Approval of all the animal experimentation was gotten from the Animal Care and Use Committee at the Sixth Affiliated Hospital of Guangzhou Medical University (ethics number: LAEC-201910-002). This study was conducted only with males, as it is well-known that atherosclerotic plaques develop more reproducibly and with less biological variability. Male apolipoprotein E deficient (apoE)^−/− mice (weight, 20–25 g; age, 4 weeks) on a C57BL/6J genetic background were obtained from Nanjing Junke Biological Engineering Co. Ltd. (Nanjing, China). A total of 24 apoE^−/− mice and 6 wild-type C57BL/6 mice (weight: 20–25 g; age: 4 weeks) were used. Following 1 week of adaptive feeding, the initiation of the experimental modeling was done. The mice were housed under a regular cycle involving 12-h of light and 12 h of darkness. The humidity was maintained at 60% and the temperature was controlled at 22°C. The mice were allowed unrestricted access to food and drinking water.

According to the transfection scheme, 24 apoE^−/− mice were randomly allocated to four cohorts, namely, adeno-associated virus (AAV, JiKai Gene, Shanghai, China)-sh-NC, AAV-sh-H19, AAV-OE-NC, and AAV-OE-H19 groups (n = 6). Male apoE^−/− mice were treated with either the AAV-sh-NC, AAV-sh-H19, AAV-OE-NC, or AAV-OE-H19 via tail vein at a titer of 1.0 × 10¹¹ vector genomes/ml. All cohorts were treated with tail vein injection of the corresponding AAV; all AAV suspensions were of the same titer and dosage. Male apoE^−/− mice were supplied with a high-fat diet containing 1.25% cholesterol (HFD, D12108C, Research Diets, New Brunswick, NJ, USA) for 1 week and subsequently treated with AAV scrambled small hairpin RNA (AAV-sh-NC), AAV-sh-H19, AAV-OE-NC, and AAV-OE-H19; the mice were continually supplied with HFD for 8 weeks. During the construction of an atherosclerosis model, AAV was injected at 5, 8, and 12 weeks. After 8 weeks rearing, male apoE^−/− mice were fasted for 6 h, then placed under isoflurane anesthesia (RWD, Shenzhen, China). Anesthetic used was 1% isoflurane, with the dosage depended on the animal's body weight (40 mg/kg). Blood samples were obtained from the retro-orbital venous plexus using a capillary tube. The mice were then sacrificed by cervical dislocation and organic tissues collected for the following experimental analyses. These analyses were performed blinded to which groups the animals were allocated.

After euthanizing the apoE^−/− mice, the entire aorta tissues were obtained, followed by the removal of the adventitial tissues. Subsequently, the aorta was opened and oil red O (Sigma, St. Louis, MO, USA) was utilized for staining. Next, the aortal was photographed using a stereomicroscope (Zeiss, Jena, Germany). The proportion of lesion area that stained after application of oil red O on the surface of the aorta was estimated utilizing the Image-Pro Plus 7.0 software (Media Cybernetics, Rockville, MD, USA).

The mice were sacrificed through exsanguination under anesthesia, and subsequently, the upper part of their hearts was separated from the proximal aorta. After washing using phosphate-buffered saline (PBS), the hearts were entrenched in an optimal cutting temperature (OCT) compound (Sakura, Tokyo, Japan) and frozen. A cryostat microtome was used to prepare serial 6-μm-thick cryosections of the 3 aortic valves, and these sections were kept on glass slides. Then, oil red O, hematoxylin–eosin (HE) and Masson staining were subsequently carried out (26). The Image J software (NIH Image J system, Bethesda, MD) was utilized for quantitative analysis.

THP-1 cells were procured from the American Type Culture Collection (ATCC: TIB 202, USA). Culturing of the THP-1 cells was done in the Roswell Park Memorial Institute (RPMI) 1640 medium (Thermo Fisher Scientific, USA) comprising 10% fetal bovine serum (FBS) in a CO₂ concentration of 5% and a temperature of 37°C. Subsequently, differentiation of THP-1 cells into macrophages was done after treating them with 80 nM of phorbol-12-myristate-13-acetate (PMA; MCE, China) for 48 h. The lentivirus used was purchased from Shanghai Jikai Gene Chemical Technology Co., Ltd. (Shanghai, China). THP-1 macrophages were transduced using the lentivirus supernatants (LV-NC, LV-H19, LV-sh-NC, or LV-sh-H19) with 8 μg/ml of polybrene (Sigma–Aldrich, St Louis, MO, USA). After 24 h of incubation, THP-1 macrophages were maintained in a complete medium (10% FBS and RPMI 1640). Real-time reverse transcription polymerase chain reaction (qRT-PCR) was performed to confirm the transfection efficiency.

The miR-146a-5p mimic, miR-146a-5p inhibitor as well as mimic/inhibitor negative control (mimic/inhibitor NC) were procured from GenePharma (Suzhou, China). Transfection of THP-1 cells was done with miR-146a-5p mimic/inhibitor or their negative controls using the lipofectamine 3000 reagent (Invitrogen, CA, USA) for 48 h as per the instructions stipulated in the kit. Subsequently, the transfection efficiency was measured using qRT-PCR. Both small interfering RNA (siRNA) targeting ANGPTL4 (si-ANGPTL4) and non-silencing siRNA were designed and synthesized by GenePharma. Transfection of both siRNAs (80 nM) into THP-1 cells was done utilizing the lipofectamine 3000 reagent for a duration of 24 h. qRT-PCR, as well as western blot (WB), were carried out to assess the efficacy of ANGPTL4 knockdown.

The online datasets such as starBase 2.0 and BiBiServ2-RNAhybrid (https://bibiserv.cebitec.uni-bielefeld.de/rnahybrid/) were utilized to forecast the interplay between miR-146a-5p and ANGPTL4 or lncRNA H19. The HEK-293T cells were added to a 24-well plate at a concentration of 2.0 × 10⁴/well. The sequences of lncRNA H19 and ANGPTL4 non-coding sequence [3′ untranslated region (3′-UTR)] comprising the putative miR-146a-5p binding site were amplified using PCR and subsequently cloned downstream of the pGL3 firefly luciferase reporter plasmids (Promega, WI, USA) labeled as lncRNA H19-WT and ANGPTL4-WT. Mutation of the putative miR-146a-5p binding sites was done to produce the matching mutant plasmids (lncRNA H19-Mut and ANGPTL4-Mut). Then, co-transfection of HEK-293T cells with the plasmids and miR-146a-5p mimic was done utilizing the lipofectamine 3000 reagent for a duration of 48 h. Subsequently, the detection of the luciferase activity was done utilizing the dual-luciferase reporter assay system (Promega) followed by the normalization of the firefly luciferase activity to the Renilla luciferase activity.

RNA fluorescence in situ hybridization (RNA-FISH) for lncRNA H19 was performed on mouse aortic root tissue and THP-1 cells. Cyanine 3 (Cy3)-labeled lncRNA H19 (human lncRNA 19: 5′-GCTGCTGTTCCGATGGTGTCTTTGATGTTGGGC-3′; mice lncRNA H19: 5′-CAGTTGCCCTCAGACGGAGATGGACG-3′) and 4′,6-diamidino-2-phenylindole (DAPI)-labeled U6 probes were gotten from Servicebio (Guangzhou, China). THP-1 cells were seeded on poly-L-lysine coated slides (Sigma-Aldrich) and fixed in 4% paraformaldehyde for 10 min, washed twice in cold PBS and permeabilized with 70% EtOH. Next, the RNA FISH assays were carried out utilizing a FISH kit (Servicebio, Guangzhou, China) as per the instructions stipulated by the manufacturer.

The isolation of the total RNA from the collected blood, tissues as well as cultured THP-1 cells was done utilizing the Blood RNA Isolation Kit (Thermo Fisher Scientific, CA, USA) and TRIzol reagent (Invitrogen) (27). Extracted RNA was quantified for purity and concentration using a NanoDrop 3000 spectrophotometer (Thermo Fisher Scientific). Subsequently, 2 μg of RNA was employed as a template to synthesize complementary DNA (cDNA; Takara, Kyoto, Japan), and qRT-PCR was carried out utilizing the SYBR Green kit (Takara) on a Bio-Rad CFX96 instrument (Bio-Rad, USA) for 32 cycles (95°C for 3 min, 95°C for 15 s and 60°C for 1 min). The U6 was employed as an internal control for glyceraldehyde 3-phosphate dehydrogenase (GAPDH), miR-146a-5p, and other genes. The primers were synthesized by Tianyi Huiyuan Biotech Co., Ltd. (Wuhan, China); they are listed in Table 1. The relative gene expression was analyzed utilizing the 2^−ΔΔCT method.

Lysing of the THP-1 cells and tissues was done utilizing the radioimmunoprecipitation assay (RIPA) buffer (Dingguo, Beijing, China) containing 1 mM of dithiothreitol (DTT) and 1 mM of protease inhibitor [phenylmethylsulfonyl fluoride (PMSF)]. The quantification of the protein extracts was done utilizing a bicinchoninic acid (BCA) assay kit (Beyotime, Shanghai, China) as per the instructions stipulated by the manufacturer. Subsequently the protein was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). This was followed by immunoblotting using the following antibodies: ANGPTL4 (abs136271, Absin, China) and GAPDH (D16H11, Cell Signaling Technology, USA). Washing of the membranes was done utilizing TBST (tris buffered saline and 0.1% Tween 20) followed by incubation using horseradish peroxidase (HRP)-labeled secondary antibodies (1:3,000, Cell Signaling Technology, USA). Visualization of the proteins was done utilizing the BeyoECL Plus kit (Beyotime), and GAPDH was employed as an internal control.

Fixing of the THP-1 macrophages was done using 4% paraformaldehyde for a duration of 10 min followed by a 3-times wash with PBS. Subsequently, the oil red O solution was utilized to stain the THP-1 cells for a duration of 20 min and afterward destained using 60% isopropanol for 10 s. A microscope (Olympus BX50) was utilized to capture the pictures of positively stained cells (red) at a magnification of 200×.

Statistical analyses were articulated as mean ± standard deviation (SD) of a minimum of 3 separate experiments. Comparison between the control cohort and the other cohorts was done utilizing the one-way or two-way analysis of variance (ANOVA). Differences between the 2 cohorts were contrasted utilizing the two-tailed Student's t-test. Statistical analysis was executed utilizing the GraphPad Prism 8.0 (GraphPad Prism Software, San Diego, CA, USA) and SPSS 19.0 statistical software (SPSS, Chicago, IL, USA). Statistical significance was delineated as ^*P < 0.05, ^**P < 0.01, and ^***P < 0.001.

Contemporary research reports have demonstrated that aberrant expression and functions of lncRNAs participate in the occurrence as well as the development of various illnesses (28–30), especially atherosclerosis. Recently, lncRNA H19 has been reported as a well-recognized lncRNA linked to the development of atherosclerosis and could be a probable treatment target for atherosclerosis (6, 16). To determine whether lncRNA H19 expression was changed during atherosclerosis, we initially analyzed its expression in the whole blood and aortic tissues of apoE^−/− mice using qRT-PCR and RNA-FISH. As illustrated in Figures 1A–C, the expression of lncRNA H19 expression was elevated in apoE^−/− mice supplied with a Western diet as opposed to that in the wild-type C57BL/6 mice. The levels of lncRNA H19 were considerably elevated in ox-LDL-treated THP-1 cells as opposed to the control cells (Figures 1D,E). These findings suggested that lncRNA H19 may be an important cause of atherosclerosis.

Atherosclerotic plaque is characterized by the accumulation of lipid as well as the formation of lipid-laden foam cells (9). It is well-known that lncRNAs facilitate lipid accumulation in multiple cells (31, 32). To determine the effects of lncRNA H19 on ox-LDL-treated THP-1 cells, lncRNA H19 was knocked down or overexpressed in THP-1 cells infected with lentivirus-mediated lncRNA H19 knockdown or overexpression. The results indicated that the knockdown of lncRNA H19 significantly decreased lncRNA H19 levels; however, the overexpression of lncRNA H19 markedly increased the levels (Figure 2A). To examine the effects of lncRNA H19 on the accumulation of lipid in ox-LDL-treated THP-1 cells, lncRNA H19 was knocked down or overexpressed in THP-1 cells. It was discovered that ox-LDL stimulation-induced foam cell formation (33), and the knockdown of lncRNA H19 significantly inhibited foam cell formation and lipid accumulation, whereas the overexpression of lncRNA H19 promoted lipid accumulation in macrophage foam cells (Figures 2B,C). These observations suggested that lncRNA H19 promoted the formation of foam cells in ox-LDL-treated THP-1 cells, implying that it aggravated the progression of atherosclerosis.

To evaluate the role and treatment prospective of lncRNA H19 in vivo, male apoE^−/− mice (weight, 20–25 g; age, 4 weeks) were supplied with a high-fat diet (HFD) for 1 week and subsequently treated with AAV scrambled small hairpin RNA (AAV-sh-NC), AAV-sh-H19, AAV-OE-NC, and AAV-OE-H19; the mice were continually supplied with HFD for 8 weeks (Figure 3A). Foremost, we ascertained the efficiency of lncRNA H19 suppression and overexpression by assessing lncRNA H19 expression in blood and the aorta after treating the mice for 8 weeks. The levels of lncRNA H19 expression identified utilizing qRT-PCR were considerably reduced in the AAV-sh-H19 cohort and elevated in the AAV-OE-H19 cohort as opposed to the matching control cohort (AAV-sh-NC or AAV-OE-NC) (Figures 3B,C). After 8 weeks of feeding HFD to apoE^−/− mice, we found a marked reduction in the number as well as the size of atherosclerotic lesions in the aortic arch sections (Figure 3D) in the AAV-sh-H19 cohort as compared with the AAV-sh-NC cohort; however, lncRNA H19 overexpression in the AAV-OE-H19 cohort significantly increased plaque formation in the aortic arch regions (Figure 3D) compared with the AAV-OE-NC cohort. Oil red O, and HE staining of the aortic root cross-sections illustrated that the overexpression of lncRNA H19 also considerably increased the lesion area, enhanced lipid deposition, and lowered collagen content (Figure 3E). The knockdown of lncRNA H19 decreased the lesion area, suppressed deposition of lipids, and elevated collagen content as demonstrated by Masson, oil red O, and HE staining (Figure 3E). Moreover, lncRNA H19 overexpression increased the en face lesions in the aorta (Figure 3F), which did not occur when lncRNA H19 was knocked down (Figure 3F). These in vivo findings collectively illustrate that lncRNA H19 serves as a vital participant in the progression of atherosclerosis.

Cumulative evidence from research reports has demonstrated that lncRNAs can act as miRNA sponges, which interfere with miRNA at the post-transcriptional level thus decreasing the binding efficiency of miRNAs to target genes (34). We used an online bioinformatic dataset (BiBiServ2-RNAhybrid) (35) for identifying the potential miRNA recognition elements of lncRNA H19. We discovered that the lncRNA H19 sequence comprised one putative miR-146a-5p site in humans (Figure 4A). The sequences of lncRNA H19 together with the miR-146a-5p binding site and mutant lncRNA H19 were embedded downstream of the luciferase gene in the reporter plasmid to be used in luciferase assays. The results from the luciferase assay illustrated that transfection with miR-146a-5p mimic suppressed lncRNA H19 expression in the wild-type cohort as opposed to the control cohort, and this suppression was inverted after mutation of the binding site (Figure 4B), indicating the equivalent sequence-specific binding of miR-146a-5p to lncRNA H19. In addition, treating THP-1 cells with LV-lncRNA H19 (OE-H19) decreased miR-146a-5p levels (Figure 4C). However, the knockdown of lncRNA H19 (sh-H19) increased miR-146a-5p levels (Figure 4C). Therefore, lncRNA H19 functioned as a contending endogenous RNA (ceRNA) for miR-146a-5p.

As illustrated in Figures 4D,E, the miR-146a-5p expression was lower in the blood and aortic tissues of apoE^−/− mice supplied with Western diet as opposed to that in the wild-type mice. Furthermore, the miR-146a-5p levels were considerably decreased in ox-LDL-treated THP-1 cells as opposed to the control cells (Figure 4F). Transfection with miR-146a-5p inhibitor promoted the accumulation of lipids in ox-LDL-treated THP-1 cells (Figure 4G). These outcomes indicated that reduced miR-146a-5p expression aggravated atherosclerosis in apoE^−/− mice and also facilitated the accumulation of lipids in ox-LDL-treated THP-1 cells.

Many studies have reported that ANGPTL4 is modulated by sponging miRNAs at the transcriptional level (36, 37). To identify prospective mRNAs that could be targeted by miR-146a-5p, we used a computational programming database (BiBiServ2-RNAhybrid) and found a putative binding site between ANGPTL4 3′-UTR and miR-146a-5p (Figure 5A). We subsequently built a luciferase reporter plasmid comprising a mutant (ANGPTL4-Mut) or wild-type (ANGPTL4-WT) miR-146a-5p binding site (Figure 5A). Transfection of these plasmids into HEK-293T cells together with miR-146a-5p mimic or mimic control was done to perform the luciferase reporter assay. Subsequently, the co-transfection of ANGPTL4-WT and miR-146a-5p mimic contributed to a considerable decline in the luciferase activity; however, this influence disappeared after the mutation of the miR-146a-5p binding site (Figure 5B). To ascertain whether miR-146a-5p might straightforwardly modulate the expression of ANGPTL4, THP-1 macrophages were transfected with miR-146a-5p mimic or inhibitor and the results indicated that miR-146a-5p mimic considerably reduced miR-146a-5p levels, whereas its inhibitor increased the levels, thus exhibiting an increased transfection efficiency (Figure 5C). Transfection with miR-146a-5p inhibitor was found to upregulate the protein and mRNA expression of ANGPTL4 (Figures 5D,E). These results revealed ANGPTL4 as a candid target of miR-146a-5p.

The analysis of publicly published datasets (GSE77104 and GSE83090) illustrated that ANGPTL4 is the highest induced gene in bone marrow-derived macrophages (BMDMs) and mouse peritoneal macrophages after treatment with oleic acid (OA) and acetylated (AC)-LDL (Figure 6A) (38–40). To scrutinize the function of endogenous ANGPTL4 in THP-1 cells, we first analyzed the regulation of ANGPTL4 expression. Treating THP-1 cells with ox-LDL markedly increased ANGPTL4 mRNA level (Figure 6B). ANGPTL4 expression was higher in the blood and aortic tissues of apoE^−/− mice fed a Western diet than that in the wild-type mice (Figures 6C,D). To examine the functional role of ANGPTL4 in lipid accumulation, we performed a loss-of-function study on THP-1 cells using ANGPTL4 siRNA. We found that the levels of expression of ANGPTL4 mRNA and protein were markedly repressed in THP-1 cells transfected with ANGPTL4 siRNA as opposed to the control cells (Figures 6E,F). Oil red O staining illustrated that ANGPTL4 knockdown reduced lipid accumulation in ox-LDL-treated THP-1 cells (Figure 6G). These results indicated that ANGPTL4 enhanced the accumulation of lipids in ox-LDL-treated THP-1 cells.

Owing to the presence of a ceRNA network among lncRNA H19, ANGPTL4, and miR-146a-5p, we investigated whether miR-146a-5p is necessary for the regulatory impacts of lncRNA H19 on ANGPTL4 expression. The knockdown of lncRNA H19 decreased protein and ANGPTL4 mRNA levels; however, the overexpression of lncRNA H19 elevated the protein and mRNA levels of ANGPTL4 (Figures 7A,B). Co-transfection of THP-1 cells with miR-146a-5p inhibitor and sh-H19 was performed. As illustrated in Figures 7C,D, transfection with miR-146a-5p inhibitor considerably abrogated the effects of lncRNA H19 knockdown on ANGPTL4 expression. However, transfection with miR-146a-5p inhibitor enhanced the impact of lncRNA H19 overexpression on ANGPTL4 expression (Figures 7E,F). Therefore, we concluded that lncRNA H19 facilitated ANGPTL4 expression by serving as a ceRNA for miR-146a-5p.