2. RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
Impact of Nanosizing on Solubility and Dissolution Rate of Poorly
Soluble Pharmaceuticals
SHARAD B. MURDANDE,1
DHAVAL A. SHAH,2
RUTESH H. DAVE2
1
Drug Product Design, Pfizer Worldwide R&D, Groton, Connecticut 06340
2
Arnold and Marie Schwartz College of Pharmacy and Health Sciences, Long Island University, Brooklyn, New York 11201
Received 1 October 2014; revised 2 February 2015; accepted 26 February 2015
Published online in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.24426
ABSTRACT: The quantitative determination of solubility and the initial dissolution rate enhancement of crystalline nanoparticles were
critically investigated using a separation-based approach (ultracentrifugation and filtration). Four poorly soluble model compounds (griseo-
fulvin, celecoxib, compound-X, and fenofibrate) were used in this investigation. The effect of the stabilizer concentration on the solubility of
the unmilled compound was determined first to quantify its impact on the solubility and used for comparing solubility enhancement upon
nanosizing. Methodologies were established for ultracentrifugation, ensuring satisfactory separation of crystalline nanoparticles. The data
obtained using separation-based methodologies proved to be accurate, reproducible, and were in fair agreement with what would be pre-
dicted from the Ostwald–Freundlich equation. The dissolution studies under sink conditions were proved to be less efficient in quantifying
the initial dissolution rate of crystalline nanoparticles. Nonsink dissolution experiments were able to reduce the high-dissolution velocity
of nanoparticles and generated the best discriminative dissolution profile. The enhancement in initial dissolution rate was significantly less
than that expected from the Noyes–Whitney equation based on surface area change. This discriminatory dissolution method can potentially
be used further in the modeling of crystalline nanoparticles during drug development. C 2015 Wiley Periodicals, Inc. and the American
Pharmacists Association J Pharm Sci
Keywords: nanoparticles; nanosuspension; solubility; nanotechnology; dissolution rate; thermodynamics
INTRODUCTION
According to recent trends, the new chemical entities com-
ing out of discovery laboratories in the pharmaceutical indus-
try frequently have low aqueous solubility because of high-
throughput screening.1,2
A number of formulation strategies
have been developed to address the low aqueous solubility of
pharmaceutical compounds.3,4
Among them, particle size reduc-
tion (i.e., via micronization) is a common strategy and has long
been used as a means for improving the oral absorption of poorly
soluble drugs. The method of particle size reduction to enhance
the performance of food, drug, and spices is well known and is
used routinely. However, the extent of dissolution rate enhance-
ment required to improve the in vivo drug absorption for such
poorly soluble drug candidates is greater than that which can
be provided by micronization.5
Therefore, the further reduction
of the particle size to the nanometer range should be considered
as a means for improving the oral absorption of these drugs.6
Several published reports have shown that the nanoparticles
have the potential to improve the oral absorption of poorly sol-
uble drugs along with providing improvements in other bio-
pharmaceutical properties (e.g., variable absorption in the fed
vs. fasted state).7–9
Nanosizing refers to the process of reducing
the active pharmaceutical ingredient (API) size to submicron
range. Nanosuspensions are aqueous dispersions consisting of
a mixture of API and stabilizers, such as surfactant and/or a
polymer in water. Nanosuspensions can be prepared by wet me-
Correspondence to: Sharad B. Murdande (Telephone: +860-715-5975;
Fax: +860-715-4473; E-mail: sharad.murdande@pfizer.com)
This article contains supplementary material available from the authors upon
request or via the Internet at http://onlinelibrary.wiley.com/.
Journal of Pharmaceutical Sciences
C 2015 Wiley Periodicals, Inc. and the American Pharmacists Association
dia milling,10
high-pressure homogenization,11
precipitation or
a combination of the two approaches mentioned previously.12
Studies conducted in this report utilized nanosuspensions pre-
pared by the wet media milling techniques.
A solubility advantage for amorphous pharmaceuticals over
their crystalline counterparts has been very well studied in
the literature.13,14
However, solubility advantage for nanocrys-
tals and their impact on thermal and kinetic properties has
not been explored in detail. Recent reports on solubility and
dissolution improvement upon nanosizing have reported vari-
able results.15,16
The most common practice for the solubility
determination reported was based on the separation by filtra-
tion to separate undissolved nanoparticles from the solution
utilizing the membrane filters of various pore sizes (0.1, 0.22,
and 0.45 :m pore size).17–19
A few reports have also shown the
use of molecular weight cut-off filters,20
dialysis bag method,21
centrifugation, and ultracentrifugation22
for the separation of
dissolved and undissolved nanocrystals during solubility mea-
surement. However, confirmation of satisfactory separation of
dissolved and undissolved fractions of crystalline nanoparticles
has been ambiguous. Determining accurate solubility is a much
needed tool in the preformulation study of any nanocrystalline
particles of a poorly soluble compound. However, it becomes
challenging to accurately measure solubility when the drug
particle size is in the nanometer range. Complicating this is
the tendency of nanoparticles that remain suspended in the
solution even after centrifugation.4
Therefore, development of
a method for precise solubility determination associated with
nanocrystals was one of the objectives of this research.
A discriminating in vitro drug dissolution test plays an im-
portant role in pharmaceutical product development as it may
be used as a representative tool for the assessment of formu-
lation performance in vivo. Results of the dissolution studies
Murdande, Shah, and Dave, JOURNAL OF PHARMACEUTICAL SCIENCES 1
3. 2 RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
for nanocrystalline formulations have been documented in
several reports where studies were carried out under sink
conditions.23–25
At sink condition, nanocrystals dissolve instan-
taneously, which hinders the discrimination of dissolution be-
havior with respect to different particle size. Recently, a few
reports have shifted the thinking in this direction and tried
other approaches for carrying out the dissolution of nanosus-
pensions. Among these new approaches, turbidimetry is a re-
ported method that can be used for real-time monitoring of
small particle dissolution. Crisp et al.26
reported that the in-
terfacial rate constant that is obtained from the turbidimetry
method is independent of particle size; hence, it can quantita-
tively predict the dissolution rate for particle sizes down to the
submicron range. In situ measurements using potentiometry27
and a solution calorimetry have also been investigated.28
In the
case of the potentiometric methods, sensors are insensitive to-
ward air bubbles and undissolved particles and can be a good
tool for accurate dissolution studies. However, interference by
the dynamics of dissolution and the requirement of precondi-
tioned sensors with API are major drawbacks for this analyti-
cal method.27
Solution calorimetry also offers advantages over
the classical dissolution testing approach by avoiding errors
because of filtration and sample preparation. It is, however, a
tedious method and requires a significant amount of additional
time compared with the conventional methods.28
In the present
study, the dissolution rate was examined in different nonsink
conditions,29
for efficient discrimination of the dissolution rate
enhancement with mean particle sizes that ranged from 80 nm
to 11 :m.
The present study primarily focused on investigating the
thermodynamic properties (the change in the melting point,
heat of fusion, and accurate solubility determination) and ki-
netic properties (the quantitative determination of initial disso-
lution rate enhancement with respect to particle size reduction
upon nanosizing) for a nanosuspension by developing the suit-
able analytical methods.
MATERIALS AND METHODS
Materials
Four poorly soluble, nonionizable crystalline model compounds
(griseofulvin, compound-X, celecoxib, and fenofibrate) were cho-
sen for this study. Celecoxib and compound-X were received
from Pfizer Inc. (Groton, Connecticut). Griseofulvin was ob-
tained from Spectrum Chemicals (New Brunswick, New Jer-
sey). Fenofibrate was purchased from Sigma–Aldrich (St. Louis,
Missouri). All model compounds were used as received. The
yttrium stabilized zirconium oxide beads were obtained from
Glen Mills Inc. (Clifton, New Jersey). Hydroxypropylcellu-
lose (HPC)–SL was obtained from Nisso America Inc. (New
York, New York). Other laboratory chemicals such as docusate
sodium, USP, acetonitrile, and trifluoroacetic acid (TFA) were
purchased from Spectrum Chemicals and were of analytical
grades.
Nanosuspension Preparation
Nanosuspensions of four model compounds were prepared by
attrition media milling. The milling vehicle was a mixture of
1.25% (w/v) HPC–SL (stearic stabilizer) and 0.05% (w/v) do-
cusate sodium (ionic stabilizer) in deionized water. Optimal
concentration of stabilizers was determined from earlier exper-
iments. Yttrium-stabilized zirconium oxide beads were used as
a grinding media (500 and 100 :m). The ratio of the grind-
ing media to API in suspension was 3:1 (w/w). The milling
speed was 1000 rpm. Temperature was maintained by the sur-
rounded cold water outside of the milling chamber to compen-
sate for the heat generated by high-speed milling. Targeted
final concentration for nanosuspension was 50 mg/mL. The av-
erage milling time for all model compounds was 4 h. Milling was
stopped after 4 h and the grinding media was allowed to settle
down for 30 min. The suspension was separated by filtration
through a coarse sintered glass filter (75 :m). The nanosus-
pension was harvested by removing the grinding media and
stored under refrigerated conditions (4°C) to evaluate physical
stability. Nanosuspensions were assayed for drug potency by
HPLC.
Characterization of Nanocrystal Particle Size and Zeta
Potential Measurements
A dynamic light scattering (DLS) instrument (Malvern Instru-
ments Inc., Malvern, UK) was used to measure particle size
and the size distribution of nanocrystalline suspensions. The
samples were prepared by diluting 5–10 :L of nanosuspension
with HPLC grade water. The sample was analyzed at 25°C, us-
ing a dispersant refractive index of 1.33. The attenuation and
measurement settings were optimized automatically by resi-
dent software. The polydispersity index (PI) and count rate
for the measurement were taken into consideration after the
measurement to assess the quality of the nanosuspensions pre-
pared. Particle size data obtained were expressed in terms of
D10, D50, and D90 (10%, 50%, and 90% of the particles’ volume
below certain sizes) along with the average particle size. In the
current study, the zeta potential of the prepared nanosuspen-
sions was measured in disposable-folded capillary cells (Model
#DTS1061) along with the particle size measurement using the
DLS instrument.
Morphological Characterization
Morphological characterization was performed using a polar-
ized light microscope (PLM; Model BX60FS; Olympus BX60,
Center Valley, Pennsylvania). The polarized microscopic images
were taken using a camera from Diagnostic Instruments Inc.
(Sterling Heights, Michigan) and the analysis was performed
with the help of SPOT
TM
imaging advanced software. Samples
were prepared by placing a 4-:L nanosuspension on a clean
glass slide and covering it carefully with a clean glass cover
slip. Nanocrystals were observed under 500x optical zoom in a
dark field.
Measurement of Thermal Properties
Sample Preparation for Differential Scanning Calorimetry
and X-ray Powder Diffraction
The initial unmilled drug suspension was prepared by mixing
micronized API with the stabilizer solution at 1000 rpm for
4 h at 25°C. The same stabilizer solution was used for both
the unmilled and the nanosuspension preparation. Thus, the
initial (unmilled) suspension was treated in the same manner
as the nanosuspensions, with respect to the stabilizer concen-
tration in the solution. Both the unmilled suspension and the
nanosuspensions were transferred via an eppendorf vial and
centrifuged using an Allegra 21 centrifuge, GS-15 (Beckman
Coulter Inc., Schaumberg, Illinois). Centrifugation was carried
Murdande, Shah, and Dave, JOURNAL OF PHARMACEUTICAL SCIENCES DOI 10.1002/jps.24426
4. RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology 3
out at 6882 RCF for 15 min (at 25°C). The supernatant was
discarded from each vial and replaced with fresh DI water for
washing the particles to remove excess stabilizer that was ad-
sorbed on the surface of the crystalline particles. The washing
procedure was repeated six times for each model compound.
After discarding the supernatant liquid, wet particles from the
bottom of the eppendorf were collected and subjected to X-ray
powder diffraction (XRPD) analysis. The remaining portion was
kept overnight in a vacuum dryer and subjected to differential
scanning calorimetry (DSC) analysis afterwards.
X-ray Powder Diffraction
The crystalline properties of formulated nanosuspensions were
assessed by XRPD using a Bruker D4 diffractometer (Bruker
Bio Spin Corporation, Billerica, Massachusetts). The XRPD
measurements were carried out in the standard measurement
mode in the 22range from 5° to 40°. The scan speed was 2°/min
and the counting step was 0.02°. The X-ray source was CuK"
(wavelength = 1.5406 ˚A); the accelerating voltage was 40 kV,
and the current was 40 mA. Samples were prepared by plac-
ing them in a silicon cavity wafer mount as described above.
Data were collected and analyzed using the Bruker DIFFRAC
Plus software suite (Bruker Bio Spin Corporation, Billerica,
Massachusetts).
Melting Point and Heat of Fusion Measurement
Thermal properties (i.e., the melting point and the heat of fu-
sion measurement) were investigated by a differential scan-
ning calorimeter (DSC Q1000; TA Instruments, New Castle,
Delaware). The instrument was calibrated using standard in-
dium samples. Calibration of the heat capacity was performed
using sapphire. The 5–10-mg samples of pure API, unmilled
suspension, and nanosuspension were sealed in standard alu-
minum pans and subjected to analysis. The DSC parame-
ters used were: a sample equilibration at 30°C, modulated
for ±0.318°C every 60 s, isothermal for 5 min, and heated to
180°C/230°C with a ramp of 2°C/min. The DSC cell was purged
with nitrogen at a flow rate of 50 mL/min. The data obtained
were analyzed with the Universal Analysis Software (TA In-
struments). All studies were performed in triplicate.
Experimental Solubility Measurement
HPLC Analysis Method
The quantitative analysis of model compounds used in the
solubility studies was performed by reverse-phase gradient
HPLC (Agilent 1100 Series and Alliance 2695 Module). Analy-
ses were performed using an Agilent 1100 Serial System (Agi-
lent Technologies, Palo Alto, California) and Alliance 2695 Sep-
aration Module (Waters Corporation, Milford, Massachusetts)
equipped with a 4.6 × 75 mm, 2.7 :m C18 column (Advanced
Materials Technology, Wilmington, DE). The temperature of
the column was kept at 45°C, and the wavelength of detection
was set to 210 nm. The mobile phase consisted of a binary gra-
dient of solvent A (acetonitrile) and solvent B (HPLC grade
water with 0.05% TFA). The linear gradient started at 5% A
and increased to 95% A in 8.5 min, followed by a return to the
starting condition within 1.6 min and equilibrated at the start-
ing condition for 2 min. The injection volume was either 5 or
10 :L. The flow rate was 1.0 mL/min, and the total run time
was set for 13.10 min. This reverse-phase gradient method pro-
vided baseline resolution and excellent peak characteristics for
each model compound. Quantification of drug concentrations
was performed by analyzing the peak area at retention using
Agilent Chemstation R
and Waters Empower Pro R
analytical
software.
Thermodynamic Solubility Determination
Solubility determination was carried out using the tempera-
ture cycling method with a rotating mixture equipped with
ENVIRO GENE software (Scientific Industry Inc., Bohemia,
NY). A three-step temperature cycling method was used for
equilibrium solubility determination in which samples were
rotated for 8 h at 40°C followed by 5 h at 15°C, which was then
followed by 12 h at 25°C. To achieve an accurate measurement,
the solubility determination was conducted in the presence of a
stabilizer solution for both unmilled and nanocrystalline model
compounds. The study provided more precise solubility data
(compared with the solubility in the water), as both unmilled
and nanocrystals were treated in the same manner.
r Aqueous solubility determination of model compounds
Solubilities of the crystalline form of the model compounds
were measured in DI water. An excess amount of API was dis-
persed in DI water and subjected to the temperature cycling
method as described above. The samples were filtered using
0.1 :m anotop syringe filters and subjected to HPLC analysis.
The aqueous solubility of all the model compounds was less
than 10 :g/mL, except for the compound-X (47.2 :g/mL).
r Effect of nanosizing stabilizer on drug solubility
The stabilizer can affect the drug solubility by altering the
wetting property of the API. Crude suspensions were prepared
with various stabilizer concentrations, that is, (1) no stabi-
lizer, (2) 0.625% HPC-SL + 0.025% docusate sodium, (3) 1.25%
HPC-SL + 0.05% docusate sodium, and (4) 2.5% HPC-SL +
0.1% docusate sodium. Samples were placed in a temperature-
controlled rotary mixture equipped with ENVIRO GENE soft-
ware for 25 h. At the end of the equilibration, it was very diffi-
cult to filter the nanosuspension samples through 0.1/0.02 :m
filters because of higher concentration of particles and also the
filter blockage was observed. Therefore, samples were subjected
to ultracentrifugation at a speed of 313130 RCF at 25°C for 60
min in a 100% vacuum using a fixed angle rotor in a Beckman
Coulter Ultracentrifuge (Beckman Coulter Inc.). Supernatants
were collected immediately and subjected to HPLC analysis for
equilibrium solubility determination. The same supernatants
were filtered from 0.1 :m anotop syringe PTFE polypropylene
filters (WhatmanTM
; GE Healthcare, Piscataway, NJ) and sub-
jected to HPLC analysis to measure the equilibrium solubility.
r Nanocrystalline particle solubility in stabilizer solution
Nanosuspensions of each compound were transferred to a
Beckman Coulter Ultracentrifuge. The ultracentrifugation pa-
rameters were kept consistent throughout the study. After
ultracentrifugation, the supernatants were collected immedi-
ately and transferred to an HPLC vial for solubility determina-
tion. The same supernatants of nanosuspensions were filtered
from 0.1 :m anotop syringe filters for the particle size greater
than 200 nm, and the supernatants for nanosuspensions with
DOI 10.1002/jps.24426 Murdande, Shah, and Dave, JOURNAL OF PHARMACEUTICAL SCIENCES
5. 4 RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
particle size less than 125 nm were filtered from 0.02 :m an-
otop syringe filters (WhatmanTM
; GE Healthcare, Piscataway,
NJ) for HPLC analysis. The first few milliliters of each sample
were discarded to minimize losses because of adsorption. The
discarded volume selected for each compound was determined
by a separate filter validation experiment (data not shown).
The solubility was also assessed from the dissolution profile by
allowing the dissolution study to proceed until equilibrium was
established. All solubility measurements were carried out in
triplicate.
Measurement of Dissolution Rate
A dissolution apparatus (Distek 2100C; Distek Inc., North
Brunswick, New Jersey) equipped with a USP type-II (pad-
dle) was used to measure dissolution rates of nanocrystalline
suspensions. Griseofulvin was chosen as a model drug for the
dissolution study. The dissolution study was carried out by com-
paring unmilled suspension (micronized mean particle size of
11.3 :m) and nanosuspensions (mean particle sizes of 362 and
122 nm). Each dissolution study was carried out at 37±0.5°C
under a 100-rpm paddle speed in a USP Type II apparatus for
2 h. The concentration of suspensions was made to 5 mg/mL by
diluting 1 mL of each suspension with 10 mL of stabilizer so-
lution [1.25% (w/v)/0.05% (w/v)-HPC-SL/docusate sodium]. Ac-
cordingly, 0.2 mL of suspension (equivalent to 1 mg of API)
was used for the dissolution study under sink condition (a) and
under reducing the sink condition factor (b).
r Dissolution under sink condition
Each dissolution study was carried out in 900 mL DI water.
The sink condition factor was maintained at 5.7× (calculated
from experimentally determined solubility value for the grise-
ofulvin). A 10-mL of a dissolution sample was withdrawn at
intervals of 0, 2, 5, 7, 10, 20, 30, 45, 60, 90, and 120 min and re-
placed with same volume of dissolution media (preequilibrated
at 37±0.5°). Samples were filtered through a 0.1-:m (unmilled
11.3 :m and 362 nm) and 0.02 :m (<125 nm) anotop syringe
filter.
r Dissolution under reducing sink condition factor
The dissolution set up and sample preparation was main-
tained as described for the sink condition dissolution experi-
ment. The only difference was with the use of the dissolution
media. In order to reduce the sink condition factor, 40% and
55% of griseofulvin (based on its experimentally determined
solubility value) were predissolved in DI water. After dissolving
the appropriate amount of griseofulvin, the media was filtered
through 0.22 :m Millipore Express Plus filter membrane. The
dissolution studies were carried out in the dissolution media
with a reduced sink condition factor of about 1.8× and 1.4×, re-
spectively. A 10 mL of a dissolution sample was withdrawn at
predefined intervals and replaced with same volume of dissolu-
tion media (pre-equilibrated at 37±0.5°). Samples were filtered
through a 0.1 :m (unmilled 11.3 :m and 362 nm) and 0.02 :m
(<125 nm) anotop syringe filter.
r Effect of griseofulvin loading dose on dissolution under
nonsink condition
The dissolution set up was maintained as described for the
sink condition dissolution experiment. The sample preparation
and the dissolution media were different as compared with pre-
vious dissolution studies. The dissolution study was performed
by generating two types of nonsink conditions.
r Dissolution media in which 25%, 40%, 55%, and 75% of
griseofulvin was predissolved in DI water and the sample
loading dose was of 10 mg of griseofulvin.
r Dissolution media in which 20%, 40%, and 75% of griseo-
fulvin was predissolved in DI water and the sample load-
ing dose was of 50 mg griseofulvin.
The dissolution media was filtered through a 0.22 :m Milli-
pore Express PLUS filter membrane. The suspension samples
were equivalent to either 10 or 50 mg of API (0.2 mL/1 mL
with the original concentration of 50 mg/mL), and were used
for the dissolution study to generate complete nonsink condi-
tions. Samples were withdrawn at 0, 2, 5, 7, 10, 20, 30, 45, 60,
90, and 120 min and replaced with the same volume of dissolu-
tion media (preequilibrated at 37±0.5°) for the analysis of the
drug concentration. Samples were filtered through a 0.1 :m
(unmilled 11.3 :m and 362 nm) and 0.02 :m (<125 nm) anotop
syringe filters. The average of the three replicate tests was re-
ported for each of the dissolution studies.
RESULTS AND DISCUSSION
Particle Size and Zeta Potential Measurement
The mean particle sizes of wet-milled crystalline nanosuspen-
sions of all four model compounds were found to be within 93–
362 nm range (Table 1) with an acceptable PI (<0.3). The av-
erage particle size data, along with their D10, D50, and D90 val-
ues, and PI and zeta potential values, were obtained from DLS.
These data indicate that the wet-milled crystalline nanosus-
pensions were homogeneous with a narrow particle size dis-
tribution. Experimentally determined zeta potential values
ranged from −22.7 to −30.3 mV, suggesting good physical
stability of the formulated nanosuspensions.30
The stabiliza-
tion of nanosuspensions can be attributed to two main factors.
(1) The adsorption of HPC-SL on the drug particle provides
stearic stabilization and (2) the docusate sodium provides elec-
trostatic repulsion by forming the ionic charges at the drug
particle surface.
Morphological Characterization
Morphological analysis by PLM micrographs demonstrated the
presence of nanosized crystalline particles (data not shown).
The images clearly show the presence of particles in the
nanometer size range.
X-ray Powder Diffraction
X-ray powder diffraction was used to determine changes in
the crystalline state of all model APIs in the nanosuspensions.
X-ray diffractograms of crystalline nanosuspensions were com-
pared with pure API and the stabilizer-treated unmilled API.
The crystallinity in the nanosuspension is confirmed by the re-
tention of characteristic peaks in X-ray diffractograms (data not
shown). Moreover, the light microscopy data are also supportive
to XRPD data for the crystallinity (data not shown). Dominant
Murdande, Shah, and Dave, JOURNAL OF PHARMACEUTICAL SCIENCES DOI 10.1002/jps.24426
6. RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology 5
Table 1. Impact of Particle Size on Melting Point and Heat of Fusion
Compounds Size (nm) PdI Sample Name MP (°C) MP (°C) Hf (J/g) Hf (J/g) Zeta Potential (−mV)
Griseofulvin 362 0.095 API 219.1 1.3 111.8 15.2 −24.1
NC 217.8 96.6
Comp.-X 243 0.210 API 207.5 1.9 101.7 10.9 −24.6
NC 205.6 90.7
Celecoxib 341 0.272 API 162.1 2.1 100.5 26.1 −29.5
NC 160 74.5
Fenofibrate 290 0.192 API 80.4 1.2 95.4 8.9 −30.3
NC 79.2 86.5
Griseofulvin 122 0.131 API 219.1 3.5 111.8 15.2 −28.2
NC 215.5 96.6
Comp.-X 93 0.172 API 207.5 5.3 101.7 34.1 −22.7
NC 202.2 67.6
MP, melting point; Hf, heat of fusion; API, active pharmaceutical ingredient; NC, nanocrystals; PdI, polydispersivity index.
peaks at 22angle were observed in the nanosuspension formu-
lations suggesting the conservation of the crystalline property
of all model compounds, though reductions in both heat of fu-
sion and in melting point were observed (Table 1).
Determination of Thermal Properties
Differential Scanning Calorimetry
A DSC analysis was performed for the investigation of ther-
mal properties and further confirmation of the crystalline state
after particle size reduction. The thermograms of the nanocrys-
talline samples were compared with the unmilled API to con-
firm the crystallinity of the API. All the nanocrystalline for-
mulations showed an endothermic melting event. Here, the
nanocrystals of griseofulvin (mean particle size 362 nm), pro-
duced a relatively sharp but slightly shifted endothermic melt-
ing peak at 217.8°C (from 219.1°C), along with a shift in Hf
(from 111.8 J/g to 96.6 J/g). The observed peak shift has been
reported to be in agreement with the particle size reduction.31,32
For all other APIs, the observed changes in the melting point
and heat of fusion with respect to particle size reduction are
presented in Table 1. For the particle sizes between 290 and
362 nm, the melting point reduction was approximately 1.2°C–
2.1°C and the heat of fusion reduction was around 8.9–26.1 J/g.
To further evaluate the thermal changes, crystalline nanosus-
pensions were formulated down to 100 nm particle size. A fur-
ther reduction in both the melting point (3.5°C and 5.3°C) and
the heat of fusion (15.2 and 35 J/g) was observed when the
particle size was reduced close to 122 and 93 nm, respectively.
The phenomena of change in the melting point and the heat
of fusion because of the reduction in particle sizes can also be
explained by the Gibbs–Thomson equation.33,34
Equilibrium Solubility
Crystalline nanoparticles of very poorly soluble compounds (sol-
ubility <10 :g/mL) for which even a small error in solubility
measurement can have a significant impact on the results. In
addition, it is also important to study the filter binding capacity
of a drug molecule to avoid underestimating the actual solubil-
ity data. Selection of the analytical method (i.e., UV/HPLC)
for quantitative analysis also plays an important role. In the
current research, we established a systematic approach to ac-
curately measure the solubility of crystalline nanoparticles in
order to estimate the improvement in solubility that can be
achieved upon nanosizing. Saturation solubilities for the pure
API were experimentally determined in DI water at room tem-
perature. The RSD values for the solubilities were within 2% of
the reported values for the crystalline drugs. Solubility studies
were also performed for unmilled drugs with different concen-
trations of the stabilizer solution. The reported data suggest
that the solubility number rises proportionally with the concen-
tration of the stabilizer. The concentration [1.25% (w/v) HPC SL
+ 0.05% (w/v) docusate sodium] at which the nanosuspension
formulation had been made have very less impact on the solu-
bility (data not shown). So, it is fair to compare the solubility re-
sults in the stabilizer solution for both the milled and unmilled
particles. The solubility enhancement observed can be consid-
ered solely because of the particle size. The average solubility
(n = 3) of unmilled API and nanocrystals in the stabilizer so-
lution [1.25% (w/v) HPC SL + 0.05% (w/v) docusate sodium]
is shown in Table 2. The solubility values determined from
the supernatant of the ultracentrifugation and ultracentrifu-
gation plus filters in the case of griseofulvin were 8.12 versus
7.63 :g/mL for unmilled API. Whereas in the case of celecoxib,
solubility values were 1.45 versus 1.00 :g/mL, and for fenofi-
brate, 1.53 versus 0.74 :g/mL. The reported data were the aver-
age number of triplicate measurements along with the standard
deviations. There was no significant difference in the solubil-
ity between nanocrystals and the unmilled API in both cases
(i.e., ultracentrifugation and ultracentrifugation +0.1 :m fil-
ter). The solubility enhancement ratio was found to be minimal
when the particle size was reduced from 292 to 362 nm. The ex-
perimental solubility enhancement ratio based on particle size
was also compared with the theoretical solubility enhancement
ratio described by the Ostwald–Freundlich equation. Note that
for the theoretical calculation, the interfacial tension is equal
to the surface tension of water of 72 dyne/cm2
has been made
in the Ostwald–Freundlich equation.35
The results for solubil-
ity enhancement were similar with what would be expected
based on the Ostwald–Freundlich equation (data not shown).
In studying the impact of particle size on solubility, when the
size is reduced to close to 100 nm, the nanosuspensions were
prepared for griseofulvin and compound-X with mean a par-
ticle size of 122 and 93 nm, respectively. The solubility mea-
surements were performed using a 0.02 :m syringe filter in-
stead of 0.1 :m because of a further reduction in the particle
size (Table 2). The solubility for griseofulvin was 10.30 versus
8.12 :g/mL (ultracentrifugation) and 9.99 versus 7.63 :g/mL
DOI 10.1002/jps.24426 Murdande, Shah, and Dave, JOURNAL OF PHARMACEUTICAL SCIENCES
7. 6 RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
Table 2. Solubility of Crystalline Nanoparticles
Compounds Sample
Solubility
Unmilled Drug
(:g/mL)
Solubility of
Nanocrystal
Drug (:g/mL)
Solubility Ratio
(Nanocrystal/
Unmilled Drug) Size (nm)
Griseofulvin UC 8.12 ± 0.14 8.17 ± 0.28 1.01 362
UC + 0.1 :m filter 7.63 ± 0.89 8.40 ± 0.25 1.1
Comp.-X UC 65.17 ± 1.58 63.41 ± 1.26 0.97 243
UC + 0.1 :m filter 74.44 ± 3.07 68.86 ± 10.73 0.93
Celecoxib UC 1.45 ± 0.51 1.50 ± 0.17 1.03 341
UC + 0.1 :m filter 1.00 ± 0.03 1.11 ± 0.03 1.11
Fenofibrate UC 1.53 ± 0.21 1.66 ± 0.66 1.08 290
UC + 0.1 :m filter 0.74 ± 0.27 0.82 ± 0.26 1.11
Griseofulvin UC 8.12 ± 0.14 10.30 ± 0.18 1.26 122
UC + 0.02 :m filter 7.63 ± 0.89 9.99 ± 0.15 1.3
Comp.-X UC 65.17 ± 1.58 89.06 ± 6.36 1.36 93
UC, ultracentrifugation.
(ultracentrifugation + filter). In the case of compound-X,
the solubility from the ultracentrifuge was 89.06 versus
65.17 :g/mL. The maximum solubility enhancement observed
upon nanosizing was about 30% compared with unmilled API.
A similar enhancement of 15% in the solubility upon nanosiz-
ing was also reported when the particle size was reduced from
158 to 700 nm.22
On the basis of the theoretical calculations
of the Ostwald–Freundlich equation, in order to get 30% en-
hancement in thermodynamic solubility, the required particle
size should be approximately 160 nm. Therefore, the observed
solubility enhancement achieved upon nanosizing is in fair
agreement with what can be expected, based on the Ostwald–
Freundlich equation. The reported solubility enhancement ra-
tios are shown in Table 2. Note that the solubility determination
studies were conducted by maintaining precision in the gener-
ated data (i.e., separate filter validation experiments; each mea-
surement was carried out running a fresh calibration curve for
all the model compounds, along with the samples during each
quantitative analysis). Therefore, the reported data were accu-
rate enough to understand the solubility behavior of nanocrys-
tals when the particle size was reduced to approximately
100 nm.
Dissolution Studies
Particle size reduction also has a major impact on the dis-
solution rate of crystalline drugs. As per the Noyes–Whitney
equation, a positive effect (i.e., increase in the dissolution rate
with increase in the surface area) can be observed in the dis-
solution rate when the particle size is reduced. Therefore, an
accurate measurement of enhancement in the dissolution rate
is important in order to understand the observed in vivo phar-
macokinetic results obtained using nanosuspension relative to
unmilled crystalline suspension.
Initially the dissolution behaviors of nanosuspensions were
determined by maintaining perfect sink condition following the
guidelines in European Pharmacopeia (Fig. 1a). A compari-
son of dissolution profiles was made among the 122, 362 nm,
and 11.3 :m size griseofulvin suspensions. From the dissolu-
tion profiles, it is clear that nanocrystals dissolved instanta-
neously (>80% API dissolved in <2 min for both nanocrys-
talline suspensions). However, the micronized unmilled
suspension showed a slow release compared with the nanosus-
pensions. It took more than 20 min for the micronized suspen-
sion to release 80% of the API. Hence, we can get a comparable
Figure 1. In vitro dissolution studies (griseofulvin: 10 mg loading dose): comparing unmilled API, 362 and 122 nm size of particles: (a) sink
condition dissolution and (b) reducing sink condition factor (1.8×, 40% saturation and 1.4×, 55% saturation).
Murdande, Shah, and Dave, JOURNAL OF PHARMACEUTICAL SCIENCES DOI 10.1002/jps.24426
8. RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology 7
Figure 2. Nonsink dissolution (griseofulvin:10 mg loading dose): comparing unmilled API, 362 and 122 nm size of particles: (a) 25% and 40%
saturation and (b) 55% and 75% saturation.
idea of the dissolution profile using the traditional sink condi-
tion dissolution method. An improvement in the dissolution ve-
locity upon nanosizing is well explained in multiple reports.36,37
However, initial linear regions in sink conditions lacked dis-
crimination for the initial dissolution rate enhancement. Also,
nanocrystals with particle sizes of 362 and 122 nm had su-
perimposable dissolution profiles. This concludes that the sink
condition dissolution profile is insufficient in providing actual
discrimination for the initial dissolution rate enhancement and
its relation to the observed improvement in various PK parame-
ters. For further discrimination and quantification of the initial
dissolution rate, studies were carried out by reducing the sink
condition factor as described in the method section. The dissolu-
tion profiles for the reduced sink condition factor are presented
in Figure 1b. A clear discrimination was observed between mi-
cronized and nanocrystals in the initial region of the dissolu-
tion profile. However, the dissolution profiles for both differ-
ent sized nanocrystals were superimposable on each other. The
initial dissolution rate ratios were calculated from dissolution
at the first couple of time points (<5 min into the dissolution
process).38
In order to further discriminate the dissolution profiles of
nanosized griseofulvin, studies were performed by generating
the two types of nonsink conditions as described in the Mate-
rials and Methods section. The dissolution profiles are shown
in Figures 2 and 3. The calculated initial dissolution rate ra-
tios are reported in Table 3. All of the dissolution profiles with
Figure 3. Nonsink dissolution (griseofulvin: 50 mg loading dose):
comparing initial dissolution rate enhancement for unmilled API, 362
and 88 nm size of particles at 20%, 40%, and 75% saturation level.
nonsink conditions showed significant discrimination between
micronized and nanocrystals. Nonsink conditions (loading dose
10 mg) with saturation levels of 25% and 40% were not suffi-
cient enough for discrimination of the initial dissolution rate
enhancement (Fig. 2a). All the dissolution profiles started to be
discriminating with each other as the saturation level increased
to 55%. When the saturation level reached 75%, maximum dis-
crimination was observed (1.8 vs. 3.6) for the initial dissolution
rate ratios between both the nanocrystals (Fig. 2b). The dissolu-
tion profiles with nonsink conditions and higher loading doses
(i.e. 50 mg) showed discrimination in all three saturation levels
(Fig. 3).
The findings of the current study justified the use and the
need for a nonsink dissolution approach to accurately evaluate
the dissolution rate enhancement upon nanosizing. Current
dissolution practices (i.e., sink condition) as reported in the
literature justifies their results by citing the Noyes–Whitney
equation,39,40
which is most popular for studying the dissolu-
tion behavior of particles greater than a few microns. Crisp
et al.26
have also reported that when particle size is reduced to
nanometer range, the interfacial phenomena plays a more im-
portant role compared with the surface area, so the dissolution
rate predicted from the Noyes–Whitney equation will not corre-
late well with the observed data. However, our findings for the
dissolution rate enhancement are not in close agreement with
respect to surface area change. The calculated surface area in-
creases upon nanosizing were much larger than the observed
dissolution rate enhancement upon nanosizing for griseofulvin
(Fig. 4).
CONCLUSIONS
We have established a reliable method for measuring the solu-
bility of crystalline nanoparticles obtained from the wet-milling
technique. The significant improvement in the separation of
undissolved nanoparticles was achieved by ultracentrifugation.
The observed increase in the solubility upon nanosizing was
only marginal and in close agreement with the theory. There-
fore, any improvement in in vivo pharmacokinetics that can be
achieved with nanoparticles can be expected to be mainly be-
cause of the dissolution rate enhancement. The enhancement
in dissolution achieved upon nanosizing is best studied un-
der nonsink conditions. However, factors that contribute to the
DOI 10.1002/jps.24426 Murdande, Shah, and Dave, JOURNAL OF PHARMACEUTICAL SCIENCES
9. 8 RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology
Table 3. Initial Dissolution Rate Enhancement at Various Saturation Levels
Ratio
Dissolved in 2 min (%) 362 nm 122 nm 122 nm
Sink Condition Factora Unmilled 362 nm 122 nm Unmilled Unmilled 362 nm
5.7× 38 79 79 2 2.1 1
1.8× 38 76 86 2 2.2 1.1
1.4× 24 86 98 3.6 4 1.11
Nonsink Conditionb
362 nm 122 nm 122 nm
Saturation Level Unmilled 362 nm 122 nm Unmilled Unmilled 362 nm
20% 14 44 49 3.1 3.5 1.1
40% 11.7 35 49 3 4.2 1.3
55% 13 36 49 2.76 3.76 1.3
75% 18 33 65 1.8 3.6 2
Nonsink Conditionc
362 nm 88 nm 88 nm
Saturation Level Unmilled 362 nm 88 nm Unmilled Unmilled 362 nm
20% 15 19 29 1.2 2 1.5
40% 12 16 30 1.3 2.4 1.75
75% 7 10 22 1.4 3.15 2.4
a
Solubility of griseofulvin in dissolution media; 6.27 :g/mL.
b
Nonsink dissolution; 10 mg griseofulvin in 900 mL of dissolution media.
c
Nonsink dissolution; 50 mg griseofulvin in 900 mL of dissolution media.
Figure 4. Comparing predictive initial dissolution rate enhancement
ratio upon surface area change with experimentally determined initial
dissolution rate enhancement ratio.
dissolution rate enhancement are not directly related to the sur-
face area increase for crystalline nanoparticles. Further studies
are needed to understand the impact of surface area change and
the surface-free energy change that occurs upon nanosizing on
the dissolution rate enhancement.
ACKNOWLEDGMENTS
The authors would like to acknowledge Ravi M. Shanker, Ph.D
and Jigna D. Patel, Ph.D for their helpful discussion.
REFERENCES
1. Lipinski CA. 2000. Drug-like properties and the causes of poor solu-
bility.pdf. J Pharmacol Toxicol Methods 44:235–249.
2. Rabinow BE. 2004. Nanosuspensions in drug delivery. Nat Rev Drug
Discov Today 3(9):785–796.
3. Dave RH, Shah DA, Patel PG. 2014. Development and Evaluation
of High Loading Oral Dissolving Film of Aspirin and Acetaminophen.
Journal of Pharmaceutical Sciences and Pharmacology 1(2):112–122.
4. Williams HD, Trevaskis NL, Charman SA, Shanker RM, Charman
WN, Pouton CW, Porter CJH. 2013. Strategies to address low drug
solubility in discovery and development. Pharmacol Rev 65:315–499.
5. Cooper ER. 2010. Nanoparticles: A personal experience for formu-
lating poorly water soluble drugs. J Control Release 141(3):300–302.
6. Sigfridsson K, Lundqvist AJ, Strimfors M. 2009. Size reduction for
improvement of oral absorption of the poorly soluble drug UG558 in rats
during early development. Drug Dev Ind Pharm 35(12):1479–1486.
7. Liversidge GG, Conzentino P. 1995. Drug particle size reduction for
decreasing gastric irritancy and enhancing absorption of naproxen in
rats. Int J Pharm 125(2):309–313.
8. Liversidge GG, Cundy KC. 1995. Particle size reduction for im-
provement of oral bioavailability of hydrophobic drugs: I. Absolute oral
bioavailability of nanocrystalline danazol in beagle dogs. Int J Pharm
125(1):91–97.
9. Kesisoglou F, Panmai S, Wu Y. 2007. Nanosizing–oral formulation
development and biopharmaceutical evaluation. Adv Drug Deliv Rev
59(7):631–644.
Murdande, Shah, and Dave, JOURNAL OF PHARMACEUTICAL SCIENCES DOI 10.1002/jps.24426
10. RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology 9
10. Merisko-Liversidge E, Liversidge GG. 2011. Nanosizing for oral and
parenteral drug delivery: A perspective on formulating poorly-water
soluble compounds using wet media milling technology. Adv Drug Deliv
Rev 63(6):427–440.
11. Keck CM, Muller RH. 2006. Drug nanocrystals of poorly soluble
drugs produced by high pressure homogenisation. Eur J Pharm Bio-
pharm 62(1):3–16.
12. Junghanns JUAH, M¨uller RH. 2008. Nanocrystal technology, drug
delivery and clinical applications. Int J Nanomed 3(3):295–309.
13. Murdande SB, Pikal MJ, Shanker RM, Bogner RH. 2010. Solubility
advantage of amorphous pharmaceuticals: I. A thermodynamic analy-
sis. J Pharm Sci 99(3):1254–1264.
14. Murdande SB, Pikal MJ, Shanker RM, Bogner RH. 2010. Sol-
ubility advantage of amorphous pharmaceuticals: II. Application of
quantitative thermodynamic relationships for prediction of solubility
enhancement in structurally diverse insoluble pharmaceuticals. Pharm
Res 27(12):2704–2714.
15. Liu G, Zhang D, Jiao Y, Guo H, Zheng D, Jia L, Duan C, Liu Y, Tian
X, Shen J, Li C, Zhang Q, Lou H. 2013. In vitro and in vivo evaluation of
riccardin D nanosuspensions with different particle size. Colloids Surf
B 102:620–626.
16. Detroja C, Chavhan S, Sawant K. 2011. Enhanced antihypertensive
activity of candesartan cilexetil nanosuspension: Formulation, charac-
terization and pharmacodynamic study. Sci Pharm 79(3):635–651.
17. Hecq J, Deleers M, Fanara D, Vranckx H, Amighi K. 2005. Prepara-
tion and characterization of nanocrystals for solubility and dissolution
rate enhancement of nifedipine. Int J Pharm 299(1–2):167–177.
18. Schmelzer JW, Schmelzer J Jr. 1999. Kinetics of nucleation at in-
creasing supersaturation. J Colloid Interface Sci 215(2):345–355.
19. Bevernage J, Brouwers J, Brewster ME, Augustijns P. 2013. Eval-
uation of gastrointestinal drug supersaturation and precipitation:
Strategies and issues. Int J Pharm 453(1):25–35.
20. Shah KB, Patel PG, Khairuzzaman A, Bellantone RA. 2014. An
improved method for the characterization of supersaturation and pre-
cipitation of poorly soluble drugs using pulsatile microdialysis (PMD).
Int J Pharm 468(1–2):64–74.
21. Cheow WS, Hadinoto K. 2012. Self-assembled amorphous drug–
polyelectrolyte nanoparticle complex with enhanced dissolution rate
and saturation solubility. J Colloid Interface Sci 367(1):518–526.
22. Van Eerdenbrugh B, Vermant J, Martens JA, Froyen L, Humbeeck
JV, Van den Mooter G, Augustijns P. 2010. Solubility increases asso-
ciated with crystalline drug nanoparticles: Methodologies and signifi-
cance. Mol Pharm 7(5):1858–1870.
23. Dolenc A, Govedarica B, Dreu R, Kocbek P, Srcic S, Kristl J. 2010.
Nanosized particles of orlistat with enhanced in vitro dissolution rate
and lipase inhibition. Int J Pharm 396(1–2):149–155.
24. Sahoo NG, Kakran M, Shaal LA, Li L, Muller RH, Pal M, Tan LP.
2011. Preparation and characterization of quercetin nanocrystals. J
Pharm Sci 100(6):2379–2390.
25. Mitri K, Shegokar R, Gohla S, Anselmi C, Muller RH. 2011. Lutein
nanocrystals as antioxidant formulation for oral and dermal delivery.
Int J Pharm 420(1):141–146.
26. Crisp MT, Tucker CJ, Rogers TL, Williams III, RO, Johnston KP.
2007. Turbidimetric measurement and prediction of dissolution rates
of poorly soluble drug nanocrystals. J Control Release 117(3):351–
359.
27. Peeters K, De Maesschalck R, Bohets H, Vanhoutte K, Nagels L.
2008. In situ dissolution testing using potentiometric sensors. Eur J
Pharm Sci 34(4–5):243–249.
28. Kayaert P, Li B, Jimidar I, Rombaut P, Ahssini F, Van den Mooter
G. 2010. Solution calorimetry as an alternative approach for dissolu-
tion testing of nanosuspensions. Eur J Pharm Biopharm 76(3):507–
513.
29. Liu P, De Wulf O, Laru J, Heikkila T, van Veen B, Kiesvaara J, Hir-
vonen J, Peltonen L, Laaksonen T. 2013. Dissolution studies of poorly
soluble drug nanosuspensions in non-sink conditions. AAPS Pharm Sci
Tech 14(2):748–756.
30. Pardeike J, M¨uller RH. 2010. Nanosuspensions: A promising for-
mulation for the new phospholipase A2 inhibitor PX-18. Int J Pharm
391(1–2):322–329.
31. Xia D, Quan P, Piao H, Piao H, Sun S, Yin Y, Cui F. 2010. Prepara-
tion of stable nitrendipine nanosuspensions using the precipitation–
ultrasonication method for enhancement of dissolution and oral
bioavailability. Eur J Pharm Sci 40(4):325–334.
32. Xu Y, Liu X, Lian R, Zheng S, Yin Z, Lu Y, Wu W. 2012. En-
hanced dissolution and oral bioavailability of aripiprazole nanosuspen-
sions prepared by nanoprecipitation/homogenization based on acid–
base neutralization. Int J Pharm 438(1–2):287–295.
33. Van Eerdenbrugh B, Vercruysse S, Martens JA, Vermant J, Froyen
L, Van Humbeeck J, Van den Mooter G, Augustijns P. 2008. Microcrys-
talline cellulose, a useful alternative for sucrose as a matrix former
during freeze-drying of drug nanosuspensions—A case study with itra-
conazole. Eur J Pharm Biopharm 70(2):590–596.
34. Hu J, Ng WK, Dong Y, Shen S, Tan RBH. 2011. Continuous and scal-
able process for water-redispersible nanoformulation of poorly aqueous
soluble APIs by antisolvent precipitation and spray-drying. Int J Pharm
404(1–2):198–204.
35. M¨uller RH, Peters K. 1998. Nanosuspensions for the formulation of
poorly soluble drugs: I. Preparation by a size-reduction technique. Int
J Pharm 160(2):229–237.
36. Singh SK, Srinivasan KK, Gowthamarajan K, Singare DS, Prakash
D, Gaikwad NB. 2011. Investigation of preparation parameters of
nanosuspension by top-down media milling to improve the dissolution
of poorly water-soluble glyburide. Eur J Pharm Biopharm 78(3):441–
446.
37. Niwa T, Danjo K. 2013. Design of self-dispersible dry nanosus-
pension through wet milling and spray freeze-drying for poorly water-
soluble drugs. Eur J Pharm Sci 50(3–4):272–281.
38. Heng D, Cutler DJ, Chan HK, Yun J, Raper JA. 2008. What is a suit-
able dissolution method for drug nanoparticles? Pharm Res 25(7):1696–
1701.
39. Li W, Yang Y, Tian Y, Xu X, Chen Y, Mu L, Zhang Y, Fang L. 2011.
Preparation and in vitro/in vivo evaluation of revaprazan hydrochloride
nanosuspension. Int J Pharm 408(1–2):157–162.
40. Lai F, Sinico C, Ennas G, Marongiu F, Marongiu G, Fadda AM.
2009. Diclofenac nanosuspensions: Influence of preparation procedure
and crystal form on drug dissolution behaviour. Int J Pharm 373(1–
2):124–132.
DOI 10.1002/jps.24426 Murdande, Shah, and Dave, JOURNAL OF PHARMACEUTICAL SCIENCES