This article is mentioned in:

Introduction

Breast cancer is the most common malignant tumor of women that originates in breast epithelial tissue, and it is found in women aged 40–60 years either before or after menopause. The incidence of breast cancer has been increasing year by year in recent years (1), and it has become the second most common cancer worldwide. Breast cancer accounts for ~25% of all cancer-associated mortalities (2). Triple-negative breast cancer (TNBC) accounts for ~12–20% of all incidences of breast cancer and is associated with early recurrence and poor prognosis (2,3). TNBC is a tumor that lacks the expression of the three common cell surface receptors-estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor-2 (HER2) (3,4). Therefore, TNBC evades effective treatment as far as the majority of the available targeted therapies are concerned (4). Owing to the high degree of malignancy, invasive ability and the risk of distant metastasis, TNBC has become a major focus of current research. As a representative cell line of TNBC, MDA-MB-231 cells have been extensively studied, achieving some meaningful results (5).

The family of bone morphogenetic proteins (BMPs) belongs to the transforming growth factor-β (TGF-β) superfamily (6). BMPs were originally viewed as being osteoinductive cytokines that promote bone and cartilage formation in vivo (7). Recently, BMPs have been shown to be involved in the regulation of tumorigenesis, development and bone metastases (8). BMP9 [also known as growth differentiation factor 2 (GDF2)] is an integral member of the BMP family that performs different functions in different tumor types (9). For example, BMP9 may promote the proliferation and migration of liver cancer cells (10), although it inhibits the growth of osteosarcoma cells (11). Our previous study revealed that BMP9 could inhibit the invasion, migration and bone metastasis of MDA-MB-231 cells via the classical BMP/SMAD signaling pathway (12). Upon performing further studies, we discovered that BMP9 may also have a role in other non-SMAD-dependent signaling pathways that are involved in breast cancer, including the phosphoinositide 3-kinase (PI3K)/Akt signaling pathway.

The PI3K/Akt pathway is one of the predominant pathways that are involved in the malignant progression of various types of tumors and mediates the proliferation, migration and invasion of breast cancer tumors. The overexpression of BMP9 reduces phosphorylation of Akt, whereas interfering with BMP9 expression increases phosphorylation of Akt (13). Furthermore, the aberrant activation of Akt is an important component in the recurrence and metastasis of tumors (14–16). Certain BMPs are able to activate the PI3K/Akt signaling pathway. Kang et al (17) identified that BMP2 may promote the invasion and migration of gastric cancer cells by activating the PI3K/Akt pathway and by upregulating matrix metalloproteinase 2 (MMP2). BMPs 4 and 7 were found to inhibit the apoptosis of ovarian granulosa cells via the PI3K/phosphoinositide-dependent kinase 1 (PDK-1)/Akt and PI3K/PDK-1/protein kinase C (PKC) signaling pathways, respectively (18). A previous study published by our research group revealed that BMP9 was able to inhibit the growth, invasion and migration of MDA-MB-231 cells (19). However, whether the inhibitory mechanism is associated with the PI3K/Akt signaling pathway has yet to be elucidated. Therefore, the present study aimed to investigate the effect of BMP9 on the proliferation of breast cancer cells and to explore the possible mechanism to potentially provide a basis for targeted therapy of breast cancer.

Materials and methods

Cell lines and adenoviral vectors

MDA-MB-231 breast cancer cells were purchased from the Cell Resource Center, Shanghai Institute for Biological Sciences (Chinese Academy of Sciences). The adenovirus that overexpresses BMP9 and the adenovirus that expresses green fluorescent protein (GFP), were obtained from the Molecular Medical Laboratory of Yongchuan Hospital of Chongqing Medical University, Chongqing, China.

Reagents

TRIzol® reagent was purchased from Takara Biotechnology Co., Ltd. (Dalian, China); L15 medium was purchased from Life Technologies (Thermo Fischer Scientific, Inc., Waltham, MA, USA). Cell protein extraction and western blot detection reagents RIPA lysis and extraction buffer were purchased from Beyotime Biotechnology Co., Ltd. (Shanghai, China). Enhanced chemiluminescence (ECL) detection kit Enlight™ was purchased from Engreen Biological Co., Ltd. (Beijing, China). Phosphorylase and protease inhibitor cocktail tablets PhosSTOP were purchased from Roche Diagnostics (Basel, Switzerland). Mouse monoclonal GAPDH antibody (cat. no. 97166), Akt total protein mouse monoclonal (cat. no. 2920) and Akt phosphoprotein rabbit monoclonal (specific to Ser473; cat. no. 4060) antibodies, cyclin E1 monoclonal antibody (cat. no. 4129), cyclin D1 rabbit monoclonal antibody (cat. no. 2978) and cyclin B1 mouse monoclonal antibody (cat. no. 4135) were purchased from Cell Signaling Technology, Inc. (Danvers, MA, USA). MM9 mouse monoclonal antibody (cat. no. ab58803) and BMP9 rabbit polyclonal antibody (cat. no. ab71809) were obtained from Abcam (Cambridge MA, USA). PV9001 horseradish enzyme anti-rabbit immunoglobulin G (IgG) polymer, PV9002 horseradish enzyme anti-mouse IgG polymer, and the DAB display reagent were purchased from Zhongshan Golden Bridge Biological Co., Ltd. (Beijing, China).

Patients and animals

A total of 23 pairs of breast cancer and adjacent normal tissues (23 females, aged 25–77 years) were obtained from the subjects (Table I), which were confirmed by pathological examination using H&E staining. All samples were collected from the Yongchuan Hospital of Chongqing Medical University from May 2016 to August 2017. All the experiments concerning the tumor specimens were approved by the Ethics Association of Yongchuan Hospital of the Chongqing Medical University, and all the patients signed the informed consent form.

Table I. Clinicopathological features of the 23 breast cancer patients recruited to the study.

Table I.

Clinicopathological features of the 23 breast cancer patients recruited to the study.

Patient	Age (years)	Sex	Tumor grade	Tumor stage	Lymph node metastasis	Tumor size (cm)
1	51	Female	II	T2N3M0	Yes	2.5×2×1
2	53	Female	I	T1N1M0	Yes	1.5×1.5×1
3	47	Female	II	T1N0M0	No	2×2×1
4	54	Female	II	T1N1M0	Yes	3×2×1
5	53	Female	II	T1N1M0	Yes	1.8×1.5×1
6	49	Female	II	T1N1M0	Yes	1.5×1×1
7	45	Female	III	T1N1M0	Yes	2.5×1×1.5
8	64	Female	II	T2N0M0	No	2×1.8×1.5
9	37	Female	II	T1N0M0	No	1.5×1.2×1
10	53	Female	II	T1N0M0	No	1×1×1
11	52	Female	II	T1N0M0	No	0.8×0.8×0.7
12	50	Female	II	T2N0M0	No	2.5×2×1.8
13	55	Female	II	T1N0M0	No	1.5×1.3×1
14	74	Female	II	T1N1M0	Yes	1.5×1.5×1.2
15	53	Female	II	T1N0M0	No	2×1.5×1
16	58	Female	II	T1N0M0	No	0.8×0.8×0.6
17	25	Female	II	T2N1M0	Yes	2.5×1.5×1
18	43	Female	III	T1N0M0	Yes	2×2×1.2
19	53	Female	II	T1N0M0	Yes	1×0.8×0.8
20	43	Female	II	T2N0M0	Yes	2.5×1.8×1.6
21	77	Female	II	T1N0M0	No	2×1.6×1.5
22	50	Female	II	T1N0M0	Yes	2×1.8×1.5
23	42	Female	I	T2N3M0	No	3×2.5×1.2

Four-week-old male nude mice (weight 21 g) (BALB/c; Beijing Hua Fukang Biological Polytron Technologies Inc., Beijing, China) were used for generation of the xenograft model. All experiments were approved by the Institutional Animal Care and Use Committee of the Chongqing Medical University, as well as regional authorities [reference number, SYXK(YU)2014-0001] in accordance with the ‘Guidelines for the Welfare of Animals in Experimental Neoplasia’ (The China Coordinating Committee on Cancer Research). The animals were housed in individually ventilated cages under sterile conditions and were granted unrestricted access to food and water. At the end of the experimental period, the animals were sacrificed by excessive ether inhalation.

Immunohistochemistry

The prepared tissue sections were dewaxed, hydrated and antigen retrieval was performed. After washing three times with PBS (5 min each wash), antibodies against BMP9, Akt and phosphorylated Akt were added dropwise followed by incubation in the humid chamber at 37°C for 2 h. Following three further washes with PBS (5 min each wash), a secondary antibody (PV9001 or PV9002) was added, and the sections were incubated at 37°C for a further 20 min. DAB and hematoxylin counterstain were added for 3 min followed by routine dehydration to maintain transparency and sealing. Each slice was photographed using the BX51P Olypmus polarizing microscope (Olympus Corp., Tokyo, Japan) at a low magnification ×100 (with 10 randomly selected fields of view), and the cumulative optical density of the dyed yellow areas was assessed using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Cell culture and infection

MDA-MB-231 cells were cultured in L15 culture medium containing 10% fetal bovine serum(FBS) (including 1% penicillin and streptomycin) at 37°C, in the absence of CO2. When the cells had reached a confluency of ~70–80%, they were digested and passaged with 0.25% trypsin (containing 0.02% EDTA 2Na), and the culture was continued. MDA-MB-231 cells in the exponential phase of growth were seeded at a density of 2×106/culture flask. After 24 h, the cells were subsequently infected with adenovirus vehicles that express GFP and recombinant BMP-9. After a further 8–12 h cultivation, the medium was replaced by fresh medium, and the fluorescence was subsequently observed 24 h later using a fluorescence microscope (Olympus Imaging IX71; Olympus Corp.).

Real-time quantitative polymerase chain reaction (RT-qPCR)

The total RNA of the MDA-MB-231 cells, and the total RNA of the MDA-MB-231/BMP9 and MDA-MB-231/GFP cells were extracted using the TRIzol® method after 24 h of adenovirus infection (12). Aliquots (2 µg) of total RNA were used to synthesize cDNA using a reverse transcription kit, and subsequently 1 µl cDNA was used as a template for PCR amplification. The following primer sequences were used: GAPDH, sense, 5′-CAGCGACACCCACTCCTC-3′ and antisense, 5′-TGAGGTCCACCACCCTGT-3′; cyclin B1 sense, 5′-GCACTTTCCTCCTTCTCA-3′ and antisense, 5′-CGATGTGGCATACTTGTT-3′; cyclin D1 sense, 5′-GCGAGGAACAGAAGTGCG-3′andantisense,5′-GGAGTTGTCGGTGTAGATGC-3′;cyclinE1sense,5′-TGGATGTTGACTGCCTTGA-3′andantisense,5′-GTCGCACCACTGATACCCT-3′;c-Mycsense,5′-ACACCCTTCTCCCTTCG-3′ and antisense, 5′-CCGCTCCACATACAGTCC-3′; and MMP9 sense, 5′-GAACTTTGACAGCGACAAGA-3′ and antisense, 5′-GGCGAGGACCATAGAGG-3′. The PCR reaction conditions were as follows: An initial step at 95°C for 30 sec; denaturation at 95°C for 5 sec; followed by annealing for 30 sec at 55°C and extension for 10 sec at 72°C in a total of 40 cycles. The experiment was repeated three times. The 2−ΔΔCq method was utilized to normalize the expression of target mRNAs relative to the reference gene, GAPDH (12).

Western blotting

A density of 5×105 MDA-MB-231 cells/culture flask were passaged into 6-cm diameter Petri dishes. When the cell confluency reached ~50–60%, the adenovirus that overexpresses BMP9 was added, and the medium was subsequently replaced with serum-free L15 medium after 24 h. The cells were cultivated for a further 48 h, at which time the culture solution was discarded, and the cells were subsequently washed twice with pre-cooled PBS. Cell lysate (150 µl/ Petri dish) was added, followed by lysis for 30 min on ice, and the cells were centrifuged at 4°C for 15 min at 12,000 rpm (10,000 × g/min). The supernatant was then transferred to a 1.5 ml EP tube, prior to utilizing the bicinchoninic acid (BCA) method for the determination of protein concentration. Aliquots (15 µl) of the protein solution were extracted for protein electrophoresis (SDS-PAGE; concentration of gel 5%, separation gel, 10%). The proteins were electrotransferred onto a polyvinylidene difluoride (PVDF) membrane, and then blocked with 5% bovine serum albumin (BSA) at 37°C for 1 h. The primary antibodies against GAPDH (dilution 1:1,000), Akt (dilution 1:1,000), phosphorylated Akt (p-Akt; dilution 1:1,000), c-Myc (dilution 1:1,000), cyclin D1 (dilution 1:1,000), cyclin B1 (dilution 1:1,000), cyclin E1 (dilution 1:1,000) and MMP9 (dilution 1:500) were added and incubated overnight at 4°C. The membranes were then washed three times (15 min each wash) in 0.1% Tris-buffered saline/Tween (TBST) solution, after which the corresponding secondary antibody (goat anti-rabbit or goat anti-mouse) (dilution, 1:5,000) was added. The cells were incubated for 1 h at 37°C, and the PVDF membranes were washed three times (15 min each wash) in 0.1% TBST solution. Finally, the ECL chromogenic solution was used for imaging. GEL EQ (Bio-Rad Laboratories, Hercules, CA, USA) was used in the gel imager, and Quantity One® software (Bio-Rad Laboratories, Inc.) was used to analyze the relative optical densities. The relative expression level of the target gene protein was calculated according to the gray value of the target gene band/the gray value of the internal reference band.

Animal models

Trypsin solution (0.25%) was used to digest the MDA-MB-231 cells, and the MDA-MB-231/GFP and MDA-MB-231/BMP9 cells in the exponential growth phase 24 h following adenovirus infection. Tumor cells were injected subcutaneously into the backs of the nude mice for each group of five nude mice (0.1 ml/nude mouse; adjusted cell density, 5×106 cells/ml). Once the tumors of any group were visible and palpable, the starting point measurements were taken. The diameters of the tumor were subsequently recorded every three days. The volume (V) was calculated based on the following formula: V = 4/3π l/2 w/2 h/2, (where π=3.14, w is the width, l is the length and h is the height), and the growth curves of the xenografts were plotted. Following inoculation of the tumor cells, all nude mice were sacrificed on the 25th day after the animal model had been successfully constructed. The tumor tissue was dissected, and the tumor size was measured. The transplanted tumors were subsequently stripped, fixed in 4% polyformaldehyde solution, paraffin-embedded, and then sectioned. The tumor tissue sections were stained with H&E prior to immunohistochemistry and analysis using ImageJ software (National Institutes of Health).

Statistical analysis

SPSS version 20.0 statistical software (IBM Corp., Armonk, NY, USA) was used to analyze the experimental data. The results are expressed as the mean ± standard deviation (SD). The expression of BMP9 or p-AKT using immunohistochemical staining in breast cancer specimens was analyzed by paired Student's t-test. The correlation between BMP9 and p-Akt was analyzed by the Spearman correlation coefficient method. The other data were analyzed by analysis of variance (ANOVA). Multiple comparison between the groups was performed using the S-N-K method. The expression levels of cyclin D1, cyclin B1, cyclin E1, c-Myc and MMP9 were analyzed using RT-qPCR and the 2−ΔΔCq method. P<0.05 was considered to indicate a statistically significant difference.

Results

Expression level of BMP9 is downregulated in breast cancer tissues, whereas the expression of p-Akt is upregulated

In the 23 clinical breast cancer specimens, immunohistochemical staining showed that the positive rate of BMP9 was 17.4% (4/23) in the tumor tissues, whereas the positive rate of BMP9 was 65.2% (15/23) in the paracancerous tissues. The expression level of BMP9 was significantly lower in the tumor tissues compared with the level noted in the paracancerous tissues (P<0.05). The positive rate of p-Akt was 82.6% (19/23) in the cancer tissues, whereas the positive rate of p-Akt was 47.8% (11/23) in the paracancerous tissues. The protein expression level of p-Akt was significantly higher (P<0.05) in cancer tissues than the level noted in the paracancerous tissues (Fig. 1A and B). Spearman correlation coefficient analysis revealed that the expression level of BMP9 was negatively correlated with the p-Akt protein expression level in tumor tissue (r=−0.422; P<0.05) (Fig. 1C). These results supported that the PI3K/Akt signaling pathway is involved in the development and progression of breast cancer, since the expression level of BMP9 was negatively correlated with the expression level of p-Akt.

Figure 1. Expression levels of BMP9, Akt, and p-Akt in breast cancer tissues and adjacent normal tissues as determined by immunohistochemistry. (A) Immunohistochemistry was used to detect the expression of BMP9, Akt, and p-Akt in breast cancer tissues and adjacent normal tissues (magnification, ×100). (B) Quantification of the positive rate of BMP9 and p-Akt in the breast cancer and adjacent non-tumor tissues. *P<0.05. (C) Correlation between p-Akt and BMP9 expression in all 23 human breast cancer tumor tissues. BMP9, bone morphogenetic protein 9; p-Akt, phosphorylated Akt.

BMP9 inhibits the PI3K/Akt signaling pathway and downregulates the expression levels of cyclin D1, cyclin B1, cyclin E1, c-Myc and MMP9 in MDA-MB-231 cells

To investigate whether BMP9 exerts an inhibitory effect on the PI3K/Akt signaling pathway, western blotting was performed. Western blotting revealed that the protein expression level of p-Akt in the BMP9 overexpression group (MDA-MB-231/BMP9) was significantly lower compared with the GFP group (MDA-MB-231/GFP) and the blank control group (MDA-MB-231) after the MDA-MB-231 cells had been infected with the adenovirus overexpressing BMP9 for 48 h. These results suggested that BMP9 inhibited the PI3K/Akt signaling pathway.

Cyclins D1, B1 and E1, c-Myc, and MMP9 have been reported to be involved in the proliferation and invasion of tumor cells (20–24). Therefore, RT-qPCR was used to estimate their expression levels in recombinant MDA-MB-231/BMP9 cells. The expression levels of cyclin D1, cyclin B1, cyclin E1, c-Myc and MMP9 were all shown to be downregulated in the MDA-MB-231/BMP9 cells (cyclin D1: 2−ΔΔCq=0.321; cyclin B1: 2−ΔΔCq=0.287; cyclin E1: 2−ΔΔCq=0.277; c-Myc: 2−ΔΔCq=0.586; MMP9: 2−ΔΔCq=0.130) (Fig. 2A). Western blotting was performed to confirm the expression levels of cyclin D1, cyclin B1, cyclin E1, c-Myc and MMP9 in the MDA-MB-231/BMP9 groups. The results demonstrated that these proteins were significantly lower in MDA-MB-231/BMP9 cells compared with these level in the MDA-MB-231/GFP and MDA-MB-231 groups (Fig. 2B). These results suggested that BMP9 inhibited the proliferation of MDA-MB-231 cells via downregulation of the expression levels of cyclin D1, cyclin B1, cyclin E1, c-Myc and MMP9.

Figure 2. Expression levels of cyclin D1, cyclin B1, cyclin E1, c-Myc and MMP9 as determined by (A) RT-qPCR and (B) western blotting. (A) The mRNA expression levels of cyclin D1, cyclin B1, cyclin E1, c-Myc and MMP9 for the MDA-MB-231/BMP9 group were significantly decreased compared with the expression levels in the MDA-MB-231/GFP and MDA-MB-231 groups (*P<0.05). (B) The protein expression level of p-Akt was significantly lower in the MDA-MB-231/BMP9 group compared with that noted in the MDA-MB-231/GFP and MDA-MB-231 groups p-Akt, phosphorylated Akt; MMP9, matrix metalloproteinase 9; RT-qPCR, real-time quantitative polymerase chain reaction.

Overexpression of BMP9 inhibits the growth of breast cancer tumors in a nude mouse xenograft tumor model

After the successful generation of the nude mouse xenograft tumor model, tumors were identified on the 10th day in the MDA-MB-231/GFP and the MDA-MB-231 groups, although tumors were found on the 16th day in the MDA-MB-231/BMP9 group. All the nude mice were sacrificed on the 25th day after the animal model had been successfully constructed, and the tumor tissue was dissected. The size of the tumors was measured using a Vernier caliper (the maximum tumour diameter: MDA-MB-231/BMP9 group was 1.08 cm; MDA-MB-231/GFP group was 1.48 cm; MDA-MB-231 group was 1.49 cm on the 25th day). The tumor volume in the MDA-MB-231/BMP9 group (0.32±0.05 cm3) was markedly lower when compared with that in the MDA-MB-231/GFP group (1.10±0.05 cm3) and MDA-MB-231 group (1.12±0.12 cm3) (P<0.001; Fig. 3A and B). At the same time, the tumors of the MDA-MB-231/GFP and MDA-MB-231 groups were found to be necrotic and ulcerated when the tumors were peeled off. Owing to the adhesion between the tumor and the skin of the nude mice, the tumors were not easy to peel off. By contrast, the tumor masses of the MDA-MB-231/BMP9 group were found to be hard, there was no obvious necrosis, and the tumors were easy to peel off. These results demonstrated that the overexpression of BMP9 may considerably inhibit the growth of breast cancer cells.

Figure 3. Effect of BMP9 on a nude mouse xenograft model. (A) The tumor volumes of xenografts formed by the MDA-MB-231/BMP9, MDA-MB-231/GFP and MDA-MB-231 breast cancer cells were determined. (B) Growth curves of xenografts formed by MDA-MB-231/BMP9, MDA-MB-231/GFP and MDA-MB-231 breast cancer cells were plotted. *P<0.001. BMP9, bone morphogenetic protein 9; GFP, green fluorescent protein.

BMP9 inhibits PI3K/Akt expression in the nude mouse xenograft tumor model

In the nude mouse xenograft tumor model, immunohistochemical staining revealed that p-Akt protein expression was obviously decreased in the MDA-MB-231/BMP9 group compared with that noted in the MDA-MB-231/GFP and MDA-MB-231 groups. The t-Akt protein expression exhibited no difference in the three groups. This result further confirmed that BMP9 inhibits the growth of breast cancer by inhibiting the PI3K/Akt signaling pathway in vivo.

Discussion

Previous studies have revealed that BMPs are able to promote bone formation, cell invasion and metastasis via the classical BMP/SMAD signaling pathway (12). Recent studies have reported the effect of BMPs on certain non-classical signal transduction pathways, including those involving PI3K/Akt and mitogen-activated protein kinases (MAPKs) (25–28). These non-classical BMP/SMAD signaling pathways have an important role in tumor growth, invasion and metastasis. Pancholi et al (29) demonstrated that the estrogen receptor (ER) could bind to both ERBB2 and PI3K, leading to the phosphorylation of Akt. Another study by Lin et al (30) confirmed that Akt is the main downstream target of PI3K, demonstrating that PI3K and Akt may work together to participate in tumorigenesis, development and metastasis. It has been reported that most BMPs can promote tumor growth, invasion and metastasis via the PI3K/Akt signaling pathway. Ren et al (13) showed that PI3K/Akt activation could promote the proliferation, migration and invasion of HER2-positive SK-BR-3 breast cancer cells, whereas Duan et al (31) showed that inhibition of the PI3K/Akt pathway could inhibit cell proliferation and migration in gastric cancer.

BMP9 was originally cloned from a fetal mouse liver cDNA. It has been shown as a pleiotropic cytokine involved in a number of physiologic events. These include hepatic reticuloendothelial system, bone morphogenesis, neuronal differentiation, hematopoiesis, angiogenesis, iron homeostasis, glucose homeostasis and cancer growth or metastasis. In our previous studies, HCT116 cells were infected with the BMP9 adenovirus to construct conditioned medium, and we found that BMP9 mRNA was readily detectable in the HCT116/BMP9 cells, and the conditioned medium indeed contained detectable levels of BMP9 protein (9). Our research confirm a previous study by Ye et al that BMP9 was released into the medium following transfection of the gene into cells (32).

BMP9, as a multifunctional cytokine, not only induces bone and cartilage formation but is also involved in the proliferation, differentiation, invasion and metastasis of tumor cells. A previous study by our research laboratory demonstrated that BMP9 could inhibit the proliferation of lung adenocarcinoma A549 cells via the SMAD-dependent BMP/SMAD signaling pathway (33). Our previous research also revealed that BMP9 can inhibit the proliferation, invasion, and migration of breast cancer cells via the activin receptor-like kinase-2 (ALK2)-dependent BMP/SMAD signaling pathway (19). This prompted us to consider whether BMP9 exerts an effect on certain non-classical signal transduction pathways, and therefore our research studies were aimed to elucidate the mechanism of BMP9 in breast cancer cells in order to provide a theoretical basis for the targeted therapy of breast cancer.

In the present study, 23 cases of patients with clinical breast cancer were first identified. Immunohistochemical staining showed that the positive rate of BMP9 expression in breast cancer was significantly lower when compared with that in adjacent normal tissues (P<0.05). In addition, the protein expression level of p-Akt was significantly reduced in the samples with BMP9 expression. Furthermore, the expression level of BMP9 in tumor tissue was negatively correlated with the expression level of p-Akt protein (r=−0.422; P<0.05) (Fig. 1C). These preliminary results suggest that BMP9 may participate in the growth, invasion, and metastasis of breast cancer cells via the PI3K/Akt signaling pathway. Subsequently, BMP9 was overexpressed in MDA-MB-231 breast cancer cells by adenovirus infection in vitro, revealing that the overexpression of BMP9 inhibited the protein expression level of p-Akt (Fig. 2B). Experiments performed in vivo further confirmed that BMP9 serves an inhibitory role in the growth, invasion, and migration of breast cancer cells, and that a clear association exists with the PI3K/Akt signaling pathway. In addition to acting via the classical BMP/SMAD signaling pathway, BMP9 also functions through the non-SMAD-dependent PI3K/Akt signaling pathway.

To further explore the carcinogenic mechanism of BMP9 in breast cancer, a nude mouse xenograft tumor model was constructed. The animal experiments demonstrated that BMP9 can inhibit breast cancer growth (Fig. 3A and B). After constructing the nude mouse model over a 28-day period, the nude mice were sacrificed and the tumor tissues were isolated and paraffin-embedded. Immunohistochemistry results showed that the positive percentage of p-Akt protein in the MDA-MB-231/BMP9 group was significantly lower compared with the MDA-MB-231 and MDA-MB-231/GFP groups (Fig. 4). These results suggested that BMP9 can indeed inhibit the growth of MDA-MB-231 cells by inhibiting the PI3K/Akt signaling pathway in vivo, thereby providing experimental data in support of the clinical application of BMP9 in the treatment of breast cancer.

Figure 4. Immunohistochemical staining results of xenograft tumors derived from the MDA-MB-231/BMP9, MDA-MB-231/GFP and MDA-MB-231 breast cancer cells. Hematoxylin and eosin (H&E) staining and Immunohistochemistry of the tumors in the different groups (magnification, ×100). BMP9, bone morphogenetic protein 9; GFP, green fluorescent protein; p-Akt, phosphorylated Akt.

As a multifunctional factor, the role of BMP9 in osteogenesis and breast cancer has been progressively determined. Accumulating evidence has shown that BMP9 may exert a role through multiple signaling pathways (34–36). The study has also, in part, explained why BMPs serve different roles in different tumors, and even for the same BMP in different tumors, its role may significantly differ. For example, BMP9 was found to have an inhibitory role in prostate cancer and breast cancer, whereas BMP9 had a role in promoting ovarian cancer (32,37). We demonstrated that BMP9 can inhibit the PI3K/Akt signaling pathway. Yet, we did not ascertain how BMP9 regulates the PI3K/Akt signaling pathway. We will knock down the BMP9 type I receptor (ALK1 or ALK2) to inhibit the BMP9-independent SMAD signaling pathways in order to determine the effect of BMP9 overexpression on breast tumor cell growth in future studies. The published data indicate that ALK1 or ALK2 belongs to BMP9 type I receptors and can bind with BMP9. Knocking down each of the BMP9 type I receptors can directly inhibit BMP9 signaling pathways. It will confirm whether BMP9 can directly target the PI3K/AKT signaling pathway or not (13).

The present study has shown that BMP9 can inhibit the PI3K/Akt signaling pathway consequently suppressing breast cancer growth. In addition, it has been shown by RT-qPCR and western blotting that BMP9 may inhibit the expression levels of cyclin D1, cyclin B1, cyclin E1, c-Myc and MMP9 (Fig. 2A and B), which subsequently leads to inhibition of the proliferation, invasion, and metastasis of breast cancer cells. These results may explain, in part, how the degree of malignancy of MDA-MB-231 cells was increased in our previous studies after the inhibitory effect of BMP9 on connective tissue growth factor (CTGF) had been restored in MDA-MB-231 cells by using an adenovirus co-infection technique in vitro (12), although certain differences existed compared with the blank control group. Since BMP9 can exert a role via the BMP/SMAD signaling pathway as well as through the non-classical PI3K/Akt signal transduction pathway, this may also explain why BMP9 considerably inhibited the growth of MDA-MB-231 cells.

In conclusion, the present study has shown that BMP9 expression was downregulated in breast cancer cells. This study has shown, to the best of our knowledge for the first time, that BMP9 inhibits the growth of breast cancer by inhibiting the PI3K/Akt signaling pathway. At the same time, BMP9 can also inhibit the proliferation, invasion, and metastasis of breast cancer by inhibiting the expression of cyclin D1, cyclin B1, cyclin E1, c-Myc, and MMP9. Therefore, BMP9-mediated PI3K/Akt signaling may serve as a novel target for potential breast cancer therapy to prevent proliferation, invasion, and metastasis of breast cancer cells.

Acknowledgements

Not applicable.

Funding

The present study was supported by the Natural Science Foundation Project of Yongchuan (grant nos. Ycstc.2016nc510 and Ycstc.2017nc5020), the Natural Science Foundation Project of Chongqing Education Committee (grant no. KJ1500202) and the Natural Science Foundation Project of Yongchuan Hospital (grant no. YJZQN201517).

Availability of data and material

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Authors' contributions

HD, SL, TZ, YW, CL and SP performed the research. KW and SL designed the research study. SL, YH and HD analyzed the data. SL wrote the paper. KW, as the corresponding author, modified the article and proposed its written content. All authors have read and approved the final version of the manuscript and agree to be accounTable for all aspects of the research in ensuring that the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Ethics approval and consent to participate

All the experiments concerning the tumor specimens were approved by the Ethics Association of Yongchuan Hospital of the Chongqing Medical University, and all the patients signed the informed consent form. All experiments were approved by the Institutional Animal Care and Use Committee of the Chongqing Medical University, as well as regional authorities [reference no. SYXK(YU)2014-0001] in accordance with the ‘Guidelines for the Welfare of Animals in Experimental Neoplasia’ (The China Coordinating Committee on Cancer Research).

Patient consent for publication

All the patients signed the informed consent form.

Competing interests

The authors state that they have no competing interests.

References