Radiation Modulates the Mechanoresponse of Bone-Homing Triple-Negative Breast Cancer Cells

Abstract

Bone is a common site of breast cancer metastasis, which dramatically increases fracture risk. Recent in vivo bone metastasis studies show that mechanical loading is osteoprotective; however, little is known about how loading regulates breast cancer cell function in the unique bone mechanical environment, especially in combination with radiotherapy, one of the first-line treatments for advanced breast cancer patients. Here, we characterize the breast cancer cell response to a range of bone-mimicking fluid shears and determine how irradiation further modulates one candidate gene: SERPINE1. We found that irradiation, regardless of dosage, modulates SERPINE1 expression and is sensitive to the timing of administration. Additionally, protein expression of SERPINE1 accompanies a protumorigenic gene expression profile, which is elevated with higher-magnitude fluid shear stresses in a bone-mimicking 3D environment. Thus, we postulate that plasminogen activator inhibitor-1 (PAI-1) (encoded by SERPINE1) is a critical growth factor contributing to osteolytic lesion development in the bone metastatic vicious cycle of breast cancer.

Impact Statement

Research on the bone metastatic vicious cycle frequently investigates how osteocytes and other cells involved in bone remodeling impact breast cancer cells, but the downstream impacts of the breast cancer cell response to the bone mechanical environment are understudied. This work is one of the few studies investigating how breast cancer cells respond to a bone-like fluid mechanical environment and will facilitate our understanding of the breast cancer bone metastatic vicious cycle. Additionally, to our knowledge, this is the first study that has examined how the timing of ionizing irradiation administration influences the breast cancer cell mechanoresponse in a 3D model. Mechanosensation and the cell mechanoresponse have largely been studied in 2D models, and an understanding of the physiologically relevant 3D environment is crucial to elucidating complex cell–extracellular matrix interactions and responses to damaging stimuli such as ionizing radiation.

Keywords

bone metastasis irradiation fluid shear vicious cycle sequencing

Introduction

Over 70% of advanced breast cancer patients are diagnosed with incurable skeletal metastasis.^1,2 Breast cancer that metastasizes to bone is generally osteolytic, destroying bone in a feed-forward vicious cycle that concurrently promotes tumor growth and bone resorption.³ Of all the breast cancer subtypes, triple-negative breast cancer (TNBC) has the lowest rate of bone metastases, but is the most aggressive subtype and has the lowest 5-year survival rate once bone metastasis occurs.^1,4 Due to localized osteolytic lesions, bone metastases require a more targeted treatment approach.⁵ Radiotherapy is one such frontline treatment modality, providing both local tumor control and palliation of skeletal pain.⁶ For TNBC, radiation therapy is a critical intervention, since standard treatments targeting estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) are ineffective.⁷ Radiation induces DNA double-strand breaks, leading to TNBC cell death.^7,8 However, advanced TNBC is particularly resistant to most therapies, especially radiation, and requires treatment combinations to improve the effectiveness of local therapies.^6,9,10

Exercise is one nonpharmacological intervention that has been rising to prominence as part of treatment strategies for cancer.¹¹ Physical activity is especially critical to maintaining bone strength. In healthy bone, mechanical stimuli are a well-documented anabolic signal for bone,^12–15 but in bone metastasis, the skeleton’s response to these signals is disrupted. Recent in vivo studies have shown that increased mechanical loading (i.e., an analog for exercise) has antitumorigenic effects on bone metastasis models in mice.^16–21 However, the skeleton’s anabolic response to loading is not completely protected in these models,^16,18 suggesting that metastasis interferes with the skeleton’s ability to functionally adapt to mechanical loads.

The skeleton, as a mechanically active organ, maintains homeostasis via the osteocyte, the primary mechanosensitive cell in bone that senses and responds to mechanical stimuli.^22–24 Osteocytes respond to force-induced interstitial fluid flow, the central mechanical signal leading to net bone formation.²⁴ Our previous in vitro studies revealed that osteocytes lost their mechanosensitivity when cultured with breast cancer cell conditioned media, pointing to a potential mechanism underlying the muted response to loading in vivo.^25,26 Furthermore, when breast cancer cells were also loaded, their downstream effects on the mechanoresponse of mesenchymal stem cells²⁷ and osteocytes^25,28 were modulated. Specifically, conditioned media from loaded breast cancer cells promoted a positive mechanoresponse from these osteoblast-lineage cells. However, none of these studies identified molecular mediators.

After metastatic colonization, cancer cells arrive in the bone marrow and often migrate to the endosteal surface. Fluid flow within the marrow and along the bone surfaces is highly variable, ranging from as low as 0.02 Pa to as high as 8 Pa based on computational estimates ex vivo.^29–33 This, combined with the increased tissue stiffness of bone, anywhere between 1 and 10 GPa, compared to the stiffness of tumors in the breast tissue on the order of 1–10 kPa, makes the skeleton a far more mechanically complex and dynamic environment for cancer cells.^34–37 The response of breast cancer cells to mechanical stimuli has been well documented in vitro, where a larger stimulus (e.g., substrate stiffness, increased fluid flow) correlates with increased migration and proliferation.^38–41 However, most of these studies were not conducted in a bone-mimicking microenvironment, nor did they account for the unique micromechanics of that environment. In a previous study from our group, breast cancer cell viability was unaffected after undergoing mechanical loading in the bone microenvironment, but expression of bone remodeling genes was mechanically regulated.¹⁶ We reasoned that the breast cancer cell response to mechanical stimuli may undermine the osteoprotective effects of increased loading. However, there are few studies characterizing the comprehensive breast cancer cell response to a bone-mimicking and mechanically similar microenvironment.

In this study, we investigated the response of TNBC cells to physiologically relevant, bone-mimicking fluid shear stresses. To address how mechanical loads may modulate TNBC resistance to common therapies like radiation, we also determined how irradiation alters the TNBC cell transcriptome. We identified several genes that may be responsible for abrogating the effects of loading, including SERPINE1. SERPINE1 encodes the serine protease inhibitor plasminogen activator inhibitor-1 (PAI-1),⁴² and our particular interest in SERPINE1 was due to its overexpression in TNBC,^43,44 its association with breast cancer cell therapeutic resistance,^45,46 and its critical role in managing bone homeostasis via fibroblast growth factor 23 (FGF23) regulation and its transcriptional activation in the transforming growth factor beta (TGF-β) signaling pathway in osteocytes.^42,47 Here, we show that irradiation, regardless of dose, shifts breast cancer cell gene expression in a timing-dependent manner, and that protein-level upregulation of many such genes accompanies a protumorigenic transcriptional profile under higher shear magnitudes in a 3D bone-mimetic culture system.

Methods

Bone-mimetic scaffold fabrication

Scaffolds were fabricated using a mixture of 152 mg of sieved NaCl, 8 mg of hydroxyapatite crystals, and 8 mg of poly(lactide-coglycolide) microspheres.⁴⁸ This mixture forms an 8mm diameter, 2 mm height structure after undergoing 2 metric tons of pressure delivered by a manual hydraulic press (Carver Press). A subsequent gas-foaming and salt-leaching procedure formed the final scaffold by revealing interconnected pores of 250–450 μm in diameter, resulting in an overall porosity of 70–90%.⁴⁹ Previous work on these scaffolds has confirmed that, while they do not mimic trabecular bone tissue microarchitecture, the scaffolds do recapitulate the bone microenvironment in important ways. Previous work shows that these scaffolds induce osteogenesis of mesenchymal stem cells²⁷ and reproduce key features of breast cancer cell behavior in the bone tumor microenvironment.^48,50

3D cell culture

A bone-homing and GFP-expressing variant of the parental MD Anderson-Metastatic Breast-231 (MDA-MB-231) (generously provided by Dr. Julie Rhoades) cell line was cultured with complete Dulbecco's Modified Eagle Medium (DMEM, with 10% fetal bovine serum, or FBS, and a 1% penicillin–streptomycin solution). Cell passage numbers ranged between P20 and P35. Scaffolds were seeded drop-wise with 1.5 million bone-homing MDA-MB-231 cells (MDA-B) and allowed to incubate on an orbital shaker (0.04 g) for 48 h at 5% CO₂ and 37°C prior to any treatments to aid adhesion and infiltration. During this time period, MDA-Bs typically distribute across all surfaces of the scaffold and infiltrate roughly 200 μm into the total scaffold depth.^16,48,49

Mechanical loading via applied perfusion

MDA-B cells seeded in scaffolds were subjected to mechanical loading via a custom direct perfusion bioreactor (Fig. 1B). DMEM flowed through the bioreactor’s tubing system to deliver cyclic inlet velocities of 1 (Low) and 15 (High) mL/min at 1 Hz to scaffolds within the bioreactor chambers. Perfusion magnitudes were chosen based on our previous work in which we simulated the fluid mechanical environment that would be experienced in future experiments by “high,” “medium,” and “low” applied flow rates that our custom bioreactor system engendered during perfusion.⁴⁹ We selected the “low” and “high” flow rates for this article because they had the largest separation in median shear stresses. Also, these flow rates correspond to shear stresses of approximately 60 and 180 mPa, both of which encompass previous studies’ estimates of physiological shear stresses from fluid flow along the bone surface during loading.^14,32 Finally, in early optimization studies, we confirmed that cell number, assessed via total DNA content, was insensitive to our range of flow rates (Supplementary Fig. S1), corroborating previous work with compressive loading of scaffolds.¹⁶ This loading method was performed for 1 h every day for 3 days. Static scaffolds were placed similarly in bioreactor chambers but did not undergo perfusion. In between loading bouts, scaffolds were returned to the orbital shaker. At the end of the 3-day loading period, scaffolds were either harvested immediately for RNA extraction (n = 3, 4 technical replicates, i.e., scaffolds, per condition per experimental “n”) or allowed to incubate on an orbital shaker for 24 h in 1% FBS before this low-serum cell culture media was collected for cytokine array analysis (n = 5).

FIG. 1.

Perfusion via custom bioreactor. (A) Perfusion of cell-seeded scaffolds is pulsatile (1 Hz) at low and high flow rates, previously validated.⁴⁹ (B) The bioreactor holds scaffolds within a chamber and permits flow through tubing below the chamber. MDA-B, bone-homing MDA-MB-231 cells.

Cell culture media was analyzed for a wide array of cytokines and chemokines (105 total) using a membrane-based immunoassay (Proteome Profiler Human XL Cytokine Array, R&D Systems, ARY022B). For sequencing, total RNA was extracted using an in-house protocol, from which poly(A)-selected RNA was converted into cDNA libraries with a size distribution of 250–350 bp using a KAPA Stranded mRNA-Seq kit (KAPA Biosystems, 07962193001). Dual-index adapters were attached at a concentration of 100 nM (KAPA Biosystems, 08861919702). Libraries were sequenced at 20 million reads per sample (paired-end) using an Illumina NovaSeq X instrument. Raw RNA-Seq data were processed via the BioFrontiers Fiji supercomputing cluster. Adapter sequences were trimmed, and low-quality bases (Phred score < 10) were filtered using BBDuk (BBMap v38.05). Read quality was evaluated after trimming using FastQC (v0.11.8). Reads were aligned to the hg38 reference genome assembly (UCSC Genome Browser) using STAR (v2.7.11a) with GENCODE v47 annotation. Gene-level counts were generated with featureCounts (v1.6.2), and differential expression analysis was performed in DESeq2 (v1.40.2). Genes with an adjusted p value < 0.05 and an absolute log₂ of the fold change (FC) > 1 were considered significantly differentially expressed. We then used the open-source graphical gene enrichment tool ShinyGO⁵¹ to identify significantly enriched biological processes and signaling pathways. Venn diagrams of differentially expressed genes were created using the open-source software “jvenn.”⁵²

Experimental design with irradiation and high-stimulus perfusion

Scaffolds seeded with MDA-B cells were divided into two groups. One group (Fig. 2, PRE) received irradiation at 0, 2, and 10 Gy 48 h after seeding and was allowed to incubate for 24 h prior to mechanical loading. After a 3-day loading period, scaffolds were then incubated for 24 h at 5% CO₂ and 37°C, and RNA was extracted for quantitative real-time polymerase chain reaction (qPCR).

FIG. 2.

Time-sequence study with irradiation and high-stimulus perfusion.

The second group of scaffolds (Fig. 2, POST) was immediately subjected to loading after seeding. Once the 3-day loading period was complete, scaffolds were incubated on an orbital shaker for 24 h and subsequently received irradiation at 0, 2, and 10 Gy. These scaffolds were again incubated for 24 h at 5% CO₂ and 37°C before RNA extraction for qPCR.

Samples were irradiated at room temperature in a Faxitron Cabinet X-ray System Model RX-650 (Lincolnshire, IL) at 115 kV and 5.33 rad/s. Loading bouts were administered at 2 h per day for 3 days, with 15 mL/min inlet flow rates for all loaded scaffolds. We applied a longer mechanical stimulation period to bolster a response to mechanical loading. Total RNA was extracted using an in-house protocol and converted into cDNA (Qiagen, 205311). cDNA was diluted to 25 ng/μL and 50 ng were quantified with qPCR. Real-time qPCR was run using a Rotor-Gene Q system (Qiagen, 9001862) and assays were assembled using primers designed in-house (Supplementary Table S1) and a QuantiNova SYBR Green qPCR kit (Qiagen, 208252).

Statistical analysis

For RNA-Seq, genes with an adjusted p value < 0.05 satisfying log₂FC > 1 were considered upregulated, and those satisfying log₂FC < −1 were considered downregulated. In our gene ontology enrichment analysis with ShinyGO, genes were sorted according to significance levels (p-adjusted < 0.05) set by Benjamini and Hochberg’s false discovery rate (FDR).⁵¹ Treatment groups were normalized to nonloaded samples (reference group).

For time-sequence experiments, qPCR results were calculated using the efficiency-corrected 2-ΔΔCt method. We then conducted a three-way ANOVA with Tukey’s HSD post hoc testing (R 4.3.3 and GraphPad Prism 10.6.1) for main effect comparisons satisfying p ≤ 0.05. Treatment groups were normalized to nonloaded samples that did not undergo irradiation (reference group).

Results

High-magnitude fluid shear stresses amplified protumorigenic breast cancer cell gene expression

High-magnitude flow tripled the number of differentially expressed MDA-B genes in response to fluid shear (Fig. 3, 4). Notably, many of these genes are protumorigenic. Low-magnitude flow upregulated PDGFB (platelet derived growth factor subunit B), NEDD9 (neural precursor cell expressed, developmentally down-regulated 9), and LOXL4 (lysyl oxidase-like protein 4) (Table 1), all of which have been implicated in cancer cell migration and metastatic progression. PDGFB, in its PDGF-BB homodimer form, induces β-catenin expression by various breast cancer subtypes in vitro.⁵³ NEDD9 is a focal adhesion protein that plays significant roles in cell attachment, migration, and invasion,^62,63 and LOXL4 is well known to be overexpressed in a number of cancers, interacting with cell-surface integrins to enhance cancer cell outgrowth.⁵⁶

FIG. 3.

High flow results in more DEGs. Volcano plots of genes differentially expressed (threshold for up/down-regulation: $|\log_{2} F C| > 1$ ) by MDA-B cells exposed to (A) Low (1 mL/min, ∼12 mPa) and (B) High (15 mL/min, ∼180 mPa) pulsatile fluid shear. DEGs, differentially expressed genes; FC, fold change.

FIG. 4.

High flow produces a larger number of unique DEGs. While 1167 DEGs are shared between low and high flow conditions, high-magnitude flow expresses many more unique DEGs (1533) than low-magnitude flow (471).

Table 1.

Notable DEGs at the Low Flow Rate $(p < 0.05, |{LOG}_{2} F C| > 1)$

Gene	log₂(FC)	Functions associated with progression of bone metastatic vicious cycle
PDGFB	2.22	Promotes osteoclast activity and Wnt/β-catenin signaling pathway activation in TNBCs^53,54
NEDD9	1.58	Activates RANKL-induced osteoclastogenesis as a TGF-β-Smad2/3 target gene⁵⁵
LOXL4	1.23	May work with other lysyl oxidases in development of a bone premetastatic niche^56,57
ID2	−1.33	Well-known BMP pathway target gene, may coexpress with RANK^58,59
RN7SK	−2.76	Essential for tumorigenesis, but knockdown also induces osteogenic differentiation and correlates with high RUNX2 expression^60,61

DEGs, differentially expressed genes; FC, fold change; PDGFB, platelet derived growth factor subunit B; NEDD9, neural precursor cell expressed developmentally down-regulated 9; LOXL4, lysyl oxidase-like protein 4; ID2, inhibitor of differentiation 2; TGF-β, transforming growth factor beta; BMP, bone morphogenetic protein; RANK, receptor activator of nuclear factor κB ligand; RANKL, RANK ligand; RUNX2, runt-related transcription factor 2.

Interestingly, high-magnitude fluid shear stress also downregulated several extracellular matrix (ECM) turnover markers, such as matrix metalloproteinases MMP25 and MMP8 (Table 2). This downregulation accompanied an upregulation of genes with important roles in ECM protein production, such as COL1A1 (pro-alpha1 chain of Type 1 collagen) and HSPA1B (heat shock protein family A member 1B). This suggests that higher-magnitude fluid flow spurs desmoplasia in the tumor microenvironment, as significant changes in MMP markers were absent in the low flow groups (Table 1).

Table 2.

Notable DEGs at the High Flow Rate $(p < 0.05, |{LOG}_{2} F C| > 1)$

Gene	log₂(FC)	Functions associated with progression of bone metastatic vicious cycle
PDGFB	3.91	Promotes osteoclast activity and Wnt/β-catenin signaling pathway activation in TNBCs^53,54
NEDD9	2.72	Activates RANKL-induced osteoclastogenesis as a TGF-β-Smad2/3 target gene⁵⁵
LOXL4	1.98	May work with other lysyl oxidases in development of a bone premetastatic niche^56,57
COL1A1	1.74	Expressed by cancer-associated fibroblasts, promotes metastasis^64,65
SERPINE1	1.15	Promotes invasion and metastasis, aids therapeutic resistance in TNBCs^43,45,46
HSPA1B	1.00	Accompanies other heat shock proteins in progression of skeletal tumor growth, correlated with immune infiltrates in various breast cancer subtypes^66,67
MMP25	−1.46	Involved in cancer progression, part of MMP family of bone metastatic vicious cycle regulators^68,69
RN7SK	−3.30	Essential for tumorigenesis, but knockdown also induces osteogenic differentiation and correlates with high RUNX2 expression^60,61
MMP8	−3.78	Osteoblast-derived MMP, involved in cancer cell invasion and proliferation, part of MMP family of bone metastatic vicious cycle regulators^69,70

DEGs, differentially expressed genes; FC, fold change; PDGFB, platelet derived growth factor subunit B; TNBC, triple-negative breast cancer; NEDD9, neural precursor cell expressed developmentally down-regulated 9; RANK, receptor activator of nuclear factor κB ligand; RANKL, RANK ligand; TGF-β, transforming growth factor beta; LOXL4, lysyl oxidase-like protein 4; COL1A1, pro-alpha1 chain of Type 1 collagen; HSPA1B, heat shock protein family A member 1B; MMP, matrix metalloproteinase; RUNX2, runt-related transcription factor 2.

Gene ontology enrichment analysis revealed that both low- and high-magnitude flow enriched pathways associated with transcriptional misregulation in cancer (Fig. 5), designated by such genes as GADD45G (growth arrest and DNA-damage inducible gamma), ID2 (inhibitor of differentiation), and BAIAP3 (BAI-associated protein 3). These results also suggest that dynamic fluid shear may induce cancer stemness in MDA-B cells. In low-flow groups, MDA-B cells demonstrated enrichment of signaling pathways regulating stem cell pluripotency (Fig. 5A), while in high-flow groups, MDA-B differential gene expression was associated more specifically with the Hippo signaling pathway. However, there was a relatively high false discovery rate (log₁₀FDR ∼1.8–2.8) associated with these stemness pathways (likely due to the low number of genes identified to be associated with these pathways), requiring more targeted investigation to confirm whether fluid shear does indeed stimulate cancer stemness. Many genes differentially expressed in the high-flow groups are also involved in pathways regulating cell motility, migration, and invasion.

FIG. 5.

Higher magnitude fluid shear is increasingly protumorigenic. (A) Low flow enriches pathways associated with stemness and metastatic progression. (B) High flow additionally enriches pathways that have significant roles in bone metastasis and the bone mechanoresponse.

One gene playing roles in many of these motility and invasion pathways, as well as cancer stemness-associated pathways, is SERPINE1. This gene encodes PAI-1, a serine protease inhibitor implicated in ECM remodeling, angiogenesis, and cancer progression.^44,71 In our enrichment analysis, we found SERPINE1 to be involved in multiple enriched pathways, including Hippo, PI3K–Akt, TGF-β, and p53. While these pathways have quite far-reaching functions, they are known to be highly activated in metastatic cancers, especially TNBC.^43,44,72

Additionally, many pathways such as TGF-β and PI3K-Akt have been implicated in their roles in mediating bone homeostasis in bone metastatic breast cancers.^47,73 Elevated PAI-1 levels, particularly, are associated with poor prognosis in breast cancer, and emerging evidence points to a potential role in modulating osteocyte signaling and impairing bone quality.^42,47,74 To confirm that flow altered PAI-1 secretion from MDA-B cells, we found that PAI-1 secretion followed SERPINE1 upregulation with high-magnitude fluid flow via an antibody array screening (Fig. 6B,C). Furthermore, SERPINE1 was one candidate gene that supported our interest in the combined effects of radiotherapy and mechanical loading on MDA-B expression. Ionizing radiation is known to induce SERPINE1 overexpression in TNBC cells in vivo, and this overexpression is generally protumorigenic, but whether this response is altered by mechanical loading is unknown. As such, we selected SERPINE1 as a target gene for subsequent experiments involving ionizing radiation.

FIG. 6.

SERPINE1 is significantly upregulated with high-stimulus fluid shear stress. (A) High flow (∼200 mPa) induces a ∼2.2-fold increase in MDA-B SERPINE1 expression (DESeq2 Wald test, *** p < 0.0005). (B, C) Cytokine array representative image and quantification also show increased secretion of PAI-1 at the high flow rate. PAI-1, plasminogen activator inhibitor-1.

Irradiation prior to flow enabled apoptosis evasion, but postflow irradiation may be more osteoprotective

Unexpectedly, we found that MDA-B gene expression of SERPINE1, Ki67 (MKI67, Marker of proliferation Ki-67), and CASP3 (caspase-3) was insensitive to irradiation dose irrespective of treatment sequence (Fig. 7, Supplementary Figures S2, S3, and S4). While previous studies show that parental MDA-MB-231 cells cultured in 2D begin undergoing apoptosis at 2 Gy of ionizing radiation, with marked apoptosis at 10 Gy, we found in our 3D model that 10 Gy did not significantly increase MDA-B CASP3 expression in nonloaded or loaded groups (Fig. 7B). There was a nonsignificant trend (p = 0.83) of decreased proliferation (downregulation of KI67) in loaded groups when loading preceded irradiation (Fig. 7C). Similarly, SERPINE1 was downregulated (ns, p = 0.89) as irradiation dose increased, but only when loading followed irradiation.

FIG. 7.

MDA-B gene expression of (A) SERPINE1, (B) CASP3, and (C) Ki67 is insensitive to the dosage of irradiation. Only loaded groups show some dose response trends (ns). Mean qPCR fold changes are shown for each time point and experiment (error bars = SEM). qPCR, quantitative polymerase chain reaction.

Though MDA-B gene expression was insensitive to irradiation dose, with or without perfusion, we observed that the timing of irradiation administration significantly altered this expression. When we combined high-magnitude mechanical loading with irradiation, loading did not upregulate SERPINE1 expression, regardless of the sequence of experiments (Fig. 8A), in contrast to our dosed loading studies (Table 2, Fig. 6). In fact, when irradiation preceded loading (PRE), fluid shear downregulated SERPINE1 relative to nonloaded groups. When irradiation followed loading (POST), however, loading neutralized this effect. Regardless of the sequence of events, based on SERPINE1 expression alone, the inclusion of irradiation in these experiments made loading antitumorigenic and osteoprotective.

FIG. 8.

Postflow irradiation may be more osteoprotective. (A) Flow downregulates SERPINE1 when applied before irradiation. (B) Loading enables MDA-B evasion of apoptosis, but only with preflow irradiation. (C) Loading decreased proliferation (a vs. b, p < 0.005). Three-way ANOVA and posthoc Tukey HSD test, ** p < 0.005, * p < 0.05.

Similar to SERPINE1, when irradiation preceded loading (PRE), MDA-B gene expression of CASP3 was downregulated (Fig. 8B), suggesting that preflow irradiation enabled cancer cell evasion of apoptosis. When irradiation followed loading (POST), CASP3 expression was unchanged with loading. Interestingly, KI67-mediated cell proliferation was consistently downregulated with loading compared to nonloaded groups, irrespective of irradiation timing (Fig. 8C).

Discussion

The objective of this study was to characterize the transcriptional mechanoresponse of breast cancer cells within a 3D bone-mimicking platform.⁴⁸ We showed that the protumorigenic response to fluid shear is sustained during bone metastasis using our 3D dynamic flow model (Fig. 5), as shown in previous studies with steady flow models.⁴¹ Our data provides evidence that high-magnitude flow may stimulate fibroblastic gene expression in the tumor-bone microenvironment due to the concurrent downregulation of MMPs and upregulation of genes coding for ECM modeling functions, e.g., collagen production and chaperone protein-mediated regulation of ECM protein folding (Table 2). We identified one candidate gene highly expressed with high-magnitude flow: SERPINE1. We also showed that MDA-B cell secretion of PAI-1 (encoded by SERPINE1) increased with high flow as well.

While we have focused on one gene in this article, SERPINE1, our initial sequencing results demonstrate that dynamic, bone-mimicking fluid flow activated several key pathways in TNBC cells that require further investigation, especially in the context of the bone metastatic vicious cycle. An important observation from our sequencing data is that high-magnitude fluid flow promoted a persistent fibroblastic gene expression pattern in our MDA-B cells. Specifically, we showed that MDA-B cells upregulate ECM production genes while downregulating MMPs. Typically, elevated interstitial flows are known to promote ECM production in TNBC cell via a fibroblast-like phenotype, while increasing expression of ECM degradation markers such as MMPs.^64,75,76 MMP upregulation is a key marker of protumorigenesis, contributing to increased invasion in TNBC cells.⁷⁷ However, our loading model for breast cancer bone metastasis downregulated MMPs at high magnitudes of flow, suggesting that breast cancer cells differentially respond to fluid shear once they have colonized the skeleton.

Our timing sequence study with irradiation and perfusion yielded surprising results. Several studies have shown that ionizing radiation (at both 2 and 10 Gy) upregulated both SERPINE1 and CASP3 in TNBC cells and induced significant amounts of apoptosis.^45,78–80 Yet, we did not observe upregulation of SERPINE1 or CASP3 expression, even in our nonloaded groups (Fig. 8B). When irradiation was paired with flow, we in fact observed a suppression of SERPINE1 upregulation for both post- and preflow irradiation treatment. Moreover, previous studies on the DNA damage response of parental MDA-MB-231 cells to radiation (8–10 Gy) show that CASP3 is regularly upregulated, a response we did not observe in our nonloaded groups. We speculate that this shift could be due to the 3D culture environment, combined with the introduction of a loading stimulus that interferes with the typical TNBC cell response to radiation. We also speculate that the more unique results we observe in this study may be due to the use of an aggressive, bone-homing variant of MDA-MB-231 cells that is likely particularly resistant to ionizing radiation. In other bone-homing cell lines generated by the same technique (repeated in vivo passaging of MDA-MB-231 cells), bone-homing cells overexpressed genes such as SMAD2 (SMAD family member 2), HRAS (Hras proto-oncogene, GTPase), and TIP60,⁸¹ all of which are associated with radioresistance in cancer cells.^82–84 In our initial dosed loading experiments, we observed that flow enriched pathways associated with cancer stemness, consistent with literature on flow inducing cancer stemness. As a more stem-like profile is also associated with radioresistance, we anticipate that flow further magnified the radioresistance of this bone-homing variant. Further mechanistic studies are required to confirm this theory.

We measured CASP3 expression in this study as our marker of radiation-induced apoptosis and only observed detectable changes (downregulation) when irradiation preceded loading (PRE). However, irradiation also causes cell death via necrosis, indirect DNA damage, and inhibited cell division.⁸ Thus, CASP3 expression alone may not fully capture MDA-B cell death due to radiation. It is possible that other mechanisms of cell death are instead leading to marked MDA-B cell death at 10 Gy, as we would expect. In our sequencing data, we observed no flow-based differential expression of markers of cell death via either apoptotic or necrotic processes, such as CASP3 and PPIF (peptidyl-prolyl isomerase F).⁸ Accordingly, the significant downregulation of CASP3 with loading in preflow irradiation groups appears to be primarily a consequence of the timing of irradiation. In future studies, we aim to determine whether necrotic and autophagy-dependent cell death markers (also activated by ionizing radiation⁸) respond similarly to CASP3. We also aim to quantify cell death via histology at various timepoints within these experiments to confirm whether evasion of cell death truly occurs in response to perfusion when the mechanical loading application follows irradiation.

Our interest in the MDA-B response to both radiation and bone-mimicking fluid shear stresses was due to the common use of radiation in bone metastasis treatment and its adverse effects on the skeleton. Clinical and in vivo studies show that high-dose therapeutic radiation results in decreased bone trabecular volume and increased bone marrow adiposity.^85,86 Though some studies show positive effects on bone cells (mainly osteoblasts) with low doses of irradiation (∼1–2 Gy), low doses over the course of several weeks still lead to increased bone loss and adiposity in vivo.^87,88 In the case of bone metastasis, ionizing radiation is known to alter cellular signaling pathways in both cancer cells and bone cells, with consequences that may intersect with or oppose mechanobiological regulation.^{85,86,89–91} As many pathways involved in the irradiation response are classically mechanical signaling pathways, we surmise that radiation sensitized MDA-B cells to fluid shear (PRE, Fig. 8) but suppressed the MDA-B response to fluid shear when administered postflow (POST, Fig. 8). Cell adhesion-mediated radioresistance, Wnt/β-catenin signaling, and regulation of nitric oxide expression and mechanosensitive ion channels are just some examples of classical mechanotransduction pathways that also participate in regulating the response to radiation.^91–94 Other studies have shed light on the direct effects that mechanical forces have on the cellular response to radiation. For example, shear stress applied to endothelial and colon cancer cells through a parallel-plate flow chamber increased radiation-induced nitric oxide expression and increased radiosensitivity by modulating integrin/FAK signaling pathways.⁹⁴ Clearly, the ECM micromechanical environment has a significant role in mediating the response to radiation. Nonetheless, the mechanisms explaining how mechanical stress affects the cellular radiation response, and vice versa, are still unclear.

Here, we applied a single dose of radiation to mimic the fractionated doses patients would receive in a single session, allowing us to create a model for immediate effects arising from a single dose of radiation coupled with an applied load.^7,9 When we applied ionizing radiation in this study, gene expression continued to be loading-dependent, and though MDA-B expression was irradiation dose-independent, the timing of irradiation administration significantly influenced our results. We found that preflow irradiation enabled cancer cell evasion of apoptosis and SERPINE1 downregulation, while postflow irradiation left markers of apoptosis unchanged alongside SERPINE1, reversing the MDA-B response to loading we observed in our sequencing data. Irradiation dose did not affect the breast cancer cell response to high-magnitude fluid shear, but timing did, and irradiation, regardless of sequence, may return an osteoprotective effect to loading in the bone-tumor microenvironment. Clinically, we suspect that, with multiple doses of radiation across several sessions, these effects could combine to result in sustained downregulation of protumorigenic genes like SERPINE1 by TNBC cells. Of the clinical studies that have examined the effects of radiotherapy combined with an exercise regimen, trials involving patients with spinal bone metastases found that this strategy generally improved bone density and did not further increase fracture risk.^95–97 The results reported from our in vitro study, combined with future work further interrogating these mechanisms, will be useful in providing context to these clinical data.

While this study identifies key mechanosensitive targets in bone-homing breast cancer cells, there are limitations to the current approach. First, our 3D model is unicellular, which precludes the investigation of reciprocal, bidirectional signaling between cancer cells and the bone microenvironment. In vivo, the cancer cell mechanoresponse is undoubtedly modulated by other cell types. For instance, studies using osteocyte-conditioned media have yielded varied results, reporting both reduced metastatic potential and increased expression of protumorigenic factors.^98,99 Notably, Verbruggen et al. demonstrated that these outcomes can differ significantly depending on whether cancer cells receive conditioned media or are in direct coculture, highlighting the critical importance of reciprocal signaling.¹⁰⁰ Furthermore, advanced multicellular platforms, such as the novel microfluidic chip developed by Mei et al. that integrates breast cancer cells, osteocytes, and an endothelialized lumen, are beginning to reveal how such complex environments regulate processes like intravasation. Second, our current scope was limited to characterizing the cancer cell’s transcriptional and secretory response.¹⁰¹ Thus, we did not investigate the downstream functional effects of these signals (e.g., PAI-1) on osteocyte activity or bone remodeling. Future studies utilizing transwell coculture or integrated multicellular scaffolds will be necessary to determine how these load-induced signals ultimately impact the “vicious cycle” of metastasis. Nevertheless, this work provides a foundational transcriptome-wide map of how radiotherapy and mechanical loading intersect to regulate the breast cancer cell phenotype.

Conclusions

This study provides additional evidence of the direct effects of mechanical loading on breast cancer cell radiosensitivity and demonstrates that the timing of radiation administration is crucial to modulating the response to those mechanical signals. A key implication of our study is that integrating radiotherapy with loading may restore the osteoprotective effects of loading, which are lost in bone metastasis by suppressing flow-induced upregulation of a protumorigenic and/or proresorptive profile. More involved mechanistic studies are necessary to improve our understanding of the cell machinery that is directly involved in these responses. Nevertheless, we present this work as an assertion that physiologically relevant models applying fluid shear stress on the cells involved in the bone metastatic vicious cycle are necessary to investigate how common therapies interact with the tumor micromechanical environment to enable or hinder metastatic progression.

Authors’ Contributions

S.V.: Conceptualization, data curation, formal analysis, investigation, validation, visualization, writing—original draft, writing—review and editing. A.O.: Data curation, formal analysis, writing—review and editing. W.R.T.: Conceptualization, writing—review and editing. E.B.C.: Conceptualization, project administration, resources, supervision, writing—review and editing. M.E.L.: Conceptualization, funding acquisition, project administration, resources, supervision, writing—original draft, writing—review and editing.

Footnotes

Acknowledgments

The authors acknowledge the Cancer Center Support Grant P30CA046934. Sequencing was performed through the University of Colorado Anschutz’s Genomics Shared Resource (RRID: SCR_021984). on DNA content as a function of loading condition were gathered by Dr. Boyuan Liu. Bone-homing and GFP-expressing MDA-MB-231 breast cancer cells in this study were kindly provided by Dr. Julie Rhoades at Vanderbilt University, Nashville, TN, USA. The authors would also like to thank Dr. Tin Tin Su for allowing them to use her irradiator to complete the experiments in this study.

Declaration of Generative AI in the Writing Process

During the revision of this article, the corresponding author used Gemini (Google) for language editing and to improve the clarity and flow of the Discussion section and responses to reviewer comments. All AI-generated suggestions were critically reviewed and verified by the authors. The AI was not used for data collection, analysis, or generation of original scientific conclusions, for which the human authors maintain full responsibility.

Data Availability Statement

The sequencing data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus^102,103 (GEO) and are accessible through the GEO Series accession number GSE310741 ().

Disclosure Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding Information

This study was supported by NSF BMMB-2047187 and the Achievement Rewards for College Scientists (ARCS).

Supplemental Material

References

Svensson

, Christiansen

, Ulrichsen

, et al. Survival after bone metastasis by primary cancer type: A Danish population-based cohort study. BMJ Open, 2017; 7(9):e016022; doi: 10.1136/bmjopen-2017-016022

Yong

, Jensen

AÖ

, Jacobsen

, et al. Survival in breast cancer patients with bone metastases and skeletal-related events: A population-based cohort study in Denmark (1999-2007). Breast Cancer Res Treat, 2011; 129(2):495–503; doi: 10.1007/s10549-011-1475-5

Guise

. The vicious cycle of bone metastases. J Musculoskelet Neuronal Interact, 2002; 2(6):570–572.

Zacharia

, Joy

, Subramaniam

, et al. Factors affecting life expectancy after bone metastasis in adults — results of a 5-year prospective study. Indian J Surg Oncol, 2021; 12(4):759–769; doi: 10.1007/s13193-021-01426-1

Arora

, Chew

. Breast Cancer Bone Metastases. In: Metastatic Bone Disease: An Integrated Approach to Patient Care. ( Randall

., ed.) Springer International Publishing: Cham; 2024; pp. 63–71; doi: 10.1007/978-3-031-52001-3_5

Marazzi

, Orlandi

, Manfrida

, et al. Diagnosis and treatment of bone metastases in breast cancer: Radiotherapy, local approach and systemic therapy in a guide for clinicians. Cancers (Basel), 2020; 12(9):2390; doi: 10.3390/cancers12092390

Kim

, Fahmy

, Haffty

. Radiation therapy for triple-negative breast cancer: From molecular insights to clinical perspectives. Expert Rev Anticancer Ther, 2024; 24(5):211–217; doi: 10.1080/14737140.2024.2333320

Jiao

, Cao

, Liu

. Radiation-induced cell death and its mechanisms. Health Phys, 2022; 123(5):376–386; doi: 10.1097/HP.0000000000001601

de Faria Bessa

, Marta

. Triple-negative breast cancer and radiation therapy. Rep Pract Oncol Radiother, 2022; 27(3):545–551; doi: 10.5603/RPOR.a2022.0025

10.

Pang

, Gan

, Xu

, et al. Bone metastasis of breast cancer: Molecular mechanisms and therapeutic strategies. Cancers (Basel), 2022; 14(23):5727; doi: 10.3390/cancers14235727

11.

Courneya

. The emerging role of exercise as a cancer treatment. J Sport Health Sci, 2024; 13(4):443–444; doi: 10.1016/j.jshs.2023.03.006

12.

Turner

, Robling

. Exercise as an anabolic stimulus for bone. Curr Pharm Des, 2004; 10(21):2629–2641; doi: 10.2174/1381612043383755

13.

Robling

, Turner

. Mechanical signaling for bone modeling and remodeling. Crit Rev Eukaryot Gene Expr, 2009; 19(4):319–338.

14.

Coughlin

, Niebur

. Fluid shear stress in trabecular bone marrow due to low-magnitude high-frequency vibration. J Biomech, 2012; 45(13):2222–2229; doi: 10.1016/j.jbiomech.2012.06.020

15.

Ozcivici

, Luu

, Adler

, et al. Mechanical signals as anabolic agents in bone. Nat Rev Rheumatol, 2010; 6(1):50–59; doi: 10.1038/nrrheum.2009.239

16.

Lynch

, Brooks

, Mohanan

, et al. In vivo tibial compression decreases osteolysis and tumor formation in a human metastatic breast cancer model. J Bone Miner Res, 2013; 28(11):2357–2367; doi: 10.1002/jbmr.1966

17.

Wang

, Pei

, Wasi

, et al. Moderate tibial loading and treadmill running, but not overloading, protect adult murine bone from destruction by metastasized breast cancer. Bone, 2021; 153:116100; doi: 10.1016/j.bone.2021.116100

18.

Pagnotti

, Chan

, Adler

, et al. Low intensity vibration mitigates tumor progression and protects bone quantity and quality in a murine model of myeloma. Bone, 2016; 90:69–79; doi: 10.1016/j.bone.2016.05.014

19.

Huang

, Liu

, Zhu

, et al. Mechanical loading attenuates breast cancer-associated bone metastasis in obese mice by regulating the bone marrow microenvironment. J Cell Physiol, 2021; 236(9):6391–6406; doi: 10.1002/jcp.30314

20.

Rummler

, Ziouti

, Bouchard

, et al. Mechanical loading prevents bone destruction and exerts anti-tumor effects in the MOPC315.BM.Luc model of myeloma bone disease. Acta Biomater, 2021; 119:247–258; doi: 10.1016/j.actbio.2020.10.041

21.

Yang

, Liu

, Zhu

, et al. Ankle loading ameliorates bone loss from breast cancer–associated bone metastasis. Faseb J, 2019; 33(10):10742–10752; doi: 10.1096/fj.201900306RR

22.

Bellido

. Osteocyte-driven bone remodeling. Calcif Tissue Int, 2014; 94(1):25–34; doi: 10.1007/s00223-013-9774-y

23.

Bonewald

. The amazing osteocyte. J Bone Miner Res, 2011; 26(2):229–238; doi: 10.1002/jbmr.320

24.

Bonewald

. Mechanosensation and transduction in osteocytes. Bonekey Osteovision, 2006; 3(10):7–15; doi: 10.1138/20060233

25.

Wang

, Sarazin

, Kornilowicz

, et al. Mechanically-loaded breast cancer cells modify osteocyte mechanosensitivity by secreting factors that increase osteocyte dendrite formation and downstream resorption. Front Endocrinol (Lausanne), 2018; 9:352; doi: 10.3389/fendo.2018.00352

26.

Sarazin

, Liu

, Goldman

, et al. Bone-homing metastatic breast cancer cells impair osteocytes’ mechanoresponse in a 3D loading model. Heliyon, 2023; 9(10):e20248; doi: 10.1016/j.heliyon.2023.e20248

27.

Lynch

, Chiou

, Lee

, et al. Three-dimensional mechanical loading modulates the osteogenic response of mesenchymal stem cells to tumor-derived soluble signals. Tissue Eng Part A, 2016; 22(15–16):1006–1015; doi: 10.1089/ten.tea.2016.0153

28.

Kumar

, Naqvi

, Verbruggen

, et al. A novel mechanobiological model of bone metastasis reveals that mechanical stimulation inhibits the pro-osteoclastogenic effects of breast cancer cells. Cell Rep, 2024; 43(5):114043; doi: 10.1016/j.celrep.2024.114043

29.

Metzger

, Vaughan

, McNamara

, et al. Altered architecture and cell populations affect bone marrow mechanobiology in the osteoporotic human femur. Biomech Model Mechanobiol, 2017; 16(3):841–850; doi: 10.1007/s10237-016-0856-4

30.

Metzger

, Schwaner

, LaNeve

, et al. Pressure and shear stress in trabecular bone marrow during whole bone loading. J Biomech, 2015; 48(12):3035–3043; doi: 10.1016/j.jbiomech.2015.07.028

31.

Metzger

, Kreipke

, Vaughan

, et al. The in situ mechanics of trabecular bone marrow: The potential for mechanobiological response. J Biomech Eng, 2015; 137(1); doi: 10.1115/1.4028985

32.

Birmingham

, Grogan

, Niebur

, et al. Computational modelling of the mechanics of trabecular bone and marrow using fluid structure interaction techniques. Ann Biomed Eng, 2013; 41(4):814–826; doi: 10.1007/s10439-012-0714-1

33.

Birmingham

, Kreipke

, Dolan

, et al. Mechanical stimulation of bone marrow in situ induces bone formation in trabecular explants. Ann Biomed Eng, 2015; 43(4):1036–1050; doi: 10.1007/s10439-014-1135-0

34.

Lorenzen

, Sinkus

, Lorenzen

, et al. MR elastography of the breast: Preliminary clinical results. Rofo, 2002; 174(7):830–834; doi: 10.1055/s-2002-32690

35.

Kim

, Kim

, Jeong

, et al. Magnetic resonance elastography of invasive breast cancer: Evaluating prognostic factors and treatment response. Tomography, 2025; 11(2):18; doi: 10.3390/tomography11020018

36.

Silver

., (ed). Viscoelastic Mechanical Properties of Tissues. In : Mechanosensing and Mechanochemical Transduction in Extracellular Matrix: Biological, Chemical, Engineering, and Physiological Aspects. Springer US: Boston, MA; 2006; pp. 181–198; doi: 10.1007/978-0-387-28176-6_7

37.

Rho

, Ashman

, Turner

. Young’s modulus of trabecular and cortical bone material: Ultrasonic and microtensile measurements. J Biomech, 1993; 26(2):111–119; doi: 10.1016/0021-9290(93)90042-D

38.

Novak

, Horst

, Taylor

, et al. Fluid shear stress stimulates breast cancer cells to display invasive and chemoresistant phenotypes while upregulating PLAU in a 3D bioreactor. Biotechnol Bioeng, 2019; 116(11):3084–3097; doi: 10.1002/bit.27119

39.

Fuh

, Shepherd

, Withell

, et al. Fluid flow exposure promotes epithelial-to-mesenchymal transition and adhesion of breast cancer cells to endothelial cells. Breast Cancer Res, 2021; 23(1):97; doi: 10.1186/s13058-021-01473-0

40.

Riehl

, Kim

, Lee

, et al. The role of fluid shear and metastatic potential in breast cancer cell migration. J Biomech Eng, 2020; 142(10):101001; doi: 10.1115/1.4047076

41.

Mitchell

, King

. Computational and experimental models of cancer cell response to fluid shear stress. Front Oncol, 2013; 3:44; doi: 10.3389/fonc.2013.00044

42.

Eren

, Place

, Thomas

, et al. PAI-1 is a critical regulator of FGF23 homeostasis. Sci Adv, 2017; 3(9):e1603259; doi: 10.1126/sciadv.1603259

43.

Huang

, Wang

, Guo

, et al. Single‐cell RNA sequencing reveals SERPINE1‐expressing CAFs remodelling tumour microenvironment in recurrent osteosarcoma. Clin Transl Med, 2024; 14(1):e1527; doi: 10.1002/ctm2.1527

44.

Wei

, Li

, He

, et al. Tumor-secreted PAI-1 promotes breast cancer metastasis via the induction of adipocyte-derived collagen remodeling. Cell Commun Signal, 2019; 17(1):58; doi: 10.1186/s12964-019-0373-z

45.

Y-H

, Wu

Y-Z

, Ann

, et al. Obesity promotes radioresistance through SERPINE1-mediated aggressiveness and DNA repair of triple-negative breast cancer. Cell Death Dis, 2023; 14(1):53; doi: 10.1038/s41419-023-05576-8

46.

Zhang

, Lei

, Jing

. Knockdown of SERPINE1 reverses resistance of triple-negative breast cancer to paclitaxel via suppression of VEGFA. Oncol Rep, 2020; 44(5):1875–1884; doi: 10.3892/or.2020.7770

47.

Dole

, Mazur

, Acevedo

, et al. Osteocyte-intrinsic TGF-β signaling regulates bone quality through perilacunar/canalicular remodeling. Cell Rep, 2017; 21(9):2585–2596; doi: 10.1016/j.celrep.2017.10.115

48.

Pathi

, Kowalczewski

, Tadipatri

, et al. A novel 3-D mineralized tumor model to study breast cancer bone metastasis. PLoS One, 2010; 5(1):e8849; doi: 10.1371/journal.pone.0008849

49.

Venkatesh

, Teeraratkul

, Rovito

, et al. High-fidelity computational fluid dynamics modeling to simulate perfusion through a bone-mimicking scaffold. Comput Biol Med, 2025; 186:109637; doi: 10.1016/j.compbiomed.2024.109637

50.

Pathi

, Lin

DDW

, Dorvee

, et al. Hydroxyapatite nanoparticle-containing scaffolds for the study of breast cancer bone metastasis. Biomaterials, 2011; 32(22):5112–5122; doi: 10.1016/j.biomaterials.2011.03.055

51.

, Jung

, Yao

. ShinyGO: A graphical gene-set enrichment tool for animals and plants. Bioinformatics, 2020; 36(8):2628–2629; doi: 10.1093/bioinformatics/btz931

52.

Bardou

, Mariette

, Escudié

, et al. jvenn: An interactive Venn diagram viewer. BMC Bioinformatics, 2014; 15(1):293; doi: 10.1186/1471-2105-15-293

53.

Yokoyama

, Mori

, Hamada

, et al. Platelet-derived growth factor regulates breast cancer progression via β-catenin expression. Pathobiology, 2011; 78(5):253–260; doi: 10.1159/000328061

54.

Zhang

, Chen

, Jin

. Platelet-Derived Growth Factor (PDGF)-BB stimulates osteoclastic bone resorption directly: The role of receptor β. Biochem Biophys Res Commun, 1998; 251(1):190–194; doi: 10.1006/bbrc.1998.9412

55.

Omata

, Nakamura

, Koyama

, et al. Identification of Nedd9 as a TGF-β-Smad2/3 target gene involved in RANKL-induced osteoclastogenesis by comprehensive analysis. PLoS One, 2016; 11(6):e0157992; doi: 10.1371/journal.pone.0157992

56.

Komalasari

NLGY

, Tomonobu

, Kinoshita

, et al. Lysyl oxidase-like 4 exerts an atypical role in breast cancer progression that is dependent on the enzymatic activity that targets the cell-surface annexin A2. Front Oncol, 2023; 13:1142907; doi: 10.3389/fonc.2023.1142907

57.

Gartland

, Erler

, Cox

. The role of lysyl oxidase, the extracellular matrix and the pre-metastatic niche in bone metastasis. J Bone Oncol, 2016; 5(3):100–103; doi: 10.1016/j.jbo.2016.04.001

58.

Varga

, Wrana

. The disparate role of BMP in stem cell biology. Oncogene, 2005; 24(37):5713–5721; doi: 10.1038/sj.onc.1208919

59.

Tarragona

, Pavlovic

, Arnal-Estapé

, et al. Identification of NOG as a specific breast cancer bone metastasis-supporting gene. J Biol Chem, 2012; 287(25):21346–21355; doi: 10.1074/jbc.M112.355834

60.

Musavi

, Kohram

, Abasi

, et al. Rn7SK small nuclear RNA is involved in cellular senescence. J Cell Physiol, 2019; 234(8):14234–14245; doi: 10.1002/jcp.28119

61.

, Ma

, Zhang

, et al. A positive feedback circuit between RN7SK snRNA and m6A readers is essential for tumorigenesis. Mol Ther, 2023; 31(6):1615–1635; doi: 10.1016/j.ymthe.2022.12.013

62.

Liu

, Luo

. NEDD9 is transcriptionally regulated by HDAC4 and promotes breast cancer metastasis and macrophage M2 polarization via the FAK/NF-κB signaling pathway. Neoplasia, 2024; 57:101059; doi: 10.1016/j.neo.2024.101059

63.

Purazo

, Ice

, Shimpi

, et al. NEDD9 overexpression causes hyperproliferation of luminal cells and cooperates with HER2 oncogene in tumor initiation: A novel prognostic marker in breast cancer. Cancers (Basel), 2023; 15(4):1119; doi: 10.3390/cancers15041119

64.

Orimo

, Weinberg

. Stromal fibroblasts in cancer: A novel tumor-promoting cell type. Cell Cycle, 2006; 5(15):1597–1601; doi: 10.4161/cc.5.15.3112

65.

Liu

, Shen

J-X

, Wu

H-T

, et al. Collagen 1A1 (COL1A1) promotes metastasis of breast cancer and is a potential therapeutic target. Discov Med, 2018; 25(139):211–223.

66.

Gibert

, Eckel

, Gonin

, et al. Targeting heat shock protein 27 (HspB1) interferes with bone metastasis and tumour formation in vivo. Br J Cancer, 2012; 107(1):63–70; doi: 10.1038/bjc.2012.188

67.

, Wang

. HspA1B is a prognostic biomarker and correlated with immune infiltrates in different subtypes of breast cancers. bioRxiv, 2019; 725861; doi: 10.1101/725861

68.

Sohail

, Sun

, Zhao

, et al. MT4-(MMP17) and MT6-MMP (MMP25), A unique set of membrane-anchored matrix metalloproteinases: Properties and expression in cancer. Cancer Metastasis Rev, 2008; 27(2):289–302; doi: 10.1007/s10555-008-9129-8

69.

Lynch

. Matrix metalloproteinases as master regulators of the vicious cycle of bone metastasis. Bone, 2011; 48(1):44–53; doi: 10.1016/j.bone.2010.06.007

70.

Juurikka

, Butler

, Salo

, et al. The role of MMP8 in cancer: A systematic review. Int J Mol Sci, 2019; 20(18):4506; doi: 10.3390/ijms20184506

71.

Shaikh

, Balaya

RDA

, Dagamajalu

, et al. A signaling pathway map of plasminogen activator inhibitor-1 (PAI-1/SERPINE-1): A review of an innovative frontier in molecular aging and cellular senescence. Cell Commun Signal, 2024; 22(1):544; doi: 10.1186/s12964-024-01910-5

72.

Hongu

, Kusunoki

, Ishimura

, et al. Permeable lung vasculature creates chemoresistant endothelial niche by producing SERPINE1 at breast cancer metastatic sites. Cancer Sci, 2025; 116(6):1604–1615; doi: 10.1111/cas.70050

73.

Lee

J-Y

, Kim

D-A

, Kim

E-Y

, et al. Lumican inhibits osteoclastogenesis and bone resorption by suppressing akt activity. Int J Mol Sci, 2021; 22(9):4717; doi: 10.3390/ijms22094717

74.

Okada

, Okamoto

, Okumoto

, et al. PAI-1 is involved in delayed bone repair induced by glucocorticoids in mice. Bone, 2020; 134:115310; doi: 10.1016/j.bone.2020.115310

75.

Chaffer

, Weinberg

. A perspective on cancer cell metastasis. Science, 2011; 331(6024):1559–1564; doi: 10.1126/science.1203543

76.

Shieh

. Biomechanical forces shape the tumor microenvironment. Ann Biomed Eng, 2011; 39(5):1379–1389; doi: 10.1007/s10439-011-0252-2

77.

Gobin

, Bagwell

, Wagner

, et al. A pan-cancer perspective of matrix metalloproteases (MMP) gene expression profile and their diagnostic/prognostic potential. BMC Cancer, 2019; 19(1):581; doi: 10.1186/s12885-019-5768-0

78.

Falcão

, Motta

, de Lima

, et al. Enhancement of viability of radiosensitive (PBMC) and resistant (MDA-MB-231) clones in low-dose-rate cobalt-60 radiation therapy. Radiol Bras, 2015; 48(3):158–165; doi: 10.1590/0100-3984.2014.0022

79.

Panzetta

, La Verde

, Pugliese

, et al. Adhesion and migration response to radiation therapy of mammary epithelial and adenocarcinoma cells interacting with different stiffness substrates. Cancers (Basel), 2020; 12(5):1170; doi: 10.3390/cancers12051170

80.

Mahmoud

, Casciati

, Bakar

, et al. The detection of DNA damage response in MCF7 and MDA-MB-231 breast cancer cell lines after X-ray exposure. Genome Integr, 2023; 14:e20220001; doi: 10.14293/genint.14.1.001

81.

Nutter

, Holen

, Brown

, et al. Different molecular profiles are associated with breast cancer cell homing compared with colonisation of bone: Evidence using a novel bone-seeking cell line. Endocr Relat Cancer, 2014; 21(2):327–341; doi: 10.1530/ERC-13-0158

82.

Choi

Y-J

, Baek

G-Y

, Park

H-R

, et al. Smad2/3-regulated expression of DLX2 is associated with radiation-induced epithelial-mesenchymal transition and radioresistance of A549 and MDA-MB-231 human cancer cell lines. PLoS One, 2016; 11(1):e0147343; doi: 10.1371/journal.pone.0147343

83.

Miser-Salihoğlu

, Karahalil

, Eşmakaya

, et al. The effect of silencing the Tip60 gene on the response to radiotherapy in breast cancer cells. Breast, 2023; 69:507–512; doi: 10.1016/j.breast.2023.04.001

84.

McKenna

, Weiss

, Bakanauskas

, et al. The role of the H-ras oncogene in radiation resistance and metastasis. Int J Radiat Oncol Biol Phys, 1990; 18(4):849–859; doi: 10.1016/0360-3016(90)90407-B

85.

Berk

. The effects of high-dose radiation therapy on bone: A scoping review. Radiat Oncol J, 2024; 42(2):95–103; doi: 10.3857/roj.2023.00969

86.

Costa

, Reagan

. Therapeutic irradiation: Consequences for bone and bone marrow adipose tissue. Front Endocrinol (Lausanne), 2019; 10:587; doi: 10.3389/fendo.2019.00587

87.

Sun

, Zhu

, Wang

, et al. Indirect effects of X-irradiation on proliferation and osteogenic potential of bone marrow mesenchymal stem cells in a local irradiated rat model. Mol Med Rep, 2017; 15(6):3706–3714; doi: 10.3892/mmr.2017.6464

88.

Bokhari

, Metzger

, Black

, et al. Positive impact of low-dose, high-energy radiation on bone in partial- and/or full-weightbearing mice. NPJ Microgravity, 2019; 5(1):13–19; doi: 10.1038/s41526-019-0074-3

89.

Pacheco

, Stock

. Effects of radiation on bone. Curr Osteoporos Rep, 2013; 11(4):299–304; doi: 10.1007/s11914-013-0174-z

90.

Bravatà

, Minafra

, Cammarata

, et al. Gene expression profiling of breast cancer cell lines treated with proton and electron radiations. Br J Radiol, 2018; 91(1089):20170934; doi: 10.1259/bjr.20170934

91.

Mottareale

, Frascogna

, La Verde

, et al. Impact of ionizing radiation on cell-ECM mechanical crosstalk in breast cancer. Front Bioeng Biotechnol, 2024; 12:1408789; doi: 10.3389/fbioe.2024.1408789

92.

Deng

, Lin

, Nowsheen

, et al. Extracellular matrix stiffness determines DNA repair efficiency and cellular sensitivity to genotoxic agents. Sci Adv, 2020; 6(37):eabb2630; doi: 10.1126/sciadv.abb2630

93.

Lacombe

, Zenhausern

. Effect of mechanical forces on cellular response to radiation. Radiother Oncol, 2022; 176:187–198; doi: 10.1016/j.radonc.2022.10.006

94.

Luo

C-W

, Wu

C-C

, Ch’ang

H-J

. Radiation sensitization of tumor cells induced by shear stress: The roles of integrins and FAK. Biochim Biophys Acta, 2014; 1843(9):2129–2137; doi: 10.1016/j.bbamcr.2014.06.007

95.

Taneichi

, Kaneda

, Takeda

, et al. Risk factors and probability of vertebral body collapse in metastases of the thoracic and lumbar spine. Spine (Phila Pa 1976), 1997; 22(3):239–245.

96.

Rief

, Petersen

, Omlor

, et al.; German Bone Research Group. The effect of resistance training during radiotherapy on spinal bone metastases in cancer patients–A randomized trial. Radiother Oncol, 2014; 112(1):133–139; doi: 10.1016/j.radonc.2014.06.008

97.

Sprave

, Rosenberger

, Verma

, et al. Paravertebral muscle training in patients with unstable spinal metastases receiving palliative radiotherapy: An exploratory randomized feasibility trial. Cancers (Basel), 2019; 11(11):1771; doi: 10.3390/cancers11111771

98.

, Prideaux

, Bonewald

, et al. Osteocytes and cancer. Curr Osteoporos Rep, 2021; 19(6):616–625; doi: 10.1007/s11914-021-00712-9

99.

Atkinson

, Delgado‐Calle

. The emerging role of osteocytes in cancer in bone. JBMR Plus, 2019; 3(3):e10186; doi: 10.1002/jbm4.10186

100.

Verbruggen

, Nolan

, Duffy

, et al. A novel primary cilium-mediated mechanism through which osteocytes regulate metastatic behavior of both breast and prostate cancer cells. Adv Sci (Weinh), 2024; 11(2):e2305842; doi: 10.1002/advs.202305842

101.

Mei

, Middleton

, Shim

, et al. Microfluidic platform for studying osteocyte mechanoregulation of breast cancer bone metastasis. Integr Biol (Camb), 2019; 11(4):119–129; doi: 10.1093/intbio/zyz008

102.

Edgar

, Domrachev

, Lash

. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res, 2002; 30(1):207–210; doi: 10.1093/nar/30.1.207

103.

Barrett

, Wilhite

, Ledoux

, et al. NCBI GEO: archive for functional genomics data sets—update. Nucleic Acids Res, 2013; 41(Database issue):D991–D995; doi: 10.1093/nar/gks1193

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.05 MB

0.23 MB

0.19 MB

0.23 MB

0.24 MB