Interactions between Drug Delivery Particles and Mucin in Solution

Feb 5, 2008 - Faculty of Health and Society, Malmö UniVersity, SE-205 06 Malmö, Sweden, and Camurus AB,. SE-230 70 Lund, Sweden. ReceiVed August ...
0 downloads 0 Views 121KB Size
Langmuir 2008, 24, 2573-2579

2573

Interactions between Drug Delivery Particles and Mucin in Solution and at Interfaces Olof Svensson,*,† Krister Thuresson,‡ and Thomas Arnebrant† Faculty of Health and Society, Malmo¨ UniVersity, SE-205 06 Malmo¨, Sweden, and Camurus AB, SE-230 70 Lund, Sweden ReceiVed August 30, 2007. In Final Form: NoVember 30, 2007 Cubosome particles were produced by fragmenting a cubic crystalline phase of glycerol monooleate and water in the presence of a stabilizing poly(ethylene oxide)-based polymer. The aim of our investigation was to study the interaction between these particles and mucin to gain information on how they would perform as a vehicle for mucosal drug delivery. Particle electrophoresis was used to investigate the interactions between particles and mucin in solution, and ellipsometry was utilized to study the interactions between particles and mucin-coated silica surfaces. The interaction studies were performed at relevant physiological conditions, and the pH and ionic strength were varied to gain more information about the driving forces for the interaction. The results from electrophoretic measurements showed that mucin in solution adsorbed to the particles at pH 4, whereas at pH 6 no clear interaction was detected. From ellipsometric measurements it was evident that the particles adsorb reversibly to a mucin-coated silica surface at pH 4, while no adsorption of particles could be detected at pH 6. The overall conclusion is that the interaction between these particles and mucin is weak and pH-dependent. These findings are in agreement with other investigations of the interactions between mucin and poly(ethylene oxide) chains.

Introduction Lipid liquid crystalline phases possess both aqueous and aliphatic nanodomains and have the ability to dissolve both hydrophilic and hydrophobic molecules as well as amphiphilic molecules. Since they have the ability to dissolve a wide range of substances, lipid liquid crystalline phases are interesting in the area of drug delivery and particularly delivery of poorly water soluble hydrophobic substances. In addition, lipid liquid crystalline phases may function as a matrix for sustained release and may offer protection from degradation of the incorporated drug.1 Cubic liquid crystalline phases can form from binary mixtures of fatty acid monoglycerides and water. Glycerol monooleate, for example, forms such phases which can coexist in equilibrium with an excess of water.2 These cubic structures are however highly viscous and hence offer a challenge for administration. One feasible way to overcome this problem is fragmentation of the liquid crystalline phase to form an aqueous dispersion of particles. It has been shown that cubic phases of glycerol monooleate can be dispersed and stabilized by block copolymers composed of poly(ethylene oxide) (PEO) and poly(propylene oxide) (PPO). Microfluidization was found to result in faceted particles in the size range of a few hundred nanometers determined from cryo transmission electron microscopy. Most particles were shown to be cubic or hexahedral in shape, reflecting the inner cubic structure, and vesicular structures were often attached to the surface of the particles.3 The block copolymer can be assumed to be located preferentially at the surface of the particle and provide steric stabilization of the dispersion. A subsequent study has shown that by careful control of process parameters particles in the range of 0.1-1 µm can be produced with good storage * To whom correspondence [email protected]. † Malmo ¨ University. ‡ Camurus AB.

should

be

addressed.

stability.4 These lipid-based particles are referred to as Cubosome particles (Cubosome is a USPTO registered trademark of Camurus AB, Sweden), which will be used throughout this paper. The aim of this work was to study the interaction between Cubosome particles and mucin. Although these nanoparticles are a promising vehicle for mucosal drug delivery, no information is available about their mucoadhesive properties and their interactions with mucin, which is the main protein constituent of the mucous gel layer. As the surface of the particle is assumed to be covered with stabilizing PEO chains, our main interest is the interaction between PEO and mucin. In a previous investigation the interactions between PEO-coated polymer particles and pig gastric mucin were studied by solution depletion techniques.5 From this investigation it was shown that mucin adsorbs on the surface of PEO-modified particles. However, it was suggested that PEO coating in a brush conformation reduced the interactions with mucin and hence facilitated the transport through the mucous gel layer. A more extensive study of the interactions between bovine mucin and PEO chains was presented by Efremova and coworkers.6 From surface force measurements and surface plasmon resonance measurements it was concluded that mucin adheres weakly and reversibly to a lipid bilayer with grafted PEO chains. Finally, the interactions between a PEO-PPO-PEO block copolymer and bovine mucin have been studied by rheological measurements, which also suggest that the interaction is relatively weak.7 In this investigation we studied the interaction between Cubosome particles and bovine submaxillary gland mucin in solution by particle electrophoresis and at interfaces by ellipsometry. The interaction studies were performed at relevant physiological conditions, and the pH and ionic strength were

E-mail:

(1) Shah, J. C.; Sadhale, Y.; Chilukuri, D. M. AdV. Drug DeliVery ReV. 2001, 47 (2-3), 229-250. (2) Larsson, K. J. Phys. Chem. 1989, 93 (21), 7304-7314. (3) Gustafsson, J.; Ljusberg-Wahren, H.; Almgren, M.; Larsson, K. Langmuir 1996, 12 (20), 4611-4613.

(4) Barauskas, J.; Johnsson, M.; Joabsson, F.; Tiberg, F. Langmuir 2005, 21 (6), 2569-2577. (5) Yoncheva, K.; Lizarraga, E.; Irache, J. M. Eur. J. Pharm. Sci. 2005, 24 (5), 411-419. (6) Efremova, N. V.; Huang, Y.; Peppas, N. A.; Leckband, D. E. Langmuir 2002, 18 (3), 836-845. (7) Huang, K.; Lee, B. P.; Ingram, D. R.; Messersmith, P. B. Biomacromolecules 2002, 3 (2), 397-406.

10.1021/la702680x CCC: $40.75 © 2008 American Chemical Society Published on Web 02/05/2008

2574 Langmuir, Vol. 24, No. 6, 2008

varied to provide more information about the driving forces for the interaction. The techniques presented here to evaluate the interaction between mucin and drug delivery particles may also be used to evaluate mucoadhesive properties of other particulate drug delivery systems. Materials and Methods Chemicals. A mixture of mono- and diglycerides (44:1 by weight), in which oleic acid constitutes approximately 90 wt % of the fatty acids, was obtained from Danisco Ingredients (Brabrand, Denmark).4 Lutrol F127 is a triblock copolymer of PEO and PPO and was obtained from BASF Svenska AB (Helsingborg, Sweden). The approximate formula of Lutrol F127 is (EO)98(PO)57(EO)98, and the average molar mass is 12 600 g/mol. All other chemicals used were of analytical grade or better, and water was treated by an ELGA UHQ PS purifying unit. Cubosome Particles. The melted lipid mixture (glycerol monooleate) was added to an aqueous Lutrol F127 solution to produce a coarse dispersion. The sample was subsequently mechanically mixed for 12-48 h, homogenized with a microfluidizer at 345 bar and 25 °C, and autoclaved for 20 min at 125 °C. The lipid:polymer ratio of the dispersion was 9:1 (w/w), and the lipid content was 50 mg/mL. The particle size distribution of the particles was determined by a laser diffraction particle size analyzer (Beckman Coulter LS230), and the mean size of the particles in pure water was determined to be 311 nm with a standard deviation of 58 nm. A more detailed description of the preparation procedure and particle size determination can be found elsewhere.4 Mucin. Bovine submaxillary gland mucin (BSM) was obtained from Sigma Chemical Co. (product no. 3895, type I-S) with a specified sialic acid content of 13%. A molecular mass of 1600 kDa has been reported for a purified fraction of this mucin,8 and the isoelectric point of bovine submaxillary gland mucin has been reported to be 3.9 ζ Potential Measurements. The electrophoretic mobility of the particles was determined using particle microelectrophoresis (Mark II, Rank Brothers, Cambridge, U.K.). A thermostated rectangular cell setup was utilized with platinum electrodes, and a ζ potential standard (MRK403-02, Malvern Instruments Ltd., U.K.) was used to determine the interelectrode distance of the cell. From the electrophoretic mobility (u) the ζ potential was subsequently calculated using the Smoluchowski equation (κR . 1). All measurements were conducted on dilute particle dispersions (0.05 mg/mL lipid) in 50 mM NaCl at 37 °C. Values of relative permittivity (r) and viscosity (η) in the calculations were 74.2 and 0.695 × 10-3 Pa S, respectively.10 Adjustment of the pH was done with small additions of aqueous solutions of NaOH (1 mg/mL) or HCl (5 mg/ mL). ζ potential measurements were conducted to characterize the Cubosome particles at different pH values. Also the interaction between mucin and Cubosome particles in solution was examined by measuring changes in the ζ potential. A stock mucin solution was prepared (2 mg/mL) and filtered three times through a filter with a pore size of 200 nm to remove mucin aggregates that would otherwise interfere with the ζ potential measurements. It has been reported previously that this mucin preparation contains large aggregates with a hydrodynamic radius of over 500 nm11 and that filtration can be utilized to remove these aggregates.6 By comparing the mucin content in lyophilized filtered and unfiltered mucin solution, it was concluded that less than 10% of the mucin is removed in the filtration procedure. After filtration the mucin solution was diluted to the desired concentration and thermostated to 37 °C. Cubosome particles were added to the mucin solution, and after an equilibrium period of 20 min, the ζ potential was determined from the electrophoretic mobility. Each data point represents the mean value (8) Shi, L.; Caldwell, K. D. J. Colloid Interface Sci. 2000, 224 (2), 372-381. (9) Perez, E.; Proust, J. E. J. Colloid Interface Sci. 1987, 118, 182-191. (10) CRC Handbook of Chemistry and Physics, 83rd ed.; CRC Press: Boca Raton, FL, 2002. (11) Bastardo, L.; Claesson, P.; Brown, W. Langmuir 2002, 18 (10), 38483853.

SVensson et al. of 10 measurements, and confidence intervals (95%) are presented in the figures. The increase in viscosity when mucin was present in the ambient solution was measured by capillary viscosimetry to be less than 2% and is not accounted for in the calculations of the ζ potential. Particle electrophoresis measurements were also performed to characterize mucin adsorption to silica spheres. Silica spheres with a diameter of 490 nm (Postnova Analytics GmbH, Germany) were washed in ethanol twice and then in water three times. The particle dispersion was then incubated with a filtered 1 mg/mL mucin solution for 2 h and diluted to a mucin concentration of 0.01 mg/mL. After 1 h the ζ potential was determined and compared with the ζ potential of the silica particles without mucin. The electrolyte concentration and pH during the measurements were 50 mM and 6, respectively. Finally, electrophoretic measurements were conducted on mucin aggregates to determine their electrophoretic mobility. Silica Surfaces. Oxidized silicon surfaces (Si/SiO2) were prepared from silicon wafers (Okmetic OY, Espoo, Finland) with an oxide layer thickness of approximately 320 Å.12 The surfaces were cleaned by plasma cleaning for 5 min in low-pressure residual air in a glow discharge unit (Plasma Cleaner PDC-32 G, Harrick Scientific Corp., New York). Plasma cleaning was followed by gentle boiling in an alkaline solution for 5 min, rinsing three times in water, and gentle boiling in an acidic solution for 5 min. The components of the first solution were NH3 (25%), H2O2 (30%), and water (1:1:5 by volume), and the second solution was composed of HCl (37%), H2O2 (30%), and water (1:1:5 by volume). Finally, the surfaces were rinsed in water three times and then in ethanol (96%) twice. The cleaned surfaces were stored in ethanol. The contact angle of silica surfaces cleaned as described above has been reported to be less than 10°.13 Before ellipsometric measurements the cleaned silica surfaces were rinsed in water, dried, and subjected to plasma cleaning for 5 min. They were then immediately transferred to the cuvette for ellipsometric measurements. Ellipsometry. The adsorption of Cubosome particles to silica surfaces and mucin-coated silica surfaces was monitored in situ by ellipsometry to obtain time-resolved values of the thickness and refractive index of the film that were used to calculate the adsorbed amount. Theoretical principles can be found elsewhere,14 and the experimental setup was based on null ellipsometry according to the principles of Cuypers.15 The instrument used was a Rudolph thin film ellipsometer (type 43603-200E, Rudolph Research, Fairfield, NJ) automated according to the concept of Landgren and Jo¨nsson.16 A xenon arc lamp was used as the light source, and light was detected at 442.9 nm using an interference filter with UV and infrared blocking (Melles Griot, The Netherlands). The 5 mL trapezoid cuvette made of optical glass (Hellma, Germany) was equipped with a magnetic stirrer and thermostated to 37 °C. Rinsing of the cuvette was done at 18 mL/min. Determination of the complex refractive index of the silicon and the thickness and refractive index of the silicon oxide layer was done using air and the aqueous phase as ambient media,16 and two zone measurements were conducted to reduce systematic errors. The refractive index for air and the aqueous phase used in the calculations was 1.000 and 1.339 (50 mM NaCl) or 1.340 (150 mM NaCl), respectively. The calculation of the adsorbed amount from the refractive index and thickness of the film was done assuming a linear increase of the refractive index with the concentration according to the formula of de Feijter.17 The refractive index increment (dn/dc) for Cubosome particles was approximated to that of pure glycerol monooleate, which has been found to be 0.169 mL/g.18 For (12) Wahlgren, M.; Arnebrant, T. J. Colloid Interface Sci. 1990, 136, 259265. (13) Malmsten, M.; Veide, A. J. Colloid Interface Sci. 1996, 178 (1), 160167. (14) Azzam, R. M. A.; Bashara, N. M. Ellipsometry and polarized light; NorthHolland: Amsterdam, 1977. (15) Cuypers, P. A. Dynamic Ellipsometry; Rijksuniversiteit Limburg: Maastricht, The Netherlands, 1976. (16) Landgren, M.; Jo¨nsson, B. J. Phys. Chem. 1993, 97, 1656-1660. (17) de Feijter, J. A.; Benjamins, J.; Veer, F. A. Biopolymers 1978, 17, 17591772.

Drug DeliVery Particle and Mucin Interactions

Figure 1. Electrophoretic mobility and ζ potential versus pH for Cubosome particles. Error bars represent 95% confidence intervals, and a dashed line is added to guide the eye. the case where mucin was first adsorbed, we also assumed a refractive index increment of 0.169 mL/g. As the mucin used has a reported refractive index increment of 0.16 mL/g,19 the mass of the mucin layer was underestimated by approximately 5%, and the underestimation in mass of the mixed layer of mucin and particles was hence 5% or less. The interaction between silica and Cubosome particles was studied by monitoring the changes in the adsorbed amount, thickness, and refractive index as the particle dispersion was added to the cuvette. The particle concentration throughout this work was 0.05 mg/mL (lipid content) if nothing else is specified, and adsorption was monitored for 1 h followed by 10 min of rinsing. The preadsorbed mucin layer was obtained by adsorption of mucin (1 mg/mL) for 2 h followed by 10 min of rinsing. After a stabilizing period of 1 h the Cubosome particles were added, and adsorption was monitored for 1 h. Subsequently, the cuvette was rinsed for 10 min.

Results and Discussion ζ Potential of Cubosome Particles. ζ potential measurements were utilized to characterize the Cubosome particles in the pH ranging from 3 to 8, and the results from the measurements are presented in Figure 1. The ζ potential was found to be negative and relatively low in magnitude, ranging from -3 mV at pH 3 to -9 mV at pH 8. To examine whether the negative ζ potential arises from the cubic glycerol monooleate phase or the stabilizing polymer (Lutrol F127), electrophoretic mobility measurements were conducted on a pure polymer particle dispersion in 50 mM NaCl. The result showed that the polymer particles have a negative ζ potential close to zero (>-2 mV at pH 5.4) in agreement with other studies.20,21 The negative charge of the Cubosome particles is thus assumed to be associated with the glycerol monooleate cubic phase. Free oleic acid present in the lipid phase may give rise to the negative charge of the particles, but the negative charge can also be explained by preferential adsorption of hydroxyl ions at the oil-water interface. A negative ζ potential is a general feature of oil droplets in water, and it is likely due to specific adsorption of hydroxyl ions at the oil-water interface, although the reason for this preferential adsorption is not clear.21 (18) Campos, J.; Eskilsson, K.; Nylander, T.; Svendsen, A. Colloids Surf., B 2002, 26 (1-2), 172-182. (19) Dedinaite, A.; Bastardo, L. Langmuir 2002, 18, 9383-9392. (20) Barnes, T. J.; Prestidge, C. A. Langmuir 2000, 16 (9), 4116-4121. (21) Marinova, K. G.; Alargova, R. G.; Denkov, N. D.; Velev, O. D.; Petsev, D. N.; Ivanov, I. B.; Borwankar, R. P. Langmuir 1996, 12 (8), 2045-2051.

Langmuir, Vol. 24, No. 6, 2008 2575

Figure 2. Electrophoretic mobility and ζ potential of Cubosome particles versus ambient mucin concentration. The interaction was monitored at pH 4.0 ( 0.1 and at pH 6.0 ( 0.2. Error bars represent 95% confidence intervals, and dashed lines represents linear fits. Key: *, statistically significant difference from particles in mucinfree solutions (P < 0.05); superscript “a”, mucin concentration calculated on the basis of unfiltered solutions.

Interaction between Cubosome Particles and Mucin in Solution. Figure 2 presents the ζ potential of the Cubosome particles versus the ambient mucin concentration. A significant decrease was observed in the ζ potential at pH 4 at a mucin concentration of 0.1 mg/mL. At pH 6 it was indicated that the ζ potential decreased, but it was not obvious when compared with the statistical deviations. The decrease can be explained by adsorption of negatively charged mucin molecules on the surface of the Cubosome particles. As the isoelectric point for bovine submaxillary gland mucin is reported to be 3, the mucin molecules carry a net negative charge at both pH 4 and pH 6. Mucin molecules can interact with PEO chains of the block copolymer as suggested by others and adsorb on the surface of the particles.6 However, there is also a possibility for mucin to adsorb in the lipid-water interface since mucins are amphiphilic molecules composed of hydrophilic glycosylated regions and regions with no or little glycosylation. It has been suggested from other investigations that these nonglycosylated regions can interact by hydrophobic interactions with other molecules11 and macroscopic surfaces.22 Measurements were also conducted on mucin aggregates in unfiltered solutions without Cubosome particles for comparison. The absolute value of the electrophoretic mobility of these aggregates was found to be 10 × 10-9 m2/(V s) at pH 4 and 14 × 10-9 m2/(V s) at pH 6. These values are substantially higher than the values for Cubosome particles in the presence of mucin, implying that the surface coverage of mucin on the particles is low. Interaction between Cubosome Particles and Silica Surfaces. Although the main objective of this work is to study the interactions between mucin and Cubosome particles, particle adsorption to silica surfaces was also investigated to gain more insight into the interfacial properties of Cubosome particles and to obtain results that could be compared with the results from mucin-coated silica surfaces. The adsorption behavior of Cubosome particles at silica surfaces has previously been investigated (22) Shi, L.; Ardehali, R.; Caldwell, K. D.; Valint, P. Colloids Surf., B 2000, 17 (4), 229-239.

2576 Langmuir, Vol. 24, No. 6, 2008

SVensson et al. Table 1. Adsorbed Amount, Ellipsometric Thickness, and Refractive Index of Cubosome Particles on Silicaa

pH 4.0

pH 6.0

Figure 3. Adsorption of Cubosome particles on silica surfaces at pH 6 ( 0.2. Adsorbed amount and thickness versus the electrolyte concentration. The dashed line is added to guide the eye, and a horizontal dashed line represents the size of the particles obtained from particle size determination.

with ellipsometry.23 It was concluded from this study that an electrostatic repulsive interaction is present between the particles and the silica surface and that the adsorption can be induced by introducing electrolytes or decreasing the pH of the system. Apart from electrostatic repulsive interactions, attractive van der Waals interactions are expected to be present between the lipid particles and the silica surface, and in addition hydrogen bonds may form between oxygen atoms in PEO chains and hydroxyl groups at the surface. Previous studies have revealed that PEO adsorption on silica may be mediated by hydrogen bonds as adsorption is only observed at neutral and acidic pH when a higher fraction of the hydroxyl groups at the surface are in the protonated state.24 A more elaborate discussion about the driving forces for hydrogen bonding between PEO chains and silica including electrolyte effects and pH can be found elsewhere.25 Figure 3 shows the adsorbed amount and ellipsometric thickness versus the electrolyte concentration at pH 6. No adsorption was detected at 20 mM NaCl and below, while a substantial increase in the adsorbed amount and thickness was observed at 40 mM NaCl and above. It is thus confirmed that an electrostatic repulsion exists between the particles and the silica surface and that the critical monovalent electrolyte concentration for adsorption is between 20 and 40 mM NaCl at our experimental conditions. Assuming that the particles are spherical in shape and that the adsorption can be described by the random sequential adsorption (RSA) model, the volume fraction of particles on the surface is calculated to be 0.33 in the adsorbed layer corresponding to a mass of approximately 70 mg/m2. The differences in the adsorbed amount between the theoretical value and values obtained from ellipsometric measurements (25 mg/m2) are thus substantial, and two possible explanations can account for this. First, the particles could be deformed upon adsorption, which reduces the free surface area, and second, constituents from the bulk or the particles could adsorb to the surface and passivate the surface with respect to further adsorption. As the thickness detected by ellipsometry (23) Vandoolaeghe, P.; Tiberg, F.; Nylander, T. Langmuir 2006, 22 (22), 91699174. (24) Van, der Beek, G. P.; Cohen, Stuart, M. A.; Cosgrove, T. Langmuir 1991, 7 (2), 327-334. (25) Iruthayaraj, J.; Poptoshev, E.; Vareikis, A.; Makuska, R.; vanderWal, A.; Claesson, P. M. Macromolecules 2005, 38 (14), 6152-6160.

50 mM NaCl before rinsing after rinsing 150 mM NaCl before rinsing after rinsing 50 mM NaCl before rinsing after rinsing 150 mM NaCl before rinsing after rinsing

adsorbed amount (mg/m2)

thickness (nm)

refractive index

7.0 6.3

99 88

1.351 1.351

7.4 7.0

105 94

1.352 1.353

22.0 21.4

225 219

1.356 1.356

25.9 25.6

215 215

1.360 1.360

a Measurements were done at pH 4.0 ( 0.2 and 6.0 ( 0.2 and at 50 and 150 mM NaCl. The mean values of two measurements are presented, and the deviations in thickness and adsorbed amount between measurements were below 25%.

was rather close to that of the particles (Figure 3), the deformation is not likely to be very pronounced, and therefore, the discrepancy may be best accounted for by assuming that small amounts of lipid or/and polymer between the particles prevents the formation of a more closely packed layer of particles. Table 1 summarizes the adsorption of Cubosome particles on silica surfaces at different pH values (pH 4 and 6) and at different ionic strengths (50 and 150 mM NaCl). At pH 4 the adsorbed amount and thickness were found to be approximately 7 mg/m2 and 100 nm, respectively, and the electrolyte concentration did not significantly affect the adsorption. The decrease in the adsorbed amount and thickness during rinsing was less than 10%, and thus, most particles were irreversibly adsorbed at the experimental time scales. At pH 6, the adsorbed amount was found to be approximately 3 times higher than at pH 4 and the thickness of the adsorbed layer was twice as high and closer to the dimensions of the Cubosome particles. The decrease in the adsorbed amount and thickness upon rinsing was less than at pH 4, and thus, most particles were found to be irreversibly bound to the surface. In addition, the electrolyte concentration had little influence on the adsorption, and the adsorbed amount was somewhat higher at 150 mM NaCl than at 50 mM NaCl. The higher thickness at pH 6 was not a result of a higher degree of surface coverage, because at equal amounts the thickness was still twice as high at pH 6 during ellipsometric measurements (data not shown). This implies that Cubosome particles are deformed upon adsorption and that the deformation is more pronounced at pH 4 than at pH 6. Deformation of Cubosome particles at surfaces has previously been reported, and atomic force microscopy investigations revealed that Cubosome particles are flattened upon adsorption on mica.26 The more pronounced deformation of the particles at pH 4 may be explained by the fact that a larger number of hydrogen bonds can form between the silica surface and the particles. At pH 6 a larger fraction of the hydroxyl groups present at the silica surface are deprotonated and can therefore not participate in hydrogen bonding. The most obvious explanation to account for the lower adsorbed amounts on silica at pH 4 is that at this pH the deformation is more pronounced, which in turn reduces the free surface area and hence the total number of particles per surface area. We also speculate that a stronger interaction at pH 4 results not only in (26) Neto, C.; Aloisi, G.; Baglioni, P.; Larsson, K. J. Phys. Chem. B 1999, 103 (19), 3896-3899.

Drug DeliVery Particle and Mucin Interactions

Langmuir, Vol. 24, No. 6, 2008 2577

Table 2. Adsorbed Amount and Ellipsometric Thickness of Preadsorbed Mucin Layers and Changes in the Adsorbed Amount and Thickness When Cubosome Particles Are Added to the Ambient Solutiona preadsorbed mucin layer

pH 4 pH 6

50 mM NaCl 150 mM NaCl 50 mM NaCl 150 mM NaCl

changes after particle addition

changes after particle addition and rinsing

amount (mg/m2)

thickness (nm)

amount (mg/m2)

thickness (nm)

amount (mg/m2)

thickness (nm)

1.0 ( 0.1 0.8 ( 0.2 0.9 ( 0.1 0.7 ( 0.4

32 ( 1 41 ( 13 20 ( 9 30 ( 14

+0.8 ( 0.1 +0.7 ( 0.2 +0.1 ( 0.0 +0.1 ( 0.0

+11 ( 9 -4 ( 7 +1 ( 7 +1 ( 6

+0.2 ( 0.1 +0.2 ( 0.2 0 ( 0.0 0 ( 0.0

0(3 -3 ( 3 0(6 +3 ( 5

a Measurements were done at pH 4.0 ( 0.1 and 6.0 ( 0.1 at 50 and 150 mM NaCl. The mean values of two measurements are presented ( the standard deviation.

more deformed particles, but also in a pronounced adsorption of lipid or/and polymer between the particles, which hinders further adsorption. A lower refractive index at pH 4 implies that the interparticle separation is higher at pH 4, supporting this assumption. The investigation of the interaction between silica and Cubosome particles provides some interesting information about the interfacial properties of the particles. It is, for example, likely that PEO chains on the surface of the particles can form hydrogen bonds. In addition, it is also indicated that the particles are deformed upon adsorption and that the degree of deformation depends on the pH. Interaction between Cubosome Particles and MucinCoated Silica Surfaces. A mucin layer was created by mucin adsorption from ambient solution. Adsorption was performed at pH 4 and 6 at 50 and 150 mM NaCl. The amount of mucin after rinsing with mucin-free solution was found to be between 0.7 and 1.0 mg/m2, and the thickness was found to be between 20 and 40 nm (Table 2). Furthermore, no correlation was found between the pH or salt concentration and the amount and thickness of the mucin layer. The irreversibly bound fraction of mucin is higher than reported previously on silica at salt-free conditions,27 which is reasonable due to screening of electrostatic repulsion between the silica surface and mucin. However, the values are lower in comparison with other results from adsorption at similar conditions,19 which might be explained by differences in surface cleaning procedures. Results from electrophoretic measurements revealed that the ζ potential of silica particles changed from -11 to -13 mV when mucin adsorbed to the particles (50 mM NaCl, pH 6). Thus, the absolute value of the ζ potential increased when mucin adsorbed, and this can be a result of the high amount of negatively charged sialic acid residues. In addition to ellipsometric and particle electrophoresis measurements, mucin adsorption on silica was also studied by atomic force microscopy. Mucin was adsorbed from a 1 mg/mL mucin solution with 50 mM NaCl at pH 6 for 2 h. After rinsing of the cell with a mucin-free solution, the surface was scanned as described in ref 27. The results showed that mucin adsorbed as closely packed disk-shaped aggregates approximately 150 nm in lateral dimensions and 20 nm in height in agreement with ellipsometric measurements. Figure 4 shows the increase in the adsorbed amount when Cubosome particles adsorb to a mucin-coated silica surface at pH 4 and 50 mM NaCl. A clear increase was detected, and the amount increased with the particle concentration. Although no plateau value was reached within this concentration range, it is indicated that the maximum in adsorbed amount is around 3 mg/m2. In the subsequent ellipsometric investigations we decided to study the interactions between mucin and the particles at a concentration of 0.05 mg/mL. Although at this concentration no (27) Svensson, O.; Lindh, L.; Cardenas, M.; Arnebrant, T. J. Colloid Interface Sci. 2006, 299 (2), 608-616.

Figure 4. Adsorption of Cubosome particles to a mucin-coated silica surface versus the particle concentration at pH 4 and 50 mM NaCl. Adsorbed amount after an adsorption time of 1 h. The maximum lipid concentration that could be used for ellipsometric measurements was 1.0 mg/mL due to light scattering of the solution. A dashed line is added to guide the eye.

plateau is reached, we chose this concentration as it is expected to be close to relevant concentrations in in vivo applications and for comparison with a previous investigation.23 Figure 5 displays representative examples of the changes in the adsorbed amount and thickness upon addition of Cubosome particles to a mucin-precoated silica surface at 50 mM NaCl, and Table 2 summarizes all results from particle adsorption to silica with adsorbed mucin. At pH 4 the addition of Cubosome particles resulted in an increase in the adsorbed amount at both 50 and 150 mM NaCl. However, the increase in the adsorbed amount was relatively small compared with that of adsorption on silica (Table 1), and after rinsing, the adsorbed amount was close to the value before addition of Cubosome particles. The changes in thickness upon addition of particles were unexpectedly small, and the deviations in thickness between measurements were substantial. As no significant changes could be detected in the thickness data (Table 2), we decided to investigate the possibility that the increase in mass could be caused by adsorption of smaller constituents in the dispersion that do not result in an increase in the ellipsometric thickness. Measurements were consequently conducted on filtered particle dispersions (200 nm pore size filter). On silica the adsorbed amount from the filtered dispersion was less than 0.2 mg/m2 compared with 7.0 mg/m2 for the unfiltered dispersion (Table 1). The increase in the adsorbed amount when a filtered dispersion was added to a mucin-precoated silica surface was 0.1 mg/m2 compared with

2578 Langmuir, Vol. 24, No. 6, 2008

SVensson et al.

Figure 5. Adsorbed amount and thickness versus time upon addition of Cubosome particles to a silica surface precoated with mucin. The figures present representative examples of the adsorption at pH 4 (left) and pH 6 (right) in 50 mM NaCl.

Figure 6. Changes in the adsorbed amount and the total thickness of the adsorbed layer after addition of Cubosome particles versus the amount of preadsorbed mucin. The pH was 6 ( 0.2 and the electrolyte concentration 50 mM NaCl. Lines are added to guide the eye.

0.8 mg/m2 for the unfiltered dispersion. Thus, we conclude that the increase in the adsorbed amount on both silica and a mucincoated silica can be assigned to adsorption of intact particles. At pH 6 the increase in the adsorbed amount was 0.1 mg/m2 or less at both 50 and 150 mM NaCl, and as expected no significant changes were detected in the ellipsometric thickness. Thus, at pH 4 we detected adhesion of particles to the mucin layer, while at pH 6 the results imply that no adhesion of particles occurred. Figure 6 illustrates how the surface coverage of mucin influences the subsequent adsorption of Cubosome particles at pH 6 and 50 mM NaCl. An increase in the adsorbed amount and the total thickness of the layer is shown versus the amount of preadsorbed mucin. The mucin adsorption time was varied between 10 min and 2 h to obtain different surface coverages. It was found that 0.30 mg/m2 mucin was sufficient to completely hinder the adsorption of the particles. At a mucin coverage of 0.25 mg/m2 the mass increased by 0.9 mg/m2 upon addition of the particles, while the ellipsometric thickness increased to 230 nm. Thus, a very low surface coverage was found to be sufficient to obtain a thickness that is close to the dimensions of the studied particles.

The results from ellipsometric measurements are in agreement with the results obtained from particle electrophoresis measurements in that we detected an interaction between the particles and mucin at pH 4, while no significant interaction was detected at pH 6. One apparent reason for this difference could be that the electrostatic repulsion between the particles and mucin is pronounced at higher pH when both the mucin and the particles have a higher charge. However, when we increase the electrolyte concentration from 50 to 150 mM NaCl at pH 6, we still do not detect any interaction by ellipsometric measurements. Thus, we cannot conclude that electrostatic repulsion hinders the interaction. The differences in binding at different pH values can however be explained by assuming that the sialic acid residues are important for the interaction, owing to the high amount in combination with the preferential location at the terminal end of the oligosaccharide side chains.28 At pH 6 the sialic acid residues have a negative charge, while at pH 4 a fraction of the carboxyl groups are protonated (pKa ) 2.629) and accessible for hydrogen bonding with the oxygen atoms in PEO. These hydrogen bonds are expected to be strong as the hydroxyl groups of carboxylic acids are more polarized than, for example, hydroxyl groups of alcohols due to the presence of adjacent carbonyl groups.30 The combined results from investigations in solution (particle electrophoresis) and at interfaces (ellipsometry) points out that the interaction between mucin and the particles is weak and pH-dependent. These findings are in agreement with the results obtained by others on the interaction between PEO and mucin. Mucin was, for example, shown to adsorb reversibly on a lipid bilayer with grafted PEO chains,6 and a PEO-PPO-PEO block copolymer was shown only to be weakly mucoadhesive.7 From a clinical viewpoint it is expected that these particles will diffuse rather unhindered through a mucous gel as the interaction with mucin is weak.

Conclusions The interaction between mucin and drug delivery particles (Cubosome particles) was studied in solution with particle (28) Strous, G. J.; Dekker, J. Crit. ReV. Biochem. Mol. Biol. 1992, 27 (1-2), 57-92. (29) De Bruyn, P.; Michelson, S. J. Cell Biol. 1979, 82 (3), 708-714. (30) Chitra, R.; Das, A.; Choudhury, R.; Ramanadham, M.; Chidambaram, R. Pramana 2004, 63 (2), 263-269.

Drug DeliVery Particle and Mucin Interactions

electrophoresis and at interfaces by ellipsometry. Results from particle electrophoresis measurements suggested that a small amount of mucin adsorbed to the surface of the particles at pH 4, while no clear adsorption was detected at pH 6. From ellipsometric measurement it was shown that the particles adsorb reversibly to a mucin-coated silica surface at pH 4, while no significant adsorption was detected at pH 6. An increase in the electrolyte concentration from 50 to 150 mM NaCl was shown not to influence adsorption to mucin-coated surfaces, and therefore, we speculate that hydrogen bonding may be important for the interactions with mucins. The overall conclusion is that the interaction between these particles and mucin is weak and pH-dependent. This finding is in agreement with other investigations of the interactions between mucin and poly(ethylene oxide) chains.

Langmuir, Vol. 24, No. 6, 2008 2579

Acknowledgment. We thank Prof. Tautgirdas Ruzgas at Malmo¨ University and Prof. Fredrik Tiberg at Camurus AB for fruitful discussions. Also we thank Dr. Justas Barauskas at Lund University for preparing the Cubosome particle dispersions and Bo Thune´r at Linko¨ping University for providing oxidized silicon surfaces. This investigation was supported by research grants from the Knowledge Foundation (BiofilmssResearch Centre for Biointerfaces). Supporting Information Available: Dark field microscopy was used to characterize a clean silica surface and a silica surface with adsorbed Cubosome particles. This information is available free of charge via the Internet at http://pubs.acs.org. LA702680X