Transforming growth factor-β1 (TGF-β1) belongs to a family of multifunctional cytokines that regulate a variety of biological processes, including cell differentiation, proliferation, and apoptosis. The effects of TGF-β1 are cell context and cell cycle specific and may be signaled through several pathways. We examined the effect of TGF-β1 on apoptosis of primary human central airway epithelial cells and cell lines. TGF-β1 protected human airway epithelial cells from apoptosis induced by either activation of the Fas death receptor (CD95) or by corticosteroids. This protective effect was blocked by inhibition of the Smad pathway via overexpression of inhibitory Smad7. The protective effect is associated with an increase in the cyclin-dependent kinase inhibitor p21 and was blocked by the overexpression of key gatekeeper cyclins for the G1/S interface, cyclins D1 and E. Blockade of the Smad pathway by overexpression of the inhibitory Smad7 permitted demonstration of a TGF-β-mediated proapoptotic pathway. This proapoptotic effect was blocked by inhibition of the p38 MAPK kinase signaling with the inhibitor SB-203580 and was associated with an increase in p38 activity as measured by a kinase assay. Here we demonstrate dual signaling pathways involving TGF-β1, an antiapoptotic pathway mediated by the Smad pathway involving p21, and an apoptosis-permissive pathway mediated in part by p38 MAPK.
- cyclin-dependent kinase
- cyclin-dependent kinase inhibitor
- extracellular regulated kinase
- transforming growth factor-β receptor
- airway epithelium
TRANSFORMING GROWTH FACTOR-B1 (TGF-β1) belongs to a family of multifunctional cytokines that regulate a variety of biological processes including cell differentiation, proliferation, and apoptosis (reviewed in Refs. 11, 56). TGF-β1 initiates signaling by binding to a heteromeric complex of transmembrane serine/threonine kinase receptors (TβR). Receptor binding induces phosphorylation of receptor Smads (Smad2 and Smad3) that then form heteromeric complexes with Smad4, a common mediator Smad. This complex translocates to the nucleus and regulates the transcription of target genes by binding to regulatory elements. Inhibitory Smads such as Smad6 and Smad7 block Smad signaling by preventing the phosphorylation of receptor Smads.
Independently of the Smad pathway, TGF-β1 also may signal via mitogen-activated protein kinase (MAPK) pathways (reviewed in Refs. 38, 67). One of these, the p38 MAPK pathway, is activated in response to chemical and physical stresses and modulates cell survival, inflammation, and proliferation. Activation of TGF-β-activated kinase kinase, a MAPK kinase kinase (75), can activate MAPKs such as p38 (20, 41) and has been shown to be involved in TGF-β-induced apoptosis.
Among the several effects of TGF-β1 is regulation of the cell cycle in the G1 phase (38). This effect is mediated through several mechanisms and is cell type dependent. Treatment with TGF-β1 elicits growth arrest in many epithelial cell lines (10, 30) principally through increased expression and/or stabilization of cyclin-dependent kinase (CDK) inhibitors (CKI) (38, 52), which in turn downregulate serine/threonine CDKs that interact with cyclins D1 and E (52, 54). TGF-β1 also inhibits the synthesis of cyclin D1 and decreases abundance of cyclin D1/CDK4 complexes in rat intestinal epithelial cells (33). Overexpression of cyclin D1 drives cells through G1 phase in many cell systems, particularly when emerging from G0 (19, 72). In a similar manner, constitutive expression of CDK4 in Mv1Lu mink lung epithelial cells confers resistance to TGF-β-induced G1 arrest (17). Growth arrest does not necessarily confer protection from apoptosis: in many instances growth arrest and apoptosis go hand in hand. TGF-β can elicit both growth arrest and apoptosis in B lymphocytes (7, 31, 48), as well as in cervical (57), uterine (71), pancreatic (65), and retinal epithelial cells (16). This effect is not uniform, however, as TGF-β1 can also suppress apoptosis both in hemapoietic cell lines (60) and in structural cells such as microglia (62). TGF-β1 protects mouse mammary epithelial cells from Forkhead response element or serum starvation-induced apoptosis (64) and myofibroblasts from IL-1β-induced apoptosis (80). The protective effect of TGF-β, much like the effects on growth inhibition and cell death, are context and cell type specific.
One clinical situation in which TGF-β may have a role is asthma, due to its protean effects on T cell function, eosinophil activation, fibroblast-mediated matrix protein production, and airway smooth muscle cell proliferation (reviewed in Ref. 27). Concentrations of the active form of TGF-β1 are higher in bronchoalveolar lavage fluid of patients with severe asthma compared with normal subjects (68). Basal TGF-β1 levels are elevated in airways of atopic asthmatics and increase further after allergen exposure (55). One potential function of TGF-β in airways is regulation of epithelial cell survival and differentiation. Injury to airway epithelium is common in asthma, even when the clinical state of disease is mild (2, 34). One possible mechanism by which epithelial damage occurs is via apoptosis, which is increased in asthmatic airway epithelium (3, 69). A substantially higher number of epithelial cells express TGF-β1 in asthmatic airways compared with airways from normal subjects (70). However, in some settings TGF-β may elicit airway epithelial cell death. TGF-β1 treatment can elicit apoptosis in the BEAS-2B airway epithelial cell line only after overexpression of either Smad2 or Smad3 (77). TGF-β1 treatment can enhance Fas-induced killing of small airway and alveolar epithelium (22), and Fas-mediated killing of alveolar epithelial cells and development of fibrosis are associated with increased expression of TGF-β mRNA (23). These studies suggest that increased TGF-β production in asthmatic airways may be deleterious to epithelial integrity.
To examine the role of TGF-β1 in central airway epithelial cells, we tested whether this growth factor would enhance or inhibit apoptosis induced by either Fas activation or corticosteroids. Our data demonstrate, in contrast to the effect of this cytokine on small airway epithelial cell survival, that TGF-β1 protects central airway epithelial cells from apoptosis. This protective effect is mediated by the Smad pathway and is associated with an increase in p21 CKI signaling. We also demonstrate a proapoptotic pathway mediated by p38 MAPK signaling. Our data suggest that TGF-β1 may have a dual role in signaling both pro- and antiapoptotic events in central airway epithelium.
MATERIALS AND METHODS
Streptomycin, l-glutamine, penicillin-streptomycin, dexamethasone, recombinant human TGF-β1, and antiactin monoclonal antibody (MAb) were purchased from Sigma (St. Louis, MO). T-Rex system, blasticidin, zeocin, tetracycline (Tet), geneticin, and subcloning efficiency DH5α competent cells were purchased from Invitrogen (Carlsbad, CA). Anti-Fas MAb (CH11) was purchased from MBL International (Watertown, MA). Anti-Fas MAb (clone ZB4) for flow cytometry was purchased from PanVera (Madison, WI). Anticyclin D1 MAb was purchased from Zymed Laboratories (South San Francisco, CA). Anticyclin E MAb was purchased from Upstate Biotechnology (Lake Placid, NY). Anti-p21 (F-5) MAb, anti-p27 (C-19) MAb, anti-CDK2 (M2) MAb, and anti-CDK4 (C-22) MAb were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). SB-203580, a p38 MAPK inhibitor, and SP-600125, a c-Jun NH2-terminal kinase (JNK) inhibitor, were purchased from CalBiochem (San Diego, CA). p38 activity assay kit was purchased from Cell Signaling (Beverly, MA). Human BEGM medium was purchased from Clonetics (Walkersville, MD). Fetal calf serum (FCS) was purchased from Hyclone (Logan, UT) and was heat inactivated before use. Terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate biotin nick end-labeling (TUNEL) TACS II fluorescent assay kits were purchased from Trevigen (Rockville, MD). Super Signal West Fempto Maximum Sensitivity Substrate was purchased from Pierce (Rockford, IL). All other reagents were obtained from Sigma and were of the highest quality available.
The use of human airways was approved by the University of Chicago Institutional Review Board. Primary airway epithelial cells were harvested from transplant donor lungs designated for research as described previously (15). These cells were grown in BEGM medium (Clonetics) containing 5 μg/ml insulin, 0.5 μg/ml human EGF, 10 mg/ml transferrin, 6.5 μg/ml triiodothyrinine, 0.5 mg/ml hydrocortisone, 0.5 mg/ml epinephrine, and 13 mg/ml bovine pituitary extract. Cells were used at passage 1. The cell lines 1HAEo− and 16HBE14o− were obtained from Dieter Gruenert (University of Vermont, Burlington, VT) and are SV40-transformed human central airway epithelial cells (21) that have cell surface markers similar to primary airway basal epithelial cells (14). NCI-H292 cells, an airway epithelial cell line derived from a lung mucoid adenocarcinoma (1, 8), were obtained from American Type Culture Cell lines were grown on collagen IV-coated chamber slides in Dulbecco's modified essential medium containing 10% FCS, 2 mM l-glutamine, 100 μg/ml streptomycin, and 100 U/ml penicillin G. All cells were used when 80–90% confluent. Cells were kept in 10% FCS during all experiments to prevent confounding of apoptosis results by withdrawal of needed growth factors (15). At the conclusion of experiments, chamber slides were washed once in fresh medium and fixed in 10% neutral buffered formalin.
The expression vector pCMV containing Smad7 as well as the empty pCMV vector were generously provided by Blanca Camoretti-Mercado (University of Chicago). The expression vector pRC/CMV containing cyclin E was generously provided by Richard Pestell (Albert Einstein College of Medicine). The expression vector pRC/Rous Sarcoma Virus (RSV) containing cyclin D1 was generously provided by Charles Sherr (St. Jude Children's Research Hospital). The expression vectors pcDNA4/TO and pcDNA6/TR were purchased from Invitrogen (Carlsbad, CA) as part of the T-Rex System, a Tet-regulated expression system for mammalian cells.
Cells were transfected to overexpress either full-length Smad7 or empty vector (EV) control by methods described previously (15). Subclones were selected after six passages on the basis of gene expression on Northern blot assay with the transfected gene as a probe. Transfected cells were maintained in 300 μg/ml geneticin until use. We created the Tet-inducible expression vectors with the T-Rex System according to the manufacturer's directions. We created the pcDNA4/TO vector containing cyclin D1 by excising the cyclin D1 cDNA from the PRC/RSV vector and inserting it into the pcDNA4/TO vector with HindIII and XbaI. We created the pcDNA4/TO vector containing cyclin E by excising the cyclin E cDNA from the PRC/CMV vector and inserting it into the pcDNA4/TO vector with KpnI and XbaI. Insertion of the fragment and correct orientation were confirmed by plasmid DNA sequencing. The pcDNA4/TO EV, in addition to the vectors containing cyclin D1 and E, was individually cotransfected with the pcDNA6/TR regulatory vector in a 1:6 ratio into 1HAEo− cells grown to 80% confluence (∼500,000 cells) with 1.2 μg total DNA, 15 μl lipofectamine, and 1,200 μl OptiMEM in six-well plates for 5 h at 37°C followed by replacement with fresh media. After 48 h, cells were selected with 7.5 μg/ml blasticidin and 50 μg/ml zeocin. Subclones as well as determination of the appropriate time needed to induce gene expression with Tet was selected based on cyclin D1/cyclin E expression by Northern blot with appropriate cDNA from each plasmid as a probe and by Western blot analysis with appropriate antibodies. The EV subclone was selected based on its ability to survive in blasticidin and zeocin. Tet-screened medium was used for these cell lines. Transfected cells were treated for 16 h with Tet before the start of experiments to induce gene expression and continued to receive Tet for the duration of the experiment.
Assay for cell DNA nicking.
We have described this method previously (15). Apoptotic cells in fixed monolayers were demonstrated by labeling free 3′-hydroxyl groups of DNA with a Trevigen TUNEL fluorescent assay kit and then stained with 1 mM Hoechst 33258 in 95% ethanol for 45 s. Representative images were collected with a 12-bit cooled charge-coupled device camera (Photometrics, Tucson, AZ) connected to a Nikon fluorescence microscope. One investigator (not the same investigator who performed the experiment) selected fields at random. For each slide, we approximately centered the well on the microscope stage under the objective without viewing the field through the eyepiece. Images then were collected in registration, and the microscope stage was moved a short distance in a random direction without observation through the eyepiece, and images again were collected. Obviously inappropriate images (e.g., noncellular debris or too few cells to count) were discarded, and the stage was moved again in a random direction if required for additional images. In this manner, observer bias was minimized. TUNEL-positive nuclei and Hoechst-stained nuclei were counted in each image as the area of the nuclei in pixels after visual thresholding and exclusion of extraneous positive pixels with Spectrum IP software (IP Labs, Vienna, VA) on a Macintosh computer. TUNEL-positive cells were expressed as the percentage of the thresholded area of the TUNEL-stained image divided by the thresholded area of the Hoechst-stained image. The TUNEL counts of two fields in the same well were averaged to produce a single n. Previous experiments (74) demonstrated a high correlation with manual counting, demonstrated that changes in cell shape or morphology alone did not significantly alter the ability to detect apoptotic nuclei, and confirmed that TUNEL-positive cells did not have the morphological features of necrosis, which may also lead to single-strand DNA nicking.
Northern blot analysis.
We have described this method previously (15). Hybridization probes for Smad7, cyclin D1, and cyclin E were generated as gel-purified restriction fragments from plasmid and labeled with Redivue [γ-32P]CTP with Rediprime II (Amersham-Pharmacia Biotech, Buckinghamshire, UK).
Western blot analysis.
We have described this method previously (15). Membranes were reprobed with an antibody for actin when appropriate to control for differences in protein loading.
p38 activity assay.
Cell lysates were collected and processed with a commercial kit (Cell Signaling) according to directions. Western blots were prepared to resolve the phospho-activating transcription factor (ATF)-2 product.
We have described this method previously (24). After interventions, ∼500,000 cells were washed and suspended in cold PBS containing 0.5% BSA and 0.2% NaN3 and blocked with 5% human serum at 4°C for 10 min. Cells were incubated for 30 min at 4°C with 2 μg/ml anti-Fas MAb (clone ZB4, PanVera) in saturating concentration diluted in fluorescence-activated cell sorting (FACS) buffer. FITC-conjugated goat anti-mouse IgG F(ab′) fragments (Becton-Dickinson) then were added at a saturating concentration in cold FACS buffer. Quantification of specific immunofluorescence was carried out with a FACScan 400 (Becton Dickinson). Dead cells were excluded by positive propidium iodide staining. Nonspecific binding was examined with isotype IgG of irrelevant antigen specificity.
TUNEL data are expressed as means ± SE. Differences were examined by analysis of variance; when significant differences were found, post hoc analysis was done with Fisher's protected least-significant difference test. P was considered significant when <0.05.
TGF-β1 protects human airway epithelial cells from apoptosis induced by Fas activation or corticosteroids.
Both corticosteroids and Fas activation induce apoptosis in central airway epithelial cells in a concentration- and time-dependent manner (15). We asked whether TGF-β1 modulated apoptosis induced by either stimulus in human central airway epithelial cells. Cells were treated with 1 μg/ml CH11 to activate Fas for 24 h concurrently with increasing concentrations of TGF-β1. Treatment was associated with a decrease in apoptosis in a concentration-dependent manner (Fig. 1, A and B). We then examined whether this protective effect was associated with changes in Fas expression. Cells were treated with TGF-β1, after which surface Fas expression was examined by flow cytometry. TGF-β1 treatment was not associated with changes in Fas surface expression (Fig. 1C). TGF-β1 treatment also attenuated substantially apoptosis elicited by 3 μM dexamethasone (Fig. 2, A and B).
To demonstrate that this effect was not cell line specific, we also confirmed this antiapoptotic effect in the central airway epithelial cell lines 16HBE14o−, an SV40 transformed cell line with basal cell characteristics (14), and NCI-H292, a lung adenocarcinoma cell line with features of mucoid epithelial cells (1, 8). Concurrent treatment with TGF-β1 blocked apoptosis elicited by either Fas activation or dexamethasone (Fig. 3A). To confirm that this effect was not specific to effects related to immortalization of cell lines, we tested the effect of TGF-β1 on primary human bronchial airway epithelial cells. TGF-β1 treatment also prevented apoptosis in primary airway epithelial cells following either Fas activation or dexamethasone treatment as done in the 1HAEo− cell line (Fig. 3B). These data suggest that TGF-β1 was protective for several central airway epithelial cell lines, as well as primary cells.
TGF-β1-mediated protection from apoptosis is mediated by the Smad pathway.
To determine whether the Smad pathway was required for signaling of the antiapoptotic effect, we asked whether overexpression of Smad7, an inhibitory Smad, would block this effect. Smad7 was stably overexpressed in the 1HAEo− cell line (1HAEo−.Smad7+), and overexpression was confirmed by Northern blot (data not shown). In contrast to the protective effect seen with the addition of TGF-β1 to control cells treated with Fas activation, TGF-β1 treatment in 1HAEo−.Smad7+ cells did not reverse apoptosis induced by treatment with the Fas-activating antibody (Fig. 4A). Similarly, TGF-β1 treatment in 1HAEo−.Smad7+ cells did not reverse apoptosis induced by dexamethasone (Fig. 4B). In both cases, apoptosis at the highest concentration of TGF-β1 was higher than in cells treated either with dexamethasone or with Fas activation alone. These data demonstrate that the antiapoptotic effect of TGF-β1 was mediated by the Smad signaling pathway.
TGF-β1 treatment is associated with an increase in p21.
We next examined whether the protective effect is associated with changes in TGF-β1-mediated cell cycle regulators, including CDK2, CDK4, p27, and p21. 1HAEo− cells were treated with CH11 MAb with and without TGF-β1 for up to 24 h. Protein abundance of CDK4, as assessed by Western blot, was not different (Fig. 5A). Protein abundance of p27 decreased over time after Fas activation; this decrease was not altered by concurrent treatment with 10 ng/ml TGF-β1 (Fig. 5B). However, Fas activation decreased p21 abundance over time, which was prevented by concurrent treatment with TGF-β1 (Fig. 5C). CDK2 abundance changed little after Fas activation and was not decreased by concurrent treatment with TGF-β1 (Fig. 5D).
Similar results were seen in CDK4 abundance in cells treated with dexamethasone with and without TGF-β1 (Fig. 6A). Unlike with Fas activation, treatment with dexamethasone did not alter p27 abundance, and abundance of this CKI was not altered by treatment with 10 ng/ml TGF-β1 (Fig. 6B). As with Fas activation, dexamethasone treatment was associated with decreased p21 abundance, which was prevented by concurrent treatment with TGF-β1 (Fig. 6C). CDK2 abundance was little changed in the presence of dexamethasone, but concurrent treatment with TGF-β1 elicited decreased abundance after 8 h (Fig. 6D). Together, these data indicate that p21 abundance decreased in the presence of an apoptotic stimulus; both apoptosis and decreased p21 abundance could be prevented by TGF-β1 treatment.
Overexpression of cyclin D1 and cyclin E reverses the rescue effect.
In cell systems in which TGF-β1 induces apoptosis, concomitant suppression of proliferation and cell cycle also may be observed. This suppression generally is observed at G1 (10, 30). In light of our data showing a protective effect with TGF-β1, we asked whether driving cells through G1 would abolish the protective effect. We therefore overexpressed key gatekeeper cyclins for the G1/S interface, cyclins D1 and E (52, 54). We overexpressed each cyclin with the Tet-inducible vector in the 1HAEo− cell line. Inducible overexpression of each cyclin was confirmed by both Northern (data not shown) and Western blot (Fig. 7A) analyses. After induced overexpression of either cyclin E or cyclin D1 with 1 μg/ml Tet for 16 h, cells were treated concurrently with the Fas-activating CH11 MAb alone or with TGF-β1 for an additional 24 h. In these experiments, the response of the cyclin E- but not cyclin D1-overexpressing cells to Fas activation was less than that seen with EV control cells (Fig. 7B). The addition of TGF-β1 did not protect either cell line from apoptosis elicited by Fas activation (Fig. 7B). Similarly, concurrent treatment with TGF-β1 did not prevent apoptosis elicited by dexamethasone after induction of either cyclin E or cyclin D1 expression (Fig. 7C).
Smad pathway blockade results in TGF-β1-induced apoptosis mediated by p38 MAPK signaling.
Given that the antiapoptotic effect of TGF-β1 was blocked by overexpression of Smad7 and in fact resulted in higher levels of apoptosis, we postulated that a possible apoptotic pathway mediated by TGF-β1 may also be present. Expression of Smad7 alone did not elicit substantial apoptosis: TUNEL labeling in control (EV) 1HAEo− cells and in the 1HAEo−.Smad7+ cell line was <1% under basal conditions. Treatment with 10 ng/ml TGF-β for 24 h elicited apoptosis that was statistically significant but was <3% in all experiments (Fig. 8A).
We then examined the potential role of the p38 MAPK signaling pathway. The 1HAEo−.Smad7+ cell line was treated with 1 μg/ml CH11 MAb to activate Fas for 24 h concurrently with 10 ng/ml TGF-β1 and with either 2 or 20 μM SB-203580, a p38 MAPK inhibitor. Apoptosis was then measured by TUNEL assay. Treatment with 20 μM SB-203580 attenuated apoptosis elicited by Fas activation in the presence of TGF-β (Fig. 8B). However, the p38 inhibitor did not block apoptosis induced by Fas activation alone. These data suggest that the presence of TGF-β1 under conditions of Smad pathway blockade facilitated apoptosis in a p38-dependent manner.
Additional experiments were done with an inhibitor of JNK, SP-600125 (4, 25). In these experiments, the 1HAEo−.Smad7+ cell line was treated with 1 μg/ml CH11 MAb alone or with 10 ng/ml TGF-β and with either 2 or 20 μM SP-600125 for 24 h. Apoptosis was then measured by TUNEL assay. Treatment with the JNK inhibitor did not change apoptosis appreciably compared with cells treated with Fas activation alone or with Fas activation and TGF-β (data not shown).
We then investigated p38 activity after TGF-β1 treatment with a specific activity assay. The 1HAEo−.Smad7+ cell line was treated with 1 μg/ml CH11 MAb alone, 10 ng/ml TGF-β1, or both for 0–120 min, after which kinase activity to phosphorylate ATF-2 (6) was assessed in collected protein lysates. In cells treated with the CH11 MAb to activate Fas, p38 activity was noted within 5 min with peak activity at 30 min (Fig. 8C). In cells treated with TGF-β alone, activity was seen only at a single time point, 15 min (Fig. 8C). In cells treated with both the CH11 MAb and TGF-β1, there was an increase in p38 activity within 5 min, with a peak at 15 min and sustained activity at 4 h (Fig. 8C). These data demonstrated that in cells in which the Smad pathway was blocked by overexpression of Smad7, TGF-β1 potentiated the activation of the p38 MAPK pathway by the Fas pathway.
We performed this study to investigate the potential effect of TGF-β1 on human central airway epithelial cell apoptosis. Apoptosis may be one mechanism by which epithelial injury occurs in asthma, and increased apoptosis in asthmatic airway epithelium has been described (3, 69). Given the substantial production of this growth factor both by airway epithelial cells (28, 37) and by inflammatory cells such as eosinophils that may be present in asthmatic airways (40, 46), we hypothesized that increased TGF-β expression could modulate apoptosis. In agreement with this hypothesis, TGF-β1 treatment reliably inhibited apoptosis induced by either Fas activation or by glucocorticoids in primary airway epithelial cells, in transformed airway epithelial cell lines of basal cell-like phenotype, and in a lung tumor cell line. Our data represent the first demonstration of a protective role of TGF-β1 against apoptotic stimuli in central airway epithelial cells.
Signaling via the Smad pathway is regulated at several points; one such point is inhibition of the pathway by Smad7, itself a downstream target of TGF-β activation, which may enhance apoptosis in several epithelial cell lines (67), by either activation of stress-activated protein kinases or inhibition of the transcription factor NF-κB (35, 39). In accordance with this, overexpression of Smad7 prevented the antiapoptotic effect of TGF-β. Overexpression of Smad7 may itself induce apoptosis in mesangial cells (47), podocytes (61), prostate epithelial cell lines (36), and Mv1Lu mink lung epithelial cells (35). However, baseline cell death in the 1HAEo−.Smad7+ cell line and death induced by either Fas activation or glucocorticoid treatment were not substantially different from that seen in cells transfected with the EV.
Stress-activated protein kinases such as JNK and p38 MAPK can mediate proapoptotic signals from the TβR. The p38 MAPK signaling pathway mediates TβR-signaled apoptosis in podocytes (61), vascular endothelial cells (29), and mouse mammary epithelial cells (78). In these studies, inhibition of p38 MAPK significantly reduced apoptosis mediated by the combination of Fas activation and TGF-β treatment, but not that elicited by Fas activation alone. p38 activation may be dispensable in situations where TβR-mediated signaling results in apoptosis (26). In our study, TGF-β1 treatment in the 1HAEo−.Smad7+ cell line concurrently with either Fas activation or glucocorticoid treatment elicited activation of p38, and blocking p38 activation with an inhibitor significantly attenuated apoptosis. In contrast, although JNK activation is associated with TβR-mediated apoptosis in several cell types (50, 79) and can mediate apoptosis elicited after Smad7 overexpression (35, 39), inhibition of JNK did not significantly change Fas- or glucocorticoid-mediated apoptosis in the 1HAEo−.Smad7+ cell line. Our data suggest that TGF-β1 signals two different pathways: an antiapoptotic signal mediated by the Smad pathway and a permissive, apoptosis enhancing signal mediated by p38 MAPK. These data represent the first demonstration of both a pro- and an antiapoptotic pathway mediated by TGF-β in the same cell system.
TGF-β is a trophic factor in the repair of airway epithelium after damage. TGF-β1 increases epithelial cell migration in rabbit tracheal explants (5) and shifts epithelial cell phenotype toward a squamous morphology (43, 59). However, the role of TGF-β in airway epithelial cell death has been less well defined. TGF-β treatment has been shown to be proapoptotic to bronchiolar (small airway) cells by enhancing the effect of Fas ligation (22). In that study, concurrent treatment of mice with parental injection of TGF-β and intratracheal instillation of anti-Fas antibody induced apoptosis only in alveolar epithelial cells and not in central airway epithelial cells. The study of Pelaia et al. (49) suggests that TGF-β treatment may elicit apoptosis in central airway epithelial cells, an effect that was mediated by p38 MAPK, JNK, and ERK 1/2, and could be attenuated by modest concentrations (10−8 M) of budesonide. In contrast, our study suggests that central airway epithelial cell lines of either basal cell or mucoid cell phenotype, as well as primary cells, are protected from either corticosteroid-induced or Fas activation-induced apoptosis by concurrent administration of TGF-β1. That the response to TGF-β was consistent in both primary cells and in several cell lines could be demonstrated over time and could be demonstrated to have a concentration-response relationship in our study strengthens our conclusion that TGF-β is protective and not proapoptotic when the Smad pathway can be activated. Pelaia et al. (49) attributed their response to an increase in p38 phosphorylation 2 h after TGF-β treatment. In contrast, TGF-β1 treatment in the Smad7-overexpressing airway epithelial cell line in our study elicited modest p38 phosphorylation within 15 min when given alone and sustained (5 min to 4 h) p38 phosphorylation when given in combination with Fas activation. Furthermore, no discernable effect of blocking either the JNK or the ERK 1/2 pathways was noted. These differences may reflect differences in cell lineage and conditions but suggest that the activation of signaling pathways, and resulting responses, from TGF-β receptor binding may be both complex and context specific. In situations such as airway inflammation, mediators present in the airway or airway lumen may upregulate p38 or downregulate the Smad pathway and thereby alter the effect of TGF-β from a protective to an apoptotic-permissive effect. The data of Pelaia et al. (49), combined with data from the present study, raise the interesting possibility that changes in downstream signaling from TGF-β in selected disease states may alter responsiveness to this factor and thus promote either cell death or cell survival.
In our experiments, TGF-β1 treatment protection from two different apoptotic stimuli could be reversed by overexpression of either cyclin D1 or cyclin E. Cell systems exist that undergo TGF-β-mediated G1 arrest independently of cyclin D-CDK4/6 and rather targeting cyclin E-CDK2 instead (42). TGF-β can inhibit either set of cyclin-CDKs, depending on the inciting proliferating stimulus as in the Mv1Lu alveolar epithelial cell line (9, 66). In our experiments, overexpression of cyclin D1 or cyclin E overcame TGF-β-mediated suppression of apoptosis. Our data suggest TGF-β-mediated protection from apoptosis may be associated with changes in cell cycle that may be overcome by expression of either G1-associated cyclin.
Two major families of CKIs oppose CDK activation (38, 52). The p21WAF1/Cip1, p27Kip1, and p57Kip2 inhibitors belong to the kinase inhibitory protein (KIP) family and inhibit CDK2/cyclin E complexes in G1 phase. The p15INK4B, p16INK4A, and p18INK4C inhibitors belong to the inhibitors of CDK4 family and inhibit both CDK4 and CDK6 (reviewed in Ref. 63). TGF-β treatment can induce expression and stabilization of one or more CKI, and conversely inactivation of CKIs may contribute to TGF-β resistance in normal and cancer cells (12). These inhibitory proteins may also mediate aspects of apoptosis. In human ectocervical cells and colon carcinoma cell lines, TGF-β1 treatment is associated with increased p21 expression and apoptosis (32, 58). However, in primary cultured bronchiolar epithelial cells and retinal endothelial cells, downregulated expression of p21 was associated with apoptosis after TGF-β1 treatment (22, 32). Alternatively, studies in neuroblastoma cells and colorectal carcinoma cells have demonstrated an antiapoptotic role of p21 (18, 51). p21 expression is increased in the bronchial epithelium of asthmatic subjects (53) and may contribute to airway inflammation and remodeling. In the present study, we demonstrated an increase in p21 expression following TGF-β1 treatment that is associated with protection from apoptotic stimuli. This increased expression was seen after either Fas activation or glucocorticoid treatment. Increased p27 expression following TGF-β treatment also may elicit growth arrest, and resistance to TGF-β may be accompanied by changes in altered activity, localization, or expression of p27 (13). In our study, TGF-β1 treatment was not associated with changes in expression of p27 over the time period in which apoptosis occurred. In this cell system, then, the antiapoptotic effect may be mediated by p21 and not p27.
TGF-β is found in increased abundance in asthmatic airways (68, 73), and the signaling pathways that may mediate TGF-β effects also may be altered. Overexpression of Smad7 in a model of mouse airway allergen challenge led to increases in markers of inflammation, including increased T helper (Th) 1- and Th2-specific cytokine production (43). Another study from the same group demonstrated that there is an inverse correlation between bronchial epithelial cell Smad7 expression and basement membrane thickening in asthmatics. However, there was no notation of epithelial damage in that study. Changes in several signaling pathways, potentially each at multiple points, make relating the increased abundance of TGF-β to changes in airway epithelial cell survival difficult. It is interesting to speculate that in asthmatic airways, epithelial damage from inflammatory and environmental factors and signals could be either attenuated or worsened by TGF-β, depending on context.
In summary, our data demonstrate that TGF-β1 elicits a protective effect against Fas activation or dexamethasone-induced apoptosis in central airway epithelial cells. This effect occurs in a concentration-dependent manner, is mediated by the Smad pathway, and can be blocked by overexpression of Smad7. This effect is associated with increased p21 signaling and can be blocked by overexpression of either cyclin E or cyclin D1. When Smad signaling is blocked by overexpression of Smad7, a second pathway mediated by p38 MAPK elicits a Smad-independent apoptotic signal. Thus TGF-β-mediated effects on apoptosis may proceed by at least two signaling pathways in the same cell.
This work was supported by National Institutes of Health Grants HL-60531, HL-63300, and AI-56352, by Canadian Institute of Health Research Grant 43898, and by institutional National Research Service Award HL-07605.
We thank Blanca Camoretti-Mercado (University of Chicago) for the Smad7 expression vector. We thank Richard Pestell Albert (Einstein College of Medicine, The Bronx, NY) for the cyclin E expression vector. We thank Charles Sherr (St. Jude Children's Research Hospital, Memphis, TN) for the cyclin D1 expression vector. We thank Pieter Hiemstra (Department of Pulmonology, Leiden University Medical Centre, Leiden, The Netherlands) for assistance with H-292 cell experiments. We thank Anne Sperling (University of Chicago) for helpful comments.
This work was presented in part at the 2002 International Conference of the American Thoracic Society, Atlanta, GA, May 19, 2002, and at the 2003 International Conference of the American Thoracic Society, Seattle, WA, May 18, 2003.
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.
- Copyright © 2004 the American Physiological Society