OL Oncology Letters 1792-1074 1792-1082 D.A. Spandidos 10.3892/ol.2019.11031 OL-0-0-11031 Articles Reduced expression of microRNA-139-5p in hepatocellular carcinoma results in a poor outcome: An exploration the roles of microRNA-139-5p in tumorigenesis, advancement and prognosis at the molecular biological level using an integrated meta-analysis and bioinformatic investigation QinHui 1 WenDong-Yue 1 QueQiao 1 ZhouChuan-Yang 1 WangXiao-Dong 1 PengYu-Ting 1 HeYun 1 YangHong 1 LiaoBo-Ming 2 Department of Medical Ultrasonics, The First Affiliated Hospital of Guangxi Medical University, Nanning, Guangxi Zhuang Autonomous Region 530021, P.R. China Department of Internal Medicine, The First Affiliated Hospital of Guangxi Medical University, Nanning, Guangxi Zhuang Autonomous Region 530021, P.R. China Correspondence to: Professor Bo-Ming Liao, Department of Internal Medicine, The First Affiliated Hospital of Guangxi Medical University, 6 Shuangyong Road, Nanning, Guangxi Zhuang Autonomous Region 530021, P.R. China, E-mail: 2369204378@qq.com 12 2019 01 11 2019 18 6 6704 6724 10052019 27092019 Copyright: © Qin et al. 2019 This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

Hepatocellular carcinoma (HCC) is generally considered one of the most common gastrointestinal malignant tumors, characterized by high invasiveness and metastatic rate, as well as insidious onset. A relationship between carcinogenicity and aberrant microRNA-139-5p (miR-139-5p) expression has been identified in multiple tumors while the specific molecular mechanisms of miR-139-5p in HCC have not yet been thoroughly elucidated. A meta-analysis of available data from The Cancer Genome Atlas (TCGA), Gene Expression Omnibus, ArrayExpress and Oncomine databases, as well as the published literature, was comprehensively conducted with the aim of examining the impact of miR-139-5p expression on HCC. Additionally, predicted downstream target genes were confirmed using a series of bioinformatics tools. Moreover, a correlative biological analysis was performed to ascertain the precise function of miR-139-5p in HCC. The results revealed that the expression of miR-139-5p was noticeably lower in HCC compared with non-tumor liver tissues according to the pooled standard mean difference, which was −0.84 [95% confidence interval (CI): −1.36 to −0.32; P<0.001]. Furthermore, associations were detected between miR-139-5p expression and certain clinicopathological characteristics of TCGA samples, including tumor grade, pathological stage and T stage. Moreover, the pooled hazard ratio (HR) for overall survival (HR=1.37; 95% CI: 1.07–1.76; P=0.001) indicated that decreased miR-139-5p expression was a risk factor for adverse outcomes. Additionally, 382 intersecting genes regulated by miR-139-5p were obtained and assembled in signaling pathways, including ‘transcription factor activity, sequence-specific DNA binding’, ‘pathways in cancer’ and ‘Ras signaling pathway’. Notably, four targeted genes that were focused in ‘pathways in cancer’ were identified as hub genes and immunohistochemical staining of the proteins encoded by these four hub genes in liver tissues, explored using the Human Protein Atlas database, confirmed their expression patterns in HCC and normal liver tissues Findings of the present study suggest that reduced miR-139-5p expression is capable of accelerating tumor progression and is associated with a poor clinical outcome by modulating the expression of downstream target genes involved in tumor-associated signaling pathways.

 hepatocellular carcinoma progressive clinicopathological features bioinformatics analysis microRNA-139-5p Introduction

The increasing incidence and mortality rates of cancer in humans worldwide have exerted serious effects on the health and quality of life of affected individuals, particularly in developed countries (1,2). Hepatocellular carcinoma (HCC), a very common malignant gastrointestinal tumor that occurs in geographically heterogeneous regions, in particular Eastern Asia and sub-Saharan Africa, where medical service is relatively limited, accounts for 75–85% of primary liver cancer cases. The major risk factors, including chronic hepatitis B and hepatitis C, excessive intake of alcohol, aflatoxin exposure and metabolic liver disease, have increased the threat of HCC worldwide (2). HCC has been revealed to have sex-specific differences, as the prevalence and mortality rates are 2–3 times higher among men than women (3,4). This predisposition may be attributed to the frequent development of chronic hepatitis B virus infection, alcoholism, fatty liver disease and hepatitis C virus infection in men (5–7). Although substantial advances have been made in the therapeutic field of HCC, frequent recurrence, advanced diagnosis and metastasis remain challenging issues to overcome and are preventative factors in obtaining a good prognosis (8,9). Thus, a useful marker for estimating the diagnosis and curative outcome of HCC is essential.

Mature microRNAs (miRNAs), commonly recognized as a major family of endogenous and small non-coding RNAs with a length of ~22 nucleotides, bind to the 3′-untranslated region (UTR) of corresponding target messenger RNAs (mRNAs), resulting in the inhibition of protein translation or degradation of mRNAs (10–12). Importantly, it is generally acknowledged that miRNAs serve important roles in biological processes, including cell proliferation, expansion, differentiation, infiltration and apoptosis, by regulating the expression of related genes in tissues or blood plasma (13–15). Moreover, miRNA dysregulation is an important factor contributing to tumorigenesis and progression, furthermore, certain specific miRNAs, such as miR-92a-2, miR-133b, miR-34a and miR-96, can act as either oncogenes or tumor suppressors (16–18).

miR-139-5p, one of the most common subtypes of miR-139, which is located on chromosome 11q13.4, has become established as a critical focus of studies examining the initiation and progression of diverse cancers, including oral squamous carcinoma (19), colorectal cancer (20,21), bladder cancer (22), endometrial cancer (23) and HCC (24–27). Wang et al (27) reported that miR-139-5p inhibits HCC development, however, the study included a limited number of HCC specimens. Moreover, a number of studies have reported a close association between differential miR-139-5p expression and the prognosis of HCC (25,26,28,29). Nevertheless, literature reporting the precise molecular mechanisms and available published studies analyzing miR-139-5p regulation in HCC are inconclusive, thus additional comprehensive studies are required.

The current meta-analysis was performed by integrating miRNA sequencing and microarray data, and comprehensively analyzing published articles that mentioned a relationship between miR-139-5p regulation and clinical data from human HCC samples. In addition, a bioinformatics analysis was conducted in order to confirm the characteristics of miR-139-5p and its targeted genes in HCC pathology- and development-related mechanisms. The present meta-analysis may provide further suggestions regarding the development of novel molecular therapeutic strategies for patients with HCC.

 Materials and methods <sec> <title>Data mining and literature search

Microarray profiles associated with miRNA expression in HCC were searched and downloaded from the Gene Expression Omnibus (GEO) (30), The Cancer Genome Atlas (TCGA) (31), ArrayExpress (32) and Oncomine (33) databases. The related records were updated on March 28, 2019 and the following search terms were used: (malignan* OR cancer OR tumor OR tumour OR neoplas* OR carcinoma) AND (hepatocellular OR liver OR hepatic OR HCC). Initially, studies were chosen based on the following inclusion criteria: i) Human HCC confirmed by pathological analysis and ii) information about miR-139-5p regulation in HCC and nontumorous samples.

For the published literature search, scientific databases, including PubMed (https://www.ncbi.nlm.nih.gov/pubmed/), ScienceDirect (https://www.sciencedirect.com/), Web of Science (http://wokinfo.com/), Wang Fang (http://www.wanfangdata.com.cn/index.html), ChongQing VIP (http://cstj.cqvip.com/) and the China National Knowledge Infrastructure (https://www.cnki.net/), were comprehensively searched to determine the most relevant studies published up to March 28, 2019. The key search terms were: (miR-139-5p OR miRNA-139-5p OR microRNA139-5p OR miR139-5p OR miRNA139-5p OR miR 139-5p OR miRNA 139-5p OR microRNA 139-5p OR microRNA-139-5p) AND (malignan* OR cancer OR tumor OR tumour OR neoplas* OR carcinoma) AND (hepatocellular OR liver OR hepatic OR HCC). The inclusion criteria regarding the published studies were: i) HCC confirmed by an operation and pathological analysis; ii) information on the miR-139-5p expression levels (high vs. low) and relative clinicopathological data, including odds ratios (ORs) with 95% confidence intervals (CIs) of some clinicopathological parameters in HCC; and iii) direct clinical prognostic information, including overall survival (OS), recurrence-free survival (RFS) or disease-free survival (DFS) with 95% CIs. Furthermore, when the Kaplan-Meier curves were drawn clearly, we could artificially extract and calculate the effect sizes, including OS, RFS, DFS with corresponding 95% CIs statistically.

The following exclusion criteria were applied: i) Lack of information on miR-139-5p expression in patients with HCC; ii) the full study was not published in English or Chinese; iii) reported data were insufficient for the calculation of clinical pathological parameters and prognostic indicators; and iv) the articles were reviews, conference abstracts or case reports.

 Data extraction

The following information was considered for microarray data selection: Dataset name, case number (N), mean (M) and standard deviation (SD) of corresponding datasets among HCC and nontumorous tissues, as well as area under the curve (AUC) values. Available parameters from the TCGA clinicopathological analysis included: Age, sex, race, neoplasm status, tumor grade, pathological stage, tumor-node-metastasis (TNM) stage, lymph node metastasis and metastasis status.

The following data were extracted from the scientific publications: First author, year published, country, sample sizes, methods, sample sources, follow-up time, outcomes and hazard ratios (HRs) with corresponding 95% CIs for OS, RFS or DFS. Kaplan-Meier curves were generated using the officially recommended method by Tierney et al (34).

 Statistical analysis

Statistical analyses were performed using Stata (version 14.0; StataCorp LP), SPSS software (version 24.0; SPSS, IBM Corp.) and GraphPad Prism software (version 9; GraphPad Software, Inc.). Initially, microarray data were separated into two groups, namely the HCC tissue group and the nontumorous liver tissue group. miR-139-5p expression levels were calculated from each included dataset, the N and M ± SD of the two groups were then recorded and compared using Student's t-test. More than two groups in the TCGA clinical parameter calculation were analyzed using one-way ANOVA. In addition, scatter plots presenting miR-139-5p expression levels were generated using GraphPad Prism software, and the receiver operating characteristic (ROC) curve analysis of each dataset was performed using SPSS software. Meta-DiSc (version 1.4; Clinical Biostatistics Unit Miscellaneous Shareware), which is capable of testing the precision of meta-analyses, was utilized to assess diagnostic efficiency, including the summary receiver operating characteristic (sROC) curve, positive likelihood ratio (PLR), negative likelihood ratio (NLR), and diagnostic sensitivity and specificity. Moreover, the merged standard mean differences (SMDs) and 95% CIs for all appropriate datasets were assessed using Stata 14.0 software to analyze miR-139-5p expression levels in patients with HCC. The merged SMD was <0 and the 95% CI did not include 0, indicating a relatively low expression level of miR-139-5p in HCC tissues. Furthermore, Kaplan-Meier survival curves were manually extracted from published studies via Engauge Digitizer software (version 4.1; http://markummitchell.github.io/engauge-digitizer/), and outcomes were recorded as HRs and 95% CIs. These values were ultimately merged using Stata 14.0 software aimed at confirming the prognostic effect of miR-139-5p. The HR was considered significant if P<0.05 and the 95% CI did not exceed 1; otherwise, no significant effect of miR-139-5p regulation was observed in patients with HCC. Heterogeneity analysis was evaluated by the Cochrane Q and inconsistency index (I2) tests (35). A random-effects model was employed if clear heterogeneity was indicated with I2>50% or P<0.05; if not, a fixed-effects model was selected. Additionally, publication bias was detected by Begg's funnel test, and influence analysis was performed to determine whether the studies or datasets included in the present study exerted excessive effects on pooled outcomes by independently removing each study.

 Bioinformatics analysis

Twelve online tools, namely miRWalk3.0 (http://mirwalk.umm.uni-heidelberg.de/), MiRTarBase (http://mirtarbase.mbc.nctu.edu.tw/php/index.php), TarBase6.0 (http://www.microrna.gr/tarbase), TargetMiner (http://www.mybiosoftware.com/targetminer-microrna-target-prediction.html), polymiRTS3.0 (http://compbio.uthsc.edu/miRSNP), RNA22 (https://cm.jefferson.edu/rna22/), microRNA.org (http://www.microrna.org/), PITA (https://genie.weizmann.ac.il/pubs/mir07/mir07_data.html), mirRNAMAP2.0 (http://mirnamap.mbc.nctu.edu.tw/), TargetScan (http://www.targetscan.org/vert_71/), miRDB (http://mirdb.org/) and PicTar-vert (http://dorina.mdc-berlin.de) (36–47), were employed to comprehensively investigate the impact of miR-139-5p on biological processes. Genes were investigated to predict possible target genes and differentially expressed genes (DEGs) in HCC tissues were downloaded from TCGA. Next, a Venn diagram (http://bioinfogp.cnb.csic.es/tools/venny/) was generated to obtain intersecting DEGs as final target genes from the aforementioned DEGs and predicted genes. Furthermore, the final target genes were subjected to Gene Ontology (GO) enrichment, Kyoto Encyclopedia of Genes and Genomes (KEGG) and Panther pathway analyses using the Database for Annotation, Visualization and Integrated Discovery (DAVID; version 6.8; http://david.ncifcrf.gov) (48). P<0.05 was considered to indicate a statistically significant difference in all pathways. Moreover, images of correlation networks were presented in a bubble diagram that was constructed using the ImageGP online website (http://www.ehbio.com/ImageGP/). The Search Tool for the Retrieval of Interacting Genes (STRING; http://string-db.org; version 11.0) was employed to explore the protein-protein interaction (PPI) network and identify the top downstream target genes with the greatest number of connections of each node that were most likely regulated by miR-139-5p in HCC (49,50). Additionally, Gene Expression Profiling Interactive Analysis (GEPIA; http://gepia.cancer-pku.cn/; version 1.0) (51) was used to determine the expression and relative prognostic roles of downstream target genes in the development of HCC. Intersecting DEGs were identified as hub genes from the top downstream genes and the target genes that were prominently enriched in the first pathway, which included the most statistically significant differences of KEGG. Additionally, the Human Protein Atlas database (52) and CBio Cancer Genomics Portal (53) were searched to determine the deregulation and mutations of hub genes. The expression patterns of hub genes were validated using immunohistochemical (IHC) staining images of pathological sections obtained from the Human Protein Atlas, and comparing normal and HCC tissue samples. The correlation between hub gene and miR-139-5p expression in TCGA data was evaluated by the calculation of Pearson's correlation coefficients using SPSS software.

 Results <sec> <title>Characteristics of the included studies, miRNA-sequencing (miRNA-seq, and microarray data

In the literature search performed using the aforementioned retrieval strategy, 205 studies were preliminarily obtained. After thorough examination of all the studies, 201 studies were excluded due to 36 articles being duplicated, and no available clinical information being explicitly provided in 165 articles. Consequently, only four studies that provided associations with the survival of 367 patients could be included (25,28,29,42). Three studies presented survival curve data that allowed OS calculation, and one directly provided accurate OS data. However, only one study separately (28,29) described DFS and RFS, which were not merged to confirm the significance of these two effect sizes. Details of the included terms are presented in Table I, and a schematic of the literature and microarray data screening processes is shown in Fig. 1.

The miRNA-seq and microarray data were initially retrieved from 90,880 datasets, and after careful examination and verification of the data, 18 datasets were identified to provide miR-139-5p expression data regarding HCC and nontumor liver tissues that conformed to the inclusion criteria of the present study. These 18 datasets were retrieved from the TCGA, GEO, ArrayExpress and Oncomine databases. Specifically, 14 datasets were obtained from GEO and Oncomine, including GSE12717, GSE21279, GSE21362, GSE36915, GSE39678, GSE40744, GSE41874, GSE50013, GSE54751, GSE64632, GSE67882, GSE69580, GSE98269 and GSE115016. Three datasets were retrieved from the ArrayExpress database: E-MTAB-511, E-MTAB-3347 and E-MTAB-4809. The remaining TCGA dataset containing available data of miR-139-5p expression profile was downloaded from TCGA miRNA-seq profiles. Detailed information is presented in Table II.

 Analysis of miRNA-seq and microarray data acquired from public databases and the scientific literature miR-139-5p expression data in HCC tissues obtained from TCGA

As presented in Table III, lower miR-139-5p expression levels were observed in HCC tissues (n=371) than in normal tissues (n=49) (6.4189±1.2376 vs. 8.5117±0.4539; t=−22.926; P<0.01). Statistically significant associations (P<0.05) were observed among miR-139-5p expression and tumor grade (G1/G2/G3/G4), pathological stage (stage I/II/III/IV) and T stage (T1-2/3–4). However, no statistically significant association was observed between miR-139-5p expression and the remaining clinicopathological characteristics.

 Meta-analysis of miRNA-seq and microarray data from four public databases

As illustrated in Fig. 2, miR-139-5p downregulation in HCC groups was clearly observed in the following datasets: GSE12717, GSE21362, GSE36915, GSE39678, GSE40744, GSE41874, GSE54751, GSE69580, GSE98269, GSE115016 and TCGA. However, higher miR-139-5p expression levels in HCC tissues were identified in the E-MTAB-4809 dataset acquired from ArrayExpress, and no statistically significant difference was detected in the remaining datasets. In addition, corresponding ROC curves were constructed and AUC values were calculated (Fig. 3). Moreover, pooled SMDs with 95% CIs for all included datasets are presented as a forest plot in Fig. 4A, suggesting reduced miR-139-5p expression in HCC tissues, as determined by a random-effects model (SMD=−0.84; 95% CI: −1.36 to −0.32; P<0.001; I2=85.7%). Additionally, the Begg's funnel plot of combined datasets revealed clear symmetry, indicating no evident publication bias in the current analysis of miRNA-seq and microarray data (Fig. 4B). Furthermore, the results from the sensitivity analysis (Fig. 4C) demonstrated no indication of any individual study significantly affecting the final SMD and its 95% CI.

 Determination of the diagnostic efficacy of miR-139-5p expression in HCC

As shown in Fig. 5, the combined diagnostic OR was 14.39 (CI: 5.96–34.75) and a slight heterogeneity was indicated by the I2, which was found to be 56.3%. Furthermore, the sensitivity and specificity of the included datasets were calculated, and the corresponding results were 0.87 (0.85–0.90) and 0.66 (0.60–0.72), respectively. Moreover, the pooled PLR and NLR values were 2.50 (CI: 1.53–4.07) and 0.23 (CI: 0.16–0.33), respectively. The AUC result of the sROC curve was 0.8978, indicating that miR-139-5p low-expression has a relatively high diagnostic efficacy to discriminatee HCC from non-tumor liver tissue based on the miRNA-seq and microarray data.

 Clinical role of miR-139-5p expression in HCC tissue samples determined from scientific publications

Four publications (25,28,29,54) offered correlative prognostic information to verify the clinical role of miR-139-5p in HCC. HRs for OS were combined with 95% CIs (HR=1.37; 95% CI: 1.07–1.76; P=0.001) and calculated using a fixed-effects model due to a lack of notable heterogeneity (I2=43.9%; P=0.148). The results indicated that decreased miR-139-5p expression was negatively associated with patient prognosis, as shown in Table IV and Fig. 6A. Furthermore, Begg's funnel test was performed to evaluate publication bias in the four included studies (Fig. 6B). No marked bias was observed based on the symmetry of the funnel plot. Moreover, as shown in Fig. 6C, it was concluded that individual studies did not affect the final outcomes. Only two studies (25,29) provided correlations among clinicopathological parameters and miR-139-5p expression in HCC tissues, while other studies reported on DFS (28) and RFS (29) rates (DFS: HR=1.67, 95% CI: 0.57–4.87; RFS: HR=1.20, 95% CI: 0.61–2.37).

 Bioinformatics analysis Target genes of miR-139-5p and their clinical roles

A total of 4,294 differentially expressed genes obtained from TCGA, 1,406 targeted genes predicted by the miRWalk website, and a total of 382 intersecting genes were downloaded as final target genes (Fig. 7A). Furthermore, DAVID was employed to conduct a GO functional enrichment analysis, in addition to KEGG and Panther pathway analyses for the purpose of identifying specific molecular mechanisms of miR-139-5p related to HCC progression and prognosis. As shown in Fig. 8 and Table V, the genes were significantly enriched in ‘response to drug’, ‘signal transduction’, ‘cell migration involved in sprouting angiogenesis’ and ‘positive regulation of cell proliferation’ (P<0.01), based on the biological process of GO. Regarding the cellular component of GO, the genes were specifically focused on ‘plasma membrane’, ‘integral component of plasma membrane’ and ‘neuron projection’ (P<0.01). Concerning the molecular function of GO, the functional genes were primarily assembled in ‘transcription factor activity, sequence-specific DNA binding’ and ‘Ras guanyl-nucleotide exchange factor activity’ (P<0.01). Moreover, as presented in Table VI, KEGG and Panther pathway analyses were performed on several signaling pathways, including ‘pathways in cancer’, ‘Ras signaling pathway’ and ‘endogenous cannabinoid signaling’ (P<0.05).

 Construction of target gene networks and identification of hub genes

The STRING database was employed in order to predict 382 intersected genes that were associated with miR-139-5p expression by constructing a PPI network. As presented in Fig. 7B, the top eight genes, namely, FBJ murine osteosarcoma viral oncogene homolog (FOS), chromatin licensing and DNA replication factor 1 (CDT1), enhancer of zeste homolog 2 (EZH2), recombination protein (RAD51), forkhead box M1 (FOXM1), forkhead box O1 (FOXO1), cell division cycle 25A (CDC25A), and guanine nucleotide binding protein alpha 12 (GNA12) with connection degrees >14 were chosen from the PPI network for further investigation. Moreover, four genes, FOS, FOXO1, GNA12 and RAD51, identified in the first signaling pathway termed ‘pathways in cancer’ were eventually determined as hub genes.

 Target gene expression and associated clinical implications

An analysis was conducted using GEPIA to determine the top eight genes associated with HCC development and verify their specific clinical implications. The expression levels of four genes, CDT1, EZH2, FOXM1 and RAD51, were substantially increased in HCC tissues, whereas decreased FOS gene expression was observed in HCC tissues (P<0.05; Fig. 9). Moreover, the prognostic value of the top eight genes was also evaluated (Fig. 10). The expression of six genes, CDT1, EZH2, RAD51, FOXM1, GNA12 and CDC25A, was negatively associated with the OS of patients with HCC. In addition, the frequency of mutations in the top eight genes (Fig. 11), and the expression patterns of the four hub genes FOS, FOXO1, GNA12 and RAD51 in HCC and normal liver tissues were explored using the Human Protein Atlas database (Fig. 12). Furthermore, based on TCGA data, a positive correlation was identified between miR-139-5p expression in HCC and the regulation of FOS and FOXO1 genes, which were also reduced in HCC tissues (Fig. 13). Consistent with the aforementioned findings, a negative correlation was detected between miR-139-5p expression and the regulation of GNA12 and RAD51 in patients with HCC (P<0.001).

 Discussion

Based on the miRNA-seq and microarray data retrieved from four public databases and data obtained from published literature (25,28,29,54), including human HCC specimens, the primary aim of this study was to perform additional bioinformatics analyses examining miR-139-5p and the corresponding target gene expression and molecular mechanism involved in HCC development. Based on the analytical miRNA-seq and microarray data, reduced miR-139-5p expression levels were observed in the HCC group (SMD=−0.84; 95% CI: −1.36 to −0.32; P<0.001). Based on the data retrieved from the TCGA database, certain clinicopathological features, including grade, pathological and T stage, were associated with miR-139-5p expression, indicating that decreased miR-139-5p expression may facilitate disease progression. Moreover, the AUC value of the sROC curve for data obtained from public databases was 0.8978, indicating a relatively high diagnostic efficacy of miR-139-5p expression. Regarding the association between clinical characteristics and miR-139-5p expressed in HCC based on relevant literature, the pooled OS (HR=1.37; 95% CI: 1.07–1.76) revealed that decreased miR-139-5p expression led to a poor prognosis in patients with HCC. Furthermore, bioinformatics software was used to identify 382 genes regulated by miR-139-5p in HCC, for which GO, KEGG and Panther pathway analyses were performed. The target genes were significantly enriched in ‘response to drug’, ‘signal transduction’, ‘plasma membrane’, ‘transcription factor activity, sequence-specific DNA binding’, ‘pathways in cancer’, ‘Ras signaling pathway’ and ‘endogenous cannabinoid signaling’ (P<0.05). In addition, the top eight target genes, FOS, CDT1, EZH2, RAD51, FOXM1, FOXO1, GNA12 and CDC25A, were selected by PPI network construction; and their clinical impacts were investigated. Moreover, FOS, FOXO1, GNA12 and RAD51, which were enriched in ‘pathways in cancer’, were determined as hub genes. The IHC staining for proteins encoded by these four genes, namely, FOS, FOXO1, GNA12 and RAD51, was further confirmed using the Human Protein Atlas database, and the expression atlas of the pathological sections supported the downregulation of FOS and FOXO1 accompanied by the upregulation of GNA12 and RAD51 in HCC tissues. Furthermore, a relativity analysis revealed that miR-139-5p expression was positively correlated with the regulation of FOS and FOXO1 and negatively correlated with GNA12 and RAD51 regulation, suggesting that miR-139-5p may mediate the expression of hub genes involved in the aforementioned signaling pathways, mainly in ‘pathways in cancer’.

Mature miRNAs, which associate with the RNA-induced silencing complex (RISC), are capable of promoting the interaction of RISC with the 3′-UTR of the downstream target mRNA, resulting in abnormal translation of downstream mRNA. In addition, miR-139-5p is regarded as a tumor suppressor due to its regulatory effects on multiple malignant tumors. Notably, miR-139-5p controls tumor progression by regulating target mRNA expression and participates in downstream signaling pathways in malignancies (55–59), including HCC. According to a report by Wong et al (29), the progression and invasion of HCC are inhibited by miR-139-5p via its effects on Rho-kinase 2 expression. In addition, Au et al (60) revealed that miR-139-5p greatly affects the metastasis-related pathways of HCC cells by increasing the expression of EZH2, which is involved in sustaining the transcriptional suppression of genes over sequential cell generations. The increased expression of EZH2 contributed to metastasis, and poor prognosis of HCC identified in our analysis (Figs. 9 and 10). As reported by Hua et al (24), miR-139-5p regulates aerobic glycolysis, cell proliferation and motility by interacting with ETS1 in HCC cells. Moreover, a meta-analysis based on published articles that explored miR-139-5p expression in digestive system tumors, including one HCC case that was analyzed prospectively, indicated an unfavorable effect on the life expectancy of patients with gastrointestinal tumors (27). Additionally, downregulation of miR-139-5p has been demonstrated to result in poor outcome and disease progression (25,26,28). This finding was confirmed in the present study as indicated by the merged results with an SMD of −0.84 and HR for OS of 1.37 (95% CI: 1.07–1.76).

To further investigate the oncogenic molecular mechanism of miR-139-5p expression, bioinformatics analysis was performed with biological software and the top eight target genes were ultimately evaluated. Regarding the GO enrichment analysis, target genes were notably enriched in ‘response to drug’ and ‘positive regulation of cell proliferation’. Certain studies have reported that miR-139-5p expression has a major impact on pesticide effects in cancer therapy (58,61). Yoon et al (62) selected certain serum miRNAs, including miR-139-5p, associated with positive radiological responses or improved survival to assess the efficacy of sorafenib in patients with HCC. However, no statistically significant difference was reported between miR-139-5p expression and sorafenib treatment in patients with HCC. Additional correlative studies focused on drug efficacy and miR-139-5p expression should be further developed.

In terms of the pathway analysis, ‘pathways in cancer’, ‘Ras signaling pathway’ and ‘endogenous cannabinoid signaling’ were strongly enriched by predictive target genes. Importantly, miR-139-5p regulation serves an important role in tumor development and progression (63). Results from the present study revealed an association of miR-139-5p expression with progression and poor outcome in patients with HCC. Mitogenic signaling cascades associated with proliferation and translation are facilitated by activated Ras (63), and Ras regulates programmed cell death (64). Thus, an association exists between Ras alterations and carcinogenesis in various malignancies (65–67). Moreover, miRNAs have the ability to initiate the Ras-mitogen activated protein kinase pathway related to cell proliferation and survival, while no publication has reported an association between the Ras signaling pathway and miR-139 expression. Endogenous cannabinoid signaling, which is mainly determined by the activity of endocannabinoids N-arachidonoylethanolamine and 2-arachidonoylglycerol, as well as cannabinoid receptor type receptors 1 and 2 has attracted increasing interest due to its antitumor effects (68,69). According to Vago et al (70), the suppression of N-acylethanolamine-hydrolyzing acid amidase activity regulates tumor cell death and migration, and controls the progression of bladder cancer, indicating a new therapeutic target for patients with bladder cancer. Furthermore, Martínez-Martínez et al (71) verified that activated CB2 accelerates the development of colon cancer via a mechanism dependent on the AKT/GSK3β signaling pathway. Nevertheless, a relationship between endogenous cannabinoid signaling and miR-139 regulation has not been experimentally confirmed, therefore further research is required.

To better understand the downstream mechanisms of the top eight target genes, biofunctional investigations of FOS, CDT1, EZH2, RAD51, FOXM1, FOXO1, GNA12 and CDC25A were conducted. CDT1, which is involved in the formation of the prereplication complex required for DNA fragment replication, functions as an oncogene in multiple tumor types, including HCC (72–75). Karavias et al (75) confirmed that increased CDT1 expression has a negative influence on the survival of patients with HCC and observed correlations between the upregulation of CDT1 and tumor grade and TNM stage by performing IHC staining in HCC tissues, supporting the results of the present study (Figs. 9 and 11). Moreover, Yu et al (76) discovered a combination of new diagnostic indicators, including CDT1, MCM7, NUDT1, CENPM and HDAC11, with favorable diagnostic efficiency in HCC. Accordingly, CDT1 expression is crucial for the development and prognostic evaluation of HCC, thus further research is required to verify the association between miR-139-5p and CDT1 expression in HCC. RAD51 recombinase (RAD51), which participates in reestablishing and repairing homologous DNA, binds to the BRCA1 and BRCA2 genes, which are involved in tumorigenesis (77). Previous studies have reported that certain pharmaceutical molecules, including corylin, melatonin and gefitinib in combination with irinotecan, were able to enhance cancer cell sensitivity to chemotherapy by inhibiting RAD51-induced DNA repair in HCC (78–80). Luo et al (81) observed increased expression of RAD51 in HCC tissues, consistent with the findings of the present study, and revealed that miRNA-146a-5p enhanced the radio-sensitivity of HCC cells via the DNA repair pathway. However, studies have not yet reported an association between miR-139-5p and RAD51 regulation in patients with HCC. FOXM1 and FOXO1, members of the forkhead family, serve vital roles in cell proliferation and gene transcription. According to certain studies (82–85), FOXM1 overexpression results in poor survival and undesirable development of HCC, consistent with the present study results (Figs. 9 and 10). Lin et al (86) confirmed the reduced expression of FOXO1 in HCC tissues using IHC, and reported that FOXO1 served an important role in the development of HCC by regulating miRNA expression (87–89). CDC25A, as a member of the CDC25 family of phosphatases, mediates the G1 to S phase transition. A strong correlation between CDC25 expression and poor prognosis of patients with HCC was demonstrated by Xu et al (90) and a relevant study reported that the suppression of CDC25 expression halts the proliferation and progression of HCC (91). GNA12 has been found to possess carcinogenic potential and accelerate the progression of certain tumors (92–95). No studies have focused on the relevancy between miR-139-5p expression and FOXM1, FOXO1, GNA12 and CDC25A in HCC. The present study may provide novel insights for subsequent studies. FOS has multiple functions in cell proliferation, differentiation, transformation and apoptosis. Moreover, an association between FOS expression and HCC development has been reported (96–98). Fan et al (98) indicated that decreased miRNA-139 expression potentially facilitates the progression and metastasis of HCC via inhibiting FOS expression. Further research is necessary.

In short, it should be noted that miR-139-5p expression has a strong impact on the progression and prognosis of HCC mainly via targeted regulation of EZH2 and FOS expression. Moreover, the results of the present study clearly demonstrated that the regulation of FOS and FOXO1 was positively associated with miR-139-5p expressed in HCC. However, GNA12 and RAD51 expression levels were negatively associated with miR-139-5p expression based on the TCGA data. However, no studies have yet demonstrated the involvement of miR-139-5p in HCC development via targeted modulation of CDT1, RAD51, FOXM1, FOXO1, GNA12 and CDC25A expression.

The present study had certain limitations. Firstly, only four publications described a relationship among miR-139-5p regulation and prognostic effect size, including OS, DFS and RFS in HCC. Secondly, the HRs and corresponding 95% CIs were directly obtained from Kaplan-Meier curves, which generated biased outcomes due to limited data. Thirdly, the included miRNA-seq and microarray data were obtained from diverse platforms and channels, suggesting clear heterogeneity and poor diagnostic efficiency. Accordingly, a uniform method for detecting miR-139-5p expression in HCC should be formulated. Moreover, evidence for signaling pathways that are regulated by miR-139-5p via the targeted modulation of hub gene expression is lacking, and further in vitro and in vivo studies are required to determine the precise biological effects of the target genes on HCC development.

In summary, the findings of the present study revealed that miR-139-5p expression was lower in HCC tissues than in nontumorous tissues and strong associations were observed for miR-139-5p expression with tumor grade, pathological stage and T stage. Additionally, reduced miR-139-5p expression generally led to an undesirable prognosis in patients with HCC. Moreover, the top eight target genes, namely FOS, CDT1, EZH2, RAD51, FOXM1, FOXO1, GNA12 and CDC25A induced by miR-139-5p expression were revealed to be involved in several pathways, including ‘response to drug’, ‘pathways in cancer’, ‘Ras signaling pathway’ and ‘endogenous cannabinoid signaling’. These genes may have a major impact on carcinogenesis and the development of HCC. The current study provided a comprehensive investigation of the role of miR-139-5p in HCC progression. Nevertheless, additional research of the underlying molecular mechanisms related to HCC, including in vitro and in vivo experiments, should be further conducted to confirm the outcomes of the present study.

 Acknowledgements

Not applicable.

 Funding

The present study was supported by the Guangxi Science and Technology Program (grant no. GuiKeAB17195020) and the National Natural Science Foundation of China (grant no. NSFC81860319).

 Availability of data and materials

The datasets used and/or analyzed during the present study are available from the corresponding author on reasonable request.

 Authors' contributions

HQ, BML, XDW, DYW, HY and YH made substantial contributions to the conception and design of the present study. HQ, QQ, YTP and CYZ assisted in the acquisition, analysis and interpretation of data. HQ, DYW, BML, YH and HY were involved in drafting the manuscript and BML, YH and HY critically revising it for important intellectual content. All authors gave final approval for the version of the manuscript to be published.

 Ethics approval and consent to participate

Not applicable.

 Patient consent for publication

Not applicable.

 Competing interests

The authors declare that they have no competing interests.

 References BrayFSoerjomataramIThe changing global burden of cancer: Transitions in human development and implications for cancer prevention and control10.1596/978-1-4648-0349-9_ch2 BrayFFerlayJSoerjomataramISiegelRLTorreLAJemalAGlobal cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countriesCA Cancer J Clin68394424201810.3322/caac.2149230207593 SiegelRLMillerKDJemalACancer statistics, 2017CA Cancer J Clin67730201710.3322/caac.2138728055103 YehSHChenPJGender disparity of hepatocellular carcinoma: The roles of sex hormonesOncology(78 Suppl 1)S172S179201010.1159/000315247 AghemoAUpdate on HCC management and review of the new EASL guidelinesGastroenterol Hepatol (NY)143843862018 SingalAGEl-SeragHBHepatocellular carcinoma from epidemiology to prevention: Translating knowledge into practiceClin Gastroenterol Hepatol1321402151201510.1016/j.cgh.2015.08.01426284591 OtedoASimbiriKOWereVOngatiOEstambaleBARisk factors for liver Cancer in HIV endemic areas of Western KenyaInfect Agent Cancer1341201810.1186/s13027-018-0214-530607173 OweiraHPetrauschUHelblingDSchmidtJMehrabiASchöbOGiryesAAbdel-RahmanOPrognostic value of site-specific extra-hepatic disease in hepatocellular carcinoma: A SEER database analysisExpert Rev Gastroenterol Hepatol11695701201710.1080/17474124.2017.129448528276812 IntaraprasongPSiramolpiwatSVilaichoneRKAdvances in management of hepatocellular carcinomaAsian Pac J Cancer Prev1736973703201627644603 Valinezhad OrangASafaralizadehRKazemzadeh-BaviliMMechanisms of miRNA-mediated gene regulation from common downregulation to mRNA-specific upregulationInt J Genomics2014970607201410.1155/2014/97060725180174 FaniniFFabbriMMicroRNAs and cancer resistance: A new molecular plotClin Pharmacol Ther99485493201610.1002/cpt.35326875151 BartelDPMicroRNAs: Target recognition and regulatory functionsCell136215233200910.1016/j.cell.2009.01.00219167326 MitchellPSParkinRKKrohEMFritzBRWymanSKPogosova-AgadjanyanELPetersonANoteboomJO'BriantKCAllenACirculating microRNAs as stable blood-based markers for cancer detectionProc Natl Acad Sci USA1051051310518200810.1073/pnas.080454910518663219 ManikandanJAarthiJJKumarSDPushparajPNOncomirs: The potential role of non-coding microRNAs in understanding cancerBioinformation2330334200810.6026/9732063000233018685719 YatesLANorburyCJGilbertRJThe long and short of microRNACell153516519201310.1016/j.cell.2013.04.00323622238 BecaFSchmittFMicroRNA signatures in cytopathology: Are they ready for prime time?Cancer Cytopathol124613615201610.1002/cncy.2172827153190 WangJChenJSenSMicroRNA as biomarkers and diagnosticsJ Cell Physiol2312530201610.1002/jcp.2505626031493 KellerARoungeTBackesCLudwigNGislefossRLeidingerPLangsethHMeeseESources to variability in circulating human miRNA signaturesRNA Biol1417911798201710.1080/15476286.2017.136788828820329 ChenZYuTCabayRJJinYMahjabeenILuanXHuangLDaiYZhouXmiR-486-3p, miR-139-5p, and miR-21 as biomarkers for the detection of oral tongue squamous cell carcinomaBiomark Cancer918201710.1177/1179299X170090000128096697 ZouFMaoRYangLLinSLeiKZhengYDingYZhangPCaiGLiangXLiuJTargeted deletion of miR-139-5p activates MAPK, NF-κB and STAT3 signaling and promotes intestinal inflammation and colorectal cancerFEBS J28314381452201610.1111/febs.1367826859226 LiuHYinYHuYFengYBianZYaoSLiMYouQHuangZmiR-139-5p sensitizes colorectal cancer cells to 5-fluorouracil by targeting NOTCH-1Pathol Res Pract212643649201610.1016/j.prp.2016.04.01127173050 HuYDengCZhangHZhangJPengBHuCLong non-coding RNA XIST promotes cell growth and metastasis through regulating miR-139-5p mediated Wnt/β-catenin signaling pathway in bladder cancerOncotarget89455494568201710.18632/oncotarget.2179129212249 LiuJLiCJiangYWanYZhouSChengWTumor-suppressor role of miR-139-5p in endometrial cancerCancer Cell Int1851201810.1186/s12935-018-0545-829618950 HuaSLeiLDengLWengXLiuCQiXWangSZhangDZouXCaoCmiR-139-5p inhibits aerobic glycolysis, cell proliferation, migration, and invasion in hepatocellular carcinoma via a reciprocal regulatory interaction with ETS1Oncogene3716241636201810.1038/s41388-017-0057-329335523 WangZDingQLiYLiuQWuWWuLYuHReanalysis of microRNA expression profiles identifies novel biomarkers for hepatocellular carcinoma prognosisTumour Biol371477914787201610.1007/s13277-016-5369-327633066 NiHDaiXLengXDengMQinYJiQXuCLiJLiuYHigher variety and quantity of microRNA-139-5p isoforms confer suppressive role in hepatocellular carcinomaJ Cell Biochem11968066813201810.1002/jcb.2687429693285 WangYHJiJWengHWangBCWangFBMiR-139 in digestive system tumor diagnosis and detection: Bioinformatics and meta-analysisClin Chim Acta4853341201810.1016/j.cca.2018.06.00629883634 WangLLiuMZhuHRongWWuFAnSLiuFFengLWuJXuNIdentification of recurrence-related serum microRNAs in hepatocellular carcinoma following hepatectomyCancer Biol Ther1614451452201510.1080/15384047.2015.107173026176380 WongCCWongCMTungEKAuSLLeeJMPoonRTManKNgIOThe microRNA miR-139 suppresses metastasis and progression of hepatocellular carcinoma by down-regulating Rho-kinase 2Gastroenterology140322331201110.1053/j.gastro.2010.10.00620951699 CloughEBarrettTThe gene expression omnibus databaseMethods Mol Biol141893110201610.1007/978-1-4939-3578-9_527008011 TomczakKCzerwińskaPWiznerowiczMThe Cancer Genome Atlas (TCGA): An immeasurable source of knowledgeContemp Oncol (Pozn)19A68A77201525691825 ParkinsonHSarkansUKolesnikovNAbeygunawardenaNBurdettTDylagMEmamIFarneAHastingsEHollowayEArrayExpress update-an archive of microarray and high-throughput sequencing-based functional genomics experimentsNucleic Acids Res39D1002D1004201110.1093/nar/gkq104021071405 RhodesDRYuJShankerKDeshpandeNVaramballyRGhoshDBarretteTPandeyAChinnaiyanAMONCOMINE: A cancer microarray database and integrated data-mining platformNeoplasia616200410.1016/S1476-5586(04)80047-215068665 TierneyJFStewartLAGhersiDBurdettSSydesMRPractical methods for incorporating summary time-to-event data into meta-analysisTrials816200710.1186/1745-6215-8-1617555582 VamvakasECMeta-analyses of studies of the diagnostic accuracy of laboratory tests: A review of the concepts and methodsArch Pathol Lab Med12267568619989701328 StichtCDe La TorreCParveenAGretzNmiRWalk: An online resource for prediction of microRNA binding sitesPLoS One13e0206239201810.1371/journal.pone.020623930335862 ChouCHShresthaSYangCDChangNWLinYLLiaoKWHuangWCSunTHTuSJLeeWHmiRTarBase update 2018: A resource for experimentally validated microRNA-target interactionsNucleic Acids Res46D296D302201810.1093/nar/gkx106729126174 VergoulisTVlachosISAlexiouPGeorgakilasGMaragkakisMReczkoMGerangelosSKozirisNDalamagasTHatzigeorgiouAGTarBase 6.0: Capturing the exponential growth of miRNA targets with experimental supportNucleic Acids Res40D222D229201210.1093/nar/gkr116122135297 BandyopadhyaySMitraRTargetMiner: MicroRNA target prediction with systematic identification of tissue-specific negative examplesBioinformatics2526252631200910.1093/bioinformatics/btp50319692556 BhattacharyaAZiebarthJDCuiYPolymiRTS Database 3.0: Linking polymorphisms in microRNAs and their target sites with human diseases and biological pathwaysNucleic Acids Res42D86D91201410.1093/nar/gkt102824163105 LoherPRigoutsosIInteractive exploration of RNA22 microRNA target predictionsBioinformatics2833223323201210.1093/bioinformatics/bts61523074262 BetelDWilsonMGabowAMarksDSSanderCThe microRNA.org resource: Targets and expressionNucleic Acids Res36D149D153200810.1093/nar/gkm99518158296 JohnBEnrightAJAravinATuschlTSanderCMarksDSHuman MicroRNA targetsPLoS Biol2e363200410.1371/journal.pbio.002036315502875 HsuSDChuCHTsouAPChenSJChenHCHsuPWWongYHChenYHChenGHHuangHDmiRNAMap 2.0: Genomic maps of microRNAs in metazoan genomesNucleic Acids Res36D165D169200810.1093/nar/gkm101218029362 AgarwalVBellGWNamJWBartelDPPredicting effective microRNA target sites in mammalian mRNAsElife4201510.7554/eLife.05005 LiuWWangXPrediction of functional microRNA targets by integrative modeling of microRNA binding and target expression dataGenome Biol2018201910.1186/s13059-019-1629-z30670076 KrekAGrunDPoyMNWolfRRosenbergLEpsteinEJMacMenaminPda PiedadeIGunsalusKCStoffelMRajewskyNCombinatorial microRNA target predictionsNat Genet37495500200510.1038/ng153615806104 DennisGJrShermanBTHosackDAYangJGaoWLaneHCLempickiRADAVID: Database for annotation, visualization, and integrated discoveryGenome Biol4P3200310.1186/gb-2003-4-5-p312734009 SzklarczykDFranceschiniAKuhnMSimonovicMRothAMinguezPDoerksTStarkMMullerJBorkPThe STRING database in 2011: Functional interaction networks of proteins, globally integrated and scoredNucleic Acids Res39D561D568201110.1093/nar/gkq97321045058 SzklarczykDMorrisJHCookHKuhnMWyderSSimonovicMSantosADonchevaNTRothABorkPThe STRING database in 2017: Quality-controlled protein-protein association networks, made broadly accessibleNucleic Acids Res45D362D368201710.1093/nar/gkw93727924014 TangZLiCKangBGaoGLiCZhangZGEPIA: A web server for cancer and normal gene expression profiling and interactive analysesNucleic Acids Res45W98W102201710.1093/nar/gkx24728407145 PonténFSchwenkJMAsplundAEdqvistPHThe human protein atlas as a proteomic resource for biomarker discoveryJ Intern Med270428446201110.1111/j.1365-2796.2011.02427.x21752111 CeramiEGaoJDogrusozUGrossBESumerSOAksoyBAJacobsenAByrneCJHeuerMLLarssonEThe cBio cancer genomics portal: An open platform for exploring multidimensional cancer genomics dataCancer Discov2401404201210.1158/2159-8290.CD-12-009522588877 LiTYinJYuanLWangSYangLDuXLuJDownregulation of microRNA-139 is associated with hepatocellular carcinoma risk and short-term survivalOncol Rep3116991706201410.3892/or.2014.303224549282 JiaoWZhangJWeiYFengJMaMZhaoHWangLJiaoWMiR-139-5p regulates VEGFR and downstream signaling pathways to inhibit the development of esophageal cancerDig Liver Dis51149156201910.1016/j.dld.2018.07.01730245290 QinLDengHYChenSJWeiWZhangYTmiR-139 acts as a tumor suppressor in T-cell acute lymphoblastic leukemia by targeting CX chemokine receptor 4Am J Transl Res940594070201728979681 ChenJYuYChenXHeYHuQLiHHanQRenFLiJLiCMiR-139-5p is associated with poor prognosis and regulates glycolysis by repressing PKM2 in gallbladder carcinomaCell Prolif51e12510201810.1111/cpr.1251030105813 ZhangHDSunDWMaoLZhangJJiangLHLiJWuYJiHChenWWangJMiR-139-5p inhibits the biological function of breast cancer cells by targeting Notch1 and mediates chemosensitivity to docetaxelBiochem Biophys Res Commun465702713201510.1016/j.bbrc.2015.08.05326299922 RenYZhuHChiCYangFXuXMiRNA-139 regulates oral cancer Tca8113 cells apoptosis through Akt signaling pathwayInt J Clin Exp Pathol845884594201526191149 AuSLWongCCLeeJMFanDNTsangFHNgIOWongCMEnhancer of zeste homolog 2 epigenetically silences multiple tumor suppressor microRNAs to promote liver cancer metastasisHepatology56622631201210.1002/hep.2567922370893 XuKShenKLiangXLiYNagaoNLiJLiuJYinPMiR-139-5p reverses CD44+/CD133+-associated multidrug resistance by downregulating NOTCH1 in colorectal carcinoma cellsOncotarget77511875129201610.18632/oncotarget.1261127738333 YoonELYeonJEKoELeeHJJeJHYooYJKangSHSuhSJKimJHSeoYSAn explorative analysis for the role of serum miR-10b-3p levels in predicting response to sorafenib in patients with advanced hepatocellular carcinomaJ Korean Med Sci32212220201710.3346/jkms.2017.32.2.21228049231 HuangLLHuangLWWangLTongBDWeiQDingXSPotential role of miR-139-5p in cancer diagnosis, prognosis and therapyOncol Lett1412151222201710.3892/ol.2017.635128789336 CoxADDerCJThe dark side of Ras: Regulation of apoptosisOncogene2289999006200310.1038/sj.onc.120711114663478 BosJLRas oncogenes in human cancer: A reviewCancer Res494682468919892547513 BosJLThe ras gene family and human carcinogenesisMutat Res195255271198810.1016/0165-1110(88)90004-83283542 PriorIALewisPDMattosCA comprehensive survey of Ras mutations in cancerCancer Res7224572467201210.1158/0008-5472.CAN-11-261222589270 SchwarzRRamerRHinzBTargeting the endocannabinoid system as a potential anticancer approachDrug Metab Rev502653201810.1080/03602532.2018.142834429390896 LuYAndersonHDCannabinoid signaling in health and diseaseCan J Physiol Pharmacol95311327201710.1139/cjpp-2016-034628263083 VagoRBettigaASaloniaACiuffredaPOttriaRDevelopment of new inhibitors for N-acylethanolamine-hydrolyzing acid amidase as promising tool against bladder cancerBioorg Med Chem2512421249201710.1016/j.bmc.2016.12.04228062195 Martínez-MartínezEMartín-RuizAMartínPCalvoVProvencioMGarcíaJMCB2 cannabinoid receptor activation promotes colon cancer progression via AKT/GSK3β signaling pathwayOncotarget76878168791201610.18632/oncotarget.1196827634891 LiontosMKoutsamiMSideridouMEvangelouKKletsasDLevyBKotsinasANahumOZoumpourlisVKouloukoussaMDeregulated overexpression of hCdt1 and hCdc6 promotes malignant behaviorCancer Res671089910909200710.1158/0008-5472.CAN-07-283718006835 BravouVNishitaniHSongSYTaravirasSVarakisJExpression of the licensing factors, Cdt1 and Geminin, in human colon cancerInt J Oncol2715111518200516273206 ArentsonEFaloonPSeoJMoonEStudtsJMFremontDHChoiKOncogenic potential of the DNA replication licensing protein CDT1Oncogene2111501158200210.1038/sj.onc.120517511850834 KaraviasDMaroulisIPapadakiHGogosCKakkosSKaraviasDBravouVOverexpression of CDT1 is a predictor of poor survival in patients with hepatocellular carcinomaJ Gastrointest Surg20568579201610.1007/s11605-015-2960-726408331 YuZWangRChenFWangJHuangXFive novel oncogenic signatures could be utilized as AFP-related diagnostic biomarkers for hepatocellular carcinoma based on next-generation sequencingDig Dis Sci63945957201810.1007/s10620-018-4961-329442275 ZhongQChenCFLiSChenYWangCCXiaoJChenPLSharpZDLeeWHAssociation of BRCA1 with the hRad50-hMre11-p95 complex and the DNA damage responseScience285747750199910.1126/science.285.5428.74710426999 ChenCCChenCYUengSHHsuehCYehCTHoJYChouLFWangTHCorylin increases the sensitivity of hepatocellular carcinoma cells to chemotherapy through long noncoding RNA RAD51-AS1-mediated inhibition of DNA repairCell Death Dis9543201810.1038/s41419-018-0575-029749376 ChenCCChenCYWangSHYehCTSuSCUengSHChuangWYHsuehCWangTHMelatonin sensitizes hepatocellular carcinoma cells to chemotherapy through long non-coding RNA RAD51-AS1-mediated suppression of DNA repairCancers (Basel)10201810.3390/cancers10090320 ShaoJXuZPengXChenMZhuYXuLZhuHYangBLuoPHeQGefitinib synergizes with irinotecan to suppress hepatocellular carcinoma via antagonizing Rad51-mediated DNA-repairPLoS One11e0146968201610.1371/journal.pone.014696826752698 LuoJSiZZLiTLiJQZhangZQChenGSQiHZYaoHLMicroRNA-146a-5p enhances radiosensitivity in hepatocellular carcinoma through replication protein A3-induced activation of the DNA repair pathwayAm J Physiol Cell Physiol316C299C311201910.1152/ajpcell.00189.201830462539 ChandVPandeyAKopanjaDGuzmanGRaychaudhuriPOpposing roles of the forkhead box factors FoxM1 and FoxA2 in liver cancerMol Cancer Res1710631074201910.1158/1541-7786.MCR-18-096830814128 LiangCZhaoJGeHLiGWuJClinicopathological and prognostic significance of FoxM1 in hepatocellular carcinoma patients: A meta-analysisOnco Targets Ther1135613571201810.2147/OTT.S15554129950861 TianCWuHLiCTianXSunYLiuELiaoXSongWDownreguation of FoxM1 by miR-214 inhibits proliferation and migration in hepatocellular carcinomaGene Ther25312319201810.1038/s41434-018-0029-429973656 SongBNChuISA gene expression signature of FOXM1 predicts the prognosis of hepatocellular carcinomaExp Mol Med50e418201810.1038/emm.2017.15929303511 LinPHeRQDangYWWenDYMaJHeYChenGYangHAn autophagy-related gene expression signature for survival prediction in multiple cohorts of hepatocellular carcinoma patientsOncotarget91736817395201810.18632/oncotarget.2408929707114 ChangYZhouCFanLQiuGWangGWeiGChangXLiXUpregulation of microRNA-300 induces the proliferation of liver cancer by downregulating transcription factor FOXO1Oncol Rep4035613572201830272296 XuHLiGYueZLiCHCV core protein-induced upregulation of microRNA-196a promotes aberrant proliferation in hepatocellular carcinoma by targeting FOXO1Mol Med Rep1352235229201610.3892/mmr.2016.515927108614 YangXWShenGZCaoLQJiangXFPengHPShenGChenDXuePMicroRNA-1269 promotes proliferation in human hepatocellular carcinoma via downregulation of FOXO1BMC Cancer14909201410.1186/1471-2407-14-90925472505 XuXYamamotoHSakonMYasuiMNganCYFukunagaHMoritaTOgawaMNaganoHNakamoriSOverexpression of CDC25A phosphatase is associated with hypergrowth activity and poor prognosis of human hepatocellular carcinomasClin Cancer Res917641772200312738732 XuXYamamotoHLiuGItoYNganCYKondoMNaganoHDonoKSekimotoMMondenMCDC25A inhibition suppresses the growth and invasion of human hepatocellular carcinoma cellsInt J Mol Med21145152200818204780 KellyPMoellerBJJunejaJBoodenMADerCJDaakaYDewhirstMWFieldsTACaseyPJThe G12 family of heterotrimeric G proteins promotes breast cancer invasion and metastasisProc Natl Acad Sci USA10381738178200610.1073/pnas.051025410316705036 KellyPStemmleLNMaddenJFFieldsTADaakaYCaseyPJA role for the G12 family of heterotrimeric G proteins in prostate cancer invasionJ Biol Chem2812648326490200610.1074/jbc.M60437620016787920 ChiaCYKumariUCaseyPJBreast cancer cell invasion mediated by Gα12 signaling involves expression of interleukins-6 and −8, and matrix metalloproteinase-2J Mol Signal96201410.1186/1750-2187-9-624976858 UdayappanUKCaseyPJc-Jun contributes to transcriptional control of GNA12 expression in prostate cancer cellsMolecules22201710.3390/molecules2204061228394299 BakiriLHamacherRGrañaOGuío-CarriónACampos-OlivasRMartinezLDienesHPThomsenMKHasenfussSCWagnerEFLiver carcinogenesis by FOS-dependent inflammation and cholesterol dysregulationJ Exp Med21413871409201710.1084/jem.2016093528356389 WatanabeTHiasaYTokumotoYHirookaMAbeMIkedaYMatsuuraBChungRTOnjiMProtein kinase R modulates c-Fos and c-Jun signaling to promote proliferation of hepatocellular carcinoma with hepatitis C virus infectionPLoS One8e67750201310.1371/journal.pone.006775023844083 FanQHeMDengXWuWKZhaoLTangJWenGSunXLiuYDerepression of c-Fos caused by microRNA-139 down-regulation contributes to the metastasis of human hepatocellular carcinomaCell Biochem Funct31319324201310.1002/cbf.290223001723

Flow chart illustrating the process for selecting published studies and microarray data. HCC, hepatocellular carcinoma; miRNA, microRNA.

Differential expressive situation of miR-139-5p between HCC and nontumor liver tissues established using 18 microarray and miRNA sequencing datasets. (A) TCGA; (B) GSE12717; (C) GSE21279; (D) GSE21362; (E) GSE36915; (F) GSE39678; (G) GSE40744; (H) GSE41874; (I) GSE50013; (J) GSE54751; (K) GSE64632; (L) GSE67882; (M) GSE69580; (N) GSE98269-GPL20712; (O) GSE115016; (P) E-MTAB-511; (Q) E-MTAB-3347; (R) E-MTAB-4809. HCC, hepatocellular carcinoma; TCGA, The Cancer Genome Atlas; miR, microRNA.

ROC curve of miR-139-5p regulation in HCC and nontumor liver tissues in each analyzed dataset. (A) TCGA; (B) GSE12717; (C) GSE21279; (D) GSE21362; (E) GSE36915; (F) GSE39678; (G) GSE40744; (H) GSE41874; (I) GSE50013; (J) GSE54751; (K) GSE64632; (L) GSE67882; (M) GSE69580; (N) GSE98269-GPL20712; (O) GSE115016; (P) E-MTAB-511; (Q) E-MTAB-3347; (R) E-MTAB-4809. HCC, hepatocellular carcinoma; ROC, receiver operating characteristic; TCGA, The Cancer Genome Atlas; miR, microRNA.

Integrated meta-analysis designed to identify the miR-139-5p expression levels in 18 analyzed datasets including HCC and nontumor liver tissues. (A) Forest plot showing that miR-139-5p is downregulated in HCC (SMD=−0.84; 95% CI: −1.36 to −0.32; P<0.001; I2=85.7%). (B) Symmetry of the funnel plot revealed no evident publication bias among the 18 included datasets. The size of each circle represents the sample size of each dataset. (C) Sensitivity analysis showing the stability of the 18 included datasets. HCC, hepatocellular carcinoma; SMD, standard mean difference; CI, confidence interval; miR, microRNA; I2, inconsistency index.

Significance of miR-139-5p expression in the diagnosis of HCC compared with non-tumor liver tissues in 18 datasets. (A) Merged sensitivity was 0.87 (0.85–0.90). (B) Merged specificity was 0.66 (0.60–0.72). (C) Pooled positive LR was 2.50 (CI: 1.53–4.07). (D) Pooled negative LR was 0.23 (CI: 0.16–0.33). (E) Pooled diagnostic OR was 14.39 (CI: 5.96–34.75). (F) AUC of the sROC curve was 0.8978. HCC, hepatocellular carcinoma; LR, likelihood ratio; CI, confidence interval; OR, odds ratio; AUC, area under the curve; sROC, summary receiver operating characteristic; miR, microRNA.

Integrated meta-analysis for determining miR-139-5p expression in HCC. (A) Forest plot demonstrating a poor prognosis for patients with HCC presenting reduced miR-139-5p expression based on four studies reporting the OS of patients with HCC. (B) Symmetry of the funnel plot revealed no evident publication bias among the four publications. The size of the circle represents the sample size of each study. (C) Sensitivity analysis shows the stability of the four publications. HCC, hepatocellular carcinoma; OS, overall survival; CI, confidence interval; miR, microRNA.

Process for determining predicted downstream target genes. (A) Venn diagram presenting the 4,294 differentially expressed genes obtained from TCGA and 1,406 targeted genes predicted by the miRWalk website; 382 intersecting genes were ultimately identified. (B) Performance of the protein-protein network for validating the connections among the top eight target genes. TCGA, The Cancer Genome Atlas.

Bubble diagrams showing the results of the enrichment analysis and signaling pathway analysis of 382 hub genes regulated by miR-139-5p in HCC. BP of GO analysis; CC of GO analysis; MF of GO analysis; KEGG signaling pathways. HCC, hepatocellular carcinoma; GO, gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; bp, biological process; CC, cellular component; MF, molecular function; miR, microRNA.

Expression of the top eight predicted target genes in HCC and nontumor liver tissues from TCGA. (A) CDC25A; (B) CDT1; (C) EZH2; (D) FOS; (E) FOXM1; (F) FOXO1; (G) GNA12; (H) RAD51. Significant upregulation of CDT1, EZH2, FOXM1 and RAD51 expression and downregulation of FOS in HCC tissues was observed (*P<0.05). No predominant differences were detected among the remaining genes, with P>0.05. LIHC, Liver Hepatocellular Carcinoma; HCC, hepatocellular carcinoma; TCGA, The Cancer Genome Atlas; CDC25A, cell division cycle 25A; CDT1, chromatin licensing and DNA replication factor 1; EZH2, enhancer of zeste homolog 2; FOS, FBJ murine osteosarcoma viral oncogene homolog; FOXM1, forkhead box M1; FOXO1, forkhead box O1; GNA12, guanine nucleotide binding protein alpha 12; RAD51, recombination protein.

Kaplan-Meier curves displaying the associations between the expression of the top target genes and OS of patients with HCC. (A) CDC25A; (B) CDT1; (C) EZH2; (D) FOS; (E) FOXM1; (F) FOXO1; (G) GNA12; (H) RAD51. These results indicate that the high expression of CDT1, EZH2, RAD51, FOXM1, GNA12 and CDC25A had a significant effect on the OS of patients with HCC (P<0.05) while no marked effect was detected in the expression of other genes in HCC (P>0.05). HCC, hepatocellular carcinoma; OS, overall survival; CDC25A, cell division cycle 25A; CDT1, chromatin licensing and DNA replication factor 1; EZH2, enhancer of zeste homolog 2; FOS, FBJ murine osteosarcoma viral oncogene homolog; FOXM1, forkhead box M1; FOXO1, forkhead box O1; GNA12, guanine nucleotide binding protein alpha 12; RAD51, recombination protein; TPM, transcripts per million.

Frequency of mutations in the top eight target genes of 360 patients with HCC from the CBio Cancer Genomics Portal. HCC, hepatocellular carcinoma; CDC25A, cell division cycle 25A; CDT1, chromatin licensing and DNA replication factor 1; EZH2, enhancer of zeste homolog 2; FOS, FBJ murine osteosarcoma viral oncogene homolog; FOXM1, forkhead box M1; FOXO1, forkhead box O1; GNA12, guanine nucleotide binding protein alpha 12; RAD51, recombination protein.

Verification of the levels of the proteins encoded by the four hub genes in normal and HCC tissues using the Human Protein Atlas database. HCC, hepatocellular carcinoma; FOS, FBJ murine osteosarcoma viral oncogene homolog; FOXO1, forkhead box O1; GNA12, guanine nucleotide binding protein alpha 12; RAD51, recombination protein.

Evaluation of the correlation between the expression of the four hub genes and miR-139-5p regulation in patients with hepatocellular carcinoma from The Cancer Genome Atlas. (A) A positive correlation between miR-139-5p and FOS regulation was detected, with a Pearson correlation coefficient of 0.350 and P<0.001. (B) A positive correlation between miR-139-5p and FOXO1 regulation was identified, with a Pearson correlation coefficient of 0.431 and P<0.001. (C) A negative correlation between miR-139-5p and GNA12 regulation was detected, with a Pearson correlation coefficient of −0.285 and P<0.001. (D) A negative correlation between miR-139-5p and RAD51 regulation was observed, with a Pearson correlation coefficient of −0.625 and P<0.001. FOS, FBJ murine osteosarcoma viral oncogene homolog; FOXO1, forkhead box O1; GNA12, guanine nucleotide binding protein alpha 12; RAD51, recombination protein; miR, microRNA.

Major characteristics of the analyzed publications reporting associations with miR-139-5p expressed in HCC.

Study Year Country N Cut-off value Method Sample Follow-up (months) Result HR (95% CI)a Source (Refs.)
Li et al 2014 China 31 NA RT-qPCR Plasma 12 OS OS: 1.56 (0.3–8.07) ED (54)
Wang et al 2016 China 153 NA RT-qPCR Tissue – OS CPP OS: 2.68 (1.42–5.05) ED (25)
Wong et al 2011 China 67 NA RT-qPCR Tissue 60 OS DFS CPP OS: 1.75 (0.36–8.48) ED (29)
DFS: 1.67 (0.57–4.87)
Wang et al 2015 China 116 NA RT-qPCR Tissue – OS RFS OS: 1.195 (0.90–1.57) RD (28)
RFS: 1.20 (0.61–2.37)

Multivariate analysis. miR, microRNA; HCC, hepatocellular carcinoma; NA, not available; RT-qPCR, reverse transcription-quantitative PCR; HR, hazard ratio; OS, overall survival; DE, data-extrapolated; CPP, clinical pathological parameters; ED, extracted; RD, reported.

Essential features of datasets involving miR-139-5p expression founded on GEO, TCGA, Oncomine and Array Express databases.

HCC Non-tumor
Dataset N M SD N M SD t P-value AUC
GSE12717 10 5.9442 1.33223 6 8.2381 0.44433 −4.036 0.001 0.917
GSE21279 4 3.3278 2.49676 12 3.3192 3.49780 0.005 0.996 0.531
GSE21362 73 4.8902 1.51216 73 6.3078 1.02803 −6.624 <0.001 0.833
GSE36915 68 12.0473 1.20972 21 13.5388 0.60042 −7.583 <0.001 0.891
GSE39678 16 7.9983 0.57305 8 8.7452 0.64627 −2.888 0.009 0.828
GSE40744 26 8.5754 1.41834 50 9.4392 0.47678 −3.018 0.005 0.735
GSE41874 3 0.8372 0.06337 4 1.1558 0.06887 −6.252 0.002 1.000
GSE50013 14 1.7929 4.95168 13 0.9769 2.13937 0.548 0.589 0.448
GSE54751 10 0.0110 0.00801 10 0.0435 0.02363 −4.119 0.002 0.920
GSE64632 3 0.3118 0.38826 3 0.1204 0.06350 0.843 0.484 0.444
GSE67882 3 12.1139 1.96061 8 9.9169 3.02709 1.149 0.280 0.250
GSE69580 5 1.3212 1.21214 5 10.8835 3.12460 −6.380 <0.001 1.000
GSE98269 3 5.4040 0.25288 3 5.9628 0.14467 −3.322 0.029 1.000
GSE115016 12 9.5731 12.11435 12 48.9893 26.09749 −4.746 <0.001 0.944
TCGA 371 6.4189 1.23762 49 8.5117 0.45390 −22.926 <0.001 0.953
E-MTAB-511 8 8.1250 10.04899 10 6.7000 9.35771 0.311 0.760 0.450
E-MTAB-3347 4 6.7444 0.34443 4 6.8057 0.38966 −0.236 0.821 0.500
E-MTAB-4809 12 9.2651 1.27329 12 5.1714 1.42375 7.424 <0.001 0.000

GEO, Gene Expression Omnibus; TCGA, The Cancer Genome Atlas; N, number; M, mean; SD, standard deviation; AUC, area under the curve; HCC, hepatocellular carcinoma.

Clinicopathological parameters of miRNA-139-5p regulation in HCC tissues from TCGA.

MiR-139-5p relevant expression (log2x)
Clinicopathological features n Mean ± SD t P-value
Tissue −22.926 <0.01
  HCC 371 6.4189±1.2376
  Normal tissue 49 8.5117±0.4539
Age (years) −0.059 0.953
  ≥60 200 6.5022±1.4621
  <60 170 6.5107±1.3107
Sex 1.196 0.233
  Female 119 6.4136±1.4532
  Male 251 6.5982±1.2365
Race F=0.662 0.576
  White 181 6.5850±1.4719
  Black or African American 17 6.6571±0.9780
  Asian 161 6.4148±1.2035
  American Indian or Alaska native 2 5.9092±2.1116
Cancer status −1.605 0.109
  Tumor free 202 6.6343±1.3739
  With tumor 150 6.3997±1.3318
Grade F=8.305 <0.001
  G1 55 7.1548±1.1886
  G2 173 6.6150±1.3285
  G3 124 6.1845±1.2038
  G4 13 6.1198±1.1988
Pathological stage F=7.575 <0.001
  Stage I 172 6.8524±1.1347
  Stage II 85 6.2293±1.3337
  Stage III 85 6.1721±1.4300
  Stage IV 5 6.3922±1.6006
T stage −2.595 0.010
  T1-2 275 6.6139±1.3039
  T3-4 93 6.1970±1.4403
Node involvement 1.506 0.133
  No 232 6.3513±1.5065
  Yes 4 7.4924±1.1574
Metastasis −0.554 0.580
  No 267 6.4695±1.3007
  Yes 4 6.1048±1.6927
Survival status −1.756 0.080
  Dead 129 6.3757±1.4070
  Alive 241 6.6261±1.2509

HCC, hepatocellular carcinoma; TCGA, The Cancer Genome Atlas; N, number; SD, standard deviation; F, One-way analysis of variance (ANOVA) test was performed to assess the difference for >2 groups.

Merged HRs and corresponding 95% CIs for OS in patients with HCC obtained using a meta-analysis based on fixed-effects model.

Study Year HR LL UL p I2 (%) P-value (Refs.)
Li et al 2014 1.56 0.30 8.07 (54)
Wang et al 2016 2.68 1.42 5.05 (25)
Wong et al 2011 1.75 0.36 8.48 (29)
Wang et al 2015 1.20 0.91 1.58 (28)
Pooled 1.37 1.07 1.76 0.001 43.9 0.148

P, P<0.05 and the 95% CI not exceeding 1 indicated HR was considered significant (HR=1.37; 95% CI: 1.07–1.76; P=0.001); P, P>0.05 indicated no significant heterogeneity was observed using fixed-effects mode (P=0.1481, I2=43.9%). HR, hazard ratio; CI, confidence interval; OS, overall survival; HCC, hepatocellular carcinoma; LL, lower confidence limit; UL, upper confidence limit; I2, inconsistency index

Enrichment of downstream target genes regulated by miR-139-5p in Gene Ontology terms in the categories of biological processes, cellular components and molecular functionsa.

Term Counts P-value
Biological process
  Response to drug 16 1.19×10−03
  Signal transduction 39 1.72×10−03
  Cell migration involved in sprouting angiogenesis 4 2.97×10−03
  Positive regulation of cell proliferation 19 5.81×10−03
  Positive regulation of neuron projection development 7 8.64×10−03
  Positive regulation of GTPase activity 21 9.52×10−03
  Regulation of insulin secretion 6 1.07×10−02
  Response to insulin 6 1.07×10−02
  Activation of protein kinase activity 5 1.21×10−02
  Positive regulation of transcription, DNA-templated 19 1.52×10−02
Cellular component
  Plasma membrane 111 1.30×10−04
  Integral component of plasma membrane 48 2.12×10−04
  Neuron projection 15 2.35×10−04
  Integral component of membrane 125 3.30×10−03
  Neuronal cell body 14 9.02×10−03
  External side of plasma membrane 11 9.07×10−03
  Postsynaptic membrane 10 2.27×10−02
  Anchored component of membrane 7 2.30×10−02
  Z disc 7 2.77×10−02
  Terminal bouton 5 3.24×10−02
Molecular function
  Transcription factor activity, sequence-specific DNA binding 36 4.37×10−04
  Ras guanyl-nucleotide exchange factor activity 8 8.28×10−03
  Steroid binding 4 1.62×10−02
  mRNA 3′-UTR binding 5 1.63×10−02
  Sequence-specific DNA binding 19 1.65×10−02
  Transcriptional activator activity, RNA polymerase II core promoter proximal region sequence-specific binding 11 2.05×10−02
  Kinase activity 11 2.33×10−02
  Pyridoxal phosphate binding 5 2.69×10−02
  DNA binding 45 3.55×10−02
  Calmodulin binding 9 3.62×10−02

P<0.05. GO, Gene Ontology.

Downstream predicted genes of miR-139-5p are principally concentrated in five KEGG pathways and five Panther pathwaysa.

Term P-value Counts Genes
hsa05200:Pathways in cancer 6.01×10−04 20 CKS1B, AR, RET, E2F3, PDGFB, GNA12, FOXO1, KITLG, FGF13, FGF12, CTNNA3, RAD51, TPM3, EDNRB, FOS, WNT7B, PAX8, TGFA, GNG4, PLCB1
hsa04720:Long-term potentiation 1.27×10−02 6 GRM5, RPS6KA6, GRIN2B, CAMK4, RAPGEF3, PLCB1
hsa04014:Ras signaling pathway 2.16×10−02 11 LAT, GAB2, GRIN2B, RASGRF2, PDGFB, EFNA3, KITLG, FGF13, FGF12, GNG4, ABL2
hsa04080:Neuroactive ligand-receptor interaction 3.31×10−02 12 GRM5, EDNRB, GABRG3, PTGIR, ADORA3, GRIN2B, P2RY4, CNR1, ADRA1A, CHRNB2, HTR1D, GABRQ
hsa04015:Rap1 signaling pathway 3.42×10−02 10 LAT, GRIN2B, PDGFB, CNR1, EFNA3, KITLG, FGF13, FGF12, RAPGEF3, PLCB1
P00026:Heterotrimeric G-protein signaling pathway-Gi alpha and Gs alpha-mediated pathway 2.26×10−02 11 GRM5, GNAL, RET, ADORA3, FGB, RGS4, RGS5, GPSM2, ADRA1A, HTR1D, PYGB
P05730:Endogenous cannabinoid signaling 3.57×10−02 4 GRM5, CNR1, PLCB1, GNG4
P00027:Heterotrimeric G-protein signaling pathway-Gq alpha and Go alpha-mediated pathway 5.59×10−02 9 GRM5, ADORA3, RGS4, RGS5, GPSM2, RHOB, PLCB1, GNG4, RHOF
P05911:Angiotensin II-stimulated signaling through G proteins and beta-arrestin 5.81×10−02 5 EGR1, RHOB, PLCB1, GNG4, RHOF
P00008:Axon guidance mediated by Slit/Robo 6.51×10−02 4 ROBO1, RHOB, ABL2, RHOF

P<0.05. KEGG, Kyoto Encyclopedia of Genes and Genomes.