The hallmarks of severe pulmonary arterial hypertension: the cancer hypothesis—ten years later

Abstract

Severe forms of pulmonary arterial hypertension (PAH) are most frequently the consequence of a lumen-obliterating angiopathy. One pathobiological model is that the initial pulmonary vascular endothelial cell injury and apoptosis is followed by the evolution of phenotypically altered, apoptosis-resistant, proliferating cells and an inflammatory vascular immune response. Although there may be a vasoconstrictive disease component, the increased pulmonary vascular shear stress in established PAH is caused largely by the vascular wall pathology. In this review, we revisit the “quasi-malignancy concept” of severe PAH and examine to what extent the hallmarks of PAH can be compared with the hallmarks of cancer. The cancer model of severe PAH, based on the growth of abnormal vascular and bone marrow-derived cells, may enable the emergence of novel cell-based PAH treatment strategies.

INTRODUCTION

One day, we imagine that cancer biology and treatment—at present a patchwork quilt of cell biology, genetics, histopathology, biochemistry, immunology and pharmacology—will become a science with the conceptual structure and logical coherence that rivals that of chemistry and physics.

—D. Hanahan and R. A. Weinberg, 2000

The formal statement of the “quasi-malignancy paradigm of severe pulmonary hypertension” was published in 2008 (138); it had been encouraged by the seminal article, published in 2000 by Hanahan and Weinberg (65) in Cell (“The hallmarks of cancer”).

For the purpose of the following discussion, we consider as forms of “severe pulmonary arterial hypertension” those diseases that are characterized by an obliterative angiopathy that frequently leads despite therapy to right heart failure and death.

The 2008 publication (138) intended to provide, borrowing heavily from Hanahan and Weinberg (65), not a unifying hypothesis but an overarching pathobiological concept of severe pulmonary arterial hypertension (PAH). The initial question was: “Is there enough evidence and data cohesiveness to adopt the idea that PAH and cancers share important pathogenic mechanisms?”

Shared between cancer and PAH are three of the hallmarks published in 2000: apoptosis-resistant cell growth, angiogenesis, and evasion of growth suppressors; whether replicative immortality and sustained proliferative signaling are shared is uncertain, as in-depth investigations of cells harvested and cultured from human PAH lung tissues are presently lacking. The hallmarks “invasion” and “metastasis” are not shared (65). Shared are the hallmarks, added in 2011: genome instability and mutations, inflammation, deregulation of cellular energetics, and avoidance of immune destruction (66). Figure 1 illustrates the similarities in the pattern of expressed genes and proteins as recognized in 2008 and the knowledge gained by 2019. “Quasi-malignancy” simply refers to the observation that there are several hallmarks shared between cancer and severe PAH (31, 174, 182). The reader is also referred to a recent publication that reflects a Pro-Con debate that took place in 2018 during the annual meeting of the American Thoracic Society (159).

Fig. 1.

Venn diagrams showing the state of knowledge in 2008 (A) and in 2019 (B). The overlapping areas represent proteins, which are expressed both in cancers and in pulmonary arterial hypertension (PAH) lung tissue. ALCAM-1, activated leukocyte cell adhesion molecule; EMT, epithelial mesenchymal transition; EnMT, endothelial mesenchymal transition; HIF-1α, hypoxia-inducible factor-1α; KDR, VEGF receptor 2; 5-LO, 5-lipoxygenase; MDRP, multiple drug resistance protein; PGI₂, prostacyclin; PPARγ, peroxisome proliferator-activated receptor-γ; TAM, tumor-associated macrophages. The overlap area in B represents protein expression data and mechanistic aspects of the “quasi-malignancy” concept like “glycolysis” and “DNA damage” that have been added to the literature since 2008. For more details, see Table 1.

Is it reasonable to compare solid tumors caused by somatic mutations due to repetitive carcinogenic stimulation (inflammation, cigarette smoke, toxins) to a nonmetastatic, primarily vascular condition? Even when admitting that cigarette smoking can cause lung diseases associated with pulmonary hypertension, cigarette smoking has been related only rarely to the development of a plexogenic pulmonary arteriopathy (24). The comparison appears to be justified, as the theoretical framework proposed here to explain the development of the angioobliterative microvasculopathy is entirely compatible with the concept of cancer as defined by Harold Dvorak: “Cancer is the wound that never heals.” (42). It is likely that the concept of “wound healing gone awry” does also apply to PAH (182). Chronic endothelial injury (due to autoantibodies, infection, or high shear stress) and increased endothelial vulnerability to apoptosis (due to BMPR2 and other gene mutations; see Ref. 44) are central in the development of PAH.

Here, we critically review to what extent recent discoveries in the pulmonary vascular research field continue to support the concept of the cancer paradigm of PAH. This approach also allows for the identification of knowledge gaps and future research goals. It is not possible to discuss each of the PAH hallmarks in detail (excellent reviews have been published recently; 18, 60, 77, 134, 167). Instead, we chose to highlight the areas of endothelial cell mesenchymal transition (EnMT) (87), DNA damage, and mutations.

Lastly, why does this conversation about the cancer model of PAH matter? PAH patients continue to die because we still have no disease-modifying treatments for severe PAH.

Tumor, Hyperplasia Apoptosis Resistance, and Cancer

Most investigators have subscribed to Bert Vogelstein's definition that cancer “is caused by a sequential series of alterations in well-defined genes that alter the function of a limited number of pathways” and “that a set of driver genes and pathways is responsible for the most common forms of cancers. These alterations directly or indirectly increase the ratio of cell birth to cell death” (185). In contrast, a “tumor” is an abnormal or malignant new growth of tissue that possesses no physiological function and arises from uncontrolled, usually rapid cell proliferation (Merriam-Webster), while “hyperplasia” is the increase in the number of normal cells due to proliferation, which ceases after the triggering stimulus has been removed.

In severe PAH, the disease-defining cells within the vascular wall are abnormal, resulting in apoptosis-resistant cell growth (149, 182), which increases the ratio of cell birth to cell death. The concept of apoptosis-resistant cell growth—as PAH disease is defined—has found wide acceptance during the past decade (50, 167). Yet, it is important to emphasize that increased cellular proliferation in itself will lead to genetic alterations that may result in a vicious feed-forward loop, thereby raising the age-old chicken-egg question (3, 29). In Table 1, we compare some of the cellular and molecular hallmarks that are shared, or not, between cancer and PAH. In Table 2, we provide a list of the features and conditions, which do not support a cancer hypothesis.

Table 1.

Comparison of the cancer hallmarks with PAH hallmarks

Hallmarks of Cancer (65)	Hallmarks of PAH
General aspects	General aspects
“Evolution gone awry” (66)*	“Wound healing gone awry” (42)**
Systemic disease***	Small lung vessels, but PAH can be a manifestation of systemic diseases or heart diseases and there are systemic disease manifestations (115, 159)
Sustained proliferative signaling
Growth factors, survival factors	Endothelial cell and vascular smooth muscle cell growth (31, 134, 138, 174)
Hormones, increased levels of receptor proteins, anti-apoptotic mechanisms (including survivin expression)	Increased sensitivity to growth factors (76, 78, 175)
Akt/PKB, Ras, c-MYC, c-KIT, p53, HIF-1α, HIF-2α, p27Kip1, FHIT (80), HSP90 (16), NFAT, β-catenin, FGF, MDRP (144), MRP4 (85), PDGF, 5-LO (89)	Increased expression of HIF-1α, HIF-2α, ARNT, VEGF, KDR (175), NFAT (15), 5-LO (190), β-catenin (138), MDRP (138), MRP4 (67)
	Survivin (103, 138), loss of p27, Kip1 (31), HSP90 (19), c-KIT (38, 45, 46, 60, 108), FHIT (41)
Genome instability, mutations
Mutations (including BMPR2, BMP9, SOX2, KDR):“corruption of the TGF-β pathway,” PARP (65), many other gene mutations (KRAS, HRAS, BRAF, BRCA, PTEN, p53) (86, 157), KDR (170)	Mutations: BMPR2 (4), BMP9, SOX2 (60), aquaporin 1, CAV1 (177), SMAD9, ACVRL1, ENG, EIF2, AK4, KCN5, KCN3, PARP (103), KDR (44)
Chromosomal abnormalities	Chromosomal abnormalities (2)
Insensitivity to growth suppressors
Cyclin-dependent kinases	Cyclin-dependent kinases (187)
	Prostacyclin induces VSMC apoptosis via ERK1/2 (97), decreased BMPR2 signaling (4, 43)
Resisting cell death
Bcl-2 family, Bax, Bim, Puma, survivin	Decreased BMPR2 signaling (4, 43) Increased survivin expression (103, 138)
	Loss of p53 function? Bax mutation (193)
Enabling replicative immortality
Telomerase (23)	Telomerase expression in PAH (111) supports hyper-proliferation of endothelial cells
	Involvement in proliferation via the β-catenin/LEF transcriptional complex? (138)
Sustained angiogenesis
Upregulated by hypoxia and oncogenes; endogenous angiogenesis inhibitors Bone marrow-derived vascular progenitors	VEGF, KDR (175) upregulated by hypoxia and oncogenes VEGF produced by inflammatory cells Angiogenic/anti-angiogenic imbalance Endostatin (39, 72), sFLT (148) and others
IL-32 (195)	IL-32 (119)
Phenotypic instability
Epithelial mesenchymal transition	Endothelial mesenchymal transition (73, 87, 139)
5-Lipoxygenase expression (107)	5-Lipoxygenase expression (190), loss of TGF-β receptor 2 (193)
Microenvironment and inflammation providing growth signals and contributing to clonal expansion (33, 34, 156)
	Smoldering inflammation, presence of cells that provide growth and survival factors (181, 182)
Il-1α, IL-1β (90)	IL-1α, IL-1β (180)
Stem cell niche? (171) Obesity (151), leptin (151) Red blood cell distribution width	Stem cells (7, 48, 196) Obesity (23), leptin (75), dyslipidemia (70) Red blood cell distribution width
Reprogramming energy metabolism
Glycolytic switch (Warburg effect) (191), glutaminolysis	Hypoxia upregulates HIF-1α and HIF-2α, which upregulate glycolysis (Warburg effect) (5, 19, 43, 196), upregulation of GLUT1
Cancer stem cells (171)
	Bone marrow-derived stem cells (7), vessel wall stem cells (48, 155), circulating endothelial cell precursors (7, 22)
Evasion of immune destruction
Deficiency in development or function of cytotoxic CD8+ CD4+ Th1 helper cells Secretion of immunosuppressive TGF-β (66)	Immunosuppressive T regs? (163)

ACVRL1, activin A receptor like type 1; AK4, adenylate kinase 4; BMP9, bone morphogenetic protein 9; BMPR2, bone morphogenic protein receptor type II; CAV1, caveolin 1; EIF2, eukaryotic initiation factor 2; ENG, endoglin; GLUT1, glucose transporter 1; HIF-1α, hypoxia-inducible factor-1α; HIF-2α, hypoxia-inducible factor-2α; HSP90, heat shock protein 90; KDR, VEGF receptor 2; MDRP, multiple drug resistance protein; PAH, pulmonary arterial hypertension; PARP, poly (ADP-ribose) polymerase; PTEN, phosphatase and tensin homolog; sFlt, soluble VEGF receptor 1; SMAD9, SMAD family member 9; TGF, transforming growth factor.

^*This concept relates to a process “formally analogous to a Darwinian evolution, in which a succession of genetic changes, each conferring one or another growth advantage, leads to progressive conversion of normal human cells” (65).

^**“Wound healing gone awry” (42); injury to the endothelium is not repaired with a return to a normal endothelial cell monolayer; instead, exuberant endothelial cell growth (174) occurs, leading to lumen obliteration and fibrosis; inflammatory and immune cells participate in this process (174).

^***Although metastatic spread turns cancer into a systemic disease, some cancers can remain localized.

^#Although stem cells in PAH are not “cancer” stem cells, they may nevertheless participate in the formation of complex vascular lesions.

Table 2.

Characteristics and conditions that do not support a cancer hypothesis of PAH

Characteristics and Conditions
In PAH, there are no metastases to the liver, bone, brain, or adrenals.
Cells harvested from human PAH lungs do not exhibit unlimited growth.*
Tumor vessels are leaky; vessels obliterated by proliferating cells are not.**
Driver mutations of KRAS, HRAS, BRAF, and BRCA are not implicated in PAH.
There are long-term survivors of severe PAH.

PAH, pulmonary arterial hypertension.

^*The cells harvested and cultured may not be the relevant vascular lesion cells; see text for more details.

^**The initial phase of the PAH lung vessel disease is unknown; whether in early PAH vessels exhibit an increase in permeability is unknown.

In the following, we take the Hanahan and Weinberg (65) cancer hallmarks as the blueprint for our discussion. It can be appreciated that the list of PAH-associated cellular and molecular abnormalities (Table 1) is most likely a work in progress. Following the publication of the second Hanahan and Weinberg (66) paper, other authors have added additional mechanistic aspects to the original “hallmarks” and emphasized telomeres, matrix-cell interactions, and other unifying aspects (22, 101, 104, 131, 156, 157, 171). Within the limits of available published reports in the pulmonary hypertension (PH) literature, these additional aspects are also considered in Table 1 (137, 139, 143, 152, 155, 175, 178, 180, 190, 193, 196).

The 2008 publication that introduced the concept of “quasi-malignancy” of severe, angioobliterative forms of PAH (138) has resonated with many investigators. A recent “Perspective” by Pullamsetti et al. (134) outlines proliferative and pro-survival signaling pathways and potential drug targets that are of interest in cancer and PAH, emphasizing mTOR and tyrosine kinase inhibitors that target signaling hubs. Boucherat et al. (18) published a comprehensive review entitled “The cancer theory of pulmonary arterial hypertension,” D’Alessandro et al. (38) recently reviewed “The Hallmarks of Pulmonary Hypertension,” and Guignabert et al. (60) discussed in their “Pathogenesis of pulmonary arterial hypertension” “the lessons learned from cancer”. A Canadian group of investigators report the aberrant expression of several genes in the lung specimens from PAH patients and appreciate the phenomenon of cell phenotype shifts as a Darwinian (evolutionary) “cancer cell identity crisis” (20, 147). Recently, Napoli et al. (113), emphasizing “the multifactorial nature of PAH”, focused on “epigenetic factors that can bridge genetic [e.g., BMPR2 (4, 43)] and environmental diseases,” favoring a concept of “dysregulation of a network-based molecular architecture”. Such an epigenomic hypothesis of PAH encourages the discovery of triggers or “hits” that are apparently required to turn BMPR2 mutation carriers into PAH patients; epigenomics can also begin to explain molecular mechanisms underlying drug-induced forms of PAH (123). For details concerning DNA methylation and histone modifications as part of the repertoire of mechanisms underlying epigenetic modifications, the reader is referred to this publication (113) and to the reviews by Boucherat et al. (18) and D’Alessandro et al. (38). A more recent paper adds to the list of epigenetic factors the aryl hydrocarbon receptor-cytochrome P450 axis (91).

Shared Expression of Proteins in Cancer and PAH: Examples

In PAH, the cellular senescence marker p16 is expressed in many of the plexiform lesions (Fig. 2) (138), and cellular senescence (142) can now be interpreted according to the Hallmarks 2 paper (66). “Cells with morphological features of senescence, including enlarged cytoplasm and the expression of the senescence-induced β-galactosidase enzyme are abundant in the tissues engineered to over-express certain oncogenes. These ostensibly paradoxical responses seem to reflect intrinsic cellular defense mechanisms designed to eliminate cells experiencing excessive levels of certain types of signaling.” Cellular senescence is an irreversible arrest of cell growth. It can be triggered by DNA damage (105), oncogene activation, and oxidative stress (176). These triggers are present in PAH. Noureddine et al. (120) reported p16 expression in the pulmonary artery smooth muscle cells from patients with COPD and pulmonary hypertension; conditioned medium from these p16+ cells stimulated the growth of nonsenescent smooth muscle cells via released paracrine factors. The role of telomerase in PH has been discussed by Mouraret et al. (111).

Fig. 2.

Immunohistochemical analysis of plexiform lesions. A: immunohistochemical stain for cMyc (brown color). Low-power view of a plexiform lesion showing near total obliteration of the lumen along the long axis of the small pulmonary artery. B: same vessel at higher power, demonstrating the nuclear staining of the endothelial cells lining the irregular lumina of the plexiform lesion. cMyc is a transcription factor regulating sets of genes involved in cell proliferation; it is an early growth response gene that is rapidly induced when quiescent cells receive a signal to divide. C: cMyc staining of a concentric obliterative and plexiform lesion from a different patient with pulmonary arterial hypertension (PAH). The concentric obliterative lesion on the left (asterisk) shows fewer positive cells than does the plexiform lesion on the right (arrow). D: immunohistochemical staining for the senescent cell marker p16 in a plexiform lesion. Note the strong nuclear endothelial cell staining. E: immunohistochemical stain for p53 (the antibody recognizes both normal and mutated p53 proteins). There are scattered positive cells demonstrating nuclear staining (arrows). There is a lack of p53 staining in the remainder of the lung parenchyma; only endothelial cells of the plexiform lesion are positive. F: a plexiform lesion stained for the HIF-1α protein (brown). There is a predominantly cytoplasmic staining pattern of the endothelial cells. Inset shows this staining pattern more clearly. A role for hypoxia-inducible factor-2α (HIF-2α) in pulmonary arterial cell proliferation has recently been described (165).

The presence of other proteins expressed in cancer cells has been documented in the vascular lesions from PAH patients (138), for example, β-catenin and the multiple drug resistance protein (MDRP). β-Catenin functions as an adherence junction protein for cell-cell adhesion and as a cancer driver via the Wnt/β-catenin pathway; β-catenin also plays a role in the T cell infiltration of tumors (54, 188). Indeed, the β-catenin/cancer literature is too voluminous to be reviewed here. The question of whether the high expression of β-catenin in plexiform lesions (138) is indicative of endothelial cells acquiring mesenchymal or stem cell properties has not been examined. However, β-catenin is required for EnMT during heart cushion development, and we postulate that high expression of β-catenin is likely part of the pulmonary vascular EnMT program (87). The multidrug resistance protein MDR1, also known as p-glycoprotein or ATP-binding cassette subfamily B member 1 (ABCB1), is expressed in macrophages and cardiac cells and highly expressed in a number of malignancies. This protein can efflux a number of cancer drugs, and its expression in cancer cells has been shown to correlate with a poor chemotherapy response and poor survival (35, 49, 64, 86). A large number of proteins expressed in cancer and PAH, including cMyc, also known as the “oncogene from hell” (189), the nuclear factor of activated T cells (NFAT) (15), STAT3, FOXO1 (153), YAP/TAZ (11), multiple drug resistance-associated protein 4 (MRP4) (67, 85, 152), and others have been added (see Fig. 1 and Table 1).

HALLMARKS OF CANCER AND PAH: THE LAST 10 YEARS

Cell-Cell Interactions

Central to the cancer hallmarks theory (66) and likewise central to the PAH hallmarks are the multiple cell-cell interactions.

Hanahan and Weinberg (65) have emphasized the importance of the tumor environment and of stromal cells, cancer stem cells, and cells of the immune system. As monocellular in vitro cancer studies are of a limited scope, it is doubtful that the investigation of cultured endothelial cells obtained from the lungs of PAH patients, which demonstrate a limited cell growth, can elucidate the cancer paradigm of PAH. While the normal wall of the small pulmonary vessels is complex, with the syncytial relationship between endothelial cells and VSMC, the “sick lung vessels” with their phenotypically changed cells and infiltration by multiple inflammatory cell types is infinitely more complex (174, 182). There are at least two reasons why cells cultured from PAH vessels do not exhibit unlimited growth. First, it is unclear whether the relevant vascular lesion cells have indeed been harvested and cultured; presently we lack important information regarding the phenotypical characterization of these cells. For example, we do not know whether the cultured endothelial cells (EC) express 5-LO, FLAP (190), and survivin (138) or lack expression of the PGI₂ synthase and of caveolin and peroxisome proliferator-activated receptor-γ (PPARγ) (181). Second, precursor and stem cells (62, 63) and factors or cofactors that drive cell proliferation in situ may be missing in a monocellular culture environment. Complex multicellular lesions have been found at the sites of small vessel bifurcations believed to be sites of high shear stress and perhaps turbulent flow (31). Another explanation regarding the localization of the complex lesions is that the EC at these sites are phenotypically distinct or that the cell proliferation is largely driven by stem cells (48, 59, 172).

Infiltration of immune and inflammatory cells is a hallmark of cancer and plays a critical role in tumorigenesis by creating an inflammatory microenvironment that promotes tumor growth (150). Analogously, PAH is characterized by the perivascular infiltration of innate and adaptive immune cells, including mast cells, macrophages, and B and T cells (21, 88). Evidence from recent preclinical and epidemiological studies suggests a critical contribution of immunity and its regulation by inflammatory cytokines [in particular IL-6 (136)], leukotrienes [LTB₄ and LTC4 (89, 169, 190)], and damage-associated molecular patterns [DAMPs; e. g. high-mobility group box 1 (HMGB1) (194)] in the pathogenesis and/or progression of PAH. These inflammatory mediators directly (and potentially, synergistically) stimulate proliferation of lung vascular cells via paracrine and endocrine pathways (88, 152). They may also drive the formation of perivascular ectopic tertiary lymphoid tissues consisting of B cells, T cells, and high-endothelial venules (130) as germinal centers for the generation of auto-antibodies (30), which in turn promote endothelial cell proliferation and PH development (99, 130). Endothelial cells of PAH patients show increased expression levels for a variety of adhesion molecules and increased interaction with peripheral blood mononuclear cells (94). Notably, expression of the adhesion molecule P-selectin has recently been shown for pulmonary artery smooth muscle cells of PAH patients, yet not in healthy controls, so that immune cells infiltrating the vascular media may potentially stimulate vascular remodeling via direct cell-cell interaction (121). While one finds detailed mechanistic studies of tumor-associated macrophages (162) and the role of neutrophiles in tumor metastasis (1), such detailed studies are still missing in PAH research.

Parallels between the roles of inflammation and immunity in cancer and PAH have recently been highlighted by the discovery of extensive vascular remodeling in lung cancer tumor tissues and signs of elevated pulmonary artery systolic pressure and pulmonary artery enlargement in 45–50% of patients with lung cancer (135). Importantly, these morphological and hemodynamic changes are suppressed by genetic inhibition of NF-κB-dependent cytokine production in lung tumor cells in mice (141). Hence, the pro-inflammatory tumor micro-milieu favors cancer growth and lung vascular remodeling in a similar fashion. Conversely, in both cancer and PAH, the body’s natural immune defense against the hyperproliferating cells is suppressed. Cancers use inhibitory checkpoints such as CTLA-4 (cytotoxic T-lymphocyte-associated protein 4) or PD1 (programed cell death) to suppress immune responses targeting cancer cells. In the normal lung, natural killer (NK) cells prevent vascular remodeling and the development of PH as demonstrated by the development of spontaneous PH in mice deficient in NK cells (141); importantly, NK cells are reduced in numbers and impaired in function in PH animal models and PAH patients (125). A similar inverse relationship exists for regulatory T cells and PH (163). As such, both cancers and PAH seem to actively suppress mechanisms of immune defenses. This notion may explain why, despite the important role of the inflammatory microenvironment outlined above, untargeted immunosuppressive therapies, e.g., corticosteroids, do not present promising therapeutic options in either disease.

Autoimmunity, Chronic Inflammation, and Cancer

Autoimmunity and cancer have a reciprocal relationship that may be germane to our general understanding of PAH triggers and progression. Like cancer, autoimmune disease can emerge in a single organ and cause chronic inflammation because of immune dysregulation. Cancer arises in part as a failure of the organism’s immune system to detect and destroy abnormal cells and grows because of a maladaptive form of immune tolerance (145). The basis of checkpoint inhibitors for cancer is to unfetter and re-engage immune surveillance and directed destruction. Not surprisingly, autoimmunity occurs after treatment with those therapeutics as collateral damage, arising as a consequence of excessive immune activation and loss of self-tolerance. A preclinical model of immune dysregulation-associated PH shows that blocking the key PD-1/PDL-1 tolerizing signal abrogates Treg-mediated protection against PH and, in this manner, shows how autoimmunity can induce PAH in a cancer-relevant context (164).

Autoimmunity may also be a stimulus for chronic inflammation, which can drive oncogenic processes (33, 57). Chronic inflammation causes DNA damage, which can lead to cancer. Malignancy arises at sites of chronic inflammation where the microenvironment fosters cellular proliferation, survival, and migration. Neoplastic cells in turn can co-opt signaling molecules of the innate immune system, including chemokines and selectins to facilitate abnormal patterns of cellular migration and invasion. In PAH, abnormal cell proliferation, survival, and migration are highly implicated in the processes, which lead to intimal obliteration. For these pulmonary vascular syndromes, the source of chronic inflammation, which drives cancer-like changes, may evolve as a consequence of the poor immune regulation of autoimmunity.

Vascular injury can transform a singular inflammatory event into a chronic inflammatory condition manifested within the pulmonary vascular lesion itself, which becomes a nidus for immunity. A notable illustration of this idea is the acquisition of intimal 5-lipoxygenase (5-LO) expression with the evolution of PAH (190, 197). This neo-expression of 5-LO may occur as a consequence of a single inflammatory vascular injury in the setting of poor immune regulation (169). As the enzyme responsible for leukotriene biosynthesis, neointimal 5-LO expression can drive the recruitment of inflammatory cells and create a self-sustained chronic inflammation that drives tissue changes reminiscent of malignant transformation (109). As we have mentioned, apoptosis-resistance, as documented by absence of protein signals in the complex vascular PAH lesions, which identify cells undergoing apoptosis, has been associated with the development of autoantibodies (6).

Limitations and Knowledge Gaps

As pointed out, autoantibodies have been associated with several forms of severe PAH (30), and mechanistic concepts are largely derived from animal models. We are uncertain whether infections trigger a lung vascular immune response and whether endothelial cell injury exposes epitopes that attract antibodies in the context of molecular mimicry.

Apoptosis, Phenotype Switch, DNA damage, Autophagy, and Anoikis

Tumor cells evolve a variety of strategies to limit or circumvent apoptosis. Most common is the loss of p53 tumor suppressor function (Figs. 2 and 3). Alternatively, tumors may achieve similar ends by increasing the expression of antiapoptotic regulators (Bcl-2, Bcl-xL) or of survival signals or by downregulating proapoptotic factors (Bax, Bim, Puma). Recent research has revealed intersections between the regulatory circuits governing autophagy, apoptosis, and cellular homeostasis (66). In the PAH vascular wall, cells undergo a phenotypic switch that generates apoptosis resistance, and endothelial cell mesenchymal transition (EnMT) is just one manifestation of the spectrum of pulmonary vascular endothelial cell phenotypes that spans from the quiescent monolayer to endothelial cell activation and inflamed endothelial cells to angiosarcomatous cells that have been detected in thromboendarterectomy samples (9).

Fig. 3.

This diagram illustrates an example of how connecting dots between inflammation and hypoxia, DNA damage and repair, glycolysis control, oncogenes, p53, apoptosis, and emergence of apoptosis-resistant proliferating cells can generate specific hypotheses. The general hypothesis underlying this concept is that the initial pulmonary vascular endothelial cell damage is the critical disease-initiating event in genetically susceptible individuals. The diagram highlights the importance of endothelial cell DNA damage.

Phenotype Switch and Endothelial Cell Mesenchymal Transition

Altered cellular energy metabolism, in particular aerobic glycolysis (the Warburg effect), is a hallmark of cancer but also of severe angioproliferative PAH (5, 16, 18, 43, 134, 196). Glycolysis is controlled, among others, by p53, the oncogene cMyc and HIF (Fig. 3) (36, 68). Hypoxia-inducible factor-1α (HIF)-1α and HIF-2α play roles in cancer and PAH (36, 37); by binding to the hypoxia response elements (HRE), these proteins control the transcription of a large number of genes (the expression of HIF-1α in PAH arterioles is shown in Fig. 2). Because the mitochondrial and metabolic abnormalities in PAH cells have been extensively reviewed (5, 196), we restrict our discussion to examine why and how the Warburg effect supports cell proliferation in PAH. Briefly, when Otto Warburg in 1924 discovered that cancer cells “fermented” glucose to lactate, neither oncogenes nor growth factor signaling pathways had been discovered. Nowadays, it is appreciated that the metabolic requirements for cell proliferation are met through several signaling pathways that regulate both cell proliferation and metabolic pathways and that oncogenic mutations enable cancer cells to metabolize nutrients for proliferation, rather than efficient ATP production [see Vander Heiden et al. (177) for more details]. This redirection of energy metabolism, the “metabolic switch,” is one characteristic underlying the phenotype switch. Figure 3 illustrates an exercise in “connecting dots” to arrive at a hypothesis that needs to be tested in the context of PAH. The possible and perhaps likely interactions between glycolysis and cell growth and the influence of p53 and cMyc can be dissected when phenotypically relevant cells have become available.

EnMT is one example of the cell phenotype switch and may be the PAH counterpart of epithelial cell transition (EMT) in cancer, although EnMT does also occur in cancers (132). EnMT is the process whereby endothelial cells lose their typical endothelial cell markers and, importantly, their functions and adopt a mesenchymal-like cell type. A comprehensive recent review article by Kovacic et al. (87) dicusses EnMT in the context of cardiovascular diseases. Partial (pEnMT) and complete EnMT (cEnMT) cells have been described; cEnMT cells exert paracrine effects, stimulating mesenchymal cell migration and endothelial angiogenic capacity (161). One driver of EnMT is TGF signaling (87, 127), and EnMT can also be stimulated by extracellular vesicles. Inflammation via downregulation of BMPR2 expression promotes EnMT (73, 152), and so does galectin 3 (99). Mechanical factors and metabolic changes and HIF-2α (165) are also triggers (191). Finally, Tsutsumi et al. (173) reported that nintedanib reduced EnMT in the Sugen/hypoxia rat model; however, this was not confirmed by the study by Rol et al. (146). An important unanswered question is whether EnMT in the “sick lung circulation” (182) is central to pulmonary vascular remodeling and amenable to treatment.

DNA Damage, Causes, and Consequences: Apoptosis and Cancer

DNA damage (DNA breaks and other chromosomal abnormalities) triggers apoptosis but also cellular senescence and the selection of apoptosis-resistant cells (Fig. 3). Indeed, DNA damage and chromosomal abnormalities in PAH have been reported (2, 26, 96, 105, 140). Federici et al. (47) examined DNA damage in lung and blood cells from patients with severe forms of PAH and also found an increased sensitivity to DNA damage in relatives of PAH patients when exposing peripheral blood monocytes to DNA-damaging agents.

DNA damage results in an abnormal chemical structure of DNA. It is caused by inflammation- and hypoxia-related reactive oxygen species (186) and by (environmental) toxins. DNA damage triggers the activation of various damage response pathways and changes the activation of a number of cell fate controlling genes, including p53, p21, NF-κB, and BRCA1 (the reader is referred to several excellent reviews of the topic of DNA damage and repair; see Refs. 10, 53, 71, 128). Most cancers are associated with p53 mutations. One of the many functions of p53 is the suppression of glycolysis (81). Figure 2 shows the expression of p53 in a plexiform lesion; because the antibody used for this immunohistochemistry (IHC) stains mutated p53 as well as wild-type p53, we thus identify a knowledge gap and the need to examine whether there are p53 mutations in PAH. Whether there are p53 polymorphisms in PAH patients is also unknown. Figure 3 is a simplified diagram that shows how oncogenes, DNA damage, glycolysis, apoptosis, and cancer are connected. This diagram may serve as an aid to further investigate how DNA damage to pulmonary vascular cells results in a phenotype switch [see also (128)]. As an aside, there is evidence that endothelin 1 promotes anti-apoptosis in keratinocytes (81, 97). It would be of interest to find out whether endothelin 1 also contributes to the apoptosis resistance of endothelial cells in PAH.

It is still unclear to what extent the Sugen 5416-based models of severe PAH (116) mimic the various forms of severe human PAH, yet the Sugen/chronic hypoxia model data suggest that the profound pulmonary arteriolar obliterations are initiated by endothelial cell apoptosis. Sugen 5416 inhibits VEGF receptor 2 (KDR) (166), and a recent report describes KDR mutations, resulting in impaired VEGF signaling, in patients with familial PAH (44). In line with this notion of pathobiologically important endothelial cell apoptosis, induction of endothelial apoptosis suffices to induce obliterative vascular lesions in mice (58). Translating these findings to the human condition lets us arrive at a model of human PAH; initial endothelial cell damage and apoptosis is followed by the evolution of apoptosis-resistant cell types. Ngo et al. (114) found obliterated lung vessels in Ku 70-knockout mice. Ku 70 (Lupus Ku antigen p70) is essential for DNA double-stranded repair; Ku 70 is anti-apoptotic by suppressing intrinsic cell death signaling via Bax. Absence of Ku 70 leads to Bax hyperactivation. This finding is likely of importance because of the loss of the Ku 70 gene expression detected by gene microarray expression analysis of lungs from patients with idiopathic PAH (56).

Autophagy (51, 79) normally operates at low basal levels but can be strongly induced by cell stress. The autophagy marker LC3B was detected in the lungs from PAH patients (93); the combination of two cell stressors caused autophagy in macaque lungs (122), and Kato et al. (83) showed that autophagy can slow PAH progression. “Paradoxically, autophagy can be protective in cancer cells; severely stressed cancer cells have been shown to shrink via autophagy to a state of reversible dormancy. This survival response may enable the persistence and eventual regrowth of some latent -stage tumors following treatment with anticancer drugs” (66).

Anoikis (Greek: “homeless”) is defined as cell death following the detachment of cells from their matrix. Anoikis resistance is part of the survival skills of cancers. Interestingly, anoikis resistance can be induced by shear stress (98), STAT3 (100), and β-catenin (95, 124). We speculate that, in analogy to cancer, the phenotype switch of the cells and also an “integrin switch” (32) in the lumen-obliterating cell aggregates also affect anoikis-resistance. In the Sugen/hypoxia model of severe, angio-obliterative PAH, induction of anoikis markers has been reported, as shown by immunohistochemistry and protein analysis (14).

Angiogenesis: Anti-angiogenesis

In severe forms of PAH, we find the pruning of the pulmonary vascular tree and angioobliteration due to a process of “misguided angiogenesis” (175). These side-by-side manifestations of vascular changes in PAH lungs have been a pathobiological puzzle. One possible explanation for the coexistence of pruning and obliteration is that both proangiogenic and anti-angiogenic factors are involved in the pathogenesis of PAH (39, 72, 148). However, if the root cause of vascular obliteration is endothelial cell apoptosis, as in the Sugen/hypoxia rat model (166), and, if indeed these findings that describe early and likely subsequent cell proliferation-initiating EC apoptosis in the animal model reflect early disease stages in human PAH, then the vascular homeostatic imbalance is initially tilted toward anti-angiogenesis. Yet functionally, the pruning of the vascular tree is not sufficient to fully explain the development of severe PAH, because many patients with end-stage emphysema and dramatic loss of lung vessels do not manifest severe PAH at rest.

VEGF is a cardinal angiogenesis and endothelial cell maintenance factor. Experimentally, inhibition of VEGF signaling causes emphysema associated with lung vessel loss (82). During the last decade, circulating anti-angiogenic proteins have been documented in PAH patients, and there are angiogenic and anti-angiogenic splice variants of the VEGF ligand (183). In this context, severe angioproliferative PAH may be a disease of anti-angiogenic loss of small lung vessels and a parallel or subsequent activation of an angioproliferative program. Anti-endothelial cell antibodies and molecular mimicry may play a role together with the anti-angiogenic circulating sFlt (soluble VEGF receptor 1) and endostatin (39, 92). While it is appreciated that angiogenesis is defined as the growth of new vessels from existing vessels in PAH, the growth of new vessels is within the vascular lumen (Fig. 2). Similarly as in tumor angiogenesis, the lumen-filling endothelial cells that form pseudo-lumina are abnormal and phenotypically switched (see above).

Histological analysis of premalignant, noninvasive lesions, including dysplasias and in situ carcinomas, has revealed the early tripping of the angiogenic switch. Some tumors, including such highly aggressive types as pancreatic ductal adenocarcinomas, are hypovascularized and replete with stromal “deserts” that are largely avascular and indeed may be actively antiangiogenic. The switching mechanism can vary in its form, even though the net result is a common inductive signal (e.g., VEGF). In some tumors, dominant oncogenes such as Ras and Myc (Fig. 2) can upregulate expression of angiogenic factors. Interestingly, wound healing is impaired by elevated expression of anti-angiogenic factor genes (66).

We conclude this discussion of the angiogenesis hallmark of PAH with a word of caution; well-intended anti-angiogenic treatment strategies designed to induce cell death of the pulmonary vascular lesion-obliterating cells may cause as collateral damage vessel loss in other organs, in particular in the failing right ventricle.

To summarize, while EnMT in PAH had previously been declared a hallmark that when present would strengthen the quasi-malignant concept of PAH (107), EnMT has now been documented (87). The inclusion of the angiogenesis/anti-angiogenesis part of the cancer-like disease model of PAH is justified, although the data in support of this concept are presently associations and correlations, and interventional studies in human PAH are lacking. Further studies, which would support a participation of autophagy and anoikis, need to be conducted using human PAH tissues.

Bone Marrow, Stem Cells, and Metastases

There is convincing evidence for a bone marrow-lung circulation axis (7, 8). Clinicians have long appreciated that patients with myelodysplastic disorders can develop severe PAH (55). On the other hand, there have been reports of myelofibrosis in patients with idiopathic PAH (133). There are also reports on patients developing severe PH following bone marrow transplantation (40). In addition, megakaryocytes (Cool CD, unpublished observations) and dendritic cells of bone marrow origin are found in and around plexiform lesions, and it remains unresolved whether some of the mast cells that are found in the plexiform lesions arrived from the bone marrow.

It has been pointed out that one of the hallmarks of cancer, “metastases,” is missing in PAH. “In addition to local signaling within the primary tumor environment, tumors also signal over long distances to sites of future metastases to promote the formation of a hospitable pre-metastatic niche that will foster growth of disseminated tumor cells upon their arrival” (129). We appreciate that avenues connecting the bone marrow and the lung vessels are the systemic bronchial vessels and the vasa vasorum (118). In severe PAH, the “sick lung circulation” (181) inflamed endothelial cells, macrophages, mast cells, and others, producing factors (for example, VEGF) that activate the bone marrow to release cells, including precursor cells. Abnormal endothelial cells are likely being sheared off and enter the systemic bronchial circulation, although the proof of such a “metastatic” lung vascular cell release is still pending. Such sheared-off cells could seed at the site of bronchial-pulmonary arterial anastomoses and form plexiform lesions (52). Perhaps liquid biopsies and single-cell RNA sequence analysis can test such a hypothesis. While there is no evidence of local vascular invasion, hematogenous and lymphatic spread of “sick lung cells” is possible.

Presently, a complete understanding of bone marrow-derived precursor cells, circulating and pulmonary resident stem cells, is lacking. Animal experiments have yielded data that demonstrate that bone marrow cells contribute to the pulmonary vascular disease but also that hematopoietic stem cells protect against PH (8, 60, 192).

Systemic Disease Components of PAH

Recent studies have drawn attention to extrapulmonary disease manifestations in patients with severe PAH. Coronary artery disease, renal disease, and skeletal muscle capillary rarefaction have been reported; a comprehensive review of these manifestations has recently been published (115), and the reader is also referred to a recently published pro and con debate (159). These systemic manifestations do not necessarily mean that severe idiopathic forms of PAH are systemic diseases. The more likely mechanistic interpretation is that the “sick lung circulation” produces “bad humors” that can cause damage in the periphery, for example, circulating anti-angiogenic proteins and lipids (115). Pathobiological contributions, either trigger factors or disease amplifiers, that emanate from the adipose tissue, as well as hormonal influences, need to be considered. Because there is an association between obesity and cancers (168), there is also an association between obesity and severe forms of PAH. Data from the REVEAL registry of pulmonary hypertension show that 30% of the PAH patients in the United States have a BMI > 35 (23). Adipose tissue is inflamed and undergoing angiogenic capillary sprouting and is a source of cytokines and growth factors (69, 74), and estrogen is produced by adiposites (102). These factors either directly or indirectly can affect injured lung vascular endothelial cells. It is also of interest that obesity has been associated with telomere shortening (112) and that adipose tissue undergoes EnMT (69).

The role of estrogen in breast and uterine cancers is well established, and in recent years the role of estrogen in the pathogenesis of severe PAH has received wide attention (27, 179). Estrogen receptor-related-α activates VEGF gene expression (160), perhaps one bridge between cancer and PAH. Estrogen metabolism is under the control of cytochrome P450 enzymes, another bridge between PAH and cancer (91).

SUMMARY AND OUTLOOK

Here, we have reexamined the quasi-malignancy concept of severe PAH. Ever since the publication of the “Cancer paradigm of PAH” article (138) in 2008, many publications provide data that support this concept (Fig. 1), including the recently pronounced “epigenomics” hypothesis (113). To summarize, there are robust hallmarks, which are shared between PAH and cancer: the phenotypic switch, the angiogenic switch, and the glycolytic switch. On the cellular level, the issue is one of cell survival and cell fate. In PAH, too many cells in the vascular wall are surviving, and too many cells are abnormal.

The past decade has evolved a more nuanced concept of the cancer hallmarks (13, 20, 25, 29, 34, 101, 156, 157, 171), and likewise, investigators in the PH research field continue to refine the hallmarks of PAH (4, 18, 38, 60, 70, 77, 106, 134). It is likely that advances in genotyping and phenotyping (61, 63) will also occur by putting to use liquid biopsies (12, 28) and analysis of circulating stem cells. For example, a recent whole genome-sequencing analysis of a large number of PAH patients discovered new rare mutations in the BMP9, aquaporin 1, and SOX17 genes (61, 62, 110), all of which have been associated with cancer and metastases. The last decade of PAH research has enlarged our knowledge of the pathobiology of severe forms of PAH by adding information relating to vascular inflammation and metabolic abnormalities of vascular cells; EnMT and DNA damage have been documented, and new gene mutations (44) have been discovered.

Are we there yet? We are not, but we are getting closer, and we are seeing the light at the end of the tunnel (184). What would novel, successful treatments for severe PAH look like? They would be disease modifying. In our opinion, there are two desirable treatment outcomes: 1) reversal of the disease in incident cases and 2) halting disease progression and stabilization of cardiac performance. For the second outcome goal, there is an example. It has not been universally appreciated that many of the aminorex-induced IPAH patients who had been diagnosed in the 1970s survived without treatment for many years. Their PAH did not disappear, but their disease did not progress.

If indeed many forms of angioobliterative PAH are quasi-malignant, as we suggest, then we can hope, and perhaps expect, that during the coming years new PAH treatment strategies will emerge that have been informed by cancer research, yet not to the exclusion of the concept of a vasoconstrictor/vasodilator imbalance. For instance, one can postulate that the disease of incident patients may be modifiable and that, as in cancer, there will be treatment responders and non-responders (184). Perhaps, informed by the hypothesis of a quasi-malignant pathobiology of severe PAH, clinical trials have begun to examine the effects of non-vasodilator drugs; examples are studies that have tested imatinib (45), bardoxolone (126), aromatase inhibition (84), and tacrolimus (158). Studies testing the treatment effect of rituximab in patients with scleroderma-associated PAH (117) and with 6-mercaptopurine in PAH patients (17) have recently been completed; cyclin-dependent kinases are also being considered as PAH treatment targets (187).

Visions of future PAH treatment strategies are to induce anoikis in anoikis-resistant lung vascular cells [see Bogaard et al. (14); this may require a detailed knowledge of integrin-matrix interactions that we are presently lacking), to correct the multilayered immunological deficiencies, and to reverse EnMT and steer stem cells toward differentiation.

Likely, in our opinion, future treatment strategies will target central disease components, like apoptosis-resistance and EnMT and immune mechanisms, rather than one particular enzyme or receptor. When strategies have been developed that reopen obliterated lung arterioles and venules, such treatments would reduce the pulmonary vascular resistance and unload the stressed right ventricle. We predict that strategies with the potential to reopen occluded pulmonary arterioles or halt disease progression will in all likelihood be informed by cancer research data and concepts.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

C.D.C., W.M.K., H.J.B., E.S., M.R.N., and N.F.V. prepared figures; C.D.C., W.M.K., H.J.B., E.S., M.R.N., and N.F.V. drafted manuscript; C.D.C., W.M.K., H.J.B., E.S., M.R.N., and N.F.V. edited and revised manuscript; C.D.C., W.M.K., H.J.B., E.S., M.R.N., and N.F.V. approved final version of manuscript.