Proliferative regulation of alveolar epithelial type 2 progenitor cells by human Scnn1d gene

Lung epithelial sodium channel (ENaC) encoded by Scnn1 genes is essential for maintaining transepithelial salt and fluid homeostasis in the airway and the lung. Compared to α, β, and γ subunits, the role of respiratory δ-ENaC has not been studied in vivo due to the lack of animal models. Methods: We characterized full-length human δ802-ENaC expressed in both Xenopus oocytes and humanized transgenic mice. AT2 proliferation and differentiation in 3D organoids were analysed with FACS and a confocal microscope. Both two-electrode voltage clamp and Ussing chamber systems were applied to digitize δ802-ENaC channel activity. Immunoblotting was utilized to analyse δ802-ENaC protein. Transcripts of individual ENaC subunits in human lung tissues were quantitated with qPCR. Results: The results indicate that δ802-ENaC functions as an amiloride-inhibitable Na+ channel. Inhibitory peptide α-13 distinguishes δ802- from α-type ENaC channels. Modified proteolysis of γ-ENaC by plasmin and aprotinin did not alter the inhibition of amiloride and α-13 peptide. Expression of δ802-ENaC at the apical membrane of respiratory epithelium was detected with biophysical features similar to those of heterologously expressed channels in oocytes. δ802-ENaC regulated alveologenesis through facilitating the proliferation of alveolar type 2 epithelial cells. Conclusion: The humanized mouse line conditionally expressing human δ802-ENaC is a novel model for studying the expression and function of this protein in vivo .


Introduction
Four subunits (α, β, γ, and δ) of the mammalian epithelial sodium channel (ENaC) have been cloned to date [1][2][3]. α ENaC and splicing δ 1 subunits form functional homomeric channels individually when expressed in oocytes [1,4], whereas co-expression with β and γ subunits can amplify the channel activity Ivyspring International Publisher up to two orders of magnitude [1,4]. The first spliced variant of the human Scnn1d gene, encoding δ 1 -ENaC, was cloned in 1995 [2]. The human Scnn1d is a homolog of degenerins (DEG) of Caenorhabditis elegans, constituting an ENaC branch of the ENaC/DEG superfamily with other three counterparts (α, β and γ subunits) [5][6][7]. Compared with the other three subunits, δ-ENaC is widely expressed in both non-epithelial (i.e., brain, heart, ganglion, placenta, and blood) and epithelial cells (i.e., trachea, kidney, pancreas, liver, and stomach) [4,8]. Functionally, classic epithelial αβγ-ENaC channels are a major regulator for maintaining electrolytes and fluid homeostasis in the lung as well as the kidney [9][10][11][12]. Even though δ-ENaC co-exists with αβγ subunits in numerous cells and tissues [4,8], and regulates the biophysical features and proteolytic properties of αβγ channels [13][14][15], the function of δ-ENaC remains uncertain. Surprisingly, Scnn1d gene is not expressed in rodents, a major obstacle for in vivo functional study [16]. The scarcity of Scnn1d in rodents may explain the discrepancies observed between mice and humans: α-ENaC deficiency results in the death of new-born mice but not human neonates due to unresolved amniotic fluid in the distal airspaces [17][18][19]. In sharp contrast, the major phenotype associated with a deletion in human chromosome 1, which is composed of the Scnn1a gene and others, is growth retardation [20,21]. The expression levels of δ-ENaC is similar to that of α-ENaC in human respiratory epithelial cells, and ~ 40% of amiloride-sensitive sodium transport is associated with δ-ENaC [22][23][24]. Moreover, children with genetic deletion of Scnn1d are predisposed to respiratory infection and nasal congestion [25]. However, the physiological role of δ-ENaC in normal lungs remains unknown. In addition to the paucity of Scnn1d in rodents, the research on Scnn1d has long been hindered by lack of pharmaceutical modulators specific for δ-ENaC activity. We have recently cloned a full-length human Scnn1d gene (δ802-ENaC). Compared with previously identified δ 1 and δ 2 splicing variants that are composed of 638 and 704 amino acid (aa) residues, respectively, this δ 802 -ENaC clone encodes 802 aa [4,8]. The aim of this study thus were twofold. First, to test the feasibility of applying α-13 inhibitory peptide to separate αand δ-type channel populations pharmacologically. α-13 inhibitory peptide is an extracellular segment released by proteolytic cleavage of α-ENaC proteins by proteases [26][27][28]. Second, to characterize the expression and function of human δ 802 -ENaC in vivo in a newly established humanized transgenic mouse line in normal animals.

Results
Cloning and characterization of human δ 802 -ENaC in Xenopus oocytes. Two spliced variants of human δ-ENaC have been reported, δ 1 and δ 2 , which are composed of 638 and 704 amino acid residues, respectively [4,29]. Based on the nucleotide and amino acid alignments of δ 802 and δ 2 clones ( Figure S1), we extended the N-terminal of δ 2 clone and substituted a few amino acid residues. As described previously, the cRNA of δ 802 -ENaC was prepared in vitro for the heterologous expression in Xenopus oocytes [30,31]. Similar to δ 1 and δ 2 clones, δ 802 -ENaC was more permeable to Na + ions over Li + ions when co-expressed with the complimentary β and γ subunits ( Figure 1A). The order of permeability was Na + >Li + >K + >Cs + ions. A linear chord conductance was observed for predominant permeants Na + and Li + ions ( Figure 1B). In contrast, outward currents carried by K + and Cs + ions were greater than those inward charge flows against the K + gradient across the plasma membrane. In agreement with predicted reversal potential by the Nernst equation, the ion selectivity of δ 802 -ENaC acted as a Na + permeable channel. Moreover, amiloride, a specific ENaC inhibitor, suppressed δ 802 -ENaC activity with a k i value of 1.69 ± 0.3 µM ( Figure 1C). The extended N-terminal tail of the δ 802 -ENaC was shown in red font ( Figure 1D). These results suggest that heterologously expressed δ 802 βγ-ENaC channels are characterized by Na + selectivity and amiloride inhibition.
Distinguishing αβγ and δ 802 βγ-ENaC sub-populations functionally with the combination of α-13 peptide and amiloride. Autoinhibitory peptides proteolytically released from αand γ-ENaC subunits functionally block the heterologously expressed mouse and human αβγ-ENaC activity [26,[32][33][34]. However, it is unknown if these peptides alter the δ-containing channels. Given the physical and functional subunit-subunit interactions between four ENaC subunits and the heterotrimeric 3D model of human ENaC protein complexes [35], inhibitory peptide released from γ-ENaC may block both αβγand δβγ-type channels. Thus, we attempted to test if α-13 peptide ( Figure S2) released from α-ENaC post proteolysis inhibits αβγbut not δ 802 βγ-ENaC channels. As shown in Figure 2A, the α-13 peptide (300 µM) has the same efficacy as amiloride (10 µM) to inhibit the human αβγ-ENaC, since subsequent addition of amiloride did not reduce channel activity further. In sharp contrast, the δ 802 βγ-ENaC activity was slightly elevated by the α-13 peptide without statistical significance. All four ENaC subunits are co-expressed in many tissues [29,36,37]. ENaC. The channel activity of heterologously expressed δ802βγ-ENaC was recorded in cells bathed with Na + -, Li + -, K + -, and Cs + -rich bath solutions, respectively. Holding potentials were stepped from -120 mV to +80 mV in an interval of 20 mV. Currents were digitized by the CLAMPEX in the presence and absence of ENaC inhibitor amiloride (10 µM) and then the amiloride-sensitive fractions at each membrane potential were generated with the CLAMPFIT. Dashed line indicates zero current level when the membrane potential was clamped to 0 mV. Scale bars show current level and recording time. (B) Current-voltage relationship of δ802βγ-ENaC. Average amiloride-inhibitable currents (Current) were plotted as a function of membrane potentials (Voltage). The reversible potentials are approximate +13 mV for Na + ions, +7 mV for Li + ions, -54 mV for K + ions, and -116 mV for Cs + ions. n=9. (C) Dose-response curve for amiloride. Accumulating doses of amiloride were perfused to oocytes expressing δ802βγ-ENaC. Current levels at each dose were plotted against applied dose of amiloride and then fitted raw data with the Hill equation to calculate IC50 value (1.69 ± 0.3 µM). n=17. Dashed line (red) is generated with the fitted parameters. (D) Comparison of amino acid sequences at the N-terminal tails of δ802 (full-length) and δ2 clones (previously known as δ2-ENaC). Letters in red font show the extended N-terminal tails and three different amino acid residues in the δ802 clone only. Data in (B) and (C) are mean ± s.e.m. The data were analyzed using one-way ANOVA followed by Tukey post hoc analysis. Average current amplitude at -100 mV for the three types of ENaC channels in the absence (basal) and presence of α-13 inhibitory peptide and amiloride. n=5. * P < 0.05 and * * P < 0.01 vs basal current levels or as indicated by the horizontal lines. (C) Concentration-effect relationship of α-13 inhibitory peptide on αβγ-ENaC channels. Dashed (for αβγ) and dotted lines (for δ802βγ) are generated by fitting the raw data points except the most right one for amiloride with the Hill equation. The retrieved IC50 values for α-13 inhibitory peptide are 0.1 ± 0.01 µM (n=4, Chi 2 = 0.44, R 2 = 0.993), and 0.04 ± 0.07 µM (n=3, Chi 2 = 0.77, R 2 = 0.86), respectively. * indicates the current levels in the presence of amiloride (10 µM). (D) Effects of δ-15 peptide corresponding to α-13 sequence. δ-15 peptide was designed by aligning the amino acid sequences of δ802and α-ENaC subunits. The sequence of 15 amino acid residues corresponding to that of α-13 inhibitory peptide was synthesized ( Figure S2). The same concentration (30 µM) was perfused to the oocytes expressing δ802βγand δ802αβγ-ENaC channels. Amiloride (10 µM) was added to confirm the expression of ENaC. Current data at -100 mV were mean ± s.e.m. * * P<0.01 vs basal levels. n=3.
Approximately half of the total activity associated with δαβγ-ENaC channels could be suppressed by α-13 peptide, whereas approximately 80% of the α-13 peptide-resistant remaining fraction of current amplitude was inhibited by amiloride ( Figure 2B). Over expression of δ 802 -ENaC increased the fraction of the amiloride-resistant current ( Figure  2B), and improved the sensitivity of αβγ channels to α-13 peptide (k i : 0.1 ± 0.01 µM for αβγ vs 0.04 ± 0.07 µM for δ 802 αβγ) ( Figure 2C). Furthermore, we tested the potential effects of a peptide in δ 802 -ENaC (δ-15 peptide, Figure S2) analogous to α-13 peptide sequence ( Figure 2D). This peptide did not significantly affect the activity associated with both δ 802 βγ and δ 802 αβγ channels. These data show that α-13 peptide can be used to separate native αβγand δβγ-ENaC channel activities in epithelial tissues.
Whether the full-length δ 802 -ENaC with extended N-terminal is catalyzed by a serine protease is uncertain. We thus incubated the oocytes expressing δ 802 βγ subunits with serine proteases. Two-chain urokinase plasminogen activator (tc-uPA) ( Figure 3A) and plasmin ( Figure 3B), two crucial molecules of the plasmin(ogen) signal pathway for fibrinolysis, activated the δ 802 βγ activity in minutes. Next, the results showed that γ-ENaC subunit could be a target of plasmin ( Figure 3C). The activation of δ 802 βγ-ENaC activity could be up and down regulated by βand γ-ENaC subunits, respectively. To test whether the fibrinolytic activity alters the blockade of α-13 peptide on ENaC channels, oocytes were pre-treated with plasmin and antiprotease aprotinin ( Figure 3D). Plasmin potentiated both αβγand δ 802 αβγ-ENaC activities significantly to a similar level. On the other hand, aprotinin did not alter the current amplitude significantly. Thus, the α-13 inhibitable fraction of the αβγ-ENaC channels was not significantly affected by fibrinolytic activity (Figure 3E), suggesting that α-13 peptide may not be cleaved by plasmin, and that the binding site of α-13 peptide may not be altered by the proteolysis-induced structural changes in ENaC protein complexes. Similarly, the activity of δ 802 αβγ channels was inhibited by α-13 peptide to approximately 50% of the total current magnitude in the presence and absence of plasmin and aprotinin. To examine the cleavage of ENaC proteins by plasmin, we incubated oocytes co-expressing δ 802 -(HA-and His-tagged) and γ-ENaC (HA-and V5-tagged) with complementary subunits with either chymotrypsin or plasmin ( Figure 3F). Neither chymotrypsin nor plasmin nor endogenous furin cleaved the δ 802 subunit. In striking contrast, γ-ENaC was catalyzed by furin-like protease and further by plasmin as shown with the cleaved band(s) in different sizes. These results suggest that activation of δ 802 βγ-ENaC channels may be due to the cleavage of γ-ENaC by plasmin. Moreover, both α-13 peptide per se and its inhibitory effects on the ENaC function were not altered by fibrinolytic activity, which is most frequently cleaves the γ subunit.
Inducible expression of human Scnn1d in mouse airway and lung epithelium. In contrast to human and other species, mice do not express Scnn1d gene [8,36]. To overcome this barrier for studying human δ-ENaC in vivo , we developed a humanized inducible transgenic mouse strain expressing newly cloned full-length human δ 802 -ENaC ( Figure S3). Because commercially available antibodies against human δ 802 -ENaC have not systematically characterized in vivo and potential post-translational modifications may modify the 3D structure to dissect or cover the antigens, we attached HA and His tags to the N-and C-termini of the δ 802 -ENaC, respectively. The tagged construct was then inserted into EGFP-tagged ROSA 26 allele post the stop codon ( Figure 4A). Sox2 Cre was cross-bred with the humanized δ 802 Tg to express human δ 802 -ENaC globally to mimic the expression pattern in human tissues. Using the primers for the stop codon, we detected a PCR product of approximate 200 bp only in induced lungs ( Figure 4B). As detected with qPCR with the primers for ENaC subunits, the expression of human δ 802 -ENaC in mouse lungs was only found in the induced animals ( Figure 4C). Comparing to the wt controls, the native αand γ-ENaC transcripts were reduced slightly but β subunit was transcribed to a greater extent post induction of the human Scnn1d expression. These observations were further corroborated by the expression of δ 802 -ENaC proteins in whole lung lysates by Western blot ( Figure 4D), with a size of 87.85 kDa band as predicted and a larger band at about 110 kDa. Both bands were confirmed as human δ 802 -ENaC proteins with mass spectrometry ( Figure S4). To locate the subcellular expression of δ 802 -ENaC in the apical membrane of the airway and lung epithelial cells, we grew primary mouse AT2 cells and mouse tracheal epithelial (MTE) cells as polarized monolayers. The results show that δ 802 -ENaC was localized at the apical membrane as recognized by anti-HA antibody in biotinylated apical proteins from both primary mAT2 monolayer cells ( Figure 4E) and MTE cells ( Figure 4F). In addition, the band at 110 kDa was only found in polarized MTE cells. These studies demonstrate the expression of δ 802 at the apical membrane of the respiratory epithelium in vivo .  Bioelectric features of δ 802 -ENaC channels in primary MTE and AT2 monolayer cells. To characterize the function of human δ 802 -ENaC expressed in the airway and alveolar epithelial cells, we harvested primary MTE and AT2 cells to grow polarized tight monolayers at the air-liquid interface ( Figure S5). The basal transepithelial short-circuit current (Isc) level of MTE monolayers was greater for δ 802 -ENaC group over the controls ( Figure 5A). The Isc currents in both control and δ 802 ENaC monolayers were gradually inhibited by accumulating concentrations of amiloride (1 -1,000 µM) and eliminated by removing Na + ions from the bath solutions. The blockade of the ENaC activity in monolayers by amiloride was concentrationdependent with a greater apparent k i value for δ 802 -ENaC group ( Figure 5B). Furthermore, we repeated these observations in the AT2 monolayers ( Figure 5C). Like those in the MTE cells, expression of the human δ 802 -ENaC in AT2 cells elevated Isc amplitude significantly. In both MTE and AT2 monolayers, the apparent k i value of amiloride was much lesser in controls than that in the δ 802 -expressing cells ( Figure 5D). We measured monolayer resistance and a significant difference was measured between controls and δ 802 -expressing AT2 cells ( Figure 5E). The results in oocytes ( Figure 2) suggest that α-13 inhibitory peptide may be used to distinguish native α-type and human δ-containing ENaC subpopulations. To test this hypothesis, we applied the α-13 autoinhibitory peptide to the apical counterpart followed by amiloride and then by Na + -free bath solution in MTE monolayers ( Figure  5F). As shown in Figure 5G, Isc values were suppressed to a different extent between control and δ 802 -ENaC group by α-13 peptide, amiloride, and Na + ion depletion. α-13 peptide-inhibitable fraction of Isc was 96% and 52%, respectively, for control and δ 802 -ENaC cells ( Figure 5H). Finally, we confirmed these observations by measuring ex vivo alveolar fluid clearance in ex vivo human lungs ( Figure 5I). Approximately half of the total alveolar fluid clearance was inhibited by α-13 peptides, suggesting that there are at least two ENaC sub-populations (i.e., α-type and δ-type) in human lungs, and that the subsequent application of α-13 peptide (αβγ-ENaC) and amiloride (both αβγ and δβγ channels) can separate them functionally in vitro and in vivo .
δ 802 -ENaC up regulated alveologenesis in vitro . ENaC is expressed in pulmonary epithelial stem and progenitor cells and implicated in the repair of injured epithelial and other tissues [40][41][42][43][44]. We deciphered the contribution of δ 802 -ENaC in mouse progenitor AT2 cells to alveologenesis in 3D organotypic cultures ( Figure 6A). Organoids were imaged under DIC and confocal microscopes ( Figure S6). A greater number of alveolar-like organoids was observed in the δ 802 -expressing AT2 cells compared to the controls ( Figure 6B), indicating an improved efficiency to generate hollow spheroids ( Figure 6C). Moreover, AT1 and AT2 cells were tracked with their specific markers pdpn and sftpc, respectively, to evaluate proliferation and differentiation in individual organoid ( Figure 6D). Intriguingly, δ 802 -ENaC augmented the efficiency of AT2 renewal significantly ( Figure 6E). This observation was further confirmed by sorting AT1 (ICAM + ) and AT2 (EpCAM + ) cells by FACS ( Figure 6F-G). In addition, we performed EdU incorporation assays in mouse AT2 monolayers ( Figure 6H). More EdU positive AT2 cells were found in cells expressing human Scnn1d over those of the wt controls ( Figure 6I). The clinical relevance of generating more alveoli by δ 802 -ENaC expression may be to re-epithelialize a larger luminal surface area than controls ( Figure 6J).

Discussion
The main findings of this study can be summarized as follows. We established a novel humanized transgenic mouse model conditionally expressing full-length human Scnn1d gene. We found that the Scnn1d gene is a regulator of alveolar progenitor AT2 cells and also functions as a critical transepithelial sodium transport pathway.
Like the previous cloned two splicing variants, the δ802-ENaC functions as a Na + selective amiloride-inhibitable Na + channel, because the selectivity filter and amiloride-binding sites are localized in the second transmembrane domain. The selectivity filter of δ-ENaC, as predicted by in silico analysis, is composed of G-A-S vertically. By comparison, the most common motifs across species for α-, β-, and γ-ENaC are G-S-S, G-G-S, and S-C-S, respectively [45,46]. Interestingly, the selectivity filter for ASIC channels is also G-A-S with a greater Na + permeability over Li + ions, and both δ 1 βγ and δ 2 βγ channels are activated by external protons. The model of ion selectivity filter is composed of three layers vertically [45,46]. The second layer is different between αβγ (S-G-C) and δβγ-ENaC (A-G-C) channels. In addition, divergent second layers for FaNaC, PPK, RPK, MEC-4, UNC-8, and BLINaC of the ENaC/DEG family are reported [9]. This could cause the diverse P Na /P Li ratio for δ 802 -ENaC (1.6 vs 0.6 for αβγ ENaC) as evidenced in ASICs, MEC-4, RPK, and BLINaC [9,47,48]. However, it has not been confirmed by mutagenesis whether the second variant layer of the δ 802 -ENaC is critical for the selectivity of Na + and Li + ions.
Although δβγ-ENaC is less sensitive to amiloride compared with αβγ channels as reported previously [2,24], the study of δ-type ENaC function, in particular for native δ-containing channels has been limited by the paucity of specific pharmacological modulators [49]. Our data provide a novel strategy to separate αβγ and δβγ ENaC activities by subsequently applying α-13 inhibitory peptide and amiloride. This novel strategy can be used for both normal and diseased lungs as the efficacy of α-13 peptide is not influenced by abnormal fibrinolytic activity. In addition, serine protease and α-13 peptide target γand α-ENaC subunits, respectively. Because α-13 peptide and amiloride have different binding sites and subunit-specific for α-13 peptide, there may not be an intermolecular regulation between them.
Our data support the hypothesis that δ-ENaC can benefit the alveolarization by re-epithelializing greater area of alveolar epithelial layer. The potential mechanisms are related to the augmented proliferation of alveolar progenitor mouse AT2 cells. Up regulation of ENaC activity by steroids, CPT-cGMP, miRNAs, and fluid flow improves stem/progenitor cell proliferation in vitro and in vivo [43,44,50,51]. In contrast, suppression of β-ENaC expression reduces cell proliferation significantly [52]. Apparently, both expression and channel activity of δ-ENaC may contribute to the increased proliferation of mAT2 cells in generating alveolus-like spheroids. In addition, increased DNA synthesis in Scnn1d cells could be associated with AT2 proliferation. An interesting topic for future research would be to characterize potential Scnn1d/DNA synthesis/ proliferation/alveologenesis cascade. On the other hand, we cannot exclude the potential contribution of Scnn1a, Scnn1b, and Scnn1g to the fate of AT2 cells. Broadly speaking, the linage of AT2 cells could be regulated by more than one gene. To the best of our knowledge, this is the first study on the regulation of AT2 fate by Scnn1d genes.
Whether the effects of δ 802 -ENaC on AT2 proliferation could be recapitulated in vivo for homeostasis and regeneration remains obscure. We did not see the significant difference in the yield of AT2 cells between adult healthy Sox2 cre and Scnn1d Tg/cre mice. These could be due to considerable loss of AT2 cells during isolation and the difference in niche for AT2 cells between in vivo and in vitro . The scenario for alveologenesis in fetal lungs and regeneration of the air sacs in injured lungs could fit the in vitro 3D cultures.
Although there is a considerable decrease in AT1 cells (pdpn positive) in Scnn1d Tg/cre organoids, as counted by the ImageJ Cell Count Plug-in, the reduction was not seen when total detached cells from all organoids in each Transwell were analyzed by FACS (ICAM positive) (Fig 6f-g). Technically, this discrepancy could be resulted from the unidentical affinity between pdpn and ICAM antibodies. We could not use the same antibody, either pdpn or ICAM, for both confocal imaging and flow cytometry for their diverse applications.
Whether the contribution of delta-ENaC to alveolar fluid clearance is mediated by its proliferative effects on AT2 cells is unclear. Considering that the uncertain life time and solubility of α-13 peptide within the Matrigel for up to 9 days, and the broad effects of amiloride on gene expression [36], other transport function (Na + /H + exchanger, Na + /Ca 2+ antiporter, Na + /Li + exchanger, Na + /K + -ATPase, etc.) [62], and anti-fibrinolytic activity (inhibitors of uPA and plasmin) [63], the results would be much less rigor due to these uncontrollable variables. In addition, the penetration of amiloride and α-13 peptide through the Matrigel is unknown if applied to the culture medium and replaced every other day. Therefore, we simply compared the difference in alveologenesis, AT2 proliferation, and channel activity between wt and Scnn1d Tg/cre cells in parallel. This well-controlled strategy provide rigorous data.
The clinical implications of Scnn1d in airway fluid homeostasis is emerging. Two mutation variants (V541L and P579L) of Scnn1d in three homozygous F508del males were identified recently [64]. They acquired P. aeruginosa and were pancreatic insufficient as diagnosed for essential hypertension, nasal polyps, and cystic fibrosis-related diabetes. δ 1(V541L) βγ mutant when expressed in Xenopus oocytes exhibited reduced channel activity. δ 1 βγ channel activity was inhibited by CFTR when co-expressed in oocytes [65], suggesting that loss of down-regulation of Scnn1d encoded ion channel activity by deficient CFTR could lead to a hyperactive δ802 channel to exacerbate dehydration of the airways. In addition to the function of Scnn1d in human nasal epithelium [22], we for the first time confirmed the contribution of Scnn1d gene to alveolar fluid clearance. Taken together, δ-ENaC may function as a critical pathway for apical fluid and electrolyte homeostasis in normal lungs.
In summary, we identified a novel role of human δ-ENaC in lung epithelial cells in a newly developed humanized mouse colony. This mouse model together with the novel combination of α-13 peptide and amiloride may pave the way for a new path for investigating the physiological roles of human δ-ENaC in vivo .

Methods
Human AT2 cell isolations. Human AT2 cells were isolated with a modified protocol [66,67]. Briefly, distal lung tissue was obtained and dissected into rough 5 cm 3 pieces. Tissues were washed in 500 ml sterile PBS for 10 minutes at 4°C at least two times, or until PBS no longer appeared bloody. An additional 10 minutes wash was then performed with Hank's buffered saline solution (HBSS). Tissue was compressed with autoclaved Kim wipes to remove as much liquid as possible and further dissected into 1 cm 3 pieces. Sterile HBSS buffer containing 5 units/ml dispase II and 0.1 mg/ml DNase I + penicillin/streptomycin was added to the small tissue pieces. The digest solution at this point was rapidly taken up by the tissues, becoming visibly engorged, and was digested 2 hours at 37°C. Fungizone (1:400) was added for the final 30 minutes of the digest. The digest solution was then stored overnight at 4°C without further degradation of cells due to lack of dispase activity at this temperature. The digested tissue was warmed to 37°C and liquefied with an Osterizer 12 speed blender as follows: 5 sec pulse, 5 sec grate, and 2-5 sec pulse. The suspension was poured through a glass funnel lined with sterile 4 × 4 gauze, applying some compression in order to recover as much of the solution as possible. The cell suspension was sequentially filtered through 100 μm, 70 μm, and 40 μm cell strainers. Finally, red blood cells were removed using the red blood cell lysis buffer (Sigma-Aldrich). Antibodies against human CD31, CD45, and EpCAM were purchased from BioLegend (San Diego, CA). Flow cytometry was performed using a FACSCanto II flow cytometer and FACSAria III sorter (BD Immunocytometry Systems, San Jose, CA) and analyzed using Flow Jo 9.6.4 software (Tree Star, Ashland, OR).
Alveolar fluid clearance in human lungs ex vivo. The studies followed the guide of the Declaration of the People's Republic of China and were approved by the Ethics Committee of the China Medical University (CMU) at Shenyang, China. All patients were given oral and written informed consent forms. Human lung tissues were obtained during pulmonary resection surgeries with lung cancer at the First Affiliated Hospital of CMU. There were no fibrous or emphysematous lesions as assessed by preoperative chest CT scanning. Human lung segments were prepared, and AFC measurements were done as described previously [68]. Briefly, the segmental bronchus was occluded by a 10-Fr. balloon catheter immediately after removal of the lung, and occluded segments that were located furthest away from the focus of tumor were chosen. A warmed physiologic saline solution (20 ml; 37°C) containing 5% BSA with or without amiloride (1 mM) was instilled into the distal air spaces through the catheter. After instillation, the lungs were inflated with 100% oxygen at an airway pressure of 7 cm H2O. Alveolar fluid was aspirated 60 minutes after instillation. Aspirated alveolar fluid was centrifuged at 3,000 × g for 10 minutes, and the supernatant was obtained for measurement of protein concentrations. AFC values were calculated as follows [68]: AFC = [(V i -V f )/V i ] × 100, where V is the volume of the instilled albumin solution (i) and the final alveolar fluid (f). V f = V i × P i /P f , where P is the concentration of protein in the instilled albumin solution (i) and the final alveolar fluid (f).
In vivo alveolar fluid clearance (AFC) in mice. In vivo AFC rate was measured as previously described [68][69][70]. Briefly, mice were placed on a continuous positive airway pressure system delivering 100% O2 at 8 cmH 2 O. All animals were maintained at a temperature of 37°C with a heating pad and ultrared bulb. An isosmotic instillate containing 5% bovine serum albumin (BSA) was prepared with saline. To maximize the collection of instilled BSA solution, the diaphragm was dissected. Methodological concerns were addressed by using amiloride to inhibit water movement to confirm that measurements accurately reflected AFC. The endogenous murine albumin would lead to an overestimation of AFC rates. This potential bias was corrected by measuring murine albumin (AssayPro, St. Charles, MO) and bovine albumin (Bethyl Laboratories, Montgomery, TX) separately using specific ELISAs following the manufacturer's instructions [69]. Final AFC values were corrected removing murine albumin present in the aspirate. In addition, we used an aliquot of the instillate that had been delivered into the lungs and removed within 5 min, rather than freshly prepared naive instillate per se. Any pre-existing edema or leaking murine plasma proteins would be excluded for it dilutes both the control at time point 0 min (Pi) and final samples (P f ). AFC rates were calculated as aforementioned for human lung AFC ex vivo.
Mouse AT2 cell isolation and monolayer cultures. Mouse AT2 cells were isolated from both wt and Scnn1d Tg/Cre strains with C57BL/6 genetic background with a modified protocol of a previous publication [71,72]. Briefly, lungs of euthanized mice were removed and incubated in dispase for 45 minutes, and then were gently teased and incubated in DMEM/F-12 + 0.01% DNase I for 10 minutes. Cells were passed through a serial of Nitex filters (100, 40, 30, and 10 microns; Corning, USA) and centrifuged at 300 × g for 10 minutes. Resuspended cells were incubated with biotin conjugated anti-CD16/32, CD45 and CD119 antibodies (BD Pharmingen, USA) and then incubated with streptavidin-coated magnetic particles. The cells were then incubated for 2 h at 37°C in a plastic culture dish precoated with mouse IgG to remove fibroblasts. Upon passing the request for both viability (> 90%) and purity assays (> 94%) ( Figures S7  & S8), mouse laminin 1 precoated transwells (Corning Costar, USA) were seeded with AT2 cells (10 6 cells/cm 2 ). Culture medium (DMEM/F12 + 2mM L-glutamine + 1% ITS + 1% BSA + 1× non-essential amino acids and 10 % new-born calf serum) was replaced with serum free media post 72 hours and then every 48 hours. Transepithelial resistance (R T ) and potential difference (E T ) were measured with an epithelial voltohmmeter (World Precision Instrument, USA). I EQ values were calculated as the ratio of E T / R T values.
Mouse tracheal epithelial cell isolations and monolayer cultures. MTE cells were isolated from C57BL/6 and Scnn1d Tg/Cre mice and cultured as reported previously [73]. Briefly, mice were anesthetized with intraperitoneal injection of ketamine HCl and xylazine. The trachea proximal to the bronchial bifurcation was isolated and removed. The resected section was immediately placed in PBS. Under a dissecting microscope, esophageal remnants and adherent adipose tissue were removed, and the tracheal sections were opened longitudinally, rinsed in PBS, and rotated in fresh DMEM containing 0.1% protease XIV, 0.01% DNase, and 1% FBS for 24 hours. Cells were pelleted by centrifugation, suspended twice in fresh DMEM containing 5% FBS, and seeded onto 6.5-mm diameter, collagen-coated transwell inserts (Corning-Costar, Lowell, MA) at a density of ∼3 × 10 5 cells/cm 2 . MTE cells were grown in a 1:1 mixture of 3T3 fibroblast preconditioned DMEM (containing 10% FBS, 1% penicillin/streptomycin) and Ham's F-12 medium, supplemented with 10 μg/ml insulin, 1 μM hydrocortisone, 250 nM dexamethasone, 3.75 μg/ml endothelial cell growth supplement, 25 ng/ml epidermal growth factor, 30 nM triiodothyronine, 5 μg/ml iron saturated transferrin, and 10 ng/ml cholera toxin. The culture medium on the basolateral side of the filters was replaced every 48 hours. Apical medium was removed 4 days post seeding and then every time the basolateral culture medium was changed, and cells were cultured for 7-9 additional days at an air-liquid interface. Transepithelial resistance was monitored every other day when culture medium was replaced by use of an epithelial voltohmmeter (WPI, Sarasota, FL). Cell-growing inserts with a resistance >1,000 Ω were used.
Organotypic cultures. AT2 cells were cultured as organoids as previously described [74]. Briefly, MLG-2908 cells (2 × 10 5 cells/ml) were mixed with 6,000 AT2 cells and pelleted down. Cells were then resuspended into a 100 µl mixture (1:1) of growth factor reduced Matrigel (Corning, USA) and organoid media (DMEM/F12 + 2 mM L-glutamine + 10 % active FBS + 1% ITS and 10 μM TGF-β rock inhibitor). The mixed cells (90 µl) were seeded on the 0.4 µm inserts (Corning Costar, USA) and incubated for 30 minutes to allow the matrix to solidify, and 410 µl of medium was added to the bottom well. One half (200 µl) of the medium was changed every other day. After ~8 days, colonies (diameter ≥ 50 μm) were visualized and counted with either a Zeiss LSM510 microscope (Carl Zeiss AG, Germany) or an Olympus IX 73 microscope combined with a microscope objective (4×, NA:0.16, FN:26.5, UPlanSAPo, Olympus, Tokyo, Japan). DIC images of all the fields across transwells were captured with a Hamamatsu photonics CMOS Camera (Orca Flash 4.0; 2,048 × 2,048 pixels) and counted for total number of colonies. Surface area of the organoids was measured by marking area across the colonies and then measured with the ImageJ software ( Figure S6). Cultured mouse organoids were incubated in PBS containing 4% paraformaldehyde for 20 minutes at room temperature for fixation. The fixed organoids were immunostained with AT1 and AT2 antibodies.
Cloning, in vitro transcription, and heterologous expression of δ 802 ENaC and two-electrode voltage clamping assay. δ 802 and other ENaC subunits were generated in human ENaC cDNAs cloned into a pGEM HE vector [4,76]. cRNAs of human α, β, and γ ENaC were prepared as described previously [77]. The stoichiometry of native ENaC complexes is unclear to date. Based on the studies in the last decade, several models have been proposed for αβγ ENaC channels: 1α1β1γ, 2α1β1γ, 3α3β3γ. Based on the EM structure of human αβγ ENaC [35], the currently acceptable model of human αβγ ENaC complexes is heterotrimeric (1α1β1γ). Thus we microinjected δ802:β:γ in a 1:1:1 subunit ratio. Oocytes were surgically removed from appropriately anesthetized adult female Xenopus laevis (Xenopus Express). Briefly, the ovarian tissue was removed from the frog under anesthesia by tricaine-S (Western Chemicals) through a small incision in the lower abdomen. Ovarian lobes were removed and digested in OR-2 calcium-free medium (in mM: 82.5 NaCl, 2.5 KCl, 1.0 MgCl2, 1.0 Na 2 HPO 4 , and 10.0 HEPES, pH 7.5) with the addition of 2 mg/ml collagenase (Roche, Indianapolis). Defolliculated oocytes were injected with ENaC cRNAs into the cytosol (25 ng per oocyte in 50 nl of RNase-free water) and incubated in regular OR-2 medium at 18°C. The two-electrode voltage-clamp technique was used to record whole-cell currents 48 h post injection. Oocytes were impaled with two electrodes filled with 3M KCl, having resistance of 0.5 -2 MΩ. A TEV-200A voltage-clamp amplifier (Dagan) was used to clamp oocytes with concomitant recording of currents. Two reference electrodes were connected to the bath. The continuously perfused bathing solution was ND-96 medium (in mM: 96.0 NaCl, 1.0 MgCl2, 1.8 CaCl 2 , 2.5 KCl, and 5.0 HEPES, pH 7.5). Whole-cell currents were recorded as previously reported [24]. Experiments were controlled by pCLAMP 10.7 software (Molecular Devices), and currents at -40 mV, -100 mV, and +80 mV were continuously monitored with data recorded at intervals of 10 sec. Data were sampled at the rate of 200 Hz and filtered at 50 Hz.
Establishment of humanized transgenic mouse strain conditionally expressing human δ 802 -ENaC. HA and His-tagged human δ 802 -ENaC encoding sequence was constructed by extending the N-terminal of the δ2-ENaC clone and substituting few amino acid residues [29]. The hSCNN1D construct was inserted into wild type ROSA26-EGFP vector. Transgenic mice on a C57BL/6/129Sv genetic background were generated by aggregation of targeted ES-cells with eight-cell stage embryos by the Biocytogen (Boston Co. USA) without marked phenotype of emphysema ( Figures S4 & S9). Germline transmitting mice were genotyped by Southern blotting and PCR for the transmission of the transgene. Conditional induction of δ 802 ENaC expression was carried out by cross bred with sox2 Cre (Jackson Laboratory, 004682) and SPC Cre (Jackson Laboratory, cat.#004682).
Surface biotinylation and immunoblotting assays. Primary MTE and AT2 cells grown on permeable transwell inserts were incubated with cell-impermeant sulfo-NHS-SS-biotin (1 mg/ml, Pierce, cat.#21331) in PBS (pH 8.0) for 30 minutes. Biotinylated cells were harvested and lysed in lysis buffer (Cell Signalling, Danvers, MA). The biotinylated membrane proteins were purified with Pierce High Capacity NeutrAvidin Agarose (Pierce, cat.#29202). Quantified proteins were probed with specific monoclonal anti-HA antibody. The blots were stripped and re-probed with an antibody against β-actin (as a loading control of cytosolic proteins). Protein signals were detected by chemiluminescence (Millipore) with Genemate Blue Light Film (ISC).
Confocal imaging of organoids. Proliferation and differentiation of AT2 cells in 3D cultures on day 7 were determined by immunostaining with rabbit sftpc (EMD Millipore, USA) and Syrian Hamster pdpn (Thermo Fisher Scientific, USA). Primary antibodies were detected by anti-rabbit Alexa Flour488, anti-Syrian Hamster Alexa Flour568 (Thermo Fisher Scientific, USA). Fluorescence images were obtained using a Zeiss LSM 510 confocal microscope and stacked with the Fiji component of the ImageJ. Monolayers and organoids were scanned for Z sections with optimal width from top to bottom. Continuous images were stacked separately for pdpn and sftpc to count the number of positive cells for both by using plugin for cell counter in the Image J. Both pdpn + and sftpc + cells were counted and analysed statistically. Each slide was scanned for 6 different fields (n = 3 different experiments). For three-dimensional reconstruction, a series of optical sections that were obtained by confocal microscope were collected at 0.5-μm intervals moving progressively across the cells. For 3D structure, all Z sections were stacked from top to bottom and saved as AVI files.
FACS assays of differentiation and proliferation. For analysis of AT2 cells proliferation and differentiation, AT2 organoids from different experimental groups were dissociated from Matrigel with dispase (10 units/ml) and digested in 0.25% trypsin-EDTA to get single cells suspension. Cells were then stained with antibodies purchased from Biolegend (AF488-EpCAM, APC-ICAM and APC-PDPN). Gates for both colours were set by unstained cells and isotype controls for each antibody. We have used a strategy to use double colour staining (AF488-EpCAM, APC-ICAM) to enhance the separation of AT1 and AT2 cells in organoid suspension.
Cells were analysed by the FACSCaliber TM (BD, USA) and the results were analysed using the FlowJow 10.1 software.
Immunohistology of lung tissues and visualization of AT1 and AT2 cells. Lung tissue sections were first deparaffinized and subjected to antigen retrieval using a citrate buffer at 95°C for 20 minutes. Immunostaining was performed by using antibodies of anti-pro-SPC (1:1000, Thermo Fisher Scientific), anti-AQP5 (1:500, EMD Millipore), and anti-His (1:1000, Abcam). Mouse tissue sections were blocked using a proprietary blocking solution from a M.O.M. kit (Vector Laboratories). The primary antibodies were then incubated overnight at 4°C in the diluent solution provided by the kit, and were visualized with Alexa Fluor 488, 568, and 647 secondary antibodies (Life Technologies). Nuclei were stained with Hoechst 33342 (Thermo Fisher Scientific). Tissue slices were mounted onto slides with ProLong Gold Antifade Reagent (Life Technologies). Tissue staining images were taken by Zeiss LSM 510 confocal systems (Zeiss) for immunofluorescence staining. After fixation, organoids were dehydrated with increasing ethanol grades and embedded as paraffin blocks. Seven-μm thick sections were cut and fixed for additional 5 minutes in 4% PFA at room temperature and stained with a routine H & E protocol.
EdU assay for DNA synthesis. AT2 monolayer cells with active DNA synthesis were detected with Click-iT TM EdU assay kit. Cell monolayers from both wt and Scnn1d Tg/cre groups were stained with the Click-iT TM EdU alexa flour 488 following the manufacturer's instructions. Images were captured and analyzed for the percentage of EdU + cells in different experimental groups. Ten randomly selected images across the monolayers from at least 3 independent experiments were captured and counted for total cells and EdU + portion using a Cell Count plug-in of the ImageJ. The percentage of EdU + cells was calculated for each group, and the difference among groups was compared statistically as described below.
Statistical analysis. All data are presented as the mean ± s.e.m. for n experiments. Differences between means were tested for statistical significance using paired or unpaired t-tests or their non-parametric equivalents as appropriate. Differences between groups were judged using analysis of variance. From such comparisons, differences yielding P < 0.05 were judged to be significant. Either GraphPad Prism or