INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: 5-lipoxygenase; bleomycin; eicosanoids; interferon-; interleukin-10

Fibrotic lung diseases are a heterogeneous group of conditions characterized by accumulation of inflammatory cells, their elaboration of proinflammatory mediators and growth factors, and expansion of the mesenchymal cell population with secretion by these cells of matrix proteins such as collagen. The resulting disruption of lung architecture impairs compliance and gas exchange. This group of diseases includes those whose etiology is attributable to a known exposure (e.g., chemotherapeutic drugs or asbestos) and those whose etiology is unknown, such as sarcoidosis and idiopathic pulmonary fibrosis (IPF). IPF is one of the most common and is certainly the best studied of the fibrotic lung diseases. Standard treatment of this disorder with corticosteroids and/or cytotoxic agents is generally disappointing, and accordingly, it tends to progress to respiratory failure within several years of diagnosis. In view of the poor outcomes and therapeutic options available in IPF and other fibrotic lung diseases, there is an urgent need for new insights into their pathobiology that can be translated into therapeutic alternatives (1).

Leukotrienes (LTs) are lipid mediators of inflammation derived from the 5-lipoxygenase (5-LO) pathway of arachidonic acid metabolism. The two classes of LTs are cysteinyl-LTs (LTs C₄, D₄, and E₄) and LTB₄. Cysteinyl-LTs are best recognized for their role in the pathogenesis of asthma. Such a role is supported by abundant data demonstrating that (1) these substances are overproduced in asthma, (2) they contribute to many of the pathophysiological features of asthma (including smooth muscle contraction, increased microvascular permeability, mucus hypersecretion, and eosinophilic inflammation), and (3) pharmacological agents that inhibit their synthesis or actions have beneficial effects in asthma (2). LTB₄, by contrast, is recognized as a potent leukocyte chemoattractant and activator (3).

We have previously reported that patients with newly diagnosed IPF exhibit overproduction of LTs in the lung. This was evidenced by demonstrating elevated levels of both LTB₄ and cysteinyl-LTs in homogenates of open lung biopsy specimens from patients, as compared with those obtained from histologically normal lung (4). Increased LTB₄ levels have also been demonstrated in the lung lavage fluid of patients with IPF (5, 6) as well as the fibrotic lung disease asbestosis (7). Although these findings are intriguing, they fail to establish a causal role for LTs in the pathogenesis of these fibrotic disorders.

Insight into the pathobiology of fibrotic lung disease has been gained from animal models. The best characterized model employs the administration to rodents of the chemotherapeutic drug, bleomycin (8, 9). We have utilized this model to investigate the potential role of LTs in the pathogenesis of pulmonary fibrosis. The involvement of these mediators was determined on the basis of comparisons between normal wild-type (WT) mice and mice rendered LT deficient by targeted deletion of the 5-LO gene (knockout [KO] mice). In this study we provide the strongest evidence to date for an important role for LTs in the evolution of pulmonary fibrosis, and elucidate some of the potential mechanisms by which LT deficiency ameliorates this process.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mice

5-LO KO (129-Alox5^tm1Fun) (10) and strain-matched WT mice were bred in the University of Michigan Unit for Laboratory Animal Medicine from breeders obtained from The Jackson Laboratory, Bar Harbor, ME. Chimeric mice generated using 129/Sv embryonic stem cells were crossed directly with wild-type 129/Sv mice, eliminating the need for further backcrossing. Mice were studied at 2-7 mo of age. Animals in different experimental groups were age matched within a given experiment, and the age of the animals did not influence the results. Separate animals were used for performance of lung lavage, lung digests, hydroxyproline determination, and histological assessments; the numbers of animals used in each experiment are detailed in the figure legends. Animal protocols were approved by the University Committee on Use and Care of Animals.

Bleomycin Injections

Mice were anesthetized and subjected to intratracheal injection with 30 µl of saline or 0.025 or 0.05 U of bleomycin (Sigma, St. Louis, MO) diluted in saline (11).

Determination of Lung Lavage Eicosanoid Levels

At various time points, animals underwent lung lavage (11). Lipids in lavage supernatants were purified using C₁₈ Sep-Pak cartridges (Waters, Milford, MA) (4). Total cysteinyl-LTs, LTB₄, and prostaglandin E₂ (PGE₂) were quantitated using specific immunoassays (Cayman Chemical, Ann Arbor, MI).

Histology and Immunohistochemical Staining

Lung sections were stained with hematoxylin and eosin for routine histology or Masson's trichrome for matrix proteins (11). Immunohistochemical staining for interferon- (IFN-) was carried out with goat anti-rat antibody (reactive with murine antigen; titer 1:100) from R&D Systems (Minneapolis, MN); nonimmune goat anti-rat serum was used as control. The secondary antibody was a biotinylated rabbit anti-goat IgG and detection was with a Vectastain alkaline phosphatase ABC kit (Vector Laboratories, Burlingame, CA).

Hydroxyproline Assays

Lung homogenates were prepared as described (11) and assayed for hydroxyproline content (9).

Collagenase Digestions of Whole Lung

Leukocytes were purified from lung digest cell suspensions by centrifugation in 40% Percoll (Sigma) (11). Cells were greater than 90% viable by trypan blue exclusion.

Differential Staining

Total numbers of monocyte/macrophages, polymorphonuclear leukocytes (PMNs), lymphocytes, and eosinophils per sample were determined from cytospins of lung digest leukocytes (11).

Flow Cytometry

Lung cells (1 × 10⁶) pooled from collagenase digests from three mice were processed for flow cytometric analysis of lymphocyte markers as described (11).

Lung Leukocyte Cultures

Mixed lung leukocytes were cultured at 5 × 10⁶ cells/ml in 24-well plates in complete RPMI medium (11). In some cases, leukocytes were adhered for 1 h to purify the macrophage population. Cells were incubated in the presence or absence of lipopolysaccharide (LPS) endotoxin (E. coli 0111.B4; Sigma) (10 µg/ml) or concanavalin A (Con A) (Sigma) (5 µg/ml). After 24 h, supernatants were collected and analyzed for cytokines.

Cytokine Assays

Levels of IFN- and interleukin-10 (IL-10) in culture supernatants or in lavage fluid were quantitated using Opti-EIA kits from Pharmingen (San Diego, CA).

Data Analysis

Data are presented as mean ± SEM. Statistical significance was determined using either Student's t test or ANOVA with a post-hoc Bonferroni test. p < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

LT Overproduction following Bleomycin Treatment

We first examined whether the bleomycin model is associated with activation of the 5-LO pathway. Levels of both LTB₄ and total cysteinyl-LTs were quantitated by immunoassays in lipid extracts of lung lavage fluid obtained from normal WT mice at various time points following intratracheal treatment with 0.05 units of bleomycin. Values were compared with baseline levels determined in untreated animals. Baseline levels of LTs in lavage fluid of untreated mice were measurable but low (Figure 1); levels of cysteinyl-LTs exceeded those of LTB₄, in keeping with the profile of LTs known to be elaborated by murine alveolar macrophages (12) but that differs from that observed in human cells (13). Following bleomycin administration, increases above baseline occurred for both LTs. A modest but nonsignificant rise in LTB₄ was observed on Day 1 postbleomycin, and the level gradually returned toward baseline by Day 7. A significant initial rise in cysteinyl-LTs was also observed at Day 1, levels declined somewhat at Day 3, and a second rise was evident at Day 7; this latter value was > 5-fold higher than the baseline level. In a separate experiment (not shown), lavage levels of the predominant cysteinyl-LTs were determined to be ~ 7-fold higher at Day 21 postbleomycin (990 ± 136 pg/ml) than under baseline conditions (146 ± 13 pg/ml; p < 0.05; n = 4). As expected, LTs were undetectable in lavage fluid from 5-LO KO mice (data not shown).

View larger version (14K): [in this window] [in a new window]   Figure 1.   Lung lavage levels of LTs in bleomycin-treated WT mice. Bleomycin (0.05 units) was administered intratracheally to WT mice on Day 0, and lung lavage fluid was obtained on the indicated days postbleomycin for quantitation of cysteinyl (cys)-LTs and LTB4. For comparison, LT levels in lavage fluid from untreated mice (U) are shown. Each data point represents the mean ± SE of four mice in a single experiment, which was representative of two others. *p < 0.05 versus untreated value.

Role of 5-LO Products in Bleomycin-Induced Fibrosis

The role of 5-LO products in the evolution of fibrosis was assessed both histologically and biochemically by comparing WT and 5-LO KO mice. Masson's trichrome staining for collagen and other matrix proteins in representative lung sections obtained at Day 21 postbleomycin (0.05 units) is presented in Figure 2. As compared with WT animals (Figure 2C), 5-LO KO mice (Figure 2F) demonstrated less intense matrix protein deposition. Similar results were seen at Day 14 (data not shown). Increases in lung collagen content with bleomycin can also be assessed by quantitating hydroxyproline levels in lung homogenates. A dose-dependent increase in lung hydroxyproline with bleomycin administration was detected in WT mice at Day 14 (Figure 3). Bleomycin at doses of 0.025 and 0.05 units/animal resulted in ~ 60% and ~ 100% increases in hydroxyproline content, relative to intratracheal saline. A doubling of lung collagen over baseline following bleomycin is typical for this model (14), and the 0.05 unit dose was used for subsequent experiments. The increase in hydroxyproline content at both doses of bleomycin was significantly less (approximately 60% less) in 5-LO KO mice than in WT mice (Figure 3). Similar results were obtained at Day 21, where a ~ 67% less hydroxyproline increment was observed following bleomycin administration in KO as compared with WT mice (data not shown).

View larger version (86K): [in this window] [in a new window]   Figure 2.   Histologic evaluation of lungs at 14 and 21 d postbleomycin in WT and 5-LO KO mice. Representative lung sections from WT (A-C ) and 5-LO KO (D-F ) mice were stained with either Masson trichrome for collagen and other matrix proteins (C and F  ) or with hematoxylin/eosin for general histopathologic assessment (A, B, D, and E ). Original magnification, ×40 for A, B, D, and E, and ×20 for C and F.

View larger version (34K): [in this window] [in a new window]   Figure 3.   Hydroxyproline content of lung homogenates in WT and 5-LO KO mice 14 d following intratracheal injection of bleomycin or saline. Hydroxyproline levels of animals administered 0.025 or 0.05 units of bleomycin were expressed as a percentage of the values determined in saline-treated animals (control). The lung hydroxyproline content of control WT animals was 218 ± 0.96 µg/lung, and of control KO animals was 221 ± 7.67 µg/lung. Each value represents the mean ± SE of five to six animals from a single experiment, which was representative of two others. *p < 0.05 versus the corresponding WT value. , WT; , 5-LO KO.

Role of 5-LO Products in Leukocyte Recruitment following Bleomycin

A comprehensive analysis of inflammation in fibrotic lung disease requires that lung leukocytes be isolated from both the alveolar space and the pulmonary interstitium. This was accomplished by a collagenase digestion technique previously described (11). In untreated animals, total lung leukocytes numbered 20.9 ± 1.82 × 10⁶ per WT mouse and 18.8 ± 0.63 × 10⁶ per KO mouse (n = 6)(p > 0.05). Also at baseline, no differences were observed between WT and KO animals in the numbers of CD8⁺ (T suppressor), CD4⁺ (T helper), CD19⁺ (B cell), or DX5⁺ (NK cell) lymphocytes, as determined by flow cytometry (data not shown).

The inflammatory response to bleomycin was quantitated by differential Diff-Quik staining of cytospins from collagenase lung digests obtained at various time points postadministration. WT animals exhibited a significant increase in PMNs and lymphocytes beginning at Day 3 and a significant increase in macrophages and eosinophils evident at Day 7 (Figure 4). By Day 7, total lung leukocytes in the WT mice had doubled over the baseline level, to 41.3 ± 6.00 × 10⁶ (p < 0.05 versus baseline). 5-LO KO mice, by comparison, exhibited dramatic reductions in the numbers of all of these cell types (Figure 4); in fact, no increases above baseline were observed in the numbers of PMNs, lymphocytes, eosinophils, or macrophages, and total lung leukocytes at Day 7 numbered 23.1 ± 1.58 × 10⁶ (p > 0.05 versus baseline). Histological evaluation of hematoxylin and eosin-stained lung sections from WT animals at Days 14 (Figure 2A) and 21 (Figure 2B) revealed extensive accumulation of inflammatory cells in alveolar spaces and lung interstitium in addition to expansion of mesenchymal cells in the alveolar walls. Confirming the quantitative data, 5-LO KO mice demonstrated less intense inflammation by histological assessment at both Days 14 (Figure 2D) and 21 (Figure 2E). These results clearly demonstrate that the 5-LO KO mice manifested a blunted inflammatory response to bleomycin.

View larger version (21K): [in this window] [in a new window]   Figure 4.   Lung leukocyte influx following bleomycin administration in WT and 5-LO KO mice. Leukocytes were separated and quantitated by differential staining from collagenase digests of lungs obtained at the time points postbleomycin indicated. For comparison, cell numbers in digests from untreated mice (U) are shown. Each value represents the mean ± SE of four to five animals from a single experiment, which was representative of two others. M/M, monocytes/macrophages; PMN, neutrophils; eos, eosinophils; lymph, lymphocytes. *p < 0.05 versus untreated value; p < 0.05 versus corresponding WT value.

Alterations in Cytokine Generation in LT-Deficient Mice

We next wished to test the possibility that the absence of LTs might influence the capacity for generation of antifibrotic cytokines. However, we were unable to detect measurable levels of immunoreactive IFN- or IL-10 in lung lavage fluid of bleomycin-treated mice of either genotype (data not shown). We therefore evaluated expression of these cytokines in situ using immunohistochemical staining of lung sections obtained from bleomycin-treated animals at Day 7. Although specific staining for IL-10 was minimal as compared with staining with negative control nonimmune serum, and not obviously different between WT and 5-LO KO animals (data not shown), we obtained highly specific staining for IFN-. As seen in Figure 5B, IFN- expression was minimal but nonetheless detectable in a patchy distribution in lungs of WT mice. By contrast, expression was diffusely greater in lungs of KO mice (Figure 5A). We next wished to further examine the capacity for cytokine generation by lung leukocytes of both genotypes. To address this, mixed leukocytes were obtained from collagenase digests of the lungs of untreated WT and 5-LO KO mice, and the cells were placed into culture and incubated for 24 h with medium alone (control), or medium containing either the T lymphocyte agonist Con A or the macrophage agonist LPS. Cytokine levels in the medium were then quantitated by immunoassays. Low levels of IFN- were elaborated by unstimulated control cultures obtained from WT animals, and levels were significantly increased in response to the addition of either Con A or LPS (Figure 6). However, levels of IFN- elaborated by cultures from KO animals were significantly greater than the WT levels under both stimulation conditions (Figure 6). IFN- has generally been regarded as an exclusive T cell product, but it has recently been recognized to also be elaborated by macrophages (17). We therefore examined in a separate experiment lung macrophages purified by adherence from the mixed leukocyte population; under basal conditions, macrophages from 5-LO KO animals generated more IFN- (259 ± 55.0 pg/ml) than did cells from WT animals (70 ± 14.1 pg/ml) (p < 0.05; n = 4). Con A-stimulated generation of IL-10 was also examined in mixed lung leukocyte cultures from WT and KO animals. Although IL-10 was undetectable in cultures from untreated WT animals, appreciable levels (932 ± 90 pg/ml; p < 0.05 versus WT; n = 4) were generated by cells from KO animals.

View larger version (95K): [in this window] [in a new window]   Figure 5.   Immunohistochemical staining for IFN- in lungs of WT and 5-LO KO mice 7 d following administration of bleomycin. Lung sections from WT (B and D) and KO (A and C) animals were stained with anti-IFN- antibody (A and B) or nonimmune serum (C and D). Original magnification, ×100.

View larger version (20K): [in this window] [in a new window]   Figure 6.   Elaboration of IFN- by lung leukocytes from WT and 5-LO KO mice. Mixed leukocytes were isolated from collagenase digests of untreated lungs from WT and KO mice; 5 × 106 cells were plated per well and cultured for 24 h with medium alone, LPS (10 µg/ml), or con A (5 µg/ml). Conditioned medium was then harvested and IFN- quantitated by immunoassay. Each value represents the mean ± SE of cultures from four separate animals. *p < 0.05 versus untreated value; p < 0.05 versus corresponding WT value. , Untreated; , con A; , LPS.

PGE₂ Levels in Lavage Fluid of WT versus 5-LO KO Mice

It was possible that in the absence of the capacity for 5-LO metabolism, arachidonic acid might be preferentially metabolized via the alternative cyclooxygenase pathway to prostanoids that possess antiinflammatory and antifibrotic activity; the best characterized of these prostanoids is PGE₂. We therefore measured PGE₂ by immunoassay in lung lavage fluid of bleomycin-treated mice of both genotypes. As shown in Figure 7, 5-LO KO mice indeed exhibited significantly greater levels of PGE₂ than did WT mice at Day 7, and their levels remained elevated through Day 21 (not shown).

View larger version (13K): [in this window] [in a new window]   Figure 7.   Lung lavage levels of PGE2 in WT and 5-LO KO mice following bleomycin administration. PGE2 was quantitated in lavage fluid obtained on the indicated days postbleomycin. For comparison, levels in lavage fluid from untreated mice (U) are shown. Each data point represents the mean ± SE of four mice in a single experiment, which was representative of two others. *p < 0.05 versus WT value.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present investigation was undertaken to critically and specifically test the hypothesis that products of the 5-LO pathway are capable of playing a causal role in the pathogenesis of fibrotic lung disease. We utilized the best studied animal model of pulmonary fibrosis, namely, intratracheal administration of bleomycin (8, 9). Previous studies utilizing the first-generation lipoxygenase inhibitor nordihydroguaiaretic acid (15) and dietary -linolenic acid (16) suggested a possible role for 5-LO products in the pathogenesis of bleomycin-induced fibrosis. However, neither of these interventions is sufficiently specific to conclusively implicate 5-LO-derived eicosanoids. For example, nordihydroguaiaretic acid inhibits arachidonate lipoxygenase enzymes other than the 5-LO enzyme, and also has generalized antioxidant actions (18). Likewise, fatty acids such as -linolenic acid may exert antiinflammatory actions by also modulating the synthesis of non-5-LO-derived eicosanoids (19) and by activating the nuclear receptor, peroxisome proliferator-activated receptor- (20). Our results with 5-LO KO mice therefore provide the strongest and most unambiguous evidence yet available that the bioactive products of this pathway can play an important causal role in the pathogenesis of fibrosis.

As an initial approach to establish the plausibility of LTs as mediators of pulmonary fibrosis in this model, we demonstrated that lung levels of these mediators do increase postbleomycin in normal WT mice. This is not a nonspecific response, as previous studies from our laboratory had established that intratracheal saline administration to these WT mice is not sufficient to increase pulmonary LT generation (21). These data represent the first demonstration of LT overproduction in bleomycin-induced fibrosis in mice. It should be noted that in the bleomycin-treated mouse lung, levels of cysteinyl-LTs exceeded levels of LTB₄, a pattern in direct contrast to the profile observed in the lungs of humans with IPF (4). However, because the lung macrophage is a key source of LT overproduction in IPF (4) and because we have previously shown that the relative generation of LTB₄ versus cysteinyl-LTs is inverted in murine (12) as compared with human (13) alveolar macrophages, we believe it likely that this difference in LT profiles represents a species phenomenon. It was also notable that bleomycin-induced generation of cysteinyl-LTs was sustained beyond the initial insult, actually reaching its maximum when fibrosis reaches its maximum at Day 21, the latest time point examined. The sustained overproduction of LTs observed in the bleomycin model is therefore similar to what has been observed in humans with IPF and asbestosis. Indeed, this may represent a fundamental feature of fibrotic lung diseases and is consistent with a role for LTs not only in the inflammatory but also in the fibrotic phase of such disorders. The local factors that drive LT generation in fibrotic pulmonary disorders have yet to be determined. However, a variety of inflammatory mediators and growth factors that are elevated in fibrotic lung disorders have been shown to amplify LT synthesis; these include IL-8 (22), endothelin-1 (23), and transforming growth factor (TGF)- (24).

Having established that overproduction of 5-LO products occurs in this model, we wished to assess their role in the evolution of the fibrotic lesion. Histological assessment indicated that KO mice exhibited less accumulation of matrix proteins such as collagen than did WT mice, and this result was confirmed by biochemical quantitation of hydroxyproline levels in lung tissue. Although protection from fibrosis was incomplete, the ~ 60% reduction in hydroxyproline levels that was observed in the LT-deficient animals compares favorably with that reported with other interventions in this model (25). These data using KO mice provide the first conclusive evidence that 5-LO products can meaningfully contribute to the pathogenesis of pulmonary fibrosis.

We next evaluated the hypothesis that protection from fibrosis in the LT-deficient animals was associated with a reduction in leukocyte recruitment to the lung. This possibility was based on the known capacity of LTs to directly promote leukocyte chemotaxis (3, 26) and inhibit leukocyte apoptosis (27, 28), as well as their indirect effects, mediated by amplification of the generation of proinflammatory substances such as IL-8 (29) and tumor necrosis factor (TNF)- (30). Indeed, pulmonary recruitment of PMNs, lymphocytes, eosinophils, and macrophages in response to bleomycin was abrogated in the 5-LO KO animals. Inasmuch as our laboratory has previously reported that LT-deficient mice demonstrated no reduction in PMN recruitment to the lung following intratracheal challenge with the pathogenic bacterium, Klebsiella pneumoniae (21), it appears that the requirement for 5-LO products in the development of a pulmonary inflammatory response is stimulus dependent. Data employing 5-LO KO mice in experimental models of inflammation in other organs such as peritoneum and skin have yielded similar conclusions (10, 31). Prior studies have suggested that the intensity of pulmonary inflammation following bleomycin administration may predict the extent of eventual pulmonary fibrosis (32). It is not possible at the present time to determine to what extent the protection from fibrosis observed in the LT-deficient animals is attributable to this attenuation of inflammation.

Fibrotic lung disease is characterized by an imbalance of cytokines that favors a fibrogenic response. For example, in both humans as well as in animal models, pulmonary fibrosis is characterized by a relative deficiency of the antifibrogenic cytokines IFN- (33) and IL-10 (34). Accordingly, we determined if the elaboration of these cytokines in lung or in lung leukocytes was regulated by 5-LO products. Indeed, we were able to detect increased expression of IFN- by immunohistochemical staining of lung sections from bleomycin-treated 5-LO KO, as compared with WT, mice. Moreover, IFN- synthesis was increased in cultures of lung leukocytes from the lungs of KO mice. On the basis of data obtained with the agonists Con A and LPS as well as with purified macrophage cultures, it is likely that both T lymphocytes and macrophages contribute to overproduction of this cytokine in the lungs of LT-deficient mice. On the other hand, T lymphocytes are apparently responsible for an increased generation of IL-10 that was observed in the KO animals. We infer that the enhanced generation of these cytokines by lung leukocytes from KO mice reflects a suppression of their synthesis by 5-LO products. To our knowledge, this is the first suggestion that 5-LO products may suppress the synthesis of the antifibrogenic cytokines IFN- and IL-10.

Administration of recombinant IFN- (35) and of IL-10 (25) are both under investigation for the treatment of IPF (1). Our data offer an alternative therapeutic approach, as pharmacological inhibition of 5-LO represents a strategy that might be expected to augment deficient lung levels of both of these antifibrogenic cytokines. Such an approach, if successful, would likely have fewer side-effects and be far less expensive than administration of recombinant cytokines.

5-LO and cyclooxygenase enzymes can compete for the metabolism of their common substrate, arachidonic acid. "Shunting" of arachidonic acid via the cyclooxygenase pathway with increased production of PGE₂ has been variably observed in cultured peritoneal macrophages from 5-LO KO mice (10, 31), and increased lung lavage LT levels have been reported in cyclooxygenase KO mice (36). We observed higher levels of PGE₂ in the lungs of 5-LO KO mice than in WT mice following intratracheal bleomycin, but this was observed only at Day 7 and later time points. The fact that such shunting was not seen in KO animals at Days 1-3, despite there being sufficient arachidonic acid mobilization to result in overproduction of LTs at those time points in WT mice, suggests that additional as yet unidentified factors must be required for enhanced PGE₂ production. PGE₂ is itself capable of inhibiting leukocyte recruitment (37), inhibiting production of proinflammatory cytokines including IL-8 (38) and TNF-a (39), inhibiting fibroblast proliferation (40) and collagen synthesis (41), and promoting the synthesis of IL-10 (42). Moreover, a deficiency of PGE₂ has been observed in patients with IPF (43, 44). It is therefore entirely possible that by such mechanisms as noted above, enhanced pulmonary synthesis of PGE₂ could contribute to the protection from bleomycin-induced inflammation and fibrosis observed in 5-LO KO mice.

To determine whether LTs play a causal role in fibrotic lung disease, we chose an interventional strategy (5-LO gene deletion) that targets both cysteinyl-LTs as well as LTB₄. This approach was selected on the basis of evidence that both classes of LTs are elevated in the bleomycin model (our data and Ziboh and coworkers [16]) as well as in human IPF (4). An inherent limitation of this approach is our inability to discern the relative contribution of cysteinyl-LTs versus LTB₄ (or, potentially, other bioactive 5-LO products) to the pathogenesis of pulmonary fibrosis. Because both classes of LTs have actions relevant to inflammation as well as fibrogenesis, we suspect that both are participants in the evolution of this disorder. Dissecting their individual and interactive contributions will require further investigations using either selective synthase gene deletions or receptor antagonists.

In summary, we have presented evidence that the murine model of bleomycin-induced pulmonary fibrosis is characterized by overproduction of LTs. Moreover, through the use of mice with a targeted deletion of the 5-LO gene that completely abrogates the capacity for LT synthesis, we have provided the first conclusive evidence for an important role for LTs in the evolution of a fibrotic lung lesion. From our data, we can further infer that LTs participate in the pathogenesis of this lesion in several ways. First, we have demonstrated that LTs are important for the inflammatory response that is generally associated with fibrosis. Second, our data suggest that LTs exert a suppressive effect on lung mononuclear cell production of antifibrogenic cytokines. Third, LTs are synthesized from arachidonic acid at the expense of potentially antiinflammatory and antifibrotic eicosanoids such as PGE₂. Finally, it is important to note that LTs have also been shown to exert direct effects on migration (45), proliferation (46), and matrix protein synthesis (47) by fibroblasts. The possible contribution of such alterations in fibroblast biology awaits further investigation, but it is likely that all of these mechanisms may be operative.

Given the critical need to identify new modes of therapy for this devastating group of fibrotic lung disorders in a timely fashion, it is attractive to consider an approach that has the potential to be rapidly translated into clinical use. Taken together, our data provide a rationale for a trial of pharmacotherapy with an inhibitor of 5-LO metabolism. This strategy offers the advantage that like the 5-LO knockout, it targets all potential mediators arising from this pathway. Zileuton, a direct 5-LO inhibitor that is already approved for the treatment of asthma (2), is a candidate agent, and is currently under investigation at our institution.