I. INTRODUCTION

Top Previous Next References

In the last 20 years, the mechanisms that regulate active salt and water transport by the alveolar and distal airway epithelium of the lung have emerged as a new area of research with important implications for understanding lung fluid balance under both normal and pathological conditions. Studies of epithelial fluid transport by the distal pulmonary epithelium have provided important new concepts regarding the resolution of pulmonary edema, a common clinical problem. Before 1982, there was no information on how lung fluid balance was regulated across the distal airway and alveolar epithelial barriers. However, in 1982, new work provided evidence that fluid balance in the lung was regulated by active ion transport mechanisms (131, 214, 229). For many years, it was generally believed that differences in hydrostatic and protein osmotic pressures (Starling forces) accounted for the removal of excess fluid from the airspaces of the lung. This misconception persisted in part because some experiments that measured solute flux across the epithelial and endothelial barriers of the lung were done at room temperature (347), and the studies were done in dogs, a species that turned out to have a very low rate of active sodium and fluid transport (28). Also, until the early 1980s, there were no satisfactory animal models to study the resolution of alveolar edema, and the isolation and culture of alveolar epithelial type II cells was just becoming a useful experimental method. Although the removal of interstitial pulmonary edema by lung lymphatics and the lung microcirculation was discussed by Staub in 1974 in his review of pulmonary edema (333), there was no information on how pulmonary edema was removed from the distal airspaces of the lung.

This review provides a perspective on how in vivo and in vitro experiments in the last 20 years have provided a new understanding of the regulation of lung fluid balance by active transport mechanisms across both the alveolar and distal airway epithelium. On balance, there is convincing evidence that the vectorial transport of salt and water across the alveolar and distal airway epithelium is the primary determinant of fluid clearance, thus accounting for the ability of the lung to remove alveolar fluid at the time of birth as well as in the mature lung when pathological conditions lead to the development of pulmonary edema. Most of this review is focused on the function of the adult, not the fetal or the perinatal, lung.

	II. SALT AND WATER TRANSPORT ACROSS THE DISTAL PULMONARY EPITHELIA

Top Previous Next References

With few exceptions, the general model for transepithelial fluid movement is that active salt transport drives osmotic water transport. This paradigm is probably correct for fluid clearance from the distal airspaces of the lung (22, 30, 91, 92, 94, 228, 229). The results of several in vivo studies have demonstrated that changes in hydrostatic or protein osmotic pressures cannot account for the removal of excess fluid from the distal airspaces (22, 28, 30, 94, 215, 219, 226-229, 343). Furthermore, pharmacological inhibitors of sodium transport can reduce the rate of fluid clearance in the lungs of several different species, including the human lung (30, 110, 164, 217, 228, 258, 266, 308, 327). In addition, there is good evidence that isolated epithelial cells from the distal airspaces of the lung actively transport sodium and other ions. Less than 10 years ago, the pathways for water clearance across alveolar and distal airway epithelium were unknown. Since then, several studies have demonstrated the presence of several functional water channels, aquaporins, in the lung.

This section reviews relevant morphological features of the distal lung epithelium with particular reference to their potential contribution to solute and water transport capacity. Next, based on in vivo and in vitro studies, there will be a discussion of some of the evidence supporting the hypothesis that active ion transport is the primary mechanism for driving vectorial fluid transport from the airspaces to the interstitium of the lung. The last section describes the evidence for the role of epithelial sodium channel (ENaC), Na⁺-K⁺-ATPase, and aquaporins in mediating salt and water transport in distal lung epithelia.

The human lung consists of a series of highly branched hollow tubes that end blindly in the alveoli, with the conducting airways (the cartilaginous trachea, bronchi, and the membranous bronchioles) occupying the first 16 generations of airways (281). Gas exchange primarily occurs in the branches that make up the last seven generations including respiratory bronchioles, alveolar ducts, alveolar sacs, and alveoli (334). The airways, ~1.4 m² in the adult human lung (376), and alveoli, ~143 m² in the adult human lung (376), constitute the interface between lung parenchyma and the external environment and are lined by a continuous epithelium. The distal airway epithelium is composed of terminal respiratory and bronchiolar units with polarized epithelial cells that have the capacity to transport sodium and chloride including ciliated Clara cells and nonciliated cuboidal cells. The alveoli themselves are composed of a thin alveolar epithelium (0.1-0.2 µm) that covers 99% of the airspace surface area in the lung and contains thin, squamous type I cells and cuboidal type II cells (334, 376) (Fig. 1). The alveolar type I cell covers 95% of the alveolar surface (376). The close apposition between the alveolar epithelium and the vascular endothelium facilitates efficient exchange of gases, but also forms a tight barrier to movement of liquid and proteins from the interstitial and vascular spaces, thus assisting in maintaining relatively dry alveoli (377).

View larger version (25K): [in this window] [in a new window]   Fig. 1. A schematic diagram of the distal pulmonary epithelium that is relevant for salt and water transport.

The tight junctions are the critical structures for the barrier function of the alveolar epithelium. Tight junctions connect adjacent epithelial cells near their apical surfaces, thereby maintaining apical and basolateral cell polarity (318). Ion transporters and other membrane proteins are asymmetrically distributed on opposing cell surfaces, conferring vectorial transport properties to the epithelium. In addition, the tight junction itself might contain discrete ion-selective pores (139). In the past, it was thought that tight junctions were rigid structures, which physically restrict passage of larger molecules; however, the permeability of tight junctions is dynamic and regulated, in part, by cytoskeletal proteins and intracellular calcium concentration (318). Physiological studies of the barrier properties of tight junctions in the alveolar epithelium indicate that diffusion of water-soluble solutes between alveolar epithelial cells is much slower than through the intercellular junctions of the adjacent lung capillaries (145, 279, 317, 320, 333). Based on tracer fluxes of small water-soluble solutes across the air-blood barrier of the distal airspaces, the effective pore radii were 0.5-0.9 nm in the distal respiratory epithelium and 6.5-7.5 nm in the capillary endothelium (346). In addition, studies of protein flux across both the endothelial and epithelial barriers of the sheep lung indicate that 92% of the resistance to albumin flux across the alveolar barrier resides in the epithelium (134). Removal of large quantities of soluble protein from the airspaces appears to occur primarily by restricted diffusion (119, 150, 151), although there is evidence for some endocytosis and transcytosis of albumin across alveolar epithelium (27, 114, 119, 150, 167, 177, 178, 221).

The most extensively studied cell in the distal pulmonary epithelium is the alveolar type II cell, partly because type II cells can be readily isolated from the lung and studied in vitro. The alveolar type II cell is responsible for the secretion of surfactant (65, 66) as well as vectorial transport of sodium from the apical to the basolateral surface (67, 132, 214, 216, 217, 228, 238). The active transport of sodium by type II cells appears to provide a major driving force for removal of fluid from the alveolar space. Sodium uptake occurs on the apical surface, partly through amiloride-sensitive and amiloride-insensitive channels. Subsequently, sodium is pumped actively from the basolateral surface into the lung interstitium by Na⁺-K⁺-ATPase (Fig. 1). An ENaC that participates in sodium movement across the cell apical membrane was cloned and characterized in 1994 (52), and work by other investigators has provided new insight into the molecular and biochemical basis for sodium uptake in alveolar epithelial cells (264, 370, 387) (see details below).

The role of the alveolar type I cell in vectorial fluid transport in the lung is unknown, although several investigators are currently trying to assess the potential contribution of the alveolar type I cell to vectorial fluid transport. On the basis of studies in freshly isolated type I cells, it is known that these cells have a high osmotic permeability to water with expression of aquaporin-5 on the apical surface (84). Immunocytochemical studies in the intact lung by some investigators failed to demonstrate the presence of Na⁺-K⁺-ATPase in these cells in vivo (319). However, more recent studies have reported the presence of the ₁- and ₂-subunits of Na⁺-K⁺-ATPase in both type I like cells in vitro (303). The presence of the Na⁺-K⁺-ATPase could be consistent with a role for this cell in vectorial fluid transport, although it is not conclusive since the Na⁺-K⁺-ATPase may be needed to maintain cell volume. Also, recent studies of freshly isolated alveolar type I cells from rats demonstrated that these cells express the Na⁺-K⁺-ATPase ₁- and ₁-subunit isoforms, but not the ₂-subunit (38). In the same study, there was evidence for ₁ Na⁺-K⁺-ATPase expression on the basolateral surface of the alveolar epithelial type I cells in situ in the rat lung. In addition, there is evidence for expression of all the subunits of ENaC in freshly isolated alveolar type I cells from two laboratories (38, 168) as well as in situ in the rat lung (168). Finally, there is some evidence that ²²Na uptake can be partially inhibited by amiloride in the freshly isolated rat alveolar type I cells (168), although definitive studies of cultured, polarized type I cells have not yet been achieved. There is also recent preliminary evidence that -receptors are expressed on type I cells (201), although this issue is not clear since evidence also has been brought forward for the lack of such expression in type I cells (249). Also, type I cells might be involved in the transport of macromolecules because of the presence of vesicles and caveolin in these cells (50, 171, 247). Although the inability to study alveolar type I cells in culture has hindered progress in assessing their capacity for ion transport and the role they may play in vectorial fluid transport across the alveolar epithelium, the new evidence provides suggestive though not conclusive evidence that alveolar type I cells may participate in vectorial salt transport in the lung.

The alveolar epithelium comprises 99% of the surface area of the lung, a finding that suggests that removal of edema fluid from the lung might primarily occur across the alveolar epithelium. However, this conclusion may be misleading. It has been demonstrated that the distal airway epithelium actively transports sodium, a process that depends on amiloride-inhibitable uptake of sodium on the apical surface and extrusion of sodium through a basolateral Na⁺-K⁺-ATPase (5, 13, 14, 41, 42, 159, 382). It appears that in airways <200 µm in diameter, sodium absorption predominates (5, 13, 14) (Fig. 1). Secretion of chloride may occur in the trachea and larger airways, and perhaps in distal airways of some animal species in the presence of cAMP stimulation (254). In support of the potential role of distal airway cells is evidence that Clara cells actively absorb sodium and transport from an apical to basal direction (358, 359). Also, there are new data on the possible role of cystic fibrosis transmembrane conductance regulator (CFTR) in upregulating cAMP fluid clearance (see sect. III). This information provides support for a possible role of distal airway epithelia in fluid clearance, since CFTR is expressed abundantly in distal airway epithelial cells. Also, there is new evidence that some alveolar fluid may be removed by convective surface active forces that might propel alveolar edema fluid into the distal airways where some absorption may occur across the respiratory bronchiolar epithelium (263, 372). Thus, even though their surface area is limited, a contribution from distal airway epithelia to the overall fluid transport is probable, especially since cells from the distal airway epithelium primarily transport salt from the apical to the basolateral surface (5, 13, 14, 41, 42, 159).

A substantial number of innovative experimental methods have been used to study fluid and protein transport from the distal airspaces of the intact lung including isolated perfused lung preparations, in situ lung preparations, surface fluorescence methods, and intact lung preparations in living animals for short time periods (30-240 min) or for extended time periods (24-144 h). The advantages and disadvantages of these preparations have been reviewed in some detail (228, 313).

The first in vivo evidence that active ion transport could account for the removal of alveolar edema fluid across the distal pulmonary epithelium was obtained in studies of anesthetized, ventilated sheep (229). In those studies, the critical discovery was that isosmolar fluid clearance of salt and water occurred in the face of a rising concentration of protein in the airspaces of the lung, whether the instilled solution was autologous serum or an isosmolar protein solution. The initial protein concentration of the instilled protein solution was the same as the circulating plasma. After 4 h, the concentration of the protein had risen from ~6.5 to 8.4 g/100 ml, while the plasma protein concentration was unchanged. In longer term studies in unanesthetized, spontaneously breathing sheep, alveolar protein concentrations increased to very high levels. After 12 and 24 h, the alveolar protein concentration increased to 10.2 and 12.9 g/100 ml, respectively (226). The overall rise in protein concentration was equivalent to an increase in distal airspace protein osmotic pressure from 25 to 65 cmH₂0 (Fig. 2). The concentration of protein in the lung lymph draining the lung interstitium declined, providing further evidence that protein-free fluid was being reabsorbed from the distal airspaces into the lung interstitium (30, 229). Also, morphological studies showed that the interstitial fluid did not contain the Evans blue dye-labeled alveolar protein (224). These data provided convincing evidence that active ion transport must be responsible for the fluid clearance, especially in the face of a rising alveolar protein osmotic pressure that was ~40 cmH₂0 greater than the protein osmotic pressure in the vascular or interstitial spaces of the lung (226, 229).

View larger version (12K): [in this window] [in a new window]   Fig. 2. In spontaneously breathing unanesthetized sheep, alveolar fluid protein concentration increased from 6.0 g/dl to a maximum of 13.8 g/dl (x-axis) after 24 h. This increase in protein concentration was paralleled by a marked increase in protein oncotic pressure (y-axis). [Data adapted from Matthay et al. (226).]

Other studies in the intact lung have supported the hypothesis that removal of alveolar fluid requires active transport processes. For example, elimination of ventilation to one lung did not change the rate of fluid clearance in sheep, thus ruling out changes in transpulmonary airway pressure as a major determinant of fluid clearance, at least in the uninjured lung (309). Furthermore, if active ion transport were responsible for fluid clearance, then fluid clearance should be temperature dependent. In an in situ perfused goat lung preparation, the rate of fluid clearance progressively declined as temperature was lowered from 37 to 18°C (323). Similar results were obtained in perfused rat lungs (306) and ex vivo human lung studies (308) in which hypothermia inhibited sodium and fluid transport.

Additional evidence for active ion transport was obtained in intact animals with the use of amiloride, an inhibitor of sodium uptake by the apical membrane of alveolar epithelium and distal airway epithelium. Amiloride inhibited 40-70% of basal fluid clearance in sheep, rabbits, rats, guinea pigs, mice, and in the human lung (21, 22, 30, 74, 94, 121, 164, 258, 266, 308, 327, 386). Amiloride also inhibited sodium uptake in distal airway epithelium from sheep and pigs (5, 13). To further explore the role of active sodium transport, experiments were designed to inhibit Na⁺-K⁺-ATPase. It has been difficult to study the effect of ouabain in intact animals because of cardiac toxicity. However, in the isolated rat lung, ouabain inhibited >90% of fluid clearance (21, 270). Subsequently, after the development of an in situ sheep preparation for measuring fluid clearance in the absence of blood flow, it was reported that ouabain inhibited 90% of fluid clearance over a 4-h period (309).

Important species differences in the basal and stimulated rates of fluid clearance have been identified; not all species respond to cAMP agonists (Table 1). To normalize for differences in lung size or the available surface area, different instilled volumes were used ranging from 1.5 to 13.0 ml/kg. The slowest fluid clearance was measured in dogs (28, 135), intermediate rates of fluid clearance in sheep and goats (30, 226, 229, 323), and the highest basal fluid clearance rates in rabbits, guinea pigs, rats, and mice (121, 124, 164, 258, 327). The basal rate of fluid clearance in the human lung has been difficult to estimate, but based on the isolated, nonperfused human lung model, basal fluid clearance rates appear to be intermediate to fast (308). In fact, the fluid clearance in the ex vivo human lung is approximately one-half of the rate in the ex vivo rat lung (307). However, as discussed in section V, studies in patients suggest that the rate of maximal fluid clearance from the human lung may be fast. The explanation for the species differences is not apparent, although it may be related to the number or activity of sodium or chloride channels or the density of Na⁺-K⁺-ATPases in alveolar epithelium in different species. Morphometric studies (76, 282) have shown no significant difference in the number of alveolar type II cells in different species. It is also possible that the contribution of the distal airway epithelium may not be uniform in all species.

The success in obtaining nearly pure cultures of alveolar epithelial type II cells from rats made it possible to study the transport properties of these cells and relate the results to the findings in the intact lung studies. The initial studies showed that when type II cells were cultured on a nonporous surface such as plastic, they would form a continuous confluent layer of polarized cells after 2-3 days (131, 214). Interestingly, after 3-5 days, small domes of fluid could be appreciated from where the substratum was detached (Fig. 3). The domes were thought to result from active ion transport from the apical to the basal surface with water following passively since they were inhibited by the replacement of sodium by another cation or by pharmacological inhibitors of sodium transport, such as amiloride and ouabain (132). More detailed information on the nature of ion transport across alveolar type II cells was obtained by culturing these cells on porous supports and mounting them in Ussing chambers and measuring short-circuit current and ion flux under voltage-clamp conditions. Confluent monolayers with tight junctions were formed by 4-5 days later, and with considerable care to the culture conditions, it has been possible in some, but not all studies, to obtain monolayers with a high transepithelial resistance (2,000 /cm²) (60). Alveolar type II cell monolayers are capable of generating spontaneous transmembrane potential difference (PD) related to electrogenic transport, and it is interesting to note that the spontaneous PD values (0.2-10 mV with apical membrane negative) (60, 73, 337) are similar to those obtained in vivo between the alveolar lumen and the pleural surface (255). When the apical and basolateral compartments are bathed with solutions with identical ionic compositions and the transepithelial potential is maintained (clamped) to zero, the measured current is termed short-circuit current (I_sc). The I_sc in alveolar type II cells monolayers corresponded closely to the net sodium absorption; substitution of sodium by equimolar concentration of choline in the apical side of monolayers or addition of ouabain that have the ability to block Na⁺-K⁺-ATPase at the basolateral side, quickly decreased I_sc to zero (60). With the use of labeled sodium (²²Na), unidirectional flux from the apical to the basolateral surface exceeded the flux in the opposite direction (176), a finding that was similar to results in isolated perfused lung preparations (21, 22, 74, 93, 94, 133).

View larger version (39K): [in this window] [in a new window]   Fig. 3. A: light micrograph of domes from primary culture of alveolar epithelial type II cells. The domes formed spontaneously after 3 days in culture. Dome formation was inhibited by amiloride. B: schematic representation of a dome of fluid accumulation between cultured alveolar epithelial type II cell monolayers. The type II cells actively transport sodium from the culture medium to beneath the monolayer (dashed arrows), which results in fluid accumulation between the cell monolayer and the culture dish. The apical cell surface (microvilli) is shown in contact with the medium; the basolateral surface is shown facing the culture dish. [From Mason et al. (214), with permission from the National Academy of Sciences USA.]

The coordinated role of apical and basolateral sodium transport has been studied in several in vitro studies. Sodium ions that enter the epithelial cells at the apical membrane are pumped out of the cells at the basolateral membrane by the Na⁺-K⁺-ATPase enzyme. Because of the continuous pumping, sodium chemical potential is lower inside the cell. The entry step is passive and then sodium flows down a chemical potential gradient through specialized pathways, where basolateral transport requires energy to move ions against the gradient. Because of the pump activity, potassium electrochemical potential is larger inside the cell, and potassium leaks through the basolateral membrane and is then recycled by the Na⁺-K⁺-ATPase. The pathways for sodium entry into alveolar type II cells are numerous. Amiloride blocked dome formation (131) and decreased I_sc in the in vitro studies (60), a finding that supported the critical importance of sodium uptake through an amiloride-sensitive pathway in the apical membrane of alveolar epithelial cells. As already discussed, the efficacy of amiloride as an inhibitor of fluid clearance in the intact lung was demonstrated in several in vivo studies, although the fraction of amiloride-sensitive transport was as low as 40-50% in some lung preparations, particularly the rat and the human lung. The amiloride-insensitive sodium influx may be represented in vivo in part by the Na⁺-glucose cotransport (21, 22). In vitro, the amiloride-insensitive sodium pathways are not clearly defined, but several sodium cotransporters such as Na-amino acid, Na-phosphate, Na-proton, as well as other cation channels may be involved (46, 68, 69, 256).

One important unresolved issue is to what extent cultured alveolar type II cells retain their in vivo phenotype. Within the first 24 h of culture, some of the morphological and functional characteristics of the alveolar type II cells are lost (81, 86). Most surfactant protein expression is downregulated (202), and some transport proteins such as Na-phosphate and Na⁺-K⁺-2Cl cotransport vanish (68, 69). However, active ion transport can be measured for up to 7 days. Also, some data indicate that culture conditions may help the cells to retain features of type II cells. For example, alveolar type II cells cultured on collagen gels with culture medium containing fetal calf serum and with an apical surface exposed to air retain morphological characteristics of alveolar type II cells for a longer duration (85). These cells express surfactant proteins and their mRNAs and express a plasma membrane marker specific for alveolar type II cells. Recent work (163) indicates that in vitro culture conditions are an important determinant for the biophysical characteristics of sodium channels. Alveolar type II cells grown on a permeable surface with an air interface and glucocorticoids in the medium maintain alveolar type II cell phenotype and express highly selective sodium channel while cells cultured on glass and in air-liquid interface express a low selective sodium channel.

A number of studies using patch-clamp techniques, measurements of ²²Na uptake, and binding of amiloride analogs have demonstrated the presence of amiloride-sensitive sodium channels in alveolar type II cells. On the basis of the biophysical and pharmacological characterization, at least three types of amiloride-sensitive sodium channels have been described in the apical membrane of cultured adult alveolar type II cells (220). The most frequently observed in both inside-out and cell-attached configurations is a nonselective cation (NSC) channel with a conductance of ~21 pS, which was equally permeable to sodium and potassium (P_Na/P_K ~1), voltage-independent calcium-activated and completely blocked by 1 µM amiloride (105). A similar channel was described in fetal distal lung epithelial cells (213, 274). Also, others have recorded a channel with a conductance of 27 pS with a high relative permeability of sodium to potassium (7:1), completely blocked by 1 µM amiloride or its analog N-ethyl-N-isopropylamiloride (EIPA) (388). Finally, the presence of highly selective cation (HSC) channels has been identified in adult alveolar type II cells cultured on permeable support, in air-liquid interface with aldosterone. These channels recorded on cell-attached configuration have a low conductance (5-7 pS) and were inhibited by a low concentration of amiloride (<0.1 µM) (163). These properties are similar to those described in fetal alveolar epithelial cells (370). Treatment of A549 cells with dexamethasone also shifted the conductance of the channel from 8 to 4 pS (194). A variety of other channels have been recorded in alveolar type II cells including ones regulated by G proteins (220).

Molecular identification of the proteins involved in amiloride-sensitive sodium influx has been achieved in the last few years. Three homologous subunits, entitled -, -, and -ENaC, correspond to the pore-forming subunits. The three subunits share in common a structure predicting two hydrophobic membrane-spanning regions, intracellular amino and carboxy termini, two transmembrane spanning domains, and a large extracellular loop with highly conserved cysteine residues (51, 298, 330). Not all investigators agree on the numbers of subunits that form functional channels; some propose that ENaC is a tetramer made of two -, one -, and one -subunit that assemble pseudosymmetrically around the channel pore to form a high amiloride affinity sodium-selective pore (111, 184), whereas others contend that ENaC is a much larger complex of nine (3 -, 3 -, 3 -) subunits (97, 329). An important role has been established for ENaC in the distal renal tubule and the collecting ducts and in the colon (15, 96, 223, 299). Some of the data suggest that individual ENaC subunits may have independent functions. For example, there are differences in the onset of expression of the three genes in the fetal lung (78, 348), and the relative levels of mRNA for the subunits can vary significantly in different regions of the lung, kidney, and colon (15, 87, 96, 104, 126, 223, 232). Xenopus oocyte studies predict a greater dependence of amiloride-sensitive sodium transport on -ENaC than on -ENaC or -ENaC (52, 237). Low levels of sodium transport (1% of maximal) could be detected in oocytes injected with -ENaC alone but not in oocytes that expressed the - or -subunits. Also, coinjection of - or -subunits with -ENaC increased sodium transport approximately fivefold, and maximal current required expression of all the three subunits. When coexpressed, the three subunits reproduce the expected properties of the HSC channel. In A549 lung epithelial cells, dexamethasone increases the expression of - and -ENaC and markedly shifts the IC₅₀ for amiloride (194).

The critical role of ENaC in the absorption of salt and fluid by lung epithelia has been confirmed by the generation of knockout mice with inactivated subunits of ENaC. After inactivation of murine -ENaC, deficient neonates develop respiratory distress syndrome and die within 40 h of birth, primarily from failure to clear their lungs of fluid (154). In contrast, - or -ENaC knockout mice were able to clear fluid from lungs at birth, although at a slower rate that in wild-type control, but these mice died from abnormal electrolyte reabsorption in the kidney resulting in fatal hyperkalemia (16, 233). These results suggest that small levels of activity provided by the remaining intact subunits (- or -channels) are enough to maintain respiratory function but are not sufficient to absorb sodium in the kidney. More definitive answers regarding the role of ENaC should be forthcoming from inducible knockout ENaC mice, i.e., CRE-lox mice in which tetracycline can transiently suppress ENaC.

In humans, loss-of-function mutations in the -, -, and -ENaC subunits have been described in systemic pseudohypoaldosteronism (173, 316). Patients had no respiratory symptoms at birth but developed, within weeks or months after birth, respiratory illnesses characterized by chest congestion and cough caused by an excessive volume of liquid that resulted from sodium channel dysfunction in the airways.

In situ hybridization studies, as well as Northern blot analyses done in mice, rats, and humans, have identified the presence of mRNA for all three subunits of ENaC in alveolar epithelial type II cells both in vivo and in vitro with usually a greater expression of - than - and -subunits (64, 75, 78, 162, 220, 223, 232, 238, 264, 315, 345, 348). The relative amounts of ENaC mRNA could explain the difference in ENaC activity and amiloride sensitivity between tissues and species, although a direct relationship between mRNA and functional protein cannot be predicted. As previously mentioned, patch-clamp studies using freshly isolated or cultured rat alveolar type II cells have reported the presence of a nonselective cation (NSC) channel and have often failed to identify the classical HSC channel that would be expected from ENaC expression. Recent studies have provided evidence that HSC channels are expressed in the apical membrane of adult alveolar type II cells under specific culture conditions (163, 194). All studies showing the presence of a 4-pS channel were done after exposure of alveolar type II cells or A549 cells to aldosterone or dexamethasone. In these cells, the number of HSC channels decreased after treatment with antisense oligonucleotides to any of the three subunits of ENaC: -antisense reduced all channel types while treatment with - or -antisense increased NSC channels (163). Therefore, in alveolar type II cells, ENaC may be expressed in different subunit combinations that determine sodium channel characteristics and render them less sensitive to amiloride (35, 268). This observation raises the possibility that ENaC-like channels could be present in vivo. As described in section IIA, new evidence has localized the three subunits of ENaC protein to alveolar type I cells also (38, 168).

In intact rat, sheep, mouse, or human lung, a high concentration of amiloride (~1 mM) was required to inhibit 50-60% of fluid clearance. The discrepancy between in vivo and in vitro studies in which 1 µM is required to completely block epithelial sodium channels (105, 388) is not fully understood but could be due in part to the binding of amiloride to protein, a rapid degradation of amiloride after instillation (386), and/or escape from the airspaces (146, 266). Nevertheless, in vivo studies suggest that amiloride has poor affinity for sodium channels that are present in alveolar epithelium. In one rat study, EIPA, a pharmacological inhibitor primarily of low-affinity amiloride channels, inhibited an equal fraction of fluid clearance as amiloride (386). These results suggest that active sodium transport in vivo involves in part low-affinity amiloride-sensitive channels, i.e., NSC channels.

Several investigators have investigated how ENaC is regulated by transcriptional, translational, and posttranslational mechanisms (275, 276, 304). It is clear that understanding the regulation of ENaC processing, trafficking to, and stability at the cell surface is of fundamental importance (304).

Studies in several species including sheep, rabbit, guinea pig, rat, and human lungs demonstrated that there was a substantial fraction of fluid clearance that could not be inhibited by amiloride (30, 164, 258, 266, 308, 327, 386). In the rat and human lung, for example, 40-50% of the basal fluid clearance is amiloride insensitive while in mice ~20% of fluid clearance is insensitive to amiloride, although the latter result has been mainly identified in CD-1 mice (121). In CD57BL/6 mice, ~40% of the fluid clearance is amiloride insensitive (146).

There is some limited evidence that the amiloride-insensitive fraction of 8-Br-cGMP stimulated I_sc and sodium uptake in rat tracheal epithelia was inhibited by dichlorobenzamil or L-cis-diltiazem, both inhibitors of cyclic nucleotide-gated cation channels (CNG) (322). In subsequent studies, dichlorobenzamil inhibited a significant fraction of lung fluid absorption in sheep (170), a finding that supported the hypothesis that CNG channels could play a role in fluid absorption from the distal airspaces. The CNG channels were originally identified and cloned from vertebrate rod photoreceptors (109) and the olfactory neuroepithelium (246). Three different isoforms have been cloned (33). The CNG1 channel has a wide tissue distribution including expression in the lung; mRNA for CNG1 has been localized to distal lung epithelium and alveolar epithelial cells (82). In a recent rat study, L-cis-dilatizem inhibited a fraction of terbutaline or cGMP-stimulated fluid clearance that was not inhibited by amiloride, suggesting that part of the amiloride-insensitive fraction fluid clearance in the distal lung could be mediated through CNG channels, activated by cGMP (82, 261). Also, there is recent data that inducible nitric oxide synthase (iNOS) (/) mice have normal rates of basal fluid clearance, but none of the basal fluid clearance is inhibited by amiloride (146).

In the intact rat, the amiloride-insensitive part of fluid clearance may be mediated in part through the Na⁺-glucose cotransporter (23). The amiloride-insensitive part of fluid clearance is abolished when sodium is replaced by choline in the instillate or when phlorizin, a specific inhibitor of the Na⁺-glucose cotransport, is added in the instillate. However, in one study, only sodium channel inhibition slowed lung water clearance in neonatal animals (265). On balance, the importance of Na⁺-glucose cotransport for fluid clearance from the airspaces remains unproven. Surprisingly, although SGLT1 mRNA is detected in mouse lungs, suggesting the presence of Na⁺-glucose transporter, luminal glucose has no influence on fluid clearance (155). In vitro there is no evidence that the Na⁺-glucose cotransporter plays a major role in transepithelial sodium transport (174). Also, the role of other cotransporters such as Na-amino acid and Na-phosphate that are specifically expressed in the apical membranes of human, mouse, and rat type II alveolar epithelial cells, remains to be defined (46, 69, 354). Some recent work indicates that SGLT1 can function as a cotransporter of water (89). Finally, regulation of vectorial fluid transport could be indirectly influenced by Na⁺/H⁺ exchange, especially in the setting of chronic metabolic acidosis (169).

The Na⁺-K⁺-ATPase enzyme is a ubiquitous plasma membrane ion-transporting ATPase that maintains electrochemical sodium and potassium gradients across the plasma membrane by pumping sodium out of the cell and potassium into the cell against their respective concentration gradients, fueled by hydrolysis of ATP. The electrochemical gradient generated by extrusion of sodium from and uptake of potassium into the cell is essential for the secondary active movement of sodium ions. In most of the epithelia, including the alveolar epithelium, the Na⁺-K⁺-ATPase is confined to the basolateral domain of the cells; this polarized distribution is essential for the vectorial transport of sodium followed isosmotically by that of water while also controlling in its housekeeping mode cell volume and composition. In the alveolar epithelium, the Na⁺-K⁺-ATPase protein is detected primarily in type II cells with a weak expression in type I cells consistent perhaps with a predominant role of type II cells in resorbing alveolar fluid. The Na⁺-K⁺-ATPase is a heterodimeric transmembrane protein composed of one - and one -subunit in a 1:1 ratio, although there are multiple isoforms of both subunits usually expressed in a tissue-specific pattern. The ₁-isoform catalyzes the movement of sodium and potassium, is phosphorylated by ATP, and binds the specific inhibitor ouabain (326). The ₁-isoform is the predominant form in alveolar type II cells, but cultured type I cells may express ₂ (303). The -subunit has three isoforms and targets the assembled molecule to the plasma membrane. It is believed that a heterodimeric form composed of the ₁- and ₁-subunits is the predominant Na⁺-K⁺-ATPase isoform in alveolar epithelial cells. It has been convincingly demonstrated that both - and -isoforms are required for functional insertion of Na⁺-K⁺-ATPase in the membrane (235). In the lung, adenoviral gene transfer of the ₁-, but not ₁-, subunit in the alveolar epithelium increases Na⁺-K⁺- ATPase expression and fluid clearance in the adult rat (102) and in confluent monolayers of fetal alveolar type II cells (351).

Removing edema fluid from the airspaces is accomplished by the polar distribution of Na⁺-K⁺-ATPases and channels on opposite poles of epithelial cells (129). Na⁺-K⁺-ATPase activity in the plasma membrane is regulated acutely by covalent or allosteric modification and/or acute trafficking of Na⁺-K⁺-ATPases between the plasma membrane and intracellular endosomal pools, and chronically by regulation of the abundance of Na⁺-K⁺-ATPases in the membrane by changes in Na⁺-K⁺-ATPase synthesis and/or degradation rates (31).

In 1993, the first transcellular water channel, aquaporin-1 (AQP1), was cloned (3). Since that time, the list of known aquaporins has grown rapidly to approximately 10 mammalian water channels (182, 366). Four aquaporins have been localized to the lung with nonoverlapping distribution (365). In rat and mouse lung, AQP1 is present in both apical and basolateral membrane of microvascular endothelial cells and in fibroblasts, while AQP3, -4, and -5 polarize to the apical or basolateral membrane at different locations in the respiratory epithelium (179). Recent work in the human lung demonstrates that there are many similarities between the distribution of aquaporins in the rodent and human lung, but there are some additional areas of expression in the human respiratory tract (185). In addition to its expression on the apical surface of alveolar type I cells, AQP5 is located in the apical membrane of airway epithelium as well as in the superficial epithelium of the nasopharynx (185). AQP3 was expressed in the apical membrane of columnar cells as well as in basal cells. There seems to be some AQP3 in type II cells and AQP4 in alveolar epithelial cells (185).

The identification of aquaporins in the lung naturally suggested the hypothesis that aquaporins might contribute to the developmental regulation of lung fluid balance in the formation or reabsorption of fetal lung fluid, the pathogenesis or resolution of pulmonary edema in the mature lung, or in the humidification of the airways. These possibilities have been studied in several lung preparations over the last few years.

The initial studies measured the water permeabilities of the endothelial and epithelial barriers in the lung in conjunction with studies that determined the cellular expression of aquaporins in the mature lung. Water permeability across a simple barrier separating two compartments is characterized by an osmotic water permeability coefficient (P_f) defined as the volume flux (J_v) induced by an osmotic gradient: P_f + J_v/(SV_wC), where S is the barrier surface area, V_w is the partial molar volume of water, and C is the osmotic gradient. High P_f (generally >0.0005-0.01 cm/s) suggests a facilitated water pathway involving molecular water channels (53). More detailed description of the biophysics of measuring water permeability is available in several sources (2, 3, 54, 362, 364).

In the lung, osmotic water permeability was found to be high in the intact sheep lung across the alveolar-capillary barrier, 0.02 cm/s (113). Additional studies were done in mice with a novel fluorescence method that provided similar data, namely, a high water permeability across the alveolar-capillary barriers, 0.017 cm/s. Aquaporin localization studies had already demonstrated that AQP1 was distributed primarily throughout the lung capillary endothelium, whereas AQP5 was localized primarily to the apical surface of alveolar epithelial type I cells (113, 181, 252, 321). Interestingly, measurements of osmotic water permeability of isolated alveolar epithelial I cells indicated that these cells have a high water permeability, 0.07 cm/s (84). A relatively high water permeability was also measured across isolated microperfused distal airways from guinea pigs, 0.005 cm/s (112). All of these studies provided indirect evidence for a possible functional role of aquaporins in lung fluid balance in the lung as potential rate-limiting channels for isosmolar fluid clearance from the distal airspaces of the lung. However, because there are no reliable, nontoxic inhibitors of water transport, it was necessary to generate specific aquaporin knockout mice to test the importance of specific aquaporins in isosmolar vectorial fluid transport under physiological conditions across the distal pulmonary epithelium.

AQP1 knockout mice had a modest decrease in lung interstitial fluid accumulation in the presence of a hydrostatic stress, but there was no effect on isosmolar fluid clearance across the distal pulmonary epithelium (9). The potential contribution of AQP5 to isosmolar fluid clearance in knockout mice was tested under both basal and stimulated conditions. Even when isosmolar fluid transport in the mouse was increased to a maximum level of 30% in 15 min, the absence of this apical type I cell membrane water channel did not slow the rate of fluid clearance (9, 331, 364, 365, 367). It is unlikely that the discovery of other water channels would alter the results of these studies because deletion of AQP1 or AQP5 each produced a 10-fold decrease in osmotically driven water transport between the airspace and capillary compartments.

Although aquaporin expression is strongly upregulated near the time of birth (180, 355, 385), AQP1, AQP4, and AQP5 knockout mice have the same ability to clear lung fluid in the perinatal period as wild-type controls (331). One study suggested that aquaporins might play some role in lung fluid balance after pneumonia because there was decreased expression of AQP4 and AQP5 in the mouse lung after acute viral pneumonia (353). However, the decreased expression may simply have reflected loss of functional epithelial cells from the pneumonia. Furthermore, knockout of AQP1, -4, and -5 had no effect on the formation or resolution of experimental pulmonary edema from hyperoxia, thiourea, or acid instillation (331). Finally, recent data indicate that the elimination of AQP3 does not delay the normal humidification process in the airways of mice (332).

The insensitivity of fluid clearance to aquaporin deletion is probably the consequence of the substantially lower rate of active fluid absorption per unit surface area in the lung compared, for example, with the proximal tubule of the kidney or the salivary gland, organs in which aquaporin deletion does alter fluid absorption or secretion (362, 363). Slower rates of active fluid transport probably do not require very high cell membrane water permeabilities.

Although a major role of AQP5 in physiological transalveolar epithelial water movement appears unlikely, aquaporins may have effects on other cell functions, especially volume regulation, particularly in the highly specialized type I cell (152, 199), in alveolar fluid homeostasis (206), or in the regulation of fluid secretion from mucus secreting cells (207).

	III. REGULATION OF SALT AND WATER TRANSPORT

Top Previous Next References

This section considers how vectorial fluid transport across the distal pulmonary epithelium can be upregulated by either catecholamine-dependent or catecholamine-independent mechanisms. The potential relevance of these mechanisms under pathological conditions is evaluated in section V.

Until recently, most studies focused primarily on the active transport of sodium as the primary determinant for regulating catecholamine-dependent transport across distal pulmonary epithelium. New evidence indicates that cAMP-stimulated uptake of chloride may be an important mechanism for regulating fluid clearance across distal lung epithelium (103, 165, 166, 267). Therefore, this section is divided into the effects of catecholamines on fluid transport based primarily on studies that examined the effects of sodium transport inhibitors. Section IIIA2 considers the new evidence regarding a potential role for chloride transport in cAMP-mediated transport.

Studies in newborn animals indicate that endogenous release of catecholamines, particularly epinephrine, may stimulate reabsorption of fetal lung fluid from the airspaces of the lung (45, 110, 259, 260, 371). In most adult mammal species, stimulation of ₂-adrenergic receptors either by salmeterol, terbutaline, or epinephrine increases fluid clearance (30, 59, 74, 164, 210, 290, 307). This stimulatory effect occurs rapidly after intravenous administration of epinephrine or instillation of terbutaline in alveolar space and is completely prevented by either a nonspecific ₂-receptor antagonist, propranolol, or in rats by a specific ₂-antagonist (210, 290). In some species, such as the guinea pig, there is also evidence that stimulation of the ₁-receptor will augment fluid clearance (258).

The increased fluid clearance by ₂ agonists can be prevented by amiloride, indicating that the stimulation was related to an increased transepithelial sodium transport (164, 258, 290). In anesthetized ventilated sheep, terbutaline-induced stimulation of fluid clearance was associated with an increase in lung lymph flow, a finding that reflected removal of some of the alveolar fluid volume to the interstitium of the lung (30). Although terbutaline increased pulmonary blood flow, this effect was not important, since control studies with nitroprusside, an agent that increased pulmonary blood flow, did not increase fluid clearance. Other studies have demonstrated that -adrenergic agonists increased fluid clearance in rat (164), dog (28, 191), guinea pig (258), mouse (121, 124, 155), as well as human lung (307, 308). The presence of ₁- and ₂-receptors on alveolar type II cells has been demonstrated in vivo by autoradiographic and immunochemistry techniques (19, 98, 201). In vitro, studies in cultured alveolar type II cells indicate that -adrenergic agonists added at the apical or the basolateral sides increased I_sc that was abolished by amiloride (132, 217). In addition, terbutaline stimulated net apical-to-basolateral flux of ²²Na, and the magnitude of this stimulation was similar to the steady-state increase in I_sc (60). Experiments in isolated lungs (130, 313) and in vitro studies provided further evidence that cAMP is a second messenger for the -adrenergic effects, whereas activation of protein kinase C does not appear to be involved (26, 29). Interestingly, -adrenergic agonist therapy does not increase fluid clearance in rabbits and hamsters (129, 327). The explanation for this lack of effect is unclear, particularly since there are -receptors in rabbit type II cells that stimulate surfactant secretion (234). One study in rabbits suggested that cAMP agonists might stimulate chloride secretion in rabbits, thus perhaps offsetting any increase in vectorial fluid clearance from the distal airspaces (254).

On the basis of studies of the resolution of alveolar edema in humans, it has been difficult to quantify the effect of catecholamines on the rate of fluid clearance (230). However, studies of fluid clearance in the isolated human lung have demonstrated that -adrenergic agonist therapy increases fluid clearance, and the increased fluid clearance can be inhibited with propranolol or amiloride (Fig. 4) (307, 308). Subsequent studies showed that long-acting lipid-soluble -agonists are more potent than hydrophilic -agonists in the ex vivo human lung (307). The magnitude of the effect is similar to that observed in other species, with a -agonist-dependent doubling of fluid clearance over baseline levels. These data are particularly important because aerosolized -agonist treatment in some patients with pulmonary edema might accelerate the resolution of alveolar edema (see sect. V).

View larger version (12K): [in this window] [in a new window]   Fig. 4. Alveolar fluid clearance (y-axis) plotted under control conditions, -agonist stimulation, and after amiloride inhibition in the human lung (A), the rat lung (B), and the mouse lung (C). Amiloride sensitivity is different depending on mouse strain (146). [Data from Fukuda et al. (121), Jayr et al. (164), and Sakuma et al. (308).]

What has been learned about the basic mechanisms that mediate the catecholamine-dependent upregulation of sodium transport in the lung? On the basis of in vitro studies, it was proposed that an increase in intracellular cAMP resulted in increased sodium transport across alveolar type II cells by an independent upregulation of the apical sodium conductive pathways and the basolateral Na⁺-K⁺-ATPase. Proposed mechanisms for upregulation of sodium transport proteins by cAMP include augmented sodium channel open probability (75, 105, 192, 212, 216, 340), increases in Na⁺-K⁺-ATPase -subunit phosphorylation, as well as delivery of more ENaC channels to the apical membrane and more Na⁺-K⁺-ATPases to the basolateral cell membrane (34, 48, 61, 106, 165, 328). Patch-clamp studies done on alveolar type II cells in the attached mode indicated that -agonists increased the open-state probability and the mean open time of the moderately selective amiloride-sensitive sodium channel without affecting single-channel conductance. This effect was completely blocked by propranolol. The observed increase in open probability by protein kinase A (PKA) was perhaps mediated by the direct phosphorylation of channel proteins. Although one study indicated that rat ENaC can be phosphorylated by PKA (324), the functional relevance of phosphorylation is not yet known, and PKA does not activate amiloride-sensitive current in Xenopus oocytes injected with ENaC mRNAs (6), suggesting that other mechanisms are involved. Sodium channels, like other ion channels and transporters, are associated with cytoskeletal proteins such as actin, ankyrin, fodrin, or spectrin, which can serve as signaling molecules. Reconstitution of ENaC into lipid bilayers resulted in a channel that is activated by PKA and ATP in the presence but not the absence of short actin filaments (25). Thus the increase in open probability may have been the result of increased phosphorylation of short actin filaments or other cytoskeleton proteins. Finally terbutaline, a ₂-adrenergic agonist, may promote insertion of new channel protein from a cytoplasmic pool to the apical membrane. In fetal distal lung cells, brefeldin A, an inhibitor of protein trafficking, prevented terbutaline-induced increase in open probability of Ca²⁺-activated nonselective channels (160).

One investigator reported that stimulation of thyroid epithelial cells transfected with -, -, and -ENaC with cAMP translocated ENaC from the cytoplasm to the cell surface, and that effect was dependent on a sequence (PPPXY) in the COOH terminus of each subunit (328). Existing evidence also suggests that agents that increase cAMP may increase sodium influx through an increase in apical chloride conductance without affecting apical membrane sodium conductance (165, 267).

One recent study using patch clamping of rat alveolar type II cells demonstrated that treatment with a ₂-agonist, terbutaline, increased the number of highly selective sodium channels, an effect that was mediated by PKA (61). Also, in the same study, terbutaline increased both intracellular calcium levels and the open probability of NSC channels, an effect that was blocked by a calcium chelator (61). The authors concluded that -adrenergic stimulation increase intracellular cAMP and the activation of PKA, and the PKA promotes an increase in the number of highly selective sodium channels and an increase in intracellular calcium. The increase in intracellular calcium increases the open probability of NSC channels. Interestingly, other investigators recently found that both extracellular and intracellular calcium may act as second messengers for -adrenergic stimulation of fluid clearance from the distal airspaces of near-term fetal guinea pigs (260). Other investigators have also demonstrated that calcium channel activation and cell volume changes alter the rate of sodium transport via NSC channels (211-213).

-Adrenergic stimulation increases extrusion of sodium at the basolateral side by increasing Na⁺-K⁺-ATPase activity. In alveolar type II cells, isoproterenol increased the maximal velocity of the enzyme independently of sodium permeability (340). The stimulation of Na⁺-K⁺-ATPase activity was associated with an increase in ouabain binding and expression of the ₁-subunit at the membrane (32). This effect occurred rapidly (<15 min) and was independent of protein kinase phosphorylation and mediated by subunit recruitment from late endosomes into the plasma membrane. There is some evidence that long-term exposure of alveolar type II cells to -agonists increased Na⁺-K⁺-ATPase and sodium channel activity, at least in part through a transcriptional effect, since cAMP increased transiently the ₁-subunit of Na⁺-K⁺-ATPase (239) and the three subunits of ENaC (77). Increased activity of PKA could lead to phosphorylation of the cAMP-responsive element protein, which could bind to the cAMP responsive element (CRE) site on the gene promoter. A CRE site has been demonstrated on the promoter region of ₁-subunit of Na⁺-K⁺-ATPase (4) and in the 5'-flanking region of -ENaC (350). However, other investigators have failed to demonstrate change in mRNA levels of Na⁺-K⁺-ATPase after long exposure to isoproterenol, whereas they found a posttranscriptional regulation via mitogen-activated protein kinase/ERK and rapamycin-sensitive pathways (280).

While most experimental studies have attributed a primary role for active sodium transport in the vectorial transport of salt and water from the apical to the basal surface of the alveolar epithelium of the lung, the potential role of chloride, especially in mediating the cAMP-mediated upregulation of fluid clearance across distal lung epithelium, has been the subject of a few recent studies. One older study of cultured alveolar epithelial cells concluded that vectorial transport of chloride across alveolar epithelium occurs by a paracellular route under basal conditions and perhaps by a transcellular route in the presence of cAMP stimulation (176). Another study of cultured alveolar epithelial type II cells suggested that cAMP-mediated apical uptake of sodium might depend of an initial uptake of chloride (165). A more recent study of cultured alveolar type II cells under apical air interface conditions reported that -adrenergic agonists produced acute activation of apical chloride channels with enhanced sodium absorption (166). However, the results of these studies have been considered to be inconclusive by some investigators (193, 378, 379), partly because the data depend on cultured cells of an uncertain phenotype. Furthermore, studies of isolated alveolar epithelial type II cells do not address the possibility that vectorial fluid transport may be mediated by several different epithelial cells including alveolar epithelial type I cells as well as distal airway epithelial cells.

It should also be appreciated that the pathways for chloride secretion and absorption in the fetal and the neonatal lung are not well understood. Recent and older studies have demonstrated a role for chloride secretion and absorption in the fetal and perinatal lung (57, 117, 127, 273) as well as in distal fetal rat lung epithelial cells (70, 195), but the molecular basis for chloride transport is not well understood (262).

To define the role of chloride transport in the active transport of salt and water across the distal pulmonary epithelium of the lung, one group has used in vivo lung studies to define the mechanisms and pathways that regulate chloride transport during the absorption of fluid from the distal airspaces of the lung. This approach may be important because studies in several species, as already discussed, have indicated that distal airway epithelia are capable of ion transport (5, 13, 14), and both ENaC and CFTR are expressed in alveolar and distal airway epithelia (104, 123, 185, 284, 299).

Both inhibition and ion substitution studies demonstrated that chloride transport was necessary for basal fluid clearance. The potential role of CFTR under basal and cAMP-stimulated conditions was tested using intact lung studies in which CFTR was not functional because of failure in trafficking of CFTR to the cell membrane, the most common human mutation in cystic fibrosis (F508 mice). The results supported the hypothesis that CFTR was essential for cAMP-mediated upregulation of isosmolar fluid clearance from the distal airspaces of the lung because fluid clearance could not be increased in the F508 mice with either -agonists or with forskolin, unlike the wild-type control mice (103). Additional studies using pharmacological inhibition of CFTR in both the mouse and human lung with gilbenclamide supported the same conclusion, namely, that chloride uptake and CFTR-like transport seemed to be required for cAMP-stimulated fluid clearance from the distal airspaces of the lung (103). Glibenclamide can also inhibit potassium channels so the inhibitory effects are not specific for CFTR, but the F508 mouse studies have provided more direct evidence. Although the absence of CFTR in the upper airways results in enhanced sodium absorption (336), the data in these studies provide evidence that the absence of CFTR prevents cAMP-upregulated fluid clearance from the distal airspaces of the lung, a finding that is similar to work on the importance of CFTR in mediating cAMP-stimulated sodium absorption in human sweat ducts (297). Because CFTR is distributed throughout the distal pulmonary epithelium in distal airway epithelium as well at the alveolar level in the human lung (95), the data also suggest that the cAMP-mediated upregulated reabsorption of pulmonary edema fluid may occur across distal airway epithelium as well as at the level of the alveolar epithelium. Finally, additional studies indicated that the lack of CFTR results in a greater accumulation of pulmonary edema in the presence of a hydrostatic stress, thus demonstrating the potential physiological importance of CFTR in upregulating fluid transport from the distal airspaces of the lung (103).

The new data on a role for CFTR in cAMP-upregulated fluid clearance from the distal airspaces raise further interest in determining how CFTR and ENaC interact and contribute to net fluid clearance under isosmolar conditions. The relative conductances for chloride and sodium are difficult to measure in the in vivo lung epithelium but conceivably the debate about the role of either ion in limiting isosmolar fluid clearance is somewhat misleading. Under open-circuit conditions, the net transfer of chloride and sodium across the distal lung epithelium must be equal, i.e., there cannot be a significant net charge accumulation. It is possible that with cAMP stimulation the conductances for both chloride and sodium increase in parallel. This hypothesis would be in accord with recent data on isolated distal lung epithelial cells in which CFTR protein was immunolocalized to these cells and forskolin-stimulated I_sc was inhibited by glibenclamide (195). Other investigators did not find evidence to support the hypothesis that cAMP-stimulated chloride conductance directly controlled sodium uptake in rat fetal distal lung epithelial cells (70), and these investigators concluded that chloride and sodium transport must be linked by some mechanism.

There are several interesting unanswered questions. What is the molecular basis for chloride transport under basal conditions across the distal lung epithelium, since it is clear that CFTR is not required for basal fluid clearance or for clearance of fetal lung fluid at birth (103)? Second, if CFTR is required for cAMP-mediated transport, does this mechanism apply to all levels of the distal lung epithelium including the alveolar epithelial type I and type II cells as well as the distal airway epithelia? Also, does upregulation of fluid clearance by cAMP-independent pathways require functional CFTR (see sect. III)?

There is new information regarding the effect of upregulating the expression of the _2-receptor (₂AR) in alveolar type II cells using the SP-C promoter in transgenic mice (236). ₂-AR expression was 4.8-fold greater than that of the nontransgenic mice (939 ± 113 vs. 194 ± 18 fmol/mg protein, P < 0.001). Basal fluid clearance in the transgenic mice was 40% greater than that of untreated nontransgenic mice (15 ± 1.4 vs. 10.9 ± 0.6%, P < 0.005) and approached that of nontransgenic mice treated with 10⁵ M of the potent -agonist formoterol (19.8 ± 2.2%). Adrenalectomy had no effect on nontransgenic mice (11.5 ± 1.0%) but decreased basal fluid clearance in the transgenic mice to 9.7 ± 0.5%. These findings demonstrated that overexpression of the ₂-AR can be an effective means to increase vectorial fluid transport in the absence of exogenous agonists.

Other investigators have used an adenoviral vector to deliver the ₂-AR to the rat lung, which results in a comparable increase in fluid clearance that was also inhibited by adrenalectomy or by -antagonists (99). In these studies, the ₂-AR may be expressed in several cells in the distal airspaces of the lung. There was evidence that the ₂-AR overexpression was associated with increased levels of ENaC (-subunit) and Na⁺-K⁺-ATPase in apical and basolateral cell membrane fractions isolated from peripheral lungs (88).

Some recent studies suggest that prolonged exposure to exogenous catecholamines may impair the ability of the alveolar epithelium to remove alveolar edema fluid and that this impairment was associated with a reduction in epithelial -receptor number (244). It is important to bear in mind that although a mild impairment was evident at clinically used doses, infused -agonist as well as instilled -agonist stimulated fluid clearance above baseline levels.

In the last few years several interesting catecholamine-independent mechanisms have been identified that can upregulate fluid transport across the distal airspaces of the lung as well as in cultured alveolar type II cells. Hormonal factors, such as glucocorticoids, can upregulate transport by transcriptional mechanisms while thyroid hormone may work by a posttranslational mechanism (197). Some growth factors can work by