The Basic Science of Pulmonary Arterial Hypertension for Clinicians: new concepts and experimental therapies

Introduction

Pulmonary arterial hypertension (PAH) is a syndrome in which pulmonary arterial obstruction increases pulmonary vascular resistance leading to right ventricular (RV) failure and a 15%/year mortality rate. This review highlights recent advances in the Basic Science of pulmonary arterial hypertension (PAH). New concepts clarify the nature of PAH and provide molecular blueprints that explain how PAH is initiated and maintained. Five basic science concepts provide a framework to understand and treat PAH. 1: Endothelial dysfunction creates an imbalance favoring vasoconstriction, thrombosis, and mitogenesis. Restoring this balance by inhibiting endothelin and thromboxane or augmenting nitric oxide and prostacyclin is the paradigm upon which most current therapy is based. 2: PAH has a genetic component. Mutations (Bone morphogenetic protein receptor-2, BMPR-2) and single nucleotide polymorphisms (ion channels and transporter genes) predispose to PAH. 3: Excess proliferation, impaired apoptosis and glycolytic metabolism in pulmonary artery smooth muscle, fibroblasts and endothelial cells suggest analogies to cancer. Many experimental therapies reduce PAH by decreasing the proliferation/apoptosis ratio, including inhibitors of: pyruvate dehydrogenase kinase, serotonin transporters, survivin, 3-hydroxy-3-methylglutaryl-coenzyme A reductase, transcription factors (hypoxia-inducible factor-1α and nuclear factor activating T-lymphocytes) and tyrosine kinases. Augmentation of voltage-gated K+ channels, Kv1.5 and BMPR-2 signaling also address this imbalance. Tyrosine kinase inhibitors used to treat cancer are in Phase-1 PAH trials. 4: Refractory vasoconstriction may occur due to rho kinase activation. Only <20% of PAH patients respond to conventional vasodilators; however refractory vasoconstriction may respond to rho kinase inhibitors. 5: The RV can be therapeutically targeted. Although increased afterload initiates RV failure, the major cause of death/dysfunction in PAH, the RV may be amenable to cardiac-targeted therapies. The RV in PAH has features of ischemic, hibernating myocardium.

Guided by these new concepts and armed with better understanding of disease mechanisms we are poised to identify new therapeutic targets. To achieve balance in a rapidly evolving field, we invited colleagues to contribute figures illustrating pathways in their area of expertise that are important to the pathogenesis and treatment of PAH. These contributors are acknowledged in the Figure legends.

Epidemiology

There are 5 categories of pulmonary hypertension in the latest World Health Organization's classification of pulmonary hypertension (PH): 1: PAH, 2: PH associated with left heart disease, 3: PH associated with lung disease/hypoxia, 4: thromboembolic PH, 5: miscellaneous¹. This review focuses on Category 1 (PAH), which includes: idiopathic and familial PAH, as well as PAH associated with a variety of conditions (including connective tissue diseases and congenital heart disease), pulmonary venoocclusive disease, pulmonary capillary hemangiomatosis, and persistent pulmonary hypertension of the newborn. The incidence and prevalence of PAH is estimated at 2.4 cases/million/year and 15 cases/million in France² and 7.6 cases/million/year and 26 cases/million in Scotland³. The global prevalence of PAH is hard to estimate because accurate diagnosis of PAH is difficult and access to care is limited in many countries. Because diseases which are risk factors for PAH, such as HIV, schistosomiasis and sickle cell disease, are more prevalent in the developing world, the global burden of PAH is likely greater than is currently recognized⁴. In developed countries prevalence will also likely increase as newer associations with PAH emerge, including dialysis⁵ and the metabolic syndrome⁶, and as widespread access to echocardiography identifies PAH earlier and in more individuals.

Definition

PAH is a small subset of pulmonary hypertensive syndromes (WHO categories 2-5). PAH is defined by a resting mean pulmonary artery pressure (PAP) >25mmHg, pulmonary vascular resistance (PVR) >3 Wood units, and pulmonary capillary wedge pressure <15mmHg (in the absence of other causes of PH). Unlike PAH, PH is ubiquitous, the sole diagnostic criterion being a resting mean PAP >25mmHg. This larger PH group often does not have intrinsic pulmonary vascular disease. Their PH is due to high flow, elevated left ventricular end diastolic pressure, lung disease/hypoxia or valve disease. There is no randomized clinical trial evidence that WHO category 2-5 patients benefit from PAH-specific therapies.

Prognosis

The 1-year, incident mortality rate of PAH remains high (15%), despite treatment with prostacyclin, endothelin antagonists and phosphodiesterase-5 inhibitors⁷. Because PAH is a syndrome, it is not surprising that the prognosis varies depending on the associated comorbid conditions. Prognosis in PAH associated with congenital heart disease tends to be better than in idiopathic PAH (iPAH) (3-year survival rates 77% versus 35%)⁸, although other cohorts found similar survival in iPAH to be 65% at 3 years⁹. PAH associated with scleroderma has 3-year survival rates of only 34-47%^{10, 11}.

Current Therapies

Treatment of PAH involves the use of prostanoids (given intravenously, by inhalation, subcutaneously or orally), endothelin receptor blockers, and/or phosphodiesterase 5 inhibitors. L-type calcium channel blockers (e.g. nifedipine) can be effective but are only safe for use in patients who respond to an acute vasodilator challenge with a >20% fall in mean PAP and no decline in cardiac output (a subset representing 12%-20% of PAH patients)¹². Most patients empirically receive anticoagulation to prevent thrombosis in situ and diuretics to limit edema. PAH treatments remain expensive and/or difficult to deliver and are more palliative than curative. A year of sildenafil is estimated to cost $13,000 versus ∼$56,000 for bosentan while costs for inhaled iloprost and intravenous prostacyclin exceed $90,000/year. The only randomized PAH study that has shown survival benefit used intravenous Flolan, which, compared to conventional therapy, decreased mortality in 81 WHO Class IV patients¹³. Thus there is a pressing need for less expensive and more effective therapies.

Most current treatments (prostacylin, endothelin antagonists and warfarin) address endothelial dysfunction by augmenting vasodilator and anti-proliferative mediators and inhibiting vasoconstrictor, prothrombotic and mitogenic pathways. Our increasing knowledge of the cellular and molecular basis of PAH suggests many potential new therapeutic agents.

Histology

The histological findings in PAH include intimal hyperplasia, medial hypertrophy, adventitial proliferation/fibrosis, occlusion of small arteries, thrombosis in situ, and infiltration of inflammatory/progenitor cells. Angioproliferative ‘plexiform’ lesions are found in PAH, but not in other PH categories (Figure 1). Plexiform lesions (and other complex lesions) are often located downstream from occluded arteries and express the transcription and growth factors typically seen in angiogenesis, including vascular endothelial growth factor (VEGF) and hypoxia inducible factor (HIF-1α)¹⁴ (Figure 2). PAH typically spares the airway, veins, bronchial circulation, capillaries and systemic vasculature (Figure 1). The various histological abnormalities of PAH are heterogeneous in their distribution and prevalence within the lungs. The natural progression of lesion severity (presumably from medial hypertrophy to plexiform arteriopathy) and the functional relevance of plexiform lesions remain uncertain, although regression of histologically proven PAH has been documented following single lung transplantation¹⁵. Human lung tissue is invaluable, offering cells for culture, histological sections for immunohistological assessment of pathogenetic pathways, and tissue to be mined by laser capture microdissection for biomarkers. It remains the gold standard against which to judge animal models.

Figure 1. Histology of PAH.

Upper panels show plexiform lesions. Left panel- note the evidence of cell proliferation (red is proliferating cell nuclear antigen, green is smooth muscle actin, blue is DAPI). Lower panels show medial hypertrophy, intimal fibrosis and adventitial proliferation.

Figure and legend contributed by Dr. S.L. Archer

Figure 2. Formation of complex and plexiform lesions in PAH.

Transformation of an arteriole, into a complex vascular lesion with near total or total lumen obliteration, usually occurs at a vessel bifurcation. The concept depicted is one of initial apoptosis of cells forming the endothelial monolayer (left). Disorganized endovascular angiogenesis results from proliferation of phenotypically abnormal cells due to 1) phagocytosis of apoptotic monolayer endothelial cells by neighboring endothelial cells 2) activation of stem cell-like endothelial cells or 3) attachment of bone-marrow derived “repair cells” to the injured endothelium. Bone marrow participation in the formation of these lesions is postulated because megakaryocytes, mast cells and dendritic cells can be released and attach to the injured vessel. Perivascular lymphocytes may cluster in the lymphatics adjacent to the adventitia. The lesion also shows a dysregulated matrix. Growth factors released by megakaryocytes and mast cells may contribute to the angiogenic growth and T and B-lymphocytes may reflect a local immune response. The table insert lists the phenotypic changes seen in plexiform lesions. Figure and legend contributed by Dr. Norbert Voelkel, Virginia Commonwealth University, Richmond, VA.

Animal Models

The evaluation of these novel targets occasionally involves the off-label use of drugs approved for another indication (e.g. Gleevec), but is largely based on studies in cellular and animal models. Cautious interpretation of preclinical studies is mandatory and one must recognize the strengths/weakness of various animal models and the risks of extrapolation to humans with PAH. Notably, no animal model completely recapitulates human PAH. Promising rodent models include: monocrotaline-treated rats ± pneumonectomy¹⁶ or abdominal aortocaval shunt¹⁷, fawn-hooded rats (which spontaneously develop PAH and are also hypoxia-sensitive)^{18, 19} and rats treated with a single dose of VEGF-receptor antagonist (SU5416) plus hypoxia²⁰. Models combining multiple insults yield more severe PAH with better hemodynamic and histological fidelity to human PAH. This may be relevant to the pathogenesis of human PAH, which also appears to require multiple “hits”. Murine models of PAH offer mechanistic insight on the relevance of single genes. Mice that transgenically overexpress serotonin transporters (SERT)²¹, bone morphogenetic protein receptor (BMPR-2) dominant-negative mutations²² or S100A4/Mts1²³ develop PH.

New Paradigms

PAH was once regarded largely as a disease of excess vasoconstriction. This view was incomplete and new concepts help us understand the fundamental causes of this syndrome.

PAH is a panvasculopathy (Figure 3)

Figure 3. PAH is a panvasculopathy.

Abnormalities can be seen at each level of the small pulmonary arteries, beginning in the blood and travelling outward to the adventitia. While most of these abnormalities are likely secondary (rather than being the initiating cause of PAH), they nonetheless offer interesting therapeutic targets

Contributed by Dr. S.L. Archer

Let's take a tour of the molecular pathology of PAH, beginning at the lumen of a small pulmonary artery (PA). In the blood, levels of serotonin, a proliferative, fibrogenic, vasoconstrictor, are elevated²⁴ (Figure 4). In the endothelium the vasodilator/vasoconstrictor ratio is decreased ^25-27 (Figure 5) whilst prothrombotic factors, including tissue factor²⁸, are increased. It is hypothesized that widespread endothelial apoptosis, early in PAH, culminates in selection of apoptosis-resistant endothelial precursor cells that proliferate and eventually form plexiform lesions²⁰ (Figure 2). In the media, pulmonary artery smooth muscle cell (PASMC) apoptosis is suppressed and proliferation is enhanced^38,41,43-45. Many factors drive PASMC proliferation, including mutation²⁹ or downregulation³⁰ of BMPR2 (Figure 6), mitochondrial metabolic abnormalities (Figure 7), de novo expression of the anti-apoptotic protein survivin^{20, 31}, increased expression/activity of SERT^{21, 32}, increased expression/activity of platelet-derived growth factor (PDGF) receptor⁴⁹, tyrosine kinase activation³³ (Figure 8) and decreased expression of Kv1.5, a voltage-gated, O₂-sensitive potassium channel. Kv1.5 downregulation occurs in human PAH³⁴, rat PAH models (whether induced by chronic hypoxia^{35, 36}monocrotaline³¹ or in FHR¹⁹) and transgenic mice with PAH due to SERT overexpression²¹ or BMPR2 mutation³⁷. Loss of Kv1.5, the same channel that is inhibited by hypoxia to initiate hypoxic pulmonary vasoconstriction⁵⁵, depolarizes the membrane and elevates cytosolic K⁺ and Ca²⁺(Figure 9). The resulting calcium overload, later reinforced by activation of transient receptor potential (trp) channels³⁸, leads to Ca²⁺-calcineurin-dependent activation of the proliferative transcription factor, nuclear factor acting T lymphocytes (NFAT)³⁹. Normoxic activation of HIF-1∝ occurs in FHR¹⁹ and human PAH^{14, 19}. In the adventitia, metalloprotease activation causes architectural disruption, permitting cell migration and generating mitogenic peptides (tenascin)⁴⁰(Figure 10). Adventitial fibroblasts are also hyperproliferative in PH, displaying increased sensitivity to serotonin⁴¹. Circulating autoantibodies⁴ and lung infiltration by inflammatory cells is common, particularly in PAH associated with connective tissue disease and schistosomiasis⁴² (Figure 11). Finally, there are increased endothelial-precursor cells, and mesenchymal and bone marrow-derived stem cells⁴³, although it is uncertain whether this is harmful or beneficial (Figure 12).

Figure 4. Serotonin (5-HT) abnormalities in PAH.

Increased bioavailability of serotonin during progression of PAH results from an increased release of serotonin from platelets and from an increased synthesis of serotonin by endothelial cells which produce serotonin and express tryptophan hydroxylase-1 (TPH1), the key enzyme controling 5-HT synthesis. Overexpression of 5-HTT (SERT) by PASMC is responsible for the increased mitogenic effect of serotonin on these cells. 5-HT receptors, including 5-HT1B/1D and 5-HT2A receptors mediate 5-HT-induced PA contraction of pulmonary vessels. 5-HT2A receptors located on platelets potentiate the aggregation response to various platelet activators. 5-HT2B receptors expressed by PASMC are also involved in the pulmonary vascular remodeling process.

Figure and legend contributed by Dr. Serge Adnot, Département de Physiologie, Hôpital Henri Mondor, CRETEIL, FRANCE

Figure 5. The endothelium and vasodilator/antiproliferative pathways.

Nitric oxide (NO), generated from L-arginine, and natriuretic peptides stimulate production of cyclic guanosine monophosphate (GMP). cGMP causes vasorelaxation and inhibits proliferation of vascular SMC. Phosphodiesterase type 5 (PDE5) inhibitors (e.g. sildenafil) enhance this vasodilatory mechanism by preventing cGMP degradation. Prostacylin from endothelial cells also promotes relaxation and inhibits cell proliferation, via a cyclic AMP-dependent mechanism. Endothelin is a potent vasoconstrictor and stimulates proliferation via ET_A receptors on SMC; while stimulating NO and prostacylin release via endothelial ET_B receptors. Adrenomedullin (AM) and vasoactive intestinal polypeptide (VIP) are additional endothelial-derived, cAMP-dependent vasodilators that are dysregulated in PAH.

Contributed by Dr. Martin Wilkins

Figure 6. BMPR2 mutations-a genetic basis for familial PAH.

BMPR2 mutations are found throughout the gene, and a universal functional consequence of these mutations has not been identified. Best studied is BMPR1 signaling through SMAD transcription factors. Mutations leading to loss of SMAD signaling decreases cell differentiation, enhance vascular tone, increase TGF-β signaling and likely increase proliferation. Signaling through XIAP (X-linked inhibitor of apoptosis), also requiring BMPR1, can impact both NFkB and MAPK pathways, leading to increased MAPK phosphorylation and presumably pro-inflammatory signaling. BMPR2 has a long, evolutionarily conserved cytoplasmic tail domain unique in the TGF-beta superfamily, which binds SRC, RACK1, and LIMK1. BMPR2 mutation in vivo leads to decreased Cofilin (Cfl1) phosphorylation by LIMK1, with the effect both of alterations in F-actin organization and defects in glucocorticoid receptor (GR) nuclear translocation.

Figure and legend contributed by Drs. James West and John H. Newman, Pulmonary Medicine, Vanderbilt University School of Medicine, Nashville, Tennessee.

Figure 7. Mitochondrial metabolism in PAH.

In aerobic metabolism, PDK is inactive, PDH is active and electron donors (mitochondrial NADH and FADH) produced by the TCA (Krebs') cycle pass electrons down a redox-potential gradient in the electron transport chain (ETC) to molecular O₂. This electron flux powers H+ ion extrusion, creating the proton-motive force responsible for creating the mitochondria's negative membrane potential (ΔΨm) and powering F1Fo-ATP-synthase. Side reactions between semiquinones and molecular O₂, accounting for ∼3% of net electron flux, create superoxide anion in proportion to PO₂. Superoxide dismutase (SOD2) rapidly converts superoxide anion (produced at complexes I and III) to H₂O₂, which serves as a redox messenger signaling “normoxia”. In hypoxia (and PAH and cancer) there is activation of HIF-1α and PDK, which inhibits PDH shifting metabolism toward glycolysis. Therapeutic implications: Dichloroacetate which inactivates PDK by causing conformational changes in its nucleotide- and lipoyl-binding pockets⁶⁰) regresses experimental PAH⁷,³³.

Contributed by Dr. Stephen L. Archer

Figure 8. Receptor tyrosine kinases and their inhibitors.

This complex kinase cascade offers many therapeutic targets to treat PAH.

Abbreviations: ATF, activating transcription factor; BAD, BCL-XL/BCL-2 associated death promoter; c-kit, CD117; DAG, diacyl glycerol; EGF-R, epidermal growth factor receptor; ErbB1,2 epidermal growth factor receptor; ERK, extracellular signal-regulated kinase; flt3, fms-like tyrosine kinase receptor-3; GSK, glycogen synthase kinase; JAK, Janus kinase; JNK, Jun N-terminal kinase; MEF, myocyte-specific enhancer-binding nuclear factor; MEK, mitogen-activated protein kinase/ERK kinase; MERM, ezrin/radixin/moezin family of cytoskeletal linkers; mTOR, mammalian target of rapamycin; NHERF, sodium-hydrogen exchange regulatory factor; P, phosphotyrosine; p70S6K, p70 ribosomal S6 kinase; PIP2, phosphatidylinositol-4,5-bisphosphate; PIP3, phosphatidylinositol-3,4,5-trisphosphate; PKB, protein kinase B; PKC, protein kinase C; PLC, phospholipase C; SOS, Son of Sevenless; STAT, signal transducer and activator of transcription; PDGF-R, platelet-derived growth factor receptor: PDK, phosphoinositide-dependent kinase; PI3K, phosphoinositide-3 kinase; PKB, protein kinase B; PKC, protein kinase C; PLC, phospholipase C; SHP, Src homology 2-containing protein tyrosine phosphatase; VEGF-R, vascular endothelial growth factor receptor.

Figure and legend contributed by Dr. Ralph Schermuly, Max-Planck-Institute for Heart and Lung Research, Bad Nauheim, Germany

Figure 9. Ion Channels in PAH.

Schematic depiction of the cellular mechanisms linked with the vasoconstriction and remodeling in pulmonary endothelial (PAEC) and PASMC in PAH. Central themes of interest for the development of PAH include: 1) impaired ion channel expression and function in PASMC (Kv, VDCC, SOC, ROC), 2) increased cytosolic calcium ([Ca2+]cyt) in PASMC (mediated by ion channel function and receptor stimulation), 3) altered signaling via membrane receptors (GPCR, TIE-2, BMPR, RTK) and transporters (i.e., SERT) in endothelial cells and PASMC, 4) changes in redox status, 5) enhanced production of vasoconstrictor or mitogenic factors, and 6) viral signaling via GPCR and RTK. Paracrine interactions between PAEC and PASMC are noteworthy.

Abbreviations: Ang-1, angiopoietin-1; ET-1, endothelin-1; GPCR, G protein-coupled receptor; 5-HT, serotonin; MAPK, mitogen-activated protein kinase; NO, nitric oxide; PDGF, platelet-derived growth factor; PGI2, prostaglandin I2; ROC, receptor-operated Ca2+ channels; ROCK, Rho-associated kinase; ROS, reactive oxygen species; RTK, receptor tyrosine kinase; SERT, 5-HT transporter; SOC, store-operated Ca2+ channels; SR, sarcoplasmic reticulum; VDCC, voltage-dependent Ca2+ channels.

Figure and Legend Contributed by Drs. Carmelle Remillard and Jason Yuan, University of California, San Diego

Figure 10. Disordered Elastin metabolism and deposition in PAH.

Elastase degrades elastin and other components of the extracellular matrix thereby releasing bound growth factors, that are both mito- and motogenic for PASMC. Heightened elastase activity also activates matrix metalloproteinases (MMPs), which upregulate the glycoprotein tenascin-C. When tenascin-C binds cell surface integrins, such as alpha-v β3 on PASMC, these integrins cluster and cell shape changes in a way that clusters and activates growth factor receptors and increases cell survival signals. Thus, pathway activation causes both release of growth factors and activation of their receptors. Transmission of cell survival signals occurs even in the absence of ligand (growth factor) binding. Blocking elastase activity or growth factor receptors can therefore arrest progression of PASMC by blocking proliferation and induce regression by enhancing apoptosis.

Figure and legend contributed by Marlene Rabinovitch, Stanford University, Palo Alto, CA.

Figure 11. The role of inflammation in the pathogenesis of PAH.

Initial inflammatory stimuli can occur in the form of infectious or foreign antigens or autoimmune disease, leading to an appropriate, but potentially excessive, immune response. The host immune response to these varied stimuli results in the release of pro-inflammatory cytokines, which can recruit bone marrow-derived cells, stimulation of resident inflammatory cells, and endothelial cell dysfunction. Endothelial cell injury and the cellular response can increase endovascular thrombosis. A network of cytokines released by the inflammatory and endothelial cells can also cause aberrant PASMC proliferation. The triad of endothelial cell proliferation, PASMC proliferation, and thrombus formation contributes to PAH. Pro-inflammatory cytokines and cell-cell interactions can potentially be therapeutically targeted.

Abbreviations: EGF: epidermal growth factor; HIMF: hypoxia induced mitogenic factor, also called RELMa and FIZZ1; NO: nitric oxide; PDGF: platelet derived growth factor; RANTES; TNFα: tumor necrosis factor-alpha.

Figure and legend contributed by Drs. Brian Graham and Rubin Tuder, University of Colorado at Denver.

Figure 12. Cellular Basis for Pulmonary Vascular Remodeling-lessons from hypoxia.

Fibroblasts, monocytes and fibrocytes play critical roles in orchestrating hypoxia-induced pulmonary vascular remodeling. Hypoxia or hypoxia-associated stimuli increase production by resident fibroblasts (and probably PASMC) of chemokines/cytokines including: monocyte chemoattractant protein (MCP)-1, stromal cell-derived factor (SDF)-1, fractalkine (CX3CL1), RANTES, VEGF, osteopontin (OPN) and endothelin. These and other factors stimulate recruitment of monocytes and monocyte-derived mesenchymal precursors (fibrocytes) to the vessel wall. Upregulation of monocyte receptors for these ligands (CCR2, CXCR4, CX3CR1, VEGFR-1 and ETR-A) occurs. Monocytes are retained in the vessel wall by the upregulation of adhesion molecules on fibroblasts, including vascular cell adhesion molecule (VCAM), intracellular adhesion molecule (ICAM) and OPN. As monocytes and fibrocytes accumulate in the vessel wall, they exert potent effects on the proliferative, migratory, matrix-producing and contractile capabilities of resident fibroblasts and PASMC through the secretion of TGF-β, PDGF-A and B, EGF, IL-6, IGF-1, MMP-9 and others. In addition, these cells produce potent proangiogenic molecules such as VEGF, S100A4, βFGF that likely play roles in stimulating further angiogenesis in the vessel wall.

Figure and legend contributed by Kurt Stenmark, University of Colorado at Denver.

PAH has a genetic component (Figure 6)

The bone morphogenetic proteins (BMP) are part of the transforming growth factor-β superfamily (TGF-β. Most (>80%) patients with familial PAH have loss of function mutations in BMPR2^44-46 that promote cell proliferation. BMPR2 is a constitutively active serine-threonine kinase receptor, which, in response to ligand (BMPs 2, 4, 6, 7, 9 and 10), forms heterodimers with any of 4 type 1 receptors (BMPR1A, BMPR1B, Alk1, Alk2), resulting in phosphorylation of the intracellular portion of the type 1 receptor by BMPR2. Receptor activation initiates a cytosolic Smad protein-signaling cascade. Receptor activated Smads, complex with common partner SMAD (Smad4) and the complex translocates to the nucleus where it regulates gene transcription (Figure 6). The inhibitors of DNA binding (Id) genes are major targets of BMP/Smad signaling⁴⁷. The Smad DNA interaction is weak and requires co-repressors or activators. BMPs can also act via an alternative, BMPR2-independent pathway involving mitogen activated protein kinases (e.g. p38MAPK, ERK1/2).

Most heterozygous BMPR2 mutations in PAH result in defective Smad signaling, while retaining p38MAPK signaling^66,48. The loss of normal BMPR2-Smad activity may exaggerate the susceptibility of vascular cells to proliferate and suppress apoptosis. BMP 2, 4 and 7 suppress PASMC proliferation in normal individuals and patients with secondary PH but are ineffective in PAH²⁹. The BMPR2-Smad pathway may display tissue heterogeneity because it can be regulated by endogenous Smad inhibitors (e.g. chordin and noggin), by inhibitory Smads (6 and 7) and also because of variable heterodimer receptor composition⁴⁹. This also may explain the restriction of the vascular disease to the pulmonary circulation.

Mice with conditional, endothelial BMPR2 deletions are predisposed to PAH, although PH occurs in only a subset, reminiscent of the incomplete penetrance seen in familial PAH⁵⁰. Mice with a SMC-specific, overexpression of a BMPR2 dominant-negative mutant develop a vasospastic form of PH that lacks vascular remodeling but is associated with downregulation of Kv1.5 expression. PH in these mice is reversed by nifedipine³⁷. Perhaps disordered BMP signaling, by reducing Kv1.5 transcription, creates an early vasospastic form of PAH that in time becomes fixed by vascular remodeling.

Initial enthusiasm that BMPR2 mutations might represent a “universal” cause of PAH has been tempered. BMPR2 mutation is uncommon (prevalence 10-20%) in the nonfamilial, Category 1 PAH population. Moreover, in familial PAH, penetrance is low (i.e. only ∼25% of carriers in affected families develop PAH)⁵¹. While modifier genes, such as SERT and TGFβ, may explain variable penetrance, aberrant BMPR2 function alone is neither a necessary nor a sufficient precondition for most cases of PAH⁵². Moreover, BMPR2 heterozygous mice do not develop PAH and are not predisposed to hypoxic PH; however they do have an exaggerated hypertensive response to serotonin⁵³.

In genetically normal animal, BMPR2 expression decreases as PH develops⁵⁴. One posttranscriptional mechanism accounting for downregulation of BMPR2 protein is activation of microRNAs (miRNAs). miRNAs regulate gene expression by inhibiting translation. A computational algorithm on the BMPR2 gene predicted that miRNAs encoded by miRNA cluster 17/92 (miR-17/92) might regulate BMPR2. Overexpression of miR-17/92 did reduce BMPR2 protein, and it appears that BMPR2 is directly targeted by miR-17-5p and miR-20a⁵⁵.

however, results of BMPR2 rescue therapy are mixed. Intravascular BMPR2 gene therapy, using an endothelial-targeted vector, reduced chronic hypoxic pulmonary hypertension in rats⁵⁶. However nebulized BMPR2 adenovirus (with a promiscuous promoter) did not regress monocrotaline-induced PAH⁵⁴. Further study is required and may be productive in identifying BMPR2-related targets for pharmacological manipulation in PAH. For example, inhibition of TGF-β signaling prevents PAH in the monocrotaline model, via inhibition of activin receptor-like kinase-5⁵⁷.

Alternative genetic mechanisms

Work is currently underway searching for modifier genes and for possible epigenetic mechanisms (gene methylation) of inheriting PAH or enhancing disease susceptibility. Single nucleotide polymorphisms (SNP) are genes that differ from normal by a single alternative nucleotide. SNPs can change the function/location of the encoded protein. SNPs occur in a significant proportion of the population and may explain susceptibility to PAH. SNP variants for PAH-relevant genes (including SERT⁵⁸, Kv1.5⁵⁹ and TRPC6) may predispose to PAH. The consequences of SNPs can be complex, for example the TRPC6 SNP not only increases TRPC6 expression but also creates a binding sequence and activates the inflammatory transcription factor, nuclear factor-kappaB⁶⁰.

Excess proliferation and impaired apoptosis suggest similarities to cancer in PAH (Figure 7)

Otto Warburg, 1931 Nobel laureate, proposed that a shift in glucose metabolism from oxidative phosphorylation to glycolysis (despite adequate oxygen supply) was central to the cause/maintenance of cancers. Several observations indicate that PAH shares this “Warburg phenotype”^{19, 61}. As highlighted by Tuder and Voelkel⁶², both cancer and PAH manifest excessive cell proliferation and impaired apoptosis. While PAH does not metastasize or disrupt tissue boundaries, emerging data show it shares a mitochondrial-metabolic abnormality with cancer (Figure 7). Pyruvate dehydrogenase kinase (PDK) is pathologically activated in both conditions^{19, 63}. This enzyme phosphorylates and inhibits pyruvate dehydrogenase (PDH)⁶⁴. PDH catalyses the irreversible oxidation of pyruvate, yielding acetylCoA and CO₂ and is a key enzyme controlling the rate of oxidative metabolism. PDK activation thus impairs the Krebs' cycle and creates a glycolytic shift in glucose metabolism. Subversion of the mitochondrial O₂-sensing mechanism, normally used to sense and respond to decreases in PO₂⁶⁵, appears to cause the sensor to signal hypoxia despite adequate PO₂. These acquired (and reversible) mitochondrial abnormalities of fusion/fission and metabolism⁶¹ are postulated to cause the observed normoxic activation of HIF-1∝ in PAH^{14, 66} and cancer ⁶³. Once active, HIF-1α turns on glycolytic genes and suppresses oxidative metabolism by increasing PDK transcription. The downstream consequences of this mitochondrial-metabolic abnormality include mitochondrial hyperpolarization, reduced production of reactive oxygen species, decreased Kv1.5 expression.

These metabolic abnormalities, which enhance cell proliferation and impair apoptosis, can be partially corrected by a simple, mitochondrial-targeted strategy. Dichloroacetate, a PDK inhibitor, restores PDH activity, increases glucose oxidation, restores mitochondrial membrane potential and reverses normoxic HIF-1α activation⁶⁷. Dichloroacetate inhibits proliferation, enhances apoptosis and can regress both human cancer, in a xenotransplantation model⁶³, and experimental PAH (chronic hypoxic PH, monocrotaline PAH and FHR PAH)^{19, 36, 67}. Dichloroacetate also regresses spontaneous PH in transgenic mice that over express SERT in PASMC⁶⁸. Dichloroacetate has been safely used in long-term treatment of inherited lactic acidosis due to mitochondrial diseases, suggesting the potential for translation to the clinic.

The right ventricle (RV) in PAH

The fetal/neonatal RV ejects blood at relatively high pressure into the pulmonary circulation. With maturation the pulmonary circulation develops into a low-pressure circuit and the RV involutes, becoming thin walled. Chronic pressure overload, as occurs in PAH, stimulates RV hypertrophy. Surprisingly little is known about the specific mechanisms underlying RVH and RV dysfunction in the setting of PAH. While the obvious approach to reducing RVH and RV failure is to treat the underlying pulmonary arterial disease, recent experimental evidence suggests that the RV can be therapeutically targeted in PAH³⁶. In RVH phosphodiesterase-5, which was expressed in the fetal RV, is selectively re-expressed. Inhibiting this enzyme (i.e. by sildenafil) enhances RV contractility without affecting the left ventricle⁶⁹ which lacks phosphodiesterase-5.

In contrast to the normal RV, which can vary its substrate utilization from fatty acids to glucose as needed, metabolism in RVH is reliant on glucose metabolism⁷⁰. In hypoxia-induced PH, expression of the glucose transporter GLUT4 is significantly increased in the RV, suggesting a metabolic switch to glycolysis. AMP-activated protein kinase (AMPK), which has a key role in the control and regulation of energy metabolism, stimulates fatty acid metabolism and glycolysis, preserving ATP production^{71, 72}. AMPK activation in ventricular hypertrophy^73-75 preserves ATP levels by increasing glucose transport and accelerating glycolysis and by inhibiting acetyl CoA carboxylase⁷². In RVH, there is systolic flow impediment in the right coronary artery that is proportional to RV pressure and mass⁷⁶. New evidence shows the RV in PAH is glycolytic, due in part to activation of PDK, and behaves as hibernating myocardium, demonstrating enhanced glucose oxidation and improved contractility in response to dichloroacetate⁷⁷. Future PAH therapies should consider the effects of agents on both the RV and the pulmonary vasculature.

Therapeutic Pathways in PAH (Table 1)

Table 1. Molecular Targets in the Treatment of Pulmonary Hypertension Current Therapies.

Target	Goal	Drug	Used in Humans	Ref
L-type Ca++ channels	Decrease SMC Ca++	L-type Ca++ channel blockers	Yes	194
Coagulation cascade	Decrease thrombosis	Warfarin	Yes	195-197
Prostacyclin receptors	Increase cAMP	epoprostenol	Yes	198
Endothelin receptors A& B	Inhibit constriction & proliferation	Bosentan	Yes	199
Endothelin receptor A	Inhibit constriction & proliferation	Sitaxsentan	Yes	200
Phosphodiesterase 5 inhibitors	Increase cGMP	Sildenafil	Yes	201
Guanylate cyclase activators	Increase cGMP	iNO	Yes	202
Novel Targets
Rho Kinase	Decrease Ca++ sensitivity in SMCs	Rho kinase inhibitors-fasudil	Yes Acute trial	140, 141
Rho A prenylation	Decrease Ca++ sensitivity in SMCs	Statins	Yes	16
Serine elastases	Decrease MMP activation	Elastase inhibitors	No	40
Kinase associated receptors	Inhibit PDGF or EGF activity	Tyrosine kinase inhibitors Publication pending	imatinib sorafenib	165, 168
PDH kinase	Normalize mitochondrial function	Dichloroacetate	Yes*	36
NFAT	Decrease antiapoptotic bcl-2; slow proliferation	Cyclosporine A	Yes*	39
Immune system	Immunosuppression	Mycophenolate mofetil	Yes*	203
Survivin	Inhibits antiapoptotic effect of survivin	Transfection of dominant-negative	Yes*	31
Guanylate cyclase	Increase cGMP	Direct sGC activators eg. Riociguat (BAY 63-2521) DHEA	Phase III trial	107, 204
Cyclooxygenase	Inhibit TxA2	aspirin	Yes*	28
Ornithine decarboxylase	Inhibit polyamine synthesis	α-difluoromethylornithine	No
Cyclin-dependent-kinase inhibitor γ27	Inhibit SMC proliferation	Heparin	Yes*	205
Peroxisome proliferator-activated receptor –Y	Increase PPARγ activity	Rosiglitazone	Yes*	206
Angiopoietin & TIE2	Inhibit SMC proliferation	Adenoviral gene transfection	No-and conflicting results	173
Serotonin transporter	Inhibit SMC proliferation	SSRI	Yes†	134
VPAC 1 & 2 receptors	Inhibit SMC proliferation	Vasoactive intestinal peptide, inhaled	Yes	125
Adrenomedullin receptors	Inhibit SMC proliferation, vasodilatation	Adrenomedullin	Yes++	207, 208
BMPR2	Enhance BMPR2 signaling	Adenoviral transfection and/or enhancing receptor trafficking to membrane	No conflicting results	30, 126, 209
eNOS	Increase cGMP and NO signaling	eNOS transfected EPCs	Phase I trial in progress	188
eNOS	Increase cGMP and NO signaling	VEGF transfected Fibroblasts	No	191

Abbreviations: DHEA: dihydroepiandrosterone. MMPs: matrix metalloproteinases. PDH: pyruvate dehydrogenase kinase. SSRI: Selective serotonin reuptake inhibitors

Human Use qualifiers:

Yes*=used in human but in a disease other than PAH

Yes†=used in human with PAH but retrospective data

Yes++†=used in human with PAH but only in an acute hemodynamic study

Prostanoids and prostanoid receptors (Figure 5, Table 1)

One of the most successful therapeutic strategies for PAH has been to augment endogenous prostacyclin production with exogenous prostanoids. Fatty acid cyclooxygenase (COX) converts arachidonic acid to prostaglandin H₂ (PGH₂), a substrate for both PGI₂ (prostacyclin) synthase and thromboxane synthase. PGI₂ synthase is expressed in pulmonary vascular endothelium and generates prostacyclin, which relaxes pulmonary artery smooth muscle cells (PASMC) and inhibits platelet aggregation through interaction with IP receptors and stimulation of cyclic adenosine monophosphate (cAMP). Thromboxane synthase, in platelets and endothelium, produces thromboxane A₂ (TXA₂). TXA₂ stimulates vasoconstriction and platelet aggregation through TP receptors. Endothelial dysfunction and platelet activation in PAH reduce prostacyclin levels and increases TXA₂ production.

Continuous intravascular infusion of epoprostenol (Flolan®) decreases pulmonary vascular resistance (PVR), increases cardiac output, and improves life expectancy¹³. Epoprostenol's poor stability, expense and requirement for continuous intravenous infusion have fostered development of more stable analogs and alternative routes of administration: iloprost (inhalation), treprostinil (subcutaneous) and beraprost (oral). New prostacylin agonists and thromboxane antagonists are in clinical trials. The combination of a prostanoid (such as iloprost) and a phosphodiesterase-5 (PDE5) inhibitor enhances pulmonary hemodynamic effects and improves exercise capacity in PAH^{78, 79}.

Nitric oxide (NO) and cyclic guanosine monophosphate (cGMP)-(Figure 5, Table 1)

NO is a radical, synthesized from L-arginine by three nitric oxide synthases (NOS). Endothelial NOS (eNOS) is the principle mediator of endothelium-dependent vasodilatation in the pulmonary circulation. Endothelium-derived NO diffuses into PASMC where it stimulates soluble guanylate cyclase (sGC) to produce cGMP. cGMP's cardiovascular effects are mediated by interaction with at least three groups of proteins: cGMP-dependent protein kinases (PKG), cGMP-regulated PDE, and cyclic nucleotide-gated ion channels. PDE5, the molecular target of sildenafil (Viagra®), decreases intracellular cGMP levels and opposes PKG-dependent signaling elicited by NO and natriuretic peptides.

The NO pathway is impaired in several ways in PAH. NOS expression⁸⁰ and NO biovailability⁸¹ are depressed. Moreover, PDE5 is induced both in PASMC⁸² and the right ventricle (RV)⁶⁹, which hastens inactivation of cGMP. Finally, production of endogenous NOS inhibitors, asymmetric and symmetric dimethylarginines (ADMA and SDMA), is enhanced in pulmonary hypertension (PH)^{83, 84}.

Pharmacological or genetic perturbations of the NO pathway demonstrate the pivotal role of the cGMP pathway in regulating PVR. Mice develop PH if they are rendered deficient in eNOS, GTP cyclohydrase-1 (the rate-limiting enzyme in synthesis of the NOS co-factor, tetrahydrobiopterin,) or dimethylarginine dimethylaminohydrolase (DDAH, the enzyme responsible for eliminating endogenous NOS inhibitors) ^85-87. Inhibition of NO production in humans, using a competitive NOS antagonist (L-NMMA), increases pulmonary vascular resistance (PVR) ^{88, 89}. Chronic pharmacological NOS inhibition causes PH in rats, although there is disproportionate systemic hypertension⁹⁰.

NONOates

Inhalation of exogenous NO gas (iNO, 0.1-100 parts per million) decrease pulmonary arterial pressure (PAP) and improves oxygenation and hemodynamics in children and adults with diverse forms of PH⁹¹. Although chronic iNO therapy for PAH is feasible⁹², delivery is complicated by the instability of NO, which mandates continuous inhalation. In addition, higher concentrations of NO and especially its oxidation products are toxic. Consequently iNO dosing must be carefully monitored to prevent exposure to toxic nitrogen oxides and methemoglobin. Alternative strategies that exploit the specificity of iNO, but utilize more stable NO sources are appealing. One such strategy uses NO/nucleophile adducts, such as diethylenetriamine/NO (DETA/NO). NONOates spontaneously release predictable amounts of NO when exposed to physiological pH. Daily nebulization of DETA/NO (half-time of NO release >20 hours) for ∼1 week reduces PH in monocrotaline-induced PAH without causing systemic hypotension⁹³. DETA/NO has been used effectively to improve pulmonary hemodynamics in intubated patients with adult respiratory distress syndrome⁹⁴. It may be valuable to investigate the many NONOates for long-term, ambulatory use in PAH.

Phosphodiesterase Inhibitors (Figure 5)

Eleven PDE families are known; however, they vary in substrate affinity, selectivity and regulatory mechanisms⁹⁵. In the pulmonary circulation, PDE5 and PDE1 are highly relevant. PDE1 has 3 isoforms that are regulated by calcium-calmodulin and can hydrolyse both cAMP and cGMP. Both PDE1A and PDE1C are upregulated in pulmonary arteries (PAs) from idiopathic (IPAH) patients⁹⁶. Infusion of the PDE1 inhibitor 8-methoxymethyl-isobutyl-1-methylxanthine reduces PVR and RV hypertrophy (RVH) in rodent PH models⁹⁶.

PDE5 expression, normally absent in cardiac myocytes, is upregulated in the right ventricle (RV) in PAH⁶⁹. PDE5 inhibition in PAH models increases RV contractility through a cGMP-mediated inhibition of PDE3⁶⁹. Thus in PAH sildenafil has an effect on the RV similar to the PDE3 inhibitor, milrinone. A single dose of sildenafil (75mg) reduces PVR without lower systemic vascular resistance in PAH patients and simultaneously lowers wedge pressure and increases cardiac output⁹⁷. Sildenafil causes sustained improvement in hemodynamics and functional capacity in PAH and it has been approved as a first-line, oral treatment for PAH (reviewed in ⁹⁸). Combining a PDE inhibitor (PDE3/4 or PDE5), even at subtherapeutic doses, with a prostanoid augments the hemodynamic/functional benefit of the prostanoid^{79, 99, 100}.

Soluble Guanylate Cyclase activators (Figure 5)

sGC is a heterodimer, consisting of α-and β-subunits. sGC expression is upregulated as a compensatory mechanism in human PAH¹⁰¹. Experimental hypoxic-PH is exacerbated in mice lacking the sGCα1¹⁰². Thus, augmentation of sGC activation is an attractive therapeutic strategy. There are NO independent, haem-dependent sGC stimulators (e.g. BAY 41-2272) and NO and haem-independent sGC activators (e.g. BAY 58-2667 and HMR-1766)¹⁰³. BAY 41-2272 stimulates sGC directly but also sensitizes the enzyme to NO, resulting in synergism. It improves pulmonary hemodynamics in models of persistent pulmonary hypertension of the newborn (PPHN)^{104, 105}. NO-independent sGC activators, such as BAY 58-2667, HMR-1766 and S-3448 provide an additive, rather than synergistic, effect when combined with NO donors. Both BAY 41-2272 and BAY 58-2667 reverse established PH in rodent models, although this benefit is partially dependent on endogenous NOS activity¹⁰⁶. BAY 63-2521 (Riociguat), a sGC stimulator, is available orally, has a favorable safety profile and has entered phase III trials in PAH¹⁰⁷.

Enhancing NOS activity – tetrahydrobiopterin and transcription enhancers (Figure 5)

Tetrahydrobiopterin (BH₄) is an important NOS cofactor, essential for dimerization and for the oxygenation of L-arginine to create NO and L-citrulline. Without BH₄, NOS becomes uncoupled and produces superoxide anion which rapidly reacts with NO, producing peroxynitrite, further attenuating NO bioavailability¹⁰⁸.

The rate-determining step for the de novo production of BH₄ is catalyzed by GTP-cyclohydrolase 1 (GTP-CH1). Mice with impaired GTP-CH1 activity exhibit reduced lung BH₄ and spontaneously develop PH with vascular remodelling⁸⁶. Conversely, congenital over-expression of GTP-CH1 in vascular endothelium protects mice from hypoxic-PH⁸⁶. In a porcine PPHN model, combined therapy with BH₄ and a superoxide dismutase mimetic (which enhances survival of endogenous NO) restores endothelial function¹⁰⁹.

While there is no evidence for GTP-CH1 deficiency in PAH, GTP-CH1 polymorphisms are associated with variations in NO bioavailability and systemic hypertension. Some PAH patients show increased markers of oxidative stress¹¹⁰ and this may result in conversion of BH₄ to dihydrobiopterin. BH₄ is a co-factor for several enzymes and is well tolerated when administered in its synthetic form, sapropterin, to patients with phenylketonuria¹¹¹. This observation paves the way for studies in PAH patients.

eNOS transcription enhancers, such as AVE9488 and AVE3085, similarly aim to increase NO signaling. Theoretically, increasing eNOS without corresponding increases in cofactors such as BH₄ could lead to uncoupling and the formation of superoxide ions. However AVE9488-treatment also increases BH₄ levels and improves eNOS coupling in apolipoprotein E-knockout mice¹¹². To our knowledge this agent has not been used in vivo in humans.

Vasoactive peptides and endopepitase inhibitors (Figure 5)

Endothelin receptor antagonists (ERA) (Figure 5)

Endothelin-1 is a vasoconstrictor that acts via two receptors, ET_A and ET_B, to regulate vascular tone and cell proliferation. Both receptor subtypes are found on PASMC and mediate vasoconstriction, while the ET_B receptor on endothelial cells mediates nitric oxide and prostacyclin release, causing vasodilatation. Lung and circulating endothelin-1 levels are increased in PAH patients¹¹³. ERAs, such as bosentan, ambrisentan and sitaxsentan cause a significant, but modest, improvement in pulmonary hemodynamics, exercise capacity (6-minute walk distance) and symptoms and are approved for management of PAH. There are no trial data to indicate whether selective ET_A antagonism offers advantages over combined ET_A and ET_B antagonism (Bosentan) nor is the relative efficacy compared to PDE5 inhibitors known (although a small trial suggests some benefits of sildenafil)¹¹⁴. Liver toxicity and teratogenicity are class effects. While comparisons with historical control data suggest that bosentan monotherapy increases survival¹¹⁵ there are no robust survival data from appropriately designed clinical trials.

The endothelins are produced from big endothelin by endothelin converting enzyme (ECE). ECE inhibitors are an alternative approach to reducing endothelin levels. Although studies with this drug class (e.g. daglutril) have been conducted in patients with systemic hypertension and heart failure, data for pulmonary hypertension are limited.

Natriuretic peptides (Figure 5)

The natriuretic peptides (ANP and BNP) are synthesized in and released from myocardial tissue in response to stretch and their elevation in the blood in PAH indicates the extent of RV dysfunction¹¹⁶. CNP is produced in vascular tissue. These peptides interact with the extracellular domain of the natriuretic peptide receptors, NPR-A and NPR-B, which are transmembrane guanylate cyclases. Upon binding, the intracellular domain hydrolyses GTP to cGMP¹¹⁷. Genetic inactivation of NPR-A is associated with PH while chronic administration of ANP attenuates PAH in animal models¹¹⁸. Short-lived natriuretic peptides are not feasible agents for chronic therapy. An alternative approach is to inhibit metabolism of endogenous natriuretic peptides using neutral endopeptidase (NEP) inhibitors. NEP inhibitors have demonstrated efficacy in animal models and in combination with PDE5 inhibition¹¹⁹, but this combination is untested in patients.

Adrenomedullin (Figure 5)

This vasodilator peptide activates several signaling pathways, such as cAMP, NO-cGMP and PI3K/Akt. It decreases mean PAP and RVH in hypoxic rats, and exhibits anti-proliferative properties¹²⁰. Adrenomedullin-2, a novel peptide, acts by the same receptors as adrenomedullin and its levels are also elevated in the RV of rats with hypoxic-PH. When aerosolized, adrenomedullin-2 reduces monocrotaline-induced PAH in rats and improves survival¹²¹. In humans with PAH, inhaled adrenomedullin causes a modest reduction in PVR and increases peak O₂-consumption during exercise without significant effects on the systemic vasculature¹²².

Vasoactive Intestinal Polypeptide (VIP)

VIP is a 28 amino-acid peptide that increases cardiac output, scavenges oxygen free radical species, inhibits platelet activation and is a potent vasodilator. VIP's effects are mediated by G protein-coupled receptors, VPAC1 and VPAC2. Receptor activation stimulates both adenylate and guanylate cyclase signaling pathways. VIP-knockout mice spontaneously develop PH¹²³. They over-express pro-inflammatory genes and genes involved in pulmonary vascular remodeling, and under-express anti-proliferative genes¹²⁴, including eNOS/NOS3, prostacyclin synthase, GTP cyclohydrolase-1 and BMP-2. Thus VIP is also a key regulator of multiple genes that control the process of vascular remodeling.

VIP receptor expression (particularly the VPAC-2 subtype) and receptor-binding affinity are increased in PASMC from PAH patients; conversely, serum and lung VIP levels are low in PAH. VIP inhibits the proliferation of PASMC from PAH patients¹²⁵. Nebulized VIP (200μg daily) improves pulmonary hemodynamics in PAH patients and, when continued for 3 months, reduces PVR and improves 6-minute walk, with little effect on the systemic circulation. The medical use of peptides in general, and VIP specifically, is complicated by their rapid degradation by endogenous proteases. A sustained release liposomal VIP preparation has extended pharmacological effects and may facilitate the development of VIP as a PAH treatment.

BMPR2-targeted treatment strategies

Loss of BMPR2 function after germ-line mutation has been linked strongly to the development and progression of familial and sporadic forms of iPAH. This has directed attention to strategies targeted at repairing BMPR2 signaling in patients with proven mutations. Gene mutations can directly inactivate BMPR2 (rescue strategies using viral vectors discussed in part 1) or can suppress function by impairing its trafficking to the cell surface. Substitution of cysteine residues in the ligand-binding domain prevents BMPR2 trafficking to the cell membrane and this can be rescued (in a cell model)¹²⁶. In cystic fibrosis, where impaired protein trafficking also occurs, sodium 4-phenylbutyrate can improve membrane trafficking of the chloride channel¹²⁷. Mutant BMPR2 protein that is trapped intracellularly can be ‘rescued’ using chemical chaperones (thapsigargin, glycerol or sodium 4-phenylbutyrate), increasing membrane expression¹²⁶. It remains uncertain how much mutant BMPR2 must reach the cell membrane to induce a clinically relevant effect.

An alternative to restoring BMPR2 function is to inhibit pro-proliferative pathways that are unchecked by BMPR2 dysfunction. PASMC in familial PAH demonstrate increased sensitivity to TGF-β/activin receptor-like kinase 5 (ALK5) signaling, suggesting TGF-β blockade as a therapeutic strategy. The ALK5 inhibitor, SB525334, reverses PAH and RVH in a rodent model, indicating that strategies that inhibit ALK5-signaling may have therapeutic benefit¹²⁸

Inhibitors of Serotonin and the Serotonin Transporter (SERT)-Figure 4

Plasma serotonin is increased in iPAH patients, even after lung transplantation¹²⁹, suggesting that serotonin is either an etiologic factor in iPAH or is associated with such a factor. In addition, PAH endothelial cells do generate more serotonin than controls. SERT expression is increased in PASMC from iPAH patients and these cells proliferate more rapidly in response to serotonin than control cells¹³⁰. In some patients with severe iPAH, the LL SERT polymorphism is associated with greater SERT expression and higher mean PAP than LS or SS genotypes¹³⁰; however, in two separate iPAH cohorts this relationship was not detectedl^{131, 132}. The proliferation of bovine and iPAH PASMC in response to serotonin depends on serotonin internalization via SERT and is blocked by selective serotonin reuptake inhibitors (SSRIs), such as fluoxetine (Prozac®)³². Fluoxetine reduces hypoxic-PH in rats¹³³. In a retrospective cohort study of PAH patients, the use of SSRIs was associated with a trend towards reduced risk of death¹³⁴. The time is right for a randomized clinical trial of fluoxetine versus placebo, on a background of conventional PAH therapy.

Another target is tryptophan hydroxylase, the enzyme that synthesizes serotonin. Deletion of tryptophan hydroxylase 1 reduces pulmonary vascular remodeling and hypoxic-PH¹³⁵. The 5-HT_2A receptor mediates serotonin-induced proliferation in rat PA fibroblasts.⁴¹ Genetic deficiency of the 5-HT_2B serotonin receptor reduces hypoxic-PH in mice¹³⁶. Terguride, a potent antagonist of 5-HT_2B and 5-HT_2A receptors, and partial dopamine agonist, is currently in a Phase II study in PAH patients and has received orphan drug status from the European Medicines Agency. PRX-08066, a selective 5-HT_2B antagonist, is in Phase II (NCT00345774), having demonstrated evidence of efficacy in inhibiting hypoxia-induced rises in PAP in humans. SERT and the 5-HT receptors may act in concert to mediate the proliferative effects of serotonin on PASMC, suggesting simultaneously inhibiting the receptor and transporter as a strategy.

Much has been written about the potential role of serotonin in the etiology of PAH associated with anorexigens, such as dexfenfluramine. Transgenic mice lacking tryptophan hydroxylase are protected from dexfenfluramine-induced PAH¹³⁷. Given that PAH was still uncommon even amongst those who consumed anorexigens¹³⁸, it seems likely that a combination of factors is required to cause disease. For example, serotonin's effects on the pulmonary vasculature are modified by interaction between the serotonin pathway and BMPR2 signaling. Chronic serotonin infusion causes exaggerated PAH and pulmonary vascular remodeling in BMPR2 haploinsufficient mice compared with wild-type mice¹³⁹. There is also a link to mitochondrial metabolism and Kv1.5 channel downregulation as SERT overexpressing mice respond favorably to therapy with the pyruvate dehydrogenase kinase inhibitor, dichloroacetate⁶⁸.

Rho Kinase Inhibitors

In response to calcium/calmodulin, MLC kinase phosphorylates myosin light chain (MLC) causing PASMC contraction; conversely, MLC phosphatase dephosphorylates MLC, causing relaxation. Rho kinase inhibits MLC phosphatase, leading to prolonged, refractory vasoconstriction. Rho kinase participates in the vasoconstriction elicited by many vasoactive agents involved in PAH, such as serotonin, endothelin-1 and thromboxane A₂. Rho-kinase inhibitors (Y-27632, fasudil) also markedly reduces PH in PAH models: the fawn-hooded rat (FHR), the chronic hypoxia/SUGEN 5416 model, and the monocrotaline model, illustrating the critical role of refractory vasoconstriction models¹⁴⁰. In humans with PAH, fasudil, a rho kinase inhibitor, causes modest, acute reductions in PVR¹⁴¹. The challenge with the use of rho kinase inhibitors is avoidance of systemic vasodilatation. Airway nebulization offers a potential means of selectively inhibiting rho kinase in the lung. Rho kinase also participates in the vascular SMC proliferation. There is a rho kinase-dependent mechanism by which serotonin transactivates the PASMC BMPR1A receptor and downstream signaling Smads 1/5/8¹⁴². In SERT over-expressing mice, Rho kinase inhibition reduces PAH and vascular remodeling and this is associated with suppression of ERK phosphorylation in PA fibroblasts¹⁴³.

Restoration of Potassium Channels (Figures 7 and 9)

Downregulation of the expression and activity of voltage-gated K+ channels, notably Kv1.5, is a finding common to human PAH and all rodent PAH models. Kv channels not only regulate the E_M but are also involved in survival signaling, suggesting that K+ channel activation or augmentation therapy could be beneficial in PAH¹⁴⁴.

Potassium channels are tetrameric, membrane-spanning proteins that selectively conduct K+. K+ leaks from PASMC down its intracellular/extracellular concentration gradient (145/5 mM) helping to establish resting membrane potential (E_M) at approximately -60mV. E_M controls vascular tone by regulating the gating of large conductance, voltage-gated calcium channels (the target of nifedipine, a clinically important PAH treatment¹⁷⁷). Depolarization, in response to K+ channel inhibition/downregulation, activates these channels, elevating cytosolic calcium and causing constriction. By regulating intracellular K+ and calcium, K+ channels also regulate cell proliferation and apoptosis and thus vascular remodeling.

PASMC express a diverse array of K+ channels (including voltage-gated (Kv) channels). Several channels are germane to PAH, most notably Kv1.5. Acute inhibition of Kv1.5 by hypoxia initiates hypoxic pulmonary vasoconstriction⁶⁵. Interestingly anorexigens such as dexfenfluramine, which promote PAH, also acutely inhibit PASMC K+ current and block Kv1.5. Expression of Kv1.5 increases longitudinally in the pulmonary circulation and is maximal in resistance arteries, the major site of pathology in PAH. Selective loss of Kv channel expression (and membrane depolarization) is a hallmark of human¹⁴⁵ and experimental^{35, 144, 146, 147} PAH. Restoring Kv1.5 expression reduces hypoxic-PH¹⁴⁴.

K+ channel downregulation increases PASMC proliferation and reduces apoptosis, contributing to obstructive vascular remodeling^{31, 36, 148, 149}. Increased cell proliferation reflects, in part, activation of the Ca²⁺-calcineurin-dependent proliferative transcription factor, Nuclear Factor for Activated T-lymphocytes (NFAT)³⁹. There are several theories for how Kv downregulation impairs apoptosis (essentially by preventing cell shrinkage and by elevating cytosolic K+, which inhibits capases). Kv channel downregulation also occurs in cancer, the prototypic proliferative, anti-apoptotic disease⁶³.

Inhibition of transcription factors

A variety of transcription factors (HIF-1α, NFAT and c-Jun¹⁵⁰) govern the expression of Kv1.5 in PASMC and regulate other factors important to the pathogenesis of PAH. HIF-1α is activated even during normoxia in the PASMC of patients and FHR with PAH¹⁹. HIF-1α activation promotes cell survival and inhibition of HIF-1α may be beneficial. Inhibition of HIF-1α restores Kv1.5 expression and Kv current in experimental PAH¹⁹. The high cytosolic calcium in PAH PASMC results in nuclear translocation (activation) of NFAT. NFAT promotes PASMC proliferation and decreases Kv1.5 expression³⁹. NFAT inhibition, using either cyclosporine or the more specific peptide inhibitor VIVIT, regresses experimental PAH³⁹. NFAT activation also likely contributes to the hyperpolarized mitochondria seen in PAH PASMC. The antiapoptotic protein, bcl-2, which promotes mitochondrial hyperpolarization, is upregulated in iPAH¹⁵¹. Inhibitors of NFAT increase Kv1.5 expression¹⁵² and inhibit bcl-2 expression in monocrotaline-induced PAH³⁹. NFAT inhibition also decreases hypoxic-PH¹⁵³. Moreover, NFATc3 knock-out mice do not show PA remodeling after chronic hypoxia¹⁵⁴. HIF-1α and NFAT inhibition are promising therapeutic strategies.

Inhibition of transient receptor potential (trp) channels (Figure 9)

Upregulation of TRPC6, a nonselective cation channel, occurs in PAH and is another mechanism by which excess amounts of extracellular calcium enter the cells in PAH, independent of L-type calcium channel^{38, 155-158}. Chronic increases in calcium, in part via trp channels, and in part via calcineurin-dependent pathways involving NFAT activation³⁹, drive PASMC proliferation, making trp channel inhibition an interesting therapeutic strategy.

Mitochondria-Metabolic dysfunction in PAH (Figure 7)

PASMC from FHR, and PASMC¹⁹ and endothelial cells¹⁵⁹ from human PAH exhibit dysmorphic and hyperpolarized mitochondria and a glycolytic shift in metabolism. Such a shift to glycolysis, occurring independent of PO₂, was first described in cancer cells (the Warburg phenotype) and is thought to confer resistance to apoptosis. Key molecular contributors to this metabolic phenotype are activation of HIF-1α, which in turn activates transcription of pyruvate dehydrogenase kinase (PDK).

Increased expression of HIF-1α activates a panel of glycolytic genes (such as the glucose transporter, glut 1). HIF-1α simultaneously suppresses the activity of the mitochondrial electron transport chain (ETC) by transactivating the PDK gene, which phosphorylates and inhibits the PDH complex⁶⁴. PDH catalyses the irreversible oxidation of pyruvate, yielding acetylCoA and CO₂. Phosphorylation of any of PDH's 3 regulatory serines by PDK completely inhibits PDH¹⁶⁰.

Dichloroacetate inhibits all four PDK isoforms, thereby activating PDH and promoting glucose oxidation. In PAH PASMC (but not normal PASMC), dichloroacetate depolarizes the mitochondria, increasing hydrogen peroxide production and restoring Kv1.5 expression. The net effect of inhibiting PDK is an induction of apoptosis and decrease in proliferation. Interestingly, there is little effect of dichloroacetate on normal cells because PDK is normally relatively inactive. Dichloroacetate regresses many forms of experimental PAH (chronic hypoxic PHT, monocrotaline PAH and FHR PAH)^{90, 19, 63}. An advantage in translating the use of dichloroacetate from rats to humans is that it has be used safely as a treatment of lactic acidosis in children and has been tested acutely in adults with heart failure. New, isoform selective PDK inhibitors are in development for diseases such as diabetes.

Tyrosine kinase inhibition (Figure 8, Table 1)

Excessive expression/activity of a variety of growth factors contribute to obstructive pulmonary vascular remodeling in PAH, including PDGF, basic fibroblast growth factor, epidermal growth factor, and vascular endothelial growth factor (VEGF). Most growth factor receptors are transmembrane receptor tyrosine kinases and they activate diverse signaling pathways (Figure 8)^{161, 162}. Inhibition of EGF and PDGF receptors have beneficial effects on hemodynamics, remodeling, and survival in experimental PAH^{163, 164}. In humans there are case reports of a beneficial effect of adding imatinib to baseline therapy¹⁶⁵. The mechanism(s) for the potential beneficial effect of imatinib are unclear because it inhibits the tyrosine kinases, PDGF receptors, BCR-ABL, and c-kit.

In addition to receptor tyrosine kinases, serine/threonine kinases, such as the Raf family and its downstream pathways, offer targets for intervention in PAH^{166, 167}. Sorafenib is a “multikinase inhibitor”, blocking the serine/threonine kinases Raf-1 and b-Raf, tyrosine kinases, PDGF and VEGF receptors, c-kit, and Flt-3, with IC₅₀ values between 6 and 70 nmol/L. Sorafenib is approved for the treatment of renal and hepatocellular carcinoma. Sorafenib prevents and reverses PAH and cardiac remodeling in monocrotaline-treated rats and may have more pronounced effects on right ventricular function than imatinib¹⁶⁸. Phase 1 clinical trials with both imatinib and sorafenib (NCT00452218) have been completed but not yet published.

Elastase and Matrix Metalloproteinases (MMPs)

Increased elastolytic activity may be an early feature of PH and serum elastase levels are elevated in experimental PAH¹⁶⁹. Endogenous elastases may contribute to the development of PAH by liberating mitogens (e.g. tenascin c⁴⁰), and growth factors from the matrix and activating growth factor receptors in a ligand-independent manner (Figure 10). Elastase inhibitors can attenuate or reverse experimental PAH¹⁷⁰ but synthetic inhibitors with acceptable toxicity in humans have yet to be developed. Augmenting expression of endogenous elastin inhibitors, such as elafin¹⁷¹, may prove a better strategy.

Peroxisome proliferator-activated receptor activation (PPARs, Table 1)

PPARs are ligand-activated transcription factors, belonging to the nuclear receptor superfamily. Upon ligand activation, PPARs heterodimerize with the retinoid X receptor and bind to PPAR response elements in regulatory promoter regions of their target genes. A series of recent observations suggests that PPARγ could be a drug target in PAH^{172, 173}. PPARγ is a downstream target of BMP2 in human PASMC¹⁷³. PPARγ is important for BMP2-mediated inhibition of PDGF-induced vascular SMC proliferation¹⁷². Mice lacking SMC PPARγ develop PAH¹⁷². PPARγ activation stimulates apolipoprotein E expression. Recombinant apolipoprotein E inhibits PDGFR-β–mediated SMC proliferation and migration¹⁷⁴ PPARγ targets, independent of apolipoprotein E, may also important in suppressing pulmonary vascular remodeling as male apolipoprotein E (-/-) mice fed a high fat diet develop PAH which is reversed by rosiglitazone (a PPARγ agonist)¹⁷³. PPARγ agonists have direct anti-inflammatory and pro-apoptotic effects. PPARs can also interact with signaling molecules to regulate gene expression, independent of DNA binding. PPARγ can impair the phosphorylation of extracellular signal-regulated protein kinase (ERK)¹⁷⁵ which is implicated in PASMC proliferation and migration. iPAH patients have reduced lung expression of PPARγ and apolipoprotein E mRNA. Because the thiazolidinedione rosiglitazone is widely used in the treatment of type II diabetes, a trial in PAH would be feasible. Despite this promise, rosiglitazone failed to ameliorate pulmonary hypertension in hypoxic-PH rats, although it did reduce RVH and pulmonary vascular remodeling¹⁷⁶.

Inflammation (Figure 11)

Aside from the association of PAH with several collagen vascular autoimmune disorders (e.g. scleroderma, systemic lupus erythematosus and mixed connective disease) and schistosomiasis, several observations argue for a role of inflammation in the pathogenesis of PAH. These include: the presence of T cells, B cells and macrophages in plexiform lesions; the detection of autoantibodies to endothelial cells and fibroblasts; raised blood cytokine and chemokine levels; and the association of PAH with certain infections such as human herpes virus 8. Mice that overexpress S100A4/Mts1 develop extensive and severe neointimal lesions following injection of the γ murine herpes virus-68 (the murine homolog of HHV-8)¹⁷⁷. PAH also develops in a subset of patients with HIV disease. The HIVnef gene was also recently implicated in plexogenic pulmonary vascular lesions associated with PAH in HIV-infected patients and SIV-infected nonhuman primates¹⁷⁸.

Athymic nude rats, lacking T cells, appear more sensitive than normal rats to the development of PAH when challenged with the VEGF-receptor antagonist SU-5416¹⁷⁹. A protective role for T cells was established by the administration of splenocytes from euthymic rats. In iPAH, regulatory T_reg cells are increased whilst CD8⁺ cytotoxic T cells are decreased¹⁸⁰. T_reg cells maintain immunotolerance and they are potent inhibitors of antitumour and possibly antiviral immune response. The increase in T_reg cells may be a normal counter-regulation or compensation for an initial inflammatory response.

Can the immune system be therapeutically targeted in PAH? Mycophenolate mofetil, a potent immunosuppressant used in humans, prevents monocrotaline-induced PAH in rats¹⁸¹; however, regression trials (a more clinically relevant standard for experimental PAH therapies) are needed.

Endothelial Progenitor Cells (EPCs)

EPCs arise from mesodermal stem cells or haemangioblasts in the bone marrow. Circulating in plasma, they home to sites of ischemia or endothelial injury and differentiate into mature endothelial cells in situ, contributing to revascularization and vascular homeostasis. EPCs can be considered a potential therapeutic target, a predictive biomarker¹⁸² or as a vector for cell-based therapy. Circulating EPC numbers (defined by CD34⁺/KDR⁺ and CD34⁺/CD133⁺/KDR⁺ positive cells) are significantly lower in patients with Eisenmenger's syndrome compared with normal controls¹⁸³. Some, but not all, investigators have reported reduced EPCs in iPAH patients. Differences in the markers used to identify and quantify EPCs complicates interpretation of the data.

The in vitro functions of endothelial-like, mononuclear cells (e.g. colony forming capacity, adherence, migration and sensitivity to apoptosis) isolated from the blood of iPAH patients differ from those of healthy controls^183-186. Whether these difference are beneficial, promoting revascularization in the hypertensive lung¹⁸⁷, or contribute to the pathology, by augmenting pulmonary vascular remodelling¹⁸⁵, is unclear. This distinction is important given that some treatments (e.g. sildenafil) are associated with a dose-dependent increase in the abundance of circulating EPCs¹⁸³ and potential new therapies for PAH, such as statins and PPAR-γ agonists, also induce the mobilization and differentiation of EPCs.

Administration of EPCs has produced improvements in pulmonary hemodynamics, vascular remodeling and survival in monocrotaline-induced PAH¹⁸⁸. Cell therapy has been less effective in hypoxia-induced PAH and may contribute to the pathologic vascular remodeling (Figure 12). The benefits of cell therapy may be enhanced by the expression of genes that inhibit smooth muscle cell proliferation and/or stimulate angiogenesis (e.g. eNOS)^{187, 189, 190}. Even fibroblasts can be made somewhat therapeutic when they are transfected with VEGF. These modified fibroblasts prevent worsening of monocrotaline-induced PAH¹⁹¹.

Two small pilot studies, in which adults and children with IPAH were given a single intravenous infusion of autologous mononuclear cells provide support for the therapeutic potential of cell-based therapy in patients^{192, 193}. A therapeutic trial (PHACeT, NCT00469027) to assesses the safety of administrating autologous, cultured, eNOS-transduced mononuclear cells in iPAH patients has commenced.

There remains much to be done in the field of cell-based therapies for PAH, particularly as it remains uncertain whether influx of progenitor cells into the lung in PAH is beneficial or harmful. In Figure 12 lessons learned from remodeling in hypoxia are reviewed. In hypoxia, inflammatory and progenitor cells appear to contribute to the pathological remodeling; however, it is not certain whether this applies to PAH.

Miscellaneous pathways with therapeutic implications

Statins, heparins, dihydroepiandrosterone (DHEA) and inhibitors of Angiopoietin 1, STAT3, Polyamines, Survivin, and the cell cycle offer potential therapeutics for PAH and are discussed, due to page limits, in the on-line supplement.

Conclusions

In this review of the basic science of PAH we assessed emerging concepts of the molecular mechanisms of PAH and identified the novel therapeutic targets suggested by this science. New therapeutic strategies include: enhancing endothelial function/vasodilatation using guanylate cyclase activators or vasodilator peptides, such as adrenomedullin and vasointestinal peptide, augmenting the BMPR2/SMAD pathway, inhibiting serotonin and the serotonin transporter, modulating expression/activity of ion channels (Kv1.5 and TRPC6), inhibiting transcription factors (NFAT and HIF-1∝), increasing apoptosis (survivin inhibitors), inhibiting tyrosine kinases, inhibiting the contractile apparatus (rho kinase inhibitors), preserving elastin, and modulating influx of inflammatory and progenitor cells. Opportunity also exists to accelerate drug development using molecules approved for the management cancer, vascular dysfunction and metabolic disorders. These conditions share PAH's pathophysiologic abnormalities (endothelial dysfunction, excessive cell proliferation, disordered apoptosis, and inflammation). Repurposed drugs that have potential in PAH include: phosphodiesterase-5 inhibitors (for erectile dysfunction), Imatinib (for chronic myelogenous leukemia), Sorafenib (for renal carcinoma) and dichloroacetate (for mitochondrial diseases). We do not endorse the off-label application of these agents in clinical practice; however there is a compelling need to study these potentially curative agents in preclinical and, when appropriate, clinical trials. This is an exciting time in the search for a cure for PAH and it is time for Physicians, armed with a Basic Science Playbook, to take the field.