processes. Many growth factors and oncogenes can stimulate Rat-1 Cells Was Dependent on Cathepsin L and B Activity. The cathepsin L expression (26–29), and increased expression of cathep- invasive potential of the AdoMetDC transformants was then studied sin L is commonly associated with malignant transformation (30, 31). in vitro by using a reconstituted basement membrane model: culturing The overexpressed procathepsin L in transformed cells is often se- the cells between two layers of three-dimensional Matrigel matrix. We creted. However, why it is secreted even though it has a lysosomal used growth factor-reduced Matrigel to avoid any possible interfer- targeting signal, mannose-6-phosphate (8), is not yet known. On the ence by exogenous growth factors. We found that, in contrast to other hand, a recent report notes that a significant proportion of the normal cells, the AdoMetDC transformants were capable of rapidly overexpressed cathepsin L remains intracellularly packed into mul- (in a day) invading the matrix, forming a more extended phenotype tivesicular endosomes (32). than in a culture on plastic tissue culture dishes. The presence of Our studies showed that AdoMetDC-transformed cells displayed cathepsin L inhibitor I inside the matrix completely inhibited this highly increased levels of cathepsin L mRNA, with no increase in invasion by AdoMetDC-transformed cells (Fig. 8A; data not shown). intracellular cathepsin L protein content or activity, whereas a high Counting the number of the invasive and proliferated cells revealed increase in extracellular cathepsin L level and activity was observed. that 5 ␮mol/L cathepsin L inhibitor I reduced invasive cell growth by These data might suggest that the main function(s) of cathepsin L Ն30%, 10 ␮mol/L cathepsin L inhibitor I reduced invasive cell regarding transformation is executed extracellularly. The procathepsin growth by 85%, and 25 ␮mol/L cathepsin L inhibitor I totally pre- vented invasive growth (Fig. 8B). Because activity of MMPs (specif- 7 A. Paasinen-Sohns and E. Ho¨ltta¨, unpublished data. 8835

Fig. 5. Immunohistochemical analysis of cathepsin L expression in AdoMetDC-induced tumors in nude mice. Paraffin-embedded tumor specimens were incubated with or without anti- cathepsin L antibody, and the signals were detected by using biotinylated antigoat immunoglobulins as a secondary antibody with avidin-biotin amplifica- tion and 3-amino-9-ethylcarbazole (AEC) as a chromogen (red). The slides were counterstained with hematoxylin (blue). Magnification, ϫ400.

L secreted by the AdoMetDC transformants existed in five forms with transformed cells showed that in addition to 39-kDa procathepsin L, different isoelectric points, which may result from differential phos- 25- to 30-kDa cathepsin L forms were present. Exposure of the CM in phorylation of the oligosaccharide moieties (33). vitro to heparin at low pH resulted in complete conversion of proca- The secreted cathepsin L in CM of the AdoMetDC-transformed thepsin L to the mature 30- and 25-kDa cathepsin L forms, as cells was predominantly in the 39-kDa precursor form. The enzyme confirmed by the two-dimensional electrophoresis and activity meas- activity measurements revealed that the inactive precursor was con- urements. These data suggest that proteolytic cleavage of secreted verted to the mature active enzyme in the presence of acidic pH and procathepsin L can occur to a small extent without added heparin but glycosaminoglycans such as heparin, as also shown by others (34, 35). that extracellular glycosaminoglycans and acidic pH are essential for Purified 39-kDa procathepsin L is known to be first cleaved to 31- to activation. Thus the extracellular glycosaminoglycans that occur at the 34-kDa intermediate forms and then further cleaved to 30-kDa single- cell surface, in the ECM, and in basement membrane proteins may chain and 25- and 5-kDa two-chain forms in vitro, and in vivo it is play an important role in the regulation and targeting of cathepsin L reported to be similarly processed intracellularly in lysosomes (21, activity. 22), but less is known about the processing of extracellular proca- Most interestingly, our data suggest that cathepsin L is a major thepsin L in vivo. Our two-dimensional electrophoresis Western blot determinant of the invasive capacity of the AdoMetDC-transformed and MALDI-TOF analyses of the CM proteins from AdoMetDC- NIH 3T3 cells because its expression was highly increased, and the

Fig. 6. A, activity of secreted cathepsin L of AdoMetDC-transformed cells compared with their normal counterparts. CM was preincubated with heparin at acidic pH for 1 hour in the absence or presence of cathepsin L inhibitor I at the concentrations indicated. Cathepsin L substrate was then added, the samples were incubated at 37°C, and A405 nm was measured in the linear range of the enzyme reaction. Variations in the activity measurements in duplicate samples were Ͻ1.5%. B, procathepsin L is processed to a mature form (30 kDa) in the presence of heparin at acidic pH as shown by SDS-PAGE and immunoblotting analyses with specific antibody to cathepsin L. C, processing of secreted cathepsin L is sensitive to cathepsin L inhibitors. Heparin-treated CM and untreated CM of AdoMetDC transformants were analyzed by two-dimensional electrophoresis using 7-cm IPG strips and immunoblotted with anti-cathepsin L antibody. Molecular mass markers (in kDa) are shown on the left.

8836

Fig. 7. A. The inhibitor of cathepsins L and B fails to prevent activation of gelatinases in AdoMetDC-transformed NIH 3T3 cells. Equal amounts of proteins (10 ␮g) from the CM of normal (4N) and AdoMetDC-transformed NIH 3T3 cells cultured with or without cathepsin L inhibitor I (25 ␮mol/L) were separated on nonreducing SDS-polyacrylamide gel impregnated with gelatin. The gel was assayed for gelatinase activity as described in Materials and Methods. The mobilities of the gelatinases expressed by HT-1080 cells used as a reference are shown on the left. B. Cathepsin L inhibitor has no significant effect on uPA activity. Equal amounts of proteins from CM of cells, cultured with or without 25 ␮mol/L cathepsin L inhibitor, were incubated at 37°C with plasminogen and its synthetic substrate. Progression of the enzymatic reactions was monitored by measuring A405 nm, and the results are from a linear range.

invasiveness of the cells in a reconstituted basement membrane (a viewed in ref. 7). The AdoMetDC-transformed cells were also found three-dimensional Matrigel matrix) was totally blocked in the pres- to display a small increase in the activities of their uPA and MMP, ence of the cathepsin L inhibitor I. However, it cannot be said that the which may have been triggered by cathepsin L. However, the activ- invasive ability of the AdoMetDC transformants is solely attributable ities of uPA and MMP-2 were not affected by the cathepsin L and B to cathepsin L because the inhibitor at higher concentrations was also inhibitor (Z-FF-FMK). It is also noteworthy that treatment of the found to inhibit cathepsin B, which is also up-regulated in these AdoMeDC-transformed cells with specific inhibitors of MMPs failed transformants. Also, it cannot be excluded that Z-FF-FMK might still to block the invasion. Hence, instead of activating other proteases, inhibit some of the as yet uncharacterized cysteine cathepsins present cathepsins L and B probably affect invasiveness through other mech- in the mouse genome (36). Considering the possible therapeutic anisms, possibly by directly degrading the ECM proteins. Indeed, applications of these inhibitors, this may yet be an advantage because cathepsin L has been found to degrade several key ECM proteins, any possible concomitant or compensatory increase in other cathep- such as collagens I and IV, fibronectin, and laminin (19, 34, 42), sins (e.g., cathepsin B) after specific cathepsin L inhibition would be whose degradation is essential for cancer cell invasion. blocked as well. At any rate, there are several studies indicating that In conclusion, in transformed rodent fibroblasts overexpressing cathepsin L by itself may also play an important role in invasion of AdoMetDC, we found a large increase in cathepsin L expression and different malignant cell lines (31, 37–39). secretion and a smaller increase in cathepsin B expression. Cathepsin Recent studies suggest that the cysteine proteases cathepsin L and L expression was found to be elevated also in AdoMetDC-induced B may act upstream of uPA and MMPs and initiate the protease tumors in nude mice. Inhibition of the activity of these cathepsins by cascade involved in the degradation of ECM (refs. 40 and 41; re- selective peptide inhibitors blocked the invasive capacity of the cells

Fig. 8. Inhibition of cathepsin L and B activity reduces invasive growth of AdoMetDC-transformed NIH 3T3 cells in the Matrigel matrix. A. Normal (4N) and AdoMetDC-transformed NIH 3T3 cells were cultured for 1 or 7 days in growth factor-reduced Matrigel matrix with or without cathepsin L inhibitor I (25 ␮mol/L) and analyzed by phase-contrast microscopy and photography. B, the number of transformed cells grown in Matrigel matrix in the absence or presence of cathepsin L or MMP inhibitors (0.5, 5, or 25 ␮mol/L) for 7 days. Cell numbers (counted in a Coulter counter) are expressed relative to the AdoMetDC cell cultures grown without the inhibitors (control).

8837

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2004 American Association for Cancer Research. CYSTEINE CATHEPSINS IN CELL INVASION in three-dimensional Matrigel. These results indicate that the activity 21. Ishidoh K, Kominami E. Multi-step processing of procathepsin L in vitro. FEBS Lett of the cathepsin L, alone or together with cathepsin B (or some as yet 1994;352:281–4. 22. Ishidoh K, Saido TC, Kawashima S, et al. Multiple processing of procathepsin L to uncharacterized cysteine cathepsin), may be responsible for the inva- cathepsin L in vivo. Biochem Biophys Res Commun 1998;252:202–7. sion potential of the transformants. Increased cathepsin L expression 23. McCoy K, Gal S, Schwartz RH, Gottesman MM. An acid protease secreted by transformed cells interferes with antigen processing. J Cell Biol 1988;106:1879–84. is also known to occur in many human tumors, such as metastatic 24. Delaisse JM, Eeckhout Y, Vaes G. In vivo and in vitro evidence for the involvement bone tumors (43), melanomas (44), malignant gliomas (45), non– of cysteine proteinases in bone resorption. Biochem Biophys Res Commun 1984; small-cell lung (46), ovarian (47), colorectal (48–50), and breast 125:441–7. 25. Kakegawa H, Tagami K, Ohba Y, et al. Secretion and processing mechanisms of cancers (48). It is also noteworthy that our preliminary immunohis- procathepsin L in bone resorption. FEBS Lett 1995;370:78–82. tochemical studies of human fibrosarcomas have shown cathepsin L 26. Dong JM, Prence EM, Sahagian GG. Mechanism for selective secretion of a lyso- to be highly expressed by the tumor cells in invasive areas of the somal protease by transformed mouse fibroblasts. J Biol Chem 1989;264:7377–83. 27. Kane SE, Troen BR, Gal S, et al. Use of a cloned multidrug resistance gene for 8 tumors. It is possible that because cathepsins L and B may be of coamplification and overproduction of major excreted protein, a transformation- general importance for cellular invasion, they may offer potentially regulated secreted acid protease. Mol Cell Biol 1988;8:3316–21. 28. Chauhan SS, Ray D, Kane SE, Willingham MC, Gottesman MM. Involvement of good targets for future cancer therapies in attempts to prevent cancer carboxy-terminal amino acids in secretion of human lysosomal protease cathepsin L. cell metastasis. Biochemistry 1998;37:8584–94. 29. Mason R, Gal S, Gottesman M. The identification of the major excreted protein (MEP) from a transformed mouse fibroblast cell line as a catalytically active precursor from cathepsin L. Biochem J 1987;248:449–54. ADDENDUM 30. Chauhan SS, Goldstein LJ, Gottesman MM. Expression of cathepsin L in human tumors. Cancer Res 1991;51:1478–81. After submission of our manuscript, Joyce et al. (51) have inter- 31. Kirschke H, Eerola R, Hopsu-Havu VK, Bromme D, Vuorio E. Antisense RNA inhibition of cathepsin L expression reduces tumorigenicity of malignant cells. Eur J estingly reported that inhibition of cathepsin cysteine proteases with a Cancer 2000;36:787–95. broad spectrum inhibitor impairs the invasive growth of pancreatic 32. Ahn K, Yeyeodu S, Collette J, et al. An alternate targeting pathway for procathepsin islet tumors in mice. L in mouse fibroblasts. Traffic 2002;3:147–59. 33. Stearns NA, Dong J, Pan J-X, Brenner DA, Sahagian GG. Comparison of cathepsin L synthesized by normal and transformed cells at the gene, message, protein, and oligosaccharide levels. Arch Biochem Biophys 1990;283:447–57. REFERENCES 34. Ishidoh K, Kominami E. Procathepsin L degrades extracellular matrix proteins in the presence of glycosaminoglycans in vitro. Biochem Biophys Res Commun 1995;217: 1. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100:57–70. 624–31. 2. Chang C, Werb Z. The many faces of metalloproteases: cell growth, invasion, 35. Mason R, Massey S. Surface activation of pro-cathepsin L. Biochem Biophys Res angiogenesis and metastasis. Trends Cell Biol 2001;11:S37–43. Commun 1992;189:1659–66. 3. Nelson AR, Fingleton B, Rothenberg ML, Matrisian LM. Matrix metalloproteinases: 36. Turk V, Kos J, Turk B. Cysteine cathepsins (proteases): on the main stage of cancer? biologic activity and clinical implications. J Clin Oncol 2000;18:1135–49. Cancer Cell 2004;5:409–10. 4. Duffy MJ. Proteases as prognostic markers in cancer. Clin Cancer Res 1996;2:613–8. 37. Frade R, Rodrigues-Lima F, Huang S, et al. Procathepsin-L, a proteinase that cleaves 5. Stetler-Stevenson WG, Yu AE. Proteases in invasion: matrix metalloproteinases. human C3 (the third component of complement), confers high tumorigenic and Semin Cancer Biol 2001;11:143–52. metastatic properties to human melanoma cells. Cancer Res 1998;58:2733–6. 6. Johnsen M, Lund LR, Romer J, Almholt K, Dano K. Cancer invasion and tissue 38. Levicar N, Dewey RA, Daley E, et al. Selective suppression of cathepsin L by remodeling: common themes in proteolytic matrix degradation. Curr Opin Cell Biol antisense cDNA impairs human brain tumor cell invasion in vitro and promotes 1998;10:667–71. apoptosis. Cancer Gene Ther 2003;10:141–51. 7. Rao JS. Molecular mechanisms of glioma invasiveness: the role of proteases. Nat Rev 39. Rousselet N, Mills L, Jean D, et al. Inhibition of tumorigenicity and metastasis of Cancer 2003;3:489–501. human melanoma cells by anti-cathepsin L single chain variable fragment. Cancer 8. Kane SE, Gottesman MM. The role of cathepsin L in malignant transformation. Res 2004;64:146–51. Semin Cancer Biol 1990;1:127–36. 40. Goretzki L, Schmitt M, Mann K, et al. Effective activation of the proenzyme form of 9. Erickson A. Biosynthesis of lysosomal endopeptidases. J Cell Biochem 1989;40:31–41. the urokinase-type plasminogen activator (pro-uPA) by the cysteine protease cathep- 10. Ren W, Sloane B. Cathepsins D and B in breast cancer. Cancer Treat Res 1996;83: sin L. FEBS Lett 1992;297:112–8. 325–52. 41. Kobayashi H, Schimitt M, Goretzki L, et al. Cathepsin B efficiently activates the 11. Paasinen-Sohns A, Kielosto M, Kaariainen E, et al. c-Jun activation-dependent soluble and the tumor cell receptor-bound form of the proenzyme urokinase-type tumorigenic transformation induced paradoxically by overexpression or block of plasminogen activator (pro-uPA). J Biol Chem 1991;266:5147–52. S-adenosylmethionine decarboxylase. J Cell Biol 2000;151:801–10. 42. Gal S, Gottesman MM. The major excreted protein of transformed fibroblasts is an 12. Pegg AE, McCann PP. S-adenosylmethionine decarboxylase as an enzyme target for activable acid-protease. J Biol Chem 1986;261:1760–5. therapy. Pharmacol Ther 1992;56:359–77. 43. Park IC, Lee SY, Jeon DG, et al. Enhanced expression of cathepsin L in metastatic 13. Regenass U, Mett H, Stanek J, et al. CGP 48664, a new S-adenosylmethionine bone tumors. J Korean Med Sci 1996;11:144–8. decarboxylase inhibitor with broad spectrum antiproliferative and antitumor activity. 44. Rozhin J, Wade RL, Honn KV, Sloane BF. Membrane-associated cathepsin L: a role Cancer Res 1994;54:3210–7. in metastasis of melanomas [published erratum appears in Biochem Biophys Res 14. Gutman M, Beltran PJ, Fan D, et al. Treatment of nude mice with 4-amidinoindan- Commun 1989;165:1444]. Biochem Biophys Res Commun 1989;164:556–61. 1-one2Ј-amidinohydrazone, a new S-adenosylmethionine decarboxylase inhibitor, 45. Sivaparvathi M, Yamamoto M, Nicolson GL, et al. Expression and immunohisto- delays growth and inhibits metastasis of human melanoma cells. Melanoma Res chemical localization of cathepsin L during the progression of human gliomas. Clin 1995;5:147–54. Exp Metastasis 1996;14:27–34. 15. O’Connell KL, Stults JT. Identification of mouse liver proteins on two-dimensional 46. Heidtmann HH, Salge U, Havemann K, Kirschke H, Wiederanders B. Secretion of a electrophoresis gels by matrix-assisted laser desorption/ionization mass spectrometry latent, acid activatable cathepsin L precursor by human non-small cell lung cancer of in situ enzymatic digests. Electrophoresis 1997;18:349–59. cell lines. Oncol Res 1993;5:441–51. 16. Vassalli JD, Belin D. Amiloride selectively inhibits the urokinase-type plasminogen 47. Nishida Y, Kohno K, Kawamata T, et al. Increased cathepsin L levels in serum in activator. FEBS Lett 1987;214:187–91. some patients with ovarian cancer: comparison with CA125 and CA72-4. Gynecol 17. Joukov V, Pajusola K, Kaipainen A, et al. A novel vascular endothelial growth factor, Oncol 1995;56:357–61. VEGF-C, is a ligand for the Flt4 (VEGFR-3) and KDR (VEGFR-2) receptor tyrosine 48. Santamaria I, Velasco G, Cazorla M, et al. Cathepsin L2, a novel human cysteine kinases. EMBO J 1996;15:290–8. proteinase produced by breast and colorectal carcinomas. Cancer Res 1998;58:1624–30. 18. Ishidoh K, Kominami E. Gene regulation and extracellular functions of procathepsin 49. Kim K, Cai J, Shuja S, Kuo T, Murnane MJ. Presence of activated ras correlates with L. Biol Chem 1998;379:131–5. increased cysteine proteinase activities in human colorectal carcinomas. Int J Cancer 19. Felbor U, Dreier L, Bryant RA, et al. Secreted cathepsin L generates endostatin from 1998;79:324–33. collagen XVIII. EMBO J 2000;19:1187–94. 50. Adenis A, Huet G, Zerimech F, et al. Cathepsin B, L, and D activities in colorectal 20. Ishidoh K, Towatari T, Imajoh S, et al. Molecular cloning and sequencing of cDNA carcinomas: relationship with clinico-pathological parameters. Cancer Lett 1995;96: for rat cathepsin L. FEBS Lett 1987;223:69–73. 267–75. 51. Joyce JA, Baruch A, Chehade K, et al. Cathepsin cysteine proteases are effectors of invasive growth and angiogenesis during multistage tumorigenesis. Cancer Cell 8 K. Ravanko, E. Ka¨a¨ria¨inen and E. Ho¨ltta¨, unpublished data. 2004;5:443–53.

8838

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2004 American Association for Cancer Research. Cysteine Cathepsins Are Central Contributors of Invasion by Cultured Adenosylmethionine Decarboxylase-Transformed Rodent Fibroblasts

Kirsi Ravanko, Kristiina Järvinen, Jari Helin, et al.

Cancer Res 2004;64:8831-8838.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/64/24/8831

Cited articles This article cites 49 articles, 14 of which you can access for free at: http://cancerres.aacrjournals.org/content/64/24/8831.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/64/24/8831.full#related-urls

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/64/24/8831. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.