Alterations of regulatory factors and DNA methylation pattern in thyroid cancer

Tumour cells might undergo various epigenetic changes, such as DNA hypermethylation, responsible for inducing an imbalance in pro- and anti-apoptotic processes regulation, and so influencing both progression and response to cancer therapy. Several studies have shown that 5-AzaC, a DNA demethylating agent, induces cell cycle arrest and cell death in various type of tumour cells [28, 29] by altering the epigenetic state, but the details of 5-AzaC-induced apoptosis or cell cycle arrest in cancer are not completely understood in thyroid cancer.

In order to evaluate the role of aberrant DNA methylation processes in thyroid cancer 5-AzaC, along with L-AA (known to have a demethylating effect) was used. Analysis of K1 tumour cells viability revealed that 5-AzaC treatment induced a significantly decrease in cell viability dependent of concentration and time (24 or 48 h), which was further enhanced by L-AA treatment combination (Fig. 3 and S1). The cytotoxicity increases with the exposure time, higher concentrations and when the cells are exposed to a combination of the two drugs.

OPEN IN VIEWER

Flow cytometric analysis showed that the early apoptosis rate of K1 tumour cells increased about 3–6-fold after exposure for 24 h or 48 h to 5 μ M and 10 μ M 5-AzaC, while the early apoptosis rate of K1 tumour cells increased only slightly (1.5-fold) upon treatment with 5 mM or 10 mM L-AA. Further analysis confirmed a synergy between L-AA and 5-AzaC effects in inducing apoptosis in K1 cell line, as evidenced by a robust increase in early apoptosis rate when the cells were co-treated with higher doses of (10 mM) L-AA and (10 μ M) 5-AzaC for 24 h, but this effect is no longer observed after 48 h treatment (Supplementary Fig. S1).

OPEN IN VIEWER

Cell cycle distribution analysis of K1 cells treated with 5-AzaC or L-AA individually indicated an increase in the population of cells in S phase after 24 h treatment. In addition, an increase in the S phase of the cell cycle was noted in cells treated 24 h with a combination of 5 μ M 5-AzaC + 5 mM L-AA without significant differences from the effect induced by each agent alone (Supplementary Fig. S2).

Furthermore, we observed an important increase of G1 phase in cells treated 48 h with both concentrations of L-AA used, accompanied by a decrease of S phase. Also, 5 μ M/48 h of 5-AzaC induces an increase of G2 phase, while 10 μ M/48 h of 5-AzaC determines cells arrest in G1 phase. When K1 cells were treated with the combination of the 5-AzaC and L-AA at higher concentrations the results indicate an increase in G2 phase of cell cycle compared with effect of either agent alone (Supplementary Fig. S2).

In K1 cells treated with both agents it was noted a significantly increase of TET1 and TET2 genes expression levels as compared with control. The best results were obtained when cells were treated with high concentrations of both substances (respectively 10 μ M 5-AzaC and 10 mM L-AA) and after 48 h (Fig. 4). Our results indicate that L-AA enhances the demethylation effect exercised by 5-AzaC this effect being noticeable especially in case of TET2 methylation (Fig. 4).

It has been found also, that combined treatments (5-AzaC and L-AA) lead to a significantly decrease in 5mC content, concomitantly with an increase in 5hmC levels in K1 cell line (Supplementary Fig. S3).

In recent years, a particularly high interest in cancer research has been attributed to epigenetic changes that occur in tumorigenesis. One of these modifications is DNA methylation and it is now acknowledged that aberrant methylation of genomic DNA represents a signature of the malignant process. Among the epigenetic events that occur in tumorigenesis are: local CpG islands hypermethylation which often takes place at genes promoter’s level (especially in the case of tumour suppressor genes) leading to loss of gene expression, and a global hypomethylation that is related with gene activation (e.g. oncogenes, DNA repetitive elements), loss of genomic imprinting and chromosomal instability [30]. In our previous study we found TP73, WIF1 and PDLIM4 genes promoters are hypermethylated in PTC [31].

Lately studies have highlighted the particular importance of epigenetic changes in thyroid cancer and their prognostic potential, among the most investigated being DNA methylation [32].

Regarding global DNA methylation patterns in thyroid cancer, the available data are rather contradictory and this is mainly due to different analysis methods used in studies. For example, it was showed that 5-methylcytidine immunostaining levels are more reduced in thyroid cancer samples as compared with benign or adjacent normal samples [33]. By contrast, Keelawat et al. found that 5-methylcytidine is significantly higher in well-differentiated thyroid cancer types (FA, FTC, cPTC) than in nontumoral adjacent tissue [34].

A study made by Rodríguez-Rodero et al. using DNA methylation arrays illustrated a different promoter methylation profile in thyroid cancer depending on cancer subtype and that promoter hypomethylation is more frequently found in non-differentiated subtypes [35].

In this study, investigating DNA methylation profile in thyroid cancer we found that PTC tumoral samples present significantly increased levels of 5mC and low levels of 5hmC. An important role for hydro-xymethylation has been showed in cancer where decreased levels of 5hmC are frequently detected [36], like in esophageal squamous cell carcinoma, parathyroid carcinoma, melanoma or glioma [37, 38, 39, 40].

It is now believed that an important feature of malignant transformation along with global reduction of 5hmC levels is the alteration of TET enzymes frequently found mutated or downregulated [10].

The present study revealed alterations in expression levels of TETs genes and DNMT1 in thyroid pathology. Data show that thyroid cancer samples display significantly reduced mRNA levels of TET1 and TET2 compared with benign samples but high mRNA levels of TET3 and DNMT1. Moreover, for PTC patients, TET1 lower levels correlate with BRAF V600E mutation and RET/PTC rearrangement. Decreased levels of TET1 and TET2 have been reported in numerous pathologies. For example, it was demonstrated that low levels of TET1 correlate with tumour grade, metastasis and a poor outcome in renal cancer patients as well in colorectal cancer [41, 42]. At the same time, an association between TET1, TET2 reduced expression levels and cancer progression and a poorer prognosis has been proved in endometrial and breast cancer [43, 44].

The current study showed a good correlation between 5mC and 5hmC levels and TETs expression levels in thyroid carcinoma patients. It was noticed that 5mC le-vels were significantly associated with TET1 and TET2 mRNA expression levels while 5hmC levels correlated with TET3 expression, thus suggesting a role for all these factors in thyroid cancer.

We highlighted for the first-time that higher degrees of methylation in TET1 and TET2 genes promoters are associated with a decrease in their expression levels in thyroid tumours.

The obtained results are supported by other reports that show that TET1 promoter is found hypermethylated in many types of cancer. One such example is breast cancer where TET1 hypermethylation is inversely correlated with its expression and associated with metastasis for which could serve as a potential diagnosis marker [45]. TET1 hypermethylation has been suggested as potential biomarker in colorectal cancer. This gene was found frequently methylated early in colorectal tumorigenesis and is associated also with BRAF mutation presence [46].

An important role in tumorigenesis has been reported for TET2 gene decreased expression or inactivation in haematological cancer [47, 48, 49]. A correlation between reduced levels of TET2 and 5hmC in cancer was found in esophageal squamous cell carcinoma (ESCC) [37] and in parathyroid carcinoma where TET2 levels are reduced due to its promoter hypermethylation [50]. Another mechanism for TET2 gene repression has been demonstrated and it is related to lysine demethylase-KDM2A action that targets and inhibits TET2 in breast cancer leading to DNA methylation and tumour progression [51].

Our study showed that the methylation of TET1 is slightly increased in FA group, but increases in malignant thyroid tumours. fvPTC patients presented lower methylation levels compared to cPTC and could be a discriminatory factor between two cancer types and benign lesions. TET2 is a poorer discriminator between FA and fvPTC, but it can be useful for cPTC identification.

In vitro experiments with demethylating agents demonstrated that the decreased TET1 and TET2 gene expression was due to aberrant gene promoter methylation, the treatments being able to increase the expression levels for both genes. Moreover, investigation of the effect of demethylating agents on K1 cell line showed that vitamin C can enhance the demethylation effect of 5-AzaC and this was best noticed for TET2 gene. Also, an increase in 5hmC levels and a decrease in 5mC content in K1 treated cells were observed. Our findings are consistent with previous studies in ES cells and melanoma cells, that indicate TETs activation and global increase of 5hmC content in treatments with ascorbic acid [52, 53]. The same effect was also noticed in HLE and Huh7 hepatocellular carcinoma cell lines where cells treated with both vitamin C and 5-AzaC displayed increased levels of TET2 and TET3 [54].

The present study shows that 5-AzaC can suppress in vitro viability, induces cytotoxicity, influences cell cycle phases distributions dependent of time and concentration, increases apoptosis of K1 tumour cells and these data may be correlated with the differences observed in the methylation pattern. L-AA supplementation amplified the 5-AzaC effect, reducing viability, increasing apoptosis and inducing G2-cell cycle arrest. Our finding has added value to the growing evidences of L-AA as an important co-factor which influences the activity of 5-AzaC and provides experimental support for novel therapies for thyroid cancer in order to ensure prolonged patient’s survival. Regarding the future therapy, further studies are necessary to evaluate the role of the combination of demethylating agents and optimal dose in vivo animal model.

5. Conclusions

The hypermethylation of TET1/TET2 genes promoters associated with down-regulation of their expression alters global DNA methylation pattern in thyroid cancer. These results may indicate global and local DNA methylation assessment as potential diagnostic tool for this pathology.

Our data demonstrated that L-AA is able to modify the activity of the 5-AzaC in K1 and induce cell cycle arrest which is considered essential in limiting tumour cell progression. The cumulative effect of L-AA and 5-AzaC as demethylating agents on methylation process, apoptosis and cell cycle could open new perspective for thyroid cancer therapy.