Personalized Therapy in Patients With Anaplastic Thyroid Cancer: Targeting Genetic and Epigenetic Alterations

Abstract

Context:

Anaplastic thyroid cancer (ATC) is the most lethal of all thyroid cancers and one of the most aggressive human carcinomas. In the search for effective treatment options, research toward targeted, personalized therapies is proving to be a path with great potential. As we gain a deeper understanding of the genetic (eg, BRAF^V600E, PIK3CA, TP53, hTERT mutations, etc) and epigenetic (eg, histone methylation, histone de-acetylation, microRNA regulatory circuits, etc) alterations driving ATC, we are able to find targets when developing novel therapies to improve the lives of patients. Beyond development, we can look into the effectiveness of already approved targeted therapies (eg, anti-BRAF^V600E selective inhibitors, tyrosine kinase inhibitors, histone deacetylase inhibitors, inhibitors of DNA methylation, etc) to potentially test in ATC after learning the molecular mechanisms that aid in tumor progression.

Design:

We performed a literature analysis in Medline through the PubMed web site for studies published between 2003 and 2014 using the following main keywords: anaplastic thyroid cancer, genetic and epigenetic alterations.

Objective:

Here, we outlined the common pathways that are altered in ATC, including the BRAF^V600E/ERK1/2-MEK1/2 and PI3K-AKT pathways. We then examined the current research looking into personalized, potential targeted therapies in ATC, mentioning those that have been tentatively advanced into clinical trials and those with the potential to reach that stage. We also reviewed side effects of the current and potential targeted therapies used in patients with advanced thyroid cancer.

Conclusions:

DNA and RNA next-generation sequencing analysis will be fundamental to unraveling a precise medicine and therapy in patients with ATC. Indeed, given the deep biological heterogeneity/complexity and high histological grade of this malignancy and its tumor microenvironment, personalized therapeutic approaches possibly based on the use of combinatorial targeted therapy will provide a rational approach when finding the optimal way to improve treatments for patients with ATC.

Anaplastic thyroid cancer (ATC) is the least common but the most aggressive of all thyroid cancers, with a median survival rate of 3–5 months (1, 2). It is thought to develop from existing papillary thyroid cancer (PTC) or follicular thyroid cancer (FTC); once ATC is established, it has an extremely high proliferative rate, it can quickly invade the neck structures and metastasize to other organs, and, more importantly, it shows resistance to radioiodine treatment (1, 3–6). ATC displays a multitude of morphological patterns, commonly presenting itself with bizarre spindle, giant, and squamoid tumor cells (7). When observed either through a biopsy or tracheostomy, it is found to be a “rock hard” mass, testing positive for keratin (1). Given the severity of the disorder, understanding genetic alterations that drive tumor progression is important when determining targets for treatment.

Genetics of ATC

There have been a multitude of genetic alterations associated with ATC, most often causing dysfunction in the ERK1/2-MEK1/2 and PI3K-AKT signaling pathways (Figure 1) (8). A mutation that has been the focus of a lot of research is BRAF^V600E. This point mutation leaves BRAF, a serine/threonine kinase involved in ERK1/2-MEK1/2 signaling, constitutively active, and it is thought to be involved with the progression of PTC to ATC and angioinvasion (6, 9–16). To cause this progression, BRAF^V600E induces changes in the tumor microenvironment, promoting tumor invasion and metastasis (17, 18). The BRAF^V600E mutation occurs in about 38% of ATC (13), and recent reports indicate that BRAF^V600E alters ATC tumor microenvironment through extracellular matrix (ECM) protein such as thrombospondin-1 (TSP-1) and ECM receptors (ie, integrins). Thrombospondin-1 is fundamental in ECM remodeling, and it has prometastatic properties in ATC, making research focusing on its mechanisms in thyroid cancers an important endeavor (11, 19).

Figure 1.

Genetic and epigenetic alterations in human ATC are potential targets for treatment options. The BRAF^V600E/ERK1/2-MEK1/2 and PI3K-AKT pathways are commonly mutated in ATC, along with epigenetic alterations on the histone proteins associated with DNA. PLX4720, PLX4032 (vemurafenib), and dabrafenib can inhibit the BRAF^V600E oncoprotein. MK2206 is an AKT inhibitor, halting downstream proliferative signaling. Imatinib works on the level of receptor tyrosine kinases (RTK) such as PDGFR. Axitinib and sorafenib inhibit angiogenic signaling through VEGFR2. PXD101 inhibits HDACs, which alter gene expression to promote tumor progression.

There are a multitude of other mutations that lead to increased PI3K-AKT signaling pathway activity. PIK3CA, which encodes the p110α catalytic subunit of PI3K, has been found to be mutated in 12–23% of ATC and in 6–13% of FTC (20), whereas this mutation is rarely seen in PTC (1–3%) (10, 20, 21). Importantly, the overactivation of the PI3K-AKT pathway can be induced by PTEN inactivation, including PTEN promoter methylation, deletion, or point mutations that occur in 10–20% of ATC (10). Furthermore, RAS mutations seem to preferentially activate the PI3K-AKT pathway and were found to be mutated in ATC 17% (NRAS mutations) and 6% (HRAS) of the time (13).

In 11% of ATC patients, a mutation in anaplastic lymphoma kinase (ALK), a protein involved with the activation of both the ERK1/2-MEK1/2 and PI3K-AKT pathways, is observed. Specifically, two point mutations have been identified (ALK^L1198F and ALK^G1201E) that result in increased tyrosine kinase activity and subsequent overactivation of the aforementioned pathways (22). On the opposite end of the spectrum, a negative modulator of these pathways, RASAL1, has been found to be mutated in about 16% of ATC, and the mutations lead to its subsequent inactivation (8). This gene is also found to be inactivated by epigenetic processes. When used in vitro and in vivo, RASAL1 inhibited thyroid cancer growth. Importantly, this was observed in human ATC cells (8). Specifically, the levels of phosphorylated AKT and ERK were decreased, indicating that RASAL1 has a direct effect in the pathways mentioned above (8). The report of this inactivation is significant because it helps bring validity to the hypothesis that RASAL1 is a true tumor suppressor gene in thyroid cancers (8).

The TP53 gene is another gene commonly mutated/inactivated in a variety of advanced human cancers and strongly involved in the ATC pathogenesis. It is mutated (12–83%) in ATC and rarely in well-differentiated thyroid carcinomas (eg, PTC), and the protein is also aberrantly overexpressed in ATC (20), causing inactivation of apoptosis and cell cycle progression. Importantly, a recent development in our understanding of ATC genetics is the mutations of the human telomerase reverse transcriptase (hTERT) promoter. This gene encodes for the catalytic domain of telomerase (23), and there are two different mutations that have been reported: C228T and C250T. About 33–50% of ATC harbor the C228T mutation in their hTERT promoter, and this is believed to result in an increase in telomerase activity (24), suggesting that it can play a role in the aggressiveness of thyroid carcinomas (24–28). The C228T mutation has been found to be associated with BRAF^V600E (29), and coexistence of these two genetic markers is a way to define the most aggressive subgroup of PTC (27–29). In PTC, the rate of recurrence is eight times greater when harboring both of these mutations, as compared to patients who lack both mutations (28). Also, the co-occurrence of hTERT C228T mutation and BRAF^V600E mutation confers a significant PTC aggressiveness and worst clinicopathological outcome (28). The synergy between the hTERT C228T and BRAF^V600E mutations may cause overexpression of an overactive hTERT, which by lengthening telomeres helps thyroid cancer cells evade apoptosis and promote cell proliferation through dysfunctional ERK1/2-MEK1/2 signaling. Given the correlation with tumor aggressiveness observed in PTC, the findings with this mutation coexistence of hTERT and BRAF bring about the possibility that the BRAF^V600E mutation with hTERT promoter mutations could be a genetic mechanism that contributes in the tumor progression from PTC to ATC.

Finally, it has also been recently described that thioredoxin-interacting protein (TXNIP) can play a role as a tumor suppressor in thyroid cancer cells. The low expression level of TXNIP in ATC is distinct when compared to differentiated thyroid cancers because they express high endogenous levels of the protein (30). Its down-regulation could, therefore, be important in the transition from differentiated thyroid cancer to ATC (30). The difference in TXNIP expression is because of the tumor suppressor's negative regulator peroxisome proliferator-activated receptor-γ (PPAR-γ). High PPAR-γ levels are observed in ATC, whereas low expression is seen in differentiated thyroid cancer (30). By contrast, the forkhead box protein M1, a member of the forkhead box family of transcription factors, has been recently identified as a novel oncogene involved in the control of ATC cell proliferation, invasion, and metastasis (31). These discoveries in ATC represent an important achievement that will improve and personalize therapy in patients with this type of aggressive malignancy.

Overall, understanding the genetic alterations of ATC is important when looking for targeted therapies that will be effective in patients who harbor this aggressive malignancy. Insights to this subject will enable us to accurately target dysfunctional signaling pathways, which could lead to improved clinical prognoses for patients.

Epigenetics and MicroRNAs (miRNAs) in ATC

Examination beyond genomic DNA alterations shows us that there are epigenetic factors that contribute to tumor progression in ATC. Enhancer of Zeste homolog 2 (EZH2) is a histone lysine-methyltransferase and is overexpressed in ATC (32). EZH2 is a member of the polycomb family of proteins, which in general are essential for the regulation of cell proliferation and differentiation. When EZH2 is overexpressed, histone methylation is altered, silencing the PAX8 gene and leading to an aggressive phenotype and poor clinical outcome (32, 33). Increased histone methylation (Figure 1) has been found to alter the expression of other genes in ATC cell lines, including p16^INK4A, DAPK, UCHL1, MGMT, TSHR, PTEN, and MAGE-A4 (20, 32, 34, 35). Many of the genes mentioned above play roles in tumor suppression, apoptosis, cell cycle regulation, and DNA repair, so with aberrant methylation at these sites, cells lose proper regulation in a multitude of areas (20, 32). Overmethylation of the TSHR gene leaves the cell incapable of concentrating iodine, rendering it resistant to radioiodide therapy and thus making the malignancy more difficult to treat (32). Additionally, histone deacetylation plays a role in ATC pathogenesis. Most ATC tumors show overexpression of histone deacetylases (HDACs), and less acetylation of histones leads to an altered expression of proteins controlling the cell cycle and proliferation (Figure 1) (36). Specifically, an overexpression of HDACs contributes to the dysregulation of both the ERK1/2-MEK1/2 and PI3K-AKT pathways (36), suggesting how these important pathways are not only altered by the aforementioned genetic mutations.

Similar to histone modifications, miRNA molecules also have the capability to alter the gene expression of ATC cells, thus making them worthy of discussion when looking into targeted therapies. miRNAs are a group of small (19–25 nucleotides) RNA molecules that can alter gene expression in a post-transcriptional manner by eliciting regulatory circuits. They function as negative regulators of the expression of protein-encoding genes involved in major processes such as development, apoptosis, cell proliferation, immune response, and hematopoiesis (32). With such a biological role, improper regulation of miRNAs can contribute to tumor progression, and this has been observed in ATC. Specifically, there are two families of miRNAs that contribute to a tumor's invasiveness via down-regulation (miR-200 and miR-30) (32, 37, 38) or up-regulation (miR-20a) (39). For example, miR-30d, a negative regulator of beclin-1, shows reduced expression in some cases of ATC, contributing to the tumor cell's inability to undergo autophagy. A down-regulation of miR-30d can also indicate insensitivity to cisplatin, a platinum-based chemotherapeutic agent that induces apoptosis in tumor cells (40).

Collectively, studies of miRNAs in thyroid tumorigenesis hold promise to improve our understanding of the biology of ATC as well as making progress in clinical prognosis.

Targeted Therapy Against Human ATC

There have been many different targeted therapies used in ATC, some not showing any success but others demonstrating a significant amount of potential. Sorafenib, a broad spectrum kinase inhibitor (multi-tyrosine kinase inhibitor [TKI]), including vascular endothelial growth factor receptor 2 (VEGFR2) inhibition (Figure 1), was recently approved by the US Food and Drug Administration (FDA) for radioiodine-resistant metastatic differentiated thyroid cancer (41). However, it did not show a significant therapeutic effect in patients with ATC, leading to additional side effects that include cardiovascular toxicity and dyspnea (Table 1) (42, 43). In particular, patients with PTC or FTC treated with sorafenib in phase II clinical trials showed either stable disease or partial responses (44, 45), whereas patients with ATC showed rapid clinical progressive disease after drug administration, making it less useful for this malignancy (44, 45). Axitinib (Figure 1), a VEGFR2 inhibitor (46), showed therapeutic efficacy only in one of two patients with ATC (42). Pazopanib, another TKI, similarly to sorafenib, also failed to significantly improve the clinical outcome of patients with ATC (42, 47). It should be noted that some of these TKIs have shown efficacy in other thyroid malignancies (eg, PTC and FTC), suggesting that their effectiveness could be limited to some histological subtypes of thyroid cancer and, more importantly, could be related to their molecular profile (ie, genetic and epigenetic alterations) (42, 48), likely because in differentiated thyroid cancers we do not observe the vast complexity of genetic and epigenetic alterations witnessed in ATC. Also, one factor that might be a reason for these TKIs to show little clinical significance in ATC is the inability to administer doses at high enough concentrations (20). Dosing of these drugs, in general, is something that needs to be worked out because different plasma levels of administered TKIs are seen among patients who received the same doses (42). With such a complex pathogenesis, many signaling pathways lose proper regulation, giving the tumor the ability to gain drug resistance. To help mitigate the lack of TKI responses in ATC, future preclinical work can test combinational therapies as a way to target many of the deregulated pathways at once, giving greater odds of responsiveness. That being said, detailed descriptions of how such drug resistance comes about are yet to be revealed. Research exploring this could give insight to disease progression that leads to an ATC phenotype.

Table 1.

Common Adverse Events of Prospective Targeted Therapies in Patients With Advanced Thyroid Cancers

Targeted Therapy (Agent)	Common Side Effects	First Author, Year (Ref)
Genetic
Sorafenib	Dyspnea, cardiovascular toxicity, diarrhea, weight loss, squamous cell carcinoma	Harris, 2011 (42); Cabanillas, 2010 (62)
Axitinib	Fatigue, weight loss, diarrhea, nausea, hypertension, mucositis, stomatitis, proteinuria	Harris, 2011 (42); Cohen, 2008 (46)
Vemurafenib	Fatigue, nausea, diarrhea, arthralgia, dermatitis, squamous cell carcinoma, alopecia	Swaika, 2014 (63)
Imatinib	Edema, fatigue, nausea, myalgia, hyperglycemia, lymphopenia, anemia, hyponatremia, abnormal liver function	Ha, 2010 (58)
Selumetinib	Fatigue, maculopapular rash, acneiform rash, edema, elevated liver enzymes	Ho, 2013 (59); Hayes, 2012 (64)
Dabrafenib	Fatigue, squamous cell carcinoma, pyrexia	Azijli, 2014 (60)
MK-2206	Rash, nausea, vomiting, fatigue, hyperglycemia	Yap, 2011 (65)
Epigenetic
PXD101 (belinostat)	Nausea, vomiting, dyspnea, edema, hematological toxicities, flushing	McDermott, 2014 (66)

Current research into potential targeted therapies has begun using specific mouse models that can more closely mimic the disease than previous in vivo systems. First, there are genetically engineered mouse (GEM) models that have been used in an effort to define the specific role a mutation plays in tumorigenesis (49). Specifically, GEM models have been used to help understand the role of BRAF^V600E. The BRAF^V600E GEM model developed tumors that appeared similar to poorly differentiated PTC (50). When this transgenic mouse was treated with BRAF^V600E inhibitors (ie, PLX4720), iodine uptake was detected (50). GEM models are useful when trying to understand the biology of specific mutations, even if they have high costs for their development (49). Another preclinical strategy to validate targeted therapy can be to develop orthotopic/xenograft mouse models that allow the study of human thyroid tumors in vivo (49). Orthotopic models of ATC that have been devised also have the capability to develop metastases, both local and distant, to more closely mimic clinical presentation (49, 51). This model has been used in successful preclinical trials aiming to inhibit BRAF^V600E with very specific inhibitors (12, 18, 51). PLX4720 is one of these very specific inhibitors, and in 2010 it was reported to inhibit tumor growth and metastasis in vivo using a human ATC cell line harboring the BRAF^V600E and TP53 mutations (12). PLX4720 has also been shown to significantly reduce tumor size in both early and late stages (tumor regression model) of human ATC in vivo (52, 53) (Figure 1). Another treatment option that has a very similar inhibitory effect is PLX4032, also named vemurafenib. When discovered in 2010, it was shown to be effective in most metastatic melanoma patients with the BRAF^V600E mutation (54). A recent study used vemurafenib in an in vivo system similar to an orthotopic model that used human ATC cells transfected with a luciferase reporter gene to make metastases more easily detectable (51). The researchers demonstrated that vemurafenib was able to slow tumor progression in their model, indicating that it could be an effective agent against ATC (51). There is yet to be any preclinical work using an orthotopic mouse model with this treatment, but such a study could help validate this as a functional therapeutic in ATC. The most promising results in humans with the inhibitor to date were in a report that indicated the treatment significantly improved (eg, by reducing lung metastasis) the outcome of an ATC patient, demonstrating potential clinical significance (55). Beyond the BRAF^V600E mutation, researchers have found success with this model looking into inhibiting the Src family kinases with dasatinib, a potential targeted therapy for ATC that is already FDA approved for patients with imatinib-resistant chronic myelogenous leukemia (56). Treating both ATC and PTC cell lines with dasatinib caused a significant induction of caspase-3/-7, indicating the agent's ability to induce apoptosis (56). Mechanistically, it also inhibited ERK1/2 activation by decreasing phosphorylation (56), considered as a potential first readout of many therapies used to treat thyroid malignancies, and partially blocked ATC proliferation and growth. The results from this study provided the first evidence that Src is a central mediator of PTC growth and metastasis, indicating that Src inhibitors may have a higher therapeutic efficacy in PTC as both antitumor and antimetastatic agents (56).

Future Directions

Currently, there are some targeted therapies being tested for their effectiveness in ATC, but they have not yet reached a clinical setting. Specifically, there has been an attempt to target the epigenetic alterations that come about with ATC. PXD101 (also known as belinostat) is an HDAC inhibitor that antagonized tumor progression of ATC in vivo, giving another prospective route for improving the lives of patients (36). Clinical validity of this drug is yet to be determined, and given the primary results, clinical trials with PXD101 could be a viable option. Another possible route would be to inhibit histone-lysine methyl-transferases, and this option should be explored when trying to alter chromatin structure in ATC. Some targeted therapies have begun to be tested in ATC, but their value is yet to be accurately assessed because of a lack of data. Imatinib, a TKI that acts on c-ABL and platelet-derived growth factor receptor (PDGFR) tyrosine kinases, was reported to have antitumor activity in ATC cell lines, making it a prospective treatment option for patients with this type of cancer (57). When used in a phase 2 clinical trial, this treatment showed some antitumor activity, with two of eight patients showing partial responsiveness; however, further trials with a larger group of patients will help determine its validity as a targeted therapy in ATC (58). There are other targeted therapies that have been tested in other types of human thyroid cancers, but these are yet to be assessed in ATC; eg, the MEK1/2 inhibitor selumetinib was used in patients with differentiated thyroid cancers that were refractory to radioiodine treatment and the effectiveness of radioiodine uptake was found to increase specifically in patients who harbored RAS mutations (59). There is also a new BRAF^V600E inhibitor that has been approved by the FDA, ie, dabrafenib, which has shown success in melanoma patients, inhibiting tumor progression and even acting on metastases (60). There is yet to be any research with this treatment in any type of thyroid cancer, but given its success in melanoma, work needs to be done to determine whether it is a viable option for ATC patients. Another target that should be considered in ATC is the PI3K-AKT pathway. MK2206, an AKT inhibitor, was shown to inhibit thyroid cancer cells in vitro (61). Future work could use this inhibitor in an in vivo ATC model to determine whether it has clinical potential. The search for new therapies could also look toward inhibiting overactive hTERT, another genetic alteration recently observed in ATC. Overactive hTERT correlates with tumor aggressiveness, and thus targeting it is a reasonable option for future research. For the current and potential targeted therapies mentioned above, side effects must be considered to determine the true potential in patients, and these are summarized in Table 1.

In summary, using specific targeted therapies in preclinical models against this lethal human carcinoma (ie, ATC) has shown to be a promising path for treatment in patients with ATC. Continuing translational research that will be crucial to make deep genomic analyses and optimize combinatorial targeted therapy approaches in patients with ATC will help develop a precise medicine and decrease the currently dismal mortality rates.