|Home | About | Journals | Submit | Contact Us | Français|
The development of effective therapies for cystic fibrosis (CF) requires animal models that can appropriately reproduce the human disease phenotype. CF mouse models have demonstrated cAMP-inducible, non–CF transmembrane conductance regulator (non-CFTR) chloride transport in conducting airway epithelia, and this property is thought to be responsible for the lack of a spontaneous CF-like phenotype in the lung. Thus, an understanding of species diversity in airway epithelial electrolyte transport and CFTR function is critical to developing better models for CF. Two species currently being used in attempts to develop better animal models of CF include the pig and ferret. In the study reported here, we sought to comparatively characterize the bioelectric properties of in vitro polarized airway epithelia—from human, mouse, pig and ferret—grown at the air–liquid interface (ALI). Bioelectric properties analyzed include amiloride-sensitive Na+ transport, 4,4′-diisothiocyanato-stilbene-2,2′-disulfonic acid (DIDS)-sensitive Cl− transport, and cAMP-sensitive Cl− transport. In addition, as an index for CFTR functional conservation, we evaluated the ability of four CFTR inhibitors, including glibenclamide, 5-nitro-2-(3-phenylpropyl-amino)-benzoic acid, CFTR inh-172, and CFTRinh-GlyH101, to block cAMP-mediated Cl− transport. Compared with human epithelia, pig epithelia demonstrated enhanced amiloride-sensitive Na+ transport. In contrast, ferret epithelia exhibited significantly reduced DIDS-sensitive Cl− transport. Interestingly, although the four CFTR inhibitors effectively blocked cAMP-mediated Cl− secretion in human airway epithelia, each species tested demonstrated unique differences in its responsiveness to these inhibitors. These findings suggest the existence of substantial species-specific differences at the level of the biology of airway epithelial electrolyte transport, and potentially also in terms of CFTR structure/function.
This comparative study of ion transport in four species of polarized airway epithelia will impact how researchers interpret present and future surrogate animal models of cystic fibrosis and other lung diseases.
Cystic fibrosis (CF) is the most common recessive genetic disease in the white population and is caused by a defect in the gene encoding the cystic fibrosis transmembrane conductance regulator (CFTR). CFTR defects result in abnormal electrolyte transport in the epithelia of many tissues. However, lung abnormalities that lead to airway infection and inflammation are the predominant cause of morbidity and mortality in patients with CF. Several lines of CFTR-deficient and -mutated mice have been generated since the CFTR gene was identified. Although these CF mouse models manifest some of the electrophysiologic characteristics of the human nasal and intestinal CF phenotypes, they have failed to accurately reproduce the spontaneous lung infections seen in patients with CF (1–5). This failure may be due to differences in airway cell biology (6), the abundance of submucosal glands (SMG) in the airway (7, 8), and/or the alternative expression or activation of pathways that control non-CFTR chloride channels in airways (4). The lack of an animal model that completely mimics the phenotype of human CF lung disease is a major barrier to understanding CF pathogenesis and to developing therapeutic approaches for this disease. This has prompted researchers around the world to attempt to identify or generate models for human CF in nonhuman primates, cows, sheep, pigs, rabbits, and ferrets.
Because of the extensive conservation in lung cell biology between the domestic ferret (Mustela putorius furos) and human, the ferret has been extensively used to study the pathogenesis and therapy of human type A and B influenza infection (9–11). Also, previous studies have provided evidence that the ferret lung is extremely similar to that of humans with respect to CFTR expression, lung/airway anatomy, airway cell types, and the distribution and abundance of submucosal glands throughout the cartilaginous airways (12–15). For these reasons, the ferret holds potential as a small animal model of CF lung disease, and efforts to produce such a model are underway (16).
The pig represents another potentially useful alternative species for generating a CF model, due to the parallels between its lung biology and that of human. Indeed, the pig has been used as an animal model for chronic bronchitis and can develop submucosal gland hypertrophy, a feature of both chronic bronchitis and CF lung disease (17). In contrast to mouse lung, the porcine lung has marked similarities to its human counterpart in terms of its tracheobronchial tree structure, lung physiology, airway morphology, and the abundance of airway submucosal glands (8, 18). The electrophysiologic properties and functions of the pig airway epithelium and submucosal gland have also been extensively studied and are thought to resemble those in humans (19–27).
The similarities between human, pig, and ferret in terms of lung anatomy, physiology, and cell biology suggest that pigs and ferrets may be two useful candidate species for generating better CF animal models. The successful generation of cloned pigs using nuclear transfer with site-specific genetically modified somatic cells (28), and the recent successful cloning of ferrets by somatic cell nuclear transfer (29), will greatly accelerate the progress of CF model development in these species. Recently, airway epithelial cells from species including human, pig, and mouse have been successfully grown and polarized in vitro, and the resulting cultures have been shown to exhibit electrophysiologic properties of the native airway epithelia (30–35). In the current study, we sought to develop a polarized ferret airway epithelium for comparison of its electrophysiologic properties with those of human, pig, and mouse airway epithelia—within a common system. Our studies demonstrate previously unanticipated diversity in the transport properties of both Na+ and Cl− in airway epithelia from these species, a finding that may reflect functional differences in the activities of ENaC, CFTR, and/or other non-CFTR chloride channels important for modeling CF.
Domestic ferrets (Mustela putorius furos) aged 6–12 mo were purchased from Marshall Farms (North Rose, NY). Animals were injected with an intraperitoneal lethal dose (60 mg/kg) of pentobarbital before harvesting tracheas. The tracheas from adult domestic pigs (~ 100 kg, ~ 6 mo old) were obtained from a local slaughterhouse.
Animal tracheas were processed as previously described for human specimens (31). The tracheas (from the larynx to the bronchial main branches) were opened longitudinally and washed in ice-cold 1% penicillin-streptomycin, 1.0 μg/ml fungizone–Dulbecco's modified Eagle's medium (DMEM) after muscle and vascular tissues were removed by dissection. Fresh medium supplemented with 1.5 mg/ml pronase (Roche Molecular Biochemicals, Indianapolis, IN) and 10 μg/ml DNase I (Sigma, St. Louis, MO) was used for dissociation, and incubation at 4°C proceeded for 36–48 h in the case of ferret tracheas, and 64–96 h for pig tracheas, with occasional inversion of the tubes. Tracheal epithelial cells were harvested after the addition of fetal bovine serum (FBS) to a final concentration of 10%. Dissociation was completed by inversion of the tube 10–20 times. Cell suspensions were transferred to fresh tubes and the tracheas washed twice with penicillin-streptomycin–fungizone-DMEM-FBS. The cell suspensions were pooled and centrifuged at 500 × g for 10 min at 4°C. The cell pellets were then resuspended in DMEM-FBS, and cells were incubated in tissue culture plates (Primera; Becton-Dickinson Labware, Franklin Lakes, NJ) for 2 h, in 5% CO2 at 37°C, to allow for fibroblast adherence. Nonadherent cells were then collected by centrifugation and resuspended in modified bronchial epithelial cell culture medium (M-BEGM) containing 5% FBS medium, after which the total cell number was calculated (31).
Primary human airway epithelial cells were obtained from lung transplant tissue using a protocol similar to that described above for ferrets and pigs (31). Six- to eight-week-old C57BL/6J mice of both sexes were used to generate primary murine tracheal epithelial cells, as previously described (32, 33, 35). These cells were used to generate polarized airway epithelia grown at an air–liquid interface (ALI) as described below.
Several previously reported airway epithelial culture media were used in this study. USG medium contains 2% Ultroser G supplement (Biosepra SA, Cergy-Saint-Chistophe, France) (31). Modified BEGM medium (M-BEGM) and ALI culture BEGM medium (ALI-BEGM) were prepared using previously described recipes (30). All chemicals used in this study were purchased from Sigma unless otherwise indicated.
In vitro polarized airway epithelia ALI cultures were generated as previously described for human and mouse airway epithelial cells (30–33, 35), with minor modification. Briefly, supported polycarbonate and polyester porous (0.4 μM pores) membranes (PCF Millicell inserts; Millipore, Bedford, MA) were pre-coated with filter-sterilized 60 μg/ml type IV human placental collagen (Sigma). Millicell insert membranes (0.6 cm2) were seeded with 2.5 × 105 cells in 5% FBS–M-BEGM and incubated in 5% CO2 at 37°C for 18–24 h. The membranes were washed with pre-warmed PBS to remove unattached cells, and cultured in 5% FBS–M-BEGM (both upper and lower chambers) for two additional days before removal of the medium. The lower chambers were then filled with 2% Ultroser G or ALI culture medium to establish an ALI, and the medium was changed twice a week. The upper chambers were emptied of medium once daily. The transmembrane resistance (Rt) was monitored using an epithelial Ohm-voltmeter (Millicell-ERS; Millipore). A polarized and highly differentiated airway epithelium was achieved ~ 2–3 wk after the ALI was established.
Millicell membranes were fixed with 2.5% glutaraldehyde, stained with 1.25% osmium tetroxide in PBS, dehydrated and sputter coated, and then visualized on a Hitachi S-450 microscope (Tokyo, Japan) for scanning electron microscopy (SEM).
For en face whole mount staining, ALI membranes with differentiated airway epithelia were fixed with 4% paraformaldehyde in PBS at room temperature for 15 min, washed in PBS three times for 5 min, and permeabilized with 0.3% Triton X-100 for 20 min at room temperature. Nonspecific antibody binding was blocked by incubation in 5% normal serum/PBS for 1–2 h at room temperature. Primary antibody against ZO-1 (Zymed Lab Inc., San Francisco, CA) was used at a final concentration of 5 μg/ml in PBS and incubated with epithelial membranes at 4°C overnight. Primary antibody binding was detected using fluorescein isothiocyanate (FITC)-labeled secondary antibody (Jackson ImmunoResearch Laboratories, West Grove, PA). After extensive washing, membranes were mounted on slides with Vectashield (Vector Laboratories, Burlingame, CA) and photographed by fluorescent microscopy.
For immunostaining of sections, tracheal tissue or membranes were embedded in Tissue-Tek OCT compound (Sakura, Torrance, CA). Immunostaining of 10-μm sections was performed as described above, but using the following primary antibody concentrations: rabbit anti-human keratin-14 (1 μg/ml dilution; NeoMarkers, Fremont, CA), mouse anti-keratin-18 (1 μg/ml dilution; Clone IB4, Lab Vision, Fremont, CA). After incubation in FITC and Texas Red secondary antibody, sections were mounted in Vectashield Mounting Medium with DAPI (H-1200; Vector Laboratories). Fluorescent images were acquired with Leica digital camera FDC-300F (Leica Microsystems, Wetzlar, Germany).
Four to six weeks after ALI was established, transepithelial short circuit currents (Isc) were measured using an epithelial voltage clamp (Model EC-825) and a self-contained Ussing chamber system (both purchased from Warner Instruments, Inc., Hamden, CT) as previously described (35, 36). Throughout the experiment the chamber was kept at 37°C, and the chamber solution was aerated. The basolateral side of the chamber was filled with buffered Ringer's solution containing 135 mM NaCl, 1.2 mM CaCl2, 1.2 mM MgCl2, 2.4 mM KH2PO4, 0.2 mM K2HPO4, and 5 mM Hepes, pH 7.4. The apical side of the chamber was filled with a low chloride Ringer's solution in which 135 mM Na-gluconate was substituted for NaCl. Transepithelial voltage was clamped at zero and the resulting Isc was recorded by a Quick DataAcq DT9800 series real-time Data Acquisition USB board (Data Translation, Inc, Marlboro, MA).
For the initial characterization of transepithelial ion transport properties of the ALI cultures, the following chemicals (all purchased from Sigma) were sequentially added into the apical chamber: (1) amiloride (100 μM) for inhibition of epithelial sodium conductance by ENaC, (2) 4,4′-diisothiocyanato-stilbene-2,2′-disulfonic acid (DIDS) (100 μM) to inhibit non-CFTR chloride channels, and (3) a mixture of the cAMP agonists forskolin (10 μM) and the phosphodiesterase inhibitor 3-isobutyl-l-methylxanthine (IBMX) (100 μM) to activate CFTR channels. This was followed by addition of bumetanide (100 μM) to the basolateral chamber, to block all transepithelial Cl− secretion.
Additional experiments were performed to assess the ability of four Cl− channel blockers to inhibit forskolin (10 μM)/IBMX (100 μM)-induced chloride currents in the presence of 100 μM amiloride and 100 μM DIDS. In these studies, the apical chamber was treated with one of the following before bumetanide was added to the basolateral side: 400 μM glibenclamide (34, 37), 100 μM 5-nitro-2-(3-phenylpropylamino)benzoic acid (NPPB) (38, 39), 20 μM CFTRinh-172 (2-thioxo-4-thiazolidinone analog) (40, 41), or 10 μM CFTRinh-GlyH-101 (N-(2-naphthalenyl)-[(3,5-dibromo-2,4-dihydroxyphenyl) methylene] glycine hydrazide) (41).
CFTR expression in differentiated primary ferret and pig trachea epithelia was investigated using RT-PCR. Epithelia were disrupted by the addition of lysis buffer (Qiagen Total RNA Isolation kit) (Valencia, CA) to Millicell apparatuses (5-wk-old ALI cultures), 60-mm dishes (monolayer unpolarized cells) or the lumen of freshly isolated tracheal segments (for native fresh epithelia). Cell lysates were collected and pooled, and total RNA was isolated using a Qiagen Total RNA isolation kit. Synthesis of complementary DNA was performed using a Qiagen Omniscript Reverse Transcript Kit. Briefly, 2 μg of total RNA from each sample were mixed with reaction reagents according to the manufacturer's instructions. Two microliters of the above RT products were used for PCR analysis in a 50-μl reaction volume, with the primers designed from a conserved region of the CFTR NBD1 domain to allow detection of products from human, pig, and ferret. The sequences of the forward and reverse primers are 5′-CTCTTGGAGCAATAAACAAAATAC and 5′-GAATGAAATTCTTCCACTGTGCTT, respectively. The PCR products were resolved by electrophoresis on 1% agarose gels, and the bands expected were 398 bp. An equal amount of total RNA not treated with RT was used as a negative control for specific amplification of cDNA.
A number of earlier studies have described methods for the generation and characterization of human and mouse polarized airway epithelia at an ALI (30–33, 35). Morphologic and electrophysiologic studies have demonstrated that primary differentiated airway epithelia generated in this manner can partially mimic the in vivo characteristics of their respective airway epithelia. More recently, a method for generating in vitro polarized pig airway epithelia has been described, but the properties of these differentiated epithelia were not fully evaluated (34). In the present study we sought to compare the bioelectric properties of polarized airway epithelia from two species (pig and ferret) that are being used in efforts to develop improved CF animal models. In addition, we wanted to reference this new information against previously well-studied human and mouse ALI models. We felt that a direct comparison of bioelectrical properties within a single study was important for the field, given that culture conditions can significantly affect the morphologic and functional characteristics of ALI cultures. Furthermore, methods for ALI cultures of the ferret airway epithelium have not yet been described.
As previously reported for human and mouse, Rt of primary tracheal epithelia stabilized after 2 wk of their establishment at the ALI, and remained stable for up to 4 mo of culture when either 2% Ultroser G or ALI-BEGM media was used (data not shown) (30–33, 35). No morphologic and electrophysiologic differences were observed between the ALI cultures established using these two media (data not shown). We used these two culturing methods as a starting point for developing polarized airway epithelia from pig and ferret tracheas.
The normal body temperatures of ferrets and pigs are 38.5°C and 39°C, respectively. To this end, we first evaluated the growth of each cultured cell type at both 37°C and the species-specific optimal temperature. As shown in Figures 1A and 1B, during the first week of culture Rt increased more rapidly at the normal body temperature for both ferret and pig than at 37°C. However, Rt stabilized by 2 wk after establishment of ALI cultures, with temperature-dependent differences disappearing. Baseline resistances at 4 wk after ALI establishment were similar for human (2.23 ± 0.40 kΩ·cm2), ferret (2.04 ± 0.41 kΩ·cm2), and mouse (1.9 ± 0.76 kΩ·cm2), but lower for pig (1.07 ± 0.26 kΩ·cm2). No morphologic differences were observed between the ALI epithelia cultured at 37°C and 38.5°C (ferret) or 37°C and 39°C (pig) (data not shown). Hence, the functional and morphologic data presented in this report were generated from epithelia cultured in 2% USG medium at 37°C for all four species. ZO-1 immunostaining demonstrated the formation of epithelial tight junctions, for all species, by 2 wk of ALI culture (Figure 1C, shown only for ferret and pig), and this remained the case for the duration of the experiment (data not shown). SEM also confirmed ciliated cell differentiation in airway epithelial ALI cultures from both ferret (Figure 2A) and pig (Figure 2B), although the abundance of cilia was significantly less than that seen in native tracheas from these species (Figures 2A and 2B). The level of ciliogenesis in ferret and pig ALI cultures was similar to that seen in human, and slightly more abundant than that seen in mouse cultures (data not shown).
Since the extent of differentiation of ferret and pig airway epithelia grown at an ALI has not previously been evaluated, we analyzed the expression of two cell-specific cytokeratins, CK14 and CK18. The expression profiles of these cytokeratins are dependent on the epithelial cell type and the stage of epithelial differentiation; in fully differentiated polarized airway epithelia, CK18 is usually expressed in columnar cells, whereas CK14 expression is restricted to basal and intermediate cells. As shown in Figure 2A and 2B, CK14 expression in ferret and pig ALI cultures mimicked the expression patterns in native tracheal epithelia from each species, with expression restricted to basal/intermediate cells. Interestingly, CK18 appeared to be expressed in nearly all cells of the ferret native tracheal epithelium, and the pattern in ALI cultures was similar. In contrast, CK18 expression in the pig native trachea was restricted to columnar cells, but as was the case in the ferret ALI cultures, CK18 also appeared to be expressed in both basal/intermediate and columnar cells of pig ALI cultures. However, due to apparent nonspecific staining of ALI membrane supports with CK18 antibody, we currently cannot rule out that staining of the basal cell layer in ALI cultures is nonspecific. Other differentiated cellular markers, such as high–molecular glycoprotein Mucin 5AC and the ciliary axonemes marker β-tubulin-IV, were also detected in the ferret and pig ALI cultures (data not shown). As previously reported for human and murine primary airway epithelial cultures (30–33, 35), the observed morphologic and immunolocalization data strongly suggest that highly differentiated ferret and pig airway epithelia can be established at the ALI, and therefore support the use of these models in functional bioelectric studies.
Expression of the CFTR gene has been previously demonstrated at the molecular level in murine and human primary airway ALI cultures (30–32, 35, 36). With the goal of evaluating pig and ferret ALI cultures as potential models for studying CF, it was critical to demonstrate that the CFTR gene was indeed expressed in these cultures. To this end, CFTR mRNA expression in pig and ferret ALI cultures was characterized by RT-PCR, using a primer set in the nucleotide binding domain-1 (NBD1) of CFTR; the sequences of these primers are common to CFTR cDNA of human, pig, and ferret. Results from this analysis revealed that the CFTR mRNA was present in primary tracheal ALI cultures from both ferret and pig, and also in monolayer cultures and epithelial cells harvested from native tracheas (Figure 3). These studies support the use of ALI models to study CFTR function in ferret and pig.
Previous studies in human and murine primary airway ALI cultures have demonstrated that cAMP agonists stimulate Cl− secretion. In humans, the sole pathway of cAMP-stimulated Cl− secretion in ALI cultures appears to be through CFTR, as indicated by a lack of transport in CF airway epithelia. However, in mouse ALI cultures the correlation between cAMP-stimulated Cl− secretion and CFTR function is less clear, given that CFTR knockout (KO) tracheal epithelia have cAMP-stimulated Cl− secretion at levels similar to those generated by wild-type animals (4, 32, 35). To compare the transepithelial ion transport properties of tracheal ALI cultures from mouse, ferret, pig, and human, 4-wk polarized tracheal epithelia from each species were mounted in Ussing chambers and changes of Isc were evaluated following the sequential addition of amiloride, DIDS, IBMX/forskolin, and bumetanide as described in Materials and Methods. Under Cl− secretory conditions, baseline Isc of ALI cultures were 14.8 ± 1.3 μA/cm2 (mouse), 16.1 ± 2.4 μA/cm2 (ferret), 60.2 ± 5.2 μA/cm2 (pig), and 42.6 ± 3.7 μA/cm2 (human). The addition of amiloride greatly inhibited Isc in pig and human cultures to similar extents (Figure 4 and Table 1). However, only modest amiloride-sensitive Isc reductions were seen in ferret and mouse cultures (Figure 4 and Table 1). These findings suggest that baseline Na+ transport (and potentially ENaC activity) is greatest in human and pig. The finding of low levels of amiloride-sensitive Na+ transport in ferret ALI cultures is consistent with previous observations in ferret tracheal xenografts that used transepithelial potential differences (PD) as an index for Na+ permeability (15). Moderate responses to DIDS (which inhibits certain non-CFTR chloride channels) were seen in mouse, pig, and human ALI cultures, and ranged from 5–7 μA/cm2. Interestingly, ferret airway epithelia demonstrated a complete absence of DIDS-sensitive chloride channels (Figure 4 and Table 1), which is consistent with the previously reported observation that DIDS does not alter transepithelial PD or chloride secretion in isolated ferret trachea (42).
CFTR-mediated chloride transport in ALI cultures is typically assessed by measuring cAMP-induced and bumetanide-inhibitable current under secretory conditions (low apical chloride) and in the presence of amiloride and DIDS. The addition of cAMP agonists (10 μM forskolin and 100 μM IBMX) significantly stimulated the transepithelial Cl− current in ALI cultures from all species. This cAMP-induced change in Isc was completely inhibited by the addition of bumetanide to the basolateral solution (Figure 4 and Table 1). Pretreatment of human ALI cultures with IBMX/forskolin overnight before assessing Isc has previously been used to amplify CFTR-mediated currents (43). Thus, we also compared bioelectrical properties of cultures that had been pretreated with IBMX/forskolin overnight. Interestingly, this treatment resulted in an ~ 2-fold increased baseline Isc under secretory conditions and, consequently, an increase in the amiloride-sensitive responses in all four species (Figures 4E and 4F). As previously reported, IBMX/forskolin pretreatment of human ALI cultures elevated the cAMP-agonist induced Cl− current typically associated with CFTR. However, a similar pretreatment of mouse, ferret and pig ALI cultures led to a reduction in the cAMP-agonist induced Cl− current (Figures 4E and 4F). The reason for this species-specific difference is currently unknown. Pretreatment of cultures with cAMP agonists overnight is thought to facilitate the accumulation of CFTR in apical epithelial membranes and, as such, to augment the capacity for subsequent activation of CFTR-mediated Cl− currents. Hence, the differences in the effect of cAMP agonists pretreatment in mouse, ferret, and pig ALI cultures may reflect a number of potential biological differences between species, including a shorter half-life of CFTR in the membrane and/or differences in the intracellular trafficking of CFTR.
Previous studies in CFTR KO murine trachea and ALI cultures have demonstrated that CFTR-mediated Cl− channels are not the sole source of cAMP agonist-induced Cl− current in murine airway epithelia (4, 32, 35). Hence, although both pig and ferret ALI cultures express CFTR mRNA and also demonstrate bumetanide-sensitive cAMP-inducible Cl− currents, it remains unclear whether this current is due to CFTR mediated Cl− channels and/or other “CFTR-like” channels as seen in mouse. Although the identity of the channel(s) mediating the cAMP-stimulated non-CFTR chloride currents in the mouse tracheal epithelium remains elusive, it is clear that it is not sensitive to DIDS (35). To assess the species-specific functional contributions of CFTR to the DIDS-insensitive, cAMP-inducible, Cl− currents measured in epithelia, we used two classes of chloride channel blockers: two general anion channel inhibitors including glibenclamide and NPPB, and two CFTR-specific inhibitors including CFTRinh-172 and CFTRinh-GlyH101. The effectiveness of these chloride channel blockers in inhibiting Isc under Cl− secretory conditions was assessed in ALI cultures treated with amiloride, DIDS, and IBMX/forskolin (Figure 5A). The fraction of IBMX/forskolin-inhibited Isc in response to each inhibitor, relative to the total level of inhibition achieved after sequential addition of bumetanide, was used as an index for the effectiveness of each inhibitor to block Cl− transport (Figure 5B).
The anion channel blockers NPPB and glibenclamide have been used to inhibit a number of different anion channels, including both CFTR and non-CFTR channels in the airway epithelium (19, 25, 38, 39). Significant inhibition of the IBMX/forskolin-stimulated Cl− current was observed in human, ferret, and mouse ALI cultures after the addition of glibenclamide or NPPB (Figure 5 and Table 2). In contrast, IBMX/forskolin-induced chloride currents in pig airway epithelia were effectively inhibited by NPPB, but not glibenclamide. However, the kinetics of NPPB inhibition in the pig was notably slower than in all other species. These findings support previous studies on submucosal gland secretion in the pig trachea, in which NPPB inhibited methacholine- and acetylcholine-induced glandular secretions more effectively than glibenclamide (25, 44).
CFTRinh-172 and CFTRinh-GlyH101 are both are high-affinity inhibitors of human CFTR-mediated Cl− transport (41, 45). CFTRinh-172 is a thiazolidinone analog that reversibly inhibits cAMP–dependent CFTR-mediated Cl− conductance by a mechanism that most likely involves direct binding to a critical site on a cytoplasmic domain of CFTR (40). In contrast, CFTRinh-GlyH-101, a glycine hydrazide analog, reversibly inhibits CFTR-mediated Cl− conductance by blocking the CFTR anion pore on the external cell surface, leading to inward rectification of the channel (41). Significant inhibition (> 75%) of IBMX/forskolin induced Cl− current by 20 μM CFTR inh-172 was observed in mouse and human ALI cultures (Figure 5, Tables 1 and and2).2). In contrast, ferret and pig ALI cultures demonstrated significantly less sensitivity to 20 μM CFTRinh-172 compared with those of humans (22.7 ± 3.9% inhibition for ferret and 55.6 ± 6.8% inhibition for pig). The addition of 10 μM CFTRinh GlyH-101 after IBMX/forskolin stimulation significantly inhibited Cl− currents (> 75%) in human, pig, and ferret ALI cultures (Figure 5, Tables 1 and and2),2), although the kinetics of inhibition were notably slower in ferret. In mouse ALI cultures, however, CFTRinh GlyH-101 demonstrated significantly less ability to inhibit IBMX/forskolin-induced Cl− currents as compared with human. This finding contrasts with the effective inhibition previously observed for CFTRinh GlyH-101 in mouse, where nasal PD was used as an index of Cl− permeability (41). These findings suggest that bioelectrical properties of the mouse nasal epithelium likely differ from those of the tracheal epithelium. Given the previously demonstrated differences between nasal and tracheal epithelia of CF mice with respect to cAMP-dependent Cl− secretion (1, 4, 35, 46), this finding is perhaps not unexpected.
The development of effective CF therapies has been hindered by the lack of suitable animal models that fully reproduce the human disease phenotype. Although mouse CF models have provided tremendous insights into CFTR biology, and under certain experimental conditions can manifest CF-like disease phenotypes in the lung, it is clear that additional models are needed to afford a more complete understanding of the human disorder. The basis of differences in CFTR-associated lung disease between humans and mice is not fully appreciated, but likely stems from variation both in lung cell biology and chloride channel composition—with mouse expressing alternative chloride channels that can compensate for the lack of CFTR. Pigs and ferrets have been proposed as alternative platforms for developing better CF models, based on the fact that their airway cell biology more closely resembles that of humans. In the context of the pig, numerous studies have begun to assess the appropriateness of this species as a CF model, using non-specific CFTR channel blockers (NPPB and glibenclamide), a non-CFTR selective chloride channel blocker (DIDS), and a CFTR-selective blocker (CFTRinh-172) to attempt to dissect the contribution of wild-type CFTR to airway fluid and/or electrolyte secretion. However, parallel studies evaluating the potential involvement of CFTR in ferret airway chloride transport have been lacking. Given that the experimental system used for any analysis of airway bioelectrical properties is critical to drawing comparative conclusions between species, we sought to compare CFTR-relevant electrophysiologic parameters between human, mouse, pig, and ferret airway epithelia within a common culture model.
Our comparative analysis reveals several interesting findings with regard to species specificity in Na+ and Cl− currents across polarized airway epithelia. First, baseline Isc in the presence of low apical chloride, as well as amiloride-sensitive Na+ currents, were significantly lower in ferret ALI cultures than in their pig and human counterparts. Both these indexes suggest that ferret tracheal epithelia have low levels of baseline ENaC activation, a finding supported by previous studies evaluating transepithelial PD in ferret tracheal xenografts (15). Given that ENaC dysregulation may be important in CF airway disease, this feature may be a disadvantage of using the ferret as a CF model. Second, ferret ALI cultures demonstrated a complete absence of DIDS-sensitive chloride channels. This was notably different than that seen in human, mouse, and pig ALI cultures. The lack of DIDS-sensitive chloride channels in the ferret tracheal epithelium is also supported by work in tracheal xenografts from this species demonstrating little change in DIDS-sensitive PD in response to low apical chloride (15). Since DIDS-sensitive chloride channels are potential modifiers of CF-associated disease, the lack of such channels in the ferret could actually be an advantage from the standpoint of CF animal modeling. One aspect of chloride channel activation demonstrating no species specificity was cAMP-inducibility—which was present in all species of ALI cultures. However, given that tracheas and ALI tracheal cultures from CF mice are known to express cAMP-inducible chloride channels (4, 35), such a finding cannot be interpreted as specific evidence for CFTR function.
In addition to examining standard electrophysiologic parameters of tracheal epithelia, our study has also attempted to address the pharmacologic profile of DIDS-insensitive, cAMP-inducible chloride currents, with the hypothesis that similarities between a given species and human might reflect the potential for that species to suitably model human CF lung disease. Mouse served as a negative control for this hypothesis, given that tracheal airway epithelia from this species is known to express cAMP-inducible non-CFTR channels that are thought to compensate for the lack of CFTR (4, 35). Results from this analysis demonstrated considerable species-specific variability in the pharmacologic profiles of the four chloride channel blockers evaluated (two nonselective CFTR blockers and two CFTR-specific blockers). Although all of the blockers significantly inhibited cAMP-inducible chloride currents in human ALI cultures, glibenclamide and CFTRinh-172 produced a significantly lower level of inhibition in pig ALI cultures compared with those from humans. Recent studies evaluating nasal potential differences in pig have demonstrated similar findings that both glibenclamide and CFTRinh-172 were ineffective at blocking cAMP-mediated changes in nasal PD (34). Similarly, this same study demonstrated that ALI cultures of pig airway epithelium required significantly higher doses of CFTRinh-172 to inhibit cAMP-induced Isc under chloride secretory conditions compared with mouse. Our study demonstrating that CFTRinh-172 effectively inhibits cAMP-induced chloride current in mouse ALI cultures confirms the latter result. Notably, the newer CFTR-specific inhibitor CFTRinh-GlyH101 was effective at inhibiting cAMP-induced Isc in pig but not mouse ALI cultures. If indeed both CFTRinh-172 and CFTRinh-GlyH101 are specific for mouse and pig CFTR, then one must conclude that the structure/function of CFTR is significantly different in these two species. Ferret ALI cultures also demonstrated a unique anion channel-blocker pharmacologic profile. Of the four blockers evaluated, only CFTRinh-172 exhibited a reduced ability to inhibit cAMP-induced Isc in ferret airway epithelia compared with those from humans. These results again emphasize the potential for unique CFTR structure/function between species and/or differing levels of DIDS-insensitive alternative (non-CFTR) chloride channels that are also uniquely susceptible to these inhibitors.
Several biological parameters could potentially affect CFTR-dependent biology in the airway of the four species studied here. These include a species-specific distribution of airway cell types and cell type–specific chloride channel composition. For example, in mouse tracheal airway epithelia the Clara cell is the predominant secretory cell type. This differs from human, pig, and ferret, in which goblet cells are the predominant secretory cell types of the proximal airway. As mentioned above, the structure/function of CFTR may also have evolved differently among various species. Indeed, studies evaluating mouse CFTR have demonstrated differences in the biological parameters of this channel relative to its human counterpart (47, 48). Given that mouse CFTR is only 88% conserved with (78% identical to) human CFTR at the amino acid level (Figure 6, Table 3), it is not surprising that such differences in channel properties exist. In contrast, pig CFTR amino acid sequence is 95% conserved with (92% identical to) human CFTR (49) (Figure 6, Table 3). Such an analysis has not been previously performed for ferret CFTR because the full-length cDNA sequence has not been reported. In an effort to generate this data, our laboratory recently collaborated with the BACPAC Resource Center (http://bacpac.chori.org) at Children's Hospital Oakland Research Institute to generate a ferret BAC genomic library. Through the NIH sequencing project, the ferret CFTR genomic sequence is now available in the NCBI database. We used this information to predict the full-length CFTR cDNA and amino acid protein sequence for comparison with human, pig, and mouse CFTR sequences (Figure 6, Table 3). We find that ferret CFTR is 95% conserved with (91% identical to) the human CFTR amino acid sequence, a level of conservation strikingly similar to that between pig and human CFTR.
In summary, we have developed methods for generating ferret and pig polarized airway epithelia as the basis for comparative bioelectric experiments with previously developed human and mouse ALI models. One of the most important findings from this study is the considerable species-specific variability in the responses of the different epithelia to four known CFTR chloride channel blockers. Whether this diversity is due to altered structure/function characteristics of CFTR among these species or to the presence of other CFTR-like channels in the epithelia remains to be evaluated once CFTR knockout pigs and ferrets are available.
The authors gratefully recognize the editorial assistance of Dr. Christine Blaumueller.
This work was supported by NIH grants DK047967 (J.F.E.) and HL061234, and the Gene Therapy Core Center (DK54759).
Originally Published in Press as DOI: 10.1165/rcmb.2006-0286OC on September 28, 2006
Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript.