Abstract
Logically, it is easy to consider RET mutation as an early event and the primary initiator of tumorigenesis; after all, mutations of this gene cause MEN2. Consistent with this hypothesis are also the earliest phenotypic studies of MEN2 suggesting that MTC, like other inherited cancers, followed Knudson's two-hit model where tumors develop after somatic loss of the normal tumor suppressor allele (7). Ironically, it was this belief that MEN2 involved a tumor suppressor gene that somewhat impeded the discovery of RET mutations once the causative gene had been linked to 10q11.2 (8). It took the realization that this region was not subject to allele loss to consider other targets as driving formation of this tumor (9,10). To this day, the finding that germline activating mutations in a proto-oncogene predispose individuals to cancer development remains an anomaly. Besides RET, MET and KIT are the only other proto-oncogenes implicated in hereditary cancer syndromes (11,12). The three genes have some common features: they are able to transform normal fibroblasts in culture, are widely expressed in numerous normal cell types, and are remarkably restricted in their tumor presentation (RET: MTC, pheochromocytoma, parathyroid hyperplasia; MET: papillary renal carcinoma; KIT: gastrointestinal stromal tumor). Therefore, how these activated proto-oncogenes function as cell-specific initiators of tumorigenesis remains unclear, but clearly at least one, and likely multiple, additional genetic hits are required. Why is this important? Because those genes responsible for initiating the genetic cascade to tumorigenesis typically are not viable therapeutic targets. For hereditary MTC, RET currently serves a diagnostic role, determining the need for prophylactic thyroidectomy.
So what is the evidence that RET serves as a “driver” of tumorigenesis and more importantly is the primary mediator of MTC cell oncogene addiction? First, the RET gene was initially defined as a proto-oncogene by the classical NIH 3T3 cell transformation assay (13,14) and shortly thereafter found to play a major role in papillary thyroid carcinoma (15). In both of these examples gene rearrangement resulting in aberrant overexpression of RET, not the presence of an activating mutation, was responsible for tumoral transformation. It was only the discovery of activating mutations in MEN2 that led to the confirmation of their transforming capacity (16). More importantly, RET M918T mutations were identified not only as the cause of MEN2B phenotype, but also a frequent somatic mutation in sporadic MTC (17). The current COSMIC somatic mutation database reports a frequency of 41% RET mutations in 1055 unique MTC tumor samples, with 72% of these being M918T (18,19). Of note, pheochromocytoma is the only other tumor observed to contain RET activating mutations. Where RET mutations exist in other tumors they are unrelated to those reported to cause MEN2, and in many cases, are consistent with inactivation. This suggests that RET activation may be unique to MTC and pheochromocytoma, while aberrant overexpression plays a primary role in other cancers; for example, papillary thyroid carcinoma (15), breast carcinoma (20), lung adenocarcinoma (21), prostate cancer (22), and pancreatic cancer (23). These seemingly disparate findings again raise the question of whether RET functions as an initiator of tumorigenesis, a driver, or perhaps both.
This issue of Thyroid includes a study by Romei et al. (24) in which the investigators evaluated the prevalence of somatic RET mutations in sporadic MTC tumors. They reasoned that if activated RET served primarily as an initiator of tumorigenesis, then mutations would exist as early event and thus be present in the smallest primary tumors available for analysis. If RET's primary role was to drive tumor growth, then they should be disproportionately absent in small tumors compared to larger tumors. It is a logical straightforward approach that relies on a large sample population in order to derive a statistically meaningful conclusion; after all, somatic RET mutations are clearly absent in the majority of sporadic MTC tumors examined (59%). In an analysis of 160 sporadic MTC tumors, the overall prevalence of RET M918T mutation was 19.4%. While this number is lower than that seen in other published series, including COSMIC, it is important to remember that the sample population is biased towards smaller tumors (75.6% of tumors were ≤2 cm). However, with that caveat, the observed frequency of mutations in large tumors (>2 cm) was statistically greater than that observed in small tumors (≤2 cm). At face value, these numbers are remarkable, 11.6% compared with 43.6%. It should be noted that at their center, tumors smaller than 2 cm are routinely embedded in paraffin for histological examination (68.5% of the tumors used in this study), which can lead to an insufficient quantity or quality of DNA for complete analysis compared with DNA isolated from fresh tissue. But, the clear distinction between the data available for the two study groups indeed supports that RET mutation may be a secondary event serving to drive tumorigenesis rather than acting as an initiator. This is not the first time somatic RET mutations have been proposed to be a secondary event in MTC (25). Paired analysis of primary and metastatic tumors from four sporadic MTC patients found the presence of RET M918T mutation in 25% of the metastatic tumors tested despite its absence from the primary tumor. Furthermore, in 14 of 22 primary tumors examined, it was clear that RET M918T mutation was not clonally distributed (25). Romei et al. did not perform a similar paired analysis of tumors samples in their study. Also missing from their study is a detailed discussion of patient clinical parameters at the time of sample acquisition. Our experience has been that size alone does not always predict tumor aggressiveness, extent of disease at diagnosis, or patient outcome. Consistent with previous studies RET M918T mutation is associated with a more aggressive tumor, greater recurrence and poorer prognosis independent of tumor size (24,26,27). In other words, a more clinically relevant conclusion from their data is that patients with RET-positive tumors have a lower “cure” rate than RET-negative patients and require close surveillance for progression.
The underlying early events responsible for the initiation of sporadic MTC remain to be uncovered, as do additional oncogenic drivers of tumorigenesis in both sporadic and hereditary MTC. The work presented by Romei et al. supports the hypothesis that mutated RET may be a secondary event, driving tumor progression, and importantly, a therapeutic oncogenic target in sporadic MTC. But we are still left with the question of whether the discovery of a somatic RET mutation provides an actionable treatment course. Until we identify other factors that initiate or drive MTC progression, we will continue to treat patients with progressive and inoperable MTC with vandetanib or other available multikinase inhibitors through clinical trials with the understanding that activated RET is an important addictive oncogene to target.