miR-30a-GNG2 and miR-15b-ACSS2 Interaction Pairs May Be Potentially Crucial for Development of Abdominal Aortic Aneurysm by Influencing Inflammation

Abstract

Abdominal aortic aneurysm (AAA) is a lethal vascular degenerative disease for the elderly, but current therapeutic options are limited. This study was to explore the molecular mechanisms of AAA to screen underlying treatment targets for AAA. The gene and microRNA (miRNA) expression profiles of human AAA were downloaded from Gene Expression Omnibus database under accession number GSE57691, GSE62179, and GSE63541. Differentially expressed genes (DEGs) and microRNAs (miRNAs; DEMs) were identified using the Linear Models for Microarray data method. Protein-protein interaction (PPI) network, module analysis, and miRNA-mRNA regulatory network analyses were performed to screen hub genes and miRNAs that regulated the hub genes. The Database for Annotation, Visualization and Integrated Discovery was used to predict the functions of genes. GEPIA and Tumor-miRNA-Pathway online software were used to validate the expressions of crucial DEMs and DEGs in other cancers, respectively. As a result, in the GSE57691 dataset, a total of 584 DEGs were found to be specific for AAA, 521 of which were used for constructing the PPI network. ACSS2 (acyl-CoA synthetase short-chain family member 2), GNG2 (G protein subunit gamma 2), and CXCL1 (C-X-C motif chemokine ligand 1) and CCR7 (C-C motif chemokine receptor 7) were believed to be hub genes by calculating their topological features in the PPI network. Upregulated GNG2 could interact with CXCL1 and CCR7 to involve in chemokine signaling pathway, while downregulated ACSS2 was associated with lipid biosynthetic process. In the miRNA-mRNA regulatory network, ACSS2 was found to be regulated by hsa-miR-15b; hsa-miR-30a could modulate the expression of GNG2. In line with our analysis in AAA, GNG2, ACSS2, hsa-miR-30a, and hsa-miR-15b were also confirmed to be significantly upregulated or downregulated in several cancer types. In conclusion, hsa-miR-30a-GNG2 and hsa-miR-15b-ACSS2 interaction pairs may represent novel mechanisms for explaining the pathogenesis of AAA. Targeted regulation of them may be potential strategies for treatment of AAA.

Introduction

Abdominal aortic aneurysm (AAA) is a common chronic vascular degenerative disease affecting about 9% of elderly men and 1% of women >65 years of age (Dereziński et al., 2017). AAA is characterized by progressive dilation of abdominal aorta (1.5 times greater than normal) due to a loss of the structural integrity of the vascular wall (Kent, 2014). Once the aortic diameter reaches the size threshold, acute rupture of AAA may possibly occur, which results in ∼80% of patients succumbing to mortality (Johansson and Swedenborg, 2010; Kent, 2014). Surgical repair is the main treatment choice for AAA, but the 5-year all-cause mortality rate remains higher (about 32%) (Lieberg et al., 2017). Therefore, there is an urgent need for the development of novel therapeutic options.

Although the pathogenesis of AAA is complex, accumulating evidence suggests that vascular smooth muscle cells (VSMCs) play important roles. VSMCs have the ability to synthesize extracellular matrix (ECM, such as collagen and elastin) to maintain the elasticity and pressure resistance of the aortic wall; therefore, the apoptosis of VSMCs can result in the degeneration of aortic walls and eventual development of AAA (Zhang et al., 2019). Infiltration of inflammatory cells (such as macrophages and T lymphocytes) is suggested to be a potential contributor for the elimination of VSMCs (Henderson et al., 1999). We, hereby, hypothesize inhibition of apoptosis or inflammation of VSMCs may be an underlying approach for treatment of AAA.

It is reported that microRNAs (miRNAs), small noncoding RNA molecules (approximately 18–25 nucleotides in length), can play crucial roles in the apoptosis or inflammation of VSMCs and the formation of AAA by negatively regulating the expression of their target genes at the posttranscriptional level. For example, Liang et al. showed the level of miR-195 was significantly increased in AAA tissues compared with the normal controls. miR-195 inhibited the proliferation and induced the apoptosis of VSMCs by indirectly downregulating SMAD family member 3 (Smad3), a key intracellular mediator of the transforming growth factor beta (TGF-β) signal pathway, for regulating ECM deposition. Overexpression of Smad3 reversed the effects of miR-195 (Liang et al., 2017). Cao et al. (2017) identified that miR-504 was downregulated in the aortic cells derived from patients with AAA when compared with controls. Transfection of VSMCs with pMSCV-miR-504 vector significantly promoted the proliferation of VSMCs and inhibited the expression levels of apoptosis-associated genes (caspase-3/9, p53 and p21), compared with nontargeting controls (Cao et al., 2017). Zhang et al. showed miR-155 was upregulated in AAA patients. Overexpression of miR-155 promoted the activation of macrophage inflammasome and then the release of cytokines, such interleukin (IL)-1, IL-6, and tumor necrosis factor (TNF)-α, while downregulation of miR-155 by using the antagomir decreased the accumulation of CD68 and elastic collagen-positive areas, which indicated the improvement of AAA (Zhang et al., 2018b). Nakao et al. (2017) observed genetic ablation of miR-33-attenuated AAA formation by inhibiting inflammatory pathways. There were also some studies to use the microarray technique to investigate the human AAA-related miRNA (Pahl et al., 2012; Cheuk and Cheng, 2014) or messenger RNA (mRNA) expression profiles (Choke et al., 2010; Wang et al., 2017) and demonstrated that the target genes of several differentially expressed miRNAs (DEMs) or the differentially expressed genes (DEGs) were involved in apoptosis and activation of inflammatory cells. These studies implied that the regulation of miRNAs or mRNAs that were associated with apoptosis or inflammation of VSMCs may be helpful for treatment of AAA. However, the crucial interaction pairs by integrated analysis of DEMs and DEGs remain rarely reported for AAA (Tian et al., 2018).

In this study, we aimed to download the relatively newly deposited microarray datasets of miRNA and mRNA expression profiles from the Gene Expression Omnibus (GEO) database compared with the study of Tian et al. (2018), and screen crucial DEMs and DEGs by constructing the protein-protein interaction (PPI) and miRNA-mRNA regulatory networks. The identified DEMs and DEGs were also validated in other cancers by using the Cancer Genome Atlas (TCGA) data. Our results may provide novel targets for treatment of AAA.

Materials and Methods

Collection of microarray datasets

The expression profiles of human AAA were downloaded from GEO database. GSE57691 microarray dataset (Erik et al., 2015) analyzed the mRNA expression profile in aortic samples from 49 (female: male, 2:47) AAA patients, 9 (female: male, 1:8) aortic occlusive disease (AOD) patients, and 10 (female: male, 4:6) controls. GSE62179 microarray dataset investigated the miRNA expression profile in aortic samples of eight AAA patients and two controls. GSE63541 microarray dataset (Spear et al., 2015) explored the miRNA expression profile in aortic samples from two AAA patients and two controls. Human AAA biopsy samples were harvested during open surgery to treat AAA. Human abdominal nonaneurysmal aortas were harvested from deceased patients providing organ retrievals and transplantation as controls.

Data preprocessing and expression analysis

The raw expression data in TXT format were downloaded from GEO database using the corresponding accession number. Array expression data were first log2 transformed to make them approximate to normal distribution and then quantile-normalized using the Linear Models for Microarray data (LIMMA) method (Ritchie et al., 2015) (version 3.34.0) in the Bioconductor R package (version 3.4.1).

The LIMMA method (Ritchie et al., 2015) with t-statistics was used to estimate the logFC (fold change) and p-value for each gene in AAA tissues compared with the controls. p-Value was adjusted to false discovery rate (FDR) using Benjamini-Hochberg method to avoid false positives. The DEGs and DEMs were defined as |logFC| > 1 and FDR <0.05. Bidirectional hierarchical clustering was performed for the DEGs and DEMs using the pheatmap in R package (version: 1.0.8) to evaluate their distinguishing performance for different sample groups.

PPI network construction and module analysis

The interaction relationships between DEGs were curated from the Search Tool for the Retrieval of Interacting Genes (STRING; version 10.0) database (Szklarczyk et al., 2015). Only the gene pairs with confidence score ≥0.8 were considered to be significant and utilized to construct the PPI network using the Cytoscape software (version 3.6.1) (Kohl et al., 2011). The hub genes were selected from the PPI network if they were in the top 10% genes according to four ranking methods: degree centrality (DC), betweenness centrality (BC), closeness centrality (CNC), and average path length (APL), which were calculated by CytoNCA plugin in the Cytoscape software (Tang et al., 2015b). Moreover, the network was divided into several function-related modules using the Molecular Complex Detection (MCODE; version, 1.4.2) (Bader and Hogue, 2003) plugin of the Cytoscape software based on the threshold value of MCODE score >3 and a node number >5.

DEMs-DEGs interaction prediction

The miRWalk database (version 2.0) (Dweep and Gretz, 2015) is the biggest collection of predicted and experimentally verified miR-target interactions using 12 algorithms (miRWalk, MicroT4, miRanda, miRBridge, miRDB, miRMap, miRNAMap, PICTAR2, PITA, RNA22, RNAhybrid, and Targetscan). The target genes of DEMs were predicted using this miRWalk2.0 database by at least one algorithm, the results of which were then overlapped with the DEGs. Only the negative expression relationships between DEMs and DEGs were collected for constructing the DEMs-DEGs network.

Function enrichment analysis

The Database for Annotation, Visualization and Integrated Discovery (DAVID) online tool (version 6.8) (Huang et al., 2009) was used for predicting the possible functions of DEGs, by which the gene ontology (GO) terms (including cellular component [CC], biological process [BP], and molecular function [MF]) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were enriched. GO and KEGG pathways with p < 0.05 were considered to be statistically significant.

Validation of crucial DEGs and DEMs in other cancers

GEPIA (Tang et al., 2017) is a newly developed interactive web server for analyzing the RNA sequencing expression data of 9736 tumors and 8587 normal samples in TCGA and the Genotype-Tissue Expression projects. In addition, tumor-miRNA-pathway (Ma et al., 2016) is a database with user-friendly web interface to display the expression profiles of miRNAs in the 20 tumor types in TCGA. The validation of the expression of crucial genes was performed by GEPIA and tumor-miRNA-pathway analyses.

Results

Identification of DEGs and DEMs

After preprocessing (Fig. 1A, B), a total of 2170 DEGs (150 upregulated and 2020 downregulated) were identified between AAA tissues and controls; and 1757 DEGs (92 upregulated and 1665 downregulated) were identified between AOD tissues and controls (Fig. 2A) in the GSE57691 dataset; after comparing the DEGs in these 2 groups, 584 DEGs were found to be specific for AAA, which were used for the subsequent analysis.

FIG. 1.

Boxplot before and after normalization. (A, B) GSE57691 microarray dataset; (C, D) GSE62179 microarray dataset; (E, F) GSE63541 microarray dataset.

FIG. 2.

Volcano plot (A) and heat map (B) of differentially expressed genes and miRNAs. (A), green points indicated differentially expressed genes or miRNAs, and black plots indicated no significant difference. (B), red represented high expression, while blue represented low expression. Four volcano plots and heat maps in (A, B) were corresponding to three datasets, the FC; AAA; AOD; CTRL. AAA, abdominal aortic aneurysm; AOD, aortic occlusive disease; CTRL, control; FC, fold change; miRNAs, microRNAs. Color images are available online.

After preprocessing (Fig. 1C, D), a total of 63 DEMs were identified between AAA tissues and controls, including 30 upregulated and 33 downregulated miRNAs (Fig. 2A) in the GSE62179 dataset; following preprocessing (Fig. 1E, F), a total of 55 DEMs were identified from GSE63541, consisting of 32 upregulated and 23 downregulated miRNAs between AAA tissues and controls (Fig. 2A); after comparing the DEMs in GSE62179 and GSE63541, 27 were shown to be common and 22 of them exhibited the similar expression trend (Table 1). The heat map was plotted to represent the differentially expressed profiles of DEGs or DEMs in AAA and control samples (Fig. 2B).

Table 1.

Differentially Expressed microRNAs and Genes

	GSE62179		GSE63541			GSE57691
	logFC	FDR	logFC	FDR		logFC	FDR
hsa-miR-29a	−1.44	2.58E-02	1.43	1.51E-02	HBA2	3.82	4.29E-07
hsa-miR-580	−1.34	4.06E-02	−1.04	3.18E-02	HBB	3.67	1.60E-06
hsa-let-7f	−1.28	3.79E-02	1.43	1.89E-02	FOSB	2.41	5.10E-03
hsa-miR-30a	−1.22	3.60E-02	−1.05	3.78E-02	IL-6	2.32	9.44E-03
hsa-miR-503	−1.15	2.38E-03	−1.10	3.47E-02	SIK1	2.21	6.20E-03
hsa-miR-1270	−1.07	4.88E-03	−1.37	3.41E-02	HBG2	1.99	1.18E-02
hsa-miR-629	1.12	3.01E-03	−1.08	3.29E-02	HBG1	1.86	2.10E-02
hsa-miR-142-5p	1.21	9.44E-03	1.41	1.18E-02	FCN1	1.80	4.21E-04
hsa-miR-494	1.22	2.10E-02	1.30	2.09E-02	CXCL8	1.77	4.20E-02
hsa-let-7g	1.23	4.05E-02	1.48	9.72E-03	HLA-DQB1	1.75	3.57E-02
hsa-miR-342-3p	1.26	3.94E-02	1.49	1.05E-02	ACTB	1.75	2.25E-02
hsa-miR-432	1.29	2.26E-02	−1.38	3.60E-02	GNG2	1.03	2.85E-02
hsa-miR-197	1.30	1.36E-02	1.13	4.64E-03	CDC16	−1.08	3.20E-02
hsa-let-7i	1.31	3.64E-02	1.38	1.41E-02	PPIE	−1.07	1.06E-02
hsa-miR-20a	1.31	2.74E-02	1.31	1.45E-02	ACSS2	−1.19	6.59E-03
hsa-miR-621	1.32	4.00E-02	−1.04	3.49E-02	MRPS9	−1.04	1.39E-02
hsa-miR-19b	1.34	2.53E-02	1.41	1.34E-02	RPL26	−2.34	3.08E-03
hsa-miR-885-5p	1.41	3.52E-03	1.07	2.88E-03	SCRG1	−2.34	1.80E-03
hsa-miR-21	1.48	3.68E-02	1.42	2.14E-02	PPP1R3C	−2.37	5.06E-04
hsa-miR-16	1.51	3.30E-02	1.44	9.66E-03	MT1X	−2.38	1.24E-03
hsa-miR-15b	1.53	1.36E-02	1.43	1.61E-02	RPL21	−2.46	1.41E-02
hsa-miR-126	1.56	1.05E-02	1.65	2.35E-02	RPS17	−2.64	1.07E-02
hsa-miR-150	1.62	2.66E-03	1.87	5.28E-03	MT1M	−2.83	2.45E-10
hsa-miR-1246	1.65	7.24E-03	1.40	2.47E-02	S100A4	−2.88	1.66E-02
hsa-miR-223	1.68	2.02E-02	1.42	1.43E-02	RPL10A	−2.91	3.29E-03
hsa-miR-142-3p	1.87	3.92E-03	1.89	7.18E-03	CTSZ	−2.91	1.74E-05
hsa-miR-451	2.45	7.48E-033	1.50	2.53E-02	RNA28S5	−3.08	1.29E-02

The common differentially expressed miRNAs in two datasets are listed. The top 10 and crucial differentially expressed genes are shown.

ACSS2, acyl-CoA synthetase short-chain family member 2; CDC16, cell division cycle 16; FC, fold change; FDR, false discovery rate; GNG2, G protein subunit gamma 2; IL, interleukin; miRNAs, microRNAs.

PPI network

Based on the threshold value of confidence score ≥0.8, 1745 interaction pairs between 521 DEGs (77 upregulated and 444 downregulated genes) were screened from the STRING database and then used to construct the PPI network. In this network, GNG2 (G protein subunit gamma 2) could interact with CXCL1 (C-X-C motif chemokine ligand 1) and CCR7 (C-C motif chemokine receptor 7) (Fig. 3). Among these 521 DEGs, ACSS2 (acyl-CoA synthetase short-chain family member 2) was considered to be a hub gene because it was overlapped in the top 50 genes according to the 3 ranking methods (BC, CNC, and APL); CDC16 (cell division cycle 16) and GNG2 may also act as hub genes because they were overlapped in DC and BC ranking; CXCL1 and CCR7 were believed to be important according to the DC ranking (Table 2).

FIG. 3.

The protein and protein interaction network. Red, upregulated; green, downregulated. The larger size of node (protein) indicated the higher degree (interaction relationships) of it. Color images are available online.

Table 2.

Hub Genes Ranked by Four Topological Features

Gene	DC	Gene	CNC	Gene	BC	Gene	APL
GNG2	55	COL13A1	1	COL13A1	1	COL13A1	1
HSPA8	38	SNTA1	1	NTF3	1	NTF3	1
NHP2L1	38	NTF3	1	GPC4	1	GPC4	1
ACACA	37	GPC4	1	CHPT1	1	CHPT1	1
CDC16	32	CHPT1	1	ATP5J2	1	ATP5J2	1
PPP2R1A	31	ATP5J2	1	SNTA1	0.666667	SNTA1	1
RPS20	31	CNOT8	1	PIGN	0.666667	CNOT8	1
RPS11	30	CNOT7	1	PIGK	0.666667	CNOT7	1
PPIE	30	CPEB3	1	CD7	0.666667	CPEB3	1
PIK3CD	30	ADAMTS8	1	RAB40B	0.6	ADAMTS8	1
CDC23	30	NAV1	1	FUT4	0.571429	NAV1	1
CCR7	29	TDG	1	GDI2	0.533333	TDG	1
ITCH	29	SMUG1	1	ISCA1	0.533333	SMUG1	1
ANAPC5	29	ZNF362	1	ACSS2	0.5	ZNF362	1
MRPS9	27	PNMA1	1	ST3GAL3	0.47619	PNMA1	1
UPF3B	26	B3GALNT1	1	RAB9B	0.333333	B3GALNT1	1
PRPF31	26	PNMAL1	1	NFU1	0.333333	PNMAL1	1
PRPF19	26	RFXANK	1	LGALS1	0.285714	RFXANK	1
RPS12	26	PDZRN3	1	NOTCH2	0.285714	PDZRN3	1
MRPS5	25	EHD1	1	HSPA8	0.186288	EHD1	1
CXCL1	24	ABCB6	1	ACACA	0.163552	ABCB6	1
MRPL11	24	H2AFY	1	GNG2	0.136162	H2AFY	1
MRPL16	24	CMBL	1	PIK3CD	0.101836	CMBL	1
PSMD4	23	TMEM5	1	PPP2R1A	0.093895	TMEM5	1
PSMD14	23	NAB2	1	ACTB	0.079388	NAB2	1
CXCL5	23	SLC25A28	1	HSPA4	0.075822	SLC25A28	1
MRPS12	23	EHD4	1	CCT6A	0.056667	EHD4	1
MRPS11	23	C1GALT1C1	1	NHP2L1	0.055815	C1GALT1C1	1
MRPL12	23	ABCB10	1	BCL2L1	0.049086	ABCB10	1
HEATR1	22	BRPF3	1	HIST2H2AA3	0.046396	BRPF3	1
WDR36	22	ADAMTSL4	1	RPS20	0.046002	ADAMTSL4	1
UBE2D4	22	PIEZO1	1	OAS1	0.043341	PIEZO1	1
MRPL1	22	A4GALT	1	PSMD4	0.042385	A4GALT	1
PSMC3	21	SCN3A	1	PRDM10	0.039216	SCN3A	1
PUF60	21	CH25H	1	POLR1E	0.037317	CH25H	1
POLR1E	21	NUP210	1	PPP2R2A	0.03345013	NUP210	1
SNRPD2	21	GTDC2	1	PPIE	0.03324088	GTDC2	1
HEBP1	21	GALNTL4	1	RPS11	0.03309413	GALNTL4	1
ITGA7	21	PIGN	2	FGF2	0.0317498	PIGN	1.333333
BOP1	21	PIGK	2	AR	0.03166632	PIGK	1.333333
PQBP1	20	SNTB1	2	HSPA9	0.03080485	SNTB1	1.333333
DHX16	20	ACSS2	2	HIF1A	0.03058983	ACSS2	1.5
HERC5	20	LPCAT3	1	RAF1	0.02964581	LPCAT3	1.5
MRPL19	20	PAFAH2	1	CDC16	0.02890363	PAFAH2	1.5
MRPL40	20	ATP6V1C1	1	ATG5	0.02844196	ATP6V1C1	1.5
ACTB	19	EXT1	1	MRPS9	0.02803152	EXT1	1.5
PSMD8	19	XYLT2	1	PSMC3	0.02785462	XYLT2	1.5
SF3A3	19	ATPIF1	1	PC	0.02675148	ATPIF1	1.5
SRSF3	19	LEPREL1	1	AP1G1	0.02505643	LEPREL1	1.666667
SNRPA1	19	CRTAP	1	RAD23A	0.02427339	CRTAP	1.666667

Top 35 were screened.

APL, average path length; BC, betweenness centrality; CNC, closeness centrality; CCR7, C-C motif chemokine receptor 7; CXCL1, C-X-C motif chemokine ligand 1; DC, degree centrality.

The 521 DEGs in the PPI network were uploaded into the DAVID database to predict their functions. As a result, 34 GO terms (14, BP; 17, CC; and 3, MF) were enriched, such as GO:0044265—cellular macromolecule catabolic process (CDC16), GO:0006511—ubiquitin-dependent protein catabolic process (CDC16), GO:0008610—lipid biosynthetic process (ACSS2), and GO:0009898—internal side of plasma membrane (GNG2). In addition, five KEGG pathways were also enriched, including hsa03040:Spliceosome, hsa03050:Proteasome, hsa04120:Ubiquitin-mediated proteolysis (CDC16), hsa00240:Pyrimidine metabolism, and hsa04144:Endocytosis (Table 3 and Fig. 4).

FIG. 4.

Function enrichment analysis for the genes in the protein and protein interaction network. BP, biological process; CC, cellular component; GO, Gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function. Color images are available online.

Table 3.

Function Enrichment for the Genes in the Protein-Protein Interaction Network

Category	Term	Count	p
BP	GO:0044265—cellular macromolecule catabolic process	49	8.32E-07
	GO:0009057—macromolecule catabolic process	52	5.48E-07
	GO:0006412—translation	29	2.05E-06
	GO:0006511—ubiquitin-dependent protein catabolic process	23	8.30E-06
	GO:0044257—cellular protein catabolic process	41	6.96E-06
	GO:0051603—proteolysis involved in cellular protein catabolic process	41	6.14E-06
	GO:0043632—modification-dependent macromolecule catabolic process	39	1.23E-05
	GO:0019941—modification-dependent protein catabolic process	39	1.23E-05
	GO:0030163—protein catabolic process	41	1.42E-05
	GO:0031400—negative regulation of protein modification process	15	2.21E-05
	GO:0051248—negative regulation of protein metabolic process	18	8.78E-05
	GO:0032269—negative regulation of cellular protein metabolic process	17	1.83E-04
	GO:0051247—positive regulation of protein metabolic process	20	2.53E-04
	GO:0008610—lipid biosynthetic process	24	2.36E-04
CC	GO:0030529—ribonucleoprotein complex	49	6.69E-12
	GO:0005739—mitochondrion	73	3.82E-10
	GO:0044429—mitochondrial part	48	3.00E-09
	GO:0005840—ribosome	26	1.25E-08
	GO:0070013—intracellular organelle lumen	96	3.54E-08
	GO:0043233—organelle lumen	97	5.33E-08
	GO:0005829—cytosol	76	1.70E-07
	GO:0031090—organelle membrane	62	4.52E-06
	GO:0031980—mitochondrial lumen	22	7.82E-06
	GO:0005759—mitochondrial matrix	22	7.82E-06
	GO:0031975—envelope	38	1.08E-04
	GO:0031967—organelle envelope	38	9.95E-05
	GO:0031966—mitochondrial membrane	27	2.42E-04
	GO:0005740—mitochondrial envelope	27	6.20E-04
	GO:0009898—internal side of plasma membrane	22	8.58E-04
	GO:0031981—nuclear lumen	67	8.13E-04
	GO:0044432—endoplasmic reticulum part	22	0.002696
MF	GO:0003735—structural constituent of ribosome	17	5.06E-05
	GO:0003723—RNA binding	43	3.12E-05
	GO:0000166—nucleotide binding	97	1.72E-04
KEGG	hsa03040:Spliceosome	15	2.09E-03
	hsa03050:Proteasome	7	2.12E-02
	hsa04120:Ubiquitin-mediated proteolysis	13	2.59E-02
	hsa00240:Pyrimidine metabolism	10	3.27E-02
	hsa04144:Endocytosis	15	4.84E-02

BP, biological process; CC, cellular component; GO, gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function.

Furthermore, four significant function-related modules were extracted from the PPI network (Fig. 5 and Table 4). The hub gene CDC16 was enriched in module 1, which was demonstrated to be associated with hsa04120:Ubiquitin-mediated proteolysis and hsa04110:Cell cycle; hub gene GNG2 with interactive genes (CXCL1 and CCR7) was enriched in module 4, and they were involved in hsa04062:Chemokine signaling pathway (Table 5).

FIG. 5.

Significant modules extracted from the protein and protein interaction network. M1, module 1; M2, module 2; M3, module 3; M4, module 4. Red, upregulated; green, downregulated. The larger size of node (protein) indicated the higher degree (interaction relationships) of it. Color images are available online.

Table 4.

Module Analysis

Cluster	Score (density × No. of nodes)	Node	Edges	Node IDs
1	9	19	19	RNF126, LRRC41, SIAH1, TRIM32, FBXL12, UBE2G2, ITCH, UBE2D4, RNF220, KEAP1, RNF115, UBE2M, FBXW5, RNF114, CDC16, HERC5, FBXW4, ANAPC5, CDC23
2	8.919	37	330	PPIE, SNRPD2, UPF3B, PUF60, GPKOW, RBM17, PRPF19, CD2BP2, SF3A3, SNRPA1, HSPA8, DHX16, DNAJC, PRPF31, PQBP1, NHP2L1, DDX46, MRPS5, MRPL53, MRPL16, MRPL11, MRPL41, MRPL37, MRPL39, MRPS11, MRPS9, MRPS12, MRPS19, MRPL14, MRPL38, MRPL46, MRPL40, MRPL1, MRPL12, MRPS26, MRPL21
3	3.5	8	28	ARPC1A, GAPVD1, OCRL, PACSIN3, SYNJ1, PACSIN1, EPS15, SH3GL1
4	4.238	21	89	HEBP1, HEATR1, GRWD1, DDX10, WDR36, ADCY9, DDX50, CCR7, CXCL5, GNG2, UTP11L, RPP40, POLR1E, GNL2, SSTR2, CXCL1, PWP1, RSL1D1, GNG12, BOP1, GNAI3

Table 5.

Function Enrichment for the Significant Modules

Module	Term	p	Genes
1	hsa04120:Ubiquitin-mediated proteolysis	3.79E-16	UBE2D4, ANAPC5, UBE2G2, TRIM32, UBE2M, CDC23, SIAH1, KEAP1, CDC16, ITCH
	hsa04914:Progesterone-mediated oocyte maturation	5.37E-03	ANAPC5, CDC23, CDC16
	hsa04114:Oocyte meiosis	8.63E-03	ANAPC5, CDC23, CDC16
	hsa04110:Cell cycle	1.07E-02	ANAPC5, CDC23, CDC16
	hsa05166:HTLV-I infection	4.12E-02	ANAPC5, CDC23, CDC16
2	hsa03040:Spliceosome	5.48E-14	PRPF19, SNRPA1, PPIE, PRPF31, DDX46, PQBP1, SNRPD2, DHX16, PUF60, HSPA8, SF3A3, RBM17
2	hsa03010:Ribosome	1.42E-10	MRPL11, MRPL1, MRPL21, MRPL12, MRPS9, MRPL14, MRPS12, MRPL16, MRPS11, MRPS5
3	hsa04144:Endocytosis	7.00E-03	EPS15, ARPC1A, SH3GL1
3	hsa00562:Inositol phosphate metabolism	4.07E-02	SYNJ1, OCRL
4	hsa04062:Chemokine signaling pathway	2.91E-07	CXCL1, CCR7, GNAI3, CXCL5, ADCY9, GNG2, GNG12
	hsa04727:GABAergic synapse	3.70E-04	GNAI3, ADCY9, GNG2, GNG12
	hsa03008:Ribosome biogenesis in eukaryotes	3.96E-04	WDR36, HEATR1, GNL2, RPP40
	hsa05032:Morphine addiction	4.52E-04	GNAI3, ADCY9, GNG2, GNG12
	hsa04713:Circadian entrainment	5.13E-04	GNAI3, ADCY9, GNG2, GNG12
	hsa04723:Retrograde endocannabinoid signaling	6.14E-04	GNAI3, ADCY9, GNG2, GNG12
	hsa04725:Cholinergic synapse	8.09E-04	GNAI3, ADCY9, GNG2, GNG12
	hsa04724:Glutamatergic synapse	8.75E-04	GNAI3, ADCY9, GNG2, GNG12
	hsa04971:Gastric acid secretion	6.84E-03	SSTR2, GNAI3, ADCY9
	hsa04726:Serotonergic synapse	1.53E-02	GNAI3, GNG2, GNG12
	hsa04728:Dopaminergic synapse	2.01E-02	GNAI3, GNG2, GNG12
	hsa05200:Pathways in cancer	2.77E-02	GNAI3, ADCY9, GNG2, GNG12
	hsa05034:Alcoholism	3.67E-02	GNAI3, GNG2, GNG12
	hsa04024:cAMP signaling pathway	4.50E-02	SSTR2, GNAI3, ADCY9

Screening of crucial DEGs by miRNA-mRNA regulatory network analysis

A total of 786 DEGs (12 upregulated and 774 downregulated) were predicted to be negatively regulated by 20 DEMs (17 upregulated and 3 downregulated), which constituted 1338 interaction relationships to construct the miRNA-mRNA regulatory network (Fig. 6). In this network, hub gene CDC16 could be regulated by hsa-miR-126; hub gene ACSS2 could be regulated by hsa-miR-15b; and hub gene GNG2 was regulated by hsa-miR-30a. All of these miRNAs belonged to the top 10 upregulated or downregulated DEMs in GSE62179 dataset, further demonstrating that these miRNAs and DEGs regulated by them may be especially crucial for the development of AAA.

FIG. 6.

miRNA-mRNA interaction network. Square nodes represented miRNAs; circular nodes represented mRNAs. Red, upregulated; green, downregulated. mRNA, messenger RNA. Color images are available online.

Validation of crucial DEGs and DEMs in other cancers

GNG2 was also found to be significantly upregulated in kidney renal clear cell carcinoma (KRIC), acute myeloid leukemia (LAML), pancreatic adenocarcinoma (PAAD), pheochromocytoma and paraganglioma (PCPG), and skin cutaneous melanoma (SKCM). ACSS2 was shown to be downregulated in most cancer types, especially significant in breast invasive carcinoma, cholangiocarcinoma (CHOL), LAML, lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), PCPG, SKCM, and testicular germ cell tumors (TGCT). CDC16 was only observed to be downregulated in kidney chromophobe (KICH) (Fig. 7). In line with the expressions of their target genes, hsa-miR-30a was found to be relatively lower expressed, while hsa-miR-15b was highly expressed in several cancer types, including the same with their target genes (Fig. 8). Unexpectedly, hsa-miR-126 was also lower expressed in KICH and thus, hsa-miR-126 and CDC16 may not negatively interact with each other.

FIG. 7.

Validation of mRNA expression in GEPIA. *p < 0.05. BRCA, breast invasive carcinoma; CHOL, cholangiocarcinoma; KRIC, kidney renal clear cell carcinoma; LAML, acute myeloid leukemia; LUAD, lung adenocarcinoma; LUSC, lung squamous cell carcinoma; N, control; PAAD, pancreatic adenocarcinoma; PCPG, pheochromocytoma and paraganglioma; SKCM, skin cutaneous melanoma; T, tumor; TGCT, testicular germ cell tumors. Color images are available online.

FIG. 8.

Validation of miRNA expression in tumor-miRNA-pathway. ACC, adrenocortical carcinoma; BLCA, bladder urothelial carcinoma; CESC, cervical squamous cell carcinoma and endocervical adenocarcinoma; COAD, colon adenocarcinoma; DLBC, lymphoid neoplasm diffuse large B cell lymphoma; ESCA, esophageal carcinoma; GBM, glioblastoma multiforme; HNSC, head and neck squamous cell carcinoma; KICH, kidney chromophobe; KIRP, kidney renal papillary cell carcinoma; LGG, brain lower grade glioma; LIHC, liver hepatocellular carcinoma; MESO, mesothelioma; OV, ovarian serous cystadenocarcinoma; PRAD, prostate adenocarcinoma; READ, rectum adenocarcinoma; SARC, sarcoma; STAD, stomach adenocarcinoma; THCA, thyroid carcinoma; THYM, thymoma; UCEC, uterine corpus endometrial carcinoma; UCS, uterine carcinosarcoma; UVM, uveal melanoma. Color images are available online.

Discussion

In this study, we identified two potentially crucial miRNA-mRNA interaction pairs (hsa-miR-30a-GNG2 and hsa-miR-15b-ACSS2) to explain the pathogenesis of AAA. GNG2 may be involved in the inflammatory pathway by interacting with chemokine CXCL1 and CCR7; ACSS2 participated in lipid biosynthetic process.

GNG2 encodes a gamma subunit of the heterotrimeric G-protein complex GαGβγ. Previous studies demonstrated that inhibition of GαGβγ signaling retarded cancer cell growth, induced cell death (Bookout et al., 2003), alleviated the invasive phenotype (Overstreet, 2012), and enhanced the chemosensitivity (Paudyal et al., 2017). Furthermore, using the zebrafish embryos as a model, Leung et al. (2006) also found that loss of GNG2 inhibited angiogenesis by attenuating vascular endothelial growth factor (VEGF)-induced phosphorylation of phospholipase C-gamma1 and serine/threonine kinase signaling (Leung et al., 2006), while blocking of VEGF-related angiogenesis has been extensively demonstrated as a promising approach for the treatment of various cancers (Li et al., 2018; Mawalla et al., 2018). These findings indicate that GNG2 may be a proto-oncogene. This hypothesis could be indirectly confirmed in a recent study that showed GNG2 was upregulated in saliva exosomes of pancreatic cancer-bearing mice compared with the control mice (Lau et al., 2013). In line these descriptions, our results also showed that GNG2 was upregulated in AAA tissues as well as in KRIC, LAML, PAAD, PCPG, and SKCM. However, the roles and mechanisms of GNG2 in cancerization remain rarely reported and even controversial (Naito, 2012; Yajima et al., 2012, 2014). In our study, we predicted that GNG2 may be involved in the development of AAA by regulating the inflammatory pathway through interacting with a series of chemokines. This prediction result was also seen in the study of colorectal cancer (Liang et al., 2016). Accumulating evidence had suggested that highly expressed proinflammatory cytokine CXCL1, which was mediated by nuclear factor-κB (NF-κB), promoted tumor proliferation, invasion, and metastasis, leading to reduced disease-specific survival (Bachmeier et al., 2008; Cheng et al., 2011; Miyake et al., 2013). The mechanism of study revealed that CXCL1 was associated with the advanced tumor stage and poorer prognosis by activating the VEGF signaling pathway (Wei et al., 2015). Similarly, a meta-analysis showed CCR7 was upregulated in gastric cancer tissues and significantly associated with deeper tumor invasion, advanced stage, vascular invasion, lymph node metastasis, lymphatic invasion, and worse 5-year overall survival rate (Du et al., 2017). CCR7 also mediated migration and invasion by regulating the VEGF pathway in LUAD (Jie et al., 2017). In line with other cancers, macrophage infiltration as well as the expression of chemokine CXCL1 and VEGF were also found to be increased in angiotensin II-induced AAA apoE(−/−) mice model (Escudero et al., 2015; Martorell et al., 2016), while immunohistochemical staining analysis showed CCR7 was obviously upregulated in AAA tissue of patients compared with the control (Wan et al., 2018). Treatment with calcitriol or bexarotene plus rosuvastatin reduced aneurysm formation, inflammation, and neovascularization by diminishing production of the above chemokines and VEGF through inactivation of extracellular signal-regulated kinases 1/2, p38 mitogen-activated protein kinase, and NF-κB (Escudero et al., 2015; Martorell et al., 2016). Furthermore, Gβ1γ2 was also implicated to promote NF-κB activation through Akt signaling (Ming et al., 2009). Accordingly, we believe GNG2 may be involved in AAA by activation of NF-κB-CCR7/CXCL1-VEGF pathway.

ACSS2 encodes one member of the acetyl-CoA synthetase family that is responsible for the synthesis of acetyl-CoA through the ligation of acetate and CoA. Acetyl-CoA is a substrate for tricarboxylic acid cycle and a key precursor of lipid synthesis. Thus, the high expression of ACSS2 may be conducive to energy production and then supports excessive growth and proliferation of cells, leading to carcinogenesis. This hypothesis had been demonstrated for renal cell carcinoma (Yao et al., 2018; Zhang et al., 2018a), glioblastoma (Tomoyuki et al., 2014), and prostate and breast cancer (Schug et al., 2015). However, recent evidence put forward opposite conclusions. For example, Hur et al. (2015) and Bae et al. (2016) identified loss of ACSS2 expression as one of the independent prognostic factors for predicting worse survival of gastric cancer and colorectal carcinoma, respectively. Sun et al. (2017) illustrated that the knockdown of ACSS2 increased the invasion and migration ability of hepatocellular carcinoma cells by promoting the epithelial-mesenchymal transition (EMT). In line with the studies of Hur et al. (2015) and Bae et al. (2016), we also found ACSS2 was downregulated in AAA samples as well as CHOL, LAML, LUAD, LUSC, PCPG, SKCM, and TGCT. These findings indicate that ACSS2 may be a tumor suppressor gene. Although the mechanism of tumor suppression remains unclear, it is speculated to be attributed to the acetylation effects on hypoxia-inducible factor (HIF)-2α. ACSS2 knockdown downregulated the acetylation levels of HIF-2α and led to enhanced HIF-2α activity, further inducing the EMT and metastasis (Sun et al., 2017). It was also reported that HIF family member, HIF-1α, was significantly increased in both human and experimental aneurysm tissues. Treatment with HIF-1α inhibitor prevented enlargement of experimental aneurysms and attenuated VSMC depletion and the accumulation of macrophages, T cells, and B cells (Tsai et al., 2016; Yang et al., 2016; Wang et al., 2018). Thus, ACSS2 mediated upregulation of HIF family genes and then proinflammation may be the potential mechanisms of ACSS2 for AAA.

In addition to the downstream targets, the changes in the upstream regulators may also be one possible reason for explaining the high expression of GNG2 and lower expression of ACSS2 in some cancers, including AAA. Increasing evidence has proved that miRNA is an important class of negative regulators for the expression of its downstream target genes. Thus, the miRNAs that could regulate the expressions of GNG2 and ACSS2 were also investigated in our study. The results showed that hsa-miR-30a (downregulated) and hsa-miR-15b (upregulated) could, respectively, regulate GNG2 and ACSS2 and the expression between miRNAs and their target genes was opposite. Our findings in AAA seemed to be in accordance with previous studies on these two miRNAs. For example, Wei et al. found miR-30a-5p was significantly downregulated in human colorectal cancer tissue specimens and cell lines compared with noncancerous tissues and cells. The overexpression of miR-30a-5p inhibited the migratory and invasive abilities of colorectal cancer cells by specifically targeting the integrin β3 gene and then suppressing the EMT (Wei et al., 2016). Using the luciferase reporter assay and transfection, Ruan et al. (2018) demonstrated that miR-30a-5p directly targeted fibroblast activation protein α gene to suppress cell propagation, migration, and invasion of oral cavity cancer cells. The study of Tang et al. (2015a) supported that miR-30a expression was significantly downregulated in non-small lung cancer tissues compared to their non-tumor lung tissues. Level of miR-30a was negatively associated with tumor size, lymphatic metastasis, clinical TNM stage, pathological grading, histological classification, and survival time (Tang et al., 2015a). There was also one study to use the polymerase chain reaction (PCR) to confirm the downregulation of miR-30a-5p in whole AAA samples and in M1 macrophages (Spear et al., 2015). Downregulation of miR-30a was also reported to exacerbate IL-17-mediated inflammation and lead to the development of autoimmune diseases (Wan et al., 2015), while upregulation of miR-30a reversed the occurrence of diseases by suppressing Th17 differentiation (Zhao et al., 2016). Moreover, knockdown of miR-30 was also proved to contribute to death of aorta VSMCs (Chen et al., 2014). Accordingly, it is theoretically confirmable that downregulated miR-30a may mediate the upregulation of GNG2 and its related downstream inflammation pathways, ultimately promoting the development and progression of AAA. Similarly, Zhang et al. (2015) elucidated that the overexpression of miR-15b promoted EMT and metastasis of pancreatic cancer, the mechanism of which was related with degrading SMAD-specific E3 ubiquitin protein ligase 2. Inhibition of miR-15b activity by adenovirus carrying antimiR-15b sequence was reported to significantly decrease the colony formation ability, invasion, and migration of HCT116 cells in vitro and liver metastasis of HCT116 tumors in vivo (Li et al., 2016). Increased expression of miR-15b was also revealed to be strongly correlated with the degree of differentiation, clinical stage, tumor diameter, lymph node metastases, and worse 5-year overall cumulative survival rates of patents with cervical carcinoma (Wen et al., 2017). In addition, overexpression of miR-15b was also demonstrated to enhance the production of proinflammatory cytokines, whereas inhibition of miR-15b decreased the inflammatory response and attenuated cell apoptosis (An et al., 2012; Zhu et al., 2015). Therefore, it is theoretically believable that upregulated miR-15b may mediate the downregulation of ACSS2 and then activate the downstream inflammation pathways, consequentially promoting the apoptosis of VSMCs and development of AAA.

However, there were some limitations in this study. First, the samples of the used datasets were not large in size and collected from different platforms with different design. This may lead to lower statistical power to obtain the results and the result difference in different datasets [GSE62179 and GSE63541; such as the expression of miR-29a was opposite in these two datasets and could not be discussed as the important targets for AAA like previous studies (Boon et al., 2011; Maegdefessel et al., 2012)]. Second, we only preliminarily screened the crucial miRNA-mRNA interaction pairs for explaining the pathogenesis of AAA. The expression of miRNAs and mRNAs, the interactions between them, and which immune cells influenced [macrophages or others (Spinosa et al., 2018)] need further experiments (luciferase reporter assay, knockout or overexpression in vitro or in vivo, and PCR) to confirm.

Conclusion

This study suggests that hsa-miR-30a-GNG2 and hsa-miR-15b-ACSS2 interaction pairs may represent novel mechanisms for explaining the pathogenesis of AAA. Both of them may be consequentially involved in AAA by influencing the inflammatory pathways and apoptosis of VSMCs. Targeted regulation of them may be potential strategies for treatment of AAA.

Availability of Data and Materials

The origin data used were all downloaded from GEO database under accession number GSE57691, GSE62179, and GSE63541.

Footnotes

Disclosure Statement

The authors declared no potential conflicts of interest.

Funding Information

The authors received no financial support for the research, authorship, and/or publication of this article.

References

, Gong

, Wang

, Yu

, Zhao

, Lin

, et al. (2012). miR-15b and miR-16 regulate TNF mediated hepatocyte apoptosis via BCL2 in acute liver failure. Apoptosis, 17, 702–716.

Bachmeier

B.E.

, Mohrenz

I.V.

, Mirisola

, Schleicher

, Romeo

, Höhneke

, et al. (2008). Curcumin downregulates the inflammatory cytokines CXCL1 and -2 in breast cancer cells via NFκB. Carcinogenesis, 29, 779–789.

Bader

G.D.

, and Hogue

C.W.

(2003). An automated method for finding molecular complexes in large protein interaction networks. BMC Bioinformatics, 4, 2.

Bae

J.M.

, Kim

J.H.

, Oh

H.J.

, Park

H.E.

, Lee

T.H.

, Cho

N.Y.

, et al. (2016). Downregulation of acetyl-CoA synthetase 2 is a metabolic hallmark of tumor progression and aggressiveness in colorectal carcinoma. Mod Pathol, 30, 267–277.

Bookout

A.L.

, Finney

A.R.

, Peppel

, Koch

W.J.

, and Daaka

(2003). Targeting Gbetagamma signaling to inhibit prostate tumor formation and growth. J Biol Chem, 278, 37569–37573.

Boon

R.A.

, Seeger

, Heydt

, Fischer

, Hergenreider

, Horrevoets

A.J.

, et al. (2011). MicroRNA-29 in aortic dilation: implications for aneurysm formation. Circ Res, 109, 1115–1119.

Cao

, Cai

, Liu

, Zhao

, Wang

, Li

, et al. (2017). miRNA-504 inhibits p53-dependent vascular smooth muscle cell apoptosis and may prevent aneurysm formation. Mol Med Rep, 16, 2570–2578.

Chen

, Ma

, Yue

, Wei

, Li

, Tong

, et al. (2014). Downregulation of the miR-30 family microRNAs contributes to endoplasmic reticulum stress in cardiac muscle and vascular smooth muscle cells. Int J Cardiol, 173, 65–73.

Cheng

W.L.

, Wang

C.S.

, Huang

Y.H.

, Tsai

M.M.

, Liang

, and Lin

K.H.

(2011). Overexpression of CXCL1 and its receptor CXCR2 promote tumor invasion in gastric cancer. Ann Oncol, 22, 2267–2276.

10.

Cheuk

B.L.

Y, and Cheng

S.W.K.

(2014). Identification and characterization of microRNAs in vascular smooth muscle cells from patients with abdominal aortic aneurysms. J Vasc Surg, 59, 202–209.

11.

Choke

, Thompson

M.M.

, Jones

, Torsney

, Dawson

, Laing

, et al. (2010). Gene expression profile of abdominal aortic aneurysm rupture. Ann N Y Acad Sci, 1085, 311–314.

12.

Dereziński

T.L.

, Fórmankiewicz

, Migdalski

, Brazis

, Jakubowski

, Woda, Ł., et al. (2017). The prevalence of abdominal aortic aneurysms in the rural/urban population in central Poland—Gniewkowo Aortic Study. Kardiol Pol, 75, 705–710.

13.

, Liu

, Ren

, Zhao

, Zhang

, Patel

, et al. (2017). Expression of chemokine receptor CCR7 is a negative prognostic factor for patients with gastric cancer: a meta-analysis. Gastric Cancer, 20, 235–245.

14.

Dweep

, and Gretz

(2015). miRWalk2.0: a comprehensive atlas of microRNA-target interactions. Nat Methods 12, 697.

15.

Erik

, Moran

C.S.

, Rush

C.M.

, Gabor

G.B.

, Charlotte

, Lindeman

J.H.N.

, et al. (2015). Differential gene expression in the proximal neck of human abdominal aortic aneurysm. Oncotarget, 6, 12984–12996.

16.

Escudero

, Navarro

, Ferrando

, Furio

, Gonzalez-Navarro

, Juez

, et al. (2015). Combined treatment with bexarotene and rosuvastatin reduces angiotensin-II-induced abdominal aortic aneurysm in apoE(−/−) mice and angiogenesis. Br J Pharmacol, 172, 2946–2960.

17.

Henderson

E.L.

, Geng

Y.J.

, Sukhova

G.K.

, Whittemore

A.D.

, Knox

, and Libby

(1999). Death of smooth muscle cells and expression of mediators of apoptosis by T lymphocytes in human abdominal aortic aneurysms. Circulation, 99, 96–104.

18.

Huang

D.W.

, Sherman

B.T.

, and Lempicki

R.A.

(2009). Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc, 4, 44–57.

19.

Hur

, Kim

Y.B.

, Ham

I.H.

, and Lee

(2015). Loss of ACSS2 expression predicts poor prognosis in patients with gastric cancer. J Surg Oncol, 112, 585–591.

20.

Jie

, Tao

, Hu

, Wang

, Fang

, Yi

, et al. (2017). CCR7 promote lymph node metastasis via regulating VEGF-C/D-R3 pathway in lung adenocarcinoma. J Cancer, 8, 2060–2068.

21.

Johansson

, and Swedenborg

(2010). Ruptured abdominal aortic aneurysms: a study of incidence and mortality. Br J Surg, 73, 101–103.

22.

Kent

K.C.

(2014). Clinical practice. Abdominal aortic aneurysms. N Engl J Med, 371, 2101–2108.

23.

Kohl

, Wiese

, and Warscheid

(2011). Cytoscape: software for visualization and analysis of biological networks. Methods Mol Biol, 696, 291–303.

24.

Lau

, Kim

, Chia

, Spielmann

, Eibl

, Elashoff

, et al. (2013). Role of pancreatic cancer-derived exosomes in salivary biomarker development. J Biol Chem, 288, 26888–26897.

25.

Leung

, Chen

, Stauffer

A.M.

, Giger

K.E.

, Sinha

, Horstick

E.J.

, et al. (2006). Zebrafish G protein gamma2 is required for VEGF signaling during angiogenesis. Blood, 108, 160.

26.

, Wang

, Shen

, Wei

, and Li

(2018). Oridonin inhibits VEGF-A-associated angiogenesis and epithelial-mesenchymal transition of breast cancer in vitro and in vivo. Oncol Lett, 16, 2289–2298.

27.

, Chen

, Guo

, Zhou

, Jia

, Tang

, et al. (2016). Inhibition of miR-15b decreases cell migration and metastasis in colorectal cancer. Tumour Biol, 37, 8765–8773.

28.

Liang

, Che

, Zhao

, Zhang

, and Shi

(2017). MiR-195 promotes abdominal aortic aneurysm media remodeling by targeting Smad3. Cardiovasc Ther 35, e12286.

29.

Liang

, Li

, and Zhao

(2016). Identification of key pathways and genes in colorectal cancer using bioinformatics analysis. Med Oncol 33, 111.

30.

Lieberg

, Pruks

L.L.

, Kals

, Paapstel

, Aavik

, and Kals

(2017). Mortality after elective and ruptured abdominal aortic aneurysm surgical repair: 12-Year single-center experience of Estonia. Scand J Surg, 107, 152–157.

31.

, Liu

, Huang

, Liu

, Zhang

H.M.

, Li

, et al. (2016). MicroRNA regulatory pathway analysis identifies miR-142-5p as a negative regulator of TGF-β pathway via targeting SMAD3. Oncotarget, 7, 71504–71513.

32.

Maegdefessel

, Azuma

, Toh

, Merk

D.R.

, Deng

, Chin

J.T.

, et al. (2012). Inhibition of microRNA-29b reduces murine abdominal aortic aneurysm development. J Clin Invest, 122, 497–506.

33.

Martorell

, Hueso

, Gonzalez-Navarro

, Collado

, Sanz

M.J.

, and Piqueras

(2016). Vitamin D receptor activation reduces angiotensin-II-induced dissecting abdominal aortic aneurysm in apolipoprotein E-knockout mice. Arterioscler Thromb Vasc Biol, 36, 1587–1597.

34.

Mawalla

, Yuan

, Luo

, and Chalya

P.L.

(2018). Treatment outcome of anti-angiogenesis through VEGF-pathway in the management of gastric cancer: a systematic review of phase II and III clinical trials. BMC Res Notes 11, 21.

35.

Ming

, He

R.L.

, Benovic

J.L.

, and Ye

R.D.

(2009). β-Arrestin1 interacts with the G-protein subunits β1γ2 and promotes β1γ2-dependent Akt signalling for NF-κB activation. Biochem J, 417, 287–296.

36.

Miyake

, Lawton

, Goodison

, Urquidi

, Gomes-Giacoia

, Zhang

, et al. (2013). Chemokine (C-X-C) ligand 1 (CXCL1) protein expression is increased in aggressive bladder cancers. BMC Cancer, 13, 322.

37.

Naito

(2012). Reduced GNG2 expression levels in mouse malignant melanomas and human melanoma cell lines. Am J Cancer Res, 2, 322–329.

38.

Nakao

, Horie

, Baba

, Nishiga

, Nishino

, Izuhara

, et al. (2017). Genetic ablation of microRNA-33 attenuates inflammation and abdominal aortic aneurysm formation via several anti-inflammatory pathways. Arterioscler Thromb Vasc Biol, 37, 2161–2170.

39.

Overstreet

R.M.

(2012). Identification of a GαGβγ, AKT and PKCα signalome associated with invasive growth in two genetic models of human breast cancer cell epithelial-to-mesenchymal transition. Int J Oncol, 41, 189–200.

40.

Pahl

M.C.

, Derr

, Gäbel

, Hinterseher

, Elmore

J.R.

, Schworer

C.M.

, et al. (2012). MicroRNA expression signature in human abdominal aortic aneurysms. BMC Med Genomics, 5, 25.

41.

Paudyal

, Xie

, Vaddi

P.K.

, Henry

M.D.

, and Chen

(2017). Inhibiting G protein βγ signaling blocks prostate cancer progression and enhances the efficacy of paclitaxel. Oncotarget, 8, 36067–36081.

42.

Ritchie

M.E.

, Phipson

, Wu

, Hu

, Law

C.W.

, Shi

, et al. (2015). limma Powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res, 43, e47.

43.

Ruan

, Tao

, and Tan

(2018). Low expression of miR-30a-5p induced the proliferation and invasion of oral cancer via promoting the expression of FAP. Biosci Rep, 38, pii:BSR20171027.

44.

Schug

, Peck

, Jones

, Zhang

, Alam

, Witney

, et al. (2015). Acetyl-coA synthetase 2 promotes acetate utilization and maintains cell growth under metabolic stress. Cancer Cell, 27, 57–71.

45.

Spear

, Boytard

, Blervaque

, Chwastyniak

, Hot

, Vanhoutte

, et al. (2015) Adventitial tertiary lymphoid organs as potential source of microRNA biomarkers for abdominal aortic aneurysm. Int J Mol Sci, 16, 11276–11293.

46.

Spinosa

, Lu

, Su

, Bontha

S.V.

, Gehrau

, Salmon

M.D.

, et al. (2018). Human mesenchymal stromal cell-derived extracellular vesicles attenuate aortic aneurysm formation and macrophage activation via microRNA-147. FASEB J 29, fj201701138RR.

47.

Sun

, Kong

, Cao

, Zhou

, Li

, Cui

, et al. (2017). Decreased expression of acetyl-CoA synthase 2 promotes metastasis and predicts poor prognosis in hepatocellular carcinoma. Cancer Sci, 108, 1338–1346.

48.

Szklarczyk

, Franceschini

, Wyder

, Forslund

, Heller

, Huerta-Cepas

, et al. (2015). STRING v10: protein-protein interaction networks, integrated over the tree of life. Nucleic Acids Res, 45243, D447.

49.

Tang

, Liang

, Luo

, Feng

, Huang

, He

, et al. (2015a) Downregulation of MiR-30a is associated with poor prognosis in lung cancer. Med Sci Monit, 21, 2514–2520.

50.

Tang

, Li

, Wang

, Pan

, and Wu

F.X.

(2015b) CytoNCA: a cytoscape plugin for centrality analysis and evaluation of protein interaction networks. Biosystems, 127, 67–72.

51.

Tang

, Li

, Kang

, Gao

, Li

, and Zhang

(2017). GEPIA: a web server for cancer and normal gene expression profiling and interactive analyses. Nucleic Acids Res, 45, W98–W102.

52.

Tian

, Hu

, He

, Wu

, Li

, and Zhang

(2018). Construction of lncRNA-miRNA-mRNA networks reveals functional lncRNAs in abdominal aortic aneurysm. Exp Ther Med, 16, 3978–3986.

53.

Tomoyuki

, Kumar

, Vamsidhara

, Hatanpaa

K.J.

, Dinesh Kumar

, Shyam

, et al. (2014). Acetate is a bioenergetic substrate for human glioblastoma and brain metastases. Cell, 159, 1603–1614.

54.

Tsai

S.H.

, Huang

P.H.

, Hsu

Y.J.

, Peng

Y.J.

, Lee

C.H.

, Wang

J.C.

, et al. (2016). Inhibition of hypoxia inducible factor-1α attenuates abdominal aortic aneurysm progression through the down-regulation of matrix metalloproteinases. Sci Rep 6, 28612.

55.

Wan

, Huang

, Ni

, and Yu

(2018). Screening key genes for abdominal aortic aneurysm based on Gene Expression Omnibus dataset. BMC Cardiovasc Disord, 18, 34.

56.

Wan

, Zhou

, Ding

, and He

(2015). The miR-30a negatively regulates IL-17-mediated signal transduction by targeting Traf3ip2. J Interferon Cytokine Res, 35, 917–923.

57.

Wang

, Bi

, Wang

, Huang

, Lu

, and Zhu

(2017). Microarray analysis to identify the similarities and differences of pathogenesis between aortic occlusive disease and abdominal aortic aneurysm. Vascular, 26, 301–314.

58.

Wang

, Xu

, Xuan

, Ge

, Wang

, et al. (2018). Hypoxia-inducible factor 1 in clinical and experimental aortic aneurysm disease. J Vasc Surg, 68, 1538.e2–1550.e2.

59.

Wei

, Yang

, Cai

, Cui

, Li

R.X.

, Wang

, et al. (2016). MiR-30a-5p suppresses tumor metastasis of human colorectal cancer by targeting ITGB3. Cell Physiol Biochem, 39, 1165–1176.

60.

Wei

Z.W.

, Xia

G.K.

, Wu

, Chen

, Xiang

, Schwarz

R.E.

, et al. (2015). CXCL1 promotes tumor growth through VEGF pathway activation and is associated with inferior survival in gastric cancer. Cancer Lett, 359, 335–343.

61.

Wen

, Xu

J.Z.

, and Wang

X.R.

(2017). Increased expression of miR-15b is associated with clinicopathological features and poor prognosis in cervical carcinoma. Arch Gynecol Obstet, 295, 743–749.

62.

Yajima

, Kumasaka

M.Y.

, and Tamura

(2012). Functional analysis of GNG2 in human malignant melanoma cells. J Dermatol Sci, 68, 172–178.

63.

Yajima

, Kumasaka

M.Y.

, Yamanoshita

, Zou

, Li

, Ohgami

, et al. (2014). GNG2 inhibits invasion of human malignant melanoma cells with decreased FAK activity. Am J Cancer Res, 4, 182–188.

64.

Yang

, Shen

, Li

, Yuan

, Jin

, and Wu

(2016). Silencing of hypoxia inducible factor-1α gene attenuated angiotensin II-induced abdominal aortic aneurysm in apolipoprotein E-deficient mice. Atherosclerosis, 252, 40–49.

65.

Yao

, Guo

, and Gui

(2018). Acetyl-CoA synthetase 2 promotes cell migration and invasion of renal cell carcinoma by upregulating lysosomal-associated membrane protein 1 expression. Cell Physiol Biochem, 44, 984–992.

66.

Zhang

, He

, Jia

, Yan

, and Yang

(2018a) Acetyl-CoA synthetase 2 enhances tumorigenesis and is indicative of a poor prognosis for patients with renal cell carcinoma. Urol Oncol, 36, 243.e9–243.e20.

67.

Zhang

W.L.

, Zhang

J.H.

, Wu

X.Z.

, Yan

, and Lv

(2015). miR-15b promotes epithelial-mesenchymal transition by inhibiting SMURF2 in pancreatic cancer. Int J Oncol, 47, 1043–1053.

68.

Zhang

, Liang

, Zou

, Chen

, Shi

, Wang

, et al. (2018b) Inhibition of miR-155 attenuates abdominal aortic aneurysm in mice by regulating macrophage-mediated inflammation. Biosci Rep, 38, pii:BSR20171432.

69.

Zhang

, Zou

, Chen

, Lu

, Liu

, Zhai

, et al. (2019). Knockdown of lncRNA PVT1 inhibits vascular smooth muscle cell apoptosis and extracellular matrix disruption in a murine abdominal aortic aneurysm model. Mol Cells, 42, 218–227.

70.

Zhao

, Sun

, Guan

, Wang

, Sang

, Liu

, et al. (2016). Disulfiram and diphenhydramine hydrochloride upregulate miR-30a to suppress IL-17-associated autoimmune inflammation. J Neurosci, 36, 9253–9266.

71.

Zhu

, Ye

, Nie

, Ashraf

, Zohaib

, Duan

, et al. (2015). MicroRNA-15b modulates japanese encephalitis virus-mediated inflammation via targeting RNF125. J Immunol, 195, 2251–2262.