Increased stromal expression of murine urokinase ... - Springer Link

Breast Cancer Research and Treatment 68: 225–237, 2001. © 2001 Kluwer Academic Publishers. Printed in the Netherlands.

Report

Increased stromal expression of murine urokinase plasminogen activator in a human breast cancer xenograft model following treatment with the matrix metalloprotease inhibitor, batimastat Claus Holst-Hansen1 , Jennifer A. Low2,3, Ross W. Stephens1 , Michael D. Johnson2,4, Peter Carmeliet5 , Thomas L. Frandsen1 , Nils Brünner1 , and Robert B. Dickson2,3 1 The

Finsen Laboratory, Copenhagen, Denmark; 2 Lombardi Cancer Center, 3 Departments of Cell Biology, Georgetown University Medical Center, Washington, DC, USA; 5 Center for Transgene Technology and Gene Therapy, Flanders Interuniversity Institute for Biotechnology, University of Leuven, Leuven, Belgium

4 Pathology,

Key words: batimastat, matrix metalloproteinase inhibitor, stromal-epithelial interactions, urokinase plasminogen activator Summary The matrix metalloprotease (MMP) family of enzymes and the urokinase plasminogen activator (uPA) pathway have both been implicated in tumor invasion and metastasis and in poor prognosis of cancer. We have previously shown that treatment with batimastat, a synthetic MMP inhibitor, leads to significant retardation but not regression of tumor growth in a human breast cancer xenograft model. In addition, batimastat treatment did not inhibit local tumor invasion, nor did it encourage stromal encapsulation of the tumor, suggesting the additional involvement of non-MMP proteolytic mechanisms. To investigate the presence of an alternative extracellular matrix protease whose activity is known to be important in breast cancer, but which is not inhibited by batimastat, expression of murine and human uPA were examined by in situ hybridization and ELISA. No differences were observed between untreated and batimastat-treated tumors regarding human uPA mRNA and protein. In contrast, murine uPA mRNA expression was increased at the tumor-stromal junction in batimastat-treated tumors in comparison with the control tumors. In agreement with these results, batimastat treatment was shown to significantly induce murine uPA protein content in the tumors. Inoculating MDA435/LCC-6 cells into immunodeficient, uPA-deficient mice resulted in tumor growth retardation as compared to tumor growth in littermate wild-type controls, while addition of batimastat treatment to uPA−/− mice did not result in further growth inhibition. The increased expression of stromal uPA may represent a cellular response to MMP inhibition and may demonstrate a new level of plasticity in the malignant progression of the disease. These results may have important implications for the clinical applications of MMP inhibitors, as well as for development of other anti-invasion drugs.

Introduction Matrix metalloproteases (MMPs) are a family of zinc endopeptidases, widely considered to be important for tumor invasion, metastasis, and angiogenesis (reviewed in Stetler–Stevenson) [1]. Inhibition of MMPs has therefore been put forward as a new anti-neoplastic treatment modality. It has already been demonstrated that a synthetic MMP inhibitor, batimastat, is capable

of retarding tumor growth, metastasis and angiogenesis in several experimental tumor models [2]. For example, Wang et al. [3], and Watson et al. [4], showed that batimastat inhibited human colon cancer tumor growth and metastasis in nude mice. Sledge et al. [5], used batimastat to retard breast cancer primary tumor regrowth and metastasis. Taraboletti et al. [6], found that batimastat inhibited the growth of a murine hemangioma, suggesting an anti-angiogenic property of

226

C Holst-Hansen et al.

the drug. In our previous study with batimastat [7], although significant differences in tumor growth rates of the human breast cancer xenograft, MDA435/LCC6, were observed between the batimastat-treated and control groups, no dramatic decrease was seen. Furthermore, the treatment had no effect on local tumor invasion, nor did it encourage stromal encapsulation of the tumor. Taken together, these observations led us to consider the presence of additional extracellular protease activity, which could functionally overlap the activity of MMP’s with regard to tumor growth and local invasion. Urokinase plasminogen activator (uPA) is a serine protease also believed to play an important role in extracellular matrix degradation. uPA expression has been correlated with tumor invasiveness and metastatic behavior, as well as with short survival of patients with a number of different cancers, including breast carcinomas [8, 9]. In several of the studied cancer types, uPA is expressed primarily by the stromal cells associated with the tumor [10, 11], and a similar stromal uPA expression was observed following implantation of a human breast carcinoma cell line MDA-MB-231 into nude mice [12]. uPA− catalyzed plasmin formation might thus contribute to the overall proteolytic activity in tumors. Moreover, plasmin has been implicated in the activation of several MMP’s [13, 14], and plasmin can, in addition, activate or release growth factors with potential implications for tumor growth [15, 16]. In order to investigate potential changes in protease expression in batimastat-treated versus control transplanted human tumors, we utilized in situ hybridization for uPA and species-specific uPA ELISA’s to examine the expression of both murine derived stromal uPA and human cancer cell derived uPA. In addition, we utilized a newly-derived combined uPA gene disrupted and immuno-deficient mouse in some of the experiments.

Materials and methods MDA435/LCC6 human breast cancer xenograft model Six to eight week old female NCr nu/nu (athymic) mice were purchased through the National Cancer Institute (Bethesda, USA) or from Bomholtgaard (Ry, Denmark), housed in the animal facility and maintained according to the regulations set forth by the United States Department of Agriculture, the American Association for the Accreditation of Laboratory Animal Care Georgetown University Animal Care and

Use Committee, and the Danish Animal Experiments Inspectorate. Athymic nude mice were implanted subcutaneously in each flank with the human breast cancer cell line MDA435/LCC6 [17]. In one experiment, tumors were excised from anesthetized animals and were halved longitudinally along the longest axis of the tumor. Half of the tumor was snap frozen in liquid nitrogen, and then stored at −80◦C until processed for northern blotting. The other half of the tumor was formalin-fixed for 24 h in 10% buffered formalin and then transferred to PBS for 24 h. The tumors were then dehydrated and paraffin-embedded. These tumors were used for in situ hybridization and trichrom staining. In another experiment, tumors for ELISA measurements were excised from anesthetized animals and snap-frozen in liquid nitrogen. Tumors were stored at −80◦ C until analyzed. Batimastat treatment Batimastat was obtained from British Biotech, Ltd. (Oxford, England). Batimastat was received as clinically formulated material (containing methylcellulose, ethanol, polyethylene glycol, and 20 mg/ml batimastat) which was diluted to 5 mg/ml batimastat in 5% dextrose. Control solutions of vehicle only were also prepared. Animals were injected intraperitoneally (i.p.) with 50 mg/kg (at 5 mg/ml) batimastat daily beginning immediately after tumor implantation and continuing for 2 or 4 weeks until the day the animals were sacrificed. Mice were weighed weekly to determine appropriate drug dosing. Control mice received vehicle 10 ml/kg/day i.p. Tumor growth in uPA deficient nude mice The generation of uPA−/− mice has been described previously [18]. Mice were kept in micro isolation cages and were fed a regular chow. In order to study the contribution of murine uPA to the growth of human MDA435/LCC-6 tumors, the uPA+/− mice (of a mixed C57Bl6J/129 background) were backcrossed 5 generations into the META/Bom substrain of BalbC background (Bomholtgaard, Ry, Denmark). Littermate offsprings were made by mating a uPA+/− META/Bom nu/nu male with a uPA+/− female of the META/Bom nu/+ genotype. In all experiments involving wild-type mice as controls, these were littermates to the uPA deficient mice. All mice used for experiments were between 8–12 weeks old at the start of the experiment. Tumor cell inoculation and batimastat treatment were performed as described above.

Anti-MMP treatment and stromal uPA expression All experimental evaluations were performed by an investigator unaware of animal genotype. All animal experiments were performed according to the guideliness published by the Danish Animal Care Committee (permission #1998/561-146). Genotyping was performed by standard procedure PCR analysis of tail DNA specimens. Genotyping of the animals was performed before and after experiments. The following primers were used: Neo: 5 ATG ATT GAA CAA GAT GGA TTG CAC G 3 , 5 TTC GTC CAG ATC ATC CTG ATC GAC 3 uPA: 5 CTG GAA TGC GCC TGC TGT CCT TCA 3 , 5 TGT CAC GAG CTG CCC TGG GAA TCA 3 In situ hybridization The probe to murine uPA, designated pMUPA07, is a 608-1642 bp cDNA fragment in pGEM5z, whereas the probe to human uPA, designated pHUPA13, is a 791-1303 bp cDNA fragment in pBluescript KS(+). Generation of 35 S-UTP-labeled anti-sense or sense probes, by transcription of the plasmid using the relevant polymerases (SP6, T3, T7), was performed as described by Kristensen et al. [19]. Before transcription, the plasmids were linearized using the following restriction endonucleases: pMUPA07: EcoRI or PstI; pHUPA13: Hind III. All probe preparations, including both sense and anti-sense probes, were adjusted to 1 × 106 cpm/µl. The uPA probes used in this study were species specific for mRNAs of mouse and human origin [12]. In situ hybridization was performed by the method of Kristensen et al. [19]. Briefly, 5-mm sections cut from formalin-fixed, paraffin-embedded tissues were treated with proteinase K. After incubation overnight at 47◦C with a hybridization solution containing radiolabeled RNA probe, the sections were washed in 50% formamide, 0.02% Ficoll 400 (w/v), 0.02% polyvinylpyrrolidone (w/v), 0.2% BSA fraction V (w/v), 0.3 M NaCl, 0.5 mM EDTA, 10 mM Tris– HCl, and 10 mM sodium phosphate (pH 6.8) twice for 1 h each time. The sections were then treated with RNase A, dehydrated, and air-dried. Autoradiographic emulsion was applied and sections developed after 1 week of exposure. Northern blot analysis Total RNA was prepared by standard methods from frozen pieces of tumor from animals treated with batimastat or the vehicle control. Aliquots (15 µg) of RNA were run on a 1.2% agarose gel containing 2.2 M

227

formaldehyde and transferred to a membrane (Nytran, S&S). The murine uPA probe was generated from the plasmid pMUPA07, by random priming. After hybridization at 65◦ C with 2.5 × 106 cpm per ml of the uPA probe, the membrane was washed in 0.5% SDS, 0.5 × SSC and exposed to X-ray film. ELISAs of uPA Human xenograft tumors or mouse organs (lungs, dermis, uterus, liver and kidney from mice treated for 2 weeks with vehicle or batimastat) were pulverized with a precooled powder pistol. The tissue powders were suspended at a ratio of 1:4 in extraction buffer (75 mM potassium acetate, 0.3 M NaCl, 0.1 M L-arginine, 10 mM EDTA, 0.25% Triton X-100, pH 4.2) [20] at 4◦ C. The suspensions were centrifuged at 105,000 × g for 1 h at 4◦ C and the resulting supernatants were stored at −80◦ C. Murine uPA ELISA Immunoassay plates (Maxisorp, Nunc, Denmark) were coated overnight at 4◦ C with 100 µl/well of the monoclonal anti-murine-uPA antibody clone H77A10 [21] (4 µg/ml) in 0.1 M carbonate buffer, pH 9.6. Before use, the assay wells were rinsed twice with 200 µl/well of SuperblockTM solution (Pierce Chemical, Rockford, IL) diluted 1:1 with PBS, followed by three washes with PBS containing 1 mg/ml Tween 20. Wells were then treated for 1 h at 37◦C with 100 µl/well of 1:20 dilutions of tissue extracts made in a sample dilution buffer of 50 M phosphate, pH 7.2, 0.1 M NaCl, 10 mg/ml bovine serum albumin (BSA; Fraction V, Boehringer Mannheim), and 1 mg/ml Tween 20. On every assay plate a series of calibrators were included that consisted of seven serial dilutions in duplicate of purified murine pro-uPA (a kind gift from Dr G. Høyer-Hansen), starting from 2 ng/ml, then 1, 0.5, 0.25, 0.125, 0.0625, 0.0313 ng/ml. Also included on each plate were duplicate blank wells containing only sample dilution buffer. After murine uPA binding, the wells were washed six times, then incubated for 1 h at 37◦C with 100 µl/well of polyclonal rabbit antibodies against murine uPA (2 µg IgG/ml) [22]. The wells were then incubated for 1 h at 37◦ C with 100 µl/well of monoclonal antirabbit immunoglobulins-alkaline phosphatase conjugate (Sigma, St Louis, MO) diluted 1:1000 in sample dilution buffer. After six washes with washing solution and three washes with water, 100 µl of freshly made p-nitrophenyl phosphate (Sigma) substrate solution

228


(1.7 mg/ml in 0.1 M Tris–HCl, pH 9.5; 0.1 M NaCl; 5 mM MgCl2 ) was added to each well and the plate was placed in a Ceres 900TM plate reader (BioTek Instruments). Yellow color development at 23◦C was monitored automatically, with readings taken at 405 nm against an air blank every 10 min for 60 min. As a test of species specificity, serial dilutions in duplicate of purified human uPA (Oncogene Science, Cambridge, USA) starting from 8.0 ng/ml, then 4.0, 2.0, 1.0, 0.5, 0.25 ng/ml were assayed. The recovery of signal from purified murine pro-uPA (muPA) was measured after addition to a 1:20 dilution of a pool of MDA-MB-435 tumor xenograft extracts. Purified muPA was added to the extract pool to final concentrations from 0 to 8 ng/ml. The recovery was calculated from the slope of the line representing the muPA signal as a function of concentration, where 100% recovery was defined as the slope obtained when muPA was diluted in the sample dilution buffer.

error rate of 0.05, with Pt (Pc ) denoting the probability of a treated (control) specimen receiving an intensity rating > 2 on a 4-point scale, where 1 = low intensity and 4 = high intensity. To determine whether treated specimens were more likely to receive a higher intensity rating than controls, a proportional-odds model was fitted to the ratings data [24]. This model was appropriate for the ratings data and avoided strong distributional assumptions such as normality. To adjust for the possible presence of intra-observer correlation, the GEE method was used to fit the proportional-odds model [25]. This statistical analysis was performed by Dr. John Hanfelt of the Biostatistics Shared Resource of the Lombardi Cancer Center. The ELISA data were analyzed by the t-test; two-sided P values below 0.05 were considered statistically significant. For comparison of the primary tumor growth curves repeated measures analysis of variance was used.

Human uPA ELISA Total human uPA concentrations in tumor extracts were measured using an ELISA kit from Oncogene Science (Cambridge, USA) [23]. Briefly, this assay used two murine monoclonal antibodies to human uPA for catching, and a rabbit polyclonal antibody to human uPA for detection. To test for species specificity, serial dilutions in duplicate of purified murine prouPA starting from 2.0 ng/ml, then 1.0, 0.5, 0.25, 0.125, 0.0625 ng/ml, were assayed. Protein concentration of the tissue extracts was determined using the Bio-Rad protein assay (Bio-Rad, Richmond, CA) with BSA as standard.

Results

Trichrome staining Paraffin-embedded tumor sections were stained for collagen using Masson’s trichrome stain. With this stain, nuclei stain black, muscle, cytoplasm and keratin stain red, and collagen stains blue to green. This staining was performed by the Lombardi Cancer Center Histopathology and Tissue Shared Resources. Statistical analysis In situ hybridization sections were scored for intensity of signal on a scale of one to four by blinded observers. Seven observers were needed to achieve 80% power to detect a significant difference in intensity between the four treated specimens and the four control specimens, assuming Pt = 0.75, Pc = 0.25, and a 2-sided Type I

Matrix deposition and local invasion around MDA435/LCC6 solid tumors MDA435/LCC6 solid tumors excised from batimastattreated and control athymic nude mice were embedded, sectioned and stained using Masson’s trichrome stain to highlight collagen within and around the tumor (Figure 1). Multiple tumor sections from both treated and control mice were examined for evidence of local invasion and stromal encapsulation. Local invasion was defined as the presence of tumor cells within the surrounding and overlying fatty and dermal tissue layers and interrupting muscular tissue fibers. Tumor encapsulation was noted as fibrous, collagencontaining material separating tumor cells from the surrounding tissue architecture. After 2 or 4 weeks of treatment, tumors from twelve control and seventeen treated mice were examined. While all tumors showed some degree of encapsulation by matrix, it was also apparent in all tumors that matrix components and surrounding tissue had been penetrated by the tumor cells. Both tumors from mice treated with batimastat as well as control tumors showed invasive properties toward the overlying dermal layer as well as toward the peritoneal musculature. Tumors 2 weeks postinjection were already invasive. No differences were apparent between tumors from treated and control groups. These results indicate that batimastat did not noticeably inhibit local invasion of MDA435/LCC6

Figure 1. Masson’s trichrome staining of MDA435/LCC6 solid tumors after two weeks of in vivo growth. This stain was used to highlight extracellular matrix components to evaluate the degree of stromal encapsulation. T indicates an area of epithelial tumor cells, while M indicates an area of musculature. In panel (a), a 2 week old control tumor is shown with adjacent surrounding musculature. In panel (b), a similar 2 week old tumor from a batimastat-treated mouse is shown. Both tumors show an invasive phenotype, with tumor cells apparent within the muscular compartment. Bars, 100 µm.

Anti-MMP treatment and stromal uPA expression 229

230


solid tumor growth nor encourage local tumor-stromal encapsulation. In situ hybridization for murine and human uPA mRNA in MDA435/LCC6 tumors In order to study the localization and relative level of expression of murine and human uPA mRNA in MDA435/LCC6 tumors in vivo, we utilized in situ hybridization. MDA435/LCC6 solid tumors from eight mice including those treated for two or four weeks with batimastat and their corresponding controls were formaldehyde-fixed, paraffin-embedded and sectioned. In situ hybridization was performed with an antisense and a sense probe for both murine and human uPA mRNA. With dark-field microscopy, distinct regions of murine uPA expression were apparent within the tumor and stroma of these tumors. Expression was localized to the stromal regions near tumor cell contact, as well as small islands of expression within the tumor (Figure 2). At higher magnification (Figure 3), murine uPA was found to be expressed by stromal cells near the tumor-stroma border, and by some endothelial-like cells. Although some endothelial cells (or cells in the vicinity of endothelial cells) did appear to express uPA, the signal was generally not associated with vascular-appearing structures, suggesting a non-endothelial stromal cell as the major expressor of uPA. Although tumors from both control and treated animals showed regions of murine uPA expression, tumors from batimastat treated animals appeared to have larger and more numerous regions of intense staining at the tumor-stromal border. This gave the appearance of a generally higher expression of murine uPA around these tumors (Figure 3). This increase in murine uPA expression was confirmed by northern blot analysis of RNA prepared from treated and untreated tumors. Murine uPA mRNA was detected in treated tumors, but was below the limit of detection in the untreated tumors (not shown). Relative levels of murine uPA expression were blindly rated by seven observers. General levels of expression were quantitated on a scale of one to four, with four representing high level of expression by in situ hybridization. Using a proportional-odds model fitted to the ratings data, the results indicated that treated specimens were 14 times (95% confidence interval, 5.5 to 37.4 times) more likely to receive a higher intensity rating than controls. This result was highly significant (p < 0.001). Expression of human

uPA mRNA was exclusively seen in the tumor cells, and no apparent difference between treated and control tumors was observed (not shown). ELISA for murine and human uPA antigen in extracts of MDA435/LCC6 tumors The murine uPA (muPA) ELISA method employed was capable of accurately measuring purified muPA over a range of 0.03 ng/ml to 8 ng/ml (Figure 4). The limit of detection, defined as the concentration of muPA corresponding to a signal 3 SD above the mean for the muPA blank, was 31 pg/ml. The ELISA signal of a standard of purified human uPA at concentrations up to 8.0 ng/ml was indistinguishable from the buffer background signal, implying high species specificity of the assay for muPA (Figure 4). Specific signal recovery was determined by addition of increasing concentrations of purified muPA to a fixed 1 : 20 dilution of a pool of MDA-MB-435 tumor xenograft extracts, and subsequent measurement of the ELISA signal. Approx. 114% recovery in diluted tumor extract of muPA signal was obtained (data not shown). Mice with MDA435/LCC6 tumors were treated with batimastat or vehicle for either 2 or 4 weeks. The tumors were then excised, extracted and measured by the murine-specific uPA ELISA for the content of murine uPA antigen. After 2 weeks of tumor growth, the mean level ± SD of murine uPA in tumor extracts from control mice was 0.69 ± 0.22 ng/mg protein (n = 8 mice), whereas batimastat-treated tumors showed a mean level of murine uPA of 1.62 ± 0.42 ng/mg protein (n = 5 mice), corresponding to a significant increase of approx. One hundred thirty five percent compared to the control group (p < 0.005). After 4 weeks of tumor growth, the mean level of murine uPA in tumor extracts from control mice was 2.12 ± 0.84 ng/mg protein (n = 9 mice), while batimastat-treated tumors showed a mean level of murine uPA of 3.54 ± 1.78 ng/mg protein (n = 8 mice) (Figure 5), corresponding to a significant increase of approx. 67% compared to the control group (p < 0.05). The tumor extracts were, in addition, measured for their content of human uPA antigen. In all samples, the level of human uPA antigen was below the assay detection limit of 25 pg/ml. In this human uPA ELISA, the signal from a standard of purified murine pro-uPA at concentrations up to 2 ng/ml was not different from buffer background (data not shown).

Figure 2. Expression of murine uPA as detected by in situ hybridization: comparison of batimastat-treated and control tumors. Control tumors are pictured in panels (a, b). Batimastat-treated tumors are pictured in panels (c, d). Matched bright and dark field images are shown in panels (a) and (b) and (c) and (d). Expression of murine uPA is seen primarily at the tumor-stromal border, with small regions of expression clustered within the tumor in both control and batimastat-treated sections. Dark field illumination reveals brighter, more dense regions of signal from murine uPA expression in treated tumors. Bars, 100 µm.

Anti-MMP treatment and stromal uPA expression 231

232


Figure 4. Species specificity of uPA ELISA. Standard curve for purified murine pro-uPA added in increasing concentrations to dilution buffer (•). A parallel experiment was performed by adding increasing concentrations of purified human uPA ( ). Values shown are means ± SD of duplicates. The linear correlation coefficient was > 0.999.

host, we examined muPA by ELISA of mouse organs from vehicle treated and batimastat treated animals. As seen from Figure 6, no differences were detected between the two treatment groups with regard to lungs, dermis and kidney. A weak, but statistically significant difference was seen when analysing the liver tissue (p = 0.04), with a slight muPA induction by batimastat. A more pronounced muPA induction was seen in the uterine tissue of batimastat treated mice (p < 0.01). Effect of batimastat on primary tumor growth in uPA – gene disrupted nude mice Figure 3. High-power magnification of murine uPA in situ hybridization, bright field. (a), expression of murine uPA in a control MDA435/LCC6 tumor. (b), expression of murine uPA in a batimastat-treated tumor. Expression is seen in areas of tumor tissue adjacent to muscular compartment invasion. (c), expression of murine uPA in a batimastat-treated tumor, with increased expression seen in an endothelial cell (indicated with arrow), but also in other stromal cells. Magnification is at 100 ×.

ELISA for murine uPA antigen in extracts of mouse organs In order to investigate whether the murine uPA upregulation observed in tumors from batimastat treated mice was exclusively a tumor-specific interaction or whether it represented a more general effect on the

Because of our observations linking tumor progression and invasion with induction of stromal uPA, we wished to further test the biological importance of stromal uPA in the system by making use of our newly developed combined uPA−/− nude mouse model. The model allows for a controlled study of the role of stromal uPA because uPA+/+ nude mouse littermates are available for direct comparisons. When MDA435/LCC-6 cells were inoculated into littermate uPA+/+ nude mice all animals developed growing tumors. Treatment with batimastat resulted in retardation of tumor growth to a similar degree as previously described [7]. When inoculating the human cancer cells into athymic uPA−/− animals, all mice developed tumors, but the growth rate was significantly slower as compared to growth in littermate uPA+/+

Anti-MMP treatment and stromal uPA expression

233

Figure 5. Murine uPA antigen concentrations in extracts of MDA435/LCC6 human breast tumor xenografts from mice treated daily i.p. with vehicle (n = 8 mice) or 50 mg/kg of batimastat (n = 5 mice) for 2 weeks. ∗ p < 0.005 versus vehicle. In another experiment, mice were treated daily i.p. with vehicle (n = 9 mice) or 50 mg/kg of batimastat (n = 8 mice) for 4 weeks. ∗∗ p < 0.05 versus vehicle. Values shown are means ± SD.

Figure 6. Murine uPA antigen concentrations in extracts of mouse organs following daily i.p. treatment of mice for 2 weeks with either vehicle (n = 5 mice) or batimastat (50 mg/kg) (n = 6 mice). ∗ p = 0.04 versus vehicle. ∗∗ p < 0.01 versus vehicle. Values shown are mean ± SD.

mice (p < 0.001) (Figure 7). In order to investigate whether the addition of batimastat treatment to the host uPA deficiency would result in a further tumor growth inhibition, tumor cells were inoculated into uPA−/− mice which received treatment with batimastat or vehicle alone. As seen in Figure 7, this combination resulted in a tumor growth inhibition which did not exceed the growth inhibition observed in vehicle treated uPA−/− mice (p > 0.1), while still being significantly different from tumors grown in wild-type littermates (p = 0.01).

Discussion A number of synthetic MMP inhibitors have been developed for clinical application as anti-cancer drugs [26]. These compounds have been designed so they fit tightly within the active site of the MMP. The zinc atom in this active site is then chelated through a zinc binding group which in many of the inhibitors under development, is a hydroxamate group. Inhibition of MMPs by low molecular weight synthetic compounds such as batimastat has been shown to be effective in

234


Figure 7. Growth curves of MDA435/LCC6 tumors in combined uPA gene disrupted and immunedeficient mice and their wild type littermates.

abrogating tumor growth, metastasis, and angiogenesis in a variety of experimental animal tumor models [3–6, 27, 28]. In addition, previously these inhibitors have been described to induce tumor encapsulation by stromal components, indicating an effect of the drug to alter the tumor-matrix environment [27, 28]. In a recent study we showed that when injected daily over a period of 4 weeks, the MMP inhibitor batimastat retarded solid tumor growth in our breast carcinoma solid tumor model [7]. Surprisingly, in this study, no discernible differences in the stromal encapsulation of the batimastat-treated tumors versus untreated tumors were seen. In fact, we noted that treated tumors appeared to be at least as invasive as their non-treated counterparts. This led us to consider the role of other matrix protease systems that may be necessary for local invasion. Expression of uPA has been correlated with invasion and metastasis in a number of tumor studies [10, 12, 29, 30]. uPA can influence tumor progression in at least three different ways: activation of plasminogen to plasmin, which is a broad spectrum proteolytic enzyme that can degrade many of the components of the extracellular matrix [31]; plasmin activation of several pro-MMPt’s to active enzymes [13, 14]; plasmin activation of or release of growth factors [15, 16]. In the present study, we investigated the expression of uPA in control and batimastat treated MDA435/LCC6 xenografted tumors. We have previously shown, in another human breast cancer xenograft (MDA-MB-231), that the human breast cancer

cells induce the appearance of murine uPA and uPAR in tumor infiltrating mouse stromal cells [12]. Thus, since the xenografted tumors included in the present study consist of both the human epithelial cancer cells and murine stromal cells, the probes used for in situ hybridization and the antibodies used for ELISA needed to exhibit strict species specificity in order to allow us to distinguish between uPA derived from the tumor cells and from the host. Using a murine specific uPA probe for in situ hybridization, we observed that expression of murine uPA occured in cells surrounding the tumor, with foci of high expression. These areas correlated with regions of stromal components within the tumor, such as fibroblast-like cells, and with regions of tumor invasiveness into surrounding tissue. Interestingly, in tumors from batimastat treated mice, we found that the mRNA expression of murine uPA appeared generally increased, with more foci of high expression at the tumor-stromal interface. While expression of murine uPA was found in regions around tumor cells in both control and treated mice, the expression was more prevalent in the treated tumors, especially in areas of tumor invasion. There were some individual or small groups of cells that appeared to express more murine uPA in or around treated tumors than in control tumors. In addition, there also appeared to be more stromal cells expressing murine uPA in the treated tumors than in control tumors. Thus, not only does murine uPA mRNA expression appear to be increased in the stromal cells, but also there appears to be an

Anti-MMP treatment and stromal uPA expression increase in the number of stromal cells expressing uPA mRNA (Figure 3). Although it is not clear exactly which stromal population of cells is expressing the increased levels of uPA, they are cells that are in direct contact with tumor cells. Specifically, the highest expression is seen at tumor-stromal borders at regions of tumor invasion, as well as within the tumor itself. It is possible that a high concentration of inflammatory cells may be found in this area. Although expression of uPA is apparent in some endothelial nuclei, the predominant expression of this protease is in non-endothelial cells (Figure 3c). Stromal fibroblasts appear to be recruited to make uPA in these regions of local tumor invasion, which is in accordance with other studies in which tumor-associated uPA expression was localized to stromal myofibroblasts [10], tissue macrophages [32, 33], and tumor endothelial cells [32, 34]. In contrast to murine derived stromal uPA, no apparent difference in expression of human cancer cell derived uPA was observed by the in situ hybridization, suggesting differences in the effect of batimastat influenced uPA expression in these two cell compartments or differences related to the species investigated. In order to substantiate the in situ hybridization data, a murine specific uPA ELISA and a human specific uPA ELISA were applied. The murine uPA ELISA, developed for this study, showed high sensitivity for detection of murine uPA, whereas no signal was detected from human uPA in concentrations up to 8 ng/ml, verifying high specificity of this assay for murine uPA. While no differences in human uPA were noted between control and treated tumors, batimastat treatment significantly induced increased murine uPA antigen levels. The protein data thus support the results from the in situ hybridization. When comparing muPA content in various organs from vehicle and batimastat treated animals, no general effect of the batimastat treatment on the host tissue was seen. However, when examining uterine tissue, a pronounced increase in muPA was seen in batimastat treated animals. The uterine tissue undergoes significant tissue remodeling during the estrous cycle and this remodeling has been shown to involve proteases [35]. Thus, it might be that in tissue which frequently and regularly undergoes physiological tissue remodeling requiring the presence and activity of various proteases, there may be an up-regulation of one protease, following treatment, that selectively suppresses the function of another protease system. In order to further investigate any biological role of host (murine) uPA in the tumor tissue, we generated

235

combined immunodeficient and uPA-gene disrupted mice. By inoculating MDA435/LCC-6 cells into the uPA−/− mice and their littermate wild-type controls, a significant reduction in tumor growth was observed in the uPA−/− mice. This observation directly points to a functional role of host uPA in regulating tumor growth. However, the combination of batimastat treatment and uPA deficiency did not result in further tumor growth retardation, further underscoring the role of the uPA system as the predominant protease system in governing progression and invasion of this tumor model. Cell surface uPA generated plasmin has been shown to activate a number of pro-MMP’s [14] and one possible explanation to our observation is that without uPA very little or no plasmin is formed on the cell surface, which in turn will result in a reduced proMMP activation. Accordingly, the effect of batimastat would be less pronounced. Inhibition of MMPs has recently received much interest in cancer research. Batimastat has been used as an anti-tumor agent in human clinical trials [36] and other MMP inhibitors, such as marimastat, are now under study for anti-tumor activity [37, 38]. Our study indicates that there may be an interaction between the MMP and uPA systems in the regulation of overall protease activity in response to batimastat treatment. This highlights the possibility of redundancy or functional overlap in tumor proteolytic systems. Specifically, uPA production could serve as a tumor escape mechanism, for example as a mechanism of drug resistance under treatment with an MMP inhibitor. Further studies into the mechanisms of this response are warranted.

Acknowledgements This work was supported by the Dagmar Marshall Foundation, Else and Mogens Wedell-Wedellsborg Foundation, The Danish Cancer Society and a project grant within the Lombardi Cancer Center (LCC) Breast Cancer SPORE Grant, NIH IP50CA58185 (R.B.D.). We would like to thank Dr. Robert Clarke (LCC) for providing the MDA435/LCC6 cell line and Dr. Elizabeth A. Bone of British Biotech, Inc. (Oxford, UK) for supplying batimastat and for her advice on these studies. We also thank Geraldine Natarajan (LCC) for performing the northern blot analysis. Finally, we appreciate the help of Dr. John Hanfelt (LCC) with the statistical analyses in the study.

236


References 1. Stetler-Stevenson WG, Hewitt R, Corcoran M: Matrix metalloproteinases and tumor invasion: from correlation and causality to the clinic. Semin Cancer Biol 7: 147–154, 1996 2. Brown PD, Giavazzi R: Matrix metalloproteinase inhibition - a review of anti-tumour activity. Ann Oncol 6: 967–974, 1995 3. Wang X, Fu X, Brown PD, Crimmin MJ, Hoffman RM: Matrix metalloproteinase inhibitor BB-94 (batimastat) inhibits human colon tumor growth and spread in a patient-like orthotopic model in nude mice. Cancer Res 54: 4726–4728, 1994 4. Watson SA, Morris TM, Robinson G, Crimmin MJ, Brown PD, Hardcastle JD: Inhibition of organ invasion by the matrix metalloproteinase inhibitor batimastat (BB-94) in two human colon carcinoma metastasis models. Cancer Res 55: 3629–3633, 1995 5. Sledge-GW J, Qulali M, Goulet R, Bone EA, Fife R: Effect of matrix metalloproteinase inhibitor batimastat on breast cancer regrowth and metastasis in athymic mice. J Natl Cancer Inst 87: 1546–1550, 1995 6. Taraboletti G, Garofalo A, Belotti D, Drudis T, Borsotti P, Scanziani E: Inhibition of angiogenesis and murine hemangioma growth by batimastat, a synthetic inhibitor of matrix metalloproteinases. J Natl Cancer Inst 87: 293–298, 1995 7. Low JA, Johnson MD, Bone EA, Dickson RB: The matrix metalloproteinase inhibitor batimastat (BB-94) retards human breast cancer solid tumor growth but not ascites formation in nude mice. Clin Can Res 2: 1207–1214, 1996 8. Stephens RW, Brünner N, Jänicke F, Schmitt M: The urokinase plasminogen activator system as a target for prognostic studies in breast cancer. Breast Cancer Res Treat 52: 99–111, 1998 9. Andreasen PA, Kjøller L, Christensen L, Duffy MJ: The urokinase-type plasminogen activator system in cancer metastasis – a review. Int J Cancer 72: 1–22, 1997 10. Nielsen BS, Sehested M, Timshel S, Pyke C, Danø K: Messenger RNA for urokinase plasminogen activator is expressed in myofibroblasts adjacent to cancer cells in human breast cancer. Lab Invest 74: 168–177, 1996 11. Grøndahl-Hansen J, Ralfkiaer E, Kirkeby LT, Kristensen P, Lund LR, Danø K: Localization of urokinase-type plasminogen activator in stromal cells in adenocarcinomas of the colon in humans. Am J Pathol 138: 111–117, 1991 12. Rømer J, Pyke C, Lund LR, Eriksen J, Kristensen P, Rønne E, Høyer-Hansen G, Danø K, Brünner N: Expression of uPA and its receptor by both neoplastic and stromal cells during xenograft invasion. Int J Cancer 57: 553–560, 1994 13. Werb Z, Mainardi CL, Vater CA, Harris ED, Jr: Endogenous activiation of latent collagenase by rheumatoid synovial cells. Evidence for a role of plasminogen activator. N Engl J Med 296: 1017–1023, 1977 14. Carmeliet P, Moons L, Lijnen R, Baes M, Lemaitre V, Tipping P, Drew A, Eeckhout Y, Shapiro S, Lupu F, Collen D: Urokinase-generated plasmin activates matrix metalloproteinases during aneurysm formation. Nat Genet 17: 439–444, 1997 15. Lyons RM, Gentry LE, Purchio AF, Moses HL: Mechanism of activation of latent recombinant transforming growth factor beta 1 by plasmin. J Cell Biol 110: 1361–1367, 1990 16. Saksela O, Rifkin DB: Release of basic fibroblast growth factor-heparan sulphate complexes from endothelial cells by plasminogen activator-mediated proteolytic activity. J Cell Biol 110: 767–775, 1990

17. Leonessa F, Green D, Licht T, Wright A, Wingate-Legette K, Lippman J, Gottesman MM, Clarke R: MDA435/LCC6 and MDA435/LCC6MDR1: ascites models of human breast cancer. Br J Cancer 73: 154–161, 1996 18. Carmeliet P, Schoonjans L, Kieckens L, Ream B, Degen J, Bronson R, De Vos R, van den Oord JJ, Collen D, Mulligan RC: Physiological consequences of loss of plasminogen activator gene function in mice. Nature 368: 419–424, 1994 19. Kristensen P, Eriksen J, Blasi F, Danø K: Two alternatively spliced mouse urokinase receptor mRNAs with different histological localization in the gastrointestinal tract. J Cell Biol 115: 1763–1771, 1991 20. Camiolo SM, Siuta MR, Madeja JM: Improved medium for extraction of plasminogen activator from tissue. Prep Biochem 12: 297–305, 1982 21. Declerck PJ, Verstreken M, Collen D: Immunoassay of murine t-PA, u-PA and PAI-1 using monoclonal antibodies raised in gene-inactivated mice. Thromb Haemost 74: 1305–1309, 1995 22. Solberg H, Rømer J, Brünner N, Holm A, Sidenius N, Danø K, Høyer-Hansen G: A cleaved form of the receptor for urokinase-type plasminogen activator in invasive transplanted human and murine tumors. Int J Cancer 58: 877–881, 1994 23. Benraad TJ, Geurts-Moespot J, Grøndahl-Hansen J, Schmitt M, Heuvel JJTM, de Witte JH, Foekens JA, Leake RE, Brünner N, Sweep CGJ: Immunoassays (ELISA) of urokinase-type plasminogen activator (uPA) – report of an EORTC/BIOMED1 work-shop. Eur J Cancer 32A: 1371–1381, 1996 24. McCullagh P, Nelder JA: Generalized Linear Models. In: Chapman, Hall (eds), New York, 1989, pp 152–152 25. Diggle PJ, Liang K, Zeger SL: Analysis of Longitudinal Data. Clarendon Press, 1994, pp 201–202 26. Brown PD: Clinical studies with matrix metalloproteinase inhibitors. Acta Pathol Microbiol Scand A 107: 174–180, 1999 27. Davies B, Brown PD, East N, Crimmin MJ, Balkwill FR: A synthetic matrix metalloproteinase inhibitor decreases tumor burden and prolongs survival of mice bearing human ovarian carcinoma xenografts. Cancer Res 53: 2087–2091, 1993 28. Chirivi RG, Garofalo A, Crimmin MJ, Bawden LJ, Stoppacciaro A, Brown PD, Giavazzi R: Inhibition of the metastatic spread and growth of B16-BL6 murine melanoma by a synthetic matrix metalloproteinase inhibitor. Int J Cancer 58: 460–464, 1994 29. Grøndahl-Hansen J, Christensen IJ, Rosenquist C, Brünner N, Mouridsen HT, Danø K, Blichert-Toft M: High levels of urokinase-type plasminogen activator and its inhibitor PAI-1 in cytosolic extracts of breast carcinomas are associated with poor prognosis. Cancer Res 53: 2513–2521, 1993 30. Sappino AP, Busso N, Belin D, Vassalli JD: Increase of urokinase-type plasminogen activator gene expression in human lung and breast carcinomas. Cancer Res 47: 4043–4046, 1987 31. Danø K, Andreasen PA, Grøndahl-Hansen J, Kristensen P, Nielsen LS, Skriver L: Plasminogen activators, tissue degradation, and cancer. Adv Cancer Res 44: 139–266, 1985 32. Jankun J, Merrick HW, Goldblatt PJ: Expression and localization of elements of the plasminogen activation system in benign breast disease and breast cancers. J Cell Biochem 53: 135–144, 1993 33. Hildenbrand R, Dilger I, Horlin A, Stutte HJ: Urokinase and macrophages in tumour angiogenesis. Br J Cancer 72: 818– 823, 1995 34. Montesano R, Pepper MS, Möhle-Steinlein U, Risau W, Wagner EF, Orci L: Increased proteolytic activity is responsible

Anti-MMP treatment and stromal uPA expression

35.

36.

37.

for the aberrant morphogenetic behavior of endothelial cells expressing the middle T oncogene. Cell 62: 435–445, 1990 Hulboy DL, Rudolph LA, Matrisian LM: Matrix metalloproteinases as mediators of reproductive function. Mol Hum Reprod 3: 27–45, 1997 Wojtowicz-Praga S., Low J, Marshall J, Ness E, Dickson R, Barter J, Sale M, McCann P, Moore J, Cole A, Hawkins MJ: Phase I trial of a novel matrix metalloproteinase inhibitor batimastat (BB-94) in patients with advanced cancer. Invest New Drugs 14: 193–202, 1996 Wojtowicz-Praga S, Torri J, Johnson M, Steen V, Marshall J, Ness E, Dickson R, Sale M, Rasmussen HS, Chiodo TA,

237

Hawkins MJ: Phase I trial of Marimastat, a novel matrix metalloproteinase inhibitor, administered orally to patients with advanced lung cancer. J Clin Oncol 16: 2150–2156, 1998 38. Benaud C, Dickson RB, Thompson EW: Roles of the matrix metalloproteinases in mammary gland development and cancer. Breast Cancer Res Treat 50: 97–116, 1998

Address for offprints and correspondence: Nils Brünner, The Finsen Laboratory, Strandboulevarden 49, Bld 86.2, DK-2100 Copenhagen Ø, Denmark; Tel.: 45 35 45 56 06; Fax: 45 35 38 54 50; E-mail: [email protected]