| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
F508CFTR on inflammatory response, ER stress, and Ca2+ of airway epitheliaDepartment of Molecular and Cell Biology, University of California-Berkeley, Berkeley, California
Submitted 11 June 2007 ; accepted in final form 3 September 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
F508CFTR in the endoplasmic reticulum (ER) and activation of cytosolic Ca2+ (Cai) and nuclear factor (NF)-
B signaling. Adenovirus infections of a human CF (F508/F508) nasal cell line (CF15) provided isogenic comparisons of wild-type (wt) CFTR and F508CFTR. In the absence of bacteria, there were no or only small differences among CF15, CF15-lacZ (
-galactosidase-expressing), CF15-wtCFTR (wtCFTR-corrected), and CF15-F508CFTR (to test ER retention of F508CFTR) cells in NF-B activity, interleukin (IL)-8 secretion, Cai responses, and ER stress. Non-CF and CF primary cultures of human bronchial epithelial cells (HBE) secreted IL-8 equivalently. Upon infection with Pseudomonas aeruginosa (PA) or flagellin (key activator for airway epithelia), CF15, CF15-lacZ, CF15-wtCFTR, and CF15F508CFTR cells exhibited equal PA binding, NF-B activity, and IL-8 secretion; cells also responded similarly to flagellin when both CFTR (forskolin) and Cai signaling (ATP) were activated. CF and non-CF HBE responded similarly to flagellin + ATP. Thapsigargin (Tg, releases ER Ca2+) increased flagellin-stimulated NF-B and ER stress similarly in all cells. We conclude that ER stress, Cai, and NF-B signaling and IL-8 secretion were unaffected by wt- or F508CFTR in control and during exposure to PA, flagellin, flagellin + ATP, or flagellin + ATP + forskolin. Tg, but not wt- or F508CFTR, triggered ER stress. Previous measurements showing hyperinflammatory responses in CF airway epithelia may have resulted from cell-specific, rather than CFTR- or F508CFTR-specific effects.
nuclear factor-B; interleukin-8; adenovirus; inflammation; endoplasmic reticulum stress; ire1
Many investigators have observed higher expression and secretion of proinflammatory mediators in CF cells than in paired CFTR-expressing cells in the control state (no bacteria), and this hyperinflammatory state of CF cells is exacerbated by exposure to Pseudomonas aeruginosa (PA) or Hemophilus influenzae (6, 7, 11, 12, 15, 34, 49, 50, 59). This has been observed repeatedly for the IB3 (CF, F508/W1282X) and C38 or S9 (CFTR-corrected; see Refs. 12, 50, and 59) pairs and also for the 9HTEo–/pCep [wild-type (wt) CFTR] vs. 9HTEo–/pCepR (overexpress R domain and do not secrete Cl–; see Refs. 7, 34, and 59) pairs. The CF cells also bind more bacteria than the non-CF or CFTR-corrected cells (7, 11, 19). Consistent with these findings, a recent study showed that a specific CFTR blocker caused, after 6 days of treatment, an increase in proinflammatory signaling and interleukin (IL)-8 secretion of CFTR-expressing but not CF airway epithelial cells (35), indicating that inhibition of CFTR channel function was sufficient to increase proinflammatory signaling. In contrast, CFT1-F508 cells exhibit smaller expression and secretion of proinflammatory cytokines than CFTR-corrected CFT1-CFTR cells in the control state and also during exposure to different strains of PA (29, 40). Ribeiro and colleagues (41) have shown that control (no bacteria) primary CF bronchial cells were hyperinflammatory during short-term (up to 11 days) culturing, but this property was lost during longer-term (>30 day) culturing. This exaggerated response of CF cells appeared to arise from sustained effects of extended exposure to inflammatory conditions characteristic of the CF lungs and not from the differential effects of CFTR expression. However, Ribeiro et al. (41) did not test specific bacterial products on the IL-8 responses of long-term CF vs. non-CF cells, although they did show that CF supernatants that likely contained such bacterial products had similar effects on long-term CF and non-CF cells. Others demonstrated different expression of only 24 out of 22,283 genes in CF vs. non-CF primary cells in control conditions (64) and no difference in intercellular adhesion molecule (ICAM)-1 and IL-8 during 6 h of bacterial exposure (2). Similarly, non-CF and CF primary cells responded similarly to cytokines a toll-like receptor (TLR) 2 agonist and to Staphylococcus aureus filtrate, although there small differences in responses to PA when cells were incubated in serum-containing medium (4). Responses of primary CF vs. non-CF airway epithelia to bacteria under conditions that would assure elimination of effects of prior CF exposure remain to be tested.
A potential problem with in vitro experiments on cell lines is that selection and passaging of cells transfected with CFTR, some by stable integration into the genome and some by episomal expression (see Ref. 2), may generate cells with different genetic backgrounds. This problem can be overcome by using one CF cell line and expressing CFTR using an adenovirus, as done two times previously. Aldallal et al. (2) showed that CF (IB3 and primary human airway epithelia) and CFTR-corrected CF cells exhibited similar expression of ICAM-1 and secretion of IL-8 under control conditions or during exposure to H. influenzae or PA strain PAO1 or to several proinflammatory cytokines. Pizurki et al. (36) found that CF15 cells [polarized human nasal CF (F508/F508) cells (24)] bound more PA than CFTR-corrected (using an adenovirus) CF15 cells, but CF15 and CFTR-corrected CF15 cells produced similar amounts of IL-8 in control conditions and during exposure to IL-1.
The present experiments used the adenoviral approach to evaluate the roles of wt- and F508CFTR in PA-triggered early innate immune responses of the CF15 cell line (21). Experiments were performed with confluent, polarized monolayers that were previously shown to elicit physiological restoration of CFTR expression in the plasma membrane and transepithelial Cl– transport after adenoviral correction (48). The role of PA binding in inflammatory initiation (7, 11) was tested by measuring binding of green fluorescent protein (GFP)-labeled bacteria to CF and CFTR-corrected cells. IL-8 secretion and nuclear factor (NF)-B activity were measured because these are key intermediates in the proinflammatory responses of airway epithelia (11, 13). We tested multiple strains of PA to address the differential responsiveness of CF vs. CFTR-expressing airway epithelia (40). We also tested flagellin, the critical bacterial activator of airway epithelial cell innate immune responses (1, 17, 31, 52, 65), to eliminate PA-secreted factors that may alter epithelial responses (8). IL-8 secretion was measured from primary human CF and non-CF bronchial cells that had been passaged to eliminate problems associated with short-term growth of primary CF airway epithelia (41).
Retention of misfolded F508CFTR in the endoplasmic reticulum (ER) could trigger ER stress (62, 63), and it has been proposed that such stress alters ER cytosolic Ca2+ (Cai) handling (59) and increases inflammatory signaling (28, 59). Ca2+ entry in CF cells may also increase in response to hyperpolarized membrane potential in CF (28, 39, 61). We tested for ER stress by measuring the activity of IRE1, an ER membrane protein receptor for ER-unfolded proteins that has both endonuclease and kinase activities that are activated when misfolded proteins accumulate in the ER (62, 63). F508CFTR with a GFP tag was overexpressed in CF15 cells and also in HeLa cells, which do not express CFTR. Finally, CFTR- and F508CFTR effects on Cai signaling and NF-B activity were tested. Although PA and flagellin activate NF-B, they do not increase Cai in CF15 cells (13, also see Ref. 25) so we stimulated cells with flagellin plus either ATP (purinergic agonist) or thapsigargin (inhibitor of Ca2+-ATPase in the ER) to increase Cai. Potential effects of wtCFTR to depolarize membrane voltage, decrease Ca2+ entry, and reduce flagellin + ATP or flagellin + thapsigargin activation of NF-B were also maximized by stimulating with forskolin to increase cellular cAMP concentration and activate protein kinase A.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
B-driven luciferase (adv-NF-Bluc; see Ref. 48) were purchased from the University of Iowa Vector Core (2, 53). Adenoviruses to express wtCFTR or F508CFTR with GFP appended to the COOH termini (wtCFTR-GFP and F508CFTR-GFP; see Ref. 54) were purchased from the University of Pennsylvania Vector Core (James Wilson).
CF15, HeLa, and Calu-3 cell culture and adenovirus infections.
CF15 CF airway cells, a continuous SV-40 large T antigen-transformed nasal epithelial cell line homozygous for F508 CFTR (21, 36), were cultured in a 1:1 mixture of DMEM and Ham's F-12 medium (DMEM-F-12) media containing 10% FBS, 2 mM L-glutamine, 1% penicillin/streptomycin, 10 ng/ml epidermal growth factor, 1 µM hydrocortisone, 5 µg/ml insulin, 5 µg/ml transferrin, 30 nM triiodothyronine, 180 µM adenine, and 5.5 µM epinephrine. All CF15 cells were grown in the absence of the antibiotic resistance marker G418.
For some experiments, cells were passaged at a 1:5–1:10 dilution, and the remaining cell suspension was seeded directly on cover glasses or on a tissue culture plate (BD Falcon, Bedford, MA). In other experiments, cells were passaged on Transwell membranes (Corning/Costar) and then grown until cells formed confluent monolayers. Control experiments were performed in Ussing chambers to assure filter-grown cells attained confluence and were polarized. As reported previously (48), CF15 cells on filters had transepithelial resistances (RT) greater than 200 
·cm2 and exhibited amiloride-sensitive, serosa-positive currents of 10–25 µA/cm2, consistent with Na+ absorption through the epithelial Na+ channel (ENaC). Forskolin treatment increased this apparent Na+ absorption but did not increase anion currents (data not shown), consistent with their CF genotype.
Calu-3 cells, a human gland epithelial cell line homozygous for wt CFTR (27), were cultured in DMEM supplemented with 10% FBS, 2 mM L-glutamine and 1% penicillin/streptomycin. Passaging and growth on filters or in plastic tissue culture dishes was the same as that described for CF15 cells.
HeLa cells were cultured in DMEM media containing 10% FBS and 1% penicillin/streptomycin at 37°C with 5% CO2 and 95% humidity.
CF15 cells, Calu-3 cells, or HeLa cells were infected with adv-CFTR, adv-lacZ, adv-wtCFTR-GFP, adv-F508CFTR-GFP, and/or adv-NF-B-luc (MOI = 100) in media for 48 h at 37°C. Epithelia grown on filters were infected on apical and basal sides. Cells grown on plastic or cover glasses were infected by additions to the media. Previous experiments showed that multiplicity of infection (MOI) = 100 for 48 h gave a maximal number of cells infected (80–100%) and increases in Cl– secretion (36, 48). Experiments with adv-wtCFTR-GFP and adv-F508CFTR-GFP showed similar frequency of cellular expression.
Primary human non-CF and CF bronchial cells.
Bronchial epithelial cells were isolated from cadavers or lungs removed at the time of transplantation (60) from one non-CF and CF individual using methods described previously. Briefly, epithelial cells were obtained by protease digestion and then grown and passaged one time, as described previously (45). These P1 ("first passage") cells were suspended in a 1:1 mixture of DMEM-F-12 containing 5% FBS and seeded at 5 x 105 cells/cm2 on Transwell polycarbonate inserts (no. 3401; Costar, Corning, NY). Growth surfaces were coated with human placental collagen. The day after plating, the DMEM-F-12 over the filters was replaced with Gray's medium, a medium whose composition is fully described elsewhere (45, 60). Once the cells had attained confluence, they were grown with an air-liquid interface (i.e., medium was only added to the basolateral surfaces). Previous experiments have shown that these monolayers contain epithelial cells with no obvious contamination from myeloid cells (45). For measurements of electrical RT, 500 µl of PBS were added to the mucosal surface, and a "chop-stick" voltmeter (Millicell ERS; Millipore Products, Bedford, MA) was used to measure RT. Experiments were performed 3–4 wk after plating (>14 days after attaining confluence). These cultures attained minimum RT >100 ·cm2 (usual RT = 400–500 ·cm2) and maintained a dry mucosal surface.
P. aeruginosa. PA strains PAK, PAO1 (serotype 05), PAC1R (serotype 03), and PA2192 (mucoid clinical isolate; see Refs. 8–10 and 55) were grown overnight in LB culture as described previously (17, 20, 26). Before experiments, PA were washed three times with PBS and resuspended in CF15 media minus antibiotics and supplements at a concentration of 108 colony-forming units/ml (optical density at 600 nm = 0.05). GFP-expressing strains PAK and PAO1 (PAK-GFP, PAO1-GFP) were generated and then grown and used as described previously (17).
Enzyme-linked immunosorbent assay of IL-8 secretion. Samples were routinely collected from the basal chamber of cells grown on filters and in some experiments from media bathing cells grown on plastic. Enzyme-linked immunosorbent assay (ELISA) methods have been published previously (13, 17). Briefly, samples were collected, centrifuged (5 min 1,000 g), stored at –20°C, and then thawed, diluted in Assay Diluent (BD Pharmingen, San Diego, CA), run in triplicate per the manufacturer's protocol (OptEIA Human IL8 Set; BD Pharmingen), and read at 450 nm with an ELX808 Ultra Microplate Reader (Bio-Tek Instruments, Winooski, VT).
NF-B-regulated luciferase.
A recombinant adenoviral vector expressing a luciferase reporter gene driven by NF-B transcriptional activation (adv-NF-Bluc) was used for functional studies as described previously (47). Adenoviral stocks were stored in 10 mM Tris with 20% glycerol at –80°C. The virus was added to CF15 cell monolayers alone or along with any other viruses at 100 MOI and returned to the incubator for 48 h. Monolayers were then washed three times to remove viruses and then exposed to PA or flagellin for 4 h. Monolayers were then washed and processed using the luciferase assay system with Reporter Lysis Buffer (Promega, Madison, WI) to measure NF-B-mediated transcriptional induction according to the manufacturer's protocol. Measurements of luciferase activity (relative light units) were performed in triplicate for each sample and normalized to the protein concentration (Bradford assay). These averages were then expressed relative to the average control values (absence of bacteria), which was set equal to 1.0.
PAK-GFP and PAO1-GFP binding.
CF15 or adenovirus-treated CF15 monolayers were incubated with either PAK-GFP (MOI = 100) or PAO1-GFP (MOI = 10) for 1 h, washed three times, and fixed and stained for NF-B (anti-p65 antibody; Santa Cruz Biotechnology) immunofluorescence to permit visualization of the cells in the fluorescence microscope. PA binding occurred in areas in which epithelial cells had overgrown in multilayers as described previously (17). Individual bacteria and epithelial cells were counted from confocal images (17), and bacterial binding was expressed as number of PA bound per 100 epithelial cells.
Cai measurements. Cells grown on cover glasses were incubated in the original growth media containing 1–10 µM fura 2-AM for 40–60 min at room temperature and then washed three times with Ringer solution to remove the extra dye. Fura 2-AM (1–3 µM) was used to load nonconfluent cells, whereas 5–10 µM fura 2-AM was used to load confluent cells. For Cai measurements, epithelial cells were incubated in solutions containing (in mM): 145 NaCl, 1.2 MgSO4, 2 CaCl2, 2.4 K2HPO4, 0.6 KH2PO4, 10 HEPES, and 10 glucose (pH 7.4). Ca2+-free Ringer was composed of the same solution without added Ca2+. Treatments of cells with ATP were made by diluting stock solutions 1,000x in Ringer solution.
Fluorescence ratio imaging measurements of Cai were performed using methods that have been reported previously (13, 20). Briefly, a Nikon Diaphot inverted microscope and charge-coupled device camera collected emission (>510 nm) images during alternate excitation at 350 ± 5 and 380 ± 5 nm using a filter wheel (Lambda-10, Sutter Instruments, Novato, CA). Axon Imaging Workbench 4.0 (Axon Instruments, Foster City, CA) controlled both filters and collection of data. Calibration of fura 2 signals was performed as described previously (13, 14), assuming the apparent dissociation constant for fura 2 was 224 nM (14). In situ calibration of fura 2 was carried out by treating the cells with ionomycin (10 µM) and then perfusing the cells with a Ca2+-free external solution to determine Rmin and with a solution containing 2 mM Ca2+ to determine Rmax. Rmin and Rmax are the minimum and maximum fluorescence (>510 nm) ratios measured during 350 nm/380 nm excitation. All images were corrected for background (region without cells).
Confocal microscopy of wtCFTR-GFP- and F508CFTR-GFP-expressing cells.
CF15 and HeLa cells grown on cover glasses and infected as described above with wtCFTR-GFP and F508CFTR-GFP were mounted live on the stage of an inverted microscope (Nikon TE2000) and observed with a spinning disk confocal attachment (488 nm excitation and 520–560 nm emission; Solamere Technology, Salt Lake City, UT). Images were collected from z-axis regions roughly in the middle of the cell nucleus.
ER stress: IRE1 activity assessed by splicing of XBP1 RNA.
ER stress was measured using a method that measures the endonuclease activity of IRE1. This ER membrane-bound enzyme is activated during increases in unfolded proteins in the ER lumen or sustained decreases of ER Ca2+ concentration or redox potential (3, 62, 63). Two RT-PCR-based assays for activation of IRE1 were performed using primers up- and downstream of the 26-bp exon of XBP1 mRNA. RT-PCR was performed on RNA isolated from CF15 and HeLa cells and also on PST-I digests of PCR products. CF15 and HeLa cells were either left untreated or treated with adenoviruses to generate expression of CFTR, CFTR-GFP, and F508CFTR-GFP. The cells were also treated with thapsigargin (1 µM, 4 h) as a positive control (63).
ER stress was quantitated from intensities of four bands of RT-PCR products that were designated as follows: "Band ," the largest product (
500 bp), resulted from full-length XBP-I; "Band " ran slightly faster because this band was full-length XBP1 – 26 bp; PST-I-treatment of the RNA cleaves XBP-1 (Band ) to produce two smaller bands (Band 
, 300 bp; Band 
, 200 bp). XBP1 PCR products from control cells with inactive IRE1 will have prominent Band and no or little Band . The Band -to-Band ratio measured (using ImageJ 1.37v, including background subtraction) should be small. PST-I digests of XBP1 PCR products from control cells will show little or no Band or but prominent Bands and , and the Band /(Bands + + ) ratio should also be small. In contrast, XBP1 products from ER-stressed cells with active IRE1 will have prominent Band , no or little Band , and the Band /Band ratio will be larger than control cells. PST-I digests of XBP1 PCR products from ER-stressed cells will show prominent Band and less or no Bands , , and , and the Band /(Bands + + ) ratio should be larger than in control cells without ER stress.
Assays for IRE1 activity have been described previously (3). Briefly, CF15 and HeLa cells grown in culture dishes over night were infected with the adenoviruses (MOI = 100) or left in culture medium. Cells were cultured for 2 days and incubated with medium or with thapsigargin (1 µM) for 4 h. Total RNA was isolated (RNeasy Isolation Kit; Qiagen). First-strand cDNA was generated by incubating 2 µg RNA with 2.5 mM dNTPs, 1 pmol random hexamer primers, 20 units RNase inhibitor, and 200 units SuperScript II reverse transcriptase (Invitrogen) for 1 h at 42°C. XBP1 DNA was PCR-amplified using 2 µl first-strand cDNA, 300 nM forward primer (5'-AAACAGAGTAGCAGCTCAGACTGC-3'), 300 nM reverse primer (5'-TCCTTCTGGGTAGACCTCTGGGAG-3'), 1 µl 10 mM dNTPs, and 2.6 units RedTaq DNA polymerase (Sigma). A thermal-cycler program consisted of 94°C for 10 min; 94°C for 20 s, 65°C for 20 s, and 72°C for 30 s for 35 times; and 72°C for 5 min. 18S RNA was assayed under identical conditions using specific primers (Ambion). PCR products were analyzed on a 3% agarose gel. Products were also purified using the Qiaquick PCR Purification Kit (Qiagen) and digested with PST-I (5 U/µg PCR product; Fermentas) for 
2 h before agarose gel analysis.
Statistics. Unpaired or paired t-tests were used to compare groups and effects, depending on the experiments (StatView; Abacus Concepts, Berkeley, CA). P < 0.05 was considered significant. Data have been presented as averages ± SD or including values from all individual experiments; n refers to the no. of experiments (for luciferase and IL-8 assays, this refers to the no. of different wells or filters sampled; for Cai measurements, n refers to the no. of different cover slips examined; and for XBP-1 assays, n refers to the no. of different cell preparations examined).
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
|
B activation, IL8 secretion, and PA binding?
NF-B luciferase assays were performed on control monolayers and on cells exposed on the apical surface to PAO1, PAK, PAC1R, or PA2192 (MOI = 100) or to flagellin (10–7 g/ml) for 4 h. Samples were also taken from the basolateral solution at time 0 and then after 4 h to measure IL-8 secretion.
NF-B activity of control cells was slightly (35%) higher in CF15 than in CF15-lacZ cells (Fig. 2A). However, there was no significant difference between CF15 and CF15-lacZ cells in IL-8 secretion (Fig. 2B). Additional of strains PAO1, PAK, PAC1R or PA2192, or flagellin to the apical surfaces of CF15 monolayers increased both NF-B activity (Fig. 2A) and IL-8 secretion (Fig. 2B), but there were no differences in responses of CF15 vs. CF15-lacZ or of CF-lacZ vs. CF15-CFTR. There was also no effect of CFTR expression on responses to flagellin. Compared with PAK and PAO1, PA2192 and PAC1R elicited larger stimulations of NF-B (Fig. 2A) and IL-8 (Fig. 2B), indicating that, as shown previously (40), there were strain-dependent differences in PA activating innate immune responses. However, unlike previous work (40), there were no differences in NF-B or IL-8 responses attributable to CFTR expression (Fig. 2, A and B).
|
|
F508CFTR in triggering ER stress?
A complication in the interpretation of data in Figs. 1–3 is that CF15 cells may express only low levels of F508CFTR protein that are insufficient to trigger an ER stress response and, consequently, not change Cai signaling. This issue was addressed by expressing F508CFTR tagged with COOH-terminal GFP (54), which enabled detection of transduced cells and evaluation of the subcellular localization of F508CFTR-GFP. Similar experiments were performed in HeLa cells, which do not express CFTR and therefore will presumably not have a stressed ER, to determine whether any activation of IRE in CF15 cells resulted from an already overloaded ER. CF15 and HeLa cells were infected with adv-F508CFTR-GFP and also adv-wtCFTR and adv-wtCFTR-GFP for 1 day followed by washing and two further days of growth; these cells were then processed for measurements of IRE1 activity.
Representative confocal images of CF15 and HeLa cells expressing wtCFTR-GFP and F508CFTR-GFP are shown in Fig. 4. Each construct transduced 70–90% of the cells. wtCFTR-GFP appeared prominently in the plasma membrane and also in a spot near the nucleus, likely the Golgi, in both CF15 (Fig. 4A) and HeLa (Fig. 4B) cells. Lower-level expression throughout the cytosol, likely the ER, was also observed. In contrast, F508CFTR-GFP was not observed in the plasma membrane or Golgi but was found in the ER in both CF15 (Fig. 4C) and HeLa (Fig. 4D) cells.
|
activity was determined in CF15 and HeLa cells from XBP1 PCR products and the same products that had been subjected to PST-I digestion. 18S RNA was used as a loading control. Results from cells expressing wtCFTR, wtCFTR-GFP, or F508CFTR-GFP –/+ thapsigargin are shown in lanes 1–8 for CF15 cells (Fig. 5, A–C) and HeLa cells (Fig. 5, E–G). Values for the Band /Band ratio for control PCR products and Band /(Bands + + ) for PST-I-digests were similar in magnitude, so these were averaged. Data were plotted for CF15 cells in Fig. 5D and for HeLa cells in Fig. 5H. Values for relative IRE1 activities were normalized to the values measured in CF15 cells or HeLa cells under control conditions.
|
and a very faint Band (Fig. 5A, lane 1). PST-I treatment yielded Bands and , whereas Bands and were now only faintly visible (Fig. 5B, lane 1). Similar patterns were observed for CF15 cells that had been transduced with wtCFTR, wtCFTR-GFP, or F508CFTR-GFP (Fig. 5, A and B, lanes 3, 5, and 7, respectively). Control HeLa cells exhibited strong Band and absence of Band (Fig. 5E, lane 1), and PST-I digests (Fig. 5F, lane 1) showed no Bands and and prominent Bands and . Expression of wtCFTR or F508CFTR-GFP had essentially no effect on these patterns (Fig. 5, E and F, lanes 3 and 7). Expression of wtCFTR-GFP elicited a small increase in Band of controls (Fig. 5E, lane 5). Thapsigargin increased Band and decreased Band for CF15 and HeLa cells and for cells transduced with wt- or F508CFTR (Fig. 5A, lanes 2, 4, 6, and 8 and Fig. 5E, lanes 2, 4, 6, and 8). PST-I digests of the thapsigargin-treated cells showed low-intensity Band , even lower-intensity Band , and higher-intensity Bands and for control CF15 and HeLa cells and for the same cells transduced with wtCFTR, wtCFTR-GFP, or F508CFTR-GFP (Fig. 5B, lanes 2, 4, 6, and 8; Fig. 5F, lanes 2, 4, 6, and 8).
PST-I treatment did not totally cleave Band in any of the preparations except the control HeLa cells (Fig. 5F, lane 1), even when PST-I digestion was performed for 12 h (data not shown). Band from lanes 1–8 of both CF15 and HeLa cells was isolated, purified, and sequenced and shown to be full-length XBP-1 containing the PST-I restriction site and suggested that the PST-I digestion was incomplete.
Quantitation showed that IRE1 activity was insignificantly (P > 0.1) higher in control CF15 cells compared with control HeLa cells (Fig. 5D, lane 1 vs. Fig. 5H, lane 1). Compared with control CF15 cells, IRE1 activity was insignificantly altered by wtCFTR (P > 0.12), wtCFTR-GFP (P > 0.43), or F508CFTR-GFP (P > 0.10; Fig. 5D, lane 1 vs. lane 3; see Refs. 5 and 7). Similar results were obtained in HeLa cells, although wtCFTR-GFP (P = 0.02) and F508CFTR-GFP (P = 0.02) both caused significant increases compared with control (Fig. 5H, lanes 5 and 7 vs. 1). This may have resulted from some virus-dependent rather than CFTR-dependent effect in the HeLa cells because the wtCFTR adenovirus did not significantly alter IRE1 activity compared with controls (Fig. 5H, lane 1 vs. 3, P > 0.28), and there was no significant difference in IRE1 activity between HeLa cells infected with adv-wtCFTR-GFP and adv-F508CFTR-GFP (P > 0.1). In contrast, the ER stress-agonist thapsigargin caused >20-fold increases in IRE1 activity of CF15 and HeLa cells in control conditions or during expression of wt- or F508CFTR (Fig. 5D, lanes 2, 4, 6, and 8; Fig. 5H, lanes 2, 4, 6, and 8).
Role for CFTR in controlling Cai and NF-B of cells with activated Cai and CFTR?
CF15 cells infected with adv-lacZ or adv-CFTR were left untreated to determine baseline Cai and then treated with flagellin (which had no effect on Cai; see Ref. 13) or with ATP. A representative single cell response to ATP in CF15 cells is shown in Fig. 6, and averages for CF15-lacZ and CF15-wtCFTR cells are summarized in Table 2: control Cai = 100 nM, and ATP caused a rapid increase in Cai ("first peak") resulting from the rapid release of ATP from the ER (13), whereas, as shown previously (13), the "second peak" and "plateau" were mostly due to ATP-induced Ca2+ entry across the plasma membrane in the cells. CFTR expression had no apparent effect on either baseline Cai or in the first peak or plateau phases of the Cai responses to ATP (Table 2). CFTR expression had a small effect on the second peak Cai response, but this second peak was variable from preparation to preparation, so it seems that this small effect may have resulted from subtle differences in this second peak rather than a difference resulting from CFTR expression.
|
|
B responses of CF15-lacZ vs. CF15-CFTR and of CF15-F508CFTR vs. CF15-wtCFTR during treatments that both activated NF-B (flagellin) and also raised Cai (ATP and thapsigargin). We used both PAK and flagellin for these experiments because some previous experiments showed differences in Cai responses to PA and flagellin in CF vs. CFTR-corrected cells (38, 59). Previous experiments showed that Cai synergized with flagellin to activate innate immune responses (13), and, since cell membrane voltage is hyperpolarized in CF (28), it might have been expected that flagellin + ATP or flagellin + thapsigargin would cause larger activation in lacZ- or F508CFTR-GFP- vs. wtCFTR-GFP-expressing cells. CF15-lacZ and CF15-CFTR cells (Fig. 7A) behaved similarly during flagellin ± ATP; CF15-F508CFTR-GFP and CF15-wtCFTR-GFP cells (Fig. 7B) also responded similarly to flagellin and flagellin + ATP. There was a larger stimulation of NF-B by flagellin + thapsigargin for CF15-wtCFTR than for CF15-F508CFTR cells (Fig. 7B). PAK also stimulated similar responses in CF15-F508CFTR-GFP and CF15-wtCFTR-GFP cells (Fig. 7B). Thus CF15 cells or CF15 cells overexpressing F508CFTR responded similarly to wtCFTR-corrected cells during flagellin + ATP. Similarly, bronchial epithelial cells from non-CF and CF individuals secreted IL-8 in the basolateral solution at similar rates during flagellin + ATP: 2,933 ± 880 pg/h in non-CF (n = 8) and 1,867 ± 1,343 pg/h in CF cells (n = 6, P > 0.3).
|
B. In contrast to the stimulatory effect of ATP on flagellin-stimulated NF-B, there was no difference in NF-B activation during flagellin vs. flagellin + forskolin (Fig. 7A). This lack of effect of forskolin was also observed in Calu-3 cells: luciferase (relative to controls) for flagellin = 4.9 ± 2.5 and flagellin + forskolin = 4.8 ± 3.6, P > 0.8 (n = 4). Forskolin had an inhibitory effect on flagellin + ATP-stimulated NF-B activity (Fig. 7A). Two other experiments showed similar inhibitory effects of forskolin on flagellin + thapsigargin-stimulated NF-B activity in CF15-lacZ and CF15-CFTR cells and also in the CFTR-expressing Calu-3 cell line (Table 3). Thus forskolin inhibited effects of flagellin + Cai-raising agonists, but this effect did not depend on the presence or activity of CFTR.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
·cm2, and generate amiloride-inhibitable cation absorption characteristic of active apical ENaC and basolateral Na+-K+-ATPase (48). The cells also exhibit Na+/H+ and Cl–/HCO3– exchange activity in the basolateral but not the apical membranes (E. Wunderlich and T. E. Machen, unpublished observations) and Ca2+-activated Cl channels in the apical but not the basolateral membrane (E. Wunderlich, T. E. Machen, and B. Illek, unpublished observations). Exposure to apical PA for 4 h caused IL-8 secretion to increase from 600 to 6,000 pg/ml in individual experiments, which was equal to or larger than that measured in other related cell types (24, 38), indicating that CF15 cells responded to PA by secreting IL-8 at normal rates.
As detailed here and elsewhere (48), adenoviruses transduced 70–90% of CF15 cells. Inflammation assays indicated that there were only small effects on NF-B activity and IL-8 secretion that could be attributed to adenovirus alone. In addition, PA binding and NF-B and IL-8 responses were equivalent in CF15 and CF15-lacZ cells. These results indicated that adenovirus expression alone had little effect on innate immune defense properties of the cells and was thus a valuable method for making comparisons between CF vs. CFTR-corrected CF epithelia. Expression of wtCFTR-GFP resulted in apical membrane expression, and wtCFTR expression in CF15 cells increased forskolin-stimulated, DPC-inhibitable Cl– secretion (48). Thus the adenoviral method was a benign approach to expressing CFTR and other genes.
Similar inflammatory responses and (lack of) ER stress of CF15 and CFTR-corrected cells.
A major conclusion from these studies was that NF-B-activated luciferase activities and IL-8 secretion were similar for CF15 cells expressing lacZ, wtCFTR, or F508CFTR both in the control state and also during treatment with mulitiple strains of PA or flagellin. IL-8 secretion was also similar in control non-CF vs. CF primary bronchial cells. PA binding was also unaltered by expression of wtCFTR.
It might have been expected that, if there were ER stress resulting from accumulation of F508CFTR, then IRE would have been active in control CF15 cells and also in CF15 cells expressing F508CFTR-GFP. However, CF15 cells exhibited low IRE1 activity that was insignificantly different from that of HeLa cells. Although F508CFTR-GFP was retained in the ER in both CF15 and HeLa cells, whereas wtCFTR-GFP was found in the plasma membrane and at lower levels in the ER, there were only small (<20%) differences in IRE1 activity among these cells. Although there were significant stimulatory effects on IRE1 activity of HeLa cells expressing wtCFTR-GFP and F508CFTR-GFP, these small effects likely resulted from virus-dependent rather than CFTR-dependent effects: IRE1 activity was the same for wtCFTR vs. controls and for wtCFTR-GFP vs. F508CFTR-GFP.
In contrast to the lack of effect of any of the CFTR constructs, thapsigargin elicited typically robust (40-fold) activation of IRE1, which occurred similarly in controls and cells expressing wt- or F508CFTR. Although it has often been assumed that ER stress contributes to exaggerated inflammatory signaling in CF (57), the lack of effect of F508CFTR or wtCFTR on ER stress was not entirely unexpected: there is only 25% difference in ER retention of F508- vs. wtCFTR (23), and the misfolded portion of F508CFTR is in the cytosol, whereas Bip, which regulates IRE1 (62, 63), is present in the ER lumen and does not interact with CFTR (24). Recent experiments (37) have similarly shown that overexpression of CFTR also has no effect on ER stress of Calu-3 cells. Expression of both CFTR and F508CFTR in vivo will likely occur at lower levels than those induced by the adenoviruses. Thapsigargin activation of IRE1, which activates NF-B signaling (63), could explain the elevated NF-B activity in cells treated with flagellin + thapsigargin.
It might have been expected if ER handling of Ca2+ had been altered in CF that baseline Cai would have been increased, and responses to ATP would have been altered (28, 58). However, CFTR expression did not alter either baseline Cai or responses to ATP. Possible explanations for the differences in the present vs. previous data (58) is use of stable cell lines vs. adenoviral expression or differences in the Cai probes fluo 3 (nonratiometric; see Ref. 58) vs. fura 2 (ratiometric; present experiments). The lack of Cai responses to flagellin and PA in CF15 (13) and 9HTEo– and 16HBE14o– (25) cells compared with 1HAEo– and 16HBE cells (1, 32, 38) may represent cell type differences. Whatever the explanation for differences, it is clear that increases in Cai are not required for airway epithelia to respond to PA or flagellin, although increases in Cai amplify responses to flagellin and PA (13).
In addition to there being no effect of CFTR in controlling Cai signaling, there were no CFTR-dependent differences in NF-B responses to flagellin + agonists known to raise Cai (ATP and thapsigargin), even under conditions in which CFTR should have been maximally activated (forskolin). An implication of these studies is that any CFTR-dependent effects to depolarize membrane potential (28) were small enough that they did not alter Ca2+ entry in the cells.
Forskolin decreases flagellin-activated NF-B.
Forskolin reduced NF-B responses to flagellin + ATP or flagellin + thapsigargin in CF15 and Calu-3 cells, indicating that increases in cAMP and protein kinase A (PKA) reduce inflammatory signaling in airway epithelia. This negative regulation of flagellin- and TLR 5 (65) signaling is consistent with previous experiments in lymphocytes showing that TLR 4-activated signaling was reduced by agonists that activated PKA (32). An implication of these and previous (13) results is that the responses of airway epithelial cells to PA and flagellin (activate TLR 5) will be increased by agonists that raise Cai (e.g., purinergic agonists and bradykinin) and decreased by agonists that raise cAMP (e.g., vasoactive intestinal peptide, adrenergic agonists and PGE2, and P. aeruginosa exotoxin Y).
Hyperinflammation in CF?
Our results demonstrated that expression of either CFTR or F508CFTR in CF15 cells had no effects on ER stress, Cai, or inflammatory responses. We acknowledge that the different results here compared with others who show exaggerated innate immune responses in CF may result from our using a CF cell line that fails to reproduce the complexities of the in vivo phenotype. However, it seems more likely that other technical problems may explain the exaggerated inflammatory responses in previous experiments. For example, it is difficult to appropriately match stable CF and non-CF primary cells or CF vs. CFTR-corrected cell lines. Ribeiro et al. (41) found persistent changes in morphology and Cai signaling in primary CF cells that disappeared after prolonged culture, indicating that these effects resulted from prior exposure of the CF cells to the inflamed CF lung environment rather than to the absence of CFTR. Consistent with this idea and with the findings here, there are now several studies on primary cells (2, 4, 64) showing that, with careful matching, there are only minor CFTR-dependent differences in inflammatory gene expression and cytokine secretion. Different responses between stable CF vs. CFTR-corrected cell lines may result from undetected genetic differences unrelated to CFTR. The apparent proinflammatory effects of 6-day treatment of non-CF cells with CFTR blocker (35) could result from long-term effects of the blocker unrelated to channel function. We have found that the cell-permeant CFTR blocker CFTRinh-172 (35) caused inconsistent increases in NF-B activation in Calu-3 cells, whereas the cell-impermeant CFTR blocker GlyH101 decreased NF-B activity (Fu and Machen, unpublished observations). It is also difficult to make precise matches of cell growth and confluence in cell lines, both of which can affect bacterial binding (7, 19, 26, 46, 53) and perhaps inflammatory responses.
Overall, our results indicate that exaggerated innate immune responses in CF airways in vivo result not from the epithelial cells being hyperinflammatory due to ER stress and exuberant Cai and NF-B signaling (12) but more likely from the absence of CFTR's Cl– and HCO3– (18, 27, 42) conductance properties and consequent effects to reduce ASL volume and change its composition (22, 58). These changes in ASL volume and composition in CF impair the mucociliary escalator and cough from eliminating PA. According to this view, CF airway epithelia respond appropriately to the large number of accumulated bacteria by producing large amounts of cytokines and recruiting large numbers of neutrophils to the airways.
|
GRANTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
