Caveolin-1: a critical regulator of lung injury

Abstract

Caveolin-1 (cav-1), a 22-kDa transmembrane scaffolding protein, is the principal structural component of caveolae. Cav-1 regulates critical cell functions including proliferation, apoptosis, cell differentiation, and transcytosis via diverse signaling pathways. Abundant in almost every cell type in the lung, including type I epithelial cells, endothelial cells, smooth muscle cells, fibroblasts, macrophages, and neutrophils, cav-1 plays a crucial role in the pathogenesis of acute lung injury (ALI). ALI and its severe form, acute respiratory distress syndrome (ARDS), are responsible for significant morbidity and mortality in intensive care units, despite improvement in ventilation strategies. The pathogenesis of ARDS is still poorly understood, and therapeutic options remain limited. In this article, we summarize recent data regarding the regulation and function of cav-1 in lung biology and pathology, in particular as it relates to ALI. We further discuss the potential molecular and cellular mechanisms by which cav-1 expression contributes to ALI. Investigating the cellular functions of cav-1 may provide new insights for understanding the pathogenesis of ALI and provide novel targets for therapeutic interventions in the future.

the purpose of this review is to explore the scientific literature supporting a role for caveolin-1 (cav-1) in the pathogenesis of acute lung injury (ALI).

Caveolae are omega-shaped invaginations of plasma membrane that were first discovered by Palade (80) and Yamada (115) in the 1950s. Caveolae reside most prominently in epithelial cells, endothelial cells, adipocytes, and fibroblasts in the lung (2, 31, 86). In the past decade, the concept has emerged that caveolae are a specialized form of lipid rafts (2, 31, 80, 86, 115). Both planar lipid rafts and caveolae provide platforms that anchor membrane proteins in the “floating sea” of lipid bilayers. The structural proteins required for caveolae formation are the caveolins. These coat proteins, i.e., caveolins, include cav-1, cav-2, and cav-3 (24, 94, 96), which facilitate membrane invagination and response to external stimuli by transducing signals that modulate cellular activity (2, 13, 31, 80, 86, 115). Cav-1 and cav-2 have been identified in the majority of tissues and organs (24, 96), whereas cav-3 is mainly are expressed in cardiac, smooth, and skeletal muscle cells (24, 97). Currently, the functional role of cav-2 remains unclear, although it is expressed and heteroligomerizes with cav-1 (24, 96, 97).

Caveolin-1

Isoforms and structure.

The cav-1 gene encodes a 178-amino acid protein in a predicted β-sheet conformation (29). Two isoforms of cav-1 (cav-1α and cav-1β) have been identified (24, 29, 88, 96, 97). The 24-kDa cav-1α (residues 1–178) and 21-kDa cav-1β isoforms (residues 32–178) are generated by alternative splicing of mRNA (47). These two isoforms differ by 31 additional amino acids present in the α-isoform at its NH₂ terminus (residues 1–31). The distinct function of these two isoforms remains unclear. However, recent studies in zebrafish, using antisense morpholino oligomers, reveal nonredundant and evolutionarily conserved functions for the individual cav-1 isoforms (17, 82).

As shown in Fig. 1, the sequence of cav-1 contains a hydrophobic central domain with a hairpin-like conformation inserted in the inner leaflet of the plasma membrane (101). 14–16 Cav-1 monomers form a single cav-1 homooligomer. Both the COOH and NH₂ terminus of the cav-1 monomer face the cytoplasm (101). Among the modular sequences, an important domain called the caveolin scaffolding domain (CSD), located between amino acid residues 82–101 in the NH₂-terminal region adjacent to the hydrophobic membrane-insertion domain, is required for caveolin dimerization and critical in controlling the interactions between cav-1 and numerous signaling proteins. The COOH-terminal region of cav-1 contains three palmitoylated cysteine residues, which are also relevant to oligomerization (101) (Fig. 1). Studies have revealed that COOH-terminal palmitoylation is crucial for cav-1 to attach the plasma membrane (32). Another key region of the cav-1 structure is tyrosine-14 (Y14). Cav-1 is phosphorylated on tyrosine-14 in response to cellular stimuli. Recent evidence indicates that phosphorylated cav-1 (pY14) is responsible for multiple biological processes including caveolae internalization and regulation of cell signaling (30).

Fig. 1.

Structure of caveolin-1. Cav-1 contains a hydrophobic central domain with a hairpin-like conformation inserted in the inner leaflet of the plasma membrane. 14–16 Cav-1 monomers form a single cav-1 homooligomer. A simplified dimmer is illustrated here. Both the COOH and NH₂ terminus of the cav-1 monomer face the cytoplasm. The caveolin scaffolding domain is located between amino acid residues 82–101 in the NH₂-terminal region adjacent to the hydrophobic membrane-insertion domain. The COOH-terminal region of cav-1 contains 3 palmitoylated cysteine residues. Inset: caveolae viewed by electron microscopy (×50,000).

Cav-1 is an integral membrane protein associated with various membranous structures, including endoplasmic reticulum (ER), Golgi, and plasma membranes (Fig. 1). After synthesis, cav-1 is inserted into the ER membrane with both its NH₂- and COOH-terminal sequences facing the cytoplasm. In the ER, cav-1 forms homogeneous SDS-resistant homooligomers of 7–14 cav-1 molecules (81). Cav-1 is then exported to the Golgi and transported to the cell surface to form caveolae. Remarkably, cav-1 can be recycled in reverse order from the cell plasma membrane (4, 81). Interestingly, cav-1 is also found in the cytoplasm as a soluble protein as well as in the mitochondria, secretary vesicles, cell-cell contact sites, and the nucleus (1). Four regions of cav-1 have been identified to influence its intracellular trafficking, including: 1) a conserved region between amino acids 66 and 70, which is necessary for exit from the ER, 2) a region between amino acids 71–80, which regulates incorporation of cav-1 oligomers into the Golgi apparatus, 3) a region between amino acids 91–100, and 4) COOH-terminal amino acids 134–154, which are thought to control cav-1 oligomerization and exit from the Golgi apparatus (96).

Regulation of cav-1 expression.

Cav-1 expression is regulated transcriptionally or posttranslationally. The cav-1 5′-flanking region includes three G+C-rich sites which are potential sterol regulatory elements (SREs). Additionally, the 5′-flanking region contains a CAAT sequence and a Sp1 consensus sequence (4). SRE-like elements are involved in the transcriptional response to stimuli, such as low-density lipoprotein-free cholesterol (LDL-FC) (89, 114). Other transcription factors reported to regulate cav-1 expression include the forkhead (FKHR) family of transcription factors, FOXO3a (20, 105), C-myc (89, 114), and NF-κB (104).

Downregulation of cav-1 can be achieved by promoting cav-1 degradation via lysosomal degradation pathways (9). For example, in intestinal epithelial cells, cav-1 protein level is regulated by another lipid raft protein, flotillin-1 (flot1), by preventing its lysosomal degradation (106). Cav-1 is also regulated by other lipid raft component proteins, for example, the cavins. Recent studies indicate that deletion of cavin-1 diminishes cav-1 protein expression without affecting cav-1 mRNA (36), and vice versa, deletion of cav-1 abolishes cavin-1 expression (34). These results suggest that cav-1 and cavin-1 are posttranslationally regulated by degradation and also by transcriptional regulation of mRNA levels.

Cav-1 functions.

Initially described decades ago, cav-1 is the primary protein of caveolae. Although the role of cav-2 remains unclear, the function of cav-1 has been studied extensively. Many of these functions, mentioned below, are regulated by cav-1 posttranslational modifications such as palmitoylation at the three cysteine sites in the COOH terminus and phosphorylation of NH₂-terminal tyrosine Y14 and serine-80 near the CSD (48).

FORMATION OF CAVEOLAE.

Cav-1 is essential for caveolae formation. Deletion of cav-1 results in absence of caveolae (84). As expected, overexpression of cav-1 leads to an increase in the number of caveolae (56). Cav-1 is thought to be the most important structural protein required for caveolae formation, although recent data suggest that the cavins also play important roles in regulating the architecture of caveolae (33).

CONTROL OF CHOLESTEROL HOMEOSTASIS.

Cav-1 directly binds cholesterol and long-chain unsaturated fatty acids and forms membrane-associated oligomers (21). Growing data suggest that cav-1 controls the import and export of cellular cholesterol by caveolae (18). Furthermore, cav-1 coordinates lipid metabolism (19). However, cav-1 has been shown to play both proatherogenic and antiatherogenic roles, depending on the cell type studied (23). In smooth muscle cells, cav-1 suppresses cell proliferation and may have antiatherogenic effects, whereas in endothelial cells, cav-1 promotes transcytosis of LDL-cholesterol particles (42).

REGULATION OF MEMBRANE TRAFFICKING, ENDOCYTOSIS, EXOCYTOSIS, AND TRANSCYTOSIS.

Cav-1 interacts with many receptor tyrosine kinases, such as EGF receptor (EGFR) as well as nonreceptor tyrosine kinases such as Src as well as serine/threonine kinases such as PKC family members (110), which play important roles in membrane trafficking. Endocytosis, exocytosis, and transcytosis of many macromolecules via caveolae require the presence of cav-1 (3). Examples of macromolecule transport include albumin, cholera toxin, and tetanus toxin (52, 71). Among these, albumin uptake by lung endothelial cells appears to be directly involved in the pathophysiology of ALI. Endothelial cell cav-1 is required for the efficient uptake and transport of albumin from the blood to the interstitium (92).

REGULATION OF CELL SIGNALING.

Cav-1 interacts with a variety of downstream signaling molecules, including endothelial nitric oxide synthase (eNOS), heterotrimeric G proteins, nonreceptor tyrosine kinases, Src-family tyrosine kinases, and p42/44 mitogen-activated protein (MAP) kinase (10, 15, 16, 20, 26, 45, 59, 98, 121). Cav-1 anchors these signal transducers in their inactive conformation until activation by appropriate stimulation (10, 15, 16, 20, 26, 45, 59, 98, 121). Many of these signaling molecules interact with the CSD directly via the hydrophobic cav-1 binding motif (ΦxxΦxxxxΦ or ΦxΦxxxxΦ, where Φ stands for aromatic amino acids). Emerging evidence demonstrates that cav-1 functions as a negative or positive regulator of cell signaling, depending on the cell type and specific cell signaling pathway investigated. For instance, as a negative regulator, cav-1 inhibits Wnt signaling by blocking β-catenin-mediated transcription (26). Cav-1 inhibits eNOS (20, 45, 98, 121), and recombinant cav-1 blocks Neu (c-erbB2)-mediated signal transduction (16). Additionally, cav-1 inhibits signaling from EGFR, Raf-1, MEK-1, and Erk2 to the nucleus. Furthermore, cav-1 peptides derived from residues 32–95 inhibit the kinase activity of purified MEK-1 and Erk2 (15, 16, 20, 26, 27, 45, 59, 98, 121). In contrast, cav-1 positively regulates integrin-dependent signaling, Shc-mediated signaling (58, 67, 112), and the phosphoinositide 3-kinase (PI3K)/Akt pathway (55, 95, 123). Cav-1 overexpression activates phospho-Akt signaling pathways in Hela cells, in prostate cancer cells, and in MCF-7 breast cancer cells (55, 95, 85).

Lung phenotype in cav-1-deficient or cav-1-transgenic mice.

Cav-1 is abundantly expressed in lung epithelia, endothelia, and fibroblasts (113). Cav-1 knockout mice (cav-1^−/− mice) exhibit significant abnormalities within the lungs (14, 83, 113). In 2001, two groups, Drab et al. (14) and Razani et al. (83), independently generated cav-1-deficient mice. These initial studies and the invaluable tool of cav-1^−/− mice have enhanced our understanding of the function of cav-1 both in vitro and in vivo. Cav-1^−/− mice, although lacking caveolae, are viable and fertile. Cav-1^−/− mice exhibit significantly lower levels of cav-2 due to posttranscriptional regulation and increased degradation of this protein in the absence of cav-1. Cav-1^−/− possess multiple defects, including impaired homeostasis of lipid, which lead to dysfunctional muscle, cardiovascular, and lipidogenic tissues (14, 83, 113). Cav-1^−/− mice also have a basal phenotype characterized by hyperlipidemia.

For the remainder of this review, we will focus on the pulmonary phenotype of cav-1-deficient mice as they have served as the primary tool in cav-1-related research. The high density of caveolae in the septa of lung tissue indicated that these structures play an important role in the respiratory system. Cav-1^−/− mice have a thickened alveolar wall leading to constricted and irregular alveolar spaces. Although these abnormal septa contain increased amounts of endothelial and hematopoietic cells, two characteristics of this “hypercellularity” need to be recognized. 1) While hypercellularity is a notable feature of cav-1^−/− mice, these cells appear to be non-differentiated cells, given that von Willebrand factor (vWF), which marks differentiated endothelial cells, is not expressed (8, 14, 83, 113). 2) The phenotype of the cav-1^−/− mice, which lacks both the cav-1α and -β isoforms, shows that only the pulmonary endothelium exhibits hyperproliferation, whereas the alveolar epithelium does not (8, 14, 83, 113). In addition to hypercellularity, recently, Le Saux et al. (53) reported a progressive increase in deposition of collagen fibrils in airways and parenchyma in cav1^−/− mice (53, 54). The increased deposition of collagen fibrils may lead to significantly reduced lung compliance and increased elasticity and airway resistance. This process is associated with increased TGF-β/Smad signaling pathways involved in tropoelastin, col1 α2, and col3 α1 gene expression in lung tissues (53, 116). As a consequence of the pulmonary abnormalities, cav-1^−/− mice are exercise intolerant and show early onset exhaustion during swimming tests (53, 116). Based on the above-described features, presumably, cav-1^−/− mice would be more susceptible to ALI. Surprisingly, numerous recent studies in cav-1^−/− mice do not fully support this hypothesis. Therefore, we will review the scientific literature exploring the function of cav-1 in the pathogenesis of ALI.

Using a different approach, Yang et al. (116) generated epithelial cell-specific cav-1 overexpressing mice via a mouse mammary tumor virus (MMTV) long terminal-repeat promoter, which is predominantly expressed in specific epithelial cells. They reported that the MMTV-cav-1(+) transgenic mice demonstrate bronchiolar epithelial hyperplasia and atypia. Additionally, the MMTV-cav-1(+) transgenic mice tend to have a greater incidence of malignancies, including lung cancer, when compared with the MMTV-cav-1(−) littermates (116).

Caveolin-1 and Lung Injury

In 1994, Lisanti et al. (58) first described the expression of caveolin proteins in lung. Since then, emerging studies have investigated the functional role of caveolins in lung diseases. ALI is a diffuse heterogeneous process characterized by diffuse alveolar cell death (DAD), acute inflammatory responses including non-cardiogenic pulmonary edema, widespread capillary leakage, cytokine release, and ROS generation leading to low lung compliance, severe hypoxia, and abnormal gas exchange (5, 41, 49, 62, 64, 69, 119). Pulmonary fibrosis develops in the later stages of ALI (62–64, 69, 119). Recent studies have explored the potential roles of cav-1 in lung injury from all the above-mentioned aspects (Fig. 2). After stimuli, various cell types participate in the pathogenesis of ALI, including polymorphonuclear neutrophils (PMNs), vascular endothelial cells, lung alveolar epithelial cells, and lung fibroblasts. In the rest of this article, we will discuss current literature on the functional roles of cav-1 in each of these components involved in ALI.

Fig. 2.

Scheme of proposed roles of caveolin-1 in acute lung injury.

Cav-1 regulates acute inflammation and capillary leakage during lung injury.

Primary ALI is caused by a direct injury to the lung, such as pneumonia, ventilation-associated injury, hyperoxic injury, trauma, and contusion (5, 41, 49, 63, 64, 69, 119). Secondary ALI is caused by an indirect insult to the lung from conditions such as pancreatitis, severe sepsis, or transfusion-related ALI (5, 41, 49, 63, 64, 69, 119).

Acute inflammation has been well described in the pathological stages of ALI/ARDS, along with increased vascular permeability, fibroproliferation with hyaline membranes, epithelial apoptosis, and varying degrees of interstitial fibrosis (5, 41, 49, 62–64, 69, 1119).

ENDOTHELIAL CELL CAV-1 PLAYS A CRUCIAL ROLE DURING ACUTE INFLAMMATION.

The transport of macromolecules from the blood-space to the tissue-space is regulated by endothelial cell caveolae (39, 40, 93, 99). For example, albumin is not endocytosed in cav-1^−/− mouse lung endothelial cells and remains in the blood vessel lumen (99). Abundant caveolae is a distinctive feature of endothelial cells. Transmembrane water channel protein aquaporin-1 is expressed in caveolae on the lung endothelial cell surface (90). This finding supports the role of caveolae in endothelia-mediated transcellular transport of water (90). The mechanical rearrangement of caveolae/lipid rafts contribute to non-cardiogenic pulmonary edema during ALI (39, 40). Unlike the above studies which focus mainly on caveolae instead of cav-1, Sundivakkam et al. (100) studied the role of CSD in Ca²⁺ entry in endothelial cells. Ca²⁺ influx has been reported to regulate pulmonary vasoconstriction (100). CSD regulates Ca²⁺ store release-induced Ca²⁺ entry in endothelial cells, suggesting a potential role in endothelial permeability (100).

Current evidence supports a dual role of cav-1 in regulating microvascular permeability: 1) as a caveolae-associated structural protein controlling caveolar transcytosis, and 2) as a tonic inhibitor of eNOS activity, which negatively regulates paracellular permeability (99). As a result of this dual function, although the cav-1^−/− mouse has lung vascular abnormalities and lung fluid balance abnormalities at baseline, upon insult, cav-1^−/− mice resist ALI compared with wild-type mice. Cav-1^−/− mice have improved survival following LPS challenge (60). Similar results were reported in several previous studies using LPS (28) and hyperoxia exposure (28, 44, 70). Since cav-1 is a negative regulator of eNOS, cav-1^−/− mice have increased plasma NO, and, paradoxically, increased pulmonary vascular resistance (PVR) (62). Elevated PVR is attributed to pulmonary precapillary vessel remodeling, and, therefore, the elevated basal plasma NO is thought to compensate for the vascular structural abnormalities in cav-1^−/− mice. Decreased pulmonary artery filling defects are found in cav-1^−/− mice (62). Furthermore, less ACE MAb binding was found in the cav-1^−/− mice, suggesting vascular surface area may also be reduced in cav-1^−/− mice (62). Cav-1^−/− mice have basal pulmonary edema, with elevated extravascular lung water, and thus this apposing tissue pressure may limit further transport and accumulation of pulmonary edema fluid from vascular damage during lung injury (unpublished observation from Dr. R. Minshall). In addition, deletion of cav-1 results in activation of eNOS and dampens Toll-like receptor 4 (TLR4) signaling, thus decreasing the innate immune response to LPS which consequently confers protection from LPS-induced inflammation (70). These results further confirmed the previous studies by Garrean et al. (28), which showed that cav-1^−/− mice have a reduced inflammatory response following LPS via NF-κB-mediated pathways. Lv et al. (61) observed that overexpression of cav-1 aggravates LPS-induced inflammation in that mouse lung alveolar type 1 cells had increased inflammatory cytokine production including IL-6 and TNFα. Instead of using cav-1^−/−, the “gain of function” approach used in this study further endorsed the hypothesis that cav-1 mediates LPS-induced inflammation during lung injury (61, 70). Similarly, Wang and coworkers (109) first observed cav-1 expression in murine alveolar macrophages and showed that cav-1 in murine alveolar and peritoneal macrophages regulated LPS-induced proinflammatory TNFα and IL-6 cytokine production. Deletion of cav-1 increased LPS-induced proinflammatory cytokine TNFα and IL-6 production but decreased anti-inflammatory cytokine IL-10 production. Overexpressing cav-1 in RAW264.7 cells gave rise to opposite results. Thus, this study suggests that cav-1 acts as a potent proinflammatory effector molecule in immune cells.

EVIDENCE FOR A ROLE OF CAV-1 IN PMN-MEDIATED INFLAMMATION DURING ALI.

There are two recognized phases of ALI. The initial phase is acute and is characterized by disruption of the alveolar-capillary barrier, leakage of protein-rich fluid, extensive release of cytokines, migration of neutrophils, and severe hypoxia. All of these processes likely contribute to non-cardiogenic pulmonary edema. During this phase, PMNs are thought to play a critical role. PMNs mediate generation of free radicals, damage of the vascular-alveolar interface, and increased protein-rich fluid permeability (5, 41, 49, 62, 63, 69, 119). Cav-1 contributes to the PMN-mediated inflammation, vascular injury, and non-cardiogenic pulmonary edema associated with ALI (28, 39, 40). In a recent study by Hu et al. (39), formyl-Met-Leu-Phe (FMLP) was used as a PMN activator and phorbol ester phorbol 12-myristate 13-acetate (PMA) as an alternative stimulator (39). In wild-type mice, infusion of cav-1^+/+ PMNs followed by FMLP and PAF increased the pulmonary capillary filtration coefficient by 150% and lung wet-to-dry weight ratio by 50% (39). Increased PMN accumulation in lung tissue was also observed. In contrast, when cav-1^−/− PMNs followed by FMLP and PAF were infused into wild-type mouse lungs, the above effects were abolished (39, 40). Furthermore, deletion of cav-1 in PMNs reduced PMN adhesion to endothelial cells, inhibited PMN chemotaxis, transendothelial migration, and reduced PMN superoxide production (28, 39, 40). In this study, the authors did not address the mortality of mice receiving infusions of cav-1^+/+ PMNs vs. cav-1^−/− PMNs, but their results support the hypothesis that neutrophil cav-1 is associated with the acute inflammatory responses during lung injury. An earlier study using a different approach drew a similar conclusion. Instead of infusing cav-1^−/− PMNs into wild-type mice, Garrean and colleagues (28) observed marked attenuation of LPS-induced PMN sequestration and pulmonary edema in cav-1^−/− mice. In this study, PMN adhesion to endothelial cells and PMN sequestration in lungs following LPS challenge was reduced in cav-1^−/− mice due to reduced ICAM-1 expression (28). Adhesion of wild-type PMN to mouse lung vascular endothelial cells (MLVEC) after exposure of endothelial cells to LPS was markedly reduced in ECs cultured from cav-1^−/− mice relative to WT mice. Diminished ICAM-1 expression in cav-1^−/− mice paralleled the reduction in PMN binding to MLVEC cultured from cav-1^−/− mice. Furthermore, lung PMN sequestration following LPS administration was reduced significantly in cav-1^−/− lungs relative to wild-type mice. Garrean et al. assessed pulmonary microvascular liquid permeability by measuring the capillary filtration coefficient (K_f,c). LPS-induced elevation in K_f,c was observed in wild-type lungs but not Cav-1^−/− lungs. Although the K_f,_c measurement relies strongly on the vascular surface area and cav-1^−/− lungs may have a reduced functional vascular surface area, as mentioned previously (62), the effect of LPS on K_f,c in the isolated perfused mouse lung was significantly reduced in cav-1^−/− mice. While it is unclear how much of the dysfunctional vascular surface area is “recruited” during the pressure pulse of the K_f,_c measurement, significantly greater wet-to-dry weight ratios were observed in wild-type lungs after LPS challenge compared with that measured in cav-1^−/− lungs. Last, this study revealed that following administration of a relatively high dose of LPS, as much as 87% of wild-type mice died within 12 h in contrast to only 40% of cav-1^−/− mice receiving the same dose of LPS. These reports demonstrate that the absence of cav-1 results in resistance to the lethal effects of LPS (28, 39, 40).

Hoetzel et al. (38) reported that in their ventilator-induced lung injury (VILI) model, in contrast to the above-mentioned studies, cav-1 confers a protective role, especially in the presence of carbon monoxide (CO). They found that cav-1^−/− mice suffered significantly more severe lung injury in the VILI model compared with the wild-type mice, which was a three- to fivefold elevation in protein and cell counts in BAL fluid. This study, however, failed to show increased neutrophil recruitment in cav-1^−/− mice exposed to high tidal volume ventilation. Furthermore, this study did not investigate the survival discrepancy between wild-type mice and cav-1^−/− in the VILI model.

Cav-1 regulates DAD during lung injury.

CAV-1 REGULATES EPITHELIAL AND ENDOTHELIAL CELL DEATH IN LUNG INJURY.

For decades, it has been appreciated that ARDS, the severe form of ALI, is characterized by DAD (5, 41, 49, 62, 63, 69, 119). Lung epithelial cell death is a main feature of DAD. Apoptosis and other causes of cell death, including necrosis and oncosis, have been observed in ALI and attributed to a number of mechanisms (5, 41, 49, 62, 63, 69, 119). Franek et al. (22) identified apoptosis as a prominent component of the acute inflammatory response in the lung. The percentage of apoptotic cells was correlated with the severity of lung injury after hyperoxia (22). Recent studies illustrated that deletion of cav-1 protects against hyperoxia-induced cell death (43, 44, 122). Using the human bronchial epithelial cell line (Beas2B) and/or primary fibroblasts and endothelial cells isolated from the wild-type and cav-1^−/− mice, Zhang et al. (122) found that deletion of cav-1 confers cytoprotection via upregulation of cytoprotective genes, such as hemeoxygenase (HO-1) and survivin (43, 44, 122). Survivin, a member of the inhibitors of apoptosis protein family, directly binds with caspase-3 and limits its activity. Thus, by regulating the level of survivin, cav-1 controls a common pathway of apoptosis in lung epithelial cells following hyperoxia exposure (43, 44, 122). However, in the above-mentioned studies, primary lung epithelial type I cells were not used, but instead, human bronchial cell lines were the major cells used for in vitro models. Thus, the above observations require further investigation using primary lung cells. Similar to survivin, deletion of cav-1 upregulates generic cytoprotective machinery (122). HO-1 confers well-known cytoprotection after hyperoxia (50, 51, 78). Compared with wild-type mice, cav-1^−/− mice showed a markedly elevated HO-1 level and significant resistance to hyperoxia-induced death in vivo (43, 44, 122). Consistent with LPS/sepsis-induced lung injury models, cav-1^−/− mice survived, on average, 2 days longer than wild-type mice after hyperoxia and showed less lung injury (43, 44, 122).

Cav-1 regulates bleomycin-induced lung injury and fibrosis.

The model of bleomycin-induced lung injury has been widely applied in pulmonary research. Bleomycin stimulates reactive oxygen species (ROS) production resulting in oxidative stress and pulmonary fibrosis (46, 72, 108). Bleomycin further induces apoptosis and senescence in epithelial and non-epithelial cells of the lung (46, 72, 108). Initial reports showed that cav-1 is upregulated 1 h after exposure to bleomycin, before the appearance of caspase-8, caspase-3, and caspase-9 cleavage products (46, 72, 108). Bleomycin leads to a partial translocation of cav-1 from lipid raft membrane fractions to non-raft fractions (46). Recently, additional studies suggested that bleomycin-induced injury of lung cells is accompanied by altered expression levels of cav-1 (73, 111). However, in the bleomycin-induced lung injury model, cav-1 was observed to play a greater role in epithelial cell growth arrest and senescence rather than promoting apoptosis (46, 72, 73, 108, 111). Reduction of cav-1 expression using siRNA and antisense oligonucleotides restored DNA synthesis and allowed senescent cells to re-enter the cell cycle (7). Dasari et al. (11) showed that oxidative stress induced premature senescence by stimulating cav-1 gene transcription through p38 mitogen-activated protein kinase/SP-1-mediated activation of cav-1 promoter elements (11). It is also known that cav-1 overexpression is sufficient to arrest mouse embryonic fibroblasts in the G₀/G₁ phase of the cell cycle, reduce their proliferative life span, and promote premature cellular senescence through activation of the p53/p21 axis (25, 107). However, despite these findings, the exact molecular mechanisms linking cav-1 to the induction of growth arrest, particularly in the G₂/M phase, remain to be determined. Linge et al. (57) confirmed that bleomycin increases the expression of cav-1 in A549 cells and that deletion of cav-1 prior to bleomycin-exposure modulates p53 and p21 levels and subsequently prevents growth arrest in the G₂/M phase. It was speculated that cav-1 inhibits cell proliferation by suppressing STAT3 activation, although details of these mechanisms require further investigation. Although this study, again, carries a limitation that A549 lung epithelial cell lines were used instead of primary lung epithelial cells, it supports a key role for cav-1 in modulating the cell cycle of bleomycin-damaged lung cells.

Multiple studies have demonstrated that cav-1 possesses an antifibrotic feature, led by Wang et al. (111), showing that cav-1 markedly ameliorates bleomycin-induced pulmonary fibrosis (7, 11, 12, 57, 73, 107, 108). This is consistent with the aforementioned characteristics of cav-1 in lung cell biology. Overall, cav-1 possesses antiproliferative and proapoptotic effects on all cell types including fibroblasts. In reflection, cav-1 shows an antifibrotic feature in lung fibrosis due to fibroblast apoptosis. Unfortunately, there are no survival data in cav-1^−/− mice in the bleomycin-induced lung injury models.

The above reports largely rely on cell culture and/or animal studies, and thus, despite the plethora of information gleaned from these studies, there is limited information from human studies to compare with these and thus validate their relative usefulness. Wang et al. (111) and Tourkina et al. (102) reported marked reduction in cav-1 expression in lung tissues or monocytes/neutrophils from patients with idiopathic pulmonary fibrosis or scleroderma, respectively, compared with controls (102, 103, 111). Moreover, Tourkina et al. demonstrated the potential role of cav-1 as a lung-protective protein by using the cav-1 CSD peptide to treat scleroderma/fibrosis (102, 103, 111).

Future directions for assessing the role of cav-1 in ALI include: 1) determination of whether cav-1 can act as a potential biomarker for lung injury, and 2) whether development of therapeutic protocols using the cav-1 CSD peptides will be practical.

Additional cellular and molecular mechanisms by which cav-1 participates in ALI.

The above-mentioned studies provide robust evidence that cav-1 is an important regulator in the pathogenesis of ALI in various models. These reports assessed lung injury using multiple parameters including wet:dry weight ratio, morphological/histological examination, non-cardiogenic pulmonary edema, protein/cell counts in the BAL fluid, and mortality. In this section, we will discuss some additional hypotheses and evidence in support of the cellular/molecular mechanisms by which cav-1 regulates lung injury.

Does cav-1 play a role in the pathogen uptake by host cells in lung injury?

There is evidence showing that cav-1 is involved in both bacteria and virus uptake by host cells. Two recent studies on the function of cav-1 in Pseudomonas aeruginosa-induced pneumonia revealed variable results. Zaas et al. (118) found that deletion of cav-1 decreases lethality from P. aeruginosa acute pneumonia (118), whereas Gadjeva et al. (25) observed that cav-1^−/− mice have increased sensitivity to P. aeruginosa infection, with an increased mortality rate. Interestingly, in the first study, Zaas et al. probably used a more cytotoxic strain of P. aeruginosa. In contrast, strains used in the second study (Gadjeva et al.) were more invasive but not cytotoxic. Furthermore, when Gadjeva et al. infected mice with the exotoxin U-positive and cytotoxic P. aeruginosa strain PA14, the previously observed survival benefits in wild-type mice diminished (25). These two studies suggest that in the model of P. aeruginosa-induced pneumonia, the cytotoxicity of the bacterial strain plays a critical role in determining the outcome observed in cav-1^−/− mice.

Although the underlying mechanisms by which cav-1 regulates pulmonary infection-associated lung injury remain unclear, studies have shown evidence supporting a role of cav-1 in pathogen endocytosis and cell signaling. Eight caveolin-binding sites have been identified in severe acute respiratory syndromes (SARS) coronavirus. These caveolin-binding sites are located in replicase 1AB, spike protein, orf3 protein, and M protein (79). Furthermore, colocalization of cav-1 and orf3a (the largest unique ORF in the SARS coronavirus genome) has been confirmed (60). These studies provide evidence that cav-1 is involved in regulating pathogen-induced lung injury. However, not all viruses rely on cav-1/caveolae to enter host cells. For instance, internalization of coxsackievirus A9 into A549 lung epithelial cells does not require cav-1 but is dependent on integrin αVβ6, β2-microglobulin, dynamin, and Arf6 (35). Additionally, whether cav-1 protein or the structure of caveolae plays a role in pathogen uptake remains unclear. Delineation of the mechanisms by which cav-1/caveolae mediate pathogen uptake by host cells will assist future development of therapeutic options for infection-associated lung injury.

Does cav-1 play a role in antigen processing and immune response in lung injury?

In non-infection-associated ALI, an unspecified immune response has been suspected to trigger severe inflammation leading to DAD. Emerging evidence suggests that cav-1 is involved in this process. Ohnuma et al. (77) first showed that T cell costimulatory molecule CD26 interacts with cav-1, resulting in antigen-specific T cell activation. A series of studies further demonstrated that CD26 binds to cav-1 in antigen-presenting cells (APC). The interaction between CD26 and cav-1 induces T cell proliferation. More interestingly, blocking CD26-caveolin-1 costimulation with soluble cav-1-Ig (Cav-Ig) inhibits T cell proliferation and cytokine production in response to recall antigen (74, 75, 76). This result potentially provides a new therapeutic approach by using soluble Cav-Ig as an immunosuppressive agent. More evidence from Medina et al. (65, 66, 67) suggests that cav-1 modulates immune responses against pathogens. In their study, cav-1^−/− mice showed increased production of inflammatory cytokines, chemokines, and nitric oxide, and cav-1^−/− macrophages displayed increased inflammatory responses to LPS.

Does cav-1 play a role in apoptotic and/or autophagic pathways in lung injury?

Numerous studies have shown that cav-1 regulates apoptotic pathways at various stages of lung injury. Death receptor agonists including Fas ligand, TNFα, and TRAIL have all been reported to interact with or upregulate cav-1 (13, 87, 120). Cav-1 protein can be significantly induced by TNFα and TRAIL (13, 87, 120), whereas Fas has been suspected to interact with cav-1 upon stimulation by FasL (122, 123). Elevated caspase activities and an increased active form of caspase-3 are known features of hyperoxia-induced lung cell apoptosis (43, 44, 122, 50, 51, 78). Deletion of cav-1 blunts these responses after hyperoxia (43, 44, 122). Further investigations suggest that cav-1 decreases survivin, the inhibitor of apoptosis, and directly regulates apoptotic pathways (43, 44, 122). Additionally, a recent report suggests that cav-1 physically interacts with LC3B, the key autophagic marker protein, at baseline. Upon exposure to ROS, this interaction diminishes, suggesting a regulatory role of cav-1 in the process of autophagic cell death during lung injury (6a).

CONCLUSIONS

Cav-1 is involved in PMN adhesion, chemotaxis, and epithelial and endothelial cell apoptosis/senescence during lung injury. These features lead to increased alveolar damage (DAD), vascular leakage (non-cardiogenic pulmonary edema), and profound inflammation during the initial stage of ALI. However, as a potential antifibrotic protein, cav-1 may be beneficial in the late stage (fibrotic phase) of lung injury, although this remains unclear and requires further investigation. The overwhelming association of caveolin-1 in the pathology of ALI suggests that additional cell- and animal-based experiments, in combination with clinical observations, have the potential to give rise to important therapeutic tools and interventions for treating ALI/ARDS in the future.

DISCLOSURES

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