Interaction of Ionic Surfactants with Cornea-Mimicking Anionic

Jul 25, 2011 - Department of Chemical Engineering, University of Florida, ... both in biological systems and also in several technological application...
0 downloads 0 Views 2MB Size
ARTICLE pubs.acs.org/Langmuir

Interaction of Ionic Surfactants with Cornea-Mimicking Anionic Liposomes Chhavi Gupta, Andrew K. Daechsel, and Anuj Chauhan* Department of Chemical Engineering, University of Florida, Gainesville, Florida 32611, United States ABSTRACT: The interaction of surface-active molecules with lipid bilayers is ubiquitous both in biological systems and also in several technological applications. Here we explore the interaction of ionic surfactants with liposomes whose composition mimics the ocular epithelia. In this study, liposomes with a composition mimicking ocular epithelia are loaded with calcein dye above the self-quenching concentration. The liposomes are then exposed to surfactants, and the rate of dye leaked from the liposomes due to the interaction of surfactants is measured. Both cationic and anionic surfactants at various concentrations and ionic strengths are explored. Results show that the liposome bilayer permeability to the dye increases on exposure to the surfactants, leading to the release of the dye trapped in the core. However, the dye release stops after a finite time, suggesting a transient increase in permeability followed by healing. The leakage profiles exhibit two different timescales for the cationic surfactant but only one timescale for the anionic surfactant. The total dye leakage increases with surfactant concentration, and at a given concentration, the dye leakage is significantly higher for the cationic surfactants. The timescale for the healing decreases with increasing surfactant concentration, and increasing ionic strength increases the dye leakage for the anionic surfactant. These results show that the surfactant binding to the lipid bilayer increases the permeability while the bilayers heal likely because of the surfactant jump from the outer to the inner leaflet and/or rearrangement into tighter aggregates.

’ INTRODUCTION Lipid bilayers are ubiquitous in all biological systems and also in several technological applications such as food, pharmaceuticals, and so forth. The permeability behavior of lipid bilayers to various biological molecules is critical in maintaining homeostasis in all tissues. A number of biological molecules such as proteins are surface-active, so understanding the interactions between surface-active molecules and lipid bilayers is important. Also, surfactants interact with biological membranes in the body in several pharmaceutical applications (e.g., surfactants increase the permeability of the intestinal epithelium1,2 and corneal epithelium3,4 to some drugs). The interaction of surface-active molecules with lipid bilayers is also relevant in the application of liposomes in drug delivery,5,6 the food industry,7 and several other industrial applications.8 In all of these cases, it is likely that liposomes contact biological fluids, which may contain amphiphillic molecules that may cause the disruption of lipid bilayers.9 The interaction of a surfactant with bilayers increases the permeability, which could possibly have a deleterious effect such as damage to the epithelium and/or a beneficial effect such as an increase in permeability to drugs. Because of these confounding effects, the increase in permeability due to surfactant exposure has been correlated with damage to the epithelium.1012 Also, surface-active preservatives in eye drop formulations that control bacterial growth13 are also known to have the potential for toxicity after eye drop use.14,15 It is also known that the damage due to surfactant exposure is transient and the membranes r 2011 American Chemical Society

recover a few hours after exposure.10,16 Some of the surfactants are also known to be excellent permeability enhancers.16,17 These observations suggest a rich array of complex interactions between surfactants and lipid bilayers, which needs to be explored because of the relevance of such interactions in biological and industrial applications. A number of direct and indirect techniques have been utilized to explore surfactant interactions with lipid bilayers. The Draize eye test has been used to correlate the exposure of surfactants to ocular toxicity.18,19 Although these direct studies with biological membranes are useful, it is difficult to understand the mechanisms of interaction from these studies. The effect of surfactants on the lipid bilayer can be discerned from the changes in permeability of the bilayer, which in turn can be determined by measuring the rate of release of a suitable dye trapped in the core. Liposomes that are spherical vesicles of lipid bilayers are frequently used in such studies. Calcein, or fluorexon dye, is ideally suited for such measurements because of its high water solubility, stability at physiological pH values,20,21 and self-quenching properties at high concentrations. Because of self-quenching, the dye present in the core of the vesicles at elevated concentrations does not contribute to fluorescence. The leaked dye, however, is at a sufficiently low concentration to be fluorescent. Received: April 20, 2011 Revised: July 5, 2011 Published: July 25, 2011 10840

dx.doi.org/10.1021/la201438s | Langmuir 2011, 27, 10840–10846

Langmuir Thus, the transient fluorescence in a suspension of liposomes after surfactant addition can be utilized to understand the mechanisms of surfactant interaction with lipid bilayers. Prior research on surfactant interactions with liposomes suggests that surfactant molecules are incorporated into the bilayer to cause increased permeability. This effect, however, is transient because the lipid bilayers heal to cause a decrease in the permeability. Several different mechanisms have been postulated to explain the phenomenon of a transient increase in permeability followed by a decrease. It has been proposed that surfactant incorporation into the bilayer causes the formation of a pore,22 which can eventually heal because of surfactant and lipid reorganization. It has also been proposed that surfactant adsorption to the outer leaflet creates pressure leading to increased permeability23 and flipping of the surfactant to the inner leaflet can reduce or eliminate the pressure difference, thus creating the healing effect. It has also been proposed that above a critical surfactant concentration the lipids can be solubilized in the surfactant micelles.24 The focus of this article is to explore the mechanisms of pore creation and healing and to try to distinguish between the two possible healing mechanisms: (i) reorganization of lipids and surfactants by diffusion along leaflets; (ii) flipping of the surfactant from the inner to the outer leaflet. We propose that if surfactant diffusion along the leaflets and subsequent assembly into a more tightly packed structure reduces the permeability, then the timescale for this healing should strongly depend on the surface concentration. If, however, surfactant flipping is the ratelimiting step for healing, then the timescale would depend weakly on the concentration. In this article, we utilize the calcein assay to measure transient leakage from anionic liposomes after exposure to surfactants. We vary the surfactant concentration, charge (cationic and anionic), and ionic strength to understand the mechanisms. Furthermore, we propose a phenomenological model for the transient permeability of liposomes after exposure to surfactants. The model is useful in illustrating the important physics and in determining the parameters associated with various aspects of the interactions. The liposomal formulation chosen here is a mimic of the corneal epithelium, which frequently contacts surface-active molecules present in cosmetics and soaps and also those surfactants that are delivered intentionally through eye drops.

’ MATERIALS AND METHODS Dulbecco’s phosphate-buffered saline (PBS), benzalkonium chloride (BC), cholesterol, sodium dodecyl sulfate (SDS), chloroform, methanol, and Sephadex (G-50) were purchased from Sigma-Aldrich (St. Louis, MO). Lipids 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE) dissolved in chloroform, 1,2-dimyristoyl-sn-glycero-3-[phosphor-rac-(1glycerol)] (sodium salt, DMPG) in powder form, 1,2-dimyristoyl-snglycero-3-phosphocholine (DMPC), and a kit for liposome preparation (mini-extruder) were purchased from Avanti Polar Lipids, Inc. (Alabaster, AL). Triton X-100 surfactant and calcein dye (Fluorexon) were purchased from Fisher Scientific (Pittsburgh, PA). Liposome Preparation. The liposomes were prepared from a mixture of 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), 1,2dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE), 1,2-dimyristoylsn-glycero-3-[phospho-rac-(1-glycerol)] (DMPG), and cholesterol in a molar ratio of 8:6:1.5:1.5. Detailed preparation methods of the liposomes have been discussed by Chauhan et al.25 and are also briefly described below. The lipid mixture was dissolved in a 9:1 mixture of chloroform and methanol. The solvent was evaporated by blowing

ARTICLE

nitrogen gas into the vial for 45 min to obtain a dried lipid layer. Finally, 1 mL of a 100 mM calcein solution in PBS was added to the dried lipid layer to obtain a mixture with 25 mg of lipid per milliliter of dye solution. After the solution was sonicated for 20 min, it was stirred for 18 to 24 h at approximately 300 rpm and 30 °C. Next, the solution was passed through the Avanti mini-extruder to obtain liposomes of a uniform size. Following extrusion, the bulk unencapsulated calcein dye was removed from the loaded liposomes by passing the solution through a column of Sephadex gel. The concentrated dye-loaded liposomes were stored at 4 °C. Calcein Leakage Assay. After preparation, liposomes were diluted 1000-fold to a final concentration of 0.025 mg of lipids/mL. The baseline fluorescence of the solution was measured in a Turner Quantech digital filter fluorometer with a 490 nm excitation filter and a 515 nm emission filter. Next, 122 μL of surfactant was added to 2 mL of the diluted liposomes, and the dynamic fluorescence was measured. The total dye loading in the liposomes was quantified by adding a concentrated Triton X-100 surfactant solution to the liposomes to solubilize the liposome membrane completely and release all of the dye. These experiments were performed at the beginning and also after a steady state had been achieved for fluorescence transients after exposure to the surfactant solutions. These experiments confirmed that the total fluorescence after exposure to Triton X-100 was not impacted by the duration of the experiment, showing the absence of any significant photobleaching effects. Each experiment was conducted in triplicate. The percentage of dye released from the liposomes was obtained from the measured dynamic fluorescence by using the following equation %release ¼

Ft  F0 Ftotal  F0

ð1Þ

where t is the time elapsed after the addition of surfactant, Ft is the fluorescence signal at time t, F0 is the initial fluorescence before the addition of surfactants, and Ftotal is the fluorescence after the addition of the Triton X-100 surfactant. All experiments were conducted in a dark room with minimal light to ensure a negligible loss of fluorescence for the duration of the experiments. It is noted that the above equation implicitly assumes linearity between the fluorescence and dye concentration in the solution. In some experiments, a new set of liposomes was added to the surfactantliposome mixture after the fluorescence signal had leveled off. Specifically, 1 mL of dye-loaded liposome solution was added to the mixture of 122 μL of surfactant and 2 mL of liposomes. The fluorescence was then monitored as a function of time. The %release was then calculated using eq 1 with the same values of F0 and Ftotal. These two parameters represent the background fluorescence and fluorescence after the entire amount of dye has leaked, and thus both remain unchanged when the new set of liposomes with the same dye loading as the original set is added to the mixture. For the experiments in which the new set of liposomes did not contain any dye, both F0 and Ftotal were assumed to be two-thirds of the original values to account for the twothirds dilution.

’ RESULTS AND DISCUSSIONS Interaction of Cationic Surfactant Benzalkonium Chloride (BC) with Anionic Liposomes. Figure 1 shows the release

profiles of calcein dye after the addition of BC surfactant solution to the liposomes. The percentage of total dye released is plotted as a function of the time elapsed after surfactant addition. The concentration of lipid is fixed at 0.025 mg/mL (= 0.25% w/w), and the surfactant concentration in the solution varies from 0.00019% (w/w) to 0.00075% (w/w). For ease of comparison, the surfactant concentrations in Figure 1 and 10841

dx.doi.org/10.1021/la201438s |Langmuir 2011, 27, 10840–10846

Langmuir

ARTICLE

Table 1. Best Fit Values of the Parameters Used to Fit the Dye Release Data after BC Addition to the Phenomenological Modela A

B

k 0.20

C/10 BC

26

17.3

C/20 BC

12.7

8.2

0.17

C/40 BC

2.

5.4

0.14

a

For ease of comparison, the concentrations are referenced to an arbitrarily chosen concentration of C = 0.0075% (w/w).

Figure 1. Release of dye from liposomes after the addition of benzalkonium chloride (BC) at concentrations (after addition to liposomes) indicated in the legend. The solid lines are fits to the phenomenological model with parameters presented in Table 1. For ease of comparison, the concentrations are referenced to an arbitrarily chosen concentration of C = 0.0075% (w/w).

subsequent figures are referenced to an arbitrarily chosen concentration of C = 0.0075% (w/w). Thus, concentrations of C/10, C/20, and C/40 refer to 0.00075, 0.00038, and 0.00019% (w/w), respectively. The addition of the surfactant clearly leads to dye leakage, which eventually stops before all of the loaded dye has leaked out of the liposomes. The fractional dye released increases with an increase in surfactant concentration. The release profiles show an initial burst of dye followed by a gradual increase on a slower timescale. On the basis of this observation, a three-parameter phenomenological model represented by eq 2 is fitted to the experimental data. %release ¼ A þ Bð1  ekt Þ

ð2Þ

Parameter A is a measure of the initial burst, and B and k are a measure of the released amount and the timescale for the slower release, respectively. The values of A, B, and k are listed in Table 1 for each concentration, and the best fit profiles are included as the solid lines in Figure 1. On the basis of the data in Figure 1, we propose the following sequence of events after liposomes are exposed to the surfactants: (i) Because of the very small sizes of the liposomes, there are no bulk diffusion limitations and thus the surfactants adsorb almost instantaneously to establish an equilibrium for partitioning into the outer leaflet of the liposomes. (ii) The adsorbed surfactants diffuse along the surface of the outer leaflet and also flip to the inner leaflet. (iii) The flipping of the molecules to the inner leaflet drives further adsorption of the free surfactant from the solution to the outer leaflet. (iv) The surfactant molecules diffusing on either the inner or the outer leaflets also assemble into tight aggregates to stop the dye release. The adsorption stops when all states of the surfactant (free surfactant and adsorbed surfactant on both inner and outer leaflets in various states of aggregation) are in equilibrium. These steps are illustrated in the schematic in Figure 2. The sequence of steps proposed above likely occurs in all systems, but the relative amount and timescales may vary across systems. For instance, the amount of surfactant that adsorbs in the first step will be much larger for systems with electrostatic attraction compared to those with repulsion between the lipids and the surfactants.

The rapid initial burst in Figure 1 is likely due to the dye release in steps iiii (i.e., the time during which the outer leaflet has significantly more adsorbed surfactant), and thus there is a pressure differential between the inner and the outer leaflets. The timescale for the burst should be comparable to the time required to flip the surfactants from the outer to the inner leaflet. The lipid bilayer is porous even after the equilibration in concentration in both leaflets but the porosity is reduced, thus the rate of dye release decreases after the initial burst. Finally, the surfactant molecules diffuse along the leaflets to form tighter aggregates (healed pores), leading to healing of the membrane, resulting in the stoppage of dye release. The second stage of the release is thus proposed to occur during step iv. The timescale for the healing of the membrane in the second phase of the release decreases with increasing surfactant concentration, supporting the hypothesis that healing occurs because of interactions among surfactants diffusing along leaflets. The reduction in the timescale of healing with increasing concentration is consistent with prior studies on Tween 60 binding to soya lechitin liposomes.24 This proposed mechanism was further explored by various experiments described below. If the initial burst is indeed caused by rapid adsorption from the solution, then its magnitude should be significantly reduced in systems with a smaller surface adsorption of surfactants. To test this hypothesis, liposome leakage was measured with anionic surfactant SDS. Interaction of Anionic Surfactant Sodium Dodecyl Sulfate (SDS) with Anionic Liposomes. Figure 3 shows the %release of dye from liposomes as a function of time after the addition of an SDS surfactant solution to the liposomes. The concentration of lipid is fixed at 0.025 mg/mL (= 0.25% w/w), but the surfactant concentration in the solution varies from 0.01% (w/w) to 0.03% (w/w). It is noted that the SDS concentration in these experiments is significantly larger than the BC concentrations in the experiments reported above, even though the %release is comparable. This data is consistent with the observation that cationic surfactants typically cause more toxicity to the cornea than do anionic surfactants.19,26 The dye release profiles in Figure 3 do not show any initial burst, which is consistent with the hypothesis proposed above. On the basis of the absence of the initial burst, the release data shown in Figure 3 was fitted to the phenomenological model (eq 2) but with the constant A = 0. The values of the best fit parameters for SDS are shown in Table 2, and the best fit curves are plotted as the solid lines in Figure 3. The timescale for the healing in this case also decreases with increasing concentration, which is consistent with the hypothesized mechanisms proposed above. The release profiles for SDS in Figure 3 are significantly different that that for BC because the electrostatic repulsion leads to very small adsorption for the anionic surfactant on the 10842

dx.doi.org/10.1021/la201438s |Langmuir 2011, 27, 10840–10846

Langmuir

ARTICLE

Figure 2. Illustration of the mechanism of surfactant adsorption leading to transient dye release from the liposome core.

Table 2. Best Fit Values of the Parameters Used to Fit the Dye Release Data after SDS Addition to the Phenomenological Modela A

B

k

3C SDS 2C SDS

0 0

73.23 53.8

0.03 0.02

C SDS

0

26.4

0.013

a

For ease of comparison, the concentrations are referenced to an arbitrarily chosen concentration of C = 0.0075% (w/w).

Figure 3. Release of dye from liposomes after the addition of sodium dodecyl sulfate (SDS) at concentrations (after addition to liposomes) indicated in the legend. The solid lines are fits to the phenomenological model, with the parameters presented in Table 2. For ease of comparison, the concentrations are referenced to an arbitrarily chosen concentration of C = 0.0075% (w/w).

anionic surfaces of the liposomes. The adsorbed surfactants, however, flip to the inner leaflet to drive more adsorption. More

importantly, the surfactants diffuse along the surface to form tight aggregates, which also drives more adsorption onto the surface. Thus, the process of surfactant adsorption from the bulk to the outer leaflet is in a pseudosteady state, with the processes of surfactant flipping, surfactant diffusion, and aggregation driving adsorption. The slowest of the timescales, which presumably is the timescale for surface diffusion, will control the overall timescale of the adsorption and also the timescale of dye leakage. The ionic repulsion between the anionic surfactant and the anionic surface of the liposomes could potentially be screened by increasing the ionic strength of the solution. The screening of electrostatics is expected to lead to increased equilibrium 10843

dx.doi.org/10.1021/la201438s |Langmuir 2011, 27, 10840–10846

Langmuir

Figure 4. Release of dye from liposomes in PBS with 1 M NaCl after the addition of benzalkonium chloride (BC) at concentrations (after addition to liposomes) indicated in the legend.

Figure 5. Release of dye from liposomes after the addition of a fresh set of dye-containing liposomes to a mixture of BC and liposomes. The surfactant concentration in the original solution is indicated in the legend. The surfactant concentration after dilution with the fresh liposomes is two-thirds of the original concentration. The arrow marks the time at which the fresh liposomes were added. The first set of data prior to the arrow is identical to that in Figure 1.

adsorption, which should cause an initial burst in dye release. To test this hypothesis, liposomes were exposed to a surfactant solution in PBS with added 1 M NaCl. The release profiles in a high-ionic-strength solution (Figure 4) exhibit a large short-time burst, followed by a slower release with eventual healing. The profiles for all concentrations are relatively similar because the concentrations are extremely high and lead to almost 100% dye release. Reversibility of BC Adsorption. The mechanism proposed above suggests that the equilibrium distribution of surfactants is mostly as aggregates (healed pores) on the surfaces of the lipid bilayers. As a further test of the hypothesis, an additional set of experiments were conducted in which a fresh batch of liposomes were added to the mixture of surfactant and liposomes after a steady state was attained in the transient fluorescence. The issue of the desorption of the surfactants is also very important in the context of ocular epithelia because continuous tear turnover after exposure to surfactant rapidly decreases the surfactant concentration in the tear film. Thus, surfactant molecules that adsorb

ARTICLE

Figure 6. Release of dye from liposomes after the addition of a fresh set of undyed liposomes to a mixture of BC and liposomes. The surfactant concentration in the original solution is indicated in the legend. The surfactant concentration after dilution with the fresh liposomes is twothirds of the original concentration. The arrow marks the time at which the fresh liposomes were added.

after exposure to high-concentration surfactants could desorb after the concentration in tears decreases because of tear turnover. The toxicity due to exposure thus depends on the dynamics of both the adsorption and desorption of the surfactants. The dye release profiles after the addition of the second batch of liposomes to a mixture of liposomes and BC are shown in Figure 5. The first phase of the release in Figure 5 is identical to the data in Figure 1. After the dye release had equilibrated, a fresh set of liposomes containing dye were added at the time indicated by the arrow in the figure. The dilution of the dyed-liposomes leads to an immediate decrease in the fluorescence due to dilution. The solid square symbols mark the predicted value of the dye release immediately after dilution, and the subsequent data are the measured values. Because a majority of the surfactants are already adsorbed on the original liposome, the rapid initial adsorption that occurs in the first step is absent, and thus there is a negligible rapid initial burst in dye release. The surfactants remaining in the bulk (or in the double layer surrounding the original set of liposomes) begin to adsorb to the fresh liposomes, thereby disturbing the equilibrium between the bulk surfactant and the surfactant on the surfaces of the original liposomes. The disturbance in equilibrium will drive the surfactants on the original liposomes to desorb, which will potentially increase the permeability, leading to dye release. The dye release after the addition of the fresh of liposomes is thus contributed by both the original liposomes because of the increase in permeability due to surfactant desorption and the fresh liposomes due to surfactant adsorption. To distinguish between the dye release from the two sets of liposomes, an added experiment was done in which the fresh liposomes were prepared without any dye. The release profile from this set of experiments is shown in Figure 6 for BC. The data shows that there is insignificant dye release from the liposomes, which suggests that either the surfactants adsorbed on the liposome surface are not desorbing or the desorption process does not significantly increase the permeability. The first possibility is negated by the fact that the experiments in Figure 5 (fresh liposomes with dye added after equilibration) show a significant release of dye without any initial burst. The absence of an initial burst in Figure 5 is a clear indication that a majority of the surfactant 10844

dx.doi.org/10.1021/la201438s |Langmuir 2011, 27, 10840–10846

Langmuir

ARTICLE

Figure 7. Release of dye from liposomes after the addition of a fresh set of undyed liposomes to a mixture of SDS and liposomes. The surfactant concentration in the original solution is indicated in the legend. The surfactant concentration after dilution with the fresh liposomes is twothirds of the original concentration. The arrow marks the time at which the fresh liposomes were added.

Figure 8. Release of dye from liposomes after the addition of a fresh set of dye-containing liposomes to a mixture of SDS and liposomes. The surfactant concentration in the original solution is indicated in the legend. The surfactant concentration after dilution with the fresh liposomes is two-thirds of the original concentration. The arrow marks the time at which the fresh liposomes were added.

molecules were adsorbed on the original set of liposomes. Other reported studies also shown that as much as 99% of the molecules could adsorb because of a combination of electrostatic and hydrophobic interactions in such cases.2730 Thus, the dye release in Figure 5 is mostly due to the adsorption of surfactant molecules on the fresh liposomes after desorption from the original liposomes. We also hypothesize that the process of desorption from the original liposomes and adsorption to the fresh liposomes is much more rapid compared to that for surface diffusion. Accordingly, the timescale for the release of the dye after fresh liposomes are added is simply the timescale for diffusion-driven aggregation. At equilibrium, the concentrations of surfactant on both the fresh and original liposomes will be the same and equal to approximately two-thirds the original surfactant concentration. The dye release after the fresh liposomes addition should thus be approximately two-thirds of the dye released from the original liposomes during the slow release phase. The total dye in the system after the addition of the fresh liposomes is three-halves of the total dye in the original set. The fractional release after the addition of the new set of liposomes is thus hypothesized to be four-ninths of the fractional release in the slow phase of the release from the original set. The dashed lines are the predicted values of the dye released after the addition of fresh liposomes on the basis of this approximate calculation. The reasonable agreement between the calculated and measured values strongly suggests that the proposed hypothesis regarding surface diffusion being the ratelimiting step is likely correct. The absence of any dye release in Figure 6 can be rationalized by the hypothesis that the timescale for the breakup of the aggregates is very slow compared to the timescale for surfactant flipping and desorption, and thus the surfactants released from an aggregate due to breakup rapidly desorb, resulting in negligible dye release during this process. These experiments thus strongly suggest that the healed pores or the aggregates of BC that form on the surfaces of the bilayers are reversible, with timescales of disassembly that are much larger than the desorption timescales. Reversibility of SDS Adsorption. Experiments similar to those described in the previous section were also conducted to at least qualitatively explore the reversibility of SDS adsorption to

the anionic liposomes. The results of dye release profiles after the addition of fresh undyed and dyed liposomes to an equilibrated mixture of SDS and liposomes are shown in Figures 7 and 8, respectively. Experiments with the addition of undyed liposomes were conducted only with concentration 3C, and those with undyed liposomes were conducted only with concentrations C and 2C. The results in Figure 7 show that the addition of undyed liposomes does not cause any increase in the dye release. This behavior is similar to that for BC, which again suggests that either the surfactants do not desorb even after dilution or surfactant desorption does not cause any significant dye release. The dye release profiles after the addition of the fresh batch of dyed liposomes to a mixture of liposomes and SDS are shown in Figure 8. The first phase of the release in Figure 8 is identical to the data in Figure 3. After the dye release had equilibrated, a fresh set of liposomes containing dye were added at the time indicated by the arrow in the figure. The dilution of the dyed liposomes leads to an immediate decrease in the fluorescence due to dilution. The solid square symbols mark the predicted value of the dye release immediately after dilution, and the subsequent data are the measured values. The results in Figure 8 show that the dye release increases after the addition of the fresh dyed liposomes, but the increase is much less than that in the first phase. This observation first shows that in spite of electrostatic repulsion a majority of the SDS molecules had adsorbed to the first set of liposomes, and thus the concentration of SDS in the bulk was too small to cause substantial leakage from the second set of liposomes. Furthermore, the SDS molecules adsorbed on the original liposomes do not appear to desorb, at least during an experimental duration of about 200 min. The electrostatic effects could increase the energy barrier for the desorption of the adsorbed molecules, leading to a long desorption time.

’ CONCLUSIONS Surfactant interaction with lipid bilayers is ubiquitous in nature and in industrial applications. The absorption of surfactants to the lipid bilayers is explored here through a calcein assay that measures the rate of release of dye trapped in the liposome core after exposure to surfactant solutions. We have explored the 10845

dx.doi.org/10.1021/la201438s |Langmuir 2011, 27, 10840–10846

Langmuir effect of electrostatic interactions by considering both cationic and anionic liposomes adsorbing on anionic liposomes. Also, the effects of surfactant concentration and ionic strength are explored. Finally, the issue of the reversibility of surfactant adsorption is explored. The results suggest that the surfactant molecules first adsorb to the outer leaflet to cause a pressure differential between the inner and outer leaflets, leading to increased permeability. The surfactants then flip to the inner leaflet to minimize the pressure differential, resulting in a reduction of the permeability, which can be considered to be partial healing. The timescale for this process appears to be too short to measure with the techniques considered here. Even after the surfactant concentration has equalized because of surfactant flipping, the dye leakage appears to continue because of the presence of the surfactant monomers. The surfactants diffuse along the surfaces of the liposomes and eventually aggregate to form a tighter structure, potentially healed pores, with minimal permeability, and thus the dye leakage stops. The timescale for the aggregation or pore healing appears to be controlled by surface diffusion because the timescale decreases with increasing concentration. The formation of the aggregates or healed pores appears to be reversible, but the timescale for desorption is much longer for the case of SDS surfactants, likely because of the repulsion between the surfactant and the liposome surface. The desorption of the surfactants is extremely relevant to potential toxicity with respect to the corneal epithelium because the surfactant concentration in the tears decreases after administration because of continuous tear turnover. The mechanisms of pore formation and healing proposed here are supported by prior studies in the literature,2224,31,32 but to our knowledge, this study is the first to explore and compare the adsorption of both cationic and anionic surfactants and to explore the effects of ionic strength. Furthermore, the studies reported here for the desorption of surfactants are not reported in previous studies. Although the results of the calcein-leakage assay are useful in unveiling the mechanisms of surfactant adsorption, direct molecular-level simulation techniques and molecular-level imaging are required to further validate the mechanisms proposed here.

ARTICLE

(10) Swenson, E. S.; Milisen, W. B.; Curatolo, W. Pharm. Res. 1994, 11, 1132–11421142. (11) Burstein, N. L. Invest. Ophthalmol. Vis. Sci. 1984, 25, 1453–1457. (12) Cadwallader, D. E.; Ansel, H. C. J. Pharm. Sci. 1965, 54, 1010–1012. (13) Noecker, R. Adv. Ther. 2001, 18, 205–215215. (14) Gasset, A. R.; Ishii, Y.; Kaufman, H. E.; Miller, T. Am. J. Ophthalmol. 1974, 78, 98–105. (15) Berdy, G. J.; Abelson, M. B.; Smith, L. M.; George, M. A. Arch. Ophthalmol. 1992, 110, 528–532. (16) Xia, W. J.; Onyuksel, H. Pharm. Res. 2000, 17, 612–618618. (17) Shin, S.-C.; Cho, C.-W.; Oh, I.-J. Int. J. Pharm. 2001, 222, 199–203. (18) Tachon, P.; Cotovio, J.; Dossou, K. G.; Prunieras, M. Int. J. Cosmet. Sci. 1989, 11, 233–243. (19) North-Root, H.; Yackovich, F.; Demetrulias, J.; Gacula, M., Jr.; Heinze, J. E. Toxicol. Lett. 1982, 14, 207–212. (20) Jaeger, D. A.; Jamrozik, J.; Golich, T. G.; Clennan, M. W.; Mohebalian, J. J. Am. Chem. Soc. 1989, 111, 3001–3006. (21) Tanaka, K.; Takeda, T.; Fujii, K.; Miyajima, K. Chem. Pharm. Bull. 1992, 40, 1–5. (22) Lesieur, S.; Grabielle-Madelmont, C.; Menager, C.; Cabuil, V.; Dadhi, D.; Pierrot, P.; Edwards, K. J. Am. Chem. Soc. 2003, 125, 5266–5267. (23) Heerklotz, H.; Seelig, J. Eur. Biophys. J. 2007, 36, 305–314. (24) Annesini, M. C.; Memoli, A.; Petralito, S. J. Membr. Sci. 2000, 180, 121–131. (25) Kapoor, Y.; Howell, B. A.; Chauhan, A. Invest. Ophthalmol. Vis. Sci. 2009, 50, 2727–2735. (26) Ikarashi, Y.; Tsuchiya, T.; Nakamura, A. Cutan. Ocul. Toxicol. 1993, 12, 15–24. (27) Howell, B.; Chauhan, A. J. Colloid Inteface Sci. 2008, 324, 61–70. (28) Howell, B. A.; Chauhan, A. Langmuir 2009, 25, 12056–12065. (29) Howell, B. A.; Chauhan, A. Anesth. Analg. 2009, 109, 678–682. (30) Howell, B. A.; Chauhan, A. J. Pharm. Sci. 2009, 98, 3718–3729. (31) Schubert, R.; Beyer, K.; Wolburg, H.; Schmidt, K. H. Biochemistry 1986, 25, 5263–5269. (32) Cocera, M.; Lopez, O.; Estelrich, J.; Parra, J. L.; Maza, A. d. Spectroscopy 2002, 16, 235–244.

’ AUTHOR INFORMATION Corresponding Author

*Phone: 352-392-2592. Fax: 352-392-9513. E-mail: chauhan@ che.ufl.edu.

’ REFERENCES (1) Walters, K. A.; H., D. P.; T., F. A. J. Pharm. Pharmacol. 1980, 33, 207–213. (2) Van Hoogdalem, E. J.; de Boer, A. G.; Breimer, D. D. I. Pharmacol. Ther. 1989, 44, 407–443. (3) Van der Bijl, P.; Engelbrecht, A. H.; van Eyk, A. D.; Meyer, D. J. Ocul. Pharmacol. Ther. 2002, 18, 419–427. (4) Fu, R. C.-C.; Lidgate, D. M. Drug Dev. Ind. Pharm. 1986, 12, 2403–2430. (5) Gregoriadis, G. Trends Biotechnol. 1995, 13, 527–537. (6) Uchegbu, I. F.; Vyas, S. P. Int. J. Pharm. 1998, 172, 33–70. (7) Kirby, C. Food Sci. Technol. 1991, 5, 74. (8) Zanten, J. H. V.; Chang, D. S. W.; Stanish, I.; Monbouquette, H. G. J. Membr. Sci. 1995, 99, 49–56. (9) Ruiz, J.; Go~ni, F. M.; Alonso, A. Biochim. Biophys. Acta 1988, 937, 127–134. 10846

dx.doi.org/10.1021/la201438s |Langmuir 2011, 27, 10840–10846