Full Terms & Conditions of access and use can be found athttps://www.tandfonline.com/action/journalInformation?journalCode=kcbt20

Cancer Biology & Therapy

ISSN: 1538-4047 (Print) 1555-8576 (Online) Journal homepage: https://www.tandfonline.com/loi/kcbt20

GSK343 induces programmed cell death throughthe inhibition of EZH2 and FBP1 in osteosarcomacells

Xifeng Xiong, Jinli Zhang, Aiguo Li, Libing Dai, Shengnan Qin, PengzhenWang, Wei Liu, Zhi Zhang, Xiaojian Li & Zhihe Liu

To cite this article: Xifeng Xiong, Jinli Zhang, Aiguo Li, Libing Dai, Shengnan Qin, PengzhenWang, Wei Liu, Zhi Zhang, Xiaojian Li & Zhihe Liu (2019): GSK343 induces programmed cell deaththrough the inhibition of EZH2 and FBP1 in osteosarcoma cells, Cancer Biology & Therapy, DOI:10.1080/15384047.2019.1680061

To link to this article: https://doi.org/10.1080/15384047.2019.1680061

Published online: 25 Oct 2019.

Submit your article to this journal

Article views: 4

View related articles

View Crossmark data

RESEARCH PAPER

GSK343 induces programmed cell death through the inhibition of EZH2 and FBP1 inosteosarcoma cellsXifeng Xionga, Jinli Zhanga, Aiguo Lia,b, Libing Daia, Shengnan Qina, Pengzhen Wanga, Wei Liuc, Zhi Zhangd,Xiaojian Lid, and Zhihe Liua

aGuangzhou Institute of Traumatic Surgery, Guangzhou Red Cross Hospital, Medical Collage, Jinan University, Guangzhou, China; bDepartment ofOrthopaedics, Guangzhou Red Cross Hospital, Medical Collage, Jinan University, Guangzhou, China; cDepartment of Breast Surgery, Guangzhou RedCross Hospital, Medical Collage, Jinan University, Guangzhou, China; dDepartment of Burns and Plastic Surgery, Guangzhou Red Cross Hospital,Medical Collage, Jinnan University, Guangzhou, Guangdong, China

ABSTRACTEnhancer of zeste homolog 2 (EZH2) is an important member of the epigenetic regulatory factor polycombgroup proteins (PcG) and is abnormally expressed in a wide variety of tumors, including osteosarcoma.Scientists consider EZH2 as an attractive target for the treatment of osteosarcoma and have found manypotential EZH inhibitors, such as GlaxoSmithKline 343 (GSK343). It has been reported that GSK343 can beused as an inhibitor in different types of cancer. This study demonstrated that GSK343 not only inducedapoptosis by increasing cleaved Casp-3 and poly ADP-ribose polymerase (PARP) expression, but alsoinduced autophagic cell death by inhibiting p62 expression. Apoptosis and autophagic cell death inducedby GSK343 were confirmed by the high expression of cleaved caspase-3, LC3-II and transmission electronmicroscopy. GSK343 inhibited the expression of EZH2 and c-Myc. Additionally, GSK343 inhibited theexpression of FUSE binding protein 1 (FBP1), which was identified by its regulatory effects on c-Mycexpression. Since c-Myc is a common target of EZH2 and FBP1, and GSK343 inhibited the expression ofthese proliferation-promoting proteins, a mutual regulatory mechanism between EZH2 and FBP1 wasproposed. The knockdown of EZH2 suppressed the expression of FBP1; similarly, the knockdown of FBP1suppressed the expression of EZH2. These results suggest the mutual regulatory association between EZH2and FBP1. The knockdown of either EZH2 or FBP1 accelerated the sensitivity of osteosarcoma cells toGSK343. Based on these results, this study clarified that GSK343, an EZH2 inhibitor, may have potential foruse in the treatment of osteosarcoma. The underlying mechanisms of the effects of GSK343 are partlymediated by its inhibitory activity against c-Myc and its regulators (EZH2 and FBP1).

ARTICLE HISTORYReceived 25 December 2018Revised 7 July 2019Accepted 6 October 2019

KEYWORDSGSK343; EZH2; FBP1;programmed cell death;osteosarcoma

Introduction

Enhancer of zeste homolog 2 (EZH2) is an important member ofthe epigenetic regulatory factor PcG (polycomb group proteins)and a catalytic subunit of polycomb repressive complex 2(PRC2).1 EZH2 transfers the methyl group of adenosine methio-nine (SAM) toH3K27 andmethylates H3K27.2,3 The Tri-methy-lation ofH3K27 can suppress the expression of target genes, suchas RUNX3 and BRCA1.4,5 It has been demonstrated that EZH2plays an important role in epigenetic regulation.6–8 EZH2 isabnormally expressed in a wide variety of tumors, and its expres-sion is unusually high in osteosarcoma.9–11 Osteosarcoma is themost common malignant tumor of mesenchymal tissues, oftenoccuring in the long bone metaphysis of adolescents, and is adifferentiation-defective disease caused by blocked osteoblastdifferentiation and/or epigenetic alterations.12 Scientists considerEZH2 as an attractive target for the treatment of osteosarcomaand have found many potential EZH inhibitors, such as 3-dea-zaneplanocin A (DZNeP), GlaxoSmithKline (GSK)126, EI1 andGSK343.13–16 The chemical name of GSK343 is N-[(6-methyl-2-oxo-4-propyl-1H-pyridin-3-yl) methyl]-6-[2-(4-methylpipera-zin-1-yl) pyridin-4-yl]-1-propan-2-ylindazole-4-carboxamide

(C31H39N7O2, MW: 541.69). GSK343 is a methionine compe-titive histone lysinemethyl-transferase inhibitor that directly andselectively inhibits the activity or expression of EZH2.16,17 It hasbeen reported that GSK343 suppresses glioblastoma, ovariancancer, liver cancer and colorectal cancer.18–21 By performingin silico and experimental analyses, Liu et al found that GSK343inhibited hepatocellular carcinoma through the induction ofmetallothionein 1 and 2A.22 In our previous study, we foundthat GSK343 not only inhibited EZH2 expression, but also inhib-ited the expression of FUSE binding protein 1(FBP1).23

FBP1 activates the transcription of the c-Myc gene by bindingto the far-upstream element (FUSE).24–26 FBP1 is overexpressedin 80% of HCC and ovarian cancer cases,27–29 physically interactswith p53 and suppresses its transcriptional activity.30,31 FBP1expression in tumor cells is linked to poor patient survival.27,32

Recently, we demonstrated that FBP1 physically interacted withEZH2 in osteosarcoma cells and that GSK343 treatment inhibitedthe expression of EZH2 and FBP1.23

In order to enhance our understanding of GSK343 as amolecule for the clinical treatment of osteosarcoma, this studyexamined the effect of GSK343 on the programmed cell death ofhuman osteosarcoma Saos2 cells, and aimed to elucidate its

CONTACT Dr Zhihe Liu zliu0731@163.com Guangzhou Institute of Traumatic Surgery, Guangzhou Red Cross Hospital, Medical College, Jinan University, 396Tongfu Zhong Road, Guangzhou, Guangdong 510220, P.R. China

CANCER BIOLOGY & THERAPYhttps://doi.org/10.1080/15384047.2019.1680061

© 2019 Taylor & Francis Group, LLC

potential underlying mechanisms. It is well known that c-Myc isthe downstream target of FBP1, while EZH2 and c-Myc are apair of mutually activating proteins.33,34 Therefore, this studyalso aimed to clarify whether FBP1 and EZH2 share mutualregulatory roles, and whether they play an important role inGSK343-induced programmed cell death.

Results

GSK343, GSK126 and DZNeP induce apoptosis in Saos2and U2OS cells

For apoptosis (type I programmed cell death) assay, AnnexinV-FITC/PI double staining and flow cytometry were used.The apoptotic rates were detected after the Saos2 and U2OScells were treated for 48 h with 0, 5, 10 and 20 μM of GSK343;0, 1, 2 and 4 μM DZNeP and 0, 5, 10 and 15 μM GSK126. ForGSK343, the apoptotic rates of the Saos2 and U2OS cellstreated with 10 and 20 μM were significantly higher thanthose of the controls (0 μM) group (P < .05). The highestapoptotic rate was observed in the 20 μM group (Figure 1(a),panels a and b, and 1B, panel a). For DZNeP, significantlyhigher apoptotic rates were only observed in the U2OS cellstreated with 4 μM. No significantly higher apoptotic rateswere found in the Saos2 cells (Figure 1(a), panels c and d,and B, panel b). For GSK126, the apoptotic rates of the Saos2and U2OS cells treated with 15 μM GSK126were significantlyhigher than those of the control (0 μM) group (P < .05)(Figure 1(a), panels e and f, and B, panel c).

Based on the aforementioned results, further experimentswere carried out using GSK343 and Saos2 cells. As apoptoticmarkers, the expressions of cleaved caspase-3 and ADP-ribosepolymerase (PARP) were detected by western blot analysis.With increasing concentrations of GSK343, the expressions ofcleaved capase-3 and PARP were markedly increased com-pared with the control group (Figure 1(c)).

GSK343 induces autophagic cell death in Saos2 cells

Autophagy refers to a process through which some proteinsand organelles are transported to lysosomes for degradationthrough the inclusion of autophagosomes.35 Autophagic celldeath, also known as type II programmed cell death, is con-sidered a possible tumor suppressor mechanism.36 In order toinvestigate whether GSK343 induces autophagy, we analyzedthe GSK343-induced activation and translocation of LC3-II tothe autophagosome. Western blot analysis revealed the accu-mulation of LC3-II in the autophagosome of Saos2 cells withincreasing concentrations GSK343. The majority of LC3-IIpunctate was found in the cells treated with 20 μM GSK343(Figure 2(a), panels a-d). The formation of autophagic vacuoleswas also detected by transmission electron microscopy to eval-uate the level of autophagic cell death. Compared with thecontrol group, the group treated with 5 μM GSK 343 exhibitedonly a mild expansion of the partial mitochondrial intracellularcleft, a mild expansion of the partial rough endoplasmic reti-culum, and occasional autophagosomes; the group treated with10 μM exhibited multiple autophagosomes in the cytoplasm,slight mitochondrial swelling and partial cavitation; the group

treated with 20 μM exhibited a large number of autophago-somes, mitochondrial swelling, internal dilatation and partialcavitation (Figure 2(a), panels e-h). Autophagosome formationincreased gradually with increasing GSK343 concentrations.The greatest degree of autophagosome formation was observedfollowing treatment with 20 μM GSK343.

Similar to LC3, p62 is recognized as amarker of autophagic celldeath. In order to evaluate the level of autophagic cell death,wemeasured the ratio of LC3-II/LC3-I and the expression of p62 inSaos2 cells treatedwith various concentrations ofGSK343 for 48 h.With increasing concentrations of GSK343,the protein expressionlevel of p62 gradually decreased (Figure 2(b,c)),while LC3-II pro-tein and the LC3-II/I ratio gradually increased (Figure 2(b,d)). Thereduction in p62 expression and the induction of LC3-II were bothGSK343 dose-dependent. A significant reduction in p62 expres-sion and the induction of LC3-II were observed following treat-ment with 5 μMGSK343.

GSK343 inhibits the expression of c-Myc, EZH2 and FBP1in Saos2 cells

The expression of EZH2 was detected by western blot analysis.As shown in Figure 3(a,b), the expression level of EZH2 wasinhibited by GSK 343 in the Saos2 cells, and this inhibitoryeffect occurred in a dose-dependent manner. As the target ofEZH2 and an oncogene, the expression level of c-Myc in theSaos2 cells was also reduced by GSK343 in a dose-dependent(Figure 3(a,d)).

Since c-Myc is up-regulated by FBP1, the expression level ofFBP1 was examined. Similar to c-Myc and EZH2, the expressionlevel of FBP1 was also inhibited by GSK343 in a dose-dependentmanner (Figure 3(a,c)).

The expressions of EZH2, FBP1 and c-Myc were signifi-cantly inhibited by 10 and 20 μM GSK343. Since c-Myc is acommon target of EZH2 and FBP1, and GSK343 inhibited theexpression of these proliferation-promoting proteins, this sug-gests a mutual regulatory association between EZH2 and FBP1.

The mutual regulatory activity of EZH2 and FBP1

Since c-Myc is the downstream target of EZH2 andFBP1,23–26,33,34 and a physical interaction has been identi-fied between EZH2 and FBP1,23 it would be of interest toclarify whether there is a mutual regulatory effect betweenFBP1 and EZH2. In this study, either EZH2 or FBP1 wassilenced, and the expressions of EZH2 or FBP1 were inves-tigated by RT-qPCR and western blot analysis in normalSaos2 cells and in Saos2 cells in which FBP1/EZH2 wasknocked down.

The knockdown of EZH2 and FBP1 was found to besuccessful by comparing lane 3 (Figure 4(a,b,d)) or lane 2(Figure 4(a,c,e)) with lane 1 (Figure 4(a–e)). By comparinglane 2 with lane 1 of Figure 4(a,b,d), it is evident that theprotein and mRNA expression of EZH2 were significantlyinhibited by FBP1 knockdown. The knockdown of FBP1 sig-nificantly accelerated the inhibitory effects of GSK343 onEZH2 at both the protein and mRNA expressions, as shownby comparing lanes 5 and 2 of Figure 4(a,b,d). The knock-down of EZH2 promoted the inhibitory effects of GSK343 on

2 X. XIONG ET AL.

Figure 1. Effects of GSK343 on the apoptosis of Saos2 cells.(a and b) Apoptotic Saos2 and U2OS cells were detected by Annexin V/PI assay following treatment with GSK343, DZNeP and GSK126 for 48 h at the indicated concentrations.Apoptotic rates are presented in (b).(c) The expression of apoptosis-associated proteins in Saos2 cells. Saos 2 cells were treated for 48 h with the indicated concentrations ofGSK343. Data are presented as themeans ± SD (n = 3). *P < .05 comparedwith the control (0 μMof GSK343). Data shown are representative of a minimum of three independentexperiments.

CANCER BIOLOGY & THERAPY 3

EZH2 expressions both at the protein and mRNA level (lanes6 and 3 of Figure 4(a,b,d)).

Similarly, the protein and mRNA expressions of FBP1 weresignificantly inhibited by EZH2 knockdown (compare lanes 1and 3 of Figure 4(a,c,e)). The knockdown of EZH2 signifi-cantly accelerated the inhibitory effects of GSK343 on FBP1expression both at the protein and mRNA level, as evidencedby comparing lanes 6 and 3 of Figure 4(a,c,e). The knockdownof FBP1 promoted the inhibitory effects of GSK343 on theFBP1 expression both at protein and mRNA level (lanes 6 and3 of Figure 4(a,c,e)).

These results indicated a mutual regulatory mechanismbetween FBP1 and EZH2 expression; the knockdown of FBP1and EZH2 enhanced the sensitivity of Saos2 cells to GSK343.

Discussion

The occurrence of a tumor is a multi-factorial, multi-stage evolu-tionary process, involving mutations and epigenetic alterations ina number of genes, including oncogenes, tumor suppressor andDNAdamage repair genes.37 EZH2 is an important component of

Figure 1. (Continued)

4 X. XIONG ET AL.

Figure 2. GSK343 induces autophagic cell death in Saos 2 cells.(a, panels a-d) LC3-II puncta formation in Saos2 cells treated with the indicated concentrations of GSK343 for 48 h, and examined by immunohistochemistry. Images wereco-stained with DAPI and observed immediately using a Zeiss confocal microscope (magnification, x400). The arrow indicates LC-II puncta.(a, panels e-h) Autophagicvacuole formation in Saos2 cells treated with the indicated concentrations GSK343 for 48 h, captured by electronic microscopy (magnification, x15,000; scale bar,500 nm).(b) The expressions of p62, LC3-I and -II in Saos2 cells treated with the indicated concentrations of GSK343 for 48 h, detected by western blot analysis. GAPDHwas used as the loading control.(c) Relative expression of p62 against GAPDH at the indicated concentrations of GSK343. (d) LC-II/LC-I ratios in Saos2 cells treated withthe indicated concentrations of GSK343 for 48 h. *P < .05 and #P < .01 compared with the control (0 μM of GSK343). Data shown are representative of a minimum of threeindependent experiments.

Figure 3. Effects of GSK343 on associated protein expression in Saos2 cells.(a) The expression of EZH2, FBP1 and c-Myc in Saos2 cells treated with the indicated concentrations of GSK343 and detected by western blot analysis. GAPDH wasused as a loading control.(b-d) The relative expressions of p62, FBP1 and c-Myc at the indicated concentrations of GSK343. #P < .01 compared with the control (0 μMof GSK343). Data shown are representative of a minimum of three independent experiments.

CANCER BIOLOGY & THERAPY 5

the epigenetic regulatory factor, PcG, and a catalytic subunit ofPRC2. It has been reported that EZH2 is abnormally and highlyexpressed in various tumor tissues,9–11 and in recent years, severalpotential inhibitors of EZH2 have been discovered.13–16 GSK343is a direct competitive inhibitor of methionine, which can inhibitthe expression of EZH2 in tumor cells to achieve an anti-tumoreffect.18–21 Currently, relatively few studies have been conductedon the efficacy of GSK343 in solid tumors, such as ovarian cancer,liver cancer and colorectal cancer.19–21

In this study, we found that GSK343 induced apoptosis,also known as type I programmed cell death, in Saos2 cells ina dose-dependent manner, and that treatment with 10 and 20μM GSK343 significantly induced Saos2 cell apoptosis. Theexpressions of apoptosis-related proteins, such as cleaved cas-pase-3 and PARP, were also increased.

Autophagy, also known as type II programmed cell death, isa subcellular degradative process where lysosomes degrade

macromolecular proteins and organelles. Autophagy is impor-tant for maintaining cell homeostasis and for adaptation toadverse internal and external environments.38 Moderate autop-hagy is considered to promote cell tolerance during environ-mental adaptation. However, the persistence of unfavorablefactors can lead to excessive autophagy, which subsequentlyinduces and accelerates autophagic cell death.39 LC3 is com-monly used to detect autophagic cell death and has type I andII forms. When autophagic cell death occurs, LC3-I and phos-phatidyl ethanolamine combine to form LC3-II and localize tothe autophagosome membrane prior to fusion with lysosomes;thus to a certain extent, LC3-II expression can reflect autopha-gic activity.40 p62 protein, an autophagic substrate, is oftenused in conjunction with LC3 to detect autophagy. This studydemonstrated a decrease in the expression of p62 and anincrease in the LC3-II/LC3-I ratio with increasing GSK343concentrations. This suggested that GSK343 promoted

Figure 4. Mutual regulatory association between EZH2 and FBP1 expression.(a) The expression of EZH2 and FBP1 in FBP1 or EZH2 normal or knockdown Saos2 cells treated with 0 or 10 μM GSK343, and detected by western blot analysis.GAPDH was used as a loading control.(b and c) The relative protein expressions of EZH2 and FBP1 in FBP1 or EZH2 normal or knockdown Saos2 cells treated with 0 or10 μM GSK343.(d and e) The relative mRNA expressions of EZH2 and FBP1 in FBP1 or EZH2 normal or knockdown Saos2 cells treated with 0 or 10 μM GSK343.*P < .05 compared with the control (0 μM of GSK343). Data are representative of a minimum of three independent experiments.

6 X. XIONG ET AL.

autophagic cell death, which was confirmed by transmissionelectron microscopy and immunofluorescence; the number ofautophagosomes increased with increasing concentrations ofGSK343, which further verified that GSK343 was able to induceautophagic cell death.

c-Myc is one of themost commonly overexpressed oncogenesin cancers, and is a transcription factor involved in DNA repli-cation and cell proliferation via regulating target gene expres-sion. EZH2 induces c-Myc expression via the inhibition of Myctargeting miR-494.33,34 As feedback, c-Myc contributes to EZH2upregulation via the repression of EZH2 targeting miR-26a.33,34

Thus, EZH2 and c-Myc activate each other via a feedback loopmechanism. On the other hand, FBP1 up-regulates c-Mycexpression by binding to the FUSE domain of the c-Myc opera-tor. Consistent with the results presented in ta study by Ezpondaet al,41 we demonstrated that c-Myc expression was decreasedfollowing GSK343 treatment, in addition to the expressions ofEZH2 and FBP1. These data demonstrated that both EZH2 andFBP1 were suppressed by GSK343. GSK343 may directly orindirectly influence c-Myc expression through EZH2 and/orFBP1. Gene knockdown experiments demonstrated that theknockdown of either EZH2 or FBP1 impaired their expressions.The current study identified that the knockdown of EZH2decreased FBP1 expression; similarly, the knockdown of FBP1also decreased EZH2 expression. Furthermore, the knockdownof EZH2 or FBP1 accelerated the reduction of FBP1 and EZH2expression induced by GSK343 treatment. These results suggestthe possibility of the existence of a mutual regulatory associationbetween EZH2 and FBP1.

In consistent to our published results,23 this study confirmedthat GSK343 induced programmed cell death in human osteo-sarcoma cells in a dose-dependent manner. The reduction inEZH2 and FBP1 expressionsmay, at least in part, be themechan-ism underlying the effects of GSK343. Additionally, this studydemonstrated the mutual positive relationship between EZH2and FBP1 basing on the fact that the knockdown of EZH2 orFBP1 suppressed the expression of FBP1 or EZH2.

The main results are summarized in Figure 5.

Materials and methods

Reagents, cells and cell culture

GSK343, GSK126 and DZNeP were provided by Selleck ChemicalCo. (Cat. nos. S7164, S7061 and S7120). The Saos2 and U2OShuman osteosarcoma cell lines were purchased from the ShanghaiCell Bank of the Chinese Academy of Sciences (Cat. no. TCHu114and TCHu88). Penicillin and streptomycin were purchased fromHyClone (Cat. no. SV30010). The Saos2 cells were cultured inhigh-glucose DMEM (Gibco, Cat. no. C11965500137) containing10% FBS (Gibco, Cat. no. 10099–141) 100 U/ml penicillin and100 U/ml streptomycin at 37°C and 5% CO2.

The apoptosis detection kits were purchased from KeyGenBiotech (Cat. No. G003-1-2). The cell total protein extractionkit and BCA protein quantification kit were purchased fromBeyotime Biotechnology (Cat. no. P00138) and ThermoFisherScientific (Cat. no. P0011), respectively. GAPDH antibody(Cat. no. 5174), EZH2 antibody (Cat. no. 5246.), p62 antibody(Cat. no. 39749), PARP (Cat. no. 9542), LC3 (Cat. no. 3868)

and cleaved caspase-3 (Cat. no. 9664) antibodies were pur-chased from Cell Signaling Technology. FBP1 (Cat. no.ab181111.) and c-Myc (Cat. no. ab320752.) antibodies wereprovided by Abcam.

Flow cytometric assay

An Annexin V-FITC/propidium iodide (PI) double stainingassay was used to measure the level of apoptosis in Saos2 cells.The cells were inoculated on a 6-well plate and treated withvarious concentrations of GSK343 for 48 h. Each group ofcells was stained following the manufacturer’s instructions.The number of apoptotic cells was detected by flow cytometryand analyzed using CellQuest™ software (BD Biosciences).The experiment was repeated 3 times.

Western blot analysis

Following treatment with various concentrations of GSK343for 48 h, the Saos2 cells were harvested. Prior to westernblotting, the cells were lysed in modified RIPA buffer[150 mM NaCl, 1% NP-40, 50 mM Tris-Cl (pH 8.0), 0.1%SDS] supplemented with protease and phosphatase inhibitor,PMSF (1 mM) ThermoFisher Scientific (Cat. no. 36978).Following rapid homogenization, the homogenate was incu-bated in ice for 30 min and centrifuged at 12,000 x g for15 min at 4°C. The protein concentration was measured byBCA assay. Approximately 40 μg of protein sample per wellwas loaded on 12% SDS-PAGE and transferred onto PVDFmembranes (Merck Millipore, Cat. no. ISEQ00010). Themembranes were blocked with 5% nonfat milk in Tris-buf-fered saline with Tween-20 (TBST) for 1 hour at room tem-perature, incubated overnight at 4°C with the designatedprimary antibodies, and incubated with secondary antibodiesfor 1 h at room temperature. An ECL luminescence assay wasused to develop the blots (Pierce, Cat. no. 32209), and relativeabundance was quantified by densitometry using QuantityOne 4.6.7 software (Bio-Rad).

Quantitative analysis of gene expression by reversetranscription-quantitative polymerase chain reaction PCR(RT-qPCR)

Approximately 2 × 105 cells were transferred to a 6-well plate(final total volume of 4 ml), and cultured for 48 h without orwith GSK343. Total RNA was extracted from the cells usingTRIzol reagent (Invitrogen; Thermo Fisher Scientific, Cat. no.15596026) and converted to cDNA using PrimeScript RTMaster Mix (Takara, Cat. no. RR036A). qPCR was performedwith SYBR Premix ExTaq (Takara, Cat. no. RR420A) in areal-time PCR system (Analytik Jena). All samples were per-formed in triplicate. GAPDH was used as an internal control.The primer sequences were as follows: FBP1-forward, 5ʹ-TGATTCCAGCTAGCAAGGCA-3ʹ and reverse, 5ʹ-CGGCCCGTCTTGAATCATAA-3ʹ; EZH2-forward, 5ʹ-ACCAGCATTTGGAGGGAGC-3ʹ and reverse, 5ʹ-TGGGAAGCCGTCCTCTTCT-3ʹ; GAPDH forward, 5ʹ-GATTCCACCCATGGCAAATT-3ʹ and reverse, 5ʹ-TCTCGCTCCTGGAAGATGGT-3ʹ. The relative gene expression was quantified using the

CANCER BIOLOGY & THERAPY 7

2−ΔΔCq method and normalized to the internal reference gene,GAPDH.

Immunofluorescence assay

The Saos2 cells were treated with various concentrations ofGSK343 for 48 h, and then fixed in 4% paraformaldehyde for15 min, washed in PBS, blocked using 10% normal goatserum, and incubated with a rabbit anti-human L3 polyclonalantibody (1:100) overnight at 4°C. Dylight 594-conjugatedanti-rabbit IgGs (Abcam, Cat. no. ab96885) were applied ata 1:200 dilution for 1 h at room temperature. The cells werecounterstained with DAPI and observed immediately using aZIESS confocal microscope (Carl Zeiss)

Transmission electron microscopy

The Saos2 cells were treated with various concentrations ofGSK343 for 48 h. Following treatment with GSK343, theSaos2 cells were fixed with 2% glutaraldehyde at 4°C for15 min. The cells were then dehydrated in a graded ethanol

series and embedded in Agar 100 epoxy resin. Ultra-thinsections were mounted on Cu grids and stained first withuranyl acetate, followed by lead citrate. Transmission electronmicroscopy (FEI) was used to detect autophagic vacuoles.

Construction of FBP1-knockdown lentivirus andgeneration of stable FBP1-knockdown cells

To knock down FBP1, a pSi-LVRH1GP vector with a puromy-cin resistance cassette (GeneCopoeia, Cat. no. LP-HSH021663)was used to express a short hairpin (sh)RNA to inhibit FBP1expression. The control vector expressed a scrambled sequence(5ʹ-GCT TCG CGC CGT AGT CTT A-3ʹ) and was designatedas pSi-LVP-FBP1-C. The FBP1 knockdown vector expressed asequence of 1671–1691 (5ʹ-GCA GGA ACG GAT CCA AATTCA-3ʹ) of FBP1 and named as pSi-LVP-FBP1-KD.

To knockdown EZH2, a psi-LVRU6MH vector with a hygro-mycin resistance cassette (GeneCopoeia, Cat. no. LPP-HSH005050) was used to express shRNA to knockdown EZH2expression. The control vector expressed a scrambled sequenceand was designated as pSi-LVH-EZH2-C. The EZH2 knockdown

Figure 5. Hypothetical model of GSK343-induced programmed cell death in osteosarcoma cells.

8 X. XIONG ET AL.

vector expressed a 695–715 sequence (5ʹ-GCCCTTGGTCAATATAATGAT-3ʹ) of EZH2 and was named pSi-LVH-EZH2-KD.

The Saos2 cells were transfected with pSi-LVP-FBP1-C/pSi-LVH-EZH2-C or pSi-LVP-FBP1-KD/pSi-LVH-EZH2-KD. First, 2 × 105 Saos2 cells were seeded in 2 ml antibio-tic-free DMEM medium supplemented with FBS, and cellswere incubated until they reached 60–80% confluence. Afterwashing the cells once with 2 ml antibiotic- and FBS-freeDMEM medium, 900 μl of the same medium was added tothe cells, alongside 1 μg pSi-LVP-FBP1-C/pSi-LVH-EZH2-Cor pSi-LV-FBP1-KD/pSi-LVH-EZH2-KD diluted in 100 μlmedium. After 6 h, the medium was exchanged for DMEMsupplemented with FBS and antibiotics. Puromycin or hygro-mycin were used for selection. Following a 48-h incubation,the medium was exchanged with DMEM containing 2.0 μg/mlpuromycin (Invitrogen; Thermo Fisher Scientific, Cat. no.A1113803) or 200 μg/ml of hygromycin B (Invitrogen;Thermo Fisher Scientific, Cat. no. 10687010). Western blot-ting was used to confirm the knockdown efficiency. The pSi-LV-FBP1-C/pSi-LVH-EZH2-C-transfected cells were desig-nated as FBP1-C/EZH2-C, and the pSi-LV-FBP1-KD/pSi-LVH-EZH2-KD-transfected cells were designated as FBP1-KD/EZH2-KD.

Statistical analysis

All statistical analyses were performed using SPSS 17.0 statis-tic software (SPSS, Inc. Chicago, IL, USA). The data areexpressed as the means ± SD. Significance differences weredetermined by one-way analysis of variance (ANOVA), andP<0.05 was considered to be statistically significant.

Acknowledgments

The flow cytometry experiment was supported by Guangdong ProvincialKey Laboratory of Malignant Tumor Epigenetics and Gene Regulation,Sun Yat-Sen Memorial Hospital, Sun Yat-Sen University.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Funding

This study was funded by the National Natural Science Foundation ofChina (grant no. 81272222, to ZL), the research grants of GuangdongBureau of Traditional Chinese Medicine (grant no. 20181206 to ZL, grantno. 20191260 to XX), the Medical Science and Technology ResearchFoundation of Guangdong (grant nos. B2016018 and A2018063, toXX), and the Guangzhou Science and Technology Program key projects(grant no. 201704020145, to SQ).

References

1. Margueron R, Reinberg D. The Polycomb complex PRC2 and itsmark in life. Nature. 2011;469:343–349.

2. Cao R, Wang L, Wang H, Xia L, Erdjument-Bromage H, TempstP, Jones RS, Zhang Y. Role of histone H3 lysine 27 methylation inPolycomb-group silencing. Science. 2002;298:1039–1043.

3. Czermin B, Melfi R, McCabe D, Seitz V, Imhof A, Pirrotta V.Drosophila enhancer of Zeste/ESC complexes have a histone H3methyltransferase activity that marks chromosomal Polycombsites. Cell. 2002;111:185–196.

4. Kodach LL, Jacobs RJ, Heijmans J, van Noesel CJ, Langers AM,Verspaget HW, Hommes DW, Offerhaus GJ, van Den Brink GR,Hardwick JC. The role of EZH2 and DNA methylation in thesilencing of the tumour suppressor RUNX3 in colorectal cancer.Carcinogenesis. 2010;31:1567–1575.

5. Rondinelli B, Gogola E, Yücel H, Duarte AA, van de Ven M, vander Sluijs R, Konstantinopoulos PA, Jonkers J, Ceccaldi R,Rottenberg S, et al. EZH2 promotes degradation of stalled replica-tion forks by recruiting MUS81 through histone H3 trimethyla-tion. Nat Cell Biol. 2017;19:1371–1378.

6. Kuzmichev A, Jenuwein T, Tempst P, Reinberg D. DifferentEZH2-containing complexes target methylation of histone H1 ornucleosomal histone H3. Mol Cell. 2004;14:183–193.

7. Simon JA, Lange CA. Roles of the EZH2 histone methyltransfer-ase in cancer epigenetics. Mutat Res. 2008;647:21–29.

8. Tsang DP, Cheng AS. Epigenetic regulation of signaling pathwaysin cancer: role of the histone methyltransferase EZH2. JGastroenterol Hepatol. 2011;26:19–27.

9. Varambally S, Dhanasekaran SM, Zhou M, Barrette TR, Kumar-Sinha C, Sanda MG, Ghosh D, Pienta KJ, Sewalt RG, Otte AP, etal. The polycomb group protein EZH2 is involved in progressionof prostate cancer. Nature. 2002;419:624–629.

10. Kleer CG, Cao Q, Varambally S, Shen R, Ota I, Tomlins SA,Ghosh D, Sewalt RG, Otte AP, Hayes DF, et al. EZH2 is a markerof aggressive breast cancer and promotes neoplastic transforma-tion of breast epithelial cells. Proc Natl Acad Sci USA.2003;100:11606–11611.

11. Lv YF, Yan GN, Meng G, Zhang X, Guo QN. 2015. Enhancer ofzeste homolog 2 silencing inhibits tumor growth and lung metas-tasis in osteosarcoma. Sci Rep. 5:12999. doi: 10.1038/srep12999..

12. Tang N, Song WX, Luo J, Haydon RC, He TC. Osteosarcomadevelopment and stem cell differentiation. Clin Orthop Relat Res.2008;466:2114–2130.

13. Crea F, Fornaro L, Bocci G, Sun L, Farrar WL, Falcone A, DanesiR. EZH2 inhibition: targeting the crossroad of tumor invasion andangiogenesis. Cancer Metastasis Rev. 2012;31:753–761.

14. Qi W, Chan H, Teng L, Li L, Chuai S, Zhang R, Zeng J, Li M, FanH, Lin Y, et al. Selective inhibition of Ezh2 by a small moleculeinhibitor blocks tumor cells proliferation. Proc Natl Acad SciUSA. 2012;109:21360–21365.

15. Verma SK, Tian X, LaFrance LV, Duquenne C, Suarez DP,Newlanger KA, Romeril SP, Burgess JL, Grant SW, Brackley JA,et al. Identification of Potent, Selective, Cell-Active Inhibitors of theHistone Lysine Methyltransferase EZH2. ACS. 2012;3:1091–1096.

16. Ferraro A, Boni T, Pintzas A. EZH2 regulates cofilin activity andcolon cancer cell migration by targeting ITGA2 gene. PloS One.2014;9:e115276.

17. Kim W, Bird GH, Neff T, Guo G, Kerenyi MA, Walensky LD,Orkin SH. Targeted disruption of the EZH2-EED complex inhi-bits EZH2-dependent cancer. Nat Chem Biol. 2013;9:643–650.

18. Yu T, Wang Y, Hu Q, Wu W, Wu Y, Wei W, Han D, You Y, LinN, Liu N. The EZH2 inhibitor GSK343 suppresses cancer stem-like phenotypes and reverses mesenchymal transition in gliomacells. Oncotarget. 2017;8:98348–98359.

19. Amatangelo MD, Garipov A, Li H, Conejo-Garcia JR, SpeicherDW, Zhang R. Three-dimensional culture sensitizes epithelialovarian cancer cells to EZH2 methyltransferase inhibition. CellCycle. 2013;12:2113–2119.

20. Liu TP, Lo HL, Wei LS, Hsiao HH, Yang PM. S- Adenosyl-L-methionine- competitive inhibitors of the histone methyltransfer-ase EZH2 induce autophagy and enhance drug sensitivity incancer cells. Anticancer Drugs. 2015;26:139–147.

21. Hsieh YY, Lo HL, Yang PM. EZH2 inhibitors transcriptionallyupregulate cytotoxic autophagy and cytoprotective unfolded pro-tein response in human colorectal cancer cells. Am J Cancer Res.2016;6:1661–1680.

CANCER BIOLOGY & THERAPY 9

22. Liu TP, Hong YH, Tung KY, Yang PM. In silico and experimentalanalyses predict the therapeutic value of an EZH2 inhibitorGSK343 against hepatocellular carcinoma through the inductionof metallothionein genes. Oncoscience. 2016;3:9–20.

23. Xiong X, Zhang J, Liang W, Cao W, Qin S, Dai L, Ye D, Liu Z.Fuse-binding protein 1 is a target of the EZH2 inhibitor GSK343,in osteosarcoma cells. Int J Oncol. 2016;49:623–628.

24. Avigan MI, Strober B, Levens D. A far upstream element stimu-lates c-myc expression in undifferentiated leukemia cells. J BiolChem. 1990;265:18538–18545.

25. He L, Weber A, Levens D. Nuclear targeting determinants of thefar upstream element binding protein, a c-myc transcription fac-tor. Nucleic Acids Res. 2000;28:4558–4565.

26. Liu J, Kouzine F, Nie Z, Chung HJ, Elisha-Feil Z, Weber A, Zhao K,The LD. FUSE/FBP/FIR/TFIIH system is a molecular machine pro-gramming a pulse of c-myc expression. Embo J. 2006;25:2119–2130.

27. Rabenhorst U, Beinoraviciute-Kellner R, Brezniceanu ML, Joos S,Devens F, Lichter P, Rieker RJ, Trojan J, Chung HJ, Levens DL, etal. Overexpression of the far upstream element binding protein 1in hepatocellular carcinoma is required for tumor growth.Hepatology. 2009;50:1121–1129.

28. Xiong X, Zhang J, Hua X, Cao W, Qin S, Dai L, Liu W, Zhang Z,Li X, Liu Z. FBP1 promotes ovarian cancer development throughthe acceleration of cell cycle transition and metastasis. Oncol Lett.2018;16:1682–1688.

29. Zhang J, Xiong X, Hua X, Cao W, Qin S, Dai L, Liang P, Zhang H,Liu Z. Knockdown of FUSE binding protein 1 enhances thesensitivity of epithelial ovarian cancer cells to carboplatin. OncolLett. 2017;14:5819–5824.

30. Dixit U, Pandey AK, Liu Z, Kumar S, Neiditch MB, Klein KM,Pandey VN. FUSE binding protein 1 facilitates persistent hepatitisC virus replication in hepatoma cells by regulating tumor sup-pressor p53. J Virol. 2015;89:7905–7921.

31. Dixit U, Liu Z, Pandey AK, Kothari R, Pandey VN. 2014. Fusebinding protein antagonizes the transcription activity of tumor

suppressor protein p53. BMC Cancer. 14:925. doi: 10.1186/1471-2407-14-925..

32. Malz M, Weber A, Singer S, Riehmer V, Bissinger M, Riener MO,Longerich T, Soll C, Vogel A, Angel P, et al. Overexpression of farupstream element binding proteins: a mechanism regulating pro-liferation and migration in liver cancer cells. Hepatology.2009;50:1130–1139.

33. Koh CM, Iwata T, Zheng Q, Bethel C, Yegnasubramanian S, DeMarzo AM. Myc enforces overexpression of EZH2 in early pro-static neoplasia via transcriptional and post-transcriptionalmechanisms. Oncotarget. 2011;2:669–683.

34. Zhang X, Zhao X, Fiskus W, Lin J, Lwin T, Rao R, Zhang Y, ChanJC, Fu K, Marquez VE, et al. Coordinated silencing of MYC-mediated miR-29 by HDAC3 and EZH2 as a therapeutic targetof histone modification in aggressive B-Cell lymphomas. CancerCell. 2012;22:506–523.

35. Lozy F, Karantza V. Autophagy and cancer cell metabolism.Semin Cell Dev Biol. 2012;23:395–401.

36. Bursch W. The autophagosomal-lysosomal compartment in pro-grammed cell death. Cell Death Differ. 2001;8:569–581.

37. Jia L, Sun X, Ding H. Research progress on the relationshipbetween EZH2 gene and tumor. J Clin Neurol. 2012;25:154–156.

38. Mizushima N, Levine B, Cuervo AM, Klionsky DJ. Autoaphagyfights disease though cellular self digestion. Nature.2008;451:1069–1075.

39. Eisenberg-Lerner A, Bialik S, Simon HU, Kimchi A. Life anddeath partners: apoptosis,autophagy and the cross-talk betweenthem. Cell Death Differ. 2009;16:966–975.

40. Kimura S, Fujita N, Noda T, Yoshimori T. Monitoring autophagyin mammalian cultured cells through the dynamics of LC3.Methods Enzymol. 2009;452:1–12.

41. Ezponda T, Dupéré-Richer D, Will CM, Small EC, Varghese N,Patel T, Nabet B, Popovic R, Oyer J, Bulic M, et al. UTX/KDM6Aloss enhances the malignant phenotype of multiple myeloma andsensitizes cells to EZH2 inhibition. Cell Rep. 2017;21:628–640.

10 X. XIONG ET AL.