Background
Chondrosarcoma is the second most common primary bone malignancy in humans [
1] and represents a heterogeneous collection of cartilage forming tumours with different outcomes depending on subtype and histological grade. Conventional central chondrosarcoma, arising centrally in the medulla of bone, accounts for ~85 % of the cases and can be histologically divided into 3 grades [
1]. Sixty-one percent of these tumours are classified as atypical cartilage tumour (ACT) (previously known as grade I), for which first line treatment is curettage with local adjuvant treatment, resulting in a 5 year survival rate of 83 %. Grade II (36 %) and grade III (3 %) tumours are more prone to metastasize and have a combined 5 year survival rate of 53 % [
1‐
3]. These tumours are treated with en bloc resection. Dedifferentiated chondrosarcoma is a highly malignant variant with an overall survival rate of 7 ~ 24 % [
4]. Mesenchymal chondrosarcoma is a rare aggressive subtype, in which distant metastasis can be identified even after 20 years [
5]. It has a 10 year survival rate between 44 and 54 % [
6,
7]. Chondrosarcoma patients with unresectable disease, due to tumour location, tumour size or extensive metastatic disease, have a 5 year survival of only 2 % [
8]. Although chondrosarcoma is known for its resistance to chemo- and radiotherapy, it was recently described that patients with inoperable disease treated with doxorubicin-based chemotherapy have a 3 year survival rate of 26 % versus 8 % in patients who did not receive systemic treatment [
8] and chemotherapy sensitivity differed between the chondrosarcoma subtypes [
9]. However, it is clear that overall efficacy of chemotherapy is limited. So far, the discovered genetic alterations and pathways involved in chondrosarcoma (reviewed in [
10] and [
11]) have not resulted in new treatment regimes
. Therefore, further unravelling of critical signalling pathways in chondrosarcoma is essential to identify new therapeutic targets.
One pathway which has been implicated in chondrosarcoma proliferation is the IGF pathway. The IGF pathway has two closely related ligands: IGF1 and IGF2 [
12]. When a ligand binds to the IGF1 receptor (IGF1R), this receptor forms homodimers or hybrid receptors with the insulin receptor (IR). The resulting autophosphorylation of the receptor recruits the insulin receptor substrate (IRS) to the membrane causing subsequent downstream activation of the PI3K/Akt/mTOR pathway and the Ras/Raf/MEK signalling pathway, which are known to be driver pathways in cancer [
12]. IGF2R functions to decrease the availability of IGF2 to IGF1R [
12].
IGF1R can be the upstream receptor that is responsible for the well known activation of the PI3K/Akt/mTOR pathway, the Src-pathway and the Ras/Raf/MEK pathway in (a subset of) chondrosarcoma cell lines and primary cultures [
13‐
17]. In a heterogeneous group of sarcoma patients, a combination of an IGF1R antibody and mTOR inhibitor has been shown to have clinical activity but the level of IGF1R expression was not predictive for response [
18]. Takigawa et al. demonstrated that cells of a clonal human chondrosarcoma-derived chondrocyte cell line produce IGF ligands and express IGF1R and IGF2R [
19]. Seong et al. and Matsumari et al. described that IGF1 increases cell proliferation in a Swarm-rat chondrosarcoma model [
20,
21]. Interestingly, Ho et al. described that IGF binding protein 3 (IGFBP3), which binds the IGF ligands thereby inhibiting their interaction with the IGF receptors, decreases with increasing histological grade of chondrosarcoma [
22]. In addition, Wu et al. demonstrated that IGF1 induced migration of chondrosarcoma cell lines which could be blocked by an IGF1R antibody [
23]. Recently, functional profiling of receptor tyrosine kinases in chondrosarcomas revealed active IGF1R signalling in one out of five chondrosarcoma cell lines [
13].
These above mentioned studies illustrate the need for a better understanding of the role of IGF1R signalling in chondrosarcoma to determine if it is a convincing target for therapy. Because chondrosarcoma is a very heterogenous disease, it is possible that the IGF1R directed therapy is only effective in a subset of patients. Hence, we used our large chondrosarcoma cell line panel, including three grade 2 and three grade 3 conventional chondrosarcomas, three dedifferentiated chondrosarcomas and one mesenchymal chondrosarcoma cell line. We analyzed expression levels of IGF1R and other important mediators of IGF1R signalling and determined the effect of IGF1R inhibitors. Our results indicate that the IGF pathway is not important for chondrosarcoma growth as IGF1R inhibition did not demonstrably impact chondrosarcoma cell line proliferation, migration and chemoresistance. In addition, IGF1R expression is low/absent in chondrosarcoma primary tumours in contrast to chondrosarcoma cell lines. This illustrates that there is limited preclinical rationale for using IGF1R inhibitors for the treatment of chondrosarcoma of bone.
Discussion
The aim of this study was to investigate whether the IGF pathway is a suitable target for therapy in chondrosarcoma. Heterogeneous expression of IGF1R, IR, IGF2R, IGF1, IGF2, IRS1 and IGFBP3 was seen, both at the mRNA and protein levels, in chondrosarcoma cell lines. This indicates that essential IGF pathway components are present in cultured chondrosarcoma cells. Furthermore, detection of phosphorylated IRS1 in two out of three chondrosarcoma cell lines demonstrates that the IGF pathway is active in a subset of chondrosarcoma cell lines. In the cell lines with phosphorylated IRS1, IGF pathway inhibition decreased phosphorylated Akt levels and increased IGF1R expression; the latter suggests activation of a feedback loop, which is further supported by the downregulation of IGF1R expression by IGF1 treatment. However, this did not influence the amount of phosphorylated S6, which is located further downstream in the PI3K/Akt/mTOR pathway. Furthermore, the activated MAPK levels were not affected by IGF pathway stimulation or inhibition, demonstrating that activity of the downstream targets is not dependent on IGF1R signalling.
In line with these findings, we demonstrate that despite activity of the pathway, IGF1R signalling is not essential for chondrosarcoma cell survival. Treatment with three different IGF1R/IR inhibitors does not have an effect on chondrosarcoma cell viability, irrespective of apparent pathway activity and stimulation with IGF1. Chondrosarcoma cell line OUMS27 was previously shown to be sensitive to IGF1R/IR inhibition by Zhang et al. [
13]. It is difficult to explain the discrepancy with the current study, as OSI-906 is a derivate of the IGF1R inhibitor used by Zang et al. with similar target potency [
41]. Moreover, we performed these assays at multiple cell densities, passage numbers and IGF1R/IR inhibitors (data not shown).
IGF1R signalling is involved in resistance to cytotoxic drugs in certain cancers [
40]. Since chondrosarcoma is resistant to chemotherapy, we explored a possible role of the IGF1R/IR pathway in chemoresistance. Doxorubicin reduced cell viability in a dose dependent manner; however, OSI-906 did not further inhibit cell viability in this cell line model. These results do not support a key role of the IGF pathway in chondrosarcoma cell survival and chemoresistance.
Our study could not confirm a role for the IGF pathway in chondrosarcoma cell migration. In contrast to the study from Wu et al., showing that IGF1 induced chondrosarcoma migration was inhibited by an IGF1R antibody [
23], we chose not to pretreat the chondrosarcoma cells with IGF1 and not to use medium supplemented with IGF1 only as chemoattractant, thereby better mimicking the
in vivo situation. These experimental differences might explain the difference in our findings.
Strikingly, we detected high expression of the IGF1R in chondrosarcoma cell lines compared to primary tumours. Moreover, we show that each of four patients with matched cell lines and primary tumours had strong membranous IGF1R expression in the cultured cells compared to absent or very weak expression in the corresponding primary tumour. The finding that cell lines are insensitive to IGF1R inhibition despite their high IGF1R expression is in line with the results from the study by Schwartz et al., which described absence of a correlation between IGF1R expression levels and responsiveness to an IGF1R targeting antibody [
18]. This series included 38 chondrosarcomas of which 53 % had immunohistochemical staining with an IGF1R antibody [
18]. Therefore, we did not anticipate to find weak (34 %) or no expression (66 %) in our cartilage tumour series. The discrepancy between our results and the study from Schwartz et al. can likely be explained by use of another antibody. Lack of reproducibility is a well described phenomenon in preclinical studies with antibodies [
42,
43]. Our study further suggests that IGF1R expression is lower in clear cell chondrosarcoma and mesenchymal chondrosarcoma compared to the other cartilage tumours. However, as the staining is very weak in the samples scored positive and IGF1R expression levels do not correlate with responsiveness to IGF1R targeting antibodies, we do not think this difference in IGF1R expression has clinical significance. Furthermore, we did not see a difference in sensitivity to IGF1R inhibition between the mesenchymal, dedifferentiated and conventional chondrosarcoma cell lines included in this study.
Increased activity of the IGF pathway is implicated in several other cancers [
12] including other bone tumours [
44]. In Ewing sarcoma, IGF binding protein 3 (IGFBP3) is downregulated by the EWSR1-FLI1 fusion gene [
45], activating the IGF pathway [
46]. Recently, aberrant expression of IGF pathway members was described in osteosarcomas and OSI-906, a dual inhibitor of the IGF1R and the IR, inhibited proliferation in 3 out of 4 osteosarcoma cell lines with IC
50 values within the therapeutic range [
24].
Clinical trials to test the safety and efficacy of IGF1R antibodies, sometimes in combination with an mTOR inhibitor, have been performed in sarcoma patients [
47], but only two trials enrolled chondrosarcoma patients [
18,
44]. In the study described by Olmos et al. one myxoid chondrosarcoma was included, which showed a small decrease in tumour size upon IGF1R inhibition [
44]. It is unclear whether this was an extraskeletal myxoid chondrosarcoma or a chondrosarcoma of bone. In addition, 1 of 17 chondrosarcoma patients showed partial response to Cixutumumab (IGF1R antibody) and Temsirolimus (mTOR inhibitor), as described by Schwarz et al. [
18]. In future studies, dual inhibitors of both the IGF1R and the IR are preferably chosen because it has been shown in osteoblasts [
48] and Ewing sarcoma cells [
49] that cells can circumvent inhibition of IGF1R by increasing IR signalling.
Abbreviations
ACT, Atypical cartilage tumour; CIM, Cell Invasion and Migration; FBS, Fetal bovine serum; FFPE, formalin-fixed, paraffin-embedded; HRP, horseradish peroxidase; IGF, Insulin-like growth factor; IGF1R, Insulin-like growth factor receptor; IGFBP, Insulin-like growth factor binding protein; IP, Immunoprecipitation; IR, Insulin receptor; IRS, Insulin receptor substrate; MAPK, Mitogen-activated protein kinases; TMA, Tissue microarray
Acknowledgements
The authors thank Dr. Joel A. Block (Rush University, Chicago, IL, USA) for providing JJ012, Prof. Antonio Llombart-Bosch (Valencia University, Spain) for CH3573 as well as CH2879 with corresponding tissue block, Dr. M. Namba (Okayama University Medical School, Japan) for OUMS27, and dr. Naoko Kudo (Niigata University Graduate School of Medical and Dental Sciences, Japan) for NDCS1. TC-32 was a gift of Dr. S.A. Burchill (St. James’ University Hospital, Leeds).
The authors also thank N. Athanasou (Nuffield Department of Orthopaedic Surgery, University of Oxford, Oxford, United Kingdom), S. Daugaard (Department of Pathology, University Hospital Copenhagen, Copenhagen, Denmark), B. Liegl-Atzwanger (Department of Pathology, Medical University Graz, Graz, Austria), and P. Picci (Laboratory of Oncologic Research, Rizzoli Orthopaedic Institute, Bologna, Italy) for contributing cases of rare chondrosarcoma subtypes. In addition, authors also thank Pauline M. Wijers-Kosters, Zuzanna Baranski and Anke H.W. Essing for expert technical assistance and Danielle Meijer, Dorien van der Geest and Jolieke G. van Oosterwijk for construction of TMAs.