INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A sustained influx of inflammatory cells, mostly neutrophils, into the airspaces is associated with lung injury from mechanical ventilation with high inspiratory oxygen concentration and with early-onset neonatal or nosocomial bacterial infection (1, 2, 3). These cells release elastase, an enzyme capable of destroying various extracellular matrix components and proteins that play a role in host defense (4). In the lungs, several protease inhibitors, including ₁-protease inhibitor (₁-PI) and the two acid-resistant inhibitors, namely mucus protease inhibitor, or MPI (also named secretory leukocyte protease inhibitor, SLPI; bronchial secretion inhibitor, BSI; or antileukoprotease, ALP) and elastase-specific inhibitor (ESI or Elafin) are very important for balancing elastase activity. Other inhibitors (4), believed to be of minor importance, include ₁-antichymotrypsin, ₂-macroglobulin, and bikunin (derived from serum by proteolytic cleavage of inter--trypsin inhibitor, formerly designated as HI-30).

₁-Protease inhibitor is mainly expressed in hepatocytes, but has also been shown to be produced to a minor extent in macrophages and neutrophils, e.g., locally in the lungs (5). A substantial fraction of the acid-resistant inhibitors is associated with the cellular fraction of lavage samples, as previously shown (6). Elastase-specific inhibitor was expressed by these cells, whereas MPI was merely bound to neutrophils or macrophages, without expression of a significant amount of mRNA. A potential compartmentalization of these inhibitors in lavage samples and their detailed and separate analysis in the cellular and soluble fractions has not been reported.

Data on elastase in newborn and premature neonates is scarce. Functional activity of ₁-PI has been found to be reduced; inactive ₁-PI was present at unchanged or increased abundance in infants who developed bronchopulmonary dysplasia (1, 7, 8). The amount of the acid-resistant inhibitor MPI increased under similar conditions after the second week of life (9, 10), and possibly during pneumonia (10). No assessments of functional activity, however, were made in these studies. In addition, no data on the relative proportions of ₁-PI and the acid-resistant inhibitors from the same lavage samples are available. The possible compartmentalization of these inhibitors has also not been investigated. Although such data for ventilated neonates would be very desirable, we lack a comparable control group (e.g., healthy neonates on ventilation), which makes it difficult for us to evaluate deviations from the normal range. The goal of this study was to determine the activity and amount of ₁-PI and of the acid-resistant inhibitors in relation to elastase activity in both the cellular and soluble compartments of airway lavage samples.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

A homogeneously distributed sample of 42 newborn infants (gestational age 24-40 wk, mean 31.3 ± 0.63 wk; birth weight 1,773 ± 129 g, range 560-3,570 g) who were intubated and mechanically ventilated for respiratory failure on the first day of life was investigated during the first 36 h of life. Reasons for respiratory failure included suspected surfactant deficiency from immaturity (respiratory distress syndrome, 21/42); suspected perinatal infection because of a history of the prolonged rupture of the membranes, maternal fever, or infection together with the positive (14/42) or negative (15/42) cultures of pathogenic bacteria from the neonates' stomach aspirates (eight times positive in the 14 infants with positive cultures) and ears (14 times positive in these 14 infants); transient tachypnea (7/42); patent ductus arteriosus (5/42); meconium aspiration syndrome (1/42); or perinatal asphyxia (7/42). Bronchopulmonary dysplasia was diagnosed in nine of the 42 neonates, the disease being defined by the characteristic radiological features and oxygen dependency at age 28 d.

Nine neonates were mechanically ventilated for 10 d or more (11.5 ± 1.2). Seven of these infants (gestational age, 27.6 ± 0.5 wk) had all the signs that characterize the clinical course of nosocomial infection; recovery from respiratory distress during the first 4-7 d of life followed by respiratory deterioration and sepsis from Days 8-12. Those neonates with an infection acquired in the hospital were investigated longitudinally in greater detail. Only two of these infants had received prenatal corticosteroids, and four infants had received surfactant therapy. Average weight was 991 ± 93 g (range, 700-1,230 g). Ventilatory pressure support was adjusted so that the arterial oxygen partial pressure was in the range of 40-70 mm Hg, oxygen saturation was more than 90%, and the carbon dioxide partial pressure in blood was between 45-55 mm Hg. Bronchopulmonary dysplasia developed in all but one of these infants. Full or partial gavage feedings were started on Day 1 or 2 of life. This was complemented as needed by parenteral nutrition to achieve adequate caloric intake. Fluid intake was varied according to fluid balance and urine production.

For comparison, another group of 10 children without lung disease (5.2 ± 0.9 yr, range 0.9-12 yr) was included. Airway lavage samples were collected during general anesthesia for adenoidectomy (n = 5) or diagnostic bronchoscopy (intermittent stridor, n = 3, and anomaly of a great vessel, n = 2).

Airway Lavage Samples and Sample Preparation

The saline airway lavage samples were obtained by the nurses by injecting 1 ml/kg body weight of 0.45% saline in the endotracheal tube. Saline 0.45% was used routinely to reduce the salt load of the infants. When compared to 0.9% saline, no differences in viability or morphology of the cells recovered were observed (data not shown). After three to five ventilator breaths, the lung wash was collected through a small catheter in a Leuken trap by suctioning. To clear the aspirates from the sides of the catheter, 0.5 ml of 0.45% saline was rinsed through the catheter into the trap. The aspirated material was diluted in phosphate-buffered saline, pH 7.4, gently vortexed for 1 min, filtered through two layers of gauze, and centrifuged at 150 g at 4° C for 10 min. The supernatant was recovered and stored at 42° C. The cellular fractions from each day were pooled, washed two times with phosphate-buffered saline, and stored frozen.

To recover the proteases and antiproteases from the cellular fraction, an extraction in a suitable volume of a "high salt" buffer (50 mM sodium phosphate, 1.5 M KCl, pH 6.5), was performed. The samples were shaken for 2 h, centrifuged for 30 min at 20,000 g, and the supernatant containing the solubilized cellular fraction was used in further experiments. The insoluble pellet was discarded. After repeated extraction cycles of such pellets, neither elastase nor inhibitors were detected.

Materials

The polyclonal antiserum developed in rabbits, with recombinant MPI as the antigen, was a gift of Dr. Heinzel-Wieland, Grünenthal, Stollberg, Germany. The immunoglobulin G (IgG) fraction was isolated by affinity chromatography on immobilized protein G (protein G-Sepharose 4 Fast Flow; Pharmacia, Uppsala, Sweden). This antibody was shown to detect both MPI and ESI. Bovine trypsin (EC 3.4.21.4, TPCK-treated), bovine chymotrypsin (EC 3.4.21.1, TLCK-treated), N-benzoyl-L-arginyl-4-nitroanilide came from Merck (Darmstadt, Germany). Porcine pancreatic elastase (EC 3.4.21.36) was obtained from Serva (Heidelberg, Germany). Neutrophil elastase (EC 3.4.21.37) was isolated from leukocytes as described earlier (6). Cathepsin G from leukocytes (EC 3.4.21.20), N-succinyl-di-alanyl-prolyl-phenylalanyl-4-nitroanilide came from Sigma (Deisenhofen, Germany). N-succinyl-tri-alanyl-4-nitroanilide and N-succinyl-di-alanyl-valyl-4-nitroanilide were from Bachem (Heidelberg, Germany). All other biochemicals were of highest purity available from Merck.

Functional Assays for the Assessment of Proteolytic and Antiproteolytic Activity

All functional assays were carried out in 96-well microtiter plates without protein-binding capacity (Nunc, Roslide, Denmark) and performed in duplicate. The molar concentrations of trypsin solutions were determined by active site titration with 4-nitrophenyl-4-guanidinobenzoate hydrochloride, those of elastase by titration with bovine bikunin, a bivalent trypsin-elastase inhibitor that was again titrated with trypsin. The concentration of MPI was determined with active site titrated trypsin.

Total elastolytic activity and elastase-inhibiting capacity of samples was assayed, using N-succinyl-di-alanyl-valyl-4-nitroanilide as substrate. The elastase from neutrophils was the target enzyme for the determination of inhibitor activity. Wells were loaded with 50 µl of target enzyme each (5 × 10¹⁰ M) dissolved in buffer (100 mM HEPES, 2 mM NaCl, 0.001% Triton X100; pH 8.2). For blanks, 100 µl of buffer was added. Suitable volumes of samples (5 µl on average), dissolved in 100 µl of buffer, were used to estimate elastase or inhibitor. The plates were incubated at 37° C for 10 min. The reaction was started by adding 50 µl of substrate from stock solutions at 5 mg/ml dimethylsulfoxide diluted 1:5 in buffer. Optical density was read immediately at 405 nm and after 5 min. All assays were linear over at least 7 min; the lower limit for detection of MPI was 50 pM. The observed changes in optical density in relation to the control incubation were used to calculate elastolytic or inhibitory activity. An increase of target enzyme activity indicated the presence of additional free elastase, whereas a decrease of target enzyme activity indicated inhibitory activity.

The contribution of ₁-PI to the total inhibitory activity was determined by the difference of inhibitory activity between the native sample and the sample after removal of ₁-PI. This was quantitatively accomplished by adsorption of ₁-PI to membranes (Immobilin; Milipore, Bedford, UK) that had previously been coated with a goat polyclonal antibody directed against human ₁-PI (Sigma, St. Louis, MO). Before use, the membranes were washed, blocked for 30 min with 3% gelatin, and washed again. After 12 h of incubation with the membranes, the sample was recovered by centrifugation (1,000 g for 10 min) and elastolytic or inhibitory activity was determined as described. The removal of ₁-PI was completely achieved (less than 2.1 ng/L remained), as demonstrated by Western blotting with ₁-PI from human serum.

The contribution of the acid-resistant inhibitors to the total inhibitory activity was determined after all the acid labile proteases and ₁-PI were inactivated by treating aliquots of the samples with acetic acid (30% by volume, 10 µl/ml supernatant) and heat (90° C, 20 min). This treatment also released the acid-resistant inhibitors that were complexed to proteases. Coagulated proteins were pelleted by centrifugation, and the supernatants were tested for elastase-inhibiting activity relative to acid-resistant inhibitors.

Enzyme-linked Immunosorbent Assays

Acid-resistant protease inhibitors. After adsorption of our polyclonal antibody against MPI and ESI to a 96-well protein-binding microtiter plate, the plate was washed, blocked with bovine serum albumin, and incubated with the various samples and the MPI standards. Bound antigen was detected by biotinylated anti-MPI IgG, followed by avidin-linked alkaline phosphatase. The lower limit of detection was 0.01 nM; the assay was linear up to 0.5 nM.

₁-Protease inhibitor. After adsorption of a goat polyclonal antibody against human ₁-PI (IBL, Hamburg, Germany) to a 96-well protein-binding microtiter plate, the plate was washed and incubated with the various samples and the ₁-PI standards for 60 min at 37° C. After washing, bound antigen was detected by biotinylated anti-₁-PI, followed by avidin-linked alkaline phosphatase. The lower limit of detection was 2 nM; the assay was linear up to 80 nM.

Secretory immunoglobulin A. A mouse monoclonal antibody directed against the secretory component of human secretory IgA (sIgA) (BioMarkor, Rehovot, Israel) was coated to the microtiter plate. The plate was washed, and samples and standard were added. After 60 min of incubation at 30° C, a second peroxidase-coupled antibody directed against the secretory component (Dako, Glostrup, Denmark) was added. After 60 min at 30° C, the plate was washed and peroxidase activity was determined from color development in the presence of H₂O₂ and phenylendiamine. The lower limit of detection was 0.14 nM, and the assay was linear between up to 9 nM. In all lavage samples, sIgA could be detected, no dependency on gestational age, birth weight, or postnatal age was observed. So, in agreement with others (11, 12), we used sIgA as a reference protein to correct unpredictable dilution of the epithelial lining fluid from the lavage procedure.

Gel Electrophoresis and Western Blotting

Samples were boiled in an equal volume of Laemli buffer, 50 µg of protein per lane was then loaded, and samples were separated by sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Five percent SDS was used in stacking gels and 15% SDS in separating gels (Midget electrophoresis unit; LKB, Hamburg, Germany). Following transfer of the proteins to nitrocellulose membranes (Milipore, Bedford, UK), the membranes were blocked for 30 min with 5% gelatin, and immunoblotting was performed with the polyclonal antibody raised against MPI for 12 h at room temperature. After washing, secondary antibody (peroxidase conjugated anti-human IgG; Dianova, Hamburg, Germany) was applied at 1:1,000 dilution in 1% gelatin for 2 h. Blots were washed three times in buffer containing 0.025% Tween 20, and peroxidase was made visible by the reaction with 4-chloronaphthol and H₂O₂.

Data Analysis

Data were expressed as mean ± standard error of the mean. As no significant deviations from normal distribution could be demonstrated by the Lilliefors modification of the Kolmogoroff-Smirnoff test, it was assumed that the data had a normal distribution pattern (13). Comparisons between two corresponding parameters were made with the two-tailed t test. The level of significance was set at p < 0.05. After calculation of Spearman rank correlation coefficients, a multiple correlation analysis was performed. Statistical evaluation was performed with SPSS (SPSS Inc., Chicago, IL). The study was approved by the authorized Human Research Committee of our University, and informed consent from the parents or guardian was obtained.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In an initial study, the supernatant of the airway lavage samples from healthy children without lung disease was investigated. No free elastase was detected in any sample, and an excess of free antiprotease activity was seen to be distributed equally among ₁-PI (0.11 ± 0.02, range 0-0.69 mol/mol sIgA; 45% of the inhibitory activity) and acid-resistant inhibitors (0.13 ± 0.03 mol/mol sIgA, range 0-1.03; 55% of the inhibitory activity). Most of the acid-resistant inhibitors were freely active (0.10 ± 0.02 mol/mol sIgA); only a small fraction was complexed to proteases, but could be released in the active state by heat and acid treatment (0.03 ± 0.01 mol/mol sIgA). The presence of the acid-resistant antiproteases was also demonstrated in all these samples by SDS-PAGE and immunoblotting with the polyclonal antibody against MPI and ESI (not shown).

Samples from neonates were then analyzed. To characterize the total antiproteolytic activity, the inhibition of a series of different proteases was assessed in a subgroup of premature infants. Native lavage supernatant contained acid labile antiprotease activity, which accounted for at least one-third of the activity (Table 1). Apart from the inhibition of human neutrophil elastase, acid-resistant activity was demonstrated against pancreatic elastase, which is specific for ESI, and against trypsin, which is specific for MPI (Table 1).

By SDS-PAGE and immunoblotting, the presence of acid-resistant protease inhibitors of 7-14 kD was also demonstrated in the cellular fraction of the airway lavage samples at all age groups investigated (24-25, 26-30, 31-35, and 36-40 wk gestation) (Figure 1). The band at about 30 kD also represented MPI associated with other higher molecular weight acid-resistant proteins, as demonstrated after exclusion of cross-reactivity with bikunin and after exclusion of self-aggregation of MPI by repeated addition of reducing agent. Based on these results, airway lavage samples were quantitatively analyzed for both the soluble fraction and the cellular fraction.

View larger version (107K): [in this window] [in a new window]   Figure 1.   SDS-PAGE and Western blotting of fractions from airway lavage samples from 42 mechanically ventilated neonates. Lavages were fractionated into cellular and soluble fractions by centrifugation, and the cellular fraction was thoroughly washed. Pooled samples (50 µg protein) from each gestational age group ( 25 wk, n = 1;  30 wk, n = 17;  35 wk, n = 14;  40 wk, n = 10) were applied. Purified, recombinant MPI was used as standard. Note additional bands at 30 kD; these were due to association of MPI with other acid-resistant proteins, as cross-reactivity of the antibody with bikunin and self-association of the inhibitors were excluded.

To characterize postnatal protease/antiprotease activity, airway lavage samples obtained from 42 neonates during the first 36 h of life were analyzed. No dependency of any of the parameters measured (including sIgA) on gestational age or birth weight was noted. When the 42 neonates were grouped according to the cause of the respiratory failure (respiratory distress syndrome, suspected perinatal infection, transient tachypnea, patent ductus arteriosus, or asphyxia), no differences among any of the parameters were noted among these subgroups of neonates. In addition, multiple correlation analysis indicated no correlations for any of the different subgroups and the parameters measured, except for a weak correlation between the development of bronchopulmonary dysplasia and the ratio of acid-resistant inhibitors and ₁-PI in the soluble fractions (r = 0.56, p < 0.0001, n = 42). Thus, data for the whole group of infants are reported (Table 2). Elastase activity was limited entirely to the cellular fraction. Although about 30% of the acid-resistant inhibitors was localized in the cellular fraction, almost no ₁-PI was found there. About 50% of the functionally active acid-resistant inhibitors was free, the other half being complexed to proteases and thus probably not available for inhibitory activity (Table 2). The difference between the ELISA tests and the functional assays indicates the amount of inactivated inhibitors, which was much greater for ₁-PI than for the acid-resistant inhibitors.

Nine of the neonates were mechanically ventilated for 10 d or more after recovery from postnatal respiratory distress syndrome. One preterm neonate had necrotizing enterocolitis and another, mature neonate was ventilated for severe lung injury following asphyxia. The other seven neonates were preterm infants and had a characteristic course. They developed the clinical signs of an infection on Day 4 to 6 of life, making increased respiratory support necessary (Figure 2, upper panel). Staphylococcus epidermidis was recovered at this time from the blood of all these neonates. Other possible causes of respiratory decompensation, such as fluid overload or persistent ductus arteriosus, were ruled out. No significant changes in elastase activity in either the cellular and the soluble fraction were noted during this period (Figure 2, lower panel). The acid-resistant inhibitors were present at all times in both of the fractions investigated (Figure 3). However, during respiratory deterioration and mechanical ventilation from Day 8 onward, decreased amounts of functionally active acid-resistant inhibitors were present in the soluble fraction (Figure 3, lower panel). This was due to an almost complete loss of free, functionally active, acid-resistant inhibitors in the presence of complexed and inactive inhibitors (Figure 4). As no group of ventilated, extremely premature neonates without respiratory deterioration and infection is available for comparison, no causal relationship can be inferred from the coincidence of changes in the inhibitors and the clinical condition. ₁-Protease inhibitor was almost completely localized in the soluble fraction during the time period investigated. In contrast to the acid-resistant inhibitors, at all times a ten- to twentyfold excess of inactivated ₁-PI was present (data not shown).

View larger version (18K): [in this window] [in a new window]   Figure 2.   Oxygenation index (upper panel ). Elastase activity (lower panel ). Seven of the 42 neonates had to be ventilated mechanically for a prolonged time after birth during nosocomial infection. Airway lavage samples were fractionated into cellular and soluble fractions by centrifugation, and the cellular fraction was thoroughly washed. Data are means ± SEM of 6-7 neonates on Days 1-11, and 4-5 neonates on Days 12-15.

View larger version (26K): [in this window] [in a new window]   Figure 3.   Total amount of acid-resistant inhibitors determined in a functional activity assay and in an ELISA system on aliquots of airway lavage samples from the lungs of seven mechanically ventilated neonates with clinical signs of septic infection. Lavages were fractionated into a cellular fraction (upper panel ) and a soluble fraction (lower panel ) by centrifugation. The cellular fraction was thoroughly washed before analysis. Data are means ± SEM of 6-7 neonates on Days 1-11, and 4-5 neonates on Days 12-15. Asterisks indicate significant differences in the soluble fraction between acid-resistant antiproteases determined by ELISA and acid-resistant antiproteases estimated by functional assay. Data pairs were analyzed by paired t tests (*p < 0.05 and **p < 0.01).

View larger version (24K): [in this window] [in a new window]   Figure 4.   Activity of free (upper panel ) and complexed (lower panel ) acid-resistant inhibitors as determined in a functional assay system on aliquots of airway lavage samples from the lungs of seven mechanically ventilated neonates with a septic infection. Lavages were fractionated into cellular and soluble fractions by centrifugation, and the cellular fraction was thoroughly washed. Data are means ± SEM of 6-7 neonates on Days 1-11, and 4-5 neonates on Days 12-15.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The main goal of our study was to define elastase and its inhibitors in both the cellular and the soluble lavage fractions in ventilated neonates with a variety of clinical conditions. For ethical reasons, we were unable to obtain airway lavage samples from healthy extremely premature infants and mature neonates during the first 36 h of life. Therefore, we can only describe these data from neonates with acute respiratory failure or nosocomial infection and assess potential correlations to various clinical parameters. The older infants, who had healthy lungs, were not an ideal comparison group, but were also reported for reference.

In the group of neonates with respiratory failure during the first 36 h of life, no elastase activity was detected in the soluble compartment, indicating no gross derangements of the protease inhibitor systems. This was true for infants of all age groups, even for those 14 of 42 infants likely to have a perinatal infection. No differences among the neonates from the various subgroups and no correlations with specific diagnoses or outcomes were observed.

When compared to older, healthy children, neonates had one-fifth to one-tenth the total activity of acid-resistant inhibitors. This might be related to the respiratory failure of the neonates, but age-related differences cannot be excluded. The latter appears less likely, as by the age of 20 wk the distribution of MPI in bronchial glands of human tissue was similar to that in adult tissue (14) and as we did not observe in our group of infants any dependency on gestational age from 24-40 wk.

Due to lack of material, we were not able to routinely differentiate between the acid-resistant inhibitors MPI and ESI. However, our pilot studies on a limited number of samples and Western blot analysis of samples from the first 36 h of life (Table 1, Figure 1) had demonstrated their presence. Although this study clearly demonstrated that about 30% of the acid-resistant inhibitors are cell-associated and that a full assessment of the protease/antiprotease balance in the lavage must include the cellular fraction, the exact origin of the inhibitors has yet to be explored. Besides synthesis by mucosal cells in the airways (5), our previous study in a comparable group of preterm neonates suggested the synthesis of ESI, but not of MPI, and the association of MPI with a substantial fraction of the neutrophils and macrophages of the cellular lavage fraction (6). We did not find expression of MPI in pilot experiments on unstimulated human neutrophils from peripheral blood or macrophages from healthy adult volunteers (6). Recently, it was demonstrated that neutrophils and peritoneal macrophages from mice synthesize MPI (or SLPI) after lipopolysaccharide (LPS) stimulation (15). Lipopolysaccharide has previously been shown to stimulate secretion of SLPI, detected by means of a polyclonal antibody and thus likely to react with both MPI and ESI from human neutrophils (16).

From a functional point of view, besides its activity as a protease inhibitor, MPI has also been suggested to play a role as a potential antiviral defense compound (17) or as an antagonist to bacterial LPS (15). Additional studies will be necessary for detailed quantitative analysis and assessment of the functional role(s) of these acid-resistant inhibitors in the lungs.

In contrast to the acid-resistant inhibitors, in the same samples ₁-PI was almost completely confined to the soluble fraction. As ₁-PI has previously been shown to be also synthesized by alveolar macrophages and neutrophils (5), some ₁-PI might have been expected in this fraction. Its lack indicated that the cells specifically synthesizing the protein were present only at a low level, or that the rate of synthesis was so low that it did not significantly contribute to total ₁-PI, or possibly that the ₁-PI synthesized was very rapidly secreted.

Interestingly, a major fraction (about 90%) of ₁-PI was functionally inactive, as shown by the difference between the amount determined by ELISA and that assessed from the functional assays. In addition, by crossed immunoelectrophoresis we demonstrated that a substantial amount of ₁-PI was bound to elastase (data not shown). These results are in agreement with previous studies by Merritt and colleagues, who found between 51 and 87% of ₁-PI to be inactivated during respiratory distress syndrome and developing bronchopulmonary dysplasia (1). ₁-Protease inhibitor as determined by ELISA may include material that has previously been complexed to elastase, oxidized, or partially degraded by proteolysis and that has been removed from the airspaces to variable extent. All these factors might explain the lack of correlation to functionally active ₁-PI.

In the group of neonates who were observed during deterioration with nosocomial infection and mechanical ventilation, a small, non-significant increase of elastase activity was noted in the soluble fraction from Day 7 to 10. The cellular fraction contained the majority of the elastase. This was likely to be associated mainly with the neutrophils. Due to relatively large scatter of the data, no significant changes with time were observed in this fraction.

Functionally active, acid-resistant protease inhibitors were very low from Day 6 to 15 in the soluble fraction. No changes with time were observed in the cellular fraction, and a substantial portion of the total inhibitor amount (as determined by ELISA), was in a functionally inactive state. Watterberg and coworkers (9) found an increased amount of the acid- resistant inhibitor MPI (called SLPI in their paper) after the second week of life in neonates developing bronchopulmonary dysplasia. The polyclonal antibody that these investigators used in their ELISA was likely to detect both MPI and ESI. Unfortunately, no functional assessments of the inhibitory activity were made.

In conclusion, our data demonstrate the presence of acid-resistant protease inhibitors in the cellular fraction of airway lavage samples from neonates and premature infants. In contrast, ₁-PI was almost exclusively limited to the soluble lavage compartment. A complete functional and biochemical analysis of the protease-protease inhibitor balance should include the cellular and soluble lavage compartments.