Colloids and Surfaces B: Biointerfaces 95 (2012) 144–153

Contents lists available at SciVerse ScienceDirect

Colloids and Surfaces B: Biointerfaces

journal homepage: www.elsevier.com/locate/colsurfb

Preparation of vesicle drug carrier from palm oil- and palm kernel oil-based
glycosides
Nurul Fadhilah Kamalul Aripin a , Jae Won Park b , Hyun Jin Park a,∗
a
School of Life Sciences and Biotechnology, Korea University, Seoul, South Korea
b
Seafood Research and Education Center, Oregon State University, United States

a r t i c l e i n f o a b s t r a c t

Article history: A new mixture of alkyl glycosides derived from palm oil (PO) or palm kernel oil (PKO) was synthesised.
Received 17 November 2011 This mixture contains glycosylated disaccharide of either maltose or lactose with aliphatic chain that
Received in revised form 21 February 2012 varies according to the PO or PKO fatty acids composition. The synthesis method produced no poly-
Accepted 21 February 2012
merised sugar unlike the production of the commercial glycosides (APG). The mixture only contains
Available online 3 March 2012
various glycosides differing by the alkyl chain and stereoisomers. Three anomeric mixtures can be pro-
duced depending on the reaction time and catalyst: ␣-dominant mixture, ␤-dominant mixture and equal
Keywords:
mixture. The PO and PKO derived glycosides were able to form a stable vesicle with a small amount of
Niosome
Glycoside
dicetyl phosphate (DCP) and showed high vitamin E encapsulation efficiency. Low packing density of
Vesicle the membrane bilayer enabled more vitamin E to participate in the membrane formation. The anomeric
Stereochemical effect mixtures of the maltosides provide no difference in membrane packing behaviour as it was governed by
Palm oil the hydrophilic region. Significant difference in membrane packing density was observed for lactosides
Palm kernel oil anomeric mixtures because the packing behaviour was influenced by the hydrophobic region. Inclusion
of cholesterol led to decrease in vitamin E encapsulation as well as reducing the stability of the vesicle sys-
tem. The vesicular formulations of the glycosides were stable for 3 months when stored at refrigeration
temperature.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction transcutaneous applications. However, the focus of research has

been on dermal and transdermal applications as niosomes have
Niosomes have been studied extensively in recent years as drug proven potential for cosmetics. Studies showed that niosomes from
carriers other than the liposomes. Non-ionic surfactants such as commercial non-ionic surfactants e.g. ethylene oxide, sugar esters
sugar esters, glycosides and polyoxyethylene glycol can form lamel- and glycosides can enhance drug delivery through the stratum
lar phase at high water concentration, some with the help of lipid corneum, the impermeable barrier in skin [2,3].
additives. The closed lamellar phase is the typical formation of the Amongst the non-ionic surfactants, sugar esters are most com-
vesicles. Non-ionic surfactants provide a few advantages over the monly used in the development of vesicular drug delivery systems
phospholipids because they are more economical and are chem- because they are inexpensive, commonly available and since no
ically more stable as they are not easily hydrolysed or oxidised toxicity has been reported [4–9]. However, the recent develop-
during storage. Moreover, most commercial non-ionic surfactants ment of drug delivery systems from glycosides has also captured
are relatively non-toxic: they are categorised as GRAS (generally the attention of the scientific community to produce novel glyco-
regarded as safe) and are listed in many pharmacopoeias such as sides and to study their properties [10–18]. Glycosides comprise of
European Pharmacopoeia 6.0 [1]. These advantages have led to the sugar head group and aliphatic chain connected via ether linkage is
wide application of non-ionic surfactants in food, cosmetic and more stable compared to the ester linkage. In nature, glycosides or
pharmaceutical industries. glycolipids are found as component of the cell membrane of living
Niosomal formulations have been studied for drug delivery of organisms such as plants, animals and microorganisms like bacteria
oral and intravenous administrations as well as cutaneous and as well as fungi [19]. A carbohydrate head group allows cell recogni-
tion and high specific interactions with the antibody receptor [20].
Commercial glycosides known as alkyl polyglucosides (APG) com-
prised of a mixture of glucosides with a degree of polymerisation
∗ Corresponding author at: #217, College of Lifescience and Biotechnology (West
approximately 1.5 sugar units. APG is considered a natural surfac-
building), Korea University, 5-Ka, Anam-dong, Sungbuk-ku, Seoul 136-701,
South Korea. Tel.: +82 02 3290 4149; fax: +82 02 953 5892. tant because it is synthesised from natural renewable resources
E-mail address: hjpark@korea.ac.kr (H.J. Park). such as starch and vegetable oil. Thus, it is readily biodegradable

0927-7765/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.colsurfb.2012.02.032
 N.F.K. Aripin et al. / Colloids and Surfaces B: Biointerfaces 95 (2012) 144–153 145

and according to the acute toxicity tests, APG is not toxic or harm- methanol was used in HPLC analyses which were all used without
ful [21,22]. It also has high skin compatibility. Therefore, APG has further purification.
been used in the formulations of cosmetic and personal skin care
products for many years.
2.2. Synthesis method
Development of glycoside-based drug delivery systems was first
reported by Kiwada et al. [23]. Results of tissue distribution and
The glycosides were synthesised according to a previously
pharmacokinetics experiments have demonstrated that glycosides
described glycosylation procedure by Hashim et al. with minor
are more effective drug carrier for liver, lung and kidney com-
modifications [27]. The synthesis method consisted of three steps;
pared to the phosphotidylcholine. Vesicles of galactoside have high
peracetylation, glycosylation and deacetylation. PO and PKO were
affinity towards the liver specifically in the binding of hepatocytes
reduced to alcohol using a general reduction method [28]. A sus-
through specific galactose interactions. Manconi et al. conducted
pension of lithium aluminium hydride (LiAlH4 ; 33 mmol) in 40 ml
studies on cutaneous delivery system of tretinoin using APG vesi-
diethyl ether was stirred in a two-neck round bottom flask in a
cles and reported increased efficacy as compared to liposome [24].
closed condition. PO or PKO (22 mmol) in 30 ml diethyl ether was
Two types of APG were studied, Oramix NS10 (HLB = 11) showed
added slowly to the suspension through a dropping funnel to avoid
the highest skin accumulation whereas more hydrophilic Oramix
vigorous reflux. Ethyl acetate followed by water was then added
CG110 (HLB = 16) displayed high fluxes for better in vitro diffusion
drop wise to deactivate excess LiAlH4 after the final addition of
through the skin layers.
the oil. Heat produced from the reaction was cooled down by an
A new mixture of alkyl glycosides derived from palm oil (PO) and
ice-water bath. The PO or PKO alcohol mixture was washed with
palm kernel oil (PKO) was synthesised. This mixture contains gly-
100 ml cold aqueous sulphuric acid (10%) and the organic solvent
cosides that vary only on the alkyl chain moiety whereas the sugar
was evaporated.
head group remains the same with initial starting sugar as no poly-
In order to have a selective glycosylation at the anomeric carbon
merisation of the sugar occurred. Disaccharides such as maltose and
as well as to activate it, hydroxyl groups of the sugar were per-
lactose were chosen based on economical reason and the availabil-
acetylated [29,30]. Sodium acetate (121 mmol) and 100 ml acetic
ity of the material. Having a disaccharide as the hydrophilic moiety
anhydride were stirred and heated to reflux. Sugar (58 mmol) was
provides better HLB for the glycosides to accommodate long alkyl
added in a small fraction to the hot suspension. Sugar and sodium
chains from plant resources.
acetate will generate heat due to the exothermic reaction therefore
PO and PKO have different fatty acid compositions. PO contains
continuous heating was not necessary. After all sugar was added,
almost 50% monounsaturated and polyunsaturated derivatives
the solution was further heated at 120 ◦ C for 1 h. The hot solution
whilst PKO is composed mostly of saturated fatty acids [25]. These
was poured into ice-water and stirred until sticky white solid was
oils are the major resources of fatty acids. A previous study charac-
formed. A white powder was obtained after several washes with
terised the liquid crystal phases of PO and PKO derived glycosides
distilled water. Pure peracetylated sugar was recrystallised from
[26]. The targeted application for these glycosides is for vesicular
ethanol.
drug delivery system as they form lamellar phases at low water con-
Peracetylated sugar (15 mmol) and the alcohol (17 mmol) were
centration. Since these glycosides are made from food materials and
dissolved in 60 ml dichloromethane and stirred in a closed appara-
contain alkyl glycosides that are similar to APG, they are assumed
tus at room temperature. Boron trifluoride (18.2 mmol) was slowly
to be non-toxic and biocompatible. Thus, preparation of vesicular
injected into the solution and the glycosylation reaction took sev-
systems from these glycosides for cutaneous drug delivery system
eral hours to complete depending on the desired mixture. The
was investigated.
␣-dominant mixture took 48 h and ␤-dominant was 6 h, whilst the
The aim of this research was to prepare vesicular formulations
equal mixture took 24 h to complete. The reaction was stopped
from PO and PKO based glycosides in order to study their per-
by neutralising the solution with saturated sodium bicarbonate
formance as drug carriers. The investigation covers the physical
solution and the organic layer was washed two times with water.
characteristic of the vesicles such as morphology, size, zeta poten-
Dichloromethane was evaporated and acetonitrile was added later
tial and membrane fluidity, as well as encapsulation efficiency and
to the product. Hexane extractions were done a few times to com-
storage stability. The main focus was to study the effect of the
pletely remove the excess alcohol. Acetonitrile was evaporated
glycosides’ chemical structures (e.g. sugar head group, aliphatic
from the peracetylated glycosides.
mixtures and stereochemical effect) on the vesicle properties.
Finally, deacetylation step was performed by adding a cat-
The glycosylation method used to synthesise these glycosides can
alytic amount of sodium methoxide to induce a basic medium
produce different anomeric mixtures depending on the reaction
into a methanolic solution of peracetylated glycosides (50 ml). The
time and catalyst. Formulation parameters were also investi-
solution was stirred for 3 h and the progress of the reaction was
gated to achieve optimised conditions for the glycoside carrier
monitored by TLC (thin layer chromatography). After the reaction
system.
was completed, methanol was evaporated and the product was
purified from unreacted sugar with n-butanol and water extrac-
tions. A small amount of diluted sulphuric acid was added to
2. Materials and methods
neutralise the excess sodium methoxide. N-butanol was evapo-
rated and the product was dried in a vacuum oven at 50 ◦ C for
2.1. Materials
48 h.
␤-d-Lactose, d(+)-maltose monohydrate and boron trifluoride
(BF3 ) required for glycosides syntheses were purchased from Sigma 2.3. Nuclear magnetic resonance (NMR)
Aldrich (USA). PO was from processed palm oil (palm olein) and
PKO was obtained from Golden Jomalina Food Industries Sdn. PO and PKO based glycosides were characterised by 1 H NMR.
Bhd. (Malaysia). Dicetyl phosphate (DCP), cholesterol and dl- NMR spectra were recorded on Varian NMR Systems spectrom-
␣-tocopherol used in the vesicle preparations were purchased eter at 500 MHz. Methanol-d4 (CD3OD) was used in the sample
from Sigma Aldrich. The fluorescence probe, 1,6 diphenyl-1,3,5- preparation. The maltosides were measured at room temperature
hexatriene (DPH) was also purchased from Sigma Aldrich. Solvents whilst the lactosides required higher temperature of 55 ◦ C due to
for synthesis and vesicle preparation were AR grade and HPLC grade insolubility problem at lower temperature.
146 N.F.K. Aripin et al. / Colloids and Surfaces B: Biointerfaces 95 (2012) 144–153

PO maltoside (MPO): ı = 5.33–5.36 (m, 1.2H, CH ), carried out using a Cary Eclipse spectrofluorometer (Varian, USA)
5.15–5.16 (d, 1H, H-1 ), 4.77 (d, ␣H-1, J = 3.8 Hz), 4.26 (d, ␤H- with a manual polariser at room temperature (25 ◦ C). The excita-
1, J = 7.6 Hz), 3.20–3.92 ((m, bulk sugar signals), 2.76–2.79 (m, 0.1H, tion and emission wavelengths were 360 and 429 nm respectively.
CH CH2 CH ), 2.01–2.07 (m, 2.5H, CH2 CH ), 1.53–1.64 (m, Membrane fluidity of the vesicles were determined from the mea-
2H, CH2 ), 1.29–1.32 (m, 23H, CH2 ), 0.90 (t, 3H, CH3 ). surement of fluorescence anisotropy (r), which was calculated by
PKO maltoside (MPKO): ı = 5.35 (m, 0.3H, CH ), 5.15 (m, 1H, the following equation:
H-1 ), 4.77 (d, ␣H-1, J = 3.6 Hz), 4.27 (d, ␤H-1, J = 8.0 Hz), 3.20–3.90
I VV − IVH
(m, bulk sugar signals), 2.02–2.07 (m, 0.6H, CH2 ), 1.53–1.64 (m, r=
IVV + 2IVH
2H, CH2 ), 1.29 (m, 20H, CH2 ), 0.90 (t, 3H, CH3 ).
PO lactoside (LPO): ı = 5.33–5.36 (m, 1H, CH ), 4.77 (d, ␣H-1, where IVV and IVH were the fluorescence intensity of the emit-
J = 4.0 Hz), 4.37 (d, 1H, H-1 ), 4.28 (d, ␤H-1, J = 7.6 Hz), 3.25–3.89 (m, ted light polarised parallel and perpendicular to the exciting light,
bulk sugar signals), 2.76–2.79 (m, 0.1H, CH CH2 CH ) 2.02–2.05 respectively.
(m, 2H, CH2 ), 1.61–1.65 (m, 2H, CH2 ), 1.29–1.33 (m, 23H,
CH2 ), 0.90 (t, 3H, CH3 ). 2.8. Determination of encapsulation efficiency
PKO lactoside (LPKO): ı = 5.35 (m, 0.2H, CH ), 4.77 (d, ␣H-
1, J = 3.6 Hz), 4.35–4.38 (m, 1H, H-1 ), 4.28 (d, ␤H-1, J = 8.0 Hz), The vesicle formulations were purified by separating the free
3.23–3.92 (m, bulk sugar signals), 2.02–2.04 (m, 0.5H, CH2 ), vitamin E using gel permeation chromatography (GPC) with
1.60–1.65 (m, 2H, CH2 ), 1.29–1.32 (m, 20H, CH2 ), 0.90 (t, 3H, Sephadex G25. The vesicles were disrupted by adding isopropanol
CH3 ). (sample to isopropanol, 1:1, v/v) into the formulation to extract
the vitamin E and samples were analysed using HPLC. Vitamin E
2.4. Vesicle preparation was determined at 295 nm via reverse phase C18 column (Mightysil
RP-18 GP, 4.6 mm × 150 mm (5 ␮m), Kanto Chemical) operating at
Glycosides vesicles were prepared by the thin film method room temperature on a Waters 2690 Separation Module (Waters
reported by Bangham et al. [31]. All formulations were prepared Associates, USA) and a Waters 996 photodiode Array Detector
at 2 ␮mol/ml total lipid concentration except for higher total lipid (Waters Associates). The mobile phase was 100% methanol at a
concentration formulations. The formulation consisted of glyco- flow rate of 1.0 ml/min. Different concentrations of vitamin E in
side, DCP and cholesterol. Ratio of glycosides to DCP was always methanol were used to construct the calibration curve. Encapsula-
1:0.2 whilst cholesterol and vitamin E were added depending on tion efficiency was defined as:
the investigated amount. First, the glycosides, DCP, cholesterol and encapsulated vitamin E (mol)
vitamin E were dissolved in a mixture of chloroform and methanol encapsulation efficiency (%) = × 100
initial vitamin E (mol)
(4:1, v/v). The solvent was then evaporated to form a thin film and
dried under vacuum condition for another hour to ensure com-
plete removal of organic solvents. Later, the thin film was hydrated 2.9. Storage stability studies
using 25 ml deionised water in a shaking water bath at 55 ◦ C (which
is above the transition temperature of the glycosides) for an hour Glycosides vesicle formulations with and without cholesterol
and hydration period was extended to overnight at room tempera- were stored in the dark at 4 ± 1 ◦ C and 25 ± 3 ◦ C for 3 months.
ture. Finally, the vesicles were sonicated to obtain small unilamellar Released vitamin E was separated from the encapsulated vitamin
vesicles (SUV) using a Sonics Vibra-Cell CV33 ultrasonic probe (Son- E by GPC. The remaining vitamin E was measured with HPLC. HPLC
ics & Materials, USA) for 10 min (pulse on, 10 s; pulse off, 10 s) at samples were prepared the same way as mentioned above.
150 W. All formulations were protected against light at all times
during the preparation process. 2.10. Statistical analysis

2.5. Transmission electron microscopy (TEM) Data analysis was performed using the software package
Microsoft Excel, version 2007. Results are expressed as the
The morphology of the glycosides vesicles was observed using mean of three experiments ± S.D. Statistical data were ana-
a Tecnai G2 F30 microscope (Philips-FEI, Holland). The dispersed lysed using Student’s t-test with p ≥ 0.05 as a minimal level of
vesicles were dropped onto a carbon film-covered copper grid. The significance.
excess dispersion was blotted off with filter paper and air-dried
overnight. TEM studies were conducted at KBSI (Seoul). 3. Results and discussion

2.6. Size and zeta potential measurements 3.1. Synthesis and characterisation of PO and PKO glycosides

Vesicle average size, polydispersity index (PDI) and zeta poten- PO and PKO based glycosides were synthesised using glycosyla-
tial were determined by quasielastic laser light scattering using a tion method adapted from the method used to synthesis branched
Zetasizer ZEN 3600 Nano Series apparatus (ZEN, UK). The measure- chain glycosides [27]. The synthesis route consists of three major
ments were made with dilute solutions at 25 ◦ C. steps: preparation of the starting materials (reduction of PO and
PKO and peracetylation of sugar), glycosylation and deacetylation
2.7. Fluorescence anisotropy measurements (Fig. 1). In this case, the purification step involving the separation
of anomers using the chromatography technique was omitted. By
DPH loaded vesicles were prepared by adding a small amount of omitting the purification step, the costs of production could be
DPH solution in tetrahydrofuran into the mixture of glycosides and reduced because chromatography consumes a lot of organic sol-
DCP in chloroform and methanol (4:1, v/v). The molar ratio of gly- vents.
cosides to DPH was 200:1. The organic solvents were evaporated Glycosylation of PO and PKO glycosides were done at a certain
and thin film was hydrated with water at 55 ◦ C for 30 min. The for- reaction time using boron trifluoride (BF3 ) as the catalyst. The reac-
mulation was protected from the light during preparation to avoid tion produced a mixture of ␣ and ␤ anomers which configuration
degradation of DPH. Fluorescence anisotropy measurements were differs at C1 carbon. Three types of mixtures could be obtained from
 N.F.K. Aripin et al. / Colloids and Surfaces B: Biointerfaces 95 (2012) 144–153 147

Fig. 1. Synthesis steps of PO and PKO based glycosides.

this synthesis method: ␣-dominant and ␤-dominant mixtures and

mixture containing equal amount of anomers (denoted as equal
mixtures). The ratio of ␣ and ␤ anomers was controlled by factors
such as reaction time and type of catalyst because the anomers
were obtained in different pathways. The ␣ anomer formed from
the thermodynamic pathway required a strong catalyst, whilst the
␤ anomer is more kinetically favoured, can be produced by a mod-
erately active catalyst. Tin tetrachloride (SnCl4 ) can also be used to
catalyse the reaction as it produced more ␣ anomers. However, this
approach is not propitious since it tends to react with the double
bond thus reducing the unsaturation degree of the PO glycosides
[26].
For a ␤-dominant mixture, a reaction time of 6–7 h gave a sat-
isfying yield of 35–75%. An equal mixture took 24 h whereas an
␣-dominant mixture needed a longer reaction time of 48 h. These
reaction times were applicable for both maltosides and lactosides
apart from a slight difference on the ␣/␤ ratio of the ␤ domi-
nant mixture. The ␣/␤ ratio was determined by comparing the
integration of nuclear magnetic resonance (NMR) spectra peak of
both anomers. An ␣ anomer showed a double peak at 4.77 ppm Fig. 2. NMR spectra showing integrations of ␣ and ␤ anomers peaks for a ␤ dominant
and ␤ anomer at 4.26–4.27 ppm (Fig. 2). Estimations were done mixture.
148 N.F.K. Aripin et al. / Colloids and Surfaces B: Biointerfaces 95 (2012) 144–153

with proper baseline and phase correction to reduce the systematic

error.
Quality of the product was checked with the NMR spec-
tra. To ensure problems from starting material contamination
did not arise, the ratio of sugar to lipid chain was calculated.
Excess alcohol was used during synthesis and was eliminated by
acetonitrile–hexane extraction. Free alcohol elimination after the
deacetylation of the sugar was impossible due to the polarity of
both compounds is quite similar. Nevertheless, alcohol peaks were
not easily detected in the spectra. This was because most of the
signals overlapped with the alkyl chain of the glycosides. Thus, the
integration of sugar peaks was compared to the CH2 peak.
Other information obtained from the NMR spectra were the
unsaturation degree and average alkyl chain length of the glyco-
sides. Analyses from the spectra showed that no degradation of
the double bond occurred during the synthesis. The unsaturation
degree of PO and PKO derived glycosides was according to the initial Fig. 3. Effect of cholesterol on the vesicle size.

amount found in the oil. The average alkyl chain for PO glycosides
as estimated from the NMR spectra was 17 carbons and PKO glyco- The effect of cholesterol addition on the vesicle sizes of the gly-
sides consisted of 14 carbons. The results were in agreement with cosides is shown in Fig. 3. MPO displayed similar sizes at mol ratio
the previous work [26]. of cholesterol of 0.2 and 0.4. But as more cholesterol was added,
sizes decreased markedly. As for MPKO vesicles, the size decreased
with addition of cholesterol at mol ratio of 0.4 and 0.8. The same
3.2. Vesicle formation, morphology and size was not found with the lactosides. Whilst LPO maintained similar
vesicle size, LPKO had an increase in size as the cholesterol amount
PO and PKO based glycosides formed vesicles by swelling the increased. It is known that cholesterol can strengthen the packing
dry surfactant mixture deposited in the round bottom flask after assembly in the hydrophobic region by filling in the crevices caused
the evaporation of the organic solvents. This was done with the by incompact surfactant packing. Therefore, in less packed mem-
addition of water at an elevated temperature along with moderate brane such as the maltosides, cholesterol provided close packing
shaking to ensure proper hydration. Hydration at higher tempera- thus reducing the size. As for lactosides particularly LPKO which
ture was necessary because the glycosides have a high gel to liquid has a compact membrane assembly, addition of cholesterol dis-
transition temperature in order to form a more fluid membrane. The turbed the packing formed by the glycosides thus increasing the
vesicular formulations were then sonicated to obtain unilamellar vesicle radius to establish a stable vesicle. A similar profile has been
and smaller-sized vesicles. observed with Span 40 vesicles [36].
Table 1 shows the characteristics of the vesicles for all glyco- TEM observations (Fig. 4) showed that vesicles of all glyco-
sides. Mean size of the PO glycosides was relatively alike, whilst side formulations were nearly spherical. LPO displayed microscopic
the LPKO formed smaller vesicles than its maltoside counterpart. aggregations apart from the vesicles. LPO was the least stable for-
In this case, vesicles formed from lactosides or maltosides did not mulations when it flocculated into large masses after 6 days. Since
differ much in terms of size. The mean sizes of octyl glucoside and the process of preparing the TEM samples involved drying of the
octyl galactoside micelles were similar despite the use of different medium, the flocculation might happen faster than when in solu-
sugar head group [32]. Furthermore, even though PO had longer tion. Drying also might have increased the vesicle sized marginally
average aliphatic chain, no significant difference in size was evi- for all samples.
dent (p > 0.05). This was probably due to the heterogeneity of the
aliphatic chains in the oil, since chain length can influence vesicle 3.3. Membrane fluidity
size of the sugar esters, as vesicles are larger with longer alkyl chain
[33]. The polydispersity index (PDI) indicated that all formulations Thermotropic liquid crystal studies have shown that packing
were reasonably multi-dispersed niosomes. The zeta potential of density of the lipophilic region is mainly attributed to the stabil-
the formulations showed more or less similar surface charges pro- ity of the glycosides’ liquid crystal phase [37]. A close or compact
vided that both sugars had a total of two sugar units. packing requires stronger energy to disrupt the assembly, thus
All anomeric mixtures of PO and PKO lactosides formed similar- increasing the thermal stability of the phase. In order to relate
sized vesicles (Table 1). On the other hand, PO and PKO maltosides this to the packing density of the vesicle membrane, membrane
showed a slight difference in size amongst the mixtures. Whilst fluidity was studied by the fluorescence anisotropy measurement.
␣-dominant and equal mixtures were similar, the ␤-dominant This method has been widely applied to monitor structural changes
mixture formed larger vesicles. In contrast, in the case of LPKO in synthetic membranes as well as biomembranes [38]. In the
maltosides, ␣ dominant mixture had larger vesicles than vesicles approach, a fluorescence probe is embedded in the hydrophobic
formed from equal and ␤-dominant mixtures’. region of the vesicle, allowing it to sense the fluidity gradient in
Inclusion of cholesterol in a vesicle formulation is required to the bilayer membrane [39]. In this study, DPH was selected, as it
form a stable vesicle in terms of reduced permeability of water is one of the most efficient probes for studying the fluidity in the
into the vesicle. Preparation of PO maltoside without cholesterol hydrophobic region [40].
demonstrated that the glycoside could not form the vesicles, but For disaccharide glycosides, ␤-anomers showed higher stability
instead formed micelles. Based on the critical packing parameter compared to the anomers [41]. This is because an aliphatic ␤ link-
(CPP) theory, a surfactant tends to assemble as vesicle when the age makes the glycoside structure more linear, thus resembling a
CPP value is 0.5–1 (a surfactant with CPP value more or less than the rod shape which enables close packing assembly. ␤-Dominant mix-
given range will formed inverted and noninverted micelles respec- tures of LPKO showed a higher value of anisotropy compared to the
tively) [34]. The reported CPP value for tetradecyl maltosides is ␣-dominant mixture (Fig. 5) implying that the former had more
0.877 [35]. a rigid membrane due to a closer packing assembly within the
 N.F.K. Aripin et al. / Colloids and Surfaces B: Biointerfaces 95 (2012) 144–153 149

Table 1
Mean size, polydispersity index (PDI) and zeta potential of the glycosides.

Mean size (nm) PDIa Zeta potentiala (mV)

␣ dominant Equal ␤ dominant

MPOb 428 ± 49 356 ± 106 585 ± 97 0.44 ± 0.09 −49 ± 4

MPKOc 454 ± 41 394 ± 37 455 ± 60 0.42 ± 0.03 −41 ± 1
LPOd 497 ± 251 287 ± 101 356 ± 25 0.38 ± 0.05 −38 ± 3
LPKOe 379 ± 80 340 ± 2 312 ± 39 0.41 ± 0.05 −44 ± 8
a
PDI and zeta potential values were of equal formulations.
b
PO maltoside.
c
PKO maltoside.
d
PO lactoside.
e
PKO lactoside.

Fig. 4. TEM micrograph of the vesicle formulations (a) MPO, (b) MPKO, (c) LPO and (d) LPKO.

membrane. However, this was not the case with anomeric mix-
tures of MPO, MPKO and LPO where no significant changes
in the membrane fluidity between ␣ and ␤ dominant mix-
tures (MPOa: 0.16 ± 0.01, MPOb: 0.16 ± 0.01; MPKOa: 0.14 ± 0.01,
MPKOb: 0.12 ± 0.01; and LPOa: 0.29 ± 0.01, LPOb: 0.27 ± 0.01)
(p > 0.05).
In order to explain this phenomenon, it is necessary to examine
the stereochemical effect of longer alkyl chain length. Stereochem-
ical effect is reduced with longer alkyl chains’ maltosides. In one
study, the difference in thermal stability decreased from 40 to 33 K
when comparing pure anomers of 12 and 14 carbon chain malto-
sides [42]. Plus both anomeric mixtures formed similar lyotropic
phases in water [26]. This can be further explained from the struc-
tural perspective. A ␣1 → 4 linkage between two glucose units in
the sugar head group led to the bending of the molecular structure
and so form a less compact assembly for maltosides. Moreover, the
packing assemblies in the membrane bilayers has weaker intralayer
Fig. 5. Fluorescence anisotropy, r of various glycoside formulations differing sugar hydrogen bonding and relies more on interlayer hydrogen bond-
head group, alkyl chain mixture and dominating anomers. ing [42]. Therefore, having a ␣ or ␤ configuration for the aliphatic
150 N.F.K. Aripin et al. / Colloids and Surfaces B: Biointerfaces 95 (2012) 144–153

chain does not provide any significant changes in membrane membranes with a lower density and so could participate in the
flexibility because the packing behaviour of the membrane is formation of vesicles membranes along with the glycosides.
governed mostly by the sugar moiety. Ribier and Handjani-Vila The encapsulation efficiency results agreed with the membrane
concluded that the fluidity is not only affected by the hydrophobic fluidity results. Mixtures with high membrane fluidity have lower
region but by the hydrophilic region as well [43]. packing density therefore provide more space for vitamin E to
As for anomeric mixtures of LPO, the anisotropy profile can- participate in the membrane formation. A closely packed mem-
not be explained in the same way as the maltosides. Contrary to brane, such as the ␤-dominant mixture of LPKO, will tend to
maltosides, lactosides display an increased difference in thermal exclude most of the vitamin E, leading to lowered encapsulation
stability as alkyl chains lengthen [41]. Structure wise, lactose is efficiency. In case of LPO, the ␣- and ␤-dominant mixtures did
more linear in shape due to the ␤1 → 4 linkage between the two not concur with the fluorescence anisotropy results. Although both
sugars. Therefore, a compact packing assembly can form. In addi- anomeric mixtures had similar membrane fluidities the encapsu-
tion, the axial hydroxyl group at the C4 of galactose unit favours lation efficiencies were totally different. The ␤-dominant mixture
intralayer hydrogen bonding. A previous study reported that the encapsulated only 38% of vitamin E, whereas encapsulation was
transition temperature for lactosides of the same alkyl chain is 100% for the ␣-dominant mixture. Despite high packing density
higher than for maltosides [30]. Anisotropy measurements indi- of the ␣ dominant mixture, somehow the vitamin E was able to
cate that the membrane of lactosides anomeric mixtures is less incorporate in the membrane as much as with the other mixtures.
fluid than that of the maltosides. However, this does not explain Apart from the impact of the surfactant structure, vesicu-
why LPO anomeric mixtures did not show differences between lar formulation preparation factors were also considered. Factors
␣- and ␤-dominant mixtures. One way to look at it is the pack- such as cholesterol amount, drug quantity and total lipid
ing arrangement in the hydrophobic region. LPO is a PO-derived concentration were observed in order to optimise encapsulation
glycoside, which contained high unsaturated alkyl chains. Vill and efficiency. Cholesterol and dicetyl phosphate were applied in the
co-workers reported a decreased in the clearing temperature of an glycosides formulations to increase the stability of the vesicle by
unsaturated derivative of 18-carbon chain maltosides compared to stabilising the membrane bilayers and to prevent the aggregation
saturated ones [41]. In one study, a difference in thermal stability within the vesicle dispersions. Cholesterol has been reported to
was more apparent for palmitoyl lactosides than oleoyl lactosides increase the encapsulation efficiency of various hydrophilic drugs
by more than 90 K [26]. This was due to the presence of a double in vesicular formulations, as it reduces the permeability of the drug
bond alkyl chain in the hydrophobic region created a disturbance from the vesicle core [44–46]. However, for hydrophobic com-
in the packing arrangement. It prevented close packing due to low pounds, it was the opposite. Lipid additives like cholesterol and
van der Waals interactions. Therefore in this case, although the ␤- dicetyl phosphate occupied the hydrophobic region, thus compet-
anomer of the lactosides had a linear shape and was able to form a ing with vitamin E for packing space in the membrane explaining
close packing, the presence of unsaturated component produced a the decreasing pattern as the cholesterol ratio increased for the
packing behaviour similar to the bent-shaped ␣-anomer. glycosides (Fig. 6b). On the other hand, LPO showed a rather incon-
Looking at equal mixtures of the maltosides, the anisotropy sistent encapsulation efficiency trend with increasing addition of
value for MPO was 0.19 ± 0.01 whilst that for MPKO was cholesterol. A formulation devoid of cholesterol was selected for
0.13 ± 0.01. The same trend was observed for the lactosides, where further investigations, as it provided the highest encapsulation effi-
LPKO (0.13 ± 0.01) had a higher membrane fluidity than LPO ciency.
(0.203 ± 0.005). For the equal mixtures, the packing behaviour was MPO and MPKO showed increased encapsulation efficiency of
likely influenced mainly by the hydrophobic region. Regardless of vitamin E up to a mol ratio of 1.0 (Fig. 6c). Decreased encapsulated
the sugar head group, glycosides derived from PO displayed higher vitamin E was observed as the vitamin E ratio exceeded the gly-
fluorescence anisotropy than the glycosides derived from PKO. cosides ratio. In case of LPO, the encapsulation efficiency started
Despite having high amount of unsaturated compounds, PO based to drop with a 1:1 mol ratio of vitamin E to glycosides. Although
glycosides managed to form a rigid membrane. This was due to the efficiency decreased, the amount of encapsulated vitamin E
the chain length effect because PO consists of long aliphatic chains was similar from ratio 0.75 to 1.25 showing that an optimal condi-
with an average of 17 carbons. Another interesting observation is tion has been achieved. The LPKO showed a similar encapsulation
that the all three lactosides showed a distinct degree of membrane efficiency of 93–98% at a different ratio. This could mean that
fluidity greater than the maltosides. It seems that the impact of LPKO is able to encapsulate more than a mol ratio 1.25. At a mol
synergistic effect was greater in lactosides. ratio of 0.75, all the glycosides formulations showed almost the
same amount of encapsulated vitamin E (MPO: 33 ± 2 ␮mol, MPKO:
36.9 ± 0.7 ␮mol, LPO: 36 ± 3 ␮mol, LPKO: 37 ± 1 ␮mol).
3.4. Encapsulation efficiency The total lipid concentration was increased from previous con-
centration of 2 ␮mol/ml to 6 ␮mol/ml to study the impact of total
In order to study the performance of these glycosides as drug lipid concentration (Fig. 6d). Formulations with concentrations
carriers, vitamin E (␣-tocopherol) was used as a model in encap- beyond this range were not favourable since the vesicles were eas-
sulation of a hydrophobic compound. Encapsulation efficiency is ily aggregated and precipitated. Even formulations of 4 ␮mol/ml
an important parameter to determine the performance of a drug and 6 ␮mol/ml were also unstable for more than a week. All gly-
vesicle. Factors such as surfactant type, formulation component cosides formulations produced high encapsulation efficiency at
and preparation method greatly influence the drug loading. In this 2 ␮mol/ml. MPO and MPKO displayed a decrease as the concen-
paper, the impacts of glycosides structure and formulation were tration increased. LPO showed decreased encapsulation efficiency
evaluated. at 4 ␮mol/ml and a slightly increased efficiency at 6 ␮mol/ml. LPKO
All maltosides mixtures showed high encapsulation efficiencies showed similar results regardless of the concentration, with a
(Fig. 6a). In spite of the dominating anomers and alkyl chain com- vitamin E encapsulation efficiency of 100%. Total lipid has been
positions, no significant differences were observed (p > 0.05). As for reported to increase the encapsulation efficiency of flurbiprofen
lactosides, the ␣-dominant formulations displayed higher encap- in Span 60 [44]. However Yoshioka et al. observed no significant
sulation efficiencies than the ␤-dominant formulations for both PO changes in the amount of carboxyfluorescein encapsulated in Span
and PKO glycosides. These results indicate the hydrophobic vitamin 80 [45]. But, in case of salicydic acid and p-hydroxyl benzoic acid in
E was encapsulated more readily within the hydrophobic region of both Span 60 and 80, an increased–decreased profile was reported
 N.F.K. Aripin et al. / Colloids and Surfaces B: Biointerfaces 95 (2012) 144–153 151

Fig. 6. Encapsulation efficiency of various glycosides’ vesicular formulations (a) different chemical structure, (b) cholesterol amount, (c) vitamin E amount and (d) total lipid
concentration.

Fig. 7. Storage stability of the vesicles encapsulating the vitamin E for various glycosides, (a) formulation prepared without cholesterol at 4 ± 1 ◦ C, (b) formulation prepared
without cholesterol at 25 ± 3 ◦ C, (c) formulation prepared with cholesterol at 4 ± 1 ◦ C and (d) formulation prepared with cholesterol at 25 ± 3 ◦ C.
152 N.F.K. Aripin et al. / Colloids and Surfaces B: Biointerfaces 95 (2012) 144–153

in another study [47] suggesting that the total lipid concentration was evident for lactosides. A mol ratio of 1:1 (glycoside:vitamin
has a different effect depending on the drug’s chemical structure. E) was the optimal amount of vitamin E encapsulation for most
glycosides. High total lipid concentration is not recommended for
vesicles preparations as it will decrease encapsulation efficiency
3.5. Storage stability
and reduce the stability of the system.
The stability of the glycoside vesicles during storage was
observed at a refrigerated temperature (4 ± 1 ◦ C) and room tem- Acknowledgements
perature (25 ± 3 ◦ C). Released vitamin E was separated using GPC
and the remaining vitamin E encapsulated within the vesicle were This work is supported by Korea University, Republic of Korea.
quantified (Fig. 7). Two types of formulations were prepared for the We are very grateful to Prof. Choi Yong Seok (Korea University) for
study: vesicular formulations with and without cholesterol. providing the facilities for the synthesis process. We are also deeply
For formulations without cholesterol, all glycosides except LPO grateful to Chemistry Department, Faculty of Science, University of
showed a similar stability profile when stored at low temperature Malaya, Malaysia for the access to the spectrofluorometer.
(Fig. 7a). More than 90% of the vitamin E was retained inside the
MPO, MPKO and LPKO vesicles. On contrary for LPO, large masses
were observed after 6 days, indicating the instability of the system References
thus the formulation was no longer observed. At room tempera-
[1] Ph. Eur. 6.0: European Pharmacopoeia 6.0, European Directorate for Quality of
ture (25 ± 3 ◦ C), all vesicles formulations became less stable. MPO Medicines, Council of Europe, Strasbourg, France, 2008.
and LPKO vesicles released more 40% vitamin E after 3 months [2] I.F. Uchegbu, S.P. Vyas, Int. J. Pharm. 172 (1998) 33–70.
storage. However MPKO still remained with 83% encapsulated vita- [3] A. Azeem, M.K. Anwer, S. Talegaonkar, J. Drug Target. 17 (2009) 671–689.
[4] A. Azeem, F.J. Ahmad, Z.I. Khan, S. Talegaonkar, J. Dispers. Sci. Tech. 29 (2008)
min E. Room temperature storage also increase the instability of 723–730.
liposomes [48]. [5] Z.S. Bayindir, N. Yuksel, J. Pharm. Sci. 99 (2010) 2049–2060.
Inclusion of cholesterol in vesicular formulations will provide [6] M. Carafa, E. Santucci, F. Alhaique, T. Coviello, E. Murtas, F.M. Riccieri, G. Lucania,
M.R. Torrisi, Int. J. Pharm. 160 (1998) 51–59.
more compact packing that reduces the permeability of the drug. [7] F. Hashim, M. El-Ridy, M. Nasr, Y. Abdallah, Drug Delivery 17 (2010) 282–287.
However, based on the findings shown in Fig. 7c, vesicles of MPO, [8] M. Manconi, C. Sinico, D. Valenti, G. Loy, A.M. Fadda, Int. J. Pharm. 234 (2002)
MPKO and LPKO remained stable up to 22 days before more vita- 237–248.
[9] A. Manosroi, P. Wongtrakul, J. Manosroi, H. Sakai, F. Sugawara, M. Yuasa, M.
min E was released. At the end of the study period, only about 50%
Abe, Colloids Surf. B: Biointerfaces 30 (2003) 129–138.
vitamin E remained encapsulated. Storage at room temperature [10] V. Vill, R. Hashim, Curr. Opin. Colloid Interface Sci. 7 (2002) 395–409.
accelerated the release of vitamin E, with decreased encapsulated [11] R. Hashim, A. Sugimura, H. Minamikawa, T. Heidelberg, Liq. Cryst. (2011),
vitamin E after 9 days for MPO, MPKO and LPKO formulations doi:10.1080/02678292.2011.614017.
[12] M. Hato, H. Minamikawa, K. Tamada, T. Baba, Y. Tanabe, Adv. Colloid Interface
(Fig. 7d). Stability of LPO vesicles remained the same. From the Sci. 80 (1999) 233–270.
observation, inclusion of cholesterol somehow affected the struc- [13] N.J. Brooks, H.A.A. Hamid, R. Hashim, T. Heidelberg, J.M. Seddon, C.E. Conn,
tural integrity of the vesicles. S.M. Mirzadeh Husseini, N.I. Zahid, R.S. Duali Hussen, Liq. Cryst. (2011),
doi:10.1080/02678292.2011.625689.
Alamelu and Panduranga Rao reported that liposomes contain- [14] J.W. Goodby, V. Görtz, S.J. Cowling, G. Mackenzie, P. Martin, D. Plusquellec,
ing unsaturated fatty acids can easily disintegrate compared to the T. Benvegnu, P. Boullanger, D. Lafont, Y. Queneau, S. Chambert, J. Fitremann,
liposomes constructed of saturated acids [49]. It was suggested that Chem. Soc. Rev. 36 (2007) 1971–2032.
[15] K. Sabah, T. Heidelberg, R. Hashim, Carbohydr. Res. 346 (2011) 891–896.
unsaturated compounds led to degradation due to the atmospheric [16] A.G. Cook, J.L. Wardell, C.T. Imrie, Chem. Phys. Lipids 164 (2011) 118–124.
oxidations occurred on the double bond. Since PO is comprised of [17] I. Nainggolan, S. Radiman, A.S. Hamzah, R. Hashim, Colloids Surf. B: Biointer-
almost 50% unsaturated compounds, it could explain the lowest faces 73 (2009) 84–91.
[18] N.I. Mat Zahid, O.K. Abou-Zied, R. Hashim, T. Heidelberg, J. Phys. Chem. C 115
remaining vitamin E of MPO vesicles for all conditions. However, it
(2011) 19805–19810.
is uncertain whether the instability shown by these glycosides was [19] V. Faivre, V. Rosilio, Expert Opin. Drug Deliv. 7 (2010) 1031–1048.
due to the same cause. Storage stability of various liposomes and [20] R. Maget-Dana, A.C. Roche, M. Monsigny, FEBS Lett. 79 (1977) 305–309.
[21] J. Steber, W. Guhl, N. Stelter, F.R. Schröder, in: K. Hill, W. von Rybinski, G.
niosomes systems also showed similar profile despite the fact that
Stoll (Eds.), Alkyl Polyglycosides Technology, Properties and Applications, VCH,
Span 60 was used in this study [50]. Weinheim, 1997, pp. 177–190.
[22] W. von Rybinski, K. Hill, Angew Chem. Int. Ed. 37 (1998) 1328–1345.
[23] H. Kiwada, H. Niimura, Y. Kato, Chem. Pharm. Bull. 33 (1985) 2475–2482.
4. Conclusions [24] M. Manconi, C. Sinico, D. Valenti, F. Lai, A.M. Fadda, Int. J. Pharm. 311 (2006)
11–19.
[25] R.J. de Vries, J. Am. Oil Chem. Soc. 61 (1984) 404–407.
Presently, equal mixtures of all glycosides possessed the best [26] N.F. Kamalul Aripin, Synthesis and characterisation of maltosides and lactosides
vesicle properties. The glycosides had high encapsulation efficien- based on alcohol mixtures obtained from palm oil and palm kernel oil, MSc
cies for vitamin E, provided by high membrane fluidity. The vesicle Thesis, University of Malaya, Kuala Lumpur, Malaysia, 2009.
[27] R. Hashim, H.H.N.Z. Abdalla Hashim, Mohd. Rodzi, R.S. Duali Hussen, T. Heidel-
sizes were comparable to commercial APG vesicles and the vesi- berg, Thin Solid Films 509 (2006) 27–35.
cles were stable for at least 3 months when stored at refrigeration [28] R.F. Nystrom, W.G. Brown, J. Am. Chem. Soc. 69 (1947) 1197–1199.
temperature. [29] A. Vogel, Textbook of Quantitative Chemical Analysis, 5th ed., Longman and
John Wiley and Sons, New York, 1989.
Packing behaviour of the maltosides was governed by the [30] V. Vill, T. Böcker, J. Thiem, F. Fischer, Liq. Cryst. 6 (1989) 349–356.
hydrophilic moieties. Therefore there was no significant difference [31] A.D. Bangham, M.M. Standish, J.C. Watkins, J. Mol. Biol. 13 (1965) 238–252.
in vesicles criteria between the anomeric mixtures. However, in the [32] T.T. Chong, R. Hashim, R.A. Bryce, J. Phys. Chem. B 110 (2006) 4978–4984.
[33] P. Balakrishnan, S. Shanmugam, W.S. Lee, W.M. Lee, J.O. Kim, D.H. Oh, D.D. Kim,
case of lactosides, packing behaviour of the hydrophobic region was
J.S. Kim, B.K. Yoo, H.G. Choi, J.S. Woo, C.S. Yong, Int. J. Pharm. 377 (2009) 1–8.
greatly affected when a perturbation such as unsaturated compo- [34] J.N. Israelachvili, D.J. Mitchell, B.W. Ninham, J. Chem. Soc. Faraday Trans. 2 (72)
nent was present. PKO glycosides were a better carrier for vitamin (1976) 1525–1568.
[35] H.J. Hinz, H. Kuttenreich, R. Meyer, M. Renner, R. Fründ, Biochemistry 30 (1991)
E compared to the PO glycosides.
5125–5138.
Addition of cholesterol to vesicular formulations reduced the [36] E.A. Essa, Asian J. Pharm. 4 (2010) 227–233.
amount of encapsulated vitamin E as it competed for packing space [37] P. Sakya, J.M. Seddon, Liq. Cryst. 23 (1997) 409–424.
within the membrane. In case of maltosides, cholesterol acted [38] G.K. Radda, in: E.D. Korrt (Ed.), Methods in Membrane Biology, vol. 4, Plenum
Press, New York, 1975, pp. 97–188.
as filler for a less compact packing, reducing vesicle sizes with [39] L. Tilley, K.R. Thulborn, W.H. Sawyer, Biol. Chem. 254 (1979) 2592–2594.
increasing amounts of cholesterol. However, the opposite effect [40] M. Shinitzky, Y. Bareholz, Biochim. Biophys. Acta 515 (1978) 367–394.
 N.F.K. Aripin et al. / Colloids and Surfaces B: Biointerfaces 95 (2012) 144–153 153

[41] H.M. von Minden, K. Brandenburg, U. Seydel, M.H.J. Koch, V. Garamus, R. [46] C. Cable, An examination of the effects of surface modifications on the physico-
Willumeit, V. Vill, Chem. Phys. Lipid 106 (2000) 157–179. chemical and biological properties of non-ionic surfactant vesicles, PhD Thesis,
[42] T.T. Chong, T. Heidelberg, R. Hashim, S. Gary, Liq. Cryst. 34 (2007) University of Strathclyde, Glasgow, UK, 1989.
349–363. [47] Y.M. Hao, F.L. Zhao, N. Li, Y.H. Yang, K.A. Li, Int. J. Pharm. 244 (2002) 73–80.
[43] A. Ribier, R.M. Handjani-Vila, Colloids Surf. 10 (1984) 155–161. [48] J. du Plessis, C. Ramachandran, N. Weiner, D.G. Müller, Int. J. Pharm. 127 (1996)
[44] M. Mokhtar, O.A. Sammour, M.A. Hammad, N.A. Megrab, Int. J. Pharm. 361 273–278.
(2008) 104–111. [49] S. Alamelu, K. Panduranga Rao, J. Microencapsul. 7 (1990) 541–551.
[45] T. Yoshioka, B. Sternberg, A.T. Florence, Int. J. Pharm. 105 (1994) 1–6. [50] R. Agarwal, O.P. Katare, S.P. Vyas, Int. J. Pharm. 228 (2001) 43–52.