Relevant Issues in the Pathology and Pathobiology of Pulmonary Hypertension

Does the pulmonary venous system play an important role in PAH and to what extent is PAH and PVOD part of the same spectrum of disease?

It is apparent that the different forms of PH present with either a predominance of pulmonary arterial remodeling or vein remodeling or a variable contribution of both. Paradigmatic of the former is idiopathic pulmonary arterial hypertension (IPAH), while pure pulmonary venoocclusive disease and PH due to left heart dysfunction are characterized predominantly by venous remodeling. Virtually all forms of PH, including those caused by interstitial lung disease, thromboembolic, hypoxia, and sarcoid may involve elements of both arterial and venous remodeling. However, more precise documentation, including morphometric analysis of pulmonary veins remodeling is still lacking in most of these conditions.

The difficulty in studying the pathology of pulmonary venous remodeling is compounded by the lack of distinct molecular markers that allow for their identification (as compared with the pulmonary arterial bed) when immersed in alveolar tissue. Normal veins can be recognized by the lack of a double elastic lamina and relatively thin muscular media; however, when remodeled, pulmonary veins become “arterialized”, making their distinction from pulmonary arteries more difficult. The only distinctive feature, which is often not present in biopsies, is the presence of pulmonary veins in interlobular septae. Moreover, molecular markers of veins, such as the ephrin B4 receptor, are often difficult to identify in human (and rat) lungs. The venous coaxial structure includes layers of cardiomyocytes arrayed externally around a subendothelial layer of typical smooth muscle cells, thus forming sphincter-like structures. In disease states this cardiac muscle layer extends further into the lung and its role in Group 2 PH merits investigation.

Given the limitations in recognizing veins histologically, most studies rely on their identification in selected regions of explant specimens. In subsets of PAH patients with scleroderma-associated PAH, pulmonary veno occlusive disease-like remodeling may predominate over arterial remodeling. The venous remodeling, if present, entails worse prognosis associated with decreased 6 minute-walk distance, PaO₂, and diffusion capacity (14). However, in a recent pathologic study of PAH, including lungs with scleroderma associated PAH, no correlations between vein remodeling and intima and media thickness in arteries could be identified (15). Recent studies have highlighted the pattern of pulmonary vein remodeling in patients with left heart failure (LHF). Lung samples obtained from patients with LHF placed on a ventricular assist device showed increased pulmonary vein and artery media thickening. Of interest, some patients who improved on the assist device and underwent repeat lung biopsy had a significant improvement of pulmonary vein remodeling (16). The molecular drivers of these processes remain largely unknown.

In summary, we continue to lack the availability of unique molecular markers that allow the specific identification of pulmonary veins. Consequently, the prevalence of venous pathology in the various forms of PH remains unknown. This is despite the potential for deleterious effects of prostacyclin and agents that increase bioavailability of nitric oxide in the context of elevated capillary pressures. Finally, little is known about pathogenetic pathways involved in pulmonary vein remodeling and how they correlate with arterial remodeling in both LHF and in forms of PH with precapillary predominance.

Are there distinct pathways in vascular cells in mild versus severe PH?

Although PH is defined as a mean pulmonary arterial pressure at right heart catheterization of higher than 25 mmHg with a normal pulmonary capillary wedge pressure, it is well known that, on the basis of clinical and hemodynamic findings, PH can range from mild to severe, even in the presence of the same stimulus (e.g. hypoxia). This spectrum of severity raises several critical pathogenetic and clinical questions in terms of the distinctive feature that differentiate the severity stages.

Increased pulmonary artery pressures occur due to sustained vasoconstriction, excessive pulmonary vascular remodeling, and in situ thrombosis, which ultimately lead to increased pulmonary vascular resistance. However, the factors responsible for the aggravation or acceleration of PH remain poorly defined (Figure 1). The contributing factors likely involve accumulation of multiple events on a background of genetic predisposition. These factors involve the action of vasoconstrictive and pro remodeling processes, including the action of inflammatory, pro-coagulant, anti-apoptotic, and autoimmune mediators, cell-cell and cell-matrix interactions and environmental factors over time (Figure 2. Although the pulmonary artery bed appears unreactive to vasodilators in advanced disease, vasoreactivity and remodeling possibly interact in disease evolution (17).

Whether the severity of pulmonary vascular disease involves a constellation of pathobiologic processes or is defined pathologically, with the hallmark finding of reduction of the pulmonary vascular lumen remains unclear. The definition of the severity of PH based on histopathological findings is complicated by the lack information on what constitutes “normal”. Surprisingly, recent analysis of unused donor “control” lungs revealed substantial neointimal formation, inflammation and venous changes (15), features usually judged to be “pathological”. This suggests a spectrum from pristine vessels (which are observed primarily in the younger controls) to vascular changes reminiscent of PH that may be present as a function of normal aging, including inflammation and left ventricular stiffening. Since a description of the pathology in mild forms of PH is largely unavailable, it is difficult to discern whether the pathological features we observe in ‘controls’ are similar but still less severe than in patients with ‘mild PH. Perhaps a better definition of severity would also incorporate the extent of the reduction in cross-sectional area of the pulmonary vascular bed. In this summary, we confine our discussion of severity to the extent of the pulmonary vascular remodeling process, though in clinical practice, the assessment of severity will also consider the function of the right ventricle. Below, we expand on how genetic factors influence with cellular and molecular pathogenetic processes to possibly account for the severity of pulmonary vascular disease (Figure 1).

Mutations in BMPR2) or ALK-1 receptor are emerging as determinants of severity of PAH. Mutations in BMPR2 (18) have been reported in more than 70% of subjects with one or more affected relatives (heritable PAH) and 11-40% of idiopathic PAH (19, 20). Mutations in several other genes have been found including mutations in the ALK1 gene (21), the endoglin gene (22), the SMAD9 gene (23), the Caveolin-1 gene (24), and recently, the KCNK3 gene (25). Patients with BMPR2 or ALK-1 mutations present with higher pulmonary vascular resistance (26). There is also evidence that these patients present and die at a younger age with more severe disease compared PAH patients without mutations, and are less likely to respond acutely to vasodilators (27). Recent evidence suggests that the degree of pulmonary vascular remodeling is greater in patients with BMPR2 mutations compared with non-BMPR2 related disease at the time of transplantation (4). In addition to these causal rare sequence variants, variable expressivity of PAH might be explained by genetic modifiers. For example, single nucleotide polymorphisms (SNPs) in several genes, such as the angiotensin-converting enzyme gene, the KCNA5 gene, the serotonin transporter gene, the serotonin 5-HT2B receptor gene have been reported to influence disease penetrance or progression. In addition, microsatellite instability of endothelial cell growth and apoptosis genes within plexiform lesions in idiopathic PAH (28) as well as somatic chromosome abnormalities (29) have been described. Most cases of PH do not have known genetic triggers; moreover vascular cells isolated from experimental PH maintain their in vivo characteristics when placed in culture, suggesting epigenetic abnormalities. An example of such a contributing factor is the silencing of mitochondrial superoxide dismutase 2 (SOD2), which alters redox signaling and therefore activates hypoxia inducible factor (HIF-1) α, leading to the aerobic glycolysis seen in PH (30, 31). However, further studies, particularly focused on epigenetic control of additional key genes, are required to verify their contribution to the pathogenesis of PH.

Apart from major effects of rare sequence variants in heritable forms of PAH, inter individual differences in response to the same stimulus are well documented in subjects exposed to environmental hypoxia at high altitude, or in the context of high pulmonary blood flow or pulmonary venous hypertension. This variable expressivity most certainly involves the impact of unknown genetic influences regulating the pulmonary vascular response; this variation in response to hypoxia is also seen in different strains of rat (32) and bovines exposed to high altitude (33). Elegant studies demonstrated that this response is inherited (34), though the identification of the genetic basis is the subject of ongoing research (35).

The variable expressivity of PH might also be due to exposure to different environmental influences or comorbidities, including: inflammation, auto-immunity, viral infection, and hormonal mediators. The specific topic of inflammation is addressed below. Furthermore, accumulating evidence indicates that metabolic dysfunction may also contribute to variability in susceptibility and expressivity of PH (Figure 2).

Although in general, animal models of PH do not recapitulate the severe pathology of human disease, these experimental models support the idea that PH can be triggered by different stimuli and that the structural changes observed in the pulmonary vasculature vary depending on the nature of the injurious stimulus (36). Several studies have underlined the importance of different signaling pathways in PH, including elastase inhibitors (37), antagonists of EGF (4), or dietary copper restriction (38). The lack of animal model fidelity to the human pathology needs to be critically analyzed. The obliterative changes in the most distal precapillary arteries in the SU5416/hypoxia rat model (36, 39), which evolves over time (39), do not fully recapitulate the large plexogenic lesions seen in patients – its potential reversibility may also indicate the model may not be fully reflective of the largely irreversible nature of the human disease. Indeed, recent studies indicate that the long telomeres present in mice may limit the extent to which animal model can phenocopy human disease.

In summary, several genetic, molecular, biochemical and environmental factors may explain the variable expressivity of PH and may contribute to its aggravation or acceleration. Severe PAH seems to represent the far end of a spectrum in which there is an augmented burden of vascular injury and dysregulated repair (Figure 1). Further studies, which likely require a “systems biology-like” approach, are needed to establish which key signaling pathways in pulmonary vascular cells which lead to severe disease. Factors that contribute to this accelerated pathology include genetic abnormalities that may affect: (1) The exuberance of the cellular response; (2) The chronicity of an inflammatory stimulus; (3) The severity of the altered metabolic state and cumulative DNA damage.

What are differences and similarities between cell proliferation in PAH when compared with traditional neoplastic disease?

The concept of a neoplastic-like pathobiology of PH (in particular in PAH) (40) has its origins in the finding that endothelial cell in IPAH plexiform lesions are clonal when compared with similar lesions in lungs of patients with congenital heart disease (41), coupled with somatic instability in PAH lesions (28) and cultured lung endothelial cells (29). Similar to the concept of an uncontrolled cell growth in PAH, early data emphasized the upregulation of HIF-1α in plexiform lesions and pulmonary arteries in PAH (42) and survivin in IPAH (43). These findings, largely centered on the angle of disorganized cell growth, favored the emergence of an apoptosis-resistant phenotype, a selection process that characterizes neoplastic processes (44). In the past 8 years, further experimental evidence emerged of metabolic plasticity and altered cell energetics in the pathogenesis of PAH (45, 46). Metabolic reprogramming (Figure 2) constitutes, with genomic instability and mutations, immune escape, and cell-growth promoting inflammation, the so called “emerging hallmarks” pertinent to the pathogenesis of cancer (47). Recent studies have delineated that cancer cells use these hallmarks as driving forces leading to cell growth and co-option of the surrounding stroma for their aggressiveness (48). Supporting documentation of metabolic adaption of pulmonary vascular cells derived from studies in rat smooth muscle cells from the fawn-hooded model of PH (45) and cultured human IPAH cells (30). In both experimental settings, the hypertensive vascular cells would preferentially use aerobic glycolysis (instead of mitochondrial metabolism). The preferential use of aerobic glycolysis by proliferating cells (including cancer cells) would endow a selective growth advantage to pulmonary vascular cells: glucose and glycolytic intermediates provide key substrates for the pentose phosphate shunt to generate reducing equivalents in the form of reduced nicotinamide adenine dinucleotide phosphate (NADPH) and nucleotides, essential for cell growth (49).

The preferential utilization of aerobic glycolysis appears to be related to upregulation of several glycolytic genes and increased expression of the subunit 4.2 of cytochrome oxidase in the mitochondrial respiratory chain, all events driven by expression of HIF-1α (50). In human PAH lungs, HIF-1α appears to be expressed by endothelial cells (42) and smooth muscle cells (45). The source of stabilization of HIF-1α remains unclear, but it appears to be related to altered cellular oxidant/antioxidant balance (45, 51) present in hypertensive cells. Moreover, HIF-1α could also account for the increased mobilization of hematopoietic precursors in PAH (52-54), which might promote further injury to the pulmonary endothelium (13) and possibly to contribute to the remodeled cell population (Figure 2).

In addition to playing roles in the metabolic adaption and mobilization of bone marrow precursors, HIF-1α also participates in the regulation of mitochondria dynamics. HIF-1α accounted for decreased numbers of mitochondria and decreased nitric oxide availability in IPAH cells when compared with control cells (42); of interest, these alterations also correlate with increased IPAH endothelial cell proliferation when compared with normal endothelial cells in culture (55). HIF-1α-dependent mitochondria plasticity also pertains to PH smooth muscle cells. Activation of HIF-1α activity with cobalt or desferrioxamine, leads to mitochondria fission (mediated by activation of dynamin related protein 1) in human PAH and rodent PH smooth muscle cells, driven by activation of dynamin related-protein (DRP)-1 by active cyclin B1 (56). Finally, consistent with the role of mitochondria fission in PH smooth muscle cell growth, inhibition of DRP-1 with the peptide Mdiv-1 blocks cell proliferation in cultured cells as well as PH caused by in vivo administration of cobalt and chronic hypoxia (56). Conversely, overexpression of mitofusin-1, which leads to mitochondria fusion, counteracts the mitochondria phenotype in pulmonary hypertensive smooth muscle cells (57).

The paradigm of metabolic plasticity and adaption of PH cells has wider implications, interfacing with controls of cellular stresses and protein misfolding, and protein sorting in cytoplasmic compartments (58). In line with these paradigms, there is evidence of an altered endoplasmic reticulum (ER) stress response in PH (59), which may ultimately contribute to both adverse remodeling and elevation of pulmonary artery pressures (60). Moreover, ER stress can lead to activation of the stress-related kinase activating transcription factor (ATF)-6, which increases the expression of Nogo-B. This contributes to an abnormal distancing and signaling coupling (via calcium fluxes) between the ER and the mitochondria (61). These events would lead to an increase in glycolysis, favoring vascular cell proliferation and ultimately PH.

The inroads made by the metabolic related studies in PH are starting to bear translational fruits. There is increased metabolic labeling of glucose uptake with 18-Fluorodeoxyglucose of IPAH lungs and right ventricle when compared with controls (30, 62), which have been also observed in animal models of PH (63, 64). Pharmacologic targeting of the preferential use of glycolysis by PH cells is being tested in humans with the small molecule dichloroacetate (clinicaltrials.gov NCT01083524), which improves PH in animals by blocking pyruvate dehydrogenase kinase, a negative regulator of pyruvate import into the mitochondria (65).

We are at the beginning of deciphering the role of cellular metabolism in the pathogenesis of PH and how this fundamental process relates to inflammation, remodeling, therapeutic response, etc. Genetic modifications that favor glycolysis in multiplying cells would certainly contribute to part of the process of remodeling. However, many of the vascular cells in established human PH are quiescent, long-lived, and probably resistant to apoptosis. These quiescent cells would have substantially different metabolic requirements from those of proliferating cells. Clarifying this metabolic heterogeneity and how it contributes to remodeling and vasoconstriction will be critical moving forward. The interface of cell metabolism with the genetic basis of PAH and inflammation will certainly constitute key areas of future investigation.

What is the role of inflammation in the initiation and progression of different PAH types?

Inflammation has been long recognized as an important pathogenetic element in PH (66). Expanding on prior observations of accumulation of inflammatory cells in PAH, it was documented for the first time that the amount of perivascular inflammatory infiltrate, largely of lymphocytes, correlated with parameters of pulmonary vascular remodeling and hemodynamics in PAH (15). Inflammation consists of a complex series of interactions among soluble factors and immunologically specialized cells triggered in response to traumatic, infectious, post-ischemic, toxic, or autoimmune injury (67). It is increasingly clear that early and persistent inflammation is present and contributes to pulmonary vascular disease, as documented in tissue-based studies and studies of circulating inflammatory cells and chemical mediators.

The presence of elevated levels of circulating inflammatory cytokines, such as interleukin-1α in PAH is well recognized (68). More recently, a wide range of cytokines has been shown to be elevated and to correlate with survival in PAH (69). At the tissue level, traditional cellular components of inflammation are described in the hypertensive pulmonary circulation (70, 71). More specifically, cellular inflammation involves increases in the number of perivascular macrophages (CD68+), macrophages/monocytes (CD14+), mast cells, dendritic cells (CD209+), T cells (CD3+), cytotoxic T cells (CD8+), and helper T cells (CD4+) in the walls of PAH vessels compared with control subjects (72). FoxP3(+) cells have been shown to be significantly decreased, suggesting a reduction in the local number of T regulatory cells (72). Although increased numbers of circulating Treg cells has been reported in patients with idiopathic PAH (73, 74), the reduced numbers of Treg in lung tissue may reflect decreased tissue recruitment of these cells. Given their immune soothing role, a diminished negative regulation on other active immune / inflammatory cells may trigger or amplify pulmonary vascular remodeling and PH. The innate inflammatory system appears to also participate in PAH. Natural killer (NK) cells, which target stressed, virally infected, or oncogenically transformed cells independent of antigen recognition, are dysfunctional, with reduced number and cytolytic capacity, in patients with idiopathic PAH and in mouse and rat models of PH (75). Neutrophils may have potential unrecognized roles as sources of powerful proteases, including elastases, which have recognised roles in pulmonary vascular remodelling (37, 76). Since hypoxia exerts profound effects on neutrophil function (e.g. resistance to apoptosis (77)) further evaluation of the role of the neutrophil in PAH is overdue. The complement system, which bridges innate and adaptive immunity, is also activated in PAH and deficiency of complement C3 protects mice from hypoxia induced PH (78). Mechanistic studies focusing on the innate immune system are needed for a better appreciation of its role in PH.

However, recent insights highlighted a potential role for acquired immunity, suggesting that the pathogenesis of PH may involve specific and targeted immune cellular responses. A recent study identified a large number of tertiary lymphoid follicles in lungs from patients with idiopathic PAH when compared with control lungs (79). These lymphoid follicles in idiopathic PAH patients displayed the canonical cellularity and structure of bona fide tertiary follicles, while lymphoid organogenic chemokines such as CXCL13 and CCL19/CCL21 were overexpressed. These data provide a structural basis for a local autoimmune response in this disease (80-82). Finally, inflammation is closely associated with pulmonary vascular disease in the setting of auto-immune diseases, such as scleroderma, and paradigmatic of the interplay of inflammation and PH, is its link with the infection with the parasite Schistosoma mansoni (the most frequent cause of PAH worldwide (83)) and HIV infection (84).

Vascular inflammation in general and PH in particular, has traditionally been considered an “inside-out” response centered on leukocyte/monocyte recruitment to the intima of blood vessels driven by the endothelium-expressed adhesion molecules. Growing experimental evidence, however, supports an alternative paradigm of an “outside-in” hypothesis, in which vascular inflammation is initiated in the adventitia and progresses inward toward the media and intima, coupled with the activation of the remodeling process. In support of the “outside-in” hypothesis of adventitial regulation of inflammation are observations that, in a wide variety of vascular injuries, including PH, there is a rapid influx of leukocytes into the adventitial compartment (85). The immediate perivascular environment in PH becomes populated by inflammatory/progenitor cells, in a niche dominated by adventitia fibroblasts (36, 70, 85-89). These adventitial fibroblasts and recruited monocytes, particularly under hypoxia, express in a time-dependent and pulmonary artery-specific manner several cytokines/chemokines, their receptors, and adhesion molecules; these cells therefore appear to initiate and perpetuate the inflammatory response in an “outside-in” fashion, possibly mediated by NF-kB signaling (90, 91). Therefore, temporal-spatial dysregulation and/or failure in the normal “switch-off signal” in fibroblasts and/or macrophages/DCs may directly contribute to the persistence of a chronic inflammatory immune response. These adventitia processes are more evident in non-mouse models and in lung of patients with PH possibly due to the presence of bronchial circulation.

The action of secreted cytokines, including TGF-β, in the adventitia mediates specific homing for leukocytes in this vascular compartment, leading to their inappropriate/pathologic retention and survival (91). These homing/recruitment cytokines include upregulated adhesion molecules, including intercellular adhesion molecule 1 (ICAM-1) and vascular adhesion molecule-1 (VCAM-1), and stromal derived factor (SDF)-1/CXCR4, which promotes adhesion of leukocytes and bone marrow precursor cells, respectively. The adventitia is therefore suited to harbor canonical innate immune cells, specifically macrophages and dendritic cells (DCs), which with adventitial fibroblasts, are all equipped with the necessary machinery, (e.g. toll-like receptors (TLRs), inflammasome components (like the nod-like receptors (NLRs)), to potently respond to a variety of exogenous and endogenous danger signals.

Importantly, there is growing evidence that epigenetic marks may “lock” innate immune cells into a distinct functional phenotype, with loss of functional plasticity and failure to respond to regulatory signals. The inflammatory microenvironment in the pulmonary circulation promotes epigenetic marks in macrophages, locking their functional phenotype into a pro-fibrogenic and pro-remodeling macrophage (92-96). Different signaling processes, mediated largely by signal transducers and activators of transcription (STAT)1, 3, and 6 may drive macrophage-promoted remodeling (97); it is likely that all these macrophages are present in PH and play complex and time dependent, sometimes overlapping or dedicated roles in the disease. Moreover, it is also likely that individual macrophages are plastic, changing their function and signaling according to the disease stage and specific molecular triggers.

Finally, emerging concepts involve the potential of organ-specific microbiota and their metabolic products to engage the inflammasome pathways and to influence pulmonary vascular responses, as shown in the liver (98). A further pathway by which remote signalling might target the pulmonary vasculature is by the release of cell-derived exosomes from distant sources, exerting anti- or pro-inflammatory depending on their cargo. Exosomes derived from mesenchymal stem cells were found to inhibit the hypoxic activation of the STAT3 pathway in hypoxic mice and prevent the development of PH (99).

In conclusion cytokines and immune cells appear to play significant but complex roles in the initiation and progression of PH. Major unanswered questions include: (1) What determines the abnormal host response to inflammation that leads to the initiation and progression of PH? (2) Is the inflammatory response in PAH caused by autoimmunity or infection? (3). Which specific features of the inflammatory response can be enhanced (if protective) or blocked (if detrimental) for therapeutic intervention?