Potassium channels in pulmonary arterial hypertension

Activation of the phospholipase C–protein kinase C cascade regulates K_V1.5 channel function

As mentioned earlier, pathogenesis of PAH is characterised by increased levels of pro-proliferative factors (serotonin, endothelin-1 and thromboxane A2 (TXA2)) and reduced levels of anti-proliferative factors, such as nitric oxide and prostacyclin, in the pulmonary vascular wall [2]. Decreased K_V1.5-mediated current in PASMCs results from both direct effects of hypoxia and indirectly from hypoxia-induced expression of mitogenic factors.

The effects of serotonin (5-HT) on the pulmonary circulation have been of major interest since reports pinpointing the use of serotoninergic appetite suppressant drugs as a risk factor for the development of PAH [21]. The implication that serotonin is a key determinant of vessel remodelling was further confirmed by several observations highlighting: 1) high platelet and plasma levels of serotonin in patients with PAH [22], 2) prevention of PAH in animal models following inhibition of serotonin receptors [23, 24], and 3) a congenital predisposition to develop PAH in fawn-hooded rats caused by a defect in platelet serotonin storage [25]. Briefly, under pathological conditions, pulmonary endothelial cells produce high levels of serotonin that acts in a paracrine fashion on adjacent PASMCs to promote cell proliferation and contraction [26–28]. Results showing that exposure of human normotensive PASMCs to aminorex, fenfluramine or dexfenfluramide (appetite suppressant drugs that interact with the metabolism of serotonin) reduces the expression of K_V1.5 and inhibits K⁺ current support the notion that serotonin signalling causes PAH, at least in part, through its effect on K_V channel gene regulation [29, 30]. Cogolludo et al. [31] reported that serotonin treatment of PASMCs or Ltk⁻ cells expressing cloned hK_V1.5 markedly inhibited K_V current and depolarised the cell membrane. Antagonists of 5-HT2A receptors abolished these effects. Under pathological circumstances, the binding of serotonin to the 5-HT2A receptors activates phospholipase C (PLC) through Gq family G-proteins that, in turn, lead to the cleavage of phosphatidylinositol-4,5-biphosphate (PIP₂) into inositol-1,4,5-triphosphate and diacylglycerol (DAG) (fig. 2) [32, 33]. The former triggers the release of Ca²⁺ from intracellular pools and the latter activates protein kinase C (PKC), which phosphorylates several intracellular substrates [33]. Likewise, inhibitors of PLC, conventional PKC and tyrosine kinase prevented the effects of serotonin on K_V currents [31]. The authors found that K_V1.5 coimmunoprecipitated with 5-HT2A receptors and caveolin-1 in native pulmonary arteries underscoring that this physical interaction could serve to cluster the 5-HT2A receptor and signalling proteins with the K⁺ channel, allowing for specific modulation [31]. More importantly, Cogolludo et al. [31], demonstrated that stimulation with serotonin leads to caveolin-dependent internalisation of K_V1.5, ensuring dynamic cell surface channel expression [31, 34]. Likewise, endocytosis of K_V2.1 and K_V1.5 channels has been reported to be a key mechanism in the control of K_V channel activity, which is regulated by tyrosine phosphorylation mediated by Src kinase [35–37]. Thus, serotonin signalling appears to modulate K_V current in PASMCs through direct alterations in K_V1.5 trafficking and surface expression. This constitutes a mechanistic insight into how the levels of K_V1.5 channels are downregulated in PAH (fig. 2).

TXA₂ and prostacyclin are major arachidonic acid metabolites produced by vascular cells. TXA₂ is a potent vasoconstrictor, whereas prostacyclin is a powerful vasodilator with antiplatelet and anti-proliferative effects. In both the primary and secondary forms of PH, the balance between these two molecules is shifted toward TXA₂ [38]. Interestingly, a TXA2 analogue (U46619), through activation of the thromboxane endoperoxide receptor, inhibited K_V current and depolarised PASMCs. Using a PKCζ pseudosubstrate inhibitory peptide, Cogolludo et al. [39] provided evidence for the role of this kinase as a link between thromboxane endoperoxide receptor activation and K_V channel inhibition. The requirement for PKCζ in mediating TXA2-induced effects on K_V current was further confirmed by the same group using PKCζ^−/− pulmonary arteries [40]. In addition, sequestosome1/p62, a stress-inducible protein regulated by the redox-sensitive transcription factor Nrf2, was shown to serve as a scaffold protein acting as a physical link in the assembly of PKCζ–sequestosome1/p62–potassium channel complexes and to facilitate phosphorylation of K_Vβ by PKCζ, which induces inhibition of pulmonary arterial K_V1.5 channels [41]. Consistent with this and similar to findings in PKCζ^−/− PASMCs, U46619 fails to inhibit K_V currents in PASMCs deficient for sequestosome1/p62 [40]. Collectively, these data provide evidence for a modulatory role of sequestosome1/p62 in the functional activity of K_V1.5. It is worth noting that the subtypes of PKC involved in the regulation of PASMC K_V1.5 channels vary among vasoconstrictors, since serotonin inhibits PASMC surface expression of K_V1.5 channels in a PKCζ-independent manner, contrary to inhibition byTXA2 [31].

The interaction between growth factor receptors and the K_V1.5 channel was examined in Xenopus oocytes, an advantageous model for studying connections between receptors and ion channels. Coexpression of cloned mouse platelet-derived growth factor (PDGF) receptor and cloned K_V1.5 in Xenopus oocytes, followed by PDGF stimulation led to a decline in the K_V1.5 current amplitude. Taking advantage of the fact that the fibroblast growth factor (FGF) and PDGF receptors share structural similarity and are thought to have a similar mechanism of activation, the involvement of PLCγ was explored using a mutated form of the FGF receptor that does not associate with PLCγ [42]. In coexpression experiments, the mutant FGF receptor was not able to modulate K_V1.5 current amplitude, providing strong evidence that increased PLCγ activity is required for the reduction in K_V1.5 current amplitude [42]. Considering the implication of FGF2 and PDGF in pulmonary vascular remodelling, the same mechanisms could conceivably occur in PAH PASMCs [43, 44].

Bone morphogenetic protein signalling and potassium channels

Several studies underscoring a critical role for bone morphogenetic protein (BMP) signalling in PAH have been published. A wide range of mutations, which are distributed throughout the coding region of the BMPR2 gene, have been identified in patients with PAH [45, 46]. Consistently, transgenic mice expressing a dominant-negative Bmpr2 gene in smooth muscle cells (SM22-tet-BMPR2delx4+) develop increased pulmonary arterial pressure, associated with a modest increase in arterial muscularisation [47]. To test whether anomalies seen in SM22-tet-BMPR2delx4+ are mediated by reduced K_V channel expression, Young et al. [48] investigated the expression levels of several K_V channels and revealed a pronounced diminution of K_V1.5 protein expression in the lungs of transgenic mice. Consistent with this, treatment of PASMCs with recombinant BMP2 stimulates K_V1.5 expression and was accompanied by increased current density in these cells; a functional effect blocked by an anti-K_V1.5 antibody [48, 49]. The mechanism of K_V1.5 regulation by BMP signalling is unclear. Considering that BMP2 induces apoptosis in human PASMCs by reducing the expression of the anti-apoptotic protein Bcl-2 and that overexpression of Bcl-2 diminishes K_V1.5 expression, it is, therefore, conceivable that the functional effects of signalling on K_V1.5 can be attributed to alteration of Bcl-2 [8, 50].

Transcriptional control: focus on hypoxia-inducible factor-1α and nuclear factor of activated T-cells

Although generation of reactive oxygen species (ROS) in mitochondria is known to be altered in PH, the direction of these changes remains debated [51]. A cancer-like metabolic shift towards a glycolytic metabolism even in the presence of adequate oxygen, also designated the Warburg effect, has been demonstrated as a pathogenic element that can contribute to the development of PAH [52]. Mitochondrial dysfunction, characterised by hyperpolarisation of mitochondrial membrane potential and fragmentation, generates a low ROS environment causing a pseudo-hypoxic state that rapidly stabilises and triggers nuclear translocation of the master transcription regulator of the adaptive response to hypoxia, hypoxia-inducible factor (HIF)-1α [18, 53]. The relationship between HIF-1α activation and K_V1.5 expression was investigated in PAH patients and in fawn-hooded rats, which develop spontaneous PAH with ageing [18]. Normoxic HIF-1α activation (e.g. nuclear localisation) was almost universal in both cultured fawn-hooded rat PASMCs and human small pulmonary arteries from PAH patients. This correlates with a significant diminution of K_V1.5 and K⁺ current, membrane depolarisation and increased cytosolic calcium. Furthermore, expression of K_V1.5 and K⁺ current were restored in fawn-hooded rat PASMCs exposed to hyperoxia or transfected with an adenovirus coding for a dominant-negative form of HIF-1α, strengthening the fact that HIF-1α activation accounts for the downregulation of K_V1.5 [18].

A sustained increase in cytosolic Ca²⁺, due to potassium-dependent or independent mechanisms, is an important activator of the nuclear factor of activated T-cells (NFAT) family of transcription factors. In resting cells, NFAT proteins are phosphorylated and reside in the cytoplasm. Upon stimulation, NFAT are dephosphorylated by calcineurin, leading to its translocation into the nucleus where it can regulate gene expression [54] (fig. 2). Persistent activation of NFATc2 is a hallmark of human idiopathic PAH (IPAH) PASMCs [55]. Inhibition of its nuclear translocation by VIVIT or cyclosporine A was shown to reverse K_V1.5 downregulation in normal PASMCs exposed to hypoxia. Moreover, in vivo NFAT inhibition by cyclosporine A treatment was shown to restore K_V1.5 expression and reverse right ventricular and pulmonary artery medial hypertrophy in the monocrotaline (MCT)-induced rat model of PAH [55]. In agreement with the direct involvement of NFATc2 in K_V1.5 regulation, serotonin treatment of PASMCs induced marked nuclear translocation of NFATc2 and a significant decrease in K_V1.5 protein expression [19]. Although analyses of the KCNA5 gene revealed multiple putative NFAT binding elements in the 5′ untranslated region, there is no evidence of direct transcriptional regulation.

Several other elements also contribute to transcriptional control of KCNA5. Among these, c-jun, a member of the transcription factor activator protein-1 family was demonstrated to influence voltage-gated K⁺ channel activity in PASMCs. By infecting PASMCs with an adenovirus expressing c-Jun, Yu et al. [56] demonstrated that overexpression of this transcription factor was associated with a significant decrease in the amplitude of K_V current by downregulating K_V1.5 expression.

Inflammation and potassium channels in PAH: an under studied field

Pulmonary vascular inflammation (chemokine-driven leukocyte accumulation in the pulmonary vascular lesions of PAH) and systemic inflammation (rise in blood levels of pro-inflammatory cytokines/chemokines, C-reactive protein and auto-antibodies) are a common feature of human and experimental PAH [57, 58]. Although inflammation is recognised to play a significant role in the pathogenesis of PAH, relative few studies have investigated the functional effects of pro-inflammatory cytokines on K⁺ current in PASMCs. Sutendra et al. [59] demonstrated that normal human PASMCs exposed to tumour necrosis factor (TNF)-α adopt a PAH phenotype with decreased K⁺ current. The molecular mechanisms mediating this effect remain unclear, but can be attributed to the activation of STAT3 (signal transducer and activator of transcription 3); a transcription factor accounting for the modulation of the expression of several proteins such as HIF-1α and NFAT (fig. 2) [60]. Indeed, increased STAT3 activation has been described in normal PASMCs treated with TNF-α or interleukin-6, suggesting that the decrease in K⁺ current elicited by these inflammatory factors is due, at least in part, to a STAT3-dependent pathway [61, 62].

Evidence for single nucleotide polymorphisms in the KCNA5 gene that underlie susceptibility to PAH

The identification of sequence polymorphisms is a key to understanding human genetic differences and diseases. The possible impacts of single nucleotide polymorphisms in KCNA5 on channel expression and function in PAH have been investigated [63, 64]. In taking advantage of large cohorts restricted to the European Caucasian population, a KCNA5 rs10744676 variant, located in the putative promoter of KCNA5, was found to be associated with systemic sclerosis (SSc). Interestingly, analysis within the SSc cohort revealed that a KCNA5 rs10744676 C allele was associated with protection from PAH [64]. It is suggested that this polymorphism within a CACCC box, normally recognised by transcription factors of the Sp/Krüppel-like factor family, could affect transcription factor binding and hence influence KCNA5 transcription and expression levels. The functional consequences of this polymorphism remain to be addressed. Nevertheless, combined deletion, site-directed mutagenesis, pharmacological and chromatin immunoprecipitation approaches have shown that the Kcna5 promoter was critically dependent on Sp1 regulation via CACCC box motifs in murine vascular smooth muscle cells, highlighting the functional importance of CACCC box motifs to drive KCNA5 gene expression [65].

Participation of microRNAs in the regulation of K_v1.5 in PASMCs

MicroRNAs are a class of short non-coding RNAs that act as negative regulators of gene expression by inhibiting the translation of target mRNAs, by promoting their degradation [66]. It is becoming increasingly apparent that the aberrant expression of microRNAs is causally related to a variety of disease states and PAH is no exception. Indeed, multiple deregulated microRNAs (e.g. miR17-92 cluster, miR-21, miR-143/145, miR-204 and miR-424/503) have been implicated in PAH [67, 68]. MicroRNA expression profiling between IPAH-PASMCs and normal PASMCs and subsequent bioinformatics microRNA target gene prediction were recently performed to examine the potential role of microRNAs in the downregulated K_V channel expression and whole-cell K_V current, characteristic of IPAH-PASMCs. Among microRNAs found to be selectively upregulated in IPAH-PASMCs, miR-29b was selected based on the predicted miR-29b–K_V1.5 mRNA relationships. Forced expression of miR-29b in normal PASMCs decreased Kv1.5 protein expression, whereas inhibition of miR-29b in IPAH-PASMCs, using antisense oligonucleotides, rescued Kv1.5 protein levels, demonstrating that K_V1.5 is a direct target of miR-29b in IPAH-PASMCs. These findings highlight an additional level of complexity to the regulation of K_V1.5 (fig. 2) [69].

Post-translational modifications of K_V1.5: implications in PAH?

Protein post-translational modification (PTM) is a highly dynamic and crucial mechanism in which the functional properties of a protein are altered by the covalent addition of a chemical group or protein to its amino-acid residues. Independently of PAH, K_V1.5 has been shown to undergo diverse, reversible or irreversible PTM, including phosphorylation, S-acylation, palmitoylation, sumoylation, ubiquitination and glycosylation, which impact on ion channel expression, channel trafficking to the membrane, stability and activity [35, 70–74]. In response to oxidative stress, a sulfenic acid modification has been reported on a cysteine residue present in the C-terminal domain of K_V1.5 that triggers internalisation of K_V1.5, decreases current density, and diverts the channel from the recycling endosome towards degradation [75]. Moreover, tyrosine phosphorylation of K_V1.5 and redox-sensitive K_V1.5 sumoylation were also reported leading to its inactivation [35, 73]. Using Xenopus oocytes as a model to study ion channels in a controlled in vivo environment, it was demonstrated that expression of the ubiquitin ligase Nedd4-2 declined K_V1.5 currents by ubiquitinating and thereby reducing K_V1.5 plasma membrane expression [74]. Finally, transfection of a K_V1.5 palmitoylation-deficient mutant into Chinese hamster ovary cells has been shown to enhance the expression of surface current through enhanced localisation of the channel to the plasma membrane, indicating that palmitoylation reduced K_V1.5 biological function [71]. No information is available on the PTM status of K_V1.5 channels in PAH. Taking into account that metabolic pathways provide substrates for PTM and that PAH is recognised as a disorder associated with significant metabolic dysfunction, it can be assumed that major alterations in PTM of K⁺ channels occur in PAH [76, 77]. Nonetheless, the likely importance of these regulatory mechanisms remains to be explored.

Together, these findings demonstrate that K_V1.5 is carefully regulated throughout the lifecycle of the channel and that alterations of these regulatory mechanisms can have important functional repercussions on K_V1.5-mediated currents and PAH progression.

KCNK3 (TASK1) as a cause of PAH

Familial cases of PAH have long been recognised and are usually due to mutations in members of the transforming growth factor (TGF) signalling cascade [45, 46, 87]. BMPR2 mutations account for ∼70% of familial PAH (HPAH) and 15% of patients with IPAH. To a lesser extent, heritable and idiopathic PAH are also associated with mutations of genes related to TGF-β signalling, such as activin receptor-like kinase 1 (ALK1), endoglin (ENG), and mothers against decapentaplegic 9 (SMAD9) [88]. To date, a significant proportion of familial cases of PAH have not been linked to a genetic cause. Recent advances in genome sequencing technologies have provided unprecedented opportunities to identify mutations. Whole exome sequencing has gained popularity and was recently and successfully applied in a family in which multiple members presented PAH without identifiable mutations in any of the genes known to be linked to the disease. This approach allowed the discovery of KCNK3 as a new predisposing gene for PAH [89]. In screening additional patients with IPAH and HPAH, a total of six KCNK3 variants were identified, all of which resulted in a loss of function of the channel at physiological pH that probably causes depolarisation of the resting membrane potential. The prevalence of KCNK3 mutations was 1.3% in IPAH and 3.2% in HPAH [89]. This study provides the first causal relationship between a K⁺ channel and PAH, and consequently PAH is now considered as a channelopathy. Surprisingly, neither right ventricular hypertrophy nor vessel remodelling were observed in Kcnk3^−/− mice underlying the fact that pulmonary arterial pressure of mutant mice was not elevated [90]. Divergence between human and rodent orthologous genes, existence of functional redundancy in rodents, and inter-species variability in expression levels may explain the apparent differences. In particular, KCNK3 does not form a functional channel in mouse PASMCs [91]. In mice, Kcnk3 function is probably replaced by other K⁺ channels, for example Kcnk6. Accordingly, Kcnk6 mutant mice are more sensitive to hypoxia-induced PH [86]. Although involvement of KCNK3 is probable in human PAH, analysis of its expression level in animal models mimicking the human disease is required. The generation of a Kcnk3 mutant rat, a popular model species in the field of PH, will constitute a powerful tool to decipher its implication in vascular remodelling.