Ewing’s Sarcoma: Overcoming the Therapeutic Plateau

INTRODUCTION

Because Ewing’s sarcoma is a disease of young patients, it stands out among the greater than 50 diverse types of sarcomas. It is, in fact, the second most common malignant bone tumor in children, adolescents and young adults(Grier et al., 2003; Herzog, 2005; Subbiah & Anderson, 2011; Subbiah et al., 2009). Recently, there has been tremendous progress in understanding the biology of Ewing’s sarcoma. Ewing’s sarcoma is a translocation-positive sarcoma subtype with the pathognomonic hallmark being fusion of EWS(22q12) and FLI(11q24). The consequent EWS/FLI1 translocation results in a hybrid transcript (Subbiah & Anderson, 2011). Additionally, several new and exciting targeted agents have entered different stages of preclinical and clinical development and have a potentially promising role in managing Ewing’s sarcoma. These advances are timely in that a clinical therapeutic plateau has been reached in treating Ewing’s sarcoma patients with cytotoxic chemotherapeutic agents that have remained the standard-of-care for the last two decades. While the effect of these agents has benefited approximately 70% of patients with Ewing’s sarcoma who present with localized disease, this is not the case for patients who present with metastatic disease at diagnosis, or for those whose localized disease relapses (Cotterill et al., 2000; Khoury, 2008; Smith et al., 2010; Stahl et al., 2012). This group of patients has an event-free survival across all studies and has remained at less than 20%. To date, survival after relapsed or metastatic Ewing’s sarcoma is poor(Cotterill et al., 2000; Hawkins, 2012; Stahl et al., 2012; Subbiah & Anderson, 2011).

Targeted agents have revolutionized the therapeutic landscape of many cancers, beginning with imatinib for bcr/c-abl-positive chronic myeloid leukemia(Kantarjian et al., 2002), c-kit-positive gastrointestinal stromal tumor(Demetri et al., 2002) and, more recently, BRAF inhibitors in melanoma(Chapman et al., 2011) and EML4-ALK translocation-directed therapy in non-small cell lung cancer(Kwak et al., 2010). Hopefully, this success will translate to Ewing’s sarcoma, whose oncogenic EWS/FLI1 translocation is present in tumor cells, but not in normal cells(Erkizan et al., 2009; Subbiah & Anderson, 2011; Uren & Toretsky, 2005). This chimeric gene presents a rational target. However, the challenge remains to develop a targeted agent for this very rare disease that can be translated into a durable clinical benefit.

Recently, inhibitors of the insulin-like growth factor 1 receptor (IGF1R) have ignited considerable excitement in early phase clinical trials(Juergens et al., 2011; Kurzrock et al., 2010; Malempati et al., 2012; Naing et al., 2011; Pappo et al., 2011; Subbiah et al., 2009; Subbiah et al., 2011b). Dramatic responses have been reported in about 10% of patients with advanced, heavily pretreated Ewing’s sarcoma. Adding an inhibitor of mammalian target of rapamycin (mTOR) to IGF1R inhibitor therapy has yielded tumor regression in approximately 25-30% of patients with refractory metastatic disease(Naing et al., 2011; Subbiah et al., 2011b). These drugs are also remarkably well tolerated. A remaining challenge is identifying the 10-30% of patients with Ewing’s sarcoma who do respond to such targeted agents, managing patients who develop secondary resistance, and addressing the approximately 75-90% of individuals who have not thus far responded to any targeted agents.

Here, we discuss advances in understanding the biology of Ewing’s sarcoma, findings from recent clinical trials that have shown promise in this disease, and catalog agents currently in early clinical trials that may be relevant to the success of targeted therapy in Ewing’s sarcoma.

Therapeutic Targets and Recent Clinical Trials

Insulin-like growth factor 1 receptor inhibitors

Recently, several studies showed the activity of single-agent IGF1R inhibitors in patients with Ewing’s sarcoma(Juergens et al., 2011; Malempati et al., 2012; Olmos et al., 2010; Pappo et al., 2011). All results demonstrated marked responses in a select subset of patients with Ewing’s sarcoma. However, majority of the patients did not respond and the studies uniformly reported “limited single-agent activity” (Pappo et al., 2011). Because of the limited response rate to single-agent IGF1R inhibitors, their development has stalled. The rarity of the disease combined with the potential efficacy of these inhibitors suggests that scientists and clinicians would be well served by collaborating to ensure that we do not lose drugs that exhibit marked activity in a subset of patients(Subbiah et al., 2011a). Since the responses can be dramatic, it is critical to determine if only a minority of patients respond because they represent a molecular subgroup or because treatment has been given so late in the disease course that resistance pathways have emerged in most patients. The key to deciphering these questions is to use next-generation deep sequencing to profile responders versus non-responders, and to perform a trial that incorporates IGF1R inhibitors early in the course of the disease.

Pilot study with combined IGF1R plus mTOR inhibition

In order to overcome resistance, several studies exploited the combination of IGF1R and mTOR inhibitors, with preliminary results showing tumor regression rates of approximately 25-30%(Naing et al., 2011; Naing et al., 2012; Subbiah et al., 2011b). For instance, among 17 patients with advanced, heavily pretreated Ewing’s sarcoma, a greater than 20% tumor regression was seen in five patients (29%) who remained on study for 8-27+ months, again demonstrating that the IGF1R pathway is active in Ewing’s sarcoma(Naing et al., 2012).

In a recent pilot study, our group highlighted two unusual responders with refractory Ewing’s sarcoma(Subbiah et al., 2011b). Both patients had marked responses to a single-agent IGF1R inhibitor and then overcame ensuing resistance with combined cixutumumab (IGF1R) and mTOR inhibitor therapy(Subbiah et al., 2011b) (Naing et al., 2011). Morphoproteomic profiling was performed to better understand response/resistance mechanisms to IGF1R–based therapy (Brown, 2009; Subbiah et al., 2011b). Briefly, we performed histopathological analysis and applied immunohistochemical probes for p-mTOR (Ser2448), p-Akt (Ser473), p-ERK1/2 (Thr202/Tyr204), nestin, and p-STAT3 (Tyr705) in initial and resistant tumor specimens(Brown, 2009; Subbiah et al., 2011b). After relapsing following multiple lines of therapy, including three different phase I studies, patient #1 with advanced Ewing’s sarcoma had a dramatic response to an IGF1R antibody [(R1507) ({"type":"clinical-trial","attrs":{"text":"NCT00400361","term_id":"NCT00400361"}}NCT00400361)], which continued for 35 months on a phase I study(Kurzrock et al., 2010) (Figure 1). After progression, the patient enrolled on another phase I study using a different IGF1R antibody, cixutumumab, in combination with temsirolimus (CCI-779) (Naing et al., 2011). After 18 months of treatment, positron emission therapy and computed tomography (PET/CT) and chest CT scans demonstrated no disease (complete response) (Figure 1). p-Akt and p-mTOR were upregulated in the tumor that relapsed after an initial response to IGF1R antibody. Upregulated p-Akt and p-mTOR is, therefore, a likely mechanism for the resistance that developed as well as the subsequent response to combined IGF1R plus mTOR inhibitor therapy. Similarly, patient #2, after failing multiple lines of therapy, was enrolled in a phase II study of another IGF1R inhibitor, SCH 717454 ({"type":"clinical-trial","attrs":{"text":"NCT00617890","term_id":"NCT00617890"}}NCT00617890), and initially demonstrated significant tumor regression. After progression, the patient was enrolled on the phase I study with the IGF1R plus mTOR combination(Naing et al., 2011), but was removed from the study because of a single treatment-resistant nodule, with three other nodules demonstrating an initial marked response. Histopathologic and morphoproteomic analysis of the resistant nodule revealed overexpressed pathways, including MAP/ERK (Subbiah et al., 2011b). These data suggest that IGF1R monoclonal antibodies combined with MEK, ERK, and mTOR inhibitors might produce optimal antitumor activity. This small study defines the significant complexity of resistance mechanisms that defy the usual linear depiction of signaling pathways involved in tumorigenesis. Correlative studies using next-generation deep sequencing in patients who responded and then acquired resistance are underway and are expected to help delineate the precise mechanisms of resistance or response to IGF1R inhibitors.

Figure 1

Responses to IGF1R +/- mTOR inhibitor therapy (Kurzrock et al., 2010; Naing et al., 2012; Subbiah et al., 2011b)

Top panel: CT of the thorax in a patient with Ewing’s sarcoma showing a response to an IGF1R antibody (R1507) alone (Kurzrock et al., 2010). Top left panel is a pre-treatment CT scan of the thorax showing metastatic Ewing’s sarcoma in the lung. Top right panel: Six weeks after IGF1R antibody (R1507) therapy, showing regression of tumor of the thorax in this patient.

Bottom panel: Bottom left panel shows a pre-treatment CT scan of the thorax in patient with metastatic Ewing’s sarcoma in the lung. Bottom right panel: Six months after IGF1R antibody plus mTOR inhibitor therapy showing continued response(Naing et al., 2012).

Our study, although limited by its small sample size, contributes to understanding the signaling cascades underlying Ewing’s sarcoma and how these might be efficaciously addressed (Subbiah et al., 2011b). As observed in other cancers, combined targeted inhibition (with IGF1R and mTOR inhibitors) yielded a superior benefit over monotherapy in some patients. MEK/ERK might be a resistance pathway abrogating successful IGF1R and mTOR inhibition. Of interest, tumors that ultimately fail to continue to respond to specific single-agent drugs may re-respond when another agent is added in a rational fashion.(Naing & Kurzrock, 2010) The latter results suggest the need to rethink the standard eligibility criteria that often exclude patients who already received a particular class of agent from trials testing the same class of agents in combination. Another general strategy that is assuming increasing practical importance is instituting biomarker analysis as an integral and dynamic component of targeted therapy studies. Morphoproteomic analysis is one such approach and deep sequencing is another.

In spite of the fact that refractory Ewing’s sarcoma is fatal(Subbiah et al., 2009), multiple IGF1R inhibitors in phase I and II studies have shown clinical benefit (Kurzrock et al., 2010; Naing et al., 2011; Pappo et al., 2011; Subbiah et al., 2011b). This recent success demonstrating “limited single-agent activity” of IGF1R therapy in Ewing’s sarcoma notwithstanding(Malempati et al., 2012), development of IGF1R inhibitors needs to be pursued. Although rare, Ewing’s sarcoma is the second most common bone malignancy, striking children, adolescents and young adults in the prime of their lives. IGF1R inhibitors are particularly attractive for these patients because of their benign side effects.

Earlier Treatment

Studies suggest that treating patients with targeted agents earlier in their disease trajectory rather than later, as has been customary, might be associated with higher response rates(Gold et al., 2011). For example, this potential is reflected by the poor response rate in advanced chronic myeloid leukemia (CML) in blast phase crisis (~15 %) treated with the Bcr-Abl inhibitor imatinib versus the excellent response rate (~90%) achieved in newly diagnosed disease. Bcr-Abl kinase inhibitors increased survival in CML blast crisis from approximately 6 months to approximately one year. But the use of these inhibitors in newly-diagnosed CML transformed the disease, increasing survival from 4 to approximately 25 years. (Gambacorti-Passerini et al., 2011) Hence, moving IGF1R and mTOR inhibitor therapy to the frontline setting, perhaps by integrating it with standard-of-care chemotherapy, might ultimately answer the question as to whether these targeted agents confer durable benefit, improving relapse-free and overall survival.

Targeted agents combined with radiation and surgery for therapeutic synergy and local control

Beginning with its original description by James Ewing in the 1920s, Ewing’s sarcoma has proved to be a radiosensitive entity, which may provide the basis for synergistic combinations with targeted IGF1R inhibitors in high-risk patients. And, in fact, several preclinical models have shown that IGF1R inhibition does behave as a potential radiosensitizing agent (Iwasa et al., 2009; Perer et al., 2000; Yavari et al., 2010). With recent unprecedented advances in the field of radiation such as proton therapy, this synergistic approach may be an exciting intervention in high-risk Ewing’s sarcoma patients. Future studies would benefit from exploring such potential in controlled clinical trials to evaluate if this approach does, in fact, work and whether optimal administration should be sequential or concurrent. Incorporating this strategy with surgery may also benefit patients, even those with advanced disease who respond in order to achieve “no evidence of disease” status.

IGF1R small molecule inhibitors

Thus far, all published studies that demonstrated IGF1R inhibitor activity in Ewing’s sarcoma have used antibodies. In addition to antibodies, several small molecule tyrosine kinase inhibitors of IGF1R/IR are in preclinical and clinical development and might have activity in patients with Ewing’s sarcoma(Subbiah & Anderson, 2011). Some of these drugs also have multi-tyrosine kinase activity, which could conceivably contribute to responses as well.

Agents targeting the EWS/FLI1 fusion protein

Theoretically, an attractive approach to eradicating Ewing’s sarcoma, especially at the stem cell level, will be targeting the tumor-specific EWS/FLI1 t(11;22) translocation. Attempts have been made in the preclinical arena to effectively target the oncogenic EWS/FLI1 fusion protein. These endeavors have been directed both at the fusion protein itself or inhibiting key signals in the process of transcription, translation, and EWS/FLI1 protein function. Several such attempts have recently been reported. One of them involves a new class of small molecule inhibitors as well as resurrecting an older drug, mithramycin, which has demonstrated activity in a preclinical model and also another multi-tyrosine kinase, which is in clinical trials for AML(Erkizan et al., 2009; Grohar et al., 2011).

Small molecule approaches

Binding of EWS/FLI1 protein to RNA Helicase A (RHA) is needed for the oncogenicity of the EWS/FLI1 fusion gene. Recently, a small molecule compound (YK-4279) was identified through surface plasmon resonance, which was shown to block in vitro and in vivo interaction between EWS/FLI1 and RNA helicase A (Erkizan et al., 2009). Subsequent research showed that only one enantiomer of this (specifically, the S–form of YK-4279) was active in disrupting the interaction(Barber-Rotenberg et al., 2012). This early identification of enantio-specific targeting is an important step in the preclinical development of this compound.

Mithramycin

After screening 50,000 compounds to target the EWS/FLI1 translocation using high-throughput technology, the group at the National Cancer Institute (NCI) identified mithramycin as the lead compound(Grohar et al., 2011). Mithramycin was shown to specifically inhibit the expression of EWS/FLI1 downstream targets in vitro and Ewing’s sarcoma xenografts in vivo(Grohar et al., 2011). In the 1960s, mithramycin demonstrated promising activity in several tumor types, including testicular germ cell tumors and Ewing’s sarcoma. Because of the introduction of several different drugs, mithramycin was largely abandoned. After this preclinical study the NCI group is attempting to resurrect mithramycin for the treatment of Ewing’s sarcoma(Grohar et al., 2011).

Midostaurin

Another study using small molecule screening technology identified a list of drugs showing promise for targeting EWS/FLI1 activity. In addition to known drugs such as doxorubicin and etoposide, as well as a multikinase inhibitor midastaurin (PKC412), other drugs have been identified(Boro et al., 2012) and showed activity in vivo and in vitro. These molecules merit clinical exploration(Boro et al., 2012).

Poly(ADP-ribose) polymerase (PARP) inhibitors

A recently published pharmacogenomic profiling study attempted to uncover biomarkers of sensitivity and resistance in 639 cell lines through screening 130 drugs in preclinical or clinical development (Garnett et al., 2012). Interestingly, EWS/FLI1 emerged as a poly(ADP-ribose) polymerase (PARP) inhibitor biomarker. Specifically, olaparib (AZD2281) had potent activity in cell lines harboring an EWS/FLI1 translocation (Figure 2) (Garnett et al., 2012). While we know that BRCA1 or BRCA2 mutations confer sensitivity to PARP inhibitors, no similar mutations have been identified in Ewing’s sarcoma. However, prostate cancer cell lines bearing ERG transcripts such as FLI1 are members of the ETS gene family and have demonstrated sensitivity to PARP inhibitors(Garnett et al., 2012). Another study in preclinical models showed synergy when combining PARP inhibitors and temozolomide in Ewing’s sarcoma(Brenner et al., 2012). In fact, a clinical trial testing these agents is currently ongoing (Clinical trials.gov ID no: {"type":"clinical-trial","attrs":{"text":"NCT01294735","term_id":"NCT01294735"}}NCT01294735) and if this rapid proof-of-principle study shows promise, the combination could be expanded to enroll additional patients.

Figure 2

Simplified Insulin-like growth factor 1 receptor (IGF1R) and mTORC1 and mTORC2 signaling [Modified from (Subbiah et al., 2011a; Subbiah et al., 2011b)]

Insulin receptor substrate 1 (IRS1) and phosphatidylinositol 3-kinases (PI3K) are activated by insulin and/or IGF1 signaling at the insulin receptor level. PDK1 and Akt are recruited to the plasma membrane by products of PI3K, PIP2, PIP3 (phosphatidylinositol 3,4,5 trisphosphate and phosphatidylinositol 3,4 bisphosphate). After this event there is phosphorylation and activation of Akt by the mTORC2 complex (mTOR + mLST8+ Rictor). This leads to a chain of activation of numerous targets by Akt. The TSC1/2 complex is also phosphorylated in this fashion. By this mechanism inactivation of TSC2’s GAP activity for the small G protein Rheb is initiated. Subsequently, the new mTORC1 complex (mTOR + mLST8+Raptor) is activated by GTP-bound Rheb and phosphorylates proteins like S6K. This starts a negative feedback loop to modulate auto-activity, through a S6K-mediated pathway decrease in the activation of PI3K. IGF1R antibody inhibits IGF1R signaling and rapalogs inhibit mTORC1 short term, and eventually inhibit mTORC 2 with chronic exposure and also suppress EWS/FLI1 which drives tumorigenesis in Ewing’s sarcoma.

The Biology of Ewing’s Sarcoma

The last two decades have witnessed the precise characterization of Ewing’s sarcoma, also referred to as the “Ewing’s sarcoma family of tumors”. Ewing’s sarcoma tumor cells are thought to arise from primitive mesenchymal stem cells that have the ability to heterogeneously differentiate into an osteogenic, adipogenic, or neurogenic lineage of cells. Historically described distinct entities such as extraskeletal Ewing’s sarcoma, Askin’s tumor, and primitive neuroectodermal tumors (PNET) are each characterized by the pathognomic EWS/FLI1 translocation, identified either by reverse transcription polymerase chain reaction (RT-PCR) or fluorescence in situ hybridization (FISH)(Khoury, 2008), and all of them belong to the Ewing’s sarcoma family of tumors(Subbiah et al., 2009). Histologically, tumor cells harboring the EWS/FLI1 fusion transcript are small, round and blue. Immunohistochemical staining positive for CD99 is considered a universal immunophenotypic hallmark. This hallmark translocation is seen in approximately 85% of Ewing’s sarcoma patient samples. The second most common translocation seen is the fusion of EWS/ERG transcript (around 5-15%). There are several other translocations, with the Ewing’s sarcoma gene as the fusion partner described in fewer than 1% of patients with Ewing’s sarcoma (Table 1)(Ordonez et al., 2009). The EWS gene also participates in other translocation-positive sarcoma subtypes (Table 1)(Ordonez et al., 2009). Ewing’s sarcoma-like small round blue cell tumors of the bone, which lack the EWS/FLI1 transcript, are considered as translocation-negative Ewing’s sarcoma and are treated as such. A recent study using RNA-sequencing technology discovered a new fusion type of bone sarcoma BCOR-CCNB3 among samples from patients that were diagnosed as “small round cell tumors of bone, possibly Ewing sarcoma”. This new finding may have significant therapeutic impact if downsteam targets are identified(Pierron et al., 2012).

Updates on the molecular biology of the EWS/FLI1 fusion transcript have identified downstream targets(Lessnick & Ladanyi, 2012; Mackintosh et al., 2010; Ordonez et al., 2009; Toomey et al., 2010). Briefly, EWS/FLI1 modulates and mediates oncogenesis in various ways, contributing to the hallmarks of Ewing’s sarcoma, namely initiation, survival, maintenance, progression, and evasion of apoptosis/ senescence and metastasis(Ordonez et al., 2009). Other mutations in these tumors, such as p53, may also contribute to tumor aggressiveness(Ordonez et al., 2009). In addition to these functions, the EWS/FLI1 fusion transcript interacts with key signaling pathways, including the IGF1/IGF1R-mediated autocrine loop, MAPK signaling and STAT 3 signaling pathways(Ordonez et al., 2009). Importantly EWS/FLI1 downregulates insulin-like growth factor-binding protein 3, which in turn, upregulates IGF/IGF1R. This effect on the IGF1R signaling machinery may explain why IGF1R inhibitors can be effective in this disease(Huang et al., 2011b).

Standard Therapy for Ewing’s Sarcoma

The current standard-of-care for Ewing’s sarcoma is cytotoxic chemotherapy with vincristine, adriamycin, and cyclophosphamide, alternating with ifosfamide and etoposide. Surgery is used for local control, with radiation in select patients. This treatment approach confers benefit to approximately 70% of patients presenting with localized disease(Khoury, 2008; Smith et al., 2010; Stahl et al., 2012; Subbiah et al., 2009). In contrast, prognosis is guarded in patients with Ewing’s sarcoma with recurrent and metastatic disease at presentation. Additionally, the five-drug standard-of-care regimen can result in serious or life-threatening cytopenias and mucositis. Improvements in supportive care using growth factors, multiple transfusions and prophylactic antibiotics have been helpful in addressing these sequelae. Second-line agents such as temozolomide/irinotecan, topotecan/cyclophosphamide, and gemcitabine/docetaxel mainly serve to prolong survival(Subbiah et al., 2009). Relapsed/refractory Ewing’s sarcoma remains uniformly fatal and novel approaches are urgently needed(Hawkins, 2012).

Therapeutic Targets and Recent Clinical Trials

Pilot study with combined IGF1R plus mTOR inhibition

In order to overcome resistance, several studies exploited the combination of IGF1R and mTOR inhibitors, with preliminary results showing tumor regression rates of approximately 25-30%(Naing et al., 2011; Naing et al., 2012; Subbiah et al., 2011b). For instance, among 17 patients with advanced, heavily pretreated Ewing’s sarcoma, a greater than 20% tumor regression was seen in five patients (29%) who remained on study for 8-27+ months, again demonstrating that the IGF1R pathway is active in Ewing’s sarcoma(Naing et al., 2012).

In a recent pilot study, our group highlighted two unusual responders with refractory Ewing’s sarcoma(Subbiah et al., 2011b). Both patients had marked responses to a single-agent IGF1R inhibitor and then overcame ensuing resistance with combined cixutumumab (IGF1R) and mTOR inhibitor therapy(Subbiah et al., 2011b) (Naing et al., 2011). Morphoproteomic profiling was performed to better understand response/resistance mechanisms to IGF1R–based therapy (Brown, 2009; Subbiah et al., 2011b). Briefly, we performed histopathological analysis and applied immunohistochemical probes for p-mTOR (Ser2448), p-Akt (Ser473), p-ERK1/2 (Thr202/Tyr204), nestin, and p-STAT3 (Tyr705) in initial and resistant tumor specimens(Brown, 2009; Subbiah et al., 2011b). After relapsing following multiple lines of therapy, including three different phase I studies, patient #1 with advanced Ewing’s sarcoma had a dramatic response to an IGF1R antibody [(R1507) ({"type":"clinical-trial","attrs":{"text":"NCT00400361","term_id":"NCT00400361"}}NCT00400361)], which continued for 35 months on a phase I study(Kurzrock et al., 2010) (Figure 1). After progression, the patient enrolled on another phase I study using a different IGF1R antibody, cixutumumab, in combination with temsirolimus (CCI-779) (Naing et al., 2011). After 18 months of treatment, positron emission therapy and computed tomography (PET/CT) and chest CT scans demonstrated no disease (complete response) (Figure 1). p-Akt and p-mTOR were upregulated in the tumor that relapsed after an initial response to IGF1R antibody. Upregulated p-Akt and p-mTOR is, therefore, a likely mechanism for the resistance that developed as well as the subsequent response to combined IGF1R plus mTOR inhibitor therapy. Similarly, patient #2, after failing multiple lines of therapy, was enrolled in a phase II study of another IGF1R inhibitor, SCH 717454 ({"type":"clinical-trial","attrs":{"text":"NCT00617890","term_id":"NCT00617890"}}NCT00617890), and initially demonstrated significant tumor regression. After progression, the patient was enrolled on the phase I study with the IGF1R plus mTOR combination(Naing et al., 2011), but was removed from the study because of a single treatment-resistant nodule, with three other nodules demonstrating an initial marked response. Histopathologic and morphoproteomic analysis of the resistant nodule revealed overexpressed pathways, including MAP/ERK (Subbiah et al., 2011b). These data suggest that IGF1R monoclonal antibodies combined with MEK, ERK, and mTOR inhibitors might produce optimal antitumor activity. This small study defines the significant complexity of resistance mechanisms that defy the usual linear depiction of signaling pathways involved in tumorigenesis. Correlative studies using next-generation deep sequencing in patients who responded and then acquired resistance are underway and are expected to help delineate the precise mechanisms of resistance or response to IGF1R inhibitors.

Figure 1

Responses to IGF1R +/- mTOR inhibitor therapy (Kurzrock et al., 2010; Naing et al., 2012; Subbiah et al., 2011b)

Top panel: CT of the thorax in a patient with Ewing’s sarcoma showing a response to an IGF1R antibody (R1507) alone (Kurzrock et al., 2010). Top left panel is a pre-treatment CT scan of the thorax showing metastatic Ewing’s sarcoma in the lung. Top right panel: Six weeks after IGF1R antibody (R1507) therapy, showing regression of tumor of the thorax in this patient.

Bottom panel: Bottom left panel shows a pre-treatment CT scan of the thorax in patient with metastatic Ewing’s sarcoma in the lung. Bottom right panel: Six months after IGF1R antibody plus mTOR inhibitor therapy showing continued response(Naing et al., 2012).

Our study, although limited by its small sample size, contributes to understanding the signaling cascades underlying Ewing’s sarcoma and how these might be efficaciously addressed (Subbiah et al., 2011b). As observed in other cancers, combined targeted inhibition (with IGF1R and mTOR inhibitors) yielded a superior benefit over monotherapy in some patients. MEK/ERK might be a resistance pathway abrogating successful IGF1R and mTOR inhibition. Of interest, tumors that ultimately fail to continue to respond to specific single-agent drugs may re-respond when another agent is added in a rational fashion.(Naing & Kurzrock, 2010) The latter results suggest the need to rethink the standard eligibility criteria that often exclude patients who already received a particular class of agent from trials testing the same class of agents in combination. Another general strategy that is assuming increasing practical importance is instituting biomarker analysis as an integral and dynamic component of targeted therapy studies. Morphoproteomic analysis is one such approach and deep sequencing is another.

In spite of the fact that refractory Ewing’s sarcoma is fatal(Subbiah et al., 2009), multiple IGF1R inhibitors in phase I and II studies have shown clinical benefit (Kurzrock et al., 2010; Naing et al., 2011; Pappo et al., 2011; Subbiah et al., 2011b). This recent success demonstrating “limited single-agent activity” of IGF1R therapy in Ewing’s sarcoma notwithstanding(Malempati et al., 2012), development of IGF1R inhibitors needs to be pursued. Although rare, Ewing’s sarcoma is the second most common bone malignancy, striking children, adolescents and young adults in the prime of their lives. IGF1R inhibitors are particularly attractive for these patients because of their benign side effects.

Insulin-like growth factor 1 receptor inhibitors

Recently, several studies showed the activity of single-agent IGF1R inhibitors in patients with Ewing’s sarcoma(Juergens et al., 2011; Malempati et al., 2012; Olmos et al., 2010; Pappo et al., 2011). All results demonstrated marked responses in a select subset of patients with Ewing’s sarcoma. However, majority of the patients did not respond and the studies uniformly reported “limited single-agent activity” (Pappo et al., 2011). Because of the limited response rate to single-agent IGF1R inhibitors, their development has stalled. The rarity of the disease combined with the potential efficacy of these inhibitors suggests that scientists and clinicians would be well served by collaborating to ensure that we do not lose drugs that exhibit marked activity in a subset of patients(Subbiah et al., 2011a). Since the responses can be dramatic, it is critical to determine if only a minority of patients respond because they represent a molecular subgroup or because treatment has been given so late in the disease course that resistance pathways have emerged in most patients. The key to deciphering these questions is to use next-generation deep sequencing to profile responders versus non-responders, and to perform a trial that incorporates IGF1R inhibitors early in the course of the disease.

Pilot study with combined IGF1R plus mTOR inhibition

In order to overcome resistance, several studies exploited the combination of IGF1R and mTOR inhibitors, with preliminary results showing tumor regression rates of approximately 25-30%(Naing et al., 2011; Naing et al., 2012; Subbiah et al., 2011b). For instance, among 17 patients with advanced, heavily pretreated Ewing’s sarcoma, a greater than 20% tumor regression was seen in five patients (29%) who remained on study for 8-27+ months, again demonstrating that the IGF1R pathway is active in Ewing’s sarcoma(Naing et al., 2012).

In a recent pilot study, our group highlighted two unusual responders with refractory Ewing’s sarcoma(Subbiah et al., 2011b). Both patients had marked responses to a single-agent IGF1R inhibitor and then overcame ensuing resistance with combined cixutumumab (IGF1R) and mTOR inhibitor therapy(Subbiah et al., 2011b) (Naing et al., 2011). Morphoproteomic profiling was performed to better understand response/resistance mechanisms to IGF1R–based therapy (Brown, 2009; Subbiah et al., 2011b). Briefly, we performed histopathological analysis and applied immunohistochemical probes for p-mTOR (Ser2448), p-Akt (Ser473), p-ERK1/2 (Thr202/Tyr204), nestin, and p-STAT3 (Tyr705) in initial and resistant tumor specimens(Brown, 2009; Subbiah et al., 2011b). After relapsing following multiple lines of therapy, including three different phase I studies, patient #1 with advanced Ewing’s sarcoma had a dramatic response to an IGF1R antibody [(R1507) ({"type":"clinical-trial","attrs":{"text":"NCT00400361","term_id":"NCT00400361"}}NCT00400361)], which continued for 35 months on a phase I study(Kurzrock et al., 2010) (Figure 1). After progression, the patient enrolled on another phase I study using a different IGF1R antibody, cixutumumab, in combination with temsirolimus (CCI-779) (Naing et al., 2011). After 18 months of treatment, positron emission therapy and computed tomography (PET/CT) and chest CT scans demonstrated no disease (complete response) (Figure 1). p-Akt and p-mTOR were upregulated in the tumor that relapsed after an initial response to IGF1R antibody. Upregulated p-Akt and p-mTOR is, therefore, a likely mechanism for the resistance that developed as well as the subsequent response to combined IGF1R plus mTOR inhibitor therapy. Similarly, patient #2, after failing multiple lines of therapy, was enrolled in a phase II study of another IGF1R inhibitor, SCH 717454 ({"type":"clinical-trial","attrs":{"text":"NCT00617890","term_id":"NCT00617890"}}NCT00617890), and initially demonstrated significant tumor regression. After progression, the patient was enrolled on the phase I study with the IGF1R plus mTOR combination(Naing et al., 2011), but was removed from the study because of a single treatment-resistant nodule, with three other nodules demonstrating an initial marked response. Histopathologic and morphoproteomic analysis of the resistant nodule revealed overexpressed pathways, including MAP/ERK (Subbiah et al., 2011b). These data suggest that IGF1R monoclonal antibodies combined with MEK, ERK, and mTOR inhibitors might produce optimal antitumor activity. This small study defines the significant complexity of resistance mechanisms that defy the usual linear depiction of signaling pathways involved in tumorigenesis. Correlative studies using next-generation deep sequencing in patients who responded and then acquired resistance are underway and are expected to help delineate the precise mechanisms of resistance or response to IGF1R inhibitors.

Figure 1

Responses to IGF1R +/- mTOR inhibitor therapy (Kurzrock et al., 2010; Naing et al., 2012; Subbiah et al., 2011b)

Top panel: CT of the thorax in a patient with Ewing’s sarcoma showing a response to an IGF1R antibody (R1507) alone (Kurzrock et al., 2010). Top left panel is a pre-treatment CT scan of the thorax showing metastatic Ewing’s sarcoma in the lung. Top right panel: Six weeks after IGF1R antibody (R1507) therapy, showing regression of tumor of the thorax in this patient.

Bottom panel: Bottom left panel shows a pre-treatment CT scan of the thorax in patient with metastatic Ewing’s sarcoma in the lung. Bottom right panel: Six months after IGF1R antibody plus mTOR inhibitor therapy showing continued response(Naing et al., 2012).

Our study, although limited by its small sample size, contributes to understanding the signaling cascades underlying Ewing’s sarcoma and how these might be efficaciously addressed (Subbiah et al., 2011b). As observed in other cancers, combined targeted inhibition (with IGF1R and mTOR inhibitors) yielded a superior benefit over monotherapy in some patients. MEK/ERK might be a resistance pathway abrogating successful IGF1R and mTOR inhibition. Of interest, tumors that ultimately fail to continue to respond to specific single-agent drugs may re-respond when another agent is added in a rational fashion.(Naing & Kurzrock, 2010) The latter results suggest the need to rethink the standard eligibility criteria that often exclude patients who already received a particular class of agent from trials testing the same class of agents in combination. Another general strategy that is assuming increasing practical importance is instituting biomarker analysis as an integral and dynamic component of targeted therapy studies. Morphoproteomic analysis is one such approach and deep sequencing is another.

In spite of the fact that refractory Ewing’s sarcoma is fatal(Subbiah et al., 2009), multiple IGF1R inhibitors in phase I and II studies have shown clinical benefit (Kurzrock et al., 2010; Naing et al., 2011; Pappo et al., 2011; Subbiah et al., 2011b). This recent success demonstrating “limited single-agent activity” of IGF1R therapy in Ewing’s sarcoma notwithstanding(Malempati et al., 2012), development of IGF1R inhibitors needs to be pursued. Although rare, Ewing’s sarcoma is the second most common bone malignancy, striking children, adolescents and young adults in the prime of their lives. IGF1R inhibitors are particularly attractive for these patients because of their benign side effects.

Earlier Treatment

Studies suggest that treating patients with targeted agents earlier in their disease trajectory rather than later, as has been customary, might be associated with higher response rates(Gold et al., 2011). For example, this potential is reflected by the poor response rate in advanced chronic myeloid leukemia (CML) in blast phase crisis (~15 %) treated with the Bcr-Abl inhibitor imatinib versus the excellent response rate (~90%) achieved in newly diagnosed disease. Bcr-Abl kinase inhibitors increased survival in CML blast crisis from approximately 6 months to approximately one year. But the use of these inhibitors in newly-diagnosed CML transformed the disease, increasing survival from 4 to approximately 25 years. (Gambacorti-Passerini et al., 2011) Hence, moving IGF1R and mTOR inhibitor therapy to the frontline setting, perhaps by integrating it with standard-of-care chemotherapy, might ultimately answer the question as to whether these targeted agents confer durable benefit, improving relapse-free and overall survival.

Targeted agents combined with radiation and surgery for therapeutic synergy and local control

Beginning with its original description by James Ewing in the 1920s, Ewing’s sarcoma has proved to be a radiosensitive entity, which may provide the basis for synergistic combinations with targeted IGF1R inhibitors in high-risk patients. And, in fact, several preclinical models have shown that IGF1R inhibition does behave as a potential radiosensitizing agent (Iwasa et al., 2009; Perer et al., 2000; Yavari et al., 2010). With recent unprecedented advances in the field of radiation such as proton therapy, this synergistic approach may be an exciting intervention in high-risk Ewing’s sarcoma patients. Future studies would benefit from exploring such potential in controlled clinical trials to evaluate if this approach does, in fact, work and whether optimal administration should be sequential or concurrent. Incorporating this strategy with surgery may also benefit patients, even those with advanced disease who respond in order to achieve “no evidence of disease” status.

IGF1R small molecule inhibitors

Thus far, all published studies that demonstrated IGF1R inhibitor activity in Ewing’s sarcoma have used antibodies. In addition to antibodies, several small molecule tyrosine kinase inhibitors of IGF1R/IR are in preclinical and clinical development and might have activity in patients with Ewing’s sarcoma(Subbiah & Anderson, 2011). Some of these drugs also have multi-tyrosine kinase activity, which could conceivably contribute to responses as well.

Agents targeting the EWS/FLI1 fusion protein

Theoretically, an attractive approach to eradicating Ewing’s sarcoma, especially at the stem cell level, will be targeting the tumor-specific EWS/FLI1 t(11;22) translocation. Attempts have been made in the preclinical arena to effectively target the oncogenic EWS/FLI1 fusion protein. These endeavors have been directed both at the fusion protein itself or inhibiting key signals in the process of transcription, translation, and EWS/FLI1 protein function. Several such attempts have recently been reported. One of them involves a new class of small molecule inhibitors as well as resurrecting an older drug, mithramycin, which has demonstrated activity in a preclinical model and also another multi-tyrosine kinase, which is in clinical trials for AML(Erkizan et al., 2009; Grohar et al., 2011).

Small molecule approaches

Binding of EWS/FLI1 protein to RNA Helicase A (RHA) is needed for the oncogenicity of the EWS/FLI1 fusion gene. Recently, a small molecule compound (YK-4279) was identified through surface plasmon resonance, which was shown to block in vitro and in vivo interaction between EWS/FLI1 and RNA helicase A (Erkizan et al., 2009). Subsequent research showed that only one enantiomer of this (specifically, the S–form of YK-4279) was active in disrupting the interaction(Barber-Rotenberg et al., 2012). This early identification of enantio-specific targeting is an important step in the preclinical development of this compound.

Mithramycin

After screening 50,000 compounds to target the EWS/FLI1 translocation using high-throughput technology, the group at the National Cancer Institute (NCI) identified mithramycin as the lead compound(Grohar et al., 2011). Mithramycin was shown to specifically inhibit the expression of EWS/FLI1 downstream targets in vitro and Ewing’s sarcoma xenografts in vivo(Grohar et al., 2011). In the 1960s, mithramycin demonstrated promising activity in several tumor types, including testicular germ cell tumors and Ewing’s sarcoma. Because of the introduction of several different drugs, mithramycin was largely abandoned. After this preclinical study the NCI group is attempting to resurrect mithramycin for the treatment of Ewing’s sarcoma(Grohar et al., 2011).

Midostaurin

Another study using small molecule screening technology identified a list of drugs showing promise for targeting EWS/FLI1 activity. In addition to known drugs such as doxorubicin and etoposide, as well as a multikinase inhibitor midastaurin (PKC412), other drugs have been identified(Boro et al., 2012) and showed activity in vivo and in vitro. These molecules merit clinical exploration(Boro et al., 2012).

Early Phase Clinical Trials for Adolescent and Young Adult Patients with Ewing’s Sarcoma

Ewing’s sarcoma straddles a critical population of children and adolescents younger than 18 years and young adults older than 18 years. A possibility to address the need for new therapies is to enroll children, adolescent and young adult patients with relapsed Ewing’s sarcoma on Children’s Oncology Group (COG) or other clinical trials of targeted agents directed at diverse tyrosine kinases believed to be deregulated in this disease. Adolescents and young adults (15-40 years), in general, have poorer outcomes compared to younger children.

Drug development is a complex and long drawn-out process. Preclinical promise often does not translate to patient benefit, as in the case of cytarabine in Ewing’s sarcoma(DuBois et al., 2009). Enrollment of patients with rare diseases having dismal outcomes on a variety of targeted therapy trials may serve to provide clinical response signals and hence “go” versus “no-go” decisions in these tumor types (Subbiah & Kurzrock, 2011; Subbiah et al., 2012). This conceptual approach provides an array of opportunities for rapidly searching for response signals with targeted agents without committing to larger trials before a clinical response signal is observed (Subbiah & Kurzrock, 2011). Eventually, combining these agents and temporally integrating them with multimodal approaches using surgery and radiation therapy may benefit patients. Some of these potential options are discussed below.

Vascular Endothelial Growth Factor (VEGF)-based therapies

Beginning with the research of Volkman (Folkman, 1971) and Fidler, (Hart & Fidler, 1980; Hart et al., 1981) VEGF has been extensively implicated as an important factor in tumor biology. Specifically, in addition to angiogenesis, vasculogenesis has a putatively important role in the biology of Ewing’s sarcoma oncogenesis, tumor growth and development. (Stewart et al., 2011) (Huang et al., 2011a; Stewart & Kleinerman, 2011; Yu et al., 2010). Also, the EWS/FL1 chimeric fusion gene is known to upregulate VEGF-A in preclinical models(Nagano et al., 2009). Many preclinical studies have shown that targeting the VEGF pathway, either using an anti-VEGF antibody (bevacizumab) or siRNA against VEGF, suppresses tumor growth(DuBois et al., 2010). One caveat to bear in mind is that bevacizumab is directed against human isoforms of VEGF and not murine VEGF(DuBois et al., 2010). Clinical experience with bevacizumab as monotherapy demonstrated stable disease at best for at least 4 months in 3 out of 5 patients with Ewing’s sarcoma enrolled in a COG phase I clinical study(Glade Bender et al., 2008). Targeting this pathway, either using antibodies (i.e., bevacizumab) in combination with other therapies or the newer multikinase inhibitors (i.e., sunitinib, pazopanib, axitinib, cedarinib) that target VEGF may be worthwhile.

Multikinase (Pazopanib-based, Axitinib) studies

Pazopanib (Votrient) is a novel small molecule multikinase inhibitor recently approved by the United States Federal Drug Administration (FDA) for the treatment of advanced and metastatic renal cell carcinoma(Kasper & Hohenberger, 2011). By co-targeting multiple kinases, including VEGF receptor 1, 2 and 3, PDGF receptors a and b and c-kit, pazopanib has been shown to exert considerable antiangiogenic effects. A phase II clinical trial using pazopanib in advanced soft tissue sarcomas (excluding Ewing’s sarcoma) conducted by the Soft Tissue and Bone Sarcoma Group (STBSG) of the European Organization for Research and Treatment of Cancer (EORTC) met its primary endpoint of a progression-free survival rate of approximately 40% at 12 weeks(Sleijfer et al., 2009). This result prompted a worldwide phase III randomized study (EORTC 62072, PALETTE) comparing pazopanib with a placebo (randomization 2:1, with no cross-over). Preliminary results were positive and demonstrated a median progression-free survival of 13 weeks in pretreated metastatic soft tissue sarcoma patients(W. T. Van Der Graaf, 2011). Most recently, the U.S. FDA approved Votrient (pazopanib) to treat patients with advanced soft tissue sarcoma who have previously received chemotherapy(2012a).

Pazopanib is available in early-phase clinical trials in combination with mTOR inhibitors, histone decetylase inhibitors and some cytotoxic agents. Pazopanib may be worth investigating in patients with Ewing’s sarcoma who relapse following treatment with standard-of-care second-line agents(Kasper & Hohenberger, 2011). Additionally, pazopanib showed a favorable toxicity profile even in heavily-pretreated patients (Sleijfer et al., 2009).

Axitinib (Inlyta), similar to pazopanib, is another novel small molecule multikinase inhibitor recently approved by the FDA for the treatment of advanced and metastatic renal cell carcinoma(Ho & Jonasch, 2011). This drug has a range of activity similar to pazopanib, and investigating this molecule in Ewing’s sarcoma may, therefore, be worthwhile.

mTOR inhibitor-based studies

mTOR is a critical pathway in Ewing’s sarcoma; however, mTOR blockade alone may not be sufficient for response. Recently, a phase II trial using an mTOR inhibitor ridaforolimus (deforolimus; AP23573; MK-8669) was conducted in advanced soft tissue sarcomas(Chawla et al., 2012). The reported median progression-free survival was 15.3 weeks and median overall survival was 40 weeks(Chawla et al., 2012). Because the trial met its endpoint a subsequent phase III study was conducted--the Sarcoma Multi-Center Clinical Evaluation of the Efficacy of Ridaforolimus [SUCCEED] study and final results have not yet been published. However, the U.S. FDA Oncologic Drugs Advisory Committee voted unfavorably for this drug(2012b).

Several mTOR inhibitors such as rapamycin, everolimus, temsirolimus are FDA-approved for specific tumors such as kidney cancer: These drugs block mTORC1 but not mTORC2 assembly. Next-generation mTOR inhibitors that inhibit mTORC2 are in development. When available for clinical trials, these agents may be worthwhile evaluating in Ewing’s sarcoma patients. There is enough preclinical and clinical evidence to demonstrate mTOR activity in Ewing’s sarcoma(Martins et al., 2006; Naing et al., 2012; Subbiah et al., 2011b; Zenali et al., 2009). For now, mTOR inhibitor-based combination studies with chemotherapy and other targeted therapeutic agents may offer prospective options for relapsing patients(Vincenzi et al., 2011).

P13K/ AKT inhibitor-based studies

Given that IGF1R-based signaling occurs through the P13K/AKT pathway, agents targeting it may perhaps be plausible additions in treating Ewing’s sarcoma(Huang et al., 2011b). Interestingly, in preclinical models of Ewing’s sarcoma, simultaneous inhibition of mitogen-activated protein kinase and phosphatidylinositol 3-kinase pathways augment sensitivity to actinomycin D in Ewing sarcoma(Yamamoto et al., 2009).

Poly(ADP-ribose) polymerase (PARP) inhibitors

A recently published pharmacogenomic profiling study attempted to uncover biomarkers of sensitivity and resistance in 639 cell lines through screening 130 drugs in preclinical or clinical development (Garnett et al., 2012). Interestingly, EWS/FLI1 emerged as a poly(ADP-ribose) polymerase (PARP) inhibitor biomarker. Specifically, olaparib (AZD2281) had potent activity in cell lines harboring an EWS/FLI1 translocation (Figure 2) (Garnett et al., 2012). While we know that BRCA1 or BRCA2 mutations confer sensitivity to PARP inhibitors, no similar mutations have been identified in Ewing’s sarcoma. However, prostate cancer cell lines bearing ERG transcripts such as FLI1 are members of the ETS gene family and have demonstrated sensitivity to PARP inhibitors(Garnett et al., 2012). Another study in preclinical models showed synergy when combining PARP inhibitors and temozolomide in Ewing’s sarcoma(Brenner et al., 2012). In fact, a clinical trial testing these agents is currently ongoing (Clinical trials.gov ID no: {"type":"clinical-trial","attrs":{"text":"NCT01294735","term_id":"NCT01294735"}}NCT01294735) and if this rapid proof-of-principle study shows promise, the combination could be expanded to enroll additional patients.

Figure 2

Simplified Insulin-like growth factor 1 receptor (IGF1R) and mTORC1 and mTORC2 signaling [Modified from (Subbiah et al., 2011a; Subbiah et al., 2011b)]

Insulin receptor substrate 1 (IRS1) and phosphatidylinositol 3-kinases (PI3K) are activated by insulin and/or IGF1 signaling at the insulin receptor level. PDK1 and Akt are recruited to the plasma membrane by products of PI3K, PIP2, PIP3 (phosphatidylinositol 3,4,5 trisphosphate and phosphatidylinositol 3,4 bisphosphate). After this event there is phosphorylation and activation of Akt by the mTORC2 complex (mTOR + mLST8+ Rictor). This leads to a chain of activation of numerous targets by Akt. The TSC1/2 complex is also phosphorylated in this fashion. By this mechanism inactivation of TSC2’s GAP activity for the small G protein Rheb is initiated. Subsequently, the new mTORC1 complex (mTOR + mLST8+Raptor) is activated by GTP-bound Rheb and phosphorylates proteins like S6K. This starts a negative feedback loop to modulate auto-activity, through a S6K-mediated pathway decrease in the activation of PI3K. IGF1R antibody inhibits IGF1R signaling and rapalogs inhibit mTORC1 short term, and eventually inhibit mTORC 2 with chronic exposure and also suppress EWS/FLI1 which drives tumorigenesis in Ewing’s sarcoma.

CONCLUSIONS

There has been a wealth of recent advances in understanding the biology of Ewing’s sarcoma, an aggressive disease that afflicts children, adolescents, and young adults in the prime of their lives. Given the dramatic responses in clinical trials of IGF1R antibodies, it was hoped that IGF1R-based therapy would be the winning strategy that could be exploited to surmount the therapeutic plateau in patients with Ewing’s sarcoma. However, these therapies benefit only a minority of patients, and even those patients who respond inevitably develop resistance. In order to overcome resistance, protocols combining IGF1R and mTOR inhibitors have been designed, and responses have been observed, even in patients whose tumors progressed after IGF1R inhibitor therapy alone. In order to increase response rates, three strategies need to be adopted: (i) incorporating additional drugs such as MEK inhibitors into trials of IGF1R and mTOR inhibitors, since preclinical data suggest that the ERK/MEK pathway may mediate secondary resistance; (ii) analyzing tumors in depth with the use of advanced molecular technologies such as exomic or full genomic sequencing so that response and resistance signatures can be identified and, hence, patients most likely to respond will be selected for future trials; and (iii) integrating these targeted agents with chemotherapy in the newly diagnosed setting where multiple aberrations promoting resistance are less likely to exist. Additional molecules that warrant testing in the clinic include multi-kinase inhibitors, antagonists of EWS-FLI itself, and PARP inhibitors. Most importantly, it is critical that molecules that show responses, such as the IGF1-R inhibitors, not be abandoned because of the rarity of Ewing’s sarcoma and the fact that only a subset of patients respond. Ultimately, in the era of personalized cancer therapy, strategies to develop drugs for uncommon tumors such as Ewing’s sarcoma and to individualize therapy will be crucial.

Acknowledgments