Is Estrogen the Answer for Osteosarcoma?

Despite being the most common primary bone tumor, patients diagnosed today with osteosarcoma are presented with the same treatment options and experience the same outcomes that they did in the 1970s (1). Given their characteristic production of abnormal bone, many researchers have suspected that these tumors result from malignant transformation of osteoblast progenitors and have sought to promote differentiation as a means of controlling disease. A few previous reports (2–4) have hinted that estrogens, which play prominent and well-described roles in osteoclast biology (5), might have a differentiating effect on osteosarcoma cells. In this issue of Cancer Research, Lillo Osuna and colleagues subject this hypothesis to more rigorous evaluation and present intriguing data that suggests an epigenetic silencing of estrogen receptors in osteosarcoma that, when reversed with clinically relevant methyltransferase inhibitors, promotes differentiation of several osteosarcoma cell lines in vitro, which can be replicated with some disease control effects in vivo (6).

Most of the previously published literature used osteosarcoma cell lines as a model for osteoblast biology and sought to study the role of estrogen in bone remodeling and homeostasis. Inherent in the design of those experiments was the observation that normal mature osteoblasts expressed relatively high levels of estrogen receptors, which was lost in the osteosarcoma cells. Thus, studying the effects of estrogen in these osteoblast models required plasmid-driven expression of the estrogen receptors. In this current work, Lillo Osuna and colleagues take the opposite approach, seeking to understand why the receptor expression is lost and showing how, in this particular malignancy, estrogen receptors may function very much like a tumor suppressor.

Lillo Osuna and colleagues make a series of important observations. First, they show that estrogen receptors are epigenetically silenced in several osteosarcoma cell lines through promoter hypermethylation. Second, they show that estrogen receptors can be reexpressed with exposure to decitabine, an FDA-approved DNA methyltransferase inhibitor. Treatment with decitabine induced differentiation, an effect replicated simply with plasmid-driven expression of ESR1 and enhanced with the addition of exogenous estrogen. The same treatment decreased proliferation and colony formation for osteosarcoma cell lines in vitro and inhibited tumor growth and decreased metastasis in a mouse xenograft model of osteosarcoma. These data present a convincing argument that reversible epigenetic silencing of estrogen receptors could be a therapeutically viable mechanism for inducing differentiation osteosarcoma tumors.

Perhaps the most impressive data presented in this article comes in the seemingly profound effect that decitabine treatment had on the formation of metastatic lesions in the lungs. The development of metastatic lung lesions represents a primary challenge in the clinical care of patients with osteosarcoma (7). Lung nodules represent the vast majority of metastatic lesions in osteosarcoma and tend to be resistant to both chemotherapy and radiation, making lung metastasis a focus of many basic and clinical scientists in the field. If the effect on lung metastasis demonstrated in this xenograft model translates accurately into other models and species, this particular observation could have significant clinical impact.

The work presented is not without caveats and shortcomings, and one should interpret some of this data with caution. Although the authors repeated a few key experiments across multiple osteosarcoma models, most of the work, and importantly, most of the in vivo work, was performed with a single cell line. Although the ESR1 knockdown experiments suggest that a large portion of the differentiating effects were driven by the reexpression of that gene, other mechanisms (off-target in the context of this hypothesis) clearly contribute, which one might expect from a drug with such broad epigenetic activity. And, for both the in vitro proliferation and the in vivo tumor growth experiments, the most profound effects observed were still merely tumor growth inhibition, which one assumes would translate to progressive disease in a clinical setting.

One can speculate that these treatments might prove more effective when combined with other interventions in synergistic ways. One obvious candidate would be estrogen itself. Indeed, the authors showed that administration of exogenous estrogen enhanced the expression of several differentiation-associated genes, but they did not carry this approach forward toward in vivo testing. It is interesting to speculate whether selective estrogen receptor modulators (SERM) might have potentiating effects when combined with treatments that drive reexpression of estrogen receptors. Although SERMs have variable pro- and antiestrogen effects in different tissues, they have long been known to promote osteoblast activity. Bazedoxifene might be a particularly attractive candidate for combination therapy, as recent work would suggest an additional off-target role in preventing metastasis through effects on IL6 signaling (8, 9). Future studies aimed at identifying effective combinations are needed and will surely be forthcoming.

The current standard-of-care first-line treatment regimen for osteosarcoma (MAP chemotherapy, which combines doxorubicin, cisplatin, and methotrexate) presents significant challenges for incorporating new modalities into first-line therapy. These challenges are well recognized by the leaders of the various cooperative groups, although divergence from the current regimen will require solid preclinical data. The activity shown for decitabine, strongest in its effect on metastasis formation, may make it particularly attractive for vetting through veterinary clinical trials. Given that canine species have high rates of osteosarcoma (10), pet dogs diagnosed with this disease can be enrolled in clinical trials to evaluate the effects of candidate therapies in ways that inform the development of pediatric trials. With a compressed natural history and high rate of metastatic recurrence, these integrated/comparative approaches lend themselves particularly well to the evaluation of therapies that target metastasis.

Although there remains work to do to determine the real potential utility of the approach put forward by Lillo Osuna and colleagues, one should recognize that this represents a novel approach to the treatment of osteosarcoma, leveraging some long-known elements of osteoclast biology in the treatment of this bone cancer. Indeed, some might say that this approach seems obvious in retrospect, and might be surprised that such work has not been done until now. That could be the hallmark of a good idea.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.