Perspective Chapter: Breast-Tumor-Derived Bone Pre-Metastatic Disease – Interplay between Immune and Bone Cells within Bone Marrow Microenvironment

Abstract

The bone marrow is a dynamic organ where osteogenesis and bone remodeling take place side by side with hematopoiesis and the maintenance of immunological memory. It provides a unique microenvironment favoring the colonization and outgrowth of breast cancer cells. The outcome of breast-cancer-derived bone metastases depends on the formation of a pre-metastatic niche, which is initiated through “education” of non-tumoral cells present in the primary cancerous niche. Among other participants, immune cells and their secreted factors can boost the successful seeding of the distant disease. In this chapter, we discuss the reciprocal interplay between bone and T and B cells, particularly in pathological contexts. In the first part, we are exploring the knowledge brought by the osteoimmunology field, especially from the best studied disease in this area, rheumatoid arthritis. In the second part, we summarize the latest findings on underlying cellular and molecular mechanisms for breast-cancer-derived bone pre-metastatic niche formation. In addition, we explore the concept that breast-tumor-primed T and B cells function as messengers from the periphery to the bone marrow, alter bone turnover homeostasis in favor of osteoclasts, before tumor colonization, leading to a pre-metastatic niche formation to further the development of bone metastases.

1. Introduction

1.1 Bone marrow: an overview

Bones provide both skeletal scaffolding and a unique microenvironment for hematopoiesis and B cell ontogenesis, osteogenesis, and also function as an immunological memory reservoir, in its marrow [1, 2, 3]. The bone marrow (BM) is a complex and dynamic structure composed of different and distinct compartments or niches, which accommodate a multitude of cell types, which functionally create an interactive network, critical for BM/bone integrity [2, 4, 5, 6, 7, 8]. These niches are composed of different stromal cell types—osteoblasts (OBs), osteocytes, reticular and perivascular cells, endothelial cells, mesenchymal cells (MSCs), smooth muscle cells, macrophages, and dendritic cells (DCs). Disruptions in these compartments can lead to aberrant pathological processes [2, 8, 9, 10, 11].

At least three niches can be identified in the BM: (i) the endosteal/subendosteal niche that supports self-renewal and differentiation of hematopoietic stem cells (HSC); (ii) the central niche for multipotent progenitors (MPP); and (iii) the perisinusoidal niche that guarantees the differentiation to the lineage committed progenitors and hematopoietic cells full commitment [4, 11, 12, 13, 14]. The endosteal/subendosteal niche contains OBs, bone-forming cells, and osteoclasts (OCs), bone-resorbing cells, as well MSCs, all collaborating to regulate hematopoietic homeostasis and osteogenesis [13]. The central and perisinusoidal niches recruit MSCs, as well as endothelial cells and their progenitors, to promote HSCs proliferation, mobilization, and differentiation,—(i) myelopoiesis, the process in which innate immune cells, such as granulocytes and monocytes, develop from a myeloid progenitor cell; and (ii) lymphopoiesis, the process in which adaptive and innate lymphocytes develop from a lymphoid progenitor cell [5, 15].

More recently, the anatomy of myelopoiesis in the BM was partially mapped in situ and the clonal relationships between myeloid progenitors and surrounding cells were assessed [16]. It was demonstrated that colony stimulating factor 1 (CSF1), also known as macrophage colony-stimulating factor (M-CSF), produced by perisinusoidal vessels provides a unique niche that regulates, spatially organizes, and controls myeloid differentiation [16]. This type of study provides valuable information on the organization of the various niches inside the BM. Dissection of these processes will allow a better understanding of their influence on bone and/or BM and vice versa, during homeostasis and local and/or systemic diseases, which directly or indirectly affect bone/BM homeostasis.

1.2 Bone metabolism and its central players

Bone tissues in adults are classified as: (i) cortical (long and compact bone) and (ii) trabecular (flat and spongy or cancellous bone) [17, 18]. During fetal development, long bones are modeled by endochondral ossification, in which the cartilage formed by chondrocytes—cells of mesenchymal origin, essential for formation and maintenance of cartilage—is replaced by bone at the edge of the growth plate [19, 20]. The cortical bone is mostly structural, supporting the stability and movement of the body, made up of compactly packed osteons—its key structural unit, formed by layers called lamellae, surrounding the Harvesian canal, which contain small blood vessels responsible for blood supply to osteocytes, former OBs embedded in the bone matrix as differentiated cells [21, 22]. The trabecular bone is highly porous and vascularized and harbors red and white bone marrow [17]. Bone matrix is composed of an organic segment, formed by type I collagen secreted by OBs, and a variety of non-collagenous proteins, such as osteocalcin and osteopontin; and an inorganic segment, also known as bone mineral, formed by calcium, phosphorus, and magnesium, which originates the hydroxyapatite [Ca₁₀(PO₄)₆(OH)₂] [17].

Even after the modeling phase, bone tissue is constantly renewed by a process called bone remodeling. Bone homeostasis is achieved by the performance of bone remodeling system, which is conducted by the synchronized activities of OBs, OCs, and osteocytes [23, 24, 25]. Osteocytes are the most abundant cells in bone tissue and play an essential role in bone homeostasis [26]. They translate mechanical—pressure and tension—low oxygen, matrix mineralization, and hormonal stimuli into biochemical signals, due to their extensive long cytoplasmic extensions. The complex network formed by osteocytes in the bone matrix, enables direct communication among them and other effector cells in the bone/BM, including OBs and OCs [21, 22, 26, 27, 28, 29]. OBs, derived from mesenchymal progenitors, promote mineralization and bone formation by secreting matrix vesicles containing type I collagen, alkaline phosphatase, and osteocalcin [22]. OCs, derived from myelomonocytic progenitors, otherwise, dissolve and absorb bone matrix by releasing hydrogen ions that acidify the bone interface and secrete lysosomal enzymes—such as tartrate-resistant acid phosphatase and cathepsin K [30, 31, 32, 33, 34].

The receptor activator of nuclear factor-κB (RANK)/receptor activator of nuclear factor-κB ligand (RANKL)/osteoprotegerin (OPG) molecular system is the most important pathway activated during bone remodeling process [35, 36, 37, 38]. Notably, BM stromal cells are responsible to initiate osteoclastogenesis, being the main sources of M-CSF and RANKL [39]. Osteocytes first release M-CSF causing myelomonocytic progenitors to commit to the OC line [26, 39]. M-CSF stimulates RANK expression in the late stages of OCs development, which interact with RANKL expressed on or secreted by OBs and osteocytes [26]. This interaction leads to the activation of mitogen-activated protein kinase (MAPK) and nuclear factor-κB (NF-κB) pathways, through tumor necrosis factor receptor–associated factor 6 (TRAF-6) and c-Fos molecules [30, 31, 33, 34, 40], giving rise to large multinucleated differentiated mature OCs [40]. RANKL activation also induces the expression of nuclear factor of activated T cells c1 (NFATc1), the master transcription factor for osteoclastogenesis [41]. B-lymphocyte-induced maturation protein 1 (Blimp1), which can be induced by NFATc1, downregulates the expression of the transcriptional factors interferon (IFN) regulatory factor 8 (IRF-8) [42, 43] and B-cell lymphoma 6 (Bcl6), in turn promoting osteoclastogenesis [43].

Notably, mice that lack RANKL or its receptor RANK develop severe osteopetrosis accompanied by a defect in tooth eruption due to a complete lack of OCs [42, 43]. Conditional deletion of RANKL in chondrocytes [44, 45] and OBs led to a severe osteopetrosis [22, 43, 45, 46, 47], whereas osteocytes-specific RANKL-deficient mice displayed a high bone mass phenotype at the adult stage [43, 45]. Thus, chondrocytes and OBs are the major source of RANKL in supporting osteoclastogenesis during skeletal development, whereas osteocyte-derived RANKL contributes to bone remodeling at the adult stage [44, 45] . In humans, loss-of-function mutations in Tnfrsf11a (gene encoding RANK) and Tnfsf11 (gene encoding RANKL) genes cause autosomal recessive osteopetrosis with a complete lack of OCs [43, 48].

OBs and osteocytes also settle the termination of osteoclastogenesis [21, 27, 29, 49]. This step initiates through the secretion of OPG—the RANKL decoy receptor, the main counter-regulator of osteoclastogenesis, which attenuates bone resorption by binding to RANKL with higher affinity than RANK and blocking RANKL osteoclastogenic effects [50]. Of note, mice lacking Tnfrsf11b (gene encoding OPG) exhibited severe osteoporosis due to an increased OC number and severe bone resorption [43, 50, 51, 52]. The same cells control the beginning of osteoblastogenesis, and several molecules regulate this next step, including parathyroid hormone (PTH), the RUNX Family Transcription Factor 2 (RUNX2), osterix transcription factor, bone morphogenetic protein (BMP), and the Wnt pathway [40, 53]. The Wnt signaling pathway is the most important player in osteoblastogenesis, preventing apoptosis of OBs and accelerating its cell cycle progression and proliferation, leading to inhibition of adipogenesis [54]. Wnt molecules activate G-protein-coupled receptors and coreceptors of the low-density lipoprotein receptor (Lrp) family, resulting in β-catenin activation, effectively upregulating aerobic glycolysis, β oxidation, and other anabolic mechanisms, through activation of the RUNX2 gene [54]. Moreover, binding of BMP to BMP receptors leads to their dimerization followed by phosphorylation of Smad proteins (main signal transducers for receptors of the TGF-β superfamily), which in turn also activate RUNX2, upregulating OB activity and differentiation [54, 55].

More recently, leucine-rich repeat-containing G-protein-coupled receptor 4 (LGR4) was reported to be another receptor for RANKL, which negatively regulates osteoclastogenesis by not only competing with RANK for RANKL binding, but also inhibiting NFATc1 activation via Gq protein alpha subunit (Gαq) [56]. Interestingly, OCs can also regulate the activity of OBs, by secreting bone morphogenetic protein-6 and sphingosine-1-phosphate, which function as coupling factors promoting OBs proliferation and bone formation [57]. Finally, osteocytes negatively regulate osteoblastogenesis by secretion of Dickkopf-1 (DKK-1) and sclerostin molecules, both antagonists of the Wnt pathway [58, 59]. Sclerostin is a marker for mature osteocytes, and its expression increases with age [60, 61], and mice deficient in this molecule show an increase in osteoblastogenesis and a decrease in the shape of BM cavities, resulting in impairment of hematopoiesis and B cells ontogenesis [62].

Taking together, we conclude that intra and intercellular and molecular interactions between osteocytes, OBs, OCs, and chondrocytes are crucial for maintaining the BM/bone niches, under physiological conditions. Currently, we know that bone remodeling process is also regulated by immune cells, residing at, or migrating to BM, such as T and B cells, innate lymphoid cells, macrophages, DCs, and other hematopoietic cells [63]. Any imbalance in one of these connections can lead to several bone pathologies, including, among others, breast-cancer-derived bone metastases [64].

2. Reciprocal interplay between bone and immune cells

2.1 An overview of the “osteoimmunology” field

The relationship between bone and immune systems has been suggested by pioneering studies reported in the early 1970s and showed that molecules secreted from immune cells were capable to induce OC activation and differentiation [65, 66]. Moreover, early studies in the immunology field, using genetically deficient mice in various immunomodulatory molecules, showed unexpected phenotypes in the skeletal systems under physiological conditions [40, 63, 67, 68]. Actually, we know that bone and immune systems share a variety of molecules, including cytokines, chemokines, transcription factors, and signaling molecules [67]. By interacting with each other in the BM, the bone and immune cells cooperatively conduct a series of bone and immune system functions [67]. Studies conducted on bone and immune phenotypes are revealing the physiological significance of the mechanisms shared by both systems [67], and the interdisciplinary field “osteoimmunology” was created to explore these mechanistic interactions, under physiological or pathological conditions [69].

The RANK/RANKL/OPG molecular system is considered the most important pathway explicitly linking immune and bone tissues [35, 38, 43, 70, 71]. Indeed, several studies are showing that RANK and RANKL, besides being the master regulatory via inducing osteoclastogenesis, also play multiple roles in the immune system, including: (i) differentiation of medullary thymic epithelial cells (mTECs) [72, 73, 74, 75]—that act as mediators of the central tolerance process, which self-reactive T cells are eliminated while regulatory T cells are generated; (ii) secondary lymphoid tissue organogenesis—the organization of the microarchitecture of lymph nodes (LNs) [42], formation of germinal centers in gut isolated lymphoid follicles [42] and Peyer’s patches [42]; and (iii) fine-tuner of adaptive immune response—enhancement of DCs longevity and survival [76], maintenance of immunological memory [77] and B cells ontogenesis [78, 79]. Of note, these molecules are expressed by cells from both systems [63]. OPG, for example, is expressed by mature B cells (accounting alone for almost 40% of OPG produced in BM. Their essential role for bone homeostasis was shown in vivo, since B-cell-deficient mice have low bone mass density associated and a marked deficit in BM OPG [80]. This homeostatic balance is achieved by B and T cells interaction, via CD40-CD40L molecules, since mice depleted of CD40 or CD40L co-stimulatory molecules presented a decline in OPG production by B cells and an increase in bone resorption and low bone mass density [80]. Also, mice depleted from T cells showed a complete suppression of OPG production by B cells followed by an increase in osteoclastogenesis and bone loss [80]. Moreover, Cytotoxic T-Lymphocyte Associated Protein 4 (CTLA-4)—a molecule expressed by T cells that helps keep immune responses in check—binds to CD80/CD86 co-stimulatory molecules expressed by OCs, leading to inhibition of osteoclastogenesis mediated by RANKL or TNF-α [81]. CTLA-4 binding to CD80/CD86 in OCs’ precursor cells induces the expression of indoleamine 2,3 dioxygenase (IDO), which in turn degrades tryptophan and leads to OCs apoptosis [82]. Consequently, mice deficient in CD80, CD86, or IDO have increased osteoclastogenesis rates and osteopenic phenotypes [82, 83] demonstrating that CTLA-4 plays important roles in the physiological regulation of bone mass preservation [81, 82, 83].

We should be aware that most of these findings were conducted in animal models; however, new indications are emerging to support the reciprocal roles of both systems in human diseases aspects [47, 67, 84]. Despite the more recent observations about the impact of immune cells for bone tissue homeostatic integrity, and vice versa, the interplay between both systems is first spotlighted by studies on bone disorders, triggered by abnormal immune responses activation, like the ones seen in rheumatoid arthritis (RA), post-menopausal osteoporosis, chronic periodontitis, multiple myeloma, fractures, HIV chronic infection, and bone metastases [67].

2.2 Role of T and B cells in bone disorders

T and B cells are derived from the same lymphoid progenitor cell during hematopoiesis and are the main cellular representatives of the adaptive immune system, so called because they do not mount an immediate response to an antigen (Ag). The Ags are recognized by specific receptors—T cell receptor (TCR) and B cell receptor (BCR), which are diverse at the population level and clonal and unique at individual cellular level. TCRs and BCRs are not conserved and are generated by gene rearrangements during T and B cell ontogenesis. T cells ontogenesis takes place in the thymus, while B cells ontogenesis is in BM—both are primary lymphoid organs.

After maturation inside BM or thymus, B and T cells gain the peripheral blood circulation and enter the secondary lymphoid organs. In lymphoid organs, as LNs and spleen, activated/educated by dendritic cells (DCs), the professional Ag-presenting cell (APC) are found. Through their ability to sense changes in their local environment and respond appropriately, DCs activate T cells by the expression of the Major Histocompatibility Molecules (MHC), in complex with linear, short, peptides Ags (9–20 amino acids long). This complex is recognized by T cells via TCR and CD3 ε and δ, ζ chains accessory molecules and their categorized cluster of differentiation (CD) surface expressed molecules, CD4 or CD8. In addition, T cells concomitantly recognize co-stimulatory molecules and cytokines, which will define their functional differentiation fates, in terms of their expression of master transcription factors and functional cytokines [85]. CD4⁺ helper T cells are divided into specialized subsets, known as: (i) T helper 1 (Th1), expressing T bet transcription factor and IFN-γ; (ii) T helper 2 (Th2), expressing GATA-3 transcription factor and IL-4, IL-5, and IL-13; (iii) T helper 17 (Th17), expressing ROR γT transcription factor and IL-17A, IL-17F, IL-22, and IL-26; (iv) T helper 22 (Th22), expressing Runx1 and RORγt transcription factors and IL-22; T follicular (Tfh), expressing B cell lymphoma 6 (Bcl6) transcription factor and IL-21; and (v) T regulatory (Treg) cells, expressing FoxP3 transcription factor and TGF-β and IL-10; while CD8⁺ T cells fall into subpopulations, known as: (i) Cytotoxic Type 1 CD8⁺ T cells (Tc1), expressing T bet and BLIMP-1 transcription factors, IFN-γ, granzyme, and perforin; (ii) Type 2 CD8⁺ T cells (Tc2), expressing GATA-3 transcription factor and IL-4, IL-5 and IL-13; (iii) Type 17 CD8⁺ T cells (Tc17), expressing ROR γT and ROR α transcription factors and IL-17A, IL-17F and IL-22 and T reg CD8⁺ T cells, expressing IL-10 [85].

B cells are also activated in secondary lymphoid organs, but, in contrast to T cells, they do not need APCs to present their cognate Ags, which will be freely recognized in linear or structural forms. At the beginning of immune responses, B cells secrete immunoglobulins M (IgM) independently of T helper cells. The T-cell-independent response is short-lived and does not result in the production of memory B cells, which will not result in a secondary response to subsequent exposures to the same Ags. However, to induce stronger B cell responses and to generate immunological memory, B cells need help from T follicular CD4⁺ T cells (Tfh). Indeed, to enable homing to B cell follicles, Tfh expresses abundant C-X-chemokine receptor type 5 (CXCR5). Another characteristic of Tfh is the expression of CD40 ligand (CD40L), inducible T cell costimulator (ICOS), programmed death-1 (PD-1), and B and T lymphocyte attenuator (BTLA). Tfh cells colocalize with Ag-specific B cells within germinal centers (GCs), which are transient structures located within B cell follicles, in secondary lymphoid tissues, in which somatic hypermutation of immunoglobulin (Ig) variable region genes and selection of high-affinity B cell clones occur. Immunoglobulin class switch (IgA, IgE, and IgG) will be defined by cytokines produced by these different specialized Tfh, at the moment of B cells activation.

It is clear now that the identity of T cell subsets is critical in guiding their role on bone remodeling system, during homeostasis or in pathological conditions [67]. In particular, Th1, Th2, Th17, Th22, and T reg CD4⁺ and CD8⁺ cells have been shown to influence bone metabolism [67, 86, 87, 88, 89]. In the RA scenario—the best studied human disease in osteoimmunology—the importance of Th17 CD4⁺ T cells is evident, beginning by their infiltration into the synovium and the association of disease susceptibility with specific variants of T-cells-related genes, such as HLA-DR (MHC class II cell surface receptor encoded by the human leukocyte Ag gene complex), Protein Tyrosine Phosphatase Non-Receptor Type 22 (PTPN22), and C-C Motif Chemokine Receptor 6 (CCR6) [40, 63]. Moreover, studies performed aiming to confirm the role of T cells showed that T cell deficient mice are protected from arthritis, and clinical trials performed to inhibit effector T cells activities demonstrate the improvement of clinical symptoms [46, 90, 91]. IL-17A, one of the cytokines secreted by Th17 CD4⁺ T cells, amplifies local inflammation and the production of TNF-α and IL-6, which in turn promote RANKL expression by induction of an intense osteoclastogenesis [37]. Th17 CD4⁺ T cells also express RANKL, but this molecule only stimulates an additive effect and is not sufficient to induce osteoclastogenesis, independently, in this disease scenario [37, 43, 46]. It was also reported that these cells stimulate the recruitment of OCs progenitors via increasing chemokine production by BM MSCs [40]. Recently, it was shown that IL-22, produced by the Th22 CD4⁺ T cells, promotes osteoclastogenesis and enhances bone destruction in arthritic mice [46, 92]. Disease severity is shown to be markedly reduced in collagen-induced arthritic mice deficient in IL-22 [92], and elevated IL-22 in serum is also associated with disease activity in patients with RA [92].

Interestingly, it has been found that a particular type of Th17 CD4⁺ T cells, derived from FoxP3⁺ Treg CD4⁺ T cells (called exFoxP3 Th17 T cells), have a much stronger pro-osteoclastogenic activity than conventional Th17 CD4⁺ T cells [86, 93]. Under arthritic conditions induced in mice model, FoxP3⁺ Treg CD4⁺ T cells lose FoxP3 by the action of IL-6 produced by synovial fibroblasts [46, 83, 94]. Indeed, FoxP3⁺ IL-17⁺ CD4⁺ T cells—a transition state during the conversion to exFoxP3 Th17 T cells—are frequently observed in synovial tissues of patients with active RA, as compared with those with inactive RA, suggesting a pathogenic role for this subset in this pathological condition [46, 95]. Equally important is the fact that Foxp3⁺ IL-17⁺ CD4⁺ T cells were also observed in periodontal tissues of patients with severe periodontal disease [96, 97]. Notably, in a ligature-induced periodontitis mouse model, it was recently shown that Th17 CD4⁺ T cells eradicate the bacteria while also inducing bone degradation and tooth loss, which is crucial for the termination of oral infection, avoiding bacterial systemic dissemination [98]. Taken together, it was concluded that Th17 CD4⁺ T cells orchestrate the host defense against oral microbiota by regulating both osteoclastic bone resorption and antimicrobial immunity [98].

It was reported that IL-4 produced by Th2 T cells inhibits OCs formation and function in vitro [86, 99, 100]; nonetheless, no functional activity has been reported in vivo. On the other hand, Th1 CD4⁺ T cells, which counter regulate Th2 cells, are found in the synovium fluid of patients with active RA [101], although it has been demonstrated that the secretion of IFN-γ by this T cell subset strongly inhibits osteoclastogenesis and protects against bone tissue degradation by OCs [102]. IFN-γ induces a strong inhibition of the RANKL-induced activation of the NF-κB, via a rapid degradation of TRAF6 [102]. In arthritic synovium, Th1 CD4⁺ T cells are not considered to be activated but often display an exhausted phenotype and express low levels of IFN-γ [86, 87, 88, 89].

It is already known that FoxP3⁺ Treg CD4⁺ cells play an indispensable role in maintaining immune homeostasis, but also exert a strict anti-osteoclastogenic activity [68, 103, 104, 105, 106]. In rheumatic patients, the number of FoxP3⁺ Treg CD4⁺ T cells is inversely related to osteoclastogenic markers and disease severity [68, 105, 107]. These results accompany findings in which mice deficient in FoxP3⁺ Treg CD4⁺ T cells were prone to arthritis, showing joint destruction and generalized bone loss, supported by higher number of OCs in joints [105]. The reintroduction of FoxP3⁺ Treg CD4⁺ T cells into these mice significantly reduced arthritic clinical symptoms [105]. As discussed in previous section, OCs express the co-stimulatory molecules CD80 and CD86, and osteoclastogenesis can be regulated via CTLA-4, promoting OCs apoptosis, and thus suppressing bone destruction [81]. Notably, BM resident FoxP3⁺ Treg CD4⁺ T cells express higher levels of CTLA-4, than peripheral FoxP3⁺ Treg CD4⁺ T cells [82]. These resident FoxP3⁺ Treg CD4⁺ T cells remove CD80/CD86 from the surface of OCs precursor cells by CTLA-4 mediated trans-endocytosis, potentially leading to reduced co-stimulation by OCs [82]. Therefore, the interaction between OCs expressing CD80/CD86 and FoxP3⁺ Treg CD4⁺ T cells expressing CTLA-4 is suggested as important player for the cross talk between these cells to support bone homeostasis [81, 82].

More recently, the term immunoporosis—a subarea under osteoimmunology’ umbrella—was proposed for the field that studies the importance of immune system for osteoporosis establishment [108]. Osteoporosis—defined by a loss of bone mass and microarchitecture, has a multifactorial etiology but endocrine factors such as hyperparathyroidism, vitamin D deficiency, and menopause are primarily implicated [109]. The disease stems mainly from the cessation of ovarian function, where declining estrogen levels result in the stimulation of bone resorption, leading to a period of rapid bone loss [109]. At the cellular level, the central mechanism by which sex steroid deficiency induces bone loss is via an increase in OC formation and life span [110].

Estrogen exhibits the potential to stimulate the differentiation and survival of regulatory T cells, which in turn suppress the expression of proinflammatory cytokines from Th17 T cells and inhibit bone resorption. In addition, many genetic and non-genetic factors intensify the negative impact of estrogen deficiency on the skeleton, including gut microbiota profile [109]. Indeed, sex steroid deficiency increases gut permeability, allowing intestinal microbiota to activate and expand Th17 and TNF-α⁺ T cells [111]. These expanded T cells increase S1PR1 (sphingosine-1-phosphate receptor 1) expression, which promotes their egress from intestine and influx into BM through CXCR3 and CCL20-mediated mechanisms [111, 112]. Additionally, this steroid deficiency-associated bone loss was prevented by probiotics administration [111, 112]. In this regard, several studies demonstrated that Lactobacillus species alleviates gut inflammation and improved barrier function of intestine [113]. Moreover, it was shown that Lactobacillus rhamnosus administration enhances bone mass in eugonadal mice [112, 114, 115], inhibits osteoclastogenesis, and skews balance of Th17 T cells to regulatory T cells, under in vitro and in vivo conditions [112, 115]. Collectively, these studies highlight the osteoprotective role of this probiotic, thereby opening novel avenues in the management and treatment of postmenopausal osteoporosis.

Finally, the effect of B cells in bone is more bidirectional, as compared with T cells, as B cells require the endosteal BM surface to their ontogenesis [78]. Indeed, B cell transcription and growth factors that control B cell differentiation play important roles in bone homeostasis, indicating the tight interaction between this immune cell lineage and bone [116]. In RA, the autoimmune process starts with the presentation of auto-Ags to CD4⁺ Th T cells, which help B cells to differentiate into plasma cells that produce auto-Abs, such as rheumatoid factor and anti-cyclic citrullinated peptide (anti-CCP), trademarks of this disease [106, 117, 118]. Of note, the substantial number of Tfh cells in the synovial tissue correlates with disease severity [119, 120]. Auto Abs and immune complexes promote bone erosion through FcRγ signaling in OC precursor cells or innate immune cells [121, 122]. More recently, it was demonstrated that plasma B cell numbers increased in BM region near the inflammatory joints during arthritis [119], due to enrichment of plasma B cells survival factors such as IL-6, BAFF, and APRIL [119]. Locally, plasma B cells provide RANKL, TNF-α, IL-17, and Ab-mediated costimulatory signals that cooperate to powerfully promote osteoclastogenesis [119]. Genetic ablation of RANKL in B cells resulted in amelioration of periarticular bone loss, but not of articular erosion or systemic bone loss, in RA [123], and was slightly but significantly protective of ovariectomy-induced bone loss [124].

After reviewing the progress on the central roles of adaptive immunity in the establishment of some bone disorders, we will now explore the knowledge behind the participation of tumor-primed T and B cells in the development of bone pre-metastatic niche, which will lead in turn to the establishment of breast-cancer-derived bone metastases.

3. Breast-cancer-derived bone metastases: molecular interactions within the BM

3.1 Preclinical and clinical implications

Breast cancer is the most frequent cancer in women, with increasing incidence and high mortality rates [125]. Breast-cancer-induced bone metastases are a frequent complication of advanced disease, with up to 70% of incidence, associated with skeletal complications, including pain, osteopenia and bone loss, pathological fracture, hypercalcemia spinal cord compression, BM aplasia, demanding surgery, and radiotherapy for bone complications, and change of antineoplastic therapy for bone pain [126, 127, 128, 129]. Collectively, these comorbidities are defined as skeletal-related events (SREs) that dramatically impair the patient’s quality of life and reduce overall survival [127, 128, 130].

The preservation of bone mass has been achieved using bone anti-resorptive bisphosphonates, such as zoledronic acid, and denosumab, an anti-RANKL monoclonal antibody, which block OC-mediated bone resorption and are approved for use in patients with cancer metastatic to bone [128]. However, these drugs only alleviate SREs complications, the development of bone metastases remains an incurable condition, and mortality rates are kept at elevated level [131]. Meta-analyses studies showed a statistically significant overall survival benefit with women treated with bisphosphonates [132, 133, 134]. It is not surprising, however, that bone-targeted therapies also display systemic immunological effects, regarding the interactions between immune and bone cells, which can partially cause eliminatory anti-tumor effects. Indeed, zoledronic acid and denosumab can modulate immune cells activity, such as γδ T cells, macrophages, and CD4⁺ Tregs, in many different types of cancer, including breast cancer, leading to an increase in T-cell-mediated anti-tumor cytotoxic effects [135]. Moreover, the knowledge about how current treatments affect the immune landscape in bone metastatic microenvironment is scarcely known. This fact could be due to our limited understanding of osteoimmunological interactions for tumor growth, the low availability of biopsies from bone metastases, and appropriate metastatic models for preclinical studies.

The risk factors for predicting breast-cancer-derived bone metastases are still controversial. In a recent study, a total of 2133 patients, including 327 with bone metastases (15.33%) and 1806 without bone metastases (84.67%), were retrospectively reviewed and showed that the spine is the most common site for bone metastases, including thoracic spine (63.61%) and lumbar spine (53.82%), followed by ribs (57.5%), pelvis (54.1%), and sternum (44.3%) [136]. The results also indicated that combined axillary LN metastases, high serum concentrations of cancer Ag 15-3 (CA15-3), alkaline phosphatase (ALP), and low level of hemoglobin have the highest predictive accuracy for bone metastases in breast cancer [136].

Breast-cancer-derived bone metastases give rise predominantly to most aggressive osteolytic lesions, although 15–20% of clinical cases present an osteoblastic pattern, resulting in a dysregulated bone deposition [126, 128, 137]. Notably, it has been shown that breast cancer osteolytic lesions may also lead to skeletal muscle atrophy and weakness, through bone-muscle cross talk, which in turn leads to a feed-forward cycle of musculoskeletal degradation [138, 139]. Osteoclastic bone resorption releases transforming growth factor-β (TGF-β), which causes oxidative stress and skeletal muscle Ca²⁺ leak and weakness, via the TGFβ-Nox4-RyR1 axis, inducing a muscle atrophy program [138, 139]. Interestingly, the same pattern was shown in both immunodeficient and immunocompetent mice, suggesting that adaptive immune system may be excluded from this pathological aspect [140, 141]. Moreover, it has been suggested that muscle dysfunction occurs prior to the loss of muscle mass—cachexia [142]. In addition, experimental strategies are being analyzed for skeletal muscle mass preservation, including: (i) the blocking of myostatin signaling [143, 144]; and (ii) antagonizing the growth hormone secretagogue receptor (GHSR)-1a [145, 146]. Both strategies showed improved survival in mice with cancer cachexia [147].

Recently, it was reported that breast cancer cell lines and human breast cancer tissue express sclerostin, suggesting that breast cancer cells impair bone formation while promoting bone resorption [140]. In a mouse model of bone metastases, the pharmacological inhibition of sclerostin by setrusumab—an anti-sclerostin monoclonal antibody, reduced bone metastatic burden and destruction, without increasing metastases at other sites [140]. Moreover, this treatment protected from induction of muscle atrophy and loss of function, leading to prolonged life span [148]. Accordingly, the expanding and maintenance of OBs functional properties were then proposed as an approach to restore bone and muscle integrity, in the context of metastases-induced osteolytic disease [140]. In parallel, it was reported that homeodomain protein TG-interacting factor-1 (Tgif1)—an inducer of osteoblastogenesis acting at Wnt and PTH1R-dependent signaling pathways, is increased in OBs upon stimulation by metastatic breast cancer cells [141]. High levels of Tgif1 were associated with poor patient survival in breast cancer [147]. The lack of Tgif1 in OBs increases Semaphorin 3E (Sema3E) expression and attenuates breast cancer cell migration as well as metastases formation, indicating that Tgif1 plays a role during the early stages of bone metastases establishment [141]. Therefore, the mechanisms driving the early steps of bone metastatic process are still not sufficiently understood and the induction of osteoblastogenesis should be analyzed with caution, since OBs and their molecules seem to play contradictory roles in breast-cancer-derived bone disease.

Finally, preclinical studies suggested that non-coding RNAs (ncRNAs) such as long ncRNAs, microRNAs, and circular RNAs are crucial regulators of breast-cancer-induced bone metastases [149, 150, 151]. Indeed, unique miRNA expression patterns were reported in different breast cancer subtypes, displaying both pro- and anti-tumorigenic functional properties [150]. In fact, lower levels of miR-34a were observed in patients suffering from later stages of breast cancer in comparison to benign breast disease and healthy controls [131], while higher expression of miR10b was observed in breast cancer patients with LN and bone metastases [148, 152]. Furthermore, lower levels of miR-124 in primary breast cancer correlate with shorter bone-metastases-free survival [153], and miR-218 serum levels are higher in patients with breast cancer bone metastases when compared with patients without metastases [154]. Currently, since altered expression of miRNAs has been associated with disease progression and clinical outcome, these molecules are emerging as potential therapeutic targets and prognostic biomarkers in the context of bone metastases.

3.2 “The vicious cycle”

Bone metastases are not established randomly, instead they request a complex reciprocal interplay between primary cancer cells and BM microenvironment stroma. Indeed, BM stroma provides an advantageous architecture for bone colonization, playing critical roles for breast cancer cells initial seeding, dormancy, and outgrowth [137]. Circulating breast cancer cells enter BM by the sinusoids—small blood vessels lined with fenestrated endothelial cells, more permissive than other types of capillaries [155, 156]. After extravasation into BM, they migrate to the perisinusoidal or to the endosteal/subendosteal niche, where OBs and other stromal cells secrete a variety of chemo-attracting factors, such as CXCL12, RANKL, osteopontin, and BMPs [155, 156]. Breast cancer cells express high levels of CXCR4—the receptor for CXCL12, which increase their ability to survive in BM and the establishment of overt metastases in this microenvironment [155, 156]. Moreover, CXCL12 stimulates PI3K-AKT signaling pathway and Src activity, which enhance cancer cell survival in challenging environments [41]. These results obtained from animal model studies were validated in clinical datasets, in which Src and CXCR4 expression in tumor cells was associated with breast cancer bone relapse [155]. In BM niches, metastatic cells adapt, survive, and reside for a prolonged period of time—possibly years or even decades [137].

When invading breast cancer cells escape from dormancy, they disrupt the normal bone remodeling process in order to promote their outgrowth, eventually leading to the development of overt bone metastases [137]. Metastatic breast tumor cells express and secrete a series of molecules, such as parathyroid hormone-related protein (PTHrP), IL-11, and TNF-α, vascular cell adhesion molecule 1 (VCAM1), intercellular adhesion molecule 1 (ICAM1), lysyl oxidase (LOX), RANK, RANKL, and IL-6, which in turn mobilize and activate OCs to resorb bone matrix and release chemotactic stimuli and additional growth factors attached to the bone matrix [126, 130, 157]. Bone matrix degradation by the hyperactivated OCs releases TGF-β, which in turn is activated due to pH changes in the local environment and proteolytic cleavage from latent peptides [128]. In sequence, this molecule triggers the production of osteolytic factors, such as PTHrP, IL-11, IL-1β, and Jagged1 from breast cancer cells [156, 158, 159, 160, 161, 162, 163, 164, 165]. Jagged1 promotes osteoclastogenesis via Notch signaling in pre-OCs, while PTHrP induces the production of RANKL by OBs [32]. Activated OCs then degrade the bone matrix on cortical and trabecular surfaces, leading to the release of numerous growth factors, including more TGF-β [158]. Consequently, TGF-β-induced Jagged1 enhances a vicious cycle between bone and tumor cells, by stimulating the expression of IL-6 from stromal cells and OBs, promoting tumor growth [158]. VCAM1 also stimulates the outgrowth of metastases, through the recruitment of OCs progenitors via expression of integrin α4β1 [166]. Importantly, several therapies using denosumab—monoclonal antibody against RANKL [167]; and monoclonal against human Jagged1 [168] or small-molecule inhibitors—have already been approved for clinical use or are under development to treat osteolytic bone metastases, by preventing progression of the vicious cycle [169, 170].

In the last few years, accumulating evidences suggest that breast cancer bone colonization is preceded by changes in BM microenvironment [171, 172, 173]. In this context, a pre-metastatic niche is established by cellular and molecular mechanisms, mostly educated by the primary breast cancer cells [64, 171, 174]. Therefore, tumor cells prepare BM microenvironment to host them, before “switching homes” and moving to bone [171]. Importantly, the pre-metastatic niche formation also leads to the disruption of bone remodeling system, in favor of osteoclastogenesis and bone consumption, but prior to metastatic cells arrival [64, 157, 174]. Accordingly, the pre-metastatic osteolytic lesions facilitate subsequent bone tumor colonization [175].

3.3 Bone pre-metastatic niche formation

3.3.1 The “seed and soil” in bone tissue adaptation

Breast cancer cells migration to bone is innately related to the molecular and cellular components provided by the pre-metastatic niche, in sequential and distinct phases [137, 172, 176]. In fact, in the nineteenth century, Stephen Paget proposed that tumor cells (“seeds”) only grow in specific and permissive microenvironments (“fertile soil”) [177]. Of note, BM is a fertile microenvironment, composed of hematopoietic cells, MSCs, endothelial cells, OBs, OCs, molecules secreted by breast primary tumor, either as soluble or contained in extracellular vesicles (EVs) or exosomes, and immune migrating cells [137, 175]. However, how and when these factors, produced locally or systemically, regulate the crucial mechanisms behind the establishment of this site remains less clear [175].

Recent studies suggest that bone pre-metastatic niche exists prior to metastatic colonization; however, disseminated breast cancer cells are detectable in BM prior to clinically detectable bone metastases [175, 178]. Interestingly, patients without any metastases harbored disseminated breast cancer cells with less genetic heterogeneity compared with the primary tumor or those disseminated cells isolated from bone metastatic patients [175, 178]. Of note, less than 0.1% of disseminated breast cancer cells survive during circulation and homing [179, 180, 181]. Based on these findings, we can speculate that bone/BM stromal cellular and molecular components probably play roles in supporting these mutations, for further licensing and selection of the best “seeds” to adapt in the pre-metastatic niche, until their overt bone colonization.

Additionally, a recent study identified LOX-derived by hypoxia condition, a factor significantly associated with bone tropism and relapse. LOX induces an intense osteoclastogenesis, through NFATc1, before, and independent of breast tumor cells arrival at BM [174]. Therefore, this study identified a previous step in bone metastases development, triggered by these osteolytic lesions, opening new opportunities for therapeutic intervention [174]. In fact, in a previous study using an intracardiac mouse model of breast-cancer-derived bone metastases, animals treated with a nonspecific LOX inhibitor—β-aminopropionitrile—reduce bone colonization when administered at the time of tumor inoculation [182].

As mentioned in the last section, recent evidence suggests that breast-cancer-derived miRNAs play key roles in tumor development and progression via exosomes transfer, regulating the outgrowth and metastases of breast cancer [183, 184]. Of note, it was described that miR-21, a highly conserved oncomicro RNA, is expressed in serum of breast cancer patients, significantly higher as compared with healthy controls [185]. Moreover, it was demonstrated that miR-21 induces OCs differentiation, by directly binding programmed cell death 4 (PDCD4), upregulation of NFATc1, and suppression of c-Fos transactivation [186, 187]. Indeed, it was showed that breast cancer cell–secreted exosomes containing miR-21 lead to an exacerbated osteoclastogenesis, which contributes to the generation of a pre-metastatic niche and further enhancing bone metastases development [188]. Importantly, the expression level of miR-21 was detected at higher level in serum exosomes of breast cancer patients with bone metastases, as compared with patients without bone metastases [188].

Almost 20 years ago, a pioneer study challenged the molecular basis for bone metastases. Using human breast cancer cell lines with elevated metastatic activity, it was determined a breast-cancer-derived bone metastases gene signature, which included genes involved in: (i) BM homing (CXCR4); (ii) extracellular matrix alteration (Matrix Metallopeptidase 1 (MMP1), ADAM metallopeptidase with thrombospondin type 1 motif 1 (ADAMTS1), and proteoglycan-1); (iii) angiogenesis (Fibroblast growth factor 5 (FGF5), and Connective tissue growth factor (CTGF); and osteoclastogenesis (IL-11) [189]. Moreover, the overexpression of this gene set is superimposed on a poor prognosis already present in the parental breast cancer population, suggesting that metastases require a set of functions beyond those underlying the emergence of the primary tumor [189]. Thereafter, several other bone metastases gene signatures were proposed, such as Src-dependent [190] or Irf7-regulated genes [191]. To date, it remains unclear the clinical significance and applicability of these gene signatures described, either by tumor heterogeneity in primary and secondary sites or by differences in tumor sources.

3.3.2 Role of breast-tumor-primed T and B cells in bone pre-metastatic niche formation

Primary breast cancer has been shown to “prepare” distant organs for tumor cell colonization even before their arrival [171, 192, 193]. Immune cells such as macrophages [194, 195], DCs [196], neutrophils [197], and T cells [64, 195, 198, 199] are associated with the formation of the pre-metastatic niches, highlighting the importance of basic mechanisms responsible for tumor cells distant establishment [171, 200]. Accordingly, it has been found that cells of the immune system acting as pro-tumor cells are enriched in the pre-metastatic niches and support cancer cell seeding via paracrine signaling and/or by suppressing anti-tumor immune cells [171, 172, 200, 201].

Particularly, our group previously showed that spontaneous bone metastases development, originated from 4 T1 triple negative breast tumor model, depends on RANKL production by tumor primed CD3⁺ T cells [64]. This conclusion was achieved by adoptive T cell transference to nude mice, which shows that 4 T1 primed T cells, in the total absence of tumor cells, induce a pre-metastatic osteolytic disease [64]. Moreover, inhibition of RANKL production (using shRNA) in fresh tumor-primed T cells does not generate osteolytic disease and the associated bone pre-metastatic niche. Consequently, development of bone metastases is completely absent. Taking together, we proposed an extra step to Mundy’s vicious cycle where initial bone consumption, mediated by pre-metastatic CD3⁺ T cells, generates a rich microenvironment that license further colonization of the bone cavity by the metastatic clones [64]. Once the initial seeding of the bone tissue is achieved, tumor cells shall continue the osteolytic process on their own, feeding themselves through the vicious cycle established within the bone microenvironment [64].

As pre-metastatic osteolytic disease happens much before metastatic colonization, it is not known how the tumor Ag would get to the BM to be recognized by T cells. This is important because T cells’ effector functions depend on peptide recognition complexed to MHC molecule, a function better exerted by DCs. Since DCs can carry Ag from peripheral tissues via lymphatics to LNs, and also travel from the peripheral tissue into the blood and to the BM [202, 203], we envisage at least two nonexclusive possibilities for Ag presentation and recognition: (i) cancer-derived exosomes could travel to the bone cavity and provide tumor Ags to be processed and presented by local resident DCs [204, 205] and/or (ii) DCs loaded with tumor Ags at the primary tumor or at the tumor draining LNs, can migrate to the BM where Ag presentation would take place [203, 206]. Moreover, it is already known that BM can prime naive T cells and recruit effector T cells, also serving as a site for CD4⁺ and CD8⁺ T cells proliferation [202].

In addition, DCs display a high developmental and functional plasticity depending on local factors and stimuli encountered during their differentiation and maturation, providing a multitude of necessary signals for shaping immune responses [207, 208, 209, 210]. Plasticity can also allow DCs to develop into other cell types, among them OCs (DC-OC), what is not unexpected considering their same origin from common myelopoietic stem cell progenitors [211, 212, 213]. Indeed, for the last 15 years, it has been reported that immature DCs can develop into OCs in vitro and in vivo, when cultured with osteoclastogenic factors, M-CSF and RANKL or RA synovial fluids containing pro-osteoclastogenic cytokines [212, 214, 215]. Independently of the presence of DCs at bone resorptive sites during inflammatory conditions [211, 216, 217, 218, 219, 220, 221], their direct contribution to bone resorption, either as APCs, keeping osteoclastogenic Th17 T cells locally activated, or overcoming their own phenotype differentiating into OCs mature functional phenotype, has yet to be solved. Indeed, it has been confirmed that multinucleated giant cells expressing markers of DCs and OCs are located next to the bone in inflammatory bone disease [222].

In fact, we recently addressed the role of DCs in breast-tumor-derived bone metastases context [223, 224]. We showed that DC-OC differentiation is induced by RANKL, either recombinant or produced by specific-tumor T cells [224], and they can act as both an APC for 4 T1 tumor-specific T cells and as an OC-like cell (DC-OC), amplifying the osteolytic phenomena before bone tumor colonization [224]. Furthermore, it is already known that the pretreatment of DCs with high levels of RANKL leads to enhancement of [76] and augments their ability to stimulate T cell proliferation [225, 226, 227]. Therefore, we can suppose that RANKL-enriched environment setup by osteoclastogenic CD3⁺ T cells located inside the BM probably contributes to a higher DC survival ratio, which in turn would support T cells’ activities in promoting the pre-metastatic niche formation [224]. Additionally, DC-OCs, but not BM-OCs, are incredibly good in activating T cell proliferation and cytokine secretion [224], and secrete high amounts of IL-23, which in turn boosts IL-17 and RANKL production by T cells, feeding the positive osteoclastogenic loop of adaptive T cell immunity [224]. This positive loop has IL-23 as one limiting step since blocking IL-23 with monoclonal antibody inhibits T cell IL-17 and RANKL production [224]. Adding more information to our work, recent data published [228] showed that monocyte-derived macrophages, rather than bone-residing macrophages, are critical for breast-carcinoma-derived bone metastases outgrowth in vivo, in IL-4R and CCR2-dependent manners [228].

More recently, we described that 67NR non-metastatic tumor cells—an in situ breast carcinoma sibling of 4 T1 tumor cell line, can modify distant sites promoting bone physiological alterations, increasing in trabecular bone mass on day 11 post-tumor implant [229]. This observation was associated with an expansion of the osteoblastic lineage cells accompanied by a reduction of OCs numbers [229]. Moreover, CD8⁺ T cells express an anti-osteoclastogenic cytokine milieu enriched by IFN-γ, IL-10 and low levels of RANKL, and the frequency of BM-derived CD8⁺ FoxP3⁺ regulatory T cells, as defined as potent suppressors of osteoclastogenesis, was also increased in such animals [229]. This milieu was capable to suppress 4 T1 tumor-specific CD4⁺ T cells phenotype in vivo and in vitro and strongly inhibited bone metastases establishment, restoring trabecular bone mass volume [229]. We concluded that the 67NR⁺ tumor derived CD8⁺ T cells phenotypes, either contributing to bone homeostasis and/or control of 4 T1 breast tumor pre-metastatic disease, interfere with OCs and OBs activities inside BM. Our study highlights the opposing roles of subverted tumor CD4⁺ and CD8⁺ T cell subtypes in directing breast cancer progression and bone metastases establishment. Furthermore, this likely reflects the fact that modification of the distant bone site by 67NR breast tumor disfavors pre-metastatic bone niche formation [229].

4. Conclusions and perspectives

Altogether, we assume that the set of our studies are revealing the cellular and molecular dynamic interactions behind breast-cancer-derived pre-metastatic bone niche formation (Figure 1). There are still many questions about the factors that determine the chemotaxis of cells, which Ags or ligands are needed, as well as the modulating elements of this distant system. So far, we know, that the immune system is central. Multiple cues need to be investigated to translate our current knowledge toward clinical impact. If these immunophenotypes patterns are confirmed in human disease, this complex network can be used either as prognostic tools or even as therapeutic targets.

Figure 1.

Left panel: anti-osteoclastogenic cytokines produced by 67NR⁺ CD8⁺ T cells keep bone homeostasis mediated by OBs and OCs cross talk. Right panel: BM pre-metastatic niche formation by 4T1 Th17 RANKL⁺ CD4⁺ T cells and RANKL⁺ CD8⁺ T cells activities. DCs loaded with tumor Ags from primary tumor growth site prime naïve CD4⁺ and CD8⁺ T cells into draining LNs to differentiate into tumor-activated T cells producing RANKL and IL-17F, which in turn migrate to BM before tumor bone colonization. Migrating DCs can now activate the maintenance of tumor RANKL⁺ CD4⁺ Th17 T cells inside BM microenvironment dependent on IL-23 production. Dysregulation of bone homeostasis by an intense activation of BM-OCs by RANKL and M-CSF under breast tumor pre-metastatic osteolytic conditions induced by RANKL⁺ CD4⁺ Th17 T cells and RANKL⁺ CD8⁺ T cells. Both T cells activities support their osteoclastogenic potential in the establishment of the pre-metastatic niche. Left panel: Bone marrow scenario after subcutaneously 67NR non-metastatic tumor cells challenge. Production of anti-osteoclastogenic cytokines, IFN-γ⁺, IL-10, and anabolic levels of RANKL expressed by CD8⁺ T cells keep bone homeostasis mediated by OCs and OBs cross talk.

Funding

This work was supported by funds from Faperj (Foundation for Research Support of the State of Rio de Janeiro, E-26/203.056/2017, E-26/010.001925/2015, E-26/311.264/2021, SEI-260003/002756/2022; Fopesq 2021/UFF; CNPq (National Research Council, 309611/2018-0); and FOCEM (Fundo para a Convergência Estrutural do MERCOSUL).