Detachable immobilization of liposomes on polymer gel particles

Colloids and Surfaces B: Biointerfaces 37 (2004) 35–42

Detachable immobilization of liposomes on polymer gel particles Md. Abdul Khaleque, Yukihisa Okumura∗ , Satoshi Yabushita, Michiharu Mitani Department of Chemistry and Material Engineering, Faculty of Engineering, Shinshu University, 4-17-1 Wakasato, Nagano 380-8553, Japan Received 9 April 2004; accepted 26 June 2004

Abstract Immobilization of liposomes on hydrophobized Sephacryl gel and controlled detachment of the liposomes from the gel were examined. The gel was chemically modified and bore octyl, hexadecyl or cholesteryl moiety via disulfide linkage as anchors to liposomal bilayer membrane. Upon interaction with the gel, egg phosphatidylcholine liposomes were successfully immobilized onto the gel. The gel with cholesteryl moiety showed 1.7 times higher liposome immobilization per anchor moiety than the gels with the alkyl moieties. The immobilization of liposomes on the gel was stable, and no significant spontaneous detachment of phospholipid or leakage of fluorescein isothiocyanate-conjugated dextran encapsulated in the immobilized liposomes was observed in 24 h. Reductive cleavage of the disulfide linkage by dithiothreitol resulted in detachment of the liposomes from the gel. The majority of the detached liposomes were found encapsulating the dextran derivative, and these liposomes should have kept their structural integrity throughout the immobilization and the detachment processes. The release of the liposomes was insignificant until the ratio of the dithiothreitol to the hydrophobic anchor reached a threshold. The presence of the threshold suggests that the immobilization of liposomes should require a certain minimum number of the hydrophobic moieties anchored in the liposomal membrane. By applying the present immobilization-detachment system, preparation of liposomes encapsulating the dextran derivative without using costly gel filtration or ultracentrifugation procedure was successfully demonstrated. © 2004 Elsevier B.V. All rights reserved. Keywords: Liposome immobilization and detachment; Cross-linked polymer gel; Hydrophobic anchor; Disulfide linkage; Liposome separation

1. Introduction Liposomes, artificial vesicles of lipid bilayer membrane, are an invaluable tool in various bio-related science and engineering fields. Encapsulation of water-soluble substances in their interior aqueous phase and accommodation of hydrophobic molecules in the lipophilic interior of the lipid bilayer membrane, with further options of the surface modification, have made liposomes a promising candidate for the sophisticated micro carrier in drug delivery systems [1]. Also, liposomes are often used as a model of biomembranes in studies of membrane proteins or cellular systems [2,3] since biomembranes and liposomes have lipid bilayer membrane structure in common. The importance of liposomes may not ∗ Corresponding author. Tel.: +81 26 269 5399; fax: +81 26 269 5424. E-mail address: [email protected] (Y. Okumura).

0927-7765/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfb.2004.06.008

be limited to the present. Liposomes have been considered as a fundamental framework of highly sophisticated chemical systems of the future [4], which would ultimately resemble biological cells. Among the various utilizations of liposomes, Sandberg et al. [5] previously described immobilization of small liposomes on cross-linked polymer gel particles. Immobilization of proteoliposomes has been also demonstrated [6]. The immobilized liposomes were tested in chromatographic applications [7–9] including studies of drug partition to lipid bilayer membrane [10–12] and interaction of proteins and the membrane [13]. These studies of liposomes immobilized on cross-linked polymer gels have demonstrated new possibilities in liposome utilization. In these previously systems, the immobilization of liposomes is intended to be permanent. For using liposomes as a stationary phase, this is appropriate since uncontrolled de-

36

Md.A. Khaleque et al. / Colloids and Surfaces B: Biointerfaces 37 (2004) 35–42

WI, USA). Fluorescein isothiocyanate conjugated-dextran (FITC-Dex, MW 45,000) was obtained from CarboMer Inc. (Westborough, MO, USA). 2.2. Chemical modification of Sephacryl S-1000 gel

Fig. 1. Immobilization and detachment of liposome. Liposome is immobilized on a polymer gel support with the hydrophobic anchors and can be liberated by reduction of the disulfide linkage.

tachment of the immobilized liposomes from the gel is simply undesirable. On the other hand, however, intentional and controlled detachment of immobilized liposomes, if possible, may further extend the application of the liposome immobilization. For example, such the system is clearly useful as a versatile support material for liposomes (typical diameter of 20–200 nm) whose small size often causes difficulty in their manipulation. In this article, in order to make the controlled detachment possible, hydrophobic octyl, hexadecyl and cholesteryl moieties, which are expected to function as the anchors by spontaneous incorporation into the hydrophobic interior of liposomal bilayer membrane, are linked via disulfide linkage to cross-linked dextran gel beads. The detachment of the immobilized liposomes can be achieved by placing the system under reductive conditions (Fig. 1).

2. Experimental

2.2.1. Reaction of Sephacryl S-1000 with 1,4-bis(2,3-epoxypropoxy)butane Sephacryl S-1000 (bed volume 39 ml) was washed successively with 50 and 25% aqueous ethanol and then with distilled water (200 ml each) on a glass filter, and the accompanying water was removed by suction with an aspirator. To the moist gel (19.6 g after the suction) in a round bottom flask was added a mixture of 14.2 g (70.0 mmol) of 1,4-bis(2,3-epoxypropoxy)butane and 42 mg (1.1 mmol) of sodium borohydride dissolved in 15 ml of 0.6 M aqueous sodium hydroxide solution. The suspension was gently stirred for 24 h at room temperature, and the gel was then washed thoroughly with 900 ml of distilled water. For the determination of the epoxy content in the gel, an appropriate amount (typically 1.2 g) of the gel was separated and suspended in distilled water making the whole volume 10 ml, and the pH of the suspension was adjusted to 7.0. To the suspension, 10 ml of 2 M aqueous sodium thiosulfate solution was added, and the liberated hydroxide was titrated with 4 mM hydrochloric acid. The gel was then washed successively with 300 ml each of distilled water, 25 and 50% aqueous acetones and dried in a vacuum chamber with phosphorus pentoxide at 100 ◦ C for 20 h. The epoxy content was calculated based on the weight of the vacuum dried gel and the amount of the hydrochloric acid consumed in the titration. 2.2.2. Conversion to thiosulfate derivative (Bunte salt) The epoxy derivative of the gel (moist, 21.0 g) was washed with and suspended in 0.02 M phosphate buffer (pH 6.86), making the whole volume 42 ml. To the gel suspension, 20 ml of 2 M aqueous sodium thiosulfate was added, and the mixture was stirred for 6 h. The resulting gel was thoroughly washed with distilled water.

2.1. Materials Sephacryl S-1000 and Sepharose 4B gels were products of Pharmacia (Uppsala, Sweden). A phospholipid assay kit (Phospholipid Test Wako), 4-(2-hydroxyethyl)piperazine1-ethanesulfonic acid (HEPES), 2,2 -dipyridyl disulfide (2-PDS), 1,4-bis(2,3-epoxypropoxy)butane, 1-octanethiol, 1-hexadecanethiol and (ethylenedinitrilo)tetraacetic acid (EDTA) were obtained from Wako Chemicals (Tokyo, Japan). Sodium borohydride and threo-1,4-dimercapto2,3-butanediol (dithiothreitol, DTT) were supplied by Kanto Chemicals (Tokyo, Japan). l-␣-Phosphatidylcholine extracted and purified from egg yolk (eggPC) was purchased from Avanti Polar Lipids (Alabaster, AL, USA). Sodium dodecyl sulfate (SDS) and 5-cholestene-3␤-thiol (thiocholesterol) were bought from Sigma (St. Louis, MI, USA). dl-Threitol was a product of Aldrich (Milwaukee,

2.2.3. Reduction of thiosulfate derivative to thiol gel The Bunte salt form of the gel (moist, 20.5 g) was suspended in 0.1 M aqueous sodium bicarbonate. To the gel suspension (whole volume, 41 ml), 13 mg (0.90 mmol) of dithiothreitol (DTT) dissolved in 9 ml of 1 mM EDTA solution was added, and the mixture was stirred for 36 h at room temperature under nitrogen atmosphere. The resulting gel (thiol gel) was washed with 500 ml of 0.1 M aqueous sodium bicarbonate (containing 1 M sodium chloride and 1 mM EDTA) and then with 100 ml of 1 mM EDTA solution. 2.2.4. Capping mercapto moieties with pyridylthio groups The thiol gel (moist, 5 g) was incubated with 2,2 dipyridyl disulfide (2-PDS; 38.8 mg, 176 ␮mol) dissolved in 200 ml of 0.1 M aqueous sodium bicarbonate for 24 h at

Md.A. Khaleque et al. / Colloids and Surfaces B: Biointerfaces 37 (2004) 35–42

room temperature under nitrogen atmosphere. The resulting gel (pyridine-capped gel, Py-gel) was washed with 500 ml of distilled water. The reaction released 2-pyridinethiol, which existed exclusively as its tautomer 2-pyridinethione [14]. The mercapto moieties in the gel were assayed based on the absorption of the released pyridinethione at 343 nm and the calibration curve obtained for l-cysteine. 2.2.5. Reaction of Py-gel with thiols Typically, 1-octanethiol (8.8 mg, 60 ␮mol) was first dissolved in 150 ml of 99.5% ethanol, and the ethanolic solution was diluted by adding 350 ml of 0.1 M HEPES buffer (pH 7.5). The thiol solution thus prepared was added to Pygel (moist, 3 g), and the mixture was stirred for 48 h at 5 ◦ C under nitrogen atmosphere. The progress of the reaction was monitored by measuring the absorbance of the supernatant at 343 nm. The resulting gel was washed with 30% aqueous ethanol until the odor of the thiol disappeared. After washing further with distilled water, the gel was stored at 4 ◦ C. The reaction with 1-hexadecanethiol was carried out in a similar manner. To introduce cholesteryl moieties, thiocholesterol (24 mg, 60 ␮mol) was dissolved in 500 ml of 90% ethanol, and the solution was mixed with the Py-gel (moist, 3.0 g) and stirred for 48 h at 60 ◦ C. By terminating the reaction before the completion, gels with less hydrophobic moieties were prepared. 2.2.6. Microscopic examination of the modified gel particles The chemically modified gel was examined under an optical microscope (Olympus IX-50, Tokyo, Japan). A small amount of moist gel particles were taken on a piece of glass slide and inspected under 20 times magnification. 2.3. Interaction of liposomes with the chemically modified gels 2.3.1. Preparation of liposomes Egg yolk phosphatidylcholine (eggPC; 60 mg) dissolved in chloroform (3.0 ml) was taken in a round bottom flask, and the solvent was gently removed on an evaporator under reduced pressure to form a thin lipid film on the flask wall. After kept in a vacuum chamber overnight, the lipid film was swelled with 0.1 M HEPES buffer (1.5 ml, pH 7.5), and the lipid was scratched off by gentle swirling with a small glass ball for 40 min. The crude lipid suspension thus formed was extruded through polycarbonate membranes (pore size 100 nm) by using a LiposoFast extruder (Avestin, Ottawa, Canada) to obtain the vesicles with approximately uniform diameter [15]. Liposomes encapsulating FITC-Dex were prepared by swelling the lipid film in FITC-Dex solution in 0.1 M HEPES buffer (1.5 ml, 0.20 mg FITC/ml, pH 7.5) followed by the extrusion. FITC-Dex not encapsulated was removed by size exclusion chromatography (see below).

37

2.3.2. Size exclusion chromatography A sample was applied on a Sepharose 4B column (length 450 mm, diameter 18 mm; eluted with 0.1 M HEPES buffer (pH 7.5)). Elution profiles were drawn for both the fluorescence (Hitachi F-3000, Tokyo, Japan; excitation at 495 nm and emission at 519 nm) and the apparent absorbance at 660 nm due to the light scattering by particles (Shimadzu UV-1700, Kyoto, Japan). The amount of the encapsulated FITC-Dex was estimated from the fluorescence intensity of a liposome sample (0.5 ml) treated with 1% SDS (3.0 ml) based on a calibration curve drawn for standard FITC-Dex solutions. 2.3.3. Phospholipid assay Throughout the study, phosphatidylcholine was assayed with Phospholipid Test Wako kit as inorganic phosphate. The absorbance of a resulting sample at 660 nm was measured, and the amount of phosphatidylcholine was determined by using a calibration curve obtained for standard phosphatidylcholine suspensions. 2.3.4. Interaction of liposomes with gel In a typical experiment, liposome suspension (15 ml, phosphatidylcholine concentration 2 mM) was mixed with the chemically modified Sephacryl gels (bearing 10 ␮mol octyl moieties, for example) in a test tube and incubated at 37 ◦ C for 24 h. The mixture was then centrifuged at 3000 rpm for 20 min, and a part of the supernatant (0.1 ml) was taken out for the phospholipid assay. In the case of liposomes encapsulating FITC-Dex, the interaction was carried out in the half scale. After the interaction, the gel was separated and washed with 0.1 M HEPES buffer (3 ml, pH 7.5) for 5–6 times using the centrifugation. A part of the gel was treated with 1% aqueous SDS for 24 h, and phospholipid and FITC-Dex released in the supernatant were assayed. 2.3.5. Treatment of liposome-interacted gel with dithiothreitol The gel separated after the interaction with liposomes was incubated with dithiothreitol (DTT) dissolved in 0.1 M HEPES buffer (5 ml, pH 7.5) for 24 h. Typically, the solution contained 20 ␮mol of DTT per ␮mol of the anchor moieties unless noted otherwise. The gel was separated from the supernatant by centrifugation, and the amounts of phospholipid and FITC-Dex were determined. The supernatant was further analyzed using size exclusion chromatography. A part of the supernatant was subjected to extraction with chloroform, and the organic phase was analyzed with thin layer chromatography for the thiol (silica gel, developed with chloroform and detected by sulfuric acid baking). In the examination of extensive interaction with DTT, the gel separated after the first DTT treatment was further incubated with a fresh batch of the DTT solution in the same manner for 24 h. Also, in the control experiments, threitol was used in place of DTT.

38

Md.A. Khaleque et al. / Colloids and Surfaces B: Biointerfaces 37 (2004) 35–42

3. Results and discussion 3.1. Chemical modification of Sephacryl S-1000 gel In prior to interaction with liposomes, commercially available Sephacryl S-1000 gel was chemically modified (Fig. 2). Sephacryl S-1000 gel has previously shown sufficient stability to similar chemical modifications, and possesses relatively large holding capacity to liposomes of the size used in the present study [6]. Mercapto moieties were introduced to the gel by modifying the procedure originally described for Sepharose 6B [16,17]. The introduced mercapto moieties were further converted to pyridyldithio groups using 2,2 dipyridyl disulfide (2-PDS) for activation and protection from undesirable oxidation. The extent of the reaction with 2-PDS revealed that approximately 100 ␮mol of the moieties was present on one gram of the dried gel (obtained after prolonged drying over P2 O5 ). The value is smaller than the one previously reported for the modification of Sepharose 6B by Ax´en et al. (270 ␮mol/g dried gel) [16]. The introduction of the thiol moieties was fairly reproducible. Hydrophobic moieties were then introduced by replacing the pyridylthio-caps with octylthio (to yield C8 -gel), hexadecylthio (C16 -gel) or cholesterylthio group (Chol-gel) by the reaction with corresponding thiols. With 1-octanethiol at 5 ◦ C, the reaction ended within 24 h, and 5.1 ␮mol of the octylthio moieties was estimated to be present on one gram of the moist swelled gel. Similarly, 3.3 ␮mol/g of hexadecylthio moieties were introduced with 1-hexadecanethiol. Under the conditions examined, the introduction of alkylthio moieties preferred low temperature. At higher temperature (30 ◦ C), the amount of the introduced alkylthio moieties dropped to approximately 1/10. In contrast, 5-cholestene-3␤-thiol (thiocholesterol), a secondary thiol, required a higher temperature (60 ◦ C) to make the reaction proceed at reasonable rate. The reaction ended after 48 h with 5.0 ␮mol/g of the

Fig. 2. Chemical modification of Sephacryl S-1000 gel.

cholesterylthio moieties on the gel. For a thiol, the maximum amount of hydrophobic moieties that can be introduced by the reaction was fairly reproducible. Also, the amount of the hydrophobic moieties on a gel can be easily controlled. For example, a C8 -gel with 65% of the maximum octyl moiety density was obtained by simply terminating the reaction at 2.5 h. The physical structure of the gel particles was stable to the series of the chemical reactions. For the three types of the gels, no degradation of the gel particles was visible with an optical microscope after the modification. 3.2. Interaction of the gel with liposomes Liposomes were prepared by the extrusion method [15] from egg yolk phosphatidylcholine (eggPC). For the estimation of the integrity of the closed liposome structure, fluorescein isothiocyanate conjugated dextran (FITC-Dex, approximate molecular weight of 45,000) was encapsulated. Liposomes thus prepared have the average diameter of approximately 150 nm [15], which was consistent with the results of the size exclusion chromatographic analysis of the liposomes using Sepharose 4B gel (data not shown). Upon coincubation of the liposome suspension with the gel that bore the hydrophobic moieties, the liposomal phosphatidylcholine in the bulk aqueous phase started to disappear. At the same time, the corresponding amount of phosphatidylcholine appeared on the gel, indicating adsorption of the phospholipid. For the three types of the gels, the time courses of the adsorption were similar (Fig. 3). The lipid adsorption completed in 24 h, and approximately 60% of the adsorption occurred in 6 h. The adsorbed phospholipid accompanied the FITC-Dex. After 24 h, FITC-Dex found on the Chol-gel corresponds to 91% of that estimated assuming the initial ratio of the encapsulated FITC-Dex and the phospholipid remained unchanged. The large holding and the small loss of the FITC-

Fig. 3. Time course of immobilization of liposomal phospholipids to gels. EggPC liposomes (phospholipid concentration, 2 mM) were incubated with the modified Sephacryl gels at 37 ◦ C. () C8 (5.0 ␮mol of anchor per gram of gel), (
) C16 (3.3 ␮mol/g) and () Chol-gels (5.0 ␮mol/g).

Md.A. Khaleque et al. / Colloids and Surfaces B: Biointerfaces 37 (2004) 35–42

Dex indicate that the interacted liposomes maintained their structural integrity and immobilized still as liposomes. For the immobilization, the hydrophobic moieties on the gel were essential. Neither unmodified nor pyridine-capped gel (Py-gel) showed the adsorption of phospholipid. The immobilization of liposomes is most likely based on the incorporation of the hydrophobic moieties into liposomal bilayer membrane (anchoring). The small leakage of the encapsulated FITC-Dex observed during the immobilization process can be explained by expected temporal disturbance of the bilayer membrane caused by the insertion of the anchor. The leakage was significantly smaller with Chol-gel than the alkyl-gels, and the less intrusive behavior of the cholesteryl moieties is consistent with the known particularly high compatibility of cholesterol molecules with phosphatidylcholine bilayer membrane [18]. Once immobilized, the liposomes firmly remained on the gel. No spontaneous release of the phospholipid from the liposome-binding gel was observed in 24 h. Also, with Cholgel, only 3% of the gel-bound FITC-Dex was liberated into the bulk aqueous phase during the same period. The immobilization of liposomes was so firm and stable that the gel can be separated from the bulk aqueous phase by ordinary filtration and gentle washing on a glass filter without causing noticeable detachment of the liposomes. Table 1 summarizes the immobilization of liposomes under the various conditions. With a sufficient amount of the gels, the immobilization can be quantitative. At the molar ratio of the liposomal lipid to the anchor moiety of 1.5, all the liposomes present in the system were immobilized. By increasing liposomes in the system, the difference among the three anchors becomes apparent. When the ratio was 3.0, the amount of the liposomes exceeded the capacity for the immobilization on C8 - or C16 -gel, and 12–18% of the liposomes were left unattached in the bulk aqueous phase. On the other hand, Chol-gel can immobilize more liposomes than the alkyl gels. At the same ratio of 3.0, the gel bearing a comparable amount of cholesteryl moieties still showed quantitative adsorption of liposomes. By increasing the ratio further to 6.0, the adsorption onto Chol-gel was finally brought to the saturation.

39

Assuming that all the anchor moieties on the gel were used for the lipid immobilization, the number of the immobilized phospholipid molecules per anchor moiety at the saturation can be calculated as 2.5 and 2.6 for C8 - and C16 -gel, respectively. Although the hexadecyl moiety is larger and more hydrophobic than the octyl one, no significant difference in the immobilization was observed between the two. For Chol-gel, 4.2 phospholipid molecules could be held per cholesteryl moiety, which were approximately 1.7 times higher than C8 -gel. Again, one can attribute the larger holding to the particularly high compatibility between phosphatidylcholine molecules and cholesteryl moiety. In fact, among the conditions examined, the highest immobilization of liposomal lipids, 23 ␮mol/g of the gel, was achieved with Chol-gel (data not shown). The incorporation of the less anchors may partly contribute to the smaller FITC-Dex leakage, thus less intrusive behavior, observed for Chol-gel. The ratio of immobilized phospholipid to anchor remained unchanged when the population density of the hydrophobic moieties on a gel was modified. Compared with the standard C8 -gel (5.0 ␮mol of octyl moieties per gram of the gel), one with the two thirds of anchor moieties (3.3 ␮mol/g) immobilized the two thirds of liposomal lipid at the saturation. Curiously, the number of immobilized phospholipid molecules per anchor for C8 -gel is well coincided with the one previously reported by Yang et al. for their non-detachable system (2.5 phospholipid molecules per octyl moiety) [6] although the structure and the anchor density of their gel were different from the present C8 -gel. The essentially same immobilization behavior was observed at 37 and 25 ◦ C (data not shown). However, the leakage of the encapsulated FITC-Dex was smaller at the lower temperature. 3.3. Treatment of immobilized liposomes with dithiothreitol The gels that were immobilizing liposomal phospholipid were brought into a solution of dithiothreitol (DTT), a reducing agent widely used for cleavage of dithio linkage in biochemical studies [19]. Upon the treatment of the gels

Table 1 Immobilization of liposomal phospholipids to gels Anchor (␮mol/g)

Phospholipid used per anchor

Phospholipid immobilized (%)

Phospholipid immobilized per anchor

FITC-Dex released (%)

C8 (5.0)

1.5 2.0 3.0 3.0

100 100 82 84

1.5 2.0 2.5 2.5

27

C16 (3.3)

1.5 3.0

100 88

1.5 2.6

14

Chol (5.0)

1.5 3.0 6.0

100 100 70

1.5 3.0 4.2

9

C8 (3.3)

Liposomal phospholipid concentration, 2 mM; interacted at 37 ◦ C for 24 h.

40

Md.A. Khaleque et al. / Colloids and Surfaces B: Biointerfaces 37 (2004) 35–42

Fig. 4. Time course of detachment of immobilized phospholipids from gels by dithiothreitol. The gels with immobilized phospholipids were treated with dithiothreitol (20 times excess of the anchors by mol) at 37 ◦ C. For the symbols, see Fig. 3.

with DTT, phospholipid appeared in the aqueous bulk phase (Fig. 4). For C8 -gel, 82–85% of the phospholipid immobilized on the gel was found detached within 6 h. For all the gels bearing liposomal lipids, similar lipid release was observed. Along with the phospholipid release, the captured FITC-Dex was also liberated from the gel. The extent of the FITC-Dex liberation approximately matched the phospholipid release. Size exclusion chromatography revealed that the released phospholipids were present as aggregates whose size was comparable to the original liposomes (data now shown). In addition, the majority (70–82%) of the liberated FITC-Dex was found associated with the lipid aggregates. These evidences indicate that the released phospholipid exists mostly as liposomes that are encapsulating FITC-Dex. The liposomes maintain their structural integrity throughout the immobilization and detachment processes. The corresponding thiol was detected in the released liposomes, attesting the cleavage of the disulfide bonds by DTT.

A similar experiment using threitol (20 ␮mol/␮mol of the anchor), which cannot attain the reduction of disulfide linkage, in place of DTT resulted in no release of the phospholipids from the gel. These observations are consistent with the present scheme (Fig. 1) in which the reductive cleavage of the disulfide bonds and the resulting detachment of the hydrophobic moieties from the gel cause the liberation of the immobilized liposomes. Table 2 summarizes the liposome release under the various conditions. In all the cases, the extent of the release fell in the range of 71–85%. Among the three different hydrophobic moieties, the release seemed to be the most extensive for octyl moieties and the least for cholesteryl although the difference was small (less than 14%). The release behavior was essentially same for the two C8 -gels different in the anchor densities. The detachment efficiency is higher than our preliminary results (58 and 52% for C8 and C16 -gel, respectively) [17]. This can be attributed to the improvement in the experimental procedure, in particular, the more strict exclusion of oxygen from the system during the detachment process. The detachment process was further examined with various concentrations of DTT (Fig. 5). For C8 -gel, the release remained at a low level until the DTT concentration reached 9 ␮mol per ␮mol of the anchor and then increased sharply to the saturation around 12 ␮mol. Clearly, a threshold at approximately 9–10 ␮mol can be seen. Similar thresholds in the DTT concentration were also observed in the detachment from C16 -gel and Chol-gel (11–12 ␮mol for C16 -gel and 12–13 ␮mol for Chol-gel). The presence of the threshold in the DTT concentration can be rationalized by our present model of the immobilization. Immobilization of a single liposome particle requires a certain minimum number of the hydrophobic moieties anchored in its membrane. At DTT concentration lower than the threshold, some disulfide linkage may be cleaved. How-

Fig. 5. Detachment of immobilized phospholipid and the amount of dithiothreitol. The gels with immobilized phospholipids were treated with various amounts of dithiothreitol at 37 ◦ C for 24 h. For the symbols, see Fig. 3.

Md.A. Khaleque et al. / Colloids and Surfaces B: Biointerfaces 37 (2004) 35–42

41

Table 2 Detachment of phospholipids from gels by dithiothreitol Anchor (␮mol/g)

Phospholipid immobilized (␮mol)

Phospholipid detached (%)

FITC-Dex detached (%)

FITC-Dex encapsulated (% of detached)

C8 (5.0)

11.2 24.7 11.2

82 83 84

84

70

C16 (3.3)

11.2 26.3

78 80

83

74

Chol (5.0)

11.2 30.0

71 76

78

82

C8 (3.3)

Dithiothreitol/anchor = 20; 37 ◦ C, 24 h.

ever, if the number of the disulfide linkages is above the minimum, no significant release of the liposomes occurs. At the concentration higher than the threshold, the number of the anchors is no longer sufficient to hold the liposomes on the gel, and the release becomes suddenly significant, even some of the anchors may still remain undetached from the gel. The release of liposomes, therefore, would accompany forced removal of the anchors out of liposomal membrane. This “ripping” of liposomes from the gel could cause disturbance to the membrane integrity and induce the leakage of the liposomal contents. In fact, a small part of the FITC-Dex (18–30% of the encapsulated) was found outside the liposomes after the detachment (Table 2). With the present system, approximately 20% of the immobilized phospholipid still remained on the gel after the treatment with DTT. Further prolonged and repeated interaction of the DTT-treated gel with a fresh DTT solution liberated only additional 2%. Along with the phospholipid, the corresponding amount of FITC-Dex was retained on the gel, indicating the presence of compartment that can still hold the aqueous FITC-Dex. The remaining phospholipid and FITC-Dex can be recovered from the gel by treatment with sodium dodecyl sulfate (SDS) although it destroys liposome structure if any. 3.4. Preparation of substance-encapsulating liposomes The present study demonstrated stable immobilization of liposomes on polymer gel particles via disulfide linkages and controlled detachment of the liposomes by reductive cleavage of the bonds. One possible application of the system is separation of liposomes from the bulk aqueous phase. Encapsulation of water-soluble substances in the interior aqueous phase of liposomes is an essential technique in any sophisticated utilization of liposomes. The protocol involves separation of liposomes from the bulk aqueous phase as a crucial step. Size exclusion chromatography or ultracentrifugation is the conventional standard procedure although those methods require costly equipments and/or a long time to completion. By attaching liposomes to a supporting material of a suitable size, liposomes may be conveniently and thoroughly sep-

arated from the bulk aqueous phase using a simpler method. This point was actually demonstrated with the present system by preparing eggPC liposomes with the extrusion method in a concentrated FITC-Dex solution. After immobilized on the gel, liposomes were conveniently separated from the unencapsulated FITC-Dex by simple filtration and washing on a glass filter. Following treatment of the gel with DTT yielded a relatively concentrated suspension of the liposomes encapsulating FITC-Dex. The amount of FITC-Dex in the bulk aqueous phase was reduced to 1/10 of the initial unencapsulated FITC-Dex. With further optimization, the procedure may be used as a novel option for separation and concentration of liposomes.

Acknowledgements This work was supported by Grant-in-Aid for Scientific Research (C) (13835003) from Japan Society for the Promotion of Science (JSPS).

References [1] D.D. Lasic, Liposomes: from Physics to Applications, Elsevier, Amsterdam, 1993. [2] R.P. Casey, Biochim. Biophys. Acta 768 (1984) 319. [3] M.K. Jain, D. Zakim, Biochim. Biophys. Acta 906 (1987) 33. [4] H. Ringsdorf, B. Schlarb, J. Venzmer, Angew. Chem. Int. Ed. Engl. 27 (1988) 113. [5] M. Sandberg, P. Lundahl, E. Greijer, M. Belew, Biochim. Biophys. Acta 924 (1987) 185. [6] Q. Yang, M. Wallsten, P. Lundahl, Biochim. Biophys. Acta 938 (1988) 243. [7] P. Lundahl, Q. Yang, J. Chromatogr. 544 (1991) 283. [8] Q. Yang, X.-Y. Liu, M. Yoshimoto, R. Kuboi, J. Miyake, Anal. Biochem. 268 (1999) 354. [9] P. Lundahl, C.-M. Zeng, C.L. Haegglund, I. Gottschalk, E. Greijer, J. Chromatogr. B 720 (1999) 103. [10] X.-Y. Liu, C. Nakamura, Q. Yang, N. Kamo, J. Miyake, J. Chromatogr. A 961 (2002) 113. [11] Q. Yang, X.-Y. Liu, S. Ajiki, M. Hara, P. Lundahl, J. Miyake, J. Chromatogr. B 707 (1998) 131. [12] X.-Y. Liu, Q. Yang, M. Hara, C. Nakamura, J. Miyake, Mater. Sci. Eng. C 17 (2001) 119. [13] M. Yoshimoto, R. Kuboi, Q. Yang, J. Miyake, J. Chromatogr. B 712 (1998) 59.

42

Md.A. Khaleque et al. / Colloids and Surfaces B: Biointerfaces 37 (2004) 35–42

[14] D.R. Grassetti, J.F. Nurray Jr., Arch. Biochem. Biophys. 119 (1967) 41. [15] R.C. MacDonald, R.I. MacDonald, B.P. Menco, K. Takeshita, N.K. Subbarao, L. Hu, Biochim. Biophys. Acta 1061 (1991) 297. [16] R. Ax´en, H. Drevin, J. Carlsson, Acta Chem. Scand. Ser. B 29 (1975) 471.

[17] Md.A. Khaleque, T. Oho, Y. Okumura, M. Mitani, Chem. Lett. (2000) 1402. [18] R.B. Gennis, Biomembranes, Springer-Verlag, New York, 1989. [19] W.W. Cleland, Biochemistry 3 (1964) 480.