Skip main navigation
×

BMP9 and BMP10 Act Directly on Vascular Smooth Muscle Cells for Generation and Maintenance of the Contractile State

Originally publishedhttps://doi.org/10.1161/CIRCULATIONAHA.120.047375Circulation. 2021;143:1394–1410

Abstract

Background:

Vascular smooth muscle cells (VSMCs) show a remarkable phenotypic plasticity, allowing acquisition of contractile or synthetic states, but critical information is missing about the physiologic signals, promoting formation, and maintenance of contractile VSMCs in vivo. BMP9 and BMP10 (bone morphogenetic protein) are known to regulate endothelial quiescence after secretion from the liver and right atrium, whereas a direct role in the regulation of VSMCs was not investigated. We studied the role of BMP9 and BMP10 for controlling formation of contractile VSMCs.

Methods:

We generated several cell type–specific loss- and gain-of-function transgenic mouse models to investigate the physiologic role of BMP9, BMP10, ALK1 (activin receptor-like kinase 1), and SMAD7 in vivo. Morphometric assessments, expression analysis, blood pressure measurements, and single molecule fluorescence in situ hybridization were performed together with analysis of isolated pulmonary VSMCs to unravel phenotypic and transcriptomic changes in response to absence or presence of BMP9 and BMP10.

Results:

Concomitant genetic inactivation of Bmp9 in the germ line and Bmp10 in the right atrium led to dramatic changes in vascular tone and diminution of the VSMC layer with attenuated contractility and decreased systemic as well as right ventricular systolic pressure. On the contrary, overexpression of Bmp10 in endothelial cells of adult mice dramatically enhanced formation of contractile VSMCs and increased systemic blood pressure as well as right ventricular systolic pressure. Likewise, BMP9/10 treatment induced an ALK1-dependent phenotypic switch from synthetic to contractile in pulmonary VSMCs. Smooth muscle cell–specific overexpression of Smad7 completely suppressed differentiation and proliferation of VSMCs and reiterated defects observed in adult Bmp9/10 double mutants. Deletion of Alk1 in VSMCs recapitulated the Bmp9/10 phenotype in pulmonary but not in aortic and coronary arteries. Bulk expression analysis and single molecule RNA–fluorescence in situ hybridization uncovered vessel bed–specific, heterogeneous expression of BMP type 1 receptors, explaining phenotypic differences in different Alk1 mutant vessel beds.

Conclusions:

Our study demonstrates that BMP9 and BMP10 act directly on VSMCs for induction and maintenance of their contractile state. The effects of BMP9/10 in VSMCs are mediated by different combinations of BMP type 1 receptors in a vessel bed–specific manner, offering new opportunities to manipulate blood pressure in the pulmonary circulation.

Clinical Perspective

What Is New?

  • Inactivation of Bmp9 and Bmp10 in developing atria results in a dramatic phenotypic shift of vascular smooth muscle cells (VSMCs) from a contractile to synthetic state, diminution of the VSMC layer, and decreased systemic blood pressure.

  • Overexpression of Bmp10 in adult mice strongly stimulates formation of contractile VSMCs and increases contractility in all vessel beds.

  • BMP type 1 (bone morphogenetic protein) receptors are expressed in different combinations in VSMCs of different vessel beds. Smooth muscle cell–specific Alk1 mutants recapitulate the Bmp9/10 mutant phenotype in pulmonary but not in aortic or coronary arteries.

  • BMP9 and BMP10 exert functions directly on VSMCs.

What Are the Clinical Implications?

  • The BMP/ALK/SMAD7 pathway is an attractive therapeutic target to modulate VSMC contractility and blood pressure.

  • Vessel bed–specific inhibition of BMP9/10 signaling may protect patients from the onset and progression of pulmonary hypertension and right ventricle dilation.

Introduction

Vascular tone is an important component of the complex regulatory network determining blood pressure. It depends critically on different vasoactive substances and on the phenotype of vascular smooth muscle cells (VSMCs),1 which show remarkable plasticity, ranging from a proliferative, undifferentiated and synthetic to a proliferation-arrested, differentiated and contractile state.2 Phenotype switching from synthetic to contractile of VSMCs and vice versa bestows blood vessels with the ability to acquire the flexibility required to perform efficiently under different long-lasting physiologic conditions. Pathologic conditions can induce pronounced phenotypic shifts in VSMCs, critically contributing to disease development. For example, accelerated migration, proliferation, and production of extracellular matrix components by synthetic VSMCs are important for development of atherosclerotic lesions.3

Numerous molecules have been shown to promote the transition of VSMCs from a contractile to synthetic state, including platelet-derived growth factor, basic fibroblast growth factor, insulin-like growth factors, epidermal growth factor, thrombin, angiotensin II, ET-1 (endothelin-1), and unsaturated lysophosphatidic acids.4,5 In comparison, the list of signaling factors that promote the contractile phenotype is relatively short, and includes soluble heparin and transforming growth factor β1 (TGFβ1).6 We and others have demonstrated that pathways essential for promoting the VSMC contractile phenotype, such as Notch and Smad-signaling, act by stimulating expression of miR-143/145, which plays a crucial role for maintenance of the contractile phenotype.7–10 In contrast, circulating, hormone-like factors that determine the contractile phenotypic state of VSMCs to a major extent are not known.

Despite intense research efforts in the past, the role of BMP9 (bone morphogenetic protein) and BMP10 for cardiovascular development, maintenance, and remodeling is unclear.11 BMP9 and BMP10 form a distinct subgroup in the TGF superfamily and share a high sequence identity at the protein level.12 Both growth factors are synthesized as inactive precursors, requiring cleavage by furin-type proteases to become activated. In adult mice, BMP9 and BMP10 are released into the bloodstream by the liver and right atrium, respectively, where they are present at high concentrations.13–15 BMP9 and BMP10 serve as high-affinity ligands for the activin receptor-like kinase 1 (ALK1) and signal via the ALK1/BMPRII complex, which also contains the accessory receptors endoglin and β-glycan.11 Mutations in components of the BMP9/10 receptor complex but also in SMAD4 give rise to different vascular disorders such as hereditary hemorrhagic telangiectasia type 1 and 2 and pulmonary hypertension (PH),16 underscoring the potentially important role of BMP9 and BMP10 for vascular biology.17–20

BMP9 and BMP10 have been described to act primarily on endothelial cells (ECs), keeping them in a resting state and regulating synthesis/release of potent vasoreactive factors such as ET-1, apelin, and adrenomedullin.16,21–23 Surprisingly, however, BMP9−/− mice are mostly normal, except for moderate defects in lymphatic vessels.24 Because BMP10−/− mice die during embryonic development, it has been difficult to determine the functional role of BMP10 in adults using genetic tools. Instead, BMP10 activity in the serum was neutralized by injection of antibodies into BMP9−/− mice, resulting in a wide range of phenotypes from hereditary hemorrhagic telangiectasia to failed closure of the ductus arteriosus.25 Although such approaches have been informative, they fail to shed light on effects caused by long-lasting absence of BMP9/10. Moreover, incomplete neutralization of BMP10 activity and potential cross-reactivity or side effects of BMP10 antibodies might compromise results. Therefore, we decided to conditionally inactivate Bmp10 in adult mice on a Bmp9 mutant genetic background to disclose the main function of BMP9/10 in the cardiovascular system. We found that induction and maintenance of the contractile state of VSMCs is the most noticeable role of BMP9 and BMP10 in vivo. Furthermore, we show that BMP9 and BMP10 act directly on VSMCs by binding to ALK1. Inactivation of the BMP type 1 receptor Alk1 in VSMCs recapitulated the Bmp9/10 phenotype in pulmonary but not in aortic or coronary arteries because of vessel bed–specific, heterogeneous expression of BMP type 1 receptors.

Methods

The data are provided with the published article or will be available together with study materials to other researchers for purposes of reproducing the results or replicating the procedure on reasonable request. Detailed methods are available in the Data Supplement.

Animals

All animal experiments were done in accordance with German animal protection laws and were approved by the local governmental animal protection committee. Generation of BMP10-LacZ reporter, BMP10loxP//loxP, BMP9/10dko, ROSA26iBMP10, and ROSA26iSmad7 mice are described in detail in the Data Supplement.

Statistical Analysis

Statistical analysis was performed by unpaired 2-tailed Student t test, paired 1-tailed Student t test when a normal distribution was assumed, or 1-way analysis of variance followed by Bonferroni multiple comparison test. Data are plotted as individual data points and means±SEM are shown. A value of P<0.05 was considered significant. N represents the number of independent experiments. All calculations were performed using GraphPad Prism software.

Data Availability

RNA sequencing data are available from the National Center for Biotechnology Information Gene Expression Omnibus database using the following accession links: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE147376 and https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE148008.

DNA microarray data have been deposited in the ArrayExpress database at the European Molecular Biology Laboratory European Bioinformatics Institute (www.ebi.ac.uk/arrayexpress) under accession numbers E-MTAB-8913 (human umbilical vein ECs) and E-MTAB-8914 (aorta).

Results

BMP9/10 Are Required for Formation of Contractile VSMCs and Normal Vessel Contractility

To study the role of BMP10 in adult mice and explore potential functional redundancy between BMP9 and BMP10, we generated germline mutants for Bmp9 and mice in which exon 2 of the Bmp10 gene is flanked by loxP sites (BMP10loxP/loxP; Figure IA in the Data Supplement). Consistent with previous reports, constitutive loss of Bmp10 initiated by CMV-Cre deleter mice (BMP10CMV) caused embryonic lethality at E10.5 attributable to arrested heart development (Figure IB in the Data Supplement).23,26 Previous studies reported a broad expression of Bmp10 in the embryonic heart that becomes restricted to the right atrium during late fetal stages.14,26 X-Gal staining of different organs from mice carrying a lacZ reporter gene inserted into the Bmp10 gene corroborated these findings and demonstrated that Bmp10 expression in adult mice is exclusively confined to the right atrium under physiologic conditions (Figure IC in the Data Supplement).

We reasoned that deletion of Bmp10 specifically in the atria might overcome embryonic lethality. Therefore, we generated BMP10ANF mice by crossing the BMP10loxP/loxP strain to ANF-Cre mice, in which the Cre recombinase is specifically expressed in atrial cardiomyocytes from at least E10.5 onwards.27 Western blot analysis confirmed that BMP10 protein expression is lost in right atria of BMP10ANF mutants (Figure ID in the Data Supplement). Phenotyping of BMP10ANF mice did not reveal any obvious morphologic alterations in the cardiovascular systems or other organs (Figure IE in the Data Supplement). Likewise, the cardiovascular system appeared grossly normal in germline Bmp9 mutants (Figure IE in the Data Supplement). Because BMP9 and BMP10 are both released into the bloodstream, we assumed that homomeric BMP9 and BMP10 molecules might substitute for BMP9–BMP10 heterodimers.15 Thus, we generated Bmp9 and Bmp10 compound mutants (BMP9−/−/BMP10ANF, hereafter called BMP9/10dko).

BMP9/10dko mice showed a dramatic vascular phenotype. The muscular layer of aortic, pulmonary, and cardiac arteries and mesenteric arteries was much thinner and the vessels were dilated compared with controls, which was confirmed by micro-computed tomography (Figure 1A and Figure IIA and IIB in the Data Supplement). The perimeter of the external elastic layer of the aorta increased by 24%. Moreover, immunofluorescence staining for ACTA2 (actin α2, smooth muscle) and PECAM (platelet endothelial cell adhesion molecule) revealed a massively reduced coverage of arteries with contractile VSMCs, as indicated by a 33% reduction of the thickness of the medial layer of the aorta. To analyze the remaining VSMCs in BMP9/10dko mice more thoroughly, we used electron microscopy. We again detected a decrease in the thickness of the muscular layer and a strong reduction of VSMCs but also dramatic ultrastructural changes such as loss of dense bodies, suggesting that VSMCs in BMP9/10dko mice acquire a synthetic phenotype (Figure 1B). Transcriptional profiling of aortae from BMP9/10dko mice followed by gene set enrichment analysis revealed a strong upregulation of genes involved in fatty acid metabolism, structural constituents of ribosome, TCA cycle, and electron transport chain (Figure 1C), which are hallmarks of synthetic VSMCs.28 Quantitative reverse transcription polymerase chain reaction (RT-qPCR) analysis and immunofluorescence analysis showed a strong downregulation of contractile markers such as Acta2, Tagln, Myh11, and Smtn, whereas expression of synthetic markers was increased, confirming the phenotype switch from a contractile to synthetic state in VSMCs (Figure 1D through 1F).

Figure 1.

Figure 1. BMP9 and BMP10 (bone morphogenetic protein) are instrumental for formation of contractile vascular smooth muscle cells (VSMCs).A, Immunofluorescence staining of aortae from 8-week-old BMP9−/− and BMP9/10dko mice for ACTA2 (actin α2, smooth muscle) and PECAM (platelet endothelial cell adhesion molecule). Quantification of medial layer thickness and external elastic lamina (EEL) perimeter (data represent mean±SEM; ****P<0.0001; n=8). B, Electron microscopy analysis of femoral arteries from BMP9−/− and BMP9/10dko mice. Higher magnification shows numerous intracellular dense bodies (white arrow) in BMP9−/− samples but virtual absence in BMP9/10dko samples (n=2). C, Gene set enrichment analysis of aortae from control and BMP9/10dko mice (n=2; y axis: enrichment score). D, Quantitative reverse transcription polymerase chain reaction (RT-qPCR) analysis of VSMC gene expression in aortae of BMP9−/− and BMP9/10dko mice (data represent mean±SEM; *P<0.05, **P<0.01; n=3). E, Immunofluorescence staining of aortae from BMP9−/− and BMP9/10dko mutants at 8 weeks for vimentin (VIM; n=3, each). F, RT-qPCR analysis of synthetic gene expression in aortae of BMP9−/− and BMP9/10dko mice (data represent mean±SEM; *P<0.05; n=3). G, RT-qPCR analysis of Bmp10 expression in the right atrium after tamoxifen injection in BMP9−/−/BMP10iMHC mice (data represent mean±SEM; ***P<0.001; n=3). H, Immunofluorescence staining of aortae from 12-week-old BMP9−/− and BMP9−/−/BMP10iMHC mice for ACTA2 and PECAM (n=6). Quantification of medial layer thickness and EEL (data represent mean±SEM; ****P<0.0001; n=6). I, RT-qPCR analysis of VSMC gene expression in aortae from BMP9−/− and BMP9−/−/BMP10iMHC mice (data represent mean±SEM; *P<0.05, **P<0.01, not significant [NS] P>0.05; n=3). J, Immunofluorescence staining for VIM in aortae from 12-week-old BMP9−/− and BMP9−/−/BMP10iMHC mutants (n=3). K, RT-qPCR analysis of synthetic gene expression in aortae from BMP9−/− and BMP9−/−/BMP10iMHC mice (data represent mean±SEM; *P<0.05, NS P>0.05; n=3). Unpaired 2-tailed Student t test was used for statistical analysis. DAPI indicates 4′,6-diamidino-2-phenylindole; FDR, false discovery rate; FWER, familywise error rate; MYH11, myosin heavy chain 11; NES, normalized enrichment score; SMTN, smoothelin; TAGLN, transgelin; and TPM4, tropomyosin α4.

To investigate physiologic consequences, we performed telemetric blood pressure measurements. Blood pressure was 11% lower in BMP9/10dko mice at daytime compared with controls and 13% at nighttime, without changes in heart rate, and did not increase after administration of the nitric oxide synthase inhibitor L-NAME (L-arginine analogue Nω-nitro-l-arginine methyl ester). We concluded that resistance vessels of BMP9/10dko mice have lost their ability to contract in response to physiologic cues but are still able to undergo further relaxation after inhibition of the angiotensin I converting enzyme by captopril (Figure IIC in the Data Supplement). Similarly, we observed a significant drop of right ventricle systolic pressure in BMP9/10dko compared with Bmp9 mutant and BMP10loxP/loxP control mice (Figure IID in the Data Supplement). We also found that BMP9−/−and BMP10ANF single mutant mice did not show a decrease in right ventricle systolic pressure compared with BMP10loxP/loxP controls (Figure IID in the Data Supplement).

Inactivation of Bmp10 expression by ANF-Cre might generate a developmental phenotype and simulate potential functions of BMP9/10 for maintaining the contractile state of VSMCs in adult mice. Therefore, we used the tamoxifen-inducible α-myosin heavy chain MerCreMer strain to inactivate Bmp10 in 8-week-old Bmp9 mutants (BMP9−/−/BMP10iMHC). The MerCreMer strain proved to be as efficient as the ANF-Cre mice to abrogate Bmp10 expression in right atria as indicated by RT-qPCR analysis (Figure 1G). We observed exactly the same vascular phenotype in BMP9−/−/BMP10iMHC as in BMP9−/−/BMP10ANF mutants. The aortae were massively dilated as indicated by an increase of the external elastic layer perimeter by 37% and thickness of the muscular layer was reduced by 32% (Figure 1H). Moreover, expression of typical contractile VSMC genes including Acta2, Tagln, Myh11, and Smtn was strongly lowered (Figure 1I). In contrast, expression of synthetic VSMC genes went up (Figure 1J and 1K). Taken together, our results demonstrate that the combined activity of BMP9 and BMP10 is instrumental for acquisition and maintenance of contractile VSMC identity.

Overexpression of Bmp10 in ECs Promotes Generation of Contractile VSMCs

Because the concomitant absence of Bmp9 and Bmp10 causes a profound loss of contractile VSMCs, we investigated whether increased expression of Bmp10 has opposite effects. Fourteen days after expression of Bmp10 in ECs from the ROSA26 locus with the help of tamoxifen-inducible Cdh5-CRE-ERT2 (ROSA26iBMP10), we observed a strong expression of Bmp10 in ECs, which massively enhanced coverage of the aorta with VSMCs marked by ACTA2 expression, resulting in an increase of medial layer thickness by 47% (Figure 2A through 2C). We observed decreased diameters of aortic lumens, indicated by a decrease of the external elastic layer perimeter by 20%, which is most likely caused by a stronger contractile force of the muscular layer (Figure 2C). RT-qPCR analysis of isolated aortae confirmed increased expression of Acta2 and other typical markers of contractile VSMCs, such as Tagln, Myh11, and Smtn (Figure 2D). Furthermore, immunofluorescence staining and RT-qPCR analysis revealed reduced expression of synthetic genes in aortae (Figure 2E and 2F), further corroborating the profound shift of VSMCs from synthetic to hypercontractile when exposed to high BMP10 concentrations.

Figure 2.

Figure 2. Overexpression of Bmp10 (bone morphogenetic protein) in endothelial cells leads to excessive formation of contractile vascular smooth muscle cells (VSMCs).A, Schematic representation of the strategy to generate the Bmp10 overexpression allele. B, Quantitative reverse transcription polymerase chain reaction (RT-qPCR) analysis of Bmp10 expression in aortae of control and ROSA26iBMP10 mice (data represent mean±SEM; **P<0.01; n=3). C, Immunofluorescence staining of aortae from control and ROSA26iBMP10 mice for ACTA2 (actin α2, smooth muscle) and PECAM (platelet endothelial cell adhesion molecule). Quantification of medial layer thickness and external elastic lamina (EEL; data are mean±SEM; ****P<0.0001; n=7). D, RT-qPCR analysis of VSMC gene expression in aortae from control and ROSA26iBMP10 mice (data represent mean±SEM; *P<0.05, **P<0.01, ***P<0.001; n=3). E, Immunofluorescence staining of aortae from control and ROSA26iBMP10 mice for vimentin (VIM; n=3). F, RT-qPCR analysis of synthetic gene expression in aortae from control and ROSA26iBMP10 mice (data represent mean±SEM; *P<0.05, **P<0.01; n=3). G, Hematoxylin & eosin staining of hearts of control and ROSA26iBMP10 mice 2 weeks after tamoxifen injection (**P<0.01; n=5). H, Right ventricle systolic pressure (RVSP) measurements of 8-week-old, Cre-recombinase expressing control (n=7) and ROSA26iBMP10 (n=3) mice 7 days after tamoxifen injection (data represent mean±SEM; unpaired 1-tailed Student t test: *P<0.05). I, Average telemetric blood pressure measurements of ROSA26iBMP10 mice 2 to 4 days after sensor implantation but before tamoxifen injection and 7 to 9 days after tamoxifen induction (data represent mean±SEM; paired 1-tailed Student t test: *P<0.05, **P<0.01; n=5). Unpaired 2-tailed Student t test was used for panels B, C, D, F, and G. DAPI indicates 4′,6-diamidino-2-phenylindole; and RV, right ventricle.

Similar observations were made in other vessels, including the pulmonary and coronary arteries. Interestingly, the increased thickness of arterial muscular layers was associated with dilation of the right ventricle 2 weeks after tamoxifen injection (Figure 2G). Hemodynamic measurements revealed an increase of the right ventricle systolic pressure in ROSA26iBMP10 mice, suggesting that increased expression of Bmp10 eventually leads to right ventricular dilation (Figure 2H). Likewise, we observed a substantial elevation of systemic blood pressure in ROSA26iBMP10 mice after initiation of increased Bmp10 expression (Figure 2I). Taken together, the findings clearly demonstrate that enhanced availability of BMP10 within arteries is sufficient to increase the number of contractile VSMCs and reduce the diameter of vessel lumens.

Smad7 Suppresses Formation and Differentiation of VSMCs

In principle, the effects of BMP9 and BMP10 on VSMCs might be mediated by ECs, which are exposed to serum components. Alternatively, BMP9 and BMP10 might exert effects directly on VSMCs, after passing through the EC layer or after delivery via the microcirculation serving larger muscularized vessels. To understand whether inhibition of SMAD signaling in VSMCs prevents formation of contractile VSMCs, we specifically expressed the inhibitory Smad7 in smooth muscle cells (SMCs).29,30 Following the same strategy as for Bmp10 overexpression, we generated SM22-Cre/ROSA26iSmad7 (ROSA26SM22-iSmad7) and SMMHC-CreERT/ROSA26iSmad7 (ROSA26SMMHC-iSmad7) strains. We found that increased expression of Smad7 in smooth muscle essentially abrogated formation of VSMCs during embryonic development, leading to embryonic lethality at E10.5. Furthermore, expression of the SMC-specific markers ACTA2 (Figure 3A) and TAGLN (transgelin; Figure 3B) was nearly gone. Conditional expression of Smad7 in SMCs of adult mice strongly decreased the coverage of arterial vessels by ACTA2-expressing VSMCs, reducing the thickness of the muscular layer by 34% and causing dilation of aortae by 31% (Figure 3C). Accordingly, expression of specific markers for contractile VSMCs in aortae was reduced (Figure 3D), whereas immunofluorescence staining and RT-qPCR indicated increased expression of vimentin and TPM4 (tropomyosin α4; Figure 3E and 3F). The nearly complete absence of contractile VSMCs in mice overexpressing Smad7 in SMCs clearly indicates that BMP/SMAD signaling is crucial for development and maintenance of VSMCs.

Figure 3.

Figure 3. Smad7 suppresses formation and differentiation of vascular smooth muscle cells (VSMCs). Immunofluorescence staining for (A) ACTA2 (actin α2, smooth muscle) and PECAM (platelet endothelial cell adhesion molecule) and (B) TAGLN (transgelin) and PECAM on transversal sections of control and ROSA26SM22-iSmad7 embryos through descending dorsal aortae at E10.5 (n=5, each). C, Immunofluorescence staining of control and ROSA26SMMHC-iSmad7 aortae for ACTA2 and PECAM after tamoxifen injection. Quantification of medial layer thickness and the external elastic lamina (EEL) perimeter (data represent mean±SEM; ***P<0.001, ****P<0.0001; n=6). D, Quantitative reverse transcription polymerase chain reaction (RT-qPCR) analysis of SMC gene expression in control and ROSA26SMMHC-iSmad7 aortae (data represent mean±SEM; unpaired 2-tailed Student t test: *P<0.05; n=3). E, Immunofluorescence staining of control and ROSA26SMMHC-iSmad7 aortae for vimentin (VIM) after tamoxifen injection (n=5). F, RT-qPCR analysis of synthetic gene expression in control and ROSA26SMMHC-iSmad7 aortae (data represent mean±SEM; unpaired 2-tailed Student t test: *P<0.05, not significant [NS] P>0.05; n=3). G, Immunofluorescence staining of collateral vessels for BMP10 after femoral artery ligation and injection of BMP10 or albumin. Noninjected BMP10ANF mutant mice served as a negative control (right panels; data represent mean±SEM; 1-way analysis of variance: ***P<0.001, NS P>0.05; scale bar, 30 μm). H, Immunofluorescence staining of phosphorylated SMAD1/5/8 after perfusion of myograph-mounted isolated mesenteric arteries with BMP9/10 (n=4). DAPI indicates 4′,6-diamidino-2-phenylindole; MYH11, myosin heavy chain 11; and SMTN, smoothelin.

Inhibition of BMP/SMAD signaling in VSMCs by Smad7 expression suggested that BMP10 acts directly on VSMCs. To prove that BMP10 binds directly to VSMCs within vessel walls, we used femoral artery ligation (FAL) as an experimental model.31 FAL causes a strong increase of contractile VSMCs in collateral vessels and therefore creates a condition facilitating detection of BMP10 bound to VSMCs. BMP10 (5 µg in phosphate-buffered saline) was intraperitoneally injected into wild-type mice twice, 1 day before and 5 days after FAL. Analysis of remodeling collateral vessels in semimembranosus and semitendinosus muscles of the proximal limb 7 days after FAL revealed a clear colocalization of BMP10 and ACTA2 signals in BMP10-injected wild-type mice, indicating that circulating BMP10 is able to bind directly to VSMCs of remodeling collateral arteries. Collateral vessels of BMP10ANF mice subjected to FAL served as negative controls (Figure 3G).

Furthermore, we isolated mesenteric arteries from wild-type mice and perfused vessels in a pressure myograph system, mimicking blood circulation.32 Perfusion of mesenteric artery with BMP9/10 induced Smad1/5/8 phosphorylation in the VSMC layer; no SMAD1/5/8 phosphorylation was observed in control-perfused arteries (Figure 3H). Because BMP9/10 cannot reach the VSMC layer via capillary vasa vasorum in arteries mounted in a pressure myograph, we concluded that circulating BMP9/10 reaches VSMCs through the endothelial barrier.

BMP9/10 Directly Act on VSMCs via the ALK1/SMAD Signaling Pathway to Induce Differentiation

To further validate the hypothesis that BMP9 and BMP10 act directly on VSMCs, we isolated pulmonary artery SMCs (PASMCs) from wild-type mice, which rapidly lose the contractile phenotype in vitro. Treatment of dedifferentiated PASMCs with BMP9/10 resulted in a strong increase of Acta2, Tagln, Myh11, and Smtn expression as measured by immunofluorescence staining (Figure 4A) and RT-qPCR (Figure 4B), indicating a switch from the dedifferentiated, synthetic to a differentiated, contractile state. RNA sequencing analysis of BMP9/10–treated PASMCs corroborated the strong upregulation of genes required for VSMC contraction including Cald1, Tpm1,3, Smtn, and Myl9,12 (Figure 4C). gene set enrichment analysis revealed enrichment of terms related to SMC contraction, elastic fiber formation, extracellular matrix organization, syndecan interactions, and vesicle-mediated transport (Figure 4D). Accordingly, we observed increased expression of typical BMP target genes such as Smurf, Fstl1, and others within the terms “smooth muscle contraction” and “signaling by BMP” (Figure 4E and 4F). The type 2 receptor Bmpr2 and the inhibitory Smad7 were upregulated as well, probably representing a cellular countermeasure to prevent excessive BMP–SMAD signaling (Figure 4C and 4F). The dramatic changes in the transcriptional profile of BMP9/10–treated VSMCs suggest that cellular processes related to maturation of the muscular layer depend on BMP9/10.

Figure 4.

Figure 4. BMP9/10 (bone morphogenetic protein) directly act on vascular smooth muscle cells (VSMCs) to induce VSMC differentiation.A, Immunofluorescence staining for ACTA2 (actin α2, smooth muscle), TAGLN (transgelin), MYH11 (myosin heavy chain 11), and SMTN (smoothelin) in cultured pulmonary artery smooth muscle cells (PASMCs) treated with BMP9/10. B, Quantitative reverse transcription polymerase chain reaction (RT-qPCR) analysis of Acta2, Tagln, Myh11, and Smtn expression in PASMCs after treatment with BMP9/10 (data represent mean±SEM; unpaired 2-tailed Student t test: *P<0.05, **P<0.01, ***P<0.001; n=3). C, RNA sequencing analysis of BMP9/10-treated PASMCs. The heatmap displays the top 50 differentially expressed genes (DEGs) based on significance (false discovery rate). D, Gene set enrichment analysis of PASMCs treated with BMP9/10 compared with nonstimulated cells by KOBAS using the Reactome database. The top 15 pathways based on P values are displayed. All DEGs in PASMCs treated with BMP9/10 compared with nonstimulated cells were used as input. Size of solid circles represents corresponding number of DEGs belonging to the term and the surrounding transparent circles represent the total number of genes assigned to each term (cutoff: false discovery rate <5%, baseMean >5, log2[fold change] >0.585). E, Heatmap for RNA sequencing analysis of BMP9/10-treated PASMCs with DEGs matching to enriched term “smooth muscle contraction” of Reactome database (n=3). F, Heatmap for RNA sequencing analysis of BMP9/10-treated PASMCs with DEGs matching to enriched term “signaling by BMP” of Reactome database (n=3). DAPI indicates 4′,6-diamidino-2-phenylindole; NGF, nerve growth factor; and TGF, transforming growth factor.

Direct comparison of BMP9 and BMP10 with other molecules inducing VSMC differentiation revealed that BMP9 and BMP10 have a remarkable potential to switch VSMCs from a synthetic to a contractile state. We found that 100 pmol/L of BMP9 were sufficient to induce strong expression of the contractile marker ACTA2 in cultured PASMCs, whereas BMP10, TGFβ1, activin A, and retinoic acid had negligible effects at this concentration (Figure IIIA through IIID in the Data Supplement). It was necessary to raise BMP10 and TGFβ1 to 500 pmol/L to achieve robust ACTA2 expression. In comparison, the potency of activin A and retinoic acid to induce contractile markers in PASMCs was much weaker, requiring 2.5 nmol/L and 10 nmol/L, respectively (Figure IIID and IIIE in the Data Supplement).

To gain further insight into the intracellular signaling pathways initiated by BMP9 or BMP10, we determined the phosphorylation levels of SMAD1/5/8; SMAD2/3; ERK5 (extracellular signal regulated kinase); ERK1/2, P38; SAPK/JNK (stress-activated protein kinases/Jun amino-terminal kinases); and CREB (cAMP response element-binding protein). BMP9 or BMP10 strongly induced phosphorylation of SMAD1/5/8 but not of SMAD2/3; TGFβ1 showed an opposite pattern. BMP9, BMP10, and TGFβ1 did not have obvious effects on ERK5; ERK1/2, P38; SAPK/JNK; or CREB phosphorylation (Figure 5A). Because overexpression of the inhibitory Smad7 in SMCs strongly suppressed formation of contractile VSMCs in vivo, we wanted to know whether SMAD7 inhibits SMAD1/5/8 and SMAD2/3 to a similar extent. We isolated PASMCs from ROSA26iSmad7 mice and activated Smad7 expression by an Adeno-Cre virus. Recombined PASMCs were exposed to BMP9/10 or TGFβ1 and subjected to Western blot analysis. Interestingly, SMAD7 inhibited phosphorylation of SMAD1/5/8 more efficiently than SMAD2/3 (Figure 5B). Although expression of Acta2, Tagln, Myh11, and Smtn was reduced after addition of TGFβ1 in Smad7 expressing PASMCs, the effects on BMP9/10-treated cells were clearly stronger, suggesting that the strong effects of Smad7 overexpression on VSMC formation and differentiation are primarily mediated by inhibition of the SMAD1/5/8 signaling axis (Figure 5C and 5D).

Figure 5.

Figure 5. BMP9/10 (bone morphogenetic protein) induce differentiation of vascular smooth muscle cells (VSMCs) via the activin receptor-like kinase 1 (ALK1)/SMAD signaling pathway.A, Western blot analysis of cultured pulmonary artery smooth muscle cells (PASMCs) treated with BMP9, BMP10, transforming growth factor β1 (TGFβ1), oncostatin M (OSM), and bovine serum albumin (BSA). B, Western blot analysis and quantification of pSMAD1/5/8, pSMAD2, and pSMAD3 in ROSA26iSmad7 PASMCs after adenoviral transduction of Cre recombinase and treatment with BMP9/10 or TGFβ1 (data represent mean±SEM; unpaired 2-tailed Student t test: *P<0.05, **P<0.01; n=2). C and D, Quantitative reverse transcription polymerase chain reaction (RT-qPCR) analysis of VSMC gene expression in PASMCs after adenoviral transduction of Cre recombinase and treatment with BMP9/10 (C) and TGFβ1 (D; data represent mean±SEM; unpaired 2-tailed Student t test: *P<0.05, **P<0.01, ***P<0.001, not significant [NS] P>0.05; n=3). E, RT-qPCR analysis of Alk1 expression in endothelial cells (ECs), VSMCs, and cardiomyocytes (CMs; data represent mean±SEM; 1-way analysis of variance: *P<0.05, NS P>0.05; n=2). F, Immunofluorescence staining and RT-qPCR for ALK1 in cultured PASMCs isolated from wild-type or SM22-Cre/ALK1ΔloxP/ΔloxP mice (data represent mean±SEM; unpaired 2-tailed Student t test: *P<0.05; n=3). ACTA2 indicates actin α2, smooth muscle; DAPI, 4′,6-diamidino-2-phenylindole; Gapdh, glyceraldehyde 3-phosphate dehydrogenase; MYH11, myosin heavy chain 11; SMC, smooth muscle cell; SMTN, smoothelin; and TAGLN, transgelin.

The strong induction of SMAD1/5/8 phosphorylation after BMP9/10 treatment and the suppression of SMAD1/5/8 phosphorylation by Smad7 overexpression, mimicking effects of BMP9/10dko, suggest that BMP9 and BMP10 might exert their effects on VSMCs via ALK1 or other members of the ALK receptor family, although ALK1 was mainly described as an endothelial receptor.33–35 However, RT-qPCR analysis revealed that Alk1 is expressed at even higher levels in VSMCs than in ECs (Figure 5E). Immunofluorescence staining and RT-qPCR for ALK1 with isolated PASMCs confirmed this assumption (Figure 5F). To address the functional relevance of Alk1 receptor expression for mediating BMP9/10 effects in VSMCs, we specifically inactivated the Alk1 gene in SMCs using SM22-Cre mice. Immunofluorescence staining revealed that Alk1-mutant PASMCs were unable to upregulate Acta2 expression and stimulate SMAD1/5/8 phosphorylation in response to BMP9/10 treatment (Figure IVA through IVC in the Data Supplement).

Deletion of Alk1 in SMCs Recapitulates the BMP9/10dko Vascular Phenotype in Pulmonary But Not in Aortic and Coronary Arteries

The strict dependence of BMP9/10 signaling on Alk1 in PASMCs was surprising, given the existence of several other ALK receptors with similar ligand binding properties. Therefore, we investigated the consequences of Alk1 receptor gene inactivation in SMCs in mice. Analysis of SM22-Cre/ALK1ΔloxP/ΔloxP mutants uncovered a massive reduction of contractile ACTA2-positive VSMCs in pulmonary arteries (Figure 6A). In stark contrast, aortae and coronary arteries did not show a reduction of VSMCs (Figure 6C and Figure VC in the Data Supplement). Similarly, RT-qPCR analysis revealed a strong decline of Acta2 expression and a significant decrease of Tagln, Myh11, and Smtn in lung samples, but only a moderate decrease of Acta2 in aortae (Figure 6B and 6D). Furthermore, immunofluorescence staining and Western blot analysis showed reduced phosphorylation of SMAD1/5/8 in pulmonary arteries but not in aortae and coronary vessels (Figure 6E and 6F and Figure VA, VB, and VD in the Data Supplement). In conclusion, inactivation of Alk1 in SMCs essentially recapitulated the vascular phenotype of BMP9/10dko mice but only in a restricted set of vessels, specifically in pulmonary arteries. To exclude any potential artifacts that might arise from differential activity of SM22-Cre in different vessel beds, we generated SM22-Cre/ROSA26-FloxStop-GFP mice. In line with previous reports,36 we found comparable activation of the GFP reporter in all vessels analyzed, including aortic, pulmonary, and coronary arteries.

Figure 6.

Figure 6. Deletion of Alk1 (activin receptor-like kinase 1) in smooth muscle cells (SMCs) recapitulates the BMP9/10dko (bone morphogenetic protein) vascular phenotype in pulmonary but not in aortic arteries.A, Immunofluorescence staining for ACTA2 (actin α2, smooth muscle) and PECAM (platelet endothelial cell adhesion molecule) in lungs of 8-week-old control, BMP9/10dko, and SM22-Cre/ALK1ΔloxP/ΔloxP mice (n=5). Quantification of medial layer thickness of pulmonary arteries (data represent mean±SEM; 1-way analysis of variance: ***P<0.001, not significant [NS] P>0.05; n=5). B, Quantitative reverse transcription polymerase chain reaction (RT-qPCR) analysis of vascular SMC gene expression in lungs of 8-week-old control and SM22-Cre/ALK1ΔloxP/ΔloxP mice (data represent mean±SEM; unpaired 2-tailed Student t test: *P<0.05; **P<0.01, ***P<0.001; n=3). C, Immunofluorescence staining of aortae from 8-week-old control, BMP9/10dko, and SM22-Cre/ALK1ΔloxP/ΔloxP mice for ACTA2 and PECAM (n=6). Quantification of aorta medial layer thickness (data represent mean±SEM; 1-way analysis of variance: ***P<0.001, NS P>0.05; n=6). D, RT-qPCR analysis of VSMC gene expression in aortae of 8 weeks old control and SM22-Cre/ALK1ΔloxP/ΔloxP mice (data represent mean±SEM; unpaired 2-tailed Student t test: *P<0.05; NS P>0.05; n=3). E and F, Immunofluorescence staining of lungs (E) and aortae (F) from 8-week-old control, BMP9/10dko, and SM22-Cre/ALK1ΔloxP/ΔloxP mice for phosphorylated SMAD1/5/8 and PECAM (n=3). Airways are labeled by (aw) and vessels by (v). DAPI indicates 4′,6-diamidino-2-phenylindole; MYH11, myosin heavy chain 11; SMC, smooth muscle cell; SMTN, smoothelin; and TAGLN, transgelin.

Transcriptional profiling revealed upregulation of the vasodilatory factor apelin in hearts from BMP9/10dko mice and strong downregulation in BMP9/10-treated human umbilical vein endothelial cells (Figure VIA through VIC in the Data Supplement). These findings are in line with previous observations demonstrating lower mRNA levels of ET-1 and higher levels of apelin and adrenomedullin in BMP9−/− mice chronically exposed to hypoxia.16 To investigate whether the increased level of apelin in BMP9/10dko mice might be involved in the loss of contractile VSMCs and increased dilation of vessels, we generated apelin/BMP9/BMP10 triple mutant mice (TKO). Surprisingly, the phenotype of apelin/BMP9/BMP10 triple mutant mice did not differ from BMP9/10dko mice. We observed the same loss of contractile VSMCs and vasodilation as in BMP9/10dko mice (Figure VID and VIE in the Data Supplement). We concluded that the upregulation of apelin is not important for the VSMC phenotype of BMP9/10dko mice, although BMP9/10-mediated regulation of apelin and other vasoactive molecules might be relevant in a different physiologic context.

Differential Expression of BMP Type 1 Receptors Reveals Heterogeneity of VSMCs in Different Vessel Beds

We reasoned that the strong reduction of contractile VSMCs in pulmonary arteries but not in aortae and coronary vessels of SM22-Cre/ALK1ΔloxP/ΔloxP mutants might be caused by differential expression of Alk receptor genes in VSMCs of distinct vessel beds, preventing the full phenocopy of the vascular phenotype of BMP9/10dko mice. Thus, we studied expression of different Alk receptor genes in different vessels by RNA fluorescence in situ hybridization at single-cell resolution. Interestingly, aortic VSMCs expressed nearly all type 1 BMP receptors at similar levels with the exception of Alk6, which was relatively low compared with Alk1, Alk2, and Alk3. A similar pattern was observed in coronary arteries. In stark contrast, VSMCs in pulmonary arteries only expressed Alk1 at high levels, but not Alk2, Alk3, and Alk6 (Figure 7A through 7C).

Figure 7.

Figure 7. Differential expression of BMP type 1 (bone morphogenetic protein) receptors in vascular smooth muscle cells (VSMCs) of different vessels. RNA fluorescence in situ hybridization (RNA-FISH) of Alk1, Alk2, Alk3, and Alk6 (activin receptor-like kinase) in (A) aortic arteries and (B) coronary arteries. Quantification of type 1 receptor expression in aortae (A) and coronary arteries (B; data represent mean±SEM; ***P<0.001, not significant [NS] P>0.05; n=4). C, RNA-FISH and quantification of type 1 receptor expression in pulmonary arteries (data represent mean±SEM; ***P<0.001, NS P>0.05; n=3). D and E, RNA sequencing analysis of type 1 receptor expression in smooth muscle cells (SMCs) from (D) aorta and (E) lung (data represent mean±SEM; **P<0.01, ***P<0.001, NS P>0.05; n=3). One-way analysis of variance followed by Bonferroni multiple comparison test was used for statistical analysis. DAPI indicates 4′,6-diamidino-2-phenylindole.

To validate these results, we isolated SMCs from lungs of SMACreERT2postdTomatopos mice by fluorescence activated cell sorting for subsequent RNA sequencing. In addition, we took advantage of published RNA sequencing data analyzing VSMCs from the medial layer of the descending thoracic aorta.37 Calculation of relative expression levels of BMP type 1 receptors revealed that Alk3 was most abundantly expressed in VSMCs from descending thoracic aortae, compared with Alk1, Alk2, and Alk6 (Figure 7D). In contrast, Alk1 was clearly the dominant receptor in SMCs of the lung, which include both VSMCs and bronchotracheal SMCs (Figure 7E). Alk3 expression was much lower compared with Alk1 and Alk2 and Alk6 were only expressed at very low levels. RNA sequencing data from pulmonary SMCs need to be viewed with caution, because isolated cells will contain a certain fraction of bronchotracheal SMCs that might express a different profile of Alk receptor genes. However, we noted that inactivation of Alk1 in SMCs prevented phosphorylation of SMAD1/5/8 not only in pulmonary VSMCs but also bronchotracheal SMCs, indicating that ALK1 is the dominant type 1 receptor in both SMC populations in the lung (Figure 6E). Antibody staining confirmed that ALK2 and ALK3 are only present in aortic but not in pulmonary arteries (Figure VIIA and VIIB in the Data Supplement). Taken together, the unexpected heterogeneity of VSMCs in different vessels in respect to Alk receptor gene expression explains why only pulmonary arteries in SM22-Cre/ALK1ΔloxP/ΔloxP mutant mice recapitulated the reduced VSMC coverage and displayed attenuated contractility observed in BMP9/10dko.

Discussion

In this study, we provide a detailed analysis of the function of BMP9 and BMP10 in the vascular system of mice. We found that the dominant effect of concomitant Bmp9 and Bmp10 inactivation is a profound loss of contractile VSMCs, resulting in dilation of major vessels and reduced blood pressure (Figure 8). This is a surprising finding, because it had been assumed that BMP9 and BMP10 act selectively on vascular ECs to maintain vascular homeostasis and to inhibit EC apoptosis, migration, proliferation, and angiogenesis.22,38,39Bmp9 mutant mice were reported to show defects in lymphatic vessel maturation and valve formation; the function of Bmp10 in adult mice was unknown.23,24,26 In our hands, neither BMP10ANF nor BMP9−/− mice exhibited any obvious phenotype under baseline conditions. Mice were healthy and fertile and the vessel morphology in adult animals appeared normal.

Figure 8.

Figure 8. Model of the role of BMP9 and BMP10 (bone morphogenetic protein) for generation of contractile vascular smooth muscle cells (VSMCs) in different vessels. BMP9 and BMP10 are released into the bloodstream by the liver (BMP9) and the right atrium (BMP10). BMP9 and BMP10 pass the endothelial barrier to directly stimulate formation of contractile VSMCs via different ALK (activin receptor-like kinase) receptors. Concomitant loss of Bmp9 and Bmp10 leads to a massive loss of contractile VSMCs in all muscularized vessels. Because pulmonary arteries express Alk1 but not Alk2 and Alk3 at high levels, inactivation of Alk1 in SMCs recapitulates the Bmp9/10 mutant phenotype only in pulmonary arteries but not aortic and coronary arteries, which also express Alk2 and Alk3, potentially compensating for the absence of Alk1.

The absence of a strong vascular phenotype in single mutants suggested that BMP9 and BMP10 might serve overlapping functions in adults, in particular because both BMP9 and BMP10 are high affinity ligands for ALK1 and endoglin and signal via BMPR2 and SMAD1/5/8.22,40 The dramatic reduction of contractile VSMCs of BMP9/10dko but not in single BMP10ANF and BMP9−/− mice proves that this assumption is correct. In addition, our results revealed that the concurrent absence of BMP9 and BMP10 foremost affected VSMCs and not ECs. This does not mean that BMP9 and BMP10 have no effects on ECs. In fact, we observed a strong downregulation of apelin and stimulation of ET-1 in ECs treated with BMP9/10, which was mirrored by corresponding expression changes in the aorta. However, the changes in ECs cannot account for the profound loss of contractile VSMCs as indicated by recapitulation of the BMP9/10dko phenotype in SM22-Cre/ALK1ΔloxP/ΔloxP mutant mice. Moreover, the regulation of apelin in ECs by BMP9 and BMP10 does not seem to be relevant for the contractile VSMC phenotype, because inactivation of apelin did not reverse the loss of contractile VSMCs and vessel dilation in BMP9/10dko mice.

Further support for this conclusion comes from in vitro experiments demonstrating that BMP9 is the strongest inducer of the contractile VSMC phenotype when compared with TGFβ1, activin A, and retinoic acid. To act directly on VSMCs, BMP9 and BMP10 have to pass the endothelial barrier. We demonstrated by coimmunofluorescence staining of collateral vessels for BMP10 and ACTA2 that BMP10 binds directly to VSMCs. Similar observations were made in isolated mesenteric arteries mounted in a pressure myograph system, in which perfusion with BMP9/10 induced SMAD1/5/8 phosphorylation in the VSMC layer. Unlike many microvessels in the liver or bone marrow, the endothelium of muscularized vessels is continuous and also lacks fenestrations.41 Thus, BMP9 and BMP10 might use transendothelial transport mechanisms via paracellular or transcellular routes, which will also allow localized regulation.

The finding that inactivation of Alk1 in SMCs only recapitulated the BMP9/10dko phenotype in pulmonary arteries but not in aortae and coronary vessels was an unexpected result and indicates a substantial heterogeneity of VSMCs in different vessels. Heterogeneity of VSMCs might reflect differential physiologic requirements and different signaling needs of vessels in different organs.37,42 Pulmonary VSMCs only express Alk1 at high levels, which matches the loss of contractile VSMCs in pulmonary arteries but not in aortae and coronary arteries of SM22-Cre/ALK1ΔloxP/ΔloxP mutant mice. The expression of Alk2 and Alk3 in VSMCs of aortae and coronary arteries might compensate for the absence of Alk1 and explain maintenance of the contractile phenotype in these vessels. Because inactivation of Bmp9 and Bmp10 caused a profound reduction of contractile VSMCs in aortic and coronary arteries, it seems reasonable that BMP9 and BMP10 also use ALK2 and ALK3 for signaling. However, it remains possible that ALK2 and ALK3 preferentially use different ligands and only become relevant for BMP9/10 signaling in the absence of ALK1. Generation of Alk1, Bmp9, and Bmp10 or Alk1/Alk3/Alk3 triple knockout mice might address this possibility. Heterogeneous expression of Alk receptor genes in VSMCs of different vessels might serve specific local physiologic requirements and occur secondarily. However, it is also possible that differential expression of ALK receptors is a consequence of developmental processes, because VSMCs originate from at least 8 different sources of progenitor cells.43,44 A comprehensive map of the developmental origin of VSMCs in pulmonary and aortic arteries is still missing, which makes it difficult to completely and precisely assign cellular sources for all VSMCs in distinct arteries. Interestingly, however, it was described that cardiopulmonary progenitor cells contribute to VSMCs of pulmonary but not aortic arteries,45 whereas VSMCs in the descending aorta seem to be exclusively derived from the paraxial mesoderm after replacing lateral plate mesoderm-derived VSMCs.44 The understanding of VSMCs heterogeneity is in its infancy but offers exciting new opportunities and enables targeted manipulation of VSMCs in different vascular beds. For example, specific inhibition of Alk1 signaling processes in VSMCs might reduce hypermuscularization of pulmonary arteries in PH without affecting VSMCs that express Alk2 and Alk3.

The proposed role of BMP9 and BMP10 for generation and maintenance of contractile VSMCs seems to oppose recent studies reporting reduced plasma levels of BMP9 and BMP10 in patients with idiopathic PH, characterized by massive hypermuscularization of pulmonary arteries.46 In addition, it has been shown that administration of BMP9 to mice carrying the heterozygous R899X mutation in the Bmpr2 gene reverses established PH in these mice.47 How is it possible that PH is associated with reduced plasma levels of BMP9 and BMP10 and administration of BMP9 improves PH when BMP9 and BMP10 are essential for formation of contractile VSMCs? We hypothesize that reduced plasma levels of BMP9 and BMP10 in PH are the consequence of a negative feedback loop restraining formation of contractile VSMCs. The excessive formation of contractile VSMCs during PH might depend on pathologic signals, involve ECs, or involve aberrant BMPR2 signaling that is not present in physiologic conditions.48 Further support for this idea comes from a recent study that describes that inhibition of BMP9 partially protects against chronic hypoxia-induced PH in mice.16 The pathophysiology of PH is highly complex and depends on multiple cellular players including endothelial, smooth muscle, and adventitial cells.49 It will be highly interesting to investigate the effect of decreased or increased BMP9/10 expression in different models of PH, including BMPR2 mutants.47

In conclusion, our study identifies BMP9 and BMP10 as hormone-like factors that circulate in the bloodstream and directly bind to VSMCs for control of the contractile state. Disbalance in this pathway enables VSMCs to escape control by BMP9 and BMP10 and initiates pathologic processes. Site-specific manipulation of BMP9/10/ALK/SMAD signaling by exploiting VSMC heterogeneity provides new means for therapeutic interventions.

Acknowledgments

The authors thank Stefan Günther and Mario Looso for performing RNA sequencing and bioinformatics analysis, Ann Atzberger for help with cell sorting, Ralf Brandes for micro-computed tomography analysis, Ulrich Gaertner for performing EM analysis, and Vincent Cristoffels and S. Paul Oh for providing ANF-Cre and Alk1-floxed mice, respectively.

Supplemental Materials

Expanded Methods

Data Supplement Figures I–VII

References 49–58

Disclosures None.

Footnotes

Sources of Funding, see page 1408

https://www.ahajournals.org/journal/circ

The Data Supplement is available with this article at https://www.ahajournals.org/doi/suppl/10.1161/circulationaha.120.047375.

Thomas Braun, MD, PhD, Department of Cardiac Development and Remodeling, Max Planck Institute for Heart and Lung Research, Ludwigstr 43, 61231 Bad Nauheim, Germany. Email

References

  • 1. Brozovich FV, Nicholson CJ, Degen CV, Gao YZ, Aggarwal M, Morgan KG. Mechanisms of vascular smooth muscle contraction and the basis for pharmacologic treatment of smooth muscle disorders.Pharmacol Rev. 2016; 68:476–532. doi: 10.1124/pr.115.010652CrossrefMedlineGoogle Scholar
  • 2. Rensen SS, Doevendans PA, van Eys GJ. Regulation and characteristics of vascular smooth muscle cell phenotypic diversity.Neth Heart J. 2007; 15:100–108. doi: 10.1007/BF03085963CrossrefMedlineGoogle Scholar
  • 3. Kawai-Kowase K, Owens GK. Multiple repressor pathways contribute to phenotypic switching of vascular smooth muscle cells.Am J Physiol Cell Physiol. 2007; 292:C59–C69. doi: 10.1152/ajpcell.00394.2006CrossrefMedlineGoogle Scholar
  • 4. Taylor DS, Cheng X, Pawlowski JE, Wallace AR, Ferrer P, Molloy CJ. Epiregulin is a potent vascular smooth muscle cell-derived mitogen induced by angiotensin II, endothelin-1, and thrombin.Proc Natl Acad Sci U S A. 1999; 96:1633–1638. doi: 10.1073/pnas.96.4.1633CrossrefMedlineGoogle Scholar
  • 5. Takahashi M, Hayashi K, Yoshida K, Ohkawa Y, Komurasaki T, Kitabatake A, Ogawa A, Nishida W, Yano M, Monden M, et al.. Epiregulin as a major autocrine/paracrine factor released from ERK- and p38MAPK-activated vascular smooth muscle cells.Circulation. 2003; 108:2524–2529. doi: 10.1161/01.CIR.0000096482.02567.8CLinkGoogle Scholar
  • 6. Beamish JA, He P, Kottke-Marchant K, Marchant RE. Molecular regulation of contractile smooth muscle cell phenotype: implications for vascular tissue engineering.Tissue Eng Part B Rev. 2010; 16:467–491. doi: 10.1089/ten.TEB.2009.0630CrossrefMedlineGoogle Scholar
  • 7. Boettger T, Beetz N, Kostin S, Schneider J, Krüger M, Hein L, Braun T. Acquisition of the contractile phenotype by murine arterial smooth muscle cells depends on the Mir143/145 gene cluster.J Clin Invest. 2009; 119:2634–2647. doi: 10.1172/JCI38864CrossrefMedlineGoogle Scholar
  • 8. Boucher J, Gridley T, Liaw L. Molecular pathways of notch signaling in vascular smooth muscle cells.Front Physiol. 2012; 3:81. doi: 10.3389/fphys.2012.00081CrossrefMedlineGoogle Scholar
  • 9. Davis-Dusenbery BN, Chan MC, Reno KE, Weisman AS, Layne MD, Lagna G, Hata A. down-regulation of Kruppel-like factor-4 (KLF4) by microRNA-143/145 is critical for modulation of vascular smooth muscle cell phenotype by transforming growth factor-beta and bone morphogenetic protein 4.J Biol Chem. 2011; 286:28097–28110. doi: 10.1074/jbc.M111.236950CrossrefMedlineGoogle Scholar
  • 10. Dahan D, Ekman M, Larsson-Callerfelt AK, Turczyńska K, Boettger T, Braun T, Swärd K, Albinsson S. Induction of angiotensin-converting enzyme after miR-143/145 deletion is critical for impaired smooth muscle contractility.Am J Physiol Cell Physiol. 2014; 307:C1093–C1101. doi: 10.1152/ajpcell.00250.2014CrossrefMedlineGoogle Scholar
  • 11. Wood JH, Guo J, Morrell NW, Li W. Advances in the molecular regulation of endothelial BMP9 signalling complexes and implications for cardiovascular disease.Biochem Soc Trans. 2019; 47:779–791. doi: 10.1042/BST20180137CrossrefMedlineGoogle Scholar
  • 12. Mueller TD, Nickel J. Promiscuity and specificity in BMP receptor activation.FEBS Lett. 2012; 586:1846–1859. doi: 10.1016/j.febslet.2012.02.043CrossrefMedlineGoogle Scholar
  • 13. Miller AF, Harvey SA, Thies RS, Olson MS. Bone morphogenetic protein-9: an autocrine/paracrine cytokine in the liver.J Biol Chem. 2000; 275:17937–17945. doi: 10.1074/jbc.275.24.17937CrossrefMedlineGoogle Scholar
  • 14. Neuhaus H, Rosen V, Thies RS. Heart specific expression of mouse BMP-10 a novel member of the TGF-beta superfamily.Mech Dev. 1999; 80:181–184. doi: 10.1016/s0925-4773(98)00221-4CrossrefMedlineGoogle Scholar
  • 15. Tillet E, Ouarné M, Desroches-Castan A, Mallet C, Subileau M, Didier R, Lioutsko A, Belthier G, Feige JJ, Bailly S. A heterodimer formed by bone morphogenetic protein 9 (BMP9) and BMP10 provides most BMP biological activity in plasma.J Biol Chem. 2018; 293:10963–10974. doi: 10.1074/jbc.RA118.002968CrossrefMedlineGoogle Scholar
  • 16. Tu L, Desroches-Castan A, Mallet C, Guyon L, Cumont A, Phan C, Robert F, Thuillet R, Bordenave J, Sekine A, et al.. Selective BMP-9 inhibition partially protects against experimental pulmonary hypertension.Circ Res. 2019; 124:846–855. doi: 10.1161/CIRCRESAHA.118.313356LinkGoogle Scholar
  • 17. Johnson DW, Berg JN, Baldwin MA, Gallione CJ, Marondel I, Yoon SJ, Stenzel TT, Speer M, Pericak-Vance MA, Diamond A, et al.. Mutations in the activin receptor-like kinase 1 gene in hereditary haemorrhagic telangiectasia type 2.Nat Genet. 1996; 13:189–195. doi: 10.1038/ng0696-189CrossrefMedlineGoogle Scholar
  • 18. Gallione CJ, Repetto GM, Legius E, Rustgi AK, Schelley SL, Tejpar S, Mitchell G, Drouin E, Westermann CJ, Marchuk DA. A combined syndrome of juvenile polyposis and hereditary haemorrhagic telangiectasia associated with mutations in MADH4 (SMAD4).Lancet. 2004; 363:852–859. doi: 10.1016/S0140-6736(04)15732-2CrossrefMedlineGoogle Scholar
  • 19. Gore B, Izikki M, Mercier O, Dewachter L, Fadel E, Humbert M, Dartevelle P, Simonneau G, Naeije R, Lebrin F, et al.. Key role of the endothelial TGF-β/ALK1/endoglin signaling pathway in humans and rodents pulmonary hypertension.PLoS One. 2014; 9:e100310. doi: 10.1371/journal.pone.0100310CrossrefMedlineGoogle Scholar
  • 20. Hong KH, Lee YJ, Lee E, Park SO, Han C, Beppu H, Li E, Raizada MK, Bloch KD, Oh SP. Genetic ablation of the BMPR2 gene in pulmonary endothelium is sufficient to predispose to pulmonary arterial hypertension.Circulation. 2008; 118:722–730. doi: 10.1161/CIRCULATIONAHA.107.736801LinkGoogle Scholar
  • 21. Budhiraja R, Tuder RM, Hassoun PM. Endothelial dysfunction in pulmonary hypertension.Circulation. 2004; 109:159–165. doi: 10.1161/01.CIR.0000102381.57477.50LinkGoogle Scholar
  • 22. David L, Mallet C, Keramidas M, Lamandé N, Gasc JM, Dupuis-Girod S, Plauchu H, Feige JJ, Bailly S. Bone morphogenetic protein-9 is a circulating vascular quiescence factor.Circ Res. 2008; 102:914–922. doi: 10.1161/CIRCRESAHA.107.165530LinkGoogle Scholar
  • 23. Chen H, Brady Ridgway J, Sai T, Lai J, Warming S, Chen H, Roose-Girma M, Zhang G, Shou W, Yan M. Context-dependent signaling defines roles of BMP9 and BMP10 in embryonic and postnatal development.Proc Natl Acad Sci U S A. 2013; 110:11887–11892. doi: 10.1073/pnas.1306074110CrossrefMedlineGoogle Scholar
  • 24. Levet S, Ciais D, Merdzhanova G, Mallet C, Zimmers TA, Lee SJ, Navarro FP, Texier I, Feige JJ, Bailly S, et al.. Bone morphogenetic protein 9 (BMP9) controls lymphatic vessel maturation and valve formation.Blood. 2013; 122:598–607. doi: 10.1182/blood-2012-12-472142CrossrefMedlineGoogle Scholar
  • 25. Levet S, Ouarné M, Ciais D, Coutton C, Subileau M, Mallet C, Ricard N, Bidart M, Debillon T, Faravelli F, et al.. BMP9 and BMP10 are necessary for proper closure of the ductus arteriosus.Proc Natl Acad Sci U S A. 2015; 112:E3207–E3215. doi: 10.1073/pnas.1508386112CrossrefMedlineGoogle Scholar
  • 26. Chen H, Shi S, Acosta L, Li W, Lu J, Bao S, Chen Z, Yang Z, Schneider MD, Chien KR, et al.. BMP10 is essential for maintaining cardiac growth during murine cardiogenesis.Development. 2004; 131:2219–2231. doi: 10.1242/dev.01094CrossrefMedlineGoogle Scholar
  • 27. de Lange FJ, Moorman AF, Christoffels VM. Atrial cardiomyocyte-specific expression of Cre recombinase driven by an Nppa gene fragment.Genesis. 2003; 37:1–4. doi: 10.1002/gene.10220CrossrefMedlineGoogle Scholar
  • 28. Shi J, Yang Y, Cheng A, Xu G, He F. Metabolism of vascular smooth muscle cells in vascular diseases.Am J Physiol Heart Circ Physiol. 2020; 319:H613–H631. doi: 10.1152/ajpheart.00220.2020.CrossrefMedlineGoogle Scholar
  • 29. Yan X, Liao H, Cheng M, Shi X, Lin X, Feng XH, Chen YG. Smad7 protein interacts with receptor-regulated Smads (R-Smads) to inhibit transforming growth factor-β (TGF-β)/Smad signaling.J Biol Chem. 2016; 291:382–392. doi: 10.1074/jbc.M115.694281CrossrefMedlineGoogle Scholar
  • 30. Zhang S, Fei T, Zhang L, Zhang R, Chen F, Ning Y, Han Y, Feng XH, Meng A, Chen YG. Smad7 antagonizes transforming growth factor beta signaling in the nucleus by interfering with functional Smad-DNA complex formation.Mol Cell Biol. 2007; 27:4488–4499. doi: 10.1128/MCB.01636-06CrossrefMedlineGoogle Scholar
  • 31. Limbourg A, Korff T, Napp LC, Schaper W, Drexler H, Limbourg FP. Evaluation of postnatal arteriogenesis and angiogenesis in a mouse model of hind-limb ischemia.Nat Protoc. 2009; 4:1737–1746. doi: 10.1038/nprot.2009.185CrossrefMedlineGoogle Scholar
  • 32. Halpern W, Osol G, Coy GS. Mechanical behavior of pressurized in vitro prearteriolar vessels determined with a video system.Ann Biomed Eng. 1984; 12:463–479. doi: 10.1007/BF02363917CrossrefMedlineGoogle Scholar
  • 33. Seki T, Yun J, Oh SP. Arterial endothelium-specific activin receptor-like kinase 1 expression suggests its role in arterialization and vascular remodeling.Circ Res. 2003; 93:682–689. doi: 10.1161/01.RES.0000095246.40391.3BLinkGoogle Scholar
  • 34. Seki T, Hong KH, Yun J, Kim SJ, Oh SP. Isolation of a regulatory region of activin receptor-like kinase 1 gene sufficient for arterial endothelium-specific expression.Circ Res. 2004; 94:e72–e77. doi: 10.1161/01.RES.0000127048.81744.31LinkGoogle Scholar
  • 35. Seki T, Hong KH, Oh SP. Nonoverlapping expression patterns of ALK1 and ALK5 reveal distinct roles of each receptor in vascular development.Lab Invest. 2006; 86:116–129. doi: 10.1038/labinvest.3700376CrossrefMedlineGoogle Scholar
  • 36. Li L, Miano JM, Cserjesi P, Olson EN. SM22 alpha, a marker of adult smooth muscle, is expressed in multiple myogenic lineages during embryogenesis.Circ Res. 1996; 78:188–195. doi: 10.1161/01.res.78.2.188LinkGoogle Scholar
  • 37. Dobnikar L, Taylor AL, Chappell J, Oldach P, Harman JL, Oerton E, Dzierzak E, Bennett MR, Spivakov M, Jørgensen HF. Disease-relevant transcriptional signatures identified in individual smooth muscle cells from healthy mouse vessels.Nat Commun. 2018; 9:4567. doi: 10.1038/s41467-018-06891-xCrossrefMedlineGoogle Scholar
  • 38. Scharpfenecker M, van Dinther M, Liu Z, van Bezooijen RL, Zhao Q, Pukac L, Löwik CW, ten Dijke P. BMP-9 signals via ALK1 and inhibits bFGF-induced endothelial cell proliferation and VEGF-stimulated angiogenesis.J Cell Sci. 2007; 120:964–972. doi: 10.1242/jcs.002949CrossrefMedlineGoogle Scholar
  • 39. Bidart M, Ricard N, Levet S, Samson M, Mallet C, David L, Subileau M, Tillet E, Feige JJ, Bailly S. BMP9 is produced by hepatocytes and circulates mainly in an active mature form complexed to its prodomain.Cell Mol Life Sci. 2012; 69:313–324. doi: 10.1007/s00018-011-0751-1CrossrefMedlineGoogle Scholar
  • 40. Castonguay R, Werner ED, Matthews RG, Presman E, Mulivor AW, Solban N, Sako D, Pearsall RS, Underwood KW, Seehra J, et al.. Soluble endoglin specifically binds bone morphogenetic proteins 9 and 10 via its orphan domain, inhibits blood vessel formation, and suppresses tumor growth.J Biol Chem. 2011; 286:30034–30046. doi: 10.1074/jbc.M111.260133CrossrefMedlineGoogle Scholar
  • 41. Wettschureck N, Strilic B, Offermanns S. Passing the vascular barrier: endothelial signaling processes controlling extravasation.Physiol Rev. 2019; 99:1467–1525. doi: 10.1152/physrev.00037.2018CrossrefMedlineGoogle Scholar
  • 42. Pfaltzgraff ER, Bader DM. Heterogeneity in vascular smooth muscle cell embryonic origin in relation to adult structure, physiology, and disease.Dev Dyn. 2015; 244:410–416. doi: 10.1002/dvdy.24247CrossrefMedlineGoogle Scholar
  • 43. Majesky MW. Developmental basis of vascular smooth muscle diversity.Arterioscler Thromb Vasc Biol. 2007; 27:1248–1258. doi: 10.1161/ATVBAHA.107.141069LinkGoogle Scholar
  • 44. Wasteson P, Johansson BR, Jukkola T, Breuer S, Akyürek LM, Partanen J, Lindahl P. Developmental origin of smooth muscle cells in the descending aorta in mice.Development. 2008; 135:1823–1832. doi: 10.1242/dev.020958CrossrefMedlineGoogle Scholar
  • 45. Peng T, Tian Y, Boogerd CJ, Lu MM, Kadzik RS, Stewart KM, Evans SM, Morrisey EE. Coordination of heart and lung co-development by a multipotent cardiopulmonary progenitor.Nature. 2013; 500:589–592. doi: 10.1038/nature12358CrossrefMedlineGoogle Scholar
  • 46. Hodgson J, Swietlik EM, Salmon RM, Hadinnapola C, Nikolic I, Wharton J, Guo J, Liley J, Haimel M, Bleda M, et al.. Characterization of GDF2 mutations and levels of BMP9 and BMP10 in pulmonary arterial hypertension.Am J Respir Crit Care Med. 2020; 201:575–585. doi: 10.1164/rccm.201906-1141OCCrossrefMedlineGoogle Scholar
  • 47. Long L, Ormiston ML, Yang X, Southwood M, Gräf S, Machado RD, Mueller M, Kinzel B, Yung LM, Wilkinson JM, et al.. Selective enhancement of endothelial BMPR-II with BMP9 reverses pulmonary arterial hypertension.Nat Med. 2015; 21:777–785. doi: 10.1038/nm.3877CrossrefMedlineGoogle Scholar
  • 48. Yu PB, Beppu H, Kawai N, Li E, Bloch KD. Bone morphogenetic protein (BMP) type II receptor deletion reveals BMP ligand-specific gain of signaling in pulmonary artery smooth muscle cells.J Biol Chem. 2005; 280:24443–24450. doi: 10.1074/jbc.M502825200CrossrefMedlineGoogle Scholar
  • 49. Tuder RM, Archer SL, Dorfmüller P, Erzurum SC, Guignabert C, Michelakis E, Rabinovitch M, Schermuly R, Stenmark KR, Morrell NW. Relevant issues in the pathology and pathobiology of pulmonary hypertension.J Am Coll Cardiol. 2013; 6225 SupplD4–12. doi: 10.1016/j.jacc.2013.10.025CrossrefMedlineGoogle Scholar
  • 50. Adams DJ, Quail MA, Cox T, van der Weyden L, Gorick BD, Su Q, Chan WI, Davies R, Bonfield JK, Law F, et al.. A genome-wide, end-sequenced 129Sv BAC library resource for targeting vector construction.Genomics. 2005; 86:753–758. doi: 10.1016/j.ygeno.2005.08.003CrossrefMedlineGoogle Scholar
  • 51. Luo F, Wang X, Luo X, Li B, Zhu D, Sun H, Tang Y. Invasive hemodynamic assessment for the right ventricular system and hypoxia-induced pulmonary arterial hypertension in mice [published online October 24, 2019].Jove J Vis Exp. doi: 10.3791/60090 https://www.jove.com/cn/t/60090/invasive-hemodynamic-assessment-for-right-ventricular-system-hypoxiaGoogle Scholar
  • 52. Wang S, Iring A, Strilic B, Albarrán Juárez J, Kaur H, Troidl K, Tonack S, Burbiel JC, Müller CE, Fleming I, et al.. P2Y2 and Gq/G11 control blood pressure by mediating endothelial mechanotransduction.J Clin Invest. 2015; 125:3077–3086. doi: 10.1172/JCI81067CrossrefMedlineGoogle Scholar
  • 53. Lee KJ, Czech L, Waypa GB, Farrow KN. Isolation of pulmonary artery smooth muscle cells from neonatal mice [published online October 19, 2013].Jove J Vis Exp. doi: 10.3791/50889 https://www.jove.com/t/50889/isolation-of-pulmonary-artery-smooth-muscle-cells-from-neonatal-miceGoogle Scholar
  • 54. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, Gingeras TR. STAR: ultrafast universal RNA-seq aligner.Bioinformatics. 2013; 29:15–21. doi: 10.1093/bioinformatics/bts635CrossrefMedlineGoogle Scholar
  • 55. Liao Y, Smyth GK, Shi W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features.Bioinformatics. 2014; 30:923–930. doi: 10.1093/bioinformatics/btt656CrossrefMedlineGoogle Scholar
  • 56. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2.Genome Biol. 2014; 15:550. doi: 10.1186/s13059-014-0550-8CrossrefMedlineGoogle Scholar
  • 57. Kuenne C, Guenther S, Looso M, Zhang T, Krueger M, Zhou Y, Braun T, Kim J. RNA-Seq analysis of isolated satellite cells in Prmt5 deficient mice.Genom Data. 2015; 5:122–125. doi: 10.1016/j.gdata.2015.05.013CrossrefMedlineGoogle Scholar
  • 58. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, et al.. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles.Proc Natl Acad Sci U S A. 2005; 102:15545–15550. doi: 10.1073/pnas.0506580102CrossrefMedlineGoogle Scholar