The CFTR and ENaC debate: how important is ENaC in CF lung disease?

James F. Collawn, Ahmed Lazrak, Zsuzsa Bebok, Sadis Matalon


Cystic fibrosis (CF) is caused by the loss of the cystic fibrosis transmembrane conductance regulator (CFTR) function and results in a respiratory phenotype that is characterized by dehydrated mucus and bacterial infections that affect CF patients throughout their lives. Much of the morbidity and mortality in CF results from a failure to clear bacteria from the lungs. What causes the defect in the bacterial clearance in the CF lung has been the subject of an ongoing debate. Here we discuss the arguments for and against the role of the epithelial sodium channel, ENaC, in the development of CF lung disease.

  • cystic fibrosis
  • sodium channel
  • mouse models
  • pathogenesis

cystic fibrosis (CF) is characterized by viscous secretions of the exocrine glands in multiple organs and elevated levels of sweat chloride. The most life-threatening features of this disease include chronic bacterial respiratory infections, airway obstruction, bronchiectasis, and respiratory failure (for an excellent review, see Ref. 7). In a landmark discovery in 1983, Paul Quinton demonstrated that CF is caused by a chloride transport defect (31). Two subsequent studies by Knowles, Boucher, and colleagues provided evidence that sodium reabsorption was elevated in the airways (4, 17). These initial studies led to the idea that both sodium and chloride transport were altered in the CF airways.

When the CF gene was cloned in 1989, it was named the cystic fibrosis transmembrane conductance regulator (CFTR) (33) and the gene product was shown to be a cAMP-regulated chloride channel (1). Furthermore, there was substantial support for the idea that part of the pathophysiology found in CF is due to enhanced amiloride-sensitive sodium currents (3, 15, 20, 21, 37, 39) and as discussed below, β-ENaC-overexpressing transgenic mice develop CF-like lung disease (25). A recent study by Grubb and colleagues (13), however, indicated that overexpression of human CFTR failed to correct the CF-like lung phenotype previously observed in the β-ENaC-overexpressing transgenic mouse (13, 25). This is a significant and important finding since the β-ENaC transgenic mouse has been utilized as a model of CF lung disease (25). Moreover, two other reports question the basic premise that loss of CFTR results in ENaC dysregulation in the respiratory system (5, 16). The focus of this Perspectives is to highlight the importance of these recent studies and to discuss how these findings impact our view of the role of ENaC in CF.

What Causes the Lung Pathology in CF?

Before discussing the function of CFTR in regulating ENaC activity, background information is needed to explain the nature of the CF mouse models and the evidence they have provided regarding the role of ENaC in CF pathogenesis. The major consequences of CF include chronic bacterial respiratory infections that are the result of compromised mucociliary clearance (25). Two models have been proposed to explain this defect. In the first model, the compositional hypothesis, loss of the CFTR chloride transport is thought to produce an airway surface liquid (ASL) composition that inactivates antimicrobial peptides (34, 40). In a second model, the low-volume hypothesis, loss of CFTR is believed to promote increased sodium and water absorption and therefore cause a corresponding depletion of ASL volume, resulting in decreased mucus transport and poor bacterial clearance (28, 40). The second model clearly supports the view that loss of CFTR results in an increase in sodium absorption through the ENaC channel.

Murine Models of CF

The controversy over the cause of the CF lung disease been argued for decades. What was needed was an animal model to clarify this issue. CF mouse models first appeared 3 years after the cloning of the CFTR gene, but, surprisingly, these CFTR knockout mice did not develop lung disease (reviewed in Ref. 12). A number of mouse models were created that included a CFTR knockout mouse (Cftrtm1Unc) (35) and a ΔF508 CFTR mouse (Cftrtm1Kth) (12, 41), the latter being a model for the most common mutation found in CF patients. Interestingly, all of the CF mouse models possess an intestinal phenotype, with some of them having a more severe small bowel obstruction than others (12). The mouse models listed above have a severe phenotype with intestinal blockage that causes a large percentage of littermates to die either soon after birth or postweaning (12, 35, 41). The ΔF508 CFTR mice exhibit growth retardation, diminished survival (only ∼40% survive the postweaning crisis), intestinal pathology, a loss of cAMP-activated chloride transport, but no obvious lung pathology (41). Autopsies of these animals indicate severe bowel obstruction, whereas the lungs appear unaffected (41). So although CF mice proved useful for studying the intestinal pathology of CF, they did not provide a good model for respiratory pathology. Unfortunately, in contrast to the mouse models, respiratory symptoms account for the vast majority of the mortality and morbidity in humans suffering from CF (7).

So why is there no lung pathology in the CF mouse? One commonly cited explanation is that epithelial cells lining the murine airways express an alternative (i.e., non-CFTR) chloride channel, whereas mouse intestinal epithelial cells do not (6). Human respiratory epithelial cells also appear to express this alternative, calcium-activated chloride channel (2). Yet the calcium-activated chloride transport mechanism fails to correct the lung pathology in CF patients. With the failure of the classic CF mouse models to mimic the human CF respiratory phenotype, it became clear that the development of novel CF models would be necessary.

The β-ENaC Mouse

In an effort to clarify conflicting hypotheses relating to the basic defect underlying CF lung pathology, Mall, Grubb, Boucher, and colleagues (25) generated transgenic mice to directly test the low-volume hypothesis. They proposed that if the CF lung disease was caused by increased sodium and water absorption from the airways, then overexpression of ENaC in airway epithelial cells should mimic the lack of CFTR-mediated ENaC inhibition in the CF respiratory epithelia (25). Significantly, overexpressing β-ENaC mimics the human lung pathology that the Cftr knockout mice lacked. Epithelia isolated from the β-ENaC mouse airways showed enhanced sodium absorption and the mice exhibited the clinical features of CF airway disease including airway obstruction with dehydrated mucus, neutrophilic inflammation, and poor bacterial clearance (25). The importance and significance of this study are best summarized by a commentary at the time entitled “Finally, mice with CF lung disease” (9).

To generate mice with elevated ENaC levels, Mall and colleagues (25) utilized a Clara cell secretory protein (CCSP) promoter to drive the expression of each subunit of ENaC [Scnn1a (α subunit), Scnn1b (β subunit), and Scnn1c (γ subunit)] to target the lower airway epithelia. Transgenic mice overexpressing each of the transgenes with high transcript levels were analyzed from two different founder lines. Surprisingly, only the β-ENaC-overexpressing animals (hereafter referred to as β-ENaC mice) had elevated ENaC activity (25). The authors also demonstrated that the cAMP-activated and calcium-activated chloride activities were not altered in the β-ENaC mice, demonstrating that the effects were due to the overexpressed β-ENaC rather than changes in chloride transport. Since the relative endogenous message levels of the three ENaC genes were Scnn1a > Scann1b > Scann1c, why the Scann1b gene was functionally limiting remained unclear. Even so, the effects of the ENaC β-subunit overexpression were unequivocal. ASL volume as monitored by ex vivo periciliary liquid (PCL) height in cultured epithelial monolayers indicated that the PCL height was reduced in β-ENaC animals compared with the controls, which is exactly what had been predicted by the low-volume hypothesis.

Complete analysis of the β-ENaC mice revealed a number additional features concordant with CF lung disease (25). In addition to the ASL volume depletion, these mice displayed reduced mucus transport, airway obstruction, and ∼50% postnatal mortality rate at 4 wk, caused by severe airway obstruction. The mice also exhibited neutrophilic lung infiltration and decreased bacterial clearance. These pathological features, found specifically in β-ENaC mice, are consistent with the hypothesis that sodium hyperabsorption is a major contributor to the CF lung phenotype (25).

Rescuing the β-ENaC Mouse Lung Pathology

The functional interaction between CFTR and ENaC is a critical element in the low-volume hypothesis. Therefore, it is notable that in the β-ENaC mouse, the endogenous mouse CFTR is still present. As in any physiological process, a fine balance of all ion transport pathways is necessary for maintaining homeostasis. Thus, in the β-ENaC mouse, the regulatory interaction between CFTR and ENaC expression may well have been altered. In this case, it appears that the endogenous mouse CFTR is unable to suppress the excess β-ENaC effect on ion transport. A simple way to explain this phenomenon is that the ratio of ENaC to CFTR is too great to overcome the enhanced sodium transport. Another possibility is that β-ENaC is expressed in cells where the endogenous CFTR is either absent or extremely low, and these are the airway epithelial cells that specifically promote the CF-like lung phenotype. A third possibility involves the significant differences between the mouse and human lung architecture, making this type of analysis more difficult. Clearly, the cartilaginous mouse lower airways (trachea and bronchi) contain greater than 50% Clara cells, whereas these regions in the human airways contain mostly ciliated cells (14) (Fig. 1).

Fig. 1.
Fig. 1.

Schematic illustration of mouse and human lower airways. Note the cellular diversity of the epithelial cell types lining the airways and submucosal glands. Clara cells are present throughout the lower airway epithelia in mice. Submucosal (mixed serous and mucous) glands are only found in the proximal trachea in mice. The number of Clara cells increases in the bronchiole. Submucosal glands are present throughout the cartilaginous airways (trachea, bronchi) in the human airways and Clara cells only appear in bronchioles. In the alveoli, type I and type II alveolar epithelial cells (AEC) are present both in mouse and human. Type I AECs provide majority of the alveoli lining surface and type II cells secrete surfactant. The β-ENaC and the hCFTR+β-ENaC mice were developed by using a Clara cell-specific promoter to express β-ENaC or hCFTR+β-ENaC only in Clara cells. The architecture of the porcine airways is similar to human with submucosal, mixed glands throughout the cartilaginous airways.

To reverse the β-ENaC mouse phenotype, Grubb, Boucher, and colleagues generated another transgenic mouse strain overexpressing human CFTR (hCFTR), using the same CCSP promoter as for the β-ENaC (13). Given that the β-ENaC mouse had CF-like lung pathology, they hypothesized that increasing CFTR expression above endogenous levels should compensate for the excess β-ENaC function. This strategy was important for two reasons. First, the idea here was to shift the balance to a more physiological CFTR-to-ENaC ratio. Second, this strategy would assure that CFTR and ENaC are expressed in the same cell type, Clara cells. Founder animals (C57Bl/6J × DBA2/J F1) were bred with CB57Bl/6J × DBA2/J F1 animals, and six mouse lines were established. Line 6 was bred with the β-ENaC murine model (C57Bl/6J × C3H/J). These crosses produced β-ENaC, hCFTR, and the hCFTR/β-ENaC (double transgenic) mice on a mixed genetic background (13). Analysis of the hCFTR-overexpressing transgenic mice revealed normal lung histology (13). The hCFTR mRNA was expressed at higher levels (∼5-fold) than endogenous murine CFTR mRNA (13). Ussing chamber analysis of isolated tracheal epithelia demonstrated that the amiloride-sensitive currents were similar in hCFTR transgenic mice and control wild-type mice. Surprisingly, the forskolin and UTP-activated currents (for activation of CFTR and Ca+-activated chloride channels, respectively) were much smaller in the hCFTR transgenic mice. However, considering that the basal currents were also dramatically higher in the tracheal epithelia of mice overexpressing hCFTR, it is likely that the overall constitutive activity of CFTR was elevated in these animals (13). Taken together, overexpression of hCFTR increased basal chloride transport without any respiratory phenotype.

On the other hand, comparisons between the hCFTR, the hCFTR/β-ENaC, and the β-ENaC transgenic mice proved to be quite interesting. For example, Ussing chamber analysis of tracheal epithelial cells from both the hCFTR and hCFTR/β-ENaC transgenic mice showed elevated basal currents and decreased responses to UTP (13). More significantly, the double transgenic mice (hCFTR/β-ENaC) demonstrated many of the same phenotypic traits found in β-ENaC mice. Specifically, the mice had a decreased survival rate, significant mucus plugging, and reduced ASL height, indicating that the transgenic coexpression of hCFTR with β-ENaC in Clara cells failed to reverse the lung pathology associated with the β-ENaC overexpression (13).

Why Doesn't CFTR Rescue the β-ENaC Mouse?

The authors provide a number of potential reasons (13). One simple explanation is that the human CFTR cannot regulate the murine ENaC channel. However, as the authors point out, this is certainly not true when both channels are expressed in oocytes (39). A second possibility is that the presence of another protein is required for a CFTR-ENaC regulatory network, and this protein is not expressed in Clara cells. Third, they noted that β-ENaC expression appeared to be higher in the upper airways, whereas hCFTR expression was higher in the distal airways, suggesting that location matters. Furthermore, the authors estimated that levels of β-ENaC were 25- to 100-fold above levels of endogenous protein expression (13), again eliciting the question “how much CFTR is enough to inhibit ENaC activity?” Both the hCFTR and mouse β-ENaC were expressed only in Clara cells, and it is possible that these cells do not endogenously expresses CFTR and, therefore, lack important binding partners that are necessary for the orchestrated regulation of the ENaC channels (13).

Perhaps the most compelling argument that may explain the failure of hCFTR overexpression to rescue the β-ENaC mouse phenotype involves studies comparing a mouse model of Liddle's syndrome with the β-ENaC mouse (27). Liddle's syndrome is caused by a mutation in the β-subunit of ENaC gene that results in a premature termination codon and creates a gain-of-function sodium channel with elevated activity. Patients with Liddle's syndrome have hypertension, but paradoxically no lung disease (27). Examination of nasal and tracheal epithelial cells from Liddle and wild-type mice revealed several critical points. First, CFTR mRNA expression in nasal and tracheal tissues is similar in Liddle and wild-type mice. Second, CFTR expression is 20-fold higher in nasal vs. tracheal tissues. Third, CFTR inhibits the enhanced ENaC function in nasal epithelia from Liddle mice, but not in the trachea, suggesting that the level of CFTR matters for suppression of ENaC activity. Mall and colleagues also suggest that endogenous human CFTR expression is high in both the upper and lower airways (19), unlike what is found in the mouse (27), and perhaps this could explain the observation that Liddle's patients do not develop lung disease associated with ENaC hyperfunction. Another interesting point raised by Mall and colleagues is that tracheal epithelia from β-ENaC mice express a pool of αβ-ENaC channels that are constitutively active and cannot be further activated by proteolysis (trypsin) (27). However, if the mechanism by which CFTR regulates ENaC activity is to suppress the proteolytic activation of ENaC (10), then presumably an αβ-ENaC channel that does not require proteolytic activation would not be regulated by CFTR, even if expressed in the same cell type. This could explain the β-ENaC mouse phenotype and why CFTR (either the endogenous or the overexpressed hCFTR) is unable to rescue the β-ENaC phenotype.

So what does this mean? Clearly, there are a large number of examples demonstrating that ENaC activity is regulated by CFTR (10, 11, 15, 18, 2024, 37, 39, 40). One argument raised suggests that a certain threshold in CFTR function is required for repressing ENaC. However, a recent study that tested ENaC activity in mouse lung slices does not support this notion (22). In this work, Lazrak, Matalon, and colleagues (22) investigated alveolar epithelial cells in freshly harvested lung slices from wild-type, CFTR heterozygous (Cftr+/−), knockout [Cftr−/− (Cftrtm1Unc)], and Cftr-ΔF508 mice (cftrtm1Kth) using patch-clamp analysis. Cftr−/− and Cftr ΔF508/ΔF508 mice were extensively backcrossed to C57BL/6 background to obtain congenic mice that varied only at the Cftr locus. Patch-clamp studies using the lung slices and single channel recordings by the cell-attached mode from alveolar type I and type II cells (ATI and ATII, distal airways, Fig. 1) revealed some striking findings (22). The open probabilities, Po, of the ENaC channels in ATII cells under basal conditions were 0.21 ± 0.02 (Cftr+/+), 0.39 ± 0.03 (Cftr+/−), 0.81 ± 0.02 (Cftr−/−), and 0.55 ± 0.01 (CftrΔF508/ΔF508). These measurements were performed on 8-wk-old mice and revealed a number of important facts. First, the basal ENaC activity was altered in different Cftr backgrounds. Second, surprisingly, the heterozygote (Cftr+/−) recordings were different from the homozygote (Cftr+/+). In addition, Cftr-null ATII cells had what appeared to be an almost fully active ENaC channel under basal conditions; finally, basal currents from the CftrΔF508/ΔF508 ATII cells were not the same as the null ATII cell currents, suggesting that even ΔF508 CFTR had some effect on ENaC activity (22). These results in murine ATII cells establish that CFTR affects ENaC activity in situ. How much CFTR was expressed in these cells? Western blot analysis failed to demonstrate CFTR expression in isolated type II alveolar cells, but CFTR mRNA was detectable, supporting the view that CFTR could regulate ENaC activity, even when the CFTR protein levels were minimal (22).

Does Dysregulation of ENaC Result in the CF Respiratory Phenotype?

Paul Quinton provides an additional hypothesis that may not rule out other ASL contributions. He suggests that the loss of bicarbonate transport rather than chloride transport through CFTR is the primary cause of increased mucus viscosity and reduced clearance (32). Furthermore, two recent studies, one performed in the neonatal CF pig and the other in human CF airway epithelial cells, dispute the role of ENaC in the development of the CF phenotype (5, 16). In the first study, Welsh, Zabner, and colleagues demonstrated that newborn CFTR−/− pigs spontaneously develop lung disease that is characterized by bacterial infections, lung inflammation, and mucus accumulation (5, 36). Interestingly, newborn ΔF508 pigs also develop lung disease, despite the fact that they have a low-level CFTR-mediated chloride conductance (29). Porcine CFTR−/− epithelial cells show a dramatically reduced chloride transport, but remarkably do not show increased sodium and water absorption or reductions in the PCL height (5) as predicted by the low-volume model. Airway epithelia from the CFTR−/− pigs also demonstrate an increase in amiloride-sensitive voltage and short-circuit currents compared with CFTR+/+ pigs. As the authors point out, it seems paradoxical that pig CF epithelia have a larger change in transepithelial voltage and short-circuit current after amiloride treatment than non-CF epithelia, given that the sodium transport is not affected (5, 36). Their argument for this “amiloride effect” is important for understanding whether sodium transport is altered in CF. Following Ussing chamber analysis of non-CF epithelia, Welsh, Zabner, and colleagues proposed that amiloride treatment “hyperpolarizes apical membrane voltage, thereby increasing the driving force for chloride secretion, whereas lack of CFTR precludes chloride secretion in CF epithelia” (5). In other words, they suggest that the greater drop in short-circuit currents in CF epithelia after amiloride treatment is due to the loss of the CFTR-mediated chloride conductance, not a greater drop in the sodium conductance.

But is the pig CF model different from human CF? Welsh, Zabner, and colleagues (16) addressed this question directly by asking whether sodium hyperabsorption is an important component in human CF lung disease. When they tested primary cultures of tracheobronchial CF and non-CF epithelia in Ussing chambers, they found no evidence for elevated sodium transport. To support these findings, they tested the hypothesis that if the greater drop in transepithelial voltage and short-circuit current in CF epithelia were due to decreased chloride conductance, they should be able to mimic this effect in normal, non-CF epithelia (16). They performed three experiments to test this idea. In the first experiment, they added forskolin and IBMX to the culture medium for 24 h and withdrew it when the non-CF epithelia were mounted in Ussing chambers. This procedure minimizes basal CFTR activity (16, 40) and therefore mimics the loss of CFTR. In the second experiment, they performed Ussing chamber experiments in chloride-free solutions, and in the third, they used the inhibitor, CFinh-172, to block CFTR channels (16). In the first two experiments, the amiloride-sensitive drop in short-circuit current was significantly elevated in non-CF epithelia to levels very similar to those observed in CF, as predicted. In the third experiment performed in non-CF epithelia, pretreatment with CFinh-172 enhanced the drop in current as predicted (16). Each of these treatments has potential limitations as the authors suggest (16). For example, withdrawal of cAMP stimulation or removal of chloride could both affect ENaC activity, and CFTRinh-172 could have other effects besides inhibiting CFTR. That being said, these studies clearly argue against the sodium hyperabsorption model.

Future Perspectives and Unanswered Questions

The introduction of the Cftr knockout mouse illustrated that loss of CFTR was not detrimental in the mouse lung, although it certainly resulted in an intestinal phenotype. Later, the generation of the β-ENaC mouse dramatically raised expectations because of the severe respiratory phenotype. But clearly there were questions that remained unanswered since in the β-ENaC mouse the endogenous mouse CFTR was present, although as noted above the endogenous CFTR and transgenic β-ENaC must have been expressed in different cell types. Correction of the β-ENaC phenotype was predicted in the hCFTR/β-ENaC mice since these mice would express hCFTR in the same Clara cells as the β-ENaC subunit. The lack of correction in the double transgenic mouse was a surprise. Does this mean that Clara cells are the wrong cell type for these studies? What cell types express both CFTR and ENaC at levels that grant measurable functional interaction? Comparisons between mouse nasal and tracheal epithelia suggest that nasal cells express 20 times more CFTR mRNA than tracheal cells (27), and mouse nasal epithelia contains ∼90% ciliated cells compared with less than 40% in the trachea (38). This suggests that the mouse upper airway epithelial cells may be the more relevant mouse model system for studying CF airway pathogenesis.

Does the cell type matter? A recent study reports ENaC expression on motile cilia in human bronchial epithelial cells, whereas CFTR was found in the apical membranes (8). Does this expression pattern eliminate the possibility of functional interactions between the two transporters in ciliated cells? Does expression of αβ-ENaC channels in the β-ENaC mouse provide an explanation for the failure of hCFTR to rescue the respiratory phenotype in the double transgenic mice? Is mimicking the human CF lung pathology so difficult because the lung architecture and distribution of the airway epithelial cell types differ fundamentally in the mouse lung compared with human? If so, why does the β-ENaC mouse model seem to mimic CF lung disease so well? If sodium hyperabsorption is really a critical component in CF lung disease, does amiloride treatment help CF patients? In an early clinical trial using inhalation therapy, amiloride treatment had no beneficial effects (30). However, it is quite possible that longer-acting and more potent analogs of amiloride will work (26). Given the recent data on the newborn CF pig and primary human tracheobronchial epithelia, the role of ENaC in the development of the CF respiratory phenotype is still being argued. Simply put, the CFTR and ENaC debate is far from over.


This work was supported by grants from the NIH (DK060065 to J. F. Collawn, HL076587 to Z. Bebok, and 5U01ES015676 and 5R01HL031197 to S. Matalon).


All authors confirm that they have no competing interests regarding the content, investigations, and results outlined in this manuscript.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
View Abstract