Extracellular Matrix Degradation by Human Prostatic Carcinoma Cells and Its Inhibition by Retinoic Acid1

Both normal and malignant prostatic epithelial cells in culture secrete urokinase-type plasminogen activator (u-PA) into the culture medium. u-PA has been shown to have a direct association with invasive and metastatic potential of many types of cancers. We propose that prostate cancer has the intrinsic ability to invade and metastasize because of its inherent ability to secrete the serine protease u-PA. We further propose that in prostate cancer, u-PA is the key enzyme which occupies a place at the apex of the proteolytic cascade and initiates the degradative process. Subsequently, collagenases are recruited after activation of procollagenases by another serine protease plasmin formed by the activation of plasminogen by u-PA. Extracellular proteolysis involving plasmin can cause massive degradation of the extracellular matrix. We show that u-PA alone can use fibronectin as a substrate and degrade it, but u-PA alone did not degrade laminin. Serum-free conditioned medium from DU-145 human prostatic carcinoma cells has the ability to degrade both fibronectin and laminin. However, treatment of cultures with 1 p all-trans retinoic acid (HA) for 48 h reduced the ability of serum-free conditioned medium to cause u-PA-mediated degradation of fibronectin and lamimn. Thus, RA had a protective effect on these extracellular matrix glycoproteins. Treatment of cells with HA also decreased their ability to invade Matrigel in the in vitro invasion assay in a dose-dependent manner. RA at the 0.5, 1, and 10 p.M level reduced invasion to 65.7%, 46.7%, and 34.3% of control, respectively. RA reduced extracellular proteolysis and thus inhibited extracellular matrix degradation and invasion. These results may also explain one mechanism by which retinoids inhibit invasion and metastasis in vitro and in vivo. These studies have important translational value in the chemoprevention of progression of prostatic intraepithehal neoplasia to invasive carcinoma. INTRODUCTION The primary cause of death from prostate cancer is invasion and metastasis. In the early stages of tumor development, cells with a metastatic phenotype may exist in the heterogeneous tumor cell population in a primary tumor. PIN3 is considered to be equivalent to carcinoma in situ. During progression from PIN to invasive carcinoma, tumor cells cross tissue boundaries and invade the surrounding stroma. Invasion of the BM is an active process involving cell adhesion to the BM, degradation of the ECM, and migration (1). Localized degradation of ECM takes place in areas where the ratio of proteolytic enzymes to their natural inhibitors, present in the surrounding ECM, shifts in favor of proteolysis. Invasion is a critical initial step in the metastatic cascade. Degradative enzymes involved in invasion and metastasis inelude senine proteases, metalloproteases, cathepsins, and heparanases. A cascade including all or some of these is probably involved in the invasion process; however, in different tumors, one type of enzymes may dominate the process. A correlation between invasion and metastatic potential and increased expression of ECM degrading proteases has been demonstrated (1). However, little is known about the interactions of normal and malignant prostatic epithelial cells with their ECM, or about the mechanisms involved in the degradation of the BM and ECM during tumor progression, invasion, and metastasis in prostate cancer. An understanding of the mechanisms regulating these interactions is essential for a complete understanding of tumor progression and for devising ways by which these degradative processes could be inhibited while still in the early PIN stage of tumor development. Retinoids play an important role in the control of normal epithelial cell proliferation and differentiation, and they inhibit carcinogenesis, growth, invasion, and metastasis of certain tumoms both in vivo and in vitro (2, 3). A reduction in the incidence of primary prostate cancer and metastases induced by an initiation-promotion protocol involving methylnitrosourea and testosterone in Lobund-Wistar rats and inhibition of angiogenesis in such tumors by N-(4-hydroxyphenyl)retinarnide has been reported (4, 5). Furthermore, inhibition of melanoma cell invasion in vivo and in vitro by RA has also been shown (6). It is interesting to note that a recent epidemiological study showed an increased risk for prostate cancer in men with low serum vitamin A levels (7). In the present study, we examined: (a) the ability of u-PA alone and that of CM from DU-145 human prostatic carcinoma cells to degrade ECM glycoproteins fibronectin and laminin; (b) the effects of RA on the degradation of ECM in prostatic carcinoma; and (c) the effects of RA on in vitro invasion by Received 1/4/95; accepted 3/2/95. 1 This work was supported by the Biotechnology Research Center (Michigan State University). 2 To whom requests for reprints should be addressed, at 5-350 Plant Biology Building, Michigan State University, East Lansing, MI 488241312. 3 The abbreviations used are: PIN, prostatic intraepithelial neoplasia; BM, basement membrane; ECM, extracellular matrix; CM, conditioned medium; FBS, fetal bovine serum; RA, all-trans retinoic acid; SF-CM, serum free-conditioned medium; u-PA, urokinase-type plasminogen adtivator. Research. on October 31, 2017. © 1995 American Association for Cancer clincancerres.aacrjournals.org Downloaded from 756 Inhibition of Matrix Degradation in Prostate Cancer DU-l45 cells. In conducting these studies, one objective was to identify agents which would decrease or block extracellular activity of degnadative proteases and thus inhibit the process of ECM degradation and invasion. MATERIALS AND METHODS Materials. The materials used were: human prostatic carcinoma cell line DU-145 HTB 81 from American Type Culture Collection; RPM! 1640 medium 320-1875AJ, antibiotic/antirnycotic mixture 600-5240AG from GIBCO; FBS from Intergen; HEMA-3 stain 122-911 from Curtin Matheson; human fibronectin 4008, mouse larninin 40232, and Matrigel 40234 from Collaborative Research; human urokinase 128 from American Diagnostica; RA R 2625, human plasrninogen 5661, mAb to human fibronectin F-7387 and to mouse larninin L-8271 from Sigma; polyclonal antibody to human u-PA 389 from American Diagnostica; aprotinin 236-624 from Boehninger Mannheim; Centniprep 10, 4304 filter units for concentrating conditioned medium from Amicon; 4-15% gradient gels from Joule; for invasion assay, Nuclepore filters 150446 (8i.m pore size) from Costar; and Immobilon-P transfer membrane IPUH-304 FO from Millipore. Cell Culture. Stock cultures of DU-145 cells were maintamed in RPMI 1640 medium containing 2 rnrvi glutarnine, 100 units penicillin, 100 ig streptomycin, 0.25 i.g Fungizone/ml medium, and 5% FBS. Cells were subcultured once per week. Collection of SF-CM. Cells (2.5 million) were plated in 150-mm culture plates and allowed to grow for 24 h in RPM! 1640 medium containing 10% FBS. Subsequently, the cultures were washed thoroughly with three changes of PBS, and 15 ml serum-free RPM! 1640 medium were added per dish. For cultures to be treated with RA, the SF-CM also contained the appropriate concentration of RA dissolved in absolute ethanol. The final concentration of ethanol in the culture medium was 0.1%. Cells were treated with RA for 48 h in all experiments. For assays involving degradation of fibronectin and laminin by SF-CM, the medium was concentrated using filter units with a molecular weight cutoff of 10,000. Gel Electrophoresis and Western Blot Analysis. SDSPAGE was performed according to Laemmli (8; also Ref. 9). Samples of pure fibronectin and laminin were incubated with pure urokinase or with CM from 48-h RA-treated and control cultures. Concentrated samples of SF-CM from cultures were standardized for SDS-PAGE on the basis of a fixed cell number (12,000 cells/lane). Samples were run on 4-15% gradient gels at 200 V at 8#{176}Cfor 45 mm to separate fibroneetin and laminin fragments after incubation. For Western blots, samples from 4 to 15% gradient gels were transferred to Immobilon-P membrane and imrnunoblotted with mAb to fibronectin or laminin and stained using the Vectastain ABC kit as described (9) to detect fibronectin and larninin and their degradation fragments. Degradation of Fibronectin and Laminin. Pure samples of human fibronectin (2.5 .tg) were mixed with 400 milliunits of pure urokinase dissolved in water and incubated for 18 h at 37#{176}C.The reaction was stopped with the addition of sample buffer without 3-mercaptoethanol, while the tubes were kept on ice. Nonreduced samples were loaded on 4-15% gradient gels without heating. Only the molecular weight markers were reduced. The gels were stained with Coomassie blue and destained in rnethanol:acetic acid:water (3: 1 :6). For laminin, 2.5-jig samples of pure laminin were mixed with 500 milliunits of urokinase and incubated as described above. The sample buffer for laminin samples contained 3-mercaptoethanol, and these samples were reduced by heating for S mm at 95#{176}C. Sample mixtures requiring plasminogen contained 1.5 .g plasminogen. Aprotinin, an inhibitor of plasmin, was used to block any plasmin activity in the CM in order to examine degradation of fibronectin or laminin caused by urokinase alone. Ten units of aprotinin in PBS were added to the conditioned medium sample and incubated for 2 h. Fibronectin or laminin was then added, and the mixture was incubated for 18 h at 37#{176}C and processed as described above. Invasion Assay. Cell invasion was assayed in Boyden blind well chambers containing Matnigel-coated filters, as descnibed by Albini et a!. (10). One million cells were plated per 100-mm culture plate. Twenty-four h later, cells were washed three times with PBS and treated with RA in 10 ml serum-free medium for 48 h and released from culture plates using I mrvt EDTA, suspended in RPM! 1640 medium containing 0.1% BSA, and counted. Cells were resuspended in medium with BSA at 1 million cells/ml. The lower chamber was loaded with 220 pi CM (ehemoattractant) from human lung fibroblasts grown for 24 h in serum-free medium containing 50 p.g/ml ascorbic acid. In the upper chamber, 200,000 cells were plated in 650 i.l RPMI 1640 with 0.1% BSA on the Nuclepore filter coated with 500 p.g/ml Matnigel. The cells were allowed to migrate for 5 h in the incubator at 37#{176}C. The filters were processed according to a method described by Grotendonst (1 1). Briefly, the migrated cells were fixed, stained with HEMA-3, and allowed to hydrate in distilled water. Nuclear stain was extracted for 15 mm with 0.1 N HC1, and absorbance was measured at 620 nm using a Titertek microplate reader. Three replicate filters were prepared per treatment, and the mean values for three such experiments were plotted. RESULTS Urokinase Degrades Fibronectin. To determine whether urokinase has the ability to degrade human fibroneetin, pure samples of fibronectin and urokinase were incubated for 18 h and analyzed by immunoblotting. Blots treated with mAb to fibronectin show (Fig. 1) that u-PA has the ability to degrade fibronectin into smaller fragments with molecular weights between ---200,000 and 25,000. However, in the presence of plasminogen, which is activated by unokinase to plasmin, further degradation of fibronectin was observed with some fragments having molecular weights different from those of fragments produced by urokinase alone. Urokinase Does Not Degrade Laminin. Experiments similar to those described above were conducted using larninin as a possible substrate for urokinase. Immunoblots treated with mAb to laminin show (Fig. 2) that pure urokinase does not degrade laminin. However, when plasminogen is added to the sample mixture, degradation of laminin occurs with the major band of laminin fragment seen at M1 -97,000 and some smaller fragments. The intensity of high molecular weight laminin bands is concomitantly decreased. Research. on October 31, 2017. © 1995 American Association for Cancer clincancerres.aacrjournals.org Downloaded from Mr kDa 0) 0. + < < 0. 0.