Volume 98, Issue 8 p. 1262-1266
Free Access

The transforming growth factor-β–bone morphogenetic protein type signalling pathway in pulmonary vascular homeostasis and disease

Paul D. Upton

Paul D. Upton

Division of Respiratory Medicine, Department of Medicine, University of Cambridge School of Clinical Medicine, Addenbrooke's Hospital, Cambridge, UK

Search for more papers by this author
Nicholas W. Morrell

Nicholas W. Morrell

Division of Respiratory Medicine, Department of Medicine, University of Cambridge School of Clinical Medicine, Addenbrooke's Hospital, Cambridge, UK

Search for more papers by this author
First published: 06 May 2013
Citations: 37
P. D. Upton or N. W. Morrell: Department of Medicine, University of Cambridge, School of Clinical Medicine, Box 157, Addenbrooke's Hospital, Hills Road, Cambridge CB2 0QQ, UK. Email: [email protected] or [email protected]

Symposium Report from the symposium Hypoxic pulmonary hypertension: mechanisms of pulmonary vascular change and their effect on the right ventricle, at IUPS in Birmingham on 22 July 2013.

New Findings

  • What is the topic of this review?

    This review summarises our current knowledge of dysregulated bone morphogenetic protein (BMP) and transforming growth factor-β1 (TGFβ1) signalling in pulmonary arterial hypertension.

  • What advances does it highlight?

    Reduced expression of the bone morphogenetic protein (BMP) type II receptor is common to the monocrotaline (MCT-PAH) and hypoxic rat models of pulmonary hypertension (PH). However, reduced BMP signalling and enhanced transforming growth factor-β1 (TGFβ1) signalling is observed only in MCT-PAH. Furthermore, TGFβ1 receptor blockade blocks MCT-PAH, but not hypoxic PH. Transforming growth factor-β1 inhibits BMP signalling in pulmonary artery smooth muscle cells.

Germ-line mutations in the bone morphogenetic protein type II receptor (BMPR2; BMPR-II) gene, a transforming growth factor-β (TGFβ) receptor superfamily member, cause the majority of cases of heritable pulmonary arterial hypertension (PAH). Pulmonary arterial hypertension is a subset of pulmonary hypertension (PH) disorders, which also encompass hypoxia-related lung diseases. Bone morphogenetic proteins (BMPs), via BMPR-II, activate the canonical Smad1/5/9 pathway, whereas TGFβs (TGFβ1–3) activate the Smad2/3 pathway via the ALK5 receptor. Dysregulated TGFβ1 signalling is pathogenic in fibrotic diseases. We compared two rat PH models, monocrotaline-induced PAH (MCT-PAH) and chronic normobaric hypoxia (fractional inspired O2 10%), to address whether BMPR-II loss is common to PH and permits pathogenic TGFβ1 signalling. Both models exhibited reduced lung BMPR-II expression, but increased TGFβ1 signalling and decreased BMP signalling were observed only in MCT-PAH. Furthermore, a pharmacological ALK5 inhibitor prevented disease progression in the MCT-PAH model, but not in hypoxia. In vitro studies using human pulmonary artery smooth muscle cells showed that TGFβ1 directly inhibits BMP–Smad signalling. In conclusion, BMPR-II loss is common to the hypoxic and MCT-PAH models, but systemic ALK5 inhibition is effective only in the MCT model, highlighting a specific role for TGFβ1 in vascular remodelling in MCT-PAH, potentially via direct inhibition of BMP signalling.

Pulmonary hypertension: classification and genetics

The Dana Point classification of pulmonary hypertension (PH) defines pulmonary arterial hypertension (PAH) as a subset of the different forms of clinical PH (Simonneau et al. 2009). Pulmonary arterial hypertension can arise as a primary disease, generally with a heritable component, or may be associated with particular drugs and toxins, connective tissue diseases and human immunodeficiency virus or schistosomiasis infection. Haemodynamically, PAH is defined as a resting mean pulmonary artery pressure (PAP) of >25 mmHg, pulmonary vascular resistance (PVR) >3 Wood units and pulmonary wedge pressure <15 mmHg in the absence of other causes of PH (Archer et al. 2010). A separate clinical subset of this classification constitutes PH owing to lung diseases and/or hypoxia, including chronic obstructive pulmonary disease, interstitial lung disease and chronic high-altitude exposure (Simonneau et al. 2009).

Pulmonary arterial hypertension is a rare disease, with 2.4–7.6 cases per million affected, depending upon the cohorts reported (Archer et al. 2010). Although rare, heritable PAH is relatively well studied, because a genetic defect has been identified. Germ-line mutations in the gene (BMPR2) encoding the bone morphogenetic protein type II receptor (BMPR-II) are associated with heritable PAH and with approximately 25% of patients with apparently sporadic PAH (Machado et al. 2009). The majority of BMPR2 mutations are predicted to cause haploinsufficiency and thus, reduced cell surface BMPR-II (Machado et al. 2009). Moreover, reduced BMPR-II may be important in other PH diseases, such as PH due to heart defects (Atkinson et al. 2002). Although BMPR2 mutations follow autosomal-dominant inheritance, only 20% of mutation carriers develop PAH, implying that additional environmental or genetic factors may precipitate disease and possibly drive disease progression (Machado et al. 2009). In the context of multifactorial events, such as inflammation, several factors could elicit inappropriate cellular responses on the background of reduced BMPR-II and/or may directly target bone morphogenetic protein (BMP) signalling. The contribution of inflammation as a ‘second hit’ in the pathogenesis of PAH has been widely proposed, with several studies exploring inflammatory mediators and responses (Dorfmuller et al. 2003). Here, we address the potential role of the cytokine, transforming growth factor-β1 (TGFβ1).

Bone morphogenetic protein and TGFβ1 signalling

Bone morphogenetic protein type II receptor is a receptor belonging to the TGFβ receptor superfamily. The ligands include the TGFβs, BMPs, growth differentiation factors, activins and inhibins. The receptors signal as heteromeric complexes of type I and type II receptors, such that a dimeric BMP ligand activates a complex comprising two type I receptors (ALKs) and two type II receptors. In smooth muscle cells, BMPR-II primarily signals with ALK3 or ALK6 in response to BMP2 or BMP4, whereas BMP6 and BMP7 activate ALK2 and BMPR-II. Transforming growth factor-β1 signals via ALK5 in a complex with the TGFβ type II receptor (TGFβR-II). Upon receptor activation, the type I receptor phosphorylates the canonical substrate proteins, the receptor Smads, at C-terminal consensus motifs (Bertolino et al. 2005). The BMP receptors typically activate Smads 1/5/9, whereas TGFβ receptors activate Smad2/3. Once activated, the BMP and TGFβ Smads translocate to the nucleus in a complex with the common Smad, Smad4, to regulate the transcription of various target genes directly. The BMP–Smad target genes include the dominant negative helix–loop–helix transcription factors, the inhibitors of differentiation (ID1–4; Id1–4 in mouse). The Smad binding element (BRE) of the ID1 promoter has been characterized and is used for Smad reporter assays (Korchynskyi & ten Dijke, 2002). Activated Smad2/3 recognizes promoter sequences with a core CAGA motif, inducing genes such as plasminogen activator inhibitor-1 (SERPINE1) and connective tissue growth factor (CTGF). In rare cases, Smad1 and Smad9 mutations have been identified in PAH patients (Nasim et al. 2011). Furthermore, mice with targeted deletion of Smad1 in the pulmonary vasculature develop PAH, implicating perturbed BMP–Smad signalling in the pathogenesis of PAH (Han et al. 2013). Indeed, we have previously reported reduced phospho-Smad1 staining in lung tissues from patients with idiopathic and heritable PAH (Yang et al. 2005).

Bone morphogenetic protein type II receptor and animal models of pulmonary hypertension

Robust and reproducible mouse models of BMPR-II-associated PAH have proved difficult to generate. The BMPR-II null homozygotes are embryonic lethal, failing to develop a vascular plexus, whereas heterozygote littermates do not develop robust PAH. This is not surprising if the penetrance of heritable PAH is only 20%, and a secondary insult may be needed. Accordingly, BMPR-II heterozygous null mice (Bmpr2+/−) develop PAH if challenged with serotonin and hypoxia (Long et al. 2006). The lethality of the homozygous knockout has been circumvented using tissue-specific promoters. Cre-dependent BMPR-II ablation in the endothelium (Hong et al. 2008) or targeted overexpression of a dominant negative BMPR-II in smooth muscle cells (West et al. 2005) causes spontaneous PAH in mice. Furthermore, mice expressing a hypomorphic Bmpr2 allele with exon 2 deletion show increased susceptibility to hypoxic PH (Frank et al. 2008).

Dysregulation of BMP signalling in hypoxia differs between mouse models and rat models, partly because mice do not readily undergo the degree of pulmonary vascular remodelling observed in rats, so PH in mice may be more related to vasoconstriction. Although BMP2 and BMP4 expression increase in hypoxia, BMP4+/lacz mice are protected from PH, whereas BMP2+/lacz mice develop more severe PH, implying that these BMPs have opposing functions in the lung (Anderson et al. 2010). Intriguingly, Smad1/5/8 signalling slightly increases, while Id1 and Id3 expression dramatically increases after 4 days of hypoxia (Anderson et al. 2010; Lowery et al. 2010). Although the hypoxic Id1 response is attenuated in the BMP4+/lacz mouse, Id1 null mice do not exhibit altered responses to hypoxia, suggesting that Id1 is dispensable or that Id3 may compensate (Lowery et al. 2010; Anderson et al. 2010). Also, Id genes are targets of other pathways, so Id responses may not represent BMP signalling in this model. In hypoxic rats, our group and others have reported a sustained reduction of BMPR-II from 7 days up to 3 weeks (Takahashi et al. 2006; Long et al. 2009). However, Smad1/5/9 activation is unchanged, whereas p38 MAP kinase activation is reduced, so non-Smad pathways may be important in hypoxia (Takahashi et al. 2006; Long et al. 2009). The central importance of BMPR-II loss has been reinforced by a study showing that adenoviral vector delivery of BMPR-II attenuates hypoxic PH in rats (Reynolds et al. 2007).

Rats develop progressive PAH after a single injection of the plant alkaloid, monocrotaline (MCT-PAH). This causes acute endothelial damage and lung inflammation over 2–3 days, followed by the progressive development of PAH with robust vascular remodelling. Our group and others have shown that MCT-PAH in rats is associated with reduced BMPR-II, BMP type I receptors, BMP Smad proteins and Id gene transcription (Morty et al. 2007; Long et al. 2009). In contrast to hypoxia, the activities of the extracellular signal-related kinases (ERK1/2) and p38 mitogen-activated protein kinase pathways are not affected in this model (Aguirre et al. 2000). Intriguingly, Morty et al. (2007) demonstrated functional differences between pulmonary artery smooth muscle cells (PASMCs) isolated from rats with MCT-PAH and normal control animals, implying that epigenetic changes, transdifferentiation or replacement with a different cell type may occur. Of note, adenoviral BMPR-II delivery into the rat pulmonary vasculature at day 14 post-MCT did not improve the pulmonary haemodynamics or pathological changes in MCT-PAH (McMurtry et al. 2007). Although this presents a case against BMPR2 gene therapy in PAH, the authors highlighted the potential limitations of the single dose delivery used. Collectively, these studies highlight the potential differences in signalling perturbations that may occur in different PAH models, yet the central observation of reduced BMPR-II expression appears common (Fig. 1).

Details are in the caption following the image

Perturbations of the bone morphogenetic protein (BMP) and transforming growth factor-β1 (TGFβ1) signalling pathways reported in the monocrotaline-treated rat model of pulmonary arterial hypertension (MCT-PAH; top panel) and the hypoxic pulmonary hypertension model (hypoxic PH; bottom panel)
Thecanonical BMP–Smad pathway is attenuated in MCT-PAH, with bone morphogenetic protein type II receptor (BMPR-II) levels being significantly reduced. Furthermore, increased TGFβ1 signalling, manifested by increased Smad2/3 activation, is also present in this model. The TGF–Smad3 pathway exerts direct negative regulation of the BMP–Smad1/5/8 pathway, which results in reduced BMP-mediated transcriptional activity. This inflammation-based model has exhibited the potential for therapies targeting TGF ligands or TGF receptor kinase activity. In hypoxic PH, BMPR-II expression is again reduced. However, canonical Smad signalling is not affected, whereas reduced BMP-mediated mitogen-activated protein kinase signalling is evident. Although evidence suggests that TGFβ1 expression increases in the lung, Smad2/3 signalling is unchanged, so no inhibition of BMP signalling is observed. Anti-TGF therapies are not effective in this model, suggesting that TGFβ1 is not instrumental in this model.

Transforming growth factor-β and pulmonary hypertension

Transforming growth factor-β1 has key physiological roles in multiple processes, including embryonic development, inflammation, angiogenesis, immune cell function and wound healing. However, excessive TGFβ1 production is associated with fibrotic diseases of several organs, including the lung (Coward et al. 2010) and kidney (López-Hernández & López-Novoa, 2012). In our study comparing hypoxic PH and MCT-PAH in rats, intraperitoneal administration of an ALK5 inhibitor prevented MCT-PAH development, but did not affect hypoxic PH (Long et al. 2009). Furthermore, Smad3 signalling was not increased in hypoxic PH, but significantly increased in the lungs of rats with MCT-PAH (Long et al. 2009). This suggests that TGFβ1 may contribute more to the pathogenesis of MCT-PAH than to hypoxic PH, and this is reinforced by two additional studies showing that ALK5 inhibition attenuates disease progression in MCT-PAH (Zaiman et al. 2008; Thomas et al. 2009). We studied PASMCs isolated from patients harbouring disease-causing BMPR2 mutations, finding that the BMPR2 mutant cells exhibit an enhanced mitogenic response to TGFβ (Morrell et al. 2001). Closer examination demonstrated that loss of BMPR-II confers resistance to the normal growth-inhibitory effects of TGFβ (Davies et al. 2012). This is not a feature of epigenetic changes or transdifferentiation, because this effect is also revealed in control PASMCs after silencing of BMPR-II with short interfering RNAs and is evident in PASMCs isolated from Bmpr2+/− mice without disease, whereas TGFβ inhibits the proliferation of PASMCs from wild-type littermates (Davies et al. 2012).

Although TGFβ responses are altered on the background of BMPR2 mutations, little is known regarding any potential interaction of the TGFβ and BMP pathways in PASMCs. Studies of fibrotic diseases have demonstrated functional inhibition between the TGFβ1 and BMP pathways. Bone morphogenetic protein 4 inhibits TGFβ1 signalling in lung fibroblasts (Pegorier et al. 2010), and BMP7 has an antifibrotic role via TGFβ1 inhibition in the kidney (Izumi et al. 2006). We have investigated whether the BMP and TGFβ1 pathways directly interact in PASMCs, primarily focusing on Smad activation and selected transcriptional targets. We present unpublished data demonstrating that TGFβ1 inhibits BMP signalling in PASMCs and may directly promote the progression of PAH. This reinforces the rationale for anti-TGFβ strategies in PAH, but these may not be applicable to hypoxic PH.


Additional information

Competing interests

None declared.


PDU is supported by a British Heart Foundation Programme Grant (RG/08/002/2718) awarded to NWM.