Metastasis, a complex multistep event

Breast cancer (BC) metastasis is the leading cause of cancer related deaths among women (metastasis; the spread of tumor cells from primary site to distant organs) [1, 2]. More than 70 % of BC patients in advanced stage develop bone metastases which are related to high rate of morbidity and mortality [3, 4]. The vast majority of researchers have focused to identify the underlying molecular and cellular mechanisms of this complex phenomenon. Developing gene expression signatures, involved in certain stages of BC metastasis, are among seminal discoveries that may let to understand the basics involved in the ‘invasion-metastasis cascade’, and consequently the generation of effective therapies against metastasis.

‘Invasion-metastasis cascade’ represents a multistep process that consists a set of sequential events, in which tumor cells invade from primary site, disseminate through circulation, and reconstitute secondary tumors at distant tissue(s) [5]. It has been investigated that breast tumor cells encounter a plenty of changes when disseminating from primary to distant sites, including changes in gene expression pattern [6], and in the states of stemness [7], that define clonal evolution and cancer stem cell (CSC) theories, respectively. Genetic alterations and subsequent shift in cellular events and states must be considered in the context of different microenvironments in different steps of metastasis in order to uncover mysteries of the complex process of metastasis cascade.

Clonal evolution and CSC theories are two models for tumor heterogeneity, cancer development, and metastasis that are widely accepted [8, 9]. Clonal evolution theory indicates that different lineages of cancer cells are developed during multistep genetic and epigenetic alterations, which lead to acquiring different features required for tumor development and metastasis [10]. On the other hand, CSCs, which have features of self-renewal, tumorigenesis, multilineage differentiation, motility, invasiveness and apoptosis resistance, are believed to be required for the development and maintenance of several forms of human cancers, including BC. Based on the CSC theory, tumor cells are not naturally alike, based on the state of stemness, and are organized in a hierarchical pattern in which CSCs are considered to be at the top of the apex [11–13]. It is believed that both models of clonal evolution and CSC can be applied in cancer development and metastasis, and that tumors heterogeneity can be generated from both of them [8].

The pattern of gene expression is likely the most critical determinant of CSC state. A quite sophisticated program that leads to expression of a group of genes, and simultaneously suppression of others, directs all features of a tumor cell in a precise time. Gene expression signatures have been developed since more than a decade for a variety of diseases like cancer metastasis. Gene signatures that have been developed for BC metastasis are ostensibly behind-the-scene forces of cellular events. In this study, we provide most recently discovered functions of sets of genes that are selected from seminal widely accepted gene signatures. We then propose a possible sequential pattern of gene expression that may lead to a benign primary breast tumor to get aggressiveness and progress toward bone metastasis. We also discuss most prominent molecular mechanisms involved in BC bone metastasis.

Gene signatures associated to breast cancer progression and bone metastasis

A number of seminal studies have uncovered gene expression signatures involved in BC development and metastasis; each of them points at a distinct step of the ‘invasion-metastasis cascade’ (Tables 1, 2, 3, 4) [14–16]. Today, these findings have entered to the diagnosis as predictors of disease outcome in BC patients [17]. Particularly, such discoveries have heralded the new era of personalized medicine, while predicting the clinical outcome of patients based on a set of distinct gene expression patterns [18]. Although improving, our understanding of the exact molecular and, most importantly, cellular mechanisms of BC metastasis is poor, and therefore reliable treatments are lacking. Analysis of data resulting from high throughput genome wide assays, and translation of the molecular pattern to cellular mechanisms/pathways may provide novel perspective to understand the complex nature of metastasis, and subsequently develop new therapeutic strategies. Among studies that have provided gene signatures for BC progression and bone metastasis, the ones by van’t Veer et al., Smid et al., and Kang et al. are of most seminal and widely accepted.

Table 1 Gene signatures associated to breast cancer metastasis
Full size table
Table 2 van’t Veer’s signature of breast cancer poor/good prognosis genes
Full size table
Table 3 Smid’s signature of breast cancer bone metastasis genes
Full size table
Table 4 Kang’s signature of breast cancer bone metastasis genes
Full size table

van’t Veer’s signature

Several studies regarding gene expression pattern in BC have been developed since more than a decade (Table 1). Earliest ones [19–23] were not quite sufficient to be utilized for predictive and therapeutic purposes. This may be because of the inconsistency in different studies (e.g., different kinds of primary tumors), and abundance of heterogeneity in tumors. The first highly applicable reported gene signature, by van’t Veer and colleagues (van’t Veer’s signature), have established a 70-gene prognosis profile from primary tumors of young BC patients. (Table 2 comprises a panel of selected genes from van’t Veer’s signature, based on their significant expression in most poor prognosed patients, and the status of being well studied). They identified a set of genes strongly predicting distant metastasis in patients who were lymph node negative, called poor prognosis signature [14]. Their finding uncovered a pattern of gene expression required for primary tumor cells to become invasive; capable to evade from primary site.

van’t Veer’s signature comprises genes associated to a well-orchestrated program for the regulation of different features required for primary tumors to grow and escape from primary site, including cell cycle, DNA replication, proliferation, tumorigenesis, survival, angiogenesis, migration, and invasion (Table 2). In particular, among those, CCNB2, CCNE2, MCM6, TSPYL5, NUSAP1, CMC2, ECT2, ORC6, DTL, PRC1, MELK, EGLN1, SLC2A3, RAB6B, ESM1, RAD21, CDC25B, CDK16, CENPA, PGK1, MAD2L1, CKS2, BUB1, FGF18, WISP1, and IGFBP5 are known regulators of cell cycle, DNA replication, proliferation, and survival [24–26]. Upregulation of these genes can be considered as the first requirements of primary tumor for its growth in order to be prepared for dissemination. Afterward, ECT2, EGLN1, ESM1, FLT1, EXT1, DIAPH3, EXOC7, NMU, CDC42BPA, VEGF, MMP9, FGF18, WISP1, and TGF-β3, are well-known pivotal elements that participate in angiogenesis, migration, and invasion. Tumor cells which express these genes seem to be adept for invasion from primary site (please see Table 2 for detail functions of gene products).

A variety of key molecular mechanisms are regulated by van’t Veer’s signature gene products. Remarkably, TGF-β signaling pathway likely plays important roles in the regulation of cell cycle, proliferation, induction of EMT, CSCs, and MMPs [27–30] (Table 2). Furthermore, NFκB signaling is also involved in this level of tumor progression taking part in tumor growth and metastasis by the mediation of ESM1, FGF18, and WISP1 [31–33]. In this step, the transcription factor MYC, which is well known for its association to breast tumor proliferation [34, 35], may also play important roles as is shown to be regulated by at least two of van’t Veer’s signature genes including RAD21 and CDC25B [36, 37]. Notably, ‘epithelial to mesenchymal transition’ (EMT), a process that is shown to be essential for tumor dissemination and metastasis of breast carcinomas [38], is enhanced through at least two of genes in this signature including MMP9 and VEGF. These two factors play determining roles in preparing a hospitable microenvironment in which emitted signals trigger EMT in order to induce/maintain CSCs [39–42]. On the other hand, two of genes in this panel may tend to trigger primary tumor cells to metastasize to bone, including FLT1 and PGK1 [43, 44]. PGK1 increases the expression of CXCR4, which is one of the most important bone metastasis factors [45]. FLT1 also provide a premetastatic niche in bone and direct bone metastasis of BC [43]. Together, above information suggest the involvement of key regulators like TGF-β, NFκB, and MYC, as well as the process of EMT in first steps of BC metastasis within the primary tumor.

Biological functions of genes in van’t Veer’s signature define the hallmarks of cancer. Tian and colleagues have recently shown that van’t Veer’s signature gene products functionally meet all the six hallmarks of cancer defined by Hanahan and Weinberg, including sustained proliferation, anti-growth signaling evasion, cell death resistance, immortality, angiogenesis, and invasion/metastasis [46]. They identified interconnected networks and showed that these genes are regulated by key tumorigenic factors like TP53, RB1, MYC, JUN and CDKN2A [47]. Interestingly, van’t Veer’s signature may also reflect the two additional hallmarks of next generation, including reprogramming of energy metabolism and evading immune destruction [1].

Adjustments of energy metabolism in order to fuel cell growth and division, is of most important features that lead to uncontrolled proliferation in neoplasms. Glycolytic fueling has been shown to be one of the most essential mechanisms in the reprogramming of energy, and associated with activated oncogenes like RAS, MYC, mutant tumor suppressors like TP53, certain signaling pathways like PI3K/Akt/PTEN, and hypoxia inducible factor 1(HIF1) [48, 49]. Now, as Tian and colleagues showed, and also from functions of certain genes including TSPYL5, CMC2, CDC25B, EGLN1, SLC2A3, RAB6B, and TGF-β3 (see details of functions in Table 2), van’t Veer’s signature likely associates with the seventh hallmark of cancer, reprogramming of energy metabolism. As for the eighth hallmark, evading immune destruction, it has been shown that tumors that produce transforming growth factor (TGF)-β escape from immune surveillance, mainly by selective and direct suppression of the T cell cytotoxic gene responses [50]. Intriguingly, TGF-β signaling is undeniably of key factors in van’t Veer’s signature. Together, it seems that functions of gene products of van’t Veer’s signature also meet the two additional next generation hallmarks of cancer, in addition to the first six ones.

Smid’s signature

Focusing on bone metastasis, Smid and colleagues have established a panel of genes in BC patients that are implicated to bone relapse (Smid’s signature). They analyzed primary tumors of lymph node negative BC patients who, subsequently, had developed metastases. A set of 69 genes was identified to be differentially expressed in patients who had experienced bone metastasis versus patients with metastasis to other sites. (Table 3 comprises a panel of selected genes from Smid’s signature that are significantly overexpressed and have been better studied). Notably, they developed classifier of tumors that metastasize to bone that was applicable in clinic [16].

Smid’s signature provides a pattern of gene expression in primary tumors that obligates them to metastasize to bone. Genes in Smid’s signature participate in essential features of metastasis including tumor growth, proliferation, survival, angiogenesis, migration, and invasion (Table 3). Among those, TFF1, TFF3, AGR2, NAT1, CRIP1, TSPAN1, FGFR3, CEACAM6, and TMSB15A may be categorized to play roles in cell growth/proliferation, and survival. Importantly, a number of genes in this signature take part in angiogenesis that include TFF1, TFF3, FGFR3, and FGFBP1. Thereupon, the vast majority of the genes in this panel which have seminal roles in migration and invasion are reported to be TFF1, TFF3, AGR2, NAT1, RND1, TSPAN1, FGFR3, CEACAM6, KRT16, FGFBP1, FOXO3A, KRT6B, and SNAI1. Interestingly, TFF1, TFF3, and FGFR3 are present in all above categories, and seemingly play pivotal roles in the process of BC metastasis to bone.

Well-characterized crucial molecular mechanisms are controlled by Smid’s signature gene products. Trefoil factor-1 (TFF1), that was the most differentially expressed gene associated to bone metastasis in Smid’s signature, is ascertained to play roles in cell survival, anchorage-independent growth, angiogenesis, migration, and invasion in breast (and also other) tumor cells [51–54]. TFF3 induces angiogenesis by the regulation of VEGF, under the control of hypoxia [55]. AGR2, NAT1, FGFR3, and TSPAN1 have shown to play roles in tumor growth/proliferation/survival and/or migration/invasion. CRIP1, CEACAM6, and TMSB15A regulate tumor growth/proliferation and survival. RND1, KRT16, KRT6B, FOXO3A, and SNAI1 induce migration/invasion. It should be noted that, SNAI1 (also SNAIL) has determined to play pivotal roles as a master regulator of EMT [38, 56–58]. On the contrary, SCUBE2, and FOXO3A have generally shown to encode suppressors of tumor growth/proliferation, inducers of apoptosis, and repressor of angiogenesis. SCUBE2, and FOXO3A, although, are distinguished to play roles in line with induction of invasiveness, and metastasis, by regulation of Hedgehog signaling and matrix metalloproteinases, respectively [59–62] (Table 3).

Critical signaling pathways are associated with the genes in the Smid’s signature. Apparently, fibroblast growth factor (FGF) signaling takes fundamental parts in this circuit, as FGFR3 and FGFBP1 that are directly linked to this pathway are two important overexpressed genes in this panel. ERK signaling is likely involved through activation by FGFBP1 [59]. In addition, epidermal growth factor (EGF) and Janus kinase-signal transducers and activators of transcription (JAK-STAT) signaling pathways are demonstrated to be enhanced by TOM1L1 [63]. Together, from the above mentioned findings, it seems that the genes in Smid’s signature direct a complex program in which tumor cells proliferate and survive, then further induce and maintain angiogenesis, and finally enhance migration and invasion. Passing these two steps of gene expression (van’t Veer’s and Smid’s signatures), tumor cells are capable to invade, intravasate, and survive in circulation, heading to bone. These two sets of genes may be regulated in an overlapped, or in a sequential pattern.

Kang’s signature

Focused on the gene expression pattern of breast tumor cells heading to bone, Kang and colleagues investigated a multigenic program in highly aggressive osteolytic BC metastatic cells (Kang’s signature) (Table 4 comprises a panel of selected genes from Kang’s signature that are significantly overexpressed and have been better studied) [15]. Their work provided a framework for the identification of genes mediating metastasis to different organs. Although Kang’s signature was established in animal model, it has recently been confirmed in human BC patients as well [64]. These genes mostly encode secreted and cell membrane proteins, and are associated to the preparation of a compatible metastatic niche.

Kang’s signature provides a panel of genes that may be categorized into four groups including: angiogenesis, migration/invasion, EMT, and growth/angiogenesis inhibition. Of those, MCAM, PTK7, CTGF, FGF5, and CXCR4 play critical roles in angiogenesis. MCAM, PTK7, RGCC, CTGF, FGF5, ADAMTS1, CXCR4, IL-11, and MMP1 are well-known essential factors for tumor migration/invasion. Notably, a number of key genes in this signature encode important inducers of EMT and CSC features including: MCAM, PTK7, RGCC, CTGF, and CXCR4. RGCC and Smad3 direct the induction of EMT through regulation of SNAIL and SLUG EMT transcription factors [65]. CXCR4 activates several signaling pathways, including AKT [66], a process in which Src plays a critical role [67]. It should be noted that, some genes in Kang’s signature controversially function against tumor growth or angiogenesis (fourth group). Those include FHL1, DUSP1, SOCS2, FST, and ADAMTS1 (see below). SLC4A7 and NCF2, which are not categorized in these groups, play roles in preparing the microenvironment and inhibition of apoptosis, respectively (Table 4). From the above mentioned information, MCAM, PTK7, CTGF, and CXCR4 are categorized into all first three groups, and likely play critical roles in the last steps of BC bone metastasis.

Genes in Kang’s signature function toward controlling principal signaling pathways, and govern a sophisticated signaling network that leads to a successful metastasis. Table 4 demonstrates undeniable deviation in the regulation of key signaling pathways, like TGF-β, WNT, NFκB, FGF, and MAPK, as underpinning functions of genes associated to bone metastasis. Almost half of the genes in this panel, including DUSP1, RGCC, FST, CTGF, CXCR4, IL-11, and MMP-1, are directly linked to the TGF-β signaling [65, 68–76]. FST also increase mTOR signaling via Smad3 [77]. PTK7, FST, CXCR4, and MMP-1 take part in the regulation of WNT signaling, and remarkably, RGCC and CTGF function via NFκB pathway. FGF5 mediate FGF signaling, and enhance MAPK pathway [78]. CTGF also act through ERK and FAK pathways [79]. Importantly, distinct gene expression pattern of Kang’s signature specifically direct disseminated tumor cells to overt bone metastasis. CTGF, ADAMTS1, CXCR4, IL-11, and MMP1 are considered crucial inducers of bone metastasis [80–85]. Together, it seems that for a successful bone metastasis such sophisticated signaling network is required, which is controlled by the power of gene expression regulation.

Breast cancer bone metastasis; a multistep cascade of events

Several studies have reported genes that are associated with BC bone metastasis, from which some prognostic tools are provided in order to obtain best available treatments for individual BC patients. However, a comprehensive understanding of the nature of metastasis is yet to be investigated. Accordingly, due to the short knowledge of this complex phenomenon, well-suited therapies are lacking. In this review, we aimed to use published gene signatures, which have been shown to be significantly linked to progression and metastasis of BC to bone, to unmask a sequential pattern of gene expression that leads to colonization of breast tumors in bone.

Primary breast tumors cells ought to arrange a programmed gene expression pattern for their growth, survival, and invasion. Therefore, programs related to cell cycle progression, proliferation, apoptosis resistant, angiogenesis, invasion, and distant colonization need to be directed by nucleus. Genes in van’t Veer’s and Smid’s signature seem to provide such well-coordinated program for tumor cells, as both signatures are derived from primary tumors of BC patients, who developed metastasis (in general) and bone metastasis, respectively [14, 16]. We thought that genes in these two signatures likely express in a time period that primary tumor: first get committed to metastasize, and second committed to metastasize to bone. Then, as is obvious in Tables 2 and 3, primary tumor first get ready for metastasis, and then get features of invasion, and intravasation. Tumor cells, then, acquire bone specific directing features of homing, extravasation, micro-colonization, and eventually macro-colonization (metastatic colonization) from the genes in Kang’s signature (Table 4) (Fig. 1). This is likely because Kang’s signature is primarily derived from metastatic breast tumor cells from bone lesions [15]. Importantly, key genes in Kang’s signature including MMP-1, CXCR4, FGF5 and CTGF, which was initially determined from study on metastatic MDA-MB-231 cells and mouse model, have been recently confirmed in patients with BC and prostate cancer [64]. Figure 2 also shows deviation toward particular functions in each signature. In van’t Veer’s signature almost two out of third number of genes (24 out of 35) are in the category of proliferation. In Smid’s signature 13 out of 17 genes are in the category of migration. In Kang’s signature 9 out of 16 genes are involved in functions of migration and invasion. This may show the evolutionarily pattern of tumor progression and metastasis in BC bone metastasis.

Fig. 1
figure 1

Sequential gene expression pattern, from primary tumor to metastatic colonization. van’t Veer’s, Smid’s, and Kang’s gene signatures shows a sequential pattern of expression that leads to breast cancer progression, metastasis. Genes in van’t Veer’s signature are mostly involved in the regulation of cell cycle, DNA replication, Proliferation, tumorigenesis, and survival, therefor, are essential for primary tumor cells to be prepared for invasion and metastasis. On the other hand some genes in this signature act toward angiogenesis, migration, and invasion. ESM1, ECT2, and EGLN1 are present in both categories and likely play important roles in primary steps of tumor metastasis. Genes in Smid’s signature also play essential roles in cell growth, proliferation, survival, angiogenesis, migration, and invasion. Tumor cells that overexpress genes in these two signatures seem to be able to invade and intravasate from primary site and disseminate through circulation. TFF1, TFF3, and FGFR3 are present in all three categories. Genes in Kang’s signature are mostly associated to angiogenesis, migration/invasion, EMT, and CSC factors. Many genes in this signature have been directly linked to bone metastasis of breast cancer. MCAM, PTK7, and CTGF are common genes in three categories. Genes in Kang’s signature likely act toward homing, extravasation, formation of micrometastasis, and eventually metastatic colonization, in order to end this journey. Abbreviations: EMT epithelial to mesenchymal transition, CSC cancer stem cells

Full size image
Fig. 2
figure 2

Genes in van’t Veer’s, Smid’s, and Kang’s signatures are each deviated toward distinct functions. Certain numbers of genes in each signature are involved in certain functions categorized into: proliferation (blue), survival (green), angiogenesis (red), migration (yellow), and invasion (brown). Several genes have common functions, which are located in interconnected territories of circles. Red highlighted genes have functions against the corresponding feature (circle). Size of the circles shows the deviation toward that function in each signature. (Color figure online)

Full size image

A set of well-defined genes govern key molecular pathways, and are the behind-the-scene forces of BC bone metastasis. Genes from van’t Veer’s, Smid’s, and Kang’s signatures control signaling pathways that have been considered as pivotal driving forces of tumor progression and metastasis. TGF-β, FGF, NFκB, WNT, PI3K, and JAK-STAT signaling pathways are induced/enhanced by genes of van’t Veer’s, and Smid’s signatures in primary breast tumor cells. On the other hand, TGF-β, FGF, NFκB, WNT, and PI3K pathways are also induced/enhanced by genes of Kang’s signature in bone colonized tumor cells. In both primary and distant tumors EMT program is induced by these pathways, and also several genes such as SNAI1 and MMP9 in primary tumor, and MCAM, PTK7, CXCR4, RGCC, and CTGF in distant metastatic tumors (Fig. 3) (Tables 2, 3, 4).

Fig. 3
figure 3

Well-orchestrated genes govern key molecular pathways, and are the behind-the-scene forces of breast cancer bone metastasis. Genes in van’t Veer’s, Smid’s, and Kang’s signatures govern a comprehensive signaling network in primary and secondary breast tumors. TGF-β, FGF, JAK-STAT, NFκB, WNT, and PI3K pathways in primary tumor, and TGF-β, FGF, NFκB, and PI3K pathways in secondary tumor are regulated by genes in these three signatures. In the primary tumor the six signaling pathways build a comprehensive signaling network that lead toward tumor growth, proliferation, survival, angiogenesis, migration, and invasion. In primary tumor (up-left), genes from van’t Veer’s, Smid’s signatures and their related signaling molecules are showed. HIF seem to have profound effects in primary tumor development and dissemination. Notably, several genes and signaling cascades induce EMT, and therefor CSC associated features. Intravasated tumor cells form the population of circulating tumor cells that disseminate, home, and extravasate into the secondary organ (bone). Importantly, the majority of differentiated circulating tumor cells (yellow) cannot survive the inhospitable environment while in circulation, and a small proportion of CSCs (red) are able to reach distant sites and form metastasis. In metastatic tumor (down-right), genes from Kang’s signature lead to activation of the five pathways, which build a comprehensive signaling network that governs features like invasion, migration, EMT, and CSC formation. It is important to mention that, tumors in primary and secondary sites can be different or identical regarding their state of differentiation. In most cases secondary tumors are at the same level of differentiation as primary tumor, or even more differentiated. But in some cases, like triple negative breast cancer, both primary and secondary tumors are mostly mesenchymal, and secondary tumors are even more mesenchymal (not shown in this figure). Abbreviations: TGF-β: transforming growth factor-beta, FGF: fibroblast growth factor, JAK-STAT: Janus kinase/signal transducers and activators of transcription, NFκB: nuclear factor kappa B, WNT, and PI3K: phosphatidylinositol 3 kinase, HIF: hypoxia inducible factor, EMT: epithelial to mesenchymal transition, CSC: cancer stem cells. (Color figure online)

Full size image

Smad-dependent and Smad-independent TGF-β signaling pathways are essential for EMT and BC metastases [86–88]. Smad3 and Smad4 dependent TGF-β signaling have been shown to be indispensable for the induction of EMT and metastasis [89–91]. Smad transcription factors orchestrate overexpression of several important genes involved in EMT and metastasis, including SNAIL, TWIST, and ZEB families of transcription factor coding genes [92]. TGF-β also participates in the activation of several key signaling pathways such as Ras/ERK, and PI3K/Akt, called Smad-independent pathways, and regulates cell growth, survival, cytoskeletal reorganization, migration, and invasion [93]. For instance, MMP9 is shown to be induced by TGF-β-induced Akt-dependent ERK pathway [94]. TGF-β stimulation leads epithelial cells to obtain mesenchymal-like features, and capability of migration, invasion, and dissemination through circulation to distant sites of metastasis, and features of stemness [95].

FGF signaling also acts as an important inducer of EMT and metastasis [96]. FGFs regulate a wide range of biological functions such as proliferation, survival, and migration [97], and likely play pivotal roles in bone metastasis of BC [16]. NFκB is also an important inducer of EMT, and act through direct activation of SNAIL and ZEB family of transcription factors [98, 99]. Interestingly, it is believed that the cooperation of NFκB and TGF-β signaling pathways is critical for EMT and cancer metastasis [100, 101]. Wnt signaling is among the most important pathways involved in the induction of EMT and breast CSCs [86, 102, 103]. Notably, WNT and TGF-β signaling pathways likely induce an mutually reinforcing autocrine signaling network that is indispensable for constant expression of EMT associated transcription factors and CSC niche [87, 104]. PI3K signaling pathway is well-known for its important roles in the induction of EMT and metastasis [105]. Intriguingly, it has been shown that PI3K signaling is essential for autocrine/paracrine TGF-β associated motility, invasiveness, and metastasis [106]. JAK-STAT signaling pathway has an essential regulatory role in growth and proliferation of breast CSCs [107]. JAK-STAT signaling is associated with essential features such as survival, cell cycle regulation, self-sufficiency in growth and metastasis [108, 109]. Importantly, JAK2 also interacts and activates PI3K and RAS signaling molecules [110].

Hypoxia and hypoxia inducible factors (HIFs) seem to have pivotal roles in BC bone metastasis. HIF directly regulates several genes from van’t Veer’s, Smid’s, and Kang’s signatures. TFF3, EGLN1, SNAI1, MMP9, TGFB3, SLC2A3, and CTGF are of genes that are directly regulated by hypoxia. In fact, HIFs are of the essential preliminary factors that trigger gene expression programs that lead to tumor progression and metastasis, and play critical roles in the induction of EMT and stemness state in CSCs [111]. Hypoxia and HIFs are likely essential factors in the regulation of on and off states of EMT between primary and secondary tumors. In primary tumors, localized hypoxia mediates HIFs to be activated, and therefore move toward EMT/CSC induction and metastasis. At the secondary sites, however, with likely no hypoxic environment, lack of hypoxia and other factors lead to the reversion of EMT and CSC features. This phenomenon is essential for metastasis of differentiated carcinomas [87, 112] (for a comprehensive review see Ref. [112]).

It seems that a comprehensive signaling network consisting of TGF-β, FGF, NFκB, WNT, PI3K, and JAK-STAT is indispensable for breast tumor cells to progress to overt bone metastasis. Essential links between key bone metastatic factors, such as vascular cell adhesion molecule 1 (VCAM1), receptor activator of nuclear factor κB ligand (RNAKL), parathyroid-hormone related peptide (PTHrP), and BACH1 with these pathways further confirms this signaling network. VCAM1 Promotes bone metastasis by attracting and tethering osteoclast progenitors that express α4 integrin and facilitating their maturation [113, 114], and induces PI3K-Akt signaling by the mediation of Ezrin [115]. RANKL regulates bone resorption [116], migration [117], invasion [117, 118], bone metastasis [3, 117, 119, 120], and induces tumorigenesis, EMT, stemness [121], and the upregulation of MMP1 [118]. PTHrP can be induced by TGF-β [122], and activates CTGF through protein kinase A/C and ERK pathways [80]. BACH1 is a common regulator of several bone metastasis genes, including MMP1 and CXCR4 [123], induced by TGF-β [124].

Genes play the central role in development and diseases. Controlling the cellular pathways is one of the most critical duties of genes, in which reciprocal feedbacks play essential roles. Gene expression pattern in a given cell likely relates the story of a journey in which the cell is born, grow, proliferate, and/or die. Regulating the cellular behavior is the most critical tasks of gene expression machinery. Malignant behavior in cancers is tightly controlled their by gene expression pattern. Connection of gene signatures discussed in this review may provide novel insight toward better understanding the journey in which tumor cells get features of malignancy and metastasize to distant sites, and therefore providing best fit treatments for any individual cancer.