INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mammalian lung surfactant is a lipoprotein mixture which possesses a unique biochemical make-up, allowing it to perform a critical role in vivo: the maintenance of alveolar stability during tidal breathing. Although surfactant apoproteins are essential for function, they represent only a small percentage of the total mass of surfactant, which is 85% to 90% lipid, the majority of which is phospholipid (PL). The principal phospholipid component of lung surfactant in virtually all species studied is phosphatidylcholine (PC), of which approximately one-half is the dipalmitoyl form (DPPC) (1). DPPC possesses a liquid-gel transition temperature above 37° C, a biophysical property which allows monolayers to achieve very high surface pressures during compression as the concentration of lipid is increased at an air-liquid interface (2). DPPC alone, however, is not an effective biological surfactant because it forms complex, thermodynamically stable micelles within the subphase when extruded from the interface. These do not readily readsorb to an interface during film reexpansion. Other secondary phospholipids and neutral lipid (NL) components, together with the surfactant apoproteins, are important for promoting reabsorption, and for allowing surfactant to function effectively during tidal respiration (3, 4).

Among the secondary phospholipids, phosphatidylglycerol (PG) and phosphatidylinositol (PI) are the most plentiful, together comprising about 10 to 15% of surfactant phospholipid content. They are also biochemically unique in that, at physiological pH, their head groups are negatively charged. This is in contrast to the remainder of surfactant phospholipids, which are zwitterionic. It has been postulated that PG and PI play a unique role in determining the biophysical properties of lung surfactant because of their unusual charge characteristics. Three potential specific functions for anionic phospholipids (APL) have been considered. It has been hypothesized that negatively charged head groups of PG and PI interact with selected basic amino acids (5, 6) near the carboxyl terminus of surfactant protein B (SP-B) to promote both adsorption and film stabilization at high surface pressures. Alternatively, it has been argued that the APLs promote selective squeeze out of less surface active lipid components during compression based on lipid interactions, independent of surfactant proteins. Finally, it has been suggested that APLs interact with the lipid-binding region of surfactant protein A (SP-A) via a calcium-dependent process to augment adsorption in the presence of SP-B.

The present study examines these potential roles of APLs by comparing the biophysical behavior of surfactant films reconstituted from phospholipids specifically deficient in PG and PI, with those containing normal surfactant phospholipids (PL).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Surfactant Isolation

Surfactant was isolated from fresh calf lungs within 3 h of procurement. Briefly, lungs were lavaged with 4 L of 0.15 M sodium chloride instilled under 25 cm H₂O pressure, and lavage return collected by passive drainage. Cells were pelleted by centrifugation at 250 × g for 8 min at 4° C and the supernatant pooled. Aliquots of 250 ml of supernatant were centrifuged at 40,000 × g for 45 min at 4° C, and the resulting crude surfactant pellets harvested. Each pellet was resuspended in 1 to 2 ml of sterile saline and dispersed by injection through a 25-gauge needle. The crude surfactant was pooled (15 ml), layered over 0.8 M sucrose (25 ml) in sodium chloride, and centrifuged at 30,000 × g for 45 min at 4° C. The pellicles present at the interface were aspirated, pooled, resuspended in saline, and ultracentrifuged at 80,000 × g for 30 min at 4° C. The resulting pellets were resuspended in 0.15 M NaCl containing 0.02% sodium azide, aliquoted into 1-ml samples containing 10 mg/ml of phospholipid each, and stored under nitrogen at 20° C. Samples stored in this fashion have retained normal biophysical function and biochemical composition for at least 3 mo.

Preparation of Surfactant Fractions

The methodology used is a modification of that previously described by Hall and coworkers (7). Aliquots of 1 ml of surfactant containing 10 mg of phospholipid were injected through a 25-gauge needle into 45 ml of butanol at room temperature. The aqueous soluble precipitate, containing predominantly SP-A and a small amount of serum proteins, was pelleted by centrifugation at 30,000 × g for 30 min at 4° C. The pellets were dried under nitrogen at room temperature, and resuspended in 20 mM octylglucopyranoside (OGP) by repeated drawing through a 27-gauge needle. The solution was then ultracentrifuged for 1 h at 100,000 × g at 4° C; the pellet, containing purified SP-A, was resuspended in 5 mM Tris-buffer (pH 7.45) and dialyzed extensively against 5 mM Tris. Total protein determinations were performed, and aliquots stored at 20° C for subsequent use.

The butanol supernatant containing the hydrophobic fractions including phospholipids, NL, and hydrophobic apoproteins (HA) was dried under vacuum and resuspended in CHCl₃:CH₃OH (1:1 by volume). Insoluble precipitate was removed by low-speed centrifugation, and phospholipid assay performed (8), to demonstrate adequate recovery before subsequent separation steps. The sample (in 2 to 3 ml total volume) was loaded onto a 1.5 × 100 cm Sephadex LH-20 column, and 1.5-ml fractions were eluted over 4 to 4.5 h at 4° C using a CHCl₃:CH₃OH:0.1 N HCl (1:1:0.1 by volume) mobile phase. Samples were neutralized with 32 µl of pH 12 0.05 M Hepes buffer to a final pH of 7.0 to 7.1. Protein, cholesterol, and phospholipid assays were performed on each fraction to verify adequate separation, and fractions containing each of these individual components were pooled for storage and subsequent analysis. PL, HA, and cholesterol-containing NL pooled samples were stored in neutral chloroform:methanol under nitrogen at 20° C.

The APL fraction was obtained by passing purified aliquots of PL fractions through a second 1.5 × 30 cm Sephadex LH-20 column equilibrated with nonacidified CHCl₃:CH₃OH (1:1) mobile phase (9). The entire eluent was collected and stored under nitrogen for biochemical analysis and subsequent reconstitution studies.

Biochemical Analysis

Total protein determinations on native surfactant, on surfactant fractions collected from Sephadex chromatography, and on pooled fractions were performed by the amido black assay using a bovine serum albumin standard curve (10). PL content was determined by molybdenum blue inorganic phosphate assay (11). Total cholesterol content was determined by ferrous sulfate reduction assay (12).

NL extraction of an aliquot of native calf lung surfactant (CLS) was performed by cold acetone extraction (13). NL composition of CLS and pooled NL fractions from Sephadex chromatography were compared qualitatively by thin layer chromatography (TLC) to verify similar composition before and after fractionation. NL and NL extracts of CLS containing 250 µg of cholesterol were separated on silica G/H thin layer plates (Fisher Scientific, Pittsburgh, PA) using hexane:diethyl ether:formic acid (80:20:2) mobile phase. Bands were detected by spraying with sulfuric acid:1.2% aqueous potassium dichromate (30:70) followed by charring for 15 to 20 min at 150° C. Spots corresponding to individual NL components were assessed in semiquantitative fashion by densitometry.

Total phospholipid extraction of native surfactant samples was performed using the method of Folch (14). Phospholipid content of CLS and pooled Sephadex PL fractions was determined by quantitation of PL subtypes after TLC separation. Four hundred micrograms of each sample, along with authentic PL standards, were spotted onto a Silica G/H TLC plate, and developed using a CHCl₃:CH₃OH:2-propanol:H₂O:triethylamine (42:12.6:9.8:35:35) mobile phase (15). After separation, PL subtypes were detected under low ultraviolet light after staining with 1,3,5 diphenylhexatriene solution. Spots were scraped, and PL extracted by resuspending in CHCl₃:CH₃OH::50:50 and vortexing vigorously for 1 h at room temperature. Silica powder was removed by centrifugation, and the chloroform fraction containing the PL withdrawn. PL content of each spot was determined by inorganic phosphate assay.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed on pooled Sephadex fractions to verify purity of the proteins, and lack of protein cross-contamination in the PL and NL fractions. All samples were run in duplicate under reducing and nonreducing conditions on 15% SDS polyacrylamide gels (16). Gels were run at room temperature at a constant voltage of 125 volts for 90 min. Each was stained with both Coomassie blue (0.1% in H₂O: MeOH:AcOH::50:40:10) (Sigma, St. Louis, MO) and silver stain reagent (Bio-Rad, Hercules, CA) to optimize detection of all components (resolution 10 to 100 ng of protein).

Sample Reconstitution for Functional Analysis

PL or phospholipids devoid of anionic species (DAPL), HA, and NL fractions were aliquoted in appropriate amounts into 1.5-ml polypropylene microfuge tubes from organic stock solutions, and dried under vacuum. Samples were then reconstituted in saline or saline with either 0.5 or 5.0 mM CaCl₂ as indicated, resuspended by repeated injection through a 27-gauge needle, and sonicated on ice for 30 s (W-375 Sonicator; Heat Systems-Ultrasonics, Farmingdale, NY). Appropriate volumes of aqueous SP-A were then added to designated samples, and mixing performed by repeated injection through a 27-gauge needle. Samples were left to equilibrate for 15 min at 37° C before making surface tension recordings.

Surfactometry

Surface tension versus surface area measurements were recorded using a pulsating bubble surfactometer (PBS) (Electronetics Corporation, Amherst, NY) operated using a modified version of the leak-free methodology previously described by Putz and colleagues (17, 18). Our modification specifically involves using putty to occlude the open end of the capillary tube, followed by sample loading using a 1-ml syringe with a 27-gauge needle. The bubble is then formed under direct visualization, and with suction applied using the needle valve control, the putty is removed, preventing fluid from entering the capillary tube. This approach prevents wetting of the capillary tube in the majority of circumstances; in those cases where fluid is observed to enter the tube, the sample is reloaded.

All surfactometry recordings were made at 37° C. Equilibrium surface tension measurements were obtained by initiating recording, and subsequently forming a fresh bubble of 0.4 mm so that details of initial adsorption dynamics could be measured. Dynamic measurements were made for samples containing 1, 0.1, and 0.01 mg/ml of phospholipid, at oscillation frequencies of 1, 20, and 100 cycles/min between 0.34 and 0.50 mm minimum to maximum bubble radius (A/A_mean = 73%). At each frequency, recordings were performed for at least 15 min to ensure that steady state had been reached. At the lowest concentration, observations were made for up to 1 h in some cases, with periodic resizing of the bubble after briefly stopping oscillations.

Leakage was readily detected by an alteration in dynamic surface tension-surface area profiles, and was confirmed by visual inspection through the microscope objective. When leakage was detected, recordings were terminated, and samples remeasured. All runs were performed in triplicate, and reproducibility was judged objectively based on maximal surface tension during film expansion, minimum surface tension during film compression, and the degree of compression required to reach minimum surface tension.

Statistics

Results are summarized as mean values plus or minus standard deviations. Comparisons of PL and DAPL sample phospholipid composition and biophysical function were performed by Student's t test. Significant differences were defined as a p value < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Biochemical Characterization of Pooled Fractions Used in Reconstitution Experiments

Protein, phospholipid, and cholesterol content of pooled PL, DAPL, HA, SP-A, and NL fractions used in reconstitution experiments are summarized in Table 1. PL and DAPL fractions contained less than 1% protein by amido black assay, and no detectable cholesterol. The HA fraction contained predominantly protein (95%), but had a small amount of detectable PL by molybdenum blue assay. However, this contributed less than 1% of the total phospholipid content in reconstitution studies given the small amount of HA fraction utilized in these experiments. Similarly, the SP-A fraction (93% protein) contained a small amount of detectable cholesterol (3%) and phospholipid (4%), but these contributed negligibly to the total amount of cholesterol and phospholipid in reconstitution experiments. The NL fraction was found to contain greater than 90% cholesterol, and a smaller percentage of phospholipid (8%).

TLC performed on the NL fraction using hexane-formic acid mobile phase followed by charring and densitometry demonstrated that the majority of NL was composed of cholesterol, with smaller contributions from cholesterol esters, free fatty acids, dipalmitin, tripalmitin, and phospholipids.

Phospholipid composition of PL and DAPL fractions determined by TLC followed by molybdenum blue assay is summarized in Table 2. The results demonstrate a 20- to 25-fold reduction in PG in DAPL samples compared with PL samples. A reduction in the combined content of phosphatidylserine (PS) and PI was similarly noted, although the relative contributions from reduced PI and PS content could not be determined owing to lack of chromatographic resolution using this separation technique. The observed increase in PC content of the APL fraction could be accounted for entirely by the corresponding reduction in APLs.

Fifteen percent polyacrylamide gel electrophoresis was performed on PL, DAPL, NL, HA, and SP-A fractions to demonstrate lack of significant protein content within lipid fractions, and purity of HA and SP-A fractions. Fifty-microgram samples of PL, APL, and NL (via cholesterol content) were analyzed under reducing conditions. Ten-microgram samples of HA and SP-A fractions were analyzed, and compared with a control of 100 µg of whole CLS lipid. No detectable protein was noted in PL, DAPL, or NL fractions by combined silver- Coomassie blue stain. In addition, no cross-contamination between HA and SP-A samples was detectable at the resolution provided by this combined staining technique (10 to 100 ng).

Equilibrium Surface Tension as a Function of Phospholipid and Calcium Concentration in Reconstituted Samples

Static equilibrium surface tension was determined as a function of phospholipid concentration for surfactant samples reconstituted with either PL or DAPL phospholipid fractions. _equil was measured as a function of calcium concentration for DAPL, (DAPL + HA), (DAPL + HA + SP-A), and (DAPL + HA + SP-A + NL) at 1 mg/ml, 0.1 mg/ml, and 0.01 mg/ml bulk phase lipid concentrations, as well as for corresponding PL samples.

Results for DAPL samples are summarized in Table 3. Even low concentrations of added calcium (0.5 mM) caused a significant reduction (5 to 8 dynes/cm) in _equil at all concentrations in samples containing DAPLs without surfactant apoproteins. The addition of HA to DAPLs had two effects on equilibrium surface tension: it caused a reduction in _equil at all bulk concentrations, and it abolished the sensitivity of _equil to calcium seen with isolated DAPLs. Addition of SP-A and NL had no further effect on equilibrium surface tension values among DAPL samples.

Samples containing a complete complement of PL demonstrated a pattern of calcium sensitivity similar to that of DAPL samples (Table 3). Equilibrium surface tensions decreased as bulk concentration increased up to 1 mg/ml, and samples containing 5 mM calcium had _equil values which were uniformly lower than those measured in calcium-free samples.

Addition of HA to PL fractions reduced equilibrium surface tension values as previously reported (19, 20), and reductions were also similar to those observed in DAPL samples (9). Addition of the HA fraction to PL at bulk phase concentrations as low as 0.01 mg/ml lowered _equil to between 22 and 24 dynes/cm, without further dependence on PL or calcium concentration. This contrasts with (DAPL + HA) samples, for which _equil values of 22 to 24 were observed only among samples at  0.1 mg/ml bulk concentration.

Equilibrium surface tension as a function of bulk concentration for whole purified CLS is also summarized in Table 3. Fully reconstituted PL and DAPL surfactant samples behaved nearly identically to whole surfactant with respect to equilibrium surface tensions at 0.01, 0.1, and 1.0 mg/ml bulk concentration in the presence of added calcium ion.

Effect of Addition of Surfactant Apoproteins and NL on the Dynamic Surface Tension Properties of Reconstituted PL and DAPL Preparations

Surface tension versus surface area for DAPL, (DAPL + HA), (DAPL + HA + SP-A), and (DAPL + HA + SP-A + NL) samples at 1.0, 0.1, and 0.01 mg/ml, and for corresponding PL samples, was measured by PBS at f = 1, 20, and 100 cycles/ min, 37° C, and area excursion (A/A_mean = 73%). The minimum radius of 0.34 mm was selected because it allowed for good centering of the bubble on the capillary tube. At larger minimum radii, bubble deformation at low surface tensions resulted in a tendency for the bubble to distort and break off with subsequent rapid development of leakage of material along the capillary. This phenomenon has been previously described by Putz and colleagues (17, 18).

Profiles for the various preparations at a single bulk concentration of 1 mg/ml and at 20 cycles/min are shown in Figures 1a-1d, and native CLS results are shown for comparison in Figure 1e. Samples with no added calcium, and samples containing 5.0 mM CaCl₂ are displayed. Minimum, maximum, and equilibrium surface tension values for samples at 1 mg/ml bulk concentration are also summarized in Table 4.

View larger version (30K): [in this window] [in a new window]   Figure 1.   Dynamic function of samples reconstituted with either PL or DAPL fractions are shown. Recordings were made by pulsating bubble surfactometry at 20 cycles/min, 37° C, and A/Amean = 73% in the absence, and presence of 5 mM exogenous CaCl2 at steady state. Panel a demonstrates that PL and DAPL fractions behave nearly identically, with or without added calcium. After addition of HA fraction (b), both PL and DAPL samples reach minimal surface tensions < 3 dynes/ cm, although their dynamic profiles are quite different. These differences persisted even after addition of SP-A (c), and were more pronounced in samples reconstituted with NL (d ) where minimal surface tensions remained near 20 dynes/cm during dynamic cycling. Profiles for whole CLS (e) are shown for comparison.

Dynamic behavior of DAPL and PL samples (Figure 1a) without added apoprotein components were similar at all lipid concentrations examined, independent of the addition of calcium. Minimal surface tensions for both DAPL (20.5 ± 0.3 dynes/cm) and PL (20.6 ± 0.9 dynes/cm) fractions were significantly greater than those observed in intact surfactant (< 1.0 dynes/cm). During dynamic oscillations at 20 cycles/min, film hysteresis was reduced in both samples compared with intact surfactant, and maximum surface tensions reached during the expansion phase of cycling were significantly greater.

The addition of 3% HA resulted in both (APL + HA) and (PL + HA) preparations reaching minimum surface tensions of less than 3 dynes/cm within 1 min of cyclic dynamic compression, with maximum surface tensions significantly lower than those observed for DAPL and PL alone (Table 4). Both minimum and maximum surface tensions were nearly identical to values observed for whole lung surfactant. However, dynamic behavior of (PL + HA) and (DAPL + HA) preparations were markedly different, with (PL + HA) samples behaving much more like whole surfactant than (DAPL + HA) samples. At 1 mg/ml bulk phase lipid concentration, (PL + HA) samples required a 45 to 55% compression ratio to reach minimal surface tension; at the same bulk concentrations, (DAPL + HA) samples required 80 to 90% film compression to reach minimal surface tension. By comparison, whole CLS reached minimal surface tensions after 25 to 30% film compression (Figure 1e). Addition of 5 mM CaCl₂ had no effect on dynamic interfacial behavior of (PL + HA) samples, whereas hysteresis was increased in (APL + HA) samples. This increase did not restore behavior to that of intact PL + HA samples, however, and did not further affect _max or _min values.

Addition of 5% purified SP-A produced a further increase in film hysteresis in (PL + HA) samples by reducing the amount of film compression required to reach minimal surface tension to 30 to 35%, nearly that of whole surfactant (Figure 1c). Calcium had no observable effect on dynamic interfacial properties of (PL + HA + SP-A) samples. By comparison, addition of SP-A to (DAPL + HA) samples had no observable effect on _min, _max, or film hysteresis, and added calcium had no effect. (DAPL + HA + SP-A) samples were indistinguishable from (DAPL + HA) samples, with both samples demonstrating only a small amount of hysteresis, and requiring between 80 to 90% film compression to achieve surface tensions below 3 dynes/cm.

The addition of NL (8% cholesterol by weight) to reconstituted surfactants produced the most striking difference between PL and DAPL preparations (Figure 1d). On the one hand, (PL + HA + SP-A + NL) preparations exhibited comparable dynamic interfacial behavior to (PL + HA + SP-A) samples devoid of NL. Both reconstituted samples behaved nearly identically to whole surfactant (Figure 1e) at all bulk concentrations and oscillation frequencies considered, displaying similar _min, _max, and film hysteresis. By contrast, addition of NL to (DAPL + HA + SP-A) samples resulted in an inability to reach surface tensions below 20 dynes/cm during film compression. Dynamic profiles (Figure 1d) of (DAPL + HA + SP-A + NL) samples showed a marked increase in minimum surface tensions, a virtual absence of hysteresis, and loss of film elastance, defined as the slope d/dA during cyclic compression. The addition of 5 mM calcium produced a small increase in film hysteresis, but failed to restore any semblance of normal dynamic interfacial behavior in these samples.

Effect on Dynamic Surface Tension Properties of Addition of Purified PG to DAPL Samples

PG was isolated and purified from CLS by preparative TLC, followed by organic phase extraction. Sample purity was documented by subsequent analytical TLC. Purified PG was added to (DAPL + HA + SP-A) and (DAPL + HA + SP-A + NL) samples at 5% and 10% (weight/weight total PL component), and static and dynamic interfacial behavior were assessed for samples at 1.0, 0.1, and 0.01 mg/ml bulk phase concentration. The results for 1 mg/ml samples at f = 20 and 100 cycles/min supplemented with 5% PG are shown in Figure 2a. Addition of 5% PG to (APL + HA + SP-A) resulted in nearly complete normalization of dynamic surface tension profiles relative to intact native CLS, and to samples reconstituted with PL fractions which contained a normal complement of APLs. At 1 mg/ml, significant hysteresis was restored after PG supplementation, and minimal surface tensions of near 0 dynes/cm were observed after 30 to 35% compression in (DAPL + HA + SP-A + 5%PG) samples, in contrast to 80 to 90% compression observed in (DAPL + HA + SP-A) samples. Dynamic interfacial properties were not affected by addition of calcium (data not shown), similar to behavior observed in preparations reconstituted with PL. Addition of 10% PG (Figure 2c) had no further effect on dynamic interfacial properties, and did not alter the calcium sensitivity of these reconstituted samples.

View larger version (20K): [in this window] [in a new window]   Figure 2.   Dynamic function of DAPL samples reconstituted with either 5% or 10% purified calf lung PG fraction are shown. Recordings were made at 20 and 100 cycles/min, 37° C, and A/ Amean = 73% at steady stae in samples containing added 5 mM CaCl2. Panel a demonstrates that 5% PG restores nearly normal function to DAPL samples reconstituted with HA and SP-A fractions at both frequencies examined. Five percent added PG had a similar effect on samples containing (DAPL + HA + SP-A + NL) (b). Increasing the PG content to 10% had little additional beneficial effect (c).

The effect of adding 5% PG to DAPL samples containing (HA + SP-A + NL) was more dramatic (Figure 2b), given the pronounced dysfunction which had been observed. PG restored near normal function to (DAPL + HA + SP-A + NL) samples, with restoration of minimal surface tensions to near 0 dynes/cm, and marked surface tension surface area hysteresis during dynamic cycling at 1 mg/ml bulk phase concentration. In samples with no added calcium, 50 to 55% film compression was required to reach _min, but addition of 5 mM CaCl₂ reproducibly reduced the compression ratio required to reach minimal surface tensions to 30 to 35%. Addition of 10% PG further decreased the compression ratio required to reach _min to 40 to 45% in samples with no added calcium, and 30 to 35% in the presence of 5 mM CaCl₂.

PC, isolated and purified from the PL fraction of calf lung using TLC in an identical fashion to PG, was also used to supplement DAPL samples. In contrast to results observed after addition of purified PG, addition of 5% and 10% PC fraction to (DAPL + HA), (DAPL + HA + SP-A), and (DAPL + HA + SP-A + NL) had no observable effect (either beneficial or inhibitory) on dynamic interfacial properties of reconstituted surfactant (profiles not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The results of the present study suggest the following: (1) Changes in the APL composition of surfactant preparations have no effect on equilibrium surface tension, and confer little if any change in sensitivity of _equil to the effects of added calcium; (2) APLs are necessary for PLs and HAs to interact in such a manner that low surface tensions can be achieved rapidly during the initial phases of film compression; (3) Calcium may promote interactions between HA components and zwitterionic phospholipids in the absence of APLs, but such effects are minimal in the presence of APLs under normal physiological conditions; (4) APLs are important for the synergistic behavior of SP-A and HA in promoting hysteresis, thereby reducing the amount of film compression required to reach minimal surface tensionthis effect is, in part, calcium-dependent; (5) APLs prevent the destabilizing effect of cholesterol on surface films composed of purified phospholipids and surfactant apoproteins; and (6) PG alone is able to restore appropriate interfacial behavior to a variety of surfactant preparations, including those containing NLs, without requiring PI, implying that either APL may be sufficient to impart normal function without regard to the molecular structure of the specific phospholipid head group.

These findings appear to confirm a role for APLs previously suggested by Cochrane and others based on theoretical grounds; that PG and PI promote interactions between SP-B and surfactant lipid components (5, 21). It has been argued that such an interaction might occur based on the potential ability of negatively charged APLs to interact with positively charged side groups of arginine and lysine residues near the carboxyl terminus of SP-B. In support of this hypothesis, it has been shown that hydrophobic extracts of surfactant, containing lipid and SP-B, function less well in an alkali environment than at pH values approaching the pKa's of the basic amino acids (22, 23). In addition, synthetic peptide homologues of SP-B have been shown to adsorb rapidly, and reach low minimal surface tensions during dynamic compression, only when basic amino acid residues are included in the sequence (24). In contrast, constructs containing acidic amino acids fail to impart normal interfacial behavior to lipid components. Taken together, our current observations, and those previously reported, support the notion that charge-related lipid-protein interactions are important in determining surfactant interfacial behavior, specifically at steady state during dynamic compression and expansion.

Our findings also suggest that APLs may facilitate interactions between SP-A and HA proteins. Samples reconstituted with PL fractions (which contain APLs) demonstrated augmented adsorption, and facilitated surface tension reduction during the initial phases of film compression compared with samples reconstituted with DAPL fractions. However, our results fail to support a specific role for APLs through direct lipid-lipid interactions, as was originally hypothesized, because static and dynamic interfacial properties of isolated PL and APL fractions were nearly identical.

These findings are largely consistent with those of Wang and coworkers who examined the interfacial properties of surfactant preparations reconstituted with either intact surfactant phospholipids or PL fractions devoid of APLs, and HA components (9). They observed similar static equilibrium surface tensions, respreading indices, and maximal and minimal surface tensions during dynamic oscillation among PL and DAPL samples, independent of whether HAs were present. These investigators also demonstrated that APL-deficient fractions restored static recoil pressures in surfactant-deficient isolated lungs, similarly to samples containing normal PL components. Based on these findings, they concluded that APLs play no significant role in determining the biophysical properties of lung surfactant.

The results of the present study suggest a very different conclusion: that APLs do play a critical role in surfactant function, despite results which do not differ substantially from those of Wang and coworkers. Our observations look more extensively at this issue, and place the former observations in a different context, however, suggesting that the crucial role of these APL components is manifest dynamically, and not evident from consideration of equilibrium, maximal, and minimal surface tensions alone.

Like Wang, we observed that static equilibrium surface tensions in both PL and DAPL samples were similar at 0.01 mg/ml bulk concentrations, and that the addition of HA significantly lowered _equil in both samples. Minimal and maximal surface tensions of PL and DAPL, and (PL + HA) and (DAPL + HA) samples measured during dynamic oscillations were also observed to be similar.

Despite similar values for _min and _max in (PL + HA) and (DAPL + HA) samples, dynamic behavior was quite different. In the absence of APLs, hysteresis during dynamic cycling was markedly reduced at 1 mg/ml bulk concentration, and the degree of film compression required to achieve minimal surface tensions was significantly greater than in (PL + HA) samples. Hysteresis was slightly increased in (DAPL + HA) samples after addition of calcium to the bulk phase, but film behavior remained markedly abnormal compared with (PL + HA) samples. These data strongly suggest that APLs do play an important role in determining the interaction between HAs and the surfactant phospholipid component during steady-state dynamic behavior.

The addition of 5% purified PG restored hysteresis, and the degree of compression required to reach minimum surface tension of (DAPL + HA + SP-A) samples to that of control (PL + HA + SP-A) samples. Addition of equivalent amounts of purified PC fraction had no beneficial effect on function, however, demonstrating that the observed abnormality in dynamic interfacial behavior was principally the result of a deficiency in anionic phospholipids. These findings further argue that PI is not specifically required for normalization of interactions between phospholipid and HA components, because only the PG fraction was restored in these studies, whereas both PG and PI were deficient in the DAPL fraction as a result of chromatographic separation (9).

The ability of APLs to interact with HAs to promote film stabilization during compression could have profound physiological significance that can go unappreciated during standard surfactometry measurements performed using large surface area excursions, or quasi-static pressure-volume measurements performed over large volume changes to total lung capacity. Schurch and coworkers have argued that the degree of compression required to achieve minimal surface tension is a useful and sensitive index of dynamic functional behavior, with increases in compression ratio indicating increasing degrees of dysfunction (4). The authors suggest that normal surfactant should reach minimal surface tensions near 0 dyne/cm after 15 to 20% area compression, equivalent to that which occurs during tidal breathing (21). Computer simulation results from our laboratory also argue that changes in the degree of film compression required to reach minimal surface tension may be important physiologically. Using a model of dynamic interfacial behavior which incorporates Langmuir kinetics, liquid-expanded to liquid-condensed to solid-phase change behavior, and squeeze-out dynamics, we have shown that surfactant displaying behavior similar to (DAPL + HA) samples by PBS will not reach _min < 5 dynes/cm during area changes approximating those of tidal breathing (25). The inability to do so could result in alveolar instability at lung volumes near FRC, and produce alveolar collapse and macroscopic atelectasis, despite retaining the ability to reduce surface tensions to near zero after large surface area excursions. Thus, the failure to observe differences in _equil, _min, and _max between (PL + HA) and (DAPL + HA) samples may not indicate equivalent physiological behavior.

The presence of 8 to 10% cholesterol by weight in native surfactant, and its profound adverse effects on surfactant function observed in DAPL samples, also suggests a potentially important physiological role for APLs not previously anticipated. Surface films with dynamic behavior similar to that observed in NL-containing DAPL preparations have little ability to change surface tension as a function of surface area, and the elevated minimal surface tensions during compression would not be expected to stabilize alveolar units at low lung volumes (26). Although the basis for this effect of APLs on NL-containing samples is unknown, these observations may be of significance in the pathophysiology of acute respiratory distress syndrome (ARDS), because ARDS surfactant appears to be deficient in APLs, whereas cholesterol content is increased (27). The combined effects of decreased APL content and increased cholesterol could result in significant surfactant film destabilization, independent of surfactant apoprotein content.

Additional studies are needed to further address the role of APLs in vivo, and the physiological significance of decreased levels of APLs in human pathophysiology, such as in ARDS. Nevertheless, the results of this in vitro study clearly demonstrate an important role for APLs in determining the dynamic properties of lung surfactant, particularly in the presence of NL components. They further suggest that this function results from a synergistic interaction with HA components.