PVECs‐Derived Exosomal microRNAs Regulate PASMCs via FoxM1 Signaling in IUGR‐induced Pulmonary Hypertension

Abstract

Background

Intrauterine growth restriction (IUGR) is closely related to systemic or pulmonary hypertension (PH) in adulthood. Aberrant crosstalk between pulmonary vascular endothelial cells (PVECs) and pulmonary arterial smooth muscle cells (PASMCs) that is mediated by exosomes plays an essential role in the progression of PH. FoxM1 (Forkhead box M1) is a key transcription factor that governs many important biological processes.

Methods and Results

IUGR‐induced PH rat models were established. Transwell plates were used to coculture PVECs and PASMCs. Exosomes were isolated from PVEC‐derived medium, and a microRNA (miRNA) screening was proceeded to identify effects of IUGR on small RNAs enclosed within exosomes. Dual‐Luciferase assay was performed to validate the predicted binding sites of miRNAs on FoxM1 3’ untranslated region. FoxM1 inhibitor thiostrepton was used in IUGR‐induced PH rats. In this study, we found that FoxM1 expression was remarkably increased in IUGR‐induced PH, and PASMCs were regulated by PVECs through FoxM1 signaling in a non‐contact way. An miRNA screening showed that miR‐214‐3p, miR‐326‐3p, and miR‐125b‐2‐3p were downregulated in PVEC‐derived exosomes of the IUGR group, which were associated with overexpression of FoxM1 and more significant proliferation and migration of PASMCs. Dual‐Luciferase assay demonstrated that the 3 miRNAs directly targeted FoxM1 3’ untranslated region. FoxM1 inhibition blocked the PVECs‐PASMCs crosstalk and reversed the abnormal functions of PASMCs. In vivo, treatment with thiostrepton significantly reduced the severity of PH.

Conclusions

Transmission of exosomal miRNAs from PVECs regulated the functions of PASMCs via FoxM1 signaling, and FoxM1 may serve as a potential therapeutic target in IUGR‐induced PH.

Nonstandard Abbreviations and Acronyms

ECs

endothelial cells

EdU

5‐ethynyl‐2'‐deoxyuridine

FBS

fetal bovine serum

FoxM1

Forkhead box M1

IUGR

intrauterine growth restriction

miRNAs

microRNAs

mPAP

mean pulmonary arterial pressure

PASMCs

pulmonary arterial smooth muscle cells

PH

pulmonary hypertension

PVECs

pulmonary vascular endothelial cells

SMCs

smooth muscle cells

UTR

untranslated region

Clinical Perspective.

What Is New?

• We provided direct evidence that pulmonary vascular endothelial cells could regulate the proliferation and migration of pulmonary arterial smooth muscle cells through FoxM1 (Forkhead box M1) signaling in intrauterine growth restriction (IUGR)‐induced pulmonary hypertension.

• We uncovered that IUGR could change the microRNA cargos enclosed within exosomes to mediate the intercellular communication that is potentially relevant to pulmonary hypertension that occurs following IUGR.

What Are the Clinical Implications?

• MiR‐214‐3p, miR‐326‐3p, and miR‐125b‐2‐3p may be triggers for the induction of exaggerated pulmonary hypertension in IUGR and served as biomarkers for this disease.

• FoxM1 inhibition might provide a novel therapeutic approach to ameliorate the effects of IUGR on exaggerated pulmonary hypertension.

The “Developmental Origins of Health and Disease” suggests that harmful exposures during early life increase the risk of disease later in life. ¹ Intrauterine growth restriction (IUGR) is generally defined as an estimated fetal weight below the 10th percentile for gestational age and sex that may be caused by maternal, placental, fetal, or genetic factors. ² Despite its multiple origins, IUGR is associated with higher risks of cardiovascular diseases, including systemic or pulmonary hypertension (PH). ³ , ⁴ PH is a complex pathology characterized by abnormal vascular remodeling and a progressive rise in the pulmonary vascular load, leading to hypertrophy and remodeling of the right ventricle. ⁵ A large prospective single‐site study found that about 18% of extremely low‐ birth‐weight infants were likely to develop PH. ⁶ Another study found that a transient perinatal insult to the pulmonary circulation left a persistent and potentially fatal imprint that may evoke the development of PH in adult life. ⁷ However, the mechanism under PH that occurs following IUGR is poorly documented.

Pulmonary vascular endothelial cells (PVECs) and pulmonary arterial smooth muscle cells (PASMCs) are 2 key cell types that play major roles in the structural remodeling process of the pulmonary vasculature. Endothelial injury or dysfunction has been believed to be the initial event. ⁸ Smooth muscle cells (SMCs) are the predominant cells of vessels, whose hypertrophy, proliferation, and migration may result in obstruction of the vascular lumen, causing elevations in mean pulmonary arterial pressure (mPAP). ⁹ Studies on the pathogenesis of PH highlighted the intercellular communication between vascular cells, especially endothelial cells (ECs) and SMCs. Prior works demonstrated that ECs can regulate the functions of underlying SMCs through cocultured SMCs with EC‐derived medium. ¹⁰ , ¹¹ In addition to such paracrine regulation, other intercellular communication via the myoendothelial injunctions and extracellular vesicles also plays an important role in the pathobiology of PH. ¹² Because of the complex signalings involved in the process of PH, identification of the critical molecular mechanism mediating vascular cell hyperproliferation and migration is vital to develop strategies to potentially reverse the PH.

Among the mechanisms evaluated, FoxM1 (Forkhead box M1) is a key transcription factor of cell cycle progression that is required for G1/S and G2/M transition, as well as the M‐phase progression. ¹³ , ¹⁴ Studies found that FoxM1 is a central downstream effector on which various pro‐proliferative and inflammatory stimuli converge. ¹⁵ , ¹⁶ , ¹⁷ FoxM1 promoted pulmonary vascular development by directly stimulating proliferation of developing SMCs, and constitutive deletion of SMC FoxM1 in mice induced arterial wall defects and premature death immediately after birth. ¹⁵ However, whether the PH risk induced by IUGR is coupled with FoxM1, and the role of FoxM1 in the crosstalk between PVECs and PASMCs, remains undefined.

Exosomes are small vesicles (traditionally considered 50–150 nm in diameter) that can be secreted by almost all cell types. ¹⁸ Exosomes have been recognized as potent vehicles of intercellular communication that can influence physiological and pathological functions of the recipient cells. ¹⁹ MicroRNAs (miRNAs) are endogenous 21‐23nt small noncoding RNAs, which are proposed to regulate over one‐half of all human genes, finetuning a diverse array of biological processes. ²⁰ Secreted miRNAs enclosed within exosomes can be protected from degradation by ribonuclease, functioning as stable diagnostic biomarkers or therapeutic targets. ²¹ Various kinds of miRNAs were reported to specifically affect vascular cells and contribute to the pathogenesis of PH. ²² Therefore, it is crucial to recognize the major mechanism underlying miRNA functions in IUGR‐induced PH.

In the current study, we established an IUGR rat model of PH to verify the effect of adverse intrauterine environment in later life. We hypothesized that aberrant crosstalk between PVECs and PASMCs, which was mediated by exosomal miRNAs through FoxM1 signaling, played an essential role in the pathology of IUGR‐induced PH, and FoxM1 inhibition might provide an effective solution for such multifactorial disease. This study described a novel, microRNA‐dependent association between PVECs and PASMCs and suggested a novel mechanism to ameliorate the effects of IUGR on exaggerated PH.

METHODS

We have full access to all the data in the study and take responsibility for their integrity and the data analysis. The data that support the findings of this study are available from the corresponding author upon reasonable request.

Ethics Statement of Animal Studies

The animal care and study protocols were approved by the Institutional Animal Care and Use Committee of Zhejiang University. All experiments were performed in accordance with the National Institutes of Health guidelines. Sprague–Dawley Rats were used for experimental data with the minimum of suffering. Wild‐type rats were purchased from Zhejiang Chinese Medical University Laboratory Animal Center.

Rat Model Establishment

We used a rodent model of nutrient restriction during pregnancy to induce IUGR in offspring as described in our previous study. ²³ , ²⁴ Pups of food‐restricted mothers, weighing <10th percentile of the controls, were considered as the IUGR group. ² Only male pups were studied to avoid hormonal cycle disturbances. ²⁵ , ²⁶ Offspring rats at 9 weeks were exposed to hypoxia with 10% oxygen concentration for 3 weeks to induce PH. Then offspring rats were randomly allocated to receive the FoxM1 inhibitor thiostrepton (10 mg/kg, Aladdin, China) or its vehicle intraperitoneally every day for 3 weeks. ¹⁷

Primary PVECs and PASMCs Culture

Rat primary PVECs were harvested as described in our previous studies. ²⁴ PVECs were cultured in EC media supplemented with 5% fetal bovine serum (FBS), 1% EC growth supplement, and 1% antibiotics (all from Science Cell, USA). Heparin (yuanye, China) was added at 90 U/mL to inhibit the growth of fibroblasts and purify ECs. The cells were authenticated by specific cell marker CD31, and the purity was over 90% (Figure S1A). PVECs were used between passages 1 and 2.

Rat PASMCs were cultured from peripheral small pulmonary artery using an enzymatic dissociation method as our previous description. ²⁷ DMEM high glucose containing 15% FBS and 1% antibiotics (all from Gibco, USA) were used to culture PASMCs, and experiments were performed with PASMCs from passages 2 to 4. Cells were positively stained for specific α‐smooth muscle actin by immunofluorescence, and the purity was >90% (Figure S1B).

Coculture System of PVECs and PASMCs

A noncontact coculture system of PVECs and PASMCs was developed through a transwell system (Figure S1C). Briefly, PVECs were seeded on the upper sides of 0.4‐μm pore cell culture inserts (Corning, USA), while PASMCs were seeded on the matched culture plates. PVECs and PASMCs were cultured in their respective complete media overnight to fully attach, and then the inserts were put into the plates and maintained in the system in DMEM supplemented with 2% FBS for 48 hours at 37 °C. ²⁸ , ²⁹

Preparation of PVEC‐Derived Medium

1×10⁶ of primary PVECs in exponential growth period were cultured in a 15‐cm plate overnight to adhere. Then the media was discarded and cells washed with PBS to remove residual FBS. PVECs continued to culture for 48 hours with 5 mL of EC media without any supplement, such as FBS, EC growth supplement, antibiotics, and heparin. Then the culture media was pipetted and centrifuged at 2000g for 30 minutes. The supernatant was collected and used as PVEC‐derived medium.

Exosomal miRNAs Isolation, Sequencing, and Differential Expression Analysis

PVEC‐derived medium from 8 plates was pooled for exosomes isolation. The isolated particles were characterized by transmission electron microscope studies, nanoparticle tracking analysis, and Western blot analysis of exosome markers CD63, TSG101, and HSP70. The differential exosomal miRNA expression analysis was performed after sequencing on an Illumina Hiseq platform.

Statistical Analysis

Statistical analysis was performed with Student t‐test for comparisons between 2 groups. One‐way ANOVA was performed for comparison of >2 groups, followed by Dunnett test compared with their corresponding controls or Tukey test. Welch ANOVA followed by Games‐Howell multiple comparison was performed if the equal variance of data was not assumed. Two‐way ANOVA was performed for comparisons with 2 variables, followed by Bonferroni multiple comparison test to analysis the effects of IUGR or hypoxia. All statistical analyses were performed using Prism8.3.0 (La Jolla, CA) and SPSS 22.0 (SPSS, Inc). Data were represented as mean±SEM. Comparisons were considered statistically significant when the P value was <0.05. Duplicated experiments were performed independently to guarantee reproducibility of findings.

Detailed methods are provided in Data S1. Additionally, all research listed in the Methods is included in the Major Resources Table and provided as Table S1.

RESULTS

IUGR‐Induced PH Rat Models Were Established

IUGR‐induced PH rat models were designed as shown in Figure 1A. IUGR pups developed significantly lower birth weights than controls, with all pups' birth weights falling below the 10th percentile of the controls, which meant that the IUGR rat models had been successfully established (Figure 1B). Both IUGR and hypoxia increased the mPAP (Figure 1C), the right ventricular (RV) hypertrophy index (Figure 1D), the medial wall thickness of vascular (Figure 1E), and the muscularization of distal pulmonary vessels (Figure 1F), with the greatest augment in the IUGR group under the hypoxic condition.

Figure 1. Intrauterine growth restriction‐induced pulmonary hypertension rat models were established.

A, Intrauterine growth restriction‐induced pulmonary hypertension rat model design and treatment. B, Birth weight (g) of intrauterine growth restriction (n=42) and control (n=40) rats. The red dotted line indicates the 10th percentile of the controls' birth weight. C, Mean pulmonary arterial pressure of the rats (n=7 per group). D, Right ventricular hypertrophy index of the rats (=right ventricular weight/[left ventricular weight plus septum weight]) (n=7–8 per group). E, Representative images of lung tissue stained with α‐smooth muscle actin and medial wall area proportion of pulmonary arteriole, which is <100 μm in diameter. Medial wall area proportion=(outer area–inner area)/outer area (n=8 per group). Scale bar=50 μm. F, Representative images of α‐smooth muscle actin staining and muscularization of distal pulmonary vessels, which is <100 μm in diameter (n=8–9 per group). Scale bar=100 μm. All data were represented as the mean±SEM. Comparisons of birth weights were made using Student t‐test; comparisons in C through F were made using 2‐way ANOVA. IUGR indicates intrauterine growth restriction; and RVHI, right ventricular hypertrophy index. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, comparison between intrauterine growth restriction and control groups as indicated; ^# P<0.05, ^## P<0.01, ^#### P<0.0001, comparison with relative group under normoxia.

FoxM1 Expression Was Remarkably Increased in IUGR‐Induced PH

Immunofluorescence of FoxM1 in the lung showed that IUGR challenged with hypoxia expressed high level of FoxM1 in vascular tissues (Figure 2A). Although there was no significant difference in the expression of FoxM1 in mRNA level, the protein level of FoxM1 was greatly augmented in lung tissues of the IUGR combined hypoxia group (Figure 2B). We next tested the expression of FoxM1 in the PASMCs and found that IUGR exaggerated the expression of FoxM1 in PASMCs, with more apparent increase after exposure to hypoxia (Figure 2C). 5‐ethynyl‐2'‐deoxyuridine and transwell assay showed that both the proliferation and migration of IUGR PASMCs were significantly augmented and greater elevations were found in hypoxic condition (Figure 2D and 2E). Similar phenomena were found in PVECs, which showed the significant elevations in FoxM1 expression, proliferation, and migration of IUGR PVECs, with more apparent increases after exposure to hypoxia (Figure 2F through 2H).

Figure 2. FoxM1 expression was remarkably increased in intrauterine growth restriction‐induced pulmonary hypertension.

A, Representative immunofluorescence photomicrographs of α‐smooth muscle actin (green) and FoxM1 (red) in small pulmonary arteries from rat lungs. Nuclei were counterstained with DAPI (blue). B, Relative mRNA expression (n=6 per group) and Western blot analysis (n=3 per group) of FoxM1 in lung tissues. C, Relative mRNA expression (n=7–9 per group) and Western blot analysis (n=9 per group) of FoxM1 in pulmonary arterial smooth muscle cells. D, Proliferation of pulmonary arterial smooth muscle cells (n=8 per group). E, Migration of pulmonary arterial smooth muscle cells (n=7–8 per group). F, Relative mRNA expression (n=7–8 per group) and Western blot analysis (n=9 per group) of FoxM1 in pulmonary vascular endothelial cells. G, Proliferation of pulmonary vascular endothelial cells (n=7 per group). H, Migration of pulmonary vascular endothelial cells (n=5–6 per group). All data were represented as the mean±SEM. Scale bar=50 μm. Comparisons in B–E were made using 2‐way ANOVA. α‐SMA indicates α‐smooth muscle actin; DAPI, 4',6‐diamidino‐2‐phenylindole; EdU, 5‐ethynyl‐2'‐deoxyuridine; FoxM1, Forkhead box M1; IUGR, intrauterine growth restriction; PA, pulmonary artery; PASMCs, pulmonary arterial smooth muscle cells; and PVECs, pulmonary vascular endothelial cells. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, comparison between intrauterine growth restriction and control groups as indicated; ^# P<0.05, ^## P<0.01, ^### P<0.001, ^#### P<0.0001, comparison with relative group under normoxia.

FoxM1 Played an Essential Role in the Crosstalk Between PVECs and PASMCs

Cocultured with PVECs from the control group, the abnormal expression of FoxM1 from the IUGR or hypoxia group was reversed in PASMCs (Figure 3A). The increases in proliferation and migration were diminished as well (Figure 3B). While cocultured with PVECs from IUGR or hypoxia‐exposed rats, FoxM1 expression of control PASMCs was augmented in the protein level (Figure 3C), in conjunction with the proliferation and migration, especially under the hypoxic condition (Figure 3D).

Figure 3. FoxM1 played an essential role in the crosstalk between PVECs and PASMCs.

A, Relative mRNA expression (n=6 per group) and Western blot analysis (n=6 per group) of FoxM1 in PASMCs isolated from intrauterine growth restriction or hypoxia‐exposed rats after being cocultured with control PVECs. B, Proliferation (n=7–8 per group) and migration (n=8 per group) of PASMCs isolated from intrauterine growth restriction or hypoxia‐exposed rats after being cocultured with control PVECs. C, Relative mRNA expression (n=6 per group) and Western blot analysis (n=6 per group) of FoxM1 in control PASMCs after being cocultured with PVECs from intrauterine growth restriction or hypoxia‐exposed rats. D, Proliferation (n=7–8 per group) and migration (n=7–8 per group) of control PASMCs after being cocultured with PVECs from intrauterine growth restriction or hypoxia‐exposed rats. All data were represented as the mean±SEM. Scale bar=50 μm. Comparisons in A through D were made using 2‐way ANOVA. co‐CN indicates cocultured with control PASMCs; CN‐co, control PVECs cocultured with; EdU, 5‐ethynyl‐2'‐deoxyuridine; FoxM1, Forkhead box M1; IUGR, intrauterine growth restriction; PA, pulmonary artery; PASMCs, pulmonary arterial smooth muscle cells; and PVECs, pulmonary vascular endothelial cells. *P<0.05, **P<0.01, ***P<0.001, comparison as indicated; ^# P<0.05, ^## P<0.01, ^### P<0.001, ^#### P<0.0001, comparison with coculture systems of relative normoxia‐derived PVECs.

Exosomes Identification and Comparison of Exosomal miRNAs Signatures Between IUGR and Control Group

Both the proliferation and migration of control PASMCs were significantly exaggerated when PASMCs were cocultured with PVEC‐derived medium of IUGR. (Figure S2A and S2B). To address how PVECs communicated with PASMCs, exosomes were extracted from PVEC‐derived medium. The morphology of isolated exosomes was observed by transmission electron microscope, which showed typical cup shape and rounded lipid bilayer particles (Figure 4A). The size distribution of exosomes was evaluated by nanoparticle tracking analysis, which confirmed that most of the particles were ≈100 nm in diameter (Figure 4B). Detailed descriptions of the vesicles were shown in Table S2. Moreover, enrichment of the exosome markers CD63, TSG101, and HSP70 were all detected in the isolated exosomes enriched fraction samples by Western blot, while calnexin was absent (Figure 4C), indicating the successful isolation of purified exosomes from PVEC‐derived medium.

Figure 4. Exosomes identification and comparison of exosomal miRNAs signatures between the intrauterine growth restriction and control groups.

A, Representative transmission electron microscope picture of exosomes derived from pulmonary vascular endothelial cell‐derived medium. Scale bar=200 nm. B, The concentration and size distribution of the exosomes detected by nanoparticle tracking analysis. C, Western blot analysis for presence of exosomes markers, CD63, TSG101, and HSP70 and absence of calnexin. D, A heatmap of the expression level of 33 upregulated miRNAs and 26 downregulated miRNAs across all 6 samples in our exosomes enriched fractions miRNA data set. E, A volcano plot of the 59 differential expression genes. F, Kyoto Encyclopedia of Genes and Genomes pathway analysis of the differential exosomal miRNAs of intrauterine growth restriction. CD indicates clusters of differentiation; FC, fold change of control; HSP, heat shock protein; IUGR, intrauterine growth restriction; MAPK, mitogen‐activated protein kinase; miR, microRNA; PVECs, pulmonary vascular endothelial cells; and TSG, tumor susceptibility gene.

The illumina HiSeq high‐throughput technology was used to sequence 2 small RNA libraries, which were separated from exosomes of the IUGR or control group. Data showed that 33 miRNAs were upregulated, while 26 miRNAs were downregulated in the IUGR group compared with the control group (Figure 4E). The mean expression levels of differentially expressed miRNAs were complemented in Table S3. The heatmap showed the differential expression genes between groups in a more visualized way and presented that the expression of exosomal miR‐214‐3p, miR‐326‐3p, and 125b‐2‐3p from PVEC‐derived medium of IUGR was greatly decreased (Figure 4D). Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis showed that differentially expressed miRNA target genes acted in many physiological and pathological processes, including Hippo, MAPK, and Wnt signaling pathway (Figure 4F). The enrichment P values of Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis were complemented as Table S4 for reference.

Exosomal miRNAs Regulated the Functions of PASMCs Through Targeting FoxM1

A significant elevation in the expression of FoxM1 in the protein level was found when control PASMCs were treated with purified PVEC‐derived exosomes of IUGR (Figure 5A). In silico analysis (www.targetscan.org) identified that from a total of 26 miRNAs that were significantly downregulated in PVEC‐derived medium of IUGR, 3 miRNAs (miR‐214‐3p, miR‐326‐3p, and miR‐125b‐2‐3p) were all predicted to target the FoxM1 3′ untranslated region (UTR) (Figure 5B). Dual‐Luciferase assay validated the predicted binding sites on the 3′UTR of FoxM1, which showed that the luciferase activity was significantly downregulated in the PASMCs transfected with the luciferase reporter plasmid in combination with the mimics of miR‐214‐3p, miR‐326‐3p, or miR‐125b‐2‐3p, while mutations of the predicted binding sites abrogated the reduction induced by the relative mimics, confirming the specific binding of miR‐214‐3p, miR‐326‐3p, and miR‐125b‐2‐3p in FoxM1 (Figure 5C).

Figure 5. Exosomal microRNAs regulated the functions of PASMCs through targeting FoxM1.

A, Relative mRNA expression (n=6 per group) and Western blot analysis (n=5 per group) of FoxM1 in control PASMCs after being treated with purified pulmonary vascular endothelial cell‐derived exosomes for 24 hours. B, Possible binding sites between microRNAs and FoxM1 3’ untranslated region obtained by Targetscan. C, Relative luciferase activity in PASMCs after cotransfections of miR‐214‐3p/miR‐326‐3p/miR‐125b‐2‐3p mimics and untranslated region of FoxM1 plasmid. WT was the wild‐type of 3’ untranslated region of FoxM1. MT was the mutant type of 3’ untranslated region of FoxM1 (n=3 per group). D, Relative microRNA expression of exosomal miR‐214‐3p, miR‐326‐3p, and miR‐125b‐2‐3p isolated from pulmonary vascular endothelial cell‐derived medium, n=10–11 per group. E, Relative microRNAs expression of exosomal miR‐214‐3p, miR‐326‐3p, and miR‐125b‐2‐3p isolated from plasma, n=8–10 per group. F, Relative mRNA expression (n=6 per group, 48 hours) and Western blot analysis (n=6 per group, 72 hours) of FoxM1 in control PASMCs transduced with miR‐214‐3p, miR‐326‐3p, or miR‐125b‐2‐3p inhibitors. G, Proliferation (n=6 per group) and migration (n=6 per group) of control PASMCs transduced with miR‐214‐3p, miR‐326‐3p, or miR‐125b‐2‐3p inhibitors for 72 hours. All data were represented as the mean±SEM. Scale bar=50 μm. Comparisons in A, D, and E were made using Student t‐test; comparisons in C and G were made using 1‐way ANOVA; comparisons in F were made using Welch ANOVA. EdU indicates 5‐ethynyl‐2'‐deoxyuridine; FoxM1, Forkhead box M1; IUGR, intrauterine growth restriction; miR, microRNA; MT, mutant type; PASMCs, pulmonary arterial smooth muscle cells; UTR, untranslated region; and WT, wild‐type. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, comparison with treatment controls as indicated.

To verify the downregulation of exosomal miRNAs in the testing set, we performed quantitative real‐time polymerase chain reaction and found that the expression of exosomal miR‐214‐3p, miR‐326‐3p, or miR‐125b‐2‐3p secreted by PVECs in the IUGR group was significantly lower than those in the control group (Figure 5D). In addition, the expression of miR‐214‐3p and miR‐326‐3p weas also greatly downregulated in the plasma‐derived exosomes of IUGR (Figure 5E), indicating that the miRNAs enclosed in exosomes can be stable in the circulation system and might be of diagnostic value.

Next, we transfected PASMCs with relative miRNA inhibitors to investigate the effects of decreased miRNAs on FoxM1 expression and cellular functions. Although transduction with the inhibitors of the 3 miRNAs in normal PASMCs showed no significant elevation in the expression of FoxM1 in the mRNA level, the protein level was greatly augmented (Figure 5F). The proliferation and migration of PASMCs were greatly exaggerated as well (Figure 5G). On the contrary, transfected PASMCs with relative mimics of miR‐214‐3p, miR‐326‐3p, and miR‐125b‐2‐3p revealed reductions of FoxM1 expression both in the mRNA and protein level (Figure S3A). The proliferation and migration of PASMCs were also greatly inhibited (Figure S3B). To see if the combination of 3 miRNA inhibitors or mimics exhibit much bigger effects, we transfected the PASMCs with combined 3 miRNA inhibitors or mimics, respectively. Interestingly, while single treatments had only partial effects, the combined treatments with 3 miRNA inhibitors or mimics were significantly more potent on the FoxM1 expression, proliferation, and migration of PASMCs (Figure S4).

FoxM1 Inhibition Blocked the PVECs‐PASMCs Crosstalk and Inhibited the Severity of PH

Transfection of small interfering RNA targeting FoxM1 can effectively knock down the expression of FoxM1 in PASMCs (Figure 6A). Under FoxM1 targeted deletion, the dysfunctions of PASMCs observed in the hypoxia‐IUGR group were greatly ameliorated, and the proliferation of transfected PASMCs was even lower than the healthy controls (Figure 6B). The effect of PVECs' cocultivation on control PASMCs was blocked when the transfected control PASMCs were cocultured with hypoxia‐IUGR PVECs (Figure 6C).

Figure 6. FoxM1 inhibition blocked the pulmonary vascular endothelial cells‐PASMCs crosstalk and inhibited the severity of pulmonary hypertension.

A, Relative mRNA expression (1=4–6 per group, 48 hours) and Western blot analysis (n=6 per group, 72 hours) of FoxM1 in PASMCs transduced with FoxM1 small interfering RNA or scrambled small interfering RNA. B, Effects of FoxM1 small interfering RNA (72 hours) on the proliferation (n=7–8 per group) and migration (n=6 per group) of PASMCs. Scale bar=50 μm. C, Proliferation (n=5–6 per group) and migration (n=6 per group) of transfected control PASMCs (siFoxM1) after being cocultured with pulmonary vascular endothelial cells from hypoxia‐exposed intrauterine growth restriction rats for 48 hours. CN indicates control. Scale bar=50 μm. D, Relative mRNA expression (n=5–6 per group, 16 hours) and Western blot analysis (n=6 per group, 24 hours) of FoxM1 in PASMCs treated with the FoxM1 inhibitor thiostrepton (4 μM) or dimethyl sulfoxide. E, Representative images of the proliferation and migration of PASMCs treated with thiostrepton (4 μM) in vitro, together with the vascular remodeling and muscularization of distal pulmonary vessels in thiostrepton‐treated rats in vivo. Scale bar=50 μm in the first 3 figures and scale bar=100 μm in muscularization. F, Effects of thiostrepton (4 μM, 24 hours) on proliferation (n=7–8 per group) and migration (n=6–7 per group) of PASMCs. G, Mean pulmonary arterial pressure (n=5–6 per group), RV hypertrophy index (n=6–8 per group), medial wall area proportion of pulmonary arteriole (n=6–8 per group), and muscularization of distal pulmonary vessels (n=6–8 per group) in the hypoxia‐exposed intrauterine growth restriction rats after treating with thiostrepton for 3 weeks. All data were represented as the mean±SEM. Comparisons in A and D were made using a Student t‐test; comparisons in B (EdU), C, F, and G were made using 1‐way ANOVA, and comparisons in B (transwell) were made using Welch ANOVA. CN‐co‐CN indicates control PVECs cocultured with control PASMCs; EdU, 5‐ethynyl‐2'‐deoxyuridine; FoxM1, Forkhead box M1; IH, intrauterine growth restriction in hypoxia; IH‐co‐CN, IUGR‐hypoxia PVECs cocultured with control PASMCs; mPAP, mean pulmonary arterial pressure; PASMCs, pulmonary arterial smooth muscle cells; RVHI, right ventricular hypertrophy index; and siFoxM1, small interfering Forkhead box M1. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, comparison as indicated; ^# P<0.05, ^## P<0.01, ^### P<0.001, ^#### P<0.0001, comparison with untreated controls.

Treatment with thiostrepton, which has been reported to be a FoxM1 inhibitor, ³⁰ can inhibit the expression of FoxM1 effectively (Figure 6D) and reverse the dysfunctions of PASMCs observed in the hypoxia‐IUGR group (Figure 6F). We also evaluated the therapeutic potential of thiostrepton in IUGR‐induced PH rats in vivo. Treatment with thiostrepton significantly ameliorated the elevated mPAP, RV hypertrophy index, vascular remodeling, and muscularization of distal pulmonary vessels in IUGR‐induced PH rats (Figure 6G). Of note, although thiostrepton could greatly lower the RV hypertrophy index, the hypertrophic right ventricle could not return to normal.

DISCUSSION

Our results provided direct evidence that PASMCs were regulated by PVECs in a noncontact way. The exosomal miR‐214‐3p, miR‐326‐3p, and miR‐125b‐2‐3p secreted by PVECs of IUGR were greatly downregulated, which was associated with overexpression of FoxM1, leading to hyperproliferation and excessive migration of PASMCs (Figure 7). Thus, FoxM1 may be a key regulator to mediate the crosstalk between PVECs and PASMCs and may be responsible for the IUGR‐induced PH.

Figure 7. Pulmonary vascular endothelial cell‐derived exosomal microRNAs regulate pulmonary arterial smooth muscle cells via FoxM1 (Forkhead box M1) signaling in intrauterine growth restriction‐induced pulmonary hypertension.

FoxM1 indicates Forkhead box M1; IUGR, intrauterine growth restriction; miR, microRNA; MVB, multivesicular body; PASMC, pulmonary arterial smooth muscle cells; PH, pulmonary hypertension; and UTR, untranslated region.

In line with the Developmental Origins of Health and Disease theory, the Lancet reported a stimulating discovery that young adults who suffered with adverse exposures during a critical fetal and perinatal period were predisposed to an augmented pulmonary arterial pressure increase at high altitude (4559 m), although there was no significant difference at low altitude. ⁷ In the present study, we induced IUGR in rats by nutrition restriction during pregnancy and found that all the pups' birth weights were below the 10th percentile of the controls. Further investigation found that the mPAP was similar in the 2 groups under normoxic conditions, while IUGR rats displayed a more significant increase in mPAP and more exaggerated pulmonary vascular remodeling compared with the controls after exposure to hypoxia. These results suggested that IUGR had a persistent effect on the pulmonary circulation that, when activated later in life, potentially developed into PH.

Prior studies identified the oncogenic role of FoxM1 in different cancers, ³¹ and FoxM1 had been shown to regulate the proliferation of EC, SMC, and fibroblast in different kinds of lung diseases. ¹⁷ , ³² , ³³ , ³⁴ , ³⁵ Recently, FoxM1 was reported to be involved in the endothelial autocrine pathway to mediate endothelial proliferation in the pathogenesis of PH. ³⁶ Our studies demonstrated that FoxM1 was highly expressed in IUGR‐induced PH that was associated with the hyperproliferation, as well as the excessive migration of PASMCs. Our findings prompted that the FoxM1 signaling may be one of the mechanisms that mediated pulmonary vascular remodeling and even IUGR‐induced PH.

Emerging evidence suggested that pathologic crosstalk between PVECs and PASMCs may contribute to the disruption of vascular homeostasis. ECs acted as key regulators to PASMCs in 3 ways mostly, including contact‐dependent signaling, ²⁹ paracrine signaling, ¹⁰ , ¹⁷ and signaling through extracellular vesicles. ³⁷ We proved that the transfer of a signal occurred in a cell contact‐independent manner between PVECs and PASMCs through a noncontact coculture system. Targeted deletion of FoxM1 in PASMCs blocked the PVECs‐PASMCs crosstalk, indicating that PASMCs were probably regulated by PVECs via FoxM1 signaling. Intercepting the aberrant crosstalk between PVECs and PASMCs may be an effective way to prevent the development of IUGR‐induced PH.

Exosomes have emerged as important transport vesicles to mediate intercellular communication. Exosomes can be integrated into the extracellular space when the multivesicular bodies fuse with the plasma membrane and are taken in by recipient cells through various processes, including direct fusion, endocytosis, or other specific recognitions. ³⁸ Therefore, bioactive molecules contained in exosomes, such as proteins, lipids, and nucleic acids (mRNA, miRNAs, and DNA) were transferred to recipient cells. ¹⁸ , ³⁸ A recent study reported that extracellular vesicles were associated with PH via modulation of tumor necrosis factor‐beta signaling. ³⁹ In the present study, we extracted vesicles from PVEC‐derived medium and identified the purified exosomes through the typical morphology, original size, and specific exosome‐associated molecules. We treated control PASMCs with purified PVEC‐derived exosomes of IUGR and found that the FoxM1 expression in PASMCs was greatly upregulated. This result suggested that exosomes may be the vital mediators of PVECs‐PASMCs crosstalk, which transferred signals from PVECs to PASMCs through FoxM1 signaling.

MiRNAs are major bioactive components of exosomes, that were reported to contribute to tumor aggressiveness and invasiveness. ⁴⁰ , ⁴¹ Our results uncovered that 33 miRNAs were upregulated, while 26 miRNAs were downregulated in PVEC‐derived exosomes of IUGR. Those differentially expressed miRNA target genes acted in many vital processes. Hippo signaling pathway, MAPK signaling pathway, and Wnt signaling pathway were also involved, which were reported to be associated with FoxM1 by a range of evidence. ⁴² , ⁴³ , ⁴⁴ Commonly, miRNAs cannot catalyze any reactions by themselves, they form RNA induced silencing complexes and guide RNA induced silencing complexes to their targets in a sequence‐specific manner, that negatively regulate their targets via post‐transcriptional regulation predominantly. ²⁰ , ⁴⁵ Our current findings demonstrated that miR‐214‐3p, miR‐326‐3p, and miR‐125b‐2‐3p, which can specifically bind to the 3’ UTR of FoxM1, were greatly downregulated in exosomes secreted by IUGR. As a result, the inhibited effect of RNA‐induced silencing complexes on the translational repression of FoxM1 was decreased in PASMCs, which generated the overexpression of FoxM1 associated with the dysfunctions of PASMCs, promoting the IUGR‐induced PH.

Studies have also proven that exosomes containing miRNAs derived from primary focus can be transferred into the body's circulation and serve as biomarkers for disease diagnosis and prognosis. ¹⁸ , ⁴⁶ Our current data uncovered that exosomal miR‐214‐3p and miR‐326‐3p in plasma showed a similar decrease tendency in keeping with those in PVEC‐derived medium, indicating the possibility of being biomarkers for clinical application. Further studies of the correlation between exosomal miRNAs and the severity of PH are needed.

Given that FoxM1 was overexpressed in IUGR‐induced PH and served as a common target of miR‐214‐3p, miR‐326‐3p, and miR‐125b‐2‐3p, FoxM1 inhibition could be an effective strategy for the treatment of this disease. Our study verified that genetic and pharmacological ablation of FoxM1 reversed the dysfunctions of PASMCs effectively. Treating IUGR‐induced PH rats with FoxM1 inhibitor thiostrepton also remarkably attenuated PH features in vivo. Of note, thiostrepton treatment in rats could not absolutely reverse the RV hypertrophy, which was consistent with previous research. ³³ A recent study found that pharmacological activation of bone morphogenetic protein signaling with FK506 (tacrolimus) improved RV function by decreasing RV afterload. ⁴⁷ More extensive researches are warranted to investigate the combinations of drugs or other therapeutic modalities to ameliorate the pathologic changes of the myocardium.

In summary, we confirmed that undesirable perinatal environments left a persistent imprint in IUGR, making them highly sensitive to hypoxia later in life and potentially developed more significant PH. Our results provided clear evidence of PVECs‐PASMCs crosstalk in the pathogenesis of IUGR‐induced PH, which was mediated by exosomal miRNAs through targeting FoxM1. This study suggested a potential mechanism for the progression from IUGR to PH and might provide a novel therapeutic approach to treat this disease.

Our study had some limitations. Although we had demonstrated that IUGR decreased the exosomal miR‐214‐3p/miR‐326‐3p/miR‐125b‐2‐3p to crosstalk with PASMCs through FoxM1 signaling, the means by which IUGR decreased the miRNAs enclosed in exosomes had not been uncovered. Furthermore, it was plausible that other pathways may participate in the pathophysiology of IUGR‐induced PH. Whether FoxM1 can interact with other mechanisms should be investigated, which may help to inspire more desirable therapeutic strategies for IUGR‐induced PH.

Sources of Funding

This work was supported by grants from the National Natural Science Foundation of China (No. 81630037, 82001588).

Disclosures

None.

Supporting information

Data S1 Tables S1–S4 Figures S1–S4 References ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵²