Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Alveolar type II cells are important for maintaining normal gas exchange in the lung and may be important in limiting fibrosis after lung injury. They produce pulmonary surfactant, proliferate to restore the alveolar epithelium after injury, and transport sodium and fluid from the alveolar space into the interstitium to keep the alveolar space relatively free of fluid. Recently, type II cells also have been shown to secrete a variety mediators, cytokines, and growth factors that are important in the inflammatory process and can serve as paracrine regulatory factors. Type II cells have been shown to produce high levels of eicosanoids, the predominant product being prostaglandin (PG) E₂ (1- 4). In addition, hypertrophic type II cells produce more PGE₂ than do normal type II cells, suggesting that PGE₂ may play an important role in lung injury (2). Rat alveolar type II cells secrete a variety of proinflammatory molecules, and these include interleukin (IL)-6 (5), monocyte chemoattractant protein-1 (6), cytokine-induced neutrophil chemoattractant (7), and granulocyte macrophage colony-stimulating factor (GM-CSF) (8). Moreover, hyperplastic type II cells express transforming growth factor (TGF)- after instillation of silica (9) or bleomycin (10). Many of these factors produced by type II cells could be important during the injury and repair process accompanying chronic inflammation and may serve as paracrine factors to alter the function of other cells in the alveolar microenvironment.

Idiopathic pulmonary fibrosis (IPF) is a common interstitial lung disease for which there is currently no effective treatment. A hallmark of this disease is type II cell hyperplasia at the sites of interstitial fibrosis. In an animal model of pulmonary fibrosis induced by bleomycin instillation, the disease is characterized by loss of alveolar type I epithelial cells, proliferation of fibroblasts, and hyperplasia of type II cells. Some fibroblasts migrate from the interstitium into the air space (11). Control of fibroblast proliferation is thought to be an important goal in the treatment of IPF. The effects of the hyperplastic type II cells on the fibrotic process are not known. One hypothesis is that type II cells reform the epithelium and limit the migration and proliferation of fibroblasts. There is some experimental evidence to support this concept in a murine butylated hydroxytoluene model of pulmonary fibrosis (12). This concept is also well supported by studies of cutaneous wounds and the observation that dissociated airway epithelial cells prevent the overgrowth of fibroblasts in transplanted tracheal explants (15). In addition, there are in vitro studies to indicate that the epithelium and substances found in alveolar fluid can limit mesenchymal cell proliferation (16, 17). Klein and Adamson demonstrated that alveolar epithelial cells exposed to silica could inhibit thymidine incorporation by fibroblasts (17). By this hypothesis, type II cells can be envisioned at the sites of pulmonary fibrosis to limit and contain the fibrotic response. A second hypothesis states that the type II cells make and secrete factors such as platelet derived growth factor (PDGF), TGF-, and TGF- that stimulate fibroblast proliferation and extracellular matrix (ECM) production (18). By this second hypothesis, type II cells could be envisioned as causing or contributing to the fibrotic reaction.

The normal alveolar wall contains relatively few fibroblasts compared with the connective tissue surrounding pulmonary bronchi, arteries, and veins. Because of the ability of type II cells to produce a variety of bioactive molecules, we favored the hypothesis that type II cells inhibit fibroblast proliferation in the normal lung and limit the fibrotic response in the injured lung. The present study was designed to investigate whether rat alveolar type II cells in primary culture produce a factor(s) that inhibits proliferation of human lung fibroblasts. The results of these studies demonstrate that epithelial control of fibroblast proliferation in vitro involves reciprocal interactions between these two cell populations and the autocrine production of PGE₂ by fibroblasts.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Adult Rat Alveolar Type II Cell Isolation and Culture

Rat alveolar type II cells were isolated from lungs of 200-g adult male, specific pathogen-free Sprague-Dawley rats (Bantin-Kingman, Fremont, CA), by tissue dissociation with porcine pancreatic elastase (Worthington Biochemicals, Freehold, NJ) followed by centrifugation over a discontinuous metrizamide gradient, as previously described (21). Cells were suspended in Dulbecco's modified Eagle's medium (DMEM) (GIBCO BRL, Grand Island). The DMEM was supplemented with 10% fetal bovine serum (FBS) (Irvine Scientific, Santa Ana, CA), 2 mM glutamine, 100 U/ml penicillin G, 100 µg/ml streptomycin, 2.5 µg/ml amphotericin B, and 10 µg/ml gentamicin (all from Sigma Chemical Co., St. Louis, MO). The cells were plated at a density of 5 × 10⁵ cells/cm² into six-well cluster dishes (Becton Dickinson, Franklin Lakes, NJ) that were coated with 1 ml of a solution of Matrigel (Collaborative Biomedical, Bedford, MA)/DMEM 2:1, (vol/vol). In some experiments keratinocyte growth factor (KGF) (10 ng/ml) (Promega, Madison, WI) was added to stimulate type II cell proliferation.

Adult Human Lung Fibroblast Culture

Adult human lung fibroblasts (No. AG02262) were obtained from the National Institute of Aging, Cell and Culture Repository, Coriell Institute for Medical Research (Camden, NJ). These fibroblasts have been demonstrated to secrete KGF and hepatocyte growth factor (HGF) (22). The cells were obtained at population doubling 2. Cells from passage 4 or 5 were used in all experiments.

Preparation of Fibroblasts from Normal Adult Rat Lung and Bleomycin-Injured Lung

The left lungs of normal Sprague-Dawley rats were minced to about 1 mm³ in a 100-mm tissue culture dish with sterile scalpel blades. Approximately 10 to 20 small pieces of lung tissue were transferred to a new dish and incubated with a minimal amount of DMEM containing 10% FBS for 30 min to permit the lung tissue to adhere to the dish. Then, 10 ml of DMEM containing 10% FBS were added. At 3 d later the pieces were removed and the adherent fibroblasts were cultured for several more days. When cells were about 50% confluent, the cells were trypsinized, centrifuged, and plated in a new dish at the density of 1 × 10⁵ cells/dish. Those cells were called passage 1. The cells were harvested and stored at 70°C for later use when the cells reached passage 3. For preparation of bleomycin-injured lung, rats were instilled with 2 U of bleomycin in 0.5 ml volume of saline into the left lung, and 21 d later the left lung was removed for isolation of fibroblasts.

Preparation of Fibroblast-Type II Cell Cocultures

On experimental Day 0, adult rat alveolar type II cells were isolated and seeded on diluted Matrigel in a six-well plate as described earlier. On the following day, adult human lung fibroblasts were mixed with rat tail collagen (23) at a density of 3.3 × 10⁵ cells/ml of collagen and 0.3 ml of the fibroblast/collagen mixture was spread on a 25-mm polycarbonate filter (Corning, Cambridge, MA) and gelled at 37°C. The medium in the type II cell cultures was changed to remove nonadherent cells and debris, then the fibroblast rafts were transferred into the well. Medium was usually changed every 4 d after initiation of coculture unless otherwise stated.

Preparation of RNA and Ribonuclease Protection Assays

Type II cells on Matrigel were directly lysed into 4 M guanidinium isothiocyanate, 0.5% N-laurylsarcosine, and 0.1 M -mercaptoethanol in 25 mM sodium citrate buffer. Total cellular RNA was isolated by ultracentrifugation for 18 h at 150,000 × g through a 5.7 M CsCl cushion as previously described (24).

Ribonuclease Protection Assay

Total RNA isolated from type II cells was analyzed for surfactant protein (SP) messenger RNA (mRNA) expression using a ribonuclease protection assay (RPA) that allowed simultaneous measurement of mRNAs for SP-A, -B, -C, and -D. Fragments of different sizes for the four SPs were isolated by polymerase chain reaction using full-length rat complementary DNAs as templates. The forward primers included a Bam H1 restriction site added to the 5' end and the backward primers included an Eco R 1 restriction site added to the 5' end to facilitate directional cloning into pGEM 4Z (Promega). The primers for SP-A were 5'-CGGATCCAGTCCTCAGCTTGCAAGGATC-3' coding sense and corresponding to nucleotides 424-444, and 5'-GGAATTCCGTTCTCCTCAGGAGTCCTCG-3', coding antisense and corresponding to nucleotides 549-569. The probe transcribed from this clone identified a fragment of 146 base pairs (bp). The primers for SP-B were 5'-CGGATCCGAGCAGTTTGTGGAACAGCAC-3', coding sense and corresponding to nucleotides 997-1017; and 5'-GGAATTCTGGTCCTTTGGTACAGGTTGC-3', coding antisense and corresponding to nucleotides 1152-1172. The probe transcribed from this clone identified a fragment of 176 bp. The primers for SP-C were 5'-CGGATCCCATACTGAGATGGTCCTTGAG-3', coding sense and corresponding to nucleotides 202-222; and 5'-GGAATTCTCTGGAGCCATCTTCATGATG-3', coding antisense and corresponding to nucleotides 381-401. The probe transcribed from this clone identified a fragment of 200 bp. The primers for SP-D were 5'-CGGATCCCGGAAGAGCCTTTTGAGGATG-3', coding sense and corresponding to nucleotides 831-851; and 5'-GGAATTCACAGTTCTCTGCCCCTCCATTG-3', coding antisense and corresponding to nucleotides 1054-1075. The probe transcribed from this clone identified a fragment of 245 bp.

The vectors were linearized with Bam H 1 and radiolabeled antisense probes were transcribed in vitro using a commercially available kit (Promega) and [-³²P]cytidine triphosphate (CTP) (800 Ci/mmol; ICN, Costa Mesa, CA). RNA probes were purified on an 8% polyacrylamide/7 M urea gel. An 18S ribosomal RNA (rRNA) probe, which was synthesized using pT7 RNA 18S template and T7 MEGA shortscript Kit (both from Ambion, Austin, TX) and [-³²P]CTP, was used as an internal standard for RNA quantitation. The quantity of 3 µg of total cellular RNA was hybridized at 45°C for 24 h with 2.4 × 10⁴ counts per min of each antisense probe. The mRNA not protected by hybridization with target probe was digested with a mixture of RNase A (0.1 U) and T1 (4 U). The protected RNA duplex fragments were precipitated, resuspended, and separated on an 8% polyacrylamide/7 M urea gel. The gel was dried and exposed to Hyper film (Amersham Life Science Products, Arlington Heights, IL) at 70°C. Radioactive counts were obtained from protected fragments for SP-A, -B, -C, and -D by ImageQuant analysis (Molecular Dynamics, Sunnyvale, CA) and normalized to 18S rRNA. In the development of the RPA, we varied the input of total RNA from 1 to 15 µg, and demonstrated that the ImageQuant values were linear.

Cell Counts

Collagen gels were digested with 1% collagenase I (Worthington Biochemicals) in DMEM containing 1% FBS at 37°C for 1 h. The cell suspension was collected and centrifuged at 500 × g for 10 min. The cell pellet was treated with trypsin-ethylenediaminetetraacetic acid (EDTA) for 15 min to produce single cells, then the digestion was stopped by adding DMEM plus 1% FBS. Total fibroblast cell number was counted by using a hemacytometer. The average of the triplicate wells represented one data point.

DNA Assay

To harvest type II cells for the DNA assay, the Matrigel was enzymatically digested by incubation in dispase (Collaborative Biochemical, Bedford, MA) for 2 h at 37°C after the medium was removed. Once released from the matrix, the cells were pelleted by centrifugation at 500 × g for 10 min. The cells were washed once with phosphate-buffered saline (PBS) and re-pelleted at 500 × g. The pellets were suspended and stored in 1 ml sterile PBS containing 5 mM EDTA at 70°C. Before analysis, the samples were added to 1 ml buffer containing 10 mM NaH₂PO₄, 40 mM Na₂HPO₄, 4 M NaCl, and 2 mM EDTA and sonicated on ice. DNA content was determined fluorometrically using the procedure of Labarca and Paigen (25).

PGE₂ Assay

PGE₂ was quantified by enzyme immunoassays using acetyl cholinesterase-conjugated tracers as described previously (26, 27). The mouse monoclonal antibody (mAb) used for the quantification of PGE₂ was purchased from Cayman Chemical Co. (Ann Arbor, MI).

Immunostaining

The cultured cells were fixed with 96% ethanol, 3% water, and 1% glacial acetic acid and embedded in paraffin (22). Macrophages were identified with the mAb MAB-1435 (Chemicon International, Inc., Temecula, CA), macrophages, and fibroblasts with an antibody to vimentin (V9; Boehringer Mannheim, Indianapolis, IN). The antibodies were detected by immunofluoresence with a secondary biotinylated donkey antimouse immunoglobulin G, and Cy3-streptavidin (Jackson ImmunoResearch, West Grove, PA). Total cells were identified by including 4', 6-diamidino-2-phenylindole in the mounting media.

Data Analysis

All data are expressed as means ± standard error of the mean (SEM). Polynomial linear regression was used to evaluate the relationship of mRNA levels of SP-A, -B, -C, and -D on time. Analysis of variance was used for all other analyses. Fisher's least significant difference method was used for multiple comparisons. When the variance was not constant, as in Figure 1 for fibroblast number and Figures 4 and 6 and Table 2 for PGE₂ concentration, a natural log transformation was used for the statistical analyses. All tests were two-sided with an -level of 0.05. Data were analyzed using the SAS statistical package (SAS Institute, Cary, NC).

View larger version (13K): [in this window] [in a new window]   Figure 1.   Type II cells inhibit fibroblast proliferation. Rat type II cells were isolated and seeded on Matrigel in six-well plates on Day 0. On the following day, fibroblast rafts containing 1 × 105 cells were transferred into wells that contained type II cells (cocultures) (filled circles) or did not contain type II cells (control) (open circles). Fibroblast number was determined on alternate days beginning on Day 2. Results represent means ± SEM from five independent experiments. (*P  0.01, compared with control for the same day.)

View larger version (14K): [in this window] [in a new window]   Figure 4.   CM from cocultures of fibroblasts/type II cells between Days 9 and 13 contains increased amounts of PGE2. Medium was collected from fibroblasts cultured alone (Fb; open bars), type II cells cultured alone (TII; hatched bars), and cocultures of fibroblasts/type II cells (Fb/TII; filled bars) between Days 9 and 13 (CM9-13) in the presence or absence of 10% FBS. Both cocultures and type II cells cultured alone secreted more PGE2 in the absence of FBS. Values represent means ± SEM of four independent experiments. (*P < 0.001 compared with open bars or hatched bars in each group.)

View larger version (15K): [in this window] [in a new window]   Figure 6.   Fibroblasts are the major site of PGE2 production in fibroblast/type II cocultures. On Day 13, fibroblasts and cocultures of fibroblasts/type II cells were washed three times with DMEM. Fibroblast rafts from cocultures were separated from type II cells and transferred to dishes containing 2 ml of DMEM plus 10 µM arachidonate. Type II cells were given 2 ml of the same medium. After 45 min, the medium was harvested and PGE2 was measured. The following designations are used: TII(Fb/TII), type II cells from cocultures; Fb(Fb/TII), fibroblasts from cocultures; Fb(Fb), fibroblasts from fibroblasts cultured alone. Columns labeled 10% FBS: cells were cultured in the presence of 10% FBS before Day 13; no FBS means that during Days 9 to 13 the cultured medium was switched to serum-free DMEM. Results represent means ± SEM from four independent experiments. (*P  0.01 compared with open bars or hatched bars.)

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Fibroblast Proliferation Is Inhibited by Coculture with Type II Cells

These studies were designed to evaluate the effects of type II cells on proliferating fibroblasts as would occur in lung injury. Hence, the fibroblasts were grown within a collagen gel and in the presence of 10% serum. As shown in Figure 1, compared with control cultures, the number of human fibroblasts was reduced when they were cocultured with rat type II cells on Days 10, 12, and 14. Because the total number of cells in the cocultures was greater than with fibroblasts alone, the possibility that these observations were due to nutritional depletion was addressed. One method to assess this possibility was by varying the frequency of the medium changes. Three groups in which the medium was changed at different time intervals were compared, and there was no effect of changing the medium every 2 d or half of the medium every day compared with media changes every 4 d (P = 0.93). These results indicate that the decreased proliferation of fibroblasts in the coculture system was not due to the medium being depleted of key nutritional components by the metabolic load present in the cocultures.

To further investigate the issue of nutritional depletion, we seeded the same cell number of fibroblasts as type II cells present on Day 9 in cocultures, then placed fibroblast rafts in the medium above the cells. As a control, fibroblast rafts were placed in medium over Matrigel on which no cells were seeded. After 4 d of culture, the fibroblast number in the rafts cocultured with fibroblasts on Matrigel increased from 1.5 × 10⁶ to 2.3 × 10⁶ cells, fibroblast number in control cultures increased from 1.5 × 10⁶ to 2.4 × 10⁶ cells. Fibroblasts cultured on Matrigel increased from 8.5 to 17.4 µg DNA per well, a value similar to that seen in type II cells after 13 d in coculture. Thus the culture medium was able to support proliferation of a pure population of fibroblasts under conditions similar to that of type II cell-fibroblast cocultures.

Type II Cell Differentiation Is Maintained in Cocultures

We determined the mRNA levels of SP-A, -B, -C, and -D from type II cells cocultured with fibroblasts as an indication of type II cell differentiation. Figure 2A shows the time course of mRNA levels of these four proteins. Linear regression analysis indicated that there were no changes for SP-A and SP-D over time, the value of SP-B peaked at approximately Day 8 and then decreased, and the amount of SP-C decreased modestly with time at a rate of 4.3% per day. However, the value of SP-C mRNA on Day 12 remained approximately 50% of that of freshly isolated type II cells. These data indicate that type II cells in this coculture system remain differentiated for an extended period. Figure 2B shows a representative picture for 18S rRNA and SP-A, -B, -C, and -D by RPA.

View larger version (74K): [in this window] [in a new window]   View larger version (19K): [in this window] [in a new window]   Figure 2.   Time course of expression of SP-A, -B, -C, and -D mRNA in type II cells cocultured with fibroblasts. In A, the data were normalized to 18S rRNA content and then expressed as percentages of those from freshly isolated type II cells (Day 0). Results represent means ± SEM from five independent experiments. (B) A representative RPA. Lane 1 is mRNA from freshly isolated type II cells; lanes 2-8 are mRNA from type II cells in coculture on experimental Day 2 to Day 14.

Conditioned Medium from Cocultures Inhibits Fibroblast Proliferation

Conditioned medium (CM) from type II cells cocultured with fibroblasts or type II cells cultured alone on Matrigel were collected and diluted 1:1 with fresh DMEM plus 10% FBS. The 50% CM was tested on fresh cultures of fibroblasts that had been grown for 6 d. Figure 3 shows that 50% CM from Days 9 to 13 of the coculture system had an inhibitory effect on fibroblast proliferation. In contrast, CM from cocultures on Days 1 through 5 and 5 through 9 had no effect. CM from type II cells cultured alone on Matrigel on Days 1 to 5, 5 to 9, or 9 to 13 had no effect on fibroblast proliferation. These results suggest that the inhibitory factor either was not present or did not reach an effective concentration until Day 9 of coculture under these conditions. Further, production of detectable concentrations of the inhibitory factor required the presence of both type II cells and fibroblasts.

View larger version (32K): [in this window] [in a new window]   Figure 3.   CM from fibroblast/type II cell cocultures inhibits fibroblast proliferation. CM from fibroblast/type II cell cocultures (filled bars) or type II cells cultured alone (hatched bars) in the presence of 10% FBS were collected on Day 5 (CM1-5), Day 9 (CM5-9), and Day 13 (CM9-13). The CM was diluted 1:1 with freshly DMEM plus 10% FBS and tested on separate fibroblasts that had been grown for 6 d. Fibroblast number in cultures treated with CM was determined on Day 11. Results represent means ± SEM from five to seven independent experiments. (*P < 0.01, compared with control.)

As a first step to characterize the factor(s) present in CM from coculture, we examined the ability of CM from serum-free cultures to inhibit fibroblast proliferation. Serum-free CM was collected from Days 9 to 13 and diluted 1:1 with DMEM containing 20% FBS. The 50% CM (containing a final concentration of 10% FBS) was tested on fresh fibroblast cultures beginning on Day 6. Fibroblast proliferation was inhibited to a similar extent with coculture, CMs containing serum, and serum-free CMs (data not shown). Therefore, further characterization of the inhibitory factor(s) was performed in the absence of serum. Treatment of serum-free CM by either boiling (10 min) or trypsin (0.1 mg/ml, 37°C, 2 h) did not affect the inhibitory effect (data not shown), which suggests that the inhibitory factor is not a peptide.

PGE₂ in CM Inhibits Fibroblast Proliferation

Previous studies have shown that alveolar type II cells in primary culture produce PGE₂ (1). Therefore, the concentrations of PGE₂ in the CM from type II cells cultured alone, or fibroblasts cultured alone, or cocultured type II cells with fibroblasts were assayed. The results (Figure 4) show that CM from Day 9 to 13 cocultures had higher levels of PGE₂ than did CM from either type II cells cultured alone or fibroblasts cultured alone. Moreover, when cocultures were switched to serum-free medium on Day 9, PGE₂ was much higher than in cocultures maintained in DMEM plus 10% FBS. These results suggest that PGE₂ was involved in the inhibition of fibroblast growth.

We then determined the effect of exogenous PGE₂ on fibroblasts cultured in these collagen gels. Varied concentrations of PGE₂ were added to fibroblasts on Day 8, and the fibroblast number and PGE₂ level were determined on Day 11 (Table 1). Inhibition of fibroblast proliferation was dose-dependent, with a minimal effective concentration falling between 120 and 340 pg/ml in different experiments.

Because of the dramatic effect of PGE₂ on fibroblast proliferation, we next determined the effect of inhibiting PG synthesis in cocultures on fibroblast proliferation. Starting on Day 9, different concentrations of indomethacin were added to the coculture system. Fibroblast number was determined on Day 12 and the CM was harvested for PGE₂ assay. The results are shown in Figure 5 and Table 2. There were no statistical differences for fibroblast number between fibroblasts cultured alone and fibroblasts cocultured with type II cells in the presence of 1 or 10 µM indomethacin (P = 0.36 and 0.33, respectively), whereas there were statistical differences in the absence of or presence of 0.1 µM indomethacin (P < 0.01). Each of the indomethacin groups (0.1, 1, and 10 µM) had a significantly lower concentration of PGE₂ than did the control group (all P < 0.0001). These data strongly support the conclusion that the inhibition of fibroblast proliferation is mediated by PGE₂.

View larger version (19K): [in this window] [in a new window]   Figure 5.   Indomethacin blocks the inhibitory effect of type II cells on fibroblast proliferation. Indomethacin was added to fibroblast-type II cocultures on Day 9, and fibroblast number was determined on Day 12. Open bars, fibroblasts cultured alone; filled bars, fibroblast number cocultured with type II cells. Control: untreated cultures. Ethanol: vehicle control containing 0.1% (vol/vol) ethanol. Values represent means ± SEM of four independent experiments. (*P < 0.01 compared with control in same group.)

Most of the PGE₂ Is Produced by Fibroblasts

To determine which cells in the coculture system produced PGE₂, the fibroblasts and type II cells were cultured separately with arachidonate. On Day 13 of coculture, the alveolar type II cells and fibroblasts were washed three times with DMEM and the fibroblast rafts were removed and placed in separate wells. The separated type II cells and fibroblasts were each incubated in 2 ml of fresh DMEM plus 10 µM arachidonate for 45 min. The CMs were harvested and assayed for PGE₂. Figure 6 shows that the fibroblasts produced significantly more PGE₂ than did alveolar type II cells. However, fibroblasts that were cultured alone produced less PGE₂ than did fibroblasts from cocultures, indicating that alveolar type II cells produced factor(s) that stimulate PGE₂ production by fibroblasts.

Rat Type II Cells also Inhibit Rat Lung Fibroblasts

All of the experiments described earlier were cocultures of human lung fibroblasts and rat alveolar type II cells. Figure 7 shows the results of cocultures of normal rat lung fibroblasts and rat type II cells. We found that normal rat lung fibroblasts were inhibited, and PGE₂ levels were increased. We also found that rat lung fibroblasts isolated from rats that had been instilled with bleomycin 21 d earlier were inhibited when cocultured with rat type II cells (data not shown). Hence, normal human lung fibroblasts, normal rat lung fibroblasts, and fibroblasts isolated from rat lungs instilled with bleomycin were all inhibited in coculture with rat type II cells.

View larger version (14K): [in this window] [in a new window]   Figure 7.   Rat type II cells inhibit rat fibroblasts. Rat lung fibroblasts cocultured with rat alveolar type II cells. Rat type II cells were isolated and seeded on Matrigel in six-well plates on Day 0. On the following day, the rafts containing 1 × 105 normal rat lung fibroblasts were transferred into wells that contained type II cells (filled bars) or did not contain type II cells (open bars). The fibroblast number (7-A) and PGE2 concentration in the media (7-B) were determined on Day 12. Results represent means ± SEM for fibroblasts isolated from four different rats. (*P < 0.01).

Fibroblast Inhibition Is Due to Type II Cells

During the last half of the culture period there was an increase in the number of cells on the Matrigel in the cocultures. The DNA/well increased from 6.27 ± 1.85 µg on Day 8 to 13.96 ± 1.11 µg on Day 12 (n = 4). Because the media contained 7.34 ± 1.69 ng/ml HGF and 0.35 ± 0.08 ng/ml KGF (n = 4), we had assumed that this was predominantly due to type II cell proliferation. However, to access purity of these cultures we used immunocytochemistry to identify any contaminating cell types. MAB-1435 was used for macrophages and dendritic cells, and vimentin for macrophages and fibroblasts. The results were similar. There were only 3.5 ± 0.7 vimentin-positive cells (n = 5) on Day 2 of the experiment, which is Day 1 for fibroblasts because type II cells were started one day previously. However, there were 20.5 ± 5.2% MAB-1435-positive cells and 30.3 ± 3.9% vimentin-positive cells on Day 12 of cultures (n = 7). Rat alveolar macrophages were therefore tested to see whether they could inhibit fibroblast proliferation in these cocultures. Rat alveolar macrophages (1 × 10⁶/well) were plated on Matrigel, and they were unable to inhibit fibroblast proliferation in coculture experiments or increase PGE₂ levels above 50 pg/ml. Finally, KGF was used to stimulate type II cell proliferation so that inhibition of fibroblast proliferation would occur earlier in culture and minimize the macrophage contamination (Figure 8). Type II cells stimulated with KGF inhibited fibroblast proliferation on Day 6 (data not shown) and Day 8, which was much earlier in culture than in previous experiments without KGF. In parallel experiments, 10 ng/ml KGF did not alter fibroblast proliferation. The contamination by macrophages and fibroblasts in the KGF experiments was markedly reduced. In cultures with rat serum and KGF the number of vimentin-positive cells (macrophages and fibroblasts) on Day 8 was 4.8 ± 1.1% (n = 4) and the number of MAB-1435-positive cells was 1.4 ± 0.1% (n = 4). There were less than 5% contaminating cells in cultures of type II cells alone or in the coculture with fibroblasts. The epithelial cells were also identified with antibodies to cytokeratin, SP-A, and their characteristic growth pattern as spherules after a few days of culture on Matrigel.

View larger version (34K): [in this window] [in a new window]   Figure 8.   KGF stimulates type II cells to inhibit fibroblast proliferation earlier in culture. Human lung fibroblasts were cocultured with rat alveolar type II cells with DMEM containing 5% rat serum. The procedure is the same as described earlier except that the seeding number of fibroblasts was 5 × 105 cells/well, KGF (10 ng/ml) was added to some cultures, and the medium was changed every 2 d. The cultures were harvested on Day 8. Open bar: fibroblasts cultured alone; hatched bar: fibroblasts plus KGF; gray bar: cocultures of fibroblasts and type II cells; filled bar: cocultures of fibroblasts and type II cells plus KGF. Values represent means ± SEM of four to eight independent experiments. (*P < 0.01 compared with other three groups.)

In This System Type II Cells Do Not Stimulate Fibroblast Proliferation

The initial studies were designed to determine whether type II cells could inhibit fibroblast proliferation in a coculture system. Therefore, the study was designed to have proliferating fibroblasts, and the fibroblasts were cultured in the presence of 10% FBS. However, we also wanted to see whether type II cells could stimulate fibroblasts if cultured in the absence of serum or in the presence of 0.4% serum. The culture system used to test whether type II cells stimulate fibroblast growth was the same as the original system except that the medium contained little or no serum. Because the type II cells were plated on Matrigel in the presence of 10% FBS, there would always be some serum components available in the cultures. As shown in Figure 9, there was no increase in fibroblast number whether the cultures were conducted in 0.4% serum or in the absence of serum.

View larger version (20K): [in this window] [in a new window]   Figure 9.   Type II cells did not stimulate fibroblast proliferation in the absence of serum or in the presence of 0.4% FBS. The experimental conditions were exactly the same as in Figure 1 except that the concentration of serum in the culture medium was varied. Approximately 100,000 fibroblasts were contained within the collagen gels on Day 0. On Day 12, the fibroblast number was determined in each group for fibroblast culture only (hatched bars) and fibroblast-type II cell cocultures (filled bars). Values represent means ± SEM of three independent experiments.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

To evaluate the effect of alveolar type II cells on lung fibroblasts in vitro, we sought a culture system that would simulate conditions in vivo and allow for analysis of mediators and growth factors. To accomplish this goal, we had several criteria that had to be met. First, type II cell differentiation should be maintained. Second, the basolateral surfaces of type II cells should face fibroblasts, which would mimic their spatial relationship in vivo. Third, fibroblasts should be cultured within a collagen gel, which maintains their interaction with ECM. Fourth, there should be a physical separation of type II cells and fibroblasts so that CM could be sampled and type II cells and fibroblasts could be analyzed separately.

In vivo fibroblasts reside within a three-dimensional ECM. In this coculture system, human lung fibroblasts were grown within a rat tail collagen gel as opposed to a monolayer on plastic dishes. This three-dimensional culture provides a microenvironment more closely resembling in vivo conditions than does the conventional plastic dish culture. We found that FBS is necessary for fibroblast proliferation in the collagen gel, and that 10% FBS is more effective than 5% FBS (data not shown). This result is consistent with the results of Mio and colleagues (28) that a higher concentration of serum is required when fibroblasts are cultured within a collagen gel.

When we observed that the number of fibroblasts was decreased in coculture, the first issue that we addressed was whether the results were due to nutritional depletion of the medium. We demonstrated that the effect was not due to nutritional depletion in four ways. First, changing medium more frequently did not affect the results. If nutrition had not been adequate in medium cocultured for 4 d, the fibroblast numbers in cocultures would have increased when medium was changed every day or every other day. Second, when fibroblasts were cocultured with fibroblasts to provide the same number of total cells as in the fibroblast-type II cell cocultures, no inhibition of fibroblast proliferation was observed. Third, the observation that indomethacin reversed the inhibition of fibroblast number proliferation provided direct evidence that nutrition was sufficient. If nutrition had been a problem, indomethacin would have not increased the numbers of fibroblasts. Fourth, type II cells and macrophages on the Matrigel proliferated in the coculture system whereas fibroblast proliferation was inhibited. Clearly, the nutrition in the medium was sufficient for proliferation of cells on the Matrigel.

The CMs from cocultures showed an inhibitory effect on fibroblast proliferation, whereas CMs from type II cells cultured alone had no effect. This suggests that interactions between the two cell types are critical for demonstrating this inhibiting factor. Our results support and expand observations made by others. Adamson and coworkers (29) cocultured rat alveolar epithelial cells in a well above rat lung fibroblasts which were cultured on plastic. They found that fibroblasts grown beneath epithelial cells showed a lower percentage of [³H]thymidine-labeled cells. However, fibroblast number and DNA content in cells cultured beneath epithelial cells were not statistically different from fibroblasts cultured alone. Klein and associates (17) reported that type II cells exposed to silica produce PGE₂, which in turn inhibited thymidine incorporation by fibroblasts. In these experiments, the type II cells were cultured on plastic and there was limited inhibition in the absence of silica (17). In our experiments we used cell number to measure fibroblast proliferation, and fibroblast proliferation was inhibited by almost 50% by coculture with alveolar type II cells. Our studies differ from others by maintaining type II cell differentiation, as evidenced by maintenance of the mRNAs for the SPs, and the longer observation time in primary culture. Without KGF the inhibition of fibroblast proliferation was not apparent until about 10 d. The specific reason for this delay is not known and cannot be determined until the factor(s) produced by type II cells are identified and quantitated. We believe that a threshold concentration of this putative factor(s) is needed and that this requires time. The time delay permits the mass of type II cells to increase, the gene products of type II cells to be expressed in culture, and the potential binding sites in the Matrigel to become saturated so that the factor accumulates in the medium. In the 12-d cultures, this factor(s) could require a complex interaction between fibroblasts, macrophages, and type II cells. Numerous types of cell- cell interactions could be envisioned. However, CMs were inhibitory only if both type II cells and fibroblasts were present. In addition, macrophages alone did not inhibit fibroblast proliferation in these cultures. To simplify this system, KGF was used to stimulate the proliferation of the type II cells (27, 30) and minimize the effect of any contaminating macrophages. In these later experiments, fibroblast proliferation was inhibited, there was an increase in PGE₂, and macrophage contamination was less than 5%. These experiments indicate that the factor comes from type II cells and not macrophages.

The addition of indomethacin indicated that a cyclooxygenase (COX) product was involved in the growth inhibition of fibroblasts. Chauncey and colleagues reported that rat type II cells produce COX products and that the major product is PGE₂ (4). Cott and associates (1) reported that type II cells in primary culture on plastic tissue culture produced about 500 pg/ml PGE₂ during 3 h culture in the absence of serum. In our study, rat alveolar type II cells cultured alone on Matrigel produced PGE₂. In the absence of serum, PGE₂ concentration was approximately 4,000 pg/ml (Figure 4) and in the presence of 10% FBS approximately 600 pg/ml. The explanation of why type II cells cultured alone in the absence of serum released much more PGE₂ than in the presence of 10% FBS is not resolved. Some have suggested that serum-free conditions maintain type II differentiation better than in 10% FBS (29, 31). However, Lipchik and coworkers (3) reported that PGE₂ synthesis by epithelial cells cultured on plastic was increased when type II cells become dedifferentiated and more similar to type I cells, and suggested the importance of type I cells as a source of PGE₂. We chose to study type II cells on Matrigel to maintain their differentiation and to simulate the cuboidal hyperplastic type II cells seen in pulmonary fibrosis.

Exogenous PGE₂ has been shown to inhibit proliferation of human lung fibroblasts (28, 32). Mio and colleagues reported that the inhibitory effect of PGE₂ on fibroblasts in collagen gels was similar to that seen in monolayer cultures in plastic dishes, although fibroblasts in the gels were less responsive to the growth stimulators, such as serum, insulin, and PDGF (28). However, the amount of PGE₂ required for inhibition of proliferation is variable (28, 32). The results from our indomethacin experiments showed that indomethacin reversed inhibition of fibroblast proliferation at concentrations of 1 and 10 µM and reduced PGE₂ concentration in the medium to less than 200 pg/ml. When indomethacin was used at a concentration of 0.1 µM, which did not reverse inhibition of fibroblast growth, the PGE₂ concentration was 311 pg/ml. Conditions with a measured media concentration of PGE₂ above 350 pg/ml consistently inhibited fibroblast proliferation.

The increase in PGE₂ in the medium effect is dependent on the cocultures of type II cells and fibroblasts. In our experiments, alveolar type II or lung fibroblasts cultured alone produced very little PGE₂, and the concentration of PGE₂ in CM from cocultures was much greater. The majority of PGE₂ in cocultures was produced by lung fibroblasts and not by the type II cells. However, because fibroblasts cultured by themselves did not produce much PGE₂, we concluded that alveolar type II cells release a soluble factor(s) that stimulate PGE₂ production by fibroblasts. PGE₂ then acts in an autocrine fashion to inhibit fibroblast proliferation. The identity of this soluble factor(s) produced by type II cells remains to be defined. Type II cells have the capacity to secrete TGF- (9, 10), IL-6 (5), GM-CSF (8) and a variety of other poorly characterized factors (29, 33, 34). Human lung fibroblasts have the ability to produce PGE₂ that can be stimulated by a variety of factors (32, 35). Sato and associates (37) reported that in a coculture system of dermal fibroblasts with epidermal keratinocytes, the production of PGE₂ was much higher, whereas PGE₂ was negligible in individual monolayer cultures. The CM from keratinocytes greatly enhanced the production of PGE₂ by fibroblasts, and IL-1 was the factor secreted by keratinocytes thought to be responsible for PGE₂ production by fibroblasts. McAnulty and coworkers (35) reported that when TGF-₁ (1 ng/ml) was added to fibroblasts, the PGE₂ production increased about 5-fold. In addition to IL-1 and TGF-₁, many other factors can stimulate fibroblasts to release PGE₂; these include lipopolysaccharide, phorbol myristate acetate, and IL-1 (35). We expect that alveolar type II cells secrete one or more of these soluble factors, including TGF-₁, IL-1, or IL-1. Recently, Wilborn and colleagues (36) reported that lung fibroblasts isolated from patients with IPF have a diminished capacity to synthesize PGE₂. Although our initial studies were done with human lung fibroblasts, in subsequent experiments fibroblast inhibition was also seen with rat fibroblasts and rat fibroblasts isolated from bleomycin-treated lung. Hence, the effect crosses species and is also likely to be important for fibroblasts in fibrotic lung.

The original design of the present study included 10% serum to stimulate fibroblast proliferation so that inhibition of growth could be evaluated, but the same culture system was used in the absence of serum to determine whether type II cells can stimulate fibroblast proliferation. In our coculture system we were not able to demonstrate that type II cells stimulate proliferation of fibroblasts in the absence of FBS or in the presence of 0.4% FBS. We interpret these experiments as supporting the hypothesis that type II cells limit or suppress the fibrotic reaction and are not the cause of fibroblast proliferation. However, it is also possible that under our culture conditions the type II cells did not produce the profibrotic growth factors that they may in vivo, that fibroblasts lacked the ability to respond to these factors, or that the matrix bound these factors so that they were not available to the fibroblasts. In addition, there was significant distance between the epithelial cells and the fibroblasts, and the putative profibrogenic factors may be labile or diluted by the media so as to be ineffective. Nevertheless, under the conditions of these experiments in vitro, type II cells did not stimulate fibroblast proliferation under serum-free or low-serum conditions.

In summary, the present study demonstrates that type II cells secrete factor(s) that inhibit fibroblast proliferation and that the inhibition is accomplished by at least two steps: factor(s) secreted by type II cells stimulate PGE₂ production by fibroblasts, and then PGE₂ directly inhibits fibroblast proliferation. Our results support the hypothesis that, on balance, the alveolar epithelium limits the fibrotic response and is antifibrogenic and not profibrogenic. These findings may have clinical significance for pulmonary fibrosis.