Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Elastin is a major component of the lung extracellular matrix, and is essential for normal lung function. Elastic fibers, together with collagen fibers, form a continuous network throughout the lung that provides the forces necessary for passive expiration (1, 2). Lung elastic fibers are predominant in the vascular media and in the medial region of conducting airways; they also line the walls of terminal bronchioles and alveolar ducts and form ringlike structures at the entrances to individual alveoli. The physiologic importance of these fibers is exemplified by the direct association between their disruption and impaired lung function in emphysema (3).

Elastic fibers are now appreciated as highly complex structures containing numerous protein and glycoprotein components. The principal component of elastic fibers, by weight, is the crosslinked form of the soluble protein tropoelastin. In the distal airways, tropoelastin is produced primarily during the alveolarization stage of lung development, by mesenchymal cells positive for -smooth-muscle actin (-SMA), located at sites of secondary crest formation (4). Gene-deletion studies have suggested a direct correlation between the differentiation of these cells from undifferentiated mesenchymal precursors and the production of elastin in the developing lung (7). In the adult lung, -SMA-expressing fibroblasts are seen at alveolar septal tips (8), but tropoelastin expression is rarely seen because of the stability of mature elastic fibers (9). Developmental induction of lung tropoelastin expression occurs at the level of gene transcription, whereas posttranscriptional mechanisms are involved in its cessation (10).

We have previously described an elastogenic response to silica instillation in an experimental model of granulomatous inflammation of the lung (8). This elastogenic response is complex, with developing granulomatous lesions devoid of elastin gene expression, and peripheral, uninvolved airways showing renewed elastin production and accumulation. Because fibrotic cells typically found in lung pathologies express -SMA (11, 12), we sought to investigate the mechanism(s) responsible for the coordinate lack of tropoelastin and -SMA expression within silicotic rat lung granulomas. We have previously shown that silicotic granulomas are rich in infiltrating, activated macrophages releasing cytokines that alter the phenotype of adjacent mesenchymal cells (13, 14). Here we report that particulate-activated macrophages produce and secrete tumor necrosis factor (TNF)-, which suppresses tropoelastin gene expression in vivo and transcription in vitro. Furthermore, we found a correlated repression of -SMA gene expression. These data extend the association between -SMA and tropoelastin gene expression in the diseased lung, and provide insight into the mechanism of their coordinate regulation. The data also identify TNF- as a physiologic repressor of endogenous tropoelastin expression during inflammatory lung disease.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals and Reagents

Normal adult (250 g) Sprague-Dawley rats were purchased from Charles River (Cambridge, MA). Silica (Min-U-Sil; 5 µm) was purchased from U.S. Silica (Berkley Springs, WV) and rendered endotoxin-free by boiling in acid. The murine monocytic cell line J774.1 was obtained from the American Type Culture Collection (Rockville, MD). Zymosan and pepsin were obtained from Sigma Chemical Co. (St. Louis, MO). Murine TNF-, (IL)-interleukin-1, TNF- and IL-1-neutralizing antibodies, and soluble TNF--receptor (sTNF-R) were purchased from R&D Systems (Minneapolis, MN).

In Situ Hybridization and Immunostaining

In situ hybridization was performed on lung sections from normal adult rats (250 g), rats receiving a single intratracheal instillation of 0.4 ml sterile saline containing 16 mg silica (Min-U-Sil; 5 µm), or rats receiving a single intratracheal instillation of 0.4 ml sterile saline as a vehicle control 7 d before being killed. A detailed description of the induction of silicosis and the methods used for in situ hybridization have previously been published (8, 13). Antisense probes for rat tropoelastin were generated from the complementary DNA (cDNA) pREL124DM. Immunostaining was performed with a Vectastain immunostaining kit (Vector Laboratories, Burlingame, CA) according to the manufacturer's instructions. Antibody clone ED1 (Harlan Inc., Indianapolis, IN) was used to identify monocyte/macrophage cells. Detection of TNF- was done with a commercially available antibody for murine TNF- (R&D Systems).

Cell Culture

Primary macrophages were obtained by bronchoalveolar lavage (BAL) from healthy or silicotic adult rat lungs 7 d after instillation of saline or silica, respectively. Briefly, animals were killed with a lethal dose of sodium pentobarbital. The trachea was cannulated and the pulmonary circulation was washed free of blood with saline. The lungs were removed from the thoracic cavity and lavaged repeatedly with a single 10-ml aliquot of normal saline. After lavage, the saline was collected and centrifuged at 500 × g for 5 min at 4°C. The cell pellet was resuspended in culture medium (Dulbecco's modified Eagle's medium supplemented with 10% calf serum, nonessential amino acids, L-glutamine, and antibiotics), and nucleated cells were counted with a hemocytometer. Cells were plated on tissue-culture-treated plastic dishes at a concentration of 0.5 × 10⁶ cells/ml and were allowed to adhere overnight. On the following day the culture dishes were washed to remove dead cells, and the remaining cells were refed with culture medium. Lavaged cells in medium-conditioning cultures consisted of more than 95% macrophages as assessed by ED1 staining. J774 cells were plated at a concentration of 0.5 × 10⁶ cells/ml and allowed to adhere overnight. On the following day, J774 cells were washed with phosphate-buffered saline and refed with culture medium. Lung fibroblasts were isolated from normal adult rats as previously described (10). Lung fibroblast gene expression was assessed in early-passage confluent cultures.

Isolation of Cells and Application of Conditioned Medium

Primary macrophages or J774 cells were washed and refed with culture medium in the absence or presence of silica (0.1 mg/ml; 12.5 µg/cm²) or zymosan (4 mg/ml; 0.5 mg/ cm²). Conditioned medium was harvested after 24 h and centrifuged to pellet particulates and cell debris. The supernatant was transferred to a sterile tube and stored at 4°C. The conditioned medium was filter-sterilized through a 0.2-µm syringe filter (Gelman Sciences, Ann Arbor, MI) prior to application to fibroblasts.

Confluent plates of lung fibroblasts at Passage 3 were washed with phosphate-buffered saline and refed with culture medium or with medium supplemented with 50% macrophage-conditioned medium. Treatment with conditioned medium was for 6-96 h, with changes of medium every 48 h. Replicate plates of lung fibroblasts were treated directly with silica (0.1 mg/ml; 12.5 µg/cm²) or zymosan (4 mg/ml; 0.5 mg/cm²) as controls. For serum-free experiments, all cells were grown as described previously, but the conditioning and treatment media used were serum-free. For each assay, at least three separate cultures of mesenchymal cells were treated as described and examined for extracellular matrix gene expression to ensure reproducibility.

Northern Blot Analysis

RNA isolation and Northern blot analysis were performed as previously described (15). Briefly, total RNA was isolated from cultured cells using a modified guanidine-phenol method. Five micrograms of total RNA were denatured by incubation for 10 min at 68°C in 50% formamide, 1 M formaldehyde, and 50 ng/ml ethidium bromide, and was separated by electrophoresis in a 1% agarose gel containing 1 M formaldehyde. RNA was transferred to Hybond N⁺ filters (Amersham, Arlington Heights, IL). Rat tropoelastin messenger RNA (mRNA) was detected with a purified insert DNA derived from the clone described earlier for in situ hybridization analysis. Clone pRGAPDH13 was used for the detection of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA as a control (9, 12). The cDNA probe for actin was obtained from Clontech (Palo Alto, CA).

Assessment of Tropoelastin Gene Transcription Rate

Levels of tropoelastin pre-mRNA in fibroblast cultures were measured as a reflection of the transcription rate for the tropoelastin gene, according to the method of Swee and colleagues (16). Briefly, total RNA isolated for Northern blot analysis was digested with ribonuclease-free deoxyribonuclease to remove residual genomic DNA, reverse-transcribed with random hexamers to make cDNA, and amplified with the polymerase chain reaction (PCR), using tropoelastin intron-specific primers (5'-GTCAGAGGTCAAGGTCTAGG-3' and 5'-TCAGTCTAGACATGCAACAC-3'). Amplification was optimized for template concentration and cycle number, and was determined to be within a linear range. Amplification products were separated in a 1.2% agarose gel, blotted to Hybond N⁺ filters and probed with [³²P]deoxycytosine triphosphate ([³²P]dCTP)-labeled internal oligonucleotide primer (5'- GACATACCACCAGGTGGCGC-3'). After stringent washing, the hybridized blot was exposed for various periods of time to X-ray film. Autoradiograms were scanned and digitally captured in a Power Macintosh computer (Apple, Inc., Cupertino, CA), and were quantified by comparing the mean pixel density of each band using the NIH Image program (National Institutes of Health, Bethesda, MD). Control reactions lacking reverse transcriptase were performed to rule out contamination of RNA by genomic DNA. Amplification of GAPDH mRNA from the fibroblast RNA samples was used to control for template concentration loading.

Analysis of the Role of TNF- in Conditioned Medium

To characterize macrophage-derived activity, serum-free, zymosan-activated J774-conditioned medium was subjected to various treatments. Acid lability was determined by adjusting the pH of the conditioned medium to 2.5 with 0.5 N HCl, incubating for 2 h at 37°C, and neutralizing with 0.5 N NaOH. Pepsin sensitivity was tested by following the same procedure, with the inclusion of 0.2 mg/ml pepsin. After each treatment, the conditioned medium was dialyzed against serum-free culture medium and filter sterilized before being added to fibroblasts.

To investigate the effects of TNF- on rat tropoelastin gene expression, primary cultures of lung fibroblasts were treated with 20 ng/ml recombinant murine TNF- for 48 h. The role of TNF- in the zymosan-activated J774-conditioned medium was assessed by neutralization with a murine TNF- neutralizing antibody. Fifty percent zymosan-activated, J774-conditioned medium was incubated with 10 µg/ml anti-TNF- antibody for 2 h at 37°C with gentle shaking. Identical treatment of conditioned medium with a murine IL-1-neutralizing antibody (10 µg/ml) served as a control.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

TNF- Is Present in Developing Granulomatous Lesions

We have previously shown that developing granulomatous lesions in an experimental model of rat lung silicosis containing macrophages are devoid of de novo elastin production and contain no -SMA-expressing cells (8, 13). Previous reports have suggested that similar lesions were rich in the proinflammatory cytokine TNF- (17, 18). Since this cytokine has been shown to be capable of inhibiting tropoelastin expression (19), we chose to assess TNF- production in our model of rat lung silicosis, and its association with tropoelastin expression. In situ hybridization for tropoelastin mRNA (Figures 1A and 1B) showed that tropoelastin expression was abundant peripherally to developing silicotic lesions, but absent within them. Immunohistochemistry for -SMA (Figure 1C) and macrophage marker antibody clone ED1 (Figure 1D) confirmed that these lesions were rich in infiltrating lung macrophages, but contained no -SMA-expressing mesenchymal cells. Immunohistochemistry for TNF- (Figure 1E) revealed abundant staining for TNF- within these granulomatous lesions in both cell-associated and matrix-associated patterns. Dual immunostaining for TNF- and macrophages (Figure 1F) showed that macrophages were abundant in the regions of granulomas rich in TNF-. Higher magnification views of sections stained for both TNF- alone (Figure 1G) or TNF- and ED1 (Figure 1H) revealed that the cell-associated pattern of TNF--staining colocalized with macrophage cells. This strongly suggests that the infiltrating macrophages produced and secreted the antielastogenic cytokine TNF- within developing silicotic granulomas. Alveolar macrophages (AM) distant from sites of granulomas also stained positively for TNF- (data not shown).

View larger version (99K): [in this window] [in a new window]   Figure 1.   Silicotic rat-lung-granuloma-infiltrating macrophages produce and secrete TNF-. Lungs were isolated from rats 7 d after receiving a single intratracheal instillation of silica. Lungs were formalin fixed, paraffin embedded, and sectioned for histologic analysis. (A and B) Paired bright- and darkfield views of a silicotic lung section hybridized in situ for tropoelastin mRNA. Expression of the tropoelastin gene is observed at alveolar septal tips adjacent to granulomas (arrowheads), but no expression is seen within the lesions. (C) Immunohistochemistry for -SMA in a silicotic rat lung section. Granulomatous lesions contain no -SMA-positive cells, whereas alveolar septae contain -SMA-positive cells (red). (D) Immunohistochemistry for macrophages (antibody clone ED1) in a silicotic rat lung section. Granulomatous lesions are rich in infiltrating macrophages (red). (E and G) Immunohistochemistry for TNF- in a silicotic rat lung section. TNF- is present within granulomatous lesions (brown), appearing in both a cell-associated (arrows) and matrix-associated pattern. (F and H) Dual immunohistochemistry for TNF- (brown) and macrophages (red) in a silicotic rat lung section. Colocalization is noted as a blending of brown and red staining. Macrophages are concentrated in TNF--rich foci of granulomatous lesions (F ), and individual cells (arrows) stain positively for both macrophage-specific markers and TNF- (H); g-silicotic granuloma. Original magnification: ×100 in (A and B); ×200 in (C and D); ×500 in (E and F); ×1,000 in (G and H).

Particulate-Activated Macrophages Repress Lung Fibroblast Tropoelastin Expression

In order to confirm and mechanistically explore the association between infiltrating macrophage TNF- production within silicotic lung granulomas and tropoelastin gene expression, we sought to develop an in vitro model (Figure 2). Initially, we isolated particulate-activated macrophages from silicotic rat lungs by BAL, allowed these cells to condition medium in culture, and assessed the effects of this medium on primary cultures of rat lung fibroblasts. As shown in Figure 2A, medium conditioned by macrophages isolated from silicotic lungs dramatically repressed lung fibroblast tropoelastin mRNA levels. This was not due to a generalized inhibition of gene expression, as evidenced by the previously described induction of collagenase-3 mRNA levels under the same conditions (14). These data indicate that macrophages activated by particulate silica in vivo elaborate factor(s) that specifically repress tropoelastin gene expression.

View larger version (22K): [in this window] [in a new window]   Figure 2.   Particulate-activated macrophages produce a tropoelastin-repressing activity. Conditioned medium from macrophages isolated from silicotic rat lungs 7 d after silica instillation (A), from healthy adult rat lung macrophages (B), or from the J774 cell line (C) was tested for tropoelastin-regulating activity in primary rat lung fibroblast cultures by Northern blot analysis. Untreated confluent cultures of adult rat lung fibroblasts produced moderate steady-state levels of tropoelastin mRNA and no collagenase-3 mRNA (A, B, and C; lane 1). Conditioned medium from silicotic rat lung macrophages strongly repressed fibroblast tropoelastin mRNA levels (A; lane 2). Conditioned medium from unstimulated macrophages isolated from normal adult rat lungs had no effect on fibroblast tropoelastin mRNA levels (B; lane 2). Normal rat lung macrophages stimulated with the particulates silica (B; lane 3) or zymosan (B; lane 4) moderately and strongly repressed fibroblast tropoelastin mRNA expression, respectively. Similarly, conditioned medium from unstimulated J774 cells (C; lane 2) and silica-stimulated J774 cells (C; lane 3) had no effect, whereas conditioned medium from zymosan-stimulated J774 cells strongly repressed fibroblast tropoelastin mRNA levels. Similar results were obtained for at least three batches of conditioned medium for each treatment group. All conditions producing repression of tropoelastin mRNA expression produced a coincident induction of lung fibroblast collagenase-3 mRNA expression.

We next determined whether we could activate normal rat lung macrophages in vitro such that they would produce the tropoelastin-repressing factor. Macrophages were isolated from normal, healthy rat lungs through BAL, and were placed in culture. These cells were allowed to condition medium in the absence or presence of particulate stimulants. As shown in Figure 2B, unstimulated healthy rat lung macrophage-conditioned medium had no effect on lung fibroblast tropoelastin mRNA levels. However, macrophages stimulated in vitro with silica (0.1 mg/ml; 12.5 µg/cm²) or zymosan (4 mg/ml; 0.5 mg/cm²) elaborated factor(s) that repressed tropoelastin expression in lung fibroblasts. Zymosan, a yeast cell-wall-derived particulate that stimulates phagocytosis, was a more potent activator of production/secretion of this activity than was silica in vitro.

Additionally, we tested the ability of the monocytic J774 cell line to respond to particulates in a similar fashion. Again, unstimulated J774-cell conditioned medium had no effect on lung fibroblast tropoelastin expression, whereas particulate-activated J774 cells elaborated factor(s) that repressed tropoelastin expression (Figure 2C). The J774 cells were unresponsive to silica in vitro, as has been previously observed (20). Nonetheless, zymosan-stimulated J774-cell-conditioned medium had tropoelastin- repressing activity identical to that observed for silicotic rat lung macrophage-conditioned medium. Zymosan-stimulated J774-conditioned medium also induced expression of the gene for lung fibroblast collagenase-3 (14). These data indicate that the conditioned medium from the different macrophage sources, whether these cells were activated in rat lungs by silica or by zymosan in vitro, was functionally equivalent. Macrophage cell death was not noted in cultures, and is apparently not involved in the production/release of the observed activity.

Macrophages Coordinately Repress Tropoelastin and -SMA Expression

We chose to further explore the temporal nature of the macrophage-dependent repression of tropoelastin gene expression in lung fibroblasts. As shown in Figure 3, this repression occurred relatively slowly over time, with effects first prominent at 24 h of treatment and maximal at 48 h of treatment. Conversely, our previously observed induction of collagenase-3 gene expression occurred much more rapidly (14). We further assessed whether this alteration in tropoelastin gene expression was epigenetic or transient. Lung fibroblasts were treated with conditioned medium for 48 h and then washed and refed with normal medium for another 48 h. After removal of conditioned medium, the fibroblasts reverted to their original gene- expression profile, showing increased tropoelastin expression and no collagenase-3 expression, indicating that the effects of the activity were not phenotypically stable. We also found that repression of tropoelastin expression (like induction of collagenase-3 expression) was dependent on new protein synthesis (data not shown).

View larger version (60K): [in this window] [in a new window]   Figure 3.   Particulate-activated macrophages coordinately repress lung fibroblast tropoelastin and -SMA mRNA levels. Confluent lung fibroblast cultures were either untreated () or treated with 50% zymosan-activated J774 conditioned medium (+) and harvested after 6, 12, 24, or 48 h for Northern blot analysis. Similar fibroblast cultures were treated for 48 h with conditioned medium, and were then washed and refed with fresh conditioned medium (+/+) or nonconditioned medium (+/) and harvested after 48 h for Northern blot analysis. The repression of tropoelastin steady-state mRNA levels was barely detectable by 12 h and was maximal at 48 h. Conversely, regulation of collagenase-3 by the conditioned medium was detectable as early as 6 h after treatment, and also persisted to 48 h. When fibroblasts were treated with conditioned medium for 48 h and then refed with nonconditioned medium for 48 h, gene expression was similar to that seen in untreated cells. As was the case with tropoelastin expression, steady-state -SMA mRNA levels were repressed by the conditioned medium, but reverted to normal levels when conditioned medium was removed. Similar results were obtained in three separate experiments.

Characteristic fibrotic lung lesions are composed of -SMA-expressing "activated myofibroblasts," which accumulate matrix (11, 12). However, in silica-induced lung fibrosis, the fibrotic cells do not express -SMA (8). Mesenchymal cell (myofibroblast) tropoelastin expression appeared to be linked to -SMA expression both in our experimental rat lung silicosis model and in numerous other in vivo and in vitro situations (7, 21). Cumulatively, these data suggest that an -SMA-expressing phenotype is permissive for tropoelastin expression in lung fibroblasts. Therefore, we assessed the level of -SMA expression in our lung fibroblasts in both the absence and presence of tropoelastin-repressing macrophage-conditioned medium. As shown in Figure 3, -SMA mRNA levels were repressed by this conditioned medium, with a time-course similar to that for tropoelastin mRNA repression. Similarly, immunocytochemistry suggested a decrease in the polymerization state of -SMA in lung fibroblasts treated with conditioned medium (data not shown). This coordinate regulation of tropoelastin and -SMA in lung fibroblast cultures paralleled our in vivo observations.

Macrophage-Derived Tropoelastin Repression Occurs at the Level of Gene Transcription

In the lung, repression of tropoelastin expression by various factors has been shown to occur both at the level of gene transcription and posttranscriptionally (19). Therefore, we assessed the mechanism of tropoelastin repression effected by medium conditioned by particulate-activated macrophages. These studies made use of a previously published reverse transcription (RT)-PCR-based assay specific for the detection of tropoelastin pre-mRNA levels (16, 22). As shown in Figure 4, the tropoelastin-repressing conditioned medium substantially decreased tropoelastin gene transcription, as evidenced by decreased steady-state levels of tropoelastin pre-mRNA. Fibroblasts treated with conditioned medium contained approximately one-third the steady-state levels of tropoelastin pre-mRNA as untreated cultures. The previously characterized tropoelastin transcriptional repressors TNF- or IL-1, administered to lung fibroblast cultures individually or in combination, produced a similar level of inhibition. These data indicate that medium conditioned by particulate-activated macrophages represses tropoelastin gene expression primarily at the level of gene transcription.

View larger version (23K): [in this window] [in a new window]   Figure 4.   Repression of tropoelastin expression occurs at the level of gene transcription. Confluent rat lung fibroblast cultures were treated for 48 h with 50% zymosan-activated, macrophage-conditioned medium (CM) or with 10 ng/ml IL-1, 10 ng/ml TNF-, or both. Total RNA was isolated from the fibroblasts and the relative levels of tropoelastin pre-mRNA were determined with an RT-PCR assay, using tropoelastin-specific oligonucleotide primers as described in MATERIALS AND METHODS. All treatments resulted in a reduced level of tropoelastin steady-state pre-mRNA levels, which is indicative of a reduced rate of gene transcription. These experiments were performed in triplicate. Bars represent the mean levels (and 1 SD) of steady-state tropoelastin pre-mRNA relative to untreated controls. Inset: Autoradiogram of Southern blot for tropoelastin pre-mRNA.

Macrophage-Derived TNF- Is Necessary and Sufficient

We sought to determine the molecular nature of the factor(s) present in particulate-activated macrophage-conditioned medium that acted to repress tropoelastin gene expression. Initially, we tested the ability of J774 cells to elaborate this factor(s) and the ability of the rat lung fibroblasts to respond to it in serum-free conditions. As shown in Figure 5A, the absence of serum had no effect on the ability of the conditioned medium to repress tropoelastin and -SMA gene expression (lanes 1 and 2). Next, we manipulated the conditioned medium before presenting it to the fibroblasts. Figure 5A also shows that the repressing activity of the conditioned medium was sensitive to pepsin digestion (lane 3), but resistant to acidification (lane 4). These data strongly suggested that the activity was proteinaceous.

View larger version (51K): [in this window] [in a new window]   Figure 5.   TNF- is necessary and sufficient for macrophage-mediated repression of lung fibroblast tropoelastin gene expression. (A) Confluent rat lung fibroblast cultures were treated in serum-free conditions for 48 h with 50% zymosan-activated, macrophage-conditioned medium (lane 2) or with conditioned medium that was previously transiently acidified and pepsin-digested (lane 3), or transiently acidified alone as a control (lane 4). Pepsin digestion abrogated the repressive activity of the conditioned medium, whereas transient acidification had no effect. (B) Confluent rat lung fibroblast cultures were treated for 48 h with 50% zymosan-activated, macrophage-conditioned medium (lane 2) or with 10 ng/ml TNF- (lane 3). Both treatments repressed fibroblast tropoelastin mRNA levels. Similarly, fibroblast cultures were treated with 50% zymosan-activated, macrophage-conditioned medium that was previously treated with anti-TNF- antibody (lane 4) or with anti-IL-1 antibody (lane 5). Neutralization of TNF- in the conditioned medium abrogated its repressive activity, indicating that this cytokine is both necessary and sufficient for tropoelastin repression.

As TNF- is a known inhibitor of tropoelastin gene transcription and is present within nonelastogenic regions of silicotic granulomas in vivo, we investigated its role in our in vitro system. As was the case with silicotic rat lung granuloma-infiltrating macrophages in vivo, particulate-stimulated AM and J774 cells stained positively for TNF- (data not shown). Furthermore, we and others have previously shown that TNF- is specifically induced in monocytic J774 cells by stimulation with zymosan, but not silica, in vitro (12, 15). As seen in Figure 5B, purified recombinant TNF- showed a similar tropoelastin-repressing activity in our primary cultures of rat lung fibroblasts to that shown by the particulate-activated macrophage-conditioned medium (Figure 5B, lanes 2 and 3). We have recently shown that in this model system, collagenase-3 induction by particulate-activated magrophage-conditioned medium is due to the combined effects of TNF- and secreted arachidonate metabolites (14). Given the differences in kinetics between collagenase-3 induction and tropoelastin repression (Figure 3), it was uncertain whether a similar mechanism was involved in the present study.

Next, we performed antibody neutralization experiments on the particulate-activated macrophage-conditioned medium, in order to assess the role of TNF- and IL-1, two proinflammatory, tropoelastin-repressing cytokines. As seen in Figure 5B, TNF--neutralizing antibody at 10 µg/ml completely abrogated the tropoelastin-repressing activity of the conditioned medium (Figure 5B, lane 4). However, a similar amount of IL-1-neutralizing antibody had no effect on the ability of the conditioned medium to inhibit tropoelastin expression (Figure 5B, lane 5). Similar neutralization of the activity of the conditioned medium was achieved with a commercially available, soluble TNF- receptor (data not shown). These data indicate that TNF- is both necessary and sufficient for the macrophage-mediated repression of tropoelastin gene expression. As in the case of tropoelastin gene expression, TNF- neutralization, but not IL-1 neutralization, abrogated the -SMA- repressing activity of the conditioned medium (data not shown). These data extend the relationship between tropoelastin expression and -SMA expression to this model, with relevance to elastin regulation in adult lung disease.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We have previously reported an elastogenic response in an experimental model of lung disease (8). In this model, granulomatous inflammation was incited by a single intratracheal instillation of silica. This led to frank fibrosis as assessed by increases in lung weight as well as in DNA, total protein, and collagen content. The elastogenic response in this model was complex and highly localized. New elastin synthesis and accumulation was observed throughout the entire rodent lung, from -SMA-containing cells at sites reminiscent of those involved in elastin production during development. Although cell proliferation and collagen accumulation occurred primarily within the granulomatous lesions (13), no elastin was produced within these lesions. This lack of expression was particularly intriguing, given that each lesion probably involved numerous alveolar septae, which are sites of -SMA-containing and tropoelastin-expressing cells. In an effort to determine the mechanisms responsible for regulating tropoelastin expression in this model, we investigated the production of known inflammatory mediators and tropoelastin gene regulators in vivo. We identified abundant production and secretion of the tropoelastin repressor TNF- by infiltrating macropohages found within developing granulomas. Therefore, we assessed the ability of macrophages to respond to particulates by secreting TNF-, which could repress tropoelastin gene expression by lung fibroblasts.

Using various cell culture model systems, we found that primary AM responded to particulates both in vivo and in vitro by secreting a tropoelastin-repressing activity in lung fibroblasts. Further, the murine monocytic J774 cell line responded to particulate zymosan in vitro by producing TNF-, which is both necessary and sufficient to emulate the tropoelastin-repressing activity. Interestingly, this activity simultaneously repressed lung fibroblast -SMA expression. We have recently described a similar mechanism that is responsible for the macrophage-dependent induction of lung fibroblast collagenase-3 gene expression in this model (14). The regulation of the tropoelastin gene in this model occurs at the level of gene transcription, as evidenced by our RT-PCR data (Figure 4). Furthermore, we have previously shown that the macrophage-conditioned medium producing this effect induces fos-related AP-1 binding in rat lung fibroblasts (14), which has previously been shown to be involved in repressing tropoelastin gene transcription (23). The longer lag time needed for a maximal response in the case of tropoelastin gene regulation (Figure 3) relative to collagenase-3 induction is probably due to the long half-life of tropoelastin mRNA. Taken together, our data indicate that factors elaborated by particulate-activated macrophages can elicit characteristics representative of lung granuloma fibroblasts.

Macrophages obtained from various sites within the lung may not be identical. For instance, the macrophages isolated from silicotic rat lungs by BAL may be somewhat different from those found within the granulomatous lesions themselves. Even so, both these populations of macrophages, along with both AM isolated from normal rat lungs and the J774 cell line, responded to particulates by producing TNF-. Although conditioned medium produced by macrophages isolated from silicotic rat lungs potently repressed tropoelastin expression by lung fibroblasts, normal AM stimulated in vitro with silica were less able to do so. Stimulation of normal macrophages and the J774 cell line with zymosan, a yeast cell-wall-derived particulate, duplicated the activity released by silicotic lung macrophages and induced expression of TNF-. Indeed, it has previously been shown that the J774 cell line will respond to zymosan but not to silica in vitro by producing TNF- (20). It is likely that in vivo, granuloma macrophages phagocytose not only silica, but also cellular and extracellular matrix debris. Although other interpretations are possible, these data led us to hypothesize that an unidentified costimulator found in the silicotic lung, which is absent in our in vitro systems, is responsible for complete macrophage activation and induction of inflammatory cytokine production.

TNF- can have pleiotropic effects on various cell types. It has been reported to have either inductive or inhibitory effects on cell proliferation. In fact, particulate-activated-macrophage-conditioned medium also inhibits lung fibroblast DNA synthesis by an undefined mechanism (our unpublished data). TNF- can have diverse effects on cell actin organization and gene expression, depending on cell type and culture conditions (24, 25). However, TNF- has been reported to destabilize actin mRNA in microvascular endothelial cells (26). Our data implicate TNF- as an inhibitor of -SMA gene expression in lung fibroblasts. This regulatory mechanism may be involved in the sequelae of fibroblast alterations that occur during inflammatory lung disease.

In the developing lung, tropoelastin expression appears to be restricted to cells that also express -SMA. These include smooth-muscle cells producing elastic fibers in the media of the vasculature and conducting airways, as well as myofibroblasts producing elastic fibers of the lung interstitium (7, 27, 28). Others have previously reported the coordinate regulation of tropoelastin and -SMA by lung fibroblasts in culture (21). Furthermore, transforming growth factor-, a potent inducer of tropoelastin expression, also induces -SMA expression in lung mesenchymal cells (29). We observed previously that silicotic rat lung granuloma fibroblasts do not express -SMA (8). We hypothesized that these cells failed to participate in the elastogenic response of the silicotic lung because of their -SMA-negative phenotype. Our current data suggest that macrophages can secrete factors that act locally on lung fibroblasts to repress both tropoelastin and -SMA gene expression. These data extend the correlation between lung mesenchymal-cell -SMA gene and tropoelastin gene expression, and further suggest that these genes are coordinately regulated in lung mesenchymal cells. Although TNF- can repress -SMA expression in lung fibroblasts, and the factor responsible for -SMA repression by conditioned medium is pepsin-labile (data not shown), the coordinate regulation of -SMA and tropoelastin in these cells may be mechanistically unrelated. However, because the expression of the two genes is so closely linked in lung fibroblasts, we propose that common regulatory pathways probably control their expression. Indeed, their coregulation by TNF- supports this hypothesis.