Expression of Heparanase by Primary Breast Tumors Promotes Bone Resorption in the Absence of Detectable Bone Metastases

Metastasis to bone and osteolytic lesions caused by the overactivity of tumor-stimulated osteoclasts are major complications of breast cancer. These events are regulated by several growth-regulatory factors, including heparan sulfates. Heparan sulfates are present within the tumor microenvironment as components of heparan sulfate proteoglycans or as free heparan sulfate chains. These highly anionic sugars act to fine-tune the activities of growth factors and chemokines, such as those that regulate tumor cell growth, angiogenesis, and osteoclastogenesis (1). Syndecan-1 is a major heparan sulfate proteoglycan of breast cancer cells (2) and is also found in the bone marrow. Clinically, high levels of syndecan-1 expression in human breast tumors are associated with poor prognosis and an aggressive phenotype (2). Syndecan-1 can also be induced in reactive stroma responding to breast cancer (3). Moreover, when breast cancer cell lines were grown together with mouse fibroblasts, the fibroblasts were induced to express syndecan-1 to high levels (4). The heparan sulfate present in tumors of the breast apparently promotes growth and may be structurally and functionally different than the heparan sulfate derived from normal breast. For example, the heparan sulfate of breast cancer promotes formation of fibroblast growth factor (FGF)-2 receptor complexes to a greater degree than heparan sulfate from normal breast epithelium (5). This may be due in part to the action of heparan sulfate–modifying enzymes, such as heparanase.

Heparanase cleaves heparan sulfate chains with an endo-β-d-glucuronidase activity releasing activated fragments of heparan sulfate that apparently mediate its growth and angiogenic effects by acting on tumor cells and endothelial cells (1, 6). Heparanase is synthesized by cells as a M_r 65,000 protein that is processed to a fully active M_r 50,000 form (7, 8). The active enzyme cleaves the glycosidic bonds of heparan sulfate at relatively few places producing fragments that are usually 10 to 20 sugar residues in length (9). These fragments are large enough to interact with growth factors but are not bound to the extracellular matrix or to cell surfaces. In addition, the cleavage of heparan sulfate contributes to erosion of basement membrane barriers, thereby facilitating invasion and metastasis (10, 11). Indeed, heparanase has been directly implicated in promoting invasiveness (7, 12, 13), angiogenesis (14, 15), and metastasis (7, 8, 16–18). Heparanase seems to play an important role in human breast carcinomas, where expression of heparanase correlates with large tumor size and enhanced metastatic potential (19). Elevated heparanase has also been reported to be associate with a poor prognosis in several other human cancers, including gastric, endometrial, pancreatic, and bladder cancers (16, 20–22).

We have shown previously that the expression of heparanase in myeloma cells implanted s.c. in severe combined immunodeficient (SCID) mice will promote tumor metastasis to bone (23). In the present study, we extended this model to test the effects of heparanase expression on metastasis of breast cancer cells to bone. Although direct evidence of metastases to bone were not detected in this model, a marked enhancement of osteoclastogenesis and bone turnover was discovered in animals bearing tumors that expressed heparanase compared with controls. This novel finding suggests an important role for heparanase in promoting bone resorption even when tumors are not evident in the bone.

Reagents, cell lines, and culture conditions. Tissue culture plasticware was purchased from Fisher (Pittsburgh, PA) and all tissue culture reagents were analytic grade and purchased from Sigma (St. Louis, MO). The MDA-MET cell lines (NEO and HPSE) were maintained in DMEM supplemented with 10% fetal bovine serum at 37°C in sterile culture dishes as described previously (24). Both cell lines were certified to be Mycoplasma free. Cells were subcultured by trypsinization in 0.5% trypsin (Sigma) and 0.5 mmol/L EDTA in HBSS without calcium or magnesium in a laminar flow hood during their logarithmic phase of growth. Conditioned medium (incubated for 48 hours; containing serum) from cells was collected, diluted 50% diluted in αMEM, and added to cultures of human peripheral blood mononuclear cells (PBMC; as described below).

Preparation of heparanase-expressing human breast cancer cells. MDA-MET cells were derived from MDA-MB-231 human breast adenocarcinoma cells based on their ability to home to and grow in bone (25). Heparanase cDNA (HPSE1) was subcloned in the sense direction into pIRES2-EGFP (Clontech, Palo Alto, CA) vector, which allowed both the heparanase gene and the enhanced green fluorescent protein (EGFP) to be translated from a single bicistronic mRNA as described previously for multiple myeloma cells (23, 26). The pIRES2-EGFP/HPSE1 construct was stably transfected into MDA-MET human breast cancer cells using Lipofectin reagent (Invitrogen, Carlsbad, CA) and Opti-MEM I (Life Technologies, Grand Island, NY) with 10 μg DNA (pIRES2-EGFP vector only for NEO transfections or pIERS2-EGFP/HPSE1 for HPSE transfections) following the manufacturer's instructions. The cells were selected with G418 (800 μg/mL), and at least 2.0 × 10⁷ cells were sorted by green fluorescence using a flow cytometer. The sorted cells were allowed to grow and again examined for green fluorescence after 1 week. A minimum of three sorting runs were done to achieve a minimum of 66% cells exhibiting green fluorescence in the transfectants.

Heparanase activity assay. The heparanase activity assay used an immobilized [³H]heparan sulfate substrate and was done as described (25, 26). Purified recombinant heparanase (46 ng) was used as the positive control and buffer was used as the negative control. Each sample was normalized to equal volume and tested in triplicate on at least two separate occasions.

Western blots. Cells were extracted in 50 mmol/L Tris-HCl (pH 7.5), 150 mmol/L NaCl, 0.5% Triton X-100, 0.2 mmol/L phenylmethylsulfonyl fluoride, and 10 μg/mL leupeptin for 30 minutes on ice. Extracts were centrifuged at 16,000 × g at 4°C for 10 minutes and the supernatants were analyzed for protein concentration using a BCA Protein kit (Pierce, Rockford, IL), and Western analysis was done using 60 μg protein per lane were added to the wells of SDS-PAGE gels and following transfer to nitrocellulose heparanase was detected with a mouse monoclonal antibody directed against recombinant human heparanase (26–28) and a horseradish peroxidase (HRP)–conjugated, sheep anti-mouse IgG (Amersham Biosciences, Piscataway, NJ). Immunoreactive bands were detected using a enhanced chemiluminescence system (Amersham Biosciences).

Cell growth assay. The number of cells present at specific times following plating was determined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay as described previously (29). Briefly, cells were seeded (5 × 10³ per well) in triplicate in 96-well plates. Plates were then incubated for 0, 24, and 48 hours in serum-containing or serum-free culture medium at 37°C at 5% CO₂. Cell number was determined using the MTT assay and reading the absorbance at 540 nm as described by others (30, 31). Growth data were analyzed using ANOVA among groups. Differences with P < 0.05 among groups were considered significant.

Tumor biology. Four- to -6-week-old female SCID mice were purchased at Harlan (Indianapolis, ID) and allowed to acclimate for 7 days. An identification chip (Pocket Scanner Systems, BioMedic Data Systems, Inc., Seaford, DE) was implanted s.c. and blood (0.05-0.2 mL) was harvested from the submaxillary artery using a 22 g needle. Bleeds were done before injecting tumor cells and at 1-week intervals following injection. In the first two experiments, animals (seven per group) received up to 2 × 10⁶ cells (NEO or HPSE) in each of four different injection sites within the mammary fat pads (two axillary and two abdominal). In the third experiment, two different doses of tumor cell inoculants were tested. Each animal was injected with either 1 × 10⁵ cells per injection site (4 × 10⁵ cells per animal) or 1 × 10⁶ cells per injection site (4 × 10⁶ cells per animal) into each of the four mammary fat pads.

Mice were examined every other day for palpable tumor masses. When tumors appeared, the height, length, and width of the tumors were measured using fine calipers. Tumor volumes were determined in live animals using the following formula: height × length × width of the individual tumors. When any tumor reached 2 cm in any dimension, all animals in that group were anesthetized and euthanized. All animal treatment and care protocols conformed to NIH guidelines and were done using a University of Arkansas for Medical Sciences (UAMS) Institutional Animal Care and Use Committee–approved protocol. After euthanasia, the mice were subjected to complete necropsy. All tumor tissue was harvested, measured, and weighed.

Immunohistochemistry. For microvessel density determinations, paraformaldehyde-fixed and paraffin-embedded sections of primary tumors were prepared, deparaffinized, and rehydrated. Epitope retrieval was accomplished by steaming slides for 20 minutes in citrate buffer (pH 6.0). Endogenous peroxidase activity was quenched with 3% H₂O₂ for 15 minutes at 22°C. Nonspecific protein-binding sites were blocked with 5% (w/v) nonfat dry milk in PBS. The sections (5 μm) were stained overnight with rat anti-mouse CD34 monoclonal antibody (Hycult Biotechnology, Uden, the Netherlands) at 1:20 dilution at 4°C. The sections were washed with PBS and stained with biotinylated rabbit anti-rat IgG (Vector, Burlingame, CA) for 30 minutes at 22°C. Microvessel densities were determined as described previously (26, 29).

To localize heparanase in tumor tissues, the same procedure as described above was followed using an antibody to heparanase (27, 28), except that a HRP-conjugated goat anti-mouse antibody (Vector) was used to detect this mouse monoclonal antibody (26).

Osteoclasts were identified in mouse long bones that were excised, fixed in 10% neutral-buffered formalin for 2 days, and decalcified in 5% formic acid with agitation until deemed clear by the ammonium oxalate end point test (18). The decalcified specimens were then dehydrated through graded ethanol and cleared in methyl salicylate before paraffin infiltration. Subsequently, they were embedded in paraffin, sectioned (5 μm), and stained with H&E as described previously (25, 32) and tartrate-resistant acid phosphatase (TRAP) using the Acid Phosphatase Leukocyte kit (Sigma) as described (24).

Bone resorption. Serum levels of the COOH-terminal type I collagen telopeptides was determined using the RatLaps ELISA kit (Nordic Bioscience Diagnostics A/S, Herlev, Denmark) following the manufacturer's instructions. Mouse serum samples (20 μL/well) were tested in triplicate. The MouseTRAP assay (Suomen Bioanalytiikka Oy, Turku, Finland) was used to determine serum levels of TRAP-5b as directed by the manufacturer. Mouse serum samples were diluted 1:4 (25 μL/well) and tested in triplicate.

Isolation and culture of peripheral blood mononuclear cells and their differentiation to osteoclasts. Peripheral blood was collected from healthy donors (approved by the UAMS Institutional Review Board) using heparin as an anticoagulant and 200 ng/mL RANK-Fc to minimize any priming of osteoclast progenitors by endogenous RANKL as described (24). Blood was diluted in sterile PBS (1:1) in a sterile hood. The blood-PBS solution was slowly layered over AccuPrep solution (Accurate Chemicals, Westbury, NY) and then centrifuged at 400 × g in swing buckets for 30 minutes at 21°C. The PBMC layer was collected and washed in 5 to 6 volumes of PBS, isolated by centrifugation at 140 × g, and resuspended in αMEM containing 10% fetal bovine serum. Cells were counted with a hemocytometer and plated in 48-well tissue culture plates at a concentration of 0.5 million cells in 0.5 mL volume per well. Macrophage colony-stimulating factor (mCSF; 25 ng/mL) was present in all treatment groups, including control. RANKL (25 ng/mL) was used as a positive control. In some treatment groups, conditioned medium from HPSE and NEO cells growing in culture was added. For this, medium was harvested 48 hours after cell plating, diluted 50% with αMEM, and added to the cultures of human PBMCs.

Mononuclear cell cultures were maintained at 37°C and half of the medium in each well was replaced with fresh medium thrice weekly. The experiment was terminated on day 10. Medium was aspirated and the cells were fixed with 10% formalin. TRAP staining was done (Sigma) for quantitation of TRAP-positive multinucleated cells. The number of osteoclasts present in the entire well was determined by manually counting TRAP-positive cells having more than three nuclei. The numerical average of cell counts from four replicate wells per treatment was determined and the results were expressed as the number of TRAP-positive multinucleated cells per well per treatment group (24).

To investigate the role of heparanase in an animal model of breast cancer, a subline of MDA-MB-231 human breast adenocarcinoma cells called MDA-MET were employed. MDA-MET cells were selected for their ability to colonize bone when injected into arterial blood (25). These cells were transfected with the cDNA for human heparanase (HPSE cells) or with the empty vector only (NEO cells). Transfected cells were enriched by G418 selection; because a bicistronic vector coding for EGFP was employed, further enrichment of transfectants was accomplished by fluorescence cell sorting. After multiple rounds of sorting, enriched cell populations containing at least 66% fluorescent cells were obtained for use in subsequent experiments (Fig. 1A). Western blots confirm that NEO cells constitutively express low levels of heparanase compared with HPSE cells that show high levels of the 50-kDa processed and highly active form of the enzyme and low levels of the 65-kDa unprocessed form (Fig. 1A).

Immediately before initiation of all in vivo experiments, cell extracts were analyzed to determine the level of heparanase activity expressed in HPSE cells and NEO controls. As expected, extracts of HPSE cells have high heparanase enzyme activity as measured by the release of immobilized [³H]heparan sulfate (Fig. 1B). The basal heparanase activity of the NEO controls is only slightly above background (buffer), whereas the activity in the HPSE cell extracts is significantly elevated (∼6-fold). The expression of active heparanase does not seem to alter growth properties in vitro as both NEO and HPSE cells exhibit identical growth rates over 72 hours (Fig. 1C).

Next, NEO or HPSE cells were injected into the mammary fat pads of female SCID mice. Despite the identical growth pattern of the HPSE and NEO cells in vitro (Fig. 1C), following fat pad inoculation, HPSE cells formed more rapidly growing tumors compared with the control (Fig. 2A). Similarly, tumor wet weights obtained at sacrifice confirmed that the HPSE tumors have a significant growth advantage over NEO tumors (Fig. 2B). Moreover, HPSE cell tumors have significantly higher microvessel densities (>2-fold higher) than those formed by NEO cells as measured by anti-CD34 antibody staining (Figs. 2C and 3B). The enhanced ability to promote angiogenesis may explain the initial growth advantage that the HPSE cells have over the NEO control cells. These growth- and angiogenesis-promoting abilities of heparanase are consistent with earlier work on heparanase-expressing tumor cells in other systems (14, 15, 26).

To determine if heparanase expression was maintained in vivo, tumors were excised after 2 weeks growth in the mammary fat pad and heparanase expression was analyzed by Western blotting. HPSE cells retain relatively high levels of enzyme expression in vivo and NEO cells maintain low levels of heparanase expression (Fig. 3A). This finding was confirmed by direct immunohistochemistry of the primary tumors using an antibody to heparanase (Fig. 3B).

We have shown previously that the expression of heparanase promotes the spontaneous metastasis of myeloma cells to bone (23). To determine if heparanase has a similar effect on the metastatic potential of human breast cancer cells, we analyzed the skeleton of animals in which tumor cells had been injected into the mammary fat pad. However, even using extensive histology, immunohistology, radiology, microcomputed tomography, and positron emission tomography, we were unable to find convincing evidence of tumor cells in the skeleton (femur, tibia, and spine; data not shown). Despite the lack of evidence for tumors growing in bone, there was consistent evidence of extensive bone resorption as measured by two serum assays. Serum levels of C-telopeptide of type I collagen, a product of osteoclastic bone resorption and an indicator of increased osteoclast activity (33) as well as osteoclast-derived TRAP-5b, an osteoclast-specific enzyme, are markedly elevated in the serum of animals bearing HPSE tumors (Fig. 4A and B). This was confirmed in a separate experiment in which serum levels of TRAP-5b are shown to be elevated in the serum of animals with HPSE tumors and not in that of animals with NEO tumors (Fig. 4C). The TRAP-5b assay is particularly useful because it measures only the active enzyme and thus provides an accurate assessment of osteoclast activity at the time the serum was harvested (34). Consistent with the elevated bone remodeling observed in animals bearing tumors of HPSE cells, abundant TRAP-positive osteoclasts were identified on the bone surface of these animals (Fig. 5A).

To test the idea that soluble tumor-derived factors are capable of stimulating osteoclastogenesis in vivo, a well-established in vitro assay using human PBMCs was employed (24). Although conditioned medium from NEO cell cultures stimulates osteoclast formation, HPSE cell conditioned medium stimulates osteoclastogenesis significantly above the level of NEO conditioned medium (P = 0.037, Fig. 5B and C). The osteoclasts formed by the conditioned medium of HPSE cells are large and multinucleated and stain positive for TRAP (Fig. 5C). These cells are functional osteoclasts with the capacity to resorb bone (data not shown), as we have described previously (24). The data shown in Fig. 5B and C are representative of two separate experiments using different harvests of conditioned medium that gave reproducible results.

This work reveals a novel role for heparanase in skeletal complications that accompany breast cancer. We have discovered that breast tumors having elevated levels of heparanase and growing in the mammary fat pads promote bone remodeling before metastases or even microscopic tumor foci can be detected in the bone marrow. In contrast, animals injected with control cells that express low levels of heparanase activity exhibit no systemic increase in bone resorption even when the primary tumor burden is large. The mechanism underlying the heparanase-mediated increase in bone resorption is the stimulation of osteoclastogenesis, because abundant osteoclasts are evident in the long bones of animals bearing tumors expressing high levels of heparanase. Moreover, osteoclastogenesis is stimulated in vitro by conditioned medium from HPSE cells. This effect is not due to a direct action of heparanase on osteoclast precursors because neither the heparanase protein nor the heparanase enzyme activity is detected in the medium of these cells (data not shown). Rather, our data indicate a role for heparanase in mediating the release of osteolytic agents into the circulation. This likely occurs via release of active heparan sulfate fragments and/or release of heparan sulfate–bound growth factors that then travel through the circulation to act on bone. The release of heparan sulfate–bound osteoclastogenic factors by tumor-generated heparanase is consistent with recent studies demonstrating the important role of heparan sulfate proteoglycans, such as syndecan-1, in binding to and regulating the activity of effector molecules, such as interleukin (IL)-8, hepatocyte growth factor, FGF, and osteoprotegerin (1, 5, 35–39).

Importantly, a role for active heparan sulfate fragments directly stimulating bone turnover is consistent with the well-established observation that treatment with heparin, a highly sulfated form of heparan sulfate, often causes bone resorption and decreased bone density in humans (40, 41). Heparin has also been shown to induce the loss of trabecular bone by increasing osteoclastogenesis and osteoclast activity in experimental animals (42, 43). These effects are mediated by a synergistic interaction of heparin with IL-11 resulting in Stat3 activation that in turn stimulates osteoclast formation (44). The breast cancer cells used in our model express elevated heparanase and are known to express the osteoclastogenesis-stimulating factor IL-8 (24, 25). IL-8 binds to heparan sulfate and is functionally modulated by heparan sulfate (35, 37). Therefore, activated heparan sulfate fragments produced by the heparanase-expressing cells may synergize with IL-8 to systemically promote osteoclastogenesis and bone resorption.

The finding that bone resorption is stimulated by breast cancer cells expressing high levels of heparanase before bone metastases are observed raises the possibility that tumors condition the bone marrow for metastases by first stimulating osteoclastogenesis and bone resorption. This increase in bone resorption releases numerous factors stored in the bone that fuels further tumor growth, thereby leading to continued stimulation of osteoclasts (24, 45). In breast cancer, bone metastases frequently result in lytic lesions with numerous osteoclasts and areas of bone erosion immediately surrounding foci of metastatic tumor cells (45). The progression of osteolytic bone metastases requires the establishment of functional interactions between metastatic cancer cells and bone cells (24, 46), which are presumably mediated by soluble stimulators of osteoclast activity (46, 47). Our results show that established foci of tumor cells are not necessary for the stimulation of bone resorption and support the idea that bone resorption is induced by the release of a systemic osteoclast-stimulating factor(s) from tumor cells expressing high levels of heparanase.

Unlike our findings with myeloma cells, expression of heparanase did not promote metastasis of the MDA-MB-231 subline to bone within the timeframe of these experiments. In the myeloma studies, metastases were apparent in 30 of 31 animals bearing s.c. tumors of myeloma cells expressing elevated heparanase, whereas only 3 of 23 animals bearing tumors of control cells expressing low levels of heparanase had bone metastases (23). There are several important differences between that study and the one reported here that may explain in part why bone metastases were not detected with the breast cancer cells. First, the s.c. tumors of myeloma cells expressing heparanase grow more slowly than their breast cancer cell counterparts, thereby allowing the myeloma experiments to be carried out for 7 and 9 weeks (23). The relatively slow growth may be associated with the fact that myeloma is primarily restricted to the bone marrow microenvironment and only become extramedullary late in disease progression. Thus, myeloma does not grow as well in nonbone locations. In the case of breast cancer cells implanted in the mammary fat pads, their rapid growth forced sacrifice of the animals by 21 days and in some experiments as early as 14 days after implantation of the tumor cells. Therefore, the animals may have been sacrificed before primary tumor was able to successfully seed metastatic cells in the bone. Second, myeloma cells implanted s.c. are at a site distant from their preferred microenvironment and may experience selective pressure to escape the s.c. site and home to bone. In contrast, the mammary fat pad is thought to resemble the microenvironment of the breast and the selective pressure to leave the breast and home to the bone may not have been as great.

In summary, this work reveals that expression of heparanase by a tumor distal to the bone can have a dramatic impact on bone turnover. In addition to affecting skeletal integrity, these events aid in preparing a growth-enriching bone microenvironment that will support metastatic tumor cells once they enter the bone. Thus, it may be beneficial to employ inhibitors of heparanase as an early therapeutic approach to impede progression of breast cancer as well as other bone-homing tumors.

Grant support: NIH grants CA10354 (R.D. Sanderson), CA68494 (R.D. Sanderson), and DK054044 (L.J. Suva, Co-I).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Terri Alpe (Micro-Positron Emission Core Facility, UAMS) for performing micro-positron emission tomography and Robert A. Skinner and Frances L. Swain (Center for Orthopedic Research, UAMS) for preparing histologic samples of bones to investigate possible bone metastases.