Chondrosarcomas of bone are malignant cartilage-forming tumors which are highly resistant to conventional chemotherapy and radiotherapy [1, 2]. However, recently various promising targets were discovered and the exploration of suitable therapies continues [3, 4]. Conventional chondrosarcomas represent about 90% of all chondrosarcomas. Most conventional chondrosarcomas are located in the medullar cavity of the bone and are called central chondrosarcoma. About 15% of conventional chondrosarcomas arise from the surface of bone and are designated as peripheral chondrosarcomas. Conventional chondrosarcomas often show local destructive growth and the high-grade tumors commonly metastasize [5].

Besides conventional chondrosarcoma, several rare chondrosarcoma subtypes are defined, together constituting 10-15% of all chondrosarcomas. Dedifferentiated chondrosarcoma (10%) is a tumor containing a high-grade dedifferentiated non-cartilaginous sarcoma next to a usually low-grade malignant well-differentiated cartilage-forming tumor, with a sharply defined junction between the two components. It bears a poor prognosis and no targets for therapy have been reported so far [6]. Mesenchymal chondrosarcoma (2%) is a highly malignant lesion occurring in the bone and soft tissue of relatively young patients. The tumor consists of differentiated cartilage mixed with undifferentiated small round cells and usually follows an aggressive course with a high rate of distant metastases, and a 5-year overall survival of 55% [7]. Clear cell chondrosarcoma (2%) is a low-grade malignant tumor, which rarely metastasizes, but commonly recurs after curettage. About 15% of the patients die as a result of the disease [8]. The lack of efficacious treatment for all different subtypes of chondrosarcomas emphasizes the need to identify new treatment strategies.

One of the potential targets for therapy is the estrogen-signaling pathway. Mutations in ESR1 and CYP19A1, the gene for aromatase, demonstrated an important role for estrogen in the proliferation and differentiation of chondrocytes in the epiphyseal growth plate [9]. Estrogen induces the pubertal growth spurt, and at the end of puberty growth plate fusion [10]. Furthermore, osteochondromas, the benign precursors of peripheral chondrosarcomas, stop growing at the end of puberty, suggesting an inhibitory effect of estrogens on these tumors. In addition, ESR1 and ESR2 expression has been shown to be a common phenomenon in chondrosarcomas [11, 12]. In a previous study, our group also demonstrated functional activity of the estrogen-producing enzyme aromatase in chondrosarcoma cells in vitro[11]. These results indicated that the ESR signaling pathway might be a potential target for endocrine treatment of metastatic or irresectable chondrosarcoma.

For already three decades endocrine therapy plays a crucial role in the treatment of women with hormone-responsive breast cancer. Breast cancer and chondrosarcomas were found to occur relatively frequently in the same patient. A population-based study by Odink et al. implicated a 7.62 times increased risk for the same female patient to have both breast cancer and a cartilaginous tumor [13]. The mean age of onset in patients with breast cancer as the first tumor and chondrosarcoma as a second tumor is nearly 10 years earlier than breast cancer in general [13]. These observations may suggest a genetic trait. Remarkably, the expression of ESR1 was significantly higher in breast cancer associated with chondrosarcoma [14].

The two strategies used for endocrine treatment are blockade of ESR1 using selective estrogen receptor modulators/downregulators like tamoxifen and fulvestrant, and deprivation of estrogen production by inhibiting aromatase with anastrozole, letrozole, and exemestane. In our above-mentioned study, we showed that the aromatase activity and proliferation of chondrosarcoma cells slightly decreased after addition of the aromatase inhibitor exemestane [11]. In our present study, we focused on further unraveling the role of estrogen in the regulation of neoplastic cartilage growth in a larger cohort of various chondrosarcoma subtypes, including conventional central and peripheral chondrosarcoma as well as dedifferentiated, mesenchymal, and clear cell chondrosarcoma. Moreover, using a larger set of drugs targeting the estrogen-signaling pathway we investigated whether interference with estrogen signaling could inhibit chondrosarcoma growth. We aimed to validate and expand our previous in vitro data by measuring the effects of estrogens, androgens, tamoxifen, fulvestrant, and aromatase inhibitors on the proliferation of various chondrosarcoma cell cultures. Furthermore, we explored the efficacy of aromatase inhibitors in a set of patients with metastatic or locally advanced chondrosarcoma.

Chondrosarcoma of bone is a malignant cartilage-forming tumor of which distinct clinical and histological subtypes are recognized. So far, for locally advanced and metastatic chondrosarcoma no treatment options are available. Previous studies have demonstrated the presence of the ESR1 and activity of aromatase in conventional chondrosarcoma [11, 12]. Furthermore, in 2005, our group showed an effect of estrogens and the aromatase inhibitor exemestane on the proliferation of chondrosarcoma cells in vitro, indicating that chondrosarcomas might be susceptible to hormonal therapy. In that study, ESR1 and CYP19A1 mRNA expression were demonstrated in a set of 23 conventional chondrosarcomas and 7 (primary) chondrosarcoma cultures. ESR1 protein expression was demonstrated in all 23 tumors tested. Addition of 17β-estradiol, 4-androstene-3,17-dione, and exemestane showed subtle effects on the proliferation of 2 cell cultures containing ESR1 and aromatase. After addition of 4-androstene-3,17-dione, an increase in proliferation was demonstrated. Proliferation was 131% of normal proliferation which decreased to 105% after inhibition with exemestane. A cell line lacking ESR1 and CYP19A1 did not show any response.

In the current study, we aimed to gain more insight into the possibility of treating chondrosarcoma patients with hormonal therapy by further investigating the expression of the hormone receptors ESR1 and AR, and of aromatase, the enzyme that mediates the last step in the biochemical formation of estrogen, in a larger set of conventional chondrosarcomas as well as three rare chondrosarcoma subtypes. In conventional chondrosarcoma, we furthermore monitored the effect of estrogen, the estrogen precursor androstenedione and the non-aromatizable androgen dihydrotestosteron, and various known estrogen signaling-inhibiting drugs on the progression of chondrosarcoma cells in vitro.

We demonstrated expression of ESR1 in a large proportion of various types of cartilaginous tumors. In conventional central and peripheral chondrosarcoma we observed immunoreactivity against ESR1 in 81% (34 out of 42) and 81% (21 out of 26) of the tumors, respectively. These results confirm and extend 2 previous studies in which the authors demonstrated nuclear expression of ESR1 in subsets of 23 [11] and 31 [12] conventional chondrosarcomas. Grifone et al. [12] suggested a decrease or loss in ESR1 expression in the higher grade or dedifferentiated chondrosarcomas. We observed such a trend in the peripheral chondrosarcomas, where only 33% of the high grade tumors show positive staining for ESR1. However, this group included only three tumor specimens. In central chondrosarcoma no correlation with grade was observed. Aromatase protein, the enzyme responsible for the conversion of androstenedione and androgens to estrogens, was expressed in 86% and 93% of the central and peripheral chondrosarcomas respectively, suggesting that tumors are capable of metabolizing estrogens from precursors. AR is another important target for hormonal therapy in for example prostate cancer. As androstenedione is a steroid precursor for estrogens as well as androgens we also investigated the possibility of AR involvement in chondrosarcoma proliferation. However, AR nuclear protein expression was observed only in a small number of cases with very few positive cells.

Besides conventional chondrosarcoma, several rare chondrosarcoma subtypes are defined. Despite aggressive therapy, approximately 90% of the patients with dedifferentiated chondrosarcoma die with distant metastasis, within 2 years after diagnosis of the disease [6, 23]. The low-grade component and the highly malignant component display ESR1 protein expression in 72% and 86% of the samples respectively. Aromatase was observed in 97% and 89%, suggesting the presence of estrogens.

Mesenchymal chondrosarcomas are usually very aggressive with a strong tendency of local recurrence and distant metastases. Patients have a 5-year overall survival of 55% [7]. Although mesenchymal chondrosarcoma of bone is generally considered to lack sex predilection [24], Fanburg-Smith et al. [25] suggested a female predominance and raised the possibility of hormonal influence in the pathogenesis of this tumor. However, all their mesenchymal chondrosarcoma cases were ESR1 negative. In our study, in 65% (15 out of 23) of the mesenchymal chondrosarcomas the small cell component was positive for ESR1, while in 33% (5 out of 15) of the tumors also the cartilaginous areas were positive. Moreover, aromatase expression was observed in the small cells of 52% (12 out of 23) of the tumors, whereas the cartilage component demonstrated aromatase expression in 77% (10 out of 13). This might indicate that these tumors do have an active estrogen signaling pathway, which might be targetable by antiestrogens or aromatase inhibitors. Discrepant results may be explained by differences in ESR1 antibody and antigen retrieval protocols.

Clear cell chondrosarcoma is a low-grade variant of chondrosarcoma, which rarely metastasizes, but has a recurrence rate of 86% after curettage. About 15% of the patients die as a result of the disease [8]. We have observed ESR1 expression and aromatase expression each in 69% of the clear cell chondrosarcomas, suggesting that also these chondrosarcoma patients potentially might benefit from antiestrogen therapy and/or aromatase inhibition.

In vitro cell models to further study the effect of estrogen signaling on chondrosarcoma are available for conventional central chondrosarcoma only. No stimulation of proliferation of central chondrosarcoma cells was observed after addition of the non-aromatizable androgen dihydrotestosterone. This suggests no significant role for AR signaling in chondrosarcoma proliferation, which is consistent with the fact that very few tumors express AR.

In addition, in spite of positive immunohistochemical staining for ESR1 protein in all in vitro cell cultures, addition of 17β-estradiol, 4-androstene-3,17-dione or drugs targeting the estrogen-signaling pathway did not have a significant effect on the proliferation of the conventional central chondrosarcoma cell cultures. These results contradict our results published in 2005, where proliferation was stimulated by 17β-estradiol and 4-androstene-3,17-dione, and inhibited by exemestane [11]. Although we included an identical experimental set up, cell culture conditions are never 100% identical. For example, each batch of FBS contains different amounts of growth factors and other components which might influence experimental outcome. Also cell characteristics might have changed over time, resulting in passages insensitive to (anti)estrogens and aromatase inhibitors, as has been described before for certain breast cancer cell lines [26–28].

Breast cancer cell line ZR-75-1 is known to be completely dependent on estrogens for its proliferation, and proliferation can be fully inhibited by abrogating the estrogen-signaling pathway [29]. Although we previously demonstrated an effect of estrogen-signaling on chondrosarcoma cell proliferation, as compared to estrogen-dependent breast cancer cell line ZR-75-1 the effects in chondrosarcoma, if present, were very subtle. As a positive control, ZR-75-1 showed a 179% increase of proliferation upon addition of 1 nM 17β-estradiol, confirming a functional experimental setup, versus a previously demonstrated 55% increase in chondrosarcoma proliferation [11] and no significant increase in the current study. Both studies clearly indicate that, in contrast to estrogen-dependent breast cancer, chondrosarcoma proliferation is not fully dependent on estrogens.

Besides investigating estrogen dependence, we tested aromatase inhibitors which block estrogen production, and the effects of tamoxifen and fulvestrant which abrogate estrogen receptor function [30, 31]. In the estrogen-dependent ZR-75-1 breast-cancer cell line proliferation was completely inhibited upon addition of tamoxifen and fulvestrant (Figure 2B). However, in the chondrosarcoma cell cultures, estrogen-signaling inhibition caused no effects on cell proliferation, suggesting that the mechanism driving proliferation in chondrosarcoma is different from the mechanism active in estrogen-dependent breast cancer. In chondrosarcoma, effects of estrogen are much more subtle and likely depend on the tissue culture conditions used, resulting in either marginal effects (in our previous study) or no effects at all.

In addition, the median time to progression in the clinical series was five months both before and after treatment. Therefore, we can conclude that aromatase inhibition was not effective in five conventional chondrosarcoma patients, nor in a patient with dedifferentiated chondrosarcoma. Although a formal prospective phase II trial would have been more suitable to prove (in)efficacy of this concept, we were not able to gain industry support without stronger preclinical data.

Since our study is limited to the effects of estrogen signaling on conventional central chondrosarcoma only, no conclusions can be drawn about the effects of estrogen signaling in the other chondrosarcoma subtypes. However, although we demonstrated the presence of aromatase and ESR1 in a majority of various chondrosarcoma subtypes, our in vitro data on conventional chondrosarcoma and our patient trial including one dedifferentiated chondrosarcoma patient suggest that effects of estrogen-signaling inhibition in other chondrosarcoma subtypes, if present at all, will be very small and that estrogen-signaling inhibition is unlikely to play a major role in chondrosarcoma management.

In summary, we demonstrated the presence of the components involved in estrogen signaling in a large majority of chondrosarcomas. However, we could not demonstrate a significant effect of estrogen or inhibitors of estrogen signaling on cell proliferation and viability in vitro using central chondrosarcoma cell lines and primary cultures. Despite the previously presented and currently confirmed biological rationale, our in vitro and pilot clinical data suggest that an active estrogen-signaling pathway might just not play a pivotal role in the development and progression of conventional chondrosarcoma and do not support the further development of therapeutic strategies including inhibition of estrogen signaling in chondrosarcoma.