In the fetal lung, endogenous transforming growth factor (TGF)-β inhibits early morphogenesis and blocks hormone-induced type II cell differentiation. We hypothesized that endogenous TGF-β inhibits type II cell differentiation and that the stimulatory effects of glucocorticoids result in part from suppression of TGF-β. Epithelial cells were isolated from human fetal lung and cultured under defined conditions with and without dexamethasone plus cAMP to promote type II cell differentiation. Control cells produced TGF-β, which was activated in part by αVβ6-integrin. Treatment with dexamethasone, but not cAMP, reduced TGF-β1 and -β2 transcripts and TGF-β bioactivity in culture medium. To examine the effects of decreased TGF-β in the absence of glucocorticoid, cells were treated with antibodies to TGF-β and its receptors. By real-time RT-PCR, antibody blockade of TGF-β reduced serpine1, a TGF-β-inducible gene, and increased gene expression for sftpa, sftpb, sftpc, and titf1, mimicking the response to hormone treatment. By microarray analysis, 29 additional genes were induced by both TGF-β antibody and hormone treatment, and 20 other genes were repressed by both treatments. For some genes, the fold response was comparable for antibody and hormone treatment. We conclude that endogenous TGF-β suppresses expression of surfactant proteins and selected other type II cell genes in fetal lung, in part secondary to increased expression of titf1, and we propose that the mechanism of glucocorticoid-induced type II cell differentiation includes antagonism of TGF-β gene suppression. Surfactant production during fetal development is likely influenced by relative levels of TGF-β and glucocorticoids.
- surfactant protein
- thyroid transcription factor-1
- gene expression
transforming growth factor (TGF)-β, a multifunctional growth factor, modulates multiple cellular functions including cell growth and differentiation, apoptosis, and extracellular matrix remodeling/deposition (12, 15). Biological effects of TGF-β are mediated by TGF-β receptors (types I and II) that signal via Smad molecules. TGF-β exists in three isoforms (1, 2, and 3) with 60–80% sequence homology. All three isoforms are highly expressed throughout development, whereas only trace amounts are normally present in adult tissues (32).
TGF-β is secreted in vivo as an inactive latent complex that requires activation to release mature, biologically active TGF-β. Mature TGF-β (25 kDa) is noncovalently associated with latency-associated peptide (LAP) (27). Removal of LAP from mature TGF-β can occur by proteolytic cleavage by matrix metalloproteinases (MMPs), plasmin, or furin, resulting in TGF-β activation. In vitro, TGF-β can be chemically activated by acidification or heating (25). Alternatively, LAP can undergo a conformational change that allows mature TGF-β access to its biological receptors (18), an event mediated by specific integrins such as αVβ6 that recognize an arginine-glycine-aspartic acid (RGD) sequence in LAP (24). The αVβ6-integrin represents one of the few integrins that is upregulated in the lung in response to acute injury (17, 31). Mice that lack the β6-subunit develop exaggerated pulmonary inflammation but are protected from fibrosis, a phenotype consistent with diminished TGF-β activity (16).
In the mouse lung, TGF-β1 expression is detected in epithelial cells as early as day 11 of gestation (14). In murine lung bud cultures, TGF-β1 and -β2 treatment inhibits lung branching morphogenesis, while abrogation of TGF-β receptors or the signaling molecules Smads accelerates lung morphogenesis (6, 30, 40). Furthermore, transgenic mice that selectively overexpress TGF-β1 in lung epithelial cells have delayed lung morphogenesis and impaired cellular differentiation as characterized by decreased expression of pro-surfactant protein (SP)-C (38). Similarly, treatment of human fetal lung explants with TGF-β suppresses hormone-induced (glucocorticoid + cAMP) type II cell differentiation (3). These and other observations support the concept that TGF-β is a negative regulator of lung development.
We have developed an in vitro model for studying lung type II cell differentiation. In this model, primary cultures of human fetal lung epithelial cells are cultured with glucocorticoid plus cAMP to stimulate type II cell differentiation as characterized by increased phosphatidylcholine synthesis, induction of SPs, and secretion of surface-active surfactant. This process involves thyroid transcription factor-1 (TTF-1, Nkx2.1), a key regulator of lung maturation, which is induced in a synergistic manner by dexamethasone and cAMP (10, 11).
In this study we hypothesized that endogenous TGF-β inhibits type II cell differentiation and that the stimulatory effects of glucocorticoids result in part from suppression of TGF-β. We examined the effects of abrogating endogenous TGF-β signaling on specific markers of lung type II cell differentiation. Our findings indicate that isolated lung epithelial cells actively produce TGF-β during culture, that the β6-integrin contributes to TGF-β activation, and that glucocorticoid-induced type II cell differentiation involves downregulation of endogenous TGF-β production. Moreover, downregulation of TGF-β by neutralizing antibodies, in the absence of hormone treatment, induces a subset of hormone-regulated genes including the SPs.
MATERIALS AND METHODS
Reagents and antibodies.
Recombinant TGF-β1, dexamethasone, and human serum albumin were purchased from Sigma (St. Louis, MO). Pan-TGF-β and TGF-β type II receptor antibodies were obtained from R&D Systems (Minneapolis, MN). GM-6001, a broad-spectrum MMP inhibitor, was obtained from Calbiochem (San Diego, CA), plasminogen activator inhibitor (PAI)-1 antibody was from B&D Transduction Laboratories (Lexington, KY), and TTF-1 antibody was from Santa Cruz Biotechnology (Santa Cruz, CA). β6-Integrin and antibodies to mature SP-B8kDa and GAPDH were obtained from Chemicon (Temecula, CA). Pro-SP-B antibody was purchased from Strategic BioSolutions (Newark, DE). Other antibodies used were anti-TTF-1 (Lab Vision), anti-CEACAM6 (Novus Biologicals), anti-DC-LAMP (Immunotech), and secondary antibodies that were tagged with Alexa 488. All culture media were provided by the Cell Center Facility at the University of Pennsylvania (Philadelphia, PA), and electrophoretic-grade reagents were obtained from Bio-Rad Laboratories (Hercules, CA).
Enriched primary cultures of lung epithelial cells were prepared from human fetal lung tissue (16–19 wk gestation) as previously described (11) under protocols approved by the Institutional Review Board at Children's Hospital of Philadelphia. Briefly, fetal lung parenchyma was chopped into 1-mm3 explants that were cultured overnight in Waymouth medium. Tissue was digested with trypsin, collagenase, and DNase, and contaminating fibroblasts were removed by a differential adherence method for monolayer culture in serum-free Waymouth medium. The nonadherent (epithelial) cells were cultured overnight in Waymouth medium supplemented with 10% fetal calf serum (FCS) and then cultured for an additional 2–5 days in serum-free medium alone or medium supplemented with 10 nM dexamethasone + 0.1 mM 8-bromo-cAMP + 0.1 mM isobutylmethylxanthine (DCI) or in the presence of two specific TGF-β-blocking antibodies (anti-pan-TGF-β1, -β2, and -β3 and anti-TGF receptor II, both at 30 μg/ml). Isotype-specific antibody controls (mouse immunoglobin and goat immunoglobin) were included in experiments with TGF-β-blocking antibodies, and, additionally, all antibody experiments included cells treated with DCI as a positive control for type II cell differentiation. The cells remained viable (Trypan blue staining and adherence to culture dishes) and hormone responsive (induction of SPs) during culture for at least 7 days.
For detection of active and latent TGF-β both in culture medium and cells, the mink lung epithelial cell (MLEC) luciferase bioassay was used according to the protocol previously described by Abe et al. (1). These cells are stably transfected with the PAI-1 promoter, a known TGF-β-inducible gene, linked to a firefly luciferase as a reporter. Briefly, MLEC were suspended in DMEM supplemented with 10% FCS, and 1.8 × 104 cells were plated out in a 96-well plate and allowed to attach for 3 h at 37°C. Medium was removed, and the cells were incubated with 30 μl of sample without or with heating (80°C, 10 min) for measurement of total and active TGF-β, respectively. Also included were serially diluted recombinant human TGF-β standards (40–2,500 pg/ml). Samples or standards were cultured overnight in a total volume of 100 μl with MLEC, and the next day medium was removed and MLEC were washed twice with PBS Ca2+, Mg2+-free. MLEC were lysed in 100 μl of a reporter cell lysis buffer, and cell extracts were then analyzed for luciferase activity with the Luciferase Assay System from Promega Biosciences (San Luis Obispo, CA).
TGF-β activation studies.
MLEC were suspended at 2 × 105 cells/ml in DMEM supplemented with 10% FCS, and 50 μl was plated in a 96-well plate and allowed to attach for 1.5 h. Medium was removed and replaced with 50 μl of DMEM only or 50 μl of medium supplemented with anti-β6 neutralizing antibody (10 μg/ml), anti-pan-TGF-β blocking antibody (30 μg/ml), or normal mouse IgG (30 μg/ml, control). To examine cell surface TGF-β activation, fetal lung epithelial cells in DMEM containing 10% FCS were added to the plated MLEC cells (at a 4-fold higher density than for MLEC cells) and cocultured overnight. Cells were collected, washed twice with PBS (Ca2+, Mg2+-free), and analyzed for luciferase activity according to manufacturer's instructions (Promega).
Total RNA was isolated with RNA-STAT-60 (Tel-Test, Friendswood, TX) per manufacturer's instructions. RNA integrity and concentrations were determined at the Nucleic Acid Core Facility at the Children's Hospital of Philadelphia with an Agilent 2100 bioanalyzer (Agilent Technologies, Palo Alto, CA). cDNA was synthesized from total RNA with a SuperScript First-Strand RT-PCR kit (Life Technologies, Carlsbad, CA) according to the manufacturer's protocol. RT-PCR reactions were performed with ABI Prism 7000 (Applied Biosytems, Foster City, CA). The Assay-on-Demand Gene Expression probe sets used for our studies were SP-A, HS00359837_ml; SP-B, Hs00167036_ml; SP-C, Hs 00161628_ml; and GAPDH, Hs99999905_gl. Included with each probe set was a set of cDNA standards that were derived from day 5 DCI-treated type II cells. The dynamic range for the cDNA standards was 0.1–100 ng RNA.
Western blot analysis.
Whole cell lysates were prepared from at least three separate epithelial cell preparations, and immunoblotting was performed with NuPAGE bis-Tris gels in MES running buffer as previously described (10). All samples were run under reducing conditions, using 10–15 μg of total protein. Electrophoresed samples were transferred to Duralose membranes (Stratagene), and immunoblotting (overnight at 4°C) was performed for mature SP-B (1:3,000), pro-SP-B (1:4,000), TTF-1 (1:2,000), PAI-1 (1:2,000), and GAPDH (1:10,000). Blots were than incubated with either goat anti-mouse or goat anti-sheep IgG-horseradish peroxidase (at 1:10,000) for 1 h at room temperature. Immunoreactive proteins were detected on X-ray film with enhanced chemiluminescence according to the manufacturer's instructions (Pierce Biologicals). Semiquantitative densitometry was performed with MacBASE version 2.4 (FUJI Film, Elmsford, NY).
Cells cultured on a cover glass were fixed in 1% paraformaldehyde in PBS and permeabilized with 0.3% Triton X-100. We carried out immunostaining as described previously (11), using polyclonal antibodies. Images were obtained under identical exposure settings for each treatment condition.
cDNA microarray analysis.
Total RNA (10–15 μg) was used to prepare cDNA and subsequently biotin-labeled cRNA per the manufacturer's protocol (Affymetrix, Santa Clara, CA; www.affymetrix.com). Microarray hybridization and analysis were carried out as recently described (34). In brief, fragmented labeled cRNA was hybridized for 16 h at 45°C with the Affymetrix human gene array U133A chip, which contains ∼22,000 oligonucleotide probes. Labeled arrays were washed and stained with streptavidin-phycoerythrin (Molecular Probes, Carlsbad, CA) according to the Affymetrix protocol. Arrays were scanned and fluorescence was quantitated with Affymetrix Microarray Suite 5.0 software. Gene expression of cells treated with TGF-β blockers was compared with the baseline gene expression in control (IgG treated) and untreated cells from the same experiments. The Microarray Suite software uses Wilcoxon signed-rank tests to evaluate whether a transcript is detectable on the array (present, marginal, absent) and the probability of a significant change between arrays (i.e., TGF-β antibody treated vs. IgG control), assigning P values and a change call (increase, decrease, no change) for each probe. The microarray data for effects of TGF-β antibody treatment represent results from 4 separate cell preparations (1 lung each) and 12 chips. Induced and repressed genes were defined as those that were detected as present and increased or decreased, respectively, with TGF-β antibody treatment for 72 h compared with IgG control in all four experiments with P < 0.003 for at least three experiments and a mean fold induction or repression ≥1.5-fold. DCI vs. control data are available at http://stokes.chop.edu/web/neoresearch and at the Gene Expression Omnibus (GEO; accession no. GSE3306) at the National Center for Biotechnology Information (www.ncbi.nih.gov).
Production and hormonal regulation of TGF-β by lung cells.
Primary cultures of fetal lung epithelial cells maintained in serum- and hormone-free medium produced and secreted TGF-β (total ≃ 3,000 pg/mg cell protein; active ≃ 250 pg/mg cell protein) over the first 4 days of culture (Fig. 1A). Production of TGF-β by cultured fetal lung fibroblasts was comparable to that by epithelial cells (data not shown). In the presence of DCI, content of both active and latent TGF-β in the medium of epithelial cells was significantly decreased compared with control cell medium (no hormones) (Fig. 1B). Decreased production of TGF-β with DCI treatment was evident at 24 h (day 2), and at 48–72 h (days 3–4) total TGF-β was decreased 84% and active TGF-β by 61%. To ascertain whether the effect observed with DCI was due to one or both hormones, cells were cultured with either cAMP or dexamethasone alone (Fig. 1C). Total TGF-β content in the medium was decreased 24 ± 9.6% by cAMP [not significant (NS)], 72 ± 9% by dexamethasone, and 76 ± 7% by DCI, indicating that the effect observed with DCI was mostly due to dexamethasone.
We used DNA microarray analysis to further characterize glucocorticoid effects on TGF-β signaling. Control cells expressed all three TGF-β isoforms, with TGF-β1 mRNA content ∼20-fold higher than for TGF-β2 and TGF-β3. After 72-h exposure of cells to DCI, contents of TGF-β1, -β2, and -β3 mRNAs were decreased to 55 ± 2% (P < 0.003), 44 ± 4% (P < 0.003), and 81 ± 6% (NS), respectively, of control (n = 5). In addition, transcript for inhibin βA (INHBA), another ligand for TGF-β receptor, was reduced by DCI to 29 ± 7% of control (P < 0.003). Hormone treatment did not decrease mRNA content for other components of TGF-β signaling: TGF-β receptors I and II and Smads 1–7 and 9 (data not shown).
Local activation of TGF-β.
To assess mechanisms of TGF-β activation, we cocultured human lung epithelial cells with MLEC as described by Munger et al. (24). In the presence of a neutralizing antibody to the β6-integrin subunit, activation of TGF-β was significantly inhibited by 50% (Fig. 2) compared with cells that were cocultured with an isotype-specific IgG antibody control. Since activation of TGF-β was only partially inhibited by anti-β6 blocking antibody, other mechanisms likely contribute to activation of TGF-β by fetal lung epithelial cells. We tested a broad-spectrum MMP inhibitor, GM-6001, which neutralizes MMP-1, MMP-2, MMP-3, MMP-8, and MMP-9. There was no significant effect of GM-6001 on TGF-β activation, suggesting that these specific MMPs were not involved in the activation of TGF-β by fetal epithelial cells; similarly, there was no effect of a blocking antibody to CD44 (data not shown).
Effect of blocking antibodies on endogenous TGF-β.
We utilized two TGF-β blocking antibodies (anti-pan-TGF-β and anti-TGF-β receptor II) to examine the role of endogenous TGF-β on epithelial cell gene expression. In initial experiments assessing TTF-1, SP-B, and/or PAI-1 (as described below), we found optimal responses to combined antibody treatment after 3-day exposure to each antibody at 30 μg/ml, a concentration that completely blocked induction of PAI-1 by exogenous TGF-β (10 ng/ml); tested individually, anti-pan-TGF-β antibody had a greater effect than anti-TGF-β receptor II antibody, with the greatest response to combined treatment (data not shown). To confirm that anti-TGF-β antibody reduced endogenous TGF-β, we measured active TGF-β content in medium of cells cultured with and without antibodies. Medium collected from cells exposed for 3 days to anti-TGF-β antibodies had significantly reduced active TGF-β content (35% of IgG-treated controls) (Fig. 3A).
To assess TGF-β signaling in the presence of blocking antibodies, we examined PAI-1, a known target protein of TGF-β. When cells were cultured with TGF-β blockers there was a 60% reduction in PAI-1 cellular content compared with control and IgG-treated cells (Fig. 3B). In these studies some cells were also cultured with DCI, which caused a comparable 70% reduction in PAI-1 (Fig. 3B).
Effect of antibodies on SPs and TTF-1.
On the basis of the known inhibitory effects of TGF-β treatment on type II cell differentiation (3), we predicted that SP gene expression would increase with reduction of endogenous TGF-β. Contents of SP-A, SP-B, and SP-C mRNAs were determined by real-time RT-PCR in cells cultured with TGF-β antibodies or normal IgG (control). SP-A and SP-C transcripts were very low to undetectable in control cells, and SP-B mRNA was present at a low level. Contents of all three SP mRNAs were greater in cells exposed to TGF-β antibodies (Fig. 4A). The magnitude of the increases for SP-A and SP-C was variable between experiments, in part attributable to very low expression in control cells; hence the fold induction is considered an estimate for these SPs. There was no significant effect of antibody treatment on content of mRNA for SP-D, which is modestly induced by cAMP + dexamethasone compared with SP-A, SP-B, and SP-C.
We next determined the content of mature (8 kDa) SP-B, which, unlike SP-A and SP-C, is detectable at a low level in control cells and is highly induced by DCI (Fig. 4B). Unexpectedly, there was no effect of TGF-β antibody treatment on mature SP-B content, although pro-SP-B (42 kDa) was slightly increased (2.1-fold; Fig. 4, C and D) by antibodies compared with 7.9-fold induction by DCI. Western blot analysis was not performed for SP-A, which is secreted and does not accumulate in cells (11), or for SP-C because of the unreliability of antibody detection for the mature protein.
Expression of SP-A, -B and -C is influenced by TTF-1, which we previously found (11) is induced in response to DCI treatment of epithelial cells. To examine the role of TTF-1 in the SP response to TGF-β antibodies, we performed immunoblotting experiments. TTF-1 content increased fivefold in cells cultured with TGF-β blocking antibodies, comparable to the response observed with DCI treatment (Fig. 5), and there was increased intensity of nuclear TTF-1 immunostaining in most but not all antibody-treated cells (Fig. 6A) compared with control (Fig. 6E). In the same experiments, staining for SP-B was slightly increased, consistent with the modest induction of pro-SP-B by Western blot analysis (Fig. 4, C and D). These findings indicate that induction of SP expression by anti-TGF-β antibodies is likely mediated, at least in part, by increased TTF-1.
Profiling of gene expression with and without TGF-β blockade.
Microarray studies were performed to identify additional DCI-regulated genes that were induced or repressed when endogenous TGF-β content was inhibited by antibody treatment. As for the previous studies, fetal lung epithelial cells were cultured for 72 h, in the absence of any hormones, with IgG (control cells) or with TGF-β-blocking antibodies. Changes in gene expression for four experiments were assessed with Affymetrix U133A chips and expressed as fold changes in treated (anti-TGF-β antibody) vs. control (IgG).
Exposure of cells to TGF-β antibodies reproducibly increased expression of 135 genes by >1.5-fold compared with IgG control. Median induction was 1.9-fold with a range of 1.6- to 16.1-fold. Thirty of the TGF-β antibody-induced genes were also responsive to DCI, with a similar level of induction for 15 of the genes (Table 1). The most highly induced genes, SP-A and SP-C, had very low expression in control cells as determined by both microarray analysis (not shown) and RT-PCR (Fig. 4A). In addition to surfactant-related genes, the list includes genes with a restricted number of biological functions: transport (7), signal transduction (7), transcription (2), and metabolic (2).
The effect of TGF-β blocking antibodies on two of the induced genes was examined by immunostaining. There was a slight increase in staining for LAMP3 with antibody treatment (Fig. 6C) compared with control (Fig. 6G) and a marked increase for CEACAM6 in most cells (Fig. 6, D vs. H). These results are consistent with the level of induction observed by microarray analysis: LAMP3 2.8-fold and CEACAM6 11-fold (Table 1).
TGF-β antibody treatment of cells repressed expression of 85 genes >1.5-fold. Median repression was 1.9-fold with a range of 1.6- to 4.9-fold. Table 2 presents 19 genes that were repressed by both TGF-β antibody and DCI; there was a similar level of repression for most of the genes. Repressed genes included functional categories of immune response (8), extracellular matrix (8), and signal transduction/regulatory (3). These results are consistent with the possibility that induction and repression of subsets of DCI-responsive genes are mediated solely by reversal of TGF-β repression and induction, respectively.
There is considerable evidence for a modulating role of endogenous TGF-β for branching morphogenesis in mouse lung (38–40). Later in fetal life, type II cell differentiation in human fetal lung explants is blocked by TGF-β treatment (3); however, the role of endogenous TGF-β in this differentiation process has not been explored. In the present work we found that blockade of endogenous TGF-β signaling induces a subset of genes that are regulated during hormone-induced type II cell differentiation. These include sftpa, sftpb, sftpc, and titf1, which are key markers for the type II cell. On the basis of these findings, we propose that glucocorticoid effects on type II cell maturation reflect in part an indirect action, namely, relieving TGF-β suppression of specific genes. To our knowledge this is the first study that demonstrates a direct association between endogenous TGF-β content and glucocorticoid-induced alveolar epithelial cell phenotype.
Activation of latent TGF-β is paramount to its ability to mediate its biological effects. There are limited data available regarding the potential mechanism(s) involved in the activation of TGF-β in human type II cells. We found that the majority of TGF-β secreted from lung epithelial cells was in the latent form, while a minor fraction was biologically active. We also found that the β6-integrin subunit is involved in the activation of latent-TGF-β by human lung epithelial cells. This finding is consistent with data in the mouse indicating that αVb6-integrin is exclusively expressed by lung epithelial cells and directly binds LAP-TGF-β via its RGD sequence (7). The αVβ6-integrin also represents one of the few integrins that can be upregulated in the airway epithelium in response to acute injury (35). The inability of the β6 antibody to fully inhibit TGF-β activation suggests that other factors are also involved in the activation of latent TGF-β by these cells, but we found no evidence by inhibitor and antibody studies for involvement of either MMPs or CD44.
In the present work we show that treatment of human lung epithelial cells with dexamethasone downregulates endogenous TGF-β production, acting at least in part at the mRNA level for β1 and β2 isoforms. Previously, Wen et al. (36) showed that dexamethasone decreased TGF-β1 and -β2 mRNA content in human fetal lung fibroblasts. As an experimental approach to blocking endogenous TGF-β signaling in the absence of glucocorticoid, we treated cells with neutralizing antibodies to both TGF-β1/-β2/-β3 and TGF-β type II receptor extracellular domain. As confirmation of antibody efficacy, expression of TGF-β-responsive PAI-1 was consistently reduced by ≥50% with antibodies. Treatment of cells with antibodies mimicked the response to DCI, albeit at reduced levels, with regard to induction of SP-A, SP-B, and SP-C. The difference in levels of induction between TGF-β antibodies and DCI likely reflects 1) failure to completely suppress TGF-β signaling, 2) the inductive effects of cAMP present in DCI, and 3) direct effects of glucocorticoids on expression of SPs or transcriptional regulators of those genes. The failure of TGF-β antibody treatment to increase mature SP-B 8-kDa protein, while modestly increasing pro-SP-B, is likely due to lack of induction of specific processing enzymes (13, 33). For example, DCI treatment highly upregulates expression of pepsinogen C (9, 34), which has a proposed role in SP-B processing, but pepsinogen C mRNA content was not affected by TGF-β antibody treatment.
Profiling of gene expression in the presence of TGF-β antibodies demonstrated regulation of subsets of both DCI-induced and DCI-repressed genes. SP-A, SP-B, and SP-C mRNAs were induced by TGF-β antibody, which was confirmed with RT-PCR, but, of interest, there were no induced genes related to phospholipid synthesis. By contrast, DCI treatment induced 13 genes involved in lipid uptake, synthesis, or remodeling (34). The major functional categories of genes induced by both TGF-β antibody and DCI were transport and signal transduction (46% of induced genes), which is comparable to the occurrence of these categories (44%) in our study of DCI-induced genes (34). Induction of CEACAM6 (carcinoembryonic antigen cell adhesion molecule) was similar with both TGF-β antibody and DCI treatment (∼10-fold; Table 1). This protein binds microorganisms, regulates cell adhesion, and is highly expressed in a variety of human malignancies, but its role in developing lung is not known (5, 8, 22, 26, 29). It is tempting to speculate that epithelial cell genes under both negative (TGF-β) and positive (DCI) regulation encode key proteins for the type II cell phenotype.
A smaller subset of genes was downregulated by both TGF-β antibodies and DCI. A number of these genes are related to the immune response and extracellular matrix and presumably represent targets of glucocorticoid anti-inflammatory action. Many of these proteins are known to be TGF-β responsive (e.g., PAI-1, collagens, lysyl oxidase, tenascin, cystatin), but others represent new candidate targets (2, 19, 23, 28). As cellular localization of expression has not been determined, it is possible that some of these genes are primarily expressed and regulated in fibroblasts that contaminate the epithelial cell preparation.
To explore potential mechanisms by which TGF-β blockade induces SP gene expression, we examined TTF-1, a known activator of SP gene transcription in vitro (4). Moreover, TTF-1 is induced by DCI in fetal lung cells (11), and TGF-β represses SP-B expression through receptor interaction and binding of Smads to TTF-1 at its regulatory element (21). As predicted, TGF-β antibody treatment increased TTF-1 content and nuclear staining. The observed variability in staining between cells is consistent with the previous findings for SP staining and indicates that only a portion of cultured epithelial cells are responsive, under the conditions used, to both hormone treatment and decreased TGF-β. We speculate that increased TTF-1 is responsible, at least in part, for induction of some (e.g., SPs, CEACAM6, LAMP3) but not all of the inductive responses observed with TGF-β antibodies. We are currently examining this topic in studies with knockdown and overexpression of TTF-1; preliminary results indicate that ∼50% of genes regulated by TGF-β antibodies are putative TTF-1 target genes (20). Thus TGF-β apparently suppresses differentiation by both TTF-1-dependent and -independent mechanisms.
In studies with cultured rat type II cells, Willis et al. (37) reported that prolonged exposure to TGF-β suppressed TTF-1 and increased expression of mesenchymal phenotype markers, consistent with epithelial-mesenchymal transformation. By contrast, we found no change with either DCI or anti-TGF-β antibody treatment on gene expression for mesenchymal cell markers other than type I collagen (2-fold decrease), which may reflect TGF-β effects in fibroblasts that contaminate the cultures (5–10%). Responses to TGF-β by lung epithelial cells may depend on developmental status, with suppression of differentiation in fetal precursor cells and profibrotic responses in adult type II cells.
There are some limitations to this study. There was considerable variability between experiments in the response of SP-A and SP-C to TGF-β antibody. We attribute this to unknown differences between lung specimens that affect the induction kinetics for these genes, which are relatively slow responders to DCI. On the basis of observation that cAMP treatment did not significantly reduce TGF-β content of cells, we expect that all genes responding to TGF-β antibody are regulated by dexamethasone, either alone or in combination with cAMP. However, we have not yet confirmed responses found by microarray analysis by other approaches nor tested effects of individual hormones. Accordingly, these findings represent candidate regulated genes. Although our model cell system has been extensively characterized and appears to closely mimic type II cell differentiation in vivo, it remains to be determined whether similar glucocorticoid-TGF-β interactions occur during lung development in vivo.
On the basis of our observations, we propose that endogenous TGF-β suppresses expression of SPs and selected other genes of epithelial cells during fetal lung development. Because TGF-β is produced by both epithelial and mesenchymal cells of fetal lung, effects on epithelial cells in vivo would likely reflect a combination of autocrine and juxtacrine regulation. Reduced TGF-β levels in response to increased glucocorticoid, of either endogenous or exogenous source, promote expression of the TGF-β-repressed genes. We suggest that this process is mediated in part by increased level and activity of TTF-1, which enhances expression of target genes including SPs and presumably others yet to be identified. The responses to TTF-1 are augmented by direct actions of glucocorticoids and cAMP on these and other target genes involved in transition of precursor epithelial cells to type II cells. This model predicts that the timing of initial surfactant production during fetal development reflects in part the relative levels of both TGF-β and glucocorticoids. Accordingly, strategies to reduce TGF-β production or signaling in combination with antenatal glucocorticoid treatment might improve efficacy of antenatal therapy to prevent respiratory distress syndrome of premature infants.
This study was supported by National Heart, Lung, and Blood Institute (NHLBI) Grants HL-19737 and HL-56401 and by the Gisela and Dennis Alter Endowed Chair in Pediatrics (P. L. Ballard) as well as NHLBI Grants HL-62472 and HL-073896 (R. C. Savani).
We thank P. Wang, V. Kolla, Y. Ning, and S. Angampalli for technical assistance, S. Guttentag for advice, and C. Dennis for assistance with the manuscript.
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 © 2007 the American Physiological Society