Review Article Dysregulated microRNAs with prognostic ... · Dysregulated microRNAs with prognostic significance in human hepatocellular carcinoma: a study based on high-throughput

Int J Clin Exp Pathol 2017;10(2):2394-2415www.ijcep.com /ISSN:1936-2625/IJCEP0040855

Review Article Dysregulated microRNAs with prognostic significance in human hepatocellular carcinoma: a study based on high-throughput microarray

Han-Lin Wang1, Zhong Tan1, Yi-Huan Luo1, Xiao-Na Liang1, Yue-Qi Cao1, Huo-Jie Xiong1, Ting-Qing Gan2, Yi-Wu Dang1, Li-Hua Yang2, Gang Chen1

Departments of 1Pathology, 2Medical Oncology, The First Affiliated Hospital of Guangxi Medical University, Nan-ning, P. R. China

Received September 27, 2016; Accepted January 5, 2017; Epub February 1, 2017; Published February 15, 2017

Abstract: Hepatocellular carcinoma (HCC) is the fifth most widespread type of malignancies with only 18% of post-operative 5-year survival. However, the prognostic value of microRNAs (miRNAs) for HCC has not been fully studied. In this study, we aimed to identify the potential candidates of dysregulated miRNAs for HCC prognosis in the profiling studies based on high-throughput microarray. The literature and NCBI Gene Expression Omnibus (GEO) database were used to identify HCC-related miRNAs. Based on the ranking results of a vote-strategy method, the most con-stantly stated dysregulated miRNAs were chosen as the predictive biomarkers for HCC patients’ survival. For meta-analysis on prognosis, eligible studies were searched and sufficient data were collected adequately. Pooled hazard ratios (HRs) with 95% confidence intervals (CIs) were used to estimate the effect size. The top six upregulated and six downregulated miRNAs were identified based on 43 published miRNA profiling studies. Additional 30 stud-ies involving 2735 HCC patients with follow-up data were also included. Among the top 12 dysregulated miRNAs, miR-222, miR-25, miR-221, miR-21, miR-214, miR-199a-3p and miR-199a-5p were closely associated with HCC patients’ survival. Both univariate and multivariate results suggested that the low quality of life for HCC patients was significantly related with high miR-21, miR-221, miR-25 and low miR-199a-3p, miR-214. In conclusion, miR-222, miR-221, miR-21 and miR-25 were significantly related with dismal prognosis for HCC patients, and miR-199a-3p and miR-214 emerged as significant predictors of unfavorable prognosis for HCC patients. These miRNAs may serve as potential candidates for HCC prognosis and therapy.

Keywords: MicroRNAs, gene expression profiling, hepatocellular carcinoma, prognosis, biological markers, meta-analysis

Introduction

Hepatocellular carcinoma (HCC) ranks the fifth in the most common types of malignant tumors all over the world and HCC is the second lead-ing cause of cancer-related death globally as well [1]. Almost 500,000 new cases of HCC occurred every year in the Asia-Pacific region, with estimated 360,000 deaths in Far East countries only [2]. China alone accounted for over 60% of the total HCC cases, most of which were related to chronic hepatitis B virus (HBV) infection [3, 4]. The high mortality with almost 0.7 million deaths from HCC each year is prob-ably attributed to the failure of survival predic-tion and the delayed treatment for patients

[5-8]. Although surgical resection, radiation and chemotherapy are optional treatments for a potential cure currently, many cases of patients may still develop recurrence, which converts the situation to a poor prognosis, lead-ing to a postoperative 5-year survival of 18% [9-13]. Due to the lack of potential predictor for diagnosis and prognosis in early stages, a majority of HCC patients miss the best opera-tion time and suffer a poor postoperative sur-vival [14-17]. It is thus urgent to identify novel prognostic targets and biomarkers for HCC and demonstrate the mechanisms underlying HCC in molecular level.

MicroRNAs (miRNAs), which belong to a cluster of endogenously expressed, non-coding small

http://www.ijcep.com

Dysregulated microRNAs with prognostic significance in HCC

2395 Int J Clin Exp Pathol 2017;10(2):2394-2415

RNAs with approximately 22 nucleotides, con-trol gene expression by targeting mRNA and trigger either RNA degradation or translation repression [18]. These multistep biological pro-grams are closely associated with the regula-tion of various cellular processes such as cell proliferation, death and tumor carcinogenesis [19]. Growing evidence indicates that dysregu-lation of miRNAs can be involved in the inva-sion and metastasis in various cancers, espe-cially in HCC [20-26]. Therefore, the roles of miRNAs in liver cancers made it possible to be promising biomarkers to accurately predict prognosis and to be potential targets for high-efficient treatment.

Nowadays high-throughput technologies were globally used to explore miRNA expression across diverse tumor and corresponding non-tumor tissues, which contributed to the identifi-cation of specific aberrant miRNAs involved in oncogenesis. However, with more complex information on the importance of miRNAs in HCC provided by accumulating findings, some miRNAs were reported to express in an incon-sistent direction among the profiling studies partly because of small sample size. To over-

prognosis and their targeting molecules with relevant pathways, which may provide potential biomarkers for HCC development.

Materials and methods

Search strategy

MiRNA profiling studies on prognosis of HCC were searched on databases including PubMed, Web of Science, Wiley Online Library, EMBASE, Ebsco, Chinese National Knowledge Infras- tructure (CNKI), Chinese Chong Qing Yangtze Island Club (VIP) and Chinese Wan Fang. The search strategy was built by means of the MeSH terms up to date Aug 16th, 2015: ‘miRNA OR microRNA OR miR’ in combination with the keyword ‘neoplas* OR tumor OR malignan* OR carcinoma OR cancer’ and ‘HCC OR hepatocel-lular OR liver OR hepatic’.

Inclusion criteria of the literature

For microarray studies, eligible studies were included if they met the principles: (i), miRNA profiling studies in HCC; (ii), tissue samples gained surgically including resected liver tumor

come the limitations in current studies and increase the statistical persuasiveness, our meta-analysis was per-formed, which com-bined the results of several individual stud-ies and improved the precision and accuracy of estimation. Firstly, we performed a sys-tematic review of 43 published studies that compared the miRNAs expression in HCC tis-sues with that in nor-mal liver tissues, and then made efforts to get complicated information of miRNA expression data from these disparate origi-nal profiling datasets and integrated them systematically. Subse- quently, our meta-anal-ysis was carried out to identify the miRNAs correlated with HCC

Figure 1. Flow diagram of liter-ature reviewing and selection for miRNA profile studies.



Table 1. Characteristics of 43 miRNA profiling studies included in this systemic review

First author (reference*) Year

HCC patients Differentially expressed miRNAs

Origin PeriodNo. of collected tissue

samples (tumor/normal)

Cut-off Criteria

Total

Up-regulated MiRNAs in HCC

Down-regulatedMiRNAs in

HCC

Platform

Zhang [1] 2014 Shanghai, China NR 85 (75/10) FC>3 P<0.05

13 8 5 MiRCURYTM

LNA Array (v.11.0)

Zhang [2] 2015 Hong Kong SAR, China Up to Feb.2015 327 (327/43) FDR<0.001 FC>3

P<0.001

50 13 37 Illumina HiSeq 2000 miRNA sequencing Plat-forms (Illumina Inc., San Diego, CA)

Park [3] 2015 Korea NR 16 (9/7) FC>1.5P<0.05

267 10+ 10+ MiRNA 3.0 Array; DNA Link

Li [4] 2013 Tianjin, China Apr.2003 to Aug.2010 22 (11/11) FC>2P<0.05

18 9 9 RT2 miRNA PCR array

Wei [5] 2013 Guangzhou, China 2004 to 2007 220 (110/110) FDR = 0 P<0.1

30 20 10 Custom microarray

Murakami [6] 2006 Japan NR 46 (24/22) P<0.0001 30 3 5 Human miRNA microarray

Bao [7] 2014 Inner Mongolia Medical University, China

Jan.2006 to Jun.2013 18 (9/9) -2<log2Ratio<2 35 19 16 Digital miRNA expression profiling

Gramantieri [8] 2007 American NR 38 (17/21) FC>1.3 P<0.05

35 1 34 KCC/TJU miRNA microarray chip

Li [9] 2008 Jiangsu Province, China NR 177 (78/88) FRD<0.01 69 29 40 NR

Li [10] 2014 Xi’an, China 2010 to 2012 6 (3/3) P<0.05 10 6 4 μParaflo microfluidic chip

Shen [11] 2015 Columbia NR 20 (10/10) FRD<0.05 25 6 19 TaqMan Low Density Arrays (TLDA, Applied Biosystems, Foster City, CA)

Liu [12] 2011 Hong Kong 1990 to 2007 188 (94/94) P<0.05 9 7 2 Custom Affymetrix array, RM-HU01Aa520485 RSTA Custom Affymetrix 1.0

Chung [13] 2010 Korea 2001 to 2004 50 (25/25) FRD<0.05 P<0.05

21 14 7 MiRNA microarray slide consisting of 308 probes for human miRNAs from Rosetta Genomics Corp

Elyakim [14] 2010 South Korea NR 60 (30/30) NR 35 27 8 Custom microarrays

Wang [15] 2012 Shanghai, China NR 6 (3/3) FC>3 P<0.01

86 10+ 13+ Taqman low density miRNA array (TLDA)

Katayama [16] 2012 Tokyo Nov.2005 to May.2008 46 (40/6) FC>1.5P<0.05

48 18 30 MiRNA microarray using 3D-Gene (Toray Industries, Tokyo, Japan)

Han [17] 2014 Chongqing, China NR 19 (10/9) FRD<0.05 |logFC|>1

32 14 18 Homo sapiens miRNA profiling platform version 4 (Luminex, New York, NY, USA)

Gao [18] 2015 Guangxi, China Jan.2012 to Dec.2013 60 (30/30) FC>2.0 33 21 12 Agilent 8x60K microarray

Huang [19] 2007 Shenzhen, China Mar.2007 to Jul.2007 20 (10/10) FC>2P<0.05

16 15 1 MiRNA microarray chips

Meng [20] 2007 American NR 6 (3/3) P<0.05 15 9 6 MirVana miRNA Bioarrays

Su [21] 2009 Guangzhou, China 2005 to 2006 8 (5/3) FC>2 P<0.05

29 14 15 CapitalBio human/mouse/rat non-coding RNA microarray

Burchard [22] 2010 Hong Kong 1990 to 2007 192 (96/96) P<0.05 OR P<0.01

1 0 1 Custom Affymetrix array, RM-HU01Aa- 520485 RSTA Custom Affymetrix 1.0



Shih [23] 2012 Taiwan 1999 to 2000 89 (68/21) FC>1.5 FDR<10-5

P<0.01

63 30 33 Sentrix Array Matrix (Illumina, San Diego, CA)

Yang [24] 2010 China NR 13 (8/5) FC>2 P<0.05

14 6 8 MiRCURYTM LNA Array (v.11.0)

Wang [25] 2012 Shanghai, China NR 6 (3/3) FC>3 20 10 10 Taqman low density miRNA array (TLDA)

Xia [26] 2013 Singapore NR 30 (20/10) NR 2 2 0 GeneChip miRNA 2.0 Array (Affymetrix, Santa Clara, CA, USA)

Li [27] 2014 Shanghai, China NR 106 (53/53) P<0.05 1 0 1 TaqMan Low Density Array (TLDA)

Chen [28] 2015 Chongqing, China Jun.2010 12 (6/6) FC>2 or <0.5 P<0.05

92 10+ 10+ MiRNA array analysis (Western Technology, Inc., Chongqing, China)

Wong [29] 2012 Hong Kong 1999 to 2009 40 (20/20) P<1.34×10-4 30 7 23 Low density microarray (LDA)

Sato [30] 2011 Japan Jan.1997 to Mar.2007 73 (69/4) P<0.01 91 42 49 Toray’s 3D-GeneTM human miRNA chips (miRBase version 12)

Li [31] 2008 Beijing, China NR 40 (20/20) P<0.01 7 3 4 MiRNA microarrays by LC sciences

Borel [32] 2011 France NR 13 (10/3) FC>2 or <0.5 P<0.05

361 11 79 Taqman Array Human MiRNA a cards v2.0 (Applied Biosystems)

Wang [33] 2010 Shanghai, China up to Jan.2010 486 (272/214) P<0.05 FRD<0.05

603 31 25 CapitalBio Mammalian miRNA Array Services V1[1].0; CapitalBio human/mouse/rat non-coding RNA microarray

S.El-Halawany [34] 2015 Egypt 2006 to 2010 15 (15/10) FC>3 P<0.05

94 53 3 NR

Yu Wang [35] 2008 Singapore NR 38 (19/19) FRD<0.05 22 19 3 TaqMan MiRNA assays human panel early ac-cess kit (Applied Biosystems)

Huang [36] 2009 Guangzhou, China 2004 to 2006 40 (20/20) P<0.05 31 12 19 Bead-based miRNA expression profiling

Murakami [37] 2014 Japan NR 17 (11/6) p<0.05 3 0 3 Agilent human miRNA microarray release 14.0

Hung [38] 2014 Taiwan NR 58 (29/29) FC>2 or <0.5P<0.05

16 8 8 Illumina BeadArray (human v2 miRNA panel)

Ma [39] 2009 China Sept.2007 to Aug.2008 2 (1/1) NR 14 6 8 LNATM miRNA array (11.0)

Sun [40] 2006 China Jan.2005 to Jul.2005 40 (20/20) P<0.01 9 2 7 Multi-analyte suspension array

Wang [41] 2010 China Jun.2006 to Oct.2008 50 (25/25) FC>4 FRD<0.05

12 3 9 NR

Xiao [42] 2012 China NR NR FC>2 P<0.01

213 14 16 ParafloTM MiRNA array by LC sciences

Fan [43] 2013 China Jul.2011 to Sept.2013 18 (9/9) FC>2 or <0.5 P<0.05

64 35 29 Affymetrix GeneChip miRNA Arrays (Genisphere, FT30AFYB)

Abbreviations: HCC, hepatocellular carcinoma; FC, fold change; FDR, false discovery rate; miRNA, microRNA; NR, not reported; +The number of the up- or down-regulated miRNAs mentioned on top in the original study; References* here were shown in Supplementary Files.



Table 2. Deregulated miRNAs (n = 12) consistently reported in 43 profiling studies (HCC tissues versus normal tissues)

Direction of expression

MiRNA name

No. of studies with same direction (reference*)

No. of tissue

samples tested

Subset of studies with fold change

No. of studies

No. of tissue samples tested

MeanFC

Range

Upregulated

MiR-222 17 (7, 9, 14, 16, 17, 20, 21, 23, 24, 28, 30, 31, 34, 35, 36, 40,43)

712 14 455 5.54 1.57~17.42

MiR-221 15 (5, 7, 8, 9, 10, 14, 16, 20, 30, 31, 34, 35, 36, 41, 43) 845 13 428 5.67 1.49~39.43

MiR-224 14 (1, 2, 6, 9, 16, 17, 21, 26, 30, 31, 32, 35, 36, 40) 982 12 487 27.80 1.81~117.46

MiR-21 14 (5, 7, 10, 14, 16, 17, 19, 20, 30, 31, 32, 34, 35, 43) 592 12 312 6.05 1.67~31.63

MiR-25 12 (4, 9, 13, 16, 21, 28, 30, 31, 35, 36, 42, 43) 524 10 336 2.38 1.32~6.55

MiR-18a 10 (4, 5, 9, 14, 16, 21, 29, 30, 34, 43) 679 8 286 3.49 1.43~9.05

Downregulated

MiR-199a-5p

23 (2, 3, 6, 8, 9, 13, 14, 15, 20, 21, 23, 24, 25, 30, 31, 32, 34, 36, 38, 39, 40, 41, 42)

1173 12 530 0.28 0.00~0.57

MiR-99a 16 (9, 10, 11, 14, 17, 18, 21, 24, 29, 30, 31, 33, 34, 39, 42, 43)

1037 11 916 0.38 0.13~0.78

MiR-195 16 (2, 6, 8, 9, 13, 14, 17, 21, 24, 30, 31, 32, 36, 39, 40, 43)

964 11 488 0.43 0.10~0.75

MiR-130a 15 (2, 8, 9, 11, 13, 14, 17, 24, 28, 29, 30, 31, 32, 34, 39)

899 11 456 0.34 0.10~0.65

MiR-199a-3p

14 (4, 6, 9, 11, 13, 14, 23, 29, 30, 31, 32, 34, 41, 43) 713 9 348 0.37 0.16~0.67

MiR-214 14 (2, 3, 4, 8, 9, 15, 17, 23 ,25, 29, 30, 31, 35, 43) 909 9 367 0.30 0.04~0.77Abbreviations: FC, fold change. References* here were shown in Supplementary Files.

lines, case reports, letters and reviews were excluded. If survival analyses were presented in the studies with inadequate information to analyze the HR value, we would make efforts to contact the authors to acquire primary survival data wherever and whenever possible.

Data abstraction

Differentially expressed miRNAs were identi-fied from each included profiling study. The sub-sequent information was gathered from each study: first author, year of publication, charac-teristics of recruited HCC patients, defined cut-off criteria, fold change (if available) of differen-tially expressed miRNAs, and platform of miRNA detected. Then, the miRNAs ranking-based method for ranking potential molecular mark-ers by Chan and Griffith [28, 29], was adopted in the systematic review, which included: (i), number of the studies that reported the same miRNA in an accordant direction; (ii), total num-ber of patients in the study; (iii), mean fold change calculated from the originally reported fold change. After that, data of miRNAs associ-ated with prognosis were extracted cautiously from included studies by another two investiga-tors (Xiao-Na Liang, Yue-Qi Cao) independently and all items reached a consistency. The

and matching noncancerous or normal tissues for comparison [27]; (iii), use of miRNA microar-ray approaches; (iv), cut-off criteria of differen-tially expressed miRNAs and the number of samples reported. The profiling studies that used HCC cell lines or serum samples from HCC patients were included, except for those comparing HCC biopsies with no adjacent non-tumorous hepatocellular tissues, or those using other technologies to detect miRNA. Review articles were also excluded.

For prognostic studies, two independent inves-tigators (Zhong Tan, Huo-Jie Xiong) reviewed the literature quantitatively with the same multi-step procedure as mentioned above. Firstly, the abstracts helped eliminate the dupli-cated or irrelevant studies. Then, full-text of the rest articles were further reviewed by the authors individualistically to make a decision whether to include based on the inclusion crite-ria listed as follows: (i), samples should be accumulated from HCC patients; (ii), studies published in either English or Chinese should investigate the relationship between miRNAs and survival data; (iii), presented data to ana-lyze hazard ratios (HRs) value and its 95% con-fidence interval (95% CI) should be offered. Additionally, trails with either animals or cell



extracted characteristics included the first author, published year, country, number and average age of the patients, sample origin, cut-off value, follow-up times, survival analysis and prognostic data. Any controversies in the data analysis were resolved by another four review-ers (Han-Lin Wang, Yi-Huan Luo, Li-Hua Yang and Gang Chen).

Statistical analysis

The HR value was selected to evaluate the cor-relations of miRNAs with outcomes of HCC patients. The higher expressed miRNAs were more likely to be correlated with poorer survival in case that HR>1 was observed. The HR val-ues and their 95% CIs of each single research were extracted and then combined into a pooled HR. For those studies that only present-ed the primary data or Kaplan-Meier curves, HR values were calculated through the soft-ware developed by Jayne F Tierney and Matthew R Sydes [30], with either the software of SPSS20.0 or Engauge Digitizer version 4.1. With respect to heterogeneity analysis, we con-ducted Cochrane Q test (Chi-squared test) to evaluate the potential heterogeneity among the

Study selection and characteristics

In the matter of miRNAs ranking, a total of 43 studies (Figure 1; Table 1, Supplementary Files) were included according to our inclusion criteria and identification. Based on the fre-quency of miRNAs mentioned in involved stud-ies and the overall sample size as described previously, we screened the top 12 most dys-regulated miRNAs with corresponding microar-ray studies (Table 2). When referring to expres-sion of miRNAs with HCC patients’ prognosis, there were 7362 studies identified in both English and Chinese databases with the search strategies (Figure 2). We took 4848 studies into consideration within this meta-analysis after duplicates were removed. One hundred and eight full-text articles were obtained for fur-ther identification after the assessment of ini-tial studies reviewed through titles and abstracts that whether or not appropriate for evaluation of prognostic miRNA biomarkers in HCC. Amongst studies with full-text, 78 studies were eliminated with definite reasons for either absence of survival data or insufficient data to determine HR. Lastly, thirty independent stud-ies published from 2008 to 2015, which ana-

Figure 2. Flow diagram of lit-erature reviewing and selec-tion for prognostic studies.

included studies. A fixed-effect model (Mantel-Haenszel method) was employed to integrate the HR values when no statis-tically significant heteroge-neity (P>0.05) existed. In addition, a random-effect model (DerSimonian and Laird method) would be used. Simultaneously, fun-nel plots were used to assess the publication bias. Begg’s test was con-ducted to assess the pos-sibility of publication bias when the meta-analysis included 10 or more stud-ies. All of the aforemen-tioned statistical analyses were performed on STA- TA12.0 (STATA Corp., Co- llege, Texas), and two-sid-ed P value <0.05 was con-sidered statistically significant.

Results



Table 3. Summary of hazard ratios of miRNA expression in HCC

MiRNA Author Year Region Patients Age Samples Assay Cut-off (ΔCt) Follow- up (mon) Suvival analysisHR/RR (95% CI)

U M

miR-21

Wang WY 2014 China 119 NR Tissue RT-PCR Median 60 OS 1.830 (1.150-2.910) 3.189 (1.911-10.012)

DFS 1.820 (1.240-2.670) 5.897 (3.009-13.763)

Gyöngyösi B 2014 Italy 20 68 Tissue RT-PCR Median 35 OS 1.030 (0.360-2.960) /

Liu M 2014 China 136 NR Serum qRT-PCR Median 50 OS 1.429 (1.080-1.890) /

He XD 2014 China 109 74 Tissue qRT-PCR Median 50 DFS 2.110 (1.296-3.508) /

Wang W 2015 China 97 50 Serum RT-PCR NR 60 OS 2.257 (1.420-3.570) 2.229 (1.328-3.743)

Shi KQ 2015 China 107 57 Tissue qRT-PCR Optimal 80 OS 1.416 (1.057-1.897) /

Huang CS 2015 China 112 NR Tissue qRT-PCR Mean 60 OS 2.417 (1.522-8.467) 2.275 (1.394-7.924)

Hu LT 2015 China 32 NR Tissue qRT-PCR Mean 30 OS 3.820 (1.720-8.480) /

Tomimaru Y 2010 Japan 30 56 Tissue qRT–PCR Median 72 OS 2.720 (1.130-6.550) /

Karakatsanis A 2011 Greece 60 NR Tissue qRT-PCR Mean 100 OS 1.978 (1.370-2.857) /

Jiang JM 2008 America 25 74 Tissue qRT-PCR NR 350 DFS 3.210 (1.120-9.160) /

Yu L 2012 China 252 NR Plasma qRT-PCR NR 120 OS 1.390 (1.128-1.713) /

DFS 1.180 (0.947-1.471) /

miR-25

Su ZX 2014 China 131 NR Tissue qRT-PCR Median 60 OS 2.390 (1.150-4.940) 2.179 (1.876-4.335)

Sadeghian Y 2015 Iran 96 NR Tissue qRT-PCR Median 80 OS 2.000 (1.080-3.680) 2.932 (1.742-5.237)

miR-199a-3p

Hou J 2011 China 152 34 Tissue qRT-PCR Median 50 OS 0.417 (0.278-0.625) 0.476 (0.303-0.714)

Hou J 2011 China 142 46 Tissue qRT-PCR Median 50 OS 0.455 (0.294-0.714) 0.476 (0.294-0.714)

Wang PP 2015 China 135 NR Tissue RT-PCR Median 60 OS 0.350 (0.170-0.730) /

Lei L 2015 China 107 48.1 Tissue qRT-PCR Median 40 OS 0.400 (0.170-0.970) /

miR-199a-5p

Song JG 2014 China 40 NR Tissue qRT-PCR Median 60 OS 0.200 (0.080-0.520) /

Guo WJ 2013 China 120 NR Tissue qPCR Optimal 90 OS 0.470 (0.310-0.710) /

Wang PP 2015 China 135 NR tissue RT-PCR Median 60 OS 0.310 (0.160-0.620) /

miR-214

Wang PP 2015 China 135 NR Tissue RT-PCR median 60 OS 0.781 (0.717-0.830) 0.799 (0.747-0.855)

Xia HP 2012 Singapore 50 NR Tissue qRT-PCR median 150 DFS 0.360 (0.170-0.780) 0.330 (0.150-0.700)

Wang J 2013 China 65 NR Tissue RT-PCR median 60 DFS 0.440 (0.210-0.910) 0.491 (0.256-0.940)

Gyöngyösi B 2014 Italy 20 68 Tissue RT-PCR median 35 OS 0.620 (0.200-1.870) /

Shih TC 2012 China 68 55.1 Tissue qRT-PCR median 120 OS 0.350 (0.120-0.990) /

DFS 0.350 (0.170-0.730) /

miR-221

Gramantieri L 2009 Italy 51 68 Tissue RT-PCR median 130 OS 1.220 (0.685-2.174) /

Li JP 2011 China 46 NR Serum qRT-PCR Mean 60 OS 1.940 (1.030-3.780) 1.903 (1.235–2.981)

Rong MH 2013 China 48 52 Tissue RT-qPCR Median 25 OS 1.066 (0.912-1.245) 1.146 (0.753-1.746)



Bae HJ 2015 Korea 154 NR Tissue qRT-PCR NR 60 OS 2.700 (1.670-4.350) /

Gyöngyösi B 2014 Italy 20 68 Tissue RT-qPCR Median 35 OS 1.920 (0.610-6.100) /

Karakatsanis A 2011 Greece 60 NR Tissue qRT-PCR Mean 100 OS 1.790 (1.500-2.130) 1.814 (1.316-2.500)

miR-222

Wong QW 2010 China 76 57 Tissue qRT-PCR Median 150 OS 3.125 (1.418-5.555) /

DFS 2.214 (1.188-3.829) /

Gyöngyösi B 2014 Italy 20 68 Tissue RT-PCR Median 35 OS 2.040 (0.640-6.460) /

Xu F 2013 China 76 NR Serum RT-PCR Mean 40 OS 2.560 (1.040-6.270) /

Zhan MX 2013 China 49 55 Serum qRT-PCR Median 40 OS 3.978 (2.061-7.678) 2.356 (1.055-5.263)

Liu K 2014 China 90 NR Tissue RT-PCR Median 60 OS 2.690 (1.440-5.040) /Abbreviations: qRT-PCR, quantitative real-time polymerase chain reaction; OS, overall survival; DFS, disease-free survival; HR, hazard ratio; RR, relative ratio; CI, confidence interval.

Table 4. Summary of hazard ratios of miRNA expression in HCC

Groups No. of studies Pooled HR 95% CI PHeterogeneity test

Statistical methodQ P I2

Overall Survival (OS) Univariate MiR-21 10 1.598 1.418-1.803 <0.001 14.310 0.112 37.1% Fixed MiR-221 6 1.623 1.155-2.281 0.005 28.030 <0.001 82.2% Random MiR-222 5 2.992 2.144-4.177 <0.001 1.380 0.847 0.0% Fixed MiR-214 3 0.777 0.723-0.836 <0.001 2.370 0.306 15.6% Fixed MiR-199a-3p 3 0.419 0.322-0.545 <0.001 0.380 0.945 0.0% Fixed MiR-199a-5p 3 0.383 0.275-0.532 <0.001 3.160 0.206 36.8% Fixed MiR-25 2 2.153 1.347-3.442 <0.001 0.130 0.714 0.0% Fixed Multivariate MiR-21 3 2.425 1.639-3.589 <0.001 0.540 0.762 0.0% Fixed MiR-221 3 1.618 1.297-2.017 <0.001 3.590 0.166 44.3% Fixed MiR-199a-3p 2 0.476 0.350-0.648 <0.001 0.000 1.000 0.0% Fixed MiR-25 2 2.424 1.734-3.390 <0.001 0.690 0.405 0.0% FixedDisease Free Survival (DFS) Univariate MiR-21 4 1.709 1.164-2.498 0.006 9.060 0.028 66.9% Random MiR-214 3 0.382 0.249-0.585 <0.001 0.220 0.895 0.0% Fixed Multivariate MiR-214 2 0.416 0.253-0.684 0.001 0.600 0.44 0.0% FixedAbbreviations: OS, overall survival; DFS, disease-free survival; CI, confidence interval.





Figure 3. Meta-analysis evaluating the relationships between miR-21 and prognosis in HCC. A. Overall survival (uni-variate); B. Disease free survival (univariate); C. Overall survival (multivariate).

Figure 4. Meta-analysis evaluating the relationships between miR-221 and survival in HCC. A. Overall survival (uni-variate); B. Overall survival (multivariate).

lyzed dysregulated miRNA levels with patients’ survival in 2735 cases of HCC, were included

in this qualitative synthesis (Table 3). In the meantime, this meta-analysis included 26 stud-



Figure 5. Meta-analysis evaluating the relationships between miR-222 and univariate overall survival in HCC.

Figure 6. Meta-analysis evaluating the relationships between miR-25 and overall survival in HCC.



ies available for overall survival (OS), 7 studies for disease-free survival (DFS), 2 studies for progression-free survival (PFS) and 5 studies for relapse-free survival (RFS). Univariate analy-sis was conducted in all researches, from which 13 were taken into account in multivariate analysis. The follow-up period varied from 25 to 350 months. Assay of all researches was quan-titative real-time polymerase chain reaction (qRT-PCR). The sample types included tissue (N = 24), serum (N = 5), and plasma (N = 1). The main characteristics involved in the survival data of the prognostic studies were shown in Table 4.

Meta-analysis

Six miRNAs that were most consistently upreg-ulated (miR-222, miR-221, miR-224, miR-21, miR-25 and miR-18a) and six most consistently downregulated miRNAs (miR-199a-5p, miR-99a, miR-195, miR-130a, miR-199a-3p and miR-214) were identified in at least ten profiling studies comparing miRNA dysregulation in HCC tissues and corresponding non-tumor liver tis-

sues (Table 2). A summary of HRs estimated and 95% CIs from the entire pooled analysis for specific miRNAs was showed in Table 4.

MiR-21

In summary, 12 studies reported survival data on miR-21 and assessed upregulated miR-21 as a predictor for survival outcome in HCC patients (Table 2). Among them, 10 performed univariate analysis on OS and four on DFS. Besides, multivariate analysis was conducted to process OS data in 3 studies. The results of univariate analysis suggested that higher miR-21 expression was significantly correlated with poorer OS, with a pooled HR of 1.598 (95% CI = 1.418-1.803, P<0.001) (Figure 3A) [31-40]. Fixed-effect model was applied since the het-erogeneity of the 10 studies was low (P = 0.112, I2 = 37.1%). On the contrary, miR-21 showed no significant association with OS in the sample set of sorafenib-treated patients (P = 0.95) in the research of Gyöngyösi et al. [36]. Multi- variate analysis was employed in 3 studies [32, 34, 37], in which no heterogeneity was found (P

Figure 7. Meta-analysis evaluating the relationships between miR-199a-3p and overall survival (univariate and mul-tivariate) in HCC.



= 0.762, I2 = 0.0%). The pooled HR was 2.425 and 95% CI was 1.639-3.589 (Figure 3C), which indicated that upregulated miR-21 may have a negative effect on patients’ prognosis. As for the DFS of miR-21 in 4 studies identified, we used a random-effect model due to the exis-tence of significant heterogeneity (P = 0.028, I2 = 66.9%) (Table 4). As shown in Table 4, overex-pression of miR-21 was significantly related with poor DFS by univariate analysis [37, 40-42], given a combined HR of 1.709 (95% CI: 1.164-2.498, P = 0.006) (Figure 3B).

MiR-221

As one of the most upregulated miRNAs in HCC, miR-221 was mentioned in 15 studies (Supp- lementary Files) including 845 tissue samples (Table 2). Survival data of OS by either univari-ate or multivariate analysis in the other 6 stud-ies indicated that high miR-221 level may func-tion as one of the candidate predictors for poor OS in HCC (n = 379) (Table 3). Five studies used HCC tissues as test samples, while Li et al. [43] tested serum sample in HCC patients. These 6 studies [36, 38, 43-46] all performed univari-ate analyses and thus a combined analysis was conducted subsequently. Random-effect model was selected as the heterogeneity of the 6 studies obviously existed (P<0.001, I2 = 82.2%). Patients with elevated miR-221 expression had a pooled HR of 1.623 (95% CI: 1.155-2.281, P = 0.005) (Figure 4A). Multivariate analysis was performed in 3 studies [38, 43, 45] with no sig-nificant heterogeneity (P = 0.166, I2 = 44.3%) in a fixed-effect model, the results of which indi-cated that miR-221 could be a candidate for poor prognosis in HCC as the pooled HR was 1.618 (95% CI: 1.297-2.017, P<0.001) (Figure 4B; Table 4).

MiR-222

MiR-222 was consistently reported upregulat-ed in HCC tissues as compared with their adja-cent non-cancerous hepatic tissues in 17 microarray studies including 712 tissue sam-ples (Table 2), which aroused our interest that whether miR-222 can serve as a predictive fac-tor for HCC patients’ poor outcome. Five stud-ies available focused on the relationship between overexpression of miR-222 and prog-nosis of HCC, 3 of which tested tissue samples and another 2 tested serum samples (Table 3). No significant inter-study heterogeneity was

found (P = 0.847, I2 = 0.0%) (Table 4) in the univariate analysis (OS) in the 5 studies [36, 47-50] , given a pooled HR estimated of 2.992 (95% CI: 2.144-4.177, P<0.001) (Figure 5). It was also proved that upregulated miR-222 can be an independent risk factor for poor HCC patients’ survival. Besides, multivariate analy-sis on OS by Zhan et al. [48] showed a HR of 2.356 (95% CI: 1.055-5.263). Wong et al. [49] conducted univariate analysis on DFS with a HR of 2.214 (95% CI: 1.188-3.829, P<0.01) and highlighted a potential prognostic value of miR-222.

MiR-25

Recent studies showed that the miR-25 was notably overexpressed in HCC tissues when compared with adjacent normal tissues (P< 0.001), suggesting an important role of miR-25 in carcinogenesis and development of HCC [51, 52]. Consistently, twelve studies with 524 sam-ples were found to conform this (Table 2). And a meta-analysis including two studies with sur-vival data was subsequently performed [51, 53]. No significant heterogeneity was found in these 2 studies (univariate: P = 0.714, I2 = 0.0%; multivariate: P = 0.405, I2 = 0.0%) and thus the fixed-effect model was used. As described in Table 4, pooled HRs of OS for univari-ate and multivariate analysis were 2.153 (95% CI: 1.347-3.442, P<0.001, Figure 6) and 2.424 (95% CI: 1.734-3.390, P<0.001) (Figure 6) respectively. The results above highlighted miR-25 could be a prognostic marker and upregula-tion of miR-25 might have predictive value for poor survival in HCC patients.

MiR-199a-3p

Among the aberrant expressed miRNAs that consistently reported in HCC, miR-199a-3p, one of the most downregulated miRNAs in HCC proved by our ranking strategy, was reported in 14 miRNA profiling studies containing 713 tis-sue samples (average FC: 0.37) (Table 2). Since miR-199a-3p was considered as a research focus and 3 studies [54-56] pointed out its dec-rement was closely associated with poor prog-nosis of HCC patients, and it was thought to be a potential protective factor for HCC. The fixed-effect model was applied since no significant inter-study heterogeneity was detected (P = 0.945, I2 = 0.0%). The combined univariate HR of OS in these 3 studies was 0.419 (95% CI:



0.322-0.545, P<0.001) (Figure 7). Another one study by Hou et al. [55] conducted two indepen-dent cohorts of HCC patients (N1 = 142; N2 = 152) to pooled the HR of OS from the multivari-ate analysis, revealing a combined HR of 0.476 (95% CI: 0.350-0.648, P<0.001) for poor OS (Fi- gure 7). The results above suggested the lower expressed miR-199a-3p featured in pathogenesis and unfavorable prognosis of HCC patients.

MiR-214

Based on the miRNA ranking results shown in Table 2, miR-214 was found to be downregu-lated in HCC tissues of 14 microarray studies with 909 samples. As for its role in HCC pro-gression and prognosis, relevant survival data to calculate the HRs of OS and DFS were pro-vided by 5 studies, all of which tested on HCC tissue samples (Table 3). As stated by our meta-analysis, univariate analysis indicated statistically significant relations of lower expressed miR-214 with OS and DFS. The pooled

HRs for OS [36, 54, 57] and DFS [57-59] were 0.777 (95% CI: 0.723-0.836, P<0.001) (Figure 8) and 0.382 (95% CI: 0.249-0.585, P<0.001) (Figure 8), respectively. Moreover, similar results were found in the multivariate analysis of DFS in HCC patients [58, 59] with a pooled HR of 0.416 (95% CI: 0.253-0.684, P = 0.001) (Figure 8). The results of pooled HRs and cor-responding heterogeneity test were listed in Table 4. The results indicated that the under-expression of miR-214 in HCC tissues might predict a more frustrating outcome for HCC patients.

MiR-199a-5p

For miRNA profiling studies, miR-199a-5p was considered to be the most downregulated miRNA in HCC tissues in our meta-analysis on 23 microarray studies with 1173 samples (Table 2). For prognosis, 3 studies [54, 60, 61] described low tumoral miR-199a-5p expression as predictor of poor survival in HCC. Moderate inter-study heterogeneity (I2 = 36.8%; P =

Figure 8. Meta-analysis evaluating the relationships between miR-214 and survival including OS (univariate) and DFS (both univariate and multivariate) in HCC.



0.206) was found thus the fixed-effect model was employed in the univariate meta-analysis, which gave a combined HR of 0.383 (0.275-0.532, P<0.001).

Publication bias

Finally, the publication bias of miR-21 in 10 studies performed by univariate analysis for OS was evaluated by funnel plots and Begg’s method. No publication bias existed in this meta-analysis since the funnel plots were sym-metry, and the P value of the Begg’s test was 0.180 (greater than 0.1) (Figure 9). We per-formed no Begg’s test for OS, DFS or RFS of other miRNAs because of the limited number of studies.

Discussion

These days, the independent prognostic bio-markers that can be easily detected in serum or tissues signify medical advancement for a growing number of HCC patients. And the high-throughput microarray has made it possible to predict prognosis for HCC patients more accu-rately depending on the genetic profile of an individual tumor. Previously, researchers had conducted certain systematic review to evalu-ate circulating miRNAs from miRNA expression profiling studies as novel potential biomarkers for HCC [9, 14, 18, 62]. Though these studies, to some extent, provided large amounts of information about biological importance of miRNAs in cancer genesis, diagnosis, progres-sion, prognosis and response to treatments,

review. Frequencies of miRNAs mentioned were taken into consideration, and we analyzed the expression levels of the top 12 miRNAs known to be dysregulated in HCC, which were then selected for further pooled analysis to uncover their roles in HCC prognosis. By combining data from several independent studies, statistical effects were increased and the consistency in association of miRNA expression with survival prognosis was evaluated, too.

MiR-21 is the well-known oncogenic miRNA in human cancers [63] and its aberrant expres-sion in HCC tissues was on average 6.05 times higher than that in non-tumorous tissues. Pre- vious meta-analyses of miR-21 had been con-ducted for diagnostic application in HCC [64-67], but relatively a few prognostic researches were published. Wang et al. [68] calculated a pooled HR of 5.77 (95% CI: 2.65-12.52) and pointed out higher expression of miR-21 was significantly correlated with worse outcomes in digestive cancer; nevertheless they failed to highlight its dysregulation significance in HCC. Besides, despite the fact that Zhou et al. [67, 69] consistently proved that overexpression of miR-21 in HCC was correlated with poor overall survival, inadequate raw survival data dented the reliability of the combined results. Based on the microarrays of HCC tissues globally and the results of miRNA ranking sort, miR-21 was selected to perform survival analysis and twelve studies including 1099 patients with comprehensive data were used to calculate the pooled HR, with no significant inter-study het-erogeneity existed. Many studies have con-

Figure 9. Begg’s funnel plot for assessing publication bias of miR-21 expression and overall survival (univariate) in HCC.

the lack of consistency in the expression direction or the differ-entially expressed miRNAs in tis-sues compared with normal tis-sues still challenged us. To deal with the obscurities of unavailable raw data and various investigat-ing methods, according to the reports of Chan and Griffith [28, 29], a vote-strategy method was applied to ensure the consistency among studies with different microarray platforms. Therefore, we collected information of all mi- RNA microarrays on HCC by searching literatures in the NCBI GEO database and eventually 43 studies with miRNA profiling data were recruited in this systematic



firmed that miR-21 overexpression might enhance HCC migration and invasion through certain mechanisms, resulting in poor survival [70-74]. They elucidated that miR-21 simultane-ously regulated multiple programs that promote cell proliferation, apoptosis or tumor invasive-ness by targeting PTEN, PDCD4, and RECK in HCC, and the results of Hu et al. [75] suggested that miR-21 participated in HCC development through promoting cell proliferation by post-transcriptionally downregulating HEPN1 expres-sion. Besides, Xu et al. [76] pointed out that miR-21 could also directly target MAP2K3 and inhibit its expression during HCC tumorigene-sis. Therefore, high tumoral miR-21 expression in HCC patients was significantly correlated with tumor progression and could be a high risk factor for poor HCC prognosis [32].

MiR-221, encoded on human chromosome X, is overexpressed in various aggressive carcino-mas [77-79]. In this meta-analysis, we observed that higher tumoral miR-221, with an average of 5.54-fold change than normal tissues, was significantly associated with poor OS in HCC. Furthermore, Yang et al. [80] found it was more efficient to detect miR-221 in serum/plasma samples than in HCC tissue samples when miR-221 was used for predicting the prognosis. The findings reminded us that circulating miR-221 is more qualified as a biomarker in HCC. It was reported that elevated miR-221 promoted growth of HCC cells by positively regulating the expression of two cyclin-dependent kinase inhibitors (CDKIs): CDKN1B/p27 and CDKN1C/p57, and it inhibited cell apoptosis by suppress-ing the expression of BMF [46, 81]. Hence, miR-221, considered as a tumor promoter, was likely to be a potential biomarker for HCC patients with poor prognosis and it could also assess therapeutic efficacy in the future.

MiR-222 was identified as one of the top five most upregulated miRNAs in our rank study. The expression of miR-222 in 16 miRNA profil-ing studies covering 632 HCC patients in HCC tissues was 5.54 times higher than that in adja-cent non-tumor tissues. As for prognosis, miR-222 was proved to be an independent risk fac-tor leading to poor outcomes for HCC patients, since the univariate pooled HR of 2.992 was calculated through combining OS data in five studies. This finding attracted our attention to the complex molecular mechanisms in liver car-

cinogenesis. As reported by Wong et al. [49], increased miR-222 expression was related to advanced stage of HCC and poor DFS, and then it was confirmed to accelerate HCC cell motility via enhancing PI3K/AKT signaling, the same oncogenic pathway identified by Liu et al. [47], too. Additionally, Zhang et al. [82] indicated that upregulation of miR-222 may cause the down-regulation of GNAI3, a member of the Gi group involved in regulating cellular proliferation, migration, invasion and apoptosis, and this aberrant regulation promoted invasion and metastasis of HCC. Since upregulation of miR-222 can promote HCC cells proliferation and migration, it is expected to be a potential bio-marker in HCC development and may provide new therapeutic targets.

MiR-25, belonging to the miR-106b~25 cluster including miR-106b, miR-93 and miR-25, is located within intron 13 of the minichromo-some maintenance protein 7 (MCM7) gene on chromosome 7q22.1 [83, 84]. In addition, miR-25 was one of the most frequently miRNAs mentioned in HCC tissues, with a fold change of 2.38 times higher than corresponding non-tumor tissues. Recently, some researchers [85-87] have found the overexpression of miR-25 was associated with the prognosis of various tumors, such as ovary, gastric and prostate cancers. Similarly, Li et al. [52] demonstrated that the miR-106b~25 cluster possessed onco-genic properties and upregulated miR-25 con-tributed to liver carcinogenesis in HCC tissues and cell lines. Despite the fact that Su et al. [51] had worked for the clinical relevance of miR-25 and its expression on the prognosis of HCC, we combined the study with Sadeghian et al. [53] in the present meta-analysis to pool the survival data and explore the feasibility of miR-25 as a potential prognostic biomarker for HCC. The result that pooled HR>1 in both univariate and multivariate analysis suggested that over-expression of miR-25 in HCC correlated with poor survival after HCC resection, which was considered to be with certain mechanisms in tumor metastasis. Activated by the WNT/β-catenin signaling pathway, overexpressed miR-25 could play its pro-metastatic role through inhibiting the Rho GDP dissociation inhibitor alpha directly, namely RhoGDI1, whose down-regulation stimulates expression of Snail, thereby enhancing epithelial-mesenchymal transition (EMT). Once EMT was activated, miR-



25 could promote cell proliferation, tumor pro-gression and invasion. Therefore MiR-25/RhoGDI1 axis promises to be a probable thera-peutic target for the treatment of HCC [88]. The results of our meta-analysis combining two prognostic studies, with no significant hetero-geneity, were consistent with the researches above. However, further experimental valida-tion in large and homogenous cohorts of HCC patients, adjusting for multiple clinic pathologic parameters, are required urgently to confirm its prognostic role in HCC.

MiR-199a-3p was downregulated in HCC [89] and it was most frequently mentioned in the microarray studies included. The expression of miR-199a-3p in HCC tissues was 0.37 times lower than that in non-tumorous tissues in 14 studies with 713 samples. One study present-ed that miR-199a-3p could be a novel serum diagnostic biomarker for HCC [90]. Never- theless, according to the meta-analysis pub-lished, only a few studies pointed out that low miR-199a-3p expression was related with the poor prognosis in HCC [55, 91, 92]. In order to strengthen the reliability of the relevance between low miR-199a-3p level and HCC patients’ survival, twelve studies were included and sufficient survival data were collected to calculate the pooled HRs. The combined results consistently suggested that miR-199a-3p was downregulated with shorter overall sur-vival for HCC. Up to now, a considerable portion of studies have demonstrated that downregu-lated miR-199a-3p might participate in direct or indirect cell proliferation, apoptosis, tumor migration and invasion in HCC through differ-ent mechanisms, which may predict poor prog-nosis in HCC patients [89, 93-95]. Shatseva et al. [94] confirmed that miR-199a-3p regulated cell proliferation and survival by targeting cave-olin-2 in HCC. Hou et al. [55] pointed out that miR-199a-3p could target tumor-promoting PAK4 to suppress HCC growth through inhibit-ing the PAK4/RAF/MEK/ERK pathway, and thus the decrement of miR-199a-3p was signifi-cantly correlated with poor survival for HCC patients. Besides, Song et al. [61] proposed that FZD7, identified as a functional target of miR-199a-3p, participated in cell proliferation and cell cycle regulation of HCC. To sum up, due to under-expression of miR-199a-3p that was significantly correlated with tumor progres-sion in HCC patients, miR-199a-3p could be

considered as a candidate biomarker indicating poor prognosis of HCC.

MiR-214 was known as a tumor suppressor of HCC [96] and its low expression in human HCC was associated with poor prognosis based on our meta-analysis above. Along with the results of fourteen miRNA profiling studies included, downregulation of miR-214 was proved in HCC according to several other studies recently [36, 97-101]. Xia et al. [59] demonstrated that low expression of miR-214 in HCC tissues was con-nected with the cell invasion, stem-like traits and early recurrence of HCC, whilst two poten-tial downstream targets of miR-214, enhancer of zeste homologue 2 (EZH2) and β-catenin (CTNNB1) were noticeably identified. Besides, Shih et al. [57] had found that hepatoma-derived growth factor (HDGF) was also a target gene of miR-214, which can activate vascular endothelial cells. While Wang et al. [58] observed that the FGFR-1 can be significantly upregulated by miR-214 and thereby promoted the aggressiveness of HCC expression. More- over, Wang et al. [91] also found that low expression of miR-214 in HCC cells correlated with inhibition of E2F2, CDK3 and CDK6. As a result, miR-214 could significantly disturb the cell-cycle and promote proliferation of HCC cells, leading to tumor angiogenesis invasion and early recurrence in HCC patients. In short, downregulation of miR-214 may provide a new therapeutic approach for cancer treatment.

To sum up, this systematic review had a well-defined search strategy, rigorous selection of studies in literature, abundant cases of HCC patients and massive sample size. The features mentioned above contributed to the improved reliability and increased statistical conviction in our meta-analysis. Moreover, the heterogeneity was decreased by ensuring the consistency of the sample types and detecting methods used in studies, and the most up- or down-regulated miRNAs in HCC in miRNA profiling studies glob-ally were then ranked and selected, so as to perform relevant survival analysis and thus we could explore the top six dysregulated miRNAs and their specific relationships with HCC patients’ prognosis independently.

However, some limitations of this study should be addressed. Above all, the lack of literatures related prognosis with enough survival data such as DFS and PFS might weaken the statisti-



cal persuasiveness of the final results to some extent. Secondly, subgroup analysis was not performed because of the lack of certain clini-copathological characteristics such as histo-logical grade, TNM stages, tumor size and the number of tumor nodes from the prognostic studies. Although the inter-study heterogeneity of some results was proved insignificant, the specific clinical characteristic that these miR-NAs related with remained unexplored. There- fore, more clinical studies with detailed clinico-pathological parameters were needed to inves-tigate the detailed and accurate functions of dysregulated miRNAs in HCC.

Conclusions

Overall, the top six upregulated miRNAs consis-tently reported in HCC tissues were miR-222, miR-221, miR-224, miR-21, miR-25 and miR-18a, all of which lead to dismal prognosis for HCC except miR-18a and miR-224. On the con-trary, the top six downregulated miRNAs were miR-199a-5p, miR-99a, miR-195, miR-130a, miR-199a-3p, and miR-214, among which miR-199a-3p and miR-214 emerged as potential predictors of favorable prognosis for HCC patients. These miRNAs can be potential candi-dates for HCC prognosis and therapy.

Acknowledgements

This study was supported by the Funds of Natural Science Foundation of Guangxi, China (2015GXNSFBA139157). The funders had no role in study design, data collection and analy-sis, decision to publish, or preparation of the manuscript.

Disclosure of conflict of interest

None.

Address correspondence to: Dr. Li-Hua Yang, Depart- ment of Medical Oncology, First Affiliated Hospital of Guangxi Medical University, 6 Shuangyong Road, Nanning 530021, Guangxi Zhuang Autonomous Region, P. R. China. E-mail: [email protected]

References

[1] Zhang J, Yang Y, Yang T, Yuan S, Wang R, Pan Z, Yang Y, Huang G, Gu F, Jiang B, Lin C and Zhou W. Double-negative feedback loop between microRNA-422a and forkhead box (FOX)G1/Q1/E1 regulates hepatocellular carcinoma tu-

mor growth and metastasis. Hepatology 2015; 61: 561-573.

[2] Lin CL and Kao JH. Risk stratification for hepa-titis B virus related hepatocellular carcinoma. J Gastroenterol Hepatol 2013; 28: 10-17.

[3] Ayub A, Ashfaq UA and Haque A. HBV induced HCC: major risk factors from genetic to molec-ular level. Biomed Res Int 2013; 2013: 810461.

[4] Robotin MC and George J. Community-based hepatitis B screening: what works? Hepatol Int 2014; 8: 478-492.

[5] Liu D, Liu A, Peng J, Hu Y and Feng X. Case analysis of complete remission of advanced hepatocellular carcinoma achieved with sorafenib. Eur J Med Res 2015; 20: 12.

[6] Cao G, Li X, Qin C and Li J. Prognostic value of VEGF in hepatocellular carcinoma patients treated with sorafenib: a meta-analysis. Med Sci Monit 2015; 21: 3144-3151.

[7] Li Q, Xing H, Liu D and Li H. Negative impact of low body mass index on liver cirrhosis patients with hepatocellular carcinoma. World J Surg Oncol 2015; 13: 294.

[8] Luo Y, Ren F, Liu Y, Shi Z, Tan Z, Xiong H, Dang Y and Chen G. Clinicopathological and prog-nostic significance of high Ki-67 labeling index in hepatocellular carcinoma patients: a meta-analysis. Int J Clin Exp Med 2015; 8: 10235-10247.

[9] Park KU, Seo YS, Lee YH, Park J, Hwang I, Kang KJ, Nam J, Kim SW and Kim JY. Altered microR-NA expression profile in hepatitis B virus-relat-ed hepatocellular carcinoma. Gene 2015; 573: 278-284.

[10] Huang Z, Huang L, Shen S, Li J, Lu H, Mo W, Dang Y, Luo D, Chen G and Feng Z. Sp1 cooper-ates with Sp3 to upregulate MALAT1 expres-sion in human hepatocellular carcinoma. On-col Rep 2015; 34: 2403-2412.

[11] Li J, Liang L, Liu Y, Luo Y, Liang X, Luo D, Feng Z, Dang Y, Yang L and Chen G. Clinicopatho-logical significance of STAT4 in hepatocellular carcinoma and its effect on cell growth and apoptosis. Onco Targets Ther 2016; 9: 1721-1734.

[12] Ren WH, Li YW, Li R, Feng HB, Wu JL and Wang HR. P15 gene methylation in hepatocellular carcinomas: a systematic review and meta-analysis. Int J Clin Exp Med 2015; 8: 4762-4768.

[13] Zhong JH, Zhong QL, Li LQ and Li H. Adjuvant and chemopreventive therapies for resectable hepatocellular carcinoma: a literature review. Tumour Biol 2014; 35: 9459-9468.

[14] Li J, Shi W, Gao Y, Yang B, Jing X, Shan S, Wang Y and Du Z. Analysis of microRNA expression profiles in human hepatitis B virus-related he-patocellular carcinoma. Clin Lab 2013; 59: 1009-1015.

mailto:[email protected]



[15] Li MX, Bi XY, Li ZY, Huang Z, Han Y, Zhou JG, Zhao JJ, Zhang YF, Zhao H and Cai JQ. Prognos-tic role of glasgow prognostic score in patients with hepatocellular carcinoma: a systematic review and meta-analysis. Medicine (Balti-more) 2015; 94: e2133.

[16] Wang L, Yao M, Dong Z, Zhang Y and Yao D. Circulating specific biomarkers in diagnosis of hepatocellular carcinoma and its metastasis monitoring. Tumour Biol 2014; 35: 9-20.

[17] Zhang L, Han J, Wu H, Liang X, Zhang J, Li J, Xie L, Xie Y, Sheng X and Yu J. The association of HMGB1 expression with clinicopathological significance and prognosis in hepatocellular carcinoma: a meta-analysis and literature re-view. PLoS One 2014; 9: e110626.

[18] Wei R, Huang GL, Zhang MY, Li BK, Zhang HZ, Shi M, Chen XQ, Huang L, Zhou QM, Jia WH, Zheng XF, Yuan YF and Wang HY. Clinical sig-nificance and prognostic value of microRNA expression signatures in hepatocellular carci-noma. Clin Cancer Res 2013; 19: 4780-4791.

[19] Murakami Y, Yasuda T, Saigo K, Urashima T, Toyoda H, Okanoue T and Shimotohno K. Com-prehensive analysis of microRNA expression patterns in hepatocellular carcinoma and non-tumorous tissues. Oncogene 2006; 25: 2537-2545.

[20] Sun J, Lu H, Wang X and Jin H. MicroRNAs in hepatocellular carcinoma: regulation, function, and clinical implications. ScientificWorldJourn-al 2013; 2013: 924206.

[21] Di Leva G, Briskin D and Croce CM. MicroRNA in cancer: new hopes for antineoplastic che-motherapy. Ups J Med Sci 2012; 117: 202-216.

[22] Gan TQ, Tang RX, He RQ, Dang YW, Xie Y and Chen G. Upregulated miR-1269 in hepatocel-lular carcinoma and its clinical significance. Int J Clin Exp Med 2015; 8: 714-721.

[23] He R, Yang L, Lin X, Chen X, Lin X, Wei F, Liang X, Luo Y, Wu Y, Gan T, Dang Y and Chen G. MiR-30a-5p suppresses cell growth and enhances apoptosis of hepatocellular carcinoma cells via targeting AEG-1. Int J Clin Exp Pathol 2015; 8: 15632-15641.

[24] Liu Y, Ren F, Luo Y, Rong M, Chen G and Dang Y. Down-regulation of miR-193a-3p dictates deterioration of HCC: a clinical real-time qRT-PCR study. Med Sci Monit 2015; 21: 2352-2360.

[25] Liu Y, Ren F, Rong M, Luo Y, Dang Y and Chen G. Association between underexpression of mi-crorna-203 and clinicopathological signifi-cance in hepatocellular carcinoma tissues. Cancer Cell Int 2015; 15: 62.

[26] Xue TM, Tao LD, Zhang M, Zhang J, Liu X, Chen GF, Zhu YJ and Zhang PJ. Clinicopathological significance of microRNA-20b expression in hepatocellular carcinoma and regulation of

HIF-1alpha and VEGF effect on cell biological behaviour. Dis Markers 2015; 2015: 325176.

[27] Boeri M, Verri C, Conte D, Roz L, Modena P, Facchinetti F, Calabro E, Croce CM, Pastorino U and Sozzi G. MicroRNA signatures in tissues and plasma predict development and progno-sis of computed tomography detected lung cancer. Proc Natl Acad Sci U S A 2011; 108: 3713-3718.

[28] Chan SK, Griffith OL, Tai IT and Jones SJ. Meta-analysis of colorectal cancer gene expression profiling studies identifies consistently report-ed candidate biomarkers. Cancer Epidemiol Biomarkers Prev 2008; 17: 543-552.

[29] Griffith OL, Melck A, Jones SJ and Wiseman SM. Meta-analysis and meta-review of thyroid cancer gene expression profiling studies iden-tifies important diagnostic biomarkers. J Clin Oncol 2006; 24: 5043-5051.

[30] Tierney JF, Stewart LA, Ghersi D, Burdett S and Sydes MR. Practical methods for incorporating summary time-to-event data into meta-analy-sis. Trials 2007; 8: 16.

[31] Hu L, Ye H, Huang G, Luo F, Liu Y, Liu Y, Yang X, Shen J, Liu Q and Zhang J. Long noncoding RNA GAS5 suppresses the migration and inva-sion of hepatocellular carcinoma cells via miR-21. Tumour Biol 2016; 37: 2691-2702.

[32] Huang CS, Yu W, Cui H, Wang YJ, Zhang L, Han F and Huang T. Increased expression of miR-21 predicts poor prognosis in patients with he-patocellular carcinoma. Int J Clin Exp Pathol 2015; 8: 7234-7238.

[33] Shi KQ, Lin Z, Chen XJ, Song M, Wang YQ, Cai YJ, Yang NB, Zheng MH, Dong JZ, Zhang L and Chen YP. Hepatocellular carcinoma associated microRNA expression signature: integrated bioinformatics analysis, experimental valida-tion and clinical significance. Oncotarget 2015; 6: 25093-25108.

[34] Wang X, Zhang J, Zhou L, Lu P, Zheng ZG, Sun W, Wang JL, Yang XS, Li XL, Xia N, Zhang N and Dou KF. Significance of serum microRNA-21 in diagnosis of hepatocellular carcinoma (HCC): clinical analyses of patients and an HCC rat model. Int J Clin Exp Pathol 2015; 8: 1466-1478.

[35] Liu M, Liu J, Wang L, Wu H, Zhou C, Zhu H, Xu N and Xie Y. Association of serum microRNA expression in hepatocellular carcinomas treat-ed with transarterial chemoembolization and patient survival. PLoS One 2014; 9: e109347.

[36] Gyongyosi B, Vegh E, Jaray B, Szekely E, Fas-san M, Bodoky G, Schaff Z and Kiss A. Pre-treatment microRNA level and outcome in sorafenib-treated hepatocellular carcinoma. J Histochem Cytochem 2014; 62: 547-555.

[37] Wang WY, Zhang HF, Wang L, Ma YP, Gao F, Zhang SJ and Wang LC. miR-21 expression pre-dicts prognosis in hepatocellular carcinoma.



Clin Res Hepatol Gastroenterol 2014; 38: 715-719.

[38] Karakatsanis A, Papaconstantinou I, Gazouli M, Lyberopoulou A, Polymeneas G and Voros D. Expression of microRNAs, miR-21, miR-31, miR-122, miR-145, miR-146a, miR-200c, miR-221, miR-222, and miR-223 in patients with hepatocellular carcinoma or intrahepatic chol-angiocarcinoma and its prognostic signifi-cance. Mol Carcinog 2013; 52: 297-303.

[39] Tomimaru Y, Eguchi H, Nagano H, Wada H, To-mokuni A, Kobayashi S, Marubashi S, Takeda Y, Tanemura M, Umeshita K, Doki Y and Mori M. MicroRNA-21 induces resistance to the an-ti-tumour effect of interferon-alpha/5-fluoro-uracil in hepatocellular carcinoma cells. Br J Cancer 2010; 103: 1617-1626.

[40] Yu L. Value of microRNAs in early diagnosis in hepatocellular carcinoma and the mecha-nisms of their effects on prognosis (PhD Thesis in Chinese). Shanghai: Fudan University; 2012.

[41] He X, Li J, Guo W, Liu W, Yu J, Song W, Dong L, Wang F, Yu S, Zheng Y, Chen S, Kong Y and Liu C. Targeting the microRNA-21/AP1 axis by 5-fluorouracil and pirarubicin in human hepa-tocellular carcinoma. Oncotarget 2015; 6: 2302-2314.

[42] Jiang J, Gusev Y, Aderca I, Mettler TA, Nagorney DM, Brackett DJ, Roberts LR and Schmittgen TD. Association of microRNA expression in he-patocellular carcinomas with hepatitis infec-tion, cirrhosis, and patient survival. Clin Can-cer Res 2008; 14: 419-427.

[43] Li J, Wang Y, Yu W, Chen J and Luo J. Expres-sion of serum miR-221 in human hepatocellu-lar carcinoma and its prognostic significance. Biochem Biophys Res Commun 2011; 406: 70-73.

[44] Bae HJ, Jung KH, Eun JW, Shen Q, Kim HS, Park SJ, Shin WC, Yang HD, Park WS, Lee JY and Nam SW. MicroRNA-221 governs tumor suppressor HDAC6 to potentiate malignant progression of liver cancer. J Hepatol 2015; 63: 408-419.

[45] Rong M, Chen G and Dang Y. Increased miR-221 expression in hepatocellular carcinoma tissues and its role in enhancing cell growth and inhibiting apoptosis in vitro. BMC Cancer 2013; 13: 21.

[46] Gramantieri L, Fornari F, Ferracin M, Veronese A, Sabbioni S, Calin GA, Grazi GL, Croce CM, Bolondi L and Negrini M. MicroRNA-221 tar-gets Bmf in hepatocellular carcinoma and cor-relates with tumor multifocality. Clin Cancer Res 2009; 15: 5073-5081.

[47] Liu K, Liu S, Zhang W, Ji B, Wang Y and Liu Y. miR222 regulates sorafenib resistance and enhance tumorigenicity in hepatocellular carci-noma. Int J Oncol 2014; 45: 1537-1546.

[48] Zhan MX, Li Y, Hu BS, Shao PJ, Meng QW, He X, Huang JW and Lu LG. [Expression of serum mi-croRNAs (miR-222, miR-181, miR-216) in hu-man hepatocellular carcinoma and its clinical significance]. Zhonghua Yi Xue Za Zhi 2013; 93: 1830-1832.

[49] Wong QW, Ching AK, Chan AW, Choy KW, To KF, Lai PB and Wong N. MiR-222 overexpression confers cell migratory advantages in hepato-cellular carcinoma through enhancing AKT sig-naling. Clin Cancer Res 2010; 16: 867-875.

[50] Xu F, Li S, Yang F, Wang Y, Xiang X, Qian NS and Li Y. Clinical significance of serum miR-29a-5p and miR-222 expression in the diagnosis and prognostic evaluation of hepatocellular carci-noma (in Chinese). Modern Oncology 2013; 21: 2513-2516.

[51] Su ZX, Zhao J, Rong ZH, Geng WM, Wu YG and Qin CK. Upregulation of microRNA-25 associ-ates with prognosis in hepatocellular carcino-ma. Diagn Pathol 2014; 9: 47.

[52] Li Y, Tan W, Neo TW, Aung MO, Wasser S, Lim SG and Tan TM. Role of the miR-106b-25 mi-croRNA cluster in hepatocellular carcinoma. Cancer Sci 2009; 100: 1234-1242.

[53] Sadeghian Y, Kamyabi-Moghaddam Z, Nodu-shan SM, Khoshbakht S, Pedram B, Yahaghi E, Mokarizadeh A and Mohebbi M. Profiles of tis-sue microRNAs; miR-148b and miR-25 serve as potential prognostic biomarkers for hepato-cellular carcinoma. Tumour Biol 2015; [Epub ahead of print].

[54] Wang P, Chen S, Fang H, Wu X, Chen D, Peng L, Gao Z and Xie C. miR-214/199a/199a* clus-ter levels predict poor survival in hepatocellu-lar carcinoma through interference with cell-cycle regulators. Oncotarget 2016; 7: 929-945.

[55] Hou J, Lin L, Zhou W, Wang Z, Ding G, Dong Q, Qin L, Wu X, Zheng Y, Yang Y, Tian W, Zhang Q, Wang C, Zhang Q, Zhuang SM, Zheng L, Liang A, Tao W and Cao X. Identification of miR-Nomes in human liver and hepatocellular car-cinoma reveals miR-199a/b-3p as therapeutic target for hepatocellular carcinoma. Cancer Cell 2011; 19: 232-243.

[56] Liu L, Li SQ, Mi XG and Fan YQ. Relationship of clinical outcomes and expression of microR-NA-199a /b-3pin HCC (in Chinese). Chinese Journal of Immunology 2015; 31: 806-808.

[57] Shih TC, Tien YJ, Wen CJ, Yeh TS, Yu MC, Huang CH, Lee YS, Yen TC and Hsieh SY. MicroR-NA-214 downregulation contributes to tumor angiogenesis by inducing secretion of the hep-atoma-derived growth factor in human hepa-toma. J Hepatol 2012; 57: 584-591.

[58] Wang J, Li J, Wang X, Zheng C and Ma W. Down-regulation of microRNA-214 and overexpres-sion of FGFR-1 contribute to hepatocellular carcinoma metastasis. Biochem Biophys Res Commun 2013; 439: 47-53.



[59] Xia H, Ooi LL and Hui KM. MiR-214 targets be-ta-catenin pathway to suppress invasion, stem-like traits and recurrence of human he-patocellular carcinoma.

[60] Guo W, Qiu Z, Wang Z, Wang Q, Tan N, Chen T, Chen Z, Huang S, Gu J, Li J, Yao M, Zhao Y and He X. MiR-199a-5p is negatively associated with malignancies and regulates glycolysis and lactate production by targeting hexoki-nase 2 in liver cancer. Hepatology 2015; 62: 1132-1144.

[61] Song J, Gao L, Yang G, Tang S, Xie H, Wang Y, Wang J, Zhang Y, Jin J, Gou Y, Yang Z, Chen Z, Wu K, Liu J and Fan D. MiR-199a regulates cell proliferation and survival by targeting FZD7. PLoS One 2014; 9: e110074.

[62] Zhang J, Chong CC, Chen GG and Lai PB. A seven-microRNA expression signature predicts survival in hepatocellular carcinoma. PLoS One 2015; 10: e0128628.

[63] Hamilton MP, Rajapakshe K, Hartig SM, Reva B, McLellan MD, Kandoth C, Ding L, Zack TI, Gunaratne PH, Wheeler DA, Coarfa C and Mc-Guire SE. Identification of a pan-cancer onco-genic microRNA superfamily anchored by a central core seed motif. Nat Commun 2013; 4: 2730.

[64] Huang JT, Liu SM, Ma H, Yang Y, Zhang X, Sun H, Zhang X, Xu J and Wang J. Systematic re-view and meta-analysis: circulating mirnas for diagnosis of hepatocellular carcinoma. J Cell Physiol 2016; 231: 328-335.

[65] Liao Q, Han P, Huang Y, Wu Z, Chen Q, Li S, Ye J and Wu X. Potential role of circulating microR-NA-21 for hepatocellular carcinoma diagnosis: a meta-analysis. PLoS One 2015; 10: e0130677.

[66] Li G, Shen Q, Li C, Li D, Chen J and He M. Iden-tification of circulating MicroRNAs as novel po-tential biomarkers for hepatocellular carcino-ma detection: a systematic review and meta- analysis. Clin Transl Oncol 2015; 17: 684-693.

[67] Yan SR, Liu ZJ, Yu S and Bao YX. Investigation of the value of miR-21 in the diagnosis of early stage HCC and its prognosis: a meta-analysis. Genet Mol Res 2015; 14: 11573-11586.

[68] Wang Y, Gao X, Wei F, Zhang X, Yu J, Zhao H, Sun Q, Yan F, Yan C, Li H and Ren X. Diagnostic and prognostic value of circulating miR-21 for cancer: a systematic review and meta-analy-sis. Gene 2014; 533: 389-397.

[69] Zhou X, Wang X, Huang Z, Wang J, Zhu W, Shu Y and Liu P. Prognostic value of miR-21 in vari-ous cancers: an updating meta-analysis. PLoS One 2014; 9: e102413.

[70] Najafi Z, Sharifi M and Javadi G. Degradation of miR-21 induces apoptosis and inhibits cell proliferation in human hepatocellular carcino-ma. Cancer Gene Ther 2015; 22: 530-535.

[71] Zhou L, Yang ZX, Song WJ, Li QJ, Yang F, Wang DS, Zhang N and Dou KF. MicroRNA-21 regu-lates the migration and invasion of a stem-like population in hepatocellular carcinoma. Int J Oncol 2013; 43: 661-669.

[72] Bao L, Yan Y, Xu C, Ji W, Shen S, Xu G, Zeng Y, Sun B, Qian H, Chen L, Wu M, Su C and Chen J. MicroRNA-21 suppresses PTEN and hSulf-1 ex-pression and promotes hepatocellular carci-noma progression through AKT/ERK pathways. Cancer Lett 2013; 337: 226-236.

[73] Zhu Q, Wang Z, Hu Y, Li J, Li X, Zhou L and Huang Y. miR-21 promotes migration and inva-sion by the miR-21-PDCD4-AP-1 feedback loop in human hepatocellular carcinoma. Oncol Rep 2012; 27: 1660-1668.

[74] Liu C, Yu J, Yu S, Lavker RM, Cai L, Liu W, Yang K, He X and Chen S. MicroRNA-21 acts as an oncomir through multiple targets in human he-patocellular carcinoma. J Hepatol 2010; 53: 98-107.

[75] Hu S, Tao R, Wang S, Wang C, Zhao X, Zhao H, Li L, Zhu S, He Y, Jiang X and Gao Y. MicroR-NA-21 promotes cell proliferation in human hepatocellular carcinoma partly by targeting HEPN1. Tumour Biol 2015; 36: 5467-5472.

[76] Xu G, Zhang Y, Wei J, Jia W, Ge Z, Zhang Z and Liu X. MicroRNA-21 promotes hepatocellular carcinoma HepG2 cell proliferation through re-pression of mitogen-activated protein kinase-kinase 3. BMC Cancer 2013; 13: 469.

[77] Galardi S, Mercatelli N, Giorda E, Massalini S, Frajese GV, Ciafre SA and Farace MG. miR-221 and miR-222 expression affects the prolifera-tion potential of human prostate carcinoma cell lines by targeting p27Kip1. J Biol Chem 2007; 282: 23716-23724.

[78] Waters PS, McDermott AM, Wall D, Heneghan HM, Miller N, Newell J, Kerin MJ and Dwyer RM. Relationship between circulating and tissue microRNAs in a murine model of breast can-cer. PLoS One 2012; 7: e50459.

[79] Wang Z, Zhang H, He L, Dong W, Li J, Shan Z and Teng W. Association between the expres-sion of four upregulated miRNAs and extrathy-roidal invasion in papillary thyroid carcinoma. Onco Targets Ther 2013; 6: 281-287.

[80] Yang J, Zhang JY, Chen J, Xu Y, Song NH and Yin CJ. Prognostic role of microRNA-221 in various human malignant neoplasms: a meta-analysis of 20 related studies. PLoS One 2014; 9: e87606.

[81] Fornari F, Gramantieri L, Ferracin M, Veronese A, Sabbioni S, Calin GA, Grazi GL, Giovannini C, Croce CM, Bolondi L and Negrini M. MiR-221 controls CDKN1C/p57 and CDKN1B/p27 ex-pression in human hepatocellular carcinoma. Oncogene 2008; 27: 5651-5661.

[82] Zhang Y, Yao J, Huan L, Lian J, Bao C, Li Y, Ge C, Li J, Yao M, Liang L and He X. GNAI3 inhibits



tumor cell migration and invasion and is post-transcriptionally regulated by miR-222 in he-patocellular carcinoma. Cancer Lett 2015; 356: 978-984.

[83] Petrocca F, Vecchione A and Croce CM. Emerg-ing role of miR-106b-25/miR-17-92 clusters in the control of transforming growth factor beta signaling. Cancer Res 2008; 68: 8191-8194.

[84] Savita U and Karunagaran D. MicroRNA-106b-25 cluster targets beta-TRCP2, increas-es the expression of Snail and enhances cell migration and invasion in H1299 (non small cell lung cancer) cells. Biochem Biophys Res Commun 2013; 434: 841-847.

[85] Kim YK, Yu J, Han TS, Park SY, Namkoong B, Kim DH, Hur K, Yoo MW, Lee HJ, Yang HK and Kim VN. Functional links between clustered microRNAs: suppression of cell-cycle inhibitors by microRNA clusters in gastric cancer. Nucleic Acids Res 2009; 37: 1672-1681.

[86] Zhang H, Zuo Z, Lu X, Wang L, Wang H and Zhu Z. MiR-25 regulates apoptosis by targeting Bim in human ovarian cancer. Oncol Rep 2012; 27: 594-598.

[87] Poliseno L, Salmena L, Riccardi L, Fornari A, Song MS, Hobbs RM, Sportoletti P, Varmeh S, Egia A, Fedele G, Rameh L, Loda M and Pan-dolfi PP. Identification of the miR-106b~25 mi-croRNA cluster as a proto-oncogenic PTEN-tar-geting intron that cooperates with its host gene MCM7 in transformation. Sci Signal

[88] Wang C, Wang X, Su Z, Fei H, Liu X and Pan Q. miR-25 promotes hepatocellular carcinoma cell growth, migration and invasion by inhibit-ing RhoGDI1. Oncotarget 2015; 6: 36231-36244.

[89] Gu S and Chan WY. Flexible and versatile as a chameleon-sophisticated functions of microR-NA-199a. Int J Mol Sci 2012; 13: 8449-8466.

[90] Yin J, Hou P, Wu Z, Wang T and Nie Y. Circulat-ing miR-375 and miR-199a-3p as potential biomarkers for the diagnosis of hepatocellular carcinoma. Tumour Biol 2015; 36: 4501-4507.

[91] Wang P, Chen S, Fang H, Wu X, Chen D, Peng L, Gao ZL and Xie C. miR-214/199a/199a* clus-ter levels predict poor survival in hepatocellu-lar carcinoma through interference with cell-cycle regulators. Oncotarget 2016; 7: 929-945.

[92] Fornari F, Milazzo M, Chieco P, Negrini M, Calin GA, Grazi GL, Pollutri D, Croce CM, Bolondi L and Gramantieri L. MiR-199a-3p regulates mTOR and c-Met to influence the doxorubicin sensitivity of human hepatocarcinoma cells. Cancer Res 2010; 70: 5184-5193.

[93] Duan Q, Wang X, Gong W, Ni L, Chen C, He X, Chen F, Yang L, Wang P and Wang DW. ER stress negatively modulates the expression of the miR-199a/214 cluster to regulates tumor survival and progression in human hepatocel-lular cancer. PLoS One 2012; 7: e31518.

[94] Shatseva T, Lee DY, Deng Z and Yang BB. Mi-croRNA miR-199a-3p regulates cell prolifera-tion and survival by targeting caveolin-2. J Cell Sci 2011; 124: 2826-2836.

[95] Zhang LF, Lou JT, Lu MH, Gao C, Zhao S, Li B, Liang S, Li Y, Li D and Liu MF. Suppression of miR-199a maturation by HuR is crucial for hy-poxia-induced glycolytic switch in hepatocellu-lar carcinoma. EMBO J 2015; 34: 2671-2685.

[96] Zhang LL, Guo YJ, Zhao CN and Gao JY. Effects and mechanism of miR-214 on hepatocellular carcinoma. Asian Pac J Trop Med 2015; 8: 392-398.

[97] Penna E, Orso F and Taverna D. miR-214 as a key hub that controls cancer networks: small player, multiple functions. J Invest Dermatol 2015; 135: 960-969.

[98] Liu WH, Chen CH, Yeh KH, Li CL, Wu YJ, Chen DS, Chen PJ and Yeh SH. ADAR2-mediated ed-iting of miR-214 and miR-122 precursor and antisense RNA transcripts in liver cancers. PLoS One 2013; 8: e81922.

[99] Wang X, Chen J, Li F, Lin Y, Zhang X, Lv Z and Jiang J. MiR-214 inhibits cell growth in hepato-cellular carcinoma through suppression of be-ta-catenin. Biochem Biophys Res Commun 2012; 428: 525-531.

[100] Magrelli A, Azzalin G, Salvatore M, Viganotti M, Tosto F, Colombo T, Devito R, Di Masi A, Antoc-cia A, Lorenzetti S, Maranghi F, Mantovani A, Tanzarella C, Macino G and Taruscio D. Altered microRNA expression patterns in hepatoblas-toma patients. Transl Oncol 2009; 2: 157-163.

[101] Tan L, Sim LK, Ng L, Toh HJ and Low JA. Ad-vance care planning: the attitudes and views of a group of catholic nuns in Singapore. Am J Hosp Palliat Care 2015; 21: 3144-3151.


1

References* in Tables 1 and 2

[1] Zhang J, Yang Y, Yang T, Yuan S, Wang R, Pan Z, Yang Y, Huang G, Gu F, Jiang B, Lin C and Zhou W. Double-negative feedback loop between microRNA-422a and forkhead box (FOX)G1/Q1/E1 regulates hepatocellular carcinoma tumor growth and metastasis. Hepatology 2015; 61: 561-573.

[2] Zhang J, Chong CC, Chen GG and Lai PB. A seven-microRNA expression signature predicts survival in hepato-cellular carcinoma. PLoS One 2015; 10: e0128628.

[3] Park KU, Seo YS, Lee YH, Park J, Hwang I, Kang KJ, Nam J, Kim SW and Kim JY. Altered microRNA expression profile in hepatitis B virus-related hepatocellular carcinoma. Gene 2015; 573: 278-284.

[4] Li J, Shi W, Gao Y, Yang B, Jing X, Shan S, Wang Y and Du Z. Analysis of microRNA expression profiles in human hepatitis B virus-related hepatocellular carcinoma. Clin Lab 2013; 59: 1009-1015.

[5] Wei R, Huang GL, Zhang MY, Li BK, Zhang HZ, Shi M, Chen XQ, Huang L, Zhou QM, Jia WH, Zheng XF, Yuan YF and Wang HY. Clinical significance and prognostic value of microRNA expression signatures in hepatocellular carcinoma. Clin Cancer Res 2013; 19: 4780-4791.

[6] Murakami Y, Yasuda T, Saigo K, Urashima T, Toyoda H, Okanoue T and Shimotohno K. Comprehensive analysis of microRNA expression patterns in hepatocellular carcinoma and non-tumorous tissues. Oncogene 2006; 25: 2537-2545.

[7] Bao L, Zhao J, Dai X, Wang Y, Ma R, Su Y, Cui H, Niu J, Bai S, Xiao Z, Yuan H, Yang Z, Li C, Cheng R and Ren X. Correlation between miR-23a and onset of hepatocellular carcinoma. Clin Res Hepatol Gastroenterol 2014; 38: 318-330.

[8] Gramantieri L, Ferracin M, Fornari F, Veronese A, Sabbioni S, Liu CG, Calin GA, Giovannini C, Ferrazzi E, Grazi GL, Croce CM, Bolondi L and Negrini M. Cyclin G1 is a target of miR-122a, a microRNA frequently down-regu-lated in human hepatocellular carcinoma. Cancer Res 2007; 67: 6092-6099.

[9] Li W, Xie L, He X, Li J, Tu K, Wei L, Wu J, Guo Y, Ma X, Zhang P, Pan Z, Hu X, Zhao Y, Xie H, Jiang G, Chen T, Wang J, Zheng S, Cheng J, Wan D, Yang S, Li Y and Gu J. Diagnostic and prognostic implications of microRNAs in human hepatocellular carcinoma. Int J Cancer 2008; 123: 1616-1622.

[10] Li T, Yin J, Yuan L, Wang S, Yang L, Du X and Lu J. Downregulation of microRNA-139 is associated with hepa-tocellular carcinoma risk and short-term survival. Oncol Rep 2014; 31: 1699-1706.

[11] Shen J, Wang S, Siegel AB, Remotti H, Wang Q, Sirosh I and Santella RM. Genome-wide expression of microR-NAs is regulated by DNA methylation in hepatocarcinogenesis. Gastroenterol Res Pract 2015; 2015: 230642.

[12] Liu AM, Zhang C, Burchard J, Fan ST, Wong KF, Dai H, Poon RT and Luk JM. Global regulation on microRNA in hepatitis B virus-associated hepatocellular carcinoma. OMICS 2011; 15: 187-191.

[13] Chung GE, Yoon JH, Myung SJ, Lee JH, Lee SH, Lee SM, Kim SJ, Hwang SY, Lee HS and Kim CY. High expression of microRNA-15b predicts a low risk of tumor recurrence following curative resection of hepatocellular carci-noma. Oncol Rep 2010; 23: 113-119.

[14] Elyakim E, Sitbon E, Faerman A, Tabak S, Montia E, Belanis L, Dov A, Marcusson EG, Bennett CF, Chajut A, Cohen D and Yerushalmi N. hsa-miR-191 is a candidate oncogene target for hepatocellular carcinoma thera-py. Cancer Res 2010; 70: 8077-8087.

[15] Wang W, Zhao LJ, Tan YX, Ren H and Qi ZT. Identification of deregulated miRNAs and their targets in hepatitis B virus-associated hepatocellular carcinoma. World J Gastroenterol 2012; 18: 5442-5453.

[16] Katayama Y, Maeda M, Miyaguchi K, Nemoto S, Yasen M, Tanaka S, Mizushima H, Fukuoka Y, Arii S and Tanaka H. Identification of pathogenesis-related microRNAs in hepatocellular carcinoma by expression profil-ing. Oncol Lett 2012; 4: 817-823.

[17] Han K, Li J, Zhao H, Liang P, Huang X, Zheng L, Li Y, Yang T and Wang L. Identification of the typical miRNAs and target genes in hepatocellular carcinoma. Mol Med Rep 2014; 10: 229-235.

[18] Gao B, Ning S, Li J, Liu H, Wei W, Wu F, Tang Y, Feng Y, Li K and Zhang L. Integrated analysis of differentially expressed mRNAs and miRNAs between hepatocellular carcinoma and their matched adjacent normal liver tissues. Oncol Rep 2015; 34: 325-333.

[19] Huang YS, Dai Y, Yu XF, Bao SY, Yin YB, Tang M and Hu CX. Microarray analysis of microRNA expression in hepatocellular carcinoma and non-tumorous tissues without viral hepatitis. J Gastroenterol Hepatol 2008; 23: 87-94.

[20] Meng F, Henson R, Wehbe-Janek H, Ghoshal K, Jacob ST and Patel T. MicroRNA-21 regulates expression of the PTEN tumor suppressor gene in human hepatocellular cancer. Gastroenterology 2007; 133: 647-658.

[21] Su H, Yang JR, Xu T, Huang J, Xu L, Yuan Y and Zhuang SM. MicroRNA-101, down-regulated in hepatocellular carcinoma, promotes apoptosis and suppresses tumorigenicity. Cancer Res 2009; 69: 1135-1142.

[22] Burchard J, Zhang C, Liu AM, Poon RT, Lee NP, Wong KF, Sham PC, Lam BY, Ferguson MD, Tokiwa G, Smith R, Leeson B, Beard R, Lamb JR, Lim L, Mao M, Dai H and Luk JM. microRNA-122 as a regulator of mitochondrial metabolic gene network in hepatocellular carcinoma. Mol Syst Biol 2010; 6: 402.

[23] Shih TC, Tien YJ, Wen CJ, Yeh TS, Yu MC, Huang CH, Lee YS, Yen TC and Hsieh SY. MicroRNA-214 downregula-tion contributes to tumor angiogenesis by inducing secretion of the hepatoma-derived growth factor in human hepatoma. J Hepatol 2012; 57: 584-591.


2

[24] Yang L, Ma Z, Wang D, Zhao W, Chen L and Wang G. MicroRNA-602 regulating tumor suppressive gene RASSF1A is overexpressed in hepatitis B virus-infected liver and hepatocellular carcinoma. Cancer Biol Ther 2010; 9: 803-808.

[25] Wang W, Zhao LJ, Tan YX, Ren H and Qi ZT. MiR-138 induces cell cycle arrest by targeting cyclin D3 in hepato-cellular carcinoma. Carcinogenesis 2012; 33: 1113-1120.

[26] Xia H, Ooi LL and Hui KM. MicroRNA-216a/217-induced epithelial-mesenchymal transition targets PTEN and SMAD7 to promote drug resistance and recurrence of liver cancer. Hepatology 2013; 58: 629-641.

[27] Li T, Xie J, Shen C, Cheng D, Shi Y, Wu Z, Zhan Q, Deng X, Chen H, Shen B, Peng C, Li H and Zhu Z. miR-150-5p inhibits hepatoma cell migration and invasion by targeting MMP14. PLoS One 2014; 9: e115577.

[28] Chen Y, Chen J, Liu Y, Li S and Huang P. Plasma miR-15b-5p, miR-338-5p, and miR-764 as biomarkers for hepatocellular carcinoma. Med Sci Monit 2015; 21: 1864-1871.

[29] Wong CM, Wong CC, Lee JM, Fan DN, Au SL and Ng IO. Sequential alterations of microRNA expression in he-patocellular carcinoma development and venous metastasis. Hepatology 2012; 55: 1453-1461.

[30] Sato F, Hatano E, Kitamura K, Myomoto A, Fujiwara T, Takizawa S, Tsuchiya S, Tsujimoto G, Uemoto S and Shimizu K. MicroRNA profile predicts recurrence after resection in patients with hepatocellular carcinoma within the Milan criteria. PLoS One 2011; 6: e16435.

[31] Li S, Fu H, Wang Y, Tie Y, Xing R, Zhu J, Sun Z, Wei L and Zheng X. MicroRNA-101 regulates expression of the v-fos FBJ murine osteosarcoma viral oncogene homolog (FOS) oncogene in human hepatocellular carcinoma. Hepatology 2009; 49: 1194-1202.

[32] Borel F, Han R, Visser A, Petry H, van Deventer SJ, Jansen PL, Konstantinova P; Reseau Centre de Ressources Biologiques Foie F. Adenosine triphosphate-binding cassette transporter genes up-regulation in untreated hepatocellular carcinoma is mediated by cellular microRNAs. Hepatology 2012; 55: 821-832.

[33] Wang W, Peng B, Wang D, Ma X, Jiang D, Zhao J and Yu L. Human tumor microRNA signatures derived from large-scale oligonucleotide microarray datasets. Int J Cancer 2011; 129: 1624-1634.

[34] El-Halawany MS, Ismail HM, Zeeneldin AA, Elfiky A, Tantawy M, Kobaisi MH, Hamed I and Abdel Wahab AH. Investigating the pretreatment miRNA expression patterns of advanced hepatocellular carcinoma patients in association with response to TACE treatment. Biomed Res Int 2015; 2015: 649750.

[35] Wang Y, Lee AT, Ma JZ, Wang J, Ren J, Yang Y, Tantoso E, Li KB, Ooi LL, Tan P and Lee CG. Profiling microRNA expression in hepatocellular carcinoma reveals microRNA-224 up-regulation and apoptosis inhibitor-5 as a microRNA-224-specific target. J Biol Chem 2008; 283: 13205-13215.

[36] Huang XH, Wang Q, Chen JS, Fu XH, Chen XL, Chen LZ, Li W, Bi J, Zhang LJ, Fu Q, Zeng WT, Cao LQ, Tan HX and Su Q. Bead-based microarray analysis of microRNA expression in hepatocellular carcinoma: miR-338 is down-regulated. Hepatol Res 2009; 39: 786-794.

[37] Murakami Y, Tanahashi T, Okada R, Toyoda H, Kumada T, Enomoto M, Tamori A, Kawada N, Taguchi YH and Azuma T. Comparison of hepatocellular carcinoma miRNA expression profiling as evaluated by next generation sequencing and microarray. PLoS One 2014; 9: e106314.

[38] Hung TM, Ho CM, Liu YC, Lee JL, Liao YR, Wu YM, Ho MC, Chen CH, Lai HS and Lee PH. Up-regulation of mi-croRNA-190b plays a role for decreased IGF-1 that induces insulin resistance in human hepatocellular carci-noma. PLoS One 2014; 9: e89446.

[39] Ma ZL, Yang L, Chen LB, Huang JM, Wang DD and Wang GB. Differential expression profile of microRNAs in the different stages of hepatitis B virus infection-related hepatocarcinogenesis (in Chinese). World Chinese Jour-nal of Digestology 2009; 17: 2112-2116.

[40] Sun K, Huang XH, C ZM and Wang Q. Analysis of microRNA expression patterns in hepatocellular carcinoma and non-tumorous tissues by liquichip assay (in Chinese). Chin J Exp Surg 2006; 23: 945-947.

[41] Wang CX, Song W, Li ZJ, Song B, Wu DH, Sun AM and Chen LH. MicroRNA profiling in patients with hepatocel-lular carcinoma (in Chinese). J South Med Univ 2010; 30: 976-980.

[42] Xiao HJ, Wang GY and Dong QH. MicroRNA expression profiles in hepatocellular carcinoma and adjacent non-tumorous tissues (in Chinese). Cancer Research on Prevention and Treatment 2012; 39: 947-949.

[43] Fan XT. Analyse the clinical risk factors of hepatocellular carcinoma based on cirrhosis and study on the mi-croRNA expression profiles of liver cancer and cirrhosis (PhD Thesis in Chinese). Urumqi: Xinjiang Medical University; 2013.

Documents

Review Article Dysregulated microRNAs with prognostic ... · Dysregulated microRNAs with prognostic significance in human hepatocellular carcinoma: a study based on high-throughput