Molecular Mechanisms and Emerging Therapeutic Targets of Triple-Negative Breast Cancer Metastasis

Breast cancer represents a highly heterogeneous disease comprised by several subtypes with distinct histological features, underlying molecular etiology and clinical behaviors. It is widely accepted that triple-negative breast cancer (TNBC) is one of the most aggressive subtypes, often associated with poor patient outcome due to the development of metastases in secondary organs, such as the lungs, brain, and bone. The molecular complexity of the metastatic process in combination with the lack of effective targeted therapies for TNBC metastasis have fostered significant research efforts during the past few years to identify molecular “drivers” of this lethal cascade. In this review, the most current and important findings on TNBC metastasis, as well as its closely associated basal-like subtype, including metastasis-promoting or suppressor genes and aberrantly regulated signaling pathways at specific stages of the metastatic cascade are being discussed. Finally, the most promising therapeutic approaches and novel strategies emerging from these molecular targets that could potentially be clinically applied in the near future are being highlighted.

Based on their molecular profile, breast cancers may also be clustered into basal-like and luminal subsets. Luminal breast cancers are more heterogeneous compared to basal cancers in terms of gene expression, mutation spectrum, copy number changes, and patient outcomes and can be further subdivided into luminal A and B subtypes (8,9). The luminal A subtype represents 50-60% of breast cancer cases and is characterized by low histological grade and good prognosis. Luminal A cancers express ER and PR and have a low frequency of P53 mutations (9). Luminal B represents 10-20% of all breast cancers; compared with the luminal A subtype, these cancers are more aggressive; they have a higher grade, worse prognosis, and worse proliferative index. Luminal B display an increased expression of proliferation genes; they are ER+, PR+/−, Her-2+/−, and EGFR+ and have a higher frequency of P53 mutation (9). Because luminal cancers have a high frequency of PIK3CA mutations, the gene that encodes the p110α catalytic subunit of the phosphatidylinositol 3-kinase (PI3K), agents targeting the PI3K/AKT/mammalian target of rapamycin pathway may be useful for their treatment (10).
The basal-like subtype represents 10-20% of breast cancer cases. They are characterized by high proliferation, high histological grade, and poor prognosis. Basal-like cancers can be triple negative and have a high frequency of P53 mutations combined with loss of Rb1 (9,11). However, not all basal-like cancers are triple negative; studies have shown that 5-45% of basal-like cancers express ER while 14% express Her2/Neu (12,13). TNBC is a diverse group of malignancies and can be further categorized to different subtypes. An analysis of 21 breast cancer data sets containing 587 TNBC cases identified seven subtypes based on differential expression of a set of 2,188 genes: two basal like (BL1 and BL2), a mesenchymal (M), a mesenchymal-stem cell-like, an immunomodulatory, a luminal androgen receptor/luminal-like, and an unclassified type (14).
The deregulation of adult mammary stem cells (aMaSC) during tumorigenesis is believed to contribute to the development of TNBC. aMaSCs give rise to common progenitor cells that can differentiate either to basal progenitors that develop mature basal cells, or luminal progenitors. Disruption in the homeostasis of luminal progenitor cells may lead to the development of TNBC. Contributors in the development of TNBC include aberrantly activated signaling pathways, such as Wnt/β-catenin and Notch, transcriptional factors, like Snail, and embryonic stem cell markers including Sox2, Nanog, and Oct4. These alterations allow the restoration of proliferation capacity as well as the de-differentiation of these progenitor cells, leading to the accumulation of mutations that give rise to TNBC (15).
Traditionally, due to the lack of ER, PR, and Her2/Neu expres sion, the ineffectiveness of current breast cancer targeted therapies as well as due to the challenges in identifying key molecular drivers of TNBC progression, chemotherapy has been the foundation of treatment for patients with this disease over the last decades. Despite its sensitivity to chemotherapy, TNBC is associated with a higher risk of distant recurrence, high rates of metastases, higher probability of relapse and worse overall survival (OS) compared to other subtypes (16,17).

COMPLeXiTY OF TnBC MeTASTASiS
The dissemination of breast cancer cells and eventual metastatic growth to distant organs-predominantly the bone, lungs, and brain-represents a significant clinical problem, as metastatic disease is incurable and is the primary cause of death for the vast majority of TNBC patients. Metastatic spread of tumor cells is a highly complex, yet poorly understood process, and consists of multiple steps, including acquisition of invasive properties through genetic and epigenetic alterations, angiogenesis, tumor-stroma interactions, intravasation through the basement membrane, survival in the circulation, and extravasation of some cancer cells to distal tissues (18). However, disseminated cells that survive pro-apoptotic signals in their new environment often remain quiescent in secondary organs undergoing long periods of latency, also known as the dormancy period (19). It is well established that the outgrowth of metastatic cells in a foreign tissue microenvironment is a highly inefficient process and is considered as the rate-limiting step of breast cancer metastasis (20) (Figure 1). During this stage, breast cancer cells are usually difficult to detect and exhibit resistance to chemotherapy due to lack of proliferation (19). This remains a major clinical problem since patients, often considered as "survivors, " can develop metastatic disease years later. Disseminated tumor cells (DTCs) can enter a state of dormancy in secondary organs by exiting the proliferative cycle for an indefinite period or by achieving a balanced state of proliferation and apoptosis. Successful emergence from dormancy is the result of further evolution of surviving DTCs, by accumulating molecular alterations as well as via permissive interactions with the tumor microenvironment (19). By acquiring these characteristics, metastatic populations can optimally adapt to the host microenvironment and initiate colonization. While significant progress has been made to highlight some of the specific processes required for the breast tumor initiation, efforts have recently been focused on elucidating the roles of critical genes, the underlying molecular mechanisms and signaling pathways involved in the fatal late stages of metastatic dissemination. These studies are of outmost importance for the development of novel effective treatments against metastasis of TNBC.

Local invasion/intravasation
Upon accumulation of genetic and/or epigenetic alterations, breast cancer cells at the primary tumor initially acquire properties, such as self-renewal, ability to migrate, and invade the surrounding normal tissues. During local invasion, breast cancer cells undergo epithelial-to-mesenchymal transition (EMT), a highly orchestrated transcriptional program, initially described during embryonic development, associated with dramatic remodeling of cytoskeleton, loss of apico-basolateral polarity, dissolution of cell-cell junctions, concomitant with downregulation of epithelial markers and upregulation of mesenchymal genes (21). This process is triggered by EMT-master regulators, such as the transcription factors Slug, Snail, and Twist to promote TNBC cell migration and intravasation in the circulation (22)(23)(24). The TGFβ pathway plays a critical role in regulating this early metastatic event. During intravasation, TGFβ promotes overexpression of musculoaponeurotic fibrosarcoma oncogene family protein K (MAFK) to induce EMT and enhance tumor formation and invasion in vivo (25). The TGFβ-Smad signaling axis controls the EMT step in the malignant progression of breast cancer cells either by inducing the expression of master transcriptional regulators of EMT, as described above, or by epigenetic silencing of epithelial genes, including CDH1 (26). The EMT program regulated by TGFβ/Smad signaling also involves WAVE3, a WASP/WAVE family actin-binding protein. In TNBC cells, depletion of WAVE3 expression prevented TGFβ-induced EMT phenotype (27). However, despite numerous studies using cell lines and animal models suggesting a functional role of EMT and EMT-inducing transcription factors in promoting breast cancer metastasis, the in vivo role and clinical relevance of this process remains controversial (28)(29)(30)(31).
Moreover, the majority of genes implicated in TNBC metastasis have been reported to play a major role at the initial stages of cancer cell dissemination which include migration, invasion, and intravasation. This is not surprising given the fact that cancer cell dissemination is thought to be an early event during breast cancer evolution and that primary and metastatic tumor growth is likely to progress in parallel (32). For example, activation of CXCR4 receptor via its ligand CXCL12 or ANGPTL2 was found to induce MLK3 and Erk1/2 signaling and promote intravasation which leads to the development of lung and bone metastases (33)(34)(35)(36)(37)(38)(39). This hyperactive signaling axis may also function in multiple stages of the metastatic cascade, including angiogenesis, extravasation, and osteolysis at the secondary organ. At the same time, it is becoming increasingly clear that trans-endothelial migration and invasion of breast cancer cells in the vasculature is inhibited by metastasis suppressors, including TP63, LIFR, lysyl oxidase-like 4 (LOXL4), FOXF2, SSBP1, RAB1B, and TIEG1 (25,(40)(41)(42)(43)(44)(45)(46)(47), suggesting that the migratory and invasive potential of breast cancer cells is ultimately determined by the balance in the activity of these molecules. The identification of numerous genes implicated in the initial stages of TNBC metastasis highlights the significant challenges for early molecular diagnosis and therapy.

Survival in Circulation
Upon entering the blood vessels, circulating tumor cells express proteins that have antiapoptotic and pro-survival functions which allow them to attach to and infiltrate specific secondary sites. Neurotrophic tyrosine kinase receptor TRKB was shown to inhibit anoikis, a form of cell death caused by lack of adhesion, via the PI3K/Akt pathway. These studies indicated that TRKB induces survival and proliferation of breast cancer cells to promote infiltration in the lymphatic and blood vessels and colonization in distant organs (48). In TNBC cells, brain-derived neurotrophic factor (BDNF) binds and activates TRKB receptor to regulate a network consisting of metalloproteases and calmodulin and thus modulate cancer-endothelial cells interaction. Importantly, Erk1/2 inhibitors were able to block the BDNFinduced phenotype, suggesting that blocking this pathway may be explored for therapeutic purposes against TNBC metastasis (49). In addition, the binding of platelets with circulating breast cancer cells has been shown to essential for their survival, evasion of pro-apoptotic signals, whereas interfering with this interaction inhibits the development of lung metastasis in TNBC mouse models (50,51).

extravasation in Distal Sites
Many of the genetic alterations found to be involved in intravasation are also implicated in extravasation ( Table 1) since, in large part, these two processes are considered "mirrored" to each other. The TGFβ pathway plays an important role in regulating both these metastatic steps. More specifically, TGFβ induces the assembly of a mutant-p53/Smad protein complex to inhibit the function of the metastasis suppressor TP63 and promote cell migration and invasion (40). During extravasation, TGFβ induces angiopoietin-like 4 (ANGPTL4) expression via the Smad signaling pathway; the increased levels of ANGPTL4 enhance the retention of cancer cells in the lungs by disrupting vascular endothelial cell-cell junctions, thus increasing the permeability of lung capillaries to facilitate trans-endothelial passage of breast cancer cells (52). Moreover, targeting the decoy interleukin-13 receptor alpha 2 (IL13Ra2) upregulates the metastasis suppressor TP63 in an IL13-mediated, STAT6dependent manner and impairs extravasation of basal-like breast cancer cells to the lungs (41). Several reports also highlight the importance of the synergistic effects of genes in promoting metastasis by regulating specific stages of the process. For example, EREG, COX2, MMP1, and MMP2 can collectively promote metastatic extravasation to the lungs. These four genes were found to be overexpressed in TNBC cells independently of VEGF. Individual reduction of each gene or their silencing in different combinations produced limited effects on tumor growth in vivo while concurrent silencing of all four achieved nearly complete growth abrogation (53).

Metastatic Colonization
Following extravasation and infiltration at the secondary site, a genetic program is initiated so that cancer cells can escape dormancy and form micro and macrometastatic tumors. Initially, EMT plasticity and the reversal to MET phenotype have been shown to be important for metastatic colonization (113). During this process, epithelial phenotype becomes re-established through miR-200-mediated downregulation of ZEB1, SIP1 to promote metastatic colonization (114,115). Also, breast DTCs in the bone marrow gain the ability to form typical osteolytic metastases by producing parathyroid hormone-related protein (PTHLH), tumor necrosis factor-α (TNFα), interleukin-6 and/or interleukin-11. These factors stimulate the release of receptor activator of nuclear factor-κB ligand (RANKL) from osteoblasts which induces osteoclast formation (33,58,83,116). Furthermore, inflammation in the lung microenvironment could also be responsible for triggering the escape of metastatic breast cancer cells from latency leading to metastatic colonization (117). A subset of genes contributing to primary tumor growth can also promote survival and growth at the secondary site. Chemokines CXCL1/2 mediate chemoresistance and lung metastasis by attracting myeloid cells into the tumor, which produce low molecular weight calcium-binding proteins S100A8/9 that enhance cancer cell survival by binding to the receptor for advanced glycation end products (RAGE) (59). Another calcium binding protein, S100A7 has been found to enhance tumor growth and metastasis, by binding to RAGE and activating Erk and NFκB signaling (88,90). Furthermore, fibroblast growth factor receptor (FGFR) was shown to trigger pro-survival signals through PI3K/Akt signaling and promote outgrowth of metastatic breast cancer cells to the lungs (62). However, it needs to be highlighted that cellular and genetic context among cancers influences whether proteins act as tumor suppressors or metastasis promoters. One controversial example is LOXL4 which has been shown to recruit bone marrow-derived cells and facilitate colonization of TNBC to the lungs via a HIF1α-dependent mechanism (118). However, in another study, knockdown of LOXL4 expression in TNBC cells promoted primary tumor growth and lung metastasis which was associated with thickening of collagen bundles and remodeling of the extracellular matrix (ECM) within tumors (25). Overall, it is noteworthy that while some genes have been associated only with TNBC metastasis so far (i.e., TIEG1, MAFK, MLK3, SDPR), the majority is also involved in other tumor types, suggesting a more fundamental role in cancer progression.     against BRCA-mutated TNBC, inducing partial responses in 54% of patients when administered as a single agent (121) and an overall response rate of 88% when combined with carboplatin (122). Anti-androgens as well as FGFR inhibitors have been tested in clinical trials against TNBCs that are androgen receptor-positive or harbor FGFR amplification, respectively (123,124). Gammasecretase inhibitors that block the NOTCH pathway are currently in clinical trials for TNBC patients with upregulated NOTCH signaling (125). All together clinical trials have shown that each agent alone provides small or no benefit in TNBC patients suggesting that further effort is needed to discover novel targets of TNBC and to identify each patient's molecular profile that will lead to a more individualized treatment. Toward this goal, some of the metastasis-promoting genes reported here could be further exploited for the future development of promising targeted therapies. Since local invasion, intravasation and possibly extravasation are thought to occur relatively early in the metastatic process (32), a plausible strategy would be to target dormancy and the outgrowth of macrometastatic tumors in distal organs. Since this final stage is considered the critical "rate-limiting" step of the "invasion-metastasis" cascade requiring even years to be completed, it provides a window of opportunity for effective therapy. Therefore, different approaches could aim against "druggable" molecules that facilitate metastatic colonization, such as overexpressed receptors or secreted molecules (i.e., CXCL1/2, FGFR, TGFβ1, WNT1, ANGPTL2, CSF2, RANKL), which target commonly deregulated signaling networks at this late-stage ( Table 1). Ongoing clinical trials are evaluating the efficacy of the TGFβR1 inhibitor LY2157299 with paclitaxel (NCT02672475), whereas the FGFR inhibitor Lucitanib is also under testing (NCT02202746) for patients with metastatic TNBC. The ultimate goal would be, if not to completely eliminate dormant metastatic breast cancer cells, to prolong dormancy period and hopefully transform this stage into a chronic inactive cancer cell state.

COnCLUDinG ReMARKS On CURRenT AnD FUTURe PeRSPeCTiveS On TnBC MeTASTASiS THeRAPY
Importantly, recent studies have shown that tumor cells are able to evade immune responses by activating negative regulatory pathways, also known as immune checkpoints, that block T-cell activation through cytotoxic T-lymphocyte protein 4 (CTLA4) or via binding of the programmed cell death protein 1 (PD1) receptor expressed on T-cell surface to the PDL1 ligand expressed by cancer cells in response to various cytokines (126). The recent development and FDA approval of anti-CTLA4, anti-PDL1, and anti-PDL1 monoclonal antibodies that elicit antitumor clinical responses in a variety of solid cancers created enthusiasm for cancer therapy (127). Currently, several clinical trials are underway to evaluate the efficacy of this approach in TNBC as well (128).
However, a major clinical problem is that breast cancer is considered one of the most desmoplastic tumor types due to the production of excessive amounts of ECM components, such as collagen and hyaluronan, which generate mechanical stresses within the growing tumor (129). This results in blood vessel compression, hypoperfusion, and hypoxia which promote cancer progression and metastasis as well as hinder drug delivery (130). Therefore, targeting components of the tumor microenvironment has also been recently proposed as another promising strategy for TNBC therapy by improving tumor penetration and delivery of cytotoxic drugs (131). For example, targeting of cancer-associated fibroblasts using pirfenidone, an FDA-approved drug for idiopathic pulmonary fibrosis, has been shown to suppress metastasis of TNBC in combination with doxorubicin (132). This effect is likely to be mediated through remodeling of tumor microenvironment which reduces ECM components through suppression of TGFβ signaling, improves perfusion and delivery of chemotherapy (133). Similar effects have also been demonstrated using the anti-fibrotic drug Tranilast or the anti-hypertensive drug Losartan in combination with chemotherapy or nanotherapy in mouse models for TNBC (134-136).
In conclusion, this evidence suggests that efforts in the near future should be focused toward the development and testing of novel anti-metastatic targeted therapies for late-stage TNBC that could be used in combination with existing chemotherapies, immunotherapies as well as with microenvironment-remodeling agents that can improve drug penetration and overall therapeutic efficacy.

AUTHOR COnTRiBUTiOnS
CN and PB wrote the paper and helped with illustrations. PP conceived the theme, wrote the paper, and prepared illustrations.