Hsa_circ_0016070/micro‐340‐5p Axis Accelerates Pulmonary Arterial Hypertension Progression by Upregulating TWIST1 Transcription Via TCF4/β‐Catenin Complex

Abstract

Background

Hypoxia is considered a major leading cause of pulmonary hypertension (PH). In this study, the roles and molecular mechanism of circ_0016070 in PH were studied.

Methods and Results

The expression of circ_0016070 in serum samples, human pulmonary artery smooth muscle cells and hypoxia/monocrotaline‐treated rats was determined by real‐time quantitative polymerase chain reaction. Cell viability, migration, and apoptosis were analyzed by Cell Counting Kit‐8, wound healing, flow cytometry, and TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) assays, respectively. The molecular interactions were validated using RNA immunoprecipitation, chromatin immunoprecipitation, and dual luciferase reporter assays. The levels of phenotype switch‐related proteins were evaluated by Western blot and immunohistochemistry. The pathological characteristics were assessed using hematoxylin and eosin staining. circ_0016070 was highly expressed in the serum samples, hypoxia‐induced pulmonary artery smooth muscle cells and pulmonary arterial tissues of PH rats. Downregulation of circ_0016070 ameliorated the excessive proliferation, migration, vascular remodeling, and phenotypic transformation but enhanced cell apoptosis in the PH rat model. In addition, micro (miR)‐340‐5p was verified as a direct target of circ_0016070 and negatively regulated TCF4 (transcription factor 4) expression. TCF4 formed a transcriptional complex with β‐catenin to activate TWIST1 (Twist family bHLH transcription factor 1) expression. Functional rescue experiments showed that neither miR‐340‐5p inhibition nor TWIST1 or TCF4 upregulation significantly impeded the biological roles of circ_0010670 silencing in PH.

Conclusions

These results uncovered a novel mechanism by which circ_0016070 play as a competing endogenouse RNA of miR‐340‐5p to aggravate PH progression by promoting TCF4/β‐catenin/TWIST1 complex, which may provide potential therapeutic targets for PH.

Nonstandard Abbreviations and Acronyms

3'‐UTR

3'‐untranslated region

BMPR2

bone morphogenetic protein receptor type 2

circRNAs

circular RNAs

FITC

fluorescein isothiocyanat

GATA6

GATA binding protein 6

HIF‐1α

hypoxia inducible factor 1 subunit alpha

IVS

interventricular septum

MCT

monocrotaline

miRNAs

microRNAs

mPAP

mean pulmonary artery pressure

PASMCs

pulmonary artery smooth muscle cells

PH

pulmonary hypertension

RT‐qPCR

real‐time quantitative polymerase chain reaction

RVI

right heart hypertrophy index

PI

propidium iodide

PVDF

polyvinylidene fluoride

TCF4

transcription factor 4

TUNEL

Terminal deoxynucleotidyl transferase dUTP nick end labelling

TWIST1

Twist family bHLH transcription factor 1

Clinical Perspective.

What Is New?

• Our data showed that a newly discovered circular RNA, circ_0016070, was highly expressed in patients with pulmonary hypertension, hypoxia‐induced pulmonary artery smooth muscle cells, and pulmonary arterial tissues of pulmonary hypertension rats.

• Circ_0016070 competitively released TCF4 (transcription factor 4) expression via sponging micro (miR)‐340‐5p, thus activating TWIST1 (Twist family bHLH transcription factor 1) transcription.

• Downregulation of circ_0016070 alleviated the excessive proliferation and migration and suppressed vascular remodeling but enhanced apoptosis triggered by hypoxia in pulmonary hypertension by regulating the miR‐340‐5p/TCF4/β‐catenin/TWIST1 axis.

What Are the Clinical Implications?

• Our results indicate that circ_0016070 may be a potential biomarker and therapeutic target for patients with pulmonary hypertension.

Pulmonary hypertension (PH) is characterized mainly by progressive elevation of pulmonary arterial resistance. ¹ Eventually, the patients died from right heart failure. The main pathophysiological changes are pulmonary vasoconstriction and pulmonary vascular remodeling. ² Although vasodilators were used to reduce pulmonary artery pressure to treat this disease, the clinical symptoms and survival of patients with PH have been only partially improved. ³ Thus, a deep understanding of the pathogenesis of PH and the development of effective drugs for the disease are urgently needed. The Warburg effect induced under hypoxic conditions is the main pathological mechanism leading to the occurrence of PH, which can promote cell proliferation, apoptosis resistance, migration, endothelial‐mesenchymal transition, and other pathological phenotypes in pulmonary artery smooth muscle cells (PASMCs). ⁴ , ⁵ Nevertheless, the underlying mechanisms of these phenotypes remain unknown.

Circular RNAs (circRNAs), as a class of noncoding RNAs, have a continuous, stable, and covalently closed ring structure, which is not easily degraded by nucleases and is crucial for many human diseases. ⁶ As in PH, dysregulation of circRNAs have been reported to take part in the development of PH. ⁷ , ⁸ Hsa_circ_0016070 is a newly discovered circRNA that is generated from ATP2B4 gene and located on chromosome 1. By University of California, Santa Cruz database analysis, the annotation for circ_0016070 depicts exons 1 to 20 of ATP2B4 (total 4384 bp). The existing literature showed that circ_0016070 was remarkably upregulated in PH and could apparently induce hypoxia‐stimulated proliferation and migration of PASMCs by targeting the miR‐223/ATR pathway. ⁹ , ¹⁰ However, the complex function and mechanism of circ_0016070 remain unknown.

MicroRNAs (miRNAs) are small noncoding RNAs (21–23 nucleotides) that can regulate gene post‐transcription level and play a critical role in various important biological processes. ¹¹ Accumulating evidence revealed that miRNAs are involved in PH progression. ¹² For instance, the level of miR‐21 was correlated with the severity of right ventricular dysfunction in patients with PH. ¹³ miR‐150 protected against pulmonary vascular remodeling and fibrosis in hypoxia‐induced PH. ¹⁴ Additionally, miR‐143‐5p promoted the phenotypic conversion, proliferation and migration of PASMCs under hypoxic conditions by targeting hypoxia inducible factor 1 subunit alpha (HIF‐1α). ¹⁵ Recently, a study showed that micro (miR)‐340‐5p acted as a tumor suppressor that inhibited glycolysis by targeting mitochondrial calcium channel protein to inhibit the proliferation and migration of breast cancer cells. ¹⁶ In addition, miR‐340‐5p was decreased in PH caused by acute pulmonary embolism and could suppress cell proliferation, migration and the inflammatory response. ¹⁷ By StarBase database predictions, a putative binding site was found between miR‐340‐5p and circ_0016070. However, whether miR‐340‐5p is correlated with circ_0016070 remains to be studied.

Herein, we explored the underlying network of circ_0016070 in a hypoxia‐induced PH model and found that circ_0016070 was increased in serum samples, hypoxia‐induced PASMCs and pulmonary arterial tissues of rats. Moreover, circ_0016070 stimulated the proliferation, migration, vascular remodeling, and vascular phenotype conversion but decreased cell apoptosis in a hypoxia‐induced PH model by interacting with the miR‐340‐5p/TCF4/β‐catenin/TWIST1 signaling pathway. Our results might shed light on novel therapeutic targets for treating PH.

METHODS

Experimental data, analytical methods, and study materials will not be made available to other researchers to reproduce the results or replicate the procedure. Other researchers can contact the corresponding authors for methodological questions.

Ethics Statement

The study was approved by the Ethics Committee of Linfen People's Hospital of Shanxi Medical University. We obtained written informed consents from all patients.

All the animal experimental protocols were permitted by the Animal Ethics Committee of Linfen People's Hospital of Shanxi Medical University.

Serum Sample Collection

A total of 32 patients suffering from PH and 32 healthy volunteers were recruited at Linfen People's Hospital of Shanxi Medical University from January 2018 to December 2019. The clinical characteristics of the patients with PH are shown in Table. All patients provided complete clinical data. Plasma from healthy volunteers and patients with PH was collected. The study was approved by the Ethics Committee of Linfen People's Hospital of Shanxi Medical University. We obtained written informed consent from all patients.

Table 1.

Clinical Characteristics of Patients With PH and Healthy Controls

	Control (n=32)	PH (n=32)
Sex, n (%)
Female sex	20 (62.5)	16 (50)
Male sex	12 (37.5)	16 (50)
Age range, y	11 (21–47)	14 (17–70)
CI, I/(min ㎡)	N/A	2.45±0.4
PVR, Wood U	N/A	13.7±6.9
mPAP, mm Hg	N/A	62.1±16.4
RAP, mm Hg	N/A	8.5±4.2
SvO2, %	N/A	64.1±10.5
NT‐proBNP, ng/L	N/A	1597±2437
6MWD, m		455.8±91.21
Medication, n (%)
ERA	N/A	13 (40.6)
PDE5 inhibitor	N/A	19 (59.4)
Epoprostenol	N/A	8 (25)

CI indicates cardiac index; ERA, endothelin receptor antagonist; mPAP, mean pulmonary artery pressure; N/A, data not available; NT‐proBNP, N‐terminal pro‐B type natriuretic peptide; PVR, pulmonary vascular resistance; RAP, right atrial pressure; SvO₂, mixed venous oxygen saturation; and PDE5, phosphodiesterase 5.

Cell Culture and Treatment

Human PASMCs were purchased from the American Type Culture Collection (ATCC, Manassas, VA). Cells were cultured in Dulbecco’s modified Eagle’s medium containing 90% high glucose, 10% fetal bovine serum (Gibco, NY), 100 U/mL penicillin and 100 mg/mL streptomycin. For cell treatment, PASMCs were cultured under hypoxic conditions with 1% O₂, 5% CO₂ and 94% N₂ at 37 °C for 1 day. For TCF4/β‐catenin complex antagonists, 10 μmol/L LF3 (Sigma, St. Louis, MO) was used to stimulate PASMCs for 24 hours.

Plasmid Construction and Cell Transfection

Small interfering RNA designed for circ_0016070 (si‐circ_0016070; sense: 5′‐CCUGUUUACACCUUCAUCUTT‐3; antisense: 5′‐AGAUGAAGGUGUAAACAGGTT‐3′) and its negative control (si‐NC) were synthesized by GenePharma (Shanghai, China). The overexpression vectors of circ_0016070, TCF4, and TWIST1 (pcDNA‐circ_0016070, pcDNA‐TCF4, pcDNA‐TWIST1) as well as the negative empty vector (pcDNA‐NC) were constructed by GenePharma. The miR‐340‐5p mimics (sense: 5′‐UUAUAAAGCAAUGAGACUGAUU‐3′, antise: 5′‐AAUCAGUCUCAUUGCUUUAUAA‐3′), miR‐340‐5p inhibitor (5′‐AAUCAGUCUCAUUGCUUUAUAA‐3′) and negative NCs including inhibitor NC: 5′‐CAGUACUUUUGUGUAGUACAA‐3′, mimics NC: (sense: 5′‐UUCUCCGAACGUGUCACGUTT‐3′, antise: 5′‐ACGUGACACGUUCGGAGAATT‐3′) were obtained from RiboBio (Guangzhou, China). Cell transfection of PASMCs was performed using Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA). After 48 hours, real‐time quantitative polymerase chain reaction (RT‐qPCR) was performed to determine the corresponding transfection efficiency.

RNA Extraction and RT‐qPCR

Total RNA was isolated using TRIzol reagent (Sigma). We synthesized cDNA by reverse transcription using a PrimeScript RT Reagent Kit (Takara, Kyoto, Japan) or a Mir‐X miRNA First Strand Synthesis Kit (Takara). The cDNA was quantified using a SYBR Green PCR Kit (Takara) on an ABI 7900 PCR system (Applied Biosystems, Foster City, CA). Primers were used in this study: circ_0016070 (F: 5′‐TGTAGCAGTTGCACCAGTCA‐3′, R: 5′‐CTAACTCGGGCCAGACTCTT‐3′); miR‐340‐5p(F: 5′‐GCGGTTATAAAGCAATGAGA‐3′, R: 5′‐GTGCGTGTCGTGGAGTCG‐3′); TCF4 (F: 5′‐CAAAAACCAGAGCCAGGTGC‐3′, R: 5′‐GGAGCATAGACTGAA GATGGCA‐3′); TWIST1 (F: 5′‐GTCCGCAGTCTTACGAGGAG‐3′, R: 5′‐GCTTGAGGGT CTGAATCTTGCT‐3′); GAPDH (F: 5′‐TGACCTCAACTACATGGTCTACA‐3′, R: 5′‐CTTCCCATTCTCGGCCTTG‐3′); U6 (F: 5′‐CTCGCTTCG GCAGCACA‐3′, R: 5′‐AACGCTTCACGAATTTGCGT‐3′). GAPDH or U6 was used as an endogenous control. All gene levels were calculated using the standard 2^−∆∆Ct method.

Cell Proliferation Assay

Cell proliferation ability was tested using the Cell counting Kit‐8 assay (Dojindo Co. Ltd., Dojindo, Japan). In 96‐well plates, 2×10³ cells/well PASMCs were seeded, and 10 μL Cell Counting Kit‐8 reagent was added to each well at the indicated time points. After incubation at 37 °C for 2 hours, the absorbance at 450 nm was determined by using a microplate reader (Molecular Device, San Jose, CA).

Flow Cytometry Assay

PASMCs were subjected to examine cell apoptosis using an Annexin V‐Fluorescein Isothiocyanat (FITC)/Propidium iodide (PI) apoptosis kit (Invitrogen). Subsequently, the fluorescence intensity of Annexin V‐FITC and PI was assessed with a flow cytometer (BD Biosciences, Franklin Lakes, NJ). Flow Jo 10.0.7 software was used to assess the apoptosis rate.

Migration Assay

PASMCs were cultured in 6‐well plates. For the cell migration assay, the cell monolayer was scratched across the center of the plates using a sterile pipette tip when the cells were grown to 90% confluence. Cells were then cultured for 24 hours. Cell photographs were captured at 0 and 24 hours after scratching using a microscope (Olympus, Tokyo, Japan). The gap distance was analyzed by Image J software.

Western Blot Assay

PASMCs or PH tissues were collected and lysed with RIPA lysis buffer to extract total proteins. The protein concentration was quantified using the bicinchoninic acid method. Furthermore, the protein samples were separated by 10% SDS–PAGE and transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, MA). After blocking with 5% nonfat milk in tris‐buffered saline with Tween 20 for 1 hour, the membranes were incubated with primary antibodies at 4 °C overnight: Ki67 (1:2000, Abcam, Cambridge, MA), calponin (1:2000, Abcam), SM22α (1:2000, Abcam), osteopontin (1:1000, Abcam), epiregulin (1:1000, Abcam), TCF4 (1:1000, Santa Cruz Biotechnology, CA), and TWIST1 (1:500, Abcam). Then, the membranes were incubated with a horseradish peroxidase conjugated secondary antibody (1:5000, Abcam) for 1 hour at room temperature. Then, the bands were detected with enhanced chemiluminescence reagents (Pharmacia, Piscataway, NJ). Finally, the optical density of the band was quantified using ImageJ. GAPDH was used as an internal control.

RNA Immunoprecipitation Assay

A Magna RIP RNA‐Binding Protein Immunoprecipitation Kit (Millipore) was purchased for examining the interaction relationship between miR‐340‐5p and circ_0016070. Briefly, cells were lysed in RNA immunoprecipitation lysis buffer and incubated with magnetic beads conjugated with anti‐Ago2 antibody (Millipore) or negative control anti‐IgG antibody (Millipore). The immunoprecipitated RNA was then extracted, and RT‐qPCR was used to detect the enrichment of coprecipitated RNA.

Dual Luciferase Reporter Assay

The binding bases “CUUUAUA” of miR‐340‐5p on circ_0016070 or TCF4 sequences was mutated to “GAAAUAU” using a point mutation kit (Agilent Technologies, Santa Clara, CA). The wild‐type (WT) or mutant‐type (MUT) binding site sequences of circ_0016070 (WT/MUT‐circ_0016070) or TCF4 (WT/MUT‐TCF4) were inserted into pSI‐check2 vectors (Promega, Madison, WI), respectively. The miR‐340‐5p mimics or mimics NC was then cotransfected with the WT‐/MUT‐ constructs into PASMCs using Lipofectamine 2000 (Invitrogen). In addition, the promoter sequences of F1 and F2 of TWIST1 were cloned into pSI‐check2 vectors. Then, the luciferase vectors and pcDNA‐TCF4 or pcDNA‐NC vectors were cotransfected into PASMCs by Lipofectamine 2000 (Invitrogen). Forty‐eight hours later, a dual luciferase reporter system (Promega) was used to assess luciferase activity.

Chromatin Immunoprecipitation Assay

PASMCs were collected for the detection of the binding relationship between TWIST1 and TCF4 or β‐catenin using a SimpleChIP Enzymatic Chromatin IP Kit (Cell Signaling Technology). Briefly, DNA proteins were chemically crosslinked using 1% formaldehyde for 10 minutes at room temperature. Crosslinked cells were then lysed using lysis buffer with protease inhibitors. The lysates were incubated with Protein A/G magnetic beads (Thermo Fisher Scientific, Waltham, MA) with control immunoglobulin G or antibodies against TCF4 or β‐catenin overnight at 4 °C. The precipitated DNA fragments were quantified using polymerase chain reaction.

Establishment of the PH Rat Model

A total of 20 male Wistar rats (9 weeks old, 180–220 g) were obtained from Shanghai Laboratory Animal Center of the Chinese Academy of Sciences (Shanghai, China). All rats were housed under pathogen‐free conditions and randomly divided into 4 groups: sham, PH, PH+siNC, and PH+si‐circ_0016070 (n=5/group). The synthesized small interfering RNA sequence was inserted into the lentiviral vector pLKO.1 to construct a lentiviral vector carrying circ_0016070 small interfering RNA. The dose of lentivirus was 1×10¹⁰ TU/mL, once a week for 2 weeks. The PH rat model were established by using hypoxia or monocrotaline treatment. For the hypoxia experiment, rats in the sham group were exposed to 20% O₂, and PH rats were exposed to 10% O₂ for 21 days. ¹⁸ , ¹⁹ For the monocrotaline experiment, monocrotaline (60 mg/kg) was injected intraperitoneally to establish the rat model of PH. Meanwhile, rats injected with normal saline were considered controls. The rats were then housed for 3 weeks. ²⁰ After anaesthetization with urethane (1 g/kg), the rats were sacrificed by cervical dislocation. Then, the rat pulmonary arterial tissues were collected and stored at −80 °C for further detections. All animal experimental protocols were approved by the Animal Ethics Committee of Linfen People's Hospital of Shanxi Medical University.

Hemodynamic Analysis

The mean pulmonary arterial pressure of rats was evaluated after the experiments. Rats were anesthetized with isoflurane inhalation. A heparinized pressure volume (PV)‐1 catheter was inserted into the pulmonary artery via the external jugular vein, right atrium, and right ventricle, and the mean pulmonary arterial pressure was measured. The fixed hearts were dissected, and the weight of the heart tissues was recorded. The right heart hypertrophy index was determined by the weight ratio of the right ventricle (RV) to the left ventricle (LV) plus the interventricular septum (IVS) (RV/[LV+IVS]).

In Situ Hybridization

In situ hybridization staining was performed using an in situ hybridization kit (Boster, China). PH tissue sections were deparaffinized, dehydrated, and pepsin‐digested. Samples were incubated with circ_0016070‐specific probe overnight at 50 °C and further incubated with biotinylated antidigoxin (Boster, China), stained with biotinylated peroxidase, and developed with 3,3′‐diaminobenzidine substrate. Then, the sections were observed by using a light microscope.

TUNEL Assay

The tissue sections were fixed with 4% paraformaldehyde for 30 minutes at room temperature. Apoptotic cells were then determined by a One Step TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) Apoptosis Assay Kit (Beyotime, Jiangsu, China) according to the manufacturer's instructions. Coverslips were incubated with DAPI (Beyotime). Finally, positive cells were imaged under a microscope (Olympus).

Hematoxylin and Eosin Staining

Samples were fixed in 10% paraformaldehyde (Sigma) and dehydrated in an ascending series of ethanol and embedded in paraffin. After slicing into sections (5 μm‐thick), the sections were dewaxed in xylene and rehydrated in a descending series of alcohol. Then, the sections were stained with hematoxylin for 5 minutes and eosin solution for 3 minutes, followed by dehydration with graded ethanol and clearing in xylene. The tissue images were captured with a microscope (Olympus).

Immunohistochemistry

Paraffin‐embedded sections of pulmonary arterial tissues acquired from rats were deparaffinized and rehydrated. After blocking in 5% bovine serum albumin for 30 minutes, the sections were incubated with osteopontin antibody (1:1500, Abcam) or Ki67 antibody (1:1000, Abcam) at 4 °C overnight. The sections were then washed with PBS 3 times and incubated with horseradish peroxidase‐labeled secondary antibody for 1 hour at room temperature. Images were taken by a light microscope (Olympus).

Statistical Analysis

Data were analyzed by GraphPad Prism 5 software (GraphPad, San Diego, CA). The mean±SD was used for data presentation. All experiments were repeated at least 3 times. The difference between 2 groups was assessed by paired /unpaired Student t test. One‐way ANOVA was performed for multiple group comparisons. The clinical correlation of gene expression was analyzed by Pearson correlation analysis. A P value of <0.05 was considered statistically significant.

RESULTS

Inhibition of circ_0016070 Inhibited the Proliferation, Migration, and Phenotype Switch but Enhanced Apoptosis in Hypoxia‐Treated PASMCs

We first detected the expression of circ_0016070 in the serum of patients with clinical PH and hypoxia‐induced PASMCs. The data revealed that circ_0016070 was highly expressed in patients with PH compared with healthy people (P<0.001; Figure 1A). In hypoxia‐treated PASMCs, circ_0016070 was increased in a time‐dependent manner within 48 hours (0 hours versus 6 hours, P<0.05; 0 hours versus 12 hours, P<0.001; 0 hours versus 24 hours, P<0.05; 0 hours versus 48 hours, P<0.01; Figure 1B). To examine the biological roles of circ_0016070, si‐circ_0016070 was transfected into PASMCs, which were then exposed to hypoxic conditions. As shown in Figure 1C, circ_0016070 silencing obviously suppressed the elevated level of circ_0016070 induced by hypoxia (P<0.001). Compared with the normoxia group, the excessive proliferation and migration of PASMCs triggered by hypoxia were greatly diminished by knockdown of circ_0016070 (P<0.001 and P<0.01; Figure 1D and 1F). Similarly, the inhibitory effect on cell apoptosis caused by hypoxia stimulation was also significantly restrained by circ_0016070 knockdown (P<0.001; Figure 1E). Additionally, the results from Western blot assay revealed that the promoting roles on the protein levels of Ki67, osteopontin, and epiregulin and the suppressive effects on calponin and SM22α caused by hypoxia stimulation were greatly abolished by knockdown of circ_0016070 (Ki67, P<0.01; calponin, P<0.001; SM22α, P<0.001; osteopontin, P<0.001; epiregulin, P<0.001; Figure 1G). Our data indicated that circ_0016070 was upregulated in PH, and silencing of circ_0016070 ameliorated the proliferation, migration, and phenotype switch but enhanced cell apoptosis in hypoxia‐induced PASMCs.

Figure 1. Inhibition of circ_0016070 inhibited proliferation, migration, and phenotype switch but enhanced apoptosis in hypoxia‐treated pulmonary artery smooth muscle cells.

A and B, Real‐time quantitative polymerase chain reaction analysis was used to determine the level of circ_0016070 in the serum of patients with pulmonary hypertension (A) and hypoxia‐induced pulmonary artery smooth muscle cells (B). C, si‐circ_0016070 or si‐NC was transfected into pulmonary artery smooth muscle cells and then exposed to hypoxic conditions. The transfection efficiency was measured by RT‐qPCR. D, A Cell Counting Kit‐8 assay was used to assess cell proliferation. E, Flow cytometry was performed to detect cell apoptosis. F, Wound healing analysis measured cell migration ability. G, Western blot analysis was performed to detect the levels of Ki67, calponin, SM22α, osteopontin, and epiregulin. Data are the means±SD from 3 independent experiments. *P<0.05, **P<0.01, ***P<0.001. PH indicates pulmonary hypertension; si‐NC, small interfering RNA targeting negative control; and inhibitor NC, inhibitor targeting negative control.

Knockdown of circ_0016070 Alleviated PH Progression In Vivo

To confirm the biological roles of circ_0016070 in vivo, the hypoxia‐ or monocrotaline ‐induced PH rat model was established. Compared with the sham group, the mean pulmonary arterial pressure and right heart hypertrophy index in the PH group were increased significantly, while circ_0016070 silencing reduced the mean pulmonary arterial pressure and right heart hypertrophy index (P<0.01 and P<0.001, Figure 2A and 2B; P<0.001 and P<0.001, Figure S1A and S1B). The results of in situ hybridization staining and RT‐qPCR showed that the expression of circ_0016070 was increased in PH tissues, while the expression of circ_0016070 was decreased greatly after transfection with si‐circ_0016070 (P<0.001, Figure 2C and 2D; P<0.001, Figure S1C). By hematoxylin and eosin staining, we found that the blood vessel wall of the PH rats was greatly thickened, and the ratio of vessel wall area/total blood vessel area increased compared with the sham group, whereas these effects were markedly diminished by circ_0016070 silencing (P<0.01, Figure 2E; P<0.001, Figure S1D). Moreover, immunohistochemical assay showed that the increased expressions of Ki67 and osteopontin in PH rats were significantly reversed by circ_0016070 downregulation (both P<0.001, Figure 2F; both P<0.001, Figure S1E). Moreover, TUNEL assay revealed that the TUNEL‐positive cell number was remarkably reduced in hypoxia‐stimulated PH rats, but it was further increased after circ_0016070 silencing (P<0.001, Figure 2G; P<0.001, Figure S1F). Similar trends were observed that inhibition of circ_0016070 markedly impeded the functional roles of hypoxia in the regulation of Ki67 and vascular phenotypic switch‐associated markers (Figure 2H: Ki67, P<0.001; calponin, P<0.001; SM22α, P<0.001; osteopontin, P<0.001; epiregulin, P<0.001; Figure S1G: Ki67, P<0.001; calponin, P<0.001; SM22α, P<0.001; osteopontin, P<0.001; epiregulin, P<0.001). Therefore, circ_0016070 silencing improved the phenotypic switch and vascular remodeling in PH in vivo.

Figure 2. Knockdown of circ_0016070 alleviated hypoxia‐induced pulmonary hypertension progression in vivo.

A and B, The values of mean pulmonary arterial pressure and right heart hypertrophy index were measured. C, In situ hybridization staining of circ_0016070 expression in pulmonary arterial tissues of rats. D, Real‐time quantitative polymerase chain reaction was used to test the expression of circ_0016070 in pulmonary arterial tissues of rats. E, Hematoxylin and eosin staining of blood vessels in the pulmonary arterial tissues of rats. F, Immunohistochemistry assessed the levels of osteopontin and Ki67 in pulmonary arterial tissues of rats. G, A TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) assay was conducted to detect cell apoptosis. H, The protein levels of Ki67, calponin, SM22α, osteopontin, and epiregulin in pulmonary arterial tissues of rats were detected by Western blot. Data are the means±SD from 5 independent experiments. *P<0.05, **P<0.01, ***P<0.001. mPAP indicates mean pulmonary artery pressure; PH, pulmonary hypertension; RVI, right heart hypertrophy index; and TUNEL, terminal deoxynucleotidyl transferase dUTP nick end labeling.

Circ_0016070 Acted as a Molecular Sponge of miR‐340‐5p

In the present section, we studied the molecular mechanism by which circ_0016070‐mediated the biological behaviors of PASMCs. miR‐340‐5p was obviously lowly expressed in the serum samples of patients with PH and hypoxia‐treated PASMCs (P<0.001 and P<0.05; Figure 3A and 3B). Interestingly, the expression of miR‐340‐5p showed a negative correlation with circ_0016070 (P<0.05; Figure 3C). By starBase software prediction, the putative binding site between circ_0016070 and miR‐340‐5p was displayed (Figure 3D). Next, the luciferase activity of PASMCs cotransfected with miR‐340‐5p mimic and WT‐circ_0016070 was notably reduced, while it was opposite in miR‐340‐5p inhibitor and WT‐circ_0016070 cotransfected cells. Meanwhile, both overexpression and knockdown of circ_0016070 had no obvious effects on MUT‐circ_0016070‐transfected PASMCs (P<0.001; Figure 3E). In addition, the expression levels of circ_0016070 and miR‐340‐5p were enriched in the RNA complex pulled down by Ago2 (P<0.01 and P<0.001; Figure 3F). Compared with the control group, si‐circ_0016070 dramatically reduced the expression of circ_0016070 but elevated miR‐340‐5p expression; however, pcDNA‐circ_0016070 increased circ_0016070 expression but decreased miR‐340‐5p expression (both P<0.001, Figure 3G and 3H). Additionally, in vivo animal experiments showed that miR‐340‐5p was decreased in PH rat tissues and further increased after knocking down circ_0016070 (P<0.001; Figure 3I). Therefore, the above results demonstrated that miR‐340‐5p is a target of circ_0016070.

Figure 3. Circ_0016070 acted as a molecular sponge of miR‐340‐5p.

A and B, Real‐time quantitative polymerase chain reaction was conducted to examine the expression of miR‐340‐5p in the serum of patients with pulmonary hypertension (A) and hypoxia‐induced pulmonary artery smooth muscle cells (B). C, Pearson analysis showed a negative correlation between miR‐340‐5p and circ_0016070 in patients with pulmonary hypertension. D, Binding sequences between circ_0016070 and miR‐340‐5p were predicted by StarBase (http://starbase.sysu.edu.cn/). E, Luciferase activity of wild‐type‐circ_0016070‐ or mutant type‐circ_0016070‐transfected pulmonary artery smooth muscle cells was measured by dual reporter assay after treatment with miR‐340‐5p mimics or inhibitor. F, RNA immunoprecipitation assay verified the direct binding relationship between circ_0016070 and miR‐340‐5p. G and H, pulmonary artery smooth muscle cells were transfected with si‐circ_0016070 or pcDNA‐circ_0016070, then the levels of circ_0016070 (G) and miR‐340‐5p (H) were tested by real‐time quantitative polymerase chain reaction. I, Real‐time quantitative polymerase chain reaction analysis was performed to measure the level of miR‐340‐5p in pulmonary arterial tissues of rats. Data are the means±SD from at least 3 independent experiments. *P<0.05, **P<0.01, ***P<0.001. IgG indicates immunoglobulin G; MUT, mutant type; PH, pulmonary hypertension; RIP, RNA immunoprecipitation; and WT, wild‐type.

miR‐340‐5p was Negatively Involved in the Regulation of circ_0016070 in Hypoxia‐Stimulated PASMCs

We next assessed the effects of miR‐340‐5p on circ_0016070‐mediated biological functions. First, the hypoxia‐induced the decrease of miR‐340‐5p was restored by circ_0016070 silencing, but this effect was further abolished by the miR‐340‐5p inhibitor (P<0.001; Figure 4A). Furthermore, miR‐340‐5p downregulation greatly reversed the inhibitory effects of circ_0016070 silencing on proliferation and migration as well as the promoting effect on cell apoptosis in hypoxia‐stimulated PASMCs (P<0.001, P<0.001, and P<0.001; Figure 4B through 4D). In addition, miR‐340‐5p downregulation increased the protein levels of Ki67, osteopontin, and epiregulin while reducing the protein levels of calponin and SM22α in circ_0016070‐silenced hypoxic‐induced PASMCs (Ki67, P<0.001; calponin, P<0.001; SM22α, P<0.001; osteopontin, P<0.001; epiregulin, P<0.001; Figure 4E). Collectively, these results implied that miR‐340‐5p downregulation significantly reversed the effects of circ_0016070 silencing.

Figure 4. MiR‐340‐5p negatively affected the biological roles of circ_0016070 in hypoxia‐stimulated pulmonary artery smooth muscle cells.

Pulmonary artery smooth muscle cells were divided into 5 groups: normoxia, hypoxia, hypoxia+siNC+inhibitor NC, hypoxia+si‐circ_0016070+inhibitor NC, and hypoxia+si‐circ_0016070+miR‐340‐5p inhibitor. A, The level of miR‐340‐5p was evaluated by real‐time quantitative polymerase chain reaction assay. B, Cell Counting Kit‐8 analysis of cell proliferation. C, Flow cytometry detected cell apoptosis. D, Wound healing analysis measured cell migration ability. E, The protein levels of Ki67, calponin, SM22α, osteopontin, and epiregulin were detected by Western blot. Data are the means±SD from 3 independent experiments. *P<0.05, **P<0.01, ***P<0.001. PI indicates propidium iodide; and OD, optical density.

TCF4 was the Downstream Transcript of circ_0016070/miR‐340‐5p Axis

TCF4 in the serum of patients with PH was found to be remarkably upregulated compared with the healthy people (P<0.001; Figure 5A). Pearson analysis showed a negative association between the levels of TCF4 and miR‐340‐5p in PH (P<0.01; Figure 5B). Importantly, StarBase analysis revealed that there were 2 potential binding sites of miR‐340‐5p in the 3'‐untranslated region (3′‐UTR) of TCF4 (Figure 5C). A dual luciferase assay revealed that overexpression and inhibition of miR‐340‐5p significantly reduced and enhanced the fluorescence activity of PASMCs containing WT‐TCF4 vectors (site 1 and site 2), respectively, whereas no alteration was observed in PASMCs containing MUT‐TCF4 vectors (site 1 and site 2) when miR‐340‐5p was overexpressed or suppressed (both P<0.001; Figure 5D). In addition, RNA immunoprecipitation immunoprecipitation experiments showed that TCF4 mRNA and miR‐340‐5p were simultaneously enriched in the anti‐Ago2 complex (both P<0.001; Figure 5E). Subsequent results showed that miR‐340‐5p upregulation markedly decreased TCF4 mRNA level, and repression of miR‐340‐5p dramatically increased TCF4 mRNA level (P<0.001; Figure 5F). Furthermore, silencing of circ_0016070 greatly suppressed the mRNA and protein level of TCF4, whereas miR‐340‐5p downregulation restored TCF4 expression (both P<0.01; Figure 5G and 5H). Likewise, a positive correlation was displayed between the expressions of TCF4 and circ_0016070 in patients with PH (P<0.001; Figure 5I). The above research results show that TCF4 is the downstream target of the circ_0016070/miR‐340‐5p axis.

Figure 5. TCF4 (transcription factor 4) was the downstream target of circ_0016070/miR‐340‐5p axis.

A, Real‐time quantitative polymerase chain reaction assay was used to examine the expression of TCF4 in the serum of patients with pulmonary hypertension. B, Pearson analysis showed a negative correlation between miR‐340‐5p and TCF4 in patients with pulmonary hypertension. C, The binding sequences between TCF4 and miR‐340‐5p were predicted by starBase (http://starbase.sysu.edu.cn/). D, Luciferase activity of wild‐type‐TCF4– and mutant type‐TCF4–transfected pulmonary artery smooth muscle cells was quantified by dual reporter assay after miR‐340‐5p mimics or inhibitor were cotransfected. E, RNA immunoprecipitation assay verified the direct interaction between TCF4 and miR‐340‐5p. F, Pulmonary artery smooth muscle cells were treated with miR‐340‐5p mimics or inhibitor, as well as their corresponding negative controls, respectively. Then, the levels of TCF4 was tested by real‐time quantitative polymerase chain reaction. G and H, Pulmonary artery smooth muscle cells were transfected with si‐circ_0016070 alone or cotransfected with si‐circ_016070 and miR‐340‐5p inhibitor, and then the expression of TCF4 was tested by real‐time quantitative polymerase chain reaction (G) or Western blot (H). I, Pearson analysis presented a positive correlation between circ_0016070 and TCF4 in patients with pulmonary hypertension. Data are the means±SD for 3 independent experiments. *P<0.05, **P<0.01, ***P<0.001. IgG indicates immunoglobulin G; MUT, mutant type; PH, pulmonary hypertension; RIP, RNA immunoprecipitation; TCF4, transcription factor 4; and WT, wild‐type.

Overexpression of TCF4 Diminished the Roles of circ_0016070 Downregulation in Hypoxia‐Induced PASMCs

In this part, we explored the effect of TCF4 overexpression on cell proliferation, apoptosis and migration mediated by circ_0016070 silencing in hypoxia‐treated PASMCs. RT‐qPCR revealed that circ_0016070 downregulation reduced hypoxia‐triggered TCF4 expression, and this effect was abolished by pcDNA‐TCF4 (P<0.001; Figure 6A). Moreover, overexpression of TCF4 significantly promoted the proliferation, migration and phenotype switch of hypoxia‐induced PASMCs and suppressed cell apoptosis, which greatly reversed circ_0016070 silencing‐mediated effects (P<0.001, P<0.001, and P<0.001; Figure 6B through 6D). Moreover, in circ_0016070‐silenced hypoxic PASMCs, overexpression of TCF4 elevated the levels of Ki67, osteopontin, and epiregulin but reduced calponin and SM22α expressions, which remarkably reversed the effect of circ_0016070 silencing (Ki67 P<0.001, calponin P<0.001; SM22α, P<0.001; osteopontin, P<0.001; epiregulin, P<0.001; Figure 6E). The above results illustrated that overexpression of TCF4 diminished the roles of circ_0016070 downregulation in hypoxia‐induced PASMCs.

Figure 6. Overexpression of TCF4 (transcription factor 4) diminished the roles of circ_0016070 downregulation in hypoxia‐induced pulmonary artery smooth muscle cells.

A, Pulmonary artery smooth muscle cells were transfected with si‐circ_0016070 and pcDNA‐TCF4 and then exposed to hypoxic conditions. The expression of TCF4 was detected by real‐time quantitative polymerase chain reaction. B, Cell Counting Kit‐8 assay was used to test cell proliferation. C, Flow cytometry detected cell apoptosis. D, Wound healing analysis measured cell migration ability. E, Western blot was conducted to quantify the protein levels of Ki67, calponin, SM22α, osteopontin, and epiregulin. Data are the means±SD from 3 independent experiments. *P<0.05, **P<0.01, ***P<0.001. TCF4 indicates transcription factor 4; PI indicates propidium iodide; and OD, optical density.

TCF4/β‐Catenin Transcriptional Complex Promoted the Transcription of TWIST1

TCF4 is one of the main regulatory transcription factors in the Wnt pathway, ²¹ and it frequently forms a transcription complex with β‐catenin to regulate downstream gene expression. ²² By previous bioinformatics analysis (Jaspar database), 2 highly binding TCF4 transcriptional recognition sites (site 1 and site 2) were found on the promoter sequences of TWIST1 (Figure 7A). RT‐qPCR assay showed that transfection of pcDNA‐TCF4 vectors enhanced TCF4 expression (P<0.01; Figure 7B). A dual luciferase reporter assay showed that TCF4 upregulation enhanced the luciferase activity of PASMCs containing TWIST1 promoters (F1, F2) (P<0.01, P<0.01; Figure 7C). In addition, chromatin immunoprecipitation assay also showed that compared with immunoglobulin G, the contents of F1 and F2 fragments were enriched in anti‐TCF4 and anti‐β‐catenin (P<0.01, P<0.01; Figure 7D). LF3 is an antagonist that blocks the interaction between TCF4 and β‐catenin. ²³ , ²⁴ As shown in Figure 7E and 7F, overexpression of TCF4 markedly elevated the mRNA and protein expressions of TWIST1, while treatment with LF3 partly abolished the impact of TCF4 overexpression on TWIST1 (TCF4 mRNA, P<0.001; TCF4 protein, P<0.01, TWIST1 protein, P<0.001), suggesting that the transcriptional activation of TWIST1 by TCF4 was related to the interaction of β‐catenin. In addition, clinical detection demonstrated that TWIST1 was overexpressed in patients with PH and positively correlated with TCF4 level (P<0.001 and P<0.001; Figure 7G and 7H). Thus, our results showed that TCF4/β‐catenin complex activated TWIST1 transcription.

Figure 7. TCF4 (transcription factor 4)/β‐catenin transcriptional complex promoted the transcription of TWIST1 (Twist family bHLH transcription factor 1).

A, JASPAR analysis (http://jaspar.genereg.net/) displayed 2 recognition sites of TCF4 on the promoter sequence of TWIST1. B, TCF4 expression in pulmonary artery smooth muscle cells after transfection with pcDNA‐TCF4 was measured by real‐time quantitative polymerase chain reaction analysis. C, A dual luciferase assay was performed to examine the targeting relationship between TWIST1 promoter sequences (F1 and F2) and TCF4. D, Chromatin immunoprecipitation assay measured RNA fragments containing F1 and F2 in the anti‐TCF4 and anti‐β‐catenin pull‐down complex. E and F, Pulmonary artery smooth muscle cells were transfected with pcDNA‐TCF4 vectors with or without LF3 (10 μmol/L) cotreatment. Then, the level of TWIST1 was tested by using real‐time quantitative polymerase chain reaction (E) or Western blot (F). G, The expression of TWIST1 in the serum of patients with pulmonary hypertension was analyzed by real‐time quantitative polymerase chain reaction. H, Pearson analysis displayed a positive correlation between TWIST1 and TCF4 in patients with pulmonary hypertension. Data are the means±SD for at least 3 independent experiments. *P<0.05, **P<0.01, ***P<0.001. IgG indicates immunoglobulin G; PH, pulmonary hypertension; TCF4, transcription factor 4; and TWIST1, Twist family bHLH transcription factor 1.

Overexpression of TWIST1 Reversed the Biological Effects of circ_0016070 Silencing in Hypoxia‐Induced PASMCs

Here, we explored the functional correlation between TWIST1 and circ_0016070. RT‐qPCR data revealed that knockdown of circ_0016070 reduced the expression of TWIST1, while cotransfection of pcDNA‐TCF4 diminished these effects in PASMCs (TCF4 mRNA, P<0.01; TCF4 protein, P<0.001; TWIST1 protein, P<0.01; Figure 8A and 8B). Furthermore, silencing circ_0016070 decreased hypoxia‐triggered TWIST1 expression, and this effect was also abolished by pcDNA‐TWIST1 (P<0.001; Figure 8C). Functional assays showed that overexpression of TWIST1 dramatically impeded the repressive impacts on cell proliferation and migration and the facilitating effect on cell apoptosis mediated by knockdown of circ_0016070 in hypoxia‐treated PASMCs (P<0.001, P<0.001, and P<0.001; Figure 8D through 8F). Additionally, in circ_0016070‐silenced hypoxic PASMCs, overexpression of TWIST1 promoted the levels of Ki67, osteopontin, and epiregulin, but decreased the levels of calponin and SM22α, which remarkably restrained the effects of circ_0016070 downregulation (Ki67, P<0.001; calponin, P<0.001; SM22α, P<0.001; osteopontin, P<0.001; epiregulin, P<0.001; Figure 8G). The above results illustrated that overexpression of TWIST1 reversed the effects of circ_0016070 silencing on the proliferation‐apoptotic imbalance, migration, and phenotypic transition of hypoxia‐induced PASMCs.

Figure 8. Upregulation of TWIST1 (Twist family bHLH transcription factor 1) reversed the biological effects of circ_0016070 silencing in hypoxia‐induced pulmonary artery smooth muscle cells.

A and B, Pulmonary artery smooth muscle cells were cotransfected with si‐circ_0016070 and pcDNA‐TCF4 (transcription factor 4). Then, the mRNA of TWIST1 was assessed by real‐time quantitative polymerase chain reaction (A), and the protein levels of TCF4 and TWIST1 were quantified using Western blot (B). C, Pulmonary artery smooth muscle cells were transfected with si‐circ_0016070 or cotransfected with si‐circ_0016070 and pcDNA‐TWIST1, and then pulmonary artery smooth muscle cells were exposed to hypoxic conditions. The expression of TWIST1 was detected by real‐time quantitative polymerase chain reaction. D, Cell Counting Kit‐8 assay was used to test cell proliferation. E, Flow cytometry detected cell apoptosis. F, Wound healing analysis measured cell migration ability. G, Western blot was conducted to quantify the protein levels of Ki67, calponin, SM22α, osteopontin, and epiregulin. Data are the means±SD from 3 independent experiments. *P<0.05, **P<0.01, ***P<0.001. TCF4 indicates transcription factor 4; and TWIST1, Twist family bHLH transcription factor 1; PI indicates propidium iodide; and OD, optical density.

DISCUSSION

PH is a chronic, progressive disease caused by various cardiac or pulmonary diseases and characterized by pulmonary vascular remodeling. ² The dysfunction of PASMCs has been reported to be a critical cause in the occurrence and development of PH. ²⁵ Studies have shown that the abnormal cellular activities of PASMCs are affected mainly by a variety of signaling pathways, noncoding RNAs, miRNAs and the external environment. ²⁶ , ²⁷ circRNAs are considered as a crucial modulator of PH development. Recently, researchers have discovered that circ_0016070 was associated with vascular remodeling in PH by promoting the proliferation of PASMCs via miR‐942/CCND1. ⁹ Exosomes or extracellular vehicles are membrane vesicles specifically secreted by cells that contain functionally active substances such as proteins, nucleic acids and lipids. ²⁸ Both exosomes and extracellular vehicles have been confirmed to participate in the transfer of intercellular substances and information and play a role in immunity and metabolism. ²⁹ Thus, we speculated that the transfer of circ_0016070 to the blood might be related to the specific vesicles or exosomes secreted by PASMCs. Consistently, our study found that circ_0016070 was upregulated in the serum of patients with PH, hypoxia‐induced PASMCs and the PH rat tissues model. Studies have shown that hypoxia stimulation could cause epigenetic modification changes in the genome, including DNA methylation, histone modification and m6A modification, ³⁰ , ³¹ , ³² as well as transcription regulatory factors such as HIF‐1α and p53. ³³ As is well known, the process of gene transcription is associated with the regulation of the chromatin state and transcription factor regulation. Therefore, we speculated that the mechanism of circ_0016070 upregulation under hypoxic stimulation might be related to epigenetic modification. In addition, we found that circ_0016070 promoted proliferation, migration, and vascular remodeling but decreased cell apoptosis in vitro and in vivo, suggesting that circ_0016070 served a key role in the pathogenesis of PH.

Accumulating data suggested that circRNAs act as miRNA sponges and participate in gene expression modulation. ³⁴ For instance, Yang et al reported that circRNA ABCB10 inhibited the occurrence and development of hepatocellular carcinoma by sponging miR‐340‐5p. ³⁵ circ_0016070 was identified as a sponge of miR‐340‐5p by directly interacting with miR‐340‐5p in PASMCs. Recent studies have explored the relationship between miRNAs and PH. ³⁶ For example, the downregulation of miR‐637 increased the risk of PH by regulating CDK6 in hypoxia‐treated PASMCs. ³⁷ Likewise, our data identified that the expression of miR‐340‐5p was downregulated in the serum of patients with PH and hypoxia‐induced PASMCs. Notably, miR‐340‐5p was a downstream gene of circ_0016070 and negatively affected the biological effects of circ_0016070 in hypoxia‐induced PASMCs, suggesting that miR‐340‐5p exerted a protective role in PH. Moreover, Ou et al also found that miR‐340‐5p suppressed the inflammatory response, proliferation, and migration of PASMCs by downregulating IL‐1β and IL‐6. ¹⁷ However, whether circ_0016070 could affect the inflammatory response of PASMCs through miR‐340‐5p remains unclear, and this is a question worth exploring in the future.

In general, miRNAs mainly affect the expression and transcription of target genes by binding to the 3‐UTR of target genes. TCF4 is a basic helix–loop–helix transcription factor that is widely expressed throughout mammalian tissue. Currently, several microRNAs have been found to target TCF4, including miR‐124, miR‐93‐5p, and miR‐137. ³⁸ , ³⁹ , ⁴⁰ In the present work, TCF4 was significantly upregulated in patients with PH and presented a negative and positive correlation with miR‐340‐5p and circ_0016070, respectively. Interestingly, our data uncovered a direct binding relationship between miR‐340‐5p and TCF4 and could form a competing endogenouse RNA network with cirrc_0016070, implying that TCF4 was a downstream molecule of the circ_0016070/miR‐340‐5p axis. Additionally, TCF4 often forms a transcription complex with β‐catenin coactivator and acts as the main regulatory transcription factor of the Wnt/β‐catenin pathway. ⁴¹ , ⁴² As Liu et al reported, TCF4 aggravated proliferation but reduced apoptosis in hepatocellular carcinoma cells by activating the Wnt/β‐catenin pathway. ⁴³ In addition, the TCF4/β‐catenin transcription complex also affected the transcription of a large number of genes, such as c‐Myc, ⁴⁴ Cyclin D1, ⁴⁵ and Bcl‐2. ⁴⁶ These downstream targeted genes were confirmed to be involved in the occurrence and development of PH. ⁴⁷ , ⁴⁸ , ⁴⁹ Moreover, LF3 (an inhibitor of the interaction between TCF4 and β‐catenin) inhibited the progression of PH. ⁵⁰ This evidence demonstrated the correlation between TCF4/β‐catenin and PH. Moreover, our results showed that overexpression of TCF4 significantly promoted the proliferation, migration, and phenotype switch of hypoxia‐induced PASMCs, and significantly reversed circ_0016070 silencing‐mediated effects. Thus, we concluded that the TCF4/β‐catenin transcriptional complex was involved in the roles of the circ_0016070/miR‐340‐5p axis in PH.

TWIST1 is a basic helix–loop–helix transcription factor that play an essential role in various cancers. ⁵¹ A previous study suggested that TWIST1 induced the growth and proliferation of dermal papilla cells by forming a complex with TCF4 and β‐catenin. ²² In this work, we validated the transcriptional effects of TCF4/β‐catenin complex on TWIST1. Notably, TWIST1 was reported to be closely linked to PH. For example, Mammoto et al reported that TWIST1 increased the expression of platelet‐derived growth factor in human pulmonary arterial endothelial cells and promoted αSMA‐positive cell accumulation, migration, and proliferation. ⁵² Mammoto et al showed that TWIST1 regulated hypoxia‐triggered PH through TGF‐β/Smad signaling. ⁵³ Fan et al demonstrated that TWIST1 induced the proliferation of PASMCs by activating GATA binding protein 6 (GATA6) and bone morphogenetic protein receptor type 2 (BMPR2). ⁵⁴ Consistently, we further found that overexpression of TWIST1 dramatically reversed the effects of circ_0016070 downregulation on the proliferation‐apoptotic imbalance, migration and vascular phenotype transition in hypoxia‐induced PASMCs, confirming that TWIST1 functioned as a functional target of the circ_0016070/miR‐340‐5p/TCF4 axis in PH.

In conclusion, our data elucidated a novel underlying network of circ_0016070 in PH pathomechanisms. Downregulation of circ_0016070 alleviated the excessive proliferation and migration and suppressed vascular remodeling but enhanced apoptosis triggered by hypoxia in PH by regulating the miR‐340‐5p/TCF4/β‐catenin/TWIST1 axis (Figure 9). Thus, our findings implied that inhibition of circ_0016070 might be a promising and novel therapeutic approach for PH treatment.

Figure 9. Schematic diagram revealing that circ_0016070 promoted TCF4 (transcription factor 4)/β‐catenin complex‐mediated transcription of TWIST1 (Twist family bHLH transcription factor 1) by inhibiting miR‐340‐5p, thereby suppressing apoptosis and promoting proliferation, migration, and phenotypic switch in pulmonary hypertension.

PH indicates pulmonary hypertension; TCF4, transcription factor 4; and TWIST1, Twist family bHLH transcription factor 1.

Sources of Funding

This work was supported by Scientific Research Fund of Coronary Heart Disease Center (GXB09).

Disclosures

None.

Supporting information

Figure S1