INTRODUCTION

Top Abstract Introduction References

LUNG SURFACTANT is a phospholipid-rich material that lines the mammalian lung and functions to promote alveolar stability by reducing the surface tension at the air-alveolar interface. Surfactant is synthesized in the large type II alveolar cell, and stored within multilamellated, membrane-bound vesicles, the lamellar bodies (LB), which are ultimately secreted to the alveolar space.

Surfactant isolated from the alveoli consists of several subtypes, or forms, which vary in density, morphological appearance, surface properties, and protein composition (for review see Ref. 11). These subtypes, which can be separated by either density gradient (12) or differential centrifugation (4, 21, 26), are believed to be in metabolic sequence, such that the heavy-aggregate forms (H), consisting of large, multilamellated structures and constituting the newly released surfactant, are a metabolic precursor of the small, light-aggregate forms (L), which consist of small, unilamellar vesicles and represent the used, non-surface-active breakdown product of H (for review see Ref. 11).

Recent studies (10) indicate that the metabolic conversion of H to L requires the activity of an enzyme, tentatively identified as a 76-kDa serine protease given the name "convertase" (14). The activity of this enzyme has been identified in H in many species, including mice (14), sheep (28), and rabbits (29), and has been localized to LB in adult mice (6, 7, 14).

Gross and Narine (13) developed a technique to study convertase activity in vitro. Referred to as surface-area cycling, this technique involves the end-over-end rotation of a tube containing resuspended large-aggregate surfactant isoforms. The end-over-end rotation repeatedly changes the surface area of the large-aggregate surfactant suspension and results in its conversion to small aggregates.

Using this technique, Ueda et al. (28) demonstrated significant developmental changes in the rate of conversion of H obtained from fetal and newborn lambs. These changes were found to correlate with the in vivo function of the preparation when tested in preterm rabbits, in that samples from very immature fetal lambs had a high conversion rate and poor in vivo function, whereas samples from late-gestation fetuses and early newborns had a low conversion rate and excellent in vivo function. These properties were attributed to the surfactant protein A (SP-A) content, which also changes dramatically during development. The subcellular source of convertase activity was not examined in their study.

The rabbit is used by many investigators as a model for studying the development of the lung surfactant system (1, 3, 23). Recent studies in our laboratory (2, 19) showed that, when tested in the preterm rabbit [27-day gestation (27d)], surfactant prepared from term fetal [31-day gestation (31d)] pups had superior in vivo properties compared with preparations obtained from adults. Preliminary studies suggested that this may be due to differences in convertase activity (unpublished observations). To date, no studies have been reported on convertase activity in the developing rabbit. The purpose of the present study was to use the in vitro cycling technique to investigate the developmental aspects of convertase activity in fetal and newborn rabbits and, then, to investigate the LB as a potential source of this activity.

	METHODS

Animals

Isolation of H

For each experiment, aliquots of H and L were removed for phospholipid analysis as previously described (25), and the percent recovery in each fraction was determined. In some experiments, the 140-g pellet was saved for analysis as described in In Vitro Conversion of H and LB to L. For each fetal experiment, lavage returns from several pups were combined for processing. For most newborn and adult experiments, individual preparations were used. In some experiments, lavage returns from several pups of the same developmental age were pooled to examine the effects of variables such as phospholipid concentration and cycling time on the percent conversion of H to L.

Isolation of LB

In Vitro Conversion of H and LB to L

For the cycling experiments, aliquots of H or LB (or 140-g pellets, where applicable) containing known amounts of lipid phosphorus (which varied according to the nature of the experiment, as described below) were made up to 2.0 ml with conversion buffer (0.15 M NaCl, 10 mM Tris, 1 mM CaCl₂, 1 mM MgCl₂, 0.1 mM EDTA, pH 7.4) in 12 × 75-mm polystyrene tubes (Becton-Dickinson, Lincoln Park, NJ) and mixed, and surface-area cycling was then performed as described by Gross and Narine (13) and Veldhuizen et al. (30). Briefly, the tubes were attached to the disk of a Rototorque rotator (Cole-Parmer, Chicago, IL) in an incubator at 37°C. The tubes were rotated for specific periods of time (up to 18 h) at 40 rotations/min so that the surface area of the contents changed from 1.1 to 9 cm² twice in each cycle. Control samples, containing the same quantity of H or LB (or 140-g pellet), were maintained without cycling at 37°C for the duration of the cycle experiment. At the end of the experiment, the tubes were centrifuged for 30 min at 10,000 g. The supernatant, containing the converted L, was carefully removed and extracted with chloroform-methanol (2:1 vol/vol), and the lipid phosphorus content was determined as described above. The pellet, containing the unconverted H or LB, was suspended in a small volume of 0.85% NaCl, and the phospholipid content was determined as described above. The percent conversion of each H or LB fraction (or, where applicable, 140-g pellet) to the L form was determined from these phospholipid values.

In a preliminary experiment, our laboratory (26) demonstrated the presence of small, unilamellar vesicles, typical of L components, thus confirming that the cycling procedure did indeed result in the conversion of H to L.

Data Analysis

	RESULTS

Variables Affecting Percent Conversion of H and LB

Phospholipid concentration. To examine the effect of phospholipid concentration on the percent conversion, samples of LB or H containing from 3 to 25 µg of lipid phosphorus were cycled for 180 min, with uncycled samples serving as controls. For each sample examined, the percent conversion varied inversely with the phospholipid concentration of the starting sample. Representative results are shown in Fig. 1 for H preparations obtained from +4d pups and adult rabbits.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Effect of phospholipid concentration on percent conversion of heavy-aggregate alveolar surfactant (H) to light-aggregate form (L). Representative samples containing 3-25 µg lipid phosphorus were cycled at 37°C for 3 h (open symbols) or maintained at 37°C for 3 h without cycling (solid symbols). Squares, 4-day-old (+4d) pups; circles, adult rabbits.

Cycling time. To examine the effect of cycling time on the percent conversion of H and LB to L, samples containing 3-4 µg of H or LB lipid phosphorus were cycled for various time periods up to 18 h. Uncycled samples, maintained at 37°C for comparable time periods, served as controls. The percent conversion of both H and LB to L increased with increasing cycling time. Representative results are shown in Fig. 2 for H preparations obtained from +4d and adult rabbits. For some samples, as shown by the +4d preparation in the diagram, maximum conversion was achieved with 3 h of cycling, whereas other samples did not achieve this until after 18 h (adult preparation in this diagram).

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Effect of cycling time on percent conversion of H to L. Each sample contained 3-4 µg lipid phosphorus and was cycled within 2 days after isolation from either +4d pups or adults. Samples were either cycled at 37°C (open symbols) for times indicated or maintained at 37°C for comparable time periods without cycling (solid symbols). Symbols are as described in Fig. 1 legend.

Storage time. Percent conversion of both H and LB decreased over time on storage at 4°C (Table 1). Loss of activity was generally observed after 7 days of storage, regardless of the phospholipid content of the starting sample. After 7 days, there was a progressive decline for at least 13 more days. This is illustrated in Table 1, which shows the decline in activity of representative samples of H and LB containing 3-4 µg of lipid phosphorus that were stored for up to 10-20 days.

LB: Percent Conversions to L

At each developmental stage, there was considerable conversion of LB to L, i.e., >82% in the 27d fetal preparation and ~78% in adult (Fig. 3). Slightly but significantly lower values were found in the term fetus and +1d newborn.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Percent conversion of lamellar bodies (LB) to L in developing rabbits. Developmental ages studied were 27- (27d) and 31-day-gestation (31d) fetuses; 1-day-old (+1d), +4d, and 7-day-old (+7d) neonates; and adults. Each sample contained 3-4 µg lipid phosphorus and was cycled for 180 min within 2 days after isolation. Nos. in parentheses, no. of determinations. * Significantly different from 27d, P < 0.05 by Duncan's multiple-range test (DMRT); t significantly different from value at previous observation time, P < 0.05 by DMRT.

There was little variation in the values obtained from individual pups of a single litter or pooled samples obtained from different litters of the same gestational age.

Alveolar Wash of H

Percentage of total alveolar pool (H + L) during development. In the 31d pup, H comprised the bulk of the alveolar surfactant pool (88.0 ± 3.2%, n = 4). This value significantly decreased to 74.9 ± 0.8% (n = 3) in the +1d newborn and to 58.9 ± 5.1% (n = 4) in the +4d newborn (P < 0.05 by DMRT). There were no further significant changes after +4d. The values for the +7d newborn and adult were, respectively, 52.5 ± 5.6% (n = 5) and 55.4 ± 12.1% (n = 7).

Percent conversion to L. As for LB, all H samples were cycled for 3 h within 2 days of preparation, and each contained 3-4 µg of lipid phosphorus. Because negligible quantities of alveolar surfactant were present in the fetal lung at 27d, there is no value for this developmental stage (Fig. 4). In the term fetal preparation, there was little conversion, i.e., <20% after 3 h of cycling. This value progressively and significantly increased to >80% by postnatal day 4. There was no further change on postnatal day 7, but the value declined somewhat in the adult. However, it is important to note that the percent conversion for adult H, as well as for each individual neonatal H, was significantly greater than the term fetal value.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Percent conversion of H to L in developing rabbits. Developmental ages studied are as described in Fig 3 legend. Each sample contained 3-4 µg lipid phosphorus and was cycled for 180 min within 2 days after isolation. Nos. in parentheses, no. of determinations. * None present; t significantly different from 31d, P < 0.05 by DMRT;  significantly different from value obtained at previous observation time, P < 0.05 by DMRT.

Protein Content of H and LB

Properties of 140-g Pellets

Results (Table 2) indicate that the 140-g pellet of the 31d fetus and +4d newborn comprised ~20% (23.8 ± 2.5 and 17.9 ± 4.4%, respectively) of the total alveolar phospholipid pool. Both had phospholipid compositions characteristic of secreted surfactants (i.e., ~80% phosphatidylcholine and 12-14% phosphatidylglycerol + phosphatidylinositol), and both demonstrated considerable conversion to an L form after in vitro cycling (i.e., almost 50% for the 31d fetus and 80% for the +4d pups).

In contrast, the 140-g pellet of the adult comprised <10% of the total alveolar phospholipid pool. Its phospholipid composition was completely uncharacteristic of surfactant, and it converted minimally to an L form (14.2 ± 6.3%).

Surface Properties of H and LB

	DISCUSSION

Alveolar surfactant exists in several subtypes or aggregate forms. The major forms are the H, constituting the newly released surface active surfactant structures, and the L, which constitutes a non-surface-active metabolic breakdown product of H (11). Evidence suggests that conversion of H to L requires the activity of an enzyme, namely convertase (10, 14). Although the enzyme has not yet been isolated in pure form (18), its activity has been demonstrated in two ways. The first involves an immunologic assay using an antibody developed in rabbits to the partially purified protein (6, 11), and the other is by an vitro surface-area cycling technique introduced by Gross and Narine (13). This latter technique has been used by several investigators to further elucidate the properties of this activity (27, 28, 30).

Our interest in this enzyme arose from recent studies in our laboratory in which we demonstrated a more prolonged survival and a more rapid clearance of lung water (19) in preterm (27d), surfactant-deficient rabbits treated with a heavy-aggregate preparation obtained from term fetal rabbits, compared with that of pups treated with preparations isolated from adults. A subsequent study (2) suggested that the superior response to the term fetal preparation might be attributed to its slower clearance, compared with the adult preparation from immature lungs. We then questioned whether the observed differences in clearance may in turn be related to developmental differences in the state of activity of the convertase enzyme, such as had been demonstrated by Ueda et al. (28) in developing lambs. In a preliminary study, using the in vitro cycling technique, we did find a significantly higher rate of conversion of adult than term fetal preparations. These preliminary findings support the concept of differences in the state of activity of the convertase enzyme in the term fetal and adult preparations. In the present study, we studied the developmental aspects of this activity more extensively and, furthermore, explored the LB as a potential source.

We prepared our H fractions from alveolar lavage by differential centrifugation. Several investigators (15, 16, 30) have reported success with this method, although they use a higher g force, i.e., 40,000 g for 15 min to harvest H. We used a 10,000-g, 30-min spin, as our laboratory had previously shown this to be optimal for pelleting the newly released multilamellated form of alveolar surfactant in fetal and newborn rabbits (25). Some investigators (11, 27-29) prefer the density-gradient procedure, first introduced by Gross and Narine (13), for preparing H. Although Putnam et al. (27) recently suggested that the density gradient procedure should be used to prevent structural distortion of H, other investigators (13, 30) reported no loss of structural integrity on electron-microscopic examination of both H and L samples, whether they were prepared by density gradient or differential centrifugation. We did not examine our samples by electron microscopy, but we did demonstrate that the cycling procedure did indeed convert the large aggregates to small vesicular structures, which, similar to the light subtype isolated from freshly prepared alveolar lavage, could be harvested as a pellet by centrifugation at 100,000 g.

We found that the cycling results were affected both by the storage conditions of the sample and, in agreement with a previous report by Gross and Schultz (10), the phospholipid concentration of the starting material. With storage at 4°C, the percent conversion was decreased but could be restored by prolonging the cycling time. Regarding the concentration of the starting sample, we found that 3-6 µg of lipid phosphorus gave optimum results. This is in contrast to the 20-25 µg used by many investigators (16, 28-30) with the same size cycling tubes, which, in this study, gave minimal conversion rates. To limit the number of variables and maintain consistency, we studied samples within a narrow, low-concentration range and within 2 days of preparation, and we recommend this approach for all studies of this type. This would minimize the difficulty of making comparisons, particularly among different studies.

Our results clearly demonstrate developmental changes in activity in the alveolar H preparations and suggest that the LB are the source of this activity. At each of the developmental stages examined, including the 27d preterm pup, the LB had high levels of activity, ranging from 70 to 80% conversion, and in all cases the activity was at least as high as, and in some cases higher than, the activity found in the alveolar preparation. These findings strongly support the concept of the LB as a major source of this activity in rabbits and are in agreement with studies by Gross and colleagues in mice (6, 7, 14).

Our data also showed that, on release of the LB to the alveoli, particularly in the term fetus and early neonate, considerable activity was lost, or at least not totally accounted for, as evidenced by the diminished activity in the alveolar P10 fraction. This diminished activity could not be attributed to a difference in protein content from the LB (29), because our data show very similar phospholipid-to-protein ratios among all of the LB and P10 fractions at all developmental ages. Activity found in the 140-g pellet, however, could account for at least some of this loss.

The recovery of activity in the 140-g pellets was of considerable interest given that these fractions are normally discarded in this type of study. On the basis of their phospholipid composition (25) and surface properties (26), the pellets from the fetus and early neonate (in contrast to the adult) probably represent the most recently released form of alveolar H (10). With the in vitro cycling technique, it was difficult to assess how much of the total convertase activity was actually recovered in 140-g pellets. Further study is warranted, especially during perinatal development when major changes appear to occur. The immunocytochemical method for assessing convertase activity would be very useful for this type of study.

Our demonstration of changes in the percent conversion of H to L in the perinatal rabbit is in agreement with the findings of Ueda et al. (28) in developing lambs. However, the patterns of change in the two species were different. In the lambs, the most significant changes occurred during fetal development; i.e., the highest percent conversion was in the most immature fetuses and it progressively decreased toward term. Although further changes were noted, i.e., a further decrease in the newborn and then an increase in the adult, these changes were not statistically significant. We did not have enough material to examine activity in H from our most immature study group (i.e., at 27d). We did, however, demonstrate high activity in the LB isolated at this developmental stage, but it is difficult to determine whether this high activity is retained on release to the alveoli. In our study group, we found that H from the term fetus had the lowest percent conversion and, in contrast to the findings of Ueda et al., that the percent conversion progressively and significantly increased in the newborn up to postnatal day 4. Although it decreased somewhat in the adult, the adult value, as well as all of the neonatal values, was at least three times greater than that obtained for the term fetus.

The difference in profiles between the two studies can be attributed in part to species differences, as well as to the fact that we focused on different developmental stages, i.e., mostly fetal for the Ueda et al. (28) lamb study and mostly neonatal for our rabbit study. There are also a few procedural differences that could contribute to these different profiles. For example, they used a 10-min rather than a 5-min 140-g preliminary spin. This could potentially remove more of the newly secreted surfactant, with associated enzyme activity, from their assay. Also they used 15-20 µg of lipid phosphorus for their cycling experiments rather than the 3-4 µg that we found optimal to give the maximum percent conversion. The very low percent conversion rates that they report for all of their samples might in fact be explained by this factor. Also, as they did not report on the storage conditions of their samples, it is difficult to determine whether this factor contributed to the differences observed in their profile.

Despite these differences, there is a major point of agreement between the study of Ueda et al. (28) and ours in that the developmental changes in the percent conversion of H to L reflect the changes in the proportion of H to L in alveolar lavage. Another major point of agreement is that the H samples with the lowest convertase activity in both species (2, 19, 28) had the greatest biological activity when tested in premature newborn rabbit pups.

From the foregoing it is evident that, at least in lambs and rabbits, convertase activity in the H, as determined by the in vitro cycling technique, undergoes dramatic changes during development. Whereas the reasons for these changes are not yet clear, they may be related to changes in SP-A levels as suggested by Ueda's group (28) or possibly to the structural and chemical changes, i.e., extensive glycosylation (5), that SP-A undergoes during perinatal development. Further study is warranted to clarify these issues.

In summary, our data suggest the LB as a major source of convertase activity in rabbits and that there are significant developmental changes in this activity once the LB are released to the alveolar space. High activity was found in LB at every developmental stage examined. In the newly released alveolar surfactant (H fraction), activity was significantly decreased in the term fetal preparations compared with those from newborns and adults. The low conversion rate of the H fraction isolated from term fetal rabbits may explain the superiority of this preparation in treating an immature rabbit model in that it prolongs the availability of the instilled preparation at the alveolar space. These findings should contribute to our understanding of developing an optimal treatment for surfactant deficiency in preterm newborns.