CBX8 Exhibits Oncogenic Activity Via AKT/Β-Catenin Activation In

Author Manuscript Published OnlineFirst on October 24, 2017; DOI: 10.1158/0008-5472.CAN-17-0700 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

1Sun Yat-sen University Cancer Center; State Key Laboratory of Oncology in South China; Collaborative Innovation Center for Cancer Medicine, Guangzhou 510060, China

2Department of Pathology, Sun Yat-sen University Cancer Center, Guangzhou 510060, China

#These authors contributed equally to this work.

Short title: CBX8 activates AKT/β-catenin pathway *To whom correspondence should be addressed: Jing-Ping Yun, M.D. Ph.D, Department of Pathology, Sun Yat-sen University Cancer Center, Guangzhou, 510060, Guangdong, China. Tel./Fax: +8620-8734-3693; Email: [email protected]. Conflict of interest: All authors declare no conflict of interest. Financial support: The study was supported by grants from the National Key R&D Program of China (2017YFC1309003 (D Xie and JP Yun)), the National Natural Science Foundation of China (No. 81572406 (JP Yun), 81572405 (CZ Zhang)).

Keywords: CBX8; EGR1, miR-365a-3p; AKT/β-catenin; hepatocellular carcinoma

Deregulation of genes involved in chromatin modification has been increasingly implicated in tumor development and progression. One of the chromatin modifiers is the Polycomb repressive complex (PRC1 and PRC2) (6). Under normal circumstances, PRC1 maintains the histone methylation induced by PRC2 to pass on the inactivation signals (7). Canonical PRC1 includes four subunits: Ring E3 ubiquitin ligase, Polyhomeotic, Posterior sex combs, and Polycomb (8). The cores of PRC1 complexes are Polycomb group (PcG) proteins, which were first identified as developmental regulators in Drosophila (9). Chromobox homolog 8 (CBX8), also known as human Polycomb 3 (HPC3), functions as a transcriptional repressor in PRC1. For example, CBX8 inhibited the expression of INK4a/ARF to bypass cell senescence in fibroblasts (10). However, a later study showed that PRC1 without CBX8 was capable of suppressing the INK4a/ARF locus (11), suggesting an unclear role of CBX8 in transcriptional regulation. Recently, CBX8 has been demonstrated to exert oncogenic functions in a non-canonical manner in human malignancies. Lee et al. reported that CBX8 cooperated with SIRT1 to suppress premature senescence and growth arrest in breast carcinoma (12). The PRC1-BCOR- CBX8 complex is required for BCL6-mediated lymphomagenesis (13). Chung and colleagues proposed that CBX8 transcriptionally activated genes involved in the Notch pathway promote breast cancer (14). However, the role of CBX8 in HCC remains unclear.

The β-catenin signaling pathway is well known for its role in driving carcinogenesis (15). β-catenin can be frequently found at the cell surface. Following stimulation, β- catenin is activated, partly by phosphorylation at Ser552, to dissociate in the cytoplasm. The cytosolic accumulation of β-catenin leads to its localization to the nucleus where it triggers the transcription of various oncogenes, including LEF1/TCF, to promote cancer initiation and progression (16). The activation of β-catenin, one of the factors contributing to hepatocarcinogenesis (17), is typically mediated by Wnt. Recent studies showed that the β-catenin signaling axis can be triggered by AKT via either phosphorylation of β-catenin at Ser552, which enhances its transcriptional activity (18), or suppression of GSK-3β (19). However, the regulation of the AKT/β-catenin 3

Using tissue microarray (TMA)-based immnunohistochemistry, gene microarrays, and biological function assays, we identified CBX8 as an oncogene in HCC. CBX8 expression is increased and associated with poor outcomes of patients with HCC. We further demonstrate that CBX8 promotes HCC progression in vitro and in vivo by up-regulating EGR1 and miR-365a-3p to subsequently activate the AKT/β-catenin pathway. Collectively, our functional and biochemical studies suggest CBX8 exhibits oncogenic activities towards HCC in a non-canonical fashion.

Materials and Methods Cell culture Human HCC cell lines Huh7, MHCC-97H, and HepG2 were purchased from the American Type Culture Collection. The HCC cell lines SK-Hep1, QGY-7701, QGY-7703, Bel-7402, Bel-7404, and SMMC-7721 were obtained from the Cell Resource Center, Chinese Academy of Science Committee. All cell lines were authenticated by Genecreate Company (Wuhan, China) 5 months before this study. Cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, Gaithersburg, MD, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS, Hyclone, Logan,

UT) at 37 °C in a humidified incubator containing 5% CO2. Cells were treated with RING inhibitor PRT4165 (MedChemExpress Company) or transfected with overexpression vectors, siRNA, shRNA, miR-365a-3p mimics, or corresponding empty vectors as described in Supplementary Table 1.

Patients and tissue specimens A cohort of 514 patients with HCC who received surgery between January 2005 and January 2009 was recruited at the Sun Yat-sen University Cancer Center, Guangzhou, China. Paraffin-embedded tissues and clinical information were collected. None of the patients had received radiotherapy or chemotherapy before surgery. Written informed consents from the patients were obtained. Another 56 pairs of fresh HCC and adjacent non-tumorous liver specimens were collected from the patients at the time of surgical resection. All fresh samples were anonymous. This study was approved by the Sun Yat-sen University Cancer Center Institute Research Ethics Committee and conducted in accordance with International Ethical Guidelines for Biomedical Research Involving Human Subjects (CIOMS). The prognostic value of CBX8 was further validated in a cohort of The Cancer Genome Atlas (TCGA) dataset (http://www.cbioportal.org).

Microarray and RNA sequencing microRNA microarrays (SYSUCC, Guangzhou, China) was used to detect the changes of microRNAs in MHCC-97H cells transfected with CBX8 siRNAs. The RNA-sequencing was performed in Bel-7402 cells with or without CBX8 4

Quantitative real-time RT–PCR (qRT-PCR) Total RNA was extracted by Trizol Reagent (Invitrogen, Carlsbad, CA, USA). For miR-365a-3p detection, reverse-transcribed cDNA was synthesized with the miRCURY LNATM universal cDNA synthesis Kit (Exiqon, Vedbaek, Denmark). For mRNA analyses, cDNA was synthesized using Moloney murine leukemia virus reverse transcriptase (Promega, Madison, WI, USA). qRT–PCR was performed with SYBR Premix ExTaq (TaKaRa, Dalian, China) with the Stratagene Mx3000P real-time PCR system (Agilent Technologies, Inc., Santa Clara, CA, USA). Expression levels of miR-365a-3p were normalized against the endogenous snRNA U6 control, while 18s rRNA was used as internal control for mRNA quantification. The relative expression ratio of genes in each paired tumor and non-tumorous tissue was calculated by the −ΔCt method. The sequences of the PCR primers are shown in Supplementary Table 1.

Immunohistochemistry (IHC) Antibodies for IHC and western analyses are listed in Supplementary Table 2. Immunohistochemical staining for CBX8, EGR1, and p-β-catenin was performed on an HCC tissue microarray. The expression levels were scored as a proportion of the immunopositive staining area (0%, 0; 1–25%, 1; 26–50%, 2; 51–75%, 3; and 76%–100%, 4) multiplied by intensity of the staining (0, negative; 1, weak; 2, moderate; and 3, intense). The scores were independently rendered by two pathologists. The median IHC score was chosen as the cut-off value for defining high and low expression.

Immunofluorescence (IF) Cells were fixed for 20 min in PBS containing 4% paraformaldehyde, permeabilized in 0.1% Triton X-100 two times, 5 min each, incubated in blocking buffer (3% donkey serum in TBS) for 1 h, and then incubated with antibody for 2 h at room temperature. After washing in PBS three times, 8 min each, cells were incubated with the appropriate fluorochrome-conjugated secondary antibody for 1 h, and observed under a fluorescence microscope.

Co-Immunoprecipitation (Co-IP) Proteins were extracted by radioimmunoprecipitation assay buffer supplemented with proteinase inhibitor cocktail. Primary antibodies were added for 2.5 h. Protein A/G beads were added for an additional 2 h. Precipitated proteins were dissolved in sodium dodecyl sulfate (SDS) loading buffer and fractionated by SDS polyacrylamide gel electrophoresis.

Tetrazolium dye (MTT) and colony formation assays For MTT assays, 3.0 × 103 cells 5

Cell migration assays About 2 × 103 cells were resuspended in 200 µL of serum-free medium and then placed in the upper compartment of a Transwell chamber (Corning; 24-well insert, pore size: 8 µm). The lower chamber was filled with 15% fetal bovine serum as a chemo-attractant and incubated for 48 h. At the end of the experiments, the cells on the upper surface of the membrane were removed, and the cells on the lower surface were fixed with methanol and stained with 0.1% crystal violet and counted. Five visual fields were randomly chosen and the number of cells was counted under a microscope.

Luciferase reporter assay For the binding of CBX8 to the EGR1 or miR-365a-3p promoter, and miR-365a-3p to the ZNRF1 3'-UTR, Bel-7402 cells were co-transfected with the overexpression or empty vector. Cells were collected 48 h after transfection and luciferase activity was analyzed with the Dual-Luciferase Reporter Assay System (Promega, CA, USA).

Chromosome immunoprecipitation (ChIP) The ChIP assay was performed using the SimpleChIP® Enzymatic Chromatin IP Kit (#9002, Cell Signaling Technology). HCC cells with CBX8 modulation were lysed using SDS lysis buffer and DNA was sheared by sonication. Protein-DNA complexes were precipitated by control immunoglobulin G or CBX8 antibody, followed by eluting the complex from the antibody. qRT-PCR was carried out with primers specific for EGR1 promoter region. The primers used in ChIP assay were described in Supplementary Table 1.

Animal experiments Stable HCC cells with CBX8 overexpression or depletion were implanted subcutaneously under the right armpits and into the flanks of male BALB/c- nude mice aged 3–4 weeks. Tumor size and body weight were measured every 4 days. Four weeks later, the mice were sacrificed and tumor weight and size were measured. Volumes were calculated using the following formula: volume = length × width2 × 0.5. For metastasis observation, four-week-old male nude BALB/c athymic mice were injected with 5 × 105 cells via the tail vein. Six weeks later, mice were killed. Lungs of the mice were fixed and stained with hematoxylin and eosin. Lung metastasis was

Statistical analysis Data from at least three experiments were detected as mean values ± SEM. Differences among various groups were compared by Student’s t-test. Spearman’s correlation test was used to evaluate the correlations between parameters. Survival was analyzed using the Kaplan-Meier method, while factors associated with survival were identified by the Cox proportional hazard regression model. Differences were considered significant when the P value was less than 0.05. All analysis was performed in statistical software SPSS Statistics 19.0 (IBM).

Results CBX8 expression is increased and associated with poor outcomes in HCC The expressions of Chromobox family genes (CBX1-8) were determined in HCC fresh tissues. Results showed that the mRNA expression of CBX8, as well as CBX1, CBX4 and CBX6, was significantly up-regulated in HCC (Supplementary Figure 1A). Data from TCGA datasets revealed that genomic amplification of the CBX8 gene being more frequently found in HCC and other cancers (Supplementary Figure 1B&2). The mRNA and protein levels were obviously higher in HCC cells than those in the nontumorous tissues (Figure 1A). In 56 pairs of fresh specimens, CBX8 mRNA was significantly up- regulated in tumor tissues, compared to that in the non-tumorous ones (Figure 1B). Similarly, CBX8 protein expression was increased by 4.35-fold in HCC cells (Figure 1C). Clear up-regulation, comparable expression and clear down-regulation of CBX8 protein expression in HCC samples were detected in 55.6% (15/27), 25.9% (7/27) and 18.5% (5/27), respectively.

TMA-based IHC analysis showed positive staining of CBX8 in 27.2% (140/514) of the HCC tissues but in only 2.5% (13/514) of the non-tumorous tissues (Figure 1D). Significantly association of CBX8 and clinical stage was found (Supplementary Table 3). Among the benign tumors in the liver, positive expression of CBX8 was found in 16.7% of PEComa, 6.7% of hepatic adenoma and 12.5% of focal nodular hyperplasia (Supplementary Figure 3A). Further study indicated that CBX8 expression was significantly increased in cholangiocarcinoma, lung cancer, colorectal cancer and breast cancer, but was decreased in renal cell carcinoma (Supplementary Figure 3B). Kaplan-Meier analysis revealed that CBX8 expression was correlated with poor post-surgical survival in patients in both our cohort and TCGA’s cohort (Figure 1E), which was validated by stratified survival analyses (Supplementary Figure 4).

CBX8 exerts oncogenic activity towards HCC cells Next, we performed in vitro functional assays to assess the role of CBX8 in HCC progression. CBX8 was either overexpressed in Bel-7402 and SMMC-7721 cells or knocked down in MHCC-97H cells (Supplementary Figure 6A). Cell viability was increased in CBX8-transfected cells but decreased in CBX8-silenced cells (Supplementary Figure 6B). Ectopic CBX8 expression significantly enhanced the ability to form foci. In contrast, the CBX8-depleted cells failed to form colonies (Figure 2A). The 5-ethynyl-2’-deoxyuridine (EdU)-positive cells were noticeably induced by CBX8 overexpression, but were reduced by CBX8 siRNAs (Figure 2B). Subcutaneous xenograft models were applied to investigate the effect of CBX8 on tumorigenesis in vivo. Tumors were found in 5/7 and 3/7 of mice injected with CBX8-expressing and empty vector- transfected Bel-7402 cells, respectively. The tumor volumes and sizes were significantly larger in the CBX8-overexpression groups (Figure 2C). Conversely, tumors formed by CBX8-silenced MHCC-97H cells were dramatically smaller than those in the control group (Figure 2C). Furthermore, the tumors were much heavier in the CBX8- overexpression groups but much lighter in the CBX8-depletion groups, compared to the corresponding control groups (Supplementary Figure 7).

We next questioned whether CBX8 could influence the metastatic ability of HCC cells. Cell shapes were identified by phallotoxin staining to visualize cytoskeleton F-actin. The formation of cell pseudopodium was enhanced by CBX8 overexpression but impaired by CBX8 silencing (Supplementary Figure 6C). The transwell assays showed that cell migration was drastically increased by CBX8 introduction, but decreased by CBX8 knockdown (Figure 2D). Wound-healing assays demonstrated that CBX8-expressing cells filled the gap faster than control cells. By contrast, CBX8 depletion hindered cell movement (Supplementary Figure 6D). A caudal vein injection model was used to evaluate the effect of CBX8 on tumor metastasis in vivo. Lung metastasis was more often detected in the CBX8-overexpression groups, but hardly found in the CBX8- 8

CBX8 augments HCC proliferative and metastatic potential via AKT/β-catenin pathway To unveil the underlying mechanisms of CBX8-mediated oncogenic features, we conducted transcriptional profiling by RNA-seq. Thirty-six genes were jointly altered by forced CBX8 expression in Bel-7402 and SMMC-7721 cells (Supplementary Figure 9). Pathway enrichment analyses showed that both β-catenin and MAPK pathways were significantly activated by CBX8 overexpression (Supplementary Figure 10). Interestingly, CBX8 overexpression caused the nuclear localization of β-catenin, supported by immunofluorescence staining (Figure 3A). The phosphorylations of β- catenin at Ser552 and AKT were increased in cells transfected with CBX8, but blocked in cells with CBX8 depletion (Figure 3B), suggesting a potential role of AKT in β-catenin activation. In clinical samples, nuclear β-catenin was significantly correlated with CBX8 expression (Figure 3C).

The profound impact of CBX8 on activation of β-catenin signaling prompted us to examine the effect of β-catenin on CBX8-mediated malignant activities. The β-catenin inhibitor (XAV-939) and siRNA were introduced into cells with CBX8 overexpression. CBX8-promoted hepatosphere formation was largely abolished by treatment with XAV-939 or β-catenin siRNA (Figure 3D). Similarly, cell motility was weakened in cells cultured with the inhibitor and siRNA (Figure 3E). Combinative suppressions of β- catenin and MAPK pathways results in complete inhibition of CBX8-mediated phenotypes (Supplementary Figure 11). These data indicate that CBX8 exerts its pro- tumor functions via activation of the AKT/β-catenin pathway.

CBX8 triggers AKT/β-catenin pathway via EGR1 Since EGR1 is interacted with β-catenin and was up-regulated by CBX8 overexpression, we next tested whether EGR1 was involved in the CBX8-mediated activation of β- catenin. The IP experiments showed that CBX8 overexpression in HCC cells strengthened the interaction between EGR1 and β-catenin (Figure 4A). Interestingly, EGR1 mRNA was positively correlated with CBX8 mRNA and protein in 56 and 27, respectively, pairs of HCC tissues (Figure 4B&C). As determined by IHC staining, high EGR1 expression was frequently found in the HCC cases with positive CBX8 expression (Supplementary Figure 12). Upon the CBX8 transfection, EGR1 mRNA expression was induced in a dose-dependent manner. In contrast, EGR1 mRNA was down-regulated by 9

In addition, endogenous EGR1 and CBX8 were detectable in the precipitate mediated by the antibody of EGR1 or CBX8 in both Bel-7402 and SMMC-7721 cells (Figure 5A). Using anti-Flag or anti-HA antibody, we found that HA-CBX8 co-precipitated with Flag- EGR1 in HEK-293T cells (Figure 5B). The co-localization of CBX8 and EGR1 in the nucleus of HCC cells was observed by confocal immunofluorescence (Figure 5C). We next intended to explore the consequences of this interaction. Using cycloheximide (CHX, 20 µg/ml), the half-life of the EGR1 protein was determined. In MHCC-97H cells, EGR1 was degraded within 30 min. Following CBX8 knockdown, the degradation of EGR1 protein was significantly enhanced (Figure 5D). In Bel-7402 and SMMC-7721 cells that express less CBX8, the half-life of EGR1 was significantly shorter than that in MHCC-97H cells with high CBX8 expression. The introduction of CBX8 resulted in a prolonged half-life of EGR1 protein (Figure 5E). Collectively, these data suggest that CBX8 promotes cell proliferation and migration via modulation of EGR1 expression.

We reasoned that if EGR1 is the downstream effector of CBX8, then the inhibition of EGR1 should rescue CBX8-promoted malignant phenotypes. EGR1 was silenced by its siRNAs in Bel-7402 and SMMC-7721 cells (Supplementary Figure 14). As expected, CBX8- induced enhancement of cell viabilities was abolished by ERG1 siRNAs (Supplementary Figure 15A). The percentage of EdU-positive cells with CBX8 overexpression and EGR1 knockdown decreased to the control levels (Supplementary Figure 15B). Furthermore, cells transfected with CBX8 and EGR1 less efficiently formed colonies, compared to those with CBX8 (Supplementary Figure 15C). The CBX8-enhanced migration ability was partly attenuated by EGR1 siRNAs (Supplementary Figure 15D). These data suggest a requirement of EGR1 in CBX8-mediated malignancy.

CBX8 triggers the AKT/β-catenin pathway via miR-365a-3p-targeting ZNRF1 microRNA microarray was used to test whether CBX8 could affect the expression of microRNAs in HCC cells. A total of 34 microRNAs were deregulated in response to CBX8 10

The effect of miR-365a-3p on CBX8-mediated phenotypes was next examined. Treatment of miR-365a-3p inhibitor in CBX8-expressing Bel-7402 cells partly attenuated cell growth and migration. In contrast, miR-365a-3p introduction into cells with CBX8 depletion partly regained the clonogenicity and cell movement (Figure 6G&H). It was noticed that combined suppressions or combined overexpressions of EGR1 and miR-365a-3p slightly rescued the CBX8-mediated phenotypes (Figure 6G&H). Following the combined treatments, the phosphorylation and the nuclear translocation of β- catenin were decreased in CBX8-expressing cells, but increased in CBX8-silencing cells, compared with the single modulation of EGR1 or miR-365a-3p expression (Figure 6I&J). These data may suggest an additive effect of EGR1 and miR-365a-3p on CBX8-mediated phenotypes.

To dissect the molecular mechanism of miR-365a-3p, we performed in silico prediction using bioinformatics tools (Targetscan, miRanda, PITA, and picTar). The Venn diagram revealed that 18 genes, including ZNRF1 that has been demonstrated in AKT/β-catenin pathway were predicted by all four bioinformatic algorithms (Supplementary Figure 16A and Supplementary Table 6). One putative binding site at the 3'-UTR of ZNRF1 was found (Supplementary Figure 16B). The luciferase activity of the wild-type reporter containing seed sequence was significantly reduced by miR-365a-3p transfection in Bel-7402 cells, whereas it remained unchanged in mutant reporters (Figure 7A). Ectopic miR-365a-3p expression resulted in a significant decrease of ZNRF1 at both the mRNA and protein levels (Figure 7B&C). In 27 pairs of clinical samples, miR-365a-3p expression was inversely associated with ZNRF1 protein expression (Figure 7D). These data suggest ZNRF1 is a direct target of miR-365a-3p.

Discussion Efforts have been increasingly made to discover the mechanisms of the unlimited growth of HCC cells over the last decades. Several biomarkers have been found and applied to the daily managements of HCC (20, 21). Here, we identified CBX8 as a promising prognostic and therapeutic factor in HCC. We also elucidated that CBX8 manifested its pro-tumor activities through activating the AKT/β-catenin pathway by simultaneously enhancing the expression of EGR1 and miR-365a-3p (Figure 7H).

Misregulation of CBX8 has been reported in human cancers, but its clinical significance is not well understood. Our data provide compelling biological and clinical evidence that CBX8 is overexpressed in HCC. Clear up-regulation of CBX8 was identified in 27.4% of paraffin-embedded samples by IHC and 55.6% of fresh tissues by western blot. Further studies showed that CBX8 expression was significantly increased in cholangiocarcinoma, lung cancer, colorectal cancer and breast cancer, but was decreased in renal cell carcinoma. In line with our study, the increase in CBX8 was also found in glioblastoma and esophageal squamous cell carcinomas in other studies. In our and TCGA’s cohorts, patients with CBX8 expression likely experienced shorter overall and disease-free survival, compared with those without CBX8 expression. A recent report showed that high expression of CBX8 was correlated with poor overall survival in breast cancer (14). On the other hand, high CBX8 expression was connected to favorable prognoses in colorectal cancer, although CBX8 was frequently up-regulated (22). Taken together, these findings suggest misregulation of CBX8 is sensitive in human cancers and CBX8 is a potential biomarker for clinical surveillance of tumor progression.

The typical manner for CBX8 to exert its functions in PRC1 is as a transcriptional repressor of other genes, such as INK4a/ARF (10). Our data demonstrate that CBX8 functions as a transcriptional activator in a PRC1-independent manner to up-regulate the expression of EGR1 and miR-365a-3p in HCC, which was also confirmed in 12

Aberrant activation of β-catenin signaling contributes to the initiation and progression of HCC. Overexpression of β-catenin has been noted in our and other studies, showing that β-catenin is up-regulated in more than half of the HCC cases (25). Accumulating cytosolic β-catenin results in its nuclear localization and transcriptional activity. Another mechanism of activation of β-catenin mediated by AKT has been proposed: phospho-AKT phosphorylates β-catenin at the Ser552 residue to increase nuclear β- catenin (18). Our data showed that β-catenin was transported to the nucleus in HCC cells with overexpression of CBX8. Ectopic CBX8 induced the phosphorylation of AKT and β-catenin. Furthermore, CBX8 protected the protein stability of EGR1 to enhance the interaction of EGR1 and β-catenin, leading to the nuclear accumulation of β- catenin. It should be noticed that PRC1 has been was essential to maintain the leukemia cell identity and intestinal stem cell identity via activation of β-catenin signaling (26, 27). Our data suggest the independency of PRC1 in CBX8-mediated phenotypes in HCC. These data may suggest that CBX8 is not involved in the regulation of PRC1 activity in HCC cells.

The relationship between Polycomb proteins and microRNA has been rarely determined. O'Loghlen et al. showed that CBX7 was down-regulated by miR-125 and miR-181 in embryonic stem cells (28). Zheng and colleagues identified miR-195 as the upstream regulator of CBX4 to suppress HCC (29). However, whether microRNAs can be regulated by Polycomb proteins remains unknown. Here, we suggest that aberrant CBX8 results in altered expression of microRNAs in HCC cells. Especially, CBX8 knockdown reduced the expression of miR-365a-3p, which targets ZNRF1 to degrade AKT. 13

In summary, we identified CBX8 as an oncogene with prognostic significance in HCC. CBX8 promotes HCC progression via a non-canonical mechanism, which triggers the AKT/β-catenin pathway through transcriptionally activating the expression of EGR1 and miR-365a-3p. In light of the development and progress of compounds with the ability to inhibit the function of CBX family proteins (30, 31), our study provides, for further evaluation, a novel therapeutic target for HCC treatment.

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. Coleman MP. Cancer survival: global surveillance will stimulate health policy and improve equity. Lancet. 2014;383:564-73. 3. Allemani C, Weir HK, Carreira H, Harewood R, Spika D, Wang XS, et al. Global surveillance of cancer survival 1995-2009: analysis of individual data for 25,676,887 patients from 279 population-based registries in 67 countries (CONCORD-2). Lancet. 2015;385:977-1010. 4. Reeves HL, Aisen AM. Hepatocellular carcinoma: optimal staging impacts survival. Gastroenterology. 2015;148:1274-6. 5. Mei Y, Yang JP, Qian CN. For robust big data analyses: a collection of 150 important pro-metastatic genes. Chinese journal of cancer. 2017;36:16. 6. Mills AA. Throwing the cancer switch: reciprocal roles of polycomb and trithorax proteins. Nat Rev Cancer. 2010;10:669-82. 7. Tavares L, Dimitrova E, Oxley D, Webster J, Poot R, Demmers J, et al. RYBP-PRC1 complexes mediate H2A ubiquitylation at polycomb target sites independently of PRC2 and H3K27me3. Cell. 2012;148:664-78. 8. Di Croce L, Helin K. Transcriptional regulation by Polycomb group proteins. Nat Struct Mol Biol. 2013;20:1147-55. 9. Peng JC, Valouev A, Liu N, Lin H. Piwi maintains germline stem cells and oogenesis in Drosophila through negative regulation of Polycomb group proteins. Nat Genet. 2016;48:283-91. 10. Dietrich N, Bracken AP, Trinh E, Schjerling CK, Koseki H, Rappsilber J, et al. Bypass of senescence by the polycomb group protein CBX8 through direct binding to the INK4A- ARF locus. EMBO J. 2007;26:1637-48. 11. Zhen CY, Tatavosian R, Huynh TN, Duc HN, Das R, Kokotovic M, et al. Live-cell single-molecule tracking reveals co-recognition of H3K27me3 and DNA targets polycomb Cbx7-PRC1 to chromatin. Elife. 2016;5. 12. Lee SH, Um SJ, Kim EJ. CBX8 suppresses Sirtinol-induced premature senescence in human breast cancer cells via cooperation with SIRT1. Cancer Lett. 2013;335:397-403. 13. Beguelin W, Teater M, Gearhart MD, Calvo Fernandez MT, Goldstein RL, Cardenas MG, et al. EZH2 and BCL6 Cooperate to Assemble CBX8-BCOR Complex to Repress Bivalent Promoters, Mediate Germinal Center Formation and Lymphomagenesis. Cancer Cell. 2016;30:197-213. 14

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2017 American Association for Cancer Research. Author Manuscript Published OnlineFirst on October 24, 2017; DOI: 10.1158/0008-5472.CAN-17-0700 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. 14. Chung CY, Sun Z, Mullokandov G, Bosch A, Qadeer ZA, Cihan E, et al. Cbx8 Acts Non-canonically with Wdr5 to Promote Mammary Tumorigenesis. Cell Rep. 2016;16:472-86. 15. Clevers H, Nusse R. Wnt/beta-catenin signaling and disease. Cell. 2012;149:1192-205. 16. Nishita M, Hashimoto MK, Ogata S, Laurent MN, Ueno N, Shibuya H, et al. Interaction between Wnt and TGF-beta signalling pathways during formation of Spemann's organizer. Nature. 2000;403:781-5. 17. Monga SP. beta-Catenin Signaling and Roles in Liver Homeostasis, Injury, and Tumorigenesis. Gastroenterology. 2015;148:1294-310. 18. Fang D, Hawke D, Zheng Y, Xia Y, Meisenhelder J, Nika H, et al. Phosphorylation of beta-catenin by AKT promotes beta-catenin transcriptional activity. J Biol Chem. 2007;282:11221-9. 19. Cross DA, Alessi DR, Cohen P, Andjelkovich M, Hemmings BA. Inhibition of glycogen synthase kinase-3 by insulin mediated by protein kinase B. Nature. 1995;378:785-9. 20. Chai S, Ng KY, Tong M, Lau EY, Lee TK, Chan KW, et al. Octamer 4/microRNA-1246 signaling axis drives Wnt/beta-catenin activation in liver cancer stem cells. Hepatology. 2016;64:2062-76. 21. Song Y, Pan G, Chen L, Ma S, Zeng T, Man Chan TH, et al. Loss of ATOH8 Increases Stem Cell Features of Hepatocellular Carcinoma Cells. Gastroenterology. 2015;149:1068-81 e5. 22. Tang J, Wang G, Zhang M, Li FY, Sang Y, Wang B, et al. Paradoxical role of CBX8 in proliferation and metastasis of colorectal cancer. Oncotarget. 2014;5:10778-90. 23. Tan J, Jones M, Koseki H, Nakayama M, Muntean AG, Maillard I, et al. CBX8, a polycomb group protein, is essential for MLL-AF9-induced leukemogenesis. Cancer Cell. 2011;20:563-75. 24. Creppe C, Palau A, Malinverni R, Valero V, Buschbeck M. A Cbx8-containing polycomb complex facilitates the transition to gene activation during ES cell differentiation. PLoS Genet. 2014;10:e1004851. 25. Shibata T, Aburatani H. Exploration of liver cancer genomes. Nat Rev Gastroenterol Hepatol. 2014;11:340-9. 26. Rossi A, Ferrari KJ, Piunti A, Jammula S, Chiacchiera F, Mazzarella L, et al. Maintenance of leukemic cell identity by the activity of the Polycomb complex PRC1 in mice. Science advances. 2016;2:e1600972. 27. Chiacchiera F, Rossi A, Jammula S, Piunti A, Scelfo A, Ordonez-Moran P, et al. Polycomb Complex PRC1 Preserves Intestinal Stem Cell Identity by Sustaining Wnt/beta- Catenin Transcriptional Activity. Cell Stem Cell. 2016;18:91-103. 28. O'Loghlen A, Munoz-Cabello AM, Gaspar-Maia A, Wu HA, Banito A, Kunowska N, et al. MicroRNA regulation of Cbx7 mediates a switch of Polycomb orthologs during ESC differentiation. Cell Stem Cell. 2012;10:33-46. 29. Zheng C, Li J, Wang Q, Liu W, Zhou J, Liu R, et al. MicroRNA-195 functions as a tumor suppressor by inhibiting CBX4 in hepatocellular carcinoma. Oncol Rep. 2015;33:1115-22. 30. Ren C, Morohashi K, Plotnikov AN, Jakoncic J, Smith SG, Li J, et al. Small-molecule modulators of methyl-lysine binding for the CBX7 chromodomain. Chem Biol. 2015;22:161-8. 31. Stuckey JI, Dickson BM, Cheng N, Liu Y, Norris JL, Cholensky SH, et al. A cellular chemical probe targeting the chromodomains of Polycomb repressive complex 1. Nat Chem Biol. 2016;12:180-7.

Figure legends Figure 1. CBX8 expression is increased in HCC and associated with poor outcome. A. The expression of CBX8 in HCC cell lines was examined. The immortalized liver cell lines (L-02 and MiHA) were used as controls. B. The mRNA level of CBX8 in 56 paired fresh HCC specimens was evaluated by qRT-PCR. U6 was used as the loading control. C. The protein expression of CBX8 in another 27 HCC samples was determined by western analysis. The relative protein densities were measured and are shown. D. CBX8 expression in 514 HCC paraffin-embedded tissues was examined by IHC. Images 15

Figure 2. CBX8 promotes tumor growth and metastasis in HCC. A. Cells with CBX8 overexpression or depletion were cultured in 6-well plates for 10 days. The number of colonies formed by HCC cells was calculated and indicated by the histogram. Data are the mean ± SEM of three independent experiments. B. Cell proliferation was measured by EdU assays. C. Stable cells were injected into the right flank of the null mice. Tumor volumes were measured every 4 days. The tumors were dissected at day 27 and weighted. Growth curves were summarized at the bottom. *P<0.05, **P<0.01. C. Transwell assays were used to determine the effect of CBX8 on cell migration. The migrated cells were stained with 0.1% crystal violet and counted under a microscope. The fold changes of cell migration were calculated and are shown in the histogram. D. Cells with CBX8 overexpression or silencing were injected into the mice through the tail vein. The metastatic nodules in the lungs were sectioned and counted. Representative micrographs of HE staining and the number of lung metastasis were shown. *P<0.05, **P<0.01.

Figure 3. CBX8 triggers the AKT/β-catenin pathway in HCC cells. A. IF staining was performed to indicate the cellular localization of β-catenin. DAPI was used to stain the nucleus. B. Proteins obtained from cells with or without CBX8 were subjected to western analysis to examine the activation of the AKT/β-catenin pathway. C. Representative IHC images for CBX8 and phosphorylated β-catenin are shown. Their correlation in 514 patients with HCC is presented below. D. Cells with CBX8 overexpression were treated with either β-catenin siRNA or its inhibitor XAV-939. Colony formation was performed to evaluate the effect of inhibition of β-catenin on CBX8-promoted cell growth. E. The impact of β-catenin suppression on cell migration was determined in cells treated as described in E. The fold change of migrated cells is indicated by histograms. *P<0.05.

Figure 4. CBX8 exhibits oncogenic effects via EGR1. A. Cells were transfected with CBX8 overexpression vector and ERG1 siRNA for 48 h. A co-IP assay, using EGR1 antibody, was used to detect the binding of EGR1 and β-catenin. B. The relationship

Figure 5. CBX8 interacts with EGR1 to maintain its protein stability. A. Proteins extracted from HCC cells were incubated with antibody for EGR1 or CBX8. After pelleting by Protein A/G-agarose, samples were subjected to western analysis to investigate the interaction of CBX8 and EGR1. B. HEK293T cells were transfected with Flag-EGR1 and HA-CBX8 overexpression vectors for 24 h. Co-IP assays were performed, using an antibody for HA or Flag. C. Bel-7402 and SMMC-7721 cells were fixed by 4% PFA and incubated with antibodies overnight at 4 °C. After staining by fluorescent secondary antibodies and DAPI, cells were observed under a confocal fluorescence microscope. D. MHCC-97H cells with CBX8 siRNAs were treated with CHX for the indicated periods. The expression of EGR1 was examined by western analysis. The relative EGR1 protein intensities were calculated and are shown by the curve. E. The half-life of EGR1 protein was measured in Bel7402 and SMMC-7721 cells with CBX8 overexpression.

Figure 6. CBX8 transcriptionally up-regulates miR-365a-3p in HCC cells. A. MicroRNA microarrays were used to examine the altered expression of microRNAs by CBX8 siRNAs in MHCC-97H cells. A total of 34 microRNAs were found to be deregulated. B. The decrease of miR-365a-3p induced by CBX8 siRNAs was confirmed in MHCC-97H cells. C. Bel-7402 and SMMC-7721 cells were transfected with CBX8 overexpression vector for 48 h. The expression of miR-365a-3p was determined by qRT-PCR. D. The correlation between CBX8 mRNA and miR-365a-3p was determined in 36 fresh HCC samples. r: Pearson correlation coefficient. E. Luciferase assays were performed to measure the effect of CBX8 overexpression or knockdown on the activity of the miR-365a-3p promoter in Bel-7402 cells. All data are means ± SEM of three independent experiments. *P<0.05, **P<0.01. F. Cells with CBX8 overexpression were treated with Ring1b siRNA or RING inhibitor PRT4165 for 24 h. The mRNA expression of miR-365a-3p was determined by qRT-PCR. G. Colony formation assays were performed. Cells with

Figure 7. miR-365a-3p targets ZNRF1 to activate the AKT/β-catenin pathway in HCC cells. A. The effect of miR-365a-3p on the promoter activity of ZNRF1 was determined in Bel-7402 cells. B. The overexpression of miR-365a-3p in HCC cell lines was confirmed by qRT-PCR. C. Cells were transfected with miR-365a-3p mimics for 48 h. The mRNA and protein expression levels of ZNRF1 were examined. D. The correlation of miR-365a-3p and ZNRF1 mRNA was tested in 27 HCC samples. r: Pearson correlation coefficient. E. The cellular localization of β-catenin was indicated by IF staining in cells incubated with miR-365a-3p mimics for 48 h. F. The activation of the AKT/β- catenin pathway was confirmed by western analysis in cells with miR-365a-3p overexpression or ZNRF1 depletion. G. Representative IHC images for ZNRF1 and phosphorylated β-catenin in HCC tissues with low or high expression of miR-365a-3p are shown. The correlation of ZNRF1 and phosphorylated β-catenin in 514 patients with HCC is presented below. H. Schematic diagram of the CBX8/EGR1 or the CBX8/ miR-365a-3p/ZNRF1 signaling axis in driving AKT/β-catenin activation in HCC cells.

CBX8 exhibits oncogenic activity via AKT/β-Catenin activation in hepatocellular carcinoma

Chris Zhiyi Zhang, Shi-Lu Chen, Chun-Hua Wang, et al.

Cancer Res Published OnlineFirst October 24, 2017.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-17-0700

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2017/10/24/0008-5472.CAN-17-0700.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/2017/10/24/0008-5472.CAN-17-0700. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.