RET signaling is crucial for mammalian embryonic development as mouse models lacking Ret function die before birth. The mutant embryos exhibit abnormal kidney development, and aberrant ureter, ovary, muscle, and intestine morphology. Ret signaling seems also essential for neurogenesis and establishment of neuronal populations of the central nervous system. In addition, Ret activation is necessary for the development of neural crest-derived lineages, in particular the enteric nervous system, adrenal medulla chromaffin cells and thyroid C-cells (Plaza-Menacho et al. 2014). Notably, loss of Ret, but not Gfra4 or Pspn, is associated with C-cell reductions in adult mice (Lindahl et al. 2000; Lindfors et al. 2006). Within the nervous system, GDNF/RET signaling is crucial to the maintenance of the nigrostriatal pathway by preventing cell death and onset of Parkinson’s disease (Kramer et al. 2007). It is also essential for the survival of motor neurons; alterations in the pattern of phosphorylation lead to amyotrophic lateral sclerosis (Luesma et al. 2014). RET signaling also maintains the enteric nervous system, and loss of RET function in these neurons leads to Hirschsprung’s disease, a condition characterized by the absence of enteric ganglia (Luesma et al. 2014). Importantly, this role of RET is reproduced in both Ret knockout (Schuchardt et al. 1994) and Ret knock-in mice with Ret C620R mutations (mixed Hirschsprung’s/MTC phenotype) (Carniti et al. 2006; Yin et al. 2007). Finally, it is important to remember that Gdnf/Ret signaling has a clearly established role in driving spermatogonial stem cell self-renewal in the testis and is therefore essential to male fertility (Meng et al. 2000; Hofmann 2008; Hofmann et al. 2005; Braydich-Stolle et al. 2007). Thus, in hereditary RETopathies or in therapies targeting RET activity, cell types beyond the thyroid C-cell must be considered.

Several known mutations can convert the RET receptor into a dominant transforming oncogene, resulting in neoplasms affecting multiple cell types and organs. While RET mutation has been primarily associated with thyroid cancer, somatic structural RET alterations or changes of its expression have been described in lung, breast, and pancreatic cancers (Esseghir et al. 2007; Gainor and Shaw 2013; Ito et al. 2005; Kohno et al. 2012; Lipson et al. 2012; Morandi et al. 2013; Plaza-Menacho et al. 2010; Sawai et al. 2005). Germline activating mutations of RET are the cause of multiple endocrine neoplasia type 2 (MEN2), which is discussed in detail within other chapters of this book. Activating mutations of the RET oncogene fall into two major classes, extracellular and intracellular (Fig. 2). The extracellular cysteine-rich domain of RET is most frequently mutated at cysteine 634 (85 % of the cases), but mutations also occur distributed among all cysteines (C609, C611, C618, C620, C630) (Figlioli et al. 2013). Intracellular RET mutations are observed in the RET kinase domain and have a variable impact on RET activity. The most frequent mutation observed is M918T, which is found in patients with MEN2B and somatically in sporadic MTC. The molecular mechanism by which RET mutations trigger transformation has been elucidated (Santoro et al. 1995; Arighi et al. 2005; Santoro and Carlomagno 2013). As discussed, the normal route to activation involves binding of GFRA1-4/ligand complex to RET, which promotes dimerization and in turn autophosphorylation of the receptor. Mutations present in the extracellular domain are known to target highly conserved cysteine residues. Because these residues are normally involved in intramolecular disulfide bond formation, mutation generates a free sulfhydryl group. Mutant RET receptors in close proximity will then form intermolecular disulfide bonds, which function to stabilize dimerization even in the absence of the ligand (Fig. 2). This will therefore constitutively activate autophosphorylation and the intracellular pathways downstream of the receptor (Santoro et al. 1995; Asai et al. 1995; Borrello et al. 1995). In contrast, mutations in the intracellular tyrosine kinase domain of RET are known to target regions associated with the ATP-binding pocket, activation loop, and substrate binding pocket (Santoro et al. 1995; Salvatore et al. 2001; Songyang et al. 1995). Depending on their specific location, they induce structural changes and alter RET substrate specificity. This results in phosphorylation of alternative intracellular proteins. Therefore, the mutated receptor no longer needs dimerization to become active. RET mutations are often an initial step in the oncogenic process.

The RAS family of proteins consists of small GTPases that act as binary molecular switches downstream of ligand–receptor interactions (Malumbres and Pellicer 1998; Wennerberg et al. 2005). They are mainly activated through growth factors binding to RTKs, such as RET. RAS proteins are anchored to the plasma membrane through prenylation, which is indispensable to their function (Buss et al. 1989; Carboni et al. 1995; Cox et al. 1992). RTK autophosphorylation enables interaction with the adaptor protein GRB2 (Fig. 3). GRB2 is bound to SOS, a guanosine nucleotide exchange factor (GEF) that activates RAS by stimulating the release of guanosine diphosphate (GDP) to allow binding of guanosine triphosphate (GTP) (Cherfils and Zeghouf 2013). Therefore, upon stimulation by an activated receptor, RAS switches from the inactive “off” form (GDP bound) to the active “on” form (GTP-bound). Binding of GTP induces a conformational change leading to an increase in RAS affinity and binding to downstream mediators such as RAF or PI3K family members depending on the cell status. RAF is a serine–threonine kinase. RAS acts to recruit it to the plasma membrane and promotes dimerization and activation (Avruch et al. 2001). Activated RAF phosphorylates and activates the kinase MEK, which in turn phosphorylates and activates ERK kinase. Activated ERK phosphorylates a number of substrates, including other kinases and transcription factors, which will induce cell cycle progression (McCubrey et al. 2007). In the other main pathway, activation of PI3K will induce phosphorylation of AKT, leading to inhibition of several tumor suppressors (e.g., p21, BAD) or activation of the mTORC1 complex (Fig. 3) (Ma and Blenis 2009; Rodriguez-Viciana et al. 1994; Sarbassov et al. 2005; Staal 1987; Vanhaesebroeck et al. 2010) (Fig. 3). Note that activation of the PI3K/AKT pathway also occurs independently of RAS through direct and indirect binding of RTKs (Castellano and Downward 2011). The ultimate outcome is regulated induction of cell survival, growth, and migration. Components of these main signaling cascades are often mutated in cancer, with approximately one-third of human tumors expressing a constitutively activated mutant form of the RAS protein (Stephen et al. 2014). For a more extensive discussion of RAS signaling and targeting in cancer, the reader is referred to the following excellent recent reviews (Stephen et al. 2014; Prior et al. 2012; Cox et al. 2014; Pylayeva-Gupta et al. 2011; Malumbres and Barbacid 2003).

RET and RAS pathways have emerged as predominant oncogenic pathways for the C-cell; however, mutation of these genes currently only explains only about 60 % of MTC. The search for other major drivers of transformation has been largely unsuccessful. This failure may be explained in part by the lack of large unbiased search for new drivers. To date, only a single study reporting exomic sequencing was performed, which included only 25 MTC tumors (Agrawal et al. 2013). An alternative explanation is that a significant proportion of C-cell oncogenesis is driven by gene copy number changes rather than gene mutation. A role for copy number changes is supported by high-density comparative genomic hybridization (CGH) array studies (Flicker et al. 2012; Ye et al. 2008) and the unexpected MTC phenotype observed in numerous mouse models with allelic loss (see Table 1). From these studies, the retinoblastoma tumor suppressor (RB1) pathway has emerged as a route to C-cell oncogenesis in mice and should be considered a potentially important pathway in humans. Indeed, a recent association study specifically identified cell cycle genes in MTC development (Barbieri et al. 2014). The pathway primarily is made up of cyclin-dependent protein kinases (CDKs), cyclin-dependent kinase inhibitors (CDKNs), Cyclins (A, D, and E), RB proteins, and the E2F—family of transcription factors (Knudsen and Wang 2010) (Fig. 4). The pathway plays a critical role in cell cycle checkpoint regulation, which safeguards appropriate cell division; alterations in one or more of various components may lead to aberrant cell proliferation and development of cancer. Cell cycle progression is positively regulated by Cyclins activated by CDKs and negatively regulated by CDKNs. The activity of both is in turn regulated by environmental and cellular signals. For example, MAPK signaling leads to increased Cyclin D expression, while p53 activation enhances p21 function. The CDKN family members thus play a central role in preventing inappropriate cell division through negative CDK regulation and G1 cell cycle arrest. The CDKNs are made up of two major families—the INK4/CDKN2 family, which includes p15, p16, p18, and p19, and the Cip/Kip/CDKN1 family that includes p21, p27, and p57. The absence of CDKN suppression leads to increased phosphorylation of RB, which disrupts an inhibitory RB-E2F interaction. Once E2F is free of RB suppression, it functions to transcribe genes that are required for progression through S phase. The aberrant loss of CDKNs and RB has been associated with the development and progression of numerous cancer types and thus these proteins serve the role of tumor suppressors (Knudsen and Wang 2010; Zhu et al. 2014; Chinnam and Goodrich 2011; Di Fiore et al. 2013). Given the fundamental role that the RB1 gene played in the establishment of “two-hit” hypothesis of retinoblastoma (Knudson 1971), it was surprising that knockout of Rb1 in mice was initially associated with the development of pituitary tumors and subsequently CCH (Jacks et al. 1992; Williams et al. 1994; Harrison et al. 1995). Indeed, it is clear from several studies that Rb1 plays a central role in mouse C-cell tumorigenesis (see next section). Despite this observation, a direct role for RB1 loss in human C-cells has not been found. Studies that including RB1 sequence analysis failed to find mutation in 46 MTC tumors examined (Agrawal et al. 2013; Kanagal-Shamanna et al. 2014; Simbolo et al. 2014). A separate study found positive RB1 staining in 46 MTC tumors (Holm and Nesland 1994). When considered together, all these studies suggest that RB1 inactivation specifically induces C-cell oncogenesis but that mutation of the RB1 gene is a low-frequency event.

In a similar manner, tumor modeling in mice has implicated roles for the RB1 regulatory CDKNs: p18 (CDKN2C), p21 (CDKN1A), and p27 (CDKN1B) in C-cell oncogenesis (see below). A role for CDKNs in MTC is supported by the observation that genetic loss of chromosome 1p (p18) is a frequent event in MTC (Mathew et al. 1987) and that reproducible losses targeting other CDKN genes are observed in recent CGH array studies (Flicker et al. 2012; Ye et al. 2008). Indeed, loss of the p18 region has been seen in 20 % of MTC tumors examined by CGH. A direct role for specific p18 loss in human MTC is also supported by a single study that found p18 mutations in 4 of 30 tumors examined (van Veelen et al. 2009). A caveat of these studies in that mutations were observed in patients with RET driver mutations, two of which were germline C634W mutations arguing against a driver role in human tumorigenesis. Additionally, an earlier study that included 50 MTC tumors found no p18 mutations (Holm and Nesland 1994), while a recent exomic study that included 17 tumors also found no mutations in any of the CDKN genes (Agrawal et al. 2013). The incongruity that exists between CGH and mutation studies suggests that as observed in mice, CDKN haploinsufficiency is sufficient to promote C-cell oncogenesis, or the inactivation of the remaining allele by mutation is a rare event. In support of a haploinsufficiency model, reduced p27 expression is significantly correlated with larger MTC tumor size and elevated serum calcitonin (Ito et al. 2005). Further investigation is needed to elucidate the role of the Rb pathway in MTC oncogenesis, a worthwhile endeavor given that targeted therapies, which function through direct CDK inhibition, are being developed to exploit these paths across many human cancers.

The first widespread animal models of MTC were the Long–Evan, Wag/Rij, and Sprague Dawley rat strains. The high incidence of spontaneous MTC tumors in these mice provided the tools for development of transplantable tumors and establishment of MTC cell lines (Muszynski et al. 1983; Zeytinoglu et al. 1980, 1983). However, despite the high frequency of MTC in these rat strains, the underlying genetic cause has only been identified in Sprague Dawley rats as mutation of the p27 gene (Cdkn1b) (Pellegata et al. 2006). For the Long–Evans and Wag/Rij rat strains, we can only say that germline mutations of the RET gene do not play a role (Fritz et al. 2002; De Miguel et al. 2003). With the development of tools to manipulate the mouse genome, rats were largely replaced as disease models for many disorders including MTC. This section focuses on engineered mouse models of MTC. A summary of mice with a reported MTC phenotype is provided in Table 1.

Unlike some cancers, a clear stepwise route to C-cell oncogenesis has not emerged. Hereditary MTC demonstrates that activating mutations of RET serve as a primary initiating event. This role of RET is further supported by the observed high prevalence (~40 %) of RET mutations in sporadic MTC (Hu and Cote 2012). Several models have been proposed for how RET might function in initiation of C-cell oncogenesis. First, inactivation of RET in MTC cell culture models causes a marked reduction in tumor cell growth supporting a mechanism of oncogene addiction (Drosten and Putzer 2006; Vitagliano et al. 2011). This model is in part supported by patient responses to targeted therapies (Elisei et al. 2013; Wells et al. 2012). A second proposal is that RET increases genetic instability. In patient tumors, RET mutation is associated with increases in copy number variants and somatic mutations (Agrawal et al. 2013; Flicker et al. 2012; Ye et al. 2008). In one report examining sporadic MTC tumors using exomic sequencing, the mutation number was more than doubled in tumors with RET mutation (21.9 in RET+ vs. 8.4 in RET−) (Agrawal et al. 2013). In hereditary tumors, somatic mutation identified by exomic sequencing appears to be less frequent (Cai et al. 2015). Finally, a role for RET in altering the stem cell population needs to be considered in both hereditary and sporadic diseases. RET is known to play a central role in regulation of human spermatogonial stem cell self-renewal (Wu et al. 2009) and neural crest stem cell lineage derivation (Iwashita et al. 2003). In fact, somatic RET M918T mutation in male germ cells is associated with enhanced stem cell proliferation and provides a plausible mechanism for the paternal origin of MEN2B (Choi et al. 2012). While it remains to be formally examined, germline mutation of RET is also predicted to increase the neural crest progenitor cell population, thereby providing a larger pool of tumor initiating or the so-called cancer stem cells. A thyroid cancer stem cell model has been proposed for both medullary and nonmedullary thyroid cancers (Thomas et al. 2008; Zhang et al. 2006; Hardin et al. 2013). Support for a role in MTC derives from the observations that C-cells or their progenitors play a role in adult thyroid regeneration (Ozaki et al. 2012) and that RET mutation is associated with EGF/FGF-independent (Zhu et al. 2010), as well as 5-fluorouracil-resistant MTC tumorsphere growth (Kucerova et al. 2014).

Because MTC occurs in the absence of RET mutation, other pathways to C-cell oncogenesis must exist. Activation of RAS gene family members has emerged as a second oncogenic pathway with mutations observed in approximately 15 % of sporadic MTCs examined (Moura et al. 2011; Agrawal et al. 2013; Boichard et al. 2012; Ciampi et al. 2013; Schulten et al. 2011). A specific role for RAS as an initiator derives from the mouse models previously discussed, as patients with germline HRAS mutations (Costello syndrome) or KRAS/NRAS mutations (Noonan syndrome) have not been reported to develop MTC (Rauen 2013). Finally, no single gene has emerged as a common driver responsible for MTC without RET or RAS mutation. However, a focused examination remains to be performed. First, a recent comparison of Sanger-based sequencing with Ion Torrent AmpliSeq found a 30 % increase in the number of RET mutations detected (Simbolo et al. 2014). Furthermore, we have recently identified a case of RET gene fusion as a driver of tumorigenesis (Grubbs et al. 2015). RET fusions would go undetected by most DNA sequencing approaches. At the time of writing, only 5 such MTC tumors had been examined by exomic sequencing (Agrawal et al. 2013). The number of reported mutations in these 5 tumors ranged from 7 to 10, with only the TDG gene found mutated in more than one tumor. Unfortunately, Sanger sequencing could not confirm these specific TDG mutations, and TDG mutation was not in a separate larger validation set. While the small sample size may have prevented the discovery of other MTC genetic drivers, it is also possible that noncoding RNAs or gene copy number changes went undetected because of the limitations associated with the sequencing approach (Berindan-Neagoe et al. 2014; Yang and Lu 1839; Davoli et al. 2013; Manikandan et al. 2012). Indeed, in mice, heterozygous loss of Rb1 or p27 in the absence of a second mutation of the remaining allele is sufficient to induce MTC (Williams et al. 1994; Harvey et al. 1995; Park et al. 1999; Zhou et al. 2005). Although clear links establishing a role for cell cycle regulators in human C-cell oncogenesis do not exist, there is supportive evidence. Copy number changes that would not be detected by exomic sequencing have been observed using CGH arrays (Flicker et al. 2012; Ye et al. 2008). There have also been association studies linking increased MTC risk to genes associated with cell cycle regulation (Barbieri et al. 2014; Ito et al. 2005; Pasquali et al. 2011; Sekiya et al. 2014), although they are not universally positive (Sekiya et al. 2014). These observations fit well with a newer model, suggesting that heterozygous gene loss is sufficient to drive cancer (Davoli et al. 2013; Manikandan et al. 2012). The observation that MTC tumorigenesis in mice is significantly enhanced when targeting multiple Rb1 pathway members, particularly in combination with the RET/RAS pathway, suggests that progression and aggressive cancer might be caused by simultaneous hits to both pathways.

The production and secretion of calcitonin in response to increased serum calcium remains as the only well-defined role of the thyroid C-cell. The specific linking of the C-cell as the origin of MTC through calcitonin production has served as the primary motivator for understanding the biology of this cell. Ironically, this role has proven to be physiologically redundant, as genetic ablation of calcitonin is relatively inconsequential (Hoff et al. 2002). Unfortunately, little is known about other possible physiological roles of the C-cell and the mechanisms behind its unique sensitivity to RET-mediated oncogenesis. Genetic modeling in mice is beginning to gain insight into both processes. Studies of thyroidal development have suggested a possible interactive relationship between C-cells and follicular cells within the thyroid microenvironment (Andersson et al. 2011). Tumor modeling in mice has demonstrated that site-directed mutation of MEN2B-equivalent M919T alone is insufficient to cause MTC, but does induce pheochromocytoma. Models with unexpected MTC phenotypes have also linked oncogenesis to the Rb1-mediated pathway of cell cycle regulation. Together these pathways appear to synergize in a manner that appears to preferentially target the C-cell. Whether it is these cellular pathways that provide the C-cell with specific sensitivity to oncogenesis remains an unanswered question. Indeed, a role for RET signaling in normal C-cell function remains to be elucidated. In the nervous system, RET plays an essential role during development and in neuronal survival. A parallel role may exist for the C-cell, or perhaps like for the spermatogonial stem cell, RET functions as a C-cell signal for stem cell maintenance (Mulligan 2014). Roles for C-cells and calcitonin in thyroid regeneration and goiter formation have been proposed (Ozaki et al. 2012; Lupulescu 1972). Such a role would certainly fit with dysregulation of these pathways being associated with oncogenesis. Thus, while we have an ever-increasing knowledge base linked to the cellular drivers of C-cell oncogenesis, the primary role of these genes in normal C-cell physiology remains unclear. A greater understanding of the normal function of these genes is likely to contribute to defining the differences observed in the biology of MTC between mouse models and human, as well as offering greater insight into pathways of tumorigenesis beyond aberrant RET activation.