Osteoprotegerin (OPG) mediates the anti-carcinogenic effects of normal breast fibroblasts and targets cancer stem cells through inhibition of the β-catenin pathway

Noura N. Alraouji a, Siti-Fauziah Hendrayani a, Hazem Ghebeh b, Falah H. Al-Mohanna c,
Abdelilah Aboussekhra a,*
a Department of Molecular Oncology, King Faisal Specialist Hospital and Research Center, Riyadh, 11211, Saudi Arabia
b Stem Cell & Tissue Re-Engineering Program, King Faisal Specialist Hospital and Research Centre, MBC#03, Riyadh, 11211, Saudi Arabia
c Department of Comparative Medicine, King Faisal Specialist Hospital and Research Center, Riyadh, 11211, Saudi Arabia
A R T I C L E I N F O

Keywords:
Breast cancer
Breast stromal fibroblasts OPG
Denosumab Cancer stem cells

A B S T R A C T

Normal breast fibroblasts (NBFs) support and maintain the architecture of the organ, and can also suppress tumorigenesis. However, the mechanisms involved are not fully understood. We have shown here that NBFs suppress breast carcinogenesis through secretion of osteoprotegerin (OPG), a soluble decoy receptor for the Receptor Activator of NF-κB ligand (RANKL). Indeed, NBFs and human recombinant OPG (rOPG), suppressed breast cancer cells proliferation and motility through inhibition of the epithelial-to-mesenchymal transition (EMT) process both in vitro and in vivo. Additionally, rOPG inhibited the IL-6/STAT3 and NF-κB pathways as well as the OPG gene, which turned out to be STAT3-regulated. This was confirmed using denosumab, a RANKL- targeted antibody, which also inhibited NF-κB, down-regulated OPG and repressed EMT in breast cancer cells grown in 2D and 3D. Importantly, both rOPG and denosumab targeted cancer stem cells (CSCs). This was mediated through inhibition of the CSC-related pathway β-catenin. Moreover, rOPG reduced tumor growth and inhibited breast CSC biomarkers in orthotopic humanized breast tumors. Therefore, normal mammary fibroblasts can suppress carcinogenesis through OPG, which constitutes great potential as preventive and/or therapeutic molecule for breast carcinomas.

Abbreviations: BCSCs, breast cancer stem cells; CAFs, cancer-associated fibroblasts; CSCs, cancer stem cells; EMT, epithelial-to-mesenchymal transition; ORF, open reading frame; Normal breast fibroblasts, NBFs; TCFs, tumor counterpart fibroblasts; TME, tumor-micro-environment.
* Corresponding author. Department of Molecular Oncology, Cancer Biology and EXperimental Therapeutics Section, King Faisal Specialist Hospital and Research Centre, MBC # 03, PO BOX 3354, Riyadh, 11211, Saudi Arabia.
E-mail address: [email protected] (A. Aboussekhra).

https://doi.org/10.1016/j.canlet.2021.08.013

Received 8 April 2021; Received in revised form 11 August 2021; Accepted 13 August 2021
Available online 17 August 2021
0304-3835/© 2021 Elsevier B.V. All rights reserved.

1. Introduction

Breast carcinomas are complex tumors that result mainly from the accumulation of genetic and epigenetic alterations in epithelial cells of the mammary gland [1–4]. However, like normal epithelial cells, tumor cells live in a complex microenvironment that includes the extracellular matriX (ECM) as well as cellular components such as immune and in- flammatory cells, endothelial cells, adipocytes, fibroblasts and bone marrow-derived cells [5,6]. Therefore, tumors consist of founder cancer cells and tumor-micro-environment (TME), which participate in all steps of carcinogenesis, including invasion and metastasis. This requires synergetic interplay between tumor cells and their microenvironment through paracrine secretions. A large amount of data indicates that cancer-associated fibroblasts (CAFs), which constitute the major component of the TME, actively participate in tumor cell proliferation and spread. CAFs are rich in active fibroblasts also called myofibroblasts [7,8].
While much of the available data have shown tumor-promoting ef- fects of CAFs on breast cancer cells, several other reports have indicated cancer-suppressive effects of normal breast fibroblasts. Indeed, the presence of normal mammary fibroblasts reverted primary breast cancer cells morphologically [9]. A different group found that fibroblasts from both normal breast tissue and breast cancer tissue suppressed prolifer- ation of MCF10A cells, but normal fibroblasts inhibited proliferation of the more transformed MCF10AT epithelial cells [10].
The soluble decoy receptor osteoprotegerin (OPG) protein is an osteoclastogenesis inhibitory factor, which belongs to the tumor ne- crosis factor (TNF) receptor superfamily, and is part of the RANKL/ RANK/OPG signaling system [11–13]. Apart from bone turnover, this system is also involved in mammary gland pathophysiology and other physiological processes, such as carcinogenesis, apoptosis and angio- genesis [11]. Several studies have shown the expression of OPG in breast cancer cell lines and tumors but not in normal breast tissue [14–16]. It has been also reported that inflammatory and invasive tumor stroma is rich in OPG [17]. Several cell culture and animal model studies have shown that, when expressed in tumor cells, OPG can promote breast tumor growth and spread [11,18,19]. Furthermore, there is clear link between OPG expression and breast cancer risk and/or prognosis. However, the nature of this correlation remains unclear owing to con- flicting reported data [20].
In the present study, we have shown that normal breast stromal fibroblast OPG as well as the pure full-length recombinant form of the protein can suppress breast carcinogenesis in a paracrine manner, through inhibition of the epithelial-to-mesenchymal transition and stemness both in vitro and in humanized orthotopic tumors.

2. Materials and methods

2.1. Cells, cell culture and reagents
Breast fibroblast cells were obtained and used as previously described [21]. They were cultured in DMEM/F12 medium supple- mented with 20% serum. MCF-7, MDA-MB-231, and BT-20 cell lines were purchased from ATCC. Cell lines were authenticated using short tandem repeat profiling by ATCC, propagated, expanded, and frozen immediately into numerous aliquots after arrival. Cells were regularly screened for mycoplasma contamination using MycoAlert Mycoplasma
Detection Kits (Lonza, Basel, Switzerland). All supplements were ob- tained from Gibco. Cells were maintained at 37 ◦C in humidified incu- bator with 5% CO2. Recombinant IL-6 was purchased from GenWay Biotech. Recombinant OPG was purchased from Sigma-Aldrich. OPG neutralizing antibody was purchased from R&D systems.

2.2. RNA purification and qRT-PCR
Total RNA, containing miRNAs, was purified using the miRNeasy mini kit (Qiagen) according to the manufacturer’s instructions and was treated with RNase-free DNase. RNA (1 μg) was used to synthesize complementary deoXyribonucleic acid (cDNA) utilizing either Advan- tage RT-PCR kit for mature mRNAs or miScript II RT kit for mature miRNAs as manufacturer’s instructions. Primer sequences are as shown in Supplementary Table 1.

2.3. siRNA transfection
The transfection with siRNA (sc-40152) and universal scrambled negative RNA (sc-37007) (Santa Cruz Biotechnology) was carried out using the RNAi-Fect reagent (Qiagen) as recommended by the manu- facturer. EXponentially growing cells (106) were transfected with 30 nM siRNA or scrambled RNA for 72 h, and then cells were harvested.

2.4. Transfections with plasmids
CTNNB1-ORF and the corresponding control (1 μg) (Origene) were utilized to transfect exponentially growing cells (5.105) using Lipofect- amine 3000 based on the manufacturer’s instructions (Invitrogen). Transfected cells were selected by neomycin sulfate (4 mg/ml).

2.5. Cellular lysate preparation and immunoblotting
This has been performed as previously described [22]. Antibodies directed against, Twist1 (10E4E6), IL-6, IL-6R, Snail (C15D3) were purchased from Abcam. ALDH1/2 (H-85), CD24 (C-20), and GAPDH (FL-335) were purchased from Santa Cruz Biotech (USA). E-cadherin (24E10), N-cadherin, AKT (C73H10), p-AKT (Thr308), JAK2 (D2E12), p-JAK2 (Tyr1007/1008), Nanog (D73G4), Oct-4 (C30A3), KLF4 (D1F2), SoX2 (D6D9), Bmi1 (D20B7), STAT3 (124H6), p-STAT3 (Tyr705), KLF-4 (D1F2), Lin-28B, NF-κB, Survivin, β catenin, α-tubulin and EpCAM (D1B3) Cell Signaling Technology (USA). CD44 was purchased from Sigma-Aldrich. ZEB1 was purchased from Abnova. ALDH1 was pur- chased from BD biosciences. All of these antibodies were used at 1:1000 dilution.

2.6. Cell proliferation, migration, and invasion assays
These assays were performed using the xCELLigence RTCA technol- ogy (Roche, Germany) as per the manufacturer’s instructions and as previously described [23]. In brief, 2X104 cells in serum-free medium were added to the upper wells of the CIM-plate coated with a thin layer of Matrigel (BD Biosciences) basement membrane matriX diluted 1:20 in serum-free medium (invasion) or non-coated (migration). Complete medium was used as a chemo-attractant in the lower chambers. For the proliferation assay, exponentially growing cells (2X104) were seeded in E-plate with a complete medium as per the manufacturer’s instruction. The obtained results were analyzed by the RTCA software. These ex- periments were performed in triplicate and were repeated several times.

2.7. Flow cytometry
Cells (2.105) were treated and then stained with CD44/CD24 as previously described [24]. Briefly, cells were washed and incubated with CD44 APC-Cy7/CD24 Pacific Blue antibodies (both from Bio- legend, USA) for surface staining (30 min at 4 ◦C). Data were acquired using the LSR II flow cytometer and the BD FACSDiva operating soft- ware. Positive staining was considered based on the negativity of an isotype control. A minimum of 10 000 events was recorded for all samples.

2.8. Cell cycle analysis by flow cytometry
Cells (5.104) were harvested, centrifuged and stained with propi- dium iodide (PI), and then were analyzed by flow cytometry using the BD FACSCalibur™ flow cytometer (San Jose, CA, USA), and the Cell- Quest™ Pro operating software (San Jose, CA, USA).

2.9. Soft agar colony formation assay
Cells were harvested and washed with serum-free medium and 4X104 cells were suspended in 4 ml of defined medium with 0.3% agarose (A9045, Sigma, MD, USA). The miXture of agarose and cells was plated in a 6-well plate containing a base layer of agarose of 0.5% (v/v). Cultures were incubated at 37 ◦C in a humidified incubator. Plates were examined with an inverted microscope to confirm only single cells were plated without any clumps. The spent medium was changed with fresh medium every 3 days and cultures were continued for 21 days. Colonies with a diameter of ≥100 μm were counted 1 week after initial plating.

2.10. 3D spheroid assay
Cells were seeded in 96 well ultra-low attachment plate at a density of 1000 viable cells/well. Cells were cultured in 171 medium supple-
mented with 1% ABM, 2% B-27, 20 ng/mL EGF, 500 ng/ml HC, 4% FBS and 5 μg/ml insulin. Cells were incubated for 10 days at 37 ◦C under 5% CO2. Mammospheres with a diameter of 100 μm were counted using OPTIKA light microscope.
Fig. 1. Normal mammary fibroblasts inhibit the tumorigenic potential of breast cancer cells.
Serum-free conditioned media (SFCM) were collected from the indicated cells, and then were applied on MDA-MB-231 cells for 24 h. SFM was used as a negative control. A and C, Whole cell lysates were prepared and were used for immunoblotting analysis using specific antibodies against the indicated proteins. GAPDH and β-actin were used as internal controls. The numbers bellow the bands represent fold change relative to the control (SFM) after correction against the internal controls. The levels of phosphorylated proteins were normalized against the total amount of their relative non-phosphorylated forms. B, Cell proliferation, migration, and invasion abilities of MDA-MB-231 cells pre-treated with the indicated media were assessed using the RTCA-DP XCELLigence system. Data are representative of different experiments performed in triplicate. *: p ≤ 0.05; **: p ≤ 0.01, as compared to SFM. D, Total RNA was prepared from MDA-MB-231 cells cultured in the indicated media, and then the Let-7b-5p and miR-21–5p expression levels were assessed using qRT-PCR. Error bars represent mean ± SD, (n = 3). **: p ≤ 0.01. E, Orthotropic breast cancer xenografts were created by injecting MDA-MB-231 cells either alone or in combination with NBF-6, TCF-180, or CAF-180 cells (n = 5 each)
under the nipple of nude mice. Histogram depicting tumor volume for each mouse. F, Graph showing tumor volumes grouped into boXes. Error bars represent mean ± SD. G, Representative photographs showing the size of the formed tumors in each group.

2.11. Three-dimensional cell culture
MDA-MB-231 cells (3X106) were seeded on Biotek 3D Insert scaffolds in 6-well plate according to the manufacturer’s instructions. Briefly, cells were re-suspended in 300 μl medium and carefully placed onto the center of the scaffold surface. After 3 h, 1700 μl medium was added. Cells were cultured in the scaffolds for another 3 days to allow the for- mation of 3D structures. The medium was changed and cells were treated for 24 h. Cells were harvested from the scaffolds, and total RNA was purified and used for qRT-PCR.

2.12. Human cytokine antibody array
Conditioned media was applied to human cytokine antibody array C5 (RayBiotech) as recommended by the manufacturer.

2.13. Orthotopic tumor xenografts
Animal experiments were approved by the KFSH&RC Institutional Animal Care and Use Committee (ACUC) and were conducted according to relevant national and international guidelines.

2.13.1. Co-injection
MDA-MB-231 cells (2.106) either alone or in combination with NBF- 6 (2.106), TCF-180, or CAF-180 (2.106) cells (n = 5 each) were injected under the nipple of nude mice (4–5 weeks) for 4 weeks.

2.13.2. Treatment with rOPG
MDA-MB-231 cells (2X106 n = 15) were injected under the right nipples of female nude mice (4–5 weeks). When the average tumor volume reached 100 mm3, mice were randomized into the following treatment conditions: PBS (control), 100 μg/kg OPG, and 300 μg/kg OPG (n 5 each) administered daily via intraperitoneal injection. Tumors were surgically retrieved after 2 weeks of treatment, were weighted and then were snap-frozen in liquid nitrogen.

2.14. Serum-free conditioned media
Cells were cultured in a serum-free medium (SFM) for 24 h, and then media were collected and briefly centrifuged. The resulting SFCM su- pernatants were used either immediately or were frozen at 80 ◦C until needed.

2.15. Quantification of protein expression level
The protein signal intensity of each band was determined using ImageQuant TL software (GE Healthcare). Next, dividing the obtained value of each band by the value of the corresponding internal control allowed a correction of the loading differences. The fold change in the protein levels was determined by dividing the corrected values by that of the control.
Fig. 2. OPG neutralization inhibits the anti-cancer effects of normal mammary fibroblasts.
A, SFM as well as SFCM from the indicated cells were applied onto human cytokine antibody array membranes (C7). B, SFCM from the indicated cells were used to assess the level of OPG by ELISA, (n = 3). Error bars represent mean ± SD. C, EXponentially-growing NBF-6 cells were transfected with OPG siRNA, while a scrambled sequence was used as control. Whole cell lysates were prepared and were used for immunoblotting analysis. D, Whole cell lysates were prepared from MDA-MB-231 cells treated with the indicated SFCM for 24 h, and then were used for immunoblotting analysis. E, IgG (0.6 μg/ml) or OPG neutralizing antibody (nOPG) (0.3 and 0.6 μg/ml) were added to NBF6-SFCM, and then the miXtures were applied on MCF-7 cells for 24 h. SFM was used as a negative control. Whole cell lysates were prepared and were utilized for immunoblotting analysis. The numbers bellow the bands represent fold change relative to the respective controls (Control-siRNA, SFM) after correction against the internal controls. The levels of phosphorylated proteins were normalized against the total amount of their relative non- phosphorylated forms.

2.16. Statistical analysis
Statistical analysis was performed using Microsoft EXcel and SPSS softwires. The results are presented as mean SED. of independent experiments. Statistical significance was determined using a two-tailed Students t-test and P values of 0.05 and less were considered as statistically significant.

3. Results

3.1. Normal mammary fibroblasts inhibit the carcinogenic potential of breast cancer cells
We started the present study by investigating the effect of normal breast fibroblasts (NBFs) on breast cancer (BC) cells. To this end, serum- free conditioned media (SFCM) were collected from 4 different NBFs: NBF-1, NBF-6, NBF-10, and NBF-11, and then were used to treat the highly aggressive BC cells MDA-MB-231 for 24 h. Serum-free medium (SFM), SFCM from CAF cells CAF-900/CAF-180 as well as their corre- sponding tumor counterpart fibroblasts TCFs (present in histologically normal breast tissues) were used as controls. Whole cell lysates were prepared and were used for immunoblotting analysis utilizing β-actin and GAPDH as internal controls. Fig. 1A and Supplementary Fig. S1A show that SFCM from the 4 NBF cells inhibited the epithelial-to- mesenchymal transition (EMT) process, via inhibition of the mesen- chymal markers (N-cadherin, Snail, Zeb1 and Twist1), and up- regulation of the epithelial markers (E-cadherin and EpCAM) as compared to controls. This effect was confirmed by showing that the SFCM from NBF-6 (NBF6-SFCM) and NBF-1 (NBF1-SFCM) inhibited the migration/invasion and proliferation capacities of MDA-MB-231 cells, relative to SFM, while SFCM from CAF-180 cells (CAF180-SFCM) enhanced these capacities (Fig. 1B). Similar anti-EMT effects were ob- tained on BT-20 cells (Supplementary Figs. S2A–B). Furthermore, while SFCM from CAF cells promoted stemness through up-regulation of ALDH1, SOX-2, NANOG and BMI-1 compared to SFM, SFCM from NBF cells down-regulated these markers relative to SFM (Fig. 1A and Sup- plementary Fig. S1A). Similar results were obtained for ALDH1 and BMI1 in BT-20 cells (Supplementary Fig. S2A). This indicates that NBFs can also inhibit the stemness characteristics of BC cells.
Next, we decided to study the effect of NBF6-SFCM on the IL-6/ STAT3 epigenetic feedback loop, which is responsible for maintaining EMT and stemness in cancer cells [25]. Fig. 1C and D and Supplementary Fig. S1B show clear NBF6-SFCM-dependent inhibition of STAT3, NF-κB, AKT, Lin-28B, IL-6 and miR-21–5p, while PTEN and let-7b-5p were up-regulated compared to SFCM. Similar results were obtained on BT-20 cells (Supplementary Fig. S2C). Together, these findings indicate that SFCM from NBFs suppresses the pro-carcinogenic/metastatic pro- cesses EMT and stemness in TNBC cell through inhibition of the IL-6/STAT3 epigenetic feedback loop.
To confirm these findings in vivo, MDA-MB-231 cells were injected under the nipple of nude mice either alone (2.106), or in combination with NBF-6 (2.106), CAF-180 (2.106) or TCF-180 (2.106) cells (n 5 for each inoculation). Inoculation of MDA-MB-231 cells alone generated tumors in 4 out of 5 animals with sizes ranging from 75 to 150 mm3 (Fig. 1E and F). However, tumors were grown in 3 and 5 animals inoculated with MDA-MB-231 and TCF-180 or MDA-MB-231 and CAF-180, respectively (Fig. 1E). As expected, tumors bearing CAF cells were much bigger (Fig. 1E, F and G). Interestingly, the presence of NBF-6 cells suppressed tumor formation and only a very small tumor was formed, as compared to the inoculation of MDA-MB-231 cells alone (Fig. 1E, F and G). This indicates tumor growth inhibitory potential of normal breast fibroblasts in vivo as well.
Fig. 3. rOPG suppresses the pro-metastatic EMT process in breast cancer cells. A, The indicated breast cancer cells were cultured in SFM either alone or containing recombinant OPG (rOPG) (100, 500 ng/ml) for 24 h. NBF6-SFCM was used as a positive control. Whole cell lysates were prepared and were used for immunoblotting analysis. The numbers bellow the bands represent fold change relative to the control (0) after correction against the internal controls. The levels of phosphorylated proteins were normalized against the total amount of their relative non- phosphorylated forms. B, MDA-MB-231 cells were cultured as 3D on 3D insert scaffolds in 6-well plate, and then were either sham-treated (0) or challenged with rOPG (500 ng/ml) for 24 h. Cells were harvested, total RNA was prepared and amplified by qRT-PCR. Error bars represent mean ± SD, (n = 3). *: p ≤ 0.05. C, Cell proliferation, migration, and invasion abilities of the indicated cells pre-treated with rOPG (100 ng/ml) were assessed using the RTCA-DP XCELLigence system. Data are representative of different experiments performed in triplicate. *: p ≤ 0.05; **: p ≤ 0.01. D, EXponentially growing MDA-MB-231 cells were either sham- treated (0) or challenged with rOPG (100 ng/ml), and then were harvested at the indicated periods of time. Cells were stained with PI, and the DNA contents were analyzed by flow cytometry. The proportion of cells in the different cell cycle phases are indicated.

3.2. Normal breast fibroblasts suppress breast carcinogenesis in an OPG- dependent manner
To determine the factor(s) responsible for NBF-dependent suppres- sion of breast carcinogenesis, the RayBiotech human cytokine array C-7 was exposed to SFM, NBF1-SFCM, NBF6-SFCM as well as CAF180-SFCM and CAF64-SFCM. Fig. 2A shows that the osteoprotegerin (OPG) protein is present in SFCM from both NBFs and CAFs with relatively similar levels. The level of secreted OPG from these cells was confirmed by the enzyme-linked immunosorbent assay (ELISA) (Fig. 2B).
To investigate the implication of OPG in NBF-related inhibition of breast carcinogenesis, we had to first knockdown this gene in NBF-6 cells, and then we tested their paracrine effect on MDA-MB-231 cells. Fig. 2C shows OPG knock-down using specific siRNA, compared to control cells (exposed to a scrambled sequence). Next, SFCM was pre- pared from these cells and used to treat BC cells. Fig. 2D shows that OPG knock-down suppressed NBF-6-related inhibition of Zeb1 and SoX2 in MDA-MB-231 cells as compared to control cells. A similar effect was observed on the IL-6/STAT3 feedback loop components (Fig. 2D, and Supplementary Fig. S3A). These results imply that OPG is responsible, at least partially, for the paracrine inhibitory effect of NBFs on BC cells. To confirm this, OPG was neutralized in NBF6-SFCM using a specific neutralizing anti-OPG antibody (0.3 μg/mL), while IgG was used as the negative control. Fig. 2E and Supplementary Fig. S3B show that treat- ment of MCF-7 BC cells with NBF6-SFCM reduced the expression levels of IL-6, ALDH1 and Snail as compared to cells exposed to SFM.
Interestingly, specific neutralization of OPG suppressed the inhibitory effect of NBF-SFCM on these genes (Fig. 2E, and Supplementary Fig. S3B). This confirms the role of OPG in the paracrine inhibitory effect of NBF cells on different BC cells.

3.3. IL-6 suppresses the paracrine anti-breast carcinogenesis effect of NBF-secreted OPG
These results prompted us to ask how OPG suppresses breast carci- nogenesis when present in NBF-SFCM but not when it’s secreted from CAF cells? One possibility is the increase in the level of several pro- carcinogenic cytokines such as IL-6 in CAF cells, which may mask the paracrine effect of secreted OPG. To test this possibility, MCF-7 cells, which express and secrete lower level of IL-6 [26], were treated with NBF6-SFCM containing or not the pure recombinant IL-6 protein (2 and 3.5 ng/mL), while SFM was utilized as a negative control. Supplemen- tary Fig. S4 shows that while NBF-6-SFCM reduced the levels of IL-6, ALDH1 and Snail relative to controls, the addition of the IL-6 protein suppressed the paracrine inhibitory effect of NBF6-SFCM in a dose-dependent manner. This indicates that an increase in the level of IL-6 suppresses the paracrine anti-breast carcinogenesis effect of NBF-secreted OPG.
Fig. 4. rOPG inhibits the IL-6/STAT3 and NF-κB pathways and downregulates OPG.
A, The indicated breast cancer cells were treated as in Fig. 3A. Whole cell lysates were prepared and were used for immunoblotting analysis. B, EXponentially growing cells were transfected with STAT3 siRNA, while a scrambled sequence was used as a control (Ctrl). Whole cell lysates were prepared and were used for immuno- blotting analysis. The numbers bellow the bands represent fold change relative to the respective controls (0, Ctrl) after correction against the internal controls. The
levels of phosphorylated proteins were normalized against the total amount of their relative non-phosphorylated forms. C, Cells were either sham-treated (0) or challenged with rOPG (100 ng/ml) for 24 h. Cells were harvested, total RNA was prepared and amplified by qRT-PCR. Error bars represent mean ± SD, (n = 3). *: p ≤0.05. D, MDA-MB-231 cells were cultured as 3D and were treated as in Fig. 3B. Total RNA was prepared and the mRNA levels of the indicated genes were assessed using qRT-PCR. Error bars represent mean ± SD, (n = 3). *: p ≤ 0.05.

3.4. Pure recombinant OPG suppresses the carcinogenic traits of breast cancer cells
To confirm the role of OPG in the inhibition of breast carcinogenesis, we treated 3 different BC cells (MDA-MB-231, BT-20 and MCF-7) with pure recombinant OPG (rOPG) that was added in the SFM. rOPG at 0.1 μg/mL potently inhibited the IL-6/STAT3 positive feedback loop in all cell lines (Fig. 3A, and Supplementary Fig. S5). Indeed, like NBF6-SFCM, rOPG upregulated PTEN and inhibited Lin28B as well as AKT phos- phorylation (Fig. 3A). Additionally, rOPG suppressed the mesenchymal markers (N-cadherin, Snail, Zeb1 and Twist1), while it upregulated the epithelial markers (E-cadherin and EpCAM) in all cells, relative to cells treated with SFM (Fig. 3A, and Supplementary Fig. S5). This effect on the EMT markers was confirmed at the mRNA level on the E-cadherin and N-cadherin coding genes (CDH1 and CDH2, respectively) when MDA-MB-231 cells were grown in 3D cultures (Fig. 3B). This potential rOPG-dependent EMT inhibition was confirmed by showing potent inhibitory effect of rOPG (0.1 μg/mL) on cellular invasion/migration as well as the proliferation of the 3 different cell lines (Fig. 3C).
To delineate the cell cycle phase that was delayed in response to rOPG, we performed a time-dependent experiment on MDA-MB-231 cells either treated or not with rOPG (0.1 μg/mL), and then cell cycle was analyzed by flow cytometry. Fig. 3D shows a clear rOPG-dependent delay in the S phase of the cell cycle. After 36 h of treatment, 42.5% of cells were in the S phase while only 11.5% of control cells were at this phase (Fig. 3D). This indicates that rOPG delays cell proliferation at the S phase of the cell cycle.

3.5. Recombinant OPG inhibits the NF-κB and the IL-6/STAT3 pathways and down-regulates OPG in breast cancer cells
Since SFCM from NBFs inhibited the IL-6/STAT3 pathway, we sought to investigate the effect of rOPG on this important BC-related pathway. Fig. 4A shows that, like NBF6-SFCM, rOPG reduced the expression of IL- 6 and its receptor (CD126) in MDA-MB-231, BT-20 and MCF-7 cells. This down-regulation was accompanied by inhibition of both JAK2 and STAT3 in the 3 cell lines (Fig. 4A, and Supplementary Fig. S6A). This clearly shows rOPG-dependent inhibition of the IL-6/STAT3 signaling pathway in BC cells. Intriguingly, the level of OPG was strongly reduced in rOPG-treated cells relative to controls (Fig. 4A, and Supplementary Fig. S6A). This indicates that rOPG represses the expression of the pro- tein in breast cancer cells, may be through inhibition of the JAK2/STAT3 pathway. To test this possibility, we investigated the effect of STAT3 down-regulation using specific siRNA on the expression of OPG. Fig. 4B, and Supplementary Fig. S6B shows that STAT3 down-regulation inhibited JAK2 and potently reduced the expression of IL-6R and OPG in MDA-MB-231 cells. Since STAT3 is a transcription factor, we checked these effects of rOPG at the mRNA level, and we have shown that rOPG reduced the level of the OPG mRNA in MDA-MB-231 and BT-20 cells (Fig. 4C). This indicates that OPG is under the control of STAT3.
These results prompted us to ask how rOPG reduces the intracellular level of the protein through inhibition of the IL-6/STAT3 pathway? OPG is a decoy receptor for Receptor Activator of NF-κB ligand (RANKL), which prevents the interaction of RANKL with its receptor RANK, and the consequent inhibition of several cancer-related pathways, including
Fig. 5. Denosumab inhibits EMT and the stemness features in breast cancer cells.
MDA-MB-231 and BT-20 cells were treated either with IgG or with denosumab (DNS) (30 and 50 μg/ml, respectively) for 24 h. A, Whole cell lysates were prepared and were used for immunoblotting. The numbers bellow the bands represent fold change relative to the control (IgG) after correction against the internal controls.
The levels of phosphorylated proteins were normalized against the total amount of their relative non-phosphorylated forms. B, Total RNA was prepared and the mRNA levels of the indicated genes were assessed using qRT-PCR. Error bars represent mean ± SD, (n = 3). *: p ≤ 0.05; **: p ≤ 0.01. C, MDA-MB-231 cells were cultured as 3D on 3D insert scaffolds in a 6-well plate, and then were treated with IgG or were challenged with DNS (50 ng/ml) for 24 h. Cells were harvested, total
RNA was prepared and amplified by qRT-PCR. Error bars represent mean ± SD, (n = 3). *: p ≤ 0.05; **: p ≤ 0.01. D, Cell proliferation ability was assessed using the RTCA-DP XCELLigence system. Data are representative of different experiments performed in triplicate. *: p ≤ 0.05. NF-κB [11,16]. Therefore, we tested the effect of rOPG on the Id2 and NF-κB (2 downstream targets of RANKL [11]). Fig. 4A and Supple- mentary Fig. S6A show that rOPG down-regulated Id2 and inhibited NF-κB through downregulation of IKKα in the 3 cell lines. These inhib- itory effects were confirmed by showing that rOPG significantly down- regulated the major downstream effector cyclin D1 (Fig. 1A, and Supplementary Fig. S6A), as well as IL-6 and IL-8 at the mRNA level (Fig. 4C). rOPG-related down-regulation of OPG, IL-6 and IL-8 was confirmed when BC cells were treated in 3D cultures (Fig. 4D).

3.6. Denosumab inhibits NF-κB, EMT and down-regulates OPG in breast cancer cells
Like rOPG, denosumab is also an inhibitor of the interaction of RANKL with its receptor RANK, which activates the NF-kB pathway [11]. Thereby, we sought to test the effect of RANKL inhibition by denosumab on NF-κB, OPG, and breast cancer cells. To this end, MDA-MB-231 and BT-20 cells were either challenged with denosumab (30 μg/mL) or with IgG, and then the expression of various genes was assessed by immunoblotting and qRT-PCR. Like rOPG, denosumab inhibited NF-κB and down-regulated its major downstream targets IL-6 and OPG at both the protein and mRNA levels (Fig. 5A and B, and Supplementary Fig. S7), as well as when BC cells were grown in 3D (Fig. 5C). Denosumab downregulated N-cadherin and upregulated E-cadherin (Fig. 5A, and Supplementary Fig. S7). The same effect was also observed at the mRNA level when BC cells were grown in 3D (Fig. 5C). In addition, denosumab inhibited the proliferative capacity of BC cells as compared to controls (Fig. 5D). This indicates that, like rOPG, denosumab can also inhibit EMT in BC cells and down-regulates OPG.

3.7. Recombinant OPG targets breast cancer stem cells through inhibition of the WNT/β-catenin pathway
Next, we investigated the effect of OPG on the stemness features of breast cancer cells. Fig. 6A, and Supplementary Fig. S8A show clear downregulation of CD44 and ALDH1, while CD24 was upregulated in response to rOPG. Concomitantly, rOPG treatment reduced the expres- sion of other important stemness markers, namely SoX2, Nanog and
BMI1 (Fig. 6A, and Supplementary Fig. S8A). Moreover, the proportion of CD44high/CD24low and ALDHhigh subpopulations in both MDA-MB- 231 and BT-20 cells decreased in rOPG-treated cells as compared to their respective controls (Fig. 6B). Interestingly, control cells demon- strated an increase in both size (growth rate) and frequency of spheroids compared to their respective rOPG-treated cells (Fig. 6C). Similarly, while OPG-treated cells did not form colonies in soft agar, control cells formed big colonies (Fig. 6D). These results indicate that rOPG inhibits stemness features and self-renewal capacity of BC cells and suppresses the formation of breast cancer stem cells (BCSCs).
To delineate the molecular pathway that mediates rOPG-related in- hibition of stemness, we have first tested the effect of rOPG on the important CSC-related pathway Wnt/β-catenin. Fig. 6A and Supple- mentary Fig. S8A shows rOPG-dependent strong down-regulation of β-catenin in both MDA-MB-231 and BT-20 cells. Since β-catenin is a nuclear protein, we tested the effect of rOPG on the nuclear level of the protein. Fig. 6E and Supplementary Fig. S8B show clear down-regulation of β-catenin in the nuclear fraction when cells were exposed to rOPG at both 0.1 and 0.5 μg/mL. A similar effect was observed on Twist1 (Fig. 6E, and Supplementary Fig. S8B). To confirm the effect of rOPG on the β-catenin pathway, we assessed the level of several β-catenin downstream targets at the mRNA level. Fig. 6F shows that all the tested genes (AXIN2, c-MYC, TWIST1 and CD44) were significantly down- regulated following exposure of BC cells to rOPG.
To confirm the implication of the β-catenin pathway, we introduced
Fig. 6. rOPG targets breast cancer stem cells by suppressing the β-catenin pathway.
A, Cells were treated as in Fig. 3A. Whole cell lysates were prepared and were used for immunoblotting analysis. The numbers bellow the bands represent fold change relative to the control (SFM) after correction against the internal controls. B, Cells were either sham-treated or challenged with rOPG (100 ng/ml) for 24 h. Top panel, cells were labeled with Aldeflour with or without the ALDH inhibitor DEAB. Bottom panel, cells were double-stained for CD24 and CD44. The proportions of ALDHhigh and CD44high/CD24low sub-populations were determined by flow cytometry, and are indicated by in the boXes as mean ± SED.; n = 3. C, Cells (1.5X103) were seeded in an ultra-low attachment 96-well plate containing the stem cell-specific medium. Tumorspheres (≥100 μm) were counted. Representative photographs of tumorspheres (top panel). Scale bar = 100 μm. Histograms depicting the number of the formed tumorspheres (bottom panel). Error bars represent (mean ± SD.; n= 3). *: p ≤ 0.05. D, Colony-forming capability of the indicated cells was assessed by seeding cells in soft agar. Representative photographs depicting the formed colonies. Scale bar = 100 μm. E, Cytoplasmic (C) and nuclear (N) lysates were prepared from the indicated cells treated as shown, and then were utilized for immunoblotting analysis. The numbers bellow the bands represent fold change relative to the control (0) for both cytoplasmic (C) and nuclear extracts (N), after correction against the internal controls α-tubulin and GAPDH. F, Total RNA was prepared from the indicated cells and was amplified by qRT-PCR. Error bars represent mean ± SD, (n = 3). *: p ≤ 0.05; **: p ≤ 0.01. G, Cells were transfected with a control vector (CT-ORF) or a vector bearing the CTNNB1-ORF (ORF), and then CT-ORF and ORF cells were either sham-treated or challenged with rOPG (100 ng/mL) for 24 h. Whole cell lysates were prepared and were used for immu- noblotting analysis. The numbers bellow the bands represent fold change relative to the control (CT-ORF) after correction against the internal control GAPDH. the CTNNB1-ORF or an empty vector into MDA-MB-231 cells (CTNNB1- ORF and CT-ORF, respectively). As expected, the level of the β-catenin protein was increased in ORF-expressing cells as compared to control cells (Fig. 6G, and Supplementary Fig. S8C). In addition, β-catenin up- regulation was accompanied by a decrease in the expression level of CD24 and an increase in the expression levels of ALDH1, CD44 and several β-catenin target genes (Fig. 6G, and Supplementary Fig. S8C). This up-regulation of β-catenin in BC cells abolished the ability of rOPG to inhibit these stemness-related markers (Fig. 6G, and Supplementary Fig. S8C). These results show that OPG suppresses the stemness char- acteristics of BC cells through inhibition of the Wnt/β-catenin pathway.

3.8. Denosumab reduces the proportion of cancer stem cells in BC cells
To further show that rOPG and denosumab have similar effects on TNBC cells, we investigated the effect of denosumab on the pool of BCSC
in both MDA-MB-231 and BT-20 cells. Fig. 7A shows that the proportion of CD44high/CD24low subpopulation decreased in denosumab-treated cells relative to respective controls. Moreover, control cells exhibited an increase in both size (growth rate) and frequency of spheroids with diameters of 100 μm compared to their respective denosumab-treated cells (Fig. 7B). This indicates that, like rOPG, denosumab targets CSC in BC cells.

3.9. rOPG suppresses tumor growth in orthotopic tumor xenografts
To explore the effect of rOPG on tumor growth of BC cells in vivo, Xenografts were generated by orthotopically injecting MDA-MB-231 cells under the nipple into mammary fat pad of SCID/NUD mice. Treatment with rOPG at 100 μg/kg and 300 μg/kg for 2 weeks strongly inhibited tumor growth compared with the control group treated with PBS (Fig. 8A). Fig. 8B shows clear shrinkage of the tumors present in the animals that were pre-treated with rOPG in a dose-dependent manner. At the end of the treatment period, tumors were excised and weighed, and a clear decrease in the weight of rOPG-treated tumors was observed relative to controls (Fig. 8C). Whole cell lysates were prepared from the retrieved tumor tissues and were subjected to immunoblotting analysis. Fig. 8D and Supplementary Fig. S9 show that rOPG had a clear inhibi- tory effect on the proliferative marker Ki-67 and the EMT process, through down-regulating the epithelial markers E-cadherin and EpCAM, while enhancing the expression of the mesenchymal markers N-cad- herin, Snail, ZEB1 and TWIST1 in breast cancer cells. In addition, rOPG suppressed the stemness process by decreasing the expression of CD44 and ALDH1, while it enhanced the level of CD24 in humanized breast tumor xenografts (Fig. 8D, and Supplementary Fig. S9). Likewise, the pluripotency markers Nanog, Oct-4 and SoX-2 were down-regulated in rOPG-treated tumors relative to controls (Fig. 8D, and Supplementary
Fig. 7. Denosumab inhibits tumorsphere formation in BC cells.
MDA-MB-231 and BT-20 cells were treated either with IgG or with denosumab (DNS) (30 and 50 μg/ml, respectively) for 24 h. A, Cells were double-stained for CD24 and CD44, the proportions of CD44high/CD24low sub-populations were determined by flow cytometry, and are indicated by in the boXes as mean ± SD.; n = 3. B, Cells (1.5X103) were seeded in an ultra-low attachment 96-well plate containing the stem cell-specific medium. Tumorspheres (≥100 μm) were counted. Representative photographs of tumorspheres (right panel). Scale bar = 100 μm. Histograms depicting the number of the formed tumorspheres (left panel). Error bars represent (mean ± SD.; n = 3). *: p ≤ 0.05.
Fig. 8. rOPG inhibits tumor growth, EMT and stemness in orthotopic breast tumor xenografts.
A, Breast cancer xenografts were created by injecting MDA-MB-231 cells (5 × 106) orthotopically into mammary fat pad of nude mice (n = 15). When the tumors reached a volume of 100 mm3, mice were randomized into 3 groups (n = 5 each), which were treated as follows: control (PBS), rOPG (100, 300 μg/kg) via intraperitoneal injection. rOPG was given thrice a week. Graph depicting tumor volume over time and error bars indicate mean ± SD (n = 5). *P ≤ 0.05, **P ≤ 0.01. B, Representative photographs showing the size of the formed tumors in each group. C, Histogram depicting tumor weights (g) upon retrieval from mice. Error bars indicate mean ± SD (n = 5). *P ≤ 0.05, **P ≤ 0.01. D, Following the treatments, tumors from 2 mice per group were excised and protein extracts were prepared and used for immunoblotting analysis. The numbers bellow the bands represent fold change relative to the respective control after correction against the internal controls GAPDH and β-catenin.
Fig. S9). These results indicate that rOPG can target breast cancer cells and their CSC subpopulation in vivo as well.

4. Discussion

In the present report, we present clear evidence that while active breast CAF cells promote carcinogenesis, breast normal fibroblasts suppress this process both in vitro and in vivo. Indeed, indirect co- culturing of MDA-MB-231 cells with NBFs suppressed EMT- and mam- mary stemness-related features. Normal fibroblasts inhibited the pro- liferative, migratory as well as invasive capacities of breast cancer cells in a paracrine fashion. Furthermore, NBFs suppressed the IL-6/STAT3 epigenetic feedback loop, which is responsible for maintaining the EMT as well as the stemness features of cancer cells [25]. Importantly, while CAFs promoted the growth of humanized breast orthotopic tu- mors, NBFs strongly inhibited the growth of these tumors when co-injected with breast cancer cells (Fig. 1E and F). Likewise, it has been previously shown that normal reduction mammoplasty fibroblasts inhibit 3-dimensional morphogenesis and growth of the non-carcinogenic MCF-10A cells and their derivative preneoplastic cells both in vitro and in vivo [10,27,28]. Moreover, Romer et al. have shown that normal mammary fibroblasts can promote acinar morphogenesis and differentiation of primary breast carcinoma cells grown in 3D cul- tures [9]. In another study, Dumont et al. presented clear evidence that normal mammary fibroblasts can suppress tumorigenic and metastatic potential of malignant cells in vivo [29]. Together, these findings indicate that NBFs can suppress both the initiation as well as the progression of breast carcinogenesis. Thereby, these stromal cells possess great therapeutic potential that needs to be exploited, either as a kind of cellular therapy or through identification and use of the secreted cancer inhibitory molecules. We have shown here that OPG plays major role in NBF-related paracrine inhibition of breast carcinogenesis.
Notably, it was striking to see that OPG is secreted not only from NBFs but also from CAF cells. However, adding recombinant IL-6 to SFCM from NBF-6 cells inhibited their paracrine anti-carcinogenic ef- fects on BC cells. This suggests that the transformation of these stromal cells from cancer-suppressive to cancer-promoting cells is mainly related to an increase in the secretion of various pro-carcinogenic molecules, which can activate several cancer-promoting pathways.
In addition, we have shown that NBFs inhibited several BC-related pathways such as NF-κB, AKT and STAT3 in an OPG-dependent manner. Indeed, we have found that, like SFCM from normal breast fi- broblasts, rOPG at low concentration (100 ng/mL) inhibits the IL-6/ STAT3 epigenetic feedback loop, promotes the anti-cancer MET pro- cess and reduces the proliferative, migratory and invasive potential of different breast cancer cell lines. The rOPG-dependent EMT inhibition was also shown on cells grown in 3D. The effect on cell proliferation was supported by showing that rOPG inhibits the cell cycle at the S phase. These results were confirmed in vivo in orthotopic tumor xenografts, wherein rOPG suppressed tumor growth and inhibited several cancer- promoting signaling pathways. Together, these results indicate that rOPG can inhibit breast carcinogenesis both in vitro and in vivo. In pre- vious studies, recombinant OPG has been shown to prevent osteolytic metastatic lesions caused by breast cancer cells in nude and ovariecto- mized mice [30,31].
Strikingly, rOPG also reduced the expression of OPG at both the mRNA and protein levels in BC cells. This could be mediated through inhibition of the NF-κB and STAT3 pathways. Indeed, after showing rOPG-dependent inhibition of the STAT3 pathway, we have shown that OPG expression is modulated in a STAT3-dependent manner. Since OPG is a decoy receptor of RANKL, a known activator of the NF-κB pathway [11,16], we confirmed rOPG-related inhibition of NF-κB, which could mediate STAT3 inhibition owing to the cross-interaction between these two oncogenic pathways. This indicates that rOPG can suppress breast carcinogenesis by inhibiting different pro-carcinogenic pathways, including NF-κB. This was confirmed by showing that denosumab, which is also an inhibitor of RANKL [11], inhibits NF-κB, down-regulates OPG, and suppresses breast carcinogenesis as well. Additionally, denosumab suppressed EMT and the proliferation capacity of BC cells.
Denosumab is currently used to treat breast cancer patients at high risk for bone fractures, and it has recently become a standard of care for patients suffering bone metastasis [11]. It’s also possible to use deno- sumab for the treatment of patients with early breast cancer in adjuvant and/or neoadjuvant settings. Several clinical trials are currently testing these possibilities [11,32,33]. Moreover, some pre-clinical studies have previously shown that denosumab can inhibit the proliferation of BRCA1-mutated mammary epithelial and progenitor cells [11,34]. The present results corroborate these findings by showing that denosumab can target the most resistant cancer cells (CSCs). Indeed, denosumab can inhibit various relevant CSC-related pathways such as NF-κB and β-catenin signaling. This indicates that denosumab could improve the neoadjuvant treatment of certain breast cancer patients through potentiating the anti-CSC effect of some chemotherapeutic drugs.
Additionally, we have shown that rOPG can target breast CSCs in the highly aggressive TNBC cell lines MDA-MB-231 and BT-20. In fact, rOPG modulated the expression of several mammary stemness-related markers through suppressing the β-catenin signaling, and the conse- quent reduction in the proportion of CD24low/CD44high as well as ALDHhigh sub-populations in these cells, and reduced their self-renewing as well as their tumorspheres formation abilities. The rOPG-dependent inhibition of EMT and breast CSC-related biomarkers was also observed in vivo in humanized breast tumor orthotopic Xenografts. These findings indicate that rOPG could be of great therapeutic value for pa- tients suffering early or locally advanced breast cancer. rOPG can target these highly resistant and recurrence-prone breast CSCs, which could enhance treatment efficiency, and consequently increases survival rate among BC patients. It has been shown in a phase I clinical trial, that the recombinant truncated OPG (AMGN-0007) suppressed bone resorption in patients with multiple myeloma or breast carcinoma [35]. Therefore, the use of rOPG has therapeutic potential for breast cancer patients, not only within the bone microenvironment but also for systemic treatment strategies.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors, and was supported in totality by King Faisal Specialist Hospital & Research Center under RAC proposal # 2180018.

Author contributions
Noura Alraouji: Performed experiments, Writing-Original draft preparation. Siti-Fauziah Hendrayani: Performed experiments, statis- tical analysis. Hazem Ghebeh: Performed flow cytometry experiments. Falah Al-Mohanna: Carried out experiments with animals. Abdelilah Aboussekhra: Conceptualization, Supervision, Writing-Reviewing and Editing.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements
We are grateful to Research Center Administration for the continuous help and support.

Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi. org/10.1016/j.canlet.2021.08.013.

References

[1] S. De Silva, K.H. Tennekoon, E.H. Karunanayake, Overview of the genetic basis toward early detection of breast cancer, Breast Cancer 11 (2019) 71–80.
[2] N. Harbeck, F. Penault-Llorca, J. Cortes, M. Gnant, N. Houssami, P. Poortmans, K. Ruddy, J. Tsang, F. Cardoso, Breast cancer, Nat Rev Dis Primers 5 (2019) 66.
[3] S. Byler, S. Goldgar, S. Heerboth, M. Leary, G. Housman, K. Moulton, S. Sarkar, Genetic and epigenetic aspects of breast cancer progression and therapy, Anticancer Res. 34 (2014) 1071–1077.
[4] L. Falzone, M. Grimaldi, E. Celentano, L.S.A. Augustin, M. Libra, Identification of modulated MicroRNAs associated with breast cancer, diet, and physical activity, Cancers 12 (2020).
[5] L.M. Arendt, J.A. Rudnick, P.J. Keller, C. Kuperwasser, Stroma in breast development and disease, Semin. Cell Dev. Biol. 21 (2009) 11–18.
[6] A.E. Place, S. Jin Huh, K. Polyak, The microenvironment in breast cancer progression: biology and implications for treatment, Breast Cancer Res. 13 (2011) 227.
[7] E. Sahai, I. Astsaturov, E. Cukierman, D.G. DeNardo, M. Egeblad, R.M. Evans, D. Fearon, F.R. Greten, S.R. Hingorani, T. Hunter, R.O. Hynes, R.K. Jain, T. Janowitz, C. Jorgensen, A.C. Kimmelman, M.G. Kolonin, R.G. Maki, R.S. Powers, E. Pure, D.C. Ramirez, R. Scherz-Shouval, M.H. Sherman, S. Stewart, T.D. Tlsty, D. A. Tuveson, F.M. Watt, V. Weaver, A.T. Weeraratna, Z. Werb, A framework for advancing our understanding of cancer-associated fibroblasts, Nat. Rev. Canc. 20 (2020) 174–186.
[8] R. Kalluri, The biology and function of fibroblasts in cancer, Nat. Rev. Canc. 16 (2016) 582–598.
[9] A.M. Romer, I. Luhr, A. Klein, A. Friedl, S. Sebens, F. Rosel, N. Arnold, A. Strauss, W. Jonat, M. Bauer, Normal mammary fibroblasts induce reversion of the malignant phenotype in human primary breast cancer, Anticancer Res. 33 (2013) 1525–1536.
[10] A. Sadlonova, Z. Novak, M.R. Johnson, D.B. Bowe, S.R. Gault, G.P. Page, J.V. Thottassery, D.R. Welch, A.R. Frost, Breast fibroblasts modulate epithelial cell proliferation in three-dimensional in vitro co-culture, Breast Cancer Res. 7 (2005) R46–R59.
[11] M. Infante, A. Fabi, F. Cognetti, S. Gorini, M. Caprio, A. Fabbri, RANKL/RANK/ OPG system beyond bone remodeling: involvement in breast cancer and clinical perspectives, J. EXp. Clin. Canc. Res. 38 (2019) 12.
[12] D.L. Lacey, E. Timms, H.L. Tan, M.J. Kelley, C.R. Dunstan, T. Burgess, R. Elliott, A. Colombero, G. Elliott, S. Scully, H. Hsu, J. Sullivan, N. Hawkins, E. Davy, C. Capparelli, A. Eli, Y.X. Qian, S. Kaufman, I. Sarosi, V. Shalhoub, G. Senaldi, J. Guo, J. Delaney, W.J. Boyle, Osteoprotegerin ligand is a cytokine that regulates osteoclast differentiation and activation, Cell 93 (1998) 165–176.
[13] H. Yasuda, N. Shima, N. Nakagawa, K. Yamaguchi, M. Kinosaki, S. Mochizuki, A. Tomoyasu, K. Yano, M. Goto, A. Murakami, E. Tsuda, T. Morinaga, K. Higashio,
N. Udagawa, N. Takahashi, T. Suda, Osteoclast differentiation factor is a ligand for osteoprotegerin/osteoclastogenesis-inhibitory factor and is identical to TRANCE/ RANKL, Proc. Natl. Acad. Sci. U. S. A. 95 (1998) 3597–3602.
[14] I. Holen, S.S. Cross, H.L. Neville-Webbe, N.A. Cross, S.P. Balasubramanian, P.I. Croucher, C.A. Evans, J.M. Lippitt, R.E. Coleman, C.L. Eaton, Osteoprotegerin (OPG) expression by breast cancer cells in vitro and breast tumours in vivo–a role in tumour cell survival? Breast Canc. Res. Treat. 92 (2005) 207–215.
[15] H.S. Park, A. Lee, B.J. Chae, J.S. Bae, B.J. Song, S.S. Jung, EXpression of receptor activator of nuclear factor kappa-B as a poor prognostic marker in breast cancer, J. Surg. Oncol. 110 (2014) 807–812.
[16] S. Goswami, N. Sharma-Walia, Osteoprotegerin rich tumor microenvironment: implications in breast cancer, Oncotarget 7 (2016) 42777–42791.
[17] S. Goswami, N. Sharma-Walia, Osteoprotegerin secreted by inflammatory and invasive breast cancer cells induces aneuploidy, cell proliferation and angiogenesis, BMC Canc. 15 (2015) 935.
[18] M. Weichhaus, S.T. Chung, L. Connelly, Osteoprotegerin in breast cancer: beyond bone remodeling, Mol. Canc. 14 (2015) 117.
[19] M. Weichhaus, P. Segaran, A. Renaud, D. Geerts, L. Connelly, Osteoprotegerin expression in triple-negative breast cancer cells promotes metastasis, NIK SMI1 Cancer Med 3 (2014) 1112–1125.
[20] D. Geerts, C. Chopra, L. Connelly, Osteoprotegerin: Relationship to breast cancer risk and prognosis, Front Oncol 10 (2020) 462.
[21] N.M. Hawsawi, H. Ghebeh, S.F. Hendrayani, A. Tulbah, M. Al-Eid, T. Al-Tweigeri,
D. Ajarim, A. Alaiya, S. Dermime, A. Aboussekhra, Breast carcinoma-associated fibroblasts and their counterparts display neoplastic-specific changes, Canc. Res. 68 (2008) 2717–2725.
[22] M.A. Al-Mohanna, H.H. Al-Khalaf, N. Al-Yousef, A. Aboussekhra, The p16INK4a tumor suppressor controls p21WAF1 induction in response to ultraviolet light, Nucleic Acids Res. 35 (2007) 223–233.
[23] G. Silva, A. Aboussekhra, p16(INK4A) inhibits the pro-metastatic potentials of osteosarcoma cells through targeting the ERK pathway and TGF-beta1, Mol. Carcinog. 55 (2016) 525–536.
[24] H. Ghebeh, G.M. Sleiman, P.S. Manogaran, A. Al-Mazrou, E. Barhoush, F.H. Al- Mohanna, A. Tulbah, K. Al-Faqeeh, C.N. Adra, Profiling of normal and malignant breast tissue show CD44high/CD24low phenotype as a predominant stem/ progenitor marker when used in combination with Ep-CAM/CD49f markers, BMC Canc. 13 (2013) 289.
[25] D. Iliopoulos, H.A. Hirsch, K. Struhl, An epigenetic switch involving NF-kappaB, Lin28, Let-7 MicroRNA, and IL6 links inflammation to cell transformation, Cell 139 (2009) 693–706.
[26] H.H. Lee, J. Jung, A. Moon, H. Kang, H. Cho, Antitumor and anti-invasive effect of apigenin on human breast carcinoma through suppression of IL-6 expression, Int. J. Mol. Sci. 20 (2019).
[27] M.P. Shekhar, J. Werdell, S.J. Santner, R.J. Pauley, L. Tait, Breast stroma plays a dominant regulatory role in breast epithelial growth and differentiation: implications for tumor development and progression, Canc. Res. 61 (2001) 1320–1326.
[28] A. Sadlonova, D.B. Bowe, Z. Novak, S. Mukherjee, V.E. Duncan, G.P. Page, A.R. Frost, Identification of molecular distinctions between normal breast-associated fibroblasts and breast cancer-associated fibroblasts, Cancer Microenviron 2 (2009) 9–21.
[29] N. Dumont, B. Liu, R.A. Defilippis, H. Chang, J.T. Rabban, A.N. Karnezis, J.A. Tjoe, J. Marx, B. Parvin, T.D. Tlsty, Breast fibroblasts modulate early dissemination, tumorigenesis, and metastasis through alteration of extracellular matriX characteristics, Neoplasia 15 (2013) 249–262.
[30] S. Morony, C. Capparelli, I. Sarosi, D.L. Lacey, C.R. Dunstan, P.J. Kostenuik, Osteoprotegerin inhibits osteolysis and decreases skeletal tumor burden in syngeneic and nude mouse models of experimental bone metastasis, Canc. Res. 61 (2001) 4432–4436.
[31] P.D. Ottewell, N. Wang, H.K. Brown, C.A. Fowles, P.I. Croucher, C.L. Eaton, I. Holen, OPG-Fc inhibits ovariectomy-induced growth of disseminated breast cancer cells in bone, Int. J. Canc. 137 (2015) 968–977.
[32] M.E. Reyes, T. Fujii, D. Branstetter, S. Krishnamurthy, H. Masuda, X. Wang, J.M. Reuben, W.A. Woodward, B.J. Edwards, G.N. Hortobagyi, D. Tripathy, W.C. Dougall, B.L. Eckhardt, N.T. Ueno, Poor prognosis of patients with triple- negative breast cancer can be stratified by RANK and RANKL dual expression, Breast Canc. Res. Treat. 164 (2017) 57–67.
[33] M. Gnant, G. Pfeiler, G.G. Steger, D. Egle, R. Greil, F. Fitzal, V. Wette, M. Balic, F. Haslbauer, E. Melbinger-Zeinitzer, V. Bjelic-Radisic, R. Jakesz, C. Marth, P. Sevelda, B. Mlineritsch, R. EXner, C. Fesl, S. Frantal, C.F. Singer, B. Austrian, G. Colorectal Cancer Study, Adjuvant denosumab in postmenopausal patients with hormone receptor-positive breast cancer (ABCSG-18): disease-free survival results from a randomised, double-blind, placebo-controlled, phase 3 trial, Lancet Oncol. 20 (2019) 339–351.
[34] E. Cuyas, B. Corominas-Faja, M.M. Martin, B. Martin-Castillo, R. Lupu, J. Brunet, J. Bosch-Barrera, J.A. Menendez, BRCA1 haploinsufficiency cell-autonomously activates RANKL expression and generates denosumab-responsive breast cancer- initiating cells, Oncotarget 8 (2017) 35019–35032.
[35] J.J. Body, P. Greipp, R.E. Coleman, T. Facon, F. Geurs, J.P. Fermand, J. L. Harousseau, A. Lipton, X. Mariette, C.D. Williams, A. Nakanishi, D. Holloway, S.W. Martin, C.R. Dunstan, P.J. Bekker, A phase I study of AMGN-0007, a recombinant osteoprotegerin construct, in patients with multiple myeloma or breast carcinoma related bone metastases, Cancer 97 (2003) 887–892.