Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The increased production of partially reduced oxygen species including superoxide anion radical (O₂ ·), hydroxyl radical (OH·), and hydrogen peroxide (H ₂O₂) during hyperoxia has been postulated to play a major role in causing pulmonary injury (1). Hyperoxia-induced lung damage can be modulated through augmentation of the pulmonary antioxidant defense. Adult rats normally develop pulmonary edema and die within 3 d when exposed to 100% oxygen, but these same animals can survive for prolonged periods of time under this concentration of oxygen if they are first pre-exposed to a sublethal concentration (85%) of oxygen for 5 to 7 d (2, 3). This apparent tolerance to hyperoxia is associated with an increase in the activities of all the major pulmonary antioxidant enzymes including copper-zinc-containing and manganese-containing superoxide dismutases (CuZnSOD and MnSOD, respectively), catalase, glutathione peroxidase, glutathione reductase, and glucose 6-phosphate dehydrogenase (3). Enhancement of pulmonary antioxidant enzyme activity through intravenous injection or intratracheal insufflation of liposomes containing CuZnSOD and catalase also provides protection to animals against hyperoxic insults (9). It has also been shown that a sublethal dose of endotoxin injected intraperitoneally prior to hyperoxia protects rats against subsequent hyperoxic exposure (12). This protection is known to be associated with a rapid and drastic induction of MnSOD gene expression (13), and to a lesser extent an increase in the pulmonary activity of CuZnSOD (16, 17).

Recently, with the aid of recombinant DNA technology, treatment of whole animals with a relatively large quantity of cytokines has become feasible. White and colleagues (18) have demonstrated that intraperitoneal or intravenous injection of both tumor necrosis factor  (TNF-) and interleukin-1 renders adult rats more resistant to oxygen toxicity. In their studies an increase in activities of several lung antioxidant enzymes was found in those rats treated with cytokines and exposed to hyperoxia (19). Furthermore, a single intratracheal administration of TNF- has been found to be sufficient to protect rats from hyperoxia-induced mortality (20). Induction of MnSOD expression by TNF- in lungs of those rats may partly contribute to the protection against the deleterious effects of high concentrations of oxygen (21). All of these observations illustrate the beneficial effects of increased antioxidant enzyme activity in modulating oxygen radical-mediated pulmonary injury.

However, this conclusion is somewhat complicated by the fact that no alteration of lung antioxidant enzyme activity has been found in rabbits developing tolerance to hyperoxia after a brief exposure to 100% oxygen (22). This finding suggests that other cellular factors besides antioxidant enzymes may also play a critical role in defense mechanisms against toxic oxygen species. Moreover, in rats developing tolerance to 100% oxygen, the relative importance of each of the antioxidant enzymes in pulmonary defense against oxygen toxicity has not been established (3). It is not clear whether an increase in the activity of a single antioxidant enzyme or of a combination of several antioxidant enzymes in the lungs is necessary to provide sufficient protection to allow the animals to survive under 100% oxygen.

To further investigate the effectiveness of each antioxidant enzyme in protecting lungs from oxygen toxicity, we have initiated the use of mice as a model. This is based on the fact that adult mice do not develop tolerance to oxygen during a hyperoxic exposure (2), and it is relatively easy to alter the levels of expression of various antioxidant enzymes through transgenic technology. In this report, we demonstrate that an increase in lung MnSOD activity, resulting from the expression of a human MnSOD transgene, provides modest protection to B6C3 transgenic mice when compared with controls during exposure to 90% oxygen but not to exposure to > 99% oxygen.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Construction of Human MnSOD Expression Vector

Two similar MnSOD expression vectors were used in this study. To construct the first expression vector (Figure 1, upper panel), the entire human MnSOD cDNA (23) fragment flanked by EcoRI restriction sites was initially converted to blunt-end by mung bean nuclease treatment, ligated to SalI linkers, followed by digestion with SalI and then insertion into the SalI site of human -actin expression vector pHAPr-1 (24; generously provided by Dr. Larry Kedes of the University of Southern California, Los Angeles, CA). Since there is no convenient restriction site 3' to the SV40 polyadenylation site, the EcoRI-HindIII restriction fragment of construct 1 containing the human -actin promoter, intron 1 of human -actin gene, and the entire MnSOD cDNA was used for generating transgenic mice.

View larger version (13K): [in this window] [in a new window]   Figure 1.   Structures of the human MnSOD transgenes and a summary of production of transgenic mice. The EcoRI-HindIII fragment of construct 1 and EcoRI-XbaI fragment of construct 2 were used to generate transgenic mice. CAP and IVS1 represent transcriptional initiation site and intervening sequence 1 of the human -actin gene, respectively.

Following completion of generation of transgenic mice with construct 1, we then realized that the apparent polyadenylation signal AATAA present in the human MnSOD cDNA may not be sufficient for cleavage of the transcript for addition of polyadenosine. A G-U-rich element, which lies downstream from the polyadenylation site, may also be critical for this process (25). We then constructed a second MnSOD expression vector. The second vector (Figure 1, bottom panel) was constructed by the following strategy: the EcoRI-PstI fragment of human MnSOD cDNA (bases 1-832, without the polyadenylation signal present in the cDNA) was initially adapted to be flanked by SalI restriction site by ligation of linkers as described above, and then cloned into the SalI site of vector pHAPr-1. The EcoRI-HindIII fragment of the resultant plasmid containing the human -actin promoter, intron 1, and MnSOD cDNA was isolated. At the same time, the HpaI site of SV40 sequence at position 2666 in plasmid pMSG (Pharmacia LKB Biotechnology, Piscataway, NJ [position 1988 of SV40]) was converted to HindIII site by linker ligation and the HindIII-PstI fragment containing the polyadenylation site of SV40 early region was isolated. These two DNA fragments were then ligated to EcoRI plus PstI digested vector pKS (Stratagene, La Jolla, CA). The EcoRI-XbaI fragment containing the entire expression construct free of plasmid sequences was isolated and used to establish transgenic mice. All the recombinant DNA procedures were performed according to the methods described by Sambrook and coworkers (26).

Production of Transgenic Mice

Purified DNA at 2-5 µg/ml in 5 mM Tris-HCl pH 7.4, 0.1 mM EDTA was injected into the pronuclei of fertilized eggs isolated from mating of (C57BL/6 X C3H) F1 (abbreviated B6C3 F1) male and female mice. Mouse eggs surviving microinjection were then implanted into the oviducts of pseudopregnant foster mothers (CD1; Charles River Laboratories, Wilmington, MA) following the procedures described by Hogan and associates (27). Transgenic mice were identified by Southern blot analysis of tail DNA probed with the entire human MnSOD cDNA (28). Transgenic founders were bred with B6C3 F1 mice to produce offspring for further studies. Homozygous transgenic mice were obtained by breeding between two hemizygous transgenic mice.

Analysis of Expression of Human MnSOD Transgene

Tissue sample preparation. Transgenic and control mice were anesthetized with pentobarbital sodium. After opening the abdomen and thorax, an incision was made on the left atrium, and the lungs were perfused through the right ventricle with 10 ml ice-cold phosphate-buffered saline (PBS). The lungs were removed, sliced with a pair of scissors to mix tissues from left and right lobes, and then divided into two parts, one for RNA analysis and the other for protein and enzyme activity assays.

RNA blot analysis. Total RNAs of various tissues were isolated by the guanidinium thiocyanate-CsCl methods as described by Chirgwin and colleagues (29). RNA blot analysis was performed according to the procedures described by Thomas (30). RNA blot filter was hybridized at 42°C with ³²P-labeled human MnSOD cDNA in 50 mM sodium phosphate buffer, pH 7.0, containing 50% formamide, 3× saline sodium citrate (SSC) (1× SSC is 0.15 M NaCl and 0.015 M Na citrate), 5× Denhardt's solution, and 200 µg/ml denatured salmon sperm DNA. Wash was performed in 0.1× SSC and 0.1% sodium dodecyl sulfate (SDS) at 50°C.

Protein blot analysis. Lung samples were homogenized in 50 mM potassium phosphate buffer, pH 7.8, containing 0.1% Triton X-100 with a Polytron homogenizer, followed by sonication on ice for 30 s with a microprobe at maximum power. Lung homogenates were clarified by centrifugation at 20,000 × g for 10 min and stored at 70°C. Protein content of lung homogenate was determined with the use of a bicinchoninic acid protein assay kit (Pierce, Rockford, IL).

CuZnSOD and MnSOD activity assay. CuZnSOD and MnSOD activities were assayed by measuring inhibition of xanthine plus xanthine oxidase-mediated cytochrome c reduction at pH 7.8 (31). All measurements were performed in the presence of 10 µM KCN to inhibit the activity of cellular cytochrome c oxidase, which interferes with the assay. Both CuZnSOD and MnSOD are insensitive to this concentration of KCN. To distinguish the contribution of CuZnSOD and MnSOD to the total SOD activity, the same measurement was also repeated in the presence of 1 mM KCN. The MnSOD activity was not inhibited, whereas the CuZnSOD was approximately 90% inhibited by this concentration of KCN. Thus the SOD activity in the presence of 1 mM KCN represents the sum of the residual CuZnSOD activity plus MnSOD activity, which permits calculation of the relative contributions of the CuZn and Mn enzymes to total cell SOD activity.

Immunocytochemical Labeling of MnSOD in Lungs of Transgenic Line TgHMS66

Tissue preparation. Lungs were fixed by intratracheal instillation of the primary fixative (consisting of 4% [wt/vol] paraformaldehyde, 0.2% [wt/vol] glutaraldehyde, 0.5% [vol/ vol] dimethyl sulfoxide, 0.5% [vol/vol] of a saturated aqueous picric acid solution, 0.05 M sucrose, and 0.08 M PIPES, pH 7.4 at 25°C) at 15 cm H₂O pressure. The tissue was fixed for 2-3 h at room temperature or overnight at 4°C. Tissue blocks (1-2 mm) were then post-fixed for 1 h in 2% (wt/vol) glutaraldehyde, 1% (wt/vol) paraformaldehyde, 0.5% (vol/vol) dimethyl sulfoxide, 0.5% (vol/vol) of saturated aqueous picric acid, 0.05 M sucrose, and 0.08 M PIPES, pH 7.4 at 25°C. The tissue was dehydrated in ascending grades of ethanol; infiltrated with an ethanol and LR-White mixture, then with LR-White (medium grade) overnight; and polymerized at 50-55°C for 24-40 h in gelatin capsules. Thin sections (80 nm) were mounted on 200-mesh copper-rhodium grids coated with collodion (7).

Immunocytochemical labeling. The MnSOD was localized by protein A-gold immunocytochemistry (7). Grids were floated on successive drops of the reagents, at room temperature, according to the following schedule: 10 min on 0.15% (wt/vol) glycine in PBS, pH 7.4; 10 min on PBS; 5 min on 1% (wt/vol) bovine serum albumin in PBS (BSA-PBS); 2 h on a 1:100 dilution of antirat MnSOD rabbit antiserum in BSA-PBS; 15 min on PBS; 5 min on BSA-PBS; 30 min on a suspension of 9 nm gold particles in BSA-PBS; and rinsed on PBS and then water. The sections were stained with 4% (wt/vol) aqueous uranyl acetate.

Sampling and quantitation of labeling. Two to three lung samples per animal, and one or two sections per tissue sample were analyzed. Grid squares were scanned from left to right and top to bottom, and the first type II cell profile encountered in each grid square was photographed at ×7,000. Five to ten cell profiles were sampled per section. Micrographs were printed at a final magnification of ×27,000 on 11-by-14-foot polycontrast paper. Gold particles were counted on mitochondria, nucleus, lamellar bodies, and the remaining cell surface, which is referred to here as the cytoplasm. The surface for each of these compartments was determined by point counting using a 224-point overlay (0.475 µm²/point). Gold particles, and points on each compartment, for all cell profiles obtained from one section were added. Labeling density (gold particles/ µm²) values were determined for each section, and the values of all sections for an animal were averaged, giving the labeling density value for the animal. Volume densities of the compartments in the type II cells (µm³/µm³) were calculated from the sum of points on all sections for each animal. Total numbers of animals, sections, and micrographs studied were: three male and three female wild-type mice, 17 sections, 129 pictures; four male and two female hemizygous mice, 20 sections, 149 pictures; three male and one female homozygous mice, 15 sections, 108 pictures. The basic assumption made in the interpretation of the results is that the labeling density over a given compartment is a direct and unbiased estimate of the concentration of immunoreactive MnSOD in this compartment (7, 8).

Hyperoxic Exposure of Mice

Eight-week-old, age-matched nontransgenic and hemizygous transgenic mice were used for exposure to > 99% or 90% oxygen in Durham, NC. Exposure of nontransgenic and homozygous transgenic mice was performed in Detroit, MI. The atmospheric pressure at both locations is around 760 torr. Hyperoxic exposures were conducted in Plexiglas chambers with gas inlet and outlet. The oxygen concentration varied less than 2%, and CO₂ concentration was maintained at less than 0.5% by providing seven to eight complete gas changes per hour. During the exposure, food and water ad libitum were provided, and the animals were kept in a 12-hours-on, 12-hours-off light cycle at all times. The number of surviving animals was counted every 8 h.

Statistical Analysis

One-way analysis of variance (ANOVA) was used to examine differences in lung SOD activity and labeling density of MnSOD in type II cells in various mice. If a significant effect was observed (P < 0.05), then pairwise comparisons among mice were made using Duncan's test. Nonparametric survival distributions were estimated separately for mice exposed to 90% or > 99% O₂ to examine differences in survival for transgenic and nontransgenic mice. The differences between the two groups of mice were analyzed using a Wilcoxon test. No comparisons were made between the 90% and > 99 % O₂ exposures.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Generation of Transgenic Mice Carrying Human MnSOD Transgenes

The EcoRI-HindIII DNA fragment of construct 1 and the EcoRI-XbaI DNA fragment of construct 2, either without or with the polyadenylation site of SV40, respectively, were isolated free of vector sequences for production of transgenic mice (Figure 1). Four transgenic mice, as determined by Southern blot analysis of tail DNA probed with human MnSOD cDNA, were produced with DNA of construct 1 and 15 transgenic mice were produced with construct 2. The human MnSOD transgenes in three out of four founder mice established with construct 1 and in 12 out of 15 founder mice established with construct 2 were found to be transmitted into offspring. Those founder mice which failed to pass the transgenes to their offspring usually contained fewer copies of transgenes than those who did. Presumably, those founder mice were mosaics with the transgenes integrated in a fraction of their cells.

Expression of Human MnSOD Transgenes in Transgenic Mice

RNA blot analyses of human MnSOD mRNA in various tissues of three transgenic mouse lines established with construct 1 were performed. Figure 2a shows that a species of 1.4-kilobase (kb) human MnSOD mRNA was present in all tissues examined from the line TgHMS66 transgenic mouse, but not in tissues from the nontransgenic littermate. The human MnSOD transgene expressed at a relatively high level in muscle, heart, lung, and tongue, and at a lower level in brain, kidney, spleen, and eye. There was a very low level of expression of the transgene in the liver. A species of hybridized mRNA, approximately 1.25 kb, was also observed in heart, liver, and kidney of the nontransgenic littermate. It probably results from cross-hybridization between the endogenous mouse MnSOD mRNA and the human probe under the stringency of hybridization used. The other two lines of transgenic mice carrying construct 1 expressed human MnSOD mRNAs of high molecular weight ranging from 2.5 to 5 kb (data not shown). Because no G-U-rich element is present in the construct microinjected, they presumably represent transcripts which are inappropriately cleaved and polyadenylated at 3' ends (25). The discrete size of the human MnSOD transcript in line TgHMS66 transgenic mice is believed to result from a mouse G-U-rich element located downstream from the site of transgene integration.

View larger version (55K): [in this window] [in a new window]   View larger version (20K): [in this window] [in a new window]   View larger version (49K): [in this window] [in a new window]   Figure 2.   Expression of human MnSOD transgene in transgenic mouse line TgHMS66. (a) RNA blot analysis of total cellular RNA (5 µg for RNA isolated from eye; 50 µg for all other tissues) probed with a human MnSOD cDNA. The positions of RNA size marker are shown to the left. Signals for human MnSOD mRNA can also be seen in eye and liver of transgenic mouse after a longer exposure of the autoradiograph. (b) Protein blot analysis of total cellular protein (30 µg) separated on a 12% SDS-polyacrylamide gel and reacted with an antirat MnSOD antiserum and 125I-protein A. (c) Protein blot analysis of human and mouse proteins separated on a 10% nondenaturing gel. Minus, plus, and double plus signs represent nontransgenic, hemizygous transgenic, and homozygous transgenic littermates, respectively. The origins of the tissues isolated are shown at the top of the figure.

Similar RNA blot analyses were also performed with RNA isolated from tissues of three transgenic lines established with construct 2. Expression of the human MnSOD transgene was found in all transgenic lines tested. The relative abundance of expression of the transgene in various tissues was very similar to that of TgHMS66 (data not shown). Further studies were then performed on transgenic line TgHMS66 which carries approximately three copies of the microinjected transgene, as determined by comparing the hybridization densities of TgHMS66 genomic DNA with those of plasmid DNA of known quantities (data not shown).

To demonstrate the presence of over-expressed MnSOD protein in transgenic mice, we performed protein blot analysis of cellular proteins separated on an SDS- polyacrylamide gel using antirat MnSOD antiserum (Figure 2b), which has been shown to react with both mouse and human MnSOD proteins. This antiserum was found to react with a protein of approximately 22-24 kilodaltons, shown to be MnSOD by the fact that this band could be competed out by an initial incubation of the antiserum with recombinant human MnSOD. Increased levels of MnSOD were found in all tissues of the transgenic mouse when compared with those of the nontransgenic littermate. As summarized in Figure 1, elevated MnSOD protein was observed in lungs of line TgHMS66 mice as well as 10 out of 12 transgenic lines established with construct 2.

To further verify that the increased MnSOD protein in transgenic mouse line TgHMS66 resulted from expression of the transgene, rather than activation of the endogenous mouse MnSOD gene, protein blot analysis of cellular proteins separated on a native gel was also performed (Figure 2c). Under nondenaturing conditions, the human MnSOD protein migrated faster than did mouse MnSOD. A species of MnSOD protein which co-migrated with the human liver MnSOD was found only in the lungs and liver of the hemizygous transgenic mouse but not in the nontransgenic littermate. The amount of the human MnSOD protein was double in the lungs of the homozygous transgenic mouse when compared with that of the hemizygous transgenic littermate.

Enzymatic Activity of MnSOD Expressed from the Transgene

The SOD activity assay was also performed to determine whether the expressed human MnSOD in lungs of transgenic mice is functional (Figure 3). While there were no statistically significant differences in the specific activity of CuZnSOD among the nontransgenic, hemizygous, and homozygous transgenic mice, the specific activity of MnSOD was found to be increased by 170% in the lungs of hemizygous transgenic mice when compared with those of the nontransgenic mice. Specific MnSOD activity was further increased in homozygous transgenic mice.

View larger version (20K): [in this window] [in a new window]   Figure 3.   SOD activities in lungs of TgHMS66 mice. Open, hatched, and solid bars represent nontransgenic (n = 6), hemizygous (n = 7), and homozygous (n = 4) transgenic mice, respectively. *P < 0.01 when compared with wild types. **P < 0.01 when compared with either wild-type or hemizygous transgenic littermates. Bars represent mean ± standard deviation.

Immunocytochemical Localization of the Expressed Human MnSOD in Lungs of Transgenic Mice

To determine the types of lung cells expressing the human MnSOD transgene, lung sections of transgenic and control mice were immunocytochemically labeled with an antirat MnSOD antiserum and protein A-gold. Control labeling experiments using the antiserum pre-absorbed with recombinant human MnSOD to deplete anti-MnSOD antibodies showed that no specific labeling remained on LR-White embedded sections (data not shown). Therefore, the antiserum used specifically labeled MnSOD on the lung sections embedded in LR-White. As expected, the density of immunolabeling was highest in homozygous animals and lowest in nontransgenic littermates (Figures 4a- 4c). Significant differences in labeling densities in mitochondria of alveolar type II cells were found between each pair of groups (P < 0.01). Quantitation of gold particles over the mitochondria of epithelial type II cells revealed a direct gene-dosage effect, the increased immunolabeling density detected in hemizygous animals over nontransgenic littermates being additive in homozygous animals (Figure 4d). The increased immunolabeling for MnSOD was highly specific for mitochondria, indicating that the expressed human MnSOD is properly targeted into the mitochondrial matrix. No significant differences were found in the labeling densities of nuclei, lamellar bodies, and cytoplasm of alveolar type II cells in these three groups of mice.

View larger version (126K): [in this window] [in a new window]   Figure 4.   Immunocytochemical labeling of MnSOD in lung type II cells. Antirat MnSOD antiserum and protein A-gold labeled the mitochondria (M) of lung type II cells in (a) nontransgenic (wild type); (b) hemizygous transgenic; and (c) homozygous transgenic mice. L represents lamellar bodies. Arrow in (c) indicates a fibroblast. (d) Quantitative analysis of the labeling density (number of gold particles/ µm2) in mitochondria, nuclei, lamellar bodies, and cytoplasm of alveolar type II cells of those three groups of mice. The volumes of subcellular organelles of alveolar lung type II cells are shown in the inset of (d). Mito, Nu, LB, and Cyto represent mitochondria, nuclei, lamellar bodies, and cytoplasm, respectively.

No differences were noted in the labeling densities between male and female mice, either nontransgenic or transgenic (data not shown). Furthermore, morphometric analysis demonstrated that no significant changes occurred in the subcellular morphometric characteristics of alveolar type II cells in the three groups of mice, and the mitochondria occupied 12-14% of the sampled cell volume (Figure 4d).

Furthermore, a high density of immunolabeling for MnSOD was also observed in mitochondria of other major types of cells in the alveolar region of the hemizygous and homozygous transgenic animals, including alveolar type I epithelial cells, capillary endothelial cells, and interstitial fibroblasts. An example of the increased immunolabeling in mitochondria of these types of cells of homozygous transgenic mice relative to that of nontransgenic mice is shown in Figure 5. These data indicated a general activity of transcription of the human -actin promoter in lung parenchymal cells of the TgHMS66 transgenic mice.

View larger version (57K): [in this window] [in a new window]   Figure 5.   Immunocytochemical labeling of MnSOD in various types of lung cells. (a) The mitochondria of lung type I (T1) and endothelial (E) cells of a nontransgenic mouse are marked by one arrow and two arrows, respectively. (b) An increase in labeling density can be found in mitochondria of endothelial cells and fibroblasts (F; the mitochondrion is marked by an arrowhead) of a homozygous transgenic mouse. (c) Similarly, an elevated labeling density can also be seen in mitochondria of type I cells and fibroblasts of a homozygous transgenic mouse.

Survival of Transgenic Mice under Hyperoxia

Eight-week-old, age-matched nontransgenic, hemizygous and homozygous transgenic mice derived from two to three litters were used for survival studies under 90% or > 99% oxygen exposure (Figure 6). There was no significant difference in the mean survival times of hemizygous transgenic and nontransgenic mice when exposed to > 99% oxygen (4.5 d ± 0.3 SE and 4.6 d ± 0.2 SE, respectively) (Figure 6a). Furthermore, the mean survival time of homo-zygous transgenic mice was not different from that of nontransgenic mice (4.1 d ± 0.1 SE and 4.2 d ± 0.1 SE, respectively) (Figure 6c). However, upon exposure to 90% oxygen, the mean survival time of hemizygous transgenic mice (6.3 d ± 0.3 SE) was significantly longer than that of nontransgenic mice (5.3 d ± 0.2 SE) (P = 0.012; Figure 6b). In addition, homozygous transgenic mice also exhibited a survival time similar to that of hemizygous transgenic mice under exposure to 90% oxygen (data not shown). It seems that a further increase in lung MnSOD activity in those homozygous transgenic mice does not prolong the survival of these mice compared with that of hemizygous mice following hyperoxic exposure.

View larger version (16K): [in this window] [in a new window]   Figure 6.   Survival analysis of hemizygous and homozygous transgenic mouse line TgHMS66 under hyperoxia. Survival of age-matched nontransgenic and hemizygous transgenic mice under (a) > 99% oxygen; and (b) 90% oxygen. (c) Survival of homozygous transgenic and age-matched nontransgenic mice under > 99% oxygen.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

To study the physiologic effects of altered antioxidant enzyme activities in transgenic mice upon exposure to hyperoxia, the ideal promoter used in construction of a transgene would be the one that directs high-level expression of the enzyme in the lung cells in which we intend to provide protection. In the rat model of oxygen toxicity, pulmonary capillary endothelial cells are thought to be a primary target for oxygen-induced lung injury, which then leads to death of the animal (3). Unfortunately, there is no adequate promoter available that would confer specific and high-level expression of a transgene in endothelial cells. The human -actin promoter was chosen based on the fact that -actin is abundantly expressed in all types of non-muscle cells (24). Indeed, an immunocytochemical study revealed increased labeling for MnSOD in various types of lung cells including endothelial cells, though quantitative analysis of labeling density was performed only on alveolar type II cells. The human MnSOD transgene was also expressed in all other types of tissues examined (32). Interestingly, the human -actin promoter seems to be more active in tissues containing a large number of muscle cells, such as heart, skeletal muscle, and tongue, than in other tissues. Our results in transcriptional activity of the human -actin promoter in vivo in transgenic animals are consistent with those reported by Gunning and associates (24) that -actin promoter is very active in vitro in rat myogenic cell lines. These observations suggest that another sequence of the human -actin gene outside that present in the expression vector is critical to downregulate the expression of -actin during myogenesis.

The magnitude of increase of MnSOD activity in each tissue of transgenic mice is determined by the level of endogenous mouse MnSOD activity and the extent of expression of the transgene. Therefore, part of the reason for a nearly 170% increase in MnSOD activity in lungs of transgenic mice compared with controls is due to the relatively low activity of MnSOD present in the mouse lungs. Furthermore, amino acid polymorphism in human MnSOD protein may also affect the activity of the expressed protein. The human MnSOD protein encoded by the cDNA cloned in our laboratory contains a threonine (Thr) at amino acid residue 58 instead of isoleucine (Ile) as reported by Beck and coworkers (33). It was found recently that replacement of Ile-58 by Thr destabilizes the 4-helix bundle of the protein, leading to a reduced tetramer stability and a lower enzyme activity (34). Indeed, recombinant human MnSOD containing Thr-58 has a specific activity of 1,500 U/mg protein, compared with 5,000 U/mg protein of the normal enzyme (Dr. Seishi Takahashi, Mitsui Toatsu Chemicals, Inc., Tokyo, Japan, personal communication). Therefore, a greater increase of pulmonary MnSOD activity in transgenic mice should be achievable when the MnSOD cDNA containing an isoleucine codon at amino acid residue 58 is used in generation of transgenic mice.

To our surprise, despite an increase of MnSOD protein in many types of lung cells, including capillary endothelial cells in transgenic mice, no tolerance to > 99% oxygen was found in these mice when compared with nontransgenic mice. The hemizygous and homozygous transgenic mice are slightly more resistant than are controls during exposure to 90% oxygen. Histopathologic studies at light microscopic level were also performed on lung samples from three nontransgenic and three hemizygous transgenic mice at 5 d of 90% oxygen exposure. Compared with air- exposed mice, both nontransgenic and hemizygous transgenic mice exhibited thickening of alveolar septa and destruction of alveolar wall, resulting in enlarged alveolar pores (data not shown). However, no significant morphologic differences in tissue injury could be found in these two groups of mice. These results are somewhat expected. Since MnSOD over-expression only marginally prolongs the survival of transgenic mice relative to controls under 90% oxygen, the difference in severity of lung injury in these mice might be difficult to detect.

Recently, White and colleagues (35) have shown that transgenic mice over-expressing human CuZnSOD are more resistant to > 99% oxygen at 630 torr than age-matched controls when young (2.5 mo). However, no resistance was observed when exposure was performed with older mice (5.5 mo old). Although the increased activities of lung glutathione peroxidase and glucose-6-phosphate dehydrogenase in young transgenic mice but not old ones may give a clue as to the age-related difference in response to hyperoxia, additional unknown cellular factors which change during the aging process may also play a critical role in providing protection to animals. In other words, a constitutively high pulmonary MnSOD activity may not be sufficient to provide a drastic protection to animals against hyperoxic insults. This notion is also supported by our results in this study. It should be noted that the increases in MnSOD immunolabeling density of transgenic type II cells (180 and 350% in hemizygous and homozygous transgenic mice, respectively) exceed that found in adult rats exposed to 85% for 7 d (67%) (7, 8). The fact that those 85%-oxygen pre-exposed rats but not the transgenic mice can survive exposure to > 99% oxygen exposure for at least 21 d (3) implicates the critical role of other cellular factors additional to MnSOD in lung antioxidant defense mechanism.

Furthermore, Wispe and associates (36) have demonstrated that transgenic mice carrying a human MnSOD transgene driven by a human surfactant protein C (SP-C) promoter are markedly more resistant to > 99% oxygen than controls. In their model of transgenic mice, a 60% increase of specific lung MnSOD activity was observed when compared with controls. Since SP-C promoter has been shown to direct high-level expression of other transgenes in both alveolar type II cells and nonciliated bronchiolar epithelial cells (37), the magnitude of increase of MnSOD in these specific cells in transgenic mice generated by Wispe and colleagues may be greater than that of our transgenic animals. However, the increase in MnSOD immunolabeling density in mitochondria of type II cells of the transgenic mice generated by Wispe and coworkers (400%) was only slightly higher than that found in our homozygous transgenic mice (350%). These observations suggest that the discrepancy in the protective effect of MnSOD may result from different strains of mice being used in Wispe and associates' (FVB/N) and our (B6C3 hybrid) laboratories. If different strains of mice do contribute to the different degrees of tolerance to hyperoxia, the different cellular factors derived from different genetic backgrounds of those mice may play an important, yet not understood, role in antioxidant defense. This notion is supported by the observation that various strains of mice exhibit different sensitivities to ozone and hyperoxia (38, 39). One possible approach to uncover the role of genetic variance in affecting pulmonary antioxidant defense is to generate FVB/N and B6C3 F1 transgenic mice with our and Wispe and colleagues' MnSOD expression constructs, respectively, and then to compare the degrees of MnSOD over-expression and oxygen tolerance of these mice.

In addition to those unknown cellular factors, a balance among various antioxidant enzymes may also be critical in antioxidant defense. It has been shown that certain cultured cells over-expressing CuZnSOD are more sensitive to oxidant stress (40). However, cells over-expressing CuZnSOD do develop resistance to oxygen radicals, and the resistance is apparently determined by the cellular activity of glutathione peroxidase (42). Omar and colleagues have also demonstrated that the protective effect of SOD against ischemia/reperfusion heart injury is highly dose-dependent (43, 44). A high dose of SOD is associated with a decrease in protection and can even enhance the injury during reoxygenation. The mechanism by which high dose of SOD causes further damage to the perfused heart is not currently understood. In our model of transgenic mice, no differences were found in the activities of lung glutathione peroxidase, glutathione reductase, catalase, and glucose-6-phosphate dehydrogenase in nontransgenic, hemizygous, and homozygous transgenic mice (data not shown). Our studies here suggest that other cellular mechanisms in addition to the over-expressed MnSOD may be necessary to provide a maximum protection to B6C3 hybrid mice against hyperoxic insults. It will be interesting to learn whether a concomitant increase of other antioxidant enzymes (such as catalase) through the transgenic technology, in addition to MnSOD, will provide effective protection to the animals against hyperoxic insults. These transgenic mice with over-expression of antioxidant enzymes in other tissues such as brain and heart may also provide opportunities for defining the role of these enzymes in other diseases whose pathogenesis is associated with overproduction of reactive oxygen species.