Airway epithelial cell responses to ozone injury.

The airway epithelial cells is an important target in ozone injury. Once activated, the airway epithelium responds in three phases. The initial, or immediate phase, involves activation of constitutive cells, often through direct covalent interactions including the formation of secondary ozonolysis products--hydroxyhydroperoxides, aldehydes, and hydrogen peroxide. Recently, we found hydroxyhydroperoxides to be potent agonists of bioactive eicosanoid formation by human airway epithelial cells in culture. Other probable immediate events include activation and inactivation of enzymes present on the epithelial surface (e.g., neutral endopeptidase). During the next 2 to 24 hr, or early phase, epithelial cells respond by synthesis and release of chemotactic factors, including chemokines--macrophage inflammatory protein-2, RANTES, and interleukin-8. Infiltrating leukocytes during this period also release elastase, an important agonist of epithelial cell mucus secretion and additional chemokine formation. The third (late) phase of ozone injury is characterized by eosinophil or monocyte infiltration. Cytokine expression leads to alteration of structural protein synthesis, with increases in fibronectin evident by in situ hybridization. Synthesis of epithelial antiproteases, e.g., secretory leukocyte protease inhibitor, may also increase locally 24 to 48 hr after elastase concentrations become excessive. Thus, the epithelium is not merely a passive barrier to ozone injury but has a dynamic role in directing the migration, activating, and then counteracting inflammatory cells. Through these complex interactions, epithelial cells can be viewed as the initiators (alpha) and the receptors (omega) of ozone-induced airway disease.


Introduction
Concerns continue to increase about whether the current ambient air quality standard for ozone (120 ppb) provides an adequate margin of safety for the U.S. population. Several clinical (1)(2)(3)(4) and epidemiological (5-7) studies, some of which are presented in detail in this paper, have This is an updated and expanded version of an article that was presented at the Environmental and Occupational Health Sciences Institute Symposium on Ozone Air Quality and Health Effects held [27][28][29] May 1992 in Piscataway, New Jersey.
The authors thank Edwin Donnelly and Thomas Bordner for technical assistance and Danny Edmonds for manuscript preparation. This study was supported in part by grants R01-ES06562, R01-ES06677, and P30-ES06096 from the National Institute of Environmental Health Sciences, the Health Effects Institute, Agreement 90-13, an organization jointly funded by the U. S. Environmental Protection Agency (Assistance Agreement X-812059) and automotive manufacturers; the Electric Power Research Institute grant RP2155-03; the Center for Indoor Air Research grant 93-08A; and the Procter and Gamble Company, UAARP-21314.
The contents of this article do not necessarily reflect the views of the Health Effects Institute nor the policies of the U.S. Environmental Protection Agency or automotive manufacturers.
Address correspondence to Dr. George D. Leikauf documented significant decrements in pulmonary airflow in persons exposed to < 120 ppb (Table 1). Such effects are more readily observed in exercising individuals when exposures are extended (> 4 hr) (8).
In related studies, another consistent finding is an increase in inflammatory cells in bronchoalveolar lavage fluid obtained after <120 ppb ozone exposure (1,3,4). Increases in inflammatory mediators (bioactive lipids and cytokines) and protease (elastase) accompany this change ( Figure 1). The cellular sources of these mediators in bronchoalveolar lavage fluid from proximal conducting airway and distal lung regions are uncertain.
Although the bronchoalveolar junction is considered the primary site of ozone attack in the lung (9), another important target site is the airway epithelium (10). Previous analyses of regional ozone deposition and absorption (11,12) suggest that exposure to high concentrations of ozone could result in deposition sufficient to induce epithelial damage (histological changes). Such ozone exposure can alter the maintenance of epithelial barrier (13-15) and particle clearance function (16). However, at lower concentrations (i.e., -100 ppb), ozone deposition in the airways is likely to be less, and health effects are thus more difficult to uncover experimentally. Nonetheless, the apparent changes in airway function (e.g., decreased forced expiratory volume [FEV10]) argues that airway perturbation (perhaps at a level undetected by conventional histology) must occur even at low level exposure.
Subtle injury to epithelial cells may play an active role in systemic airway responses following low-level ozone exposure (5,17,18). In evaluating these effects, it is important to consider both constitutive functions and elicited defense mechanisms of the epithelium ( Table 2). Ozone may influence cellular function directly by chemical modification of molecular constituents of the cell membrane or indirectly by the interaction of its primary reaction products   with the cell. Although the anatomical, cellular, and biochemical sites of reaction of ozone in the lungs are not known, it is generally believed that ozone does not penetrate further than the airway epithelium because of its high reactivity (19). This would confine ozone to react with substances in the extracellular fluid in the airway lumen and in the apical plasma membrane of the epithelial cell. The resultant ozonolysis products are generally more stable than ozone itself and can diffuse farther into the airway epithelium, where they may activate one or more mediator cascades and thereby elicit a variety of cellular responses. Here, we divide the likely events associated with ozone injury to the airway epithelium into immediate, early, and late phases (Table 3).

Immediate Phase
The immediate phase occurs during (or within 2 hr after) exposure and is charac- terized by activation of constitutive cells in the airway epithelium. Because it is a highly reactant irritant (20), ozone's attack may be limited to components of the respiratory tract fluid and the apical plasma membrane, with little unreacted ozone penetrating into the cytosol (19). Examination of ozone reactions with amino acids (21,22), proteins (23-28), unsaturated fatty acids (29)(30)(31), and phospholipids containing fatty acids (32)(33)(34)(35) in vitro indicates that ozone is likely to react with cysteine, methionine, tyrosine, tryptophan, and histidine residues in proteins and carbon-carbon double bonds of unsaturated fatty acids in membrane lipids ( Figure 2). The latter reactions yield hydrogen peroxide, aliphatic and allelic aldehydes and hydroxyhydroperoxides of various chain lengths (32,35). Recently, we have investigated the ability of selected ozonolysis products of fatty acids to activate membrane phospholipases in human airway epithelial cells in culture (36). Compounds with longer chain lengths were more potent than compounds of shorter chain length (C9 >C6 >C3), and hydroxyhydroperoxides were more potent than corresponding aldehydes or hydrogen peroxide in initiating eicosanoid release from human airway epithelial cells. Significant effects were observed at concentrations of .3 pM 1hydroxy-I-nonanehydroperoxide.
To place these findings into perspective with regard to in vivo exposure concentration, we estimate the rate of hydroxyhydroperoxide production at the airway surface based on available estimates of ozone deposition in the lung. Many difficulties are associated with such an estimate, so an initial point is to apply the dosimetric model developed by Miller and Overton (11,12) that predicts a dose of up to approximately 2 pg 03 deposited/cm tissue x min per pg ambient 0 3/m3 air. This value can be higher or lower, depending on breathing rate and site of regional deposition, respectively. At an ambient concentration of 120 ppb (2.4x 108 pg/mi3 air: the current ambient air quality standard), the rate of adsorption of ozone by the airway epithelium would be 10 pmole/cm2 x min (480 pg/cm2 x min ÷ 48 pg/pmole) From the stoichiometry of the reaction, we know that 1 molecule of hydroxyhydroperoxide will be produced from every molecule of ozone that reacts with a carbon-carbon double bond in an unsaturated fatty acid. If we limit the amount of ozone that reacts with unsaturated fatty acids to 20% of the absorbed dose, then the hydroxyhydroperoxides formation rate would be approximately 2 pmole/cm2 x min. Assuming these products remain within a 10 pm layer on the airway surface, the concentration of hydroxyhydroperoxide is approximately equal to the lowest dose (3 ,uM) that elicited [3H]-activity release from airway epithelial cells.
These findings suggest that secondary ozonolysis products also have biological activity. Because they are more stable than ozone, these compounds can move from their formation site and react within the cytoplasm, activating other enzymes (e.g., cyclooxygenase). Thus, inhaled ozone can generate chemical intermediates that can react with and activate epithelial cells. Possible reaction sites of ozone with cellular macromolecules. The three preferred sites are free sulfhydryl groups in membrane proteins, carbon-carbon double bonds in unsaturated fatty acids, and disulfide crosslinks in membrane proteins. Ozonolysis of membrane fatty acids leads to the formation of an aldehyde and a hydroxyhydroperoxide, one retained in the membrane the other free to exist the membrane and react with cytoplasmic constituents. When a hydroxyhydroperoxide is formed, it subsequently reacts in the presence of water to form a second aldehyde and hydrogen peroxide (H202). Thus, the net reactions yield two aldehydes and H202' In addition to immediate activation of phospholipases (responsible for arachidonic acid release and subsequent metabolism by specific pathways), ozone could possibly alter other enzymes responsible for biotransformation capacity within the epithelium. Again, this could result from direct ozone reactions with amino acid in membrane-bound proteins (24,28) or through the action of reactive intermediates (such as hydroxyhydroperoxides) that can act at sites distant (cytoplasmic, nucleus) to the site of initial ozonolysis (35). Thus both surface and nonmembrane-bound proteins and enzymes could be activated or inactivate by ozone exposure.
The biological consequences of other inhaled materials, when absorbed subsequently, can also be influenced by this process because they can be activated by epithelial enzymes (e.g., aromatic hydrocarbon by cytochrome P4 ) or inactivated by other enzymes, (e.g., Mehydes by aldehyde dehydrogenases). Thus, alteration in constitutive "metabolic" functions within the epithelial cell may affect local response to subsequent stimuli. Experimental details are yet to be obtained with ozone in this area, this may be a fertile subject for further research.
An example of inactivation/activation of epithelial enzymes has been clarified following toluene diisocyanate inhalation. This irritant can inactivate neutral endopeptidase, contained on the plasma membrane surface of airway basal cells. Loss of neutral endopeptidase can in turn heighten airway smooth muscle reactivity to substance P and other neuropeptides (37). Similarly, loss of neutral endopeptidase activity caused by toluene diisocyanate may influence subsequent responses to irritants such as compounds that induce mucosal edema or mucus gland secretion (through mechanisms involving retrograde neurotransmitter release from sensory neurons). Data on ozone's effects on this important enzyme are preliminary (38,39), but recent evidence suggest that neuropeptide release and inactivation of neutral endopeptidase can also occur following ozone exposure.
Early Phase During 2 to 24 hr after acute injury, the next (early) phase is initiated and characterized by infiltration of polymorphonuclear leukocytes (predominately neutrophils). Epithelial cells, along with other resident cells, can synthesize and release chemotactic factors thereby directing migration and activation of neutrophils (17,40,41). This, in part, results from augmented bioactive lipid mediator release from both resident and migratory cells, with elevations present in pro-inflammatory eicosanoid (e.g., leukotriene B4 or prostaglandin F2Q). As noted previously, ozone activates eicosanoid metabolism in airway epithelial cells (17,36).
Additional elevations of cytokine concentrations, demonstrated to be derived from airway epithelial cells, may also play an important role in orchestrating this early response to ozone. Elevated synthesis and release of interleukins (IL) and possibly tumor necrosis factor (TNF ) during this early period are thought to activate the subsequent release of secondary mediators. For example, IL-1 and TNFa may be responsible for secondary IL-6 and IL-8 expression.
Recently, we have begun to examine the effects of ozone inhalation on macrophage inflammatory protein-2 (MIP-2), a chemokine, in mouse lung (42). These results indicate that MIP-2 mRNA transcript levels are increased (as determined by reverse transcription-polymerase chain reaction) at doses that produce increases in neutrophils in lavage fluid. Increases in MIP-2 transcript levels also precede the neutrophil infiltration. Our initial conclusion is that ozone-induced chemokine (including MIP-2 and related proteins, IL-8, RANTES, etc.) expression has a clearer association with neutrophil infiltration than does chemotactic eicosanoid (leukotriene B4 or 1 5-diHETE) formation.
Once leukocytes are present, the profile of eicosanoid metabolites within the epithelium also may change through transcellular eicosanoid metabolism. For example, leukotriene A 4 (LTA ) released from activated neutrophils can be subsequently metabolized to leukotriene B (LTB4) by adjacent cells. Recently, we ave begun examining transcellular metabolism with airway cells and neutrophils stimulated with Ca-ionophore, a stimulus of phospholipase. Leukotriene B4 production is increased when cells are cocultured. Since ozone also stimulates phospholipase, it is likely that it could also affect LTA4-LTB4 co-metabolism.
Finally, during this period, elastase and cathepsin G can be released from incoming cells. Neutrophil elastase is clearly elevated in bronchoalveolar lavage in persons exposed to ozone (Figure 1). Elastase is a potent agonist for mucus secretion (43) and chemokine formation (41).

Late Phase
The third phase of ozone injury is characterized by a period of eosinophil (important in asthma) or monocyte (important in bronchitis) infiltration. Clear (44), and possibly granulocyte, macrophage-colony stimulating factor (45,46) or transforming growth factory (47)(48)(49). During this time, factors like transforming growth factormay markedly enhance fibronectin mRNA expression. In human airway epithelial cells, as evidenced by cells in culture, TGF -induced fibronectin mRNA expression can increase with a time course consistent with a late phase response (maximal levels occur after 18 to 24 hr). Synthesis of other structural proteins (important in tissue repair), antiproteases (including secretary leukocyte protease inhibitor) (50), and antioxidants (superoxide dismutase, catalase and glutathione reductase) (51) also increase during this period. The loss of anti-inflammatory eicosanoids through suicidal, cyclooxygenase inactivation may aggravate leukocyte activation, protease release, and tissue damage. Clearly several of these possibilities merit further investigation.

Summary
Inhalation of ozone in the past has been associated with airway dysfunction marked by physiologic and histological change. These changes are typically noted only after high level exposure. More subtle changes have been noted in recent years which may or may not be associated with histological changes. Through a clearer understanding of ozone's preferred targets and the role of subtle biochemical changes in the airway epithelium, the early events in a host-initiated cascade have been uncovered. At present, aspects of the above discussion are still speculative due to a lack of experimental details. (For example, does the increase in MIP-2 observed in mouse lung occur in humans?) One purpose of this presentation is to present future research directions that may increase our knowledge of the cellular and subcellular mechanisms responsible for ozone toxicity.
Nonetheless, at present we can conclude that epithelium is not passive in airway injury/response to ozone. It has an active role in directing the migration of inflammatory cells and in the release and co-metabolism of inflammatory mediators. Thus, it is partially responsible for changes in the microenvironment surrounding the epithelial cells. Ultimately, these changes can alter the phenotype of the epithelial cell population, possibly leading to persistent changes in cell metabolism (e.g., loss of neutral endopeptidase) or in mucin-producing cell types important in hypersecretory diseases (bronchitis). In this way, epithelial cells can be viewed as the alpha (initiator) and the omega (receptor) cell in ozone-induced airway disease.