Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Cellular inflammation orchestrated by cytokines secreted from T helper (Th) 2 cells underlies the pathogenesis of asthma (1). It is now established that in animals repeated airway antigen challenge results in a chronic disease associated with widespread narrowing of the bronchial airways, massive eosinophil and mast cell infiltration (2), epithelial shedding (3), high interleukin (IL)-4 and immunoglobulin (Ig) E production (4), and remodeling of the airways (1). Within this context, airway remodeling refers to structural changes that occur over time, perhaps due to such chronic airway inflammation. Airway remodeling comprises alterations in the airway epithelium, lamina propria and submucosa, leading to thickening of the airway wall. The consequences include partially reversible airway narrowing, bronchial hyperresponsiveness, airway edema, and mucus hypersecretion (1).

Members of the transforming growth factor (TGF)- family, which include TGF-s, activins, and bone morphogenic proteins (BMPs), are pluripotent cytokines exerting an array of biologic effects in a large variety of cell types. They orchestrate events vital to the initiation, progression, and resolution of inflammatory responses. TGF- has been shown to create chemotactic gradients and upregulate expression of adhesion molecules, resulting in recruitment of monocytes and neutrophils at the initiation of the inflammatory response. In addition, these cytokines are also involved in the termination of the responses by inhibiting proliferation of many cell types, including lymphocytes (5). Finally, members of the TGF- family regulate deposition of extracellular matrix (ECM) components like collagen, fibronectin, and proteoglycans that contribute to fibrosis and scar formation (6).

Dysregulated expression or function of the TGF- signaling system may result in pathophysiologic conditions. Indeed, several studies have implicated the TGF- family in the pathophysiology of autoimmune, neurodegenerative, and fibrotic diseases as well as in carcinogenesis and chronic inflammation (7, 8). Previous studies have provided indirect evidence indicating important roles for these cytokines during airway inflammation and remodeling. For instance, neutralization of TGF-1 and TGF-2 in the bleomycin-induced fibrosis model counteracted development of fibrosis (9). Moreover, activin A was shown to promote differentiation of fibroblasts to myofibroblasts in vitro and to be upregulated during this fibrotic response in the airways in vivo (10).

TGF- family members are produced as larger precursor dimeric proteins that are processed proteolytically into mature proteins (11, 12). These biologically active dimers bind to specific, constitutively active type II serine/threonine kinase receptors, which can subsequently recruit and phosphorylate specific type I serine/threonine kinase receptors (13). Downstream signals are propagated through the type I receptors, which phosphorylate receptor-regulated Smad proteins (R-Smads). Whereas Smad2 and Smad3 are phosphorylated by activin and TGF- type I receptors, Smad1, Smad5, and Smad8 act downstream of BMP type I receptors. Type I receptor-induced phosphorylation occurs at the two most carboxy-terminal serine residues in R-Smads; for Smad2, the phosphorylation occurs in its SSMS C-terminal motif. Activated R-Smads form heterocomplexes with the common mediator Smad4 and these heteromeric complexes efficiently translocate to the nucleus where they participate in the regulation of the transcription of target genes (14).

The TGF- family has been associated with disease pathophysiology to a great extent by monitoring changes in the levels of TGF- cytokines or receptors during the progression of the disease. However, these studies were unable to determine whether ligand-induced receptor activation occurred in vivo. The recent development of reagents that specifically recognize Smad proteins, and in particular antibodies against R-Smads in their phosphorylated and active configuration, has provided an alternative method to demonstrate involvement of the TGF- superfamily/Smad signaling during pathophysiologic conditions in vivo. We have used antibodies specific for TGF- and activin type I and type II receptors, and their downstream Smad proteins as well as reverse transcriptase/polymerase chain reaction (RT-PCR) with specific primers for TGF-/activin receptors, Smads, and ligands to monitor their expression during the development of an inflammatory response. Our data demonstrate for the first time, in situ overactive TGF- family-mediated signaling during antigen-induced airway inflammation. This suggests that TGF-s and activins, their receptors, and downstream Smad effector proteins could be important contributors to the pathophysiology of allergic asthma.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals

Female Balb/C (H-2^d, Mls-1^b) mice were obtained from Bomholtgård (Ry, Denmark) and were routinely used at the age of 8 to 12 wk. The experiments were conducted with the approval of the Swedish ethical committee.

Allergen-Induced Airway Inflammation

Mice were sensitized by an intraperitoneal injection of 7.5 µg ovalbumin (OVA) (Sigma, St. Louis, MO) adsorbed to 1.5 mg aluminum hydroxide F-2200 (Reheis, Dublin, Ireland) dissolved in 0.3 ml phosphate-buffered saline (PBS) (Life Technologies Ltd., Paisley, UK). Animals were boosted on Day 7 using the same administration. On Days 14, 21, and 22, mice were challenged with 10 mg/ ml aerosolized OVA for 1 h in a Plexiglas chamber (20 × 15 × 15 cm). The aerosol was generated with the Bird 500-ml Inline Micronebulizer 890 (Rium, Sollentuna, Sweden) driven at 4 bar. Control animals were sensitized intraperitoneally with PBS or OVA and challenged with aerosolized PBS alone. Animals were terminated by an intraperitoneal injection of 0.15 ml (60 mg/ml) sodium pentobarbital (Apoteksbolaget, Umeå, Sweden).

Tissue Collection

Before resection, lungs were inflated with a solution consisting of 0.2 ml Tissue-Tek ornithine carbamyl transferase (OCT) (Sakura Fintek, Torrance, CA) and 0.4 ml PBS. A small incision was made in the proximal area of the trachea allowing insertion of a plastic cannula attached to a syringe through which the OCT/PBS solution was administered. The trachea was then tied off and the entire lung removed. Each lobe was dissected and individually placed in plastic boats covered with OCT and frozen in isopentane kept on dry ice. The specimens were stored at 70°C until used.

Immunohistochemistry

Cryopreserved tissues were cut in 7-µm-thick sections and fixed in 2% paraformaldehyde (Sigma, Steinheim, Germany) for 20 min. Endogenous peroxidase activity was quenched by a 20-min incubation in the dark with 1% H₂O₂ (Sigma) dissolved in methanol (Histolab, Gothenburg, Sweden). After blocking with avidin-biotin (Vector Laboratories, Burlingame, CA) and normal goat serum (5%) for 60 min, the sections were incubated with primary antibody dissolved in 0.5 M NaCl Earle's balanced salt solution (EBSS)/saponin buffer for 45 min (Life Technologies Ltd.; ICN Biomedicals Inc., Aurora, OH). Biotinylated goat antirabbit IgG (Vector Laboratories) secondary antibodies were added for 30 min in EBSS/saponin buffer supplemented with 2% normal goat serum. Vectastain Elite kit in 0.2 M methyl -D-mannopyranoside (Sigma) EBSS/saponin buffer was used as a reagent to avidin-biotin. Slides were developed for 5 min in diaminobenzidine (DAB) (Vector Laboratories) and counterstained with Mayer's hematoxylin (Histolab). Images were analyzed using the Leica IM1000 and the Leica Qwin computer program (Leica Microsystems, Heerbrugg, Switzerland). To verify our results, we pre-incubated our primary antibody with a 10-fold excess of the relevant or irrelevant peptide overnight at 4°C and then followed similar procedures as described previously.

Antibodies

Polyclonal antisera were raised in rabbits against synthetic peptides as previously described (15, 16). Briefly, for antibodies against activin type IB receptor (ActR-IB), also termed activin receptor-like kinase (ALK)-4, TGF- type I receptor (TR-I), also known as ALK-5, and activin type IIB receptor (ActR-IIB), peptides that corresponded to the divergent intracellular juxtamembrane domains were used. For antibodies against the TGF- type II receptor (TR-II) and activin type IIA receptor (ActR-IIA), peptides that corresponded to the divergent carboxy-termini were used. For antibodies against Smad2, Smad3, and Smad4, peptides that corresponded to the variable proline-rich linker region were used. The antibody against the phosphorylated Smad2 was raised by coupling KKK-SSpMSp (where Sp stands for phosphorylated serine residue) to keyhole limpet hemocyanin with glutaraldehyde, mixing it with Freuds adjuvant, and immunizing rabbits with it. All the antisera were tested for specificity by immunoprecipitation and Western blot analysis on COS cells transfected with different receptors and Smads. The antisera were purified using ImmunoPure IgG (protein A) Purification Kit (Pierce, Rockford, IL) according to the manufacturer's instructions. Alternatively, antibodies were affinity-purified using the corresponding immobilized peptides on cyanogen bromide (CNBr)-activated sepharose 4B beads (Amersham Pharmacia Biotech, Uppsala, Sweden). Briefly, sera were cycled three times over the column, eluted at pH 3, and adjusted with 1 M Tris-HCl, pH 7.5. Relevant fractions were dialyzed using membranes with a 10,000-mol wt cutoff in Slide-A-Lyzer Dialysis Cassettes (Pierce) and supplemented with 0.1% sodium azide. Polyclonal antibodies against phosphorylated Smad2 (Upstate Biotechnology, Lake Placid, NY) and normal rabbit IgG (Dako, Alvsjo, Sweden) were used as positive or negative controls in the experiments, respectively.

Western Blotting Analysis

Lung tissue from mice challenged three times either with PBS or OVA was homogenized in liquid nitrogen using a mortar and pestle. The proteins were then extracted in 1% TritonX-100, 25 mM Tris-HCl (pH 7.6), 0.1 M NaCl, 1 mM ethylenediaminetetraacetic acid, 5 mM NaF, 0.1 mM Na₃VO₄, 1 mM phenylmethylsulfonyl fluoride, and complete protease inhibitor cocktail (Boehringer Mannheim, GmbH, Germany) diluted 1:25. Equal amounts of lung protein were run under reduced condition on a 10% NuPage Bis/ Tris gel (Novex/Invitrogen, Groningen, the Netherlands) and transferred to polyvinylidene difluoride membranes (Novex/Invitrogen) by electroblotting using transfer buffer supplemented with 20% methanol (Novex/Invitrogen). Blots were blocked overnight at 4°C in PBS/0.1% Tween 20/1% bovine serum albumin and then incubated with 1 µg/ml of the primary antibody for 1 h at room temperature using the DecaProbe system (Hoefer, Amersham Pharmacia). Thereafter, the membranes were washed six times for 10 min with blocking buffer, incubated for 1 h with the secondary horseradish peroxidase-linked donkey-antirabbit antibody at 1:5,000 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), washed six times for 10 min with blocking buffer, incubated with enhanced chemiluminescence substrate (Amersham Pharmacia), and finally exposed to radiographic film.

RNA Isolation and RT-PCR

Total RNA was isolated from whole lung tissue with RNAqueous (Ambion, Austin, TX) according to the manufacturer's instruction. Complementary DNA (cDNA) was synthesized from 2 µg total RNA using the First-Strand cDNA synthesis kit (Amersham Pharmacia, Piscataway, NJ) according to the protocol provided by the manufacturer. Lung tissue from two animals challenged three times with PBS and two animals challenged three times with OVA was homogenized separately, and semiquantitative RT-PCR amplifications were performed using different concentrations of cDNA template (3:1, 1:1, 1:3, 1:9, 1:27, 1:81, 1:243, 1:729, and 1:2187). Amplification of the cDNA template was performed in a 25-µl PCR mix containing 0.5 µl cDNA template, 1 µl 10 mM deoxynucleotide triphosphate, 2.5 µl 10× PCR buffer, 0.5 µl MgCl₂, 250 U Taq polymerase (Boehringer Mannheim), 1 µl 10 pmol/µl of each primer (DNA Technology, Aarhus, Denmark), and 18.5 µl H₂O. The reactions were performed using the DNA Engine/Tetrad (MJ Research Inc., Watertown, MA) and the amplification conditions were 10 min at 95°C followed by 30 cycles each consisting of 60 s at 95°C, 90 s at 55°C, 60 s at 72°C, and finally followed by 10 min at 72°C. The PCR products were separated on 1% agarose gels and visualized by ethidium bromide staining, and the relative intensity of the obtained products was measured. The gels were scanned using the Fluor-S MultiImager (BioRad Hercules, CA) and the relative intensity of the obtained products was measured using the QuantityOne software (BioRad). Titration curves from the serial diluted templates were drawn and the bestfit equation was obtained with Excel software (Microsoft, Redmond, WA). The fold modulation, between PBS treated to 3× OVA treated sample for each marker, was calculated using the obtained formula. Statistical power was determined by the Student's t test. The PCR primers used were: ALK-1 forward, 5'-GAGCCGTGTTCATGGTAGT 3'; ALK-1 reverse, 5'-GGA GGAGCCAGAAGTTGAT-3'; ALK-2 forward, 5'-GGCGGG GTCTTACACGTAA-3'; ALK-2 reverse, 5'-CTG GACCAGAGGAACAAAGG-3';ALK-4forward, 5'CTGA GGACTGCTA CGGGAA-3'; ALK-4 reverse, 5'-TAAGCGT GCAGGAAGAT GT-3'; ALK-5 forward, 5'-ATGGGCAAT AGCTGGTTTT-3'; ALK-5 reverse, 5'-GCC ATAACCGCAC TGTC-3'; TR-II forward, 5'-AATTTCTGG GCGCCCTC GG TCT-3'; TR-II reverse, 5'-CCCGGGGCATC GCTCATCT-3'; ActR-IIA forward, 5'-ACACAGCCCACTTCAAATCC-3'; ActR-IIA reverse, 5'-CT GACAGTGAGC CCTTTTC-3'; ActR-IIB forward, 5'-GCTTAA AGGAGTCCG CACA-3'; ActR-IIB reverse, 5'-TCAATTGCTAC GGGCA-3'; Smad2 forward, 5'-GC CGTCTTCAGGTTTCACA-3'; Smad2 reverse, 5'-TAGTATGC GATTGAACACC-3'; Smad3 forward, 5'-CGCCAGTTCTAC CTCCAGTG-3'; Smad3 reverse, 5'-AAA GACCTCCCCTCC GATGT-3'; Smad4 forward, 5'-AGC CGTC CTTAC CCAC TGA-3'; Smad4 reverse, 5'-CTCAATCGCT TCT GTCCTG 3'; TGF-1 forward, 5'-CTCGGGGGCTGCG GC TACTG-3'; TGF-1 reverse, 5'-GGCGTATCAGTGGGGG TCA-3'; TGF-2 forward, 5'-CGAGCGGAGCGAGCAGGAG-3'; TGF-2 reverse, 5'-TAGGAGGGCAACAACATTA-3'; TGF-3 forward, 5'-GGT CCTGGCACT TTACAAC-3'; TGF-3 reverse, 5'-GGCGTACACAGCAGTTCTC-3'; activin A forward, 5'-GAGGACGACATTGGCAGGA-3'; Activin A reverse, 5'-GCA CTAGACTGGCACCACT-3'; Activin B forward, 5'-GCAGGC AACAGTTCTTCAT-3'; Activin B reverse, 5'-CTCCCTCTG GTCCTGACTG-3'; Activin C forward, 5'-CACAATGCCAC CCAGACC-3'; Activin C reverse, 5'-CAGCCAATCTCACG GAAGT-3'; Activin E forward, 5'-CAGCCGTCCCAGAA TAACT-3'; Activin E reverse, 5'-CAACATAAGGGGGTCT CAG-3'; glyceraldehyde-3-phosphate dehydrogenase (GAPDH) forward, 5'-GAAGGGTGGAGCCA-3'; GAPDH reverse, 5'-TGCCAGCCCCGGCA-3'.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

R-Smad2 Is Phosphorylated and Translocated to the Nucleus during Airway Inflammation

To validate the role of the TGF-/activin signaling pathway during the development of airway inflammation, mice were sensitized by two intraperitoneal injections of OVA, then challenged three times by inhalation of aerosolized OVA or PBS, after which the expression of Smads and TGF-/ activin receptors in the lung was analyzed by immunochemistry. Because all parameters monitored were modulated progressively in the course of the three challenges, the results presented and discussed were derived from animals killed 24 h after the third challenge.

Hallmarks of Smad activation are the carboxy-terminal phosphorylation and nuclear accumulation. Therefore, the effect of OVA challenge on the subcellular distribution of phosphorylated Smad2 was investigated. Smad2 is a downstream effector for both TGF- and activin, and thus provides as indicator of active TGF- and/or activin signaling in situ. Bronchial epithelial cells in PBS-treated animals had a weak cytoplasmic expression of phosphorylated Smad2 (Figures 1A and 1B, Table 1). Occasionally, a few of these cells (~ 4%) displayed nuclear translocation in healthy animals. Similarly, in all animals analyzed, scattered alveolar cells were positive (~ 4%) and some showed a nuclear localization of phosphorylated Smad2 (~ 0.3%), indicating a small but detectable level of ongoing TGF- and/or activin signaling in peripheral lung tissue (Figures 1A-1D).

View larger version (76K): [in this window] [in a new window]   Figure 1.   Markedly elevated expression of phosphorylated Smad2 in allergic lungs. Mice were challenged with three aerosolic inhalations of PBS (A-D) or OVA (F-H ). Lungs were dissected 24 h after the last inhalation and cryopreserved until use at 70°C. Lung sections (7 µm thick) were stained and visualized as brown color by DAB and counterstained with hematoxylin. Sections were stained with 0.3 µg of protein A-purified antibody against phosphorylated Smad2 (pSmad2) and shown at ×20 (A, C, E, and G) or ×100 (B, D, F, and H ) magnification. Expression in a representative intermediate airway (A, B, E, and F  ) and around a representative small vessel (C, D, G, and H ) is shown. Results from one of four similarly analyzed lungs are shown.

In sharp contrast, strong expression of phosphorylated Smad2 in almost all bronchial epithelial cells and intermediate staining in vascular endothelial and fibroblast-like cells were observed in mice challenged with OVA, especially after three inhalations (Figures 1E-1H). In addition, a substantial fraction of the alveolar (~ 75%) and infiltrating (~ 50%) cells did express phosphorylated Smad2 (Figures 1E-1H, Table 1). Most importantly, the majority of these cells showed nuclear translocation, indicating increased TGF- or activin signaling (Figures 1F and 1H). The nuclear immunoreactivity was efficiently inhibited by a tenfold excess of the phosphorylated Smad2 peptide used to produce the antisera but not by similar amounts of nonphosphorylated Smad2 peptide (data not shown).

Modulation of Smad Expression during Allergen-Induced Airway Inflammation

Changes in the level of Smad proteins induced by the OVA challenges were monitored by immunohistochemistry. Smad2 protein was detected in the bronchial epithelium and in the vascular endothelium in both healthy and OVA-challenged lungs (Figures 2A and 2E). A moderate induction was noted in the alveolar cells in the inflamed lung compared with the healthy lungs (Figures 2A and 2E). Very low immunostaining of Smad3 was detected in the bronchial epithelium in one of three lungs investigated, whereas the other cell compartments were negative in all three PBS-treated healthy control animals (Figure 2B). In contrast, expression of Smad3 was significantly upregulated in the bronchial epithelium after allergen challenge. Furthermore, the great majority of the infiltrating cells and the alveolar cells expressed Smad3 (Figure 2F). The common signaling mediator Smad4 was expressed in the bronchial epithelium and to a lower extent in the vascular endothelium, whereas the alveolar cells only expressed a minimal level of Smad4 in healthy animals (Figure 2C). The Smad4 levels were strongly upregulated in both bronchial and vascular endothelia in the allergen-challenged animals (Figure 2G). In addition, a moderate immunostaining was also noted on alveolar and infiltrating cells (Figure 2G).

View larger version (75K): [in this window] [in a new window]   Figure 2.   Induction of Smad3 and Smad4 expression in allergic lungs. Mice were challenged with three aerosolic inhalations of PBS (A-D) or OVA (E-H ). Sections were stained with 0.5 µg of protein A-purified antibody against Smad2 (A and E ), Smad3 (B and F ), and Smad4 (C and G), or 3 µg of normal rabbit serum antibodies (D and H ). Expression in a representative intermediate airway is shown at ×20 magnification. Results from one of three similarly analyzed lungs are shown.

Changes in the Pattern of Expression of TGF- and Activin Receptors during the Development of Airway Inflammation

To investigate which type I and II receptors were responsible for the observed Smad activation, and thus gain insights into the potential TGF- ligands triggering these events, we analyzed the expression of all the known TGF-/ activin responsive type I and II receptors in inflamed and healthy lungs by immunohistochemistry.

ALK-1, a candidate type I receptor for TGF- in endothelial cells (17), was found on all cell types in the lung with high expression on endothelial cells (Table 1). The expression pattern for ALK-2, initially described as a type I receptor for activin but recently found to be more important for BMP signaling (18), was found to be more restricted and only bronchial epithelial cells and a few alveolar cells were positive (Table 1). The expression pattern of ALK-1 was not altered notably during the progression of airway inflammation, whereas ALK-2 expression was upregulated moderately in fibroblasts and bronchial epithelial cells (Table 1). The expression pattern of ALK-4 and ALK-5 showed more pronounced changes after allergic challenge (Figures 3 and 4, Table 1), which suggests that these receptors might play a more determinative role during the development of airway inflammation.

View larger version (90K): [in this window] [in a new window]   Figure 3.   Strong induction of ALK-4 during airway inflammation. Mice were challenged with three aerosolic inhalations of PBS (A-C ) or OVA (D-F ). Sections were stained with 0.3 µg of protein A-purified antibody against ALK-4 and shown in ×20 (A, B, D, and E) or ×40 (C and F ) magnification. Expression in a representative vessel (A and B) and around a representative intermediate airway (B, C, E, and F ) is shown. "Attenuated fibroblasts" are indicated by red arrows. Results from one of four similarly analyzed lungs are shown.

View larger version (80K): [in this window] [in a new window]   Figure 4.   Redistribution of the TGF- type I receptor (ALK-5) during airway inflammation. Mice were challenged with three aerosolic inhalations of PBS (A-C ) or OVA (D-F ). Sections were stained with 2 µg of protein A-purified antibody against ALK-5 and shown in ×20 (A, B, D, and E ) or ×40 (C and F ) magnification. Expression in a representative vessel (A and D) and around a representative intermediate airway (B, C, E, and F ) is shown. Results from one of four similar experiments are shown.

Low levels of ALK-4 were observed in healthy lungs at the luminal edge of the bronchial epithelial cells (Figures 3B and 3C). A weak staining was also noted in vessels (Figure 3A), the subepithelial fibroblast-like cells around the bronchioles, and on some alveolar cells (Figures 3A-3C). After OVA challenge, ALK-4 was highly expressed on alveolar cells (Figures 3E-3F). A more pronounced upregulation was recorded when animals were challenged repeatedly than after a single challenge (data not shown). Similar increased expression was noted on both vascular endothelial and bronchial epithelial cells, especially in late disease (Figures 3D-3F). Interestingly, a large fraction of the infiltrating cells (> 80%) was ALK-4 positive (Figure 3D). Most importantly, all of the fibroblasts surrounding the bronchioles were strongly positive, indicating that behavior of these cells could be regulated by activins (Figures 3E- 3F). The ActR-IIA showed high expression on epithelial and endothelial cells in healthy lungs, whereas the ActR-IIB showed a weaker expression (Table 1). Infiltrating, alveolar, endothelial, and bronchial epithelial cells expressed ActR-IIB in levels that were not significantly changed on exposure to OVA. A modest induction of ActR-IIA expression was noted in the inflamed lungs (Table 1).

Earlier studies have demonstrated that in healthy lungs, bronchial epithelial as well as vascular endothelial cells are strongly positive for ALK-5, whereas alveolar cells express moderate levels of this receptor. Our data confirmed these findings (Figures 4A-4C) and, furthermore, demonstrated that during the progression of airway inflammation epithelial cells express further elevated levels of ALK-5 (Figures 4E and 4F). In addition, a fraction of the infiltrating cells was found to express moderate levels of ALK-5 (Figure 4D). Unexpectedly, both alveolar and especially vascular endothelial cells consistently showed a clearly reduced expression of ALK-5 after repeated OVA challenges compared with the healthy lungs (Figures 4A, 4C, 4D, and 4F). Ubiquitous expression of TR-II, with high expression on bronchial epithelial cells and on vessels, and moderate levels on alveolar and infiltrating cells, was seen in both healthy and OVA-challenged lungs. These findings indicate that regulation of TGF- responses at the receptor expression level occurs by induction of ALK-5 rather than TR-II (Table 1).

Western Blot Analysis of TGF- Signaling Components in Healthy and Inflamed Lungs

To further support the conclusions derived from the immunohistochemistry experiments, protein extracts from PBS- and OVA-challenged lungs were analyzed by Western blotting. In accordance with the immunohistochemistry findings (Figure 2), similar levels of total Smad2 protein were detected in extracts from PBS- and OVA-challenged animals (Figure 5A). The antiphosphorylated Smad2 antibody used in immunohistochemistry and a commercially available antiphosphorylated Smad2 antibody, which only works in Western blots, were used to monitor the presence of phosphorylated Smad2. Whereas only minute levels of phosphorylated Smad2 were detected in protein extracts from PBS-challenged lungs, a dramatic increase was noted in protein extracts from OVA-challenged animals (Figure 5A). Under the conditions employed, phosphorylated Smad2 migrated with a slower electrophoretic mobility than did Smad2 (Figure 5A). The phosphorylated Smad2 signal was competed efficiently by a tenfold excess of phosphorylated Smad2 peptide, but not with similar levels of unphosphorylated Smad2 peptide (Figure 5A). Conversely, the Smad2 signal was competed by the Smad2 peptide and not the phosphorylated Smad2 peptide, thus, further verifying the specificity of these reagents (Figure 5A). Taken together, these results demonstrate that the nuclear signal detected in bronchial epithelial cells and alveolar cells is due to the recognition of receptor-regulated, nuclear-translocated phosphorylated Smad2 protein.

View larger version (39K): [in this window] [in a new window]   Figure 5.   Expression of phosphorylated Smad2 (pSmad2) and TGF-/activin receptors in lung tissue. Mice were treated with three aerosolic inhalations of PBS or OVA and total lung homogenates were analyzed by Western blotting to monitor TGF- family elements. (A) Western blot analysis of lysates with the anti-pSmad2, anti-Smad2 antibodies used in immunohistochemistry, and an independent anti-pSmad2 antibody (Upstate Biotech). Tenfold excess of pSmad2 or Smad2 peptides were used, as indicated, to compete the immunoreactivity. One of three similar experiments is shown. (B) Western blot analysis of cell lysates with antibodies against ALK-4, ALK-5, ActR-IIA, ActR-IIB, and TR-II. One of two similar experiments is shown.

Similar levels of ALK-5 and TR-II protein were present in healthy and OVA-challenged animals, whereas both ActR-II receptors were upregulated during inflammatory conditions (Figure 5B). Most importantly, we were only able to detect ALK-4 in protein extracts from OVA-challenged lung, whereas healthy lungs expressed this receptor in levels below the detection limit of our assay (Figure 5B).

Significant Induction of Activin-Specific Messenger RNA and Modulation of TGF- Signaling Element Messenger RNA during Airway Inflammation

To further support the conclusions derived at protein level using immunohistochemistry and Western blot analysis, messenger RNAs (mRNAs) from PBS- and OVA-challenged lungs were analyzed by semiquantitative RT-PCR. The analyses demonstrated presence of similar levels of TR-II, ActR-IIA, and ActR-IIB mRNA in healthy and diseased lungs, moderate induction of ALK-1 (approximately twofold), moderate reduction of ALK-5 (approximately twofold), and a significant upregulation of ALK-2 (approximately fivefold) and ALK-4 (approximately threefold) transcripts in the inflamed lungs (Figures 6B and 6C). A moderate increase of Smad2 (approximately twofold) and Smad3 (approximately twofold) mRNA levels was noted in OVA-challenged lungs (Figures 6B and 6C). In contrast, significant induction of Smad4 transcripts (approximately fourfold) was evident during airway inflammation (Figures 6B and 6C). Both TGF- and activin ligands have been described to be upregulated during inflammatory responses in lung (19, 20). To investigate how these cytokines were modulated in response to OVA challenge in the experimental model used in our studies, the levels of mRNA encoding TGF- and activins in healthy and challenged lungs were monitored by semiquantitative RT-PCR. TGF-1, TGF-2, and TGF-3 were detected in healthy lungs (Figure 6A). During the inflammatory response, a moderate increase (~ 2.5-fold) of TGF-3 mRNA was noted, whereas TGF-1 and TGF-2 remained stable (Figures 6B and 6C). Low levels of activin B, C, and E mRNA were found in healthy lungs, whereas only minute levels of activin A could be detected (Figures 6B and 6C). Interestingly, a moderate induction of activin A (approximately twofold) and activin C (~ 2.5-fold), and significant induction of B (approximately fourfold) and especially E transcripts (~ 5.5-fold) were noted in the OVA-challenged lung (Figures 6B and 6C). Similar levels of follistatin (1.5-fold induction), an endogenous inhibitor of activins, were present in healthy and inflamed tissues (data not shown).

View larger version (14K): [in this window] [in a new window]   View larger version (32K): [in this window] [in a new window]   View larger version (22K): [in this window] [in a new window]   Figure 6.   Expression of TGF- family mRNA transcripts in lung after antigen-induced airway inflammation. mRNA from PBS- or OVA-challenged lungs was analyzed by RT-PCR. Serial dilutions of the cDNA were used as templates and the PCR products were processed as described in MATERIALS & METHODS. Titration curves of RT-PCR products at different template concentrations for a representative nonmodulated (GAPDH) and a representative upregulated (ALK-2) gene (A). Fold modulation in mRNA expression was derived from the statistical analysis of the corresponding titration curves (B). Statistical analysis was carried out with the Student's t test. *0.05 < P > 0.01; **0.01 < P > 0.001. RT-PCR products obtained using cDNA templates at 1:3 dilution (C).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, we demonstrate in situ, with antibodies recognizing phosphorylated/activated Smad2, that the TGF-/activin-Smad2 signaling pathway is active in resident pulmonary epithelial and endothelial cells, and within inflammatory cells migrating to the site of allergic challenge. A dramatic increase in the number of cells containing nuclear phosphorylated Smad2 and a significant modulation of the expression pattern for activin and TGF- receptors, noted in the lungs of animals challenged by OVA inhalation, raise the possibility that active TGF-/activin signaling might be associated with the development of allergic inflammation in this animal model system. Few scattered alveolar cells positive for phosphorylated Smad2 were found in normal lungs, whereas bronchial epithelial and endothelial cells were negative. The low level of phosphorylated Smad2 detected in healthy lungs could either reflect a low level of inflammation due to ubiquitous environmental antigens (e.g., dust) or a homeostatic response to suppress uncontrolled extensive branching similar to that noted by Zhao and coworkers (21) in embryonic cultures.

The relationship between TGF- signaling and airway inflammation has been extensively studied in the bleomycin- induced pulmonary fibrosis model. Alveolar macrophages, infiltrating eosinophils, mast cells, bronchial epithelial cells, as well as myofibroblasts, showed expression of TGF-1 during bleomycin-induced fibrotic changes (19). Experiments in which TGF-1 and TGF-2 were neutralized in vivo showed that deposition of ECM components were counteracted in the bleomycin model (9). Furthermore, animals deficient in v6 integrin, an activator of latent TGF-, were resistant to the development of bleomycin-induced fibrosis (22). Thus, there is good evidence demonstrating induction and suggesting involvement of TGF- during bleomycin-induced fibrosis.

However, the situation appears to be different in allergic airway inflammation. Adoptive transfer of OVA-specific Th2 cells engineered to produce latent TGF-1 demonstrated a protective role during OVA-induced airway inflammation. Furthermore, analysis of TGF-1 knockout animals revealed bronchial epithelial hyperplasia and metaplasia, thus indicating a potential role for TGF-1 in regulating homeostasis and thus guarding the integrity of the airway epithelium (23). The data presented herein indicate a clear participation of the TGF- family during the progression of allergic airway inflammation and, furthermore, implicate the activins as potential modulators of this process. It is interesting that analysis of biopsies from asthmatic patients has demonstrated increased production of both TGF-s and activins (24). Neither bleomycin- nor OVA-induced inflammation systems recapitulate the full spectrum of asthma pathophysiology, and it is thus probable that different members of the TGF- family regulate the inflammation induced by bleomycin or OVA in these experimental models.

The combination of increased phosphorylation of Smad2, the upregulation of type I activin receptors on epithelial, subepithelial, alveolar, and infiltrating cells, and the strong induction of transcripts for activins, described in the present report, raises the possibility that activins play a determinative role during antigen-induced airway inflammation. Indeed, production of activin A has been detected in smooth muscle cells in vessels and in bronchial epithelial in healthy lung (20). Furthermore, secretion of activin A was increased in alveolar cells in the bleomycin-induced lung fibrosis model, especially in areas around infiltrating cells (20). In vitro studies have shown that activin A stimulates the proliferation and differentiation of lung fibroblasts into myofibroblasts, thus advocating a critical role in lung tissue remodeling for this cytokine (10).

It is intriguing that among the cells exhibiting the most pronounced increase in ALK-4 expression was a layer of subepithelial fibroblast-like cells (indicated by arrows in Figures 4C and 4F) that were also -smooth muscle actin positive (data not shown). This anatomically distinct layer of fibroblasts has been previously recognized as the "attenuated fibroblast sheath" (25) and in conjunction with the overlaying epithelial layer, has also been referred to as the "epithelial-mesenchymal trophic unit" (26). Cells of the attenuated fibroblast sheath cover a large surface area and can respond to various stimuli due to their close proximity to the epithelial/environment interface (27). It has been speculated that during the lung fibrosis process, these attenuated fibroblasts have a distinct role compared with the fibroblasts present more deeply in the tissue controlling inflammatory responses and production of ECM, respectively (27, 28). Activins might thus play an important role in the crosstalk between the different cellular layers during the antigen-induced airway inflammation.

A surprising finding in the present study was the altered expression pattern of the TGF--responsive ALK-5 receptor. We confirmed earlier studies and demonstrated the presence of ALK-5 and TR-II in healthy lung (17, 29). Whereas the type II receptor showed similar ubiquitous expression patterns in the inflamed lung as in the healthy, the pattern of expression of ALK-5 was clearly altered in the allergic lung. Markedly reduced expression was noted on vessels, myofibroblasts, and alveolar cells, whereas increased expression was observed on bronchial epithelial cells. The implication of the ALK-5 downregulation is that although TGF- is present in the allergic lung, it most likely does not propagate signal through the ALK-5/ Smad2 pathway in vessels, fibroblasts, alveolar, and infiltrating cells. One possibility could be that TGF- signals through the ALK-1 receptor, which has been shown to be expressed in healthy, and in the present study also in inflamed, lung (17). Alternatively, the reduction of ALK-5 expressed on the surface of these cells could be the result of ligand-mediated receptor internalization and degradation that could occur during very early stages of OVA challenge.

Despite the fact that both Smad2 and Smad3 function as intracellular effector components of TGF-/activin receptor-mediated signaling (13, 14, 30, 31), certain TGF-- responsive promoters can be either Smad2-independent (32) or Smad3-independent (33). In some cases, Smad2 and Smad3 can antagonize functionally (34) or, alternatively, induce each other's expression (35). The phenotypes of Smad2 and Smad3 knockouts are different (33, 36), and in some cases in Smad2 and Smad3 genes can be regulated differentially at the transcriptional level. Thus, it could be anticipated that regulating the relative abundance of the different Smad proteins in a given cell could provide an additional mechanism by which signaling initiated by the same ligand could result in different responses. Furthermore, different cell types in the same tissue could respond in completely different ways to the same ligand if they would contain different relative levels of Smad2, Smad3, and Smad4.

We have observed clear differences in the expression pattern of Smad2, Smad3, and Smad4 before and after OVA challenge. The levels of Smad2 were similar in the epithelium of healthy and diseased airways and were moderately increased in the alveolar cells of OVA-challenged animals. Smad3 levels, which were almost undetectable by immunohistochemistry or by RT-PCR analysis in healthy lungs, were strongly upregulated in inflamed lungs. Finally, Smad4, which was readily detected only in the bronchial epithelium in healthy lungs, was strongly upregulated in the alveolar cells of OVA-challenged lungs both at the RNA and protein levels. Our findings strengthen the notion that Smads can be differentially regulated and thus raise the possibility that differential transcriptional activation of Smad genes could provide an additional level of regulation for biological processes modulated by TGF-s. This hypothesis echoes the ideas that Massagué and Wotton (37) have recently put forward and have so illustratively summarized in the sentence "don't ask what a signal can do with a cell, but what a cell can do with a signal."

Collectively, our results demonstrate that signaling by members of the TGF- family occurs locally in tissues during allergen-induced airway inflammation accompanied by modulation in expression of the type I receptors. Therefore, we suggest a model in which TGF- plays an important role in normal homeostasis in lung and other members, like activins, contribute to the pathophysiology of allergy. In this context, it is interesting to note that activin has an important role in Th2-driven airway inflammation because it has been shown to promote an IL-4-induced shift toward IgE secretion, which is one of the features of airway inflammatory responses (38).

Despite the fact that our experiments revealed a number of intriguing alterations in the expression pattern of several TGF- family components, we are unable to distinguish which of these changes are part of the tissue's attempt to protect itself and which contribute to tissue-damaging processes. Future experiments using adenoviral-mediated gene transfer to locally administer constitutive active or dominant negative forms of these receptors will provide more clear information and indicate which of the TGF- receptors and ligands play a determinative role during the development of allergic airway inflammation. Furthermore, the animal model used in the present study does not recapitulate the entire spectrum of changes that underlie the pathophysiology of human asthma or chronic obstructive pulmonary disease (COPD) and reflects only the changes occurring during acute allergic inflammation in asthma. Therefore, only the analysis of biopsy material from patients with respiratory diseases such as asthma and COPD will illuminate the relevant importance of different TGF- family members in respiratory pathophysiology. The availability of the antibody reagents and the optimized protocols for using them in immunohistochemistry, described in the present study, should greatly facilitate such an endeavor.