Future Research Directions in Idiopathic Pulmonary Fibrosis

MOLECULAR MECHANISMS

Fibrogenesis

Because tissue fibrosis is a prominent feature of progressive diseases in a number of other organs, important clues to the pathogenesis of pulmonary fibrosis are likely to come from analysis of shared and divergent features of tissue fibrosis in other disease models.

Myofibroblasts

In virtually every tissue, myofibroblasts are widely accepted to play a significant role in fibrosis (23, 24). Previous studies have characterized these cell types as having a phenotype intermediate between muscle and nonmuscle cells (24). Myofibroblasts are contractile, and they express members of the myogenic regulatory family (MyoD, myf5, and myogenin) and muscle structural proteins (25). A key marker for myofibroblasts is alpha-smooth muscle actin (26). Cell culture models of myofibroblasts, including hepatic stellate cells, kidney mesangial cells, cardiac valvular interstitial cells, and pulmonary myofibroblasts, all express one or more of these myogenic regulators (L. A. Leinwand, personal communication). Consistent with the concept that these myogenic regulators are functional are experiments demonstrating that a reporter construct with the skeletal myosin light chain promoter is as active in liver and kidney myofibroblasts as it is in differentiated skeletal muscle (25).

One intriguing feature of myofibroblasts is their refractoriness to the cell cycle withdrawal that normally accompanies expression and function of myogenic regulatory factors. Because myofibroblast proliferation can become pathogenic, a better understanding of the myofibroblast cell cycle would be useful.

The contractile properties of myofibroblasts are thought to be very important for their function. Contractility can be in­duced by a number of cytokines, including endothelin-1 and TGF-ß (24).

Pulmonary myofibroblasts appear to have greater contrac­tility than their kidney, liver, or cardiac counterparts (L.A. Leinwand, personal communication). The role that muscle sarcomeric proteins play in contractility of myofibroblasts is unclear. Experiments with dominant-negative mutants of sarcomeric myosin heavy chain and troponin T suggest that sarcomeric protein function is required for myofibroblast contractility. Although myofibroblasts from different organs share many features, they are clearly diverse. Future research efforts should be made to understand the signatures of myofibroblasts from different uninjured and injured organs.

Molecular Mechanisms of Hepatic Fibrosis

Significant advances in understanding hepatic fibrosis have been made in the past decade, many of which may be highly instructive for clarifying pulmonary fibrosis (27).

The liver as a model for the pathogenesis of fibrosis pro­vides several advantages that account for this progress: (1) clear identification of the cellular target of organ injury, typically the hepatocyte or epithelial compartment of liver due to toxic or viral insult, (2) identification of activated stellate cells/ myofibroblasts as the cellular source of hepatic matrix in liver injury, (3) reproducible methods to isolate and characterize hepatic stellate cells, and (4) well-validated markers to identify stellate cells in situ and track their response to liver injury.

Hepatic stellate cells are the primary source of ECM in liver fibrosis after their -activation- into hepatic myofibroblasts. These stellate cell-derived myofibroblasts bear remarkable similarity to their counterparts in lungs, including the expression of contractile proteins and responsiveness to proliferative, fibrogenic, and contractile cytokines. Activated stellate cells/myofibroblasts synthesize a large array of ECM molecules, including collagens, glycoproteins, and proteoglycans. Sinusoidal endothelial cells make a small but important contribution through their early production of cellular (EIIIA) fibronectin during liver injury. They also can activate latent TGF-ß1, which enhances fibrogenesis by stellate cells.

Activation of stellate cells is a central event in the pathogenesis of hepatic fibrosis. The fundamental features of stellate cell activation are similar regardless of the initial cause of injury (e.g., alcohol, hepatitis C). Stellate cell activation occurs in two general phases, an "initiation" or preinflammatory phase, followed by a "perpetuation" phase in which inflammatory mediators and cytokines play an important role. Release of cellular fibronectin by sinusoidal endothelial cells and production of lipid peroxides by injured hepatocytes and Kupffer cells contribute to initiation. Recent studies have defined many phenotypic features of perpetuation, including (1) proliferation caused by induction of platelet-derived growth factor receptors and upregulation of autocrine platelet-derived growth factor as well as other mitogens, including hepatocyte growth factor and insulin-like growth factor; (2) chemotaxis, also in response to platelet-derived growth factor; (3) contractility caused by upregulation of endothelin-1 production, a potent vasoconstrictor; (4) fibrogenesis, attributable to increased TGF-ß activity because of increased secretion and activation of the cytokine, combined with enhanced TGF-ß binding; (5) release of interleukin-10, colony stimulating factor, and leukocyte chemoattractants, including macrophage chemotactic peptide-1, which amplify the inflammatory response; (6) loss of retinoids caused by hydrolysis of retinyl esters; and (7) enhanced degradation of basement-membrane matrix as a result of increased secretion of matrix metalloproteinases and reduced production of tissue inhibitors of metalloproteinases. Continuing progress in understanding the molecular events driving stellate cell activation is likely to identify key pathways for further exploration in pulmonary fibrosis.

TGF-ß and Its Activation

Extensive experimental evidence has identified a transforming growth factor(s) as a central regulator of tissue fibrosis at multiple sites. With respect to pulmonary fibrosis, studies using blocking antibodies (28), chimeric TGF-ß receptors (29), and inhibitors of TGF-ß signaling have all been found to inhibit bleomycin-induced fibrosis dramatically in rodent models. The three mammalian TGF-ß isoforms, TGF-ß1, TGF-ß2, and TGF-ß3, are members of a superfamily that includes more than 40 members, all of which signal through heterodimeric transmembrane serine/threonine kinases. TGF-ß1, TGF-ß2, and TGF-ß3 all bind to the same receptors, composed of the TGF-ß type I receptor (Alk5) and the TGF-ß type II receptor, and signal by activating the cytoplasmic transcription factors SMAD2 or SMAD3. When phosphorylated by the type I receptor, these SMAD proteins bind a co-SMAD (SMAD4) and translocate to the nucleus, where they bind to TGF-ß response elements in regulatory regions of TGF-ß-responsive target genes. Recent findings make it clear that different cells can employ multiple mechanisms to modulate this canonical sig­naling pathway, including induction of inhibitory SMAD (e.g., SMAD7) proteins that compete with SMAD2/3 for interaction with activated receptors, and positive and negative regulators of SMAD-mediated transcriptional activation and repression. The importance of these modulatory signals is made clear by the dramatic differences in the patterns of gene expression induced by TGF-ß in lung fibroblasts, epithelial cells, and airway smooth muscle cells.

Although TGF-ß appears to be a potent inducer of fibrosis in the lungs and other organs, large amounts of TGF-ß protein are present in the lungs and other tissues of healthy adults without apparent TGF-ß effect. This is due, in large part, to the fact that TGF-ß isoforms are secreted as larger complexes that include TGF-ß latency-associated peptide and latent TGF-ß binding proteins (30). These latent complexes, often stored cross-linked to components of the ECM, prevent TGF- from binding to its receptors and inducing subsequent signals. One of the principal pathways for regulating TGF-ß effects is the regulation of the activity of these latent complexes. Latent complexes can be activated in vitro by extremes of temperature or pH, and by a variety of proteases that cleave the latency-associated peptide, but the mechanisms responsible for activation in vivo in health and disease have not been fully elucidated. At least three mechanisms of activation have been implicated in pulmonary fibrosis: binding to the matrix protein thrombospondin-1 (31, 32), macrophage-mediated activation by plasmin (33), and binding of latent complexes to the inducible epithelial integrin alpha v beta 6 (34). This last mechanism may be especially useful as a therapeutic target, as alpha v beta 6-mediated TGF-ß activation is not critical for developmental effects of TGF-ß but appears to have a restricted role in the enhanced activation that characterizes pulmonary and renal fibrosis.

Based on currently available evidence, it is likely that systemic global inhibition of TGF-ß would be effective in inhibiting pulmonary fibrosis but would be too toxic for long-term treatment. Targeting mechanisms of TGF-ß activation or fibrosis or pulmonary-specific downstream targets of TGF-ß might therefore be attractive alternatives. To realize the po­tential of these strategies, it will be important to identify the critical downstream targets and to develop methods to block specifically TGF-ß activation pathways (e.g., thrombospondin-1 or integrin alpha v beta 6 binding) in vivo. Testing these ideas (and even the feasibility of affecting ongoing fibrosis by blocking TGF-ß) would be greatly facilitated by the development and characterization of animal models in which pulmonary fibrosis is persistent and/or progressive.

Interrelationships between Apoptosis and Proliferation in the Development of Fibrosis

Apoptosis (i.e., programmed cell death), necrosis, and proliferation are dynamic and coexisting processes that are commonly noted in wound repair and in epithelial cells of the lung after exposure to a number of toxicants leading to fibrosis. Alterations in apoptosis and proliferation of myofibroblasts and fibroblasts have been noted in isolated cells and in lung tissues of patients with IPF, although several reports have yielded contradictory results (35, 36). Other data suggest the importance of cross-talk between epithelial cells and fibroblasts/ myofibroblasts in both the development of proliferation and apoptosis. Recent studies indicate that apoptosis of epithelial cells by bleomycin or Fas-Fas ligand cross-linking is causally related to fibrosis (37, 38). Preventive or therapeutic strategies aimed at inhibition of apoptosis in these cell types or augmentation of apoptosis in myofibroblasts may be possible. However, understanding the complex and critical pathways of proliferation and apoptosis in each cell type in situ in models of fibrosis or patient tissues is essential for targeting key regulatory pathways. More importantly, deciphering the eventual outcomes of apoptotic and proliferative responses in terms of either promotion and/or inhibition of lung inflammation and fibrosis is necessary in well-defined models of disease.

Several immunocytochemical and molecular approaches are now available to study the sequence of apoptotic and proliferative events in the lung, and these can be coupled with imaging techniques to allow colocalization of signals in key cell types (39).

Complementary approaches such as laser capture microdis­section (LCM) and reverse transcription-polymerase chain reaction (TaqMan), genomics, and proteomics might be used to decipher critical genes and signaling proteins involved in apoptosis and/or proliferation in pulmonary epithelial cells and myofibroblasts. Finally, comprehensive studies using pharmacologic inhibitors of specific signaling cascades and transgenic approaches are needed to determine whether phenotypic and functional outcomes of the fibrotic response can be modified in animal models of fibrosis.

Proteases and the ECM

The initiation and progression of lung fibrosis is determined at least in part by how anchorage-dependent lung cells interact with their surrounding ECM. In the setting of IPF, epithelial cells and fibroblasts (or myofibroblasts) contribute to and respond to provisional matrix deposition at sites of injury. Although the initiation of injury is not clear, the outcome of provisional matrix turnover significantly determines the extent of focal lung fibrosis. This is most clearly demonstrated in mice with deficiency of plasminogen activator inhibitor-1, which is protected from fibrotic responses, and is consistent with human studies showing excessive plasminogen activator inhibitor-1 in the setting of IPF (40, 41). Modifying the activation and inhibition of the set of "wound repair enzymes," which include the plasmin and thrombin systems (42), is an attractive therapeutic approach to patients with IPF. However, realization of this approach would require new tools that target successfully one or more of the key determinants of provisional matrix resolution, especially plasminogen activator inhibitor-1 and thrombin.

Although determinants of the resolution of provisional ma-trices are reasonably well-defined, how lung epithelial and fibroblastic cells respond to these matrix proteins is much less clear. Whether activation of epithelial and/or fibroblastic cells is a key initiator of remodeling and matrix modification is uncertain. A number of cytokines, proteolytic enzymes, growth factors, and their receptors are implicated in this pathobiology, but the key pathways to injury and fibrosis in IPF remain uncertain. Better understanding of how cells coordinate their signaling responses to matrix remodeling, stress, cytokines, and growth factors is probably crucial to developing new insights into progressive fibrosis. In part, this is best explored at the individual investigator level using modern cell and molecular biology approaches and appropriate animal models. Such recent studies of mouse knockouts have revealed presenilin 1 and matrix metalloproteinase 7 as previously unsuspected proteolytic enzymes involved in the pathobiology of pulmonary fibrosis (43, 44). However, application of modern methodology for high-throughput pattern recognition at the DNA and protein level to patients with IPF is also a particularly attractive approach to developing relevant new hypotheses and to better understanding coordinated responses at the whole cell level. This approach requires both an infrastructure for high-throughput testing and a support network for the ascertainment and assessment of patients and patient material. Such a network could also serve as a coordinating mechanism for future clinical trials.

Roles of TGF-ß and Connective Tissue Growth Factor in Pathological Scarring

Fibrosis, or pathological scarring, occurs in many tissues, in­cluding the peritoneal cavity, the eye, and the lung. Healing of injuries in the peritoneal cavity and eye appears to share some important common molecular and cellular aspects with healing of lung injury. For example, after an injury to the single layer of mesothelial cells that line the peritoneal cavity, a fibrin clot forms on the injured surface. If two injured areas come into contact, the surfaces adhere via a provisional fibrin matrix. If the fibrin adhesion is not lysed in a few days by plasmin, the fibroblasts in the underlying connective tissue will migrate into the fibrin matrix and begin synthesizing collagen, which converts the provisional fibrin matrix into a permanent scar (adhesion) that can cause infertility, pelvic pain, or bowel obstruction (45). Experiments in animal model of peritoneal adhesions have shown that TGF-ß is produced by macrophages and fibroblasts that populate the injured areas. More importantly, exogenous TGF-ß increases the formation of adhesions, whereas a single injection of a 20-mer antisense oligo­nucleotide targeting TGF-ß or repeated injections of neutralizing antibody to TGF-ß reduces the formation of peritoneal adhesions (46). Similar observations have been reported in animal models of glaucoma surgery to reduce intraocular pressure. In this procedure, called trabeculectomy filtering surgery, a port is created through the sclera and aqueous humor flows out of the anterior chamber into a filtering bleb in the conjunctival tissue. Exogenous TGF-ß promotes rapid failure of filtering blebs, and antisense oligonucleotides or neutralizing antibodies to TGF-ß prolong bleb survival (47). A clinical trial is currently underway in the United Kingdom that evaluates a humanized neutralizing monoclonal antibody to TGF-ß in trabeculectomy patients that are at high risk for bleb failure. Connective tissue growth factor is known to mediate TGF-ß induction of collagen synthesis in cultured fibroblasts (48). Thus, connective tissue growth factor may be a better gene to target for reducing excessive scar formation than TGF-ß. In summary, excessive scarring that leads to fibrosis in several different tissues appears to share some important common pathways, including regulation by TGF-ß and connective tissue growth factor, and agents that are designed to target these genes by selectively reducing their expression may be beneficial in preventing fibrosis from occurring. A similar approach in IPF might be beneficial.

Advances in our understanding of the basic biology of tissue injury and repair and specific advances relevant to lung injury and fibrosis suggest that epithelial injury and/or death, induction and activation of TGF-ß and/or connective tissue growth factor, and induction and activation of myofibroblasts are all likely to play critical roles in the initiation and progression of pulmonary fibrosis. It is now appropriate to begin translating these advances into improvements in treatment of this currently devastating disease. To realize the potential of these advances, it will be important to refine existing animal models and or to design new ones that will specifically allow the evaluation of treatment interventions during the sustained and/or progressive fibrotic stages of this disease. It will also be important to encourage the rapid development of drugs suitable for administration to animals and patients that will block the molecular targets outlined previously here.

Several recommendations were made that would address the issues raised related to investigating the molecular mechanisms in fibrogenesis, including the following:

ETIOLOGY AND DISEASE HETEROGENEITY

Not since the central dogma of modern biology was formulated and the genetic code elucidated has the opportunity to catapult our understanding of human disease forward been greater. This opportunity derives from timely advances in digital technology, analytical biochemistry, structural chemistry, medicinal chemistry, biostatistics, and sample acquisition, together with the sequencing of the human genome. During the session "Etiology and Disease Heterogeneity," advances in genomics, microarray analysis, and proteomics were reviewed, and the application of these technologies in identifying informative differences between biologic specimens (blood, body fluids, tissue) that are normal or diseased was highlighted. Based on available published and emerging data, a set of recommendations to use these powerful new approaches to understand and treat IPF was developed. Emphasis was given to creating easily accessed, high-quality core clinical and analytical resources and to funding mechanisms encouraging high-risk exploratory studies of etiology.

Lessons about Disease Etiology/Pathobiology from Genetic Epidemiologic Studies

Complex disorders are influenced by the actions of multiple genes and their interactions with each other and with the environment. New statistical genetic methods have been devel­oped to facilitate the search for the genes influencing susceptibility to these complex diseases (49, 50). In contrast to classic genetic methods, these new techniques do not require specification of unknown properties of the underlying genetic model, such as allele frequencies and inheritance patterns at the disease loci. Using these methods, the genetic influences on a complex trait can be quantified as heritabilities (51) or as relative risk of disease in the family members of affected individuals as compared with the general population (52). The results of such systematic family studies can then be used to design linkage studies to localize the specific genes influencing risk of disease (53, 54). Linkage methods for complex traits generally use identity-by-descent allele sharing and are based on finding regions of the genome where individuals who are phenotypically similar share more identity-by-descent alleles than expected, given their kinship. Once susceptibility loci have been localized to a particular chromosomal region, association methods can be used to assess candidate genes and polymorphisms within this region. More intensive statistical genetic analyses may also be done to assess heterogeneity, gene-gene, or gene-environment interaction.

IPF is currently at the very beginning of this process. Al­though genetic studies are currently underway with a subset of patients who appear to have simpler Mendelian forms of the disease (see subsequent passages), no systematic assessment has been made of the heritability of IPF in the general patient population or the risk to relatives of affected individuals. Such analyses are a necessary first step in determining whether genetic influences are important in general pulmonary fibrosis and in designing future linkage studies. They may also be used to investigate questions regarding the spectrum of phenotypes related to IPF (e.g., whether other types of fibrosis appear in these families or whether subclinical markers of disease susceptibility can be identified).

Experimental Approaches to Searching for Fibrotic Genes

The following complementary lines of logic strongly suggest that inherited genetic factors play a role in the development of IPF:

Although complementary genetic and pathogenic studies in animals may provide clues to the genetic determinants of IPF in humans, it is now technologically feasible to apply the recent advances in the human genome to the identification of fibrosis susceptibility genes in humans. Using standard genetic methodology (linkage analysis), one can investigate the distribution of polymorphisms for anonymous genetic markers in individuals with familial pulmonary fibrosis and then use this information to identify loci and genes that play a role in the susceptibility of IPF. Although a traditional candidate gene approach is limited by our current understanding of disease pathogenesis and is likely to pursue many unrelated genes, a

genome-wide search for linkage is more comprehensive and may identify novel genes not currently considered to be involved in the pathogenesis of IPF. Moreover, one could use these genetic findings to develop new strategies for screening and diagnosis, to identify novel therapeutic targets for this progressive life-threatening disorder, and to understand further the pathogenesis of IPF. These genetic studies provide obvious opportunities for collaboration with basic pathogenic research and expression-based research projects in interstitial lung disease.

Use of Expression Microarrays

Expression microarrays are essentially a method to perform Northern blots on multiple genes simultaneously using far less biologic material than that required by other techniques. Within the next few years, it should be feasible to perform such analyses on all genes in the human or murine genome. Despite the power of these methods, it is important to keep in mind that they only measure mRNA abundance and that study design (e.g., adequate repeats to accommodate expected biologic variability) needs to be as rigorous as that employed for any other methods that measure mRNA abundance. There are currently two widely used methods for microarray analysis: one that employs a series of short oligonucleotide probes directly synthesized onto slides by masked photolithography and the other that robotically spots individual cDNA probes on slides or filters (60). The first method has the advantage of being easy to use and commercially available in a format that provides broad coverage of human or murine genomes, but has the disadvantage of high cost. The second method can be done less expensively, but access to high-throughput robots, difficulties in cDNA clone management, and lack of specificity of long cDNAs for distinguishing closely related family members and splice variants are all disadvantages. These latter problems can probably be overcome by using spotted long synthetic oligonucleotides (e.g., 50-70 mers), but these points need to be experimentally verified.

Despite these caveats, expression microarrays have already provided some useful insights into molecular mechanisms of pulmonary fibrosis. In studies using genetically identical mice that differed only in the expression of a single integrin gene (the ß6 subunit), distinct clusters of genes were identified that were predicted to regulate lung inflammation and pulmonary fibrosis differentially (34, 61). In each case, roughly two-thirds of the genes in the cluster were already known to regulate inflammation and fibrosis, respectively, providing an additional group of genes that are candidates to regulate each of these responses. The "fibrosis" cluster included a substantial number of genes regulated by TGF-ß, providing further support for the central role of this cytokine in the pathogenesis of pulmonary fibrosis. Furthermore, careful analysis of the time course of induction of these genes identified unexpected effects of TGF-ß in the early stages of the response to injury and led to the identification of a role for this cytokine in the pulmonary edema that follows acute lung injury (62), as well as the expected role in the subsequent fibrotic response.

Application of Expression Microarrays to Tissue Samples from Patients with IPF

Analysis of a small number of such samples using informatics tools designed to evaluate the information content of expression data for each gene on the array was reported at the work-shop. Analysis of samples from five patients with well-characterized IPF suggests that there are approximately 100 informative genes that can be evaluated in more detail. At least one gene was identified for which the encoded protein has now been shown to be functionally important in bleomycin-induced fibrosis in mice. These preliminary results suggest that this technology could identify unexpected molecular participants in IPF and might help in the development of novel targets for improved treatment. The method may also allow molecular fingerprinting that could improve the ability to identify sub-classifications of pulmonary fibrosis that might be more informative than the current classification based primarily on histologic and radiographic patterns. Use of expression microarrays with microdissected samples containing specific cell types from a number of patients with a range of histologic diagnoses could be especially informative, as could analysis of samples before and after therapeutic intervention. Optimal use of this technology will require design and evaluation of a uniform set of oligonucleotide probes encompassing the entire human and mouse genome, improved methods for unbiased amplification of small amounts of input RNA, and access to appropriately sampled tissue from a series of well-characterized patients.

Using Microarrays to Elucidate the Pathobiology of Lung Disease and to Identify Molecular Targets for Drug Discovery

Microarrays have also been useful in elucidating the patterns of gene expression in the lung in health and disease, with the goal of identifying specific genes or pathways that play critical roles in defending the lung and in modulating abnormalities in disease states. An example presented at the workshop was an analysis of gene expression in the human airway epithelium in normal nonsmokers compared with 20 pack/year smokers with no evidence of lung dysfunction. The airway epithelium (10⁶ to 10⁷ cells per individual, more than 98% epithelium) recovered by fiberoptic bronchoscopy and brushing was assessed for levels of mRNA expression of 7,000 genes using the Affy­matrix human chip.

Assessment of the right and left lung in the same individual and among different individuals demonstrated that there was more variation in gene expression from individual to individ­ual than within the same individual. Gene expression among smokers was greater than that among nonsmokers, but the individual-to-individual variation was greater than that among smokers; that is, the stress of cigarette smoking was not major when assessed in a global fashion. Antioxidant genes were assessed as an example of a class of genes relevant to how the lung defends itself from the environment. Evaluation of airway epithelial expression of the superoxide dismutase genes, catalase, and the many genes involved in the glutathione pathway demonstrated only minor (1.5- to 2.5-fold) increases in smokers compared with nonsmokers in some of these genes. However, when each of the genes was assessed on an individ­ual basis, it became apparent that, for several genes, there was a subset of smokers that exhibited significant elevation of gene expression, whereas others were in the same range as nonsmokers. This observation leads to the interesting hypoth­esis that there are genetic variations in the population with regard to the ability to upregulate genes that function to protect the lung from environmental stress such as cigarette smoke; that is, those capable of upregulating these genes are protected from such stress, whereas those unable to do so are vulnerable to environmental stress-induced injury.

These observations serve as a paradigm to using array-based assessment of gene expression in humans relevant to fibrotic lung disease. For IPF, the cells that can be readily obtained for gene expression analysis by fiberoptic bronchoscopy include alveolar macrophages (a major source of mediators that stress lung parenchymal cells in the fibrotic disorders) and the airway epithelium (the cell source of bronchogenic carcinoma, known to affect 10 to 20% of individuals with IPF). For gene expression in the lung parenchyma, lung biopsy via video-assisted thorascopy, together with laser capture dissection of specific foci in the biopsy (e.g., foci of mesenchymal cells), provides a cell source of specific groups of cells within the lung parenchyma that plays a central role in the pathogenesis of the fibrotic state.

The strategy of array-based assessment of gene expression represents a technology that is readily adaptable to evaluation of the lung in health and disease and should lead to major new insights into pathogenesis and therapy of the fibrotic disorders.

LCM: Obtaining and Analyzing Tissue Samples Exhibiting Heterogeneity of Disease Expression

LCM is a powerful method for procuring pure cells from specific microscopic regions of tissue sections (63, 64) that has been under commercial development since 1997 through collaborative efforts among the National Institutes of Health (NIH) and Arcturus Engineering Inc. (http://www.arctur.com). A description of LCM, the tissue processing protocols, the facilities at the NIH and opportunities for interacting with NIH staff and use of the facilities are available through links online at http://dir.nichd.nih.gov/lcm/lcm.htm. LCM has been developed to automate and standardize microdissection and has greatly increased reproducibly and accuracy of selecting specific, targeted cells from a complex tissue for subsequent molecular analysis.

Most tissues are heterogeneous complicated structures with many different cell types locked in morphologic units exhibiting strong adhesive interactions with adjacent cells, con­nective stroma, blood vessels, glandular and muscle components, adipose cells, and inflammatory or immune cells. In normal or developing organs, specific cells express different genes and undergo complex molecular changes both in response to internal control signals, signals from adjacent cells, and humoral stimuli. In disease pathologies, the diseased cells of interest, such as precancerous cells or invading groups of cancer cells, are surrounded by these heterogeneous tissue elements. Cell types undergoing similar molecular changes, such as those thought to be most definitive of the disease progression, may constitute less than 5% of the volume of the tissue biopsy sample. Therefore, microdissection is essential to apply molecular analysis methods to study evolving disease lesions in the context of intact organs. The cDNA libraries generated from microdissected samples approximate the true pattern of gene expression of the pure cell subpopulations in their actual tissue context.

The process of LCM involves use of a special transfer film that is applied to the surface of the tissue section. Under the microscope, the thin tissue section is viewed through the glass slide on which it is mounted and microscopic clusters of cells are chosen to study. When the cells of choice are in the center of the field of view, the operator pushes a button, which acti­vates a near infrared laser diode integral with the microscope optics. The pulsed laser beam activates a precise spot on the special transfer film (either 30 or 60 micrometer in diameter), which melts and fuses with the underlying cells of choice. When the film is removed, the chosen cell(s) is tightly held within the focally expanded polymer, and the rest of the tissue is left behind. The process allows for multiple homogeneous samples within the tissue section or cytologic preparation to be targeted and pooled for extraction of molecules and analysis. In the commercial system, the process of handling the transfer film has been simplified and automated, and computer software to control the tissue microdissection process and to store data records, including digital images of the microdissected cells before and after transfer, has been developed.

This technique has proved useful in cancer research in demonstrating that genotype precedes phenotype and cancer arises in discreet tissue fields, and in discovering new genes such as tumor suppressors. Coupled with the use of protein microarray analysis and artificial intelligence, it has been possible to identify proteomics patterns that can be used to distinguish normal and precancerous tissue.

The power of LCM is the ability to obtain DNA, RNA, or protein from selected pure cells that are not damaged by the process. In the case of IPF, this technique could be very infor­mative in examining the expression of DNA, RNA, or proteins in various cells (e.g., fibroblasts, epithelial cells) in the lesion, for example, to look for molecules important in cell-cell interaction. A comparison to levels in cells in normal tissue, either from the patient lung or from nonfibrogenic lung from another source, could be used to identify those that are expressed at increased or decreased levels in fibrotic lesions.

Another encouraging new technology is the identification of proteomics signature patterns, discovery of patterns of expression of new low molecular weight molecules, and identification of serum protein patterns associated with disease. These strategies could be readily applied to IPF as new approaches to diagnosis and responses to therapy.

During the discussion of this portion of the meeting, sev­eral suggestions were made that would address the issues raised related to the etiology and disease heterogeneity in IPF, which include the following:

DIAGNOSIS AND TREATMENT

Because it appears that IPF is associated, in part, with a dysregulated reparative response of the lung resulting in the relentless accumulation of ECM, it is logical to propose that the fibroproliferative process occurring in the distal airspaces represents a promising molecular therapeutic target for treatment of IPF. Although it is likely that in some cases of IPF a genetic predisposition exists, in the majority of cases, the inciting injury remains unknown. Moreover, corticosteroid and immunosuppressive therapy, although still recommended, does not prevent ECM accumulation in the majority of affected individuals (65).

Hepatitis C virus produces both liver inflammation and fibrosis (66). Treatment with antiviral drugs and alpha-interferon can prevent and even reverse the fibrosis induced by hepatitis C virus (67). The myofibroblast (hepatic stellate cell) is thought to play a central role in the development of cirrhosis, as it does in IPF. Although lessons can be learned from hepatic fibrosis in terms of potential antifibroproliferative therapies, there are clear distinctions from IPF. The cause of IPF is unknown. Inflammation, a prominent feature of hepatitis C, is variable in IPF. The fibroblastic foci, essential features for progression in IPF, are not present in hepatic fibroblastic proliferation.

There are several approaches to the treatment of IPF. Interferon-gama, which has been shown in laboratory studies to induce myofibroblast apoptosis and to suppress the myofibro­blast phenotype, has been evaluated in a small number of patients with IPF (68). The results of a large placebo-controlled multicenter trial are being assessed.

There is evidence that TGF-ß may be central to the fibroproliferative response in skin wound healing, cirrhosis, and IPF (69-71). TGF-ß is a product of inflammatory and damaged epithelial cells. Intermittent injury to lung epithelial cells, which is presumed to occur in IPF, may lead to enhanced local levels of TGF-ß, which is capable of multiple profibrotic actions described earlier in this report. Potential approaches to anti-TGF-ß therapy include the use of neutralizing anti-bodies, soluble receptors, and decorin, a naturally occurring matrix protein capable of antagonizing TGF-ß actions (33, 34, 72). Antibodies to connective tissue growth factor, another fibroproliferative cytokine, are also being developed (68, 73, 74). Endothelin-1, in addition to its vasoactive properties, also has potent fibroproliferative potential, particularly as a myofibroblast inducer (75). If relevant, endothelin-1 receptor antag­onists, some that are approved, and others that are undergoing trials in primary pulmonary hypertension, could eventually be available for IPF treatment trials.

Another interesting approach is the development inhibitors to enzymes necessary for collagen synthesis, such as prolyl hydroxylase (76). Studies have shown that inhibition of specific enzymatic functions prevents excessive collagen accumulation in models of dermal and cardiac scarring. Studies in lung models of fibrosis are underway.

Data suggest that arachidonic acid metabolites can act to increase or decrease pulmonary fibrosis. Prostaglandin E₂ in­hibits fibroblast proliferation (77, 78) and fibroblast collagen synthesis (79, 80). Its level in pulmonary fibrosis patients' lungs is reduced (81), and lung fibroblasts from pulmonary fibrosis patients are less able to produce it in vitro (81). In addition, levels of leukotriene B₄ are higher in pulmonary fibrosis patients' lungs (82), and elimination of leukotriene production has been shown to protect against pulmonary fibrosis (83). Thus, inhibition or stimulation of the appropriate metabolite is a possible point for intervening in this disease (84).

A different approach to IPF therapy includes strategies to promote matrix resorption by enhancing the activity of certain matrix metalloproteinases, the interstitial collagenases, whose levels are reduced in IPF (36, 85). This could be accomplished by augmenting their activity either endogenously (e.g., interferon-gama) or exogenously by the administration of supplemental enzymes. There is an imbalance between the natural inhib­itors of the matrix metalloproteinases (tissue inhibitors of metalloproteinases) and the matrix metalloproteinases favor­ing matrix accumulation rather than degradation in IPF. Tissue inhibitors of metalloproteinases are antagonized by anti-TGF-ß therapeutic strategies, thereby increasing collagenolytic activity. The development of antagonists to the tissue inhibitors of metalloproteinases should be considered.

There are approaches to identifying therapeutic targets and therapeutics that provide a conceptual framework that could be applied to IPF. The molecular target must be accessible to chemical genetics using chemistry, structural biology, and library screening techniques to design small agonists or antagonists for the molecular target. There must be in vitro and in vivo assay systems for evaluating the effectiveness of potential therapeutics. The potential toxicity, pharmacokinetics, and pharmacodynamics of potential therapeutics should be con­sidered at the beginning of the process of discovery.

Within the area of IPF, it should be possible to identify potential therapeutic targets that, if stimulated or blocked by bioprobes, could interfere with or reverse fibrogenesis. In or­der to develop such bioprobes, it will be necessary to establish cell-free and cell-based assays for testing their agonistic or antagonistic behavior. By identifying libraries of privileged structures that have high potential to be used as nontoxic bioprobes in vitro and in vivo, a wide range of potential therapeutics can be screened in the appropriate animal models in which their effect on at least one aspect of the disease can be addressed. Collaborations between clinical investigators and molecular targeting programs (academic and industrial) would facilitate development.

During the discussion, several suggestions were made that would address the issues raised related to diagnosis and treat­ment in IPF, including the following:

WORKSHOP RECOMMENDATIONS

After discussion of the suggestions that arose during the meet­ing, the following overall recommendations of the Working Group to the NHLBI were formulated:

References

1. American Thoracic Society. Idiopathic pulmonary fibrosis: diagnosis and treatment. International consensus statement. Am J Respir Crit Care Med 2000;161:646-664.

2. Coultas DB, Zumwalt RE, Black WC, Sobonya RE. The epidemiology of interstitial lung diseases. Am J Respir Crit Care Med 1994;150:967-972.

3. Coultas DB, Hughes MP. Accuracy of mortality data for interstitial lung diseases in New Mexico, USA. Thorax 1996;51:717-720.

4. Bjoraker JA, Ryu JH, Edwin MK, Myers JL, Tazelaar HD, Schroeder DR, Offord KP. Prognostic significance of histopathologic subsets in idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 1998;157:199-203.

5. Nicholson AG, Colby TV, Dubois RM, Hansell DM, Wells AU. The prognostic significance of the histologic pattern of interstitial pneumonia in patients presenting with the clinical entity of cryptogenic fibrosing alveolitis. Am J Respir Crit Care Med 2000;162:2213-2217.

6. Douglas WW, Ryu JH, Schroeder DR. Idiopathic pulmonary fibrosis: impact of oxygen and colchicine, prednisone, or no therapy on survival. Am J Respir Crit Care Med 2000;161:1172-1178.

7. Nicholson AG, Colby TV, Dubois RM, Hansell DM, Wells AU. The prognostic significance of the histologic pattern of interstitial pneumonia in patients presenting with the clinical entity of cryptogenic fibrosing alveolitis. Am J Respir Crit Care Med 2000;162:2213-2217.

8. Katzenstein AL, Myers JC. Idiopathic pulmonary fibrosis: clinical relevance of pathologic classification. Am J Respir Crit Care Med 1998; 157:1301-1315.

9. Travis WD, King TE Jr, Bateman ED, Lynch DA, Capron F, Center D, Colby TV, Cordier JF, du Bois RM, Galvin J, et al. American Thoracic Society/European Respiratory Society international multidisciplinary consensus classification of idiopathic interstitial pneumonias: general principles and recommendations. Am J Respir Crit Care Med 2002;165:277-304.

10. Kuhn C III, Boldt J, King TE Jr, Crouch E, Vartio T, McDonald JA. An immunohistochemical study of architectural remodeling and connective tissue synthesis in pulmonary fibrosis. Am Rev Respir Dis 1989; 140:1693-1703.

11. Kuhn C III. Patterns of lung repair: a morphologists view. Chest 1991; 99:11S-14S.

12. Kuhn C III, McDonald JA. The roles of the myofibroblast in idiopathic pulmonary fibrosis: ultrastructural and immunohistochemical features of sites of active extracellular matrix synthesis. Am J Pathol 1991;138: 1257-1265.

13. Selman M, King TE Jr, Pardo A. Idiopathic pulmonary fibrosis: prevailing and evolving hypotheses about its pathogenesis and implications for therapy. Ann Intern Med 2001;134:136-151.

14. Uhal BD, Joshi I, True AL, Mundle S, Raza A, Pardo A, Selman M. Fibroblasts isolated after fibrotic lung injury induce apoptosis of alveolar epithelial cells in vitro. Am J Physiol 1995;269:L819-L828.

15. Uhal BD, Joshi I, Hughes WF, Ramos C, Pardo A, Selman M. Alveolar epithelial cell death adjacent to underlying myofibroblasts in advanced fibrotic human lung. Am J Physiol 1998;275:L1192-L1199.

16. Leslie KO, Mitchell J, Low R. Lung myofibroblasts. Cell Motil Cytoskel­eton 1992;22:92-98.

17. Izumi T. Proceedings of the international congress on bronchiolitis obliterans organizing pneumonia. Chest 1992;102(Suppl.):1S-50S.

18. Peyrol S, Cordier JF, Grimaud JA. Intraalveolar fibrosis of idiopathic bronchiolitis obliterans-organizing pneumonia: cell-matrix patterns. Am J Pathol 1990;137:155-170.

19. Fukuda Y, Ishizaki M, Kudoh S, Kitaichi M, Yamanaka N. Localization of matrix metalloproteinases-1, -2, and -9 and tissue inhibitor of metal-loproteinase-2 in interstitial lung diseases. Lab Invest 1998;78:687-698.

20. Lappi-Blanco E, Soini Y, Pääkkö P. Apoptotic activity is increased in the newly formed fibromyxoid connective tissue in bronchiolitis obliterans organizing pneumonia. Lung 1999;177:367-376.

21. Selman M, Ruiz V, Cabrera S, Segura L, Ramirez R, Barrios R, Pardo A. TIMP-1, -2, -3, and -4 in idiopathic pulmonary fibrosis: a prevailing nondegradative lung microenvironment? Am J Physiol Lung Cell Mol Physiol 2000;279:L562-L574.

22. Hayashi T, Stetler-Stevenson WG, Fleming MV, Fishback N, Koss MN, Liotta LA, Ferrans VJ, Travis WD. Immunohistochemical study of metalloproteinases and their tissue inhibitors in the lungs of patients with diffuse alveolar damage and idiopathic pulmonary fibrosis. Am J Pathol 1996;149:1241-1256.

23. Brenner DA, Waterboer T, Choi SK, Lindquist JN, Stefanovic B, Burchardt E, Yamauchi M, Gillan A, Rippe RA. New aspects of hepatic fibrosis. J Hepatol 2000;32(Suppl 1):32-38.

24. Powell DW, Mifflin RC, Valentich JD, Crowe SE, Saada JI, West AB. Myofibroblasts. I. Paracrine cells important in health and disease. Am J Physiol 1999;277:C1-C9.

25. Mayer G, Leinwand LA. Sarcomeric gene expression and contractility in myofibroblasts. J Cell Biol 1997;139:1477-1484.

26. Roy SG, Nozaki Y, Phan SH. Regulation of alpha smooth muscle actin gene expression in myofibroblast differentiation from rat lung fibroblasts. Int J Biochem Cell Biol 2001;33:723-734.

27. Friedman SL. Molecular regulation of hepatic fibrosis: an integrated cellular response to tissue injury. J Biol Chem 2000;275:2247-2250.

28. Giri SN, Hyde DM, Hollinger MA. Effect of antibody to transforming growth factor beta on bleomycin induced accumulation of lung collagen in mice. Thorax 1993;48:959-966.

29. Wang Q, Wang Y, Hyde DM, Gotwals PJ, Koteliansky VE, Ryan ST, Giri SN. Reduction of bleomycin induced lung fibrosis by transforming growth factor beta soluble receptor in hamsters. Thorax 1999;54: 805-812.

30. Munger JS, Harpel JG, Gleizes P, Mazzieri R, Nunes I, Rifkin DB. La-tent transforming growth factor-beta: structural features and mechanisms of activation. Kidney Int 1997;51:1376-1382.

31. Schultz-Cherry S, Murphy-Ullrich JE. Thrombospondin causes activation of latent transforming growth factor-beta secreted by endothelial cells by a novel mechanism. J Cell Biol 1993;122:923-932.

32. Crawford SE, Stellmach V, Murphy-Ullrich JE, Ribeiro SM, Lawler J, Hynes RO, Boivin GP, Bouck N. Thrombospondin-1 is a major acti­vator of TGF-beta1 in vivo. Cell 1998;93:1159-1170.

33. Khalil N, Corne S, Whitman C, Yacyshyn H. Plasmin regulates the activation of cell-associated latent TGF-beta 1 secreted by rat alveolar macrophages after in vivo bleomycin injury. Am J Respir Cell Mol Biol 1996;15:252-259.

34. Munger JS, Huang XZ, Kawakatsu H, Griffiths MJD, Dalton SL, Wu JF, Pittet JF, Kaminski N, Garat C, Matthay MA, et al. The integrin alpha v beta 6 binds and activates latent TGF-1: a mechanism for regulating pulmonary inflammation and fibrosis. Cell 1999;96:319-328.

35. Raghu G, Chen Y, Rusch V, Rabinovitch PS. Differential proliferation of fibroblasts cultured from normal and fibrotic human lung. Am Rev Respir Dis 1988;138:703-708.

36. Ramos C, Montano M, Garcia-Alvarez J, Ruiz V, Uhal BD, Ramos M, Pardo A. Fibroblasts from idiopathic pulmonary fibrosis and normal lungs differ in growth rate, apoptosis and tissue inhibitor of metallo­proteinases expression. Am J Respir Cell Mol Biol 2001;24:591-598.

37. Hagimoto N, Kuwano K, Nomoto Y, Kunitake R, Hara N. Apoptosis and expression of Fas/Fas ligand mRNA in bleomycin-induced pulmonary fibrosis in mice. Am J Respir Cell Mol Biol 1997;16:91-101.

38. Hagimoto N, Kuwano K, Miyazaki H, Kunitake R, Fujita M, Kawasaki M, Kaneko Y, Hara N. Induction of apoptosis and pulmonary fibrosis in mice in response to ligation of Fas antigen. Am J Respir Cell Mol Biol 1997;17:272-278.

39. Taatjes D, Palmer CJ, Pantano C, Buder-Hoffmann S, Cummins A, Moss-man BT. Laser-based microscopic approaches: application to cell signaling in environmental lung disease. Biotechniques 2001;31:880-894.

40. Eitzman DT, McCoy RD, Zheng X, Fay WP, Shen T, Ginsburg D, Si­mon RH. Bleomycin-induced pulmonary fibrosis in transgenic mice that either lack or overexpress the murine plasminogen activator inhibitor-1 gene. J Clin Invest 1996;97:232-237.

41. Chapman HA Jr, Allen CL, Stone OL. Abnormalities in pathways of alveolar fibrin turnover among patients with interstitial lung disease. Am Rev Respir Dis 1986;333:437-443.

42. Bernard GR, Vincent JL, Laterre PF, LaRosa SP, Dhainaut JF, Lopez-Rodriguez A, Steingrub JS, Garber GE, Helterbrand JD, Ely EW, et al. Recombinant human protein C worldwide evaluation in severe sepsis (PROWESS) study group: efficacy and safety of recombinant human activated protein C for severe sepsis. N Engl J Med 2001;344:699-709.

43. Powell WC, Fingleton B, Wilson CL, Boothby M, Matrisian LM. The metalloproteinase matrilysin proteolytically generates active soluble F ligand and potentiates epithelial cell apoptosis. Curr Biol 1999;9:1441-1447.

44. Herreman A, Hartmann D, Annaert W, Saftig P, Craessaerts K, Ser­neels L, Umans L, Schrijvers V, Checler F, Vanderstichele H, et al. Presenilin 2 deficiency causes a mild pulmonary phenotype and no changes in amyloid precursor protein processing but enhances the em­bryonic lethal phenotype of presenilin 1 deficiency. Proc Natl Acad Sci USA 1999;96:11872-11877.

45. Schultz GS. Comparative tissue repair. In: diZerega GS, editor. Proceed­ing of the 3rd international congress on pelvic surgery and adhesion formation. New York: Springer Verlag; 1997. p. 26-36.

46. Williams RS, Schultz GS, Chegini N, Rossi AM. Effect of transforming growth factor beta on post operative adhesions formation and intact peritoneum. J Surg Res 1992;52:65-70.

47. Cordeiro MF, Gay JA, Khaw PT. Human anti-transforming growth factor-beta2 antibody: a new glaucoma anti-scarring agent. Invest Ophthalmol Vis Sci 1999;40:2225-2235.

48. Duncan MR, Frazier KS, Abramson S, Williams S, Klapper H, Huang X, Grotendorst GR. Connective tissue growth factor mediates trans-forming growth factor beta-induced collagen synthesis: down-regula­tion by cAMP. FASEB J 1999;13:1774-1786.

49. Almasy L, Blangero J. Linkage strategies for mapping genes for complex traits in man. In: Crusio WE, Gerlai RT, editors. Handbook of molecular-genetic techniques for brain and behavior research. Amsterdam: Elsevier Science; 1999. p. 100-112.

50. Terwilliger JD, Goring HHH. Gene mapping in the 20th and 21st centuries: statistical methods, data analysis, and experimental design. Hum Biol 2000;72:63-132.

51. Falconer DS. Introduction to quantitative genetics, 3rd ed. New York: John Wiley & Sons; 1989.

52. James JW. Frequency in relatives of an all-or-none trait. Ann Hum Genet 1971;35:47-49.

53. Risch N. Linkage strategies for genetically complex traits. II. The power of affected relative pairs. Am J Hum Genet 1990;46:229-241.

54. Blangero J, Williams JT, Almasy L. Quantitative trait locus mapping using human pedigrees. Hum Biol 2000;72:35-62.

55. Rosenberg DM. Inherited forms of interstitial lung disease. Clin Chest Med 1982;3:635-641.

56. Raghu G, Mageto YN. Genetic predisposition of interstitial lung disease. In: Schwarz MI, King TE Jr, editors. Interstitial lung disease. Hamil­ton, ON: B.C. Decker; 1998.

57. Bitterman PB, Rennard SI, Keogh BA, Wewers MD, Adelberg S, Crys­tal RG. Familial idiopathic pulmonary fibrosis: evidence of lung inflammation in unaffected family members. N Engl J Med 1986;314: 1343-1347.

58. Fujita J, Yoshinouchi T, Ohtsuki Y, Tokuda M, Yang Y, Yamadori I, Bandoh S, Ishida T, Takahara J, Ueda R. Non-specific interstitial pneumonia as pulmonary involvement of systemic sclerosis. Ann Rheum Dis 2001;60:281-283.

59. White B, Moore WC, Wigley FM, Xiao HQ, Wise RA. Cyclophosphamide is associated with pulmonary function and survival benefit in patients with scleroderma and alveolitis. Ann Intern Med 2000;132:947-954.

60. Albelda S, Sheppard D. Functional genomics and expression profiling: be there or be square. Am J Respir Cell Mol Biol 2000;23:265-269.

61. Kaminski N, Allard J, Pittet J-F, Zuo F, Griffiths MJD, Morris D, Huang XZ, Sheppard D, Heller RA. Global analysis of gene expression in pulmonary fibrosis reveals distinct programs regulating lung inflammation and remodeling. Proc Natl Acad Sci USA 2000;97:1778-1783.

62. Pittet J-F, Griffiths MJD, Geiser T, Kaminski N, Dalton SL, Huang X, Brown LAS, Gotwals PJ, Koteliansky VE, Matthay MA, et al. TGF-beta is a critical mediator of acute lung injury. J Clin Invest 2001;107:1529-1536.

63. Emmert-Buck MR, Bonner RF, Smith PD, Chuaqui RF, Zhuang Z, Goldstein SR, Weiss RA, Liotta LA. Laser capture microdissection. Science 1996;274:998-1001.

64. Bonner RF, Emmert-Buck M, Cole K, Pohida T, Chuaqui R, Goldstein S, Liotta LA. Cell sampling: laser capture microdissection: molecular analysis of tissue. Science 1997;278:1481-1483.

65. King TE, Costabel U, Cordier JR. Idiopathic pulmonary fibrosis: diagnosis and management: international consensus statement. Am J Respir Crit Care Med 2000;161:646-664.

66. Albanis E, Friedman SL. Hepatic fibrosis: pathogenesis and principles of therapy. Clin Liver Dis 2001;5:315-334.

67. Manabe N, Chevallier M, Chossegros P, Causse X, Guerret S, Trepo C, Grimaud JA. Interferon-alpha 2b therapy reduces liver fibrosis in chronic non-A, non-B hepatitis: a quantitative histological evaluation. Hepatology 1993;18:1344-1349.

68. Ziesche R, Hofbauer E, Wittmann K, Petkov V, Block LH. A prelimi­nary study of long-term treatment with interferon gamma-1b and low-dose prednisolone in patients with idiopathic pulmonary fibrosis. N Engl J Med 1999;241:1264-1269.

69. Broekelmann TJ, Limper AH, Colby TV. Transforming growth factor beta 1 is present at sites of extracellular matrix gene expression in human pulmonary fibrosis. Proc Natl Acad Sci USA 1991;88:6642-6646.

70. Gauldie J, Sime PJ, Xing Z, Marr B, Tremblay GM. Transforming growth factor-beta gene transfer to the lung induces myofibroblast presence and pulmonary fibrosis. Curr Top Pathol 1999;93:35-45.

71. Sime PJ, Xing Z, Graham FL, Csaky KG, Gauldie J. Adenovector-medi­ated gene transfer of active transforming growth factor-beta 1 induces prolonged severe fibrosis in rat lung. J Clin Invest 1997;100:768-776.

72. Zhao J, Sime PJ, Bringas P Jr, Gauldie J, Warburton D. Adenovirus­mediated decorin gene transfer prevents TGF-beta induced inhibition of lung morphogenesis. Am J Physiol 1999;277:L412-L422.

73. Lasky JA, Ortiz LA, Tothat B, Hoyle GW, Corti M, Athas G, Lun­garella G, Brody A, Friedman M. CTGF mRNA is upregulated in ble­omycin-induced lung fibrosis. Am J Physiol 1998;275:365-371.

74. Shi-Wen X, Pennington D, Holmes A, Leask A, Bradham D, Beau-champ JR, Fonseca C, du Bois RM, Martin GR, Black CM, et al. Au­tocrine overexpression of CTGF maintains fibrosis. Exp Cell Res 2000;259:213-224.

75. Teder P, Noble PW. A cytokine reborn? Endothelin-1 in pulmonary in­flammation and fibrosis. Am J Respir Cell Mol Biol 2000;23:7-10.

76. Franklin TJ, Morris WP, Edwards PN, Large MS, Stephenson R. Inhibi­tion of prolyl-4-hydroxylase in vitro and in vivo by members of a novel series of phenanthrolinones. Biochem J 2001;353:333-338.

77. Elias J. Tumor necrosis factor interacts with interleukin-1 and interfer­ons to inhibit fibroblast proliferation via prostaglandin-dependent and independent mechanisms. Am Rev Respir Dis 1988;138:652-658.

78. Bitterman P, Wewers N, Rennard S, Adelberg S, Crystal R. Modulation of alveolar macrophage-driven fibroblast proliferation by alternative macrophage mediators. J Clin Invest 1986;77:700-708.

79. Korn J, Halushka P, Leroy E. Mononuclear cell modulation of connec­tive tissue function: suppression of fibroblast growth by stimulation of endogenous prostaglandin production. J Clin Invest 1980;65:543-554.

80. Goldstein R, Polgar P. The effect and interaction of bradykinin and pros­taglandins on protein and collagen production by lung fibroblasts. J Biol Chem 1982;257:630.

81. Wilborn J, Crofford L, Burdick M, Kunkel S, Strieter R, Peters-Golden M. Cultured lung fibroblasts isolated from patients with idiopathic pulmonary fibrosis have a diminished capacity to synthesize _{prostaglandin E2} and to express cyclooxygenase-2. J Clin Invest 1995; 95:1861.

82. Wilborn J, Bailie M, Coffey M, Burdick M, Strieter R, Peters-Golden M. Constitutive activation of 5-lipoxygenase in the lungs of patients with idiopathic pulmonary fibrosis. J Clin Invest 1996;97:1827-1836.

83. Peters-Golden M, Bailie M, Marshall T, Wilke C, Phan SH, Toews GB, Moore BB. Protection from pulmonary fibrosis in leukotriene-deficient mice. Am J Respir Crit Care Med 2002;165:229-235.

84. Zeldin DC. The 5-lipoxygenase pathway: a new therapeutic target for the treatment of pulmonary fibrosis. Am J Respir Crit Care Med 2002;165: 146-147.

85. Lemjabbar H, Gosset P, Lechapt-Zalcman E, Franco-Montoya ML, Wal­laert B, Harf A, Lafuma C. Overexpression of alveolar macrophage gelatinase B (MMP-9) in patients with idiopathic pulmonary fibrosis: effects of steroid and immunosuppressive treatment. Am J Respir Cell Mol Biol 1999;20:903-913.

Correspondence and requests for reprints should be addressed to Robert A. Musson, Room 7126, MSC 7922, 6701 Rockledge Drive, Bethesda, MD 20892. E-mail: rmusson@nih.gov