circRNA/lncRNA-miRNA-mRNA Network in Oxidized,Low-Density,Lipoprotein-Induced Foam Cells

Abstract

Although long noncoding RNAs (lncRNAs) and circular RNAs (circRNAs) have been suggested to play important roles in the pathogenesis of diseases, atherosclerosis-related lncRNAs and circRNAs remain rarely reported. This study aimed to explore the underlying molecular mechanisms of atherosclerosis based on the competing endogenous RNA (ceRNA) regulatory hypothesis of lncRNAs and circRNAs. The expression profiles of circRNAs, lncRNAs, and mRNAs in human THP-1 macrophages treated with oxidized low-density lipoprotein (an in vitro atherosclerosis model), or not, were obtained from the Gene Expression Omnibus database under accession numbers GSE107522, GSE54666, and GSE54039, respectively. The present study identified 29 differentially expressed circRNAs in GSE107522, 544 differentially expressed genes (DEGs) in GSE54666, and 502 DEGs and 231 differentially expressed lncRNAs in GSE54039 datasets by using the Linear Models for Microarray Data method. Eight DEGs were found to be shared and expressed with the consistent trend in GSE54666 and GSE54039 datasets. Two of them (ASPH, aspartate beta-hydroxylase; and PDE3B, phosphodiesterase 3B) were suggested to be crucial based on functional enrichment, protein–protein interaction, and ceRNA network analyses. ASPH, through interaction with CACNA2D4 (calcium voltage-gated channel auxiliary subunit alpha2delta 4), may be associated with atherosclerosis by regulating the cellular response to calcium ion; and PDE3B may exert roles in negative regulation of angiogenesis through cross talk with ELMO1 (engulfment and cell motility 1). Furthermore, the expression of ASPH and PDE3B may be regulated by hsa_circ_0028198/hsa_circ_0092317/XIST-miR-543; PDE3B expression may be also modulated by hsa_circ_0092317/hsa_circ_0003546/H19/XIST-miR-326. In conclusion, our identified ceRNA interaction axes may possibly be important targets for treatment of atherosclerosis.

Introduction

Atherosclerosis is an important risk factor to induce ischemic cardiovascular diseases, including myocardial infarction and stroke, which are two leading causes for mortality and morbidity worldwide (Mathers and Loncar, 2006; Barquera et al., 2015). The hallmark of early atherosclerosis is the formation of foam cells that are resulted from uptake and accumulation of oxidized low-density lipoprotein (ox-LDL) in arterial macrophages (Yu et al., 2013; Delgadomaroto et al., 2017). Thus, how to prevent foam cell formation and slow the progression of arteriosclerotic processes has been a rising public concern.

Recently, emerging evidence has suggested that noncoding RNAs (ncRNAs), including microRNAs (miRNAs), long noncoding RNAs (lncRNAs), and circular RNAs (circRNAs), play crucial roles in development of various diseases (Manel, 2011; Beermann et al., 2016). miRNAs negatively regulate the expression of mRNAs of protein-coding genes by binding to the complementary sequences located in the 3′-untranslated region (UTR) (Hausser and Zavolan, 2014); lncRNAs/circRNAs could function as competing endogenous RNAs (ceRNAs) to competitively bind with miRNAs through their miRNA response elements and then indirectly regulate the expression of miRNA-targeted mRNAs (Salmena et al., 2011). Therefore, lncRNA/circRNA-miRNA-mRNA interaction may be an important mechanism for initiation and development of atherosclerosis. This hypothesis has been preliminarily demonstrated by the following studies: Ye et al. (2019) identified that upregulated lncRNA myocardial infarction associated transcript (MIAT) could sponge miR-149-5p and prevent the inhibition effects of miR-149-5p on CD47, leading to upregulation of the antiphagocytic molecule, CD47, in advanced atherosclerosis. Knockdown of MIAT attenuated the progression of atherosclerosis, reduced the necrotic core size, increased plaque stability in vivo, and promoted the clearance of apoptotic cells by macrophages in vitro (Ye et al., 2019). Yan et al. (2017) found that ox-LDL increased MEG3 expression in Raw264.7 macrophage cells. MEG3 could sponge miR-204, while miR-204 could regulate the expression of cyclin-dependent kinase inhibitor 2A in Raw264.7 cells treated with ox-LDL. By RNA-Seq analysis and ceRNA network construction, Zhang et al. (2018a) identified ocu-cirR-novel-18038, -18298, -15993, -17934, -17879, -18036, and -14389 as crucial circRNAs related to atherosclerosis. However, atherosclerosis-related lncRNA/circRNA-miRNA-mRNA ceRNA regulatory mechanisms remain rarely reported until now, and no lncRNAs and circRNAs that interact with the same miRNAs and targets were investigated.

The goal of this study was to further screen crucial lncRNA/circRNA-miRNA-mRNA ceRNA axes in ox-LDL-induced foam cells by using microarray data collected from a public database and a series of bioinformatic analysis. The results of the present study may deepen the understanding of the molecular mechanisms of atherosclerosis and provide some targets for developing therapeutic approaches.

Materials and Methods

Downloading public microarray datasets

By searching the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus repository (GEO) database with the key words, ox-LDL and foam cells, three microarray datasets involving human THP-1 macrophages were obtained, including GSE107522, GSE54666 (Reschen et al., 2015), and GSE54039 (Hu et al., 2014). In the GSE107522 dataset, the authors analyzed the circRNA expression profiles in three THP-1 macrophages and three foam cells induced by stimulation of THP-1 macrophages with ox-LDL (50 mg/mL) for 24 h using Agilent-069978 Arraystar Human CircRNA microarray, V1 (platforms: GPL19978). In the GSE54666 dataset, six independent biological replicates, including half of the THP1 macrophages treated with ox-LDL (50 μg/mL) for 48 h to form foam cells and the other half treated with the control buffer, were obtained to detect the mRNA expression profiles using Illumina HumanHT-12, V4.0, Expression BeadChip (platforms: GPL10558). In the GSE54039 dataset, lncRNA/mRNA expression profiling analysis for three THP-1 macrophages and three THP-1 macrophage-derived foam cells (50 μg/mL of ox-LDL treatment for 48 h) was performed using Arraystar Human LncRNA microarray, V2.0 (Agilent-033010 Feature Number version) (platforms: GPL13825).

Differential analysis

The preprocessed expression matrix of each dataset was downloaded from GEO, with the quantile normalization method using the GeneSpring GX, v11.5.1, software package for GSE107522 and GSE54039, while robust spline normalization was performed for GSE54666. The gene differential expression analysis was carried out with the Linear Models for Microarray Data (LIMMA) method software (version 3.34.0) (Ritchie et al., 2015). The differentially expressed genes (DEGs), differentially expressed circRNAs (DECs), and differentially expressed lncRNAs (DELs) were annotated according to the corresponding platform information. A p-value <0.05 and |logFC (fold change)| > 0.5 were defined as statistical threshold values due to the fact that results of the false discovery rate were not perfect. Furthermore, this cutoff point was selected to prevent the rare interaction pairs screened between RNAs. A heatmap was rendered using the R package pheatmap (version: 1.0.8) based on Euclidean distance to visualize the distinguishing ability of DEGs, DECs, and DELs for samples.

Protein–protein interaction network analysis

The Search Tool for the Retrieval of Interacting Genes/Proteins (STRING; version 10.0) database (Szklarczyk et al., 2015) was used to retrieve the protein interactions between DEGs identified in GSE54666 and GSE54039 datasets. Only STRING-derived interactors with high confidence (≥0.700) were imported into the Cytoscape software (version 3.6.1) (Kohl et al., 2011) to construct the protein–protein interaction (PPI) network.

circRNA/lncRNA-miRNA-mRNA ceRNA network construction

The interaction relationships between miRNAs and DEGs were predicted using the miRWalk database (version 2.0), which contained 12 algorithms (Dweep and Gretz, 2015). miRNAs that interacted with DECs were predicted using the CircInteractome database (version 1.0) (Dudekula et al., 2016). miRNAs that interacted with DELs were predicted using the miRcode database (version 1.0) (Jeggari et al., 2012). DEC and DEL-related ceRNA networks were constructed based on the commonly interactive miRNAs with DEGs, DECs, and DELs. The circRNA/lncRNA-miRNA-mRNA ceRNA network was visualized using Cytoscape software (version 3.6.1) (Kohl et al., 2011).

Functional enrichment analysis

Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway and gene ontology (GO) enrichment analyses were done using the Database for Annotation, Visualization and Integrated Discovery (DAVID) online tool (version 6.8) (Huang et al., 2009). Pathways or GO biological process terms with p-values of less than 0.05 were considered to be significantly enriched.

Results

Differential expression analysis

According to the cutoff criteria (p < 0.05 and |logFC| > 0.5), a total of 29 DECs (21 upregulated and 8 downregulated) and 544 DEGs (236 upregulated and 308 downregulated) were, respectively, identified between macrophages treated with and without ox-LDL in GSE107522 and GSE54666 datasets. The GSE54039 datasets included mRNA and lncRNA data, in which 502 DEGs (264 upregulated and 238 downregulated) and 231 DELs (63 upregulated and 168 downregulated) were, respectively, examined when comparing macrophages undergoing ox-LDL treatment with ox-LDL nonstimulation. Furthermore, 12 DEGs were found to be shared in GSE54666 and GSE54039 datasets, among which 8 (FARP1, FERM, ARH/RhoGEF, and pleckstrin domain protein 1; FCGR2B, Fc fragment of IgG receptor IIb; ASPH, aspartate beta-hydroxylase; NAV2, neuron navigator 2; OGFRL1, opioid growth factor receptor-like 1; ST14, suppression of tumorigenicity 14; HMGCS1, 3-hydroxy-3-methylglutaryl-CoA synthase 1; and PDE3B, phosphodiesterase 3B) were expressed with the consistent trend, while 4 (IGF1, insulin-like growth factor 1; DUSP1, dual-specificity phosphatase 1; ATP2B4, ATPase plasma membrane Ca2+ transporting 4; and HS3ST2, heparan sulfate-glucosamine 3-sulfotransferase 2) were opposite in expression trend in 2 databases (Table 1). The hierarchical clustering heatmap visually revealed that THP-1 macrophages and foam cells displayed distinguishably characteristic circRNA (Fig. 1A), mRNA (Fig. 1B, C), and lncRNA (Fig. 1D) expression profiles.

FIG. 1.

Hierarchical clustering and heatmap analysis of differentially expressed circRNAs (A), mRNAs (B, C), and lncRNAs (D). Green, low expression; and red, high expression. circRNA, circular RNA; lncRNA, long noncoding RNA. Color images are available online.

Table 1.

Crucial Differentially Expressed Genes, lncRNAs, and circRNAs Between Macrophages Treated With and Without Oxidized Low-Density Lipoprotein

mRNA					lncRNA			circRNA
	GSE54039		GSE54666			logFC	p		logFC	p
SYMBOL	logFC	p	logFC	p	AC005392.13	−2.05	2.89E-03	hsa_circ_0008896	2.19	1.31E-02
NAV2	1.65	1.16E-02	0.76	3.69E-03	C10orf55	−2.29	3.11E-03	hsa_circ_0003645	2.01	4.30E-02
HMGCS1	−1.39	1.78E-02	−1.54	2.78E-08	LINC00355	−1.47	2.29E-02	hsa_circ_0050486	1.79	2.93E-02
OGFRL1	1.45	2.41E-02	0.55	1.49E-05	OGT	−1.35	4.15E-02	hsa_circ_0005699	1.68	3.51E-02
HS3ST2	−1.40	2.74E-02	0.92	9.86E-03	PCA3	−3.81	2.64E-03	hsa_circ_0007478	1.40	3.57E-03
FCGR2B	−2.00	3.13E-02	−0.71	3.11E-02	PI4KA	−2.19	9.09E-03	hsa_circ_0064924	1.17	3.35E-02
FARP1	1.27	3.48E-02	0.65	2.56E-03	SERHL	−2.72	4.52E-02	hsa_circ_0026218	1.14	5.96E-03
ATP2B4	1.45	3.54E-02	−0.65	1.31E-04	COX10-AS1	1.26	4.53E-02	hsa_circ_0092283	1.10	3.75E-02
IGF1	1.99	3.58E-02	−0.66	1.38E-03	FAM60A	1.87	2.25E-02	hsa_circ_0092327	1.02	1.26E-02
PDE3B	1.67	3.95E-02	0.52	2.26E-02	MALAT1	2.68	1.20E-02	hsa_circ_0029605	0.88	3.56E-02
ST14	1.27	3.96E-02	0.83	9.61E-04	RAD54L2	1.6	3.95E-02	hsa_circ_0001292	0.84	3.28E-02
DUSP1	−1.24	4.50E-02	0.51	6.26E-04	XIST	1.63	4.62E-02	hsa_circ_0055201	0.75	1.11E-02
ASPH	3.10	4.73E-02	0.55	6.70E-03	ZBTB20-AS1	2.1	4.85E-02	hsa_circ_0092317	0.72	4.41E-02
JUN			0.58	8.98E-05	ZNF224	1.66	1.01E-02	hsa_circ_0028198	0.71	4.66E-02
HMGCR			−1.40	4.58E-08	C8orf31	1.47	2.31E-02	hsa_circ_0070284	0.69	2.61E-02
FABP1	−3.07	4.62E-02			KIAA1033	2.23	3.07E-02	hsa_circ_0011692	0.65	3.05E-02
MSMO1	2.35	7.69E-03			TTTY2B	1.93	1.34E-02	hsa_circ_0007085	0.63	3.56E-03
MVD			−0.76	2.70E-07	DBF4B	1.52	1.40E-02	hsa_circ_0008659	0.615	2.98E-02
SQLE			−2.03	1.83E-09	H19	2.16	7.38E-03	hsa_circ_0003546	0.57	1.59E-02
GMPR2	−1.11	4.63E-02			AC004893.11	1.94	2.04E-02	hsa_circ_0003505	0.56	3.30E-02
ADCY10	1.37	1.50E-02			CLPTM1	1.62	3.75E-02	hsa_circ_0026978	0.54	4.76E-02
ENTPD1			−0.74	2.09E-03	EFCAB6-AS1	1.55	3.96E-02	hsa_circ_0065149	−0.53	1.77E-02
CAMK2D	1.365	3.86E-02			SYMPK	1.60	2.73E-02	hsa_circ_0006501	−0.53	5.97E-03
CAMK2B			−0.56	1.34E-03	RP11-262H14.3	3.35	2.95E-04	hsa_circ_0049700	−0.62	4.33E-02
ELMO1			−0.50	9.87E-05	NR2F1-AS1	2.97	1.71E-02	hsa_circ_0048492	−0.63	2.05E-02
ACAT2			−0.85	3.34E-02	AC009299.3	2.71	1.22E-02	hsa_circ_0011279	−0.65	2.20E-02
CACNA2D4			−0.67	2.04E-04	RP11-158H5.8	2.69	2.0E-02	hsa_circ_0082139	−0.68	1.89E-02
LILRA5			−1.00	4.91E-06	AC008268.2	2.68	3.60E-04	hsa_circ_0040921	−0.82	3.90E-02
PLIN2			2.68	4.36E-06	RP11-276H19.1	2.32	3.86E-02	hsa_circ_0079485	−0.86	4.44E-02

circRNA, circular RNA; FC, fold change; lncRNA, long noncoding RNA.

Functional enrichment for DEGs

Considering that common DEGs were less in GSE54666 and GSE54039 datasets, KEGG pathway enrichment was performed for DEGs identified in these two datasets, respectively. The results showed that DEGs in GSE54666 were enriched into 20 KEGG pathways, such as hsa04923:Regulation of lipolysis in adipocytes (PDE3B); hsa04921:Oxytocin signaling pathway (CAMK2B, CACNA2D4); and hsa01212:Fatty acid metabolism (ACAT2); while 5 KEGG pathways were obtained for the DEGs in GSE54039 such as hsa05205:Proteoglycans in cancer (CAMK2D) (Table 2).

Table 2.

Kyoto Encyclopedia of Genes and Genomes Pathway Enrichment Genes for Differentially Expressed Genes

Category	Term	p	Genes
GSE54666	hsa00100:Steroid biosynthesis	1.32E-07	TM7SF2, SOAT1, EBP, SQLE, CYP51A1, DHCR7, NSDHL, FDFT1, DHCR24
	hsa04145:Phagosome	1.04E-05	OLR1, NCF1, NCF4, COLEC12, HLA-DQA1, CORO1A, CD36, FCGR2B, CD209, HLA-DPA1, HLA-DPB1, CLEC7A, HLA-DOA, DYNC1H1, TUBA1A, THBS2, CD14, TUBB3
	hsa00900:Terpenoid backbone biosynthesis	6.54E-05	MVD, HMGCR, FDPS, HMGCS1, MVK, ACAT2, IDI1
	hsa04380:Osteoclast differentiation	4.43E-04	IL1R1, NCF1, NCF4, PPARG, TGFBR2, LILRB2, CYBB, TNFRSF11A, FCGR2B, LILRB5, JUN, LILRA5, NFATC1, CSF1R
	hsa01130:Biosynthesis of antibiotics	7.68E-04	TM7SF2, MVD, CYP51A1, HMGCR, ALDOC, HMGCS1, FDPS, ACAT2, ALDH3A2, FDFT1, G6PD, SDS, SQLE, PHGDH, PLA2G7, MVK, IDI1, NSDHL
	hsa05202:Transcriptional misregulation in cancer	1.46E-03	MAF, SLC45A3, MEF2C, NFKBIZ, PPARG, TGFBR2, FOXO1, TSPAN7, IGF1, BCL2L1, JUP, NUPR1, ETV5, CD14, CSF1R
	hsa03320:PPAR signaling pathway	1.74E-03	CD36, OLR1, CYP27A1, PPARG, FABP4, FADS2, DBI, FABP5, NR1H3
	hsa05150:Staphylococcus aureus infection	2.08E-03	C1QA, C1QB, FCGR2B, C5, HLA-DPA1, HLA-DPB1, HLA-DOA, HLA-DQA1
	hsa04640:Hematopoietic cell lineage	2.51E-03	CD38, IL1R1, CD36, CD33, KITLG, CD4, CD1A, ITGA3, CD14, CSF1R
	hsa05321:IBD	5.48E-03	MAF, JUN, IL21R, HLA-DPA1, HLA-DPB1, HLA-DOA, HLA-DQA1, NFATC1
	hsa04923:Regulation of lipolysis in adipocytes	1.09E-02	ADCY1, ADRB2, PTGS1, ABHD5, PDE3B, MGLL, FABP4
	hsa04921:Oxytocin signaling pathway	1.23E-02	MEF2C, KCNJ5, CD38, ADCY1, RGS2, JUN, RCAN1, CAMK2B, KCNJ2, CACNA2D3, NFATC1, CACNA2D4
	hsa05152:Tuberculosis	1.62E-02	CORO1A, CARD9, FCGR2B, CD209, CTSD, CAMK2B, HLA-DPA1, CLEC7A, HLA-DPB1, HLA-DOA, CD14, CD74, HLA-DQA1
	hsa05310:Asthma	1.74E-02	FCER1A, HLA-DPA1, HLA-DPB1, HLA-DOA, HLA-DQA1
	hsa01212:Fatty acid metabolism	2.20E-02	ACADVL, FADS1, ACACA, FASN, FADS2, ACAT2
	hsa05200:Pathways in cancer	2.39E-02	ADCY1, EPAS1, PDGFB, PPARG, TGFBR2, SKP2, FZD1, KITLG, IGF1, FOXO1, GNG11, ITGA3, RB1, BCL2L1, JUP, LPAR5, CXCR4, ETS1, JUN, GNG2, LAMC1, CSF1R
	hsa05020:Prion diseases	2.65E-02	C1QA, C1QB, C5, LAMC1, CCL5
	hsa04610:Complement and coagulation cascades	2.82E-02	C1QA, C1QB, VWF, F13A1, C5, TFPI, PLAUR
	hsa05140:Leishmaniasis	3.19E-02	NCF1, JUN, NCF4, HLA-DPA1, HLA-DPB1, HLA-DOA, HLA-DQA1
	hsa05166:HTLV-I infection	4.77E-02	IL1R1, ADCY1, PDGFB, TGFBR2, FZD1, FDPS, RB1, BCL2L1, HLA-DQA1, ETS1, JUN, HLA-DPA1, HLA-DPB1, HLA-DOA, NFATC1
GSE54039	hsa05205:Proteoglycans in cancer	2.47E-02	FGFR1, WNT7B, TNF, ROCK2, ITGAV, HCLS1, TP53, CAMK2D, IGF1, IGF2, FRS2, WNT8A
	hsa04730:Long-term depression	2.59E-02	PPP2R1B, PLA2G4A, LYN, GNA11, IGF1, CACNA1A
	hsa04721:Synaptic vesicle cycle	3.12E-02	ATP6V1A, SYT1, SLC18A2, RIMS1, SNAP25, CACNA1A
	hsa05218:Melanoma	4.85E-02	FGF19, FGFR1, TP53, IGF1, FGF13, EGF
	hsa05200:Pathways in cancer	4.90E-02	FGF19, DCC, FGFR1, ROCK2, GNA11, TP53, TGFB3, IGF1, BDKRB1, FGF13, STK4, WNT7B, NCOA4, ITGAV, CASP8, EGF, TRAF5, WNT8A

IBD, inflammatory bowel disease.

In addition, the GO enrichment study was also performed. The results indicated that DEGs in GSE54666 were enriched into 215 GO biological process terms, such as GO:0071277∼cellular response to calcium ion (ASPH), GO:0010524∼positive regulation of calcium ion transport into cytosol (ASPH), GO:0016525∼negative regulation of angiogenesis (PDE3B), and GO:0006915∼apoptotic process (ELMO1), while 75 GO biological process terms were enriched for the DEGs in GSE54039, also including GO:0016525∼negative regulation of angiogenesis (PDE3B), GO:0071277∼cellular response to calcium ion (ASPH), and GO:0070588∼calcium ion transmembrane transport (ASPH) (Table 3).

Table 3.

Gene Ontology Term Enrichment Analysis for Differentially Expressed Genes

	Term	p	Genes
GSE54666	GO:0006695∼cholesterol biosynthetic process	2.88E-16	TM7SF2, EBP, CES1, MVD, CYP51A1, HMGCR, HMGCS1, FDPS, FDFT1, G6PD, SQLE, DHCR7, INSIG1, MVK, IDI1, NSDHL, DHCR24
	GO:0008299∼isoprenoid biosynthetic process	1.90E-08	COQ2, MVD, HMGCR, FDPS, HMGCS1, MVK, IDI1, FDFT1
	GO:0007165∼signal transduction	4.49E-05	IMPA2, ACVRL1, S100A9, PPARG, PDE3B, SDC4, ARHGAP4, TNFRSF11A, EVI2A, ZFYVE16, WWP1, MICAL1, ARHGAP9, NLRP3, ARHGAP23, PLAUR, JUP, CD38, LILRB2, NPC1, ADM, CD33, TXNRD1, NRGN, TNFRSF21, NDP, KITLG, NEDD9, NMB, GNG11, CXXC5, CD74, CCL7, CD4, CAMK2B, INPP5D, HLA-DOA, CSF1R, EBP, EPAS1, TP53BP2, IGF1, RCAN1, LEP, P2RX4, TNFSF10, LYVE1, FCGR2B, CD300A, CHN2, ANTXR1, IGFBP4, VLDLR
	GO:0006629∼lipid metabolic process	6.75E-04	LEP, CD36, G6PD, LDLR, FADS1, PTGS1, PPARG, ABHD5, HMGCS1, MGLL, FADS2, ACAT2, FABP5
	GO:0071277∼cellular response to calcium ion	1.77E-03	AKR1C3, MEF2C, ADCY1, ALOX5AP, JUN, SLC25A23, ASPH
	GO:0006915∼apoptotic process	5.40E-03	MEF2C, TNFRSF21, S100A8, NCF1, ARHGEF6, TGFBR2, S100A9, PIM1, FOXO1, RB1, SGMS1, NLRP3, ELMO1, MEF2D, TNFSF10, TMEM173, NISCH, CXCR4, BOK, FAIM, PAK1, INPP5D, GADD45B, BMF, CD14, DHCR24
	GO:0016525∼negative regulation of angiogenesis	2.04E-02	SERPINF1, STAB1, PDE3B, TCF4, THBS2, VASH1
	GO:0033197∼response to vitamin E	3.04E-02	LEP, MEF2C, HMGCS1
	GO:0010524∼positive regulation of calcium ion transport into cytosol	3.58E-02	P2RX4, CD4, ASPH
	GO:0032869∼cellular response to insulin stimulus	4.59E-02	PPARG, DENND4C, PDE3B, FOXO1, PAK1, LPIN1
GSE54039	GO:0045840∼positive regulation of mitotic nuclear division	4.88E-04	PHIP, EPGN, BTC, IGF1, IGF2, EGF
	GO:0042981∼regulation of apoptotic process	1.42E-03	EGR1, SGK1, IKZF3, GDF6, DEDD, TP53, TGFB3, ATM, DIP2A, DUSP1, TNFRSF10D, CASP8, CHM, TRAF5, FRS2
	GO:0070374∼positive regulation of ERK1 and ERK2 cascade	2.14E-03	FGF19, CCL3, TNF, GCNT2, NPNT, CCL18, CCL25, FGA, TEK, NPTN, CAMK2D, TREM2, PLA2G5
	GO:0010976∼positive regulation of neuron projection development	2.21E-03	FGFR1, LYN, DMD, NTRK2, NPTN, PTN, DPYSL3, RAPGEF1, ABL2
	GO:0018108∼peptidyl-tyrosine phosphorylation	2.23E-03	NTRK3, EPHA4, FGFR1, FLT1, LYN, NTRK2, TEK, BTC, NPTN, DYRK3, EGF, ABL2
	GO:0033138∼positive regulation of peptidyl-serine phosphorylation	2.25E-03	OSM, NTRK3, TNF, ATP2B4, HCLS1, NTRK2, STK4, OPRD1
	GO:0016525∼negative regulation of angiogenesis	5.40E-03	ROCK2, TNMD, NF1, TEK, CX3CR1, PDE3B, PTN
	GO:0007626∼locomotory behavior	6.24E-033	NAV2, SLC6A3, SLC18A2, VPS13A, DBH, SNAP25, MEIS1, CDH23
	GO:0071277∼cellular response to calcium ion	1.03E-02	SYT1, CAMK2D, CPNE1, EEF2K, ASPH, JUNB
	GO:0070588∼calcium ion transmembrane transport	1.26E-02	TRPM3, ATP2B3, ATP2B4, PANX1, CHRNA9, ATP2C1, ITGAV, ASPH, CACNA1A

Only top 10 are listed.

GO, gene ontology.

PPI network

When uploading the DEGs identified in both GSE54666 and GSE54039 datasets to STRING, 1217 high-confidence interaction relationships were found to be present in 448 DEGs, which were used to construct the PPI network. In this network, ASPH could interact with CACNA2D4; PDE3B could interact with ELMO1, CAMK2D, and CAMK2B; and HMGCS1 and HMGCR were interacted (Fig. 2), further revealing the underlying upstream or downstream mechanisms for crucial genes.

FIG. 2.

Protein and protein interaction network for differentially expressed genes. Color images are available online.

ceRNA network construction

There were 585 interaction pairs predicted between 29 DECs and 230 miRNAs by the CircInteractome database; there were 3995 interaction pairs predicted between 56 DELs and 153 miRNAs by the miRcode database; and there were 3208 interaction pairs predicted between 8 common DEGs in 2 datasets and 931 miRNAs by the miRWalk 2.0 database based on at least 6 algorithms. By integrating with the common miRNAs, circRNA-miRNA-mRNA (including 94 miRNAs, 8 DEGs, and 29 DECs) (Fig. 3) and lncRNA-miRNA-mRNA (including 54 miRNAs, 8 DEGs, and 29 DELs) (Fig. 4) regulatory networks were constructed.

FIG. 3.

ceRNA interaction network of circRNA-miRNA-mRNA. Circular nodes represent circRNAs; regular triangle nodes represent mRNAs; and inverted triangle nodes represent miRNAs. ceRNAs, competing endogenous RNAs. miRNA, microRNA. Color images are available online.

FIG. 4.

ceRNA interaction network of lncRNA-miRNA-mRNA. Square nodes represent lncRNAs; regular triangle nodes represent mRNAs; and inverted triangle nodes represent miRNAs. Color images are available online.

In the circRNA-related ceRNA network (Fig. 3), ASPH can be targeted by 43 miRNAs, while 41 of them could also interact with 18 upregulated DECs (such as hsa_circ_0028198 and hsa_circ_0092317), resulting in the ceRNA relationship pairs of hsa_circ_0028198/hsa_circ_0092317-miR-543-ASPH; PDE3B can be regulated by 15 miRNAs, all of which could also interact with 17 upregulated DECs (such as hsa_circ_0092317, hsa_circ_0028198, and hsa_circ_0003546), forming the following ceRNA relationships of hsa_circ_0092317/hsa_circ_0003546-miR-326-PDE3B and hsa_circ_0028198/hsa_circ_0092317-miR-543-PDE3B.

In the lncRNA-related ceRNA network (Fig. 4), ASPH can be regulated by 30 miRNAs, while these 30 miRNAs could also interact with 18 upregulated DELs (such as XIST), which constituted the ceRNA axis of XIST-miR-543-ASPH. PDE3B can be regulated by 11 miRNAs; meanwhile, all these miRNAs could also interact with 17 upregulated DELs (such as H19 and XIST), with the ceRNA relationships of MALAT1/XIST-miR-543-PDE3B and H19/XIST-miR-326-PDE3B. These crucial ceRNA axes are shown in Figure 5 after integration of lncRNA and circRNA ceRNA networks.

FIG. 5.

Integrated competing endogenous RNA (ceRNA) interaction network of circRNA/lncRNA-miRNA-mRNA based on the common miRNAs. Square nodes represent lncRNAs; circular nodes represent circRNAs; regular triangle nodes represent mRNAs; and inverted triangle nodes represent miRNAs. Color images are available online.

Discussion

Although eight common DEGs were identified in two microarray datasets, two DEGs (ASPH and PDE3B) may be particularly crucial based on functional enrichment and ceRNA network analyses. ASPH, through interaction with CACNA2D4, may be associated with atherosclerosis by regulating the cellular response to calcium ion; and PDE3B may exert roles in negative regulation of angiogenesis through cross talk with ELMO1. Furthermore, the expression of ASPH and PDE3B may be regulated by hsa_circ_0028198/hsa_circ_0092317/XIST-miR-543; PDE3B expression may also be modulated by hsa_circ_0092317/hsa_circ_0003546/H19/XIST-miR-326.

Phosphodiesterase (PDE) is an enzyme that is responsible for hydrolysis of cyclic adenosine monophosphate (cAMP) (Soderling and Beavo, 2000). cAMP is an important second messenger that activates cAMP-dependent protein kinases (such as PKA) and then contributes to phosphorylation of ABCA1 (ATP-binding cassette transporter A1) to mediate the cholesterol transport process and prevent the formation of foam cells (Liao et al., 2017). Thus, the upregulation of phosphodiesterase genes may be an underlying mechanism to promote the development of atherosclerosis. This hypothesis has been demonstrated previously (Nagaoka et al., 1998; Priksz et al., 2018). For example, Priksz et al. (2018) observed that the level of PDE9A enzyme was significantly elevated in left ventricle samples of animals that received atherogenic rabbit chow supplemented with 1% cholesterol and 1% fat compared with control. Nagaoka et al. (1998) found that the expression and activity of the PDE3A gene were significantly increased in the aortas of atherosclerosis-prone insulin-resistant rats. In addition, Kilanowska and Szkudelski (2019) also proved that inhibition of PDE3B promoted insulin release from isolated pancreatic islets and possibly decreased the risk of diabetes-induced arteriosclerosis (Muhammed et al., 2012; Kilanowska and Szkudelski, 2019). In line with these studies, we also identified that PDE3B was an upregulated gene in macrophages treated with ox-LDL. However, the mechanisms of PDE3B in atherosclerosis remain rarely investigated. In this study, we found that PDE3B could interact with ELMO1. It has been reported that ELMO1/Dock180 proteins form an evolutionarily conserved signaling module to function as a bipartite guanine nucleotide exchange factor for the small GTPase Rac that mediates physical engulfment of apoptotic cells by macrophages (Elliott and Ravichandran, 2010; Schäker et al., 2015). The reduced macrophages due to transformation into foam cells may lead to accumulation of apoptosis, which was also an important pathogenesis for atherosclerosis (Li et al., 2017). Thus, ELMO1 may be downregulated in foam cells, which was also demonstrated in our study. Furthermore, deregulation of PDE3B expression might promote pancreatic islet cell apoptosis (Muhammed et al., 2012). These findings indicate that upregulated PDE3B may be involved in atherosclerosis by downregulation of ELMO1 and then promoting cell apoptosis.

Considering the important roles of PDE in atherosclerosis, PDE inhibitors are hypothesized to be potential drugs for alleviation of atherosclerosis. This hypothesis has been confirmed in current studies by several scholars: clinical trials showed that use of the PDE III selective inhibitor, cilostazol (100–200 mg/day), significantly reduced the mean intima-media thickening of both common carotid arteries and internal carotid arteries, total plaque area (TPA) in patients with a TPA less than 0.5 cm², and the noncalcified plaque component (Katakami et al., 2010; Kim and Cho, 2013; Lee et al., 2019). In vivo experiments showed that cilostazol decreased triglycerides, increased HDL cholesterol levels, and inhibited foam cell formation (Tani et al., 2000; Okutsu et al., 2009; Park and Heo, 2018). However, attention was also paid to drug complications (myocardial infarction, stent thrombosis, bleeding, rash, and gastrointestinal side effects) and long-term use-induced drug resistance (Sakurai et al., 2013). Therefore, it is necessary to further develop therapeutic approaches that targeted suppression of PDE3B. miRNAs are recently extensively studied ncRNAs that inhibit the expression of target genes by binding to their 3′ UTR (Hausser and Zavolan, 2014). Due to endogenous characteristics, miRNAs have potential for development of novel drugs without major complications. In this study, we also predicted the upstream miRNAs for PDE3B, such as miR-543 and miR-326. Although there was no direct evidence to show the roles of miR-543 and miR-326 in atherosclerosis, they may be speculated to be upregulated according to reports on diabetes (Santovito et al., 2014; Azhir et al., 2018) and inflammatory diseases (Navarro-Quiroz et al., 2017; Wu et al., 2018), which are risk factors for atherosclerosis (Pleskovič et al., 2017). Furthermore, lncRNAs/circRNAs could function as ceRNAs to influence the inhibition effects of miRNAs on target genes (Qu et al., 2015). Hereby, the investigation of lncRNAs/circNAs may also provide targets for development of novel drugs. In this study, we predicted that hsa_circ_0003546/hsa_circ_0092317 and H19/XIST could interact with miR-326, while hsa_circ_0028198/hsa_circ_0092317 and XIST could interact with miR-543. There were studies to report that lncRNA H19 promoted atherosclerosis by regulating MAPK and NF-κB inflammatory signaling pathways (Pan, 2017); knockdown of XIST and H19 may suppress development of coronary atherosclerotic heart disease by suppressing cell apoptosis induced by ox-LDL (Xu et al., 2018; Zhang et al., 2018b). The roles of these circRNAs have not been explored previously, revealing that they may be novel targets for atherosclerosis. The hsa_circ_0003546 parental gene is DENND5A, and the study by Zhang et al. (2018a) showed that ocu-cirR-novel-10230 (corresponding to DENND5A) was significantly upregulated in high-fat diet-induced atherosclerosis rabbits (logFC = 2.59, p = 0.0347), indirectly proving that hsa_circ_0003546 was highly expressed in foam cells, which was in line with our study. No studies were performed to validate the association of hsa_circ_0028198/hsa_circ_0092317 with atherosclerosis and thus they may be novel targets.

ASPH encodes aspartate beta-hydroxylase, which is a type II transmembrane protein and a kind of α-ketoglutarate-dependent dioxygenase. Recently, most of the studies focused on the cancer-promoting effects of ASPH (Sturla et al., 2016; Huang et al., 2018) and only a small portion of studies indicated that recombinant ASPH reduced the viability of natural killer (NK) cells and influenced its cytotoxicity (Huyan et al., 2014). Decreased NK cell activity was reported to be associated with atherosclerosis in elderly humans (Bruunsgaard et al., 2002). Thus, ASPH may be involved in atherosclerosis by decreasing the numbers of NK cells. However, the roles of ASPH remain rarely investigated. In our study, we predicted that ASPH may interact with CACNA2D4 to participate in foam cell formation. CACNA2D4 is a part of the CACN gene family responsible for encoding voltage-dependent calcium channels; its deletion was implicated in bipolar disorder and schizophrenia (Van Den Bossche et al., 2012). The youth with major bipolar disorder are predisposed to develop accelerated atherosclerosis and early cardiovascular diseases (Goldstein et al., 2016). In line with these studies, we also found that CACNA2D4 was downregulated in foam cells. Therefore, targeted inhibition of ASPH may be an underlying approach for treatment of atherosclerosis. Furthermore, we also demonstrated that hsa_circ_0028198/hsa_circ_0092317/XIST-miR-543 may be involved in regulating ASPH through the ceRNA mechanism, indicating that hsa_circ_0028198/hsa_circ_0092317/XIST-miR-543 may be also potential targets for inhibition of ASPH and alleviation of atherosclerosis, as described above.

However, there were some limitations in this study. First, although there is evidence to reveal the associations of H19/XIST/hsa_circ_0003546/miR-326/PDE3B with atherosclerosis, no experiments (dual-luciferase reporter assay, overexpression, or knockout of genes) were performed to validate the ceRNA relationships among them. Second, although our study preliminarily screened ASPH, which may be involved in atherosclerosis, further clinical in vitro and in vivo experiments are necessary to confirm its expression and function mechanisms.

Availability of Data and Materials

The sequencing datasets, GSE107522, GSE54666, and GSE54039, were downloaded from the GEO database in NCBI.

Footnotes

Disclosure Statement

No competing financial interests exist.

Funding Information

No funding was received for this article.

References

Azhir

, Dehghanian

, and Hojati

(2018). Increased expression of microRNAs, miR-20a and miR-326 in PBMCs of patients with type 1 diabetes. Mol Biol Rep, 45, 1973–1980.

Barquera

, Pedroza-Tobías

, Medina

, Hernández-Barrera

, Bibbins-Domingo

, Lozano

, et al. (2015). Global overview of the epidemiology of atherosclerotic cardiovascular disease. Arch Med Res 46, 328.

Beermann

, Piccoli

M.T.

, Viereck

, and Thum

(2016). Non-coding RNAs in development and disease: background, mechanisms, and therapeutic approaches. Physiol Rev 96, 1297.

Bruunsgaard

, Pedersen

A.N.

, Schroll

, Skinhøj

, and Pedersen

B.K.

(2002). Decreased natural killer cell activity is associated with atherosclerosis in elderly humans. Exp Gerontol, 37, 127–136.

Delgadomaroto

, Benitez

, Fortelago

, Morell

, Maganto-Garcia

, Souza-Moreira

, et al. (2017). Cortistatin reduces atherosclerosis in hyperlipidemic ApoE-deficient mice and the formation of foam cells. Sci Rep 7, 46444.

Dudekula

D.B.

, Panda

A.C.

, Grammatikakis

, De

, Abdelmohsen

, and Gorospe

(2016). CircInteractome: a web tool for exploring circular RNAs and their interacting proteins and microRNAs. RNA Biol, 13, 34–42.

Dweep

, and Gretz

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

Elliott

M.R.

, and Ravichandran

K.S.

(2010). ELMO1 signaling in apoptotic germ cell clearance and spermatogenesis. Ann N Y Acad Sci, 1209, 30–36.

Goldstein

B.I.

, Carnethon

M.R.

, Matthews

K.A.

, McIntyre

R.S.

, Miller

G.E.

, Raghuveer

, et al. (2016). Major depressive disorder and bipolar disorder predispose youth to accelerated atherosclerosis and early cardiovascular disease: a scientific statement from the American Heart Association. Circulation, 132, 965–986.

10.

Hausser

, and Zavolan

(2014). Identification and consequences of miRNA-target interactions—beyond repression of gene expression. Nat Rev Genet 15, 599.

11.

Y.W.

, Yang

J.Y.

, Ma

, Chen

Z.P.

, Hu

Y.R.

, Zhao

J.Y.

, et al. (2014). A lincRNA-DYNLRB2–2/GPR119/GLP-1R/ABCA1-dependent signal transduction pathway is essential for the regulation of cholesterol homeostasis. J Lipid Res, 55, 681–697.

12.

Huang

C.K.

, Iwagami

, Zou

, Casulli

, Lu

, Nagaoka

, et al. (2018). Aspartate beta-hydroxylase promotes cholangiocarcinoma progression by modulating RB1 phosphorylation. Cancer Lett, 429, 1–10.

13.

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.

14.

Huyan

, Li

, Ye

L.J.

, Yang

, Xue

X.P.

, Zhang

M.J.

, et al. (2014). Inhibition of human natural killer cell functional activity by human aspartyl β-hydroxylase. Int Immunopharmacol, 23, 452–459.

15.

Jeggari

, Marks

D.S.

, and Larsson

(2012). miRcode: a map of putative microRNA target sites in the long non-coding transcriptome. Bioinformatics, 28, 2062–2063.

16.

Katakami

, Kim

Y.S.

, Kawamori

, and Yamasaki

(2010). The phosphodiesterase inhibitor cilostazol induces regression of carotid atherosclerosis in subjects with type 2 diabetes mellitus: principal results of the Diabetic Atherosclerosis Prevention by Cilostazol (DAPC) study: a randomized trial. Circulation, 52, 1421–1422.

17.

Kilanowska

, and Szkudelski

(2019). Effects of inhibition of phosphodiesterase 3B in pancreatic islets on insulin secretion: a potential link with some stimulatory pathways. Arch Physiol Biochem [Epub ahead of print]; DOI: 10.1080/13813455.2019.1628071.

18.

Kim

S.M.

, and Cho

K.I.

(2013). Impact of cilostazol on the progression of carotid atherosclerosis in patients with retinal vascular occlusion. Cardiovasc Ther, 31, e94–e101.

19.

Kohl

, Wiese

, and Warscheid

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

20.

Lee

D.H.

, Chun

E.J.

, Oh

T.J.

, Kim

K.M.

, Moon

J.H.

, Choi

S.H.

, et al. (2019). Effect of cilostazol, a phosphodiesterase-3 inhibitor, on coronary artery stenosis and plaque characteristics in patients with type 2 diabetes (ESCAPE study). Diabetes Obes Metab, 21, 1409–1418.

21.

Ying.

, Yang

C.Y.

, Zhang

L.L.

, and Yang

(2017). MicroRNA-210 induces endothelial cell apoptosis by directly targeting PDK1 in the setting of atherosclerosis. Cell Mol Biol Lett 22, 3.

22.

Liao

, Wan

, Zhang

, Shi

, Zhu

, and Chen

(2017). Intermedin ameliorates atherosclerosis by increasing cholesterol efflux through the cAMP-PKA pathway in macrophage RAW264.7 cell line. Med Sci Monit, 23, 5462–5471.

23.

Manel

(2011). Non-coding RNAs in human disease. Nat Rev Genet, 12, 861–874.

24.

Mathers

C.D.

, and Loncar

(2006). Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med 3, e442.

25.

Muhammed

S.J.

, Lundquist

, and Salehi

(2012). Pancreatic β-cell dysfunction, expression of iNOS and the effect of phosphodiesterase inhibitors in human pancreatic islets of type 2 diabetes. Diabetes Obes Metab, 14, 1010–1019.

26.

Nagaoka

, Shirakawa

, Balon

T.W.

, Russell

J.C.

, and Fujita-Yamaguchi

(1998). Cyclic nucleotide phosphodiesterase 3 expression in vivo: evidence for tissue-specific expression of phosphodiesterase 3A or 3B mRNA and activity in the aorta and adipose tissue of atherosclerosis-prone insulin-resistant rats. Diabetes, 47, 1135–1144.

27.

Navarro-Quiroz

, Pacheco-Lugo

, Navarro-Quiroz

, Lorenzi

, España-Puccini

, Díaz-Olmos

, et al. (2017). Profiling analysis of circulating microRNA in peripheral blood of patients with class IV lupus nephritis. PLoS One 12, e0187973.

28.

Okutsu

, Yoshikawa

, Nagasawa

, Hirose

, Takase

, Mitani

, et al. (2009). Cilostazol inhibits modified low-density lipoprotein uptake and foam cell formation in mouse peritoneal macrophages. Atherosclerosis, 204, 405–411.

29.

Pan

J.X.

(2017). LncRNA H19 promotes atherosclerosis by regulating MAPK and NF-kB signaling pathway. Eur Rev Med Pharmacol Sci, 21, 322–328.

30.

Park

K.Y.

, and Heo

T.H.

(2018). Combination therapy with cilostazol and pravastatin improves antiatherogenic effects in low-density lipoprotein receptor knockout mice. Cardiovasc Ther 36, e12476.

31.

Pleskovič

, Letonja

.Š., Vujkovac

A.C.

, Nikolajević Starčević

, Gazdikova

, Caprnda

, et al. (2017). C-reactive protein as a marker of progression of carotid atherosclerosis in subjects with type 2 diabetes mellitus. Vasa, 46, 187–192.

32.

Priksz

, Bombicz

, Varga

, Kurucz

, Gesztelyi

, Balla

, et al. (2018). Upregulation of myocardial and vascular phosphodiesterase 9A in a model of atherosclerotic cardiovascular disease. Int J Mol Sci 19, 2882.

33.

, Li

, Zhong

, and Hu

(2015). Competing endogenous RNA in cancer: a new pattern of gene expression regulation. Int J Clin Exp Med 8, 17110.

34.

Reschen

M.E.

, Gaulton

K.J.

, Da

, Soilleux

E.J.

, Morris

A.J.

, Smyth

S.S.

, et al. (2015). Lipid-induced epigenomic changes in human macrophages identify a coronary artery disease-associated variant that regulates PPAP2B expression through Altered C/EBP-beta binding. PLoS Genet 11, e1005061.

35.

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.

36.

Sakurai

, Koo

B.K.

, Kaneda

, Bonneau

H.N.

, and Nagai

(2013). Cilostazol added to aspirin and clopidogrel reduces revascularization without increases in major adverse events in patients with drug-eluting stents: a meta-analysis of randomized controlled trials. Int J Cardiol, 167, 2250–2258.

37.

Salmena

, Poliseno

, Tay

, Kats

, and Pandolfi

P.P.

(2011). ceRNA hypothesis: the Rosetta Stone of a hidden RNA language?. Cell, 146, 353–358.

38.

Santovito

, De Nardis

, Marcantonio

, Mandolini

, Paganelli

, Vitale

, et al. (2014). Plasma exosome microRNA profiling unravels a new potential modulator of adiponectin pathway in diabetes: effect of glycemic control. J Clin Endocrinol Metab,, 99, 1681–1685.

39.

Schäker

, Bartsch

, Patry

, Stoll

S.J.

, Hillebrands

J.L.

, Wieland

, et al. (2015). The bipartite rac1 Guanine nucleotide exchange factor engulfment and cell motility 1/dedicator of cytokinesis 180 (elmo1/dock180) protects endothelial cells from apoptosis in blood vessel development. J Biol Chem, 290, 6408–6418.

40.

Soderling

S.H.

, and Beavo

J.A.

(2000). Regulation of cAMP and cGMP signaling: new phosphodiesterases and new functions. Curr Opin Cell Biol, 12, 174–179.

41.

Sturla

L.M.

, Tong

, Hebda

, Gao

, Thomas

J.M.

, Olsen

, et al. (2016). Aspartate-β-hydroxylase (ASPH): a potential therapeutic target in human malignant gliomas. Heliyon 2, e00203.

42.

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 43, D447.

43.

Tani

, Uehara

, Sudo

, Marukawa

, Yasuda

, and Kimura

(2000). Cilostazol, a selective type III phosphodiesterase inhibitor, decreases triglyceride and increases HDL cholesterol levels by increasing lipoprotein lipase activity in rats. Atherosclerosis, 152, 299–305.

44.

Van Den Bossche

M.J.

, Strazisar

, De Bruyne

, Bervoets

, Lenaerts

A.S.

, De Zutter

, et al. (2012). Identification of a CACNA2D4 deletion in late onset bipolar disorder patients and implications for the involvement of voltage-dependent calcium channels in psychiatric disorders. Am J Med Genet B Neuropsychiatr Genet, 159B, 465–475.

45.

C.T.

, Huang

, Pei

Z.Y.

, Xi

, and Zhu

G.F.

(2018). MicroRNA-326 aggravates acute lung injury in septic shock by mediating the NF-κB signaling pathway. International Int J Biochem Cell Biol, 101, 1–11.

46.

, Ma

, Liu

, Duan

, and Zhang

(2018). Knockdown of long noncoding RNA XIST alleviates oxidative low-density lipoprotein-mediated endothelial cells injury through modulation of miR-320/NOD2 axis. Biochem Biophys Res Commun, 503, 586–592.

47.

Yan

, Liu

, Yin

, Guo

, and Luo

(2017). Silencing of MEG3 inhibited ox-LDL-induced inflammation and apoptosis in macrophages via modulation of the MEG3/miR-204/CDKN2A regulatory axis. Cell Biol Int, 43, 409–420.

48.

Z.M.

, Yang

, Xia

Y.P.

, Hu

R.T.

, Chen

, Li

B.W.

, et al. (2019). LncRNA MIAT sponges miR-149-5p to inhibit efferocytosis in advanced atherosclerosis through CD47 upregulation. Cell Death Dis 10, 138.

49.

X.H.

, Fu

Y.C.

, Zhang

D.W.

, Yin

, and Tang

C.K.

(2013). Foam cells in atherosclerosis. Clin Chim Acta, 424, 245–252.

50.

Zhang

, Zhang

, Wu

, Li

, Zhang

, et al. (2018a). Comprehensive analysis of circRNA expression pattern and circRNA-miRNA-mRNA network in the pathogenesis of atherosclerosis in rabbits. Aging, 10, 2266–2283.

51.

Zhang

, Cheng

, Yue

, Li

, Zhang

, and He

(2018b). H19 knockdown suppresses proliferation and induces apoptosis by regulating miR-148b/WNT/β-catenin in ox-LDL-stimulated vascular smooth muscle cells. J Biomed Sci 25, 11.