Molecular Weight Dependence of the Depletion Attraction and its Effects on the Competitive Adsorption of Lung Surfactant

Abstract

Albumin competes with lung surfactant for the air-water interface, resulting in increased surface tension. Polyethylene glycol (PEG) and other hydrophilic polymers restore the normal rate of surfactant adsorption to the interface, which re-establishes low surface tensions on compression. PEG does so by generating an entropic depletion attraction between the surfactant aggregates and interface, reducing the energy barrier to adsorption imposed by the albumin. For a fixed composition of 10 g/L (1 wt.%), surfactant adsorption increases with the 0.1 power of PEG molecular weight from 6 kDa - 35 kDa as predicted by simple excluded volume models of the depletion attraction. The range of the depletion attraction for PEG with a molecular weight below 6 kDa is less than the dimensions of albumin and there is no effect on surfactant adsorption. PEG greater than 35 kDa reaches the overlap concentration at 1 wt% resulting in both decreased depletion attraction and decreased surfactant adsorption. Fluorescence images reveal that the depletion attraction causes the surfactant to break through the albumin film at the air-water interface to spread as a monolayer. During this transition, there is a coexistence of immiscible albumin and surfactant domains. Surface pressures well above the normal equilibrium surface pressure of albumin are necessary to force the albumin from the interface during film compression.

Introduction

Lung surfactant (LS) lines the alveolar air-liquid interface and lowers the interfacial tension in the lungs [1, 2]. The surface tension control imposed by LS is likely compromised during acute respiratory distress syndrome (ARDS) which afflicts 140,000 people annually in the US with a 40% mortality rate [2, 3]. The clinical similarities between ARDS and neonatal RDS suggest that the lack of functional surfactant that causes NRDS might be a common factor. However, replacement surfactant treatments that have been successful in treating NRDS have been disappointing when used to treat ARDS [4]. The replacement LS, as does the native surfactant, loses its ability to reduce surface tension and is said to be “inactivated” [2, 5-7]. In addition to inactivated LS, serum and inflammatory protein levels in the bronchial and alveolar fluid of ARDS patients are elevated {Spragg, 1991 #431; Gunther, 2001 #434; Ishizaka, 2004 #162; Notter, 2000 #126; Nakos, 1998 #25}. In vitro, there is an ARDS-like depression of LS activity when serum proteins are added to a surfactant-covered interface [12], surfactant is added to a serum-covered interface [13] or both surfactant and serum proteins are presented simultaneously [14]. Surfactant inactivation caused by serum leakage into the alveoli is hypothesized to be a contributing factor to the lack of success in ARDS treatment [15, 16], although the detailed biophysical mechanisms of inactivation are unknown.

We have shown that one cause of surfactant inactivation is the inhibition of surfactant adsorption to the air-water interface caused by the competitive adsorption of serum proteins (such as albumin, immunoglobulin and fibrinogen) [12, 13, 17, 18]. Many serum proteins are surface-active and have a surface pressure, Π, (Π = γ_w-γ ; γ_w is the surface tension of a clean air-water interface, 72 mN/m, and γ the measured surface tension) that is a logarithmic function of protein concentration up to a saturation concentration, which is ∼1 mg/mL for albumin [12, 19]. The surface pressure at the saturation concentration for albumin and many other serum proteins is between 18 and 25 mN/m (γ ∼ 47-54 mN/m) [12, 19], which is much lower than Π ∼70 (γ near zero) required for proper respiration. Albumin adsorbed at the alveolar air-water interface induces a steric and electrostatic energy barrier to LS adsorption [12, 13, 17, 18, 20], which inhibits surfactant transport to the interface [17, 21], thereby reducing surfactant adsorption. Insufficient surfactant at the interface does not allow the low surface tensions required for proper lung function to be reached and the work of breathing increases. This increases the potential for further inflammation and injury, consistent with the development of ARDS.

Several hydrophilic polymers, including polyethylene glycol (PEG), dextran, and hyaluronic acid enhance the ability of clinical surfactants to resist inactivation by serum proteins and other substances in vitro [13, 14, 17, 18, 20, 22-25]. Lung function of animals with lung injury is markedly improved when polymers are added to clinical surfactants used for treatment of NRDS [6, 7, 26-29]. Theory [17] and experiment [13, 18] suggest that hydrophilic polymers induce a “depletion attraction” [17, 30-34] between LS aggregates in solution and the interface, thereby increasing LS adsorption to the interface and restoring surfactant function [13, 18]. The depletion attraction arises from the decrease in the polymer excluded volume as the surfactant aggregates approach the interface [17, 30, 31]; in a dilute solution, the translational entropy of non-entangling, non-adsorbing polymer molecules increases with increasing solution volume [30]. The increased entropy of the polymer molecules acts like an attractive force between the surfactant aggregates and the interface (or between the aggregates) that pushes the surfactant towards the interface, thereby enhancing adsorption [13, 17, 30, 31, 34].

The depletion attraction model provides quantitative predictions regarding the molecular weight dependence of PEG on adsorption (Eqns. 1-5) [17]. At low polymer concentrations, the magnitude of the depletion attraction increases linearly with the polymer weight fraction below the entanglement concentration of the polymer [13, 17, 30, 31, 34]. However, the range of the depletion attraction is limited by the hydrodynamic diameter of the polymer [17, 30, 31, 34]. If the molecular weight, and hence the diameter of the polymer, is too small, the range of the depletion attraction may not overlap the range of the repulsive electrostatic and steric interactions. As a result, a minimum molecular weight is necessary to enhance surfactant adsorption. For molecular weights above this minimum, the depletion attraction increases weakly (MW^0.1) with the molecular weight of the polymer [17] below the entanglement concentration (which also depends on molecular weight). Above the polymer entanglement concentration, the magnitude of the depletion attraction decreases because the excluded volume effects and the entropy increase of the polymer decrease, leading to less enhancement of surfactant adsorption [35, 36].

To test these predictions of the depletion attraction/energy barrier model, we examine the inactivation induced by albumin of the clinical lung surfactant, Survanta, and the enhancement of surfactant adsorption induced by PEG with molecular weights ranging from 1.45 kDa - 200 kDa. Several studies have shown that 10 kDa PEG reverses surfactant inactivation both in vitro and in vivo [6, 18, 24, 27]. We previously showed that surfactant adsorption increased exponentially with the weight fraction of 10 kDa PEG, but only linearly with surfactant concentration as predicted by the depletion attraction theories and the modified Smolukowski equations (Eqns. 1-5) [13, 17]. Here we show that there is a minimum molecular weight of PEG needed to reverse inhibition; PEG below 6 kDa does not reverse inhibition because the range of the depletion attraction, (twice the polymer radius of gyration), is less than the thickness of the albumin layer. At 1 wt %, PEG above 100 kDa is above the overlap concentration, at which the depletion interaction decreases and the solution viscosity increases, both of which reduce surfactant adsorption. For 1% wt. PEG, surfactant adsorption is proportional to (MW)^0.1 from 6 - 35 kDa as predicted by the depletion attraction model.

Fluorescence images show that surfactant must break through the albumin film to the air-water interface to spread as a monolayer and lower surface tension; in the absence of polymer, the surfactant cannot reach the interface due to the albumin. Once surfactant has broken through locally, a coexistence of immiscible albumin and surfactant domains are present at the interface and the polymer increases the displacement of albumin for surfactant at the interface. Upon compression, the albumin cannot withstand high surface pressures and is removed in favor of the surfactant; surface pressures well above the 18 - 25 mN/m equilibrium surface pressure of albumin [12, 19] are necessary to completely force it from the interface.

Theory

The effects of albumin on surfactant adsorption can be quantified using a variant of the classic Smolukowski analysis of colloid stability [17, 37]. Charged, surface-active serum proteins adsorbed at an interface induce an electrostatic and steric barrier to surfactant adsorption at that interface, similar in magnitude to the energy barriers responsible for colloidal stability against flocculation [38]. The addition of hydrophilic polymers generates a “depletion attraction” which offsets the otherwise repulsive potential induced by the proteins at the interface.

The change in surfactant interfacial concentration with time, $\frac{d\Gamma }{dt}$, is driven by the gradient of a generalized chemical potential of surfactant in the subphase. The presence of a charged, protein layer adsorbed to the interface induces a potential energy barrier to surfactant adsorption, as in the Smolukowski theory of colloid stability [17]:

$$\frac{d\Gamma }{dt}=C_o{\left(\frac{D_{\mathit{eff}}}{\pi t}\right)}^{1∕2}\exp \left[-\left(V_{\max }\right)∕k_{B}T\right]$$ (1)

C_o is the bulk surfactant concentration, D_eff is the effective surfactant diffusivity, V_max is the maximum height of the potential energy barrier (located a distance l from the interface), k_B is Boltzmann’s constant and T is the temperature. The magnitude and location of the potential energy barrier are both important when determining the impacts on surfactant adsorption. The polymer generated depletion attraction modifies this the potential energy barrier to enhance surfactant adsorption. The potential energy barrier results from several effects which are taken to be additive:

$$V_{\max }\cong (E_1-E_0)+\Pi \Delta A+E_{\mathit{Elect}}+W$$ (2)

(E₁-E₀) is the energy difference between surfactant molecules at the interface and in the bulk, which drives surfactant adsorption to a clean interface. This is opposed by a steric term, ΠΔA , which accounts for the energy necessary to clear an interfacial area, ΔA, of serum protein at a surface pressure of Π for surfactant adsorption. The repulsive double-layer electrostatic potential, E_Elect, is due to the negatively charged albumin and its magnitude depends on the ionic strength of the subphase and the net charge at the interface and the surfactant bilayers [21].

In the presence of non-adsorbing polymers such as PEG, these two repulsive terms are offset by the depletion attraction, W. Moving a large sphere of radius R (the surfactant aggregates) to within l of an interface (or toward another large sphere) frees excluded volume for the polymer of radius R_g, and thereby increases the polymer entropy, leading to the depletion attraction [17,31]:

$$\begin{array}{c}\hfill l\le R_g:W=-3(R∕R_g){\varphi }_{\mathrm{p}}{k}_{B}T{[1-(l∕2R_g)]}^2\hfill \\ \hfill l>2R_g:W=0\hfill \end{array}$$ (3)

ϕ_p is the volume fraction of the polymer in solution. The depletion attraction is very short-ranged and has no effect for distances greater than the polymer diameter. As the surfactant particle comes into contact with the air-water interface, W_max =-3(R/R_g)ϕ_pk_BT . For the PEG used here, R_g ranges from 1.5 nm (1.45 kDa PEG) to 22.7 nm (200 kDa PEG). Bovine serum albumin is a prolate spheroid of dimensions 4 × 4 × 14 nm and self assembles at the air-water interface with a thickness of 4 nm; infrared spectroscopy measurements suggest that denaturation of the albumin does not occur on a time scale of hours or days [39]. The surfactant aggregates are of order microns in diameter [17, 40]; hence the depletion attraction is significantly greater for the surfactant aggregates than for albumin. Due to its entropic origin, the depletion attraction is independent of the surfactant, protein and polymer as long as the polymer does not adsorb to the protein, surfactant or interface. This explains why PEG [6, 13, 20, 27], dextran, [7, 14] and hyaluronan [25] are all effective at enhancing surfactant adsorption.

Eqns. 1 - 3 predict that the net rate of surfactant adsorption is dominated by the magnitude of the energy barrier, V_max, and hence on W, the magnitude of the depletion attraction [13, 17, 18]. Experimentally, we compare the rate of surfactant adsorption with albumin and PEG in the subphase, $d\Gamma ∕d{t}_{\mathit{Alb}+\text{Poly}}$, to surfactant adsorption from a clean subphase, $d\Gamma ∕dt$. From Eqns. 1-3, the enhancement in surfactant adsorption due to the polymer should scale as:

$$\frac{d\Gamma ∕d{t}_{\mathit{Alb}+\text{Poly}}}{d\Gamma ∕dt}\cong \exp \left[\frac{-(\Pi \Delta A+E_{\text{Elect}}+W_{\max })}{k_{B}T}\right]=\exp \left[\frac{-(\Pi \Delta A+E_{\mathit{Elect}})}{k_{B}T}\right]\exp \left[3\frac{R}{R_g}{\varphi }_p\right]$$ (4)

As the subphase concentration of albumin is fixed, (ΠΔA +E_Elect) is constant. Previous work has verified the exponential dependence of surfactant adsorption on polymer volume fraction (ϕ_p) at a fixed molecular weight (constant R_g) and surfactant aggregate size (R) [13]. Here we vary the molecular weight, and hence R_g, at a constant weight fraction of polymer. Water is a good solvent for PEG over the entire experimental range and several authors have shown R_g = αMW^0.55 where MW is the polymer molecular weight and α is a constant [41, 42]. In Eqn. 4, ${\varphi }_p\sim N_p{R}_g^3∕V$ in which N_p is the number of polymer molecules in a volume V. For a fixed polymer weight fraction of ρ, $N_p=\rho V∕MW$, hence the depletion attraction for PEG should scale as W_max ∼ MW^0.1. The range of the depletion attraction is also proportional to R_g and hence should increase as MW^0.55.

Materials and Methods

The clinical replacement surfactant Survanta (Abbott Laboratories, Columbus, Ohio) was a generous gift of the Santa Barbara Cottage Hospital nursery. Survanta is an organic extract of minced bovine lungs which has been fortified with dipalmitoylphosphatidylcholine (DPPC), tripalmitin and palmitic acid. It contains 80 - 90 wt.% phosphatidylcholine, of which, ∼70 wt.% is saturated DPPC, approximately 5 wt.% negatively charged phospholipids, primarily phosphatidylglycerol and phosphatidylserine, and about 10 wt.% palmitic acid [40, 43]. There is < 2 wt.% of the lung surfactant specific proteins SP-B (0.04-0.13 wt. %) [44] and SP-C (0.9-1.65 wt.%) [40, 43, 45]. Survanta and other clinical lung surfactants form multi-micron bilayer aggregates in buffered saline solution [40] and Survanta adsorbs to a clean interface from solution as a combination of monolayer and multilayers [46]. Polyethylene glycol (PEG) of 1.45 kDa, 3.35 kDa, 6 kDa, 10 kDa, 20 kDa, 35 kDa, 100 kDa and 200 kDa and bovine serum albumin were obtained from Sigma (St. Louis, MO) and used as received.

Isotherms were recorded at 25°C (No significant changes are seen from 23 - 37°C [46]) using a custom stainless steel ribbon trough (Nima, Coventry, England) designed to minimize film leakage at high surface pressures (low surface tensions). Surface pressure was monitored during compression and expansion using a filter paper Wilhelmy plate. The trough had a surface area of 130 cm², a subphase volume of 150 mL and a typical compression/expansion cycle took 8 min (∼0.42 cm²/sec). Albumin and polymer were dissolved in the same buffer used as the subphase (NaCl 150mM, CaCl₂ 2mM, NaHCO₃ 0.2mM and pH=7) at the stated concentrations for all experiments. To initiate each experiment, a saline buffer subphase containing albumin and polymer was added to the Langmuir trough and allowed to equilibrate for 10 minutes; for albumin containing subphases the surface pressure was ∼18 mN/m, consistent with the known surface activity of albumin [12, 19]. Survanta was diluted in the same buffer to a lipid concentration of 2 mg/mL and was deposited as microliter drops from a syringe by touching the drop to the air/water interface of the open trough. The drops primarily wound up in the subphase adjacent to the interface; drop diffusion into the subphase was easily monitored as the Survanta usually was labelled with a fluorescent dye. The drops did not spread appreciably at the interface, in contrast to when surfactants are spread from organic solvents; essentially all of the Survanta adsorbed from the subphase. No stirring of the subphase was used and the first compression began 20 minutes after deposition of a fixed quantity of Survanta. The amount of Survanta chosen for the inhibition experiments was such that collapse would occur at about 50% trough compression in the absence of albumin or PEG; this was then held fixed for all remaining experiments except where noted.

A Nikon Optiphot optical microscope (Nikon, Tokyo, Japan) was positioned above the trough with either a 10X or 50X extra long working distance objective designed for fluorescent light [47-50]. Full-length movies and individual frames were recorded directly to computer (Moviestar, Mountain View, CA). Contrast in the images was due to segregation of 1% mol fluorescent lipid Texas Red-DHPE (Molecular Probes, Eugene, OR) which causes the Survanta monolayer to appear a light grey in images [46]. Larger aggregates of Survanta have significantly more dye and appear bright white, leading to an overall mottled texture for the surfactant film. The albumin was not labeled, does not fluoresce and appears black in the images.

Zero-shear viscosities at 25°C were measured with a capillary viscometer (#100 C503, Cannon Instrument Co., State College, PA) calibrated for a viscosity range of 0.8 to 100 mPa-s (water has a viscosity of ∼ 1 mPa-s at 25°C) [51]. Samples containing PEG of varying molecular weights were dissolved in buffer (NaCl 150mM, CaCl₂ 2mM, NaHCO₃ 0.2mM and pH=7). Three measurements were taken and averaged for each sample with a typical standard deviation of 0.7%. Overlap is generally taken to be at polymer concentrations that increase the solution viscosity by more than a factor of two [52].

Results

Fig. 1a shows a typical compression-expansion cyclic isotherm for 800 μg of the clinical lung surfactant, Survanta, spread on a clean saline buffered subphase (no albumin or PEG) [13, 46]. The isotherm traces over itself on subsequent cycles, and on compression exhibits a characteristic shoulder at Π ∼ 40 mN/m which corresponds to rearrangement of the unsaturated lipids and surfactant proteins SP-B and SP-C in the film [46]. On further compression, the surface pressure rises abruptly to the collapse pressure, Π_max ∼ 65 mN/m, where the film begins to “collapse” and forms cracks and folds [13, 46]. Film collapse determines the minimum surface tension possible for a given surfactant. The hysteresis between compression and expansion cycles is typical of Survanta and other clinical and natural lung surfactant isotherms [2, 53] and is due to the partial re-adsorption of the collapse structures into the monolayer [12]. On a clean subphase, re-expanding the interface after monolayer collapse leads to a rapid drop in surface pressure to about 5-10 mN/m which is maintained until compression is resumed [2, 46]. In general, collapse structures do not re-adsorb to the interface until the surface pressure is well below the collapse pressure, resulting in the compression and expansion hysteresis [12]. There is no significant change in Survanta isotherms from 23 - 37°C [46].

Figure 1.

Cyclic isotherms of Survanta on buffered saline subphase containing albumin and/or polymer. (a) 800 μg Survanta on a clean buffered saline subphase (no albumin or polymer). On compression, the isotherm exhibits a characteristic shoulder at 40 mN/m. The collapse plateau at Π_max ∼ 65 mN/m determines the maximum surface pressure (minimum surface tension) possible for a given surfactant. On expansion, the surface pressure immediately drops to 10 mN/m; lung surfactant isotherms exhibit significant hysteresis between the compression and expansion parts of the cycle as collapsed material re-adsorbs at lower surface pressures [12].

(b) Black curve: 800 μg Survanta deposited onto a saline buffer subphase containing 2 mg/mL albumin and 1% wt. 1.45 kDa PEG. The characteristic shoulder and collapse plateau on compression seen in (a) cannot be reached with albumin in the subphase, despite the presence of low molecular weight PEG. Red Curve: The isotherm for the albumin subphase, with no Survanta or PEG. The two curves trace over each other, indicating that the interfacial film is dominated by albumin and the surfactant has been prevented from reaching the surface as shown in the fluorescence image, Fig. 2b.

(c) 800 μg Survanta on saline buffer containing 2 mg/mL albumin and 1% wt. 10 kDa PEG. The characteristic shoulder and collapse plateau have been restored at similar trough areas as (a) with little change in surface pressure showing that the presence of 10 kDa PEG completely reverses the surfactant adsorption inhibition. The only difference is that the expansion cycle in Fig. 1c has a minimum surface pressure that is about 15-20 mN/m (vs. 5-10mN/m in Fig. 1a) which is a result of enhanced surfactant adsorption during the expansion due to 10 kDa PEG. Compression cycles are labeled 1-4 in chronological order.

(d) 800 μg Survanta on saline buffer containing 2 mg/mL albumin and 1% wt. 200 kDa PEG. In this case, the characteristic shoulder and collapse plateau of (a) are obtained at higher trough compression indicating that 1% wt. 200 kDa PEG partially reverses the adsorption inhibition. Compression cycles are labeled 1-4 in chronological order.

However, when the same amount of Survanta (800 μg) is deposited on a buffered subphase containing 2 mg/mL albumin and 1% wt. PEG of molecular weight 1.45 kDa (Fig. 1b, black curve), the surface pressure does not increase above 35 mN/m even at the smallest trough area. Albumin concentrations in ARDS alveolar fluid may reach 100 mg/mL with an average concentration of 25 mg/mL [10] The concentrations used here are significantly lower than typically found in ARDS patients, but are at or above the saturation concentration at which the albumin surface pressure no longer increases with concentration [12]. The characteristic shoulder and collapse plateau on compression seen in Fig. 1a cannot be reached with albumin in the subphase, despite the presence of low molecular weight PEG. Both the compression and expansion isotherm are not different than that of albumin alone (Fig. 1b, red curve). The minimum surface pressure (∼15 mN/m) during expansion of the Survanta-albumin-1.45 kDa PEG isotherm (Fig 1b, black curve) is higher than Fig. 1a and is set by the re-adsorption of albumin to the interface at its saturation surface pressure. This shows that Survanta on an albumin subphase does not adsorb to the interface in appreciable quantity and that 1.45 kDa PEG does not increase adsorption. Survanta inactivation under these conditions results from an inability of the surfactant to adsorb to the interface.

In contrast to 1.45 kDa PEG, for 1% wt. 10 kDa PEG and 2 mg/mL albumin in the subphase (Fig. 1c) restores the isotherm to that of the clean interface after two compression-expansion cycles (Fig. 1a), showing that the higher molecular weight PEG reverses surfactant inactivation. The compression cycle in Fig. 1c displays the characteristic shoulder at Π ∼ 40 mN/m and a collapse plateau at Π_max ∼ 65 mN/m at similar trough areas as Fig. 1a, confirming that PEG 10 kDa does not alter the properties of the surfactant monolayer. Rather, the PEG enhances the adsorption of surfactant to the interface and the displacement of albumin from the interface. The only difference between Fig. 1a and Fig. 1c is that the expansion cycle in Fig. 1c has a minimum surface pressure of about 15 mN/m (vs. 5-10mN/m in Fig. 1a). While the increased minimum surface pressure could be set by albumin re-adsorption at its equilibrium surface pressure, Survanta on subphases containing PEG but no albumin also shows an increased minimum surface pressure during the expansion. For example, Survanta (800 μg) deposited onto a subphase containing 5% wt. 10 kDa PEG (no albumin) shows a minimum surface pressure during expansion of ∼23 mN/m suggesting that the increased minimum surface pressure in Fig. 1c is a result of PEG-enhanced surfactant adsorption (data not shown). The restoration of the Survanta isotherm (Fig. 1a) during subsequent compression cycles in Fig. 1c shows that the rate of surfactant adsorption to the interface is sufficient to prevent any appreciable albumin re-adsorption. 1% wt. 200 kDa PEG also enhances surfactant adsorption (Fig. 1d) though not as efficiently as 10 kDa PEG (Fig. 1c). Fig. 1d shows that the characteristic shoulder at Π ∼ 40 mN/m and the collapse plateau at Π_max ∼ 68 mN/m are present, but at a significantly smaller trough area (greater compression) indicating less surfactant has adsorbed to the interface [13] than in Fig. 1a or 1c.

Fig. 2a shows a fluorescence micrograph of a typical interface after Survanta adsorption. Survanta (doped with 1% mol Texas Red-DHPE) adsorbs to the interface of a saline buffer subphase as a mixture of monolayers (mottled light grey and dark grey) typical of a phase separated lipid/protein monolayer [49] along with bright, three-dimensional aggregates that appear to be attached to the interface [46]. This characteristic mottled texture is found at all surface pressures from 0 to collapse [49] and is indicative of the coexistence of solid-like, semi-crystalline domains (darker gray) and less ordered, fluid domains (lighter gray) [54]. In contrast, Fig. 2b shows that images of Survanta on an albumin subphase consist of isolated, out-of-focus bright regions with an overall dark homogeneous background. While the Survanta aggregates appear to be able to diffuse toward the interface, they apparently cannot reach the interface and spread as a monolayer due to the adsorbed albumin film. The Survanta does not form a monolayer at the interface and the surface pressure remains low throughout the compression/expansion cycle. The interfacial film is dominated by albumin; the albumin prevents the surface pressure from increasing by preventing surfactant adsorption to the interface.

Figure 2.

Fluorescence images of 800 μg Survanta spread at varying subphase compositions. Images are 1023 μm by 789 μm.

(a) Survanta on a clean, buffered subphase at Π = 43 mN/m during compression. The image shows the mottled texture typical of a phase separated lipid/protein monolayer. The bright spots are Survanta aggregates partially adsorbed to the interface and partially in solution.

(b) Survanta on a subphase containing 2 mg/mL albumin at Π = 25 mN/m during compression. The weak out of focus fluorescence on a black background shows that Survanta aggregates come close to the interface, but cannot spread due to the albumin film at the interface.

The remaining images show Survanta on a subphase containing 2 mg/mL albumin and 1% wt. 200 kDa PEG in successive compression/expansion cycles. Real time fluorescence movies of the isotherms are available in the supplementary materials.

Row 2-First Cycle (c) Π = 18 mN/m during compression. At low surface pressure, no fluorescence is visible showing that the albumin prevents Survanta from adsorbing to the interface. (d) Π = 38 mN/m during compression. Similar to (b), Survanta (out of focus bright spots) cannot break through the albumin film even at a surface pressure well above the equilibrium spreading pressure of albumin (∼ 20 mN/m).

Row 3-Second Cycle (e) Π = 18 mN/m during compression. Survanta breaks through the albumin film; extended (>1000 μm) immiscible Survanta (mottled grey) and albumin (black) domains coexist on the interface. (f) Π = 16 mN/m during expansion. The Survanta and albumin coexistence persists as the second cycle Π_max (∼ 55 mN/m) is not sufficient to squeeze out all of the albumin.

Row 4-Third Cycle (g) Π = 26 mN/m during compression. Albumin domains exist at low surface pressure but are squeezed out at high surface pressure (Third cycle Π_max ∼ 65 mN/m). (h) Π = 15 mN/m during expansion. From the third expansion onward, only Survanta is observed in the film.

Row 5-Fourth Cycle (i) Π = 43 mN/m during compression. The mottled texture of the film is similar to Survanta (a) however the bright domains are significantly larger in (h-j) due to the PEG induced depletion attraction leading to the flocculation and growth of the Survanta aggregates. (j) Π = 68 mN/m during compression. The film reaches a collapse pressure similar to Fig. 1a.

Figs. 2c-j show the gradual displacement of albumin from the interface by Survanta in the presence of 200 kDa PEG in the subphase. Survanta was deposited onto a subphase containing albumin and 1% wt. 200 kDa PEG and the interface was compressed and expanded. Images from the first cycle show a featureless albumin-covered interface at low surface pressure (Fig. 2c, Π = 18) and out of focus bright spots at a higher surface pressure (Fig. 2d, Π = 38), indicating that Survanta aggregates approach the albumin-covered interface but cannot spread. It appears from these and other images that break-up of the Survanta aggregates from bilayer to monolayer is required for the aggregates to reach the air-water interface. At maximum compression on the first cycle, (Fig 1d, 1) the maximum surface pressure is ∼ 40 mN/m, which is similar to an albumin-covered interface (Fig. 1b). However, images from the second cycle show a coexistence between Survanta (mottled bright texture) and albumin (black) with a well-defined interface between the materials. Survanta has broken through the albumin film, perhaps through defects in the monolayer [55] and occupies an increasing fraction of the interfacial area (See Movie #1 in Supplemental Materials). The coexistence between the extended (>1000 μm) interfacial domains of Survanta and albumin occurs both during compression (Fig. 2e, Π = 18) and expansion (Fig. 2f, Π = 16) of the film. Even though the surface pressure reaches a Π_max of ∼ 55 mN/m (Fig 1d, 2), which is much greater than the equilibrium surface pressure of a 2 mg/ml albumin solution (∼ 20 mN/m), albumin remains at the interface. The removal of albumin from the interface may be limited by a high surface viscosity that retards the motion of the albumin (See Movie #2 in Supplemental Materials). However, as Survanta can maintain a much higher dynamic surface pressure on compression than albumin, after repeated compressions the albumin is forced from the surface into the subphase at surface pressures of 50-60 mN/m (See Movie #3 in Supplemental Materials). Additional Survanta likely adsorbs from the bright aggregates that appear to be attached to the interface to occupy the new interface created as the trough expands.

Images from the third cycle show Survanta/albumin coexistence at low surface pressure (Fig. 2g, Π = 26) but now adequate surfactant has adsorbed to raise the maximum surface pressure to 65 mN/m (Fig 1d, 3) which is sufficient to force all of the albumin from the interface. Only Survanta is observed in the interfacial film during the third cycle expansion (Fig. 2h, Π = 15). On the fourth cycle, the mottled Survanta texture similar to Fig. 2a is seen exclusively at all surface pressures (Fig. 2i, Π = 43) and the system forms a collapse plateau (Fig. 2j, Π = 68). Due to the extremely short collapse plateau, the cracks and folds typically associated with film collapse [13, 46] are not seen in the 200 kDa PEG system. Images from systems with a collapse plateau similar in length to Survanta on a saline subphase (Fig. 1a) such as 1% wt. 10 kDa PEG show cracks and folds. While the texture is similar between Figs. 2a and 2h-j, the bright aggregates are significantly larger in Figs. 2h-j; the PEG-induced depletion attraction also leads to flocculation of the Survanta aggregates [17, 40]. In addition to helping the Survanta break through the albumin monolayer, the depletion attraction helps maintain a larger reservoir of surfactant in close proximity to the interface during collapse and re-adsorption cycles. The mechanism of surfactant displacement of the albumin at the interface is similar for all PEG molecular weights which restore surfactant adsorption to the interface (6-200 kDa). The process is most easily imaged for the 200 kDa PEG because it takes three compression-expansion cycles to complete, images from 10 kDa PEG demonstrate a similar surfactant displacement although here the albumin is eliminated by the second cycle.

Fig. 3a shows the fourth cycle of compression isotherms for varying amounts of Survanta deposited onto a clean saline subphase. The characteristic shape of the isotherms in Fig. 3a are translated unchanged (note the shape of the collapse plateau and the shoulder at ∼ 40 mN/m) from low to high trough area as the amount of Survanta added to the trough is increased from 3 μg up to 800 μg. As the relationship between surface pressure and area/molecule is fixed for a given surfactant composition and temperature, an increase in surface pressure at a given trough area means that the total amount of surfactant at the interface has increased [46]. Therefore increasing surfactant adsorption translates the isotherm at a given surface pressure from low to high trough area. Eventually, further increases in surfactant concentration have diminishing effect (note the smaller offset between 300 μg to 800 μg) as the adsorption saturates. Any additional surfactant that does not adsorb to the monolayer remain near the interface as bilayer aggregates as in Fig. 2; this implies that there is more material near the interface for the higher concentration sample compared to the lower concentration sample.

Figure 3.

Fourth cycle compression isotherms of (a) varying concentrations of Survanta on a clean buffered subphase and (b) 800 μg Survanta on a saline buffered subphase containing 2 mg/mL albumin and 1% wt. PEG of varying molecular weights.

(a)▽ 3 μg Survanta; △ 8 μg Survanta; ⎔ 30 μg Survanta; ◇ 80 μg Survanta; ○ 300 μg Survanta; □ 800 μg Survanta; At a given surface pressure, the isotherms are translated unchanged from low trough area to high trough area with increasing Survanta concentration (note the characteristic shoulder at ∼ 40 mN/m and the collapse plateau at ∼ 65 mN/m). The interface becomes saturated for concentrations greater than about 300 μg; the 800 μg isotherm is not displaced significantly to higher trough areas. Increasing surfactant adsorption to the interface leads to this gradual shift of the isotherms from left to right.

(b) □ Survanta only; ○ Survanta-albumin; △ Survanta-albumin-PEG 1.45 kDa; ▽ Survanta-albumin-PEG 3.35 kDa; ◁ Survanta-albumin-PEG 6 kDa; ◇ Survanta-albumin-PEG 35 kDa; ⎔ Survanta-albumin-PEG 100 kDa. Except for PEG 1.45 kDa, the presence of 1% wt. PEG in the subphase shifts the isotherms to higher trough areas, the same effect as increasing the Survanta concentration in (a). For readability, 10 kDa and 20 kDa curves were omitted as they nearly overlap with 35 kDa; 200 kDa was similarly omitted as it overlapped with 100 kDa. Cumulative results from all experiments are presented in Fig. 4. The shaded area denotes the trough area over which the surface pressure was averaged for each PEG molecular weight to obtain the relative surfactant adsorption plotted in Fig. 4.

Fig. 3b shows the effect of albumin on the isotherms is similar to decreasing the Survanta concentration (Fig. 3a), even though the same amount of Survanta (800 μg) was deposited to initiate each experiment. The shifting of the isotherms confirms that albumin inhibits the adsorption of surfactant to the interface as predicted by Eqn. 1. With no added polymer, the isotherm of Survanta deposited on a subphase containing albumin (Fig. 3b, circles) differs little from the 800 μg Survanta-albumin-PEG 1.45 kDa curve (Fig. 3b, up triangles), or from that of albumin by itself (Fig. 1b). In all the isotherms in Fig. 3b, the minimum surface pressure during compression is about 15 mN/m when albumin and PEG are present; this is set by some combination of albumin and PEG-enhanced surfactant adsorption. Compressing the film and comparing to films of pure Survanta or pure albumin can show the relative amount of Survanta and albumin in the film. Fig. 3b shows that even though 800 μg of Survanta was added to the subphase, as well as 1 wt% 1.45 kDa PEG, the albumin prevents surfactant adsorption to the interface (as in Fig. 1b). The maximum surface pressure obtained by the Survanta-albumin isotherm (Fig. 3b, circles) is between that of the 3 μg and 8 μg Survanta isotherms in Fig 3a, which gives an estimate of the amount of surfactant actually adsorbed at the interface. Hence, the effective decrease in surfactant adsorption in the presence of the energy barrier induced by albumin for a Survanta concentration of 800 μg is similar to that from a solution with a concentration 3 μg to a clean interface with no energy barrier. From Eqn. 1, this decrease in adsorption can be used to estimate the magnitude of the energy barrier:$\ln \left(\frac{800\mu g}{3\mu g}\right)=\frac{V_{\max }}{k_{B}T}$, or V_max ∼ 5.6 k_BT.

Adding PEG with molecular weights of 3.35 kDa or greater (Fig. 3b) to the subphase containing both albumin and PEG has the same effect as increasing the surfactant concentration on a clean interface (Fig. 3a). As in Fig. 3a, the shapes of the isotherms are unchanged, just shifted to the right to larger trough areas with increasing PEG molecular weight. This similarity between adding polymers and increasing the surfactant concentration confirms that surfactant behavior at the interface is not affected by the albumin and polymer, but rather only the net rate of adsorption is modified. The Survanta isotherms shift to the right with increasing PEG molecular weight yielding a maximum in the surfactant adsorption for 6 kDa up to 35 kDa. For 100 kDa PEG and above, the isotherm shifts back to the left, indicating that the higher molecular weight PEG is not as effective at promoting Survanta adsorption.

Discussion

Eqn. 4 gives the theoretical prediction of the change in surfactant adsorption relative to a clean interface as a function of the magniture of the depletion attraction, which in turn, depends on the polymer molecular weight and concentration. Fig. 4 shows this relative rate of surfactant adsorption as a function of PEG molecular weight [13]. We define the relative adsorption (RA) as the difference between the sample surface pressure (Π) and the surface pressure of the albumin only isotherm (Π_Alb, red curve in Fig. 1b), divided by the difference between the surface pressure for the clean surface isotherm, Π_Sat (> 1% PEG 10 kDa [13]) and Π_Alb, $RA={\phantom{\mid }\frac{\Pi -{\Pi }_{\mathit{Alb}}}{{\Pi }_{\mathit{sat}}-{\Pi }_{\mathit{Alb}}}\mid }_{A_0}$. All surface pressures were evaluated by averaging over the same trough area (A_o) denoted by the shaded area in Fig. 3b. This region showed the maximum variation in surface pressure. The area/molecule is uniquely determined by the surface pressure for a given surfactant composition and temperature, which means that the amount of surfactant adsorbed is proportional to the surface pressure at a fixed trough area [46].

Figure 4.

Relative adsorption (RA) of 800 μg Survanta on subphases containing 2 mg/mL albumin at varying PEG molecular weights and concentrations. ○ 1% wt. PEG; □ 0.5% wt. PEG; △ 0% wt. PEG which has been plotted for comparison purposes. RA is the difference between the sample surface pressure (Π) and the surface pressure of the albumin only isotherm (Π_Alb, red curve in Fig. 1b), divided by the difference between the surface pressure for the saturated isotherm (Π_Sat, > 1% PEG 10 kDa [13]) and Π_Alb, $RA={\phantom{\mid }\frac{\Pi -{\Pi }_{\mathit{Alb}}}{{\Pi }_{\mathit{Sat}}-{\Pi }_{\mathit{Alb}}}\mid }_{A_0}$. All surface pressures were evaluated by averaging over the same trough area, A₀, denoted by the shaded area in Figure 3b. Region I (PEG 1.45 - 3.35 kDa) corresponds to minimal reversal of surfactant adsorption inhibition, Region II (PEG 6 - 35 kDa) corresponds to complete reversal of adsorption inhibition and Region III (PEG 100 - 200 kDa) corresponds to partial reversal of adsorption inhibition. The dashed line, where RA depends on the MW^0.1 as predicted by Eqns. 1-5, is a good fit to the PEG 1% wt. data in Region II, consistent with the depletion attraction lowering the energy barrier to surfactant adsorption. Data points were offset vertically (maximum 5%) to enhance graph readability.

Fig. 4 shows three distinct responses as a function of PEG molecular weight. Region I corresponds to minimal reversal of surfactant adsorption inhibition, Region II corresponds to complete reversal of adsorption inhibition and Region III corresponds to partial reversal of adsorption inhibition. In Region I, for the lowest molecular weight PEG’s, the range of the depletion attraction is less than the range of the repulsive interactions; V_max in Eqn. 1 is not decreased resulting in minimal effects on surfactant adsorption. For the intermediate molecular weight range in Region II, the depletion interaction is sufficiently long-ranged that it overlaps the repulsive interactions and leads to greatly enhanced surfactant adsorption. From Eqns. 1-4, the relative adsorption with albumin and PEG in the subphase compared to a clean (diffusion-limited) interface should be an exponential function of PEG molecular weight:

$$\ln \left(\frac{d\Gamma ∕d{t}_{\mathit{Alb}+\text{Poly}}}{d\Gamma ∕dt}\right)=\ln \left(RA\right)=-(\Pi \Delta A+E_{\mathit{Elect}}+W_{\max })∕k_{B}T\sim \gamma +\beta M{W}^{0.1}$$ (5)

in which β and γ are constants for a given Survanta and albumin concentration. For 1% wt. PEG over the molecular weight range of 6-35 kDa (Region II), the dashed line in Fig. 4, ln (RA) ∝ MW^0.1, fits the data quite well. Though there is more scatter in the data, the scaling law also holds reasonably well for 0.5% wt. PEG over this range of molecular weights. For the highest molecular weights (Region III), surfactant adsorption decreases for the higher molecular weight PEG at both 0.5 and 1.0% wt. indicating a smaller depletion attraction. The net decrease at higher molecular weights might also be due to the effects of the polymer on the solution viscosity (Table 1); the effective diffusivity, D_eff, in Eqn. 1 is inversely proportional to the solution viscosity.

Table 1.

Physical properties of varying molecular weight PEG in saline buffer. R_g is calculated from the scaling law for PEG in water [41]. W(0 nm) and W(4 nm) are the magnitudes of the depletion attraction in units of k_BT at contact (0 nm) and at 4 nm (roughly the thickness of an albumin film) from Eqn. 3 for molecular weights up to 35 kDa. *W(0) for 100 and 200 kDa are estimates based on the polymer reference interaction site model (PRISM) theory of Fuchs and Schweizer for overlapping, entangled polymers where the maximum value of the depletion attraction occurs at the polymer overlap concentration [35, 36]. Above overlap, the relevant length scale is no longer R_g but polymer mesh length, ξ, which decreases with increasing concentration, reducing the range of the depletion attraction. However, this theory does not predict a distance dependence of the interaction. For each PEG molecular weight, η/η_sol is the ratio of the zero shear viscosity of the 1% wt. polymer solution to the viscosity of the buffer measured with a capillary viscometer. η/η_sol> 2 is typically taken to be a rough indication that the polymer solution is above the overlap concentration. A radius of surfactant aggregate, R, of 500 nm is used for all calculations.

MW (kDa)	2 Rg (nm)	W (0 nm) (kBT)	W (4 nm) (kBT)	η/ηsol
1.45	3.0	-60.1	0	1.12
3.35	4.8	-65.4	-1.8	1.14
6.0	6.6	-69.3	-10.7	1.19
10.0	8.7	-72.9	-21.4	1.24
20.0	12.8	-78.2	-36.9	1.30
35.0	17.4	-82.7	-49.0	1.43
100.0	31.0	-68.2*	-----	2.40
200.0	45.4	-52.1*	-----	3.30

Both the magnitude and range of the depletion attraction are important to surfactant adsorption. In Region I, the range of the depletion attraction (l < 2R_g < 5 nm) does not sufficiently overlap with the maximum in the repulsive potential V_max and surfactant adsorption is not significantly increased. Bovine serum albumin is a prolate spheroid of dimensions 4 × 4 × 14 nm and forms a monolayer at the air-water interface with a thickness of 4 nm [39]. The Debye length for electrostatic repulsion in physiological saline is about 1 nm [38]. Some combination of the albumin film thickness and the Debye length likely defines the location of the maximum in the repulsive potential, and hence the minimum range of a depletion attraction necessary to enhance surfactant adsorption. Table 1 shows 2R_g for the different molecular weight polymers as well as the magnitude of the depletion attraction at 4 nm. The range of the depletion attraction induced by 1.45 kDa PEG (3.0 nm) is less than the axial dimension of the albumin molecule, so we expect minimal effects on surfactant adsorption as observed (Fig. 4). The range of the 3.35 kDa PEG depletion attraction (4.8 nm) is about the same as the albumin dimensions, but W(4 nm) is only -1.8 k_BT which is smaller than the ∼5 k_BT repulsive energy barrier determined from previous experiments [13], so the reversal of inhibition is less than complete, consistent with Fig. 4.

While the magnitude and range of the depletion attraction in Eqn. 3 was derived for hard spheres, Survanta aggregates are generally non-spherical, rough and deformable (Fig. 2) [40]. Even so, for the intermediate molecular weight PEGs (Region II), diffusion-limited surfactant adsorption is restored. As predicted by the depletion interaction model (Eqn. 5), ln (RA) ∝ MW^0.1 and 1 wt. % PEG (Fig. 4, circles) completely reverses the albumin inhibition as predicted by Eqns. 1-5. For 0.5% wt. PEG (Fig. 4, circles), none of the PEG molecular weights provide a relative adsorption greater than 0.5, consistent with the concentration dependence of the depletion attraction [13]. Similar ln (RA) ∝ MW^0.1 scaling is apparent for PEG MW 6 kDa-35 kDa at 0.5 wt. % PEG.

In Region III, RA drops from Region II for both 1 and 0.5% wt., in contrast to the scaling predicted by Eqn. 5. From Table 1, the ratio of polymer solution viscosity to the solvent (water) viscosity, η/η_sol, for 1% wt. polymers in Region III exceeds the overlap criterion (η/η_sol > 2) for these PEG molecular weights. As the polymer solution approaches the overlap concentration and crosses into the semi-dilute regime, the polymer is no longer isolated random coils with characteristic length scale R_g but instead an entangled polymer mesh with characteristic length scale ξ. Above overlap, the Asakura and Oosawa (AO) hard sphere model for the depletion attraction must be modified by the polymer reference interaction site model (PRISM) of Fuchs and Schweizer to account for the competition between the increasing polymer concentration and decreasing polymer mesh length, ξ [35, 36]. For a dilute colloidal system with a large aspect ratio between the colloid (R ∼ 500 nm) and polymer (R_g ∼ 20 nm), PRISM predicts that the magnitude of the depletion attraction will plateau at roughly the polymer overlap concentration. 200 kDa PEG reaches its overlap at significantly lower concentrations (∼0.5% wt.) than 10 kDa PEG (∼4.0% wt.), which explains why the PEG 200 kDa depletion attraction is lower in magnitude (Table 1) and yields lower RA values for PEG molecular weights in Region III. The plateau of the depletion attraction at roughly the overlap concentration demonstrates an upper limit for the effectiveness of a polymer of a given molecular weight; any additional polymer above the overlap concentration no longer increases surfactant adsorption. Increasing the polymer concentration greatly increases the solution viscosity; as D_eff in Eqn. 1 is inversely proportional to the solution viscosity the net rate of adsorption will decrease as shown in Fig. 4 (Eqn. 1).

Our results are consistent with those of Yu et. al. who showed that increasing PEG molecular weight (from 3.35 -35 kDa) enhanced the rate of bovine lipid extract surfactant absorption to a clean interface [20], while higher molecular weight PEG (300 kDa) did not enhance adsorption. Surface force apparatus (SFA) measurements between mica-supported lipid bilayers in 10% wt. 1 kDa PEG yielded a force-distance profile similar to pure water, indicating at this low molecular weight, the PEG does not generate a depletion attraction [56]. Kuhl et. al. also showed a concentration dependant increase in the adhesion force between lipid bilayers in the SFA for solutions containing 8 or 10 kDa PEG, which quantitatively agrees with that expected for a depletion attraction between the bilayers [56]. Alig et. al. have shown the re-adsorption of anionic lipids to the interface after collapse was proportional to the subphase ionic strength, showing that an electrostatic barrier exists for anionic lipids and that this barrier could be modified via charge screening [21]. Though all experiments in this work were performed with Survanta, the depletion attraction theory predicts enhanced surfactant adsorption is independent of surfactant composition. This is consistent with results which showed that clinical and native lung surfactants of varying composition all show resistance to serum inhibition in the presence of PEG [40]. However, the theory does predict a surfactant aggregate size dependence (R) and so if surfactant preparations have different aggregate sizes, differences in polymer enhanced adsorption would be expected between different clinical and native lung surfactants.

Though the compression and expansion cycles on the Langmuir trough are slow (8 min/cycle) compared to physiological rates (3 sec/cycle), previous work performed on the pulsating bubble surfactometer shows that 10 kDa PEG reverses serum protein inhibition at physiological rates [40]. This correspondence likely holds for over the entire PEG molecular weight range and the same Relative Adsorption vs. PEG MW results presented in Fig. 4 would likely hold at higher rates. In a clinically relevant situation, the polymer is added to replacement lung surfactants and the resulting preparation is instilled intratracheally as rescue therapy after lung injury. Lung function of animals with lung injury is markedly improved when polymers are added to clinical surfactants [6, 7, 27, 29]. In our experiments, the PEG is added to the subphase instead of the surfactant to ensure a uniform PEG subphase concentration. There is good correspondence between the concentrations of PEG necessary to reverse serum protein inhibition in our Langmuir trough experiments (1% wt.) and PEG concentration that is utilized in animal injury models (5% wt.) [6]. While the polymer concentration is higher in the animal injury models, this concentration is based on the instilled surfactant volume (∼1.2 mL [6]) and will be reduced by the lining fluid in the airways and alveoli.

An additional factor to consider in future treatments is that the osmotic pressure of the polymer-surfactant solution must be minimized to prevent infiltration of liquid into the lungs during any potential ARDS treatment [57]. Hence, the lowest concentration of the highest molecular weight polymer that provides the necessary inhibition reversal should be used; the optimal PEG molecular weight for surfactant inhibition reversal is therefore likely to be around 35 kDa; which provides the greatest depletion attraction/surfactant adsorption with the smallest osmotic pressure [28]. A caveat to this optimal molecular weight is that it depends on the concentration of the polymer; the use of higher polymer concentrations (>1% PEG) results in overlap and a corresponding plateau in enhanced surfactant adsorption at a lower polymer molecular weight. A further advantage of using 35 kDa PEG compared to smaller molecular weights is the increased range of the depletion attraction (2R_g). Other serum proteins such as fibrinogen have also been shown to competitively adsorb with lung surfactant lipids; fibrinogen is much larger than albumin with dimensions of 5 × 5 × 46 nm [58]. While the exact orientation of fibrinogen at the air-liquid interface is unknown, the higher molecular weight PEG can likely generate a depletion attraction with sufficient range and magnitude to enhance surfactant adsorption. Considering all of these factors, the data suggest that for a given polymer concentration, using the largest molecular weight polymer which does not reach the overlap concentration will optimally enhance surfactant adsorption with smallest osmotic pressure and largest range.

The results presented here show that for a given polymer and surface-active inhibitor, an optimum concentration/molecular weight exists for enhancing surfactant adsorption and reversing inhibition. While molecular weights of 6 -35 kDa are optimal for PEG, other polymers may exhibit optimum molecular weights which are larger or smaller depending on composition, overlap concentration and charge. For example, hyaluronan (HA), an anionic natural polysaccharide that is secreted by alveolar epithelial cells, of molecular weight 100-1240 kDa has been shown to reverse surfactant inhibition in vitro [25] at much lower concentrations than PEG. If both surfactant aggregates and polymer have the same charge (net anionic), the effective depletion layer can extend further. The electrostatic repulsion between the large and small particles increases the effective radius of both the surfactant aggregates and polymer by roughly the Debye length, additionally increasing surfactant adsorption [17]. HA occurs in the lung epithelial fluid at concentrations of 4000 μg/L with a molecular weight of 220 kDa, in contrast to the 2000 kDa HA in cartilage and 7000 kDa HA in synovial fluid [59], suggesting an optimized HA molecular weight in the lung. During lung injury and disease, HA can be broken down by enzymatic action to produce smaller molecular weight fragments (1.6 kDa-10 kDa) [59]. However, similar to Region I of Fig. 4, these fragments may also generate depletion forces that lack sufficient range to insure surfactant adsorption, especially in the presence of serum proteins. Our results demonstrate the existence of an optimum PEG molecular weight range for depletion-enhanced surfactant adsorption and suggest that deviating from that optimum range can have deleterious results, especially in the case of existing epithelial fluid constituents such as HA.

Conclusions

Fluorescence images confirm that albumin on the interface prevents Survanta bilayer aggregates from reaching the air-water interface, which appears to be necessary to initiate the conversion from bilayers to monolayers. Survanta can approach the albumin-covered interface, but without the added push from the polymer-induced depletion attraction, Survanta remains below the interface indefinitely and never spreads to form a monolayer. However, once the Survanta has broken through to the albumin monolayer, albumin and surfactant coexist at the air-water interface in immiscible domains. The albumin domains remain at the interface up to surface pressures (∼55-65 mN/m) well above the equilibrium surface pressure of a pure albumin film (∼ 20 mN/m) before finally being squeezed-out from the interface. Survanta aggregates remain tethered to the interface even after monolayer formation, providing a reservoir of surfactant for replenishing the interface on expansion. In the presence of polymer, the Survanta aggregates are significantly larger than without polymer; the depletion attraction leads to surfactant aggregation and flocculation in solution in addition to enhancing surface adsorption [17, 40]. The larger reservoir of surfactant at the interface likely also acts to enhance surfactant adsorption, even in the absence of albumin inhibition, as observed by Yu et al. [20].

The dependence of surfactant adsorption on the molecular weight of PEG can be quantitatively described using the depletion attraction-modulated energy barrier to surfactant diffusion. Intermediate molecular weight PEG’s (6 - 35 kDa) provide the optimal reversal of surfactant inhibition induced by the competitive adsorption of albumin. For smaller molecular weight PEG’s, the range of the depletion attraction (2R_g) is insufficient and surfactant adsorption is not significantly increased. At 1% wt., PEG greater than 35 kDa reaches the overlap concentration resulting in decreasing surfactant adsorption; at the overlap concentration the simple model of depletion attraction is no longer valid and the magnitude of the attraction decreases. Fluorescence images confirm that albumin on the interface prevents Survanta bilayer aggregates from reaching the air-water interface; Survanta only breaks through the albumin interface with the added push from the polymer-induced depletion attraction. After Survanta breakthrough, the albumin and surfactant coexist at the air-water interface in immiscible domains; surface pressures well above the normal equilibrium surface pressure of albumin are necessary to force the albumin from the interface during compression.

Reversing surfactant inhibition due to the competitive adsorption of serum proteins is likely one of many necessary steps to address the treatment of ARDS. However, the physiological processes that accompany ARDS development and severity are consistent with a process of inhibition of surfactant adsorption by surface active contaminants in which: (1) surface active proteins resulting from the combination of inflammation and increased alveolar epithelial permeability adsorb quickly to the alveolar interface, (2) the protein layer creates an electrostatic and steric energy barrier that exponentially decreases surfactant adsorption (3) less surfactant adsorption means that low surface tensions cannot be reached during normal breathing resulting in increased lung injury and more inflammation; and (4) hydrophilic, non-adsorbing polymers can provide sufficient depletion attraction to overcome the energy barrier and re-establish normal surfactant adsorption. The attractive depletion forces generated by the hydrophilic polymers are concentration and molecular weight dependent and balance the repulsive serum protein induced steric and electrostatic interactions to restore diffusion limited surfactant adsorption.