Humoral Bone Morphogenetic Protein 2 Is Sufficient for Inducing Breast Cancer Microcalcification

Abstract

Microcalcifications are an important diagnostic marker for breast cancer on mammograms, yet the mechanism of their formation is poorly understood. Indeed, there is presently no short-latency, high-yield, syngeneic rodent model of the process. Bone morphogenetic protein 2 (BMP-2) is a key mediator of physiologic bone formation and pathologic vasculature calcification, but its role in breast cancer microcalcification is unknown. In this study, R3230 rat breast tumors were adapted to cell culture, transduced with adenoviral BMP-2, and inoculated into a syngeneic host. Tumor growth and calcium salt deposition were quantified in living animals over time using micro–computed tomography and probed chemically using near-infrared fluorescence. Plasma BMP-2 levels were quantified over time by enzyme-linked immunosorbent assay. Within 3 weeks, 100% of the breast tumors developed microcalcifications, which were absent from all normal tissues. Importantly, when two tumors were initiated in a single host, the ipsilateral tumor expressing BMP-2 was able to induce microcalcification in the contralateral tumor that was not expressing BMP-2, suggesting that BMP-2 can act humorally. Taken together, we describe the first reproducible rodent model of breast cancer microcalcification, prove that BMP-2 expression is sufficient for initiating the process, and lay the foundation for a new generation of targeted diagnostic agents.

BREAST CANCER is a leading cause of cancer-related deaths, with 182,460 women in the United States expected to be diagnosed with the disease in 2008.¹ Despite its deficiencies, diagnostic x-ray mammography² remains the most widely available screening tool for the early detection of breast cancer and reduces breast cancer mortality rates by 20 to 35% in women aged 50 to 69 years and slightly less in women aged 40 to 49 years.³

A hallmark of breast cancer on diagnostic mammography is microcalcification, which occurs in 30 to 50% of patients.⁴ Although mammography cannot distinguish the chemical form of the calcium salts present and therefore relies on the pattern of crystal deposition,⁵ microcalcification is actually of two major types. Type I crystals, found more frequently in benign ductal cysts, are birefringent and colorless and are composed of calcium oxalate.⁴ Type II crystals, most often seen in proliferative lesions and associated with breast cancer cells, are nonbirefringent and basophilic and are composed of calcium hydroxyapatite (HA).⁶ HA itself appears to have potent bioactivity, inducing mitogenesis and increased enzymatic activity of matrix metalloproteinases (MMPs) MMP-2 and MMP-9.⁷ These responses can be blocked by inhibiting calcium phosphate crystallization with phosphocitrate.⁸

Despite its prevalence and importance in breast cancer, the mechanism of HA deposition is poorly understood. In fact, there is presently no reproducible, high-penetrance animal model available for studying the molecular pathways of breast cancer microcalcification. Although multiple rodent models of breast carcinogenesis have been developed with the use of mouse mammary tumor virus (MMTV) vectors or BRCA1/BRCA2 knockouts,^9–12 microcalcification of these tumors is an extremely rare and unpredictable event. A straightforward, reproducible, syngeneic, short-latency rodent model of breast cancer microcalcification could potentially have a major impact on the development of improved diagnostic imaging for the disease.

Given that many physiologic processes are conserved in malignancies, a signal transduction pathway of particular interest in breast cancer microcalcification is that mediated by bone morphogenetic protein 2 (BMP-2). Synthesized as a 60 kDa precursor, BMP-2 is processed intracellularly to a small 18 kDa monomer that is secreted.^13,14 BMP-2 is a member of the transforming growth factor β superfamily and is involved in embryogenesis and morphogenesis of various tissues and organs.^15–17 It is also a key mediator of calcium salt mineralization. Recombinant adenoviruses expressing BMP-2 can effectively induce orthotopic ossification when either adenovirus-transduced osteoblast progenitors or the viral vectors themselves are injected into the quadriceps of athymic mice.^18,19 Preclinical studies have also shown that ectopic BMP-2 expression can be used in various therapeutic interventions, such as bone defects, nonunion fractures, spinal fusion, osteoporosis, and root canal surgery.^20–23 Locally overexpressed BMP-2 can also induce calcification in MCF-7 xenograft breast tumors,²⁴ although the kinetics, efficiency, chemical form, and reproducibility of this model have not been described.

The present study was designed to determine if BMP-2 could be playing a role in breast cancer microcalcification, to determine whether it exerted its biological activity systemically or only locally, and to develop a reproducible, robust animal model for the testing of new diagnostic agents for breast cancer.

Materials and Methods

Adaptation of the R3230 Mammary Adenocarcinoma to Cell Culture

Serially passaged R3230 mammary tumors²⁵ were the generous gift of Dr. Ralph Weissleder (Center for Molecular Imaging Research, Massachusetts General Hospital, Boston, MA). To adapt the cells to cell culture, fresh tumor (measuring approximately 1 cm in diameter) was harvested from a Fischer 344 female rat host. Within 30 minutes of tumor resection, the tumor was homogenized with a tissue grinder (Model 23, Kontes Glass Co., Vineland, NJ) using aseptic technique and suspended in 7 mL of RPMI 1640 (INC Biomedicals, Aurora, IL) in a sterile 50 mL conical centrifuge tube. The tumor cell suspension was centrifuged at 700g for 10 minutes and washed twice with 20 mL of phosphate-buffered saline (PBS). Cell pellets were suspended in 4 mL of 0.25% w/v trypsin-versene (Cambrex, Walkersville, MD) solution and incubated at 37°C with 5% CO₂ for 30 minutes. The trysinversene digestion was stopped by adding 30 mL of complete RPMI 1640 media containing 10% fetal bovine serum (FBS) (Gemini Bio-products, Calabasas, CA) and antibiotics (penicillin [100 U] and streptomycin [100 mg]). The cell suspension was transferred to a sterile 15 mL conical centrifuge tube, and the cells were centrifuged at 700g for 10 minutes at room temperature. The supernatant was discarded carefully so as not to disturb the cell pellet. Cell pellets were resuspended in 5 mL of RPMI + 10% FBS, and 100 µL was used for Coulter counting of cell number. Cells (2 × 10⁶) were seeded onto a T-25 flask with 10 mL of RPMI 1640 medium supplemented with 10% FBS. Cells were fed twice weekly, during which time, R3230 tumor cells began to outgrow other contaminating cells, such as fat cells, blood cells, endothelial cells, and primary cells. Cells were split weekly at a 1:3 ratio into new flasks, and within 2 to 3 weeks, a uniform monolayer of R3230 cells, with their distinctive morphology, was obtained. Cells were further expanded and stocks stored in 90% FBS + 10% dimethyl sulfoxide in LN₂.

Polymerase Chain Reaction Amplification and Isolation of the Human BMP-2 Coding Sequence

The complementary deoxyribonucleic acid (DNA) for human BMP-2A (originally cloned from osteosarcoma cell line U-2 OS) was purchased from the American Type Culture Collection (ATCC, Rockville, MD) as a lambda gt10 phage. Host strain Escherichia coli c600 (ATCC) was used for phage plaque formation, and the human BMP-2 coding sequence was directly polymerase chain reaction (PCR)-amplified from phage plaques. Briefly, a plaque was touched with a sterile tip and transferred into 10 µL of 10 µM MgSO₄ in H₂O. Each 50 µL PCR reaction contained 1 µM of each primer, 0.5 mM each deoxyribonucleotide triphosphate (dNTP), 1 µL (5 U/µL) of Deep Vent DNA polymerase, and 2 µL of phage plaque suspension. PCR primers were 5′-GGGTTAACCTCGAGATGGTGGC ***CGGGACCCGCTG-3’ and 5′-GGGTTAACTCTAGA GCGACACCCACAACCCTCCA-3′. XhoI and XbaI sites, respectively, are shown underlined within the sequences. PCR was performed as follows: 95°C for 2 minutes, 28 cycles of PCR (94°C for 20 seconds, 55°C for 30 seconds, and 72°C for 90 seconds), extension at 72°C for 10 minutes, and stopping at 4°C. PCR products were purified using a Qiaquick PCR purification kit (Qiagen, Valencia, CA) for DNA sequencing, restriction enzymes digestion, and subsequent cloning.

Production of AdGFP/BMP-2 Adenoviruses and Cell Transduction

To facilitate the expression of BMP-2 in breast cancer cells, human BMP-2 was first cloned into a shuttle vector, pAdTrack-CMV.²⁶ pAdTrack-CMV also coexpresses the enhanced Aequorea victoria green fluorescent protein (GFP), which can be used as a convenient surrogate for adenovirus infection and gene transduction. The resultant plasmid (pAdTrackCMV-BMP2) was verified by bidirectional DNA sequencing, linearized by digestion with restriction endonuclease PmeI, and cotransfected with the adenoviral backbone vector pAdEasy-1²⁶ into E. coli BJ5183 competent cells. Recombinants were selected for kanamycin resistance, and recombination was confirmed by PCR and multiple restriction endonuclease analysis.

The resultant adenoviral vector pAdETC/BMP2 was linearized with PacI and transfected into HEK293 cells for adenovirus production. Two weeks after initial transfection, cells were harvested, pelleted, suspended in PBS (4 mL for up to 10 10-cm plates of HEK293 cells), and lysed by three freeze/thaw cycles. The lysate was then used to reinfect HEK293 cells for virus amplification. For large-scale production, final lysates were placed in preprepared dialysis tubes (Spectra/Por 12–14 kDa molecular weight cut off [MWCO], Spectrum Rancho Dominguez, CA), which were placed in dry PEG 8000 powder on ice for several hours until the desired volume (2 mL) was reached. This method concentrated the virus without loss of titer. Virus titration was performed as described previously.²⁷ Typical titers of final adenovirus lysate preparations used for R3230 cell infection were 1.4 × 10¹¹ PFU/mL. For long-term storage, lysates were adjusted to 10% final glycerol, flash-frozen in LN₂, and stored at −80°C.

R3230 cells grown at 70 to 80% confluence on 10 cm diameter tissue culture plates were washed once with Dulbecco's Modified Eagle's Medium (DMEM), incubated with adenovirus at a multiplicity of infection (MOI) of 500:1 for 2 hours, and then returned to DMEM containing 10% FBS (Gemini Bio-products, Calabasas, CA) for 48 hours.

Breast Tumor Induction in Syngeneic Hosts

All animals were used under the supervision of an approved institutional protocol. Anesthesia was induced using an intraperitoneal (IP) injection of a mixture of 50 mg/kg ketamine hydrochloride (Ketaject, Phoenix Pharmaceutical, St. Joseph, MO) and 5 mg/kg xylazine hydrochloride (Bayer, Shawnee Mission, KS). Booster anesthesia injections at one-tenth the dose were administered every 30 to 60 minutes IP as needed.

After harvesting adenovirus-transduced cells with trypsin/ethylenediaminetetraacetic acid and washing with PBS, pH 7.4, 2 × 10⁷ adenovirus-transduced R3230 cells suspended in 350 µL of DMEM were slowly implanted subcutaneously into the mammary fat pad of syngeneic Fischer 344 female rats (Taconic Farms, Germantown, NY) using a 1-cc syringe and a 11/2-inch 21-gauge needle. Tumor growth was initially assessed by palpation every 3 days, with tumors forming a raised hard nodule over the otherwise smooth breast plate, and by micro-CT and fluorescence imaging as described throughout. At the time of injection, rats averaged 7 to 9 weeks of age and weighed 150 ± 20 g. For comparison of microcalcification formation in adenoviral green fluorescent protein (AdGFP)/BMP-2 versus AdGFP-transduced tumors, eight rats in each group were used. Four rats were used for bilateral tumor formation. Eight rats per group were used for the time course study of GFP expression. Resected tumors were fixed in formalin, processed for histology, and stained using hematoxylin and eosin or the method of von Kossa.²⁸

Immunofluorescence Analysis of BMP-2 and GFP Expression in R3230 Cells

For immunofluorescence analysis of cellular BMP-2 expression, R3230 cells were seeded at 3 × 10⁴ into a 12-well plate onto glass cover slips and cultured for 24 hours in DMEM containing 10% FBS and penicillin (100 U/mL) and streptomycin (100 µg/mL). Cells were then infected with AdGFP/BMP-2 or AdGFP at an MOI of 10⁶ and 0 (mock) for 24 hours, fixed with 2% paraformaldehyde in PBS for 15 minutes, washed with PBS, and then incubated with 5 µg/mL of an affinity-purified goat polyclonal antibody specific for human BMP-2/4 (Fitzgerald Industries Intl., Concord, MA), in PBS supplemented with 1% nonfat dried milk, for 1 hour at room temperature. Cells were then thoroughly washed with PBS and stained with 5 µg/mL of a Cy3-conjugated donkey antigoat IgG (Jackson ImmunoResearch Laboratories Inc., West Grove, PA) for 1 hour in the dark, washed with phosphate-buffered saline containing 0.05% Tween 20 (PBS-T) and DAPI for 5 minutes, and mounted with Fluoromount-G for observation under fluorescence microscopy.

Immunoblotting Analysis of BMP-2 Expression in R3230 Cells

For immunoblot analysis of cellular BMP-2 expression, R3230 cells were seeded onto six-well plates until 60 to 80% confluent and infected as described above. Within 24 hours of infection, GFP signal could be seen in ≈80% of cells. Cells were then washed once with PBS, refed with 750 µL of PBS per well, and incubated at 37°C for an additional 24 hours (ie, 48 hours postinfection); 750 µL of cell culture supernatant was collected, desalted, and concentrated fivefold in PBS with a 10 kDa MWCO Vivaspin filter (VivaScience Sartorius Group, Hannover, Germany). For measurement of BMP-2 levels, the equivalent of 75 µL of supernatant was mixed with 4× sample loading buffer (PAGEgel Inc, San Diego, CA), heated at 90°C for 10 minutes, and separated by 16% Tris-Tricine SDS-PAGE. Comparison was made to recombinant human BMP-2 produced in bacteria. After electrophoretic transfer (100 V for 1 hour), polyvinylidene fluoride (PVDF) membranes were dried at 37°C for 1 hour, washed in methanol for 2 seconds, and blocked with 2% nonfat dry milk in PBS-T overnight at 4°C. After washing the membrane with PBS-T, 2 µg/mL of mouse monoclonal antihuman BMP-2 antibody (R&D Systems, Minneapolis, MN) in PBS-T containing 1% nonfat dry milk was added and incubated for 1 hour at room temperature. Bound antibodies were detected using horseradish peroxidase (HRP)-conjugated donkey antimouse IgG secondary antibody (Jackson ImmunoResearch Laboratories Inc., West Grove, PA) and detected using the ECL detection system (Pierce, Rockford, IL) according to the manufacturer's instructions. Band density was quantitated by densitometric analysis. Known amounts of pure recombinant human BMP-2 (R&D Systems) were used in parallel lanes to create a calibration curve for quantitation of supernatant protein concentration.

Micro-CT

Micro-CT scans of the entire animal (snout to tail), at 90 µm resolution, and micro-CT scans of breast tumors at 45 µm resolution were performed weekly for 5 weeks using an eXplore Locus micro-CT scanner (GE Healthcare, Milwaukee, WI). General imaging system parameters included 80 kV, 450 mA, bin mode 2 × 2, and field of view 8.0 × 4.5 cm. For 90 µm resolution scans, pixel size was set to 95 µm, the angle of increment was 0.9 mm, scan time was 20 minutes, and reconstruction time was 10 minutes. For 45 µm resolution scans, pixel size was set to 48 µm, the angle of increment was 0.5 mm, scan time was 45 minutes, and reconstruction time was 45 minutes. Visualization and quantification were performed using MicroView (GE Healthcare) software.

Near-Infrared Fluorescence Imaging Using Pam800

The HA-specific, targeted near-infrared (NIR) fluorophore Pam800 has been described in detail previously.²⁹ Four rats (two rats with AdGFP/BMP-2 R3230 tumor, two rats with control R3230 tumor) were administered Pam800, diluted into 100 µL of PBS, pH 7.4, intravenously at a dose of 20 nmol per rat. Five hours postinjection, resected tumors were imaged with our previously described NIR fluorescence imaging system.³⁰

Visible Wavelength Fluorescence Imaging for GFP

All visible fluorescence data were collected using a Maestro (CRI, Woburn, MA) in vivo imaging system. Hyperspectral imaging was performed daily for the first 13 days post–tumor inoculation, weekly for 5 weeks, after animal sacrifice, and after tumor resection. The hair of rats was removed either by shaving or using a depilatory cream (Nair). Multispectral imaging data sets (cubes) were acquired with images typically spaced every 10 nm throughout the desired spectral range. For GFP fluorescence, animals were imaged for fluorescence emission from 500 to 720 nm using a blue excitation filter (445–490 nm) and a 515 nm long-pass secondary emission filter.

Quantitation of Serum BMP-2 Levels

Serum levels of human BMP-2 were measured on days 0, 7, 14, 21, 28, and 35 using an enzyme-linked immunosorbent assay (ELISA) specific for human BMP-2 (Antigenix America, Huntington Station, NY), which does not cross-react with rat BMP-2. Briefly, the plates were coated with anti-BMP-2 antibody (1 µg/mL, 100 µL/well) in carbonate buffer, pH 9.6, and blocked with ELISA coating stabilizer (200 µL/well) (Antigenix America). The collected sera with serial dilutions were added to the coated plates (100 µL/well) in triplicate and incubated at room temperature for 2 hours. After four washes of PBS-T, biotinylated anti-BMP-2 antibody (1:1,000 dilution [1 µg/mL], 100 µL/well) was added and incubated for 2 hours. After four washes of PBS-T, streptavidin-HRP (1:1,000 dilution, 100 µL/well) was added and incubated for 30 minutes. After five washes with PBS-T, 3,3′,5,5’ tetramethylbenzidine (TMB) substrate (100 µL/well) was added and incubated at room temperature for 10 minutes. Reactions were stopped using 2N sulfuric acid (100 µL/well), and absorbance at 450 nm was measured with a VersaMax Tunable Microplate Reader (Molecular Devices, Sunnyvale, CA).

Reverse Transcription–PCR of Excised Breast Tumors

Bilateral tumors were excised from anesthetized rats, flash-frozen in LN₂, and stored at −80°C. Total ribonucleic acid (RNA) was extracted from tumors using Trizol reagent (GIBCO/BRL, Grand Island, NY) according to the manufacturer's instructions. Sample RNA (5 µg) was treated with amplification grade DNase I (Invitrogen, Carlsbad, CA) as described by the manufacturer's protocol before reverse transcription (RT). First-strand complementary DNA was synthesized from RNA by RT using a Qiagen OneStep RT-PCT kit (Qiagen GmbH, Hilden, Germany) and the primers listed below. In a parallel control reaction, RT was performed without specific primers and processed for PCR to exclude DNA contamination. The PCR mixture, in a total volume of 25 µL, contained 50 µM dNTPs, 0.5 U Taq DNA polymerase, 1.5 mM MgCL₂, 1× PCR buffer, and the primer pair at 1 µM. The primers used for amplification are as follows, with product sizes in parentheses:

Human BMP-2 (377 bp)

Sense: 5′-TCATCCCAGCCCTCTGACGA-3′

Antisense: 5′-GAAACTGCTATTGTTTCCTA-3′

Enhanced Green Fluorescent Protein (EGFP) (312 bp)

Sense: 5′-TTCAAGGACGACGGCAACTA-3′

Antisense: 5′-GTGCTCAGGTAGTGGTTGT-3′

AdBMP-2/SV40pA (280 bp)

Sense: 5′-GCTATTGCTTTATTTGTAACCATT-3′

Antisense: 5′-AAGAACTATCAGGACATGGT-3′

Rat BMP-2 (374 bp)

Sense: 5′-TTATCCCGGCCTTCGGAGGAC-3′

Antisense: 5′-GAAACTACTATTTCCCAAAGCTTC-3′

Rat GAPDH gene (270 bp)

Sense: 5′-GTGAACGGATTTGGCCGTAT-3′

Antisense: 5′-ACTCCACGACATACTCAGCA-3′

For PCR, the reaction mixture was initially denatured at 95°C for 2 minutes and then subjected to 28 cycles of 94°C for 30 seconds, 54°C at 60 seconds, and 72°C for 60 seconds. After final extension at 72°C for 7 minutes, products were separated on a 2.5% agarose gel along with base-pair markers, visualized with ethidium bromide, and quantified for band intensity. The GAPDH gene was used as an appropriate reference in RT-PCR studies of R3230 tumor cells with or without viral infection.

Statistical Analyses

Two-tailed Student t-tests were used for significance testing. Means are presented along with standard errors.

Results

Adaptation of R3230 Rat Breast Tumors to Cell Culture

R3230 breast adenocarcinomas are typically passaged serially in syngeneic Fischer 344 female rats, which precludes genetic manipulation. As described in Materials and Methods, we adapted R3230 cells to tissue culture while maintaining their high (43 of 46 inoculations; > 93%) tumor induction rate after inoculation subcutaneously into the mammary fat pad.

Adenoviral Coexpression of BMP-2 and GFP

R3230 cells were infected at a MOI of 500:1, with adenoviruses expressing enhanced GFP alone (controls) or coexpressing GFP and human BMP-2. To confirm that BMP-2 was being secreted extracellularly as a mature protein, immunofluorescence and immunoblotting were performed. As shown in Figure 1A, BMP-2 deposited extracellularly within 48 hours after adenoviral transduction, even when cells were coexpressing high levels of GFP. Immunoblotting (Figure 1B) confirmed that BMP-2 was indeed secreted, showing the expected molecular weight of 18 kDa. Quantitative immunoblotting using recombinant BMP-2 as a calibration standard revealed levels of 0.2 µg/mL in the cell supernatant secreted between 24 and 48 hours after transduction.

Figure 1.

Adenoviral coexpression of bone morphogenetic protein 2 (BMP-2) and green fluorescent protein (GFP). A, Immunofluorescence detection of BMP-2 and GFP in adenovirus-transduced R3230 cells. R3230 cells infected with AdGFP/BMP-2 virus (top) and control AdGFP viruses (bottom) 48 hours previously were assessed for expression of BMP-2 (left column) and GFP (second column). A phase contrast (third column) and merge of BMP-2 and phase contrast images (right column) are also shown. B, Secretion of BMP-2 by adenovirus-transduced R3230 cells. The equivalent of 75 µL of cell culture supernatant collected between 24 and 48 hours from R3230 cells transduced with AdGFP/BMP-2 (left lane) or control GFP viruses (middle lane) and 50 ng of recombinant human BMP-2 (rhBMP-2; right lane) were resolved on 16% Tris-Tricine SDS-PAGE gel and subjected to immunoblot analysis using a BMP-2-specific antibody.

Breast Cancer Microcalcifications Induced by BMP-2 Expression

R3230 cells inoculated into the mammary pad of Fischer 344 rats formed large, histology-consistent breast tumors over the course of 5 weeks. Interestingly, cells infected with AdGFP/BMP-2 for 48 hours prior to inoculation formed tumors that produced microcalcifications in 100% (8 of 8 tumors) of cases (Figure 2A). In contrast, small-volume (< 1%) microcalcifications occurred in only a single control (AdGFP) tumor of eight inoculated (see Figure 2A), and this was in an area of apparent tumor necrosis. Importantly, full-body micro-CT failed to demonstrate microcalcifications in any normal tissue, even as microcalcifications in tumors approached high levels.

Figure 2.

Molecular imaging of breast cancer microcalcifications. A, Longitudinal micro–computed tomography (CT) of Fischer 344 rats with R3230 breast tumors. Representative micro-CT scans (maximal intensity projections of 90 µm resolution scan) of week 4 animals with a single tumor composed of R3230 cells transduced preinjection with either AdGFP/bone morphogenetic protein 2 (BMP-2) (left) or AdGFP (right). Insets show high-resolution (45 µm) maximal intensity projection of the tumors. White arrows indicate microcalcifications. B, Hydroxyapatite (HA) content of breast tumor microcalcifications. Five hours after intravenous injection of 20 nmol of HA-targeted near-infrared (NIR) fluorophore Pam800, AdGFP/BMP-2 tumors (left) or control AdGFP tumors (right) were excised and subjected to 45 µm micro-CT (top; maximal intensity projections) and planar reflectance NIR fluorescence (bottom). C, Tumor histology. Shown are representative sections from AdGFP/BMP-2 tumors (left) or control AdGFP tumors (right) stained using hematoxylin and eosin (H&E) (top) or the method of von Kossa (bottom). Magnification = 40X.

To determine whether the calcium salt HA was a component of the microcalcification seen by micro-CT, a cohort of tumor-bearing animals was injected intravenously with a NIR fluorophore specific for HA.^29,31–33 As shown in Figure 2B, intense NIR fluorescence signal (signal to background ratios greater than 5:1) in AdGFP/BMP-2 tumors, but not in AdGFP tumors, suggested that HA is, indeed, a major component of BMP-2-induced microcalcification.

As shown in Figure 2C, hematoxylin-eosin histology of AdGFP/BMP-2 tumors revealed an intense osteoblast-like reaction. Von Kossa staining confirmed the presence of abundant calcium salts in AdGFP/BMP-2 tumors but not control AdGFP tumors.

Kinetics of Tumor Growth, Microcalcification Formation, and Serum BMP-2 Levels

After a lag period of approximately 2 weeks, R3230 breast tumors exhibited exponential growth in vivo (Figure 3A). There was no statistical difference between tumors transduced prior to inoculation with AdGFP or AdGFP/BMP-2. Beginning at day 21 after R3230 cell inoculation, however, microcalcifications were easily detectable by micro-CT, accounting for up to 15% of tumor volume (see Figure 3A). Absolute levels of microcalcification stabilized by week 4. Percent tumor calcification actually decreased during week 5 as tumor volume increased exponentially, suggesting that the process of calcification is a complex function of cell proliferation and local environment. Intriguingly, serum levels of human BMP-2 in rats inoculated with AdGFP/BMP-2 peaked at week 1 and became undetectable by week 3 (Figure 3B).

Figure 3.

Kinetics of breast tumor growth, microcalcification formation, and serum bone morphogenetic protein 2 (BMP-2) levels. A, Rats with AdGFP-transduced breast tumors (white bars) or AdGFP/BMP-2-transduced breast tumors (black bars) were scanned weekly by micro–computed tomography, and breast tumor volume (top), microcalcification volume (middle), and percentage of microcalcification per tumor (bottom) were quantified. Shown are means ± SEM for eight rats per condition. B, BMP-2 serum enzyme-linked immunosorbent assay from rats bearing AdGFP/BMP-2 tumors. Shown are mean ± SEM (three rats per condition). Asterisks denote statistically significant differences from the control (p < .01) as measured using a two-tailed Student t-test.

BMP-2 as a Humoral Mediator of Breast Cancer Microcalcification

The above experiments confirmed that local secretion of BMP-2 was sufficient to produce microcalcification of breast cancers. To determine whether BMP-2 could also act as a humoral mediator of the process, bilateral tumors were grown in the same animal. The ipsilateral fat pad was inoculated with R3230 cells transduced with AdGFP/BMP-2 and the contralateral fat pad with mock-infected (untransduced) R3230 cells. The kinetics of tumor growth and microcalcification in the BMP-2-expressing tumor followed the pattern described for single tumors (data not shown). However, in 100% (4 of 4) of the animals with bilateral tumors, microcalcifications were also found in the contralateral, untransduced tumor, albeit with statistically significant lower percentages of calcification (Figure 4A).

Figure 4.

Microcalcification formation in untransduced, contralateral tumors. A, Micro–computed tomography (CT) of bilateral tumors. Rats inoculated with bilateral tumors were scanned using micro-CT at 4 weeks. B = ipsilateral tumor whose R3230 cells were transduced with AdGFP/bone morphogenetic protein 2 (BMP-2) prior to inoculation; U = contralateral tumor whose R3230 cells were untransduced prior to inoculation. The percentage of microcalcifications per tumor (mean ± SEM) is shown for four animals. B, Reverse transcription–polymerase chain reaction (RT-PCR) analysis of bilateral tumors. Tumors excised 4 weeks after inoculation were subjected to RT-PCR analysis on a 2.5% agarose gel. Lane 1 = human BMP-2 (377 bp); lane 2 = green fluorescent protein (GFP) (312 bp); lane 3 = partial of BMP-2 sequence and SV40 polyA sequence from adenovirus (280 bp); lane 4 = rat BMP-2 (374 bp) and rat glyceraldehyde-3-phosphate dehydrogenase (270 bp). C, Expression of GFP in ipsilateral and contralateral tumors. Cohorts of eight Fischer 344 rats with bilateral R3230 tumors as described in Figure A were sacrificed weekly. Tumors were bisected and imaged using white light (left) or GFP fluorescence (right). B = ipsilateral tumor whose R3230 cells were transduced with AdGFP/BMP-2 prior to inoculation. GFP fluorescence images have identical exposure times and normalizations; U = contralateral tumor whose R3230 cells were untransduced prior to inoculation.

To rule out infection of the contralateral tumor by adenovirus, RT-PCR was employed. No human BMP-2, GFP, or adenovirus transcripts could be amplified from untransduced, contralateral tumors, whereas ipsilateral BMP-2-expressing tumors had easily detectable levels of all three (Figure 4B). Importantly, neither tumor had detectable levels of the rat homologue of BMP-2. To ensure that viral replication was not a confounding variable, R3230 cells were infected in tissue culture with AdGFP/BMP-2 and AdGFP at a MOI of 500:1. Although all cells were highly fluorescent within 24 hours, supernatant harvested at 48 hours postinfection did not transfer GFP expression to any other cell line tested (data not shown). Moreover, in cohorts of rats bearing bilateral tumors (as described above) that were sacrificed weekly after R3230 cell inoculation, GFP fluorescence could be found only in the ipsilateral, BMP-2-expressing tumors and not in the contralateral untransduced tumor (Figure 4C). Unlike serum BMP-2 levels, which peaked by week 1, GFP levels were relatively constant, peaked at week 4, and then decreased. Taken together, these data suggest that BMP-2 is capable of acting as a humoral factor to produce microcalcifications in breast cancer.

Discussion

Although microcalcification is an important diagnostic sign for breast cancer on mammography, very little is known about the mechanism of calcium salt deposition. Our data suggest that, at least in a syngeneic model system, BMP-2 is sufficient for inducing high levels of microcalcification in breast adenocarcinoma, while sparing normal tissues. Although these results are consistent with the fact that BMP-2 messenger RNA is expressed in normal human mammary epithelial cells and normal mammary glands and in up to 85% of human breast cancers,³⁴ an unanswered question is whether BMP-2 is exerting its effect on microcalcification through the breast cancer cells themselves, rare osteoblast-like cells within the tumor, or the surrounding stroma. For example, the kinetics of BMP-2-induced microcalcification and the presence of HA as a major component are highly analogous to ectopic bone formation,²³ and it is possible that pericyte-derived, osteoblast-like cells contribute to ossification, as they do in a model of BMP-stimulated skull wound repair.³⁵

Even under normal circumstances, the relationship among BMP-2, the vasculature, and osteoprogenitor cells is complex. Biologically active BMP-2 is a disulfide-linked homodimer that binds to heterodimeric type 1 and type 2 bone morphogenetic protein receptors (BMPRs), which themselves have serine/threonine kinase activity.³⁶ On BMP-2 binding, BMPR2 phosphorylates and activates BMPR1, which, in turn, phosphorylates the regulatory Smads and modulates target gene expression.^37,38 BMPR activation promotes physiologic vascularization and appears to also be involved in tumor angiogenesis by stimulating the Id1 and p38 MAPK pathway.^39,40 There is also feedback built into the system since it was recently demonstrated that phosphatidylinositol 3-kinase contributes to lovastatin-induced MAPK activation, which, together with Akt kinase, regulates BMP-2 expression.⁴¹ Using our animal model, it should now be possible to identify the signal transduction pathways that mediate breast cancer microcalcification under our particular experimental conditions, thus laying a foundation for future studies aimed at understanding human breast cancer calcification.

Although multiple studies in bone healing and remodeling have used adenovirus-transduced BMP-2,^22,42 ours is the first to use it for breast cancer cells. Previously, BMP-2 has been expressed stably in human MCF-7 breast cancer cells using a plasmid and cell selection,^24,40 and the BMPR2 has been expressed both transiently and stably using plasmids.⁴³ The former studies provided important insight into the role of BMP-2 in tumorigenicity and angiogenesis and suggested that BMP-2 was also playing a role in calcification. Our study used adenovirus-based transduction of BMP-2 so that protein levels, and the timing of its expression, could be controlled precisely and without the need for long-term aminoglycoside selection.

The kinetics of tumor growth, BMP-2 expression, and microcalcification (see Figure 3) are extremely interesting. Tumor growth is exponential over the 5 weeks studied, with no statistical difference in tumor volume or visual difference in in vivo pattern of growth over this time. This latter result is in contrast to previous, contradictory studies in which BMP-2 expression has improved the tumorigenicity²⁴ and angiogenesis⁴⁰ of xenograft tumors but inhibited the proliferation of breast cancer cells in culture.⁴⁴ It is likely that in tumors, such as our syngeneic model, in which the tumor take rate is close to 100%, the overall contribution of BMP-2 is negligible. With respect to calcification, however, BMP-2's effects are profound. At the time of tumor cell inoculation, BMP-2 levels are at their highest and decline rapidly to baseline by week 3. Microcalcification, however, is only measurable starting in week 3, increases rapidly, and plateaus by week 4. Our data prove that even transient exposure to BMP-2 is sufficient to trigger molecular changes in the calcification process that persist long after the exposure. Consistent with our kinetic data is the possibility that this event occurs during angiogenesis, at which time, pericytes are active. Our data also suggest that calcification in human breast cancer, which occurs in only 30 to 50% of women, might be an epiphenomenon of a similar transient event.

Independent of molecular mechanism, our study provides the first straightforward, reproducible, syngeneic, and short-latency rodent model of breast cancer microcalcification. Although by week 3, the percentage of tumor volume harboring calcifications is much higher (≈15%) than that seen in human breast cancer, earlier time points provide clinically realistic levels of ≈1% (week 2) and any other desired level selectable based on time. Our animal model is of paramount importance for the development of novel contrast agents that could provide improved sensitivity and specificity for breast cancer screening. For example, our group has previously shown that the bisphosphonate pamidronate can be used as a modular ligand to create targeted contrast agents specific for HA.^29,32,33 Using this new animal model, one could develop microcalcification-targeted contrast agents for magnetic resonance imaging and positron emission tomography, two modalities that may play an increasing role in breast cancer screening over the coming years. One could also use it to optimize NIR fluorescent bisphosphonates for tomographic optical imaging of breast cancer, with micro-CT serving as the ground truth.

Footnotes

Acknowledgments

We thank Ralph Weissleder, MD, PhD (Massachusetts General Hospital), for the generous gift of the initial R3230 tumors; Barbara L. Clough for editing; and Eugenia Trabucchi and Alice Gugelmann for administrative assistance.

References

American Cancer Society . Cancer facts and figures. Available at: http://www.cancer.org/downloads/STT/2008CAFFfinalsecured.pdf. (accesssed March 27, 2008)

van Ongeval

Bosmans

van Steen

. Current status of digital mammography for screening and diagnosis of breast cancer. Curr Opin Oncol 2006;18:547–54.

Elmore

Armstrong

Lehman

Fletcher

. Screening for breast cancer. JAMA 2005;293:1245–56.

Morgan

Cooke

McCarthy

. Microcalcifications associated with breast cancer: an epiphenomenon or biologically significant feature of selected tumors? J Mammary Gland Biol Neoplasia 2005;10:181–7.

Stomper

Geradts

Edge

Levine

. Mammographic predictors of the presence and size of invasive carcinomas associated with malignant microcalcification lesions without a mass. AJR Am J Roentgenol 2003;181:1679–84.

Haka

Shafer-Peltier

Fitzmaurice

. Identifying microcalcifications in benign and malignant breast lesions by probing differences in their chemical composition using Raman spectroscopy. Cancer Res 2002;62:5375–80.

Morgan

Cooke

Christopherson

. Calcium hydroxyapatite promotes mitogenesis and matrix metalloproteinase expression in human breast cancer cell lines. Mol Carcinog 2001;32:111–7.

Cooke

McCarthy

Sallis

Morgan

. Phosphocitrate inhibits calcium hydroxyapatite induced mitogenesis and upregulation of matrix metalloproteinase-1, interleukin-1beta and cyclooxygenase-2 mRNA in human breast cancer cell lines. Breast Cancer Res Treat 2003;79:253–63.

Kim

O'Hare

Stein

. Models of breast cancer: is merging human and animal models the future? Breast Cancer Res 2004;6:22–30.

10.

Pasini

Casalis Cavalchini

Genovese

Manoukian

. Evaluating breast cancer risk: available models to assess individual breast cancer risk and probability to be a BRCA mutation carrier. J Exp Clin Cancer Res 2002;21:23–9.

11.

Tsukamoto

Grosschedl

Guzman

. Expression of the int-1 gene in transgenic mice is associated with mammary gland hyperplasia and adenocarcinomas in male and female mice. Cell 1988;55:619–25.

12.

Tyrer

Duffy

Cuzick

. A breast cancer prediction model incorporating familial and personal risk factors. Stat Med 2004;23:1111–30.

13.

Urist

Sato

Brownell

. Human bone morphogenetic protein (hBMP). Proc Soc Exp Biol Med 1983;173:194–9.

14.

Wang

Rosen

Cordes

. Purification and characterization of other distinct bone-inducing factors. Proc Natl Acad Sci U S A 1988;85:9484–8.

15.

Vainio

Karavanova

Jowett

Thesleff

. Identification of BMP-4 as a signal mediating secondary induction between epithelial and mesenchymal tissues during early tooth development. Cell 1993;75:45–58.

16.

Wozney

Rosen

Celeste

. Novel regulators of bone formation: molecular clones and activities. Science 1988;242:1528–34.

17.

Yamaguchi

Katagiri

Ikeda

. Recombinant human bone morphogenetic protein-2 stimulates osteoblastic maturation and inhibits myogenic differentiation in vitro. J Cell Biol 1991;113:681–7.

18.

Sasaki

. Osteogenic potential of five different recombinant human bone morphogenetic protein adenoviral vectors in the rat. Gene Ther 2003;10:1735–43.

19.

Kang

Sun

Cheng

. Characterization of the distinct orthotopic bone-forming activity of 14 BMPs using recombinant adenovirus-mediated gene delivery. Gene Ther 2004;11:1312–20.

20.

Feeley

Conduah

Sugiyama

. In vivo molecular imaging of adenoviral versus lentiviral gene therapy in two bone formation models. J Orthop Res 2006;24:1709–21.

21.

Helm

Sheehan

. Utilization of type I collagen gel, demineralized bone matrix, and bone morphogenetic protein-2 to enhance autologous bone lumbar spinal fusion. J Neurosurg 1997;86:93–100.

22.

Olmsted-Davis

Gugala

Gannon

. Use of a chimeric adenovirus vector enhances BMP2 production and bone formation. Hum Gene Ther 2002;13:1337–47.

23.

Varady

Alden

. CT and radionuclide study of BMP-2 gene therapy-induced bone formation. Acad Radiol 2002;9:632–7.

24.

Clement

Raida

Sanger

. Bone morphogenetic protein 2 (BMP-2) induces in vitro invasion and in vivo hormone independent growth of breast carcinoma cells. Int J Oncol 2005;27:401–7.

25.

Goldberg

Girnan

Lukyanov

. Percutaneous tumor ablation: increased necrosis with combined radio-frequency ablation and intravenous liposomal doxorubicin in a rat breast tumor model. Radiology 2002;222:797–804.

26.

Zhou

da Costa

. A simplified system for generating recombinant adenoviruses. Proc Natl Acad Sci U S A 1998;95:2509–14.

27.

Anderson

Haskell

Xia

. A simple method for the rapid generation of recombinant adenovirus vectors. Gene Ther 2000;7:1034–8.

28.

Thompson

Hunt

. Selected histochemical and histopathological methods. Springfield (IL): Charles C. Thomas; 1996.

29.

Bhushan

Tanaka

Frangioni

. Synthesis of conjugatable bisphosphonates for molecular imaging of large animals. Angew Chem Int Ed Engl 2007;46:7969–71.

30.

Nakayama

del Monte

Hajjar

Frangioni

. Functional near-infrared fluorescence imaging for cardiac surgery and targeted gene therapy. Mol Imaging 2002;1:365–77.

31.

Lenkinski

Ahmed

Zaheer

. Near-infrared fluorescence imaging of microcalcification in an animal model of breast cancer. Acad Radiol 2003;10:1159–64.

32.

Zaheer

Lenkinski

Mahmood

. In vivo near-infrared fluorescence imaging of osteoblastic activity. Nat Biotechnol 2001;19:1148–54.

33.

Zaheer

Murshed

De Grand

. Optical imaging of hydroxyapatite in the calcified vasculature of transgenic animals. Arterioscler Thromb Vasc Biol 2006;26:1132–6.

34.

Alarmo

Kuukasjarvi

Karhu

Kallioniemi

. A comprehensive expression survey of bone morphogenetic proteins in breast cancer highlights the importance of BMP4 and BMP7. Breast Cancer Res Treat 2006;103:239–46.

35.

Sato

Urist

. Induced regeneration of calvaria by bone morphogenetic protein (BMP) in dogs. Clin Orthop Relat Res 1985;197:301–11.

36.

Hoodless

Haerry

Abdollah

. MADR1, a MAD-related protein that functions in BMP2 signaling pathways. Cell 1996;85:489–500.

37.

Liu

Yue

Frey

. Transforming growth factor beta signaling through Smad1 in human breast cancer cells. Cancer Res 1998;58:4752–7.

38.

Wang

Wei

Zhu

. BMP-2 activates Smad6 gene transcription through bone-specific transcription factor Runx2. J Biol Chem 2007;282:10742–8.

39.

Kimura

Matsuo

Shibuya

. BMP2-induced apoptosis is mediated by activation of the TAK1-p38 kinase pathway that is negatively regulated by Smad6. J Biol Chem 2000;275:17647–52.

40.

Raida

Clement

Leek

. Bone morphogenetic protein 2 (BMP-2) and induction of tumor angiogenesis. J Cancer Res Clin Oncol 2005; 131:741–50.

41.

Ghosh-Choudhury

Mandal

Ghosh Choudhury

. Statin-induced Ras activation integrates PI 3 kinase signal to AKT and MAPK for BMP-2 expression in osteoblast differentiation. J Biol Chem 2006;282:4983–93.

42.

Ishihara

Shields

Litsky

. Osteogenic gene regulation and relative acceleration of healing by adenoviral-mediated transfer of human BMP-2 or -6 in equine osteotomy and ostectomy models. J Orthop Res 2008;26:764–71.

43.

Pouliot

Blais

Labrie

. Overexpression of a dominant negative type II bone morphogenetic protein receptor inhibits the growth of human breast cancer cells. Cancer Res 2003;63:277–81.

44.

Arnold

Tims

McGrath

. Identification of bone morphogenetic proteins and their receptors in human breast cancer cell lines: importance of BMP2. Cytokine 1999;11:1031–7.