An evenly distributed tear film sustains a proper degree of hydration of the cornea, contributing to overall ocular health and comfort.1–11 Specifically, the tear lipid layer plays an integral role in maintaining ocular surface homeostasis by facilitating the spreading of the aqueous layer over the cornea after a blink.1,2,6,8,10 The interactions between the lipid layer and tear fluid constituents (e.g., proteins) are essential for maintaining tear film stability.1–11 The human tear film is a complex biological system, and its composition must be delicately balanced to maintain its optimal functions.1–4,7,10 Unfortunately, this intricate balance of essential components can be easily disrupted by many factors, including contact lenses and lens care solutions. For example, many multipurpose solutions (MPS) contain surface-active ingredients (i.e., surfactants) added to enhance their cleaning efficiency and lens-surface wettability. Some surfactants may destabilize the superficial lipid layer of the tear film by lowering the oil-water interfacial tension to nearly zero, which leads to emulsification of the lipid layer of the tear film.1 Therefore, these surfactants introduced into an eye upon lens insertion can weaken already susceptible prelens tear film.1,12
The stability and spreading ability of thin liquid films in general, tear films included, are governed by the balance of the interfacial forces acting at a three-phase contact line, namely, the tensions at liquid-air, liquid-solid, and solid-air interfaces.1,13 Therefore, evaluation of the equilibrium surface tension/pressure at the air- aqueous interface is important for understanding tear film stability and spreading behavior.1,3–5,9–11 Studies involving tear lipids confront a great challenge because only tiny amounts of the material can be collected harmlessly from a single human eye. As a result,6 most studies have focused on model or animal meibomian lipid monolayers (one molecule thick), and their equilibrium surface pressures have been investigated using the Langmuir trough technique. However, recent findings show that human tear lipids form multilayers, not monolayers, and are of a 20-molecule thickness at minimum.14–17 Several investigations of tear lipids, carried out in the presence of model proteins (e.g., lysozyme, mucin, and lactoferrin), have shown that proteins change the surface pressure of lipid monolayers and can penetrate into Meibomian lipid monolayers.18–22 As all the aforementioned studies have emphasized, the interaction between tear proteins and lipids can affect the equilibrium surface tension of the lipid layers, thus suggesting that a similar mechanism could impact tear film stability in vivo. However, human tear film stability is likely dependent not only on the equilibrium surface tension but also on the mechanical properties of films operating under dynamic conditions and high stresses exerted by the eyelids in a blinking cycle.23,24 Recent clinical studies have shown that in vivo relaxation of tear lipid layers after a blink can be described as a viscoelastic process.24 Unfortunately, the information regarding interfacial viscoelastic properties of lipid films is scarce.23 Clearly, to advance the understanding of human tear lipid film dynamics and the role of proteins in the mechanical properties of complex tear films, a new experimental approach was needed to create thick tear lipid films at an air-aqueous interface and then to investigate the interfacial properties of these multilayered films under dynamic conditions that mimic human eye conditions more realistically.
In contrast with most previous studies that investigated lipid monolayers at equilibrium, we have developed an experimental technique allowing assessment of both the equilibrium and the dynamic interfacial properties of lipid multilayers with accurately controlled and varying thickness. This was achieved by using sessile bubble tensiometry, in which the air-aqueous interface area was 15 to 20 mm2, approximately 300 times smaller than that in a typical Langmuir trough, and a few micrograms of ex vivo human tear lipids were sufficient to deposit thick films. With this method, we investigated the equilibrium surface tension and dynamic mechanical properties of thick lipid films by studying relaxation after instantaneous expansion and compression (i.e., dilatational interfacial rheology), mimicking human blink cycles. We also examined how the interaction of lipid multilayers with the model tear protein lysozyme influenced their mechanical properties. Lysozyme was chosen not only because it is the most abundant protein in human tears25,26 but also because it has recently been shown that interaction between model lipids and lysozyme alters the mechanical properties of model lipid films.24 In addition to examining the effect of lysozyme on the dynamic properties of lipid multilayers, this novel experimental approach provided information regarding the possible mechanisms involved in clinically observed human tear film destabilization and related discomfort induced by surfactant- containing MPS during contact lens wear.12
MATERIALS AND METHODS
Distilled and deionized water from a MilliQ filter system (Millipore Co., Bedford, WA, resistivity >18 MΩ cm) was used for solutions preparation. The aqueous phase in all experiments was buffered saline solution composed of 5 g/l NaCl with 4 g/l of sea salts (Sigma, Milwakee, WI) added to supplement the aqueous phase with other ions found in human tears (e.g., K+, Ca2+, and Mg2+, phosphate and bicarbonate), which are essential for maintenance of the corneal epithelial surface.27 The pH of the solution was adjusted to 7.3 by addition of small amount (e.g., 0.5 to 0.8 ml) of 250 M KH2PO4.
The protein studied was hen egg-white lysozyme from Sigma-Aldrich Co. (Milwakee, WI), used without further purification. Lysozyme has a relatively low molecular weight of 14 kDa; its isoelectric point is 11; and therefore, the lysozyme net charge was positive under experimental conditions.
Lipids Extraction Procedure
Tear lipids were extracted from lotrafilcon A [Focus Night & Day (PreAqua formulation); CIBA] contact lenses that had been worn by 25 neophyte human subjects. Each participant first went through a period of adaptation to the day-wear modality with lotrafilcon A lenses, followed by 1 month of continuous wear. All subjects were recruited from the campus of the University of California at Berkeley and were free from ocular disease or any ocular abnormality that contraindicated contact lens wear. A thorough explanation of the study protocols was given to each prospective subject, and informed consent was obtained thereafter. This study observed the tenets of the Declaration of Helsinki and was approved by the University of California, Berkeley Committee for Protection of Human Subjects.
Lotrafilcon A lenses were chosen for this study because they are approved by the Food and Drug Administration for continuous overnight wear and because these lenses were reported by the manufacturer (CIBA) to be free of any surface-active modifiers; the blister packaging solution has a surface tension of 71.5 mN/m, which is close to pure water, as confirmed by our measurements. Furthermore, it has been shown that lotrafilcon A lenses worn for 30 days of continuous wear accumulated up to 640 μg of tear lipids per lens.28 Taking all these factors into consideration, we concluded that lotrafilcon A lenses worn for 30 days would be a suitable source of tear lipids in an amount sufficient for the interfacial properties studies using sessile bubble tensiometry.
The worn lenses were collected by research optometrists wearing powder-free gloves. The lenses were rinsed in deionized distilled water and blotted with filter paper. Each lens was placed in a separate 20-ml glass scintillation bottle containing 4 ml of solvent. The solvent used for extraction consisted of five parts of toluene and one part of isopropyl alcohol (5:1 v/v), both were supplied from Sigma, spectroscopic grade. Toluene is widely used for lipid extraction from tissues, and systematic studies of the solubility of lipids in a toluene-ethanol mixture have shown that this combination has superior lipid-extraction properties compared with chloroform-methanol.29 The lenses immersed in solvent were subjected to 30-min sonication in a water-filled bath. The solvent with extracted lipids was then transferred into smaller (5 ml) clean glass vials and evaporated under vacuum. Special care was taken to avoid contact of solvent and lipid extracts with human skin, any plastic parts, or gloves. All glassware was cleaned with saturated KOH solution in ethanol, copiously rinsed first with distilled and then with deionized water, and dried under vacuum before use.
Dry lipid extracts were stored in a freezer at −20°C. Extracted samples were redissolved in 200 μl of the same solvent before use. Fifteen unworn blank lotrafilcon A lenses identical to the lenses distributed to the subjects (from the same lot number) were subjected to exactly the same procedures. The reconstituted extracts from blank unworn lenses were tested for dry residues using ellipsometry and for surface activity using tensiometry in exactly the same manner as with the lipid extracts. The dry residues from all blank lenses were <0.5 nm in thickness, practically indistinguishable from the dry residues left after evaporation of an equal amount of pure solvent. For all blank extract samples, the surface tension was 71.6 ± 0.8 mN/m, and it remained constant for 24 h after deposition and solvent washout. The blank experiments confirmed that no dissolution or degradation of the lotrafilcon A lenses occurred during treatment with solvent and that the lenses did not contain any surface-active ingredients, which would contaminate the collected lipid extracts.
Lipid Film Thickness Measurement
An SE-400 Ellipsometer (Sentec Instruments GmbH, Germany) was used to measure the thickness of the lipid films. Five drops of Millipore water, 10 μl each in volume, were placed on a silicon wafer with a 2-nm thick layer of native SiO2. A 1-μl drop of the reconstituted lipid extract was then deposited and spread over the surface of each water droplet. The silicon wafers were kept in covered Petri dishes to reduce the evaporation rate and thus to get relatively uniform lipid film thickness. After slow complete evaporation of both solvent and aqueous drops, the thickness of the lipid films left on the silicon wafer surface after water evaporation was measured with an ellipsometer across the entire film area (total number of measurements was 50 to 75 points) and averaged. The variations in thickness for each film did not exceed 15%. The mean thickness of the dry remnants left by pure solvent drops evaporated from the water droplet surface was used as a baseline. The average thickness of the lipid films varied from 8 to 25 nm between extract samples. On the basis of the results of the ellipsometric measurements, we calculated the amount of the lipid-extract solution necessary to build the interfacial layer around an air bubble of known surface area with a thickness of 8 to 12 nm and a surface tension of 25 ± 1.2 mN/m.
To deposit a multilayer with varying thickness up to 30 nm, the lipid extracts collected from six worn lenses were pooled together and then reconstituted in 200 μl of solvent.
There are several methods available to measure surface tension. A Wilhelmy plate is used in the Langmuir trough technique for tension measurements at a planar air-water interface,6,20,21 whereas automatic axisymmetic drop/bubble shape analysis is used for tension determination at the curved surfaces of pendant drop18,30 or sessile bubble.31–33 Miano et al.18 used the pendant drop technique to investigate the interaction of model tear proteins with monolayers of bovine meibomian lipids. Others have used the Langmuir trough with a Wilhelmy plate to examine the physical properties of lipids.34,19–21
In this study, the sessile bubble configuration was used. A small air bubble was formed at the underside of a straight hydrophobic capillary vertically immersed into an aqueous phase, and lipids were deposited on the surface of this bubble. A similar method was used in earlier studies investigating surface activity and protein-lipid interactions in pulmonary surfactant systems.30,31 The major advantage of the sessile bubble in comparison with the widely used Langmuir trough coupled with a Wilhelmy plate is that the surface area of the sessile bubble can be very small (i.e., 15 ± 2 mm2 in this study). Thus, a few micrograms of lipids were sufficient to form multilayers up to 30 nm thick. The advantage of the sessile bubble configuration over the pendant drop was faster equilibration of the interface because of continuous stirring of the aqueous phase and larger volume-to-area ratio.32,33 Another advantage was high stability of the bubbles, when compared with the pendant drop, especially at low surface tension. In addition, this method eliminated the influence of aqueous phase evaporation. Lastly, the sessile bubble technique was suitable for studies of the interfacial dilatational rheology by the step-strain relaxation method.32,33,35–37
The sessile bubble setup was augmented with a bulk phase exchange option allowing displacement (washout) of one aqueous phase (for instance, MPS solution) in the measuring cell with another fluid such as a pure buffer solution while maintaining the interface intact during displacement.36 This “washout” option was used to validate the irreversible character of lysozyme binding to the reconstituted ex vivo lipids and to evaluate the effect of MPS on the lipid multilayer's interfacial properties.
A Ramé-Hart tensiometer (Ramé-Hart Instrument Co., Netcong, NJ) with DropImage Advanced (v.2.2) software and automated dispensing system was used for real-time surface tension data acquisition. Fig. 1 displays the schematics of the experimental setup. A magnified image (1 μm = 10 pixels) of a bubble was displayed on the computer screen. A more detailed description of this experimental setup is available elsewhere.32,33
First, 18 to 20 ml of aqueous phase was placed into an optical glass cell (Hellma, NY). The precalculated volume (typically in the range 1 to 2 μl) of lipid extract was manually deposited onto the air bubble surface (15 ± 2 mm2) from underneath using a 10 μl high-precision glass syringe (Hamilton, NV) to produce the lipid film of desired thickness. The aqueous phase in a cell was stirred to provide uniform lipid distribution at the air-aqueous interface and to accelerate solvent dissolution into the aqueous phase. The aqueous phase was then displaced with pure buffered saline to remove all traces of organic solvents. After this first washout, the lipid-coated bubble was left to rest for 17 to 24 h (overnight) without stirring. For elasticity studies, lipid films with constant initial surface tension 25 ± 1 mN/m were deposited; the working concentration of the lipid extracts, assuming lipid density 0.9 μg/μm3, was 0.8 to 1.5 μg/μm3; and resulting lipid film thickness at the air-liquid interface was 8 to 12 nm. After the lipid film was aged overnight, the model tear protein lysozyme solution, 40 to 50 mg in 2 ml of buffered saline, was injected into a cell and stirred for additional 16 to 19 h. The working concentration of lysozyme in the cell was 2 to 2.5 mg/ml, which was within the concentration range found in normal human tears.25,26 Interfacial tension was monitored during these experimental procedures. After lysozyme-lipid film equilibration, the rheological parameters of the mixed layers were determined.
The transient step-strain relaxation technique was used to investigate the interfacial dilatational rheological properties of the individual and mixed lysozyme-lipid layers.32,33,35–37 An air-aqueous interface, previously equilibrated with a surface-active substance (lipids, protein, lipid-protein complex, or MPS) was instantaneously expanded or contracted, then held constant while the interfacial stress (tension) was measured. The amplitude of the interfacial area change (ΔA) was <10% of the initial surface area (A0). After the area perturbation, the interfacial tension was measured for 1 h during which it relaxed toward the equilibrium value, decreasing after expansion and increasing after contraction. This time was sufficient for the tension to reach almost constant value. The transient elasticity E(t) and its decay with time were then determined as follows:
where Δγ(t) is the change in surface tension induced by an area change. Single-exponential decay (Eq. 2) was fit (KaleidaGraph, version 4.03, software, Least Squares method, 100 iterations using all 1860 data points collected during an experimental run) to the experimentally observed transient elasticity, and the viscoelastic parameters of the interfacial layer were estimated according to the Maxwell model35–37
where τr represents the characteristic time of viscoelastic relaxation, E∞ is the elasticity at time approaching infinity or storage modulus, and A1 is the amplitude of response to the initial surface perturbation. The response of each interfacial film to the area expansion and contraction was measured 10 times.
At this initial exploratory stage of the project, all the experiments were performed at ambient temperature, 22 ± 1°C.
Surface Tension Measurements
Surface tension of the reconstituted ex vivo lipid films was first studied as a function of film thickness. Fig. 2 shows that the surface tension of the lipid layers was reduced from 32 to 22.5 mN/m as the layer thickness increased from 2 to 15 nm and did not change with further thickening of the lipid film.
Fig. 3 shows the surface tension history at the air-saline interface after injection of lysozyme with continuous stirring of the aqueous phase; the working concentration of lysozyme in the cell in this experiment was 2 mg/ml. The surface tension decreased from 72.5 to 53 mN/m during the first hour, then continued to descend slowly for over 16 h. After 16 to 24 h of interface aging, the final tension was 50 ± 1 mN/m. This experiment was repeated for lysozyme concentrations ranging from 2.0 to 2.5 mg/ml, and the results (not shown here) suggested that the final surface tension was independent of the concentration. This observation was in agreement with data reported in the literature.4,5
In these experiments, after 16 h of equilibration, the aqueous lysozyme solution in the cell was displaced with buffered saline. The washout provided complete protein removal from the bulk phase inside the cell.30,31 The surface tension did not change after washout and remained 50 ± 1 mN/m, thus indicating an irreversible character of the protein binding to the air-water interface.
Fig. 4 illustrates the surface tension history of the mixed lipid-protein layer. An arrow indicates the moment when lysozyme was injected into the cell. Surface tension decreased by 1.5 mN/m immediately after lysozyme injection, then increased back to 26 mN/m and remained constant for another 10 h of measurements.
The results of the surface tension measurements during the injection, 30-min equilibration, and then washout of opti-free express (OFX) are presented as a function of time in Fig. 5. In this set of experiments, an air bubble was formed in 19 ml of buffered saline with a surface tension of 72.5 mN/m. A few minutes later, as indicated by the arrow in Fig. 5, 1 ml of OFX was injected into the cell with continuous stirring for 30 min. The surface tension dropped to approximately 40 mN/m almost instantaneously upon OFX injection. Continuous stirring and buffer flow through the cell caused the scatter (±1 mN/m during washout) of the surface-tension data during the washout period shown in Fig. 5. Before washout, stirring was stopped for 2 to 3 min, and the surface tension at the air-OFX (1:20 dilute) interface was found to be 36.5 ± 0.5 mN/m. This value was in agreement with that reported by Dalton et al.38 for undiluted OFX. Stirring was then resumed, and 500 ml of buffered saline was pumped through the cell. The final tension at the air-saline interface increased by 3 mN/m after OFX was washed out from the bulk phase but still remained low at 39.5 ± 0.5 mN/m. Similar behavior was observed upon injection and washout of opti-free replenish (OFR). The surface tension for OFR (1:20 dilute) was 35.2 ± 0.5 mN/m, which then increased to 38.5 ± 0.5 mN/m after washout. The surface tension at the air-aqueous interface after a complete washout of both OFX and OFR from a bulk solution remained much lower than 72.5 mN/m, which is typical for a pristine air-water interface. These observations suggested that both MPSs used in this study contained some irreversibly adsorbing components, which remained attached to the interface after bulk-surfactant washout.
Fig. 6 illustrates the surface tension history of the mixed lipid-protein layer that was exposed to OFX for 30 min then rinsed with buffered saline. Injection of OFX into the cell caused a slight decrease (∼1.5 mN/m) of the mixed film surface tension. After displacement of the aqueous phase with pure buffered saline, the surface tension returned to its original value of 26.1 ± 0.3 mN/m. The same trend was observed upon injection of OFR followed by washout with buffered saline. Unchanged surface tensions indicated that MPS did not displace, solubilize, or destroy the predeposited ex vivo lipids during the 30-min exposure. However, these observations do not provide sufficient information regarding specific interactions between MPS and the mixed lipid-protein films.
To acquire more information in relation to the interactions between MPS and lipid-protein complexes, we used the interfacial step-strain relaxation method to study in detail the rheological properties of the layers formed by each of these substances alone and of the mixed interfacial layers. Rheological parameters are often used to characterize the interfacial properties of thin films under dynamic conditions and are relevant for many practical applications where the interfacial film stability is involved.33,35–37
Interfacial Elasticity Measurements
The interfacial viscoelastic behavior of the individual layers formed by lysozyme and the reconstituted ex vivo lipids was studied. Fig. 7 shows the interfacial elasticity history of the layer formed by lysozyme (bulk concentration 2.5 mg/ml), before (Fig. 7a) and after (Fig. 7b) washout of protein from the bulk solutions, which served as control experiments for further studies of MPSs influence on the interfacial elasticity of mixed lipid-protein films.
Fig. 7a shows that before washout (i.e., in the presence of lysozyme in a bulk solution), the responses of the interfacial protein layer to expansion and contraction of surface area completely coincide. The single-exponential decay model (Eq. 2) was fit to the experimental data, and rheological parameters were thus estimated. E∞ was found to be 25.6 ± 0.5 mN/m, and τ was 460 ± 20 s for both expansion and contraction of the protein interfacial layers.
As seen in Fig 7b, protein washout from the bulk phase changed the viscoelastic response of the interfacial layer, and for postwash dilatations, some hysteresis arose between the expansion and contraction steps. E∞ increased after washout, becoming 34.5 ± 0.5 mN/m and 33.2 ± 0.5 mN/m for expansion and contraction, respectively. The postwash relaxation time τ also increased from 460 to 760 ± 10 s and 630 ± 15 s for expansion and contraction, respectively.
Fig. 8 depicts typical transient elasticity vs. time curves for the interfacial layers formed by reconstituted ex vivo lipids (curve 1), for the same lipids after injection and equilibration of this layer with lysozyme (curve 2), and after washout of the mixed layer (curve 3). Elastic response of the lipid layer alone was independent of whether expansion or contraction was applied, i.e., no hysteresis was observed. E∞ of the ex vivo lipid layers was close to that of the lysozyme, 25.3 ± 2.5 mN/m; τ for the lipid multilayers was 90 ± 25 s, 6 to 7 times shorter than that of the lysozyme layers. For the mixed ex vivo lipid-lysozyme interfacial films, E∞ was found to be 24.8 ± 3.7 mN/m; it did not change upon lysozyme adsorption at the lipid-aqueous interface. This result is as expected, as both of the components, lipids and lysozyme, have similar values of E∞. However, the relaxation time τ for the mixed lipid-protein films was 250 ± 35 s, 2.7 times higher than for the lipids alone and 1.8 times lower than for lysozyme alone. In the control run (curve 3), where a lipid-protein mixed film was washed with buffered saline without exposure to MPS, E∞ increased slightly by 2 ± 1.2 mN/m, after protein removal from the bulk phase.
Fig. 9 depicts an example of the transient elasticity of a postwash OFX interfacial layer at the air-water interface, without preadsorbed lipids or protein. One can see from this plot that the interfacial rheological behavior of postwash OFX was distinctly different from that of the lipids, protein, or the mixed lipid-protein layer. The OFX layer exhibited a threefold lower E∞ than lysozyme or lipids. For both postwash OFX and OFR layers, E∞ was 8.3 ± 0.5 mN/m. In contrast, τ was 950 ±50 s for the postwash MPS layers; these layers relaxed 2 times slower than protein and 9.5 times slower than the lipids. MPS formed relatively weak, viscous layers relaxing slowly after expansion or contraction of the interface.
Fig. 10 shows the elasticity history of the mixed lipid-protein layers first exposed to either OFX or OFR for 30 min, then washed with buffered saline. MPS induced noticeable changes in the viscoelastic behavior of the mixed lipid-protein layers. After exposure to OFX, the postwash E∞ of the mixed layers was reduced from approximately 25 ± 2 to 13 ± 4.5 mN/m (p < 0.001, pair t test), almost equal to the value observed for postwash OFX alone. The τ of the films treated with OFX increased from 250 to 450 ± 30 s (p < 0.001, pair t test), suggesting that some of OFX remained adsorbed onto the initial predeposited lipid-protein layer.
Exposure of the mixed lipid-protein films to OFR produced a weaker effect on their viscoelastic behavior. The E∞ of the mixed layers was reduced from 25.6 to 21 ± 3.0 mN/m after 30-min exposure to OFR (p = 0.004, pair t test). In contrast, τ increased almost threefold to 730 ± 55 s (p < 0.001, pair t test). Once again, these results suggest that the surface-active components of MPS partially displaced the lysozyme and became irreversibly attached to the predeposited lipid-protein layer.
Previous investigations of human tear constituents and whole tear surface properties such as surface tension have shown that the tear lipid layer plays a significant role in tear film surface tension reduction and tear film stability.4–6,10,11 McCulley and Shine15,16 have developed a composition-based model for the thick tear lipid layer, proposing a complex stratified multilayered structure. According to this model, polar amphiphilic lipids (mostly phospholipids comprising 6 to 16 percent of the total lipids) with surfactant characteristics serve as a structure upon which the stability of tear lipid film is largely dependent, although most of the tear lipid film consists of non-polar components (84 to 94%). The thickness of the tear lipid layer is one of its most extensively studied properties, and the estimates range from 13 to 100 nm depending on the method used for measurements.14,17 These measurements were consistent with the thick multilayer model of McCulley and Shine.15,16 The thickness of a dipalmitoyl phosphatidyl choline (DPPC, a phospholipid, which is the major constituent of pulmonary surfactant and is most commonly used as a model for tear lipids) monolayer was in the range 2 to 2.5 nm as determined by neutron reflectivity measurements.22 Even at the lowest estimated thickness of 13 nm, the tear lipid film is at least 6 times thicker than a monolayer; therefore, monolayers of meibomian lipid previously investigated may not be an adequate model of the human tear lipid layer. In a subsequent investigation, Shine and McCulley39 conducted a detailed study of polar lipids in human meibomian secretion and suggested that the polar lipid layer was most likely 1 to 3 molecules thick and served as a surfactant between the aqueous tear and the nonpolar part of the lipid layer. Recently, Butovich et al.17,34 summarized the most current data regarding the structure and chemical composition of meibomian and tear lipids. The authors proposed that tear lipids form a thick blanket, with a minimum thickness of 20 molecules, over the aqueous layer of the tear film.
The interfacial pressure of meibomian lipid monolayers has been examined in a number of articles.6,18–22 In these publications, lipid monolayers were chosen to model the tear lipid layer and to investigate in greater detail the interaction of lipid monolayers with model tear proteins. Tragoulias et al.20 found that model tear proteins possessed significant surface activity; from these results, the authors concluded that tear proteins were the major contributors to the surface pressure of a tear film. Several studies18–20 have explored monolayers of meibomian lipids at relatively low surface pressures corresponding to low surface coverage by polar lipids. Under these conditions, proteins were able to penetrate into the lipid monolayer and reduce the surface tension.18
In this study, we focused on the interfacial properties of relatively thick ex vivo human lipid multilayers. We believe that the behaviors of thick lipid films provide a more realistic model of the human tear lipid layer than the lipid monolayers studied previously. The results shown in Fig. 2 clearly indicate that surface tension strongly depended on ex vivo lipid layer thickness. The thinnest lipid film we studied was a 2 nm thick monolayer. The surface tension of this layer was close to 32 mN/m, corresponding to the surface pressure of 40 mN/m reported by other researchers.18–20 However, as the thickness of the deposited lipid layers increased, the surface tension decreased until it reached a constant value of 22 ± 1 mN/m at a lipid layer thickness of 15 nm or greater. These results suggested that the saturation, or coverage, of the lipid-aqueous interface with polar lipids increased in concert with the total ex vivo lipid film thickness. At a total lipid thickness between 15 and 25 nm, when a completely saturated, densely packed polar lipid layer was formed, the films exhibited a very high surface pressure of ∼ 50 mN/m, which is in agreement with the values found for condensed phospholipid layers and lung surfactants.40,41 The important implication for lipid film dynamic behavior in vivo arises from these observations. The curve in Fig. 2 can be dissected into two linear parts: one for thin films between 2 and 10 nm, and the second for films thicker than 20 nm. The first part of the linear fit gives a surface tension gradient, dγ/dh = 1.0 (mN/m)/nm, whereas for the second part, the tension gradient is 30 times lower, 0.033 (mN/m)/nm. The tension gradients are known to be the driving force for the mass transfer along a liquid-air interface, which causes flows in liquid films. This phenomenon is known as the Marangoni effect. In thick lipid films, small tension gradients lead to sluggish flows and slow film thickness rearrangement. Therefore, clinically observed interference patterns24 persist in the case of a thick lipid film. However, in thinner lipid films, a slightest decrease in the film thickness gives rise to high surface tension gradients, resulting in strong Marangoni flows, which lead to thickness reequilibration.
As shown in Fig. 4, lysozyme injected into the aqueous phase surrounding an ex vivo lipid film did not significantly change the surface tension at the air-aqueous interface even after long-time equilibration with continuous stirring. This observation was consistent with the results reported by Miano et al.22 for a mixed DPPC-lactoferrin layer at an initial DPPC surface pressure of 35 mN/m or higher. In that study, neutron reflection measurements were performed on a DPPC-lactoferrin layer, and it was found that at high surface pressures, protein adsorption into the lipid layer was very low. The results of our surface tension measurements led us to the conclusion that lysozyme, similar to lactoferrin, could not penetrate into or displace the condensed, thick layer of reconstituted ex vivo lipids. The evidence of lipid-lysozyme interaction and formation of an interfacial lipid-lysozyme complex in this study was obtained using an interfacial relaxation technique in combination with a washout option.
We found significant differences between the interfacial rheological parameters of the ex vivo lipid layers alone and of the mixed lipid-lysozyme layers. These differences persisted even after complete protein removal from the bulk phase, indicating that lysozyme was irreversibly bound to the lipid layer. At high surface pressures, the presence of lysozyme in the mixed interfacial layers was undetectable by surface tension measurements. However, it manifested itself by an approximately twofold increase in mixed film relaxation time compared with that of a layer formed by ex vivo lipids alone. The interfacial rheological properties of the reconstituted ex vivo lipids in the thick mixed layers were much more sensitive indicators of their interaction with protein than were surface tension or surface pressure. Changes in interfacial rheological parameters after protein washout were most likely caused by the removal of loosely bound protein molecules from the adsorbed interfacial layer, leading to increased interfacial viscosity.
The adsorption of lysozyme into the ex vivo lipid layers was likely driven by electrostatic interactions. Lysozyme has an isoelectric point of 11 and a net positive charge at pH 7.3. Polar tear lipids are known to contain a significant amount of the species with negatively charged head groups, such as anionic phospholipids and fatty acids.15,17 These anionic head groups are likely to interact with the positively charged groups of the lysozyme molecules, forming complexes through an ion-pairing mechanism.
Lysozyme is one of the most abundant proteins in human tears.25,26 We hypothesize that the binding of lysozyme molecules to anionic lipid head groups may play a very important role in maintaining tear film integrity. Lysozyme and other tear proteins, coupled with polar lipids in the tear lipid film, may provide additional tethering and stronger linkage between the hydrophobic lipid layer and the underlying tear aqueous.
MPS, with its relatively high concentration (0.1% by weight) of Tetronic T1304 and other surface-active ingredients reduced the surface tension at the air-aqueous interface by ∼32 to 34 mN/m. Our previous studies of Pluronics (nonionic surface-active polyethylene-polypropylene block copolymers) adsorption at air-water and water-mineral oil interfaces have shown that these polymers bind irreversibly to the interfaces.32 At the water-oil interface, the polymeric surfactants were found to compete successfully with a synthetic double-tailed surfactant Aerosol OT (bis-ethyl-hexyl sulfosuccinate, sodium salt), whose structure resembles that of natural phospholipids.33 Tetronic T1304 also belongs to the family of nonionic surface-active polyethylene-polypropylene block copolymers. Several experiments conducted with aqueous solution containing only 0.1 wt% of Tetronic T1304 have revealed that, similar to Pluronics, Tetronic binds irreversibly to the air-aqueous interface. We hypothesize that in MPS, Tetronic T1304 is responsible for an irreversible reduction of surface tension at the air-solution interface. Moreover, interfacial rheology studies conducted with mixed lipid-lysozyme layers exposed to MPS (Fig. 10) have shown that both OFX and OFR alter the rheological characteristics of mixed lipid-lysozyme films. The changes caused by OFX were found to be significant and persistent, whereas OFR produced milder effects. Besides Tetronic T1340 and other components present in both OFX and OFR, OFX contains an unspecified amount of AMP 95, amino propanol, which is a polar organic solvent, whereas OFR contains nonanoyl ethylenediaminetriacetic acid, which is a surface-active fatty acid. The differences in formulations between OFX and OFR, slight as they may be, were the major reason for the differences in MPS performance. It is conceivable that MPS cytotoxicity42 might be associated with high surface activity of some ingredients and their interaction with the phospholipids on corneal epithelium cells.
The results of this study show that MPS surface tension (Fig. 5) was at least 10 mN/m lower than that of lysozyme (Fig. 4), suggesting that MPS contains ingredients that are more surface active than lysozyme. The interfacial layers became more plastic after treatment with OFX (Fig. 10), indicating that OFX partially displaced protein from the mixed lipid-protein interfacial film. The difference in the responses of lipid films to exposure to OFX and OFR suggests a new explanation for documented differences in their cytotoxicities.42 It seems likely that OFX ingredients interacting with phospholipids comprising cell walls might compromise the cell-wall protective function by making the lipid membranes more permeable to potentially cytotoxic preservatives. The milder disruptive effect of OFR on lipid-protein mixed films correlates with its lower cytotoxicity level.
Tear films are complex, multilayered biological colloid systems. The composition of the tears is delicately balanced and regulated by ocular tissues to maintain optimal function under the high stress exerted by the eyelids during blinking. It is reasonable to assume that the interfacial rheological properties of healthy tear films are also optimally adjusted to provide timely and full recovery after very rapid compression (closed eye) followed by expansion (open eye). Soaps and shampoos contain highly concentrated surfactants that are much more surface active than tear proteins. These substances are able to solubilize and destroy tear lipid films,1 causing the familiar, extremely unpleasant sensation when soap or shampoo comes into contact with the ocular surface. As we have shown above, MPS containing some highly surface-active ingredients was able to displace lysozyme from mixed ex vivo lipid-protein layers, leading to alterations of lipid-protein film interfacial rheological characteristics. In an eye, the delicate natural balance of tear film interfacial properties might become compromised by MPS leaching from a lens. There are indications in the literature that highly porous soft pHEMA-MAA lenses (Acuvue 2), immersed in solution overnight, were able to accumulate up to 36 μg/lens of Tetronic and then release this surfactant into a buffered saline.43,44 A similar process may occur when a contact lens wearer inserts such a lens on an eye after overnight storage in MPS. The surface active components trapped inside the lens matrix may be slowly released into the eye, interacting with the mixed tear lipid-protein films and displacing proteins from this layer. Such displacement might produce undesirable alterations of tear film interfacial rheological properties, reduced tear film stability, and discomfort after several hours of lens wear. These findings help to explain the reduced tear film thickness and discomfort reported by patients who used OFX.12 It should be emphasized that in clinical practice, special measures must be taken to find a proper combination of lens material and lens care solution that will be the safest, most suitable, and comfortable for each patient, especially for those with compromised tear film stability and short tear break-up time.
We thank C. J. Radke, PhD, Department of Chemical Engineering, University of California, Berkeley, for providing laboratory space and access to equipment. We also thank H. Green, OD, PhD, T. Sanders, OD, and A. Graham, MS, for their valuable feedback during manuscript preparation.
Meng C. Lin
Clinical Research Center
School of Optometry
University of California
Berkeley, California 94720-2020
1. Holly FJ. Formation and rupture of the tear film. Exp Eye Res 1973;15:515–25.
2. Lemp MA, Holly FJ, Iwata S, Dohlman CH. The precorneal tear film. I. Factors in spreading and maintaining a continuous tear film over the corneal surface. Arch Ophthalmol 1970;83:89–94.
3. Holly FJ, Patten JT, Dohlman CH. Surface activity determination of aqueous tear components in dry eye patients and normals. Exp Eye Res 1977;24:479–91.
4. Tiffany JM, Winter N, Bliss G. Tear film stability and tear surface tension. Curr Eye Res 1989;8:507–15.
5. Nagyova B, Tiffany JM. Components responsible for the surface tension of human tears. Curr Eye Res 1999;19:4–11.
6. Holly FJ. Surface chemistry of tear film components analogs. J Coll Interface Sci 1974;49:221–31.
7. Lemp MA. Advances in understanding and managing dry eye disease. Am J Ophthalmol 2008;146:350–6.
8. Craig JP, Tomlinson A. Importance of the lipid layer in human tear film stability and evaporation. Optom Vis Sci 1997;74:8–13.
9. Miller D. Measurement of the surface tension of tears. Arch Ophthalmol 1969;82:368–71.
10. Bron AJ, Tiffany JM, Gouveia SM, Yokoi N, Voon LW. Functional aspects of the tear film lipid layer. Exp Eye Res 2004;78:347–60.
11. Kaercher T, Mobius D, Welt R. Biophysical characteristics of the Meibomian lipid layer under in vitro conditions. Int Ophthalmol 1992;16:167–76.
12. Nichols JJ, Mitchell GL, King-Smith PE. The impact of contact lens care solutions on the thickness of the tear film and contact lens. Cornea 2005;24:825–32.
13. de Gennes PG. Wetting: statics and dynamics. Rev Mod Phys 1985;57:827–63.
14. King-Smith PE, Fink BA, Hill RM, Koelling KW, Tiffany JM. The thickness of the tear film. Curr Eye Res 2004;29:357–68.
15. McCulley JP, Shine W. A compositional based model for the tear film lipid layer. Trans Am Ophthalmol Soc 1997;95:79–88.
16. McCulley JP, Shine WE. The lipid layer: the outer surface of the ocular surface tear film. Biosci Rep 2001;21:407–18.
17. Butovich IA, Millar TJ, Ham BM. Understanding and analyzing meibomian lipids: a review. Curr Eye Res 2008;33:405–20.
18. Miano F, Calcara M, Millar TJ, Enea V. Insertion of tear proteins into a meibomian lipids film. Colloids Surf B Biointerfaces 2005;44:49–55.
19. Tragoulias ST, Anderton PJ, Dennis GR, Miano F, Millar TJ. Surface pressure measurements of human tears and individual tear film components indicate that proteins are major contributors to the surface pressure. Cornea 2005;24:189–200.
20. Millar TJ, Tragoulias ST, Anderton PJ, Ball MS, Miano F, Dennis GR, Mudgil P. The surface activity of purified ocular mucin at the air-liquid interface and interactions with meibomian lipids. Cornea 2006;25:91–100.
21. Mudgil P, Torres M, Millar TJ. Adsorption of lysozyme to phospholipid and meibomian lipid monolayer films. Colloids Surf B Biointerfaces 2006;48:128–37.
22. Miano F, Zhao X, Lu JR, Penfold J. Coadsorption of human milk lactoferrin into the ipalmitoylglycerolphosphatidylcholine phospholipid monolayer spread at the air/water interface. Biophys J 2007;92:1254–62.
23. Nishimura SY, Magana GM, Ketelson HA, Fuller GG. Effect of lysozyme adsorption on the interfacial rheology of DPPC and cholesteryl myristate films. Langmuir 2008;24:11728–33.
24. Yokoi N, Yamada H, Mizukusa Y, Bron AJ, Tiffany JM, Kato T, Kinoshita S. Rheology of tear film lipid layer spread in normal and aqueous tear-deficient dry eyes. Invest Ophthalmol Vis Sci 2008;49:5319–24.
25. Ohashi Y, Dogru M, Tsubota K. Laboratory findings in tear fluid analysis. Clin Chim Acta 2006;369:17–28.
26. Sariri R, Khamedi A. Variations in electrophoretic tear protein pattern due to contact lens wear. J Chromatog A 2007;1161:64–6.
27. Bachman WG, Wilson G. Essential ions for maintenance of the corneal epithelial surface. Invest Ophthalmol Vis Sci 1985;26:1484–8.
28. Jones L, Senchyna M, Glasier MA, Schickler J, Forbes I, Louie D, May C. Lysozyme and lipid deposition on silicone hydrogel contact lens materials. Eye Contact Lens 2003;29:S75–9.
29. Christie WW. Preparation of lipid extracts from tissues. In: Christie WW, ed. Advances in Lipid Methodology, Vol. 2. Dundee, Scotland: Oily Press; 1993:195–213.
30. He Q, Zhang Y, Lu G, Miller R, Mohwald H, Li J. Dynamic adsorption and characterization of phospholipids and mixed phospholipids/protein layers at liquid/liquid interfaces. Adv Coll Interface Sci 2008;140:67–76.
31. Schurch S, Bachofen H, Possmayer F. Surface activity in situ, in vivo, and in the captive bubble surfactometer. Comp Biochem Physiol A Mol Integr Physiol 2001;129:195–207.
32. Svitova TF, Wetherbee MJ, Radke CJ. Dynamics of surfactant sorption at the air/water interface: continuous-flow tensiometry. J Colloid Interface Sci 2003;261:170–9.
33. Svitova TF, Radke CJ. AOT and pluronic F68 co-adsorption at fluid/fluid interfaces: a continuous flow tensiometry study. Industr Eng Chem Res 2005;44:1129–38.
34. Butovich IA. On the lipid composition of human meibum and tears: comparative analysis of nonpolar lipids. Invest Ophthalmol Vis Sci 2008;49:3779–89.
35. Lucassen J, van den Temple M. Dynamic measurements of dilational properties of a liquid interface. Chem Eng Sci 1972;27:1283–91.
36. Freer EM, Yim KS, Fuller GG, Radke CJ. Shear and dilatational relaxation mechanisms of globular and flexible proteins at the hexadecane/water interface. Langmuir 2004;20:10159–67.
37. Myrvold R, Hansen FK. Surface elasticity and viscosity from oscillating bubbles measured by automatic axisymmetric drop shape analysis. J Colloid Interface Sci 1998;207:97–105.
38. Dalton K, Subbaraman LN, Rogers R, Jones L. Physical properties of soft contact lens solutions. Optom Vis Sci 2008;85:122–8.
39. Shine WE, McCulley JP. Polar lipids in human meibomian gland secretions. Curr Eye Res 2003;26:89–94.
40. Rugonyi S, Biswas SC, Hall SB. The biophysical function of pulmonary surfactant. Respir Physiol Neurobiol 2008;163:244–55.
41. Li JB, Miller R, Vollhardt D, Weidemann G, Mohwald H. Isotherms of phospholipid monolayers measured by a pendant drop technique. Colloid Polym Sci 1996;274:995–9.
42. Siwale RC, Sills AM, Tan D, Wright AM, McKee MM. Biocompatibility of silicone hydrogel contact lens materials and contact lens care products using in vitro cytotoxicity assays. Invest Ophthalmol Vis Sci 2008;49:E-Abstract 4863.
43. Ketelson HA, Meadows DL, Stone RP. Dynamic wettability properties of a soft contact lens hydrogel. Colloids Surf B Biointerfaces 2005;40:1–9.
44. Copley KA, Wu C, Chen L, Radke CJ. Polymeric-surfactant adsorption onto and absorption into soft contact lenses. Invest Ophthalmol Vis Sci 2007;48:E-Abstract 5415.