Airway defenses are regulated by a complex purinergic signaling network located on the epithelial surfaces, where ATP stimulates the clearance of mucin and pathogens. The present study shows that the obstructive disease cystic fibrosis (CF) affects the activity, expression, and tissue distribution of two ectonucleotidases found critical for the regulation of ATP on airway surfaces: NTPDase1 and NTPDase3. Functional polarities and mRNA expression levels were determined on primary cultures of human bronchial epithelial (HBE) cells from healthy donors and CF patients. The in vitro model of the disease was completed by exposing CF HBE cultures for 4 days to supernatant of the mucopurulent material (SMM) collected from the airways of CF patients. We report that NTPDase1 and NTPDase3 are coexpressed on HBE cultures, where they regulate physiological and excess nucleotide concentrations, respectively. In aseptic conditions, CF epithelia exhibit >50% lower NTPDase1 activity, protein, and mRNA levels than normal epithelia, whereas these parameters are threefold higher for NTPDase3. Exposure to SMM induced opposite polarity shifts of the two NTPDases on both normal and CF epithelia, apical NTPDase1 being mobilized to basolateral surfaces and bilateral NTPDase3 to the apical surface. Their immunolocalization in human tissue revealed that NTPDase1 is expressed in epithelial, inflammatory, and endothelial cells, whereas NTPDase3 is restricted to epithelial cells. Furthermore, the SMM-exposed CF HBE cultures reproduced the impact of the disease on their in vivo distribution. This study provides evidence that an extensive remodeling of the enzymatic network regulating clearance occurs in the airways of CF patients.
- mucociliary clearance
- P2Y2 receptors
inherited respiratory diseases are particularly complex because genetic mutations, pathogens, and inflammatory mediators interact to perpetuate the chronic nature of the pathogenesis. For instance, cystic fibrosis (CF) is diagnosed by mutations in the cystic fibrosis transmembrane regulator (CFTR) (4). The resulting failure to regulate Cl− secretion and Na+ absorption on epithelial surfaces causes airway dehydration by a reduction in the osmotic force driving water into the lumen. The depletion of the airway surface liquid (ASL) layer promotes the formation of viscoelastic mucus fostering the establishment of bacterial infection. Signaling events are triggered on both epithelial surfaces to clear the pathogens by mucociliary clearance (MCC), cell lysis, and phagocytosis (17). The inability of CF patients to clear infections is responsible for the chronic and damaging inflammation causing the loss of lung function (1). Microarray analysis of human bronchial epithelial (HBE) cultures revealed that the absence of functional CFTR modifies the expression of genes involved in airway defenses (60, 77), the list being considerably extended by chronic infection and inflammation (52). Hence, the airway complications emanate from complex interactions between the CF genetic defect and the recurrent episodes of infection and inflammation.
Airway clearance is supported by an intricate purinergic network concentrated on the epithelial barrier (14). The rhythmic mechanical stress caused by breathing induces ATP release from the epithelia (10), which binds P2Y2 receptors (P2Y2Rs) to induce fluid secretion via Ca2+-activated Cl− channels, and mucin secretion to shield the surface and entrap the pathogens. Surface enzymes, named ectonucleotidases, convert a fraction of the ATP into adenosine (ADO) (80), which binds A2B receptors (A2BRs) to stimulate cilia beating and CFTR activity (4). Purinergic signaling also regulates inflammatory responses, as ATP induces cytokine secretion from the epithelium (16, 64) and regulates the activities of inflammatory cells (32). Given the extensive ramifications of purinergic signaling in airway defenses, anomalies in the regulation of ATP, in CF, may contribute to the chronic airway obstruction and inflammation.
The mechanisms regulating ATP on airway epithelial surfaces were recently summarized in a mathematical model, which combines several ATP-hydrolyzing enzymes: nonspecific alkaline phosphatase (NSAP), nucleotide pyrophosphatase/phosphodiesterases (NPPs), and nucleoside triphosphate diphosphohydrolases (NTPDases) (80). These ectonucleotidases convert ATP into ADP and/or AMP; then, AMP is converted into ADO by NSAP and ecto 5′-nucleotidase (CD73). NTPDases have the unique ability to dephosphorylate nucleoside triphosphates (i.e., ATP, UTP) and nucleoside diphosphates (i.e., ADP, UDP) with distinct catalytic properties. For instance, NTPDase1 is considered a low-capacity, high-affinity ectonucleotidase [Km = 10–20 μM (11)], whereas NTPDase3 falls in the high-capacity, low-affinity category regulating higher nucleotide concentrations [Km = 90–130 μM (72)]. We recently reported a strong sensitivity of nucleotide metabolism to azide on HBE cultures (>40%) (8), supporting a major role for NTPDase1 (42) and/or NTPDase3 (72), but not the other surface NTPDases (26, 71), NSAP (30), or NPPs (69). Human NTPDase1 has been extensively investigated and assigned critical roles in the prevention of ischemia-reperfusion injury (31) and transplant rejection (2, 67), in thrombus formation, and inflammatory responses (6, 20). In contrast, few studies addressed human NTPDase3 (49, 73). In the respiratory system, NTPDase1 was detected at the mRNA level in mouse lung extract (33, 61) and immunolocalized on murine alveolar cells (21). The recent development of a selective inhibitor of human NTPDase3 (49) provided the necessary tool to distinguish the contributions of NTPDase1 and NTPDase3 to the regulation of the P2Y2R agonist in human airways.
In the present study, we combined in vitro and in vivo approaches to determine the impact of CF on NTPDase1 and NTPDase3. The contribution of chronic infection and inflammation was examined in vitro by exposing normal and CF HBE cultures to supernatant of the mucopurulent material (SMM) collected from the airways of CF patients (63). These data were validated in vivo by immunolocalization in airway tissue from healthy donors and CF patients. First, we demonstrate that the two NTPDases are coexpressed on human airway epithelia, where they regulate different concentration ranges of the P2Y2R agonists. Second, we provide evidence that CF causes an extensive remodeling of the enzymatic network regulating clearance. The aseptic HBE cultures revealed that the CF genetic defect causes a downregulation of NTPDase1 and an upregulation of NTPDase3. Chronic SMM exposure enhanced these effects and induced opposite polarity shifts of the two NTPDases. This systematic approach provides invaluable information on the respective contributions of inherited and acquired components to their expression and distribution in the airways of CF patients. Such distinct deregulation of NTPDase1 and NTPDase3 highlights the sophistication and plasticity of the purinergic network, reorganized to address local changes in the sources and concentrations of the signaling molecules.
MATERIALS AND METHODS
All nucleotides, sodium azide, KH2PO4, TBASH, and HEPES were purchased from Sigma-Aldrich (St. Louis, MO), and radiolabeled CTP for RNase protection assays was from Amersham Biosciences (Piscataway, NJ). Culture media and reagents were bought from Gibco-BRL Life Technologies (Rockville, MD). All other reagents were of analytical grade.
Airway tissue was obtained from healthy donors and CF patients during transplant under the auspices of protocols approved by the Institutional Committee on the Protection of the Rights of Human Subjects. The CF patients presented the most common CFTR mutation (ΔF508/ΔF508), which nearly obliterates the channel activity. The epithelial cells were isolated and cultured as previously described (28). In brief, the tissue was digested and the epithelial cells plated on T-clear filters coated with type IV collagen (Sigma-Aldrich). The cultures were grown in Ham's F-12-based medium supplemented with 0.5 ng/ml epithelial growth factor, 50 nM retinoic acid, 0.5 mg/ml bovine serum albumin, 0.8% bovine pituitary extract, 50 U/ml penicillin, and 50 μg/μl streptomycin (43). After 3 days, the confluent cultures were maintained at air-liquid interface. After 4 wk, they were composed of columnar-ciliated (>70%) and mucin-secreting cells over basal-like cells, with transepithelial electrical resistance ≥300 Ω·cm2.
Exposure protocol for SMM.
The SMM stock solution was prepared from mucopurulent material collected from the airways of five adult CF patients infected with Pseudomonas aeruginosa and Staphylococcus aureus, as we previously described (63). Normal and CF HBE cultures were exposed from 0 to 4 days to the vehicle control (20 μl PBS) or 20 μl SMM, added daily to the apical surface without wash to allow the development of inflammatory responses. The epithelial response was documented in terms of general morphology by hematoxylin and eosin (H&E) staining and quantification of mucus cell metaplasia using Alcian blue periodic acid-Schiff (AB-PAS) staining (45).
The NTPDase1 and NTPDase3 activities were measured on both surfaces of airway epithelial cultures in Krebs buffer (pH 7.4) at 37°C (5% CO2/95% O2) (56). They were preincubated either with the vehicle, an inhibitor of both NTPDases (1–20 mM azide), or 1–100 ng/ml monoclonal antibodies demonstrated to selectively inhibit the activity of human NTPDase3 (hN3-H10s) (49). All reactions were initiated with the substrate and stopped by periodically collecting aliquots to be processed for HPLC analysis (56). NTPDase1 activity was calculated from the azide-sensitive activity remaining after pretreatment with an optimal concentration of hN3-H10s.
Protocols for RT-PCR.
Total RNA was extracted from cultured and freshly excised human airway epithelial cells (9). Reverse transcriptase (RT) reactions were conducted with the Superscript III First Strand kit (Invitrogen, Carlsbad, CA), and specific antisense primers for NTPDase1 (5′-TGTGGACAATGGTTGCTCAGC-3′), NTPDase3 (5′-CAAGAGACCAGGCTATGCTGCTAT-3′), and internal standard β-actin (Ambion, Austin, TX), as we previously described (8). The cDNA was amplified using the Platinum PCR Supermix (Invitrogen) and specific primers for NTPDase1 (sense, 5′-AAGAGATTTGAGATGACTCTTCCATTC-3′; antisense, 5′-AGCCTTGCAGAAGGAGGGAGA-3′), NTPDase3 (sense, 5′-TGGCAGGAG AGAAGATGGATC-3′; antisense, 5′-AAGTGGTAGAGTTAGTTGGCTGAGAAG-3′), and β-actin (Ambion). The amplified cDNA products were separated on 1% agarose gel, purified using a kit (Qiagen, Valencia, CA) and tested by sequencing. The absence of genomic contamination was verified by reactions run without RT. The expected product sizes were 372, 608, and 294 bp for NTPDase1, NTPDase3, and β-actin.
RNase protection assays.
The assays were conducted as we previously described (56). In brief, the PCR products for NTPDase1, NTPDase3, or β-actin were ligated into pCR II plasmids (TA Cloning Kit, Invitrogen) and cloned in DH5α-T1 competent cells. The plasmids were purified and linearized, and the [32P]CTP-labeled antisense probes transcribed using the MAXIscript in vitro transcription kit (Ambion). The probes were purified by electrophoresis on 5% acrylamide/8 M urea gels (250 V, 1 h), eluted, and hybridized (8 × 104 disintegration/min) with 20 μg of total RNA (RPA III kit, Ambion). The protected RNA fragments were separated by electrophoresis (5% acrylamide/8 M urea gels; 250 V, 1 h), and the gels were dried and exposed to a phosphor screen (−20°C). After 4 days, the screen was scanned by Storm phosphoimager software (Molecular Dynamics, Sunnyvale, CA), and the signal was quantified by the software ImageQuant (Molecular Dynamics).
Taqman real-time PCR.
The cDNA was prepared by reverse transcriptase using the iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA). The cDNA was amplified with a 7500 Fast Real-Time PCR system (Applied Biosystems, Foster City, CA), and Taqman probes were designed by the manufacturer. The final reaction volume was 20 μl in the TaqMan Fast Universal PCR Master Mix, and the cycling conditions were: 20 s at 95°C (1 cycle), followed by 3 s at 95°C (40 cycles), then 30 s at 60°C. The internal standard 18S was used to normalize the expression levels.
SDS-PAGE was conducted on membrane proteins from HBE cultures under nonreducing conditions (70). In brief, the cells were scraped off the T-clears and ruptured by sonication (4°C) in Tris-NaCl buffer (pH 7.6). Membrane proteins were enriched by centrifugation at 2,000 g (15 min; 4°C) to remove organelles and debris, followed by centrifugation of the supernatant at 150,000 g (60 min; 4°C). The proteins were separated on 4–12% acrylamide-SDS gel gradients under nondenaturing conditions and transferred to an Immuno-Blot polyvinylidene difluoride membrane (Bio-Rad). The membranes were probed with monoclonal antibodies to human NTPDase1 (BU61; AnCell, Bayport, MN) (70) or NTPDase3 (hN3-B3s) (49), followed by HRP-conjugated goat anti-rabbit secondary antibodies (Pierce, Rockford, IL). The chemiluminescence was detected using the SuperSignal West Dura Extended Duration Substrate (Pierce Biotech).
The tissue distribution of NTPDase1 and NTPDase3 was established in human airways as previously described (49). In brief, 6-μm-thick sections of snap-frozen airway tissue from healthy donors or CF patients were fixed in an ice-cold solution of 95% acetone and 5% of 10% formalin (Fisher Scientific, Pittsburgh, PA) and then blocked in 0.1% PBS containing 7% goat serum. The sections were incubated with anti-NTPDase1 (BU61) or anti-NTPDase3 antibodies (hN3-H10s) (49) overnight, treated with 0.15% H2O2 in PBS to inactivate endogenous peroxidase, followed by avidin/biotin (Vector Labs, Burlington, Ontario, Canada) to prevent nonspecific staining. The sections were incubated with biotin-labeled goat anti-mouse secondary IgG (Jackson Immuno Research Laboratories, West Grove, PA) in the presence of the avidin-biotinylated HRP complex (VectaStain Elite ABC kit, Vector Laboratories) to optimize the reaction. Peroxidase activity was detected with 3,3′-diamino benzidine (Sigma-Aldrich).
All experiments were conducted on cultures or tissue samples from at least three different healthy subjects or CF patients. The cell surface ectonucleotidase activities were calculated from the initial linear rate of substrate decay determined by HPLC (nmol.min−1.cm−2). All data were expressed as means ± SE. Unpaired Student's t-tests were used to test the significance between measurements performed on different cultures. Paired t-tests were used to compare data obtained on the same culture before and after inhibitor treatments. Linear regressions, curve fits, and transformations were performed with SigmaPlot (Systat Software, Chicago, IL) or Prism (Graphpad Software, La Jolla, CA). Digital imaging was prepared with Photoshop CS3 (Adobe Systems).
Contribution of the azide-sensitive NTPDases.
We recently developed a mathematical model for the multi-enzyme network regulating extracellular nucleotides on airway epithelia (80). The simulations and experimental validations showed that the ectonucleotidases interact via substrate competition. Therefore, the following protocols were carefully designed to account for enzymatic interactions while establishing the contribution of the azide-sensitive NTPDases to the regulation of the P2Y2R agonists. The general properties of ASL nucleotide metabolism were determined on HBE cultures using 0.1 mM nucleotides. Both the triphosphates and diphosphates were metabolized (Fig. 1, A and B), as reported for purified NSAP and NTPDases (78). In an earlier study, we determined that NSAP accounts for ∼20% of the total surface activity of HBE cultures toward 0.1–1.0 mM ATP (56). Since the mathematical model showed that this concentration range is regulated by the high-capacity enzymes NTPDase3 and NSAP (80), the former is expected to account for most of the hydrolytic activity toward excess nucleotides on airway epithelia. Incidentally, the metabolic profile of ATP exhibits a transient accumulation of ADP (Fig. 1A) typical of NTPDase3 transfected in COS-7 cells (41). Also, the triphosphates were hydrolyzed three times more rapidly than the corresponding diphosphates (Fig. 1B), as reported for recombinant NTPDase3 (41). These data suggest that NTPDase3 regulates excess nucleotides, as locally released by cell lysis.
The identity of the NTPDases expressed on human airway epithelia was further addressed using azide. This agent was originally described as an inhibitor of an ATP-diphosphohydrolase (35) later identified as NTPDase1 (CD39) (33, 46). In the past decade, additional enzymes with high homology to NTPDase1 were sequenced (66), which required reassessment of azide specificity. Recent studies showed that azide inhibits NTPDase1 (42) and NTPDase3 (72), but not the other surface NTPDases (26, 71), NSAP (30) or NPPs (69). Therefore, this inhibitor would target specifically NTPDase1 and NTPDase3 on human airway epithelia.
The azide sensitivity of NTPDases is usually assayed by dose-response curves using 0–20 mM azide and a saturable substrate concentration (0.1–1.0 mM) (41). Therefore, HBE cultures were pretreated for 30 min with this azide concentration range and then assayed with 0.1 mM ATP. Figure 1C shows that azide reduced ATP metabolism in a dose-dependent manner, by up to 35%. Unexpectedly, azide generated the opposite effect on ADP metabolism, which increased twofold. This response was explained by substrate competition between dephosphorylating enzymes and the phosphorylating activity of ecto adenylate kinase (ecto-AK: ATP + AMP ↔ 2ADP) expressed on these cells (54). Analysis of the HPLC traces confirmed that ADP metabolism generated AMP, ADO, and ATP (Fig. 1D). The accumulation of ATP was accelerated by azide, but prevented by the ecto-AK inhibitor, diadenosine pentaphosphate (Ap5A) (54). Based on these findings, we repeated the dose-response curves for ADP in the presence of 0.5 mM Ap5A to isolate the NTPDase activities. Under these conditions, azide reduced the rate of ADP hydrolysis by up to 40% (Fig. 1C). Similar results were obtained for UTP and UDP metabolism, with maximal inhibitions of 46% and 38% (Fig. 1E). These values are within the range reported for purified NTPDase1 and NTPDase3 (45–60%) (53, 58), thereby supporting a major role for azide-sensitive NTPDases in the regulation of ASL nucleotides. The fact that UDP metabolism was not raised by azide is compatible with the substrate specificity of ecto-AK for adenine nucleotides (15, 54). Consequently, UTP was used to identify the azide-sensitive NTPDases expressed by human airway epithelia to avoid interference by ecto-AK.
Identification of the azide-sensitive NTPDases.
The identity of the azide-sensitive NTPDases expressed by airway epithelia was first determined by RT-PCR. These assays were conducted with total RNA from freshly excised and cultured nasal and bronchial epithelial cells. Figure 2A shows that both culture types coexpress NTPDase1 and NTPDase3. Since their expression is preserved in culture, the surface activities were investigated on HBE cultures. The activity of NTPDase3 was measured by dose-response curves using an antibody (hN3-H10s) that selectively inhibits this enzyme activity (49). The antibody reduced the metabolism of 1.0 mM UTP by up to 70% (Fig. 2B). Since the other high-capacity enzyme expressed by HBE cells, NSAP (80), accounts for ∼20% of the total activity toward excess nucleotides (56), these data suggest that 30 ng/ml hN3-H10s completely inhibits NTPDase3. This finding allowed us to circumvent the unavailability of specific inhibitors for NTPDase1. While the NTPDase inhibitor POM-1 was used to identify the protective role of NTPDase1 in ischemia-reperfusion injury (36), it was found to inhibit NTPDase1 and NTPDase3 with comparable efficiency (47). Therefore, the activity of NTPDase1 was calculated as the azide-sensitive activity remaining after complete inhibition of NTPDase3 with 30 ng/ml hN3-H10s. Following a 30-min treatment with the antibody, azide reduced the remaining activity by up to 28% (Fig. 2C), which corresponds to 15% of the total ectonucleotidase activity toward 0.1 mM UTP. These assays identify NTPDase1 as a minor regulator of excess nucleotides on human airway epithelia. The contribution of NTPDase1 and NTPDase3 to the regulation of physiological nucleotide concentrations was determined by repeating the assays with 1 μM UTP. The inhibition of NTPDase3 with 30 ng/ml hN3-H10s did not significantly affect UTP metabolism. In contrast, the azide-sensitive NTPDase1 activity remaining after NTPDase3 inhibition accounted for 40% of the total activity toward 1 μM UTP (Fig. 2D). Together, these data demonstrate that both azide-sensitive NTPDases are expressed on human airway epithelia. The fact that NTPDase1 and NTPDase3 target physiological and excess nucleotides, respectively, is consistent with their distinct substrate affinities, with Km values of 10–20 μM (11) and 90–130 μM (72, 80).
The distribution of the surface proteins involved in nucleotide-mediated MCC varies considerably along the airways. For instance, CFTR is restricted to ciliated cells (39), whereas different aquaporin isoforms are expressed in proximal and distal airways (38). Therefore, we hypothesized that the distribution of the two azide-sensitive NTPDases may also vary along the airways. First, we compared their mRNA levels in cultures of human nasal, bronchial, and bronchiolar epithelial cells. RNase protection assays indicated that all cultures express both NTPDases (Fig. 2E). Quantitative analysis of the hybridization gels revealed the existence of opposite expression gradients for NTPDase1 (nasal > bronchial > bronchiolar) and NTPDase3 (bronchiolar > bronchial > nasal) along the airways (Fig. 2F). We tested whether these patterns are reproduced functionally by comparing the activities of NTPDase1 and NTPDase3 on nasal and bronchial cultures. Since the total RNA originates from multi-layered epithelia, these assays were conducted on both surfaces with a saturable substrate concentration (1 mM) to obtain a quantitative measure of total surface activity. Figure 2G shows that NTPDase1 activity was detected on both nasal and bronchial epithelia, and was restricted >90% apically. The lower mRNA levels of the bronchial cultures (Fig. 2F) translated as 50% lower apical activities (Fig. 2G) compared with nasal epithelia. In contrast, NTPDase3 was significantly active on both surfaces of nasal and bronchial epithelia (Fig. 2G). The higher mRNA levels of the bronchial cultures resulted in >2-fold higher bilateral activities than on the nasal cultures. These assays also exposed local variations in the relative contributions of the two NTPDases to the regulation of excess nucleotides along the airways (Fig. 2G). While they share the responsibilities on nasal surfaces, bronchial epithelia rely mainly on NTPDase3. It is intriguing that the expression of high-capacity enzymes NSAP (57) and NTPDase3 (Fig. 2, E–G) increase toward alveoli, whereas high-affinity enzymes NTPDase1 (Fig. 2, E–G) and ecto 5′-NT (AMP → ADO) (57) share the opposite gradient. These data suggest that the composition of the multi-enzyme network is customized along airway surfaces to meet local requirements in the regulation of the signaling molecules.
The vast majority of the purinergic events are communicated by nucleotide concentrations within the 0.1- to 10.0-μM range. Therefore, the bilateral activity of the NTPDases on nasal and bronchial epithelia was also assessed with 1 μM UTP (Fig. 2H). On the apical surface, NTPDase1 accounted for nearly 90% of the azide-sensitive activity of nasal epithelia, whereas both enzymes shared responsibilities on bronchial epithelia. On basolateral surfaces, these enzymes conferred a relatively weak regulation of the P2Y2R agonist. These data are consistent with the fact that airway epithelia release nucleotides mainly into the ASL layer on the apical surface (14). The predominant role of NTPDase1 toward physiological nucleotide concentrations suggests that P2Y2R activation is regulated more stringently in the nose, whereas bronchi regulate a wider concentration range using both NTPDases.
Immunolocalization of NTPDase1 and NTPDase3 in healthy airways.
The in vivo distribution of NTPDase1 and NTPDase3 was established by immunolocalization in human airway tissue. A general survey at low magnification showed that NTPDase1 is expressed throughout the respiratory system, along superficial epithelia, in submucosal glands, fibroblasts, and alveolar cells (Fig. 3, A, C, E, G, I). In bronchial epithelia, NTPDase1 was detected intracellularly as a diffuse signal, whereas the infiltrated cells were heavily stained (Fig. 3B), consistent with its widespread expression in immune and inflammatory cells (23). In lower airways, NTPDase1 shifted to basolateral surfaces (Fig. 3, D–F). The submucosal glands presented weak staining for NTPDase1 compared with interstitial cells (Fig. 3H). Of particular interest is the alveolar region, which was heavily stained by the NTPDase1 antibodies (Fig. 3J), as reported in the murine lungs (21). This survey suggests that NTPDase1 is concentrated in epithelial barriers particularly involved in the regulation of leukocyte infiltration: distal airways and alveoli.
Contrary to NTPDase1, the airway distribution of NTPDase3 appears to be restricted to epithelial cells (Fig. 4, A, C, E, G, I). Higher magnification showed that the ectoenzyme exhibits a bilateral distribution throughout the airways (Fig. 4, B, D, F). On the apical surface, the antibodies labeled specifically mucin-secreting cells and mucin granules (Fig. 4D). NTPDase3 was also highly expressed in basal cells lining basolateral surfaces. The submucosal glands were heavily stained, both intracellularly and on luminal surfaces (Fig. 4H). In contrast, NTPDase3 was not detected in the alveolar region (Fig. 4J). In summary, NTPDase3 appears restricted to mucin-secreting cells in superficial epithelia and submucosal glands.
CF reorganizes the enzymatic network.
CF is an inherited disease aggravated by the establishment of recurrent bacterial infection (5). Microarray analysis of HBE cultures and cell lines showed that the lack of CFTR activity affects the expression of genes involved in airway defenses in the absence of infection (60, 77), a list considerably extended during chronic infection (52). Therefore, we hypothesized that the genetic defect and chronic infection both affect the regulation of purinergic signaling on airway surfaces. We reported previously that CF HBE cultures eliminate 1 mM nucleotides three times more rapidly than normal HBE cultures and that high-capacity NSAP accounts only for 30% of the enhanced activity (56). In the present study, we tested whether the remaining high-capacity enzyme, NTPDase3 (80), would be responsible for the accelerated elimination of the signaling molecules in CF. We first tested whether CF affects the activity of azide-sensitive NTPDases toward 1 mM UTP in HBE cultures. Figure 5A shows that the total azide-sensitive activity is twofold higher on both surfaces of CF epithelia. Interestingly, the discriminative assays showed that the CF genetic defect raises NTPDase3, but reduces NTPDase1, activity (Fig. 5B). These results suggest that NTPDase3 is largely responsible for the rapid elimination of excess nucleotides on CF airway epithelia. In these conditions, high-capacity NTPDase3 would mask the downregulating effect of CF on low-capacity NTPDase1.
The mechanism by which the CF genetic defect affects the expression of NTPDase1 and NTPDase3 was investigated at the protein and mRNA levels. Western blots conducted on normal and CF HBE cultures showed a single reactive band of 70–80 kDa (Fig. 5, C and D) for each enzyme, consistent with the size of their native form (49, 70). Gel band analysis supported a direct relationship between surface activity and protein expression, as CF reduced NTPDase1 expression by >50% and increased NTPDase3 expression fourfold (Fig. 5E). Similar results were obtained at the mRNA level by RNase protection assay. Quantitative analysis of the hybridization gels revealed a 50–70% reduction in NTPDase1 mRNA levels in both cultured and freshly excised CF bronchial epithelia (Fig. 5, F and G). In contrast, the mRNA levels of NTPDase3 were >4-fold higher in CF, both in vitro and ex vivo (Fig. 5G). These results, confirmed by Taqman PCR (data not shown), demonstrate that the CF genetic defect downregulates NTPDase1 and upregulates NTPDase3 in the absence of infection.
Chronic infection and inflammation are critical components of the CF airway disease, the ensuing damage and remodeling being responsible for the progressive loss of respiratory function (44). We tested the possibility that the bacterial products and inflammatory mediators accumulating in the airways of CF patients modify the regulation of the P2Y2R agonists by the azide-sensitive NTPDases. This hypothesis was tested by exposing normal and CF HBE cultures to SMM. Treatment efficiency was assessed by the development of mucus cell metaplasia, which appeared after 2 days and reached maximum expression after 4 days (Fig. 6A). Both normal and CF epithelia presented threefold higher numbers of mucus-secreting cells than their respective vehicle controls. SMM also affected the expression levels of both NTPDases in normal and CF epithelia, but with the same directionality as caused by the CF genetic defect. Taqman PCR shows that SMM downregulated NTPDase1 in normal epithelia, but did not further decrease the enzyme expression in CF epithelia (Fig. 6B). In contrast, the expression of NTPDase3 was raised in both normal and CF epithelia, with apparent additivity between the effects of the CF genetic defect and SMM. The global impact of the CF disease was determined by expressing the mRNA levels of SMM-exposed CF cultures relative to those of the vehicle-treated normal epithelia. Overall, the chronic airway disease downregulated NTPDase1 by 60% and upregulated NTPDase3 by threefold. These data are consistent with the RNase protection assays conducted on freshly excised normal and CF bronchial epithelia (Fig. 5, F and G). As such, the opposite changes in expression of the NTPDases caused by the lack of CFTR activity are expected to be aggravated in young CF patients by the onset of chronic infection and inflammation.
The functional consequences of aberrant expressions of the NTPDases in the CF airway disease were investigated by testing the impact of SMM on their surface activity and polarity on normal and CF HBE cultures. Bilateral assays conducted with 1.0 mM UTP revealed that airway epithelia respond to chronic infection and inflammation by an extensive reorganization of the enzymatic network on both surfaces. First, SMM reduced the total bilateral activity of NTPDase1 in normal cultures (Fig. 6D), in agreement with the expression data (Fig. 6B). In contrast, SMM also reduced NTPDase1 activity on CF epithelia, which did not respond by changes in mRNA level (Fig. 6B). This finding raises the possibility that components of the SMM may act directly on the surface protein, as reported for TNFα on vascular NTPDase1 (65). On the other hand, the impact of SMM on NTPDase3 activity (Fig. 6E) was consistent with the additive enhancing effects of the CF genetic defect and chronic infection on the mRNA levels (Fig. 6B). The enzymes also responded to SMM by opposite polarity shifts, as NTPDase1 was relocated from the apical to basolateral surfaces (Fig. 6D), and bilateral NTPDase3 was concentrated on the apical surface (Fig. 6E). Collectively, these data suggest that airway epithelia respond to chronic infection and inflammation by enhancing their capacity to absorb excess nucleotides in the lumen while promoting fine-tuning of purinergic signals in the interstitium, such as nucleotide-mediated leukocyte infiltration (6, 20). In CF, the downregulation of NTPDase1 may impair the regulation of critical inflammatory responses.
Impact of CF on the tissue distribution of NTPDases.
The in vitro model of the CF disease was tested against the tissue distribution of the NTPDases in the airways of CF patients. All tissue samples exhibited late-stage characteristics of the disease (27). The airways were sporadically obstructed by mucopurulent material and lined by epithelia presenting mucus cell metaplasia and hyperplasia (Fig. 7, A and B). The interstitial milieu contained neutrophil aggregates and developed tissue fibrosis. Submucosal glands developed hyperplasia and occlusion of the gland ducts (Fig. 7, G and H). In the small airways, mucosal edema caused narrowing of the lumen (Fig. 7, C and D).
CF dramatically raised the expression of NTPDase1 throughout the respiratory system, except for the epithelial barrier. The antibodies heavily stained inflammatory cells recruited in the airways, infiltrating the epithelial barrier and the interstitial and alveolar compartments (Fig. 7, A–J). This distribution is consistent with the widespread expression of NTPDase1 in immune and inflammatory cells, including neutrophils and dendritic cells (6, 20). In bronchial epithelia, the diffuse apical signal (Fig. 3F) was lost in CF (Fig. 7F), whereas basolateral membranes were more heavily stained. In contrast, NTPDase3 remained restricted to epithelial cells in CF (Fig. 8, A–J) and was upregulated especially in submucosal glands (Fig. 8H) compared with healthy airways (Fig. 4H). Also, the basolateral or bilateral polarity of NTPDase3 observed in normal epithelia (Fig. 4, B, D, F) was shifted predominantly to apical surfaces (Fig. 8, B, D, F). This immunological survey validates our in vitro model of the CF airway disease in terms of gene expression and epithelial polarity of the NTPDases.
Cell surface NTPDases have been exhaustively investigated in the vasculature (2, 67), and their critical role in the regulation of immune and inflammatory cells is emerging (6, 20). In most instances, NTPDase1 was identified as the main enzyme regulating ATP-mediated responses. The present study demonstrates, for the first time, that two NTPDases are functionally coexpressed on human airway epithelia, where they provide nonredundant regulation of the P2Y2R agonists. We identified NTPDase1 and NTPDase3 in primary cultures of airway epithelial cells at the mRNA and protein levels, and functionally by discriminative assays. Measurements of their surface activities over a wide range of nucleotide concentrations suggest that NTPDase1 predominates (40%) under physiological conditions (<1 μM ATP), whereas NTPDase3 predominates (70%) in the presence of excess nucleotides, as generated by cell lysis. These findings are consistent with their biochemical properties, as NTPDase1 was characterized as a low-capacity, high-affinity enzyme (11) and NTPDase3 as a high-capacity, low-affinity enzyme (72). Their contributions to the regulation of P2Y2R agonists are expected to vary along the airways and over time with the nucleotide concentrations.
Distinct purinergic microenvironments along the airways.
An immunological survey conducted in human airway tissue provided spatial information on the NTPDases. We show that the two NTPDases are coexpressed on the epithelial barrier lining the entire respiratory system, except for NTPDase3 in the alveolar region. Interestingly, their distribution varied in terms of expression level and polarity along the airways, thus creating enzymatic microenvironments. In the tracheo-bronchial area, the pseudostratified epithelium presented little overlap between the enzymes, as NTPDase1 was detected apically as a diffuse intracellular signal, and NTPDase3 was restricted to basolateral surfaces and basal cells. In distal airways, the NTPDases were coexpressed on basolateral surfaces, and NTPDase3 was also detected apically. Such polarity shifts along the airways were reported for other proteins, including the water channel aquaporin-3 (38). These NTPDases also followed opposite expression gradients on the apical surface, as reported for NSAP and ecto 5′-NT (57). As a result, the high-affinity enzymes NTPDase1 and ecto 5′-NT (57) are concentrated in the upper airways, and the low-affinity enzymes NTPDase3 and NSAP (57) in the distal airways. These data suggest that the composition of the enzymatic network is adjusted along the airways to follow local requirements in ASL nucleotide regulation and/or specific epithelial cell types. For instance, NTPDase3 was immunolocalized preferentially on mucin-secreting cells, which are more abundant in distal airways (13). Human airway epithelial cells were recently shown to package nucleotides into mucin granules, which are released into the ASL by P2Y2R activation (40). High-capacity NTPDase3 may be mobilized to regulate the amplification cycles of P2Y2R activation resulting from ATP-induced mucin and nucleotide secretion (14). Hence, the distinct expression gradients of NTPDase1 and NTPDase3 along the airways suggest that they regulate different defense mechanisms.
CF deregulates the NTPDases.
The vast majority of the in vitro studies on airway clearance in CF are derived from aseptic primary cultures of HBE cells from CF patients (4, 5). Using this model, we previously reported that the metabolism of P2Y2R agonists is accelerated on CF airway epithelia (56). However, chronic infection and inflammation are considered major determinants in the severity of lung complications in CF (7, 34). In the present study, a systematic approach was developed to examine both aspects of the CF airway disease: the genetic defect and the recurrent cycles of chronic infection and inflammation. First, we tested the impact of the disease on NTPDase1 and NTPDase3 using HBE cultures grown under aseptic conditions, allowing sufficient time to abolish any residual response to the infection and inflammation experienced in vivo (63). These cultures provided evidence that the genetic defect profoundly modifies the regulation of the P2Y2R agonists on airway epithelial surfaces. The elimination rate of excess ASL nucleotides was raised >3-fold in CF, most of which was caused by an upregulation of high-capacity NTPDase3. In contrast, high-affinity NTPDase1 responded by a 50% decrease in activity and expression. This metabolic imbalance may result from modifications in the ASL properties, as the lack of HCO3− secretion by CFTR has been shown to acidify airway surfaces (12). The fact that purified NTPDase1, not NTPDase3, is sensitive to the acidic pH range measured in the ASL of CF patients (pH 6.0–6.5) (12, 41) may contribute to its reduced activity. The lack of CFTR activity was also proposed to maintain the ASL in a state of oxidative stress in the absence of infection (68). Several studies reported that oxidative stress reduces the activity and expression of NTPDase1. In the kidney, NTPDase1 was inhibited by reactive oxygen species (59) or γ-irradiation (76). Oxidative stress-mediated tissue injury also plays a pivotal role in delayed graft function. Allograft neuropathy is associated with a downregulation of NTPDase1, and an upregulation of ecto 5′-NT, in tissue biopsies (48). These data are consistent with the low NTPDase1 and high ecto 5′-NT (55) activities of aseptic CF HBE cultures. Since NTPDase3 does not regulate P2Y2R agonists in the physiological concentration range, chronically low NTPDase1 activity in CF may allow ASL nucleotide levels to rise and promote MCC (4, 5). However, mucin hypersecretion is considered an early event precipitating the course of the disease, as the airways fail to clear the dehydrated and adhesive mucin (5). Since P2Y2R activation is a powerful inducer of mucin secretion (4), high ASL ATP levels may predispose CF patients to the development of chronic airway obstruction and infection.
In adult CF patients, the recurrent cycles of infection and inflammatory responses concentrate myriad mediators into the airways (5). This aspect of the disease was examined by exposing normal and CF HBE cultures to SMM collected from the airways of CF patients (63). After 4 days, both culture types developed typical mucus cell metaplasia (44). They also responded to SMM by reducing NTPDase1, and raising NTPDase3, activity and/or expression. In a recent study, microarray analysis indicated that 24-h exposure to SMM affects the expression of several genes in HBE cells, but not the enzymes regulating ASL nucleotides: NTPDase1, NTPDase3, NSAP, and ecto 5′-NT (62). However, our chronic exposure conditions allowed sufficient time for the development of mucus cell metaplasia, which may be required to affect the expression of the NTPDases. For instance, immunohistochemistry showed that NTPDase3 is preferentially expressed on mucin-secreting epithelial cells, which increase in number during mucus cell metaplasia. Since SMM amplifies the effects of the CF genetic defect on the NTPDases, the onset of chronic infection and inflammation in young patients may exacerbate the aberrant regulation of P2Y2R agonists on airway surfaces.
Numerous studies support a role for the hypoxic conditions maintained in the obstructed airways of CF patients in the regulation of NTPDases. For instance, the endothelial barrier responds to hypoxia by a coordinated upregulation of NTPDase1 and ecto 5′-nucleotidase (25) to optimize the production of ADO and reduce vascular leakage (24). The upregulation of ecto 5′-nucleotidase during hypoxia was found to be mediated by the transcription factor, hypoxia-inducible factor-1 (HIF-1) (74). In addition, Eltzschig and collaborators (25) recently showed that Sp1 plays a critical role in the hypoxia-induced upregulation of NTPDase1. Interestingly, this transcription factor is upregulated in HBE cultures infected with P. aeruginosa (3), where it plays a major role in the initiation of inflammatory responses through an upregulation of proinflammatory cytokines (3) and Toll-like receptor-2 (TLR2) (29). We also observed by real-time PCR that hypoxia raises the expression of both NTPDase1 and NTPDase3 in HBE cultures (M. Picher, personal communication). The fact that CF is associated with a reduction in the expression of NTPDase1 on airway epithelial surfaces suggests that other factors, such as oxidative stress, may drive the deregulation of this enzyme.
Chronic infection mobilizes the NTPDases to distinct areas.
The most striking outcome of chronic infection and inflammation was the opposite polarity shifts of NTPDase1 and NTPDase3 observed on normal and CF HBE cultures, which closely reproduced the impact of the disease in vivo. On the HBE cultures, SMM triggered the mobilization of apical NTPDase1 to basolateral membranes, whereas bilateral NTPDase3 became primarily located on the apical membrane. This type of response does not appear to be limited to ectonucleotidases, as a similar polarity shift was reported for hepatocyte growth factor activator inhibitor-1 on the HBE cultures (75). Like NTPDase3, this proteinase inhibitor reacted to inflammatory mediators by a polarity shift in the basolateral-to-apical direction. Intrahepatic cholestasis, induced in the rat by LPS, modified the polarity of NTPDase and NSAP activities in liver epithelia within 2 h, which persisted >7 days (79). LPS relocated the canalicular NTPDase activity to basolateral surfaces, which was recently identified as high-affinity NTPDase8 (26). In contrast, LPS caused an opposite polarity shift of NSAP activity (79). Hence, LPS and SMM both recruit high-affinity enzymes (NTPDase8 and NTPDase1) to basolateral membranes and high-capacity enzymes (NSAP and NTPDase3) to apical membranes. These studies suggest that the kinetic properties of ectonucleotidases influence their polarity under stress conditions.
Little is known of the mechanisms directing NTPDases to apical or basolateral membranes in polarized epithelia. In airway epithelia, lipid rafts constitute sorting platforms for apical membrane proteins, including CFTR (18). Koziak et al. (37) showed that NTPDase1 undergoes a posttranslational modification that involves the palmitoylation of a cysteine group (Cys13) found essential for protein insertion into lipid rafts. This mechanism could account for the apical location of NTPDase1 in proximal airways and the diffuse intracellular signal, as proteins sorted via this route are recycled by the activities of palmitoyl transferases and palmitoyl-protein thioesterases (19). Other factors likely influence the polarity of NTPDases since the sequence of NTPDase3 also contains this cysteine residue. Infection of intestinal epithelia by Escherichia coli allows basolateral proteins to reach the apical surface by disruption of the tight junctions (50, 51). Consequently, the polarity shifts of NTPDase1 and NTPDase3 in infected CF airways may result from interference with cargo delivery and/or barrier disruption.
The present work reveals the complexity and adaptability of the enzymatic network regulating airway clearance. The information acquired from carefully designed in vitro assays and in vivo immunological surveys demonstrate that NTPDase1 and NTPDase3 work in concert to regulate the availability of P2Y2R agonists. Their distinct biochemical properties, epithelial polarities, and expression gradients along the airways alleviate the redundancy anticipated for the coexpression of two closely related enzymes. Also, the fact that normal and CF epithelia responded similarly to SMM raises the possibility that other airway diseases may affect the regulation of nucleotide-mediated airway defenses. Such extensive remodeling of the activities and polarities of the NTPDases during chronic infection and inflammation may constitute an attempt to strengthen airway defenses. On apical surfaces, the mobilization of high-capacity NTPDase3 may improve the regulation of excess ATP accumulating in the mucopurulent material from bacterial lysis and damaged epithelia. On basolateral surfaces, high-affinity NTPDase1 may regulate inflammatory responses, such as neutrophil recruitment. The recent finding that A2BRs preserve epithelial barrier integrity in the lung (22) suggests that the colocalization of NTPDase1 and ecto 5′-nucleotidase (57) on basolateral surfaces may contribute to this defense mechanism by locally enhancing ADO availability. Future studies will identify the epithelial functions regulated by these NTPDases and the consequences of their altered regulation for CF.
This work was supported by Cystic Fibrosis Foundation Grants PICHER07I0 (M. Picher) and RIBEIRO3FG0 (C. M. Ribeiro); National Heart, Lung, and Blood Institute Grant P01-HL-034322 (M. Picher); and Canadian Institutes of Health Research Grants MOP-49460 and MOP-93683 (J. Sévigny). M. Fausther was a recipient of a scholarship from the government of Gabon. J. Sévigny was a recipient of a New Investigator Award (CIHR) and a scholarship from the Fonds de la Recherche en Santé du Québec.
No conflicts of interest are declared by the author(s).
We thank Dr. Scott Randell (UNC CF Center Tissue Core, Director) for providing the airway epithelial cells and airway tissue, and Kimberly Burns and Tracy Eldred (UNC CF Center Histology Core) for the preparation of the frozen tissue blocks.
- Copyright © 2010 the American Physiological Society