Abstract
Pulmonary arterial hypertension (PAH) is a devastating disease that involves pulmonary vasoconstriction, small vessel obliteration, large vessel thickening and obstruction, and development of plexiform lesions. PAH vasculopathy leads to progressive increases in pulmonary vascular resistance, right heart failure and, ultimately, premature death. Besides other cell types that are known to be involved in PAH pathogenesis (e.g. smooth muscle cells, fibroblasts and leukocytes), recent studies have demonstrated that endothelial cells (ECs) have a crucial role in the initiation and progression of PAH. The EC-specific role in PAH is multi-faceted and affects numerous pathophysiological processes, including vasoconstriction, inflammation, coagulation, metabolism and oxidative/nitrative stress, as well as cell viability, growth and differentiation. In this review, we describe how EC dysfunction and cell signalling regulate the pathogenesis of PAH. We also highlight areas of research that warrant attention in future studies, and discuss potential molecular signalling pathways in ECs that could be targeted therapeutically in the prevention and treatment of PAH.
Abstract
Dysfunctional endothelial cells play an important role in initiating and progressing PAH by impacting on multi-aspects of vascular remodelling and vasoconstriction. EC-targeted therapy may be effective in inhibiting vascular remodelling to treat PAH. https://bit.ly/3nXPvMB
Introduction
Pulmonary hypertension (PH) is defined as an increase in mean pulmonary arterial pressure of >20 mmHg from the normal range of 10–20 mmHg at rest, as assessed by right heart catheterisation [1]. PH is a heterogeneous cardiopulmonary disease that is divided into five groups, including pulmonary arterial hypertension (PAH), PH due to left heart disease, PH due to lung disease and/or hypoxia, chronic thromboembolic PH and PH with unclear multifactorial mechanisms [2]. PAH, including idiopathic PAH (IPAH), is characterised by a progressive rise in pulmonary vascular resistance and occlusive vascular remodelling, which leads to right heart failure and premature death. The histopathological features of PAH include intima and media thickening, muscularisation of distal pulmonary arteries, vascular occlusion and complex plexiform lesions [3–5]. Some of these histological features are also present in other groups of PH, but to a lesser extent. Despite major advances in the field over recent years, the underlying molecular mechanisms of obliterative vascular remodelling remain largely unknown. Current therapies are based on concepts of endothelial dysfunction developed almost three decades ago, targeting the endothelin, nitric oxide (NO) and prostacyclin pathways, and they do not address the fundamental disease-modifying mechanisms. These therapies have only resulted in a modest improvement in morbidity and mortality; therefore, the ultimate treatment remains lung transplantation [6–8].
The healthy endothelial monolayer lining the inner wall of blood vessels regulates the flux of fluid, proteins and blood cells across the vessel wall into parenchymal tissue; maintains vascular tone and integrity; and exerts anti-thrombotic and anti-inflammatory influences on the vascular bed [9]. However, endothelial cells (ECs) are damaged and/or dysfunctional in PAH patients [10–13]. Factors that can cause EC injury include hypoxia, toxins, inhibition of survival signalling (e.g. vascular endothelial growth factor (VEGF) antagonists), recreational drug use, inflammatory cytokines, pathological shear stress and fluid mechanics in the pulmonary circulation raised by left to right shunts in congenital heart disease. Shear stress is of particular importance in the pathogenesis of PAH, given the dramatic changes in arterial pressure and fluid dynamics that occur in the pulmonary circulation of PAH patients. Many characteristics of PAH are consequences of dysfunctional EC signalling; these characteristics include pulmonary inflammation and coagulation, oxidative/nitrative stress, altered vascular cell viability (e.g. apoptosis resistance), proliferation, metabolic shift and accumulations of inflammatory cells and fibroblasts [14–16]. Mice with the encoding prolyl-4 hydroxylase-2 (PHD2) (Egln1) deletion in ECs and bone marrow cells exhibit unprecedented severe PH, recapitulating many features of clinical PAH such as occlusive vascular remodelling and right heart failure [17]. This supports the critical role of EC dysfunction in the pathogenesis of PAH (figure 1). It has been proposed that EC phenotypic changes contribute to the onset of PAH, e.g. in the cases of smoking-induced lung EC apoptosis or inherited epigenetic EC dysfunction [18–20]. However, dysfunctional EC phenotypes can manifest in parallel with PAH or after the onset of PAH, e.g. in the instances of chronic vascular inflammation or proliferative vasculopathy following anti-angiogenic therapy [21, 22]. In other words, the precise timing of the phenotypic changes from healthy to dysfunctional endothelium during the pathogenesis of PAH is unclear. The evolution of PAH-associated EC phenotypes likely depends on multiple variables such as the EC phenotype and subpopulation being studied, the disease type and severity, and the patient's genetic inheritance and demographics, including age, sex and other co-existing PAH risk factors.
Although current treatments for PAH can reduce disease symptoms and delay clinical worsening in PAH patients, there is no therapy or combination of therapies that prevents the onset of PAH or completely alleviates the disease. This review will describe the abnormal endothelial signalling pathways that contribute to the initiation and development of PAH and describe how a dysfunctional endothelium regulates PAH pathogenesis and progression. We will also highlight areas of research that could ultimately support the development of EC-targeted therapies against PAH and identify future studies that could improve understanding of obliterative vascular remodelling and PAH pathogenesis.
Pulmonary EC phenotypes
The early stage of PAH development involves EC injury and apoptosis, while apoptosis-resistant ECs emerge later as PAH progresses [23–25]. In a transgenic mouse model with Fas-induced EC apoptosis, PH and pulmonary arteriopathy are observed, providing direct evidence that lung EC damage acts as a trigger to initiate PAH [26]. In the late stages of PAH, hyper-proliferative and apoptosis-resistant ECs predominate, contributing to the formation of plexiform lesions [24, 25, 27, 28]. In distal pulmonary arteries of lungs from IPAH patients, there are increased numbers of proliferating ECs and decreased numbers of apoptotic ECs [29]. These observations are also seen in vitro, with pulmonary ECs from IPAH patients exhibiting increased proliferation and reduced sensitivity to apoptosis [29, 30]. In the end stage of PAH, there is evidence of endothelial senescence. A switch from a proliferative to a senescent vascular phenotype contributes to the loss of reversibility of PAH [31]. Dysfunctional EC signalling also results in increased coagulability and decreased EC integrity, which contribute to the development of PAH [32]. PAH pathogenesis is often associated with aberrant EC barrier integrity and IPAH patients commonly demonstrate a hypercoagulable phenotype [33]. Additionally, it is increasingly being recognised that alterations in multiple metabolic and epigenetic pathways are driving the development of PAH [34]. Survival and hyperproliferation of the PAH endothelium requires increased glutamine metabolism through the tricarboxylic acid cycle; PAH patients exhibit systemic and lung-specific changes in glutamine metabolism, with PAH lung vasculature taking up more glutamine than healthy controls [35].
However, it is important to note that pulmonary ECs comprise separate subpopulations of ECs, including proximal pulmonary artery ECs (PAECs) and distal microvascular ECs, which may be subject to different injurious stimuli and mechanical forces according to their position in the pulmonary vasculature [36, 37]. Moreover, the alveolar endothelium can be resolved by single-cell transcriptomics into at least two specialised capillary EC phenotypes characterised either by expression of apelin or its receptor, that play specialised roles in gas exchange and repair [38]. This EC heterogeneity could therefore affect the severity of the aberrant EC phenotypes observed in PAH, including EC proliferative potential. This section provides an overview of EC-expressed factors that control the aberrant EC phenotypes seen in PAH (figure 1).
Factors affecting EC survival
PAH is hereditable in ∼10% of cases, and the vast majority of patients with hereditary PAH harbour heterozygous mutations in bone morphogenetic protein receptor type 2 (BMPR2) [39]. Loss of BMPR2 is an initiating factor for PAH [39]: heterozygous Bmpr2 knockout causes EC injury and persistent PH in mice [40, 41], and genetic ablation of Bmpr2 in pulmonary ECs predisposes to PH [42, 43]. BMPR2 mediates pro-survival signalling in PAECs [41, 44]. Overexpression of mutant BMPR2 in human PAECs increases susceptibility to apoptosis [39]. Thus, as in the experimental models [26], EC apoptosis appears to represent a potential initiating mechanism in the pathogenesis of human PAH as well. Meanwhile, the BMPR2 ligand BMP9 prevents EC apoptosis, enhances EC monolayer integrity and inhibits PH induced by BMPR2 mutations, monocrotaline (MCT) or Sugen/hypoxia (SuHx) [45]. Furthermore, the BMPR2 activator FK506 improves endothelial function, inhibits apoptosis and reverses PH in hypoxic mice [46]. However, another study has demonstrated that inhibition of BMP9 signalling partially protects against experimental PH [47]. These data suggest that the transforming growth factor-β (TGF-β)/BMP pathway is highly complex and the effects of BMP9 may depend on other factors [48]. BMPR2 also mediates the transcriptional complex between peroxisome proliferator-activated receptor-γ (PPARγ) and β-catenin in PAECs [49]. Apelin is a downstream target of this complex. Apelin-deficient PAECs are prone to apoptosis and promote pulmonary artery smooth muscle cell (PASMC) proliferation [49]. Apelin treatment can increase CD39 ATPase enzymatic activity in PAECs [50], whereas repression of CD39 in vitro results in an ATP-enriched environment that acts as a phenotypic switch, promoting apoptosis resistance in PAECs via the P2Y11 receptor [50]. Indeed, genetic deletion of CD39 [51] and the apelin/APJ system [52] augments hypoxia-induced PH in mice. Other factors, including chloride intracellular channel 4 (CLIC4), platelet-derived growth factor B (PDGFB) and hypoxia-induced mitogenic factor (HIMF), also regulate PAEC survival and PH development [53–57].
Factors affecting EC proliferation
In later stages of PAH, EC proliferation is a dominant feature that leads to complex arterial remodelling. Several pathways are involved in this transition. The loss of PPARGγ is associated with PAH development [58, 59]. Inhibition of PPARγ in PAECs upregulates expression of cell cycle genes, worsens VEGF-induced EC barrier dysfunction [60] and attenuates the migration and angiogenic capacity of pulmonary microvascular ECs [61]. PPARγ also maintains EC homeostasis via ubiquitin protein ligase E3 component n-recognin 5 (UBR5)/ATM interactor (ATMIN)-mediated DNA repair [62]. Accordingly, endothelial deletion of PPARγ induces spontaneous PH and impairs recovery from hypoxia-induced PH in mice [63]. The role of endothelial PHD2 in the development of PAH has recently been studied [17, 64, 65]. Mice with Tie2Cre-mediated disruption of PHD2 in ECs and haematopoietic cells exhibit severe PH and occlusive vascular remodelling [17]. Marked increases in EC proliferation are seen in the pulmonary vascular lesions of these mice. In IPAH patients, PHD2 expression is diminished in ECs of the occlusive vascular lesions. PHD2 deficiency-induced PAH is mediated by endothelial activation of hypoxia-inducible factor-2α (HIF-2α) [17, 64, 66], which alters the expression of many of the PAH-causing factors. Genetic deletion of endothelial HIF-2α inhibits PH development in hypoxic mice [66, 67]. Pharmacological inhibition of HIF-2α inhibits PH in experimental mouse and rat models and promotes survival [66, 68]. HIF2A mutation has been identified in IPAH patients [69] and mice with the mutation exhibit PH [70]; thus, HIF-2α is emerging as a promising target of PAH therapy.
Caveolin1 expression is markedly decreased in pulmonary vascular ECs of IPAH patients [71, 72]. Inheritable mutations have been reported in CAV1 in PAH patients [73] and Cav1−/− mice develop PH [74], while re-expression of Caveolin1 in endothelium rescues PH in Cav1−/− mice [75]. Treatment with Cavtrin, a Caveolin1 mimic peptide, inhibits EC proliferation and promotes apoptosis [76], whereas Caveolin1 deficiency induces PAEC proliferation [72]. Consistently, disruption of Cav1 in ECs augments hypoxia-induced PH [72]. Mammalian target of rapamycin (mTOR) [77] and Nur77 [78] are also negative regulators of PAEC proliferation and protect against PH development.
A number of other factors are also involved in PAEC proliferation contributing to the pathogenesis of PAH. For example, granzyme B cleaves intersectin-1s, generating an N-terminal fragment that enhances EC proliferation [79]. Interleukin-6 (IL-6) stimulates EC proliferation and increases endothelin 1 expression in ECs [80, 81]. p130Cas may modulate PAEC migration and proliferation by acting as an amplifier of receptor tyrosine kinase downstream signals [82]. Upregulation of growth differentiation factor 11 (GDF11) enhances the aberrant angiogenesis and proliferation in PAECs induced by hypoxia or VEGF treatment [83]. Endothelial dysfunction is strongly associated with oxidative and nitrative stress, and the antioxidant TEMPOL or MitoQ decreases migration and proliferation of ECs [84]. Inhibition of reactive oxygen species (ROS)-induced calcium entry also attenuates EC migration and proliferation [84].
Van der Feen et al. [31] showed that the loss of reversibility of pulmonary arterial remodelling in a congenital heart disease PAH model induced by MCT and aortocaval shunt is related to an EC phenotypic switch from proliferation to senescence. Cultured pulmonary ECs from PAH patients are more prone to becoming senescent in response to shear stress and the senescent cells are more sensitive to senolytic ABT263-induced apoptosis. Treatment of end-stage PH rats with ABT263 to target vascular cell senescence reversed the haemodynamic and structural changes. These studies demonstrate a new way to reverse end-stage PAH.
Factors affecting both EC survival and proliferation
Studies have also identified several factors that affect both EC proliferation and survival. For example, PAECs from IPAH patients exhibit increased fibroblast growth factor 2 (FGF2) expression. Disruption of FGF2 signalling normalises IPAH EC sensitivity to apoptosis and proliferation [29]. Apelin also regulates EC survival and proliferation [49]. Kim et al. [85] described a microRNA (miRNA)-dependent association between apelin and FGF2 in PAECs in which apelin deficiency leads to increased expression of FGF2 as a result of decreased expression of miR-424 and miR-503, mediated by myocyte enhancer factor 2 (MEF2). MEF2 activity is impaired in PAECs from IPAH patients owing to excessive nuclear accumulation of the class IIa histone deacetylases HDAC4 and HDAC5. Indeed, pharmacological inhibition of class IIa HDACs restored MEF2 activity, decreasing cell migration and proliferation in PAECs, and rescued experimental PH. A recent study showed that endostatin, a cleavage product of collagen type XVIII α1 chain (Col18A1), inhibits EC migration via inhibitor of DNA binding 1 (ID1)/thrombospondin-1 (TSP-1)/CD36 signalling and proliferation and apoptosis through CD36 and CD47 [86]. Elevated serum endostatin is associated with increased mortality and disease severity in PAH and a COL18A1 variant is associated with survival difference in PAH patients [87]. In a separate study, Notch1 increased PAEC proliferation and inhibited apoptosis, and pharmacological inhibition of Notch1 reduced PH in SuHx rats [88]. However, genetic deletion of endothelial Notch1 in mice worsens hypoxia-induced PH, possibly by increasing EC monolayer vulnerability [89], demonstrating that the relationship between EC survival and proliferation in PAH is complex.
Factors affecting EC activation and thrombogenicity
p-selectin is a pro-coagulant factor that is present on pulmonary ECs and platelets, and its expression reflects the extent of pulmonary EC injury [90]. Increased levels of p-selectin appear in the plasma of PAH patients, which can be decreased by infusion of the vasodilator prostacyclin [90]. Furthermore, soluble p-selectin levels in PAH patients are associated with EC dysfunction [91]. Levels of von Willebrand factor (VWF), another pro-coagulant factor, are also increased in the plasma [92] and pulmonary ECs [93] of PAH patients and correlate with risk of death [94, 95], EC damage and dysfunction [92]. These studies suggest that p-selectin and VWF could act as prognostic markers in PAH, e.g. to predict EC dysfunction and likelihood of disease onset or progression [96]. Intriguingly, the plasma levels of the anti-coagulant factor thrombomodulin are decreased in PAH patients [90, 97], and can be restored by infusion of the vasodilators prostacyclin [90] and tadalafil [97]. These reports suggest that the pulmonary vascular endothelium in PAH patients is prothrombogenic with increased expression of pro-coagulant molecules and decreased expression of anti-coagulant factors.
Factors affecting EC metabolism and epigenetics
Abnormal metabolism, especially aerobic glycolysis or the Warburg effect, has been proposed as an important pathogenic mechanism in the development of PAH. Pulmonary vascular ECs from PAH patients rely heavily on glycolysis (a shift from oxidative phosphorylation) for increased growth [98–101]. 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase (PFKFB) is a key regulator of glycolysis. Mice with EC-targeted Pfkfb3 deficiency exhibit attenuated PH or slowed PH progression with less EC inflammation and leukocyte recruitment to the lungs [102]. BMPR2-mediated Notch activation increases mitochondrial mass and expression of PFKFB3, which is necessary for citrate-dependent acetylation of H3K27, leading to expression of Notch1 target genes such as c-Myc and thus EC proliferation [89]. Overexpression of miR-124 or knockdown of polypyrimidine tract-binding protein 1 (PTBP1) restores normal levels of proliferation and glycolysis in ECs from PAH patients [103]. BMPR2 positively regulates miR-124 expression in ECs that targets PTBP1. Increased PTBP1 expression results in alternative splicing of pyruvate kinase muscle isoforms 1 and 2 (PKM1 and 2), leading to increased PKM2 expression. Thus, BMPR2 mutation or deficiency increases EC glycolysis via miR-124/PTBP1/PKM2 signalling [103]. Endothelin 1/endothelial nitric oxide synthase (eNOS) signalling is also involved in the glycolytic shift [104]. Endothelin 1 disrupts carnitine homeostasis and mitochondrial bioenergetics, which correlate with uncoupled eNOS redistribution from the plasma membrane to the mitochondria. The glycolytic switch appears to be dependent on mitochondrial-derived ROS that activate HIF signalling [104].
Studies also show that bolA family member 3 (BOLA3) regulates glycolysis and mitochondrial respiration [105]. Bola3 knockdown in mice or BOLA3 mutations in human decreases glycine cleavage system protein H, and thus enhances intracellular glycine. Bola3 deficiency enhances EC proliferation, survival and vasoconstriction, leading to PH. Iron-sulfur deficiency and changes in electron transport/cellular respiration have also been demonstrated in PAH via deficiencies in iron-sulfur cluster assembly enzyme (ISCU) signalling [106]. White et al. [106] showed in mouse and human vascular and endothelial tissues that miR-210 level was elevated in PAH samples, accompanied by reduced ISCU1/2 and iron-sulfur integrity. In mice, miR-210 repressed ISCU1/2 and enhanced PH. Conversely, mice deficient in EC-specific miR-210 showed increased ISCU1/2 levels and were resistant to PH, while ISCU1/2 knockdown promoted PH. Thus, the miR-210/ISCU1/2 axis causes iron-sulfur deficiency and PH. [106]. Other miRNAs that have been shown to regulate PH-associated dysfunctional phenotypes in ECs include miR-126 and miR-140-5p [107, 108]. Although the mechanisms through which each miRNA regulates PH remain incompletely understood, it is possible that miRNAs regulate EC PH phenotypes in an endocrine manner [109].
Peroxisome proliferator-activated receptor-γ coactivator α (PGC1α) is a master regulator of cellular metabolism and mitochondrial biogenesis [110]. Reduced PGC1α expression in PAECs by hypoxia leads to decreased oxidative metabolism, mitochondrial function and ROS generation, as well as increased ATP formation and eNOS phosphorylation, while upregulated PGC1α restores mitochondrial function. Another study demonstrated that uncoupling protein 2 (Ucp2) is also involved in EC mitochondrial function [111]. Cobalt chloride treatment (which mimics hypoxia) of Ucp2-deficient ECs increases mitophagy and decreases mitochondrial biogenesis. Thus, the loss of endothelial Ucp2 leads to inadequate mitochondrial biosynthesis, which may cause EC apoptosis.
Epigenetic mechanisms are also important in the regulation of EC metabolism. Glutamine carbon is required for endothelial survival in PAH, the maintenance of energetics and the hyperproliferative phenotype, and its delivery into the tricarboxylic acid cycle is increased in ECs with BMPR2 mutations [35]. The strict requirement for glutamine is driven by the loss of deacetylase sirtuin 3 activity. Preservation of sirtuin 3 function restores glutamine metabolism and prevents PH [35]. It has also been shown that vascular stiffness activates glutaminolysis to drive PH [112]. In the MCT-induced PH rat model, pharmacological targeting of pulmonary vascular stiffness and yes-associated protein (YAP)-dependent mechano-transduction modulated glutaminolysis, pulmonary vascular proliferation and PH. Furthermore, pharmacological targeting of glutaminase reduced MCT-induced PH progression [112].
PAH ECs exhibit altered DNA methylation in many of the genes related to lipid metabolism, including ATP binding cassette subfamily A member 1 (ABCA1) [113]. In rats, treatment with an ABCA1 agonist reduces MCT-induced PH. Histone methylation in ECs is also involved in PH development [114]. Myocardin-related transcription factor A (MRTFA)/megakaryoblastic leukemia 1 (MKL1) regulates expression of cell adhesion molecules, including intercellular adhesion molecule 1 (ICAM1) and vascular adhesion molecule 1 (VCAM1), through recruitment of H3K4 methyltransferase to the promoters. Mrtfa−/− mice exhibit hypoxia-induced PH with decreased expression of cell adhesion molecules. Endothelial-specific knockdown of absent, small, or homeotic discs 2 (ASH2) and WD repeat domain 5 (WDR5), two components of the H3K4 methyltransferase complex, reduces hypoxic PH in mice [114]. As described above, increased nuclear accumulation of HDAC4 and HDAC5 is also observed in PAH ECs, which impairs MEF2 activity leading to decreased miR-424 and miR-503 expression and increased EC proliferation [85, 115].
Factors affecting EC dedifferentiation
Under pathological conditions, ECs may undergo mesenchymal cell transition (EndoMT). Previous studies provide circumstantial evidence that EndoMT contributes to PAH directly, by EC transformation into smooth muscle-like cells with higher proliferative and migratory potential, or indirectly, through paracrine effects on vascular intimal and medial proliferation [116]. A recent study employing genetic lineage tracing demonstrated that EndoMT does not contribute to neointimal formation in a chronic inflammatory PH mouse model, but rather this results from a subpopulation of Notch3-expressing SMCs, a finding which raises questions about the direct contribution of EndoMT to PAH pathogenesis [117]. EndoMT markers are observed in complex vascular lesions in PAH patients and rats with BMPR2 mutation [116]. In normal PAECs, BMPR2 knockdown leads to increased expression of high mobility group AT-hook 1 (HMGA1) and EndoMT markers. The expression of EndoMT markers can be largely reversed by double knockdown of BMPR2 and HMGA1 or slug [118]. Also, rapamycin treatment inhibits expression of EndoMT markers, improves PH in BMPR2 mutant rats and decreases human PAEC migration [116]. In lungs of Egln1Tie2Cre mice, EndoMT marker expression is increased along with snail family transcriptional repressor (SNAI)1/2 in a HIF2α-dependent manner [65]. In IPAH lung ECs, PHD2 is downregulated, HIF2α expression is increased and expression of EndoMT markers is enhanced [65]. Future studies using genetic lineage tracing approaches in various animal models of severe PH, such as Egln1Tie2Cre mice and SuHx rats, are warranted to investigate the role of EndoMT in occlusive vascular remodelling and the pathogenesis of severe PAH.
Pulmonary EC crosstalk with SMCs
Heightened vasoconstrictor activity or reduced vasodilator activity contribute to PAH [119–122], and multiple EC-derived factors, including endothelin 1, NO and prostacyclins, regulate vascular tone. A key early component of PAH pathogenesis involves SMC vasoconstriction in response to increased endothelin 1, reduced NO bioavailability and low prostacyclins. Paracrine factors released from pulmonary ECs may also regulate SMC survival, proliferation and their functional phenotype, i.e. contractile versus synthetic, possibly contributing to the emergence of apoptosis-resistant hyperproliferative SMCs as PAH progresses [23–25], and ultimately remodelling of the pulmonary vasculature. This section provides an overview of EC-dependent mechanisms that control the aberrant SMC phenotypes seen in PAH.
EC regulation of SMC vasomotion
Endothelium-dependent pulmonary vasodilator signalling involves three main pathways: endothelium-derived hyperpolarising factor (EDHF), NO and prostacyclins. EDHF requires activation of calcium-sensitive potassium channels and cytochrome metabolites [123]. Impaired NO synthesis and bioavailability has been described in PH animal models and PAH patients [124–128]. In experimental studies, a wide variety of treatments that increase eNOS activity directly or indirectly have been shown to attenuate PH [129–134] and the evidence that NO signalling plays a crucial role in PAH is reviewed in detail elsewhere [135]. Prostacyclins are also potent vasodilators that are generated by vascular ECs as well as SMCs and endothelial progenitor cells (EPCs). The efficacy of prostacyclins for the treatment of PAH patients is well established [136, 137]. Endothelin 1, predominantly expressed in ECs, is a potent vasoconstrictor that plays an important role in the pathogenesis of PAH [138–140], as evidenced by its marked upregulation, particularly associated with complex arterial lesions, in lungs from patients with PAH [140]. Hypoxia-induced PH, for example, is suppressed in EC-specific Edn1 knockout mice [141]. This and many other studies have led to the development of drugs that target the vasoconstrictive actions of endothelin 1 and this area of research has been thoroughly reviewed by others [142–144]. Endothelial-derived oxidative/nitrative stress, e.g. secondary to Caveolin1 deficiency in ECs [71, 127], is another vasoconstriction mechanism that induces protein kinase G (PKG) tyrosine nitration, leading to impaired NO signalling due to a reduction in PKG activity, thereby inducing vasoconstriction and vascular remodelling [127, 145, 146]. Accordingly, PKG nitration is a prominent feature of IPAH lungs [127, 147], and targeting endothelial nitrative stress-induced PKG dysfunction may represent a novel therapeutic strategy for PAH treatment.
EC regulation of SMC proliferation, migration and survival
Culture of PASMCs in medium conditioned by IPAH ECs results in increased proliferation [148]. PAECs release a variety of growth factors and chemokines, including PDGFB, C-X-C motif chemokine ligand 12 (CXCL12), FGF2, macrophage migration inhibitory factor (MIF) and endothelin-1, that stimulate PASMC proliferation and pulmonary vascular remodelling [17, 63, 85, 149–153], likely through the transcription factor forkhead box M1 (FoxM1) [154]. Genetic deletion of Foxm1 in SMCs prevents hypoxia-induced PH in mice and pharmacological inhibition of FoxM1 inhibits severe PH in experimental PH models [154]. FoxO1 is a negative regulator of SMC proliferation in response to some angiocrine factors [155]. Apoptosis-resistant ECs from PAH patients also release miRNA1-95-5p to promote SMC proliferation via HIF1α and Smad7 [156]. AMP-activated protein kinase (AMPK) expression is decreased in PAECs from PAH patients [157]. Endothelial AMPK deficiency augments hypoxia-induced PASMC proliferation through phosphorylation and stabilisation of angiotensin converting enzyme 2 (ACE2), which increases eNOS-derived NO bioavailability and reduces PH [158].
Several factors released from pulmonary ECs can induce SMC migration. In IPAH patients, C-C chemokine ligand 2 (CCL2) release by pulmonary ECs is enhanced [159]. PASMCs from IPAH patients exhibit greater migratory and proliferative responses to CCL2. CXCL12 is another potent chemokine derived from ECs that may play an important role in promoting SMC migration contributing to vascular remodelling [17].
EC-specific gene transfer of indoleamine-2,3-dioxygenase attenuates PH in preclinical models [160]. Specifically, EC-derived indoleamine-2,3-dioxygenase promotes PASMC apoptosis via depolarisation of mitochondrial transmembrane potential and inhibits PASMC proliferation via a paracrine mechanism that remains to be elucidated [160]. In response to injury, apoptotic ECs release TGF-β1 and VEGF, which induce SMC proliferation [161]. Thus, EC death induced by inflammation and proinflammatory cytokines could activate SMC proliferation, leading to progression of pulmonary vascular remodelling and PAH. Tumour protein, translationally controlled 1 (TPT1) (also called translationally controlled tumour protein) is a potent anti-apoptotic factor that has been implicated in malignant cell transformation. TPT1 is released by ECs undergoing apoptosis in apoptotic nanovesicles, which are taken up by SMCs and directly induce SMC apoptosis resistance and growth dysregulation [162–164].
SMC regulation of EC proliferation
Recent studies provide some intriguing findings about SMC regulation of EC proliferation. Activation of Notch1 by BMPR2 leads to EC proliferation in SMC-EC co-cultures that is mediated by direct SMC-EC contact [89]. BMPR2 is required by both cell types to produce collagen IV to activate integrin-linked kinase, leading to stabilisation of presenilin 1 and activation of Notch1. This maintains the EC proliferative capacity by increasing mitochondrial mass and inducing PFKFB3. EC-targeted deletion of Notch1 in mice worsens hypoxia-induced PH in association with impaired EC proliferation and regeneration, and thus loss of pre-capillary arteries [89]. This study provides direct evidence that SMCs promote EC proliferation and regeneration to maintain monolayer integrity and vascular homeostasis in response to injury. miR-143-3p released from SMC exosomes is another mechanism promoting EC migration and proliferation [165]. However, in this case, EC proliferation is pathological given that inhibition of miR-143-3p reduces hypoxic PH in mice [165].
Pulmonary EC crosstalk with non-SMCs
Besides the direct effects of EC injury and dysfunction in the pathogenesis of PAH, crosstalk of ECs with SMCs and non-SMCs is increasingly recognised to play an important role in PAH progression. PAH is characterised by fibro-proliferative changes in the adventitia and immune cell accumulation in pathologically remodelled pulmonary vessels [27, 166]. (Myo)fibroblasts and inflammatory leukocytes are recruited to the lung through EC-dependent signalling mechanisms [166]. Several proinflammatory adhesion molecules and proinflammatory cytokines are abundantly expressed in activated ECs in experimental PH models and in the lungs of IPAH patients, which leads to inflammatory cell binding and recruitment [167]. Infiltrating inflammatory cells release cytokines including IL-1β and TNF-α, which activate ECs to express adhesion molecules, chemokines and cytokines and promote EC proliferation and death. Furthermore, crosstalk between ECs and other non-SMCs such as pericytes or T-cells can contribute to the pathogenesis of PAH. In this section, signalling mechanisms that occur in PAH pathogenesis between ECs and non-SMCs are described.
Inflammatory cells and immune (T-)cells
Accumulation of inflammatory cells in vascular lesions is a characteristic feature of clinical PAH. The expression levels of proinflammatory adhesion molecules such as ICAM1, VCAM1 and E-selectin are markedly elevated in the pulmonary vascular endothelium of IPAH patients and in cultured ECs from IPAH patients [167]. Activated ECs release granulocyte–macrophage colony-stimulating factor (GM-CSF) [168], CCL2 [159], CXCL12 [17, 169], connective tissue growth factor (CTGF) [170], IL-6 [80] and leptin [171, 172] to promote leukocyte recruitment and accumulation. The accumulated leukocytes release other factors such as macrophage-derived leukotriene B4 (LTB4) [173] that induces PAEC apoptosis, and T-cell lymphocyte-derived MIF [167] that induces the proinflammatory phenotype of ECs and the further recruitment of inflammatory cells. In contrast, regulatory T-cells function to limit endothelial injury and inflammation. VEGFR2 inhibition with SU5416 alone induces severe PH with pulmonary EC apoptosis in T-cell-deficient rats and in a sub-strain of “hyper-responder” Sprague-Dawley rats [174]. Immune reconstituted nude rats exhibit limited lung perivascular inflammation and EC apoptosis and attenuated PH [175].
Pericytes
Pericyte numbers are increased in PAH and pericyte–EC crosstalk also contributes to pulmonary vascular remodelling in PAH [176]. IPAH ECs promote pericyte migration via the release of FGF2 and IL-6 and proliferation via the release of FGF2 [176]. EC-specific disruption of PHD2 increases pericyte coverage of pulmonary arteries [177]. By contrast, pericytes induce the expression of Wnt5a in normal ECs, which promotes the recruitment of pericytes and thereby stabilises the distal arteriolar bed, but not in ECs derived from PAH patients [178]. Accordingly, pulmonary microvascular ECs from PAH patients have a reduced capacity to recruit pericytes. EC-targeted deletion of Wnt5a reduces microvessel pericyte coverage and induces vessel loss, resulting in persistent PH and right heart failure after cessation of hypoxia. Thus, endothelium Wnt5a plays an important role in pericyte recruitment and microvessel stabilisation [178]. Additionally, PAH pericytes have increased levels of pyruvate dehydrogenase kinase 4 (PDK4) [179], correlating with their reduced mitochondrial metabolism, higher rates of glycolysis and hyperproliferation; reducing PDK4 restores pericyte mitochondrial metabolism and cell proliferation, and enhances EC–pericyte interactions, stabilising small vessels [179]. Thus, genes that regulate EC–pericyte interactions could represent novel therapeutic targets to prevent small vessel loss in PAH.
Fibroblasts
EC–fibroblast crosstalk also plays a pathogenic role in PAH. As mentioned above, ECs may undergo EndoMT to become fibroblast-like cells. ECs also secrete factors such as endothelin-1, PDGF and CXCL12 [17, 180] that induce fibroblast migration/recruitment and proliferation. Furthermore, ECs can release factors such as endothelin-1 and IL-6 that induce fibroblast differentiation to myofibroblasts [181, 182], which are highly proliferative, proinflammatory, invasive and produce collagen and other extracellular matrix proteins and a variety of factors contributing to pulmonary vascular remodelling [166, 183]. Adventitial fibroblasts contribute to pulmonary vascular remodelling through several mechanisms, e.g. accumulation of myofibroblasts increases extracellular matrix stiffness, which leads to activation of PAEC proliferation [183]. (Myo)fibroblast-derived matrix metalloproteinase 2 (MMP2) and MMP9 and 15-hydroxyeicosatetraenoic acid (15-HETE) can induce EC proliferation [184]. Fibroblast-released thrombospondin-1 can destabilise EC–EC interactions [185], leading to an injured endothelium, which contributes to pulmonary vascular remodelling.
Endothelial progenitor cells
EPC markers were markedly increased in remodelled arteries from PAH patients, particularly in plexiform lesions that displayed increased stromal cell-derived factor 1 (SDF1) expression [186]. Circulating angiogenic EPC numbers are also increased in PAH patients, and EPCs from PAH patients with BMPR2 mutations have a hyperproliferative phenotype with impaired vascularisation, suggesting that dysfunction of circulating EPCs contributes to PAH vascular remodelling [186]. Clinical studies provide evidence of beneficial effects of EPC transplantation to PAH patients [187–189]. EPC-conditioned medium inhibits EC apoptosis via VEGF-A or VEGF-B and EC proliferation by VEGF-A or IL-8 [180]. Treatment with EPCs or EPC-conditioned medium improves pulmonary artery relaxation, suggesting that paracrine mechanisms promote vasoprotection, such as through the release of prostacyclin and cAMP [190]. This is consistent with work showing that the therapeutic potential of EPCs can be enhanced by transfection of eNOS, another important paracrine signalling pathway [134]. This study also provides evidence that eNOS-transfected EPCs can induce regeneration of lung microvasculature. Together, these studies demonstrate that EPCs can attenuate pulmonary vascular remodelling and PAH development through paracrine mechanisms.
Future perspectives
Despite major advances in the understanding of the pathophysiology of PAH, only a limited number of these mechanistic insights have been translated into approved therapies for PAH patients, mainly relating to the three major endothelial vasoactive pathways. Although such therapies can reduce the symptoms of PAH, they do not prevent progression or cure the disease. The 5-year mortality of PAH is still as high as 40% [7, 191]. Here, we summarise the recent findings on the role of ECs in the pathogenesis of PAH (figure 2). Novel therapeutic advances could occur from an improved understanding of EC-mediated mechanisms that regulate PAH (tables 1 and 2). In this regard, many of the EC signalling pathways described above now represent novel therapeutic targets for the prevention or treatment of PAH (table 3). Novel PAH treatments could aim to 1) inhibit EC injury/apoptosis in the early stages of disease and promote EC regeneration and repair, 2) inhibit apoptosis-resistant EC hyperproliferation, 3) target the EC secretome, 4) inhibit EndoMT or 5) target EC-derived oxidative/nitrative stress. The ideal targets are those nodal signalling molecules that regulate multiple pathways in the pathogenesis of PAH. For example, endothelial PHD2/HIF-2α regulates EC release of PDGF-B, SDF1, endothelin-1 and apelin, among others, and affects BMPR2 signalling and Caveolin1 and PKG expression [17, 64–68]. This option is particularly appealing given that HIF-2α inhibitors are already in clinical trials of renal cancer patients [192–194]. Targeting the mechanisms that regulate metabolic reprogramming of pulmonary ECs may represent a novel therapeutic approach [35, 102, 105]. The epigenetic manipulation of key pathways and miRNAs [85, 103, 113] could also represent a new therapeutic strategy. In terms of specifically targeting lung ECs in the treatment of PAH, several therapeutic options deserve attention in future experimental and clinical studies. It may be possible to deliver PAH treatments in nanoscale delivery vehicles that target the lung endothelium by coating them with antibodies or receptors that enhance EC uptake and retention. For example, small molecule inhibitors or other pharmacological agents could be encapsulated in nanoparticles that have been coated with anti-VWF or anti-CD31 antibodies. Endothelial-enriched nanoparticles can be employed to deliver small interfering RNA (siRNA) oligoes to disrupt gene expression in vascular ECs [195]. Our unpublished data (manuscript under revision) show that a novel nanoparticle can efficiently deliver plasmid DNA targeting the vascular ECs with high genome editing efficiency, which holds great potential for non-viral gene therapy in PAH by genome editing-mediated gene disruption and correction of genetic mutations, etc.
Given the heterogeneity in EC phenotypes [196], future studies should define the subpopulations of ECs in the pathogenesis of PAH by employing single-cell RNA sequencing analysis coupled with genetic lineage tracing and depletion studies. Future research could also employ computational modelling techniques to improve understanding of the relationships between these heterogeneous EC subpopulations and their interactions with neighbouring cell types. Similarly, the potential role of EndoMT in the mechanisms of obliterative vascular remodelling should also be defined with similar genetic lineage tracing approaches. Although PAH has an inflammatory component, it is unclear whether inflammation is a cause or consequence of this disease. Future studies should aim to determine whether and how inflammation triggers the EC dysfunction that contributes to vascular remodelling. Future studies could assess the timing of the EC phenotypic changes in relation to PAH pathogenesis and progression by employing extensive time-course experiments or longitudinal experiments of labelled ECs from healthy to diseased states. Our knowledge of the mechanisms of vascular fibrosis in PAH is also limited; thus, studying the role of EC dysfunction in vascular fibrosis is another important research direction. Targeting EC-derived oxidative/nitrative stress may provide a novel therapeutic approach for treatment of PAH [132, 145].
Conclusions
Dysfunctional EC signalling pathways tightly regulate multiple aspects of PAH, including pulmonary vascular tone, inflammation, coagulation, metabolism and remodelling. Given the increasingly large body of evidence demonstrating that ECs are crucial mediators of PAH initiation and progression, novel therapies for PAH could aim to target multiple aspects of EC dysfunction and EC signalling, especially those nodal signalling molecules that regulate multiple pathways in the pathogenesis of PAH. Improved understanding of the EC signalling pathways responsible for the initiation and progression of PAH will facilitate the development of effective treatments for PAH.
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-03957-2020.Shareable
Footnotes
Author contributions: C.E. Evans wrote the manuscript; N.D. Cober, Z. Dai and D.J. Stewart edited the manuscript; and Y-Y. Zhao revised and finalised the manuscript. All authors approved the final version.
Conflict of interest: N.D. Cober reports grants from Canadian Institute of Health Research and Canadian Vascular Network, during the conduct of the study.
Conflict of interest: Z. Dai reports grants from National Institutes of Health, American Heart Association and American Thoracic Society, during the conduct of the study.
Conflict of interest: D.J. Stewart reports other (founding member, equity stake) from Northern Therapeutics, outside the submitted work.
Conflict of interest: Y-Y. Zhao reports grants from National Institutes of Health/National Heart, Lung, and Blood Institute (R01HL123957, R01HL133951, R01HL140409 and R01HL148810), during the conduct of the study.
Conflict of interest: C.E. Evans reports grants (Career Development Award, 19CDA34500000) from American Heart Association, during the conduct of the study.
Support statement: This work was supported in part by NIH grants R01HL123957, R01HL133951, R01HL140409 and R01HL148810 to Y-Y. Zhao and by an American Heart Association Career Development Award (19CDA34500000) to C.E. Evans. Funding information for this article has been deposited with the Crossref Funder Registry.
- Received October 27, 2020.
- Accepted January 13, 2021.
- Copyright ©The authors 2021. For reproduction rights and permissions contact permissions{at}ersnet.org