Transforming growth factor-beta stimulation of lung fibroblast prostaglandin E 2 production

Transforming growth factor-beta (TGF beta) stimulated the production of total protein, collagen, and fibronectin by normal human lung fibroblasts. The stimulatory response was maximal at 100 pM TGF beta and reversed toward control at higher concentrations. Inhibition of fibroblast prostaglandin (PG) synthesis enhanced TGF beta-induced stimulation of total protein, collagen, and fibronectin production and reversed the negative slope of the dose-response curve at high concentrations of TGF beta. Determination of the steady-state levels of Types I and III procollagens and fibronectin mRNAs employing specific cDNA probes demonstrated that inhibition of fibroblast PG production increased the stimulatory effect of TGF beta on the levels of these transcripts. Exogenous PGE2 abrogated the stimulatory effects of TGF beta. These findings suggest that fibroblast stimulation by TGF beta may be down-regulated by endogenous PG synthesized in response to TGF beta. This notion was supported by the demonstration that TGF beta markedly stimulated fibroblast PGE2 production. These results indicate that TGF beta-induced stimulation of fibroblast PGE2 production may be an autoregulatory control mechanism to limit the effects of TGF beta on connective tissue protein synthesis.

Transforming growth factor-@ (TGFB) stimulated the production of total protein, collagen, and fibronectin by normal human lung fibroblasts. The stimulatory response was maximal at 100 p~ TGFP and reversed toward control at higher concentrations. Inhibition of fibroblast prostaglandin (PG) synthesis enhanced TGFB-induced stimulation of total protein, collagen, and fibronectin production and reversed the negative slope of the dose-response curve at high concentrations of TGFB. Determination of the steady-state levels of Types I and I11 procollagens and fibronectin mRNAs employing specific cDNA probes demonstrated that inhibition of fibroblast PG production increased the stimulatory effect of TGFB on the levels of these transcripts. Exogenous PGE, abrogated the stimulatory effects of TGFB. These findings suggest that fibroblast stimulation by TGFB may be down-regulated by endogenous PG synthesized in response to TGFB. This notion was supported by the demonstration that TGFB markedly stimulated fibroblast PGE, production. These results indicate that TGFB-induced stimulation of fibroblast PGE, production may be an autoregulatory control mechanism to limit the effects of TGFB on connective tissue protein synthesis.
The extracellular matrix of connective tissue consists of complex and highly organized macromolecules, including collagen and fibronectin. Synthesis of connective tissue components is necessary for maintenance of the structural and functional integrity of most parenchymal organs as well as for a variety of dynamic events such as wound healing, repair, and development. The net accumulation of extracellular matrix is dependent on a precise balance between the synthesis and the degradation of connective tissue components. Excessive deposition of collagen and fibronectin is characteristic of pathological states of fibrosis, including scleroderma and pulmonary fibrosis (1,2). Fibroblast connective tissue production must, therefore, be self-limited in order to prevent abnormal fibrogenesis. Thus, these cells must be responsive to both stimulatory and inhibitory influences. The frequent observa-* This work was supported in part by grants from the Scleroderma Research Foundation and the Scleroderma Society, Inc. and by National Institutes of Health Grants AM19616 and HL31650. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18  tion of a proximity of fibroblasts and chronic inflammatory cells in areas undergoing active fibrogenesis has suggested that cytokines may play an important role in modulation of fibroblast functions (3, 4).
Transforming growth factor$ (TGFP),' a polypeptide produced by neoplastic and normal cells including macrophages and lymphocytes (5), has been shown to stimulate fibroblast protein, collagen, and fibronectin production in vitro (6)(7)(8) and to accelerate wound healing and angiogenesis in experimental animals in vivo (9,10). Fibroblasts cultured in the presence of TGFP display an increase in the steady-state levels of fibronectin and Types I and I11 procollagen mRNAs (11-13), which remain elevated even after removal of TGFP from the culture medium (12). It therefore appears that TGFP can stimulate and perpetuate augmented connective tissue biosynthesis and may, thus, play a major role in the development of fibrosis. On the other hand, fibroblast collagen and fibronectin production is inhibited by prostaglandins (PG) (14,15), suggesting that endogenous PG production by fibroblasts may participate in the autocrine modulation of TGFP effects. To examine experimentally this hypothesis, we investigated the effect of TGFP on lung fibroblast PGE, production and the influence of endogenous and exogenous PGE, on TGFP-stimulated total protein, collagen, and fibronectin synthesis by these cells. We found that TGFP markedly stimulated the production of PGE, by the fibroblasts and that the stimulatory effects of TGFP on connective tissue production were further enhanced by inhibition of endogenous PGE, synthesis. Measurement of the steady-state levels of Types I and I11 procollagen and fibronectin mRNAs showed that inhibition of fibroblast prostaglandin production resulted in augmentation of the stimulatory effect of TGFP on the levels of these transcripts. Endogenous and exogenous PGE, abrogated the stimulatory effects of TGFP on collagen and fibronectin production. Our results indicate that under normal circumstances, endogenous PGE, may play an important role in limiting or terminating the stimulation of fibroblast connective tissue synthesis induced by TGFP.

Fibroblast Culture and Labeling
Conditiom-Human lung fibroblast cell lines were established using explant techniques from histologically normal areas of lungs surgically resected for diagnostic reasons. Early passage cells were seeded at a density of 5 X lo4 cells/ well in 24-well flat-bottom plates and cultured in Eagle's minimal essential medium supplemented with 10% fetal calf serum and 1% vitamins, and incubated at 37 "C in 5% CO,. After the cultures reached confluency, the media were removed and fresh media containing 1% fetal calf serum, ascorbic acid (50 pg/ml), and various concentrations of PGE, (Sigma) and human TGFP (Collaborative Research Inc., Bedford, MA) were added to the wells. Appropriate cultures were preincubated for 30 min with indomethacin (1 pg/ml). NaCl, 50 mM Tris-HC1, and protease inhibitors, and the cells were detached mechanically and then sonicated in the same buffer. Control cultures were handled exactly as described above, except that the culture media did not contain TGFB or PGE,. All experiments were performed in triplicate.
Analysis of Lubeled Proteins and PGE, Assay-Aliquots of media and cell layers were exhaustively dialyzed to remove unincorporated radioactive precursors. Total incorporation of [U-'4C]proline into macromolecules was measured in a scintillation spectrophotometer. The amount of radioactive collagen synthesized by the fibroblasts was determined by a bacterial collagenase assay (16) and the relative proportion of collagen calculated following the formula of Breul et al. (17). The labeled proteins in the media and cell layers were analyzed by electrophoresis on SDS-polyacrylamide slab gels (7%) under reducing conditions. After electrophoresis, the gels were processed for fluorography, and the relative proportion of radioactivity in each band was determined by densitometric scanning (540 nm). In addition, newly synthesized fibronectin was quantified by immunoprecipitation with a specific anti-human fibronectin antibody (Bethesda Research Laboratories) as described previously (18). PGE, production by control and treated fibroblast cultures was determined by a specific ELISA assay of the undialyzed media (19) of confluent fibroblast cultures incubated with TGFp for 24 h as described.
Isolation and Analysis of Total RNA from the Lung Fibroblasts-Lung fibroblasts were cultured in T-75 flasks in the presence or absence of 500 p~ TGFP and indomethacin as described above. At the end of the incubation period the cell layers were washed in Hanks' solution and harvested in 4 M guanidinium isothiocyanate, and total RNA was isolated in a CsCl, discontinuous gradient as described by Maniatis et al. (20). For Northern blot analysis, aliquots of 10 pg of total RNA were denatured in formaldehyde, electrophoresed in 0.8% agarose gels, and then transferred to nitrocellulose filters. Human cDNA clones Hf677 specific for d ( 1 ) procollagen (21), S318 specific for al(II1) procollagen (a gift from Dr. Leena Ala-Kokko, Jefferson Institute of Molecular Medicine, Philadelphia, PA), and pFHl specific for fibronectin (22) were nick-translated with [cY-~'P]~ATP to a specific activity of 2.1 to 7.2 X lo-' dpm/pg as described by Rigby et al. (23). The filters were hybridized for 24 h in a 50% formamide solution. Autoradiographs were scanned with a densitometer. For quantitative estimation of specific mRNA levels hybridized to each cDNA clone the integrated areas were corrected for total RNA and of RNA hybridized/pg of RNA electrophoresed X total RNA/total DNA using the following equation: specific mRNA = integrated area DNA as described (24). The amount of DNA was quantified by a fluorimetric method (25).

Effect of Inhibition of Fibroblast PGE, Synthesis on the
Stimulation of Collagen and Fibronectin Production by TGFP-Treatment of lung fibroblasts with TGFp up to a concentration of 250 pM resulted in stimulation of total protein, collagen and fibronectin production. At higher TGFp concentrations, however, the amount of newly synthesized macromolecules declined toward control values (Table I, Fig.  1, A and B ) . To investigate the mechanisms responsible for this decline, we examined the effects of inhibition of endogenous lung fibroblast PG production by indomethacin. The rationale to perform these experiments was the well known inhibitory effect of PGE, on fibroblast connective tissue biosynthesis (14,15). The addition of TGFp to cultures pretreated with indomethacin resulted in progressive and quantitatively higher stimulation of fibroblast total protein, collagen, and fibronectin production by TGFB concentrations of up to 500 PM (Table I, Fig. 1, A and B ) . These effects on fibronectin production were confirmed by immunoprecipitation (data not shown).
To examine if this finding was reproducible, we evaluated the response of four separate lung fibroblast cell lines to TGFP (500 PM) in the presence and absence of indomethacin (data not shown). TGFB in the absence of indomethacin caused stimulation of the production of total protein, collagen, and fibronectin by three cell lines whereas a fourth cell line showed a moderate decrease in collagen but an increase in fibronectin production. Pretreatment of cultures with indomethacin enhanced TGFP effects in the three stimulated cell lines and reversed the decrease in the fourth with a net stimulatory effect.
Effect of TGFP on PGE, Production by the Fibroblasts- The finding of a negative slope on the dose-response curve of lung fibroblasts stimulated with high concentrations of TGFP in the absence of indomethacin suggested that TGFp at these concentrations might induce an endogenous inhibitor of collagen and fibronectin production by these cells. The restoration of a stimulatory response to TGFP by preincubating the cultures with indomethacin indicated that this inducible endogenous inhibitor might be a cyclooxygenase product. To confirm this possibility directly, the amount of PGE, in the media of control and TGFP-treated cultures was measured by an ELISA. The addition of TGFp to the cultures caused a dose-dependent stimulation in the synthesis of PGE,. Although a 4-fold increase in the production of PGE, was noted at low doses of TGFP, dramatic stimulation was apparent when TGFP concentration exceeded 100 PM (Fig. IC). In a separate experiment, we tested if the TGFP effect on PGE, production was saturable and compared PGE, production by cultures exposed to 500 and 1000 PM TGFP. We found that TGFP at a concentration of 1000 pM did not cause further stimulation In these experiments endogenous fibroblast PG production was blocked by preincubation with indomethacin. As shown in Table 11, the addition of increasing concentrations of PGE, to the cultures exposed to TGFp plus indomethacin caused a reversal of TGFp effects. At a concentration of 100 ng/ml of PGE,, the TGFP-induced stimulation of collagen and fibronectin production was almost completely abrogated. The amount of PGE, added to the cultures (10-100 ng/ml) was in the range of that produced by the TGFPtreated fibroblasts. Effects of TGFP on Procollagen and Fibronectin mRNA Levels in the Presence and Absence of Indomethacin-In order to investigate the mechanism of augmentation of TGFP effects on fibroblast collagen and fibronectin synthesis by cyclooxygenase inhibition, the steady-state mRNA levels for al (1) and al(II1) procollagens and fibronectin were examined by Northern blot hybridization using specific cDNA probes for these transcripts (Fig. 2). The autoradiographs of the Northern blots were quantified by densitometric analysis (Table 111). The results indicate that TGFp caused a 3-fold increase in the amount of Type I and Type I11 procollagen mRNAs and a 5-fold increase in fibronectin mRNA. When endogenous PG synthesis was inhibited by the addition of indomethacin to the cultures, TGFp caused a further increase in the levels of these mRNAs.

DISCUSSION
In these experiments we found that TGFP induced a reproducible stimulation of the production of total protein, collagen, and fibronectin by several cultured human lung fibroblast cell lines. Blockade of endogenous PG synthesis with indomethacin potentiated the stimulatory effect of TGFp on the production of these proteins. These findings suggested that PG may function to down-modulate the effects of TGFp on

TABLE I Effects of indomethacin on TGFP-induced stimulation of total protein, collagen, and fibronectin synthesis by lung fibroblasts
Confluent human lung fibroblasts were incubated for 24 h with control medium or with media containing different concentrations of TGFp with or without indomethacin (1 pg/ml) and labeled with [U-"Clproline. The amount of newly synthesized total protein, collagen, and fibronectin in the media and cell layers was determined as described under "Materials and Methods." Values represent the average from triplicate experiments which varied less than 10% from each other. The numbers in parentheses show the percentage relative to control values. collagen and fibronectin production. When higher concentrations of TGFP (250 and 500 PM) were used, the dose-response curve of the production of these proteins displayed a downward slope. Furthermore, the potentiation of this TGFP effect by indomethacin was much more pronounced at high TGFP concentrations. These findings suggested that the higher concentrations of TGFP resulted in the production of an endog-   (lane 4 ) . Total RNA was extracted from each culture as described under "Materials and Methods," and after denaturation, samples containing 10 pg of RNA were electrophoresed in each lane of 0.8% agarose gels and transferred to nitrocellulose filters. The filters were hybridized to radiolabeled human cDNA probes specific for pro-al(1) collagen (Hf677), pro-al(II1) collagen (S318), and fibronectin (pFH1). Fig. 2 were scanned with a densitometer and the areas quantified employing a planimeter. The values of integrated areas were corrected for the total yield of RNA and DNA according to the equation: specific mRNA = integrated area of RNA hybridized/pg of RNA electrophoresed X total RNA/ total DNA and are expressed as a -fold increase relative to control. enous inhibitory factor which reversed the stimulatory effect of TGFP on collagen and fibronectin production. This inhibitory activity was completely abolished by pretreatment of cultures with indomethacin. The addition of exogenous PGE, in the range of concentrations produced by the fibroblasts abrogated the TGFP stimulation of collagen and fibronectin production. Furthermore, the measurement of PGE, levels in the media of cultures treated with TGFP showed a dosedependent increase in fibroblast PG production. These observations suggest that the marked elevation in PGE, levels induced by high concentrations of TGFP was responsible for the negative portion of the TGFP dose-response curve of collagen and fibronectin production. In addition, our observation that in some cell lines pretreatment with indomethacin permitted the expression of the stimulatory effect of TGFP on collagen synthesis implicated TGFP-induced endogenous prostaglandin production as a possible explanation for the variable response to TGFP occasionally noted by us and others (11,12). The levels of mRNAs for al (1) and al(II1) procollagens and for fibronectin were increased in TGFP-treated lung fibroblasts. The TGFP-induced elevation of the levels of these transcripts was increased further when endogenous PG production was inhibited by indomethacin. These findings provide evidence that the down-regulation of collagen and fibronectin production in TGFP-treated fibroblasts by endogenous PGE occurs largely at a pretranslational level.

TABLE I11 Densitometric analysis of total RNA from TGFPand indomethacin (ZND)-treated lung fibrobhts hybridized to procollagen al(Z), procollagen al(III), and fibronectin cDNA clones Northern blot autoradiographs shown in
Our results suggest that the stimulation of endogenous fibroblast PGE, production by TGFP may serve as a negative feedback mechanism to limit the increase in extracellular matrix protein production induced by this cytokine. Similar autocrine modulation of collagen biosynthesis mediated by endogenous PG has been demonstrated in fibroblasts stimulated with bradykinin (26) and interleukin 1 (27). The observations described in this paper raise the possibility that endogenous fibroblast PG may play a role in limiting the stimulation of extracellular matrix proteins deposition induced by TGFP in vivo. A derangement in this autocrine regulatory mechanism could contribute to the initiation or progression of pathological states characterized by fibrosis.