role of polymer and surfactant on...
TRANSCRIPT
PHYSICOCHEMICAL CHARACTERIZATION OF SELF-ASSEMBLED
NANOSTRUCTURES BY NMR SPECTROSCOPY
JUNYING ZHANG
A DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENT FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
(PHARMACEUTICAL SCIENCES)
DEPARTMENT OF PHARMACEUTICAL TECHNOLOGY
GRADUATE SCHOOL OF MEDICAL AND PHARMACEUTICAL SCIENCES
CHIBA UNIVERSITY
JAPAN
2012
CONTENTS
ABSTRACT………………………………………………………………… ……… 1
INTRODUCTION………………………...………………………………………… 3
EXPERIMENTAL
PART I NMR characterization of self-aggregation behavior of
-glucosylhesperidin in water………………………………………8
PART II Molecular-level characterization of probucol nanocrystal in water by
in situ solid-state NMR spectroscopy……….......................... ……11
RESULTS AND DISCUSSION
PART I NMR characterization of self-aggregation behavior of
-glucosylhesperidin in water…………………………………… 15
PART II Molecular-level characterization of probucol nanocrystal in water by
in situ solid-state NMR spectroscopy……….............................. …34
CONCLUSIONS …………………….…...…………………………………………45
ACKNOWLEDGEMENTS………………………………………………….. …… 47
REFERENCES…………………………………………….……………...…………48
LIST OF PUBLICATIONS……………………………………………….....………56
THESIS COMMITTEE.…………………………………………………………… 57
1
ABSTRACT
Part I
α-Glucosylhesperidin (Hsp-G), a functional food additive, significantly
enhances the solubility and bioavailability of poorly water-soluble drugs despite little
surface activity. Herein we present investigations into the underlying mechanism by
NMR techniques. A concentration dependence of the chemical shift of Hsp-G protons
correlated well with a mass-action law model, indicating self-association of Hsp-G
molecules. The critical micelle concentration was 5.0 mg/mL (6.5 mM) at 37 ºC. The
aggregation number was calculated around 2-5 according to aggregation equilibrium
and mass action law. The gradual rather than abrupt chemical shift variation upon Hsp-
G aggregation would be different mode to conventional surfactants. Dynamic light
scattering and 2D NOESY measurements demonstrated that Hsp-G molecules self-
associated into particular small micelles, with the flavanone skeleton forming a
hydrophobic core, and surrounding sugar groups working as a shell. The packing of the
hydrophobic portion is not strictly oriented and the intermolecular arrangement of
micelle shell is loose. Solubility enhancement was due to the incorporation of drugs into
Hsp-G micelle, with naringenin being more soluble than flurbiprofen. This difference is
possibly related to the structural similarities between the hydrophobic portion and the
micelle core. Hsp-G micellization process with little loss of surface tension is a unique
observation in surface and interface science.
Part II
The molecular state of colloidal probucol nanoparticles with additives was
evaluated by 13
C in situ solid-state NMR spectroscopy. The nanoparticles were obtained
by dispersing a ternary co-ground mixture of probucol/polyvinylpyrrolidon
2
(PVP)/sodium dodecyl sulfate (SDS) in water. The mean particle size was found to be
approximately 150 nm by dynamic light scattering and cryogenic-scanning electron
microscopy measurements. The results of the 13
C in situ solid-state NMR spectroscopy
showed that probucol existed in the crystalline state (form I) in water. 13
C liquid-state
NMR results indicated that PVP and SDS interacted with probucol in water. Their broad
signals suggested that the surface interaction of the probucol nanocrystal with PVP and
SDS stabilized the suspension. In addition, a freeze-dried sample of the suspension was
studied by 13
C solid-state NMR and powder X-ray diffraction experiments, which
confirmed the presence of the probucol nanocrystals. The combination of the in situ
solid-state, solid-state, and liquid-state NMR measurement results provided molecular-
level insights about the role of intermolecular interactions in the design of
nanoformulations.
3
INTRODUCTION
PART I
Oral administration is the most preferred and easiest route of drug delivery, but
its use is limited when drug molecules have poor water-solubility and permeability. One
approach to enhance the solubility and bioavailability of hydrophobic drugs taken orally
is to use a particulate delivery system soluble or dispersible in water, e.g. using
cyclodextrins1, solid dispersion systems
2-3, nanoparticles
4, and micelles
5.
Since the late 1960s, micelles have drawn considerable attention as drug carriers
owing to their ability to increase the solubility of sparingly soluble drugs in water, and
their easily controlled properties for drug release and targeting6-7
. Micelles are formed
from unique hydrophilic-hydrophobic combinations when amphiphiles are placed in
water. They consist of an inner core of assembled hydrophobic segments capable of
solubilizing poorly water-soluble drugs and an outer hydrophilic shell serving as a
stabilizing interface between the hydrophobic core and the external aqueous
environment. However, low molecular weight surfactants, especially ionic surfactants
such as sodium dodecyl sulfate (SDS), often have very high toxicity8. This delivery
system is subject to severe dilution in the gastrointestinal fluid upon oral administration.
Another amphiphiles assembly type is polymer micelles9-11
, which usually consist of
several hundred block copolymers and have diameters of approximately 10–100 nm.
The molecular weight of the hydrophobic block is usually less than 2,000, while that of
the hydrophilic part is about 1,000–12,0009. These polymeric micelle systems are
currently a growing research field for drug delivery, with the small size and the high
structural stability of polymeric micelles considered ideal for the attainment of stable,
long-term circulation of the carrier system in the bloodstream7.
4
Recently, much attention has been paid to transglycosylated food additives as
new pharmaceutical excipients in order to improve the dissolution and bioavailability of
poorly water-soluble drugs. Hesperidin, an abundant flavonoid in citrus fruits, is well
known as vitamin P. Glucosylhesperidin (Hsp-G), the transglycosylation product of
hesperidin, is superior in water-solubility, as well as being substantially tasteless and
odorless, free of toxicity, and readily hydrolyzable in vivo to hesperidin and D-glucose
to exhibit the physiological activity inherent to hesperidin12
. Tozuka et al. reported that
spray-dried powders of a water-insoluble drug and Hsp-G showed a pronounced
enhancement of dissolution when compared to solid dispersions of the drugs with
hydrophilic polymers13-14
. Moreover, in comparison with the untreated flurbiprofen (FP)
crystals, the bioavailability of FP from the prepared spray-dried powders revealed 2.5-
and 2.8-fold improvements in the Cmax and area under the curve (AUC) values,
respectively13-14
. We confirmed that 10% (w/v) concentrations of Hsp-G solution did
not have any toxic effect to Caco-2 cells, while a 0.1% (w/v) concentration of SDS
solution had higher cytotoxicity15
. The surface tension of Hsp-G at 37 °C decreased
from 72 mN/m of purified water to only 65.5 mN/m in concentrated Hsp-G solution15
,
indicating that the surface activity of Hsp-G is remarkably weaker than that of
conventional micelles such as SDS (35 mN/m). Thus, we postulated that Hsp-G formed
a specific aggregation structure in concentrated solutions, which could create a micro-
environment suitable for the entrapment and solubilization of hydrophobic drug
molecules. However, the detailed structure of this aggregate and the solubilization
mechanism of the encapsulated drugs in water has not been clearly understood.
In recent decades, NMR has become a powerful and reliable method for
structural investigation and determination. Various one-dimensional (1D) and 2D NMR
techniques have been successfully applied to study both the static (chemical
environment, degree of association, size and shape) and the dynamic properties
5
(molecular mobility, kinetics of aggregation, solubilization) of micelles16
. The 1D 1H
chemical shift appears very sensitive to subtle changes in the local environment of the
proton and proves to be a useful tool for the rapid detection of molecular association.
2D 1
H-1H nuclear Overhauser effect spectroscopy (NOESY), based on the dipole-dipole
interactions between nuclei in spacial proximity, is an effective method of studying the
relative arrangement of the aggregates and their related structure17-19
.
The aim of the present study was to clarify the specific aggregation behavior of
Hsp-G in aqueous solution showing the least surface activity, and to understand the
solubilization mechanism of Hsp-G for poorly water-soluble drugs by NMR
spectroscopy. In this study, two model hydrophobic drugs, flurbiprofen and naringenin,
were used. 1D 1H NMR and 2D
1H-
1H NOESY spectra were employed to evaluate the
aggregation of Hsp-G in water and the spatial localization of drugs.
6
INTRODUCTION
PART II
Drug nanoparticle formation has emerged as a promising strategy for enhancing
the bioavailability of hydrophobic drugs 20-22
. Information on the size or morphology of
drug nanoparticles can be obtained by analytical techniques such as light scattering or
by microscopic measurements 23
. However, only a little research has been done on the
molecular state of drug nanoparticles in a suspension 24-26
; it is considered that a large
amount of liquid water or solvent molecules obstruct the direct evaluation of the solid
molecules 24-26
. It is important to understand how drugs and additives coexist in a
suspension because the molecular state of the drugs and additives and the interaction
between them determine the stability and solubility of drug nanoparticles in the
suspension. The stability and solubility significantly affect the bioavailability and
development of the drug. Although it is possible to characterize freeze-dried samples of
a nanosuspension by various solid-state analytical techniques 27
, the molecular state of
dissolved or colloidal components in the suspension cannot always be evaluated
accurately because the freeze-drying process converts the liquid sample into solid and
sometimes causes aggregation of the colloidal particles.
In situ solid-state NMR spectroscopy involves the magic angle spinning 28
technique that reduces the dipole-dipole interactions and chemical shift anisotropy of
the solid components dispersed in water, thus enabling the measurement of an NMR
spectrum of the suspension 29-30
. Recently, in situ solid-state NMR experiments in
various fields have revealed unknown aspects of suspensions 28,31-32
. The in situ solid-
state NMR spectroscopy is a non-invasive, non-destructive tool for direct observation of
solid materials in suspensions, which may provide us useful information on the
crystallinity, type of crystal, mobility, and molecular interaction of drugs, additives, etc.
7
However, the inherent insensitivity of NMR necessitates a highly concentrated
suspension and a longer duration to perform the measurement. Moreover, in MAS,
which exerts a strong centrifugation force, the suspension is required to retain its
physical and chemical properties for a long time 24-26
. For the abovementioned reason,
the application of this approach has been restricted to the field of pharmaceutics. In this
research, we successfully prepared a stable and highly concentrated nanosuspension for
in situ solid-state NMR evaluation.
We have reported that the co-grinding of poorly water-soluble drugs in the
presence of polyvinylpyrrolidon (PVP) and sodium dodecyl sulfate (SDS) could induce
drug nanoparticle formation 33-36
. Probucol nanosuspensions showed high stability and
their bioavailability was significantly enhanced 22
. Solid-state 13
C-NMR studies
revealed that when a drug/PVP/SDS ternary ground system was dispersed in water,
grinding-induced solid-state interactions among the components of the ternary system
played an important role in nanoparticle formation 33,35
. Atomic force microscopy
proved that colloidal probucol nanoparticles were coated with the PVP-SDS complex 37
.
However, the molecular details of nanoparticles dispersed in water are still unknown.
The aim of this research was to obtain molecular-level information on probucol
and additives in a nanosuspension by in situ solid-state and liquid-state NMR
spectroscopies. In this study, probucol was used as a poorly water-soluble drug. PVP
(PVP K12, Mw ≈ 2500) and SDS were employed as the water-soluble polymer and ionic
surfactant, respectively (Figure 2). In situ solid-state and liquid-state NMR
measurements were performed for evaluating the molecular state of the nanosuspension.
A freeze-dried sample of the nanosuspension was also evaluated by powder X-ray
diffraction (PXRD) and 13
C solid-state NMR measurements, and the results were
compared with those of in situ solid-state and liquid-state NMR measurements.
8
EXPERIMENTAL
PART I NMR characterization of self-aggregation behavior of -
glucosylhesperidin in water
Materials
Flurbiprofen (FP) and naringenin (NRG) were purchased from Tokyo Kasei Co.,
Ltd. (Tokyo, Japan) and used without further purification. -Glucosylhesperidin was a
gift from Toyo Sugar Refining Co., Ltd. (Tokyo, Japan). Commercial 99.9% deuterium
oxide (D2O) (Aldrich, St. Louis, USA) was used as received. All other chemicals and
solvents were of reagent grade quality and used without further purification. The
chemical structures and proton numbering of Hsp-G, FP, and NRG are depicted in
Figure 1.
Preparation of spray-dried powders
Spray-dried powders of FP/Hsp-G or NRG/ Hsp-G were prepared using a spray-
drying method. To prepare samples by the spray-drying method, 500 mg of FP or NRG
and defined ratios of Hsp-G were dissolved in an ethanol/water solution (8:2 v/v). The
loading ratio of drug/Hsp-G was varied from 1:10 to 1:30 (w/w). This solution was fed
to a spray dryer (GS31; Yamato, Tokyo, Japan) at a rate of 10 mL/min and sprayed into
the chamber from a nozzle with a diameter of 406 μm at a pressure of 0.13 MPa. The
inlet and outlet temperatures of the drying chamber were maintained at 120 and 70 °C,
respectively. All spray-dried powders were dried in a desiccator with silica gel under
reduced pressure for 1 day before their physicochemical properties were tested.
Particle size analysis:
The volumetric particle size distribution was determined by a dynamic light
scattering method using Microtrac UPA
(Nikkiso, Japan). The detection range of
UPA is 0.0008–6.5 m.
9
NMR measurements
All NMR spectra were recorded on a JEOL Lambda 400 MHz spectrometer
(Tokyo, Japan). The samples were dissolved in D2O and measured at 37 °C. Chemical
shifts are referenced to the internal signal of 0.05% 3-(trimethylsilyl) propionic-2,2,3,3-
d4 acid, sodium salt (TSP) at 0.0 ppm. Resonance assignments for Hsp-G were made by
comparison with referenced data38-40
and by 2D 1H-
1H correlation spectroscopy (COSY)
NMR. The peaks of FP and NRG were assigned by comparison with the literature38-40
.
The chemical shifts of various protons in the Hsp-G units were monitored as a function
of concentration (1, 2, 3, 4, 5, 6, 7, 10, 20, 30, and 40 mg/mL). 1H-NMR spectra of
drug-loaded Hsp-G solutions (concentration equivalent to 20 mg/mL of Hsp-G) with
different loading ratios were also recorded.
The 2D 1H-
1H NOESY spectra for 1.0 and 20 mg/mL Hsp-G solutions were
measured to verify the solution structure of Hsp-G. The spectra were recorded with 512
data points in the t2 time domain, 256 t1 increments and 64 scans. The mixing time was
1.0 s. A relaxation delay of 6.3 s was used between the scans. A sine apodization
function was applied in both dimensions before Fourier transformation. The spectra
were zero-filled in the f1 dimension to give a 512 × 512 data matrix in the frequency
domain. Under similar conditions, the NOESY spectra of 20 mg/mL FP-loaded Hsp-G
and NRG-loaded Hsp-G solutions were also recorded to validate the interaction
positions between drug and Hsp-G molecules.
10
Figure 1. Chemical structures and atom numbering of (A) -glucosylhesperidin (Hsp-
G), (B) flurbiprofen (FP), and (C) naringenin (NRG).
11
EXPERIMENTAL
PART II Molecular-level characterization of probucol nanocrystal in water by in
situ solid-state NMR spectroscopy
Materials
Probucol was supplied by Daiichi-Sankyo (Japan). PVP (Kollidon
12 PF, Mw ≈
2500) was obtained from BASF (Japan). SDS was purchased from Wako Pure Chemical
Industries (Japan). All other chemicals used were of reagent grade.
Preparation of physical mixture (PM) and ground mixture (GM)
It is reported that the suitable weight ratio for drug/PVP/SDS is 1:3:1 as it shows
the highest recovery of drug nanoparticles 33-36
. The PM was prepared by physically
mixing probucol, PVP, and SDS at a weight ratio of 1:3:1 in a glass vial using a vortex
mixer. For the preparation of the ternary GM, the PM was ground in a TI-500ET
vibrational rod mill (CMT, Japan) for 30 min. The temperature during the grinding
process was controlled and maintained at 0 ± 5 °C using a nitrogen-gas flow cooling
system. The grinding cell and rod were made of stainless steel.
Preparation of freeze-dried mixture (FD)
In this research, to achieve enough the signal to noise (S/N) ratio and avoid
precipitation during long in situ solid-state NMR measurement, a stable and highly
concentrated nanosuspension was prepared for NMR evaluation and other
measurements. To prepare the suspension, the GM was dissolved in 20 mL of distilled
water with a probucol concentration of 5 mg/mL and the mixture was then sonicated for
2 min. The suspension was stored for 24 h at 25 °C and then freeze-dried under 15 Pa
and a trap temperature of -30 °C for 72 h using the EYELA freeze dryer FD-1000
(Tokyo Rikakikai, Japan).
Analysis of particle size
12
The PM, GM, and FD were dispersed in distilled water and then sonicated for 2
min to form the suspension. The drug concentration in the suspension was fixed at 5
mg/mL. The particle size was determined by the dynamic light scattering method using
Microtrac UPA
(Nikkiso, Japan; measurement range: 0.0008–6.5 m) and by the light
scattering method using Microtrac HRA
(Nikkiso, Japan; measurement range: 0.1–700
m).
Morphology observation by cryogenic-scanning electron microscopy (cryo-SEM)
The morphology of the suspension was investigated using a scanning electron
microscope (JSM-6510A) equipped with a cryo-SEM unit (JEOL, Japan). The GM with
5 mg/mL probucol was dispersed in water. The sample was loaded onto the cryo-
specimen holder, cryo-fixed in slush nitrogen, and then quickly transferred to the cryo-
SEM unit in a frozen state. The frozen sample was then fractured by striking them with
a pre-cooled razor blade, at the point on the sample surface where the fracture plane was
required. The specimen was then sputter-coated with gold, and a cryo-SEM image was
obtained at an acceleration voltage of 1.5 kV.
Powder X-ray diffraction (PXRD) measurement
PXRD measurements were performed using a Rigaku Miniflex diffractometer
(Rigaku, Japan) with a CuK radiation source at a voltage of 30 kV, a current of 15 mA,
and a scanning speed of 4° min-1
.
13C solid-state and in situ solid-state nuclear magnetic resonance (NMR) spectroscopy
13C NMR spectra were measured on a JNM-ECX400 NMR spectrometer (JEOL
Resonance, Japan). The solid-state and in situ solid-state NMR spectra were recorded
with a 6-mm HX probe. The spectra were obtained by cross-polarization/magic angle
spinning/total spinning sideband suppression (CP/MAS/TOSS) techniques. MAS was
carried out at a rotational speed of 4 kHz. The pertinent parameters included a 1H 90°
pulse of 6 µs and relaxation delay of 2 s. The CP contact time was 10 ms for in situ
13
solid-state NMR and 2 ms for solid-state NMR. For solid-state NMR spectroscopy,
powder samples weighing ca. 150 mg were placed in a 6-mm zirconia rotor. For the in
situ solid-state NMR measurement, a GM sample with 5 mg/mL probucol was dispersed
in water such that the signal to noise (S/N) ratio was sufficient. The measurements were
performed by filling the rotor with the suspension (160 μL). To achieve a suitable S/N
ratio, the numbers of accumulations were 20,000 for in situ solid-state NMR and 12,000
for solid-state NMR. All spectra were externally referenced by setting the methine peak
of hexamethylbenzene at 17.3 ppm.
13C liquid-state NMR spectroscopy
13C liquid-state NMR spectra were acquired on a JNM-ECX400 NMR
spectrometer (JEOL Resonance, Japan) by a conventional single pulse decoupling
method as follows: the samples with 5 mg/mL probucol were dissolved in D2O. The
measurements were performed at 37 °C. Chemical shifts were referenced to an internal
signal of 0.05% 3-(trimethylsilyl) propionic-2,2,3,3-d4 acid sodium salt (TSP) at 0.0
ppm. A relaxation delay of 2.0 s, scan of 8,000 accumulations, and rotational speed of
15 Hz were used.
14
Figure 2. Chemical structures and carbon atom numbering of (A) probucol, (B) PVP
(K12, Mw ≈ 2,500), and (C) SDS.
S S
OH
C(CH3)3
C(CH3)3
HO
(H3C)3C
C(CH3)3
CH3H3CA
B
C
Na+
SO O
O
O-
1"3"5"
6"
7"
2"4"8"
9"
10"
11"
12"
1
2
34
5
6
7 8
N
CH
O
CH2
1'2'
3'
4' 5'
6'
n
S S
OH
C(CH3)3
C(CH3)3
HO
(H3C)3C
C(CH3)3
CH3H3CA
B
C
Na+
SO O
O
O-
1"3"5"
6"
7"
2"4"8"
9"
10"
11"
12"
1
2
34
5
6
7 8
N
CH
O
CH2
1'2'
3'
4' 5'
6'
n
15
RESULTS AND DISCUSSION
PART I NMR characterization of self-aggregation behavior of -
glucosylhesperidin in water
Aggregation behavior of Hsp-G in aqueous solution
The 1H chemical shift is often sensitive to subtle changes in the local
environment of the molecule and is usually adopted to study molecular association39,42-
43.
1H-NMR was performed on gradual concentrations (1, 2, 3, 4, 5, 6, 7, 10, 20, 30, and
40 mg/mL) of Hsp-G as shown in Figure 3. The results revealed that the chemical shifts
(ppm) of all protons decreased with increased concentration of Hsp-G, suggesting that
some aggregation structure was formed as the Hsp-G concentration was increased. The
flavanone skeleton protons (H-2`, H-5`, H-6`, H-2, H-6 and H-8) were considerably
more shielded. Meanwhile the sugar protons (H-1```, H-1```` and H-6```) were less
shielded, except H-1`` which is close to the flavanone skeleton. The significant upshift
of flavanone skeleton protons could be caused by the formation of intermolecular
interactions, such as π-π interaction44
during the process of Hsp-G aggregation.
Consequently, the flavanone skeleton of the Hsp-G molecule should play an important
role in its aggregation. The chemical shifts of sugar protons were further influenced by
the effects of ring currents and thus exhibited small change in chemical shift during the
aggregation process.
Figure 4 shows concentration-dependent changes of 1H chemical shifts for some
key Hsp-G signals that were typically affected by the aggregation. NMR chemical shift
values corresponding to different nuclei in molecular units can be used to detect self-
aggregation behavior, e.g. micellization, has been previously documented42,45
. This
method could also apply to the investigation of large molecular weight polymer
micelles16
. Here, the value of critical micelle concentration (cmc) was used as the index
16
parameter of the self-assembly. Below the cmc, the average environment of various
protons in the molecule was surrounded by solvent molecules and the observed
chemical shift, δobs, was the chemical shift of the monomer, δmon. Above the cmc, δobs
was the weighted average of the monomer and micelle chemical shifts, based on the
assumption that the exchange of monomers and micelles in the solution is
comparatively fast on the NMR time scale. If the monomer concentration was constant
above the cmc, the observed chemical shift could be described as follows using the
mass-action law model (Eq. (1)) 42,45-46
:
δobs = δmic –(cmc/C)(δmic–δmon) (Eq. (1))
A plot of δobs as a function of 1/C should consist of two straight lines-one above
and the other below the cmc. The intersection of the two lines is the point corresponding
to 1/C = 1/cmc. Figures 4A and 4B represent plots of δobs (H) at H-2` and H-1`` as a
function of the reciprocal of the total Hsp-G concentration, respectively. The lines of the
experimental data fitted to the theoretical mode. The chemical shift variation broke
abruptly at a specific concentration. The cmc value determined from these chemical
shift dependences was about 5.0 mg/mL (6.5mM) at 37 ºC. This value was in
accordance with one previously obtained by surface tension measurements, although the
decrease in surface tension is very small15
. The critical aggregation of conventional
surfactants, e.g., dodecyldimethylbenzylammonium chloride, provides an abrupt
chemical shift variation corresponding to the critical micelle concentration42,45-46
. The
gradual chemical shift upon Hsp-G aggregation, on the contrary, would be different
mode of aggregation to conventional surfactants. The proposed Hsp-G aggregation may
not be categorized as a strict definition of “critical micelle concentration”. It has been
reported that a self-aggregation of bile salts surfactants provides a gradual change in
17
NMR chemical shift47
. This was explained by their specific structures, which are rather
different from those of the conventional surfactants47
. Similarly to bile acids, there
exists a range of concentration where Hsp-G aggregation took place. A negligible shift
was observed for the terminal methyl group in hydrophilic sugar groups of Hsp-G as
shown in Figure 4C. It confirmed that this proton does not change its chemical
environment through aggregate formation of Hsp-G molecules.
The Hsp-G aggregation number was further calculated by the chemical shift
variation according to aggregation equilibrium and mass action law (Eq. (2))48
, where n
and K are the aggregation number and aggregation equilibrium constant, respectively.
The plot of log (C(δobs-δmic) versus log( C (δmon-δobs) ) should give a straight line. The
slope yields the value of the aggregation number n.
log( C (δmon-δobs) )=n log (C(δobs-δmic) +log (nK)+(1-n)log(δmon-δobs) (Eq. (2))
At the Hsp-G concentration below 5 mg/mL, the aggregation number was
calculated as 2. It could show a dimer formation (Figure 5). At the Hsp-G concentration
above 5 mg/mL, the aggregation number was increased to 4-5. It suggested that Hsp-G
molecules self-associated into particular small micelles with low aggregation number.
It should be noted that the surface tensions at 37 °C decreased slightly from 72
mN/m to 65.5 mN/m of purified water when the concentration of Hsp-G was increased.
Hsp-G had surface activity although the surface activity was considerably weaker than
that of conventional micelles such as SDS (35 mN/m) or fluorinated random
copolymers (20~30 mN/m)49
. It is currently accepted that common surfactants or
amphiphiles are first adsorbed at the water surface to form a Gibbs monolayer and then
form micelles in the solution above cmc. However, Matsuoka et al., reported that a new
class of amphiphiles, which consisted of hydrophobic and ionic polymer chains, showed
18
“micelle formation without Gibbs monolayer formation”. Its origin was suggested to be
image charge repulsion at the air/water interface. Owing to its hydrophobicity, the block
copolymer would be adsorbed at the air/water interface. However, near the interface a
strong electrostatic repulsion from the interface was induced by the image charge effect.
As the hydrophilic chain consisted of polyions, this electrostatic repulsion was so strong
that the polymers could not be adsorbed but formed micelles in bulk solution. This was
explained as a totally new observation in the field of physical chemistry of surfaces and
interfaces50-54
. Since Hsp-G is non-ionic material, the surface activity could not be
explained by an image charge effect. It could be mentioned that the Hsp-G is
energetically more favorable in an aggregation, compared with in a surface monolayer.
The modest decrease in surface tension probably corresponds to a distribution of
“gaseous” like Hsp-G molecules at the surface. For an appreciable reduction in surface
tension, a higher surface concentration of Hsp-G at the surface is required. However, it
seems that the large polar head groups of Hsp-G geometrically preclude the possibility
of a significant association of the non-polar rings at the planar air/water interface. Thus,
a high surface concentration can not be achieved, since it would be unfavorable to have
sugar moieties at the surface. In contrast, a small aggregation would provide such an
environment to decrease the unfavorable hydrophobic interaction between water and the
rings. This assumption is supported by the results from dynamic light scattering (DLS)
method that the particle size in Hsp-G solution increases from ca. 1 nm to 3 nm as a
function of the concentration (Figure 6), although it is not easy to determine the
accurate size due to the principle of the measurement.
To further probe the conformation of Hsp-G molecules in either non-aggregated
or aggregated states, 2D 1H-
1H NOESY experiments were performed at Hsp-G
concentrations of 1 and 20 mg/mL, which are below and above the cmc value (= around
5.0 mg/mL (6.5 mM)) (Figure 7). The correlation peaks observed in the 2D 1H-
1H
19
NOESY spectra are the result of cross-relaxation between neighboring protons that are
spatially close to each other (roughly below 5 Å). The NOESY spectrum of a 1 mg/mL
Hsp-G solution shown in Figure 7A gave a few weak cross-peaks only among sugar
proton pairs, which could be due to intramolecular interaction. In addition, we did not
observe any NOE interactions between flavanone skeleton protons of the Hsp-G
molecule. These results showed that in a non-aggregated or monomolecular state, the
Hsp-G molecules are probably in an extended rather than folded conformation. By
contrast, when the concentration of Hsp-G was increased from 1 to 20 mg/mL, many
more additional correlation peaks were observed, implying much denser packing of the
molecules in the aggregates. The peaks were ascribed to the formation of the
hydrophobic core because most of new cross-peaks appeared in the hydrophobic region,
confirming the hydrophobic interaction of the flavanone skeleton. At the hydrophilic
portion, the intensity of all cross-peaks among sugar protons was also increased in
comparison with the non-aggregated state. Additionally, we did not observe any NOE
interaction between flavanone skeleton and the hydrophilic portion of the Hsp-G
molecule. These results offer insight into the spatial proximity between the respective
molecular fragments and give information about the preferred conformation of Hsp-G.
In many articles, the self-aggregations of amphiphilic materials are generally
described as micelles even when the aggregation number is lower than 1055-57
. For
instant, bile salts such as sodium cholate (NaC) and sodium taurocholate (NaTC) form
dimmer and trimmer aggregations, which are widely accepted as micelles or small
micelles47,56
. In this study, it is most probable that due to steric hindrance, amphiphilic
Hsp-G molecules self-associate into small micelles. We speculate the Hsp-G micelle
structure to be core-shell like where the flavanone skeleton is segregated from the
aqueous exterior to form an inner core encased by a shell of sugars. The conventional
core-forming segments, such as alkyl or aryl chains, could form the core through a
20
combination of intermolecular forces58
. Interestingly, flavanone skeleton moieties show
appreciable self-associating behavior, driven by hydrophobic interactions, which are
substantial for the core formation. The micropolarity of such a hydrophobic core
comprising flavanone skeletons may provide a better solubilizing microenvironment for
the desired encapsulation of a specific drug with a similar flavanone skeleton or
aromatic structure, since the loading capability was mainly dependent upon the
compatibility of the drug and the hydrophobic segment59
.
To further understand the Hsp-G micelle, the key cross sections from the 2D
NOESY spectra were taken and are shown in Figure 7C. Figures 7C (a, c, e) display
cross-sections along the f2 dimension from the NOESY spectrum of Figure 7A taken at
7.05 ppm, 5.11 ppm, and 1.09 ppm, which are the chemical shifts of H-2`, H-1`` and H-
6```, respectively. Correspondingly, Figures 7C (b, d, f) show cross-sections from
Figure 7B taken at 6.88 ppm, 4.88 ppm, and 1.08 ppm representing H-2`, H-1`` and H-
6``` resonance. These three proton environments represent an aromatic proton, anomeric
proton, and terminal sugar proton, respectively. At low Hsp-G concentration, no NOE
interactions can be detected for these protons. At the higher Hsp-G concentration of 20
mg/mL, numerous intra- and intermolecular NOE interactions can be seen in this trace.
For H-2` (Figure 7C (b)), several intra- and intermolecular NOESY cross-peaks were
detected involving H2`-H8,6, H2`-H2, H2`-H4`(OCH3), H2`-H3 pairs. These groups
participate in the formation of the hydrophobic core and are thus in close proximity with
each other. In the hydrophobic portion of the Hsp-G molecules, the distance between H-
2` and H-8,6 is too large for intramolecular interactions to be significant. The observed
cross-peak could be explained as the result of intermolecular interactions between
different Hsp-G molecules. Emin et al. , and Yuan et al.
reported the surfactant
structures of Hyamine-M and Triton X-100, respectively, by NMR spectroscopy. The
packing of the hydrophobic chains was not strictly oriented during aggregation, this
21
being confirmed by NOE contacts16,60
. Hence, we speculated that the packing of the
hydrophobic part in Hsp-G micelles is similarly not strictly oriented, i.e. H-2` group of
one molecule might be packed near the H-8,6 groups of another molecule. The
additional cross-peak between H-2` and H-4`(OCH3) can similarly be attributed to this
disorder. For the anomeric signal H-1``(Figure 7C (d)), sugar proton accumulation
could also be observed. For terminal sugar signal H-6``` (Figure 7C (f)), no NOE
correlation peak can be detected in the observed range. In the micellization of alkyl or
alyl surfactants with low molecular weight, aggregation of the hydrophilic portion can
occur in accordance with the concentration changes42
. The least dense cross-peaks
indicate less intermolecular interaction at the surface of Hsp-G micelles even at high
concentration. Thus, we assumed that in Hsp-G micelles, the sugar protons are loosely
packed. Although the correlation of surface activity and micelle structure has not been
extensively discussed in the literature, the loose packing of the sugar shell appears to be
related to little surface activity of Hsp-G.
Solubilization mechanism of poorly water-soluble drugs
Figure 8 shows the 1H-NMR spectra of Hsp-G in D2O at 37 °C before (8A) and
after (8B, or 8C) being spray-dried with FP or NRG, together with assignment of the
most prominent resonance signals. The peaks of FP and NRG were clearly detected. In
general, solution NMR cannot be used to analyze molecules with low mobility e.g.
crystals due to the broad signals. The solution NMR properties suggest that these drugs
in Hsp-G solution are in molecular-dispersed state rather than crystal state. The
solubilities of FP and NRG were calculated at 37 °C from the integration values of 1H
peaks using 0.05% TSP as a reference. The concentration of NRG was calculated as
1.42 mg/mL (untreated NRG: 0.05 mg/mL), indicating almost 30 times greater
solubility when compared to NRG crystals. This result was in accordance with our
previously published data, which indicated the enhancement effect of Hsp-G on the
22
solubility of NRG by HPLC13
. The solubility of FP prepared as spray-dried powders of
FP/Hsp-G also showed a remarkable increase in water, with the concentration of FP
calculated as 0.49 mg/mL (untreated FP: 0.04 mg/mL). The enhancement of the drug
solubility is due to the incorporation of the drug into the Hsp-G micelle via
intermolecular interaction, as Hsp-G micelles have an inner core of assembled
flavanone skeleton segments capable of solubilizing poorly water-soluble substances.
They also have an outer hydrophilic shell serving as a stabilizing interface between the
hydrophobic core and the external aqueous environment. Since both FP and NRG
molecules possess a hydrophobic character, they are incorporated in the Hsp-G
hydrophobic core. The intermolecular interactions in the spray-drying process facilitate
the drug-solubilized Hsp-G micelle formation in water. The solubility enhancement of
NRG was significantly higher than that of FP. Many studies indicate that the most
important factor related to drug solubilization capacity of a micelle is the compatibility
between the drug and the core-forming segment59
. For this reason, the choice of core-
forming segment is very important; the more similar the polarity of the hydrophobic
micelle core to the solubility characteristics of the drug, the more effective the drug
incorporation should be61
. It can be rationalized that NRG could more easily gain
ingress to the hydrophobic part of micelles compared with FP, primarily due to
similarity in the polarity of NRG with Hsp-G micelle cores.
We applied this same approach using the chemical shifts of some key signals
from drug/Hsp-G spray-dried powders with different drug-loaded ratios. Figures 9A-B
show chemical shift variation of the 1H chemical shifts in the Hsp-G unit on the
different drug-loaded ratios (1:10, 1:20, and 1:30) of the same Hsp-G concentration.
Compared with untreated Hsp-G, drug-loaded Hsp-G exhibited upshift variation in the
hydrophobic part (H-4`, H-5`6`, H8 and H-3), whereas no significant change was
observed in the hydrophilic part (H-6```). This upshift is probably due to interaction of
23
the flavanone skeleton of Hsp-G with the FP or NRG molecules. Additionally, the
chemical shift variation of the Hsp-G signals showed a good relationship with the
amount of drug loaded. It is clear that drug inclusion influenced the structure of Hsp-G
micelle. Furthermore, the overall effect of the drug on the chemical shift of Hsp-G was
in agreement with the above-mentioned solubilization enhancement.
2D NOESY spectroscopy was used to confirm the localization of FP and NRG.
Figure 10 show NOESY spectra of FP-loaded Hsp-G and NRG-loaded Hsp-G. The key
cross-peaks to deduce the possible interaction sites between the Hsp-G molecules and
FP or NRG molecules are highlighted. A significant number of NOE interactions
connecting drugs and Hsp-G were found. These indicated that the FP and NRG were
bound to the Hsp-G micelles. These new cross-peaks presented were not observed in the
NOESY spectrum of the pure micelle without the drugs (Figure 7). Figure 10A shows
that the aromatic protons of FP exhibited NOE contacts with the flavanone skeleton as
well as with methoxy group (H-4`OCH3) of Hsp-G. NOE interactions between the
aromatic proton (H-c, H-f, H-b`, and H-f`,) of FP and Hsp-G (H-2`,5`,6`) were observed.
Additionally, the methyl group of FP (H-h), and the methoxy group of Hsp-G (H-
4`OCH3) exhibited cross-peak, whereas no NOE interactions connecting FP to the
hydrophilic moiety of Hsp-G could be found. It was confirmed that the intermolecular
interactions between FP and Hsp-G were between the flavanone skeleton and the FP
molecules. The same NOESY experiment was carried out with NRG-loaded Hsp-G as
shown in Figure 10B. Since there is a high degree of similarity between the flavanone
skeleton of NRG and the Hsp-G structure, a number of signals showed significant
overlap, making the assignment of the NOE cross-peaks uncertain. However, a well-
defined NOE cross-peak indicates the proximity of two flavanone skeletons between
Hsp-G (H-2`) and NRG (H-B`, H-F`). The observed NOE interaction can be interpreted
as a result of the insertion of NRG into the core interior, and the flavanone skeletons of
24
Hsp-G and NRG were relatively close to each other. All NMR observations
undoubtedly confirmed our assumption that the solubilized FP and NRG mainly
accumulated and was stabilized in the Hsp-G hydrophobic core by hydrophobic
interaction.
The micellization of Hsp-G indicated by NMR measurements is somewhat
similar to that of bile salts surfactants. The bile salts surfactants, with bulky molecules
and large molecular weight, have a gradual rather than abrupt cmc, and aggregate with
low numbers around 2-5, which are different from other alkyl surfactant47, 55-57
. The
particle size of ca. 1-3 nm from DLS measurements supported that Hsp-G self-
associates into small micelles, in which the aggregate number of Hsp-G molecules was
around 2-5. A simplified schematic representation of the Hsp-G micelles is given in
Scheme 1. Interestingly, Hsp-G molecules which show less surface activity are shown
to undergo self-assembly to form small micelles with a core-shell like architecture in
aqueous solution. Flavanone skeletons are segregated from the aqueous exterior to form
a drug-loading core through hydrophobic driving forces. The core is stabilized by shell-
forming segments and the packing of these hydrophobic parts is not strictly oriented
with the intermolecular arrangement of micelle surface being loose. Furthermore, poorly
soluble drugs could be effectively incorporated into the Hsp-G hydrophobic core
through hydrophobic interactions.
25
Fig
ure
3.P
arti
al 1
H N
MR
spec
tra
of
Hsp
-G p
roto
n s
ignal
s at
dif
fere
nt
con
centr
atio
ns
ran
gin
g f
rom
1 t
o 4
0
mg
/mL
reco
rded
in
D2O
at
37 º
C. T
he
con
cen
trat
ion
s ar
e giv
en o
n t
he
rig
ht
han
d s
ide
of
the
spec
tra.
26
Figure 4. Variation of chemical shifts versus reciprocal Hsp-G concentrations ranging
from 1 to 40 mg/mL: (A): H-2`, (B): H-1``, and (C): H-6```. The chemical shift data
used in this graph were indicated in Figure 3 at the concentration of 1 and 40 mg/mL.
27
Figure 5. The plot of log (C (δobs-δmic)) versus log (C (δmon-δobs)) of the H-2` signal of
Hsp-G in D2O at 37 °C.
-3.8 -3.2 -2.6 -2.0-5
-4
-3
-2
-1
cmc5 mg/mL
Lo
g(
C (δ
mon-δ
ob
s) )
Log( C(δobs-δmic))
28
Figure 6. Particle size distribution of Hsp-G at different concentrations ranging from 1
to 50 mg/mL measured by DLS.
Particle size (m)
0
50
0.001 0.01 0.1 1 10
0
50
0
50
0
50
0
50
0
50
0
100
0
100
0
100
0
100
0
100
0
100
Con.=1 mg/mL
mv = 1.3 nm
Frequency (%)
Cumulative(%)
Con.=2 mg/mL
mv = 2.0 nm
Con.=5 mg/mL
mv = 1.7 nm
Con.=10 mg/mL
mv = 2.1 nm
Con.=20 mg/mL
mv = 2.5 nm
Con.=50 mg/mL
mv = 3.0 nm
Fre
quen
cy (
%)
Cum
ulativ
e (%)
Particle size (m)
0
50
0.001 0.01 0.1 1 10
0
50
0
50
0
50
0
50
0
50
0
100
0
100
0
100
0
100
0
100
0
100
Con.=1 mg/mL
mv = 1.3 nm
Frequency (%)
Cumulative(%)
Con.=2 mg/mL
mv = 2.0 nm
Con.=5 mg/mL
mv = 1.7 nm
Con.=10 mg/mL
mv = 2.1 nm
Con.=20 mg/mL
mv = 2.5 nm
Con.=50 mg/mL
mv = 3.0 nm
Particle size (m)
0
50
0.001 0.01 0.1 1 10
0
50
0
50
0
50
0
50
0
50
0
100
0
100
0
100
0
100
0
100
0
100
Con.=1 mg/mL
mv = 1.3 nm
Frequency (%)
Cumulative(%)
Con.=2 mg/mL
mv = 2.0 nm
Con.=5 mg/mL
mv = 1.7 nm
Con.=10 mg/mL
mv = 2.1 nm
Con.=20 mg/mL
mv = 2.5 nm
Con.=50 mg/mL
mv = 3.0 nm
Fre
quen
cy (
%)
Cum
ulativ
e (%)
29
Figure 7. 2D 1H/
1H NOESY spectra and cross sections of Hsp-G (A) NOESY spectrum
of 1 mg/mL (below cmc) and (B) NOESY spectrum of 20 mg/mL (above cmc). (C)
cross sections taken at proton resonance of (a): H-2` (1 mg/mL, 7.05ppm), (b): H-2` (20
mg/mL, 6.88ppm), (c): H-1`` (1 mg/mL, 5.11ppm), (d): H-1`` (20 mg/mL, 4.88ppm),
(e): H-6``` (1 mg/mL, 1.09ppm), and (f): H-6``` (20 mg/mL, 1.08ppm).
A B
C
30
Fig
ure 8
.1H
NM
R s
pec
tra
of
(A)
Hsp
-G, 2
0 m
g/m
L;
(B)
FP
-load
ed H
sp-G
, 22
mg
/mL
; an
d (
C)
NR
G-l
oad
ed H
sp-G
, 2
2 m
g/m
L. S
tars
ind
icat
ed p
eak-o
ver
lap
bet
wee
n H
sp-G
an
d d
rug
sig
nal
s.
31
Figure 9. Chemical shift variation at the different chemical moieties of Hsp-G after
spray-drying with different weight ratios (drug:Hsp-G) of (A): FP and (B): NRG.
32
Fig
ure
10.2D
1H
/1H
NO
ES
Y s
pec
tra
of
(A)
FP
-load
ed H
sp-G
and (
B)
NR
G-l
oad
ed H
sp-G
. N
ew c
ross
pea
ks
are
hig
hli
ghte
d.
4`(
OC
H3)
33
SCHEME 1. Schematic representation of the micelle structure of Hsp-G based on
NMR constraints.
34
RESULTS AND DISCUSSION
PART II Molecular-level characterization of probucol nanocrystal in water by in
situ solid-state NMR spectroscopy
Size measurement
Figure 11 shows the mean particle size of the ternary probucol/PVP/SDS system.
The mean particle size of probucol in the PM, measured by the laser diffraction method,
was 53.7 μm (Figure 11A). The mean particle size of the GM was approximately 40 nm
after it was dispersed into water (primary nanoparticles) (Figure 11B) and 150 nm after
it was stored at 25 °C for 24 h (Figure 11C). These particle sizes were different from
previously reported data33,35
. This difference was attributed to the difference in the
concentration of probucol; the concentration of probucol in this study, i.e., 5 mg/mL,
was 10 times higher than that in previous studies. Figure 11C suggests that secondary
nanoparticles were formed when the suspension was stored for 24 h. The secondary
nanoparticle suspension showed superior stability; its particle size was maintained at
approximately 150 nm even after one month of storage (data not shown). For in situ
measurement, this concentrated suspension was spun at 4 kHz in an NMR sample tube
for 96 h. The mean particle size of the suspension was 154 nm after spinning (Figure
11D). It should be noted that centrifugation for a long duration also had no effect on the
particle size of the suspension. From these results, we inferred that this concentrated
nanosuspension from ternary GM could be successfully evaluated by in situ solid-state
NMR spectroscopy. The FD suspension showed a mean particle size of ca. 178 nm
(Figure 11E), indicating that the size of the particles only slightly increased after the
freeze-drying process. Hence, the probucol in the FD powder was expected to exhibit
the same physiochemical properties as that in the GM suspension.
Morphology observation by cryo-SEM measurement
35
Cryo-SEM images of the GM suspension after storage at 25 °C for 24 h are
shown in Figure 12. Cryo-SEM is used for direct viewing of hydrated (wet) samples,
thus providing information about the solid or semisolid components in the suspension.
The images show that the nanoparticles in water were spherical with a size of
approximately 150–200 nm, which was in agreement with the DLS results. The particles
appeared to be an agglomeration of primary nanoparticles.
PXRD characterization
Figure 13 shows PXRD patterns of the probucol/PVP/SDS system. The PXRD
pattern of the PM showed the superimposition of the diffraction peaks of probucol and
SDS crystals (Figure 13A). For the GM, the diffraction peaks disappeared and a halo
pattern was observed (Figure 13B). This result suggests that crystalline probucol may
have been transformed into amorphous or very small crystals with some disorder by co-
grinding. Diffraction peaks of both crystalline probucol and SDS were observed in the
PXRD pattern of the FD samples, as shown in Figure 13C. It is supposed that the
crystallization of probucol occurred after the GM was dispersed in water, and SDS was
crystallized from its dissolved state during the freeze-drying process.
NMR characterization
Figures 14~16 show the 13
C solid-state, in situ solid-state, and liquid-state NMR
spectra of the probucol/PVP/SDS system. The solid-state spectrum of the ternary PM
(Figure 14A) showed superimposition of the spectrum of each component in the PM
(data not shown), indicating little interaction among probucol, PVP, and SDS. The
solid-state spectrum of the GM (Figure 14B) showed broadening of the peaks of
probucol, and a new peak was observed at around 143 ppm, shown by a diamond. The
peak of SDS (C-1") at around 66.7 ppm was upshifted. It has been reported that the new
peak and upshift variation occur because of probucol-PVP and PVP-SDS interactions
during co-grinding 33
. The solid-state NMR spectrum of the FD sample (Figure 14C)
36
was quite similar to that of the PM in Figure 14A, thus confirming the results of the
PXRD measurements that probucol and SDS existed in a crystalline state. However, a
little broadening of the probucol peak (i.e., C-3, 5) for the FD sample was observed
compared with the peak shape for the PM sample. This difference could be attributed to
the difference between the particle size of probucol crystals in the FD sample
(nanocrystals) and that in the PM sample (several m size crystals). Figure 15 shows an
in situ solid-state NMR spectrum of the GM dispersed in water and stored at 25 °C for
24 h. In situ solid-state NMR techniques have expanded the application scope of NMR
spectroscopy from liquid- or solid-state materials to suspended-state materials. During
CP, the transfer of nuclear magnetization from 1H to
13C occurs mainly via the
1H-
13C
through space dipolar interactions, which are present in both solids and liquids. In
isotropic liquids, these dipolar interactions are averaged to zero, and therefore, the
liquid component cannot be detected 29
. In this study, the 13
C peaks of samples in a
relatively rigid solid state were prominent in the spectrum. The probucol signals could
be detected in the spectrum of the GM suspension, possibly indicating the presence of a
probucol nanocrystal core in water. These results coincided with the solid-state NMR
results of the FD sample. Probucol has two polymorphs, named form I and form II 62
.
The chemical shift of the characteristic peak at 122 ppm coincided with probucol form I,
while the characteristic peak of probucol form II at 117 ppm was not shown 33
. Further,
no peaks were recorded from the PVP and SDS molecules. These results indicated that
most of the PVP and SDS molecules were either freely mobile or soluble in water. The
broad amorphous peaks of probucol and a new peak at 143 ppm in the solid-state GM
spectrum (Figure 14B) disappeared in the NMR spectrum of the GM suspension. It was
speculated that the spectral differences for probucol were induced by the transformation
of amorphous/or very small crystals with some disorder into a nanocrystalline state after
dispersion in water. Probucol nanocrystals in water could be stabilized by having PVP
37
and SDS on the surface, although the strength of interaction on the surface was too low
to be detected.
The suspensions were further investigated by 13
C liquid-state NMR spectroscopy
to study the interactions at the surface of the nanoparticles (Figure 16). No peaks were
observed for the drug in both the PM and GM spectra despite the high drug
concentration. The results suggested almost all probucol existed in a crystalline state
with low mobility, which could not be detected in the liquid-state NMR spectra. The
peaks of PVP and SDS were clearly observed in the spectra. In addition, the spectrum of
the GM (Figure 16B) was slightly different from that of the PM (Figure 16A). The
broader peaks of PVP (C-2', C-3') and SDS (i.e. C-1'', C-4"~9") were observed for the
GM suspension. In the PM suspension, a PVP/SDS complex was expected to exist
where some SDS micelles were bound to the PVP chain. In the GM suspension, solid-
liquid interaction between probucol crystals and the PVP/SDS complex were expected
at the surface. The PVP/SDS complex was found to alternate between its bound and
unbound forms. This exchange in the solution was fast, as inferred from the NMR time
scale, and thus, the NMR signals of PVP and SDS were broadened 63
. Such interaction
on the probucol crystal surface stabilizes the suspension 37
.
Thus, this study confirmed the proposed nanoparticle structure in the ternary
probucol/PVP/SDS GM suspension 37
and that the probucol nanocrystals were covered
by PVP/SDS complexes. Moreover, the direct observation of the nanosuspension at a
molecular level not only elucidated the crystal structure of probucol but also indicated
that the PVP-SDS complex achieved equilibrium by alternating between its bound and
unbound forms on the surface of the probucol nanocrystal. For in situ solid-state NMR
spectroscopy, the suspension is required to maintain its stability at high concentration or
during the long MAS process. The knowledge acquired by in situ solid state NMR
spectroscopy was useful for understanding the nanoparticle structure in water.
38
Furthermore, a combination and comparison of the results obtained by solid-state and
liquid-state NMR spectroscopies can provide insight about the molecular states of drug
and additives in a suspension.
39
Figure 11. Particle size distribution of probucol/PVP/SDS systems in water: (A) PM
suspension (measured by laser diffraction method), (B) GM suspension after dispersion,
(C) sample B storage at 25 °C for 24 h, (D) sample C after spinning at 4 kHz for 96 h,
and (E) FD suspension (B~E measured by dynamic light scattering method.).
Fre
qu
ency (%
) Cum
ulativ
e (%)
A
0.1 1 10 100 10000
2
4
6
8
10
0
20
40
60
80
100
0
6
12
18
24
30
0
20
40
60
80
100
D
C
0
5
10
15
20
25
0
20
40
60
80
100
MV= 53.7 μm
MV= 150 nm
MV= 154 nm
0.0001 0.001 0.01 1 10
Particle size (μ m)
0.1
E
0
5
10
15
20
25
0
20
40
60
80
100
MV= 178 nm
B
0
5
10
15
20
25
0
20
40
60
80
100
MV= 40 nm
Fre
qu
ency (%
) Cum
ulativ
e (%)
AA
0.1 1 10 100 10000
2
4
6
8
10
0
20
40
60
80
100
0
6
12
18
24
30
0
20
40
60
80
100
DD
CC
0
5
10
15
20
25
0
20
40
60
80
100
MV= 53.7 μm
MV= 150 nm
MV= 154 nm
0.0001 0.001 0.01 1 10
Particle size (μ m)
0.1
E
0
5
10
15
20
25
0
20
40
60
80
100
MV= 178 nm
B
0
5
10
15
20
25
0
20
40
60
80
100
MV= 40 nm
40
Figure 12. Cryo-SEM images of probucol/PVP/SDS GM suspension after 24 h storage
at 25 °C. A and B indicate a varying scale bar.
B 1 m
A 1 m
B 1 m
A 1 mA 1 m
41
Figure 13. PXRD patterns of probucol/PVP/SDS system: (A) PM, (B) GM, and (C) FD.
Filled squares and unfilled circles represent the characteristic peaks of probucol and
SDS, respectively.
2 10 20 30 35
2 ( )
3000
0
10000
01400
0
A
B
C
Inte
nsi
ty
(cp
s)
2 10 20 30 35
2 ( )
3000
0
10000
01400
0
A
B
C
Inte
nsi
ty
(cp
s)
42
80
60
40
02
0
pp
m
6'
6
5,3
4
7,2
,8
1',
5'
2"~
10
"
19
01
70
15
01
10
13
0
pp
m
6'
65
,34
6'
6
5,3
4
A B C
1"
12
', 3
'1
2"
4'
11
"
1"
12
', 3
'1
2"
4'
11
"
1"
12
', 3
'1
2"
4'
11
"
80
60
40
02
0
pp
m
80
60
40
02
0
pp
m
6'
6
5,3
4
7,2
,8
1',
5'
2"~
10
"
19
01
70
15
01
10
13
0
pp
m
19
01
70
15
01
10
13
0
pp
m
6'
65
,34
6'
6
5,3
4
A B C
1"
12
', 3
'1
2"
4'
11
"
1"
12
', 3
'1
2"
4'
11
"
1"
12
', 3
'1
2"
4'
11
"
Fig
ure
14
. 13C
soli
d-s
tate
CP
/MA
S/T
OS
S N
MR
spec
tra
of
pro
buco
l/P
VP
/SD
S s
yst
ems
bet
wee
n 0
-80
ppm
(rig
ht)
and 1
10-1
90 p
pm
(lef
t):
(A)
PM
, (B
) G
M,
and (
C)
FD
. T
he
dia
mond d
enote
s th
e new
pea
k.
43
20
00
50
10
01
50
pp
m
6
5,3
4
72,8
1
20
00
50
10
01
50
pp
m
20
00
50
10
01
50
pp
m
6
5,3
4
72,8
1
Fig
ure
15.
13C
in
sit
uso
lid
-sta
te C
P/M
AS
/TO
SS
NM
R s
pec
tru
m o
f p
rob
uco
l/P
VP
/SD
S t
ern
ary
GM
su
spen
sio
n a
fter
sto
rag
e at
2
5ºC
fo
r 2
4 h
.
44
80
60
40
02
0
pp
m
19
01
70
15
01
10
13
0
pp
m
A B
6'
6'
1"
2',
3'
12
"
4'
3"
1'
4"-
9"
11
"
1"
2',
3'
12
"
4'
3"
1'4"-
9"
11
"
80
60
40
02
0
pp
m
80
60
40
02
0
pp
m
19
01
70
15
01
10
13
0
pp
m
19
01
70
15
01
10
13
0
pp
m
A B
6'
6'
1"
2',
3'
12
"
4'
3"
1'
4"-
9"
11
"
1"
2',
3'
12
"
4'
3"
1'4"-
9"
11
"
Fig
ure
16
.13C
liq
uid
-sta
te s
ing
le p
uls
e w
ith
dec
ou
pli
ng
NM
R s
pec
tra
of
pro
bu
col/
PV
P/S
DS
sy
stem
s b
etw
een
0-8
0
ppm
(rig
ht)
an
d 1
10-1
90 p
pm
(lef
t):
(A)
PM
su
spen
sio
n a
nd (
B)
GM
su
spen
sio
n a
fter
sto
rag
e at
25
ºC
fo
r 2
4 h
.
45
CONCLUSIONS
Part I
Concentration-dependent variation of proton chemical shift in amphilic Hsp-G
revealed the self-association of these molecules in solution. The gradual chemical shift
variation indicated the specific Hsp-G aggregation, which was different from that of
conventional alkyl surfactants. The cmc for Hsp-G was determined to be around 5.0
mg/mL (6.5 mM). The aggregation number was calculated around 2-5 according to
aggregation equilibrium and mass action law. Further 2D NOESY and DLS
measurements confirmed that Hsp-G molecules self-associated into particular small
micelles with a core-shell like architecture, in which the flavanone skeleton is
segregated from the aqueous exterior to form a novel drug-loading core surrounded by a
hydrophilic shell of sugar groups. As Hsp-G has little surface activity, Hsp-G
micellization should be a unique phenomenon. Moreover, we found that two model
drugs, FP and NRG, were effectively trapped into the core, as the structure dependent
manner. The chemical shift variation of the Hsp-G signals showed a good relationship
with the quantity of drug loaded, indicating that drug loading affects the Hsp-G micelle.
Hsp-G, a food additive, could prove attractive as non-toxic and less expensive excipient
for drug delivery in pharmaceutical industry.
Part II
Nanoparticles were obtained by dispersing the probucol/PVP/SDS ternary co-
ground mixture in water. The nanoparticles with a mean particle size of approximately
150 nm were directly evaluated by 13
C in situ solid-state and liquid-state NMR
spectroscopies. The 13
C in situ solid-state NMR spectra showed that probucol existed as
a nanocrystal (form I) in water. The 13
C solid-state NMR and powder X-ray diffraction
46
characterization of the FD sample confirmed the presence of the probucol nanocrystal.
PVP and SDS observed by liquid-state NMR spectroscopy suggested that they
interacted on the surface of the probucol nanocrystals to provide stability to the
nanoparticles. The advantage of in situ solid-state and liquid-state NMR spectroscopies
is that the drug and additives in a nanosuspension can be observed at a molecular level.
We will perform further in situ solid-state NMR measurements for monitoring the
dynamic interactions that occur at the surface of nanoparticles.
47
ACKNOWLEDGEMENTS
Words can never describe my sense of gratitude for Prof. Keiji Yamamoto.
Throughout my thesis-writing period, he provided warm encouragement, thoughtful
guidance, good company, and lots of splendid ideas. I am also quite impressed at his
enthusiasm in the science.
I thank Asso. Prof. Kunikazu Moribe and Assist. Prof. Kenjirou Higashi for their
kindness, helpful discussions, as well as their academic experience, which have been
invaluable to me. I would have been lost without them.
I am grateful to Assist. Prof. Waree Limwikrant for her support and hospitality.
I am also grateful to the members of the Yamamoto labs. They have been excellent
sources of scientific advice and support, and I am lucky to have met several great
friends there as well.
I wish to thank Prof. Hirofumi Takeuchi, Asso. Prof. Yuichi Tozuka, and Mr.
Hiromasa Uchiyama at Gifu Pharmaceutical University, Assist. Prof. Takeshi Morita at
Graduate School of Science and Technology of Chiba University, for their invaluable
help and encouragement.
I would also like to thank the Monbusho (Monbukagakusho) SCHOLARSHIP
PROGRAMME for financial support during my graduate studies.
I also owe deep thanks to my teachers and friends from China Pharmaceutical
University in Nanjing who supported me during my studies in Japan.
I feel blessed to have had the love and support of my families throughout this
arduous journey. I am very grateful to my husband, Chunyong Wu, for all that he has
done for me. Finally, this work is dedicated especially to my son, Wenbo Wu.
48
REFRENCES
1. Higashi K, Ideura S, Waraya H, Moribe K, Yamamoto K: Incorporation of
salicylic acid molecules into the intermolecular spaces of γ-cyclodextrin-
polypseudorotaxane. Cryst Growth Des 9:4243–4246 (2009)
2. Dahlberg C, Millqvist-Fureby A, Schuleit M, Furo I: Relationships between
solid dispersion preparation process, particle size and drug release - an NMR and NMR
microimaging study. Eur J Pharm Biopharm 76:311–319 (2010)
3. Sinha S, Ali M, Baboota S, Ahuja A, Kumar A, Ali J: Solid dispersion as an
approach for bioavailability enhancement of poorly water-soluble drug ritonavir. AAPS
PharmSciTech 11:518–527 (2010)
4. Letchford K, Burt H: A review of the formation and classification of
amphiphilic block copolymer nanoparticulate structures: micelles, nanospheres,
nanocapsules and polymersomes. Eur J Pharm Biopharm 65: 259–269 (2007)
5. Mikhail AS, Allen C: Poly(ethylene glycol)-b-poly(ε-caprolactone) micelles
containing chemically conjugated and physically entrapped docetaxel: synthesis,
characterization, and the influence of the drug on micelle morphology.
Biomacromolecules 11:1273–1280 (2010)
6. Wiedmann TS, Kamel L: Examination of the solubilization of drugs by bile salt
micelles. J Pharm Sci 91:1743–1764 (2002)
7. Yokoyama M: Polymeric micelles as a new drug carrier system and their
required considerations for clinical trials. Expert Opin Drug Deliv 7:145–158 (2010)
8. Dickson FM, Lawrence JN, Benford DJ: Surfactant-induced cytotoxicity in
cultures of human keratinocytes and a commercially available cell line (3T3). Toxicol
In Vitro 7:381–384 (1993)
49
9. Kwon GS, Okanob T: Polymeric micelles as new drug carriers. Adv Drug
Deliver Rev 21:107–116 (1996)
10. Torchilin VP: Micellar nanocarriers: pharmaceutical perspectives. Pharm Res
24:1–16 (2007)
11. Blanco E, Bey EA, Dong Y, Weinberg BD, Sutton DM, Boothman DA, Gao J:β-
lapachone-containing PEG-PLA polymer micelles as novel nanotherapeutics against
NQO1-overexpressing tumor cells. J Control Release 122:365–374 (2007)
12. Hiromi H, Toshio M: α-glycosyl hesperidin and its uses. United States Patent
5627157 (1997)
13. Tozuka Y, Kishi J, Takeuchi H: Anomalous dissolution property enhancement
of naringenin from spray-dried particles with α-glucosylhesperidin. Adv Powder
Technol 21:305–309 (2010)
14. Uchiyama H, Tozuka Y, Imono M, Takeuchi H: Improvement of dissolution and
absorption properties of poorly water-soluble drug by preparing spray-dried powders
with alpha-glucosyl hesperidin. Int J Pharm 392:101–106 (2010)
15. Uchiyama H, Tozuka Y, Imono M, Takeuchi H: Transglycosylated stevia and
hesperidin as pharmaceutical excipients: Dramatic improvement in drug dissolution and
bioavailability. Eur J Pharm Biopharm 76:238–244 (2010)
16. Emin SM, Denkova PS, Papazova KI, Dushkin CD, Adachi E: Study of reverse
micelles of di-isobutyl-phenoxy-ethoxy-ethyl-dimethyl-benzyl ammonium methacrylate
in benzene by nuclear magnetic resonance spectroscopy. J Colloid Interface Sci
305:133–141 (2007)
17. Denkova PS, Van Lokeren L, Willem R: Mixed micelles of Triton X-100,
sodium dodecyl dioxyethylene sulfate, and synperonic l61 investigated by NOESY and
diffusion ordered NMR spectroscopy. J Phys Chem B 113:6703–6709 (2009)
50
18. Schedlbauer A, Coudevylle N, Auer R, Kloiber K, Tollinger M, Konrat R:
Autocorrelation analysis of NOESY data provides residue compactness for folded and
unfolded proteins. J Am Chem Soc 131:6038–6039 (2009)
19. Yang W, Li Y, Cheng Y, Wu Q, Wen L, Xu T: Evaluation of phenylbutazone
and poly(amidoamine) dendrimers interactions by a combination of solubility, 2D-
NOESY NMR, and isothermal titration calorimetry studies. J Pharm Sci 98:1075–1085
(2009)
20. Zhang Z, Bu H, Gao Z, Huang Y, Gao F, Li Y: The characteristics and
mechanism of simvastatin loaded lipid nanoparticles to increase oral bioavailability in
rats. Int J Pharm 394:147–153 (2010)
21. Ruenraroengsak P, Cook JM, Florence AT: Nanosystem drug targeting: facing
up to complex realities. J Control Release 141:265–276 (2010)
22. Shudo J, Pongpeerapat A, Wanawongthai C, Moribe K, Yamamoto K: In vivo
assessment of oral administration of probucol nanoparticles in rats. Biol Pharm Bull
31:321–325 (2008)
23. Sharma P, Zujovic ZD, Bowmaker GA, Denny WA, Garg S: Evaluation of a
crystalline nanosuspension: polymorphism, process induced transformation and in vivo
studies. Int J Pharm 408:138–151 (2011)
24. Mayer C: NMR characterization techniques- application to nanoscaled
pharmaceutical carriers. Chem Inform 38 (2007)
25. Mayer C: NMR studies of nanoparticles. Annu Rep NMR Spectrosc 55:205–
258 (2005)
26. Mayer C: Nuclear magnetic resonance on dispersed nanoparticles. Prog Nucl
Mag Res Sp 40:307–366 (2002)
51
27. Abdelwahed W, Degobert G, Stainmesse S, Fessi H: Freeze-drying of
nanoparticles: formulation, process and storage considerations. Adv Drug Deliver Rev
58:1688–1713 (2006)
28. Xu M, Harris KDM, Thomas JM: In situ solid-state 1H NMR studies of
hydration of the solid acid catalyst ZSM-5 in its ammonium form. Solid State Nucl Mag
35:93–99 (2009)
29. Crisp JL, Dann SE, Edgar M, Blatchford CG: The in-situ solid-state NMR
spectroscopy investigation of alcoholic lactose suspensions. Solid State Nucl Mag
37:75–81 (2010)
30. Friebolin V, Marten S, Albert K: Characterization of binding affinities in a
chromatographic system by suspended state HR/MAS NMR spectroscopy. Magn Reson
Chem 48:111–116 (2010)
31. Skogsberg U, Meyer C, Rehbein J, Fischer G, Schauff S, Welsch N, Albert K,
Hall AJ, Sellergren B: A solid-state and suspended-state magic angle spinning nuclear
magnetic resonance spectroscopic investigation of a 9-ethyladenine molecularly
imprinted polymer. Polymer 48:229–238 (2007)
32. Blasco T, Corma A, Iborra S, Lezcano-González I, Montón R: In situ
multinuclear solid-state NMR spectroscopy study of Beckmann rearrangement of
cyclododecanone oxime in ionic liquids: The nature of catalytic sites. J Catal 275:78–
83 (2010)
33. Pongpeerapat A, Higashi K, Tozuka Y, Moribe K, Yamamoto K: Molecular
Interaction among probucol/PVP/SDS multicomponent system investigated by solid-
state NMR. Pharm Res 23:2566–2574 (2006)
34. Pongpeerapat A, Wanawongthai C, Tozuka Y, Moribe K, Yamamoto K:
Formation mechanism of colloidal nanoparticles obtained from probucol/PVP/SDS
ternary ground mixture. Int J Pharm 352:309–316 (2008)
52
35. C.Wanawongthai, Pongpeerapat A, Higashi K, Tozuka Y, Moribe K, Yamamoto
K: Nanoparticle formation from probucol/PVP/sodium alkyl sulfate co-ground mixture.
Int J Pharm 376:169–175 (2009)
36. Itoh K, Pongpeerapat A, Tozuka Y, Oguchi T, Yamamoto K: Nanoparticle
formation of poorly water-soluble drugs from ternary ground mixtures with PVP and
SDS. Chem Pharm Bull 51:171–174 (2003)
37. Moribe K, Wanawongthai C, Shudo J, Higashi K, Yamamoto K: Morphology
and surface states of colloidal probucol nanoparticles evaluated by atomic force
microscopy. Chem Pharm Bull 56:878–880 (2008)
38. Ficarra R, Tommasini S, Raneri D, Calabro ML, Di Bella MR, Rustichelli C,
Gamberini MC, Ficarra P: Study of flavonoids/beta-cyclodextrins inclusion complexes
by NMR, FT-IR, DSC, X-ray investigation. J Pharm Biomed Anal 29:1005–1014
(2002)
39. Saveyn P, Cocquyt E, Zhu W, Sinnaeve D, Haustraete K, Martins JC, Van der
Meeren P: Solubilization of flurbiprofen within non-ionic tween 20 surfactant micelles:
a 19
F and 1H NMR study. Phys Chem Chem Phys 11:5462–5468 (2009)
40. Wade KE, Wilson ID, Troke JA, Nicholson JK: 19
F and 1H magnetic resonance
strategies for metabolic studies on fluorinated xenobiotics: application to flurbiprofen
[2-(2-fluoro-4-biphenylyl)propionic acid]. J Pharm Biomed Anal 8:401–410 (1990)
41. Cumming JL, Hawker DW, Nugent KW, Chapman HF: Ecotoxicities of
polyquaterniums and their associated polyelectrolyte-surfactant aggregates (PSA) to
Gambusia holbrooki. J Environ Sci Health A Tox Hazard Subst Environ Eng 43:113–
117 (2008)
42. Okano LT, Seoud OAE, Halstead TK: A proton NMR study on aggregation of
cationic surfactants in water: effects of the structure of the headgroup. Colloid Polym
Sci 275:138–145 (1997)
53
43. Shimizu S, Pires PAR, Fish H, Halstead TK, Seoud OAE: Proton and carbon-13
NMR study of the aggregation of benzyl-(2-acylaminoethyl)dimethylammonium
chloride surfactants in D2O. Phys Chem Chem Phys 5:3489–3497 (2003)
44. Shetty AS, Zhang J, Moore JS: Aromatic π-stacking in solution as revealed
through the aggregation of phenylacetylene macrocycles. J Am Chem Soc 118:1019–
1027 (1996)
45. Andrade-Dias C, Lima S, Teixeira-Dias JJ: From simple amphiphilic to
surfactant behavior: analysis of 1H NMR chemical shift variations. J Colloid Interface
Sci 316:31–36 (2007)
46. Farías T, Ménorval LCd, Zajac J, Riveraa A: Solubilization of drugs by cationic
surfactants micelles: Conductivity and 1H NMR experiments. Colloid Surface A:
Physicochem Eng Aspects 345:51–57 (2009)
47. Funasaki N, Fukuba M, Hattori T, Ishikawa S, Okuno T, Hirota S: Micelle
formation of bile salts and zwitterionic derivative as studied by two-dimensional NMR
spectroscopy. Chem Phys Lipids 142:43–57 (2006)
48. Luchetti L: NMR investigation on the various aggregates formed by a gemini
chiral surfactant. Langmuir 16:161–165 (2000)
49. Sharma SC, Acharya DP, Garcia-Roman M, Itami Y, Kunieda H: Phase
behavior and surface tensions of amphiphilic fluorinated random copolymer aqueous
solutions. Colloid Surface A: Physicochem Eng Aspects 280:140–145 (2006)
50. Matsuoka H, Matsutani M, Mouri E, Matsumoto K: Polymer micelle formation
without Gibbs monolayer formation: synthesis and characteristic behavior of an
amphiphilic diblock copolymer having strong acid groups. Macromolecules 36:5321–
5330 (2003)
54
51. Matsuoka H, Maeda S, Kaewsaiha P, Matsumoto K: Micellization of non-
surface-active diblock copolymers in water: special characteristics of poly(styrene)-
block-poly(styrenesulfonate). Langmuir 20:7412–7421 (2004)
52. Matsumoto K, Ishizuka T, Harada T, Matsuoka H: Association behavior of
fluorine-containing and non-fluorine-containing methacrylate-based amphiphilic
diblock copolymer in aqueous media. Langmuir 20:7270–7282 (2004)
53. Kaewsaiha P, Matsumoto K, Matsuoka H: Non-surface activity and
micellization of Ionic amphiphilic diblock copolymers in water: Hydrophobic chain
length dependence and salt effect on surface activity and the critical micelle
concentration. Langmuir 21:9938–9945 (2005)
54. Kaewsaiha P, Matsumoto K, Matsuoka H: Sphere-to-rod transition of non-
surface-active amphiphilic diblock copolymer micelles: a small-angle neutron scattering
study. Langmuir 23:9162–9169 (2007)
55. Coello A, Meijide F, Nunez ER, Tato JV: Aggregation behavior of bile salts in
aqueous solution. J Pharm Sci 85:9–15 (1996)
56. Alvarez Alcalde M, Jover A, Meijide F, Galantini L, Viorel Pavel N, Antelo A,
Vazquez Tato J: Aggregation behavior of tetracarboxylic surfactants derived from
cholic and deoxycholic acids and ethylenediaminetetraacetic acid. Langmuir 25:9037–
9044 (2009)
57. Dormidontova EE: Micellization kinetics in block copolymer solutions: scaling
model. Macromolecules 32:7630–7644 (1999)
58. Kataoka K, Harada A, Nagasaki Y: Block copolymer micelles for drug delivery:
design, characterization and biological significance. Adv Drug Deliv Rev 47:113–131
(2001)
59. Rangel-Yagui CO, Junior AP, Tavares LC: Micellar solubilization of drugs. J
Pharm Pharm Sci 8:147–165 (2005)
55
60. Yuan H-Z, Cheng G-Z, Zhao S, Miao X-J, Yu J-Y, Shen L-F, Du Y-R:
Conformational dependence of Triton X-100 on environment studied by 2D NOESY
and 1H NMR relaxation. Langmuir 16:3030–3035 (2000)
61. Francis MF, Cristea M, Winnik FM: Polymeric micelles for oral drug delivery:
why and how. Pure Appl Chem 76:1321–1335 (2004)
62. Gerber JJ, Caira MR, Lotter AP: Structures of two conformational polymorphs
of the cholesterol-lowering drug probucol. J Cryst Spectrosc 23:863–869 (1993)
63. Stockman BJ, Dalvit C: NMR screening techniques in drug discovery and drug
design. Prog Nucl Mag Res Sp 41:187–231 (2002)
56
LIST OF PUBLICATIONS
This thesis is based on the following publications:
[1] J. Zhang, Y. Tozuka, H. Uchiyama, K. Higashi, K. Moribe, H. Takeuchi, K.
Yamamoto. NMR investigation of a novel excipient, α-glucosylhesperidin, as a suitable
solubilizing agent for poorly water-soluble drugs. J. Pharm. Sci., 100:4421–
14431(2011).
[2] K. Moribe, M. Masaki, R. Kinoshita, J. Zhang, W. Limwikrant, K. Higashi, Y.
Tozuka, T. Oguchi, K. Yamamoto. Guest molecular size-dependent inclusion
complexation of parabens with cholic acid by cogrinding. Int. J. Pharm., 420:191–
197(2011).
[3] J. Zhang, K. Higashi, W. Limwikrant, K. Moribe, K. Yamamoto. Molecular-level
characterization of probucol nanocrystal in water by in situ solid-state NMR
spectroscopy. Int. J. Pharm., in press.
57
LIST OF COMMITTEE
This thesis, conducted for the Degree of Doctoral of Philosophy (Pharmaceutical
Sciences) was examined by the following committee, authorized by the Graduate
School of Pharmaceutical Sciences, Chiba University, Japan.
Professor Yasushi Arano, Ph.D., Chairman
(Graduate School of Pharmaceutical Sciences, Chiba University)
Professor Toshihiko Toida, Ph.D.
(Graduate School of Pharmaceutical Sciences, Chiba University)
Professor Toshiharu Horie, Ph.D.
(Graduate School of Pharmaceutical Sciences, Chiba University)
Professor Katsuhide Terada, Ph.D.
(Faculty of Pharmaceutical Sciences, Toho University)