Main

RDS remains a leading cause of morbidity and mortality in newborns despite the use of innovative approaches such as antenatal corticosteroids and TSH-releasing hormone, and exogenous surfactant administration to neonates(1–3). RDS is caused by a deficiency of pulmonary surfactant(4) and is characterized histologically by the presence of fibrinous exudates and fibrin-rich hyaline membranes lining the alveoli. Excessive alveolar deposition of fibrin and its degradation products may interfere with surfactant function, injure vascular endothelium, serve as chemotactic factors for neutrophils, and function as immunoregulatory peptides(5–9). Fibrin deposition and inadequate breakdown may also accelerate the inflammatory cascade, which may be important in the pathogenesis of BPD(10). Fibrin breakdown or fibrinolysis is catalyzed by plasmin, which is a broad-spectrum proteinase. Plasmin is activated from its zymogen form plasminogen by two specific endopeptidases: tPA or uPA(11). Once plasmin is generated, it can also act directly on plasminogen to generate additional plasmin.

PA and their inhibitors, which influence the extent of fibrinolysis, have been demonstrated in both normal and injured lungs of laboratory animals and humans(12–15). Urokinase and PAI are also produced by alveolar epithelial cells and by pulmonary capillary endothelial cells(6, 16–18). We have previously shown that PAI type-1 (PAI-1) is expressed by type II pneumocytesin vitro(19). Alveolar macrophages also release PA and PAI and may also influence fibrin degradation(20–22). Adult RDS is also characterized by alveolar fibrin deposition and has been associated with depressed activity of alveolar uPA(14, 15).

We hypothesized that preterm neonates with RDS would have lower alveolar levels of fibrinolytic activity. In addition, those neonates with RDS who progress to BPD would have an even greater impairment in alveolar fibrinolysis. Because plasmin generation leading to fibrinolysis is regulated by the balance between PA and their inhibitors, we postulated that impaired fibrinolysis would be due to either a depletion of PA, an excess of PAI, or a combination of these two. We investigated this by measuring PA activity as well as immunoreactivity of the PA and PAI found in TA from neonates with RDS.

TA sampling has a number of drawbacks. First, the volume of ELF recovered varies with each sampling. Second, the central airways are sampled along with the alveoli. Third, alveolar proteins cannot be separated from plasma proteins, which are particularly prominent during the enhanced capillary leakage of inflammation(31). Little can be done to influence the latter two drawbacks. Both serum and TA urea were measured to correct for the dilution of ELF by saline in the TA, to minimize the first of these drawbacks(24, 25).

METHODS

Materials. Plasminogen, uPA, tPA, and ELISA kits for uPA, tPA, PAI-1, and PAI-2 were obtained from American Diagnostica (Greenwich, CT). Protein assay reagent was obtained from Bio-Rad Laboratories (Hercules, CA). Chromogenic substrate (D-Val-Leu-Lys-pNA), nitrocellulose membranes, fibrinogen, urea assay reagents, and all other biochemicals of the highest grade available were obtained from Sigma Chemical Co. (St. Louis, MO).

Study population. Forty-six infants intubated at birth for RDS(clinical and radiographic criteria) and nine intubated for nonpulmonary causes were enrolled in the study, after obtaining informed parental consent. Consent was refused in one case. Infants were enrolled from the Neonatal Intensive Care Unit at University Hospital, Stony Brook, NY. The study protocol was approved by the institutional review board of the hospital. Patients with the following conditions at birth were not included:1) birth weight not appropriate for gestational age, 2) documented infection, 3) congenital lung or heart anomalies; and4) pulmonary disease complicated by meconium aspiration, pulmonary hemorrhage, or pneumothorax. BPD was defined as requiring supplemental oxygen at 28 d of life to maintain Pao2 ≥ 50 mm Hg and radiographic findings of BPD according to Northway's classification(23).

Radiographic staging. Chest radiographs were obtained for routine clinical indications. All radiographs were reviewed by a pediatric radiologist, unaware of the patient's clinical diagnosis. Each radiograph was staged according to Northway's criteria for BPD(23).

TA. Secretions from TA were collected from intubated infants on d 1, 7, 14, 21, and 28 or until extubation. TA collections were discontinued if the patient developed complications such as pneumothorax, pulmonary hemorrhage, or the need for placement on high frequency jet ventilation. Aspirates were obtained during routine suctioning. The infant was positioned in a supine posture with head midline. One milliliter of saline was instilled, two to four breaths were delivered, and the airway was suctioned one or two times with a suction catheter at pressures ≤80 cm H2O. The catheter was rinsed with 1 mL of saline, and the aspirate was collected in a Leuken's trap. The infant's heart rate, respiratory rate, and oxygen saturation were monitored and allowed to stabilize during the suctioning procedure. TA samples were kept on ice and processed within 1-2 h. Total cell count was performed using a hemocytometer under phase contrast microscopy. Specimens were then centrifuged at 1500 × g for 10 min at 4°C, and the cell-free supernatants were stored at -70°C until further analysis.

Urea assay. Quantitative assay for urea in the TA was performed using the commercially available kit (66-UV; Sigma Chemical Co.) with slight modifications(24, 25). The amount of urea present is directly proportional to the decrease of NADH as measured by the change of extinction at 340 nm. The ratio of the sample to reagent was adjusted to obtain linearity in the range of 0.06-0.90 mmol/L. Interassay and intraassay variability was <10%. TA samples were analyzed for urea in duplicate. Serum urea was measured using the same assay on a blood sample taken for routine clinical management within 6 h of lavage. ELF volume was calculated from theformula: The volume of fluid recovered from each TA sample varied from 0.6 to 1.0 mL. Concentrations of the measured constituents of TA were normalized to the ELF to avoid errors from dilution during the sampling procedure.

Amidolytic assay for net PA/plasmin activity. The assay is based on the cleavage of plasminogen by PA to form plasmin, which then converts the chromogenic substrate (D-Val-Leu-Lys-pNA) to a colored end product(26). The assay was performed in triplicate for each TA sample. This assay is positively modulated by PA and plasmin, whereas it is inhibited in the presence of anti-plasmins or PAI.

Immunoblotting. Aliquots of TA were denatured by heating to 80°C for 1 min in the presence of 4% β-mercaptoethanol(27). Samples were separated on SDS-polyacrylamide minigels with 4% stacking and 8% separating gels(28). Proteins were then transferred electrophoretically onto a nitrocellulose membrane. uPA and tPA were identified using alkaline phosphatase-conjugated secondary antibodies after incubation with primary antibodies. Color development was achieved with nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate (p-toluidine salt)(19).

RFA. RFA was used to detect PAI, as well as PA and PA-PAI complexes, according to the method previously described(19). Basically, proteins were separated by SDS-PAGE, the gels were cleared of SDS by washing with Triton X-100, and these gels were overlaid onto a cloudy matrix containing fibrin. Proteins in the gel which accelerated fibrinolysis (PA and PA-PAI complexes) caused visible lysis at their apparent molecular weights, whereas proteins capable of retarding fibrinolysis resulted in persistent opaque zones of spared lysis. Molecular weights were interpolated from concurrently run markers (not shown).

Quantitative assay for uPA, tPA, PAI-1, and PAI-2. ELISA for uPA, tPA, PAI-1, and PAI-2 were performed in duplicate, according to the instructions of the manufacturer (American Diagnostica, Greenwich, CT). The uPA immunoassay measures both single chain uPA and high molecular weight uPA, in free form and in complex with PAI-1 and PAI-2. The tPA ELISA measures tPA in complex with α2-antiplasmin or PAI. There is no cross-reactivity between the ELISA for urokinase and for tPA. The assay range for tPA ELISA is 0.015-0.3 ng/well with a maximal detection limit of 1.5 ng/mL. PAI-1 ELISA assay has a detection limit of 1 ng/mL and PAI-2, of 6 ng/mL.

Data analysis. For comparison of clinical characteristics, theχ2 with Yates' correction was applied for the nominal data and the one-way analysis of variance with Bonferroni posttest was applied for continuous data. Results are expressed as mean ± SD. Intergroup comparisons for the TA assays were done using the two-tailed, unpaired Mann-Whitney U test. A nonparametric test was used because the data were not normally distributed. Results were expressed as the median value with the 25th to 75th quartile values in parentheses. A p value of<0.05 was considered to be statistically significant.

RESULTS

Characteristics of the study population. Forty-six premature infants with RDS were enrolled at birth, of which 26 had self-resolved RDS and 18 progressed to develop BPD. Two infants died at the ages of 3 and 8 d. Autopsy consent was obtained in one case and revealed disseminated intravascular coagulation with multiple organ hemorrhage. These two infants were excluded from the final analysis because group assignment could not be determined. Two other infants died in the BPD group at the ages of 28 and 69 d from respiratory failure. Autopsy was done on one infant, and it showed fungal sepsis. Another nine infants intubated for nonpulmonary causes (one each with gastroschisis, imperforate anus, biliary atresia, apnea, choanal atresia, hydrocephalus, PDA ligation, diaphragmatic hernia, and esophageal atresia) were enrolled as controls.

The demographic characteristics of the infants are presented inTable 1. There were no statistically significant differences in maternal factors between the three groups except for antenatal dexamethasone therapy. Of infants developing BPD, 12 (67%) had been exposed to antenatal steroid therapy, whereas only 11 (42%) presented with self-limited RDS, and 1 (11%) was in the control group. As would be expected, the infants developing BPD had a lower birth weight and a younger gestational age than those with self-limited RDS. Infants in the control group were of significantly higher birth weight and gestational age than either of the other two groups. It would have been desirable to have infants in the control group with similar birth weights and gestational ages as in the self-limited RDS and RDS + BPD groups, but it is unusual to have intubated premature infants without RDS, except for those with apnea. Infants developing BPD had lower Apgar scores and required additional resuscitation at birth compared with those with self-limited RDS or no pulmonary disease (control). Groups were similar for sex distribution and for inborn or outborn births.

Table 1 Demographic characteristics
Full size table

Table 2 shows the clinical characteristics of the infants. All patients with RDS required similar mean airway pressures on d 1, regardless of the outcome. Groups were also similar for the incidence of patent ductus arteriosus requiring treatment, pulmonary interstitial emphysema, and intraventricular hemorrhage ≥ grade II. Infants developing BPD had a higher incidence of culture-proven sepsis and necrotizing enterocolitis than those without BPD. More infants developing BPD also developed retinopathy of prematurity ≥ stage II. A sole patient in the self-resolved RDS group received dexamethasone therapy starting at age 21 d and was weaned to room air by age 25 d. Sixty-one percent of infants developing BPD were treated with dexamethasone, which was started on or after 21 d of age. The samples obtained on d 28 may therefore be affected by this treatment. As expected, duration of positive pressure ventilatory support, exposure to supplemental oxygen, and duration of hospitalization were longer in the BPD group than in the other two groups.

Table 2 Clinical characteristics
Full size table

Detection of PA and PAI in TA. TA samples were screened for the presence of PA, PAI, and PA-PAI complexes by RFA In serial photographs of the same RFA (Fig. 1), clear bands of lysis, indicating PA activity, were visible by 18 h (Fig. 1A) and were more pronounced by 24 h (Fig. 1B). After additional incubation(33 h, Fig. 1C), persistence of a zone of inhibition (PAI activity) was documented only in the PAI-1 standard lane, with the remainder of the gel completely cleared. PAI present in the TA samples was detectable only in complex formation with PA, but not as free PAI. Confirmation of the PA activity seen by RFA was demonstrated with immunoblot analysis. PA was detected as both high molecular weight uPA (HMW, Mr= 53 000, Fig. 2) and tPA (Mr = 65 000,Fig. 3). Low molecular weight uPA was not seen in TA by either RFA or immunoblot.

Figure 1
figure 1

RFA of TA samples after 18 h (A), 24 h(B), and 33 h (C) of incubation. Volumes of 10, 20, 30, and 40 ÎĽL (lanes 2-5) from the same TA; and known standards of PAI (lane 1, obtained from human plasma) and uPA (lane 6) were separated by SDS-PAGE and overlaid onto an indicator gel consisting of uPA, plasminogen, human fibrinogen, and agarose. Zones of lysis appear dark and indicate PA activity (which is also seen in the presence of PA-PAI complexes); whereas inhibition of lysis zones remain opaque (white) and localize PAI activity. Apparent molecular weights (Mr) are indicated on the left margin in thousands. PA activity is demonstrated in: standard lane 1 (PA-PAI complex Mr = 120 000); sample lanes 2-5 (high molecular weight (HMW) uPAMr = 53 000, tPA Mr = 65 000, and PA-PAI complex Mr = 120 000); and standard lane 6 (HMW uPAMr = 53 000 and low molecular weight (LMW) uPAMr = 35 000). Free PAI activity was seen only in lane 1 (Mr = 50 000).

Full size image
Figure 2
figure 2

Immunoblot of TA samples for uPA. Samples were separated by SDS-PAGE, electroblotted onto nitrocellulose membranes, and tested with specific rabbit anti-human uPA antiserum followed by alkaline phosphatase-conjugated goat anti-rabbit antibody. Lanes 1-4, TA sample aliquots from four patients, lanes 5 and 6, known standards of high molecular weight (HMW) uPA and low molecular weight (LMW) uPA.

Full size image
Figure 3
figure 3

Immunoblot of TA samples for tPA. Samples were separated by SDS-PAGE, electroblotted onto nitrocellulose membranes, and tested with specific goat anti-human tPA antiserum followed by alkaline phosphatase-conjugated rabbit anti-goat antibody. Lanes 1-4, TA sample aliquots from four patients; lane 5, known standard of tPA.

Full size image

Assay for fibrinolytic components in TA. Quantitative ELISA were performed for uPA, tPA, PAI-1, and PAI-2, in TA from d 1 to 28(Table 3). No significant differences were noted between patients with self-resolved RDS and those who progressed to BPD for any of the assays. To obtain an estimate of normal values in neonates, TA were collected on d 1 of intubation from nine control infants, receiving mechanical ventilation for nonpulmonary conditions. As noted previously, these babies weighed more, were more mature, and received higher Apgar scores than either of the other two groups. ELISA and PA/plasmin activity assay results on d 1 in control patients are compared with patients with self-resolved RDS and those who progressed to BPD in Table 4. Net PA/plasmin activity was significantly depressed in patients with self-resolved RDS (median = 20.9 ng/mL, p < 0.05) and RDS + BPD (median = 4.97 ng/mL, p< 0.001) compared with control patients (median = 87.1 ng/mL)(Fig. 4). PAI-1 was also elevated in self-resolved RDS patients (median = 111.1 ng/mL) compared with control patients (median = 10.9 ng/mL, p < 0.05). Net PA/plasmin activity was found to be significantly depressed in infants developing BPD (median = 4.97 ng/mL) compared with those with self-resolved RDS (median = 20.9 ng/mL, p< 0.001) on d 1 (Table 5). No significant differences were noted from d 7 to 28 (Fig. 4). Net PA/plasmin activity was undetectable on d 1 in two infants. Both of the infants developed BPD, and one died at 69 d of age.

Table 3 Comparison of PA and PAI antigen concentrations in patients with self-limited RDS and those who progressed to BPD
Full size table
Table 4 Fibrinolytic activities in tracheal aspirates on d 1 of intubation (ng/ml)
Full size table
Figure 4
figure 4

PA/plasmin activity in TA samples over the first month of life. PA/plasmin activity for the group with self-resolved RDS (median values, closed circles) is compared with the PA/plasmin activity for the group with RDS who progressed to BPD (median values, open circles). TA samples obtained from intubated neonates without RDS (control) on d 1 are shown as a hatched bar. A significant difference (#) between the neonates without RDS and both the self-resolved RDS group(p < 0.05) and the RDS group who progressed to BPD (p< 0.001) was noted. A significant difference (*) between the self-resolved RDS group and the RDS group that progressed to BPD was noted on d 1(p < 0.001).

Full size image
Table 5 Comparison of net PA activity in patients with self-limited RDS and those progressed to BPD
Full size table

Supernatants from centrifuged samples were stored at -70°C awaiting batch assay. To test whether TA samples might be affected by recent surfactant administration, ELISA and PA/plasmin activity assays were carried out in the presence of varying concentrations of surfactant (Survanta, Ross Laboratories, Columbus, OH). No differences were noted in the results of ELISA or PA/plasmin activity assays (results not shown).

Because the control group infants had higher gestational ages than both other groups, net PA/plasmin activity on d 1 was plotted against the gestational age for all patients in the study (Fig. 5). If maturation positively influences net PA/plasmin activity expression after acute lung injury, one might expect slopes within each group to reflect this positive trend. Regression lines for each group were compared and were not found to be significantly different from a slope of zero. This suggests that differences between the groups were not due to maturation.

Figure 5
figure 5

PA/plasmin activity in TA samples by gestational age on d 1. Individual values of PA/plasmin activity from TA samples obtained on d 1 is plotted by the gestational age of the patient. Regression lines have been drawn for the control, self-resolved RDS, and RDS progressing to BPD groups. Slopes of the three regression lines were not significantly different from a slope of zero.

Full size image

DISCUSSION

Avery and Mead(29) proposed that surfactant deficiency in the premature newborn initiated a series of changes at the level of the alveolar-capillary unit which resulted clinically in RDS. Despite postnatal administration of exogenous surfactant, the incidence of BPD has been largely unaffected(30).

The treatment of preterm infants with oxygen and mechanical ventilation is now known to initiate a cascade of inflammation which cannot be mitigated by the adequate replacement of surfactant. The factors that initiate acute inflammation and lung injury may then destabilize the alveolar-capillary unit leading to the development of BPD. These factors are not completely understood. In adults who develop RDS, analysis of bronchoalveolar lavage fluid has shown depressed fibrinolytic activity(14, 15). This depressed activity was attributed to decreased uPA activity thought to be secondary to increased anti-plasmin or PAI activity. Incomplete fibrinolysis may result not only in surfactant inactivation, but may also initiate acute lung injury directly (acting as a nidus for fibroblast ingrowth) or indirectly (i.e. chemotactic or immunologic mechanisms). Thus, these early changes in the fibrinolytic pathway may be part of the “trigger” events which contribute to the acute lung injury response seen in premature neonates.

Significant lung injury in adults can be investigated with bronchoscopy and bronchoalveolar lavage analysis. In the premature neonate, in vivo collections of TA can be made. Quantification of recovered ELF in the TA specimen is crucial to accurately evaluate the cellular and molecular components of the lining fluid. The use of albumin or protein in TA to normalize for dilution has been shown to have many disadvantages, because plasma proteins may leak from the capillaries into the alveoli and result in a disporportionately larger concentration in the ELF compared with the plasma, particularly under inflammatory conditions(31). A secretory component of IgA has also been used to normalize for ELF dilution, but a standardized assay is not described(35). Urea is a relatively small molecular weight compound that is freely diffusible throughout the compartments of the body and which has been used to normalize the constituents of bronchoalveolar lavage as well as TA(24, 25).

In the present study we focused on the regulators of fibrinolysis by quantifying PA and their inhibitors, as well as by investigating net PA/plasmin activity, in neonates with RDS. At the earliest time point (d 1), net PA/plasmin activity was decreased in infants with RDS compared with controls. In addition, there was a significant difference seen between those who had self-limited RDS compared with those who progressed to BPD. Infants developing BPD had the most severe inhibition of net PA/plasmin activity. This inhibition cannot be explained by the excess of either PAI-1 or PAI-2, and may indicate the presence of other inhibitors such as anti-plasmins. Whether this depressed fibrinolytic activity is intrinsic to neonates predisposed to BPD, or whether it is reflective of a pattern of developmental immaturity in this smaller and less mature group of infants, cannot be conclusively demonstrated in this study. Although it would have been desirable to compare these groups using gestational age-, birth weight-, and sex-matched patients, the relatively small number of patients made this unfeasible. Additional TA collections from neonates 28 wk or less with RDS will be needed to obtain a sufficient sample size for these comparisons.

Depressed PA/plasmin activity has been previously demonstrated in preterm infants with RDS when compared with controls(22). The presence of significant concentrations of PAI, predominantly PAI-2, was noted, along with a trend of lower PA activities in BPD infants compared with the RDS group. We found a similar trend in depressed PA/plasmin activities when comparing the two groups, a trend which reached significance on d 1. We did not find any differences between the PAI antigen concentrations for the self-limited RDS group as compared with the RDS progressing to BPD group. However, the absolute magnitude of PAI-2 values were increased in both groups, compared with PAI-1 values, which diminished over the study period. We would reiterate their finding of significant variation within groups, which occurs in most studies examining TA in a diverse group of infants.

Acute lung injury may be initiated shortly after the exposure to oxygen and mechanical ventilation in preterm infants. The major determinants responsible for the progression from acute lung injury to BPD may already be present shortly after this exposure. Merritt et al.(33) found that infants with RDS who subsequently developed BPD had greater numbers of neutrophils, elastase activity, and albumin concentration in their TA by d 2-3 of life compared with infants who did not develop BPD. Bagchi et al.(34) have shown that, in infants with RDS, those who develop BPD have significantly higher levels of IL-6 in their TA on d 1 of life compared with infants who do not. Evidence of depressed PA/plasmin activity on d 1 may be helpful in identifying patients at high risk for developing BPD. Day 1 PA/plasmin activity (normalized with urea) was ≤10.0 ng/mL in 86% (12/14) of infants developing BPD compared with 24% (5/21) in infants with self-resolved RDS. In infants of ≤30-wk gestation at birth with RDS, a d 1 PA/plasmin activity≤10.0 ng/mL had a positive predictive value of 80% (12/15) and a negative predictive value of 82% (9/11) that there will be progression to BPD.

In summary, this study focused on the mediators of alveolar fibrinolysis isolated from TA of premature neonates with RDS. A depressed net PA/plasmin activity was demonstrated in neonates on d 1 of life with RDS that progressed to BPD, compared with neonates with self-limited RDS. This decreased PA/plasmin activity could not be attributed to decreases of PA or to corresponding increases in PAI. This implies that decreased plasmin or increased anti-plasmin or other inhibitory components may have been responsible for this depression of net PA/plasmin activity. The lack of strict matching for gestational age and birth weight limits the ability to determine whether this depressed fibrinolytic activity is a result of a developmentally immature response to lung injury or an intrinsic depression of this activity in the group that progresses to BPD. Additional TA collections from neonates 28 wk or less with RDS, particularly on d 1, will be needed to clarify this issue. The finding that the degree of depression of net PA/plasmin activity in TA from neonates with RDS on day one may predict whether the RDS will be self-resolved, or will progress to BPD warrants further investigation to identify the particular component(s) of suppression in this PA/plasmin pathway responsible for initiating the cascade of events of acute lung injury in the premature neonate leading to BPD. This may also be critical in the development of interventional strategies to prevent acute and chronic lung disease in infants.