Sensors and signals: the role of reactive oxygen species in hypoxic pulmonary vasoconstriction

Abstract

When lung cells experience hypoxia, the functional response, termed hypoxic pulmonary vasoconstriction, activates a multitude of pathways with the goal of optimizing gas exchange. While previously controversial, overwhelming evidence now suggests that increased reactive oxygen species – produced at complex III of the mitochondrial electron transport chain and released into the intermembrane space – is the cellular oxygen signal responsible for triggering hypoxic pulmonary vasoconstriction. The increased reactive oxygen species (ROS) activate many downstream targets that ultimately lead to increased intracellular ionized calcium concentration and contraction of pulmonary arterial smooth muscle cells. While the specific targets of ROS signals are not completely understood, it is clear that this signalling pathway is critical for development and for normal lung function in newborns and adults.

The smooth muscle cells of pulmonary arteries contract during acute hypoxia in a response called hypoxic pulmonary vasoconstriction (HPV). It is likely that this process evolved in mammals from the need to maintain fetal oxygenation during gestation. During fetal development, the lung is filled with hypoxic amniotic fluid. Gas exchange occurs in the placenta, so the blood returning to the fetal heart is relatively well oxygenated. Perfusion of the hypoxic lung with blood that has a high $P_{{\mathrm{O}}_2}$ would result in loss of O₂ from pulmonary capillary blood to the amniotic fluid. HPV augments pulmonary vascular resistance to a point where it exceeds systemic vascular resistance. This limits blood flow to the lungs and diverts most of the output of the right ventricle through the ductus arteriosus to the aorta. At birth, the initiation of lung ventilation results in a rapid increase in alveolar $P_{{\mathrm{O}}_2}$. The vascular cells of the distal pulmonary arteries detect the change in O₂ tension and trigger relaxation of the pulmonary arteries, while the increase in $P_{{\mathrm{O}}_2}$ in blood triggers contraction of smooth muscle cells in the ductus arteriosus. These responses enable the shift from placental to lung respiration at birth. O₂ sensing by lung and vascular cells is therefore required for that transition.

In the adult lung, smooth muscle cells retain the ability to constrict in response to hypoxia. HPV can help to optimize pulmonary gas exchange by matching pulmonary perfusion to ventilation when hypoxic regions in the lung develop. Pulmonary smooth muscle cells sense a decrease in alveolar O₂ tension and decrease local perfusion, thereby redirecting blood flow to other regions with higher alveolar $P_{{\mathrm{O}}_2}$. HPV, however, plays a minimal role in normal healthy lungs where basal pulmonary vascular tone is low. Few regions of hypoxia exist in the normal lung, such that application of pulmonary vasodilators does not alter the efficiency of alveolar ventilation–perfusion matching or arterial $P_{{\mathrm{O}}_2}$. HPV may play a more important role in matching pulmonary perfusion to ventilation in diseases such as asthma and acute respiratory distress syndrome where regional differences in alveolar $P_{{\mathrm{O}}_2}$ can develop. However, HPV can also be detrimental in cases of generalized alveolar hypoxia when the HPV response can cause vasoconstriction throughout the lung, resulting in pulmonary hypertension. This can occur in healthy lungs at high altitude, as well as in diseases that produce large areas of alveolar hypoxia such as chronic obstructive pulmonary disease, sleep apnoea and pulmonary fibrosis.

Acute HPV and chronic hypoxia‐induced vascular remodelling

The hypoxic responses of pulmonary vessels have been studied in a wide variety of species and experimental conditions (Sylvester et al. 2012). It has been established that exposure of pulmonary arteries to hypoxia results in a biphasic vasoconstrictor response (Bennie et al. 1991; Leach et al. 1994, 2001; Weissmann et al. 2001, 2004; Sylvester et al. 2012). Phase 1 of the pressor response involves an early vasoconstriction corresponding to acute HPV, while phase 2 represents a subsequent sustained and progressive increase in pulmonary arterial pressure (Weissmann et al. 2001, 2004). Phase 1 of HPV generally begins within seconds of the hypoxic exposure and peaks within 5 min, while phase 2 develops gradually, reaching its sustained maximum at 30–60 min (Weissmann et al. 2001; Sylvester et al. 2012). The degree of constriction during phase 1 can be quite variable, but several studies demonstrated an increase in vascular resistance by 10–300% over basal levels (Rodman et al. 1989; Sylvester et al. 2012). Many studies have demonstrated that the amplitude of the phase 2 contraction is smaller than in phase 1, but other studies suggest that the phase 2 contraction could equal or even surpass that of phase 1 (Sylvester et al. 2012). Investigators simultaneously measuring tension and intracellular Ca²⁺ concentration ([Ca²⁺]_i) have demonstrated that phase 1 of HPV is associated with an increase in [Ca²⁺]_i, whereas the tension increase in phase 2 occurs without a further increase in [Ca²⁺]_i (Robertson et al. 1995). These data indicate that Ca²⁺ sensitization is likely involved in phase 2 of HPV. Experimental evidence has shown that prolonged hypoxia can cause an elevation in pulmonary arterial pressure that persists after return to normoxia (Weissmann et al. 2001), which is important from a clinical standpoint because lung diseases associated with chronic hypoxia lead to prolonged elevation of pulmonary arterial pressures that are refractory to increases in the fraction of inspired O₂. In addition to changes in Ca²⁺ sensitivity, chronic hypoxia leads to downregulation of K⁺ channel expression and upregulation of transient receptor potential canonical (TRPC) ion channel protein expression, which have been interpreted as being associated with store‐operated Ca²⁺ channel (SOCC) activity (Wang et al. 1997, 2006). These increases in [Ca²⁺]_i lead to pulmonary arterial smooth muscle cell (PASMC) contraction, proliferation and pulmonary vascular remodelling. Chronic hypoxia and sustained HPV can therefore lead to increased pulmonary vascular resistance, pulmonary vascular remodelling and increased right heart afterload.

Experimental studies using isolated PASMC and isolated pulmonary vascular rings denuded of endothelium have demonstrated that the phase 1 constriction response is intrinsic to smooth muscle cells (Murray et al. 1990; Demiryurek et al. 1991; Madden et al. 1992; Sham et al. 2000). While these studies indicate a lack of need for circulating mediators, studies using pulmonary vascular rings with intact endothelium or precision‐cut lung slices demonstrate an enhanced response to hypoxia (Murray et al. 1990; Leach et al. 1994; Paddenberg et al. 2006), which suggests that phase 2 hypoxic contraction requires an intact endothelium (Sylvester et al. 2012). Studies in several species have demonstrated that endothelin‐1 (ET‐1) released from endothelial cells enhances the hypoxic reactivity of PASMCs (Oparil et al. 1995; Liu et al. 2001). Indeed, treatment with ET‐1 antagonists inhibits HPV in intact animals suggesting that release of ET‐1 may be required to achieve a full HPV response (Oparil et al. 1995; Liu et al. 2001). Additionally, production of nitric oxide (NO), which promotes vasodilatation, by endothelial cells is decreased in hypoxia (Le Cras & McMurtry, 2001). Together, these data suggest that pulmonary arterial endothelial cells modulate the response of PASMC via the release of vasoactive factors to complement the overall functional response to changes in O₂ tension. Indeed, hypoxia also increased endothelial monolayer permeability allowing vasoactive factors to reach PASMC (Partridge, 1995; Yang et al. 2016a). Thus, O₂ sensing mechanisms in multiple vascular cell types contribute to the activation of diverse responses that collectively define the responses to acute and chronic hypoxia (Gao et al. 2016).

Characteristics of an O₂ sensor

Since its discovery by von Euler and Liljestrand in 1946, HPV has been studied by many investigators and its relevance to health and disease has been defined. The exact mechanism by which cells in the lung detect a decrease in $P_{{\mathrm{O}}_2}$ and translate that into a biological response is still not fully established. Identifying how lung cells sense changes in O₂ will enhance our understanding of the physiology and pathophysiology of pulmonary circulation and offer insight into how other tissues sense and respond to changes in $P_{{\mathrm{O}}_2}$. In light of this, the use of HPV as a model provides a unique opportunity to study and identify the underlying O₂ sensor(s).

Complex organisms cannot rely on diffusion alone to supply cells with molecular O₂ and nutrients. Evolution from single celled organism to metazoan species was associated with the development of elegant and elaborate systems to ensure adequate delivery of O₂ to meet metabolic demand. Sudden changes in metabolic demand in multicellular organisms require a dynamic system with the ability to respond rapidly (López‐Barneo et al. 2001). Responses to hypoxia can be generalized into two categories. Post‐translational responses limit metabolic energy consuming processes, such as protein synthesis, and enhance glycolysis by translocating glucose transport proteins to the cell membrane. Transcriptional responses involve de novo expression of glycolytic enzymes, glucose transporters and genes that enhance the ability of cells to maintain ATP production in the absence of O₂. De novo gene expression requires transcription, translation and often post‐translational modification. Completion of these processes takes a significant amount of time. In general, a response to hypoxia requiring cell proliferation or remodelling is not likely to protect the organism from acute hypoxic stress that can develop in a matter of seconds. Additionally, post‐translational modifications that allow for rapid translocation of glucose transporters to the cell surface are not an optimal response to chronic hypoxia where glucose levels may become depleted (Zhang et al. 2015). Therefore, O₂ sensors must be able to respond accordingly to both acute and chronic hypoxic stress in order to coordinate responses with differing time constraints (Schumacker, 2014). Indeed, the ability of these sensors to detect early initial decreases in $P_{{\mathrm{O}}_2}$ and to trigger adaptive responses quickly is essential to prevent cellular injury and to lessen the decline in $P_{{\mathrm{O}}_2}$.

Mitochondria as O₂ sensors

Mitochondria are responsible for the vast majority of the O₂ consumed by the cell. Therefore, they represent an appealing site for an O₂ sensor. Given that a major role of mitochondria is the production of ATP, it is conceivable that decreased ATP production due to decreased oxidative phosphorylation may be the signal that initiates the hypoxic response. However, studies in isolated cells as well as intact lungs have demonstrated that hypoxia responses are initiated at $P_{{\mathrm{O}}_2}$ values ranging from 25 to 75 mmHg (Murray et al. 1990; Madden et al. 1992). However, mitochondrial respiration can continue normally even at very low O₂ levels, and only begins to become O₂ supply‐limited at $P_{{\mathrm{O}}_2}$ values less than ∼7 mmHg (Chandel et al. 1997). Given that mitochondrial respiration can continue normally at such low O₂ levels without a decrease in ATP production, it seems unlikely that ATP concentration could trigger cellular responses to hypoxia (Buescher et al. 1991). If decreases in ATP are not the trigger for hypoxic responses, another possibility is that O₂‐dependent production of reactive oxygen species (ROS) by mitochondria is the initiating signal for the hypoxic response.

ROS as signalling molecules

Mammalian cells have developed many signalling systems to relay information including post‐translational modifications such as phosphorylation, acetylation and ubiquitinylation, as well as signalling systems regulated by Ca²⁺. ROS generated by mitochondria have historically been seen as toxic by‐products of the electron transport chain (ETC) that cause cell damage and injury, such as in ischaemia–reperfusion injury or other disorders. However, significant evidence points to the role of low levels of ROS as signalling molecules that play important physiological roles in a variety of biological processes, such as O₂ sensing (McCord, 1985; Cross et al. 1987; Brigelius‐Flohe & Flohe, 2011).

The mitochondrial ETC is composed of four multiprotein complexes in the mitochondrial inner membrane. It generates an electrochemical gradient across the membrane which is used by the ATP synthase to generate ATP during oxidative phosphorylation. Complexes I and II of the ETC each pass a pair of electrons to the electron carrier ubiquinone, which then becomes ubiquinol. The two electrons are sequentially removed from ubiquinol at the Q_o ubiquinol binding site in complex III. The Riske iron–sulfur protein (RISP) in complex III receives the first electron from ubiquinol and passes the electron on to cytochrome c ₁, to cytochrome c and finally to complex IV where it is transferred to molecular O₂ to form H₂O. Removal of the first electron from ubiquinol results in the formation of the transient free radical ubisemiquinone at the Q_o site. The second electron is then rapidly removed by the b cytochromes and ubiquinone returns to the membrane pool.

Mitochondria can generate ROS at complexes I, II and III (Fig. 1; Misra & Fridovich, 1972; Turrens & Boveris, 1980; Turrens et al. 1985; Garcia‐Ruiz et al. 1995; Genova et al. 2001; Leach et al. 2001). In complexes I and II, escape of electrons from flavins and iron–sulfur groups results in superoxide formation and release into the mitochondrial matrix. In complex III, an increased lifetime of ubisemiquinone due to delayed removal of the last electron represents a possible mechanism for increased ROS generation during hypoxia. Oxidants formed at the Q_o site of complex III are released into the intermembrane space, due to the high electrical field strength within the inner membrane (Sabharwal et al. 2013). The principal ROS species produced in cells is superoxide (O₂ ⁻), which is dismuted to hydrogen peroxide (H₂O₂) either enzymatically or spontaneously. Superoxide dismutases (SODs) combine two superoxides to produce H₂O₂ in the cytosol or intermembrane space (via Cu–Zu SOD) and in the mitochondrial matrix (via MnSOD) (Murphy, 2009; Murphy et al. 2011). H₂O₂ is an important cellular signalling molecule that interacts with proteins by reversibly oxidizing thiol groups on cysteine or methionine residues resulting in altered protein structure and function (Brigelius‐Flohe & Flohe, 2011). Increases in ROS production are thought to be the triggering mechanism of HPV (Rounds & McMurtry, 1981; Weir et al. 1983; Leach et al. 2001; Waypa et al. 2001, 2006; Liu et al. 2003; Wang et al. 2007).

Figure 1. Signalling in hypoxic pulmonary vasoconstriction.

Hypoxia increases production at mitochondrial complex III of reactive oxygen species (ROS), which are released into the mitochondrial intermembrane space. ROS signals (superoxide and H₂O₂) can move from the intermembrane space to the cytosol where the superoxide is converted to H₂O₂. H₂O₂ can then activate many downstream signalling pathways. H₂O₂ can activate phospholipase C (PLC) leading to production of inositol 1,4,5‐trisphosphate (IP₃) and diacylglycerol (DAG) from phosphatidylinositol bisphosphate (PIP₂). IP₃ activates IP₃ receptors resulting in Ca²⁺ release from the sarcoplasmic reticulum (SR) and DAG activates receptor‐operated Ca²⁺ channels (ROCC). H₂O₂ also directly activates ryanodine receptors (RyR) by oxidizing cysteine residues, leading the Ca²⁺ release from SR. Release of Ca²⁺ from the SR causes store depletion, which activates Ca²⁺ entry via activation of calcium release‐activated Ca²⁺ (CRAC) channels. H₂O₂ also inhibits K_V channels, which increases the membrane potential (E _m) and opens voltage‐dependent Ca²⁺ channels (VDCC). Together, the increased intracellular Ca²⁺ concentration ([Ca²⁺]_i) causes Ca²⁺ to bind calmodulin (CaM) and activate myosin light chain kinase (MLCK), which phosphorylates myosin (MLC) and leads to contraction of pulmonary arterial smooth muscle cells (PASMC).

Hypoxia‐induced increases in mitochondrial ROS

The redox theory of HPV originally proposed by Archer and colleagues postulated that ROS generated by the mitochondrial ETC decreases in response to hypoxia in pulmonary arterial smooth muscle cells as a consequence of decreased O₂ availability (Archer et al. 1989; Archer et al. 1993). However, since then a significant body of evidence has emerged which indicates that increases in mitochondrial ROS develop in response to hypoxia, which trigger HPV (Leach et al. 2001; Liu et al. 2003; Guzy et al. 2005; Waypa et al. 2006, 2010, 2013; Wang et al. 2007; Rathore et al. 2008; Jung et al. 2013; Yang et al. 2016b; Song et al. 2017). The controversy surrounding these opposing mechanisms stems mainly from the different techniques used to measure ROS generated in response to hypoxia. In the late 1980s and early 1990s, Archer and colleagues employed luminol and lucigenin in isolated perfused lungs (Archer et al. 1989, 1993). As these reagents may be membrane impermeant and also reside in the extracellular space, the decline in chemiluminescence they detected during hypoxia, which was interpreted by that group as reflecting a decrease in mitochondrial ROS generation, instead likely reflects decreases in extracellular superoxide levels, which are less relevant for intracellular signalling.

More recently, several reagents have been developed to measure intracellular ROS in specific subcellular compartments. Using a FRET‐based redox sensor, several studies demonstrated an increase in oxidation following acute hypoxia (Guzy et al. 2005; Mansfield et al. 2005). To further understand the changes in oxidant signalling following hypoxia, Waypa and colleagues targeted redox‐sensitive proteins to specific subcellular compartments of PASMCs (Waypa et al. 2010). These sensors (roGFPs) are mutants of green fluorescent protein that provide ratiometric assessments of thiol oxidation; they reveal that oxidation is significantly decreased in the mitochondrial matrix during hypoxia but is increased in the mitochondrial intermembrane space and cytosol (Waypa et al. 2010). Similar results were seen using these sensors in precision cut murine lung slices (Desireddi et al. 2010). In cultured PASMCs, scavenging of ROS during hypoxia inhibited the hypoxia‐induced increase in Ca²⁺ in the cytosol ([Ca²⁺]_i) (Waypa et al. 2001, 2002, 2006). Moreover, adenoviral expression of peroxiredoxin‐5, an H₂O₂ scavenger, in the intermembrane space attenuated ROS signalling in the intermembrane space and cytosol, and also inhibited stabilization of hypoxia‐inducible factor 1α (HIF‐1α) in acute hypoxia in PASMCs (Sabharwal et al. 2013). Together, these studies demonstrate that ROS signalling is required for the activation of pathways linking hypoxia to the contractile response in PASMCs.

It has been suggested that the increase in ROS generation during hypoxia is due to increased ROS production at complex III (Waypa et al. 2016). However, several studies suggest that inhibition of complex I or complex III can inhibit HPV in rat lungs (Leach et al. 2001; Waypa et al. 2001, 2006). Genetic studies have identified RISP as essential for ROS production at complex III in hypoxia, because in the absence of this protein complex III cannot oxidize ubiquinol to ubisemiquinone (Guzy et al. 2005). Depletion of RISP from PASMCs has been shown to abolish hypoxia‐induced increases in ROS generation in the mitochondrial intermembrane space, and hypoxia‐induced increases in [Ca²⁺]_i (Guzy et al. 2005). Additionally, deletion of RISP blocks the acute increase in right ventricular systolic pressure in response to hypoxia (Waypa et al. 2013). As decreases in electron flux through complex I can decrease electron entry into complex III, these findings suggest that electron flux through complex III is a critical signalling event in the response of pulmonary vascular cells to hypoxia.

While excessive oxidant stress is known to contribute to cellular dysfunction and death under conditions such as ischaemia–reperfusion (Schriewer et al. 2013), the levels of ROS involved in redox signalling in hypoxia are much lower and their effects on protein thiol redox state are eventually reversed by reductases such as thioredoxin and glutaredoxin (Sabharwal & Schumacker, 2014). Similar low levels of oxidant signalling mediate other physiological effects, such as the mitogenic response to growth factors (Finkel, 2011).

Other sources of ROS

Although an abundance of data support the role of mitochondria as the site of the increase in ROS signalling during hypoxia, other sources of ROS may also be engaged to amplify that response. NADPH oxidases are membrane‐bound multiprotein complexes that transfer electrons from NADH or NADPH to molecular O₂ to produce superoxide. These systems have been suggested to function as O₂ sensors in a variety of cell types (Wolin et al. 2007). However, studies have demonstrated that inhibition of NOX2 in mice did not alter the pulmonary response to hypoxia (Archer et al. 1999; Weissmann et al. 2006b). These studies suggest that while NADPH oxidases may contribute to the signalling pathways activated in response to hypoxia, they are not essential for triggering the response.

Downstream mechanisms of HPV

In biology, many pathways have redundant and compensatory mechanisms that protect the organism and allow for amplification of the signal. Increased ROS generation is appropriate to consider as a signalling mechanism in HPV because ROS are involved in regulating many processes including Ca²⁺ release, activation of AMP kinase, activation of Ca²⁺ channels and other signalling pathways. The following sections will highlight the role of ROS in the downstream mechanisms of HPV.

K_V channels

The debate as to whether hypoxia increases or decreases ROS notwithstanding, it has long been known that hypoxia inactivates K_V channels (Post et al. 1992; Yuan et al. 1993). Inhibition of K_V channels results in depolarization of the plasma membrane which can induce Ca²⁺‐dependent action potentials, increases in [Ca²⁺]_i and vasoconstriction. Several studies have suggested that increased ROS production leads to inhibition of K_V channels (Cogolludo et al. 2006; Mittal et al. 2012; Sahoo et al. 2012). These studies demonstrated that increased production of H₂O₂ plays a role in K_V channel inhibition and the contractile response of pulmonary arteries (Cogolludo et al. 2006; Mittal et al. 2012). Additionally, a recent study demonstrated that hypoxia causes inhibition of K_V channels, which can be mimicked by application of H₂O₂ in PASMCs (Sommer et al. 2017). Together, these data suggest that hypoxia‐induced increases in mitochondrial ROS production can inhibit K_V channels resulting in cellular membrane depolarization and HPV (Fig. 1). These data also challenge the idea that decreases in ROS lead to K_V channel inhibition through a reductive (as opposed to oxidative) stress pathway (Michelakis et al. 2004).

Increased [Ca²⁺]_i

Hypoxia causes a rapid increase in [Ca²⁺]_i in PASMCs that leads to smooth muscle contraction. As discussed above, increased ROS production following hypoxia inhibits K_V channels and causes cellular membrane depolarization. This leads to the opening of voltage‐dependent Ca²⁺ channels (VDCCs) (Post et al. 1992; Yuan et al. 1993). VDCCs are important for Ca²⁺ influx following membrane depolarization, but are not the only source of increased [Ca²⁺]_i in response to hypoxia. Studies have shown that blockade of VDCCs can only partially inhibit HPV in rat pulmonary arteries, suggesting that other sources of increased [Ca²⁺]_i play a role in the response to hypoxia (Robertson et al. 2000b). The hypoxia‐induced increase in [Ca²⁺]_i is also mediated by ryanodine receptors, which, when activated, release Ca²⁺ from the sarcoplasmic reticulum. ROS have been shown to interact with and activate ryanodine receptors (RyRs) in PASMCs leading to increased [Ca²⁺]_i (Lin et al. 2007; Liao et al. 2011). Ca²⁺ release from the sarcoplasmic reticulum can also come from activation of inositol 1,4,5‐trisphosphate (IP₃) receptors. H₂O₂ has been shown to activate phospholipase C, which generates diacylglycerol (DAG) and IP₃ from phosphatidylinositol bisphosphate (PIP₂) (Gonzalez‐Pacheco et al. 2002). This suggests a mechanism by which hypoxia‐induced ROS can cause Ca²⁺ release from IP₃‐sensitive intracellular stores. Additionally, DAG activation of receptor‐operated Ca²⁺ channels (ROCCs), such as transient receptor potential channel 6 (TRPC6), can also increase [Ca²⁺]_i either directly by conducting ionized calcium, or indirectly by conducting sodium ions that contribute to the membrane depolarization. Studies show that application of a DAG analogue caused vasoconstriction under normoxic conditions, but had no effect in TRPC6 knockout mice (Weissmann et al. 2006a; Fuchs et al. 2011; Smith et al. 2015). The release of Ca²⁺ from the sarcoplasmic reticulum from RyR and IP₃ receptors can lead to store depletion and the opening of store‐operated Ca²⁺ channels (SOCCs). A study by Mungai and colleagues demonstrated that hypoxia induces ROS‐dependent release of calcium from intracellular stores, leading to the activation of SOCCs and increased [Ca²⁺]_i in PASMCs (Mungai et al. 2011). Several studies have reported that inhibition of SOCCs can inhibit HPV, which underscores the importance of this mechanism in mediating the pulmonary vascular response to hypoxia. Collectively, these data demonstrate that several mechanisms involving hypoxia‐induced increases in ROS can lead to increased [Ca²⁺]_i and HPV.

Contraction mechanisms

An increase in [Ca²⁺]_i is a major trigger for PASMC contraction and acute pulmonary vasoconstriction (Fig. 1). When [Ca²⁺]_i is increased, Ca²⁺ binds to calmodulin (CaM) and forms a Ca²⁺–CaM complex. The Ca²⁺–CaM complex activates myosin light chain kinase (MLCK), which phosphorylates myosin and leads to contraction of PASMCs. The Ca²⁺–CaM complex can also coordinate with RhoA to activate Rho kinase (ROK), which phosphorylates myosin light chain phosphatase (MLCP). Phosphorylation inactivates MLCP also causing smooth muscle contraction. Hypoxia has been shown to activate RhoA in PASMC and endothelial cells leading to activation of ROK (Robertson et al. 2000a; Wang et al. 2001; Gosal et al. 2015). ROK induces Ca²⁺ sensitization of contractile proteins by activating the MLCP‐inhibitor protein, CPI‐17 (Koyama et al. 2000). This allows for a small increase in [Ca²⁺]_i to cause a significant increase in contraction of PASMCs.

Chronic hypoxia‐induced signalling

Hypoxia activates hypoxia‐inducible factors (HIFs) which are oxygen‐sensitive transcription factors involved in regulating a variety of physiological and pathological mechanisms (Shimoda & Laurie, 2014). HIF transcriptional activity requires formation of a heterodimer composed of an oxygen‐regulated component (HIF‐1α) and an O₂‐independent component (HIF‐1β, ARNT). Under normoxic conditions, HIF‐1α degradation is initiated by HIF prolyl hydroxylase enzymes. Multiple studies have shown that mitochondrially derived ROS signals regulate HIF‐1α stabilization by controlling the activity of HIF prolyl hydroxylase (Chandel et al. 1998; Guzy et al. 2005). However, given that pulmonary arteries contract almost immediately after sensing a decrease in O₂ tension, HIF‐mediated changes in transcription are unlikely to regulate the acute phase of HPV (Shimoda & Laurie, 2014). Some support for a role for HIF in phase 1 of HPV comes from patients with Chuvash polycythemia, a rare genetic condition which results in stabilization of HIF under normoxic conditions, who exhibit enhanced HPV (Smith et al. 2006; Shimoda & Laurie, 2014). The exact role of HIF in the acute phase of HPV remains to be determined, but it seems likely that basal HIF activity contributes to the permissive transcriptional control of multiple components that contribute to the HPV response. Indeed, HIF‐1α heterozygous mice show blunted hypoxia responsiveness in multiple tissues (Shimoda et al. 2001). HIF also plays a role in the development of pulmonary hypertension, which arises from various aetiologies including chronic hypoxia. HIF mediates the hypoxia‐induced increase in TRPC channel expression and the decrease in K⁺ channel expression which leads to pulmonary vasoconstriction and pulmonary vascular remodelling (Wang et al. 1997, 2006; Shimoda & Laurie, 2014). Additionally, ET‐1 is a HIF target (Bodi et al. 1995) that contributes to sustained activation of ROK. Prolonged hypoxia can thereby increase ROS signalling, leading to HIF stabilization, ET‐1 release, ROK activation, Ca²⁺ sensitization and pulmonary vasoconstriction (Chi et al. 2010). Together, these data demonstrate a role of HIF in the pulmonary vascular remodelling and the development of pulmonary hypertension in response to chronic hypoxia.

Conclusion

Hypoxia triggers a diverse set of cellular responses involving transcriptional and post‐translational mechanisms in the lung, which are triggered through the activation of O₂ sensors. Abundant evidence supports the concept that increased ROS production at complex III is a required event in the activation of these mechanisms that comprise the functional responses to hypoxia in the pulmonary circulation. While the molecular details linking hypoxia to the increase in ROS following hypoxia are not yet fully elucidated, it is clear that release of ROS into the intermembrane space is involved in the hypoxia response. Further understanding of these pathways and how they are activated may lead to the identification of targets that could be exploited for treatment of lung disorders that involve cellular hypoxia, including lung cancer, pulmonary hypertension and high‐altitude pulmonary oedema.

Additional information

Competing interests

The authors have no competing interests to report.

Funding

This work was supported by NIH grants HL35440, HL122062 and HL109478 (P.T.S.) and a Parker B. Francis Fellowship (K.A.S.).

Biographies

Kimberly A. Smith studied biology at Illinois Wesleyan University and earned her Master's degree from Southern Illinois University, and her PhD from Northwestern University. During her postdoctoral fellowship in P.T.S.’s lab, she studied the molecular mechanisms of hypoxia and pulmonary hypertension.

Paul T. Schumacker is the Patrick M. Magoon Professor of Neonatology Research in the Department of Pediatrics at Northwestern University Feinberg School of Medicine and the Ann and Robert H. Lurie Children's Hospital, with joint appointments in the Departments of Medicine (Pulmonary and Critical Care), Cellular and Molecular Biology, and the Northwestern University Comprehensive Cancer Center. His laboratory studies oxygen sensing, mitochondrial redox signalling, and cellular responses to hypoxia in health and disease. He has published more than 200 articles, mostly in the field of respiratory biology.