INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Bronchopulmonary dysplasia (BPD), a condition of chronic lung injury in the premature neonate, is associated with increased morbidity and mortality (1). The factors implicated in its pathogenesis include prematurity of the lung with a lack of pulmonary surfactant and acute injury induced by barotrauma and oxygen toxicity. About 90% of pulmonary surfactant is lipid in nature and 5 to 10% is protein (2, 3). This study concentrates on two of the proteins in alveolar lining material, SP-A and SP-D. Both SP-A and SP-D proteins are hydrophilic collagenous glycoproteins (collectins) expressed in type II epithelial cells and nonciliated cells of airways, and SP-A may be important in determining the physical properties of the surfactant complex (4, 5). In addition, both SP-A and SP-D may have important roles in nonimmune host defense (6). Comparative studies suggest overlapping functions. These include the induction of respiratory burst in macrophages, the binding of viruses and bacteria through carbohydrate groups, the interaction with receptors in macrophages to stimulate phagocytosis, and the inhibition of viral infectivity (11). Most extracellular SP-A in lavage from adults is bound to surfactant; nearly all SP-D is unbound. Whether there is a physiologic advantage in these respective physical states is unknown.

Infection is a significant problem in patients with BPD who require prolonged periods of ventilation (12), and this could be influenced by changes in collectins. Results from studies using animals with acute and chronic injury induced with high oxygen indicate that SP-A mRNA and protein levels increase during acute hyperoxia in neonatal rabbits and adult rats (13, 14), whereas the levels decrease during prolonged exposure (15). In contrast, the regulation of SP-D protein in neonatal injury has largely been unstudied. To our knowledge, there are no available reports on the effects of chronic lung injury on the expression of SP-D mRNA and SP-D protein in lungs of premature infants. Further, there is only one report on the developmental expression of SP-D in primates (16), and because this was done using human abortuses, the gestational time-span of the observations was necessarily limited.

This study was undertaken to quantify SP-A and SP-D in the baboon during normal gestational development, and to compare changes in a baboon homologue of neonatal chronic lung injury (BPD). Our results indicate that in gestational development, SP-D appears earlier in the distal respiratory epithelium than does SP-A. With chronic lung injury induced with 100% O₂, SP-A and SP-D concentrations in tissue exceed those of the normal adult by 5- and 16-fold, respectively. Lavage concentration of SP-A, however, is 7% of that of normal adults and lavage SP-D just equals the amount in normal adults. The combined lavage SP-A/SP-D pool is only 12% of normal adult despite this normal level of SP-D. Because premature animals are already less able than adults to mount an immunologic response, this decreased concentration of surfactant host-defense proteins may further augment a proclivity to infection and worsening injury.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

The animal experiments were performed according to the National Institutes of Health Animal Use and Care Guidelines and were approved by the Institutional Animal Care Committee of the University of Texas Health Science Center at San Antonio and Southwest Foundation for Biomedical Research, San Antonio.

A detailed description of the baboon model of BPD has been reported elsewhere (17). Gestational ages (ga) of the baboons were determined by matings timed by the observation of female perineal sex skin changes and were confirmed by ultrasound examination during pregnancy.

A total of 16 baboons (delivered at 140 d ga) were divided into two groups. Group A was managed on a pro re nata (PRN) basis, i.e., ventilated with oxygen levels to maintain appropriate PO₂ and PCO₂ levels for either 6 d (6-d PRN O₂ group; n = 4) or 10 d (10-d PRN O₂ group; n = 4) prior to being killed. The Group B baboons were placed on positive pressure ventilation with 100% oxygen for a period of either 6 d (6-d 100% O₂ group; n = 4) or 10 d (10-d 100% O₂ group; n = 4). The nonventilated gestational control groups of animals were delivered at 125 d (n = 4), 140 d (n = 4), or 175 d (n = 4) ga. For comparison with adults, normal adult baboons (n = 4) were also included in the study. All the manipulations of the animals were done under ketamine immobilization (10 mg/kg, given intramuscularly) and the animals were killed by an overdose of nembutal (25 mg/kg, given intravenously). All animals received Travisol-based total parenteral nutrition (TPN) (2.25 g/kg body weight/d). Antibiotics were administered as follows: gentamicin (2.5 mg/kg), beginning at 6 h of life, then daily through Day 10; ampicillin (50 mg/kg), beginning at 7 h of life and then every 12 h through Day 10. Less than 10% of the animals developed metabolic acidosis, and there were no significant patent ductus arteriosus (PDA) problems.

At necropsy, the right lung was weighed (wet weight), and a segment of the lung was removed and immediately frozen in liquid nitrogen for RNA and protein analysis. To study the protein levels in epithelial lining fluid, a preweighed lobe of lung was lavaged with 0.15 M NaCl (pH, 7.4) with a recovery of 70 to 80% of the instilled volume. The lavage fluid was centrifuged at 1,500 rpm for 15 min to remove macrophages and cellular material. The total volume of cell-free lavage obtained was measured and stored at 80° C. The left lower lobe was removed, weighed, and intrabronchially fixed with phosphate-buffered 4% paraformaldehyde and 0.1% glutaraldehyde at 20 cm H₂O constant pressure for 48 h. It was then sectioned into three equally spaced transverse levels, embedded in Paraplast, sectioned at 4 µm, and stained for histopathologic and immunocytochemical studies.

Preparation of Riboprobes

For measuring SP-A mRNA by northern blots, a cDNA clone was prepared by subcloning a 165 bp fragment of human SP-A cDNA. The human cDNA was generously provided by Dr. Brad Benson. The 165 bp sequence of human cDNA was chosen for its homology to exon III of baboon SP-A cDNA (18). Human SP-D cDNA (1.4 kb), cloned in pGEM3Z plasmid and containing the entire coding sequence for human SP-D protein, was used for SP-D northern blots (19). After characterization and purification of the plasmids, they were linearized using appropriate restriction enzymes. Antisense riboprobes for SP-A and SP-D mRNA were prepared by in vitro transcription using T7 RNA polymerase enzyme (MAXIscript kit; Ambion Inc., Austin, TX) in the presence of [-³²P]UTP (800 Ci/mmol; Dupont NEN Research Products, Boston, MA). The [-³²P]UTP labeled riboprobes (SP-A- and SP-D-riboprobes) were purified on a denaturing 5% acrylamide/ 8 M urea gel and were eluted in elution buffer containing 0.5 M ammonium acetate, 1 mM ethylene diamine tetraacetic acid (EDTA), and 0.1% sodium dodecyl sulfate (SDS). Similarly, single-stranded, unlabeled, sense SP-D cRNA (1.4 kb) was prepared in the same manner and served as a negative control for northern blots. Northern blots for SP-A and SP-D mRNA show that the SP-A riboprobe hybridized with a major SP-A transcript of about 3.0 kb, whereas the SP-D riboprobe hybridized with a 1.4 kb single transcript of SP-D mRNA. The respective riboprobes were specific and did not cross-react.

RNA Extraction and Norther Blot Analysis

Total RNA from lung tissues was extracted using Trireagent (Molecular Research Center, Cincinnati, OH), according to the manufacturer's instructions. Briefly, lung tissue (100 to 250 mg) was crushed with a mortar and pestle on dry ice and homogenized in 1 ml Trireagent solution using tissue homogenizer. The homogenate was centrifuged in the presence of chloroform at 12,000 g at 4° C for 15 min and precipitated with isopropanol at room temperature. The RNA pellet was suspended in diethyl pyrocarbonate-treated water and quantified spectrophotometrically at 260 nm.

For northern blot analysis, 10 µg and 20 µg of total lung RNA were electrophoresed on 1% formaldehyde-agarose gel and transferred onto nylon membranes (Nytran; Schleicher & Schuell, Keene, NH). Total lung RNA (10 and 20 µg) of a normal adult baboon was also run each time. The membranes were baked at 80° C for 1 h and prehybridized at 60° C for 4 h in hybridization buffer containing 50% formamide. The membranes were probed overnight with [-³²P]UTP- labeled antisense SP-A and SP-D cRNA probes (0.5 × 10⁶ and 0.2 × 10⁶ cpm/ml, respectively) together in hybridization buffer at 60° C. The membranes were washed three times for 20 min each with buffer containing 0.18 M NaCl, 10 mM NaH₂PO₄, and 1 mM EDTA at pH 7.7 (SSPE) and 0.1% SDS at 65° C followed by washing with 0.1× SSPE and 0.1% SDS for 50 min at 60° C. The blots were subsequently exposed to X-ray films (Kodak, XAR-5) at 70° C for 72 to 96 h. For quantitation of SP-A and SP-D mRNA levels, autoradiograms were scanned by densitometer using NIH Image 1.60 software for Macintosh computers. Densitometric readings of SP-A mRNA and SP-D mRNA were normalized with 18s rRNA, which was quantified by densitometric scanning of photographic negatives.

Quantitation of SP-A and SP-D Protein Content

The lung tissue and lavage fluid samples were quantified by enzyme-linked immunosorbent assay (ELISA). Antibody specificity was verified by western blots. Samples of lung tissues were prepared by thawing 100 to 200 mg tissue on ice and homogenizing in ice-cold lysis buffer (1 ml/100 mg tissue) containing 1% igepal CA-630 (similar to NP-40), 0.1% SDS, 1 mM EDTA, 1.1 µM leupeptin, 1 µM pepstatin, and 0.2 mM phenylmethyl sulfonyl fluoride. The tissue homogenates were centrifuged at 2,000 rpm for 10 min to remove debris, and the supernatants were stored at 80° C. The total protein concentration in homogenates and lavage fluids was estimated by micro bicinchoninic acid method (Pierce Chemical Co., Rockford, IL).

Western Blots

The purified SP-A protein and rabbit antihuman SP-A antibody were provided by Dr. Brad Benson. The purified recombinant human SP-D protein and rabbit antihuman SP-D antibody were prepared as previously described (20). To check the specificity and titers of these antibodies, purified SP-A and SP-D proteins and lung tissue and lavage samples of adult baboons were run on 12% reducing SDS polyacrylamide gels. The proteins were transferred overnight onto polyvinylidene fluoride membranes and the membranes were blocked by 15% nonfat milk and 4% goat serum. The membranes were then incubated for 1 h with either antihuman SP-A or SP-D antibodies, washed, and further incubated for 1 h with goat antirabbit IgG conjugated to horseradish peroxidase before detection by enhanced chemiluminescence reagents (Amersham, Arlington Heights, IL). The antibodies, rabbit antihuman SP-A, and antihuman SP-D were specific for their respective antigens. Antihuman SP-A antibody recognized bands of 31 and 66 kD likely to be SP-A monomer and dimer. Antihuman SP-D antibody recognized a 43-kD SP-D and its probable dimeric and trimeric forms.

ELISA

Lung tissue homogenates and lavage samples, diluted in 0.1 M NaHCO₃ buffer at pH 9.6 were incubated overnight in multiwell Immulon-4 polystyrene strips (Dynatech Laboratories, Alexandria, VA). The wells were rinsed three times with deionized water, and nonspecific sites were blocked with a buffer containing 0.25% bovine serum albumin, 0.05% Tween-20, 0.17 M boric acid, and 0.12 M NaCl at pH 8.5. The wells were washed and incubated for 2 h with either rabbit antihuman-SP-A or SP-D antibody. After washing, 50 µl of alkaline-phosphatase-conjugated antirabbit IgG antibody were added and incubated for 2 h. The wells were washed again and 75 µl of 4-methyl-umbelliferyl phosphate (4-MUP) substrate solution containing 0.2 mM 4-MUP, 0.05 M Na₂CO₃ and 0.05 mM MgCl₂ was added. The fluorescence was read in a spectrofluorometer (Perkin-Elmer Medical Instruments, Oakbrook, IL) at an excitation wavelength of 365 nm and an emission wavelength of 450 nm. The regression coefficients for a least-squares linear fit to the standard curves of both SP-A and SP-D were 0.99. Intra-assay variabilities ranged from 3.5 to 9.8% (SP-A) and 4.3 to 10.3% (SP-D). Inter-assay variabilities were 3.0 to 14.1% (SP-A) and 12.8 to 15.6% (SP-D). The limits of detection in both ELISAs were comparable: 2 to 4 ng/ml.

To estimate the effectiveness of the assay in recovering the total amount of surfactant protein in the samples, we added known amounts of SP-A to samples of lung homogenates and lavages and measured the incremental increases in SP-A concentration. We detected 65% (SEM, 2.7%) of the added SP-A in homogenates and 89% (SEM, 4.9%) in lavages, suggesting comparable efficiencies of detection of the endogenous pools.

In Situ Hybridization

Frozen sections 5-µm thick were obtained from tissue specimens of the infracardiac or the right middle lobes that had been stored in liquid nitrogen. Thereafter, the procedures for cryosections followed those published by Zeller and Rogers (21). Riboprobes were used for the in situ hybridization studies. The specific DNA fragments of SP-A and SP-D were subcloned into pGEM4Z (Promega, Madison, WI), an appropriate transcription vector that contains promoters for SP6 and T7 RNA polymerases. The cDNA probes used in these studies contained the entire coding regions for human SP-A and SP-D. The orientation of the inserts dictated which enzymes were used to transcribe the probe from the appropriate linearized plasmids in order to obtain both an anti-mRNA transcript (the antisense probe) and a sense transcript (the control). In addition to the sense transcript (negative control), an antisense probe for ceruloplasmin was used as a positive control. Ceruloplasmin is known to have a specific expression in the airway epithelium, and its expression is not related to that of the surfactant proteins. The transcription reaction was performed using a Promega kit. Twenty-four microliters of [5'--³⁵S]-UTP (specific activity, 1,000 to 1,500 Ci/mmol) (Dupont NEN Research Products) were added in each reaction. The RNA pellets were precipitated with ethanol, dissolved in 3.3 mM dithiothreitol, and then adjusted so that each microliter of the hybridization buffer with the probe contained about 80,000 to 200,000 cpm. All of the hybridized slide preparations were exposed for 5 d and then developed.

Immunocytochemistry

Paraffin sections 4-µm thick were used for localizations of surfactant apoproteins. Blocking procedures for endogenous activity of peroxidase and avidin were utilized, and the labeled avidin biotin (LAB) method was performed. Controls included: (1) omission of the primary antibody to check for endogenous peroxidase activity and nonspecific binding of secondary antibody; (2) replacement of the primary antibody with nonspecific immune serum from the same species; and (3) replacement of the primary antiserum with another primary antiserum to a totally different antigen than the one under investigation.

Antibodies against human SP-A (kindly provided by Dr. J. Whitsett) and SP-D were used to immunostain the proteins. Details of the purification of SP-A and generation of antibody against it have been described by Khoor and colleagues (22). Rabbit antihuman SP-A antibody reacts strongly with both glycosylated and nonglycosylated isoforms of SP-A. The rabbit antihuman-SP-D antibody was prepared as previously described (20).

Statistical Analysis

One-way analysis of variance (ANOVA) was used to assess the statistical significance. Significance was assumed for p  0.05, with p  0.1 also noted.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pathologic Changes

These findings have been detailed previously (23), but are briefly reviewed. After 6 d of PRN O₂ ventilation, specimens show mild changes of microatelectasis and rare foci of edema and/or residual hyaline membranes. When treated with 100% O₂ for 6 d, an early exudative diffuse alveolar lesion (DAD) of focal edema in saccular (saccule = primitive alveolar structure) walls and air spaces, increased numbers of alveolar macrophages, and occasional neutrophils were evident.

After 10 d of ventilation with 100% O₂, specimens had hyperplastic and/or metaplastic epithelial changes in airways, an altered inflation pattern of alternating sites of atelectatic and overexpanded saccules/alveoli, and increased fibrosis around airways and within saccular walls. PRN O₂-treated specimens lacked the airway lesions and distal fibroproliferative lesions and showed only focal saccular hemorrhages and scattered alveolar macrophages. No evidence of pneumonia was present in any of the specimens studied.

In Situ Hybridization and Immunocytochemistry

SP-A mRNA expression in specimens from the 6-d PRN O₂ group was present in bronchial and bronchiolar epithelia and scattered type II epithelial cells, whereas SP-D mRNA was localized in bronchiolar and type II epithelial cells. In the 6-d 100% O₂ group, the SP-A mRNA expression in bronchiolar epithelium was similar to that seen in the PRN O₂-treated specimens, but more type II cell localization was evident. The specimens from the 6-d 100% O₂ group showed heavy SP-D mRNA expression in the bronchiolar and type II epithelia.

The animals ventilated with PRN O₂ for 10 d had more type II epithelial localization of SP-A mRNA, but it was still not intense. Specimens from the 10-d 100% O₂ animals showed a very patchy expression of SP-A mRNA, which, when present, was heavily localized in cells that tended to be subjacent to bronchioles and/or vessels. In much of the parenchyma, there was no expression of SP-A mRNA. SP-D mRNA in specimens from both PRN O₂- and 100% O₂-treated animals showed bronchiolar and type II epithelial cellular localization, which overall tended to be comparable in intensity.

SP-A immunocytochemical preparations of 6-d PRN-treated and 100% O₂-treated specimens were similar; protein was seen in scattered type II and in some bronchiolar epithelial cells. Conversely, SP-D stained preparations from animals ventilated for 6 d were dramatically different, with much more staining present in bronchiolar and type II cells in the 100% O₂-treated specimens than in the PRN O₂-treated specimens. In the 10-d ventilated animals, this same pattern of more SP-D cellular protein staining was evident in the 100% O₂-treated specimens (Figure 1). Interestingly, the SP-A-stained preparations from the 10-d 100% O₂ group also showed more protein staining in airway and type II cells when compared with those of PRN O₂-treated (for 10 d) preparations (Figure 2).

View larger version (44K): [in this window] [in a new window]   Figure 1.   10-d PRN and 100% O2 SP-D immunostains. The PRN control shows scattered positive bronchiolar epithelial cells and the cuboidal alveolar type II epithelial cells subjacent to the bronchiole are positive (A). The 100% O2-treated specimen shows interstitial fibroproliferation and enlarged hyperplastic type II epithelial cells, which are positively immunostained with SP-D (arrows) (B). b = bronchiole. Immunostains; original magnification: ×150.

View larger version (48K): [in this window] [in a new window]   Figure 2.   10-d PRN and 100% O2 SP-A immunostains. SP-A immunopositivity is seen in only scattered epithelial type II cells and on the surface of the ciliated airway epithelium (A). In the 100% O2-treated specimen, the hyperplastic type II epithelial cells that line the thickened saccular walls contain abundant SP-A protein (arrows) (B). Small bronchus = B. Immunostains; original magnification: ×150.

Changes in SP-A and SP-D mRNA Expression

The steady-state levels of mRNAs for SP-A and SP-D in nonventilated gestational control animals are shown in Figure 3. SP-A mRNA was barely detectable in baboons of 125 and 140 d ga, but at 175 d ga it was 62% of adult SP-A mRNA (p < 0.05). In contrast to SP-A, SP-D mRNA in baboons of 125 and 140 d ga was 3 and 12% of adults, respectively. It increased sharply to about 200% of that of adult level in baboons of 175 d ga (p < 0.05).

View larger version (29K): [in this window] [in a new window]   Figure 3.   SP-A and SP-D mRNA levels in lung tissue samples of nonventilated fetal baboons of 125, 140, and 175 d ga. Adult levels are assigned a value of 1 and are shown for comparison. *Values different from those of adults (p < 0.05).

When premature baboons of 140 d ga were ventilated with PRN or 100% O₂, both SP-A mRNA and SP-D mRNA increased from control levels (Figure 4); the increases after 6 d of ventilation were much greater using 100% O₂ than with PRN O₂ (p < 0.05). By 10 d, however, SP-A mRNA levels were comparable in both groups. Significant differences (reductions) from normal adults were found in SP-A mRNA levels at all the time points in both ventilation protocols; SP-D was reduced only after 6 d of PRN O₂.

View larger version (41K): [in this window] [in a new window]   View larger version (30K): [in this window] [in a new window]   Figure 4.   SP-A (upper panel ) and SP-D (lower panel ) mRNA levels in baboons ventilated with either PRN O2 or 100% O2. The mRNA levels are shown relative to adults, which are given a value of 1. *Values different from those of adults (p < 0.05).

Tissue Pools of SP-A and SP-D

Western blots of the isoforms of SP-A detected in tissue and lavage samples of animals ventilated for 10 d with 100% O₂ and PRN O₂ are shown in Figure 5. There were no major differences in the detectable isoforms in the ventilated animals as compared with normal adults or with nonventilated gestational controls (not shown). Likewise, SP-D-detected isoforms were identical to those in adult tissues (not shown).

View larger version (43K): [in this window] [in a new window]   Figure 5.   Western blots of SP-A protein in baboons ventilated for 10 d with either PRN O2 or 100% O2. Upper panel: 1 mg total protein from lung tissue. Lower panel: 1 mg total protein from lavage. Within each panel: lane 1, blank lane; lanes 2-4, three samples ventilated with 100% O2; lanes 5-8, four samples ventilated with PRN O2; lane 9, SP-A standard. The upper band in lanes 2-8 (both panels) is probably an artifact, which is also observed in lane 1 (blank).

Tissue pools of SP-A and SP-D proteins were quantitatively analyzed by ELISA and the results are shown in Table 1. The data are presented in four ways, but the discussion in the text is based on quantitation by (g protein/g lung dry wt). In nonventilated fetuses of 125 and 140 d ga, SP-A protein concentration was below the lower limit of detection of ELISA, whereas SP-D concentration at 125 d ga was about 50% of the levels in adult baboons: 125 d ga comparison based on (g protein/g lung wet wt). Tissue pools of both SP-A and SP-D increased with gestational age, approximately following changes in mRNA. By 175 d ga, SP-A was 8-fold greater than adult; SP-D 6-fold greater.

Tissue pools of both SP-A and SP-D increased after ventilation with both PRN and 100% O₂ ventilation, but the increases with 100% O₂ were substantially higher than those in the PRN O₂ groups for the same days of treatment (p < 0.05). With 100% O₂ ventilation, SP-A in tissue was more than five times greater than that in adults (p < 0.05); SP-D in tissue was more than 16 times greater.

Lavage Pools of SP-A and SP-D

The results are shown in Table 2. Neither SP-A nor SP-D were present in lavage in significant amounts in nonventilated fetuses of 125 and 140 d ga; by 175 d ga they were both about equal to adult levels.

Ventilation of the 140-d ga animals with PRN O₂ for 10 d resulted in modest increases in lavage SP-A to 9% (p < 0.05), whereas SP-D was not different from that of the adult. Ventilation with 100% O₂ for 6 d resulted in an increase in lavage SP-A to about 20% of the adult, but this decreased to 7% after 10 d as the injury worsened (p < 0.05). In contrast, lavage SP-D was equal or exceeded the adult in both groups of animals. Thus, in both of the ventilated groups, the amounts of SP-A found in the lavage pools after 10 d of ventilation were nearly the same, and they never exceeded 10% of that present in normal adults. In contrast, SP-D equaled or exceeded adult amounts (Table 2). However, because of the relatively small amounts of SP-D compared with SP-A, the combination of SP-A and SP-D in lavage in animals ventilated 10 d with 100% O₂ was 12% of the normal adult.

The ratios of lavage to tissue pools of SP-A and SP-D in different groups of animals are shown in Table 3. In normal adult, lavage SP-A and SP-D are about 50% of tissue pools. SP-A was not secreted in detectable amounts at 125 or 140 d ga. By 175 d ga, SP-A was 6% and SP-D was 10% of the tissue level.

Ventilation of the 140-d ga animals with 100% O₂ for 10 d resulted in only modest increases in the fraction of the tissue pool of SP-A that was releasedless than 1%. SP-D fared only slightly better3% of the tissue level after 10 d. However, probably because of the large store of SP-D in the tissue (relative to adult), the amount of SP-D in lavage after 10 d of 100% O₂ ventilation equaled the amount found in the lavage of normal adults (Table 2).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Studies of changes in surfactant proteins in neonatal chronic lung injury (BPD) are infrequent; however, the results of the study of Hallman and colleagues (24), who noted a correlation between low concentrations of SP-A in tracheal aspirates and mortality, suggest that changes in SP-A may participate in the pathogenesis of the injury. There are no similar studies for SP-D. Changes in SP-A also occur in related diseases. SP-A is reduced in human infants with RDS (25) and in adult patients with ARDs (26, 27) and bacterial pneumonia (28). Other components of surfactant are changed in BPD as well. Griese and colleagues (29) found decreases in the fraction of highly active forms of surfactant. Obladen (30) observed that survivors of BPD had higher concentrations of acetone-precipitable phosphatidylcholine in tracheal fluid that did nonsurvivors.

In addition to these clinical studies, there are now several reports that show that chronic lung injury affects surfactant constituents in experimental animals. Preterm lambs, adult rats, and neonatal rabbits show increases in SP-A mRNA during lung injury induced with hyperoxia for 1 to 3 d (14, 31, 32), but reductions in SP-A after 6 to 8 d (32). In premature baboons with 100% oxygen exposure for 11 d followed by PRN O₂ ventilation for 5 d, lung tissue SP-A was normal but SP-A protein in lavage fluid was found decreased (15). This contrasts with experiments using relatively short acute exposures to 100% O₂, in which SP-A increases (13, 14). Thus, the response of SP-A to hyperoxia is time-dependent; there are early time increases, followed by decreases with extended injury (32). There is no information on SP-D in neonatal injury.

The present study shows that both the expression and the protein secretion of saccular/alveolar SP-D precede that of SP-A in normal gestational development in the baboon but are comparable or exceed adult levels by 175 d ga. The earlier onset of SP-D in these baboons is consistent with the findings of Dulkerian and colleagues (16) in studying human fetuses and in studies using fetal mice and rats (33, 34).

SP-D is not only more precocious than SP-A in normal development but the responses of its mRNA and protein to changes sustained in chronic lung injury are greater. Tissue pools of SP-D increase to higher levels with ventilation with 100% O₂ than does SP-A, although after 10 d with 100% O₂, both are considerably higher than those of adults. Despite these ample tissue pools, lavage SP-A is significantly reduced relative to SP-A in normal adults, and SP-D in the neonatal baboons with injury just equals the concentration of SP-D in normal adults. The lavage pools of both proteins, expressed as a percentage of the tissue pools, are 1 and 3%; considerably less than the 50% found with normal adult baboons (or term baboons breathing spontaneously for 2 d; data not shown). These data suggest that the release of these surfactant proteins may be inhibited under these conditions. An alternative explanation of these findings is that the reuptake of SP-A is somehow stimulated by oxidant injury, and secretion can not keep pace. However, we know of no biologic precedent for such an effect. With PRN O₂ ventilation lavage pools more closely follow those in tissue, but still the ratio of lavage to tissue is low. Moreover, since tissue pools of SP-A are themselves relatively low, even after 10 d, the amounts of SP-A in the lavage are comparable to those in the 100% O₂ group (about 8% of that found in adult lavage).

We considered whether the decreased SP-A found in lavage was the result of SP-A binding to cells and microorganisms and being excluded from the quantified pool rather than from a real deficiency in secretion. To test for this possibility, we measured the amount of SP-A associated with the cell pellets from four animals experiencing the following protocols: 175 d ga nonventilated, 125 d ga ventilated for 14 d with PRN O₂, and 140 d ga ventilated with PRN and 100% O₂. The amounts of SP-A associated with the cell pellets ranged from 0.5 to 1.5%. Our analysis is based on the assay of all of the remaining pool of surfactant proteins found in the lavage, proteins both bound and unbound to lipid, and we can find no evidence that cell-absorbed protein is a factor.

We note that although SP-D concentration in lavage is about equal to those in adults, the combined pool of SP-A and SP-D is still only about 18% (PRN-ventilated) or 12% (100% O₂-ventilated) of the normal adult. The question becomes: How much is enough?, particularly in the premature infant where maternal IgG is low (35) and there is a reduced innate ability to mount an immunologic defense. The closest comparable experimental situation may be that of the SP-A knockout mice, and these animals show greater susceptibility to infection than do the normal controls (36). These observations suggest that the total lack of SP-A is not compensated by presumably normal amounts of SP-D. In the premature baboons studied here, SP-A in lavage is only slightly higher than that in the SP-A knockout mouse (7% of adult). We suspect that they too may be immunocompromised but remain relatively free of infection because of diligent clinical management and the use of antibiotics in this controlled experimental setting.

In summary, we find that the amount of SP-A is significantly reduced in the premature neonate with chronic lung injury despite elevated tissue pools of SP-A protein, and this would likely diminish the ability of the neonate to respond to a microbial challenge. Because infection remains a considerable problem in the management of the premature infant (37- 39), there may be merit in including SP-A and/or SP-D in the next generation of surfactant replacement drugs.