Genome-Wide Screening and Functional Analysis Identifies Tumor Suppressor Long Non-Coding Rnas Epigenetically Silenced in Hepatocellular Carcinoma

Author Manuscript Published OnlineFirst on February 4, 2019; DOI: 10.1158/0008-5472.CAN-18-1659 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Genome-wide screening and functional analysis identifies tumor suppressor long non-coding RNAs epigenetically silenced in hepatocellular carcinoma

Feiyue Xu1, Chi Han Li1, Chi Hin Wong1, George G Chen2, Paul Bo San Lai2, Shengwen Shao3, Stephen L. Chan4, Yangchao Chen1,5*

1School of Biomedical Sciences, Faculty of Medicine, The Chinese University of Hong Kong, Shatin, NT, Hong Kong 2Department of Surgery, Prince of Wales Hospital, The Chinese University of Hong Kong, Shatin, Hong Kong 3Institute of Microbiology and Immunology, Huzhou University, Huzhou, Zhejiang, China 4Department of Clinical Oncology, State Key Laboratory in Oncology of South China and Institute of Digestive Disease, Prince of Wales Hospital, The Chinese University of Hong Kong, Shatin, Hong Kong 5Shenzhen Research Institute, The Chinese University of Hong Kong, Shenzhen 518087, China

Running titles: Tumor suppressor lncRNAs epigenetically silenced in HCC

Keywords: Hepatocellular carcinoma; EZH2; long non-coding RNA, Epigenetics; RNA-binding protein

Contact Information Yangchao Chen, PhD, School of Biomedical Sciences, Faculty of Medicine, Chinese University of Hong Kong, Shatin, Hong Kong, Phone: +852 39431100; Fax: +852 26035123; Email: [email protected]

Conflict of interest There is no conflict of interest.

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 2019 American Association for Cancer Research. Author Manuscript Published OnlineFirst on February 4, 2019; DOI: 10.1158/0008-5472.CAN-18-1659 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

List of Abbreviations 3-Deazaneplanocin A (DZNep); Chromatin immunoprecipitation (ChIP); DZnep- upregulated 2 (TCAM1P-004); DZnep-upregulated 5 (RP11-598D14.1); Enhancer of zest homolog 2 (EZH2); Hepatocellular Carcinoma (HCC); polycomb repressive complex 2 (PRC2); Long non-coding RNA (LncRNA); Quantitative reverse- transcription PCR (qRT-PCR); RNA immunoprecipitation (RIP)

Number of figures: 7

Abstract Long non-coding RNAs (lncRNA) play critical roles in the development of cancer including hepatocellular carcinoma (HCC). However, the mechanisms underlying their deregulation remain largely unexplored. In this study, we report that two lncRNA frequently downregulated in HCC function as tumor suppressors and are epigenetically silenced by histone methyltransferase EZH2. lncRNAs TCAM1P-004 and RP11-598D14.1 were inhibited by EZH-mediated trimethylation of H3K27me3 at their promoters. Downregulation of TCAM1P-004 and RP11-598D14.1 were frequently observed in HCC tumors compared to adjacent normal tissues. Both lncRNAs inhibited cell growth, cell survival, and transformation in HCC cells in vitro as well as tumor formation in vivo. Using RNA pull-down and mass spectrometry, we demonstrated that TCAM1P-004 bound IGF2BP1 and HIST1H1C, while RP11-598D14.1 bound IGF2BP1 and STAU1. These lncRNA-protein interactions were critical in regulating p53, MAPK, and HIF1-α pathways that promoted cell proliferation in HCC. Overexpression of EZH2 was critical in repressing TCAM1P-004 and RP11-598D14.1, and EZH2-TCAM1P-004/RP11- 598D14.1-regulated pathways were prevalent in human HCC. Aberrant suppression of TCAM1P-004 and RP11-598D14.1 led to loss of their tumor suppressor effects by disrupting the interaction with IGF2BP1, HIST1H1C and STAU1, which in turn promoted HCC development and progression. Collectively, these findings demonstrate the role of TCAMP1P-004 and RP11-598D14.1 in suppressing tumor growth and suggest that EZH2 may serve as a therapeutic target in HCC.

Statement of significance: EZH2-mediated loss of lncRNAs TCAM1P-004 and RP11-598D14.1 hinders the formation of tumor suppressor lncRNA-protein complexes and subsequently promotes HCC growth.

Introduction Liver cancer has one of the lowest cancer survival rates (cancer statistics 2018) (1), and more than 80% of liver cancer incidences are hepatocellular carcinoma (HCC)

(2). According to the World Health Organization, liver cancer is the second leading cause of death worldwide in 2015 (3). The prognosis of HCC is very poor with approximately 5–6% 5-year survival rate (4). Almost 80% of HCC cases occurred in East Asia and sub-Saharan Africa (5) and more than 50% of diagnosed HCC patients are Chinese (6).

EZH2, a core component of polycomb repressive complex 2 (PRC2), silences gene expression via its histone methyltransferase activity (7). EZH2 plays essential roles in cancer initiation, progression and metastasis (8-10). Its expression is elevated aberrantly in certain types of cancers including HCC. EZH2 frequently functions as an oncogenic factor and heavily involved in the silencing of critical tumor suppressor genes such as HOX, CCN3/NOV, DAB2IP, TIMP2/3, p16, KLF2 and RUNX3 (11, 12). Also, EZH2 participated in the deregulations of miRNAs including miR-218, miR-26a, miR-1246 and miR-4448 (13, 14).

Long non-coding RNAs (LncRNAs) are non-coding RNA molecules with over 200 nucleotides, which are also transcribed by RNA polymerase II similar to mRNAs (15). LncRNAs play diverse roles in biological process such as epigenetic regulation, translational regulation, post-transcriptional processing, imprinting, apoptosis and cell cycle (15, 16). Several aberrantly expressed lncRNAs have been studied, but only a few lncRNAs have been characterized comprehensively, including HOTAIR, HOTTIP, HULC and MALAT1 (17-20). These lncRNAs have crucial regulatory roles in cancer biology. In HCC, aberrant expression of lncRNAs are frequently observed (21, 22) that resulted in the deregulation of gene expressions through remodeling of chromatin, regulation of gene transcription, control of posttranscriptional mRNA processing, assistance of protein localization or function and acting in intercellular signaling. The importance of lncRNAs in cancer is clearly understood, however, the underlying mechanism of their deregulations is still largely unexplored.

Both lncRNAs and EZH2 are important epigenetic regulators in cancer. Numerous

protein-coding genes repressed by EZH2 were reported, and the interactions between lncRNAs and EZH2 have been widely studied, but there are few reports studying the regulation of lncRNAs by EZH2. Therefore, we attempted to identify the lncRNAs directly regulated by EZH2, reveal the regulating mechanisms and explore the roles of these lncRNAs in HCC.

Materials and methods Clinical Samples, Cell Lines, and Drug Treatment Fifty pairs of HCC tumor tissues and adjacent normal tissues were obtained from HCC patients who received surgery of liver cancer resection at the Prince of Wales Hospital, Hong Kong. The study was carried out according to the ethical guidelines and with the approval of the Joint CUHK-NTEC Clinical Research Ethics Committee in accordance with Declaration of Helsinki and the written informed consent was obtained from all patients recruited. All specimens of HCC tumors and normal tissues were stored at -80°C. Each sample was divided into two parts. One was used for RNA extraction, and the other was fixed and embedded into paraffin. The nontumorigenic human hepatocyte cell line MIHA was obtained from Dr. J.R. Chowdhury's laboratory at Albert Einstein College of Medicine (New York, NY). The human HCC cell line Huh7 (kindly provided by Dr. H. Nakabayashi, Hokkaido University School of Medicine, Sapporo, Japan), Bel-7404 and L02 (Cell Bank of the Chinese Academy of Sciences) were authenticated with short-tandem repeat profiling by the vendors. The human HCC cell lines HepG2, PLC/PRF/5 (PLC), and Hep3B (American Type Culture Collection) were verified by short-tandem repeat profiling at the GENEWIZ, Inc. within 6 months of use. Cell lines were maintained in DMEM containing 100 μg/ml streptomycin and 100 unit/ml penicillin with 10%

(v/v) FBS at 37°C in a humidified chamber with 5% CO2. All cell lines undergo routine mycoplasma testing. For drug treatment, Deazaneplanocin A (DZnep) was dissolved in ethanol (stock solution concentration 1 mg/ml). 5×105 to 8×105 Hep3B and Huh7 cells were seeded in a 12 well plate and were treated with 10μM 3- Deazaneplanocin A (DZnep) for 48 hours. Cells treated with ethanol only were used

as the control.

Packaging of lentivirus and cell transduction Packaging of lentivirus was performed according to our protocol (10). In brief, packaging plasmids pMDLg/pRRE, pRSV-REV, pCMV-VSVG and transgene lentiviral were transfected into 293T cells using PEI (Polyetherimide, 1mg/ml). Before transduction, HCC cells (1x104) were seeded on 24 well overnight. 293T cell supernatant containing lentivirus particles were added to the wells with 10μg/ml polybrene. Cells were transduced for 24 hrs before replenishing fresh medium with G418 (800 ng/mL) mammalian selection. Lentiviral vector carrying MYC-tagged full length EZH2 was prepared previously (14). Short-hairpin RNA (shRNA) targeting EZH2, TCAM1P-004 and RP11-598D14.1 were cloned into H1 promoter containing lentiviral vectors constructed by our team that harbored G418-resistant genes and EGFP according to our previous publication (10). Full length TCAM1P- 004 and RP11-598D14.1 were amplified and cloned into a same lentiviral backbone contained CMV promoter instead of H1 promoter for lncRNA expressions, and coexpressed G418-resistant gene and EGFP (10).

Chromatin immunoprecipitation (ChIP) assay ChIP assay was performed by using SimpleChIP® Enzymatic Chromatin IP Kit Magnetic Beads (Cell signaling Technology, Danvers, MA) according to the manufacturer’s protocol (23). For immunoprecipitation, anti-EZH2 (Cell Signaling; #5246), anti-H3K27me3 (Millipore; 07-449), anti-EED (Millipore; 05-1320), anti- SUZ12 (Millipore; 05-1317) antibody and negative control Normal Rabbit IgG antibody were incubated with cross-linked chromatin at 4 °C overnight with rotation. The precipitated DNA enrichment was quantified by qPCR and normalized by respective 2% input.

siRNAs transfection Transfections of siRNAs were conducted using DharmaFECT 1 transfection reagent (Thermo Scientific) according to the manufacturer’s protocol. During transfection,

100 nM of siRNAs was used to treat the cells for 72 hrs before RNA or protein extraction.

RNA pull-down assay and Liquid chromatography mass spectrometry (LC-MS) RNA pull-down assay was performed using Pierce Magnetic RNA-Protein Pull- Down Kit (Thermo Fisher Scientific) according to the manufacturer’s protocol. RNAs were labeled with Biotin and were purified by streptavidin magnetic beads. MIHA cell protein lysates were prepared by using Pierce IP Lysis Buffer and incubated with mixture of RNAs and magnetic beads. Unbound proteins were removed by washing. RNA binding Proteins were purified with acetone and then incubated with sequencing grade trypsin (Promega, Madison, WI) to digest protein into peptides at 37 °C overnight. Peptides were purified by ZipTip Pipette Tips (Merck Millipore, Billerica, MA) according to the manufacture’s protocol. The peptides were dried using DNA Savant™ SpeedVac DNA110 (Thermo Fisher Scientific) and detected by LC-Triple TOF 5600 (SCIEX, Framingham, MA). Results were analyzed by ProteinPilot TM software.

RNA immunoprecipitation (RIP) assay RNA immunoprecipitation was performed with Magna RIP™ RNA-Binding Protein Immunoprecipitation Kit (Merck Millipore) according to the manufacture’s protocol. Cells were lysed in RIP Lysis Buffer with protease inhibitor cocktails and RNase inhibitor. The precipitated RNA was extracted with phenol: chloroform: isoamyl alcohol or by RNeasy Mini Kit (QIAGEN, Hilden, Germany). Quantitative reverse- transcription PCR (qRT-PCR) was performed to determine the RNA enrichment with the binding proteins.

Results Identification of upregulated lncRNAs in HCC cells treated with EZH2 inhibitor DZNep. To identify lncRNAs regulated by EZH2, Huh7 cells and Hep3B cells were treated with 10 μM EZH2 inhibitor DZnep. Treatment with DZnep effectively depleted

EZH2 proteins, and H3K27 trimethylation (H3K27me3) levels were notably reduced in both Huh7 and Hep3B cells (Fig1A). Arraystar Human lncRNA/mRNA expression profiling was performed to identify lncRNAs that were differentially expressed after DZnep treatment in Hep3B cells. We focused on the lncRNAs that passed volcano plot filtering with statistical significance (p<0.05) and had fold change > 2.0. Both upregulation and downregulation of lncRNAs were observed in DZnep-treated cells (Supplementary table 1). As we attempted to identify tumor suppressor lncRNAs repressed by EZH2, we only focused on the lncRNAs that were upregulated after EZH2 was inhibited by DZnep. We firstly numbered them with the prefix DZnep-upregulated (DN). Ten lncRNAs that had the highest fold increase from the microarray data were measured in Hep3B and Huh7 cells. We showed that most of them were upregulated after DZnep treatment (Sfig1), suggesting their specificity to DZnep treatment. We then measured the upregulated lncRNAs in a panel of HCC and non-tumor cell lines, and human HCC tissues. LncRNAs that were frequently downregulated in HCC cells and tissues were selected for subsequent analysis (Sfig2-3).

TCAM1P-004 and RP11-598D14.1 (DN2 and DN5 respectively) were two of the ten lncRNAs validated to be upregulated in Huh7 and Hep3B cells treated with DZnep (Fig1B). Their expressions were measured in a panel of HCC cell lines including Hep3B，Huh7，Bel7404 and PLC, and in non-tumor cell lines MIHA and L02. Compared to MIHA and L02 cells, both TCAM1P-004 and RP11- 598D14.1 expressions were lower in HCC cells (Fig1C). The levels of TCAM1P- 004 and RP11-598D14.1 in 50 pairs of HCC tumors and corresponding adjacent normal tissues were measured by qRT-PCR. Their expression levels were reduced significantly in HCC tumor tissues (p<0.001) (Fig1D). In addition, Kaplan-Meier plots revealed an association of higher TCAM1P-004 and RP11-598D14.1 expression with longer overall survival of HCC patients (n=50) (Fig1E). Among all ten candidate lncRNAs, only TCAM1P-004 and RP11-598D14.1 were significantly downregulated in HCC tissues. Therefore, we omitted other lncRNAs because their importance was highly diminished as no dysregulation in human HCC tissues

observed. As such, we focused on TCAM1P-004 and RP11-598D14.1 in this study.

EZH2 suppressed TCAM1P-004 and RP11-598D14.1 expression by inducing histone 3 lysine 27 trimethylation. Full-length EZH2 with MYC-tag was cloned into an overexpression lentiviral vector (14). EZH2 was ectopically overexpressed in MIHA cells by transducing EZH2 overexpressing lentivirus. Western blot showed that the EZH2 protein level was higher in the MIHA-EZH2 cells than in the vector control cells. MYC-tag expression was observed in the MIHA-EZH2 cells only, indicating that full-length EZH2 was overexpressed in MIHA-EZH2 cells (Fig2A). TCAM1P-004 and RP11- 598D14.1 levels were reduced significantly in EZH2-overexpressing MIHA cells. These indicated that EZH2 was able to repress TCAM1P-004 and RP11-598D14.1 expressions in non-tumor hepatocytes.

Moreover, we constructed Hep3B-shEZH2 cells transduced by lentivirus carrying shRNA targeting EZH2 (14). Western blot showed that EZH2 protein was effectively reduced in EZH2-inhibited Hep3B cells (Fig2B). The expressions of TCAM1P-004 and RP11-598D14.1 were significantly increased in EZH2-inhibited cells compared to the control cells (Fig2B). To further confirm that TCAM1P-004 and RP11-598D14.1 were the putative targets of EZH2, we measured the expression levels of TCAM1P-004 and RP11-598D14.1 in Hep3B and Huh7 cells that were transfected with EZH2 siRNAs. The transient inhibition of EZH2 by siRNAs also resulted in the upregulation of TCAM1P-004 and RP11-598D14.1 in both cell lines (Fig2C-D). As the gene silencing effect induced by EZH2 was dependent on the polycomb repressive complex 2 (PRC2), we inhibited two PRC2 core members EED and SUZ12 by siRNAs in Hep3B and Huh7 cells in order to study their effect on TCAM1P-004 and RP11-598D14.1 levels. Inhibition of EED and SUZ12 led to the upregulation of TCAM1P-004 and RP11-598D14.1 in both cell lines (Fig2E-F). Taken together, our results suggested that TCAM1P-004 and RP11-598D14.1 were the putative targets repressed by EZH2 in HCC cells.

EZH2 couples with PRC2 partners to the promoters of protein-coding genes, thus leading to trimethylation of H3K27 that silences target genes. We hypothesized that EZH2 was recruited together with PRC2 members EED and SUZ12 to regulate TCAM1P-004 and RP11-598D14.1. To prove our hypothesis, ChIP assays were performed in MIHA and Hep3B cells. ChIP assays showed that EZH2, EED and SUZ12 were enriched in the promoter regions of TCAM1P-004 and RP11-598D14.1 in Hep3B cells, and the levels were higher than in MIHA cells. Epigenetic marker H3K27me3, which was catalyzed and maintained by PRC2, was also detected in the promoters of TCAM1P-004 and RP11-598D14.1 in Hep3B but not in MIHA cells. This indicated that TCAM1P-004 and RP11-598D14.1 were both repressed by PRC2 in HCC cells, whereas TCAM1P-004 and RP11-598D14.1 were not deregulated in MIHA cells with low EZH2 level (Fig3A).

ChIP assays were also carried out in EZH2-inhibited cells. Depletion of EZH2 led to the reduced occupancies of SUZ12, EED and H3K27me3 in the promoters of TCAM1P-004 and RP11-598D14 (Fig3B). Overexpression of EZH2 induced the downregulation of TCAM1P-004 and RP11-598D14.1 in MIHA-EZH2 cells. Upon EZH2 overexpression, the occupancies of SUZ12, EED and EZH2, together with H3K27me3 levels at the promoters of TCAM1P-004 and RP11-598D14.1 were also elevated (Fig3C). To demonstrate the involvement of PRC2 complex during the repression of TCAM1P-004 and RP11-598D14.1, we transiently inhibited EED and SUZ12 by transfecting siRNAs into Hep3B cells to study the effects on the targeting of EZH2. ChIP assay showed that inhibition of EED and SUZ12 could hinder the binding of EZH2 on the promoters of TCAM1P-004 and RP11-598D14.1 (Fig3D-E).

To show that our ChIP assays could identify true interaction between EZH2 and target promoter, we attempted to detect the promoter of a gene nearby TCAM1P-004 genomic region that named SMARCD2. We showed that SMARCD2 was not regulated by EZH2, as both DZnep treatment and EZH2 siRNAs transfection in Hep3B cells were unable to alter SMARCD2 expression (Sfig4A). As expected, we failed to detect the presence of SMARCD2 promoter in the anti-EZH2 precipitated

DNA as expected (Sfig4B). It proved that our ChIP assay was able to measure the genuine interaction between EZH2 and its targeted promoter regions. These ChIP assays confirmed our hypothesis that EZH2 was dependent on the PRC complex to catalyze the formation of H3K27me3 and inhibit the transcription of TCAM1P-004 and RP11-598D14.1.

TCAM1P-004 and RP11-598D14.1 acted as tumor suppressors. To characterize the biological functions of TCAM1P-004 and RP11-598D14.1 in HCC, TCAM1P-004 and RP11-598D14.1 were inhibited by at least two independent siRNAs in MIHA cells. MTT cell proliferation assay showed that inhibition of TCAM1P-004 and RP11-598D14.1 expression promoted MIHA cell proliferation on day 4 (Sfig5A -B). We further knocked-down TCAM1P-004 and RP11-598D14.1 in MIHA and L02 cells by lentivirus transduction of shRNAs transgenes. MTT assays showed that the proliferation rates of MIHA and L02 cells were increased after knockdown of TCAM1P-004 and RP11-598D14.1 (Fig4A-D), suggested that depletion of TCAM1P-004 or RP11-598D14.1 could promote the growth of non- tumor hepatocytes.

Annexin V apoptosis assay was performed to measure the number of apoptotic TCAM1P-004-inhibited or RP11-598D14.1-inhibited MIHA cells. There was a nearly 4-fold decrease in the number of post-apoptotic cells in TCAM1P-004- inhibited and RP11-598D14.1-inhibited cells compared to the controls. This indicated that the inhibitions of the lncRNAs contributed to the repression of apoptosis in MIHA cells (Sfig6A). In turn, we would like to study the role of TCAM1P-004 and RP11-598D14.1 in the colony forming ability of cancer cells. Colony formation assays showed that inhibition of TCAM1P-004 and RP11- 598D14.1 could promote colony formation in MIHA and L02 cells (Sfig6B-C). Moreover, soft agar assay was used to study the anchorage-independent growth ability of the transformed cells. Inhibition of TCAM1P-004 or RP11-598D14.1 promoted the anchorage-independent colony formation in MIHA cells (Sfig6D), indicating that TCAM1P-004 and RP11-598D14.1 played a role in inhibiting cell

transformation.

Subsequently, we would like to study the tumorigenic ability of TCAM1P-004 and RP11-598D14.1 in vivo. Hep3B and Huh7 cells were transduced by lentivirus carrying TCAM1P-004 or RP11-598D14.1 transgenes or blank vector control. MTT assay showed that there were significant growth-inhibiting effects after overexpression of TCAM1P-004 and RP11-598D14.1 in both cell lines (Fig4A-D). In addition, overexpression of TCAM1P-004 and RP11-598D14.1 could suppress the clonogenic and transformation ability of HCC cells, as revealed by colony formation assay and soft agar assay respectively (Sfig7A-B). More importantly, overexpressing TCAM1P-004 and RP11-598D14.1 in Hep3B cells inhibited the rate of tumor formation in vivo. We showed that the size of tumors derived from TCAM1P-004 and RP11-598D14.1 overexpressing Hep3B cells were significantly smaller than control group (Fig4E, lower left). Immunochemical staining of proliferation marker KI67 showed that the number of KI67 positive cells were reduced in tumors derived from TCAM1P-004 and RP11-598D14.1 overexpressing Hep3B cells (Fig4E, lower right). These studies demonstrated that TCAM1P-004 and RP11-598D14.1 played important roles in tumorigenesis. In turn, terminal dUTP nick-end labeling (TUNEL) assay was performed to study the effect of TCAM1P-004 or RP11-598D14.1- overexpression in the subcutaneous tumors derived. TUNEL assay showed that apoptosis frequently occurred in tumors derived from TCAM1P-004- or RP11- 598D14.1-overexpressing Hep3B cells (Sfig8). Furthermore, MIHA cells expressing shTCAM1P-004, shRP11-598D14.1 and shSCR were injected subcutaneously into nude mice. The formation of tumor was observed after inhibition of TCAM1P-004 and RP11-598D14.1, but not in the control MIHA cells (n=6) (Sfig9). Collectively, we demonstrated that TCAM1P-004 and RP11-598D14.1 negatively regulate HCC growth by inhibiting cell proliferation and inducing apoptosis.

Microarray analysis revealed genes and signaling pathways related to TCAM1P-004/RP11-598D14.1. To investigate the mechanisms of TCAM1P-004 and RP11-598D14.1, we measured

the changes of gene expression in MIHA cells with TCAM1P-004 or RP11- 598D14.1 knockdown by gene expression microarrays. The lists of genes differentially expressed in MIHA cells with knockdown of TCAM1P-004 and RP11- 598D14.1 were listed (Supplementary table 2 and 3, supplementary file 1 and 2). KEGG pathway analysis showed that the genes differentially expressed after TCAM1P-004 knockdown may be involved in the p53 and MAPK signaling pathway (Sfig10A). Similarly, RP11-598D14.1 regulated genes that were associated with the p53, HIF-1 and MAPK pathways (Sfig10B). As they were both repressed by EZH2, it was interesting to analyze the pathway commonly inhibited. Overlapping genes between TCAM1P-004 and RP11-598D14.1 were used for KEGG pathway analysis, and showed that p53 signaling pathway was commonly enriched (Sfig10C), suggesting that EZH2 heavily deregulated p53 pathway through the repression of lncRNAs. qRT-PCR validation also indicated that p53-associated genes (GADD45A, CDK6, MDM2, PMAIP1 and TSP1) were downregulated whereas CCNG2 expression was increased. GADD45A and TSP1 functioned to suppress angiogenesis (24-25); CDK6, PMAIP1 and CCNG2 regulated cell proliferation and apoptosis (26-28). For MAPK pathways-associated genes, RAP1A, DUSP4, STK3, DDIT3, and HSPA1A were downregulated while FOS was upregulated (Sfig11A). Previous studies reported that RAP1A, DDIT3 and HSPA1A regulated cell proliferation (29-30) and DUSP4 and FOS regulated metastasis (31- 32). In RP11-598D14.1-inhibited cells, genes involved in the p53 pathway (CDK6, GADD45A, PMAIP1, TSP1 and CCNG2) were changed. For genes associated with the HIF-1 pathway, only the expression of PDK4 and PFKFB4 (33-34) was consistent with the microarray data (Sfig11B).

TCAM1P-004 interacted with IGF2BP1 and HIST1H1C and promoted DDIT3 expression. LncRNAs can function as signals, decoys, guides and scaffolds for proteins (35). To determine whether TCAM1P-004 and RP11-598D14.1 exerted their biological functions through protein interaction, RNA pull-down assays followed by LC-MS were performed to identify RNA binding proteins that interacted with TCAM1P-004

and RP11-598D14.1. Proteins bound to TCAM1P-004 with positive detection and showed greater than 95% confidence was listed (Supplementary table 4). By literature review, we selected the potential binding proteins that were associated with cancer for further validation. Western blot validated that IGF2BP1 and HIST1H1C were detected from TCAM1P-004-bound proteins but not the negative control (Fig5A). RIP assay also showed that TCAM1P-004 was obviously enriched with IGF2BP1 and HIST1H1C compared to IgG control (Fig5B).

IGF2BP1 is a RNA-binding protein that functions as a protumorigenic factor in HCC (36). It was reported that IGF2BP1 could bind with c-MYC and MKI67 mRNAs to enhance their protein expressions (36). IGF2BP1 also bound with lncRNA HCG11 to suppress HCC cell apoptosis (37). Meanwhile, the role of HIST1H1C in cancer is not explored. To identify the functional roles of the TCAM1P-004-IGF2BP1, we inhibited IGF2BP1 and HIST1H1C in Hep3B cells with siRNAs (Fig5C) and measured the levels of TCAM1P-004 potential targets revealed from previous microarray. DDIT3 was upregulated after knockdown of IGF2BP1 in Hep3B cells (Fig5D), so we focused on the regulatory functions of IGF2BP1 on DDIT3. It has been reported that IGF2BP1 controls translation of IGF2 (38). IGF2 is associated with the MAPK pathway in which DDIT3 is also involved. (39) Western blot showed that inhibition of IGF2BP1 by transfecting siRNAs significantly upregulated DDIT3 protein levels (Fig5E). After revealing the association between IGF2BP1 and DDIT3, we studied the changes in protein levels of IGF2BP1, IGF2 and DDIT3 in MIHA and L02 cells after TCAM1P-004 was inhibited. While inhibition of TCAM1P-004 had no effect on IGF2BP1, the IGF2 level was increased while DDIT3 was reduced in both TCAM1P-004-inhibited MIHA and L02 cells. (Fig5E). In turn, we measured DDIT3 expression in HCC cell lines and HCC tissues. Compared to MIHA cells, mRNA expression of DDIT3 was downregulated in five different HCC cells lines (Sfig12A). DDIT3 mRNA levels were also significantly decreased in 50 pairs of HCC tumor compared with adjacent normal tissues (Sfig12B). Correlation analysis showed that TCAM1P-004 expression was positively associated with DDIT3 expression (R2=0.1975, p<0.0001)

(Sfig12C). Collectively, our data suggested that TCAM1P-004 interacted with IGF2BP1 to repress IGF2 translation and in turn promoted DDIT3 expression.

We next explored the interaction between TCAM1P-004 and HIST1H1C. First, we knocked-down HIST1H1C in Hep3B and Huh7 cells with siRNAs. Then, we measured the expression of TCAM1P-004 potential target genes revealed from previous gene expression microarray. We showed that DDIT3 expression was again downregulated in Hep3B and Huh7 cells after inhibition of HIST1H1C. Western blot results confirmed the inhibitory effects of HIST1H1C on DDIT3 in HCC cells, indicating that the interaction between TCAM1P-004 and HIST1H1C also contributed to the repression of DDIT3 (Fig5F).

RP11-598D14.1 interacted with IGF2BP1 and STAU1. Proteins bound to RP11-598D14.1 that were detected by RNA pull-down-LC-MS were listed (with positive detection and over 95% confidence) (Supplementary table 5). Subsequent Western blot validated that RP11-598D14.1 was able to bind with IGF2BP1 and STAU1 (Fig6A). RIP assay showed that the endogenous interaction of IGF2BP1 and STAU1 with RP11-598D14.1 could be observed in MIHA cells (Fig6B). Since RP11-598D14.1 was shown to bind with IGF2BP1, we speculated that RP11-598D14.1 could also regulate IGF2 translation and DDIT3 expression. First, we measured the mRNA level of DDIT3 in RP11-598D14.1-inhibited MIHA cells and found that DDIT3 expression was reduced compared to that in the control cells (Fig6C). Inhibition of RP11-598D14.1 had no effect on IGF2BP1 protein expression, while IGF2 protein levels were increased and DDIT3 protein levels were reduced upon the depletion of RP11-598D14.1 expression (Fig6C). Correlation analysis showed that RP11-598D14.1 expression was positively associated with DDIT3 expression (R2=0.2091 and p<0.0001) (Sfig12D). Therefore, we suggested that RP11-598D14.1-IGF2BP1 regulated IGF2 expression that led to the downregulation of DDIT3 expression.

STAU1 is another RNA-binding protein that has important role in cancer. It was

reported that STAU1 could bind to lncRNA TINCR, and promoted gastric cancer progression by regulating KLF2 mRNA stability (40). STAU1 also bound to lncRNA SNHG5 to promote colorectal cancer cell survival (41). However, the role of STAU1 in HCC is not studied. To investigate the biological function of the RP11- 598D14.1-STAU1 interaction, we measured the changes in expression of RP11- 598D14.1-targeted genes after knockdown of STAU1 in Hep3B. Expression of SLC2A3, EGLN3, TNSF15, CCNG2 and PFKFB4 were significantly upregulated after inhibition of STAU1 (Fig6D). Western blot analysis revealed that STAU1 protein level did not change after the knockdown of RP11-598D14.1 in MIHA cells, whereas PFKFB4 expression was upregulated upon the knockdown of RP11- 598D14.1 expression (Fig6E). This provided evidence that RP11-598D14.1-STAU1 complex could repress PFKFB4 expression. PFKFB4 was reported to be an oncogene in cancer (42-43), and TCGA data analysis showed that PFKFB4 level was higher in HCC tumor tissues compared to non-tumor tissues (Sfig13). It indicated that PFKFB4 should play an important role in HCC and the silencing of RP11-598D14.1 might contribute to the upregulation of PFKFB4 in human HCC.

IGF2BP1 and HIST1H1C promoted cell proliferation in HCC To show that the lncRNA-regulatory axis was relevant in vivo, we measured the expression of downstream genes in Hep3B xenograft tumors. Total RNA was extracted from tumors derived from TCAM1P-004- and RP11-598D14.1- overexpressed Hep3B cells and the vector control cells. qRT-PCR showed that TCAM1P-004 and RP11-598D14.1 were upregulated in the TCAM1P-004- overexpressing and RP11-598D14.1-overexpressing tumors respectively (Fig7A). We had previously shown that RP11-598D14.1 could couple with STAU1 to inhibit PFKFB4. qRT-PCR showed that there was a significant downregulation of PFKFB4 in RP11-598D14.1-overexpressing tumor (Fig7B), suggesting that RP11-598D14.1 inhibited PFKBF4 level in vivo. Moreover, TCAM1P-004 and RP11-598D14.1 could couple with IGF2BP1 to reduce the protein level of IGF2, and subsequently increase the expression of DDIT3 in vitro. qRT-PCR showed that DDIT3 expressions were significantly increased in both TCAM1P-004-overexpressed and

RP11-598D14.1-overexpressed tumors (Fig7C). These evidences suggested that the lncRNA-regulatory pathways identified were functioning in vivo. Furthermore, we also investigated the biological functions of IGF2BP1 and HIST1H1C in HCC cells through knockdown of their expression by siRNAs. The cell proliferation of Hep3B and Huh7 cells were hindered after the knockdown of IGF2BP1 or HIST1H1C (Fig7D and 7E). Taken together, we demonstrated that IGF2BP1 and HIST1H1C promoted cell proliferation in HCC and their functions could be regulated by TCAM1P-004 and RP11-598D14.1.

Discussion In this study, we revealed a novel function of EZH2 in the repression of lncRNAs TCAM1P-004 and RP11-598D14.1 in HCC and expanded the landscape of the epigenetic regulatory roles of EZH2 and lncRNAs in HCC. The schematic diagram demonstrated the mechanism in promoting HCC growth via the aberrant suppression of TCAM1P-004 and RP11-598D14.1 by EZH2 (Fig7F). EZH2 within the PRC2 complex bound to the promoters of TCAM1P-004 and RP11-598D14.1, resulted in the epigenetic silencing of both lncRNAs. Downregulation of TCAM1P-004 and RP11-598D14.1 reduced the level of lncRNA-protein complexes including TCAM1P-004-IGF2BP1, TCAM1P-004-HIST1H1C, RP11-598D14.1-IGF2BP1 and RP11-598D14.1-STAU1. IGF2BP1, HIST1H1C and STAU1 released from the interactions from TCAM1P-004 and RP11-598D14.1 could then repressed pro- apoptotic factor DDIT3 or promoted anti-apoptotic factor PFKFB4 that heavily contributed to HCC tumorigenesis (Fig7F).

IGF2BP1 belongs to the conserved protein family IGF2BP (44). The IGF2BP family proteins were heavily involved in cancer biology, and high expression of IGF2BP causes an aggressive malignancy phenotype (45). Consistently, we also observed that IGF2BP1 could promote cell proliferation in HCC cells. As TCAM1P-004 and RP11-598D14.1 participated in the suppression of cell proliferation, we hypothesized that TCAM1P-004 and RP11-598D14.1 could block the biological function of IGF2BP1 through physical interaction. The ability of TCAM1P-004 and

RP11-598D14.1 competing to bind with oncogenic RNA-binding proteins underlay parts of their tumor suppressor functions. Here, the interactions of TCAM1P-004 and RP11-598D14.1 with IGF2BP1 resulted in the disruption of IGF2BP1-promoted IGF2 translation that led to promotion of cell apoptosis and reduction of cell proliferation. As both TCAM1P-004 and RP11-598D14.1 were able to bind with IGF2BP1, so they were expected to share certain degree of sequence homology. However, alignment of the two lncRNAs showed low similarity (29.7%) on their RNA sequences (Sfig14), suggesting that they might not share similar RNA structure or motif. The IGF2BP1 has two RNA recognition motifs and a RGG RNA-binding domain, which show different degree of preference for RNA but often display degenerate binding specificity. Given that IGF2 signaling was critical in carcinogenesis, multilevel of regulations should be present to tightly control it in non-tumor cells. Thus, it is not surprising that multiple lncRNAs could function to regulate the IGF2 signaling.

The other TCAM1P-004-binding protein HIST1H1C, which belongs to the histone H1 family, plays important roles in various cancers. Song et al. observed that HIST1H1C is the downstream targeting gene of heat shock protein 90 in pancreatic cancer (46). In human breast cancer cells, HIST1H1C was found to have low abundance at the transcription start sites of the inactive genes and was proven to be intensively correlated with low gene expression (47). In HCC, we observed that knockdown of HIST1H1C caused a decreased expression of DDIT3, indicating that HIST1H1C interaction with TCAM1P-004 may promote DDIT3 expression. Although the TCAM1P-004-HIST1H1C complex could downregulate DDIT3 expression which induced cell apoptosis, inhibition of HIST1H1C suppressed cell proliferation in both Hep3B and Huh7 cells. HIST1H1C may exhibit alternative molecular functions independent of TCAM1P-004 that promoted cell proliferation.

Here, we demonstrated that DDIT3 was one of the critical proteins repressed upon the loss of TCAM1P-004 and RP11-598D14.1. DDIT3, also called CHOP and GADD153, is a key regulator under cellular stress. Studies showed that DDIT3

regulates cell migration, proliferation, cell apoptosis and survival in many cancers (48-50). Low DDIT3 expression in cell lines and human samples may imply a critical role of DDIT3 in HCC. Furthermore, we found that PFKFB4 may be the downstream targeting gene of the RP11-598D14.1-STAU1 complex. PFKFB4 is reported to be an oncogene in glioblastoma and bladder cancer (42-43). TCGA analysis also showed that upregulation of PFKFB4 expression was frequently observed in HCC tumor samples. Currently, the role of PFKFB4 in HCC was largely unexplored. Further characterization of this protein may reveal its importance in human HCC.

In addition to EZH2, EZH1 was another mammalian homolog in the EZH family that served similar molecular functions but showed distinct expression pattern in different developmental stages. There was no study reporting that EZH1 had any role in the development or progression of HCC. To exclude the potential involvement of EZH1 in human HCC, we measured the EZH1 mRNA expression in a panel of HCC cell lines and non-tumor hepatocyte cell lines, and showed that the level of EZH1 expression had no significant difference between HCC and non-tumor cells (Sfig15A). We also attempted to exclude the unspecific effect of DZnep on EZH1, and showed that DZnep treatment had no effect on the mRNA level of EZH1 (Sfig15B). Indeed, we had performed alternative approaches to specifically inhibit EZH2 in HCC cells (e.g. shRNAs and siRNAs targeting EZH2 specific regions), and demonstrated that the affected lncRNAs were putative EZH2 regulating targets.

Collectively, this work further elucidates that the biological roles of EZH2. EZH2 not only regulates protein-coding genes and miRNAs expression, but also suppresses the expression of lncRNAs TCAM1P-004 and RP11-598D14.1. Since TCAM1P-004 and RP11-598D14.1 were significantly reduced in human HCC samples and negatively associated with survival time of HCC patients, TCAM1P- 004 and RP11-598D14.1 could be potential diagnostic markers for HCC, which could help to identify HCC at the early stage. More importantly, TCAM1P-004 and RP11-598D14.1 played tumor suppressor functions by inhibiting cell proliferation

and tumor growth. The association of cell apoptosis with the EZH2-TCAM1P- 004/IGF2BP1-IGF2-DDIT3, EZH2-RP11-598D14.1/IGF2BP1-IGF2-DDIT3 and EZH2-TCAM1P-004/HIST1H1C-DDIT3 pathways suggests that TCAM1P-004 and RP11-598D14.1 could be used in the development of a novel therapeutic strategy.

Acknowledgements The work described in this paper was supported by grants from the General Research Fund, Research Grants Council of Hong Kong (CUHK462713, 14102714 and 14136416), National Natural Science Foundation of China (8142730) and Direct Grant from CUHK to YC.

Reference: 1. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2018. CA Cancer J Clin 2018;68:7-30. 2. El-Serag HB. Hepatocellular Carcinoma. N Engl J Med 2011;365:1118-1127. 3. Tagliamonte M, Petrizzo A, Tornesello ML, Ciliberto G, Buonaguro FM, Buonaguro L. Combinatorial immunotherapy strategies for hepatocellular carcinoma. Curr Opin Immunol 2016;39:103-13. 4. Fong ZV, Tanabe KK. The Clinical Management of Hepatocellular Carcinoma in the United States, Europe, and Asia A Comprehensive and Evidence-Based Comparison and Review. Cancer 2014;120:2824-2838. 5. Luk JM, Liu AM. Proteomics of Hepatocellular Carcinoma in Chinese Patients. Omics 2011;15:261-266. 6. Mittal S, El-Serag HB. Epidemiology of Hepatocellular Carcinoma Consider the Population. J Clin Gastroenterol 2013;47:S2-S6. 7. Sparmann A, van Lohuizen M. Polycomb silencers control cell fate, development and cancer. Nat Rev Cancer 2006;6:846-56. 8. Kim K, Jutooru I, Chadalapaka G, Johnson G, Frank J, Burghardt R, et al. HOTAIR is a negative prognostic factor and exhibits pro-oncogenic activity in pancreatic cancer. Oncogene 2013;32:1616-25.

9. Bachmann IM, Halvorsen OJ, Collett K, Stefansson IM, Straume O, Haukaas SA, et al. EZH2 expression is associated with high proliferation rate and aggressive tumor subgroups in cutaneous melanoma and cancers of the endometrium, prostate, and breast. J Clin Oncol 2006;24:268-273. 10. Chen Y, Lin MC, Yao H, Wang H, Zhang AQ, Yu J, et al. Lentivirus-mediated RNA interference targeting enhancer of zeste homolog 2 inhibits hepatocellular carcinoma growth through down-regulation of stathmin. Hepatology 2007;46:1314- 1314. 11. Gao SB, Zheng QF, Xu B1, Pan CB, Li KL, Zhao Y, et al. EZH2 Represses Target Genes through H3K27-Dependent and H3K27-Independent Mechanisms in Hepatocellular Carcinoma. Mol Cancer Res. 2014;12:1388-1397. 12. Gao SB, Xu B, Ding LH, Zheng QL, Zhang L, Zheng QF, et al. The functional and mechanistic relatedness of EZH2 and menin in hepatocellular carcinoma. J Hepatol 2014;61:832-839. 13. Au SL, Wong CC, Lee JM, Fan DN, Tsang FH, Ng IO, et al. Enhancer of zeste homolog 2 epigenetically silences multiple tumor suppressor microRNAs to promote liver cancer metastasis. Hepatology 2012;56:622-631. 14. Li CH, To KF, Tong JH, Xiao Z, Xia T, Lai PB, et al. Enhancer of zeste homolog 2 silences microRNA-218 in human pancreatic ductal adenocarcinoma cells by inducing formation of heterochromatin. Gastroenterology 2013;144:1086-1097.e9. 15. Mercer TR, Dinger ME, Mattick JS. Long non-coding RNAs: insights into functions. Nat Rev Genet 2009;10:155-159. 16. Li CH, Chen Y. Targeting long non-coding RNAs in cancers: progress and prospects. Int J Biochem Cell Biol 2013;45:1895-910. 17. Li CH, Xiao Z, Tong JH, To KF, Fang X, Cheng AS, et al. EZH2 coupled with HOTAIR to silence MicroRNA-34a by the induction of heterochromatin formation in human pancreatic ductal adenocarcinoma. Int J Cancer 2017;140:120-129. 18. Quagliata L, Matter MS, Piscuoglio S, Arabi L, Ruiz C, Procino A, et al. Long noncoding RNA HOTTIP/HOXA13 expression is associated with disease progression and predicts outcome in hepatocellular carcinoma patients. Hepatology 2014;59:911-923.

19. Panzitt K, Tschernatsch MM, Guelly C, Moustafa T, Stradner M, Strohmaier HM, et al. Characterization of HULC, a novel gene with striking up-regulation in hepatocellular carcinoma, as noncoding RNA. Gastroenterology 2007;132:330-342. 20. Xu C, Yang M, Tian J, Wang X, Li Z. MALAT-1: a long non-coding RNA and its important 3' end functional motif in colorectal cancer metastasis. Int J Oncol 2011;39:169-175. 21. Yang X, Xie X, Xiao YF, Xie R, Hu CJ, Tang B, Li BS, Yang SM. The emergence of long non-coding RNAs in the tumorigenesis of hepatocellular carcinoma. Cancer Lett. 2015;360:119-124. Review. 22. Klingenberg M, Matsuda A, Diederichs S, Patel T. Non-coding RNA in hepatocellular carcinoma: Mechanisms, biomarkers and therapeutic targets. J Hepatol 2017;67:603-618. 23. Xiao Z, Li CH, Chan SL, Xu F, Feng L, Wang Y, et al. A small-molecule modulator of the tumor-suppressor miR34a inhibits the growth of hepatocellular carcinoma. Cancer Res 2014;74:6236-6247. 24. Zhang D, Zhang W, Li D, Fu M, Chen R, Zhan Q. GADD45A inhibits autophagy by regulating the interaction between BECN1 and PIK3C3. Autophagy 2015; 11:2247-58. 25. Lawler J. Thrombospondin-1 as an endogenous inhibitor of angiogenesis and tumor growth. J Cell Mol Med 2002;6:1-12. 26. Kikkawa N, Kinoshita T, Nohata N, Hanazawa T, Yamamoto N, et al. microRNA-504 inhibits cancer cell proliferation via targeting CDK6 in hypopharyngeal squamous cell carcinoma. Int J Oncol 2014;44:2085-92. 27. Ishida M, Sunamura M, Furukawa T, Lefter LP, Morita R, Akada M, et al. The PMAIP1 gene on chromosome 18 is a candidate tumor suppressor gene in human pancreatic cancer. Dig Dis Sci 2008;53:2576-82. 28. Sun GG, Zhang J, Hu WN. CCNG2 expression is downregulated in colorectal carcinoma and its clinical significance. Tumour Biol 2014;35:3339-46. 29. Lu L, Wang J, Wu Y, Wan P, Yang G. Rap1A promotes ovarian cancer metastasis via activation of ERK/p38 and notch signaling. Cancer Med 2016;5:3544-3554. 30. Balko JM, Schwarz LJ, Bhola NE, Kurupi R, Owens P, Miller TW, et al.

Activation of MAPK pathways due to DUSP4 loss promotes cancer stem cell-like phenotypes in basal-like breast cancer. Cancer Res 2013;73:6346-58. 31. Fan Q, He M, Deng X, Wu WK, Zhao L, Tang J, et al. Derepression of c-Fos caused by microRNA-139 down-regulation contributes to the metastasis of human hepatocellular carcinoma. Cell Biochem Funct 2013;31:319-24. 32. Lee JH, Kim EJ, Kim DK, Lee JM, Park SB, Lee IK, et al. Hypoxia induces PDK4 gene expression through induction of the orphan nuclear receptor ERRγ. PLoS One 2012;7:e46324. 33. Zhang H, Lu C, Fang M, Yan W, Chen M, Ji Y, et al. HIF-1α activates hypoxia- induced PFKFB4 expression in human bladder cancer cells. Biochem Biophys Res Commun 2016;476:146-152. 34. Xie M, Sun M, Zhu YN, Xia R, Liu YW, Ding J, et al. Long noncoding RNA HOXA-AS2 promotes gastric cancer proliferation by epigenetically silencing P21/PLK3/DDIT3 expression. Oncotarget 2015;6:33587-601. 35. Wang KC, Chang HY. Molecular mechanisms of long noncoding RNAs. Mol Cell 2011;43:904-14. 36. Gutschner T, Hämmerle M, Pazaitis N, Bley N, Fiskin E, Uckelmann H, et al. Insulin-like growth factor 2 mRNA-binding protein 1 (IGF2BP1) is an important protumorigenic factor in hepatocellular carcinoma. Hepatology 2014;59:1900-1911. 37. Xu Y, Zheng Y, Liu H, Li T. Modulation of IGF2BP1 by long non-coding RNA HCG11 suppresses apoptosis of hepatocellular carcinoma cells via MAPK signaling transduction. Int J Oncol 2017;51:791-800. 38. Nielsen J, Christiansen J, Lykke-Andersen J, Johnsen AH, Wewer UM, Nielsen FC. A family of insulin-like growth factor II mRNA-binding proteins represses translation in late development. Mol Cell Biol 1999;19:1262–1270. 39. Scott DW, Longpre JM, Loo G. Upregulation of GADD153 by butyrate: involvement of MAPK. DNA Cell Biol 2008;27:607-14. 40. Xu TP, Liu XX, Xia R, Yin L, Kong R, Chen WM, et al. SP1-induced upregulation of the long noncoding RNA TINCR regulates cell proliferation and apoptosis by affecting KLF2 mRNA stability in gastric cancer. Oncogene 2015;34:5648-5661.

41. Damas ND, Marcatti M, Côme C, Christensen LL, Nielsen MM, Baumgartner R, et al. SNHG5 promotes colorectal cancer cell survival by counteracting STAU1- mediated mRNA destabilization. Nat Commun 2016;7:13875. 42. Goidts V, Bageritz J, Puccio L, Nakata S, Zapatka M, Barbus S, et al. RNAi screening in glioma stem-like cells identifies PFKFB4 as a key molecule important for cancer cell survival. Oncogene 2012;31:3235-43. 43. Yun SJ, Jo SW, Ha YS, Lee OJ, Kim WT, Kim YJ, et al. PFKFB4 as a prognostic marker in non-muscle-invasive bladder cancer. Urol Oncol 2012;30:893-9. 44. Bell JL, Wächter K, Mühleck B, Pazaitis N, Köhn M, Lederer M, et al. Insulin- like growth factor 2 mRNA-binding proteins (IGF2BPs): post-transcriptional drivers of cancer progression? Cell Mol Life Sci 2013;70:2657-2675. 45. Bell JL, Wächter K, Mühleck B, Pazaitis N, Köhn M, Lederer M, et al. Insulin- like growth factor 2 mRNA-binding proteins (IGF2BPs): post-transcriptional drivers of cancer progression? Cell Mol Life Sci 2013;70:2657-75. 46. Song D, Chaerkady R, Tan AC, García-García E, Nalli A, Suárez-Gauthier A, et al. Antitumor activity and molecular effects of the novel heat shock protein 90 inhibitor, IPI-504, in pancreatic cancer. Mol Cancer Ther 2008;7:3275-84. 47. Millán-Ariño L, Islam AB, Izquierdo-Bouldstridge A, Mayor R, Terme JM, Luque N, et al. Mapping of six somatic linker histone H1 variants in human breast cancer cells uncovers specific features of H1.2. Nucleic Acids Res 2014;42:4474-93. 48. Jauhiainen A, Thomsen C, Strömbom L, Grundevik P, Andersson C, Danielsson A, et al. Distinct Cytoplasmic and Nuclear Functions of the Stress Induced Protein DDIT3/CHOP/GADD153. Plos One 2012;7. 49. Marciniak SJ, Yun CY, Oyadomari S, Novoa I, Zhang Y, Jungreis R, et al. CHOP induces death by promoting protein synthesis and oxidation in the stressed endoplasmic reticulum. Genes & Development 2004;18:3066-3077. 50. He K, Zheng X, Li M, Zhang L, Yu J. mTOR inhibitors induce apoptosis in colon cancer cells via CHOP-dependent DR5 induction on 4E-BP1 dephosphorylation. Oncogene 2016;35:148-57.

Fig1. Identification of two upregulated lncRNAs in HCC cells treated with EZH2 inhibitor DZNep. (A) Protein levels of EZH2 and H3K27me3 were decreased in Huh7 cells and Hep3B cells treated with 10 μM DZnep (B) qRT-PCR validation of microarray analysis showed that levels of TCAM1P-004 and RP11- 598D14.1 were increased in DZnep-treated HCC cells. (C) Their expressions were frequently reduced in HCC cells compared to non-tumor MIHA and L02 cells. (D) he abundance of TCAM1P-004 and RP11-598D14.1 were lower in human HCC tumor compared to non-tumor tissues with statistical significance. (E) Kaplan-Meier plots revealed the associations of high TCAM1P-004 and RP11-598D14.1 expressions with a longer overall survival for HCC patients. Significant differences were analyzed using Mantel-Cox test (n=50). Results were expressed as the mean ± SEM.

Fig2. EZH2 and PRC2 complex inhibited TCAM1P-004 and RP11-598D14.1 expressions. (A) Overexpression of EZH2 downregulated TCAM1P-004 and RP11- 598D14.1 in MIHA cells. (B) Inhibition of EZH2 by shRNAs upregulated both lncRNAs in Hep3B cells. (C-D) Transient inhibition of EZH2 by siRNAs upregulated both lncRNAs in (C) Hep3B and (D) Huh7 cells. (E-F) Disrupting the formation of PRC2 through inhibiting EED and SUZ12 by siRNAs in Hep3B and Huh7 cell led to the upregulation of TCAM1P-004 and RP11-598D14.1 in both (E) Hep3B and (F) Huh7 cells.

Fig3. EZH2 recruited PRC2 partners and induced H3K27 trimethylation to suppress TCAM1P-004 and RP11-598D14.1 expression. (C) Enrichment levels of EZH2, H3K27me3, SUZ12 and EED at the promoter regions of TCAM1P-004 and RP11-598D14.1 were higher in Hep3B cells compared to MIHA cells. (D) Inhibition of EZH2 in Hep3B reduced the occupancies of EZH2, H3K27me3, SUZ12 and EED at the two lncRNAs’ promoters. (E) Overexpression of EZH2 in MIHA cells increased occupancies of EZH2 and PRC2 partners in the promoters of TCAM1P- 004 and RP11-598D14.1. Results are expressed as the mean ± SEM. * p<0.05, ** p<0.01, *** p<0.001

Fig4. TCAM1P-004 and RP11-598D14.1 functioned to regulate cell proliferation in hepatocytes and HCC cells. Knockdown of (A) TCAM1P-004 or (B) RP11-598D14.1 promoted cell proliferation in MIHA and L02 cells. Results are expressed as the mean ± SEM. * p<0.05, *** p<0.001. Growth rates of (G-H) TCAM1P-004-overexpressed or (D) RP11-598D14.1-overexpressed HCC cells were decreased compared to control cells. (E, upper panel) Tumor growth for Hep3B cells overexpressed with TCAM1P-004 or RP11-598D14.1 was decreased compared to vector control cells. (Lower left panel) The tumors derived from TCAM1P-004 and RP11-598D14.1 overexpressed cells were smaller than control cells macroscopically (Lower right panel). Immunohistochemical staining showed that tumors derived from TCAM1P-004 and RP11-598D14.1 overexpressed Hep3B cells had reduced number of KI67 positive cells.

Fig5. TCAM1P-004 interacted with IGF2BP1 and HIST1H1C, and activated DDIT3 expression. (A) In vitro RNA-protein binding assay and mass spectrometry was conducted to identify TCAM1P-004 interacting proteins. IGF2BP1 and HIST1H1C were validated to be the binding proteins of TCAM1P-004 by Western blot. (B) RIP showed that IGF2BP1 and HIST1H1C interacted endogenously with TCAM1P-004. (C) Inhibition efficiency of siRNAs targeting IGF2BP1 in Hep3B cells were validated by qRT-PCR. (D) TCAM1P-004-related genes involved in MAPK pathway were measured in IGF2BP1-inhibited Hep3B cells. Knockdown of IGF2BP1 increased mRNA levels of GADD45A and DDIT3 in Hep3B cells. (E) Upon knockdown of IGF2BP1, the DDIT3 protein level was upregulated while IGF2 protein level was downregulated in Hep3B cells. In TCAM1P-004-inhibited MIHA and L02 cells, IGF2 protein level was increased while DDIT3 protein level was reduced, but there was no change in IGF2BP1 expression. (F) mRNA and protein levels of DDIT3 were downregulated in HIST1H1C-inhibited Hep3B and Huh7 cells. Results are expressed as the mean ± SEM.

Fig6. RP11-598D14.1 interacted with IGF2BP1 and STAU1. (A) In vitro RNA-

protein binding assay and mass spectrometry were conducted to identify RP11- 598D14.1 interacting proteins. Western blotting validated that IGF2BP1 and STAU1 were the binding proteins of RP11-598D14.1. (B) RIP showed that IGF2BP1 and STAU1 interacted endogenously with RP11-598D14.1. (C) mRNA level and protein level of DDIT3 were reduced in RP11-598D14.1-inhibited MIHA cells. (D) Knockdown of STAU1 increased PFKFB4 mRNA expression in Hep3B cells. (E) Knockdown of RP11-598D14.1 increased PFKFB4 protein level. STAU1 protein had no change in RP11-598D14.1-inhibited MIHA cells. Results are expressed as the mean ± SEM.

Fig7. Knockdown of IGF2BP1 and HISIT1HIC suppressed HCC cell growth. (A) qRT-PCR showed that TCAM1P-004 and RP11-598D14.1 were upregulated in the TCAM1P-004-overexpressed tumor and RP11-598D14.1-overexpressed tumor respectively. (B) PFKFB4 was significantly downregulated in RP11-598D14.1- overexpressed tumor. (C) DDIT3 expressions were significantly increased in both TCAM1P-004-overexpressed and RP11-598D14.1-overexpressed tumors. (D) MTT assay showed that inhibition of IGF2BP1 could repress cell proliferation in Hep3B and Huh7 cells. (E) Cell proliferation of Hep3B cells and Huh7 cells was suppressed after treated with siRNAs targeting HIST1H1C. Results are expressed as the mean±SEM. ** p<0.01, *** p<0.001. (F) Hypothetical model of tumor suppressive roles of TCAM1P-004 and RP11-598D14.1 in HCC. EZH2 coupled with PRC2 partners to silence TCAM1P-004 and RP11-598D14.1 through inducing H3K27me3. Downregulation of TCAM1P-004 and RP11-598D14.1 resulted in the suboptimal level of TCAM1P-004-IGF2BP1, RP11-598D14.1-IGF2BP1 and TCAM1P-004- HIST1H1C, which subsequently decreased DDIT3 expression and promoted PFKFB4 expression.

Genome-wide screening and functional analysis identifies tumor suppressor long non-coding RNAs epigenetically silenced in hepatocellular carcinoma

Feiyue Xu, Chi Han Li, Chi Hin Wong, et al.

Cancer Res Published OnlineFirst February 4, 2019.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-18-1659

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2019/02/02/0008-5472.CAN-18-1659.DC1

Author Author manuscripts have been peer reviewed and accepted for publication but have not yet been Manuscript edited.

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2019/02/02/0008-5472.CAN-18-1659. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.