NHBE cells (Cambrex, Walkersville, MD) were grown on Transwell membranes as described previously 35. Media was changed every other day until the cells reached confluence, at which time the apical medium was removed to establish an ALI. Thereafter, the basolateral medium was changed daily. All experimentation was carried out on day 7–9 after ALI establishment. At this point, mature secretory cells are present in these differentiating cultures and the cells respond with maximal proliferation to IL-13 (10 ng/ml) as determined previously 6. Concentrations of TGFα (5 ng/ml) and neutralizing antibodies (0.2 μg/ml) used were based on studies utilizing similar compounds in NHBE cells (6; X Fu and LD Martin, unpublished results). A range of concentrations of rhADAM17 (50 - 0.1 ng/ml) as well as TIMP1 and TIMP3 (100 - 0.5 μg/ml; R&D Systems, Minneapolis, MN) were examined for effectiveness in modulating IL-13-induced proliferation or TGFα shedding in NHBE cells. The lowest possible concentrations that yielded repeatable results with little impact on constitutive growth or growth factor release were used for final experiments in this study. All experiments were repeated a minimum of three times using cells from at least two human donors (except the RT-PCR which was done once). One representative experiment is shown in each Figure.

Following experimental treatments, media samples were collected and analyzed with commercially-available TGFα or IL-8 ELISA kits according to manufacturer's instructions (R&D Systems, Minneapolis, MN).

[³H]-thymidine incorporation assays were performed as described previously 6. Cells were exposed for 24 hrs to IL-13 (10 ng/ml) and/or specific reagents as described. To perform manual cell counts, NHBE cells were liberated from the Transwell membranes with warm Versene (Invitrogen, Grand Island, NY) for 5–10 min at 37°C and counted using a hemacytometer.

Antisense oligonucleotides were utilized following a protocol modified from Li et al 36. Briefly, NHBE cells in ALI culture were exposed to varying concentrations of antisense oligonucleotides directed against ADAM17, scrambled oligonucleotides as a control, or transfection reagent alone (FuGene6; Roche, Indianapolis, IN). All cells were treated for 3 days with the oligonucleotides, with FuGene6 added only on the first day at the manufacturer's suggested concentration. On the third day, the cells were exposed to IL-13, media (control) or TGFα for 24 hrs, with media samples collected and cells counted. Phosphorothioate-modified oligonucleotides were synthesized by Invitrogen (Rockville, MD). ADAM17 antisense sequence was 5'-CCG CCT CAT GTT CCC GT-3' [Genbank: NM_003183]. The scrambled sequence was 5'-TGC GCC ATC TCG CTC TC-3'.

Total protein was extracted from NHBE cells using RIPA buffer containing Roche Complete protease inhibitor cocktail (1 mM EDTA; 1% NP-40; 0.5% sodium deoxycholate, 0.1% SDS, 30 μg/ml pancreas extract, 3 μg/ml pronase, 0.8 μg/ml thermolysin, 1.5 μg/ml chymotrypsin, 0.2 μg/ml trypsin, and 1.0 mg/ml papain). These lysates were incubated overnight with primary antibody at 4°C with shaking. A 50% slurry of Protein A was then added and incubated for 3 hrs. The resulting pellet was washed 5 times in buffer and mixed 1:1 with 2× SDS gel loading buffer (100 mM Tris-Cl, pH 6.8; 4% SDS, 0.2% bromophenol blue, 20% glycerol, 200 mM β-mercaptoethanol). Western analysis was then performed.

Total protein in 2× SDS gel loading buffer was boiled for 5 min, and separated via SDS-PAGE on 10–20% precast gradient gels (Bio-Rad, Hercules, CA). Proteins were transferred to a nitrocellulose membrane (Bio-Rad, Hercules, CA) that was then blocked in 5% nonfat milk/PBS for 1 hr at room temperature. Membranes were hybridized with primary anti-ADAM17 antibody (R&D Systems, Minneapolis, MN) at a concentration of 1:1000 in 5% nonfat milk/PBS overnight at 4°C. The membranes were then washed twice (30 min each) with 0.01% Tween-20/PBS at room temperature. After the second wash, the membrane was exposed to HRP-conjugated secondary antibody diluted 1:5000 in 5% nonfat milk/PBS for 45 min at room temperature. Washes were repeated and bands visualized with ECL (Amersham, Buckinghamshire, UK). The blots were stripped using a commercially available kit (Chemicon International, Temecula, CA) and then rehybridized with an anti-actin primary antibody (Santa Cruz Biotechnology, Santa Cruz, CA) to verify equal protein loading.

Total RNA was extracted from NHBE cells with TRI Reagent (Sigma, St. Louis, MO) and reverse transcribed using specific oligonucleotides and the First Strand cDNA Synthesis Kit for RT-PCR (AMV) (Roche, Indianapolis, IN) in accordance with manufacturer's guidelines. Effort was made to use the amount of cDNA in each PCR that would provide a product in the linear range of the reaction in 35 cycles. PCR reactions were carried out using Taq polymerase (Boeringher Mannheim, Mannheim, Germany) in a Perkin Elmer GenAmp PCR System 2400. PCR products were separated by electrophoresis through a 2% agarose gel and visualized by staining with ethidium bromide. Primers used were ADAM17 forward-ACCTGAAGAGCTTGTTCATCGAG, ADAM17 reverse-CCATGAAGTGTTCCGATAGATGTC [Genbank: NM_003183]; β-actin forward-TCGACAACGGCTCCGGCA, β-actin reverse-CGTACATGGCTGGGGTGT [Genbank: BC014861].

At each time point, 2 control cultures were exposed to media and 2 experimental cultures to IL-13 (10 ng/ml). Following treatment, the NHBE cells were fixed on the Transwell inserts with 4% formalin. All staining was carried out in the Transwell inserts. Cells were washed with PBS, permeabilized with 0.2% Triton X-100 in PBS, and reacted with primary antibodies, either anti-TGFα or anti-ADAM17, followed by a 45 min incubation in the dark with appropriate secondary antibodies tagged with Alexa 488 (for use with anti-ADAM17) or Alexa 594 (for use with anti-TGFα) (Molecular Probes, Eugene, OR). Membranes containing the cells were then removed from the Transwell inserts and mounted on glass slides in Vectashield mounting media (Vector Laboratories, Burlingame, CA). Cells were visualized with a Nikon Eclipse TE2000-E confocal microscope via a Plan Apo 60× water immersion objective. The entire experiment, from cell growth through microscopy, was repeated 3 times, resulting in 6 samples per experimental and 6 samples per control, time point. Each sample was divided into quadrants and 250 to 300 cells per quadrant were examined qualitatively to gain a general understanding of the expression patterns at each time point.

Six to nine scans per control or experimental time point were chosen randomly from the captured Z-stack confocal microscopy images. Five to 10 cells per scan were examined. Three areas [apical/middle/basal] within each cell were inspected to determine whether more TGFα or more ADAM17 was expressed in each area. The Z-stack images had been generated using a constant Z-stack interval. In each Z-stack, the first image was from just above the transwell membrane at the basal cellular surface and the last image was at the cell's apical surface. Thus, "apical" and "basal" refer to the apical-most and basal-most images in the Z-stack from a single cell, while "middle" is defined as the image halfway between the apical-most and basal-most images from a single cell. With examination of approximately 100 cells (50 control and 50 experimental) per time point, about 97% of the cells were found to have essentially two expression patterns [apical/middle/basal]: [TGFα/TGFα/ADAM17] or [ADAM17/ADAM17/TGFα ]. Using only these 97% of cells, final percentages of cells exhibiting each pattern were calculated.

Experimental data were analyzed for significance by one-way analysis of variance (ANOVA), with post-test correction for multiple comparisons where appropriate. Differences between treatments were considered significant at p < 0.05. Data are shown as mean ± standard error of the mean (SEM).

Research from our laboratory indicates that IL-13 initiates proliferation of NHBE cells via a TGFα/EGFR (epidermal growth factor receptor) autocrine/paracrine growth loop 6. Since ADAM17 is known to cleave membrane-inserted pro-TGFα to its mature form in a number of cell types 253738, we determined whether ADAM17 could act similarly in NHBE cells to mediate proliferation in a TGFα-dependent manner. Treatment of NHBE cells with exogenous recombinant human (rh) ADAM17 resulted in an increase of soluble TGFα in the surrounding medium (Fig. 1a). ADAM17 also induced cellular proliferation as did IL-13 and TGFα (Fig. 1b). These results indicate that NHBE cells express TGFα on the extracellular membrane in a form that is amenable to proteolytic cleavage by ADAM17. Next we determined if the proliferation observed following exposure to rhADAM17 was occurring via cleavage of surface expressed TGFα, rather than via cleavage of another growth factor. The addition of neutralizing anti-TGFα antibody attenuated the proliferative effect induced by exogenous rhADAM17 (Fig. 1c) suggesting that rhADAM17 cleaves surface-expressed TGFα, that in turn induces proliferation of the epithelial cells.

After determining that exogenous ADAM17 can induce cellular proliferation mediated by TGFα in NHBE cells, we determined whether endogenous ADAM17 is involved in IL-13-induced proliferation of these cells. First, the effects of various inhibitors of ADAM17 on IL-13-induced shedding of TGFα were examined. Tissue inhibitor of metalloproteinase (TIMP)-3 is a documented inhibitor of ADAM17 3940, while a related family member, TIMP-1, has been found to have no effect on ADAM17 41. Furthermore, the differential inhibition of ADAM17 by the two TIMPs is useful to distinguish the action of ADAM17 from that of ADAM10, whose activity can be inhibited by both TIMP-3 and TIMP-1 41. In the current study, TIMP-3 was found to attenuate IL-13-induced TGFα shedding, while TIMP-1 did not have an inhibitory effect (Fig. 2a). Additionally, anti-ADAM17 antibodies also blocked IL-13-induced TGFα shedding (Fig. 2b). Thus, these data support the role of ADAM17 in mediating IL-13-induced TGFα shedding in NHBE cells.

To confirm the requirement of ADAM17 in mediating IL-13-induced TGFα shedding, and to determine whether ADAM17 is similarly required for IL-13-induced NHBE cell proliferation, cells were exposed to antisense oligonucleotides directed against ADAM17 or to scrambled oligonucleotides as a control. Scrambled oligonucleotides had little effect on ADAM17 expression in a culture exposed to media and in another culture exposed to IL-13; however, in the same experiment, decreased expression of ADAM17 was easily discernible in comparable cultures exposed to antisense oligonucleotides directed against the protease (Fig. 3a). In cultures similarly exposed in this same experiment, ADAM17 antisense oligonucleotides inhibited IL-13-induced NHBE cell proliferation (Fig. 3b) and inhibited IL-13-induced, as well as constitutive, shedding of TGFα (Fig. 3c). ADAM17 antisense oligonucleotides, however, did not inhibit TGFα-induced proliferation (Fig. 3b). In all experiments, scrambled oligonucleotides had no significant effect on growth of control cells or on their constitutive release of TGFα (Figs. 3b and 3c). Furthermore, while the presence of scrambled or ADAM17 antisense oligonucleotides reduced the maximal level of proliferation inducible by IL-13 or TGFα, only the ADAM17 antisense oligonucleotides were capable of blocking IL-13-induced proliferation with specificity, as these oligonucleotides had no effect on TGFα-induced proliferation (Fig. 3b). Taken together, these results support the requirement of endogenous ADAM17 for IL-13-induced proliferation of NHBE cells, and confirm that ADAM17 plays a role in the shedding of TGFα in NHBE cells.

Since ADAM17 appeared to mediate IL-13-induced TGFα shedding and proliferation in NHBE cells, we wanted to determine whether these effects were due to a simple IL-13-induced increase in ADAM17, or its activity. The amount of steady-state mRNA coding for ADAM17 in control or IL-13-treated cells was found to be the same following 4 or 24 hrs of treatment (Fig. 4a). Next the amount of ADAM17 protein was examined. This protein exists in two forms, an inactive, latent form and an active form 32. Conversion to the active form requires proteolytic cleavage of the enzyme, resulting in removal of a 20-kDa section of the protein. The amount of latent ADAM17 in NHBE cells varied little in response to control media or IL-13 over a time course of 5 min to 24 hrs (Fig. 4b). The amount of active ADAM17 in control cells during this time period also varied little, while slightly less active ADAM17 was observed at early time points in IL-13-treated cells. The amount of active ADAM17 in these treated cells, however, was similar to control levels at the latter time points (1 to 24 hrs) (Fig. 4b). Thus, while IL-13 may induce a small, transient decrease in the amount of active ADAM17, the quantity of active protein is no greater than that observed in control cells at time points when IL-13 induces an increase in soluble TGFα (i.e. approximately 60 min in this study (Fig. 4c), and as early as 15 min in a previous study 6). These data show that IL-13 does not induce a dramatic alteration in the amount of ADAM17 mRNA, latent ADAM17, or active ADAM17 in NHBE cells.

Since activation of ADAM17 and ADAM17-mediated shedding can be induced via PKC stimulation 2642, we tried to enhance the shedding of TGFα by exposing NHBE cells to phorbol-12-myristate 13-acetate (PMA), a known activator of PKC and well-characterized inducer of TGFα ectodomain shedding 43, at a concentration shown previously to enhance TGFα shedding in a pulmonary mucoepidermoid carcinoma cell line (NCI-H292) 30. Exposure of the NHBE cells to PMA, however, did not yield an increase in soluble TGFα (Fig. 4c) or cellular proliferation (Fig. 4d), even though IL-13 could still induce these events. The NHBE cells did respond to the PMA, however, as secretion of IL-8, a process known to be PKC-dependent in NHBE cells 44, was enhanced while IL-13 had no effect on IL-8 secretion (Fig. 4e). Thus, these results suggest that the mechanism mediating IL-13-induced release of soluble TGFα from NHBE cells differs from the PKC-mediated mechanism responsible for TGFα shedding in NCI-H292 cells, an event which appears to involve direct activation of ADAM17 by PKC 30. Thus, it appears that although the IL-13-induced increase in TGFα shedding, as well as the IL-13-induced proliferation, is mediated by ADAM17 in NHBE cells, these events do not occur solely via an IL-13-induced increase in ADAM17 or its activity.

An alternate mechanism whereby IL-13 could increase the amount of TGFα shed from NHBE cells would be for the cytokine to promote the release of pre-formed, intracellular growth factor. NHBE cells are already known to release pre-formed mucin proteins (the glycoprotein component of airway mucus) upon stimulation with various inflammatory mediators 3645. Under such conditions, granules containing the mucin proteins are thought to be mobilized rapidly to the cell surface where the proteins are secreted 36. To determine whether a similar mechanism mediates IL-13-induced release of TGFα, confocal microscopy was used to examine the location of TGFα and its sheddase, ADAM17, in NHBE cells exposed to IL-13 or control media over a 4-hr time course. (Quantitative results from this study are shown in Table 1.)

Untreated NHBE cells (data not shown), or NHBE cells exposed only to control media (Figs. 5 and 6a; Table 1), were found to express TGFα and ADAM17 constitutively. The majority of the growth factor (TGFα) was localized to the interior of the epithelial cells, with ample expression observed in the basal and central regions of the cells. Little expression of TGFα was observed in apical cellular regions. By contrast, ADAM17 was expressed throughout the cytoplasm, although the majority of this enzyme was expressed in portions of the cytoplasm adjacent to the cell membrane, with expression particularly prominent in the apical region of the epithelial cells. In fact, about 80% of control cells exhibited this pattern of expression which remained relatively unchanged as NHBE cells were exposed to fresh media for 15 min, 30 min, 1 hr, or 4 hrs (Figs. 5 and 6a; Table 1). More precisely, the percentage of media-exposed, control cells exhibiting this expression pattern (TGFα interior with ADAM17 highly expressed in the apical region) at these time points was 81%, 82%, 71%, and 88%, respectively (see Table 1). The cross-section and Z-stack video images of the control cells (Fig. 6a and Additional file 1, respectively) as well as an illustration of a control cell (Fig. 6b) summarize the observed location of TGFα (red) and ADAM17 (green) in cells without IL-13 stimulation.

While exposure of NHBE cells to IL-13 for 15 min did not alter the location of TGFα expression compared to its location within control cells (Fig. 5), continued exposure to IL-13 for 30 min or more did induce an alteration in the location of TGFα expression. Specifically, at 30 min, patches of TGFα were less defined within the cytoplasm, with almost no TGFα expression detectable in the basal areas of IL-13-exposed cells. Rather, the majority of the growth factor was expressed in the apical region and on the apical surface of the NHBE cells (Fig. 5). This pattern of apical TGFα localization was observed in 46% of the IL-13-treated cells compared to just 18% of the control cells (Table 1). While IL-13 induced increased apical localization of TGFα, apical localization of ADAM17 was observed in fewer cells (54% compared to 82% of control cells) following IL-13 exposure, with the enzyme now found to a greater extent in the middle and basal regions of the NHBE cells. Thus, it would appear that when NHBE cells are exposed to IL-13, localization of TGFα shifts to the apical region of these cells within 15 to 30 min. Such a finding would be consistent with the movement of TGFα from its intracellular region of constitutive expression (middle and basal) into the apical region of these cells, a region where prominent ADAM17 expression is observed constitutively.

Following exposure of NHBE cells to IL-13 for 60 min, the expression patterns of both TGFα and ADAM17 remained similar to those observed in cells exposed to IL-13 for 30 min (Fig. 6a; Table 1; see Additional file 2); more treated cells expressed TGFα in their apical regions (35% compared to 29% of control cells) while fewer treated cells expressed ADAM17 apically (65% compared to 71% of control cells). However, the percentage of affected cells appeared somewhat intermediate between the 15 min and the 30 min-treated values. This finding may suggest that the TGFα, whose apical expression was induced by IL-13, is beginning to be cleaved from the cell, while ADAM17 is being internalized.

Following a 4-hr exposure to IL-13, little TGFα remains within most of the NHBE cells. In fact, 98% of the treated cells, compared to 88% of the control cells, express mainly ADAM17 with little to no TGFα expression found at any level within the cells. The relatively small amount of growth factor that is present appears to be expressed in the intracellular regions where TGFα was maintained prior to stimulation (middle or basal region of the cells). Conversely, more control cells (12%) express TGFα in their apical regions compared to IL-13-treated cells (2%). This dramatic shift from 35 – 46% of IL-13-treated cells expressing TGFα apically at 30 – 60 min, to just 2% of the treated cells expressing it at 4 hrs, is consistent with the apical TGFα being cleaved and released from the cells.

Taken together, the confocal images (examples provided in Figs. 5 and 6a) and quantitative analysis (Table 1) of TGFα and ADAM17 expression in NHBE cells support the conclusion that IL-13 can induce movement of a stored growth factor (TGFα) from the central and basal cytoplasmic regions to the apical region of airway epithelial cells, where it is cleaved by ADAM17. Fig. 6b illustrates the timing of this inducible translocation, with an increase in TGFα near the apical surface observed by 30 – 60 min of IL-13 exposure, with the growth factor co-localizing with ADAM17 in this region. By 4 hrs of IL-13 exposure, very little TGFα is observed within the cells, likely due to its being cleaved from the apical surface by ADAM17, following its IL-13-induced translocation.