Cellular Pathways Promoting Pulmonary Vascular Remodeling by Hypoxia

Abstract

Exposure to hypoxia increases pulmonary vascular resistance, leading to elevated pulmonary arterial pressure and, potentially, right heart failure. Vascular remodeling is an important contributor to the increased pulmonary vascular resistance. Hyperproliferation of smooth muscle, endothelial cells, and fibroblasts, and deposition of extracellular matrix lead to increased wall thickness, extension of muscle into normally non-muscular arterioles, and vascular stiffening. This review highlights intrinsic and extrinsic modulators contributing to the remodeling process.

Introduction

In the normal adult lung, the pulmonary circulation is a low-resistance vasculature with thin-walled vessels. For over 100 years, it has been appreciated that acute challenge with hypoxia causes pulmonary vessels to contract, a phenomenon termed hypoxic pulmonary vasoconstriction (reviewed in Ref. 155), which optimizes gas exchange by diverting perfusion from areas of inadequate ventilation. Decades later, the migration of low-altitude dwelling individuals into high-altitude regions uncovered the deleterious effects of prolonged hypoxia on the pulmonary vasculature (22a, 146). It is now recognized that long-term exposure to hypoxia, resulting from residence at high altitude or chronic lung disease, causes sustained contraction and vascular remodeling in the lung, leading to increased pulmonary arterial pressure (Ppa) and pulmonary vascular resistance (PVR), subsequent right heart enlargement, and, if severe enough, right ventricular failure and death (4, 8, 22a, 53, 101, 112, 113, 118, 152). Given the important contribution of pulmonary vascular remodeling to the increase in PVR observed with chronic hypoxia (CH), substantial efforts have been invested in understanding the cellular basis for these changes.

Hypoxia-Induced Changes in the Pulmonary Vascular Wall

Following the initial findings of altitude-induced pulmonary dysfunction in the early 1960s, research focused on characterizing the effects of CH on pulmonary vascular structure. Histological analysis of postmortem or postoperative tissue specimens from native high-altitude populations in South America, Tibet, and the U.S. revealed that long-term exposure to hypoxia was associated with thickening of the walls of the small arteries due to pulmonary arterial smooth muscle cell (PASMC) hyperplasia and muscle extension into precapillary arterioles (8, 53, 101, 112, 152). These human anatomic data provided compelling evidence of vascular remodeling in response to hypoxia. Subsequent studies in a variety of species confirmed and extended our understanding of the effect of long-term hypoxia on the structure of the pulmonary circulation (159). For example, experiments conducted in the high mountains of Colorado identified altitude as the primary cause of “brisket disease,” or heart failure, in cattle (176). Susceptible cattle exhibited extensive vascular remodeling, characterized by collagen deposition, thickened adventitia, increased intimal and medial thickness, and extension of muscle in small pulmonary arteries (126, 149, 151), which was particularly robust in neonatal calves (148, 149). Like humans or cattle experiencing hypoxia due to altitude, pigs also develop significant remodeling with altitude exposure (reviewed in Ref. 126).

Although early studies used large animals in a high-altitude setting, in the laboratory, just a few weeks in an experimental hypoxic environment (typically 10% O₂) predictably and reproducibly increased vascular wall thickness in rats due to PASMC hypertrophy, and hyperplasia and muscularization of typically non-muscular arterioles (30, 93, 124). In contrast, wall thickening is minimal in mice, and remodeling more often takes the form of distal muscularization (reviewed in Ref. 151) and functional stiffening of proximal, conduit arteries (171). Some species, such as llamas, coati, and yaks, are completely protected from hypoxia-induced vascular remodeling (35, 51, 52). The explanation for species-specific differences in degree of remodeling (FIGURE 1) is still under investigation, although with respect to cattle, genetic mutations that confer susceptibility and/or the greater presence of PASMCs in peripheral arteries of normotensive animals may underlie some of these differences (64, 106).

FIGURE 1.

Relative effect of hypoxia on pulmonary vascular remodeling in various species

Hyperproliferation of PASMCs and fibroblasts can account for medial and adventitial thickening, respectively; however, extension of muscle along the vascular tree likely includes a component of cell migration as well. Supporting this concept, cells exposed to hypoxia in vitro or isolated from CH animals are hypermotile (39, 73, 163, 178). Elegant fate-mapping experiments indicate that the cells surrounding the small arterioles following exposure to hypoxia derive from a smooth muscle lineage and are suggestive of cell migration followed by clonal expansion (135, 136). Unfortunately, the lack of specific markers for migration and inability to track single cells over time in vivo create practical difficulties in measuring the exact contribution of cell migration to the remodeling process.

In addition to wall thickening resulting from cellular hyperproliferation, enhanced deposition and/or cross-linking of extracellular matrix components (i.e., collagen) causes vascular stiffening of the large, conduit pulmonary arteries (34, 105, 170, 171). Reduced distensibility of the proximal vasculature contributes to increased Ppa via loss of right ventricle-pulmonary artery coupling, increased right ventricular afterload, and increased pulsatility of flow in the distal vasculature (34, 156).

Given the substantial CH-induced structural changes observed in the pulmonary circulation (FIGURE 2) and that conventional vasodilators had little acute effect on Ppa (10, 19, 97, 108), it was widely regarded that “fixed” remodeling was the underlying cause of increased PVR. Indeed, imaging studies in chronically hypoxic lungs often used vascular infusion of agents to visualize vessels and appeared to show reduced luminal diameter and decreased numbers of small-diameter vessels, or vascular pruning, suggesting remodeling occurred in an inward fashion to occlude vessels (55, 123). However, it is possible that these observations may have reflected robust vasoconstriction or variable perfusion (12, 91, 150), since careful morphological examination in fully vasodilated lungs revealed medial expansion was directed outward and was associated with angiogenesis (58). Later studies demonstrating that inhibitors of Rho kinase (ROCK), a cytoskeletal regulator that promotes contraction, acutely normalized Ppa in chronically hypoxic rats (40, 102, 103) suggested that constriction was not permanent and that remodeling instead may augment contraction secondary to increased muscularity of the small pulmonary arteries.

FIGURE 2.

Effects of chronic hypoxia on the pulmonary arterial wall

Under normoxic conditions, main and resistance level arteries consist of a single layer of endothelial cells, and thin layers of smooth muscle and fibroblasts. Exposure to prolonged hypoxia causes vessels to be exposed to elevated levels of local and circulating growth factors, and increased mechanical forces such as shear stress. Hypoxia also stimulates production of bone morphogenetic proteins (BMPs) and activates hypoxia-inducible factors (HIFs). The resulting changes in the vascular wall include proliferation of fibroblasts and smooth muscle cells, smooth muscle extension into non-muscular arterioles, increased deposition and cross-linking of extracellular matrix components (i.e., collagen), and angiogenesis.

Endothelial cells (ECs) exposed to in vitro hypoxia, especially severe levels (i.e., ≤1% O₂), exhibit increased proliferation (50, 158, 178). However, since intimal thickening is rarely observed with CH outside of some cattle, the relevance of this response to vascular remodeling and development of hypoxia-induced pulmonary hypertension in most species is unclear. One possibility is that microvascular EC proliferation in response to hypoxia may contribute to angiogenesis (58, 114, 158), perhaps as a compensatory mechanism to increase gas exchange.

Extrinsic Factors That Promote Pulmonary Vascular Remodeling

The mechanisms underlying smooth muscle and fibroblast hyperproliferation include both intrinsic and extrinsic factors. As noted earlier, one of the first pulmonary responses to hypoxia is vasoconstriction. Reduced diameter increases the mechanical forces exerted on the pulmonary vascular wall, both perpendicular strain from elevated pressure and increased shear stress (or drag) along the endothelium. A role for altered hemodynamic stress influencing remodeling is supported by the fact that augmenting pulmonary blood flow, alone or in combination with hypoxia, increased remodeling (125). Moreover, supernumerary vessels, small-diameter arteries that branch from the parent artery at a 90° angle and contain strictures at the opening that protect against high hemodynamic stress, do not exhibit wall thickening in chronically hypoxic rats (109).

As the interface with the bloodstream, ECs serve a critical barrier function and are ideally positioned to respond to changes in mechanical forces (i.e., shear stress). Under normal conditions, excess endothelial production of anti-proliferative vasodilators, such as nitric oxide and prostacyclin, maintain low PVR. With hypoxia and increased shear stress, levels of endothelial-derived pro-proliferative circulating factors (i.e., endothelin-1) are increased (134, 142), whereas synthesis of anti-proliferative factors is reduced (reviewed in Ref. 148). When applied in vitro, ET-1 induces migration and proliferation of PASMCs (60, 92, 134), and proliferation and collagen production in lung fibroblasts (3). In vivo, ET receptor inhibitors prevented vascular remodeling in several animal models (29). Platelets can also respond to changes in shear stress with release of serotonin (5-HT), possibly accounting for elevated circulating levels observed in CH (5, 22). Similar to ET-1, 5-HT uptake promotes PASMC growth and vascular remodeling (37, 84, 85). Some of these responses can be attributed to internalization of 5-HT via the serotonin transporter (SERT) and receptor-independent signaling (85, 87), although inhibiting 5-HT_1B receptors by silencing or pharmacological inhibitors reduced PASMC proliferation and CH-induced remodeling (85), suggesting potential cooperativity between uptake and receptor signaling.

A major signaling pathway regulating PASMC growth involves bone morphogenetic protein (BMP) signaling. BMPs are part of the transforming growth factor-β (TGF-β) super family and bind to complexes containing type I and type II receptors (145) to activate SMAD1/5/8 signaling and downstream transcriptional responses (79). BMPs also induce SMAD-independent signaling, involving MAP kinases (MAPKs), phosphatidylinositol 3-kinase/AKT, and protein kinase C (179). Most work examining the role of BMPs in pulmonary vascular remodeling has focused on ligand (BMP2, BMP4, and BMP7) and receptor (BMPR1 and BMPR2) binding. Importantly, BMPs exert differential effects on PASMC growth depending on the location within the vasculature; BMP4 was anti-proliferative in PASMCs from proximal vessels (99) but induced proliferation (180) in PASMCs from distal vessels. Mice with partial deficiency for BMP2 and BMP4 were more susceptible to and protected against hypoxia-induced vascular remodeling, respectively (6, 44). With respect to lung fibroblasts, BMP4 may play a protective role by antagonizing TGF-β-induced proliferation and collagen production (111), whereas, in ECs, activation of BMPR2 by BMP2 or BMP4 was linked to nitric oxide production (45), providing a paracrine effect whereby normal BMP signaling exerts anti-proliferative effects on PASMCs and/or fibroblasts.

Infiltrating immune cells provide another source of paracrine factors that can promote vascular remodeling. Recruited by pro-inflammatory signals from adventitial fibroblasts (38, 75), macrophage accumulation around pulmonary vessels in chronically hypoxic animals in turn stimulates fibroblast growth, collagen production, and secretion of growth factors that promote smooth muscle proliferation (75, 119, 129).

In both PASMCs and fibroblasts, hypoxia induces the expression and activity of lysyl oxidases (Loxs) (105), enzymes that oxidize lysine residues in elastin and collagen to promote formation and repair of the extracellular matrix. Upregulation of Loxs contributes to vascular stiffness via promoting extracellular matrix deposition and collagen cross-linking. Recently, galectin-3, a β-galactoside-binding lectin, was found to be induced in the hypoxic lung and to contribute to collagen production from adventitial fibroblasts (83). Increased deposition and/or cross-linking of extracellular matrix components not only increases vascular stiffness but may also induce intracellular signaling that promotes growth/motility via matrix-cell interactions.

Hypoxia Alters Expression and Activity of Membrane Channels/Transporters

A variety of intrinsic cellular mechanisms are activated by hypoxia to promote a pro-proliferative, pro-migratory phenotype in PASMCs and/or fibroblasts. One of the earliest changes noted in PASMCs from chronically hypoxic animals was membrane depolarization (154), subsequently identified to be due to repressed expression/activity of several types of K⁺ channels (94, 140, 166, 169). Augmenting PASMC K⁺ channel expression/activity reduced remodeling in chronically hypoxic rats (16, 94, 116), but the exact mechanism by which depressed K⁺ channel activity and/or depolarization causes remodeling is still being explored. One possibility is that increased intracellular K⁺ confers resistance to apoptosis (21). Another possibility is that depolarization might drive activation of voltage-gated calcium channels (VGCC) in the L-type channel family, leading to elevated basal intracellular calcium concentration ([Ca²⁺]_i). Consistent with this hypothesis, increased [Ca²⁺]_i is required for PASMC growth and migration (47, 72, 73) and has been documented in cells from hypoxic animals (143, 168). However, inhibitors of L-type channels have little effect on PASMCs from hypoxic animals (76, 143, 168). Rather, the elevated basal [Ca²⁺]_i appears to occur primarily via upregulation of canonical transient receptor potential (TRPC) proteins (76, 168), which form Ca²⁺-permeable nonselective cation channels that are not activated by depolarization but can be modulated by phosphorylation, receptor activation, or store depletion (reviewed in Ref. 36). Nonselective cation channels are required for elevated [Ca²⁺]_i (73, 76, 168), migration (73), and proliferation (88) in PASMCs exposed to hypoxia in vitro or isolated from CH animals. Interestingly, VGCCs in hypoxic PASMCs can be activated by mitogenic agonists (82, 164) and contribute to stimulated proliferation (128), suggesting a potential role in the setting of excessive growth factors. Whether [Ca²⁺]_i is increased in hypoxic adventitial fibroblasts is unknown, but hypoxia-induced growth requires activation of Ca²⁺-dependent PKC isoforms (28), suggesting elevated [Ca²⁺]_i might also be a feature of hypoxic exposure in these cells.

Following a rise in [Ca²⁺]_i, several downstream signal transduction pathways and transcription factors are activated that could be involved in hypoxia-induced proliferation (reviewed in Ref. 70) and/or migration (73, 178). In particular, Ca²⁺ activates nuclear factor of activated T cells (NFAT), which in turn reduces K⁺ channel expression and increases proliferation (18), providing a link between alterations in [Ca²⁺]_i, dysregulated K⁺ channel expression/activity, and PASMC growth. Accordingly, remodeling was reversed when NFAT was pharmacologically inhibited (18) or genetically deleted (13). Moreover, Ca²⁺-dependent increases in the expression of the water channel, aquaporin 1 (AQP1), are required for hypoxia-induced migration of PASMCs (73). In endothelial and tumor cells, increased AQP1 levels may result in localized control of water flux across the cell membrane, possibly allowing for directed cell movement (131); however, in PASMCs, the actions of AQP1 appear to be independent of water transport (71) and instead require the COOH-terminal tail portion of the protein, which regulates the levels of β-catenin, a dual-function protein, to control both migration and proliferation (185). AQP1 may also participate in cytoskeletal rearrangement (98) to facilitate cell movement.

Another contributor to control of migration and proliferation in PASMCs is the Na⁺/H⁺ exchanger (NHE), a main contributor to pH homeostasis (86, 122). NHE isoform 1 (NHE1) is upregulated in PASMCs by hypoxia, leading to increased NHE activity and an alkaline shift in pH (127, 139). Activation of NHE is required for growth factor-induced proliferation (120), whereas loss of NHE activity (121) or genetic deletion of NHE1 (183, 184) attenuated hypoxia-induced vascular remodeling and PASMC proliferation and migration. NHE1 activation represses a growth inhibitor pathway by increasing p27, a cyclin-dependent kinase inhibitor, while simultaneously stimulating proliferation by decreasing the nuclear transcription factor E2F1 (183). NHE1 may also facilitate cell growth and migration via regulation of cytoskeletal arrangement (31); in fibroblasts, NHE1 binds to actin filaments through the adaptor protein, ezrin, providing a link between NHE1 and cytoskeletal rearrangement (31). Preliminary studies indicate similar NHE1-actin interactions occur in PASMCs (90). Hypoxia increased NHE1 expression in ECs (26), but whether NHE1 contributes to hypoxia-induced EC migration, proliferation, or angiogenesis is unknown.

Reactive Oxygen Species and Altered Metabolism

Changes in mitochondrial-derived reactive oxygen species (ROS) are well documented in response to hypoxia and are essential for initiation of vasoconstriction (reviewed in Ref. 155); thus it is not unreasonable to suspect that mitochondrial ROS might also play a role in the remodeling process. Indeed, mitochondrial fission, which is typically associated with increased ROS production from mitochondria (177), has been reported in PASMCs from CH rats, with inhibitors of fission attenuating remodeling (89). However, observations that ROS levels were reduced and MitoQ, a mitochondrial-targeted antioxidant, had no effect on hypoxia-induced remodeling (110) would appear to argue against a role for mitochondrial ROS in CH-induced remodeling. Cytosolic ROS are also generated by NADPH oxidase (Nox). Nox1 (25, 161), Nox2 (7, 77, 100), and Nox4 (11, 48, 96) are expressed in pulmonary vascular tissue and have been studied with respect to hypoxia-induced remodeling. For example, female mice with genetic deficiency for Nox1 were protected from the development of CH-induced pulmonary vascular remodeling (57), whereas male mice instead developed PH in the absence of hypoxia, highlighting the importance of sex hormones in the pulmonary vasculature. Indeed, Nox1 mediates the effects of estrogen and serotonin on PASMC proliferation (56). Deficiency of the Nox2 subunit, gp91^phox, also protected against medial wall thickening in CH mice (77). Of note, since Nox2 is highly expressed in immune cells and these studies were conducted in global knockout animals, whether Nox2 activity in phagocytic cells, vascular cells, or both is necessary for hypoxia-induced vascular remodeling is unclear. Abundant data demonstrate hypoxia increases Nox4 expression in PASMCs, pulmonary arteries, or lungs from CH mice and rats (48, 50, 96, 107, 172), where Nox4 activation is necessary for hypoxia-induced reduction of K_v currents (95) and mediates BMP4-induced increases in protein expression of nonselective cation channels (66). Nox4 activation increases H₂O₂ production and cyclin D1 expression, a factor important in the control of cell cycle (173), and enhances proliferation by inhibiting peroxisome proliferator-activated receptor gamma (PPARγ) (80, 81). Since Nox4 is repressed by PPARγ (49), downregulation of PPARγ sets up a feed-forward mechanism to further enhance Nox4 activity. NF-κB has been identified as both an upstream regulator (14, 81, 173) and a downstream effector (15, 172) of Nox4, suggesting an additional NF-κB/Nox4 feed-forward loop.

Despite substantial evidence implicating Nox4 interactions with several pathways critical to pulmonary vascular remodeling during hypoxia, results from in vivo Nox4 inhibition have proved equivocal. GKT137831, which inhibits both Nox4 and Nox1, attenuated hypoxia-induced wall thickening but not muscularization of small arteries (50), yet Nox4 deficiency had no effect on pulmonary vascular remodeling in CH mice (57, 162). Whether these data suggest that Nox4 is not necessary for CH-induced pulmonary vascular remodeling in vivo or plays a less important role in mice than other species remains to be determined.

Signaling Pathways That Form Nodes of Interaction

Many of the pathways outlined above are linked through regulation by the oxygen-sensitive transcription factors, hypoxia-inducible factor 1 and 2 (HIF-1 and HIF-2, respectively). HIF-1 was originally identified as a heterodimeric transcription factor consisting of a constitutively expressed β (HIF-1β) subunit and an oxygen-sensitive α (HIF-1α) subunit (132), with subsequent studies identifying HIF-2α as structurally similar to HIF-1α and also binding HIF-1β to form the HIF-2 transcription factor (117). Both α subunits are hydroxylated on two proline residues via prolyl hydroxyalse domain (PHD) proteins using molecular oxygen as a substrate, allowing binding to the von Hippel-Lindau protein, ubiquitination, and proteasomal degradation (reviewed in Refs. 133, 117). Reductions in oxygen levels limit hydroxylation, allowing HIF-1α/HIF-2α accumulation.

HIF-1α is found in all cells, whereas HIF-2α expression is more restricted. In the lung, hypoxia induces HIF-1 (39, 41, 68, 69, 115, 157, 181) and HIF-2 (39, 69, 157) in PASMCs, fibroblasts, and ECs. Initial evidence that HIFs contribute to pulmonary responses to CH came from transgenic animals, where complete loss of HIF-1α or HIF-2α was lethal (23, 63), whereas heterozygosity for the null alleles (Hif1a^+/− and Hif2a^+/−) impaired development of hypoxia-induced PH and vascular remodeling (20, 182). Targeting HIF-1 (2) or HIF-2 (27, 59) activity pharmacologically or with knockdown approaches also reduced CH-induced vascular remodeling. Use of inducible Cre recombinase strategies to selectively delete HIFs in adult animals revealed that complete loss of smooth muscle Hif1a (9) or endothelial Hif2a (157), or partial global loss of Hif2a (59), attenuated hypoxia-induced vascular remodeling. Similar benefits were achieved with constitutive HIF-2α homozygous deletion targeted to ECs (67, 147). Consistent with these findings, HIF-1 mediates hypoxia-induced PASMC proliferation (138) and HIF-2 mediates both migration and proliferation in fibroblasts (39) in vitro. Surprisingly, mice with constitutive homozygous deletion of HIF-1α targeted to vascular smooth muscle cells (68) or complete global (59) or EC-targeted deletion of HIF-2α (147) exhibited enhanced PH or death during hypoxia, suggesting that, in these mice, HIFs played a beneficial role. Inconsistency in results across these studies highlight the complexity of HIF signaling and indicate further investigation will be required to determine whether homozygous versus heterozygous genetic modifications, long-term (constitutive) versus short-term (inducible) deletions, off-target effects of the drivers used, and/or sex- or strain-dependent differences account for the variability observed.

The preponderance of evidence appears to indicate that HIFs mediate pulmonary vascular remodeling during CH, although the downstream mechanisms are still being defined. In the case of HIF-1, both Ca²⁺ and pH homeostasis appear to be involved (139, 141). HIF-1 also regulates other factors known to promote hypoxia-induced remodeling, including the expression of K⁺ channels (17, 140, 175), BMP4 (165), mitochondrial fission (89), and vascular endothelial-derived growth factor (43). Moreover, HIF-1 binds to the Nox4 promoter, leading to Nox4-dependent increases in proliferation and migration (33). Nox4 upregulation further increases HIF-1α transcription via activation of NF-κB (15) and increases HIF activation by interfering with hydroxylation (32), creating a feed-forward loop to augment and/or maintain HIF activity.

Plasma levels of ET-1, a well-known HIF target, were reduced in Hif2a^+/− mice (20). Although HIFs induce ET-1 production, ET-1 in turn upregulates HIF-1 selectively in PASMCs by downregulating PHD2 expression (74, 115), creating feed-forward enhancement of HIF-1 not observed in systemic smooth muscle cells (115). Since ECs serve as a primary source of ET-1 in vivo, these data suggest a possible model whereby ET-1 production enhanced by hypoxia-induced HIF-2α expression in ECs subsequently augments HIF-1α in PASMCs, initiating a mechanism for upregulation of HIF-1 and, consequently, HIF target genes to promote PASMC remodeling. HIF-2 in ECs also upregulates arginase 1 (24), an enzyme responsible for converting the NO precursor L-arginine to L-citrulline, perhaps explaining the reduced bioavailability of NO during hypoxia. Arginase 2 is likewise upregulated by hypoxia in pulmonary ECs (158), but whether this response is HIF-dependent is unclear. Finally, the effects of hypoxia on production of pro-inflammatory mediators and growth factors in fibroblasts are also HIF-dependent and serve to both recruit macrophages to the vessels and stimulate fibroblast, PASMC, and possibly EC proliferation (38, 129).

Rho-associated protein kinase (ROCK) is another signaling molecule interconnecting several pathways. A major consequence of ROCK activation is enhancement of the Ca²⁺ sensitivity of the contractile apparatus (65, 174) to render pulmonary arteries more sensitive to vasoconstrictor stimuli; however, ROCK is also implicated in the remodeling process (42, 137). For example, ROCK activation is necessary for migration and proliferation in a variety of vascular cell types, including PASMCs (46, 60, 78, 137, 163). Upstream activators of ROCK, RhoA and RhoB, mediate cytoskeletal rearrangement, a process necessary for cell movement, and contribute to hypoxia-induced migration and proliferation (178). Consistent with these findings, treatment with ROCK inhibitors reduced neovascularization and remodeling in hypoxia models (1, 40, 62). Activated by various growth factors (144) and by ROS (65), ROCK may thus represent a node of convergence for numerous inputs to augment opening of Ca²⁺ channels (54, 82, 167) and activity of NHE1 (60, 160).

Last, mammalian target of rapamycin (mTORC) is required for proliferation induced by both in vivo and in vitro hypoxic exposure (69). mTORC2 is activated in CH cells and is a critical regulator of PASMC proliferation and survival via downregulation of AMP-activated protein kinase (AMPK) (48). In PASMCs where HIF-1α levels are increased, depletion of mTORC2 reduced HIF-1α protein, suggesting cross-regulation between these pathways (48). Finally, increased Nox4 expression is required for mTORC2 activation (48), placing Nox4 upstream of another central coordinator of PASMC function.

Future Directions

The aforementioned studies confirm that pulmonary vascular remodeling in response to hypoxia is a multifaceted process (FIGURE 3). Although significant progress has been made in the past century to characterize the structural changes and underlying cellular mechanisms induced by hypoxia, much is yet to be learned in unraveling the complicated process controlling pulmonary vascular cell proliferation and migration. Adding complexity is apparent cross-talk between critical signaling pathways and conflicting data regarding the precise roles HIFs, EC proliferation, and ROS production play in vivo. Many of the differences observed across studies and between in vivo and in vitro results may stem from limitations in current experimental approaches. For example, considerable heterogeneity exists within cell types isolated from different portions of the pulmonary circulation, yielding variable responses to the same stimuli. Thus use of cells from proximal vessels may be applicable for exploring the mechanisms underlying arterial stiffening, whereas cells from distal regions of the lung may be more appropriate for examining mechanisms of migration and/or proliferation. In addition, cells in culture may not entirely reflect the in vivo situation for a variety of reasons, with traditional cell-culture experiments conducted with cells in isolation, on hard cultureware, and under static conditions. In vivo, cell-cell communication, a flexible extracellular matrix, and constant exposure to mechanical forces may elicit very different responses to stimuli. Recent advances with microfluidics (104, 153) and “lung-on-a chip” technology (61) may provide improved approaches for studying interactions between cells under conditions that more faithfully simulate the in vivo environment. And finally, even as new factors are identified for targeting as treatments for pulmonary vascular remodeling, several important obstacles remain in translating this information to therapies. Many of the molecules and signaling pathways recognized as being involved in the remodeling process have other important physiological functions. Thus clinical trials will need to be performed to assess the extent of their benefit in the patient population compared with unwanted side effects. Potentially, use of inhibitors for any of the described pathways to target pulmonary vascular remodeling may only be realized if cell-specific targeting can be achieved. Thus further investigation is clearly needed to identify additional mechanisms involved in the pathogenesis of vascular remodeling and to develop new pharmacological agents aimed at precisely targeting molecules in particular cell populations to reverse the remodeling process.

FIGURE 3.

HIF-dependent pathways involved in pulmonary vascular remodeling in response to chronic hypoxia

HIFs have been shown to increase under hypoxic conditions in endothelial and smooth muscle cells and fibroblasts. In endothelial cells (ECs), HIF-2 appears to be the predominant HIF upregulated by hypoxia, leading to upregulation of arginase, and subsequent reduction in nitric oxide (NO) production, and endothelin-1 (ET-1). Both hypoxia and ET-1 can augment HIF-1 levels in pulmonary arterial smooth muscle cells (PASMCs). Downstream consequences of HIF-1 activation in PASMCs include upregulation of Na⁺/H⁺ exchanger causing an alkaline shift in pH, induction of NADPH oxidase, bone morphogenetic protein 4 (BMP4) and ET-1, increased mitochondrial (mito) fragmentation and subsequent reduction in voltage-gated K⁺ channel expression, and upregulation of Ca²⁺-permeable channels, which facilitate accumulation of aquaporin 1 (AQP1), a water channel that in turn upregulates the pro-growth transcription factor β-catenin. PASMC-derived ET-1 can feed back to augment HIF-1 levels. In fibroblasts, both HIF-1 and HIF-2 levels increase, leading to production of pro-inflammatory cytokines that recruit immune cells and enhanced growth factors that can stimulate EC and PASMC proliferation.