Organosilicate Nanoparticles as Gene Delivery Vehicles for Bone Cells
by
Suniti Moudgil
B. S., Chemical Engineering
University of Florida, 1998
Submitted to the Department of Chemical Engineering in Partial Fulfillment
of the Requirements for the Degree of
Doctor of Philosophy in Chemical Engineering
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
February 2004
© Massachusetts Institute of Technology 2004. All rights reserved.
ARCVES
I
-
_
i
I
_
,,A
MASSACHUSETTS INSTITUTE
OF TECHNOLOGY
JAN 15 2004
LIBRARIES
Author:
roe
Department
Chemical
f
Engineering
January 7, 2004
Certified by:
/Trofessor'Jackie
[. Ying
Professr of Chemical Engineering
Thesis Supervisor
Accepted by:
.
Professor Daniel Blankschtein
Professor of Chemical Engineering
Chairman, Committee for Graduate Studies
ArtGItVES
,,,
8
Organosilicate Nanoparticles as Gene Delivery Vehicles for Bone Cells
by
Suniti Moudgil
B.S., Chemical Engineering
University of Florida, 1998
Submitted to the Department of Chemical Engineering
on January 7, 2004 in Partial Fulfillment of the Requirements for
the Degree of Doctor of Philosophy in Chemical Engineering
Abstract
While bone has a substantial capacity to heal itself, there are approximately 1 million
fractures that occur in the U.S. annually that are difficult to heal. These include fractures that
occur at sites of low vascularity, fractures that result in a large amount of tissue loss, and
fractures that result from bone fragility syndromes such as osteoporosis. There has been a
great deal of interest in the tissue engineering of bone in order to treat such fractures. One
major aspect of the tissue engineering approach involves the addition of growth factors or
proteins to synthetic grafts to accelerate bone regeneration. However, delivering these
proteins at the appropriate times and therapeutic levels poses significant challenges.
Alternatively, delivering the genes that encode for these proteins could offer a more effective
treatment, since proper incorporation of the appropriate genes into cellular nuclei would allow
the cells to manufacture authentic protein products.
The motivation of this research was to design new materials for gene delivery to bone
cells. Conventional non-viral vectors are plagued by toxicity and low transfection efficiencies.
The purpose of this work was to design bioactive nanoparticles that could enter the osteoblast
membrane without inducing toxicity. These materials were silicate-based, since doped
silicates have been shown to possess osteogenic properties.
A method to synthesize
monodisperse, spherical organosilicate nanoparticles using a surfactant-stabilized sol-gel
technique was developed. Surface dopants such as Ca, Mg and Na were found to influence
cellular response to nanoparticles. In addition to particle composition, particle size was also
found to have a significant effect on osteoblast uptake and cell proliferation.
The metabolic response of these cells after particle ingestion was also characterized in
order to ensure that the osteoblasts retained their phenotype. The expression of various
proteins involved in bone formation, such as alkaline phosphatase, osteocalcin, osteopontin
and fibronectin was quantified. The results indicated that osteoblasts retained their phenotype
after organosilicate nanoparticle ingestion. The levels of various cytokines expressed during
inflammatory response remained low due to the biocompatibility of amorphous silica.
An optimized Ca-SiO 2 nanoparticulate system was developed with maximum particle
uptake that enhanced cell viability. A model gene delivery system was created by complexing
these nanoparticles with plasmid DNA that encoded for green fluorescent protein (GFP). The
effects of nanoparticle size, composition and surface charge on complex size, DNA binding
affinity and subsequent GFP expression in osteoblasts were investigated in detail.
2
In addition to primary and immortalized osteoblasts, we have studied the effect of this
gene delivery system on two other control cell lines: fibroblasts (connective tissue cells) and
The Ca-SiO2-DNA complexes displayed
hepatocytes (non-connective tissue cells).
significantly higher transfection efficiencies in osteoblasts and fibroblasts relative to
hepatocytes compared to lipofectamine-DNA complexes. In addition, Ca-SiO2-DNA
complexes enhanced osteoblast cell proliferation while achieving successful transfection. In
clinical applications, such characteristics would allow viable cells to remain at the fracture
site while delivering growth factors via transfected cells. Properties such as calcium loading
and particle size could be tailored to achieve the desired amount of growth factor delivery.
The ability of Ca-SiO2-DNA complexes to transfect bone cells selectively without inducing
cytotoxicity suggested that this nanoparticulate system has exciting potential for clinical
applications for bone regeneration.
Thesis Supervisor: Jackie Y. Ying
Title: Professor of Chemical Engineering
3
Acknowledgments
I wish to thank Prof. Jackie Ying for being my advisor
and mentor.
Her
encouragement has contributed greatly to my professional development, and I look forward to
our continued association and friendship. I would also like to acknowledge the members of
my thesis committee, Prof. Paula Hammond, Prof. Myron Spector and Prof. Dane Wittrup, for
their guidance and suggestions during the course of my graduate studies. I would especially
like to acknowledge Prof. Spector for many interesting research discussions.
I would also like to acknowledge my colleagues in the Nanostructured
Materials
Research Laboratory (NMRL). I had the pleasure of working with some of the earlier
doctoral students of the NMRL, Dr. Larry Panchula, Dr. Mark Fokema, Dr. Andrey Zarur, Dr.
Michael Wong, Dr. Chen-Chi Wang, Dr. John Lettow, Dr. Justin McCue and Dr. Jason
Sweeney, and I thank them for their help. I would especially like to acknowledge Dr. Edward
Ahn for serving as my mentor when I first entered the group. I thank Steven Weiss, Dr.
Neeraj Sangar and Yee Su for their great friendship and support over the past few years. I
would also like to acknowledge two undergraduate students, Frederick Tan and Henry Tsang,
for their contributions to my research. I am grateful to the current members of the NMRL for
creating such an enjoyable and memorable work environment: Dr. Javier Garcia-Martinez, Dr.
Xiaohua Huang, Todd Zion, Tom Lancaster, Pemakorn Pitukmanorom, Tseh-Hwan Yong,
Noreen Zaman, Jianyi Cui, Yuhua Hu, Hong He and Cindy Ren.
I would like to acknowledge Prof. Clark Colton's lab for generously sharing their cell
culture facilities, and Hao Wang of the Brigham and Women's Hospital for providing us with
primary cells for our experiments. I would also like to thank Nicki Watson of the Whitehead
Institute for her assistance with electon microscopy.
I thank Mike Frongillo and Dr. Libby
Shaw of MIT CMSE and Dr. Steven Kooi of ISN for their help with various characterization
techniques. I thank Arline Benford and Barbara Driscoll for their support and constant
cheerfulness, and I especially thank Linda Mousseau for her friendship and for helping me
with so many administrative tasks during my time here.
Thanks also to Suzanne Easterly,
Anne Fowler, Jennifer Shedd and Mary Keith for their help with numerous departmental
responsibilities. Financial support for my research from the Singapore-MIT Alliance and the
NSF Graduate Research Fellowship is greatly appreciated.
I especially thank my family for their unconditional love and support. I would like to
thank my father, Brij Moudgil, for his perpetual optimism and for encouraging me to keep a
positive attitude throughout my graduate career. He sets professional and personal examples
that serve as constant inspirations. I would also like to thank my mother, Sheela Moudgil, for
her love and concern, and for teaching me to do my best. I thank my sister, Sarika, and my
brother, Bharat, for keeping my life filled with humor and laughter. Finally, I thank my
fianc6, Ravi Amaravadi, for his love and friendship. His companionship has made the last
couple of years not only the best years of my graduate career, but the best years overall. I
look forward to our future together.
4
Contents
Chapter 1 - Background and Research Motivation
13
1.1. Bone Regeneration and Fracture Repair
13
1.2. Synthetic Bone Grafts
13
1.3. Tissue Engineering of Bone
14
1.4. Gene Therapy in Orthopedic Medicine
15
1.5. Research Objectives
17
1.5.1. DopedSilicatesas OrthopedicBiomaterials
17
1.5.2. Organosilicate Nanoparticlesfor Gene Delivery to Osteoblasts
18
1.6. References
19
Chapter 2 - Synthesis, Cytoxicity and Cellular Uptake of Organosilicate
Nanoparticles
22
2.1. Introduction
22
2.1.1. Doped Silicate Synthesis via Sol-Gel Processing
22
2.1.2. Initial Bioactivity Studies with Organosilicate Gels
23
2.2. Experimental
23
2.2.1. Synthesis of Organosilicate Microspheres
23
2.2.2. Synthesisof OrganosilicateNanoparticles
24
2.2.3. Materials Characterization
24
2.2.4. Cell Culture Experiments
25
2.3. Results and Discussion
27
2.3.1. Effect of Synthesis Conditions on Particle Size Distribution
27
2.3.2. Effect of Nanoparticles on Cellular Response
30
2.3.3. Characterization of Ca-SiO2 Nanoparticles
32
2.3.4. Effect of Ca-Si0 2 Nanoparticles on Cellular Uptake
33
2.3.5. Intracellularvs. ExtracellularSignaling
37
2.3.6. Effect of Ca-Si02 Nanoparticles on Other Cell Types
39
2.3.7. Effect of Other Compositions on Osteoblast Uptake and Proliferation
40
2.3.8. Calcium Transporters in the Osteoblast
42
2.4. Summary
43
2.5. References
44
5
Chapter 3 - Osteoblast Protein Expression in Response to Organosilicate
Nanoparticles
3.1. Introduction
46
46
3.1.1. Proteins Involved in Bone Formation
46
3.1.2. Cytokine Expression
47
3.2. Experimental
47
3.3. Results and Discussion
49
3.3.1. ALP Activity Measurements
49
3.3.2. Osteocalcin Expression Measurements
51
3.3.3. Osteopontin Expression Measurements
52
3.3.4. Fibronectin Expression Measurements
54
3.3.5. Cytokine Production
55
3.4. Summary
61
3.5. References
62
Chapter 4 - Synthesis, Physicochemical Characterization and In Vitro
Transfection Studies of Organosilicate Nanoparticle-DNA Complexes
64
4.1. Introduction
64
4.2. Experimental
65
4.2.1. Synthesis of Nanoparticle-DNA Complexes
65
4.2.2. Physicochemical Characterization of Vector-DNA Complexes
66
4.2.3. Transfection Experiments
66
4.3. Results and Discussion
67
4.3.1. Synthesis and Physicochemical Characterization of Ca-SiO2 -DNA Complexes
67
4.3.2. Synthesisand PhysicochemicalCharacterizationof NH3+-SiO
2-DNA
Complexes
72
4.3.3. Transjfction Studies
74
4.4. Summary
82
4.5. References
83
Chapter 5 - Recommendations for Future Work
85
Chapter 6 - Conclusions
87
6
List of Figures
Figure 1.1. Screw fixation of a fracture in the lateral tibial plateau, augmented
with Norian SRS® carbonated apatite cement [14].
14
Figure 1.2. Schematic of the major intracellular barriers that DNA encounters
before entry into the nucleus, including complex formation, initial uptake and
endosomal release.
16
Figure 1.3. Insertion of Bioglass® (BG) implant into a rat tibial defect results
in surface formation of calcium phosphate (CaP) layer, followed by osteoblast
(O) adhesion and bone (B) formation [11].
17
Figure 2.1. SEM images of organosilicate nanoparticles synthesized (a) without
surfactant (mean particle size = 1.5 ptm) and (b) with surfactant (mean particle
size = 7 nm).
27
Figure 2.2. Effects of pH and water/alkoxide ratio on the particle size distribution
of organosilicate nanoparticles.
28
Figure 2.3. DLS particle size distributions of 10-nm organosilicate nanoparticles
(a) before and (b) after 5 wt% Ca addition.
30
Figure 2.4. Effect of the surface composition of 10-nm organosilicate nanoparticles
on osteoblast proliferation after 7 days. Nanoparticle loading: 0.15 wt%; nominal
salt loading: 2.5 wt%. Values are mean + standard error of the mean; n = 3.
31
Figure 2.5. Effect of nominal Ca loading and particle size of Ca-SiO2 on osteoblast
proliferation. Particle sizes of () 10 nm, () 50 nm, (A) 100 nm and () 200 nm
are examined at a particle loading of 0.15 wt%. Values are mean ± standard error
of the mean; n = 3.
31
Figure 2.6. Transmission electron micrographs of primary osteoblasts that have
ingested: (a) 10-nm, (b) 50-nm, (c) 100-nm Ca-SiO 2 nanoparticles containing
2.5 wt% Ca.
Figure 2.7. Effect of particle loading on the cellular uptake of ()
34
10-nm,
(*) 50-nm and (A) 100-nm Ca-SiO 2 particles. All particles had a nominal Ca
loading of 2.5 wt%. Values are mean + standard error of the mean; n = 3.
35
Figure 2.8. Effect of Ca loading and particle size on the cellular uptake of
(I) 10-nm, () 50-nm, () 100-nm and () 200-nm Ca-SiO2 particles. Particle
loading: 0.15 wt%. Values are mean ± standard error of the mean; n = 3.
35
Figure 2.9. Effect of Ca loading and particle size on the osteoblast proliferation for
(I) 10-nm, () 50-nm and () 100-nm Ca-SiO2 nanoparticles. Particle
loading: 0.15 wt%; n = 3.
36
Figure 2.10. DLS particle size distributions of 100-nm organosilicate nanoparticles
(a) before and (b) after 15 wt% Ca addition.
37
7
Figure 2.11. Osteoblast proliferation in the presence of fresh and conditioned cell
culture media. Values are mean + standard error of the mean; n = 3. *Nominal
Ca loading: 2.5 wt%.
38
Figure 2.12. Basic mechanisms of calcium signaling [20].
39
Figure 2.13. Proliferation of osteoblasts, fibroblasts and hepatocytes in the presence
of () 10-nm, (E) 50-nm and () 100-nm Ca-SiO 2 nanoparticles.
0.15 wt%; Ca loading: 2.5 wt%.
Particle loading:
Figure 2.14. Effect of () Ca-SiO 2, (+) Mg-SiO 2 and (A) Na-SiO 2 nanoparticle
size on the proliferation of osteoblasts. Particle loading: 0.15 wt%; nominal salt
loading: 2.5 wt%. Values are mean ± standard error of the mean; n = 3. Note: cell
density in the absence of nanoparticles is - 230 cells/mm 2.
40
41
Figure 2.15. Effect of () Ca-SiO2 , () Mg-SiO2 and (A) Na-SiO2 nanoparticle
size on the uptake of osteoblasts. Particle loading: 0.15 wt%; nominal salt
loading: 2.5 wt%. Values are mean ± standard error of the mean; n = 3.
41
Figure 2.16. Zeta potential vs. pH for nanoparticle suspensions of () Ca-SiO2,
(A) Mg-SiO 2, (e) Na-SiO 2 and () SiO2 in complete cell culture medium.
Particle loading: 0.15 wt%; nominal salt loading: 2.5 wt%.
Figure 2.17. Transport systems in osteoblasts [21].
42
43
Figure 3.1. Effect of nanoparticle composition on the ALP activity of osteoblasts.
Particle size: 50 nm; nominal salt loading: 2.5 wt%. Values are expressed as
nanomoles of p-NP produced by 1 ig of total protein during 1 hr. Values are mean ±
standard error of the mean; n = 3.
50
Figure 3.2. Effect of Ca-SiO 2 particle size on the ALP activity of osteoblasts.
Nominal Ca loading: 2.5 wt%. Values are expressed as nanomoles of p-NP produced
by 1 ig of total protein during 1 hr. Values are mean + standard error of the mean;
n=3.
50
Figure 3.3. Optical micrographs of osteoblasts (a) before and (b) after incubation with
50-nm Ca-SiO 2 nanoparticles. The darker regions indicate ALP expression.
51
Figure 3.4. Effect of nanoparticle composition on the osteocalcin expression of
Osteoblasts. Particle size: 50 nm; nominal salt loading: 2.5 wt%. Values are
expressed as proteins produced by 106cells. Values are mean ± standard error of the
mean; n = 3.
52
Figure 3.5. Effect of Ca-SiO2 particle size on the osteocalcin expression of
osteoblasts. Nominal Ca loading: 2.5 wt%. Values are expressed as proteins
produced by 106 cells. Values are mean standard error of the mean; n = 3.
52
Figure 3.6. Effect of nanoparticle composition on the osteopontin expression of
osteoblasts. Particle size: 50 nm; nominal salt loading: 2.5 wt%. Values are
expressed as proteins produced by 106 cells. Values are mean standard error
of the mean; n = 3.
8
53
Figure 3.7. Effect of Ca-SiO2 particle size on the osteopontin expression of osteoblasts.
Nominal Ca loading: 2.5 wt%. Values are expressed as proteins produced by 106
cells. Values are mean + standard error of the mean; n = 3.
53
Figure 3.8. Effect of nanoparticle composition on the fibronectin production of
osteoblasts. Particle size: 50 nm; nominal salt loading: 2.5 wt%. Values are
expressed as proteins produced by 106 cells. Values are mean + standard error of the
mean; n = 3.
Figure 3.9. Effect of Ca-SiO 2 particle size on the fibronectin production of
osteoblasts. Nominal Ca loading: 2.5 wt%. Values are expressed as proteins
produced by 106 cells. Values are mean ± standard error of the mean; n = 3.
54
55
Figure 3.10. Effect of nanoparticle composition on the interleukin levels of
osteoblasts. Particle size: 50 nm; nominal salt loading: 2.5 wt%. Values are
expressed as proteins produced by 106 cells. Values are mean ± standard error of the
mean; n = 3.
56
Figure 3.1 1. Effect of nanoparticle composition on the protein depletion from cell
culture medium. Particle size: 50 nm; nominal salt loading: 2.5 wt%. Values
are expressed as protein depleted after one week of incubation with nanoparticles.
Values are mean + standard error of the mean; n = 3.
57
Figure 3.12. Effect of Ca-SiO 2 particle size on the IL-6 levels of osteoblasts. Nominal
Ca loading: 2.5 wt%. Values are expressed as proteins produced by 106 cells. Values
are mean + standard error of the mean; n = 3.
58
Figure 3.13. Effect of Ca-SiO 2 nanoparticle size on the IL-6 levels of fibroblasts.
Nominal Ca loading: 2.5 wt%. Values are expressed as proteins produced by 106
cells. Values are mean ± standard error of the mean; n = 3.
59
Figure 3.14. Effect of nanoparticle composition on the TNF-a levels of osteoblasts.
Particle size: 50 nm; nominal salt loading: 2.5 wt%. Values are expressed as
proteins produced by 106 cells. Values are mean ± standard error of the mean; n = 3.
60
Figure 3.15. Effect of Ca-SiO2 particle size on the TNF-a levels of osteoblasts.
Nominal Ca loading: 2.5 wt%. Values are expressed as proteins produced by 106
cells. Values are mean + standard error of the mean; n = 3.
60
Figure 3.16. Effect of nanoparticle composition on the TNF-a levels of fibroblasts.
Particle size: 50 nm; nominal salt loading 2.5 wt%. Values are expressed as
proteins produced by 106 cells. Values are mean + standard error of the mean; n = 3.
61
Figure 4.1. Complex size distributions of complexes formed by addition of
(a) Ca-DNA solution to 50-nm SiO 2 nanoparticles and (b) 50-nm SiO 2 nanoparticles
to Ca-DNA solution. Nominal Ca loading: 1.25 wt%; DNA loading: 24 jgg/ml.
68
Figure 4.2. Agarose gel electrophoresis of nanoparticle-DNA complexes prepared with
50-nm SiO2 with nominal Ca loadings of 1 wt% Ca (Lane 1), 0.5 wt% Ca (Lane 2)
and 0 wt% Ca (Lane 3). Lane 4 represents free DNA without complex formation, and
Lane 5 is the DNA marker. Particle loading: 0.15 wt%; DNA loading: 240 [tg/ml.
69
9
Figure 4.3. Effects of particle size and Ca loading on the extent of DNA binding in
Ca-SiO 2-DNA complexes. Particles of (M) 10 nm, () 50 nm and (A) 100 nm
are examined at a particle loading of 0.15 wt% and a DNA loading of 24 glg/ml.
70
Figure 4.4. Effects of particle size and Ca loading on the size distribution of
Ca-SiO2-DNA complexes. Particles of (I) 10 nm, () 50 nm and
(A) 100 nm are examined at a particle loading of 0.15 wt% and a DNA loading of
24 tg/ml. Values are mean
i
standard error of the mean; n = 3.
71
Figure 4.5. AFM image of pure plasmid DNA.
71
Figure 4.6. AFM images of Ca-SiO2-DNA complexes formed with (a)
10-nm and (b) 50-nm particles. Nominal Ca loading: 1.25 wt%; DNA loading:
24 gtg/ml.
Figure 4.7. The effect of particle
(0) Ca-SiO 2 nanoparticles and
(-) Ca-SiO 2-DNA complexes.
Ca-SiO 2 particles, 1.0 wt% for
size on the DNA binding of (o) NH 3+-SiO 2 and
the size distributions of (0) NH 3+-SiO 2-DNA and
Nominal Ca loading: 1.5 wt% for 10-nm and 50-nm
100-nm Ca-SiO 2 particles.
72
74
Figure 4.8. Transfection efficiencies of lipofectamine-DNA, Ca-SiO2-DNA and
NH 3+-SiO2 -DNA complexes in osteoblasts and hepatocytes. SiO 2 particle size: 50
nm; particle loading: 0.15 wt%; nominal Ca loading: 1.5 wt%; DNA loading: 24
gtg/ml. Values are mean - standard error of the mean; n = 3.
75
Figure 4.9. Effect of Ca loading on the osteoblast transfection efficiency of
Ca-SiO2-DNA complexes formed with () 10-nm, () 50-nm and (A) 100-nm
particles. Particle loading: 0.15 wt%. Values are mean ± standard error of the mean;
n =3.
76
Figure 4.10. Fluorescence micrographs of osteoblasts transfected with Ca-SiO2-DNA
complexes formed with 50-nm Ca-SiO 2 particles with a nominal Ca loading of (a)
0.50 wt%, (b) 0.75 wt%, (c) 1.00 wt% and (d) 1.25 wt%.
Figure 4.11. Effect of Ca loading on the uptake of ()
77
10-nm and (o) 50-nm
Ca-SiO2 nanoparticles and the osteoblast transfection efficiency of complexes formed
with () 10-nm and (o) 50-nm Ca-SiO 2 particles. Particle loading: 0.15 wt%; DNA
loading: 24 g/ml; n = 3.
78
Figure 4.12. Transfection efficiencies in osteoblasts, fibroblasts and hepatocytes
with () lipofectamine-DNA, and Ca-SiO 2-DNA complexes formed with ()
and () 50-nm Ca-SiO 2 particles. DNA loading: 24 Ctg/ml;vector loading:
0.15 wt%. Values are mean ± standard error of the mean; n = 3.
10-nm
79
Figure 4.13. Effect of Ca loading on the cellular proliferation (closed symbols) and
the transfection efficiency (open symbols) of osteoblasts transfected with
Ca-SiO 2-DNA complexes formed with ()
10-nm and () 50-nm Ca-SiO 2 particles.
10
80
-
Figure 4.14. Effect of Ca loading on the GFP expression of osteoblasts transfected
with Ca-SiO 2 -DNA complexes formed with (
)O10-nmand (m) 50-nm particles.
GFP expression of osteoblasts transfected with lipofectamine-DNA is indicated
81
by (]).
11
List of Tables
Table 1.1. Effect of growth factors in various tissues of the musculoskeletal system [18]. 15
Table 2.1. Elemental analysis of Ca-SiO2 nanoparticles.
32
Table 2.2. XPS data for 50-nm Ca-SiO 2 nanoparticles.
32
Table 4.1. Zeta potential values of SiO 2, Ca-SiO 2 and NH 3+-SiO 2 nanoparticles.
73
12
Chapter 1 - Background and Research Motivation
1.1. Bone Regeneration and Fracture Repair
Bone is a dynamic
tissue that is constantly renewing
itself through a natural
remodeling process. This lifelong process involves the resorption of old bone by osteoclast
cells and the deposition of new bone by osteoblast cells.
remodeling
process
is involved in fracture healing.
When a bone breaks, the same
While normal bone possesses
a
substantial capacity to regenerate itself following a fracture, orthopedic medicine lacks an
effective treatment for the nearly 1 million fractures that occur each year in the United States
that heal with difficulty [1].
These injuries include: fractures that occur at sites of low
vascularity, fractures that result in large tissue loss, and fractures that arise from bone fragility
syndromes such as osteoporosis [2]. The cost of treatment for such fractures is extremely
high; for example, the annual cost to treat fractures caused by osteoporosis in the U.S. alone is
approximately $10 billion [3].
While external or internal fixation is the most common treatment for normal fractures,
bone graft augmentation is the current gold standard for acute fracture repair whereby fixation
alone is not sufficient for complete bone restoration [4]. Specifically, autogenous bone grafts
are the most common. However, disadvantages of autografting include: the need for a
second surgical site, donor site morbidity, and the limited availability of autogenous bone,
especially in elderly patients [5]. While allografts can also be used from other donors, they
are less efficient and can potentially transfer disease and trigger host immune responses [6].
1.2. Synthetic Bone Grafts
In order to overcome these challenges, much research has been focused on synthetic
bone grafts to aid the regeneration of bone [7,8].
In particular, orthopedic medicine has
focused on bioactive ceramics, such as calcium phosphates [9,10] and bioactive glasses [1113], which encourage bone cell adhesion and skeletal regeneration.
Figure 1.1 shows the
insertion of a calcium phosphate cement for tibial fracture augmentation. In addition to
ceramic grafts, polymers, metals, and various composite materials have also been studied
extensively [15-17].
13
L
Figure 1.1. Screw fixation of a fracture in the lateral tibial plateau, augmented with Norian
SRS® carbonated apatite cement [14].
Despite these developments, synthetic bone grafts are currently used in only 10% of
the 2.2 million bone graft procedures worldwide [4]. This is because these materials often do
not stimulate bone regeneration rapidly enough or at large enough length scales, as they do
not possess the proteins and osteoinductive cells required for complete repair [4].
1.3. Tissue Engineering of Bone
In order to accelerate bone regeneration, the addition of growth factors to synthetic
bone grafts has also been studied.
Growth factors are proteins capable of stimulating cell
proliferation, migration, differentiation, and matrix synthesis.
The stimulating effect of a
number of growth factors has been demonstrated in a variety of tissues of the musculoskeletal
system (Table 1.1), and the gene that encodes for most of the known growth factors has been
determined [18].
In particular, osteogenic proteins such as bone morphogenetic proteins (BMPs) have
received attention as a promising therapeutic alternative for fracture repair [19,20]. However,
delivering these proteins at the appropriate time and at the appropriate therapeutic level poses
a significant challenge [21]. Alternatively, delivering genes that encode for these proteins
could be a more effective treatment since proper incorporation of the appropriate genes into
cellular nuclei would result in the authentic production of proteins [22,23].
14
Table 1.1. Effect of growth factors in various tissues of the musculoskeletal system [18].
Skeletal
Muscle
Articular
Cartilage
IGF-1
b-FGF
+
+
+
+
NGF
+
aFGF
+-
Growth
Factor
PDGF AA
PDGF AB
PDGF BB
EGF
TGFTGF-[
BMP-2
CDMP (1-3)
-
+/+
+
+
+
+/-
+/-
Ligament/
Tendon
+
+
+
-
+
+
+/+
+
+
+
+
+
+/-
/
+-
+
+
+
BMP-4
BMP-7
VEGF
+
+
+/-
1.4. Gene Therapy in Orthopedic Medicine
Gene therapy involves the transfer of genetic material into cells to alter their function,
that is, to allow cells to produce proteins to treat and potentially cure acute and chronic
conditions [24]. In order for the target cells to manufacture the protein products of the
introduced gene, the genetic material or DNA must be delivered to the nucleus of the cells.
While naked DNA has been shown to successfully transfect certain cell types [25,26], the
transfection efficiency for this approach is relatively low, especially in systemic applications.
In general, DNA requires the aid of a viral or non-viral vector in order to enter the cell
membrane and transfect the nucleus [27]. Although viral vectors are highly efficient in gene
delivery, their major disadvantages include toxicity and the difficulty associated with largescale production [24].
For non-viral vectors, negatively charged DNA molecules are complexed with a
delivery vehicle or transfection agent, such as DEAE dextran [29], calcium phosphate [29,30],
lipids [31,32], proteins, and other polymers [33-35].
The resulting complexes are then
internalized by the cell, generally through endocytosis (Fig. 1.2).
Following uptake, the
complexes enter acidic endosomal compartments that may degrade DNA and its associated
15
complexes. DNA that has survived endocytosis and cytoplasmic nucleases can then enter the
nucleus.
Once inside the nucleus, the biological
processes of gene transcription
and
expression would proceed [36].
Initial uptake
DNA
-
delivery
DNA release in cell
Entry of DNA into
!?.?:a
S.:
.
*'
;
vehicle
.:
:
X·~·
cell membrane
Figure 1.2. Schematic of the major intracellular barriers that DNA encounters before entry
into the nucleus, including complex formation, initial uptake, and endosomal release.
Because of the multiple barriers that the DNA encounters between the cell membrane
and the nucleus, the transfection efficiency of non-viral systems is fairly low, and a number of
strategies have been used to improve the transfection efficiency of gene delivery vehicles by
targeting these various barriers [24,36]. Fortunately, while a high transfection efficiency may
be necessary for gene therapy treatment of systemic diseases, this may not be the case for
bone regeneration.
In the case of bone repair, a relatively small amount of growth factors
produced by a few transfected cells can be of significant value [38]. For example, research
directed toward enhancing bone regeneration using collagen matrices as carriers for specific
genes have shown promising therapeutic results despite exhibiting a relatively low
transfection efficiency [38,39].
While
such matrices have shown some clinical potential with DNA alone, the
incorporation of non-viral transfection agents into these matrices would aid DNA transfection
and result in more rapid bone regeneration. However, since conventional non-viral agents
exhibit toxicity, they cannot be incorporated into clinical implants. A non-viral gene delivery
system that is capable of transfecting bone cells without any toxicity could be combined with
a matrix system for more effective local delivery. Furthermore, a non-viral system that
16
targets bone cells could be administered systemically, thus reducing the need for invasive
surgery. In either case, achieving successful transfection without toxicity would allow viable
bone cells to remain at the fracture site during treatment, thus accelerating fracture healing.
1.5. Research Objectives
In order for materials to enter the cell membrane, they need to be in the submicron size
range.
Many groups have studied the effect of particle size on the cellular uptake and
endocytosis
of different cells lines, and have shown that smaller particles are generally
favorable for cellular uptake [40-44].
Nanoparticles are ideal candidates for biological
delivery vehicles, since they are able to bypass cells of the immune system [45,46]. The focus
of this research is to synthesize bioactive nanoparticles that can enter osteoblasts without
inducing toxicity. These particles are then bound to DNA and tested for their ability to
deliver DNA to osteoblasts.
1.5.1. Doped Silicates as Orthopedic Biomaterials
Doped silicates such as Bioglass® have been recognized for their ability to bond
quickly and strongly to bone [11].
The surface chemistry of these silicates promotes the
formation of a calcium phosphate layer that is able to integrate with living bone, as shown in
Figure 1.3. However, the clinical applications of bioactive glasses as bulk implants have been
limited because of their weak mechanical properties. A nanoparticulate formulation of a
similar composition may provide a bioactive delivery system for bone cells.
~ ~ ~~~~.
I..~
.....
.yii
1/
.-.
r
CP:i,:'.§,
,io
Ca.'D,
%
Figure 1.3. Insertion of Bioglass® (BG) implant into a rat tibial defect results in surface
formation of calcium phosphate (CaP) layer, followed by osteoblast (0) adhesion and bone
(B) formation
[ 11 ].
17
1.5.2. Organosilicate Nanoparticles for Gene Delivery to Osteoblasts
Rather than synthesizing doped silicates using conventional high-temperature methods,
sol-gel processing can be employed to produce silicate monoliths and particles.
This low-
temperature approach provides for greater chemical flexibility and functionalization.
Osteoblasts, like most mammalian cells, are on the order of microns, so they would likely
behave very differently with nanoparticles vs. bulk materials. The particle size may also be
used to control cellular response. While tailoring the composition of organosilicate monoliths
affects their in vitro bioactivity, bulk gels may not be fully reacted, which is detrimental in
cell culture.
To attain a fully reacted material, the sol-gel synthesis parameters could be
modified to attain fully condensed particles in ultrafine domain sizes.
In this research, organosilicate nanoparticles of various compositions and particle sizes
are obtained using sol-gel processing. Bioglass®-like surface compositions can be achieved
by adding various salts to the synthesis scheme. Since Bioglass® elicits a bioactive response
as a bulk
material,
nanoparticles
of similar
compositions
are also expected
to be
biocompatible with no cytotoxicity. These particles may be optimized for osteoblast cell
viability by manipulating their particle size and composition.
Specifically, the effects of Ca-SiO 2, Mg-SiO 2, and Na-SiO 2 nanoparticles on cellular
uptake and proliferation
are studied in this research. Ca surface doping elicits the most
dramatic increase in cellular proliferation and uptake. The amount of cellular uptake can be
correlated to calcium surface loading and particle size, and the optimized system for cellular
uptake will be a likely candidate for a successful gene delivery system for osteoblasts.
The effect of nanoparticle uptake on cellular metabolic response will also be
characterized
in order to ensure that the cells retain their osteoblastic phenotype.
Since
amorphous silica is a bioinert material, cells that have ingested nanoparticles are expected to
secrete proteins at similar levels as cells that are not incubated with particles.
Alkaline
phosphatase (ALP) expression will be quantified along with expression of bone-specific
proteins such as osteopontin and osteocalcin.
measured.
Collagen and fibronectin synthesis will also be
Any increase in cytokine levels associated with particle
ingestion will be
determined as well.
Once an optimized particulate system is developed that shows no cytotoxicity and
maximizes cellular uptake, it will be complexed with DNA to transfect osteoblast cells.
18
Factors that may influence transfection efficiency include complex size as well as the
particles' efficiency to bind and release DNA. Smaller complex sizes may result in greater
initial uptake. However, small complexes that condense DNA too strongly may not release
the DNA effectively to the nucleus.
These characteristics can be tailored through calcium
loading and other surface functionalization of organosilicate particles.
Lastly, complexes that transfect osteoblasts may also transfect other connective tissue
cells like fibroblasts.
The selectivity of nanoparticles towards connective tissue vs. a non-
connective tissue cell line, e.g., hepatocytes, will be tested in vitro through cell culture studies.
The transfection efficiencies of nanoparticle-DNA complexes in osteoblasts and fibroblasts as
compared to hepatocytes will be investigated.
1.6. References
[1]
Bonadio, J., Goldstein, S.A., and Levy, R.J., Adv. Drug Delivery Rev. 33, 53 (1998).
[2]
Gazit, D., Turgeman, G., Kelley, P., Wang, E., Jalenak, M., Zilberman, Y., and
Moutsatso, I., J. Gene Med. 1, 121 (1999).
[3]
Riggs, B., and Melton, L., Bone 17, 505 (1995).
[4]
Lewandrowski,
K., Gresser, J.D., Wise, D.L., and Trantolo, D.J., Biomater. 21, 757
(2000).
[5]
Arrington, E.D., Smith, W.J., Chambers, H.G., Bucknell, A.L., and Davino, N.A., Clin.
Orthop. Rel. Res., 329, 300 (1996).
[6]
Vangsness, C.T., Garcia, I.A., Mills, C.R., Kainer, M.A., Roberts, M.R., and Moore,
T.M., Am. J. Sports Med. 31, 474 (2003).
[7]
Niyibizi, C., and Kim, M., Expert Opin. Inv. Drug 9, 1573 (2000).
[8]
Behairy, Y., and Jasty, M., Orthop. Clin. N. Am. 30, 661 (1999).
[9]
Niedhart, C., Maus, U., Redmann, E., and Siebert, C.H., J. Biomed. Mater. Res. 55,
530 (2001).
[10]
Cook, S.D., Dalton, J.E., Tan, E.H., Tejeiro, W.V., Young, M.J., and Whitecloud, T.S.,
Spine 19, 1856 (1994).
[11]
Hench, L.L, J. Am. Ceram. Soc. 81, 1705 (1998).
[12]
Wheeler, D.L., Eschbach, E.J., Hoellrich, R.G., Montfort, M.J., and Chamberland,
D.L., J. Orthop. Res. 18, 140 (2000).
19
[13]
Schepers, E., Barbier, L., and Ducheyne, P., Int. J. Oral. Max. Impl. 13, 655 (1998).
[14]
Synthes® Company, Synthes-Stratec/Norian SRS®, November 2002,
<http://www.synthes-stratec.com/html/NORIAN
[15]
SRS .148.0.html > (5 August 2003).
Huang, M., Feng, J.Q., Wang, J.X., Zhang, X.D., Li, Y.B., and Yan, Y.G., J. Mater.
Sci. Mater. Med. 14, 655 (2003).
[16]
Wang, M., Biomater. 24, 2133 (2003).
[17]
Yin, Y.J., Zhao, F., Song, X.F., Yao, K.D., Lu, W.W., and Leong, J.C., J. Appl. Polym.
Sci. 77, 2929 (2003).
[18]
Huard, J., Li, Y., Peng, H. and Fu, F.H., J. Gene. Med. 5, 93 (2002).
[19]
Lu, H.H., Kofron, M.D., El-Amin, S.F., Attawia, M.A., and Laurencin, C.T., Biochem.
Biophys. Res. Commun. 305, 882 (2003).
[20]
Lieberman, J.R., Daluiski, A., and Einhorn, T.A., J. Bone Joint. Surg. Am. 84A, 1032
(2002).
[21]
Wu, D., Razzano, P., and Grande, D.A., J. Cell. Biochem. 88, 467 (2003).
[22]
Alden, T.D., Pittman, D.D., Hankins, G.R., Beres, E.J., Engh, J.A., Das, S., Hudson,
S.B., Kerns, K.M., Kallmes, D.F., and Helm, G.A., Hum. Gene Ther. 10, 224 (1999).
[23]
Niyibizi, C., Baltzer, A., Lattermann, C., Oyama, M., Whalen, J.D., Robbins, P.D.,
and Evans, C.H., Clin. Orthop. Rel. Res. 355, S148 (1998).
[24]
Huang, L., and Viroonchatapan,
E., in "Nonviral Vectors for Gene Therapy,"
(L.
Huang, M.C. Hung, and E. Wagner, Eds.), Academic Press, San Diego, CA, 1999, p. 4.
[25]
Schratzberger, P., Krainin, J.G., Schrazberger, G., Silver, M., Ma, H., Kearney, M.,
Zuk, R.F., Brisken, A.F., Losordo, D.W., Isner, J.M., Molec. Ther. 6, 576 (2002).
[26]
Maruyama, H., Higuchi, N., Nishikawa, Y., Hirahara, H., Ilino, N., Kameda, S.,
Kawachi, H., Yaoita, E., Gejyo, F., and Miyazaki, J.I., Hum. Gene Ther. 13, 455
(2002).
[27]
Liu, F., and Huang, L., J. Control. Release 78, 259 (2002).
[28]
Gregory, L.G., Harbottle, R.P., Lawrence, L., Knapton, H.J., Themis, M., and
Coutelle, C., Molec. Ther. 7, 19 (2003).
[29]
Roy, I., Mitra, S., Maitra, A., and Mozumdar, S., Int. J. Pharm. 250, 25 (2003).
[30]
Howcroft, T.K., Kirshner, S.L., and Singer, D.S., Anal. Biochem. 244, 22 (1997).
20
[31]
Wang, D., Jing, N.H., and Lin, Q.S., Biochem. Biophys. Res. Commun. 225, 450
(1996).
[32]
Kumar, V.V., Singh, R.S., and Chaudhuri, A., Curr. Med. Chem. 10, 1297 (2003).
[33]
Kim, H.H., Lee, W.S., Yang, J.M., Shin, S., Biochim. Biophys. Acta-Mol. Cell Res.
1640, 129 (2003).
[34]
Azzam, T., Raskin, A., Makovitzki, A., Brem, H., Vierling, P., Lineal, M., and Domb,
A.J., Macromolecules 35, 9947 (2002).
[35]
Putnam, D., Gentry, C.A., Pack, D.W., and Langer, R., Proc. Natl. Acad. Sci. 98, 1200
(2001).
[36]
Luo, D., and Saltzmann, W.M., Nat. Biotechnol. 18, 33 (2000).
[37]
Niidome, T., and Huang, L., Gene Ther. 9, 1647 (2002).
[38]
Samuel, R.E., Lee, C.R., Ghivizzani, S.C., Evans, C.H., Yannas, I.V., Olsen, B.R., and
Spector, M., Hum. Gene Ther. 13, 791 (2002).
[39]
Bonadio, J., Smiley, E., Patil, P., and Goldstein, S., Nat. Med. 5, 753 (1999).
[40]
Desai, M.P., Labhasetwar, V., Walter, E., Levy, R.J., and Amidon, G.L., Pharm. Res.
14, 1568 (1997).
[41]
Foster, K.A., Yazdanian, M., and Audus, K.L., J. Pharm. Pharmacol. 53, 57 (2001).
[42]
Almofti, M.R., Harashima, H., Shinohara, Y., Almofti, A., Baba, Y., and Kiwada, H.,
Arch. Biochem. Biophys. 410, 246 (2003).
[43]
Langer, K., Balthasar, S., Vogel, V., Dinauer, N., von Briesen, H., and Schubert, D.,
Int. J. Pharm. 257, 169 (2003).
[44]
Bettinger, T., Remy, J.S., and Erbacher, P., Bioconjugate Chem. 10, 558 (1999).
[45]
Fontana, G., Licciardi, M., Mansueto, S., Schillaci, D., and Giammona, G., Biomater.
22, 2857 (2001).
[46]
Gref, R., Domb, A., Quellec, P., Blunk, T., Muller, R.H., Verbavatz, J.M., and Langer,
R., Adv. Drug. Deliver. Rev. 16, 215 (1995).
21
Chapter 2 - Synthesis, Cytotoxicity and Cellular Uptake of Organosilicate Nanoparticles
2.1. Introduction
Doped silicate materials such as Bioglass® have received much attention because of
their ability to bond quickly and strongly to bone [1]. Various compositions of these silicates
have been found to encourage osteoblast proliferation, and the bioactivity of these materials
has resulted in their clinical use as dental and cochlear implants [2].
We will attempt to
synthesize nanoparticles of similar compositions for use as gene delivery vehicles that can
enter bone cells without inducing cytotoxicity.
2.1.1. Doped Silicate Synthesis via Sol-Gel Processing
The first bioactive glasses were prepared via a melt-derived fusion method [3]. While
glasses prepared by this method have shown promise, more recent research has employed the
use of sol-gel chemistry [3,4]. Sol-gel processing is a wet-chemical synthesis technique that
can be used to fabricate inorganic oxides via the hydrolysis and polycondensation
of metal
alkoxides. This approach provides a great deal of flexibility in tailoring structural
characteristics,
such as compositional homogeneity, pore size, surface area, grain size, and
particle morphology. The first step in this reaction involves the hydrolysis of alkoxides,
-SSi-OR + H20 - -Si-OH + ROH
Polymerization results from the condensation reactions as follows,
-Si-OR + HO-Si- -
-Si-O-Si- + ROH
-Si-OH + HO-Si- -- -Si-O-Si- + H20
Lower pH conditions favor the hydrolysis reaction and gel formation, while higher pH
conditions favor condensation and particle formation. Organic functional groups may be
introduced via the sol-gel reaction of organically modified silicate precursors, R'nSi(OR) 4 -n
[5].
The presence
of organic
groups results in a more hydrophobic
material,
and
consequently,
greater protein adsorption when the material is exposed to a physiological
environment.
Enhanced protein adsorption may be more desirable in a bulk orthopedic
implant since protein adsorption is a precursor to osteoblast adhesion and matrix deposition
[6].
It may not be desirable for injectable nanoparticles, however, since excess protein
adsorption may result in subsequent clearance from the circulation [7].
22
Nevertheless, the
increase in surface hydrophobicity from organic functionalization may allow the nanoparticles
to pass through the phospholipid bilayer of the cell membrane more easily than hydrophilic
silica [8].
2.1.2. Initial Bioactivity Studies with Organosilicate Gels
While organosilicate gels synthesized via sol-gel chemistry have shown interesting
properties, their surface reactivity can be a detriment to biological applications.
Many
inorganic materials that are produced by sol-gel processing must undergo heat treatment to
fully condense the gel network and burn out solvents and residual organics.
While small
amounts of residual reactants may not affect many non-biological applications, they can be
extremely harmful in a sensitive biological environment. During our preliminary studies on
organosilicate monoliths, an increase in cell proliferation was noted upon the incorporation of
various salts. However, when these materials were placed in aqueous in vitro systems, any
unreacted precursors would alter the pH of the surrounding medium, resulting in cytotoxicity.
Thus,
the goal of this study is to synthesize
fully condensed
organosilicate
nanoparticles through low-temperature sol-gel processing. The chemical composition of these
nanoparticles will be modified so as to elicit a bioactive in vitro response similar to that of
Bioglass®.
The effects of various synthesis conditions on material properties and subsequent
cellular response will be examined.
2.2. Experimental
2.2.1. Synthesis of Organosilicate Microspheres
As described earlier, increasing the pH promotes the polycondensation
reactions and
particle formation. Synthesis of monodisperse silica particles is a well-documented process
that was initially developed by Stober et al. [9]. The sol-gel synthesis parameters used to
control particle size include precursor concentration and pH. We used a basic catalyst of
sodium hydroxide to produce organosilicate particles.
A high water/alkoxide ratio was
employed to ensure complete hydrolysis once the organosilane came into contact with the
aqueous reaction medium.
Monodisperse organosilicate microparticles were initially synthesized using a method
developed by researchers at the Shin-Etsu Chemical Company [10]. Methyltrimethoxysilane
23
(CH 3Si(OCH 3) 3, Gelest) was used as the organosilane precursor to synthesize micron-sized
polymethylsilsequioxane
(CH 3SiO 1. 5) particles. It was slowly added dropwise to a solution of
water and sodium hydroxide, and the particles were allowed to nucleate and grow. The
derived particles could then be surface functionalized using silane precursors with the desired
functional groups [10].
2.2.2. Synthesis of OrganosilicateNanoparticles
While micron-sized spherical particles were easily obtained by the above method, they
tended to agglomerate and were difficult to disperse.
To obtain dispersed nanoparticles, we
dissolved Tween-80® (Sigma), an FDA-approved surfactant, in the aqueous solution before
methyltrimethoxysilane addition. Tween-80® is a polyoxyethylene (20) sorbitan monooleate
surfactant with a hydrophilic-lipophilic
balance (HLB) of 15.0.
This surfactant-stabilized
suspension prevented particle growth and increased dispersion, resulting in much finer
particle sizes.
The following salts were introduced to achieve bioactive glass compositions: calcium
nitrate
tetrahydrate
(Ca(NO3) 2'4H0,2
Fluka),
magnesium
nitrate
hexahydrate
(Mg(NO 3) 2'6H20O, Fluka), sodium nitrate (NaNO 3, Sigma) and potassium nitrate (KNO 3,
Sigma). The salt was added to the particle suspension after the particles were aged for 1 hr.
After salt addition, the particles were allowed to age for - 24 hr before further processing.
2.2.3. Materials Characterization
Dynamic light scattering (DLS) experiments were performed at 250 C using a Lexel 95
Argon-ion laser and a Brookhaven high-precision photomultiplier at 900. Data were acquired
using a Brookhaven 9000AT correlator. The periods for acquisition of the autocorrelation
function were 0.1 gs to 0.5 s, and the particle size distributions were obtained from non-
negatively constrained least-squares (NNCLS) analysis of the data. Phase identification of
organosilicate powders was performed by powder X-ray diffraction (XRD) using a Siemens
D5000 0-0 diffractometer
(45 kV, 40 mA, Cu-Koc). Scanning electron microscopy (SEM)
was performed on a JEOL 6320FV field emission high-resolution microscope, and
transmission electron microscopy (TEM) was performed on a Philips EM 410 high-resolution
microscope.
Elemental analysis was performed by Desert Analytics, Inc. (Tuscon, AZ).
24
Surface characterization was performed with X-ray photoelectron spectroscopy (XPS) on a
Kratos AXIS Ultra Imaging X-ray Photoelectron Spectrometer using a Mono (Al) anode. The
binding energy range of 340-350 eV was analyzed to determine the amount of Ca species
present on the surface. An additional sample of CaSiO 3 (Alfa Aesar) was run as a standard.
Zeta potential measurements (Brookhaven ZetaPALS Zeta Potential Analyzer) were
performed on 0.15 wt% nanoparticle suspensions in cell culture medium. Solution pH was
measured using an Orion model 420A pH meter, and adjusted by the addition of KOH or HCI.
2.2.4. Cell CultureExperiments
Four different cell types were investigated for their response to organosilicate
nanoparticles.
The osteoblast cell culture experiments were performed with both an
immortalized (MC3T3-E1) cell line and with primary osteoblasts obtained from neonatal
mouse calvaria. Preliminary cell culture experiments were also performed on a cell line from
connective tissue (C3HIOT1/2 fibroblasts) and another cell line from non-connective tissue
(HepG2
hepatocytes).
The MC3T3-E1
cells obtained
from ATCC were cultured
in
Dulbecco's Modified Eagle medium (DMEM, Invitrogen) supplemented with 10% fetal
bovine serum (Hyclone) and 2% antibiotic-antimycotic (Invitrogen). Primary osteoblasts
obtained from neonatal mouse calvaria (courtesy of Hao Wang, Brigham and Women's
Hospital, Boston, MA) were cultured similarly, except that minimum essential alpha medium
(oc-MEM, Invitrogen) was used instead of DMEM.
The culture medium was also
supplemented with 1% ascorbic acid (Sigma), as recommended by the protocol.
C3H cells
(ATCC) were cultured with Basal Eagle Medium (ATCC) supplemented with 10% fetal
bovine serum (ATCC) and 1% antibiotic (ATCC). HepG2 cells (ATCC) were cultured with
Minimum Essential Eagle Medium (ATCC) supplemented with 10% fetal bovine serum and
1% antibiotic.
The particles were ultrafiltered three times in serum-free cell culture medium using an
Amicon Stirred Ultrafiltration Cell (Millipore) with a 50K polyethersulfone filtration
membrane (Millipore) before suspending in complete cell culture medium.
25
2.2.4.1. Proliferation Studies
50-mm polystyrene petri dishes were seeded at an initial cell density of 80 cells/mm 2 .
The petri dishes were incubated at 37°C in 5% CO 2 and the cells were allowed to attach
overnight.
suspension
On day 1, the medium was aspirated and replaced with a stock solution of particle
and cell culture medium.
Subsequent
cell proliferation
was measured by
passaging the cells with 0.25% Trypsin-EDTA (Invitrogen), lysing the cells and performing a
Cy-Quant DNA Assay (Molecular Probes). The proliferation of fibroblasts and hepatocytes
was measured
hemacytometer.
by staining
the cells with trypan blue and counting
the cells
in a
Tissue-culture polystyrene (TCPS) was used as a positive control.
2.2.4.2. Uptake Studies
Samples for cellular uptake were prepared by seeding cells in 75-cm2 tissue culture
flasks and adding the particles after day 1 of cell attachment. The cell-particle samples were
incubated for at least 24 hr to allow for maximum uptake. The cells were then washed once in
phosphate
buffered
saline
(PBS,
Invitrogen),
passaged,
and
resuspended
in a 2%
paraformaldehyde (Sigma) solution. They were allowed to incubate at room temperature for
2 hr in order to fix them in solution. The fixed cells were then stained with SYBR Green I
(Molecular Probes), a fluorescent probe that would emit in the FITC region upon binding to
silica nanoparticles.
Fluorescent particle uptake was quantified through fluorescence-
activated cell sorting (FACS) of the stained cell suspensions using a Facscan cytometer
(Becton-Dickinson)
equipped with an argon laser.
At least 10,000 events were counted for
each sample. The software package Cell Quest (Becton-Dickinson) was used to determine the
events of interest from forward and side scatter parameters. The mean fluorescence intensity
of the cells was obtained from the mean channel number of the fluorescence histograms of the
gated population.
Uptake values were calculated as shifts in the mean fluorescence of cell
populations relative to the control cell population with no particles.
Calibration curves for
particle fluorescence were generated by measuring fluorescence vs. particle wt% in an Fmax
microplate spectrofluorometer (Molecular Devices) using a 485-nm excitation/538-nm
emission filter. The particle fluorescence data were analyzed using Softmax PRO (Molecular
Devices), and all results were normalized to the fluorescence signal of fluorescent 10-nm CaSiO 2 particles.
26
For TEM studies, the cells were grown in Petri dishes and incubated with particles for
at least 24 hr. These cells were fixed in a 2% paraformaldehyde solution for 1 hr. The fixed
cells were
then
(C 2H 6AsNaO 2 3H
0, 2
scraped,
pelletized,
and washed
three times
in sodium
Electron Microscopy Sciences or EMS) buffer.
cacodylate
The washed pellets
were stained with 1% osmium tetroxide (OS0 4, EMS) solution and placed on ice in the dark
for 1 hr. The stained pellets were incubated in uranyl acetate ((CH 3 COO) 2UO22H0, 2
Sigma)
overnight, washed in H 20, and gradually dehydrated with 50%, 75%, 90%, 95% and 100%
ethanol washes. The dehydrated pellets were incubated in propylene oxide (Sigma) solution
overnight. They were embedded in Epon (EMS), and sectioned for microscopy.
2.3. Results and Discussion
2.3.1 Effect of Synthesis Conditions on Particle Size Distribution
2.3. 1.1. Effect of Surfactant Addition
While the formation of microspheres could be obtained by increasing the
water/alkoxide ratio and increasing the pH, monodisperse nanoparticles were difficult to
achieve by just varying these two parameters, since particle agglomeration and growth would
occur.
To prevent further particle growth, a biocompatible
surfactant was added as a
dispersing agent. SEM and DLS indicated that the surfactant addition resulted in a substantial
reduction in particle size from - 1.5 Rm to < 10 nm (Fig. 2.1). It was found that the surfactant
concentration did not affect the particle size. Thus, the surfactant concentration was kept at 6 wt%, which was well above the critical micelle concentration of 0.01 wt%.
(a)
1-
(b)
k-1
Figure 2.1. SEM images of organosilicate nanoparticles synthesized (a) without surfactant
(mean particle size = 1.5 rm) and (b) with surfactant (mean particle size = 7 nm).
27
2.3.1.2. Effects of Sol pH and Water/Alkoxide Ratio
The effects of both sol pH and alkoxide concentration on particle size are shown in
Figure 2.2. For a given water/alkoxide ratio, increasing the pH resulted in smaller particle
sizes. An increase in pH would increase the kinetics of the condensation reaction [11], and
the presence of the surfactant would prevent any further growth of the condensed particle. An
increased water/alkoxide ratio would promote nucleation versus particle growth, leading to
finer particle sizes.
12.b
12.0 50- 100
I a 1.0
-
iO nm
11.0 -
10.5
4
5
6
I
I
I
7
8
9
1C
Water/Alkoxide Ratio
Figure 2.2.
Effects of pH and water/alkoxide
ratio on the particle size distribution of
organosilicate nanoparticles.
2.3.1.3. Effect of Alkoxide Addition Rate
While the high sol pH greatly favors particle formation over gel formation, we found
that the addition rate of the alkoxide precursor under basic pH conditions greatly affected the
monodispersity of the particles produced. The results indicated that under certain conditions,
increasing the alkoxide addition rate led to an increase in particle size as well as a broader
particle size distribution. This phenomenon was most likely due to the competition between
nucleation and particle growth; increasing the alkoxide addition rate favored particle growth
since more alkoxide came rapidly into contact with the existing particles to promote growth
before the new precipitates had time to nucleate.
28
The effect of alkoxide addition rate was
more pronounced with finer particle sizes. To ensure a narrow particle size distribution, the
alkoxide addition rate was kept at a slow rate of - 0.1 ml/s.
2.3.1.4. Effect of Salt Addition
In order to incorporate
the bioactive components
composition, we attempted various methods of salt addition.
into the nanoparticle
surface
One method was to predissolve
the salts in the surfactant-stabilized aqueous solution before addition of the alkoxide. This
approach resulted in an extremely broad particle size distribution. Another method was to
form the particles first and then immediately add them to an aqueous salt solution.
This
method also resulted in polydisperse particles. From these studies, we inferred that salt
addition prior to sufficient particle growth would disturb the delicate balance between the
hydrolysis and polycondensation
reactions.
Thus, the salt addition step was introduced only
after sufficient aging of the particles ( 3 hr). This approach allowed the narrow particle size
distribution to be maintained. Some particle agglomeration and broadening of the particle
size distribution initially occurred due to the salt addition, which was to be expected [12].
However, after the dilution and sonication, the particles were effectively redispersed, and
there was only a slight broadening of their size distribution. Figure 2.3 shows the particle size
distributions for 10-nm particles before and after the addition of calcium nitrate.
29
(a)
100 -
" 80-
.2
_.
o
a)
a.
20
I
0
0
I
I
5
I
I
10
I
15
I
20
I
25
30
35
30
35
Hydrodynamic Diameter (nm)
(b)
100
80
.3
0.
0
a- 60
0
C)
0 40
')
o
34.
U)
IL
20
0
0
5
10
15
20
25
Hydrodynamic Diameter (nm)
Figure 2.3. DLS particle size distributions of 10-nm organosilicate nanoparticles (a) before
and (b) after 5 wt% Ca addition.
2.3.2. Effect of Nanoparticleson CellularResponse
Bioactive silicate glasses containing various combinations of calcium, magnesium,
and sodium oxides have been shown to stimulate cell proliferation [13-15]. To determine if
similar cellular response could be achieved with nanoparticles
of such salts, we have
examined the effects of nanoparticle composition on cell proliferation (see Fig. 2.4). While
MC3T3-E1 and primary cells exhibited similar trends, the osteoblast results discussed in this
30
chapter were associated with experiments done with primary cell cultures. This study showed
that while certain compositions enhanced cell proliferation more than others, Ca-containing
particles had the most positive effect on cell viability.
600
I.. .....
I...
....I......
I....
........
1:
Ca
E 500
E
T
(0
a) 400
Mg
I.-
_?! 'nn
K
,0
._
0)
200
en
100
0
Nanoparticles
No particles
Figure 2.4. Effect of the surface composition of 10-nm organosilicate nanoparticles on
osteoblast proliferation after 7 days. Nanoparticle loading: 0.15 wt%; nominal salt loading:
2.5 wt%. Values are mean + standard error of the mean; n = 3.
After the initial screening study, the Ca-containing organosilicate particles (Ca-SiO2)
were investigated in detail as a function of particle size and Ca loading. Figure 2.5 shows that
an increase in osteoblast proliferation was achieved with Ca-SiO 2 particles of < 100 nm with
low nominal calcium loadings.
600
N500
E
.'
400
>, 300
ID 200
0 100
0
5
10
15
20
Nominal Ca Loading (wt%)
Figure 2.5. Effect of nominal Ca loading and particle size of Ca-SiO2 on osteoblast
proliferation. Particle sizes of () 10 nm, () 50 nm, (A) 100 nm and () 200 nm are
examined at a particle loading of 0.15 wt%. Values are mean
= 3.
31
standard error of the mean; n
2.3.3. Characterization of Ca-SiO2 Nanoparticles
The composition
of the Ca-SiO 2 nanoparticles
was characterized
by elemental
analysis.
All materials characterization was done after particle ultrafiltration in cell culture
medium.
Table 2.1 shows the actual calcium contents in the 10- and 50-nm particles.
The
results indicated that the upper limit for total calcium loading was - 0.5 wt%.
Table 2.1. Elemental analysis of Ca-SiO2 nanoparticles.
Nominal Ca
Actual Bulk Ca Loading (wt%)
Loading (wt%)
10-nm Particles
50-nm Particles
0.50
1.25
2.50
0.13 +0.09
0.54 + 0.14
0.51 + 0.20
0.11 0.10
0.57 + 0.12
0.53 + 0.18
The particles were amorphous by XRD analysis, which was to be expected since the
effective Ca loading was very low.
XPS was used to provide insight into the surface
composition of Ca-SiO 2. Table 2.2 shows the surface Si/Ca ratios as measured by XPS. The
effective surface Ca loadings calculated from XPS results were higher than the bulk Ca
loadings shown in Table 2.1. This indicated that the Ca species were mainly located on the
surface of the nanoparticles, which was to be expected since the Ca precursor was introduced
to the organosilicate nanoparticles after the latter were formed and aged for > 3 hr.
Table 2.2. XPS data for 50-nm Ca-SiO2 nanoparticles.
Nominal Ca
Surface Si/Ca
Surface Ca
Loading (wt%)
0.00
1.25
2.50
Atomic Ratio
89.6
2.43
2.56
Loading (wt%)
0.39
6.50
7.13
The XPS data also indicated that the phosphorous loading on the particles was < 0.5
wt%, suggesting that the formation of a calcium phosphate layer on the surface of the
particles due to reaction with phosphate ions from cell culture medium was highly unlikely.
While bioactive glasses would typically form a calcium phosphate layer on their surface, the
time scale for that reaction to occur on the surface of these particles would be - 2 weeks, as
indicated by XPS (data not shown). This was in fair agreement with Kokubo et al. [16], who
32
reported that formation of a calcium phosphate layer on the surface of sodium silicate glasses
took about 240 hr.
2.3.4. Effect of Ca-SiO2Nanoparticleson Cellular Uptake
The effect of Ca-SiO2 nanoparticles on cellular uptake was also studied. TEM results
indicated that 10-, 50- and 100-nm Ca-SiO 2 particles were taken up by cells (Fig 2.6).
Multiple micrographs illustrated that cellular uptake was extremely selective; so that only
particles with surface dopants of Ca, Mg and Na were ingested, while undoped SiO 2 particles
were not.
For Ca-SiO 2 nanoparticles, only particles of < 100 nm were ingested, while for
Mg- and Na-SiO 2 particles, only particles of < 50 nm were ingested.
Nanoparticle uptake was quantified by the FACS assay so as to study the effects of
various parameters on cellular uptake. Since particle uptake would reach a saturation after a
certain particle loading (Fig. 2.7), all subsequent experiments were performed at a particle
loading of 0.15 wt% to ensure that the experiments were conducted in this region of
maximum uptake.
Figure 2.8 shows the effect of calcium loading on the cellular uptake of Ca-SiO2
particles of various sizes.
The FACS data supported the TEM findings, showing that
organosilicate particles of < 100 nm with nominal calcium loadings of > 1 wt% were ingested
by osteoblasts. Figure 2.9 shows that cell proliferation increased for Ca-SiO 2 nanoparticles
with low nominal Ca loadings. This could be attributed to particle ingestion by cells. At
higher nominal Ca loadings (> 10 wt%), however, cellular uptake (Fig. 2.8) did not
correspond to cell proliferation (Fig. 2.9).
33
a)
w.
i:; -
5
UA.
..
~
·
..
..
l
..
b*)
~','
,,,e;;.
\5!~
500 nm
'
b)
;;
...
k
' '4
%
s
' '
,
500 nm
c)
A
A*.w
k
WN
q
S
.,
Figure 2.6. Transmission electron micrographs of primary osteoblasts that have ingested (a)
10-nm, (b) 50-nm and (c) 100-nm Ca-SiO 2 nanoparticles containing 2.5 wt% Ca. Particle
0.15
loading: 0.
15 wt%.
34
_
fI:
OUU
.
500
D
400
LL
O 300
- 200
100
SO
0.00
0.05
0.10
0.15
0.20
Particle Loading (wt%)
Figure 2.7. Effect of particle loading on the cellular uptake of ()
0.25
10-nm, () 50-nm and (A)
100-nm Ca-SiO 2 particles. All particles had a nominal Ca loading of 2.5 wt%.
mean ± standard error of the mean; n = 3.
"7
IUU
Values are
.% ..
600
. 500
LL.
Cr 400
m 300
200
100
0
0
5
10
15
Nominal Calcium Loading (wt %)
20
Figure 2.8. Effect of Ca loading and particle size on the cellular uptake of ()
10-nm, () 50-
nm, (A) 100-nm and (e) 200-nm Ca-SiO 2 particles. Particle loading: 0.15 wt%. Values are
mean standard error of the mean; n = 3.
35
i-
%^
3DUU
cm
E
400
E
in
'
0 300
-
',
200
0=
100
0
0
5
10
15
20
Nominal Calcium Loading (wt%)
Figure 2.9. Effect of Ca loading and particle size on the osteoblast proliferation for () 10-nm,
(*) 50-nm and (A) 100-nm Ca-SiO 2 nanoparticles. Particle loading: 0.15 wt%; n = 3.
The decrease in proliferation at higher nominal Ca loadings could be due to the greater
degree of agglomeration caused by higher salt loadings. This agglomeration led to broader
particle size distributions (see Fig. 2.10), so despite the fact that there were enough small
particles at a particle loading of 0.15 wt% to result in maximum uptake, there were probably
also some larger particles that did not enter the cell and simply depleted protein from the cell
culture medium, resulting in a slight decrease in cell proliferation.
36
(a)
.1
......
I....
I....
.........
......
........
I....I.
...........
--.....
........
...--......-.....
. - . -I..-,-,-,
- ,-,-.-..,..........
. .......
I UU
..-.1--- -.- -.- ,
80
OC
,,
ou
0
a.
At
0
40 -
20
0
l
I
0
1
50
Il
100
*
I
150
Hydrodynamic
(b)
I
i
200
250
300
Diameter (nm)
100
0
o
a.
0
a-
80
60
o
0
-
40
C
a)
L
20
a0
I
....
1
50
100
150
I
0
-
nI
-T
D
I........
H
n
1
200
250
300
Hydrodynamic Diameter (nm)
Figure 2.10. DLS particle size distributions of 100-nm organosilicate nanoparticles (a) before
and (b) after 15 wt% Ca addition.
2.3.5. Intracellular vs. ExtracellularSignaling
The increase in osteoblast proliferation might be because particle ingestion induced a
paracrine response, prompting cells to proliferate faster [17]. In order to test this hypothesis
of extracellular signaling, an experiment was performed with conditioned medium. First, the
cells were cultured in the presence of nanoparticles (particle loading: 0.15 wt%) and allowed
to proliferate.
The conditioned medium was collected and then used to culture a fresh batch
37
of cells in the presence of nanoparticles. Presumably, if this conditioned medium contained
signaling molecules released by the cells, the same molecules would also cause the fresh
batch of cells to proliferate at an increased rate.
However, as Fig. 2.11 indicates, cells
.....
cultured in the presence of the conditioned medium proliferated at a normal rate. This result
suggested
that the increase
in proliferation
was not associated
mechanism, but rather with an intracellular mechanism.
600
....
LI Fresh medium
NV
E
with an extracellular
E Conditioned medium
500
E
400
U
.)
= 300o
T
s
CT
-
I
200
100
C:
_
No Particles
I
Ii
i
SiO 2
Ca-SiO2*
Figure 2.11. Osteoblast proliferation in the presence of fresh and conditioned cell culture
media. Values are mean + standard error of the mean; n = 3. *Nominal Ca loading: 2.5 wt%.
Once the particles entered the cells, some of them were surrounded by vesicles,
possibly endosomal vesicles (see TEM images in Fig. 2.6). Endosomal pH is < 6, slightly less
than that of the cytoplasm (pH = 7.4) [18]. Under these slightly acidic conditions, the once
insoluble calcium on the surface of the particles might become slightly soluble and Ca2+ could
be released.
XPS analysis before and after acidification indicated that the surface calcium
loading was 2.43 wt% at pH = 5.5 for 50-nm particles, which was dramatically lower than the
Ca loading at pH = 7.4 (7.13 wt%) (see Table 2.2).
Ca2 + is ubiquitous in various cellular processes from contraction to cell proliferation
(see Fig. 2.12).
Ca 2+ serves as an intracellular messenger in many signaling pathways, and
cellular responses to changes in Ca2 + levels depend on both the cell type and the nature of
38
stimulation [19]. An intracellular increase in Ca 2+ due to the presence of the particles may
lead to a cascade of events involved in a mitogenic pathway.
Grco;th tactors
Hornmones
.. Ney
+-:'':""+"'+'.'ww"
NJeurotrarnsmitters
I.++,:
- -.1....
..1- . I......I...........
%-, 11
+.t1,1.1'.-1,-'
"",,
':::R
2&,M
iA
..a,'110,'*'+''
--,, " +..--''.66. . ;.+C
I 61 .1"··.·:
-'11
1
F, :-g
-,
0
ll-lIl:rl -,PLC
, ea; l.
( a n .%
t'dli
14,D
t4.b~~[t
,,,1-
all~8
+4
4*
lI I Y
7s1$
i
'
7.4I
I
1---
Endop lasrn ict
reticulu ll
E>
4.% k.
*..
r.,~.
Mito ho d rion
Bcl-2 -
I
'7'
F.rilization
Pro liferationor
Sr:;cration
M.1':tabolisn
C::ntr action
-
Ap° pltosis
Figure 2.12. Basic mechanisms of calcium signaling [20].
2.3.6. Effect of Ca-SiO2 Nanoparticles on Other Cell Types
While the increase in cell proliferation due to ingestion of Ca-SiO 2 nanoparticles was
evident in osteoblasts, this phenomenon did not extend to other cell types. Cell proliferation
in fibroblasts was fairly unaffected by the presence of Ca-SiO 2 nanoparticles, as shown in Fig.
2.13. On the other hand, the presence of Ca-SiO 2 nanoparticles led to decreased proliferation
of hepatocytes, which might be due to a rise in Ca2 + , since an increase in Ca2 + above a certain
threshold could lead to cell death.
This threshold might vary depending on the cell type.
Since Ca2 + played a significant role in mineralization and bone formation, the threshold for
Ca2 + concentrations could be higher in osteoblasts than in hepatocytes [20,21].
This might
allow us to target Ca-SiO 2 nanoparticles for gene delivery to osteoblast cells selectively.
39
---
ZbU
.
.....
...............
..................
.............
.
...I........
i
200
._
'-
150
"
100
50
50
F_17
I
Osteoblasts
Fibroblasts
Hepatocytes
Figure 2.13. Proliferation of osteoblasts, fibroblasts and hepatocytes in the presence of () 10nm, ()
50-nm and ()
loading: 2.5 wt%.
100-nm Ca-SiO 2 nanoparticles.
Particle loading: 0.15 wt%; Ca
2.3.7. Effect of Other NanoparticleCompositionson OsteoblastUptakeand Proliferation
A slight increase in osteoblast proliferation was noted with Mg-SiO2 particles, but not
with Na-SiO 2 particles (Fig. 2.14). There has been some prior experimental evidence that
osteoblasts would respond to cations other than calcium [22-23].
Mailland et al. have
suggested that this "cation-sensing mechanism" might participate in the regulation of the
skeletal apposition of calcium circulating in the plasma [24].
Compared to Ca-SiO 2 particles, Mg-SiO 2 and Na-SiO
2
particles gave rise to lower
osteoblast proliferation and uptake. We noted that in all three systems, finer particle sizes led
to higher cell density and cellular uptake. Mg-SiO 2 and Na-SiO 2 particles of < 100 nm and <
50 nm, respectively, were taken up by osteoblasts (see Fig. 2.15). The presence of Mg, Na
and Ca all resulted in a slight decrease in the negative surface charge of the organosilicate
nanoparticles (Fig. 2.16). This decrease might have facilitated particle attachment to the cell
membrane and subsequent endocytosis, if the particles were small enough in dimension.
40
-T\T\
....................................................
...............................
.......
............
......... -. .......
..-................. . ...
uu
.I.. .
...I....
........
I...I.....I.....i
500
E
E
>' 400-
~~_._
4-_--
>300
u,
200
100
l
-
-
I
0
1
50
100
1
150
200
Particle Size (nm)
Figure 2.14. Effect of () Ca-SiO 2, () Mg-SiO 2 and (A) Na-SiO 2 nanoparticle size on the
proliferation of osteoblasts.
Particle loading: 0.15 wt%; nominal salt loading: 2.5 wt%.
Values are mean + standard error of the mean; n = 3. Note: cell density in the absence of
nanoparticles is - 230 cells/mm 2 .
600
500
'
U-
400
J
300
-
200
100
0
0
50
100
150
200
Particle Size (nm)
Figure 2.15. Effect of ()
Ca-SiO 2, () Mg-SiO 2 and (A) Na-SiO
2
nanoparticle size on the
uptake of osteoblasts. Particle loading: 0.15 wt%; nominal salt loading: 2.5 wt%. Values are
mean ± standard error of the mean; n = 3.
41
E
ax
0
I.
0
N
2
3
4
5
6
7
8
9
10
pH
Figure 2.16. Zeta potential vs. pH for nanoparticle suspensions of () Ca-SiO 2, (A) Mg-SiO2,
(e) Na-SiO 2 and () SiO 2 in complete cell culture medium.
nominal salt loading: 2.5 wt%.
Particle loading: 0.15 wt%;
2.3.8. Calcium Transportersin the Osteoblast
While the slight decrease in the negative particle surface charge due to the presence of
Ca2+ might facilitate particle attachment and subsequent endocytosis, the increase in surface
charge alone would not explain the enhanced uptake of the Ca-SiO2 nanoparticles. If surface
charge was the only mechanism, then uptake for Mg-SiO 2 and Ca-SiO 2 nanoparticles would
be similar since they displayed similar zeta potentials at pH = 7.4 (Fig. 2.16). While much
research has been done on the transport systems of cells from various tissues such as the
nervous system, cardiac and skeletal muscle, and epithelia [25], information on ion pumps,
transporters and channels in bone cells still remains limited compared to other tissues. This
lack of information on bone cells is largely because the methods for isolation and culture of
bone cells have only been established recently [26], whereas the other tissues mentioned
earlier can be manipulated with more ease. Recent research done by groups such as Francis et
al. has just begun to elucidate some of the transport mechanisms involved in bone cells [21].
There is a particularly high density of Ca-dependent transporters involved in osteoblasts (Fig.
2.17), suggesting the relative importance of calcium in osteoblast mineralization and
subsequent bone formation.
The presence of Ca-dependent tranporters indicates that
42
mechanisms other than non-specific endocytosis may be involved in the transport of Ca-SiO2
particles across the plasma membrane. While the specific mechanism of cellular uptake is not
explored in detail in this thesis, the enhanced uptake of these Ca-SiO 2 nanoparticles would be
very useful towards the delivery of DNA to osteoblasts.
ROJOiER.
Ca
Ca
NKa
t
OSTEOBLAST
NK'
Na
s
4
-
Cu
Na Ca
:.'
CB
c
N
Na...
,
H
Ca
BONE MARROW
Figure 2.17. Transport systems in osteoblasts [21].
2.4. Summary
We have synthesized monodisperse organosilicate particles of various sizes and
compositions via alkoxide sol-gel processing. The effects of these nanoparticles on osteoblast
response
were investigated,
and Ca-SiO 2 nanoparticles
were found to be taken up by
osteoblasts and to stimulate osteoblast proliferation. The increase in proliferation was found
to be osteoblast-specific
and did not extend to fibroblasts or hepatocytes.
The significant
uptake of Ca-SiO 2 nanoparticles by osteoblasts and the resulting enhancement in osteoblast
43
proliferation suggest that these nanoparticles are well-suited for gene delivery applications to
bone cells.
2.5. References
[1]
Hench, L.L., and Paschall, H.A., J. Biomed. Mater. Res. 7, 25 (1973).
[2]
Hench, L.L., Am. Ceram. Soc. Bull. 72, 93 (1993).
[3]
Sepulveda, P., Jones, J.R., and Hench, L.L., J. Biomed. Mater. Res. 58, 734 (2001).
[4]
Ramila, A., Balas, F., and Vallet-Regi, M., Chem. Mater. 14, 542 (2002).
[5]
Schmidt, H.K., in "Inorganic and Organometallic Polymers," (M. Zeldin, K.J. Wynne,
and H.R. Alcock, Eds.), American Chemical Society, Washington, D.C., 1988.
[6]
El-Ghannam, A., Ducheyne, P., and Shapiro, I.M., J. Orthop. Res. 17, 340 (1999).
[7]
Gref, R., Domb, A., Quellec, P., Blunk, T., Muller, R.H., Verbavatz, J.M., and Langer,
R., Adv. Drug. Deliver. Rev. 16, 215 (1995).
[8]
Tamai, I., and Tsuji, A., J. Pharm. Sci. 89, 11 (2000).
[9]
Stober, W., Fink, A., and Bohn E., J. Coll. Interf Sci. 26, 62 (1968).
[10]
Shimizu, T., Okon, T., Ohba, T., and Inokuchi, Y., "Process of Preparing Surface-
Modified Polymethylsilsequioxane Spherical Fine Particles," US Patent 5149748
(1992).
[11]
Brinker, C.J., and Scherer, G.W., "Sol-Gel Science," Academic Press, San Diego, CA,
1990.
[12]
Iler, R.K., "The Chemistry of Silica," John Wiley & Sons, New York, 1979, p. 382.
[13]
Hench, L.L., J. Am. Ceram. Soc. 81, 1705 (1998).
[14]
Kim, H.M., Miyaji, F., Kokubo, T., Ohtsuki, C., and Nakamura, T., J. Am. Ceram.
Soc. 78, 2405 (1995).
[15]
Jallot, E., Benhayoune, H., Kilian, L., Irigaray, J.L., Barbotteau, Y., Balossier, G., and
Bonhomme, P., J. Coll. Interf Sci. 233, 83 (2001).
[16]
Takadama, H., Kim, H.M., Kokubo, T., and Nakamura, T., Chem. Mater. 13, 1108
(2001).
[17]
Dixon, S.J., and Sims, S.M., Drug Devel. Res. 49, 187 (2000).
[18]
Rudenko, G., Henry, L., Henderson, K., Ichtchenko, K., Brown, M.S., Goldstein, J.L,
and Deisenhofer, J., Science 298, 2353 (2002).
44
[19]
Berridge, M.J., Trends Pharmacol. Sci. 15, 419 (1980).
[20]
Berridge, M.J., Bootman, M.D., and Lipp, P., Nature 395, 645 (1998).
[21]
Francis, M.J.O., Lees, R.L., Trujillo, E., Martin-Vasallo,
P., Heersche, J.N.M., and
Mobasheri, A., Int. J. Biochem. Cell B 34, 459 (2002).
[22]
Jones, T.R., Antonetti, D.L., and Reid, T.W., J. Cell. Biochem. 30, 31 (1986).
[23]
Quarles, L.D., Wenstrup, R.J., Castillo, S.A., and Drezner, M.K., Endocrinology 128,
3144 (1991).
[24]
Mailland,
M., Waelchli,
R., Ruat, M., Boddeke,
H.G.W.M.,
Endocrinology 138, 3601 (1997).
[25]
Hasselbach, W., Ann. New York Acad. Sci. 853, 1 (1998).
[26]
Gundle, R., and Beresford, J.N., CalcifJ Tissue Int. 56, S8 (1995).
45
and Seuwen,
K.,
Chapter 3 - Osteoblast Protein Expression in Response to Organosilicate Nanoparticles
3.1. Introduction
While cellular proliferation is a fundamental indicator of material performance in cell
culture systems, the expressions of cellular proteins and other biochemical markers are also
important parameters by which to evaluate material biocompatibility [1]. Once osteoblast
cells have ingested nanoparticles, it is essential that these cells retain their normal metabolic
function in order to synthesize new bone.
This metabolic function can be evaluated by
quantifying the osteoblast expression of proteins that are involved in the bone formation
process. In addition, the expression levels of certain cytokines, or inflammatory markers, can
be quantified
to assess any potential trauma that the cells might experience
due to
nanoparticle uptake. The expression of certain cytokines often leads to bone resorption and
fibrous tissue formation [2], so expression of such cytokines should be minimal.
The goal of this chapter will be to study the osteoblast expression of a wide variety of
proteins in response to different nanoparticles in order to further elucidate the cellnanoparticle interactions in our gene delivery system. The proteins of relevance in our study
will range from those involved in bone mineralization to those expressed during inflammatory
response.
3.1.1. Proteins Involved in Bone Formation
The first subset of proteins we examined is specific to the osteoblast and pertinent to
new bone formation.
Bone-specific alkaline phosphatase (ALP) is a tetrameric glycoprotein
that is found on the surface of osteoblasts [3]. The function of the ALP enzyme is not clearly
understood, although it has been shown to be a biochemical marker of bone turnover.
ALP
expression is also considered to be a strong indicator of osteoblastic phenotype [4-6]. We will
measure the ALP expression of osteoblasts in response to different nanoparticles to ensure
that the osteoblastic phenotype is retained after particle ingestion.
Another bone-specific protein studied is osteocalcin. Osteocalcin is a vitamin Kdependent protein produced by osteoblasts. This protein contains residues that are thought to
be involved in calcium ion and hydroxyapatite
binding [7].
While the specific role of
osteocalcin in vivo is not known, its affinity for bone mineral constituents implies a role in
46
bone formation [7]. Osteopontin is another protein expressed by matrix-producing osteoblasts.
It is an extracellular glycosylated bone phosphoprotein, which binds calcium and interacts
with the vitronectin receptor that is involved in cell adhesion [8]. In addition to osteopontin
production, we will also study the osteoblast expression of fibronectin.
Fibronectin is a
glycoprotein with various binding sites that allow it to act as an integrating protein in the
extracellular matrix of osteoblasts [9].
3.1.2. Cytokine Expression
Cytokines are a large array of proteins that are secreted by a cell for the purpose of
altering either its own functions (autocrine effect) or those of adjacent cells (paracrine effect)
[10]. The cytokine network is extremely important in the regulation of inflammatory and
immune response. Different cytokines may have multiple biological activities. We examined
the expression of a variety of cytokines, including interleukins such as IL-2, IL-4 and IL-6, as
well as Tumor Necrosis Factor (TNF-a).
Since consistently elevated levels of various
cytokines would be indicative of a potential inflammatory response in vivo, expression of
these cytokines should be minimal for a biocompatible gene delivery system. Inflammation
has been known to be induced when crystalline silica-based particles are inhaled [11-13]. In
contrast, amorphous silica has widespread use in the food and pharmaceutical industries due
to its non-toxicity and bioinertness [14-16].
For this reason, our amorphous organosilicate
nanoparticles are expected to elicit minimal inflammatory response, if any.
3.2. Experimental
Cell culture experiments were performed on primary osteoblasts and C3H10T1/2
fibroblasts. Cells were cultured according to established protocols (see Section 2.2) [17]. All
absorbance measurements were read in a VersaMax microplate spectrophotometer (Molecular
Devices), and the absorbance data were analyzed using Softmax PRO software (Molecular
Devices).
For ALP activity measurements, the protocol of Sampath et al. was followed [18]. Cells
were seeded in 24-well culture plates at a density of 80 cells/mm2 and incubated overnight.
On day 1, the nanoparticle suspensions were added at a concentration of 0.15 wt%. After one
week, the cells were passaged and lysed. 50 jiL of cell lysates were assayed for enzymatic
47
activity in 96-well plates with p-nitrophenyl phosphatase (p-NP, Sigma) as a substrate. After
60 min of incubation at 37°C, the reaction was stopped, and the absorbance was measured at
405 nm with p-nitrophenol as a standard. ALP activity was expressed as nanomoles of p-NP
produced by 1 [tg of protein/hr. The protein concentration of the samples was determined by
a total protein assay (Pierce).
ALP staining was done by seeding cells in 8-well chamber slides and incubating
overnight.
On day 1, nanoparticle suspensions were added at a concentration of 0.15 wt%.
After one week, the cells were fixed in 2% paraformaldehyde and stained for ALP using a
Fast Red Violet Liquid Substrate Kit (Sigma). The stained cells were viewed using an optical
light microscope (Nikon).
Osteocalcin measurements were obtained using an Enzyme-Linked Immunosorbent
Assay (ELISA) kit for mouse osteocalcin (Biomedical Technologies, Inc.) For protein
measurements, cells were seeded in 75-cm 2 culture flasks and incubated overnight. On day 1,
nanoparticle suspensions were added at a concentration of 0.15 wt%.
After one week, the
medium was collected for testing. This assay used a sandwich-ELISA format. 25 [pL of
samples were added to a 96-well plate coated with antibody. The plate was then incubated for
2.5 hr at 37°C.
The medium was then aspirated and the wells were washed three times.
Another antibody was added and allowed to incubate for 30 min at room temperature. After
washing, 100 [tL of substrate were added to the samples and allowed to react. After 15 min of
reaction, 100 pL of stop solution were added and the absorbance was measured at 450 nm.
For osteopontin measurements, a very similar procedure was performed using a mouse
osteopontin ELISA kit obtained from Assay Designs, Inc.
This assay also involved a
sandwich-ELISA experiment, with different incubation times as specified by the protocol.
Fibronectin expression was measured using a human fibronectin ELISA kit
(Quantimatrix).
Rather than using a sandwich-ELISA format like the osteocalcin and
osteopontin kits, this assay uses a competitive inhibition ELISA.
Samples were diluted and
preincubated with a polyclonal rat antibody to human fibronectin, which binds with any
fibronectin present in the samples. 100-gtL sample mixtures were then transferred to a human
fibronectin coated plate. After incubating at room temperature for 1 hr, the wells were
washed four times with wash buffer. 100 pL of a Goat anit-Rabbit IgG-HRP Conjugate were
then added to each well and incubated at room temperature for 30 min.
48
The wells were
washed before the addition of 100 pL
of substrate to each well.
1
After the reaction was
complete, 100 [L of stop solution were added to each well and the absorbance was measured
at 450 nm.
For the cytokine measurements, sandwich-ELISA kits (Ebioscience) were used to
detect all interleukin levels.
coated with 100
The general procedure was as follows: a 96-well plate was
L/well of capture antibody.
The plate was then sealed and incubated
overnight at 4 0 C. The wells were then aspirated and washed three times, and then blocked
with 200 tL/well of assay diluent. The blocked wells were incubated at room temperature for
1 hr. After incubation, the wells were washed and then 100 [tL of sample were added to each
well. Binding of any free cytokine to the capture antibody was allowed to occur over the next
2 hr at room temperature.
After binding, the wells were washed, 100 i[L/well of detection
antibody were added, and the plate was incubated for 1 hr. The wells were then washed and
100 tL/well of a second antibody were added and allowed to incubate for 30 min at room
temperature.
Finally, the wells were washed, 100 jiL/well of substrate solution were added,
50 FL/well of stop solution were introduced, and the absorbance of the plate was read at 450
nm.
3.3. Results and Discussion
3.3.1. ALP Activity Measurements
The effect of various nanoparticles on ALP expression was measured.
As with the
proliferation and uptake experiments in the previous chapter [17], all protein expression
experiments were done at a nanoparticle loading of 0.15 wt% to ensure maximum uptake. All
control samples were cells grown under normal conditions without any particles.
Fig. 3.1
shows that the ALP activity of primary osteoblasts incubated with nanoparticles of different
compositions remained fairly constant. In other words, the phenotype of osteoblasts was
retained in the presence of these nanoparticles.
49
0.5 -
0.4-
X,
0.3-
.0
········-·...:
·-··.·
z 0.2-
·.::··:1·1.:·:1::·.
E
E
a
:....:...
.:
.:-·::··::.··.·.:.
0.1 -
·.·i:.: r···::
·· ·· r··
·· ···-·
,·i,·,,·ii··,·,.,·,,.
u.U
-
-
-
,,%',,,
i,,!
-
----
Control
"i%'[
-----;i
-"-,,
!·---
ihW"
SiO 2
Ca-SiO 2
Mg-SiO 2
......
i'?'
Na-SiO 2
Figure 3.1. Effect of nanoparticle composition on the ALP activity of osteoblasts. Particle
size: 50 nm; nominal salt loading: 2.5 wt%.
Values are expressed as nanomoles of p-NP
produced by 1 tg of total protein during 1 hr. Values are mean + standard error of the mean;
n = 3.
To study the effect of the Ca-SiO 2 system in more detail, ALP activity was also
measured in the presence of Ca-SiO 2 nanoparticles
of different sizes (Fig. 3.2).
The
phenotype of osteoblasts was retained regardless of the Ca-SiO2 particle size.
t'
-
U.b - ............
.............. I .
...
...
..........................................
I.-........." -
........ - .:
...I.....I..............
.................................
s 0.4.-
z0. 0.2 0
E 0.1 6c
0.0
--1-
2-
"
Control
I
.
'
I
I-
-
10 nm
50 nm
100 nm
Figure 3.2. Effect of Ca-SiO 2 particle size on the ALP activity of osteoblasts. Nominal Ca
loading: 2.5 wt%. Values are expressed as nanomoles of p-NP produced by 1 jig of total
protein during 1 hr. Values are mean + standard error of the mean; n = 3.
50
These measurements also indicated that osteoblastic phenotype was retained
regardless of particle uptake, which was different for different nanoparticle composition and
size. This consistency of ALP expression after particle ingestion was a significant finding,
since it was imperative that these osteoblasts retained their phenotype in order to continue
their function of synthesizing healthy bone. Figure 3.3 optically depicts the retention of
osteoblastic phenotype in the presence of Ca-SiO2 nanoparticles. The cells stained positive
for ALP after nanoparticle ingestion.
(ae) "
.
.1,
.
. i.
~~~~~~b)~~~~~·:·'
:
.'Z,...
~:
...
.
,
.,.
..
S
I
:av" t~~A,4
6:?"
:··: .·a·
X:.I
':
"
··.
'
a·m·
:····
o-
CLlrdj
:::;
Figure 3.3. Optical micrographs of osteoblasts (a) before and (b) after incubation with 50-nm
Ca-SiO2 nanoparticles. The darker regions indicate ALP expression.
3.3.2. OsteocalcinExpressionMeasurements
Figure 3.4 shows the effect of nanoparticle composition on osteocalcin expression. All
samples continued to secrete osteocalcin in the presence of nanoparticles, with a noticeable
increase in osteocalcin expression in the Ca-SiO 2 samples. A rise in osteocalcin expression is
often associated with a mature osteoblastic phenotype and the onset of mineralization [19],
both of which are desirable for bone regeneration. The effect of calcium loading for Ca-SiO2
nanoparticles on osteocalcin expression was not significant (data not shown), but there was a
slight effect of particle size, with 50-nm particles showing higher osteocalcin levels (Fig. 3.5).
Like the ALP activity findings, these results were encouraging
since they showed the
retention of osteoblastic phenotype and normal osteoblastic activity in the presence of
organosilicate-based nanoparticles.
51
ii
30
E
25
0)
S
20
-j
15
C
0 10
(0
0
5
n
v
I1
Control
SiO 2
Ca-SiO
2
Mg-SiO
2
Na-SiO 2
Figure 3.4. Effect of nanoparticle composition on the osteocalcin expression of osteoblasts.
Particle size: 50 nm; nominal salt loading: 2.5 wt%. Values are expressed as proteins
produced by 106 cells. Values are mean + standard error of the mean; n = 3.
"U
l
25
E
c 20
3 15
o
0
C)
o
5
0
Control
10 nm
50 nm
100 nm
Figure 3.5. Effect of Ca-SiO2 particle size on the osteocalcin expression of osteoblasts.
Nominal Ca loading: 2.5 wt%.
Values are expressed as proteins produced by 106 cells.
Values are mean + standard error of the mean; n = 3.
3.3.3. OsteopontinExpressionMeasurements
The effect of nanoparticle composition on osteopontin expression is shown in Fig. 3.6.
The experiments indicated that the ingestion of nanoparticles did not result in any significant
change in osteopontin production. Osteopontin expression was also unaffected by Ca-SiO2
particle size (Fig. 3.7). These results suggested that the doped SiO 2 nanoparticle-based gene
52
delivery system would allow for normal levels of osteopontin production and potential bone
formation.
nr%
3UU 250 -
T
T
E
...
- 200
:~......
....
f~~
..-..
..·,.·· ,.., .., <,
:. .' -~ . .
150 -
'.'.'~:·;Nl:.
._
Co
/
·
:.:.
· ..·
:" "TV......
100 -
.:.......
::·
.:.::
· ·, . 2r::.
8
:· .
...
50 '·'··
0
-
-
1_1:___
_
Control
.-
.
-:
:
-
-
SiO 2
u
Ca-SiO 2
Mg-SiO 2
Na-SiO 2
Figure 3.6. Effect of nanoparticle composition on the osteopontin expression of osteoblasts.
Particle size: 50 nm; nominal salt loading: 2.5 wt%. Values are expressed as proteins
produced by 106 cells. Values are mean ± standard error of the mean; n = 3.
300 E
T
250 -
0I,
a
T
:I::
:,·
200 -
··
`·:'·:`:'··.:
'::I··"'··
·
._ 150 "':::::":.`·I
0
e.
BC)
o
100 -
`''''' '"
.·'
::
I·:
...r·. -·,:·.1·
:·ci·c·:·:s:
·
50 -
i···
········
;.:i;::::::
·
::':
:.:·,..:::::·(··:::,:::
::::
... .::.(:(·:::-:·:
:··':':·.·::,'I·::i:i:
·:··:-·
0
I
-
-
Control
-
10 nm
50 nm
i
·-·····
······-
100 nm
Figure 3.7. Effect of Ca-SiO2 particle size on the osteopontin expression of osteoblasts.
Nominal Ca loading: 2.5 wt%. Values are expressed as proteins produced by 106 cells.
Values are mean ± standard error of the mean; n = 3.
53
3.3.4. Fibronectin Expression Measurements
Fibronectin expression also remained fairly constant in the presence of nanoparticles
of different compositions (Fig. 3.8), as well as in the presence of Ca-SiO 2 nanoparticles of
different sizes (Fig. 3.9). This result was not surprising, since fibronectin production has been
known to often depend on the biomaterial surface [20-21], and all the cells in these
experiments
were grown on the same polystyrene surface.
The consistency of fibronectin
production was also an encouraging finding that suggested that the osteoblasts would continue
to adhere and migrate normally in vivo even after the ingestion of nanoparticles.
L3
20
E
_
.........
··
15
T
II
n-
...
I
T
I
Tr
I
77.....
..
"'''' ''.::'
.
I....
10
.
:,,.:,: :,.
!.,,..
Io
".a.::·
'''"'
"i·'''
:
0
I
''''''
'
·:':'
·:
''.
.
' '''''..:''
Control
1
·i·.·i.,
·i,,i:··
*·····
·:I::.
·
L
SiO 2
Ca-SiO
I
. .:!, ..... ....
' ::
'.''''"'11
::·il·:.::·
·······..
·······
·······-· ···· ·
····-·i·····
·.:··ri.
..
.:'. '<;.:
.:':.::,7-···-·
····-·......
··-i·s.:.,
··i·-··
:;:······-·· ·
····
,···
···
··· · ·
)_"·
·-··.-···.·..-.·
· ·· ·
2
Mg-SiO
2
Na-SiO 2
Figure 3.8. Effect of nanoparticle composition on the fibronectin production of osteoblasts.
Particle size: 50 nm; nominal salt loading: 2.5 wt%. Values are expressed as proteins
produced by 10 6 cells. Values are mean + standard error of the mean; n = 3.
54
20
---
i
i 15
C
~..v.
.....
..
-..v..
.."
..........
,
,:,:::.:
10
.. fM
i./'
a:1
U,
C
·. ..... .
A
U
Control
10 nm
50 nm
100 nm
Figure 3.9. Effect of Ca-SiO2 particle size on the fibronectin production of osteoblasts.
Nominal Ca loading: 2.5 wt%. Values are expressed as proteins produced by 106 cells.
Values are mean ± standard error of the mean; n = 3.
3.3.5. Cytokine Production
In order to further elucidate the biocompatibility of the nanoparticles, the expression
levels of various cytokines were studied. Consistently elevated cytokine levels in vitro would
indicate the potential of in vivo trauma. A number of groups have documented inflammatory
response to small particles in various cell lines and through different routes of administration
[22-25], particularly
through inhalation.
However, we expected that both the bioactive
composition of these nanoparticles and their administration directly to the site of connective
tissue would minimize any inflammatory response.
In addition to the effects of particle composition and mode of administration, the
effects of parameters such as particle size, shape and crystallinity on inflammatory response
have also been examined [23, 26]. Specifically, some groups have shown that while smaller
particle sizes could result in more cytokine secretion, this effect was secondary to particle
morphology. For example, Lacquierrier et al. found that needle-like hydroxyapatite particles
resulted in more cytokine secretion than spherical particles [23]. This was because needle-
like particles caused substantial mechanical damage to the cell membrane during uptake [23].
Our organosilicate particles were spherical in morphology, which would minimize such
damage.
In addition,
any inflammation
observed with amorphous
55
particles was often
temporary [27-29], whereas crystalline particles have been shown to elicit a more severe and
persistent inflammatory response.
3.3.5.1. Interleukin Levels
The effect of various parameters on the presence of a number of different interleukins
was investigated.
Figure 3.10 shows the effect of nanoparticle composition on the osteoblast
production of IL-2, IL-4 and IL-6. These results indicated that while production of IL-2
remained low, elevated levels of IL-4 and IL-6 were evident in the presence of certain
nanoparticles. Specifically, the presence of undoped SiO2 nanoparticles gave rise to the most
dramatic effect on IL-4 and IL-6 production.
/U
-
E
60
50
o
0 40
t5
a. 30
a
= 20
*- 10
n
Control
SiO 2
Ca-SiO 2
Mg-SiO
2
Na-SiO 2
Figure 3.10. Effect of nanoparticle composition on the interleukin levels of osteoblasts.
Particle size: 50 nm; nominal salt loading: 2.5 wt%. Values are expressed as proteins
produced by 106 cells. Values are mean ± standard error of the mean; n = 3.
While SiO2 particles did not enter cells, their presence in the culture medium could
have resulted in the adsorption of proteins onto their surface, leading to protein depletion from
the medium itself. Cell culture medium was supplemented with a variety of growth factors
and proteins to nourish the cells under culture. If a foreign material in the medium caused a
deficiency of such proteins, cell trauma and possibly elevated cytokine production might
occur. Fig. 3.11 shows the protein depletion from cell culture medium due to the presence of
56
nanoparticles of various compositions. This experiment was done without any cells in order
to measure only the effect of protein adsorption on the particle surface. Protein depletion was
calculated by measuring the total protein in the cell culture medium before and after a weeklong incubation with nanoparticles. Doped SiO2 particles essentially resulted in similar
depletion of protein over a week-long period as compared to the undoped SiO2 particles (Fig.
3.11).
However, when these particles were incubated with cells, some of the doped SiO 2
nanoparticles were ingested by cells within the first 24 hr, and would not contribute to further
protein depletion from the medium. Consequently, only the excess particles that have not
been ingested would cause protein depletion that actually affected the cells over the course of
one week. This depletion of nutrients could explain the substantial increase in IL-4 and IL-6
production due to the undoped SiO 2 nanoparticles.
v
vv....v
......................................
........................
..........................
MM
v v..
- ..
v..
.......
25
0
= 20
0
.a
a 15
-....
~:
ID ..I
......
a
.,::.
T
'·- . o:
.
'"'''`""'"""
,::.::::::.
0
i:":·:'::·'
r.·:.:·,::,:·s;::r·:i·:·
"'`'"'*""
.
· ······ · ·· · ·· ··,·::,·
···
;:.:··:.:.:
·
.··. ·:m
··· ·········
··· ·······-·-· ···········.··
··· ·-·
·· ·· ······················
······ ········· ·
· ····· · ··········
.·i ,···::
·.:::·::·-::·:.:···
.·. ·
····i··.···
: :·::::·:·):(:)·
:I'··:·
: :r··.·:··..:i
.
· "''··""
5
n
SiO 2
Ca-SiO
i:,P
·.:.·:
·
j
2
Mg-SiO
2
Na-SiO
2
Figure 3.11. Effect of nanoparticle composition on the protein depletion from cell culture
medium.
Particle size: 50 nm; nominal salt loading: 2.5 wt%.
Values are expressed as
protein depleted after one week of incubation with nanoparticles. Values are mean + standard
error of the mean; n = 3.
Ingestion of nanoparticles has been shown to stimulate cytokine levels in a variety of
cell types, which would explain the elevated IL-6 levels associated with Ca-SiO 2 , Mg-SiO 2,
and Na-SiO
2
nanoparticles.
Many groups have reported the downregulation of osteoblastic
phenotypic marker expression under the influence of various cytokines [6, 30].
However,
while excess IL-6 production was observed due to nanoparticle ingestion (Fig. 3.10),
downregulation
of osteoblastic phenotype was not noted (Figs. 3.1-3.3).
57
The fact that the
slightly elevated IL-6 levels were not significant enough to result in a loss of osteoblastic
phenotype was a very positive finding, since it suggested that osteoblasts would continue to
function normally after particle ingestion.
The effect of Ca-SiO 2 particle size on osteoblast IL-6 production is shown in Fig. 3.12.
Smaller particle sizes resulted in greater IL-6 production.
This effect was most likely due to
the increased cellular uptake of smaller particles, as reported in Chapter 2 [17], since
inflammatory response due to particle ingestion was often dependent on particle dosage [26].
While an elevated level of IL-6 was observed due to Ca-SiO 2 nanoparticle ingestion, the
osteoblasts continued to express phenotypic markers at control levels (Figs. 3.2 and 3.5). In
addition, there was enough evidence in the literature to indicate that this cytokine response
might not have a permanently damaging effect in vivo, since mild inflammatory response due
to amorphous materials was often temporary [27-29]. The retention of osteoblastic phenotype
despite the elevated IL-6 expression supported the use of Ca-SiO2 nanoparticles as osteoblast
gene delivery vehicles.
A
"
/4U
..................
30
E
I
a
20
:·:·'·
-ja,
···
:···:
· · ·· :,::,
···:I·I:
·
10
:::::
····
A)
Control
10 nm
50 nm
100 nm
Figure 3.12. Effect of Ca-SiO2 particle size on the IL-6 levels of osteoblasts. Nominal Ca
loading: 2.5 wt%. Values are expressed as proteins produced by 106 cells. Values are mean +
standard error of the mean; n = 3.
Fibroblasts
were also assayed for interleukin expression. As with osteoblasts, the
presence of Ca-SiO 2 nanoparticles led to slightly elevated IL-6 levels in fibroblasts (Fig. 3.13).
58
100
80
0
.0
·
0
40
20
nv
Control
10 nm
50 nm
100 nm
Figure 3.13. Effect of Ca-SiO 2 nanoparticle size on the IL-6 levels of fibroblasts. Nominal
Ca loading: 2.5 wt%. Values are expressed as proteins produced by 106 cells. Values are
mean ± standard error of the mean; n = 3.
3.3.5.2. TNF-a Levels
In addition to our detailed study of the interleukin expression levels, we have also
quantified the cellular levels of TNF-a in response to various nanoparticles. TNF-a has been
cited to play an integral role in bone resorption; it has also been implicated to stimulate the
production of other pro-inflammatory cytokines [6,30]. Figure 3.14 shows the TNF-a levels
for osteoblasts cultured with nanoparticles of various compositions. These results indicated
that there was no substantial induction of TNF-a due the ingestion of organosilicate
nanoparticles. The expression of TNF-a also remained unaffected by variations in calcium
loading and particle size (Fig. 3.15).
Fibroblasts were also evaluated for TNF-a induction, as shown in Fig. 3.16.
The
results indicated that various compositions of nanoparticles did not substantially stimulate
TNF-a secretion by fibroblasts. These data suggested that any inflammatory response caused
by nanoparticles administered to the connective tissue would be minimal.
The fact that the
Ca-SiO2 nanoparticles caused virtually no increase in TNF-a expression was a very positive
result. Minimization of this particular cytokine in orthopedic procedures is particularly
important, since a rise in TNF-a often stimulates the secretion of other cytokines.
59
160 -
r
140 -
E
S
0.
120-
100 -
-
·:-:····
·! ·::j\.v::
80
:
::
60
L
:M
::%
::
::U
40 -
:
::
20 -
'b.N
· ········
:.··.(
·· · :'·· · ······ ··
f
:gND N
MSj:
.D
/:'::
0-
!
!
Control
................
i
SiO2
:..................
Ca-SiO
': ·· · ······
...I·
-111
-
Mg-SiO 2
2
-
Na-SiO2
Figure 3.14. Effect of nanoparticle composition on the TNF-a levels of osteoblasts. Particle
size: 50 nm; nominal salt loading: 2.5 wt%. Values are expressed as proteins produced by 106
cells. Values are mean ± standard error of the mean; n = 3.
rrrr
L
..r-
."!...?..
80 -
......·... .
E
0
0.
-J
:.,·:-·:.··:· :. .·.·
·········
·· ·-··
:......
.·liI···
··· ·
.·
60-
·:: :·:
·-····· · ··
·,...I·
...::
i.::·.·
:·· ·
40 -
··::·:
z
·· · :·
::·::
I-
20 -
:::.
.· .: .
.-· .·
·.:::·
0
Control
10 nm
50 nm
100 nm
Figure 3.15. Effect of Ca-SiO2 particle size on the TNF-a levels of osteoblasts. Nominal Ca
loading: 2.5 wt%. Values are expressed as proteins produced by 106 cells. Values are mean +
standard error of the mean; n = 3.
60
-7 .'
/I
-
60 -
T
.
T
I
E 507a)
'
z1- 20-
.: ....
,...: ..
. .1::'
i
.:·
::
o
'... .... ,...
40
30-
I
...... ....
T
40 -
T
Fi TIi
I
I
':X"'"
.::....i..
..b. *`"
.`"
.'.. ... i
'':'. .'
". '. ' ' '
.::::::
'~
''"
'''
'
i:.
...!··
'· '·'v/.' j ,,v,o
, '.'
"'a',
'. ....
,..
'' ' "*
'i
....:,.:j ,',..
:, '/.
,
.¢
:
:':
':
I
.. ',' j Ni`'' Sj'S
10 -
s:' i: ;: 'i.'i
::.
: :' ':'
' .''.i'
i·
.'..:'·-.:.
:...i
nControl
SiO 2
Ca-SiO 2
Mg-SiO 2
Na-SiO 2
Figure 3.16. Effect of nanoparticle composition on the TNF-a levels of fibroblasts. Particle
size: 50 nm; nominal salt loading: 2.5 wt%. Values are expressed as proteins produced by 106
cells. Values are mean ± standard error of the mean; n = 3.
3.4. Summary
We have examined osteoblast protein expression in response to cellular uptake of doped
organosilicate nanoparticles. Osteoblastic phenotype was retained after nanoparticle ingestion,
as evidenced by the normal expression of alkaline phosphatase, osteocalcin, osteopontin and
fibronectin. A detailed study of cytokine production showed that while IL-2, IL-4 and TNFa levels were not substantially affected by particle uptake, there was an increase in IL-6
production in both osteoblasts and fibroblasts.
This increase, however, did not affect
osteoblastic phenotype, induce cell toxicity, or stimulate production of the other cytokines.
The non-toxicity and preservation of osteoblastic phenotype after particle uptake were very
positive in vitro results that provided strong support for the further examination of these
nanoparticles as gene delivery vehicles for osteoblasts.
61
3.5. References
[1]
Windeler, A.S., Bonewald, L., Khare, A.G., Boyan, B., and Mundy, G.R., in "The
Bone-Biomaterial Interface," (J.E. Davies, Ed.), University of Toronto Press, Toronto,
Canada, 1991, p. 205.
[2]
Yang, S.-Y., Ren, W., Park, Y., Sieving, A., Hsu, S., Nasser, S., and Wooley, P.H.,
Biomater. 23, 3535 (2002).
[3]
Broyles, D.L, Nielsen, R.G., Bussett, E.M., Lu, W.D., Mizrahi, I.A., Nunnelly, P.A.,
Ngo, T.A., Noell, J., Christenson, R.H., and Kress, B.C., Clin. Chem. 44, 2139 (1998).
[4]
Oreffo, R.O.C., Driessens, C.M., Planell, J.A., Triffitt, J.T., Biomater. 19, 1845 (1998).
[5]
Wang, M.L., Nesti, L.J., Tuli, R., Lazatin, J., Danielson, K.G., Sharkey, P.F., and
Tuan, R.S., J. Orthop. Res. 20, 1175 (2002).
[6]
Tryoen-Toth,
P., Vautier, D., Haikel, Y., Voegel, J.-C., Schaaf, P., Chluba, J., and
Ogier, J., J. Biomed. Mater. Res. 60, 657 (2002).
[7]
Wu, H.C., Lin, C.C., Chen, W.C., Chen, HY., and Tsai, F.J., Eur. Urol. 43, 197 (2003).
[8]
Reinholt, F.P., Hultenby, K., Oldberg, A., and Heinegard, D., Proc. Natl. Acad. Sci. 87,
4473 (1990).
[9]
Moursi, A.M., Damsky, C.H., Lull, J., Zimmerman, D., Doty, S.B., Aota, S., and
Globus, R.K., J. Cell Sci. 109, 1369 (1996).
[10]
Lee,T., Cytokines, January 1997, <http://microbiology.medicine.dal.ca/education/
pimunit/cytok.htm>
[11]
(10 September 2003).
Porter, D.W., Ye, J., Ma, J., Barger, M., Robinson, V.A., Ramsey, D., McLaurin, J.,
Khan, A., Landsittel, D., Teass, A., and Castranova, V., Inhal. Toxicol. 14, 349 (2002).
[12]
Hetland, R.B., Schwarze, P.E., Johansen, B.V., Myran, T., Uthus, N., and Refsnes, M.,
Hum. Exp. Toxicol. 20, 46 (2001).
[13]
Savage, S.T., Lawrence, J., Katz, T., Stearns, R.C., Coull, B.A., and Godleski, J.J., J.
Air Waste Manage. 53, 1088 (2003).
[14]
Rabovsky, J., Scand. J. Work Environ. Health 21, 108 (1999).
[15]
Lewinson, J., Mayr, W., and Wagner, H., Regul. Toxicol. Pharm. 20, 37 (1994).
[16]
Johnston, C.J., Driscoll, K.E., Finkelstein, J.N., Baggs, R., O'Reilly, M.A., Carter, J.,
Gelein, R., and Oberdorster, G., Toxicol. Sci. 56, 405 (2000).
[17]
Moudgil, S., and Ying, J.Y., to be submitted to Biomater.
62
[18]
Sampath, T.K., Maliakal, J.C., Hauschka, P.V., Jones, W.K., Sasak, H., Tucker, R.F.,
White, K.H., Coughlin, J.E., Tucker, M.M., Pang, R.H.L., Corbett, C., Ozkaynak, E.,
Oppermann, H., and Rueger, D.C., J. Biol. Chem. 267, 20352 (1992).
[19]
Siggelkow,
H., Rebenstorff,
K., Kurre, W., Niedhart, C., Engel, I., Schulz, H.,
Atkinson, M.J., and Hufner, M., J. Cell Biochem. 75, 22 (1999).
[20]
Stephansson, S.N., Byers, B.A., and Garcia, A.J., Biomater. 23, 2527 (2002).
[21]
Clarke, S.A., and Revell, P.A., J. Biomed. Mater. Res. 57, 84 (2001).
[22]
Richard, R., Mineral. Mag. 67, 129 (2002).
[23]
Laquerriere, P., Grandjean-Laquerriere,
A., Jallot, E., Ballosier, G., Frayssinet, P., and
Guenounou, M., Biomater. 24, 2739 (2003).
[24]
Hubbard, A.K., Timblin, C.R., Shukla, A., Rincon, M., and Mossman, B.T., Am. J.
Physiol.-Lung Cell. Molec. Physiol. 282, L968 (2002).
[25]
McCusker, C., Chicoine, M., Hamid, Q., and Mazer, B., J. Allergy Clin. Immun. 110,
891 (2002).
[26]
Schins, R.P.F., Inhal. Toxicol. 14, 57 (2002).
[27]
Murphy, S.A., Berube, K.A., Pooley, F.D., and Richards, R.J., Life Sci. 62, 1789
(1998).
[28]
Warheit, D.B., McHugh, T.A., and Hartsky, M.A., Scand. J. Work Environ. Health 21,
19 (1995).
[29]
Yuen, I.S., Hartsky, M.A., Snajdr, S.I., and Warheit, D.B., Am. J. Resp. Cell. Mol. 15,
268 (1996).
[30]
Quintero, J.C., Piesco, N.P., Langkamp, H.H., Bowen, L.L., J. Dent. Res. 74, 1802
(1995).
63
Chapter 4 - Synthesis, Physicochemical Characterization and In Vitro Transfection
Studies of Organosilicate Nanoparticle-DNA Complexes
4.1. Introduction
Osteogenic proteins such as bone morphogenetic proteins (BMPs) have received
attention as promising therapies for fracture repair [1]. The pioneering work in this area was
done by Urist and coworkers who first identified BMP activity in tissue extracts [2].
However, it has been almost 40 years since that discovery and there is still no effective
recombinant protein therapy available for bone repair. The reasons for this slow progress lie
in the multiple challenges associated with protein delivery. These challenges include the
development of a delivery system that can maintain the structural integrity and functional
activity of the protein in vivo, and deliver the therapeutic dosage of the protein required to
induce bone formation. The tissue engineering of bone involves the incorporation of such
proteins into cell-seeded three-dimensional scaffolds to restore tissue function. While this
approach has had some success, maintenance of cell mass at the fracture site is difficult, and
the mass transfer limitations associated with large and complex fractures continue to be a
problem [3].
To overcome these challenges, researchers have recently begun to consider gene
therapy for fracture repair [4-5]. Rather than delivering osteogenic proteins to the fracture site,
delivering the genes that encode for these proteins would allow cells to produce such proteins
at therapeutic times and levels. Fang et al. have shown that direct transfer of naked DNA via
a polymer matrix can induce bone formation [4]. However, naked DNA transfection is often
associated with low transfection efficiencies. In most cases, DNA requires a delivery vehicle
or vector to aid its entry into the cell. While the development of non-viral vectors for gene
therapy has been an active area of research, non-viral vectors are plagued by cytotoxicity and
lower transfection efficiencies than their viral counterparts. The development of non-viral
vectors that can transfect bone cells at higher efficiencies than naked DNA without inducing
cytoxicity would allow for authentic growth factor production and cell maintenance at the
fracture site, thus accelerating fracture healing.
Our previous work has shown that surface-doped organosilicate nanoparticles were
able to enter the osteoblast cell membrane without inducing cytotoxicity [6-7]. We have
quantified the effects of particle size and composition on cellular uptake, cell proliferation and
64
cellular protein expression. In particular, Ca-SiO2 nanoparticles were found to timulate cell
proliferation and maximize cellular uptake, making them ideal candidates for intracellular
delivery vehicles.
In this chapter, we will explore the potential of these nanoparticles as non-
viral vectors to deliver plasmid DNA to osteoblasts. We will investigate how properties such
as particle size and composition affect nanoparticle-DNA complex formation.
understanding
This
is then used to develop a DNA delivery system that transfects osteoblasts
without inducing cytotoxicity.
4.2. Experimental
4.2.1. Synthesis of Nanoparticle-DNA Complexes
The plasmid gWizTM GFP (5757 bp) encoding the green fluorescent protein (GFP)
gene was obtained from Aldevron, Inc. This plasmid contains the GFP gene under the control
of a modified promoter from the cytomegalovirus (CMV) immediate early (IE) gene. The
ready-to-use reporter plasmid was obtained as a 1.0 mg/ml stock solution in aqueous TrisHCL/EDTA
buffer.
The GFP produced has an excitation peak at 470-480 nm and an
emission peak at 510 nm.
For Ca-SiO2 nanoparticle-DNA complex synthesis, Ca(NO3) 2'4H02
(Fluka) was
added to plasmid DNA diluted in a buffer solution and quickly vortexed. Ca loading was
varied through the addition of different dilutions of a 250 mM Ca(NO3) 2'4H0 2 stock solution
to the DNA solution. SiO2 nanoparticles (synthesized according to protocol in Section 2.2
[6]) were ultrafiltered three times in buffer in an Amicon Ultrafiltration Cell (Millipore) using
a 50K PES membrane (Millipore). The nanoparticles were slowly added to the Ca-DNA
solution, vortexed and allowed to form complexes for 30 min at 25°C.
NH3+-SiO2 nanoparticles were synthesized via sol-gel surface functionalization of
SiO2 nanoparticles. SiO2 nanoparticles were prepared as described in Section 2.2 [6] and
aged in reaction solution. After 3 hr, 3-aminopropyltrimethoxysilane
((CH 3 0)Si(CH
3
2) 3NH 2
or aminoTMOS, Gelest) was slowly added to the particle suspension at an aminoTMOS to
methyltrimethoxysilane (Gelest) molar ratio of 0.2:1. The resulting particles were aged for 3
hr before ultrafiltration.
For complex formation, the NH3+-SiO 2 nanoparticles were added to a
plasmid DNA solution and mixed for 30 min.
65
The cationic lipid lipofectamine
2000 (Sigma) was used as a positive
control.
Lipofectamine was diluted in a buffer solution, added to a plasmid DNA solution, and mixed
for 30 min to form lipofectamine-DNA complexes.
4.2.2. PhysicochemicalCharacterizationof Vector-DNAComplexes
Complex size was measured using dynamic light scattering (DLS) at 250 C using a
Lexel 95 Argon-ion laser and a Brookhaven high-precision photomultiplier at 90 °. Data were
acquired using a Brookhaven 9000AT correlator.
autocorrelation
The periods for acquisition of the
function were 0.1 gs to 0.5 s, and complex size distributions were obtained
from non-negatively constrained least-squares (NNCLS) analysis of the data. Zeta potential
measurements were performed on 0.15 wt% nanoparticle suspensions in cell culture medium
using a ZetaPALS Zeta Potential Analyzer (Brookhaven).
Solution pH was measured using an Orion model 420A pH meter, and adjusted by the
addition of KOH or HC1. Agarose gel electrophoresis was used to determine the ability of the
nanoparticles to bind DNA. It was performed in a 1% (w/v) gel, ethidium bromide induced
for visualization for 3 hr at 60 V. DNA binding efficiency was measured by centrifuging the
complexes at 4000 g and measuring the amount of unbound DNA in the supernatant using the
PicoGreen® DNA Quantitation Kit (Molecular Probes).
Atomic force microscopy (AFM) samples were prepared by incubating synthesized
complexes on freshly cleaved mica discs obtained from Ted Pella, Inc. (Redding, CA) for 20
min. After sample adsorption, the discs were washed three times with 200 pL of deionized
water to remove any unbound sample.
AFM was performed on a Multimode AFM (Digital
Instruments/Veeco, Santa Barbara, CA) controlled by a Nanoscope IV Controller (Digital
Instruments/Veeco).
Topographic and phase images were collected in the tapping mode in air.
4.2.3. TransfectionExperiments
Primary osteoblasts, MC3T3 osteoblasts, HepG2 hepatocytes, and C3H1OTl/2
fibroblasts were cultured according to established protocols (see Section 2.2) [6]. Cells were
grown in 8-well chamber slides for microscopy experiments, and 75-cm2 flasks for
transfection efficiency experiments. Vector-DNA complexes were prepared as described
above. Various compositions of complexes were added to nearly confluent cell layers 2-3
66
days after cell seeding. After addition of the complexes, cells were incubated at 37C in 5%
CO2 atmosphere. After 3 hr, the complexes were aspirated and replaced with culture medium.
After 48 hr, cells were assayed for GFP expression
both qualitatively
through
fluorescence microscopy and quantitatively through fluorescence-activated cell sorting
(FACS).
For microscopy,
the cells grown in chamber
slides were fixed with 2%
paraformaldehyde. The slides were then viewed under a Zeiss Axiovert 200 fluorescence
microscope equipped with a SPOT digital camera. FACS experiments were performed using
a Facscan cytometer (Becton-Dickinson) equipped with an argon laser. At least 10,000 events
were counted for each sample.
The software package Cell Quest (Becton-Dickinson)
was
used to determine the events of interest from forward and side scatter parameters. The mean
fluorescence intensity of the cells was obtained from the mean channel number of the
fluorescence histograms of the gated population. For transfection efficiency measurements,
total fluorescence intensity was plotted versus the GFP fluorescence intensity on a log scale
and cell auto-fluorescence was eliminated. Total GFP expression was quantified from a shift
in the mean fluorescence of cell populations relative to the control cell population with no
DNA.
4.3. Results and Discussion
4.3.1. Synthesis and Physicochemical Characterization of Ca-Si02 -DNA Complexes
4.3.1.1. Effect of Precursor Addition Sequence
Addition of DNA to the preformed Ca-SiO2 nanoparticles did not result in complex
formation, due to charge repulsion between the negatively charged DNA molecules and the
negative surface charge of the Ca-SiO2 nanoparticles.
In order to form Ca-SiO2-DNA
complexes, the DNA was premixed with the Ca(NO 3) 2 4H20 solution and then mixed with the
SiO2 nanoparticles. This synthesis method is modified from the concept used to form calcium
phosphate-DNA complexes [5], in which the DNA is premixed with one of the precursors and
complex formation is initiated by the addition of the other precursor to form a precipitate.
While the Ca-DNA solution addition to the nanoparticle suspension resulted in some
complex formation, this sequence of addition led to a broader complex size distribution
compared to the slow addition of nanoparticles to the Ca-DNA solution (Fig. 4.1). Plasmid
DNA molecules alone exhibited a broad DLS distribution with a mean of - 350 nm.
67
(a)
100
0..................v....._............A.....................................................................^...^...^.^........-.''.'.'^.'.'.'^'.v.''............
'-'-'-'-''-'-'-'^-'-^
-'--^-'-'-^-'
'''''''^''^......^.'.^.'.'.^.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'..
100
80
Vu 60
0
-
'4-
60
40
0
o
I
0
0
100
200
300
400
Hydrodynamic
500
600
700
800
Diameter (nm)
100
(b)
800
0
=
60
0
I
40
0
0
100
200
300
400
500
600
700
800
Hydrodynamic Diameter (nm)
Figure 4.1. Complex size distributions of complexes formed by addition of (a) Ca-DNA
solution to 50-nm SiO 2 nanoparticles and (b) 50-nm SiO2 nanoparticles to Ca-DNA solution.
Nominal Ca loading: 1.25 wt%; DNA loading: 24 pg/ml.
The data indicated that the Ca-DNA disrupted the colloidal suspension of SiO2 more
than the addition of the stable SiO 2 suspension to the Ca-DNA mixture.
68
Other groups have
also reported the preparation of nanoparticle-DNA complexes with narrow size distributions
through the addition of nanoparticles to DNA [8].
4.3.1.2. Effects of Nanoparticle Size and Composition on DNA Binding Affinity
In order to confirm
that DNA was complexed
electrophoresis was performed.
separate macromolecules
with the particles,
agarose gel
This technique uses a matrix of highly purified agar to
based on charge and size.
Electrophoretic
immobilization of the
DNA results in gel retardation of the complexes, implying binding of the DNA to the
nanoparticle
surface [9,10].
Figure 4.2 shows gel electrophoresis
results for complexes
formed with 50-nm SiO2 particles of various Ca loadings. The results indicated that no
complex
formation
occurred with Ca-free SiO 2 nanoparticles,
since the electrophoretic
mobility of the DNA was almost identical to that of the free plasmid. This was most likely
due to the lack of charge attraction between SiO2 and DNA molecules. However, the DNA
appeared to form complexes with the SiO2 nanoparticles of increased Ca loadings, as
indicated by the DNA retention.
Ca-SiO 2
Ca-SiO 2
(1 wt% Ca)
(0.5 wt% Ca)
SiO2
Control
Marker
Figure 4.2. Agarose gel electrophoresis of nanopartice-DNA complexes prepared with 50-nm
SiO 2 with nominal Ca loadings of 1 wt% Ca (Lane 1), 0.5 wt% Ca (Lane 2) and 0 wt% Ca
(Lane 3). Lane 4 represents free DNA without complex formation, and Lane 5 is the DNA
marker. Particle loading: 0.15 wt%; DNA loading: 240 [tg/ml.
The extent of DNA binding was measured using co-sedimentation analysis by
ultracentrifuging the complexes and measuring the amount of bound DNA through
fluorescence. Figure 4.3 indicates that as Ca loading increased, the amount of bound DNA
69
1200
1000
E
I
a)
N
800
xn 600
a.
E
400
0
200
TO
0.0
0.5
1.0
1.5
2.0
Nominal Ca Loading (wt%)
Figure 4.4. Effects of particle size and Ca loading on the size distribution of Ca-SiO2-DNA
complexes. Particles of () 10 nm, () 50 nm and (A) 100 nm are examined at a particle
loading of 0.15 wt% and a DNA loading of 24 lg/ml. Values are mean i standard error of the
mean; n = 3.
5 nm
0 nm
Figure 4.5. AFM image of pure plasmid DNA.
71
also increased.
formation.
This was to be expected, since the addition of Ca promoted complex
The further addition of Ca led to agglomeration between complexes, and these
large aggregates bound even more DNA.
The data also indicated that as nanoparticle size
increased, bound DNA increased. The colloidal suspensions of larger particles were less
stable than those of smaller particles, and the addition of DNA to the larger particles would
lead to greater aggregate formation, binding more DNA.
tU
50
C 40
z
30
20
1
10
0
0.0
0.5
1.0
1.5
2.0
Nominal Ca Loading (wt%)
Figure 4.3. Effects of particle size and Ca loading on the extent of DNA binding in Ca-SiO2DNA complexes. Particles of () 10 nm, () 50 nm and (A) 100 nm are examined at a
particle loading of 0.15 wt% and a DNA loading of 24 [pg/ml.
4.3.1.3. Effects of Particle Size and Ca Loading on Complex Size
While the DNA binding affinity was shown to increase with increasing particle size
and Ca loading, the complex size analysis showed that above certain parameters, extremely
large complexes
were formed (see Fig. 4.4).
Addition of Ca and DNA to the SiO 2
nanoparticle suspension resulted in an increased complex size and in a broadening of the
complex size distribution.
The broad complex size distribution was caused by the structural
range of complexes that were formed when the nanoparticles were mixed with Ca and DNA.
This phenomenon was illustrated by AFM (see Figs. 4.5-4.6).
The image of pure DNA (Fig.
4.5) depicted the various plasmid conformations that often led to the polydispersity of DNA
complexes (Fig. 4.6).
70
90 nm
...
0 nm
Figure 4.6. AFM images of Ca-SiO 2-DNA complexes formed with (a) 10-nm and (b) 50-nm
particles. Nominal Ca loading: 1.25 wt%; DNA loading: 24 ptg/ml.
It is interesting
to note that while 10- and 50-nm Ca-SiO
2
particles possessed
discretely different size distributions when prepared as colloidal suspensions, the addition of
DNA not only led to a broadening of their size distributions, but also resulted in substantial
overlap of their size distributions (Fig. 4.4). For example, at a Ca loading of 1.5 wt%, the
complex size distributions for 10- and 50-nm particles were almost identical, indicating that
the addition of Ca and DNA to the 10-nm SiO 2 particle suspension resulted in a significant
degree of agglomeration.
This agglomeration might be due to the energetically
favorable
reduction in surface free energy that occurred from the decrease in surface area [11].
However, while the 10- and 50-nm particles formed complexes of similar sizes, the 100-nm
particles led to much larger complexes (see Fig. 4.4.), indicating that the 100-nm particle
suspension was less stable from agglomeration.
Many groups have shown this agglomeration
trend in pure colloidal silica, whereby as particle size increased, the critical salt concentration
required for aggregation decreased [12]. The complex size data agreed with the DNA binding
data, suggesting that agglomeration caused increases in both complex size and DNA binding.
4.3.2. Synthesis and PhysicochemicalCharacterizationof NH3+-SiO2-DNA Complexes
Although we were able to successfully form Ca-SiO 2-DNA complexes, we wanted to
investigate if we could further reduce complex size in order to deliver more DNA into the cell
cytoplasm.
The ability to control complex size is extremely important, since smaller
complexes have been shown to enhance cellular uptake [13-15]. Our research with doped
72
SiO2 nanoparticles has also displayed the significant effect of particle size on cellular uptake
[6]. Researchers have shown that the amount of DNA delivered to the cell is an important
determinant of transfection efficiency [16]. Since stronger electrostatic interactions between
the nanoparticles and plasmid DNA would likely result in greater DNA condensation and
smaller complex sizes, we have functionalized
the SiO 2 nanoparticles with -NH
+
3
surface
groups.
Surface functionalization of silica particles has been studied extensively using various
organosilane surface modifiers [17]. NH3 -SiO2 nanoparticles were synthesized via sol-gel
functionalization of SiO2 nanoparticles using aminoTMOS. To determine if the aminoTMOS
precursors were indeed reacting with the surface of SiO 2 particles and not condensing with
other aminoTMOS, the particle size distribution was analyzed after the aminoTMOS
precursors had been introduced. The retention of a single normal particle size distribution
(not shown) confirmed that aminoTMOS only reacted with the SiO 2 surface.
The zeta potential of the particles indicated that unlike SiO2 and Ca-SiO 2 nanoparticles,
the NH 3+-SiO 2 nanoparticles were positively charged at neutral pH (Table 4.1). This positive
surface charge might allow for greater DNA condensation.
which
shows
that for similarly
sized SiO 2-based
This was confirmed by Fig. 4.7,
nanoparticles,
the NH 3+-SiO 2-DNA
complexes were smaller than the Ca-SiO 2-DNA complexes.
Table 4.1. Zeta potential values of SiO 2, Ca-SiO 2 and NH 3 +-SiO 2 nanoparticles.
Particle
Zeta Potential
Composition
SiO 2
Ca-SiO 2*
NH 3 +-SiO2
at pH = 7.4
-30.4
-15.8
21.3
*Nominal Ca loading for Ca-SiO2 nanoparticles: 2.5 wt%.
73
70 -
500
60 -
400
50 -
E
-
- 300 ·
o 40 -
,en
._
z 30-
200 _)
E
20-
100 oS
1r I
0 -
I
10
50
-O
100
Particle Size (nm)
Figure 4.7. The effect of particle size on the DNA binding of (o) NH 3+-SiO2 and (El) Ca-SiO 2
nanoparticles and the size distributions of (e) NH 3+-SiO2 -DNA and () Ca-SiO 2-DNA
complexes. Nominal Ca loading: 1.5 wt% for 10-nm and 50-nm Ca-SiO 2 particles, 1.0 wt%
for 100-nm Ca-SiO 2 particles.
The extent of DNA complexed with the NH3+-SiO2 nanoparticles was measured using
the DNA binding
assay described
earlier.
Figure 4.7 indicated
that the NH 3+-SiO2
nanoparticles bound more DNA than the Ca-SiO2 nanoparticles.
4.3.3. Transfection Studies
4.3.3.1. Effect of Nanoparticle Composition on Transfection Efficiency
Since the NH3+-SiO2-DNA complexes bound more DNA and formed smaller
complexes than the Ca-SiO 2-DNA complexes, they might be able to deliver more DNA to the
cells. However, Fig. 4.8 indicates that the transfection efficiencies of NH 3+-SiO2-DNA were
very low in both osteoblasts and hepatocytes.
The transfection efficiency of Ca-SiO 2-DNA in
osteoblasts was much higher than that of NH3+-SiO2-DNA.
74
A A
o Osteoblasts
35
O
30
r_
T
o
o
Hepatocytes
_T
0
u 15-
.2 10-
I-
···-:··
T
·--......·····-:,-·.
-:····
·-·· ·i··:-..
·.,.........
..
.·.·.
......
:...:.·.....
.,.····
.......-.
i::··-·:i
I
5·· ···· · ···
:····
n-
;i~i!!ii.,,?::?i'::lI
Lipofectamine-DNA
Ca-SiO 2-DNA
rT_
~
~
i?:i::i!~:i!~i~i¥
NH3+-SiO2 -DNA
Figure 4.8. Transfection efficiencies of lipofectamine-DNA, Ca-SiO2-DNA and NH3 +-SiO2DNA complexes in osteoblasts and hepatocytes. SiO 2 particle size: 50 nm; particle loading:
0.15 wt%; nominal Ca loading: 1.5 wt%; DNA loading: 24 g/ml. Values are mean
standard error of the mean; n = 3.
These data suggested that factors other than cellular uptake and DNA delivery were
controlling the ultimate transfection efficiency. There are many intracellular barriers to gene
delivery, including release of the DNA inside the cell and entry of DNA into the nucleus. In
some cases, gene expression can occur in the cytoplasm [18-19].
However, for the gWizTM
GFP plasmid used in our experiments, GFP production would only occur if the plasmid has
entered the nucleus. While some groups have reported nuclear entry of entire vector-DNA
complexes [20-21], our previous microscopy data showed that organosilicate nanoparticles
did not enter the nucleus [6]. Thus, the major barrier preventing gene expression in the case
of the NH 3 -SiO2-DNA complexes was most likely the intracellular release of DNA, since the
DNA would have to make its way to the nucleus in order for GFP to be expressed. Schaffer
et al. have stressed the importance of "vector unpackaging" [22] as a major barrier to
transfection. In our NH3 +-SiO2-DNA system, while stronger electrostatic interactions resulted
in smaller complex sizes, the increased DNA binding strength might have hindered DNA
release from the complexes, thus preventing the entry of DNA into the nucleus.
Some groups have reported successful transfection of other cell lines with silica [23]
and gold [24] nanoparticles functionalized
with quaternary ammonium groups.
The alkyl
spacer length between the nanoparticle surface and the ammonium group has been suggested
75
to play a role in DNA condensation and subsequent release, with a longer chain length
allowing for steric hindrances to prevent the DNA from binding too tightly. While these
approaches have had some success, they continue to exhibit some cytotoxicity.
Our Ca-SiO2 -DNA complexes exhibited lower transfection efficiencies in osteoblasts
and hepatocytes than lipofectamine-DNA complexes (Fig. 4.8).
However, the former
demonstrated excellent selectivity towards transfecting osteoblasts versus hepatocytes
compared to lipofectamine-DNA
complexes.
content in the Ca-SiO 2-DNA complexes.
This was most likely due to the surface Ca
This selectivity could be very useful in the systemic
administration of these complexes to a fracture site.
4.3.3.2. Effects of Particle Size and Ca Loading on Transfection Efficiency
Since the Ca-SiO 2-DNA complexes exhibited successful transfection in osteoblasts,
the effects of particle size and Ca loading in this system on transfection
efficiency in
osteoblasts were investigated in detail (see Fig. 4.9). Figure 4.10 provides a visual depiction
of some of the transfected cells.
20
0
15
.
0
u)
C
.
5
0
0.0
0.5
1.0
1.5
2.0
Nominal Ca Loading (wt%)
Figure 4.9. Effect of Ca loading on the osteoblast transfection efficiency of Ca-SiO2-DNA
complexes formed with () 10-nm, () 50-nm and (A) 100-nm particles.
0.15 wt%. Values are mean + standard error of the mean; n = 3.
Particle loading:
The data indicated that Ca-SiO 2-DNA complexes formed with 10-nm and 50-nm Ca-
SiO2 particles resulted in successful transfection. Complexes formed with 100-nm particles
76
resulted in almost no transfection. The inability of these complexes to transfect osteoblasts
could be attributed to their effective complex size. Figure 4.4. shows that the complexes
formed with 100-nm Ca-SiO 2 nanoparticles were extremely large, with a mean complex size
of > 450 nm. These large aggregates have major difficulty in entering the cell to deliver
DNA to the nucleus.
Fig. 4.10.
Fluorescence micrographs of osteoblasts transfected with Ca-SiO2-DNA
complexes formed with 50-nm Ca-SiO2 particles with a nominal Ca loading of (a) 0.50 wt%,
(b) 0.75 wt%, (c) 1.00 wt% and (d) 1.25 wt%.
Figure 4.9 also shows that the transfection efficiencies exhibited by complexes formed
with 10-nm and 50-nm Ca-SiO 2 particles were fairly similar. This similarity was somewhat
surprising, since Section 2.3.4 indicated that the cellular uptake of 10-nm Ca-SiO2 particles
was significantly higher than that of 50-nm Ca-SiO2 particles [6]. Because of this difference
in uptake, one would expect to see significantly higher transfection efficiencies with the
77
complexes formed with 10-nm Ca-SiO2 particles as compared to those formed with 50-nm
Ca-SiO2 particles. However, the sizes of the Ca-SiO2-DNA complexes formed with these two
different Ca-SiO 2 particle sizes were quite similar (see Figure 4.4).
This similarity in
complex sizes led to the similarity in their overall transfection efficiencies.
An optimal Ca
content of 1.25 wt% led to transfection efficiencies of 14.7% and 17.4% in osteoblasts for the
complexes formed with 10-nm and 50-nm Ca-SiO 2 particles, respectively.
This optimal Ca
loading gave rise to one of the smallest average complex sizes formed with 10-nm and 50-nm
Ca-SiO 2 particles (Fig. 4.4).
In order to further examine the correlation between nanoparticle uptake and
nanoparticle-DNA transfection efficiency, the effects of particle size and Ca loading on
uptake and transfection efficiency were investigated (Fig. 4.11). The data showed that a
higher nominal Ca loading was required for optimal transfection efficiency vs. optimal uptake.
These results suggested that in the case of the nanoparticles alone, Ca merely modified the
surface composition (not the particle size) to enhance cellular uptake. On the other hand, in
nanoparticle-DNA complexes, Ca played an additional role of reducing the complex size.
Since Ca interacted with the anionic groups of DNA in order to condense the large DNA
molecules, more Ca was required to achieve optimal transfection.
bUU
4U
35 _
500
30 >,
0
" 400
w
25 .._o
300
20
W
: 200
5
15
100
0
0u
0
0.00
0.25
0.50
0.75
1.00
1.25
1.50
Nominal Calcium Loading (wt%)
Figure 4.11.
Effect of Ca loading on the uptake of ()
10-nm and (o) 50-nm Ca-SiO 2
nanoparticles and the osteoblast transfection efficiency of complexes formed with ()
and ()
10-nm
50-nm Ca-SiO 2 particles. Particle loading: 0.15 wt%; DNA loading: 24 jig/ml; n = 3.
78
4.3.3.3. Transfection
Transfoction
tionEf
Efficiencies of Ca-SiO2-DNA Complexes
Ef
lexesin Other
Other Cell
Cell Types
ypes
ypes
Besides osteobla
()steobla
osteoblasts,
)steoblasts, other connective and non-connective
:)nnective tissue cell types were also
tr;
insfection studies.
examined in transfecti(
examined
transfection
insfecti(
Figure 4.12 shows
vs the transfection efficiencies for
)blasts
fibroblasts
osteoblasts, fibr(
)blasts aand hepatocytes. Lipofectamine,
Lipofectarine, the cationic lipid used as a positive
ted
all cells
c(
transfected
control, transfected
control,
transfecl
tedall
due to non-specific interactions.
)ns. While
While Ca-SiO2-DNA
Ca-SiO2-DNA complexes
complexes
exhibited
exhibited lower
lower transfec
transfection efficiencies in all cell lines as compared
compared to lipofectamine,
lipofectamine, they
or
selectively transfected
selectively
trans-.'ected
'ected o:
osteoblasts and fibroblasts versus hepatocytes. This indicated that the
connective tissue.
Ca-SiO 2 -DNA complex
complexes
selective toward
toward cellss of the connective
Ca-SiO2-DNA
-omplexes were selective
tissue.
Ca-SiO2-DNA
-omplex
Such
Such
osteoblasts,fibr
selectivity woul
would
extremely useful
useful towards
towards a clinical
clinical
application for bone, which
which is
selectivity
wouldJ be ec
e)
extremely
ical application
oth ost
comprised of both
osteoblasts
,oth
ost
and fibroblasts.
Ca-SiO2-DNA
The selectivity of the Ca-SiO2-DNA
cells would be especially valuable forr gene
mnecells
gene delivery
delivery to a fracture
fracture site.
site.
complexes for bone
Nemne
80
80
o
:.-, ---- .......------......
^~~~`~
~
'' '-~" ^~x~`~"*"'"""~""'1'""
........
"~~'I~'""`~~"~'"~"~
60
.2
m
40
0
·
20
0
osl
Osteoblasts
osl
Fib rob lasts
Hepatocytes
Hepa
ato cytess
ransfectFigure.
4.12. Transfection
T
Figure. 4.12.
efficiencies in osteoblasts, fibroblasts
TransfectFibroblastsand hepatocytes with ()
(1
lipofectamine-D'.1A,and Ca-SiO 2-DNA complexes formed
lipofectamine-DNA,
with (0)
() 1-nm
10-nm and
and (0
() 50-nm
lipofectamine-DNA,
-d with
50-nm CaaSiO 2 particles.
particles. DNA loading: 24 g/ml; vector loading:
0.15
wt%.
Values
are
mean
+
SiO2
,ig: 0. 15
standard error
error of the
the mez
mez n = 3.
standard
mean;
4.3.3.4. Cellular
Cellular Prolifet
Proliferationafter Transfection
4.3.3.4.
Despite its
i is signit
significant transfection efficiency, lipofectamine's
Despite
signit
)ofectamine's toxicity would prevent
applii nation [.[[25-26]. In contrast, the Ca-SiO 2 nanoparticles
nanoparticles entered
entered cells
cells without
without
its clinical application
exhibiting any
any toxicity
exhibiting
toxicity [6-7].
In this section, the potential toxicity of Ca-SiO2-DNA
Ca-SiO 2-DNA
:valuate( by conducting proliferation studies
complexes was
was evaluated
complexes
evaluate(
idies in conjunction with transfection
79
experiments. Figure 4.13 shows the effects of particle size and Ca loading on both cellular
proliferation and transfection efficiency.
The data indicated that for a given particle size, the Ca loading required for optimal
cell proliferation was lower than that required for optimal transfection.
Optimal proliferation
required a smaller Ca loading, which corresponded well with that required for optimal cellular
uptake (see Fig. 4. 1). As discussed earlier, Ca played the additional role of condensing DNA
in nanoparticle-DNA complexes, so a higher Ca loading was needed for optimal transfection.
We also note that greater cell proliferation was achieved with complexes formed with 50-nm
Ca-SiO 2 particles than with 10-nm Ca-SiO 2 particles. This was because for the same nominal
Ca loading, the sizes of Ca-SiO 2-DNA complexes formed with 10-nm and 50-nm Ca-SiO
particles were similar (see Fig. 4.4).
2
This indicated that in addition to uptake of the
complexes, it was likely that a greater number of free 10-nm Ca-SiO2 particles (that were not
bound to DNA) were ingested relative to free 50-nm Ca-SiO2 particles. This increased
ingestion of free 10-nm particles as opposed to free 50-nm particles might have effectively
resulted in a higher intracellular rise in Ca2 + concentration for the former. This rise in Ca2+
concentration for the cells that were transfected with complexes formed with 10-nm Ca-SiO2
particles might have been above the threshold of optimal Ca loading required to stimulate cell
proliferation.
onn%
C nr
160
40
LUU
0
§
120
30
.o
i'
,-
-
80
20 0*
40
10
0
L
I-
0
0.0
0.5
1.0
1.5
2.0
Nominal Ca Loading (wt%)
Figure 4.13. Effect of Ca loading on the cellular proliferation (closed symbols) and the
transfection efficiency (open symbols) of osteoblasts transfected with Ca-SiO2-DNA
complexes formed with ()
10-nm and () 50-nm Ca-SiO 2 particles.
80
4.3.3.5. Total GFP Expression due to Transfection
While transfection efficiency is a very useful parameter for evaulating gene delivery
systems, the most therapeutically relevant parameter for our application is the amount of
protein produced due to transfection. Transfection efficiency is given by (the number of
transfected cells)/(the number of viable cells). A transfected cell could be transfected with
one or more copies of a plasmid, and as a result, each transfected cell could produce a
different amount of the expressed protein [27]. In the case of our model system, the protein of
interest was GFP. In a system of interest to fracture repair, the protein of relevance would be
a growth factor such as bone morphogenetic protein (BMP).
A
fnnn
'tVVV
C
" 3000
o
E
:L. 2000
C
0
·
1000
o
aO
0
0.00
0.25
0.50
0.75
1.00
1.25
Nominal Ca Loading (wt%) in Ca-SiO 2
1.50
Lipofectamine
Figure 4.14. Effect of Ca loading on the GFP expression of osteoblasts transfected with CaSiO2 -DNA complexes formed with (i)10-nm and () 50-nm particles. GFP expression of
osteoblasts transfected with lipofectamine-DNA is indicated by ().
The total GFP expression was calculated by FACS as the total shift in the mean
fluorescence distribution of the transfected cell population compared to the control cell
population with no DNA.
Figure 4.14 indicated that the combined effect of transfection
efficiency and cellular proliferation for some Ca-SiO 2-DNA complexes contributed to a total
GFP expression that approached that of cells transfected with lipofectamine.
The fact that
certain Ca-SiO 2-DNA complexes could cause osteoblasts to produce such high levels of GFP
81
without any toxicity in vitro suggested that they would be extremely useful as in vivo gene
delivery systems.
4.4. Summary
We have used our previous understanding of cell-nanoparticle interactions to develop
a gene delivery system based on Ca-SiO2-DNA complexes. A model system was constructed
using plasmid DNA encoding for GFP. Ca-SiO2 parameters such as particle size, Ca loading
and surface charge were examined for their effects on DNA binding affinity and complex size.
Specifically,
osteoblasts
10-nm and 50-nm Ca-SiO 2 particles led to the successful transfection
without any toxicity.
of
The resulting Ca-SiO 2 -DNA complexes were able to
transfect connective tissue cells selectively versus hepatocytes, unlike lipofectamine-DNA
complexes.
The optimal Ca-SiO2-DNA complexes induced GFP expression levels
comparable to that of lipofectamine-DNA complexes. The ability of Ca-SiO 2 nanoparticles to
transfect osteoblasts efficiently and selectively without any cytotoxicity effects makes a very
compelling case for their use as gene delivery vehicles for bone regeneration.
82
4.5. References
[1]
Cheng, J.W., Jiang, W., Phillips, F.M., Haydon, R.C., Peng, Y., Zhou, L., Luu, H.H.,
An, N.L., Breyer, B., Vanichakarn, P., Szatkowski, J.P., Park, J.Y., and He, T.C., J.
Bone Joint Surg. Am. 85A, 1544 (2003).
[2]
Urist, M., Science 150, 893 (1965).
[3]
Bonadio, J., Goldstein, S.A., and Levy, R.J., Adv. Drug Deliv. Rev. 33, 53 (1998).
[4]
Fang, J., Zhu, Y.-Y., Smiley, E., Bonadio, J., Goldstein, S.A., McCauley, L.K.,
Davidson, B.L., and Roessler, B.J., Proc. Natl. Acad. Sci. 93, 5753 (1996).
[5]
Jordan, M., Schallhorn, A., and Wurm, F.M., Nucleic Acids Res. 24, 596 (1996).
[6]
Moudgil, S., and Ying, J.Y., to be submitted to Biomater.
[7]
Moudgil, S., and Ying, J.Y., to be submitted to Biomater.
[8]
Olbrich, C., Bakowsky, U., Lehr, C.M., Muller, R.H., and Kneuer, C., J. Control.
Release 77, 345 (2001).
[9]
Luk, K.D.K., Chen, Y., Cheung, K.M.C., Kung, H.F., Lu, W.W., and Leong, J.C.Y.,
Biochem. Biophys. Res. Commun. 308, 636 (2003).
[10]
Jeong, J.H., and Park, T.G., J. Control. Release 82, 159 (2002).
[11]
Tiyaboonchai, W., Woiszwillo, J., and Middaugh, C.R., Eur. J. Pharm. Sci. 19, 191
(2003).
[12]
Iler, R.K., "The Chemistry of Silica," John Wiley & Sons, New York, 1979, p. 382.
[13]
Kneuer, C., Sameti, M., Haltner, E.G., Schiestel, T., Schirra, H., Schmidt, H., and
Lehr, C.-M., Int. J. Pharm. 196, 257 (2000).
[14]
Liu, G., Li, D.S., Pasumarthy, M.K., Kowalczyk, T.H., Gedeon, C.R., Hyatt, S.L.,
Payne, J.M., Miller, T.J., Brunovskis, P., Fink, T.L., Muhammad, O., Moen, R.C.,
Hanson, R.W., and Cooper, M.J., J. Biol. Chem. 278, 32578 (2003).
[15]
Pang, S.W., Park, H.Y, Jang, Y.S., Kim, W.S., and Kim, J.H., Colloid. Surface. B 26,
213 (2002).
[16]
Tachibana, T., Harashima, H., Shinhara, Y., and Kiwada, H., Adv. Drug Deliv. Rev. 52,
219 (2001).
[17]
Shimizu, T., Okon, T., Ohba, T., and Inokuchi, Y., US Patent 5149748 (1992).
[18]
Frolov, I., Hoffman, T.A., Pragai, B.M., Dryga, S.A., Huang, H.V., Schlesinger, S.,
and Rice, C.M., Proc. Natl. Acad. Sci. 93, 11371 (1996).
83
[19]
Mizuguchi, H., Nakagawa, T., Morioka, Y., Imazu, S., Nakanishi, M., Kondo, T.,
Hayakawa, T., and Mayumi, T., Biochem. Biophys. Res. Commun. 234, 15 (1997).
[20]
Godbey, W.T., Wu, K.K., and Mikos, A.G., Proc. Natl. Acad. Sci. 96, 5177 (1999).
[21]
Pollard, H., Remy, J.-S., Loussouarn, G., Demolombe, S., Behr, J.-P., and Escande,
D., J. Biol. Chem. 273, 7507 (1998).
[22]
Schaffer, D.V., Fidelman, N.A., Dan, N., and Lauffenburger, D.A., Bioeng. Biotech.
67, 598 (2000).
[23]
Kneuer, C., Sameti, M., Bakowsky, U., Schiestel, T., Schirra, H., Schmidt, H., Lehr,
C.M., Bioconjugate Chem. 11, 926 (2000).
[24]
Sandhu,
K.K.,
McIntosh,
C.M.,
Simard,
J.M.,
Smith,
S.W.,
Rotello,
V.M.,
Bioconjugate Chem. 13, 3 (2002).
[25]
Gebhart, C.L, and Kabanov, A.V., J. Control. Release 73, 401 (2001).
[26]
Wu, J., Lizarzaburu, M.E., Kurth, M.J., Liu, L., Wege, H., Zern, M.A., and Nantz,
M.H., Bioconjugate Chem. 12, 251 (2001).
[27]
Batard, P., Jordan, M., and Wurm, F.M., Gene 270, 61 (2001).
84
Chapter 5 - Recommendations for Future Work
In this thesis, organosilicate nanoparticles were developed as gene delivery vehicles
for bone cells.
Fundamental cell-nanoparticle interactions were studied in vitro to help
determine which material properties had the most significant effects on cellular behavior.
Specifically, the effects of particle size and composition on osteoblast uptake, proliferation
and protein expression were investigated. While the cell culture experiments performed on
three different cell types yielded useful information, future research should use these in vitro
results to formulate various in vivo studies.
The cellular proliferation and protein expression studies indicated that the
nanoparticles were not toxic to osteoblasts, allowed the cells to retain their osteoblastic
phenotype, and did not induce any significant inflammatory response. To confirm these
results in vivo, the nanoparticles should be administered to an animal bone defect site, and
histological and biochemical characterization should be performed. The increased osteoblast
proliferation and the phenotype retention that were evident in vitro would presumably lead to
increased bone formation and increased production of bone phenotypic markers in vivo. Also,
the retention of healthy cell morphology and the absence of macrophage cells at the defect site
would suggest minimal inflammatory response.
The in vitro results also indicated that osteoblast uptake was maximized with Ca-SiO2
nanoparticles
of < 100 nm.
For in vivo uptake studies, the histology sections should be
visualized for any nanoparticle ingestion by cells to see if the trends that were evident in cell
culture would correspond to an animal model. While in vitro uptake was quantified precisely
through a fluorescence-based assay, in vivo uptake would most likely have to be performed
more qualitatively using analytical microscopy techniques.
The understanding gained by studying the effect of nanoparticle parameters on cellular
behavior was used to design a model gene delivery system for bone cells. While the plasmid
used in our model system encoded for green fluorescent protein (GFP), future research should
use a plasmid that encodes for a bone morphogenetic protein (BMP) that would aid bone cell
growth and differentiation.
BMP-2 and BMP-7 have been shown to induce significant
osteogenic activity, so a plasmid that encodes for one of these proteins would be ideal. The
extension of our model system to such a plasmid should be fairly straightforward, since the
portion of the plasmid that contains the gene of relevance is extremely small. Thus, it is very
85
likely that a plasmid that encodes for BMP-2 would interact with the nanoparticles similar to
the GFP-encoding
plasmid.
In vivo studies should be performed
on the Ca-SiO 2-DNA
complexes by injecting them into a defect site and assaying for BMP production and bone
formation.
Histology should also be performed to assess the targeting ability of these
complexes for cells of the connective tissue by visualizing nanoparticle uptake through
analytical microscopy techniques.
The in vitro results obtained in this thesis are very useful in that they quantified cellmaterial interactions in detail.
While animal studies are often limited by sample size and
number, this was not the case for the cell culture experiments.
While the clinical potential of
these nanoparticles can only be verified and assessed through in vivo studies, the extensive in
vitro characterization performed in this thesis would set the framework and understanding for
the design of an appropriate in vivo model.
86
Chapter 6 - Conclusions
We have designed bioactive organosilicate nanoparticles for gene delivery to bone
cells. Monodisperse organosilicate nanoparticles were synthesized via surfactant-stabilized
sol-gel processing. The composition of these nanoparticles was modified through the addition
of various surface dopants, and particle size was controlled via sol pH and alkoxide
concentration.
Ca-SiO2 nanoparticles of < 100 nm were found to maximize osteoblast
proliferation and cellular uptake.
The effects of particle size and composition on cellular protein expression were also
characterized. Osteoblastic phenotype was retained in the presence of various organosilicate
nanoparticles, and cytokine expression was found to be minimal. These results provided
strong support for the futher investigation of these nanoparticles as gene delivery vehicles for
osteoblasts.
The nanoparticles were complexed with plasmid DNA encoding for green fluorescent
protein (GFP). The effects of nanoparticle size, composition and surface charge on complex
formation were studied.
These findings were used to design a Ca-SiO 2-DNA gene delivery
system that transfected osteoblasts and fibroblasts selectively versus hepatocytes. The ability
of these nanoparticles to transfect osteoblasts selectively without inducing any in vitro
toxicity makes them very attractive for gene delivery applications aimed at bone regeneration.
87