Page 1
Dynamic Article LinksC<Nanoscale
Cite this: Nanoscale, 2012, 4, 7475
www.rsc.org/nanoscale PAPER
Dow
nloa
ded
by S
eoul
Nat
iona
l Uni
vers
ity o
n 22
Nov
embe
r 20
12Pu
blis
hed
on 0
9 O
ctob
er 2
012
on h
ttp://
pubs
.rsc
.org
| do
i:10.
1039
/C2N
R31
775C
View Article Online / Journal Homepage / Table of Contents for this issue
Capacity of mesoporous bioactive glass nanoparticles to deliver therapeuticmolecules†
Ahmed El-Fiqi,abc Tae-Hyun Kim,ab Meeju Kim,ab Mohamed Eltohamy,abc Jong-Eun Won,ab Eun-Jung Leeab
and Hae-Won Kim*abcd
Received 8th July 2012, Accepted 6th October 2012
DOI: 10.1039/c2nr31775c
Inorganic bioactive nanomaterials are attractive for hard tissue regeneration, including
nanocomponents for bone replacement composites and nanovehicles for delivering therapeutics.
Bioactive glass nanoparticles (BGn) have recently gained potential usefulness as bone and tooth
regeneratives. Here we demonstrate the capacity of the BGn with mesopores to load and deliver
therapeutic molecules (drugs and particularly genes). Spherical BGn with sizes of 80–90 nm were
produced to obtain 3–5 nm sized mesopores through a sono-reacted sol–gel process. A simulated body
fluid test of the mesoporous BGn confirmed their excellent apatite forming ability and the cellular
toxicity study demonstrated their good cell viability up to 100 mg ml�1. Small molecules like chemical
drug (Na-ampicillin) and gene (small interfering RNA; siRNA) were introduced as model drugs
considering the mesopore size of the nanoparticles. Moreover, amine-functionalization allowed
switchable surface charge property of the BGn (from �20–30 mV to +20–30 mV). Loading of
ampicillin or siRNA saturated within a few hours (�2 h) and reflected the mesopore structure. While
the ampicillin released relatively rapidly (�12 h), the siRNA continued to release up to 3 days with
almost zero-order kinetics. The siRNA–nanoparticles were easily taken up by the cells, with a
transfection efficiency as high as �80%. The silencing effect of siRNA delivered from the BGn, as
examined by using bcl-2 model gene, showed dramatic down-regulation (�15% of control), suggesting
the potential use of BGn as a new class of nanovehicles for genes. This, in conjunction with other
attractive properties, including size- and mesopore-related high surface area and pore volume, tunable
surface chemistry, apatite-forming ability, good cell viability and the possible ion-related stimulatory
effects, will potentiate the usefulness of the BGn in hard tissue regeneration.
1. Introduction
Regeneration of hard tissues including bone has been signifi-
cantly facilitated by the use of materials developed and engi-
neered to react actively with biological environments where the
bone-associated cells are favored to anchor and differentiate to
secrete bone extracellular matrices (ECMs).1–3 Among this class
of bone-bioactive materials, the group of bioactive glasses (BGs)
has been the most actively and popularly studied. Initially
developed to specific compositions by the melt-quenching, the
BGs take the form of blocks, microfibers and crushed granules in
aDepartment of Nanobiomedical Science and WCU Research Center,Dankook University, South KoreabInstitute of Tissue Regeneration Engineering (ITREN), DankookUniversity, South Korea. E-mail: [email protected] ; Fax: +82 41 5503085; Tel: +82 41 550 3081cGlass Research Department, National Research Center, EgyptdDepartment of Biomaterials Science, School of Dentistry, DankookUniversity, South Korea
† Electronic supplementary information (ESI) available. See DOI:10.1039/c2nr31775c
This journal is ª The Royal Society of Chemistry 2012
micron-size.4–6 A sol–gel technique has spurred the development
of BGs with compositions in a broader range compared to melt-
quenching, enabling the incorporation of bioactive molecules
and allowing nanoscale formulations such as nanofibers and
nanoparticles.7–12
In particular, the BGs developed at the nanoscale (possibly
tens to hundreds of nanometers in tuned dimensions) possess
substantially increased surface properties, accelerating the
possible interactions with other materials/molecules and bio-
logical reactions, which compel their use for either nano-
composites in concert with biopolymers13–15 or delivery systems
incorporating candidate therapeutic molecules.16,17 A nano-
fibrous form of the sol–gel BGs, previously developed by elec-
trospinning and studied as the nanocomponent of the inorganic
phase within biopolymeric matrices, including collagen, poly-
(lactic acid) (PLA) and poly(caprolactone) (PCL), has been
demonstrated to play a significant role in accelerating acellular
mineral formation as well as the in vitro cellular proliferation and
osteogenic differentiation.18–20 Some recent studies have shown
the development of nanoparticulates of BG, with sizes ranging
from tens to hundreds of nanometers.21,22 This class of BG
Nanoscale, 2012, 4, 7475–7488 | 7475
Page 2
Dow
nloa
ded
by S
eoul
Nat
iona
l Uni
vers
ity o
n 22
Nov
embe
r 20
12Pu
blis
hed
on 0
9 O
ctob
er 2
012
on h
ttp://
pubs
.rsc
.org
| do
i:10.
1039
/C2N
R31
775C
View Article Online
nanoparticles has also shown high surface bioreactivity inducing
ionic release and mineral precipitation, which is highly attractive
for fabrication of nanocomposites with the biopolymers.23,24
Furthermore, BG nanoparticles are of potential interest in
dentistry because of their antibacterial properties25,26 and the
ability to remineralize dentine.27
One attractive application of the BG nanoparticles is as a
delivery system for therapeutic molecules. The size of BG
nanoparticles can be modulated to allow intracellular uptake by
endocytosis, while not harming the tissue cells. In tandem with
excellent bone-bioactivity, BG nanoparticle-mediated delivery of
therapeutic molecules will potentiate the capacity of these
nanomaterials for hard tissue regeneration.
To this end, the present study developed spherical BG nano-
particles destined for utilization as delivery vehicles for thera-
peutic molecules. For this we provided a level of mesoporosity to
the nanoparticles using surfactant templates during an ultra-
sound assisted base-catalyzed sol–gel process applied newly in
this study for the nanoparticle preparation. As the model mole-
cules, we applied either chemical drug (antibiotic) or nucleic acid
(small interfering RNA, siRNA) and examined the capacity to
load and release those biomolecules.
Although previous studies have reported the potential of
mesoporous BG in bulk, scaffold or granular form for loading
and delivery of therapeutic molecules, such as drugs and
proteins,16,28 there have been few works using the nanoparticle
form of BG. This nanoparticle form is considered much more
intriguing than other forms as it is possibly taken up by the cells
to carry therapeutic molecules and temporarily release them
inside of cells, directly regulating genetic functions. This allows
more potential use of the BGn in gene delivery systems, which
was demonstrated for the first time in this study. This report on
the utilization of a novel inorganic mesoporous BG nanodelivery
system should provide valuable information for the development
of bioactive delivery systems, particularly for the hard tissue
regeneration.
2. Materials and methods
2.1. Preparation of mesoporous BG nanoparticles
2.1.1. Mesoporous structuring. Tetraethyl orthosilicate
(TEOS, C8H20O4Si, 98%), calcium nitrate tetrahydrate (Ca(N-
O3)2$4H2O, 99%), poly(ethylene glycol) (PEG; (C2H4)nH2O,Mn:
10 000), hexadecetyltrimethyl ammonium bromide (CTAB,
C19H42BrN, $98%), ammonium hydroxide (NH4OH, 28.0%
NH3 in water, $99.99% metal basis), methanol anhydrous
(CH4O, 99.8%), toluene anhydrous (C7H8, 99.8%), 3-amino-
propyl triethoxysilane (APTES, C9H23NO3Si, $98%), and
ampicillin sodium salt (C16H18N3NaO4S) were all purchased
from Sigma-Aldrich and were used as-received without any
further purification.
Mesoporous binary 85SiO2/15CaO (mol%) BG nanoparticles
with well-developed spherical morphology were synthesized by a
novel ultra-sound assisted base-catalyzed sol–gel method using
PEG and CTAB as templates (coded as ‘BGn1’ and ‘BGn2’,
respectively). In a typical synthesis, 5 g PEG for BGn1 or 5 g
CTAB for BGn2 was completely dissolved by stirring in 120 ml
absolute methanol. The solution pH was then adjusted to 12.5 by
7476 | Nanoscale, 2012, 4, 7475–7488
adding about 30 ml NH4OH. To this clear solution, 0.179 g
Ca(NO3)2$4H2O was dissolved while gently stirring. In a sepa-
rate vessel, 0.895 g TEOS was diluted with 30 ml absolute
methanol and then added drop-wise to the vigorously stirred pH
12.5 solution with the simultaneous application of a high-power
ultra-sound using a Sonoreactor, LH700S ultra-sonic generator
(Ulsso Hitech, South Korea) operating at 20 kHz and 700 W.
The output power was 220 W in a 10 s on/10 s off cycle for
20 min. After 24 h of vigorous stirring, the white precipitate was
separated by centrifugation at 5000 rpm for 5 min in aMega17 R
centrifuge (Hanil Science, South Korea), and washed, and re-
dispersed three times with de-ionized water and twice with
absolute ethanol. The final white precipitate was dried at 70 �Cfor 12 h. Finally, organics and nitrates were removed from the
dried powder by calcination at 600 �C under air for 5 h at a
heating rate of 1 �C min�1.
2.1.2. Surface functionalization. The calcined BGn1 and
BGn2 were surface-functionalized with aminopropyl groups
through a post-synthesis procedure. Briefly, 100 mg nanopowder
was dispersed in 50 ml anhydrous toluene and refluxed with 1 ml
APTES at 60 �C by stirring for 6 h. The solution was then
allowed to cool to room temperature and the nanoparticles were
collected by centrifugation at 5000 rpm for 5 min, washed,
re-dispersed three times with toluene, and dried in an oven at
80 �C overnight. The amine-functionalized BGn1 and BGn2
samples were designated as BGn1(A) and BGn2(A), respectively.
2.2. Characterizations of BG nanoparticles
2.2.1. Physicochemical properties. The morphology of the
mesoporous BG nanoparticles was observed by field-emission
scanning electron microscopy (TESCAN,MIRA II LMH, Czech
Republic). The particle size and mesoporous structure of the
nanoparticles were also observed by a high-resolution trans-
mission electron microscope using a JEM-3010 apparatus
(JEOL, Japan).
The phase of the samples was analyzed by X-ray diffraction
(XRD) using an Ultima IV apparatus (Rigaku, Japan) with
CuKa radiation (l ¼ 1.5418 �A). X-ray was generated at 40 mA
and 40 kV, and data were obtained at diffraction angles (2q) from
4–70� with a step size of 0.02� and a scanning speed of 2� min�1.
Small angle X-ray diffraction was also performed using a Rigaku
D/Max-2500 diffractometer in the 2q range of 0.5–5�, to examine
the characteristic of mesopores. Infrared spectra of the samples
were obtained with a resolution of 4 cm�1 on a Varian 640-IR
Fourier transform infrared (FT-IR) spectrometer (Varian, Aus-
tralia) in the range 4000–400 cm�1 using the KBr method.29 The
particle size was also analyzed by a Zetasizer Nano ZS, dynamic
light scattering (DLS) instrument (Malvern Instruments, UK).
Thermogravimetric analysis of the samples was carried out on a
TGA N-1500 instrument (Scinco, South Korea) at a heating rate
of 10 �C min�1 and a nitrogen flux of 40 ml min�1.
The specific surface area, pore volume, and pore size distri-
bution were determined based on N2 adsorption–desorption
measurements. The N2 adsorption–desorption isotherms were
obtained at 77 K using a Quadrasorb SI automated surface area
and pore size analyzer (Quantachrom Instruments, UK).
Samples were degassed under vacuum at 300 �C for 12 h prior to
This journal is ª The Royal Society of Chemistry 2012
Page 3
Dow
nloa
ded
by S
eoul
Nat
iona
l Uni
vers
ity o
n 22
Nov
embe
r 20
12Pu
blis
hed
on 0
9 O
ctob
er 2
012
on h
ttp://
pubs
.rsc
.org
| do
i:10.
1039
/C2N
R31
775C
View Article Online
analysis. The specific surface area was determined according to
the Brunauer–Emmett–Teller (BET) method.30 The pore size
distribution was determined from the N2 desorption branch of
the obtained N2 adsorption–desorption isotherms on the basis of
the density functional theory (DFT) method.31 The total pore
volume was calculated from the amount adsorbed at a maximum
relative pressure (P/P0).
The surface electrical potential and dispersion stability of the
nanoparticles were investigated by zeta (z) potential measure-
ments. The z-potential of the nanoparticles was measured with a
Zetasizer Nano ZS laser Doppler electrophoresis (LDE) instru-
ment (Malvern Instruments, UK). The nanoparticles were
dispersed in deionized water at different pH values. The
z-potential was measured at 25 �C with an applied field strength
of 20 V cm�1 five times (each measurement being the average of
40 runs) and the mean � standard deviation (n ¼ 5) was calcu-
lated. The instrument automatically calculates electrophoretic
mobility (U), and z-potential according to the Helmholtz–
Smoluchowsky equation: z ¼ Uh/3, where z is the zeta potential,
U is the electrophoretic mobility, h is the dispersing medium
viscosity, and 3 is the dielectric constant.
2.2.2. In vitro apatite forming ability. The in vitro hydroxy-
apatite forming ability of the prepared BG nanoparticles was
tested in a simulated body fluid (SBF).32 SBF has an ionic
composition and concentration similar to those of the human
body plasma. Briefly, SBF was prepared by dissolving NaCl,
NaHCO3, KCl, K2HPO4$3H2O, MgCl2$6H2O, CaCl2,
and Na2SO4 in deionized water and buffering at pH 7.4 with
tris(hydroxymethyl) aminomethane (HOCH2)CNH2 and HCl.
The BG nanoparticles were well-dispersed in SBF at a concen-
tration of 1 mg ml�1 in clean, sterile (sterilized with 70% ethanol
and washed with deionized water), tightly stopped 100 ml poly-
ethylene bottles. The bottles were placed inside an incubator at a
controlled temperature of 37 �C for different times up to 28 days,
and the SBF medium was refreshed regularly. At each selected
time point, a bottle was removed from the incubator and the
powder was separated by centrifugation at 5000 rpm for 5 min,
washed, redispersed gently with deionized water and pure
ethanol, and dried at 50 �C overnight. For the next run, the
medium was refreshed. The samples were examined by scanning
electron microscopy (SEM) to detect crystallite formation. The
phase and chemical bond structure of the SBF-treated nano-
particles were examined using XRD and FT-IR, respectively.
2.3. In vitro cytotoxicity assays
The in vitro cytotoxicity of the BGn1(A) and BGn2(A) was
assessed by using different types of cells: HeLa cell line, pre-
osteoblastic MC3T3-E1 cells, and rat bone marrow mesen-
chymal stem cells (rMSCs). The procedures of isolation and
maintenance of rMSCs were described in detail elsewhere.33 Cells
were maintained in Dulbecco’s modified Eagle’s medium
(DMEM; Welgene, Korea) containing 10% fetal bovine serum
(FBS; Gibco, USA) and 1% penicillin–streptomycin under an
atmosphere of 5% CO2 at 37�C. One hundred microliter aliquots
of the cells prepared at a density of 1 � 105 ml�1 were plated in
each well of 96-well plates. After culture for 24 h, the culture
medium was refreshed with that containing the nanoparticles
This journal is ª The Royal Society of Chemistry 2012
prepared at specific concentrations (0, 5, 10, 20, 40, 60, 80 and
100 mg ml�1). After incubation for further 24 h, the cells were
collected and analyzed for the viability using a CCK-8 cell
counting kit (Dojindo, Japan). According to the manufacturer’s
instructions, the reaction medium was added to each well and the
cells were incubated for 4 h at 37 �C, after which 100 ml of each
cell culture supernatant was collected and the optical density was
measured at 450 nm using an iMark microplate reader (BioRad,
USA). Absorbance values were normalized to those of control
free of BG nanoparticles for each test group, and data were
averaged from triplicate samples (n ¼ 3).
2.4. Drug loading and release studies
2.4.1. Sodium ampicillin loading and release. Na-ampicillin
was used as a model drug for the loading and release tests of
antibiotics. Trials were first made to load Na-ampicillin into BG
nanoparticles, lacking surface functionalization. No significant
loading was detected. This may reflect charge effects, with the
negatively charged Na-ampicillin and BG nanoparticles repelling
each other. Therefore, the loading of Na-ampicillin was con-
ducted on aminopropyl surface-functionalized BG nanoparticles,
BGn1(A) and BGn2(A). The amount of loaded Na-ampicillin
was measured by a depletion method, in which the difference in
Na-ampicillin concentration in the loading solution before and
after the loading process was ascertained. Na-ampicillin was
assayed by UV-VIS spectroscopy using a Libra S22 apparatus
(Biochrom, UK) by monitoring the changes in the absorbance at
a characteristic wavelength, 230 nm. A series of standard Na-
ampicillin solutions in deionized water (10–100 mg ml�1) were
prepared to obtain a linear calibration curve (r2 ¼ 0.99) that
obeys the Beer–Lambert law A¼ abc, where A is the absorbance,
a is a constant known as absorptivity coefficient, c is the
concentration, and b is the cell bath length, which is constant.
The loading procedure was as follows. Five milligrams of amine-
functionalized nanoparticles, BGn1(A) or BGn2(A), were
completely dispersed in Na-ampicillin solutions (1–20 mg). These
concentrations were selected to investigate the effect of the initial
Na-ampicillin concentration on the adsorption kinetics. The
solutions were incubated at 37 �C for up to 4 h. At each selected
time point, 1 ml of each solution was withdrawn and centrifuged
at 10 000 rpm for 3 min using a model 5415D centrifuge
(Eppendorf, Germany). The separated clear solution was
analyzed for the Na-ampicillin concentration by UV-VIS spec-
trophotometry. A blank solution of the same pH as the sample
was used as a reference. After determining the optimal time point
for the loading of Na-ampicillin, the same procedure was
repeated for all other solutions. For further release tests, the Na-
ampicillin loaded nanoparticles were washed with distilled water
and dried at 37 �C overnight.
The Na-ampicillin release profile was estimated in vitro using
phosphate buffered saline (PBS, pH 7.4). For the in vitro release
test, 5 mg of each Na-ampicillin loaded sample was dispersed in
10 ml PBS and the solution was incubated at 37 �C for up to 24 h.
At each selected time point, 1 ml of each solution was withdrawn
and centrifuged at 10 000 rpm for 3 min. The clear solution was
analyzed for the Na-ampicillin concentration by UV-VIS spec-
trophotometry using PBS as a blank. The same procedure was
repeated for all other samples.
Nanoscale, 2012, 4, 7475–7488 | 7477
Page 4
Dow
nloa
ded
by S
eoul
Nat
iona
l Uni
vers
ity o
n 22
Nov
embe
r 20
12Pu
blis
hed
on 0
9 O
ctob
er 2
012
on h
ttp://
pubs
.rsc
.org
| do
i:10.
1039
/C2N
R31
775C
View Article Online
2.4.2. siRNA loading and release. The siRNA loading test
onto the nanoparticles of either BGn1(A) or BGn2(A) was per-
formed using fluorescein isothiocyanate (FITC)-conjugated
siRNA for the nanoparticle detection by fluorescence. 10 mg of
FITC-conjugated siRNA was mixed with varying quantities (10,
20, 50 and 100 mg) of BG nanoparticles (resultant
RNA : nanoparticles ¼ 1 : 1, 1 : 2, 1 : 5 and 1 : 10) within 1 ml
release medium (RNase-free TE buffer) to find out an optimal
mixing ratio for high loading efficiency of siRNA. Tests gave a
ratio of 1 : 5 with the highest loading. Next, 10 mg of siRNA was
mixed with 50 mg of BG nanoparticles for different times (30 min,
1 h, and 2 h) at room temperature. The highest loading quantity
was attained at 2 h. For the loading quantity assay, the siRNA–
nanoparticle mixture was centrifuged at 10 000 rpm at 4 �C, andthe supernatant was collected and the siRNA quantity was
assessed by using a SpectraMax M2e multi-detection microplate
reader (Molecular Devices, USA) at an absorbance of 210 nm.
Subtraction of the quantity from the initial amount of siRNA
resulted in the loaded amount of siRNA onto the nanoparticles.
For the siRNA release study, the siRNA-loaded nanoparticles
were pooled into 10 ml of release medium (RNase-free TE buffer)
and stored at 37 �C for various time points (up to a saturation
point). The supernatant was collected by centrifugation at 10 000
rpm and then assessed using a multi-detection microplate reader.
2.4.3. siRNA delivery to cells. Intracellular uptake experi-
ments of the siRNA–nanoparticles as well as the functional
activity of gene silencing were performed using BGn2(A) as the
representative sample as this type presented higher loading
capacity of siRNA. FITC-conjugated AccuTarget Validated
siRNA specific to human bcl-2 mRNA was designed (Bioneer,
South Korea). As a negative control, the same nucleotides were
scrambled to form a non-genomic combination and conjugated
to FITC (scRNA). As the siRNA transfection mixture, 10 mg of
BGn2(A) loaded with siRNA (either bcl-2–siRNA or scrambled
RNA; scRNA) in Opti-MEM (Invitrogen, USA) was used. HeLa
cells were seeded (1 � 105 cells) in wells of 6-well plates and the
transfection mixture was added to each well and incubated for
4 h. As the negative control, 1 mg ml�1 of free siRNA (i.e.,
without loading onto the nanoparticles) was also used for the
transfection. After the siRNA transfection, the cells were har-
vested and used for the following experiments. For the quanti-
fication of bcl-2 mRNA expression, the transfected HeLa cells
were cultured in a-MEM and harvested 24 h later.
The harvested cells were fixed with 4% paraformaldehyde
solution for 30 min on a coating slide glass, which was followed
by washing with cold (4 �C) PBS and mounting on a cover glass.
The fluorescent images were observed and analyzed by confocal
laser scanning microscopy (CLSM; model LSM 510, Carl Zeiss,
Germany). Cells were counterstained with propidium iodide (PI;
Invitrogen, USA) to observe the nucleus of the cells. For the
quantification of the transfection efficiency, cells were analyzed
using a fluorescence activated cell sorting (FACS) Calibur flow
cytometer (BD Biosciences, USA). The data acquired for 10 000
cells in each sample were analyzed using the CellQuest Pro
software (BD Biosciences, USA).
For examination of the ultrastructure of cells, the transfected
cells with bcl-2 siRNA loaded BGn2(A) were harvested and fixed
in a mixture of 2% paraformaldehyde and 2.5% glutaraldehyde
7478 | Nanoscale, 2012, 4, 7475–7488
with 0.2 M PBS for 8 h at pH 7.2. Post-fixation was conducted
using 1% osmium tetroxide in PBS for 2 h. Subsequently, the
fixed cells were dehydrated in ethanol of ascending concentra-
tions (70%, 80%, 90%, 95%, and 100%) and embedded in EMbed
812 resin (EMS, USA) via propylene oxide. Ultrathin sections
were obtained using an ultramicrotome (Leica, USA) and were
double-stained with uranyl acetate and lead citrate. The stained
sections on the grids were then examined with a H7000 TEM
(Hitachi, Japan) operating at an acceleration voltage of 80 kV.
The expression of bcl-2 in the cells treated with bcl-2 siRNA
loaded BGn2(A) was confirmed by quantitative real-time RT-
PCR. The first strand cDNA was synthesized from the total
RNA (2 mg) using a SuperScript first strand synthesis system for
real-time PCR (Invitrogen, USA) according to the manufactur-
er’s instructions. The reaction mixture was made up to 50 ml.
Real-time PCR was conducted using SYBR GreenER qPCR
SuperMix reagents (Invitrogen, USA) and a Bio-Rad iCycler.
The relative transcript quantities were calculated using the DDCt
method with glyceraldehyde 3-phosphate dehydrogenase
(GAPDH) as the endogenous reference gene amplified from the
samples.
2.4.4. Statistical analysis.Datawere expressed as themeans�one standard deviations. Statistical analysis was carried out by
one-way analysis of variance (ANOVA) and p-values <0.05 were
considered significant.
3. Results
3.1. Properties of mesoporous BG nanoparticles
The morphology of the mesoporous BG nanoparticles (BGn1
and BGn2) prepared via a sol–gel route using different templates
(PEG and CTAB) was observed by high resolution SEM and
TEM. Spherical nanoparticles mono-dispersed with sizes
<100 nm were abundant in SEM examination (Fig. 1, panels a1
and b1). TEM images show that the nanoparticles contained
mesopores throughout their inner structure (Fig. 1, panels a2, a3,
b2, and b3). The size of the nanoparticles was measured from
arbitrarily selected TEM images (Fig. 1c); the average size of the
BGn1 and BGn2 nanoparticles was approximately 85 � 15 nm
and 95 � 15 nm, respectively. DLS instrumental analysis showed
that the particle sizes in ranges slightly increased, 105 nm for
BGn1 and 115 nm for BGn2 (Fig. 1d), possibly due to the
aggregation effect of nanoparticles.34 The XRD patterns
confirmed the amorphous phase for both BG nanoparticles
(Fig. 1e). Small angle XRD patterns (inset in Fig. 1e) revealed a
single peak (2q ¼ 2.959�) for BGn2 whilst no peaks for BGn1.
Results demonstrate that the mesopores in BGn2 are disordered
but uniform-sized; on the other hand, those in BGn1 are not
uniform-sized.16,28,35,36 The chemical compositions of BG nano-
particles were confirmed by EDS analyses, which showed Si and
Ca peaks at similar atomic ratios to the glass stoichiometry
(Fig. 1f).
The pore characteristics of the nanoparticles were evaluated
from the N2 adsorption–desorption isotherms by the BET
method. Fig. 2 shows the N2 adsorption–desorption isotherms
and the corresponding pore size distributions (inset graphs) of
BGn1 and BGn2. The N2 isotherms exhibited characteristics of
This journal is ª The Royal Society of Chemistry 2012
Page 5
Fig. 1 (a and b) Nano-structural morphologies of the mesoporous BG nanoparticles (BGn1 and BGn2) developed by a sol–gel method; (a1–a3) BGn1
and (b1–b3) BGn2; (a1 and b1) SEM and (a2, a3, b2 and b3) TEM images. Spherical nanoparticles were mono-dispersed (SEM images) and contained a
large amount of mesopores inside the structure (TEM images). (c and d) Particle size distributions of the BGn1 and BGn2, as measured from the TEM
images arbitrarily chosen (c); average sizes were 85 nm (�15 nm) and 95 nm (�15 nm), respectively, for BGn1 and BGn2, as well as from the DLS
measurement (d), which gave slightly increased sizes (105 nm for BGn1 and 115 for BGn2). (e) XRD patterns showing the amorphous state of the
mesoporous BG nanoparticles, moreover, small angle XRD patterns in the inset revealed one peak at 2q ¼ 2.959� for BGn2, a characteristic of
disordered but uniform-sized mesopores, whilst no peaks were noticed for BGn1. (f) EDX analysis confirming the glass composition (BGn1 repre-
sentatively shown).
Dow
nloa
ded
by S
eoul
Nat
iona
l Uni
vers
ity o
n 22
Nov
embe
r 20
12Pu
blis
hed
on 0
9 O
ctob
er 2
012
on h
ttp://
pubs
.rsc
.org
| do
i:10.
1039
/C2N
R31
775C
View Article Online
the type IV isotherm associated with mesoporous materials
according to the IUPAC (international union of pure and
applied chemistry) classification. BGn1 exhibited a small
hysteresis loop of the desorption branch, indicative of the exis-
tence of large pores. BGn2 showed a little hysteresis loop,
This journal is ª The Royal Society of Chemistry 2012
indicating that the adsorption process was completely reversible.
The pore size distribution in BGn1 showed multi-sized pores
with broad distribution and relatively low pore volume, whilst
BGn2 showed mono-sized pores with narrow distribution and
large pore volume.
Nanoscale, 2012, 4, 7475–7488 | 7479
Page 6
Fig. 2 Nitrogen adsorption–desorption isotherms of the BG nanoparticles after the amine functionalization; (a) BGn1 and (b) BGn2. The pore size
distribution is shown in the insets of both graphs, as determined by the DFT method. Both isotherms exhibited a type IV isotherm characteristic of
mesoporous materials. While the pore size distribution in BGn1 was relatively broad with diffused pore sizes towards large pores, that in BGn2 was
much narrower with a sharp peak at �3 nm. The average pore size was 4.9 and 3.2 nm, respectively, for BGn1 and BGn2.
Table 1 Summary of the mesopore structures of BG nanoparticles,including pore size, specific surface area and pore volume, as determinedfrom N2 adsorption–desorption isotherms
Parameter BGn1 BGn2
Pore size (nm) 4.9 3.2Specific surface area (m2 g�1) 54 830Specific pore volume (cm3 g�1) 0.133 0.415
Fig. 3 Characteristics of the mesoporous BG nanoparticles. (a) FT-IR spect
functionalization. FT-IR results showed that the aminated nanoparticles, B
TGA curves showed weight losses at 250–580 �C, which is ascribed to the a
potentials of the nanoparticles at pH 7.0 changed from highly negative (�29
7480 | Nanoscale, 2012, 4, 7475–7488
Dow
nloa
ded
by S
eoul
Nat
iona
l Uni
vers
ity o
n 22
Nov
embe
r 20
12Pu
blis
hed
on 0
9 O
ctob
er 2
012
on h
ttp://
pubs
.rsc
.org
| do
i:10.
1039
/C2N
R31
775C
View Article Online
The average pore size of BGn1 and BGn2 was 4.9 nm and 3.2
nm, respectively, as determined from the DFT method (Table 1).
The pore volume of BGn1 and BGn2 was 0.133 cm3 g�1 and
0.415 cm3 g�1, respectively, and the corresponding surface area
of BGn1 and BGn2 was 54 m2 g�1 and 830 m2 g�1, respectively,
as determined on the basis of the BET approximation. Conclu-
sively, BGn2 showed a greater level of pores with more uniform
pore size and higher surface area than BGn1. These mesopore
ra, (b) TGA and (c) zeta-potential of the samples before and after amine
Gn1(A) and BGn2(B), revealed characteristic N–H vibration bands (a).
mine groups present in the nanoparticles (b). After amination, the zeta-
mV and �19.5 mV) to highly positive (+28.5 mV and +22.5 mV) (c).
This journal is ª The Royal Society of Chemistry 2012
Page 7
Fig. 4 Cytotoxicity tests of the BGn1(A) and BGn2(A) using different
types of cells including (a) bone marrow MSCs, (b) pre-osteoblast
MC3T3-E1 cells, and (c) HeLa cells. Varying concentrations (0, 5, 10, 20,
40, 60, 80 and 100 mg ml�1) of either BGn1(A) or BGn2(A) pooled in cell
growth medium were applied to the cells for 24 h and the cell viability was
assessed by an MTS assay. Data were normalized to those control media
without the addition of nanoparticles. Data are shown as means and
standard deviations from triplicate samples.
Dow
nloa
ded
by S
eoul
Nat
iona
l Uni
vers
ity o
n 22
Nov
embe
r 20
12Pu
blis
hed
on 0
9 O
ctob
er 2
012
on h
ttp://
pubs
.rsc
.org
| do
i:10.
1039
/C2N
R31
775C
View Article Online
configurations of the BG nanoparticles are of special importance
in loading and releasing therapeutic molecules.
The surface of the mesoporous BG nanoparticles was modified
with amine groups, which can impart a positive charge to the
surface. The prepared nanoparticles had abundant OH groups
on their surface, conferring a net negative charge. We aminated
BGn1 and BGn2 nanoparticles with APTES to allow the meso-
porous surface to contain NH2 functional groups, which are
considered highly useful for the loading of negatively charged
molecules including nucleic acids. FT-IR spectra of the aminated
nanoparticles, BGn1(A) and BGn2(A), revealed N–H vibration
bands characteristic of NH2 functional groups (Fig. 3a). Fig. 3b
shows the TGA curves of the aminated and non-aminated BG
nanoparticles. For BG1, there was only a 4% weight loss before
approximately 250 �C, being mainly associated with the loss of
water molecules (from about 30–120 �C) adsorbed onto the
surface of the nanoparticles and the condensation of surface
silanol groups (from 120–250 �C). In the range of 250–550 �C,only 1% weight loss was recorded, which may be due to the
residual PEG template, after which the weight became stable
above 550 �C. As a result of its large surface area, BG2 showed a
15% weight loss before 250 �C, and in the range of 250–550 �Conly 2% weight loss was recorded, which might also be from the
residual CTAB template, after which the weight stabilized above
550 �C. For aminated BG nanoparticles, the TGA curves showed
different weight loss behavior compared to non-aminated ones.
BG1(A) showed 10% weight loss between 250 �C and 550 �C and
BG2(A) showed a similar weight loss step with 13% weight loss
between the region, which was attributed to the decomposition of
NH2 functional groups.37
After the amination, the surface charge of the nanoparticles
was totally changed. Fig. 3c shows the z-potential of the BG
nanoparticles with variation in pH. At pH 7.0 the z-potentials for
BGn1 and BGn2 before amination were highly negative, being
�29.0 mV and �19.5 mV, respectively. However, after amina-
tion, the z-potentials of the nanoparticles became positive within
the pH range measured (from pH 3.0–9.0). In particular, at
pH 7.0, the z-potentials were highly positive, +28.5 mV and
+22.5 mV, for BGn1(A) and BGn2(A), respectively. The results
obtained by FT-IR and z-potential analyses demonstrate that the
amination process was properly complemented to provide posi-
tively charged amine groups to the surface of BG nanoparticles.
The mesoporous BG nanoparticles, with their ultra-high
specific surface area resulting from the small particle size and
inner mesopores, should have excellent bone bioactivity. As a
first index, we investigated the hydroxyapatite forming ability in
vitro. Samples were incubated in SBF for periods up to 28 days,
and the development of crystalline phases, changes in nano-
morphology, and chemical structures were observed (ESI 1†).
The XRD phase change of the BGn2(A) samples after immersion
in SBF showed the evolution of typical hydroxyapatite peaks,
and the intensity at the main peak 2q ¼ 32� increased with
increasing immersion time (ESI 1a†). FT-IR spectra also showed
the development of bands related to carbonated hydroxyapatite
(PO4 bands at 565, 605 and 964 cm�1 and CO3 bands at 1420 and
1458 cm�1) (ESI 1b†). The high resolution field emission-SEM
images show the changes on the surface of the BGn2(A). During
SBF incubation as short as 1 day, some crystallites started to
form (ESI 1†, panel c1). With prolonged immersion for 7 and
This journal is ª The Royal Society of Chemistry 2012
14 days (ESI 1†, panels c2 and c3, respectively), the formation of
hydroxyapatite crystallites was pronounced, to the point where
the spherical shape of the nanoparticles completely disintegrated.
High resolution TEM images taken at 1 and 7 days also sup-
ported the development of hydroxyapatite crystallites on the
surface of the BG nanoparticles (ESI 1†, panels c4 and c5).
Comparing the two types of BG nanoparticles, no significant
differences in XRD and FT-IR could be found.
To utilize the BG nanoparticles as a delivery system for
biomolecules, their cellular toxicity needed to be first clarified.
We assessed the cytotoxicity of the nanoparticles using several
Nanoscale, 2012, 4, 7475–7488 | 7481
Page 8
Dow
nloa
ded
by S
eoul
Nat
iona
l Uni
vers
ity o
n 22
Nov
embe
r 20
12Pu
blis
hed
on 0
9 O
ctob
er 2
012
on h
ttp://
pubs
.rsc
.org
| do
i:10.
1039
/C2N
R31
775C
View Article Online
types of cells including rat bone marrow mesenchymal stem cells
(rMSCs), pre-osteoblastic MC3T3-E1 cells and HeLa cells. Fig. 4
shows the relative cell viability in response to varying concen-
trations (0, 5, 10, 20, 40, 60, 80 and 100 mg ml�1) of the BGn1(A)
and BGn2(A) during culture for 24 h. In rMSCs, the cell viability
appeared to be dose-dependent for both types of nanoparticles;
the cell viability up-regulated at low doses (<20 mg ml�1) was
reduced at higher doses. The up-regulation at low doses
(particularly in BGn2(A)) was more apparent in MC3T3-E1
cells. In the HeLa cell line, the viability was largely maintained at
all the doses used for both types of nanoparticles. In particular,
the cell viability at a dose of 10 mg ml�1, which was chosen for
further drug and gene delivery tests, was preserved well (equal to
or even higher than the control).
3.2. Capacity to load and deliver therapeutic molecules
As a first model therapeutic molecule, we selected the hydrophilic
antibiotic, Na-ampicillin. First, drug loading tests were per-
formed. Ampicillin loading was not possible with the BGn1 or
BGn2 nanoparticles without amine-functionalization. Conversely,
Fig. 5 (a and b) Na-ampicillin loading tests onto the aminated-BG nanoparti
presented as the cumulative total amount or the relative amount to the initial
different times to observe the optimal loading time (a); based on this, the load
by varying the initial concentration of ampicillin, resulting in about 180 and 3
the higher pore volume and surface area of BGn2 allowed better drug loadin
recorded in cumulative after testing in PBS for different time points; drug relea
d). Three replicate samples were tested for each condition.
7482 | Nanoscale, 2012, 4, 7475–7488
when loaded onto the amine-functionalized BGn1(A) or
BGn2(A), we could detect the loading quantities with increasing
time of loading. The ampicillin loading was shown to saturate at
approximately 120 min for both types of nanoparticles, which was
chosen as the optimal loading time (Fig. 5a). After the loading, the
mesopore structure of the BG nanoparticles was observed to
change (shown in ESI 2†). Surface areas and pore volumes were
reduced significantly, and the pore sizes were also decreased. In
particular, the decrease in the pore volume of BGn1(A) and
BGn2(A) caused by the ampicillin loading was as high as �30 to
40%, demonstrating the incorporation of ampicillin molecules
within the mesopore structure (summarized in ESI 3†). Further-
more, we sought to find the loading capacity by obtaining an
adsorption isotherm. The initial concentrations of ampicillin were
varied and the corresponding loaded amounts were measured
(Fig. 5b); the maximal loading was obtained at approximately 180
and 300 mg of ampicillin per gram of BGn1(A) and BGn2(A),
respectively, both of which were reached at the maximal concen-
tration of ampicillin that has been tested (500 mg ml�1). It can be
deduced that the high pore volume and surface area of BGn2(A)
allowed enhanced loading capacity of ampicillin. The release
cles, BGn1(A) and BGn2(A) and (c and d) the release profiles for the drug
loading. Na-ampicillin was loaded onto either BGn1(A) or BGn2(A) for
ing time was determined at 120 min. The adsorption isotherm was plotted
00 mg of ampicillin per gram of BGn1(A) and BGn2(A), respectively (b);
g capacity. The release profiles of ampicillin from the nanoparticles were
se was almost completed at�12 h for both BGn1(A) and BGn2(A) (c and
This journal is ª The Royal Society of Chemistry 2012
Page 9
Fig. 6 Loading quantity of siRNA onto BGn1(A) and BGn2(A); 10 mg
of siRNA, initially used onto 50 mg of nanoparticles with varying loading
time, provided a loading saturation of 1.93 mg (�19% of initial loading)
for BGn1(A) and 2.59 mg (�26% of initial loading) for BGn2(A) attained
at 2 h. The loading efficiency (maximal quantity of siRNAwith respect to
nanoparticles) was thus 3.86% for BGn1(A) and 5.18% for BGn2(A),
demonstrating significantly higher loading efficiency in BGn2(A) than in
BGn1(A).
Dow
nloa
ded
by S
eoul
Nat
iona
l Uni
vers
ity o
n 22
Nov
embe
r 20
12Pu
blis
hed
on 0
9 O
ctob
er 2
012
on h
ttp://
pubs
.rsc
.org
| do
i:10.
1039
/C2N
R31
775C
View Article Online
profiles of ampicillin from the nanoparticles were further recorded
in a cumulative manner after incubation in PBS for different time
points (Fig. 5c and d). The drug release pattern showed an initial
linear profile (up to a few hours), a further slow-down in the
release rate, and then final saturation at approximately 12 h.
Although the release patterns of BGn1(A) and BGn2(A) were
similar, the release amount of ampicillin with time was substan-
tially different because the loading amount was initially different.
Next, we sought to find the potential of the BG nanoparticles
in loading and delivery of genes. We selected siRNA as a model
Fig. 7 TEM ultrastructure of the delivered BG nanoparticles (BGn2(A) sh
particles taken up by the cells, distributed in cytoplasm around the cellular c
cells) are shown for comparison.
This journal is ª The Royal Society of Chemistry 2012
gene, as this has been one of the most fascinating genetic mole-
cules used in many diseases and injuries of target tissues.38,39 For
the loading tests, we used standard siRNAwithout endowing any
genetic functions to the sequence, but we conjugated FITC to
assess the fluorescence intensity and further to detect the intra-
cellular process of the siRNA–nanoparticle complex. First, we
determined the maximal loading quantity of siRNA onto the
nanoparticles and the proper loading time. A preliminary test
revealed that the mixing ratio of siRNA–nanoparticles affected
the loading quantity of siRNA (not shown here), and the optimal
loading was determined at the mixing ratio of 1 : 5 (siR-
NA : nanoparticles). We further changed the time of complexa-
tion at a ratio of 1 : 5. Amaximal loading quantity of 1.93 mg and
2.59 mg siRNA per 50 mg BGn1(A) and BGn2(A), respectively,
was found to be attained at approximately 2 h, after which a
loading saturation was observed, indicating that BGn1(A) and
BGn2(A) had a loading capacity of siRNA of approximately
3.86% and 5.18%, respectively (�1.3 times higher in BGn2(A))
(Fig. 6). The loading conditions were thus determined at a ratio
of siRNA : nanoparticles 1 : 5 and a loading time 2 h for further
gene delivery and cell studies.
The typical ultrastructure of the cells that received the deliv-
ered siRNA–BGn2(A) complex was observed by TEM (Fig. 7).
Many electron dense nanoparticles with sizes <100 nm were
readily apparent in the cytoplasm around intracellular compo-
nents, including the Golgi apparatus and the rough endoplasmic
reticulum. There was no appearance of cellular stressing associ-
ated with the intracellular uptake of the nanomaterials, such as
signs of apoptosis and necrosis.
The intracellular uptake of the siRNA–BGn2(A) complex was
analyzed by CLSM and FACS (Fig. 8). Both scramble RNA
(scRNA) and bcl-2 target gene siRNA complex with BGn2(A)
own representatively) within HeLa cells showing electron dense nano-
omponents. Cell images without the treatment of nanoparticles (control
Nanoscale, 2012, 4, 7475–7488 | 7483
Page 10
Dow
nloa
ded
by S
eoul
Nat
iona
l Uni
vers
ity o
n 22
Nov
embe
r 20
12Pu
blis
hed
on 0
9 O
ctob
er 2
012
on h
ttp://
pubs
.rsc
.org
| do
i:10.
1039
/C2N
R31
775C
View Article Online
were used for the experimental groups. On the other hand,
control groups without any addition (w/o), siRNA only, and
BGn2(A) only were also tested for comparison purposes. For the
experimental groups, a large number of cells were stained posi-
tive (green) for fluorescence FITC dye, as well as co-stained
positive (red) for the nucleus-specific dye PI (Fig. 8a). This was
not readily observed in other control groups; w/o, siRNA only,
or BGn2(A) only. The fraction of the cells that were positive for
the FITC from the whole cell mass was examined by FACS
analysis. Compared to the cells in control groups, where no
positive fractions were noticed, those with the siRNA–BGn2(A)
or scRNA–BGn2(A) complex presented as high as approxi-
mately 80% of FITC-positive cells (Fig. 8b), confirming the high
intracellular delivery capacity of the complexes.
We further sought to find the functional genetic effects of the
delivered siRNA–BGn2(A). Prior to this, we investigated the
Fig. 8 (a) FITC-conjugated siRNA loaded within BGn2(A) was transfected
BGn2(A) complex in green as well as co-staining of nuclei with PI in red; in
siRNA only, or with BGn2(A) only, there was no green fluorescence (only nucl
significant transfection of siRNA–BGn2(A) or scRNA–BGn2(A) delivery int
7484 | Nanoscale, 2012, 4, 7475–7488
release profile of the siRNA from the nanoparticles. The siRNA
release in PBS exhibited a pattern of almost linear increase for up
to 3 days (Fig. 9a). When the release amount was converted to
the percentage of initial loading amount the total release corre-
sponded to approximately 45%. The release profile of siRNA for
up to 3 days indicated the potential value of the BGn2(A) system
in the delivery of siRNA and its possible genetic regulation at
least for up to those periods, i.e., the temporal expression of
target genes and further silencing effects for a certain period was
expected. To elucidate the functional activity of the siRNA, we
designed a simple well-known target gene siRNA that silences
bcl-2 gene; although not specified for bone cells, it is considered
proper as a pilot study to find out functional gene-silencing
efficacy of siRNA when delivered from the mesoporous BG
nano-vehicle. The silencing effect of the siRNA delivery by
the designed system is shown in Fig. 9b. A significant effect
into HeLa cells, which was visualized under CLSM to reveal the siRNA–
contrast, other control cells, including those without any additions, with
ei with PI in red). (b) FACS analysis of the cells positive for FITC showing
o the cells (approximately 80%), in direct contrast to other cases (<1%).
This journal is ª The Royal Society of Chemistry 2012
Page 11
Fig. 9 (a) Release profile of siRNA from the BGn2(A) delivery system, showing a continual release (zero-order kinetics) for up to 3 days with a final
quantity of approximately 45% of the initial loading and then almost saturation. (b) Assay on biological activity of the siRNA delivered from the
BGn2(A) nanocarrier, designed to silence bcl-2 gene, demonstrated clearly the silencing effect down to approximately 15% maintained, in contrast to
other control cases (either siRNA only, approximately 89%; BGn2(A) only, approximately 79%; or scramble siRNA–BGn2(A), approximately 90%
maintained). Data are shown as means and standard deviations from triplicate samples.
Dow
nloa
ded
by S
eoul
Nat
iona
l Uni
vers
ity o
n 22
Nov
embe
r 20
12Pu
blis
hed
on 0
9 O
ctob
er 2
012
on h
ttp://
pubs
.rsc
.org
| do
i:10.
1039
/C2N
R31
775C
View Article Online
(down-regulation to 15%) was evident, when compared to other
cases used as controls (siRNA only; 89%, BGn2(A) only; 79%, or
scramble siRNA–BGn2(A); 90% remained).
4. Discussion
Development of proper delivery systems for the regeneration of
tissues including bone is of special importance to maximize the
therapeutic efficacy of bioactive factors by delivering them in a
timely and sustainable way.40,41 The currently introduced BG
nanoparticles, especially those retaining the large volume of
mesopores within the structure, satisfy some promising aspects
targeting bone tissue, as their composition has already been
proven as a suitable platform for novel inorganic bioactive
nanomaterials.42,43 Endowing a capacity to deliver therapeutic
biomolecules improves the bone regenerative potential of the
bioactive nanomaterials, which seeks to target and modulate
specific cellular behaviors.40,41
Here, we developed the BG nanoparticles into spherical mes-
oporous structures using two different templates: PEG as a non-
ionic surfactant and CTAB as a cationic surfactant. The char-
acteristics of mesopores, including pore size, volume, and surface
area, are very important in determining the capacity of the
nanoparticles to incorporate and release candidate therapeutic
molecules. The use of two different templates to generate mes-
opores within the BG nanoparticles resulted in mesopore volume
(PV) and surface area (SA) with different levels depending on the
template type. The use of CTAB produced significantly higher
levels of PV and SA than those of PEG. This difference originates
from the differences in chemical structure and properties of
CTAB and PEG in the micelle formation, such as chemical
structure and size, and the maximum number of surfactant
molecules per micelle. The large quantity of mesopores, espe-
cially in the case of BGn2(A), is considered highly beneficial for
providing enough space to home small molecules in large
amounts, including chemical drugs and genes.43 In fact, the
effectiveness of the BG nanoparticles with such a mesoporous
structure in the loading of biomolecules can be presumed when
This journal is ª The Royal Society of Chemistry 2012
referenced from the accumulating studies on pure mesoporous
silica nanoparticles (MSNs).44–46 MSNs generally developed by
means of CTAB as surfactants present typical values of large
PV in the range 0.5–1 cm3 g�1 and high SA of approximately
900 m2 g�1.45,47 A very close range (PV: 0.415 cm3 g�1 and SA:
830 m2 g�1) was obtained for our BGn2 templated with the same
surfactant. Along with the total space volume and area of pores,
the size of individual pores is also crucial, determining the size of
biomolecules that can be allowed to enter into the mesopores.
Both BGn1 and BGn2 showed average pore sizes of 4.9 and
3.2 nm, respectively, values considered to facilitate loading of
relatively small molecules like chemical drugs and small sized
nucleic acids.48–52 Therefore, here we employed candidate (rela-
tively small) biomolecules of antibiotic and siRNA. In fact, the
mesopore size can be increased by using different types of
templates and with the help of some auxiliary organic molecules,
as have been well-studied in the case of MSNs.48,53 Thus, when
this will be exploited, large molecules such as growth factors can
also be considered as the candidate molecules to be delivered by
the engineered mesoporous BG nanoparticles, which remain as a
further interesting study.
Another factor to consider in tailoring the mesopores of the
BGn is the surface chemistry, particularly the surface charge, as
this can directly affect the nanoparticle–biomolecule interac-
tions.54,55 Biomolecules such as chemical drugs, proteins, and
genes can have a certain level of surface net charge under bio-
logical fluid conditions (generally at pH �7); drugs can be either
ionic or non-ionic, proteins are dependent on the sequence of
amino acids comprising the whole structure, and genes are highly
negatively charged due to the existence of a bunch of phosphate
groups.38 Here, the mesoporous BG nanoparticles initially
synthesized without surface treatment possessed a relatively high
negative charge; e.g., at pH 7, the z-potential of BGn1 and BGn2
before amination is in the range from �20 to �30 mV, which is
due to the presence of a high concentration of hydroxyl groups
on the surface of nanoparticles and, thus, it is considered
appropriate to afford chemically relevant sites for the ionic
interaction with positively charged drugs or biomolecules. For
Nanoscale, 2012, 4, 7475–7488 | 7485
Page 12
Dow
nloa
ded
by S
eoul
Nat
iona
l Uni
vers
ity o
n 22
Nov
embe
r 20
12Pu
blis
hed
on 0
9 O
ctob
er 2
012
on h
ttp://
pubs
.rsc
.org
| do
i:10.
1039
/C2N
R31
775C
View Article Online
negatively charged drugs or biomolecules (e.g., Na-ampicillin
and siRNA), we aminated the surface with APTES. The func-
tionalization clearly showed the presence of amine groups on the
surface (from FT-IR spectra and TGA weight loss), and
dramatically changed the z-potential to high positive values (+20
to +30 mV).
We next sought to utilize the aminated BGn in the loading and
delivery of biomolecules. As the candidate molecules, we used
either chemical drug (anionic drug Na-ampicillin) or small
nucleic acid (siRNA). At first guess, the loading of siRNA was
presumed to be more effective than the ionic Na-ampicillin; while
the siRNA designed here with a base pair of �20 (20 bp) was
highly negatively charged, the Na-ampicillin is an amphoteric,
namely it forms dissociated ionic species in solution, either in
cationic, zwitterionic, and anionic form depending on the solu-
tion pH.56
Loading of Na-ampicillin within the nanoparticles was shown
to be attained completely at around 120 min. The loading
capacity of the nanoparticles showed a marked difference
between the two types: approximately 180 mg g�1 and 300 mg
g�1 of BGn1 and BGn2, respectively, with approximately
1.7 times higher in BGn2. This was primarily due to the differ-
ence in the total pore volume. Taken from the BET results, the
mesopores in BGn2 provided a large pore volume and high
surface area compared to those in BGn1 (0.415 cm3 g�1, 830 m2
g�1 vs. 0.131 cm3 g�1, 54 m2 g�1, respectively). The average pore
sizes of the BGn (4.9 nm and 3.2 nm) are much larger than the
molecular width of ampicillin (0.77 nm) reported.57 Thus we can
imagine the presence of ampicillin molecules inside the meso-
pores of both nanoparticles, presumably adsorbed onto the
mesopore channel surface. In fact, after the ampicillin loading,
the pore volumes of the BGn were significantly reduced (30–40%,
in ESI 3†), suggesting that a large number of ampicillin molecules
have filled the mesopore channel. Interestingly, although the
pore space of both nanoparticles shows big differences (3-times
in pore volume and 15-times in surface area), the difference in
ampicillin loading capacity was relatively small (1.7-times),
falling short of our expectation. A more detailed examination of
the pore size distribution demonstrated a broader pore size
distribution in BGn1 than in BGn2, particularly towards much
larger mesopores than the average value, which should be more
sufficient to incorporate the drug. The results explain well the
importance of the mesopore properties, including the size and
distribution of pores, and their volume and surface area, in
determining the loading capacity of drugs within the mesoporous
BGn. Although here we utilized the positive-charged mesopore
surface in capturing negatively charged drugs, cationic or
hydrophobic non-charged drugs can also be applied if the surface
of the nanoparticles is functionalized relevantly, namely using
hydroxylated nanoparticles (as-prepared without modification)
or after the carboxylation process for cationic drugs, and after
tailoring with alkyl groups for hydrophobic drugs.58
The loaded ampicillin was released over 12 h, presenting a
typical profile dominated by a diffusion process. This was simi-
larly observed for both types of nanoparticles, although the
BGn2 released more ampicillin, reflecting the higher loading
quantity. The release of ampicillin was not complete (80–90%
release). Possibilities for the retention of the drug are loss or
degradation of the drug during the test or the actual retention of
7486 | Nanoscale, 2012, 4, 7475–7488
10–20% of ampicillin within the mesopores. Loss/degradation is
more likely, but retention due to bonds between ampicillin and
the mesopore surface that are essentially irreversible is very
unlikely. In fact, although a somewhat sustainable release profile
was apparent over 12 h, the mesoporous BG nanoparticles are
not considered effective in retaining the ampicillin within the
pore structure in a manner that will prolong the release period for
days to weeks. It can be assumed that the Na-ampicillin would be
quite susceptible to ionic exchanges in saline solution, so the
positively charged amine groups will not allow strong bonds with
added ampicillin. However, such a delivery pattern (short but
diffusion-controlled and predictable) is beneficial for therapeutic
functions of certain drugs, like antibiotics.59 Strategies to prolong
the release period (days to weeks) would be needed to enhance
the potential of mesoporous BG nanoparticles as a delivery
system of small chemical drugs with weak ionic interactions. A
prolonged release will also allow better control of the drug
release kinetics.
The present study also assessed the potential of BGn as a gene
delivery vehicle. Genetic modification of cells is possible by the
proper delivery of genes (such as miRNA, siRNA and pDNA)
within cells, targeted to the cytoplasm or the nucleus. As the size
of the mesoporous BGn was presently in the range of 80–100 nm,
we consider that particle entry into the nucleus does not readily
occur, given the presence of the nuclear membrane.60 Therefore,
a strategy to deliver genes into the cytoplasm in a manner that
retains the functional activity of the delivered genes was
employed, utilizing siRNA. Once the designed siRNA is present
in the cytoplasm assisted by the synthetic nanocarrier after being
taken up by the cell, it is able to abrogate fundamental cellular
pathways through the well-established mechanism of RNA
interference (RNAi).60 To realize the therapeutic potential of
siRNA, safe and effective delivery systems are required. The
delivery of naked siRNA into target cells and tissues is not easily
implemented, mainly due to the degradation by endogeneous
enzymes, and the difficulties in penetrating cell membranes
imposed by large size (impermeable to ion channels) and highly
negative charge.59 For this reason, for the delivery of siRNA,
viral vectors were among the first vehicles introduced. However,
their tendency to produce unacceptable toxic effects associated
with immune rejection61 negated their routine use. Synthetic
nanoparticles from polymers and inorganics have thus gained
great attraction.44,62–64 Although many compositions and
formulations have been studied, there have been few reports on
the gene delivery with inorganic nanoparticles compared to
polymeric materials.44,64,65 One of the most fascinating non-viral
inorganic vectors, mesoporous silica nanoparticles (MSNs), have
been investigated in terms of their ability to incorporate genes
and deliver them into intracellular compartments.44,64 The mes-
oporous structure of the nanoparticles allows the uptake of a
large quantity of nucleic acids, and the particle size (<100 nm)
allows effective intracellular uptake through an endocytosis
mechanism. Furthermore, the entrapped genes within mesopores
can be secured and released sustainably.66 The presently devel-
oped mesoporous BGn share these basic concepts. Moreover,
calcium-containing BGn have beneficial effects on hard tissue
regeneration. Studies have suggested significant roles of such
nanoparticles in bone cell functions including stem cell differ-
entiation and hard tissue mineralization.1,3,22 Therefore, the
This journal is ª The Royal Society of Chemistry 2012
Page 13
Dow
nloa
ded
by S
eoul
Nat
iona
l Uni
vers
ity o
n 22
Nov
embe
r 20
12Pu
blis
hed
on 0
9 O
ctob
er 2
012
on h
ttp://
pubs
.rsc
.org
| do
i:10.
1039
/C2N
R31
775C
View Article Online
presently developed mesoporous BGn are a promising candidate
non-viral vector for the delivery of genes targeting hard tissues
like bone and teeth.
Cytotoxicity tests confirmed that the mesoporous BGn were
very useful for cells in interest, including MSCs and osteoblasts.
Cellular toxicity is greatly influenced by factors of mesoporous
nanomaterials including size, pore structure, and surface chem-
istry.67 BGn1 and BGn2 with sizes of approximately 80–100 nm
containing mesoporous structures and having aminated
surface chemistry were not appreciably toxic (>90%, compared
to control) to a variety of cells, at doses ranging from 5 to 100 mg
ml�1. Compared to conventional nonbioactive silica nano-
particles, such as MSNs, the results on the BGn confirmed their
similar or even less toxic nature.67 Although more in-depth
investigations need to support, the observation of up-regulated
cell viability at very low doses particularly in the MC3T3-E1
osteoblastic cells makes it worthwhile to do further research on
the effective roles of the BGn in proliferation of bone-related
cells. The presently developed BGn are considered favorable for
preserving cell viability, at least to the level of MSNs which are
regarded as effective cell transfection nanocarriers.44 Specifically,
at the low dose of 10 mg ml�1 used for drug and gene delivery, the
cell viability level was almost equal to or even higher than that of
control.
If toxicity is not a concern, the efficacy of BGn for use as
siRNA delivery carriers relies on their ability to incorporate
genes and release them in a proper manner. Surface functional-
ization with amine groups is critical to utilize BGn in capturing
polyanionic nucleic acids. Presently, 2.59 mg of siRNA could be
loaded into 50 mg of BGn2, demonstrating a loading capacity of
�5.18%, and we used 10 mg of BGn2 for siRNA transfection,
which allowed a sufficient quantity of siRNA to be taken up into
cells if the transfection efficiency is high enough. Cells treated
with a FITC-conjugated siRNA–BGn2 complex demonstrated a
transfection efficiency as high as 80% in FACS analysis. This
cellular uptake phenomenon in BGn is considered to be
explained by the endocytosis mechanism similar to that reported
in MSNs. Although MSNs lack cell membrane-bound receptors,
such as low density lipoprotein or transferrin receptors, they are
known to have a great affinity for adsorbing onto cell membrane
surfaces, particularly the head groups of various phospholipids,
which thus leads to ‘adsorptive’ endocytosis of the nano-
particles.68–70 TEM also revealed pronounced intracellular
uptake of the siRNA-loaded nanoparticles, which were present
within the cytosol, localized in some intracellular compartments
such as mitochondria and rough endoplasmic reticulum. While
these sites are not exactly target-specific, it is presumed that the
internalized BG nanoparticles will further release siRNA in the
cytosol to allow the biological function of target-gene silencing.
The release profile of siRNA gene from the nanoparticles is
consistent with the view that very strong interactions are created
between the aminated-surface of BGn2 with siRNA than with
Na-ampicillin, facilitating effective complexation of the siRNA
within the mesopores, which is mainly due to the highly negative-
charged polyanion characteristic of nucleic acids. Thus, relatively
sustained and linear release of siRNA up to about 3 days could
be attained, and the total quantity released was almost 45% of
the initial loading quantity. Taken the result that an almost
saturation was attained after 3 days, parts of siRNA might be
This journal is ª The Royal Society of Chemistry 2012
degraded in the complexation process (but counted as the loaded
gene) or during the release period. However, a significant portion
of the siRNA was shown to release, profiling zero-order kinetics
up to 3 days, and this release time should be critically considered
in designing the siRNA delivery system, as we can presume the
possible intracellular action periods of gene silencing based on
the siRNA expression period.
We further investigated the possible biological interference of
siRNA within cells using a model silencing gene, bcl-2. The bcl-2
biological function in HeLa cells was almost completely
knocked-down when the target siRNA-loaded BGn2 nano-
particles were transfected, which was not readily recognizable in
other comparison groups including target siRNA alone, BGn2
alone, or scramble RNA-loaded BGn2, confirming that the
intracellular transfection of the siRNA–BGn2 complex was in
effect in silencing target gene. The bcl-2 gene is not specified for
hard tissue regeneration, but was utilized to confirm the parallel
action of transfection and the subsequent gene-silencing effect.
More relevant applications are possible and wait for further
studies.
Together with the excellent in vitro bone-bioactivity (namely
apatite forming ability) and the low cellular toxicity, the ability
to load therapeutic molecules and deliver them in a proper
manner as proved using model biomolecules demonstrates the
potential usefulness of the mesoporous BG nanoparticles for the
regeneration of bone. Specific diverse applications that are
immediately apparent include the use of nano-vehicles in the
osteogenesis of stem cells by the direct treatment of the gene-
loaded BG nanoparticles as well as in the bone regeneration as
implantable biomaterials and tissue engineering scaffolds when
incorporating drug-loaded BG nanoparticles.
5. Conclusions
Here we show for the first time the performance of mesoporous
BG nanoparticles as the delivery vehicles of biomolecules,
including drugs and genes. The particle sizes (<100 nm) and
extremely high surface area and large volume of mesopores
inside the nanoparticles, as well as the nontoxic and bioactive
traits are very suitable for such a purpose, by incorporating
candidate biomolecules and releasing them in a sustainable
manner. After the surface functionalization with amine groups,
specific utility of the mesoporous BG nanoparticles as a gene
delivery vector was demonstrated. As a model gene studied,
siRNA was shown to be effectively loaded within the nano-
particles, and then further released in vitro over a period of
3 days. The siRNA-loaded BG nanoparticles were shown to be
transfected efficiently into cells and to have a gene-silencing
function as well. The results support the possible promising uses
of the novel non-viral vector in delivering target genes such as
siRNA for the disease treatment or regeneration of hard tissues.
Acknowledgements
This study was supported by grants from the Priority Research
Centers Program (2009-0093829) and the WCU program (R31-
2008-000-100069-0), National Research Foundation, South
Korea.
Nanoscale, 2012, 4, 7475–7488 | 7487
Page 14
Dow
nloa
ded
by S
eoul
Nat
iona
l Uni
vers
ity o
n 22
Nov
embe
r 20
12Pu
blis
hed
on 0
9 O
ctob
er 2
012
on h
ttp://
pubs
.rsc
.org
| do
i:10.
1039
/C2N
R31
775C
View Article Online
References
1 J. R. Jones, O. Tsigkoua, E. E. Coates, M. M. Stevens, J. M. Polakand L. L. Hench, Biomaterials, 2007, 28, 1653–1663.
2 P. Ducheyne and Q. Qiu, Biomaterials, 1999, 20, 2287–2303.3 M. Bosetti and M. Cannas, Biomaterials, 2005, 26, 3873–3879.4 L. L. Hench, J. Am. Ceram. Soc., 1998, 81, 1705–1733.5 M. Marcolongo, P. Ducheyne, J. Garino and E. Schepers, J. Biomed.Mater. Res., 1998, 39, 161–170.
6 T. Turunen, J. Peltola, A. Yli-Urpo and R. P. Happonen, Clin. OralImplants Res., 2004, 15, 135–141.
7 R. Li, A. E. Clark and L. L. Hench, J. Appl. Biomater., 1991, 2, 231–239.
8 E. M. Santos, S. Radin and P. Ducheyne, Biomaterials, 1999, 20,1695–1700.
9 H. W. Kim, H. E. Kim and J. C. Knowles, Adv. Funct. Mater., 2006,16, 1529–1535.
10 W. Xia, D. Zhang and J. Chang, Nanotechnology, 2007, 18, 5601–5607.
11 S. A. Oh, S. H. Kim, J. E. Won, J. J. Kim, U. S. Shin and H. W. Kim,J. Tissue Eng., 2010, 2010, 475160.
12 X. Chen, B. Lei, Y. Wang and N. Zhao, J. Non-Cryst. Solids, 2009,355, 791–796.
13 S. K. Misra, D. Mohn, T. J. Brunner, W. J. Stark, S. E. Philip, I. Roy,V. Salih, J. C. Knowles and A. R. Boccaccini, Biomaterials, 2008, 29,1750–1761.
14 D. S. Couto, Z. Hong and J. F. Mano, Acta Biomater., 2009, 5, 115–123.
15 B. Marelli, C. E. Ghezzi, D. Mohn, W. J. Stark, J. E. Barralet,A. R. Boccaccini and S. N. Nazhat, Biomaterials, 2011, 32, 8915–8926.
16 W. Xia and J. Chang, J. Controlled Release, 2006, 110, 522–530.17 L. Z. Zhao, X. X. Yan, X. F. Zhou, L. Zhou, H. N. Wang,
H. W. Tang and C. Yu, Microporous Mesoporous Mater., 2008,109, 210–215.
18 H.W. Kim, J. H. Song andH. E. Kim, J. Biomed.Mater. Res., Part A,2006, 79, 698–705.
19 H. W. Kim, H. H. Lee and G. S. Chun, J. Biomed. Mater. Res., PartA, 2008, 85, 651–663.
20 H. H. Lee, H. S. Yu, J. H. Jang and H.W. Kim, Acta Biomater., 2008,4, 622–629.
21 G. M. Luz and J. F. Mano, Nanotechnology, 2011, 22, 494014.22 S. Labbaf, O. Tsigkoua, K. H. M€uller, M. M. Stevens, A. E. Porter
and J. R. Jones, Biomaterials, 2011, 32, 1010–1018.23 S. I. Roohani-Esfahani, S. Nouri-Khorasani, Z. F. Lu,
R. C. Appleyard and H. Zreiqat, Acta Biomater., 2011, 7, 1307–1318.24 S. K. Misra, T. Ansari, D. Mohn, S. P. Valappil, T. J. Brunner,
W. J. Stark, I. Roy, J. C. Knowles, P. D. Sibbons, E. V Jones,A. R. Boccaccini1 and V. Salih, J. R. Soc. Interface, 2010, 7, 453–465.
25 T. Waltimo, T. J. Brunner, M. Vollenweider, W. J. Stark andM. Zehnder, J. Dent. Res., 2007, 86, 754–757.
26 V. Mortazavi, M. M. Nahrkhalaji, M. H. Fathi, S. B. Mousavi andB. N. Esfahani, J. Biomed. Mater. Res., Part A, 2010, 94, 160–168.
27 M. Vollenweider, T. J. Brunner, S. Knecht, R. N. Grass, M. Zehnder,T. Imfeld and W. J. Stark, Acta Biomater., 2007, 3, 936–943.
28 W. Xia and J. Chang, J. Non-Cryst. Solids, 2008, 354, 1338–1341.29 A. A. Ahmed, A. A. Ali, D. R. Mahmoud and A. M. El-Fiqi, Solid
State Sci., 2011, 13, 981–992.30 L. D. Gelb and K. E. Gubbins, Langmuir, 1998, 14, 2097–2111.31 V. Cauda, B. Onida, B. Platschek, L.M€uhlstein and T. Bein, J.Mater.
Chem., 2008, 18, 5888–5899.32 T. Kokubo, H. Kushitani, S. Sakka, T. Kitsugi and T. Yamamuro,
J. Biomed. Mater. Res., 1990, 24, 721–734.33 S. A. Oh, S. H. Kim, J. E. Won, J. J. Kim, U. S. Shin and H. W. Kim,
J. Tissue Eng., 2010, 475260.34 S. Pabisch, B. Feichtenschlager, G. Kickelbick and H. Peterlik, Chem.
Phys. Lett., 2012, 521, 91–97.
7488 | Nanoscale, 2012, 4, 7475–7488
35 X. Yan, X. Huang, C. Yu, H. Deng, Y. Wang, Z. Zhang, S. Qiao,G. Lu and D. Zhao, Biomaterials, 2006, 27, 3396–3403.
36 H. C. Lee, H. J. Kim, C. H. Rhee, K. H. Lee, J. S. Lee andS. H. Chung, Microporous Mesoporous Mater., 2005, 79, 61–68.
37 J. Sun, Y. Li, L. Li, W. Zhao, L. Li and J. Gao, et al., J. Non-Cryst.Solids, 2008, 354, 3799–3805.
38 K. Gao and L. Huang, Mol. Pharmaceutics, 2009, 6, 651–658.39 D.M. Dykxhoorn, C. D. Novina and P. A. Sharp,Nat. Rev. Mol. Cell
Biol., 2003, 4, 457–467.40 H. E. Davis and J. K. Leach, Ann. Biomed. Eng., 2011, 39, 1–13.41 S. H. Lee and H. Shin, Adv. Drug Delivery Rev., 2007, 59, 339–359.42 C. Wu and J. Y. Xiao, Ther. Delivery, 2011, 2, 1189–1198.43 X. Li, X. Wang, Z. Hua and J. Shi, Acta Mater., 2008, 56, 3260–3265.44 X. Li, Q. R. Xie, J. Zhang, W. Xia and H. Gu, Biomaterials, 2011, 32,
9546–9556.45 I. I. Slowing, J. L. Vivero-Escoto, C. W. Wu and V. S. Y. Lin, Adv.
Drug Delivery Rev., 2008, 60, 1278–1288.46 J. Lu, M. Liong, J. I. Zink and F. Tamanoi, Small, 2007, 3, 1341–
1346.47 M. Eltohamy, U. S. Shin and H. W. Kim, Mater. Lett., 2011, 65,
3570–3573.48 M. Vallet-Regi, A. R�amila, R. P. Del Real and J. P�erez-Pariente,
Chem. Mater., 2001, 13, 308–311.49 F. Qu, G. Zhu, S. Huang, S. Li and S. Qiu, ChemPhysChem, 2006, 7,
400–406.50 B. Munoz, A. Ramila, J. Perez-Pariente, I. Diaz and M. Vallet-Regi,
Chem. Mater., 2003, 15, 500–503.51 F. Torney, B. G. Trewyn, V. S. Y. Lin and K. Wang, Nat.
Nanotechnol., 2007, 2, 295–300.52 H. Meng, M. Liong, T. Xia, Z. Li, Z. Ji, J. I. Zink and A. E. Nel, ACS
Nano, 2010, 4, 4539–4550.53 S. K. Jana, R. Nishida, K. Shindo, T. Kugita and S. Namba,
Microporous Mesoporous Mater., 2004, 68, 133–142.54 C. H. Lee, L. W. Lo, C. Y. Mou and C. S. Yang, Adv. Funct. Mater.,
2008, 18, 3283–3292.55 Q. He and J. Shi, J. Mater. Chem., 2011, 21, 5845–5855.56 J. W. Peterson, R. S. Burkhart, D. C. Shaw, A. B. Schuiling,
M. J. Haserodt and M. D. Seymour, Chemosphere, 2010, 80, 1268–1273.
57 A. Sarkar, K. C. Anderson and G. E. Kellogg, Eur. J. Med. Chem.,2012, 52, 98–110.
58 B. G. Trewyn, I. I. Slowing, S. Giri, H. T. Chen and V. S. Y. Lin, Acc.Chem. Res., 2007, 40, 846–853.
59 A. C. Queiroz, J. D. Santos, F. J. Monteiro, I. R. Gibson andJ. C. Knowles, Biomaterials, 2001, 22, 1393–1400.
60 K. A. Whitehead, R. Langer and D. G. Anderson, Nat. Rev. DrugDiscovery, 2009, 8, 129–138.
61 J. Barquinero, H. Eixarch and M. Perez-Melgosa, Gene Ther., 2004,11, S3–S9.
62 D. Cun, C. Foged, M. Yang, S. Frøkjær and H. M. Nielsen, Int.J. Pharm., 2010, 390, 70–75.
63 X. Yuan, L. Li, A. Rathinavelu, J. Hao, M. He, V. Heitlage, L. Tam,S. Viqar and M. Salehi, J. Nanosci. Nanotechnol., 2006, 6, 2821–2828.
64 C. Hom, J. Lu, M. Liong, H. Luo, Z. Li, J. I. Zink and F. Tamanoi,Small, 2010, 6, 1185–1190.
65 C. Hom, J. Lu and F. Tamanoi, J. Mater. Chem., 2009, 19, 6308–6316.
66 J. L. Vivero-Escoto, I. I. Slowing, B. G. Trewyn and V. S. Y. Lin,Small, 2010, 6, 1952–1967.
67 Z. Tao, B. B. Toms, J. Goodisman and T. Asefa, Chem. Res. Toxicol.,2009, 22, 1869–1880.
68 S. Mornet, O. Lambert, E. Duguet and A. Brisson, Nano Lett., 2005,5, 281–285.
69 X. Xing, X. He, J. Peng, K. Wang and W. Tan, J. Nanosci.Nanotechnol., 2005, 5, 1688–1693.
70 I. I. Slowing, J. L. Vivero-Escoto, C. W. Wu and V. S. Y. Lin, Adv.Drug Delivery Rev., 2008, 60, 1278–1288.
This journal is ª The Royal Society of Chemistry 2012