Periostin

Periostin—a matricellular ligand for integrins originally identified in periosteum—promotes cell adhesion and migration upon activation following tissue injury and remodeling.^1,2 Periostin is upregulated in asthma, idiopathic pulmonary fibrosis, and cardiac fibrosis, serving as a biomarker of treatment-responsive endophenotypes and lineage marker of effector fibroblasts.^2–6 Periostin is increased in the lungs of rodents with hypoxia-induced experimental pulmonary hypertension (PH) and in hypoxia-treated pulmonary artery (PA) smooth muscle cells.^7,8 A seminal proteomic analysis of lung tissues from patients with idiopathic or heritable PA hypertension (PAH) revealed increased expression of periostin, suggesting a role in human pulmonary vascular disease (PVD).⁹ Modulation of periostin-expressing lineages has been proposed for the treatment of PH.¹⁰ However, the significance of periostin itself as a pathogenetic factor, biomarker, or therapeutic target in PVD has yet to be demonstrated.

In the current issue, Nie et al¹¹ establish periostin as a marker and mediator of pulmonary vascular remodeling associated with hypoxic lung disease. Periostin was overexpressed in lung homogenates, PAs, and PA endothelial cells (PAECs) from patients with World Health Organization group 3 PH due to hypoxic lung disease, including idiopathic pulmonary fibrosis and chronic obstructive pulmonary disease. Elevated plasma levels of periostin were associated with worse outcomes, including hospitalization for PH, lung transplantation, or death, while periostin expression in lung tissues correlated with hemodynamics, suggesting periostin is a prognostic marker of group 3 PH. Toward demonstrating a mechanistic role, the authors documented overexpression of periostin in mice exposed to hypoxia or combined exposure to SU5416 and hypoxia. Depletion of periostin using periostin-deficient mice or siRNA was protective in these two mouse models and diminished release of several cytokine growth factors including TGF-β1 (transforming growth factor beta 1). Improvements in pulmonary vascular remodeling were accompanied by decreased proliferation and increased apoptosis of PAECs, suggesting beneficial effects were due to cell-autonomous antiproliferative and proapoptotic effects of periostin depletion in endothelium. The effects were not limited to endothelial cells, as hypoxia promoted periostin secretion from PAECs into culture supernatants that could elicit proliferation and migration of PA smooth muscle cells.

Importantly, the authors identified several synergies and feedback mechanisms linking hypoxia and periostin, involving HIF1α (hypoxia-inducible factor 1-alpha) and TrKB (tropomyosin receptor kinase B). In PAECs, hypoxia enhanced the expression of periostin and TrkB, while the combination of periostin and hypoxia enhanced the production of TrkB ligand BDNF (brain-derived neurotrophic factor), each of which in turn increased HIF1α expression, establishing HIF1α as a critical node of this positive feedback loop. This signaling mechanism also regulated the expression of the vasoactive protein endothelin-1 and the angiogenic vascular endothelial growth factor, suggesting periostin may be a key mediator by which hypoxic signaling disrupts the homeostasis of endothelial cells.

Given that periostin regulates TGFβ and BMP (bone morphogenetic protein) signaling in the fibrotic remodeling of other organs,^7,8,10 and the importance of BMPR2 (BMP type 2 receptor) and downstream SMAD1/5/9 signaling in heritable PAH, the authors investigated a potential link between periostin and BMPR2 expression. Silencing of periostin normalized the diminished BMPR2 expression and BMP-SMAD1/5/9 signaling seen in experimental PH and endothelial cells from PH patients. Conversely, silencing of BMPR2 in PAECs did not alter periostin expression, suggesting periostin acts upstream of BMPR2 in this context. These findings could suggest the potential trans latability of periostin, as other promising experimental therapies for PAH attenuate TGFβ or preserve BMP signaling to restore their balance in the pulmonary vasculature.^12–14 Periostin is reported to function both up- and downstream of TGFβ signaling. For example, an inducible dominant negative mutant TGFβ type II receptor allele abrogated the induction of lung periostin in mice exposed to chronic hypoxia.⁸ The development of hypoxia-induced PH was blunted in mice lacking the TGF-β type I receptor specifically in periostin-expressing cells.¹⁰ In idiopathic pulmonary fibrosis patients, monocytes and fibrocytes release periostin and TGFβ induces periostin in lung mesenchymal cells.³ While silencing periostin reduced expression of TGFβ1 from PAECs in the current study,¹¹ it is not known if depletion of periostin similarly attenuates downstream SMAD2/3 signaling, as such a mechanism could also contribute to therapeutic effect. It is also unclear how the impact of periostin on BMP and TGFβ signaling integrates with feedback mechanisms involving hypoxia and HIF1α.

Whereas the earlier proteomic study reported prominent periostin expression in the airways and neointima of remodeled lobar PAs in patients with World Health Organization group 1 PAH,⁹ the current study found periostin most abundantly expressed in the endothelium of distal vessels in group 3 PH, suggesting that anatomic localization of periostin expression may depend on the etiology of PH or PAH. Periostin appeared to exert potent cell-autonomous functions in vascular endothelium, as hypoxia-induced periostin in PAECs potentiated HIF1α responses to enhance proliferation, migration, and angiogenesis while reducing apoptosis. Hypoxia also enhanced the secretion of periostin from these cells to promote the proliferation and migration of PA smooth muscle cells. Considering the elevated plasma periostin levels in World Health Organization group 3 PH, these data are compatible with endothelial-derived periostin exerting cell-autonomous, autocrine, paracrine, and endocrine effects in mediating PVD. Therapeutic approaches for inhibiting periostin function, such as neutralizing antibodies to the ligand or its α_Vβ₃ or α_Vβ₅ integrin receptors would thus need to consider whether secretory versus nonsecretory functions of periostin are most critical in its pathogenic function. The potential impact of periostin on the phenotypic plasticity of vascular lineages should also be investigated, including the endothelial-to-mesenchymal transition of PAECs, and myofibrogenic phenotypic switching of PA smooth muscle cells previously implicated in PVD pathogenesis. Given the importance of periostin-expressing fibroblasts in promoting fibrosis, it is likely that periostin functions as an integrator of hypoxic and fibrotic signaling in the context of PVD.

Periostin appeared to have similar contributions in hypoxia and SU5416/hypoxia models of PH, but it is unclear if the role of periostin derives solely from the hypoxia component common to both models or if periostin contributes to the PAH-like phenotypes of SU5416/hypoxia–treated mice. Taken together with the known overexpression of periostin in idiopathic PAH patients,⁹ one could speculate that periostin represents a therapeutic target for World Health Organization group 1 PAH in addition to group 3 PH. With more translatable pharmacological inhibitors, the therapeutic potential of periostin in PAH could be explored further in models of severe obliterative PH such as the SU5416/hypoxia–exposed rat.

While the current study was focused on the pulmonary vasculature, periostin is also known to be highly upregulated in the right ventricle (RV) of rats with PH due to monocrotaline,^15,16 suggesting a role of periostin in RV fibrosis analogous to its known role in left ventricular fibrosis. Increased periostin levels enhance expression of inducible NO synthase and subsequent NO production in RV fibroblasts, which promote RV failure through the suppression of L-type Ca2+ channel activity of cardiomyocytes in PH rats.¹⁷ It would be important to discern whether the impact of systemic periostin depletion derives from its role in the pulmonary circulation, RV fibrosis, or both and if therapeutic inhibition of periostin improves PH/PAH by RV-directed, as well as pulmonary vascular, mechanisms.

In summary, the study by Nie et al represents the most extensive characterization of periostin in PH to date, revealing mechanisms of periostin and HIF1α signaling with pleiotropic effects on various pulmonary vascular cell lineages. These important and novel findings demonstrate that periostin is a key integrator of hypoxia-mediated pulmonary vascular remodeling. The potential use of periostin as a biomarker of PH deserves further investigation in prospective cohorts, and development of novel pharmacological tools targeting periostin are certainly warranted.