• Open Access

Inhibitory effect of soluble platelet-derived growth factor receptor β on intraosseous growth of breast cancer cells in nude mice


To whom correspondence should be addressed.
E-mail: uehara@basic.med.tokushima-u.ac.jp


Bone metastasis is a frequent complication of advanced breast cancer. On the basis of functional and molecular evidence, signaling mediated by the binding of platelet-derived growth factor (PDGF)-BB and -DD to PDGF receptor β (PDGFRβ) is critical for the survival and growth of metastatic breast cancer cells within the bone microenvironment. In this study, we propose a new approach to blocking PDGFRβ signaling using soluble PDGFRβ (sPDGFRβ) as a decoy receptor for PDGF-BB and -DD secreted from tumor cells and bone marrow stromal cells. A bone-seeking TNBCT/Bo cell line was established by in vivo selection from TNBCT human breast cancer cells, which are negative for estrogen receptor, progesterone receptor, and human epidermal growth factor receptor 2 protein expression. The TNBCT/Bo cells were transfected with a mammalian expression vector encoding the extracellular domain of PDGFRβ. A stable transfectant (TNBCT/Bo-sPDGFRβ) grew at a similar rate to that of control cells under normal culture conditions, although growth stimulation of human fibroblasts with PDGF-BB was neutralized by the culture medium from TNBCT/Bo-sPDGFRβ cells. Intratibial injection of TNBCT/Bo-sPDGFRβ cells into athymic nude mice resulted in a significant decrease in tumor incidence compared with control mice (< 0.01). This attenuated growth correlated with decreased cancer cell proliferation, angiogenesis, and recruitment of stromal cells, and with an increase in the number of apoptotic cells. These findings suggest that sPDGFRβ is useful for the treatment of breast cancer bone metastasis. (Cancer Sci 2011; 102: 1904–1910)

Breast cancer is the leading cause of cancer death in women,(1) and the skeleton is the preferred site for breast cancer metastasis.(2,3) Breast cancer bone metastases are usually incurable, and only 20% of patients with breast cancer are still alive 5 years after the discovery of bone metastases.(4) Of the several subtypes of breast cancer, triple-negative breast cancers do not express estrogen receptors (ER), progesterone receptors (PR), or human epidermal growth factor receptor 2 (HER2),(5) and are associated with a poor prognosis as they do not respond to endocrine therapy or HER2-targeted therapy.(6) Thirteen percent of triple-negative breast cancer patients have bone metastases and the median survival is only 0.8 years.(6)

Metastatic cancer cells in the bone microenvironment release factors and cytokines that accelerate bone destruction. In turn, other factors liberated from the bone matrix promote cancer cell growth and proliferation.(7) Of these cytokines and growth factors, the platelet-derived growth factor (PDGF) family is thought to contribute to the regulation of bone turnover.(8) Members of the PDGF family are well-known inducers of cell migration, proliferation, and transformation through activation of their cognate receptors.(9–11) Interactions between PDGF and PDGF receptor (PDGFR) are multiple and complex, however, PDGF-AA and PDGF-CC interact with PDGFRα,(12–14) and PDGF-BB and PDGF-DD interact with PDGFRβ.(15,16) Both PDGF-BB and PDGF-DD are abundantly expressed in many human cancer cells and tissues, including prostate,(17–19) kidney,(20–22) lung,(23) and breast.(24) In addition, a large number of breast tumors display a positive PDGFR phenotype. It is reported that invasive mammary carcinomas express PDGFRα in 65% and PDGFRβ in 75% of cases,(25) and PDGFRβ expression is significantly higher in tumor-associated stromal cells.(26–28) Both PDGF-BB and -DD enhance tumor angiogenesis by stimulating vascular endothelial growth factor expression.(29,30) In addition to its angiogenic effects, PDGF also recruits cancer-associated stromal cells (CAS). Increasing evidence suggests that cancer cells recruit stromal cells and endothelial cells into the tumor, and tumor–stromal cell interactions in the microenvironment are essential for directing tumor progression.(31,32) Human bone matrix is a reservoir of growth factors such as transforming growth factor β (TGF-β), PDGF,(33) and basic fibroblast growth factor.(34) During the process of bone remodeling, TGF-β and PDGF are released from the bone matrix and TGF-β signaling through Smad2/3/4 promotes PDGF-BB secretion by tumor cells.(35) Furthermore, several studies show that PDGFRβ is activated by both autocrine and paracrine mechanisms.(9,26) These data support a critical role of PDGFRβ signaling in cancer bone metastasis.

Several strategies to block PDGF/PDGFR-mediated tumor growth are currently being tested in basic research and clinical trials. These strategies fall into two categories: (i) neutralizing antibodies directed against PDGF ligands or receptors;(9) and (ii) inhibitors of PDGFR kinase activity (e.g. imatinib), which inhibit phosphorylation of PDGFR and increase apoptosis of tumor cells.(36) However, neutralizing antibodies are not completely specific,(9) and imatinib has potential disadvantages in that it is active against multiple receptor tyrosine kinases(9) and no benefit was reported in metastatic breast cancer patients.(37) Therefore, more specific and effective strategies must be developed.

Here, we propose a novel approach for neutralizing PDGFRβ signaling by using soluble PDGFRβ (sPDGFRβ), and show evidence of inhibited growth, angiogenesis, and bone lysis of triple-negative breast cancer cells after blockade of PDGFRβ signaling.

Materials and Methods

Cells and animals.  The triple-negative breast cancer Tokushima (TNBCT) cell line, established from lymph node metastases from triple-negative breast cancer patients, was kindly provided by Dr. H. Yamai and Dr. A. Tangoku (Department of Oncological and Regenerative Surgery, University of Tokushima Graduate School, Tokushima, Japan). This cell line was maintained in RPMI-1640 supplemented with 10% FBS, 100 U/mL penicillin G, and 0.1 mg/mL streptomycin sulfate. The in vitro and in vivo characteristics of the cell line are shown in the supporting information (Fig. S1). Primary human fibroblasts were kindly provided by Dr. Y. Nishioka (Department of Respiratory Medicine and Rheumatology, University of Tokushima Graduate School). Culture conditions were the same as for the TNBCT cell line. Five-week old female athymic nude mice were purchased from Charles River Japan (Yokohama, Japan). Mice were housed and maintained under specific pathogen-free conditions. Experimental procedures were carried out according to the Guidelines for the Care and Use of Laboratory Animals of the University of Tokushima School of Medicine and were approved by the Animal Care and Use Committee.

In vivo selection.  To obtain a subclone that prefers to grow intraosseously, TNBCT cells were injected into the tibias of female nude mice as described previously.(36) Nine weeks after injection, tumor lesions were separated, minced with a surgical scalpel, and filtered with a cell strainer (φ = 40 μm). The resulting cell suspension was reseeded and cultured for 1 week. Culture was continued until passage five. This selected subclone was designed TNBCT/Bo.

Construction of mammalian expression vector.  The extracellular region of PDGFRβ contains five Ig-like domains (Ig1–Ig5). As the Ig2–Ig3 region is essential for ligand binding,(38) we constructed a mammalian expression vector expressing a human PDGFRβ fragment covering the N-terminus to the Ig3 region (hPDGFRβIg1–3). Human PDGFRβ Ig1–3 cDNA was amplified by PCR using a plasmid containing the complete hPDGFRβ cDNA (kindly provided by Dr. Y. Takuwa, Department of Physiology, Kanazawa University Graduate School of Medicine, Kanazawa, Japan) as a template. The primers used were: forward, 5′-CCATGGTTATGCGGCTTCCGGGTGCG-3′; and reverse, 5′-GGATCCCCGCACGTAGCCGCTCTC-3′. The PCR product was digested with NcoI and BamHI and subcloned into the pFUSE-hFc2 (IL2ss) vector (InvivoGen, San Diego, CA, USA). The resulting product was expressed as a human IgG Fc-fusion protein at the C-terminus. The construct generated using the above procedure was designated pFUSE-sPDGFRβIg1–3. Successful construction was confirmed by direct sequencing using an ABI PRISM 3100-Avant Genetic Analyzer (Applied Biosystems, Carlsbad, CA, USA).

Stable transfection and detection of mRNA and protein expression levels.  Cells were stably transfected using the TransFast transfection reagent (Promega, Madison, WI, USA) as previously described.(39) The transfectants were designated TNBCT/Bo-mock and TNBCT/Bo-sPDGFRβ. After selection by Zeocin (InvivoGen), the transfectants and their culture media (CM) were examined for expression of sPDGFRβIg1–3 mRNA and secreted protein. Total RNA from the transfectants was isolated using an RNeasy mini kit (Qiagen, Valencia, CA, USA). Aliquots of these RNAs (1 μg/sample) were subjected to RT-PCR as described previously.(39)β-actin was amplified as an internal standard. The primers used are listed in Table 1. Culture media were concentrated using Amicon Ultra Centrifugal Filter Units (10K; Millipore, Billerica, MA, USA), then sPDGFRβIg1–3 protein secreted into the CM was detected using a hIgG ELISA quantification kit (Bethyl Laboratories, Montgomery, TX, USA).

Table 1.   Primer sequences for RT-PCR analysis
TargetForward (5′→3′)Reverse (5′→3′)
  1. †This sequence was derived from pFUSE-hFc2 (IL2ss) to distinguish it from endogenous platelet-derived growth factor receptor β (PDGFRβ).


Proliferation assays.  The TNBCT/Bo-mock and TNBCT/Bo-sPDGFRβ cells (1 × 105) were plated onto 6-well plates and preincubated at 37°C overnight. Next, the CM was replaced with 2 mL fresh complete medium and the number of viable cells was counted after incubation for 48 h using the Trypan blue exclusion method.

The effect on cell growth of the PDGF-BB in the CM from the transfectants was then assessed using primary human fibroblasts. The TNBCT/Bo-mock and TNBCT/Bo-sPDGFRβ cells (1 × 106) were plated on 100 mm culture dish and incubated for 4 days at 37°C in 10 mL complete culture medium. Cell debris was removed from the CM by centrifugation before concentration using Amicon Ultra Centrifugal Filter Units. Primary human fibroblasts (2 × 103) were plated on 96-well microplates and incubated at 37°C overnight. Next, the CM was replaced with 100 μL RPMI-1640/1% FBS containing CM from the TNBCT/Bo-mock or TNBCT/Bo-sPDGFRβ cells (adjusted for 10 ng/mL each fusion protein) and incubated either with or without recombinant PDGF-BB (10 or 30 ng/mL; R&D Systems, Minneapolis, MN, USA) for 24 h at 37°C. After incubation, viable cells were determined by the MTT method.

Nude mouse bone metastasis model.  After 1 week of acclimatization, TNBCT/Bo-mock or TNBCT/Bo-sPDGFRβ cells (5 × 105 cells/mouse) were injected proximal to the tuberositas tibiae of nude mice, as described previously.(36) After 12 weeks, the hind limbs were analyzed by X-ray radiography. Subsequently, the mice were killed and the hind limbs collected for histological analysis.

Histological analysis.  After fixation in 10% phosphate-buffered formaldehyde at room temperature for 48 h, the hind limbs were decalcified in 10% EDTA (pH 7.4) for 12 days, embedded in paraffin, and sectioned at 4–6 μm. Tumors were confirmed in H&E-stained sections. Expression of molecular immunology borstel-1 (MIB-1), cleaved caspase-3, CD31, α-smooth muscle actin (SMA), and human IgG was also examined by immunohistochemistry. The following antibodies were used: rabbit polyclonal anti-MIB-1 antibody (1:100 dilution; Dako, Glostrup, Denmark); rabbit polyclonal anti-mouse CD31 antibody (1:50 dilution; Abcam, Cambridge, UK); polyclonal rabbit anti-α-SMA antibody (1:50 dilution; Dako); rabbit polyclonal anti-cleaved caspase-3 antibody (1:200 dilution; Cell Signaling Technology, Danvers, MA, USA); and goat polyclonal anti-human IgG-Fc antibody (1:200 dilution; Bethyl Laboratories). In brief, sections were deparaffinized and rehydrated. After endogenous peroxidase activity was blocked by 5% H2O2 in methanol for 15 min at room temperature, antigen retrieval was carried out by autoclaving the sections in 0.01 mol/L citrate buffer (pH 6.0) for 15 min. All primary antibodies were incubated with 5% goat serum for 1 h at room temperature. Horseradish peroxidase-conjugated secondary antibodies (ChemMate Envision kit; Dako) were then applied for 1 h at room temperature. The sections were visualized with DAB and counterstained with Mayer’s hematoxylin (Muto Pure Chemicals, Tokyo, Japan).

Quantitative morphometric analysis of cells positive for MIB-1, cleaved caspase-3, or α-SMA in the tumor samples was carried out as follows: five areas/slide were photographed under a DXM1200 camera (×200 magnification; Olympus, Tokyo, Japan) and the fraction of MIB-1-positive cells or caspase-3-positive cells was determined after analyzing at least 1000 cells in all tumors from each group.(40) Microvessel density (MVD) was calculated as follows: any brown-staining endothelial cells or endothelial cell cluster that was clearly separate from the adjacent microvessel, tumor cells, and other connective tissue elements was considered to be a single countable microvessel. The number of CAS was counted on immunostained sections using an anti-α-SMA antibody in five randomly selected tumor fields.

Statistical analysis.  Results were expressed either as the median (95% confidence interval) or the mean ± SD. Comparison of tumor incidences between TNBCT/Bo-sPDGFRβ-injected mice and control mice was carried out using Fisher’s exact test. For immunostaining, statistical analysis of the results was carried out using Student’s t-test. P-values were considered to be statistically significant when < 0.05 in experiments comparing two groups of animals.


Generation of a human breast carcinoma cell line stably expressing sPDGFRβ.  We first tried to establish a stable transfectant by selecting with Zeocin followed by expression checks. Analyses using RT-PCR showed that TNBCT/Bo-sPDGFRβ cells successfully expressed mRNAs encoding the sPDGFRβIg1–3 sequence (Fig. 1a). The expression of endogenous PDGFRα and PDGFRβ mRNA was also examined. Both cell lines constitutively expressed PDGFRα and PDGFRβ mRNA (Fig. 1a). The ELISA analysis showed that the concentration of the IgG-fusion protein was 38.4 ± 16.8 ng/mL in TNBCT/Bo-mock cells and 4.5 ± 0.7 ng/mL in TNBCT/Bo-sPDGFRβ cells (data not shown). The growth rate of the cells was also examined using the Trypan blue exclusion method. After 48 h incubation in complete medium, the number of TNBCT/Bo-mock cells increased from 1 × 105 to 1.95 × 105 and the number of TNBCT/Bo-sPDGFR-β cells increased from 1 × 105 to 2.15 × 105 (Fig. 1b). There was no significant difference between the cell growth rates.

Figure 1.

 Characteristics of TNBCT/Bo-mock and -sPDGFRβ cells. (a) Triple-negative human breast cancer TNBCT/Bo cells were stably transfected with a mammalian expression vector encoding the extracellular domain of platelet-derived growth factor receptor β (TNBCT/Bo-sPDGFRβ). The expression of sPDGFRβIg1–3, and endogenous PDGFRα and PDGFRβ mRNA was examined in TNBCT/Bo-mock and -sPDGFRβ cells by RT-PCR. (b) After 48 h incubation of TNBCT/Bo-mock and -sPDGFRβ cells (1 × 106) in complete medium, the cell numbers were counted by the Trypan blue exclusion method. Columns, mean numbers (n = 4) of two independent experiments; bars, SD.

Effect of TNBCT/Bo-sPDGFRβ CM on PDGF-BB-induced human fibroblast proliferation.  Proliferation assays using primary human fibroblasts were carried out to examine the biological activity of sPDGFRβIg1–3. The growth rate of the fibroblasts was accelerated by PGDF-BB, but not by PGDF-AA (data not shown). Treatment with PDGF-BB (10 and 30 ng/mL) significantly increased the rate of cell growth in the presence of TNBCT/Bo-mock CM (from 100.0 ± 3.4% in untreated cells to 115.7 ± 5.6% in 10 ng/mL-treated cells and 114.6 ± 4.4% in 30 ng/mL-treated cells; Fig. 2). In contrast, no increase was observed in the presence of TNBCT/Bo-sPDGFR-β CM (100.0 ± 16.1% in untreated cells, 103.0 ± 4.2% in 10 ng/mL-treated cells, and 90.7 ± 3.7% in 30 ng/mL-treated cells; Fig. 2).

Figure 2.

 Effect of culture media (CM) from triple-negative breast cancer TNBCT/Bo cells transfected with soluble platelet-derived growth factor receptor β (sPDGFRβ) on PDGF-BB-induced human fibroblast proliferation. Human primary fibroblasts were incubated for 24 h with or without recombinant PDGF-BB (10 or 30 ng/mL) in the presence of CM from TNBCT/Bo-mock or -sPDGFRβ. Cell growth was measured by MTT assay. The A550 value for untreated cells was assigned as 100% and the relative percentages of PGDF-BB-treated cells are shown. Columns, mean percentages (n = 4) of two independent experiments; bars, SD. *P < 0.005.

Decreased tumor incidence and osteolysis in TNBCT/Bo-sPDGFRβ-injected mice.  To assess the therapeutic benefit of sPDGFRβIg1–3, TNBCT/Bo-mock or TNBCT/Bo-sPDGFRβ were injected into the tibias of nude mice (n = 13 for TNBCT/Bo-mock and n = 15 for TNBCT/Bo-sPDGFRβ). X-ray radiographs taken at Week 12 showed significant lysis of tibial bone in the TNBCT/Bo-mock-injected mice compared with the TNBCT/Bo-sPDGFRβ-injected mice (Fig. 3a). Subsequently, all mice were killed and the hind limbs histologically examined to determine tumor incidence. Bone tumors were found in 10/13 (77%) TNBCT/Bo-mock-injected mice and in 1/15 (7%) TNBCT/Bo-sPDGFRβ-injected mice (P < 0.01; Table 2). Histologically, TNBCT/Bo-sPDGFRβ cells showed limited growth with minimal osteolysis. However, TNBCT/Bo-mock cells showed invasive growth with severe osteolysis (Fig. 3b). Successful transfection was confirmed by positive immunostaining of TNBCT/Bo-mock and -sPDGFRβ tumors for human IgG (Fig. 3c).

Figure 3.

 Effects of soluble platelet-derived growth factor receptor β (sPDGFRβ) on growth of cancer cells in the bone. (a) Representative X-ray findings of hind limbs in TNBCT/Bo-mock- and TNBCT/Bo-sPDGFRβ-injected mice. White arrow, bone lesion. (b) Representative histological findings of bone lesions in TNBCT/Bo-mock- and TNBCT/Bo-sPDGFRβ-injected mice. H&E staining. Scale = 100 μm. (c) Immunohistochemical analysis of human IgG expression in bone tumors for the confirmation of successful transfection. TNBCT/Bo was used as negative control. Scale = 50 μm.

Table 2.   Bone tumor incidence in mice injected with triple-negative human breast cancer TNBCT/Bo-mock cells and TNBCT/Bo cells transfected with soluble platelet-derived growth factor receptor β (sPDGFRβ)
Injected cellsTotal no. of miceNo. of mice with bone tumors (%)
  1. †Significantly different from TNBCT/Bo-mock at P < 0.01 by Fisher’s exact test.

TNBCT/Bo-mock1310 (77)
TNBCT/Bo-sPDGFRβ151 (7)†

Increased apoptosis and decreased cell proliferation, MVD, and CAS recruitment in TNBCT/Bo-sPDGFRβ tumor.  We next assessed the number of MIB-1-positive cells (proliferating cells) and caspase-3-positive cells (apoptotic cells) in the TNBCT/Bo-mock and -sPDGFRβ tumors. As shown in Figure 4, the proliferation rate was significantly lower in TNBCT/Bo-sPDGFRβ cells than in the TNBCT/Bo-mock cells (16.7 ± 9.4%vs 39.9 ± 4.8%; P < 0.01). In contrast, the rate of apoptosis in TNBCT/Bo-sPDGFRβ cells was significantly higher than that in TNBCT/Bo-mock cells (16.0 ± 5.8%vs 7.0 ± 3.9%; P < 0.05).

Figure 4.

 Immunohistochemical staining for cleaved caspase-3, molecular immunology borstel-1 (MIB-1), CD31, and α-smooth muscle actin (SMA). Mice were injected with triple-negative breast cancer TNBCT/Bo-mock cells, or TNBCT/Bo cells transfected with soluble platelet-derived growth factor receptor β (sPDGFRβ). Bone tumors were fixed in formaldehyde, decalcified with EDTA, embedded in paraffin, and sectioned. The sections were then subjected to immunohistochemical staining for cleaved MIB-1, caspase-3, CD31, and α-SMA. The percentage of MIB-1-positive and cleaved caspase-3-positive cells, and the numbers of CD31-positive microvessels and α-SMA-positive cells per field were calculated. Scale = 100 μm. Columns, mean percentages or numbers from five random fields per group; bars, SD. MVD, microvessel density. *P < 0.05; **P < 0.01; ***P < 0.001.

Next, microvessels were immunostained with an anti-CD31 antibody and MVDs assessed in TNBCT/Bo-mock and -sPDGFRβ tumors. Microvessel density was significantly lower in TNBCT/Bo-sPDGFRβ tumors (4.6 ± 1.9 vs 1.2 ± 0.8; P < 0.01, Fig. 4).

As PDGFRβ signaling is important for chemotaxis of CAS, we further examined the recruitment of CAS by immunostaining with anti-α-SMA antibody, a marker for CAS.(41) The numbers of CAS per field in the TNBCT/Bo-mock and -sPDGFRβ tumors was 18.0 ± 3.8 and 7.2 ± 2.8, respectively (P < 0.001). Although only one tumor from the TNBCT/Bo-sPDGFRβ-injected mice was not enough for histological analysis, the experiment was repeated with similar results.


In this study, the bone-seeking triple-negative breast cancer cell line, TNBCT/Bo, was transfected with a mammalian expression vector encoding the extracellular domain of PDGFRβ, and a stable transfectant, TNBCT/Bo-sPDGFRβ, was established. The CM from TNBCT/Bo-sPDGFRβ cells neutralized the growth of human fibroblasts stimulated with PDGF-BB, and the tumor incidence in TNBCT/Bo-sPDGFRβ-injected athymic nude mice was significantly lower than in control mice (< 0.01), as was cancer cell proliferation, angiogenesis, and the recruitment of stromal cells. However, the number of apoptotic cells increased.

Many characteristics of bone tissue make it an ideal environment for breast cancer cell migration and colonization.(42) When tumor cells reach the bone, stromal cells within the bone microenvironment (including fibroblasts, myofibroblasts, pericytes, and endothelial cells) contribute to the development of micrometastatic foci.(43,44) Because the reciprocal interaction between breast cancer cells and the bone microenvironment results in a vicious cycle that increases the tumor burden,(45) blocking this interaction is one potential therapeutic strategy for treating bone metastases. The present study focused on PDGFRβ signaling for the following reasons: (i) PDGFRβ is expressed in human mammary carcinomas and correlates with a malignant, invasive phenotype,(25) and PDGF-BB expression is also increased in advanced breast cancer patients;(16) (ii) paracrine PDGFR signaling is commonly observed in breast cancer, which triggers stromal cell recruitment, thereby affecting tumor growth, angiogenesis, and metastasis;(9) and (iii) the PDGFR tyrosine kinase inhibitor, imatinib, has been used to treat breast cancer bone metastases; however, the side-effects included myelosuppression.(46) Therefore, we attempted to specifically block PDGFRβ-signaling to treat breast cancer bone metastasis.

The triple-negative breast cancer cell line TNBCT was chosen because neither endocrine therapy nor HER2-targeted therapy is effective against triple-negative breast cancer;(6) therefore, other treatment strategies must be developed. The TNBCT cells showed epithelial-like growth in vitro, and formed bone tumors in nude mice after direct intratibial injection of cancer cells. Expression of ER, PR, and HER2 was not detected on the tumor cells by immunohistochemistry (Fig. S1). Because the incidence of TNBCT bone tumors was not very high, we tried to select bone-seeking cancer cells in vivo. The selected subclone was designated TNBCT/Bo, and showed a bone tumor incidence greater than that of TNBCT (data not shown).

Using TNBCT/Bo, we established a stable transfectant, TNBCT/Bo-sPDGFRβ. This cell line expressed mRNA encoding the sPDGFRβIg1–3 sequence (Fig. 1a). Secretion of IgG-fusion proteins by TNBCT/Bo-mock and TNBCT/Bo-sPDGFRβ cells was confirmed by ELISA, although the concentration of the protein from TNBCT/Bo-mock cells was approximately 8.5-fold higher than that from TNBCT/Bo-sPDGFRβ cells.

Next, a proliferation assay was carried out to examine the biological activity of sPDGFRβIg1–3. Primary human fibroblasts were used for this assay, in which growth was accelerated by PGDF-BB but not by PGDF-AA. Treatment with PDGF-BB significantly induced cell growth in the presence of CM from TNBCT/Bo-mock cells. In contrast, no growth induction was observed in the presence of CM from TNBCT/Bo-sPDGFRβ cells (Fig. 2). These results indicated that the PDGF-BB–PDGFRβ interaction was blocked by sPDGFRβIg1–3. The growth rates of TNBCT/Bo-mock and -sPDGFRβ cells were also examined. Despite constitutive expression of endogenous PDGFRα and PDGFRβ mRNA (Fig. 1a), both cells lines grew at similar rates under normal culture conditions (Fig. 1b). A PDGFRβ-mediated autocrine loop might not exist in these cells, as sPDGFRβIg1–3 did not show any inhibitory effect on the growth of TNBCT/Bo-sPDGFRβ cells.

The TNBCT/Bo-mock or TNBCT/Bo-sPDGFRβ cells were injected into the tibias of nude mice to assess the therapeutic benefit of sPDGFRβIg1–3. The histologically-determined tumor incidence at week 12 was 10/13 (77%) in TNBCT/Bo-mock-injected mice and 1/15 (7%) in TNBCT/Bo-sPDGFRβ-injected mice (P < 0.01; Table 2). Radiographic and histological analyses revealed that TNBCT/Bo-sPDGFRβ cells showed limited growth with minimal osteolysis. In contrast, TNBCT/Bo-mock cells showed invasive growth with severe osteolysis (Fig. 3a,b). Positive immunostaining of TNBCT/Bo-mock and -sPDGFRβ cells for human IgG (Fig. 3c) confirmed continued expression of the IgG-fusion proteins. Taken together, these results suggest that sPDGFRβ is useful for the treatment of triple-negative breast cancer bone metastases.

Immunohistochemical analysis also showed that sPDGFRβ inhibited the proliferation of tumor cells and induced apoptosis in vivo; however, in vitro data suggested that the proliferation of TNBCT/Bo cells was independent of autocrine PDGFRβ signaling. Taken together, these results suggest that PDGFRβ-mediated paracrine stimulation in the bone microenvironment facilitates the progression of TNBCT/Bo tumors. In agreement with a direct role for PDGFRβ signaling as a switch toward endothelial cell development and induction of endothelial cell differentiation,(47,48) MVD was significantly decreased in TNBCT/Bo-sPDGFRβ tumors compared with that in the controls. As PDGFRβ expression in tumor-associated endothelial cells was upregulated in the bone microenvironment in our previous study,(36) sPDGFRβ could be an effective inhibitor of tumor-associated angiogenesis, especially in bone metastases. In addition to its pro-angiogenic effects, PDGFRβ signaling is associated with the recruitment of stromal cells.(31,32) The tumor stroma has been viewed as a passive component of tumors in bone metastasis.(29) It is widely accepted that tumor–stromal cell cross-talk promotes a tumor-promoting phenotype in the recruited stromal cells. Paracrine stimulation of tumor pericytes is associated with a decrease in tumor cell apoptosis.(40) Emerging data suggest, however, the CAS may affect the malignant behavior of neoplastic cells and contribute to tumor progression.(49) Expression of PDGFRβ is significantly higher in tumor-associated stromal cells.(26,27) Based on these data, we analyzed stromal cell recruitment by TNBCT/Bo-mock and -sPDGFRβ tumors. The number of CAS per field was significantly less in the TNBCT/Bo-sPDGFRβ tumor, suggesting another mechanism underlying the antitumor effects of sPDGFRβ: inhibition of tumor-associated stromal cell recruitment.

In conclusion, we have shown that sPDGFRβ inhibits the growth of triple-negative breast cancer cells in the bone microenvironment through the inhibition of tumor cell proliferation, induction of tumor cell apoptosis, and reduced CAS recruitment and neovascularization. These findings suggest that sPDGFRβ could be a useful agent for blocking the vicious PDGFRβ-mediated cycle of bone metastasis.


We thank Megumi Kume and Hitomi Umemoto for technical assistance.

Disclosure Statement

The authors have no conflicts of interest.