Targeting IGF-1R in the treatment of sarcomas: past, present and future

ARTICLE

Auteur(s) : Su Young Kim, Xiaolin Wan, Lee J Helman

Pediatric Oncology Branch, National Cancer Institute, National Institutes of Health, 10, Center Drive, CRC 1w-3750, Bethesda, Maryland, MA 20892, USA

Article reçu le 22 Decembre 2008, accepté le 9 Juin 2009

Background

Early epidemiological studies suggested the connection between high-levels of IGF and increased incidence of many carcinomas and sarcomas. The converse also appears to be true, in that patients with congenital IGF-I deficiency do not develop cancer [3]. Various transgenic murine models have confirmed these findings [4]. However, there is an important caveat in interpreting these results. In mice, IGF-I levels are low in the prenatal period but high in adult mice [2]. In contrast, IGF-II levels are very high before birth and almost absent thereafter. In comparison, in human adults, both IGF-I and IGF-II are expressed, with the latter expressed at slightly higher levels. Therefore, preclinical murine studies should be interpreted with the knowledge that these experiments do not account for possible effects of IGF-II.

IGF-I and IGF-II constitute the main ligands that are involved in IGF signalling (figure 1). Other factors that influence IGF ligand levels are six distinct IGF-binding proteins (IGFBP) [5]. The majority of IGF exists as a complex in the circulation, bound to IGFBP-3 and a molecule called acid labile subunit, which attenuates transport of the complex out of the vasculature (figure 1). IGFBPs have higher affinity for IGFs than their cognate receptors. In general, IGFBPs inhibit the action of IGFs by sequestering them from binding to their receptors. However, other experimental models suggest that IGFBPs can potentiate IGF action in certain systems, possibly by increasing their half-life in circulation [5]. Another possibility is that IGFBPs bring IGF closer to their receptors by binding to proteglycans, thereby enhancing IGF action. Recent studies showing that free IGFBPs can bind to a host of macromolecules, including extracellular matrix components and cell membrane associated proteins, has added to the growing list of IGF-independent physiological functions of IGFBPs [5].

Six combinations of receptors contribute to IGF signalling: IGF receptors type 1 and type 2 (IGF-1R and IGF-2R), insulin receptors type A and type B (IR-A and IR-B) and two hybrid receptors (IGF-1R/IR-A and IGF-1R/IR-B) (figure 1) [2]. The data suggesting that both insulin receptors and hybrid receptors are implicated in tumorigenesis has been reviewed previously [6]. IGF-2R binds IGF-II preferentially, but the lack of an intracellular signalling domain in IGF-2R precludes downstream signalling. Binding of ligand to IGF-2R leads to endocytosis of the complex and targeted degradation of both the receptor and the ligand, resulting in modulation of extracellular IGF-II concentrations. IGF-1R is therefore the main receptor that is involved in IGF signalling and it will be the focus of this review.

IGF-1R is a member of the receptor tyrosine-kinase family. For most members of this family, binding of ligand leads to dimerization of the receptor. In contrast, both IGF-1R and insulin receptor exist as dimeric structures [2]. Binding of IGF-I or IGF-II to IGF-1R is followed by phosphorylation of key tyrosine residues and activation of the IGF-1R-kinase domain (figure 1) [2]. In addition, tyrosine phosphorylation in other regions of the receptor allows for recruitment of adaptor proteins that contain SH2 domains, which recognize and bind phosphorylated tyrosine residues [7]. Insulin receptor substrate (IRS) family members are key contributors to IGF signalling, by acting as adaptor proteins. The combination of tyrosine-kinase domain activation and recruitment of adaptor proteins results in a cascade of downstream signalling.

Two of the more important pathways that mediate IGF signalling are RAS/RAF/MAPK and PI3K/AKT/mTOR [7]. Although the sequence of signalling events is complex, the end result is cell proliferation and inhibition of apoptosis. In normal cells, these pro-growth signals are tightly regulated, resulting in growth, only when required. However, in many sarcomas, IGF signalling is activated, either by loss of genomic imprinting, or by abnormal stimulation by endocrine, paracrine and autocrine mechanisms. This results in abnormal growth and contributes to the subsequent formation of sarcomas.

Preclinical highlights

ESFT express high-levels of IGF-1R [8]. In addition, examination of cell lines and tumors reveal high-levels of IGF-I and IGF-II in many samples. These growth factors are secreted and have the potential to bind back to IGF-1R, resulting in autocrine growth stimulation [8]. In other tumor types, such as neuroblastoma, this effect may be enhanced by secretion of IGF-II from stromal cells that surround the tumor [9]. In these cases, paracrine stimulation may potentiate autocrine effects. The contributions of paracrine, autocrine and endocrine effects vary in different tumor types. However, all of the sarcomas mentioned previously have abnormalities in at least one, if not all three of these routes, ultimately increasing IGF ligand levels.

Loss of imprinting is another mechanism by which the IGF pathway is activated. Precise expression of genes is important at many levels: quantitatively, spatially and temporally. In almost all tissues, only the paternally derived allele of the IGF-II gene is expressed due to imprinting of the maternal allele [10]. Beckwith-Wiedemann syndrome is the best example of how high-expression of IGF-II due to loss of imprinting results in marked phenotypic abnormalities is characterized by fetal and neonatal overgrowth. One very common abnormality in cancers is loss of imprinting. Specifically, this has been found in a host of pediatric tumors including neuroblastoma, rhabdomyosarcoma, synovial sarcoma and Wilms’ tumor. In these tumors, loss of imprinting results in bi-allelic expression of IGF-II, and subsequent higher levels of circulating IGF-II [10]. Loss of imprinting of genes in the IGF pathway thereby provides another mechanism by which these cells are subjected to higher levels of IGF ligands. This is slightly more complicated in ESFT. Approximately, 30% of ESFT samples demonstrate bi-allelic IGF-II expression, but this does not result in higher levels of IGF-II transcription [11]. These results suggest that in ESFT there are other IGF-II independent genetic or epigenetic abnormalities that are driving tumorigenesis.

Modulation of IGFBP-3 may also play an important role in regulation of the IGF signalling pathway. EWS/FLI-1, which is the most common translocation fusion product in patients with ESFT, directly binds the IGFBP-3 promoter and represses its activation, subsequently resulting in decreased expression of IGFBP-3 [12]. Measurement of serum IGFBP-3 and IGF-I levels in a small number of patients with ESFT has shown a trend towards increased survival in those with a high ratio of IGFBP-3 compared to IGF-I [13]. These findings suggest the possibility that EWS/FLI-1 mediated down regulation of IGFBP-3 may result in decreased serum levels of IGFBP-3, thus allowing more unbound IGF-I ligand to exit the circulation and interact with IGF-1R. To our knowledge this hypothesis has not been tested.

Studies in the pediatric kidney neoplasm Wilms’ tumor demonstrate a different mechanism of IGF modulation in tumors. IGF-1R mRNA levels are six times higher in Wilms’ tumor samples than in adjacent normal kidney tissue [14]. An interesting finding is that high levels of IGF-1R are inversely correlated with levels of WT1. The latter is a tumor suppressor gene, inactivation of which is an important step in the development of some Wilms’ tumors. Studies have demonstrated that introduction of exogenous WT1 into cells that lack WT1 results in a significant decrease in IGF-1R mRNA levels [14]. These results strongly suggest that WT1 represses IGF-1R gene expression, providing another regulatory mechanism that affects IGF family members. Synovial sarcomas activate the IGF pathway slightly differently. These tumors are characterized by a specific chromosomal translocation that brings together the genes SYT and SSX1 or SSX2. Knockdown of the fusion protein in synovial sarcoma cells leads to deceased IGF-II levels, whereas forced overexpression of the fusion product results in increased IGF-II levels [15]. The presence of the pathognomonic fusion protein in these tumors results in IGF-II activation.

As described above, activation of the IGF signalling pathway can occur by many different mechanisms. However, the repercussions following its activation are concordant with what is expected. Levels of phosphorylated PI3K (phosphoinositol-3-kinase), MAPK (mitogen-activated protein-kinase) and AKT, are all increased, which agrees with results showing that these genes play an important role in cell growth and survival. Furthermore, treatment of ESFT cells with the PI3K inhibitor LY294002 impairs proliferation and cell cycle progression by inducing G1 arrest [16]. The addition of exogenous IGF is able to reverse this impaired proliferative ability. This is one of many findings that demonstrate the existence of multiple signalling pathways (both PI3K-dependent and PI3K-independent) that are affected by IGF.

Murine studies provide further evidence that ESFT is reliant upon the IGF pathway. Early studies utilized αIR3, a murine IGF-1R neutralizing monoclonal antibody. Treatment of mice with αIR3 following subcutaneous injection of ESFT cells showed that 56% of treated mice were tumor free, whereas all of the untreated mice developed tumors [17]. In addition, the tumors that formed in the treated mice were twice as small as those in the control group. These exciting findings were recently extended with the use of humanized and fully human IGF-1R antibodies. One of these antibodies was used to evaluate activity in a series of pediatric tumor xenografts [18]. Mice treated with the SCH717454 IGF-1R monoclonal antibody had an increase in event free survival in 20 of 35 solid tumor xenografts, which included regression in ESFT and osteosarcoma models. Further support of the effectiveness of targeting IGF-1R comes from studies showing that small molecule inhibitors of the IGF-1R-kinase domain also reduce the size of EFST xenografts [19]. In total, these studies demonstrate that one component of ESFT tumor formation is activation of the IGF pathway, a phenotype that can be abrogated by blockade of the IGF-1R receptor.

Recent studies suggest that therapy that targets IGF-1R increases the susceptibility of ESFT cells to chemotherapy agents such as doxorubicin and vincristine [19]. Both of these agents are mainstays in sarcoma chemotherapy regimens. Combination treatment, using chemotherapy and IGF-1R-kinase domain inhibitors, potentiated reductions in proliferation and motility and increases in apoptosis, in patterns that were dependent on the levels of IGF-1R in these cells [19]. The combined use of chemotherapy and IGF-1R antibodies in murine models has not been reported yet, but it would be interesting to determine if the synergistic effects seen in vitro are also seen in vivo.

The mTOR signalling pathway plays a central role in the regulation of cancer cell growth by controlling mRNA transcription and protein translation [20]. It is aberrantly activated in many human cancers and as a result, the mTOR pathway has been an important target for cancer drug development. There are several mTOR inhibitors, including rapamycin and its derivatives, CCI-779, RAD001 and AP23573. To date, clinical results have demonstrated tolerability and therapeutic efficacy in some patients. However, cancer cells often develop resistance to treatment with mTOR inhibitors, showing paradoxical hyperphosphorylation of AKT mediated through IRS-1, suggesting that feedback activation of this pathway may be contributing to treatment resistance [21]. This is also true in rhabdomyosarcoma, but siRNA mediated suppression of IRS-1 does not alleviate AKT hyperphosphorylation following sustained treatment with CCI-779, suggesting an IRS-1 independent mechanism [22]. However, combined treatment using an IGF-1R antibody and an mTOR inhibitor results in abrogation of rapamycin-induced feedback activation of AKT, and enhances the effect of rapamycin on growth inhibition [22]. Moreover, combining treatment with IGF-1R antibody and rapamycin results in a marked inhibition of rhabdomyosarcoma tumor growth in mice, as compared to mice given either agent alone [23]. These results suggest that combining IGF-1R antibody therapy with mTOR inhibitor treatment may overcome the problem of development of resistance with mTOR inhibitors alone.

GIST is receiving increased attention as a tumor that may prove to be sensitive to IGF pathway targeted therapy. Most adults with GIST have activating mutations in the cell surface tyrosine-kinase receptors KIT or PDGFRA [24]. Twenty years ago, prolonged survival for these patients was less than 25%. Recently, treatment of patients with imatinib and sunitinib (tyrosine-kinase inhibitors that target KIT and PDGFRA) has resulted in a marked increase in overall and progression free survival [24].

A vast majority of younger patients with GIST do not have any detectable molecular abnormalities (mutations, deletions, duplications) in KIT or PDGFRA, and are therefore termed wild type GIST. Surprisingly though, patients with wild type GIST have high-levels of KIT, even higher than that of patients with mutated (activated) KIT [25]. Despite this fact, patients with wild type GIST have only limited susceptibility to tyrosine-kinase inhibitors. GIST, both KIT mutated and wild type forms, is unresponsive to chemotherapy [24]. In addition, there is a high-rate of recurrence following complete surgical resection [24]. All of these factors point to the need for alternative therapies.

Almost all samples from wild type GIST tumors have markedly higher levels of IGF-1R compared to those from KIT mutated GIST tumors [26]. One explanation for this increase is amplification of the IGF-1R locus in some cases [27]. Inhibition of IGF-1R, using either IGF-1R-kinase domain inhibitors or siRNA directed to IGF-1R, results in cytotoxicity in GIST cell lines [27]. It is important to note that these results were obtained in KIT mutated cell lines, due to the fact that there are no cell lines or xenografts from wild type GIST tumors. However, another factor that may increase the therapeutic role of IGF-1R antibodies is the finding that many patients with GIST, either wild type or KIT mutated, have high IGF-1R levels [27]. In addition, patients with increased levels of IGF-I or IGF-II have significantly worse disease free survival [28]. As is the case in ESFT and the other sarcomas mentioned above, these findings support the concept that GIST tumors have activation of the IGF pathway and that they may be susceptible to its blockade.

Current clinical trials

A number of phases I and II trials are actively recruiting patients. A list of studies in the United States can be accessed on-line {http://www.clinicaltrials.gov}. A search using the terms “IGF-1R antibody” reveals 18 citations for current trials. However, a more refined search using the specific name of the compounds reveals the status of a host of additional trials [29]. {AMG-479, AVE-1642, BIIB-022, CP-751-871, IMC-A12, MK-0646, R-1507, SCH-717454}. Similar data are available for the small molecule IGF-1R-kinase inhibitors [29].

The cumulative results from phase I studies have shown that the response to treatment with IGF-1R antibody has varied. Not surprisingly, the majority of patients have progressed. There have also been several cases of patients who have had marked initial response to therapy, followed by progression. Some patients have had slowing of the rate of progression. There have also been several patients who continue to have sustained regression of tumors. To date, most clinical responses have been seen in patients with ESFT. These responses have occurred despite the fact that these tumors have proven refractory to primary and salvage therapies and are also rapidly growing tumors. The fact that ESFT appears to be exquisitely sensitive to IGF-1R blockade has sustained the initial excitement of the use of IGF-1R antibodies as a therapeutic agent. It has also generated many concepts for future clinical trials on how to best expand its use, in efforts to potentiate its activity and to maximize its efficacy in cases where there is some initial response.

In the preclinical section, we described in vitro and murine studies demonstrating that treatment with IGF-1R antibody increases tumor susceptibility to chemotherapy. These findings suggest that additive therapy using IGF-1R antibody may improve tumor response by combining antiproliferative and pro-apoptotic mechanisms with the tumor toxic effects of chemotherapy. The current salvage regimens that are most effective in ESFT include the use of cyclophosphamide and topotecan or temozolomide and irinotecan. Addition of IGF-1R to either regimen has the potential to improve response rates, and also to transform initial responses into sustained responses. A more ambitious, but much more difficult study, would be to assess the addition of IGF-1R to standard front-line five-drug chemotherapy. In order to avoid the possibility that unforeseeable side effects with additive therapy may delay the administration of chemotherapy, it is likely that such an ambitious study would begin as a feasibility study by enrolling newly diagnosed ESFT patients who have poor prognostic features. Since we do not currently have the ability to identify patients who are expected to respond to IGF pathway directed therapy, the sensible alternative is to eliminate patients who we expect to have a good chance of cure with standard therapy alone.

The preclinical evidence that IGF-1R antibody potentiates the antitumor effects of rapamycin and suppresses AKT feedback activation, suggests that blockade aimed at the AKT pathway should be pursued. This can be performed in many ways, beginning with combination therapy using IGF-1R antibody and rapamycin. Alternatively, studies could employ one of the rapamycin analogues (CCI-779, RAD001 and AP23573), since the doses and the timing of administration have been tested in patients with sarcomas. As additional data on the safety of AKT-specific inhibitors and PI3K inhibitors become available, combination therapy using these agents can also be explored.

If implemented, all of these trials will have to be undertaken by large cooperative groups. Due to the relative rarity of pediatric sarcomas, this is the only mechanism by which a sufficient number of patients can be enrolled in a reasonably short-time frame. As results from phase II studies for specific pediatric sarcomas emerge, it will be evident if patients with tumor types other than EFST are responsive to IGF-1R antibody therapy. If other responsive tumor types are identified, additional innovative clinical trials are destined to emerge.

Potentiating success

• – one. There are a percentage of cells that are susceptible to IGF-1R blockade and they undergo cell death, but there are also some cells that are intrinsically resistant, and they continue to grow. As the number of susceptible cells decreases, the number of resistant cells increases, slowly changing the balance from sustained equilibrium to slow proliferation;
• – two. The cells in the tumor are susceptible to IGF-1R blockade. However, they quickly develop acquired resistance to the antibody by co-opting other pathways, allowing cells to resume unregulated growth. Any number of cell surface receptors and alternative pathways can be implicated, such as EGFR, MET and PDGFR. Data from breast and prostate cancer cell lines suggest that a similar mechanism exists. Specifically, resistance to EGFR inhibition may be a function of activation of the IGF pathway [30].

Even when considering just these two possibilities, the clinical ramifications are quite different. If the former is correct, then continued treatment with IGF-1R antibody in the face of progressive disease is not warranted. If the latter is correct, then treatment should focus on continued treatment with IGF-1R antibody, and augmentation with another agent that targets the co-opted pathway. Conventional practice dictates that patients discontinue use of an experimental agent at the time of progression. However, it is possible that if alternative growth promoting pathways are responsible for IGF-1R resistance, then substitution therapy may be less effective than additive therapy. Hypothetically, if the co-opted pathway involved EGFR, then changing from IGF-1R antibody to EFGR antibody therapy may prove to be effective. However, it is also possible to imagine a scenario in which the therapeutic effect of IGF blockade is being masked by supplemented growth from the alternative pathway. In this case, the possibility then exists that discontinuation of IGF-1R therapy would allow the cell to revert back to utilizing the IGF pathway for continued growth. If this is indeed the case, then continuing IGF-1R antibody in the face of progression, while adding EGFR directed therapy, may lead to a dramatic secondary clinical response. Currently, we can only guess as to which of the two possibilities is correct. However, animal models provide the means by which we can begin to answer this question.

Shown in figure 2 are luminescent images of mice, after injection of a rhabdomyosarcoma cell line into the gastrocnemius muscle (figure 2). One cohort was treated twice a week with IGF-1R antibody. The other cohort was treated with a vehicle control. Control mice had rapid growth of primary tumors. In comparison, mice treated with IGF-1R antibody had a marked period of no tumor growth. Unfortunately, for reasons that are currently unknown, this was followed by rapid tumor formation that occurred at the same rate as the control mice. Although the time to tumor formation and the latency period with treatment varied with different rhabdomyosarcoma cell lines, the general trend remained consistent, represented quantitatively in figure 2 (quantitative values are shown to demonstrate trend and are not actual values from any single experiment). How? And why? How does IGF-1R antibody keep primary tumor formation in check for a prolonged period of time? And why does this anti-tumor activity then become ineffective?

In this case, we are fortunate that there is an animal model that closely recapitulates the clinical course of prolonged latency. The ability to obtain tumor samples from different time points provides a wealth of reagents that is currently not available through clinical trials. The large number of mice in each cohort allows for their sacrifice at different time points. Specifically, tumor samples can be harvested from representative mice prior to the start of treatment, and then at weekly intervals, up to the time of high tumor burden. Tumors can then be analyzed by a number of methods, including DNA analysis to detect mutations and copy number changes, RNA microarray to assess transcriptional differences, proteomics to determine differences in proteins or activated phospho-proteins. In addition, when candidate genes are suspected, samples can be subjected to immunohistochemistry to identify any changes.

A second group of patients who have the potential to provide a wealth of knowledge on tumors that respond to IGF-1R directed therapy are those who have an extraordinary clinical response, but then progress. Shown in figure 3 are CT-scans of one such patient with metastatic and recurrent ESFT who was treated with an IGF-1R antibody. The CT-scan prior to treatment revealed a 9 × 4 cm pericardial mass and multiple smaller lung nodules (figure 3). Restaging scans at week 6, following weekly therapy with the antibody, demonstrated a dramatic reduction in the size of the pericardial mass, now 4 × 1 cm. Unfortunately, further imaging at week 12 demonstrated regrowth of the tumor. Again, how and why does such a dramatic clinical response, cease?

The simplest explanation for this finding is that the tumor is a heterogeneous mass, composed of both IGF-1R antibody sensitive and resistant cells. In the case above, 90% of the tumor was composed of cells that were sensitive to blockade of the IGF pathway. As these cells underwent apoptosis during treatment, the small minority of cells that were resistant to IGF pathway blockade then continued to grow unimpeded. This would explain the radiographic findings. But only rarely is biology this simple. And currently, animal models that mimic the above clinical scenario do not exist. The ideal model would begin with a mouse that has multiple gross metastatic lung nodules as determined by non-invasive luminescent or fluorescent imaging. These mice would then undergo treatment with IGF-1R antibody. Those that demonstrate tumor shrinkage would continue to be treated and monitored through the period of progression. Examination of samples from pretreated mice, those at maximal therapy response, and following progression would be the most efficient method to determine the genetic components that mediate IGF resistance.

We have purposely emphasized the need for tumor samples as reagents for future studies. Murine models represent valuable surrogates, but they will never supplant the value of actual patient samples. In an ideal research environment, we would have access to the tumor at all stages of treatment: pretherapy, at the time of clinical response and at progression. Clearly, it is unreasonable and unethical to expect to obtain all of these samples, since biopsy involves high-risk with little benefit at this point, and especially since it would not change the patient’s medical management. However, there are many opportunities to obtain samples during clinically indicated procedures. For example, metastatic tumors are many times associated with malignant pleural or pericardial effusions that compromise respiratory or cardiac function. Pleurocentesis and pericardiocentesis are both highly therapeutic in these cases. Cytological analysis provides verification of the malignant nature of these effusions. Obviously, tumor tissue cannot be obtained in these cases. However, malignant effusions provide the perfect milieu from which tumor cells can be cultured into cell lines. We have recently established a cell line following culture of cells obtained after pleurocentesis from a patient with rhabdomyosarcoma who progressed while receiving IGF-1R antibody. This sample has the potential to answer many basic science questions pertaining to IGF-1R resistance. Other occasions when samples become available include tumor resection for pain palliation or neurovascular compromise and biopsies to differentiate tumor from infection. It is imperative that clinicians be able to collect and distribute these samples to researchers, to allow laboratory studies to progress rapidly.

Conclusion

References

2 LeRoith D, Roberts CT. The insulin-like growth factor system and cancer. Cancer Lett 2003; 195: 127-37.

3 Shevah O, Laron Z. Patients with congenital deficiency of IGF-I seem protected from the development of malignancies: a preliminary report. Growth Horm IGF Res 2007; 17: 54-7.

4 Wu Y, Cui K, Miyoshi K, Hennighausen L, Green JE, Setser J, et al. Reduced circulating insulin-like growth factor I levels delay the onset of chemically and genetically induced mammary tumors. Cancer Res 2003; 63: 4384-8.

5 Bach LA, Headey SJ, Norton RS. IGF-binding proteins: the pieces are falling into place. Trends Endocrinol Metab 2005; 16: 228-34.

6 Belfiore A. The role of insulin receptor isoforms and hybrid insulin/IGF-1 receptors in human cancer. Curr Pharm Des 2007; 13: 671-86.

7 Butler AA, Yakar S, Gewolb IH, Karas M, Okubo Y, LeRoith D. Insulin-like growth factor-I receptor signal transduction: at the interface between physiology and cell biology. Comp Biochem Physiol B Biochem Mol Biol 1998; 121: 19-26.

8 Scotlandi K, Benini S, Sarti M, Serra M, Lollini PL, Maurici D, et al. Insulin-like growth factor I receptor-mediated circuit in Ewing’s sarcoma/peripheral neuroectodermal tumor: a possible therapeutic target. Cancer Res 1996; 56: 4570-4.

9 El-Badry OM, Helman LJ, Chatten J, Steinberg SM, Evans AE, Israel MA. Insulin-like growth factor II-mediated proliferation of human neuroblastoma. J Clin Invest 1991; 87: 648-57.

10 Chao W, D’Amore PA. IGF2: epigenetic regulation and role in development and disease. Cytokine Growth Factor Rev 2008; 19: 111-20.

11 Zhan S, Shapiro DN, Helman LJ. Loss of imprinting of IGF2 in Ewing’s sarcoma. Oncogene 1995; 11: 2503-7.

12 Prieur A, Tirode F, Cohen P, Delattre O. EWS/FLI-1 silencing and gene profiling of Ewing cells reveal downstream oncogenic pathways and a crucial role for repression of insulin-like growth factor binding protein 3. Mol Cell Biol 2004; 24: 7275-83.

13 Toretsky JA, Steinberg SM, Thakar M, Counts D, Pironis B, Parente C, et al. Insulin-like growth factor type I (IGF-1) and IGF binding protein-3 in patients with Ewing sarcoma family of tumors. Cancer 2001; 92: 2941-7.

14 Werner H, Shen-Orr Z, Rauscher FJ, Morris JF, Roberts CT, LeRoith D. Inhibition of cellular proliferation by the Wilms’ tumor suppressor WT1 is associated with suppression of insulin-like growth factor I receptor gene expression. Mol Cell Biol 1995; 15: 3516-22.

15 Sun Y, Gao D, Liu Y, Huang J, Lessnick S, Tanaka S. IGF2 is critical for tumorigenesis by synovial sarcoma oncoprotein SYT-SSX1. Oncogene 2006; 25: 1042-52.

16 Benini S, Manara MC, Cerisano V, Perdichizzi S, Strammiello R, Serra M, et al. Contribution of MEK/MAPK and PI3-K signalling pathway to the malignant behavior of Ewing’s sarcoma cells: therapeutic prospects. Int J Cancer 2004; 108: 358-66.

17 Scotlandi K, Benini S, Nanni P, Lollini PL, Nicoletti G, Landuzzi L, et al. Blockage of insulin-like growth factor-I receptor inhibits the growth of Ewing’s sarcoma in athymic mice. Cancer Res 1998; 58: 4127-31.

18 Kolb EA, Gorlick R, Houghton PJ, Morton CL, Lock R, Carol H, et al. Initial testing (stage 1) of a monoclonal antibody (SCH 717454) against the IGF-1 receptor by the pediatric preclinical testing program. Pediatr Blood Cancer 2008; 50: 1190-7.

19 Martins AS, Mackintosh C, Martin DH, Campos M, Hernandez T, Ordonez JL, et al. Insulin-like growth factor I receptor pathway inhibition by ADW742, alone or in combination with imatinib, doxorubicin, or vincristine, is a novel therapeutic approach in Ewing tumor. Clin Cancer Res 2006; 12: 3532-40.

20 Wan X, Helman LJ. The biology behind mTOR inhibition in sarcoma. Oncologist 2007; 12: 1007-18.

21 Haruta T, Uno T, Kawahara J, Takano A, Egawa K, Sharma PM, et al. A rapamycin-sensitive pathway down-regulates insulin signalling via phosphorylation and proteasomal degradation of insulin receptor substrate-1. Mol Endocrinol 2000; 14: 783-94.

22 Wan X, Harkavy B, Shen N, Grohar P, Helman LJ. Rapamycin induces feedback activation of AKT signalling through an IGF-1R-dependent mechanism. Oncogene 2007; 26: 1932-40.

23 Cao L, Yu Y, Darko I, Currier D, Mayeenuddin LH, Wan X, et al. Addiction to elevated insulin-like growth factor 1 receptor and initial modulation of the AKT pathway define the responsiveness of rhabdomyosarcoma to the targeting antibody. Cancer Res 2008; 68: 8039-48.

24 Demetri GD, Benjamin RS, Blanke CD, Blay JY, Casali P, Choi H, et al. NCCN task force report: management of patients with gastrointestinal stromal tumor (GIST) — update of the NCCN clinical practice guidelines. J Natl Compr Canc Netw 2007; 5: s1-s31.

25 Janeway KA, Liegl B, Harlow A, Le C, Perez-Atayde A, Kozakewich H, et al. Pediatric KIT-wild type and platelet-derived growth factor a wild type gastrointestinal stromal tumors share KIT activation but not mechanisms of genetic progression with adult gastrointestinal stromal tumors. Cancer Res 2007; 67: 9084-8.

26 Agaram NP, Laquaglia MP, Ustun B, Guo T, Wong GC, Socci ND, et al. Molecular characterization of pediatric gastrointestinal stromal tumors. Clin Cancer Res 2008; 14: 3204-15.

27 Tarn C, Rink L, Merkel E, Flieder D, Pathak H, Koumbi D, et al. Insulin-like growth factor 1 receptor is a potential therapeutic target for gastrointestinal stromal tumors. Proc Natl Acad Sci USA 2008; 105: 8387-92.

28 Braconi C, Bracci R, Bearzi I, Bianchi F, Sabato S, Mandolesi A, et al. Insulin-like growth factor (IGF) 1 and 2 help to predict disease outcome in GIST patients. Ann Oncol 2008; 19: 1293-8.

29 Feng Y, Dimitrov DS. Monoclonal antibodies against components of the IGF system for cancer treatment. Curr Opin Drug Discov Dev 2008; 11: 178-85.

30 Jones HE, Goddard L, Gee JMW, Hiscox S, Rubini M, Barrow D, et al. Insulin-like growth factor-I receptor signalling and acquired resistance to gefitinib (ZD1839; Iressa^®) in human breast and prostate cancer cells. Endocr Relat Cancer 2004; 11: 793-814.