Dormancy in cancer bone metastasis

Immunological surveillance

Immune surveillance describes a host's protective mechanism whereby the immune system recognizes and eliminates harmful pathogens (e.g., viral and bacterial infections), and also includes cancer immunoediting [ ]. Cancer immunoediting refers to the process that the host immune system exerts over the tumor to shape its progression. Three main phases are postulated to describe cancer immunoediting: (1) elimination, where the immune system inhibits tumor proliferation; (2) equilibrium, characterized by a balance of immune activity against the tumor and active proliferation of cancer cells; (3) escape, where the immune factors fail to diminish tumor growth, which later becomes a metastatic and clinical lesion [ ]. The immune dormancy fits into the equilibrium phase, as the tumor mass proliferates and also undergoes apoptosis due to the immune surveillance, resulting in a “dormant” tumor mass [ ].

Among the immune cells controlling the tumor fate, the CD8+ T cells are considered to be the key antitumor effectors due to their ability to recognize cancer antigens and induce cancer cytolysis, working in pair with the CD4+ T cells [ ] ( Fig. 29.1 ). The bone marrow of BCa patients presents higher CD4+/CD8+ cell quantities compared to healthy controls, while in in vivo models where the CD8+ T cells were depleted, a switch from dormancy to proliferation was acquired in cancer cells, implying the importance of the immune system in tumor dormancy [ , ].

In a recent study, Owen et al. demonstrated that tumor intrinsic type 1 interferon (IFN) signaling regulated the immune-mediated dormancy in PCa bone metastases, with enrichment of interferon regulatory factor 7 (IRF7) and enhanced CD8+ and CD4+ T cell activation [ ]. In a study using a spontaneous BCa bone metastasis model, bone metastatic BCa cells showed downregulation of IRF7 and its up/downstream genes, such as STAT1 [ ]. Mice inoculated with IRF7-overexpressed 4T1.2 cells benefit from significantly lower CD11b+ Ly6G+ cells; enhanced CD8+, CD4+, and NK cells; and diminished bone metastasis. In human BCa sample databases, low IRF7 expression was associated with increased bone metastasis development. The role of IFN/IR7 immune dormancy axis was also observed in ER- BCa cells that underwent dormancy after chemotherapy treatment and maintained a quiescent status by increased recruitment of CD8+, CD4+, dendritic cells, B lymphocyte, and reduced myeloid-derived suppressor cells (MDSCs) (CD11b+/Gr1+) [ ]. Altogether, these provide evidence to support that IFN plays a pivotal role as a main factor mediating immunogenic dormancy in bone.

Evidence suggested that CD8+ T cells also maintain dormancy via IFN-γ or type II interferons [ , ]. The IFN-γ/STAT1 signaling pathway has been linked to downregulation of cyclin E and cyclin A markers required for proliferation, as well as interaction with D1 cyclin and CDK4, leading to p21 and p27 upregulation, and CCA [ , ]. Paradoxically, chronic exposure to IFN-γ could promote tumor immune–mediated escape by inducing PD-L1/2 expressions and CTL antigen 4 (CTLA4). PD-L1 expressed by cancer cells binds to PD-1 receptor in T cells leading to T cell immune suppression and tumor-immune evasion, while CTLA4 is an inhibitor of the T cell response [ , ].

The CD11b+/Gr1+ MDSCs have also been found to be expanded in BCa cells in murine bone marrow [ ]. They represent a major source of transforming growth factor beta (TGF-β) that interacts with BCa cells, releasing GLI2 and PTHrP (factors intimately associated with osteoclast differentiation and osteolysis). Moreover, these MDSCs have the potential to differentiate into osteoclasts, while osteoclast-mediated bone resorption has been suggested to reactivate dormant cells [ ]. Therefore, in addition to their immunosuppressive role, the MDSCs could contribute indirectly to reawaken dormant cells.

More recently, macrophages have attracted attention due to their dual role in cancer. Macrophages can be classified into M1 and M2 phenotypes. The classic M1 macrophages (activated via IFN-γ, TLR, and LPS) are associated with inflammation, bacterial and virus phagocytosis, and antitumor responses. M1 macrophages release IL-1, IL-6, IL-23, IFN-γ, and IL-12 cytokines to recruit and activate CD8+ and NK cells to kill cancer cells [ , ]. The M2 macrophages (activated by IL-4, IL-10, and IL-13), also postulated as tumor-associated macrophages (TAMs), are involved in wound healing and can promote tumor growth, angiogenesis, metastasis, and immune suppression [ ]. TAMs release factors (CCL3, CCL4, CCL5, CCL22, IL-10, and TGF-β) that recruit regulatory T cells (nTreg) into the tumor microenvironment and inhibit CD8+ and CD4+ T cell functions [ ]. TAMs are also related to bone metastasis potentials and patients with PCa bone metastasis display increased CD206+ M2 macrophages in primary tumor biopsies [ ]. Intriguingly, new evidence indicates that M1/M2 macrophages may play unconventional roles in cancer dormancy in bone. As assessed by Walker et al., M2 macrophages in the bone marrow stroma can connect to BCa cells by forming gap junctional intercellular communication (GJIC). This binding reduces BCa cell proliferation, leading to quiescence and chemoresistance to carboplatin. In contrast, M1 macrophages released exosomes that activated NF-κB in the dormant cells, shifting their dormant status into proliferation [ ]. This evidence regarding M1/M2 macrophages' roles in cancer dormancy is contradictory to their general biological functions in cancer. Nonetheless, the M2 macrophages polarize into further subsets (M2a, M2c, M2d) [ ] and it is likely that different M2 subsets, or even a novel subset not fully yet characterized, play a different role in tumor dormancy, which warrants further investigation.

Angiogenesis

To gain active proliferation, the tumor mass requires a large amount of nutrients derived from blood vessels surrounding the tumor microenvironment. A lack of blood supply into the tumor will result in “angiogenic dormancy,” where the tumor mass maintains a balance between proliferation and apoptosis due to impaired energy intake. In fact, it is estimated that tumors cannot grow more than 1–2 mm ³ without angiogenesis. The development of blood vessels is controlled by pro- and antiangiogenic factors. To escape dormancy, the tumor utilizes proangiogenic factors, stimulating the development of novel blood vessel networks, perpetuating its growth and expansion ( Fig. 29.2 ). This balance toward proangiogenic factors is known as “angiogenic switch” [ , ].

The heat shock proteins (HSPs), a family of chaperone proteins secreted by cells under stress conditions, are involved in angiogenesis and cancer dormancy. In cancer, HSPs are secreted by tumor cells to ensure adequate protein folding and expression, secure tumor proliferation, and survival [ ]. Studies imply a role for HSP27 to promote bone physiology and BCa bone metastasis [ , ]. Inhibition of the HSP27 has been found to induce dormancy in BCa cell lines [ ]. Of interest, suppression of HSP27 did not affect dormant tumor proliferation but increased the apoptosis rate. This is consistent with the angiogenic tumor mass dormancy concept, showing a balance among proliferation and apoptosis due to a lack of nutrients provided by the vasculature. Furthermore, Huston et al. examined the functional inhibition of HSP90 and its client protein Akt in multiple myeloma (MM), in a bone marrow–like environment [ ]. Suppression of HSP90 and Akt led to MM CCA, osteoclastogenesis inhibition, and remarkably, angiogenesis. Another HSP related to cancer progression is the HSP70-2. In samples obtained from BCa patients and BCa cell lines, overexpression of HSP70-2 was identified [ ]. Moreover, inhibition of HSP70-2 induced CCA and significantly reduced tumor growth in vitro and in vivo .

Furthermore, the preferential localization of BCa cells near the perivascular region and their associated dormant status have been supported by real-time in vivo imaging in xenograft models and in patient samples [ ]. However, whether neoangiogenesis reawakens or maintains dormant cancer cells in bone is still under debate. Allocca et al. found that BCa cells were preferentially located near type H (CD31 ^hi Emcn ^hi ) micro blood vessels [ ]. The type H microvessels are the main contributor to neoangiogenesis and found in reduced quantity in aged mice [ ]. This is correlated with the new finding by Singh et al. that aged bone blood vessels promote BCa dormancy exit, a possible consequence of the loss of PDGF-B expressing, prodormant type H microvessels with aging [ ]. This observation is contradictory to studies by Ghajar et al. which suggest that while stable microvasculature promotes dormancy, sprouting neovasculature disrupts dormancy, where endothelial-derived thrombospondin-1 (TSP-1) plays critical roles [ ]. Both TSP-1 and thrombospondin 2 (TSP-2) are potent angiogenic suppressors derived from endothelial cells. Their antiangiogenic property arises from their interaction with CD36 expressed on endothelial cells or interaction with the very low-density lipoprotein receptor. TSP-1 also inhibits angiogenesis via integrin-associated protein (IAP or CD47) interaction, leading to decreased nitric oxide, imperative for vasculature homeostasis [ ]. TSP-1 not only limits angiogenesis but, in addition, induces dormancy of BCa DTCs residing in the perivascular niche of bone tissue. The quiescent BCa cells were found surrounding the mature and stable endothelium vessels characterized by high TSP-1 levels, whereas tumor outgrowth was found in sprouting neovascular tips with low TSP-1 expression [ ]. The molecules POSTN, tenascin, versican, fibronectin, and TGF-β1 were found enriched on these neoangiogenic tips, which are known for encouraging tumor growth in bone [ ]. This debate encourages further studies to clarify the role of neoangiogenic factors in tumor dormancy, combining with other prodormancy factors, particularly the osteoblastic niche.