Cell Oncol. (2014) 37:363–375 DOI 10.1007/s13402-014-0197-1

ORIGINAL PAPER

BMP9 regulates cross-talk between breast cancer cells and bone marrow-derived mesenchymal stem cells Shaoheng Wan & Yuehong Liu & Yaguang Weng & Wei Wang & Wei Ren & Chang Fei & Yingying Chen & Zhihui Zhang & Ting Wang & Jinshu Wang & Yayun Jiang & Lan Zhou & Tongchuan He & Yan Zhang

Accepted: 1 September 2014 / Published online: 11 September 2014 # International Society for Cellular Oncology 2014

Abstract Purpose Breast cancer cells frequently metastasize to distant organs, including bone. Interactions between breast cancer cells and the bone microenvironment are known to enhance tumor growth and osteolytic damage. Here we investigated whether BMP9 (a secretary protein) may change the bone microenvironment and, by doing so, regulate the cross-talk between breast cancer cells and bone marrow-derived mesenchymal stem cells. Methods After establishing a co-culture system composed of MDA-MB-231breast cancer cells and HS-5 bone marrowderived mesenchymal stem cells, and exposure of this system to BMP9 conditioned media, we assessed putative changes in migration and invasion capacities of MDA-MB-231 cells and concomitant changes in osteogenic marker expressionin HS-5 cells and metastases-related genes in MDA-MB-231 cells. Results We found that BMP9 can inhibit the migration and invasion of MDA-MB-231 cells, and promote osteogenesis Shaoheng Wan and Yuehong Liu these authors contributed equally to this work. S. Wan : Y. Liu : Y. Weng : W. Wang : C. Fei : Y. Chen : Z. Zhang : T. Wang : J. Wang : Y. Jiang : L. Zhou : Y. Zhang (*) Key Laboratory of Diagnostic Medicine designated by the Ministry of Education, Chongqing Medical University, 1 Yixueyuan Road, Yuzhong District, Chongqing 400016, China e-mail: [email protected] Y. Zhang e-mail: [email protected] W. Ren Department of General Surgery, The First Affiliated Hospital of Chongqing Medical University, 1 Youyi Road, Yuzhong District, Chongqing 400042, China T. He Molecular Oncology Laboratory, Department of Surgery, The University of Chicago Medical Center, Chicago, IL, USA

and proliferation of HS-5 cells, in the co-culture system. We also found that the BMP9-induced inhibition of migration and invasion of MDA-MB-231 cells may be caused by a decreased RANK ligand (RANKL) secretion by HS-5 cells, leading to a block in the AKT signaling pathway. Conclusions From our data we conclude that BMP9 inhibits the migration and invasion of breast cancer cells, and promotes the osteoblastic differentiation and proliferation of bone marrow-derived mesenchymal stem cells by regulating crosstalk between these two types of cells through the RANK/ RANKL signaling axis. Keywords BMP9 . Breast cancer . Bone marrow-derived mesenchymal stem cells . Cross-talk . RANKL/RANK axis Abbreviations BMP9 Bone morphogenetic protein 9 BCC Breast cancer cells BMSC Bone marrow –derived mesenchymal stem cells MSCs Mesenchymal stem cells OPG Osteoprotegerin RANK Receptor activator of nuclear factor-kappa B RANKL The RANK ligand Ad-BMP9 Adenovirus expressing BMP9 protein Ad-GFP Adenovirus expressing green fluorescent protein ERK1/2 Extracellular signal regulated kinases p-ERK1/2 Phospho-ERK1/2 OPN Osteopontin OCN Osteocalcin IL-6 Interleukin-6 IL11 Interleukin-11 PTH-rp Parathyroid hormone-related protein MMP9 Matrix metallopeptidase 9 MMP2 Matrix metallopeptidase 2

364

DKK1 ALP MAPK GSK-3β pGSK-3β CTGF LRP-6 BMP9-CM GFR-CM

S. Wan et al.

Dickkopf WNT signaling pathway inhibitor 1 Alkaline phosphatase Mitogen-activated protein kinase Glycogen synthase kinase-3β Phospho-GSK-3β Connective Tissue Growth Factor Low density lipoprotein receptor-related protein 6 BMP9 condition medium GFP Vector control condition medium

what role BMP9 (as a secretary protein) plays in the bone microenvironment and whether it does affect the cross-talk between BCCs and BMSCs. In the past, it has been reported that BMP9 can increase the osteogenic differentiation of BMSCs [7, 12]. Here, we focused on the effect of BMP9 on the interaction between BCCs and BMSCs, aiming to unravel the role of BMP9 in decreasing the occurrence of osteolytic lesions in the BCC microenvironment. To this end, a coculture system composed of MDA-MB-231 BCCs and HS-5 BMSCs was employed to investigate the effects of BMP9 on the interaction between these two cell types, and to clarify the action of BMP9 in inhibiting BCC osteolytic metastases in the bone microenvironment.

1 Introduction Breast cancer is a common malignancy in women world-wide, more than 1.38 million new cases are diagnosed and about 500,000 individuals die of the disease [1]. Breast cancer cells (BCCs) frequently metastasize to distant organs such as bone, lung and liver [2]. When BCCs metastasize to bone, they disturb the intricate equilibrium of the bone microenvironment, resulting in osteolytic injury [3]. The process of osteolytic metastasis is closely related to the bone microenvironment, and involves osteoblasts, osteoclasts and bone marrow-derived mesenchymal stem cells (BMSCs). In the past, it has been shown that the interaction between BCCs and bone microenvironment cells enhances tumor growth and osteolytic damage [4, 5]. BCCs, on one hand, not only secrete various cytokines to activate osteoclasts, but also activate osteoblasts to secrete osteoclast-activating factors, which results in the formation of osteolytic lesions. Osteoclasts and osteoblasts, on the other hand, secrete cytokines that promote the metastasis and proliferation of BCCs, which aggravates the osteolytic damage [4, 5]. It thus appears that their crosstalk generates a vicious cycle which is sustained by cytokines. Recently, several studies have focused on ways to interfere with the cross-talk between BCCs and bone microenvironment cells [3–5]. Mesenchymal stem cells (MSCs) are present in many tissues, including bone marrow, and they have the capacity to differentiate into three different lineages, i.e., osteogenic, chondrogenic and adipogenic lineages [6]. Luo et al. recently found that BMP9 can induce the osteogenic differentiation of MSCs [7]. In the bone microenvironment, BCCs suppress the differentiation and proliferation of bone marrow-derived stem cells (BMSCs) [8], and BMSCs promote the invasion and migration of BCCs [9]. Previously, we found that BMP9 promotes the apoptosis of MDA-MB-231 breast cancer cells, inhibits the proliferation and invasion of these BCCs, and decreases the occurrence of osteolytic lesions elicited by these BCCs, both in vitro and in vivo [10, 11]. As yet, however, the mechanisms underlying these phenomena remained unclear. We specifically asked

2 Materials and methods 2.1 Cell culture HCT116 colon cancer cells and HS-5 bone marrowderived mesenchymal stem cells were maintained in complete DMEM (Dulbeccos modified Eagle medium) supplemented with 10 % fetal bovine serum (FBS; GIBCO, USA) and 100 units/ml streptomycin/penicillin at 37 °C in a humidified atmosphere of 5 % CO2. MDA-MB231breast cancer cells (purchased from Shanghai Institute for Biological Sciences, Chinese Academy of Science) [13, 14] were maintained in complete L-15 medium supplemented with 10 % FBS and 100 units/ml streptomycin/ penicillin at 37 °C in a humidified atmosphere without CO2. 2.2 Preparation of BMP9 conditioned medium Sub-confluent HCT116 cells cultured in 75 cm2 flaks were infected with optimal titers of recombinant adenoviruses expressing BMP9 and GFP (control), respectively (Ad-BMP9 and Ad-GFP were kindly provided by Professor TC He, Chicago University, USA). 24 h after infection, the culture medium was changed to serum-free medium composed of a mix of L-15 and DMEM. Conditioned medium (CM) was collected at 48 h after infection and used immediately. The experimental group was treated with MDA-MB-231+HS-5+ BMP9 CM, while MDA-MB-231+HS-5+GFP CM treated cells served as a control group and MDA-MB-231+HS-5 CM treated cells as a blank group. 2.3 Western blotting For Western blotting, cells were collected and lysed in RIPA buffer. Total cell lysates were denatured by boiling and loaded onto an 8–15 % gradient SDS-PAGE gel. After electrophoretic separation, proteins were transferred to an Immobilon-P

BMP9 regulates cross-talk between breast cancer cells

365

membrane. This membrane was blocked with Super-Block Blocking Buffer, and probed with primary antibodies, followed by incubation with a secondary antibody conjugated with horseradish peroxidase. The proteins of interest were detected using a SuperSignal West Pico Chemiluminescent Substrate kit. The following primary antibodies were obtained from Santa Cruz or Abcam: anti-BMP9(1:500), anti-phosphor-p38 (1:1500), anti-p38 (1:1000), anti-phosphor-ERK1/2 (1:1500), anti-ERK1/2 (1:1000),anti-phosphor-AKT (1:1500), antiAKT (1:1000), anti-β-actin (1:500), anti-RANKL (1:2000), anti-OPG (1:1000), anti-OPN (1:1000), anti-OCN (1:1000), anti-IL-6 (1:1000), anti-PTH-rp (1:1000) and antiMMP9 (1:1000). 2.4 RNA isolation and RT-PCR Total RNA was isolated using TRIZOL Reagents (Invitrogen) according to the manufacturer’s instructions, and employed to generate cDNA using a Prime Script Kit (TAKARA, Dalian, China). PCR Primers (β-actin, BMP-9, RANKL,OCN and OPN; listed in Table 1) were designed using the Primer 5 program. Gene expression levels were quantified by semiquantitative reverse transcriptase polymerase chain reaction (RT-PCR) for which β-actin was used as an endogenous control. The cycling program used was: initial denaturation at 94 °C for 5 min, followed by 28–34 cycles of denaturation

Table 1 Primer sequences and product sizes of different genes

Gene

at 94 °C for 30 s, annealing for 30 s at 52–58 °C, and extension at 72 °C for 30 s. The PCR products (5 μl) were separated in 1.5 % agarose gels, which were subsequently stained with GoodView Nucleic Acid Stain. Images were acquired using the Quantity One Software package (BIORAD, USA). All samples were run in triplicate and gene expression analyses were performed using the Quantity One Software package (BIO-RAD, USA). 2.5 Quantitative real-time RT-PCR Total RNA was isolated using TRIZOL Reagents (Invitrogen) and used to generate cDNA with a Prime Script Kit (TAKARA, Dalian, China). PCR Primers (GAPDH, IL6, IL11, PTH-rp, RANK, RANKL, MMP2 and MMP9; listed in Table 1) were designed using the Primer 5 program. SYBR Green-based qPCR analyses were performed to amplify the cDNAs of interest using RG-6000 (Corbett Research, Australia). Triplicate reactions were carried out for each sample. The cycling program used was: 94 °C for 2 min for 1 cycle and 30 cycles at 92 °C for 20 s, 57 °C for 30 s, and 72 °C for 20 s, followed by a plate read at 78 °C for each cycle. The relative expression of target mRNAs was normalized to a reference (GAPDH) using the 2-△△CT method, and expressed as the foldchange relative to the control. All samples were run in triplicate.

Primer sequence

Product size (bp)

β-actin

Forward: 5′-CACCACACCTTCTACAATGAGC-3′

695

BMP9

Reverse: 5′-GTGATCTCCTTCTGCATCCTGT-3′ Forward: 5′-GAGCAGTCACGAGGAGGA-3′

322

OCN

Reverse: 5′-ACCCGCAGGGAGGTCTTT-3′ Forward: 5′-GGCAGCGAGGTAGTGAAGAG-3′

230

OPN

Reverse: 5′-CTGGAGAGGAGCAGAACTGG-3′ Forward: 5′- TTGCTTTTGCCTCCTAGGCA-3′

381

GAPDH

Reverse: 5′- GTGAAAACTTCGGTTGCTGG-3′ Forward: 5′- CAGCGACACCCACTCCTC-3′

120

IL-6

Reverse: 5′-TGAGGTCCACCACCCTGT-3′ Forward: 5′- TAGTGAGGAACAAGCCAGAG-3′

234

IL-11

Reverse: 5′-TACATTTGCCGAAGAGCC-3′ Forward: 5′- GCTGACGGGGACCACAAC-3′

124

PTH-rp

Reverse: 5′-GCCGCAGGTAGGACAGTAGG-3′ Forward: 5′- CAGCGACACCCACTCCTC-3′

217

RANKL

Reverse: 5′-TGAGGTCCACCACCCTGT-3′ Forward: 5′- TGCCAACATTTGCTTTCG-3′

127

RANK

Reverse: 5′-TCCTCCTTTCATCAGGGTAT-3′ Forward: 5′- GCAGATCGCTCCTCCAT-3′

132

MMP2

Reverse: 5′-CCACAGGGCAGACATACAC-3′ Forward: 5′- TGGCAAGGAGTACAACAGC-3′

174

MMP9

Reverse: 5′- TGGAAGCGGAATGGAAAC-3′ Forward: 5′- GGGACGCAGACATCGTCATC-3′

139

Reverse: 5′-TCGTCATCGTCGAAATGGGC-3′

366

S. Wan et al.

2.6 Wound healing assay

2.10 Alizarin Red S staining

A wound healing assay was performed to assess the migration of MDA-MB-231 cells. To this end MDA-MB-231 cells, seeded in 6-well plates, were co-cultured with HS-5 cells seeded in Anopore tissue culture plate inserts (0.4 μm, Millipore). When MDA-MB-231 confluent cell areas reached ~80 % of the culture plate, conditioned medium was added to the co-culture system and a wound was created at the center of the culture plate by a pipette tip. Photographs were taken under a microscope at indicated time points (0 to 24 h). The incision width was measured at different sites, and the average wound-closure rate was calculated as:(W1–W2)/W1× 100 %, where W1 is the 0 h incision width and W2 is the 24 h incision width. This assay was independently repeated three times.

Calcium deposition was detected using Alizarin Red S staining as described by Luu et al. [15]. Briefly, HS-5 cells were cultured in medium containing 50 μg/ml α-ascorbic acid, 100 mM β-glycerophosphate and 100 nM dexamethasone. On the 19th day after treatment with conditioned medium, the cells were fixed with 0.05 % (v/v) glutaraldehyde at room temperature for 10 min. After being washed with distilled water, the fixed cells were incubated in 0.4 % Alizarin Red S for 5 min, followed by extensive washing with distilled water. Finally, stained calcium mineral deposits were recorded under bright field microscopy.

2.7 Transwell invasion assay

A MTT assay was performed in quintuplicate to assess the viability of HS-5 cells in the co-culture system. To this end, HS-5 cells were seeded at a concentration of 5×104 cells per well into a 6-well culture plate and MDA-MB-231 cells were seeded at a concentration of 1×104 per well into a 25-mm Anopore tissue culture plate (0.04 μm pore size; Millipore). The cells were co-cultured in conditioned medium from 1 to 5 days. The absorbance of the HS-5 cells was measured on 5 consecutive days at 492 nm using a microplate reader, after which a growth curve was made. This assay was independently repeated in triplicate.

A transwell invasion assay was performed to assess the invasion of MDA-MB-231 cells. To this end, MDA-MB-231 cells were seeded at a density of 4×105/well into the upper chamber of type I-collagen-coated 6-well culture inserts and HS-5 cells were seeded at a density of 2×106/well in to 6-well plates. Duplicate reactions were carried out for each group. The cells were co-cultured in conditioned medium. After 24 h, MDAMB-231 cells were dried for 5 min, fixed with dehydrated alcohol, and stained with hematoxylin-eosin. Next, the cells that invaded the collagen-coated-inserts were counted. Mean values from five randomly selected fields were obtained for each well. This assay was independently repeated three times. 2.8 Quantitative ELISA Three days after BMP9 CM treatment the cytokines RANKL, IL-6, DKK1, PTH-rp and IL-11 in the culture supernatants were measured by quantitative ELISA using a RD Systems kit (R&D) following the manufacturer’s protocol. 2.9 ALP activity measurement On the indicated day after exposure to the conditioned medium, the alkaline phosphatase (ALP) activity of HS-5 cells was measured using a modified Great Escape SEAP Chemiluminescence Assay (BD Clontech, Mountain View, CA, USA). Additionally, ALP expression by HS-5 cells was assessed by a histochemical staining assay (using a mixture of 0.1 mg/ml naphthol AS-MX phosphate and 0.6 mg/ml Fast Blue BB salt), as described by Luu et al. [15]. Each assay was performed in triplicate, and the ALP activity was normalized against the total cellular protein concentrations of the respective samples.

2.11 Cell viability assay

2.12 Animal models Human femur fragments obtained from freshly discarded tissue at the time of surgery from female patients undergoing total hip replacement were washed with PBS and dissected into 1 cm×1 cm pieces. Then the femur pieces were subcutaneously implanted in female nude mice. After 1 week the embedded bone was successfully implanted, as evidenced lack of inflammation and/or necrosis. Female nude mice were randomly divided into three groups comprising three mice each. Simultaneously, MDA-MB-231 breast cancer cells were infected with Ad-BMP9/Ad-GFP virus and a mixed cell suspension of infected MDA-MB-231 cells (5×106 ~5×108) and HS-5 cells (5 × 106 ~ 5 × 108, cell ratio ~1) was injected directly onto the implanted femur fragments in the nude mice. MDA-MB-231+HS-5 and MDA-MB-231/AdGFP+HS-5 served as two control groups, whereas MDA-MB-231/ AdBMP9+HS-5 served as the treatment group. After 2 weeks, when tumors were palpable, tumor diameters were recorded every 3 days. Tumor volumes (V in cm3) were then calculated as: (4π/3)×[(a+b)/4]3, whereπ=3.14. After 5 weeks, tumors and implanted femur fragments were harvested to assess the morphological and immunohistochemical changes. All experiments were approved by the Institutional Animal Care and

BMP9 regulates cross-talk between breast cancer cells

Use Committee of the Chongqing Medical University, as well as by regional authorities. 2.13 Immunohistochemistry Paraffin sections of the subcutaneous tumors were prepared following standard methods. The expression of osteopontin (OPN), a late osteogenic marker, and p-AKT in the tissues was examined by immunohistochemistry (IHC). To this end, the sections were rehydrated and heat-treated for antigen retrieval with citric acid buffer as usual, and then incubated with the respective primary antibodies at 4 °C overnight. On the following day the sections were incubated with a secondary antibody and visualized using 3, 3-diaminobenzidine tetrachloride (DAB) until the desired brown reaction product was obtained. All slides were observed under a Nikon E400 Light Microscope and representative photographs were taken. 2.14 Statistical analyses The results are expressed as means±standard deviations (SD). All statistical analyses were performed by SPSS 17.0 using the independent sample t-test for comparing the two sample groups, where P<0.05 was taken as significance level.

367

12.5 (Fig. 1d). From the metastases-related genes tested in MDA-MB-231 cells in the co-culture system, the mRNA expression levels of IL-6, PTH-rp and MMP9 were found to be decreased by 43 %, 38 % and 51 %, respectively, while there were no significant changes in mRNA expression levels of IL-11and RANK (Table 2). We next set out to examine the protein expression levels of IL-6, PTH-rp and MMP9. As illustrated in Fig. 1e, we found that BMP9 also downregulates the protein expression levels of PTH-rp, IL-6 and MMP9. It is well-documented that cellular signaling pathways, such as the MAPK and AKT pathways, regulate metastases-related genes (i.e., IL-6, PTH-rp and MMP9) in BCC [17–20]. To ascertain whether BMP9 changes these signaling pathways in MDA-MB-231 cells in the co-culture system, the total amounts and phosphorylated forms of P38, ERK1/2 and AKT were determined by Western blotting. As illustrated in Fig. 1f, we found that BMP9 suppresses the activation of P38, ERK1/2 and AKT at different time points without affecting the total amounts of these proteins. It is welldocumented that BMP9 can block the MAPK and AKT signaling pathways. Our data suggest that BMP9 may inhibit the malignant phenotype of MDA-MB-231 cells through these pathways in the co-culture system. 3.2 BMP9 promotes osteogenesis and proliferation of HS-5 cells in the co-culture system

3 Results 3.1 BMP9 inhibits the malignant phenotype of MDA-MB-231 cells in a co-culture system After HCT116 cells were infected with the Ad-BMP9 adenovirus vector (see materials and methods), we verified whether the cells exhibited BMP9 mRNA expression and whether their conditioned medium contained excreted BMP9 protein. By using RT-PCR and Western blotting, we found that BMP9 mRNA was readily detectable in the HCT116/BMP9 cells, and that the conditioned medium indeed contained detectable levels of BMP9 protein (Fig. 1a and b). Next, we set out to assess the effect of BMP9 on MDA-MB-231 cells in the coculture system by analyzing malignancy-associated characteristics such as migration and invasion, and their concomitant signaling pathways. These characteristics are known to play significant roles in osteolytic metastasis [16]. Migration and invasion were assessed by a wound healing and a transwell invasion assay, respectively. After the BMP9 conditioned medium was added to the co-culture system, the healing rate of the MDA-MB-231 cells in the experimental group was found to be significantly decreased from 78.3±3.2 to 28.4± 0.7 % as compared to the cells in control group (Fig. 1c). Also, the number of invading cells which moved across the matrix barrier was found to be decreased from 408.7±17.2 to 169.0±

Osteolytic metastasis is not only related to malignancy, but also to the normal biological characteristics of BMSCs such as osteogenic differentiation and proliferation. We reasoned that uncovering the effects of BMP9 on the normal biological characteristics of BMSCs might contribute to our understanding of the mechanism of bone damage caused by BCCs. To assess the effect of BMP9 on osteogenic differentiation and proliferation of HS-5 cells in the co-culture system, BMP9 conditioned medium was added to this co-culture system. By doing so, we found that the alkaline phosphatase (ALP) activity of HS-5 cells in the experimental group increased in a time-dependent manner and reached a peak on the 9th day. Subsequent cytochemical staining confirmed this observation (Fig. 2a and b). ALP serves as an early osteogenic marker. We also sought to determine whether BMP9 exerts an effect on late osteogenic markers such as osteocalcin (OCN), osteopontin (OPN), or on matrix mineralization. After being treated with the BMP9 conditioned medium for 12 days, total RNA and protein were isolated from HS-5 cells for qPCR analysis and Western blotting, respectively. As illustrated in Fig. 2c, d and e, the mRNA and protein expression levels of OCN and OPN increased upon BMP9 treatment. Additionally, we found through Alizarin Red S staining that after BMP9 treatment for 19 days also calcium deposition was increased (Fig. 2f). Since the Wnt/β-catenin signaling pathway is critical for osteogenic differentiation of MSC [21], we set out to assess the protein

368 Fig. 1 Inhibition of the malignant phenotype of MDAMB-231 cellsBMP9 in the coculture system. a HCT116 cells were infected with Ad-BMP9 or Ad-GFP and BMP9 mRNA expression by the cells was detected using RT-PCR. b HCT116 cells were infected with Ad-BMP9 or Ad-GFP and BMP9 protein in the media supernatants was detected using Western blotting. c, d after adding BMP9 conditioned media to the co-culture system the ability of MDA-MB-231 cells to migrate was assessed using a wound healing assay. Photographs were taken at the indicated time points. The wound closure rate was calculated as: (W1–W2)/W1× 100 %, where W1 is the 0 h incision width and W2 the 24 h incision width. The invasive ability of MDA-MB-231 cells at 24 h was assessed by a transwell invasion assay. Group 1: MDA-MB-231+ HS-5; Group 2: MDA-MB-231+ HS-5+GFP CM; Group 3: MDAMB-231+HS-5+BMP9 CM. e IL6, PTH-rp and MMP9 expression analyzed by Western blotting at day 3. f expression of the MAPK signaling pathway proteins P38 and ERK1/2, and the AKT signaling pathway protein AKT, analyzed by Western blotting at the indicated time points. Data are presented as means±SD of three independent experiments. * P<0.05. Magnification,×100

S. Wan et al.

BMP9 regulates cross-talk between breast cancer cells

369

Table 2 Changes in mRNA expression levels of metastases-related genes by Q-PCR Genes

mRNA expression in experimental group Up or Down (%)a

IL-6 IL-11 PTH-rp RANK

0.57 1.08 0.62 0.94

RANKL not detectable MMP2 0.92 MMP9 0.49 a

43 %↓ 8%38 %↓ 6%not detectable 8%51 %↓

↓ down, ↑ up, - no change

expression levels of GSK-3β, pGSK-3β and β-catenin in HS5 cells by Western blotting. As illustrated in Fig. 2g, BMP9 treatment suppressed pGSK-3β expression without affecting the total amount of GSK-3β, but enhanced β-catenin expression, indicating that the Wnt/β-catenin signaling pathway is activated by BMP9. Next, a MTT assay was employed to assess the proliferation of HS-5 cells after BMP9 treatment for 1 to 5 days. Fig. 2h shows that BMP9 could stimulate the proliferation of HS-5 cells in the co-culture system. These results imply that BMP9 can promote osteogenesis and proliferation of HS-5 cells in the co-culture system. 3.3 BMP9 inhibits MDA-MB-231 cell invasion by blocking the AKT signaling pathway It has been reported that the MAPK and AKT signaling pathways can be activated by environmental cytokines such as IL-6, DKK1, RANKL, PTH-rp and IL-11 [22–25]. Therefore, we tested whether BMP9 may affect the concentration of these cytokines in the co-culture supernatant. After the BMP9 conditioned medium was added to the co-culture system, we found by ELISA that the concentration of IL-6 decreased from 6,500±590 to 3,800±390 pg/ml, DKK1 from 2,150±230 to 1,550 ± 170 pg/ml, RANKL from 4,160 ± 430 to 2,460 ± 220 pg/ml and PTH-rp from 4,900±520 to 3,170±410 pg/ ml. In contrast, we found that the concentration of IL-11 increased from 2,660±300 to 3,600±400 pg/ml. Taken together, these data indicate that BMP9 can decrease the concentration of IL-6, DKK1, RANKL and PTH-rp, but increase the concentration of IL-11 in the co-culture system. Next, we explored the observed RANKL concentration decrease in further detail. To this end, the mRNA and protein expression levels of RANKL were assessed by RT-PCR and Western blotting, respectively. We found that, compared to the control group, the mRNA and protein expression levels of RANKL in HS-5 cells in the experimental group were decreased from 0.070±0.008, 0.099±0.034 to 0.036±0.004, 0.055±0.006, respectively (Fig. 3a and b). In MDA-MB-231 cells no RANKL mRNA or protein expression was detected. Thus,

the decline in expression of RANKL in HS-5 cells very likely caused the decrease in RANKL concentration in the coculture supernatant. The receptor activator of nuclear factorkappa B (RANK) and the RANK ligand (RANKL) pathway is a key molecular axis which promotes the migration and invasion of BCCs by activating the AKT signaling pathway. Osteoprotegerin (OPG), as a soluble decoy receptor for RANKL, can bind RANKL [25]. To determine whether RANKL down-regulation may have an effect on the AKT signaling pathway, 6 ng/ml anti-OPG antibody was added to the co-culture system in the presence of BMP9. In the experimental group, the p-AKT protein was found to increase from 0.027±0.005 to 0.041±0.010 in the MDA-MB-231 cells, while the total amount of AKT protein was not affected. Concomitantly, the number of invading MDA-MB-231 cells, which moved across the matrix barrier, was increased from 169.0±12.5 to 408.7±17.2, whereas no significant change in the wound healing rate of MDA-MB-231 cells was observed (Fig. 3d, e and f). Collectively, these results suggest that BMP9 can inhibit the invasion of MDA-MB-231 cells in the co-culture system due to a decrease in RANKL secretion by HS-5 cells which, in turn, blocks the AKT signaling pathway. 3.4 BMP9 inhibits tumor formation and reduces bone damage in vivo To explore the in vivo effects of BMP9 on breast cancer cells and bone, we subcutaneously implanted bone fragments on the back of nude mice (Fig. 4a, Left). After 1 week, a mixed cell suspension of MDA-MB-231 cells infected with AdBMP9/Ad-GFP and HS-5 cells was injected nestled against the bone (Fig. 4a, Right). We found that the tumor volume decreased in the BMP9 over-expression group after 30 days (Fig. 4b). Subsequent histological and histomorphometric analyses showed that BMP9 reduced the damage of the embedded bone (Fig. 4c). Immunohistochemistry revealed that OPN, a late osteogenic marker, was significantly increased in the BMP9 over-expression group (Fig. 4d). p-AKT staining in the BMP9 over-expression group was found to be decreased compared to the control group.

4 Discussion BMP9 (also known as growth differentiation factor 2 or GDF2) is known to exert several biological effects on tumor cells. It has been reported that it can promote the growth of ovarian cancer cells through the BMP/SMAD pathway, inhibit the growth and invasion of prostate cancer PC-3 cells through the BMP/SMAD and a non-canonical BMP/SMAD pathway, enhance the apoptosis of PC-3 cells through the prostate apoptosis response protein 4 (PAR4), and enhance the

370 Fig. 2 BMP9 promotes osteogenesis and proliferationof HS-5 cells in the co-culture system. a, b after the BMP9 conditioned media were added to the co-culture system, ALP activity and ALP cytochemical staining were assessed at the indicated time points. c, d, e mRNA and protein expression levels of OCN and OPN detected by RT-PCR and Western blotting at 3 days and 12 days, respectively. f calcium deposition analyzed by Alizarin Red S staining at 19 days. g protein expression levels of GSK-3β, pGSK-3β and β-catenin detected by Western blotting. h cell proliferation from 1 day to 5 days assessed by MTT. *P<0.05

S. Wan et al.

BMP9 regulates cross-talk between breast cancer cells Fig. 3 BMP9 inhibits the invasion of MDA-MB-231 in the co-culture system. a after exposure to BMP9 conditioned medium for 4 days, the co-culture supernatant was harvested. Chemokine levels were quantified by ELISA. b, c mRNA and protein expression levels of RANKL in HS-5 cells analyzed by RT-PCR and Western blotting. d, e, f after anti-OPG antibody was added to the co-culture system containing BMP9, p-AKT and AKT were detected by Western blotting and the concomitant invasion and migration capacities of MDA-MB231 cells were assessed. Data are presented as means±SD from three independent experiments. * P<0.05

371

372

S. Wan et al.

Fig. 4 BMP9 over-expression inhibits tumor formation and reduces bone damage in vivo. a cancellous bone was implanted subcutaneously on the back of nude mice (Left) and cells were inoculated after a week (Right). b when tumors were palpable, the tumor diameters were recorded every 3 days. The tumor volume V (in cm3) was then calculated as: (4π/3)×[(a+b)/4]3, where π= 3.14. c, d Histologic and histomorphometric bone damage (arrows) in each group; enlarged view on the right. p-Akt and OPN staining of the various groups using immunochemistry

apoptosis of MDA-MB-231 breast cancer cells and repress their invasion and migration capacities via the downregulation of CTGF [26, 27, 10, 11]. In addition, it has been shown that BMP9 can induce the osteogenic differentiation of mesenchymal stem cells [7, 12]. Breast cancer cells (BCC) preferentially metastasize to bone tissue and cause local hypercalcemia, osteolytic bone fractures and other symptoms [28, 29]. In general, patients with breast cancer metastases have a poor prognosis and a low quality of life. The occurrence of osteolytic metastases is closely related to the local bone microenvironment, especially the presence of bone marrow stem cells (BMSC). BMSC can

promote the migration and invasion of BCC and, thereby, aggravate bone injury. Previously, is has been shown that BMP9 can not only inhibit osteolytic injury caused by MDA-MB-231breast cancer cells through down-regulation of the connective tissue growth factor CTGF [10, 11], but also inhibit the growth, invasion and bone metastasis of MDAMB-231 cells through down-regulation of the SMAD signaling pathway [30]. Here, we addressed the question whether BMP9 (as a secreted protein) can exert an effect on the bone microenvironment invaded by MDA-MB-231 breast cancer cells and the cross-talk between MDA-MB-231 cells and HS5 bone marrow-derived mesenchymal stem cells.

BMP9 regulates cross-talk between breast cancer cells

First of all, we investigated whether BMP9 can affect several malignant properties of MDA-MB-231 cells, such as migration and invasion, in the presence of HS-5 cells. Our results indicate that BMP9 can indeed inhibit the migration and invasion of MDA-MB-231 cells in the presence of HS-5 cells. In order to shed further light on the mechanism(s) underlying these inhibitory effects, the expression of several metastases-related genes and their effect on related signaling pathways were assessed in MDA-MB-231 cells by qPCR and Western blotting, respectively. Our data show that a decrease in the expression of IL-6, PTH-rp and MMP9 may be related to blocking of the MAPK and AKT signaling pathways. According to recent reports, cytokines such as PTH-rp, IL-6, IL-8 and RANKL can directly or indirectly promote the migration and invasion of BCCs by potentiating the activities of the MAPK and AKT signaling pathways [31–33, 25]. In this present study, we found that the concentrations of IL-6, RANKL and PTH-rp decreased, putatively leading to a block of the MAPK and AKT signaling pathways. RANKL, in conjunction with RANK, activates the AKT signaling pathway which, in turn, can be suppressed by binding of osteoprotegerin (OPG). In the bone microenvironment invaded by BCC, RANKL can be secreted by osteoblasts, mesenchymal stem cells (MSC) and BCC [29]. We found that RANKL in the co-culture medium was only derived from HS-5 cells. Additionally, after adding anti-OPG antibody to the coculture system, we found that the AKT signaling pathway was recovered and the invasive capacity of MDA-MB-231 cells increased next to its ability to migrate. From these results, we conclude that BMP9 reduces RANKL secretion of HS-5 cells and inhibits the invasion of MDA-MB-231 cells by blocking the AKT signaling pathway. Osteolytic metastasis is not only related to the malignant degree of BCCs, but also to the biological characteristics of BMSCs. Das et al. [34] reported that bone integrity is maintained by an intricate equilibrium between the activities of bone-forming osteoblasts and bone-resorbing osteoclasts. As osteoblasts derive from BMSCs, these cells play a decisive role in maintaining bone homeostasis. The activity of osteoblasts can, however, be hampered by metastatic tumor cells within the bone microenvironment, and the resulting bone defects are difficult to restore [35]. Therefore, to further our understanding of the mechanism underlying osteolytic lesion formation and to provide a basis for the treatment of osteolytic bone fractures, we assessed whether BMP9 can promote the osteogenic differentiation and proliferation of BMSCs. Our results indicate that BMP9 can not only increase the expression and activity of alkaline phosphatase (ALP) in HS-5 cells on the 9th day, but also increase the mRNA and protein expression levels of osteocalcin (OCN) and osteopontin (OPN) on the 12th day, as well as the calcium deposition on the 19th day by HS-5 cells. ALP is an early osteogenic marker and its activity starts to decrease during the late differentiation

373

period when the expressions of late osteogenic markers such as OCN and OPN start to increase [36]. We found that the activity of ALP began to decrease after treatment with BMP9 conditioned medium after 9 days. In addition, we found that the concentration of DKK1 decreased in the co-culture system and that the Wnt/β-catenin signaling pathway recovered. DKK1, in conjunction with LRP-6, blocks the Wnt/βcatenin signaling pathway, which is suppressed in the bone microenvironment of breast cancer patients [37]. Therefore, BMP9 may promote the osteogenesis of HS-5 cells through a decrease in DKK1 caused by the recovered Wnt/β-catenin signaling pathway. Moreover, we found that BMP9 enhances the proliferation of HS-5 cells in the co-culture system. Our in vivo results showed that the tumor volume decreased in the BMP9 over-expression group. In addition, histomorphometric and immunohistochemical analyses showed that BMP9 can promote osteogenesis by increasing the expression of OPN and inhibiting bone damage. This effect may relate to the observed decreases in RANKL, IL-6 and PTH-rp. These latter factors are known to be crucial for osteoclast activation and breast cancer cell proliferation [38–40]. Thus, BMP9 may serve as therapeutic agent for breast cancer bone metastases. In summary, we have shown that BMP9 may regulate the cross-talk between MDA-MB-231 breast cancer cells and HS5 bone marrow-derived mesenchymal stem cells in a coculture system. Our results indicate that BMP9 can reduce RANKL secretion by HS-5 cells and inhibit the invasion of MDA-MB-231 cells through blocking the AKT signaling pathway. Simultaneously, BMP9 can promote the osteogenic differentiation and proliferation of HS-5 cells in the tumor microenvironment. Acknowledgments This work was supported by the Nature Science Foundation of China (81172017) and the National Basic Research Program of China (973 Program, 2011CB707906). Conflict of interest There is no conflict of interest to declare.

References 1. A. Halon, P. Donizy, P. Surowiak, R. Matkowski, ERM/Rho protein expression in ductal breast cancer: a 15 year follow-up. Cell Oncol (Dordr) 36, 181–190 (2013) 2. A. Lipton, R. Uzzo, R.J. Amato, G.K. Ellis, B. Hakimian, G.D. Roodman, M.R. Smith, The science and practice of bone health in oncology: managing bone loss and metastasis in patients with solid tumors. J Natl Compr Cancer Netw 7(Suppl 7), S1–S29 (2009). quiz S30 3. I. Zinonos, K.W. Luo, A. Labrinidis, V. Liapis, S. Hay, V. Panagopoulos, M. Denichilo, C.H. Ko, G.G. Yue, C.B. Lau, W. Ingman, V. Ponomarev, G.J. Atkins, D.M. Findlay, A.C. Zannettino, A. Evdokiou, Pharmacologic inhibition of bone resorption prevents cancer-induced osteolysis but enhances soft tissue metastasis in a mouse model of osteolytic breast cancer. Int. J. Oncol. 45, 532–540 (2014)

374 4. R.L. Theriault, R.L. Theriault, Biology of bone metastases. Cancer Control 19, 92–101 (2012) 5. R. Krawetz, Y.E. Wu, D.E. Rancourt, J. Matyas, Osteoblasts suppress high bone turnover caused by osteolytic breast cancer in-vitro. Exp. Cell Res. 315, 2333–2342 (2009) 6. P.O. Favaron, A. Mess, S.E. Will, P.C. Maiorka, M.F. de Oliveira, M.A. Miglino, Yolk sac mesenchymal progenitor cells from New World mice (Necromys lasiurus) with multipotent differential potential. PLoS One 9, e95575 (2014) 7. J. Luo, M. Tang, J. Huang, B.C. He, J.L. Gao, L. Chen, G.W. Zuo, W. Zhang, Q. Luo, Q. Shi, B.Q. Zhang, Y. Bi, X. Luo, W. Jiang, Y. Su, J. Shen, S.H. Kim, E. Huang, Y. Gao, J.Z. Zhou, K. Yang, H.H. Luu, X. Pan, R.C. Haydon, Z.L. Deng, T.C. He, TGFbeta/BMP type I receptors ALK1 and ALK2 are essential for BMP9-induced osteogenic signaling in mesenchymal stem cells. J. Biol. Chem. 285, 29588– 29598 (2010) 8. D. Mendoza-Villanueva, L. Zeef, P. Shore, Metastatic breast cancer cells inhibit osteoblast differentiation through the Runx2/CBFβdependent expression of the Wnt antagonist, sclerostin. Breast Cancer Res. 13, R106 (2011) 9. A.P. Molloy, F.T. Martin, R.M. Dwyer, T.P. Griffin, M. Murphy, F.P. Barry, T. O’Brien, M.J. Kerin, Mesenchymal stem cell secretion of chemokines during differentiation into osteoblasts, and their potential role in mediating interactions with breast cancer cells. Int. J. Cancer 124, 326–332 (2009) 10. K. Wang, H. Feng, W. Ren, X. Sun, J. Luo, M. Tang, L. Zhou, Y. Weng, T.C. He, Y. Zhang, BMP9 inhibits the proliferation and invasiveness of breast cancer cells MDA-MB-231. J. Cancer Res. Clin. Oncol. 137, 1687–1696 (2011) 11. X. Sun, K. Wang, H. Feng, Effects and possible mechanism of BMP9 on the bone metastasis of human breast cancer cells MDA-MB- 231. Chin Biotechnol 32, 7–13 (2012) 12. N. Tang, W.X. Song, J. Luo, X. Luo, J. Chen, K.A. Sharff, Y. Bi, B.C. He, J.Y. Huang, G.H. Zhu, Y.X. Su, W. Jiang, M. Tang, Y. He, Y. Wang, L. Chen, G.W. Zuo, J. Shen, X. Pan, R.R. Reid, H.H. Luu, R.C. Haydon, T.C. He, BMP-9-induced osteogenic differentiation of mesenchymal progenitors requires functional canonical Wnt/betacatenin signalling. J. Cell. Mol. Med. 13, 2448–2464 (2009) 13. N. Serna-Marquez, S. Villegas-Comonfort, O. Galindo-Hernandez, N. Navarro-Tito, A. Millan, E.P. Salazar, Role of LOXs and COX-2 on FAK activation and cell migration induced by linoleic acid in MDA-MB-231 breast cancer cells. Cell Oncol 36, 65–77 (2013) 14. J.S. Park, S.Y. Choi, J.H. Lee, J.H. Lee, M. Lee, E.S. Nam, A.L. Jeong, S. Lee, S. Han, M.S. Lee, J.S. Lim, Y. Yoon do, Y. Kwon, Y. Yang, Interleukin-32β stimulates migration of MDA-MB-231 and MCF-7cells via the VEGF-STAT3 signaling pathway. Cell Oncol 36, 493–503 (2013) 15. H.H. Luu, W.X. Song, X. Luo, D. Manning, J. Luo, Z.L. Deng, K.A. Sharff, A.G. Montag, R.C. Haydon, T.C. He, Distinct roles of bone morphogenetic proteins in osteogenic differentiation of mesenchymal stem cells. J. Orthop. Res. 25, 665–677 (2007) 16. J.A. Sterling, S.A. Guelcher, Bone structural components regulating sites of tumor metastasis. Curr Osteoporos Rep 9, 89– 95 (2011) 17. J.L. Gilmore, J.A. Scott, Z. Bouizar, A. Robling, S.E. Pitfield, D.J. Riese 2nd, J. Foley, Amphiregulin-EGFR signaling regulates PTHrP gene expression in breast cancer cells. Breast Cancer Res. Treat. 110, 493–505 (2008) 18. S. Ley, A. Weigert, B. Weichand, N. Henke, B. Mille-Baker, R.A. Janssen, B. Brune, The role of TRKA signaling in IL-10 production by apoptotic tumor cell-activated macrophages. Oncogene 32, 631– 640 (2013) 19. J.Y. Jang, Y.K. Jeon, C.W. Kim, Degradation of HER2/neu by ANT2 shRNA suppresses migration and invasiveness of breast cancer cells. BMC Cancer 10, 391 (2010)

S. Wan et al. 20. K. Tawara, J.T. Oxford, C.L. Jorcyk, Clinical significance of interleukin (IL)-6 in cancer metastasis to bone: potential of anti-IL-6 therapies. Cancer Manag. Res. 3, 177–189 (2011) 21. Y. Guo, P.F. Li, X.C. Shu, H. Deng, H.L. Ma, L. Sun, Involvement of Wnt/beta-catenin signaling in the osteogenesis of bone marrow mesenchymal stem cells induced by drynaria total flavonoids. Zhonghua Yi Xue Za Zhi 92, 2288–2291 (2012) 22. L. Zhang, Y. Teng, Y. Zhang, J. Liu, L. Xu, J. Qu, K. Hou, X. Yang, Y. Liu, X. Qu, C-Src-mediated RANKL-induced breast cancer cell migration by activation of the ERK and Akt pathway. Oncol Lett 3, 395–400 (2012) 23. Y. Zhang, L. Wang, M. Zhang, M. Jin, C. Bai, X. Wang, Potential mechanism of interleukin-8 production from lung cancer cells: an involvement of EGF-EGFR-PI3K-Akt-Erk pathway. J. Cell. Physiol. 227, 35–43 (2012) 24. S. Ibaragi, T. Shimo, M. Iwamoto, N.M. Hassan, S. Kodama, S. Isowa, A. Sasaki, Parathyroid hormone-related peptide regulates matrix metalloproteinase-13 gene expression in bone metastatic breast cancer cells. Anticancer Res. 30, 5029–5036 (2010) 25. L. Zhang, Y. Teng, Y. Zhang, J. Liu, L. Xu, J. Qu, K. Hou, Y. Liu, X. Qu, Proteasome inhibitor bortezomib (PS-341) enhances RANKLinduced MDA-MB-231 breast cancer cell migration. Mol Med Rep 5, 580–584 (2012) 26. L. Ye, H. Kynaston, W.G. Jiang, Bone morphogenetic protein-9 induces apoptosis in prostate cancer cells, the role of prostate apoptosis response-4. Mol. Cancer Res. 6, 1594–1606 (2008) 27. L. Ye, H. Kynaston, W.G. Jiang, Bone morphogenetic protein-10 suppresses the growth and aggressiveness of prostate cancer cells through a Smad independent pathway. J. Urol. 181, 2749–2759 (2009) 28. W. Irvin Jr., H.B. Muss, D.K. Mayer, Symptom management in metastatic breast cancer. Oncologist 16, 1203–1214 (2011) 29. Y.C. Chen, D.M. Sosnoski, A.M. Mastro, Breast cancer metastasis to the bone: mechanisms of bone loss. Breast Cancer Res. 12, 215 (2010) 30. W. Ren, X. Sun, K. Wang, H. Feng, Y. Liu, C. Fei, S. Wan, W. Wang, J. Luo, Q. Shi, M. Tang, G. Zuo, Y. Weng, T. He, Y. Zhang, BMP9 inhibits the bone metastasis of breast cancer cells by downregulating CCN2 (connective tissue growth factor, CTGF) expression. Mol. Biol. Rep. 41, 1373–1383 (2014) 31. Y.H. Wang, Y.Y. Dong, W.M. Wang, X.Y. Xie, Z.M. Wang, R.X. Chen, J. Chen, D.M. Gao, J.F. Cui, Z.G. Ren, Vascular endothelial cells facilitated HCC invasion and metastasis through the Akt and NF-κB pathways induced by paracrine cytokines. J. Exp. Clin. Cancer Res. 32, 51 (2013) 32. A. De Luca, L. Lamura, M. Gallo, V. Maffia, N. Normanno, Mesenchymal stem cell-derived interleukin-6 and vascular endothelial growth factor promote breast cancer cell migration. J. Cell. Biochem. 113, 3363–3370 (2012) 33. L.E. Wright, J.B. Frye, A.L. Lukefahr, B.N. Timmermann, K.S. Mohammad, T.A. Guise, J.L. Funk, Curcuminoids block TGF-β signaling in human breast cancer cells and limit osteolysis in a murine model of breast cancer bone metastasis. J. Nat. Prod. 76, 316–321 (2013) 34. S. Das, R.S. Samant, L.A. Shevde, Hedgehog signaling induced by breast cancer cells promotes osteoclastogenesis and osteolysis. J. Biol. Chem. 286, 9612–9622 (2011) 35. S. Das, J.A. Tucker, S. Khullar, R.S. Samant, L.A. Shevde, Hedgehog signaling in tumor cells facilitates osteoblast-enhanced osteolytic metastases. PLoS One 7, e34374 (2012) 36. T.A. Owen, M. Aronow, V. Shalhoub, L.M. Barone, L. Wilming, M.S. Tassinari, M.B. Kennedy, S. Pockwinse, J.B. Lian, G.S. Stein, Progressive development of the rat osteoblast phenotype in vitro: reciprocal relationships in expression of genes associated with osteoblast proliferation and differentiation during formation of the bone extracellular matrix. J. Cell. Physiol. 143, 420–430 (1990)

BMP9 regulates cross-talk between breast cancer cells 37. G. Bu, W. Lu, C.C. Liu, K. Selander, T. Yoneda, C. Hall, E.T. Keller, Y. Li, Breast cancer-derived Dickkopf1 inhibits osteoblast differentiation and osteoprotegerin expression: implication for breast cancer osteolytic bone metastases. Int. J. Cancer 123, 1034–1042 (2008) 38. Z. Zhai, X. Qu, W. Yan, H. Li, G. Liu, X. Liu, T. Tang, A. Qin, K. Dai, Andrographolide prevents human breast cancer-induced osteoclastic bone loss via attenuated RANKL signaling. Breast Cancer Res. Treat. 144, 33–45 (2014)

375 39. N. Kurio, T. Shimo, T. Fukazawa, M. Takaoka, T. Okui, N.M. Hassan, T. Honami, S. Hatakeyama, M. Ikeda, Y. Naomoto, A. Sasaki, Anti-tumor effect in human breast cancer by TAE226, a dual inhibitor for FAK and IGF-IR in vitro and in vivo. Exp. Cell Res. 317, 1134–1146 (2011) 40. C. Bolin, K. Tawara, C. Sutherland, J. Redshaw, P. Aranda, J. Moselhy, R. Anderson, C.L. Jorcyk, Oncostatin m promotes mammary tumor metastasis to bone and osteolytic bone degradation. Genes Cancer 3, 117–130 (2012)