KDM6B promotes ovarian cancer cell migration and invasion by induced transforming growth factor-β1 expression

Shumei Liang1 | Qingmin Yao2 | Deying Wei1 | Ming Liu1 | Feng Geng1 | Qin Wang4 | Yun‐shan Wang3

Abstract

KDM6B, also known as JMJD3, is a member of the family of histone lysine demethylase (KDMs), which is closely related to many types of cancers. However, its role and the underlying mechanisms in ovarian cancer remain unknown. Here we show that KDM6B is elevated in epithelial ovarian cancer and its expression level is closely related with metastasis and invasion. In addition, survival analysis showed that high expression of KDM6B was associated with low overall survival in ovarian cancer patients. Overexpression of KDM6B in epithelial ovarian cancer cells promoted proliferation, epithelialmesenchymal transition (EMT), migration and invasion in vitro, and enhanced metastatic capacities in vivo. On the contrary, silencing KDM6B in invasive and metastatic ovarian cancer cells inhibited these processes. Mechanistically, we found that KDM6B exerts its function by modulating the transforming growth factor‐β1 (TGF‐β1) expression, and TGF‐β1 signal pathway inhibitor LY2157299 significantly inhibited KDM6B‐induced proliferation, migration, metastasis, and EMT in ovarian cancer cells. Our findings, for the first time, reveal the pivotal role of KDM6B in the invasion and metastatic behavior of epithelial ovarian cancer. Thus, targeting KDM6B may be a useful strategy to interfere with these behaviors of epithelial ovarian cancer.

KEYWORDS
EMT, KDM6B, ovarian cancer, TGF‐β1

1 | INTRODUCTION

Ovarian cancer is a serious malignancy, and its mortality ranks first in gynecological cancer.1,2 There are no early symptoms and most patients are diagnosed in an advanced stage, when treatment is not very effective. The 5‐year survival rate of ovarian cancer is about 30% to 40%. This poor prognosis is also due to high recurrent and metastatic rates.3-5 Therefore, it remains important to identify novel and effective therapeutic targets to improve the diagnosis and treatment of ovarian cancer patients. Therefore, we have been committed to the understanding of the mechanism of ovarian cancer to provide a more effective method for the treatment of metastatic ovarian cancer. However, the metastatic mechanism of ovarian cancer remains less understood.
Recent studies showed that epithelial cancer, including epithelial ovarian cancer, induction of epithelial‐mesenchymal transition (EMT) is a major event that promotes cancer cell migration, and thus metastasis.6-8 EMT is the process of polarized epithelial cells transforming into viable mesenchymal cells and gain invasion and migration ability, which exist in many physiological and pathological processes of the human body.9-11 The abnormal expression of EMT‐related regulatory factors or inducing factors in cancer cells is closely related to the malignant phenotype and poor prognosis.12,13 Therefore, clarification of these mechanisms will be of great benefit to our understanding of ovarian cancer metastasis.
KDM6B is a member of the family of histone lysine demethylase (KDMs), which specifically catalyzes the removal of trimethylation of histone H3K27me3.14-16 Many studies showed that KDM6B is closely associated with tumor progression.17-20 KDM6B was highly expressed in invasive breast cancer compared with normal breast cancer.21,22 The knockdown of KDM6B drastically reduced breast cancer cell invasion by inhibiting Snail1 expression.21 The same mechanism was also seen in colon cancer.18 Thus, KDM6B seems to play an important role in malignant tumor cell metastasis, but the role of KDM6B in the formation and metastasis of ovarian cancer is still unknown.
In the current study, we show that KDM6B is elevated in epithelial ovarian cancer and its expression level is closely related with metastasis and invasion. KDM6B plays an oncogenic role in ovarian cancer by modulating the transforming growth factor‐β1 (TGF‐β1). Our results, for the first time, depict the pivotal role of KDM6B in ovarian cancer. Thus, targeting KDM6B may be a useful strategy for intervention in ovarian cancer invasion and metastasis.

2 | MATERIALS AND METHODS

2.1 | Reagents and cell cultures

We purchased antibodies against KDM6B (mouse monoclonal antibody [mAb]) and E‐cadherin (mouse mAb) from BD Transduction Labs (Gibbstown, NJ), antibodies against N‐cadherin, α‐catenin, and vimentin antibodies from Cell Signaling Technology (Danvers, MA), TGF‐β1 and β‐actin (mouse mAb) from Sigma (St Louis, MO). Six human ovarian cancer cell lines were used. The cell lines UACC‐1598 and SW‐626 were purchased from the Cell Bank of Shanghai (Shanghai, China). The SK‐OV‐3, Caov‐3, UCC‐2727, and PA‐1 cell lines were purchased from American Type Culture Collection (ATCC, South San Francisco, CA). Cells were cultured in Dulbecco modified Eagle medium (DMEM; Life Technologies, Grand Island, NY) with 10% fetal bovine serum (Life Technologies) and were maintained in a humidified 5% CO2 incubator at 37°C.

2.2 | Patients and tissue specimens

This study was scrutinized and approved by the Hospital Bioethics Committee, and patient consent was obtained before the initiation of the study. The prospective study group comprised 62 patients who had primary epithelial ovarian cancer and underwent ovariectomy at the Department of Obstetrics and Gynecology, Shandong Provincial Hospital Affiliated to Shandong University. Tissue samples of ovarian cancer as well as adjacent noncancerous (normal appearance) ovarian tissues were fixed in 10% neutral formalin, processed for paraffin sections, and used for histopathology and immunohistochemistry studies. All available hematoxylin and eosinstained slides of the surgical specimens were reviewed.

2.3 | Immunohistochemical analysis

Using the immunohistochemical analysis to examine the expression of the KDM6B. Normal ovarian tissue and ovarian cancer tissues were fixed with 4% paraformaldehyde and then overnight in phosphate‐buffered saline, and then embedding in paraffin to cut into a thickness of 4 μm. Then they were stained with hematoxylin and eosin for the histological analysis. The different markers in these arrays were subjected to immunohistochemical analyses. The proportion of stained cells (lower, <30% staining; higher, ≥30% staining) was semiquantitatively determined according to the protocol published by ZVG.

2.4 | RNA extraction, reverse transcription, and real‐time reverse transcription polymerase chain reaction

Total RNA was extracted using Trizol reagent and using SuperScript II Reverse Transcriptase (Invitrogen, Shanghai, China) to synthesize the complementary DNA. Using the ABI PRISM 7900HT Sequence Detection System (Waltham, MA) to carry out the real‐time reverse transcription polymerase chain reaction (RT‐PCR) and the data collection. Primers used to amplify a given gene can be obtained as needed.

2.5 | Western blot analysis

We first lysed the cells in lysis buffer, and then measured the total protein content using the Bradford method. Thirty micrograms of lysis were isolated using by sodium dodecyl sulfate polyacrylamide gel electrophoresis, then transfers it onto a polyvinylidene fluoride membrane. Using 2% albumin from bovine serum (BSA) blocked the membranes for 30 minutes, and then incubated at room temperature for 4 hours in the primary antibody and 1 hour in the secondary antibody. Immunodetection was accomplished using enhanced chemiluminescence (Amersham Biosciences, Beijing, China).

2.6 | Establishment of KDM6B stable expression and silently expressed cell lines

As described previously, we construct the retroviral containing the human KDM6B complementary DNA and the pSuper.retro.puro with short hairpin RNA (shRNA) against human KDM6B, respectively, and then transfected the plasmid into ovarian cancer cells to generate the retrovirus supernatants as described previously. After 6 hours of incubation at 37°C for 6 hours, transferred the transfected cells to fresh culture medium for overnight. In the following days, media were collected three times a day to gather the produced virus until the cells reached to the total confluency. The stable expression of KDM6B was analyzed by real‐time RT‐PCR and Western blot analysis.

2.7 | Proliferation assay

For cell proliferation assay, cells were seeded in triplicate in 96‐well plates, and the density was 1×103 per well by 3‐(4,5dimethylthiazol‐2‐yl)‐2,5‐dimethylthiazol 5‐diphenyltetrazolium‐2‐yl)‐2,5‐diphenyltetrazolium bromide (MTT) assay. In brief, the proliferation of cells was observed at appropriate time points using MTT. The MTT assay was performed 4 hours later by addition of 20μL of MTT (5mg/mL) to each well, discarded the supernatants and added 100μL dimethyl sulfoxide per well to dissolve the remains. After shaken for 15 minutes, and measured the light absorption of the solution at 570nm on a microplate reader.

2.8 | Transwell and Matrigel assay

The Matrigel‐coated transwell inserts which contains the 8‐μm pore size polycarbonate filter were used to perform the cell invasion assay. According to the manufacturer’s recommendations, adding 50 μL 1 mg/mL Matrigel matrix into the insert, making Matrigel spread evenly distributed and covered the all pores of the bottom surface of the upper chambers. At 48 hours after transfection, cells cultured in serum‐free medium were plated in the upper chamber, the 600 μL medium supplemented with 10% lethal bovine serum were added to the lower chamber. After 24 hours incubation, cells that migrated to the lower surface of the membrane were fixed and stained. After 24 hours, collected the number of cells through Matrigel in each group to evaluate their invasion ability. Five randomized regions were counted on each membrane at magnification x10 times.

2.9 | In vivo tumor metastasis model

All animals experiment was carried out with the approval of the Animal Care Use Committee. In the metastasis assay, the cells were suspended with phosphate‐buffered saline and at the concentration of 1 × 107/mL. Subsequently, injected the cell suspension (0.1 mL) into the nude mice via the tail vein. All mice were killed with CO2 60 days after inoculation to observe the metastasis.

2.10 | Gene expression profiling

Total RNA quality and quantity were determined using Agilent 2100 Bioanalyzer and NanoDrop ND‐1000 (Beijing, China). Affymetrix HU U133 plus 2.0 arrays were used according to the manufacturer’s protocol. The data were initially normalized by robust multiarray average normalization algorithms in expression console software (Affymetrix, Waltham, MA). Significantly altered genes between KDM6B overexpression and its control cells were considered by scatter plots and the genes upregulated and downregulated greater than or equal to 5‐fold. Clustering analysis was done using gene list by Gene Cluster v3.0 software (developed by Michael Eisen, Stanford University), and heat maps were visualized using Java TreeView v1.1.4r3 software (Written by Alok Saldanha at Stanford University). Gene set enrichment analysis was carried out using ConceptGen (http://conceptgen.ncibi.org/core/ conceptGen/index.jsp). Gene sets were either obtained from the ConceptGen or from published gene signatures.

2.11 | Chromatin immunoprecipitationquantitative PCR

Chromatin immunoprecipitation kit (Cat. 17‐371) was purchased from Millipore (Darmstadt, Germany) and chromatin immunoprecipitation (ChIP) experiments were carried out essentially as described.8,23 Immunoprecipitated DNA was analyzed by the ABI PRISM 7900HT. The primers used for detection of promoters after ChIP are available upon request.

2.12 | Statistical analysis

SPSS13.0 statistical software (Armonk, N.Y.) was used to analyze the results statistically. The quantitative variables were analyzed using t test, and the correlation between the expression level of KDM6B and TGF‐β1 was evaluated using the χ2 test. Survival rates were estimated according to the Kaplan‐Meier method, and the survival curves were compared by log rank test. P less than 0.05 was used to determine whether the difference in mean values was statistically significant.

3 | RESULTS

3.1 | KDM6B is highly expressed in ovarian cancer and is associated with invasion and metastasis of ovarian cancer

To investigate whether histone lysine demethylase might be involved in ovarian cancer metastasis, the messenger RNA (mRNA) expression level of JMJD1A, JMJD2A, JMJD2B, JMJD2C, JMJD2D, KDM6B, and JMJD6 was detected in three nonmetastatic ovarian cancers and three metastatic ovarian cancers by quantitative RT‐PCR. As compared with nonmetastatic ovarian cancer, the expression level of KDM6B significantly increased in metastatic ovarian cancer (Figure 1A). We then analyzed the mRNA expression level of KDM6B in four noninvasive ovarian cancer cells and two invasive ovarian cancer cells. We found that KDM6B overexpression was significantly associated with aggressiveness in ovarian cancer (Figure 1B). Protein levels of KDM6B in these cancer cells was also analyzed by Western blot analysis (Figure 1C), and the protein level of KDM6B was also upregulated in invasive ovarian cancer cells compared with noninvasive ovarian cancer cells. These data suggest that upregulation of KDM6B might be associated with progression of ovarian cancer.
The mRNA expression level of KDM6B was detected by quantitative RT‐PCR in 62 cases of ovarian cancer and adjacent tissues. Statistical analysis showed that the expression level of KDM6B was significantly higher in ovarian cancer tissues compared with adjacent normal tissues (Figure 1D and 1F), and that of KDM6B in metastatic ovarian cancer was significantly higher than that in nonmetastatic ovarian cancer (Figure 1G).
We detected the protein expression level of KDM6B in normal ovarian tissues and ovarian cancer tissues by immunohistochemistry. The results showed that the protein level of KDM6B in ovarian cancer tissues was significantly higher than that in normal ovarian tissues (Figure 1H). We further analyzed the expression level of KDM6B in ovarian cancer and the survival time of the patients (Figure 1I). The overall survival time of the high expression group was significantly shorter than that of the low KDM6B expression group (P < 0.001). These results suggest that KDM6B plays an important role in the invasion and metastasis of ovarian cancer and is associated with prognosis.

3.2 | KDM6B promotes the proliferation of ovarian cancer cells

To test the carcinogenic activity of KDM6B in ovarian cancer, we established PA‐1 and SW‐626 cell lines that silently expressed KDM6B (Figure 2A‐D) and SK‐OV‐3 and Caov‐3 cell lines stably overexpressing KDM6B (Figure 2E‐H), the KDM6B levels in resulting cell lines were examined by Western blot analysis and quantitative RT‐PCR. The proliferation of ovarian cancer cells was measured by MTT assay. The results showed that the proliferation capacity of SK‐OV‐3‐KDM6B and Caov3‐KDM6B cells significantly increased compared to the control group (Figure 2K and 2L). In contrast, the proliferation capacity of PA‐I‐shKDM6B and SW‐626shKDM6B cells was significantly reduced (Figure 2I and 2J). Our results show that KDM6B is an important regulator of ovarian cancer cell proliferation.

3.3 | KDM6B regulates epithelial and mesenchymal transition of ovarian cancer cells

It is well known that EMT is the loss of epithelial cell characteristics and the acquisition of a mesenchymal phenotype that confers the ability of cancer cells to invade adjacent tissues and migrate to distant sites. To investigate whether KDM6B is involved in the regulation of migration and invasion of ovarian cancer cells, we analyze epithelial SW‐626 cell lines silently expressed KDM6B (A‐D), SK‐OV‐3 and Caov‐3 cell lines stably overexpressing KDM6B (E‐H). The protein and messenger RNA levels of KDM6B in resulting cell lines and their vector control cells were examined by Western blot analysis (A,B,E, and F) and quantitative reverse transcription polymerase chain reaction (C,D,G, and H). I‐L, Cell proliferation was examined by MTT assays in resulting cell lines. I and J, Proliferation capacity of PA‐I‐shKDM6B＃ 1＃2, SW‐626‐shKDM6B＃1＃2, and their control cells. K and L, Proliferation capacity of SK‐OV‐3‐KDM6B, Caov‐3KDM6B, and their vector cells. **P < 0.01 is based on the Student t test. All results are from three independent experiments. Error bars, SD. MTT, 3‐(4,5dimethylthiazol‐2‐yl)‐2,5‐dimethylthiazol 5‐diphenyltetrazolium‐2‐yl)‐2,5diphenyltetrazolium bromide and mesenchymal markers of established cell lines by Western blot analysis. We found that the levels of epithelial markers (E‐cadherin and α‐catenin) of PA‐I and SW‐626 which silencing express KDM6B were elevated compared to their respective control cells, and the levels of mesenchymal markers (N‐cadherin and vimentin) decreased (Figure 3A and 3B). In contrast, overexpression of KDM6B in both SK‐OV‐3 and Caov‐3 cell lines reduced the levels of epithelial markers and increased levels of mesenchymal markers (Figure 3C and 3D). These results suggest that KDM6B plays an important role in the regulation of EMT‐mesenchymal‐epithelial transition (MET) transduction in ovarian cancer cells.

3.4 | KDM6B promotes ovarian cancer cell migration and invasion in vitro

We evaluated the effect of KDM6B on the migration and invasion of ovarian cancer cells by transwell and Matrigel. Migration and invasion capacity of PA‐I‐shKDM6B and SW‐626‐shKDM6B cells were significantly decreased compared to the control group (Figure 4A and 4B), suggesting that the recovery of epithelial phenotype inhibits cell migration. In contrast, overexpression of KDM6B significantly increased migration and invasion of SK‐OV‐3 and Caov‐3 cells, suggesting that acquisition of mesenchymal phenotype promotes cell invasion and motility (Figure 4C and 4D). These results indicate that KDM6B promotes migration and invasion in ovarian cancer cells.

3.5 | KDM6B promotes tumor metastasis in vivo

We further investigated whether KDM6B could modulate ovarian cancer cell migration in vivo. SK‐OV‐3‐KDM6B cell and its control cell were injected into the nude mice via tail vein. The number of lung metastatic tumors was significantly higher in which receiving SK‐OV‐3‐KDM6B cell than that in the control group. Our results in vivo transforming growth factor‐β1 (TGF‐β1) expression in ovarian cancer cells. A, Supervised hierarchical clustering of the genes differentially expressed after KDM6B overexpression in ovarian cancer cells. B, Gene set enrichment analysis was carried out using ConceptGen. C and E, The messenger RNA (mRNA) and protein expression levels of TGF‐β1 was detected in PA‐I‐shKDM6Bs and their control cells by quantitative reverse transcription polymerase chain reaction (RT‐PCR) and Western blot analysis. D and F, The mRNA and protein expression levels of TGF‐β1 were detected in SW‐626shKDM6Bs and their control cells by quantitative RT‐PCR and Western blot analysis. G and I, The mRNA and protein expression levels of TGF‐β1 were detected in SK‐OV‐3‐KDM6B and their vector cells by quantitative RT‐PCR and Western blot analysis. H and J, The mRNA and protein expression levels of TGF‐β1 were detected in Caov‐3‐KDM6B and their vector cells by quantitative RT‐PCR and Western blot analysis. K, Correlation analysis showed that KDM6B and TGF‐β1 were significantly positively correlated. **P < 0.01 is based on the Student t test. All results are from three independent experiments. Error bars, SD further confirm the key role of KDM6B in ovarian cancer metastasis (Figure 4E and 4F).

3.6 | KDM6B exerts its function through TGF‐β1

To understand the mechanism of KDM6B in promoting the development of ovarian cancer, we detected the target gene of KDM6B with gene chip technology. The result showed that TGF‐β1 may be the main regulator of KDM6B (Figure 5A and 5B). To verify this conclusion, we further evaluated the expression levels of TGF‐β1 in the cells with altered KDM6B expression by quantitative RT‐PCR and Western blot analysis. TGF‐β1 mRNA and protein levels were significantly increased in SK‐OV3‐KDM6B and Caov‐3‐KDM6B cells (Figure 5G‐J), and dramatically decreased in PA‐I‐shKDM6B and SW‐626shKDM6B cells (Figure 5A‐F). Correlation analysis showed that KDM6B and TGF‐β1 were also significantly positively correlated in ovarian cancer (Figure 5K).
KDM6B is frequently involved in chromatin regulation.24,25 To determine whether KDM6B regulates specific histone modifications in ovarian cancer cells, histone modification patterns were measured after modulation of KDM6B expression. Among histone H3K4 and H3K27, we found that only H3K27me3 was affected by KDM6B (Figure 6A‐D). Ectopic expression of KDM6B decreased H3K27me3 while silencing of KDM6B increased this modification. To detect the correlation between TGF‐β1 and H3K27me3 in cells, ChIP‐quantitative PCR was applied. Results showed that silencing KDM6B significantly increased the binding of H3K27me3 to the promoter of TGF‐β1, while, overexpression KDM6B decreased the binding of H3K27me3 to the promoter of TGF‐β1 (Figure 6E‐G). Thus, taken together through this way, KDM6B modulates the expression of TGF‐β1.
We further treated SK‐OV‐3 and Caov‐3 cells which stable overexpression KDM6B and their respective control cells with TGF‐β1 inhibitor LY2157299. The levels of obtained interstitial markers in both cells decreased and the epithelial marker levels increased (Figure 7A and 7B). Transwell and Matrigel experiments also showed that the migration and invasion abilities of the two cells were decreased compared to the control group (Figure 7C and 7D). Taken together, these results indicate that TGF‐β1 mediates KDM6B‐induced migration, invasion, and EMT of ovarian cancer cells.

4 | DISCUSSION

Ovarian cancer is one of the most common female malignant tumors which mortality ranks first in gynecological malignancies.1,4 The poor prognosis is due to high recurrent and metastatic rates.5 At present, the treatment options for metastatic ovarian cancer are limited and ineffective. The identification of new and effective therapeutic targets remains a clinically important task.
Histone methylation is a kind of epigenetic modification, the change of methylation status plays an important role in the genesis and development of cancer.25,26 KDM6B is a member of the family of histone lysine demethylase (KDMs), which specifically acts on H3K27me3 and is reported to participate in a variety of organizational differentiation process.27,28 The role of KDM6B in malignant tumors has also been reported in recent studies. Several studies reported that KDM6B acts as a tumor suppressor and is downregulated in human cancers.29,30 However, there were also some studies suggesting that KDM6B is highly expressed in human cancers.28,31 In this regard, our studies confirmed the oncogenic role of KDM6B in ovarian cancer. To our knowledge, this is the first research to show that KDM6B plays an important role in epithelial ovarian cancer. Here we found that KDM6B was elevated in ovarian cancer and its expression level was closely related with metastasis and invasion. High expression of KDM6B was associated with low overall survival. Overexpression of KDM6B in ovarian cancer cells promoted proliferation, migration and invasion in vitro, and enhanced metastatic capacities in vivo. On the contrary, silencing KDM6B in invasive and metastatic ovarian cancer cells inhibited these processes.
EMT is a critical event that occurs in embryonic microRNA, and epigenetics.32-34 During the process of development, tissue repair control, organ fibrosis, and tumor cell metastasis, the tumor cells lose their polarity and carcinoma invasion and metastasis, which process is intercellular adhesion through EMT, and thus gaining the regulated by many factors, such as transcription factors, ability of invasion and migration. In the new metastatic sites, tumor cells undergo epithelial‐mesenchymal transition (MET) to restore epithelial cell status and form new tumor foci.32,35 Therefore, the EMT‐MET process is thought to play a decisive role in the invasion and metastasis of malignant tumor cells. Interestingly, our results indicate that KDM6B not only promotes EMT, but silencing of KDM6B also leads to MET.
The occurrence of EMT involves multiple signal transduction pathways and complex molecular mechanisms, such as calcium connexins, growth factors, transcription factors, and microenvironment and so on. In the current study, we tried to elucidate the mechanism of KDM6B regulating EMT in ovarian cancer. We identified a significant positive correlation between KDM6B and TGF‐β1 expression. TGFβ1 signaling pathway inhibitor LY2157299 could significantly inhibit KDM6B‐induced EMT and invasion ability in ovarian cancer. Therefore, we conclude that TGF‐β1 is an effective mediator of KDM6B‐induced EMT, invasion and metastasis in ovarian cancer, but the specific target of KDM6B needs further study confirmed.
Our results revealed the role of KDM6B in the proliferation, invasion and migration and EMT of ovarian cancer and provide a new therapeutic option for the clinical treatment. On this basis, expect to find the corresponding drugs or agents for ovarian cancer.

REFERENCES

1. Torre LA, Bray F, Siegel RL, Ferlay J, Lortet‐Tieulent J, Jemal A. Global cancer statistics, 2012. CA Cancer J Clin.2015;65:87‐108.
2. Creasman WT, Odicino F, Maisonneuve P, et al. Carcinoma ofthe corpus uteri. FIGO 26th Annual Report on the Results of Treatment in Gynecological Cancer. Int J Gynaecol Obstet.2006;95(suppl 1):S105‐S143.
3. da Costa AA, Valadares CV, Mantoan H, et al. The value ofsecondary cytoreductive surgery in recurrent ovarian cancer and application of a prognostic score. Int J Gynecol Cancer.2016;26:449‐455.
4. Sehouli J, Grabowski JP. Surgery for recurrent ovarian cancer: options and limits. Best Pract Res Clin Obstet Gynaecol.2016;41:88‐95.
5. Bickell NA, Egorova N, Prasad‐Hayes M, et al. Secondary surgery versus chemotherapy for recurrent ovarian cancer. Am J Clin Oncol. 2016
6. Giannelli G, Koudelkova P, Dituri F, Mikulits W. Role ofepithelial to mesenchymal transition in hepatocellular carcinoma. J Hepatol. 2016;65:798‐808.
7. Jayachandran A, Dhungel B, Steel JC. Epithelial‐to‐mesenchymal plasticity of cancer stem cells: therapeutic targets in hepatocellular carcinoma. J Hematol Oncol. 2016;9:74.
8. Wang Y, Wen M, Kwon Y, et al. CUL4A induces epithelialmesenchymal transition and promotes cancer metastasis by regulating ZEB1 expression. Cancer Res. 2014;74:520‐531.
9. Lee JM, Dedhar S, Kalluri R, Thompson EW. The epithelialmesenchymal transition: new insights in signaling, development, and disease. J Cell Biol. 2006;172:973‐981.
10. Beerling E, Seinstra D, De Wit E, et al. Plasticity betweenepithelial and mesenchymal states unlinks EMT from metastasis‐enhancing stem cell capacity. Cell Rep. 2016;14:2281‐2288.
11. Chaffer CL, San Juan BP, Lim E, Weinberg RA. EMT, cellplasticity and metastasis. Cancer Metastasis Rev. 2016;35:645‐654.
12. Przybyla L, Muncie JM, Weaver VM. Mechanical control ofepithelial‐to‐mesenchymal transitions in development and cancer. Annu Rev Cell Dev Biol. 2016;32:527‐554.
13. Thiery JP, Sleeman JP. Complex networks orchestrate epithelial‐mesenchymal transitions. Nat Rev Mol Cell Biol. 2006;7:131‐142.
14. De Santa F, Totaro MG, Prosperini E, Notarbartolo S, Testa G,Natoli G. The histone H3 lysine‐27 demethylase Jmjd3 links inflammation to inhibition of polycomb‐mediated gene silencing. Cell. 2007;130:1083‐1094.
15. Lan F, Bayliss PE, Rinn JL, et al. A histone H3 lysine 27 demethylase regulates animal posterior development. Nature. 2007;449:689‐694.
16. Xiang Y, Zhu Z, Han G, Lin H, Xu L, Chen CD. JMJD3 is ahistone H3K27 demethylase. Cell Res. 2007;17:850‐857.
17. Agger K, Cloos PAC, Rudkjaer L, et al. The H3K27me3 demethylase JMJD3 contributes to the activation of the INK4AARF locus in response to oncogene‐ and stress‐induced senescence. Genes Dev. 2009;23:1171‐1176.
18. Pereira F, Barbáchano A, Silva J, et al. KDM6B/JMJD3 histonedemethylase is induced by vitamin D and modulates its effects in colon cancer cells. Hum Mol Genet. 2011;20:4655‐4665.
19. Anderton JA, Bose S, Vockerodt M, et al. The H3K27me3 demethylase, KDM6B, is induced by Epstein‐Barr virus and over‐expressed in Hodgkin’s lymphoma. Oncogene. 2011;30: 2037‐2043.
20. van Haaften G, Dalgliesh GL, Davies H, et al. Somaticmutations of the histone H3K27 demethylase gene UTX in human cancer. Nat Genet. 2009;41:521‐523.
21. Ramadoss S, Chen X, Wang CY. Histone demethylase KDM6Bpromotes epithelial‐mesenchymal transition. J Biol Chem.2012;287:44508‐44517.
22. Bae WK, Yoo KH, Lee JS, et al. The methyltransferase EZH2 isnot required for mammary cancer development, although high EZH2 and low H3K27me3 correlate with poor prognosis of ERpositive breast cancers. Mol Carcinog. 2015;54:1172‐1180.
23. Wang Y, Zhang P, Liu Z, et al. CUL4A overexpression enhanceslung tumor growth and sensitizes lung cancer cells to erlotinib via transcriptional regulation of EGFR. Mol Cancer. 2014;13:252.
24. Agger K, Cloos PAC, Christensen J, et al. UTX and JMJD3 arehistone H3K27 demethylases involved in HOX gene regulation and development. Nature. 2007;449:731‐734.
25. Manna S, Kim JK, Baugé C, et al. Histone H3 lysine 27 demethylases Jmjd3 and Utx are required for T‐cell differentiation. Nat Commun. 2015;6:8152.
26. Yang D, Okamura H, Teramachi J, Haneji T. Histone demethylase Jmjd3 regulates osteoblast apoptosis through targeting anti‐apoptotic protein Bcl‐2 and pro‐apoptotic protein Bim. Biochim Biophys Acta. 2016;1863:650‐659.
27. Park WY, Hong BJ, Lee J, Choi C, Kim MY. H3K27 demethylase JMJD3 employs the NF‐kappaB and BMP signaling pathways to modulate the tumor microenvironment and promote melanoma progression and metastasis. Cancer Res.2016;76:161‐170.
28. Perrigue PM, Najbauer J, Barciszewski J. Histone demethylaseJMJD3 at the intersection of cellular senescence and cancer. Biochim Biophys Acta. 2016;1865:237‐244.
29. Tokunaga R, Sakamoto Y, Nakagawa S, et al. The prognosticsignificance of histone lysine demethylase JMJD3/KDM6B in colorectal cancer. Ann Surg Oncol. 2016;23:678‐685.
30. Yamamoto K, Tateishi K, Kudo Y, et al. Loss of histonedemethylase KDM6B enhances aggressiveness of pancreatic cancer through downregulation of C/EBPalpha. Carcinogenesis.2014;35:2404‐2414.
31. Na J, Lee K, Na W, et al. Histone H3K27 demethylase JMJD3 incooperation with NF‐kappaB regulates keratinocyte wound healing. J Invest Dermatol. 2016;136:847‐858.
32. Hanna A, Shevde LA. Hedgehog signaling: modulation ofcancer properties and tumor mircroenvironment. Mol Cancer.2016;15:24.
33. Serrano‐Gomez SJ, Maziveyi M, Alahari SK. Regulation of epithelial‐mesenchymal transition through epigenetic and posttranslational modifications. Mol Cancer. 2016;15:18.
34. Sun SS, Zhou X, Huang YY, et al. Targeting STAT3/miR‐21 axis inhibits epithelial‐mesenchymal transition via regulating CDK5 in head and neck squamous cell carcinoma. Mol Cancer.2015;14:213.
35. Baum B, Settleman J, Quinlan MP. Transitions betweenepithelial and mesenchymal states in development and disease. Semin Cell Dev Biol. 2008;19:294‐308.