Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The pathogenesis of chronic lung injury following mechanical ventilation, oxygen exposure, or infection is complex. Recent evidence suggests that the inflammatory response of the lung to these noxious factors is an important component of the injury. Increased concentrations of proinflammatory mediators such as interleukin (IL)-1 and tumor necrosis factor (TNF)- are found in bronchoalveolar lavage fluid from infants that develop chronic lung disease and adults with acute respiratory distress syndrome (ARDS) (1). In contrast, the anti-inflammatory cytokines IL-10 and IL-11 may ameliorate lung injury (2, 3). Sources of cytokine production in lung include alveolar macrophages, fibroblasts, epithelial cells, and endothelial cells. Local release of cytokines leads to an amplification of the inflammatory response through recruitment and adhesion of peripheral inflammatory cells to the alveolar epithelium. Inflammatory mediators released at sites of injury interact with receptors on lung fibroblasts to increase cell proliferation and protein production. In this way, the inflammatory response generated during acute injury may contribute to the excessive fibroblast proliferation and extracellular matrix production that are hallmarks of diseases, such as bronchopulmonary dysplasia in the infant and pulmonary fibrosis in the older child and adult.

Experimental evidence supports a role for insulin-like growth factor-I (IGF-I) and IGF-II in the lung pathology associated with chronic lung disease. IGF-I and IGF-II are small peptides produced within lung that are important regulators of cell proliferation and differentiated cell function (4). In patients with inflammatory lung disease, alveolar fluid contains IGF-I that is mitogenically active (5). Several cell types may produce IGF-I following an inflammatory response. Lung biopsies from patients with idiopathic pulmonary fibrosis show abundant IGF-I expression in interstitial mesenchymal cells, epithelial cells and macrophages (6). In addition, alveolar macrophages isolated from lungs of patients with pulmonary fibrosis or various animal models of pulmonary fibrosis express IGF-I (6, 7). The interaction of IGF-I with a cell surface receptor, the type 1 IGF receptor (IGF-1R) results in increased collagen formation (8) and pulmonary fibroblast proliferation (9), thus potentially contributing to the pathogenesis of inflammatory lung injury.

The activity of IGF-I and IGF-II is regulated by a group of six structurally related binding proteins, IGF binding proteins-1 through -6 (IGFBP-1 to -6) (10). IGFBPs are secreted by lung cells (11) and tightly bind IGFs in the pericellular space, controlling the availability of free IGF and access of IGFs to the IGF-1R. Recent studies have demonstrated that many trophic factors and hormones modulate IGF activity through regulation of IGFBP abundance (12, 13). Various proinflammatory cytokines regulate IGFBPs in cells derived from bone, breast carcinoma, and reproductive tissues (14). Whether cytokines regulate IGFBP production by lung cells is unknown. Because the IGF system regulates mesenchymal cell growth and lung fibroblast overproliferation is a key component of inflammatory lung injury, we chose to investigate the regulation of lung fibroblast IGFBPs by various cytokines associated with the pathogenesis of lung injury.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials

The following reagents and materials were purchased: calf serum (CS) from Hyclone (Logan, Utah); minimal essential medium, mIL-1, and mTNF- from GibcoBRL (Gaithersburg, MD); mIL-1 from Endogen (Woburn, MA); rIL-10 and mIL-11 from Peprotech (Rocky Hill, NJ), trypsin and DNase I from Sigma Chemical Co. (St. Louis, MO); IGF-I and IGFBP-4 from GroPep (Adelaide, Australia); IGFBP-3 from Upstate Biotechnology (Lake Placid, NY); and tissue culture flasks and multiwell plates from Costar (Cambridge, MA). The IGFBP-4 antisera was a gift from Dr. David Clemmons (University of North Carolina at Chapel Hill, Chapel Hill, NC).

Cell Culture

Primary cultures of fetal rat lung fibroblasts were prepared using differential adherence as described previously (11) with some modification. Rat pups were delivered aseptically from 19-d gestation timed pregnant Sprague-Dawley rats (Hilltop Laboratory Animals, Scottdale, PA). The lungs were removed, minced, and incubated with trypsin (0.05%) and DNase (10 mg/ml) for 20 min at 37°C. Trypsin activity was stopped by adding MEM with penicillin (100 U/ml), streptomycin (100 U/ml), and 10% charcoal-stripped calf serum (MEM/CS-). The cells were filtered through a 50-µm Nitex mesh and collected by centrifugation, then resuspended in a RBC lysis solution (15 mM Tris and 150 mM NH₄Cl, pH 7.2) and incubated for 10 min at 37°C. Cell clumps trapped by the mesh filter were again digested with trypsin and DNase (20). After centrifugation, the cells were resuspended in MEM/CS- and incubated in 75 cm² tissue culture flasks. After 45 min, the nonadherent cells were removed and fresh MEM/CS- was added to the flask. The adherent cells are predominantly fibroblasts and are termed primary fibroblast cultures in this study. Previous studies have demonstrated that this method yields primary cultures of fibroblasts having ~ 85-90% purity as judged by morphology and vimentin immunostaining (20). Cells were grown in 5% CO₂/95% air at 37°C.

Preparation of Conditioned Medium and Cell Lysate

Cells were detached by trypsinization within 48 h of the initial harvest and plated at a concentration of 3.5 × 10⁴ cells/cm² in 6- or 24-well well plates. After plating, the cells were washed twice with serum-free MEM (SFM), and incubated in SFM, with or without added cytokines, for up to 48 h. At the end of the experimental period, conditioned media (CM) were collected and frozen at 20°C. For assessment of cellular DNA content, 0.1% SDS was added to the cell layer in each well and collected after incubation for 1 h at 42°C. The DNA content for each well was determined using a DNA fluorimeter (Hoeffer Scientific, San Francisco, CA) (21). For each treatment condition tested, DNA content varied by less than 15% from untreated control cells.

To examine IGFBPs in cell lysates (22), cell monolayers were washed twice with SFM, then 1% Triton X-100 containing 5 mM 1,10-phenanthroline was added to each well. The resulting cell lysate was centrifuged at 12,000 × g for 30 min at 4°C and the protein content of the supernatant measured using a Bradford assay. IGFBP abundance in cell lysate was determined by ligand blot analyses.

To prepare cell-free CM for IGFBP-4 proteolysis experiments, CM was collected and centrifuged at 300 × g for 10 min at 4°C to pellet cells and cell debris. Aliquots of the cell-free CM were then incubated at 4°C or 37°C for the indicated times and IGFBP-4 abundance analyzed by ligand blot or immunoblot.

Ligand Blot Analysis of CM

Ligand blot analysis was performed as described (11). CM samples were lyophilized and resuspended in water. For cell extracts, 70 µg of supernatant protein was aliquoted for each lane. After addition of Laemmli buffer and heating to 95°C for 3 min, samples were centrifuged and the supernatant subjected to electrophoresis through a 12.5% SDS-polyacrylamide gel under non-reducing conditions. The proteins were transferred to nitrocellulose membranes (Schleicher and Schuell Inc., Keene, NJ) and incubated with 100,000 cpm/ml of [¹²⁵I]IGF-I. After washing, the nitrocellulose membranes were exposed to Kodak Biomax MS autoradiography film (Eastman Kodak, Rochester, NY) at 70°C. Autoradiograms were analyzed by densitometry using the Image-Pro system (Media Cybernetics, Silver Spring, MD) to determine the relative abundance of each IGFBP species.

Immunoblot Analysis of CM

After ligand blotting nitrocellulose membranes were washed in 50 mM Tris, 0.2 M NaCl, and 5% nonfat milk and then incubated with a 1:2,000 dilution of rabbit antiserum against human IGFBP-4 (23). After washing, blots were incubated with goat anti-rabbit IgG biotin-SP conjugate (1:5,000; Jackson Laboratories, West Grove, PA). Blots were again washed and incubated with a streptavidin-peroxidase conjugate (1:5,000, Jackson Laboratories). Antibody complexes were visualized with Amersham ECL detection reagents (Arlington Heights, IL) and Kodak Biomax MS autoradiographic film. Fetal rat serum was used as a positive control for IGFBP-4. Negative controls (no primary antisera) recognized no proteins of M_r less than 50,000. There is no significant crossreactivity between IGFBPs for the antisera used (23, 24).

RNA Isolation and RNase Protection Assay

RNA was isolated from cells using Trizol reagent (Gibco-BRL, Gaithersburg, MD) according to the manufacturer's instructions. RNase protection assays were performed using the RPAIII kit (Ambion, Austin, TX). A 220 nucleotide biotin-labeled antisense riboprobe was transcribed using a rat IGFBP-4 cDNA fragment that contained 202 nt of the IGFBP-4 cDNA and a 301 nucleotide biotin-labeled riboprobe was transcribed using a rat IGFBP-3 cDNA fragment that contained 257 nt of the IGFBP-3 cDNA. For -actin the 150 nucleotide probe contained 125 nucleotides of the rat -actin cDNA. Five to ten micrograms of total RNA was hybridized with the probes overnight at 45°C, followed by RNase A/RNase T1 digestion. The protected fragments were separated on a 5% polyacrylamide/8 M urea gel and then transferred to BrightStar nylon membranes (Ambion Incorporated, Austin, Texas). Labeled probe was detected using the BrightStar BioDetect kit (Ambion Incorporated) according to the manufacturer's directions after exposure to Biomax MS film (Eastman Kodak). Liver RNA was used as a positive control. Labeled probe hybridized with yeast RNA resulted in no protected fragments. The relative abundance of the IGFBP-4 RNA fragments was standardized for the abundance of -actin mRNA using densitometry as described for ligand blots.

Statistical Analysis

Each experiment was performed on at least three separate occasions using different primary cell culture preparations for each experiment. Statistical differences between groups within an experiment were determined by ANOVA and Dunnett's test for multiple comparisons to a control group using SigmaStat (version 1.0; Jandel Corporation, San Rafael, CA). Results were considered significant for P values less than 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

IGFBPs Secreted by Fetal Rat Lung Fibroblasts in Response to Cytokines

To determine if exposure of fetal rat lung fibroblasts to cytokines results in alteration of IGFBP accumulation in CM, cells were plated and exposed to the following cytokines implicated in the pathogenesis of lung disease: TNF- (20, 100, and 200 ng/ml), IL-1 (2, 10, and 20 ng/ml), IL-1 (10 and 20 ng/ml), IL-10 (2, 10, and 20 ng/ml), or IL-11 (2, 10, and 20 ng/ml) for 48 h. The abundance of IGFBPs in CM was analyzed by ligand blot (Figure 1, Table 1). Cells incubated with TNF-, IL-1, IL-1, or IL-11 demonstrated increased abundance of CM IGFBP-3. Cells incubated with TNF-, IL-1, IL-1, IL-10, or IL-11 showed an increase in IGFBP-4 abundance in CM.

View larger version (47K): [in this window] [in a new window]   Figure 1.   Representative autoradiographs of ligand blots showing IGFBP abundance in CM from fetal rat lung fibroblasts following treatment with cytokines for 48 h. To determine if exposure of fetal rat lung fibroblasts to cytokines resulted in alteration of IGFBP accumulation in CM, cells were plated and exposed to various cytokines for 48 h and the abundance of IGFBPs determined by ligand blot.

To examine the response of fetal rat lung fibroblasts to IL-1 and TNF- in more detail, cells were isolated, plated, and then exposed to various doses of IL-1 (0.1 to 20 ng/ml) or TNF- (2-200 ng/ml) for 48 h. CM was collected and analyzed by ligand blot. Figure 2 shows representative ligand blots and densitometric analyses of experiments demonstrating that both IGFBP-3 and IGFBP-4 increase in a dose-dependent manner in response to TNF- and IL-1. The lowest concentration of IL-1 that stimulated IGFBP-3 and IGFBP-4 was 2 ng/ml and the lowest effective concentration for TNF- was 20 ng/ml for IGFBP-3 and 60 ng/ml for IGFBP-4. Immunoblot analysis confirmed that the changes in the 24 kD ligand blot band from CM samples corresponded to changes in the 24 kD IGFBP-4 band identified by the IGFBP-4 antisera (data not shown).

View larger version (39K): [in this window] [in a new window]   Figure 2.   Autoradiograph of ligand blot and graphic representation of a representative experiment showing IGFBP abundance in CM from fetal rat lung fibroblasts following treatment with various doses of IL-1 or TNF- for 48 h. Fetal rat lung fibroblasts were treated with SFM alone or indicated cytokines for 48 h and CM analyzed by ligand blot (upper panels). The graphs show densitometric analyses of ligand blot bands. Bars represent the mean ± SEM percent of SFM control samples for IGFBP-3 and IGFBP-4. *P < 0.05 versus SFM control. Solid bars, IGFBP-3; open bars, IGFBP-4.

Mechanisms Regulating IGFBP-3 and IGFBP-4 in Fetal Lung Fibroblasts

To determine if the increase in IGFBP-3 and IGFBP-4 following cytokine exposure is due to release of IGFBPs from the cell surface, fetal rat lung fibroblasts were plated, exposed to IL-1 and TNF-, and cell lysate analyzed by ligand blot. There was a large increase in cell-associated IGFBP-3 and a small increase in cell-associated IGFBP-4 in lysates from cells exposed to IL-1 and TNF-, indicating that increased CM IGFBP-3 and IGFBP-4 are not due to a release of cell-associated IGFBP (data not shown).

To determine whether the increase in IGFBP-3 or IGFBP-4 is associated with an increase in mRNA abundance, cells were treated with IL-1 (10 ng/ml) or TNF- (50 ng/ml) for 18 or 48 h and mRNA abundance assessed using RNase protection assays. Figure 3A shows the small increase in IGFBP-4 and larger increase in IGFBP-3 mRNA following exposure to either cytokine. After 48 h exposure, IGFBP-3 but not IGFBP-4 mRNA was increased following exposure to TNF- or IL-1. More detailed examination of IGFBP-3 mRNA transcript abundance demonstrated that the increase in IGFBP-3 mRNA was evident at doses of IL-1 and TNF- as low as 10 ng/ml and 20 ng/ml, respectively (Figure 3B). Addition of cycloheximide (CHX) at a dose that prevented new protein synthesis (10 µg/ml) did not alter baseline IGFBP-3 but blocked the increase in IGFBP-3 mRNA following addition of IL-1 and TNF- (Figure 3C).

View larger version (28K): [in this window] [in a new window]   Figure 3.   RNase protection assays for IGFBP-3 and IGFBP-4 using mRNA from cells treated with cytokines. (A) Autoradiograph of RNase protection assay showing the protected IGFBP-3 and IGFBP-4 bands in RNA samples from cells treated with SFM, IL-1 (10 ng/ml) or TNF (100 ng/ml) for 18 h. Liver RNA was used as a positive control (+Con). -actin was used as an internal control. Undigested IGFBP-3, IGFBP-4 and -actin probes (Probes) are shown at right. (B) RNase protection assays showing the IGFBP-3 protected band in samples of RNA from cells treated with various doses of IL-1 or TNF- for 48 h. Liver RNA was used as a positive control (+C). Densitometric analysis of IGFBP-3 bands is expressed as a percent of SFM control (S) after correcting for -actin abundance. (C) Autoradiograph of RNase protection assay showing the protected IGFBP-3 bands in RNA samples from cells treated with SFM, IL-1 (10 ng/ml, IL), or TNF- (60 ng/ml, Ta) with and without cycloheximide (10 µg/ml, CHX) for 18 h. Densitometric analysis of IGFBP-3 bands is expressed as a percent of SFM control after correcting for -actin abundance.

IGFBP-4 Proteolysis is Regulated by Cytokines

For rat lung fibroblasts, the clearance of IGFBP-4 from CM is regulated by an IGFBP-4-specific protease (25). The activity of the IGFBP-4 protease in CM is accentuated in the presence of IGF-I and inhibited by 1,10-phenanthroline, an inhibitor of metalloproteinases. To examine alterations in IGFBP-4-specific proteolysis in fetal lung fibroblasts, CM was collected from 19-d fetal rat lung fibroblasts after incubation with SFM, IL-1 (10 ng/ml) or TNF- (50 ng/ml) for 48 h. Aliquots of cell-free CM (SFM-CM, IL-CM and TNF-CM) were then incubated at 4°C or 37°C for 24 h and IGFBP abundance determined by ligand blot analysis. Incubation of cell-free 48-h SFM-CM at 37°C for 24 h decreased IGFBP-4 abundance (Figure 4, Table 2). The decrease in IGFBP-4 was accentuated by the addition of IGF-I during the cell-free incubation period and was inhibited by 1,10-phenanthroline (not shown). Immunoblot analysis of SFM-CM confirmed that the decrease in the 24,000 M_r ligand blot band was associated with a reduction in intact immunoreactive IGFBP-4 (Figure 4). In addition, the IGFBP-4 antisera identified an ~ 18,000 M_r IGFBP-4 fragment in the SFM-CM samples. Of note, IGFBP-3 was not different between samples, indicating that the loss of IGFBP-4 was not due to generalized proteolysis in the sample. In contrast, CM from cells exposed to IL-1 or TNF- incubated at 37°C for 24 h did not show a decrease in the 24,000 ligand blot band or the immunoreactive IGFBP-4 band. Addition of IGF-I fully restored, and 1,10-phenanthroline inhibited, IGFBP-4 proteolysis in IL-CM and TNF-CM. We next determined whether the decrease in IGFBP-4 proteolysis in IL-CM and TNF-CM was a result of the high concentration of IGFBP-4 in these samples, making the proteolysis step rate-limiting. To test this, exogenous IGFBP-4 (1.25 nM) was added to SFM-CM (SFM+4-CM) to approximate the IGFBP-4 abundance in IL-CM and TNF-CM. Under these conditions, a 24-h incubation at 37°C resulted in a similar degree of IGFBP-4 proteolysis in SFM-CM and SFM+4-CM samples (65% and 62% decrease in IGFBP-4 abundance after a 24-h incubation at 37°C in SFM-CM and SFM+4-CM samples, respectively). These results suggest that, under these conditions, the increased concentration of IGFBP-4 in CM induced by IL-1 and TNF- is not a rate-limiting step in IGFBP-4 proteolysis.

View larger version (60K): [in this window] [in a new window]   Figure 4.   Autoradiograph of representative ligand blot demonstrating proteolysis of IGFBP-4 in CM. Medium was collected from fetal rat lung fibroblasts after a 48-h incubation in SFM (SFM-CM), IL-1 (IL-1-CM), or TNF- (TNF-CM). (A) Ligand blot and corresponding IGFBP-4 immunoblots of CM incubated for 24 h at 4°C or with and without IGF-I at 37°C. In SFM-CM, the 24,000 Mr intact IGFBP-4 band is diminished following incubation at 37°C. Also visible in SFM-CM samples is an 18,000 Mr IGFBP-4 fragment. Note that the SFM-CM ligand blot autoradiograph is overexposed compared with the other autoradiographs to better show the IGFBP-4 band. (B) Ligand blot analysis of aliquots of pooled cell-free CM incubated at 4°C or incubated with and without IGF-I (75 ng/ml) or 1,10 phenanthroline (1,10 PT) for 24 h at 37°C. Addition of IGF-I fully restored and 1,10-phenanthroline inhibited IGFBP-4 proteolytic activity in IL-CM and TNF-CM.

Because both TNF- and IL-1 upregulate IGFBP-3, and IGFBP-3 is known to regulate IGFBP-4 protease activity (26, 27), we next designed experiments to determine whether IGFBP-3 abundance in lung cell CM regulates IGFBP-4 proteolysis. To dilute the influence of endogenous IGFBP-3 in the cell-free proteolysis assays, we next performed experiments using small aliquots of CM containing protease activity. Exogenous rhIGFBP-4 (3 nM) was added to a pool of 48-h SFM-CM (SFM+4-CM) and aliquots of this medium then used in cell-free proteolysis experiments. The presence of excess IGFBP-4 during the 18-h incubation at 37°C resulted in a smaller decrease in IGFBP-4 abundance compared with the experiments shown in Figure 4 (decrease of 34 ± 17% compared with 4°C samples; n = 4 experiments). Addition of IGF-I (6 nM) accelerated the proteolysis of IGFBP-4, as expected (Figure 5A). Addition of IGFBP-3 to the cell-free aliquots of 48-h SFM+4-CM reversed the IGF-I-mediated acceleration of IGFBP-4 proteolysis.

View larger version (39K): [in this window] [in a new window]   Figure 5.   Autoradiographs of representative ligand blots showing the abundance of IGFBP-4 in CM following the addition of IGFBP-3. (A) The effect of exogenous IGFBP-3 on IGF-dependent clearance of IGFBP-4 in cell-free CM. To dilute the influence of endogenous IGFBP-3, IGFBP-4 (6 nM) was added to a pool of 48-h SFM-CM and small aliquots (80 µl) were then incubated with IGF-I (6 nM) and various amounts of IGFBP-3 at 4°C or 37°C for 18 h. Note that the addition of exogenous IGFBP-4 reduces the loss of IGFBP-4 when IGF-I is not present. Addition of exogenous IGFBP-3 to cell-free CM reverses the IGF-dependent clearance of exogenous IGFBP-4 in a dose-dependent manner. (B) Exogenous IGFBP-3 added to cells in culture. Cells were plated, exposed to SFM for 24 h, then SFM or IGFBP-3 was added to the incubating cells for an additional 24 h. At the end of the incubation period, medium was collected and IGFBPs analyzed by ligand blot. Addition of exogenous IGFBP-3 to cells in culture reverses the time-dependent clearance of IGFBP-4 in a dose-dependent manner. (C) IGF-I reverses the IGFBP-3-mediated inhibition of IGFBP-4 proteolysis. Cells were plated, exposed to SFM for 24 h, then SFM, IGFBP-3 (2 nM) or IGFBP-3 plus IGF-I (2 nM) was added to the incubating cells for an additional 24 h. At the end of the incubation period, medium was collected and IGFBPs analyzed by ligand blot.

To determine if IGFBP-3 similarly regulates CM IGFBP-4 abundance for cells in culture, primary fetal rat lung fibroblasts were plated, washed, and then exposed to SFM for 24 h to allow the IGFBP-4 protease to accumulate. Without removing the medium, IGFBP-3 (0.1-2 nM) or IGFBP-3 and IGF-I (2 nM each) were then added to the medium and the cultured cells incubated for an additional 24 h. As expected, the abundance of IGFBP-4 declined between 24 and 48 h of incubation due to accumulating IGFBP-4 protease activity (25). This decline was not observed in CM from cells to which exogenous IGFBP-3 had been added (Figure 5B). Immunoblot analysis confirmed that the increased abundance of the 24,000 M_r ligand blot band was due to increases in intact IGFBP-4 (data not shown). Addition of both IGF-I (2 nM) and IGFBP-3 (2 nM) again resulted in decreased CM IGFBP-4 compared with CM with no additives (Figure 5C). Smaller amounts of exogenous IGF-I did not induce IGFBP-4 proteolysis (data not shown), suggesting that IGFBP-3 is able to inhibit IGF-mediated IGFBP-4 proteolysis only when sufficient IGFBP-3 is present to sequester free IGF-I.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

These studies demonstrate that various cytokines regulate IGFBP accumulation in fetal rat lung fibroblast CM. TNF-, IL-1, IL-1, and IL-11 increase both IGFBP-3 and IGFBP-4 accumulation in CM and IL-10 increases IGFBP-4 accumulation. Regulation of lung cell IGFBPs by cytokines is consistent with various studies showing that IGFBPs are controlled by hormones, growth factors, or other agents important to the growth or development of that tissue (10). Previous studies of cytokine-IGFBP interactions have examined bone and reproductive tissues (14, 28) but not lung, and have not included studies of the anti-inflammatory cytokines IL-10 and IL-11. Regulation of IGFBPs by cytokines suggests a mechanism by which cytokines modulate IGF activity in lung.

Increased IGFBP-3 abundance in response to growth factors and cytokines may result from increased production, decreased proteolysis, or altered cell localization. The present study and our previous investigations show that there is no significant proteolysis of IGFBP-3 in CM from unstimulated fetal lung fibroblasts or from fibroblasts exposed to the cytokines tested (21). Also, our results did not demonstrate a decrease in IGFBP-3 associated with the cell surface following cytokine exposure, suggesting that altered cell localization does not contribute to increased CM IGFBP-3 following cytokine exposure. Coupled with the observed increase in IGFBP-3 mRNA transcript abundance following IL-1 and TNF-, these findings indicate that increased accumulation of IGFBP-3 in CM is due to increased production of IGFBP-3. Increased IGFBP-3 mRNA in response to cytokines also has been shown following exposure of Leydig cells to IL-1 (15) and exposure of breast cancer and Sertoli cells to TNF- (14, 16). The mechanism by which cytokines induce IGFBP-3 production is not known. IL-1 is known to act through cAMP (29) and increased cAMP has been shown to increase IGFBP-3 transcription and IGFBP-3 mRNA stability independent of new protein synthesis (30). In contrast, IGFBP-3 does not accumulate in fetal rat lung fibroblast CM following an increase in intracellular cAMP (11). Moreover, our results demonstrate that cycloheximide blocks the increase in IGFBP-3 mRNA following either IL-1 or TNF-. Taken together, these findings suggest that induction of IGFBP-3 by IL-1 is likely not dependent on increased intracellular cAMP, but is dependent on synthesis of new protein.

The increase in fetal rat lung fibroblast CM IGFBP-4 in response to both IL-1 and TNF- resulted, in part, from decreased IGFBP-4 proteolysis. Several mechanisms may explain the cytokine-associated change in IGFBP-4 protease activity. Our experiments demonstrated that the decrease in IGFBP-4 proteolysis following cytokine exposure occurred concomitantly with an increase in IGFBP-3, a known inhibitor of IGFBP-4 protease activity (26). Moreover, the decrease in IGFBP-4 is inhibited following the addition of exogenous IGFBP-3 to cells in culture or to cell-free CM, suggesting that IGFBP-3 directly or indirectly inhibits IGFBP-4 proteolysis. One mechanism proposed to explain the regulation of IGFBP-4 proteolysis by IGFBP-3 is that the binding of IGF-I to IGFBP-4 changes the conformation of IGFBP-4 making it more susceptible to proteolytic cleavage (27, 31, 32). Thus, excess IGFBP-3 may indirectly inhibit IGFBP-4 proteolysis by sequestering IGF-I and decreasing the amount of IGFBP-4 susceptible to proteolysis. Our results support the lack of available IGF-I as a mechanism for decreased IGFBP-4 proteolysis, as exogenous IGF-I reversed the ability of IGFBP-3 to protect IGFBP-4 from proteolysis. However, these results do not exclude the possibility that the binding of IGF-I to IGFBP-3 alters IGFBP-3 so that IGFBP-3 can no longer directly inhibit the IGFBP-4 protease (26). The decrease in IGFBP-4 proteolysis following incubation with IL-1 or TNF- is not likely a result of decreased protease production or lack of a protease cofactor because addition of IGF-I to cell-free CM completely restores IGFBP-4 proteolytic activity in TNF-CM and IL-CM. Both IL-1 and TNF- have been shown to decrease IGF-I production (18, 32). Taken together, these findings suggest that cytokines regulate fetal rat lung IGFBP-4 protease activity through multiple mechanisms including decreasing IGF-I production, increasing IGFBP-3 and by decreasing the amount of free IGF-I available to bind IGFBP-4. In addition, the regulation of IGFBP-4 by IL-1 and TNF- differs for cells derived from other organs (15, 19, 28), highlighting the importance of cell-specific factors in regulating IGFBP-4.

Several observations suggest that regulation of IGFBPs by cytokines has important biologic significance. Evidence from in vitro studies strongly suggests that one mechanism by which growth factors and cytokines affect cell proliferation and cell function is through regulation of IGFBPs and subsequent modulation of IGF activity. The growth inhibitory activity of TNF- as well as other agents involves increased IGFBP-3 accumulation in conditioned medium (12, 13, 16). Likewise, IGFBP-4 inhibits cell proliferation (33). IGFBPs may sequester IGF-I and prevent IGF-I from activating the IGF-1R (33) or may regulate cell proliferation directly, independent of IGF-I, likely through interaction with cell surface receptors and activation of a growth inhibitory signaling pathway (10). Expression of an IGFBP-3 receptor by lung fibroblasts has been reported (37) although its role in lung cell proliferation is unknown.

Inflammatory cytokines have been implicated in the pathogenesis of chronic lung injury due to barotrauma, oxygen toxicity, and infection. TNF- and IL-1 enhance fibroblast proliferation and IL-1 stimulates collagen and fibronectin production by fibroblasts, although the mechanism by which cytokines regulate these cell processes is unknown (17). Inadequate IL-10 following lung injury has been implicated in the pathogenesis of lung inflammation and ARDS (38, 39). Several studies also suggest a role for increased IGF-I in the pathogenesis of pulmonary fibrosis. IGF-I increases pulmonary fibroblast collagen formation (8) and proliferation (9), is present in alveolar fluid from patients with inflammatory lung disease (5) and is abundantly expressed in lung biopsies from patients with pulmonary fibrosis (6). IGFBPs are known to modulate IGF-I activity and are increased at sites of inflammation (6). In the present study, we add to the link between the IGF system and inflammation by demonstrating that various cytokines regulate IGFBP production. Interestingly, both pro- and anti-inflammatory cytokines similarly upregulate lung fibroblast IGFBP-3 and IGFBP-4. It is important to recognize that IGF activity is determined by the sum of many different inputs, such as IGFBP accumulation and localization, IGF receptor expression, and IGF production. Thus, these diverse cytokines may have differing effects on IGF activity by differentially regulating other components of the IGF system. These observations lead us to speculate that cytokines regulate cell proliferation, in part, by altering IGFBP accumulation in the pericellular space.