CircNAP1L4 regulates pulmonary artery smooth muscle cell proliferation via the NAP1L4‐mediated super‐enhancer‐driven glycolysis gene hexokinase II (HK II) in pulmonary hypertension

Abstract

Glycolysis is a major determinant of pulmonary artery smooth muscle cell (PASMC) proliferation in pulmonary hypertension (PH). Circular RNAs (circRNAs) are powerful regulators of glycolysis in multiple diseases; however, the role of circRNAs in glycolysis in PH has been poorly characterized. The aim of this study was to uncover the regulatory mechanism of a new circRNA, circNAP1L4, in human pulmonary artery smooth muscle cell (HPASMC) proliferation through the host protein NAP1L4 to regulate the super‐enhancer‐driven glycolysis gene hexokinase II (HK II). CircNAP1L4 was downregulated in hypoxic HPASMCs and plasma of PH patients. Functionally, circNAP1L4 overexpression inhibited glycolysis and proliferation in hypoxic HPASMCs. Mechanistically, circNAP1L4 directly bound to its host protein NAP1L4 and affected the ability of NAP1L4 to move into the nucleus to regulate the epigenomic signals of the super‐enhancer of HK II. Intriguingly, circNAP1L4 overexpression inhibited the proliferation but not the migration of human pulmonary arterial endothelial cells (HPAECs) cocultured with HPASMCs. Furthermore, pre‐mRNA‐processing‐splicing Factor 8 (PRP8) was found to regulate the production ratio of circNAP1L4 and linear NAP1L4. In vivo, targeting circNAP1L4 alleviates SU5416 combined with hypoxia (SuHx)‐induced PH. Overall, these findings reveal a new circRNA that inhibits PASMC proliferation and serves as a therapeutic target for PH.

Splicing factor PRP8 regulated the expression of circNAP1L4 and linear NAP1L4 in hypoxic PASMCs. The binding of the circNAP1L4 and NAP1L4 protein is reduced in the cytoplasm, leading to entry of the NAP1L4 protein into the nucleus. Then, NAP1L4 increases the SE modification of HK II to promote the transcription of HK II, leads to PASMC glycolysis and proliferation. In addition, the small amount of circNAP1L4 released by PASMCs does not protect PAECs from proliferation.Finally, increased cell proliferation triggers PH.

Abbreviations

AAV5

serotype 5 adenovirus‐associated virus

CEBPB

CCAAT enhancer binding protein beta

ChIP

chromatin immunoprecipitation assay

CircRNAs

circular RNAs

ECAR

extracellular acidification rate

FISH

fluorescent in situ hybridization

HK II

hexokinase II

HPAECs

human pulmonary arterial endothelial cells

HPASMC

human pulmonary artery smooth muscle cell

NC

negative control

PAEC

pulmonary artery endothelial cell

PASMC

pulmonary artery smooth muscle cell

PFKFB3

endothelial 6‐phosphofructo‐2‐kinase

PH

pulmonary hypertension

PKM2

pyruvate kinase 2

PRP8

pre‐mRNA‐processing‐splicing factor 8

SE

super‐enhancer

SuHx

SU5416 combined with hypoxia

α‐SMA

α‐smooth muscle actin

1. INTRODUCTION

Pulmonary hypertension (PH) is a malignant pulmonary vasculature disease characterized by pulmonary vasoconstriction and pulmonary vascular remodeling, eventually leading to right ventricular hypertrophy. ¹ , ² Pulmonary artery smooth muscle cell (PASMC) hyperproliferation caused by hypoxia is the pathological basis of pulmonary vascular remodeling. ³ At present, due to limited clinical treatment strategies, the survival and prognosis of PH patients are limited. ⁴ This highlights the need to further elucidate the underlying mechanisms of excessive PASMC proliferation and develop novel effective therapies to treat patients with PH.

Glycolysis is a process by which glucose converted into lactic acid. Accumulated evidence has unveiled the significance of glycolysis in PH progression. Multiple signaling molecules and signal transduction pathways are involved in the regulation of glycolysis. For instance, endothelial 6‐phosphofructo‐2‐kinase (PFKFB3)‐mediated glycolysis promotes the development and progression of PH. ⁵ MiR‐125a‐5p can inhibit glycolysis and PASMC proliferation to alleviate PH by targeting hexokinase‐II (HK‐II). ⁶ In view of these regulatory networks, further insight into other regulators of glycolysis in PH is needed.

Circular RNAs (circRNAs), a distinct class of noncoding RNAs formed by back splicing of pre‐mRNAs, have recently been identified as important regulators at the posttranscriptional level involved in the pathological process of many diseases. By interacting with microRNAs (miRNAs), DNA, and proteins, circRNAs can exert regulatory functions. Notably, unlike other noncoding RNAs, circRNAs can play regulatory roles by influencing the expression of parental genes. For example, circRNA‐ENO1 promoted lung adenocarcinoma cell proliferation through upregulating its host gene ENO1. ⁷ The antisense circular RNA circSCRIB enhanced breast cancer cell proliferation by suppressing parental gene splicing and translation. ⁸ However, the regulatory mechanism through which circRNAs act on parental genes in PH development and progression has not been determined.

Super‐enhancers (SEs) are a large cluster of multiple enhancers with transcriptional activity, which can be highly enriched in transcription factors, cofactors, and histone modifiers, thus exerting powerful transcriptional effects. ⁹ , ¹⁰ Accumulating evidence shows that SEs are capable of driving the expression of genes that play prominent roles in regulating the development of tumors, cardiovascular disease, and diabetes. ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ In addition to these important findings, a recent study showed that the SE‐associated circular RNA circKrt4 regulates pulmonary artery endothelial cell (PAEC) dysfunction in PH. ¹⁶ Studying the role of SE‐associated genes in PASMC proliferation will be highly interesting and valuable.

Our present study revealed that a circRNA derived from NAP1L4 (circNAP1L4) is significantly decreased in hypoxia‐induced HPASMCs. Overexpression of circNAP1L4 attenuated chronic hypoxia and SuHx (SU5416 combined with hypoxia)—induced PH in mice by regulating the host protein NAP1L4‐mediated SE modification of HK II. Our data demonstrated a novel regulatory mechanism involved in pulmonary vascular remodeling, and circNAP1L4 may be a potential target involved in the progression of PH.

2. MATERIALS AND METHODS

2.1. Human plasma

A total of 40 (control = 15, PH = 25) plasma samples from healthy controls and patients with PH were obtained at the Second Affiliated Hospital of Harbin Medical University. Clinical information on the patients in the plasma sample cohort is shown in Table S2. Written informed consent was obtained from participants in this study. Whole blood samples (5 mL) were centrifuged at 1500 g for 15 min at 4°C, immediately collected in RNase/DNase‐free tubes, and stored at −80°C until further analysis.

2.2. Animal treatments and lung tissue preparation

Since female sex hormone can affect chronic hypoxia and SuHx‐induced PH, ¹⁷ , ¹⁸ adult male C57BL/6J mice with a mean weight of 20 to 25 g were used in this study. C57BL/6J mice were obtained from the Laboratory Animal Center of the Second Affiliated Hospital of Harbin Medical University. The sequences used for circNAP1L4 cloning and for the serotype 5 adenovirus‐associated virus (AAV 5) packaging experiment were packaged by Genechem (Shanghai, China). An aliquot of the vector at 10¹¹ genome equivalents was prepared in 20 μL of HBSS and isoflurane anesthesia followed by nasal drops. After 7 to 14 days, the mice were randomly divided into normoxic (Fi, O₂ 0.21) and hypoxic (Fi, O₂ 0.10) groups. SU5416 (S8442; Sigma Aldrich, Darmstadt, Germany) was injected subcutaneously at a dose of 20 mg/kg, and control mice were injected with the same volume of vehicle alone. Animals injected with Su5416 were exposed to normobaric chronic hypoxia (10% O₂) for the next 3 weeks, followed by preexposure to normoxia for 2 weeks. All mice were anesthetized through an intraperitoneal injection of avertin (200 mg/kg i.p., Sigma Aldrich, St. Louis, USA), after which the right ventricular (RV) hypertrophy index (the ratio of the RV free wall weight to the sum of the septum plus left ventricular (LV) free wall weight) and the RV/(LV+S) ratio were calculated.

2.3. Echocardiography and right ventricular systolic pressure (RVSP) measurements

The right ventricular systolic pressure (RVSP) and echocardiography results were measured as previously described. The right ventricular systolic pressure (RVSP) was measured with PowerLab monitoring equipment (AD Instruments, Colorado Springs, CO). A 1.2 French pressure catheter (Scisense Inc, USA) was inserted into the superior vena cava and finally into the right ventricular vein, and the RVSP was continuously recorded for a period of at least 5 min. Mice were subjected to echocardiography using a Vevo2100 imaging system (VisualSonics, Toronto, Ontario, Canada). The pulmonary artery velocity time integral (PAVTI), pulmonary artery acceleration time (PAAT), and left ventricular ejection fraction (LVEF) were obtained from stable images. All the measurements were performed by the same experienced investigator. Off‐line data analysis was performed using a LabChart 8 Reader and Vevo2100 software by an investigator in a blinded manner.

2.4. Morphometric analysis

In brief, mouse lung tissues were immersed in 4% paraformaldehyde for 48 h. After dehydration, clearing, and embedding in paraffin wax, the blocks were cut into 5‐μm‐thick sections and stained as appropriate. Hematoxylin and eosin (HE) staining and Masson trichrome staining were performed according to the manufacturer's instructions. For HE staining quantification, approximately 6 arteries (>50 or <50 μm external diameter) per animal were randomly examined and the measurements were made at three points randomly for each vessel under microscope at magnification × 200, and an average was calculated. The medial thickness was quantified with high‐resolution images of individual vessels using a color‐recognition algorithm in Image‐Pro Plus 6.0 software. The percentage of medial thickness (%thickness/diameter) was determined from the thickness of the vessel wall relative to the outer diameter of pulmonary arteries. The medial thickness is calculated by using the following formula: medial thickness = external diameter − inner diameter; medial thickness % = medial thickness × 100/external diameter.

For Masson trichrome quantification, approximately 6 arteries (<150 μm external diameter) per animal were randomly examined under microscope at magnification × 200, and an average was calculated. The percentage of fibrosis area (%fibrosis area) was determined from the blue area of the perivascular relative to the total area using a color‐recognition algorithm in ImageJ software. All the analyses were performed by another person who was blinded to the experimental protocol.

2.5. Cell culture

HPASMCs and HPAECs were obtained from Procell Life Science & Technology (Wuhan, China) and cultured in medium (endothelial cell medium (ECM) for HPAECs, Sciencell, 1001, CA, USA; smooth muscle cell medium (SMC) for HPASMCs, Sciencell, 1101, CA, USA) containing 15% fetal bovine serum and 1% penicillin‐streptomycin at 37°C, 5% CO₂, and 100% relative humidity. For the hypoxia exposure experiments, HPASMCs were incubated in a Tri‐Gas Incubator (Heal Force) in a water‐saturated atmosphere containing 3% O₂ and 5% CO₂ for 24 h. All the experiments were performed with 6 batches of HPASMCs that were less than passage 5.

2.6. Fluorescent in situ hybridization (FISH)

Cy3‐labeled circNAP1L4 probes were synthesized by GeneChem (Shanghai, China). FISH experiments for HPASMCs and lung tissues samples were performed using a Fluorescent In Situ Kit (GeneChem Shanghai, China) according to the manufacturer's instructions. In brief, HPASMCs cultured on coverslips were fixed in 4% paraformaldehyde, and then permeabilized with 0.3% Triton X‐100. Cy3‐labeled circNAP1L4 probes were incubated with HPASMCs or 5‐μm‐thick lung tissue sections overnight at 37°C in the dark. 18S and U6 probes were used as internal controls. Finally, 4′,6‐diamidino‐2‐phenylindole (DAPI) was added to label the nuclei at 37°C for 10 min. Images were captured with a live cell workstation (AF6000; Leica, Wetzlar, Germany). The sequences of the circNAP1L4 probes used are listed in Table S1.

2.7. Western blot analysis

The proteins were extracted from HPASMCs, HPAECs, and lung tissues by using RIPA (P0013B; Beyotime Biotechnology, Shanghai, China) buffer supplemented with PMSF (ST506; Beyotime Biotechnology, Shanghai, China). The 30 μg protein samples were fractionated by 10% SDS–PAGE, transferred onto nitrocellulose membranes, and subsequently blocked with 5% nonfat milk at room temperature for 1 h. Anti‐PCNA antibody (1:1000, #2586, Cell Signaling Technology, MA, US), anti‐cyclin A antibody (1:1000, BM4674, Boster, Wuhan, China), anti‐cyclin D antibody (1:500, A11022, ABclonal, Wuhan, China), anti‐NAP1L4 antibody (1:1000, 16018‐1‐AP, Proteintech, IL, USA), anti‐HKII antibody (1:5000, 22029‐1‐AP, Proteintech, IL, USA), anti‐PKM2 antibody (1:500, A0268, ABclonal, Wuhan, China), anti‐PRP8 antibody (1:500, A6053, ABclonal, Wuhan, China), anti‐H3K27ac antibody (1:500, A7253, ABclonal, Wuhan, China), anti‐H3K4me1 antibody (1:500, A2355, ABclonal, Wuhan, China), and anti‐CEBPB antibody (1:500, GTX100675, GeneTex, Southern California, USA) were used as primary antibodies and were incubated overnight at 4°C, followed by incubation with appropriate horseradish peroxidase‐conjugated secondary antibodies at room temperature for 1 h and enhanced chemiluminescent reagents imaging.

2.8. Quantitative real‐time polymerase chain reaction (qRT–PCR)

Total RNA was extracted from HPASMCs, HPAECs, and lung tissues with TRIzol reagent (Invitrogen, CA, USA) according to the manufacturer's protocol. RNA was extracted from blood/plasma samples using a Blood/Plasma RNA Extraction Kit (HaiGene, Harbin, China) according to the manufacturer's protocol. Cytoplasmic and nuclear RNAs were isolated and purified using a Norgen Cytoplasmic and Nuclear RNA Purification Kit (Thorold, ON, Canada). cDNA was synthesized from 2 μg of RNA using the Superscript First‐Strand Complementary DNA Synthesis Kit (HaiGene, Harbin, China). To measure the RNA expression levels, the cDNA products were quantified using SYBR Green real‐time PCR (Toyobo, Japan) in a Roche LightCycler 480II instrument. β‐actin was used as the internal control. The threshold cycle (Ct) was determined, and the data were analyzed using the 2^−ΔΔCT method. The nucleotide sequences of the primers used are listed in Table S1.

2.9. CCK8 assay

HPASMCs were cultured in 96‐well plates at a density of 5000 cells/well and treated with different reagents. After incubating at 37°C for an additional 24, 48, and 72 h, ten microliters of CCK8 reagent were added to each well,which was subsequently incubated for 2–4 h at 37°C, after which the absorbance at 450 nm was measured using a spectrophotometric microplate reader.

2.10. Plasmid and siRNA construction and transfection

An overexpression vector for wild‐type and mutant circNAP1L4 was constructed using the vector GV486, and an empty vector alone was used as a negative control (Genechem, Shanghai, China). Small interfering RNAs (siRNAs) were designed and synthesized by GenePharma (Shanghai, China), and negative control siRNA (siNC) was used as an NC. The sequences are listed in Table S1. Transfections in HPASMCs were performed using Lipofectamine 2000 and X‐tremeGene siRNA following the manufacturer's instructions. HPASMCs were transfected with 3 μg of circNAP1L4 plasmid or 2 μg of siRNA. The transfection reagents were removed after 4–6 h, after which the medium was replaced with 5% serum‐containing medium, and the cells were cultured under normoxic or hypoxic growth conditions for another 24 h.

2.11. 5‐ethynyl‐2′‐deoxyuridine (EdU) incorporation assay

HPASMCs and HPAECs in different groups were treated with 50 μmol/L 5‐ethynyl‐2′‐deoxyuridine (C0071S; Beyotime, Shanghai, China) and incubated for 2 h at 37°C. Then the cells were fixed with 4% formaldehyde for 10 min and exposed to 0.3% Triton X‐100 for 10 min. DAPI (5 μg/mL) was used to stain the DNA of the cells in each well for 10 min, after which the cells were observed under a live cell workstation (AF6000; Leica, Wetzlar, Germany).

2.12. Cell cycle and DNA analysis

Cell cycle progression was measured by flow cytometry. Briefly, HPASMCs were treated as indicated and then harvested and fixed using 70% ethanol at 4°C. The cells were stained using a Cell Cycle and Apoptosis Analysis Kit (C1052; Beyotime, Shanghai, China) following the manufacturer's instructions. Finally, stained cells were detected by using a FACSCalibur flow cytometer (BD Biosciences, Bedford, MA, USA).

2.13. Glycolysis assay

A Seahorse XF24 Extracellular Flux Analyzer (Seahorse Biosciences, North Billerica, MA) was used to detect the extracellular acidification rate (ECAR, indicative of glycolysis). Experiments were performed according to the manufacturer's instructions. ECAR was measured using Seahorse XF Glycolysis Stress Test Kit (Agilent Technologies). Briefly, HPASMCs were cultured in XF24 plates at a density of 8000 cells/well and treated with different reagents. On the day of the measurement, the cells were incubated at 37°C for 1 h in the CO₂‐free incubator to balance the media pH and temperature, and then detected following sequential additions of glucose (10 mM), oligomycin (1 μM) and 2‐DG (50 mM). Results were normalized to total cell number of each well.

2.14. Lactate, pyruvic acid, and ATP content assay

Lactate content was measured using a Lactic Acid (LA) content assay kit (BC2230, Solarbio, Beijing, China). Briefly, HPASMCs sample was mixed with the reducing agent solution and reduction reaction color developing solution, and the lactate content was determined by detecting the absorbance at 570 nm. Pyruvic acid content was measured using a Pyruvate (PA) content assay kit (BC2200, Solarbio, Beijing, China) according to the manufacturer's instructions. The reaction mixtures were measured at 520 nm. The intracellular ATP content was detected via luminometer using the ATP assay kit (S0026, Beyotime, Shanghai, China) following the manufacturer's instructions.

2.15. Cell coculture assay

A noncontact transwell coculture model was used to evaluate the effect of HPASMCs on HPAECs. Well inserts (Costar, Corning, NY, USA) for 12‐well plates with a 0.4 μm/8 μm pore‐sized filter were used for the EdU and cell migration assays, respectively. Before starting the coculture experiments, HPASMCs in 12‐well plates were transfected with NC or circNAP1L4 plasmids, and HPAECs were seeded into the well inserts. Then the inserts were subsequently placed in a 12‐well plate containing HPASMCs to establish coculture conditions for the next 24 h. A 1:1 mixture of ECM: SMC was used during all the coculture experiments.

2.16. Migration assay

HPAECs were resuspended in serum‐free medium and placed in the upper chamber of a transwell plate. The migrated cells were fixed with 4% paraformaldehyde for 10 min and then stained for 30 min with 0.4% crystal violet at room temperature. The number of migrated HPAECs was counted under an inverted microscope (Nikon, Japan).

2.17. Coimmunoprecipitation (Co‐IP)

HPASMCs were lysed in lysis buffer (Tris 50 mM, pH 7.4; NaCl 150 mM; Triton X‐100 1%; EDTA 1 mM; and PMSF 2 mM) and then incubated with 5 μg of the target antibody or IgG. Protein A + G agarose beads were added overnight at 4°C. The antibody‐protein complexes were washed, and the pellet was resuspended in protein loading buffer (2×) and subjected to Western blot.

2.18. Cell and tissue immunofluorescence staining analysis

HPASMCs were cultured on coverslips in 12‐well plates and treated with different reagents. Prepared cells were fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X‐100, and blocked with 5% normal bovine serum. HPASMCs were incubated with anti‐NAP1L4 antibody (1:100, 16 018‐1‐AP; Proteintech, IL, USA) and anti‐Ki67 antibody (1:100, M00254‐3; Boster, Wuhan, China) at 4°C overnight. After washing three times with PBS, the cells were incubated with Cy3‐conjugated goat anti‐rabbit antibody (1:50, A0516; Beyotime Biotechnology, Shanghai, China) and FITC‐conjugated goat anti‐mouse antibody (1:50, A0568; Beyotime Biotechnology, Shanghai, China) followed by DAPI in the dark. Finally, the coverslips were incubated with anti‐fade mounting medium (P0126; Beyotime Biotechnology, Shanghai, China), and the cells were examined with a live cell workstation. Frozen sectioning of mouse lung tissues was performed in the same manner. The sections were incubated with anti‐PCNA antibody (1:100, #2586; Cell Signaling Technology, MA, US), anti‐α‐SMA antibody (1:100, #19245; Cell Signaling Technology, MA, US), and anti‐HKII antibody (1:50, MA01389; Boster, Wuhan, China).

The positive staining area in the vascular walls was quantified with high‐resolution images of individual vessels using a color‐recognition algorithm in Image‐Pro Plus 6.0 software. For small vessel muscularization quantification, data are expressed as the number of vessels per high‐power field. Approximately 6 arteries (>50 μm external diameter) or 6 high‐power fields (<50 μm external diameter) per animal were randomly examined under microscope at magnification × 200, and an average was calculated.

2.19. RNase R treatment

Total RNA (2 μg) was incubated with or without 3 U/μg of RNase R (Epicenter Biotechnologies, Madison, WI, USA) at 37°C for 15 min. Then, the expression levels of circNAP1L4 and linear NAP1L4 were detected via qRT–PCR.

2.20. Actinomycin D assay

HPASMCs were treated with actinomycin D (2 μg/mL, HY‐17559, MedChem Express, NJ, USA) for 0, 4, 8, and 12 h. Total RNA was subsequently extracted with TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. The expression levels of circNAP1L4 and linear NAP1L4 were detected via qRT–PCR.

2.21. Chromatin immunoprecipitation–quantitative PCR (ChIP–qPCR)

ChIP assays were performed using ChIP Assay Kits (P2078, Beyotime Biotechnology, Shanghai, China) according to the manufacturer's instructions. Briefly, HPASMCs were cross‐linked with 1% formaldehyde, followed by ultrasonic shear. The fragments of the genomic DNA were broken into 200–1000 bp size. Sonicated lysates were incubated with magnetic beads combined with anti‐H3K27ac antibody (1:100, A7253, ABclonal, Wuhan, China), anti‐H3K4me1 antibody (1:100, A2355; ABclonal, Wuhan, China), and anti‐CEBPB antibody (1:100, GTX100675; GeneTex, Southern California, USA) at 4°C overnight. Finally, the purified DNA was detected via qRT–PCR.

2.22. RNA immunoprecipitation assay

The RNA immunoprecipitation assay was performed using an RNA Immunoprecipitation (RIP) Kit (Bes5101, BersinBio, Guangzhou, China). Briefly, HPASMCs were washed with PBS three times, and the cell pellet was resuspended and lysed in RIP lysis buffer. Then, the samples were incubated with 20 μL of protein A/G bead‐conjugated anti‐NAP1L4 antibody (5 μL, 16 018‐1‐AP; Proteintech, IL, USA) and anti‐PRP8 antibody (5 μL, A6053; ABclonal, Wuhan, China) for 16 h. Normal rabbit IgG was used as a negative control. The precipitated RNA samples were extracted and purified with phenol:chloroform:isoamyl alcohol (25:24:1), and the immunoprecipitated RNA was subsequently analyzed via qRT–PCR.

2.23. RNA pull‐down assay

An RNA pull‐down assay for the circNAP1L4 and NAPIL4 proteins was performed using an RNA pull‐down kit (Bersinbio, Guangzhou, China). The biotinylated circNAP1L4 probe (989 nt–1040 nt) was synthesized by GenePharma (Shanghai, China). HPASMCs were harvested and lysed with lysis buffer. The cell extract was incubated with magnetic beads coupled with the circNAP1L4 probe or a negative control (NC) probe for 2 h at 25°C. The beads were then thoroughly washed with eluent, and the proteins pulled down by the probe were collected using protein elution buffer for 2 h at 37°C. The retrieved proteins were visualized by Western blot.

2.24. Cy3 fluorescence labeling of circNAP1L4

circNAP1L4 plasmid was labeled with Label IT Tracker Intracellular Nucleic Acid Localization Kit according to the manufacturer's instructions (Mirus, Madison, WI, USA). Then HPASMCs containing Cy3‐labeled circNAP1L4 were subjected to coculturing system for evaluating the process of circNAP1L4 transported to HPAECs.

2.25. Nuclear and cytoplasmic protein extraction

Nuclear and cytoplasmic protein extraction from HPASMCs and lung tissue was performed using a nuclear and cytoplasmic protein extraction kit (P0027; Beyotime Biotechnology, Shanghai, China) according to the manufacturer's instructions. The protein samples were detected by Western blot.

2.26. Statistical analysis

The data were analyzed using GraphPad Prism 8.0 and are expressed as the means ± SEM. All expression values were checked for a normal distribution and equal variance (F test) before statistical testing. Student's t test was used to compare the data between two groups with equal variance, and the Welch correction test was used for 2‐group analysis with unequal variance. One‐way or two‐way ANOVA with Tukey post hoc test was used to compare multiple groups with equal variance, and Brown Forsythe and Welch ANOVA with Tamhane T2 post hoc test was used to compare multiple groups with unequal variance. For nonnormally distributed data, we performed nonparametric analyses such as the Mann–Whitney U test for two groups or the Kruskal–Walli's test followed by the Dunn post‐test for multiple groups. A 2‐tailed of p < .05 was considered to indicate statistically significant.

3. RESULTS

3.1. CircNAP1L4 is downregulated in hypoxia‐induced HPASMCs and PH plasma of patients

To identify novel circRNAs involved in PH, we screened 4 circRNAs (circPAIP2, circRSRC1, circMAN1A2, and circNAP1L4) that are expressed in HPASMCs and are homologous to those expressed in rats or mice by using Cancer‐Specific CircRNA Database (http://gb.whu.edu.cn/CSCD/) and circAtlas 2.0 (http://circatlas.biols.ac.cn/) and combined with reports from the literature ¹⁹ (Figure S1A). Then, only circNAP1L4 (hsa_circ_0020816) was found to be circularized in HPASMCs by Sanger sequencing (Sanger sequencing of Figure 1A; other data not shown). CircNAP1L4 is generated from the NAP1L4 gene and is located on human chromosome 11:2972488‐3000467.

FIGURE 1.

The expression and characteristics of circNAP1L4 in HPASMCs and PH patients' plasma. (A) Genomic location and back‐splice junction site of circNAP1L4. (B) Quantitative RT–PCR (qRT–PCR) analysis of circNAP1L4 expression in hypoxic HPASMCs (n = 7). (C) Analysis of the RNA levels of circNAP1L4 and linear NAP1L4 after treatment with RNase R (n = 3). (D) Analysis of the RNA abundance of circNAP1L4 and linear NAP1L4 treated with actinomycin D (2 μg/mL) at the indicated time points (n = 6). (E) FISH was performed to determine the circNAP1L4 distribution and expression in HPASMCs. Scale bar, 50 μm. (F) qRT–PCR analyses of circNAP1L4 in plasma from control (n = 15) and PH patients (n = 25). All tests were performed at least three times, and the values are presented as the mean ± SEM. Statistical analysis was performed with one‐way ANOVA or Student's t‐test. *p < .05, **p < .01, ***p < .001. HYP, hypoxic; NOR, normoxia; PH, pulmonary hypertension.

To investigate the significance of circNAP1L4 in hypoxic HPASMCs, the expression of circNAP1L4 was determined. The results showed that circNAP1L4 was significantly downregulated in hypoxic HPASMCs but not in HPAECs (Figure 1B, Figure S1B). Compared with linear NAP1L4, circNAP1L4 was more resistant to RNase R digestion and had a longer half‐life (Figure 1C,D). Furthermore, the subcellular localization of circNAP1L4 was detected by RNA fluorescence in situ hybridization (RNA‐FISH), and the results confirmed that circNAP1L4 was completely distributed in the cytoplasm of HPASMCs (Figure 1E). In addition, we measured circNAP1L4 expression in plasma samples from patients with PH, and the plasma circNAP1L4 expression level was lower in the PH group than in the control group (Figure 1F).

3.2. Overexpressing circNAP1L4 inhibits hypoxia‐induced HPASMC proliferation and glycolysis

To assess the functional role of circNAP1L4 in regulating HPASMC proliferation under hypoxic conditions, we constructed a circNAP1L4 overexpression plasmid. The efficiency of overexpression was verified by qRT–PCR (Figure S1C). CCK8 assays showed that circNAP1L4 overexpression had no significant effect on the basic proliferation of HPASMCs but did reduce the proliferation of HPASMC triggered by hypoxia (Figure 2A, Figure S1D). Next, Ki67 and EdU assays were performed and showed that circNAP1L4 overexpression inhibited hypoxia‐induced proliferation of HPASMCs (Figure 2B). In addition, we used flow cytometry to evaluate the effects of circNAP1L4 on cell cycle progression. The results showed that circNAP1L4 overexpression reduced the percentage of cells in the G2/M phase during cell cycle progression under hypoxic conditions (Figure 2C). We then measured the levels of proliferating cell nuclear antigen (PCNA) and cell cycle‐related proteins by Western blot. The expression levels of PCNA, cyclin A, and cyclin D increased significantly under hypoxic conditions, and these effects were reversed in the presence of circNAP1L4 overexpression (Figure 2D).

FIGURE 2.

CircNAP1L4 inhibits hypoxia‐induced HPASMC proliferation and glycolysis. (A) CCK8 assays were performed for HPASMCs overexpressing circNAP1L4 (n = 6). (B) EdU (red) incorporation and immunofluorescence analysis of Ki67 (green) in HPASMCs overexpression circNAP1L4 (n = 5–6). Scale bar, 50 μm. (C) Fluorescence‐activated cell sorting analyses were used to detect the number of HPASMCs in each phase of the cell cycle. The bar graph shows the number of cells in each cell cycle phase. (D) Protein levels of cyclin A, cyclin D, and PCNA in HPASMCs overexpressing circNAP1L4 (n = 6). (E) Protein levels of HK II and PKM2 in HPASMCs transfected with circNAP1L4 overexpression (n = 6). (F) Extracellular acidification rate (ECAR) of HPASMCs overexpressing circNAP1L4 was measured via the Seahorse XFe24 platform (n = 3). (G) Lactate, pyruvate production, and ATP content in HPASMCs treated circNAP1L4 (n = 6). All tests were performed at least three times, and the values are presented as the mean ± SEM. Statistical analysis was performed with one‐way ANOVA. *p < .05, **p < .01, ***p < .001. HYP, hypoxia; NC, negative control; NOR, normoxia.

As glycolysis is an important factor that contributes to cell proliferation, ²⁰ the expression of key glycolytic enzymes, including hexokinase II (HK II) and pyruvate kinase 2 (PKM2), was detected. Hypoxia upregulated the expression of HK II and PKM2, which were reversed by circNAP1L4 overexpression in HPASMCs (Figure 2E). Furthermore, the extracellular acidification rate (ECAR) was monitored in real time to estimate glycolysis using a Seahorse XFe24 Extracellular Flux Analyzer to elucidate the effects of circNAP1L4 on glycolysis in HPASMCs. The results showed that overexpression of circNAP1L4 significantly decreased the glycolytic capacity (Figure 2F). In addition, the overexpression of circNAP1L4 reduced lactate and pyruvate production and increased ATP levels (Figure 2G). Taken together, these findings imply that circNAP1L4 suppresses hypoxia‐induced HPASMC proliferation and glycolysis.

3.3. CircNAP1L4 inhibits NAP1L4 protein nuclear translocation by interacting with NAP1L4 protein

It is axiomatically known that circRNAs can exert their functions through regulating their parental gene expression. ²¹ To clarify the regulatory mechanisms of circNAP1L4 on HPASMC proliferation and glycolysis, the effects of circNAP1L4 on its parent gene NAP1L4 were investigated. We found that NAP1L4 exhibited higher expression in HPASMCs under hypoxic conditions (Figure S2A,B). Moreover, overexpression of circNAP1L4 did not affect the mRNA and protein expression of NAP1L4 (Figure 3A,B). However, it inhibited the translocation of endogenous NAP1L4 protein from the cytoplasm to the nucleus under hypoxic conditions, but did not affect NAP1L4 mRNA translocation (Figure 3C, Figure S2C). The same results were also obtained by nuclear and cytoplasmic protein extraction (Figure 3D).

FIGURE 3.

CircNAP1L4 inhibits NAP1L4 nuclear translocation by interacting with NAP1L4. (A) qRT–PCR was used to detect NAP1L4 mRNA expression in HPASMCs transfected overexpressing circNAP1L4 (n = 6). (B) Protein levels of NAP1L4 in HPASMCs overexpression circNAP1L4 (n = 6). (C) Immunofluorescence analysis of NAP1L4 expression in HPASMCs overexpression circNAP1L4. Scale bar, 50 μm. (D) The translocation of NAP1L4 between the nucleus and cytoplasm in HPASMCs overexpression circNAP1L4 (n = 4). (E) Interactions between circNAP1L4 and the NAP1L4 protein were predicted by using catRAPID website (http://s.tartaglialab.com/page/catrapid_omics2_group). (F) RNA pull‐down assays were used to detect the interaction between circNAP1L4 (nucleotide position [nt] 989–1040) and the NAP1L4 protein. (G) RIP enrichment was determined as circNAP1L4 associated with the NAP1L4 IP relative to the input control (n = 5). (H) Immunofluorescence was used to observe the colocalization of circNAP1L4 and NAP1L4 proteins in the cytoplasm. Scale bar, 50 μm. All tests were performed at least three times, and the values are presented as the mean ± SEM. Statistical analysis was performed with Student's t‐test. *p < .05, ***p < .001. HYP, hypoxia; NC, negative control; NOR, normoxia.

Next, we assessed the likelihood of protein–RNA interactions between circNAP1L4 and the NAP1L4 protein using the catRAPID website. The heatmap results showed that circNAP1L4 can interact with the NAP1L4 protein, and nucleotides 989 to 1040 of circNAP1L4 bind to NAP1L4 with greater interaction propensity (Figure 3E, Figure S3A). To clarify the possibility that circNAP1L4 interacts with NAP1L4, RNA pull‐down assay and subsequent Western blot analysis were performed, and the results confirmed that nucleotides 989 to 1040 of circNAP1L4 interact with NAP1L4 (Figure 3F). Furthermore, RNA immunoprecipitation (RIP), qRT–PCR, and agarose gel electrophoresis assays demonstrated that circNAP1L4‐NAP1L4 complex was increased in circNAP1L4‐overexpressed compared with that in the NC group under hypoxia (Figure 3G, Figure S3B). In addition, colocalization of circNAP1L4 and NAP1L4 was detected via implementing FISH and immunofluorescence staining colocalization assay. When circNAP1L4 was overexpressed, colocalization of circNAP1L4 and NAP1L4 in the cytoplasm was increased compared with NC group in hypoxia (Figure 3H). Finally, we used the HNADOCK server to predict and analyze the 3D structural docking of circNAP1L4 and NAP1L4 and visualized the structure of the binding complex between circNAP1L4 and NAP1L4 (Figure S3C). Taken together, our data show that circNAP1L4 binds to NAP1L4 and modulates its nuclear translocation.

3.4. NAP1L4 is involved in circNAP1L4‐mediated HPASMC proliferation and glycolysis under hypoxia

To further elucidate the role of NAP1L4 in HPASMC proliferation and glycolysis, we transfected NAP1L4 siRNA into HPASMCs. The interference efficiency was verified by Western blot (Figure S2D). CCK8, EdU, and Ki67 immunofluorescence assays showed that NAP1L4 siRNA compromised hypoxia‐induced HPASMC proliferation (Figure S4A,B). We detected the effect of NAP1L4 on cell cycle progression, and the results showed that interference with NAP1L4 decreased the percentage of G2/M + S phase cells under hypoxic conditions, accompanied by an increase in the number of cells in the G0/G1 phase (Figure S4C). The expression of PCNA and cell cycle‐related proteins (cyclin A and cyclin D) was downregulated in hypoxia‐exposed HPASMCs by silencing NAP1L4 (Figure S4D). Additionally, glycolysis was monitored in real time using a Seahorse XFe24 Extracellular Flux Analyzer. As expected, NAP1L4 knockdown decreased the glycolytic stress activation stimulated by hypoxia (Figure S4E). Besides, hypoxia upregulated the production of lactate and pyruvate and decreased the expression of ATP, which were reversed by NAP1L4 knockdown (Figure S4F).

To investigate whether the effect of circNAP1L4 on proliferation and glycolysis is dependent on NAP1L4 in HPASMCs, we constructed a circNAP1L4 plasmid (MUT) which the mutated nucleotides 989 to 1040 of circNAP1L4 interacted with NAP1L4. When circNAP1L4 was mutated, NAP1L4 translocation to the nucleus was increased compared with that of wild‐type circNAP1L4 under hypoxic conditions, suggesting that MUT‐circNAP1L4 did not bind to NAP1L4 (Figure S5A). Moreover, the circNAP1L4‐mediated attenuation of hypoxia‐induced HPASMC proliferation was rescued by MUT‐circNAP1L4 (Figure S5B–D). In addition, Seahorse and Western blot experiments showed that hypoxia significantly enhanced the expression of glycolysis, which was inhibited by the overexpression of wild‐type circNAP1L4 but not by the overexpression of MUT‐circNAP1L4 (Figure S5E,F). Similar results were also obtained in Figure S5G. These findings demonstrated that circNAP1L4, which directly binds to NAP1L4, blocks hypoxia‐induced HPASMC proliferation and glycolysis.

3.5. NAP1L4 promotes the expression of the SE‐modified glycolysis gene HK II in hypoxic HPASMCs

As a histone chaperone, NAP1L4 has been shown to regulate transcription by interacting with the SE‐related protein p300. ²² The Co‐IP assay results verified the interaction of NAP1L4 with the SE marker H3K27ac (Figure 4A). Subsequently, we explored how NAP1L4 affects the SE modification of HK II or PKM2, two key enzymes in glycolysis. To determine whether hypoxia‐induced upregulation of HK II or PKM2 expression is associated with SE in HPASMCs, we treated cells with JQ1, a SE inhibitor, and found that JQ1 specifically inhibited the expression of HK II but not the PKM2 in hypoxia, suggesting that HK II was driven by SE (Figure 4B). Next, through joint analysis of the SEdb and WashU Epigenome Browser databases, we further found that the HK II transcript was marked by the monomethyl H3K4 (H3K4me1) and acetyl H3K27 (H3K27ac) signaling peaks in the SE region of lung tissue. We divided the SE region of HK II into three constituents (E1–E3) (Figure 4C). Furthermore, the expression of HK II was downregulated in hypoxia‐exposed HPASMCs by silencing NAP1L4 (Figure 4D).

FIGURE 4.

NAP1L4 enhances the expression of the SE‐modified glycolysis gene HK II in hypoxic HPASMCs. (A) Coimmunoprecipitation (Co‐IP) assay verified the interaction of H3K27ac and NAP1L4. (B) qRT–PCR was used to detect HK II and PKM2 mRNA expression levels in HPASMCs treated with JQ1 (250 nM; S7110; Selleck, USA) (n = 6). (C) Gene tracks depicting the SE region of HK II in lung primary cells and lung tissue with measured H3K27ac, or H3K4me1 marks. (D) Protein levels of HK II in HPASMCs with NAP1L4 knockdown (n = 6). (E) CEBPB was identified as a candidate transcription factor, and Western blot analysis was used to verify the expression of CEBPB in hypoxia (n = 6). (F) HPASMCs were subjected to ChIP analysis using antibodies against H3K27ac, H3K4me1, and CEBPB. The association with the SE region (E1–E3) of HK II was quantified by qRT–PCR (n = 3–4). (G) HPASMCs were subjected to ChIP analysis using antibodies against H3K27ac, H3K4me1, and CEBPB. The association with the promoter region (△1–△5) of HK II was quantified by qRT–PCR (n = 3). All tests were performed at least three times, and the values are presented as the mean ± SEM. Statistical analysis was performed with one‐way ANOVA or Student's t‐test. *p < .05, **p < .01, ***p < .001. HYP, hypoxic; NC, negative control; NOR, normoxia.

As SE can bind to transcription factors to stimulate transcription across large genomic distances, ²³ we identified candidate transcription factors that could bind to both of SE region and promoter region of HK II using SE databases (SEanalysis and SEdb) and promoter databases (JASPAR and AnimalTFDB). The transcription factor CEBPB was noted, and its expression was upregulated in hypoxic HPASMCs (Figure 4E, Figure S3D).

To clarify the direct regulatory effects of H3K27ac, H3K4me1, and CEBPB on HK II at the transcriptional level through the SE, ChIP–qPCR was performed to quantify the occupancy of H3K27ac, H3K4me1, and CEBPB on the HK II SEs (EI–E3). We found a significant association between CEBPB and the E1–E3 SE regions of HK II in hypoxic HPASMCs (Figure 4F). According to the results of the other two direct ChIP–qPCR studies, the results showed that siNAP1L4 under hypoxic conditions decreased H3K27ac enrichment at the E1 SE region of HK II and decreased H3K4me1 enrichment at the E1 and E2 SE regions of HK II (Figure 4F). Therefore, the results indicate that CEBPB, H3K27ac, and H3K4me1 mainly bind to the E1 SE region of HK II to play regulatory roles.

SEs often form long‐range chromatin loops with the promoters of target genes to control gene expression. ²⁴ To clarify the specific mechanism responsible for chromatin loop formation, we divided the HK II promoter into five equal segments (+100 – −2000 △1–△5) according to the distribution of binding sites between CEBPB and the HK II promoter. We performed ChIP with an antibody against CEBPB followed by qPCR and observed that hypoxia increased CEBPB enrichment at the △3 promoter region of HK II in HPASMCs (Figure 4G). Additionally, ChIP–qPCR results demonstrated that H3K27ac specifically bound to △1, △3, and △4 within the HK II promoter in hypoxic HPASMCs, which was reduced by silencing NAP1L4 (Figure 4G). Further ChIP–qPCR analysis revealed that NAP1L4 knockdown strongly decreased H3K4me1 levels at △1 and △3 of the HK II promoter in hypoxic HPASMCs (Figure 4G). Taken together, these data demonstrate that H3K27ac, H3K4me1, and CEBPB interact with both the E1 SE and △3 promoter regions of HK II and enhance its transcription during glycolysis.

3.6. CircNAP1L4 regulates HPASMC proliferation through HK II under normoxic conditions

We performed rescue experiments to determine whether HK II is involved in circNAP1L4‐mediated HPASMC proliferation. circNAP1L4 and HK II were knocked down in HPASMCs using specific siRNAs, and the transfection efficiency was verified by qRT–PCR and Western blot (Figure S6A,B). CircNAP1L4 knockdown promoted cell viability, Ki67‐ and EdU‐positive cells in HPASMCs, but this effect was reversed by HK II siRNA (Figure S6C,D). Moreover, HK II knockdown in HPASMCs partially restored the increase in the protein levels of PCNA, cyclin A, and cyclin D induced by circNAP1L4 siRNA (Figure S6E). Taken together, these results imply that HK II, a downstream gene of circNAP1L4, is involved in circNAP1L4‐mediated HPASMC proliferation.

3.7. CircNAP1L4 in HPASMCs affects the glycolysis and proliferation but not the migration of HPAECs

Because interactions between PAECs and PASMCs contribute to pulmonary vascular remodeling, we assessed the interactions of HPASMCs with HPAECs using a noncontact transwell coculture model in the presence/absence of circNAP1L4 (Figure 5A). In order to visualize whether circNAP1L4 released from HPASMCs in extracellular vesicles are transported to HPAECs, we transfected HPASMC with a Cy3‐labeled circNAP1L4, and co‐culturing with HPAECs for 24 h. The fluorescence of Cy3‐circNAP1L4 was shown in HPAECs, indicating that circNAP1L4 was transported from HPASMCs to HPAECs (Figure 5B). The levels of circNAP1L4 were examined in HPAECs cocultured with HPASMCs, and we found that circNAP1L4 was decreased in HPAECs of HPASMCs co‐culture under hypoxia (Figure 5C). Moreover, lactate content and EdU staining showed that circNAP1L4 inhibited the glycolysis and proliferation of HPAECs under both normoxia and hypoxia when HPASMCs were cocultured with HPAECs (Figure 5D,E). To further assess the functional consequence of circNAP1L4 overexpression in HPAECs, migration was evaluated. However, overexpression of circNAP1L4 did not regulate HPAEC migration when HPASMCs were cocultured with HPAECs under either normoxia or hypoxia (Figure 5F). Collectively, our data demonstrate that circNAP1L4 in HPASMCs plays a critical role in regulating HPAEC proliferation.

FIGURE 5.

CircNAP1L4 in HPASMCs affects the glycolysis and proliferation but not the migration of HPAECs. (A) Schematic of the transwell coculture model. (B) Fluorescence images of Cy3‐circNAP1L4 fluorescence in HPAECs cultured without (CON) or with HPASMCs transfected with Cy3‐circNAP1L4. Scale bar, 50 μm. (C) qRT‐PCR showed the circNAP1L4 expression level in HPAECs after co‐culture under hypoxic conditions. (D) Lactate production in transwell coculture model (n = 6). (E) EdU (red) incorporation in transwell coculture model (n = 6). Scale bar, 50 μm. (F) Cell migration assay in transwell coculture model (n = 6). Scale bar, 100 μm. All tests were performed at least three times, and the values are presented as the mean ± SEM. *p < .05, **p < .01, ***p < .001. Statistical analysis was performed with two‐way ANOVA. CON, control; HYP, hypoxic; NC, negative control; NOR, normoxia.

3.8. The splicing factor PRP8 is required for the hypoxia‐induced increase in proliferation and glycolysis in HPASMCs via circNAP1L4

Previous reports suggest that the production of linear RNAs shifts to circular RNAs when the core spliceosome components affect pre‐mRNA processing. ²⁵ To better understand the mechanisms underlying the inconsistent levels of circNAP1L4 and linear NAP1L4 expression in HPASMCs under hypoxia, we first selected five core spliceosome proteins, including pre‐mRNA‐processing‐splicing factor 8 (PRP8), pre‐mRNA‐splicing factor SLU7 (SLU7), pre‐mRNA‐processing factor 17 (PRP17), splicing factor 3B subunit 1 (SF3B1), and putative RNA‐binding protein Luc7‐like 1 (LUC7L). Only the expression of PRP8 in hypoxic HPASMCs was significantly enhanced (Figure 6A). Moreover, we examined the protein expression of PRP8 and found that the PRP8 protein level was elevated in hypoxic HPASMCs compared with that in the normal group (Figure 6B). To further confirm the roles of PRP8 in modulating the expression of the circNAP1L4 and linear NAP1L4 under hypoxic conditions, we transfected cultured HPASMCs with PRP8 siRNA. The efficiency of transfection was verified by a significant decrease in the expression of endogenous PRP8 after siRNA transfection, and we found that siPRP8‐3 had the most obvious effects (Figure S7). qRT–PCR revealed decreased expression of circNAP1L4 and increased expression of linear NAP1L4 under hypoxia, which was reversed by PRP8 siRNA transfection, suggesting that the output of circNAP1L4 and linear NAP1L4 was mediated by PRP8 under hypoxia (Figure 6C). In addition, we performed an RIP assay, and the results suggested that PRP8 directly interacts with NAP1L4 pre‐mRNA to play a splicing role (Figure 6D). In general, our data showed that linear NAP1L4 becomes the preferred mode of NAP1L4 pre‐mRNA processing when the spliceosome components PRP8 is abundant under hypoxia, thereby causing decreased production of circNAP1L4.

FIGURE 6.

PRP8 is involved in HPASMC proliferation and glycolysis through circNAP1L4. (A) qRT–PCR was used to detect the mRNA expression of PRP8, SLU7, PRP17, SF3B1, and LUC7L in hypoxic HPASMCs (n = 6). (B) Protein levels of PRP8 in HPASMCs (n = 10). (C) qRT–PCR was used to detect the expression levels of circNAP1L4 and NAP1L4 in HPASMCs with PRP8 knockdown (n = 6). (D) RIP enrichment was determined as the NAP1L4 pre‐mRNA associated with the PRP8 IP relative to the input control (n = 4). (E) CCK8 assays of HPASMCs cotransfected with PRP8 and circNAP1L4 siRNA (n = 6). (F) EdU (red) incorporation in HPASMCs cotransfected with PRP8 and circNAP1L4 siRNA (n = 6). Scale bar, 50 μm. (G) Immunofluorescence analysis of Ki67 (green) expression in HPASMCs cotransfected with PRP8 and circNAP1L4 siRNA (n = 6). Scale bar, 50 μm. (H) Fluorescence‐activated cell sorting analyses detected the number of HPASMCs in each phase of the cell cycle. The bar graph shows the number of cells in each cell cycle phase. (I) Extracellular acidification rates (ECAR) of HPASMCs with PRP8 and circNAP1L4 siRNA were measured via a Seahorse XFe24 (n = 3). (J) Lactate, pyruvate production, and ATP content in HPASMCs treated PRP8 and circNAP1L4 siRNA (n = 6–8). All tests were performed at least three times, and the values are presented as the mean ± SEM. Statistical analysis was performed with one‐way ANOVA or Student's t‐test. *p < .05, **p < .01, ***p < .001. HYP, hypoxia; NC, negative control; NOR, normoxia.

To validate the role of the PRP8/circNAP1L4 axis in hypoxia‐induced HPASMC proliferation and glycolysis, we performed functional recovery experiments after the knockdown of PRP8. CCK8 results revealed that transfection with PRP8 siRNA reversed hypoxia‐induced HPASMC proliferation, and this effect was eliminated after silencing circNAP1L4 (Figure 6E). Then, we determined that the Ki67‐ and EdU‐positive cells were reduced by PRP8 siRNA and enhanced by sicircNAP1L4 cotransfection (Figure 6F,G). Flow cytometry assays demonstrated that cell proliferation was minimized after transfection with PRP8 siRNA in response to hypoxic stimulation, and this decrease was reversed by circNAP1L4 siRNA (Figure 6H). Moreover, HPASMC glycolysis was abrogated by PRP8 siRNA under hypoxia, and circNAP1L4 siRNA enhanced this effect (Figure 6I). PRP8 siRNA was sufficient to inhibit lactate and pyruvate production and increase the ATP content of HPASMCs in hypoxia, which were reversed by circNAP1L4 knockdown (Figure 6J). Taken together, these findings suggest that PRP8 leads to abnormal activation of HPASMC proliferation and glycolysis under hypoxia, which is achieved mainly through circNAP1L4.

3.9. CircNAP1L4 alleviates SuHx‐induced PH in mice

We found that the rat homolog of circNAP1L4, but circNAP1L4 was not circularized in rat PASMCs by Sanger sequencing. Previous studies have shown that human noncoding RNAs can interact with mouse proteins to play a therapeutic role in some diseases. ²⁶ , ²⁷ Although we did not find a mouse homolog of circNAP1L4, the protein sequence of NAP1L4 is highly conserved among humans and mice (Figure S8A). More importantly, human circNAP1L4 could bind to NAP1L4 in mouse PASMCs (Figure S8B). We used AAV5 to overexpress circNAP1L4 and constructed a SuHx‐PH model (Figure 7A). Subsequently, circNAP1L4 overexpression was verified by qRT–PCR (Figure 7B). We performed FISH analysis to determine the distribution of circNAP1L4 in lung tissues and found that circNAP1L4 was expressed in pulmonary artery smooth muscle layer (Figure S8C). We found that circNAP1L4 overexpression inhibited the right ventricle‐to‐left ventricle plus septum (RV/LV+S) ratio and RVSP induced by SuHx (Figure 7C,D). According to echocardiographic measurements, circNAP1L4 attenuated the decrease in the pulmonary artery velocity time integral (PAVTI) and pulmonary artery acceleration time (PAAT) induced by SuHx, which had no effect on the left ventricular ejection fraction (LVEF) and heart rate (Figure 7E, Figure S8D). Furthermore, HE staining, α‐SMA staining, and Masson staining revealed that circNAP1L4 reversed SuHx‐induced distal pulmonary vascular remodeling, small muscularized vessels, and collagenation (Figure 7F–H, Figure S8E).

FIGURE 7.

Delivery of human circNAP1L4 alleviated Su5416+hypoxia (SuHx)‐induced pulmonary hypertension in mice. (A) Serotype 5 adenovirus‐associated virus (AAV5) with the NC or circNAP1L4 sequence was used to treat mice via dropwise intranasal instillation before intraperitoneal injection of Su5416. (B) Overexpression efficiency of circNAP1L4 was quantified by qRT–PCR in lung tissues (n = 8). (C) The right ventricle (RV)/(left ventricle (LV)+septum) weight ratio was calculated (n = 8). (D) Indices of right ventricular systolic pressure (RVSP) (n = 8). (E) Pulmonary artery velocity time integral (PAVTI), pulmonary artery acceleration time (PAAT), left ventricular ejection fraction (LVEF), and heart rate in AAV5–circNAP1L4‐treated mice (n = 8). (F) Pulmonary arterial morphological analysis was performed by using hematoxylin and eosin (HE) staining. Scale bar, 200 μm (n = 8). (G) α‐SMA immunostaining of a midsized pulmonary artery. Scale bar, 50 μm (n = 8). (H) Muscularized small vessels (<50 μm external diameter) identified by positive α‐SMA immunostaining. Arrows indicate muscularized small vessels. Data are expressed as the number of vessels per high‐power field (HPF). Scale bar, 50 μm (n = 8). All tests were performed at least three times, and the values are presented as the mean ± SEM. Statistical analysis was performed with one‐way ANOVA. *p < .05, **p < .01, ***p < .001. HPF, high‐power field; HYP, hypoxia; NC, negative control; NOR, normoxia.

In addition, immunofluorescence staining for PCNA revealed that overexpression of circNAP1L4 prevented the proliferation of both the smooth muscle layer and endothelial layer in SuHx‐treated mice (Figure S9A,B). We examined the effect of circNAP1L4 overexpression on NAP1L4 in lung tissues, and the translocation of NAP1L4 into the cytoplasm was significantly increased when circNAP1L4 was overexpressed under SuHx conditions, suggesting that circNAP1L4 inhibited the translocation of endogenous NAP1L4 from the cytoplasm to the nucleus (Figure S9C). Immunofluorescence staining and Western blot for HK II revealed that overexpression of circNAP1L4 prevented increases in HK II expression in SuHx‐treated mice (Figure S9D,E). The above results confirmed that circNAP1L4 plays a critical role in inhibiting PASMC glycolysis, proliferation, and PH progression associated with hypoxia.

4. DISCUSSION

In this study, we investigated the involvement of circNAP1L4, a novel circRNA, in the pathological process of hypoxia‐induced PASMC proliferation. This involvement was confirmed by the following experimental findings. First, circNAP1L4 was frequently downregulated in HPASMCs, and an increase in PRP8 contributed to the downregulation of circNAP1L4. Second, overexpression of circNAP1L4 inhibited hypoxia‐induced HPASMC proliferation and PH progression by regulating glycolysis, and circNAP1L4‐mediated cell–cell communication between HPASMCs and HPAECs. Third, circNAP1L4 interacts with its host gene NAP1L4 to regulate the nuclear translocation of NAP1L4. Finally, NAP1L4 increased the level of the SE modification of the glycolytic enzyme HK II and thus activated the transcription of HK II under hypoxic conditions. Overall, our study revealed that circNAP1L4 inhibits hypoxia‐induced PASMC glycolysis and proliferation, which are novel mediators of PH.

Increasing evidence indicates that the expression level of circRNAs is involved in PASMC proliferation in PH. Our previous study showed that the circ‐calm4/miR‐337‐3p/myosin 10 (Myo10) signal transduction axis promoted the PASMC proliferation. ²⁸ Furthermore, hsa_circ_0002062 was reported to significantly increase and isolate miR‐942‐5p under hypoxic conditions, eventually activating CDK6 and thus promoting the proliferation of PASMCs. ²⁹ In the present study, we found that decreased circNAP1L4 was significantly associated with cell proliferation. Since glycolysis has been increasingly revealed to be a hallmark of cell proliferation in diverse diseases, it is possible that circNAP1L4 mediates PASMC proliferation through the glycolysis pathway. Using Seahorse and the key enzyme involved in glycolysis, we proved that circNAP1L4 can inhibit the expression of glycolysis‐related enzymes and the HPASMC ECAR. These results suggest that circNAP1L4 mediates HPASMC proliferation partially through the glycolysis pathway.

CircRNAs can play vital biological roles by regulating parental genes in different ways, such as by forming an R‐loop with parent DNA to regulate the splicing of their parent mRNAs ³⁰ , ³¹ and suppressing the transcriptional activity of transcription factors to regulate the transcription of their parent DNA. ³² In our study, we determined that circNAP1L4 regulates HPASMC proliferation through its parental protein NAP1L4. In addition to the reported mechanism through which circRNAs regulate parental genes, we identified a new mechanism by which circNAP1L4 reduces the nuclear translocation of NAP1L4 by forming circNAP1L4/NAP1L4 complexes. We also provide evidence that the 989 nt–1040 nt region of circNAP1L4 is the main binding site for the NAP1L4 protein. Therefore, our data reveal a previously unrecognized mechanism through which circRNAs participate in biological functions.

NAP1L4, a histone chaperone, is involved in the process of integrating coactivators with chromatin. ²² , ³³ A recent study showed that NAP1L4 can regulate cell proliferation through Lys320 acetylation of p53, ³⁴ but circNAP1L4 cannot affect the expression of p53 in HPASMCs caused by hypoxia, suggesting that circNAP1L4 inhibited hypoxia‐induced HPASMC proliferation via NAP1L4 without relying on the regulation of p53 (Figure S7B). Our present study revealed that NAP1L4 translocation into the nucleus was inhibited when circNAP1L4 was overexpressed in hypoxia‐exposed HPASMCs, thus decreasing SE modification of HK II. Indeed, the silencing of NAP1L4 in HPASMCs resulted in the attenuation of glycolysis as well as significant decreases in hypoxia‐mediated cell proliferation. Nevertheless, we did not clarify the mechanism of NAP1L4 translocation from the cytoplasm to the nucleus, and future studies are necessary to investigate whether circNAP1L4 regulates the nuclear localization signal of NAP1L4 protein.

SEs are a vital type of transcriptional regulation involved in cancer and cardiovascular disease development. ³⁵ , ³⁶ The SE inhibitor JQ‐1 has been proven to significantly suppress proliferation in many tumors and cardiovascular diseases, including PH, whereas how SE triggers PASMC proliferation in PH needs to be further explored. ³⁷ , ³⁸ , ³⁹ In this study, we revealed that SE modification (in the E1 region) was enriched in the promoter region (in the △3 region) of the glycolysis‐related enzyme HK II, after which the transcription factor CEBPB was recruited to binding sites to promote the transcription of HK II in hypoxic HPASMCs. The activity of SE modification in HK II was regulated by the interaction between circNAP1L4 and its host protein NAP1L4, ultimately affecting glycolysis in HPASMCs. Interestingly, the SE inhibitor JQ‐1 did not affect PKM2 expression, the possible mechanism by which PKM2 is regulated by circNAP1L4 likely involves other regulation pathway of NAP1L4, such as acetylation. Apparently, more rigorous investigations are needed to elucidate this mechanism.

Previous studies suggest a shift from canonical splicing that generates linear to back‐splicing that results in circRNA production when pre‐mRNA processing is limited by the inhibition of the spliceosome. ²⁵ , ⁴⁰ , ⁴¹ In addition, the splicing events driven by some upstream regulators are influenced by hypoxia, including RNA‐binding‐motif protein 25 (RBM25), polypyrimidine tract binding protein 1 (PTBP1), scaffold attachment factors B (SAFB), and serine and arginine rich splicing factor 2 (SRSF2). ¹⁶ , ¹⁷ , ⁴² , ⁴³ Here, we showed that the expression of PRP8, the largest protein component in the spliceosome, was triggered by hypoxia in HPASMCs. As expected, silencing PRP8 resulted in decreased linear NAP1L4 and increased circNAP1L4 expression under hypoxic condition, as demonstrated by qRT–PCR. Importantly, our results also showed that PRP8 siRNA reversed hypoxia‐induced HPASMC glycolysis and proliferation, suggesting that PH might be due in part to the aberrant NAP1L4 pre‐mRNA splicing caused by PRP8 in PASMCs. However, exactly how PRP8 is involved in NAP1L4 pre‐mRNA splicing is unclear and needs to be studied in the future.

In conclusion, our data demonstrated that overexpression of circNAP1L4 inhibits hypoxia‐induced PASMC proliferation and PH progression by reducing the expression of the SE‐associated glycolysis gene HK II via its host protein NAP1L4. More importantly, the circNAP1L4 level was found to be decreased in the plasma of PH patients and involved in the crosstalk between PASMCs and PAECs (Figure 8). These findings broaden the knowledge of circRNA action in PH progression and provide potential therapeutic targets for its treatment.

FIGURE 8.

A schematic diagram illustrating the hypothetical role of circNAP1L4 in regulating vascular remodeling and PH. The splicing factor PRP8 was triggered by hypoxia in PASMCs, linear NAP1L4 (NAP1L4 mRNA) expression was upregulated, and circNAP1L4 expression was downregulated. The binding of the circNAP1L4 and NAP1L4 protein is reduced in the cytoplasm, leading to entry of the NAP1L4 protein into the nucleus. Moreover, NAP1L4 increases the SE modification of the glycolytic gene HK II and further recruits the transcription factor CEBPB to promote the transcription of HK II, leading to PASMC glycolysis and proliferation. In addition, a small amount of circNAP1L4 released by PASMC is transported to PAEC and promotes proliferation. Finally, increased cell proliferation leads to vascular remodeling and PH.

AUTHOR CONTRIBUTIONS

D.Z. and X.W. performed study concept and design; D.Z., C.M., L.Z., and X.W. performed development of methodology and writing, review, and revision of the paper; X.W., X.G., X.Z., L.Z., J.B., J.M., and S.H. provided acquisition, analysis, and interpretation of data, and statistical analysis; Q.L., N.S., and B.W. provided technical and material support. All authors read and approved the final paper.

FUNDING INFORMATION

This study was supported by the National Natural Science Foundation of China [31820103 007, 31971057, 31 771 276 to DZ, 82 170 059 to CM and 81 870 043 to BW], Natural Science Foundation of Heilongjiang Province [ZD2023H003 to CM], and the Fundamental Research Funds for the Provincial Universities [JFQN202301 to XW].

DISCLOSURES

The authors declare no competing interests.

ETHICS STATEMENT

All experimental procedures were performed in accordance with the ethical standards in the 1964 Declaration of Helsinki and its later amendments and approved by the Ethics Committees of Harbin Medical University (HMUDQ20220527001).

Supporting information

Appendix S1..

Wang X, Guan X, Zhu X, et al. CircNAP1L4 regulates pulmonary artery smooth muscle cell proliferation via the NAP1L4‐mediated super‐enhancer‐driven glycolysis gene hexokinase II (HK II) in pulmonary hypertension. The FASEB Journal. 2024;38:e23868. doi: 10.1096/fj.202400585RRR

DATA AVAILABILITY STATEMENT

All data and materials generated and analyzed during the course of this study are included in this article and its supplementary information files.