Journal of Applied Physiology
Vol. 83, No. 2, pp. 559-568, August 1997
CELLULAR ASPECTS OF LUNG FUNCTION

Aerosolized manganese SOD decreases hyperoxic pulmonary injury in primates. II. Morphometric analysis

Karen E. Welty-Wolf, Steven G. Simonson, Yuh-Chin T. Huang, Stephen P. Kantrow, Martha S. Carraway, Ling-Yi Chang, James D. Crapo, and Claude A. Piantadosi

Department of Medicine, Duke University Medical Center, Durham, North Carolina 27710

ABSTRACT

INTRODUCTION

METHODS

RESULTS

DISCUSSION

ACKNOWLEDGEMENTS

FOOTNOTES

REFERENCES

ABSTRACT

Welty-Wolf, Karen E., Steven G. Simonson, Yuh-Chin T. Huang, Stephen P. Kantrow, Martha S. Carraway, Ling-Yi Chang, James D. Crapo, and Claude A. Piantadosi. Aerosolized manganese SOD decreases hyperoxic pulmonary injury in primates. II. Morphometric analysis. J. Appl. Physiol. 83(2): 559-568, 1997.Hyperoxia damages lung parenchyma via increased cellular production of reactive oxygen species that exceeds antioxidant defenses. We hypothesized that aerosolized human recombinant manganese superoxide dismutase (rhMnSOD) would augment extracellular antioxidant defenses and attenuate epithelial injury in the lung during hyperoxia in primates. Twenty-four adult male baboons were anesthetized and mechanically ventilated with 100% oxygen for 96 h. The baboons were divided equally into four groups. Oxygen alone and oxygen plus rhMnSOD given at 3 mg · kg¹ · day¹ were compared to assess efficacy of the drug. Subsequently, aerosolized rhMnSOD was given at 1 or 10 mg · kg¹ · day¹ to study dose effects and toxicity. Quantitative morphometry showed protection of alveolar epithelium from hyperoxia by 3 mg · kg¹ · day¹ rhMnSOD (P < 0.05). In addition, interstitial fibroblast volumes were increased in the treatment group (P = 0.06). This effect appeared greater at the two higher doses of the rhMnSOD. The aerosolized drug was localized to the surface of airways and air spaces and macrophages by immunolabeling studies, suggesting efficacy via physicochemical properties that localize it to cell surfaces or by effects on alveolar macrophage function.

superoxide dismutase; antioxidant enzymes; oxygen; acute lung injury; acute respiratory distress syndrome; lung ultrastructure

INTRODUCTION

OXIDATIVE STRESS plays an important role in the pathogenesis of many pulmonary diseases, including the acute respiratory distress syndrome, drug-induced lung injury, and ischemia-reperfusion injury of the lung. Hyperoxic lung injury, in particular, is thought to be related directly to the generation of reactive oxygen species (ROS), including superoxide (O₂·), hydrogen peroxide (H₂O₂), and hydroxyl radical (· OH). The rate of production of these ROS is thought to exceed the capacity of the lung's endogenous antioxidant defenses to detoxify them. Progressive exposure to 100% oxygen results in injury to endothelium and alveolar type 1 epithelium, with an accompanying proliferation of type 2 epithelial cells and interstitial cells (10, 12, 17). In primates, the damage to type 1 epithelium results in partial denudation of the basement membrane by 4 days of exposure, with exudative changes in the alveolar space consisting of edema, cellular debris, and inflammatory cell infiltrate (12, 17). Polymorphonuclear cells are increased in lavage fluid at this point and are thought to play a role in amplification of the lung injury (10).

Exactly how an increase in ROS translates into the histological changes characteristic of oxygen toxicity is unknown. In small-animal models of pulmonary oxygen toxicity, depletion of antioxidants increases lung injury (11), whereas exposure to conditions such as endotoxemia (4, 13, 26) or 85% oxygen (7), which increase cellular antioxidants, decreases pulmonary injury. Important enzymatic antioxidant defenses in the lung include the superoxide dismutases (SOD), which catalyze the dismutation of O₂ · to H₂O₂ and molecular oxygen. Experiments with aerosolized bovine copper-zinc SOD (CuZnSOD) failed to modify hyperoxic lung injury despite achieving high levels of enzyme activity in lung tissue (6). This lack of protection was postulated to be due to lack of delivery of the protein to the intracellular compartments, although studies using tracheally instilled CuZnSOD show that deposition of that enzyme in small terminal airways may protect against airway complications of hyperoxia (8). Recently, a human recombinant manganese SOD (rhMnSOD) has been shown to have efficacy in improving biochemical and physiological indicators of lung injury when given to hyperoxic mice via nasal insufflation (32). Topical application of the enzyme elevated lung levels, whereas systemic administration did not (24). The basis for the effect is hypothesized to be the cationic nature of the enzyme, which causes it to adhere to cell membranes (32). Such localization should lend itself to protection of the alveolar epithelial surface in hyperoxia.

We used aerosolized delivery of rhMnSOD in this study to test the hypothesis that this enzyme would attenuate pathological injury to the lung during hyperoxia in baboons. In this animal model, hyperoxia produces an injury pattern with pathological and physiological characteristics similar to these seen in the acute respiratory distress syndrome in humans, including inflammatory cell infiltrates and epithelial damage (10, 22). This animal species is particularly well suited to the study of aerosolized therapy because of the anatomic similarity of the airways of the baboon to the human. We demonstrated that aerosolized rhMnSOD instituted at the onset of exposure to high inspired concentrations of oxygen protects the alveolar epithelium from oxygen injury. The physiological and gas-exchange effects of the treatment on hyperoxic lung injury are reported in the accompanying study (24a).

METHODS

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

RESULTS

Exposure of baboons to 100% oxygen for 96 h produced a histological pattern of lung injury typical of oxygen-induced pulmonary damage, as previously reported (12, 17, 23). The hyperoxic lung injury in our oxygen-exposed animals was characterized by severe swelling and vacuole formation in type 1 alveolar epithelial cells. The epithelium was injured so severely that large portions of the alveolar basement membrane surface were completely denuded of epithelial cells. The epithelial injury was accompanied by thickening of the interstitial compartment with edema and an inflammatory cell infiltrate consisting of both mononuclear and polymorphonuclear cells. Some swelling and vacuolization of endothelial cells was present as well. In contrast, lungs from animals exposed to oxygen and 3 mg · kg¹ · day¹ rhMnSOD had preserved integrity of the epithelial surface. Ruffling of the epithelial cell surface was present with minor edema, but the cell layer was intact along the majority of the alveolar surface. In these animals, some of the type 2 epithelial cells appeared hypertrophied and showed blebbing. Others were seen spreading in low profiles along the basement membrane. Neither of these findings is typical of lungs exposed to oxygen alone. The interstitium was thickened in animals treated with oxygen plus 3 mg · kg¹ · day¹ rhMnSOD as well, but, unlike animals exposed to oxygen alone, this was due in large part to an increase in fibroblasts and other noninflammatory cells. Less interstitial edema was present. Endothelial cell changes persisted in the treated group, but the overlying alveolar epithelial cells in these septae were relatively undamaged. Low-magnification electron micrographic views of an oxygen-injured lung and a lung from an animal receiving 3 mg · kg¹ · day¹ rhMnSOD during hyperoxia are shown in Fig. 1.

1

1

A summary of the results of the morphometric analysis is shown in Table 1. As noted in the accompanying paper (24a), three animals did not survive (to 96 h), one each in the untreated oxygen-exposed group, in the oxygen plus 1 mg · kg¹ · day¹ rhMnSOD group, and in the oxygen plus 10 mg · kg¹ · day¹ rhMnSOD group. All animals were included in the morphometric analysis (n = 6 in each group). The lungs of the oxygen-exposed control group were compared with animals treated with oxygen plus 3 mg/kg rhMnSOD to assess protective effects of the drug. The rhMnSOD-treated animals showed decreased epithelial injury compared with untreated oxygen-exposed animals. Total epithelial cell volume normalized to alveolar basement membrane was 50% greater in hyperoxic animals treated with 3 mg/kg rhMnSOD (P = 0.009). This increase was due primarily to changes in normalized type 1 cell volume density, which was also 50% higher in the treated group (P = 0.002). Normalized type 2 cell volume density also tended to increase, although this was less significant (P = 0.14). The composition of the alveolar epithelial surface also was affected by rhMnSOD treatment. Bare basement membrane, an important indicator of alveolar epithelial injury in hyperoxia, was present over 24.44 ± 8.64% of the alveolar surface in oxygen-injured lung, but with concomitant rhMnSOD treatment it was found to represent only 5.03 ± 1.38% of the alveolar surface (P = 0.05). The percentage of the alveolar surface covered by type 1 cells was preserved in rhMnSOD-treated animals (P = 0.09). The percent type 2 cell coverage tended to increase in animals who received 3 mg · kg¹ · day¹ rhMnSOD plus oxygen (P = 0.13) (see Fig. 2).

Table  1.   Summary of morphometric analysis Air Oxygen rhMnSOD dose, mg · kg1 · day1 1 3 10 Epithelium, total 0.31 ± 0.036  0.250 ± 0.015  0.392 ± 0.094  0.387 ± 0.040  0.515 ± 0.069  (0.009)   Type 1  0.202 ± 0.020  0.149 ± 0.016  0.223 ± 0.403  0.232 ± 0.012  0.264 ± 0.027  (0.002)   Type 2  0.099 ± 0.018  0.11 ± 0.10  0.169 ± 0.054  0.155 ± 0.033  0.251 ± 0.053  (0.14)    %BBM 1.56 ± 0.58  24.44 ± 8.64  16.53 ± 12.43  5.02 ± 1.38  14.86 ± 6.50  (0.05)    %Type 1  95.14 ± 0.57  70.76 ± 8.81  77.00 ± 11.96  87.62 ± 2.63  75.13 ± 7.32  (0.09)    %Type 2  3.30 ± 0.57  4.80 ± 0.80  6.47 ± 1.66  7.35 ± 1.35  10.01 ± 2.74  (0.13)  Interstitium, total 0.631 ± 0.055  0.824 ± 0.076  0.806 ± 0.073  0.883 ± 0.056  0.936 ± 0.053  (0.54)    Fibroblasts 0.208 ± 0.023  0.283 ± 0.022  0.300 ± 0.036  0.351 ± 0.026  0.363 ± 0.034  (0.06)    Mononuclear cells 0.006 ± 0.002  0.020 ± 0.003  0.007 ± 0.002  0.021 ± 0.004  0.012 ± 0.004  (0.91)  Endothelium 0.309 ± 0.024  0.260 ± 0.028  0.300 ± 0.033  0.296 ± 0.022  0.314 ± 0.018     (0.33)    %DCBM 0.11 ± 0.06% 2.87 ± 1.33  2.21 ± 0.91  2.62 ± 1.64  1.05 ± 0.50  (0.90)  PMN, total 0.029 ± 0.004  0.085 ± 0.016  0.127 ± 0.034  0.105 ± 0.008  0.104 ± 0.012  (0.28)    Intravascular 0.026 ± 0.005  0.051 ± 0.011  0.082 ± 0.024  0.060 ± 0.010  0.072 ± 0.012  (0.26)    Interstitial 0.002 ± 0.001  0.011 ± 0.002  0.015 ± 0.003  0.012 ± 0.004  0.009 ± 0.003  (0.89)    Alveolar 0.002 ± 0.001  0.023 ± 0.008  0.030 ± 0.013  0.033 ± 0.015  0.023 ± 0.007  (0.27)  Data are means ± SE; n = 6 animals in each group. Values are normalized volume density (in µm3/µm2) except for percentages of surface area as indicated. BBM, bare basement membrane; %Type 1, %basement membrane covered by type 1 epithelium; %Type 2, %basement membrane covered by type 2 epithelium; DCBM, disrupted capillary basement membrane; PMN, polymorphonuclear cell. Oxygen and oxygen plus 3 mg · kg1 · day1 recombinant human manganese superoxide dismutase (rhMnSOD) values were compared by using Student's unpaired t-test. P values are shown in parentheses. Data from air-ventilated animals are shown for reference.

The increase in thickness of the pulmonary interstitium seen in hyperoxia was not prevented by 3 mg/kg rhMnSOD. Normalized total interstitial volume density was similarly increased in the two groups (P = 0.54). The normalized volume density of fibroblasts in the interstitium, however, increased 27% in animals treated with oxygen and rhMnSOD compared with hyperoxic controls (P = 0.06). Similarly, the inflammatory cell infiltrate was not diminished by rhMnSOD. Normalized total polymorphonuclear cell volume density increased threefold during hyperoxia in both untreated and treated animals. Normalized mononuclear cell volume density also was increased in the interstitium in both groups. Normalized endothelial cell volume density was unaffected by 3 mg/kg rhMnSOD (P = 0.33).

On the basis of the epithelial protection seen at 3 mg/kg, additional studies were done at 1 and 10 mg · kg¹ · day¹ of rhMnSOD to explore dose-response relationships. Exploratory data analysis was performed in these additional groups. Data (means ± SE) from the morphometric analysis of animals treated with 1 and 10 mg · kg¹ · day¹ of rhMnSOD are summarized in Table 1. As with the 3 mg · kg¹ · day¹ dose, type 1 epithelial volume density appeared preserved at the other two doses of rhMnSOD. There was a suggestion, however, that rhMnSOD was only partially protective at 1 and 10 mg because the percentage of alveolar epithelial surface covered by bare basement membrane increased to 16.53 ± 12.43% in the 1 mg/kg group and to 14.86 ± 6.50% in the 10 mg/kg group. Graphs in Fig. 3 show the individual data points and median values for the three treatment groups and the oxygen control group. The data from air-ventilated animals are also plotted for comparison. Type 2 epithelial cell volume density, which tended to increase in the 3 mg/kg rhMnSOD group compared with oxygen control animals (P = 0.14), was increased in the two additional treatment groups as well (see Table 1). Data points demonstrating some individual values as high as twice those seen in any air or oxygen-exposed animal are shown in Fig. 4. The normalized volume density of inflammatory cells in all compartments of the lungs was increased equally throughout the dosage range of rhMnSOD (Table 1). Interstitial and fibroblast volume densities increased progressively with increases in rhMnSOD dose. Individual data and medians are illustrated in Fig. 5.

1

1

Immunolocalization of the aerosolized rhMnSOD showed deposition of the protein along the epithelial surface and in airways (Fig. 6). Little staining was seen within epithelial or interstitial cells or in the interstitial matrix, but significant accumulation of rhMnSOD was present within alveolar macrophages. Notably, at the highest dose and, therefore, high concentrations of rhMnSOD, a few small airways at lung sectioning were found to be occluded by proteinaceous plugs.

DISCUSSION

During continuous hyperoxia, concomitant treatment with aerosolized rhMnSOD at a dose of 3 mg · kg¹ · day¹ decreased histological lung injury in the baboon. The lung was protected at the level of the alveolar epithelial surface by preservation of type 1 cell volume and the integrity of the epithelial cell surface on the basement membrane. Type 2 epithelial cell morphology was altered qualitatively because the cells seemed to spread over the alveolar surface. Morphometric analysis showed a trend toward increase in normalized volume density in this cell type. A proliferative response in the interstitial space was suggested by both the qualitative appearance and the increase in normalized fibroblast volume density. The epithelial protection did not require inhibition of the inflammatory cell influx and did not extend to the capillary endothelium.

Given the evidence that increases in MnSOD expression as well as other antioxidant enzymes such as catalase and glutathione peroxidase are important in protecting the lung against oxidative injury, previous therapeutic efforts have attempted to target exogenous antioxidant enzymes to intracellular compartments. Polyethelene glycol-conjugated CuZnSOD and catalase augment antioxidant activity in both alveolar type 2 and endothelial cells in culture, conferring resistance to oxidant stress (2, 29) and decreasing lung injury from 100% oxygen in small animals when given together (31). Administration of liposome-encapsulated CuZnSOD and/or catalase also is protective in rat and rabbit models of oxidant lung injury (1, 27). Liposomes, however, are technically difficult to prepare, and the effects of large amounts of lipid on the lung and reticuloendothelial system are not well understood.

Our results show that aerosolized rhMnSOD, delivered without a specific targeting mechanism, was successful in preventing epithelial injury during hyperoxia. The immunolocalization of rhMnSOD in our study shows that the protein was located almost exclusively along the airway and alveolar epithelial surfaces, in keeping with studies in mice where the enzyme was delivered via nasal insufflation. These latter studies, performed with gold immunolabeling at the EM level, showed that only very small amounts of the enzyme were taken up into epithelial cells or the interstitial spaces, although there was significant accumulation within alveolar macrophages by 48 h (9). We also found significant uptake of biotinylated rhMnSOD by alveolar macrophages within an hour, and there was only minor labeling of the protein within the epithelial cells at this point. If earlier doses of the drug are also distributed without later uptake into epithelium, then extracellular mechanisms may account for the protective effects of the enzyme. Unlike CuZnSOD, which is negatively charged at physiological pH, rhMnSOD carries a positive charge. When deposited in the lung, the positive charge of rhMnSOD should allow it to bind to negatively charged cell membrane and matrix surfaces. The protection afforded by nebulized rhMnSOD may, therefore, be in large part due to the augmentation of extracellular antioxidant defenses.

The protection of type 1 alveolar epithelium afforded by rhMnSOD may have contributed to better physiological outcome by decreasing lung edema and, thereby, improving intrapulmonary shunt, as reported in the accompanying paper (24a). The primary physiological effect of the treatment was prevention of shunt during hyperoxia. This effect may have been due to several possible mechanisms, including prevention of microatelectasis or preservation of the alveolar epithelium. Our inflation-fixation technique precluded quantitative assessment of atelectasis; however, preservation of epithelial integrity can be implicated from our data as a mechanism to prevent loss of gas-exchange units. By protecting type 2 epithelium and type 2 cell function, shunt may be prevented by preventing alveolar collapse. Although normalized type 2 epithelial volume was not significantly different in animals treated with rhMnSOD at a dose of 3 mg/kg, there were qualitative structural differences present, suggesting that rhMnSOD may preserve the ability of type 2 cells to produce surfactant as well as protect them from oxidative damage.

Notably, rhMnSOD did not affect inflammatory cell recruitment into the lung, but it may exert a protective effect by detoxifying extracellular O₂· generated by those cells. It is also possible that the uptake of rhMnSOD into alveolar macrophages alters their inflammatory function in a manner that attenuates lung injury in this model (9). For example, antioxidants decrease the production of tumor necrosis factor and prostaglandin E₂ by alveolar macrophages after lipopolysaccharide stimulation consistent with a role for ROS in intracellular signaling pathways (20). MnSOD has also been shown to play a role in cellular differentiation (25). Therefore, uptake of aerosolized rhMnSOD by alveolar macrophages may be a significant factor in the protection by modulating intra-alveolar inflammatory events.

rhMnSOD did not prevent interstitial thickening and increases in fibroblast and other interstitial cells during hyperoxia. It may be that its deposition along the epithelial surface positions it ideally to protect epithelial cells from oxidant damage, but the interstitial compartment continues to suffer from oxidative damage. This proliferative response may be a primary manifestation of hyperoxic lung injury, but the suggestion of progressive change with increasing dose of rhMnSOD implicates other factors in the process. The interstitial changes may represent a repair process in the face of partial protection of epithelial and/or endothelial cells. This type of proliferative response is a feature of sublethal oxygen exposure (5), and the combination of epithelial protection with persistent interstitial change has been reported with artificial surfactant treatment of oxygen-exposed lungs in this species (23). A third possibility is a dose-related effect of the rhMnSOD on fibroblast proliferation under hyperoxic conditions.

Although a complete dose-response analysis is subject to type I statistical error due to the small number of animals in each treatment group, our data suggest that increasing doses of rhMnSOD result in progressive increases in interstitial thickness and fibroblast cell volume density. Because the data shown in Figs. 3, 4, 5 suggested dose-related differences in histopathology, a post hoc analysis of variance was performed comparing the morphometric analysis of all animals studied: air, oxygen control, and rhMnSOD at 1, 3, and 10 mg · kg¹ · day¹. Post hoc analysis of variance (Fisher's post hoc test) shows that the increase in interstitial thickness that occurs with hyperoxia (P = 0.04) is persistent at 1 mg/kg rhMnSOD (P = 0.06), 3 mg/kg rhMnSOD (P = 0.01), and 10 mg/kg rh MnSOD (P = 0.002). The increased normalized volume density of fibroblasts is seen in oxygen controls (P = 0.08) and in hyperoxia plus 1 mg/kg (P = 0.03), 3 mg/kg (P = 0.002), and 10 mg/kg rhMnSOD (P < 0.001). The results of this analysis suggest that fibroblast volume is greater in hyperoxic animals receiving 10 mg/kg rhMnSOD than in those exposed to oxygen alone (P = 0.06). It should be noted that the power of this analysis is limited by the small size of the groups relative to the number of comparisons and that, because it was not part of the original study design, definite conclusions regarding drug dose effects cannot be drawn. The data strongly suggest, however, that this compound may have effects beyond epithelial protection.

An increase in type 2 cell volume is not a prominent feature of hyperoxic lung injury in our 96-h primate model, although it has been described at later time points in other animal models. A trend toward type 2 cell response, however, was found in all three rhMnSOD treatment groups. The post hoc analysis showed significant increase in type 2 epithelium in the oxygen plus 10 mg/kg rhMnSOD group compared with both air and oxygen alone groups (P = 0.01). Whether this represents a response to rhMnSOD alone or an interaction between hyperoxia and rhMnSOD cannot be determined from our study, but type 2 cell hypertrophy and proliferation, with subsequent differentiation into type 1 cells, is a well-known response to acute lung injury. It is possible that protection from MnSOD in this model permits reparative processes in the lung to proceed at an accelerated rate, e.g., by protection or stimulation of the type 2 cell. The apparent increase in fibroblast cell volume we found with increasing rhMnSOD dose supports this concept.

A bell-shaped dose-response curve for rhMnSOD has been described in several previous studies (21, 22). In the rabbit heart during ischemia-reperfusion, decreased lactate dehydrogenase release was seen when perfusates included 2 or 5 mg/l rhMnSOD. When the concentration was increased to 50 mg/l, however, the extent of injury was increased over controls. The investigators suggested that O₂· acts as a terminator as well as an initiator of lipid peroxidation (21, 22). Overscavenging of O₂· by SOD, then, may result in a loss of protection against oxidant generation and even increase some types of cellular injury. In addition, increased production of · OH from H₂O₂ and O₂· has been hypothesized to be a potential consequence of excess SOD in the presence of reduced transition metals such as iron (33). Because we delivered the enzyme by nebulizer, we cannot directly compare our doses with those used in the cardiac reperfusion studies. At the highest dose we tested, however, some animals demonstrated increased denudation of the epithelial surface. Loss of the protective effect of the treatment is one possible explanation for this finding.

Another potential contributor to the lung changes we found at high doses of rhMnSOD is manganese toxicity. The immunolabeling suggests that MnSOD is taken up at least partly by alveolar macrophages, but the amount of free manganese in these animals and its fate is unknown. In humans, manganese toxicity has been well characterized in terms of its neurological effects after chronic inhalational exposure. The mechanisms of cellular damage and level of exposure required for toxicity have not been determined, but manganese accumulates in mitochondria, affecting oxidative phosphorylation and interfering with calcium homeostasis (15). In vitro, it also has effects on integrin-dependent cell adhesion (10). Like other transition metals, manganese may enhance the production of oxygen radical species via the Haber-Weiss reaction. Whether manganese in the concentrations used here has any effect on cell structure or function is unknown, and it cannot be excluded as a contributor to the effects of the enzyme in animals treated with high doses of rhMnSOD.

In summary, we have shown that aerosol application of a positively charged enzyme that scavenges O₂· can successfully protect alveolar epithelium from acute injury induced by hyperoxia. If little uptake actually occurs into epithelial cells, the mechanism of protection is likely to relate to detoxification of ROS generated by leukocytes in the extracellular space. It is also possible that the enzyme alters alveolar macrophage function. rhMnSOD may provide better protection than the CuZn isoform of SOD if the positive charge of the molecule allows it to more closely adhere to anionic surfaces. We also found evidence of either toxicity or decreased efficacy at high doses, consistent with the bell-shaped dose-response curve reported for SOD in other experimental models of oxidative injury (21, 22). This effect will be an important consideration in future applications of this and other drugs with SOD-like activity to treat acute lung injury.

ACKNOWLEDGEMENTS

We acknowledge the technical expertise of John Patterson and Craig Marshall and the secretarial assistance of Louise Wilson.

FOOTNOTES

Address for reprint requests: K. E. Welty-Wolf, Dept. of Medicine, Box 3518, Duke Univ. Medical Center, Durham, NC 27710.

Received 19 August 1996; accepted in final form 31 March 1997.

REFERENCES