Mechanisms of Acute Lung Injury and Repair

TEACHING POINTS.

• •Acute lung injury represents a common pathway of cellular and chemical processes despite a wide array of underlying causes.
• •Inflammation causes alveolar permeability and leads to extravasation of protein-rich fluid into the alveolar space.
• •Leukocytes, endothelium, and epithelium all actively contribute to the injury process.
• •Mediators of injury are also mediators of host defense, which makes physiologic studies (and treatment) complex.

BACKGROUND

The term acute lung injury (ALI) refers to a syndrome of diffuse pulmonary inflammation and increased capillary permeability that manifests in acute refractory hypoxemia and lung infiltrates. Although references date through the last century, the first formal description of the syndrome is attributed to Ashbaugh in 1967.¹ Historically, reports of the syndrome have included terms such as adult respiratory distress syndrome and shock lung, thus emphasizing specific patient populations or predisposing conditions. More recently, hyaline membrane disease of newborns has been recognized as sharing the same radiographic and histologic findings, and being mediated by the same cellular and soluble factors, as the syndrome seen in adults with septic shock. To incorporate all continuums of patient populations and the vast variety of primary insults that lead to the final common pathway of diffuse pulmonary parenchymal damage, the more inclusive terms acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) are now favored. The first American European Consensus Conference² defined ALI as (1) acute onset of bilateral infiltrates consistent with pulmonary edema; (2) the absence of evidence for left atrial hypertension; and (3) reduced ratio of arterial oxygen tension (Pao₂) to the fraction of inspired oxygen (Fio₂), Pao₂/Fio₂. The syndrome is called ALI when Pao₂/Fio₂ >300 and ARDS when this ratio falls below 200. The consensus definitions of ALI/ARDS have allowed for extensive clinical and translational research into the mechanisms of lung injury and repair.

STRUCTURE AND FUNCTION

Radiographic and Pathologic Aspects

Acute lung injury's hallmark finding of noncardiac pulmonary edema from alveolar capillary disruption usually appears several hours after an initial predisposing insult, but may not be detected for up to 72 hours. Clinical aspects of ALI are covered in more detail in Chapter 19. The syndrome of ALI has been attributed to an extensive and broad list of inciting causes that are frequently divided into direct (or primary pulmonary insult) and indirect (extrathoracic insult with systemic involvement) insults. A direct insult to the lung stimulates alveolar macrophages to produce a cascade of cytokines including tumor necrosis factor-alpha (INF-α) and interleukins (IL),3, 4, 5 which in turn recruit further cellular and biochemical responses that characterize the clinical effects of acute lung injury including fever, epithelial injury, and increased endothelial permeability.⁶ The pulmonary response to an indirect insult is commonly considered to be part of the so-called systemic inflammatory response syndrome (SIRS), and mediated by migration of proinflammatory cytokines7, 8 and microbes9, 10, 11 through the systemic circulation. Many of the local and systemic mediators of ALI are included in Table 6-1 . Regardless of the nature and anatomic location of the initial insult, the clinical physiology and pathologic findings within the lungs are remarkably similar, indicating a common final pathway of injury.

Table 6-1.

Mediators of Acute Lung Injury

Factor	Class	Main Cell of Origin	Major Effects
Tumor necrosis factor-alpha (TNF-α)	Cytokine	Macrophage	Activate endothelium
			Induce nitric oxide synthesis
			Stimulate IL-1 production
			Fever, mobilize metabolites
Interleukin-1 β (IL-1β)	Cytokine	Neutrophil, endothelial cell	Activate endothelium
			Stimulate IL-6 production
			Local tissue destruction
			Fever, mobilize metabolites
IL-6	Cytokine	Macrophage, endothelial cell	Activate lymphocytes
			Stimulate antibody production
			Fever
IL-8	Cytokine	Many	Stimulate neutrophil transmigration
			Degranulate neutrophil (oxidative burst)
IL-10	Cytokine	Many	Suppress proinflammatory cytokine expression
Vascular endothelial growth factor (VEGF)	Soluble protein	Many	Activate endothelium
			Induce permeability
			Stimulate adhesion molecules
Intracellular adhesion molecule (ICAM)	Adhesion molecule	Endothelial cell	Attract leukocytes to injured endothelium
Vascular adhesion molecule (VCAM)	Adhesion molecule	Endothelial cell	Attract leukocytes to injured endothelium
Nuclear factor kappa-B (NF-κB)	Transcription factor	Many	Increase transcription of proinflammatory mediators
CD14	Cell surface receptor	Macrophage	Bind bacterial endotoxin
			Activate macrophage
Toll-like receptor (TLR)	Cell surface receptor	Macrophage	Bind bacterial endotoxin
			Activate macrophage

The acute phase of lung injury is also characterized by reduced respiratory system compliance and pulmonary hypertension. The latter, in conjunction with increased capillary permeability, leads to pulmonary edema, which is contrasted to congestive heart failure by the finding of protein-rich fluid in the alveoli of those with ALI.¹² Despite vast differences in the molecular characteristics of noncardiogenic edema fluid, chest radiograph findings are indistinguishable from those depicting cardiogenic pulmonary edema¹³—ranging from asymmetrical patchy infiltrates to dense consolidation and may also include pleural effusions.14, 15 Chest computed tomography indicates heterogeneous involvement of injured lungs with a dependent gradient of consolidation that results in reduced effective alveolar surface area.¹⁶ This consolidation relates to prognosis in that the percentage of lung units that can be recruited by increasing ventilatory support correlates with mortality.¹⁷ In parallel, measure ments of physiologic dead space correlate inversely with the prognosis of patients with acute lung injury.¹⁸ It is important to note that lung areas that appear radiographically normal exhibit significant biochemical abnormalities on analysis of bronchoalveolar lavage fluid.¹⁹

Post-mortem examination of lungs from ALI patients reveals heavy, congested, and atelectatic lungs. The majority of autopsy organs show evidence of infection²⁰ even though pre-mortem studies reveal pneumonia at a much lower rate21, 22, 23; whether this is due to sampling errors or antibiotic effects on clinical evaluation is not clear. During the acute or exudative phase of ALI, biopsies display hyperemia and evidence of both epithelial and endothelial cell injury.24, 25 Endothelial cells are swollen, with decreased cell-cell adhesion leading to extravasation of microthrombi and polymorphonuclear neutrophils (PMNs) into the interstitium. Denudation of the alveolar epithelium leads to flooding of the alveoli with proteinaceous fluid, immune cells (largely PMNs) and red blood cells. Alveoli may be atelectatic and hyaline membranes are seen on the epithelial side of the basement membrane. The exudative phase of ALI usually persists for several days, after which many patients have rapid clinical improvement with resolution of parenchymal injury. This resolution is marked by a return of macrophages as the prominent luminal cell,²⁶ whereas type II pneumocytes re-epithelialize the alveolus, differentiate into type I pneumocytes, and restore epithelial barrier function. Other patients develop a prolonged fibroproliferative phase characterized by the absence of type II pneumocytes and the proliferation of myofibroblasts, fibronectin, and collagen within the alveoli.²⁷ It is not clear what triggers this prolonged recovery phase in some individuals with ALI, but these findings on lung biopsy signal an ominous prognosis. Attempts at reducing or reversing the fibroproliferative phase with corticosteroids showed initially promising results,28, 29 but ultimately corticosteroids have yet to be proven beneficial in larger studies of patients with ALI.

CELL-MEDIATED LUNG INJURY

Neutrophils

Biopsy and bronchoalveolar lavage specimens from patients with ALI are dominated by neutrophils, and intense investigations into the function of these cells during clinical illness have uncovered numerous processes mediated by PMNs. Neutrophils are recruited to injured sites by the expression of adhesion molecules including intracellular adhesion molecule (ICAM)-1 and E-selectin on activated endothelial cells, and plasma levels of these adhesion molecules correlate with the degree of organ dysfunction.³⁰ Once recruited to the alveolus, neutrophils in patients with ALI demonstrate activation of the transcriptional regulatory unit nuclear factor kappa-B (NF-κB), which in turn increases neutrophil expression of IL-1β and other proinflammatory mediators implicated in tissue injury. Accordingly, both the persistence of alveolar neutrophils and the activation of NF-κB within neutrophils³¹ correlate inversely with survival. Following activation, neutrophils orchestrate a process of oxidative burst which generates hydrogen peroxide and other reactive oxygen and nitrogen species that destroy invading pathogens. This free-radical stress also culminates in oxidized host phospholipid membranes that alter mitochondrial and cellular function and compound tissue injury. Neutrophils may play a key role in the apoptosis (or programmed cell death) of immune cells in the lung through regulation of phosphatidylinositol-3 kinase (PI-3K) pathways.³²

The association of neutrophil number and activation with outcome provides strong evidence of the involvement of these leukocytes in the injury pathways, yet tremendous controversy persists regarding the importance of PMNs in the pathophysiology of ALI. Attempts to mediate neutrophil activation have led to conflicting results, raising questions regarding PMN function. This controversy is at least in part due to the recognition that these processes contribute to organ injury but are also key mediators of host defense. For these reasons, clinical and translational research continues in this important area.

Macrophages

The predominance of neutrophils in lung specimens obtained during ALI has drawn attention away from monocytic cell lines, including the macrophage. Alveolar macrophages are the resident immune cell in the lung, and represent the innate host defense system and as such initiate many of the processes leading to the intense inflammatory response of ALI (Fig. 6-1 ). Importantly, key macrophage activity may precede the clinical recognition of the disease. Alveolar macrophages recognize pathogens or their products through various cell surface receptors. In the most well-characterized system, pathogens or their products initially bind to CD14 which in turn recruits Toll-like receptor (TLR) subtypes that are specific for the toxin.33, 34, 35 Together, toxin binding to CD14/TLR leads to activation of the macrophage and a subsequent intense inflammatory response. In addition to CD14 on the cell surface, macrophages release a soluble form of the receptor, sCD14, which activates cells that do not express CD14 (most notably endothelial and epithelial cells). The importance of sCD14 is demonstrated by the finding that the concentration of this protein in alveolar fluid is highly associated with the number of neutrophils in the lung.³⁶

Figure 6-1.

The alveolar macrophage in acute lung injury. Pathogenic bacteria are recognized by a family of Toll-like receptors (TLRs) which present the pathogen to the cell surface receptor CD14. Together, these receptors mobilize NF-κB to the nucleus where it facilitates gene transcription of several proinflammatory factors. Tumor necrosis factor-alpha (INF-α) stimulates other immune cells and leads to the production of nitric oxide, which promotes both killing of engulfed bacteria as well as tissue injury. Stimulated macrophages also produce interleukin (IL)-6, which activates endothelial cells, and IL-8, which recruits neutrophils to the site of injury.

Once stimulated, macrophages display mobilization of NF-κB to the cell nucleus.³⁷ In contrast to neutrophil activation, macrophage activation is more characterized by the production of INF-α³⁸ although certainly IL-1β is secreted as well. Like the macrophages themselves, INF-α levels peak³⁸ and wane early in the disease process, making investigations, or manipulations, of INF-α-mediated processes in humans difficult. In addition to INF-α, the secretory products of activated macrophages form a list that is extensive,³⁹ redundant, and interactive in function, and summate to sequester and stimulate neutrophils in the lung early in ALI.³⁸

The return of macrophages as the dominant alveolar cell line signals resolution of ALI,²⁶ supporting the notion that macrophages are also involved in the regulation of tissue injury. Further support is garnered by findings that alveolar macrophages possess surface receptors for neutrophil proteases,⁴⁰ and can scavenge hydrogen peroxide and limit oxidant-mediated injury.⁴¹ Finally, macrophages phagocytose neutrophils in vitro in a time-dependent manner that corresponds to the clinical time course of resolution.⁴² Taken together, these observations lend credence to the importance of alveolar macrophages in the resolution of ALI.

Endothelial Cells

The histology of ALI clearly documents altered capillary endothelial cells, and some consider ALI as a continuum of “panendothelial disease” resulting from the systemic inflammatory response syndrome.⁴³ This not only pertains to indirect, but also direct causes of ALI, as alveolar INF-α affects the adjacent endothelium.⁴⁴ Once considered a relatively static cell line, the endothelium is now recognized as an active tissue that regulates blood flow, immune function, and solute transport. Whether the noted changes represent an injury to the endothelial cell or an activation of this cell line is controversial. Most likely, there is a continuum of altered endothelial processes that, if allowed to persist unabated, lead to irreversible loss of normal function. Because endothelial cell changes in ALI involve a loss of normal cell function and the extent of endothelial changes is related to the severity of disease, the term endothelial injury will be used throughout this chapter. This view is exemplified by findings that von Willebrand factor antigen (vWf, normally found in large concentrations only within endothelial cells) is released from injured cells into the vessel lumen and into the alveolar space. The extracellular concentrations of vWf are predictive of the development of ALI in those at risk,⁴⁵ and of outcome in patients with established ALI.46, 47

Systemic inflammation leads to the secretion of vascular endothelial growth factor (VEGF) from many different cell lines. VEGF, also known as vascular permeability factor, induces many of the endothelial changes seen in ALI. Once injured, the endothelium transforms from a flat monolayer with tight intracellular junctions to an irregular surface of rounded endothelial cells and a loss of cell-cell interactions. This state creates a permeable surface such that fluid can escape capillaries into the interstitium and ultimately the alveoli. Extravascular lung water is clearly deleterious to gas exchange, but the injured endothelium also allows plasma proteases to exit the vessel and impair alveolar surfactant function and contribute to atelectasis.⁴⁸

The role of endothelial cells in lung injury extends beyond a passive loss of barrier function leading to extravasation of vessel contents into the alveoli. VEGF stimulates the expres sion of several adhesion molecules on the luminal surface of endothelial cells, particularly intracellular adhesion molecule (ICAM), vascular adhesion molecule (VCAM), and the selectin family of glycoproteins. Together, these adhesion molecules function to slow neutrophil transport within the vessel and initiate rolling and adhesion of neutrophils on the endothelial surface (Fig. 6-2 ). Additional chemotactic molecules on the basolateral surface and beyond then promote transmigration through the permeable endothelium. VEGF is downregulated by endothelial-derived factors and this function is lost during sepsis; hence, the loss of normal endothelial function contributes to further endothelial injury. Importantly, manipulations to decrease expression of VEGF lead to decreased organ injury and improved mortality in live-infection models of sepsis, demonstrating the importance of VEGF and the related adhesion molecules in the progression of disease.⁴⁹ Consistent with these findings, the concentrations of endothelial-neutrophil adhesion molecules are more strongly associated with mortality in humans than are measures of neutrophil activation.³⁰

Figure 6-2.

Endothelial-neutrophil interactions in acute lung injury. Endothelial cells become activated through a variety of factors, including tumor necrosis factor-alpha, angiotensin II, and vascular endothelial growth factor. Once activated, endothelial cells lose cell-cell interactions and the monolayer is permeable. Activated endothelial cells express P- and E-selectins which form weak interactions with passing neutrophils (via L-selectin) and initiate leukocyte rolling. Stronger interactions with intracellular adhesion molecule (ICAM) and vascular adhesion molecule (VCAM) function to adhere leukocytes to the monolayer, where chemokines such as interleukin-8 (IL-8) stimulate transmigration into the tissue.

Beyond neutrophil recruitment, endothelial cells propagate the inflammatory response by secreting cytokines, including IL-1 and IL-6. These inflammatory mediators further stimulate endothelial cells to decrease tissue-type plasminogen activator and increase plasminogen activator inhibitor activities⁵⁰ as well as decrease thrombomodulin secretion,⁵⁰ and thus induce the procoagulant state found in the alveolar fluid of patients with ALI.⁵¹ Injured endothelial cells have diminished capacity to secrete endogenous vasoconstrictors and vasodilators necessary to regulate blood flow.⁵² Endothelial cells exhibit injury that is evident on pathologic studies, but they clearly mediate the injury pattern and contribute to the morbidity and mortality of ALI.

Epithelial Cells

Histologic analysis shows diffuse alveolar damage during clinical ALI, and altered epithelial structure is evident. The lung epithelium has many functions during health, and many of these functions are lost during acute inflammatory processes. Epithelial damage clearly contributes to the pathogenesis and morbidity of ALI.

Increased capillary permeability may present a conduit for vascular contents to extravasate into alveoli, but there is a growing body of evidence describing ineffective clearance of alveolar fluid by epithelial cells leading to the clinical findings of noncardiogenic pulmonary edema. The epithelial surface is lined mainly with type I pneumocytes that maintain the structural integrity of the alveolus through barrier function. The remaining cells in the lung epithelium are type II pneumocytes, whose diverse functions include ion transport regulation, surfactant production, and regeneration of type I pneumocytes. Defects in epithelial function lead to alveolar fluid accumulation in two ways—increased permeability and decreased fluid transport out of the alveoli. However, the epithelial barrier is less permeable than the endothelium, even after injurious exposure,⁵³ suggesting that defects in alveolar liquid clearance play a large role in the accumulation of extravascular lung water. Indeed, the rate of alveolar fluid clearance is impaired in patients with ALI, and inversely related to prognosis.⁵⁴ Transepithelial transport of fluid occurs in several fashions, the best described being along an osmotic gradient formed by active transport of sodium via a Na⁺/K⁺-ATPase on the basolateral surface of type II pneumocytes.⁵⁵ Experimental evidence shows that hypoxia leads to displacement of the Na⁺/K⁺-ATPase from the basolateral surface—a condition that is rapidly reversible with alveolar instillation of the beta-adrenergic agonist, terbutaline.⁵⁶ Although the role of Na⁺/K⁺-ATPase in human ALI is still unclear, beta-agonist administration decreases total lung water in ALI patients, lending support to this hypothesis. Alveolar fluid clearance is impaired by several factors in addition to hypoxia. The generation of reactive oxygen and nitrogen species, possibly caused by free-radical deactivation of transport proteins, diminishes epithelial fluid transport.57, 58 Epithelial function is likely impaired through a loss of epithelial cells through the process of apoptosis.⁵⁸

With the progression of ALI and loss of epithelial integrity, inflammatory mediators and bacteria normally contained by the intact epithelium can enter the lung parenchyma and circulation.⁵⁹ The importance of this phenomenon is highlighted by a study showing that mortality-lowering ventilator strategies reduce the incidence of bacteremia in animal models.⁶⁰ However, compelling data of translocation across human lung epithelia are lacking. In contrast, there is a significant body of evidence that surfactant, a secretory product of the epithelium, is altered in composition and function in human ALI. Normal lung surfactant is composed of phospholipids, neutral lipids, apoproteins, and the surfactant proteins (SP)-A, B, C, and D. During ALI, the total amount of surfactant phospholipids is reduced, with marked decreases in phosphatidylcholine and phosphatidylglycerol61, 62; these changes are associated with increased surface tension of surfactant liquid⁶¹ and the severity of respiratory failure.⁶³ In addition to deficiencies in phospholipids, there is marked depletion of SP-A and SP-B during ALI. The net deficiency of proteins and phospholipids is due to decreased production by type II pneumocytes as a result of inflammation⁶⁴ and increased destruction by oxidant stress⁶⁵ and proteolytic cleavage.⁶⁶ The findings of abnormal surfactant composition and function have led to more than 200 clinical evaluations of exogenous surfactant,⁶⁷ yet no trial has provided a convincing mortality benefit. Several variables in previous trials of exogenous surfactant therapy include issues of dose, timing, and composition of the applied therapy—leading to continued debate regarding the future of this mode of treatment.

Soluble Mediators of Inflammation

CYTOKINE-MEDIATED LUNG INJURY

Cytokines are low-molecular-weight soluble proteins that transmit signals within cells; therefore they mediate many of the interactions between immune cells and lung tissue in the pathogenesis of ALI. Despite the intense effort spent in researching the role of cytokines in the pathogenesis of ALI, criticism of human studies has evolved and some consideration of sampling techniques must be mentioned. In most instances, cytokines participate in cell signaling through interaction with cell surface receptors and in this regard may have very localized effects. In parallel, because various anatomic barriers compartmentalize the immune response,⁶⁸ cytokine concentrations may be quite different in the serum, alveolar space, or lung parenchyma. Furthermore, systemic cytokines may represent “the tip of the iceberg”⁶⁹ or overflow of cytokines from various sources,⁷⁰ making interpretation of these factors complex. There is evidence, however, that the alveolar compartment is disrupted during intense inflammation and airway sampling yields cytokine concentrations that are similar to more invasive sampling.⁷¹ With this in mind, as well as the capacity to perform repeated measurements in patients safely and easily, most regard bronchoscopic sampling of alveolar fluids as the best method available to assess the role of cytokines in ALI.⁷²

Investigations into the role of cytokines in ALI initially focused on INF-α and IL-1β—primarily because bacteria stimulate production of these molecules, whereas in the absence of bacteria, these cytokines are capable of initiating an inflammatory response identical to ALI. INF-α has been identified in BAL fluid in some, but not all, studies of human ALI. This discrepancy may be related in part to timing of samples because this cytokine may be prevalent in the airway fluid for less than 24 hours. INF-α exerts its effects through interaction with TNF receptors I and II on the surface of macrophages and other immune cells,⁷³ or is shed as a soluble receptor. INF-α interactions with these soluble receptors are complex because the receptors may either potentiate INF-α effects by stabilizing the cytokine, or attenuate INF-α effects by interfering with INF-α binding to active receptors.⁷⁴ The local effects of INF-α collectively function to increase inflammation within the tissue by activating endothelial cells, thereby increasing vascular permeability and allowing the passage of immune cells, immunoglobulin, and complement into the tissue. In addition to these effects, INF-α also stimulates the production of itself and IL-1β and thus further stimulates inflammation.

Like INF-α, IL-1β has been identified in lavage fluid obtained during ALI, and through binding to the interleukin-1 receptor (IL-1r)—a potent stimulator of vascular endothelium. This cytokine also activates lymphocytes and as such has many of the same effects as INF-α. However, IL-1β is commonly thought to be more responsible for tissue destruction than is INF-α, as evidenced in a study that showed that inhibition of IL-1β decreased endothelial activation caused by BAL fluid from ALI patients, whereas inhibition of INF-α had no effect.⁷⁵ In addition to the effects on tissue injury, IL-1β stimulates the production of IL-6. It has been proposed that IL-6 integrates the inflammatory response through diverse actions including differentiation of lymphocytes, induction of immunoglobulin production, and induction of many proteins found during the acute phase of inflammation.⁷⁶ Unlike INF-α and IL-1β, IL-6 is found in the BAL of patients throughout the course of ALI, but the concen-trations of the latter cytokine also inconsistently predict outcome.72, 77 All three of these early cytokines can be detected in the blood of patients with ALI, and may be responsible for systemic disease.

While INF-α, IL-1β, and IL-6 have local and systemic effects, the effects of the chemotactic cytokine IL-8 are largely local in nature. This cytokine is the most abundant product secreted by alveolar macrophages followed by bacterial toxin stimulation, and it is the predominant neutrophil chemoattractant in BAL fluid.78, 79 It is worth noting, however, that although IL-8 concentrations correlate with BAL neutrophil concentrations, this strength of relationship varies over time, and IL-8 concentrations at any time are poor predictors of outcome. This observation clearly indicates that additional neutrophil chemoattractants are involved in the pathogenesis of ALI.

Immune cells also secrete a number of anti-inflammatory cytokines that regulate inflammation. In keeping with its role as an integrative cytokine, IL-6 reduces the effects of INF-α and IL-1β by inhibiting their production⁸⁰ and stimulating their natural antagonists.⁸¹ However, the most widely described anti-inflammatory cytokine is IL-10. This counter-regulatory protein is synthesized in lymphocytes and monocytic cells in response to bacteria and inflammatory cytokines, and inhibits inflammatory cytokines, inhibits class II major histocompatibility protein expression, and suppresses monocyte procoagulant activity in experimental systems. IL-10 is found in BAL fluid during ALI, and lower levels of this cytokine are associated with a poor prognosis.⁸² Although this suggests that failure to decrease inflammation leads to increased injury, increasing IL-10 in pneumonia models leads to impaired bacterial clearance and increased mortality.⁸³

Tremendous efforts have been made to identify which cytokines may be primary mediators of inflammation during ALI. Although these studies provide irrefutable evidence of the involvement of many biochemical and cytologic processes, none has clearly prevailed as primary mediators amenable to therapy in human disease. Attempts at blocking or reversing suspected agents in patients have been generally unsuccessful and, at times, lethal. At least four issues may explain this apparent inconsistency. First, ALI induces a complex and redundant inflammatory milieu, and ameliorating any one factor is unlikely to stop the cascade of events. Second, inflammation and organ injury likely begin hours or days prior to the clinical recognition of the disease such that the primary mediators are no longer actively involved at the time of therapy. Third, the inflammatory process is, in fact, a reaction to a primary inciting event; a reduction in inflammation impairs host defense mechanisms. Fourth, and related to the latter issue, the host response is a tightly controlled response involving an initial proinflammatory response rapidly followed by an exuberant anti-inflammatory response. Both responses are increased during the first days of lung injury; however inflammatory antagonists predominate in the alveoli within 24 hours of disease onset.⁸⁴ It is becoming clear that the balance of inflammatory agonists and antagonists in patients at risk for lung injury may be more important than any one factor.85, 86 Further, exogenous attempts at altering inflammation may disrupt this balance and lead to either excessive inflammation or impaired host defenses, both of which may increase mortality.⁸⁷

In summary, cytokine measurements in the BAL fluid of patients with ALI have provided insight into the complexity of interactions between lung epithelium, endothelium, and immune cells. However, despite intense research we have yet to discover a consistent pattern of cytokine regulation in ALI. Further research into cohesive groups of cytokines, coupled with markers of epithelial and endothelial barrier injury, is needed to provide a more comprehensive understanding of immune regulation in ALI.

RENIN-ANGIOTENSIN MEDIATED INJURY

The emergence of the severe acute respiratory syndrome (SARS) and the subsequent identification of an angiotensin converting enzyme (ACE) subtype as the receptor for the causative coronavirus have increased our understanding of the role of the renin-angiotensin-system (RAS) in the pathogenesis of ALI. Briefly, renin is produced by a variety of stimuli, including decreased glomerular pressure as encountered during shock. Renin cleaves angiotensinogen to angiotensin, which is subsequently transformed to angiotensin II (Ang II) by ACE subtype I (ACE-1). Ang II then exerts multiple effects through interactions with the AT1 receptor, including vasoconstriction and sodium resorption, that counteract the hemodynamic state during shock.

The association of the lung and RAS was first recognized anatomically and may be traced to findings that ACE-1 is produced extensively on the luminal surface of pulmonary capillary endothelial cells, and thus the lungs are a major source of systemic Ang II.⁸⁸ Furthermore, ACE-1 is found in lung lavage fluid during experimental lung injury⁸⁹ and as well as in serum from patients with lung diseases not typically associated with shock. These findings were initially explained as a demonstration of shedding of ACE-1 from injured endothelium, but there has been gradual acceptance that RAS is actively involved in the process of acute lung injury. Certainly RAS is upregulated in critical illness and patients have increased concentrations of Ang II in serum. It is also clear that this peptide mediates many of the pathogenic processes implicated in lung injury. Ang II promotes inflammation through activation of the NF-κB pathways,⁹⁰ as well as through the recruitment of immune cells via increased expression of VEGF⁹¹ and of intercellular adhesion markers.⁹² Furthermore, through the AT-1 receptor, Ang II promotes collagen formation in lung fibroblasts⁹³ and apoptosis in epithelial cells⁹⁴ and hence is responsible for features seen during the fibroproliferative phase of ALI. Most strikingly, AT-1 receptor blockers may decrease neutrophil infiltration, improve oxygenation, and prolong survival in an animal model of ALI.⁹⁵

The lungs are a significant source of Ang II during critical illness as evidenced by increased concentrations found in arterial compared to mixed venous blood96, 97, 98; presumably this is related to an increase in ACE-1 activity within the lungs. In fact, patients with ACE-1 polymorphisms associated with increased ACE-1 activity appear to have an increased risk of both the development of ALI as well as mortality from the syndrome.⁹⁹ ACE-1 activity is not limited to pulmonary endothelial cells as previously thought—alveolar macrophages, neutrophils, and alveolar epithelial cells also produce ACE-1. These additional sources of Ang-II may, in fact, be more clinically relevant during critical illness because endothelium-bound ACE-1 activity is actually decreased in patients with ALI.¹⁰⁰ Because inhibition of ACE-1 attenuates endothelial activation¹⁰¹ in patients at risk of ALI and decreases INF-α activation in animal models,¹⁰² persistent expression of angiotensin I is likely to contribute to the pathogenesis of ALI.

Recently, a homologue of ACE-1, or ACE-2, was discovered. This subtype of ACE cleaves Ang I and Ang II to additional angiotensin species that do not act through the AT-1 receptor. In so doing, ACE-2 appears to be protective in ALI; loss of ACE-2 activity via genetic manipulation of experimental sepsis leads to increased vascular permeability, lung edema, and neutrophil accumulation.¹⁰³ The finding that ACE-2 is an essential receptor to the coronavirus responsible for SARS, a severe demonstration of clinical ALI, lends clinical credence to the protective function of this enzyme.

CLINICAL APPLICATIONS

Decades of research have provided insight into many mechanisms of the pathogenesis of ALI, but have failed to identify disease-specific mechanisms that are amenable to therapy. As such, therapy is considered supportive, and most patients require some form of mechanical ventilation, parenteral fluids and nutrition, and intensive monitoring. The largest clinical trials of ALI104, 105, 106 have described excessive mortality and morbidity because of the supportive care that allows patients to survive beyond the first several hours of disease. Patterns of injury related to ICU therapy should, therefore, be discussed.

Ventilator-Associated Lung Injury

Soon after the initial description of ALI and its pathology, many animal models of positive-pressure mechanical ventilation have demonstrated that high peak inspiratory pressures induce injury in previously normal lungs, including changes in pulmonary mechanics,¹⁰⁷ alveolar integrity,¹⁰⁸ and histology¹⁰⁹ that are identical to those seen in ALI. Subsequently, investigators found that limiting chest wall excursion during high pressure ventilation reduced the magnitude of this injury, demonstrating that the high lung volumes induced by high inspiratory pressures are responsible for this injury.¹¹⁰ Furthermore, low tidal volume strategies decrease the translocation of bacteria⁶⁰ and toxins¹¹¹ from the lungs to the systemic circulation, showing that high tidal volumes contribute to a loss of epithelial barrier function. Edema and atelectasis in ALI could culminate in a significant decrease in effective alveoli, the so-called “baby lung,”112, 113 such that standard tidal volumes produce overdistention and alveolar wall shear stress and injury of the remaining functional lung.¹¹⁴

Clinical investigations of patients with ALI demonstrate that those ventilated with lung-protective strategies that included limited tidal volumes had lower concentrations of neutrophils and inflammatory cytokines in BAL fluid and decreased systemic inflammatory cytokines compared to those ventilated with traditional settings.¹¹⁵ An early study of lung-protective ventilation strategies showed mortality lower than that predicted by severity of illness scoring,¹¹⁶ but the initial randomized controlled trials of lung-protective ventilatory strategies that followed were small and produced conflicting results.117, 118, 119, 120 However, in the largest clinical trial of ventilator strategies in ALI, the group of patients ventilated with small tidal volumes (6 mL/kg) experienced lower mortality than those ventilated with higher tidal volumes (12 to 15 mL/kg).¹⁰⁵ Although there is controversy regarding both the choice of control group strategy and whether a tidal volume between the two strategies would be even more effective, experimental and clinical evidence clearly shows that mechanical ventilation with high tidal volumes contributes to lung injury and mortality.

Extravascular lung water is a hallmark finding in ALI, and as mentioned, indicates both increased alveolar-capillary permeability and impaired resorption of alveolar edema. In addition, hydrostatic pressure¹²¹ and loss of capillary oncotic pressure owing to hypoproteinemia¹²² contribute to noncardiogenic pulmonary edema. Logically, decreasing capillary hydrostatic pressure may decrease edema and improve outcomes. Conversely, impaired left ventricular stroke volume from inadequate filling pressures may contribute to the inflammatory state resulting from inadequate organ perfusion.

Several studies have examined the relation between extravascular lung water and outcomes in ALI. An early study showed that patients managed with attempts to decrease extravascular lung water had more ventilator-free days than those whose fluid management was based on pulmonary capillary wedge pressures¹²³—whether this was an effect of net fluid management or of the monitoring techniques was unclear. Increasing plasma oncotic pressure while decreasing total body water by co-infusing furosemide and albumin may also improve ICU-related outcomes.¹²⁴ The large ARDSnet Fluid and Catheter Treatment Trial utilized a 2 × 2 factorial design including conservative or liberal fluid treatment strategies guided by either central or pulmonary artery catheters.104, 106 Patients who were randomized to conservative fluid management experienced more ventilator-free days and fewer ICU days—this effect was true regardless of how fluid management was monitored. Although mortality did not differ between the two catheter groups, those managed with pulmonary artery catheters experienced more complications, illustrating the potential for iatrogenic complications in ALI. Taken together, multiple trials have shown that minimizing hydrostatic pressure improves physiology and ICU-related outcomes; the assuredness of this conservative fluid management is likely more important than the methods used to achieve it. These findings add support to the significance and complexity of pathology leading to the accumulation of noncardiogenic pulmonary edema fluid in ALI.

CONCLUSION

Acute lung injury is a syndrome resulting from a variety of causes that involves diffuse inflammation leading to proteinaceous alveolar edema. The mechanisms involved in the injury are remarkably similar regardless of the underlying etiology, leading to damage of both epithelial and endothelial surfaces. These surfaces are not only passively injured, however, and along with leukocytes contribute to the pathogenesis of the syndrome. Considerable overlap exists between mediators of injury and mediators of host defense and repair such that attempts at intervening in the injury pathway have been troublesome.