surfactant stabilizes the alveoli of mammalian lungs preventing them from collapse at end expiration. This is due to the reduction of surface tension to values below 5 mN/m on compression of the air-liquid interface during expiration (11). Recently, the importance of surfactant to stabilize small airways has been postulated (11), and impaired function of this airway surfactant may contribute to the airway obstruction and compromised lung function characteristic of asthma (18). Airway secretions contain significant amounts of surfactant material, with a phospholipid composition very similar to alveolar surfactant (4, 33, 41). However, surfactant entering the airways of mammalian lungs from the alveolar spaces appears to have a different surface tension function than that of alveolar surfactant and cannot generate minimal surface tension values (γ_min) below 20 mN/m. Whereas surfactant, presumably derived from proximal airways, can be purified by density gradient centrifugation from induced sputum or tracheal aspirates (4,33), it has not proved possible to purify and study airway surfactant from small bronchioli of mammals without substantial contamination from alveolar-derived material (18).

Bird lungs, however, possess air capillaries instead of alveoli, and, unlike the alveoli of mammalian lungs, these tubular structures are not ventilated by convection. Instead, they are aerated during inspiration and expiration via a highly differentiated system of parabronchi and cranial and caudal air sacs. This leads to an air flow in a single direction with a highly efficient, cross-flow-mediated oxygenation of the capillary blood (23, 34). Granular (type II) pneumocytes with morphological characteristics of surfactant-producing cells are located in the atrial walls, air sacs, and parabronchi but not in the air capillaries (36). Surfactant secreted from the atrial type II pneumocytes then has to spread into the air capillaries, a process that appears analagous to airway surfactant in the mammalian lung, which has to enter bronchioli after being secreted from type II pneumocytes into the alveolar space (28).

Consequently, we used surfactant from avian lungs to evaluate functional and compositional characteristics of surfactants predominantly designed to keep open small tubules (air capillaries in avian lungs) compared with a surfactant designed to prevent end-expiratory collapse of pulsating alveoli (in mammalian lungs) (11, 16, 22, 32). We purified surfactant from duck, chicken, and pig lungs and compared properties of morphology, biochemical composition, and static as well as dynamic surface tension function between avian and mammalian surfactants. This study aimed to determine the fundamental properties required for a surfactant designed primarily to keep open small tubules rather than alveoli.

MATERIALS AND METHODS

Materials.

Phospholipid standards were from Sigma-Aldrich (Deisenhofen, Germany). Hydrogen peroxide (30%, analytic grade) was from Boehringer Ingelheim (Ingelheim, Germany), and perchloric acid (70%, analytic grade) was from Merck (Darmstadt, Germany). All organic solvents were of HPLC grade and from Baker (Deventer, Netherlands). All other chemicals were of analytic grade and obtained from various commercial sources.

Isolation of surfactant from lung lavage.

Lungs of freshly slaughtered pigs were obtained from the local slaughter house. Surfactant was isolated from these lungs by differential ultracentrifugation of blood-free lung lavage fluid (4). Chicken and ducks, purchased from our local animal house, were fed ad libitum and had free access to tap water. Lavage fluid was harvested from chicken and ducks after death by cerebral disintegration. A catheter was inserted into the trachea, the cranial air sacs were ligated, and the lungs together with the other air sacs were flushed with 1–1.5 liters of ice-cold 154 mM saline. Fluid recovery was 50–70% of instilled saline. Lungs were subsequently lavaged twice with up to 1 liter of saline, at the fractions were pooled to give a final volume of 1.0–1.3 liters. Lavage fluid was centrifuged at 500 g × 15 min at 4°C to remove cells. The supernatant was then centrifuged at 27,000 g× 3 h at 4°C to give a pellet (P27), which was resuspended in 154 mM saline supplemented with 1.5 mM calcium chloride. In some control experiments, P27 was further purified by density gradient centrifugation with sodium bromide as previously described for the pig (4), which did not alter surface tension function or phospholipid composition. Surfactant preparations were stored at −80°C until functional and biochemical measurements.

Isolation of surfactant from lung tissue.

Intracellularly stored surfactant (ICS) was prepared from lavaged lung tissue according to Shelley et al. (35). In brief, aliquots (5 g) of lavaged lung tissue were homogenized with an Ultra-Turrax homogenizer (IKA Labortechnik, Staufen, Germany) at 4°C after the addition of an equal volume of 154 mM saline. The homogenate was then centrifuged (580 g × 10 min at 4°C), and the resulting pellet was washed twice with 3 ml of 154 mM saline and repeated centrifugation. Surfactant was then isolated from the pooled supernatants by sodium bromide density-gradient centrifugation as described previously (4). The resulting ICS fraction was resuspended in 154 mM saline supplemented with 1.5 mM calcium chloride.

Electron microscopy.

Chicken lung tissue and lavage surfactant preparations (P27) from pig, duck, and chicken lungs were fixed for morphological analyses in the presence of tannic acid (20). Pieces (1 × 1 × 5–7 mm) were cut from the periphery of the lungs and were fixed overnight at 4°C in 0.1 mol/l sodium cacodylate-HCl buffer pH 7.2 containing 3% (wt/vol) glutaraldehyde and 1% (wt/vol) tannic acid. Surfactant suspensions were centrifuged at 60,000 g for 1 h, and the pellets were similarly fixed for 6 h. After postfixation in 2% osmium tetroxide for 2 h, the samples were dehydrated in ethanol dilutions and embedded in Epon (Serva, Heidelberg, Germany). Ultrathin sections, ∼60 nm in thickness, were cut with a diamond knife and examined in a Zeiss EM10 electron microscope (Zeiss, Oberkochen, Germany).

Analysis of phospholipids and hydrophobic surfactant proteins.

Total lipid and surfactant proteins (SP)-B and -C were extracted from the various surfactant preparations using chloroform and methanol according to Bligh and Dyer (5) and from lavaged lung according to Folch et al. (13). The concentration of total phospholipid in these lipid extracts was quantified as inorganic phosphate (3) after digestion of the organic components at 190°C for 35 min in the presence of 500 μl 70% perchloric acid (wt/vol) and 200 μl 30% hydrogen peroxide (wt/vol). For calculation of results relative to dry weight of the lungs, aliquots (500 mg) of lung homogenate were dried at 85°C until constant weight. The fractions of surfactant phospholipid in the P27 pellet and ICS suspension were then calculated relative to phospholipid concentrations of the 27,000 g supernatant and lavaged lung tissue, respectively.

For analysis of the molecular species compositions of phosphatidylcholine (PC) and sphingomyelin (SPH), lipid extracts from surfactants (P27, ICS) or tissue homogenate containing 0.5 or 1.0 μmol phospholipid, respectively, were dissolved in 1 ml chloroform containing 50 nmol dimyristoyl-PC (PC14:0/14:0) as internal standard (31). The fraction containing PC and SPH was then isolated by solid-phase extraction using Bondelut NH₂disposable cartridges (Varian Associates, Harbor City, CA) as described previously (4). Individual molecular species of PC and SPH were resolved by reverse-phase HPLC using a 4.6 × 250-mm SpherImage ODS II column (Schambeck, Bad Godesberg, Germany) at 50°C and detected by postcolumn fluorescence-derivative formation with 1,6-diphenyl-1,3,5-hexatriene (30). Eluted molecular species were quantified by integration of the fluorescence signal response, and total PC concentration was calculated as the sum of the signal responses of all PC species relative to that of the internal standard.

SP-B and SP-C concentrations were analyzed in total lipid extracts (0.5- to 1.2-μmol phospholipid aliquots) as described by van Eijk et al. (38). Briefly, a glass column (1 × 30-cm inner diameter) was filled with Sephadex LH-60 (Pharmacia, Uppsala, Sweden) as the stationary phase (total bed volume = 20 ml), which was equilibrated with a mobile phase of dichloromethane-to-methanol-to-0.1 mM HCl mixture (30:65:5, vol/vol) at a flow of 0.4 ml/min. Dried lipid extracts were dissolved in 50 μl dichloromethane-to-methanol-to-0.1 mM HCl mixture (70:25:5, vol/vol) and applied to the column using a 100-μl injection slope. SP-B and SP-C protein in the eluate was detected by ultraviolet absorbance at 228 nm and quantified in collected eluate fractions using the fluorimetric assay of Böhlen et al. (6).

Analysis of SP-A.

The SP-A concentration of surfactant samples (P27) was determined by immunoblot analysis (4) after SDS-PAGE (24), with purified porcine SP-A from lung lavage fluid as a standard. After the samples were heated under reducing conditions with Laemmli buffer (100 mM Tris · HCl, pH 6.8, 20% glycerol, 2% SDS, 10% 2-mercaptoethanol, 0.001% bromphenol blue), SP-A was separated on a 12% polyacrylamide gel with a 3% stacking gel. Prestained low molecular mass standards (lysozyme, 20.9 kDa; soybean trypsin inhibitor, 28.1 kDa; carbonic anhydrase, 35.6 kDa; ovalbumin, 50.8 kDa; bovine serum albumin, 77.0 kDa; phosphorylase B, 106.0 kDa) were from BioRad (Munich, Germany). After SDS-PAGE, the proteins were blotted onto a nitrocellulose membrane (Protran BA85, Schleicher & Schuell, Dassel, Germany) and efficiency of transfer was assessed using Ponceau S red dye. The primary antibody was polyclonal rabbit anti-human SP-A antibody (kindly provided by Byk Gulden, Konstanz, Germany). This antibody cross-reacted with porcine and bovine SP-A and was used at 1:2,000 dilution in phosphate-buffered saline with 0.04% bovine serum albumin. A porcine HRP-labeled anti-rabbit IgG antibody (1:2,000 in 10 mM Tris, pH 7.4, with 1% bovine serum albumin, DAKO, Hamburg, Germany) served as the secondary antibody. Detection was performed with chemiluminescence using the enhanced chemiluminescence detection system (Amersham, Freiburg, Germany). Total protein in samples was determined according to Lowry et al. (25).

Analysis of surfactant function.

Surfactant function was assessed with the pulsating bubble surfactometer (PBS; Electronetics, Amherst, NY) (10) and the capillary surfactometer (CS; Calmia Medical, Toronto, Canada) (12). For the PBS, surfactant preparations were adjusted with 154 mM NaCl supplemented with 1.5 mM CaCl₂ to give two phospholipid concentrations, either 1.33 or 4 mM, and were vortexed for 30 s to ensure complete homogenization. The resulting surfactant suspensions were instilled into the sample chamber of the PBS. Because body temperatures are higher in birds than in mammals (usually 41–42°C), we investigated surface tension function at both 37°C and 41.5°C. A bubble communicating with ambient air was created in the surfactant suspension, and surfactant was allowed to adsorb to the air/liquid interface for 10 s. After this time, the bubble was pulsated between a minimum radius of 0.4 mm and a maximum radius of 0.55 mm. Because the tubular bird lung is not subjected to the rapid surface area changes characteristic of mammalian lung, dynamic surface tension function was investigated over a wide range of pulsation frequencies (1, 5, and 20 pulsations/min). The pressure across the bubble was measured by a pressure transducer, and the surface tension was calculated using the Laplace equation. The surface tension after 10 s adsorption (γ_ads) and the γ_min and maximum surface tension (γ_max) after 1–10, 21, 50, and 100 cycles were determined.

The CS was used to model surfactant function in the cylindrical surface of a narrow tube such as conducting airways of mammalian or the air capillaries of avian lungs (12). Briefly, surfactant samples were adjusted to 1.33 or 0.133 mM phospholipid and a small volume, sufficient to occlude the capillary (0.5 μl), was placed into the thin (0.2-mm inner diameter) section of a glass capillary. The capillary was placed into the CS, and a stream of air was applied at one end of the capillary. As the pressure rises, the liquid will be pushed out of the thin section of the capillary and lead to a drop of pressure. Poor surfactant will not be able to prevent reformation of a liquid droplet, which reoccludes the capillary, whereas good surfactant will be able to prevent the reformation of an occluding droplet (11, 12). Openness of the capillary, as a parameter of the surfactant's ability to keep small tubules open, was determined as the time of zero pressure in the capillary and expressed as “percent open” of a total measuring time of 120 s.

Statistics.

Data are expressed as means ± SD. One-way analyses of variance were calculated using GraphPad Instat Version 1.11a (GraphPad Software, San Diego, CA). Group differences were tested by a two-tailed Student's t-test using the Bonferroni correction for multiple-group comparison, with P < 0.05 being considered significant.

RESULTS

Light and electron microscopy.

The distribution of granular (type II) pneumocytes and the location of surfactant in the gas-exchanging air capillaries of the avian lungs were examined by light and transmission electron microscopy. Granular pneumocytes were evident in the epithelial lining of the atria of chicken lungs, which connect the lumen of parabronchi through infundibula with the air capillaries as the main locus of gas exchange (Fig. 1A). These atrial granular pneumocytes contained the microvilli and cytoplasmic lamellar bodies characteristic of type II pneumocytes from mammalian lungs (Fig.1, B and C). However, none of these putative type II pneumocytes were detectable in the walls of the tightly packed air capillaries (Fig. 1, D and E) of avian lungs. Nevertheless, surfactant was present in these air capillaries as a mostly continuous thin layer, which was no thicker than the plasma membrane of the underlying epithelial (type I) pneumocytes (Fig.1F). Electron microscopy of surfactant isolated from lung lavage fluids (P27) showed the expected morphology of pig surfactant (Fig. 2A), containing tubular myelin, granular and amorphous material, and secreted lamellar bodies. In contrast, the surfactant pellet (P27) from both duck (Fig. 2B) and chicken (Fig. 2C) lungs was composed mainly of oligo- and multilamellar structures. Electron microscopy provided no evidence that avian surfactant contained any tubular myelin.

Biochemical composition of lavage fluid, surfactant, and lung tissue.

The proportion of lavage fluid phospholipid that sedimented after centrifugation at 27,000 g (P27 surfactant) comprised 74.1 ± 6.6, 69.4 ± 16.7, and 39.8 ± 10.3% of total lavage phospholipid for pig, duck, and chicken, respectively (Table1). Despite this wide variation, the fractional content of PC in P27, expressed relative to total phospholipid, was relatively constant for all three animal species (Table 1). The principle component of PC in P27 surfactant from the pig (n = 8) was dipalmitoyl-PC (PC16:0/16:0, 48.9 ± 3.4%) followed by palmitoyloleoyl-PC (PC16:0/18:1, 15.9 ± 3.9%), palmitoylmyristoyl-PC (PC16:0/14:0, 13.0 ± 4.5%), and palmitoylpalmitoleoyl-PC (PC16:0/16:1, 11.3 ± 2.9%). Compared with the pig, P27 surfactant from both duck (n = 6) and chicken (n = 11) contained significantly (P < 0.001) more PC16:0/16:0 (66.3 ± 2.2% and 67.4 ± 5.6%, respectively), at the expense of lower amounts (P < 0.001) of PC16:0/14:0 (1.7 ± 0.7% and 2.9 ± 1.1%, respectively) and PC16:0/16:1 (5.5 ± 1.1% and 4.6 ± 0.7%, respectively; Fig.3A). The content of PC16:0/18:1 was comparable for surfactant PC from all three animals studied, ranging from 14 to 16 mol%.

Although the concentration of total phospholipid in lung tissue was lower in avian than in porcine lung tissue (Table 1), the fraction of ICS in relation to total lung phospholipid was comparable in lavaged lungs from all three animal species. The PC composition of this intracellular surfactant fraction was identical to that of lavage (P27) surfactant for each animal (Fig. 3, A andB), in contrast to the PC composition of lavaged lung tissues (Fig. 3C). Compared with surfactant fractions, lavaged lung tissue from all three animal species contained a lower proportion (P < 0.001) of PC16:0/16:0 (pig: 33.6 ± 4.4%; duck: 43.1 ± 2.4%; chicken: 42.2 ± 3.1%) and greater amounts (P < 0.001) of PC16:0/18:1 (pig: 23.0 ± 3.1%, P < 0.01; duck: 27.5 ± 1.2%, P < 0.001; chicken: 25.5 ± 1.4%,P < 0.001). Similarly, sphingomyelin, which is essentially a cell membrane component, was increased in lung tissues (10–15%; Fig. 3C) over the corresponding mean values in P27 surfactant (0.5–1.6%, P < 0.001; Fig. 3A) and ICS (1.3–4.9%, P< 0.001; Fig. 3B).

Analysis of SPs.

The concentration of total protein (μg protein/μmol phospholipid) was higher in duck (14.5 ± 3.7, P < 0.05) and chicken (18.1 ± 2.7, P < 0.05) P27 surfactant than in porcine material (4.3 ± 2.3). Western blot analysis could not demonstrate any immunologically detectable SP-A in duck and chicken surfactant. Essentially identical negative results were obtained when immunoblots were probed either with a polyclonal anti-human SP-A that cross-reacts with human, porcine, and bovine SP-A (Fig.4) or with a polyclonal anti-rat SP-A antibody (kindly provided by Prof. L. M. G. van Golde, Utrecht, Netherlands), which detected rat, porcine, and human SP-A (results not shown). As expected, both surfactant and lavaged tissue from pig lung contained significant amounts of SP-B and SP-C (Fig. 5 and Table2), but, whereas SP-B was abundant in avian surfactant and lung tissue, neither duck nor chicken surfactant contained detectable amounts of SP-C. Absence of a silver-stained band at the molecular mass range of SP-C (4–5 kDa) in avian but not in porcine surfactant was confirmed by SDS-PAGE of surfactant samples (results not shown).

Dynamic and static surface tension functions of avian vs. mammalian surfactant.

Porcine P27 surfactant easily reached γ_min values below 5 mN/m at 20 (Fig. 6, Aand B) and 5 (not shown) pulsations per minute within 5 cycles at both 37°C and 41.5°C. At one pulsation per minute, porcine surfactant was equally effective at 37°C, whereas at 41.5°C, it only reached γ_min values of 8 ± 5 mN/m (Fig. 6, C and D). In contrast, γ_min values of duck and chicken surfactant remained at 15–20 mN/m throughout the initial 10 cycles, and, even after 100 pulsations, γ_min was higher for duck and chicken compared with pig surfactant (Table 3). However, surface tension at the maximal radius of the pulsating bubble (γ_max) was lower (P < 0.01) during the first 10 pulsations in the bird surfactants (25–30 mN/m) than in pig surfactant (31–33 mN/m; Fig. 7). Surface tension after 10 s static adsorption was in the range 24–27 mN/m for all tested surfactants at 37°C as well as 41.5°C (Table 3). Moreover, avian and porcine surfactants were equally potent when assessed in the CS at phospholipid concentrations of 1.33 mM (not shown) and at a low concentration of 0.133 mM (Table3). Both avian and porcine surfactants kept the capillary more than 90% open at both 37 and 41.5°C.

DISCUSSION

In this study, morphological, biochemical, and functional aspects of avian lung surfactant have been compared with those of a typical mammalian surfactant. As birds have tubular lungs with no alveoli, bird lung surfactant functions to maintain air flow in the air capillaries rather than to prevent alveolar collapse at end expiration (9, 15). Surface tension is important for tubular structures of low radius; application of the law of Laplace to tubules (P = γ/r; P, opening pressure, γ, surface tension,r, radius of the tubule) implies that if surface tension were high, there would be a significant negative pressure across the endothelial/epithelial barrier leading to influx of fluid or collapse of the tubule. Surfactant acts in the air capillaries of avian lungs both as an “antiglue” to prevent adhesion of respiratory surfaces and to prevent liquid influx into the lungs (9). An analagous and essential role of surfactant for small airway patency in alveolar lungs (11) was recently highlighted by the finding of impaired surface tension function of surfactant from asthmatic patients after local allergen challenge (17,18). Consequently, as it is not possible to investigate the properties of surfactant from terminal airways of alveolar lungs without contamination from alveolar material, we investigated the properties of avian surfactant as a prototype of a surfactant designed to keep small tubules open instead of preventing collapse of an oscillating alveolus (9, 15, 23).

Morphology of the avian lung and its surfactant.

As previously shown for other avian species (36), microscopic analysis of chicken lungs demonstrated type II pneumocytes in the atria close to the entrances of air capillaries but not in the air capillaries themselves (see Fig. 1, D and E). Nevertheless, an almost continuous surfactant layer covered the luminal surface of the air capillaries, suggesting spreading of surfactant from the atria into the air capillaries. The mechanism for such efficient spreading of surfactant into the bird air capillaries is not clear. In mammalian lungs, the high surface pressure of surfactant at end expiration in the alveolus is proposed to drive surfactant into the airways (28), although there is no direct experimental evidence to support this concept. The absence of alveolar compression in the avian lung (9, 23) shows clearly that such expiration-dependent surface pressure forces are not essential to promote surfactant spreading, at least for bird surfactant. However, it remains to be established whether this conclusion has any direct relevance for understanding of mechanisms of spreading of mammalian surfactants in airways.

Surfactant function.

The surface tension properties of both bird surfactants are consistent with their lack of any role in alveolar function and may provide a good explanation of their ability to spread into the air capillaries. Whereas pig surfactant readily generated γ_min values <5 mN/m on compression, consistent with its role in preventing alveolar collapse at end expiration, neither avian surfactant could generate such low surface tension values under cyclical compression. These properties of avian surfactant function are directly comparable with those we have reported previously for surfactant purified from conductive airways of mammalian lungs, which only reaches γ_min values around 20 mN/m (4, 33). As the surfactant surface layer in the bird lungs is not subject to high compression forces, there is probably little surface sorting and squeeze-out of material and no formation of a collapse phase. Consequently, avian surfactant will be more disordered and effective at respreading than porcine surfactant, thus explaining its lower γ_max value. As both avian and porcine surfactant easily reached a surface tension close to equilibrium (γ_ads), they were also both equally effective at adsorption to the air/liquid interface. This combination of efficient respreading and adsorption of avian surfactant may contribute both to its effective transport into the air capillaries and to its ability to maintain their patency. It also explains why the avian capillary surfactant and mammalian alveolar surfactant were equally efficient in keeping open the glass restriction of the CS, even at the low concentration of 0.133 mmol phospholipid/l.

Phospholipid biochemistry of avian lung surfactant.

The detailed analysis of phospholipid compositions of lavage and lung tissue fractions (Fig. 3, A and C) shows that overall properties of the surfactant system were similar for bird and pig lungs. The comparable contents of ICS in avian and pig lungs, expressed as a fraction of total lung phospholipid (Table 1), together with the similar PC molecular species compositions of lavage and tissue surfactants (Fig. 3, A and B) demonstrate the selective storage and secretion of a surfactant of a characteristic phospholipid composition, differing substantially from the composition of the underlying tissue. The lower total phospholipid concentration in avian compared with pig lung tissue was probably due to the increased concentration of connective tissue fibers in avian lungs (14). The most probable sites for the storage of this ICS in bird lungs are the lamellar bodies that are apparent in type II cells of the lung atrium (Fig. 1, A-C).

Comparison of PC molecular species compositions between avian and mammalian surfactants, however, highlighted more subtle differences that may have functional significance. Whereas PC16:0/16:0 was enriched in secreted (P27) and ICS from both avian and porcine lungs compared with lung tissue, the content of PC16:0/16:0 was higher in both chicken and duck surfactants (Fig. 3) than the 50–55% of total PC typically reported for both porcine and human surfactant PC (4,31, 41). It is intriguing that this increased content of PC16:0/16:0 in avian surfactants was offset by decreased fractional contents of PC16:0/14:0 and PC16:0/16:1, both of which are integral components of all mammalian lung surfactants so far analyzed (4,31, 41). Whereas no defined role in mammalian surfactant function has yet been identified for either of these minor PC species, it is worth noting that their concentrations increase during fetal development in parallel to that of PC16:0/16:0 (21) and that their combined concentration is typically equivalent to that of phosphatidylglycerol (1, 4). Phosphatidylglycerol promotes adsorption of PC16:0/16:0 to the air/liquid interface in the mammalian alveolus, and it is also only present at low concentration in avian surfactant (14, 29). Consequently, it is possible that PC16:0/14:0 is essentially absent from avian surfactants, and PC16:0/16:1 is significantly diminished, because their presence in mammalian lungs has evolved to serve some specialized function within the alveolus such as potentiating the compressibility of the surface film.

It is also important to realize that this potentially significant distinction between avian and typical mammalian surfactants would not have been evident if PC analysis had been restricted to the determination of disaturated PC as the residue after oxidation of unsaturated PC species using osmium tetroxide (26). In addition to PC16:0/16:0, PC16:0/14:0 is also a disaturated species, and the osmium tetroxide techniques typically include contributions from unsaturated PC species such as PC16:0/16:1 and PC16:0/18:1 (19). This comparison highlights the value of molecular species analysis and emphasizes that disatured PC is not equivalent to PC16:0/16:0.

SPs of avian lung surfactant.

Alveolar surfactant contains at least four SPs, SP-A, -B, -C, and -D, and SPs are essential components of mammalian surfactant that promote surface adsorption and facilitate the highly dynamic properties of surfactant in the alveolus (16, 22, 39). Our results demonstrating the presence of SP-B but not of SP-C in either duck or chicken surfactant (Table 2, Fig. 5) are consistent with their proposed function in air capillaries rather than in alveoli. In mammalian lungs, SP-B is required for the formation of lamellar bodies (7), contributes to the ordering of phospholipids at air-liquid interfaces (2, 8), and its absence causes severe alveolar dysfunction (27). As SP-B promotes film formation at the air/liquid interface (29), its presence in bird lungs is consistent with the good adsorption parameters of avian surfactant. The absence of SP-C may contribute to the reduced compressibility of avian surfactant, because for mammalian surfactants, SP-C acts to promote reinsertion into the surface layer of surfactant lipids from the collapse phase. Moreover, selective insertion of PC16:0/16:0 into the surface layer depends on the cooperative action of surfactant apoproteins (29).

Because SP-A is essential for the formation of tubular myelin in mammalian surfactant (40), its absence from bird surfactant is consistent with the observed morphological structure (Fig. 2). This contained no tubular myelin but was characterized by numerous multilamellar structures that were similar to the structures generated by the addition of pure SP-B to phospholipid mixtures in vitro (40). However, expression of SP-A and SP-B has recently been demonstrated for chicken as well as nonmammalian vertebrate lungs (37, 42). SP-A of porcine and avian surfactant samples (P27) was determined by immunoblot analysis (4) after SDS-PAGE (24), using purified porcine SP-A from lung lavage fluid as a standard. Zeng et al. (42) demonstrated the presence of SP-A (cross-) reactivity in epithelial cells of major airways but not in either atrial type II pneumocytes or in air capillaries of chicken lung. Because our lavage procedure included material from air sacs and major airways, the absence of SP-A from P27 bird surfactants was somewhat surprising. Whereas the use of an inappropriate antibody is possible, we employed two different polyclonal anti-SP-A antibodies that cross-reacted with SP-A from all mammalian surfactants so far tested. One possible explanation for this discrepancy is that Sullivan et al. (37) measured immunoreactive SP-A in total lavage fluid, and its molecular structure was not characterized. Consequently, the protein they detected may have had a different oligomeric structure or an altered affinity for phospholipid, causing it not to associate with the P27 aggregate surfactant of birds. This difference of results mirrors our observations for mammalian airway surfactant, which showed SP-A to be reduced in surfactant purified from pig trachea (4) but present in supernatants from induced sputum samples obtained from human volunteers (41). Further work is evidently required to reconcile these different observations.

In conclusion, surfactant from duck and chicken lungs shares many properties with mammalian alveolar surfactants, such as a good ability to spread from its site of secretion and a high concentration of PC16:0/16:0, but also displays potentially important differences. At the very least, our study has demonstrated that poor compressibility and absences of PC16:0/14:0, PC16:0/16:1, SP-A, and SP-C are compatible with the function of a surfactant designed to keep open the air capillaries of bird lungs. Whether these observations have direct relevance to the composition and function of surfactant in the airways of mammals now deserves closer investigation.

Perspectives

Pulmonary surfactant in the mammalian lung is synthesized by and secreted from type II epithelial cells in the alveolus, where it opposes surface tension forces and enables alveolar expansion and contraction during the breathing cycle. All available evidence suggests that alveolar surfactant is the source of surfactant recovered from airways, where it is thought to maintain patency and prevent edema. The suggestion that dysfunction of this conducting airway surfactant contributes to the pathology of asthma has proved difficult to study directly, however, because recovered lavage material contains large amounts of alveolar surfactant. By contrast, the bird lung has a very different morphology and physiology and is composed of air capillaries that are aerated with no cyclical change to surface area during respiration. Consequently, birds may be good animal models to study the properties of a pulmonary surfactant designed to keep tubules open rather than to prevent alveolar collapse. Avian surfactant is secreted from type II cells located in the atria outside the air capillaries. The functional analyses presented in this paper show clearly that this bird surfactant is capable of adsorbing and spreading rapidly at an air/liquid interface, consistent with effective spreading into the air capillaries. In the absence of a requirement to stabilize alveoli, it is perhaps not surprising that avian surfactants displayed impaired compressibility compared with pig surfactant. In this context, the compositional differences among duck, chicken, and pig surfactant are intriguing. They suggest that an airway surfactant requires a high concentration of PC16:0/16:0 and SP-B but that other components characteristic of mammalian alveolar surfactant are not essential to maintain airway patency. The implication is that such components, including PC16:0/14:0, PC16:0/16:1, phosphatidylglycerol, and SP-C, may exert a primary role in alveolar but not in airway surfactant. These conclusions provide a good theoretical framework for future studies on the interactions of phospholipids and proteins in airway surfactant in relation to the pathology of asthma and other airway diseases.

FOOTNOTES

• This work was in part supported by the Deutsche Forschungsgemeinschaft Grant Ha1959/2.

• Address for reprint requests and other correspondence: W. Bernhard, Dept. of Pediatric Pulmonology & Neonatology, Hannover Medical School, Carl-Neuberg-Str. 1, 30625 Hannover, Germany (E-mail:bernhard.[email protected]de).

• The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.