Pulmonary hypertension begets pulmonary hypertension: mutually reinforcing roles for haemodynamics, inflammation, and cancer-like phenotypes

This editorial refers to ‘Haemodynamic unloading reverses occlusive vascular lesions in severe pulmonary hypertension’ by K. Abe et al., pp. 16–25.

Pulmonary arterial hypertension (PAH) is a devastating disease of the pulmonary vasculature in which vasoconstriction, inflammation, and a cancer-like phenotype, defined as excessive cell proliferation, impaired apoptosis, and mitochondrial metabolic abnormalities, contribute to vascular obstruction. The resulting increase in right ventricular (RV) afterload ultimately leads to death from RV failure. The histology of PAH includes a variable mixture of lesions in small pulmonary arteries (PAs), including medial hypertrophy, intimal hyperplasia, adventitial fibrosis, and plexiform lesions, as well as fibrosis and reduced compliance of large PAs.

PAH is a heterogeneous syndrome; however, several phenotypes are recognized that offer mechanistic and therapeutic insights. PAH has a haemodynamic phenotype [increased vasoconstriction and mean pulmonary artery pressure (mPAP), due in part to endothelial dysfunction and changes in ion channel function and calcium homeostasis in pulmonary artery smooth muscle cells, PASMC], a cancer-like phenotype (increased PASMC proliferation and apoptosis resistance associated with a glycolytic metabolic shift, epigenetic silencing of key redox regulatory genes such as SOD2,¹ and an increased mitochondrial fission/fusion ratio²), and an inflammatory phenotype [increased perivascular accumulation of proinflammatory macrophages and changes in the population of immunoregulatory T-cells (T-Reg)³ and natural killer cells].⁴

Current PAH therapies (prostanoids, phosphodiesterase-5 inhibitors, endothelin receptor antagonists, and soluble guanylate cyclase activators) are vasodilators and primarily address PAH's haemodynamic phenotype. They improve symptoms and reduce hospitalization but, on average, only decrease mPAP ∼5 mmHg.⁵ Their limited haemodynamic efficacy and lack of effects on PAH's inflammatory and cancer-like phenotypes suggest that they are unlikely to reverse the histological lesions of PAH. Indeed, chronic intravenous epoprostenol, the gold standard of haemodyamically active therapeutics, does not regress PAH's complex pulmonary vascular lesions.⁶ In this context, the observation by Abe et al. that an intervention which has a profound haemodynamic effect (left pulmonary artery banding, LPAB) regresses distal PAH histology is remarkable.⁷

In this issue of Cardiovascular Research, Abe et al. investigated mechanisms and potential reversibility of the vascular remodelling in PAH.⁷ They used a well-accepted rodent PAH model to show that abnormal haemodynamics (once established) drive adverse vascular remodelling. They also suggest that the haemodynamic stress of PAH drives the proinflammatory milieu. Abe et al. assessed the effects of reducing pressure and flow on the formation of complex vascular lesions in the rat Sugen5416/hypoxia/normoxia model of PAH (Su/Hx/Nx). They did this in a clever manner, comparing vascular remodelling in the right lung, which was exposed to the full haemodynamic stress of PAH, with the left lung, which was protected from increased pressure and haemodynamic stress by LPAB. LPAB dramatically reduced adverse vascular remodelling (but only in the left lung). LPAB also significantly reduced left lung inflammation, evident as a decrease in perivascular macrophage accumulation, NFκB activity, and inflammatory cytokines [such as interleukin-6 (IL-6)]. The authors conclude that the pulmonary vascular pathology in PAH is primarily sustained by haemodynamic stress and excessive vasoconstriction. They also suggest that this haemodynamic view of PAH may be sufficient to explain the vascular remodelling in PAH and question the role of PAH's cancer-like phenotype.⁸

The Abe paper raises the possibility that if PAH-targeted therapeutics could more effectively lower mPAP, they might relieve inflammation and regress PAH. Supporting the proposed importance of haemodynamics in sustaining PAH is the excellent prognosis seen in 5–10% of the idiopathic PAH patients who have a vasodilator-responsive haemodynamic phenotype (defined as a 10 mmHg fall in mPAP to <40 mmHg in response to acute vasodilator challenge).⁹ These patients can be maintained on calcium channel blockers and have excellent long-term survival.¹⁰

However, because PAH patients tolerate systemic hypotension poorly, any vasodilator must be quite specific for the pulmonary circulation (although selectivity may be achieved by nebulization). An example of the potential value of a potent vasodilator is fasudil, a Rho kinase inhibitor. Acutely, fasudil causes profound pulmonary vasodilation and improves acute haemodynamics in PAH patients.¹¹ When given chronically in rats with monocrotaline-induced PAH (MCT-PAH), fasudil improves RV function and decreases pulmonary vascular pathological changes.¹²

While Abe's finding that pulmonary hypertension begets pulmonary hypertension is an innovative contribution, there are limitations to this study. First, the study design was not parallel between control and LPAB groups. It would have been important (and logical) to have haemodynamic and histopathologic measurements from a control group studied at the same time point in control and LPAB groups.

While it is tempting to infer that the histologic improvement derived from LPAB might have therapeutic benefit, the authors find the opposite. LPAB, despite its histological benefits, elevated RV systolic pressure, increased right ventricular hypertrophy (RVH), and worsened RV failure, perhaps reflecting the hyperperfusion of the non-banded right lung. This highlights another study limitation, the relatively limited haemodynamic and functional characterization of these animals. In order to parse the effects of LPAB on the RV, an important predictor of prognosis in PAH,¹³ it would be important to document cardiac output, exercise capacity, and brain natriuretic peptide levels.

Abe et al. appear to assert a central role for vasoconstriction in the adverse vascular remodelling and infer that PA banding is somehow reducing vasoconstriction (Abe Figure 1); however, they did not measure dynamic vascular tone. Moreover, LPAB would be expected to reduce shear stress, pressure, and flow; but it is unclear why it would alter vasoconstriction.

Figure 1.

A mechanistic figure proposing integration of the role of pulmonary hypertension in progression of vascular disease in PAH as reported in the study by Abe et al., with the neoplastic paradigm of PAH pathogenesis. PA, pulmonary artery; PVR, pulmonary vascular resistance; RV, right ventricle; IL-1ra, interleukin-l receptor antagonist; LPAB, left pulmonary artery banding.

The authors' conclusion that pulmonary hypertension drives cellular inflammation, cell proliferation, and adverse vascular remodelling is consistent with published data. Perivascular inflammation occurs in all subgroups of PAH,¹⁴ and numerous cytokines and chemokines are involved in disease pathogenesis. In MCT-PAH, PASMCs generate increased IL-1, and both PA pressure and RVH are reduced by a recombinant human IL-l receptor antagonist. Likewise, in the Su/Hx/Nx model, inhibition of NFκB, using pyrrolidine dithiocarbamate, reduces obliterative lesions and improves cardiac output.¹⁵ Similarly, in hepatopulmonary syndrome, intravascular accumulation of activated CD68(+) macrophages plays a central role in adverse vascular remodelling, and in vivo depletion of macrophages regresses the syndrome.¹⁶

Based on the ability of their haemodynamic intervention to reduce inflammation and regress vascular lesions, Abe et al. suggest that PAH's cancer-like phenotype, with its mitochondrial metabolic abnormalities, has limited relevance; however, this phenotype was not assessed. The neoplastic view of PAH invokes acquired changes in mitochondrial metabolism and mitochondrial dynamics as an explanation for the increases in cell proliferation, apoptosis resistance, and the Warburg metabolic phenotype (a shift of mitochondrial oxidative metabolism to favour uncoupled glycolysis).^8,17 Metabolic changes in PAH partially reflect enhanced expression and activity of pyruvate dehydrogenase kinase (PDK), a family of four enzymes that inhibit pyruvate dehydrogenase in PAH PASMC and RV myocytes.¹⁸ Inactivation of mitochondrial metabolism in PAH PASMC reduces apoptosis, permitting unchecked cell proliferation, and vascular obstruction, while in RV myocytes it impairs ATP generation, causing RV failure.

PAH's neoplastic phenotype also encompasses abnormalities of mitochondrial dynamics, marked by mitochondrial fragmentation. Mitochondria normally exist in networks; however, in PAH PASMC, there is increased mitochondrial division, occurring in coordination with cell division. This is called mitotic fission. Mitochondrial fragmentation in PAH reflects increased fission (due to activation of dynamin-related protein 1) and decreased fusion (due to down-regulation of mitofusin-2). Mitotic fission is precipitated by the same mitogens and kinases that drive mitosis and support excessive rates of cell proliferation.² These same abnormalities of mitochondrial metabolism and dynamics are seen in cancer (reviewed in²). The evidence supporting PAH's cancer-like phenotype is substantial. Dichloroacetate (DCA), a PDK inhibitor, reduces cell proliferation, restores apoptosis sensitivity, and improves RV function leading to regression of PAH in multiple animal models.^17,19,20 However, this does not refute Abe's view that haemodynamics are of central importance. PAH's metabolic abnormalities can also be corrected with effective haemodynamic intervention. For example, epoprostenol reduces the elevated RV uptake of [¹⁸F] fluorodeoxyglucose (¹⁸FDG) in PAH patients, suggesting that glucose metabolism can be corrected by a haemodynamically targeted treatment.²¹ Moreover, imatinib (which improves pulmonary haemodynamics by inhibiting the platelet-derived growth factor pathway) normalizes lung ¹⁸FDG uptake and regresses adverse pulmonary vascular remodelling in rodent PAH models.²²

The gene-environment milieu in PAH, composed of genetic (BMPR2 mutations), epigenetic, and environmental (drugs and toxins) factors, predisposes to endothelial injury. This promotes vasoconstriction and cell proliferation, which activates a downstream cascade of abnormalities, including changes in mitochondrial metabolism and dynamics, alteration in fibrosis and elastin metabolism, changes in the expression and function of ion channels and inflammation.¹ We suggest that PAH's cancer-like and inflammatory phenotypes may both be exacerbated by elevated mPAP (Figure 1).

In conclusion, the work of Abe et al. adds to our knowledge by showing that pulmonary hypertension itself drives adverse vascular remodelling, perhaps by increasing perivascular macrophage accumulation and cytokine release.⁷ This suggests that substantially improving haemodynamics might reverse pulmonary vascular remodelling in PAH. The possibility that PAH's haemodynamic, inflammatory, and cancer-like phenotypes are mutually reinforcing merits further study.

Conflict of interest: none declared.

Funding

S.L.A. received grants from CIHR Foundation, NIH-RO1-HL071115 and 1RC1HL099462; a Tier 1 Canada Research Chair in Mitochondrial Dynamics; and the William J Henderson Foundation.