I. Introduction

Top Next References

The biochemical literature abounds with reviews of drug oxidation and conjugation by and toxic effects in liver (Gillette, 1974a,b). Although it was once believed that drug-induced hepatotoxicity was the result of some inherent property of the native drug, more modern data have implicated chemically reactive intermediates as being ultimate toxicants (Guengerich and Shimada, 1991; Conney, 1982; Hinson and Roberts, 1992). It was suggested that chemically inert drugs were activated in vivo to metabolic products that were capable of forming covalent bonds with proteins, nucleic acids, and other endogenous substances, and the adducts were capable of inducing carcinogenesis or tissue (and cellular) necrosis (Brodie, 1967). In fact the large research field, which is now known under the generic term "drug metabolism," was presaged by investigations in the field of chemical carcinogenesis. Pioneer work initiated by the Millers and their associates (reviewed with verve and humanity by J.A. Miller, 1994) with the aminoazodyes revealed that dye residually bound to rat liver microsomes could not be removed by extraction with hot or cold mineral acids or alkali or organic solvents but were released upon digestion of liver proteins by proteases. It was thus concluded that the dye must be bound in liver by covalent bonds. After withdrawal of rats from N,N-dimethylaminoazobenzene (DAB^b), dye disappeared from livers with a time required for plasma (or tissue) drug content to decline to 50% of peak value (t_1/2) of approximately 4 days, which was in accord with the turnover times of most rat liver proteins. In addition, DAB is selective in causing only liver tumors, and no covalently bound dye could be found in other organs. Moreover, very low or undetectable levels of bound dye could be found in livers of mice, guinea pigs, rabbits, or chickens, species that are essentially resistant to the carcinogenic effects of DAB (Miller, 1994).

In approximately 1968 or 1969, B.B. Brodie, while vacationing in Australia, was visiting a sheep station where oral doses of carbon tetrachloride (CCl₄) were routinely administered to the animals as an anthelmintic. He was told that almost all the sheep tolerated a standard dose of CCl₄ well, but an occasional animal died of hepatic necrosis. It was known that pretreatment of sheep with phenobarbital (PB) greatly increased the hepatotoxicity of CCl₄ (Costa et al., 1989). This information promptly ushered into existence, in pharmacology and toxicology, the enormously productive area of the activation of chemically inert drugs and their conversion in vivo into highly reactive metabolites that evoked cyto- and histotoxicity either by forming covalent bonds with critical cellular macromolecules or by stimulating the formation of highly toxic oxygen metabolites such as hydrogen peroxide (H₂O₂), superoxide anion radical (O₂), hydroxyl radical (·OH), or singlet oxygen (¹O₂). Reactive oxygen species (ROS) usually intitiate toxicity by stimulating lipid peroxidation of highly unsaturated fatty acids (e.g., 20:4, 22:6) present in phospholipids that form an integral part of the biomolecular leaflet present in virtually all biological membranes. Cytotoxins may also act by altering cellular glutathione (GSH), calcium, and energy homeostasis.

If, indeed, ontogeny recapitulates phylogeny, cytochrome P450 enzymes should be practically absent in newborn mammals as well as more phylogenetically primitive animals, such as reptiles and fish (Brodie and Maickel, 1962). This is, indeed, the case.

In evolutionary terms, as animals left sea water as a habitat and began to inhabit the land (fish to amphibians; frogs and reptiles to mammals), they encountered a new population of pernicious substances such as alkaloids, terpenes, and steroids, lipid-soluble compounds present in their new diet of aerial plants. It has been argued on teleological grounds that plants, in turn, evolved to produce these noxious and toxic substances to ward off their herbivorous intruders and thus insure their own survival. Many of these phytotoxins were lipid-soluble, and because the mammalian kidney behaves like a large lipid membrane, mammals had to evolve a mechanism for converting lipophilic compounds into hydrophilic ones that could be readily excreted. In liver, this occurred by evolution of a hydroxylating system, a hydrolytic system, and a conjugation system. In addition to converting a lipophile into a hydrophile, these reactions converted the substrate from one possessing significant biological activity (morphine, atropine, ethanol, and cocaine) to products that were essentially biologically inert. Oxidation (hydroxylation) requires cytochrome P450, nicotinamide adenine dinucleotide phosphate (NADPH), O₂, and NADPH cytochrome c reductase, is the most common primary detoxication mechanism, and occurs with hundreds of drugs. Hydrolysis totally inactivates esters, drugs such as procaine, cocaine, and succinylcholine. Finally, once a functional group (such as hydroxyl or primary amine) has been added or unmasked, conjugation occurs using cosubstrates such as glucuronic acid, sulfate, acetyl groups, GSH, or free amino acids. Many xenobiotics are pharmacologically active but toxicologically inactive. Some of these compounds are metabolized to biologically inert products whose physical properties (e.g., water solubility) allow them to be rapidly cleared from the body in the urine or bile. Biological inactivation of xenobiotics often occurs by oxidation, often hydroxylation, which is followed by conjugation to glucuronides, sulfates, GSH (or other amino acid) derivatives. Occasionally, however, xenobiotic substrates are converted to highly chemically reactive products that can react covalently with vital cellular macromolecules such as nucleic acids, proteins, and lipids and dramatically alter or inactivate the normal physiological (biochemical) function of the critical cellular macromolecule. Chemically reactive foreign compounds may produce the whole plethora of cytotoxicities including dedifferentiation, transformation to benign or malignant tumors, teratogenesis, mutagenesis, or necrosis. In most cases however, the mechanisms and the chemical nature of the adducts formed between reactive intermediates and the tissue macromolecules that produce these pathological changes are unknown.

Early work resulted in the facile conclusion that covalent binding of xenobiotics in tissues resulted a priori in necrosis. These conclusions, however, proved to be overly simplistic. In the case of bromobenzene, it was shown (Monks et al., 1982) that p-bromophenol, a metabolite of bromobenzene, binds covalently in both liver and kidney in vivo but is not toxic to either organ. Thus, it was evident that the amount of covalently bound metabolite, per se, in a tissue could not predict toxicity of any series of compounds. Further, it was found that the antioxidant N,N'-diphenyl-p-phenylenediamine (DDPD) prevented bromobenzene-mediated cell death in cultured hepatocytes without affecting the extent of covalent binding (Casini et al., 1982). Using acetaminophen (Labadarios et al., 1977; Devalia et al., 1982), it was reported that pretreatment of animals with certain chemicals could prevent acetaminophen hepatotoxicity but had little effect on its covalent binding in tissues.

Nonetheless, there is a mountain of data indicating a close correlation between covalent binding of several exogenous compounds and cellular necrosis (Hinson and Roberts, 1992; Conney et al., 1994; Cho et al., 1995). Thus, factors that reduce xenobiotic activation and binding such as CoCl₂, 2-diethylaminoethyl-2,2-diphenylvalerate (SKF 525-A), and piperonyl butoxide also reduce necrosis. In addition, factors that increase binding such as pretreatment with PB or 3-methylcholanthrene (3-MC) tend to increase toxicity. Moreover, covalent binding is highest over frankly necrotic regions and dramatically less over cells that are spared. Treatments that deplete cells of endogenous protectants such as GSH and -tocopherol markedly enhance covalent binding and cellular and tissue toxicity, whereas repletion of tissues with GSH or -tocopherol reduces toxicity.

Thus, it is given that, in general, there is a high degree of correlation between the demonstration of covalent binding of a xenobiotic in an organ or cell type within that organ and cellular necrosis. There exist, however enough discrepancies between covalent binding and toxicity that should be explored to allow the careful scientist to tease apart and to reveal more precisely the biochemical mechanisms of how and why these correlations exist.

The purpose of this review is, therefore, to reveal situations in which correlations between covalent binding and toxicity exist and situations in which the two may be divorced.

Type I alveolar cells are elongated and flattened and have corresponding nuclei (fig. 1a). Figure 1a also shows a capillary that is enclosed by a vascular endothelial cell and an erythrocyte (original magnification × 13,000). Type I cells have a limited cytoplasmic volume and a distinct paucity of organelles; they are attached to the basal lamina that they share with the capillary endothelium. Approximately 95% of the rat lung alveolar surface is covered by type I cells, the remainder being covered by type II cells (Crapo et al., 1980).

View larger version (170K): [in this window] [in a new window]   Fig. 1.   (a) An electron micrograph showing a type I alveolar epithelial cell (original magnification × 13,800) and a capillary endothelial cell containing an erythrocyte. The type I cell is a long, flat cell with a large, flat nucleus containing sparse cytoplasmic organelles including few mitochondria, RER, ribosomes, and lysosomes. The endothelial cell cytoplasm is virtually devoid of organelles. Reproduced with permission from Kuhn (1976). (b) An electron micrograph showing a type II alveolar epithelial cell (original magnification ×17,000). The cytoplasm contains mitochondria, RER, numerous free ribosomes, and a Golgi apparatus. The most prominent feature of the type II cell is the electron-dense osmiophilic lamellar bodies thought to be involved in surfactant biosynthesis and/or storage. Reproduced with permission from Kuhn (1976). (c) An electron micrograph of a human pulmonary alveolar macrophage (original magnification × 13,500). The cytoplasm is richly endowed with lysosomes and contains numerous mitochondria, RER, and ribosomes. Reproduced with permission from Kuhn (1976). (d) An electron micrograph of a ciliated bronchiolar epithelial cell from mouse lung (original magnification × 12,000). Numerous mitochrondria and free ribosomes are seen. Reproduced with permission from Boyd et al. (1980b). (e) A scanning electron micrograph of a rat lung bronchiole (original magnification × 2,400) showing Clara cells (domes) as apical bulges amid the grain-like stalks of ciliated cells. Reproduced with permission from Gandy et al. (1983). (f) An electron micrograph of Clara cells of rat bronchiole (original magnification × 2,700). The most prominent organelle is the extensive network of (mainly tubular) SER particularly in the apex of the cells. This is known to be the site of CYP450 enzymes. Other characteristic organelles, particularly in the dome of Clara cells, are numerous electron-dense granules that are not mucin and do not contain phospholipid (surfactant). Reproduced with permission from Boyd et al. (1980b).

Type II alveolar epithelial cells have a more complex cytoplasm consisting of rough endoplasmic reticulum (RER), free ribosomes, mitochondria, and Golgi and multivesicular bodies (Kuhn, 1976; fig. 1b; original magnification ×17,000). The dominant organelles are the electron-dense osmiophilic lamellar bodies, thought to be the biosynthetic or storage sites of pulmonary surfactant. RER and ribosomes are also present. The lamellar bodies consist primarily of the dipalmitoyl ester of phosphatidylcholine, which is highly surface-active, and probably keep the alveoli from collapsing during expiration. Also abundant are enlongated mitochondria, RER, ribosomes, a small Golgi apparatus, and multivescular bodies. Esterases and acid phosphatases have been reported; activities of CYP450 enzymes and conjugating enzymes are measurable in type II cells but are much lower than in Clara cells.

Pulmonary alveolar macrophages can be found in the interstitium or alveolar spaces and have numerous villous projections on their outer surfaces. They exhibit short segments of RER along with ribosomes and small mitochondria. Numerous lysosomal granules fill the cytoplasm and are endowed with a plethora of hydrolytic enzymes such as esterases, acid phosphatases, lysozyme, cathepsins, RNase, elastase, hyaluronidase, and -glucuronidase, a complement to degrade almost any necrotic cell (fig. 1c).

Capillary endothelial cells (fig. 1a) consist of a large basal nucleus, a lumen, and an exquisitely thin double membrane that forms the barrier for diffusion of oxygen from the alveolar space to the blood capillary (Ryan and Li, 1993). Recent work has revealed metabolic and uptake functions for endothelial cells for endogenous and exogenous amines and angiotensin converting enzyme (ACE) is known to be localized on the vascular surface of the capillary endothelial cells.

Ciliated bronchiolar epithelial cells have been characterized physiologically, strangely by the fact that they are not Clara cells. Little experimental data have accumulated on the cytochemistry or enzymology of ciliated bronchiolar cells, and methods have not been devised to purify them in high yield (fig. 1d).

Nonciliated bronchiolar epithelial (Clara) cells (figs. 1e,f) are cytologically and enzymatically heterogeneous (see Plopper et al., 1991b; Cho et al., 1995). Clara cells exist in somewhat different functional forms throughout the respiratory tree (in nasal mucous, trachea, bronchi, and bronchioles). In most animal species, however, they tend to be concentrated in the bronchioles. The most conspicuous morphological characteristic of the Clara cell is its bulbous configuration, the bulb or apex projecting into the bronchiolar lumen. In most species, the apex is rich in smooth endoplasmic reticulum (SER), known to be the site of P450 isozymes. More than 40% of the cytoplasmic volume is composed of SER. CYP450 isozymes are concentrated 5- to 10-fold in Clara cells over other pulmonary cell types and enzymes that conjugate xenobiotics with GSH, glucuronic acid, or sulfate less so. RER is generally restricted to the basolateral portions of the cytoplasm surrounding the nucleus. The nucleus is centrally placed. The apical portion of the cytoplasm contains a small number of ovoid, electron-dense secretory granules whose function is not known. Typical mitochondria and Golgi apparatus are found. Glycogen may or may not be present (Gould et al., 1972).

	II. Hyperoxic Lung Injury

Top Previous Next References

Attempts to specifically delineate the causative agent responsible for O₂-induced lung damage have not been rewarding. In vitro, oxygen-free radicals (O₂ and ·OH) and other metabolites [¹O₂, H₂O₂, and hypochlorous acid (HOCl)] are known to produce cytotoxic damage to isolated lung cells. Many of these oxygen-derived products are interconvertible in vivo and in vitro, and the widespread use of enzymes and their substrates do not inspire confidence in this area of investigation. Examples of such enzymes are substrates are: xanthine and xanthine oxidase thought to general O₂; glucose and glucose oxidase thought to produce H₂O₂; and a witch's brew of other substances, such as myeloperoxidase, lactoperoxidase, and galactose oxidase (which are thought to produce specific metabolic products that cannot be measured directly but rely on such crude measures as chemiluminescence) or by "specific radical scavengers," such as mannitol and thiourea.

Thus, until such time when methodology is available possessing both sensitivity and specificity for assays of single, discrete reactive oxygen metabolites, it is not considered useful to attempt to distinguish oxygen-free radicals and other metabolites (see last paragraph). Therefore, for purposes of this review, oxygen will be defined as O₂, O₂, ·OH, ¹O₂, H₂O₂, and HOCl, plus any underdiscovered forms of oxygen formed in vivo.

When adult rats are exposed to 100% O₂ at 1 atmosphere (atm) pressure, almost all the animals die within 60 to 72 h. Death is preceded by dyspnea, a bloody muzzle, alveolar hemorrhage, pulmonary edema and hypoxia. Prior exposure of rats to 85% O₂, however, allows animals to adapt, and after 5 to 7 days at 85%, they can be transferred to 100% O₂ and survive for long periods of time in pure O₂ (Crapo et al., 1980). Early work on the effects of pure O₂ on the lungs of monkeys (Macaca mulatta) for periods of up to 12 days (Kapanci et al., 1969) traced both the time course and fine structural changes that characterize O₂ toxicity. Electron micrographs of control animals (exposed to air) revealed that the alveolar wall is lined by two types of epithelial cells (type I and II) and the capillaries by endothelium. The endothelium overlies a basement membrane that fuses with the epithelial cells. Therefore, the alveolar membrane that separates blood from air consists of a "sandwich" composed of type I (or type II) cells, the basement membrane, and the capillary endothelial cells. Endothelial cells possess sparse cytoplasmic structures. The alveolar epithelium consists of type I cells that are broad and flat and possess few organelles and thin cytoplasmic extensions. Type I cells are cytologically similar to endothelial cells. Type I cells cover approximately 95% of the alveolar surface. Type II cells have extensive cytoplasm and organelles, mitochondria and osmiophilic lamellar bodies that are thought to be the biosynthetic precursor and storage form of pulmonary surfactant are prominent (Frank and Massaro, 1979).

Four days of O₂ exposure (Kapanci et al., 1969) severely damaged the alveolar walls. Ninety percent of the type I cells were necrotic, and some exfoliated, leaving a denuded basement membrane covered by fibrin strands. Seven days of O₂ exposure revealed a starkly different picture. Type I cells were replaced by primitive type II cells containing fewer lamellar bodies than controls. Type II cells now constituted 95% of the alveolar epithelium as compared with 5% in controls. Type II cells contained numerous mitotic figures. In essence, then, type I cells had undergone necrosis and were replaced by hyperplasia of type II cells that resulted in a thickening of the alveolar epithelium. Of equal importance, endothelial cells had been reduced to approximately 50% those of controls, suggesting massive endothelial cell necrosis.

Using a combination of morphometric and biochemical techniques, Crapo et al. (1980) reported that after 60 h of exposure of rats to 100% O₂, the most significant injury occurred to the capillary endothelium. The latter cell type was reduced in number by approximately 45%. In contrast to Kapanci et al. (1969), Crapo et al. (1980) found no significant changes in type I cells in total volume, total number, and mean surface area after exposure to 100% O₂ for 60 h (96% of a LT₁₀₀). Therefore, Crapo et al. (1980) attributed death to lethal changes in endothelial cells and alterations in type II cells, consisting of swelling of RER or granular endoplasmic reticulum and mitochondria, and ruputure of the plasma membrane and pynknotic nuclei. Lethality was not reflected in significant changes in total superoxide dismutase (SOD), SOD containing copper and zinc as part of the molecule, or levels of SOD containing manganese as part of the molecule.

Employing a different experimental design, Adamson et al. (1970) exposed adult male mice to 90% O₂ for 7 days. Most (90%) of the animals developed respiratory distress and died, but the remainder survived to make a complete recovery on return to air. The earliest ultrastructural change in poisoned animals was focal cytoplasmic swelling of capillary endothelial cells. Animals from the group that died displayed progressive cytoplasmic swelling of type I epithelial cells. This caused disintegration of the basement membrane. Type II cells were well preserved. Capillary distension was a prominent feature. In the 10% of mice that survived, destruction of type I epithelium did not occur. Type II cells were normal until the sixth day of 90% O₂, at which time no lamellar bodies were observed, and the mitochondria displayed swelling and polymorphism. Two days after return to air, focal hyperplasia of type II cells was observed. Surfactant in dying animals was either destroyed or inactivated.

Barry and Crapo (1985) reported that the pulmonary capillary endothelium was a primary target for injury in animals exposed to hyperoxia at normobaric pressure. A key factor in death or survival associated with exposure to rats to either 85% or 100% O₂ was maintenance of the pulmonary capillary bed (endothelium). Adult Swiss mice were exposed to 90% O₂ for periods ranging from 24 h to 14 days, and the only pathological lesions were consistent ultrastructural alterations in the capillary endothelium (Bowden et al., 1968). No consistent alterations were noted in alveolar epithelium; the number of type I alveolar cells did not differ from controls.

In a carefully controlled study of the effects of 99% O₂ on the fine structure of adult male rat lungs (Kistler et al., 1967), it was reported that the most profound alteration was a 50% loss in capillary endothelial cells. Rats were exposed to O₂ for 6, 24, 48, and 72 h at 1 atm pressure. The basement membrane, i.e., the air-blood barrier, approximately doubled in thickness. In contrast to the striking destruction of capillary endothelium, the vast majority of epithelial type I and II cells revealed normal fine structure. A functional correlate of the dramatic loss of capillary endothelial cells, thickening of the basement membrane, and reduction of gas-diffusing capacity, the authors noted marked dyspnea and cyanosis of animals returned to room air. This study clearly identified capillary endothelium as the primary pulmonary target of hyperoxia. Similar findings were reported by Crapo et al. (1978), who investigated fine structural aspects of adaption to hyperoxia. When adult rats were exposed to 100% O₂, nearly all died within 60 to 72 h, whereas animals exposed to 85% O₂ survived and after 5 to 7 days could be transferred to 100% O₂ and survive for prolonged periods, i.e., they become O₂-tolerant. This study compared the morphometric fine structure of lungs from control rats maintained in room air with those from rats maintained in 85% O₂ for 7 days. The O₂-adapted animals had a normal number of alveolar type I epithelial cells, and a moderate increase (two-fold) in the number of type II cells. A large increase was noted in numbers of interstitial cells (five-fold). The major area of pulmonary damage occurred in the vascular compartment, where entire segments of the capillary bed were lost, and the total number of endothelial cells decreased by 45%. Corresponding reductions were noted in capillary volume.

Thet and coworkers (1983), reported that small doses of endotoxin protected rats against O₂-induced lung injury. Bacterial endotoxin (Salmonella typhimurium lipopolysaccharide) was injected into rats that were then exposed to 100% O₂ for 72 h. At the end of O₂ exposure, seven of the eight endotoxin injected rats were still alive as opposed to one of the eight saline-injected, O₂-treated rats, thus confirming the protective effect of endotoxin against O₂ toxicity. This protection against O₂ lethality was not reflected by alterations in pulmonary fine structure. Capillary endothelial cells were reduced approximately 40% by 100% O₂, and this change was not attenuated by endotoxin. Capillary surface area, increased cellularity, and marked edema of the interstitium produced by O₂ were not influenced by endotoxin. Types I and II epithelial cells in hyperoxic animals were not significantly altered by endotoxin. The authors concluded that although the gross toxicity (survival) produced by endotoxin against O₂, poisoning was not accompanied by dramatic changes in lung ultrastructure; this protection might be reflected in enzymatic changes that parallel O₂ toxicity.

Harrison (1971) reminded us that the air-blood barrier consists of three components: the alveolar epithelium (type I and type II cells), the basement membrane, and the capillary endothelium. She studied the effects of 100% O₂ (1 atm) in young rats (90-100 g) for 3 to 6 weeks (immature rats are relatively resistant to high O₂ tensions) and described ultrastructural changes. In the first stage of O₂ poisoning, the mitochondria of epithelial and endothelial cells appeared swollen, and the cristae were compressed or destroyed. The cisternae of the endoplasmic reticulum were also swollen. The basement membrane became edematous and thickened. At the same time, discontinuities occurred in the endothelium. As damage progressed, the endothelial lining disappeared, and the basement membrane began to disintegrate, leaving the capillary space in direct contact with the epithelial cells. Soon, all that remained was a strand of epithelium and a few shreds of basement membrane that ultimately ruptured, and the destruction of the alveolar-capillary wall was complete. In addition to total loss of this barrier, alveolar spaces coalesced and accumulated exudate that contained cellular debris and myelin figures. Gas diffusion became impossible. Thus, destruction of the air-blood barrier started with ultrastructural changes in the endothelial cells that line the capillary bed.

In an analysis of the literature up to 1979, Frank and Massaro (1979), concluded that the pulmonary capillary endothelial was the first lung cell type to be seriously damaged by hyperoxia. I concur with this analysis, and a predominant systemic effect should be hypoxemia.

B. Pulmonary Enzymes That Alter Oxygen Toxicity

1. Enzymes that promote formation of reactive oxygen species. In mitochondria, oxygen undergoes stepwise and sequential one-electron reduction with the formation of water. These reactions are catalyzed by a series of enzymes collectively termed "cytochrome oxidase," and the reactions are exquisitely sensitive to cyanide and other inhibitors. These reactions are tightly linked to oxidative phosphorylation. In addition to this normal function of oxygen in intermediary metabolism, it may undergo a second type of metabolism that produces ROS that range from slightly to enormously cytotoxic (Bishop et al., 1984; fig. 2). O₂ and ·OH are radicals, and ¹O₂ and H₂O₂ are not, but these four species collectively are known as ROS (Halliwell and Gutteridge, 1984). ROS can enter into a wide variety of toxic reactions, some functional (e.g., strand scission of DNA and enzyme inhibition) and some structural (lipid peroxidation of polyunsaturated fatty acids in phospholipids that constitute most biological membranes (Bishop et al., 1984). Table 1 shows the cellular consequences of oxygencentered free radicals. ROS can react with protein, lipid (membranes), DNA base pairs, and the deoxyribose phosphate backbone of DNA. Lipid peroxidation can be cytotoxic as chain reactions can be initiated (Comporti, 1987).

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Reduction of oxygen to water. Tight coupling to oxidative phosphorylation is not presented. Adapted with permission from Bishop et al. (1984).

2. Enzymes that catalyze inactivation of reactive oxygen species. The chemistry and biochemistry of ROS have been discussed in numerous reviews (Freeman and Crapo, 1982; Kehrer, 1993; Doelman and Bast, 1990; Halliwell and Gutteridge, 1986; Halliwell and Gutteridge, 1984; fig. 2) and will not be detailed further here. Some of the biological effects of ROS are presented in table 1. Many of these effects result in cytotoxicity.

3. Role of cytochrome P450 and glutathione in pulmonary oxygen toxicity. A major tool in understanding the mechanism of cytochrome P450 has been the use of various mouse strains that are either "responsive" or "unresponsive" to its inducers, benzo[a]pyrene (BP) or 3-MC (Okey et al., 1989). Lungs of "responsive" animals usually respond with 2- to 4-fold induction of cytochrome P450 but are unresponsive to the inductive effects of PB as are lungs of most other species that have been examined (Gelboin, 1993). Gonder et al. (1985) studied pulmonary O₂ toxicity in genetically responsive (to aromatic hydrocarbons) and unresponsive mouse strains. Hydrocarbon responsive (C3H/HeJ) and unresponsive mouse (DBA/2J) strains were tested for their susceptibility to O₂ toxicity by exposing these and other strains to O₂ (>98%) for periods up to 96 h. DBA/2J mice lived significantly longer than C3H/HeJ animals in hyperoxia (122 h versus 92 h). Lung microsomal cytochrome P450 increased approximately 3.5- to 4-fold in responsive mice after 72 h of hyperoxia but was not induced in unresponsive animals. Histological evaluation of lungs revealed no significant alterations in DBA/2J animals, whereas lungs of C3H/HeJ mice displayed profound septal thickening and pulmonary edema at 96 h. The authors concluded that O₂ toxicity parallels the genetic control of cytochrome P450 induction and that O₂ toxicity develops at a time when O₂ has induced cytochrome P450.

4. Age and species differences in the susceptibility to oxygen toxicity. An interesting ontogenetic pecularity has been revealed in that neonatal rats are quite resistant to the pneumotoxic effects of prolonged hyperoxia compared with adults. This was explained by Stevens and Autor (1980) with the observation that lung tissue of neonatal rats was capable of a rapid, oxygen-induced enhancement of SOD activity (see fig. 2). This response was age-dependent, the maximum effect occurring in 10-day-old animals, and the enzyme change resulted from increased oxygen-mediated protein synthesis. Accordingly, cycloheximide, an inhibitor of protein synthesis, inhibited the oxygen-enhanced incorporation of [³H]leucine into semi-purified SOD. The activities of both mangano-SOD and the cupro-zinc SOD were induced in lungs of oxygen-exposed neonates. Catalase and GSH-Px were also rapidly induced in the lungs of these animals during hyperoxia. The enzymatic response to hyperoxia thus consists of induction of four antioxidant enzymes. A correlation was found between age-dependent tolerance to hyperoxia and enzyme induction in young rat lungs. In hyperoxic animals, SOD activity peaked abruptly at 10 days of age, whereas both GSH-Px and catalase activities increased slowly from birth to peak at 20 to 25 days of age and then fell abruptly. By 34 days of age, little or no oxygen-mediated enzyme responses were detectable; similarly, rats older than 34 days were no longer resistant to oxygen-provoked lung damage. These observations provided compelling evidence for the hypothesis that the resistance of neonatal rats to hyperoxia may result from rapid induction of pulmonary enzymes that catalyze the detoxication of ROS.

	III. Paraquat and Nitrofurantoin-Induced Lung Damage

Top Previous Next References

A. Paraquat Lung Toxicity

1. Morphological effects. The morphological effects of a single dose of paraquat have been reported and are well documented (Smith and Heath, 1976). They occur in humans and in most common laboratory species except rabbits; to produce "paraquat lung" in rabbits requires subacute or chronic dosing (Smith et al., 1979). The lung is the primary target of paraquat toxicity, but the kidney, liver, and thymus are also affected. The dose-response curve for the pneumotoxicity of paraquat in rats is quite steep: a dose of 30 mg/kg was lethal to all animals, whereas 20 mg/kg produced no pulmonary lesions (Vijeyaratnam and Corrin, 1971). Thus, 25 mg/kg administered intraperitoneally produced the following effects. Lungs of rats dying within 4 days were from grossly dyspneic animals. They were markedly congested and plum-colored. There was perivascular edema and hemorrhage into the alveolar spaces. Frequent mitoses were observed in alveolar cells and large amounts of interstitial infiltration with large mononuclear cells and lymphocytes were noted. At 14 days, the cellularity of the alveolar walls was further increased by profibroblasts and fibroblasts, the cells were now arranged in whorls, and large increases in interstitial and intraalveolar reticulin and especially collagen were noted. Ultrastructurally, 3 days after paraquat, there was complete loss of alveolar epithelium in both type I and II cells. In contrast to these dramatic changes in epithelial cells, the capillary endothelium was unaltered. Therefore, large areas of the alveolar surface consisted merely of basement membrane and endothelium. In animals injected with paraquat and sacrificed shortly antemortem, lungs did not collapse upon thoractomy, appeared nearly solid and rubbery, and sank when placed in water. Upon cross-section, normal lung architecture was not recognizable, interstitium was markedly hypercellular and thickened, alveolar spaces obliterated mainly by the influx of fibroblasts, with their abundant RER and their extensive forest of secreted collagen fibers. Incubation of lung explants from paraquat-intoxicated rats with [³H]hydroxyproline revealed a 10- to 20-fold increase in collagen biosynthesis (Greenberg et al., 1978). In those animals that survived, type II epithelial cells repopulated the alveoli, and some of these cells redifferentiated into type I cells (Smith and Heath, 1974). BALB/c mice were exposed to a paraquat aerosol and killed 1 to 28 days later. Initial necrosis and sloughing of the bronchiolar and alveolar epithelium with intact endothelium were followed by type II cell hyperplasia, fibroblast proliferation, and increased synthesis of collagen. Thus, inhaled paraquat produces pulmonary fibrosis that resembles that administered systemically (Popenoe, 1979).

2. Biochemical pharmacology. Paraquat is selectively concentrated in mammalian lungs as are many basic amines (Bend et al., 1985). After intravenous administration to rats, lung/plasma ratios for paraquat were approximately 17:1 at 5 h, 31:1 at 12 h, and 42:1 at 24 h (Witschi et al., 1977). Similarly, incubation of lung slices with paraquat reveals that slice/medium ratios ranging from 5:1 to 15:1 can readily be obtained. Accumulation does not occur by covalent binding (Rose and Smith, 1977). Paraquat uptake by lung is an active energy-dependent process that occurs against a concentration gradient and can be blocked by anaerobiosis, cold, iodoacetate, etc., and also by a variety of other amines that are themselves accumulated by lung slices such as 5-hydroxytryptamine, norepinephrine, spermine, propanolol, imipramine, chlorpromazine, diphenhydramine, and chlorphenterine (Bend et al., 1985). It is interesting to note that the administration of most of these amphiphilic amines does not produce the pulmonary fibrosis characteristic of paraquat but instead causes pulmonary phospholipidosis. In this disorder, apparently not a disease, per se, one observes a dramatic proliferation of osmiophilic lamellar bodies in the cytoplasm of pulmonary epithelial type II cells. This increased synthesis of lamellar bodies is followed by their secretion, exocytosis, into the alveolar space where it is attended by some blockade of the alveolar space and a dramatic increase in pulmonary surfactant that can be removed by lavage. These lamellar bodies are thought to be precursors of pulmonary surfactant that prevent the alveoli from collapsing upon expiration. Pulmonary surfactant is composed of dipalmitoyl phosphalidyl choline with some associated proteins.

2·

2·

2·

Nitrofurantoin is an agent that has been used clinically to treat urinary tract infections. Its use, particularly when administered chronically, is accompanied by pulmonary reactions, once thought to be "hypersensitivity reactions" (Sovijarui et al., 1977), ranging from cough and dyspnea to infiltrates, effusion, and pulmonary fibrosis, which were confirmed by pulmonary biopsy. Administration of large doses of nitrofurantoin to rats produced severe respiratory compromise that caused death in 12 to 36 h. Tachypnea and cyanosis were conspicuous symptoms; at autopsy the lungs were grossly distended, edematous, and hemorrhagic. Histologically, there was widespread interstitial and alveolar edema, vascular congestion, and hemorrhagic consolidation (Boyd et al., 1979). Perhaps more importantly, the lethality of nitrofurantoin could be manipulated broadly by factors known to be involved in lipid peroxidation. For example, the lethality and pulmonary damage produced by nitrofurantoin were significantly enhanced in rats maintained in 100% oxygen after drug administration (Boyd et al., 1979) and in rats maintained on a vitamin E deficient diet that was rich in polyunsaturated fat (corn oil). The LD₅₀ of nitrofurantoin in the oil group was 35 mg/kg compared with 400 mg/kg in controls. If vitamin E-deficient rats were repleted with vitamin E, the lethality of nitrofurantoin returned to that of controls.

These findings should be integrated with those of Mason and Holtzman (1975) that under aerobic conditions and in the presence of NADPH and microsomes catalyze a one-electron reduction of the nitro group of nitrofurantoin to yield a nitro free radical (R-NO₂) that spontaneously reacts with oxygen to regenerate the parent nitro compound and reduces oxygen to O₂. Although covalent binding of nitrofurantoin can be demonstrated under anerobic conditions in vitro, under aerobic conditions little covalent binding occurs but large amounts of superoxide are formed in the presence of lung microsomes, NADPH, and nitrofurantoin. Thus it seems likely that the pneumotoxicity of nitrofurantoin is mediated through superoxide and its secondary metabolites H₂O₂ and ·OH. The mechanism of toxicity therefore seems to bear a strong resemblance to that of paraquat and high concentrations of oxygen itself (Martin, 1983).

	IV. Pulmonary Neoplasia Associated with Benzo[a]pyrene

Top Previous Next References

BP is but one of many polycyclic aromatic hydrocarbons that rely on metabolic activation to produce ultimate carcinogens in vivo (Conney et al., 1994). Administration of BP to mice produces neoplastic changes in lung which may be benign (adenomas) or malignant (adenocarcinomas).

The basic hypothesis of chemical carcinogenesis is the formation of a covalent bond between an exogenous chemical and DNA and is the essential first step in the tumor initiation process (Conney, 1982). Most chemicals require metabolic activation to bind covalently to cellular macromolecules, and the ultimate reactive species, which are electrophilic in character, react with nucleophilic groups of cellular macromolecules to form adducts.

Among the earliest clear demonstrations of the metabolic activation of a xenobiotic to a chemically reactive intermediate was that by Miller (1951), who applied [³H]BP to mouse skin and found covalent binding of its metabolites to skin proteins. Similarly, Brooks and Lawley (1964) painted [³H]BP onto the shaved skin of mice and recovered the radiolabel covalently bound to DNA. The extent of binding of [³H]BP to DNA but not to protein correlated with carcinogenic potential. Later, Grover and Sims (1968) reported that rat liver microsomes incubated with exogenous DNA, NADPH and [³H]BP resulted in covalent binding of radioactivity to DNA. Gelboin (1969) confirmed this finding and further reported that pretreatment of rats with 3-MC increased the binding of [³H]BP to DNA by 2- to 4-fold. He also reported that although metabolic activation of [³H]BP was required, the metabolites of BP were sufficiently stable so that the same level of DNA-bound radioactivity was found whether the DNA was added before or after the incubation. Purification of the DNA in CsCl gradients revealed that the [³H] BP-derived radioactivity was, in fact, bound covalently to DNA.

The efforts of many workers have focused on the chemical nature of the ultimate carcinogenic species. Kapitulnik et al. (1978) studied the pulmonary tumorigenic activities of BP (±)-trans-7,8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (diol epoxide 1), (±)-trans-7,8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (diol epoxide 2), (±)-trans-7,8-dihydroxy-7,8-dihydrobenzo[a]pyrene (BP 78-dihydrodiol), and the tetraols derived from the hydrolysis of diol epoxide 2 in newborn Swiss-Webster mice. Animals were injected intraperitoneally with 4 nmol of each compound on day 1 of life, 8 nmol on day 8 and 16 nmol on day 15, and were killed at 28 weeks of age. Diol epoxide 1 was highly toxic in newborn mice, and most of the animals treated with this compound died before weaning (25 days). Histological examination failed to reveal the cause of death. Diol epoxide 2 and BP 7,8-dihydrodiol were approximately 40 and 15 times more active than BP in causing pulmonary adenomas. The tetraols derived from diol epoxide 2 did not induce pulmonary adenomas (table 2). These data showed that BP diol epoxide 2 derived from BP 7,8-dihydrodiol is a highly active pulmonary carcinogen in newborn mice.

Because the studies of Kapitulnik et al. (1978) were conducted with optical racemates, the next logical step was to investigate the tumorigenicity of the pure stereoisomers. This study was conducted by Buening et al. (1978). BP and each of the enantiomers of the diastereoisomeric BP 7,8-diol-910-epoxides derived from trans-78-dihydroxy-78-dihydrobenzo[a]pyrene were tested by intraperitoneal (ip) administration to mice of 12 and 4 nmol or 24 and 8 nmol of each compound on days 18 and 15 of life. The animals were killed at 34 to 37 weeks of age. (+)-7,8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene [(+) diol epoxide 2; compound No. 6] had exceptional tumorigenicity, whereas BP and the other optical isomers had little or no activity (table 3).

More recently (+)diol epoxide 2 has been found (Jeffrey et al., 1977) as a major adduct bound to DNA and RNA of bronchial explants after treatment of the cultures with BP. Also (+)diol epoxide 2 is the major enantiomer formed stereoselectivity in vivo from BP. It is concluded that (+)-BP-7,8-diol-9,10 epoxide 2 is a major ultimate carcinogenic metabolite of BP in mice (fig. 3). These oxidative steps are thought to be catalyzed by cytochrome P450 1A1 and others (Foth, 1995).

View larger version (13K): [in this window] [in a new window]   Fig. 3.   (+) BP-7,8-diol-9,10-epoxide 2, an ultimate carcinogenic metabolite of BP formed in mouse lung. Reproduced from Buening et al. (1978).

Ross et al. (1995) injected male mice with BP in doses ranging from 5 to 200 mg/kg and sacrificed them 1 to 21 days later. DNA was purified from whole lung homogenates and lung adenomas were counted 240 days after BP injection. DNA adduct analysis was conducted using ³²P postlabeling; time-integrated DNA adduct levels (TIDAL) were determined at each dose level. A strong correlation was found betwen lung ademona induction and the TIDAL levels. After BP administration, the major adduct isolated from lung was N (2)-(10,7,8,9-trihydroxy-7,8,9,10-tetrahydrobenzo[a]pyrene]yl deoxyguanosine, from the binding of the diolepoxide of BP to deoxyguanosine. The induction of adenomas as a function of TIDAL values suggests that the formation and persistence of DNA adducts determines the carcinogenic potency in a series of PAH. This relationship had been predicted earlier (Harvey, 1982) with deoxyguanosine as the DNA adduct. Earlier (Feldman et al., 1980) work had reported that reaction of the racemic ± BP diolepoxide with cultured human lung cells (A549) resulted predominately in the formation of the deoxyguanosine adduct.

Recently, the distribution of BP diol epoxide adducts along exons of the p53 gene were examined in HeLa cells and bronchial epithelial cells was mapped at nucleotide resolution (Denissenko et al., 1996). Strong and selective adduct formation occurred at guanine positions in codons 157, 248 and 273. These same positions are the major mutational hotspots in human lung cancers. Thus, targeted adduct formation rather than phenotypic selection appear to shape the P53 mutational spectrum in lung cancer.

Only (+) diol epoxide 2 (also referred to as the R, S,S, R enantiomer) demonstrated exceptional mutagenic activity in Chinese hamster V-79 cells, approximately eight times that of the other three isomers (Conney et al., 1994) which was in general accord with its higher tumorigenicity in newborn mouse lung (50 to 100 times) (Buening et al., 1978). The mutagenicity of three concentrations of (+) BP diol epoxide 2 [(+)BPDE2] was evaluated at the hypoxanthine (guanine) phosphoribosyntransferase (HPRT) locus in V-79 cells. Exposure of these cells to low (0.01 µM), intermediate (0.04-0.10 µM) or high (0.30-0.48 µM) concentrations of (+)BPDE2 resulted in 97%, 100% and 32% cell survival and mutation rates were increased 9-, 51- and 513-fold respectively compared with solvent (dimethyl sulfoxide) controls. Surviving colonies were isolated, complementary DNAs (cDNAs) were prepared, amplified by the polymerase chain reaction, and the coding region of the HPRT cDNA was sequenced (Conney et al., 1994). Base substitution mutations in the coding region of the HPRT gene were most common and were found in approximately 70% of the mutant clones. Exon deletions were observed in approximately 25% of the mutants and frame shift mutations were noted in approximately 5% of the mutant clones. Reducing the concentration of (+) BPDE2 decreased the proportion of base substitutions at guanine and cytosine base pairs (GC) and increased substitutions at adenine-thymine base pairs (AT). At the high concentration, 7 of 120 base substitutions occurred at AT (6%) and 113 at GC (94%). At the intermediate concentration, 20 of 82 base substitutions occurred at AT (24%) and 62 at GC (76%). At the low concentration, 27 of 76 base substitutions were at AT (36%) and 49 were at GC (64%). An evaluation of the frequency of mutations at specific bases in the coding region of the HPRT gene indicated differences in the profile of hot spots induced by different concentrations of (+) BPDE2. Eleven hot spots were observed at high concentrations, 7 hot spots at the intermediate concentration and 6 hot spots were noted at the low concentration. Thus, at a high cytotoxic concentration of (+) BPDE2, most base substitutions were at GC (predominantly GC  AT transversions), with only an occasional mutation at an AT. As the concentration of (+) BPDE2 was reduced, there was an increase in the base substitution at AT and a corresponding decrease in substitutions at GC (Conney et al., 1994).

	V. Pneumotoxicity of 4-Ipomeanol and Other Furans

Top Previous Next References

4-Ipomeanol [1-(3-furyl)-4-hydroxypentanone] is a natural product produced by the fungus Fusarium solani when it infects sweet potatoes. Cattle consuming such sweet potatoes developed severe, occasionally fatal, pulmonary insufficiency; necropsy revealed severe pulmonary edema and hemorrhage. Most of the published data on the biochemical toxicology of 4-ipomeanol are from the laboratory of Boyd and his coworkers (Boyd, 1976, 1980a).

Administration of 4-ipomeanol to rats, rabbits, guinea pigs and hamsters (Dutcher and Boyd, 1979) produces an organ- and cell-specific lesion, viz, selective necrosis of the nonciliated bronchiolar epithelial (Clara) cells of the lung. Larger doses cause less specific effects in which ciliated bronchiolar cells are also involved and which eventually affect the alveolar epithelium and vascular endothelial cells and larger doses evoke massive intra-alveolar edema and hemorrhage together with pleural effusion. Higher doses produce hepatic and renal lesions. In contrast to rabbits, rats and hamsters, and other species that display pulmonary lesions in response to 4-ipomeanol, adult male mice exhibit renal cortical necrosis as a primary lesion; female mice and immature mice of either sex are markedly resistant to this renal insult. Taking advantage of an interesting phylogenetic peculiarity, Buckpitt et al. (1982a,b) demonstrated that birds (Japanese quail, chickens), whose respiratory tracts lack Clara cells, fail to develop lung damage after 4-ipomeanol. Instead these species develop severe hepatic injury with no evidence of pulmonary involvement. In accord with these findings microsomes prepared from lung and liver of rats, guinea pigs, hamsters, and rabbits catalyzed the metabolic activation and covalent binding of 4-ipomeanol in vitro as did microsomes from kidneys of adult male mice. In each case, metabolic activation in vitro correlated with target organ toxicity in vivo. Similarly chicken liver microsomes activated and covalently bound 4-ipomeanol in vitro, whereas chicken lung microsomes were devoid of activity.

Dutcher and Boyd (1979) looked for possible species and strain differences in the covalent binding of 4-ipomeanol. Three strains of rat (Lewis, Sprague-Dawley, and Fisher-344), Hartley guinea pigs, albino New Zealand rabbits, Golden Syrian hamsters and six strains of mouse (BALB/CJ, C3H-HeJ, NIH Swiss, DBA/2J, C57BL/6J, and A/J) were examined. In all species tested, the lung was the major target for 4-ipomeanol covalent binding and toxicity as determined histologically. In the hamster and the mouse, 4-ipomeanol caused liver and kidney necrosis in addition to pulmonary damage. Corresponding high levels of covalent binding of 4-ipomeanol occurred in liver and kidney in these species.

The lung was the major site of damage in rats, guinea pig, and rabbit and this selectivity correlated with covalent binding. All six mouse strains exhibited some lung damage and striking renal toxicity: renal cortical necrosis occurred together with bronchiolar lesions. Covalent binding of 4-ipomeanol in the six mouse strains was roughly twice as high in kidney as in lung.

Boyd et al. (1975) showed that after administration of [¹⁴C]4-ipomeanol to rats, radioactivity was concentrated in lung (per gram of organ), 90% of which was covalently bound. In this paper, the term "covalent binding" refers to material remaining bound to tissues or cell homogenates after their extraction with strong mineral acids and alkali, numerous hot and cold organic solvents ranging in polarity from methanol to hexane, boiling, and lyophilization. Employing electron microscopic autoradiography, Boyd (1977) demonstrated that after the administration of [¹⁴C]4-ipomeanol to rats, mice and hamsters, electron-dense granules were found specifically localized over Clara cells which later became necrotic. In contrast, the adjacent ciliated bronchiolar cells and other major pulmonary parenchymal cells were neither