Bone Sarcomas in Pediatrics: Progress in Our Understanding of Tumor Biology and Implications for Therapy

IGF-1R mediates cell differentiation, proliferation, and apoptosis in human cancer by activating two major oncogenic signaling cascades: the phosphoinositide 3-kinase (PI3K) pathway and the mitogen-activated protein kinase (MAPK) pathway [23, 24]. Elevated expression of IGF-1R has been observed in most OS and ES cell lines and tumor samples [25‐27]. Overexpression of IGF-1R and its ligand IGF-1 in pediatric bone sarcomas is correlated with a poorer prognosis, and IGF pathway inhibition impeded tumor growth and metastasis in preclinical models [28, 29].

Current therapeutic approaches directed against the IGF-1R pathway can be grouped into three categories: monoclonal antibodies targeting IGF-1R, IGF ligand-neutralizing antibodies, and small-molecule tyrosine kinase inhibitors. At present, eight different anti-IGF-1R monoclonal antibodies (mAbs) have been or are currently being evaluated in phase I/II clinical trials, and one of which is being evaluated in pediatric patients (Table 1). Although their safety has been proven in pediatric patients, these IGF-1R antagonists displayed limited or no clinical benefit as monotherapy for patients with advanced bone sarcomas [30‐37]. Investigations using IGF-1R-targeted agents in combination are ongoing, though it is not clear what benefit these studies will show. Further, since these agents showed no benefit in common adult malignancies, their development has been abandoned by the pharmaceutical industry, suggesting that they are unlikely to be available for future bone sarcoma patients.

An alternative approach to inhibit IGF signaling is to neutralize the bioactivity of IGF ligands IGF-I and -II with mAbs. In preclinical studies, these agents achieved more effective inhibition of IGF signaling than IGF-1R mAbs by blocking binding of IGF-I and -II ligands to IGF-1R and insulin receptor (IR)-A [38, 39]. Two neutralizing antibodies against IGF-I/II are available: MEDI-573 and BI836845. A phase I clinical trial for MEDI-573 in adult patients with advanced solid tumors demonstrated stable disease in 13 of 39 patients [40]. Currently, five phase I clinical trials of BI836845 are ongoing in adult patients with various solid tumors, but there have been no specific studies in OS or ES patients. Since IGF-II and IGF-2R can also be overexpressed in OS and ES, these patients might benefit from therapies that target both IGF ligands [27, 41].

In addition to mAbs, small-molecule inhibitors of IGF-1R are also being developed. Some of these agents also inhibit IR-A-dependent tumor growth [42]. Novel IGF-1R tyrosine kinase inhibitors include linsitinib, XL-228, INSM-18, GSK1904529A, GSK1838705A, and BMS-554417, all of which have shown promising results in pediatric sarcoma models during preclinical studies [22]. As yet, no pediatric clinical trial data have been reported for these agents.

HER2 is one of the four RTKs in the epidermal growth factor receptor (HER/EGFR/ERBB) family and has an essential role in tumor growth. In recent years, targeted therapies against HER2 have achieved significant therapeutic benefits in the treatment of several solid tumors. However, data have been conflicting regarding expression of HER2 in OS and ES and its association with clinical outcome. While some reports demonstrated minimal expression of HER2 in tumor samples of pediatric bone sarcomas or lack of correlation between HER2 expression level and patient outcome [43‐48], other studies have suggested that HER2 is highly expressed in up to 40 % of OS cases and 20 % of ES cases and its overexpression is correlated with metastases and poor prognosis [46, 49‐55]. One possible reason for these disparate results may be purely technical: the Her-2 antigen is susceptible to oxidative degradation, such that it is essentially undetectable 6 months after slides are cut [56, 57]. The safety of the HER2 mAb trastuzumab in combination with standard chemotherapy has been shown in a phase II clinical trial for OS, but no benefit was seen [58]. There is no clear evidence of therapeutic benefit for this agent in bone sarcoma, and no basis for treating bone sarcoma patients with it except in the context of a clinical trial. In addition to the antibody approach, small-molecule tyrosine kinase inhibitors of the ERBB family such as erlotinib, lapatinib, afatinib, neratinib and dacomitinib are currently in clinical development [59]. Since HER4, another member of the HER family, has emerged in recent years as an essential regulator in OS, ES, and other pediatric solid tumors [60‐62], the pan-Her small-molecule inhibitors (afatinib, dacomitinib, and neratinib) may represent a more effective approach in treating pediatric bone sarcomas than EGFR-specific small-molecule inhibitors such as erlotinib [63, 64].

The PDGF family of signaling molecules consists of five ligands (PDGF-AA, -BB, -AB, -CC, -DD), and two RTKs (PDGFR-α and -β) [65]. In OS and ES, PDGF/PDGFR signaling has a central role in tumor growth and metastasis, and overexpression of PDGFR-α and -β is often correlated with poor prognosis [66‐69]. Imatinib, a potent inhibitor of c-Kit and PDGFR, has been evaluated in phase II clinical trials for treating bone sarcomas. However, this compound failed to demonstrate significant antitumor activity as a single agent in children with recurrent OS and ES [70, 71]. Since blocking PDGF/PDGFR signaling is not sufficient to inhibit tumor progression in patients, other multi-targeted RTK inhibitors such as dasatinib are currently being studied in phase I/II studies for patients with advanced sarcomas (Table 1).

As a member of the ezrin/radixin/moesin (ERM) family, ezrin links the actin cytoskeleton to the plasma membrane. Gene microarray studies demonstrated increased Ezrin expression in metastatic OS lesions [72]. Further, high ezrin expression in OS patients, both human and canine, was correlated with poor overall survival [72, 73]. Khanna and colleagues [73, 74] demonstrated that early steps in OS pulmonary metastases are dependent on ezrin-mediated protein kinase B (AKT) and MAPK signaling, and reduction of ezrin expression by a short hairpin RNA (shRNA) decreased the survival of metastatic cells in the lung. The relevance of ezrin in metastatic disease has been validated for other sarcomas, including ES, although in this model ezrin mediates metastasis by signaling through the AKT/mammalian target of rapamycin (mTOR) pathway [75]. Two small-molecule ezrin inhibitors have been successfully studied in vitro and in vivo using OS models, but these agents still await testing in clinical trials [76].

The mTOR is a serine/threonine kinase and integral effector of the PI3K–AKT signaling pathway. It regulates cell cycle progression and protein synthesis among other steps during carcinogenesis [77]. Rapamycin (sirolimus) and its derivatives have been effective at reducing tumor growth in OS and ES murine models and in clinical trials [78‐82] and has been used as a radiosensitizer for OS [83]. A recent phase III clinical trial tested ridaforolimus in adult sarcoma patients who had achieved objective responses with prior chemotherapy [84]. For the 702 patients treated on that study (only 10 % had bone sarcoma), ridaforolimus increased progression-free survival by 28 % (p < 0.001), but greatly increased grade 3 or higher toxicities, especially stomatitis, cytopenias, and infection. The report does not provide a subset analysis for the bone sarcoma patients. Several clinical trials using mTOR inhibitors in combination therapies are in progress (Table 1).

The steroid receptor co-activator (Src) family of kinases is expressed at high levels and is constitutively active in many cancers, including OS and ES. Pharmacologic inhibition of Src in vitro led to apoptosis and decreased invasion, migration, and adhesion of OS and ES cells; however, these results were not reproducible using OS in vivo models, pointing at possible redundancy in activation of downstream effectors like focal adhesion kinase (FAK) [85‐88]. A dual inhibitor of BCR-Abl and Src, dasatinib has been used in one clinical trial where the maximum tolerated dose was determined but no objective responses were observed [89]. Clinical trials are in progress with either dasatinib alone or in combination therapy or with saracatinib, an Src-specific inhibitor.

Signaling via the Notch pathway is essential for the development of most organ systems, including for both neurogenesis [90] and osteoblast maturation [91]. Activation of the Notch pathway is required for vasculogenesis during tumor progression in ES [92]. Notch has been linked to increased invasion and metastasis in OS, in part through promoting a tumor-initiating cell phenotype [93‐95]. Membrane-bound Notch activation upon ligand binding occurs through a two-step proteolytic process carried by ADAMs family proteases followed by gamma-secretase cleavage, releasing a soluble intracellular Notch that can regulate transcription [96]. A phase I clinical trial in advanced solid malignancies using a gamma-secretase inhibitor (GSI) showed anti-tumor activity and a low toxicity profile [97]. A phase I/II clinical trial using a GSI in combination with an inhibitor of the hedgehog pathway for the treatment of metastatic sarcomas is currently recruiting patients. In considering the effects of GSI, one should recall that GSIs inhibit the processing of several receptors that effect metastasis, including Her-4, CD44, E-cadherin, and N-cadherin [98].

The hedgehog pathway is important for embryonic development and is dysregulated in various cancers. High expression of the hedgehog ligands and targets are observed in both OS and ES models, where this pathway is activated in both a ligand-dependent and a ligand-independent manner [99‐101]. Interestingly, EWS-FLI1 signaling is mediated through GLI, an effector and transcription regulator in the hedgehog pathway [102]. Inhibition of the hedgehog pathway in vitro and in ES and OS xenografts has been successful and warrants further research [100, 103]. In a recent clinical trial in adult patients with advanced solid tumors, an oral inhibitor of the hedgehog pathway was fairly well tolerated [104]. While the skeletal abnormalities seen in young mice briefly treated with hedgehog pathway inhibitors might raise concerns about pediatric applications for these agents [105], most OS and ES patients are close to their expected adult size at diagnosis, suggesting that these concerns should not preclude study.

Histone deacetylase inhibitors (HDACi) have been studied in cancer due to their effects in promoting transcription of tumor suppressor genes silenced during malignant transformation. Phase I clinical trials in pediatric patients with relapsed or refractory solid tumors using pracinostat or vorinostat monotherapy showed no tumor responses [106, 107]. Patient trials are underway using combination therapy with HDACi and adjuvant chemotherapies and may have greater promise (Table 1). Additionally, treatment with HDACi in preclinical models caused upregulation of natural killer (NK) cell recognition markers and of the apoptosis-promoting Fas receptor, resulting in increased sensitivity to NK-mediated killing [108, 109]. These results warrant further investigation in clinical trials of HDACi plus NK cells.

Ras proteins are small GTPases that regulate cell proliferation, apoptosis, and survival by activating multiple downstream signaling pathways, including MAPK. Though constitutively active, Ras mutations are uncommon in pediatric sarcomas; targeting Ras reduced tumor growth, possibly due to the many pathways requiring Ras relay signals [110‐113]. Reolysin is an oncolytic virus that selectively targets Ras transformed cells, and xenografts showed tumor growth inhibition by reolysin used alone or with chemotherapy agents [114]. A phase II study in sarcoma patients has been completed. While partial results were presented at the American Society of Clinical Oncology (ASCO) annual meeting in 2009 [115], there have been no peer-reviewed publications for sarcoma since that abstract was presented.

MDM2 is a ubiquitin ligase that regulates p53 activity by targeting this tumor suppressor for proteasomal degradation. Nutlins are small molecules that inhibit MDM2 and p53 binding, leading to increased availability of p53. Treatment using nutlins have been effective in OS and ES models, inducing apoptosis and cell cycle arrest [116‐118]. RG7112, a nutlin family member, induced tumor regression in ES models, but no objective response was observed with OS models [119]. Recently, a phase I clinical trial using RG7112 in patients with relapsed or refractory tumors was completed, though results have yet to be reported.

Bisphosphonates, which inhibit the mevalonate pathway at high concentrations and impede osteoclast-mediated bone resorption through induction of osteoclast apoptosis, have been shown to suppress tumor growth and pulmonary metastasis of ES in preclinical models [124‐128]. To date, several types of bisphosphonates, including zoledronate, pamidronate, and alendronate, displayed significant anti-tumor activity in vitro and in vivo [129‐132]. A phase II study evaluating the combination of chemotherapy and pamidronate for patients with OS demonstrated little impact on patient survival [133]. However, pamidronate has been shown to improve the durability of limb reconstruction [133]. In a recently completed phase I study, the addition of zoledronate to conventional multi-agent chemotherapy was safe but failed to demonstrate statistically significant differences in event-free or overall survival in patients with newly diagnosed metastatic OS [134]. However, our clinical team has treated many patients with bone metastasis of OS with zoledronate, and we have found that patients usually do not develop new bone metastases after receiving four to six doses of monthly zoledronate. We also have the impression that the need for opiates during palliation is reduced after patients receive bisphosphonates, suggesting that the clinical trials performed to date may not have looked at the correct endpoints. Currently, three phase II/III trials that evaluate the efficacy of zoledronate as a single agent or as an adjuvant to chemotherapy in localized and metastatic OS and ES are ongoing (Table 1).

Conjugated radioisotopes such as Samarium (¹⁵³Sm) lexidronam (Samarium-153 EDTMP) and radium-223 dichloride (Xofigo) have high specificity for bone uptake, which allows for the local delivery of high-dose radiation in bone tumors [135, 136]. Standard dose of Samarium-153 EDTMP was originally approved by the US FDA for pain management in patients with bone metastases, and radium-223 was recently approved for the treatment of castration-resistant prostate cancer patients with symptomatic bone metastases. Although radiation therapy is not widely used in treatment for OS, high-dose conjugated radioisotopes are under clinical investigation for their anti-tumor activities against OS. In a follow-up study of 14 patients with osteoblastic OS, Samarium-153 EDTMP in combination with the radiosensitizer gemcitabine induced short-term anti-tumor response in eight patients [137]. Thus far, conjugated radioisotopes have no clear role in Ewing sarcoma. The ongoing clinical trials for this class of agents include a phase I/II study for radium-223 dichloride and a phase II study for Samarium-153 EDTMP in combination with external radiotherapy in high-risk OS (Table 1).

Among the signaling molecules that have been associated with worse outcome in OS is the receptor activator of nuclear factor-kb (RANK), along with its ligand (RANKL) and decoy osteoprotegerin (OPG), which normally are essential for regulation of the homeostasis between bone lysis and formation during bone remodeling [138, 139]. High expression of RANKL is associated with reduced survival in OS [140], and some OS cell lines have functional RANK expression [141], allowing for possible autocrine stimulation of this pathway. Inhibition of RANK with shRNA reduced motility and anoikis resistance in OS cell lines, while overexpression of RANK using a retroviral vector increased OS cell motility without affecting proliferation [142].

Denosumab is an mAb specific for human RANKL and was developed initially to treat osteoporosis [143] and was later found effective in treating painful bone metastasis [144‐147]. It was subsequently found to be an effective treatment for giant cell tumor of bone [148], a benign but destructive neoplasm in which transformed mononuclear cells secrete RANKL, causing osteoclast hyperactivity. We have found that denosumab can be effective in treating painful bone metastasis in OS, which is in line with the FDA-approved indication for the drug. Whether it will have any direct effect against OS in patients remains to be seen.

Part of the pathogenesis of bone sarcomas includes the ability to invade through extracellular matrix tissues and to recruit a new blood supply as tumors grow [166]. As a part of hematogenous metastasis, tumor cells must also gain access to the endovascular space [91]. These activities typically proceed through hijacking normal biological processes that are then exploited by tumor cells to facilitate their growth and spread [167].