microRNA-124 inhibits bone metastasis of breast cancer by repressing Interleukin-11

Human breast cancer cell lines (BT-549, MDA-MB-231, Hs578T, MDA-MB-468, MDA-MB-436, MCF7, T47D and BT-474) as well as RAW264.7 and MC3T3-E1 were obtained from ATCC. The strongly bone-metastatic MDA-MB-231-derived subline, MDA-MB-231-luc-D3H2LN, was purchased from Xenogen Corporation (Alameda, CA, USA). Bone marrow-derived macrophages (BMMs) were isolated from C57/BL6 mice as described previously [30]. All of the cell lines were tested using MycoAlert Mycoplasma Detection Kit (Lonza, Switzerland) according to the manufacturer’s instructions and were free of mycoplasma contamination.

Primary breast cancer tissues and adjacent non-tumorous mammary tissues were obtained from patients who received breast cancer surgery and were histopathologically verified at Changhai hospital. Bone metastases from breast cancer were obtained from patients that received bone metastasis resection at Changzheng hospital (Shanghai, China). All of the subjects provided written informed consent. Ethical consent was granted by the Committees for Ethical Review of Research involving Human Subjects of Second Military Medical University (Shanghai, China).

Fluorescence ISH was used to detect miR-124 expression in tissue microarray slides containing 79 paired primary breast cancer tissues and adjacent non-tumorous mammary tissues as well as 34 bone metastases from breast cancer. ISH was performed using a miR-124 locked nucleic acid probe (5′-digoxigenin-GGCATTCACCGCGTGCCTTA-3′-digoxigenin) and the microRNA ISH Optimization Kit (Exiqon, Vedbaek, Denmark) according to the manufacturer’s instructions as described previously [31]. The signals were examined with a BX51 fluorescence microscope (Olympus) and quantified using the Aperio Spectrum® software with a pixel count algorithm.

To construct the lenti-miR-124 plasmid (pLenO-DCE-Puro-miR-124), cDNA encoding pri-miR-124 was appended with EcoRI and BamHI sites and cloned into pLenO-DCE-Puro Vector (Bio-link, Shanghai, China). The backbone plasmid expressing Green Fluorescent Protein (GFP) was used as a negative control (NC). Lentiviruses were produced by four-plasmid cotransfection of 293 T cells with the packaging helper plasmid pRSV-Rev., pMDLg/pRRE, pMD2.G and pLenO-DCE (Transfer Vector). The lentivirus carrying miR-124 inhibitor and its negative control (NC) were constructed by Obio Technology (Shanghai, China) according to the method described previously [32]. TuD-miR-124 fragment containing two miR-124 binding sequence or its control fragment was synthesized and cloned into the plasmid pLKD-CMV-DsRed2-U6-shRNA digested with AgeI and EcoRI.

To investigate the effect of miR-124 on breast cancer cell survival in the bone microenvironment, luciferase-labeled MDA-MB-231 cells stably expressing miR-124 or NC were transplanted into female Balb/c nude mice via the intratibia route. To investigate the effect of miR-124 on the bone metastasis of breast cancer cells in vivo, luciferase-labeled MDA-MB-231 cells stably expressing miR-124 or NC, as well as luciferase-labeled MCF7 cells stably expressing miR-124 inhibitor or NC were inoculated into the left ventricle of nude mice. To explore if miR-124 can prevent bone metastasis, the left ventricle of nude mice were inoculated with luciferase-labeled MDA-MB-231 cells and then treated with injections of 10 nmol ago-miR-124 (miR400004422, RiboBio, Guangzhou, China) or NC (miR04201) via tail vein. Bone metastases were analyzed by X-ray and histopathologically confirmed with H&E staining. To determine the role of IL-11 in the suppression of bone metastasis by miR-124, MCF7 cells stably expressing miR-124 inhibitor or NC were firstly inoculated into the left ventricle of nude mice, and then 5 μg IL-11 neutralizing antibody or the control IgG were injected into the tail veins of the mice twice a week for up to 3 weeks in both groups. All experimental mice were monitored with an ex vivo imaging system (IVIS200, Caliper LS, Hopkinton, MA, USA). All of the mouse experiments were performed according to protocols reviewed and approved by the Institutional Animal Care and Use Committee at the Second Military Medical University.

miR-124 mimic and negative control (NC) or miR-124 inhibitor and inhibitor NC were purchased from RiboBio (RibiBio, Guangzhou, China). MDA-MB-231 cells were transfected with miR-124 mimic and NC, and MCF7 cells were transfected with miR-124 inhibitor and inhibitor NC using Lipofectamine2000 (Invitrogen) according to the manufacturer’s instructions. Twenty-four hours after transfection, the media from MDA-MB-231 cells transfected with miR-124 mimic or NC and the media from MCF7 cells transfected with miR-124 inhibitor or inhibitor NC were collected as conditioned media.

BMMs (5 × 10³ cells/well) were seeded into 96-well plate and incubated with M-CSF (10 ng/ml) and the receptor activator of nuclear factor-κB ligand (RANKL) (50 ng/ml) in the presence of 30% conditioned media, which was replaced every 48 h. The proliferation of BMM cells was analyzed using Cell Counting Kit-8 (Dojinodo, Shanghai, China) according to the manufacturer’s protocol. After 5–7 days, BMM differentiation was analyzed by tartrate resistant acid phosphatase (TRAP) staining and actin-ring formation assays.

To construct the luciferase reporter plasmid encoding IL-11 3’untranslated regions (3′UTR), a 1562 bp fragment of the 3′UTR from human IL-11 was sub-cloned into the psiCHECK2 vector (Promega, Madison, WI, USA) using XhoI and NotI restriction sites. Mutation of the miR-124 binding site on the IL-11 3′UTR reporter vector was performed using overlap extension by PCR as described previously [33]. The luciferase reporter assays were performed with the Dual-Luciferase Reporter Assay (Promega, Madison, WI, USA) using a luminometer (Synergy™ 4 Hybrid Microplate Reader, BioTek, Winooski, VT, USA).

Human IL-11 neutralizing antibody (AF-218-NA), normal goat IgG control antibody (AB-108-C) and recombinant human IL-11 protein (218-IL-005) were purchased from R&D systems (Minneapolis, MN, USA).

Additional details of methods are described in the Additional file 1.

We examined miR-124 expression in normal human mammary tissue and a panel of human breast cancer cell lines with different invasive properties. miR-124 levels were significantly decreased relative to the control in all of the eight tested breast cancer cell lines (Fig. 1a). Furthermore, miR-124 expression was substantially lower in the invasive cell lines (BT-549, MDA-MB-231, Hs578T, MDA-MB-468 and MDA-MB-436) than in the non-invasive cell lines (MCF7, T47D and BT-474) (Fig. 1b). MDA-MB-231-luc-D3H2LN (MDA-MB-231-B) is a commonly used metastatic breast cancer cell line that has strong bone-metastatic characteristics. miR-124 expression was significantly down-regulated in MDA-MB-231-B compared to parental MDA-MB-231 cells (MDA-MB-231-P) (Fig. 1c). We also detected miR-124 expression in 79 paired primary breast cancer tissues and adjacent non-tumorous mammary tissues as well as 34 bone metastases from breast cancers using fluorescence ISH. miR-124 expression was substantially repressed in primary lesions compared to paired non-tumor tissues (median, 0.0313 and 0.05225, respectively; P < 0.001; Mann-Whitney’s U test) and was further reduced in metastatic bone tissues (median, 0.0173; P < 0.001; Mann-Whitney’s U test) (Fig. 1d, e).

To investigate the effect of miR-124 on survival of breast cancer cell in the bone microenvironment, luciferase-labeled MDA-MB-231 cells infected with lentivirus expressing miR-124 or control lentivirus (negative control, NC) (Additional file 1: Figure S1) were transplanted into Balb/c nude mice via the intratibia route. Luciferase signals detected in tibias using ex vivo imaging 4 weeks after injection were significantly lower in mice expressing miR-124 than in mice from control group (Fig. 2a, b). Moreover, X-ray analysis indicated that cancer cell-induced osteolysis was repressed in the miR-124 group (Fig. 2c). TRAP staining showed that both the number and activity of osteoclasts were markedly reduced at the boundary in the miR-124 group (Fig. 2d).

We further evaluated the effect of miR-124 on bone metastasis of breast cancer cells in vivo by inoculating luciferase-labeled MDA-MB-231 cells stably expressing miR-124 or NC, as well as luciferase-labeled MCF7 cells stably expressing miR-124 inhibitor or NC, into the left ventricles of Balb/c nude mice. In mice administered MDA-MB-231 cells stably expressing NC, 25% (2 of 8) showed bone metastasis after 1 week by ex vivo imaging, including metastases to the skull, spine and tibia, and all of the subjects exhibited bone metastases 3 weeks after inoculation, with one mice died in the third week (Additional file 1: Figure S2a, Figure S3a). However, no signal was observed until day 21 in mice injected with cells stably expressing miR-124, and significantly lower signals were identified in 37.5% (3 of 8) of mice 5 weeks after inoculation (Additional file 1: Figure S2a, Figure S3a, b). X-ray analysis indicated inhibition of cancer cell-induced osteolysis in the miR-124 group (Fig. 3a), and the following hematoxylin-eosin (H&E) staining confirmed bone lesions in the control group mice (Additional file 1: Figure S2b). Moreover, luciferase signals were significantly higher in mice administered MCF7 cells stably expressing miR-124 inhibitor than those in mice administered cells stably expressing NC (Fig. 3c, d). Metastases to the bones including ribs, spine and tibias, as detected by luciferase signals, were also significantly increased in mice administered MCF7 cells stably expressing miR-124 inhibitor than those in mice from NC group (Additional file 1: Figure S3a, b). In addition, metastases to the lungs, as detected by luciferase signals and H&E staining, were identified in 60% (3/5) of the mice of the miR-124 inhibitor group, while no lung metastasis was observed in the mice of the NC group (Additional file 1: Figure S3c, d). The luciferase signals in the hearts were emitted by residual cells after ventricle inoculation. To explore whether miR-124 might be applied to prevent bone metastasis, luciferase-labeled MDA-MB-231 cells were inoculated into the left ventricle of nude mice followed by injections of ago-miR-124 or NC via the tail vein. Luciferase signals were significantly lower in the ago-miR-124 group compared to the control group (Fig. 3e, f). Strikingly, only 50% (3 of 6) of mice in the agomiR-124 group developed tiny bone metastases, whereas all of the mice in the control group developed obvious bone metastases (Fig. 3e, f). X-ray analysis and H&E staining showed a further reduction of tumor-induced osteolysis in the ago-miR-124 group (Fig. 3e, Additional file 1: Figure S4).

The role of miR-124 in the metastasis of breast cancer cells to bone is largely unknown. The reduced osteolysis observed in mice injected with miR-124 expressing cells raises the possibility that cancer cell-derived miR-124 might inhibit osteoclastogenesis by regulating osteoclast and osteoblast activity. To verify this hypothesis, we selected BMMs with M-CSF stimulation as an in vitro osteoclast differentiation model. We first cultured BMMs, using conditioned media from MDA-MB-231 (with low expression of miR-124) transfected with miR-124 mimic (miR-124-CM) or NC (NC-CM) as well as conditioned media from MCF7 (with high expression of miR-124) transfected with miR-124 inhibitor (in-miR-124-CM) or NC inhibitor (NC-CM). Cultured with miR-124-CM significantly decreased, while cultured with in-miR-124-CM increased the number of BMMs (Fig. 4a, b). Second, after treatment with M-CSF and RANKL, BMMs cultured in miR-124-CM showed fewer cells differentiated into TRAP-positive multinucleated osteoclasts than BMMs cultured in NC-CM (Fig. 4c, d), and miR-124-CM also reduced the formation of actin-ring structures (Additional file 1: Figure S5). In contrast, in-miR-124-CM significantly enhanced the differentiation of BMMs as detected by TRAP staining (Fig. 4c, d). Cathepsin K, NFATc1, c-fos and TRAP are important transcription factors involved in osteoclastogenesis [34]. Cultured with miR-124-CM markedly promoted, whereas cultured with in-miR-124-CM suppressed the expression of these transcription factors in osteoclast cell line RAW264.7 (Fig. 4e, f). RANKL is the most prominent cytokine inducer of osteoclastogenesis [7], and RANKL activity is balanced in normal bone homeostasis by osteoprotegerin (OPG), a decoy RANKL receptor secreted by osteoblasts [7]. Therefore, in the bone microenvironment, tumor cell-induced osteoclastogenesis involves a reduction of the OPG/RANKL ratio in osteoblasts. In this study, we cultured osteoblast-like cells MC3T3-E1 with miR-124-CM and NC-CM or in-miR-124-CM and NC-CM, and real-time RT-PCR showed increased OPG and decreased RANKL expression as well as an increased OPG/RANKL ratio in MC3T3-E1 cultured with miR-124-CM (Fig. 4g). miR-124-CM also inhibited MMP13 expression, which is secreted from osteoblasts and stimulates osteoclast differentiation and activation [35], in MC3T3-E1 (Additional file 1: Figure S6). In contrast, in-miR-124-CM reduced the expression of OPG, enhanced the expression of RANKL, thus decreasing the ration of OPG/RANKL (Fig. 4h). We further collected the conditioned media from MC3T3-E1 cells cultured with in-miR-124-CM and NC-CM from MCF7 cells. The results showed that the conditioned media from MC3T3-E1 cells cultured with in-miR-124-CM significantly enhanced the differentiation of BMMs as detected by TRAP staining (Fig. 4i, j). In addition, the cancer cell-derived osteoclast-activating factors MCSF, IL-6, IL-8, RANKL, MMP2, MMP13, RUNX2 and PTHrP [6] are significantly down-regulated in MDA-MB-231 cells transfected with miR-124 mimic (Additional file 1: Figure S7a) and up-regulated in MCF7 cells transfected with miR-124 inhibitor (Additional file 1: Figure S7b). Collectively, these results suggest that miR-124 inhibits osteoclastogenesis both directly and indirectly by acting on osteoclasts.

To investigate the underlying molecular mechanisms by which miR-124 exerts its anti-bone-metastasis effect, we screened for putative miR-124 targets by performing an in silico complementarity search using TARGETSCAN-VERT (www.​targetscan.​org/), MICRORNA.​ORG (www.​microrna.​org/), MIRDB (http://​mirdb.​org/) and TARBASE (http://​diana.​imis.​athena-innovation.​gr/) on key regulators of osteoclastogenesis. These screening algorithms overlapped on IL-11, indicating that IL-11 is a potential downstream miR-124 target. Additional assays showed that IL-11 mRNA and protein levels were decreased in MDA-MB-231 by ectopic miR-124 expression and increased in MCF7 cells transfected with miR-124 inhibitor (Fig. 5a, b). Furthermore, IL-11 level was reduced in the tibia of mice injected with MDA-MB-231 cells stably expressing miR-124 (Fig. 5c). Western blotting indicated that IL-11 protein expression was substantially decreased in non-invasive cell lines (MCF7, T47D, BT-474) compared with invasive cell lines (BT-549, MDA-MB-231, Hs578T, MDA-MB-468 and MDA-MB-436) and was lowest in normal human breast tissue (Fig. 5d), which is the inverse of miR-124 expression (Fig. 1a). Indeed, an association study showed a negative correlation between miR-124 and IL-11 protein expression (Fig. 5e). Moreover, IL-11 expression was enhanced in MDA-MB-231-B compared to MDA-MB-231-P (Additional file 1: Figure S8), and a reporter assay revealed that miR-124 overexpression reduced luciferase activity from the wild-type (WT) IL-11 3′ untranslated region (UTR) by 57.4% (Fig. 5f). Point mutation of the target sequence in the IL-11 3′ UTR diminished the miR-124 effect, indicating that IL-11 is a direct downstream target of miR-124 (Fig. 5f).

To verify that IL-11 is not only a direct target but also a functional effector of miR-124 when regulating osteoclastogenesis, we performed a series of rescue assays using the commercial IL-11 neutralizing antibody and recombinant human IL-11 both in vitro and in vivo. First, BMMs treated with M-CSF and RANKL and cultured in conditioned medium from MDA-MB-231 cells transfected with miR-124 inhibitor (i-miR-124-CM) displayed higher proliferation and more differentiation to TRAP-positive multinucleated osteoclasts than those cultured in the control group, whereas IL-11 neutralizing antibody treatment efficiently reversed the effect of i-miR-124-CM on proliferation and differentiation of osteoclast progenitor (Fig. 6a-c). Consistently, BMMs cultured in conditioned medium from MDA-MB-231 cells transfected with miR-124 mimic (miR-124-CM) displayed lower proliferation and less differentiation to TRAP-positive multinucleated osteoclasts than those cultured in the control group, whereas recombinant IL-11 treatment substantially reversed these effects (Fig. 6d-f). Next, IL-11 neutralizing antibody reversed i-miR-124-CM mediated increase of Cathepsin K, NFATc1, c-fos and TRAP expression in RAW264.7 cells, while recombinant IL-11 reversed miR-124-CM mediated decrease of these genes levels in RAW264.7 cells (Additional file 1: Figure S9a, b). Thirdly, the suppressive effect of i-miR-124-CM on the ratio of OPG to RANKL was substantially inhibited by IL-11 neutralizing antibody treatment, while the promoting effect of miR-124-CM on the ratio of OPG to RANKL was partially reversed by recombinant IL-11 (Additional file 1: Figure S9c, d). Importantly, in vivo assay showed that IL-11 neutralizing antibody could partially reverse the promoting effect of miR-124 inhibitor on the bone metastasis and lung metastasis of breast cancer cells in mice (Fig. 6g, h, Additional file 1: Figure S10, Figure S11). Together, these findings demonstrate that IL-11 down-regulation contributes to the function of miR-124 in bone metastasis of breast cancer cells both in vitro and in vivo.

To further validate the clinical relevance of miR-124 in bone metastasis of breast cancer patients, we first performed a Kaplan-Meier analysis in patients who received breast cancer surgery and then a follow-up of the development of bone metastasis. The results showed lower miR-124 levels in primary breast cancer tissues were correlated with shorter bone metastasis-free survival (Fig. 7a). Consistently, correlation analysis revealed that miR-124 levels in metastatic bone tissues were positively correlated with the time from the primary breast cancer surgery to the occurrence of bone metastasis, and the progression time in patients with high miR-124 expression was longer than in patients with low miR-124 expression (Fig. 7b, Additional file 1: Figure S12), indicating that patients with lower miR-124 expression might progress to bone metastasis earlier. Of interest, clinicopathological analysis demonstrated that miR-124 down-regulation in human metastatic bone tissues was significantly correlated with aggressive clinicopathological characteristics, including poorer quality of life as evaluated by the Karnofsky performance score (KPS) (P = 0.0324), a larger number of metastatic bone lesions (P = 0.0366) and a higher proliferation potential as assessed by Ki67 staining (P = 0.0381) (Fig. 7c-e and Additional file 1: Table S1). Importantly, Kaplan-Meier analysis revealed that lower miR-124 levels in metastatic bone tissues were correlated with shorter overall survival of patients (Fig. 7f). As a functional downstream miR-124 target, IL-11 protein expression detected by IHC analysis was enhanced in the primary breast cancer compared to paired non-tumor tissues and was further elevated in metastatic bone tissues (Additional file 1: Figure S13) as opposed to the miR-124 expression in human tissues (Fig. 1d, e). Indeed, IL-11 expression was negatively correlated with miR-124 expression in paired primary breast cancer tissues and adjacent non-tumorous mammary tissues as well as bone metastases from the breast cancer (Fig. 7g). Furthermore, IL-11 expression was negatively correlated to the time from breast cancer surgery to bone metastasis development (Additional file 1: Figure S14a). In contrast to miR-124, the progression time in patients with high IL-11 expression was shorter than in those with low IL-11 level (Additional file 1: Figure S14b), indicating that patients with higher IL-11 expression might develop bone metastasis earlier. Clinicopathological analysis also indicated that IL-11 up-regulation in human metastatic bone tissues was significantly related to aggressive clinicopathological characteristics, including a larger number of metastatic bone lesions (P = 0.0366) and a higher proliferation potential as assessed by Ki67 staining (P = 0.0004) (Additional file 1: Figure S15, Table S2). Intriguingly, in contrast to the prognostic significance of miR-124, higher IL-11 expression in metastatic bone tissues was correlated with shorter overall survival for patients with bone metastasis (Fig. 7h). Overall, these findings suggest that dysregulation of the miR-124/IL-11 axis affects tumor-stromal interactions during the bone metastasis of breast cancer, thus exerting a negative influence on clinicopathological characteristics and the survival of breast cancer patients with bone metastasis (Fig. 7i).