Abstract
Several key oncogenic pathways that initiate or promote malignancy in thyroid tissue have been identified. A large percentage of these are driven by somatic mutations that constitutively activate the RAS/RAF/MAPK pathway, which promotes uncontrolled cell proliferation, invasion, and survival. Mutations of BRAF and the tyrosine kinase, RET, are predominant in papillary thyroid cancer cases, while mutations in RAS family members are more common in FTC (4 –6). A high prevalence of RAS mutations, however, have also been noted in benign thyroid adenomas, indicating that a RAS mutation alone is likely not sufficient to induce transformation to the malignant state (7 –9).
Unlike the BRAF V600E mutation that results in the specific activation of the MEK/ERK pathway, activation of RAS regulates additional protumorigenic pathways. Of these, the phosphatidylinositol-3-kinase (PI3K) pathway is a major downstream pathway stimulated by RAS activation, resulting in increased growth and survival. In addition to activation by RAS family members, the PI3K pathway is constitutively activated by somatic mutations in genes encoding for the PI3K catalytic subunit (PIK3CA), or its downstream effector AKT1, as well as copy number gain in components of the pathway. A major negative regulator of the PI3K pathway is the lipid phosphatase and tensin homolog (PTEN), which antagonizes the pathway by dephosphorylating phosphatidylinositol-3,4,5-triphosphate. Thus, inactivation of PTEN, results in activation of the PI3K pathway, which is a common event in cancer (4,9 –14).
Somatic mutations in the RAS and PI3K pathways are particularly common in FTC and ATC, and both can occur within the same tumor (4,9,12 –14). Recent studies have shown their increasing prevalence in the progression of well-differentiated thyroid cancer to ATC, suggesting that concurrent activation of both pathways may provide a selective advantage for the development of aggressive thyroid cancers (4,9 –13). Consistent with this is the finding that ∼80% of undifferentiated thyroid cancers harbor genetic alterations activating the MEK/ERK and the PI3 kinase pathways (11). Activation of the PI3K pathway is more prevalent in poorly differentiated thyroid cancers (4,9 –13), suggesting that this is a late event that may enhance transformation by increasing signaling of other key pathways, such as oncogenic RAS. The fundamental role for the RAS and PI3K pathways in thyroid tumorigenesis is further underscored by recent studies demonstrating synergistic inhibition of thyroid cancer growth and survival by simultaneously targeting both pathways with specific agents (15 –17). Thus, dual inhibition of the RAS and PI3 kinase pathways is a potentially powerful strategy for treatment of advanced thyroid cancer.
To further understand the role of RAS and PI3K signaling in the initiation and progression of thyroid cancer, a group led by Dr. Antonio Di Crisofano recently generated a mouse model with conditional physiologic expression of oncogenic Kras G12D (See Miller et al. [18] for references). Physiologic expression of the Kras G12D allele was achieved by targeting Kras G12D to its endogenous genomic locus with thyroid-specific expression being achieved by the human thyroid peroxidase promoter, which was expressed at day E14.5 (18). This study was the first report relating to a role for endogenously expressed oncogenic KRAS in thyroid cancer. Previous studies using thyroid cancer transgenic mouse models achieved expression of supra-physiologic levels of RAS, which does not necessarily recapitulate endogenous expression of oncogenic RAS (19 –21). Interestingly, when the product of oncogenic Kras G12D was expressed at physiologic levels in the thyroid, no alterations in the thyroid gland were observed, even after a follow-up period of 1 year (18). This contrasts with studies of other tumor types in which physiological expression levels of the same Kras oncogenic allele were associated with rapid induction of intraepithelial neoplasia in the pancreas (22), colonic hyperplasia (23), and lung hyperplasia (24).
Nonetheless, the lack of thyroid gland transformation in the Kras G12D mutant is not necessarily surprising since RAS signaling can lead to senescence in some cell types via a complex negative feedback mechanisms that function to downregulate the pathway (25 –27). Indeed, other studies (28) have shown that endogenous expression of mutant Kras failed to induce any detectable changes in the mouse small intestine after 1 year. Consistent with this, RAS mutations are frequently observed in benign thyroid adenomas, suggesting that activation of RAS is an early event in thyroid tumorigenesis and may require additional genetic events for malignant transformation (7 –9). Although the development of senescence was not investigated in the Miller et al. study (18), it would not be surprising if this tumor suppressor mechanism plays a role in the lack of transformation observed in the Kras G12D mice.
Since RAS can also activate the PI3K/Akt pathway by interacting with the p110 catalytic subunit of PI3K, and activation of both pathways plays a key role in thyroid cancer development, Miller et al. (18) further tested whether oncogenic Kras expression along with PI3K pathway activation via PTEN deletion was sufficient to transform thyroid epithelial cells in vivo. As previously noted, PTEN is a key negative regulator of the PI3K/AKT signaling cascade. Using this approach, Miller et al. (18) showed that all Kras G12D/Pten −/− double mutant mice rapidly develop FTC, with 50% of the mice dying at 7 weeks of age and none surviving for more than 4 months. Thyroglobulin-positive lung metastases developed in all of the mice that survived at least 12 weeks (18). Interestingly, serum thyroxine concentrations were increased and serum thyrotropin (TSH) concentrations were suppressed in the double Kras G12D/Pten −/− mutant mice. Although the reasons for these hormonal changes are not clear, the data for serum TSH suggest that thyroid tumor development in this mouse model is independent of TSH signaling (18). Notably and in contrast, BRAF V600E and the thyroid hormone β receptor (TRβPV/PV) transgenic mice exhibit elevated TSH signaling that may help drive thyroid tumorigenesis (29 –32).
Surprisingly the Di Cristofano group noted that while the downstream effector, MEK1/2 was activated in Kras G12D mice, there was little activation of ERK1/2, a direct downstream target of MEK1/2 (18). These results raise the possibility that MEK1/2 signaling may be independent of ERK1/2 in this model (18,33). Alternatively, mutational activation of KRAS in this model may result in the up-regulation of ERK1/2 phosphatases, creating a negative regulatory loop (23,34). Miller et al. further showed that in the Kras G12D/Pten −/− mice (e.g. with concurrent activation of PI3K and KRAS), ERK1/2 as well as MEK1/2 was activated, indicating that PI3K signaling blocks a putative negative feedback loop regulating ERK1/2 phosphorylation (18). A further observation in Kras G12D/Pten −/− mice was up-regulation of the cell cycle gene, Ccnd1, apparently resulting from cooperation between oncogenic KRAS and PI3K signaling (18).
To determine the specific roles of KRAS and PI3K in Kras G12D/Pten −/− mice, Miller et al. (18) tested the effects of selective inhibitors of PI3K (LY294002) and MEK1/2 (PD98059) using primary cell lines isolated from the tumors. Using this approach, they showed that inhibition of the KRAS-MEK pathway using the MEK1/2 inhibitor, PD98059, resulted in a modest inhibition of cell growth, while inhibition of PI3K with LY294002 was associated with a robust inhibition of cell growth. Concurrent inhibition of MEK1 and PI3K with PD98059 and LY294002 completely abolished the growth of the thyroid cancer primary cell cultures (18). To determine the permissive role of PI3K signaling in thyroid tumor progression in the intact Kras G12D/Pten −/− mice, treatment with the PI3K inhibitor was initiated at 3 weeks of age. Consistent with the cell culture studies, PI3K-inhibitor–treated mice had significantly longer survival than untreated mice (18). Taken together, these data demonstrate that concomitant PI3K activation is necessary for KrasG12D to induce thyroid transformation and metastases, which is consistent with mouse models of tumors in non-thyroid organs (35 –39). Furthermore, these studies indicate that pharmacological inhibition of the PI3K pathway may be a promising approach to treating RAS-mutant thyroid tumors, in addition to thyroid tumors with dual activation of the RAS and PI3K pathways.
In this issue of Thyroid, Dima et al. (40) extend these important studies with a report of the establishment and characterization of three novel mouse cell lines, T683, T691, and T826, derived from these poorly differentiated, double-mutant Kras G12D/Pten −/− thyroid carcinomas. The development of these novel transformed thyroid cell lines provide a key model to study the molecular mechanisms and pathways regulated by RAS and PI3K. In their study, Dima et al. (40) confirmed expression of the putative mutations and performed karyotype analysis, revealing a normal chromosome number. However, two of the cell lines exhibited distinct rearrangements in chromosome 4. This normal chromosome number contrasts with many human thyroid cancer cell lines, which are typically nondiploid, likely due to many years of in vitro culture growth.
To further characterize their new cell lines, Dima et al. (40) tested genes commonly involved in thyroid tumor progression, including the tumor suppressors, p16 and p19ARF. The T826 cell line exhibited a homozygous deletion of the entire locus, while the T691 cell line lost both exons 1 and 2, indicating that both cell lines have lost expression of the p16 and p19ARF tumor suppressors. The T683 cell line, on the other hand, was characterized by the presence of normal p16 and p19ARF loci. Similar to most undifferentiated thyroid tumors, Dima et al. (40) showed that all three cell lines express functional p53. Furthermore, they demonstrated that expression of key genes involved in thyroid differentiation and function were virtually absent, including Foxe1, Nkx2-1, Pax8, Duox-1/2, Nis, Pds, TG, Tpo, and Tshr, consistent with a poorly differentiated phenotype. Expression of the cyclin Ccnd1, but not Ccnd2 or Ccnd3, was also increased in all three double-mutant cell lines, consistent with expression levels observed the Kras G12D/Pten −/− mice (18,40). Similar to their previous study using primary cultures derived from the double mutant tumors (18), treatment of these permanent cell lines derived from the double mutant tumors with inhibitors of MEK1 (PD98059), PI3K (LY294002), or mTOR, a key effector of the PI3K pathway (RAD001), resulted in a partial inhibition of growth, while dual inhibition of both pathways was required to induce cell death (40). Thus, the permanent cell lines derived from the double mutant mice appear to retain the molecular features of the primary tumors from which they were derived. The double mutant cell lines further exhibit a highly glycolytic phenotype compared to wild-type thyrocytes, suggesting that energy production in these tumors has switched from an oxidative phosphorylation mode to a glycolytic mode, a common event in tumorigenesis, and another potential pathway for therapeutic intervention.
Permanent thyroid cancer cell lines remain a critical tool to understand the oncogenic mechanisms that drive thyroid tumorigenesis, as well as a necessary preclinical model to test new therapies using in vivo and in vitro approaches. The recent discovery that the majority of thyroid cancer cell lines (17 out of 40) are either redundant or misidentified with cell lines derived from other tumor types (e.g., melanoma and colon cancer), has further increased the need for new, validated thyroid cancer cell line models (41).
The development of cell lines derived from the Kras G12D/Pten −/− genetically engineered mice by Dima et al. (40) represents an important preclinical tool to study thyroid cancer biology in the context of two major oncogenic pathways, RAS and PI3K. Furthermore, implantation of these cell lines into immunocompetent mice will provide an opportunity for efficient orthotopic thyroid tumor formation as well as the development of metastases, providing an excellent model to study both thyroid tumorigenesis and metastasis in the presence of an intact immune system. Another advantage to these new cell lines is the stable introduction of the luciferase reporter gene which allows for noninvasive bioluminescence imaging of the primary tumor over time and the development and progression of metastases, the latter of which are typically difficult to monitor. Therefore, the creation of this new genetic model, which closely recapitulates the clinical progression of human thyroid cancer progression and metastases, provides a major new tool to study the mechanisms involved in the metastatic process and, ultimately, the development of new therapeutic strategies to inhibit thyroid cancer growth and metastasis.