Airway surface liquid (ASL) absorption is initiated by Na+ entry via epithelial Na+ channels (ENaC), which establishes an osmotic gradient that drives fluid from the luminal to serosal airway surface. We and others have recently reported that a protease/anti-protease balance regulates ENaC in human airway epithelial cells (HAEC) and provides a mechanism for autoregulation of ASL volume. In cystic fibrosis (CF), this balance is disturbed, leading to constitutive proteolytic activation of ENaC and the pathological Na+ hyperabsorption characteristic of this airway disease. Prostasin is a glycosylphosphatidylinositol-anchored serine protease that activates ENaC and is expressed on the surface epithelium lining the airway. In this report we present evidence that prostasin expression is regulated by the ASL volume, allowing for increased proteolytic activation of ENaC when the ASL volume is high. Prostasin activity is further regulated by the cognate serpin protease nexin-1 (PN-1), which is expressed in HAEC and inhibits Na+ absorption by forming an inactive complex with prostasin and preventing the proteolytic processing of prostasin. Whereas these mechanisms regulate prostasin expression in response to ASL volume in non-CF epithelia, HAEC cultured from CF patients express >50% more prostasin on the epithelial surface. These findings suggest that a proteolytic cascade involving prostasin, an upstream prostasin-activating protease, and PN-1 regulate Na+ absorption in the airway and that abnormal prostasin expression contributes to excessive proteolytic activation of ENaC in CF patients.
- epithelial sodium channels
- protease nexin 1
- bronchial epithelial cultures
amiloride-sensitive epithelial sodium channels (ENaC) mediate Na+ absorption across the apical (luminal) epithelial surfaces lining the lung, colon, sweat duct, and kidney (15). In airway epithelium, Na+ absorption via ENaC in conjunction with a basolateral Na+-K+-ATPase establishes an osmotic gradient that drives airway surface liquid (ASL) absorption (29–31, 41). ENaC activity is increased in cystic fibrosis (CF), which, together with impaired CFTR-mediated anion secretion, leads to pathological ASL depletion and impaired mucus clearance from the lung (20, 23, 27, 29–31, 41). Previous work from our group (34) and Tarran et al. (42) has indicated that a balance between the protease activity of membrane-tethered channel-activating proteases (CAPs) and soluble protease inhibitors in the ASL modulates Na+ absorption in human airway epithelial cells (HAEC). When the ASL volume is low, as in the steady state, the soluble protease inhibitor concentration is sufficient to minimize constitutive activation of ENaC by CAPs in non-CF epithelium. However, when the ASL volume is expanded, the soluble protease inhibitors are diluted, and this allows for CAP-mediated ENaC activation. More importantly, recent evidence suggests that excessive activation of ENaC by CAPs contributes to Na+ hyperabsorption in CF airways (5, 34, 42). However, the soluble protease inhibitor concentration in the CF ASL would be predicted to be high, because the low ASL volume of CF airways would lead to concentration of the ASL proteins. To reconcile the apparent contradiction between a high degree of proteolytic ENaC activation and the predicted high level of protease inhibitor in the CF ASL, we questioned whether the protease/anti-protease balance present at the airway surface is modulated directly by changes in the expression of CAPs.
Prostasin, the human ortholog of CAP1, is a glycosylphosphatidylinositol (GPI)-anchored serine protease that activates ENaC (4, 8, 13, 44). Donaldson et al. (13) demonstrated that prostasin is expressed in airway epithelium and increases the Na+ current by ∼75% when coexpressed with ENaC in Xenopus oocytes. Furthermore, Tong et al. (43) reported an ∼75% reduction in the Na+ conductance of an immortalized airway epithelial cell line following small interference (si)RNA-mediated prostasin knockdown. Recent evidence from Bruns et al. (4) demonstrated that prostasin cleaves the extracellular loop of the γENaC subunit and thereby increases the open probability of the channel. Although the activating effects of prostasin on Na+ channels are well known, little is known regarding the regulation of prostasin expression and activity in the airway.
To investigate whether prostasin expression is regulated to maintain ASL volume homeostasis, we examined the expression of prostasin in primary HAEC cultured on an air-liquid interface with basal and expanded ASL volumes. We observed that prostasin expression and processing are regulated by changes in the ASL volume. Accordingly, the surface expression of prostasin increases under conditions when the ASL volume is expanded, presumably allowing for augmented Na+ and fluid absorption. In CF epithelia, these regulatory mechanisms are altered, leading to increased expression of processed prostasin on the luminal airway surface. Prostasin is synthesized as an inactive zymogen and has not been shown to be capable of autocatalysis (2, 39); activation requires its cleavage by an upstream protease, which is believed to be matriptase (10, 25, 26, 37). Protease nexin-1 (PN-1), an associated inhibitor of prostasin (9), inhibited the amiloride-sensitive Na+ current by formation of an inactive prostasin complex and by preventing the processing of prostasin zymogen to active enzyme. These results suggest that ENaC activity in airway epithelium is determined by a proteolytic cascade involving prostasin, an upstream prostasin-activating protease, and PN-1 and suggest that excessive Na+ absorption in CF airways is caused by abnormal prostasin regulation.
MATERIALS AND METHODS
All cell culture medium was obtained from GIBCO (Invitrogen, Carlsbad, CA), except bronchial epithelial growth medium (Clonetics, San Diego, CA) and Ultroser G (BioSepra, Cedex, France). Recombinant human PN-1 and recombinant human matriptase were purchased from R&D Systems (Minneapolis, MN). Sulfo-NHS-SS-biotin and streptavidin beads were obtained from Pierce Biotechnology (Rockford, IL). Unless otherwise specified, all other reagents were obtained from Sigma (St. Louis, MO).
HAEC were cultured from excess pathological tissue following lung transplantation and organ donation under a protocol approved by the University of Pittsburgh Investigational Review Board. HAEC were cultured on human placental collagen-coated Costar Transwell filters (catalog no. 3470; 0.33 cm2, 0.4-μm pore) as previously described (12, 35) and used for experimentation following 4–6 wk of culture at an air-liquid interface. Where indicated, 20–100 μl of PBS were gently pipetted onto the apical surface of differentiated HAEC to expand the ASL volume. Non-CF HAEC were obtained from donors with chronic obstructive pulmonary disease (11 patients), idiopathic pulmonary fibrosis (6 patients), scleroderma (3 patients), or primary pulmonary hypertension (3 patients). Qualitative differences due to disease state were not observed. CF HAEC were obtained from 12 donors with the following CFTR genotypes: ΔF508, G551D, G542X, N1303K, and two unknowns. Because the majority of the patients had at least one ΔF508 allele, there was insufficient power to assess for differences due to CFTR mutation severity.
Western blotting and surface biotinylation.
Before experimental use, the apical surface of differentiated HAEC cultures was washed once with PBS plus 5 mM dithiothreitol and then three times with PBS to remove the accumulated cellular debris and mucus. After the indicated period of time, the apical secretions were collected in 100 μl of PBS and concentrated by acetone precipitation overnight; the basolateral medium was collected and similarly concentrated. After collection of the secretions, surface biotinylation and Western blotting were performed as previously described (34). Briefly, the HAEC cultures grown on filter supports were placed on ice, and the apical surface was washed with ice-cold PBS plus 1 mM CaCl2 to remove cellular debris. Subsequently, the apical or basolateral surface of the HAEC was incubated with 0.5 mg/ml sulfo-NHS-SS-biotin in borate buffer (85 mM NaCl, 4 mM KCl, and 15 mM Na2B4O7, pH 9). After 20 min, the biotinylation reaction was quenched with PBS plus 10% fetal bovine serum, and the cells were rinsed with ice-cold PBS plus Ca2+. The cells were then lysed in cell lysis buffer (10 mM Tris-Cl, 50 mM EGTA, 0.4% sodium deoxycholate, and 1% NP-40, pH 7.4). Cellular debris was removed by centrifugation, and protein concentration was determined using the Bradford method (Bio-Rad, Hercules, CA). The biotinylated proteins in 100 μg of cellular lysate were recovered by incubation with streptavidin beads overnight at 4°C. The proteins were resolved using standard SDS-PAGE under reducing conditions and transferred to nitrocellulose. The membrane was then immunoblotted using monoclonal anti-prostasin (1:500; BD Transduction Laboratories, San Jose, CA) or monoclonal anti-protease nexin antibodies (1:500; R&D Systems). Subsequently, the blots were stripped and reprobed with monoclonal β-actin antibodies (1:3,000) or monoclonal α1 Na+-K+-ATPase antibodies (1:1,000; Millipore, Billerica, MA). Band intensity was quantified by densitometry (Quantiscan; Biosoft, Cambridge, UK).
Human sputum collection and processing.
Sputum samples were collected from adult patients with CF (mean age: 22 yr) under a protocol approved by the local Institutional Review Board. Sputum samples were processed using Sputolysin (Dade Behring, Deerfield, Illinois) as previously described (33). Briefly, 1 ml of 10% Sputolysin was added per 1 mg of sputum, and the sample was incubated for 5 min at 37°C with vigorous shaking and mixed with a transfer pipette. Samples were then centrifuged at 2,000 rpm for 5 min at 4°C, and the supernatants were used for Western blotting.
Biochemical characterization of prostasin in cultured human airway epithelium.
To determine whether prostasin is attached to the cell surface via a GPI anchor, we exposed the cultured HAEC to 1 U/ml recombinant phosphatidylinositol-specific phospholipase C (PI-PLC; Molecular Probes, Carlsbad, CA) for 1 h before collection of the secreted proteins, surface biotinylation, and lysis (45). Enzymatic digestion of the N-glycans was performed using N-glycosidase F (PNGase F) per the manufacturer's instructions (New England Biolabs, Ipswich, MA).
TaqMan probes for human prostasin and β-actin were purchased from Applied Biosystems (Foster City, CA). The ASL of differentiated HAEC was expanded for the indicated period of time, and the total RNA was subsequently extracted using Trizol (Invitrogen). After purification (RNeasy; Qiagen, Valencia, CA), the RNA was reverse-transcribed to cDNA using the High Capacity cDNA reverse transcription kit (Applied Biosystems). Real-time PCR was performed with an ABI Prism 7900HT sequence detection system (Applied Biosystems), and the relative gene expression was determined using the ΔΔCt method, where Ct is the threshold cycle: relative expression = 2e − (ΔCt sample − ΔCt average of control), where the ΔCt for an individual sample is Ct prostasin − Ct actin.
Short-circuit current recordings.
Before short-circuit current (Isc) recording, the apical surface of differentiated HAEC was exposed to increasing concentrations of recombinant human PN-1 or vehicle control in 20 μl of PBS for 4 h. The addition of 20 μl of PBS for 4 h to the apical surface caused the baseline amiloride-sensitive Isc to increase approximately twofold compared with control filters that remained at an air-liquid interface (data not shown). The Isc were measured as previously described (34). In brief, cells cultured on filter supports were mounted in modified Ussing chambers, and the cultures were continuously short-circuited with an automatic voltage clamp (Department of Bioengineering, University of Iowa, Iowa City, IA). Transepithelial resistance was measured by periodically applying a 2.5-mV bipolar voltage pulse and was calculated using Ohm's law. The bathing Ringer solution was composed of 120 mM NaCl, 25 mM NaHCO3, 3.3 mM KH2PO4, 0.8 mM K2HPO4, 1.2 mM MgCl2, 1.2 mM CaCl2, and 10 mM glucose. Chambers were constantly gassed with a mixture of 95% O2 and 5% CO2 at 37°C, which maintained the pH at 7.4. After a 5-min equilibration period, the baseline Isc was recorded. Subsequently, 1 μM trypsin was added to the apical surface for 5 min, providing a measure of protease-activatable channels (1, 5, 22, 34). The amiloride-sensitive Isc (INa) was determined by the addition of amiloride to the apical cell chamber to a concentration of 10 μM.
In vitro biochemical characterization of the PN-1 and matriptase interaction.
As an initial test for matriptase inhibition, 10 nM recombinant human matriptase was incubated with a 10-fold molar excess of recombinant human PN-1, aprotinin, 4-(2-aminoethyl)-benzenesulfonyl fluoride (AEBSF), or ovalbumin in assay buffer (150 mM NaCl, 50 mM Tris, 10 mM CaCl2, and 0.05% Brij 35, pH 7.5) for 30 min at 25°C. Residual enzyme activity was measured by adding the substrate Boc-Gln-Ala-Arg-AMC (Bachem, King of Prussia, PA) and monitoring the change in relative fluorescence units over time with the FluorImager 575 (Molecular Dynamics, Sunnyvale, CA). To determine the stoichiometry of inhibition (SI), we incubated increasing molar ratios of PN-1 with 10 nM matriptase in assay buffer for 30 min at 25°C. The SI was determined by plotting the ratio of the velocity of the inhibited enzyme to the velocity of the uninhibited control versus the molar ratio of PN-1 to matriptase, and extrapolating to zero activity. PN-1 and matriptase complex formation was determined by incubating 250 nM PN-1 with 50 nM matriptase in assay buffer for 30 min at 25°C, resolving the product by SDS-PAGE, and immunoblotting for PN-1 and matriptase (R&D Systems).
Results are means ± SE. Significance was determined using unpaired Student's t-tests unless otherwise indicated. A P value <0.05 was considered statistically different. For analysis of Western blotting data, the densitometry units from the blots containing both experimental and control conditions were normalized to the darkest band, and the provided n value indicates the number of tissue donors. For each tissue donor, the mean densitometry values were obtained from more than three individual cultures. The IC50 for PN-1 was estimated from the amiloride-sensitive current (normalized to values obtained from matched HAEC treated with vehicle control) plotted as a function of PN-1 concentration fitted to the Hill equation. All statistical analysis and data fitting were performed using SigmaPlot 10 (SPSS).
Prostasin is expressed in a nonpolarized distribution but only secreted and processed on the apical membrane of primary HAEC.
To localize prostasin expression in primary HAEC, we performed Western blotting on the secreted proteins, surface biotinylated proteins, and whole cell lysate of HAEC. Before selective biotinylation of the apical and basolateral surfaces, the secretions into the apical and basolateral compartments were collected and concentrated by acetone precipitation. As shown in the Western blots in Fig. 1A, prostasin migrates as 37- and 40-kDa bands in the apical secretions; however, no prostasin was observed in the basolateral secretions. These results are consistent with previous reports of prostasin secretion from human prostate tissue (48) and from various immortalized mouse cell lines (36, 45) but differ from a study in the immortalized bronchial epithelial cell line JME/CF15, in which prostasin was not present in the secretions (43). To clarify whether prostasin is secreted in the airway, we also performed Western blotting on sputum from a CF patient (Fig. 1B). These data demonstrate that prostasin is apically secreted from HAEC in vivo and in culture.
Because the GPI anchor acts as a strong signal for apical sorting (32), we anticipated that prostasin would only be present in the apically biotinylated proteins. As expected, both the 37- and 40-kDa prostasin molecular species were present in the biotinylated proteins from the apical membrane (Fig. 1A). Surprisingly, the 40-kDa prostasin molecular species was also present in the proteins recovered from the basolateral membrane, suggesting some degree of basolateral delivery. To determine whether the prostasin present on the basolateral surface is attached to the cell membrane via a GPI anchor, we treated the HAEC with PI-PLC. After PI-PLC treatment, the amount of cell surface prostasin decreased and prostasin correspondingly increased in the secretions (Fig. 1C), suggesting that prostasin is attached to both the apical and basolateral surfaces via a GPI anchor. To confirm that the selective apical and basolateral biotinylation reactions were segregated, we stripped the blots and reblotted for Na+-K+-ATPase, a basolaterally expressed membrane protein; as anticipated and as shown in Fig. 1D, Na+-K+-ATPase was only present in the basolateral biotinylated proteins of HAEC. In the whole cell lysate, prostasin migrated primarily as a 40-kDa band with a less prominent 37-kDa band. These results suggest that although prostasin is present on both the apical and basolateral membranes, it is only secreted and processed to a lower molecular mass species on the apical surface.
Next, we assessed whether differential posttranslational modification accounts for the difference in electrophoretic mobility between the 37- and 40-kDa prostasin bands. Because prostasin is known to be N-glycosylated, we examined the effects of glycan digestion in CF and control HAEC (Fig. 2A). After digestion with PNGase, both prostasin molecular species migrated faster, consistent with the previous reports of prostasin glycosylation (2, 38, 39, 45). However, differential glycosylation does not account for the difference between the 37- and 40-kDa prostasin bands, because both molecular species were present following N-glycan digestion.
Prostasin is synthesized as a zymogen and as such requires proteolytic processing to be active (2). Therefore, we questioned whether the 37-kDa band in HAEC represents a proteolytically processed form of prostasin. We reasoned that serine protease inhibition would prevent the formation of the 37-kDa band if this molecular species represents prostasin that has undergone proteolytic processing. As shown in Fig. 2B, 24 h of apical aprotinin exposure increased the abundance of the 40-kDa band and correspondingly decreased the intensity of the 37-kDa band. This suggests that the 37-kDa molecular species represents a form of prostasin that has been processed by a serine protease. The molecular masses of the two prostasin bands are consistent with a recently published report of prostasin zymogen processing by matriptase (CAP3) in mouse epidermis (37). Therefore, both prostasin proteolytic processing and secretion occur selectively on the apical cell surface. This suggests that polarized processing of prostasin is not due to its GPI anchor but, rather, a consequence of polarized protease and/or phospholipase activity.
Prostasin expression is increased in CF HAEC.
The results comparing the posttranslational modification of prostasin in CF and non-CF HAEC (Fig. 2) suggested that epithelia cultured from CF patients express more processed prostasin on the apical membrane than control HAEC. Furthermore, Tarran et al. (42) recently reported that prostasin gene expression was increased nearly fivefold in CF compared with non-CF HAEC in three tissue donors. To determine whether CF epithelia express more prostasin, we performed semiquantitative Western blotting on the apical secretions, apically biotinylated proteins, and whole cell lysate of nine CF tissue donors and morphologically matched non-CF donors (Fig. 3). More prostasin was present in the CF HAEC apical secretions (5,747 ± 213 vs. 4,677 ± 372 mean normalized densitometry units, P = 0.027, n = 9 tissue donors) and apical biotinylated proteins (4,399 ± 387 vs. 2,760 ± 472 mean normalized densitometry units, P = 0.025, n = 9 tissue donors). These results confirm that primary cultured CF airway epithelium expresses increased amounts of prostasin on the cell surface, where the protease is believed to activate ENaC. Furthermore, as shown in Fig. 3D, the ratio of processed (37 kDa) to total (37 kDa + 40 kDa) prostasin was significantly higher in CF HAEC when the two bands were clearly discernable by densitometry (0.78 ± 0.02 vs. 0.54 ± 0.08, P = 0.004, n ≥ 3 tissue donors). These results demonstrate that there is an increased amount of processed prostasin on the CF cell surface.
HAEC coordinate prostasin expression with ASL volume.
Next, we determined whether HAEC regulate prostasin expression in response to changes in the ASL volume. To expand the ASL volume, we applied 100 μl of PBS to the apical surface of differentiated HAEC for 24 h. The apical secretions that had accumulated over the 24-h interval were collected, and the apical surface was biotinylated. As shown in the representative Western blot in Fig. 4A, prostasin expression increased in HAEC under ASL volume expansion conditions, suggesting that cellular mechanisms coordinate prostasin expression with ASL volume. To confirm these findings, we performed semiquantitative Western blotting for prostasin on HAEC cultured from additional tissue donors under air-liquid and ASL volume expansion conditions. Although the amount of secreted prostasin between HAEC under control and ASL volume expansion conditions was not statistically increased (12,493 ± 1,223 vs. 9,029 ± 1,770 normalized densitometry units, P = 0.13), there was a large increase in the cell surface prostasin expression with ASL expansion (5,537 ± 356 vs. 3,547 ± 570 normalized densitometry units, P = 0.007, n = 12 tissue donors). Similarly, prostasin expression in the whole cell lysate also increased with ASL expansion (0.9 ± 0.03 vs. 0.62 ± 0.06 densitometry units normalized to β-actin, P < 0.001 by Mann-Whitney U-test, n = 12 tissue donors). ASL volume expansion had a similar effect on prostasin expression in CF HAEC (data not shown). Therefore, airway surface epithelium regulates prostasin expression in response to changes in the ASL volume.
Next, we assessed the time dependence of the increase in prostasin expression following ASL expansion. Our previous results indicate that ASL expansion rapidly increases Na+ absorption with a half time (t1/2) of ∼30 min and a maximal response after 24 h (34). Real-time PCR and Western blotting for prostasin were performed on HAEC after 0, 2, 6, 12, and 24 h of ASL expansion (Fig. 5, A and B). Prostasin expression increased significantly after 12 h of ASL expansion. Because prostasin expression increased on a protein and message level over the same time interval, the increase in prostasin expression associated with ASL expansion appears to be transcriptionally mediated. These results suggest that ASL expansion may increase Na+ absorption by at least two mechanisms: 1) an initial rapid increase due to the dilution of soluble protease inhibitors in the ASL (34, 42) and 2) a slower increase due to upregulation of prostasin expression.
Endogenous PN-1 regulates ENaC activity through prostasin interactions.
Previously, PN-1, or SERPINE2, has been shown to act as a suicide inhibitor of prostasin through formation of an inactive SDS-stable complex with prostasin (8, 9). In Xenopus oocytes, coexpression of PN-1 with αβγENaC leads to an inhibition of the amiloride-sensitive current INa (47). Furthermore, PN-1 knockdown, using siRNA, augments Na+ absorption in the mouse cortical collecting duct cell line M-1 (47). On the basis of these studies, we hypothesized that PN-1 may act as an endogenous regulator of prostasin and ENaC activity in the ASL of airway epithelium. To confirm the expression of PN-1 in HAEC, we cloned PN-1 from HAEC using real-time PCR, as described by Chen et al. (9), and the sequence was confirmed to be PN-1 (data not shown). As shown in Fig. 6A, Western blotting of HAEC lysate with anti-PN-1 monoclonal antibody revealed an ∼45-kDa band with electrophoretic mobility similar to that of recombinant human PN-1 (rhPN-1), confirming PN-1 expression in HAEC on a protein level. To determine whether the increased prostasin expression in CF HAEC could be explained by a deficiency in PN-1 expression, we compared PN-1 expression between CF and control HAEC (Fig. 6B). There was no difference in PN-1 expression between control and CF HAEC (1.442 ± 0.044 vs. 1.342 ± 0.038 PN-1 densitometry units normalized to β-actin, P = 0.097, n = 3 tissue donors). Thus PN-1 is expressed similarly in CF and non-CF HAEC, suggesting that the increased prostasin present in CF is not due to deficiency of PN-1.
Next, we determined whether PN-1 forms a complex with prostasin in HAEC. As shown in Fig. 6C, when prostasin Western blots of HAEC secretions were exposed for a prolonged period, an ∼82-kDa SDS-stable complex was observed. This suggests that PN-1 forms a complex in HAEC, because the prostasin complex with recombinant PN-1 migrates as an 82-kDa band (8, 9). To confirm that the 82-kDa band on the prostasin Western blots represents a complex with PN-1, we replaced the endogenous apical secretions of HAEC with 100 μl of PBS for 24 h. After apical washout, the 82-kDa band in the apical secretions was no longer present (Fig. 6D), suggesting that this molecular species represents a prostasin complex with a factor in the ASL. The 82-kDa band returned when 1 μM PN-1 was included in the apical solution, suggesting that this band represents a prostasin-PN-1 complex. The addition of other serine protease inhibitors, such as bikunin, known to inhibit both prostasin and Na+ current (3), did not lead to the formation of the 82-kDa prostasin complex. These findings suggest that PN-1 is an endogenous regulator of prostasin in the airway.
Because prostasin inhibition would be expected to inhibit ENaC activity, we determined whether PN-1 has an inhibitory effect on the Na+ conductance of HAEC. Increasing concentrations of rhPN-1 were applied to the apical surface of HAEC for 4 h before measurement of the amiloride-sensitive short-circuit current (INa). As demonstrated in Fig. 7A, PN-1 induced a dose-dependent decrease in INa with an IC50 of 5.5 × 10−7 M. The effects of PN-1 on INa were similar in CF HAEC (IC50 = 2.6 × 10−7 M, n = 4 cultures), suggesting that CF epithelia respond normally to protease inhibition by PN-1. As a protein control, we compared the effects of PN-1 with those of ovalbumin, which is a serpin with no inhibitory activity, on INa. Whereas 1 μM PN-1 decreased INa to 40.4 ± 1.6% of control, 1 μM ovalbumin did not affect Na+ conductance (100.9 ± 0.5% of control, P < 0.001, n = 3 cultures). These data indicate that PN-1 inhibits Na+ transport in airway epithelium.
Proteases activate ENaC by limited proteolysis of the extracellular loop of the α- and γ-subunits, which increases the open probability of the channel (4, 6, 7, 17–19). Therefore, if the inhibitory effect of PN-1 on Na+ current is due to prevention of protease-mediated ENaC activation, one would expect to find an increase in the number of unprocessed inactive channels and a corresponding decrease in the number of processed active channels. To assess this, we measured the INa in HAEC exposed to increasing concentrations of PN-1 before and after apical exposure to 1 μM trypsin (Fig. 7B). The increase in INa following trypsin exposure (Δtrypsin) provides a measurement of the number of unprocessed channels present on the apical surface (1, 5, 22, 34). In parallel to the decrease in INa associated with increasing PN-1 concentration, there was a reciprocal increase in the Δtrypsin (P = 0.01 by ANOVA for both the INa and Δtrypsin trends, n = 4 cultures). After trypsin treatment, there was no difference in the INa among the HAEC exposed to the increasing PN-1 concentrations (P = 0.59 by ANOVA, n = 4 cultures), suggesting that the PN-1 exposure decreased Na+ absorption by interfering with channel gating and not by altering channel number. These results indicate that PN-1 inhibits proteolytic processing of ENaC on the cell surface, leading to a reduction in Na+ conductance.
To determine whether PN-1 inhibits INa though prostasin inhibition, we assessed the effect of PN-1 on prostasin processing. If the inhibitory effects of PN-1 are solely due to formation of an inactive prostasin complex, then the amount of the PN-1-prostasin complex should increase in proportion to INa inhibition. To assess this, we applied increasing concentrations of PN-1 to the apical surface of differentiated HAEC and subjected the apical secretions to prostasin Western blotting. As shown in Fig. 7C, there was a dose-dependent increase in the abundance of the 82-kDa complex over the same PN-1 concentration that caused a large decrease in INa. However, the potent inhibition of ENaC by PN-1 was greater than what would be expected based on the ratio of inactive 82-kDa prostasin complex to total prostasin. This suggests that the decrease in Na+ conductance following PN-1 treatment may not be fully explained on the basis of the formation of an inactive complex of PN-1 with prostasin.
Because prostasin is synthesized as an inactive proenzyme, proteolytic processing is required to activate the enzyme. Therefore, we examined whether PN-1 prevents prostasin processing as an additional means to inhibit prostasin activity. As demonstrated in Fig. 7D, treatment with 1 μM PN-1 caused an increase in appearance of the 40-kDa band and a decrease in the 37-kDa band in the apically biotinylated proteins of HAEC. These results suggest that PN-1 inhibits prostasin by preventing the conversion of zymogen to active enzyme in addition to formation of an inactive prostasin complex. Furthermore, because prostasin has not been shown to be capable of autoactivation (2), these results suggest that PN-1 inhibits an upstream prostasin-activating protease.
To further characterize the mechanism by which PN-1 inhibits INa and prostasin processing, we performed a series of experiments to determine whether PN-1 inhibits an additional protease involved in ENaC channel activation and prostasin processing. Because the type II membrane serine protease matriptase is expressed in airway epithelium (21), has broad serpin reactivity (40), and is believed to be the human ortholog of CAP3 (46), we reasoned that matriptase might be an additional target for PN-1. Furthermore, matriptase is a key physiological activator of prostasin, and this newly described matriptase-prostasin cascade has been demonstrated to be fundamental to normal mouse epidermal differentiation (10, 24–26, 37). Initial tests demonstrated that the enzymatic activity of recombinant human matriptase was abolished by a 10-fold molar excess of aprotinin, AEBSF, or PN-1 (Fig. 8A). These results suggest that PN-1 is capable of inhibiting matriptase, which is a protease that activates both ENaC and prostasin.
To further characterize the interaction between PN-1 and matriptase, we determined the molar ratio at which PN-1 inhibits matriptase and assessed whether PN-1 forms a classic covalent complex with matriptase. The stoichiometry of inhibition (SI) was determined by incubating increasing molar ratios of PN-1 with matriptase and measuring the residual enzymatic activity (Fig. 8B). By extrapolation, the enzymatic activity of matriptase was abolished at an ∼5.5:1 molar ratio. These results indicate that PN-1 acts as a matriptase inhibitor, albeit at a high molar ratio in these in vitro conditions. We next determined whether PN-1 forms a covalent complex with matriptase by incubating a fivefold molar excess of PN-1 with matriptase. As shown in Fig. 8C, the product of the PN-1 and matriptase reaction contained the catalytic domain of matriptase (26 kDa), PN-1 (45 kDa), and the 71-kDa SDS-stable matriptase complex. In addition, an ∼40-kDa band was resolved, presumably due to removal of the PN-1 COOH terminus at the reactive site. These data demonstrate that PN-1 is capable of inhibiting matriptase through formation of a covalent serpin complex and suggest that the proteolytic activation of ENaC on the airway surface involves a proteolytic cascade involving prostasin, matriptase, and PN-1.
An accumulating body of research supports the hypothesis that the activity of Na+ channels is regulated by limited proteolysis of the extracellular loops of the α- and γENaC subunits (4, 7, 14, 16–19, 22, 28). Previous work from our laboratory (34) and Tarran et al. (42) has demonstrated that the degree of proteolytic activation of ENaC is modulated by the balance between channel-activating protease activity and soluble protease inhibitor concentrations present on the apical surface of airway epithelium. In the current studies, we have demonstrated that the expression of prostasin, a channel-activating protease, is regulated by changes in the ASL volume. Accordingly, 1) prostasin transcription and apical expression increase in response to ASL volume expansion, and 2) prostasin processing occurs on the apical cell surface and is mediated by a proteolytic cascade involving PN-1 and upstream prostasin-activating proteases, including matriptase. Most importantly, increased expression of processed prostasin is present on the CF cell surface, suggesting that abnormal regulation of prostasin contributes to the excessive Na+ and ASL absorption characteristic of CF airways.
Under air-liquid conditions, normal airway epithelia possess a reserve pool of inactive Na+ channels that are activated when exposed to CAPs such as trypsin (6), neutrophil elastase (5), or endogenous proteases when the ASL volume is expanded (34, 42). The ability of HAEC to increase Na+ absorption in response to changes in the ASL volume has been attributed to the dilution of endogenous soluble protease inhibitors and a corresponding CAP-mediated ENaC activation (34, 42). In CF HAEC, the reserve pool of unprocessed or partially processed Na+ channels appears to be constitutively activated, leading to an impaired ability to regulate Na+ and ASL absorption. The mechanism responsible for the protease-protease inhibitor imbalance in CF epithelia remains uncertain, since the concentrating effects of excessive ASL absorption would be predicted to increase the protease inhibitor concentration and mitigate protease-mediated ENaC activation. Na+ currents in CF epithelia are inhibited by both PN-1 (Fig. 7) and Kunitz-type protease inhibitors (3, 34, 42), suggesting that the CAPs in CF HAEC are responsive to protease inhibition. The current studies suggest that the increased level of proteolytic ENaC activation in the CF airways is, at least in part, the result of increased expression of prostasin.
The mechanism by which prostasin activates Na+ channels has proven to be complex. ENaC processing by prostasin leads to cleavage of the γ-subunit and a corresponding dramatic increase in the open probability of the channel (4). This increase in channel activity appears to be due to the release of an inhibitory 43-amino acid tract (γ-43) that lies between the furin cleavage site and the putative prostasin cleavage site (4, 17). In support of the theory that prostasin activates ENaC through proteolytic cleavage, prostasin fails to activate ENaC when the prostasin cleavage site on γENaC is mutated or the protease activity is inhibited by aprotinin (4, 13). Interestingly, Andreasen et al. (2) reported that when the catalytic triad of prostasin is mutated, the protease maintains its ability to activate the channel, suggesting that perhaps prostasin's proteolytic activity is not required for channel activation. Furthermore, the Na+ current of oocytes expressing ENaC is increased by prostasin coexpression despite the biochemical absence of proteolytically processed prostasin and under neutral pH conditions where the catalytic activity of the protease is dramatically decreased (2, 48). Subsequently, Bruns et al. (4) reported that γENaC processing occurs in oocytes expressing the catalytic site mutant prostasin, suggesting that either the catalytic mutant prostasin retains a portion of its protease activity or prostasin may facilitate the action of another channel-activating protease. Therefore, although it remains unclear whether the catalytic activity of prostasin is required, γENaC processing near the first membrane-spanning region appears to be critical for channel activation by proteases (1, 4, 16).
The kinetics of increased prostasin expression and INa with ASL volume expansion in this and our previous report (34) suggest two mechanisms whereby ASL volume regulates ENaC activity. The increase in prostasin expression (Fig. 5) occurs at a slower rate than the increase in Na+ absorption associated with ASL expansion [t1/2 of ∼30 min (34)]. We hypothesize that the rapid increase in INa during ASL expansion occurs as a result of protease inhibitor dilution and subsequent CAP-mediated ENaC activation. After prolonged periods of ASL expansion, however, prostasin expression increases as an additional mechanism to maintain heightened levels of Na+ absorption. Further studies are needed to elucidate the mechanism by which HAEC “sense” the ASL volume and to determine the cellular mechanisms that regulate prostasin transcription and protein processing.
PN-1 appears to be a critical regulator of a proteolytic cascade that leads to Na+ channel activation. PN-1 knockdown led to a 1.6-fold increase in INa in a mouse cortical collecting duct cell line (47), and extracellular exposure to PN-1 inhibited INa by ∼60% in HAEC (Fig. 7A). PN-1 appears to decrease INa by multiple mechanisms: 1) direct prostasin inactivation, 2) prevention of prostasin zymogen processing, and 3) direct inhibition of matriptase. Thus the recently described matriptase-prostasin cascade, counterbalanced by PN-1, appears to regulate Na+ channel activity.
PN-1 has recently been shown to be a susceptibility gene for the development of chronic obstructive pulmonary disease in several large population studies (11, 49). Although the functional significance of these single-nucleotide polymorphisms on PN-1 function or expression remain to be determined, it is possible that the airflow obstruction reported in these studies may be due to Na+ hyperabsorption as a result of altered PN-1 activity or expression associated with these polymorphisms. Whether similar polymorphisms could account for a portion of the heterogeneity in CF lung disease among patients with the ΔF508 mutation remains to be determined.
In summary, our data suggest that the airway epithelium regulates prostasin expression as a means of achieving ASL volume homeostasis. These regulatory mechanisms involve prostasin transcription, processing, and inactivation by PN-1. As a result of these complex regulatory networks, prostasin expression increases when the ASL volume is expanded. Thus Na+ and ASL absorption in the airway is regulated by a proteolytic cascade involving prostasin and matriptase, as well as the serpin PN-1. CF epithelia fail to properly regulate prostasin, leading to an increase in the cell surface expression of processed prostasin. These results suggest that abnormal prostasin regulation in CF epithelia leads to excessive proteolytic activation of ENaC and that this plays a significant role in the Na+ hyperabsorption characteristic of CF airway disease. Further defining the regulation of proteolytic cascades on the airway cell surface may reveal novel targets for CF therapeutics beyond the protease inhibitors currently under development.
This work was supported by the Cystic Fibrosis Foundation Shwachman Award (M. M. Myerburg) and the Research Development Program to the University of Pittsburgh (R. A. Frizzell and J. M. Pilewski), Institutional National Research Services Award T32 HL007653 (M. M. Myerburg), National Institutes of Health Grants R01 DK065161 (T. R. Kleyman), R01 DK054814 (R. A. Frizzell), and P30 DK072506 (R. A. Frizzell and J. M. Pilewski), and the American Lung Association (J. M. Pilewski).
We thank Dr. Kenneth McCurry and the Lung Transplant Program at the University of Pittsburgh Medical Center for facilitating tissue acquisition and Joseph Latoche and Elisa Heidrich-Ohare for technical assistance. Real-time quantitative PCR was conducted by the TaqMan core facility of the Genomics and Proteomics Core Laboratories of the University of Pittsburgh.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society