Mechanisms of Organ-Specific Metastasis of Breast Cancer

Abstract

Cancer metastasis, or the development of secondary tumors in distant tissues, accounts for the vast majority of fatalities in patients with breast cancer. Breast cancer cells show a striking proclivity to metastasize to distinct organs, specifically the lung, liver, bone, and brain, where they face unique environmental pressures and a wide variety of tissue-resident cells that together create a strong barrier for tumor survival and growth. As a consequence, successful metastatic colonization is critically dependent on reciprocal cross talk between cancer cells and host cells within the target organ, a relationship that shapes the formation of a tumor-supportive microenvironment. Here, we discuss the mechanisms governing organ-specific metastasis in breast cancer, focusing on the intricate interactions between metastatic cells and specific niche cells within a secondary organ, and the remarkable adaptations of both compartments that cooperatively support cancer growth. More broadly, we aim to provide a framework for the microenvironmental prerequisites within each distinct metastatic site for successful breast cancer metastatic seeding and outgrowth.

Distant metastasis is the leading cause of cancer-related mortality in patients with breast cancer. Metastatic breast cancer carries a poor prognosis, with an overall 5-yr survival rate of ∼27% (Yousefi et al. 2018). Breast cancer exhibits a distinct metastatic pattern, with patients most frequently presenting with metastasis in the bone, liver, lung, and brain. Different breast cancer subtypes also show preferences for distinct secondary sites, with luminal breast tumors commonly metastasizing to bone. Triple-negative breast cancer (TNBC) displays lung tropism while HER2⁺ breast cancers exhibit a high propensity for brain and liver metastasis (Wu et al. 2017). Metastasis is a complex multistep process that begins with local invasion of tumor cells through the basement membrane, intravasation, and then dissemination via the circulation and/or lymphatic systems. After extravasation at the target distant organ, disseminated tumor cell (DTC) seeding and outgrowth leads to the formation of micrometastatic lesions, and eventually life-threatening macrometastases.

In this review, we will focus on the major rate-limiting step of the metastatic cascade, the successful colonization of distant organs (Vanharanta and Massagué 2013). DTCs arriving in the secondary organ are vulnerable to immune surveillance, host-tissue defense, as well as the distinct nutrient and environmental constraints in the foreign environment. Thus, the generation of a favorable metastatic niche is fundamental for successful metastatic seeding and the expansion from micrometastases to overt lesions. Cancer cells have the remarkable ability to modify metastatic sites, both at the systemic level in preparation for metastatic colonization (premetastatic niche), and after their arrival at the secondary organ to create a specialized environment more suitable for outgrowth. Cancer cells themselves also undergo profound adaptations, such as metabolic reprogramming, to adapt to the specific requirements of the distant tissue. This review will explore the extraordinary coevolution of metastatic breast cancer cells and their target organ, with a special focus on the intricate tumor–niche interactions that nurture the survival and outgrowth of DTCs and govern metastatic success. Increasing our understanding of the drivers of cancer-niche evolution, as well as the unique microenvironmental prerequisites for successful colonization, may shed new light on effective anti-metastasis therapies.

LUNG NICHE

TNBC, which accounts for ∼10%–15% of all breast cancers, are especially prone to metastasize to the lung, with an incidence of pulmonary metastasis in up to 40% of patients (Foulkes et al. 2010). Lung metastases are a major cause of breast cancer–related mortality, with patients typically facing a median survival of 22 mo after treatment (Smid et al. 2008). The lung metastatic niche represents a remarkable example of the complex and constantly evolving symbiosis between tumor cells and host cells during metastatic colonization. Breast DTCs evoke profound phenotypic changes in lung-resident and recruited cells, shaping the formation of a tumor-supportive niche, which in turn fuels metastatic progression (Fig. 1). In this section, we unravel this ultimate partner-in-crime relationship, identifying the key players and mechanisms that determine metastatic success in the lung environment.

Figure 1.

Overview of the mechanisms governing organ-specific metastasis in breast cancer. Successful metastatic seeding in the lung, liver, bone, and brain environment is highly dependent on intricate cross talk between cancer cells and tissue-specific niche cells. Depicted here are the main players, biological processes, and adaptations that contribute to the evolution of a protumorigenic niche within each distinct metastatic site. Cancer cells are capable of modifying their intended secondary organ at both the systemic level in preparation for metastatic colonization (premetastatic niche; lower right panel) and after their arrival in the distant tissue. (ECM) Extracellular matrix, (CAFs) cancer-associated fibroblasts, (DCs) dendritic cells.

Tumor–Immune Cell Interplay in the Lung Niche

The cooperation between newly arriving breast cancer cells and the immune compartment of the lung showcases the ability of tumor cells to fine-tune their environment and, in doing so, license their growth. Macrophages, an abundant component of the lung niche, play a critical role in supporting the extravasation, seeding, and outgrowth of malignant cells (Pollard 2004). A distinct subset of tumor-promoting macrophages has been identified in the metastatic lung, which are preferentially recruited by extravasating breast cancer cells (Qian et al. 2009). These originate from CCR2-expressing inflammatory monocytes that infiltrate the lung in response to cancer cell–derived CCL2, where they differentiate into macrophages and directly support metastatic seeding (Qian et al. 2011). In addition to promoting macrophage infiltration, lung-trophic breast cancer cells are intrinsically primed to exploit macrophage-derived protumorigenic signals (Song et al. 2016). By up-regulating their expression of GALNT14, arriving breast cancer cells can maximize their responsiveness to macrophage-derived fibroblast growth factor (FGF) signaling, increasing their chance of successful colonization. Interestingly, GALNT14 up-regulation is also an adaptive mechanism to overcome the inhibitory effect of anti-metastatic bone morphogenetic protein (BMP) signaling in the lung (Song et al. 2016).

Perhaps the most well-established driver of pulmonary metastasis are neutrophils, a major component of the lung immune microenvironment. Both pro- and antitumoral functions of neutrophils have been described, with recent reviews extensively covering this context-dependent dichotomy (Jaillon et al. 2020; Hedrick and Malanchi 2022). In metastasis, neutrophils typically behave as tumor accomplices, secreting an array of cytokines and chemokines that support cancer cell seeding, survival, immune-evasion, and outgrowth. Neutrophils also promote metastasis via neutrophil extracellular traps (NETs), the web-like lattice of DNA coated in proteolytic enzymes that trap and kill pathogens during an infection. Arriving metastatic breast cancer cells can induce NETosis in lung neutrophils in the absence of pathogens, fostering invasion and the expansion of disseminated cells (Park et al. 2016). Thus, cancer cells are able to hijack this intrinsic neutrophil activity to support their growth in noninfectious conditions. NET-driven metastasis is profoundly exacerbated by sustained lung inflammation induced by tobacco smoke exposure or lipopolysaccharide (LPS), in a mechanism linked to extracellular matrix (ECM) remodeling by NET-DNA-associated proteases (Albrengues et al. 2018). More recently, NET-fueled metastasis was linked to cancer-cell-secreted protease cathepsin C, which promotes breast cancer metastatic outgrowth by enhancing IL-6/CCL3-mediated neutrophil recruitment and NETosis (Xiao et al. 2021). Importantly, targeting cathepsin C effectively suppressed lung metastasis in mice, arguing for the use of niche-modifying agents for the treatment of metastasis.

Recent studies have highlighted the dynamics and complexity of the tumor–immune cell interplay in the lung niche, suggesting multiple layers of cell cross talk that create a “domino effect.” For example, the prometastatic effect of lung neutrophils is tunable, dictated by the status of host natural killer (NK) cells (Li et al. 2020a). In an elegant study using mice with varying levels of immune system integrity, the authors showed that while neutrophils themselves display ROS-mediated tumoricidal activity, they simultaneously suppress the more potent tumoricidal ability of NK cells. Thus, in the lung tissue of an NK-competent host, the net contribution of neutrophils is prometastatic. Interestingly, the NK cell status of the host has also recently been found to impact clonal evolution in metastatic breast cancer (Lo et al. 2020). Circulating tumor cell (CTC) clusters were shown to be less sensitive to NK-mediated immunosurveillance compared to single CTCs; thus, NK cells confer a selective advantage for polyclonal seeding in the lung. Another example of coordinated tumor–immune cell interplay is T-cell immunoregulation, in which cancer and immune cells conspire to create an immunologically permissive environment for metastatic colonization. T-cell immunosuppression can be driven by γδ-T-cell-induced neutrophil recruitment and polarization (Coffelt et al. 2015), myeloid-derived osteopontin production (Sangaletti et al. 2014), or via a T-cell-intrinsic mechanism of oxygen sensing specific to the lung environment (Clever et al. 2016). In all cases, the induction of immunotolerance licensed breast tumor colonization in the lung, highlighting the power of immunoregulation to either restrain or unleash metastatic outgrowth.

Importantly, some of the mechanisms enabling immune cells to support metastatic breast cancer cell growth within the lung are influenced by systemic perturbations orchestrated by primary tumor-derived signals. Those systemic changes influencing distant organs drive the formation of a “premetastatic niche,” and they will be discussed later in this article.

Tumor-Mediated Education of Lung Stromal Cells

Breast DTCs arriving in the lung elicit a profound rewiring of the host stromal cells such as fibroblasts, an obligatory event that nurtures their expansion from micrometastases to overt macrometastases. The activation and perturbation of lung fibroblasts and other mesenchymal cells lead to the generation of carcinoma-associated fibroblasts (CAFs), a general term now used to define a highly heterogenous collection of cells with distinct features and activities within the metastatic environment (Sahai et al. 2020; Chen et al. 2021b). At the very early stages of metastatic colonization, lung-infiltrating breast cancer cells trigger the induction of stromal periostin (POSTN) expression, supporting cancer stem cell (CSC) maintenance and metastatic colonization by augmenting Wnt signaling (Malanchi et al. 2012). Indeed, human breast metastatic cells with enhanced stem cell features have been linked with a higher metastatic-initiating capacity, with their early arrival into distant tissues preceding advanced metastatic disease (Lawson et al. 2015). Another matricellular protein, tenascin C (TNC), is highly implicated in promoting cancer cell stemness and metastatic lung colonization (Oskarsson et al. 2011). While initially secreted by tumor cells themselves to support their survival in the new host tissue, lung stromal cells eventually take over as the source of TNC and foster metastatic outgrowth. More recently, transcriptomic analysis has painted a picture of a dynamic global reprogramming of fibroblasts in the lung metastatic niche. Interestingly, these studies showed that fibroblast evolution is influenced by both the stage of metastatic progression and by the metastatic potential of the arriving breast cancer cells (Pein et al. 2020; Shani et al. 2021). Breast DTC seeding in the lung evoke an IL-1α/IL-1β-mediated proinflammatory phenotype in lung fibroblasts, reminiscent of fibroblast activation in wound healing and primary tumors. Cancer cells then exploit this reactive fibroblast niche to fuel their colonization (Pein et al. 2020). Using transgenic mice at defined stages of breast-to-lung metastases, Erez and colleagues suggested that fibroblasts functionally adapt to the evolving lung niche via distinct Myc-regulated transcriptional programs (Shani et al. 2021). ECM remodeling and cellular stress programs were up-regulated in fibroblasts in early micrometastases, while inflammatory signaling was instigated in advanced metastases to perpetuate growth. This demonstration of coevolution of lung stromal and breast cancer cells is in line with previous work showing a reciprocal cross talk between cancer cells and the lung stromal niche, modulating the epithelial-to-mesenchymal (EMT) state of cancer cells, and dictating the outgrowth potential of DTCs (Martin et al. 2015). Overall, these studies reveal the fascinating ability of lung-infiltrating breast cancer cells to trigger the rewiring of their local stromal niche, which, in turn, plays a crucial role in shaping their progression into overt metastases.

New Players in the Lung Niche

More recently, in addition to the more abundant and well-characterized tumor niche cell types, the engagement of other organ-specific cellular components in the lung metastatic niche has emerged. Metastasis in the lung initiates from cancer cells arriving via the microvasculature, and outgrowth mainly occurs within the alveolar environment. During metastatic expansion, malignant cells can reenter the alveolar wall and migrate between endothelial and epithelial layers, whereby they coopt the existing pulmonary microvasculature to enable tumor vascularization (Szabo et al. 2015). Recently, an active participation of lung alveolar cells was discovered in the breast cancer metastatic niche (Ombrato et al. 2019; Montagner et al. 2020). Using a novel niche-labeling system, the arrival of breast DTCs in the lung was found to trigger the regenerative-like activation of neighboring alveolar type II cells (Ombrato et al. 2019). This epithelial response, characterized by the acquisition of stem cell–like features, directly supported the growth of cancer cells ex vivo, suggesting it may contribute to the formation of a favorable metastatic niche. Importantly, cross talk between DTCs and the lung epithelium, in particular alveolar type I cells, may also mediate breast cancer dormancy and the survival of indolent cancer cells (Montagner et al. 2020). Tumor cell dormancy is highly problematic for the clinical management of cancer, a topic covered elsewhere in the literature (see, for example, Dalla et al. 2022); thus, exposing niche factors that mediate tumor cell latency could have major implications for cancer treatment.

In addition to the epithelium, Zheng et al. (2019) recently identified a novel population of atypical cytokine-producing large platelets specifically enriched in lung metastases but not in primary mammary tumors. These are locally produced in response to lung-infiltrating DTCs, likely from lung-resident megakaryocytes that egress from extrapulmonary sites such as bone marrow (Lefrançais et al. 2017). While the contribution of these megakaryocytes to metastatic growth remains to be determined, their identification further shows the complexity of the tumor microenvironment in metastases. A recent study revealed a method for in situ transcriptomic profiling of rare cell populations (“Flura-seq”) to examine organ-specific metabolic reprogramming during the early stage of breast cancer metastatic colonization (Basnet et al. 2019). The authors showed that the pulmonary microenvironment induced specific cancer cell adaptations, in particular an enrichment of mitochondrial electron transport, oxidative stress, and antioxidant programs, when compared to primary tumors or brain micrometastases. The high-sensitivity of Flura-seq could reveal the existence of rare cell populations that, despite representing a small fraction of the metastatic niche, may still be of clinical significance.

BONE NICHE

Bone marrow is a preferred metastatic site for multiple solid tumors including breast, with ∼70% of advanced breast cancer patients harboring bone metastasis (Coleman 2001). The vast majority of breast cancer bone metastases are classified as bone-destructive osteolytic lesions, which are associated with multiple bone complications that can profoundly impact the patient's quality of life, such as bone pain, fracture, hypercalcemia, and paralysis. The success of breast DTC colonization of the bone is dictated by the intrinsic features of tumor cells, favorable tumor–niche interactions, and by the cancer cell–directed protumorigenic rewiring of the local bone environment. There are several recent reviews that extensively dissect the bone marrow metastatic niche, providing a wealth of knowledge on the role of the specialized bone environment in supporting tumor colonization (Esposito et al. 2018; Chen et al. 2021a; Satcher and Zhang 2022). In this section, we provide an update on the intricate tumor–stromal interactions governing the progression from initial seeding to overt macrometastasis, focusing on the evolving symbiosis between tumor cells and the bone microenvironment during colonization (Fig. 1).

Tumor–Niche Interactions

After extravasation into the bone marrow compartment, cancer cells preferentially occupy one of two specialized bone microenvironments or niches—the vascular niche and the osteogenic (or endosteal) niche. The vascular niche harbors bone marrow stromal cells, pericytes, and endothelial cells that closely line the sinusoids, while the osteogenic niche primarily includes all osteoblast and osteoclast lineage cells and is the central hub of bone remodeling. Both niches play a key role in shaping the metastatic process, and are capable of providing both a safe haven for indolent cancer cells or supplying a fertile soil to unleash metastatic growth.

In the osteogenic niche, early-stage colonization of breast DTCs is dependent on the formation of heterotypic adherens junctions between cancer-derived E-cadherin and osteogenic N-cadherin, triggering mTOR activation in cancer cells (Wang et al. 2015). Direct calcium influx from the osteogenic niche to cancer cells is also essential for the proliferation of early-stage micrometastatic lesions (Wang et al. 2018). Once cancer cells have established a foothold within the bone marrow, metastatic progression is driven by the osteolytic vicious cycle, a formidable positive-feedback cycle composed of tumor cells, osteoclasts, osteoblasts, and the bone matrix. Differentiation and maturation of osteoclasts from their monocytic precursors critically depends on the cytokine RANKL, which is produced by osteoblasts. Signaling of RANKL is antagonized by its decoy receptor osteoprotegerin (OPG). Therefore, imbalance in the level of RANKL and OPG may result in disruption of bone homeostasis, leading to osteoporosis or osteopetrosis conditions. In osteolytic bone metastasis of breast cancer, elevated osteoclast activity can be induced by cancer-derived secreted factors such as PTHrP and prostaglandin E2, which stimulate dysregulation of osteoblastic RANKL/OPG expression (for reviews, see Esposito et al. 2018; Wu et al. 2020). Breast cancer cells themselves can also directly secrete RANKL in response to enhanced BMP signaling (Liu et al. 2018) or CCL20 secreted by both cancer cells and osteoblasts (Lee et al. 2017). Alternatively, tumor-derived matrix metalloproteinases can suppress OPG expression by proteolytically releasing membrane-bound EGF family growth factors and subsequent downstream EGFR signaling in osteoblasts (Lu et al. 2009). Activation of osteoclastogenesis can occur independently of RANKL via IL-11-mediated JAK1/STAT3 signaling (Liang et al. 2019) or VCAM1-induced NF-κB signaling (Lu et al. 2011). Bone destruction triggers the release of growth factors deposited in the bone matrix such as IGF1, PDGF, and TGF-β, which in turn stimulates tumor growth and further accelerates bone loss. TGF-β plays a crucial role in perpetuating the osteolytic vicious cycle, and an enrichment of TGF-β signaling is a well-established feature of highly bone-metastatic breast cancer cells (Kang et al. 2003; Korpal et al. 2011; Sethi et al. 2011; Esposito et al. 2021a). TGF-β stimulates the expression of PTHrP, Jagged1, CTGF, and IL-11, all of which are strong activators of osteoclastogenesis (Yin et al. 1999; Kang et al. 2005; Shimo et al. 2006; Sethi et al. 2011). TGF-β-mediated induction of DACT1 is critical for bone metastatic progression via suppression of Wnt signaling, which is important for initial cancer cell seeding but is constrained during macrometastasis outgrowth (Esposito et al. 2021a). The temporal modulation of Wnt signaling in both cancer cells and the bone niche is indeed one of the key aspects in bone metastasis. During the early stage of metastasis, Wnt activity in breast cancer cells induced by IL-1β derived from the bone microenvironment can promote breast cancer stemness and metastatic outgrowth (Eyre et al. 2019). Moreover, breast cancer cells were also reported to secrete DKK1, which boosts osteoclastogenesis by suppressing canonical Wnt signaling–mediated expression of OPG, increasing osteolytic bone metastatic progression (Zhuang et al. 2017).

In another demonstration of bidirectional cross talk involving TGF-β, its release by bone destruction was shown to trigger cancer cell secretion of Jagged1, which in turn engages Notch signaling in tumor-associated osteoblasts and preosteoclasts. This confers both a proliferative advantage to tumor cells by enhancing osteoblast-derived protumorigenic growth factors IL-6 and CTGF and stimulates osteoclast maturation, together fostering osteolytic metastasis (Sethi et al. 2011). Interestingly, chemotherapy appears to induce the inverse cross talk, in which treatment-induced Jagged1 secretion by osteoblasts creates a prosurvival niche for breast DTCs via Notch activation in cancer cells. Anti-Jag1 therapy sensitizes bone metastasis to chemotherapy, with a nearly 100-fold reduction in bone metastatic burden demonstrated following combination therapy in vivo (Zheng et al. 2017). In line with this study, cancer treatment–induced alterations in the bone niche were recently linked to augmented bone loss and breast-to-bone metastasis (Zuo et al. 2020). The dual PARP1/2 inhibitor olaparib was found to promote osteoclast differentiation and bone resorption and foster the recruitment of immune-suppressive immature myeloid cells to the bone niche. The overall effect was exacerbated bone metastasis, warranting a careful examination of the use of current PARP inhibitors in the clinic.

The perivascular niche has increasingly been recognized to play a key role in the development of bone metastasis. Stable microvasculature has been reported to promote the dormancy of DTCs via endothelial-derived thrombospondin-1, TGF-β1, and POSTN as active tumor-promoting factors (Ghajar et al. 2013). The perivascular niche also protects DTCs from chemotherapy through integrin-mediated interactions. Furthermore, vascular niche–derived E-selectin promotes bone metastasis by directly engaging with cancer cells and inducing mesenchymal-to-epithelial transition and Wnt-dependent stemness (Esposito et al. 2019). Using quantitative 3D imaging of mouse bone marrow, Yip and colleagues recently showed that mouse mammary DTCs preferentially home to a distinct vascular niche comprised of type H capillaries during bone colonization. Metastatic progression was critically dependent on cancer cell–directed remodeling of the local vasculature, creating a self-sustaining microenvironment (Yip et al. 2021).

Other important players fueling metastatic growth in the bone environment are emerging. Sympathetic nerves have been shown to play an active role in pathogenesis of bone metastasis (Elefteriou 2016). Ma et al. recently identified a specific subset of protumorigenic macrophages derived from CCL2-recruited inflammatory monocytes, which promote breast cancer bone metastasis in an IL-4R-dependent manner (Ma et al. 2020). Bone marrow adipocytes, the most abundant component of the bone marrow environment, can provide a supportive niche for metastatic cancer cells via the secretion of a plethora of adipokines such as leptin and IL-6. Adipokines play an assertive role in metastatic colonization and progression in the bone marrow through the regulation of tumor cell recruitment, invasion, survival, proliferation, angiogenesis, and immune modulation (for reviews, see Liu et al. 2020; Soni et al. 2021).

Cancer Cell Adaptations in the Bone Environment

Metastatic breast cancer cells are capable of expressing genes normally restricted to bone cells to support their growth in the bone niche. This is known as osteomimicry, in which breast tumor cells acquire an osteocyte-like phenotype and secrete osteomimetic factors that enhance their survival, proliferation, and colonization of the bone environment. These factors include cadherin-11 (Tamura et al. 2008), osteoactivin (Rose et al. 2007), BSP (Wang et al. 2013), Runx2 (Tan et al. 2016), Src (Zhang et al. 2009), cathepsin K (Gall et al. 2007), and osteopontin (Anborgh et al. 2010), among others (Brook et al. 2018). The master switch for this organ-specific cancer cell reprogramming has been accredited to forkhead box F2 (FOXF2), a transcription factor that drives pleiotropic transactivation of the BMP4/SMAD1 signaling pathway (Wang et al. 2019). Epigenetic rewiring of ER⁺ breast cancer cells in the osteogenic niche also influences clonal evolution and fosters endocrine therapy resistance (Bado et al. 2021). Osteoblast–cancer cell interactions during early metastatic colonization drive a transient EZH2-dependent loss in ER expression, allowing cancer cells to escape endocrine therapy–mediated destruction. This partially recovers as the lesion progresses beyond the osteogenic niche, contributing to tumor heterogeneity in late-stage bone metastasis (Bado et al. 2021). Interestingly, EZH2-mediated epigenetic reprogramming in the bone environment also plays a role in metastasis-to-metastasis multiorgan seeding, by enhancing cancer cell stemness and invigorating their dissemination (Zhang et al. 2021).

Cancer cells are also able to adapt to the bone microenvironment by adjusting their metabolic activities. Bone metastatic breast cancer cells are found to overexpress three enzymes required for de novo serine synthesis: phosphoglycerate dehydrogenase (PHGDH), phosphoserine aminotransferase 1 (PSAT1), and phosphoserine phosphatase (PSPH), as well as the serine transporter SLC1A4 (Pollari et al. 2011). These enzymes promote the production of L-serine, an essential amino acid for differentiation of osteoclasts. Bone metastatic tumor cells also release large amounts of lactate to fuel osteoclastogenesis and osteolytic progression (Lemma et al. 2017).

BRAIN NICHE

Brain metastasis is associated with poor survival outcomes and poses unique clinical challenges such as poor drug delivery across the blood–brain barrier (BBB). The distinct microenvironment of the brain, characterized by highly specialized resident cells and metabolic constraints imposes intense selective pressures on breast DTCs. Thus, successful metastatic colonization is dictated by the ability of cancer cells to adapt to the brain microenvironment and orchestrate coevolution of brain-resident cells. In addition, the anatomical location of the brain in which the metastatic cells colonize profoundly influences the metastatic process. The specific microenvironmental requirements associated with each location results in distinct types of metastatic lesions in the brain: parenchymal metastases, leptomeningeal metastasis (in which cancer cells colonize the cerebral spinal fluid), and the rare form, choroid plexus metastasis. The molecular mechanisms underlying breast cancer brain metastases have been recently reviewed in detail (Wang et al. 2021b). In this section, we will focus on the interplay between arriving breast DTCs and the brain microenvironment postextravasation, a relationship that profoundly shapes the metastatic process and empowers tumor colonization (Fig. 1). While we will focus on parenchymal metastasis due to the body of literature in this area, we will also touch on leptomeningeal metastases, a less common site but which is associated with an alarmingly poor prognosis (Franzoi and Hortobagyi 2019).

Tumor–Niche Cross Talk in the Brain Microenvironment

The survival and outgrowth of breast parenchymal metastases critically depends on favorable interactions between cancer cells and the unique cellular compartments within the neural niche. Astrocytes are the most abundant glial cells in the brain and become activated by direct contact with breast DTCs following extravasation through the BBB. Reactive GFAP⁺ astrocytes secrete a plethora of soluble proteins including IL-1, IL-6, IGF-1, TNF-α, and TGF-β that directly supports the invasion, growth, and survival of metastatic cancer cells (Sierra et al. 1997; Wasilewski et al. 2017). Recent studies have highlighted an array of additional metastasis-promoting functions of reactive astrocytes. A subpopulation of pSTAT3⁺ astrocytes found at the periphery of breast, lung, and melanoma brain metastatic lesions form an immunosuppression barrier for metastatic-initiating cells, via the regulation of innate and adaptive immune cells (Priego et al. 2018). Astrocyte–cancer cell cross talk can shape metastatic progression through the regulation of tumor cell gene expression. Astrocyte-derived miR-19a-containing exosomes induce PTEN loss in breast DTCs, priming metastatic outgrowth by boosting CCL2-mediated recruitment of protumorigenic myeloid cells (Zhang et al. 2015). Astrocyte-induced epigenetic regulation of the glycoprotein Reelin specifically in HER2⁺ metastatic breast cancer cells leads to increased proliferation and supports their prometastatic phenotype (Jandial et al. 2017). This may also help to explain the high incidence of brain metastasis in patients with recurrent HER2⁺ breast cancer (Heitz et al. 2011; Martin et al. 2017). In an elegant demonstration of reciprocal cross talk, research from the Massagué laboratory showed that transfer of cGAMP from brain metastatic breast cancer cells to astrocytes triggers the secretion of proinflammatory cytokines, which in turn activates STAT1/NF-κB signaling in cancer cells. The overall effect is augmented metastatic outgrowth and increased chemoresistance (Chen et al. 2016). Bidirectional tumor–astrocyte interactions were also shown to support the growth of metastatic breast cancer cells in the brain in a study by Watabe and colleagues. Cancer cell–derived COX2 and prostaglandins activate astrocytes to secrete chemokine CCL7, which in turn augments the self-renewal activity of metastatic-initiating cells in the brain (Wu et al. 2015). Importantly, astrocytes may also foster metastatic progression by allowing cancer cells to take advantage of extrinsic factors and nutrients in the brain microenvironment. For example, astrocytes can serve as a source of polyunsaturated fatty acids for arriving breast DTCs, promoting cancer cell proliferation via the activation of peroxisome proliferator–activated receptor γ (PPARγ) (Zou et al. 2019). Astrocytes also act as paracrine mediators of estrogen signaling in TNBC brain metastasis, transmitting mitogenic signals to otherwise unresponsive ER^– breast cancer cells and supporting prometastatic behaviors (Sartorius et al. 2016). These studies suggest that cancer cell–astrocyte cross talk allows metastatic cancer cells to exploit the lipid and hormone-rich milieu of the brain microenvironment, fueling their transition to overt metastasis.

Microglia are the unique resident macrophages of the brain parenchyma, and become activated in response to physical contact from arriving DTCs. Activated microglia are capable of carrying out cytotoxic functions, but this can be cleverly thwarted by the secretion of neurotrophin-3 (Louie et al. 2013) or MYC-induced production of the antioxidant glutathione peroxidase 1 by brain metastatic cancer cells (Klotz et al. 2020). On the other hand, tumor-educated microglia can favor metastatic outgrowth by promoting mesenchymal-to-epithelial transition via E-cadherin up-regulation (Louie et al. 2013), or supporting proinvasive behavior in a Wnt-dependent manner (Pukrop et al. 2010). Macrophage-derived cathepsin C supports tumor cell survival and outgrowth in the brain (Sevenich et al. 2014). Interestingly, breast cancer cells themselves are the initial source of cathepsin C to promote BBB transmigration, while macrophages take over as the predominant supplier after extravasation and foster metastatic expansion. More recently, the loss of X-inactive-specific-transcript (XIST) in brain metastatic breast cancer cells was shown to trigger prometastatic microglia reprogramming via cancer cell–derived exosomal miRNA-503 (Xing et al. 2018). Augmented secretion of immunomodulating cytokines by the rewired microglia led to suppression of local T-cell immunity, thereby enhancing metastatic growth (Xing et al. 2018). Interestingly, brain-resident microglia may also be responsible for the more aggressive disease commonly observed in younger breast cancer patients and in mouse models of both TNBC and luminal B breast cancers (Evans et al. 2004; Hung et al. 2014; Wu et al. 2021a). The increase in brain metastatic propensity in younger versus older mice, an effect not observed in liver or lung metastasis, was recently attributed to age-associated loss of protumoral-resident microglia, and myeloid cell depletion preferentially reduced brain metastatic burden in younger mice (Wu et al. 2021a). Importantly, the complexity and plasticity of the immune landscape in the neural niche was recently unveiled through single-cell analyses combined with transgenic mouse models (Guldner et al. 2020). Central nervous system (CNS)-native myeloid cells, primarily microglia, were found to cultivate an immunosuppressive niche via Cxcl10 signaling, promoting the outgrowth of breast cancer brain metastasis (Guldner et al. 2020). More recently, Gonzalez and colleagues generated an extensive single-cell transcriptomic data set for both malignant and nonmalignant niche cells from 15 human brain metastases, including three breast metastatic lesions (Gonzalez et al. 2022). Interestingly, comparable metastatic niches were observed across cancers of different primary tumor origin, which included two functionally distinct subsets of immune-modulating macrophages.

A critical step in successful brain metastatic colonization is vascular cooption, in which interactions between DTCs and endothelial cells allow newly arriving cancer cells to exploit the preexisting blood supply, supporting their outgrowth (for review, see García-Gómez and Valiente 2020). This is unlike leptomeningeal metastasis, in which de novo angiogenesis is a key feature (Lorger and Felding-Habermann 2010). In vivo experimental metastasis models using both human and mouse breast cancer cells showed vascular cooption in over 95% of early brain parenchymal micrometastases, rather than neoangiogenesis. This process is dependent on integrin β1–mediated breast tumor cell adhesion to the vascular basement membrane (Carbonell et al. 2009), while others have also implicated integrin β4 (Fan et al. 2011) and adhesion molecule L1CAM (Valiente et al. 2014). Importantly, niche cells can foster cancer cell–endothelial interactions, elegantly demonstrated by the JAK2/STAT3-dependent promotion of cooption by tumor-recruited macrophages (Wang et al. 2017).

Cancer Cell Adaptation to the Neural Niche

Successful metastatic colonization of the brain requires cancer cells to undergo metabolic reprogramming to adapt to nutrient deprivation in the brain environment. Patient-derived brain metastatic breast cancer cells exhibit neural characteristics and coopt γ-aminobutyric acid (GABA) as an oncometabolite, providing a proliferative advantage (Neman et al. 2014). More recently, brain-colonizing breast DTCs were shown to coopt glutamate-mediated neuronal signaling to support metastatic growth (Zeng et al. 2019). The formation of pseudosynapses between tumor cells and glutamatergic neurons enabled seeding cancer cells to access a supply of glutamate, enhancing their proliferation and supporting their outgrowth into overt metastasis.

Similarly, increased expression of the fatty acid–binding protein 7 in brain-trophic HER2⁺ breast cancer cells enhances fatty-acid utilization and supports the adoption of a glycolytic phenotype in the brain environment (Cordero et al. 2019). In response to glucose deprivation in the neural niche, GRP9A-expressing breast cancer cells can induce a prosurvival autophagy response, relieving metabolic stress and enhancing survival (Santana-Codina et al. 2020). Brain metastatic breast cancer cells also display enhanced gluconeogenesis and enhanced oxidation of glutamine and branched chain amino acids, facilitating glucose-independent growth and enhancing brain metastasis in vivo (Chen et al. 2015). Finally, as well as supporting the nutrient requirements of metastatic cells, metabolic reprogramming in the neural niche can increase cancer cell resistance to ROS-mediated oxidative stress. Increased LEF1 expression by brain-colonizing breast cancer cells boosts levels of glutathione, improving the antioxidant capability of metastatic cells (Blazquez et al. 2020).

Brain-colonizing breast DTCs also adapt protective mechanisms to shield themselves from brain defenses. By producing high levels of the antiplasminogen activator inhibitory Serpins, brain metastatic cancer cells can subvert the lethal action of plasmin produced by the reactive brain stroma, in particular by activated astrocytes (Valiente et al. 2014). Brain metastatic cells can dampen the adaptive immune response in the neural niche through the recruitment of PD-L1⁺ immunosuppressive neutrophils into the brain TME (Zhang et al. 2020). This is instigated by Src-dependent phosphorylation of EZH2 in brain metastatic cells, activating the c-Jun/G-CSF/neutrophil axis to deter T-cell functions and foster metastatic colonization. Interestingly, in leptomeningeal metastases, breast and lung metastatic cancer cells can also orchestrate protumorigenic modifications of their local environment (Boire et al. 2017). Through the secretion of complement component 3, cancer cells disrupt the blood–CSF barrier and facilitate the entry of tumor-supportive mitogens into the nutrient-poor CSF. Thus, by adapting to the hostile environment of the CSF, cancer cells are able to overcome microenvironmental challenges and thrive against the odds (Boire et al. 2017). Interestingly, studies analyzing the CSF metabolome before and during progression of brain metastasis have shown potential for CSF sampling for the early diagnosis of cancer patients with leptomeningeal metastasis (Dekker et al. 2005; Yoo et al. 2017).

HEPATIC NICHE

The liver is a frequent site of metastasis for the majority of solid tumors including breast. This metastatic propensity can be partly attributed to its unique biological structure: the liver is a highly vascularized organ, characterized by a fenestrated endothelium that lacks a subendothelial basement membrane. These features, combined with an exceptionally low blood-flow rate, make the organ intrinsically susceptible to DTC extravasation. After extravasation, successful metastatic outgrowth is then dependent on reciprocal cross talk between tumor cells and the unique network of highly specialized tissue-resident cells, including parenchymal hepatocytes, liver sinusoidal endothelial cells, hepatic stellate cells, and Kupffer cells. Recruited and resident immune cells such as neutrophils can also influence the metastatic potential of DTCs in the liver. In this section, we summarize the requirements for the formation of a hospitable hepatic niche in metastatic breast cancer (Fig. 1).

Cancer Cell Adaptations to the Liver Niche

Metastatic breast cancer cells can achieve angiogenesis-independent growth in the liver by replacing hepatocytes and coopting sinusoidal blood vessels (Stessels et al. 2004). This hijacking may be crucial for overt liver metastasis, since lesions that use the existing blood supply were shown to thrive in the liver microenvironment (Martin et al. 2010; Frentzas et al. 2016). Postextravasation, liver-metastatic breast cancer cells undergo metabolic reprogramming to acclimatize to the hepatic microenvironment. In a mechanism dependent on the HIF-1α target pyruvate dehydrogenase kinase-1 (PDK1), breast cancer cells adapt a specific glycolytic phenotype in the liver in response to nutrient limitations and hypoxia (Dupuy et al. 2015). This distinct metabolic profile is crucial for efficient hepatic metastasis, since PDK1-knockdown resulted in an 80% reduction in liver metastasis in vivo but had no effect on outgrowth in the lung (Dupuy et al. 2015). Niche-induced metabolic plasticity may also drive endocrine resistance in liver metastatic ER⁺ breast cancer (Zuo et al. 2022). In response to fulvestrant treatment, ER⁺ metastatic tumors from the liver but not lung demonstrated increased glucose metabolism in vivo, differential ERα activity, and reduced treatment response. Interestingly, this metabolic vulnerability could be exploited by giving mice a fasting-mimicking diet to block glycogen accumulation in the liver, resulting in an improved fulvestrant response (Zuo et al. 2022).

Cancer Cell–Niche Interactions

Interactions between breast cancer cells and the liver ECM is a crucial step in the formation of a favorable hepatic metastatic niche. This is mediated by the tight-junction protein claudin-2, which is up-regulated in liver-trophic breast cancer cells and liver metastases from breast cancer patients (Tabariès et al. 2011). By enhancing cell surface integrin expression, claudin-2 drives cancer cell adhesion to the liver ECM and promotes metastatic colonization. Expression of the adhesion molecule CD44 is also highest in liver-trophic breast cancer cells, and its induction fosters breast cancer liver metastasis in vivo (Ouhtit et al. 2007; Erin et al. 2013). In addition, ECM remodeling in response to liver fibrosis creates a growth-permissive microenvironment capable of enhancing metastatic colonization (Cox et al. 2013). This is mediated by enhanced lysyl oxidase (LOX) expression in activated hepatic stellate cells, which triggers collagen cross-linking and favors the outgrowth of breast DTCs.

Cross talk between liver-infiltrating breast cancer cells and tissue-resident hepatocytes has also been linked to successful metastatic colonization. As with matrix adhesion, hepatocyte–breast cancer cell interactions are facilitated by claudin-2, although this heterotypic interaction occurs independently of integrin complexes (Tabariès et al. 2012). Reduction of claudin-2 levels in liver-aggressive breast cancer cells significantly decreases metastasis in vivo, which can be rescued by specifically restoring breast–hepatocyte interactions using chimeric claudin-2 constructs. This study supports the historic observation that breast cancer cells, upon seeding the liver, extend cellular projections through the fenestrated endothelium and make direct contact with hepatocytes (Roos et al. 1978). Mechanistically, breast DTCs and hepatocyte cross talk may support metastatic outgrowth by triggering mesenchymal-to-epithelial transition. Hepatocytes have been shown to directly induce E-cadherin reexpression in neighboring breast cancer cells, driving reversion to an epithelial phenotype and boosting postextravasation survival and chemoresistance (Chao et al. 2012).

The liver-metastatic potential of breast DTCs is highly reliant on interactions with infiltrating prometastatic neutrophils within the liver microenvironment (Tabariès et al. 2015). A distinct subset of immature low-density neutrophils (iLDNs) is preferentially mobilized in mice bearing liver metastases and is required for efficient hepatic colonization (Hsu et al. 2019). This metabolically distinct subset exhibits an enhanced bioenergetic capacity, enabling the execution of prometastatic functions such as NETosis under conditions of nutrient deprivation. Both this group and others (Yang et al. 2020) have demonstrated the functional importance of NETosis during breast cancer liver metastasis, with a significant reduction in metastatic colonization observed following the treatment of mice with the nuclease DNase1. In the latter study, the authors demonstrated a profound abundance of NETs in liver metastases of patients with breast and colon cancers, and identified a specific NET-DNA receptor, CCDC25, that mediates NET-dependent liver metastasis (Yang et al. 2020).

Hepatic Niche Perturbations

Recent evidence suggests the hepatic metastatic niche can be greatly influenced by systemic perturbations in the host. For example, a high-fat diet was shown to enhance spontaneous breast cancer metastasis to the liver, but not the lungs, by enhancing the accumulation of myeloid cells with immune-suppressive functions that dampen antitumor immunity (Clements et al. 2018). In a mouse model of obesity, metastatic breast cancer cells triggered triglyceride lipolysis in adjacent hepatocytes, the products of which were then transferred to tumor cells via fatty acid transporter protein 1 and used as a source of energy for tumor growth. This reciprocal relationship led to a dramatic increase in liver metastasis (Li et al. 2020b).

Interestingly, it was recently shown that weaning-induced involution in the postpartum liver conveys a metastatic advantage to seeding breast cancer cells via the induction of immune tolerance (Bartlett et al. 2021). Mechanistically, the massive influx of immune cells, hepatocyte cell death, and substantial ECM remodeling that accompanies liver involution resulted in the formation of a prometastatic hepatic niche, which favored the transition of micrometastases to overt metastatic lesions (Bartlett et al. 2021). This finding is in line with a study that found an increased propensity for liver metastasis in young postpartum breast cancer patients compared to age-, tumor-, and subtype-matched nulliparous women (Goddard et al. 2019). Thus, in addition to tumor-induced niche evolution, normal physiological remodeling of secondary sites may also support the formation of a prometastatic niche.

PREMETASTATIC NICHE

In the previous sections, we have outlined the symbiosis of metastatic cancer cells and their host organ, which is a prerequisite for cancer cells to thrive in a secondary site. However, it is well established that distant tissues are not idle, and niche evolution does not just begin upon the arrival of DTCs. Instead, the engraftment of cancer cells is preceded by active modifications in the target tissue to create a permissive microenvironment, the premetastatic niche, for the subsequent establishment of metastatic foci (for review, see Peinado et al. 2017). Premetastatic niches are generated through a complex interplay of soluble factors and exosomes released from the primary tumor, the mobilization and recruitment of immune cells to the intended site, and tumor-elicited reprogramming of tissue-resident cells (Fig. 1). This specialized microenvironment sets the stage for subsequent metastasis, allowing newly arriving DTCs to gain a foothold in the foreign tissue by supporting their survival and growth during seeding.

Tumor-Driven Education of a Distant Site

Cross talk between cancer cells and secondary organs is a requirement for the formation of a tumor-supportive premetastatic niche and is augmented by soluble factors released from cancer cells. An excellent example of tumor-mediated niche preparation is the systemic mobilization and accumulation of myeloid cells in intended metastatic sites such as the lung and brain, prior to cancer cell infiltration (Liu et al. 2013; Wculek and Malanchi 2015). In the premetastatic lung, CD11b⁺Ly6G⁺ neutrophils support metastatic initiation by selectively expanding the subpool of arriving breast DTCs with high metastatic potential (Wculek and Malanchi 2015). Similarly, neutrophil accumulation in the premetastatic liver, triggered by primary tumor-derived tissue inhibitor of metalloproteinases (TIMP)-1, increases its susceptibility to breast cancer metastasis (Seubert et al. 2015).

The preparation of secondary tissues for metastatic growth can also be augmented by hypoxia in the primary tumor, triggering the release of protumorigenic factors. Hypoxic breast tumors secrete LOX, which accumulates in premetastatic lungs and catalyzes collagen remodeling, facilitating the recruitment of metastasis-promoting myeloid cells (Erler et al. 2009). The secretion of LOX and other ECM-modifying family members is critically regulated by hypoxia-inducible factor (HIF-1), and the suppression of HIF activity abrogates premetastatic niche formation and metastatic colonization (Wong et al. 2011). More recently, HIF-dependent expression of ADAM12 was found to promote breast-to-lung metastasis through the activation of EGFR signaling, endowing cancer cells with an increased capability for migration, invasion, and outgrowth in distant tissues (Wang et al. 2021a). In the premetastatic bone, hypoxia-induced LOX secretion by primary tumor cells can also modulate bone homeostasis, triggering RANKL-independent osteoclastogenesis and the formation of bone lesions (Cox et al. 2015). The requirement for premetastatic osteolytic lesions for successful bone metastasis was also demonstrated by an earlier study, in this instance triggered by RANKL production by tumor-educated T cells (Monteiro et al. 2013). More recently, breast cancer–derived RSPO2 and RANKL was shown to recruit osteoclast progenitors and promote osteoclastogenesis in the premetastatic niche via Wnt signaling modulation (Yue et al. 2022). Regardless of their origin, focal osteolytic bone lesions serve as a highly supportive niche to foster the colonization of newly arrived breast DTCs.

Nonmalignant cells within the primary tumor microenvironment can also prime cancer cells for organ-specific metastasis. CXCL12 and IGF-1 secreted by CAFs in the primary site can select for cancer cells fit to thrive in the bone environment, leading to gain in the predisposition of cancer cells for bone metastasis (Zhang et al. 2013). Similarly, microenvironment-derived TGF-β primes departing cancer cells for lung metastasis by inducing the expression of ANGPTL4, facilitating lung seeding via destabilization of the pulmonary vasculature (Padua et al. 2008).

Exosome-Mediated Niche Preparation

Tumor-derived exosomes are small membrane-encapsulated vesicles (30–100 nm) that carry a cargo of functional biomolecules including DNA, RNA, proteins, and lipids. By facilitating intercellular communication between cancer cells and their intended niche, they play a crucial role in shaping distant microenvironments for subsequent metastatic colonization. Seminal work by the Lyden group demonstrated the influence of exosomal priming on metastatic organotropism. Exosomal integrins a₆β₄ and a₆β₁ from lung-tropic breast cancer cells was found to fuse preferentially with lung fibroblasts and epithelial cells, promoting lung metastasis; whereas, integrin a_vβ₅ specifically binds to Kupffer cells, mediating liver tropism (Hoshino et al. 2015). Thus, the preferential transfer of exosomal integrins to resident cells in intended metastatic sites elicits organ-specific niche preparation. Breast tumor–derived exosomes can also foster immune suppression in the premetastatic microenvironment. NK cell attenuation and reduced adaptive immune surveillance favored metastatic growth in the lung and liver following the uptake of breast tumor–derived exosomes (Wen et al. 2016), while exosome uptake has also been linked to proinflammatory activation of distant macrophages (Chow et al. 2014). In brain metastasis, exosomal cell migration–inducing and hyaluronan-binding protein (CEMIP) secreted by brain metastatic cancer cells induces inflammation and vascular remodeling in the perivascular niche, supporting brain metastatic colonization (Rodrigues et al. 2019). In addition, metastatic outgrowth in the brain and lung niche can be fostered by exosomal annexin II, which promotes tPA-dependent angiogenesis and macrophage-mediated premetastatic niche formation (Maji et al. 2017).

Tumor-driven education of distant tissues can also be achieved by exosome-mediated transfer of cancer cell–derived microRNAs (miRNAs). In the bone, the uptake of miR-21-containing exosomes by osteoclasts was recently shown to promote premetastatic niche formation, priming metastasis through increased osteoclastogenesis and osteolysis (Yuan et al. 2021). Osteoclastogenesis can be induced by exosomal miR-19a from bone-trophic ER⁺ breast cancer cells, suggesting a potential mechanism for the increased incidence of bone metastasis in ER⁺ patients (Wu et al. 2021b). In an elegant demonstration of the diversity of biological processes influenced by miRNAs, exosomal uptake of miR-122 was found to orchestrate metabolic reprogramming in lung and brain-resident premetastatic niche cells. This fueled metastatic outgrowth by increasing the glucose availability for arriving cancer cells (Fong et al. 2015). In a more direct example of miRNA-induced metastasis, the uptake of miR-105 (Zhou et al. 2014) and miR-181c (Tominaga et al. 2015) can promote metastatic seeding through the destruction of endothelial tight junctions and loss of barrier function in distant tissues, including the lung and brain.

Importantly, chemotherapy was recently shown to stimulate the release of tumor-derived prometastatic extracellular vesicles (EVs), including exosomes. Enriched in annexin-6, chemotherapy-elicited EVs induced endothelial cell activation, Ccl-2 induction, and Ly6C⁺CCR2⁺ monocyte expansion in the premetastatic lung niche, resulting in augmented breast-to-lung metastasis (Keklikoglou et al. 2019). The monitoring of cancer-associated EVs has shown promise as biomarkers for the management of metastatic breast cancer. Protein profiling of EVs from the plasma of cancer patients could accurately monitor and predict treatment responses in metastatic breast cancer patients, as well as serve as an independent prognostic factor for progression-free survival (Tian et al. 2021).

Distant Tissues as Active Players

Host cells in the premetastatic niche are themselves capable of influencing the metastatic cascade and encouraging the recruitment and outgrowth of breast DTCs. Thus, the remodeled secondary organ serves as a tumor accomplice, coordinating a second line of attack to facilitate successful metastatic colonization of that tissue. For example, the accumulation of NETs in the premetastatic liver and lung acts as a chemotactic factor to attract breast cancer cells and enhance distant metastasis (Yang et al. 2020). This is orchestrated via the specific NET-DNA receptor CCDC25 present on cancer cells, which senses extracellular DNA and responds by activating cancer cell motility. In the premetastatic lung, mesenchymal stromal cells display potent metastatic-promoting activity via the production of complement C3 (Zheng et al. 2021). C3 promotes neutrophil recruitment and NET formation, which fosters subsequent breast-to-lung metastasis. The role of premetastatic niche cells in aiding metastasis has also been demonstrated in the bone microenvironment. Bone-derived CXCL12 preferentially recruits tumor cells expressing CXCR4 (Müller et al. 2001; Devignes et al. 2018). In the latter study, the boost in CXCL12 in the circulation was attributed to HIF activation in osteoprogenitors residing in hypoxic bone marrow niches. This led to a CXCR4-dependent increase in breast cancer cell proliferation and dissemination (Devignes et al. 2018). Bone metastasis is also promoted by bone-derived RANKL, which favors the recruitment and colonization of RANK⁺ breast DTCs (Jones et al. 2006).

Recent studies have shown that premetastatic conditioning of secondary organs can also be elicited by tumor-independent mechanisms, via local or systemic perturbations in the host. Obesity, a condition associated with chronic, low-grade inflammation, enhances neutrophil accumulation in the premetastatic lung niche and supports subsequent lung metastatic seeding via elevated IL-5 and GM-CSF (Quail et al. 2017). In addition to neutrophil recruitment, obesity can drive neutrophil reprogramming in the premetastatic lung, enhancing ROS production and NETosis (McDowell et al. 2021). This results in a loss of vascular integrity, and boosts extravasation of breast DTCs. Even in obese tumor-free mice, their lungs demonstrate key features reminiscent of the premetastatic niche formed by primary tumors, including myeloid cell accumulation, increased expression of inflammatory cytokines, and elevated collagen deposition (Hillers-Ziemer et al. 2021). This response, orchestrated by obesity-activated lung stromal cells, suggests obesity can create a premetastatic niche that is conducive for metastatic growth in the absence of a primary tumor. Local tissue insults can also elicit a tumor-supportive environment in a secondary organ that fosters subsequent metastasis. Chronic nicotine exposure promotes the influx of prometastatic neutrophils that secrete lipocalin 2 (LCN2), promoting breast tumor cell colonization and metastatic outgrowth (Tyagi et al. 2021). Similarly, activated neutrophils recruited to the lungs following radiation-induced injury elicit tissue perturbations that strongly fuel metastatic colonization (Nolan et al. 2022). This tumor-supportive preconditioning of the lung environment was governed by enhanced regenerative Notch signaling in the lung epithelium, a process dependent on neutrophil activation (Nolan et al. 2022). Together, these findings suggest the status of the host tissue, independent of tumor-elicited changes, significantly influences its capacity to support metastatic growth.

CONCLUDING REMARKS

Metastatic disease is a devastating and usually incurable complication of cancer that remains the underlying cause of death for the majority of breast cancer patients. In this work, we have provided an overview of organ-specific breast cancer metastasis, with a focus on the site-dependent interactions between tissue-resident cells and arriving cancer cells that govern successful colonization of the lung, bone, brain, and liver environments (summarized in Table 1). The ability of both metastatic cancer cells and their target organ to coevolve and cooperatively support outgrowth is perhaps the most remarkable and intriguing step in the metastatic cascade. This partner-in-crime relationship is complex, dynamic, and context-dependent, influenced by tumor intrinsic factors, the unique players, and environmental constraints within each organ, as well as by local and systemic perturbations of the host. Undoubtedly, a deeper understanding of the tissue-specific prerequisites for metastatic growth will lay the foundation for the development of more effective therapies for metastatic breast cancer.

Table 1.

Comparison of metastatic niches in breast cancer, highlighting the key processes and cellular subsets that govern successful metastatic seeding and outgrowth

Metastatic niches	Lung	Bone	Brain	Liver
Immune recruitment and activation	TAMs (CCL2-recruited) Neutrophils (pro- and antitumor) NK cells, T cells	TAMs (CCL2-recruited) Myeloid-derived suppressor cells	Myeloid cells (CCL2-recruited) Protumor neutrophils Distinct TAM subsets	Protumor neutrophils (iLDNs)
NETosis	Enhanced by inflammation			NET-DNA receptor CCDC25
T-cell suppression	γδ-T-cell/neutrophil interplay Myeloid-derived osteopontin Oxygen sensing	Immune-suppressive myeloid cells (chemotherapy) Adipokines	Astrocyte immunomodulation Arg1/PD-L1+ neutrophils	Immune-suppressive myeloid cells (high-fat diet, involution)
Vascular engagement	Distinct growth pattern: angiogenic or vascular cooption		Distinct growth pattern: angiogenic or vascular cooption	Distinct growth pattern: angiogenic or vascular cooption
Metabolic reprogramming	↑ Electron transport, oxidative stress, antioxidant programs	↑ Serine synthesis, lactate production	↑ FA utilization, gluconeogenesis, antioxidants, autophagy Neuron cooption (GABA, glutamate)	↑ Glycolysis (PDK1-dependent) Endocrine resistance ↑ FA utilization (obesity)
Resident cell interactions	Lung alveolar cells Stromal cells Endothelial cells	Osteogenic niche Perivascular niche Adipocytes	Reactive astrocytes Microglia reprogramming Neurons, endothelial cells	Hepatocytes Hepatic stellate cells Endothelial cells
ECM remodeling	↑ ECM proteins (TNC, POSTN)	Bone resorption (vicious cycle) RANKL/OPG dysregulation		Matrix adhesion LOX-induced collagen cross-linking, fibrosis
Fibroblast activation	CAF-induced stemness Dynamic reprogramming EMT modulation			Stellate cell activation, ↑ LOX expression
Premetastatic niche	Myeloid cell infiltration LOX-induced ECM remodeling Exome-mediated priming NET accumulation	LOX-induced ECM remodeling Exome-mediated priming Bone-derived factors (CXCL12/RANKL)	Myeloid cell infiltration Exome-mediated priming	Myeloid cell infiltration Exome-mediated priming NET accumulation

(TAM) Tumor-associated macrophage, (iLDNs) immature low-density neutrophils, (FA) fatty acid, (GABA) γ-aminobutyric acid, (PDK1) pyruvate dehydrogenase kinase-1, (TNC) tenascin C, (POSTN) periostin, (CAF) cancer-associated fibroblast, (EMT) epithelial-to-mesenchymal transition, (LOX) lysyl oxidase, (ECM) extracellular matrix.

There is still much to be learned about the molecular mechanisms of breast cancer metastasis. Recent studies have begun unraveling the influence of tumor extrinsic factors such as tissue injury, metabolic diseases, physical exercise, and normal physiological processes (such as postpartum involution) on the development of tumor-supportive niches, a list certain to grow in the future. More work is needed to understand the complexity and evolution of metastatic niches, for example, the existence of spatially and temporally dynamic niches within the same organ, and whether an established metastatic niche in one organ can support niche development in other tissues, facilitating multiorgan metastasis. The influence of cancer-niche evolution on treatment response and the development of resistance is a topic of great interest to the field. Indeed, a recent study using optical barcoding revealed that the metastatic niche can profoundly influence the degree of intrametastatic breast tumor heterogeneity, a known barrier for treatment efficacy (Berthelet et al. 2021). Finally, the concept of “personalized niches” needs exploring, in which the intrinsic genomic profile of cancer cells plus the extrinsic host environment together dictate the development of patient-specific personalized niches.

It remains a daunting challenge to develop TME-targeting agents for cancer therapy, since it requires identification of the fundamental drivers of tumor-promoting niches. Progress has been made in breast-to-bone metastasis, with bone resorption inhibitors such as bisphosphonates (such as zoledronic acid) and denosumab (an anti-RANKL antibody) now widely used for the management of metastatic breast cancer. Other bone-modifying agents are emerging that have demonstrated potential clinical utility, including cathepsin K and Src inhibitors (for review, see Brook et al. 2018). The complexity and diversity of organ-specific niches present as major challenges to develop therapies that can be broadly applicable to a large fraction of patients. One possible approach is to identify common stresses experienced by metastatic cancer cells at different organ sites, and the shared fitness-promoting pathways that promote their survival under stress (Esposito et al. 2021b). Although rapid improvement has been made in our understanding of organ-specific breast cancer metastasis, it is clear that we have only scratched the surface of this remarkably complicated process. Probably an even bigger challenge will be translating our growing knowledge of cancer–niche interactions into better preventative and curative options for metastatic breast cancer patients.