INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Commercially available surfactant preparations have been investigated repeatedly with regard to their in vitro properties, and conclusions have been drawn from these in vitro data to their in vivo physiologic effects. These in vitro properties are usually assessed with reference to static surface adsorption and dynamic changes in surface tension during cyclic film compression using dynamic systems such as the pulsating bubble surfactometer (PBS). However, surfactant concentrations, usually expressed as milligrams phospholipid (PL) per milliliter, and calcium concentrations in the sample buffers differ widely between studies, ranging from 0.007 to 35 mg/ml for PL and from zero to 6 mmol/L for calcium concentration. Moreover, a variety of different surfactant preparations was used (1). These variations in at least three experimental parameters render it almost impossible to compare their results. Hence, we set out to determine the effect of surfactant PL concentration on the in vitro properties of four commercially available surfactant preparations at physiologic calcium concentrations and to compare these with the physical properties of native bovine (NBS) and porcine (NPS) lung lavage surfactant as biologic standards. Additionally, we systematically determined the effect of calcium on surfactant function in vitro, as some experimental protocols do not fit with the in vivo concentrations (1, 4, 5, 10, 11). As the parameters of surface tension function of a surfactant depend on apoproteins (2, 11), we (1) used SP-A containing native surfactants as a physiologic standard and (2) measured the concentrations of the hydrophobic surfactant proteins SP-B and SP-C in the different formulations. Moreover, although it is known that phosphatidylcholine (PC) molecular species composition modifies surface tension functions of surfactant (13), there is little information on the detailed molecular composition of therapeutic surfactants. Hence, we (3) determined the molecular composition of PC species as the major surfactant PL and correlated our data with their surface tension function. We were particularly interested in determining to what extent the physical properties required for a "good" surfactant, namely, an equilibrium surface tension after 10 s adsorption (_ads) of 25 to 28 mN/m (14), and a minimum surface tension (_min) of < 5 mN/m during cyclic film compression (6, 14), are dependent on PL concentration, and to what extent surface tension function of surfactants can be correlated with their respective content in surfactant apoproteins and their different PC compositions.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Commercially Available Therapeutic Surfactants

Four commercially available surfactant preparations were used: Alveofact, Curosurf, Exosurf, and Survanta. Alveofact is a surfactant isolated from bovine lung lavage and then prepared by lipid extraction and precipitation steps, which essentially remove hydrophilic proteins, including surfactant apoprotein A (SP-A) (15, 16). Curosurf is a lipid extract surfactant from porcine lungs. However, it is not prepared from lung lavage but from whole minced porcine lung tissue. Survanta is produced from minced bovine lung extracts with added dipalmitoylphosphatidylcholine (DPPC), palmitic acid, and triacylglycerol. The respective concentrations of total PC are 82, 79, and 74% of total PL and 72, 78, and 62% of total mass for Alveofact, Curosurf, and Survanta, respectively (16). Exosurf is a purely synthetic and protein-free surfactant containing DPPC, hexadecanol, and tyloxapol in a relation of 13.5:1.5:1, containing 84% DPPC in relation to mass as the only PL (15). The three surfactants being derived from biologic sources contain SP-B and SP-C, but no SP-A, which is removed during lipid extraction (15, 16). All surfactants were used prior to expiration date for functional analyses and stored as recommended by the manufacturer. For biochemical analyses sample aliquots were stored at 80° C.

Native Lung Lavage Surfactants

Bovine and porcine lung surfactants were obtained from lungs of freshly slaughtered cattle and pigs and isolated by density gradient centrifugation of blood-free bronchoalveolar lavage fluid (BALF) without organic extraction to preserve the SP-A in the preparations (17). Lungs were transported within 30 min to the laboratory, and lavage was subsequently performed with 154 mmol/L saline at room temperature. Pig lungs were lavaged sequentially with 3 × 1 L, cattle lungs with 3 × 5 L saline. The BALF fractions were pooled and the liquid was centrifuged in 250-ml aliquots at 4° C and 270 × g for 15 min to remove cells. The supernatant was then centrifuged at 27,000 × g and 4° C for 2.5 h. From the 27,000 × g pellet, surfactant was isolated as previously described (17). In brief, the pellet was resuspended in 3.75 ml 154 mmol/L saline to a concentration of 15 to 20 mg PL/ml. To these 3.75 ml of suspension, 1.25 ml 64% NaBr in 154 mmol/L saline were added. These suspensions were mixed and sequentially overlayered with 5 ml 13% NaBr in 154 mM saline and 1.5 ml 154 mmol/L saline. Samples were centrifuged at 114,000 × g for 100 min and the surfactant band between 13% NaBr and saline was then aspirated with a 5-ml disposable syringe. The band was transferred to a 25-ml Corex centrifugation tube, diluted with double-distilled water to a final volume of 25 ml, and centrifuged at 27,000 × g at 4° C for 60 min. The supernatant was discarded, the pellet was resuspended in 6 to 10 ml 154 mmol/L saline to which 1.5 mmol/L calcium chloride had been added, and three 10-µl aliquots were taken for the determination of phospholipid phosphorus concentration in the preparation. PL concentration of such stock suspensions was 15 to 20 mg/ml. Surfactant preparations were stored at 80° C until measurement and then diluted to the desired concentrations.

Preparation of Surfactant Samples for Surface Tension Measurement

Commercial and native surfactants were adjusted to an initial PL concentration of 6 and 8 mg/ml (8.0 and 10.7 µmol/ml), respectively, and then geometrically diluted in 1.5 ml Eppendorf tubes so that measurements could be performed at 8.0, 6.0, 4.0, 3.0, 2.0, 1.5, 1.0, 0.75, 0.5, 0.375, 0.25, 0.125, and 0.063 mg/ml. To investigate the effect of surfactant concentration on surface tension function in different preparations, measurements were performed at a final concentration of 1.5 mmol/L calcium chloride. For determining the effects of calcium, resuspension of surfactant was performed with 154 mmol/L saline in the absence of calcium, and the concentration of the latter then adjusted with calcium chloride to 0.0, 0.375, 0.75, 1.5, 3.0, or 6.0 mmol/L added calcium at the desired final PL concentrations. Calcium tightly bound to surfactant preparations and not removed by purification steps was neglected.

Surface Tension Measurements

Prior to the measurements, surfactant samples were vortexed for 30 s and then sonicated for 60 s in a sonication bath to ensure complete homogenization of the samples. Although sonication at high energies may affect surface tension function of surfactant (18), pilot studies had shown that this was not true for the procedure used in this study, probably because of the brief sonication period and the use of the relatively soft Eppendorf tubes, which do not transmit much sonication energy to the sample. Surface tension was determined in the pulsating bubble surfactometer (PBS; Electronetics, Amherst, NY) as described by Enhorning (19). Briefly, 36 µl of the surfactant suspension were instilled into the sample chamber of the PBS at 37° C. A bubble communicating with ambient air was created in the surfactant suspension and surfactant allowed to adsorb to the air/liquid interface for 10 s. After this time the bubble was pulsated at 20 oscillations per min between a minimum radius of 0.4 mm and a maximum radius of 0.55 mm. The pressure across the bubble was measured by a pressure transducer and the surface tension calculated using the LaPlace equation. The surface tension after 10 s adsorption (_ads) and the minimum (_min) and maximum (_max) surface tensions after 1, 3, 9, 30, and 100 cycles were determined. All measurements were performed five times and the mean and standard deviation (SD) calculated.

Analysis of Phospholipids and Hydrophobic Surfactant Proteins

For PL and protein analysis the surfactant preparations were extracted according to Bligh and Dyer (20) and the PL phosphorus quantified from sample aliquots as described by Bartlett (21) 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). Total PC and sphingomyelin were isolated from 500-nmol PL aliquots using 100 mg Varian Bondelut NH₂ disposable cartridges (Varian, Hamburg, Germany) (22). Individual PC molecular species and sphingomyelin were then resolved by reverse-phase HPLC on a 4.6 × 250 mm Spherimage ODS II column (Schambeck, Bad Godesberg, Germany) and eluted PC species quantified by postcolumn fluorescence derivative formation in the presence of 1,6-diphenyl-1,3,5-hexatriene (22). For the determination of total PC as a percentage of total PL, the samples were spiked with 50 nmol dimyristoyl-PC (PC14:0/14:0) as a standard prior to PC isolation. Total PC concentration in PL was then calculated as the sum of the PC peaks in relation to the standard (22). SP-B and SP-C were analyzed from 0.5 to 1.2 µmol PL aliquots using a Sephadex LH-60 column with UV detection at 228 nm as described by van Eijk and colleagues (23). Quantitation of proteins was performed using the fluorimetric assay by Böhlen and colleagues (24).

Electron Microscopy

Transmission electron microscopy was performed from all native and commercial surfactant preparations. As Exosurf, a nearly pure formulation of DPPC, could not successfully be fixed in the absence of tannic acid, all surfactant preparations were fixed for comparative analyses in the presence of tannic acid (25) as follows: Surfactant suspensions were centrifuged down at 60,000 × g for 1 h, and the pellets were fixed in a solution containing 3% glutaraldehyde and 1% tannic acid in 0.1 mol/L sodium cacodylate-HCl buffer at pH 7.2 and 4° C 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 about 60 nm thick were cut with a diamond knife and examined in a Zeiss EM10 electron microscope (Zeiss, Oberkochen, Germany). A point-counting method was used to determine the relative amounts of the different types of precipitates in the surfactant preparations (26). Using a square grid (mesh width, 10 mm) and prints of electron micrographs (final magnification: 63,000:1) the frequencies of the following types were quantified: (1) lamellar bodies, (2) tubular myelin, (3) multilamellar vesicular aggregations (less than 10 lipid layers), (4) crystalloid aggregations (10 layers or more). Other material such as amorphous substances and obliquely sectioned lipid layers were excluded from quantification. According to Weibel (26) the percentages of the point-countings represent the volume fractions of the four types.

Statistics

One-way analyses of variance were calculated using GraphPad Instat Version 1.11a (GraphPad Software, San Diego, CA), and results corrected using the Bonferroni method for multiple group comparisons.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Influence of PL Concentration on Static and Dynamic Surface Tension Functions

Physical properties required for a "good" surfactant (see above) at physiologic calcium concentrations (1.5 mmol/L) (27) were achieved with lung lavage surfactant preparations from both bovine and porcine lungs even at a PL concentration of 0.25 mg/ml (Figure 1). In contrast, with Alveofact and Curosurf, PL concentrations three to 12 times higher (0.75 and 3 mg/ml, respectively) were required to achieve a _ads of < 29 mN/m, and six to 12 times higher (3 and 1.5 mg/ml, respectively) to achieve a _min of < 5 mN/m (Figure 1, top and middle panels). Interestingly, Alveofact and Curosurf showed distinct plateaus, i.e., _min stayed at around 22 to 24 mN/m below a critical PL concentration of 3 and 1.5 mg/ml, respectively, and reached values below 5 mN/m only with PL concentrations above these values. For Survanta and Exosurf, the lowest _ads measured was 36 (SD, 2.9) and 49 (SD, 0.5) mN/m, respectively, even at a PL concentration of 8 mg/ml (Figure 1, top panel). Also, neither surfactant reached _min values < 5 mN/m, although values close to this (8.3 mN/m) were achieved with Survanta even at a PL concentration of 0.375 mg/ml. Exosurf showed a peculiar behavior: its _min fell to 22 mN/m at a PL concentration of 0.25 mg/ml, but reproducibly increased again with increasing PL concentration.

View larger version (27K): [in this window] [in a new window]   Figure 1.   Influence of surfactant concentration on surface tension function. Native (bovine, porcine) and commercial (Alveofact, Curosurf, Survanta, Exosurf) surfactant preparations were suspended in buffer to give the desired PL concentrations in 154 mmol/L saline + 1.5 mmol/L calcium chloride. Static adsorption (ads; top panel ) was measured after 10 s, and minimum (min; middle panel ) and maximum (max; bottom panel ) surface tension after 100 cycles in the pulsating bubble surfactometer (PBS) at 20 cycles/min and 37° C. Data are shown as means and standard deviations (error bars) from five individual measurements.

The data for _max were similar (Figure 1, bottom panel): with both native bovine and porcine lavage surfactant, _max fell to < 35 mN/m at a PL concentration of 0.25 mN/m, whereas considerably higher concentrations (1.0 and 4.0 mg/ ml) were required to achieve this with Alveofact and Curosurf. Interestingly, the minimal concentrations required to achieve _ads values below 29 mN/m and _max values below 35 mN/m were lower with Alveofact than with Curosurf (Figure 1, top and bottom panels), whereas the minimal concentration to achieve _min values below 5 mN/m were lower with Curosurf (Figure 1, middle panel ). Survanta and Exosurf showed _max values of more than 47 mN/m at all concentrations measured (Figure 1, bottom panel ).

To investigate if the better surface tension function of the native surfactants compared with Curosurf and Alveofact was related to the lipid extraction process with removal of hydrophilic components, which the latter undergo, comparative analyses of surface tension function were performed with lipid extracts of native bovine lung lavage surfactant. These showed that values required for a "good" surfactant were now only achieved at concentrations of 2 mg/ml (_ads, 24.6 mN/m [SD, 0.5]); _min, 0.9 mN/m [SD, 0.1]; _max 33.9 mN/m [SD, 0.7]), whereas at 1 mg/ml, the corresponding values were 25.8 (1.0), 21.4 (0.5), and 31.3 (0.5) mN/m. The time required to achieve _min values < 5 mN/m during repeated cycling also varied considerably between preparations. We investigated this at the minimal surfactant concentrations necessary to achieve _min values < 5 mN/m within 5 min of cycling. With both native lung lavage surfactants, _min fell to < 5 mN/m already after the first compression, whereas with Curosurf and Alveofact, nine cycles were required (Figure 2). As neither Survanta nor Exosurf reached _min values < 5 mN/m, the number of cycles required to reach such values could not be determined with these preparations.

View larger version (21K): [in this window] [in a new window]   Figure 2.   Number of cycles necessary to produce min values < 5 mN/m in the PBS. Suspensions of native bovine surfactant, native porcine surfactant, Alveofact, and Curosurf were diluted in 154 mmol/L saline with 1.5 mmol/L calcium chloride to those minimal PL concentrations that resulted in min values below 5 mN/m (see Figure 1). Note logarithmic scale on x-axis. Data are shown as means and standard deviations (error bars) from 5 individual measurements.

Influence of Calcium Concentration

Knowing that at a PL concentration of < 1.5 mg/ml and a calcium concentration of 1.5 mmol/L, _min values < 5 mN/m could not be achieved with any commercially available surfactant, we determined to what extent this could be modified by changes in calcium concentration. At a PL concentration of 3 mg/ml, calcium had little effect, i.e., _min remained < 5 mN/m with Alveofact and Curosurf and was around 30 mN/m with Exosurf (data not shown). Only Survanta showed a decrease in _min, from 6.3 (SD, 0.5) to 3.1 (SD, 0.3) mN/m when calcium concentration was raised to 6 mmol/L (p < 0.05). At a PL concentration of 1 mg/ml, however, surface tension function of both Alveofact and Curosurf was markedly improved by increasing calcium concentration. Hence, we investigated in more detail the influence of calcium on these lipid extract surfactants in comparison with native surfactants. As demonstrated in Figure 3, there was an abrupt fall in _min from > 20 to < 5 mN/m when calcium concentration was increased from 2.5 to 3 mmol/L for Alveofact, and from 3 to 4 mmol/L for Curosurf.

View larger version (18K): [in this window] [in a new window]   Figure 3.   Influence of calcium concentration on minimal surface tension (min). min was determined for Alveofact, Curosurf, and porcine lung lavage surfactant at a PL concentration of 1 mg/ml in the PBS after 5 min of cycling. Data are shown as means and standard deviations (error bars) from five individual measurements. With Alveofact and Curosurf, there is a brisk decrease in min to < 5 mN/m once calcium concentration is raised to above 2.5 and 3.0 mmol/L, respectively.

In contrast, with native porcine lung lavage surfactant, _min stayed at < 5 mN/m independent of calcium concentration in the suspension buffer. Because calcium had little effect on the comparatively high surface tension of Survanta and Exosurf at 3 mg/ml, these two surfactants were not included in the measurements performed at a PL concentration of 1 mg/ml.

Biochemical Composition: Hydrophobic Proteins and PC Molecular Species

Native surfactants contained 1.5% SP-B and 2.8 to 4.5% SP-C per µmol PL (Table 1 and Figure 4). Commercial surfactants without supplementation by exogenous DPPC (Alveofact, Curosurf) contained only one half to one third of these proteins as compared with their progenitors NBS and NPS, whereas Survanta contained only 1/10 of the SP-B, but 1/2 of the SP-C found in NBS. PC was the only (Exosurf) or major (other surfactants) phospholipid in all therapeutic surfactants (Table 2). Bovine and porcine native surfactants contained 85.5 (SD, 4.7) and 84.7 (SD, 3.7) mol% PC in relation to total PL. Similarly, Alveofact, as a lung lavage surfactant, contained 83.7 (SD, 3.9) mol% PC. Comparable values were achieved with Survanta (79.1 [SD, 8.0]%), whereas PC content was lower for Curosurf (73.0 [SD, 2.7]%; p < 0.01). Sphingomyelin as a membrane component was present in low concentration in all natural surfactants investigated. Notably, Curosurf as a minced lung tissue extract contained significantly more sphingomyelin than did the corresponding lavage surfactant (Table 2), whereas Exosurf as a preparation of pure DPPC contained no sphingomyelin (Figure 5).

View larger version (27K): [in this window] [in a new window]   Figure 4.   SP-B and SP-C in native and commercial surfactant preparations. Surfactant material was extracted with chloroform:methanol, 1 µmol PL material applied onto a 18 × 1.6 cm Sephadex LH-60 column and SP-B and SP-C separated with dichloromethane:methanol:0.1 N hydrochloric acid (30:65:5; vol/vol). Representative separations are shown from bovine native surfactant, Alveofact, Survanta, porcine native surfactant, and Curosurf.

View larger version (28K): [in this window] [in a new window]   Figure 5.   Composition of phosphatidylcholine molecular species in native and commercial surfactant preparations. Phosphatidylcholine (PC) and sphingomyelin (SPH) were isolated from 500 nmol phospholipid and separated with HPLC as described in METHODS. Representative HPLC traces are shown. Abbreviations: 1: sphingomyelin; 2: palmitoylmyristoyl-PC; 3: palmitoylpalmitoleoyl-PC; 4: palmitoylarachidonoyl-PC; 5: palmitoyllinoleyl-PC; 6: dipalmitoyl-PC; 7: palmitoyloleoyl-PC; 8: stearoylarachidonoyl-PC; 9: stearoyllinoleyl-PC.

As demonstrated in Table 3 and Figure 5, there were remarkable differences in the composition of PC molecular species of both native and commercial surfactant preparations. Two of the bovine surfactants contained 40% DPPC (native, 40.2% [SD, 2.8]; Alveofact, 39.4% [SD, 2.1]), compared with more than 50% for the porcine surfactants (native, 59.5% [SD, 3.2]; Curosurf, 50.2% [SD, 2.2]). Survanta, as a bovine surfactant with exogenous DPPC added, contained 74.9% (SD, 5.0) DPPC and Exosurf 99.5% (SD, 0.8). Palmitoylmyristoyl-PC (PC, 16:0/14:0) and palmitoylpalmitoleoyl-PC (PC 16:0/16:1), the two other PC species, relatively unique to mammalian surfactant, did not differ significantly between native surfactants, Alveofact, and Curosurf. However, they were decreased in Survanta, because of the addition of exogenous DPPC, and absent in Exosurf. On the other hand, palmitoyloleoyl-PC (PC 16:0/18:1; POPC) was significantly increased in bovine compared with porcine surfactant and in the commercial compared with the corresponding native surfactants, possibly because of the manufacturing processes of extraction and precipitation (Table 3 and Figure 5). Again, Survanta contained much less POPC than the other surfactant preparations. All other PC species, particularly the highly unsaturated ones such as palmitoylarachidonoyl-PC (PC 16:0/20:4) and stearoylarachidonoyl-PC (PC 18:0/20:4) were only minor components in any surfactant preparation.

Electron Microscopy

Native surfactant from porcine and bovine lungs showed a similar ultrastructure, mainly consisting of multilamellar and vesicular material (Figures 6A and 6C). Some tubular myelin was found in both preparations, but it was more prominent in bovine surfactant (Figure 6A). Native surfactant before and after sonication showed only marginal alterations of its ultrastructure (Figures 6B and 6D). Commercial surfactant preparations differed from native surfactants in the general absence of tubular myelin and in the structure of the lamellar and vesicular material (Figures 7 and 8). As demonstrated, Alveofact and Curosurf mainly consisted of lamellar and vesicular structures (Figures 7A and 7D and Figure 8). In contrast, Exosurf showed crystalline formations only and may be interpreted as DPPC crystals (Figures 7B and 8). Such structures were only a minor compound in native (Figure 6) and lipid extract surfactants (Figures 7 and 8). Survanta, a mixture of lipid extract surfactant supplemented with DPPC, showed vesicular structures (Figures 7C and 8) similar to those found in Alveofact or Curosurf, and also crystal-like structures similar to those found in Exosurf. This finding suggests that DPPC molecules were not or incompletely integrated into the surfactant structures.

View larger version (207K): [in this window] [in a new window]   Figure 6.   Thin-section electron microscopy of native surfactant before and after sonication. Surfactant material was fixed in glutaraldehyde/tannic acid and postfixed in osmium tetroxide. Both bovine (A, B) and porcine (C, D) surfactant mainly consist of multilamellar structures and some tubular myelin (lower left corner in A). Only minor structural alterations are found after sonication for 60 s (B, D) in comparison with preparations without sonication (A, C ). A-D, original magnification: ×50,000; insets ×3,800.

View larger version (206K): [in this window] [in a new window]   Figure 7.   Ultrastructure of the commercial surfactant preparations Alveofact, Exosurf, Survanta, and Curosurf. Although Alveofact (A) and Curosurf (D) predominantly show oligolamellar and vesicular formations, Exosurf (B) consists of straight crystalline structures, which probably represent DPPC. As Survanta (C ) is a mixture of bovine lipid extract surfactant with DPPC added, it shows both the vesicular/oligolamellar structures typical for Alveofact and Curosurf as well as the crystalline structures characteristic of Exosurf. A-D, original magnification: ×50,000; insets, original magnification: ×3,800.

View larger version (38K): [in this window] [in a new window]   Figure 8.   Ultrastructural composition of the commercial versus native surfactants. Transmission electron microscopy was performed as described in METHODS, and ultrastructure of surfactant preparations was determined by point-counting of photographs. Data are means of 136 to 321 point-countings.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study has for the first time defined parameters of surface tension function under standardized experimental conditions in four commercial surfactant preparations used clinically, comparing these with each other and with native surfactants in vitro. Moreover, it quantified the major biochemical components of surfactants such as SP-B and SP-C and individual molecular species of PC as functionally distinct components and defined the effect of nonphysiologically high calcium concentrations in the suspending medium as a potential source of misinterpreting surfactant activity in vitro.

Pharmacodynamics of Surfactants

As with other therapeutic agents, surfactants can be characterized by their own maximal effects (intrinsic activity) and by their concentration-effect curves (potency) in relation to a standard (28), in this case native surfactants. Potency usually describes the concentration of a therapeutic agent required to achieve half its maximal effect (28). In this study, a 50% effect was impossible to define, as _min of both the native (NBS, NPS) and the corresponding nonmodified lipid extract surfactants (Alveofact, Curosurf) showed no concentration-dependent continuum in the PBS, but two distinct, concentration-dependent plateaus (at ~ 20 and < 5 mN/m). Thus, we could define only the concentrations above which _min fell to the lower plateau. Interestingly, the minimal concentration necessary to achieve _min values < 5 mN/m was inversely correlated with the concentration of DPPC in the SP-B/C containing "natural" surfactants Alveofact and Curosurf. As DPPC is the essential surfactant molecule to lower surface tension at the air-liquid interface (29), this correlation is not surprising, and it confirms previous measurements by Gail and colleagues (30), who demonstrated that the amount of DPPC in alveolar surfactant closely correlates with the surface area of lungs in vivo. Hence, in the presence of sufficient concentrations of SP-B/C as promotors of surface activity under dynamic conditions, and possibly other adsorption-promoting surfactant components like POPC, palmitoylmyristoyl-PC (PC 16:0/14:0) and palmitoylpalmitoleoyl-PC (PC 16:0/16:1) (13), the concentration of DPPC in a therapeutic surfactant is critical for its potency, both in a standardized in vitro protocol and, possibly, also in vivo.

Surface Tension Function in Relation to Surfactant Proteins

Measurement of surface tension function strongly depends on the technique applied. For example, Exosurf easily lowers surface tension to below 5 mN/m in the Wilhelmy balance, but not under the dynamic conditions of the pulsating bubble surfactometer (31). Our data fit with those of other investigators (11), which showed poor surface tension function at low concentrations or absence of SP-B/C. Whereas native surfactant reached _ads values of < 29 mN/m and _min values of < 5 mN/m at concentrations of 0.25 mg/ml, neither therapeutic surfactant was able to achieve such values at these concentrations. Native surfactant contains specific proteins, particularly SP-A, SP-B, and SP-C, and is believed to be transformed to tubular myelin under the influence of SP-A and -B, whereas SP-B and -C promote surface adsorption (11, 32). None of the commercial surfactants contains SP-A since all are lipid extracts (15, 16). Although lung function is still maintained with a surfactant devoid of SP-A at sufficient concentrations (33), our data together with those of Ingenito and colleagues (11) clearly show that in the absence of SP-A surfactant potency is impaired. In addition to this, therapeutic surfactants, including Alveofact and Curosurf, contain less SP-B/C than NBS and NPS. The importance of SP-B and SP-C in promoting the rapid adsorption of surfactant PL to the air-liquid interface has been clearly demonstrated (34) and the deleterious effects of SP-A, -B, and -C deficiency on surface tension function have also been shown for lipid extracts from NBS and for conductive airway surfactant from pigs, which does not contain SP-B or -C (17, 35). Hence, we conclude that in commercial surfactants the reduced SP-B/C content, together with a complete absence of SP-A, is an important biochemical parameter of impaired surface tension function. Whether the dosage of therapeutic surfactants containing SP-A or nonimmunogenic SP-A analogs can be lower under clinical conditions is still somewhat hypothetical (36), but it is of clinical interest because of the high costs of surfactant treatment in adults.

Surface Tension Function in Relation to PC Molecular Species

Although the lower potencies of Curosurf and Alveofact compared with native surfactants can be explained by their differences in surfactant protein content, the higher potency of Curosurf over Alveofact in reaching _min values < 5 mN/m, and the better static (_ads) and dynamic (_max) adsorption properties of Alveofact over Curosurf (Figure 1, top and bottom panels), are not associated with such differences. However, PC molecular species composition strongly influences surface tension function of surfactants (13), which is distinctly different in Curosurf when compared with that in Alveofact. Bovine surfactant contains less DPPC and more POPC than does porcine (17) and human (37) surfactant. Although DPPC is responsible for lowering surface tension at the air-liquid interface at end-expiration (32), POPC accelerates the adsorption of surfactant to the interface (13). Consistent with these experimental findings, Curosurf, containing more DPPC than Alveofact, reached _min values < 5 mN/m at lower concentrations than did Alveofact (1.5 versus 3 mg/ml). On the other hand, Alveofact, which contains more POPC, displayed better static and dynamic adsorption rates. Hence, the static and dynamic differences between Curosurf and Alveofact seen in vitro can be explained by their different PC molecular species composition. Although the practical relevance of such molecular differences remains to be elucidated, we conclude from our data that porcine surfactant better meets the physiologic PC composition of human surfactant (37).

The other commercial surfactants investigated are either artificially enriched in DPPC (Survanta) or contain DPPC as the only original surfactant component (Exosurf). Their _min values were impaired in relation to both Alveofact and Curosurf. However, they not only contain less or no hydrophobic surfactant proteins (SP-B and SP-C) than Alveofact and Curosurf, but are also impoverished in (Survanta) or free of (Exosurf) POPC, PC 16:0/16:1, and PC 16:0/14:0, three surface adsorption promoting PC species (13). In addition, the ultrastucture of Survanta and Exosurf considerably differed from that of other surfactants. Exosurf contained only large crystalline structures (apparently DPPC), which were present in only small concentrations in NBS, NPS, Alveofact, or Curosurf. In Survanta, both the vesicular/oligolamellar structures of Alveofact and the crystalline structures characteristic of Exosurf were present. Hence, incorporation of the added DPPC into the natural surfactant components seemed to be either incomplete in Survanta or not to have occurred at all. As rapid DPPC adsorption requires the integration and assistance of surfactant proteins (11, 38, 39) and unsaturated PC (13), exogenous DPPC probably did not improve the surface tension function of Survanta under the dynamic in vitro conditions represented by the PBS. In agreement with our quantitative morphologic data, Survanta and Exosurf were not only less potent than Alveofact or Curosurf, but also displayed an impaired intrinsic activity, as _min values below 5 mN/m were not achieved. These findings contrast to those of McMillan and colleagues (40), who found _min values < 5 mN/m for Exosurf in the captive bubble surfactometer, but agree with recent data on the poor adsorption properties of DPPC or isolated PL mixtures to gas-liquid interfaces in the absence of SP-B/C (11). On the basis of these findings, we speculate that the clinical response to exogenous Survanta or Exosurf seen in patients with respiratory distress syndrome is not predominantly due to the intrinsic surface activity of these therapeutics, but involves the contribution of other, endogenous, mechanisms such as those occurring during reprocessing of surfactant phospholipids (41).

Calcium as a Factor Affecting Surface Tension Measurement

Comparability of results from studies on functional differences between surfactant preparations (1) is hampered by the large variability in the experimental conditions under which these were performed. One of the experimental conditions that varies considerably between studies is calcium concentration. Calcium is required for an in vitro assembly of tubular myelin (42) and improves surface tension function in vitro (1). Optimal adsorption of lipid extract surfactants requires the presence of calcium ions in the subphase (43) and, also, the ability of SP-A, -B, and -C to enhance PL adsorption from the subphase into the air/liquid interphase was shown to be calcium-dependent (44). In addition, calcium is important for maintaining large aggregate formation during surface-area cycling by stabilizing tubular myelin in native surfactant (48). As demonstrated by our data, function of native surfactant was independent from additional calcium in the suspending buffer at the concentrations tested. In contrast, surface tension function of commercial lipid extract surfactants was strongly influenced by calcium. The physiologic calcium concentration in the alveolus is approximately 1.5 mmol/L (27), which is considerably below that used in some in vitro studies (1, 4, 5, 10, 11). We found that surface tension function required for a "good" surfactant was not achieved with either Exosurf or Survanta at physiologic calcium concentrations (1.5 mmol/L), but it could be achieved with Survanta at nonphysiologically high calcium concentrations (6 mmol/L). In contrast, Curosurf and Alveofact met these requirements already at 1.5 mmol/L calcium, but only at PL concentrations that were three to 12 times higher than those of native surfactants. However, when calcium was artificially increased, the critical surfactant concentration required to achieve _min values < 5 mN/m was diminished to 1 mg/ml for both Alveofact and Curosurf. Hence, the therapeutic surfactants, which are less effective than native surfactants, may erroneously be regarded as equivalent to native surfactants if investigated at high calcium concentrations.

Clinical Relevance of In Vitro Surfactant Analyses

All commercial surfactant preparations investigated in this study are widely used and have been clinically shown to be effective in infant respiratory distress syndrome (IRDS) (43). However, their functional differences under standardized conditions in vitro, both compared with each other and with their native progenitors, and the putative consequences of their biochemical differences, are less clear. In animal experiments surfactant containing a synthetic SP-A analogue was superior to surfactant lacking this component (36). In the future, surfactant may become a routine treatment of diseases other than IRDS, e.g., the acquired respiratory distress syndrome (ARDS) or acute asthma (49). Functional requirements of a therapeutic surfactant under these conditions may be different from those in IRDS since in the former, alveolar or bronchiolar dysfunction is due to inhibition rather than to lack of surfactant. Moreover, it is still important to consider SP-A- or SP-A-analogue-containing surfactants with respect to dosage reductions caused by their higher intrinsic activities (36, 50) because SP-A may help to prevent surfactant inhibition (50) and because SP-A, added to a commercial surfactant, improves dynamic compliance and lung recruitment (51). Finally, there is increasing evidence that additional components such as individual PL molecular species exert distinct effects on both surfactant function and on lung cells (12, 13). Hence, we postulate that despite the overall clinical success of both natural and synthetic surfactants in IRDS, a standardized and detailed protocol for functional and biochemical analyses is a prerequisite for the characterization of surfactants as therapeutic drugs.

We conclude that unmodified lipid extract surfactants display an intrinsic activity comparable to that of native surfactant. However, their potency is impaired compared with native surfactants, likely because of the absence of SP-A and a reduced concentration of SP-B/C. Differences in the concentrations of Alveofact and Curosurf necessary to reach _min values below 5 mN/m and adequate _ads or _max values can be explained by their differences in DPPC and POPC concentrations, which is in agreement with the experimental data on the functional properties of individual PL components. In contrast, semisynthetic (Survanta) or synthetic (Exosurf) surfactants were unable to demonstrate surface tension properties comparable to those of native or lipid extract surfactants. Finally, the surface tension function of "natural" surfactant preparations (Alveofact, Curosurf) can be erroneously regarded as equivalent to native surfactant if in vitro calcium concentrations are above the physiologic range. Our data underscore the importance of a standardized protocol for the assessment of surfactant function and for a detailed biochemical analysis, using native surfactant as a reference.