INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Despite recent advances in neonatal intensive care, bronchopulmonary dysplasia (BPD) remains a major cause of infant morbidity and mortality (1, 2). The mechanisms for the development of BPD are not yet clearly defined, although oxidant stresses and volutrauma/barotrauma in ventilated infants with immature lungs are critical components of the acute injury process that leads to BPD (3). There are many potential sources of oxidant exposures in immature lungs, including increased intracellular production of oxidants from exposure to supplemental oxygen (6) and oxidants from activated inflammatory cells (7). Lung inflammation is a common feature in the development of BPD (7). Observation of greater levels of neutrophil chemoattractants and neutrophil accumulation in the tracheal secretions in infants who develop BPD has led to efforts to inhibit lung inflammatory responses as potential therapies for halting the progression to BPD (11). Unfortunately, the events initiating lung inflammation have not been defined sufficiently to suggest specific anti-inflammatory interventions, and interventions to date have used agents that have effects other than inhibition of inflammatory responses and have been associated with untoward clinical effects (14).

Prematurely born infants are ill equipped to manage the oxidant stresses encountered immediately after birth. The speculation that prematurely born infants have limited abilities to defend themselves against oxidant stresses, and that the levels of susceptibility are inversely related to gestational age first received recognition from animal studies by Frank and coworkers, who observed acute increases in lung antioxidant enzyme activities in the later days of gestation (15, 16). Studies in premature human infants supported this theory, in that activities of antioxidant enzymes and concentrations of antioxidants were lower in premature infants than in term infants, including lower cord blood copper-zinc superoxide dismutase activities (17), plasma glutathione concentrations (18), and hepatic capacities for GSH synthesis (19). The lower capacities of antioxidant systems in prematurely born infants impose an increased risk for deleterious molecular oxidations of biologically important macromolecules in premature infants, which could lead to cellular dysfunction and cell death. In addition, specific oxidations could alter normal cellular or extracellular homeostasis, which could initiate inflammatory responses and inflammation-dependent lung injury.

Oxidations of protein and lipid have been observed in tracheal aspirate fluids of prematurely delivered infants and have been associated with adverse pulmonary outcomes (20, 21). Studies of relevant oxidations in premature infants have focused on analyses of tracheal aspirate fluids (20). A recent study in prematurely born infants revealed increases in serum and urine levels of o-tyrosine, an oxidation product of phenylalanine, suggesting that potentially important oxidations occur also in the vascular compartment (23).

The present study was designed to test the hypothesis that specific protein oxidations in the tracheal fluids of prematurely delivered infants in the first week of life would be associated with the subsequent development of BPD. We obtained tracheal aspirate fluids from infants born at  29 wk gestation and derivatized the samples with 2,4-dinitrophenylhyrazine (DNPH), which reacts with aldehydes and ketones to form hydrazones (24) and similar derivatives with other products of protein oxidation (25). The proteins were separated by electrophoresis, and DNPH-reactive products of oxidation were detected by western blotting with an antibody that detects the dinitrophenyl derivatives. Using these methods, we found that Clara cell secretory protein (CCSP) was oxidized more frequently in infants who developed BPD than in infants who did not develop BPD. In view of the possibility that oxidation of CCSP might diminish the anti-inflammatory functions of the protein, we also tested the hypothesis that immunoreactive CCSP levels would be lower in tracheal aspirate fluids of infants who subsequently developed BPD than in those who would not develop BPD.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

The institutional review boards at Texas Children's Hospital and Baylor College of Medicine approved this study. Forty-five infants  29 wk gestation were enrolled in this prospective study at Texas Children's Hospital from July to October 1997. Infants were eligible for enrollment on Day of life 1 if they were  29 wk gestation. Infants born at  29 wk are routinely intubated, ventilated, and treated with exogenous surfactant. Infants were excluded from the study if they had congenital anomalies or congenital sepsis proven by culture, or if they were admitted to the NICU after Day of life 1. Infants extubated prior to Day of life 6 remained in the study, and samples were analyzed from Days of life 1 and 3 only.

Birth histories were obtained at enrollment and assessed for presence or absence of maternal hypertension, multiple gestation, preterm labor, suspected or proven chorioamnionitis, exposure to prenatal steroids, and the route of delivery. Neonatal histories were assessed for gestational age at delivery, birth weight, Apgar scores at 1 and 5 min, and numbers of exogenous surfactant doses administered. At 28 d of life, the medical histories were reviewed to assess the pulmonary courses, including need for supplemental oxygen, treatment with postnatal steroids, and radiographic evidence of chronic lung disease as assessed by staff pediatric radiologists unaware of the individual infant's participation in the study. Infants were defined as having BPD if they required supplemental oxygen at 28 d of life and they had an abnormal chest X-ray. If both criteria for the definition of BPD were not met the patients were assigned to the no BPD group. Infants who died prior to 28 d of life were analyzed separately, and infants who died after 28 d of life were placed in the BPD or no BPD groups based on criteria assessed at 28 d of life and were assigned appropriately. At discharge, medical records were reviewed for long-term outcomes, including duration of oxygen therapy, length of hospital stay, and the need for home oxygen therapy.

Study Design

Eligible infants enrolled on Day of life 1 had tracheal aspirate samples collected on Days of life 1, 3, and 6. Tracheal aspirate samples were assayed for DNPH-reactive proteins by incubation with DNPH, followed by western blotting, utilizing an antibody directed against DNPH derivatives using the methods described by Knight and coworkers (26). DNPH-reactive proteins that were observed to have markedly different signal intensities in infants who developed BPD than in those who did not develop BPD and individual proteins with different signal intensities were identified visually. For DNPH-reactive proteins that demonstrated consistent differences in signal intensities, the corresponding proteins on Coomassie blue-stained membranes were excised and analyzed by N-terminal amino acid sequencing in the Amino Acid Sequencing Core Laboratory at Baylor College of Medicine. Infants were followed clinically and at 28 d of life were evaluated for clinical and radiographic evidence of BPD, as described above. Each infant was followed until discharge. At no time during the infants' hospitalizations were the caretakers aware of the study results.

Collection of Tracheal Aspirate Samples

Tracheal aspirate samples were collected only when tracheal suctioning was indicated clinically. When suctioning was indicated, 0.5 ml of sterile isotonic saline was instilled into the infant's endotracheal tube and the infant was bagged manually through the endotracheal tube, for three breaths, after which the fluid was suctioned into a sterile mucus trap. This procedure was repeated two more times, and the three tracheal aspirate samples were pooled, placed at 4° C immediately, and subsequently stored at 70° C. Samples were not spun prior to determinations of protein concentrations and DNPH reactivity.

Total Protein Assessment of Tracheal Aspirate Samples

Protein concentrations of tracheal aspirate samples were determined using the technique described by Bradford (27).

Identification of DNPH-reactive Proteins in Tracheal Aspirate Samples

Thirty-five µl of tracheal aspirate samples was combined with 15 µl of 18% sodium dodecyl sulfate (SDS) and heated at 37° C for 30 min. Fifteen µl of 10 mM DNPH in 10% trifluoroacetic acid was added to the sample solutions, producing final volumes of 65 µl. Two 30-µl aliquots of each sample were electrophoresed through 4% acrylamide/ 0.1% SDS stacking gels and 8-20% acrylamide/0.1% SDS gradient gels at constant voltage for 2 h. The separated proteins were transferred electrophoretically to PVDF membranes (Immobilon-P, Millipore Corporation, Bedford, MA). After transfer, one membrane was prepared for analysis of DNPH-reactive protein by western blotting, and the second membrane was stained with Coomassie blue for a visual assessment of total proteins in each sample. The first membrane was washed in a blocking solution of 5% nonfat milk in Tris-buffered saline (TBS) (20 mM Tris, 150 mM NaCl, pH 7.5) for 30 min, then incubated for 16 h on a rocker table in 50 ml of a solution containing a rabbit polyclonal anti-DNPH antibody diluted 1:1,000 (Sigma Chemical Company, St. Louis, MO) in 5% milk/TBS. After incubation with the antibody solution, the membrane was washed three times with TBS, followed by a 2-h incubation on a rocker table in 50 ml of a solution containing a horseradish peroxidase-conjugated polyclonal antirabbit antibody made in goats diluted 1:2,000 (Bio-Rad Laboratories, Hercules, CA) in 5% milk/TBS. The membrane was then washed three times with TBS, stained with a solution containing enhanced chemiluminescence western blotting detection reagents (Amersham Pharmacia Biotech UK Limited, Buckinghamshire, England) for 1 min. The membrane was then blotted dry and exposed to X-ray films for 5, 15, 60, or 300 s intervals. X-Ray films were developed and assessed visually to determine the sequential time points at which the signal intensities did not change between films, and the film with the shorter of the two exposure times was analyzed for total DNPH-reactive proteins by densitometry. For quantitative comparison of DNPH-reactive protein levels, the signals generated from samples obtained on Day of life 1 from infants who did not develop BPD were normalized to 100% in each western blot and utilized for comparison of signals generated for all other samples obtained from infants who did develop BPD, from infants who died over the prior to 28 d of life, and on Days of life 3 and 6 from infants who did not develop BPD.

Identification and Quantification of Clara Cell Secretory Protein (CCSP) in Tracheal Aspirate Samples

In initial studies, DNPH-reactive CCSP was more prevalent in infants who developed BPD than in infants who did not develop BPD, so we performed western blot analyses for levels of immunoreactive CCSP protein in the same tracheal aspirate samples. Equal volumes of tracheal aspirate samples were utilized for the assessment of CCSP protein expression. Each western blot was loaded so that the blot contained samples from at least one infant who did not develop BPD. Ten µl of tracheal aspirate samples was electrophoresed on 10% SDS-polyacrylamide denaturing gels and transferred to PVDF membranes electrophoretically using methods described previously (28). The membranes were then incubated with a polyclonal antibody to CCSP (rabbit antihuman CCSP was generously provided by Dr. G. Singh, V.A. Medical Center, Pittsburgh, PA) that was diluted 1:1,000 in 5% milk/TBS. The reactivity and specificity of the antibody preparation against human CCSP were reported previously (29). After incubation with the primary antibody, membranes were washed three times with TBS, followed by a 2-h incubation on a rocker table in 50 ml of a solution containing a horseradish peroxidase-conjugated polyclonal antirabbit antibody made in goats diluted 1:10,000 (Bio-Rad Laboratories) in 5% milk/TBS. Signals for immunoreactive CCSP were generated on autoradiograms by enhanced chemiluminescence, and CCSP contents were quantitated using methods identical to those described above for DNPH-reactive proteins. For quantitative comparison of CCSP protein levels, the signals generated from samples obtained on Day of life 1 from infants who did not develop BPD were normalized to 100% in each western blot and utilized for comparison of signals generated for all other samples obtained from infants who did develop BPD, from infants who died prior to 28 d of life, and on Days of life 3 and 6 from infants who did not develop BPD.

Statistical Analyses

The infant data, acute pulmonary course data, and late pulmonary outcome data were compared by chi-square analyses of proportional variables, t tests for comparison of continuous variables between two groups, one-way analysis of variance (ANOVA) of continuous variables containing more than two groups, and repeated measures ANOVAs for analyses of serial measurements in two or more groups (30). The statistical analyses were performed on SPSS version 8.0 for windows (SPSS Inc., Chicago, IL). Statistical significance was attributed to p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Forty-five infants were enrolled in the study; 19 infants developed BPD, 16 infants did not develop BPD, and 10 infants died prior to 28 d of postnatal life. The groups were assessed for differences in gestational age, birth weight, receipt of prenatal steroids, and Apgar scores (Table 1). The infants who did not develop BPD were larger with more advanced gestational ages at birth than were the infants who died secondary to acute respiratory failure in the first 28 d of life. The proportion of infants exposed to prenatal steroids and the Apgar scores at 1 and 5 min were similar in all groups. There were no deaths in the non-BPD group before hospital discharge. All of the infants who died in the first 28 d of life died of respiratory failure, with 50% of the deaths occurring in the first 3 d of life. In the BPD group of infants, there were three deaths that occurred after 28 d of life and prior to discharge from the hospital. All deaths after 28 d of life in the infants with BPD were secondary to nonpulmonary complications of prematurity. There were no differences among groups in the number of surfactant doses administered, nor were there any differences in fraction of inspired oxygen (FI_O2) requirements between infants who subsequently developed BPD and those who did not (Figure 1). Infants who died in the first 28 d of life did have higher FI_O2 requirements in the first week of life than infants who did not develop BPD.

View larger version (19K): [in this window] [in a new window]   Figure 1.   FIO2 requirements in infants born at < 29 wk gestation associated with pulmonary outcome. Infants born at  29 wk gestation were treated with mechanical ventilation and appropriate FIO2s. Data are expressed as means ± SD; n = 16 in no BPD, n = 19 in BPD, and n = 10 in infants who died in the first month of life. Repeated measures ANOVA indicated that FIO2 requirements were higher in infants who died in the first 28 d of life than in infants who did not develop BPD. There was no effect of day of life, nor was there an interaction between patient group and day of life.

In the first week of life, total protein concentrations in the tracheal aspirate samples of infants who subsequently developed BPD or died prior to 28 d were higher than in infants who did not develop BPD (Figure 2). All tracheal aspirate samples from all infants had multiple DNPH-reactive proteins as indicated by the western blot illustrated in Figure 3. Please note that the western blot depicted in Figure 3 contains samples from one patient who developed BPD and two who did not. Interestingly, despite having higher protein concentrations in their tracheal aspirates than patients who did not develop BPD, infants who developed BPD did not have higher levels of total DNPH-reactive proteins in their tracheal aspirates than did patients who did not go on to develop BPD (Figure 4). Figure 4 also contains data from western blots that included samples from infants who died prior to 28 d of life. Infants who died had higher levels of total DNPH-reactive proteins than did infants who survived 28 d and did or did not develop BPD.

View larger version (25K): [in this window] [in a new window]   Figure 2.   Protein concentrations in tracheal aspirate samples. Tracheal aspirate samples were collected on Days of life 1, 3, and 6 from infants born  29 wk gestation. Total protein concentrations in the samples were determined. Data are expressed as means ± SEM; n = 16 in no BPD, n = 19 in BPD, and n = 10 in infants who died in the first month of life. Repeated measures ANOVA indicated that protein concentrations were higher in samples from infants who developed BPD or died than in those patients who did not develop BPD. There was no effect of day of life, nor was there an interaction between patient group and day of life.

View larger version (76K): [in this window] [in a new window]   Figure 3.   Western blot analysis of tracheal aspirate samples for the detection of 2,4-dinitrophenylhydrazine (DNPH)-reactive proteins. Tracheal aspirate samples were collected on Days of life 1, 3, and 6 from infants born at  29 wk gestation. Samples were incubated with DNPH, which reacts with aldehydes, ketones, and certain other products of protein oxidation, and DNPH reactivities were assessed by western blotting with an anti-DNPH-reactive antibody. This figure is a representative western blot that contains samples from two infants who did not develop BPD and one infant who did develop BPD as indicated at the top of the figure. There were many DNPH-reactive proteins noted in all samples, and similar levels of total DNPH-reactive proteins in the various samples. However, the DNPH-reactive signal at approximately 9 kD (arrow) was observed more frequently in samples from infants who developed BPD than in samples from the infants who did not develop BPD.

View larger version (33K): [in this window] [in a new window]   Figure 4.   Total 2,4-dinitrophenylhydrazine (DNPH)-reactive proteins in tracheal aspirate samples. Tracheal aspirate samples were collected on Days of life 1, 3, and 6 from infants born at  29 wk gestation. Samples were incubated with DNPH, which reacts with aldehydes, ketones, and certain other products of protein oxidation, and DNPH reactivities were assessed by western blotting with an anti-DNPH-reactive antibody. Total anti-DNPH immunoreactivities were quantified by densitometry. The results were normalized to levels from samples obtained on Day of life 1 from infants who did not develop BPD. Data are expressed as means ± SEM, n =16 in non BPD, n = 19 in BPD, and n = 10 in infants who died in the first month of life. Repeated measures ANOVA indicated that infants who died prior to 28 d of life had higher levels of DNPH-reactive proteins in their tracheal aspirates than did infants who survived for 28 d and did or did not develop BPD. There was no effect of day of life, nor was there an interaction between patient group and day of life.

Despite not having a difference in total levels of DNPH- reactive proteins, infants who subsequently developed BPD were more likely to have specific DNPH-reactive protein migrating at approximately 9 kD (Figure 3, see arrow) than infants who did not develop BPD. The DNPH-reactive protein migrating at approximately 9 kDa on any of the 3 d assessed was observed in 31% of the infants who did not develop BPD, 68% of the infants who did develop BPD, and 80% of the infants who died earlier than 28 d of life (Figure 5A). In each group, the proportion of infants who exhibited DNPH-reactive protein at 9 kD in tracheal aspirate samples varied as a function of postnatal age, and the data for Days 1, 3, and 6 are displayed in Figure 5B (note that there were smaller numbers of infants on Days of life 3 and 6 in the group of infants who died, secondary to early deaths). In infants who did not develop BPD, the proportion of infants who had observable DNPH-reactive protein at 9 kD decreased from 19% on Day of life 1 to 6% on Day of life 6, whereas in infants who developed BPD, the proportion increased from 31% to 47% over the same time frame. In the infants who died, the DNPH-reactive protein migrating at 9 kD increased from 60% on Day of life 1 to 100% on Day of life 3. At 6 d of life, however, only five of the original infants in this group were still alive, and two of the five surviving infants showed DNPH-reactive protein at 9 kD. The corresponding protein band on the membrane stained with Coomassie blue was excised and analyzed by N-terminal amino acid sequencing. The amino acid sequence of the excised protein was 100% identical through 12 amino acids to the previously published N-terminal sequence of human CCSP (31).

View larger version (31K): [in this window] [in a new window]   View larger version (34K): [in this window] [in a new window]   Figure 5.   DNPH-reactivity of a 9-kD protein in tracheal aspirate samples. Tracheal aspirate samples were collected on Days of life 1, 3, and 6 from infants born at  29 wk gestation. The samples were incubated with DNPH, and the reactivity with DNPH of a protein of approximately 9 kD was assessed. (A) The proportion of infants who were observed to have DNPH reactivity, at a signal-to-noise ratio of five or greater, at 9 kD on western blots from tracheal aspirate samples collected on Day of life 1, 3, or 6 were determined and analyzed by chi-square analysis. (B) The proportion of infants who were observed to have DNPH reactivity at 9 kD present in the tracheal aspirate fluids collected Days of life 1, 3, or 6, as indicated. In both A and B, chi-square analyses indicated that infants who developed BPD or died of respiratory failure prior to 28 d of life more commonly exhibited observable DNPH reactivity at 9 kD than did infants who survived and did not develop BPD. Data are expressed as percent of the total infants from each group; n = 16 in non BPD, n = 19 in BPD, and n = 10 in infants who died in the first month of life. *Different from the non-BPD group, p < 0.05.

To determine whether the higher levels of DNPH-reactive CCSP in the tracheal aspirate fluids of infants who subsequently developed BPD or died were associated with increases in the amounts of immunoreactive CCSP in the tracheal aspirate fluids, samples were assayed for CCSP by western blotting (Figure 6). On Day of life 1, the level of immunoreactive CCSP was higher in the infant who did not develop BPD than in the infant who did develop BPD and in the infant who died. Furthermore, the levels remained higher in the infants who did not develop BPD than in the other two groups over the first 6 d of life.

View larger version (48K): [in this window] [in a new window]   Figure 6.   Western blot for Clara cell secretory protein (CCSP) immunoreactivity in tracheal aspirate fluids. Tracheal aspirate samples were collected on Days of life (DOL) 1, 3, and 6 from infants born at  29 wk gestation. CCSP immunoreactive protein levels were assessed by western blotting. CCSP immunoreactivities were higher on Day of life 1 in aspirate samples from infants who survived and did not develop BPD than in samples from infants who developed BPD or from infants who died of respiratory failure prior to 28 d of life. In addition, levels of CCSP immunoreactivity increased in the non-BPD infants, whereas similar increases in CCSP levels are not observed in the other two groups.

The results from all of the western blots were combined and analyzed (Figure 7). Infants who did not develop BPD had higher levels of CCSP on Day of life 1 than were observed in infants who developed BPD and in infants who died in the first month. In addition, infants who did not develop BPD showed significant increases in the expression of immunoreactive CCSP on Days of life 3 and 6, whereas the infants who did develop BPD or died showed no increases in CCSP immunoreactivities over the first week of life.

View larger version (20K): [in this window] [in a new window]   Figure 7.   Immunoreactive CCSP in tracheal aspirate fluids as a function of postnatal age. Tracheal aspirate fluids were collected on Days of life 1, 3, and 6 from infants born at  29 wk gestation. Samples were analyzed for CCSP immunoreactive protein by western blotting. CCSP immunoreactivities were quantified by densitometry, values in the non-BPD patients on Day of life 1 were normalized to 100%, and the values in the other groups were compared. On Day of life 1, the CCSP expression levels in tracheal aspirate samples were more than twice as high in the non-BPD group than in the other two groups. CCSP immunoreactivities were increased markedly on Days of life 3 and 6 in infants who did not develop BPD, whereas there was no induction in infants who developed BPD or in infants who died of respiratory failure prior to 28 d of life. Data are expressed as means ± SEM; n = 16 in non-BPD, n = 19 in BPD, and n = 10 in infants who died in the first month of life. *Different from other groups at the same time point, p < 0.05.

The surviving infants in the no BPD group and the BPD groups were evaluated at the time of discharge from the hospital for exposure to postnatal steroids, length of supplemental oxygen treatment, the need for oxygen therapy at 36 wk corrected postconceptual age, the need for home oxygen therapy at the time of discharge, and for total hospital stay (Table 2). Three infants in the BPD group expired prior to discharge, secondary to nonpulmonary complications of prematurity. The infants who developed BPD were more likely to have been treated with postnatal steroids. Thirteen of the 19 infants who developed BPD received steroids at some point in their hospital course. However, the course of steroids was administered after the first week of life in every case and was used in an attempt to wean from the ventilator. Four out of 10 of the infants who died prior to 28 d of life received steroids before 3 d of life for refractory hypotension. Infants who developed BPD were exposed to supplemental oxygen longer, were more likely to need oxygen therapy at 36 wk postconceptual age, and, despite longer hospital stays, were more likely to go home being treated with supplemental oxygen than were the infants who did not develop BPD.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The incidence of classical BPD induced by aggressive treatment of hyaline membrane disease (HMD) is declining, secondary to advances in perinatal management of premature infants. However, the incidence of chronic lung disease in the absence of acute lung disease remains a major cause of infant morbidity and mortality (32). The high incidence of BPD in our study population reflects the high survival rates of very small newborns in our nursery and is consistent with an inverse relationship of birth weight and BPD. The differences in weights and gestational ages at birth we observed in our study groups (Table 1) reflect recognized risk factors for development of BPD and infant mortality (2, 3, 5). However, biomarkers are needed for the specific mechanisms through which tissue damage is expressed, to elucidate critical mechanisms contributing to disease evolution and to guide efforts to develop and apply therapeutic interventions.

The important findings in this study are that immunoreactive CCSP levels in tracheal aspirates are lower on Day of life 1 in infants who subsequently develop BPD than in infants who do not develop BPD, and CCSP levels are not induced in the first week of life in the BPD group, whereas CCSP levels in the infants who do not develop BPD are induced dramatically. Furthermore, these alterations in CCSP immunoreactivity occur in the context of more frequent DNPH reactivity of CCSP in infants who develop BPD than in infants who do not. The extent to which oxidation of CCSP contributes to loss of immunoreactivity, if at all, is not clear from the data available.

The greater concentrations of proteins in tracheal aspirate samples obtained during the first week of life from the infants who would develop BPD or die within the first month than from the infants who would survive and not develop BPD (Figure 2) suggest a potential screening biomarker of risk, but greater specificity is needed to unravel possible mechanisms of disease. In contrast to the differences in total protein concentrations, levels of total DNPH-reactive proteins were not different in the tracheal aspirates of infants who developed BPD and infants who did not. This discrepancy suggests that the increase in total protein concentrations in the BPD infants is due to a larger leakage of serum proteins, in which a relatively low proportion is oxidized, into the tracheobronchial fluid than in infants who do not develop BPD. The fact that the levels of total DNPH-reactive proteins were not different in the two groups of infants is consistent with nearly equivalent levels of supplemental oxygen and supports the idea that loading equal volumes in western blotting for DNPH-reactive proteins represents loading relatively equal volumes of protein derived from tracheobronchial secretions. The fact that infants who died prior to 28 d had levels of DNPH-reactive proteins that were markedly higher than the other two groups as early as Day of life 1 suggests that these infants were developmentally ill prepared for exposure to mechanical ventilation and/or supplemental oxygen administration, and that total DNPH-reactive proteins in tracheal aspirates are a biomarker for risk of death. It is also important to note that some of the differences between the group that died prior to 28 d of life and the other two groups may be secondary to the significant differences in gestational age.

Oxidative alterations of critical biological molecules are logical contributing mechanisms for the development of BPD, but consistent differences in overall DNPH reactivities were not observed between the infants who would develop BPD and those who would not. Different biological molecules are unlikely to be equally critical to normal host function, and different covalent modifications of the same molecule will not necessarily effect similar physiological consequences.

Oxidation of proteins in tracheal fluids has been associated with poor neonatal outcomes (20). These oxidative modifications may arise from activation of inflammatory responses (7, 33) and/or inadequate functioning of primary antioxidant regulatory systems in the lung (17, 34). However, the roles that protein oxidation plays in lung injury are not clear, in part because of the lack of specific assessments of protein oxidation. Enhanced lung inflammation and deficient antioxidant enzyme systems are processes that have been noted in acute hyperoxia-induced lung injury and in human infants who were at high risk for the development of BPD.

Oxidation of CCSP could be particularly damaging to the lung, if the oxidative modifications disrupt the antiinflammatory functions of this protein in the airway tree. CCSP is the major secretory product of Clara cells, which are nonciliated epithelial cells localized to airways down to the level of terminal bronchioles. Clara cells are thought to be pleuripotent cells that function as progenitor cells for ciliated and nonciliated cells in the airway epithelium (35). CCSP is abundant in tracheal fluid (29) and has been reported to function as an antiinflammatory agent in the lung by inhibition of phospholipase A₂ (36). Thus, oxidation of CCSP and/or diminished expression of CCSP may lead to deleterious lung inflammatory responses, lung damage, and the subsequent development of BPD. Transgenic mice devoid of CCSP expression are more susceptible to hyperoxic lung injury (37), suggesting that CCSP plays important roles in protecting lungs exposed to hyperoxia, perhaps by attenuating or regulating lung inflammatory responses. Additionally, CCSP has been reported to function as a binding protein for both phosphatidylcholine and phosphatidylinositol and has been hypothesized to function as a transporter of these molecules for phospholipid recycling in Clara cells (38).

In addition to the increased oxidation of CCSP, in the present studies, we clearly observed that infants who developed BPD had lower tracheal aspirate contents of immunoreactive CCSP on the first day of life than did infants who did not develop BPD. Furthermore, infants who developed BPD had no increase in immunoreactive CCSP in tracheal aspirate samples, whereas infants who did not develop BPD had marked induction of CCSP expression in the first week of life (Figure 7). The correlation of poor pulmonary outcome of prematurely born infants with the marked attenuation of levels of immunoreactive native CCSP underscores the possibility that CCSP may have a key role in lung homeostasis and lung responses to oxidant stresses and mechanical ventilation. Diminution of CCSP expression could result from decreased protein production, Clara cell destruction, or decreased immunoreactivity of oxidized CCSP. Although the mechanisms for attenuated postnatal induction of CCSP in infants who develop BPD or die secondary to respiratory failure have not been defined, it is clear from the present study that diminished postnatal CCSP expression is a predictor for poor neonatal lung outcomes and that the reasons for the differences in CCSP expression between groups should be pursued.

Infants who developed BPD had longer hospitalizations and increased need for home therapy with supplemental oxygen (Table 2). These findings underscore the importance of determining the specific mechanisms involved in the development of BPD. The association of CCSP oxidation and abnormalities in CCSP expression with BPD development suggests that alterations in CCSP in the lung may play a significant role in the development of BPD. Furthermore, investigations into the mechanisms regulating CCSP expression and oxidation are needed and may lead to mechanism-based interdictions designed to improve pulmonary outcomes in infants at high risk for developing BPD or respiratory failure.

In summary, our findings suggest that deficiencies in CCSP resulting from diminished production and/or oxidative inactivation, perhaps mediated by inflammatory processes, may contribute to development of lung injury. These results also raise the possibility that CCSP may play a central role in normal airspace homeostasis. Early alterations in CCSP oxidation or expression may be useful for the identification of subsets of prematurely born infants who have greater sensitivities to lung oxidant stresses and for the development of adverse pulmonary outcomes. Further studies to delineate the mechanisms for diminished CCSP expression in infants who develop BPD are needed for the rational design of therapeutic interventions to preserve CCSP expression and function.