Expression and Functional Analysis of lncRNAs Involved in Platelet-Derived Growth Factor-BB-Induced Proliferation of Human Aortic Smooth Muscle Cells

Abnormal proliferation of vascular smooth muscle cells (VSMCs) is a common feature of many vascular remodeling diseases. Because long non-coding RNAs (lncRNAs) play a critical role in cardiovascular diseases, we analyzed the key lncRNAs that regulate VSMC proliferation. Microarray analysis identified 2,643 differentially expressed lncRNAs (DELs) and 3,720 differentially expressed coding genes (DEGs) between fetal bovine serum (FBS) starvation-induced quiescent human aortic smooth muscle cells (HASMCs) and platelet-derived growth factor-BB (PDGF-BB)-stimulated proliferative HASMCs. Gene Ontology and pathway analyses of the identified DEGs and DELs demonstrated that many lncRNAs were enriched in pathways related to cell proliferation. One of the upregulated lncRNAs in proliferative HASMC was HIF1A anti-sense RNA 2 (HIF1A-AS2). HIF1A-AS2 suppression decreased HASMC proliferation via the miR-30e-5p/CCND2 mRNA axis. We have thus identified key DELs and DEGs involved in the regulation of PDGF-BB induced HASMC proliferation. Moreover, HIF1A-AS2 promotes HASMC proliferation, suggesting its potential involvement in VSMC proliferative vascular diseases.


INTRODUCTION
Vascular smooth muscle cells (VSMCs), the main cells that constitute blood vessels, play a critical role in maintaining their normal physiological function (1). Abnormal proliferation of VSMCs is a common feature of many vascular remodeling diseases, including atherosclerosis (2), hypertension (3), and vascular aneurysms (4). Thus, regulation of VSMC proliferation has major implications for the prevention of pathological vascular conditions (5).
Vascular injury induces the release of platelet-derived growth factor (PDGF) by activated inflammatory cells, platelets, and VSMCs, resulting in a switch from VSMC differentiated phenotype to a proliferative phenotype (14)(15)(16). However, lncRNAs involved in human aortic smooth muscle cell (HASMC) proliferation activated by the PDGF-BB isoform remain obscure.
In this study, using microarrays, we analyzed differentially expressed lncRNAs (DELs) and differentially expressed coding genes (DEGs) in proliferative HASMCs induced by PDGF-BB and quiescent HASMCs induced by fetal bovine serum (FBS) starvation. Our data showed that many lncRNAs were enriched in Gene Ontology (GO) terms and pathways related to cell proliferation. In addition, knockdown of HIF1A antisense RNA 2 (HIF1A-AS2) inhibited HASMC proliferation, at least in part, via the miR-30e-5p/CCND2 mRNA axis. Our study provides vital clues for elucidating the lncRNAs exert in VSMC abnormal proliferation, as it relates to VSMC proliferative vascular diseases.

RNA Extraction and Hybridization
Total RNA was isolated from proliferative and quiescent HASMCs using TRIzol reagent (Life Technologies, Carlsbad, USA

GO and Pathway Analyses
GO analysis was a functional analysis associating DEGs with GO categories. GO categories were derived from Gene Ontology (www.geneontology.org) and divided into biological processes (BP), cellular components (CC), and molecular functions (MF). Pathway analysis is an effective method to uncover the underlying biological functions in response to DEGs. Based on the latest Kyoto Encyclopedia of Genes and Genomes (http://www. genome.jp/kegg) database, we performed pathway analysis for DEGs. p < 0.05 was the threshold for statistical significance.

Construction of DELs-DEGs Co-expression Networks and DELs-miRNAs Interaction Networks
Pearson correlation coefficient (PCC) was calculated; R-value was utilized to calculate the PCC correlation coefficient between six DELs and DEGs from microarray data. Based on PCC (using the selection parameter PCC ≥ 0.90 as meaningful), the coexpression networks were constructed using Cytoscape_v3.7.1.

Cell Viability Assay
Cell viability was measured using CCK-8 assay kit (Wanleibio). In brief, HASMCs transfected with control or HIF1A-AS2 siRNA were seeded into 96-well plates and cultured in a medium containing 10 ng/mL PDGF-BB for 48 h. HASMCs were incubated with 10% CCK-8 solution 2 h in a humidified incubator containing 95% air and 5% CO 2 at 37 • C, and absorbance was measured at 450 nm. The experiment was repeated three times.

EdU Incorporation Assay
Incorporation assay of 5-Ethynyl-2 ′ -deoxyuridine (EdU) was performed using BeyoClick TM EdU Cell Proliferation Kit with Alexa Fluor 594 (Beyotime, Shanghai, China). HASMCs were seeded in 24-well plates at a density of 2 × 10 5 cells/well, and each sample contained 6 duplicate wells. SiRNA-Control or siRNA-HIF1A-AS2 were transfected into HASMCs and cultured with or without 10 ng/mL PDGF-BB. After 48 h of siRNA treatment, each well was incubated with SMCM containing 10 µM EdU for 2 h. HASMCs were fixed by 100 µL 4% paraformaldehyde for 15 min. After washing with PBS, HASMCs were incubated with 100 µL PBS containing 0.3% TritonX-100 for 15 min. After washing with PBS, HASMCs were stained with 100 µL Click Additive Solution for 30 min. After washing with PBS, nuclear staining was performed using 1 × Hoechst 33,342 incubation for 30 min. After washing with PBS, the positive cells were observed by fluorescence microscope (OLYMPUS IX71). The experiment was repeated three times.

Nuclear and Cytoplasmic RNA Detection
Nuclear/cytoplasmic components of proliferative HASMCs induced by 10 ng/mL PDGF-BB were isolated utilizing the Nuclear/Cytosol Fractionation Kit (BioVision, California, USA) according to the manufacturer's instructions. Extraction, quantification, and integrity detection of RNAs were the same as above. Reaction condition of qRT-PCR was in line with the above. U6 and GAPDH acted as positive controls for the cytoplasm and nucleus, respectively. Primers were exhibited in Supplementary Table 1. The experiment was repeated three times.

Dual-Luciferase Reporter Gene Assay
RNAhybrid was performed to analyze the potential binding sites between HIF1A-AS2 or CCND2 and miR-30e-5p. Using GPtransfect-Mate (GenePharma), luciferase reporter gene vectors (psi-CHECK TM -2 Vector, Promega, USA) containing wild-type (WT) or mutant (MUT) HIF1A-AS2 and the 3 ′ -UTR of WT or MUT CCND2 were co-transfected into 293A cells with miR-30e-5p mimics or miR-Control (GenePharma). After 24 h, Dual-Luciferase R Reporter Assay System (Promega) was performed based on the manufacturer's instructions. For each analysis, the Renilla luciferase signal was normalized to the firefly luciferase signal. The experiment was repeated three times.

Statistical Analysis
Data presented as bar graphs are the means ± SEM or SD of at least three independent experiments. Statistical analysis between two groups was performed using the Student's t-test. Differences between multiple groups were assessed by one-way ANOVA with Tukey's multiple comparisons test. Value of p < 0.05 was considered statistically significance.

Identification of DEGs Between Proliferative and Quiescent HASMCs
A total of 26,109 coding genes were analyzed by microarrays in HASMCs; from these, 3,720 DEGs in proliferative and quiescent HASMCs were identified (fold  change ≥2, p < 0.05). DEGs (Supplementary Table 3) included 3,019 up-regulated and 701 down-regulated DEGs (Figure 3).

GO and Pathway Analyses of DEGs Neighboring DELs
GO analysis showed that 251 up-regulated neighboring DEGs and 260 down-regulated neighboring DEGs were enriched in BP (Figures 5A,B), 64 up-regulated neighboring DEGs and 63 down-regulated neighboring DEGs were enriched in CC (Figures 5C,D), and 27 up-regulated neighboring DEGs and 48 down-regulated neighboring DEGs were enriched in MF (Figures 5E,F). Cell proliferation-related GO terms, such as RNA-induced silencing complex (RISC) associated-CC ( Figure 5D), PDGF receptor binding associated-MF (Figure 5E), and TGF-beta receptor binding associated-MF ( Figure 5F) were significantly enriched. Pathway analysis (Figures 5G,H) revealed some cell proliferation-related pathways, such as the regulation of actin cytoskeleton.

DISCUSSION
In this study, we investigated the function of lncRNAs in HASMC proliferation stimulated by the growth factor PDGF-BB.
LncRNAs can cis regulate the levels of their neighboring genes (40,41). Our GO and pathway analyses showed that DEGs neighboring DELs were enriched in the RNA-induced silencing complex (RISC; Figure 5D) that is required for miRNAs binding FIGURE 9 | HIF1A-AS2 restraint inhibits HASMC proliferation via the miR-30e-5p/CCND2 axis. (A) Venn diagram revealed unique one shared DEG between potential target genes of miR-30e-5p/miR-25-3p/miR-200c-3p/miR-490-3p/miR-34b-5p and DEGs co-expressed with HIF1A-AS2. (B,C) Western blot analysis of CCND2 in HASMCs transfected with control or HIF1A-AS2 siRNA (B) and transfected with control or miR-30e-5p mimics (C) under 10 ng/mL PDGF-BB treatment; β-tubulin and GAPDH were utilized as a control, respectively. Data are shown as mean ± SD of three independent experiments, *p < 0.05 vs. siRNA-Control or miRNA-Control group. (D) HIF1A-AS2 subcellular localization prediction utilizing the LncLocator database. (E) HIF1A-AS2 was abundant in the cytoplasm of HASMCs. U6 and GAPDH acted as positive controls in the nucleus and cytoplasm, respectively. Data are shown as mean ± SD of three independent experiments. (F) Binding sites of HIF1A-AS2 and miR-30e-5p were predicted by the RNAhybrid database. (G) Confirmation of miR-30e-5p as a sponged target of HIF1A-AS2 using the dual-luciferase reporter gene assay. Data are presented as mean ± SD of three independent experiments, ***p < 0.001. (H) Sequence of the 3 ′ -UTR of CCND2 mRNA that matches the miR-30e-5p seed sequence was predicted by the RNAhybrid database. (I) Verification of CCND2 as a target gene of miR-30e-5p utilizing the dual-luciferase reporter gene assay. WT: wild-type; MUT: mutant. Data are shown as mean ± SD of three independent experiments, ***p < 0.001.
to target genes. LncRNAs have been identified to exert as ceRNAs to sponge miRNAs and regulate RISC, thus regulating VSMC proliferation, differentiation, and apoptosis (42)(43)(44). Hence, DELs might target VSMC proliferation-related genes by sponging miRNAs, regulating RISC, and cis regulating neighboring DEGs. In addition, DEGs neighboring DELs were enriched in PDGF receptor binding associated-MF ( Figure 5E) and TGFbeta receptor binding associated-MF ( Figure 5F). Both PDGF (45,46) and TGF-beta (47) signaling pathways function in VSMC proliferation. Another pathway enriched was the regulation of actin cytoskeleton (Figure 5H). Actin cytoskeleton remodeling is necessary for VSMC phenotypic switch (48,49). Our group previously demonstrates that smooth muscle 22 alpha (SM22α), an actin-binding protein, participates in the organization of actin cytoskeleton in differentiated VSMCs by inducing F-actin bundling, thereby increasing VSMC contractility and mobility, and ultimately maintaining the differentiated phenotype of VSMCs (48). Over-expression of SM22α inhibits VSMC proliferation and neointima hyperplasia via reducing the response to mitogen stimuli (49). Thus, DELs might regulate VSMC proliferation by targeting the cytoskeleton-associated proteins, and by cis regulating their neighboring DEGs.
To our knowledge, this is the first report demonstrating that lncRNAs and mRNAs are differentially expressed in proliferative HASMCs induced by PDGF-BB. Furthermore, our results indicate that lncRNA HIF1A-AS2 promotes HASMC proliferation by the miR-30e-5p/CCND2 mRNA axis in some degree and highlight that HIF1A-AS2 might serve as a new therapeutic target for VSMC proliferative vascular diseases.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.