Content uploaded by Zimei Wu
Author content
All content in this area was uploaded by Zimei Wu on Dec 31, 2014
Content may be subject to copyright.
RESEARCH PAPER
Strategies to Maximize Liposomal Drug Loading for a Poorly
Water-soluble Anticancer Drug
Wenli Zhang &Guangji Wang &James R. Falconer &Bruce C. Baguley &John P. Shaw &Jianping Liu &Hongtao Xu &Esther See &
Jianguo Sun &Jiye Aa &Zimei Wu
Received: 6 July 2014 /Accepted: 10 October 2014
#Springer Science+Business Media New York 2014
ABSTRACT
Purpose To develop a liposomal system with high drug loading
(DL) for intravenous (i.v.) delivery of a poorly water-soluble basic
drug, asulacrine (ASL).
Methods A thin-film hydration and extrusion method was used
to fabricate the PEGylated liposomal membranes followed by a
freeze and thaw process. A novel active drug loading method was
developed using ammonium sulphate gradient as an influx driving
force of ASL solubilized with sulfobutyl ether-β-cyclodextrin
(SBE-β-CD). DL was maximized by optimizing liposomal prepa-
ration and loading conditions. Pharmacokinetics was evaluated
following i.v. infusion in rabbits.
Results Freeze-thaw resulted in unilamellar liposome formation
(180 nm) free of micelles. Higher DL was obtained when dialysis
was used to remove the untrapped ammonium sulphate com-
pared to ultracentrifuge. The pH and SBE-β-CD level in the
loading solution played key roles in enhancing DL. High DL
ASL-liposomes (8.9%w/w, drug-to-lipid mole ratio 26%) were
obtained with some drug “bundles”in the liposomal cores and
were stable in a 5% glucose solution for >80 days with minimal
leakage (<2%). Surprisingly, following administration of ASL-
liposomes prepared with or without SBE-β-CD, the half-lives
were similar to the drug solution despite an increased area under
the curve, indicating drug leakage from the carriers.
Conclusions High liposomal DL was achieved with multiple strat-
egies for a poorly-water soluble weak base. However, the liposo-
mal permeability needed to be tailored to improve drug retention.
KEY WORDS active loading .cyclodextrin .liposomes .
pharmacokinetics .supersaturated
ABBREVIATIONS
ASL Asulacrine
ASL-L Asulacrine liposomes
Cryo-TEM Cryo-Transmission electron microscopy
DL Drug loading
DLS Dynamic light scattering
EE Entrapment efficiency
EPR Enhanced permeability and retention
PDI Polydispersity index
PEG Polyethylene glycol
PIP Post-injection precipitation
RES Reticuloendothelial system
SBE-β-CD Sulfobutyl ether-β-cyclodextrin
TFH Thin-film hydration
INTRODUCTION
Liposomes are established carriers for anticancer drugs due to
their potential of targeting solid tumors and reducing systemic
toxicity by exploiting the enhanced permeability and retention
(EPR) effect [1,2]. Controlled drug release from the liposomes
during circulation is essential to ensure enough drug to reach
the tumor site; and a sufficient drug content in the carriers
could lead to efficient drug uptake by the cancer cells [3,4].
Therefore, controlled drug release and high drug loading are
vital in liposomal preparation. However, both have often
posed a number of technical challenges to the formulation
scientists.
Drugs can be incorporated into liposomes by either a
passive or active (remote) loading method. With the former
approach, drugs are loaded during liposome formation [5];
whereas in remote loading, drugs are loaded in the preformed
liposomes with an additional trans-membrane potential as the
W. Zhang :J. R. Falconer :J. P.Shaw:H. Xu :E. See :Z. Wu (*)
School of Pharmacy, The University of Auckland, 1142 Auckland
New Zealand
e-mail: z.wu@auckland.ac.nz
W. Zhang :G. Wang :J. Liu (*):J. Sun:J. Aa
China Pharmaceutical University, 210009 Nanjing, People’s Republic of
China
e-mail: liujianpingljp@hotmail.com
B. C. Baguley
Auckland Cancer Society Research Centre, The University of Auckland
1142 Auckland, New Zealand
Pharm Res
DOI 10.1007/s11095-014-1551-8
driving force(s). Compared to passive loading, the remote
loading method has been proven to be more effective in
achieving a higher drug-to-lipid ratio due to its active
pumping mechanism and more controlled drug release by
“locking”the drug inside liposomes [6]. Various chemical
gradients can be used as an influx driving force depending
on the drug properties. Ammonium sulphate [7] and calcium
acetate [8] are employed for weak bases and acids, respective-
ly. EDTA [9], cyclodextrin [10]andtransitionmetals[11]are
used for drugs with which a less soluble complex can be
formed in the vesicles.
Although remote loading methods are superior to passive
loading, a poorly water-soluble drug seems to defy ready
encapsulation via the trans-membrane transport due to its
low concentration gradient. Zucker et al. [12] reported a
working model to predict loading efficiency based on a study
of nine model drugs with different physicochemical properties.
As described in this model, the poor aqueous solubility of drug
candidates is usually a major limitation for efficient active
loading. Boman et al. [13] have constructed an equation
describing that the drug uptake rate into liposomes was pro-
portional to the extra-liposomal concentration of the neutral
species of the drug. In practice, a high trans-membrane con-
centration gradient is difficult to achieve given the poor water
solubility of compounds especially in neutral species. To ad-
dress this problem, supersaturated drug solutions with or
without the use of cyclodextrin (CD) have been recently
explored by Anderson’s group [14] and have achieved a high
drug-to-lipid ratio (17% mole ratio). This novel strategy has
not been widely applied to different poorly water-soluble
drugs. Also, the effects of CD on the liposome physicochem-
ical stability and particularly the in vivo retention properties
have not yet been reported.
Asulacrine (ASL), 9-[2-methoxy-4-methylsulphonylamino)
phenylamino]-N, 5-dimethyl-4-acridinecarboxamide, also
known as SN 21407 or CI-921 (Fig. 1), is an analogue of
amsacrine synthesized by The University of Auckland, New
Zealand [15]. ASL is an inhibitor of topoisomerase II [16]and
has shown greater activity than amsacrine against breast and
lung cancer in clinical trials Phase I and II [17,18]. However,
phlebitis resulting from intravenous (i.v.) infusion hampered its
further development. ASL is a weakly basic drug with a
reported pK
a
value of 6.7 [19]andwasadministrated
as a 1 mg/ml solution of isethionate salt (pH ~ 4). Our
preformulation data [19] showed that the water solubility of
ASL was high in acidic solutions, however, it was remarkably
reduced at physiological pH (843 μg/l at pH 7.4), possibly
resulting in post-injection precipitation (PIP) in the vein [20,
21]. Therefore, it was envisioned that the irritancy of drug as
well as PIP [22] were the probable main causes of phlebitis for
ASL. Based on their ability to separate the encapsulated drugs
from the surrounding tissues and to prevent PIP, reducing the
degree of tissue damage [23,24], liposomes are proposed as a
new formulation for parenteral administration to improve the
venous tolerance with favorable pharmacokinetic behavior
and tumor targeting.
The aim of this paper was to develop a stable long-
circulating liposome system containing a high content of
ASL for i.v. administration. ASL also served as a “challeng-
ing”model drug with low water solubility, high log P and a
pK
a
(6.7) close to intra-liposomal pH (pH=5.6 with 250 mM
ammonium sulphate for developing strategies to improve
drug loading in liposomes for active reagents of its kind. To
obtain a high drug-to-lipid ratio (drug loading, DL) and
entrapment efficiency (EE), the remote loading using an am-
monium sulphate gradient was employed, and a negatively
charged cyclodextrin, sulfobutyl ether-β-cyclodextrin
(SBE-β-CD) was utilized to create a highly concentrated
supersaturated solution for drug loading [14]inadditionto
acidification by using its isethionate salt. The preparation
process was optimized to maximize the volume of aqueous
cores in the liposomes and trans-membrane ammonium sul-
phate gradient prior to drug loading. DL and EE were max-
imized through manipulation of the key factors affecting
extra-liposomal drug concentration and pH, SBE-β-CD levels
and loading duration. Finally, the effects of SBE-β-CD on the
liposome stability, particularly the in vivo drug retention, were
investigated following a one-hour i.v. infusion in New Zealand
white rabbits.
MATERIALS AND METHODS
Materials
The phospholipids, 1,2-dipalmitoyl-sn-glycero-3-
phospocholinemonohydrate (DPPC), N-(carbonyl-methoxy-
polyethyleneglycol 2000) -1,2- distearoyl-sn-glycero-3-
phosphoethanolamine (DSPE-mPEG 2000) were purchased
Fig. 1 Chemical structure of ASL free base.
Zhang et al.
from Lipoid GmbH (Ludwigshafen, Germany). Cholesterol was
obtained from Sigma-Aldrich Co., Ltd. Asulacrine isethionate
salt (99% pure) was synthesized and kindly provided by
Auckland Cancer Society Research Centre, The University of
Auckland. Sulfobutyl ether-β-cyclodextrin (Captisol®) (SBE-β-
CD) was a gift sample from Captisol Technology (La Jolla,
USA).Allotherreagentsusedinthisstudywereof
analytical grade except methanol and acetonitrile of
chromatographic grade.
Animals
New Zealand white rabbits were housed in individual cages in
an animal roommaintained at 23± 3°C and 55±10% relative
humidity with ventilation 13-16 times per hour and a 12-h
light–dark cycle. The rabbits were allowed free access to diet
and water. This animal study was approved by the Committee
on Animal Experiments of The University of Auckland (Ethics
Approval No. C881).
Preparation of Liposomes
with and without Freeze-thaw
The preformed liposomes (empty) were prepared according to a
thin-film hydration (TFH) method followed by extrusion. Briefly,
DPPC, DSPE-mPEG 2000 and cholesterol (mole ratios 6:1:3,
total mass 20 mg) were dissolved in 1 ml of chloroform: methanol
(3:1, v/v) and dried in an eggplant-shaped flask using a rotary
evaporator under vacuum condition (R-215, Büchi,
Switzerland). The thin film obtained was hydrated with 1 ml of
250 mM ammonium sulphate solution at 45°C for 10 min,
followed by ultrasonication for 180 s (70 amplitude)
using a probe sonicator (UP200S Ultrasonic processor,
Germany) or subjected the hydrated suspension to 3, 5,
7or10freeze-thawcyclestooptimize the process for
maximized formation of unilamellar liposomes [25]and
absorbing any potential micelles formed by the PEG-
lipid during hydration [26]. Each cycle consisted of
freezing in liquid nitrogen at−180°C for 3 min and thawing
in water bath at 45°C for 7 min. Thereafter, the liposome
suspension was extruded through 0.2 μm and then 0.1 μm
pore sized polycarbonate membrane filters (Whatman, UK)
with a stainless steel Avanti extruder fitted with two gas-tight
1 ml syringes (Hamilton, USA) in formulation development,
or a 10 ml LIPEX™Extruder (Northern Lipids Inc, Burnaby,
Canada) for animal studies. After extrusion, the size, polydis-
persity index (PDI) of the liposomes, and the mass of the
liposome pellet (after ultracentrifugation) with different num-
ber of cycles of freeze-thaw were compared.
The optimum process to prepare liposomes to load the
drug was chosen after confirmation of the absence of micelles
with cryo-transmission electron microscopy (cryo-TEM).
Optimization of Remote Drug Loading Conditions
To maximize DL and EE, various factors were investigated
including methods to remove free ammonium sulphate, con-
centrations of drug loading solutions and the presence or
absence of SBE-β-CD, loading duration and extra-liposomal
pH, as well as SBE-β-CD concentration, as follows:
(1) A trans-membrane ion gradient of ammonium sulphate
was generated by removing the untrapped ammonium
sulphate using either dialysis against an iso-osmotic NaCl
solution (100 ml for 1 ml liposomes each time) 4 times for
a total of 20 h at 37°C or ultracentrifuge at 188, 272×g
(4°C) for 1 h in a F50L-24×1.5 rotor within Sorvall®
WX 80 Ultra centrifuge (Thermo Scientific, Auckland,
New Zealand).
(2) The above blank liposomes with higher trans-membrane
ion gradient were incubated at 37°C with the same
volume of saturated (with excess drug in solid) or super-
saturated drug solution established by 5% w/w SBE-β-
CD. The final drug-to-lipid weight ratio (including the
excess drug in solid) was kept at 1:10 in both cases and the
loading solutions were adjusted to different pHs. After
incubation for 1.5 h, excess drug precipitate was removed
firstlybycentrifugeat700×gfor 10 min. Then the
supernatant containing free drug, ASL-CD complex
and SBE-β-CD was removed immediately by ultracen-
trifuge at 188, 272×g(at 4°C) for 1 h. EE and DL were
compared with different loading solutions.
(3) After the loading solution conditions were selected, the
maximum concentration of SBE-β-CD in supersaturated
drug loading solution was determined at which the integ-
rity of liposomes was not damaged. The above blank
liposomes were incubated at 37°C with the same volume
of supersaturated drug solution established by 4%, 5%,
6%, 7% w/w SBE-β-CD with drug-to-lipid weight ratio
kept at 1:10. After incubation for 1.5 h, the EE was
determined as described below.
(4) To optimize the loading pH and duration, various con-
centrated supersaturated drug solutions in the presence
of SBE-β-CD were used for drug loading, and dynamic
drug uptake process was monitored over 3 h. Blank
liposomes with a trans-membrane ammonium sulphate
gradient were incubated with drug solutions containing
the optimal ratio of SBE-β-CD with pH adjusted to 4.25
(the initial pH of drug solution as salt), 5, 5.4 and 5.8,
respectively. At each pre-determined time interval,
100 μl aliquots of ASL-L suspension were taken to deter-
mine the EE.
The finished ASL-liposomes (ASL-L) were centrifuged and
the pellets were re-suspended with a glucose solution
(5%, w/v)forfurtherstudies.
Strategies to Maximize Liposomal Drug Loading
Size, Zeta Potential, Entrapment Eff iciency
and Drug-loading of ASL-L
Particle sizes and PDI of ASL-L were measured by dynamic
light scattering (DLS) using a Malvern Nano ZS (Malvern
Instruments, UK). Samples were diluted with distilled water
to obtain liposome suspensions with lipid concentration below
10 mg/ml. Zeta potentials of liposomes were measured in
glucose solution using the same instrument. All measurements
were conducted at 25°C in triplicate.
To determine EE and DL, unencapsulated drug existed in
the form of solute and/or crystal was separated from lipo-
somes by 2 steps. Firstly, the drug precipitate was removed
from the nano-liposome suspension by centrifuge at 700×gfor
10 min. Then the supernatant containing ASL-L (C
1
) was
ultra-centrifuged at 188, 272×g(4°C) for 1 h allowing lipo-
somes to be separated from the drug solution (C
f
)including
free soluble ASL and ASL-CD complex. The concentration of
drug was determined by HPLC after extracting the drug with
acetonitrile. EE and DL were calculated using the following
formulae:
EE %ðÞ¼C1−Cf
Vtotal
Mdrug
100 ð1Þ
DL %ðÞ¼C1−Cf
Vtotal
Mlipids þMdrug
100 ð2Þ
where V
total
is the total volume of original liposomes, M
drug
and
M
lipid
is the mass of the drug and total lipids used in the
liposome preparation, respectively.
The concentration of ASL was analyzed by a validated
stability-indicating HPLC method. The HPLC method
employed an Aglient 1200 instrument, a Phenocifcifmenex
RP18 column (150× 4.6 mm, 5 μm), and a 20 mM
monopotassium phosphate (pH 2.5) buffer-acetonitrile
(72:28, v/v) mobile phase with a flow rate of 1 ml/min. The
detection wavelength was set at 254 nm and the injection
volume was 20 μl.
Cryo-Transmission Electron Microscopy
Morphology of empty and drug-loaded liposomes prepared
with different freeze-thaw cycles, were analyzed by cryo-
TEM. A drop of the sample (theoretical lipid concentration
10 mg/ml) was placed on the copper grid in the climate
chamber and blotted, leaving a thin film stretched over the
holes. The samples were shock-frozen by dipping into liquid
ethane and cooled to 90 K by liquid nitrogen. The samples
were transferred to the Tecnai 12 electron microscope (FEI,
Hillsboro, USA) operating at 120 KV. Liposomal membrane
lamellae, presence of micelles and the drug form in the aque-
ous cores were observed.
ASL-liposomes Long-term Stability
The optimized ASL-L in 5% glucose solution were stored at
4°C in the dark. The size, zeta potential and drug leakage
were monitored over 80 days. The drug leakage ratio was
calculated as the increased percentage of free drug over
storage.
Pharmacokinetics Following Intravenous Infusion
The pharmacokinetics of ASL-L prepared with SBE-β-CD
(DL =8.9% w/w) was compared with an ASL solution in 5%
glucose (pH =4), simulating that used in the clinical trial [17,
18]. To investigate the effect of SBE-β-CD on membranes,
liposomes actively loaded using a saturated drug solution at
pH 4.25 (1 mg/ml) without SBE-β-CD (DL =4.9% w/w)was
also tested. All the formulations were adjusted with 5% glu-
cose solution to the same drug concentration of 0.5 mg/ml
and sterilized by filtration. Twelve rabbits (body weight
3.0–3.5 kg) were randomly divided into 3 treatment
groups (n =4). The formulations were administrated by
a 1 h-infusion via the ear vein at ASL dose of
6.67 mg/kg through a 23 gauge plastic catheter with outer
diameter of 0.65 mm (Terumo, Tokeyo, Japan). The rate of
administration was set and maintained at 20 mg/h by a
syringe pump (Model KDS200, KD Scientific Inc., USA).
At the end of infusion (0 h) and 0.17 (10 min), 0.5, 1, 1.5, 2,
3, 5, 8, 12 and 24 h post infusion, 1 ml blood samples were
collected from the contralateral ear vein.
The blood samples in heparinized tube were immediately
centrifuged (700×g, 10 min, 25°C) to obtain the plasma. Then
100 μl of plasma was mixed with 700 μl acetonitrile and
vortex mixed for 2 min. After centrifuging the mixture, all
the supernatant was taken and dried in a SpeedVac (SVC
100 H; Savant Instruments Inc.) at 25°C. HPLC mobile phase
was added to re-dissolve the residue. The concentration of
ASL was analyzed by HPLC as described for EE determina-
tion. The drug concentration was linear ranging from 0.1 to
5μg/ml with absolute recoveries >90%.
The non-compartmental pharmacokinetic parameters
were calculated using DAS software program (version1.0,
Pharmacometrics Professional Committee of China,
Shanghai, China).
Statistical Analysis
Student T-test for data comparison was performed using
GraphPad Prism 6, version 6.01 (GraphPad Software, Inc.).
The pvalue for significance was set at 0.05.
Zhang et al.
RESULTS
Effect of Freeze-thaw on Liposome Formation
The results (Table 1and Fig. 2) showed that with the increase
of freeze and thaw cycles the size of liposomes became more
uniform (PDI values reduced). Increased volume and
mass of liposome pellets were obtained as the cycles
number increased, reaching a maximum at 7 cycles with
no significant increase at 10 cycles (p>0.05). Therefore,
7 cycles of the freeze and thaw process was used in the
rest of the studies. In addition, no significant changes to
zeta potential occurred between samples subjected to all
the cycles (Table 1).
The Cryo-TEM micrographs (Fig. 3) revealed that most
liposomes had unilamellar structures. The blank liposomes with-
out freeze-thaw treatment had fewer vesicles and exhibited a
‘sun’like image: each of the liposomes was surrounded with tiny
disks. These disks were reported to be micelles formed by PEG
lipids [26].However,therewerenodisksfoundinliposomes
prepared with 7 or 10 cycles of freeze-thaw (Fig. 3b).
The DLS (Nano ZS) gave a larger average size than TEM
as it measured the hydrodynamic diameter of the particles
[27], which could be dehydrated and therefore shrunk
(~150 nm) during TEM sample preparation.
Effect of Method of Removing Ammonium Sulphate
on EE and DL
The dialysis method appeared to be more efficient in remov-
ing the free ammonium sulphate and resulted in higher EE
and DL (98.20± 0.10 and 8.93 ±0.08% respectively) com-
pared to the ultracentrifuge method (80.43±1.25 and
7.07±0.08% respectively). Therefore, the dialysis meth-
od was employed to remove ammonium sulphate for later
drug loading.
Determination of the ASL Concentration for EE and DL
Both the pH and presence of SBE-β-CD in the loading
solution affected ASL concentration in the loading solution
directly, resulting in different EE and DL (Table 2). When
saturated solutions (with excess solid drug) were used for
drug loading, low EE (49%) and DL (4.5%) were ob-
tained at pH 4.25 and decreased dramatically as con-
centration dropped when pH increased to 5.8. In con-
trast, only when SBE-β-CD was used in drug loading,
supersaturated solutions at higher drug concentrations at
different pH were created. Drug loading was maximized
by 5.6 times with the aid of SBE-β-CD, 8.9% w/w at
pH 4.25 with no change when pH increased to 5.4.
However, EE and DL decreased as pH increased to 5.8.
Therefore the supersaturated drug solution with SBE-β-
CD as loading solution at low pH was better for drug
loading.
Optimization of SBE-β-CD Concentration
Maximized EE was achieved when 5% SBE-β-CD were
employed. Although 4% SBE-β-CD could keep ASL solubi-
lized at a supersaturated state initially, precipitation was ob-
served after loading for 1.5 h. As SBE-β-CD concentration
rose to 6 and 7%, EE dropped to 72.59± 0.23 and 23.23±
0.12% respectively without drug precipitation and in a
concentration-dependent manner along with a reduced mass
of liposome pellets.
Loading Kinetics at Different Extra-liposomal pH
As seen from Fig. 4, the maximal uptake of ASL by
liposomes was accomplished at 1 h with EE above 90%
regardless of the pH used, and the initial drug influx
was faster with elevated pH. After 1 h, however, there was a
drug efflux observed at pH 5.8 resulting in a low EE with drug
precipitation observed at this pH (where more than 10% drug
was unionized).
Therefore a loading duration of 1 h with pH at 4.25 (the
initial pH of drug solution as salt) with addition of 5% SBE-β-
CD was selected for the preparation of final formulation for
cryo-TEM observation, stability and pharmacokinetic study.
The physicochemical properties were shown in Table 3.
Figure 3C showed that drug-loaded liposomes appeared
more electron-dense under Cryo-TEM compared with empty
liposomes and a large proportion of trapped drug in liposomes
was globular. Some formed bundled or fibre-like struc-
tures (precipitates) which induced a slight change in
liposomal shape into a ‘coffee bean’appearance, similar
to doxorubicin-liposomes loaded via an ammonium sulphate
gradient method [28].
Ta b l e 1 The effect of Freeze-thaw Cycles on the Formation of Liposomes as
the Mass Collected from Tubes (mean ± SD, n=3). Liposomes Contained
10% PEGylated Lipid and no Drug
Cycles Size PDI Zeta
potential
Recovery as pellet
mass (mg)
0
a
184.7±1.53 0.089± 0.003 -47.3±1.5 2.4 ±0.6
3 183.9 ±1.37 0.060± 0.002 -48.8 ±2.1 2.8± 0.4
5 183.5 ±1.70 0.058± 0.001 -47.4 ±1.9 3.7± 0.1
7 185.2±1.27 0.047± 0.001* -48.6 ±2.0 7.0± 0.2*
10 189.2±0.36 0.044±0.001* -49.3±2.4 7.2± 0.2*
a
by ultrasonication without freeze-thaw treatment
*represent significant differences from 0, 3 or 5 cycles (p<0.05)
Strategies to Maximize Liposomal Drug Loading
Long-term Stability of ASL-L
The optimized ASL-L were prepared as following: the empty
liposomes were prepared with 7 cycles of freeze-thaw and
dialysis, then incubated with supersaturated drug solution
with 5% SBE-β-CD at 37°C for 1 h with extra-liposomal
pH at 4.25. ASL-L were centrifuged immediately and resus-
pended in 5% glucose solution (pH ~6), an i.v. infusion fluid
frequently used in clinic. Table 4shows that the size and zeta
potential did not change significantly after storage for 80 days
(p>0.05). The total drug concentration in the samples was
maintained above 95% of the original, and the drug leakage
ratio was minimal, suggesting satisfactory physical and chem-
ical stability of ASL-L.
Pharmacokinetics in Rabbits
Figure 5depicts the pharmacokinetic profiles of ASL-
liposomes prepared with (ASL-L + CD) and without (ASL-
L-CD) SBE-β-CD in comparison with the ASL solution fol-
lowing 1 h i.v. infusion in the rabbits. The drug concentration-
time profile of the ASL solution followed a one-compartment
model with logarithms of all the concentration data vs time
linear (R
2
>0.99). Although the onset concentrations of both
ASL-L + CD and ASL-L-CD were higher than those of ASL-
solution,showingabiggerareaunderthecurve(AUC)
(p>0.05), a smaller volume of distribution (V
d
)(p>0.05) and
a slower distribution process from the central bloodstream,
interestingly, the clearance phases of both ASL-Ls, prepared
with CD and without CD, conincided with that of ASL-
solution with a similar short half-life (p>0.05). Also, no signif-
icant differences (p>0.05)werefoundonallthe
phamacokinetic parameters between ASL-L prepared with
or without SBE-β-CD despite the different DL in the two
formulations.
DISCUSSION
In this study combination strategies were used to improve
drug loading into the PEGylated liposomal system for a poorly
water-soluble weak base. With the aid of a cyclodextrin as a
solubilization enhancer and the manipulation of the condi-
tions of active drug loading, a maximal DL of 8.9% w/w
(drug-to-lipid mole ratio 26%) with high EE (>98%) were
achieved. The liposomes showed little leakage over 80 days.
These results may also be extended to other drugs with similar
Fig. 2 Size and size distribution
of the liposomes subjected to
different cycles of freeze-thaw,
0cyclemeanssampleswere
treated by ultrasonication
without freeze-thaw.
Ta b l e 2 Effects of ASL Concentration on EE and DL in Preparation of ASL-liposomes. The Drug Loading Time and Temperature was 1.5 h and 37°C
Respectively (Data are means ± SD, n=3)
External drug loading medium Extra-liposomal pH
4.25 5.4 5.8
Saturated solution Initial soluble drug
concentration (μg/ml)
934.25±0.47 223.95± 0.50 26.50±0.18
DL (%) 4.53±0.21 1.60±0.04 1.57±0.06
EE (%) 49.83±0.16 17.60±0.15 17.27±0.12
Supersaturated
solution with SBE-β-CD
Initial soluble drug
concentration (μg/ml)
2,000.24±0.41 2,000.15±0.28 2,000.19± 0.36
DL (%) 8.93±0.08 8.95±0.09 6.44±0.07
EE (%) 98.20±0.10 98.47±0.14 70.84±0.18
Zhang et al.
physicochemical properties. However, the short half-life in vivo
indicated poor drug retention in the carriers despite the for-
mation of drug ‘bundles’(precipitates) in the liposomal cores.
The possible mechanisms for efficient drug loading and the
absence of long circulation are discussed below.
In a liposomal active drug loading model with an acidic
interior, the quantity of loaded amino-containing drugs was
reported to be proportional to 1) the volume of aqueous phase
of liposomes, 2) the extra-liposomal concentration of neutral
drug and 3) the trans-membrane proton gradient with a
favorable interior concentration of H
+
[29]. In the present
study of maximization of the drug loading, various factors
were investigated and optimized (Fig. 6).
Firstly, the maximum entrapment volume of aqueous cores
which accommodate the drug, was achieved by reducing the
number of lamellae of the liposomes [25]. A freeze-thaw
process was applied and unilamellar nano-sized liposomes
were obtained (Fig. 3). It was also found that sufficient freeze
and thaw was effective in preventing formation of micelles by
the PEGylated lipid, increasing the amount of liposomes. It is
well known that modification of liposomes with polyethylene
glycol (PEG) can protect liposomes from elimination by the
reticulo-endothelial system (RES) and achieve prolongation of
blood circulation lifetime. The in vivo circulation time of
PEGylated liposomes was prolonged with increased PEG
content when the ratio of PEG is within 10% (mole) [30].
However, the use of >5% PEG lipid could result in formation
of micelles [26], which were also shown in cryo-TEM photos
in this study where disk-like images were revealed, similar to
those reported [26], in samples without freeze-thaw treat-
ment. These disks disappeared when more than 7 cycles of
freeze-thaw were applied, leading to an increase in the mass of
resulting liposomes.
Of the two methods to remove the untrapped ammonium
sulphate, dialysis appeared to be more efficient than ultracen-
trifuge resulting in a higher EE and DL. With the ammonium
sulphate gradient method to load the drug, after removing the
extra-liposomal ammonium sulphate, the higher concentra-
tion of ammonium inside the liposomes caused efflux of the
neutral ammonia molecules. For every ammonium molecule
that leaves the liposome, one proton is left inside. Thus, an
imbalanced pH gradient is formed. The free uncharged ASL
base passing through the membrane would be protonated
intra-liposomally. Then the new extra-liposomal, non-
protonated ASL can diffuse into the liposome due to the
concentration gradient. Therefore a ‘proton pool’inside the
liposomes is necessary as a driving force for ASL loading and
accumulation. Apart from the protonation and charging ef-
fect, precipitation of ASL (Fig. 3) in the hydrophilic interior of
the vesicle also served as a driving force. Unlike ultracentrifu-
gation, dialysis does not only remove the external ammonium
sulphate, but it also allows an efflux of ammonium from
liposomes leaving more protons in the liposomal cores prior
to the active drug loading, which highly affects the rate of drug
uptake by liposomes. For most soluble drugs, the rate of
uptake was not so important because a similar high EE can
be achieved by extending the loading time [12]. However, it
would be crucial when a supersaturated solution was used, as
Fig. 3 Cryo-TEM micrographs of
blank liposomes (aand b), and
liposomes containing 8.9% of ASL
(c). ais sample prepared with
ultrasonication without freeze and
thaw, showing the presence of
micelles with less number of
liposomes; band cwere typical
liposomes prepared with 7 or 10
freeze-thaw cycles.
Fig. 4 Kinetic profiles of ASL uptake by liposomes at different extra-liposomal
pH in the presence of 5% SBE-β-CD (means ± SD, n=3).
Ta b l e 3 Physicochemical Stability of ASL-liposomes Stored in 5% Glucose
and Kept in the Dark at 4°C (means ± SD, n=3)
Storage time
(days)
Size (nm) PDI Zeta potential
(mV)
Leakage ratio
(%)
0 182.3±0.6 0.065± 0.020 -49.3 ±0.1 –
10 180.0±1.2 0.092±0.015 -40.8± 0.2 0.93±0.01
20 179.9±1.2 0.104±0.008 -45.6± 5.1 1.01±0.05
30 179.6±0.7 0.110±0.011 -49.2± 0.6 1.45±0.03
80 180.2±0.5 0.107±0.006 -40.2± 0.5 1.80±0.03
Strategies to Maximize Liposomal Drug Loading
a slow transport of the less soluble neutral species would result
in drug precipitation in the extra-liposomal medium.
To overcome the limitations of the low aqueous solubility
of ASL, a supersaturated drug solution at a high concentration
of 2 mg/ml was created by addition of SEB-β-CD. The
negatively charged SEB-β-CD (as sodium salt) with a hydro-
philic exterior and large molecular weight (2163 Da) could not
cross the lipid bilayers but acts as a ‘bridge’to deliver drug to
liposomes. The most important mechanism by which SBE-β-
CD increased DL and EE is its well-known solubilization
effect through the inclusion of the neutral species in the
hydrophobic cavity, allowing a rich drug solution reservoir
for drug loading. Furthermore, SEB-β-CD was able to stabi-
lize the supersaturated ASL solutions, similar to other findings
[14] which allowed time for drug to be taken by liposomes.
Secondly, ASL was a weak base (pK
a
=6.7) and the majority
was ionized in the drug loading settings (pH <5.8); one
SBE-β-CD molecule (3° substitution) could provide three
anions which may bond with a maximum of three positively
charged ASL
+
, reducing the ratio of neutral ASL and further
increasing the solubility. Drug loading could complete very
well when pH was lower than 5.8. However, an extra-
liposomal pH of 5.8 resulted in incomplete drug loading due
to drug precipitate in the supersaturated solution. This may be
explained by the fact that SBE-β-CD could not maintain the
high drug concentration with less ASL in charged form at
higher pH. In the contrast, when the liposomes were loaded
with saturated solutions, EE and DL were much lower at all
pHs and decreased as pH increased, this is probably due to the
lower solubility of ASL at high pH, leading to a lower con-
centration gradient.
The drug uptake kinetic study allowed optimization of pH-
dependent loading duration. For poorly water-soluble drugs,
the concentration of neutral species was extremely limited by
their solubility. Since ionization is a dynamic process, a suit-
able extra-liposomal pH can be utilized at which a proportion
of the drug to be loaded is ionized to warrant a high solubility
acting as a reservoir providing unionized species for liposomal
uptake. However, the drug uptake rate slowed at low pH
when less drug molecules existed as neutral form. The higher
liposomal uptake rate of drug was found with higher pH. The
results also revealed that supersaturated drug solution could
provide sufficient driving force for loading, even with less than
1% of neutral drug species with pH at 4.25 within 1 h.
Despite the long-circulating property of the PEGylated
carriers as designed, the half-life of ASL-L in rabbits was far
shorter than that of Doxil [31]. Unfortunately, a half-life less
than6h[32] is generally considered not long enough to
exploit the EPR effect for tumor-targeted drug delivery.
One-hour post i.v. infusion, the elimination phase half-lives
of both ASL-L prepared with or without SBE-β-CD (DL was
8.9 and 4.9% respectively) was almost identical with the free
drug solution, suggesting rapid drug leakage from liposomes,
although ASL-L in 5% glucose solution had a leakage ratio
less than 5% even after 80 days. Complement activation by
first injection of PEGylated liposomes has been reported in
human [33] which could result in opsonization with C3b and
iC3b and a subsequent more rapid clearance than even con-
ventional liposomes. However, if this were the case, the elim-
ination phase of the two liposome formulations (containing
different amount of total PEG-lipid) was unlikely to be iden-
tical to the free drug solution.
The use of cyclodextrins in liposomal drug loading has
significant potential to increase drug to lipid ratio and encap-
sulation efficiency. However, care should be taken as high
concentration of CDs may destabilize the liposomal mem-
brane by forming inclusion complexes with phospholipids or
cholesterol in the membrane [34,35]. It has also been report-
ed that (α-, β-, γ-) cyclodextrins can be incorporated into
phospholipid bilayers of the liposomes and lowered the mo-
lecular packing in the DPPC-membranes [35]. Therefore in
this study, the concentration of SBE-β-CD was carefully tai-
lored and SBE-β-CD was removed immediately after loading.
A concentration of 5% (DPPC/SBE-β-CD mole ratio = 3:5)
was found to preserve the integrity of liposomes while
stabilizing the drug solution for efficient loading. The
pharmacokinetic profile of this formulation showed no
statistical difference from that prepared without SBE-β-
CD, strongly supporting the view that the leakage was pre-
dominately due to the liposome membrane, rather than the
damage by SBE-β-CD.
Ta b l e 4 Pharmacokinetic Parameters of ASL Formulations Following i.v.
Infusion to Rabbits at a Dose of 6.67 mg/kg (Means ± SD; n=4)
ASL-S ASL-L+CD ASL-L-CD
T
1/2
(min) 84.28±18.41 81.70±16.77 83.81± 14.58
V
d
(L/kg) 0.71±0.20 0.56±0.22 0.46±0.20
AUC
0-t
(mg· h/L) 19.12±2.27 27.29±10.20 28.81± 4.03
Fig. 5 Pharmacokinetics profiles in rabbits after 1-h i.v. infusion of ASL
formulations at a rate of 20 mg/h (the dose was approaximatley 6.67 mg/kg)
(means ± SD, n=4).
Zhang et al.
Drug release from liposomes can be influenced by both the
membrane composition and the choice of drug [36], and
could be drug-dependent for each membrane [38]. For in-
stance, vincristine (log P=2.82) was found to leak out much
faster than doxorubicin (log P= 1.26) even using the same
remote loading approach and lipid composition, despite the
larger molecular weight of vincristine [13]. Liposome leakage
has been reported widely for lipophilic drugs, such as vinca
alkaloids [37], ciprofloxacin [38], and idarubicin [9] even
using an active loading method. In vitro drug leakage was
found, in hindsight, when pH 7.4 PBS was used as release
medium especially in the presence of serum (data were not
shown). ASL is a lipophilic molecule with logD >1 at pH 3.6
[19], which is the minimum pH within the liposomes contain-
ing 250 mM ammonium sulphate in theory [7]. Therefore,
the drug concentration in the bilayer membrane is more than
10 times higher than that in aqueous cores (not including the
precipitate as shown in TEM). Drug partitioned into the
membrane could diffuse out quickly following the large con-
centration gradient in the medium and precipitate in a pH 7.4
Fig. 6 Mechanisms of active loading using (NH
4
)
2
SO
4
gradient coupled with SBE-β-CD and various strategies used to maximize DL.
Fig. 7 Possible mechanisms of ASL
leakage from liposomes in the
blood. B is ASL base in neutral form,
BH
+
is ASL in protonated form.
Strategies to Maximize Liposomal Drug Loading
medium. The rapid drug precipitation may accelerate drug
leakage. The dynamic intra-liposomal drug dissociation equi-
librium (BH
+
and B in Fig. 7) would shift to supply more
neutral species to the membrane, eventually making drug
precipitate in the cores re-dissolve. A close pK
a
of ASL (6.7)
to intra-liposomal pH (5.5) facilitated ionization of drug (be-
come soluble) and accelerated leakage [39]. In addition, lipo-
philic drugs repartitioned within the liposomal membrane are
most likely susceptible to interaction with blood components
[19,38], resulting in a short circulation time. Therefore, the
drug retention should be optimized in further study. Attempts
to improve drug retention in liposomes have been made in the
literature, including optimization of lipid component(s) [38],
coating [40] or cross-linking membrane and forming a less-
soluble drug complex inside liposomes [9,11].
CONCLUSIONS
By optimizing the method of preparation and manipulating
the loading conditions including the use of SBE-β-CD, a high
drug-to-lipid ratio (26% by mole) was achieved with
drug precipitate being observed in the liposomal cores.
The multiple strategies to achieve high DL and EE
could be applied to other poorly-water soluble weak bases.
However, the in vivo stability of the liposomes was compro-
mised. The half-life of ASL-L was identical with that of free
drug, suggesting drug leakage from liposomes. Further study
to improve drug retention by reducing liposomal permeability
is under way.
ACKNOWLEDGEMNTS AND DISCLOSURES
This study is a New Zealand-China Research Alliance Project
funded by the New Zealand Ministry of Science and Innova-
tion (MSI) (UOAX1102) and International Science and Tech-
nology Cooperation Program of China (2011DFG33380). The
consumables were supported by a Faculty Research Develop-
ment Fund from the University of Auckland to Dr Zimei Wu.
The authors declare that they have no conflicts of interest to
disclose.
REFERENCES
1. Peer D, Karp JM, Hong S, Farokhzad OC, Margalit R, Langer R.
Nanocarriers as an emerging platform for cancer therapy. Nat
Nanotechnol. 2007;2(12):751–60.
2. Maeda H, Wu J, Sawa T, Matsumura Y, Hori K. Tumor vascular
permeability and the EPR effect in macromolecular therapeutics: A
review. J Control Release. 2000;65(1):271–84.
3. Taurin S, Nehoff H, Diong J, Larsen L, Rosengren RJ, Greish K.
Curcumin-derivative nanomicelles for the treatment of triple nega-
tive breast cancer. J Drug Target. 2013;21(7):675–83.
4.XuH,PaxtonJ,LimJ,LiY,ZhangW,DuxfieldL,et al.
Development of high-content gemcitabine pegylated liposomes and
their cytotoxicity on drug-resistant pancreatic tumour cells. Pharma
Res. 2014;1–10.
5. Cui Y, Wu Z, Liu X, Ni R, Zhu X, Ma L, et al. Preparation, safety,
pharmacokinetics, and pharmacodynamics of liposomes containing
Brucea javanica oil. AAPS PharmSciTech. 2010;11(2):878–84.
6. Gubernator J. Active methods of drug loading into liposomes: recent
strategies for stable drug entrapment and increased in vivo activity.
Expert Opin Drug Deliv. 2011;8(5):565–80.
7. Bolotin EM, Cohen R, Bar LK, Emanuel N, Ninio S, Barenholz Y,
et al. Ammonium Sulfate Gradients for Efficient and Stable Remote
Loading of Amphipathic Weak Bases into Liposomes and Ligando
liposomes. J Liposome Res. 1994;4(1):455–79.
8. Clerc S, Barenholz Y. Loading of amphipathic weak acids into
liposomes in response to transmembrane calcium acetate gradients.
Biochim Biophys Acta. 1995;1240(2):257–65.
9. Gubernator J, Chwastek G, Korycińska M, Stasiuk M, Grynkiewicz
G, Lewrick F, et al. The encapsulation of idarubicin within liposomes
using the novel EDTA ion gradient method ensures improved drug
retention in vitro and in vivo. J Control Release. 2010;146(1):68–75.
10. Li C, Cui J, Wang C, Li Y, Zhang L, Xiu X, et al.Novelsulfobutyl
ether cyclodextrin gradient leads to highly active liposomal irinotecan
formulation. J Pharm Pharmacol. 2011;63(6):765–73.
11. Taggar AS, Alnajim J, Anantha M, Thomas A, Webb M, Ramsay E,
et al.Copper–topotecan complexation mediates drug accumulation
into liposomes. J Control Release. 2006;114(1):78–88.
12. Zucker D, Marcus D, Barenholz Y, Goldblum A. Liposome drugs’
loading efficiency: A working model based on loading conditions and
drug's physicochemical properties. J Control Release. 2009;139(1):
73–80.
13. Boman NL, Mayer LD, Cullis PR. Optimization of the retention
properties of vincristine in liposomal systems. Biochim Biophys Acta.
1993;1152(2):253–8.
14. Modi S, Xiang T-X, Anderson BD. Enhanced active liposomal
loading of a poorly soluble ionizable drug using supersaturated drug
solutions. J Control Release. 2012;162(2):330–9.
15. Baguley BC, Denny WA, Atwell GJ, Finlay GJ, Rewcastle GW,
Twigden SJ, et al. Synthesis, antitumor activity, and DNA
binding properties of a new derivative of amsacrine, N-5-
dimethyl-9-[(2-methoxy-4-methylsulfonylamino) phenylamino]-
4-acridinecarboxamide. Cancer Res. 1984;44(8):3245–51.
16. Schneider E, Darkin SJ, Lawson PA, Ching L-M, Ralph RK,
Baguley BC. Cell line selectivity and DNA breakage properties of
the antitumour agent N-[2-(Dimethylamino) ethyl] acridine-4-
carboxamide: role of DNA topoisomerase II. Eur J Cancer Clin
Oncol. 1988;24(11):1783–90.
17. Fyfe D, Price C, Langley R, Pagonis C, Houghton J, Osborne L, et al.
A phase I trial of amsalog (CI-921) administered by intravenous
infusion using a 5-day schedule. Cancer Chemother Pharmacol.
2001;47(4):333–7.
18. Sklarin NT, Wiernik PH, Grove WR, Benson L, Mittelman A,
Maroun JA, et al. A phase II trial of CI-921 in advanced malignancies.
Inves New Drugs. 1992;10(4):309–12.
19. See E, Zhang W, Liu J, Svirskis D, Baguley BC, Shaw JP, et al.
Physicochemical characterization of asulacrine towards the develop-
ment of an anticancer liposomal formulation via active drug loading:
Stability, solubility, lipophilicity and ionization. Int J Pharm.
2014;473(1–2):528–35.
20. Yalkowsky SH, Valvani SC, Johnson BW. In vitro method for
detecting precipitation of parenteral formulations after injection. J
Pharma Sci. 1983;72(9):1014–7.
21. Wu Z, Hassan D, Shaw JP. In-vitro prediction of bioavailability
following extravascular injection of poorly soluble drugs: an insight
into clinical failure and the role of delivery systems. J Pharm
Pharmacol. 2013;65(10):1429–39.
Zhang et al.
22. Powis G, Kovach JS. Disposition of bisantrene in humans and
rabbits: evidence for intravascular deposition of drug as a cause of
phlebitis. Cancer Res. 1983;43(2):925–9.
23. Hu K, Zhu L, Liang H, Hu F, Feng J. Improved antitumor efficacy
and reduced toxicity of liposomes containing bufadienolides. Arch
Pharm Res. 2011;34(9):1487–94.
24. Liu X, Sun W, Zhang B, Tian B, Tang X, Qi N, et al.
Clarithromycin-loaded liposomes offering high drug loading and less
irritation. Int J Pharm. 2013;443(1):318–27.
25. Sriwongsitanont S, Ueno M. Effect of Freeze-Thawing Process on
the Size and Lamellarity of PEG-Lipid Liposomes. Open Colloid Sci
J. 2011;4:1–6.
26. Johnsson M, Edwards K. Liposomes, disks, and spherical micelles:
aggregate structure in mixtures of gel phase phosphatidylcholines and
poly (ethylene glycol)-phospholipids. Biophys J. 2003;85(6):3839–47.
27. Taratula O, Dani RK, Schumann C, Xu H, Wang A, Song H, et al.
Multifunctional nanomedicine platform for concurrent delivery of
chemotherapeutic drugs and mild hyperthermia to ovarian cancer
cells. Int J Pharm. 2013;458(1):169–80.
28. Fritze A, Hens F, Kimpfler A, Schubert R, Peschka-Süss R. Remote
loading of doxorubicin into liposomes driven by a transmembrane
phosphate gradient. Biochim Biophys Acta. 2006;1758(10):1633–40.
29. Harrigan P, Wong K, Redelmeier T, Wheeler J, Cullis P.
Accumulation of doxorubicin and other lipophilic amines into large
unilamellar vesicles in response to transmembrane pH gradients.
Biochim Biophys Acta. 1993;1149(2):329–38.
30. Lee C-M, Choi Y, Huh EJ, Lee KY, Song H-C, Sun MJ, et al.
Polyethylene Glycol (PEG) Modified 99mTc-HMPAOLiposome
for Improving Blood Circulation and Biodistribution: The Effect of
the Extent of PEGylation. Cancer Biother Radiopharm. 2005;20(6):
620–8.
31. Vail DM, Amantea MA, Colbern GT, Martin FJ, Hilger RA.
Working PK, editors Pegylated liposomal doxorubicin: proof of
principle using preclinical animal models and pharmacokinetic stud-
ies. Semin Oncol. 2004;31:16–35.
32. Fenske DB, Cullis PR. Liposomal nanomedicines. Expert Opin Drug
Deliv. 2008;5(1):25–44.
33. Moghimi SM, Hunter AC, Murray JC. Long-circulating and target-
specific nanoparticles: Theory to practice. Pharmacol Rev.
2001;53(2):283–318.
34. Piel G, Piette M, Barillaro V, Castagne D, Evrard B, Delattre L.
Study of the relationship between lipid binding properties of cyclo-
dextrins and their effect on the integrity of liposomes. Int J Pharm.
2007;338(1):35–42.
35. Puskás I, Csempesz F. Influence of cyclodextrins on the physical
stability of DPPC-liposomes. Colloids Surf B: Biointerfaces.
2007;58(2):218–24.
36. Lindner LH, Hossann M. Factors affecting drug release from lipo-
somes. Expert Opin Drug Discov Devel. 2010;13(1):111–23.
37. Zhigaltsev IV, Maurer N, Akhong Q-F, Leone R, Leng E, Wang J,
et al. Liposome-encapsulated vincristine, vinblastine and vinorelbine:
a comparative study of drug loading and retention. J Control
Release. 2005;104(1):103–11.
38. Maurer-Spurej E, Wong KF, Maurer N, Fenske DB, Cullis PR.
Factors influencing uptake and retention of amino-containing drugs
in large unilamellar vesicles exhibiting transmembrane pH gradients.
Biochim Biophys Acta. 1999;1416(1):1–10.
39. Lasic D, Čeh B, Stuart M, Guo L, Frederik P, Barenholz Y.
Transmembrane gradient driven phase transitions within vesicles:
Lessons for drug delivery. Biochim Biophys Acta. 1995;1239(2):145–56.
40. Mady MM, Darwish MM. Effect of chitosan coating on the charac-
teristics of DPPC liposomes. J Advanc Resb. 2010;1(3):187–91.
Strategies to Maximize Liposomal Drug Loading
