Pathogenesis of Fibrosis in Acute Lung Injury PETERB.
BITrERMAN, M.D., Minneapolis, Minnesota
The anatomic changes that occur in response to acute lung injury significantly impair gas exchange. As is the case with skin wounds, a fibroproliferative response follows lung injury. In the lungs, however, this can result in lifethreatening obliteration of alveolar air spaces. A better understanding of the mechanisms involved in lung repair may allow the development of therapies that regulate the fibroproliferative response. Studies from our laboratory have identified a peptide in bronchoalveolar lavage fluid from patients with acute lung injury that promotes the migration and replication of lung fibroblasts. This peptide is related to platelet-derived growth factor (PDGF) antigenically as well as by receptor-binding criteria; its molecular weight is 14 kilodaltons (kDa) as compared to 29 kDa for PDGF. Despite the potent activity of the 14 kDa peptide, however, such a growth signal may not be absolutely required for tissue granulation. The possibility that lung fibroblasts from patients with acute lung injury might be capable of dividing without exogenous stimulation will be examined. Another theoretical consideration is the signals that regulate termination of the fibroproliferative response. Insights into the molecular mechanisms involved in lung repair may result in therapies that modulate the sometimes maladaptive fibroproliferative response following acute lung injury.
From the Department of Pulmonaryand Critical Care Medicine, University of Minnesota Medical Center, Minneapolis, Minnesota. Requests for reprints should be addressedto Peter B. Bitterman, M.D., University of Minnesota Medical Center, 420 Delaware Street, SE, Minneapolis, Minnesota 55455.
cute lung injury is a dramatic and common clinical disorder, with an estimated incidence of 100,000-200,000 cases per year in the United States . Acute lung injury is defined as a rapid alteration of the alveoli that results in impairment of gas exchange. The most severe form of acute lung injury is the adult respiratory distress syndrome (ARDS). Despite vast improvements in supportive care for patients with acute lung injury, the mortality rate for ARDS has hovered in the 50% range for the past decade . Recent studies have suggested that once a patient has survived the initial injury, subsequent gas exchange problems may arise in part from an inadequately regulated healing response . A key part of integumentary wound repair is a fibroproliferative response, which rapidly fills the wound with new capillaries, mesenchymal cells, and their connective tissue products. When this process takes place in the lungs following injury, it can sometimes result in the obliteration of alveolar air spaces. A more complete understanding of the fibroproliferative response may lead to the development of therapies that are capable of modulating lung repair.
PHYSIOLOGY OF ACUTE LUNG INJURY Acute lung injury can be likened to a biologic explosion in the lung; the outcome is catastrophic damage to the alveolar wall. In a normal alveolus (Figure 1, left), type I epithelial cells facilitate the exchange of gas across their thin cytoplasm while, at the same time, preventing the flux of interstitial fluid into the air space . Acute lung injury results in a severe disruption of this epithelial layer (Figure 1, right), thus removing the major barrier to the flow of interstitial fluid and allowing the air space to flood . This pulmonary edema results in gas exchange impairment and microanatomic shunting of blood. In addition, the endothelial cells lining the microcirculation begin to die, exposing a bare endothelial basement membrane. This permits the products of coagulation to accumulate, leading to microthrombi and impaired microcirculation. The interstitium, located between the epithelial and endothelial basement membranes, swells with fluid. All of these events often occur just as the patient begins to complain of breathlessness. Early in the course of acute lung injury, severe inflammation is observed. The epithelial disruption
June 22, 1992 The American Journal of Medicine Volume 92 (suppl 6A)
SYMPOSIUM ON ANTIMICROBIAL THERAPY / BITrERMAN
allows the products of inflammation to enter the response that fills the alveolar air spaces and may alveolar air space, further compromising the air- ultimately result in death. Our understanding of lung interface. In bronchoalveolar lavage fluid ob- the molecular basis for these responses is still in its tained from patients at this point in the disease pro- infancy. Nevertheless, as knowledge of the signals cess, red cells outnumber white cells by a ratio of that influence the fibroproliferative response exapproximately 10 to 1. Almost all of the white cells pands, the development of therapies based on conpresent are neutrophils, a type of cell that is only trolling this poorly regulated healing response berarely found in the noninjured lung. Mononuclear comes an increasingly promising option. phagocytes also enter the lung, followed by a marked accumulation of mesenchymal cells in the LUNG REPAIR The events of lung repair that take place followalveolar air space. Because of the explosive nature of the transmu- ing lung injury occur in each of the three major alral injury to the alveolar wall, anti-inflammatory veolar compartments: the interstitium, the airtherapies are likely to be of little use unless pa- lung interface, and the blood-lung interface . tients at risk for developing ARDS can be identified When all three of these compartments have been in advance. Even when at-risk patients are identifi- properly repaired, the permeability defect that ocable, anti-inflammatory intervention may be of only curs in response to acute lung injury is corrected limited use, as the lungs have efficient defense and normal gas exchange can resume. Conversely, mechanisms in place to cope with the inflammatory a lack of effective repair in these compartments results in continued impairment of gas exchange. response . An alternative approach to therapeutic interven- Morphologic lung analyses have revealed that most tion is to influence mechanisms involved in the fl- cases of ineffective repair are associated with exbroproliferative response that takes place during cessive accumulation of fibrotic tissue in the air lung repair. For patients with acute lung injury, space. survival depends on the rapid and orderly reconstruction of the gas exchange apparatus. In some Effective Repair Normal repair of the air-lung interface and interpatients, this process occurs without any complications, resulting in a complete recovery. Other pa- stitium involves the re-epithelialization of the lung, tients experience an extensive fibroproliferative primarily by type II epithelial cells. The two major
Figure 1. Cellular anatomy of acute lung injury. A normal alveolus (left) and an alveolus following acute lung injury (right). EP] = epithelial cell type 1; A = alveolar air space; C = capillary; BM = basement membrane. Reprinted with permission from . 6A-40S
June 22, 1992 The American Journal of Medicine Volume 92 (suppl 6A)
SYMPOSIUM ON ANTIMICROBIALTHERAPY/ BITrERMAN
functions of these cells are to release surfactant and to serve as progenitors of type I epithelial cells, which constitute the normal mediators of gas exchange. Repair of the interstitium also involves restoration of the extracellular matrix, during which connective tissues are deposited in an ordered fashion and the replication of fibroblasts is tightly regulated. These events result in restoration of the physical attributes of the alveolar wall, a prerequisite to normal gas exchange. Repair of the microcirculation of patients with acute lung injury is also critical. Damage to the blood-lung interface results in microvascular coagulation and an influx of proinflammatory molecules into the lung. Repair of the blood-lung interface requires the recanalization and re-endotheliatization of the microcirculatory apparatus.
Maladaptive Repair The complicated repair process involved in healing the lung sometimes goes awry, as is reflected in the 50% mortality rate of patients with ARDS. The characteristic lesion observed in patients who die from ARDS is granulation of the alveolar air space. During the fibroproliferative response, myofibroblasts enter the air spaces by migrating through breaks in the epithelial basement membrane. The air spaces subsequently become filled by replicating mesenchymal cells and developing capillary networks. This intra-alveolar granulation tissue produces a fibrotic lung that is deficient in gas exchange. The clinical condition of patients with granulated air spaces is marked by hypoxemia associated with severe shunt physiology. Ironically, although this fibroproliferative response can impair function when it occurs in the lungs and other organs, it is the same response that prevents most surface wounds from becoming life threatening. Indeed, the ability to heal surface wounds is a highly conserved biologic response that is present throughout the plant and animal kingdoms. When a fibroproliferative response occurs after a cut on the skin, it is an appropriate and necessary means by which to restore the barrier between the organism and the sometimes inhospitable environment. However, the same response in the lung can lead to the presence of cells and capillaries in air spaces that must remain free of such tissue in order for gas exchange to occur. A maladaptive fibroproliferative response can also cause problems in other organs. A survey by the Department of Health and Human Services  revealed the astonishing fact that approximately 45% of all deaths in the United States during 19841989 were due to fibroproliferative diseases. The largest contributor to this category was atherosclerosis (Table I), with fibroproliferative diseases of
TABLEI FibroproliferativeDiseases(AnnualDeath Rate per 100,000 Population) Atherosclerosis MI Stroke Other Acutelunginjury Cirrhosis Nephrosclerosis TOTAL
335 230 75 30 38 13 8 394•878
38~ 26~ % 3~ 5~ 1~ 1~ 45~
Adaptedwith permissionfrom ,
the lung, liver, and kidney accounting for the majority of the rest of the cases. Although these diseases seem dissimilar, the fibroproliferative aspects of each suggest that they may share some common pathogenic factors. As knowledge of the molecular events responsible for these diseases continues to grow, such shared disease mechanisms may become more evident. An enhanced understanding of these events may also lead to possible therapeutic interventions to interrupt or prevent a maladaptive fibroproliferative response. Two avenues of research may prove fruitful in the pursuit of potential therapeutic modalities: the characterization of signals that promote the fibroproliferative response and the study of signals that normally terminate it .
MOLECULAR EVENTS THAT PROMOTE THE FIBROPROLIFERATIVE RESPONSE Growth Factors In order to understand further the fibroproliferative response involved in lung granulation, our laboratory examined bronchoalveolar lavage fluid obtained from patients with acute lung injury for the presence of substances capable of stimulating mesenchymal cell migration and replication. We devised a simple in vitro assay for growth factors in which lavage fluid was added to nonreplicating lung fibroblasts that were being maintained in a defined medium. The replication of cells cultured from lavage fluid of patients with ARDS was compared with cells cultured from lavage fluid of patients who were receiving mechanical ventilatory support due to hypermetabolism (mainly neurosurgery patients with head injuries). It was noted that the lavage fluid from the ARDS patients had high levels of biologic activity, whereas the lavage fluid from the control patients had almost none (Figure 2) . One of the best-characterized factors that promotes wound healing is platelet-derived growth factor (PDGF), which appears to act as a chemoattractant and growth factor for mesenchymal cells. After ascertaining that factors in lavage fluid from ARDS patients could indeed satisfy the require-
June 22, 1992 The American Journal of Medicine Volume 92 (suppl 6A)
SYMPOSIUMON ANTIMICROBIALTHERAPY/ BITTERMAN
ment for PDGF in cell cultures, we performed a Western blot analysis with a polyclonal antibody to PDGF to see whether cationic proteins present in the bioactive lavage fluid contained significant quantities of this growth factor. To our surprise, only trace amounts of the 29 kilodalton (kDa) PDGF peptide were detected in the lavage fluid. Instead, the predominant PDGF immunoreactive moiety had a molecular mass of 14 kDa. Small amounts of a 38 kDa PDGF-immunoreactive peptide were also present in some samples . Additional studies of the 14 kDa peptide revealed that it is closely related to the 29 kDa PDGF molecule as assessed by its ability to bind to surface receptors on adult lung fibroblasts and to induce migration and replication of these cells. Preliminary studies in our lab in which lung tissue was hybridized with a probe recognizing a region common to several different species of PDGF mRNA indicated that cells producing this protein are localized to the air-lung interface where granulation is occurring. At least some of the cells containing PDGF-related mRNA are alveolar macrophages. The 14 kDa peptide characterized in our studies is also likely to play a role in the healing of wounds at locations outside the lung. A study of factors present in human wound fluid found that a bioactive, PDGF-like peptide with a molecular mass of approximately 16 kDa was present during the healing of integumentary wounds . It will be of great interest to analyze the degree of similarity between these two small peptides and to unravel their relation to 29 kDa PDGF.
200 "O .
I.J._ (1) O') 100 -
0 0 O0 0000 00000 0000 0000
0 00000 0000000
ARDS n =25
Controls n = 12
Rgure 2. Quantitation of the platelet-derivedgrowth factor (PDGF)-Iikeactivity in bronchoalveolarlavagefluid from patients with adult respiratory distress syndrome (ARDS) as compared with non-ARDS controls. One unit = the reciprocal of the dilution of test fluid that induces 50°~ of the maximum fibroblast growth response. Reprintedwith permission from . 6A-42S
Once the factor responsible for the PDGF-like activity observed in bronchoalveolar lavage fluid from patients with ARDS has been further characterized and its receptor on the target cells identified, it may be possible to develop receptor antagonists that prevent the growth-promoting activity of this factor. Such agents could provide a means to curtail the fibroproliferative response that occurs during the repair phase following acute lung injury. Other growth factors may also participate in the fibroproliferative response. In our studies of the bronchoalveolar lavage fluid from patients with ARDS, we found that in addition to PDGF-related peptides, this fluid contained substances capable of replacing two other growth factors that promote fibrobiast replication: epidermal growth factor and insulin-like growtl~ factor . We have also identified a substance in the bronchoalveolar fluid of patients with acute lung injury that promotes angiogenesis, a key component of the granulation process . These other growth-promoting substances provide additional possible targets for therapeutic intervention. Fibroblasts: An Altered Differentiated State? Response to a growth factor commonly involves two types of cells: the cell producing the growth factor and the cell that responds to it. Having detected a growth factor that could promote the migration and growth of normal tung fibroblasts, we have begun to investigate the biologic behavior of the cells involved in the actual fibroproliferative process, that is, lung fibroblasts obtained from patients with ARDS. In order to study the biologic characteristics of lung flbroblasts, mesenchymal cells from the intra-alveolar granulation tissue of patients who died from ARDS were isolated and grown in culture. Preliminary studies suggest that, unlike normal lung flbroblasts, the fibroblasts taken from the lungs of patients with ARDS were capable of dividing in a defined tissue culture medium that contained no exogenous growth factors. Similar changes in the growth and connective tissue-producing properties of mesenchymal cells have also been reported in cells obtained from lesions in patients with other fibroproliferative diseases, such as atherosclerosis, interstitial renal fibrosis, and idiopathic pulmonary fibrosis [10-15] (Table II). Thus, there is a biologic precedent for the conversion of fibroblasts to a new differentiated state. Once converted to this state, the cells may be able to perpetuate the fibroproliferative process even in the absence of external signals. It is not yet clear what form of cellular communication is required to shift fibroblasts into this biologic mode or to return the cells to their normal pattern of controlled growth and connective tissue production.
June 22, 1992 The American Journal of Medicine Volume92 (suppl 6A)
SYMPOSIUM ON ANTIMICROBIALTHERAPY/ BITTERMAN
The elucidation of such signals may lead to therapies aimed at regulating the fibroproliferative response.
TERMINATIONOF THE FIBROPROLIFE~TIVE RESPONSE: A HYPOTHESIS
Despite the consequences inherent in having a poorly controlled fibroproliferative response occur in the lungs, it is important to keep in mind that some level of response may be required for lung repair. Furthermore, nearly half of the patients with ARDS survive and regain normal lung function, even though in at least some cases the alveolar air spaces of these patients have undergone extensive tissue granulation. In order for this granulation tissue to be removed, the excess fibroblasts must somehow be eliminated. One possible mechanism by which the body could accomplish this feat is through apoptosis, or programmed cell death. This form of ordered cell removal, which occurs in tissue remodeling during ontogeny, is capable of the specific removal of large quantities of cells without concomitant inflammation. Our laboratory is currently exploring the hypothesis that programmed cell death is the principal process responsible for the removal of intra-alveolar lung fibroblasts in patients who successfully recover from acute lung injury. CONCLUSION
The healing process that takes place following acute lung injury is a double-edged sword. Although the fibroproliferative response may be required to repa~ the damage done to the lung, this response can also cause the obliteration of alveolar air spaces, resulting in severe impairment of gas exchange and, for many patients, death. In order for clinicians to counteract overstimulated fibroproliferative responses, the molecular mechanisms involved in the healing process must be understood. Studies have shown that the body promotes the flbroproliferative response through the production of growth factors and may also have the means to stabilize changes in the differentiation state of participant flbroblasts. The termination of the fibroproliferative response in survivors appears to be carefully regulated, perhaps through the production of one or more signals that mediate programmed cell death. Accordingly, there may be several points at which therapeutic intervention may be successful. Modulation of the granulation process may be possible by inhibiting the production of growth factors or, if confirmatory data become available, by modulating the flbroblast's ability to differentiate. Theoretically, it may be possible to regulate the onset of programmed fibroblast death so that sufficient time is allowed for the lung
TABLE II Examples of Clinical Disorders Associatedwith a Stable Change in Mesenchymal Cell Differentiated State Enhancedproliferativecapacity Atherosclerosis Idiopathicpulmonaryfibrosis Interstitial renalfibrosis Enhancedconnectivetissuerelease Liver cirrhosis Progressivesystemicsclerosis Idiopathicpulmonaryfibrosis Autocrineproductionof growthfactors Atherosclerosis Idiopathic pulmonaryfibrosis
to be healed, but not for the air spaces to be completely obliterated. Although these approaches will require a far greater understanding of the biologic mechanisms responsible for lung repair, initial studies supportthe premise that modulation of the flbroproliferative response may provide a means of therapeutic intervention for patients with acute lung injury. REFERENCES 1. Fowler AA, Hamman RF, Zerbe GO, et aL Adult respiratory distress syndrome: prognosis after onset. Am Rev Respir Dis 1985; 132: 472-8. 2. Marinelli WA, Henke CA, Harmon KR, et aL Mechanisms of alveolar fibrosis after acute lung injury. Clin Chest Med ]990; 11: 657-72. 3. Bachofen M, Weibel ER. Structural alterations of lung parenchyma in the adult respiratory distress syndrome. Clin Chest Med 1982; 3: 35-56. 4. Gadek JE. Adverse effects of neutrophils on the lung. Am J Med 1992; 92 (Suppl 6A): 27S-31S. 5. Kozak U, Bacon WE, Krzyzanowski M, et al. Hospital use in Poland and the United States. US Department of Health and Human Services Publication No. (PHS} 88-1478. Vital and Health Statistics, Series 5, No. 2. Washington, DC: Government Printing Office, 1988. 6. Bitterman PG, Henke CA. Fibroproliferative disorders. Chest 1991; 3 (Suppl): 81-4. 7. Snyder LS, Hertz MI, Peterson MS, etal. Acute lung injury: pathogenesis of intraalveolar fibrosis. J Clin Invest 1991; 88: 663-73. 8. Matsuoka J, Grotendorst GR. Two peptides related to platelet-derived growth factor are present in human wound fluid. Proc Natl Acad Sci USA 1989; 86: 441620. 9. Henke C, Fiegel V, Peterson M, et al. Identification and partial characterization of angiogenesis bioactivity in the lower respiratory tract after acute lung injury. J Clin Invest 1991; 88: 1386-95. 10. Benditt EP, Benditt JM. Evidence for a monoclonal origin of human atherosclerotic plaques. Proc Natl Acad Sci USA 1973; 70: 1753-6. 11. Jordana M, Schulman J, McSharry C, et aL Heterogeneous proliferative characteristics of human adult lung fibroblast lines and clonally derived fibroblasts from control and fibrotic tissue. Am Rev Respir Dis 1988; 137: 579-84. 12. LeRoy EC. Increased collagen synthesis by scleroderma skin fibroblasts in vitro: a possible defect in the regulation or activation of the scleroderma fibrobtast. J Clin Invest 1974; 54: 880-9. 13. Parkes JL, Cardell RR, Hubbard FC Jr, Hubbard D, Meltzer A, Penn A. Cultured human atherosclerotie plaque smooth muscle cells retain transforming potential and display enhanced expression of the myc protooncogene. Am J Pathol 1991; 138: 765-75. 14. Rodemann HP, Muller GA. Abnormal growth and clonal proliferation of fibroblasts dervived from kidneys with interstitial fibrosis. Proc Soc Exp Biol Med 1990; 195: 57-63. 15. Yoshida Y, Mitsumata M, Yamane T, Tomikawa M, Nishida K. Morphology and increased growth rate of atherosclerotic intimal smooth-muscle cells. Arch Pathol Lab Meal 1988; 112: 987-96.
June 22, 1992 The American Journal of Medicine
Volume 92 (suppl 6A)