Long-Hui Zhng Dong Wng Zho Li Gng Wng Ding-Bo Chen Qin Cheng bcShi-Hu Hu bcJi-Ye Zhu bc*

a Department of Hepatobiliary Surgery, Peking University People’s Hospital, Beijing 10 0 044, China

b Peking University Institute for Organ Transplantation, Beijing 10 0 044, China

c Beijing Key Laboratory of Liver Cirrhosis and Liver Cancer, Beijing 10 0 044, China

d Department of Pathology, Peking University People’s Hospital, Beijing 10 0 044, China

Keywords:Anillin DNA methylation Copy number variation Hepatocellular carcinoma Prognosis

ABSTRACT Background: Anillin (ANLN) is required for tumor growth.It has been proven that knockdown of ANLN effectively reduces the occurrence of hepatocellular carcinoma (HCC) in transgenic mice.However,the functional role of ANLN in HCC patients remains to be elucidated.Methods: Both microarray and TCGA project were used for the analyses of ANLN expression and regulation in HCC.The effect of ANLN on proliferation and cell cycle was detected by CCK-8,colony formation assay and flow cytometry.ANLN expression was measured by immunohistochemistry.Correlation between ANLN expression and clinicopathological features was assessed by Pearson Chi-square test and 5-year overall survival after liver resection was evaluated by Kaplan-Meier method.Results: Increased copy number,decreased methylation levels in the CpG island and upregulated histone hypermethylation of ANLN were found in HCC.Knockdown of ANLN inhibited proliferation and induced G2/M phase arrest in SMMC-7721 cells.ANLN was mainly expressed in the nucleus and showed significantly higher expression levels in cancerous tissues than those in paired adjacent tissues.Moreover,nuclear ANLN expression levels in HCC metastases were significantly higher than those in primary HCC.The results of Cox proportional hazards regression model suggested that ANLN nuclear expression in HCC was an independent risk factor for poor 5-year overall survival of patients after liver resection.Conclusions: ANLN is a potential therapeutic target for HCC.Patients with nuclear ANLN overexpression in HCC tissue may need adjuvant therapy after liver resection.

Introduction

The incidence of primary hepatic carcinoma (PHC) ranks sixth,and the mortality ranks second of all malignant tumors worldwide [1] .Hepatocellular carcinoma (HCC) accounts for 85% of PHC cases [2],and 50% of global HCC cases occur in China [1] .Despite a wide range of therapeutic options,the long-term mortality is still as high as 95% due to HCC recurrence and metastasis [1] .We urgently need effective therapeutic targets to improve adjuvant treatment after surgical operation.

Anillin (ANLN) is an actin-binding protein essential for assembly of cleavage furrow during cytokinesis which is critical for cell division.Field and Alberts initially discovered that ANLN participates in cytokinesis and cell cycle regulation [3] .ANLN is recruited to the cleavage furrow and links to RhoA,septins,myosin,and actin during the anaphase and telophase,affecting the cell’s polarity and motor ability [4-6] .ANLN mRNA expression has been observed in diverse human tissues,and it increases with increasing malignancy [7] .Overexpression of ANLN contributes to poor prognosis in breast cancer [8],non-small cell lung cancer(NSCLC) [9],colorectal cancer [10],pancreatic ductal adenocarcinoma (PDAC) [11] and bladder urothelial carcinoma [12] .It has been proven that knockdown of ANLN effectively inhibits the tumorigenesis of HCC in transgenic mice [13] .Recently,ANLN has been proven necessary for the growth of HBV-related HCC [14] .However,the functional role of ANLN in HCC patients remains to be elucidated.

The large amounts of transcriptomic data currently available,both from microarrays and next generation sequencing (NGS),have made it possible to systematically investigate the functional role of a single gene in cancers.In the current study,we systematically examined the dysregulation of ANLN using large numbers of data sets from both microarray and NGS data.

Methods

Data set

All of the array data related to cancers from the Affymetrix human genome U133 plus 2.0 platform were downloaded from the GEO database (http://www.ncbi.nlm.nih.gov/geo/) based on the methods mentioned previously [15] .Average rank score and percentile rank profile were used for the measurement of the general gene expression intensity across samples and for expression profile comparison as described previously [16] .Higher average rank score indicates higher expression level.RNA-seq,methylation data,together with clinic data from TCGA (The Cancer Genome Atlas) cohorts were directly downloaded from the cBioPortal (http://www.cbioportal.org/) [17] and UCSC Xena functional genomics (http://xena.ucsc.edu/) databases.For RNA-seq,gene expression levels were measured using normalized RSEM (RNA-Seq by Expectation Maximization) value [18] .DNA methylation was described as beta value,which is a continuous variable between 0 and 1.Higher beta values represent higher level of DNA methylation.

The public data for histone modification were from GSE112221 [19] .The SRA (Sequence Read Archive) Toolkit(https://trace.ncbi.nlm.nih.gov/Traces/sra/) was used to download the data set and transform data formats.Bowtie 2 [20] was used for reads mapping.Data were normalized for graph illustration based on sample reads counts.In the data set,one normal and two HCC tissues were measured by ChIP-Seq.

Patient selection and clinicopathological information collection

The samples consisted of 20 metastatic lesions of HCC and 197 HCC lesions and adjacent tissue specimens.The latter were from resected livers of 41 recurrent HCC patients and 156 primary HCC patients.They were collected from the pathology department from January 2008 to December 2014.

The medical records of patients enrolled were composed of demographic information,clinicopathological features and follow-up date.All postoperative patients were followed up by telephone or at the outpatient clinic.

Immunohistochemistry (IHC)

The thickness of tissue sections on slides from Paraffin blocks was 4 μm.The IHC procedure consisted of de-paraffin,blockage of endogenous peroxidase activity,antigen retrieval in EDTA buffer,incubation of anti-ANLN antibody (rabbit anti-human,1:50,HPA005680,Sigma-Aldrich,St Louis,USA) and secondary antibody(PO-7000,Xi Ya Biology,Beijing,China),10 min of staining development with DAB (ZLI-9409,ZSGB-BIO,Beijing,China),counterstain with hematoxylin,dehydration and mounting.

Immunohistochemical staining of ANLN including tumor lesions and adjacent tissue specimens was evaluated by two senior pathologists using the blind method.The nuclear and cytoplasmic staining were both evaluated.A 4-point scale according to the relative number of ANLN-positive staining cells [21] was used:0,≤10%positive;1,11%-25% positive;2,26%-50% positive;3,≥51% positive.All samples were divided in to ANLN(-) (0) and ANLN(+) (1-3).

Cell culture and transfection

HCC SMMC-7721 cell line obtained from the Cell Bank of Beijing Biology Institute,Chinese Academy of Science (Beijing,China)was cultured in RPMI-1640 medium supplemented with 100 μg/mL streptomycin (HyClone,Logan,UT,USA),100 U/mL penicillin (Hy-Clone,Logan) and 10% fetal bovine serum (FBS;Gibco,Life Technologies,Carlsbad,CA,USA) at 37 °C in an atmosphere of 5%CO2.ANLN was knockdown by the small interfering RNA (siRNA)(Genepharma,Suzhou,China).The target sequence for ANLN is:siANLNa,5’-GCA GCA GAU ACC AUC AGU GTT-3’ siANLNb,5’-GCU ACA UUC UGU UCC CAA ATT-3’ and siANLNc,5’-CCA GAC CUC UGC UUU CAA ATT-3’.A random sequence was used as a negative control (NC) with sequence:5’-UUC UCC GAA CGU GUC ACG UTT-3’.SMMC-7721 cell line was selected for transfection and LipofectamineTM3000 Transfection Reagent (Thermo,South Logan,UT,USA) was used according to the manufacturer in all transfections.

RNA extraction, reverse transcription and qPCR

According to the manufacturer’s protocols,the total RNA of HCC cell lines was extracted using TRIzol reagent (Invitrogen,Life Technologies) and performed reverse transcription with QuantiTect Reverse Transcription Kit (Qiagen).And then,QuantiFast SYBR Green PCR kit (Qiagen,Dusseldorf,Germany) was used for qPCR on ABI 7000 real-time PCR machine (Applied Biosystems,Foster City,CA,USA).The primers of ANLN used were:5’-TGC CAG GCG AGA GAA TCT TC-3’ (sense),5’-CGC TTA GCA TGA GTC ATA GAC CT-3’ (antisense);GAPDH:5’-AAG GTG AAG GTC GGA GTC AA-3’ (sense),5’-AAT GAA GGG GTC ATT GAT GG-3’ (antisense).The GAPDH was used as an internal control.The reactions conditions conformed to the instruction of QuantiFast SYBR Green PCR kit.

Western blotting analysis

The Cell Lysis Buffer supplemented with 1% PMSF (Cell Signaling Technology,Boston,MA,USA) was used for protein extraction,10% SDS-PAGE gels electrophoresis for separation.And then the proteins were transferred to polyvinylidene fluoride (PVDF) membranes (PALL Life Science,New York City,NY,USA).The membranes were blocked with 5% nonfat milk for 1 h at room temperature,incubated with anti-ANLN (rabbit anti-human,1:500,HPA005680,Sigma-Aldrich) over night at 4 °C .The GAPDH was used as an internal control.The incubation with secondary antibody (1:1000;Xi Ya Biology) conjugated to horseradish peroxidase was for 1 h.And then,we detected the immunolabeled protein with the electrochemiluminescence (ECL) system (Thermo).All experiments were repeated independently for three times [22] .

Cell proliferation and colony formation assay

After transfected with NC,siANLNa or siANLNb for 48 h,SMMC-7721 cells were seeded in 96-well plates (2000 cells per well) and the Cell Counting Kit-8 (CCK-8) (Dojindo,Kyushu,Japan) solution(10μL per well) was added.The optical density values (OD) were measured at 450 nm by microplate reader (Thermo) after incubation at 37 °C in an atmosphere of 5% CO 2 for 2 h.The measurement was repeated every 24 h until a total of 4 times.The whole experiment was repeated three times.

After transfection with NC,siANLNa or siANLNb for 48 h,SMMC-7721 cells were seeded in 6-well plates (150 cell per well).The renewal of medium was performed every three days.Colonies were stained and counted two weeks later.Experiments were performed in triplicate as well.

Cell cycle assay

SMMC-7721 cells (2 × 105cells per well) were seeded in 6-well plates and then transfected with NC,siANLNa or siANLNb for 48 h.According to the manufacture we collected the cells and treated them with cell cycle staining solution (BD,Franklin Lakes,NJ,USA).Then the DNA content was analyzed by the flow cytometer (BD,Franklin Lakes,NJ,USA).Experiments were performed in triplicate.

Statistical analysis

Omic data analysis,statistics and graphing were performed under R software environment (https://www.r-project.org/).The database of all patients was built and analyzed by IBM SPSS version 20.0 (SPSS Inc.Chicago,IL,USA).All continuous variables were evaluated with the Student’sttest or Wilcoxon rank-sum test and the qualitative variables were carried out by Pearson Chi-square test or Fisher’s exact test.Survival curves were plotted using the Kaplan-Meier method and compared with log-rank test.Cox proportional hazards regression model was used to estimate the impact of ANLN expression on overall survival (OS) in multivariable analysis.Two-sidedP＜0.05 was considered statistically significant.

Results

ANLN was upregulated in HCC

According to the microarray data,compared to normal liver,ANLN was upregulated at the RNA level in HCC (average rank score:72.66 vs.40.96,P＜0.001).In accordance with the different average rank scores,higher expression was clearly illustrated by the rank-based gene expression profiles,such as in HCC(Fig.1 A),showing different expression levels.The upregulation was also confirmed by TCGA data (Fig.1 B).

ANLN overexpression at the mRNA level was observed from both microarray and TCGA data.ANLN expression at the protein level was further evaluated by IHC in 156 pairs of primary HCC and adjacent tissues,41 pairs of recurrent HCC and adjacent tissues and 20 HCC metastatic lesions (Fig.1 C).The results confirmed ANLN upregulation in cancerous tissues (Fig.1 D).Moreover,we found that in cancerous tissues,ANLN was mainly expressed in the nucleus (Fig.1 E),and the ratio of nuclear expression significantly increased with increasing malignancy (primary,metastatic)(Fig.1 F).

Upregulation of ANLN in HCC involves copy number variation, DNA and histone methylation alteration

There were 31% of ANLN copy numbers in HCC manifesting as amplification and gain (Fig.2 A).ANLN mRNA expression levels were significantly increased in the samples with increased copy number compared to those with normal copy number (Fig.2 B).ANLN upregulation was also observed in cancerous samples with normal copy number (Fig.2 C).One of the seven CpG sites,cg19529675,which located within 200 bp from the transcription start site,showed significantly lower DNA methylation levels(P=2.58E -09) (Fig.2 D).In liver cancer,the low methylation levels for this site were further confirmed by non-TCGA data,GSE54503(Fig.2 E) [23] .Hypomethylation was not detected in the nine segments located in the ANLN gene body (Fig.2 F).Upregulated histone methylation was found at ANLN gene locus in HCC from public non-TCGA data,GSE112221 (Fig.2 G) [19] .These data suggested that hypomethylation in some CpG sites and histone hypermethylation might also play important roles in ANLN upregulation in HCC.

Downregulation of ANLN expression induced G2/M phase arrest in SMMC-7721 cells and inhibited proliferation

ANLN mRNA expression levels in SMMC-7721 cells transfected with siANLNa,siANLNb,or siANLNc cultured for 24 h were reduced by 95.5%,94.4%,and 52.9% compared with the NC,respectively.Western blotting results also showed stronger downregulation by siANLNa and siANLNb than by siANLNc (Fig.3 A).Next,we examined the effects of ANLN on growth through CCK-8 and colony formation assays.The viability of SMMC-7721 cells transfected with siANLNa or siANLNb significantly decreased from the second day(Fig.3 B) and the colony formation rate significantly decreased after downregulation of ANLN by siRNA (Fig.3 C).Flow cytometry was used to detect the potential reason.After transfection with siRNAs,the percentage of cells in the G2/M phase increased from 4.08%to 34.32% or 28.07%,and the percentage of cells in the G0/G1 or S phase concomitantly decreased (Fig.3 D).As suggested in Fig.3,ANLN may promote HCC development through affecting the transition from the G2 stage to the M stage.

Positive ANLN expression predicted poor prognosis

ANLN protein expression levels were evaluated by IHC in 156 primary HCC patients after liver resection.

As shown in Table 1,the preoperative serum alpha-fetoprotein(AFP) (＞400 ng/mL),alkaline phosphatase (＞135 U/L),and albumin (＜35 g/L) levels,multiple tumor number,tumor size (＞8 cm),poor histological tumor differentiation,peritumor intravascular cancer emboli,satellite lesions,positive ANLN expression,liver capsule invasion,and invasion of portal vein,biliary duct,or hepatic vein were all potential poor prognostic factors for 5-year OS as evaluated by the log-rank test (P＜0.05).The multivariate analysis showed that positive ANLN expression (OR=3.884,95% CI:1.4 45-10.4 41,P=0.007) and liver capsule invasion (OR=10.168,95% CI:1.195-86.534,P=0.034) were independent risk factors of poor 5-year OS.

Table 1 Prognostic factors for 5-year survival after liver resection evaluated by univariate and multivariate analyses.

In order to estimate the prognostic value of ANLN expression for HCC patients,the 5-year OS after liver resection was assessed.As shown in Fig.1 G,the median survival time of ANLN-positive patients is much shorter than that of ANLN-negative patients after liver resection (16.7 vs.46.7 months,P＜0.001).For further verification of the independent prognostic value,the 5-year OS in the subgroups of other independent factors was analyzed.In the subgroup of liver capsule invasion (Fig.1 H and I),patients with nuclear ANLN expression in HCC tissue were associated with poorer long-term outcome than those with negative ANLN expressions.

Discussion

In this study,we found that ANLN was significantly upregulated in HCC.ANLN had increased copy numbers and decreased methylation levels in the CpG island,partly explaining the ANLN overexpression observed in HCC.We also investigated the clinical value of ANLN in primary HCC at the protein level,and found that nuclear overexpression was a marker of poor prognosis in patients after liver resection.

Fig.1.ANLN is upregulated at the mRNA level and protein levels in HCC tissue.A:ANLN is upregulated in HCC based on array data.The X-axis represents the expression levels.B:ANLN is upregulated based on TCGA data.C:HE and IHC results show nuclear overexpression of ANLN in primary HCC (5 and 6),recurrent HCC (7 and 8),and HCC metastases (9 and 10).Negative in adjacent normal tissues (1 and 2) and positive in cytoplasm in adjacent normal tissues (3 and 4).The scale bar represents 100 μm.D:Positive ANLN expression in HCC tissue (119/197,60.4%) is higher than that in adjacent tissue (22/197,11.2%).E:In those tissues expressing ANLN (119 out of 197 in HCC and 22 out of 197 in adjacent normal tissue),ANLN is mainly expressed in nucleus (nucleus expression:114/119,95.8%;cytoplasmic expression:5/119,4.2%) in HCC tissue,as compared to that in adjacent normal tissue (nucleus expression:0/0;cytoplasmic expression:22/22,100%).F:The ANLN positive rate in HCC.Recurrent HCC 68.3%,primary HCC 58.3% and metastasis 85.0%.G:Overall survival rate (n=156).H and I:The effect of ANLN expression on overall survival in HCC patients with different liver capsule invasion result.*P ＜ 0.05;*** P ＜ 0.001.ANLN:anillin;HCC:hepatocellular carcinoma;IHC:immunohistochemistry.

Fig.2.Copy number variation and the methylation levels of seven CpG sites in the CpG island,gene body and histone of ANLN in HCC.A:The percentages of amplification,gain,normal,shallow deletion and deep deletion of ANLN copy number in HCC.B:Increased copy numbers are correlated with ANLN upregulation.C:ANLN is upregulated in samples with normal copy number.D and E:Methylation levels of seven CpG sites based on TCGA and microarray data.F:Methylation levels of nine ANLN gene body segments based on TCGA data.G:Histone methylation level at ANLN gene locus in HCC based on microarray data.H3K4me1 and H3K4me3 represent mono-and tri-methylation levels of H3 at Lys4,respectively.ANLN:anillin;HCC:hepatocellular carcinoma.

The uncontrolled proliferation of cells is the essence of cancer.During cytokinesis,ANLN is involved in the assembly of cleavage sulcus,which is essential for cell division [24] .The downregulation of ANLN inhibits cell division,resulting in multinucleate cells [4] .Therefore,it is important to explore the role of ANLN in tumor.There are only three literatures on the role of ANLN in HCC [7,13,14] .Hall et al.first found that the expression level of ANLN mRNA in liver tumors is higher than that in normal tissues [7] .In mice,H2.35 hepatocyte with downregulated ANLN expression forms tumors more slowly than control cells,and liver tumors occur less frequently in mice [13] .The study by Lian et al.showed that the deficiency of ANLN results in the activation of DNA damage checkpoints and apoptosis signaling,increasing multinucleated cells,reducing colony formation,and growth restraint.HBV infection increases ANLN expression via the inhibition of microRNA-15a and microRNA-16-1.HBV-associated HCC patients with high expression of ANLN have poor prognosis [14] .

Fig.3.Effects of downregulation of ANLN on the proliferation and cell cycle of SMMC-7721 cells.A:Effective downregulation of ANLN expression at mRNA and protein levels.B:The effect of ANLN downregulation on proliferation.C:The colony formation rate markedly decreased after ANLN downregulation.D:Downregulation of ANLN leads to a marked increase in G2/M-phase cells and a marked decrease in cells in the G0/G1 or S phase.NC:negative control;ANLN:anillin.*P ＜ 0.05;**P ＜ 0.01;*** P ＜0.001.

We reported some new findings about ANLN in HCC.Our results indicated that increased copy number,decreased methylation levels at certain CpG sites and histone hypermethylation might result in upregulation of ANLN.Compared to samples with normal copy numbers,the increase in number of ANLN copies were associated with an increase in gene expression levels.However,we found that in cancerous samples with normal copy numbers,ANLN was also upregulated.This suggested that in addition to copy number variation,other regulatory mechanisms also played a role.DNA methylation,especially the presence of multiple methylated CpG sites in promoter CpG islands,usually leads to gene silencing [25] .Therefore,we further evaluated methylation levels at seven CpG sites on ANLN CpG island based on TCGA methylation data from the Illumina Infinium Methylation 450K platform.We found that the overall methylation levels at these sites were very low in both cancerous and para-cancerous normal tissues.For example,the quantile values of beta values in all seven sites were 0.0235 (Q1,25th percentile),0.0524 (Q2,median),and 0.0687 (Q3,75th percentile),while the quantiles of normal tissue were 0.0231,0.0565 and 0.0727,respectively.Since beta values generally presented a bimodal distribution with two peaks around 0.1 (low methylation)and 0.9 (high methylation),these data revealed that ANLN was lowly methylated in the promoter region.However,we still observed that one CpG site,cg19529675,located in the CpG island within 200 bp from the transcription start site,showed a significant methylation difference in HCC,although the level of difference in DNA methylation between HCC and normal tissue observed in the promoter region was not significant.Methylation in gene body can also affect gene expression [26] .However,theANLNgene body methylation level did not change significantly in our study.Histone modification is one of the main factors that regulate gene expression.Therefore,we further investigated the differences in histone methylation between HCC and normal liver tissues using publicly available data GSE112221 [19],and found that both mono-and trimethylation levels of H3 at Lys4 were enhanced in HCC compared to normal tissues,suggesting that histone modification at this locus could be a key factor to regulate ANLN expression.

We had two interesting findings about the expression of ANLN in HCC.Firstly,the vast majority of ANLN proteins (95.8%) in HCC tissues were present in the nucleus,which was significantly different from the cytoplasmic expression in adjacent tissues.This suggested that,at least in HCC,nuclear ANLN was the dominant active form.A previous study [27] has revealed that subcellular localization of ANLN predicts different survival outcomes in urothelial carcinoma of the upper urinary tract.The entry of ANLN into nucleus can be blocked by LY294002,a specific inhibitor of the catalytic subunit of PI3K,or another selective AKTinhibitor,NL71-101 through the PI3K/AKT signaling pathway [9] .Studies [28,29] have shown that the participation of both PI3K and AKT can regulate cell cycle progression and apoptosis.Furthermore,patients with the nucleus-ANLN-positive NSCLC showed shorter cancer-specific survival times [9] .Unfortunately,there were not enough cytoplasm-ANLN-positive HCCs as control for the analysis of clinical significance.In addition,our study found that the expression of ANLN in distant HCC metastasis was much higher than that in primary HCC,while no significant difference appeared between recurrent and primary HCC,showing that ANLN mRNA expression in liver tissue increases with increasing malignancy (primary,recurrent,metastatic) [7] .

Our results indicated that downregulation of ANLN can induce G2/M phase arrest in SMMC-7721 cells.By the CCK-8 and colony formation arrays,we demonstrated that ANLN could promote HCC proliferation.Knockdown of ANLN has been proven to inhibit the proliferation and migration of the PDAC cell line PANC-1,the NSCLC cell line LC319,the breast cancer cell lines MDA-MB-231 and ZR-75-30 [9,11,30] .When ANLN is lacking,the rapid oscillatory movement of cytosol and DNA across the furrow could occur,leading to the failure of cytokinesis [6,24] .These data further suggest that ANLN is a potential therapeutic target for cancer therapy.

In summary,in HCC,overexpression of ANLN involving copy number variation,DNA and histone methylation alteration is a marker of poor prognosis.Downregulation of ANLN expression inhibited proliferation and induced G2/M phase arrest in SMMC-7721 cells.ANLN is a potential therapeutic target for HCC patients.

Acknowledgments

We thank Professor Ping-Zhang Wang from Department of Immunology,School of Basic Medical Sciences,Peking University Health Science Center for the guidance and contribution about bioinformatics analysis.

CRediTauthorshipcontributionstatement

Long-HuiZhang:Conceptualization,Data curation,Formal analysis,Investigation,Methodology,Software,Validation,Visualization,Writing -original draft,Writing-review &editing.Dong Wang:Conceptualization,Data curation,Investigation,Project administration.ZhaoLi:Conceptualization,Formal analysis,Funding acquisition,Project administration,Resources,Supervision.Gang Wang:Data curation,Investigation.Ding-BaoChen:Investigation,Visualization.QianCheng:Data curation,Project administration,Supervision.Shi-HuaHu:Data curation,Methodology,Visualization.Ji-YeZhu:Conceptualization,Funding acquisition,Methodology,Project administration,Resources,Supervision,Writing -review &editing.&editing.

Funding

This study was supported by the grants from National Natural Science Foundation of China (81570590 and 81502509).

Ethicalapproval

This retrospective study was approved by the Ethics Committee of Peking University People’s Hospital.Written informed consents were obtained from all patients.

Competinginterest

No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.