INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

There is increasing recognition that infection plays an important role in the pathogenesis of bronchopulmonary dysplasia (BPD) and neonatal chronic lung disease (CLD) in premature infants (1). In these infants, IgG levels are low at birth because of premature interruption of transplacental transfer of antibodies, since efficient transfer of maternal antibodies occurs only after 34 wk of gestation (2). Even the full-term neonate is known to be vulnerable to infection (3). Neonates, particularly those born before term, are poor at mounting an immunologic response to polysaccharide antigen (4), and have diminished serum opsonic capability (5). Neonatal CLD is the newer clinical name for the disease originally called BPD. Although both diseases require mechanical ventilation and supplemental oxygen therapy for prematurely born infants with respiratory failure, improvements in respiratory care and the use of exogenous surfactant have reduced the severity of lung disease in this population. Extremely immature infants who do not experience respiratory distress syndrome (RDS) are those who usually develop CLD (1).

The clinical course of infants with the formerly named BPD was often complicated by bacterial, fungal, or viral infections, which induced respiratory failure and death in those infants with severe disease (6). Patients with CLD may also experience this complication. Recently, Rojas and colleagues identified nosocomial infection as an important pathogenic factor in the development of CLD (1), and others have implicated colonization by Ureaplasma urealyticum as a major contributor to the development of CLD (7).

Surfactant proteins A and D (SP-A and SP-D), designated lung collectins, have been implicated in the host defense against viral, fungal, and bacterial microorganisms (8, 9). A deficiency in SP-A and/or SP-D as a consequence of lung immaturity or of injuries associated with ventilatory support (e.g., from oxygen radicals or inflammatory mediators) might influence the susceptibility of the neonate to infection and thereby interfere with clinical recovery. This possibility is consistent with the results of a study of newborn infants with CLD conducted by Hallman and colleagues (10). These investigators collected tracheal aspirates from infants of 24 to 29 wk gestational age over a 2-wk period, and measured SP-A and saturated phosphatidylcholine (SatPC). They found that a low ratio of SP-A to Sat PC was predictive of mortality, although a direct correlation with indices of infection was not developed. There is no information on the importance of SP-D in the clinical course of CLD.

The experiments described in this report were done to examine changes in surfactant-associated host-defense proteins that accompany the development of CLD in the premature neonatal baboon. In our previous studies, baboons of 140 d gestational age (g.a.) developed BPD in response to aggressive ventilatory support with 100% oxygen (11, 12). For these studies, however, we used conditions similar to those experienced by premature human infants with CLD, who develop infectious complications despite postnatal management with antibiotics (1). Animals were delivered at 125 d g.a. (68% of term) after prenatal steroid treatment of the dam, received exogenous surfactant, and were ventilated for extended periods of 1 to 2 mo with appropriate oxygenation and a low-volume ventilatory strategy. These primates spontaneously developed CLD complicated by infection, as is often seen with extremely immature human infants delivered at 24 to 28 wk gestation (1).

We report here the results of analyses of SP-A and SP-D from two groups of these animals. The first group of animals (Group I) was ventilated without clinical incident for up to 2 wk. Lavage pools of SP-A and SP-D after 6 d of ventilation of these animals were relatively small compared with the amounts for adult (or term newborn) animals, but in both cases increased progressively with time of ventilation. The animals of the second group (Group II) were also delivered at 125 d g.a., but were ventilated for periods ranging from 5 d to 2 mo. These animals were killed because of a deteriorating disease course that could not be reversed despite vigorous respiratory management. Deaths before 27 d were invariably due to severe infections, and infection was usually a factor in those animals with CLD that succumbed after 1 mo. SP-A in lavage fluid from these animals was severely reduced even after 2 mo of ventilation. We also demonstrated a correlation between a low concentration of lavage fluid SP-A and the release of interelukin (IL)-8; and with an "infection index" based on histopathology, microbiologic cultures, and clinical indications of sepsis. Our data suggest that the amount of SP-A in lavage fluid is an indicator of the risk of infection in the evolution of neonatal CLD, and that a reduced amount of SP-A may diminish the ability to resolve this condition.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

General Animal Care

The study was done at the Southwest Foundation for Biomedical Research in San Antonio, Texas. All animal husbandry and animal handling and procedures were reviewed and approved to conform to guidelines of the American Association for Accreditation of Laboratory Animal Care. Details of the animals' care have been published (13). Pregnancies were dated with cycle dates, and growth parameters were obtained from prenatal ultrasound examinations at 70 and 100 d estimated fetal g.a. Infants used in the study were delivered by hysterotomy at 125 ± 6 d (68% of a term gestation of 185 d). At birth, all animals were weighed, sedated with intramuscular ketamine hydrochloride (10 mg/kg), and intubated with a 2.5-mm I.D. endotracheal tube. Tracheal lung fluid was collected before the infants' first breath. All animals received 4 ml/kg of exogenous surfactant (Survanta; donated by Ross Laboratories, Columbus, OH) before ventilatory support was begun. Ventilation was initiated with a humidified, pressure-limited, time-cycled infant ventilator (InfantStar; donated by Infrasonics, San Diego, CA) with an initial rate of 40 breaths/min, a peak inspiratory pressure (PI_max) adequate to move the chest, a positive end-expiratory pressure (PEEP) of 4 cm H₂O, and an inspired oxygen fraction (FI_O2) of 0.40. Animals had an umbilical arterial catheter and percutaneous central venous catheter inserted, and were nursed in a servo-controlled, infrared-heated, whole-body plethysmograph (VT1000; VitalTrends Technology, New York, NY). Subsequent ventilator adjustments were made on the basis of chest radiographs, clinical examination, arterial blood gas measurement, and tidal volume (VT) measurement, as subsequently described. Intermittent sedation was provided as needed with ketamine (5 mg/kg) and/or diazepam (0.1 mg/kg).

The ventilatory approach used in the study was based on a strategy designed to maintain VT at 4 to 6 ml/kg as continuously measured with the VitalTrends system and associated with adequate chest motion by physical examination. The respiratory rate (RR) was adjusted as required to regulate the animals' arterial carbon dioxide tension (Pa_CO2) at 45 to 55 mm Hg. High-frequency oscillatory ventilation (HFOV; SensorMedics 3100A; SensorMedics, Anaheim, CA) was instituted as rescue therapy if an air leak developed (pneumothorax, pneumatocele, interstitial emphysema) or if the Pa_CO2 rose above 65 mm Hg with a PI_max above 40 cm H₂O and an RR of 60 breaths/min.

Target goals for Pa_O2 were 55 to 70 mm Hg. Oxygenation was primarily manipulated through changes in PEEP and its effect on mean airway pressure (Paw), and by alterations in FI_O2. HFOV was applied as rescue therapy if the arterial oxygen tension (Pa_O2) was below the target goal at a Paw > 16 cm H₂O and FI_O2 = 1.00.

During the first 24 h of life, all animals received heparinized normal saline via the umbilical artery catheter, and a 5% dextrose-in- water infusion with supplemental calcium via the central venous catheter. Initial fluid volume intakes for the first day of life were calculated to provide 250 to 300 ml/kg/d, but were subsequently decreased over the first 3 to 4 d to approximately 175 to 200 ml/kg/d. These initial fluid intakes were necessary to maintain electrolyte homeostasis, to provide a minimal urine output of 1 to 2 ml/kg/h, to maintain an acceptable blood pressure, and to minimize metabolic acidosis. Parenteral nutrition was initiated at 24 h of life with amino acids at 1.25 g/kg/d (Trophamine; McGaw, Inc., Irvine CA), and electrolytes, vitamins (Pediatric MVI; Astra, Westborough, MA; or Cernevit; Clintec, Deerfield, IL), and trace elements (MTE-5; Fujisawa USA, Deerfield, IL). The amino acid input was increased to 2.5 g/kg/d at 48 h of life, and L-cysteine (0.60 mmol/kg/d) was added at 72 h of life.

Arterial blood gases were measured hourly for the first 24 h of life, and then at 2-h intervals from 24 to 48 h, 4-h intervals from 48 h to 96 h, and 6- to 12-h intervals as determined by clinical needs. Electrolytes and hematocrit were monitored every 12 to 24 h. Complete blood chemistry analyses and cell counts were done weekly. To maintain the hematocrit at 30% to 45%, we administered packed red blood cells periodically, using fresh heparinized blood from adult baboon donors. Local anesthesia with 2% lidocaine and additional ketamine hydrochloride were administered for any invasive procedures.

Care Unique to Group II

Dams of infants in Group II were treated with 6 mg of intramuscular betamethasone at 24 h and 48 h before hysterotomy. Extubation of the infants was attempted after 7 d of age, in the absence of respiratory distress, and if target blood gases were maintained with a ventilatory rate < 15 breaths/min, PI_max < 20 cm H₂O, PEEP < 4 cm H₂O, and FI_O2 < 0.30. To support long-term parenteral nutrition, a 20% lipid emulsion (Intralipid; Pharmacia & Upjohn, Clayton, NC) was initiated on Day 7 of life, and was increased to 2.5 g/kg/d. If the infant was clinically stable, enteral nutrition was initiated on Day 7 of life. Donated human breast milk was given by intermittent gastric infusion at an initial volume of 10 ml/kg/d, and was increased by 5 to 10 ml/kg/d as tolerated. Once enteral intakes of 100 ml/kg/d were tolerated, enteral feeding was changed to Primilac (Bio-Serv, Frenchtown, NJ). Nutritional goals included a volume intake of 150 to 200 ml/kg/d, an energy intake of 80 to 120 cal/kg/d, and a protein inake of 3.0 g/kg/d. Because of early colonization with Pseudomonas, antibiotics were changed in Group II animals to amikacin/ceftazidime or amikacin/ piperacillin from Days 4 to 7, and were then discontinued. Subsequent antibiotic therapy, as needed for clinically suspected sepsis, included the use of vancomycin and the anti-Pseudomonas regimen just described. Because of several early deaths associated with Candida sepsis, prophylactic fluconazole was initiated in all animals at 6.0 mg/kg/ dose at 12 h, 96 h, and 168 h of age. Doses were then given every other day until Day 28 of life. Prophylactic intravenous immunoglobulin (Sandoglobulin; Swiss Red Cross, Berne, Switzerland) was also given, at a dose of 400 mg/kg, on Days 5 and 21 of life.

Light Microscopy and Immunocytochemistry

At the time of necropsy, the right lower lobe of each infant baboon was removed, weighed, and fixed intrabronchially for 24 h with phosphate-buffered 4% paraformaldehyde and 0.1% glutaraldehyde at 20 cm H₂O constant pressure. The lobe was cut into three serial, equally spaced, horizontal tissue sections. The entire cut surfaces of all three horizontal sections were processed for light-microscopic study. These specimens were dehydrated in alcohol, embedded in paraffin, cut at a thickness of 4 µm, and stained with hematoxylin and eosin.

Assessment of Infection

Microbiologic cultures were obtained when infection was clinically suspected. The pattern of tracheal colonization with microbes was found to be similar to that in the 140-d baboon CLD model characterized fully in an earlier study (14) (i.e., initially sterile, with subsequent acquisition of gram-positive bacteria, especially Staphylococcus epidermidis, followed by gram-negative bacterial colonization). Clinical chart reviews noting radiographic worsening of an animal's clinical course, microbiologic results of tracheal aspirate and blood cultures examination, and lung tissue cultures obtained under sterile conditions at necropsy were evaluated for most of the animals. A clinical assessment of ongoing sepsis was accepted if at least three of the following criteria were met: acute change in cardiovascular stability; deterioration in pulmonary function accompanied by nonhomogeneous radiographic densities; neutropenia; thrombocytopenia; or increasing metabolic acidosis.

We defined a semiquantitative infection index that was calculated as follows: Animals were assigned an index grade of from 0 (normal) to 4 on the basis of the histopathologic evaluation done at the time of killing. These criteria are defined in Table 1. Mild, focal bronchopneumonia was defined as the presence of scattered neutrophils within occasional bronchioles and the immediately subjacent alveoli/saccules. Moderate bronchopneumonia was diagnosed when more bronchioles and peribronchiolar sites showed neutrophilic and fibrinous exudates. Confluence or consolidation, and/or necrosis of the lung parenchyma by a fibrinopurulent exudate, was designated as severe or necrotizing bronchopneumonia. In addition, the animals were graded for clinical indications of sepsis on a scale of 0 = none, and 1 = probable sepsis. The Infection Index was the sum of the histopathologic and sepsis grades. Blood and tissue cultures were analyzed under both aerobic and anaerobic growth conditions. Specific fungal cultures were done on blood and tissue specimens. Viral cultures were not done.

Assay of IL-8 in Bronchoalveolar Lavage Fluids

At necropsy, the left lower lobe was lavaged with 0.9% saline. An enzyme immunoassay (PE Biosystems, Foster City, CA) was used to measure IL-8. The assay sensitivity was 100 pg/ml and the intra- and interassay coefficients of variation were 10% and 24%, respectively. Although there is no consensus about the appropriate reference protein for protein analysis of bronchoalveolar lavage fluid (BALF), the latter was done with the Pierce Bicinchoninic Acid Protein Assay (Pierce Chemical Co., Rockford, IL).

Preparation of Riboprobes

The preparation of riboprobes has been described (15). For SP-A, a 165-bp sequence of human complementary DNA (cDNA) was chosen for its homology to the exon III region of baboon SP-A cDNA. For SP-D we used the entire cDNA for human SP-D. After characterization and purification of the plasmids, they were linearized with appropriate restriction enzymes. Antisense riboprobes for SP-A and SP-D messenger RNA (mRNA) were prepared by in vitro transcription, using T7 RNA polymerase enzyme (MAXIscript kit; Ambion Inc., Austin, TX) in the presence of [-³²P]uridine triphosphate ([-³²P]UTP) (800 Ci/mmol; NEN Dupont, Boston, MA). Sense probes of the same sequences were used as negative controls. The [-³²P]UTP-labeled riboprobes (SP-A and SP-D riboprobes) were purified on denaturing 5% acrylamide/8 M urea gels and were eluted in elution buffer containing 0.5 M ammonium acetate, 1 mM ethylenediamine tetraacetic acid, and 0.1% sodium dodecylsulfate (SDS).

The properties of the probes have been described (15). Northern blots for SP-A and SP-D mRNA showed 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 Northern Blot Analysis

Total RNA from lung tissues was extracted with Trireagent (Molecular Research Center, Cincinnati, OH) according to the manufacturer's instructions.

For Northern blot analysis, 10 µg and 20 µg of total lung RNA was electrophoresed on 1% formaldehyde-agarose gel, and was transferred onto nylon membranes (Nytran; Schleicher & Schuell, Keene, NH). Total lung RNA (10 µg and 20 µg) of a normal adult baboon was also run in each analysis for standardization. The membranes were probed overnight with [-³²P]UTP-labeled antisense SP-A and SP-D cRNA probes (0.5 × 10⁶ cpm/ml and 0.2 × 10⁶ cpm/ml, respectively) used together in hybridization buffer at 60° C, and were washed and exposed to X-ray film at 70° C for 72 to 96 h. SP-A and SP-D mRNA levels were quantified by densitometric scanning of autoradiograms, using NIH Image version 1.60 software for Macintosh computers. Densitometric readings of SP-A mRNA and SP-D mRNA were normalized with 18s ribosomal RNA that was quantified by densitometric scanning of photographic negatives.

Enzyme-Linked Immunosorbent Assay of SP-A and SP-D

Lung tissue homogenates and lavage samples, diluted in 0.1 M NaHCO₃ buffer, pH 9.6, were incubated overnight in multiwell Immulon-4 polystyrene strips (Dynatech, Chantilly, 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, pH 8.5. The wells were washed and incubated for 2 h with rabbit antibody to human-SP-A or SP-D. After washing, 50 µl of alkaline phosphatase-conjugated antirabbit IgG antibody was added and incubated for 2 h. The wells were washed again, and 75 µl of 4-methylumbelliferyl 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, Norwalk, CT) at an excitation wavelength of 365 nm and 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 (SD = 0.01) and 0.97 (SD = 0.01), respectively. Intraassay variabilities ranged from 3.5% to 9.8% (SP-A) and 4.3% to 10.3% (SP-D). Interassay variabilities were 3.0% to 14.1% (SP-A) and 12.8% to 15.6% (SP-D). The lower limits of detection in enzyme-linked immunosorbent assays (ELISAs) for both proteins were comparable at 2 to 4 ng/ml. The curve of serial dilutions of lung homogenate was approximately linear over a 1,000-fold range (r² = 0.98); lavage fluid was tested over a 1,000-fold range (r² = 0.99).

The ELISA antibodies (to human SP-A and -human SP-D) were developed in rabbits, using the full-length surfactant proteins. The antibodies were specific and did not cross-react. Both antibodies recognized both monomeric and multimeric forms of their respective proteins (16).

To estimate the effectiveness of the assay in quantitating the total amount of surfactant protein in the samples, we added known amounts of SP-A to samples of lung homogenates and lavage fluids, and measured the incremental increases in SP-A concentration. We detected 75% (SEM = 2.7%) of the added SP-A in homogenates and 89% (SEM = 4.9%) in lavage fluids, suggesting comparable efficiencies of detection of the endogenous pools of these surfactant proteins.

Data Analyses

Statistical results were generated with Statview (Abacus Concepts, Berkekey, CA) or SAS (Cary, NC) software. Significance was taken at p < 0.05 unless otherwise indicated.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Group I Animals

Morphology and infection. Table 1 shows the IL-8 levels in necropsy BALF, results of pathologic evaluation, presence of positive blood and/or lung cultures, and results for clinically suspected sepsis (change in ventilatory requirements, decrease in platelet count, onset of acidosis, X-ray change) in the infants in Group I. None of the animals in Group I had any clinical indication of pulmonary infection or sepsis. Only one of the four 6-d-study animals had specimens examined for histopathology, and results were normal. In prior studies, 11 6-d- old animals were examined after ventilation with the same protocol. Only one showed any histopathologic evidence of pulmonary infection, which was designated as a mild bronchopneumonia. No organisms were identified with special stains.

Lung infection was not seen on gross examination of the 14-d lungs at necropsy, nor was it evident histopathologically. Concentrations of IL-8 were below 1,000 pg/ml except in one animal (Animal I H). Despite the elevated IL-8 level, the lung histopathology of this animal was negative for infection, and the animal's clinical course was normal. The IL-8 data are consistent overall with our earlier finding that IL-8 levels are low in animals without severe intrapulmonary infection (13).

Steady-State Levels of SP-A and SP-D mRNA

Both SP-A and SP-D mRNA levels increased from control levels (125 d g.a.) after 6 d and 14 d of ventilation. However, the rate of development of SP-A and SP-D mRNA differed, with that of SP-D mRNA preceding that of SP-A mRNA. SP-D mRNA was 50% of the value for adult baboons after 14 d of ventilation. In contrast, SP-A mRNA was only 13% of the adult level (p < 0.05) (Figure 1).

View larger version (20K): [in this window] [in a new window]   Figure 1.   Steady-state levels of SP-A and SP-D mRNA in 125-d g.a. animals and animals ventilated for 6 d and 14 d (Group I animals). *p < 0.05 and  p < 0.10 as compared with adult SP-A and SP-D mRNA levels.

Tissue pools of SP-A and SP-D. Tissue pools of SP-A and SP-D were quantified with ELISA, with the results shown in Table 2. The data are presented in three ways: the amount of SP-A and SP-D per unit of lung wet weight, the amount of SP-A and SP-D per unit of lung dry weight, and the amount of SP-A and per SP-D per total lung (animal). All comparisons are reported on the basis of µg protein/g lung dry weight. In nonventilated infants of 125 d g.a., the SP-A concentration was below the lower limit of detection of the ELISA, whereas that of SP-D was over 75% of the amount in adult baboons, previously reported (13). Tissue pools of both SP-A and SP-D, and especially of SP-D, increased with ventilation. After 6 d of ventilation, the amount of SP-D was over 20 times (comparison by µg/g lung dry weight) and that of SP-A was 125% of the values for adult animals (Table 2). With further ventilation, the amount of SP-D decreased, but after 14 d was still fivefold greater than for adults.

Lavage pools of SP-A and SP-D. Neither SP-A nor SP-D was present in significant amounts in nonventilated infants of 125 d g.a. (Table 3). Ventilation for 6 d resulted in an increase in lavage SP-A to about 14% of the adult value; with 14 d of ventilation this increased to 35%. In contrast, lavage SP-D exceeded the adult value by threefold after 6 d, and by more than 10-fold after 14 d. Combined SP-A and SP-D values were 23% and 88% of the respective adult values after 6 and 14 d of ventilation. In calculating the amounts of both proteins in lavage fluid as fractions of the amounts in tissue, it is evident that the distribution between lavage fluid and tissue of both proteins, and especially of SP-D, increased with postnatal maturity (Table 4).

Group II Animals

Morphology and infection. The deaths that occurred before 21 d of ventilation in Group II were primarily sustained during the first year of development of our long-term baboon model of CLD. These early deaths were due to severe infections, and provided a one-time source from which surfactant data might be analyzed over the early course of development of CLD. Thereafter, management of infection improved, with survival to 30 d being about 70%. The data are summarized in Table 1, and the major pathogens isolated from each of the Group II animals are presented in Table 5. Histopathologically, bronchopneumonia of fungal or gram-negative bacterial origin was seen in the lungs of nine of the 18 animals in Group II. Only four animals did not have an active infectious process in the lungs at death. With few exceptions, IL-8 levels of > 1,600 to 1,800 pg/ml correlated with diffuse consolidating lung infections. Limited lung involvement by an infectious lesion (e.g., one septic embolus, a single microscopic abscess) was not associated with increases in IL-8 above 1,800 pg/ml, with one exception. Animal II K, clinically thought to be septic, died at 28 d and did not have gross or microscopic evidence of infection in the right lung used for histopathologic study. However, on gross inspection, the lavaged left lung showed gross sites of consolidation, indicating that a sampling difference was probably responsible for the disparity.

Tissue pools of SP-A and SP-D. The SP-A and SP-D concentrations (µg/g lung dry weight) in tissue for each of the 18 animals in Group II are shown in Figure 2. For most animals, the tissue SP-A concentration was equal to or exceeded that of adult animals within 14 d of postnatal ventilation, and was unchanged through 2 mo. The appearance of SP-D in tissue was relatively precocious, and in most animals its concentration equaled or exceeded that of adult animals within 5 d.

View larger version (17K): [in this window] [in a new window]   Figure 2.   SP-A (left panel ) and SP-D (right panel ) concentrations (µg/g lung dry weight) in lung tissues of Group II animals. Amounts of SP-A and SP-D in adult animals are also shown.

Lavage Pools of SP-A and SP-D

The SP-A and SP-D concentrations (µg/g lung dry weight) in lavage fluid for each of the 18 animals in Group II are shown in Figure 3. The lavage fluid SP-A content in 14 of the 18 animals was, on average, 20% of the amount in adult lavage fluid, even after 2 mo of ventilation. Only four of the 18 animals in Group II had lavage fluid SP-A concentrations greater than that of the Group I animals ventilated for 14 d. In contrast to those of SP-A, concentrations of SP-D were generally equal or greater than those of adults within 11 d, but were markedly less than those of the Group I animals ventilated for 14 d. However, the combined lavage pool of surfactant collectins (SP-A and SP-D) never exceeded 60% of the combined pool in normal adults, except in three animals (Animals II E, II M, and II R), and two of these three had very low IL-8 concentrations.

View larger version (19K): [in this window] [in a new window]   Figure 3.   SP-A (left panel ) and SP-D (right panel ) concentrations (µg/g lung dry weight) in lavage fluids of Group II animals. Amounts of SP-A and SP-D in adult animals are also shown.

The ratios of lavage fluid to tissue SP-A and SP-D are shown in Figure 4. The averages of both the SP-A and SP-D ratios were relatively low compared with those of adults; the average SP-A ratio was about 0.16 and the average SP-D ratio about 0.17 of adult ratios (adult ratios for both SP-A and SP-D are about 0.5). However, the ratios were highly varied among individual animals, and there was no trend with time of ventilation. About 60% of the animals had ratios of SP-A that were less than the 14-d Group I average; all animals had ratios of SP-D that were less than the Group I average. This result is considerably different than that observed in the Group I animals who did not develop infection and CLD, in which it was found that ratios of both SP-A and SP-D increased with time of ventilation.

View larger version (19K): [in this window] [in a new window]   Figure 4.   Lavage/tissue ratios of SP-A (left panel ) and SP-D (right panel ) of Group II animals on the basis of lung dry weight. Lavage/tissue ratios of SP-A and SP-D in adult animals are also shown.

Relationship of IL-8 concentration and infection index to SP-A concentration. A scattergraph relating lavage SP-A concentration to the concentration of lavage IL-8 for all of the Group II animals ventilated for more than 14 d is shown in Figure 5A. All three animals with SP-A concentrations exceeding 72 µg/g lung dry weight (the average concentration of SP-A in the 14-d Group I animals) also had IL-8 concentrations of less than 1,600 pg/ml. Only three of six animals with IL-8 concentrations below 1,600 pg/ml had SP-A concentrations below 72 µg/g lung dry weight. The chi-square distribution was significant (p < 0.025). The mean SP-A concentrations (µg/g lung dry weight) in animals with IL-8 concentrations < 1,600 pg/ml and those with IL-8 concentrations > 1,600 pg/ml were 75.7 (SEM = 18.3) and 17.5 (SEM = 3.8), respectively (p < 0.01 by analysis of variance [ANOVA]).

View larger version (13K): [in this window] [in a new window]   View larger version (10K): [in this window] [in a new window]   View larger version (10K): [in this window] [in a new window]   Figure 5.   Scattergraphs of Group II animals ventilated for > 14 d, showing chi-square distributions of values. (Left panel ) IL-8 (pg/ml) and SP-A concentrations in lavage fluids. Line parallel to x-axis indicates average lavage fluid SP-A concentration (72 µg/ g lung dry weight) found in Group I animals ventilated for 14 d. Line parallel to y-axis indicates IL-8 level of 1,600 pg/ml in lavage fluid. (Right panel ) Infection indices and SP-A concentrations in lavage fluids of Group II animals ventilated for > 14 d. Line parallel to x-axis indicates average Group I lavage fluid SP-A concentration (72 µg/g lung dry weight). Line parallel to y-axis indicates infection Index of 3.0. Animal IIK was not used in this evaluation because of obvious differences in pathology between the right lung, used for histopathology, and the left lung, used for lavage (see text). (Bottom panel ) Scattergraph of lavage fluid SP-A concentrations versus days of ventilation. Line parallel to y-axis shows average Group I lavage fluid SP-A concentration (72 µg/g lung dry weight).

A scattergraph of lavage fluid SP-A concentration versus the infection index is shown in Figure 5B. There is a correlation between low concentrations of SP-A and higher infection indices in Group II animals ventilated for more than 14 d (chi-square distribution; p < 0.1). The amounts of SP-A (µg/g lung dry weight) in animals with infection indices < 3.0 and > 3.0 were 66.3 (SEM = 18.1) and 19.6 (SEM = 4.9), respectively (p < 0.05, ANOVA). In contrast, there was no correlation of SP-A concentration with time of ventilation (Figure 5C).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this report, we present data that support a defective host-defense role for SP-A and/or SP-D in the lungs of immature baboons with CLD that developed naturally acquired lung infections. Histopathologic evaluation, as well as lung and blood cultures, indicated that most of the animals in Group II in our study were infected, which led to their unplanned deaths. In addition, these animals showed increased concentrations of IL-8 in the lung. This cytokine, which can attract and activate neutrophils, has been shown to be a good indicator of long-term infection, and is increased in other disorders associated with a neutrophilic alveolitis (13, 17, 18). Recently, Bancalari and Gonzalez implicated acute pulmonary infections as a major complication in the clinical course of human infants with either mild or severe CLD (19), a finding that strengthens the findings in earlier studies in which bacterial, fungal, and viral infections were shown to worsen the evolution of BPD (6).

The best evidence for a host-defense role of SP-A has come from studies of SP-A-null transgenic mice (16, 20). These animals show a defective response to a variety of microorganisms instilled into the respiratory tract. However, increased susceptibility to naturally acquired bacterial or fungal infections with SP-A deficiency has not been described, and host-defense functions of SP-D have not yet been definitively demonstrated in vivo.

We cultured both gram-positive and gram-negative organisms from specimens taken from baboons with CLD. However, it was the gram-negative bacteria, and especially Pseudomonas aeruginosa, Staphylococci, and Candida species that were most clearly implicated as the significant pathogens in this model. This is particularly interesting, given the increased susceptibility of SP-A-null mice to challenge with P. aeruginosa. SP-A-null animals show increased proliferation and decreased phagocytosis of P. aeruginosa following intratracheal instillation of bacteria, and the wild-type response is restored in the presence of purified SP-A (16). SP-D could also contribute to the host response to this organism. For example, purified SP-D enhances the internalization and killing of at least one mucoid strain of P. aeruginosa by macrophages in vitro (21). Both SP-A and SP-D are decreased in the lavage fluid of patients with cystic fibrosis (22), a condition known to be complicated by chronic infections with specific strains of P. aeruginosa. SP-A-null mice also show increased susceptibility to challenge with certain gram-positive organisms, including Group B streptococci and Staphylococcus aureus (20). In addition, there is in vitro evidence that SP-A and SP-D contribute to the host defense against fungi (23, 24).

Abnormalities in Collectin Pools

Group I. As expected, we found that nonventilated fetal baboons at 68% of term had very little SP-A. These animals, however, were able to develop, within 6 d of postnatal ventilation, tissue pools of SP-A that were comparable to those of normal adults. Alveolar pools of SP-A lagged behind the tissue pools; nevertheless, after 6 d and 14 d of ventilation, lavage fluid SP-A was about 14% and 35%, respectively, of that in adults or term neonates, clearly demonstrating a progression toward normal adult levels. We suspect that with longer ventilation, alveolar SP-A would continue to increase. In contrast to SP-A, the appearance of SP-D was precocious, and even by 6 d of ventilation SP-D in lavage fluid was about equal to that of adults.

Group II. In the animals in Group II, the deficiency in lavage fluid levels of SP-A was considerably more severe and long-lasting than that seen in the Group I animals. Lavage fluid SP-A was generally less than 20% of that in adults, and the combined lavage pool of surfactant collectins was less than 25% of the adult value even after 2 mo of ventilation. Tissue levels of SP-A were relatively high, suggesting that there is no defect in its synthesis upon ventilation.

These data indicate that changes in SP-A induced in the condition of premature injury/infection are remarkably long-lived, possibly contributing to compromised host defense and a self-reinforcing cycle of enduring infection. Our results further indicate that the lavage pool of SP-A is markedly reduced in the setting of lung infection and sepsis, even though tissue pools contain normal amounts of the protein. The alveolar concentrations of SP-D were comparable to those of normal adult baboons, but we do not interpret this finding as indicating that SP-D homeostasis is normal in prematurity with lung infection/sepsis. The ratio of alveolar to tissue pools of SP-D was as low and varied as that of SP-A, even though the tissue levels were much higher than those of normal adults.

Potential Mechanisms for Alterations in Collectin Pools

The mechanism responsible for the observed differences in lavage and tissue pools of SP-A and SP-D is uncertain. The homeostasis between cellular and alveolar pools of surfactant lipid and protein involves a complicated interplay between synthesis, secretion, reuptake, and intra- and extracellular degradation (25). Thus, decreased steady-state levels of airspace SP-A could reflect decreased entry into and/or increased removal of proteins from the airspace compartment. As indicated earlier, the tissue pools of SP-A in our premature baboons were high, suggesting that there is no defect in its biosynthesis. However, there are at least two possibilities that could explain a hypothetical decrease in the release of SP-A into lavage fluid in the face of normal or increased SP-A production.

First, SP-A in the tissue pool could be damaged or incorrectly folded, thereby preventing its secretion. Many studies have shown that the intracellular transport and release of collagen is impaired if its structure is perturbed (26). In our experiments, the antibody we used recognized SP-A in ELISAs, Western blots, and immunoprecipitation procedures (data not shown). In addition, we were unable to detect differences in the size distribution of SP-A in blots of any sample of fetal, postnatal, or adult lung tissue or lavage fluid. However, these analyses cannot exclude more subtle structural alterations in SP-A.

Second, increased amounts of cytokines and growth factors are released in association with infection, and consequently altered cytokine/growth factor environment could affect SP-A metabolism. SP-A production is inhibited by tumor necrosis factor (TNF)- (27, 28) and transforming growth factor- (29) in cultured cells. However, little is known about the regulation of surfactant metabolism either by growth factors and cytokines or by the myriad of other mediators that are released in altered amounts at various times following lung injury. We found a statistical correlation between IL-8 and SP-A concentration, but know of no metabolic studies to support a direct cause-and- effect relationship for this finding. Although the concentrations of certain other proinflammatory cytokines, including TNF-, IL-6, and macrophage inflammatory protein, are increased in the lungs of SP-A-null mice, we suspect that the relationship between IL-8 and SP-A is indirect, and that the increased IL-8 simply reflects the presence of lung infection.

As suggested earlier, increased loss of collectins from the lavage pool could also have contributed to the observed deficiency in SP-A, particularly in Group II animals. Although we did not observe high-molecular-weight degradation products on Western blots, degradation to smaller fragments or loss of antigenic determinants cannot be excluded. In this regard, Wright and coworkers have shown that both SP-A and SP-D can be internalized and degraded to trichloroacetic acid soluble peptides by alveolar macrophages in vitro (30, 31). Such degradation might be further increased in the setting of infection. Although SP-A might also be lost by adsorption to cell surfaces, this cannot explain most of the differences observed in our study, since cell-associated SP-A in our hands accounted for only 0.5% to 1.5% of the total alveolar pool of SP-A (15).

The observed differences in SP-A and SP-D pools suggest that the production, secretion, and/or turnover of the two proteins are differentially regulated. This interpretation is confounded by uncertainty about the predominant site(s) of synthesis and mechanisms of secretion of SP-A and SP-D. A number of cell types synthesize and secrete SP-A and SP-D in human and rodent lung. For example, SP-A and SP-D are synthesized and secreted by both type II cells and nonciliated bronchiolar cells (32, 33). SP-A is also synthesized by human bronchial submucosal glands (34), and by baboon bronchial epithelium and ductal cells of submucosal glands (32). Both SP-A and SP-D are probably subject to regulated secretion by granule exocytosis from Clara cells (33). In some studies, SP-A secretion by isolated type II cells is stimulated by phorbol 12-myristate 13-acetate (PMA) (35), but the published findings regarding this are not consistent (36, 37). SP-D appears to be constitutively secreted by type II cells as a soluble protein (38), and is not stimulated by PMA (35). In addition, developmental differences in the time course of human SP-D and SP-A expression have been described. For example, SP-D mRNA appears earlier in gestation than that of SP-A (39), and is preferentially increased by glucocorticoids (40).

In summary, the study reported here is the first to implicate lung collectin deficiency in the host defense against bacterial and fungal organisms in a model of naturally acquired infection in premature neonates. Given this, the findings may be relevant in the clinical management of premature human infants. In the first week of life of the very immature neonate, the normal alveolar pool of SP-A/SP-D may be substantially reduced with or without infection. It may still be low even after 2 wk of ventilation. If the neonate develops CLD, the reduced collectin pool could be even more persistent. The interruption of normal gestation by delivery, and the ventilatory support needed to sustain the extremely immature baboon, arrests ongoing developmental processes in the lung (e.g., vasculogenesis and alveolization), and arrested alveolization has also been documented in the human infant with CLD (41). We suspect that surfactant protein homeostasis may be dysfunctional for prolonged periods in infants receiving prolonged ventilatory support, and may in part explain the extreme susceptibilty of these infants to microbial infections. This scenario is different from that of older children and adults in whom normal amounts of collectins are present before the onset of a viral infection, but are decreased during the infectious episode (42). These findings suggest that replacement therapy with lung collectins, particularly SP-A, may be efficacious in the management of neonatal CLD, and that further studies in thist direction may be warranted.