INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Inflammatory lung injury may result in an exaggerated reparative response with accumulation of extracellular matrix components. Lung fibroblasts play important roles in these processes (1). Differences in activation and functional phenotypes of fibroblasts from different fibrotic processes have been demonstrated in previously published studies (4). What has not been established to date is a comparison of functional phenotypes in fibroblasts obtained from patients with pulmonary fibrosis with a difference in the heterogeneity of fibrotic process.

During wound-healing processes, fibroblast-mediated contraction may favor repair at injured sites, and the strength of cellular traction differs greatly between cell types (7, 8). In pulmonary fibrosis, especially in chronic progressive pathologic processes such as usual interstitial pneumonia (UIP), contraction of acini contributes to functional deterioration (9).

Pathologic studies of the patterns of fibrotic response have demonstrated differences in the degree of variation in the age of histopathologic differences between homogeneity and temporal heterogeneity (10). Repair response to acute lung injury is likely to result in temporally homogeneous forms of fibrosis that can be found in acute interstitial pneumonia (AIP), bronchiolitis obliterans organizing pneumonia (BOOP), and nonspecific interstitial pneumonia (NSIP), whereas exaggerated reparative processes in response to chronic lung injury are more likely to result in temporally heterogeneous forms of fibrosis such as that found in UIP (10).

In a previous study we found that the rates of proliferation of cultured fibroblasts obtained from patients with idiopathic pulmonary fibrosis (IPF) pathologically diagnosed as UIP by surgical lung biopsy were no different from those of fibroblasts obtained from normal regions of resected lungs (15). Furthermore, we also showed that there was no difference in apoptotic processes in lung fibroblasts from the same two groups (16).

In a more recent study, we succeeded in demonstrating an effect of glucocorticoids on lung fibroblast contractility which suggested that the responsiveness to glucocorticoids may reflect distinct differences in fibroblast phenotype (17).

This study was designed to test the hypothesis that fibroblasts obtained from biopsies of NSIP display different functions from those obtained from biopsies of UIP.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

Fibroblasts from surgical lung biopsies obtained from five patients with IPF/UIP (five male; mean age, 59.4 yr) and five patients with NSIP (1 male, 4 female; mean age, 56.0 yr) were studied. All patients were classified histopathologically according to the criteria of Katzenstein and Fiorelli (10). Biopsies from five adult patients who underwent thoracotomies for clinically relevant reasons (three primary lung adenocarcinomas, one metastatic lung cancer, one lung hamartoma) were selected as controls (5 male; mean age, 58.4 yr). Informed consent was obtained from each patient according to the Helsinki Declaration. The study was approved by the Ethical Committee of the Graduate School of Medicine, Kyoto University.

Cell Cultures

Lung specimens of the pulmonary parenchyma examined under microscope after carefully removing pleura were minced into 1 to 2 mm pieces in a sterile procedure (6, 18). The minced pieces were washed with Dulbecco's modified Eagle's medium (DMEM; Life Technologies, Inc., Rockville, MD) and then plated in 100-mm dishes (Asahi Techno Glass Corp., Tokyo, Japan). The specimens were cultured with DMEM supplemented with 10% FCS, 2 mM glutamine, 50 U/ml penicillin, and 50 µg/ml streptomycin at 37° C in an atmosphere of 5% CO₂. When fibroblasts reached confluence, the cells were detached by a 10-min treatment with 0.05% trypsin/0.53mM EDTA·4Na (Life Technologies) and then subcultured at a 1:4 ratio. The cells were studied at passages 4 or 5. To identify each cell strain as fibroblast, cultured cells were immunostained with antismooth muscle -actin, antivimentin, antidesmin, and antikeratin antibody, respectively (Sigma Chemical Company, St. Louis, MO) (17, 19).

Preparation of Fibroblasts and Their Fractions in Conditioned Media

Fibroblasts (1 × 10⁵ cells/ml) were plated in 100-mm dishes until they achieved confluence. After several washes, cells were cultured with 4 ml of serum-free DMEM for 24 h, debris was removed by centrifugation, and the supernatant was stored at 80° C until use. Cell viability was confirmed before and after incubation with conditioned media using MTT [3-(4,5-dimethylthiazol-2yl)-2,5-diphenyltetrazolium bromide] (Sigma Chemical) (20). A 5 cm² × 40 cm column of Sephacryl HR 100 (Amersham Pharmacia Biotech, Inc., Piscataway, NJ) at a flow rate of 20 ml/h at 4° C, equilibrated in Hanks' balanced salt solution (pH, 7.4), was used for fractionation.

Measurement of Gel Contraction

Fibroblast contractility was assessed by measuring changes in the surface area of type I collagen gels mediated by fibroblasts (17, 21). Serum-free DMEM was used in contraction assay in order to exclude the modulation of growth factors contained in serum (22). Change in surface areas (% Contraction) was expressed as: ( change in surface gel area)/(initial surface gel area) × 100. The gels consisted of type I collagen (0.75 mg/ml) and cell suspension (5 × 10⁵ cells/ml) in HEPES-buffered DMEM (pH, 7.4). After fibroblast culture in the gels for variable times, the gels were released from the plate and measured by an image scanner connected to a computer running a public domain NIH image program (version 1.61; available by anonymous FTP from ).

Measurement of F-actin Content

Fixed and permeabilized fibroblasts (5 × 10⁵ cells/ml) within collagen gels were incubated for 30 min with rhodamine-labeled phalloidin at the concentration at which F-actin is fully saturated (2 × 10⁷ M) (17, 24, 25). The bound phalloidin was extracted by adding 500 µl of 0.1 N NaOH and neutralized with 1.0 M TRIS-HCl at pH 7.4. After centrifugation at 13,000 rpm at 4° C for 15 min, the intensity of rhodamine fluorescence was measured (excitation at 540 nm; emission at 575 nm) using a fluorescence spectrophotometer (model F-3000; Hitachi Inc., Tokyo, Japan). Contractility and F-actin content were assessed in separate experiments.

Assessment for Fibronectin

The concentration of fibronectin in conditioned media was measured using an enzyme-linked immunosorbent assay (ELISA) kit (Biomedical Technologies Inc., Stoughton, MA). The sensitivity was approximately 25 ng/ml of fibronectin. There was no cross reaction with other proteins. Conditioned media were treated through a gelatin-Sepharose 4B affinity column (Amersham Pharmacia Biotech; bed volume, 1 ml, 0.9 × 1.6 cm) equilibrated in Hanks' balanced salt solution (pH, 7.4) at room temperature. Sepharose 4B columns (Amersham Pharmacia Biotech) were used for controls. Binding of fibronectin to the affinity column was confirmed by washing with urea.

Assessment for TGF-1

TGF-1 concentrations in conditioned media were measured by growth inhibitory assay (26) using mink lung epithelial cells (ATCC CCL-64). The cells were plated at 3 × 10⁴ cells per well and allowed to attach for 24 h. The medium was replaced with DMEM containing 0.1% BSA and TGF- (R&D Systems, Inc., Minneapolis, MN) at 5, 10, 25, 100, 250, or 500 pg/ml for 14 h at 37° C, or replaced by test samples under the same conditions. Cell growth was examined using MTT assay. The sensitivity of the assay was approximately 50 pg/ml of TGF-1. The absorbance at 595 nm was measured using a Multiscan MCC/340 (Labsystems, Helsinki, Finland).

To examine whether TGF- contributes to the increase in fibroblast contractility, polyclonal antihuman TGF- rabbit antibody (IgG1, pan-specific neutralizing antibody; R&D systems) was added at a final concentration of 0.1 µg/ml, the gels were incubated for 12 h, and then the contraction was assessed. Type- and class-matched nonimmune Ig controls (rabbit antihuman albumin antibody) were used to demonstrate specificity of the antibody.

Immunoblotting for Cellular Fibronectin and TGF-1

Samples were mixed with Laemmil's sample buffer (27), separated by electrophoresis on a sodium dodecyl sulfate-10% agarose gel, and electrotransferred onto nitrocellulose membrane (Hybond ECL; Amersham Pharmacia Biotech). Detection was performed by an initial immunostaining with either a monoclonal antibody specific for the ED-A epitope of human cellular fibronectin (CHEMICON International, Inc., Temecula, CA) or a monoclonal antibody against human-TGF-1 (R&D systems), a second immunostaining by peroxidase-conjugated goat antimouse IgG, and development using an enhanced chemiluminescence kit (Amersham Pharmacia Biotech).

Statistics

Data were expressed as mean ± SEM of not less than three measurements and analyzed by ANOVA with Scheffe's post-hoc test for comparisons of any two groups using StatView software (Abacus Concepts, Inc., Berkely, CA). Analysis of correlation was achieved using Pearson's correlation coefficients. A p value of less than 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical Characteristics: UIP versus NSIP

Each biopsy specimen underwent strict histopathologic diagnosis, equal numbers of patients were alotted to each disease group, and the groups were matched for age but not sex (Table 1). All patients had dyspnea and fine end-inspiratory crackles. All four patients with UIP and one of five NSIP presented with clubbed fingers at biopsy. Three of the five patients with NSIP had complaints, which suggests collagen vascular disease (Patients 6, 7, and 8, Raynaud phenomenon; Patient 6, myalgia and finger swelling; Patients 7 and 8, dry eye and dry mouth), and all five patients with NSIP were positive for antinuclear antibody in serum. None of the patients with UIP complained of symptoms suggestive of collagen vascular diseases, although two were positive for antinuclear antibody in serum. Only one patient with UIP was treated with glucocorticoid before undergoing surgery (Patient 6), whereas two patients were administered both glucocorticoid and cyclophosphamide after biopsies (Patients 8 and 9). The two groups differed in both the duration of symptoms (UIP, 76.8 ± 20.2 mo; NSIP, 4.0 ± 1.7; p < 0.001) and prognosis (three worsened and two stable in UIP; two stable and three improved in NSIP).

Contractility and F-Actin Content

Lung fibroblasts cultured for 24 h were identified as fibroblasts by negative staining with both antikeratin and antidesmin antibody, and by positive staining with both antivimentin and antismooth muscle -actin antibody. There was no difference between the groups in the percentage of smooth muscle -actin-positive cells (UIP, 90 ± 9% versus NSIP, 93 ± 6 and control, 92 ± 5).

The cells obtained from the UIP group (n = 5) showed increased contractility compared with those obtained from the NSIP group (n = 5) or the control group (n = 5) (Figure 1A). The decrease in the percentage of gel area was significantly higher in the UIP group than in the NSIP group and the control group (mean ± SEM UIP, 72.7 ± 6.2% versus NSIP, 32.8 ± 5.5 and control, 28.5 ± 3.5, p < 0.01).

View larger version (11K): [in this window] [in a new window]   Figure 1.   Comparison of the contractility and intracellular F-actin content in 15 cell strains. (A) Fibroblast contractility in each cell strain was examined by type I collagen gel contraction assay. Fibroblasts within gels were cultured with serum-free DMEM for 24 h. Change in surface areas (% Contraction) was expressed as: ( change in surface gel area)/(initial surface gel area) × 100. (B) F-actin content in each cell strain was determined by phalloidin-binding assay. Fibroblasts within gels were cultured with serum-free DMEM for 24 h. The cells were then fixed and permeabilized, and the binding of phalloidin to F-actin was measured using a fluorescence spectrophotometer. Values indicate mean ± SEM of triplicates.

The relationship between fibroblast-mediated gel contraction and polymerized actin content of the 15-fibroblast strains was investigated in separate experiments under the same culture conditions. The increase in contractility was paralleled by an increase in F-actin content (UIP, 0.35 ± 0.03 expressed by relative fluorescence intensity versus NSIP, 0.21 ± 0.03 and control, 0.21 ± 0.02, p < 0.01) (Figure 1B). There were no significant differences between NSIP and control groups.

Conditioned Media-stimulated Contractility

Conditioned media were prepared by harvesting the supernatant culture medium from 24 h cultures of each fibroblast cell line. After control fibroblasts within gels were cultured in three differently conditioned media for 24 h, the contractility of the control fibroblasts was measured. Conditioned media from UIP fibroblast cultures (n = 5) enhanced control fibroblast contractility, whereas those obtained from NSIP (n = 5) or controls (n = 5) did not (Figure 2A). The decrease in initial gel area was greater in the UIP samples than in the samples from the other two groups (UIP, 49.6 ± 3.8% versus NSIP, 28.3 ± 4.4, control, 28.4 ± 3.4; or medium alone, 18.5 ± 2.3, p < 0.01).

View larger version (14K): [in this window] [in a new window]   Figure 2.   Control fibroblast contractility stimulated by fibroblast-conditioned media. (A) Control fibroblasts obtained from control lung specimens were embedded in collagen gels, and 500 µl of conditioned media obtained from either UIP fibroblasts (n = 5), NSIP fibroblasts (n = 5), control fibroblasts (n = 5), or serum-free DMEM were overlaid on the gels. Contractility was measured after a 24 h incubation with condition medium. (% Decrease in initial gel area is explained in Figure 1.) Values indicate mean ± SEM of triplicates. (B) Conditioned media from UIP or NSIP fibroblasts were fractionated using Sephacryl HR 100 gel chromatography. All fractions were evaluated for contractility in triplicate. Closed squares indicate the conditioned media obtained from UIP, and open circles indicate those obtained from NSIP. (a, b, c, d, and e on the horizontal axis indicate the molecular weights of 158, 67, 43, 25, and 14 kD, respectively.)

Two major peaks of contractility at approximately 180 and 25 kD were observed in both UIP and NSIP samples using gel chromatography (Figure 2B).

There were no differences in the cell viability of each cell strain before and after incubation with conditioned media (UIP, 102.5 ± 14.5%; NSIP, 99.8 ± 15.8; control 101.5 ± 12.6; medium alone, 102.4 ± 9.5) when expressed as the ratio of optical density: (after incubation for 24 h)/(before adding medium) × 100.

Possible Mediators Responsible for Changes in Contractility

Two peaks of the activity were identified as fibronectin (ED-A domain) and TGF-1 by immunoblots, respectively (Figures 3A and B). In all five UIP samples, the amounts of both fibronectin and TGF-1 were increased compared with the amounts in the NSIP and control samples.

View larger version (42K): [in this window] [in a new window]   Figure 3.   Immunoblot analysis of fibroblast-conditioned media from UIP (n = 5: lane 3-7), NSIP (n = 5: lane 8-12), and control samples (n = 5: lane 13-17) using either (A) anti-ED-A fibronectin antibody or (B) anti-TGF-1 antibody. Recombinant human cellular fibronectin (A: lane 1, 250 ng/ml; lane 2, 50 ng/ml) or TGF-1 (B: lane 1, 1,000 pg/ml; lane 2, 200 pg/ml) was used as positive control. Medium alone was used as negative control (lane 18).

Higher amounts of both fibronectin and TGF-1 were contained in the UIP-conditioned media than in the NSIP- and control-conditioned media (fibronectin; UIP, 289 ± 47.1 ng/ml versus NSIP, 121 ± 23.0 and control, 118 ± 16.0; p < 0.01, [Figure 4A]; TGF-1; UIP, 798 ± 119 pg/ml versus NSIP, 246 ± 69.1; and control, 247 ± 53.6; p < 0.01, [Figure 4B]. The contractility was positively correlated with the amounts of fibronectin (Figure 5A: r = 0.867, p < 0.001, n = 15) and TGF-1 (Figure 5B: r = 0.939, p < 0.001, n = 15).

View larger version (13K): [in this window] [in a new window]   Figure 4.   Amounts of either fibronectin or TGF-1 in the fibroblast-conditioned media from 15 cell strains. The amounts of fibronectin were measured by ELISA kit (A), and the amounts of TGF-1 were measured by growth inhibitory assay (B) as described in METHODS. Values indicate mean ± SEM of three determinations.

View larger version (12K): [in this window] [in a new window]   Figure 5.   Relationship between fibroblast contractility and the amounts of fibronectin (r = 0.867, p < 0.001) (A) or TGF-1 (r = 0.939, p < 0.001) (B) in all 15 subjects (five UIP, five NSIP, and five control). Each square represents results from one subject.

The effects of TGF-1 and fibronectin on fibroblast contractility were evaluated independently. Measurement of trace amounts of fibronectin in ELISA assay confirmed that gelatin affinity chromatography did, in fact, deplete the amount of fibronectin in the conditioned media. The contractility in the conditioned media obtained from IPF/UIP patients was reduced by exposure to either the gelatin-Sepharose 4B affinity column or the anti-TGF-1 antibody (Figure 6). The reduction through the affinity column was greater than that obtained by the treatment with anti-TGF-1 antibody (Fibronectin inhibition, 42.5% versus TGF-1 inhibition, 8.5%). Similarly, greater reduction of contractility in NSIP-cultured-condition media was also observed by exposure to the gelatin-Sepharose 4B column (Fibronectin inhibition, 22.2% versus TGF-1 inhibition, 3.2%). It was also confirmed that type- and class-matched control antibodies did not influence contractility.

View larger version (37K): [in this window] [in a new window]   Figure 6.   Contribution of either fibronectin or TGF- to the fibroblast contractility. Conditioned media were initially passed through the gelatin-Sepharose column to remove fibronectin, and next they were exposed to either TGF- antibody (open column), the gelatin-Sepharose column without TGF- antibody (hatched column), the control column without gelatin followed by TGF- antibody (dotted column), or no treatment at all (closed column). Values indicate mean ± SEM of triplicates.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study, we have demonstrated that lung fibroblasts obtained from biopsies with a pathophysiologically demonstrated UIP pattern had greater contractility than fibroblasts obtained from NSIP or control biopsies. Both the contractility and F-actin content were greater in UIP fibroblasts than in NSIP fibroblasts. Conditioned media from UIP fibroblasts contained factors that could increase the control fibroblast contractility. The presence of cellular fibronectin (ED-A domain) and TGF-1 in the conditioned media was identified using immunoblot analysis. Anti-TGF-1 antibody decreased the ability of conditioned media from UIP fibroblasts to contract the collagen gels. Gelatin-Sepharose affinity chromatography of UIP- fibroblast-conditioned media also reduced the gel contraction. These results suggest that fibronectin and TGF-1 are responsible for the contractility. The amounts of both fibronectin and TGF-1 were higher in the UIP-conditioned media than in the NSIP- or control-condition media. The degree of contractility was positively correlated with the amounts of fibronectin and TGF-1 in the conditioned media.

On the basis of current understanding of the spectrum of interstitial pneumonia and fibrosis, there is a critical need to pathophysiologically examine whether there are functional differences between the lung fibroblasts of patients with IPF/ UIP and those with NSIP. We selected patients with IPF/UIP and NSIP who were diagnosed by surgical lung biopsy, and their histologic criteria were intensively reviewed by well-trained pathologists (Kitaichi M and Colby TV). The histopathologic criteria are utilized for differentiation between the UIP and NSIP, according to the review of Katzenstein and Fiorelli (10).

There were sex differences between disease groups. If sex (or hormonal state, for example) modulates fibroblast contractility, then the fact there are little differences between the NSIP group and the control group may be explained in part by sex and not only any severity of disease.

In clinical findings, both groups (UIP versus NSIP) showed clear differences in the disease duration and presence of radiographic honeycombing. The marked difference in the clinical duration may have been related to the differences in the process of pulmonary fibrosis from NSIP to UIP. If UIP biopsies were taken early in the disease process, they might be less contractile like early stage NSIP fibroblasts, and with increasing duration of disease, they might become more contractile. Therefore, current experimental data suggest that differences seen in fibroblast contractility may not necessarily be because of different disease processes (UIP versus NSIP) but that the fibroblast phenotype might be a function of disease duration when the biopsy was taken.

With regard to clinical evidence of collagen vascular disease, all patients with NSIP and two patients with IPF/UIP had serologic evidence of positive antinuclear antibody. However, it remains to be solved whether the presence of collagen vascular disease influences the lung fibroblast functions that were investigated in this study.

In our attempts to identify the culture cells that induce the gel contraction, immunocytochemical staining suggested that the cell strains in the current study had the features of myofibroblasts (4, 19). It is possible that the alteration of contractility may occur during the passage process or culture. However, we compared all fibroblast cell strains in the same culture conditions. The significant increase in contractility found in UIP fibroblasts, therefore, may reflect an altered functional phenotype of fibroblasts in association with the disease itself.

Conditioned media from UIP fibroblasts enhanced control lung fibroblast contractility, whereas those obtained from either NSIP or control fibroblasts did not. The effects manifested were partially due to fibronectin and TGF-1, as shown by the reproducible measurement of TGF-1 and fibronectin. Both factors play critical roles in fibroproliferative disorders (28). When we further investigated whether TGF-1 and fibronectin influence contractility independently, we discovered that fibronectin had a stronger role than TGF-1 in influencing fibroblast contractility in our assay system. The data suggested that TGF-1 upregulated the products of ED-A fibronectin in the fibroblasts (31) and that these products then upregulated the fibroblast contractility dramatically.

In conclusion, UIP fibroblasts showed greater contractility than did NSIP fibroblasts and upregulated control. These findings may be due to a difference in fibroblast phenotype between UIP and NSIP, though this difference may be partly due to a difference in disease duration between both groups.