Oxidative stress is implicated in the pathogenesis of many inflammatory pulmonary diseases, including cystic fibrosis (CF). Delineating how oxidative stress stimulates CF transmembrane conductance regulator (CFTR) in airway epithelial cells is useful, both to increase the understanding of airways host defense and suggest therapeutic approaches to reduce the oxidant stress burden in the CF lung. Using the airway epithelial cell line Calu-3, we investigated the hypothesis that hydrogen peroxide (H2O2), which stimulates anion efflux through CFTR, does so via the production of prostaglandin E2 (PGE2). Using iodide efflux as a biochemical marker of CFTR activity and short circuit current (Isc) recordings, we found that the H2O2-stimulated efflux was abolished by cyclooxygenase-1 inhibition and potentially also involves microsomal prostaglandin E synthase-1 activity, implicating a role for PGE2 production. Furthermore, H2O2 application resulted in a rapid release of PGE2 from Calu-3 cells. We additionally hypothesized that the PGE2 subtype 4 (EP4) receptor was involved in mediating this response. In the presence of (4Z)-7-[(rel-1S,2S,5R)-5-((1,1′-biphenyl-4-yl)methoxy)-2-(4-morpholinyl)-3-oxocyclopentyl]-4-heptenoic acid (AH23848) (which blocks the EP4 receptor), the H2O2-stimulated response was abolished. To investigate this finding in a polarized system, we measured the increase in Isc induced by H2O2 addition in the presence and absence of AH23848. H2O2 induced a robust increase in Isc, which was significantly attenuated in the presence of AH23848, suggesting some role for the EP4 receptor. In conclusion, with H2O2 as a model oxidant stress, stimulation of CFTR seems to involve PGE2 production and likely EP4 receptor activation in Calu-3 airway epithelial cells. This mechanism would be compromised in the CF airways.
Oxidative stress has been implicated in the pathogenesis and progression of inflammatory pulmonary diseases, including cystic fibrosis (CF) (Cantin et al., 2006). CF arises as a result of dysfunctions in the CF transmembrane conductance regulator (CFTR) ion channel, whose normal role is to mediate anion movement across the apical membrane of epithelial cells throughout the body (Sheppard and Welsh, 1999). At present, the major morbidity and mortality associated with CF results from pulmonary infections and seems to be related to a failure of the mucociliary clearance system to effectively remove opportunistic bacterial pathogens, which are able to colonize the airways for extended periods (Boucher, 2007). Thus, there is an increased burden of oxidant stress in the CF airways, as activated neutrophils recruited to the airways in response to the presence of bacteria release reactive oxygen species (ROS) in an attempt to kill bacteria (Konstan and Berger, 1993). ROS are unstable compounds capable of initiating tissue damage via their actions on cellular lipids, proteins, and DNA. Normally, extensive antioxidant defense mechanisms exist in the airways that neutralize ROS before they are capable of damaging cellular components (Rahman et al., 2006). However, in the CF lung, antioxidant defenses seem to be significantly compromised because of reduced systemic antioxidant levels, arising from decreased absorption of fat-soluble antioxidant vitamins from the diet (Farrell et al., 1977), as well as decreased glutathione transport via the CFTR (Linsdell and Hanrahan, 1998). This, together with the increased exposure to ROS from activated neutrophils (Chmiel et al., 2002), causes the equilibrium between oxidant stress and antioxidant defenses to become unbalanced, and tissue damage can result.
We and others have previously reported that CFTR activity is increased by exposure to acute oxidant stress (Nguyen and Canada, 1994; Cowley and Linsdell, 2002). For example, exposure to hydrogen peroxide (H2O2) results in a rapid increase in transepithelial anion secretion across monolayers of the model human airway submucosal gland serous cell line Calu-3, a cell type widely used to investigate wild-type CFTR activity. This increase in secretion is mediated via coordinated effects on both basolateral potassium channel activity and apical CFTR activity (Cowley and Linsdell, 2002). Rapid CFTR activation is also seen in the presence of the Pseudomonas aeruginosa virulence factor pyocyanin, likely via increased H2O2 production (Schwarzer et al., 2008). It is noteworthy, however, that this initial stimulation by pyocyanin later gives way to an inhibition of CFTR activity, which is in line with other reports that longer-term exposure to oxidant stress can actually decrease CFTR activity and expression (Cantin et al., 2006; Schwarzer et al., 2008; Qu et al., 2009). Thus, there is a complex physiological response to oxidant stress, consisting of an acute stimulatory effect, followed by a suppression of CFTR activity.
One mechanism by which H2O2 can reportedly acutely stimulate CFTR is via cyclooxygenase (COX)-dependent production of prostaglandin E2 (PGE2) (Soodvilai et al., 2007). PGE2 mediates its effects via activation of four distinct PGE2 receptors, termed EP1-4, which are coupled to a variety of signaling mechanisms including alterations in intracellular cAMP and Ca2+ levels (Matsuoka and Narumiya, 2007) and have all been identified in Calu-3 cells (Joy and Cowley, 2008). Because CFTR is regulated by protein kinase A activity (Gadsby et al., 2006), activation of cell surface receptors coupled to cAMP generation will increase CFTR activity and anion efflux from the cell. We have previously reported that only the PGE2 subtype 4 (EP4) receptor is capable of eliciting a CFTR-mediated anion efflux from Calu-3 cells, despite the presence of other EP prostanoid receptor subtypes on Calu-3 cells (Joy and Cowley, 2008).
Another potential way that oxidant stressors such as H2O2 could stimulate CFTR is via the generation of a class of prostaglandin-like molecules termed isoprostanes, which also predominantly mediate their effects via prostanoid receptors (Janssen, 2001). Isoprostanes are generated when ROS react with the unsaturated bonds of polyunsaturated fatty acids in the cell membrane (Jahn et al., 2008). One example of such an isoprostane, 8-iso-PGE2, is capable of stimulating CFTR-mediated transepithelial anion secretion across Calu-3 cells (Cowley, 2003). We previously proposed that this isoprostane produced its secretory response via action at the EP4 receptor (Joy and Cowley, 2008). Thus, there is evidence that the EP4 receptor plays an important role in mediating the effects of oxidant stress.
In the present study, we wanted to investigate whether the H2O2-stimulated increase in CFTR-mediated anion efflux from Calu-3 cells might be mediated via activation of the EP4 receptor via direct PGE2 production. To do this, we have predominantly used a biochemical marker of CFTR activity, the iodide efflux assay, and investigated pharmacologically aspects of this proposed pathway. When appropriate, our findings are supported by measurements of short circuit current (Isc) recordings across polarized monolayers of Calu-3 cells. Our hypothesis is that activation of the EP4 receptor, whether that is via isoprostane generation or the generation of PGE2 in response to oxidant stress, plays an important role in the CFTR-mediated response to oxidant stress, which would be compromised in the CF airways.
Materials and Methods
Calu-3 cells (American Type Culture Collection, Manassas, VA) were cultured in 1:1 Dulbecco's modified Eagle's medium/Ham's F-12 nutrient mixture supplemented with 15% fetal bovine serum (FBS Gold; PAA Laboratories, Inc., Etobicoke, ON, Canada), 1% penicillin, and 1% streptomycin (Invitrogen, Burlington, ON, Canada). Cells were incubated at 37°C in humidified 5% CO2/95% air. For RNA extraction and iodide efflux studies, cells were cultured on 35-mm-diameter culture dishes (BD Biosciences, San Jose, CA). For Isc recordings, cells were plated on Snapwell inserts (Corning Life Sciences, Lowell, MA) and maintained at an air-liquid interface with medium present only on the basolateral side as described previously (Cowley and Linsdell, 2002).
RNA Extraction, RT-PCR, and Quantitative PCR.
Total RNA was extracted by using TRIzol reagent (Invitrogen), and 2 μg of RNA was reversed-transcribed to produce cDNA by using Moloney murine leukemia virus reverse transcriptase (Invitrogen) in the presence of 5 mM deoxynucleoside-5′-triphosphate (Invitrogen) and 1 μM oligo(dT) (Promega, Madison, WI).
RT-PCR was performed to confirm the expression of COX-1, COX-2, and microsomal prostaglandin E synthase-1 (mPGES-1) in Calu-3 cells by using the primers described in Table 1. All primers were purchased from Invitrogen and used at a final concentration of 400 nM. PCR experiments were performed in the presence of 5 mM deoxynucleoside-5′-triphosphate (Invitrogen), 25 mM MgCl2, 10× Taq buffer with KCl, and 2.5 U Taq polymerase (all from Fermentas, Burlington, ON, Canada) in a total reaction volume of 25 μl. PCR products were visualized on a 1.5% agarose gel containing 250 μg · l−1 ethidium bromide, alongside a 100-base pair ladder (Fermentas). Each PCR was performed at least three times on different passages of Calu-3 cells.
Quantitative PCR (qPCR) was used to investigate changes in mPGES-1 gene expression after exposure to 4-(benzo[b]thiophen-2-yl)-3-bromo-5-hydroxydihydrofuran-2(3H)-one (Cay10526) (5 μM for 24 h). RNA was extracted from three different passages of cells, and reactions were performed in quadruplicate. Changes were detected by normalizing to the amount of transcript of the housekeeping gene hypoxanthine guanine phosphoribosyltransferase (HPRT) and expressed as a percentage of control. Primers used are given in Table 1 with conditions as follows: 10 min at 95°C, followed by 45 cycles of 95°C for 10 s, 60°C (for HPRT) or 52°C (for mPEGS-1) for 5 s, and 72°C for 10 s. qPCR was performed by using a Lightcycler thermal cycler system (Roche Applied Science, Laval, PQ, Canada). Data were subjected to a Student's t test with statistical significance reported at P < 0.05.
Calu-3 cells were cultured to confluence on 35-mm-diameter plates. Cells were loaded with iodide by incubation with 2 ml of the iodide loading buffer (136 mM NaI, 3 mM KNO3, 2 mM Ca(NO3)2, 11 mM glucose, and 20 mM HEPES, pH 7.4) for 1 h at room temperature. The loading buffer was then removed by rapidly washing the cells three times with iodide efflux buffer (136 mM NaNO3, 3 mM KNO3, 2 mM Ca(NO3)2, 11 mM glucose, and 20 mM HEPES, pH 7.4). Samples were then collected by repeatedly replacing the efflux buffer with fresh solution every 1 min, and the iodide concentration was determined by using an iodide-sensitive electrode (Orion Research Inc., Boston, MA). Generally, pharmacological agents under investigation were added to the efflux buffer at 3 min, and 1.5-ml samples were then collected every minute for another 12 min in the presence of the agent under investigation. In addition, some inhibitors required a 1-h incubation period, in which case they were added to the loading buffer. For each experiment, a negative control was run that had no agonist added and represented basal iodide efflux, while the cAMP-elevating agent forskolin (7β-acetoxy-8,13-epoxy-1α,6β,9α-trihydroxylabd-14-en-11-one) was added as a positive control to test the viability of the cells. Stimulated effluxes were measured in triplicate or quadruplicate per experiment, for a total of at least three experiments.
Data are represented both as examples of individual iodide efflux experiments (mean ± S.E.M of the duplicate or quadruplicate samples run during that individual experiment) and as the cumulative percentage of total iodide efflux released by the cells over the course of the experiment in which the percentage of iodide efflux was calculated from the cumulative iodide efflux up to time point X divided by the total amount of iodide effluxed from the cells over the course of the experiment (mean ± S.E.M of at least three different experiments). Finally, the absolute iodide efflux rate was calculated at the time of peak efflux (2–4 min after application of the agonist) by calculating the slope of percentage of iodide efflux/time (Joy and Cowley, 2008). Data were subjected to analysis of variance (ANOVA) followed by a Bonferroni correction to adjust for multiple comparisons.
Measurement of Transepithelial Isc.
Calu-3 cell monolayers were mounted in an Ussing chamber (World Precision Instruments, Inc., Sarasota, FL), and the transepithelial potential difference was clamped to zero by using a DVC-1000 voltage-clamp apparatus (World Precision Instruments, Inc.). The transepithelial Isc was recorded by using Ag-AgCl electrodes in agar bridges and reflects the net movement of ions across the epithelial monolayer. Apical and basolateral solutions were maintained at 37°C by heated water jackets and separately perfused and oxygenated with a 95% O2/5% CO2 mixture. Bath solutions for intact monolayers were 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 (basolateral) or mannitol (apical), pH 7.4 at 37°C, when gassed with 95% O2/5% CO2.
After application of H2O2 (1 mM) to the apical face of the monolayers, the peak Isc was recorded. Data reported reflect the measured difference in Isc between the baseline and the peak Isc (ΔIsc).
Measurement of PGE2 by ELISA.
Cells were grown to confluence in 12-well plates. Cells were exposed to H2O2 (1 mM) for either 5, 10, or 30 min in serum-free media. The media were then collected, cleared by centrifugation at 2000g, and stored at −80°C before measurement of PGE2 levels using an ELISA kit (Cayman Chemical, Ann Arbor, MI), according to the manufacturer's instructions, and with a sensitivity of 30 pg/ml. In all cases, triplicate samples were collected on three separate occasions (n = 3). The amount of protein in the treated cells was normalized to that in the controls, and the normalized data were subjected to ANOVA followed by a Bonferroni correction. Values are reported as mean ± S.E.M. with statistical significance reported at P < 0.0125 or P < 0.016 as appropriate.
H2O2-Stimulated Efflux Is Abolished by COX-1, but Not COX-2, Inhibition.
Application of H2O2 at a concentration previously determined to produce CFTR-mediated transepithelial anion secretion across Calu-3 cells (1 mM; Cowley and Linsdell, 2002) likewise induced a strong stimulation of iodide efflux from the nonpolarized cells. Figure 1A shows a representative example of the increased iodide efflux seen in response to H2O2 applied in the fourth minute of the experiment. This increase was completely abolished when cells were preincubated for 1 h with the COX inhibitor indomethacin [1-(4-chlorobenzoyl)-5-methoxy-2-methyl-1H-indole-3-acetic acid] (10 μM; Fig. 1, B, C, and H), implicating the production of prostaglandins in the response. Because indomethacin nonselectively inhibits both COX-1 and COX-2, we next used the reportedly more selective inhibitors 5-(4-chlorophenyl)-1-(4-methoxyphenyl)-3-(trifluoromethyl)-1H-pyrazole (SC560) and 3-(4-methylsulphonylphenyl)-4-phenyl-5-trifluoromethylisoxazole (Cay10404) to examine the potential contributions of these enzyme isoforms. Incubation with the COX-1 inhibitor SC560 (0.5 μM) effectively abolished the H2O2-stimulated efflux (Fig. 1, D, E, and H), whereas the COX-2 inhibitor Cay10404 (1 μM) had no effect (Fig. 1, F-H). When the efflux rates are compared (Fig. 1H), it can clearly be seen that indomethacin and SC560 significantly reduce the H2O2-stimulated response to a rate comparable with basal efflux, whereas Cay10404 has no effect.
In addition, every time an iodide efflux experiment was performed the rate of basal efflux was obtained by collecting samples from nonstimulated cells (n = 66), as well as the rate of CFTR-mediated efflux in response to the cAMP-elevating agent forskolin (at 10 μM) (n = 72), to confirm the integrity of our cells. These rates are represented in Fig. 1H. Finally, to confirm our previous observation that the H2O2 response is mediated via the CFTR (Cowley and Linsdell, 2002), we performed experiments in the presence of the CFTR inhibitor N- (naphthalen-2-ylamino)-acetic acid(3,5-dibromo-2,4,-dihydroxy-benzylidene)-hydrazide (GlyH101) (20 μM; n = 8), which effectively abolished the H2O2 stimulation of efflux (Fig. 1H).
To confirm these findings, we recorded the increase in Isc across polarized monolayers of Calu-3 cells after apical application of H2O2 (1 mM), because the H2O2-stimulated increase in Isc is larger when the H2O2 is applied apically (Cowley and Linsdell, 2002). Application of H2O2 induces a robust increase in the Isc (Fig. 2A) (ΔIsc = 67.8 ± 6.8 μA · cm−2; n = 4), which was significantly reduced after preincubation with 0.5 μM SC560 (ΔIsc = 6.6 ± 2.3 μA · cm−2; n = 3), but unaffected by preincubation with Cay10404 (ΔIsc = 72.7 ± 21.1 μA · cm−2; n = 3; Fig. 2, B-D). These results suggest a role for COX-1, rather than COX-2, in the response.
Microsomal Prostaglandin E Synthase-1 Mediates PG Production in Calu-3 Cells.
The attenuation of the H2O2-mediated response seen in the presence of indomethacin and SC560 implicates a role for COX in the generation of some biological molecule involved in mediating the oxidant stress response. COX mediates the conversion of arachidonic acid to PGH2, the precursor of a plethora of biologically active molecules including prostacyclin, thromboxanes, and prostaglandins. Of these arachidonic acid derivatives, probably the most studied has been PGE2. The report of Soodvilai et al. (2007) concluded that H2O2 stimulated CFTR in renal cells via the generation of PGE2 through both COX and mPGES-1. Therefore, we were interested in seeing whether a similar mechanism might be present in Calu-3 cells. In addition to COX, PGE2 can be produced via the activity of the enzyme mPGES-1, which converts PGH2 to PGE2. mPGES-1 is increasingly being investigated as a potential anti-inflammatory agent (Iyer et al., 2009). Therefore, we wanted to investigate whether mPGES-1 was involved in mediating the response to H2O2. Initially, we confirmed the presence of mPGES-1 mRNA in Calu-3 cells via PCR; COX-1 and COX-2 expression was also confirmed (Fig. 3A). We next investigated the potential role for this enzyme by using 2-[[4-[([1,1′-biphenyl]-4-ylmethyl)amino]-6-chloro-2-pyrimidinyl]thio]-octanoic acid (Cay10589), a reported mPGES-1 inhibitor, and Cay10526, which decreases mPGES-1 expression (Guerrero et al., 2009). In the presence of Cay10589 (5 μM), the H2O2-stimulated response was significantly inhibited (Fig. 3, B and D); similarly, treatment of the cells with Cay10526 for 24 h (5 μM) abolished the response (Fig. 3, C and D). In addition, we confirmed the down-regulation of mPGES-1 expression in response to treatment with Cay10526 for 24 h by using quantitative PCR (Fig. 3E).
The effect of Cay10589 was additionally investigated by using Isc recordings from polarized monolayers. Preincubation with Cay10589 resulted in a significant decrease in the subsequent H2O2-stimulated response (ΔIsc = 11.1 ± 2.7 μA · cm−2; n = 3; Fig. 3, F and G). Therefore, these results indicate that, in addition to COX-1 playing a role in the generation of the biologically active molecule mediating the response to H2O2, mPGES-1 may potentially be involved.
Application of H2O2 Results in the Rapid Release of PGE2.
The experiments described above (Figs. 1⇑–3) suggest to us that the generation of PGE2 might be one, if not the exclusive, mechanism by which the oxidant stress response is mediated in this system. However, we decided to directly investigate whether H2O2 application did indeed result in PGE2 generation on a time scale that would be rapid enough to explain the effects we see. Therefore, we investigated whether application of H2O2 resulted in the generation of PGE2, which would then be available to act at prostanoid receptors at the cell surface. Cells were exposed to H2O2 (1 mM) for either 5, 10, or 30 min, after which the media were removed and assayed for PGE2 by using a commercially available ELISA. Significant increases were detected over basal levels at all three time points, with the maximum amounts detected at 10 min. After 30 min, levels seemed to be returning toward baseline and were no longer statistically significant (p = 0.02; Fig. 4A).
Because our H2O2-mediated anion efflux was inhibited by SC560 and Cay10589, we also investigated the effects of these agents on PGE2 production. Preincubation for 60 min with SC560 (0.5 μM) before exposure to H2O2 for 10 min significantly reduced the PGE2 production seen in Calu-3 cells in response to this oxidant stress (Fig. 4B; p = 0.015). However, in the presence of Cay10589, PGE2 production was not statistically different from H2O2 alone (p = 0.05).
The H2O2-Stimulated Efflux Probably Is Mediated via the EP4 Prostanoid Receptor.
Increased iodide efflux from Calu-3 cells can be stimulated by activation of the EP4 receptor, which despite the presence of all four PGE2 receptors on Calu-3 cells, is the only PGE2 receptor subtype capable of mediating a CFTR-mediated anion efflux (Joy and Cowley, 2008). The EP4 receptor also mediates the anion secretory response to 8-iso-PGE2, which is an example of an isoprostane, produced by the activity of ROS on cell membranes (Milne et al., 2008). We next wanted to investigate the possibility that some of the H2O2-stimulated response we see, which we find clearly linked to increased PGE2 production, might be mediated by the EP4 receptor. To do this, we used (4Z)-7-[(rel-1S,2S,5R)-5-((1,1′-biphenyl-4-yl)methoxy)-2-(4-morpholinyl)-3-oxocyclopentyl]-4-heptenoic acid (AH23848) (at 10 μM); in its presence, the H2O2-stimulated response was significantly inhibited, effectively to basal levels (Fig. 5). AH23848 also acts to inhibit the thromboxane (TP) A2 receptor (Abramovitz et al., 2000), and there are functional TPα receptors on Calu-3 cells capable of mediating a modest secretory response (Cowley, 2003). It is possible therefore that at least some of the H2O2-stimulated response we see is mediated via TP receptors. To address this question, we additionally performed iodide efflux experiments in the presence of [1S-[1α,2α(Z),3α,4α]]-7-[3-[[2- [(phenylamino)carbonyl]hydrazino]methyl]-7-oxabicyclo[2.2.1]hept-2-yl]-5-heptenoic acid (SQ29548), a TP receptor antagonist. The presence of 1 μM SQ29548 had no effect on the H2O2-stimulated iodide efflux response, with the rate of iodide efflux being 6.6 ± 0.2 (n = 6) for H2O2 alone versus 6.4 ± 0.4 (n = 6) for H2O2 in the presence of SQ29548 (results not shown). This result supports our contention that it is the EP4 receptor mediating the response rather than TP receptors.
Finally, because activation of the EP4 receptor has been associated with activation of the phosphotidylinositol-3-kinase (PI3K) pathway, both by our laboratory and others (Regan, 2003), we also investigated the H2O2 response after preincubation of cells for 30 min in the presence of the PI3K inhibitor 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY-290042) (10 μM). The response was significantly reduced after exposure to LY-290042, although not entirely abolished (Fig. 5B). Because there is no report that the TP receptor is associated with PI3K signaling, we believe these results again suggest the EP4 receptor is involved in mediating at least some of the anion efflux response from these cells.
To investigate the effect of EP4 receptor inhibition in a polarized system, we measured the increase in Isc induced by the application of 1 mM H2O2 to the apical face of Calu-3 cells in the presence and absence of AH23848. In these experiments, H2O2 induced a robust increase in Isc, which was significantly attenuated in the presence of 10 μM AH23848 (65.6 ± 10.9 μA · cm−2 for H2O2 alone versus 22.6 ± 3.0 μA · cm−2 in the presence of AH23848; Fig. 5, C-E). Thus using a different experimental system suggests a role for the EP4 receptor in mediating the anion secretory response to H2O2.
Direct Activation of the EP4 Receptor Is Unaffected by COX Inhibition.
The above experiments suggest that the H2O2-mediated anion secretory response depends on PGE2 production and is mediated, to some extent, via the EP4 receptor. To eliminate the possibility that direct activation of the EP4 receptor somehow depended on COX and/or mPGES-1 activity and consequent PGE2 production, we performed iodide efflux experiments by using the selective EP4 receptor agonist PGE1-OH. Application of this agonist induces a robust increase in anion efflux (Joy and Cowley, 2008); however, the presence of indomethacin, SC560, Cay10404, or Cay10589 (as described above) had no inhibitory effect on the PGE1-OH response (Fig. 6), suggesting PGE2 generation is not essential for a EP4 receptor-mediated secretory response in these cells.
The Secretory Response Is Reduced by Phospholipase A2 Inhibitors.
PGE2 production depends on the activity of phospholipase A2 (PLA2), an enzyme implicated in mediating inflammatory processes (Linkous and Yazlovitskaya, 2010). PLA2 activity cleaves membrane-bound phosphatidylcholine to form arachidonic acid, the precursor of prostaglandins. Additionally, however, isoprostanes generated by the activity of ROS on such polyunsatured fatty acids as arachidonic acid can remain in situ in the plasma membrane to be subsequently released via the activity of PLA2 (Rokach et al., 1997). Therefore, we wanted to investigate whether the H2O2 anion secretion was affected by PLA2 inhibition. Preincubation with the reported PLA2 inhibitor methyl arachidonyl fluorophosphate (MAFP; 5 μM for 20 min) completely abolished the H2O2-mediated response (Fig. 7; n = 8). MAFP is a reported irreversible inhibitor of both the cytosolic (cPLA2) and Ca2+-independent forms of the enzyme. We additionally measured the Isc in response to H2O2 in polarized cell monolayers after preincubation with MAFP as above. This resulted in a significant decrease in the H2O2-stimulated increase in Isc (ΔIsc = 6.7 ± 1.6; n = 3; results not shown).
We additionally used the more selective cPLA2 inhibitor arachidonil-trifluoromethyl-ketone (ATK) in iodide efflux experiments to see whether it also affected the response; preincubation with 5 μM for 20 min had no effect (n = 6). However, increasing the concentration to 50 μM significantly reduced the ability of H2O2 to induce anion efflux from Calu-3 cells (Fig. 7B; n = 6). These results suggest a role for cPLA2 in mediating this response; however, it is impossible to determine whether isoprostane release plays a role in mediating this response.
Oxidant stress is a well accepted component of the pathogenesis of inflammatory pulmonary lung diseases and represents a potentially important target for novel therapeutic approaches. The work presented here helps increase our understanding of how airway epithelial cells cope with an acute oxidant stress insult. We provide evidence that application of the model stressor H2O2 induces PGE2 production via COX-1 and possibly mPGES-1 activity, and we additionally propose that the EP4 receptor plays an important role in mediating the secretory response to H2O2.
H2O2 has a multitude of diverse physiological and toxicological effects on cells, which can be broadly classified into its ability to oxidize chemical structures (for example, direct oxidative effects on cysteine residues in cell membrane receptors or ion channels) or its ability to function as an intracellular signaling molecule (for example, effects on cell proliferation and transcription factors). H2O2 is a molecule of particular relevance to study in the context of airway epithelial host defense, because it is toxic to bacteria and protects the lung from invading pathogens. When activated inflammatory cells are recruited to a site of infection, they release the superoxide radical (O2̇̄). This is quickly converted to H2O2, which kills bacteria, probably by affecting their membrane integrity (Bonvillain et al., 2011). However, in cases of chronic or severe pulmonary infection, the numbers of activated inflammatory cells producing H2O2 may be extremely high, and in addition to affecting bacteria, the oxidant produced may overwhelm innate antioxidant defenses and tissue damage may result. This system may be additionally compromised in the CF lung, where antioxidant defenses are compromised and neutrophil counts as high as 38 × 106 per milliliter of extracellular lining fluid have been reported (Konstan et al., 1993), which would be sufficient to generate the concentration of 1 mM H2O2 used in this study (Test and Weiss, 1984).
H2O2 application to polarized Calu-3 cell monolayers results in a rapid increase in the transepithelial movement of anions across the cell, mediated at the apical membrane by CFTR (Cowley and Linsdell, 2002). Thus, there seems to be an important CFTR-mediated dynamic response to oxidant stress in airway epithelial cells, which we have proposed would act to increase innate host defense responses and help clear bacteria (Cowley and Linsdell, 2002). Of course, such a response would be absent in CF airway cells, resulting in a propensity for oxidant-induced tissue damage in this disease. In the present study, we were interested in how H2O2 was able to rapidly mediate this anion secretion from Calu-3 cells, because there has been no report to date indicating that CFTR structure can directly be affected by H2O2. Of interest, however, was the observation that H2O2 induces PGE2 production in the kidney (Han et al., 2005), because PGE2 is also a potent mediator of anion efflux from Calu-3 cells (Joy and Cowley, 2008). Using iodide efflux as a measure of CFTR-mediated anion release, in combination with Isc recordings, we found that the rapid efflux in response to H2O2 is abolished by indomethacin, SC-5690, Cay10589, and Cay10526 (Figs. 1⇑–3), suggesting a role for PGE2 generation. PGE2 is generated from arachidonic acid by the activity of both COX and mPGES-1; COX, of which there are two isoforms, generates PGE2 by mediating its conversion from arachidonic acid to PGG2 and then PGH2 (Smith et al., 1996). The COX-1 isoform seems to be relatively constitutively active in most cell types, whereas COX-2 is inducible by inflammation and mitogens (Süleyman et al., 2007). We find both isoforms expressed in Calu-3 cells (Fig. 3A), at least at the mRNA level; however, our finding that SC560 reduced the H2O2-mediated response, whereas Cay10404 had no effect, suggests a possible role for COX-1.
PGE2 production also depends on PGE synthases, which catalyze the conversion of PGH2 to PGE2. mPGES-1 expression is induced by inflammation, and since its initial description, mPGES-1 inhibition has been proposed as a potentially useful anti-inflammatory therapy (Jakobsson et al., 1999). To date, there have been relatively few studies investigating the role of mPGES-1 in human airway epithelial cells. We demonstrate here mPGES-1 mRNA expression in Calu-3 cells, as well as suggest a possible role in mediating the H2O2-stimulated efflux response. In the present study we used two compounds previously reported as selective inhibitors of mPGES-1: Cay10526 and Cay10589 (Guerrero et al., 2009). These compounds have different mechanisms of action, Cay10589 by direct inhibition of the enzyme and Cay10526, which decreases gene expression and thus inhibits production. However, the recent report by Yu et al., (2011) reveals significant non-mPGES-1-related effects of these drugs, because both agents also inhibited PGD synthase, an enzyme responsible for conversion of PGH2 to PGD2, and thus crucially involved in the generation of the D- and J-series of prostaglandins. Although we have not confirmed PGD synthase expression in Calu-3 cells, inhibition of this enzyme could theoretically shift PGH2 metabolism away from PGD2-derived products, instead favoring, for example, prostacyclin production. Therefore, although we clearly demonstrate a down-regulation of mPGES-1 gene expression via qPCR after Cay10526, we can not be certain that the effects we see on iodide efflux with Cay10589 are exclusively caused by the effects on mPGES-1 inhibition and may rather reflect a combination of inhibitory effects on other enzymes involved in this cascade.
Although the effects of Cay10526 and Cay10589 may provide potential evidence that mPGES-1 activity is involved in the production of PGE2 underlying the H2O2-secretory response we see, this result was, in fact, a little surprising, because mPGES-1 activity is usually preferentially coupled to COX-2 (Murakami et al., 2000). Because here we are examining the basal, nonstimulated situation, rather than after a sustained proinflammatory stimulation, the mechanism underlying how mPGES-1 inhibition blocks the H2O2 response is not immediately apparent. One possible explanation is that COX-1 can couple with mPGES-1 under very specific circumstances, namely in the presence of large amounts of arachidonic acid, or if there is a burst in activity of cytosolic PLA2 induced by calcium mobilizers (Murakami et al., 2000). Because the application of H2O2 would be expected to induce lipid peroxidation and degradation, it is possible that arachidonic acid is being released, as has been reported from renal epithelia (Salahudeen, 1995). It is noteworthy that our findings mirror those of Soodvilai et al., (2007) using inner medullary collecting duct cells. Those authors reported that H2O2 stimulated Cl− secretion via PGE2 production that was blocked by COX-1, but not COX-2 inhibition, and which they also linked to mPGES-1 activity. Therefore, the possibility that COX-1 and mPGES-1 are coupled to produce PGE2 is not without precedent. However, the possibility of nonspecific effects of Cay10526 and Cay10589 must be raised, and we may be seeing a non-mPGES-1-related effect on PGD synthase or some other unknown enzyme.
Application of H2O2 to Calu-3 cells did result in a significant increase in PGE2 detectable in the extracellular milieu within 5 min, which was the shortest time period we were able to sample (Fig. 4A). This increase in PGE2 was, perhaps not surprisingly, inhibited in the presence of the COX-1 inhibitor SC560; however, inhibition with Cay10589 failed to decrease PGE2 production significantly. Soodvilai et al. (2007) demonstrated H2O2-induced PGE2 release by using mouse inner medullary collecting duct cells; therefore, one possibility is that we are observing a more universal mechanism of oxidant stress handling. However, the lack of effect seen with the purported mPGES-1 inhibitor Cay10589 may again reflect either that it is not selectively inhibiting mPGES-1 in our system or may simply be a reflection of our relatively small sample size.
A second major component of this work was to evaluate the extent to which the EP4 receptor was mediating the secretory response to H2O2, because we have found this receptor to be the only PGE2 receptor subtype associated with CFTR-mediated anion secretion from Calu-3 cells (Joy and Cowley, 2008). When H2O2 was applied in the presence of AH23848 the iodide efflux response was entirely abolished. Because AH23848 has antagonistic effects at both the EP4 and TP receptor (Abramovitz et al., 2000) we performed iodide efflux experiments in the presence of SQ29548, a TP receptor antagonist. This agent had no effect on the H2O2-mediated response, suggesting that we were indeed looking at a predominantly EP4 receptor-mediated event, rather than one involving TP receptors.
Initially, we found the complete loss of iodide efflux response in the presence of AH23848 somewhat surprising and therefore decided to investigate the role of this receptor in more detail by moving to a polarized cell system. H2O2 applied to the apical membrane produces a large increase in the Isc, reflecting increased transepithelial anion movement mediated by the coordinated activity of apical CFTR and basolateral potassium channels (Cowley and Linsdell, 2002). When H2O2 was applied in the presence of AH23848 there was a significant attenuation of the response, although it was not abolished (Fig. 4, D and E). Together, these results would seem to suggest some role for the EP4 receptor in mediating the response to H2O2. Because the EP4 receptor is expressed at both the apical and basolateral membranes of polarized Calu-3 cells (Joy and Cowley, 2008), application of H2O2 and AH23848 to the apical face of polarized cells will predominantly affect EP4-CFTR mediated events; however, apically applied H2O2 could move into the cell and have effects at basolateral EP4 receptors or K+ channels, both of which would be expected to increase transepithelial anion secretion, as we see. Application of AH23848 in the nonpolarized system would inhibit all EP4 receptors, abolishing effects on both CFTR and K+ channels, and perhaps explaining the total lack of response we see with the efflux assay.
Direct stimulation of the EP4 receptor with the agonist PGE1OH stimulates iodide efflux, independent of PGE2 production (Fig. 6). In the final series of experiments we investigated the effect of PLA2 inhibition. PLA2 controls eicosanoid production by catalyzing the hydrolysis of membrane phospholipids to release arachidonic acid. cPLA2 has been reported to directly interact with CFTR via formation of a complex, also including annexin-1 and S100 calcium-binding protein A10 (Borot et al., 2009). In their article, application of tumor necrosis factor α to Calu-3 cells induced cPLA2 activation and increased PGE2 production. Furthermore, pharmacological inhibition of CFTR also increased PGE2 production. Therefore, it seems that a CFTR-cPLA2 complex may be a critical determinant of PGE2 production, at least in response to tumor necrosis factor α stimulation. We report that inhibition of cPLA2 plays a role in mediating the oxidant stress response, which again may implicate PGE2 (Fig. 7). Alternatively, the potential role of a cPLA2/CFTR/annexin-1/S100 calcium-binding protein A10 complex in inhibition of the iodide efflux response in the presence of cPLA2 inhibitors must also be considered.
CF airway epithelia reportedly demonstrate defective redox regulation, producing elevated H2O2 and an associated production of the proinflammatory cytokines IL-6 and IL-8, which has been linked to the transcription factor Nrf-2 (Chen et al., 2008). Our present findings suggest that H2O2 activates CFTR via PGE2 generation, and the EP4 receptor is a likely candidate for mediating these events. We would propose that this represents an acute host defense response to activate CFTR, releasing antioxidants and antimicrobial agents. However, this important mechanism is lost in CF epithelia, which, already burdened with excessive H2O2 production (Chen et al., 2008), will demonstrate a more inflammatory milieu via increased PGE2 production and EP4 receptor activation. We have recently reported that chronic stimulation of the EP4 receptor is associated with increased levels of IL-6 and IL-8, as well as PGE2 (Li et al., 2011). Therefore, our data add to our knowledge of physiological responses in the normal airways and perhaps suggest potential defects in the CF airways that could be targeted.
One final consideration is that isoprostanes generated by the activity of ROS on arachidonic acid and other membrane-bound polyunsaturated fatty acids can remain in situ in the plasma membrane to be subsequently released via the activity of PLA2 (Rokach et al., 1997). We have previously concluded that the activity of at least one isoprostane, 8-iso-PGE2, is reportedly mediated by the EP4 receptor (Joy and Cowley, 2008), whereas Sametz et al. (2000) reported that 8-iso-PGE2 achieves its effects in smooth muscle via the TP receptor. The release of this molecule would be predicted to be PLA2 dependent; however, it is unfortunately impossible to determine the extent to which the secretory response we see is caused by isoprostane generation. However, our finding that the TP receptor antagonist SQ29548 had no effect on the response, whereas AH23848 provided significant inhibition, suggests to us that we are dealing with EP4 receptor-mediated effects, be they mediated by one or several agents. The actions of ROS in polyunsatured fatty acids can potentially result in the generation of an enormous number of metabolites including isoprostanes, isolevuglandins, isothromboxanes, and isoleukotrienes (Janssen, 2001). Identification and investigation of these factors are hindered by the lack of available selective and sensitive research tools. It is possible, indeed maybe even probable, that some of the H2O2-mediated secretion we report here is caused by the generation of other eicosanoids or reactive by-products, such as isoprostanes. These present studies are unable to conclusively determine the extent to which these other compounds may contribute, although we believe it is highly probable that they play a role.
In summary, application of the model oxidant stressor and biologically relevant molecule H2O2 results in CFTR-mediated anion efflux from Calu-3 airway epithelial cells. We provide evidence suggesting that this response may involve the rapid generation of PGE2 and likely subsequent activation of the EP4 receptor. Anion secretion, which would be accompanied by water movement, would help maintain a healthy environment at the airway surface and facilitate innate host defense mechanisms to clear the oxidant stress source. Clearly, this mechanism would be lost in the CF airway, predisposing these lungs to oxidant damage. Because chronic activation of the EP4 receptor results in the increased release of inflammatory mediators (Li et al., 2011), prolonged exposure to ROS would be expected to produce the same effect. Therefore, our data underscore the significance of devising strategies to combat oxidant stress in the CF lung.
Participated in research design: Jones and Cowley.
Conducted experiments: Jones and Li.
Performed data analysis: Jones, Li, and Cowley.
Wrote or contributed to the writing of the manuscript: Cowley.
We thank Dr. Eileen Denovan-Wright for the use of equipment.
This work was supported by Cystic Fibrosis Canada.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
- cystic fibrosis
- CF transmembrane conductance regulator
- prostaglandin E2
- PGE2 subtype 4
- reactive oxygen species
- hydrogen peroxide
- short circuit current
- microsomal prostaglandin E synthase-1
- hypoxanthine guanine phosphoribosyltransferase
- polymerase chain reaction
- quantitative PCR
- reverse transcription-PCR
- enzyme-linked immunosorbent assay
- methyl arachidonyl fluorophosphate
- analysis of variance
- phospholipase A2
- cytosolic PLA2
- 2-[[4-[([1,1′-biphenyl]-4-ylmethyl)amino]-6-chloro-2-pyrimidinyl]thio]-octanoic acid
- (4Z)-7-[(rel-1S,2S,5R)-5-((1,1′-biphenyl-4-yl)methoxy)-2-(4-morpholinyl)-3-oxocyclopentyl]-4-heptenoic acid
- N-(naphthalen-2-ylamino)-acetic acid(3,5-dibromo-2,4,-dihydroxy-benzylidene)-hydrazide
- [1S-[1α,2α(Z),3α,4α]]-7-[3-[[2-[(phenylamino)carbonyl]hydrazino]methyl]-7-oxabicyclo[2.2.1]hept-2-yl]-5-heptenoic acid
- Received August 19, 2011.
- Accepted February 13, 2012.
- Copyright © 2012 by The American Society for Pharmacology and Experimental Therapeutics