Tight junctions, but not too tight: fine control of lung permeability by claudins

surviving acute lung injury requires a functional alveolar epithelial barrier to prevent flooding of the air space (25). The alveolar epithelium is not a static impenetrable seal; instead, it serves as a dynamic, selectively permeable barrier that actively optimizes the water content of lung lining fluid. The preeminent role of several classes of epithelial ion channels in mediating fluid reabsorption from the distal air spaces is well-established (6, 11, 17). However, the contribution of ion movement between alveolar epithelial cells via the paracellular route in maintaining lung fluid balance is only beginning to be delineated (14, 21).

Paracellular permeability is regulated by tight junctions, structures localized at sites of direct contact between neighboring cells (18). Tight junctions are composed of several different components, including transmembrane, peripheral, and cytoskeletal proteins, which act in concert to control paracellular permeability. Although tight junction integrity requires the coordinated activity of several different proteins, the ability to regulate fluid movement between cells is directly mediated by a family of two dozen transmembrane proteins known as claudins (1, 9). Claudins not only restrict bulk fluid movement between cells, but also selectively permit paracellular ion diffusion. This is best characterized in the kidney, where claudin-16 and claudin-19 create a paracellular barrier preferentially permeable to sodium and other cations to promote ion resorption (12).

The overall paracellular ion permeability characteristics of an epithelium are determined by the pattern of claudin expression. At least seven different claudins have been found to be expressed by alveolar epithelium where type I and type II cells express different claudin isoforms. Although differences in claudin expression have been linked to differences in alveolar cell phenotype (3, 4, 15) and pathological status of the lung (8, 13), assigning functional roles to specific individual claudins in regulating lung fluid balance has proven difficult.

In this issue of AJP-Lung, Wray et al. (27) used gene expression microarrays to determine that claudin-4 expression was acutely upregulated by mice in response to 3 h of ventilator-induced lung injury (VILI). The increase in claudin-4 mRNA and protein expression correlated with the extent of injury (based on tidal volume, peak airway pressure, and whether positive end-expiratory pressure was used in the ventilation protocol). The increase in claudin-4 was specific, since mRNA expression of other claudins, including claudin-3 and claudin-18, as well as other tight junction proteins was unchanged in response to VILI. Interestingly, cultured alveolar epithelial cells show considerable variation in claudin-4 expression, even within the same monolayer (15, 24), suggesting that claudin-4 expression is more sensitive to cell phenotype or microenvironment as opposed to other claudins that are more uniformly expressed and regulated, such as claudin-18 (10).

The finding of a specific increase in claudin-4 expression was fortuitous, since it enabled a peptide fragment derived from Clostridium perfringens enterotoxin (CPE) to be used as a probe for claudin-4 function in vivo. Full-length CPE is cytotoxic. In the gut, CPE causes a common form of food poisoning by first binding to intestinal epithelial claudin-4, followed by formation of large multiprotein complexes that compromise plasma membrane integrity, which ultimately leads to necrotic cell death (19). The precise mechanisms that lead to CPE toxicity remain an active area of study and have led to the discovery of recombinant CPE fragments that lack the cytotoxic NH₂ terminus yet retain the ability to specifically bind with high affinity to claudin-3 and claudin-4 (7, 16, 26). In fact, cultured epithelial cells treated with CPE-derived peptides show specific depletion of claudin-3 and claudin-4 from tight junctions, leaving other claudins to remain assembled into functional tight junction strands (20). The functional effect varies depending on the claudin composition of the cells challenged with CPE, where cells with tight junctions high in claudin-3 or claudin-4 content will show large decreases in barrier function in response to CPE treatment. Consistent with this, Wray et al. (27) found that treatment of alveolar epithelial cells in culture with a CPE-claudin binding domain peptide (CPE_BD) decreased claudin-3 and claudin-4 content of tight junctions, whereas claudin-18 was unaffected. These changes in claudin composition were accompanied by a ∼20% decrease in transepithelial resistance, but paracellular permeability to a 40-kDa tracer was largely unchanged, suggesting a specific effect of CPE_BD on ion permeability of unstressed, cultured alveolar epithelial cells.

Consistent with the effect on cultured cells, CPE_BD administered in vivo significantly decreased claudin-4 content of the lung. Functionally, CPE_BD dramatically increased pulmonary edema and bulk alveolar protein permeability (leak) in response to severe VILI, underscoring a role for upregulated claudin-4 in protecting the lung from mechanical injury to tight junctions (2). By contrast, alveolar protein permeability was low and largely unaffected by CPE_BD in both the unstressed lung and lungs subjected to moderate VILI. However, CPE_BD treatment increased the amount of lung edema in response to moderate VILI and, importantly, diminished the fluid clearance capacity of unstressed lungs both at baseline and in response to β-adrenergic receptor stimulation. This represents the first in vivo evidence in support of a role for claudins in regulating the fine control of fluid balance in the lung. Considering that claudin-4 preferentially restricts paracellular cation permeability while having little effect on chloride diffusion (22), Wray et al. (27) propose a model that claudin-4 could restrict paracellular backflow of sodium while permitting paracellular chloride diffusion and thus help maintain the active vectorial salt transport needed for optimal fluid efflux from the alveolus. Rigorously testing this model electrophysiologically presents a future challenge.

Although CPE_BD has considerable utility as a tool for manipulating claudins in vivo, one consideration is that CPE_BD binds to both claudin-3 and claudin-4, which are present in the alveolus. Thus the mode of action for CPE_BD may be due to an effect on both claudins, particularly in the unstressed lung. Nonetheless, claudin-3 and claudin-4 might serve similar roles in regulating paracellular ion permeability, given that their first extracellular loop domains are nearly identical (1, 5). Also, claudin-3 is primarily expressed by type II cells (15, 24) raising the possibility that type II-type I cell junctions might be more anion selective in the unstressed lung compared with type I-type I cell junctions. If so, then an overall increase in claudin-4 expression by alveolar epithelial cells in response to mechanical stress could potentially equalize tight junctional ion permeability characteristics throughout the alveolus, as part of a putative mechanism to enhance fluid efflux.

Recent structural analysis of CPE-claudin complexes has identified key motifs that determine the specificity of claudin binding (16, 23, 26). Thus, the possibility of developing CPE-based peptides to specifically interact with different claudins seems feasible. Undoubtedly, targeted claudin-deficient transgenic mice will be needed to better define roles for individual claudins in lung fluid clearance and in the pathology of VILI. Given the relatively mild effect of CPE_BD on unstressed mice, claudin-4 deficient mice should be viable and would be predicted to be hypersensitive to mechanical injury. Transgenic mice would also help determine whether claudin-4 deficiency increases susceptibility to other forms of acute lung injury beyond VILI, such as lung injury driven by inflammation. If this is the case, then identifying strategies to upregulate claudin-4 could provide an attractive approach to help treat pulmonary edema.

GRANTS

This research was supported by the Emory Alcohol and Lung Biology Center [National Institutes of Health (NIH) P50-AA013757] and NIH Grant HL-083120.