Targeting EZH2 as a novel therapeutic strategy for sorafenib‐resistant thyroid carcinoma

Abstract Thyroid carcinoma is the most common endocrine malignancy. Surgery, post‐operative selective iodine‐131 and thyroid hormone suppression were the most common methods for the therapy of thyroid carcinoma. Although most patients with differentiated thyroid carcinoma (DTC) showed positive response for these therapeutic methods, some patients still have to face the radioactive iodine (RAI)‐refractory problems. Sorafenib is an oral multikinase inhibitor for patients with advanced RAI refractory DTC. However, the side effects and drug resistance of sorafenib suggest us to develop novel drugs and strategies for the therapy of thyroid carcinoma. In this study, we firstly found that patients with sorafenib resistance showed no significant change in rapidly accelerated fibrosarcoma and VEGFR expression levels compared with sorafenib sensitive patients. Moreover, a further miRNAs screen by qRT‐PCR indicated that miR‐124‐3p and miR‐506‐3p (miR‐124/506) were remarkably reduced in sorafenib insensitive patients. With a bioinformatics prediction and functional assay validation, we revealed that enhancer of zeste homolog 2 (EZH2) was the direct target for miR‐124/506. Interestingly, we finally proved that the sorafenib resistant cells regained sensitivity for sorafenib by EZH2 intervention with miR‐124/506 overexpression or EZH2 inhibitor treatment in vitro and in vivo, which will lead to the decreased tri‐methylation at lysine 27 of histone H3 (H3K27me3) and increased acetylated lysine 27 of histone H3 (H3K27ac) levels. Therefore, we conclude that the suppression of EZH2 represents a potential target for thyroid carcinoma therapy.

rate of the patients with locally advanced or metastatic RAI-refractory differentiated thyroid cancer is significantly reduced, which is the focus and difficulty in clinical diagnosis and treatment and needs to be further explored. 7,8 Sorafenib, an oral multikinase inhibitor, was initially developed to target rapidly accelerated fibrosarcoma (RAF) kinase in the MAPK pathway. 9 Additionally, sorafenib was found to have potent inhibitory roles of RTKs, VEGFR and RET kinases, leading to the approval of the United States Food and Drug Administration in November 2013 for patients with advanced RAI therapy refractory differentiated thyroid carcinoma (DTC). [9][10][11][12] Considering the side effects, cost and lack of overall survival benefit of sorafenib during the treatment of metastatic DTC patients, it is essential and urgent to develop novel drugs and strategies for the therapy of thyroid carcinoma. 9,13 The dysregulation of microRNA (miRNA) expression is involved in variety of human malignancies and emerging miRNAs have been demonstrated playing important roles during the tumorigenesis of thyroid carcinoma, including miR-199b-5p, 14 miR-214, 15 miR146b-5p, 16 miR-129 17 ; miR-124-3p 18 ; miR-506-3p 19 ; miR-150-5p 20 and miR-205. 21 Enhancer of zeste homolog 2 (EZH2), a subunit of polycomb repressor complex 2 (PRC2), is important for the gene transcriptional regulation through H3K27me3. 22,23 Many studies have revealed the inappropriate expression of EZH2 in various aggressive cancers, including thyroid carcinoma. 24,25 However, there is currently no information on the relationship between EZH2 expression and the cancer therapy.
In this study, we firstly focused on the gene expression profile between the sorafenib sensitive and resistant tumour tissues collected from 16 patients with thyroid carcinoma. Then, tumour cell lines were established from sorafenib sensitive and resistant thyroid carcinoma tissues to further confirm the resistance of sorafenib in vitro. To reveal the detailed molecular mechanism for the resistance of sorafenib in the sorafenib resistant patients, we also investigated the differential expressed miRNAs and their target genes between the sorafenib sensitive and resistant tumour tissues. Most importantly, we aim to discover viable therapeutic strategy to be offered to patients with thyroid cancer to decrease tumour burden and progression.

| Clinical samples collection
A total of 16 tumour tissues from patients with thyroid carcinoma used in this study were collected from the Shanghai Tenth People's Hospital, Tongji University School of Medicine hospital. Informed consent was obtained from the patient and all the tissue collection procedure was approved by Shanghai Tenth People's Hospital, Tongji University School of Medicine committee. Samples were put immediately into the pre-cold cell culture medium after surgery. Tissues will be divided into three parts for the RNA and protein extraction and the establishment of patients-derived cell lines.

| The establishment of patients-derived cell lines
Tumour tissues collected from clinic were subsequently rinsed in PBS containing penicillin and streptomycin and cut into small pieces.
The tumour fragments were then digested with 1 mg/mL collagenase, 0.02 mg/mL DNAase and 0.01 mg/mL hyaluronidase for 16-20 hours at room temperature. Then, the specimens were cultured using DMEM/F12 medium containing 10% FBS, thyroid stimulating hormone, triiodothyronine (T3), dermal growth factor (EGF) and hydrocortisone. The tissues were then kept at 37°C in a humidified atmosphere with 5% CO 2 and the patients-derived cell lines will be established a few days later.

| The miRNA mimics and transfection
The miR-124/506 mimics and the control (NC) sequence were synthesis in Shanghai Liangtai Biotechnology (Shanghai, China). Cells were seeded into 60 mm dish and the mimics were transfected into the cells 24 hours later with Lipofectamine ® 2000 Reagent (Life Technologies, Pleasanton, CA) according to the manufacture's protocol.

| Total RNA extraction and qRT-PCR procedure
The total RNA from tumour tissues and cells were extracted with

| Cell viability assay
Transfected cells were counted and seeded into 96-well plates at a density of 2000 cells per well. After cultured for indicated time, the cells were changed with 100 μL fresh medium containing 10 μL CCK-8 reagent (Beyotime, Haimen, China) and 90 μL complete medium. The absorbance at the wavelength of 450 nm was finally examined after 1 hour incubation. Data were collected for 5 days and three replicate wells were set in each group.

| Statistical analysis
The Kaplan-Meier estimator and GraphPad Prism 6 were used to generate the survival curve. Differences between survival plots were calculated using a log-rank test. Other experiments in this study were analysed by the Student's t test for the significance. P < 0.05 were considered to be statistically significant. ***P < 0.001, **P < 0.01, *P < 0.05.

| No significant difference is observed for the expression levels of RAF and VEGFR in the thyroid tumour tissues from sorafenib sensitive and resistant patients
To  Figure 1C). Taken together, all these data illustrate that there is no significant difference for the expression levels of RAF and VEGFR in the thyroid tumour tissues from sorafenib sensitive and resistant patients.  Figure 4A). Interestingly, a bioinformatics analysis predicted that miR-124 and miR-506 shared a common target gene, EZH2.

| Sorafenib inhibits the proliferation ability of tumour cells derived from sorafenib sensitive patients not the sorafenib resistant patients
A further Western blot result put evidence on this prediction that the protein level of EZH2 was significantly highly expressed in the TC-13 cells compared with the TC-07 cells ( Figure 4B). Subsequently, the expression level of H3K27me3 was up-regulated, whereas, the H3K27Ac was down-regulated in TC-13 cells compared with the TC-07 cells ( Figure 4B). Consistently, the relative mRNA levels of EZH2 and H3K27me3 were markedly elevated and H3K27Ac was dramatically reduced in the TC-13 cells compared with the TC-07 cells ( Figure 4C). In summary, all these data indicate that the EZH2 was modulated by miR-124/506 in sorafenib resistant cells.

| Targeting EZH2 by overexpression of miR-124/506 or EPZ-6438 treatment inhibits the proliferation ability of sorafenib resistant thyroid tumour cells via epigenetic regulation
To identify whether the regulation of sorafenib resistance by

| Combination of sorafenib with miR-124/506 overexpression or EZH2 inhibitor improves the survival in mice model
We next investigated whether miR-124/506 overexpression or

| D ISCUSS I ON
Thyroid carcinoma is the most common and worldwide endocrine malignancy, the incidence of which has been rapidly increasing all over the world. 1,3,4 Although most of the patients with thyroid carcinoma felt in remission following therapies, the recurrence rate still remains high, leading to a decreased survival rate. 7 Sorafenib was a well-known inhibitor targeting RAF, RTK, VEGFR and REK signalling pathways for the therapy of multiple cancers. 9 Even tough emerging studies have reported the therapeutic effects for the DTC by sorafenib treatment, the side effects and drug resistance are still the big problems remained to be solved. 9 Therefore, the detailed mechanism and new therapeutic methods are still needed to be further explored. In this study, we aim to investigate novel target to broaden the choices for thyroid carcinoma therapy.
In this study, we firstly found that some of the patients with thyroid carcinoma showed insensitivity for the sorafenib treatment. Further The sorafenib treatment is already widely used for the therapy of DTC with promising preclinical studies. 9 However, few researches were reported for its side effects and drug resistance. In our study, we found that some patients showed resistance to sorafenib and no significant expression difference of sorafenib targeted genes was overserved, which prompts us to seek new strategies to choose for the DTC therapy with sorafenib resistance.
The dysregulation of EZH2 in tumour cells contributes to inhibitory roles of tumour suppressor genes, leading to the carcinogenesis. 26 The miR-124-3p and miR-506-3p were reported as potential biomarker for the DTC. 18,19 In our study, we found that overexpres- However, there still exist some limitations for this research.
According to the research conditions in the hospital, the main limitation involved into this study is that the number of clinical samples in this study is relatively small, leading to the difficulties for the credibility and generalization of the experimental result. Therefore, further research should be conducted with more clinical samples to reveal the detail mechanism of sorafenib resistance in some of patients with thyroid carcinoma.
In summary, we found out that sorafenib was not suitable for all of the thyroid patients and revealed the reasons why they responded differently to sorafenib treatment, which was the abnormal regulation of PRC signalling pathway. Furthermore, we demon- 3. Our findings suggested a potential therapeutic role of EZH2 in thyroid carcinoma therapy.

D I SCLOS U R E
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y
Data could be obtained upon request to the corresponding author.