Harvesting the Low-Hanging Fruit: Kinase Inhibitors for Therapy of Advanced Medullary and Nonmedullary Thyroid Cancer

A pproximately 37,000 new cases of thyroid cancer are diagnosed annually in the United States, and of these, about 10–20% present with or ultimately develop distant metastases (1). The prognosis of patients with distant metastases from non-medullary thyroid cancer (non-MTC) remains favorable when these are responsive to radioactive iodine (RAI). By contrast, the 10-yr survival after discovery of distant metastases is only 10% if these have lost the ability to trap or retain iodine. Patients with metastatic RAI-refractory lesions that are also positive on fluoro-deoxyglucose-positron emission tomography (FDG-PET) have a particularly poor prognosis. In addition to patients with distant metastatic disease, those with unresectable local recurrences also have a higher mortality. Palliative external beam irradiation plays an important role in local disease control; however, conventional chemotherapeutic agents are relatively ineffective in thyroid cancer (2). Recent clinical trials with several kinase inhibitors have shown promising evidence of activity in thyroid cancer of follicular origin (3,4,5,6) as well as in MTC (7,8).

The rationale for use of kinase inhibitor therapy in advanced forms of non-MTC is grounded on the genetics and the biology of the disease, in which oncogenic kinases play a prominent role. Metastatic RAI-refractory papillary and poorly differentiated thyroid cancers that are positive on FDG-PET scans have a high frequency of mutations of BRAF and RAS, which do not overlap in the same tumor specimen. In addition, mutations of kinase-encoding genes that activate signaling along the phospho-inositol 3-kinase pathway are found in a significant proportion of these lesions (9).

Sorafenib was originally developed as a Raf inhibitor but is now known to block the activity of multiple molecular targets, including Raf-1, B-Raf, vascular endothelial growth factor receptor (VEGFR)-1, VEGFR-2, VEGFR-3, platelet-derived growth factor receptor β, fms-like tyrosine kinase 3, c-KIT, and RET (REarranged during Transfection) (10). This compound is approved by the U.S. Food and Drug Administration (FDA) for use in renal cell carcinoma, where it has been proposed to exert its beneficial effects through its antiangiogenic properties (11,12), as well as for unresectable hepatocellular carcinoma. Of note, neither of these tumor types is associated with a particularly high prevalence of activating mutations of any of the therapeutic targets of this compound. Two recent phase II clinical trials showed that sorafenib had significant activity in RAI-refractory thyroid cancer, with partial responses ranging from 15–23% of patients, and stabilization of the disease in an additional 53–56% (3,5). Although the FDA has not yet approved sorafenib for this indication, in response to these findings the National Comprehensive Cancer Center Network (NCCN) Guidelines in Oncology currently recommend consideration of small-molecule kinase inhibitors or systemic therapies in patients with progressive metastatic disease who are not willing or able to enter into a clinical trial. Accordingly, these drugs are being used at cancer centers off-label for this indication. In this issue of JCEM, Maria Cabanillas et al. (13) from the MD Anderson Cancer Center in Houston report on their experience in 15 patients with progressive, RAI-refractory thyroid cancer, 13 of which were treated with sorafenib and two with sunitinib. This study provides valuable insights about the use of these drugs in real practice. Any retrospective report is subject to patient selection concerns, but most exclusions in this study related to patients that were treated with other kinase inhibitors or whose follow-up was conducted at other sites. All patients had evidence of progressive disease before the start of treatment, and response rates were very consistent with those reported in the two phase II studies mentioned above.

The authors found disparity in response according to the site of metastases: i.e. target lesions in lung were more responsive than in lymph nodes, with bone metastases being the least sensitive to the drug. The apparent refractoriness of thyroid cancer bone metastases to sorafenib was also reported in a recent study from The Netherlands (14) and has also been observed in other cancer types (15). There is as yet no clear explanation for this phenomenon, which could be due to the pharmacokinetics of the drug or to metastatic niche-specific microenvironmental factors influencing tumor growth or viability. If confirmed in appropriately designed trials, this property will affect how sorafenib and perhaps other compounds of this class may be used in practice because metastatic bone disease is often an important driver of disease prognosis.

Of the 15 patients described in the Cabanillas report (13), eight had papillary thyroid cancer and seven had follicular thyroid cancer (two of these were Hurthle cell carcinomas). Neither the histological subtype nor the presence of BRAF mutations predicted response to therapy. This argues against the beneficial therapeutic effects of sorafenib being due to inhibition of B-Raf kinase activity and is consistent with the lack of efficacy of this compound in patients with metastatic melanoma, including those with BRAF mutations (16). This should not be interpreted as evidence that cancers with mutant BRAF are not “addicted” to the oncoprotein and dependent on ERK signaling. Much to the contrary, a recent phase I clinical trial of patients with metastatic melanoma with PLX4032, a relatively selective Raf inhibitor with preferential activity on mutant B-Raf, showed that only those patients with BRAF mutations responded to the compound (17). Three patients with BRAF (+) papillary thyroid cancer also showed favorable responses. The efficacy of PLX4032 is believed to be due in part to its low toxicity, which allows the administration of a sufficiently high dose to fully block B-Raf kinase activity. However, there is an important cautionary note. Raf inhibitors paradoxically activate MAPK kinase (MEK)-ERK signaling in cells with wild-type BRAF, or with mutations of RAS (18,19). Thus, ATP-competitive Raf inhibitors inhibit their catalytic activity, but mimic the active, phosphorylated conformation of the kinase. In the presence of activated Ras, the drug-bound Raf isoform can therefore activate MEK-ERK signaling through transactivation of its drug-free C-Raf or B-Raf dimerization partner. This paradoxical activation of the pathway does not take place in cells with mutant B-Raf, presumably because Ras activity is down-regulated, and Raf dimerization is therefore not promoted (18).

These findings may help explain some of the side effects and potential concerns associated with the use of Raf inhibitors. Several studies (20,21), including the Cabanillas report (13), point to development of squamous cell carcinomas and keratoacanthomas in patients on sorafenib. It is reasonable to speculate that these tumors may have arisen through promotion of MAPK signaling, and it will be interesting in due course to determine whether squamous cell carcinomas arising in the context of sorafenib therapy are particularly enriched for oncogenic RAS mutations.

One of the patients who failed to respond to sorafenib in the Cabanillas study (13) had a significant response to sunitinib. Hence, despite an overlapping profile of kinase targets, patients who fail to respond to one of these drugs should not be assumed to be unresponsive to the other. Moreover, clinical trials designed to explore their activity should not exclude patients who received prior therapy with the other inhibitor.

Despite these encouraging reports, one important message that should be drawn from the sorafenib and sunitinib experience in thyroid cancer so far is that patients must be carefully selected before initiation of therapy. It is prudent to recommend that only patients with more rapidly progressive RAI-refractory metastatic disease be started on these drugs because there is no proof that they confer patients with an improved overall survival. These metastatic lesions will often be positive on FDG-PET scans, whereas indolent, slowly developing metastases are not. Patients with unresectable, locally recurrent cancer that threatens vital structures and/or is unresponsive to external beam radiotherapy are also candidates for therapy. Patients with progressive metastatic disease that is primarily confined to bone may not benefit from treatment with sorafenib, although this needs to be formally investigated further. Median progression-free survival in the Cabanillas study (13) was 19 months, and undesirable effects, although manageable in most cases, were quite significant. Moreover, the paradoxical effects of sorafenib to activate MAPK signaling in cells that are wild-type for BRAF raises the theoretical concern that under certain conditions this drug may serve as a tumor promoter. Until we have definitive data showing that one or more of these drugs extend survival, we believe patients needing systemic therapies should be entered into a clinical trial if one is available. If not, they should be offered empiric treatment with small-molecule kinase inhibitors or other systemic therapies, as recommended in the NCCN guidelines.

Hereditary MTC is the prototypic solid tumor type that should, at least in principle, be an ideal candidate for targeted therapy directed against an oncogenic kinase. This is because germline activating mutations of RET, a protooncogene that encodes the tyrosine kinase receptor for a family of neurotrophic ligands, is mutated in virtually all patients with familial forms of MTC and is thought to be causally involved in tumor initiation. There are multiple lines of evidence from cell lines and animal models that show that the constitutively active RET kinase is a key driver of the disease (22,23,24). Moreover, a number of small molecules with strong inhibitory activity on RET kinase have been described (25). Among these, vandetanib, an inhibitor of VEGFR, epidermal growth factor receptor, and RET, has been studied in greatest detail (26,27). Recently, Wells et al. (7) reported the results of an open-label phase II study of vandetanib in patients with advanced hereditary MTC, treated with a maximal tolerated dose of the compound (300 mg/d). Twenty percent of patients had a confirmed partial response, and an additional 53% had stable disease lasting at least 24 months. Of note, several patients required a reduction of the dose because of toxicity, such as skin rash and QT prolongation. In this issue of JCEM, Bruce Robinson et al. (28) report on an open-label phase II trial of patients with advanced metastatic hereditary MTC treated with a lower starting dose of vandetanib (100 mg/d). They observed a partial response in three of 19 patients (16%) and stable disease lasting a minimum of 24 wk in an additional 10 of 19 (53%), for an overall clinical benefit of 69%. Because patients were not confirmed to have experienced disease progression (as determined by response evaluation criteria in solid tumors) before entry into the trial, the significance of disease stabilization at 24 wk cannot be assessed. The study was not designed to compare efficacy between the doses used in the two trials. Similarly, the side effect profile appeared consistent with what has been reported for vandetanib in other trials, and it is not clear whether the lower dose was better tolerated. However, it is reassuring that an antitumor effect is still observed at a lower dose of this compound. The authors acknowledge that at this point there is no information to conclude whether the effects of vandetanib are mediated through inhibition of RET kinase activity, or of its other molecular targets.

Physicians treating patients with thyroid cancer rely on biomarkers as one of the key tools to monitor disease progression and response to therapy. In MTC, calcitonin determinations have long been used as a useful surrogate measurement of tumor burden. However, it is now apparent that RET signaling plays a significant role in controlling calcitonin gene expression in normal and tumoral C cells (29). Hence, in patients treated with inhibitors of RET kinase activity, serum calcitonin and tumor cell mass may not necessarily go hand in hand, and measurement of carcinoembryonic antigen levels may provide a better indicator of the response to therapy (7).

Kinase inhibitors are now at the forefront of developmental therapeutics for patients with MTCs and non-MTCs. One of the important challenges ahead will be to establish the mechanism of action of some of the compounds that are showing efficacy in the clinic. Virtually without exception, these molecules are ATP competitors (i.e. axitinib, sunitinib, sorafenib, vandetanib, and motesanib), and as such inhibit several molecular targets, which likely results in a worse profile of side effects and poorer tolerance. Development of allosteric kinase antagonists, such as the MEK inhibitor AZD6244, or even ATP-competitive compounds with a narrower spectrum, such as the Raf inhibitor PLX4032, provide a roadmap for the future but will rely on identifying the key drivers of the disease and inhibiting them with greater selectivity.