Induction of Superoxide Dismutase by Oxygen in Neonatal Rat Lung*

Neonatal rats are markedly resistant to oxygen toxicity when compared with adult animals. This study was designed to understand better some of the underlying biochemical processes in the lung which accompany oxygen exposure of the neonatal rat with the aim of developing an explanation for the oxygen tolerance. Superoxide dismutase activity was determined in neonatal rat lung tissue after either exposure of animals to a normobaric atmosphere of 95% oxygen for 24 h or incubation of excised minced lung in an atmosphere of 100% oxygen for 5 h. Enzyme activity of this oxygen-treated tissue was compared with values obtained from air-exposed lung tissue. Hyperoxic exposure either in uiuo or in vitro produced an increase in total lung superoxide dismutase activity. The increased activity resulting from both types of exposure conditions showed identical characteristics upon column chromatography when compared with the unstimu-lated activity. Cellular integrity was found to be necessary for the effect to occur in the in vitro exposure system since hyperoxic incubation of broken cell preparations did not result in enhanced superoxide dismutase activity. The increase in enzymatic activity was accounted for solely by a change in the mitochondrial manganosuperoxide dismutase. The oxygen-mediated

one such adaptive process. The molecular nature of this special effect has been further studied and is the subject of this report.

Animals
-The Sprague-Dawley rats utilized for these studies were bred and raised at The University of Iowa Animal Care Facility. Mature female rats were exposed to a 12-h breeding period. The onset of pregnancy was timed from the midpoint of this 12-h period. An earlier report from this laboratory showed that the maximum increase in superoxide dismutase activity due to either in viva or in vitro oxygen exposure occurred in the lungs of neonatal rats 10 days after birth (5). Only animals of this age, therefore, were used for these studies.

Oxygen
Exposure-For the experiments in which neonatal rats were exposed in viva to hyperoxia, animals were maintained for 24 h in a controlled atmosphere chamber filled with 95% oxygen. The OX concentration was monitored with a Beckman model OM-11 gas analyzer.
The temperature of the chamber was maintained at 24-26", water vapor pressure was less than 10 mm Hg. The CO, concentration was monitored with a Beckman model LB-2 gas analyzer and kept below 0.6%. After the exposure period, the animals were decapitated and the lungs immediately removed en bloc. The tissue was perfused with ice-cold isotonic potassium phosphate buffer, pH 7.4, blotted dry, and weighed.
In studies in which lung tissue was exposed to hyperoxia in vitro, neonatal rats were killed by decapitation, and the lungs immediately removed en bloc. The tissue was perfused with ice-cold isotonic potassium phosphate buffer, pH 7.4, blotted dry, and minced with a razor blade to a size of 1 to 2 mm3. Lung tissue was obtained from several litters of animals and was pooled prior to each experiment. Minced lung tissue (0.1 g per sample) was then added to incubation vessels containing 1 ml of adult rat plasma and 3 ml of isotonic potassium phosphate buffer, pH 7.4, according to our previously reported procedure (51. The tissue was incubated in either air or 100% 0, at 37" for 5 h. Homogenate Preparation -Potter-Elvehjem homogenates (lo%, wi v) were prepared from both the air-and the O,-incubated tissue samples by three or four consecutive passes through this tissue homogenizer.
Omni-Mixer homogenates were prepared from the above described homogenates using a 50-ml capacity Sorvall Omni-Mixer.
Homogenization was continued for 3 min at O-4" at a speed setting of 6.5. Low speed centrifugation was performed in a Sorvall RC-2B refrigerated centrifuge equipped with a SS34 rotor. High speed centrifugation was performed in a Beckman L-3 centrifuge equipped with a SW 56 rotor. Protein concentration was assessed by the micro-biuret method (10).

RESULTS
An earlier report from this laboratory described an increase in superoxide dismutase activity which occurred in neonatal rat lung tissue after both in uiuo and in vitro exposure to 95 to 100% oxygen (5). Initially, the in vitro exposure studies were conducted on minced lung tissue. To determine whether this increase in enzymatic activity required the presence of whole lung cells, various lung tissue preparations were exposed to either air or 100% oxygen under the above described conditions. Table I shows that, as previously reported, superoxide dismutase activity increased appreciably (43%) when minced lung tissue was incubated under hyperoxic conditions. In contrast to this, however, incubation of homogenized lung tissue prepared with either a Potter-Elvehjem tissue disrupter or a Sorvall Omni-Mixer failed to produce an oxygen-mediated increase in enzymatic activity. Furthermore, incubation of either the supernatant fraction or the resuspended pellet obtained after centrifugation of each homogenate did not produce a significant increase in enzymatic activity. The small increase in activity that was observed after incubation of the 15,000 X g pellet fraction may represent activity from cells which escaped disruption. From these data, we have concluded that maintenance of cellular integrity during hyperoxic exposure is necessary in order that the increase in superoxide dismutase activity will be obtained.
Since superoxide dismutase is known to be present in two different forms in the cells of eukaryotes (16, 171, studies were conducted to ascertain whether the observed increase in activity after oxygen exposure was a result of a change in the cytoplasmic cuprozinc superoxide dismutase activity, the mitochondrial manganoenzyme activity, or both. After in vitro incubation of minced lung tissue, both a Potter-Elvehjem homogenate and an Omni-Mixer homogenate were prepared in order to determine the degree of tissue disruption that was necessary to measure the oxygen-stimulated increase in activity. As shown in Table II, the oxygen-mediated increase in  activity could be demonstrated only in the Omni-Mixer-prepared homogenate. No increase in superoxide dismutase activity above the basal level could be measured when the minced tissue was homogenized by the relatively gentle Potter-Elvehjem procedure. The oxygen-mediated effect had occurred, however, in this latter tissue preparation since further tissue disruption of the Potter-Elvehjem homogenate in the Omni-Mixer yielded preparations in which both the basal and incremental activity was seen. The resuspended pellet obtained after centrifugation of the Omni-Mixer homogenate at 15,000 x g for 15 min (a fraction which roughly corresponds to isolated mitochondria), was the only subcellular fraction in which both the basal and the oxygen-mediated increase in enzyme activity could be detected. The resuspended pellet from the Potter-Elvehjem homogenate showed only the basal activity. However, Omni-Mixer homogenization of this pellet, as with similar treatment of the tissue, yielded a fraction in which the increase in activity could be observed. It appears, therefore, that the oxygen-enhanced enzyme activity is detected only after vigorous homogenization.
Analysis of lung tissue obtained from animals exposed in uivo to either air or 0, gave results similar to those described above.
The two forms of superoxide dismutase differ with respect to their sensitivity to cyanide (16). Cytoplasmic superoxide dismutase is markedly inhibited in the presence of 1 mM cyanide, whereas the mitochondrial enzyme is unaffected. By utilizing this distinguishing characteristic, we sought confirmation that only the mitochondrial superoxide dismutase was affected by hyperoxia. In Table III it can be seen that in the whole tissue homogenate as well as in the crude mitochondrial fraction, the activity increase is not diminished in the presence of cyanide. We conclude therefore that only the mitochondrial enzyme is associated with the oxygen effect.
Previous data from this laboratory reported that the oxygenmediated increase in enzymatic activity was observed regardless of whether the specific activity was calculated in terms of lung weight, lung protein, or lung DNA content (5). These results as well as the observation that the activity of the cuprozinc enzyme is unchanged strongly suggested that oxygen-stimulated cell proliferation was not the cause of the hyperoxic effect. However, oxygen-stimulated mitochondrial proliferation could explain the results. To test this possibility, the activity of certain mitochondrial marker enzymes was determined in Omni-Mixer preparations of minced lung tissue after incubation in 100% oxygen and compared with the activity of similar preparations that were incubated in air. Neither cytochrome oxidase and succinate dehydrogenase (enzymes associated with the mitochondrial inner membrane), nor isocitrate dehydrogenase and glutamate dehydrogenase (enzymes associated with the mitochondrial matrix) change under conditions where superoxide dismutase activity is increased (Fig. 1).
To ascertain whether the increase in superoxide dismutase activity was the result of the synthesis of a new form of the mitochondrial enzyme, a comparison was made of the chromatographic characteristics of basal and oxygen-stimulated mitochondrial superoxide dismutase. After exposure to 95% oxygen or air, either in vitro or in uiuo, excised lung tissue was homogenized with the Potter-Elvehjem homogenizer and the crude mitochondrial fraction prepared. After sonication, the preparation was centrifuged at 105,000 x g for 60 min and the pellet discarded. As seen in Table IV, this procedure released both the basal and the oxygen-stimulated mitochondrial enzyme activity from the membrane.
Ammonium sulfate was added to the supernatant fraction to 60% of saturation and the resultant suspension was centrifuged at 105,000 x g for 60 min after which the pellet was discarded. Following dialysis of the supernatant fraction against 0.05 M Tris/HCl, pH 7.8, the solution containing cyanide-resistant superoxide dismutase activity was applied to a Sephadex G-100 column and the activity eluted with 0.05 M Tris/HCl. Fractions containing the cyanide-resistant superoxide dismutase activity were pooled and dialyzed overnight against 0.005 M Tris/HCl. This crude preparation was subjected to ion exchange chromatography on a column of DEAE-cellulose with an applied Tris/HCl gradient of increased ionic strength from 10 to 100 mM. This procedure represents a modification of that developed by Weisiger and Fridovich (16). The results depicted in Figs. 2 and 3 show that the activity eluted in a single peak at the same ionic strength regardless of whether the initial hyperoxic exposure of the pulmonary tissue occurred in uivo or in vitro. In addition,  b These preparations were obtained by resuspension of the 15,000 x g pellet following centrifugation of homogenized (Potter-Elvehjem), previously incubated minced lung tissue of lo-day-old rats. c The preparation was similar to that described above but subjected to additional homogenization in the Omni-Mixer before enzymatic assays were conducted.
" The preparation was similar to that described in Footnote b but subjected to sonic oscillation four times for 3 min each at a power setting of 20.  Fig. 2, except that lung tissue was excised from lo-day-old rats, minced, and then exposed to either air or 95% oxygen in vitro as described in the text.
inhibitors of RNA-dependent protein synthesis (19). Following hyperoxic pressure, the tissue was homogenized and assayed for superoxide dismutase activity. These results were compared with data obtained from similarly treated tissue incubated in air. Actinomycin D greatly diminished the oxygenmediated increase in superoxide dismutase activity in the in vitro system (Fig. 4). The oxygen effect was completely abolished when the incubation medium contained either puromytin or cycloheximide.
A similar loss of the oxygen effect was seen when cycloheximide was administered intraperitoneally to g-day-old rats prior to in uiuo exposure to hyperoxia. 21% oxygen FIG. 4 (left). Effect of protein synthesis inhibitors on the oxygenmediated increase in pulmonary superoxide dismutase activity in loday-old rat lungs. Minced lung tissue samples, pooled from 16 loday-old rats, were incubated in 100% oxygen in the presence of the indicated inhibitors. The results are reported as the percentage change in enzymatic activity, compared with air-incubated minced lung tissue. FIG. 5 (center). Effect of hyperoxic exposure on leucine incorporation into mitochondrial superoxide dismutase with and without puromycin. Pooled minced tissue from ten IO-day-old rats were used in each experimental group. Minced tissue samples were incubated in 21% oxygen (air) or 95% oxygen in the presence of 10 PCi of hyperoxia in a similar manner (20). In the yeast, both the cuprozinc and the manganoenzyme increased in activity. The results reported here clearly show that in the lung tissue of neonatal rats, a specific protein synthetic capacity is linked to the oxygen tension to which the pulmonary tissue is exposed and that a rapid oxygen-directed synthesis of the mitochondrial superoxide dismutase occurs. This response appears somewhat different than that described for adult rat lung tissue in that (a) it occurs very rapidly (within 24 h) and (b) can be reproduced in excised tissue. Both characteristics make it unlikely that the response could be associated with oxygen-induced, pulmonary reparative-proliferative changes as suggested by Cross and his associates (4).
Since the neonatal lung is known to be very resistant to the toxic effects of hyperoxia (6-81, it is possible that this rapid enzyme synthesis more nearly represents the adaptive and defensive response to hyperoxia described by Fridovich and his colleagues in the prokaryotic and simple eukaryotic systems. In their studies, enhanced intracellular levels of superoxide dismutase, as well as of catalase and peroxidase, were directly associated with protection against oxygen toxicity (1,2,20,21). In higher animals such a defense mechanism would be important primarily to the lung since this organ is one of those directly exposed to the environment. Because certain animals such as rats and humans are abruptly thrust at birth from a semianaerobic environment of the uterus to the relative hyperoxia of air and, because the lung would necessarily be required to cope rapidly with such an environmental change, it is not unreasonable to propose that rapidly responding endogenous protective mechanisms should be present in neonatal lung. Available evidence indicates that of all organs of adult animals tested only the lung is capable of enhanced superoxide dismutase activity following in vivo exposure to hyperoxia (3,22). A similar test has not been conducted, as yet, for neonatal animals.
Although pulmonary morphologic and functional changes accompanying oxygen toxicity have been described (8, 91, the underlying molecular events are not well understood either as 13Hlleucine. Puromycin was added at a concentration of 200 pg/ml. Results are reported as counts per min of 13Hlleucine per unit of mitochondrial superoxide dismutase activity from the semipurified fraction obtained after ion exchange chromatography on DEAEcellulose as previously described. FIG. 6 (right). Effect of hyperoxic exposure on leucine incorporation into mitochondrial superoxide dismutase in the presence and absence of cycloheximide.
Nine-day-old rats were injected intraperitoneally with 50 &i of 13HIleucine 2 h before injection of 0.5 mg of cycloheximide/kg of body weight. In uiuo 24-h exposure to 21% oxygen (air) or 95% oxygen commenced immediately. Results are reported as described in Fig. 5.
to the primary location or the initiating agent. It is possible that both the primary toxic agent and the enzyme inducer could be either molecular oxygen or a product of its metabolism. Our work clearly shows that one event accompanying hyperoxic exposure, that of specific enzyme induction, occurs in the mitochondria of lung tissue. We have no evidence as yet whether this protein synthesis is directed from the nuclear genome or mitochondrial DNA although present evidence favors the former location for the synthesis of mitochondrial superoxide dismutase (17).
Maintenance of viable mitochondria in the presence of potentially damaging agents such as oxygen is important for lung function. Superoxide-free radicals are known to be generated by many enzymes (11,(23)(24)(25) including those located in the mitochondria (26). In the presence of high concentrations of oxygen, the flux of oxygen free radicals generated by enzymatic reactions is elevated (27). It follows then, as has already been suggested for liver cells (261, that pulmonary mitochondrial superoxide dismutase may function to preserve the molecular integrity of the oxygen-metabolizing system against the by-products of this system. In response to hyperoxic exposure, the protein synthetic apparatus of the affected cells may be programmed to elevate the levels of this endogenous protective enzyme, superoxide dismutase, in the appropriate subcellular location. The ultimate effect could be, in concert with other such endogenous protective mechanisms, to diminish the toxic effect of hyperoxia.