Plasma membrane wounding and repair in pulmonary diseases

Abstract

Various pathophysiological conditions such as surfactant dysfunction, mechanical ventilation, inflammation, pathogen products, environmental exposures, and gastric acid aspiration stress lung cells, and the compromise of plasma membranes occurs as a result. The mechanisms necessary for cells to repair plasma membrane defects have been extensively investigated in the last two decades, and some of these key repair mechanisms are also shown to occur following lung cell injury. Because it was theorized that lung wounding and repair are involved in the pathogenesis of acute respiratory distress syndrome (ARDS) and idiopathic pulmonary fibrosis (IPF), in this review, we summarized the experimental evidence of lung cell injury in these two devastating syndromes and discuss relevant genetic, physical, and biological injury mechanisms, as well as mechanisms used by lung cells for cell survival and membrane repair. Finally, we discuss relevant signaling pathways that may be activated by chronic or repeated lung cell injury as an extension of our cell injury and repair focus in this review. We hope that a holistic view of injurious stimuli relevant for ARDS and IPF could lead to updated experimental models. In addition, parallel discussion of membrane repair mechanisms in lung cells and injury-activated signaling pathways would encourage research to bridge gaps in current knowledge. Indeed, deep understanding of lung cell wounding and repair, and discovery of relevant repair moieties for lung cells, should inspire the development of new therapies that are likely preventive and broadly effective for targeting injurious pulmonary diseases.

the plasma membrane of eukaryotic cells is composed of a phospholipid bilayer decorated with proteins and transporters that are selectively permeable to gases, ions, nutrients, hormones, solid particles, and macromolecules in a tightly regulated fashion. Various functional tasks stress living cells in tissue on a daily basis, such as abrasion and acid erosion in the gastrointestinal tract, stretching and compression in the muscle tissues, and hydrostatic pressure in the cardiovascular system. In addition, bacterial toxins, inflammatory cytokines, immune complexes, and harmful environmental exposures directly damage the plasma membrane or impede the cellular repair process. Severe or irreversible injuries to the plasma membrane result in intracellular components escaping the cell, and poisonous extracellular components flushing into the cell, leading to compromised cellular functions or even cell death. Thus, similar to the need for the well-elucidated DNA repair mechanisms (284), intrinsic membrane repair mechanisms to rescue the deteriorating fate of a broken cell is an evolutionally favorable and likely conserved trait among different cell types, tissues, and species (162, 174, 175).

The concept of membrane repair was put forward by McNeil and colleagues in the 1990s (245), and extensive studies were conducted to reveal the cellular events of plasma membrane repair (48, 173). Generally, it is believed that lateral flow of the lipid bilayer alone, driven by thermodynamic gradient, is sufficient to reseal small membrane disruptions (nm range) in nucleated cells and most membrane damage in nonnucleated cells. On the other hand, exocytotic lipid trafficking plus lateral lipid flow could prevent permanent cell damage upon membrane disruptions of the micromolar range while remedy of even larger membrane wounds requires fusion of exocytotic lipid vesicles into a repair “patch” through highly choreographed steps. Furthermore, it has been shown that removal of the disorganized plasma membrane wounds either through endocytosis (50, 242) or exocytotic budding (114) is also critical for effective membrane repair. Meanwhile, depolymerization of subcortical cytoskeleton at the time of exocytotic vesicle trafficking and repolymerization afterward to restore cell morphology are shown to be integral steps of the membrane repair process (83).

Lung cells are exposed to both environmental insults and physical-biological stresses and thus are highly susceptible to plasma membrane injuries. Pertinent to the focus on pulmonary diseases in this review, disruptions of lung cell membrane following mechanical stresses were first captured on transmission electron microscopic (EM) images in the 1990s (52, 63). Later studies established an association between lung cell injury and death with pathogenesis of injurious pulmonary diseases, including acute respiratory distress syndrome (ARDS) (58, 197) and idiopathic pulmonary fibrosis (IPF) (33). Currently, the study of how lost cells replenish from progenitor cells has become a trendy field in lung biology. However, the mechanisms of membrane repair at molecular and cellular levels are underappreciated and understudied in pulmonary diseases. Here we attempt to summarize relevant literature revealing the mechanisms of lung cell wounding and repair with a focus on biophysical injury mechanisms and discuss the therapeutic perspectives of targeting lung cell repair for the treatment of ARDS and IPF.

Two Devastating Lung Injurious Diseases: ARDS and IPF

ARDS and IPF are two distressing pulmonary diseases where cell and tissue injury may be part of their possible etiology. ARDS is a syndrome associated with diffuse alveolar damage, severe hypoxemia, noncardiogenic pulmonary edema, and bilateral pulmonary infiltrate (232). The 2012 Berlin definition (208) replaced the original American-European Consensus Conference definition of acute lung injury and ARDS (21) and categorized both acute lung injury and ARDS into mild, moderate, and severe ARDS based on the level of hypoxemia. ARDS has a reported incidence of 86 per 100,000 people per year, and there are ~190,000 cases of ARDS in the United States (219). Despite the enormous efforts to improve and develop new therapeutics for the treatment of ARDS based on emerging knowledge on the pathogenic mechanisms of ARDS, most clinical trials failed with the exception that low tidal volume ventilation therapy showed benefits on early and late mortality (38, 246a). In addition, early application of prolonged prone positioning (87), or neuromuscular blocker cisatracurium besylate (199), was also reported to reduce 28- and 60-day mortality of the ARDS patients. As a result, the mortality rate for ARDS is still as high as 21~58% (11), and significant comorbidities occur among survivors (91, 178). Thus, new therapeutic target(s) covering broad pathology in ARDS is in high demand.

IPF is the most common form of idiopathic interstitial lung disease that causes lung tissue damage and scarring (239). It is characterized by progressive loss of respiratory lobules due to fibrosis and remodeling of alveolar units, manifested as the decline of pulmonary function in the clinic (56). IPF affects ~500,000 patients in Europe and the United States (53) and is the most fatal fibrotic disease, since it only has a median survival of three to five years after diagnosis. There were no treatments for IPF until the recent Food and Drug Administration (FDA) approval of two drugs, i.e., pirfenidone (127) and nintedanib (212). Significantly, pirfenidone reduced the percentage of patients with disease progression from 31.8 to 16.5% and increased the percentage of patients with stable pulmonary function from 9.7 to 22.7% at 52 wk of treatment. The all-cause mortality rate of patients was also reduced from 7.2 to 4%. Similarly, nintedanib reduced pulmonary function decline by ~39%, but there was no significant improvement on mortality rate. Despite that this great breakthrough ended the era of no treatment for IPF, the fact that these two drugs only slow down disease progression in some patients warrants further development of combinational therapies for the treatment of IPF.

Experimental Evidence of Lung Cell Injury in ARDS and IPF

Adult lung epithelial cells have relatively low cell turnover rate and facultative postinjury regeneration capacity (132). Its main physiological function of gas exchange is fulfilled by a specialized structure, i.e., the blood-gas barrier, that is composed of a thin layer of basal membrane (~0.3 $\stackrel{․}{\mathrm{A}}$) and three types of cells [the thin and flat type I alveolar epithelial cells (ATI) covering 90% of the alveolar lining surface, the corner-residing cuboidal type II alveolar epithelial cells (ATII) producing surfactant, and the endothelial cells lining the microvasculature wall]. Interstitium is the space between endothelium and alveolar basement membrane at the vascular side where interstitial fibroblasts and macrophages reside. Although alveolar epithelial cells have membrane lipid folds that may partially mitigate the likelihood of cell breakage (258) at noninjurious lung volume changes, or fix small disruptions at the plasma membrane (50), cell stress failures do occur at high lung parenchymal strains (52, 73, 74) or when exposed to excessive biotoxins. As a result, alveolar epithelial and endothelial cells are wounded and consequently trigger cell death (150) without effective repairs. In addition, the plasma membrane bilayer is highly susceptible to injury by free radicals, so various internal and external oxidizing insults (95, 103, 223, 290), such as radiation, ischemic reperfusion, hyperoxia, ozone, and heavy metals, etc., cause compromised membrane integrity and cell death partially through increasing oxidative stresses.

Studies have shown that epithelial and endothelial cell wounding occurs upon physical stresses to the lung (52, 99). For example, Costello et al. applied high hydrostatic pressure to lungs of rabbits, and membrane porations and breakages affecting alveolar epithelial cells and endothelial cells were captured on transmission EM (52, 63, 88). Hall et al. showed that alveolar cell blebbing occurs during reperfusion of transplanted rat lungs by EM (88). Furthermore, studies by Gajic et al. (74) and Kim et al. (125) demonstrated reversible and irreversible forms of alveolar cell membrane stress failure in situ following stretch injury on optical lung slices. It is worth noting that mechanical stresses not only cause compromised cell integrity but also alter cellular architecture. For example, Dreyfuss et al. (61) caught endothelial blebs, and detachment from the basement membrane, following 5–10 min of short ventilation with 45 cmH₂O peak airway pressure in rats. In addition, early studies by Egan et al. (64, 65) showed that mechanical stress caused an increase in alveolar permeability, likely due to perturbation of intercellular tight-junction and alveolar-capillary barrier structures in endothelial cells (37). These structure perturbations all directly and indirectly contribute to the conventional injurious manifestations of the lung such as histological alteration, lung edema, and levels of protein and lactate dehydrogenase, etc., in bronchoalveolar lavage fluid (BALF).

In injurious pulmonary diseases, cell death may occur through apoptosis, necrosis, necroptosis, and other undefined mechanism(s), following plasma membrane injury. In ARDS, cell necrosis may be a direct consequence of bacteria toxin (183) and mechanical stress (125, 204). Morphological characteristics of apoptosis (12) and DNA fragmentation (16) were also detected in alveolar epithelial cells of the ARDS patients. Ligands of the apoptotic pathway were shown to accumulate in edema fluid and lung tissues (2), as well as BALF of the ARDS patients (172). Complicating factors often contribute to the extent of cell death in ARDS. For example, increased lung cell apoptosis was found under injurious mechanical ventilation in conjunction with hyperoxia (27) or hypoxia (135), suggesting that a cascade of cell injury and cell death may contribute to the loss of pulmonary function in ARDS. Necrosis is featured by plasma membrane disruption and release of intracellular components, including endogenous “danger” signal molecules that cause immune responses and pathological changes (76). Delayed clearance of apoptotic cells may induce secondary necrosis with their intracellular contents released to induce subsequent immune responses, so the maintenance of cell membrane integrity is crucial for avoiding undesired immunity (168, 202). Furthermore, repeated microinjury to alveolar cells is viewed as an etiological factor for pulmonary fibrosis (22, 225) that causes epithelial cell loss on one hand and stimulates uncontrolled fibroblast proliferation on the other hand. For example, changes in function and morphology of alveolar epithelial cells were consistent features of early phase pulmonary fibrosis in human patients (42, 121), and many studies (15, 138, 203) detected apoptotic markers at proximal and distal epithelial cells of the IPF lung samples. As detailed in the section below, chronic epithelial cell injury and abnormal repair may be the initiating factor to stimulate mesenchymal cell proliferation of fibrogenesis, eventually leading to IPF.

Pathogenic Mechanisms of Cell Wounding in ARDS and IPF

The pathophysiological hallmarks of ARDS are diffuse alveolar damage, surfactant dysfunction, and activation of inflammatory responses (289). Injury to pulmonary epithelial and endothelial cells was recognized as a contributing factor for the pathogenesis of ARDS (58, 197), particularly when mechanical ventilation is used (52). Progressive lung cell injury and cell death in ARDS are likely initiated from heterogeneity in lung parenchymal stiffness (215, 241) resulting from local surfactant inactivation and inflammation, leading to regional acinar overdistention. Overdistension causes vascular leakiness by breaking or remodeling tight junctions. The resulting alveolar edema further inactivates surfactant, which in turn contributes to epithelial injury by interfacial stress, starting a vicious cycle of progressive lung injury. In many cases, fibrin and collagen deposition originated from denuded basement membrane may serve as the sites of fibrotic signaling cascades for the future development of pulmonary fibrosis (272).

On the other hand, although the exact causes for IPF remain elusive and its pathogenesis incompletely understood, a hallmark of this disease is recurrent or persistent alveolar epithelial injury and epithelial cell apoptosis (15, 138, 203). Prolonged cell injury in conjunction with aberrant epithelial cell repair stimulates fibrogenesis of mesenchymal cells through inappropriate epithelium and mesenchyme cross talk, whereas mesenchymal-like cells may also feed back on epithelial cells to promote epithelial-mesenchymal transition (EMT) (118), forming a self-amplifying loop of progressive fibrosis (56). Furthermore, it is hypothesized that sudden increases in epithelial injury may be one of the underlying causes for acute exacerbation of IPF (117). The injurious insults associated with IPF are thought to be diverse in nature. In addition to well-recognized genetic factors for IPF susceptibility (137), many biological factors and environmental exposures such as virus infection, gastresophageal reflux, radiation, cigarette smoke, and exposure to dusts, fumes, or chemicals had been reported to be risk factors for IPF (17, 78). Prolonged mechanical ventilation may also lead to development of local fibrotic changes in the lung (256). In the following section, we will elaborate on the biophysical, inflammatory, and environmental mechanisms through which each of the above pathogenic factors injures lung cells, and thus contributes to the development of ARDS and IPF (Fig. 1). It is important to remember that these pathogenic factors are interdependent of each other.

Fig. 1.

Interdependence of biophysical injury stimuli. Internal and external pathogenic factors, including surfactant dysregulation, regional distention, proinflammatory cytokines, bacterial toxins, viral infection, gastric acid aspiration, and cigarette smoking, collectively and interdependently injure the epithelial-endothelial barrier and further induce the pathogenesis of acute respiratory distress syndrome (ARDS) and idiopathic pulmonary fibrosis (IPF). ROS, reactive oxygen species; RNS, reactive nitrogen species; ER, endoplasmic reticulum. Unless an arrow pointing down is present, the contents of each box are increased by its predecessor.

Surfactant dysfunction.

Surfactant is a lipoprotein complex that reduces surface tension and assists with innate host defense of the lung (254). It is composed of ~90% lipids and ~10% surfactant-associated proteins (SPs), including SP-A, -B, -C, and -D. SPs are mainly produced by ATII cells, and surfactant dysfunction is viewed as an important contributing factor in the complex pathophysiology of ARDS and IPF. Indeed, various pathological changes occurred during ARDS and IPF, e.g., plasma proteins in extravasated edema fluid, proteolytic enzymes from recruited immune cells, or free oxygen radicals can all inactivate surfactant in the airspace (216). In addition, study suggests that surfactant dysfunction can be induced by transforming growth factor (TGF)-β overexpression before the presence of profibrotic lung remodeling (157). Mechanical ventilation has also been shown to alter the organization and function of surfactant complex (254).

The overall biophysical consequence of surfactant dysfunction is heterogeneous compliance at local lung regions, in conjunction with the loss of lung unit due to local edema, which causes lung cell injury as described in Overdistension and interfacial stress. In addition, it was shown that surfactant dysfunction contributes to alveolar epithelial cell apoptosis (72) through upregulation of endoplasmic reticulum (ER) stress (137) as detailed in ER stress. Surfactant has also been shown to have a protective effect on the barrier function (134), and thus surfactant dysfunction may aggravate severity of lung edema and inflammatory cell infiltration. Furthermore, because a previous report showed that loss of SP-D may enhance sensitivity of the lung to hyperoxia-, ozone-, and bleomycin-induced injury through inhibiting inducible nitric oxide synthase-mediated reactive nitrogen species production (169), it is reasonable that surfactant dysfunction may substantially increase susceptibility of lung cells to injurious stimuli. Interestingly, large genome-wide sequencing studies identified an association between mutations in SFTPC and SFTPA2, the gene encoding SP-C and -A2, and familial IPF (137), which accounted for up to 20% of IPF cases (137). Nevertheless, lack of functional surfactant alone does not recapitulate the fibrotic changes of IPF but rather aggravates bleomycin-induced lung pathology in mouse models (8, 142), supporting that surfactant dysfunction is mainly a disease modifier rather than a direct cause of IPF.

Overdistension and interfacial stress.

It is known that biophysical mechanisms alone can cause physiologically significant cellular injury in the lung. Approximately 19% of mechanically ventilated patients develop de novo ARDS (112). In addition, in absence of previous lung injuries, mechanical ventilation could induce lung pathologies in animal models that are similar to those observed in ARDS patients (277). Indeed, a decline in hospital-acquired ARDS was reported since the introduction of lower tidal volume ventilation strategy. Recently, it was also demonstrated that even spontaneous breathing with increased tidal volumes, particularly when associated with mechanical ventilation, could injure normal lungs and exacerbate edema and inflammation in diseased lungs (26). The biophysical mechanisms of lung cell injury, in particular overdistension and interfacial stress, in disease states such as ARDS and IPF can be summarized as follows. Although both biophysical mechanisms contribute to the pathogenesis of ARDS (241), we speculate that overdistension-induced lung injury may be more relevant for IPF due to lack of pulmonary edema secondary to poor blood circulation in scarred lung tissues.

In the uninjured lung, alveolar walls more or less unfold and refold during normal breathing. Neither epithelial nor endothelial cells undergo large elastic deformations as alveolar surface area changes under this condition. Only at lung volumes approaching total lung capacity are alveolar walls and adherent cells stretched to a significant degree, leading to an increase in tensile stress on cell-cell junctions. However, the above elastic buffering capacity of lung parenchyma is greatly reduced in fibrotic lungs, rendering susceptibility to tensile stress. If tidal volumes remain substantially elevated for extended periods of time, this tensile stress triggers the remodeling or frank breakage of cell-cell junctions, which in turn is associated with an increase in paracellular and possibly transcellular fluid flux (overdistension-induced lung injury). In ARDS, the egress of proteinaceous fluid in the alveolar space in conjunction with large alveolar area oscillations impair surfactant function and thereby raise local surface tension. An increase in alveolar surface tension represents a collapse force that decreases lung compliance. Similar changes in lung compliance occur in IPF lungs largely due to surfactant dysfunction and lung scarring. Increased alveolar surface tension further stresses the cell and tissue attachments at the boundary of the collapsing lung segments and augments the hydrostatic pressure gradient favoring cell injury and alveolar flooding. As long as the fluid in the airspace remains liquid (in contrast to an organized semisolid), breathing will cause the agitation and movement of air-liquid interfaces along small airways and airspaces. In vitro microchannel experiments have shown that the stresses associated with such movements are large enough to deform adjacent epithelial cells and cause plasma membrane bleb formation and bleb rupture (198) (interfacial stress). Moreover, the agitation of alveolar exudate during breathing and thus mixing of gas and fluid in the presence of surfactant favors the formation of foam. Foam is unstable and releases energy during foam fracture, which again if sufficiently large would wound adjacent cells (81) (again, interfacial stress). It should be emphasized that, when presented, surfactant dysfunction, alveolar overdistension, small-scale heterogeneity of the lung mechanical properties, and the cyclic recruitment and derecruitment of lung unit are interdependent rather than distinct and independent biophysical injury mechanisms (241). The ARDS lung is particularly vulnerable to all above biophysical injury mechanisms because impaired barrier properties are one of the hallmarks of the syndrome. Moreover, the need to support gas exchange through oxygen supplementation (119) and to support ventilation by mechanical means (112) increases the risk of iatrogenic lung insults.

Proinflammatory cytokines.

Proinflammatory cytokine tumor necrosis factor-α (TNF-α), interleukin (IL)-1, IL-6, and IL-8 levels were detected in BALF from early and late-stage ARDS patients (176). In addition, IL-17A and IL-1β were elevated in BALF of IPF patients (276), whereas expression of IL-17A and interferon-γ were found in lungs of bleomycin-injected mice. Proinflammatory cytokines often turn on downstream signaling pathways that either produce membrane-injury products or cause compromised plasma membrane integrity as part of the cell death mechanism. For example, TNF can mediate cell injury and death through well-known and newly emerging mechanisms. Typically, TNF binds to either TNF receptor (TNFR) 1 or TNFR2 and activates nuclear factor-κB (NF-κB), Jun NH₂-terminal kinases, or p38 mitogen-activated protein kinases to escalate inflammatory responses (265) and produce mediators with high plasma membrane and systemic toxicity such as reactive oxygen species (ROS) and reactive nitrogen species (23). If NF-κB activation is inhibited by cytosolic factors, binding of TNF to TNFR1 may also trigger cell death pathways (89) through apoptosis in cells with low levels of receptor-interacting protein (RIP) 3 or necroptosis in cells with a sufficient amount of RIP3 (89), the latter being a particular type of cell death mode characterized by cytoplasm granulation, cell swelling, and loss of plasma membrane integrity (187). Furthermore, proinflammatory chemokines attract immune cells, including neutrophils to infiltrate the lung, and persistent neutrophil inflammation is an important indicator of tissue injury (75) and multiorgan failure in ARDS (170), since sequestered neutrophils may injure epithelial and endothelial layers through degranulation, oxidative burst, and extracellular traps (275). Indeed, neutrophil transmigration across lung epithelial cell monolayer causes a denudation type of “wound” (288), whereas inhibition of CXC/C-X-C chemokine receptor 2 interaction significantly reduced neutrophil sequestration and lung injury in mice following injurious ventilation (20). Increased neutrophil levels in IPF lungs (126) were also shown to correlate with the alveolar epithelial injury marker cytokeratin 19 (107) as well as early mortality of IPF patients (126), and neutrophil elastase inhibitor ameliorated bleomycin-induced pulmonary fibrosis in mice (244).

On the other hand, cell injury and cell death in ARDS and IPF are capable of stimulating immune responses and hence production of more proinflammatory cytokines. Slutsky and colleagues put forward the notion of “biotrauma” based on their experimental findings that mechanical ventilation causes release of proinflammatory mediators of the innate immunity into the airspace (231, 248), and Andrews hypothesizes that prolonged injury may provide both the primary antigens and immune adjuvants to sufficiently activate the adaptive immunity (7). Although apoptosis largely maintains cell membrane integrity, delayed clearance of apoptotic cells may induce secondary necrosis with the release of intracellular contents to stimulate immune response (168, 202). Necrosis is featured by plasma membrane disruption, and release of intracellular components, including endogenous danger signal molecules, elicits immune responses and pathological changes (76). Similar to necrosis, necroptotic cells swell and rupture, leading to the release of damage-associated molecular patterns (77).

Bacterial toxins.

Nearly all types of bacteria produce pore-forming toxins (PFTs) that cause damage to a plasma membrane by direct pore formation. Bacterial pneumonia and sepsis are the leading cause and frequent complication of ARDS (62). Cell lysis by PFTs may be an important pathogenic mechanism for epithelial and endothelial barrier disruption and infection spreading in ARDS of infectious origins (104, 159). In addition, evidence showed that bacterial infection may be a mortality risk (97), and treatment by antibiotic septrin led to a reduction in infections and mortality in IPF (229). Bacteria, including Hemophilus, Streptococcus, and Pseudomonas were also isolated from BALF culture of IPF patients (213). Thus direct cell injury by bacterial products is not to be overlooked in injurious pulmonary diseases.

PFTs account for 25~30% cytotoxic bacterial products (104), which are often used by bacteria as virulent factors and drug-resistant mechanisms (104). PFTs are capable of recognizing host plasma membrane components, inserting in plasma membranes and forming stable pores of 2~40 nm in size (235). Take a few common pulmonary pathogens, as examples, the gram-positive cocci Staphylococcus aureus produce α-toxin, which has been shown to create small pores of ~2 nm in diameter on host plasma membrane that are not permeable to Ca²⁺ (98). Consequently, host cells likely use Ca²⁺-independent endocytic and exosomal shedding mechanisms to close these pores in a slow process (6 h or more) for cell survival (98). Studies showed that Rab-5 and Rab-11 may be key regulators of the Ca²⁺-independent removal of bacterial pores (158). S. pyogenes is another gram-positive cocci that secretes PFT streptolysin O (SLO). While creating much larger pores on host membranes than α-toxin (~40 nm), cells mainly use rapid (within minutes) Ca²⁺-dependent mechanisms to remove SLO pores (51, 249). As described in Caveolin1, caveolar endocytosis was shown to be critically important for repairing SLO perforation on cultured cells (50) while another study found that the endosomal sorting complex for transport (ESCRT) machinery was required for Ca²⁺-dependent exosomal shedding of SLO pores (114). P. aeruginosa are opportunistic gram-negative rods causing severe lung infections in immunocompromised patients (221). P. aeruginosa secretes PFTs that create approximately 3.5- to 8-nm pores on host cell membranes (55, 230), which was used as a type III secretion system for injection of other injurious cytotoxins in the host cells (221, 222). Notably, ExoU, a cytotoxin injected through this system, disrupts lipid metabolism and membrane integrity of the targeted lung epithelial cells and macrophages (221). Thus, P. aeruginosa strains possessing ExoU, such as PA103, often cause widespread systemic infection and high mortality of hosts (3, 69). Another common gram-positive rod, Bacillus anthracis, produces a lethal toxin that destroys lung epithelial barrier function and interferes with injury repair by immobilizing the actin and microtubule network (145). In addition to direct injury to host cells, PFTs also induce cell death secondary to K⁺ efflux, Ca²⁺ influx, ATP depletion, mitochondrial damage, and cell swelling (123). PFTs of certain bacteria are also capable of triggering necroptosis, the regulated form of necrosis (187). For example, RIP kinase 3-mediated necroptosis of macrophages was activated by PFTs from pulmonary pathogens such as S. aureus (128), S. pneumonia, and Serratia marcescens (85), whereas inhibition of necroptosis regulators improved the outcome of pneumonia in animal models (85).

Viral infections.

Certain types of community viral pneumonia may evolve into ARDS (165), notably influenza viruses H5N1 and H1N1, and coronavirus, which caused a severe acute respiratory syndrome epidemic in 2003. Nosocomial herpesviridae (HV) infections causing viral pneumonia can similarly evolve into ARDS (165). H5N1 was also used to create an ARDS model in mice (281). With regard to infectious predispositions for IPF, more association studies have pointed to a viral source rather than a bacterial source, e.g., human HV were often detected in serum, BALF, or biopsy of IPF but not non-IPF patients (122, 171, 237). HV infection worsened fibrosis in animals administered with fibrogenic agents such as bleomycin or fluorescein isothiocyanate, although viral infection alone did not trigger fibrosis in young adult mice (156, 186, 238), supporting the hypothesis that HV infection, latent or activated, serves as a “second hit” for developing fibrotic changes in the lung (136). Nevertheless, combination of HV infection and aging did trigger fibrotic changes in experimental mice (191), implying that IPF is a nonspecific injurious disease and that a repair threshold of the lung had been overwhelmed by a combination of injurious insults. Because most viruses enter the host cells via regulated endocytosis (195), plasma membrane integrity of the host cell is usually preserved during viral infections. However, surfactant dysfunction, ER stress, and cell death are common consequences of viral infections, which are all attributing factors for plasma membrane damage and cell injury in IPF (108, 141). In addition, virus infection triggers cytokine and chemokine release from activated innate and adaptive immune cells, including neutrophil-attracting chemokines (10). As a result, proinflammatory cytokines and tissue-infiltrated neutrophils cause lung injury following viral infections. As elaborated above, injured cells in turn release danger signals such as uric acid and activate the inflammasome (79), further promoting inflammatory responses and aggravating lung injury.

Cigarette smoking.

Among environmental risk factors for pulmonary diseases, cigarette smoking is a main one. Study by Ekstrom showed that cumulative exposure to smoke led to increased risk for IPF, and disease severity was linearly associated with the dose of exposure (66). In addition, male gender and occupational exposures (including fibers, fumes, gas, mineral dust, organic dust, and vapors), two other risk factors for IPF, had synergistic effects with cigarette smoking (66). Furthermore, introduction of tobacco exposure biomarker rather than solely relying on smoking history provided more definitive proof for the link between smoking exposures and development of ARDS (32). Cigarette smoking causes a collection of pathophysiological changes such as inducing profibrotic gene expression, weakening of the immune system, and stimulating proinflammatory responses (40, 46, 182). Existing data showed that smoking can increase permeability of alveolar epithelial cells (116) and that of the endothelium (93, 224), and such action was largely attributed to ROS production (41). Specifically, ROS is capable of interacting with plasma membrane lipids and protein moieties and was shown to increase “leaking” of the plasma membrane directly (234). Thus, the oxidative agent H₂O₂ is often used as a reagent to induce nonspecific cell injury via oxidative stress. In addition, ROS can also activate the unfolded protein response (UPR) and induce ER stress and ultimately apoptosis in lung cells (124, 264). Indeed, antioxidants had been a long-term interest for developing new therapies against IPF until the recent clinical trials on acetylcysteine failed to show any significant benefits on lung function among IPF patients (171a), possibly because ROS-induced lung cell injury may only be relevant for a small portion of IPF cases.

Gastric acid aspiration.

Gastric acid aspiration is a common complication of general anesthesia or among unconscious trauma or ICU patients. It is recognized as a direct cause and an independent risk factor for ARDS (206). In addition, multivariate analysis also identified that gastresophageal reflux is associated with a higher risk for IPF (207, 247). Interestingly, acid aspiration has also been linked to acute exacerbation of IPF (47), highlighting the importance of injurious factors in the disease state of IPF. The injurious effects of acid on lung epithelium are physical and can be a direct effect of protons to some extent (274). Studies showed that contact with concentrated acid causes membrane pore formation in alveolar epithelial cells, and this transient poration sufficiently induced an unusual and transitory proinflammatory response of the alveolar epithelium through mediating Ca²⁺ entry and NADPH oxidase 2-dependent H₂O₂ release (274). In addition, acid aspiration may also work in concert with lipopolysaccharide (LPS) to disrupt barrier function and alter lung mechanics (246). Low pH of other origins was shown to similarly injure the lung cells. Doerr et al. showed that hypercapnic acidosis significantly increased the percentage of cell wounding in ventilated lungs (57), and buffering hypercapnic acidosis to neutral pH rescued cell wounding (36), supporting the direct cell lytic effect of protons. To this end, although the cell mending effect of hypercapnic acidosis inspired many studies to explore its therapeutic role for ARDS in vivo, the whole body effects of hypercapnic acidosis are more complex and somewhat contradictory, precluding its possible clinical application (54). Last, other gastric contents may also play a role in lung cell injury. Felder et al. fabricated a microfluidic channel device for exposing A549 cells to acidified pepsin-containing solution, and their results found that, while low pH alone can alter cell morphology, additional pepsin was required to cause cell monolayer disruption, suggesting that multiple erosive factors in the gastric content and weakened barrier mechanisms contributed to lung epithelial wounding much like the formation of gastric ulcers (67).

State of the art and perspective.

In this section, we described clinically identified risk factors for ARDS and IPF and summarized current knowledge on the biophysical mechanisms through which they eventually cause plasma membrane damage and cell death. We aim to collect a holistic view of these injurious stimuli under the premise that cell wounding and repair is critically important for the pathogenesis and possibly prognosis of both ARDS and IPF, with the ultimate goal of developing a versatile adjuvant therapy targeting cell injury. Under that premise, perhaps the current experimental models of ARDS and IPF using a single or two injurious stimuli, such as ventilation, acid aspiration, or bleomycin injection, can be replaced with a model of unified cell wounding. This would be ideal to reduce the vast amount of variations inherent within these models. However, before that, comparison of the extent and characterization of cell wounding following different injurious stimuli needs to be done. In addition, the distribution of cell wounding and injured cell types in the lung would determine disease symptoms and pathology severity. It is known that the base of the lung experiences more strain in a standing individual. Coincidentally, IPF has characteristic lesion distribution at the base of the lung. Thus, it would be very interesting to study how regional strain distribution, or accessibility of injurious stimuli, correlates with cell wounding in the lung. Last, it is intriguing why in most cases cell wounding resolves well, as in ARDS, but others go on to develop fibrotic diseases. Because the current theory is that repetitive injury makes the difference, knowledge on the threshold of injury severity and length to trigger unresolvable fibrosis is of great value.

Repair Mechanisms for Injured Lung Cells

In response to a variety of insults as discussed above, cells engage repair mechanisms to rapidly restore membrane integrity and prevent apoptosis, necrosis, or necroptosis (239, 258). Indeed, defects in membrane repair have been shown to be involved in a number of human diseases, including muscular dystrophy, cardiovascular diseases, and neuronal injury (48). The membrane repair process is important for maintaining lung cell function and alveolar architecture (209). Indeed, ultrastructure microscopy captured gradual restoration of repaired pulmonary vasculature structure following ischemic and reperfusion injury (180). Lung cell regeneration has been an intensive area of research for ARDS and IPF in recent years. Despite studies demonstrating that ATI cells are capable of proliferating in vitro (267), and a defined population of them can self-renew in vivo (109), full replenishment of permanently injured ATI cells also relies on activation of other lung progenitor cells. Although a number of heterogeneous populations of multipotent cells and progenitors have been identified in the lung (252), their intrinsic regeneration ability does not seem to meet the demands of excessive and progressive cell loss in ARDS and IPF. Furthermore, the harsh airspace environment in diseased lungs such as the presence of persistent injurious factors, excessive inflammation, and low oxygenation greatly inhibits cell repair by proliferation (143). Thus, membrane repair mechanisms to rescue the fate of mildly injured cells are necessities for normal function of the lung. Previous studies indicated that lung cells engage plasma membrane repair in a similar manner comparable to cells in other tissue, such as those in striated muscle and kidney (125, 190, 267). Here we will discuss a few of these repair mechanisms that had been studied for lung cell repair (Fig. 2); it is important to understand these repair steps work in an orchestrated fashion to collaboratively rescue the falling fate of a cell.

Fig. 2.

Normal repair mechanisms for lung cells. Lateral flow of lipid bilayer is sufficient to reseal a small membrane wound with diameter <1 µm. Membrane defects are repaired by deformation-induced lipid trafficking (DILT) and lipid vesicles from lysosome. Ca²⁺-dependent exosomal shedding of pore-forming toxins (PFT) pores are removed by endosomal sorting complex for transport (ESCRT) machinery. With facilitation from acid sphingomyelinase (ASM), tripartite motif-containing protein (TRIM) 72 is transported to caveolin (Cav) 1 on the membrane wound site. TRIM72 mediates caveolar endocytosis through its physical interaction with Cav1 in a Src-dependent fashion. TRIM72 polymerizes through its cysteine residues, and nonmuscle myosin type IIA (NMIIA) serves as motor to translocate TRIM72 during membrane repair. Cell membrane repairs are either enhanced or impaired by actin depolymerization. Increasing extracellular ATP inhibits translocation of lysosomal-associated membrane protein 1 (LAMP1) to membrane wound site by deactivation of ATP receptor, purinergic receptor P2Y (P2Y2R), and affects the activity of neighbor cells. Ca²⁺-mediated repair patch exocytosis is essential to remove the lung cell membrane defects with coordination of LAMP1, TRIM72, and de novo lipid synthesis. ASM, acid sphingomyelinase; DILT, deformation-induced lipid trafficking; CSK, subcortical cytoskeleton; PM, plasma membrane; T72, TRIM72.

Ca²⁺ in epithelial cell repair.

As described above, generally thermodynamic lateral flow of the lipid bilayer is sufficient to reseal small membrane disruptions in cells, likely in a Ca²⁺-independent process (158). When larger plasma membrane disruption occurs, three principal steps are engaged to repair the membrane wound. First, the cell senses disruption at the plasma membrane and locates the wound. Second, the plasma membrane signals to mobilize endomembrane lipid vesicles and translocate to the injury site. Third, endocytosis or exocytosis is engaged to remove membrane pores formed by bacterial toxins or large irregular membrane wounds that are hard to patch. Fourth, repair vesicles fuse together with the plasma membrane wound edges to form a repair patch at the injury site. Ca²⁺ had been shown to be a key component for overall repair success and for the implementation of the above processes, including fusion of lipid vesicles, vesicle translocation to wound sites, and removal of membrane pores (173, 245). Consistently, a few essential membrane repair proteins discovered in muscle cells possess Ca²⁺-binding domain and translocate to membrane injury sites in a Ca²⁺-dependent manner, including dysferlin, synaptotagmin, and annexin, etc., which was described in detail in the comprehensive reviews by Blazek et al. (24) and Cooper and McNeil (48). Although only scarce experimental evidence showed that extracellular Ca²⁺ plays a role in repair of lung cells (179, 228), we argue that the importance of Ca²⁺ in membrane repair is applicable to lung cells as well, since key repair steps are shared among different cell types.

Lipid trafficking.

In lung cells, stretching induces lipid trafficking to and from the plasma membrane in a process termed deformation-induced lipid trafficking (DILT) (177, 259). Studies using immortalized lung cell lines and primary alveolar epithelial cells showed that this dynamic lipid trafficking is essential for both injury prevention by membrane unfolding and injury repair by membrane patching (70, 84, 260, 269). Vlahakis et al. showed that stretching of human alveolar epithelial cell line A549 at 25% strain triggers a significant increase in cell surface area and incorporation of intracellular lipid vesicles in the existing plasma membrane as a cytoprotective mechanism (260). In addition, low temperature and cholesterol depletion, two perturbations that affect plasma membrane fluidity and thus lipid trafficking (240), inhibited lipid trafficking to the plasma membrane and compromised the ability of immortalized and primary lung epithelial cells to repair membrane wounds (261). Nevertheless, the injury resistance and repair ability of cell lines and different types of primary lung cells differ substantially (82, 261), suggesting that, although the main repair steps may be conserved, differences in dynamics of the repair process and/or repair machinery may exist among diverse cell types and injury modes. One question under debate is the exact source of endomembrane contributing to patching of wounded plasma membranes (39, 242). The lysosome is the first suggested source of exocytic vesicles for membrane repair (5, 101, 110, 242). However, conflicting results were reported on whether the lysosome fusion inhibitor vacuolin can interfere with membrane repair and Ca²⁺-dependent exocytosis (39, 100), undermining the exclusive role of lysosomes in repair-directed exocytosis. Furthermore, prior studies suggest that endocytosis is also indispensable for successful repair of plasma membrane wounding (50, 106, 242). Studies in primary ATI cells found that caveolar endocytosis, but not clathrin-mediated or fluid-phase endocytosis, was involved in repairing plasma membrane wounding (268). Vaughan et al. also pointed out the potential importance of de novo lipid synthesis at the wound site after cell damage through analysis of lipid domain components (253). It is worth noting that the exocytosis and endocytosis processes in response to cell injury are not disengaged, since it was shown that injured cells release lysosomal enzyme acid sphingomyelinase in an exocytosis process to promote endocytosis during plasma membrane repair (106, 242).

The potential beneficial role of endocytosis inspired an experimental therapy for overdistension-induced lung injury, i.e., hypertonic saline. Aerosolized hypertonic saline is one of the approved therapies for cystic fibrosis that facilitates mucociliary clearance by restoring the liquid layer lining the airways and promoting anti-infective and anti-inflammatory activities (211). It was shown that hypertonic saline could also increase membrane-retrieval response primarily via the caveolar endocytic pathway and thus enhance cellular repair in primary and cultured alveolar epithelial cells (269). In addition, other agents capable of modulating lipids had also been tested in models of membrane injury, notably poloxamer 188 (P188, also referred to as Pluronic F68). P188 is a biocompatible nonionic amphiphilic copolymer that is able to insert lipid bilayer (184) and thus increase the plasma membrane lipid reservoir. It was approved by the FDA ~50 years ago as a therapeutic reagent to reduce blood viscosity for transfusions. Studies showed that P188 was effective in improving cell injury and tissue pathology in muscular dystrophy, heart failure, neurodegenerative disorders, and electroporation damage to cells (184). To our interest, Plataki et al. reported that P188 can promote alveolar cell repair, rendering cell membranes more resilient to mechanical stress and enhancing cell survival in alveolar resident cells in isolated perfused rat lungs subjected to ventilation injury (204). However, this protective effect was not seen in live animals, perhaps reflecting other in vivo injury mechanisms that inhibit cell repair or prolong injury in stress-activated cells.

Subcortical cytoskeleton.

During wounding and repair of plasma membrane wounds, forces like edge energy and adhesive interactions of the lipid bilayer and underlying subcortical cytoskeleton (CSK) are balanced at the wound margins. In addition, endocytic and exocytic mechanisms work together and maintain a favorable force balance between edge energy and plasma membrane-cytoskeleton adhesion (83). As one of the main components to maintain cell stiffness and support cell structure, CSK is a key player in plasma membrane wounding and repair (83). The specific role of the cytoskeleton in wound repair is likely epiphenomena of the complex biology and dependent on the mode of injury. For example, CSK disruption not only leads to a softer cell that renders differential susceptibility to injury by different types of insult, it was also shown to disrupt DILI that is essential for successful cell repair (261). In response to stretch injury, a softer cell is less stressed due to the effective release and spreading of applied lateral loads, but the inhibition on DILI by cytochalasin D to depolymerize actin was overwhelming. Hence, Vlahakis et al. observed increased cell wounding and reduced cell repair in lung epithelial cells under these conditions (261). Nevertheless, the differences in orientation, magnitude, topographical distribution of the stress within the cell, and duration of the stress would likely also modify the cell’s response to deformation. Therefore, because hypertonic saline also increases CSK and plasma membrane adhesion in addition to stimulating caveolar endocytosis (198), it is not surprising that its protective effect would be more effective for stretch-induced lung wounding. On the contrary, because interfacial stress mainly injures the cells by creating local blebs following detachment of plasma membrane from the cytoskeleton network (198), a softer cell would instead be more susceptible to interfacial stress due to air-liquid movements or fracture of liquid bridges. In addition, the role of CSK in cell wounding repair is also dynamic in nature. Specifically, Godin found that, in the first 28 s after needle wounding of primary ATI cells, actin depolymerization occurs, followed by a much longer repolymerization and remodeling phase of the actin network (83). The initial depolymerization of the subcortical cytoskeleton network is thought to remove the barrier for endomembrane lipid trafficking (83) while the reestablishment of the actin network is likely for anchoring repair patches and restoring cell morphology. Furthermore, changes in cytoskeletal structure also affect cell attachment to extracellular matrix structure. Yalcin showed that stabilization of actin improved cell adhesion due to increase in the size and strength of focal adhesions, whereas more detachment is observed in cells with disrupted actin cytoskeletons (282). Thus, when evaluating the impact of altered subcortical cytoskeleton on cell wounding and repair, all above-mentioned factors need to be integrated and carefully considered.

ATP.

Studies on the contribution of ATP release in wound repair have just started. It is known that ATP is released from contracting muscles, inflamed cells (92), and wounded cells (286). Previous studies showed that noninjurious stretching is sufficient to cause low-level ATP release and initiate purinergic signaling in alveolar epithelial cells (19, 86). Belete et al. showed that added extracellular ATP could promote plasma membrane repair while enzymatic removal of extracellular ATP with apyrase inhibited stretch-induced wound repair of alveolar epithelial cells (19). Both extracellular ATP and stretching trigger translocalization of lysosomal-associated membrane protein 1 (LAMP1) to the plasma membrane, indicating endomembrane trafficking and cell repair. ATP signaling transduction to trigger cell repair is dependent on its conventional receptor purinergic receptor P2Y (P2Y) 2 (18), since silencing of P2Y2 receptors inhibited LAMP1 translocalization to the plasma membrane induced by ATP or stretching, as well as repair in A549, rat lung epithelial cells, and primary ATI cells. Nevertheless, the multiple mechanisms that purinergic signal transduction may interact with at large, for example, the mechanotransduction pathways, mechanisms that are parallelly activated by stretching, are not clear due to limited availability of relevant studies. Other advantages of ATP release are its ability to impact neighboring cells via paracrine mechanisms. ATP binding to P2Y2 receptors on neighboring cells induces protein kinase C-dependent oxidative activation of TNF-α-converting enzyme, which releases membrane-bound ligands of epidermal growth factor receptor (EGFR) (250), and initiates EGFR-dependent cell proliferation and migration (250). ATP is also thought to regulate endothelial barrier function. Extracellular ATP produced by inflammation and injured cells could induce a rapid and dose-dependent increase in transendothelial electrical resistance across pulmonary endothelial cells, indicating increased endothelial barrier permeability (163).

Furthermore, released ATP from stressed cells may be able to turn on certain fibrotic pathways (161, 250), and signals conducted by binding of UTP/ATP to P2Y2 receptor, and binding of ADP/UTP/UDP to P2Y6 receptors, were shown to have profibrotic effects (68). Recent study showed that, although ATP release has no effect on mechanical stretch-induced Ca²⁺ influx in human lung fibroblasts (189), ATP could release intracellular Ca²⁺ through ryanodine channels and increase the expression of profibrotic genes such as TGF-β, collagen A1, and fibronectin (111). Therefore, ATP release from stressed cells may be a critical factor that transduces chronic injurious stimuli to fibrotic changes during the pathogenesis of IPF.

Tripartite motif-containing protein 72.

As the subcellular steps of cell repair are largely revealed, investigators spent much effort to identify biologically active protein components that carry out critical functions of each step, such as a wound sensor, a driver for vesicle translocation, and patch fusion. Inspirations from neurotransmitter secretion and vesicle fusion had led to the findings of soluble N-ethylmaleimide-sensitive factor attachment protein receptors, synaptotagmins, and ESCRT as repair proteins, and genetic mutations causing repair-defective muscle diseases have propelled the discovery of new membrane repair proteins such as dysferlin, caveolin (Cav) 3, dystrophin, and annexins, etc., as detailed elsewhere (4, 14, 24, 48, 160, 173, 218). In this review, we will highlight the repair molecules that were indicated in the lung. Tripartite motif-containing protein (TRIM) 72 (also referred to as mitsugumin 53) was first identified from skeletal muscle and was thought to be a muscle-specific protein (31). TRIM proteins were named for their conserved RING, B-box, and coiled-coil domains at the NH₂-terminus, and TRIM72 contains an additional PRY-SPRY domain at its COOH-terminus. The RING domain of TRIM72 was indicated in E3 ubiquitination of insulin receptor substrate 1 (151, 233, 285) and focal adhesion kinase (194) and was therefore involved in the development of insulin resistance and myogenesis. Recent studies reported that TRIM72 was expressed in lung tissue, and its presence is broad in lung resident cells, including ATI, ATII cells, endothelial cells, and alveolar macrophages (125, 190). Kim et al. showed that absence of TRIM72 increases susceptibility to ventilation-induced lung injury, whereas overexpression via an inducible transgenic model is protective for ventilation-induced lung cell death (125). Specifically, results using fluorescent dyes that separately label nonwounded, wounded but repaired, and necrotic cells showed that ablation of endogenous TRIM72 specifically affected the repair process without rendering the primary alveolar epithelial cells to be more resistant to mechanical injury (125). During repair of lung cells, a facultative interaction between TRIM72 and Cav1 was shown to be a critical step (125, 190), which will be discussed in Caveolin1.

Previous studies showed that TRIM72 preferably localizes at plasma membrane and vesicle complexes close to the membrane, possibly due to its higher affinity to the biological membrane component phosphatidylserine (31). Disruption of the plasma membrane induces rapid translocation of intracellular TRIM72 to the wounding site and repair of myogenic C₂C₁₂ cells. Interestingly, this was not abolished by removal of extracellular Ca²⁺ but by scavenging of ROS with dithiothreitol (31). Controversially, Howard et al. showed that dithiothreitol did not alter dye entry via laser-created wounds in either HeLa or kidney epithelial BSC1 cells, whereas vitamin E improved repair of muscle cells in an anti-redox-dependent manner (94), suggesting that, although the role of TRIM72 in plasma membrane repair is established, the specific trigger(s) to mobilize TRIM72 for repair are under debate. It was observed that recombinant TRIM72 tends to polymerize through its cysteine residues under native condition (102), and S-nitrosylation of TRIM72 at cysteine-144 was shown to be important for protection of cardiomyocytes from oxidative insults (129, 279). Other studies also reported that nonmuscle myosin type IIA served as a motor to translocate TRIM72 during repair (149), and Zn²⁺ binding to TRIM72 was important for its repair function (30). Furthermore, cholesterol content in the plasma membrane was shown to greatly affect injury-induced membrane translocation of TRIM72 and cell repair in ischemia-reperfusion injury (271). Nevertheless, these repair modifiers of TRIM72 have yet to be verified by other independent studies.

Developing a TRIM72 augmentation therapy for the treatment of tissue injurious diseases had been a continuing research area of interest. Recombinant TRIM72 protein was shown to have notably therapeutic effects in mouse models of muscular dystrophy (90, 273) and for ischemic reperfusion injury in muscle and heart (49, 152, 291). Because striated muscles have the highest levels of endogenous TRIM72 expression in mice, it is not clear why the abundant repair protein is not sufficient to rescue tissue injuries, since there were no mutations or polymorphisms of TRIM72 identified so far in tissue injurious diseases. It was reported that human heart tissue had no TRIM72 expression unlike the mouse heart (146), suggesting species disparities in tissue repair. Curiously, in a recent study by Jia et al. (113), the authors observed an increase in mean linear intercept measurements (L_m) of alveolar spaces in the trim72^−/− lung compared with the wild-type lung, and recombinant TRIM72 was able to reduce the increase in end-expiratory lung volume and L_m in mouse models of emphysema. However, Kim et al. reported repair defects in primary epithelial cells isolated from trim72^−/− lungs but normal histology under resting conditions (125). It is not clear how defective membrane repair affects lung structure and remodeling, so further verification of the observation is needed.

Caveolin1.

Caveolins are protein components of the plasma membrane invaginations, caveolae (217), and play major roles in endocytosis and cell signal transduction (201). There are three subtypes of caveolins, i.e., Cav1, Cav2, and Cav3, while Cav1 is the main type of caveolar protein in ATI cells, fibroblasts, and endothelial cells in the lungs. Mutations in Cav3 cause several forms of dystrophic myopathies (80, 181), and Cav1 was long thought to be a critical regulator of acute lung injury (115). Major histological abnormalities in lung tissue were observed in Cav1^−/− knockout mice (60, 210). Specifically, the alveolar walls of the Cav^−/− lung were thicker and filled with extracellular fibrillary deposits, whereas irregular alveolar space and hypercellularity were observed throughout the lung (60). Studies by Corrotte et al. (50) showed that caveolar endocytosis is an indispensable step for the success of plasma membrane repair by removing pores formed by PFTs on the plasma membrane. Previous studies also found that caveolae structures were disturbed during acute kidney injury by oxidative stress, and upregulation of Cav1 was linked to survival of renal cells (167, 287). Using primary ATI cells, Wang et al. (268) also showed that caveolar endocytosis, but not the clathrin- and fluid-phase endocytic pathways, was important for repair of lung cells. Their results indicated that hypotonic exposure increased the probability of plasma membrane wound repair through upregulation of caveolar endocytosis in a Src-dependent fashion (269).

Interestingly, previous studies identified a direct physical interaction between TRIM72 and Cav1 (125, 190). The interaction between TRIM72 and Cav1 led to enhanced membrane distribution of Cav1 in cotransfected HEK cells, and primary ATI cells from trim72^−/− lung had reduced membrane localization of Cav1 (125, 190). Through pulse-chase experiments, a significant reduction in caveolar endocytosis was also detected in primary trim72^−/− lung cells compared with the wild-type cells, whereas cells overexpressing TRIM72 had increased cargo uptake through caveolar but not the clathrin- and fluid-phase pathways (190). In addition, TRIM72-Cav1 double-knockout mice showed increased apoptotic cell death in the lung following overventilation compared with either single knockout alone or the wild-type lung (190). This finding suggests that a direct link between TRIM72 and caveolar endocytosis collaboratively repairs wounding of lung epithelial cells. Echoing this observation is the lower repair capacity of ATII cells compared with ATI cells (82) where caveolae was lacking but TRIM72 is present (125), suggesting that a dual repair protein team is more effective than either alone in lung cell repair. Interestingly, mutagenesis studies found that all four domains of TRIM72, including RING, B-box, coiled-coil, and PRY-SPRY, were required for TRIM72 and Cav1 interaction, indicating the conformational nature of this interaction (190).

State of the art and perspective.

In this section, we described membrane repair mechanisms that were studied in lung epithelial cells. Because this research area is seriously understudied, more detailed investigation is warranted. Generally, the internal and external triggers for a cell to initiate membrane repair, the interaction between repair molecules with persistent presence of injury stimuli, long-term cell fate following successful membrane resealing, and behavior of neighboring cells of an injured or repaired cell are interesting areas to study. For lung cell repair, discovering members of the entire repair team and their spatial and temporal interaction during the repair process are of great interest for latching onto tangible therapeutic targets for the treatment of ARDS and IPF. In addition, preliminary investigation by Godin showed that ATIs are more likely to repair PM defects than ATII cells (82), whereas coincidentally ATII cells are historically known to lack Cav1 (34). Thus, the repair phenotypes of different types of lung cells that harbor variable repair molecules need to be carefully characterized. Furthermore, determining whether targeting membrane repair effectively reduces the pathology of ARDS and IPF is an integral part of our long-term research interests.

Activation of Aberrant Tissue Repair Pathways Following Cell Injury

Repair of individual cells cannot be taken out of the tissue context. We learned a great deal about tissue healing from repair of wounded skin, which is largely comprised of three stages, the coagulation stage, the granulation stage, and the maturation stage (133). The coagulation stage serves to stop bleeding, form scaffolding for migrating epithelial cells, and nucleate activated growth factor-secreting platelets/macrophages; the granulation stage is dominated by fibroblast proliferation, angiogenesis, and collagen deposition to form scars; the maturation stage is characterized by debris removal and scar remodeling by macrophages and proteases and retraction of fibroblasts. During normal healing of small wounds, minimal scar tissues are formed due to limited activity and early withdrawal of fibroblasts while large-scale fibroblast activation is usually necessary for repair of large tissue wounds with extensive epithelial destruction. The process of skin wound healing inspired hypotheses to explain the pathogenesis of injurious lung diseases such as ARDS and IPF. Because ARDS is often characterized by extensive lung injury, it is conceivable that a percentage of ARDS patients develop fibrosis long after the respiratory distress is relieved (29). However, severe lung injury was not evidenced in most IPF cases, but large-scale fibroblast activation was often seen (56). A prevailing theory for IPF etiology is that repeated or persistent microinjury of the lung epithelium can also cause overactivation of the profibrotic signaling pathways (225), with the aid of reactivation of developmental signaling pathways. Because in-depth knowledge on epithelial-to-mesenchymal cross talk in IPF has been reviewed previously (220, 226), we will briefly describe a few key signaling pathways that could be induced by epithelial injury as an extension of our cell injury and repair focus (Fig. 3).

Fig. 3.

Activation of aberrant tissue repair pathways following injury. On injured epithelial cell membrane, latency-associated peptide (LAP) binds to integrins αvβ₆ to expose and activate the transforming growth factor (TGF)-β. Increased activated TGF-β binds to TGF-β receptor (TGFβRI and TGFβRII). Smad was phosphorylated and moved to nucleus to promote fibrotic gene expression. For Wnt signaling, injured alveolar epithelial cells secrete Wnt ligands. Wnt binds to cell surface receptor Frizzled to expose β-catenin. β-Catenin moves to nucleus and promotes the paracrine of Wnt and WISP1-mediated epithelial-mesenchymal transition (EMT) and collagen production. Epithelial injury promotes the secretion of sonic hedgehog (SHH) from injured epithelial cells. SHH releases glioma-associated oncogene homolog (GLI) by inhibiting transmembrane protein smoothened (SMO). GLI moves to nucleus to promote fibroblast proliferation, migration, and extracellular matrix (ECM) deposition. Notch signaling promotes EMT in epithelial cells and activation of fibroblasts via the intracellular domain of the notch protein (NICD). ER stress induces the expression of binding immunoglobulin protein (BiP). BiP binds and activates protein kinase R-like endoplasmic reticulum kinase (PERK), iron-responsive element 1 (IRE1), and activating transcription factor-6 (AFT6), which eventually lead to cell apoptosis.

TGF-β signaling pathway.

Among many profibrotic mediators, TGF-β is one of the initiating, and most essential, fibrogenetic cytokines for the pathogenesis of IPF. TGF-β was secreted as an inactive complex with latency-associated peptide (LAP) and tethered to the extracellular matrix (ECM) by latent TGF-β-binding protein (278). Binding of LAP to integrin exposes and activates TGF-β, which binds to TGF-β receptor (TGFβR) I and II to turn on transcription of profibrotic genes via phosphorylation of SMAD. Previous studies showed that alveolar epithelial cells serve as a primary source of TGF-β (283). In particular, following injury, epithelial cells are induced to contract their subcortical actin-myosin cytoskeletons, applying physical force on latent TGF-β complex, and activate TGF-β (227). Meanwhile, injurious stimuli also promote the expression of epithelial integrins αvβ₆ (227). Neverthless, blocking TGF-β signaling in either epithelial cells or mesenchymal cells attenuated bleomycin-induced lung fibrosis. Li et al. (148) reported that epithelial-specific TGFβRII deficiency protects the lung from bleomycin-induced fibrosis, which is associated with improved epithelial survival and increased fibroblast apoptosis. Recently, Luo et al. (164) reported that blocking TGF-β signaling in lung mesenchymal cells by targeting TGFβRII with Tbx4 lung enhancer-driven Tet-On system also protected the lung from bleomycin-induced fibrosis, mainly through inhibiting TGF-β-induced collagen production. Therefore, both epithelial and mesenchymal TGF-β signaling play important but distinct roles in mediating lung fibrosis. It is worth mentioning that TGF-β alone is not sufficient to promote permanent fibrosis, and TGF-β-induced fibrotic responses are mostly transient and reversible, which was completely recovered within a month once induction of active TGF-β was stopped (144).

Wnt signaling pathway.

Wnt is a family of cysteine-rich lipid-modified signaling glycoproteins that regulate embryo architecture through coordinating proliferation and differentiation of adjacent cells (147, 262). The Wnt signaling pathway starts with secretion and binding of Wnt ligands to its surface receptor Frizzled on target cells. The canonical Wnt pathway includes binding of coreceptor Lrp5/6, cytosolic stabilization, and nuclear translocation of β-catenin, whereas the noncanonical pathway induces β-catenin-independent signals to trigger expression of target genes (166). Under normal development and homeostasis, canonical Wnt signaling is essential for specification, expansion, and differentiation of multiple lung epithelial and mesenchymal lineages (262, 266), whereas noncanonical Wnt signaling controls vascular formation, ECM production, smooth muscle cell proliferation and migration, and epithelial cell maintenance and differentiation (147). Given the critical role of the Wnt pathway in directing alveolar cell fate (43, 214, 270), the observed overactivation of both canonical and noncanonical Wnt/β-catenin pathways in IPF (1, 43, 130, 193, 263) may reflect an attempt of compensatory response following tissue injury (59). However, aberrant Wnt upregulation turns out to be more deleterious than beneficial. First, constitutively Wnt upregulation may inhibit ATII to ATI differentiation and thus prevent proper reepithelization of the lung. Second, Wnt/β-catenin activation promotes fibroblast expansion via the paracrine mode and thus increases collagen accumulation in the ECM (263). Third, Wnt/β-catenin signaling is also a central trigger of EMT, a mechanism providing an additional source of mesenchymal cells in IPF (44, 205).

Sonic Hedgehog signaling pathway.

The Hedgehog (HH) signaling pathway is a developmental pathway that finely controls tissue polarity and architecture by distance-dependent ligand dispensing on target cells (257). Of the three HH family members (sonic, Indian, and desert HH), sonic HH (SHH) is expressed in the lung. Secreted SHH binds with its receptor patched, which releases the transmembrane protein smoothened (SMO) from inhibition by patched. SMO activates glioma-associated oncogene homolog (GLI), allowing it to translocate into the nucleus and turn on genes involved in cell proliferation, migration, angiogenesis, and stem cell regeneration (292). SHH may be secreted from lung epithelial cells and acts on a variety of target cells, and thus it was considered to be critically important for epithelium-to-mesenchyme cross talk in IPF. For example, a few studies found upregulated SHH in epithelial cells of the fibrotic human lung, whereas its downstream signaling partners patched-1, smoothened, and GLI1 were mainly present in fibroblasts and inflammatory cells (25, 236). SHH overexpression was shown to increase fibroblast proliferation, migration, and deposition of ECM (25, 153), whereas inhibition of GLI decreases lung fibrosis and collagen accumulation (188). In mouse pulmonary fibrosis models, an increased number of GLI1-positive cells was observed in the fibrotic tissue of bleomycin-injected lungs 4 wk after injury (153). Interestingly, a recent study by Peng et al. (200) showed that GLI1-positive cells were decreased in the parabronchial mesenchymal cells following chronic bleomycin injury to the lung and proposed a role of epithelial SHH expression in maintaining mesenchymal quiescence. These studies revealed a complex functional role of SHH signaling in pulmonary fibrosis, which may be spatial, temporal, and cell-type dependent.

Notch signaling pathway.

The Notch signaling pathway plays a major role in embryonic development through direct cell-cell communication to influence fine binary cell fate decision and tissue polarity (71, 280). The inactive transmembrane Notch is activated by the binding of ligands to its extracellular portion, followed by enzymatic cleavage and nuclear translocation of its intracellular domain to turn on transcription of genes in apoptosis, cell cycle, proliferation, differentiation, transcription regulation, and neurogenesis. Recent studies highlighted the postinjury activation of Notch signaling in lung injurious models (35, 251), which was shown to be critical to awake progenitor cells (251). Nevertheless, persistent Notch activity at mesenchymal cells seems to be a culprit that drives normal proliferative repair to a fibrotic process (35, 251). Indeed, inhibition of Notch signaling promoted alveolar repair and reduced fibrosis in bleomycin-induced lung injury (35). Another noticeable role of Notch in pulmonary fibrosis is to promote EMT. For example, ectopic expression of the Notch intracellular domain, or treatment of alveolar epithelial cells with Notch ligand, induced expression of smooth muscle α-actin (SMA), collagen I, and vimentin and reduced the expression of E-cadherin, occludin, and zonula occludens–1, indicating EMT (9).

ER stress.

ER stress is technically not a signaling pathway, but it plays a pivotal role in multiple pathogenic aspects of ARDS and IPF such as epithelial cell apoptosis, EMT, and inflammation (118). During ER stress, ER capacity to properly fold nascent proteins is reduced, and UPR occurs to regulate the expression of ER chaperons, additional protein quality control genes, genes of the ER-associated protein degradation, or proapoptotic genes (131). As described above, ER stress is a direct consequence of various injuries to lung cells, including environmental insults, viral infection, and cigarette smoking (196, 243), presumably even in cells that successfully resealed their membrane wounds. Although all injured lung cells undergo some degree of ER stress (13, 141), the ability to restore homeostasis in certain cells, such as aging ATII cells harboring misfolded surfactant proteins, may be severely compromised. Therefore, it is not surprising to find that ER stress was increased in ARDS models induced either by oleic acid (OA) or LPS (154, 155, 192). Furthermore, OA-induced lung injury was exacerbated by ER stress inducer, but ameliorated by ER stress inhibitor (154), suggesting that ER stress is critically involved in the pathogenesis of ARDS. ER stress may be aggravated by several disease modifiers of ARDS, such as alcohol abuse and hyperoxia (120, 185), as well as by many disease modifiers of IPF, such as aging, oxidative stress, mutations in surfactant protein, telomerase and mucin, or simply energy deficiency (243). Although ER stress is not specific to ARDS or IPF, severe ER stress is associated with lung epithelial cell death and fibrogenesis (131). For example, Flodby et al. (72) created conditional binding immunoglobulin protein knockout specifically at ATII cells and found increased ATII cell apoptosis, oxidative stress, Smad3 phosphorylation, and expression of Smad3-target genes in the knockout lung. Bueno et al. showed that ER stress triggers profibrotic factor expression in epithelial cells (28) and promotes EMT transition, providing a direct source of myofibroblasts in IPF (118). In addition, EMT-derived cells further promote fibrogenesis through aberrant epithelial-mesenchymal cross talk (118), forming a positive feedback in IPF.

Signaling pathway network.

It should be emphasized that the above signaling pathways evidently cross talk with one another during fibrogenesis. For example, study by Akhmetshina et al. showed that TGF-β stimulates canonical Wnt signaling in a p38-dependent manner by decreasing the expression of the Wnt antagonist Dickkopf-1 (1), and the study by Lam et al. highlighted the role of TGF-β as a downstream factor of Wnt and coreceptor Lrp5 activation in bleomycin-induced pulmonary fibrosis (140). In addition, SHH was shown to play a mutually simulative role with TGF-β and create a positive feedback loop in the fibrotic process (45, 96). Furthermore, Notch-induced myofibroblast differentiation is dependent on the TGF-β-Smad3 pathway, and inhibition of endogenous Notch with small molecule inhibitor attenuated TGF-β-induced SMA expression (9). Interestingly, Cao et al. (35) proposed a complex signal relay network among pulmonary endothelial cells, epithelial cells, macrophages, and perivascular fibroblasts for the pathogenesis of IPF in which Wnt ligand secreted from activated macrophages induces Notch ligand Jagged1 expression in endothelial cells that in turn stimulates the Notch signaling pathway in the neighboring perivascular fibroblast.

State of the art and perspective.

In this section, we described prevailing signaling pathways known to be activated following lung epithelial injury to facilitate a big-picture view on the injury-repair-fibrosis axis. What is shared in common is that these signaling pathways largely cause progenitor cell activation, mesenchymal cell expansion, and fibrogenesis, in favor of IPF pathogenesis. We attempt to review the mechanisms through which injurious stimuli may activate these pathways, but this is not clear with the exception of the TGF-β signaling pathway, and partly the ER stress pathway. This would be of great interest to the field. In addition, information on how cellular wounding (loss of cell function, leakage of cellular content, or activation of stress genes) and repair moieties interact with components of these signaling pathways, how these interactions direct the entirety of these pathways, and how these pathways affect the membrane repair process would bridge gaps in current knowledge. Furthermore, because activation of proliferative and fibrotic signals is part of the normal tissue healing process, whereas failure of resolution leads to disease state, the impact of cell wounding and repair on the “off switch” of these above signaling pathways may shed light on future research.

Perspectives and Significance

Plasma membrane repair is a fundamental biological process, and key membrane repair steps are largely conserved, although tissue and cell type differences may exist due to variations in organ function, cell turnover rate, and existence of different repair machineries. Complex physical and biochemical tissue injury factors injure lung cells under pathological conditions, and lung cells engage lipid trafficking, subcortical cytoskeleton rearrangement, and repair machineries to mend plasma membrane wounding. Lung cell wounding and repair occurs throughout injurious lung diseases and is an integrated part of ARDS and IPF etiology. Defective lung cell repair, or chronic lung cell wounding, also leads to activation of aberrant tissue repair pathways following injury, contributing to the pathogenesis of fibrotic lung diseases. Deep understanding of the mechanisms of membrane repair provides additional insights for lung biology and new therapeutic target for ARDS and IPF. The early presence of cell injury, persistence of injurious factors, and broad indications of membrane repair therapies make targeting membrane repair a particularly attractive direction for the treatment of lung injurious diseases, and thus further research efforts are necessary for propelling this promising field.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grant R01-HL-116826 and Eastern Virginia Medical School start-up.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

X.C. and X.Z. prepared figures; X.C., R.D.H., and X.Z. drafted manuscript; X.C., R.D.H., C.L., and X.Z. edited and revised manuscript; X.C., R.D.H., C.L., and X.Z. approved final version of manuscript.