Clin. Exp. Metastasis, 1992, 10,211-220

Expression of gelatinase/type IV collagenase in tumor necrosis correlates with cell detachment and tumor invasion R. Daniel Bonfil*t, Paula A. Medina*t, Daniel E. G6mez$, Eduardo Farias$, Alberto Lazarowski*, M. Fernanda Lucero Gritti*t, Roberto P. Meiss* and Oscar D. Bustuoabad* *IIHEMA, Academia Nacional de Medicina; ?Fundaci6n de Investigaci6n del Cdncer; and ~:Instituto de Oncologfa A. H. Roffo, Facultad de Medicina, Universidad de Buenos Aires, Argentina (Received 13 May 1991; accepted 5 March 1992)

We have previously observed that acellular extracts from necrotic areas (NE) of the non-metastatic murine mammary adenocarcinoma M3, enhance in vitro cell detachment and spontaneous lung metastases. In the present study, using different proteinase inhibitors along with NE, only the calcium chelator EDTA could significantly abrogate the enhanced cell detachment from M3 produced by NE. The typical cleavage products of type IV collagenase were detected inside the tumor necrotic area, mainly in association with necrobiotic cells, as evaluated by Western blot analysis and immunohistochemical assays. Zymography revealed the presence of 72- and 92-kDa gelatinase/type IV collagenase in NE. Moreover, NE increased the in vitro invasive ability of cultured M3 cells. The use of specific antibodies against both 72- and 92-kDa type IV collagenases in the invasion assay showed that only the latter was able to revert the enhanced invasiveness to the baseline. It can be concluded that tumor necrosis is an important source of gelatinase/type IV collagenase, mainly in its 92 kDa form, and plays a major role in tumor invasion. Keywords: gelatinase, metalloproteinases, tumor invasion, tumor necrosis, type IV collagenase

Introduction Neoplastic cells need to detach from the primary tumor in order to reach their target organ and establish secondary tumor foci. An increased rate of malignant cell detachment has been associated, among other factors, with tumor necrosis [1, 2], decreased cellular expression of fibronectin or laminin [3, 4] and enhanced activity of different enzymes [5, 6]. Serine, thiol-, carboxy-, and metalloproteinases have been implicated in the invasive process [7]. Since type IV collagen is exclusively found in basement membranes [8], type IV collagenase been the object of studies which Correspondence to: D r R. Daniel Bonfil, Fundaei6n de Investigaci6n del C~incer, F. D. Roosevelt 2408 1° " C " , 1428 - Buenos Aires, Argentina.

(~ 1992 Rapid Communications of Oxford Ltd

strongly correlated increased levels of this protease with enhanced invasion and/or metastasis [9-12]. MetaUoproteinase inhibitors [13-15] or type IV collagenase antibody [16] abrogated the in vitro invasive capacity of some tumor cells, as a consequence of the absence of enzymes available to degrade extracellular matrix. We have previously reported that acellular extracts from necrotic areas (NE) of both M3 (nonmetastatic) and MM3 (metastatic) murine mammary adenocarcinomas, increased in vitro cellular detachment from the primary tumor [1]. Moreover, inoculations of NE into M3 tumors led to the formation of spontaneous lung metastases in 100% of the mice so treated [1]. In the present paper, we have used different

Clinical & Experimental Metastasis Vol 10 No 3 211

R. D. Bonfil et al.

enzyme inhibitors in order to determine which protease, if any, might be present in this acellular extract. We also have investigated the type IV collagenolytic activity in NE, as well as its effect on the in vitro invasive capacity of cultured M3 cells. Our results demonstrate the existence of gelatinase/type IV collagenase with a molecular weight of 92kDa in NE; this may explain the increased cellular detachment and invasive potential seen in vitro.

Materials and methods Mice Female BALB/c mice, 10-weeks-old, raised in our Laboratory Animal Facility were used throughout. They were maintained on specially formulated pellets (Nutric, C6rdoba, Argentina) and water acl libitum, and were microbiologically tested for pathogenic murine virus and bacteria. Tumor M3 is a BALB/c mammary adenocarcinoma of spontaneous origin [17]. It has been maintained in our laboratory by serial s.c. passages in syngeneic female mice for 6 years without any apparent behavioral change. It does not give rise to metastasis and kills the mice 35-40 days after s.c. inoculation [18]. Cell line The cell line M3-TC was established from a s.c. M3 tumor (M.F. Lucero Gritti et al., unpublished data) and maintained in RPMI 1640 culture medium (Sigma Chemical Co., St Louis, MO, USA) containing 10% fetal bovine serum (FBS; BIONOS, Buenos Aires, Argentina) at 37°C in a 5% CO2 humidified incubator. This cell line has been in continuous culture for more than 90 passages and proved tumorigenic in female BALB/c mice. M3-TC was routinely screened and found to be negative for Mycoplasma spp. contamination, using the Gen-Probe Rapid Detection System (Fisher, Pittsburgh, PA, USA). Acellular tumor necrosis extract (NE) This was prepared as described previously [1]. Briefly, central necrosis from M3 was homogenized with 4 ml of saline per gram and centrifuged at 4000 and 8000 r.p.m. The protein concentration of the final supernatant was measured

212

Clinical & Experimental Metastasis Vol 10 No 3

and used as NE; its pH was 7.2. The extract was stored at -18°C until use. In some experiments NE was again centrifuged at 10000 g to obtain a particulate fraction (consisting mainly of plasma membrane and nuclei) and a supernatant. The toxicity of NE on M3-TC was measured using a colony-forming efficiency assay. Subconfluent M3-TC cells were exposed to different dilutions of NE in FBS-free culture medium with 0.1% bovine serum albumin (BSA, fraction V), for 3 h. Once treated, the cells were rinsed with RPMI, harvested, and finally plated at a concentration of 1000 cells/60 mm 2 Petri dish. Seven days later cells were fixed and stained with Leukostat Stain kit (Fisher Scientific, Orangeburg, NY, USA) and colonies of 50 or more cells were counted. The surviving fraction was calculated as a fraction of the control.

Detachment assay Non-necrotic 30 mm 3 M3 tumor fragments were placed in vials containing 2ml of NE with a protein concentration of 3.7 mg/ml, as measured by the method of Bradford with the Bio-Rad protein assay (Bio-Rad Chemical Division, Richmond, CA, USA). After incubation for 45 min at 37°C without shaking, the tumor fragments were agitated for 45 min at 29°C and 225 strokes/minute. The number of cells detached in each case was counted with a hemocytometer, and cell viability measured by Tryptan Blue exclusion test. Controls consisted of M3 tumor fragments incubated in saline instead of NE. The role of proteinases in in vitro detachment from M3 was evaluated by incubating NE with different proteinase inhibitors. The metalloproteinases were inhibited with EDTA (Sigma) [7, 19]. Aprotinin and soybean trypsin inhibitor (SBTI) (both from Sigma) were employed as serine proteinase inhibitors [7, 19]. The concentrations used were based on their documented biologic activity in the absence of cytotoxicity [13, 19, 20].

Attachment assay Multiwell tissue culture plates were coated with 0.5 ml of basement membrane Matrigel (Collaborative Research, Bedford, MA, USA; 0.5 mg/ml), and left at 37°C to polymerize. One day later subconfluent M3-TC cultures were harvested and washed three times in FBS-free culture medium. These cells were resuspended in RPMI-1640 supplemented with 0.1% BSA, seeded

Gelatinase/type IV coUagenase in tumor necrosis at a density of 3 x 105 onto the Matrigel-coated wells, and incubated at 37°C in a 5% CO2 humidified atmosphere. After 60 and 120 min, non-attached M3-TC cells were collected and counted. Attached cells were enzymatically harvested and the percentage of attachment was expressed as the number of adherent cells divided by the total number of cells in each well.

Type IV collagenolytic assay The ability of NE to degrade type IV collagen was measured using [3H]proline-labeled type IV collagen purified from Engelbreth-Holm-Swarm sarcoma as a substrate. The assay was carried out essentially as described by Nakajima [11], but using collagen in suspension instead of collagen films. Briefly, aliquots of trypsin-activated NE were incubated along with the radioactive substrate (5000 c.p.m./tube) at 37°C for 18 h. Products which did not precipitate in 10% trichloroacetic acid and 0.5% tannic acid, measured by liquid scintillation counting, represent type IV collagen digests. Bacterial collagenase (Sigma) was used as a positive control. Invasion and chemotaxis assays These were carried out essentially as described previously [9]. Polycarbonate filters with 8 #m pore diameter (Nuclepore, Pleasanton, CA, USA), were coated with 50/~g total protein of Matrigel (Collaborative Research). The Matrigel-coated filters were used to separate both compartments of a modified Boyden chamber, which contained in its lower portion 5 #g of laminin (Gibco Laboratories, Grand Island, NY, USA) dissolved in RPMI 1640 with 0.1% BSA as a chemoattractant. The upper compartment was inoculated with 1.5 x 105 M3-TC cells suspended in RPMI 1640 plus 0.1% BSA or the same medium containing NE at a final concentration of 3.7 mg/ml, and the whole was incubated at 37°C in a humidified 5% CO 2 atmosphere. After an 8-h incubation the polycarbonate filters were removed, their upper surface swabbed, and fixed and stained with Leukostat stain kit to count cells using a light microscope. The number of cells that traversed the filter and remained attached to its lower surface was used as a measure of their invasive ability. The chemotaxis assay (non-invasive migration assay), was carried out under the same conditions described for the invasion assay, except that 5 #g of type IV collagen (Gibco) was used to coat the filters to permit cell attachment. Unlike Matrigel,

type IV collagen does not cover the filter pores, thus allowing free movement of the cells through the pores towards the chemoattractant. All assays were performed in quadruplicate.

Gelatin zymography Gelatinolytic activity in NE was identified by an assay based on the method of Heussen and Dowdle [21], and adapted to use a Mini Protean II dual slab system (Bio Rad). Non-heated NE aliquots were mixed with SDS sample buffer without fl-mercaptoethanol, and applied to 7.5% acrylamide separating gels co-polymerized with 1 mg/ml of gelatin (Sigma). After electrophoresis, gels were rinsed with 2.5% (v/v) Triton X-100 in 50 mM Tris-HC1 buffer, pH 7.5, to remove SDS, and then incubated overnight in 0.15MNaC1, 10 mMCaC12 and 50 mMTris-HC1 buffer, pH 7.5. Gels were stained with 0.05% (w/v) Coomassie brilliant Blue G-250 in a mixture of methanol: acetic acid:water (5:1:4), and destained in the same solution without dye. Gelatinolytic enzymes were detected as transparent bands on the background of Coomassie Blue-stained gelatin. Prestained standard molecular weight marker proteins (Amersham Corp., Arlington Heights, IL, USA) were used as reference. Western blot analysis NE aliquots were analyzed as for zymography, except that they were previously reduced with /~-mercaptoethanol followed by boiling for 3 min, and then transferred to nitrocellulose filters (Schleicher and Schuell, Keene, NH, USA) at 100 V for 1 h at 4°C. Non-specific binding sites were blocked with 5% skim milk. The nitrocellulose sheets were then incubated overnight at 4°C with rabbit antibodies raised against type IV collagen (used at 1:60) or against 92-kDa type IV collagenase (1:250). In the first case NE was electrophoresed in a 7.5% polyacrylamide gel, while for antibodies to type IV collagenase a 6% polyacrylamide gel was used. After washing in phosphate buffer solution (PBS), the sheets were treated for 1 h at room temperature with horseradish peroxidase-conjugated goat anti-rabbit IgG (Sigma) at a dilution of 1:2500. Peroxidase activity was visualized using as chromogen substrate 3,3'-diaminobenzidine tetrahydrochloride (S~gma) in the presence of hydrogen peroxide. Affinity purified antibodies to type IV collagen were kindly supplied by Dr Hynda Kleinman (NIDR, NIH). Anti-92-kDa type IV collagenase was a gift from

Clinical & Experimental Metastasis Vol 10 No 3

213

R. D. Bonfil et al. Results

Dr Gregory Goldberg (Washington University School of Medicine). Immunohistochemical technique The classical 3-step peroxidase-anti-peroxidase (PAP) staining method was used to localize 92-kDa type IV collagenase on frozen sections from M3 tumors. Briefly, after blocking of endogenous peroxidase activity, sections were incubated with anti-type IV collagenase for the first antibody, anti-rabbit IgG for the bridge antibody, and finally the PAP complex. Diaminobenzidine (containing hydrogen peroxide) was used as chromogen substrate. In some cases, before mounting, slides were counterstained for 30 seconds with hematoxylin. Statistical analysis Non-parametric Mann-Whitney U-test, Student's t-test and variance analysis were used; P < 0.05 was considered to indicate statistical significance. In the latter the Bonferroni T-test was performed to identify which pairs of groups were significantly different (on the ranks, P -- 0.05).

Effect of proteinase inhibitors on in vitro M3 tumor cell detachment Figure 1 summarizes the degree of cell detachment seen when non-necrotic M3 tumor fragments were treated with the different agents. As previously observed NE enhanced spontaneous cell release [1]: this was not due to a cytotoxic effect of NE, since there was no alteration in colony-forming ability at the doses used. Furthermore, the viability of cells released from the tumor fragments, as evaluated by Trytpan Blue, was always greater than 90%, with or without NE. We studied whether this cell detachment enhancement produced by NE could be caused by enzymatic action, using proteinase inhibitors. Only the metal chelator EDTA could abrogate the effect induced by NE on cell detachment; SBTI and aprotinin had no inhibitory effect. We speculate that the enhanced number of cells released by these serine proteinase inhibitors when used together with NE, was not due to their own effect since they did not alter the baseline cell detachment when used alone.

NE

102-123)

NE+APROTiNIN

(61-140)

SALI NE-~APROTINtN

(39-59)

NE+EDTA

(28-42)

SALJNE+EDTA

(18-87)

NE+SBTI

(64-109)

SALI NE+SBTI

(28-44)

SALINE

30-65) I

0

I

I

I

20 40 60 80 100 120 140 MEDIAN NUMBER OF CELLS (range) x 103

Figure 1. Cell detachment from M3 tumor fragments treated in vitro with NE and proteinase inhibitors. Doses used were: aprotinin 500 KIU/ml, EDTA 0.5 mg/ml, SBTI 100 #g/ml, NE 3.7 mg protein/ml. Results were analyzed by non-parametric Mann-Whitney U-test: *P = 0.0095, **P = 0.0011 compared to NE values.

214 Clinical & Experimental Metastasis Vol 10 No 3

Gelatinase/type I V collagenase in tumor necrosis Influence of N E on in vitro tumor invasiveness, adhesion to extracellular matrix (ECM), type I V collagenolysis and chemotaxis Since many metaUoproteinases have shown to be of importance in invasion and metastases [22, 23], we decided to study the effect of N E on tumor invasive ability. When fibronectin and laminin were tested as chemoattractants in the chemoinvasion assay, M3-TC cells only migrated in response to the latter. Those cells, when suspended in culture medium containing NE, showed a significantly enhanced invasion with respect to the control (median n u m b e r of invasive cells 143 (87-180) vs. 45 (21-69), P = 0.028 by non-parametric M a n n - W h i t n e y U-test). The invasive process is supposed to occur in at least three steps: adhesion to E C M , degradation of E C M , and cell motility [24]; therefore we studied the effect of the acellular extract in each one of these. NE, used at the same concentration as in the chemoinvasion assay, did not modify the adhesive rate of M3-TC cells to Matrigel: 8 4 . 1 + 0 . 5 vs. 7 9 . 4 + 1.4 at 60 min, and 76.9 + 1.8 vs. 87.0 + 1.4 at 120 min, N E and control respectively 0 ] _ DS, expressed as percentage of attached cells). Moreover, the chemotactic response p r o m o t e d by laminin did not significantly differ in NE-treated (median n u m b e r of cells 194 (84-230)) or non-treated M3-TC cells (180 (75-201)). However, N E showed a type IV collagenolytic activity that represented 62% of that generated by bacterial collagenase, and was almost three times higher than the control (Table 1). This effect could be reversed by preheating N E at 100°C or incubating it together with E D T A , which again demonstrates that the factor responsible is a

Table 1. Type IV collagenolytic activity of NE Treatment a NE NE + EDTA NE pretreated at 100°C Bacterial collagenase Saline + EDTA Saline

Type IV collagen degrading activity (c.p.m.) b 2187.0 699.1 813.6 3505.7 742.5 784.5

+ 125.3" + 72.3 + 42.0 ___108.8"* ___ 37.2 + 54.5

aNE: 3.7 mg protein/ml; bacterial collagenase: 0.1 mg/ml; EDTA: 0.5 mg/ml. bData represent mean + S.D. of triplicate samples. Total c.p.m./tube = 5000. *P = 0.003; **P = 0.0003 compared to saline by Student's t-test.

metalloproteinase. T o determine more accurately which collagenolytic metalloproteinase was responsible for the enhancement of invasion shown by NE, rabbit antibodies against 72- or 92-kDa type IV procollagenases, kindly supplied by D r Motowo Nakajima, were added at a dilution of 1:250 to the NE-treated groups. As can be seen in Figure 2, only the antibody against the 92-kDa form was capable of inhibiting the NErelated enhanced invasiveness; no alteration in chemotaxis was observed.

Detection of 92-kDa type I V collagenase/gelatinase enzyme in N E Matrix metalloproteinases with type IV collagendegrading ability include not only 72- and 92-kDa type IV collagenases but also stromelysin and stromelysin-2 (also known as transin and transin-2) [25, 26]. The type IV collagenases degrade the basement membrane collagen as well as having gelatinolytic activity [27-29]. Thus we decided to study the gelatin degrading capacity of N E by zymography. Two major negatively stained bands of 72 and 92 kDa were observed when whole N E or particulate fraction from NE, obtained at 10000 g, were electrophoresed in the S D S - p o l y acrylamide system including the gelatin substrate (Figure 3). The gelatinolysis shown by the supernatant obtained by centrifugation of N E at 10000 g was only due to a band of 72 kDa. No activation by aminophenylmercuric acetate ( A P M A ) or trypsin was necessary for the gelatinase activity. Gelatinolytic activities were completely inhibited by including E D T A in the incubation buffer used to reveal the enzymatic reaction. Immunoblotting of the gelatin/type IV collagen degrading enzyme detected in NE, using polyclonal antibodies to 92-kDa type IV collagenase, recognized two proteins of approximately 72 and 92 kDa (Figure 4). Type I V collagenase degrades basement membrane collagen in vivo In order to demonstrate that the collagen degrading activity effectively occurs in vivo, NE-reduced aliquots were run in a S D S - P A G E system and antibodies to type IV collagen were added for Western blot analysis. As can be seen in Figure 5, degradation of collagenous proteins naturally taking place in tumor necrotic areas could be detected by the antibody. The typical cleavage products [30, 31] of approximately 135 and 125 kDa corresponding to 75% from a~l (185 kDa) and or2 chains (170 kDa) were obtained. Other type IV collagen

Clinical & Experimental Metastasis Vol 10 No 3

215

R. D. Bonfil et al.

800

_I.

600 _T_

m

.~,

I,¢1

~ 400 4-1

(D I,X~ _j ._1 LI.I 0

200

0

CONTROL

m

NE

INVASION

NE 4. ant;-72Kd

NE 4.

ant;-g2Kd

NE 4. n o r m a l IgG

f ~ l CHEMOTAXIS

Figure 2. Effect of rabbit antibodies against 72- and 92-kDa type IV procollagenases on the in vitro invasion and chemotaxis of M3-TC cells treated with NE. Antibodies and rabbit non-immune IgG were used at a dilution of 1:250, NE at a concentration of 3.7 mg protein/ml in RPMI-1640 supplemented with 0.1% BSA. Each bar represents mean number of cells migrating through porous filter. The values are calculated from quadruplicate determinations. Difference in invasion among the groups is extremely significant (variance analysis, P = 0.0001). The Bonferroni T-test showed significant differences only between control and NE, and NE and anti-92-kDa groups (P < 0.05).

A

B

C

D 97K

'q69K

216

Clinical& ExperimentalMetastasisVol 10 No 3

Figure 3. Zymographic analysis of NE. Negative-stained bands represent proteinases with gelatinolytic activity, and their disappearance indicates specific inhibition. Lane A: whole NE; lane B: supernatant obtained after centrifuging NE at 10000 g; lane C: pellet obtained at 10000 g from NE; lane D: whole NE when E D T A was added to the incubation buffer. Positions of the migration of standard protein markers are indicated according to their molecular weight: Phosphorylase b = 97 kDa; bovine serum albumin = 69 kDa.

Gelatinase/type I V collagenase in tumor necrosis the dilutions used the antibody only detected the enzyme in necrobiotic regions (Figure 6).

-97K -69K Figure 4. Western blot analysis of type IV collagenase in NE. Bands detected represent 72-kDa and 92-kDa type IV collagenases. Positions of 69-kDa and 97-kDa weight markers are shown.

185kD 17ok1>

o200K

135kD 125k1>

Discussion Necrosis is a relatively c o m m o n event in solid tumors. Although this can be induced by different factors, as tumors grow larger oxygen and nutrients b e c o m e unavailable to central areas because of peripheral vascularization: the m a x i m u m distance across which oxygen and nutrients can diffuse would be 100-150/~m f r o m blood vessels [32]. This results in a peripheral well-perfused region, a central necrotic area and an intermediate seminecrotic zone [33]. We have previously observed that acellular extracts f r o m necrotic areas of the M3 murine m a m m a r y adenocarcinoma could enhance in vitro cell d e t a c h m e n t and even generate spontaneous lung metastases in this non-metastatic t u m o r m o d e l [1]. In the present study, we have examined the nature of N E obtained f r o m M3, concluding that its cell-detaching capacity could

o97K

Figure 5. Western blot analysis of reduced NE using antibodies against type IV collagen. Conversion of type IV collagen to lower molecular species is observed. 170 and 185 kDa represent o:1 and 0:2 chains; 125 and 135 kDa corresponding to the typical cleavage products generated by type IV collagenase are indicated. 97 and 200 kDa represent phosphorylase b and myosin, used as molecular weight standards.

degradation products were observed; these could be due to other enzymes present in physiological conditions but not active in vitro. It is important to emphasize that all the degradations detected occurred inside the t u m o r necrotic area in the absence of any enzyme inhibitor, and that type I V collagen was part of the t u m o r extracellular matrix. In order to localize enzyme activity in situ, M3 frozen sections were incubated with antibody to 92-kDa type I V collagenase and immunohistochemically detected using the P A P technique. A t

Figure6. Immunodetection on frozen sections (PAP technique) of type IV collagenase: a = positivity around nest of ceils in necrobiotic areas; b = negative in wellpreserved areas. Hematoxylin counterstain, x40.

Clinical & ExperimentalMetastasisVol 10 No 3 117

R. D. Bonfil et al.

only be reversed using a metalloproteinase inhibitor. More detailed studies showed that NE contained 72- and 92-kDa gelatinases/type IV collagenases; the 92-kDa species was mainly located in the particulate fraction of NE, presumably bound to the plasma membrane debris. Unlike lysosomal enzymes, collagenases are secreted by the cell or localized on its surface [6, 7]. In a recent study no collagenolytic activity could be detected in conditioned media from M3 cultures [34]. However, focal clearing of matrigel could be observed in the vicinity of M3 pseudopodia by electron microscopy [34], which supports our presumption that membrane-bound collagenase is involved. Moreover, NE was able to enhance in vitro invasiveness of cultured M3 cells and to degrade type IV collagen. That effect would be due mainly to the activity of 92-kDa type IV collagenase, since the use of the specific antibody could revert the effect of NE. This would be of importance, taking into consideration that the 92-kDa form of the enzyme was essentially associated with malignant transformation [19, 27, 35], while it was only found in a few normal cells [22, 27]. Immunohistochemical analysis detected the gelatinase/type IV collagenase enzyme mainly in semi-necrotic areas and not in regions composed of morphologically wellpreserved cells. Under no circumstance can it be assumed that viable cells cannot express the same enzyme activities as cells in necrobiosis, but it is possible that higher levels become concentrated in the latter and are therefore detected more easily than in other areas. Hypothetically, it could be that acquisition of collagenolytic abilities would be of great advantage to cells with a deficit of oxygen and nutrients, permitting cells in necrobiosis to escape from such an inadequate environment. Their metastatic success would depend on the availability of lymphatic or blood vessels close to the necrotic areas. This was confirmed, in part, by the fact that MM3 tumor, which alternates necrotic and well-perfused non-necrotic areas, gives rise to metastases, while M3, which does not contain vessels near the necrotic zones, does not [1]. It is difficult to imagine how cells committed to die are able to synthesize and/or secrete collagenase in huge amounts. However, Young [36, 37] observed that tumor hypoxia induces D N A over-replication and enhances metastatic potential in murine systems, supporting our hypothesis. It is important to emphasize that type IV collagen degradation could be detected in NE without previous activation and without adding exogenous type IV collagen, as in in vitro assays.

218

Clinical & Experimental Metastasis Vol 10 No 3

Collagenases have been found secreted or membrane-bound in transformed cells [6, 9-12, 19, 22, 23, 27], macrophages [38], polymorphonuclear leukocytes [39], or even in serum from tumorbearing hosts [40]. However, to our knowledge, none of the above studies demonstrated in vivo collagen-degrading activity, as observed herein. B a s e d on these results, one can assume that tumor necrosis could enhance invasion and metastasis through gelatinase/type IV collagenase activity, mainly in its 92-kDa form. Thus, if 99.9% of the cells contained in a 1 cm 3 tumor, which represents approximately 109 cells [41], are killed by any therapeutic modality, the necrosis generated would be sufficient to enhance the invasive and metastatic potentials of the 106 cells remaining alive. Perhaps this could explain the many cases in which initially successful radiotherapy or chemotherapy are followed by a prompt metastatic development.

Acknowledgements We thank Dr Christiane Dosne Pasqualini for her critical reading of this manuscript and helpful suggestions; Dr Hynda Kleinman for providing the antibody to type IV collagen, Dr Gregory Goldberg for the anti-92-kDa type IV collagenase antibody and Dr Motowo Nakajima for the polyclonal anti-72- and 92-kDa type IV coUagenase antibodies. This work was supported by C O N I C E T (Consejo Nacional de Investigaciones Cientificas y Trcnicas).

References 1. Bonfil RD, Bustuoabad OD, Ruggiero RA, Meiss RP and Pasqualini CD, 1988, Tumor necrosis can facilitate the appearance of metastases. Clinical and Experimental Metastasis, 6,121-129. 2. Weiss L and Ward PM, 1983, Cell detachment and metastasis. Cancer and Metastasis Reviews, 2, 111-127. 3. Hayman EG, Engvali E and Ruoshlati E, 1981, Concomitant loss of cell surface fibronectin and laminin from transformed rat kidney cells. Journal of Cell Biology, 88,352-357. 4. Terranova VP, Hujanen ES and Martin GR, 1986, Basement membrane and the invasive activity of metastatic tumor cells. Journal of the National Cancer Institute, 77,311-316.

Gelatinase/type I V collagenase in tumor necrosis 5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

Dano K, Andreasen PA, Grondahl-Hansen J, Kristensen P, Nielsen IS and Skriver L, 1985, Plasminogen activators, tissue degradation and cancer. Advances in Cancer Research, 44, 146-239. Woolley DE, 1984, Collagenolytic mechanisms in tumor cell invasion. Cancer and Metastasis Reviews, 3, 361-372. Zucker S, 1988, A critical appraisal of the role of proteolytic enzymes in cancer invasion: emphasis on tumor surface proteinases. Cancer Investigation, 6, 219-231. Inoue S, 1989, Ultrastructure of basement membranes. International Review of Cytology, 117, 57-98. Bonfil RD, Reddel RR, Ura H, et al., 1989, Invasive and metastatic potential of a v-Ha-ras-transformed human bronchial epithelial cell line. Journal of the National Cancer Institute, 81,587-594. Liotta LA, Tryggvason K, Garbisa S, Hart I, Foltz CM and Shafie S, 1980, Metastatic potential correlates with enzymatic degradation of basement membrane collagen. Nature, 284, 67-68. Nakajima M, Welch DR, Belloni PN and Nicolson GL, 1987, Degradation of basement membrane type IV collagen and lung subendothelial matrix by rat mammary adenocarcinoma cell clones of differing metastatic potentials. Cancer Research, 47, 4869-4876. Turpeenniemi-Hujanen T, Thorgeisson UP, Hart IR, Grant SS and Liotta LA, 1985, Expression of collagenase IV (basement membrane collagenase) activity in murine tumor cell hybrids that differ in metastatic potential. Journal of the National Cancer Institute, 75, 99-103. Reich R, Thompson EW, Iwamoto Y, et al., 1988, Effects of inhibitors of plasminogen activator, serine proteases, and collagenase IV on the invasion of basement membranes by metastatic cells. Cancer Research, 48, 3307-3312. Schultz RM, Silberman S, Persky B, Bajkowski AS and Carmichael DF, 1988, Inhibition by human recombinant tissue inhibitor of metalloproteinases of human amnion invasion and lung colonization by murine B16-F10 melanoma cells. Cancer Research, 48, 5539-5545. Yagel S, Khokha R, Denhart DT, Kerbel RS, Parhar RS and Lala PK, 1989, Mechanisms of cellular invasiveness: a comparison of amnion invasion in vitro and metastatic behavior in vivo. Journal of the National Cancer Institute, 81, 768-775. Wang M and Stearns ME, 1988, Blocking of collagenase secretion by extramustine during in vitro tumor cell invasion. Cancer Research, 48, 6262-6271. Colombo LL, Bonaparte YP, Klein S and StillitaneD'Elia I, 1980, Seleccion in vivo de una linea tumoral con alia incidencia de metastasis pulmonares. Medicina (Buenos Aires), 40,827-828. Bonfil RD, Ruggiero RA, Bustuoabad OD, Meiss RP and Pasqualini CD, 1988, Role of concomitant

resistance in the development of murine lung metastases. International Journal of Cancer, 41,415-422. 19. Bernhard EJ, Muschel RJ and Hughes EN, 1990, Mr 92,000 gelatinase release correlates with the metastatic phenotype in transformed rat embryo cells. Cancer Research, 50, 3872-3877. 20. Turner GA and Weiss L, 1981, Analysis of aprotinin-induced enhancement of metastasis of Lewis lung tumors in mice. Cancer Research, 41, 2576-2580. 21. Heussen, C and Dowdle EB, 1980, Electrophoretic analysis of plasminogen activators in polyacrylamide gels containing sodium dodecyl sulfate and copolymerized substrates. Analytical Biochemistry, 102, 196-202. 22. Murphy G, Reynolds JJ and Hembry RM, 1989, Metalloproteinases and cancer invasion and metastasis. International Journal of Cancer, 44, 757-760. 23. Liotta LA and Stetler-Stevenson W, 1989, Metalloproteinases and malignant conversion: does correlation imply casuality? Journal of the National Cancer Institute, 81,556-557. 24. Liotta LA, 1986, Tumor invasion and metastases -role of the extracellular matrix: Rhoads Memorial Award Lecture. Cancer Research, 46, 1-7. 25. Birkedal-Hansen H, 1988, From tadpole collagenase to a family of matrix metalloproteinases. Journal of Oral Pathology and Medicine, 17,445-451. 26. Matrisian LM, 1990, Metalloproteinases and their inhibitors in matrix remodeling. Trends in Genetics, 6, 121-125. 27. Ballin M, Gomez DE, Sinha CC and Thorgeirsson UP, 1988, Ras oncogene mediated induction of a 92 kDa metalloproteinase; strong correlation with the malignant phenotype. Biochemical and Biophysical Research Communications, 154,832-838. 28. Collier IE, Wilhelm SM, Eisen AZ, et al., 1988, H-ras oncogene-transformed human bronchial epithelial cells (TBE-1) secrete a single metalloprotease capable of degrading basement membrane collagen. Journal of Biological Chemistry, 263, 6579-6587. 29. Wilhelm SM, Collier IE, Marmer BI, Eise EZ, Grant GA and Goldberg GI, 1989, SV40-transformed human lung fibroblasts secrete a 92-kDa type IV collagenase which is identical to that secreted by normal human macrophages. Journal of Biological Chemistry, 264, 17213-17221. 30. Garbisa S, Ballin M, Daga-Gordini D, et al., 1986, Transient expression of type IV collagenolytic metalloproteinase by human mononuclear phagocytes. Journal of Biological Chemistry, 261, 2369-2375. 31. Turpeenniemi-Hujanen T, Thorgeirsson UP, Rao CN and Liotta LA, 1986, Laminin increases the release of type IV collagenase from malignant cells. Journal of Biological Chemistry, 261, 1883-1889. 32. Brown JM, 1990, Tumor hypoxia, drug resistance, and metastases. Journal of the National Cancer Institute, 82,338-339. 33. Jain RK, 1989, Delivery of novel therapeutic agents

Clinical & Experimental Metastasis Vol 10 No 3 219

R. D. Bonfil et al. physiological barriers and strategies. the National Cancer Institute, 81,

induces D N A overreplication and enhances metastatic potential of murine tumor cells. Proceedings of

the National Academy of Science of the United States of America, 85, 9533-9537.

37.

570-576. G6mez D, Farias E, Ballin M, Puricelli L, Bal de Kieffer Joff6 E and Torgheirsson UP, 1990, Correlation between metalloproteinases and the metastasizing ability of two murine mammary adenocarcinomas. Journal of Cancer Research and Clinical Oncology, 116, 173. Yamagata S, Ito Y, Tanaka R and Shimizu S, 1989, Gelatinases of metastatic cell lines of murine colonic carcinoma as detected by substrate-gel electrophoresis. Biochemical and Biophysical Research Communications, 158,228-234. Young SD and Hill RP, 1990, Effects of reoxygenation on cells from hypoxic regions of solid tumors: anticancer drug sensitivity and metastatic potential. Journal of the National Cancer Institute, 82, 371-380. Young SD, Marshall RS and Hill RP, 1988, Hypoxia

220

Clinical& ExperimentalMetastasisVol 10 No 3

in tumors:

Journal 34.

35.

36.

of

38. Hibbs MS, Hoidal JR and Kang A H , 1987, Expression of a metalloproteinase that degrades native type V collagen and denatured collagens by cultured human alveolar macrophages. Journal of Clinical Investigation, 80, 1644-1650. 39. A e e d PA, Nakajima M and Welch DR, 1988, The role of polymorphonuclear leukocytes (PMN) on the growth and metastatic potential of 13762NF mammary adenocarcinoma cells. International Journal of Cancer, 42,748-759. 40. Ishihara A, Nabeshima K and Koono M, 1986, Partial purification and characterization of serum protease from tumor-bearing rats which cleaves type IV collagen. Invasion and Metastasis, 6,225-245. 41. Killion JJ and Fidler I J, 1989, The biology of tumor metastasis. Seminars in Oncology, 16, 106-115.