Pathogenesis of osteoblastic bone metastases from prostate cancer

Authors

Errata

This article is corrected by:

  1. Errata: Erratum: Pathogenesis of osteoblastic bone metastases from prostate cancer Volume 116, Issue 10, 2503, Article first published online: 2 March 2010

  • We thank Dr. Ian Seymour for help with editing the article.

Abstract

Prostate cancer is the second leading cause of cancer-related death in men. A typical feature of this disease is its ability to metastasize to bone. It is mainly osteosclerotic, and is caused by a relative excess of osteoblast activity, leading to an abnormal bone formation. Bone metastases are the result of a complex series of steps that are not yet fully understood and depend on dynamic crosstalk between metastatic cancer cells, cellular components of the bone marrow microenvironment, and bone matrix (osteoblasts and osteoclasts). Prostate cancer cells from primary tissue undergo an epithelial-mesenchymal transition to disseminate and acquire a bone-like phenotype to metastasize in bone tissue. This review discusses the biological processes and the molecules involved in the progression of bone metastases. Here we focus on the routes of osteoblast differentiation and activation, the crosstalk between bone cells and tumor cells, and the molecules involved in these processes that are expressed by both osteoblasts and tumor cells. Furthermore, this review deals with the recently elucidated role of osteoclasts in prostate cancer bone metastases. Certainly, to better understand the underlying mechanisms of bone metastasis and so improve targeted bone therapies, further studies are warranted to shed light on the probable role of the premetastatic niche and the involvement of cancer stem cells. Cancer 2010. © 2010 American Cancer Society.

Solid tumors, such as breast and prostate cancer, have an affinity to metastasize to bone, causing osteolysis and abnormal bone formation. Bone metastasis starts with the tropism of cancer cells to the bone through specific migratory and invasive processes. Once in the bone marrow, metastatic cells are able to survive and grow. Here, they actively interact with bone marrow stem cells and hematopoietic progenitors in the so-called metastatic niche, where they acquire a bone-like phenotype. This leads to the formation of bone lesions (lytic or osteoblastic), obtained through reciprocal paracrine amplification and cell-to-cell communication with bone cells. The complex molecular pathogenesis mechanisms of bone metastasis offer several potential targets for prevention and therapy.

Prostate cancer is the second leading cause of cancer-related death in men, and a typical feature of this disease is its ability to metastasize to bone. Indeed, it has been estimated that >80% of men who die from prostate cancer develop bone metastases.1, 2 Although most bone metastases from prostate cancer have been classified as osteoblastic, based on the radiographic appearance of lesions, it is clear that bone resorption and bone formation are dysregulated.3-5

The Metastatic Process: From Primary Tumor to Growth in Bone Tissue

The first step in metastasization is the acquisition of motility and invasiveness; capabilities that are not compatible with normal tissue. Cancer cells must therefore shed many of their epithelial characteristics, detach from epithelial sheets, and undergo a drastic alteration, a process referred to as the epithelial-mesenchymal transition.6 Acquisition of this invasive phenotype is reminiscent of that which occurs during early embryonic development.7 In malignancy, genetic alterations and the tumor environment can both induce epithelial-mesenchymal transition (EMT) in tumor cells. The important steps that facilitate metastasis appear to be reversible,7 and cannot be explained solely by irreversible genetic alterations, indicating the existence of a dynamic component to human tumor progression. In cancer, although the PI3K/Akt pathway is the primary inducer of epithelial-mesenchymal transition, the Wnt/B-catenin, Notch, Ras, integrin-linked kinase, and integrin signaling pathways are also involved.8-11 The main characteristics of epithelial-mesenchymal transition in prostate cancer cells, and the complex molecular interactions that facilitate this process, are summarized in Figure 1.

Figure 1.

Genetic modifications of prostate cancer cells from primary tumors to bone tissue are shown. UPA indicates urokinase-type plasminogen activator; PTHrP, parathyroid hormone-related protein; IGFBF, insulin-like growth factor-binding protein; TGF-β, transforming growth factor-β; IGF, insulinlike growth factor; BMP, bone morphogenetic protein; ET-1, endothelin 1; PGDF, platelet-derived growth factor; FGF, fibroblast growth factor; VEGF, vascular endothelial growth factor; RANKL, receptor activator of the nuclear factor-κB ligand.

After epithelial-mesenchymal transition, prostate cancer cells must go through a multistep process to metastasize to bone, which involves dislodgement from a primary site, survival in the circulation, binding to the resident cells in bone, and survival and proliferation in the bone and bone marrow.12-18 The dissemination of prostate cancer cells may take place early in disease progression with tumor cells preferentially engaged in the bone marrow, and a subset of cells surviving and evolving into clinically apparent disease. These cells then enter a period of dormancy in which they either stop proliferating, or proliferate at a reduced rate before showing evidence of metastasis; a process that can sometimes exceed 10 years.19-23 However, in some situations, there is at least 1 further and crucial event that takes place, the trigger that reactivates tumor cell dormancy. However, the mechanisms that facilitate this process remain unknown.

Circulating prostate cancer cells preferentially adhere to bone marrow endothelial cells and then migrate through the endothelial layer.24 This process involves a variety of adhesion molecules (eg, selectins, integrins, and cadherins) present on the surfaces of endothelial and metastatic prostate cancer cells. Cell adhesion and migration are mediated in part by many integrin-extracellular matrix interactions, especially those involving integrin αvβ3, whose expression is enhanced by stromal cell-derived factor-1 (SDF-1).25, 26 Several lines of evidence suggest that SDF-1 contributes to the pathogenesis of prostate cancer metastases. Inhibition of chemokines reduces the in vitro proliferation of PC-3 cells. Expression of the SDF-1 receptor, CXCR4, has been detected in prostate cancer cells, and SDF-1 has also been reported to increase their migration capacity.1, 27 This evidence implies that several bone paracrine factors regulate the expression of adhesion molecules in disseminated prostate cancer cells, and supports the interaction between cancer cells and resident bone cells, thereby contributing to the tropism of circulating prostate cancer cells toward bone. Cadherin 11 has been implicated in this process because of the finding that it is overexpressed in prostate cancer bone metastases but not in normal tissue or nonmetastatic cancer. Furthermore, in a mouse model, injection of PC3 cells did not lead to any bone metastasis if cadherin 11 was silenced.28 Monocyte chemotactic protein-1 (MCP-1) is chemotactic for prostate cancer cells, and the expression of its receptor, CCR2, correlates with pathologic stage.29-31 Lu et al reported that the activation of MCP-1/CCR2 axis promotes prostate cancer growth in bone.32 Once metastatic prostate cancer cells arrive in the bone, they are stimulated by growth factors present in the noncellular fraction of bone marrow, improving their interactions with the resident bone cells.

Metastasis suppressor genes encode a class of proteins that block the metastatic process without affecting primary tumor progression. These genes are rarely mutated, but are instead thought to be controlled by epigenetic events such as gene methylation. The first metastasis suppressor gene characterized, Nm23,33 when overexpressed in metastatic cell lines, has been shown to decrease metastatic competency in vivo, and motility and invasion ability in vitro.33-36 There is mounting evidence that many metastasis suppressor genes, including KAI1, CD44, and MAPK4, are able to block metastasis in prostate cancer cell lines, and several clinical studies have also supported their involvement.37-39KAI1, the first prostate cancer metastasis suppressor gene identified, is down-regulated in both metastatic and high-grade prostate cancer.37, 40 Although the mechanism of suppression is still unclear, it is believed that KAI1 interacts with several membrane proteins implicated in metastasis progression, such as E-cadherin, beta1 integrins, and epidermal growth factor receptor. CD44 is a widely expressed cell adhesion protein, and its down-regulation has been correlated to prostate cancer grade and metastatic stage.38, 41, 42 Similar results have also been reported for MAPK4.39

Osteoblast Function and Their Regulation

The same regulatory pathways essential for bone development and remodeling are probably also involved in prostate cancer bone metastases. Osteoblast growth and differentiation are regulated by complex signaling pathways mediated by growth factors such as bone morphogenetic proteins (BMPs), insulinlike growth factor (IGF)-I and IGF-II, transforming growth factor-β1 (TGFβ1) and TGFβ2, fibroblast growth factor (FGF), platelet-derived growth factor (PDGF), and Wnt.43-48 Moreover, other factors, such as vascular endothelial growth factor (VEGF), are thought to have an indirect action on osteoblast function through their modification of the bone microenvironment.

The transcription factor RUNX2 (also called core binding factor 1) stimulates BMP2 and FGF, and is essential for osteoblast differentiation.49 In mice, inactivation of RUNX2 results in a skeleton made entirely of cartilage, indicating that RUNX2 is necessary for bone formation.49 Another transcription factor, Osterix, is believed to control osteogenesis, based on the evidence that Osx null mice do not undergo any bone development.50 Furthermore, β-catenin, a factor that mediates the Wnt signaling pathway, has also been implicated in osteogenesis regulation. A Wnt coreceptor, which regulates bone formation through β-catenin, interacts with Wnt by forming a complex with the Wnt receptor, Frizzled.51 This complex activates the canonical Wnt signaling pathway, in turn stabilizing free cytoplasmic β-catenin, which is then translocated to the nucleus. Here, it heterodimerizes with transcription factors of the lymphoid enhancer-binding factor/T-cell factor (LEF/TCF) family to regulate unknown genes that control bone formation.52, 53

Interactions between prostate cancer and bone cells

Histopathologic analysis of prostate cancer bone metastases typically shows a large number of osteoblasts adjacent to prostate cancer cells. In contrast, osteoblasts are almost always absent in normal bone or in bone metastases from other cancers (such as breast, lung, and kidney), which largely contain osteoclasts.51 The increase in prostate cancer bone-forming activity gives rise to a woven bone, characterized by an osteosclerotic appearance distinct from the typical lamellar structure seen in normal bone. These lesions are associated with an increase in bone mass at the lesion site, and often have an elevated osteoid surface area, osteoid volume, and mineral apposition rate.54-56

Supporting the observation that prostate cancer bone metastasis is associated with increased osteoblast activity, serum levels of osteoblast proliferation markers, such as bone-specific alkaline phosphatase and type 1 protocollagen C-propeptide, have been observed to be higher in metastatic prostate cancer patients.57 The increased number of osteoblasts often present in the woven bone surrounding the metastatic lesions suggests a corresponding increase in osteoblast proliferation and differentiation. This may contribute to prostate cancer cell survival and invasiveness by providing abundant extracellular matrices.58 In vivo experiments by Gleave et al59 confirmed osteoblast involvement in prostate cancer bone metastasis, showing that factors secreted by bone fibroblasts, but not other cell types, accelerate prostate cancer growth. In vitro, Fizazi et al60 established the stimulation of prostate cancer cells by the culture medium of osteoblast-like cell lines. In these studies, primary mouse osteoblasts, cocultured with MDA prostate cancer 2b cells, led to cancer cell proliferation, hinting at the possibility of paracrine interactions between osteoblasts and the prostate cancer cell line.60 In addition, osteoblast-conditioned medium was found to stimulate prostate cancer cells into producing matrix metalloproteinase-9 (MMP-9) and urokinase-type plasminogen activator (uPA), while at the same time increasing the rate of prostate cancer proliferation.61 Furthermore, newly formed bone may secrete chemoattractants that favor migration and enhance the invasion capability of prostate cancer cells. Osteopontin, osteonectin, and bone sialoprotein can also modulate the properties of prostate cancer cells.62 Osteonectin increases the migration and invasiveness of PC3 and DU145 cells,63 and bone sialoprotein facilitates the attachment of cancer cells to the bone and enhances their metastatic potential.64 These observations indicate that bone-derived factors are able to initiate a series of cellular events that collectively help in cancer cell proliferation. Indirect evidence has also shown that this effect is specific to prostate cancer cells, because conditioned medium from osteoblasts does not enhance the growth of other cancer cell types.65 Clinical trials with specific bone-homing radiopharmaceuticals such as strontium-89, samarium, and rhenium, with deposition at sites of increased osteoblast activity and bone-matrix synthesis, have provided significant palliative benefits.66, 67 Taken together, these observations indicate that osteoblast inhibition has a direct effect on the pathogenesis of osteoblastic lesions.

Receptor Activator of the Nuclear Factor-κB/RANK Ligand/Osteoprotegerin Axis

Osteoblasts also control osteoclast activity through the expression of cytokines such as the receptor activator of the nuclear factor-κB (RANK) ligand (RANKL), a key activator of osteoclast differentiation, and osteoprotegerin, a soluble decoy receptor that inhibits RANKL.68, 69 An increase in the level of serum osteoprotegerin (OPG) has also been established in advanced prostate cancer patients.57, 70 These findings show that osteoblasts, by means of their ability to control prostate cancer cell and osteoclast proliferation, operate as the “master switch” for prostate cancer progression in bone. Previous breast and prostate cancer studies in rodent models have proved that RANKL inhibition decreases bone lesion development and tumor growth in bone.71-73 Furthermore, RANK is expressed in prostate cancer cells and promotes invasion in a RANKL-dependent manner.74

Acquisition of a Bone-Like Phenotype

The acquisition of a bone cell-like phenotype allows prostate cancer cells to home, survive, and proliferate in bone. It has yet to be confirmed whether cancer cells already possess osteomimetic phenotypes when they detach from primitive sites, or whether some of these characteristics are instead acquired in the bone marrow. Prostate cancer cells express several factors involved in normal bone development and remodeling (Fig. 1 and Table 1). These factors have been implicated in osteoblastic lesions, either by affecting osteoblast function directly, or by influencing bone formation indirectly through modification of the bone matrix or microenvironment.

Table 1. Molecules Produced by Prostate Cancer Cells Involved in Bone Metastases
MoleculesEffects on Osteoblasts
DifferentiationProliferationMigration
  1. ET1 indicates endothelin 1; BMP, bone morphogenetic protein; IGF, insulin-like growth factor; FGF, fibroblast growth factor; PDGF, platelet-derived growth factor; PSA, prostate-specific antigen; uPA, urokinase-type plasminogen activator; VEGF, vascular endothelial growth factor; TGF-β, transforming growth factor-β; PTHrP, parathyroid-hormone-related protein.

ET1 
BMPs  
IGF 
FGF 
MDA-BF-1 
PDGF  
PSA 
uPA 
VEGF  
TGF-β  
Wnt  
PTHrP  

BMPs

BMPs belong to the TGFβ superfamily, and are the most powerful inductive bone factors enriching the bone matrix.75 Primary and metastatic prostate cancer have different phenotypic patterns of BMP expression, and adopt discrete downstream signaling pathways.76 BMPs demonstrate a clear and close relationship with the development and progression of primary prostate cancer, and also contribute to the onset and development of bone metastases. As a result, it has been postulated that BMPs play an important role in the etiology of the osteoblastic bone metastasis phenotype. Indeed, BMP-secreting prostate cancer cell cultures have been shown to promote in vitro mineralization (potency from high to low: BMP6 > BMP7 > BMP4).77 Furthermore, several BMPs are known angiogenic factors that indirectly facilitate the development of bone metastases via the angiogenic route. On 1 hand, BMPs produced by prostate cancer cells are able to activate osteoblasts, leading to lesions that are predominantly osteoblastic in nature. On the other hand, BMPs synthesized by osteoblasts, or released from the bone matrix, subsequently enhance the growth and aggressiveness of prostate cancer cells, which in turn increases the production of BMPs by tumor cells.

Growth Factors

The IGF system, comprising 3 receptors, 3 ligands, and 6 IGF-binding proteins (IGFBPs), is involved in the mitogenic, transformational, and antiapoptotic activities of several cell types. In the prostate, this system plays an important role in both normal and tumoral cells. IGF-I is secreted by the prostatic epithelium, probably in response to growth hormones, and even in hypertrophy, high levels of IGF-I have been reported.78 Prostate cancer cell lines have been shown to express a range of different IGF components.79 IGF-I and II are abundant in the bone environment, and are known to stimulate osteoblasts, increasing bone matrix apposition, while at the same time decreasing collagen degradation.80 Several studies have shown a possible relationship between IGFs and bone metastasis from prostate cancer.81, 82 IGF-I and II increase the proliferative and chemotaxis activities of prostate cancer cells in vitro.81 Moreover, the IGF-I pathway is up-regulated in prostate cancer cells metastasized to bone.82 Several in vivo studies have added further weight to this important link. Serum IGF-I levels are correlated with the risk of developing prostate cancer,83 plasma IGFBP-III is lower in patients with bone metastases, and IGFBP-II is elevated in prostate cancer patients.84 In addition, high IGF-I and low IGFBP-III levels are associated with the risk of developing advanced prostate cancer.85 Furthermore, polymorphisms of IGF-I and cytochrome P450 enzyme 19 have been reported to accurately predict bone metastasis in prostate cancer patients.86 However, Rubin et al87 recently showed that IGF-I is neither necessary nor sufficient for the osteoblastic response to prostate cancer metastases. For this reason, it is still not clear how the IGF system participates in the formation of osteoblastic prostate cancer bone metastasis, and additional studies are required to delineate the role of the IGF system in this process.

Parathyroid Hormone-Related Protein and Endothelin 1

The osteolytic factor parathyroid hormone (PTH)-related protein (PTHrP) is a homolog of PTH that has a direct action on PTH receptors, increasing bone resorption and renal tubular calcium reabsorption.88 In bone metastases, the release of PTHrP by cancer cells, together with other factors, contributes significantly to metastatic spread.89, 90 PTHrP is abundant in prostate cancer bone metastases, and in these tumors osteoblastic lesions tend to predominate.91 One possible explanation for this paradox is that prostate cancer-derived PTHrP mediates the interactions between the bone marrow microenvironment and prostate cancer, which further facilitates the establishment of skeletal metastases and osteoblastic alterations. Liao et al92 have provided evidence supporting this opinion, reporting that PTHrP increases osteoblastogenesis-stimulating osteoblast progenitor cell proliferation and induces early osteoblast differentiation.

Another theory is that NH2-terminal fragments of PTHrP share a strong sequence homology with endothelin (ET) 1, thus stimulating new bone formation through activation of the ETA receptor.93 As an osteoblast mitogenic factor, ET1 promotes osteoblast growth at metastatic sites.94 Nelson et al95 showed that plasma ET1 concentrations are significantly higher in metastatic prostate cancer patients, suggesting that ET1 may be secreted by prostate cancer cells. A recent study has also suggested that ET1 increases osteoblast proliferation and new bone formation through activation of the Wnt signaling pathway and suppression of the Wnt pathway inhibitor, DKK1.96

uPA

Prostate cancer cells produce other factors, including uPA and prostate-specific antigen (PSA) that, although not directly involved in normal skeletal development, affect osteoblast function by modifying the bone microenvironment.97 The serine protease uPA is involved in degradation of the extracellular matrix, which facilitates tumor cell invasion by increasing osteoblastic prostate cancer bone metastases. The pathogenetic model of osteoblastic prostate cancer metastases can be summarized as follows. The neoplastic cell produces high molecular weight uPA (HMW-uPA) that binds uPA receptor (uPAR) on the tumor cell surface and exerts its proteolytic action, leading to tumoral invasion. HMW-uPA is then split into low molecular weight uPA and amino terminal fragment, which binds uPAR on the osteoblast cell surface, helping to trigger the cascade of events leading to osteoblast proliferation and activation. Furthermore, uPA can cleave and active TGFβ, which is produced in a latent form by osteoblasts. TGFβ regulates osteoblast and osteoclast differentiation as well as the growth of tumor cells themselves. In addition, uPA stimulates osteoblast proliferation, probably by hydrolyzing IGFBPs, thereby increasing free IGF levels.98-100 PSA belongs to the kallikrein serine protease family, is secreted by prostate cancer cells, and is used routinely as a marker of prostate cancer progression. PSA not only cleaves PTHrP to release osteoblastic PTHrP fragments, but also activates osteoblast growth factors such as TGFβ.100 Consequently, PTHrP degradation by PSA decreases bone resorption,101 allowing the osteoblast response to predominate. As with uPA, PSA can also cleave IGFBP3, thereby freeing up IGF-I to bind to its receptor and stimulate osteoblast proliferation.102

Other Molecules

The growth factor MDA-BF-1 has recently been identified in the supernatant of bone marrow specimens from patients with prostate cancer bone metastases.103 This secreted form of the ErbB3 growth factor receptor is expressed in prostate cancer cells that metastasize to bone, but not in the primary tumors of patients with localized disease, nor in prostate cancer cells that metastasize to other sites. Its function is mediated by an osteoblast-expressed receptor,104 and through specific interactions between prostate cancer cells and bone. Scientific literature is rich in papers describing new proteins that seem to be involved in prostate cancer bone metastases. One of these is the tyrosine kinase receptor c-Kit, and its ligand stem cell factor (SCF).105 In an experimental model, prostate cancer bone metastasis was found to strongly express c-Kit, but prostate cancer cell lines were c-Kit negative. In addition, immunohistochemical analysis showed a higher expression of this protein in bone metastasis compared with primary tumors, and SCF was often overexpressed.105

Prostate cancer cells produce a range of additional osteoblast-regulatory growth factors, including PDGF, FGF, and VEGF. The dimeric polypeptide growth factor PDGF has 2 subunits, A and B, that form AA, BB, and AB isoforms. The BB isoform is known to be a potent osteotropic factor, contributing to the formation of osteoblastic lesions through the promotion of osteoblast migration and proliferation.106, 107 Although FGF1 (acidic) and FGF2 (basic) are both able to increase osteoblast proliferation, only FGF2 suppresses osteoclast formation.108 It is clear that further investigations are necessary to fully understand the complex FGF interplay in bone metastasis. VEGF appears to have both a direct and an indirect effect on bone growth by activating osteoblasts and promoting angiogenesis, respectively.77, 109 The Wnt family ligand has been reported to be up-regulated in the tumor cells of advanced metastatic prostate cancer patients.110 Furthermore, Wnt production can act in a paracrine way to induce osteoblastic bone metastases. As mentioned previously, the Wnt signaling pathway can be blocked by the Wnt antagonist, DKK1. This molecule is usually expressed early in the development of skeletal metastases, resulting in the masking of osteogenic Wnts, which favors metastatic site osteolysis. As metastasis progresses, a decrease in DKK1 expression unmasks Wnt-mediated osteoblastic activity, leading to osteosclerosis.111, 112

The Role of Osteoclasts

Although prostate cancer bone metastasis is osteoblastic in nature, as revealed by radiographic imaging, the presence of osteolytic-osteogenic bone lesions in osteoblastic cases may account for the increase in observed fractures.113 Key steps in the successful establishment of prostate cancer bone metastases appear to be osteoclast formation and bone resorption, followed by the release of several growth factors from the bone matrix.114 However, the mechanisms by which prostate cancer cells promote this process remain unclear. It has been suggested that RANKL, interleukin (IL) 6, IL-8, and C-C chemokine ligand 2 mediate osteoclast formation from human mononuclear bone marrow cells.31, 32, 99 Furthermore, prostate cancer cells have been shown to secrete factors that promote human bone marrow mononuclear cell osteoclastogenesis.31 Other authors have observed that IL-6 affects fusion but not resorption, and that osteoclastogenesis in prostate cancer bone metastasis is induced by a RANKL-independent mechanism. A possible hypothesis put forward is that prostate cancer cells, by producing PTHrP, can facilitate both osteoclastogenesis and osteoblastogenesis.115 Another supposition is focused on C-C chemokine ligand 2, which may serve 2 distinct roles in the development and promotion of prostate cancer: a direct effect on epithelial prostate cancer cells, and an indirect effect on osteoclasts and endothelial cells at the metastatic site, thereby supporting tumor growth. The presence of C-C chemokine ligand 2 in the bone microenvironment may contribute to prostate cancer growth by initiating osteoclastogenesis and angiogenesis, creating a favorable niche for metastases.116

The Vicious Cycle of Osteoblastic Bone Metastasis

The complex interactions between tumor cells, bone cells, and the bone matrix constitute a vicious cycle of osteoblast-mediated bone metastasis (Fig. 2). In the early stages, prostate cancer cells produce osteogenic factors such as PDGF, ET1 and BMPs, which activate osteoblasts. Once differentiated from their progenitor cells, osteoblasts deposit a new matrix for bone formation. However, this unmineralized matrix provides a more fertile soil for tumor cells, enriched with growth factors and noncollagen proteins. Moreover, newly formed bone may provide additional factors to attract prostate cancer cells, allowing them to survive and proliferate in the bone environment, which in turn further activates osteoblasts. However, the osteoblastic nature of bone metastases may represent a double-edged sword regarding the development of lesions; the initial increase in bone volume may limit the space available to cancer cells, and therefore help to confine the tumor. This may explain why osteoblastic prostate cancer metastases seem to progress more slowly than osteolytic metastases from other tumors. In addition, especially at the beginning of the process, tumor-derived factors and RANKL-secreting osteoblasts can both activate osteoclasts, leading to some level of bone resorption, which subsequently creates more space for dominant osteoblastic lesions. Cytokines released from the bone matrix during bone resorption can enhance this vicious cycle by facilitating the continued proliferation of prostate cancer cells and osteoblasts. Regrettably, as soon as the metastatic disease enters the osteoclastic stage, disease progression is rapid, and survival times are short.54, 117, 118

Figure 2.

The vicious cycle of osteoblastic bone metastasis is shown. (A) Prostate cancer cells secrete osteogenic growth factors, activating osteoblasts to deposit new bone matrix. (B) Osteoblasts secrete a range of additional factors such as insulin-like growth factor (IGF), fibroblast growth factor (FGF), and transforming growth factor-β (TGFβ). (C) These factors attract prostate cancer cells, further enhancing their proliferation and growth. ET1 indicates endothelin 1; PDGF, platelet-derived growth factor; BMP, bone morphogenetic protein; RANK, receptor activator of the nuclear factor-κB; RANKL, RANK ligand; OPG, osteoprotegerin.

Clinical Applications

A deeper understanding of the molecular mechanisms of prostate cancer bone metastases will hopefully identify new biological targets for innovative drugs, besides zoledronic acid, which has become the standard treatment for this disease, to be used in combination with conventional therapies, as well as new predictive or prognostic markers. One of the most extensively studied pathways in recent years, the RANK/RANKL/OPG axis, has been advocated as a potential therapeutic target because it governs bone homeostasis, both under normal physiologic conditions, and during bone metastasis progression. In preclinical and clinical trials, inhibition of RANKL with OPG reduces bone turnover markers,119 suggesting that RANKL, in addition to having an effect on bone lysis, also has a direct action on tumor cells.120, 121 The anti-RANKL monoclonal antibody, denosumab, has recently been tested against zoledronic acid in a randomized prostate cancer bone metastases phase 2 study,122 and was found to decrease urinary bone turnover markers in zoledronic acid resistant patients. Phase 3 bone metastases studies are ongoing with skeletal-related events (SREs) as the primary objective, and results should be available shortly.

Atrasentan, an ETA receptor antagonist, is being tested in clinical trials for the treatment of prostate cancer, and it has been shown to prevent osteoblastic bone metastases in mouse models, as well as reduce disease progression in advanced phase 2 prostate cancer trials.123, 124 However, a recent phase 3 study of atrasentan versus placebo reported no significant differences in disease progression,125 and docetaxel/atrasentan prostate cancer studies are still ongoing. The monoclonal antibody vitaxin binds integrin αvβ3, and androgen-independent prostate cancer clinical studies are currently in progress.126 This molecule has demonstrated both in vivo and in vitro antitumor activities, affecting bone resorption by impairing osteoclast attachment, whereas osteoclast formation and multinucleation processes remain unaffected.

Other targets of bone metastases include Akt, cyclooxygenase-2, and MMP-9. Diaz et al127 recently investigated the Akt inhibitor Palomid 529 in vitro, and found that it had a synergistic effect with radiotherapy, enhancing the antiproliferative effect of radiation in prostate cancer cells, while decreasing p-Akt, VEGF, MMP-9, MMP-2, and Id-1 levels. Another drug target of MMP-9, 3,3′-diindolylmethane,128, 129 is an inhibitor of the DNA binding activity of nuclear factor-κB, which mediates the expression of many genes such as VEGF, IL8, and uPA. Antibodies against IGF-I receptor, and IGF-I receptor-specific tyrosine kinase inhibitors, have been developed as anticancer agents, and some have already been tested in clinical trials against solid tumors, showing promising results.130, 131 These are just a few of the vast array of new drugs currently under preclinical and clinical experimentation.

Regarding the monitoring of metastatic prostate cancer, the most commonly cited markers seem to be the products of collagen degradation.132-135 This category includes N-telopeptide (NTx) or C-telopeptide type I collagen (CTX), amino-terminal procollagen propeptides (PINP), and cross-linked C-terminal telopeptides (ICTP). Other common markers of osteoblastic bone metastases include total and bone-specific alkaline phosphatase, serum tartrate-resistant acid phosphatase, and PSA.132, 133 In a recent study of 100 patients, Klepzig et al134 observed that procollagen propeptides could reliably predict prostate cancer bone metastases. Another important study by Lein et al133 has focused further attention toward procollagen propeptides, NTx, and PSA, establishing bone markers as useful tools for the prediction and diagnosis of SREs in patients with prostate cancer bone metastases undergoing zoledronic acid therapy. NTx has recently undergone validation in 3 large retrospective phase 3 trials.135 This marker was found to decline to normal physiologic concentrations within 3 months, and, compared with high NTx levels, was associated with a reduced risk of skeletal complications and death. In contrast, PSA has been reported to be a poor marker, both for identifying patient populations at high risk of metastatic disease and for monitoring skeletal progression during treatment.132 A further study by Jung et al57 tested 10 serum markers in 117 prostate cancer patients, citing OPG and bone sialoprotein as excellent bone metastases markers in addition to being independent prognostic factors for prostate cancer-related death.

Conclusions

Cancer cells can only metastasize in organs where the microenvironment accommodates their growth. In other words, the complex interactions between cancer cells and the bone/bone marrow microenvironment are fundamental to the establishment of bone metastases. Prostate cancer cells from primary tissue undergo epithelial-mesenchymal transition to disseminate and acquire a bone-like phenotype to metastasize in bone tissue. After the arrival of prostate cancer cells in bone, their crosstalk with the bone microenvironment facilitates bone metastases, and defines their osteoblastic pattern. Finally, recent discoveries on the mechanisms of bone metastasization, for example epithelial-mesenchymal transition and osteomimicry, will certainly help to improve bone target therapies. Without doubt, there is an urgent need for further work, particularly on the potential role of the premetastatic bone niche, and on the involvement of cancer stem cells in bone metastases.

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

Ancillary