This review summarizes the highlights of the EB2004 symposium that dealt with the integrated aspects of the lung fluid balance. It is apparent that maintenance of lung fluid balance requires the proper functioning of vascular endothelial and alveolar epithelial barriers. Under physiological conditions, the transcytotic pathway requiring repeated fission-fusion events of the caveolar membrane with other caveolae solely transports albumin. Caveolin-1, which forms caveolae, and albumin-binding proteins play a central role in signaling the transcytosis of albumin. Signals responsible for increasing endothelial permeability in lung microvessels in response to inflammatory mediators were also described. These studies in gene knockout mouse models revealed the importance of Ca2+ signaling via store-operated transient receptor channel 4 and the activation of endothelial myosin light chain kinase isoform in mediating the increase in microvessel permeability. Increases in the cytosolic Ca2+ in situ in microvessel endothelia can occur by mitochondria-dependent as well as mitochondria-independent pathways (such as the endoplasmic reticulum). Both these pathways, by triggering endothelial cell activation, may result in lung microvascular injury. The resolution of alveolar edema, requiring clearance of fluid from the air space, is another area of intense investigation in animal models. Although β-adrenergic agonists can activate alveolar fluid clearance, signaling pathways regulating these events in intact alveoli remain to be established. Development of mouse models in which the function of regulatory proteins (identified in cell culture studies) can be systematically analyzed will provide a better and more integrated picture of lung fluid balance. In vivo veritas!
- albumin binding proteins
- intracellular calcium
- store-operated transient receptor channel 4
- microvessel permeability
normal pulmonary function and tissue integrity depend on the maintenance of fluid balance. According to Starling's equation, tissue fluid balance is dependent on two countering forces: the filtration pressure, vascular hydrostatic pressure plus interstitial colloid osmotic pressure, which determines the filtration of fluid, and absorptive pressure, plasma colloid osmotic pressure plus interstitial hydrostatic pressure, which determines the reabsorption of fluid (38, 45). Given the vast surface area of the pulmonary microcirculation and its close proximity to the gas-exchanging alveolar epithelial barrier, the integrity of both microvessel endothelial and alveolar epithelial barriers is essential for lung fluid balance (38, 40). In the normal lung, the active transport of albumin across the endothelial barrier via the transcellular pathway, requiring the transcytosis of albumin via caveolae, is likely responsible for maintaining the tissue oncotic pressure and thus helps to maintain the hydration state of the interstitium (49, 91). However, the proteins that do leak across endothelium are cleared by the extensive system of lymphatics. Inflammatory mediators such as thrombin and TNF-α, generated in response to sepsis and intravascular coagulation, disrupt the endothelial barrier by forming intercellular gaps, which are formed in part from actinomyosin-regulated contraction of endothelial cells (8, 38). These gaps permit the passage of albumin and other plasma proteins in an unrestricted manner, resulting in increased protein concentration in the interstitial tissue. This condition clinically manifests as high-permeability, protein-rich pulmonary edema. This fluid may enter the air space leading to alveolar “flooding.” Resolution of air space pulmonary edema requires a functionally active alveolar epithelial barrier (40). The inability to clear alveolar protein results in the accumulation of fluid and macromolecules in the alveolar space and severe gas exchange impairment. Although many of these advances made through studies on endothelial and epithelial cells in isolation or cultured monolayers have been seminal, the picture at the level of microvessels and the alveolar space in lungs is less clear. This symposium served to highlight newer studies employing animal models (e.g., genetically modified mice). The specific topics covered are presented below.
ENDOTHELIAL BARRIER AND ALBUMIN TRANSCYTOSIS1
The cells lining the intima of blood vessels comprise an endothelial cell monolayer that functions as a semipermeable barrier, controlling the exchange of macromolecules and fluids between the blood and interstitial space. This tenet is based on experimental data that demonstrate that under basal conditions, the endothelium functions as a barrier to fluid exchange (38), while at the same time it actively translocates diverse macromolecules [e.g., albumin (22), insulin (4), transferrin (12), ceruloplasmin (31), LDL (50), angiotensin (47), and orosomucoid (64)]. Albumin, as the most abundant plasma and interstitial fluid protein, is responsible for maintaining the plasma oncotic pressure as well as acting as a carrier for an array of molecules (e.g., fatty acids, amino acids, cholesterol, bilirubin, hormones) (61). It is transported primarily in continuous endothelia via the transcellular route under basal conditions (21, 44, 46). Albumin was initially thought, based primarily on theoretical interpretations of physiological observations, to transit the endothelial barrier through “pores” or filtration slits, assumed to be located along interendothelial junctions (IEJs) (57). The advent of electron microscopy (EM) coupled with increasingly sophisticated biochemical and molecular biological techniques have led to the current hypothesis that many of these proteins are transported across the endothelial cell barrier in an energy-dependent and a receptor-mediated process. Data obtained using tracers showed that endothelial caveolae function in the transport of macromolecules from the blood to the interstitial fluid (82). This mechanism of transcytosis has been demonstrated for albumin whose transendothelial pathway has been shown to be dependent on caveolae microdomains found in large number on the surface of endothelial cells. Caveolae are rich in cholesterol and sphingolipids (22, 29, 48, 63). This mechanism of macromolecular egress from the vasculature has been described as “receptor-mediated transcytosis” (83). This model of transcytosis was initially proposed on the basis of EM analysis (83) and subsequently gained support from a series of functional studies (28, 29, 48, 63, 78, 93, 95). Subsequent to receptor-ligand binding and the initial stages of internalization, the cargo is thought to move across the cell guided by repeated fission-fusion events of caveolar membranes at the plasmalemma and subsequently with other caveolae and ultimately with the apposing plasmalemma (83). Experimentally it has been shown that vesicular carriers internalize tracers (through endocytosis), first loading the cargo on one side of the cell and then discharging it through exocytosis on the opposite side of the cell (80).
Recent work by our group (29, 63) (Fig. 1) and others (21, 25, 54, 101) showed that albumin traverses the intact endothelial barrier primarily via caveolae. Experiments with fluid phase markers horseradish peroxidase (30), dextran, and glycogen (84) and blood-borne proteins immunoglobulin G (62), fibrinogen (35), and albumin have demonstrated that a significant component of macromolecular transport occurs by fluid phase. Therefore, whereas albumin can be transported after binding albumin-binding proteins (ABPs), it can also move across the endothelium by fluid phase (29, 81). Thus the permeability of the vessel wall to albumin has the opportunity to be regulated at many levels.
Endothelial cells typically contain a large number of caveolae (10,000–30,000 caveolae/cell) having a diameter ranging from 50 to 100 nm (Fig. 2A). These microdomains are contiguous with the surrounding plasmalemma as long as the vesicles are open to either the luminal or abluminal cell surface. As the cell surface caveolae begin to internalize and form vesicles they continue to communicate with the vascular lumen or perivascular space through developing necks of variable dimensions (56). Vesicular caveolae are found in the cytoplasm and have long been thought to be the result of internalized caveolae that have pinched off from the plasma membrane (55). Interestingly, the large numbers of caveolae present in the capillaries of lung (23) effectively double the endothelial cell surface (68). In an attempt to compare the caveolar density of heterogeneous vasculature in the lung, we found, using a monomeric albumin tracer, dinitrophenyl albumin, that the highest number of caveolae was present in capillaries (Fig. 2B). These findings support the previous hypothesis that the lung demonstrates segmental variation with respect to caveolar density (85, 86). Although a rare event, it was consistently seen that under basal conditions, a single caveola or a short chain of two to four caveolae will fuse with both sides of the plasmalemma generating transendothelial channels (83). In pathological conditions, such as in inflammation, multiple caveolae have been reported to aggregate to form large vacuoles or vesiculo-vacuolar organelles (VVOs) (9). It has been suggested that these channels or VVOs function as the originally postulated “pores” (9).
As mentioned above, a significant facet of caveola-based transport of albumin is its association with endothelial cell surface receptors (20, 29, 48). Albumin has been shown to bind to endothelial cells by a receptor-mediated process: the binding is saturable, subject to competition by unlabeled albumin and is temperature sensitive (29, 102). These and other studies led to the search for endothelium-specific ABPs that regulate the receptor-mediated component of albumin transcytosis. Four such proteins were identified (19); however, at present these proteins remain only partially characterized (19, 78). Using albumin anti-idiotypic antibodies developed in our laboratory, we immunoprecipitated four proteins that migrated by SDS-PAGE as the previously described ABPs (i.e., 16, 32, 60, 72 kDa) (19, 22, 78, 88). Immunostaining of cultured lung microvascular endothelial cells with the anti-idiotypic antibodies showed a diffuse membrane distribution of ABPs at the light microscopy level. Morphometric analysis of the ABP distribution showed that these proteins were concentrated approximately twofold in caveolar regions of the plasma membrane. Future studies using anti-idiotypic antibodies will, hopefully, seek to more clearly define the function of ABPs in albumin transcytosis, in particular albumin interactions as regulated by ABPs in endothelial cells and the role of albumin binding to ABP in regulating transcellular transport of albumin.
The major caveolar component of caveolae, caveolin-1 (Cav-1), also known as VIP-21, is the canonical biochemical marker of caveolae (24, 36). At the ultrastructural level, Cav-1 decorates the cytoplasmic aspect of a caveolae and appears as a ridged structure (60, 70). Genetic analysis has identified three members of the caveolin family (Cav-1, Cav-2, and Cav-3) (77), each encoded by its own gene (13). Cav-1 contains three exons, whereas Cav-2 and Cav-3 have two exons each. The boundary of exon 3 is identical in all three genes, and this exon encodes most of the functional domains of caveolin (i.e., homo-oligomerization, scaffolding, membrane spanning, and COOH-terminal domains). Cav-1 and -2 interact specifically with one another; however, the three genes are differentially expressed in tissues (76). Cav-1 and Cav-2 are coexpressed in all tissues; their gene products form hetero-oligomers in many cell types and are particularly abundant in endothelial cells (59). The expression of Cav-3 is muscle specific (87). Using cells expressing exogenous Cav-1, several groups showed that there is a direct correlation between the level of Cav-1 expression and the appearance of caveolae (14, 37, 96). However, there are exceptions as in some cell types (e.g., endothelial cells from brain microvessels and type 1 alveolar epithelial cells): Cav-1 is present, but there are relatively few caveolae (1, 5). If a correlation between the amount of Cav-1 protein present in a cell and the number of caveolae present in that cell is valid, we predict that there are likely important quantitative differences in Cav-1 expression along successive lung vascular segments that reflect caveolae number and transcytosis. This is based on data as in Fig. 2 that in lung there are more caveolae in capillaries than in arteries or veins.
Results have shown that caveolae are absent in vascular endothelia of Cav-1−/− mice (66). Studies also showed impairment in the endocytosis of FITC-albumin and transendothelial transport of 125I-labeled albumin in Cav-1−/− mice (66, 79). Using our in situ system of perfusion, we followed the pathway of the albumin-Au tracer in Cav-1−/− mice. We also observed that caveolae were absent in endothelial cells of all microvascular beds examined and there was also no evidence of albumin transcytosis in these areas (Fig. 3). Although deletion of Cav-1 in mice resulted in the absence of caveolae in endothelial cells (7, 66), studies also noted prominent invaginations or caveola-like structures with electron-dense diaphragms at their necks in endothelia of some vessels (7, 104). These “Cav-1-independent vesicles” may in their own right be important in mediating transcellular albumin permeability in lungs, but their molecular composition and function in vessel endothelia are unknown. There may be a fundamentally important inverse relationship between Cav-1 expression and endothelial permeability via IEJs. This concept is based on the evidence that the transcellular transport of albumin is impaired in Cav-1 null mice, but IEJ albumin permeability is increased in small veins and capillaries (79). However, the mechanisms by which Cav-1 can regulate the permeability of the IEJ barrier remain unclear.
In summary, the vessel wall permeability to albumin, via transcytosis, is responsible for maintaining endothelial barrier function under basal conditions. Transcytosis of albumin requires repeated fission-fusion events of caveolar membranes with the plasmalemma and with other caveolae. Cav-1 and ABPs appear to play a central role in albumin transcytosis. An important issue that remains unresolved is the quantification of albumin transcytosis in intact microvessels.
COMPARTMENTALIZED CA2+ SIGNALING IN LUNG MICROVESSELS2
The pathological consequences of increasing the lung capillary pressure (Pc) include major clinical conditions such as pulmonary edema and lung injury. The underlying pathology is attributed classically, to increases in the transmicrovascular fluid transport. However, the extent to which Pc elevation induces direct pathogenic effects on capillary endothelial cells (EC) remains inadequately understood. We addressed endothelial cell responses to Pc elevation in venular capillaries of the ex vivo, blood-perfused rat lung by means of real-time optical imaging (33, 34, 103). We determined cytosolic Ca2+ by the fura-2 method and EC expression of the leukocyte adhesion receptor P-selectin by in situ immunofluorescence. This approach quantifies the surface expression of endothelial cell P-selectin, resulting from the Ca2+-induced exocytosis of P-selectin-containing Weibel-Palade bodies (WPBs). In lung capillaries, increases of Pc from a baseline of 5 cmH2O progressively increased endothelial cell Ca2+ (33). A sustained increase to 20 cmH2O evoked marked Ca2+-dependent enhancement of surface P-selectin (103), indicating that even modest Pc increases causes proinflammatory endothelial cell EC activation. Surprisingly, the associated Ca2+ increases were relatively small, being in the 50–60 nM range, different from the micromolar increases reported for endothelial cells in vitro. However, it was noted in these studies that the major effect of Pc elevation on the cytosolic Ca2+ was to increase the amplitude of Ca2+ oscillations.
Continuing experiments in lung venular capillaries, Ichimura et al. (27) instituted a 10-min Pc increase to 15 cmH2O to double the Ca2+ oscillation amplitude without changing the mean Ca2+. Concomitantly, mitochondrial Ca2+ increased in endothelial cells, as detected in fluorescence increases in capillaries loaded with the mitochondrial dye rhod 2. The mitochondrial Ca2+ increase occurred with a slight delay following the amplitude increase in cytosolic Ca2+. Both responses decayed immediately after Pc was decreased to baseline, and both were repeatable in two successive pressure challenges, indicating that the responses were not attributable to capillary injury.
The mitochondrial inhibitors rotenone, the complex 1 blocker, and FCCP, the mitochondrial depolarizer, each inhibited the mitochondrial, but not the cytosolic, Ca2+ response, reaffirming the mitochondrial origins of the rhod 2 responses. However, xestospongin C, the inhibitor of the inositol trisphosphate (IP3) receptor on the endoplasmic reticular (ER) membrane, inhibited both the cytosolic and mitochondrial Ca2+ responses. Together, these findings indicate that Pc increase caused IP3-induced store release of Ca2+ from the ER, then possibly, to Ca2+ uptake in juxta-ER mitochondria. A major consequence of the mitochondrial Ca2+ increase was the increased production of reactive oxygen species (ROS), which Ichimura et al. (27) detected in capillaries loaded with the ROS-detecting dye dichlorofluorescin. Ruthenium red, the inhibitor of mitochondrial Ca2+ uptake, and the inhibitors of mitochondrial electron transport, rotenone and FCCP, each blocked the pressure-induced increases in ROS production, indicating that the ROS were of mitochondrial origin. Because mitochondrial ROS may damage the mitochondrial inner membrane, Ichimura et al. (27) determined that the fluorescence of the dye tetramethylrhodamine methyl ester, which detects the mitochondrial membrane potential, remained unchanged during the pressure challenge, confirming the absence of mitochondrial damage. Importantly, the membrane-permeable form of catalase, polyethylene glycol (PEG)-conjugated-catalase, which inhibits H2O2, completely abolished the pressure-induced fluorescence increase, identifying H2O2 as the major mitochondrial ROS produced by Pc elevation. PEG-catalase, as also the mitochondrial inhibitors ruthenium red, rotenone, and FCCP, each blocked the pressure-induced P-selectin expression. The novel conclusions to be drawn from these findings are that pressure-induced deformation stress on endothelial cells induces mitochondrial H2O2 production and that the mitochondrial H2O2 diffuses to the cytosol to cause WPB exocytosis, resulting in P-selectin expression on the endothelial cell surface.
An important insight into the manner by which the endothelial cell compartmentalizes Ca2+-induced proinflammatory responses was revealed through the studies of Parthasarathi et al. (58). Also working on lung venular capillaries, these workers report that although TNF-α induced similar mitochondrial Ca2+ and H2O2 responses to Pc elevation, arachidonate induced ROS production through nonmitochondrial sources. The conclusion to be drawn is that whereas receptor-mediated signaling recruits mitochondria for proinflammatory endothelial cell responses, nonreceptor agonists such as arachidonate short circuits the mitochondrial route for other ROS-producing mechanisms. Preliminary data from the Bhattacharya laboratory suggest that nonreceptor Ca2+ mobilization utilizes the NAD(P)H oxidase system for Ca2+ mobilization.
In summary, real-time optical imaging of lung venular capillaries has revealed the feasibility of determining compartmentalized Ca2+ signaling in lung endothelial cells in situ. Studies conducted thus far indicate that increases in the cytosolic Ca2+ are sufficient for proinflammatory activation of EC. However, this activation is highly regulated by mitochondria-dependent and mitochondria-independent pathways (Fig. 4). Although both pathways lead to ROS production and P-selectin exocytosis, they are activated by receptor-induced and receptor-independent pathways, respectively. The physiological significance of this compartmentalization may relate to the regulation of endothelial cell activation. One possibility is that the receptor-mediated mitochondrial pathway restricts activation to the few mitochondria-rich endothelial cells of lung capillaries (58), whereas the receptor-independent pathway establishes a more widespread inflammatory response in the microvascular bed.
ROLE OF CONTRACTILE MECHANISMS IN REGULATING LUNG MICROVASCULAR PERMEABILITY3
Endothelial cells regulate their barrier properties by altering their cell shape, which was earlier described ultrastructually as “cell rounding” by Majno and Palade in 1961 (39). Cell shape change occurs as a result of increased contraction of endothelial cells via actinomyosin-generated tension activated by myosin light chain (MLC) phosphorylation and actin polymerization (8, 38). However, little is known about the significance of this mechanism in the intact microcirculation. Thrombin, a serine protease, proteolytically activates the protease activating receptor-1 (PAR-1) and increases endothelial permeability (6). PAR-1 activation was shown to form junctional gaps in endothelial monolayer and does so by activating actinomyosin-generated contractile tension (8). It is known that PAR-1 increases contractile force in endothelial cells by inducing MLC phosphorylation through activation of heterotrimeric Gαq and Gα12/Gα13 proteins that are coupled to PAR-1 (8, 72, 90).
One of the mechanism by which PAR-1 activation induces MLC phosphorylation is through activation of endothelial isoform of myosin light chain kinase (MLCK) referred to as EcMLCK (17, 92). EcMLCK contains all the domains present in smooth muscle cell isoform but also contains a long NH2 terminus, which has an Src binding site and Src homology (SH) 2 and SH3 domains (17, 92). MLCK activation requires increased cytosolic calcium concentration. PAR-1 through Gαq activates MLCK by resulting in a transient increase through depletion of cytosolic Ca2+ stores and activation of Ca2+ entry via transient receptor potential channels (TRPC) (51, 65, 90). In addition, Src-induced tyrosine phosphorylation of MLCK has been shown to regulate MLCK activity (18). The inhibition of MLCK using pharmacological agents inhibited PAR-1-induced endothelial cell contraction and the loss of barrier function (8). These findings indicate that MLCK activation plays an important role in regulating the endothelial barrier function.
The other mechanism by which PAR-1 activates MLC phosphorylation is by activating a small GTPase, RhoA. Studies from our labs and others showed that PAR-1 through Gα12/Gα13 and Gαq proteins stimulates RhoA activity (26, 32, 72, 97). RhoA increases MLC phosphorylation via Rho-kinase-mediated inhibition of MLC phosphatase (8, 10). In addition, we showed that RhoA regulates the increase in Ca2+ by activating Ca2+ entry through TRPC1 (42). As the increase in cytosolic Ca2+ is critically required for activation of MLCK it is possible that an effector arm of the Rho signaling may involve EcMLCK to regulate endothelial cell contractility. Evidence in support of this conclusion has been reported by Garcia et al. (18). They showed that Src-dependent tyrosine phosphorylation of EcMLCK occurred through a Rho-dependent pathway. RhoA activation requires upstream effectors that convert Rho-GDP (inactive state) to Rho-GTP (active state) (52, 69). Three different classes of proteins are required for this regulation: 1) guanine nucleotide exchange factors, which stimulate the GTP-GDP exchange reaction; 2) GTPase-activating proteins, which stimulate the GTP-hydrolytic reaction; and 3) guanine nucleotide dissociation inhibitors (GDIs), which, by binding to Rho, block dissociation of GDP from Rho GTPases. Furthermore, GDI is also capable of inhibiting GTP hydrolysis by Rho family GTPases, as well as stimulating release of Rho-GTPases from cellular membranes, and thereby shutting off the Rho cycle. In cultured cells, we showed that PAR-1 activation rapidly induced the phosphorylation of GDI-1 and activation of RhoA, indicating PAR-1 may activate Rho and thus the increase in endothelial permeability by modulating GDI function (43). Coincident with this idea, recent data show that thrombin failed to induce Rho activation and loss of endothelial barrier function in endothelial cells overexpressing a GDI-1 mutant (71). Thus RhoA and MLCK, activated sequentially or in parallel upon PAR-1 activation, regulate the status of MLC phosphorylation in the endothelial cells and therefore may be intimately involved in regulating lung microvessel permeability.
Recent studies in knockout mice have begun to uncouple the relevance of PAR-1 signaling events in regulating permeability of microvessels. We showed that deletion of PAR-1 prevented thrombin-induced increase in lung microvessel permeability and MLC phosphorylation (94). We also showed the importance of Ca2+ entry in the PAR-1-induced increase in lung microvessel permeability (89). In wild-type mice, PAR-1 receptor activation induced increase in lung microvessel permeability was reduced to 50% by lanthanum (La3+), a known Ca2+ channel blocker, indicating Ca2+ entry contributes to the increased lung vascular permeability. Deletion of the TRPC4, a component of store-operated channels (SOC) in mouse endothelial cells, inhibited thrombin-induced Ca2+ entry, indicating TRPC4 is an essential constituent of the SOC in mouse endothelial cells and thereby for Ca2+ entry (89). Deletion of the TRPC4 gene in mice significantly interfered with the thrombin-induced increase in lung microvessel permeability, which was not affected by La3+. Thus these findings indicate that PAR-1-induced activation of Ca2+ store depletion and calcium entry via TRPC4 is a determinant of liquid permeability in the lung microvasculature. Preliminary findings indicate that deletion of EcMLCK prevented thrombin-induced increase in microvessel permeability. Thrombin also failed to induce MLC phosphorylation in lung homogenates from MLCK knockout mice. Thus EcMLCK appears to be the key effector mediating the PAR-1-induced increase in lung vascular permeability in vivo. Future studies will be required to establish the role of Rho and MLCK activation in PAR-1 regulation of endothelial barrier function in an intact microcirculation.
In summary, studies from gene knockout mice give credence to a model based on cell culture data in which cell surface receptor activation triggers both the rise in cytosolic Ca2+ via TRPC and Rho activation (after its liberation from GDI) (Fig. 5). These events then result in engagement of the contractile machinery via the activation of the EcMLCK, which activates the actin-myosin contractile apparatus, thus resulting in increased lung microvessel permeability.
MECHANISMS OF ACTIVE FLUID CLEARANCE FROM THE LUNG4
Remarkable progress has been made in the last two decades in understanding the basic mechanisms responsible for removal of excess fluid from the distal air spaces of the lung. A recent major review summarized much of the progress in studies of the mature lung (40). This portion of the symposium on integrated lung fluid balance focused on new information related to the regulation of alveolar fluid clearance, under both normal and pathological conditions. There is a recent comprehensive review of lung epithelial fluid transport in the developing lung (53).
Alveolar fluid clearance is driven by active ion transport (40). Currently, the molecular basis for sodium uptake across the alveolar epithelium is reasonably well understood. Epithelial sodium channels (ENaC) are located on the apical surface of both alveolar epithelial type II cells as well distal airway cells. Sodium is absorbed down a concentration gradient through ENaC and then is pumped by Na+-K+-ATPase across the basolateral membrane into the interstitial space (Fig. 6). This transport of sodium creates a miniosmotic gradient that functions to reabsorb water from the air spaces of the lung in an isosmolar fashion. The pathways for chloride transport have not been established. There is also new information that suggests that alveolar epithelial type I cells may participate in sodium transport and vectorial fluid clearance. ENaC and Na+-K+-ATPase have been identified at both the mRNA level as well at the protein level in type I alveolar epithelial cells. However, it has not yet been possible to study these cells under polarized conditions in cell culture.
Alveolar fluid clearance is regulated by both catecholamine-dependent and catecholamine-independent mechanisms. Catecholamine-dependent mechanisms depend on cAMP-stimulated transport. This is an important pathway, which is present in most species, including the human lung (73, 74, 99). There is impressive evidence that, under some circumstances, endogenous catecholamine stimulation can provide short-term upregulation of alveolar fluid clearance in the presence of severe septic or hypovolemic shock. However, new information from clinical studies of patients with pulmonary edema indicates that the levels of plasma catecholamines, specifically plasma epinephrine and plasma norepinephrine, probably are not sufficiently elevated in most patients to stimulate alveolar fluid clearance. This is understandable because most patients with severe pulmonary edema require intubation, ventilation, and treatment with both narcotics and sedatives, therapies that will reduce the release of endogenous catecholamines. Therefore, it is likely that aerosolized therapy will be needed to upregulate alveolar fluid clearance in patients with hydrostatic or increased permeability pulmonary edema. One clinical study has established that therapeutic concentration of a β2-adrenergic agonist can be achieved with aerosolization every 4 h of a standard β2-adrenergic agonist (2). Also, administration of an inhaled β2-adrenergic agonist was successful in decreasing the incidence of high-altitude pulmonary edema in a double-blind, placebo-controlled study (40, 75).
There is new information that cAMP-stimulated fluid clearance may be mediated both by ENaC and CFTR. The role of ENaC has been well established in several studies. More recently, work from our group indicated that deficiency of CFTR in ΔF508 mice is associated with the inability to stimulate fluid clearance (11). Additional pharmacological studies in mouse and human lungs have also supported the conclusion that CFTR plays an important role in cAMP-mediated fluid clearance across alveolar epithelium (11) (Fig. 6). This information provides a potential molecular basis for chloride uptake under cAMP-stimulated conditions and matches well with other studies of ENaC and CFTR interdependence in other epithelia, such as sweat glands (67).
Under pathological conditions, impaired alveolar fluid clearance is associated with worse clinical outcomes in patients with acute lung injury (41, 100). There are several potential mechanisms that may explain why alveolar fluid clearance is submaximal in patients with alveolar fluid clearance (40). First, the degree of lung endothelial permeability may be so severe that the extent of transvascular fluid flux and subsequent alveolar flooding overwhelms alveolar epithelial fluid transport mechanisms. Second, alveolar epithelial injury may be so extensive that there is an insufficient epithelial surface area to effectively remove alveolar edema fluid. Pathology studies have shown that denuding of alveolar epithelium is characteristic of severe acute respiratory distress syndrome (3) in patients who died with severe lung injury. In addition, alveolar epithelial transport may be diminished because of several other mechanisms, including alveolar hypoxia, the presence of injurious oxygen and nitrogen products, and the presence of specific cytokines that may downregulate alveolar fluid clearance, such as transforming growth factor-β (15, 40). On the other hand, there is some evidence that proinflammatory molecules may upregulate fluid clearance, including TNF-β and leukotriene D (98). Potentially injurious effects of unfavorable ventilatory strategies further complicate the potentially complex interplay of soluble factors that may upregulate or downregulate fluid clearance in the setting of lung injury. It is well known that excessive tidal volume and airway pressures contribute to the severity of lung injury clinically, and experimental studies have demonstrated that alveolar epithelial fluid clearance is impaired significantly in the presence of unfavorable ventilatory strategies. By contrast, a reduced tidal volume and airway pressure result in better preservation of alveolar epithelial fluid transport (16).
Thus a considerable body of new knowledge has accumulated in the last 25 years regarding the basic mechanisms that regulate alveolar epithelial fluid clearance under both normal and pathological conditions. From a therapeutic perspective, it is plausible that aerosolized delivery of β-adrenergic agonists may be effective in upregulating alveolar fluid clearance in the presence severe pulmonary edema, either from a hydrostatic mechanism or from increased permeability edema. It is also possible that in some patients the extent of alveolar epithelial injury and the severity of endothelial lung injury may preclude a therapeutic benefit of β-adrenergic agonists, at least in the early stages of the lung injury. However, as repair of the alveolar epithelium occurs, it is possible that the epithelial barrier may become responsive to cAMP stimulation, thus hastening the resolution of alveolar edema. Well-designed clinical studies will be needed to test the potential therapeutic value of β2-agonist therapy in patients with pulmonary edema.
In conclusion, the symposium advanced the notion that it has become necessary to address the control of lung fluid balance at the level of intact lung by exploiting approaches in molecular imaging, genetics (to modify or ablate specific proteins), and rigorous quantification of endothelial and epithelial barrier functions in animal models, in particular genetically modified mice. These advances will be extremely useful from both scientific and a clinical perspective.
- Copyright © 2004 the American Physiological Society