Pulmonary hypertension: therapeutic targets within the serotonin system

Abstract

Pulmonary arterial hypertension (PAH) is characterized by a sustained and progressive elevation in pulmonary arterial pressure and pulmonary vascular remodelling leading to right heart failure and death. Prognosis is poor and novel therapeutic approaches are needed. The serotonin hypothesis of PAH originated in the 1960s after an outbreak of the disease was reported among patients taking the anorexigenic drugs aminorex and fenfluramine. These are indirect serotinergic agonists and serotonin transporter substrates. Since then many advances have been made in our understanding of the role of serotonin in the pathobiology of PAH. The rate-limiting enzyme in the synthesis of serotonin is tryprophan hydroxylase (Tph). Serotonin is synthesized, through Tph1, in the endothelial cells of the pulmonary artery and can then act on underlying pulmonary arterial smooth muscle cells and pulmonary arterial fibroblasts in a paracrine fashion causing constriction and remodelling. These effects of serotonin can be mediated through both the serotonin transporter and serotonin receptors. This review will discuss our current understanding of ‘the serotonin hypothesis' of PAH and highlight possible therapeutic targets within the serotonin system.

Introduction

Pulmonary arterial hypertension (PAH) is characterized by constriction and remodelling of the pulmonary arterial bed, leading to a progressive increase in pulmonary arterial pressure, right heart failure and death. Idiopathic PAH (iPAH) (formerly known as primary PAH) describes a form of the disease for which there is no known cause. PAH can also occur secondary to global hypoxia as seen in patients with chronic obstructive pulmonary disease or following chronic exposure to high altitude. Various drugs and toxins have also been associated with the development of PAH, as has HIV infection (Simonneau et al., 2004). Familial PAH (fPAH) transmits as an autosomal dominant trait that exhibits genetic anticipation but also markedly reduced penetrance (20%) (Loyd et al., 1995). The primary genetic defect of fPAH, (present in ∼70% of cases), is a mutation in the gene-encoding bone morphogenetic protein receptor type II (BMPR-II), a member of the transforming growth factor-β (TGF-β) superfamily (Lane et al., 2000; Machado et al., 2001). Genetic heterogeneity may occur in some cases of severe unexplained PAH. For example, mutations in the TGF-β receptor, activin receptor-like kinase-1, have been reported in families with hereditary haemorrhagic telangiectasia and severe pulmonary hypertension (Trembath et al., 2001). The true prevalence of BMPR-II mutations in iPAH is unknown, with reports ranging from 10 to 40% of patients (Thomson et al., 2000; Sankelo et al., 2005; Baloira et al., 2008; Fujiwara et al., 2008). The cause of the variable phenotypic expression of PAH among carriers of mutated BMPR-II genes and patients is unclear, and likely related to additional environmental and/or genetic modifiers. Increased activation of the serotonin system has been proposed as a ‘second hit' risk factor.

The ‘serotonin hypothesis of PAH' arose in the 1960s when patients taking the anorexigen aminorex fumarate were associated with an increased risk of developing PAH. Subsequently, in the early 1980s French investigators reported a cluster of PAH among patients using fenfluramine derivatives (Abenhaim et al., 1996; Kramer and Lane, 1998). Both aminorex and fenfluramine are serotonin transporter (SERT) substrates and increase extracellular concentrations of serotonin (Rothman et al., 1999). Serotonin is thought to mediate PAH by promoting both vasoconstriction and remodelling of the pulmonary vasculature (MacLean et al., 2000). Serotonin induces proliferation of both pulmonary arterial fibroblasts and pulmonary arterial smooth muscle cells (Lee et al., 1994; Welsh et al., 2004). This results in a thickening of the medial layer and a narrowing of the lumen of the pulmonary artery and contributes to the pulmonary vascular remodelling associated with PAH. Clinically, pulmonary arterial smooth muscle cells from PAH patients proliferate faster than those from controls when stimulated with serotonin, and elevated circulating peripheral serotonin has been associated with the development of PAH (Herve et al., 1995; Eddahibi et al., 2001). Experimentally, exogenous serotonin can potentiate the development of hypoxia-induced PAH in rats (Eddahibi et al., 1997), and inhibition of serotonin receptors or SERT can inhibit the development of PAH in animal models (Keegan et al., 2001; Hironaka et al., 2003; Marcos et al., 2003; Guignabert et al., 2005). Fawn hooded rats, which have an inherited platelet storage defect to serotonin and increased expression of SERT, have increased susceptibility to PAH (Morecroft et al., 2005).

Although many advances have been made in recent years, the pathophysiology of PAH is still not fully understood. Prognosis is poor and current therapies are not effective in the long-term. This review will examine our current understanding of the role of serotonin in the pathobiology of PAH, and discuss possible therapeutic targets.

Tryptophan hydroxylase 1

Tryptophan hydroxylase (Tph) is the rate-limiting enzyme in the synthesis of serotonin. Recently, it was demonstrated that there are two isoforms of Tph, now termed Tph1 and Tph2. Tph1 is present mainly in the gut and mediates generation of serotonin in the periphery, whereas Tph 2 is present exclusively in the central nervous system (Walther and Bader, 2003; Walther et al., 2003a). Expression of the Tph1 gene is increased in lungs and pulmonary arterial endothelial cells from patients with iPAH. Medium collected from human pulmonary arterial endothelial cell cultures induces marked proliferation of human pulmonary arterial smooth muscle cells, an effect reduced in the presence of the Tph inhibitor p-chlorophenylalanine. This effect is greater when using pulmonary artery endothelial cells and pulmonary arterial smooth muscle cells from iPAH patients than when using cells from control subjects (Eddahibi et al., 2006). Taken together these data suggest that the increased Tph1 expression/activity in pulmonary artery endothelial cells, and subsequent paracrine effects of serotonin on pulmonary arterial smooth muscle cells, may play a role in PAH (see Figure 1 for summary). Moreover, there is also evidence to suggest that Tph1 is involved in PAH secondary to hypoxia. Hypoxia-induced increases in both right ventricular pressure and pulmonary vascular remodelling are severely ablated in mice deficient in Tph1 indicating that de novo synthesis of serotonin is essential for the development of hypoxia-induced PAH (Izikki et al., 2007; Morecroft et al., 2007). Both hypoxia and mechanical stretch have been shown to increase Tph1 expression and serotonin release in rabbit lung (Pan et al., 2006). These observations suggest that chronic hypoxia itself may induce Tph1 synthesis in the pulmonary arterial endothelium, which can then act on underlying pulmonary arterial smooth muscle cells and pulmonary arterial fibroblasts in a paracrine fashion. Mice lacking Tph1 have a reduced risk of thrombosis and thromboembolism (Walther et al., 2003b) and so targeting Tph1 for the treatment of PAH is of interest.

Figure 1.

Serotonin is synthesized in the pulmonary arterial endothelial cells by tryptophan hydroxylase1 (Tph1). Serotonin can then influence pulmonary vascular smooth muscle proliferation and/or contraction via activity at the serotonin transporter (SERT) and serotonin receptors (particularly the 5-HT_1B receptor in humans). Intracellular accumulation of serotonin and activation of the 5-HT_1B receptor may induce reactive oxygen species (ROS), activation of Rho-kinase (ROCK), leading to phosphorylation and nuclear translocation of extracellular-regulated kinase (ERK)1/2. Once inside the nucleus phosphorylated ERK1/2 can increase transcription of nuclear growth factors and mediate cellular proliferation. Activation of the bone morphogenetic receptor type II (BMPR-II) leads to signalling through the Smad 1/5/8 pathway. Smads1, 5 and 8 must dimerize with Smad 4 to enter the nucleus where they can activate anti-proliferative transcription factors. Serotonin may antagonize BMPR-II signalling as phosphorylated ERK1/2 can phosphorylate the linker region of Smad 1and inhibit nuclear translocation.

Serotonin transporter

The gene-encoding SERT is located on chromosome 17q11.2 and has a variant in the promoter region. This polymorphism affects SERT expression and function. The long (L) allele induces an increased rate of SERT transcription over the short (S) allele (Lesch et al., 1996). The LL genotype was found to be more prevalent in a small sample of primary PAH patients than in the control group (Eddahibi et al., 2001). However, other groups studying larger cohorts of PAH patients suggest that variation of the SERT gene alone is unlikely to mediate susceptibility to PAH, although PAH patients with the LL genotype may present earlier than those without (Machado et al., 2006; Willers et al., 2006). The LL genotype has also been associated with exaggerated PAH in patients with chronic obstructive lung disease (Eddahibi et al., 2003), and with an increased risk of developing PAH at high altitudes (Long et al., 2002). Further evidence for a role for SERT in PAH exists in animal models of the disease. Mice overexpressing SERT (SERT+ mice) develop increased pulmonary pressures and are more susceptible to hypoxia-induced PAH, whereas mice deficient for the SERT are less susceptible (Eddahibi et al., 2000; MacLean et al., 2004; Guignabert et al., 2006). Inhibition of SERT protects against PAH secondary to both hypoxia (Marcos et al., 2003) and monocrotaline injection (Guignabert et al., 2005). Moreover, drugs such as aminorex, dexfenfluramine and methamphetamine, all of which are SERT substrates and compete for SERT, mediating release of serotonin via SERT, have been associated with an increased risk of developing PAH (Rothman et al., 1999; Chin et al., 2006). Indeed, dexfenfluramine-induced pulmonary vascular remodelling is exaggerated in SERT+ mice compared with wild-type controls (Dempsie et al., 2008). There is evidence to suggest that SERT may mediate the proliferative effects of serotonin, as inhibition of SERT reduces proliferation of human and bovine pulmonary arterial smooth muscle cells (Lee et al., 1994; Marcos et al., 2004), and rat pulmonary arterial fibroblasts (Welsh et al., 2004). Moreover, both serum- and serotonin-induced proliferation is increased in pulmonary arterial smooth muscle cells from PAH patients compared with controls, and this is due to increased expression of SERT (Eddahibi et al., 2001). Although the precise mechanism by which SERT mediates proliferation of pulmonary arterial smooth muscle cells and pulmonary arterial fibroblasts still remains unclear, evidence suggests that serotonin transport into the cell via SERT results in the production of reactive oxygen species (ROS). ROS then mediates phosphorylation and /or nuclear translocation of extracellular signal-regulated kinase 1/2 (ERK1/2), which leads to activation of transcription factors, such as GATA 4, and cellular proliferation (Lee et al., 1999, 2001a; Suzuki et al., 2003; Liu et al., 2004; Lawrie et al., 2005).

Classical vasodilators cannot be used to treat PAH due to their systemic hypotensive effects. To avoid systemic hypotension, pharmacological targets for PAH need to be pulmonary-specific. Serotonin-induced proliferation of systemic SMCs is through the 5-HT_2A receptor (Sharma et al., 1999). Hence, serotonin-induced proliferation mediated by SERT activity is a pulmonary-specific mechanism, making SERT an attractive target for PAH. However, we have observed increased serotonin-induced contraction in the presence of SERT inhibitors in isolated pulmonary arteries from animals showing increased SERT expression, that is, fawn hooded rats and SERT+ mice (Morecroft et al., 2005; Dempsie et al., 2008). Therefore, if SERT inhibitors are considered for the treatment of PAH, these may have to be administered in conjunction with serotonin receptor antagonists. This is discussed further below.

The 5-HT_1B receptor

The 5-HT_1B receptor mediates serotonin-induced constriction in human small and large pulmonary arteries (MacIntyre et al., 1992; MacLean et al., 1996; Morecroft et al., 1999). Moreover, pulmonary arteries removed from PAH patients show increased expression of the 5-HT_1B receptor (Launay et al., 2002). The 5-HT_1B receptor is a Gi-coupled receptor and we have previously demonstrated that 5-HT_1B receptor-mediated vasoconstriction can be increased through pharmacological synergy (MacLean, 1999). For example hypoxia, reduced cGMP levels and raised vascular tone all markedly increase 5-HT_1B-mediated responses in pulmonary arteries. As all these stimuli exist in PAH, it is likely that there will be increased 5-HT_1B-mediated responses in pulmonary arteries from PAH lungs. Experimentally, the 5-HT_1B receptor has been shown to be involved in the development of PAH in rats and mice exposed to chronic hypoxia, as well as in the increased contractile response to serotonin observed in the pulmonary arteries from these animals (Keegan et al., 2001). 5-HT_1B receptor expression is enhanced in pulmonary arteries from monocrotaline-treated rats (Wang et al., 2001), and in a pig model of pulmonary hypertension (Rondelet et al., 2003). Serotonin mediates vasoconstriction of most systemic resistance arteries via the 5-HT_2A receptor (Hoyer et al., 1994). Therefore, the 5-HT_1B receptor could be a pulmonary selective target for PAHs. The 5-HT_1B receptor and SERT activity have been shown to co-operate in pulmonary vascular contraction of rat pulmonary arteries (Morecroft et al., 2005). In pulmonary resistance arteries isolated from fawn-hooded rats, the vascular response to serotonin can be enhanced by the SERT inhibitors fluoxetine and citalopram. However, in the presence of the combined 5-HT_1B receptor/SERT antagonist LY393558 the contractile response to serotonin is reduced. In pulmonary resistance arteries from both normoxic and chronically hypoxic rats the 5-HT_1B receptor antagonist SB224289 and SERT inhibitor fluoxetine both inhibited the serotonin-induced contractile response. However, LY393558 was the most potent inhibitor of serotonin-induced constriction and there was synergy between the effects of fluoxetine and SB224289 when given simultaneously. Synergy between the 5-HT_1B receptor and SERT has also been described in proliferation of bovine and human pulmonary arterial smooth muscle cells (Liu et al., 2004; Lawrie et al., 2005). Therefore, dual blockade of the 5-HT_1B receptor and SERT would likely be the optimal therapeutic approach. Indeed unpublished data from our laboratory suggests that LY393558 is more effective than citalopram at preventing PAH secondary to SERT overexpression and hypoxia.

The 5-HT_2A receptor

The 5-HT_2A receptor has also been implicated in PAH. Antagonism of the 5-HT_2A receptor inhibits monocrotaline-induced PAH in mice (Hironaka et al., 2003), and also inhibits serotonin-induced pulmonary vasoconstriction in vessels from both normoxic and hypoxic rats (Morecroft et al., 2005; Cogolludo et al., 2006). Moreover, the 5-HT_2A receptor mediates serotonin-induced proliferation of rat pulmonary arterial fibroblasts (Welsh et al., 2004). The 5-HT_2A receptor is present in human pulmonary arteries; however, it only contributes to vasoconstriction when serotonin concentrations are much higher than the physiological range (Morecroft et al., 1999). Moreover, the 5-HT_2A receptor also mediates vasoconstriction of the systemic circulation. The 5-HT_2A receptor antagonist ketanserin has proved to be clinically effective in the treatment of systemic hypertension. Hence, there is no specificity for the pulmonary circulation and the systemic effects have limited its use in either primary or secondary pulmonary hypertension, where it fails to improve pulmonary haemodynamics significantly (Frishman et al., 1995).

The 5-HT_2B receptor

The 5-HT_2B receptor is upregulated in pulmonary arteries removed from pulmonary hypertensive patients, and the development of hypoxia-induced PAH is inhibited in mice deficient for the 5-HT_2B receptor (Launay et al., 2002). This receptor may also control plasma serotonin levels in mice (Callebert et al., 2006). However, there is currently no evidence of 5-HT_2B receptor-mediated constriction of human pulmonary arteries or 5-HT_2B receptor-mediated proliferation of human pulmonary arterial smooth muscle cells or pulmonary arterial fibroblasts. In addition, loss of 5-HT_2B receptor function may pre-dispose to fenfluramine-associated PAH in man (Blanpain et al., 2003).

Serotonin signalling in PAH

Much progress has been made in recent years in identifying mechanisms downstream of serotonin that are involved in the pathogenesis of PAH. Unravelling the serotonin-induced signalling pathways is important to identify further possible therapeutic targets. As discussed above, serotonin transport through SERT has been shown to be involved in pulmonary arterial smooth muscle cell and pulmonary arterial fibroblast proliferation (Lee et al., 1994; Marcos et al., 2004; Welsh et al., 2004), and this involves ROS and ERK1/2 activation (Figure 1). Studies on human pulmonary arterial smooth muscle cells provide evidence that it may be the breakdown of serotonin by monoamine oxidase which results in the generation of ROS (Lawrie et al., 2005), whereas studies on bovine pulmonary arterial smooth muscle cells suggest serotonin induces ROS through activation of NADPH oxidase (Lee et al., 1998; Liu and Folz, 2004). Serotonin-induced activation of NADPH oxidase has been postulated to occur via activation of small GTPase proteins such as Rac-1 (Lee et al., 1998). In platelets, internalized serotonin is transamidated to small GTPases by transglutaminases, rendering these GTPases constitutively active (Walther et al., 2003b). The mechanism by which ERK1/2 is phosphorylated and translocated to the nucleus is also controversial and may be species-specific. In bovine pulmonary arterial smooth muscle cells, ROS has been proposed to mediate phosphorylation of ERK1/2, whereas activation of Rho-kinase (ROCK) via the 5-HT_1B receptor mediates nuclear translocation of phosphorylated ERK1/2 (Lee et al., 1999, 2001a; Liu et al., 2004). Studies in human pulmonary arterial smooth muscle cells show phosphorylation of ERK1/2 to occur via activation of the 5-HT_1B receptor, with ROS responsible for nuclear translocation of phosphorylated ERK1/2 (Lawrie et al., 2005). After phosphorylated ERK1/2 has been translocated to the nucleus it can increase DNA binding of transcription factors such as GATA-4, Egr-1 and Elk-1, and thus increase expression of proteins, which are involved in cellular proliferation (Liu et al., 2004). One such protein is S100A4/Mts1, a calcium-binding protein that is involved in proliferation of human pulmonary arterial smooth muscle cells (Lawrie et al., 2005). S100A4/Mts1 is upregulated in the neointima in remodelled vessels from patients with PAH and a subset of mice overexpressing S100A4/Mts1 develop occlusive pulmonary arterial lesions similar to those seen in patients with PAH (Greenway et al., 2004).

Under hypoxic conditions it would appear that serotonin-induced proliferation of pulmonary arterial fibroblasts is mediated via the p38 mitogen-activated protein (MAP) kinase pathway, rather than the ERK1/2 pathway (Welsh et al., 2004). The p38 MAP kinase pathway has frequently been associated with hypoxic-induced proliferation of pulmonary arterial fibroblasts and has recently been linked to activation of the transcription factor hypoxia-inducible factor-1α (Das et al., 2001; Welsh et al., 2001, 2006). It is interesting to note that hypoxia mediates upregulation of the α and γ isoforms of p38 MAP kinase and also of hypoxia-inducible factor-1α in the pulmonary but not the systemic circulation, thus providing possible future pulmonary selective targets for the treatment of PAH (Welsh et al., 2006; Mortimer et al., 2007).

ROCK has been implicated in both pulmonary vascular contraction and pulmonary vascular remodelling (Fukumoto et al., 2007). Inhibition of ROCK attenuates chronic hypoxic-induced PAH in mice, chronic hypoxic, monocrotaline and high flow-induced PAH in rats, decreases susceptibility to PAH in fawn-hooded rats, and is involved in the beneficial effect of sildenafil on PAH (Abe et al., 2004; Fagan et al., 2004; Guilluy et al., 2005; Hyvelin et al., 2005; Nagaoka et al., 2006; Li et al., 2007a). In two small groups of PAH patients, the ROCK inhibitor fasudil decreased pulmonary arterial pressure and pulmonary vascular resistance. Although decreased systemic vascular resistance was also reported in one of these studies, neither study reported systemic hypotension (Fukumoto et al., 2005; Ishikura et al., 2006). Interestingly, conventional pulmonary vasodilators such as prostacyclin do not mediate vasodilatation of the pulmonary vasculature through inhibition of ROCK (Abe et al., 2005), and the combination of ROCK inhibitor fasudil with beroprost, an oral prostacyclin analogue, was more effective than each drug alone in attenuating monocrotaline-induced PAH in rats (Tawara et al., 2007). The ROCK pathway can also be inhibited by members of the statin family of drugs, and various statins have beneficial effects in animal models of PAH. Atorvastatin, which can inhibit monocrotaline-induced PAH (Laudi et al., 2007), also inhibits serotonin mediated activation of the ROCK pathway and subsequent nuclear translocation of phosphorylated ERK1/2 in bovine PASMCs (Li et al., 2007b). Simvastatin can inhibit chronic hypoxic-induced PAH and also hypoxic-induced ROCK expression (Girgis et al., 2007). Statins are widely used as cholesterol-lowering agents, and the safety and tolerability of these drugs have been established making statin treatment of PAH an interesting possibility.

Dexfenfluramine and PAH

The original ‘serotonin hypothesis of PAH' was derived from the observation that obese patients using appetite suppressants such as aminorex and dexfenfluramine were at increased risk of developing PAH (Abenhaim et al., 1996; Kramer and Lane, 1998). The appetite-suppressant effect of fenfluramine derivatives is thought to be dependent upon inhibition of neuronal serotonin reuptake, increased serotonin release and subsequent serotonin receptor stimulation (Caccia et al., 1993; Rothman et al., 1999). As neurones, platelets, pulmonary endothelial and smooth muscle cells share the same SERT encoded by a single gene (Lesch et al., 1996), one school of thought is that it is this action of dexfenfluramine that promotes PAH (Celada et al., 1994). Indeed, increased serotonin plasma levels have been observed during treatment with fenfluramine derivatives (Celada et al., 1994). In addition, the metabolite, nor-dexfenfluramine is an agonist at 5-HT_2A, 5-HT_2B and 5-HT_2C receptors (Rothman and Baumann, 2002). However, dexfenfluramine also has direct effects on pulmonary vessels including inhibition of potassium channels (Weir et al., 1996), increased intracellular calcium (Reeve et al., 1999), vasoconstriction (albeit with very low potency) (Higenbottam et al., 1999; Patnaude et al., 2000) and proliferation (Lee et al., 2001b). Intriguingly, attempts to elucidate the mechanism of dexfenfluramine-induced PAH in animal models had produced varied results, with dexfenfluramine being reported not only to exacerbate PAH (Launay et al., 2002) but also to protect against both hypoxia- and monocrotaline-induced PAH (Mitani et al., 2002; Rochefort et al., 2006) However, we have now shown that although dexfenfluramine induced PAH in wild-type mice, dexfenfluramine had no effects on the development of PAH in mice deficient in Tph1, the rate limiting enzyme in the synthesis of peripheral serotonin (Dempsie et al., 2008). Hence, we have definitively shown that dexfenfluramine mediates the development of PAH through a peripheral serotinergic mechanism rather than through non-serotinergic effects. This provides further evidence that peripheral serotonin plays a causative role in the development of PAH. Moreover, dexfenfluramine inhibited hypoxic-induced proliferation of pulmonary arterial fibroblasts derived from SERT+ mice, through inhibition of p38 MAP kinase, providing a possible mechanism for the protective effects of dexfenfluramine on hypoxic-induced PAH (Dempsie et al., 2008). As p38 MAP kinase has also been linked to vascular proliferation associated with monocrotaline-induced PAH (Lu et al., 2004), and SERT is overexpressed in pulmonary arteries from monocrotaline-treated rats (Laudi et al., 2007) our results also provide a potential mechanism for the protective effects of dexfenfluramine on PAH secondary to monocrotaline injection. Understanding the pharmacology of dexfenfluramine is essential to identify other drugs, which may also be risk factors for PAH. This is of current importance in light of recent reports on the increased abuse of amphetamine-like stimulants, such as methamphetamine, which have also been associated with the development of PAH (Chin et al., 2006).

Serotonin and the BMPR-II receptor

Bone morphogenetic proteins are members of the TGF-β family of cytokines and act on a variety of cell types to regulate growth, differentiation and apoptosis. The BMPR-II receptor is a member of the TGF-β superfamily type II receptors, and forms a complex with an associated type I receptor on ligand stimulation (Miyazono et al., 2005; Morrell, 2006). Activated type I receptors can then phosphorylate and thus activate the Smad-signalling pathway. Bone morphogenetic proteins signal through Smads 1,5 and 8 which must dimerize with Smad 4 to enter the nucleus, and regulate transcription of target genes (Massague et al., 2005). Pulmonary arterial smooth muscle cells from patients with BMPR-II mutations have a reduced capacity to activate Smads 1 5 and 8 and BMP-4 has a reduced ability to suppress proliferation in these cells (Yang et al., 2005). Thus a decrease in signalling through the BMPR-II receptor may increase pulmonary arterial smooth muscle cell proliferation and mediate PAH. However, as mentioned previously, disease penetrance in carriers of a mutation in the BMPR-II gene is low. Therefore, it has been postulated that exposure to another risk factor is necessary for development of PAH (Newman et al., 2004). In line with this, male mice deficient in BMPR-II-signalling (BMPR-II+/− mice) do not develop spontaneous PAH, but rather need an added stimulus to uncover a PAH phenotype. Indeed, an infusion of serotonin has been shown to uncover a PAH phenotype in BMPR-II+/− mice and this was associated with a serotonin inhibition of phosphorylation of Smad 1/5/8. Expression of the inhibitor of DNA-binding 3 mRNA in response to BMP2 was lower in BMPR-II+/− mice than their wild-type counterparts and this was inhibited by serotonin in both wild-type and BMPR-II+/− mice (Long et al., 2006). These effects of serotonin might be expected to lead to a proliferative phenotype (decreased Smad 1/5/8 and decreased inhibitor of DNA binding 3). Interestingly, although receptor-mediated phosphorylation of the Smads at the carboxyl terminal enables association with Smad 4 and thus nuclear translocation, ERK1/2 can phosphorylate the linker region of Smad 1 thus inhibiting nuclear translocation (Morrell, 2006). As ERK1/2 functions downstream of serotonin, this may provide the mechanism by which serotonin can antagonize the BMP pathway (Figure 1). Indeed, pulmonary arterial smooth muscle cells from BMPR-II+/− mice show increased proliferation in response to serotonin, and this is associated with increased ERK1/2 signalling. Moreover, pulmonary resistance arteries from BMPR-II+/− mice show an increased contractile response to serotonin (Long et al., 2006). Interestingly, patients with the BMPR-II mutation who had developed PAH after exposure to fenfluramines, had a shorter duration of fenfluramine exposure before illness than patients without the mutation (Humbert et al., 2002). As fenfluramines mediate PAH through a serotinergic mechanism (Dempsie et al., 2008), this observation may provide further evidence for a link between serotonin and the BMPR-II pathways.

Conclusions

There is now a considerable and growing body of evidence to support the ‘serotonin hypothesis of PAH'. Antagonistic interactions between the serotonin and the BMP system have been reported, thus increased serotonin signalling may provide a ‘second hit' risk factor which can mediate PAH in patients with a mutation in the BMPR-II gene. The effects of serotonin on the pulmonary vasculature are mediated through both SERT and the serotonin receptors (particularly the 5-HT_1B receptor). Indeed, SERT and the 5-HT_1B receptor synergize in mediating constriction of the pulmonary artery and emerging evidence now suggests that SERT and the 5-HT_1B receptor co-regulate components of the serotonin-signalling system. Therefore, targeting both of these may provide optimal therapy for PAH. Also emerging as a possible therapeutic target is the rate-limiting enzyme in the synthesis of serotonin, Tph1. Advances have been made in unravelling the serotonin-signalling system, uncovering further possible therapeutic targets. Much work remains to be carried out in this area and the challenge will be to specifically target signalling molecules within the pulmonary vasculature.

Abbreviations

BMP

bone morphogenetic protein

BMPR-II

bone morphogenetic protein receptor type II

ERK

extracellular signal-regulated kinase

fPAH

familial pulmonary arterial hypertension

iPAH

idiopathic pulmonary arterial hypertension

MAP kinase

mitogen-activated protein kinase

PAH

pulmonary arterial hypertension

ROCK

Rho-kinase

ROS

reactive oxygen species

SERT

serotonin transporter

TGF-β

transforming growth factor beta

Tph

tryptophan hydroxylase

Conflict of interest

The authors state no conflict of interest.