The role of stretch-activated ion channels in acute respiratory distress syndrome: finally a new target?

Abstract

Mechanical ventilation (MV) and oxygen therapy (hyperoxia; HO) comprise the cornerstones of life-saving interventions for patients with acute respiratory distress syndrome (ARDS). Unfortunately, the side effects of MV and HO include exacerbation of lung injury by barotrauma, volutrauma, and propagation of lung inflammation. Despite significant improvements in ventilator technologies and a heightened awareness of oxygen toxicity, besides low tidal volume ventilation few if any medical interventions have improved ARDS outcomes over the past two decades. We are lacking a comprehensive understanding of mechanotransduction processes in the healthy lung and know little about the interactions between simultaneously activated stretch-, HO-, and cytokine-induced signaling cascades in ARDS. Nevertheless, as we are unraveling these mechanisms we are gathering increasing evidence for the importance of stretch-activated ion channels (SACs) in the activation of lung-resident and inflammatory cells. In addition to the discovery of new SAC families in the lung, e.g., two-pore domain potassium channels, we are increasingly assigning mechanosensing properties to already known Na⁺, Ca²⁺, K⁺, and Cl⁻ channels. Better insights into the mechanotransduction mechanisms of SACs will improve our understanding of the pathways leading to ventilator-induced lung injury and lead to much needed novel therapeutic approaches against ARDS by specifically targeting SACs. This review 1) summarizes the reasons why the time has come to seriously consider SACs as new therapeutic targets against ARDS, 2) critically analyzes the physiological and experimental factors that currently limit our knowledge about SACs, and 3) outlines the most important questions future research studies need to address.

Current Knowledge Gaps and Obstacles in Studying Stretch-Activated Ion Channels in ARDS

The biophysical characterization of ion channel properties in living biological systems represents one of the most difficult challenges in modern medical research. The multitude of different ion channels expressed on a given cell makes it notoriously difficult to isolate the functional and biophysical characteristics of a single channel. To circumvent this problem, electrophysiologists routinely overexpress channels of interest in heterologous expression systems such as the Xenopus laevis oocyte. This approach allows for a detailed characterization of biophysical channel properties in a controlled, well-defined environment but the native intra- and extracellular milieus of the channel are lost. Conversely, the study of a given channel in a native cell is often limited by low expression levels and interference from the variety of other channels present on any cell. Although many of these limitations apply to ion channel research in general, they are particularly relevant to stretch-activated ion channels (SACs), since not only the intra- and extracellular environments but also the amount and direction of mechanical forces that SACs are exposed to in vivo are difficult to standardize.

Nevertheless, the central role of SACs in mechanotransduction processes in the lung is nowadays undisputed. Understanding the biophysical properties and contribution of SACs to both the development and resolution of ARDS is particularly important, since by definition the lungs of all ARDS patients are exposed to mechanical stretch induced by positive pressure ventilation. Decades of intense research have substantially expanded our knowledge to the point that we can now confidently assign a role to SACs in the development of the two main clinical features of ARDS: 1) alveolar inflammation and 2) loss of alveolar-capillary barrier function. To advance this field, it is vitally important to combine our knowledge from three different areas of research: 1) the detailed understanding of biophysical SAC properties, unfortunately often obtained in rather artificial expression systems; 2) our knowledge of ARDS pathophysiology derived from in vitro and in vivo lung injury models; and 3) changes in ARDS biomarker profiles and physiological variables obtained from clinical studies, e.g., oxygenation and ventilation indexes, plasma and bronchoalveolar lavage (BAL) pH, BAL cytokines and cell counts, lung compliance and resistance measurements, and morbidity/mortality data. The creation of unifying ARDS models that integrate all of these research areas represents an extremely challenging task given the complexity of each individual field. This review highlights the most relevant topics in each field and unravels common areas of interest to facilitate collaborations between investigators from these different, often even opposing fields. Combining our knowledge will undoubtedly unify these fields, and together we can target SACs as a very promising new therapeutic approach against ARDS.

Background

Acute respiratory distress syndrome (ARDS) was first described in 1967 and to this day remains a devastating illness characterized by diffuse alveolar injury and inflammation caused by neutrophil recruitment to the lungs, macrophage activation, inflammatory cytokine release, and loss of alveolar-capillary barrier function (177). Despite our best efforts, mortality rates for ARDS patients remain high, and more patients die from ARDS than any other illness in our intensive care units (ICUs), accounting for 196,600 cases, 74,500 deaths, and 2.2 million ICU days per year for adults (163), and 7,700 cases, 1,400 deaths, and 62,000 ICU days per year for children in the US (225). Current treatment regimens for ARDS primarily consist of oxygen supplementation therapy (hyperoxia; HO) and mechanical ventilation (MV) (65). Unfortunately, the combined effects of HO and MV cause even more lung damage than each stimulus alone (119, 176). For example, HO exposure of alveolar epithelial cells (AECs) alters their mechanical properties, resulting in increased cell detachment during mechanical stretch (159). Although the signaling cascades employed by HO and inflammatory cytokines are relatively well established, less is known about “mechanosensing” and “mechanotransduction” processes in the lung, including the mechanisms responsible for the simultaneous integration of HO-, MV- and cytokine-mediated signaling during epithelial, endothelial, inflammatory, and immune cell activation. Nevertheless, our knowledge in this field is rapidly expanding. The combined effects of HO and MV activate c-Jun NH₂-terminal kinase (JNK) and apoptosis signal-regulating kinase (ASK)-1 (117, 118), but other mitogen-activated protein kinases, tyrosine kinases, transcription factors (including NF-κB), and protein kinase C also mediate mechanical coupling processes in ARDS lungs (7, 61, 126). Therefore, the identification of specific targets within these complex and interrelated signaling pathways constitutes a major challenge in the development of new therapeutic strategies against ARDS. This is further evidenced by the fact that, besides the introduction of “lung-protective,” low tidal volume ventilation strategies and possibly prone positioning, few if any interventions have improved the outcomes or survival rates of ARDS patients over the past two decades (16, 51). Importantly, in tissues other than the lung, SACs are well known to function as so-called “signal integrators” and it is becoming increasingly evident that in the lung these channels could constitute the link between HO, MV, and cytokine effects (76, 166, 174).

SACs Are Mechanosensors and Mechanotransducers in the Lung

The effects of mechanical stretch on many biological lung functions, including fetal lung growth, surfactant metabolism, extracellular matrix and cytoskeleton turnover, cell proliferation, apoptosis, mediator release, and alveolar-capillary permeability, have been recognized for over two decades (3, 38, 43, 47, 48). Lansman et al. (104) described the first stretch-activated cation channel in 1987 in the vascular endothelium. Advances in nanotechnology techniques now enable us to image and measure changes in physical forces at subcellular level, improving our understanding of membrane deformation processes in health and disease. These advances led to a renewed interest in the molecular mechanisms underlying mechanotransduction processes in living organisms (91). SACs are among the best-studied examples of mechanosensors and mechanotransducers in nature and can mediate both mechanoelectrical and mechanochemical signals (139). Although mechanosensitive ion channels were originally identified and cloned in bacteria (25), in humans various types of SACs can be found in almost any body tissue and dysregulation of mechanical responses often results in major pathology (137, 191). SACs are usually involved in the earliest stage of mechanosensing and can alter the cell membrane potential within micro- to milliseconds. Such a rapid perturbation can lead to either membrane depolarization via Na⁺, Ca²⁺, or nonselective cation channels (such as Piezo 1 and 2), or to membrane hyperpolarization via K⁺ channels. To add further complexity to this system, in mammalian cells not only stretch-activated but also some stretch-inactivated ion channels have been characterized (166). While certain SACs can be directly activated by membrane tension, others require force transduction via anchors in the extracellular matrix, the cytoskeleton, or both (139) (Fig. 1), and any damage to the cortical cytoskeleton, as induced by MV, can severely affect SAC gating and function (199).

Fig. 1.

SACs are stretch sensors and stretch integrators. Mechanical stretch can either activate or inhibit SAC function (1) and mobilize intracellular signaling cascades (2) affecting gene transcription (3) and protein translation (4), which ultimately regulate secretory transport pathways (5), cytoskeletal rearrangements, and capillary-alveolar barrier function (6). In addition, SACs may also directly modulate nongenomic downstream effector functions (dotted arrows). The amount and vector of stretch forces applied to a plasma membrane depend on the integrity of tight junctions (TJ), cell adhesion to the extracellular matrix (ECM), and cell-cell interactions (epithelial/endothelial-to-inflammatory cell interactions). Adapted from Ref. 115.

To date, no unique protein structure has been found to be inherent to SACs. The term SAC does not apply to a specific ionic conductance but includes Na⁺ (221), K⁺ (158), Ca²⁺ (112), Cl⁻ (75), and nonselective cation channels (35) (Fig. 2). The main unifying feature of SACs is their ability to respond to membrane stretch with a configuration change between open and closed states (122) and, with few exceptions, their inhibition by gadolinium (Gd³⁺) (215). However, just because mechanical stretch changes the open probability of a given channel, this does not imply any physiological relevance. Therefore, to understand the pathophysiological implications of changes in SAC expression and function in ARDS lungs, each channel of interest has to be characterized in the appropriate microenvironment consisting of HO, mechanical stretch, and inflammatory cytokines.

Fig. 2.

Expression and role of SACs. A major unifying feature of SACs is a configuration change upon mechanical stretch exposure and inhibition by gadolinium (Gd³⁺). All major classes of ion channels contain members that are stretch-sensitive. SACs are almost ubiquitously expressed in body tissues and regulate a variety of cellular functions. CNS, central nervous system; PNS, peripheral nervous system.

The ability of SACs to sense changes in membrane shear stress, tension, and curvature, which are all features associated with MV, makes SACs the ideal candidates for mechanotransducers and signal integrators in ARDS, even though the specific mechanisms of channel activation (or inhibition) by stretch may vary substantially between channels. Stretch can activate a given SAC via 1) direct tension on the lipid bilayer, 2) transduction of stretch via a tethering mechanism from cytoskeletal or extracellular matrix structures, or 3) release or binding of another molecule that in turn activates a channel (e.g., a phosphokinase), as has been proposed for the epithelial Na⁺ channel (ENaC) (20). However, to assign a role to SACs in the pathogenesis of ARDS, it is important to recognize that SACs are not exclusively activated by mechanical stretch but also by a variety of other stimuli that are altered in ARDS lungs, including changes in pH or temperature, drugs (e.g., volatile anesthetics), biological ligands (e.g., ATP, lipids), and changes in the membrane potential (voltage dependency) (40, 103). Over the past decade, we accumulated evidence that pharmacological and molecular modulation of SACs alters inflammatory processes in the lung. Nevertheless, some of the specific mechanisms that link SAC activation (or inhibition) to inflammatory mediator release and alveolar-capillary barrier function, the two hallmarks of ARDS, are still under investigation. Due to our limited knowledge about SAC expression and function in the human lung, many of our mechanistic hypotheses rely on SAC characteristics described in other organs and on data extrapolated from animal models. In rat lungs, so-called rapidly adapting receptors (RARs) transduce mechanical stimuli to nerve fibers via activation of voltage-dependent Na⁺ and K⁺ channels, N-type Ca²⁺ channels, and Cl⁻ channels (GABA A- and glycine-activated) (36). The complex interplay between these channels excites the afferent nerve terminal and regulates airway tone and fluid homeostasis (36). This is only one example of a well-defined pathway in rodents that, if confirmed in the human lung, could become tremendously useful in designing new therapies against ARDS. Other exciting candidates are so-called 2-pore domain K⁺ channels (K2P), which are not only stretch sensitive (73, 223) but are also regulated by HO (172) and affect cytokine secretion from mouse ARDS lungs (174). We are now finally at a stage where many of the characteristics assigned to individual SACs can be exploited in the context of ventilator-induced lung injury (VILI), which allows us to specifically target these channels as new therapeutic options against ARDS.

The Role of SACs in Inflammatory Mediator Release and Loss of Capillary-Alveolar Barrier Function

Two hallmarks of ARDS are 1) increased levels of inflammatory mediators in the BAL fluid and plasma (19) and 2) the loss of alveolar-capillary barrier function (33), both of which correlate with patient mortality rates (56, 125). Historically, ion channel activity has been closely linked to mediator release, particularly in neuronal cells where cell depolarization leads to Ca²⁺ influx followed by neurotransmitter secretion (134), but it is still controversial whether this model also applies to nonneuronal cells. Today we know that ion channels regulate mediator secretion in many other cell types, including lung-resident, inflammatory, and immune cells (94, 152), although the secretory mechanisms employed by nonexcitatory cells are not as clearly defined. In general, an increase in intracellular Ca²⁺ as the final common pathway precedes mediator release (160). K⁺ and Cl⁻ currents affect Ca²⁺ entry via alteration of the membrane potential, thereby increasing or decreasing the electrical driving force for Ca²⁺ to enter a cell (53, 121). In many tissues K2P channels, including the TREK subfamily, are important stabilizers of the resting membrane potential and prevent cell depolarization (59). In contrast, inhibition of K2P channels results in decreased “leak currents” and accumulation of positive K⁺ charges inside the cell resulting in cell depolarization (18). Importantly, in lung epithelial, inflammatory, and immune cells not all secretory pathways are Ca²⁺ dependent (173, 190). Furthermore, certain meditators such as macrophage-inflammatory protein-2 (MIP-2) are secreted upon stretch exposure without activation of any particular SACs, as suggested by a lack of effect of the SAC blocker gadolinium on stretch-induced MIP-2 secretion (130).

From both in vitro and in vivo ARDS studies we learned that mechanical forces, HO exposure, and proinflammatory cytokines such as TNF-α alter transepithelial ion transport. The net result is a decrease in alveolar fluid clearance (68, 154, 206), which is directly associated with an increased length of MV and patient mortality (83, 177). Undoubtedly, SACs are important in both alveolar edema formation and resolution, but their regulation in the complex inflammatory environment of ARDS lungs is somewhat difficult to predict (Fig. 3). Nevertheless, in rat lungs, Gd³⁺ (an inhibitor of stretch-activated ion channels) prevents stretch-induced loss of capillary-alveolar barrier function (142), and HO, stretch, and TNF-α all regulate barrier function by altering the phosphorylation status of tight junction (TJ) proteins (58). In general, epithelial TJ phosphorylation is a Ca²⁺-dependent process and several Ca²⁺ channels, including transient receptor potential (TRP) channels, contribute to alveolar-capillary barrier dysfunction in ARDS (78, 129) but other conductances such as the CFTR Cl⁻ channel may be involved as well (108). Besides ventilator-induced stretch, mechanosensitive ion channels in the lung are also exposed to stretch forces evoked by cell swelling, a process underlying cell apoptosis, detachment, proliferation, and repair processes in ARDS lungs. Na⁺, K⁺, Ca²⁺, and Cl⁻ channel families all include members that are activated by cell swelling-induced membrane stretch (180).

Fig. 3.

SACs as orchestrators of alveolar inflammation in ARDS. Hyperoxia, mechanical stretch, and inflammatory cytokines alter SAC expression and function, which impacts the pathogenesis of ARDS through downstream effects on cytokine secretion from epithelial, endothelial, inflammatory, and immune cells and alveolar-capillary barrier function (epithelial and endothelial cells). While some of these processes are mediated via changes in intracellular Ca²⁺ concentrations ([Ca_i²⁺]), others appear Ca²⁺ independent.

The enthusiasm about targeting SACs as mechanosensors and mechanotransducers in ARDS stems from the possibility of specifically altering stretch-induced lung injury processes without interfering with innate immune mechanisms responsible for host defenses and lung repair.

Stretch-Activated Na⁺ Channels

Stretch-activated Na⁺ channels are important regulators of alveolar fluid clearance and inflammatory mediator secretion by regulating the plasma membrane potential of lung-resident and inflammatory cells. One of the earliest models of a mechanotransducing complex was described in the nematode Caenorhabditis elegans, where the key molecules (so-called MEC and DEG proteins) forming a mechano-gated ion channel are related to the α, β, and γ subunits of the mammalian epithelial Na⁺ channel ENaC (54, 114, 185). Despite some controversy in the 1990s over the mechanosensitivity of ENaC, which derived from differences in biophysical channel properties between lipid bilayer and oocyte experiments (161), nowadays the stretch-activated nature of ENaC is widely accepted (4, 20, 66, 67, 175). In fact, the channel appears intrinsically mechanosensitive since it can be directly activated by mechanical stretch without cytoskeletal interactions (100). Besides direct ENaC activation by membrane tension, other models propose stretch-mediated ENaC activation via a tethering mechanism to the cytoskeleton or extracellular matrix, or via stretch-mediated release of another molecule from the cell that in turn activates ENaC. Evidence actually exists to support all three mechanisms (44, 77, 209). Importantly, stretch activation of ENaC plays a role not only for lung epithelial but also for inflammatory cells, likely via channel interactions with cytoskeletal structures (1). Interestingly, in lymphocytes (116) and osteoblasts (101) only the α subunit is required to form a functional mechanosensitive channel.

ENaC plays a central role in ARDS pathogenesis as a key regulator of alveolar fluid clearance, as required during the exudative phase of ARDS (5, 6, 69, 110). Furthermore, in the developing lung stretch increases epithelial ENaC expression via p38, MAPK, and JNK pathways and improves alveolar fluid clearance (14, 102, 133), which could accelerate recovery from the exudative phase of ARDS. In contrast, elevated serotonin levels, as commonly observed in ARDS lungs, inhibit ENaC activity and correlate with worsening pulmonary edema and ARDS outcomes (72, 128). In an inhalation (hydrogen sulfide) lung injury model, downregulation of α-ENaC also correlated with worsening pulmonary edema, which was prevented by steroid therapy (97). In airway neurons, activation of an ENaC-like, mechanosensitive Na⁺ channel contributes to the propagation of lung inflammation (36). ENaC is also expressed in B lymphocytes and can be activated by membrane stretch, but, in contrast to epithelial ENaC, in lymphocytes an association of ENaC with the cytoskeleton is mandatory for current activation (1, 116). ENaC activity is also sensitive to stretch induced by cell swelling and shrinkage (96), as observed during cell death or due to osmotic changes caused by protein-rich alveolar exudate in ARDS lungs, and could thus affect both the early exudative phase and late fibrotic phase of ARDS. In addition to stretch, ENaC can also be activated by reactive oxygen species (ROS) (10, 188), which are routinely elevated in ARDS lungs due to HO and MV exposure. In a murine LPS-induced ARDS model, ROS increased ENaC activity and alveolar fluid clearance (71). Therefore, ROS-induced upregulation of ENaC activity may constitute a protective mechanism to counteract increased pulmonary edema formation during the exudative phase of ARDS.

Besides ENaC, stretch can also activate voltage-sensitive Na⁺ channels (Na_v1.5–1.9) (23, 179, 201). In the lung, Na_v channels are expressed in bronchial smooth muscle (29) and afferent nerve endings in the airways (132). Interestingly, calpain, which is elevated in lungs exposed to MV and promotes lung injury via neutrophil recruitment (113) and loss of epithelial barrier function (203), can cleave Na_v channels (197). Na_v channels should be considered as potential targets during the exudative and proliferative phases of ARDS, since stretch activation of Na_v channels in colonic epithelial cells promotes cell migration and steroid metabolism (85), while in neurons it leads to Ca²⁺ influx-mediated cell death (201).

Furthermore, the recent characterization of slowly and rapidly adapting stretch receptors (SARs and RARs, respectively) in the lung has greatly improved our understanding of mechanosensing (30). SAR and RAR activation is regulated by Na⁺ channels and is responsible for adaptation of hemodynamic responses to breathing patterns. Importantly, SAR and RAR activity is also regulated by hypoxia, a key feature of ARDS pathology. Harvesting the potential of SARs and RARs could be particularly important in light of the increasing utilization rates of airway pressure release ventilation (APRV) for ARDS. This ventilation mode exposes lungs to prolonged inflation and deflation pressures, which is the main stimulus for SAR and RAR activation. In the future, modulation of these receptors could not only provide us with novel ways to target neuronal regulation of airway inflammation but also lead to new lung recruitment strategies to improve ventilation-perfusion mismatching in ARDS.

Besides ion channels, several Na⁺ transporters and exchangers (Na⁺/Ca²⁺, Na⁺/H⁺, Na⁺/K⁺ ATPase) are mechanosensitive and thus attractive targets for new ARDS therapies (123, 196, 206). In fact, the Na⁺/K⁺ ATPase promotes alveolar edema clearance in ARDS and may be stimulated by Na⁺ influx via stretch-activated Na⁺ channels, but the results of clinical trials remain disappointing (32).

Stretch-Activated K⁺ Channels

K⁺ channels are typically classified according to their number of transmembrane domains (TMD): voltage-dependent (K_v) and Ca²⁺-activated (K_Ca) channels have six TMDs; K2P have four TMDs; inwardly rectifying K⁺ channels (Kir; including K_ATP) have two TMDs. Although each class includes channels with stretch-sensitive characteristics, K2P channels are particularly well-known for their mechanosensing and mechanotransducing capabilities (107). The K2P channel family consists of 15 members and is divided into six subfamilies (TWIK, THIK, TREK, TASK, TALK, TRESK) (144). What distinguishes K2P channels from almost any other SAC is that their intrinsic mechanosensitivity, independent of an intact cytoskeleton. Evidence from our group suggests that K2P channels, especially TREK-1, are important regulators of the inflammatory processes observed in ARDS since they are expressed in lung epithelial cells and macrophages and are regulated by both stretch and HO (172). In fact, stimulation of TREK-1-deficient (TREK-1 ko) AECs with TNF-α, the main cytokine in the BAL fluid of ARDS patients, decreases IL-6 and RANTES secretion but increases MCP-1 secretion, while KC/IL-8 release is not affected. Although several TNF-α-mediated signaling pathways including p38 kinase, PKC, and c-Jun NH₂-terminal kinases (JNK1/2/3) are altered in TREK-1 ko AECs, decreased IL-6 secretion is associated with impaired PKCθ phosphorylation, whereas increased MCP-1 secretion occurs unrelated to the increased JNK1/2/3 phosphorylation found in TREK-1 ko AECs (171, 173).

Based on these in vitro findings we expected that TREK-1 ko mice would be protected from HO- and MV-induced lung injury. To our surprise, however, exposure of TREK-1 ko mice to HO increases lung injury as evidenced by gross lung anatomy, hematoxylin and eosin staining of lung sections (increased inflammatory cell infiltration and hemorrhage), increased lung injury scores, decreased lung compliance, and increased macrophage and neutrophil accumulation in the BAL fluid of TREK-1 ko mice (174). Interestingly, exposure of these mice to HO+MV results in no further injury when compared with HO exposure alone. In accordance with our in vitro findings (173), IL-6 levels decrease in HO+MV-treated TREK-1 ko mice, but, in contrast to our in vitro findings, MCP-1 levels also decrease in HO+MV treated TREK-1 ko mice, while TNF-α levels are unchanged between control and TREK-1 ko mice (174). Therefore, in our mouse ARDS model alterations in BAL cytokine concentrations are unlikely the cause for the increased lung injury observed in TREK-1 ko mice. We cannot, however, exclude that intraparenchymal or plasma cytokine levels are altered in TREK-1 ko mice contributing to the observed compliance and lung injury changes. Interestingly, alveolar barrier function appears uncompromised by TREK-1 deficiency since BAL protein levels are similar between control and TREK-1 ko mice (174). One explanation for the increased lung damage observed in HO-exposed TREK-1 ko mice could be an increase in alveolar type II (AT I) and AT II cell apoptosis as evidenced by TUNEL staining and corroborated by increased PARP-1 cleavage (174).

Importantly, TREK channel gating is strongly affected by pH changes with intracellular acidification leading to TREK activation, whereas extracellular acidification results in TREK inhibition (49). Patients with severe ARDS suffer from a significant metabolic acidosis caused by multiorgan failure, which is also the main cause of mortality in these patients. In addition, although never documented in ARDS, the pH of exhaled water vapor of asthmatic subjects is 5.2 compared with a pH of 7.6 in healthy controls (89). Therefore, the acidotic extracellular environment in ARDS could easily inhibit TREK-1 channels and potentially worsen ARDS consistent with the increased lung injury seen in TREK-1 ko mice (174).

Besides, K2P channels, other stretch-sensitive K⁺ channels cannot be overlooked. For instance, K_ATP channels can directly be activated by local bilayer tension and cortical F-actin (87). It is also known that cyclic stretch depletes intracellular ATP stores in AECs, which in turn could inhibit K_ATP activation (37). In addition to the epithelium, K_ATP channels are also mechanosensors in lung endothelial cells where they regulate ROS formation, Ca²⁺ influx, and inflammatory cytokine secretion (39). Although K_ATP channels decrease lymphocyte TNF-α production (208), they promote neutrophil influx and TNF-α and IL-6 release in the lung (149). A variety of K_v channels promote lymphocyte activation and proliferation, and K_v1.3 currents promote LPS- and TNF-α-induced macrophage activation (192). Large conductance K_Ca channels (BK_Ca) are sensitive to both oxygen and stretch and regulate pulmonary blood flow (15), potentially contributing to ventilation-perfusion mismatching in ARDS.

Therefore, K_ATP, K_v, and K_Ca channels may all play important roles in the different stages of ARDS. During the exudative phase, the Kir6.1/6.2 subunits of K_ATP channels can upregulate transepithelial Na⁺ transport via ENaC and Cl⁻ transport via CFTR in AECs (106). In primary lung epithelial cells, K_ATP openers (pinacidil, YM934) and blockers (glibenclamide) can regulate fluid clearance by activation or inhibition of ENaC and cAMP-activated Cl⁻ channels, respectively (106, 168), and inhibition of stretch-activated K_ATP channels impairs fluid clearance in frog lungs (26). Besides K_ATP channels, activation of KvLQT1 and KCa3.1 channels also has a positive effect on alveolar fluid clearance (12). Similar to stretch-activated Na⁺ channels, several of the mechanosensitive K⁺ channels linked to alveolar fluid clearance are also activated by cell swelling, including the K2P channel TASK-2 (136), K_Ca/I_K (200), and K_ATP (39) channels. During the proliferative and fibrotic phases of ARDS, activation of K_ATP currents in lung epithelial cells by both physiological (lipoxin A4) and pharmacological (pinacidil) stimuli can promote cell migration, proliferation, and epithelial repair via ERK-MAP kinase phosphorylation (31). Furthermore, deficiency of Kir7.1 in the lung epithelium delays lung development (194) and KCa3.1-anti-β1 integrin interactions regulate alveolar epithelial wound healing (70).

Stretch-Activated Ca²⁺ Channels

Stretch-activated Ca²⁺ conductances gained increasing attention in the 1980s as endothelial regulators of vascular tone (104), but their contribution to inflammatory processes was not much explored until the 1990s. Initially the molecular identity of these Ca²⁺ conductances remained largely elusive. These channels were simply termed “stretch-activated channels” and thought to activate K_Ca channels, which then would hyperpolarize the cell membrane and promote further Ca²⁺ influx (210). In fact, this mechanism was later recognized as an important regulator of the pulmonary vascular tone (24) and of several immune cell functions, including lymphocyte adhesion (184) and endothelial MCP-1 secretion (212), which is a key neutrophil attractant in ARDS.

Unfortunately, the regulation of mediator secretion from AECs is not as well understood as in neuronal or inflammatory cells, but the activation of stretch-activated Ca²⁺ channels is a common trigger for cytokine release (204). In our search for new therapeutic approaches against ARDS, the TRP superfamily of cation channels has emerged as one of the most promising targets. TRP channels are key regulators and integrators of several major features of ARDS, including mechanosensing and mechanotransduction (217), O₂ sensing (186), inflammatory cell activation and mediator release (111), alveolar-capillary barrier function (8), epithelial wound healing and fibrosis (214), and cell apoptosis and necrosis (79). Although TRP channels are nonselective cation channels, they perform their effector functions largely via Ca²⁺ inward currents (147). To date, the mammalian TRP superfamily comprises seven subfamilies (TRPC, TRPM, TRPV, TRPA, TRPP, TRPML, and TRPN) with a total of 33 members (147). Although several TRP channels are expressed in lung-resident, immune, and inflammatory cells, only four channelopathies have been associated with human disease (137). In the lung, members of the TRPA, TRPC, TRPM, and TRPV subfamilies have been linked to asthma and COPD based on studies in knockout mouse models where these channels regulate bronchoconstriction, airway hyperreactivity, fibrosis, airway remodeling, chronic cough, and mucus hypersecretion (9, 50, 109). Evolutionarily, TRP channels are among the oldest mechanosensors and TRP-4 was identified in C. elegans where it acts as the pore-forming subunit of a native mechanotransduction channel (99). Of all TRPs, the mechanosensing apparatus of TRPC1 and TRPC6 is particularly well characterized (74, 143). However, most studies were performed in cardiac and smooth muscle cells (205), and the applicability of these findings to lung injury models needs to be determined. Of note, while TRP channels can be quite insensitive to stretch in heterologous expression systems, in vivo their low threshold to stretch is well documented (74).

Interestingly, the gating mechanism of several TRP channels is oxygen sensitive and TRP channels themselves can promote ROS production. In fact, TRPC3 and -C4 are endothelial redox sensors (205). In macrophages and endothelial cells, oxidative stress activates Ca²⁺ influx via TRPC1, -C4, and TRPM2 channels causing endothelial barrier disruption (42, 79). Similarly, oxidative stress-induced activation of TRPM2 induces zonula occludens and claudin-2 degradation, as well as neutrophil sequestration, resulting in loss of epithelial tight junctions (78, 213). In contrast, in lung macrophages activation of TRPM2 inhibits further ROS production and prevents endotoxin-induced lung injury (52). In lung nerve fibers, activation of TRPV1 and TRPA1 channels by ROS results in apnea (162). On the other hand, hypoxia, a key feature of ARDS, causes TRPC6- and TRPV4-mediated pulmonary vasoconstriction (120). TRPC1 activation contributes to alveolar-capillary barrier dysfunction and TRPC1 protein is highly upregulated by TNF-α (141), which is routinely elevated in the patients with ARDS. TRPA1 has received attention in ARDS research since its inhibition in vagal nerve afferents decreases ventilator-induced inflammation in a rat model of ARDS (202). TRPM8 activation is associated with IL-1α, IL-1β, IL-4, IL-6, IL-8, IL-13, GMCSF, and TNF-α release from human bronchial epithelial cells (164, 165), all of which contribute to the pathogenesis of ARDS. TRPC1 activation counteracts the development of sepsis-induced ARDS by promoting bacterial clearance and reducing proinflammatory cytokine secretion from macrophages and epithelial cells (224), while TRPM2 deficiency correlates with increased mortality rates in sepsis patients (151). On the other hand, TRPV1 activation promotes somatostatin release from sensory nerve endings and attenuates LPS-induced lung injury (81, 189) but also triggers IL-6 and IL-8 secretion from bronchial epithelial cells (155). TRPM7 induces mast cell degranulation (88), while TRPML2 causes chemokine release from macrophages upon TLR stimulation (182). Clearly, TRP channels affect a variety of mechanisms that regulate barrier function and cytokine secretion in ARDS, and we are in the midst of unraveling these complex networks. It appears, therefore, that while several TRP channels are important regulators of host defense mechanisms, they may also trigger downstream effects that ultimately worsen lung inflammation and barrier function.

Of all TRP channels, TRPV4 has received the most attention as a potential therapeutic target against ARDS. In mouse aspiration and inhalation models of ARDS, TRPV4 inhibition decreases neutrophil and macrophage influx and inflammatory cytokine release, reduces lung injury scores, prevents epithelial and endothelial cell detachment (193), improves lung compliance, and increases blood O₂ saturations, even if administered after acid instillation (8). Recently, another group showed in a similar hydrochloric acid-induced ARDS model that genetic deficiency of TRPV4 and prophylactic administration of a TRPV4 inhibitor prevented neutrophil activation and disease development, including impairment of gas exchange and lung function, inflammation, lung edema, and histological lung damage (218). In contrast to the previous study, in this model the protective effect was no longer observed when the TRPV4 inhibitor was administered after acid instillation. The potential of preserving alveolar-capillary barrier function and reducing macrophage-derived reactive oxygen and nitrogen species by inhibiting TRPV4 was also confirmed in other inhalation- (129) and ventilator-induced (76) ARDS models. Further support for a role of TRPV4 in ARDS pathology is provided by a recent report showing TRPV4 activation by ROS and subsequent Src kinase-dependent endothelial barrier disruption (183). Interestingly, TRPV4 inhibition also prevents heart failure-induced pulmonary edema formation (187), which is important because most ARDS patients die from multiorgan failure, not respiratory failure.

In addition, TRPs may also play a role in adaptive connective tissue responses and epithelial wound healing in ARDS. Inflammatory stimuli found in ARDS, such as TNF-α, LPS, and IL-1α, induce TRPV1 expression in fibroblasts (167), whereas TRPM7 and TRPV4 inhibition decreases lung fibroblast proliferation, differentiation, and fibrogenesis (153, 219). Interestingly, blocking TRPC1 inhibits lung epithelial cell proliferation, while blocking TRPM8 and TRPA1 promotes epithelial cell proliferation (55, 220). Furthermore, several TRPs, including TRPC, TRPV, TRPM, TRPA, and TRPP, play important roles in the regulation of cell apoptosis and necrosis as they promote Ca²⁺ influx during cell shrinkage and cell swelling (146). Cell death via both of these mechanisms plays an important role in ARDS. Interactions of TRPM2 with its splice variant TRPM2-S induces apoptosis in endothelial cells via ROS-induced PKC activation (79), whereas degradation of TRPC1 by caspase-11 controls IL-1β release from macrophages (150).

Stretch-Activated, Nonselective Cation Channels

Piezo 1 and Piezo 2.

Mechanosensitive, stretch-activated, nonselective cation channels (NSCCs) assemble as tetramers with a Ca²⁺>Na⁺=K⁺ selectivity and constitute the largest ion channels complexes in nature (MmPiezo1 assembles as a 1.2 million Dalton homo-oligodimer). This fairly new group of Piezo 1 and 2 channels conducts large whole cell currents upon pressure stimulation, is inhibited by Gd³⁺ and ruthenium red, and shows rapid activation kinetics and conformational change-dependent gating, which are all typical features of SACs (46). Their role in mechanosensing and mechanotransduction is nowadays widely recognized and their activation leads to a 17- to 300-fold increase in mechanically activated whole cell currents (45). In lung epithelial cells, Piezo 1 deficiency decreased cell adhesion and increased cell migration, which could impact epithelial wound healing in ARDS (124, 195).

Other Ca²⁺-permeable NSCCs.

Mechanosensitive, Ca²⁺-permeable NSCCs act as volume sensors and volume regulators in many body tissues and are blocked by Gd³⁺, a typical feature of SACs (86). Inflammatory and epithelial cell volume changes, commonly seen in ARDS, can trigger cellular dysfunction, apoptosis, necrosis, cell detachment, and abnormal cell proliferation (86, 127, 131, 198). Inhibition of stretch-activated NSCCs could be particularly important in the exudative phase of ARDS since in rat lungs Ca²⁺ influx via these channels regulates microvascular permeability (142). Other features relevant to ARDS pathophysiology include their activation by oxidative stress resulting in cell necrosis (17) and their role as mechanotransducers (via stretch-induced cell adhesion kinase-β/CAK-β) in vascular smooth muscle cells (92).

Na²⁺- and K⁺-permeable-permeable NSCCs.

In AT II cells high level expression of the classical, stretch-sensitive, and Na⁺-selective ENaC channel depends heavily on culture conditions, including the presence of steroid hormones, permeable supports, and an air-liquid interface (93). Without these culture conditions this highly Na⁺-selective channel (HSC; Na⁺:K⁺ selectivity ratio >80, conductance 4–6pS) is not observed in AT II cells and instead a 21pS-conductance, nonselective cation channel (NSC; Na⁺:K⁺ selectivity ratio ∼1) can be detected. Furthermore, the ENaC α subunit alone (without β and γ) can act as a stretch-activated, nonselective cation channel and mechanosensor (101). The main importance of these channels in ARDS could relate to the exudative phase and alveolar fluid clearance and may support steroid therapy for ARDS patients. In fact, dexamethasone reduces pulmonary edema formation by upregulating α-ENaC in hydrogen sulfide-induced ARDS (97).

Stretch-Activated Cl⁻ Channels

Mechanosensitive Cl⁻ channels were described in the late 1990s in vascular endothelial cells and neurons where they counteract K⁺ efflux-induced membrane hyperpolarization and promote cell spreading (11, 57). Nowadays the cystic fibrosis transmembrane conductance regulator (CFTR) is the best characterized mechanosensing ion channel and is known to be expressed in AT I, AT II, and immune cells (63, 98, 135). Since its discovery in 1989 by Riordan et al. (156), an incredible amount of information has been published about the functional properties of CFTR and its regulation. However, the field entered a new era in 2010 with the discovery of the stretch-activated gating properties of CFTR (221). Zhang et al. (221) showed in non-CF, human airway epithelial cells (Calu-3) that the normal cellular machinery that controls CFTR can be completely bypassed simply by changes in membrane tension. This mechanosensitivity was confirmed in cell-free, excised membrane patches, demonstrating that neither an intact cell nor ATP was required for this effect. However, since membrane tension was less potent at activating CFTR in these cell-free experiments, an intracellular factor or cytoskeletal contributions to the mechanosensing properties could not be excluded (75). This discovery sparked new enthusiasm in studying CFTR as a regulator of acute lung injury and ARDS, as this channel could affect not only the exudative phase of ARDS by regulating alveolar fluid clearance but also the proliferative and fibrotic phases via regulation of inflammatory cell activation. In fact, deficiency of CFTR in neutrophils (181) and platelets (222) accelerates LPS-induced ARDS development. In contrast, upregulation of CFTR by lipoxin A4, an endogenous “breaking signal” for inflammation, improves alveolar fluid clearance (216). Of importance, the regulation of alveolar fluid clearance by CFTR is a highly complex process (21). Although most would agree that in general CFTR activation promotes fluid clearance (62), Solymosi et al. (178) showed that under conditions of hydrostatic pressure CFTR-dependent fluid flux can reverse from an absorptive into a secretory mode, thus promoting pulmonary edema formation. These phenomena may, at least in part, be explained by the differential expression of cotransporters in AT I vs. AT II cells, which may determine the ultimate direction of fluid flux (82). AT I cells preferentially express the K⁺-Cl⁻ cotransporter, resulting in apical to basolateral Cl⁻ movement and fluid absorption, while AT II cells mostly express the Na⁺-K⁺-2Cl⁻ cotransporter, resulting in Cl⁻ internalization, which creates the driving force for Cl⁻ (and subsequent fluid) secretion upon CFTR activation. In ARDS, destruction of AT I cells could thus contribute to decreased alveolar fluid clearance (105), while expansion of the AT II cell pool in an effort to replace injured epithelial areas may lead to further fluid secretion. Interestingly, in influenza models of ARDS, CFTR inhibition improves pulmonary edema formation and prevents ARDS development by upregulating macrophage and IL-6 responses (2, 211), although CFTR deficiency at the same time can also activate proinflammatory JAK3-dependent pathways (34). The astonishing complexity of CFTR regulation is further evidenced by studies reporting that CFTR inhibition promotes lung development and protects lung epithelial cells from ROS-induced damage by regulating intracellular glutathione levels, but also increases p38- and ERK-mediated IL-6 and IL-8 release and stimulate neutrophil infiltration (22, 27). Therefore, CFTR acts as an oxygen-sensitive mechanosensor and mechanotransducer in the lung with wide-ranging effects on alveolar fluid clearance and immune cell function. In addition, CFTR can directly regulate other Cl⁻ channels (ClC3B, VRAC, CaCl), K⁺ channels (Kir, KvLQT1, ROMK2), Na⁺ channels (ENaC), water channels (aquaporins), ATP transporters, and antiporters (Na⁺/H⁺ and Cl⁻/HCO⁻) (138, 170). Volume-regulated anion channels (VRACs) are other mechanosensitive Cl⁻ channels that partially overlap with the ClC Cl⁻ channel family (95, 138). The importance of VRACs in ARDS is supported by their role as mechanosensors in endothelial cells, cell apoptosis, and ATP, lactate, and HCO₃⁻ transport (145).

Limitations, Challenges, and Future Research

Challenges inherent to ARDS pathophysiology.

Similar to all other treatment options for ARDS, the multifactorial etiology and pathophysiological heterogeneity of ARDS represents therapeutic challenges (157, 207). Nonuniform alveolar expansion creates multiple different biophysical microenvironments within an ARDS lung with potentially opposing stretch forces (atelectatic vs. emphysematous areas, edematous vs. nonedematous areas, fibrotic vs. nonfibrotic areas, more vs. fewer cellular infiltrates) (148). Although in ARDS alveolar cells are simultaneously exposed to a combination of stretch forces produced by MV, pulmonary blood flow, and local changes in surface tension caused by surfactant inactivation, the ramifications of this complex, heterogeneous microenvironment (mechanical stretch, hyperoxia, and inflammatory mediators) on SAC expression and function are now being addressed (13, 76, 174). In addition, clinical interventions such as neuromuscular blockade or prone positioning of patients can also affect stretch forces in a patient's lung. Future studies should also explore the possibility of inhibiting the ATP-driven inflammasome activation by modulating SAC expression and activity. We have already found TREK-1 expression on alveolar and interstitial macrophages (174) and inhibition of TRP and ENaC channels abolishes stretch-induced ATP release from epithelial cells (140). Another important area for research is the effect of metabolic acidosis on SAC function, which can be profound in ARDS patients suffering from multiorgan failure. In fact, gating of many SACs is pH dependent as seen with K2P and BK_Ca channels (41, 84). In addition, the role of calpain, which is increased during MV (113), in cleaving Na_v should be further explored. It is conceivable that stabilization or overexpression of Na_v could inhibit calpain-induced lung injury. As with all known ion channel modulators, drug specificity issues will continue to constitute challenges for future studies since many SACs are quite ubiquitously expressed. Nevertheless, specific inhibitors for some of the most promising SAC targets such as TREK-1 and TRPV4 are already commercially available and have been tested in humans, although not in the context of ARDS. The TREK-1 inhibitor Spadin is currently being used as an antidepressant (28), while TRPV4 inhibitors (GSK193874, GSK2263095) are used for prevention and resolution of cardiogenic pulmonary edema (187). Given the varying degrees of effect of TRPV4 inhibitors in different ARDS models (8, 218), future studies will need to determine which ARDS patients (possibly based on ARDS etiology) may benefit most from TRPV4 inhibition.

Difficulties in measuring stretch forces in vivo.

Despite tremendous technological advances, we are still limited in our ability to quantify alveolar wall stress in situ, especially when it comes to measuring stretch transduction between interacting epithelial/endothelial and inflammatory/immune cells (115). These are important considerations when translating the stretch forces required to activate a SAC in vitro to an in vivo system. If supraphysiological stretch forces are necessary to alter a given channel's function, then stretch activation may well be a biophysical property of this channel but potentially of no physiological relevance. Another important question for future ARDS research is how SACs react to compression forces, such as hydrostatic pressure from pulmonary edema or alveolar compression from atelectasis, since SAC activation by hydrostatic pressure has been reported in other organs (90, 169). In addition to cyclic stretch and compression forces, the effects of tonic stretch on SACs in the lung need further investigation. High positive end-expiratory pressure is an important component of ARDS treatment and exposes lungs to high tonic stretch levels. Tonic stretch forces are also caused by APRV-type MV, often employed in ARDS patients (60). Importantly, in AECs tonic stretch induces unfolding and insertion of additional phospholipids into the plasma membrane and reduces membrane tension, which could modify SAC activation during MV (64).

Patch-clamp controversies.

Whole cell and single channel recordings using multiple variations of the patch-clamping technique are the gold standard of defining the biophysical profiles of SACs. Of importance, since a patch recording is derived from a small area of a single cell under gigaohm seal conditions (necessary to minimize leak currents), it is challenging to extrapolate such a one-time measurement to the highly dynamic and heterogeneous microenvironment of ARDS lungs. Furthermore, during patch formation the outer part of the lipid bilayer is more stretched than the inner part and, depending on the location of the gating sensor within a channel, it may affect its mechanosensitivity (166). Nevertheless, despite the recent expansion of fluorescent tracer probes to detect ion currents, patch clamp will remain the gold standard for ion current measurements for the foreseeable future due to its vast superiority in isolating and delineating the biophysical properties of a channel. Further technological advances allowing to patch single cells within a native lung tissue slice have already opened exciting new opportunities to study cell currents in more physiological (and pathological) environments (80).

From heterologous expression systems to in vivo ARDS models.

Heterologous expression systems, such as the Xenopus laevis oocyte, remain an invaluable tool for electrophysiologists to study the details of an ion channel's biophysical properties in a controlled environment, especially to initially characterize newer groups of ion channels such as K2P channels or Piezos. Species differences in that regard are of limited concern since ion channel sequences are usually well preserved throughout nature. Limitations of the oocyte system include the potential lack of channel insertion into the cytoskeleton and cytoskeletal rearrangements caused by channel overexpression (26). Heterologous expression systems have been indispensable in determining the biophysical properties of ENaC but its subunit composition in lung-resident and inflammatory cells needs to be confirmed. In most cell types, the α subunit forms the core of the channel and γ-ENaC seems important for alveolar fluid clearance (14). The contribution and combination of other subunits (β, γ, δ) to create a mechanosensitive channel in the lung are important for the design of targeted therapeutic strategies against specific ENaC subunits rather than altering global fluid homeostasis by modifying all ENaC channels in ARDS lungs. The development of genetic technology such as the CRISPR systems now allows us to develop lung cell type-specific conditional knockout models in a much cheaper and less time-consuming manner. These approaches clearly aid us in clarifying some of the inconsistencies in SAC characteristics found between cultured cell models and in vivo systems, as described for example for TREK-1 (171, 174).

Conclusion

As efficient mechanosensors, mechanotransducers, and signal integrators, SACs are prime candidates for much-needed new therapeutic approaches against ARDS. A better integration of the vast amounts of basic science and clinical knowledge concerning SAC physiology and ARDS pathology will undoubtedly revitalize old and spark new collaborations between electrophysiologists and clinician-scientists. The fruits of this work will eventually allow us to counteract the life-threatening inflammatory processes in ARDS lungs at the earliest stages, rather than attempting to ameliorate already existing lung damage.

GRANTS

This research was supported by National Heart, Lung, and Blood Institute Grant 7KO8HL118118-03.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

A.S. prepared figures; A.S. drafted manuscript; A.S. edited and revised manuscript; A.S. approved final version of manuscript.