Control of vascular smooth muscle function by Src-family kinases and reactive oxygen species in health and disease

Abstract

Reactive oxygen species (ROS) are now recognised as second messenger molecules that regulate cellular function by reversibly oxidising specific amino acid residues of key target proteins. Amongst these are the Src-family kinases (SrcFKs), a multi-functional group of non-receptor tyrosine kinases highly expressed in vascular smooth muscle (VSM). In this review we examine the evidence supporting a role for ROS-induced SrcFK activity in normal VSM contractile function and in vascular remodelling in cardiovascular disease. VSM contractile responses to G-protein-coupled receptor stimulation, as well as hypoxia in pulmonary artery, are shown to be dependent on both ROS and SrcFK activity. Specific phosphorylation targets are identified amongst those that alter intracellular Ca²⁺ concentration, including transient receptor potential channels, voltage-gated Ca²⁺ channels and various types of K⁺ channels, as well as amongst those that regulate actin cytoskeleton dynamics and myosin phosphatase activity, including focal adhesion kinase, protein tyrosine kinase-2, Janus kinase, other focal adhesion-associated proteins, and Rho guanine nucleotide exchange factors. We also examine a growing weight of evidence in favour of a key role for SrcFKs in multiple pro-proliferative and anti-apoptotic signalling pathways relating to oxidative stress and vascular remodelling, with a particular focus on pulmonary hypertension, including growth-factor receptor transactivation and downstream signalling, hypoxia-inducible factors, positive feedback between SrcFK and STAT3 signalling and positive feedback between SrcFK and NADPH oxidase dependent ROS production. We also discuss evidence for and against the potential therapeutic targeting of SrcFKs in the treatment of pulmonary hypertension.

Introduction: SrcFKs as ROS effectors in VSM

The purpose of this review is to highlight the interactions between reactive oxygen species (ROS) and Src-family kinases (SrcFKs), a family of non-receptor tyrosine kinases, in the regulation of vascular smooth muscle (VSM) function. We will examine evidence supporting an important role for this interaction in normal excitation–contraction coupling. We will also provide details of the role of SrcFKs in oxidative stress-related VSM proliferation and migration signalling associated with vascular remodelling, with a focus on pulmonary hypertension, and briefly comment on the potential therapeutic use of SrcFK inhibitors against this group of diseases. Firstly, however, we will set the scene by describing ROS production in VSM, their role as second messengers and mechanism of action on target proteins, and evidence supporting SrcFKs as key proximal ROS effectors in VSM.

Vascular ROS production

ROS are now considered as bona fide second messenger molecules, being produced within cells in response to physiological and patho-physiological stimuli, acting on cellular target proteins to reversibly alter cellular function. There are two main sources of ROS in VSM. Firstly from cytoplasmic oxidoreductase enzymes, most notably NADPH oxidase (NOX), which transfers an electron from cytosolic NADPH to molecular oxygen, generating superoxide (O₂^•−) (Bedard & Krause, 2007). Secondly, electrons leaking from the mitochondrial electron transport chain form superoxide in the mitochondrial inter-membrane space (Turrens, 2003; Waypa et al. 2013). Superoxide readily reacts with nitric oxide (NO•), producing peroxynitrite (ONOO⁻), or is converted to the more stable hydrogen peroxide (H₂O₂) by superoxide dismutase (SOD) (Bedard & Krause, 2007). H₂O₂ is in turn fully reduced to H₂O by catalase or converted to the highly reactive hydroxyl radical via the Fenton reaction (Fig.1).

Figure 1.

Generation, metabolism and action of ROS in VSM

Superoxide (O₂^•−) is generated by oxido-reductase enzymes in the plasma membrane, such as NADPH oxidase (NOX). A reducing substrate, in this case NADPH, donates an electron which is transferred to molecular oxygen on the extracellular surface (Bedard & Krause, 2007). Superoxide is generated in a similar manner by the mitochondrial electron transport chain, principally complex III (Waypa et al. 2013). Superoxide can re-enter the cell or exit the mitochondria through non-selective anion channels (NSAC), or is converted to H₂O₂ by cytosolic or mitochondrial superoxide dismutase (SOD). Alternatively, superoxide may react with nitric oxide (NO•) to form the similarly reactive peroxynitrite (ONOO⁻). H₂O₂ is converted to the highly reactive hydroxyl radical (OH•) via the Fe²⁺-catalysed Fenton reaction or is fully reduced to H₂O by catalase (CAT). Superoxide and H₂O₂ reversibly modify protein function principally via oxidation of the sulfhydryl (R-SH) group of cysteine residues to sulfenic acid (R-SOH). Peroxynitrite S-nitrosylates cysteine (R-SH → R-SOH + NO₂⁻) and nitrates the hydroxyl groups of serine, threonine and tyrosine residues (R-OH → R-NO₂), and the latter is less readily reversible (Janssen-Heininger et al. 2008). OH• is highly reactive, much less discriminating, and therefore potentially damaging to the cell.

Vascular NOX activity is stimulated by various G-protein-coupled receptor (GPCR) agonists including angiotensin II, endothelin-1, thrombin, ATP and thromboxane (Banes-Berceli et al. 2005; Madamanchi et al. 2005; El-Awady et al. 2011; Chakraborti et al. 2012). It is generally agreed that mitochondrial ROS production is enhanced by hypoxia both in the short term and the long term and, acutely, this is considered an important trigger for the normal physiological vasoconstrictor response to regional hypoxia in the pulmonary vascular bed (Waypa et al. 2013). GPCR-induced arterial constriction is generally relaxed by antioxidant enzymes, NOX inhibitors, or small molecule antioxidants (El-Awady et al. 2011; Connolly et al. 2013), while hypoxic pulmonary vasoconstriction is enhanced by inhibitors of endogenous SOD (Abdalla & Will, 1995) and inhibited by antioxidants and inhibitors of mitochondrial electron transport function (Connolly et al. 2013; Waypa et al. 2013). Conversely, exogenous ROS generally enhance VSM proliferation (Wedgwood et al. 2001; Madamanchi et al. 2005) and cause constriction, especially in pulmonary artery, although they may relax other vascular beds depending perhaps on the nature of the pre-constricting agent (Knock et al. 2009; Snetkov et al. 2011).

REDOX regulation of SrcFKs

ROS alter protein function by chemically modifying specific amino acid residues. Cysteine is relatively readily oxidised by superoxide and H₂O₂, and S-nitrosylated by peroxynitrite. Peroxynitrite also nitrates serine, threonine and tyrosine residues (Fig.1). Oxidation of paired cysteine residues within a protein or in adjacent proteins may also result in the formation of disulphide bridges, but is not an abso-lute requirement for a change in protein function (Janssen-Heininger et al. 2008). Cysteine oxidation and nitrosylation, and to a lesser extent nitration, is readily reversible through the action of REDOX buffer systems, most notably glutathione reductase/peroxidase (Janssen-Heininger et al. 2008). The susceptibility of individual cysteine residues to oxidation varies between proteins and between residues within the same protein, depending on the ionisation pK_a of the residue in question (Chiarugi et al. 2003; Gusan & Anand-Srivastava, 2013).

c-Src is the principle member of a family of closely related tyrosine kinases collectively known as the Src-family kinases (SrcFKs). There are nine in total, of which three (c-Src, Fyn and Yes) are highly expressed in VSM (Nakao et al. 2002; Knock et al. 2008a). Note that in this review we will refer to these kinases collectively as SrcFKs, since Src, Fyn and Yes are difficult to separate pharmacologically, and rarely has genetic manipulation been used to determine their individual contributions to VSM function. SrcFKs feature prominently among the various classes of proteins that possess cysteine residues with a pK_a low enough for them to be oxidised by physiological levels of ROS (Knock & Ward, 2011). Although there may be specific exceptions, cysteine oxidation is generally stimulatory of tyrosine kinase activity and inhibitory of protein tyrosine phosphatase (PTP) activity (Knock & Ward, 2011; Funato & Miki, 2014). Thus, it is no surprise that increased ROS production in response to various vasoactive stimuli is associated with an early increase in total cellular tyrosine phosphorylation (Uzui et al. 2000; Chiarugi et al. 2003; Knock et al. 2008a,b) and that exogenous ROS enhance tyrosine phosphorylation in vascular preparations, whilst antioxidants do the opposite (Frank et al. 2001, 2003). Correspondingly, PTP inhibitors tend to promote constriction, while tyrosine kinase inhibitors, including those selective for SrcFKs, tend to relax pre-constricted arteries (Jin & Rhoades, 1997; Knock et al. 2008a, 2009). SrcFKs are normally activated by de-phosphorylation of the auto-inhibitory C-terminal phospho-tyrosine (Tyr-527 in c-Src), and subsequent auto-phosphorylation of the catalytic subunit activation loop (Tyr-418 in c-Src) (Xu et al. 1999). ROS activate SrcFKs through multiple mechanisms, firstly by oxidising cysteine residues in the SH2 domain, causing inter-molecular disulphide bridge formation and disruption of internal Tyr-527–SH2 domain interaction, secondly by oxidising cysteine residues on the SH2 and kinase domains of the already active kinase, further activating it, and thirdly by inhibiting their inactivation through oxidative inhibition of an Src-specific PTP, and of C-terminal Src-kinase (CSK) (Akhand et al. 1999; Giannoni et al. 2005; Roskoski, 2005; Mills et al. 2007) (Fig.2).

Figure 2.

Activation of Src by ROS in VSM

Src activation is normally promoted by de-phosphorylation of the auto-inhibitory Tyr-527, disrupting intra-molecular binding with the SH2 domain. This promotes auto-phosphorylation of Tyr-418 on the activation loop, opening up the kinase domain for substrate binding. Additionally, SrcFKs are both directly and indirectly REDOX sensitive, being activated by: (i) ROS-induced intermolecular disulphide bridge formation which promotes trans-phosphorylation of Tyr-418 (Akhand et al. 1999), (ii) oxidation of cysteine residues in the SH2 (Cys-245) and kinase (Cys-487) domains, which further activates an already active Src (Giannoni et al. 2005), (iii) via oxidative inactivation of the inhibitory c-Src kinase (CSK), which phosphorylates the auto-inhibitory C-terminal Tyr-527 (Mills et al. 2007), or (iv) oxidative inactivation of Src-specific PTPs which de-phosphorylate Tyr-418 on the activation loop (Roskoski, 2005). Specific amino acid numbers given refer to the sequence in human c-Src.

SrcFKs are activated in smooth muscle by various vasoconstrictor stimuli including GPCR agonists (Nakao et al. 2002; Knock et al. 2008a), stretch (Gui et al. 2010; Gonzales et al. 2014) and hypoxia (Knock et al. 2008b), while VSM contraction induced by exogenously applied H₂O₂, or by artificially enhanced superoxide production, is sensitive to SrcFK inhibition (Oeckler et al. 2003; Knock et al. 2009). Thus, VSM responses to various stimuli may be mediated via an interaction between ROS and SrcFKs. Indeed, as discussed in the following sections, many of the numerous SrcFK phosphorylation targets, including various classes of ion channel, cytoskeletal proteins associated with small G-proteins of the Rho family, growth-factor receptors, other tyrosine kinases, and transcription regulators, are phosphorylated in a ROS-dependent manner.

SrcFKs and ROS in normal contractile function

The principle function of vascular smooth muscle is to contract or relax in response to circulating or local factors, thus contributing to the control of tissue blood flow and mean arterial blood pressure. Smooth muscle contraction is caused by the ratchet-like movement of the molecular motor myosin along actin fibres. Myosin ATPase activity, which triggers this movement, is initiated by Ca²⁺/calmodulin-dependent phosphorylation of myosin light-chain-20 (MLC₂₀) by myosin light-chain kinase (MLCK). Polymerisation of actin filaments and their bundling with myosin into contractile fibres is also dynamically regulated, as is the association of these fibres with integrin-rich focal adhesions which permit the transmission of contractile force to the extracellular matrix, as well as acting as platforms for multiple signalling pathways (Ridley & Hall, 1992; Min et al. 2012; Sreenivasappa et al. 2014). In this section, we describe how ROS and SrcFKs regulate contractile force via changes to the free intracellular Ca²⁺ concentration ([Ca²⁺]_i) to influence MLCK activity, by inhibiting the activity of myosin light-chain phosphatase (MLCP) which de-phosphorylates MLC₂₀, and by regulating actin polymerisation and focal adhesion signalling through the activation of the small monomeric G-protein RhoA and various adaptor proteins associated with integrins.

Regulation of ion channels

Many vasoconstrictors that bind Gq-coupled receptors activate phospholipase C-β (PLC-β) to induce inositol 1,4,5-trisphosphate (IP₃)-dependent Ca²⁺ release from the sarco-endoplasmic reticulum and open diacylglycerol (DAG)-sensitive transient receptor potential (TRP) non-selective cation channels. The resultant depolarisation then triggers influx through voltage-dependent Ca²⁺ channels. Exogenous ROS enhance agonist-induced increases in [Ca²⁺]_i and contraction in pulmonary artery (Snetkov et al. 2011) and these effects may be in part via activation of SrcFKs, since GPCRs stimulate SrcFKs to further activate TRP channels through tyrosine phosphorylation of PLC-γ (Vazquez et al. 2004; Gonzales et al. 2014), or via direct phosphorylation of canonical TRP family members on the cytosolic N-terminus, which has been shown as an obligatory step in their activation (Kawasaki et al. 2006). Alternatively, SrcFKs also directly phosphorylate voltage-gated Ca²⁺ channels, specifically the α1c pore-forming subunit of Ca_V1.2, which facilitates its binding to α5β1 integrin, a step which is necessary for its contribution to myogenic contraction (Gui et al. 2010) (Fig.3).

Figure 3.

SrcFKs and ROS modulate ion channel activity and contraction in VSM

GPCR stimulation activates Gq and phospholipase C-β (PLC-β) to generate diacylglycerol (DAG), which opens C-type transient receptor potential (TRP) channels. GPCRs also induce ROS-dependent activation of SrcFKs which phosphorylate and activate phospholipase C-γ (PLC-γ) or directly phosphorylate TRP channels on the cytoplasmic N-terminal domain (e.g. Tyr-226 on TRPC3; Kawasaki et al. 2006), further enhancing their activity. Resultant cationic influx depolarises the cell (+V_m), which in turn opens voltage-gated Ca²⁺ channels (VGCC). SrcFKs also phosphorylate and further activate VGCCs (e.g. Tyr-2122 on the α1c pore-forming subunit of Ca_V1.2; Gui et al. 2010), promoting their association with integrins (e.g. α5β1; Gui et al. 2010). VGCC opening elevates intracellular [Ca²⁺] ([Ca²⁺]_i); Ca²⁺ binds calmodulin (CAM), which activates myosin light-chain kinase (MLCK) which in turn phosphorylates myosin light-chain-20 (MLC) to initiate contraction. SrcFKs may also phosphorylate K⁺ channels of the Ca²⁺-activated (BK_Ca), voltage-gated (K_V) and two-pore acid-sensitive (TASK) types, thus limiting depolarisation. This phosphorylation may be stimulatory or inhibitory and in pulmonary artery, whether hypoxia enhances (Waypa et al. 2013) or inhibits (Wu et al. 2007) ROS production, determines whether SrcFKs are activated or inhibited by hypoxia, thus enhancing or limiting cell depolarisation, which is of potential importance in pulmonary hypoxic vasoconstriction (see main text for explanation).

K⁺ channels are also phosphorylation targets for SrcFKs and in the systemic circulation there is evidence that this phosphorylation may be stimulatory or inhibitory. In mesenteric artery, serotonin-induced SrcFK activity results in inhibition of K_V current, thus contributing to contraction (Sung et al. 2013). There is conflicting evidence over the effects of SrcFK-dependent phosphorylation on large conductance Ca²⁺-activated potassium (BKCa) channels, with inhibition being reported in rat aorta (Alioua et al. 2002) and activation being reported in rat cremaster VSM, which also promotes association of the channel with integrins (Gui et al. 2010). The reasons for these discrepancies are unclear, but perhaps relate to differences in vascular bed or experimental conditions. In some vascular beds where K⁺ channel activity has a particularly strong influence on vascular tone, a stimulatory effect of SrcFKs on K⁺ currents may account at least in part for the occasionally observed relaxing effects of ROS. For example, H₂O₂ enhances large conductance Ca²⁺-activated K⁺ (K_Ca) channel activity in coronary VSM, accounting for its endothelium-derived hyperpolarising factor-like activity in that vascular bed (Barlow et al. 2000), and mesenteric artery pre-constricted with a GPCR agonist, but not with 30 mm KCl, is relaxed by superoxide and this is associated with reduced [Ca²⁺]_i and enhanced voltage-gated K⁺ (K_V) current (Snetkov et al. 2011) (Fig.3).

As described in the introductory section, pulmonary artery differs from most systemic vascular beds by constricting in response to hypoxia and this constriction is generally considered to occur as a result of increased mitochondrial ROS production (Waypa et al. 2013). Coupled with this, SrcFKs are activated by hypoxia (Sato et al. 1999, 2005; Knock et al. 2008b), and SrcFK inhibition reduces hypoxic constriction in pulmonary artery (Knock et al. 2008b). A recent study on the two-pore acid-sensitive K⁺ channel TASK-1 in pulmonary VSM has shown this channel to be phosphorylated and opened by SrcFKs, and using this model to explain how SrcFKs contribute to hypoxic pulmonary vasoconstriction, SrcFKs are shown to be inhibited by hypoxia thus reducing TASK-1 current and causing depolarisation (Nagaraj et al. 2013). Paradoxically, this latter study is only consistent with the wealth of evidence that SrcFKs are activated by ROS if hypoxia inhibits mitochondrial ROS production (Wu et al. 2007), contradicting the aforementioned work of Waypa et al. Perhaps direct evidence for an involvement of SrcFKs in the relationship between ROS and K⁺ channels in pulmonary VSM, and the relative importance of K⁺ channels to pulmonary vascular tone would help to resolve this apparent discrepancy. Interestingly, ROS do not relax pre-constricted pulmonary artery, despite K⁺ current being enhanced by ROS to a degree similar to that in mesenteric VSM (Snetkov et al. 2011), perhaps due to a relative dominance of ROS-induced Rho-kinase activity in pulmonary VSM (Knock et al. 2009; Snetkov et al. 2011), as discussed below.

Regulation of RhoA/ROCK

The monomeric G-protein RhoA and its effector kinase, Rho-kinase (ROCK), are essential regulators of actin polymerisation, focal adhesion assembly and MLCP activity. Many contractile stimuli, including various GPCR agonists, hypoxia and exogenous ROS, activate RhoA/ROCK in VSM, and there is pharmacological evidence that this activity is in part dependent on SrcFKs (Nakao et al. 2002; Knock et al. 2008a,b), presumably downstream of RhoA activation. RhoA is activated by exchange of bound GDP for bound GTP and this is promoted by guanine nucleotide exchange factors (GEFs) (Bos et al. 2007). Several RhoA-specific GEFs (ARHGEFs) are expressed in VSM (Cario-Toumaniantz et al. 2012) and they are activated directly by G_12/13 and/or by tyrosine phosphorylation, which may be promoted by ROS (Chandra et al. 2012). It has been hypothesised that SrcFKs may activate RhoA via GEF phosphorylation (Knock et al. 2008a; Sreenivasappa et al. 2014), but to date this has not been demonstrated directly. Instead, RhoGEFs are known to be phosphorylated and activated by three other non-receptor tyrosine kinases, focal adhesion kinase (FAK), protein tyrosine kinase-2 (PYK2) or Janus kinase (JAK2) (Lim et al. 2008; Guilluy et al. 2010; Gadepalli et al. 2012). Although these three kinases are not known to be directly ROS sensitive, they are activatable by GPCR agonists in the vascular wall in a NOX-dependent manner, as well as by exogenous ROS (Vepa et al. 1999; Frank et al. 2003; Daou & Srivastava, 2004) and through SrcFK-mediated tyrosine phosphorylation (Calalb et al. 1995; Andreev et al. 2001; Singh et al. 2011). JAK additionally appears to require prior ROS-dependent activation of PYK2 (Frank et al. 2002) (Fig.4).

Figure 4.

SrcFKs mediate ROS-induced activation of RhoA and Rho-kinase in VSM

(i) GPCRs and hypoxia induce ROS production (via NOX and mitochondrial electron transport chain, respectively, see Fig.1), which activate SrcFKs. GPCRs also activate G_12/13. RhoA is activated by exchange of bound GDP for bound GTP and this is promoted by guanine nucleotide exchange factors (RhoGEFs). RhoGEFs may be activated by G_12/13 and/or by tyrosine phosphorylation. SrcFKs activate other non-receptor tyrosine kinases focal adhesion kinase (FAK), protein tyrosine kinase-2 (PYK2) or Janus kinase (JAK2), all of which may activate RhoGEFs. SrcFKs may also directly activate RhoGEFs but this has not been clearly demonstrated (dotted line). JAK2 activity may also require prior PYK2 activity. (ii) Contraction requires polymerisation and bundling of actin filaments into contractile fibres and their attachment at focal adhesions, which are composed in part of integrins, talin and vinculin, and various adapter proteins including FAK, paxillin and Cas, which are phosphorylated by SrcFKs. RhoA at focal adhesions promotes actin polymerisation and bundling, in part through association with paxillin and activation of the actin chaperone protein mDia. (iii) RhoA also activates Rho-kinase (ROCK) which further promotes actin polymerisation by activating LIM-kinase (LIMK), which in turn prevents cofilin from severing actin filaments. ROCK also phosphorylates the myosin targeting subunit of myosin phosphatase (MLCP) resulting in reduced myosin light-chain de-phosphorylation, thus enhancing net MLC phosphorylation and contraction.

Active RhoA promotes actin fibre nucleation, polymerisation and bundling into contractile fibres in association with SrcFKs, FAK, paxillin and other adapter proteins at focal attachments as well as chaperone proteins such as mDia (Geneste et al. 2002; Lim et al. 2008). Also, ROCK phosphorylates and activates LIM-kinase, which in turn phosphorylates and inhibits the actin-severing protein cofilin, thus promoting actin fibre stability (Geneste et al. 2002) (Fig.4). Contractile stimuli induce SrcFKs to phosphorylate several proteins at focal adhesions, including FAK (Tyr-925), paxillin (Tyr-118) and Cas (Tyr-165) (Min et al. 2012). MLCP activity is constitutive and does not require active stimulation, but contractile stimuli that raise [Ca²⁺]_i can further enhance MLC₂₀ phosphorylation by actively inhibiting MLCP. The resultant further enhancement of MLC₂₀ phosphorylation above that induced by raising [Ca²⁺]_i is known as ‘Ca²⁺ sensitisation’. Rho-kinase phosphorylates myosin phosphatase targeting subunit-1 (MYPT-1) on at least two threonine residues, which result in disassembly of the MLCP holoenzyme (Thr-850), or a reduction in its phosphatase activity (Thr-696), respectively (Ichikawa et al. 1996; Velasco et al. 2002) (Fig.4).

SrcFKs and ROS in pulmonary hypertension

Pulmonary hypertension is a spectrum of diseases which includes pulmonary hypertension due to hypoxia and/or respiratory disease (PH) and pulmonary arterial hypertension (PAH) which has various causes, both known and unknown. Both PH and PAH are characterised by remodelling of the pulmonary vasculature and hyper-responsiveness to constrictor stimuli, resulting in increased pulmonary vascular resistance and subsequent right heart failure and death (Simonneau et al. 2009). Vascular remodelling includes smooth muscle proliferation and migration which normally contribute to tissue growth and repair, but otherwise are usually suppressed in favour of a contractile phenotype. Occlusive cardiovascular diseases on the other hand, are associated with excessive or dysregulated VSM proliferation and migration. These changes have been strongly linked with excessive and poorly regulated ROS production, with contributions from altered expression of oxidant and antioxidant enzymes, endothelial dysfunction and increased sensitivity to, or increased availability of ROS-stimulating factors, resulting in an abnormal shift in smooth muscle REDOX state in favour of oxidation, commonly described as ‘oxidative stress’. Enhanced mitochondrial and/or NOX-derived ROS coupled with reduced expression of antioxidant enzymes have been implicated in the pathogenesis of both PH and PAH (Liu et al. 2006; Crosswhite & Sun, 2010; Reis et al. 2013).

SrcFKs, as key ROS targets in VSM, and known mediators of smooth muscle proliferation and migration (Krymskaya et al. 2005; Pullamsetti et al. 2012), are therefore well placed to contribute to oxidative stress-related vascular remodelling. ROS promote cell migration (Wang et al. 2014), acting in part through activation of SrcFKs and FAK at focal adhesions (Timpson et al. 2001; Chiarugi et al. 2003; Giannoni et al. 2005). As discussed in the previous section, and shown in Fig.4, SrcFKs promote actin contractile fibre formation through tyrosine phosphorylation of multiple protein targets including paxillin, Cas, RhoGEFs and activation of RhoA. This is also important in cell migration, but occurs alongside disassembly of cell–cell junctions (Frame et al. 2002) and activation of the monomeric G-proteins cdc42 and Rac-1, which generate cycles of filopodia formation and cell spreading, respectively, at the leading edge of the migrating cell (Nobes & Hall, 1995). In the remaining sections of this review, with a particular focus on pulmonary hypertension, we will examine how ROS and SrcFKs interact to induce VSM proliferation, via transactivation of growth factor receptors, signal transducer and activator of transcription-3 (STAT3) and hypoxic inducible factors, and through positive feedback, further enhance ROS production. Finally, we will discuss evidence for and against targeting SrcFKs as treatments for PH/PAH.

SrFK, ROS and growth factor transactivation

ROS-generating stimuli use multiple signalling pathways to induce VSM growth and proliferation. An early step following activation of SrcFKs is the activation of receptor tyrosine kinases, and here we will focus on two of the most important of these, the epidermal growth factor receptor (EGFR) and the platelet-derived growth factor receptor (PDGFR). In response to ROS-generating stimuli, receptor tyrosine kinases are activated in the absence of their natural ligands through a process known as transactivation (Saito & Berk, 2001). After growth factor binding, the receptor forms homodimers and auto-phosphorylates multiple tyrosine residues on its cytoplasmic regions. During transactivation in response to GPCR agonists such as endothelin-1 and angiotensin II, these residues are instead phosphorylated in a ROS-dependent manner by SrcFKs (Heeneman et al. 2000; Oeckler et al. 2003; Chen et al. 2006; Nakashima et al. 2006). The resultant phospho-tyrosines then act as docking sites for multiple SH2-domain-containing proteins, which in turn trigger multiple downstream signalling events. Here and in Fig.5, using PDGFR as an example, we describe four of the most fully characterised of these signalling pathways. Transactivated PDGFR recruits the adapter proteins Shc and growth-factor receptor bound protein-2 (GRB2) (Ward et al. 1996; Ravichandran, 2001). SrcFKs subsequently phosphorylated Shc to provide further binding sites for GRB2 (Ravichandran, 2001). GRB2 can then bind and activate the guanine nucleotide exchange factor for Ras, son-of-sevenless (SOS). Activated Ras then initiates the Raf/MEK/ERK mitogen-activated protein (MAP) kinase cascade, promoting cell division. PLC-γ also binds PDGFR in VSM and is activated by ROS-dependent SrcFK-mediated phosphorylation (Saito et al. 2002; ten Freyhaus et al. 2011), and the subsequent sustained elevation in [Ca²⁺]_i contributes to sustained ERK activation (Egan et al. 2005). Phosphatidyl inositol-3-kinase (PI3K) is also phosphorylated by SrcFKs following binding to the transactivated PDGFR (ten Freyhaus et al. 2011; Karoor et al. 2013) and this phosphorylation relieves intra-molecular PI3K auto-inhibition (Cuevas et al. 2001; Choudhury et al. 2006). PI3K then generates phosphatidyl inositol trisphosphate (PIP₃), which through membrane localisation and activation of Akt (protein kinase B) and PDK initiates the Bad and S6 kinase anti-apoptotic and protein synthesis pathways (Eguchi et al. 1999). SrcFKs further enhance this signalling by phosphorylating and inhibiting phosphatase and tensin homologue (PTEN), preventing it from de-phosphorylating PIP₃ (Lu et al. 2003; Karoor et al. 2013).

Figure 5.

Role of SrcFKs in growth factor receptor transactivation and mitogenic signalling in VSM

(i) GPCR-induced or hypoxia-induced ROS stimulate SrcFKs to phosphorylate platelet-derived growth factor receptor (PDGFR) on multiple tyrosine residues. This occurs in the absence of native PDGF ligand (transactivation). (ii) PDGFR recruits the adapter proteins Shc and growth factor receptor bound protein-2 (GRB2) on phosphorylated Tyr-557 and Tyr-684, respectively (Ward et al. 1996). SrcFKs subsequently phosphorylate Shc to provide further binding sites for GRB2. GRB2 can then bind and activate the guanine nucleotide exchange factor for Ras, son-of-sevenless (SOS). Activated Ras then initiates the Raf/MEK/ERK mitogen-activated protein kinase cascade, promoting cell division. In addition, PLC-γ binds PDGFR at phosphorylated Tyr-1021 (Saito et al. 2002) and is then itself phosphorylated and activated by SrcFK. Activated PLC-γ induces a sustained elevation in [Ca²⁺]_i which contributes to sustained ERK activation. (iii) Phosphatidyl inositol-3-kinase (PI3K) binds PDGFR at phosphorylated Tyr-751 (Karoor et al. 2013), following which it is phosphorylated and activated by SrcFKs. PI3K then phosphorylates phosphatidyl inositol bisphosphate (PIP₂) to generate phosphatidyl inositol trisphosphate (PIP₃). SrcFKs also phosphorylate and inhibit phosphatase and tensin homologue (PTEN), preventing de-phosphorylation of PIP₃. PIP₃ promotes the membrane localisation of Akt and PDK, allowing PDK to phosphorylate and activate Akt. Activated Akt initiates the Bad and S6 kinase (S6K) anti-apoptotic and protein synthesis pathways.

Reciprocal relationship between SrcFKs and ROS

Chronic hypoxia triggers a sustained enhancement in mitochondrial ROS production in VSM, resulting in enhanced SrcFK activity (Sato et al. 2005). However, as well as being a key downstream effector of ROS, SrcFKs also contribute to the activation of NOX. In response to GPCR stimulation, SrcFKs phosphorylate the p47^phox subunit of NOX and trigger the activation of Rac-1, which is required for holoenzyme assembly (Seshiah et al. 2002; Chowdhury et al. 2005). Coupled with the steady increase in basal mitochondrial ROS production described above, increased expression of NOX, and reduced expression of antioxidant enzymes in PH and PAH (Liu et al. 2006; Reis et al. 2013), this apparent positive feedback relationship between ROS and SrcFK activity could not only facilitate the prolonged sustained increases in ROS characteristic of oxidative stress in PH/PAH, but may also be required for the long-term changes in VSM phenotype, migration and proliferation induced downstream of SrcFKs (Seshiah et al. 2002) characteristic of pulmonary hypertension, acting via growth-factor receptor transactivation, and as discussed below, hypoxia-inducible factor (HIF)-1α/2α and STAT3 activation (Fig.6).

Figure 6.

Role of SrcFKs in pulmonary hypertension

(i) Pulmonary hypertension induced by chronic hypoxia (PH) is associated with enhanced mitochondrial ROS production. PH and pulmonary arterial hypertension (PAH) are both also associated with enhanced NOX expression and activity and reduced antioxidant expression and activity, further elevating ROS production. In addition to being activated by ROS, SrcFKs also in turn enhance NOX activity through phosphorylation of the p47^phox subunit and activation of Rac-1 which is required for NOX holo-enzyme assembly, therefore generating a positive feedback loop between ROS and SrcFKs. (ii) SrcFKs phosphorylate and inactivate two enzymes, prolyl hydroxylase (PLH) and von Hippel-Lindau tumour suppressor protein (VHL), thus preventing the prolyl hydroxylation and degradation of hypoxia inducible factor-1α or -2α (HIF1α/2α). HIF enhances the expression of platelet-derived growth factor (PDGF) and its receptor (PDGFR) and inhibits the expression of tyrosine phosphatases that oppose the activity of SrcFKs and PDGFR. (iii) SrcFKs phosphorylate and activate the transcription factor, signal transducer and activator of transcription-3 (STAT3). STAT3 inhibits the expression of microRNA-204 (miR-204) and bone morphogenetic protein receptor-II (BMPRII) which inhibit the expression and activity of SrcFKs, respectively. STAT3 therefore indirectly enhances expression and activity of SrcFKs, generating a second positive feedback loop. STAT3 also up-regulates three additional mitogenic signals: nuclear factor of activated T-cells (NFAT), Survivin and Pim-1. This activity, coupled with HIF activation, oxidative stress and direct SrcFK transactivation of PDGFR (Fig.5), results in greatly enhanced VSM proliferation, migration and inhibition of apoptosis.

ROS, SrcFKs and altered gene expression via hypoxia-inducible factor and STAT3

SrcFKs are implicated in several additional signalling pathways considered important in VSM remodelling in PH/PAH. For example, they also contribute to VSM migration and proliferation by increasing the availability of hypoxia-inducible factors (HIF-1α and HIF-2α) (Sato et al. 2005). In normoxia, HIF levels are kept low by prolyl hydroxylation and subsequent proteolysis; however, in response to chronic hypoxia, ROS-induced SrcFKs phosphorylate and inhibit prolyl hydroxylase (PLH) and/or von Hippel-Lindau tumour suppressor protein (VHL), thus increasing the availability of HIF (Chan et al. 2002; Sato et al. 2005; Chou et al. 2010). Activated HIF alters the expression of a wide variety of proteins that contribute to the enhanced proliferation, migration and hyper-contractility associated with PH, including increases in expression of growth factors and their receptors (Shimoda & Laurie, 2014; Smith & Yuan, 2014) and decreases in expression of protein tyrosine phosphatases that oppose the actions of SrcFKs and PDGFR (ten Freyhaus et al. 2011) (Fig.6).

STATs are activated by tyrosine phosphorylation, which triggers their dimerisation and translocation to the nucleus. This phosphorylation is typically mediated by Janus kinase following cytokine receptor activation. However, in PH, ROS-induced SrcFKs may also phosphorylate and activate STAT3 independently of JAK, perhaps in association with growth-factor receptors. Activated STAT3 subsequently up-regulates the expression of nuclear factor of activated T-cells (NFAT), Pim-1 and Survivin, promoting VSM proliferation and inhibiting apoptosis (Paulin et al. 2011). STAT3 also inhibits the expression of the anti-proliferative bone morphogenetic protein receptor-II (BMPRII) and micro-RNA-204 (miR-204) (Paulin et al. 2011), and since BMPRII normally inhibits SrcFK auto-phosphorylation (Wong et al. 2005), while miR-204 inhibits its expression, STAT3 indirectly stimulates expression and activity of SrcFKs. Furthermore, some forms of familial PAH are also associated with a constitutive deficiency in miR-204 levels and/or mutations in BMPRII that render it unable to inhibit SrcFKs (Wong et al. 2005; Courboulin et al. 2011). Thus, through STAT3, there is the potential for a second positive feedback loop enhancing the activation of SrcFKs and downstream signalling in both PH and PAH (Fig.6).

SrcFKs as a therapeutic target in pulmonary hypertension?

Considering the evidence in support of a key role for SrcFKs in smooth muscle function, both normal (constriction) and abnormal (oxidative stress, vascular remodelling), one might expect them to be potential therapeutic targets for the treatment of cardiovascular disease in general and pulmonary hypertension in particular. SrcFK-selective inhibitors have been tested against some cancers but so far the focus of the search for effective treatments against pulmonary hypertension has been on a group of mixed specificity inhibitors exemplified by imatinib and related compounds that target PDGFR, c-Kit and c-Abl, originally designed as anti-cancer drugs (Iqbal & Iqbal, 2014). These have been tested against animal models of PAH and PH with promising results (Schermuly et al. 2005; Berghausen et al. 2013), and a recent clinical trial also suggests that imatinib improves cardiac function in a cohort of patients with PAH (Shah et al. 2014). Interestingly, although specific SrcFK inhibitors have yet to be tested in this context, a comparison of two PDGFR antagonists, imatinib and dasatinib, showed dasatinib to be the most potent against hypoxia-induced VSM migration and proliferation and morphological changes associated with PH, and the difference was attributed to the additional inhibition of SrcFKs by dasatinib (Pullamsetti et al. 2012). However, the impact of this finding has been muddled somewhat by recent reports that dasatinib, while in use as an anti-cancer drug actually induces pulmonary hypertension in human subjects, presumably through vasoconstriction rather than remodelling since it is readily reversible (Godinas et al. 2013; Seferian et al. 2013). This is consistent with a reported acute enhancement of pulmonary perfusion pressure in isolated mouse lung by both dasatinib and the selective SrcFK inhibitor PP2 (Godinas et al. 2013; Nagaraj et al. 2013) and consistent with the finding that SrcFKs phosphorylate and activate TASK channels in pulmonary VSM (Nagaraj et al. 2013), but not with the weight of evidence showing SrcFK to be both pro-contractile and pro-proliferative in its actions. Nevertheless, whether these unwanted effects of dasatinib in the clinical setting are indeed mediated through SrcFKs or are independent of tyrosine kinase inhibition remains to be confirmed, considering that no reports of selective SrcFK inhibitors inducing PH in humans have been made.

Conclusions

In summary, there is clear evidence placing SrcFKs as key ROS effectors in VSM, acting both as mediators of normal smooth muscle contractile responses, via modulation of ion channel and RhoA/Rho-kinase activity, and as mediators of uncontrolled VSM proliferation and migration in response to oxidative stress, acting upon multiple downstream signalling pathways including growth-factor receptor transactivation, STAT3 and hypoxia-inducible factors. More specifically, there is also considerable evidence implicating SrcFKs in the pathogenesis of pulmonary hypertension, but more research is required to attribute the experimental and clinical effects of mixed-specificity kinase inhibitors to specific tyrosine kinases such as SrcFKs.

Glossary

BMPRII

bone morphogenetic protein receptor-II

FAK

focal adhesion kinase

GEF

guanine nucleotide exchange factor

GPCR

G-protein-coupled receptor

GRB2

growth factor receptor bound protein-2

MLC₂₀

myosin light-chain-20

MLCK

myosin light-chain kinase

MLCP

myosin light-chain phosphatase

NOX

NADPH oxidase

PAH

pulmonary arterial hypertension

PDGFR

platelet-derived growth factor receptor

PH

(hypoxic) pulmonary hypertension

PLC-β/γ

phospholipase C-β/γ

PTP

protein tyrosine phosphatase

ROCK

Rho-kinase

ROS

reactive oxygen species

SOD

superoxide dismutase

SrcFKs

Src-family kinases

STAT3

signal transducer and activator of transcription-3

TASK channel

two-pore acid-sensitive K⁺ channel

VSM

vascular smooth muscle

Biographies

Charles MacKay is currently in the second year of his PhD at King’s under the supervision of Dr Knock and presented his work to date at the aforementioned symposium.

Greg Knock is a Lecturer in Physiology at King’s College London, appointed in 2010. His research interests are signal transduction in vascular and respiratory smooth muscle, with a focus on protein tyrosine kinases, small G proteins of the RhoA family and reactive oxygen species. He was the organiser of the Physiological Society Research Symposium entitled ‘Tyrosine Kinases in Smooth Muscle Function'’ held in London in July 2014.

Additional information

Competing interests

None declared.

Funding

C. Mackay is funded by British Heart Foundation studentship (FS/12/43/29608).