Costimulation of fibroblast collagen and transforming growth factor beta1 gene expression by monocyte chemoattractant protein-1 via specific receptors.

Recent studies indicate potential roles of monocyte chemotactic protein-1 (MCP-1) in recruitment of monocytes to sites of inflammation. However, their increased expression does not always correlate with monocyte influx, suggesting other possible biological activities for this member of the C-C chemokine family. In view of its potential role in regulating extracellular matrix expression in fibrotic disorders, the effects of MCP-1 on lung fibroblast collagen expression were evaluated. Isolated rat lung fibroblasts were treated with increasing doses of MCP-1 for variable periods of time and examined for effects on collagen synthesis and expression of procollagen α1(I) mRNA expression. The results show that MCP-1 was able to stimulate collagen expression in these cells in a dose-dependent manner but required over 24 h for significant elevation to occur. In view of this delayed time course, the possibility of mediation via endogenous transforming growth factor β (TGFβ) was tested by the ability of anti-TGFβ antibody to inhibit this MCP-1 stimulation of collagen expression. Significant but incomplete inhibition by this antibody was observed. Pretreatment of the cells with antisense but not by sense or missense TGFβ1 oligodeoxyribonucleotides caused essentially complete inhibition of this MCP-1 stimulatory effect. Furthermore, MCP-1 treatment was found to also stimulate TGFβ secretion and mRNA expression, which was also abolished by pretreatment with antisense TGFβ1 oligodeoxyribonucleotides. The kinetics of TGFβ expression indicates that significant increase preceded that for collagen expression. Binding studies using 125I-labeled MCP-1 indicated the presence of specific and saturable binding sites with a dissociation constant consistent with the dose response curves for stimulation of fibroblast collagen synthesis and TGFβ activity by MCP-1. These results taken together suggest that MCP-1 stimulates fibroblast collagen expression via specific receptors and endogenous up-regulation of TGFβ expression. The latter then results in autocrine and/or juxtacrine stimulation of collagen gene expression.

Recent studies indicate potential roles of monocyte chemotactic protein-1 (MCP-1) in recruitment of monocytes to sites of inflammation. However, their increased expression does not always correlate with monocyte influx, suggesting other possible biological activities for this member of the C-C chemokine family. In view of its potential role in regulating extracellular matrix expression in fibrotic disorders, the effects of MCP-1 on lung fibroblast collagen expression were evaluated. Isolated rat lung fibroblasts were treated with increasing doses of MCP-1 for variable periods of time and examined for effects on collagen synthesis and expression of procollagen ␣ 1 (I) mRNA expression. The results show that MCP-1 was able to stimulate collagen expression in these cells in a dose-dependent manner but required over 24 h for significant elevation to occur. In view of this delayed time course, the possibility of mediation via endogenous transforming growth factor ␤ (TGF␤) was tested by the ability of anti-TGF␤ antibody to inhibit this MCP-1 stimulation of collagen expression. Significant but incomplete inhibition by this antibody was observed. Pretreatment of the cells with antisense but not by sense or missense TGF␤ 1 oligodeoxyribonucleotides caused essentially complete inhibition of this MCP-1 stimulatory effect. Furthermore, MCP-1 treatment was found to also stimulate TGF␤ secretion and mRNA expression, which was also abolished by pretreatment with antisense TGF␤ 1 oligodeoxyribonucleotides. The kinetics of TGF␤ expression indicates that significant increase preceded that for collagen expression. Binding studies using 125 I-labeled MCP-1 indicated the presence of specific and saturable binding sites with a dissociation constant consistent with the dose response curves for stimulation of fibroblast collagen synthesis and TGF␤ activity by MCP-1. These results taken together suggest that MCP-1 stimulates fibroblast collagen expression via specific receptors and endogenous up-regulation of TGF␤ expression. The latter then results in autocrine and/or juxtacrine stimulation of collagen gene expression.
Monocyte chemoattractant protein-1 (MCP-1 1 ) belongs to the C-C chemokine supergene family that is related to other families of proinflammatory cytokines, including the neutrophilspecific chemoattractant interleukin 8 (known as NAP-1) (1). The C-C chemokine supergene family has been cloned and characterized, yielding many putative mediators of macrophage-, lymphocyte-, and granulocyte-derived responses in animal models of human diseases (2-3). The C-C chemokine family includes MCP-1, MCP-2, MCP-3, HC-14, Clo, I-309, macrophage inflammatory protein-1␣, macrophage inflammatory protein-1␤, and RANTES (regulated on activation normal T cell expressed). MCP-1 has also been characterized as the murine JE gene product. In vitro studies demonstrated stimulus-specific induction of MCP-1 mRNA from human fibroblasts and endothelial cells (4), originally described for its monocyte chemotactic activity, but as a more potent chemoattractant for lymphocytes in vitro (5). Numerous cell types, including endothelial cells, epithelial cells, mononuclear cells, fibroblasts, and smooth muscle cells, are now known to be capable of expressing MCP-1 in the presence of serum or specific stimuli (6 -11).
Several studies have documented that MCP-1 is up-regulated in rat and other rodent models of pulmonary fibrosis (12)(13)(14)(15)(16). Passive immunization of granuloma-bearing CBA/J mice with anti-MCP-1 antibody decreases the size and cellularity of the inflammatory granulomatosis lesion (17)(18) as well as human idiopathic pulmonary fibrosis (19 -20). These results provide evidence that increases in MCP-1 and macrophage inflammatory protein-1␣ expression contribute to the development of both acute and chronic inflammatory lesions (21). Pulmonary fibrosis is the end point of a chronic inflammatory process characterized by leukocyte recruitment and activation, fibroblast proliferation, and increased extracellular matrix production with deposition of collagen type I along with increased TGF␤, MCP-1, and interleukin 1 (22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35)(36)(37)(38). Although MCP-1 expression is elevated in affected tissues, their kinetics is more prolonged or sustained than that for mononuclear cell influx into the lung and extended to the period of maximal lung collagen expression (12,22,30). These results suggest that MCP-1 may have activities beyond the mere recruitment of leukocytes and may have more direct stimulatory effects on collagen expression as well. This study was designed to determine whether MCP-1 could stimulate rat lung fibroblast collagen gene expression and to provide insights as to mechanisms.

EXPERIMENTAL PROCEDURES
Cells-Rat lung fibroblasts were isolated from 150 -175 g male Fisher 344 rats by trypsinization of lung tissue as described previously (20). Dulbecco's modified Eagle's medium was supplemented with 10% fetal bovine serum and passed by trypsinization. Only cells earlier than the tenth passage from primary culture were used for these studies.
A mink lung epithelial cell line was used for TGF␤ assay. These cells were obtained from the American Type Culture Collection, Rockville, MD, and were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 1% antibiotics, and 2 mM L-glutamine. For preparation of rat lung fibroblast conditioned media, cells were grown to confluence and washed; then the medium was replaced with serum-free Dulbecco's modified Eagle's medium containing 2 mg/ml bovine serum albumin (Bovuminar reagent pure powder, Intergen Company, Purchase, NY), supplemented without or with the indicated concentrations of recombinant human (R&D Systems, Minneapolis, MN) or rat (PeproTech, Inc., Rocky Hill, NJ) MCP-1. They were incubated at 37°C in a humidified atmosphere at 5% CO 2 for the indicated times. At the end of each indicated incubation period, the medium was collected for TGF␤ assay. The conditioned medium was stored frozen at Ϫ70°C until assay and characterization. After removal of the medium, total RNA was extracted from the cells. Some cells also were used for slide preparation for in situ hybridization. The rat and human MCP-1 gave similar responses; therefore, unless otherwise stated, all experiments described used human MCP-1.
Collagen Synthesis Assay-Rat lung fibroblasts were grown to confluence in 12-well (22-mm diameter) tissue culture dishes and treated with the indicated concentration of recombinant human MCP-1 diluted in serum-free Dulbecco's modified Eagle's medium containing 2 mg/ml bovine serum albumin and 50 g/ml ascorbic acid for the indicated times. The cells were pulsed for the final 6 h with [ 3 H]proline in the presence of ascorbic acid and ␤-aminopropionitrile, as described previously (39). Collagen synthesis was expressed as the collagenase-susceptible incorporated radioactivity in dpm/10 5 cells.
TGF␤ Assay-This assay was performed on fibroblast conditioned medium prepared as described above using methods described previously (40 -42). Samples were activated by acidification prior to assay, as described previously (40 -42). Activity was expressed as ED 50 units, the dose necessary to cause 50% inhibition of [ 3 H]thymidine incorporation by the mink lung epithelial cells, which was reversible by specific anti-TGF␤ antibody (mouse monoclonal antibody 1D11.6, which neutralizes TGF␤ 1 with weak cross-reactivity to TGF␤ 2 ) (Celtrix Pharmaceuticals, Santa Clara, CA). When compared to a standard curve generated using porcine TGF␤1 (R&D Systems), one unit is approximately equivalent to 1-5 pg of the porcine protein. Given the isoform specificity of the antibody, the activities reported by this assay were specific for TGF␤1, and to a lesser extent TGF␤ 2 , but did not detect the TGF␤ 3 isoform. Activities were normalized to the volume of conditioned medium corresponding to equal numbers of cells that were directly counted from the plate after removal of the conditioned medium.
mRNA Isolation and Northern Blot Hybridization-Rat lung fibroblasts were grown in 75-cm 2 tissue culture flasks to confluence as described above and treated with or without the indicated concentration of recombinant human MCP-1 for the indicated times. After removal of medium, total RNA was extracted from the cells as described previously (41)(42)(43)(44)(45)(46). Briefly, the cells were lysed with 4 M guanidine isothiocyanate solution and then extracted twice with phenol:chloroform 1:1 and twice with chloroform:isoamyl alcohol. Equal amounts (20 g) of total RNA were separated by electrophoresis in formaldehydecontaining 1% agarose gels. Equal loading of RNA was judged based on the ethidium bromide-stained gels under ultraviolet light as well as by hybridization with a cyclophilin probe. The mRNA was transblotted onto nitrocellulose membranes, baked for 2 h at 80°C in a vacuum oven, prehybridized, and hybridized with a synthetic oligonucleotide antisense probe. The 30-base oligodeoxyribonucleotide antisense probes for procollagen ␣ 1 (I), TGF␤ 1 , and MCP-1 were synthesized by an automated DNA synthesizer (Gene Assembler) and purified by high performance liquid chromatography before use. The sequences for these probes were as follows: 5Ј-AGGGCCAGTCTCAGCACGGTCACCCTT-GGC-3Ј for procollagen ␣ 1 (I) (22); 5Ј-GAAGTTGGCATGGTAGCCCTT-GGGCTCGTG-3Ј for TGF␤ 1 (42); and 5Ј-AGTGAATGAGTAGCAGGAG-GTGAGTGGGGC-3Ј for MCP-1 (12). The probes were labeled with 32 P using 3Ј end-labeling, as described previously (41). After overnight hybridization at 56°C, blots were washed as described previously (22). Autoradiographs developed from these blots were quantitated using the Ambis Optical Imaging System (Ambis, San Diego, CA) (49), and the results were expressed as random integration units. For quantitative analysis, results were normalized to the signal produced with the rat cyclophilin probe (16).
In Situ Hybridization-In situ hybridization was undertaken using the same 30-mer oligonucleotide antisense probes for rat procollagen ␣ 1 (I) and TGF␤ 1 , whereas the corresponding sense probes were used for negative control purposes. Probes were labeled with 35 S-labeled dATP by 3Ј-end labeling and purified by electrophoresis.
Confluent rat lung fibroblast as described above and treated with the indicated dose of MCP-1 for the indicated times were detached by brief treatment with 0.01% trypsin and 1 mM EDTA and neutralized with 1% fetal bovine serum. After cell counting, the suspension was adjusted to 5 ϫ 10 5 cells/ml, and 100 l were applied per slide by centrifugation. In situ hybridization was performed as described previously (47)(48)(49). Briefly, the cytospun slides were fixed in 4% paraformaldehyde for 20 min, washed in 2 ϫ standard saline citrate, treated with 0.2 N HCl, and 0.25% acetic anhydride (v/v) in triethanolamine after sequential washes in 2 ϫ standard saline citrate. Slides were prehybridized and hybridized with the indicated radiolabeled oligonucleotide probe, followed by sequential washing with various dilutions of standard saline citrate and dehydration with alcohol. Then slides were air-dried and dipped in Kodak nitro blue tetrazolium emulsion solution. The emulsion-coated slides were air-dried in the dark and stored in a desiccating chamber at 4°C for 2 days for collagen and 10 days for TGF␤ and MCP-1. They were then developed in Kodak D-119 developer, washed and counterstained with hematoxylin and eosin, and coverslipped. Controls for in situ hybridization included the following: (a) hybridization using the corresponding sense probes instead of the antisense probe; (b) pretreatment of slides with 100 g/ml of RNase at 37°C for 45 min before the addition of antisense probe; and (c) slides made from normal rat lung fibroblasts treated with interleukin 1␤ (2 ng/ml) for 24 h as positive control for in situ hybridization analysis of TGF␤ 1 and MCP-1 mRNA.
Anti-TGF␤ and Antisense Experiments-The role of TGF␤ in mediating the MCP-1 stimulation of fibroblast collagen synthesis was examined in experiments using either specific TGF␤ antibody or TGF␤ 1 antisense phosphorothioate-derivatized oligodeoxyribonucleotides (ODNs). To examine the effects of TGF␤ neutralization using specific antibody, confluent cells were pretreated for 2 h with 100 g/ml of anti-TGF␤ antibody (see above for description of antibody) or nonimmune IgG (negative control), treated with MCP-1 at the indicated doses and times, and then assessed for collagen synthesis as described above.
The inhibition of TGF␤ 1 expression was also attempted by pretreatment of cells with antisense TGF␤ 1 ODN. Where indicated, confluent fibroblasts were pretreated for 24 h with 10 M sense, missense, or antisense ODN in serum-free medium supplemented with 2 mg/ml bovine serum albumin. The cells were then treated with the indicated agonists, the conditioned medium was harvested, and the cells were extracted for total RNA for Northern analysis or detached for in situ hybridization analysis as described above. The ODNs used for these studies were: sense, 5Ј-CACGAGCCCAAGGGCTAC-3Ј; antisense, 5Ј-GTAGCCCTTGGGCTCGTG-3Ј; and missense, 5Ј-AACGATTTCCAG-AGCCAC-3Ј.
Preliminary studies indicated that ODN doses Ͼ10 M did not cause significant further inhibition of TGF␤ 1 expression in these cells, whereas doses Ͻ7.5 M caused significantly less inhibition. None of the doses tested caused detectable toxicity, as judged by trypan blue exclusion, or impaired protein synthesis, measured by radioactive proline incorporation (data not shown).
MCP-1 Binding Studies-125 I-Labeled recombinant human MCP-1 (2200 Ci/mmol, Dupont NEN) was used to assess the presence of MCP-1 binding sites on rat lung fibroblasts. Confluent cells in 6-well (35-mm diameter wells) tissue culture plates were pre-chilled on ice in serumfree medium supplemented with 2 mg/ml bovine serum albumin for 1 h. The indicated concentrations of radiolabeled MCP-1 were then added with or without 100-fold excess cold MCP-1. After incubating on ice for the indicated times, an aliquot of the medium was sampled and counted for radioactivity. The remainder of the medium was removed, and the cell layer was washed with cold PBS three times. The cells were then dissolved with a buffer containing 2% sodium dodecyl sulfate, 1 mM EDTA, and 50 mM Tris-HCl, pH 7.6, and counted. Assays were done in triplicate.
Statistical Analysis-All of the data were expressed as means Ϯ S.E., and differences between means of various treatment and control groups were assessed for statistical significance by analysis of variance, followed by post hoc analysis using Scheffé's test for comparison of any two groups (30). Differences with p Ͻ 0.05 were considered significant.

Effects of MCP-1 on Collagen Expression-Both human and
rat MCP-1 caused significant stimulation of rat lung fibroblast collagen synthesis. Human MCP-1 did this at doses Ն100 ng/ ml, approaching maximal stimulation at 400 ng/ml (Fig. 1). Peak stimulation, however, required Ͼ16 h of treatment. This stimulation was also observed at the mRNA (for ␣ 1 (I) procollagen) level, as shown in the Northern blots in Fig. 2. Here again, over 24 h were required for maximal stimulation by 200 ng/ml MCP-1. In view of this delay in the time course for stimulation of collagen gene expression, the possibility that this could be mediated by an intermediate signal also stimulated by MCP-1 treatment was explored. Since these fibroblasts are known to produce TGF␤, a potent stimulator of extracellular matrix synthesis, the effects of anti-TGF␤ antibody on this stimulation of collagen expression was examined. The results show that MCP-1 stimulation of fibroblast collagen synthesis was partially inhibited by 52 Ϯ 8% in the presence of anti-TGF␤ antibody at concentrations found to completely inhibit TGF␤ activity in the conditioned medium. Doubling the dose of antibody did not cause further inhibition. Nonimmune IgG had no significant effects on collagen synthesis (data not shown).
Given this incomplete inhibition by anti-TGF␤ antibody treatment, an alternative approach was also used to abrogate endogenous TGF␤ expression. When cells were pretreated with a TGF␤ 1 antisense but not sense ODN, MCP-1 stimulation of collagen synthesis was completely suppressed (Fig. 3). This inhibition was also seen at the ␣ 1 (I) procollagen mRNA level as well, without significant effects on cyclophilin mRNA (Fig. 4). To confirm that the TGF␤ 1 antisense ODN caused inhibition of MCP-1 stimulation of collagen expression was due to suppression of endogenous TGF␤ expression, the antisense studies were repeated to evaluate for effects on fibroblast TGF␤ production. Fig. 4 also shows that TGF␤ 1 antisense but not sense ODN pretreatment inhibited TGF␤ 1 steady-state mRNA levels.
Similar inhibitory effects on TGF␤ activity secretion were seen with antisense but not sense or missense ODN (data not shown). TGF␤ 1 antisense ODN treatment also inhibited basal levels of collagen synthesis and mRNA expression (Figs. 3 and  4). In situ hybridization is useful for evaluation of distribution and localization of mRNA expression (50 -53), and in this study, the results confirm the Northern blot data (Fig. 5) and further show that the cells responded in a heterogeneous manner (Fig. 5C), with some cells showing striking increases in collagen mRNA while others barely responded.
Effects of MCP-1 on TGF␤ Expression-Given the dramatic inhibitory effects of TGF␤ 1 antisense ODN on MCP-1 stimulation of collagen gene expression, the effects of MCP-1 on fibroblast TGF␤ expression itself was examined as a way of confirming the importance of this autocrine/juxtacrine loop in the MCP-1 effects on collagen expression. Fig. 6 shows that MCP-1 also stimulated secretion of TGF␤ activity by rat lung fibro- Fibroblasts were treated as described in the legend to Fig. 3. Total RNA was then isolated and subjected to Northern analysis for ␣ 1 (I) procollagen, TGF␤ 1 , and cyclophilin mRNA, as described under "Experimental Procedures." Pretreatment with TGF␤ 1 antisense, but not sense, ODN caused significant inhibition of the MCP-1-stimulated increases in both procollagen and TGF␤ 1 expression while also inhibiting basal expression. Cyclophilin mRNA was unaffected by any of the treatments. blasts in a dose-dependent manner. The activities detected in the conditioned medium were in the range of 1-50 ng/ml in equivalent porcine TGF␤ 1 protein mass. The response, however, was more rapid compared to that for stimulation of ␣ 1 (I) procollagen gene expression, with significant stimulation as early as 12 h (Fig. 7). When the kinetics were compared at the mRNA level, a similar correlation was found (Fig. 8). Basal ␣ 1 (I) procollagen mRNA levels gradually increased over time in culture; however, treatment with MCP-1 caused a substantially greater stimulation in mRNA levels. Steady-state TGF␤ 1 mRNA levels were also stimulated by MCP-1 with a kinetics that correlated well with that for the ␣ 1 (I) procollagen mRNA (Fig. 8). Cyclophilin mRNA was not significantly affected by the MCP-1 treatment over the same time points.
MCP-1 Binding Studies-In view of the above functional responses to MCP-1, rat lung fibroblasts were evaluated for the presence of MCP-1 receptors. Incubation of these cells with increasing doses of radiolabeled MCP-1 showed dose-dependent increases in specific binding that approached saturation at 20 nM (Fig. 9A). Scatchard analysis revealed K d ϭ 5.0 Ϯ 1.3 ϫ 10 Ϫ8 M and B max ϭ 7.4 Ϯ 1.1 ϫ 10 Ϫ11 M (Fig. 9B), which translates to 2.23 ϫ 10 4 binding sites per cell. Binding at this temperature showed a lag time of about 30 min, before it rapidly approached peak values at 60 min (Fig. 10). These binding kinetics suggest that MCP-1 stimulated fibroblast collagen and TGF␤ expression via these receptors. DISCUSSION MCP-1 is a specific and potent chemoattractant for monocytes and has been implicated in a variety of inflammatory and fibrotic diseases, the pathogenesis of which is known to involve infiltration and activation of monocytes (12). Up-regulation of MCP-1 production has been shown in pulmonary fibrosis in animal models (12). A number of cell types, including macrophages, lymphocytes, and fibroblasts, have been shown to both produce and/or be modulated by MCP-1 and another C-C chemokine macrophage inflammatory protein-1␣ (21). In this context, the specific roles of MCP-1 in these diseases have not been definitively explored. Thus, the basis for increased MCP-1, TGF-␤ collagen synthesis and deposition in pulmonary fibrosis remains unclear. In this study, we have examined the possibility that MCP-1 may be involved in the production of mediators or cytokines with fibrogenic properties. To test this hypothesis, we examined first the effect of MCP-1 on rat lung fibroblast collagen synthesis. The results show that MCP-1 increased fibroblast collagen synthesis, with highest responses observed at a dose of MCP-1 Ն 200 ng/ml. These doses of MCP-1 are comparable to those found to stimulate monocyte chemotaxis (54,55) and basophil (56 -58) and monocyte activation (59). Since mixed monocyte and lymphocyte cultures easily secrete a minimum of 40 nM (approximately 400 ng/ml) MCP-1 into their conditioned medium (60), it is not inconceivable that these effective doses of MCP-1 could be encountered in vivo.
On the basis of the selective stimulatory effect on collagenous protein synthesis, the activity could be mediated by costimulation of TGF␤ expression. This possibility was supported by the following findings. (a) TGF␤ activity was significantly elevated in conditioned medium of MCP-1-stimulated fibroblasts with a time course which correlated with the increases in collagen synthetic rate. (b) Fibroblasts stimulated with MCP-1 contained significantly higher steady-state levels of TGF␤ 1 mRNA, which again preceded the peak increase in procollagen ␣ 1 (I) gene expression. (c) The levels of TGF␤ activity in conditioned medium of MCP-1-stimulated cells were in the range known to stimulate collagen synthesis in these cells (42). (d) MCP-1induced increases in collagen and TGF␤ 1 gene expression were inhibited by both anti-TGF␤ antibody and TGF␤ 1 antisense ODN. (e) The presence of specific MCP-1 binding sites with comparable kinetics as that for the observed functional effects confirm these new biological activities of MCP-1. Taken together, these results suggest that MCP-1 stimulates collagen synthesis indirectly via endogenous up-regulation of TGF␤ gene expression, which could then stimulate collagen gene expression via an autocrine and/or juxtacrine loops. These effects appear to be mediated by specific receptors expressed by these cells. These novel findings indicate potentially important new roles for this C-C chemokine in the pathogenesis of a number of fibrotic diseases and in tissue repair.
Acknowledgments-We thank Celtrix Pharmaceuticals for the gift of the 1D11.16 monoclonal anti-TGF␤ antibody and acknowledge the excellent technical assistance of Bridget McGarry and Lisa Riggs.