Bad to the Bone: The Role of the Insulin-Like Growth Factor Axis in Osseous Metastasis

Bone metastases are a frequent complication of cancer, and the skeleton is the site of the most significant tumor burden in many patients with advanced disease (1). Prognosis after the development of bone metastases varies by tumor type, but in most cases, bone metastases lead to considerable morbidity, with devastating consequences including intractable bone pain, pathologic fractures, hypercalcemia, and spinal cord compression (2). Existing treatments for bone metastases are not curative, but may slow the progression and provide symptom palliation. Consequently, there is an unmet need for new therapeutic interventions (2).

Over the last few decades, extensive research has identified some of the molecular mechanisms that promote bone metastases, including bone-derived growth factors such as the insulin-like growth factors (IGF; ref. 3). Research suggests that IGFs play a fundamental role in bone development, remodeling, and repair, and contribute to key hallmarks of primary cancers, via alteration of stem-cell renewal/differentiation, epithelial–mesenchymal transition (EMT), and treatment resistance (4). In this review, we discuss the clinical significance of bone metastases, review the normal physiologic role of IGFs in bone development and homeostasis, and consider the potential role of the IGF axis in the metastasis of primary tumors to bone. Finally, we evaluate the available evidence suggesting that targeting this signaling network may lead to potential therapeutic strategies for the prevention and/or treatment of bone metastases.

Bone metastases are a common feature of solid tumors, and in one U.S. study, were found to affect almost 7% of patients overall; the highest incidence of bone metastases was in prostate cancer (18%–29%), followed by lung (up to 13%), renal (up to 10%), and breast cancers (up to 8%; ref. 5). In some tumor types such as hepatocellular carcinoma, the incidence of bone metastases is increasing, likely due to improved tumor control at other disease sites (6, 7). Bone metastases can cause severe bone pain, which can be treated with palliative radiotherapy but is often difficult to control (8). Further, by promoting bone resorption, bone metastases can lead to bone fragility and skeletal-related events (SRE), including pathologic fracture, spinal cord compression, and hypercalcemia (1, 9, 10). SREs are typically associated with reduced quality of life and substantially higher healthcare costs (11). Tumors that have metastasized to bone are generally incurable (11); survival varies with primary cancer type and is lower if patients have skeletal complications (11). Indeed, in a large population-based cohort study, the median overall survival (OS) for patients with breast cancer with bone metastases was 16 months but only 7 months for patients with bone metastases and an SRE (11).

The main goals of treatment are to palliate bone pain, alleviate and prevent bone complications, maintain quality of life, and slow metastatic progression (12, 13). Symptom palliation can be achieved using analgesia, endocrine therapy, chemotherapy, localized radiation, and surgery. However, the only agents specifically designed to treat bone metastases are bone-targeting agents, including the bisphosphonates zoledronate and pamidronate; the targeted alpha therapy radium-223 dichloride; and the receptor activator of NF-κB ligand (RANKL)–inhibiting antibody denosumab (9, 12). Although radium-223 has been shown to prolong survival and delay time to first symptomatic skeletal event in metastatic prostate cancer, a recent phase III trial was prematurely unblinded because of an increase of fractures and deaths in patients treated with radium-223 in combination with abiraterone and prednisone (14). Bisphosphonates reduce skeletal morbidity in advanced breast cancer and can also decrease the number of bone metastases (15–17). However, 30% to 50% of patients treated with bone-targeting agents develop new bone metastases, skeletal complications, and disease progression, highlighting the need for more effective therapies (6).

The IGF system is critical for skeletal growth and maintenance (18). All skeletal cells express IGF-1 and its receptor, IGF-1R, and require IGF-1 for normal development and function (19). IGF-1 and IGF-2 play an important role in regulating growth and development of normal human tissues by promoting cellular proliferation and differentiation via activation of IGF-1R and IGF-1R/insulin receptor (INSR) hybrids (4, 20). In addition, IGF-2 binds to the A isoform of the INSR (INSR-A), a fetal variant of the classical metabolic INSR (INSR-B), which is frequently overexpressed in tumors (4).

Bone development and remodeling are driven by the coordinated activity of osteoclasts and osteoblasts, which regulate bone resorption and formation, respectively (21, 22). Osteoblast proliferation is promoted by IGFs, transforming growth factor β (TGFβ), fibroblast growth factors (FGF), and platelet-derived growth factor (PDGF), and osteoclast apoptosis by TGFβ (1). Of these, IGF-1 and IGF-2 are the most abundant growth factors in bone (18, 23) and have been shown to directly influence both osteoclasts and osteoblasts (24). In particular, the IGF signaling pathway is thought to play a critical role in regulating osteoblast function (25). IGF-1 also helps maintain the normal interaction between osteoblasts and osteoclasts to support osteoclastogenesis (26), and in vitro data suggest that IGF-1 stimulates osteoclastic bone resorption by supporting the generation, differentiation, and activation of osteoclasts (26–29). IGF-2 also promotes differentiation and function of primary human osteoblasts and osteoclasts but is less potent than IGF-1 (27, 28).

The liver is the major source of circulating IGF-1, producing approximately 75% of plasma IGF-1 (30). IGF-1 is also produced locally by bone cells such as osteoblasts, chondrocytes, osteoclasts, and osteocytes, thereby functioning as an autocrine/paracrine effector to regulate bone turnover (19, 21, 24, 30, 31). In addition, an inactive form of IGF-1 is stored within the bone matrix, and is activated and released during osteoclast-mediated bone resorption (18). Data from human biopsies suggest that bone volume is positively associated with the IGF-1 content of bone matrix, but not IGF-2 content or serum IGF-1 (32). Notably, IGF-2 is nine times more abundant in adult human bone than IGF-1, unlike in mice, where IGF-1 content is higher postnatally (23).

Bones undergo constant and dynamic cell growth and turnover, and so provide a permissive environment for proliferation of tumor cells (1). A growth factor–rich environment contributes to the tropism of primary tumor cells for bone or bone marrow (22). For bone metastasis to occur, circulating cancer cells must arrest in the bone marrow cavity before migrating into the marrow stroma where they generate their own blood supply (1). Consistent with the seed and soil hypothesis proposed by Paget in 1889 (33), primary tumor cells (seeds) prime the bone microenvironment (soil) for metastasis by secreting factors that stimulate osteoblast activity and bone formation, thus enriching the bone microenvironment with osteoblast-derived growth factors that support the local growth of tumor cells (34). Consequently, both tumor cells and bone cells produce factors that promote malignant growth (13).

Interaction between tumor cells and the bone microenvironment occurs via sequential stages: tumor-cell homing, dormancy, colonization, and expansion (6). Tumor cells preferentially home and adhere to bone marrow endothelium, enabling their invasion into the bone marrow microenvironment via extravasation. Metastatic tumor cells then compete for occupancy of the hematopoietic stem cell (HSC) niche in bone marrow. Tumor cells may subsequently remain dormant in the HSC niche, or begin growth and colonization. Metastatic colonization may be delayed for years, resulting in relapse many years after diagnosis of the primary tumor (11, 13, 35, 36). Upon recovery from dormancy, tumor cells colonize bone and induce differentiation, recruitment, and activation of osteoclasts and/or osteoblasts (6, 37). Secretion of proosteoclastogenic factors such as TNFα, interleukin-8, and parathyroid hormone (PTH) leads to osteoclast-mediated bone resorption and further release of bone matrix–stored growth factors including IGFs, TGFβ, FGFs, and bone morphogenetic proteins (BMP). This sets up a vicious cycle of bone colonization and metastasis, known as the expansion stage (6, 22, 36). Such osteolytic bone metastases are frequently associated with breast and lung tumors (36). Conversely, the release of bone morphogenetic proteins such as TGFβ and FGF, in addition to the IGFs, enhances osteoblast differentiation and activity; the resulting osteoblastic bone metastases are commonly seen in prostate cancer (36, 38). Regardless of the mechanism, this reciprocal cross-talk between tumor and bone cells creates an environment ripe for metastatic growth (6, 37).

Although not all the key players in the development of bone metastases are yet known, multiple molecules and signaling pathways appear to be important for metastasis. Indeed, in addition to the IGF pathway, NF-κB, matrix metalloproteases, and the Wnt, stromal cell–derived factor 1 (CXCL12)/C-X-C chemokine receptor type 4 (CXCR4) and PI3K/AKT signaling axes may also play key roles in the development and progression of bone metastases (39–46).

Altered IGF axis signaling is thought to contribute to the pathogenesis of primary bone tumors including Ewing sarcoma (47) and osteosarcoma, with evidence of IGF axis activation due to downregulation of IGF-binding proteins observed in osteosarcomas as compared with osteoblasts and mesenchymal stem cells (48). Another osteosarcoma study identified recurrent mutations of IGF axis genes that were predicted to result in the activation of IGF-1R signaling in 7%, and IGF1R gene amplification in 14% of tumors (49).

The IGFs can also encourage metastasis by promoting anchorage-independent growth, motility, and invasion (50, 51). In particular, IGF-1R signaling via INSR substrate 2 (IRS-2) has been implicated in mediating invasion and metastasis versus proliferation (51, 52). Interestingly, IGF-1R inhibition via dominant negative IGF-1R, or an IGF-1R–blocking antibody inhibits metastasis without affecting primary tumor growth (53, 54).

IGFs have been implicated in each of the key stages of bone metastasis: homing, dormancy, colonization, and expansion (6). In triple-negative breast cancer, stromal cancer–associated fibroblasts were identified as the source of IGF-1 and CXCL12, which were shown to prime cells to home to the CXCL12- and IGF-1–rich bone microenvironment, in a process dependent on CXCR4 and IGF-1R expression by cancer cells (55). IGF-1 also enhances homing of myeloma cells to bone marrow, and acts as a chemo-attractant that increases adherence to the extracellular matrix via activation of beta integrin and PI3K/AKT (56). IGF-1R upregulation was shown to promote adherence of neuroblastoma cells to human bone marrow endothelial cells, and transendothelial migration was decreased by IGF-1R inhibition (57).

IGF-1 and IGF-2 have been shown to mediate tumor cell dormancy in bone in models of pancreatic cancer and osteosarcoma (58, 59). In a murine pancreatic cancer model, the activation of IGF-1/AKT signaling was a common survival mechanism in dormant cancer cells that had survived ablation of the oncogenic drivers KRAS and MYC (58). In clinical osteosarcomas and a mouse model of MYC-driven osteosarcoma, IGF-2 expression increased following chemotherapy, and prolonged IGF-2 exposure induced a dormancy-like state, characterized by attenuated responsiveness of the IGF axis to IGF-2, with low levels of AKT–mTOR activity, cell-cycle arrest, and autophagy (59).

Finally, IGF-1 and -2 appear to play important roles in bone colonization and expansion by metastasizing tumor cells. In one study, bone-derived IGFs stimulated metastasis of breast cancer to bone by increasing cancer cell proliferation and survival, via AKT activation and recruitment of NF-κB (60). Further, culture medium from cells stimulated to undergo bone resorption was found to contain high concentrations of IGF-1; notably, the anchorage-independent growth of human breast cancer cells cultured in this medium was inhibited by the IGF-1R–neutralizing antibody αIR3, but not by antibodies against TGFβ, FGF-1 or -2, or PDGF-BB (60). Reflecting cross-talk between the IGF axis and hypoxia signaling, IGF-1 released during RANKL-induced bone resorption was shown to cooperate with tumor cell–derived hypoxia-inducible factors (HIF) in the hypoxic bone microenvironment, to promote bone marrow colonization and establishment of osteoblastic bone metastases in a murine osteosarcoma model (61). Similarly, growth of human breast cancer cells in a human adult bone model was facilitated by active osteoclasts induced by RANKL, and IGFs released following bone resorption (62). The IGF axis may also be involved in bone invasion in oral squamous cell carcinoma, with the depletion of IGF-2 mRNA-binding protein-3 inhibiting regional bone destruction in a xenograft mouse model (63).

Given the apparent involvement of the IGF axis in the generation and growth of bone metastases, targeting key molecules involved in this pathway represents a rational approach to help prevent and/or manage bone metastases. Indeed, considerable preclinical evidence supports this hypothesis, with most available evidence coming from models of metastatic prostate and breast cancer.

Several preclinical studies have investigated the effects of antibodies or tyrosine kinase inhibitors (TKI) directed at IGF-1R, the results of which suggest that IGF axis inhibition can at least partially prevent the establishment and progression of bone tumors (Table 1). Furthermore, whole-genome expression analysis of bone metastasis biopsies from patients with prostate cancer suggested an inverse relationship between IGF-1R expression and immune cell function, (64) suggesting that IGF-1R inhibition has the potential to stimulate an antitumor response from endogenous immune cells, a concept that is also supported by models of other tumor types (65–68).

Therapeutic targeting of the ligands IGF-1 and IGF-2 has been investigated in preclinical models of prostate and breast cancer metastasis (Table 1). KM1468, an anti–IGF-1/2 antibody, suppressed metastatic development and progression of prostate cancer cells into implanted human adult bone (69). Using the same model, the specific anti–IGF-2 antibody m610 significantly reduced the growth of prostate cancer cells after injection into human bone implanted in the mammary fat pad, but not when the same cells were injected without human bone. These data support the importance of IGF-2 in promoting the progression of bone metastases, and reinforce the relevance of paracrine IGF signaling in bone metastases (23). Similarly, KM1468 reduced the growth of human breast cancer cells in human adult bone implanted into mice but had no effect on growth of the same cells subcutaneously injected in the absence of bone (62). These data are consistent with reports that IGF-1R inhibition can inhibit metastases without affecting the primary tumor (53, 54). Likewise, the transfection of human breast cancer cells with dominant-negative IGF-1R markedly inhibited the development of osteolytic bone metastases, whereas lung metastases were largely unaffected (60). Such findings suggest that IGF inhibition may have more impact in the bone microenvironment (where IGFs play an important role in the normal homeostatic functioning of the tissue) than in the primary tumor.

It should be noted that although the findings from these studies are encouraging, the IGF system of mice is substantially different from that in humans; thus, in contrast to mouse bone cells, human bone cells produce less IGF-1 than IGF-2 (23, 70). Thus, the roles of the IGFs and the effects of their inhibition may also differ between species. Indeed, because IGF-2 also induces proliferative/antiapoptotic signaling via INSR-A (4), the contribution of IGF-1R–mediated signaling may be less important in humans than in mice.

To date, no clinical trials have specifically examined the impact of IGF inhibition on the development and/or progression of bone metastases. Previous studies of monoclonal antibodies and TKIs targeting IGF-1R in patients with a range of solid tumors have yielded mixed, mostly negative, results, both as monotherapy and when combined with established therapies (see Supplementary Table S1 for details). Objective responses have been low and mainly limited to studies in osteosarcoma, Ewing sarcoma, and other soft tissue sarcomas (71–75). The failure of IGF-1R–inhibitory drugs to impact tumor growth in unselected patients may be attributable to compensatory activation of other signaling pathways, including those involving the epidermal growth factor receptor (EGFR; ref. 76) and Wnt (77), and by the lack of a validated biomarker for patient selection (4). Further, IGF-1R inhibitors are commonly associated with hyperglycemia, due to coinhibition of INSR-B (71, 72, 78–85), which may preclude their use at therapeutic doses; consequently, most agents targeting IGF-1R are no longer in clinical development.

Despite these limitations, there remains considerable interest in the therapeutic targeting of the IGF axis, particularly in view of the recent development of IGF ligand–neutralizing antibodies. The issue of dose-limiting hyperglycemia may have been overcome by the development of such antibodies, which do not interfere with glucose metabolism (86, 87). These agents inhibit proliferative/antiapoptotic signaling via both IGF-1R and INSR-A, without interference with the metabolic INSR-B isoform (4). Given the disappointing clinical results with the IGF-1R–targeted agents, it is possible that IGF-1R is activated and critically involved in oncogenic signaling pathways in only a minority of patients (88). However, the evidence summarized herein supports a clear role of the IGF axis in bone metastasis development, which suggests that the presence of bone metastases could potentially indicate involvement of an IGF-dependent process.

Bones are the sites of dynamic cell turnover that provide a supportive environment for the establishment and growth of tumor metastases (1). Bone metastases are a frequent complication in patients with advanced cancer and a significant cause of morbidity (5, 8). The IGF axis plays an important role in bone development and remodeling (18), and is implicated in metastasis of primary tumors to bone. Preclinical data demonstrate that inhibition of IGF axis signaling can suppress metastasis and tumor cell proliferation in bone (23, 60, 62, 64). Most evidence related to the value of targeting IGF comes from the models of breast and prostate cancer bone metastases, probably because these cancers are commonly associated with metastatic spread to the bone (5). Preliminary evidence suggests that IGF inhibition may also be relevant to the treatment of bone metastases from other primary tumors, such as oral squamous cell carcinoma (63). As yet, there is no clinical evidence to support the use of IGF-targeted therapies as specific treatments for bone metastasis, and there are differences in IGF production between human and rodent bone cells (70) that may impact clinical relevance. Nevertheless, given the known involvement of the IGF axis in bone metastasis (6, 60, 61), and initial preclinical data supporting the utility of IGF blockade in this context (23, 60, 62, 64), we suggest that exploration of the potential benefits of IGF inhibition may be warranted in patients with bone metastases, particularly for breast and prostate cancers.

V.M. Macaulay is a consultant/advisory board member for Boehringer Ingelheim. A.V. Lee is Senior Advisor, Translational Sciences at UPMC Enterprises. T. Bogenrieder holds ownership interest (including patents) in Roche, Seattle Genetics, Celgene, Gilead, and Immunogen. No potential conflicts of interest were disclosed by the other authors.

The authors were fully responsible for all content and editorial decisions, were involved at all stages of manuscript development, and have approved the final version for submission.

Medical writing assistance, supported financially by Boehringer Ingelheim, was provided by Katharine Williams, of GeoMed, an Ashfield company, part of UDG Healthcare plc, during the preparation of this manuscript.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.