Tailoring the (bio)activity of polymeric and

metal oxide nano- and microparticles in

biotic and abiotic environments

Sathish Ponnurangam

Submitted in partial fulfillment of the

requirements for the degree of

Doctor of Philosophy

in the Graduate School of Arts and Science

Columbia University

2012

© 2012

Sathish Ponnurangam

All Rights Reserved

ABSTRACT

Tailoring the (bio)activity of polymeric and metal oxide nano- and microparticles in biotic

and abiotic environments

Sathish Ponnurangam

Polymeric and metal oxide micro- and nanoparticles are being increasingly introduced into

biomedical applications such as tissue engineering as well as in consumer products, which has boosted

extensive research towards developing predictive paradigms of their (bio-)activity. The core hypotheses

which are tested in the four interrelated studies of this work is that the (bio-)activity of the particles is

defined not only by their intrinsic properties such as the composition/structure, functional groups, surface

charge, and size/morphology, but also on the concentration of particles which in turn is determined by

specific applications. In addition, the (bio)activity of the particles can be controlled by the application-

specific biomolecules or surfactants. These hypotheses are tested on polymeric and metal oxide particles

from the perspective of their application in tissue engineering of articular cartilage and consumer products

(antioxidant additives and dyes), respectively. The modeling of the transport properties of biomaterials, as

well as of the adsorption properties of metal oxide nanoparticles can help to determine or interpret the

observed relationships.

To test this hypothesis, the first study addresses the effect of the concentration of the polysebacic

anhydride microparticles on the cellular viability of chondrocytes (cartilage cells) and on the properties of

in vitro grown tissue. The polysebacic anhydride based microparticles, which are one most common

biocompatible and biodegradable polymer, is used in this study for short term delivery of bioactive

molecules such as growth factors and Vitamin C in articular cartilage tissue engineering. An in situ short

term delivery of nutrients and growth factors is hypothesized to overcome the problem of the decrease in

the diffusion caused by the progressive elaboration of extracellular matrix in 3D agarose scaffolds with

time. I found that, contrary to the reported safety of polyanhydrides, the polysebacic anhydride -based

microparticles are cytotoxic to chondrocytes (cartilage cells) at concentrations relevant to delivery in

articular cartilage engineering. The cytotoxicity is explained by the lipotoxicity from the polysebacic

anhydride degradation products. The available data allow suggesting that the bioactivity of polysebacic

anhydride polymer is collectively and uniquely determined by its degradation rate and hydrophobicity

which requires further verification. It is found that the cytotoxicity of polysebacic anhydride can be

mitigated significantly by administering of bovine serum albumin. The bovine serum albumin-based

mitigation strategy makes polysebacic anhydride microparticles effective in the short-term delivery of

growth factors in articular cartilage engineering. With this bovine serum albumin-based mitigation

strategy, the loading concentrations of fast degrading polyanhydride-based particles can be increased to

levels relevant to in situ delivery requirements for articular cartilage tissue engineering. This opens up

possibility of obtaining tissues with spatially homogenous properties.

Another critical factor that arises due to the presence of the polymeric particles is the modification

of transport properties of agarose hydrogel scaffold. To address this issue, a model is developed to

extract the pore/void size as a function of agarose concentration from the turbidity measurements.

The second study verifies the importance of functional groups of polymer particles in the model

system of polyacrylate-based nanoparticles and chondrocytes in a 3D agarose hydrogel scaffold. From

this study, the charge on the functional groups of the polymers is found to have a significant effect on

their bioactivity. Specifically, the polyacrylate-based nanoparticles are shown to be biocompatible (at

0.2% w/v) to chondrocytes in terms of cellular viability. However, they exhibit bioactivity which is

detrimental to the synthesis of extracellular matrix even at such low concentrations. An inverse correlation

is established between mechanical strength and the negative surface charge of these nanoparticles. The

cause for poor matrix synthesis is suggested to be the disruption of the inter-cellular signaling process by

higher anionicity. The main impact of this study is that the utility of anionically charged polymeric particles

in articular cartilage engineering should be scrutinized carefully, since it is shown that for a clinically

relevant scaffold system such as agarose hydrogels, these particles may act detrimentally.

Study 3 focuses on the (bio-)activity of redox-active metal oxide nanoparticles toward

chondrocytes. For the first time, nanoceria is shown to greatly improve the biochemical as well as

mechanical properties of the cartilage tissue grown in vitro. The effect is explained by the unique redox

properties of the nanoparticles. Specifically, nanoceria is biocompatible toward chondrocytes at

concentrations at least up to 1000 μg/mL, while the constructs cultured with embedded nanoceria at 1000

μg/mL have mechanical and biochemical properties significantly better than those of the control.

Interestingly, the bioactivity of nanoceria in combating inflammatory chemical such as interleukin-1α

depends on the mode of its administration. In particular, when nanoceria is suspended in the growth

medium, it offered significant chondroprotectivity against interleukin-1α and partial prevention of the

matrix degradation. However, when nanoceria is directly embedded into the constructs, it does not offer

any protection against interleukin-1α. These results show the great potential of nanoceria in improving

tissue properties and combating the arthritic inflammation. If the mode of nanoceria administration and

cellular uptake are further optimized, the long term protection of cartilage can be achieved.

In contrast to nanoceria, ferric (hydr)oxide nanoparticles, which are also redox active,

detrimentally affect the growth of cartilage tissue. The adverse effect increases with increasing

nanoparticle size and confirms the initial hypothesis of the size dependence of bioactivity of ferric

(hydr)oxide nanoparticles, which is based on earlier observation from our group of the catalytic activity of

these nanoparticles. Specifically, the mechanical properties of the tissue constructs embedded with

hematite and ferrihydrite nanoparticles degrade with increasing nanoparticle size. On the other hand, the

cellular viability remains similar among all these sample sets. These two results together imply that the

bioactivity is directly related to the redox potential and the catalytic redox mechanism (including either the

deactivation or generation of reactive oxygen species) of the metal oxide nanoparticles toward redox

active constituents at the biointerface.

Study 4 addresses the effects of nanoparticle size and morphology on the adsorption and

dispersion of the ferric (hydr)oxide nanoparticles in the presence of surfactants, which is relevant to the

dispersion of the nanoparticles in their application in image and contrast agents, cancer therapy,

ferrofluids, paints, and cosmetics. The first experimental evidence is obtained for the dependence of the

surface density, speciation, and packing order of adsorbed fatty acids on the nanoparticle size and

morphology. The conditions under which fatty acids form self-assembled monolayers and bilayers on

such nanoparticles in water are distinguished and the electric polarization of the nanoparticles is

demonstrated to be a powerful tool for manipulating the interfacial properties of the nanoparticles.

Specifically, an increase in nanosize improves the adsorption capacity and affinity of hematite

nanoparticles, in agreement with the nanosize-induced changes in the structural and electronic properties

of the nanoparticles. Consequently, an increase in nanosize of hematite nanoparticles improves the

packing order of laurate and hence hydrophobicity of the nanoparticles provided a similar hexagonal habit

of these nanoparticles. It was also observed that, independent of morphology, the dispersibility of the

nanoparticles by fatty acids in water degrades as nanoparticle size decreases. This effect is attributed to

the decrease in the adsorption capacity which ends up in the formation of a loosely packed bilayer. This

implies that the surface electronic and structural properties, which can be controlled by size for metal

oxide nanoparticles, offer important ways to manipulate adsorption properties.

Morphology of the nanoparticles is critical for the self-assembly if (i) the majority of the molecules

are chemisorbed and (ii) the morphology-determined surface arrangement of the adsorption sites is

geometrically incompatible with a surfactant ordering. The surface ratio of the physisorbed/chemisorbed

surfactant depends on the surface charge of nanoparticles. Negative polarization suppresses

chemisorption. This effect can be used to remove the geometrical constraints for self-assembly. The

dominance of the chemisorbed (inner-sphere) complexes results in the isotropic surface filling by fatty

acids, while the prevalence of outer-sphere complexes favors the cluster-like filling pattern.

A new approach to density functional theory modeling of the in situ FTIR spectra of adsorbed

species is developed to solve the technically challenging problem of determining the adsorbed structures

especially at the magnetic oxide-water interfaces. This new approach constitutes searching of the

adsorbed structure that provides the best correspondence of the trends of the theoretical and

experimental spectra upon variations of the solution pH, ionic strength and protonation-deprotonation of

the carboxylate group, instead of conventional approach of searching for the structure that provides the

best one-to-one correspondence between the theoretical and experimental spectra. The new density

functional theory approach is used to interpret the vibrational spectroscopic data in obtaining information

on adsorption forms and packing of fatty acids on ferric (hydr)oxide nanoparticles. These molecular

information is used to interpret the dispersion properties of ferric (hydr)oxide nanoparticles. The

developed approach advances the utility of density functional theory modeling of adsorption phenomena

by incorporating physical variables such as pH, surface charge and ionic charge. This new approach, if

applied, can potentially solve many of controversies in literature in interpreting the adsorption forms

obtained from spectroscopic and density functional theory data.

Although there is still no general paradigm predicting the (bio-)activity for both the polymeric and

metal oxide micro- and nano-particles, the experimental, methodological and theoretical results presented

in these studies are an important step and base towards developing such a paradigm. In this context, the

following factors determine the bioactivity of the polymeric particles considered (polyanhydrides and

polyacrylates): degradation rate, hydrophobicity, functional group and surface charge. In addition to

above parameters, concentration of these particles with respect to cell density is critical factor determining

their (bio-)activity. Among the numerous physico-chemical variable that can affect the (bio-)activity of

metal oxide nanoparticles, the catalytic redox properties and the redox mechanism are general predictors

for ceria and ferric (hydr)oxide nanoparticles. These properties can be varied by varying nanosize of the

nanoparticles. Both nanosize and morphology of nanoparticles can also significantly affect their

adsorption properties and, through them, their bioactivity and bioavailability. In addition, the density

functional theory-based methodology to interpret the adsorption phenomena developed in this work can

further augment the possibility of formulating such a general paradigm for (bio-)activity.

i

Table of Contents

Table of Contents ........................................................................................................................................... i

List of abbreviations ...................................................................................................................................... v

List of Figures ............................................................................................................................................... vii

List of Tables ................................................................................................................................................ xii

List of Schemes ........................................................................................................................................... xiii

Acknowledgements ..................................................................................................................................... xiv

Chapter 1 Introduction, Literature review, and Goals ................................................................................... 1

1.1 Literature review and gaps in the current knowledge ......................................................................... 1

1.1.1 Bio-activity of polysebacic anhydride microparticles used for in situ delivery in articular cartilage

engineering............................................................................................................................................ 1

(a). Challenges in articular cartilage engineering.............................................................................. 2

(b). Rationale for using polymeric microparticles for in situ delivery of bioactive molecules ............ 5

(c). Comparison of biomedical properties of natural and synthetic polymeric biomaterials .............. 6

(d). Sebacic acid based polyanhydride polymer for delivery device in articular cartilage

engineering: Advantages and challenges ......................................................................................... 8

1.1.2. Bioactivity of polyacrylate nanoparticles in articular cartilage tissue engineering: Effects of

functional groups ................................................................................................................................. 11

1.1.3 (Bio-)chemical activity of metal (hydr)oxide nanoparticles ......................................................... 13

Current paradigm of the cytotoxicity and biocompatibility of metal oxide nanoparticles................. 14

1.1.4 (Bio-)activity of engineered ceria oxide nanoparticles: Anti-inflammatory and reactive-oxygen

species scavenging properties ............................................................................................................ 16

(a). Structural, electronic, and abiotic redox properties of CeO2 nanoparticles .............................. 16

(b). Rationale for selection of nanoceria to combat inflammation in articular cartilage .................. 18

1.1.5 (Bio-)activity of ferric (hydr)oxide nanoparticles: adsorption, dispersion and biocompatible

properties ................................................................................................................................................ 20

(a) Structural, electronic, and abiotic redox properties of hematite nanoparticles .......................... 20

(b). Size, shape and morphology dependence of (bio-)activity of ferric (hydr)oxide nanoparticles 23

1.1.6. Challenges in modeling the adsorption of fatty acids at the metal oxide nanoparticle-water

interface ............................................................................................................................................... 28

1.2. Goals and objectives ........................................................................................................................ 30

Chapter 2 Materials and Methods ............................................................................................................... 33

2.1 Materials ............................................................................................................................................ 33

2.2 Tissue culture studies ....................................................................................................................... 33

2.3 Characterization and physico-chemical experimental methods ....................................................... 35

2.4 Density Functional Theory (DFT) Methodology ................................................................................ 38

ii

2.5 Error Analysis of data ........................................................................................................................ 39

Chapter 3 Synthesis and characterization of polymers and polymeric micro- and nano-particles ............. 40

3.1 Synthesis of polysebacic anhydride polymer using melt condensation method ............................... 40

3.2. Functionalization of polysebacic anhydride with fluorescein label and vitamin C ........................... 42

3.2.1 Functionalization with Fluorescein ............................................................................................. 42

3.2.2 Functionalization with vitamin C (L-ascorbic acid) ..................................................................... 42

3.3 Characterization of polysebacic anhydride and its derivatives using NMR and Raman

spectroscopies ........................................................................................................................................ 44

3.4 Fabrication of polysebacic anhydride microparticles and encapsulation of bioactive molecules using

double emulsion method ......................................................................................................................... 46

3.5 Characterization and release kinetics of polysebacic anhydride microparticles using optical,

electron, and confocal fluorescence microscopies, as well as UV-Vis absorption spectroscopy. .......... 48

3.6. Synthesis and characterization of acrylate nanoparticles ................................................................ 50

3.6.1. Ternary phase diagrams ........................................................................................................... 50

3.6.2.Synthesis of poly(acrylic acid) nanoparticles and its functional derivatives ............................... 51

3.6.3.Size determination using SEM and dynamic light scattering ..................................................... 53

3.6.4 Atomic composition using X-Ray photoelectron spectroscopy .................................................. 54

3.6.5 Surface charge using electrophoretic mobility (Zeta potential) measurements ......................... 54

3.7 Pore-size determination of agarose-based hydrogel as a function of agarose concentration and

presence of polysebacic anhydride microparticles ................................................................................. 55

Chapter 4 Synthesis and characterization of ferric (hydr)oxide nanoparticles ........................................... 62

This chapter presents the synthesis procedures used to obtained hematite nanoparticles of different

sizes, as well as the characterization of these nanoparticles using TEM, HRTEM, XRD, Raman, XPS,

acid-base titration, and zeta potential methods. ..................................................................................... 62

4.1 Synthesis of hematite nanoparticles ................................................................................................. 62

4.2 Characterization of ferric (hydr)oxide nanoparticles ......................................................................... 62

Chapter 5 Bioactivity of polymeric particles in articular cartilage engineering: In situ delivery using

poly(sebacic anhydride) microparticles and functional group effects of poly(acrylate) based nanoparticles

.................................................................................................................................................................... 68

5.1 Effects of in situ delivery of bioactive molecules on the chondrocytes cultured in 3D scaffold ........ 68

5.1.2. Dose dependent cellular viability in the presence of poly(sebacic anhydride) ......................... 69

5.1.3 Mechanisms of the cytotoxicity of poly(sebacic anhydride) microparticles ................................ 69

(a). Are cytotoxic effects of PSA-microparticles on chondrocytes due to local pH changes? Testing

the hypothesis through increased buffer capacity........................................................................... 71

(b). Are cytotoxic effects due to cellular uptake PSA-microparticles or its degradation products and

the resulting lipotoxicity? Confocal fluorescence studies ................................................................ 72

5.1.4 Mitigation of the cytotoxicity of PSA on chondrocytes: positive impact of bovine serum albumin

(BSA) ................................................................................................................................................... 73

iii

5.1.5 Delivery of TGF-β3 using PSA-microparticles in BSA-based system ........................................ 74

5.2 The effects of functionalization of polyacrylate nanoparticles on the cellular viability and metabolism

................................................................................................................................................................ 76

5.2.1 Sterilization tests: ....................................................................................................................... 76

5.2.2 Biocompatibility tests: ................................................................................................................. 76

5.2.3 Three-dimensional tissue culture studies with acrylate nanoparticles ....................................... 77

5.3 Conclusions: ...................................................................................................................................... 80

Chapter 6 Anti-inflammatory and anti-oxidant bioactivity of ceria nanoparticles in articular cartilage

engineering ................................................................................................................................................. 82

6.1 Experimental design .......................................................................................................................... 82

6.2 Effects of nanoceria embedded into the 3D-agarose scaffold on tissue growth as well as combating

inflammation due to interleukin-1α .......................................................................................................... 84

6.3 Effects of nanoceria suspended in the growth medium on tissue growth as well as combating

inflammation due to interleukin-1α .......................................................................................................... 87

6.4 Conclusions ....................................................................................................................................... 88

Chapter 7 Dependence of the (bio)activity of hematite NPs on their size, morphology, and surface charge

.................................................................................................................................................................... 89

7.1 Size, shape and surface-charge dependent bioactivity of hematite NPs on chondrocytes .............. 89

7.2 Size, shape and surface-charge dependent interaction of hematite NPs with sodium laurate

(dodecanoate) ......................................................................................................................................... 93

7.2.1 Dispersion and adsorption properties of hematite NPs in the presence of sodium laurate. ...... 93

7.2.2 Speciation of carboxylic groups in the first monolayer. .............................................................. 97

a. Inner-sphere monodentate mononuclear complex (ISMM): ....................................................... 99

b. Outer-sphere hydration shared complex (OS): ........................................................................... 99

c. Protonated outer-sphere hydration shared complex (OS-H): ................................................... 100

7.2.3 Packing order of hydrocarbon chains. ..................................................................................... 101

7.2.4 Controls of NP dispersion by fatty acids. Formation of a bilayer. ............................................ 102

7.2.5 Effect of negative polarization on the structure of the first monolayer. .................................... 104

7.2.6 Speciation of adsorbed fatty acid. Effects of the oxide acidity. ................................................ 106

7.2.7 Conditions for SAM formation. Effects of the NP size, morphology, and surface charge. ...... 107

7.3 Conclusions ..................................................................................................................................... 109

Chapter 8 A new approach to DFT modeling of the adsorbed species at the metal (hydr)oxide NPs-water

interfaces ................................................................................................................................................... 111

8.1 A combined DFT-FTIR approach to resolve the adsorption forms at the oxide-water interface .... 111

8.2 A combined DFT-FTIR approach to resolve the adsorption forms at the oxide-water interface .... 113

8.3 Identification of adsorption forms of fatty acid on ferric (hydr)oxide NPs ....................................... 116

8.2.1 FTIR results .............................................................................................................................. 117

iv

8.2.2 DFT modeling of carboxylate binding to hematite surface........................................................... 119

8.4 Conclusions ..................................................................................................................................... 123

Chapter 9 Summary .................................................................................................................................. 124

Chapter 10 Suggested Future Studies ...................................................................................................... 131

REFERENCES .......................................................................................................................................... 133

APPENDIX ................................................................................................................................................ 150

A. Estimation of buffer capacity needed to offset the drop in pH due to polsebacic anhydride

degradation ........................................................................................................................................... 150

B. Determination of wavelength exponent ............................................................................................ 151

v

List of abbreviations

AFM Atomic force microscopy

ANOVA Analysis of variance

AOT Aersol-OT

ATR Attenuated total reflection

BET Brunauer-Emmett-Teller

BSA Bovine serum albumin

CM Culture medium

CMC Critical micellar concentration

COSMO Continuum solvation model

DFT Density functional theory

DNP Double numerical polarization

ECM Extracellular matrix

FH Ferrihydrite

fPSA Fluoresceinated-polysebacic anhydride

FTIR Fourier transform infrared spectroscopy

GAG(s) Glycosaminoglycan(s)

GRGD Glycine-arginine-glycine-aspartic acid

HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid)

HLB Hydrophilic-lipophilic balance

IEP Isoelectric point

IL-1α Interleuklin-1alfa

ISBB Innersphere bidentate binuclear

ISMM Innersphere monodentate mononuclear

IRE Internal reflection element

ML Monolayer

MSC Mesenchymal stem cells

NMR Nuclear magnetic resonance

NP(s) Nanoparticle(s)

OS Outersphere

PAA Polyacrylic acid

PAM Polyacrylamide

PAAP Polyacrylic acid-phosphonate

P(CPP-SA) Poly(carboxy phenoxy propane-co-sebacic anhydride)

vi

PEG Poly(ethylene glycol)

PEGF Poly(ethylene glycol-co-fumarate)

PLGA Poly(lactic-co-glycolic acid)

PSA, P(SA) Polysebacic anhydride

PSA-M Polysebacic anhydride-microparticles

PZC Point of zero charge

RGD Arginine-glycine-aspartic acid

ROS Reactive oxygen species

RNS Reactive nitrogen species

SA Sebacic acid

SAED Small area electron diffraction

SAM Self-assembled monolayer

SEM Scanning electron microscope

SL Sodium laurate

SOD Superoxide dismutase

TDW Triple distilled water

TEM Transmission electron microscope

TGF-β3 Transformer growth factor beta 3

THF Tetrahydrofuran

TMAO Trimethyl-amine oxide

XPS X-ray photoelectron spectroscopy

XRD X-Ray Diffraction

vii

List of Figures

Figure 1.1 A typical cross section of articular cartilage showing four zones4 .............................................. 2

Figure 1.2 The fcc cell of CeO2 with the fluorite structure. Ce is white, O is red. ...................................... 16

Figure 1.3 A model of the reaction mechanism for the oxidation of hydrogen peroxide by nanoceria and

the regeneration via reduction by superoxide. An oxygen vacancy site on the nanoceria surface (1)

presents a 2Ce4+

binding site for H2O2 (2), after the release of protons and two-electron transfer to the two

cerium ions (3) oxygen is released from the now fully reduced oxygen vacancy site (4). Subsequently

superoxide can bind to this site (5) and after the transfer of a single electron from one Ce3+

, and uptake of

two protons from the solution, H2O2 is formed (6) and can be released. After repeating this reaction with a

second superoxide molecule (7) the oxygen vacancy site returns to the initial 2Ce4+

state (1). It is also

possible that the third Ce3+

indicated, which gives rise to the oxygen vacancy, could participate directly in

the reaction mechanism. The square Ce–O matrix is shown here only to illustrate the model and does not

correspond to the actual spatial arrangement of the atoms in the crystal structure. Reproduced from

Ref.70

........................................................................................................................................................... 18

Figure 1.4 Crystal structure of α-Fe2O3. Red atoms are O; Blue atoms are Fe. ........................................ 21

Figure 1.5 Conceptual scheme of electrochemical mechanism of redox reaction Ox2 + Red1 Ox1 +

Red2 in the dark.116

FL = Fermi level, VB = valence band, CB = conduction band, △s = gradient of

surface potential. ......................................................................................................................................... 22

Figure 3.1 NMR spectra of polymers on their precursors; (a) Sebacic acid, (b) Poly(sebacic anhydride)

(inset: expanded and curve fitted region from 2.18-2.48 ppm), (c) Ascorbic acid (d) Poly(Sebacic

anhydride)- Ascorbic acid. ........................................................................................................................... 45

Figure 3.2 Raman spectra of polymers on their precursors; Poly(sebacic anhydride) and Sebacic acid.. 46

Figure 3.3 SEM images of PSA-microparticles encapsulated with casein-FITC after 24 hours in aqueous

suspension. All the images are from the same batch of microparticles but at different magnifications. .... 48

Figure 3.4 Particle size distribution of PSA-microparticles with casein-FITC after 24 hours in aqueous

suspension. The size distribution was obtained from SEM image analysis performed using ImageJ

program. ...................................................................................................................................................... 49

Figure 3.5 L-ascorbic acid loaded polyanhydride microparticles embedded in agarose scaffold. Green

color indicates the fluorescein-functionalized PSA polymeric shell. Black core is the encapsulated L-

ascorbic acid. Both the images are from the same microparticle sample .................................................. 49

Figure 3.6 Release profile of amitriptyline drug from poly(sebacic anhydride)-based microparticles........ 50

Figure 3.7 Phase diagrams of (a) hexane/water/AOT (b) hexane /(water+acrylic acid)/AOT systems at

room temperature. Only inverse emulsion region is indicated by colors .................................................... 52

viii

Figure 3.8 Phase diagram of hexane /(water+acrylic acid)/AOT systems at 50 C. Only inverse emulsion

region is indicated by the enclosed region. ................................................................................................. 52

Figure 3.9 Synthesis of functionalized polyacrylate nanoparticles, (a) polyacrylic acid (b) polyacrylamide

(c) poly(acrylic acid-phosphonate) (d) poly(acrylic acid-sulfonate); Bisacrylamide is the crosslinker and

ascorbic acid is the initiator ......................................................................................................................... 53

Figure 3.10 SEM micrographs of poly(acrylic acid) particles synthesized using concentrations at (a)

phasepoint-1 (b) phasepoint-2 .................................................................................................................... 53

Figure 3.11 Size of nanoparticles using dynamic light scattering technique, (a) poly(acrylic acid), 130-

340-nm (b) poly(acrylic acid-acrylamide(10%)), 100-320-nm (c) poly(acrylic acid-phosphonate(10%)),

140-260-nm (d) poly(acrylamide) 70-90-nm nanoparticles ......................................................................... 54

Figure 3.12 X-ray photoelectron spectroscopy of poly(acrylic acid) nanoparticle. Carbon 1S spectral

region is shown. .......................................................................................................................................... 55

Figure 3.13 Zeta potential of poly(acrylamide), poly(acrylic acid-phosphonate), poly(acrylic acid)

nanoparticles ............................................................................................................................................... 55

Figure 3.14 Theoretical calculated wavelength exponent as a function of pore size for wavelength range

of 700-800 nm. ............................................................................................................................................ 59

Figure 3.15 Pore size of agarose hydrogel as function of cooling time. The temperature at time t=0 is 95 0C................................................................................................................................................................. 61

Figure 3.16 Effect of agarose concentration (C, % w/v) on the pore size (ξ, nm) of agarose hydrogel..... 61

Figure 4.1 TEM images of hematite and FH NPs. NPs are labeled according to the average size as

measured by the analysis of TEM images. ................................................................................................. 64

Figure 4.2 Typical (a) HR-TEM image and (b) SAED pattern of 38-nm hematite (H38) ........................... 65

Figure 4.3 (a) XRD and (b) Raman spectra of dry hematite and ferrihydrite NPs. “T” marks the peak due

to tetrahedral defects, LO Eu is the peak activated by the lattice disorder.222

............................................ 65

Figure 4.4 Effect of size of hematite nanoparticles on their point of zero charge and isoelectric point ..... 66

Figure 5.1 Effect of polymer concentration on chondrocytes viability on day-1 in 3D agarose scaffolds.

PSA concentrations are (a) 0.2% (b) 2% and (c) 10% w/v. Green color indicates cytoplasm of the live

cells and red color indicates dead cells and probably microparticles ......................................................... 70

Figure 5.2 Equilibrium Young’s modulus of hydrogel constructs with different concentrations of

poly(sebacic anhydride) microparticles. The y-axis values are normalized by modulus of the

corresponding control constructs (without microparticles). The concentration of polymeric microparticles in

x-axis is represented as w/v% of hydrogel constructs ................................................................................ 70

Figure 5.3 Effect of increased in buffer capacity (5mM HEPES) of growth medium on chondrocytes

viability on day-23 in 3D agarose scaffolds using Live/Dead cytotoxicity kit. The agarose constructs

contain (a) no PSA (control) (b) 2 % PSA microparticles. The microparticles are encapsulated with

chondroitinase-ABC. Green color indicates cytoplasm of the and red color indicates dead cells and

probably microparticles ............................................................................................................................... 72

ix

Figure 5.4 Cellular uptake of fluoresceinated-PSA (F-PSA) by chondrocytes. F-PSA is added nearby a

chondrocytes aggregate at time t=0 and images were obtained at time, t= a) 5 min b) 20 min. c) 60 min.

Gradual increase in the fluorescence intensity is due to intracellular accumulation of F-PSA ................... 73

Figure 5.5 Mechanical properties of control without BSA (BSA-), control with BSA (BSA+) and TGFβ3

microparticles-containing (ugel-TGF-β3-BSA+) constructs. (a) Equilibrium Young’s modulus, and dynamic

modulus at (b) 0.1 Hz. Using two way ANOVA analysis, there was no statistically significant difference

was found between Control BSA+ vs. ugel-TGF-β3-BSA+ for p< 0.05 ..................................................... 75

Figure 5.6 Mechanical properties of control, polyacrylamide containing (PAM) and polyacrylic acid-

Phosphonate (PAA-P) nanoparticles-containing constructs. (a) Equilibrium Young’s modulus, and (b)

dynamic modulus at 0.1 Hz. ........................................................................................................................ 78

Figure 5.7 Total GAG & DNA content (in mg) of control, polyacrylamide containing (PAM) and polyacrylic

acid-Phosphonate (PAA-P) containing scaffolds ........................................................................................ 79

Figure 5.8 Live/Dead assay of scaffolds using fluorescence microscopy, for control, polyacrylamide

containing (PAM) and polyacrylic acid-Phosphonate (PAA-P) containing scaffolds on day0 and day33... 79

Figure 6.1 Mechanical properties of control and nanoceria-containing constructs on day 16 and day 30.

Interleukin-1α was added from day 16 to 30. Nanoceria concentration in constructs are 10, 100 and 1000

μg/mL. Control constructs contain no nanoceria (a) Equilibrium Young’s modulus, (b) dynamic modulus at

0.1 Hz. The significance are *p<0.001 and **p<0.001 ................................................................................ 85

Figure 6.2 Biochemical composition of control and nanoceria-containing constructs on day 16 and day 30

as a fraction of wet weight. Interleukin-1α was added from day 16 to 30. Nanoceria concentration in

constructs are 10, 100 and 1000 μg/mL. (a) Glycasoaminoglycan (GAG) (b) Collagen (c) DNA .............. 85

Figure 6.3 Live/dead analysis of control and nanoceria-containing constructs on day 30 using cytotoxicity

kit and confocal fluorescence microscopy. Green color indicates live and red color indicates dead cells.

Top row contains images of control, nanoceria-containing constructs at 10, 100, 1000 μg/mL

concentration without interleukin-1α insult. Bottom row has the same sets with interleukin-1α insult from

day 16 to day 30. ......................................................................................................................................... 86

Figure 6.4 Mechanical and biochemical properties of constructs at day 34 as a function of nanoceria

concentration in the growth medium. (a) compressive Young’s modulus, (b) dynamic modulus at 0.1 Hz,

(c) glycosaminoglycan. Interleukin-1α is added from day 21 to 34 and its concentrations are given on the

top of the bars (0, 0.5, 10 ng/mL). For IL-1α 10ng/mL and 20000 μg/mL vs. 0 μg/mL *p<<0.00001 ......... 88

Figure 7.1 Mechanical properties of control and feriic (hydr)oxide NPs-containing constructs on day

30.The concentration of the NPs is plotted as total surface area. (a) Equilibrium Young’s modulus,

dynamic modulus at (b) 1 Hz (c) 0.5 Hz (d) 0.1 Hz. .................................................................................... 92

Figure 7.2 Live/dead analysis of control and ferric (hydr)oxide NPs-containing constructs on day 72 using

cytotoxicity kit and confocal fluorescence microscopy. Green color indicates live cells, red color indicates

dead cells. Top row contains images of control, FH, and H9-containing constructs at 25 μg/mL

concentration and bottom row that of H38 and H150 containing-constructs at two different NP

concentrations (25, 50 μg/mL) .................................................................................................................... 93

Figure 7.3 Suspensions of FH and hematite NPs at initial SL concentrations of 3 mM and ionic strength

of 5 mM NaCl. ............................................................................................................................................. 94

x

Figure 7.4 Adsorption isotherms of sodium laurate on FH (wine) and H150 (blue) NPs at pH ~ 7.2. ....... 95

Figure 7.5 Contact angles of SL-coated H150 nanoparticulate film. Formal monolayer coverage of 4.3

molecules/nm2 was estimated from the adsorption isotherm. Hydrophobicity of 150-nm hematite

increases as the surface coverage of laurate increases up to 1 ML. At this point, the particulate film is

superhydrophobic (water droplets are repelled from the surface). With a further increase of surface

coverage, hydrophobicity gradually decreased, indicating that laurate is adsorbed with the head group

pointing out into the solution. ...................................................................................................................... 96

Figure 7.6 Effect of the initial concentration of laurate (written at the bottom as well as on the vials) on the

dispersion of H150 particles at pH 7.2 ........................................................................................................ 96

Figure 7.7 (a) Adsorption isotherms of sodium laurate on FH (wine), H9 (orange), H38 (red) and H150

(blue) NPs at pH ~ 7.2. The last data point corresponds to 1 mM initial laurate concentration (CMC of

laurate) (b) Zeta potential of H150 and FH as a function initial laurate concentration in 5mM NaCl

aqueous solution. Both electrophoretic (red and blue curves) and electroacoustic (green curve) results

are shown. Connecting lines are drawn for reading ease. .......................................................................... 97

Figure 7.8 in situ FTIR-HATR spectra of FH and hematite NPs in contact with a 10–3

M solution of sodium

laurate at pH 10.00 ± 0.05 with no added electrolyte in spectral region of (a) COO and (b) CH

vibrational modes: (olive) H150, (red) H38, (blue) H9. Conditioning time is 2530 min. ........................... 98

Figure 7.9 in situ FTIR-HATR spectra of FH and hematite NPs in contact with a 10–3

M solution of sodium

laurate at pH 7.15 ± 0.05 with no added electrolyte in spectral region of (a) COO and (b) CH

vibrational modes: (olive) H150, (red) H38, (blue) H9, (magenta) FH. Conditioning time is 25 min.

Intensity of the spectra in the 12001800-cm–1

region is normalized by the intensity of the sCH2 band

(see text for more detail). ............................................................................................................................ 99

Figure 7.10 Correlation between frequency of sCH2 band of ISMM laurate in the spectra shown in

Figure 7.9a and PZC of FH and hematite NPs measured in open air (Figure 7.7b) ................................ 100

Figure 7.11 Effect of surface coverage on the position of the sCH2 band of laurate adsorbed on (a)

H150, (b) H38, (c) H9 hematite, and (d) FH from 1 mM solution at pH 10.0 and 7.15. Surface coverage,

evaluated using the sCH2 peak intensity, increases from (red) ~ 0.030.1 to (magenta) 1.0 formal ML

which is assumed to be formed at the maximum interaction time. It is increased with increasing the

conditioning time which is indicated in the graph. The band intensity is normalized to the maximal value in

the series. Color codes for H150, H38, and H9 are the same. FH was conditioned only at pH 7.15

because it does not adsorb 1 mM laurate at pH 10. ................................................................................. 103

Figure 7.12 Arrangement of surface cations on (001) basal, (101) hexagonal as well as on (012) and

(104) rhombohedral hematite surfaces. .................................................................................................... 104

Figure 7.13 Effect of the substrate on in situ FTIR-HATR spectra of H38 hematite NPs in contact with a

10–3

M solution of sodium laurate at pH 7.15 ± 0.05 with no added electrolyte in spectral region of (a)

COO- and (b) CH vibrational modes: NPs were deposited on a (green) Ge and (red) ZnSe IRE.

Conditioning time is 25 min. ...................................................................................................................... 105

Figure 7.14 Comparison of in situ FTIR-HATR spectra in the CH region of (red) H38 deposited on Ge

and (violet) H150 deposited on ZnSe. The NPs were conditioned similarly with a 10–3

M solution of

sodium laurate at pH 7.15 ± 0.05 with no added electrolyte for 2530 min. ............................................. 106

xi

Figure 8.1 A general predictive paradigm for identifying adsorption forms on metal (hydr)oxides using

density functional theory simulations and FTIR spectroscopy. ................................................................. 115

Figure 8.2 Ferric trioctahedral cluster extracted form hematite (104) surface. The cluster contains both

edge-shared and corner-shared oxygen ions. The surface terminal OH/H2O groups that are involved in

coordination with carboxylate group are shown in yellow spheres. (Iron – blue spheres; oxygen – large

red spheres; carbon – large gray sphere; hydrogen – small green spheres) ........................................... 116

Figure 8.3 in situ FTIR-HATR spectra in region of (a) CO2 and (b) scisCH2

vibrational modes of (1)

hematite reacted for 25 min with a 10–3

M solution of sodium laurate at pH 7.15 ± 0.05 with no added

electrolyte. For comparison, shown are spectra of (2) aqueous 10–2

M sodium laurate at pH 10.2

measured with a Ge IRE, (3) solid sodium laurate solid, and (4) solid lauric acid. Baseline in the

16001660-cm-1

region of the adsorbed laurate was corrected by adding spectrum of water (water is

removed from the interface upon the adsorption). This figure is taken from Ref. 190

. Spectral regions of the

C=O and scisCH2 modes of the surfactant are shadowed. Band assignment is taken from Ref.253

....... 118

Figure 8.4 in situ FTIR-HATR spectra of 150-nm hematite reacted successively with SL at (a) pH 7.15

and SL concentration of (black) 0.1 mM, (blue) 0.5 mM and (green) 1.0 mM and (b) pH 10.0 and SL

concentration of 1 mM. Time of the interaction is shown in the graphs. It was measured from the moment

of the introduction of a SL solution into the FTIR-ATR cell till the beginning of the spectrum scanning

(which takes 5 min). This figure is taken from Ref. 190

. The background spectrum was that of the

deposited particles pre-equilibrated with water of pH 7.15 for 1 hour. Spectra are shifted along the Y axis

for clarity. ................................................................................................................................................... 119

Figure 8.5 Geometry optimized DFT models of a butanoate ion attached as ISMM and OS hydration-

shared complexes to a ferric trioctahedral cluster with charge q = 0 which represents the (104) hematite

surface. Numbers are the CO and CC and H bond lengths in Å Dashed lines show the surface

coordinating H bonds. ............................................................................................................................... 120

Figure 8.6 The predictive methodology applied in this work for identifying adsorption forms of sodium

laurate on 150-nm hematite NPs using density functional theory simulations and FTIR spectroscopy ... 122

xii

List of Tables

Table 1.1 Characteristics of human articular cartilage using zonal representation4 .................................... 3

Table 4.1 Morphological characteristics of hematite NPs .......................................................................... 67

Table 6.1 Two way ANOVA for all the mechanical properties (equilibrium Young’s modulus, dynamic

modulus at 0.1, 0.5, and 1 Hz) on day 30 of control and nanoceria-containing constructs those were not

insulted with interleukin-1α. ......................................................................................................................... 86

Table 8.1 Calculated vibrational frequencies of IS and OS complexes of butanoate .............................. 121

xiii

List of Schemes

Scheme 1.1 In situ delivery of bioactive molecules using encapsulated polymeric microparticles

embedded in a hydrogel matrix. The surface eroding polymeric microparticles allows release the bioactive

molecules in a uniform and sustained manner. ............................................................................................ 5

Scheme 1.2 Hydrolysis of acetyl-endcapped poylanhydride polymer into polyacid-anhydride oligmers and

acetic acid. The red ellipsoid is a lipophilic backbone chain with or without anhydride interlinks ................ 9

Scheme 1.3 Cluster models of different laurate adsorption complexes. ISBB-Inner sphere bidentate

bridging; ISMM – Innersphere monodentate mononuclear; OS – Outersphere; Blue octahedra – Fe3+

; red

spheres – oxygen; gray spheres – carbon; green spheres – hydrogen. .................................................... 29

Scheme 3.1 Synthesis of polyanhydrides using melt condensation method ............................................. 41

Scheme 3.2 Synthesis of L-ascorbic acid functionalized PSA polymer ..................................................... 43

Scheme 3.3 Synthesis of microparticles using double emulsion and solvent evaporation method ........... 47

Scheme 6.1 Two different ways of introducing nanoceria in a 3D-agarose based culture of chondrocytes.

(a) nanoceria is embedded in 3D agarose constructs during casting on day-0 (b) nanoceria was added in

to the growth medium along with interleukin from day-21 .......................................................................... 83

Scheme 7.1 Cluster models of surface binding of fatty acid: (a) inner sphere monodentate mononuclear

(ISMM) carboxylate complex with H-bonded second oxygen (b) outer sphere (OS) surface-hydration

shared ion pair, and (c) OS fatty acid. Inner-sphere monodentate mononuclear complex: specific

adsorption of one of the oxygen of laurate molecule with one of surface ferric ion, whereas the second

oxygen H-bonds with other surface hydroxyl or water. Outer-sphere surface-hydration complex: non-

specific adsorption of laurate (or lauric acid) on surface ferric ions mediated by a hydration layer. .......... 98

Scheme 7.2 Stretching of the sCO2– dipole of adsorbed carboxylate by positive charge of the NP

surface. ...................................................................................................................................................... 100

Scheme 7.3 Deprotonation of lauric acid by basic surface sites upon adsorption. ................................. 107

Scheme 8.1 Cluster models of different laurate adsorption complexes. Blue octahedra – Fe3+; red

spheres – oxygen; gray spheres – carbon; green spheres – hydrogen. ................................................. 117

xiv

Acknowledgements

I would like to express my most sincere thanks and intense gratitude to my advisor Prof. Ponisseril

Somasundaran for his guidance, motivation and support. His extraordinary ability to motivate, energize

and train me to work hard helped me in solving different problems. His ‘Saturday sermons’ are very

inspiring and are valuable as advice and benefits from his decades of experience. There is never a dull

moment when Prof. Som is around and is always pleasant to work with. However careful I was, I did

become part of large repertoire of his jokes, which of course, I enjoyed thoroughly. He propelled and

inspired me to attack and tackle problems far from my expertise and continuously encouraged me

through the difficult times in the course of this work. I am highly grateful to Prof. Somasundaran, for

teaching and training in various concepts and skills, in particular where I was found wanting. I convey my

deepest thanks to him for giving me great freedom to pursue my ideas.

I would like to take this opportunity to convey my profound gratitude to Prof. Clark T. Hung for

agreeing to collaborate with me in the tissue engineering part of this work, guiding and motivating me on

this part. In particular, this work would not have been possible without his encouragement and

commitment. I would also like to thank him greatly for taking time to teach me and for access to his

laboratory facilities. I would also like to specially show my indebtedness to Dr. Grace O’Connell for her

constant guidance in tissue engineering fundamentals and experimental methods. I enjoyed greatly

working and communicating with her.

I express my deepest appreciation for Dr. Irina Chernyshova for her collaborating with me during

my graduate study and teaching me and educating me not only in the spectroscopic methods, but also in

accurate and reproducible experimentation, as well as writing scientific papers and proposals. She is a

constant source of inspiration and criticism, and I am greatly indebted to her for her crucial help in

achieving and contextualizing the spectroscopic interpretations presented in the work.

I would like to extend my special thanks to Dr. Raymond Farinato for his teaching, advice and

guidance which helped me immensely in choosing my research topics and overcoming the problems in

xv

particle dispersions and polymeric particle synthesis. I would also like to thank Dr. Sun-yi Huang whose

advice solved problems that I was facing in microemulsion systems. I express my sincere thanks to Dr.

Nagaraj for his constant guidance in critical thinking and for further providing me with avenues where I

could contribute. I would like to thank Prof. Paul Duby, Prof. Nick Turro, Drs. Pradip, Venkat Runkana and

K. Ananthapadmanabhan for their advice, discussions, and motivation throughout this work.

I would also like to take this opportunity to acknowledge the financial support provided by NSF

Center for Particulate and Surfactant Systems, Columbia University and Tata Research Design and

Development Center.

Not least of all, I wish to acknowledge and thank all my current and previous lab members for

constant willingness to help me. In particular, I appreciate contributions, experimental help, and very

useful comments of Drs. P Purhoit, X. Fang, B. Li, S. Lu, and J. Wu. I also would express my genuine

appreciation for all my friends, in particular Jonathan Levine and Karthik for their constant interest in my

progress, support, encouragement, and great time spent together. I deeply appreciate the administrative

personnel from Earth and Environmental Engineering for the sustained support, and Katherine Hickey for

proof reading.

Finally, I would like to convey my love and gratitude towards my parents, Ponnurangam

Muniswamy and Banumathi Renu, for without their love, support and encouragement none of my

endeavors would have been possible.

xvi

Dedicated to

Tamil people who lost their lives in the genocide in Srilanka

1

Chapter 1 Introduction, Literature review, and

Goals

Polymeric nano- and micro-particles have attracted attention as promising candidates for

biomedical applications, such as encapsulation and drug-delivery. It has been demonstrated that these

particles can be biocompatible, biodegradable, and, moreover, obtained from renewable resources. They

have already found applications in medical imaging and diagnostic devices, tissue engineering, and cell

therapeutics. Another important relevant area, with no general or simple solution, is the safe, uniform and

controlled delivery of drugs, nutrients and growth factors using polymeric particles. Metal (hydr)oxide NPs

are one of the main nanomaterial in the current and researched nanotechnologies.1 The societal concern

about the safety of these NPs for human health and the environment has recently triggered an additional

interest to the nano-biointerfaces.2

In general, structural, compositional, dimensional, and morphological properties of nano- and

microparticles determine their technological and biomedical performance, as well as cytotoxicity.

However, each particular application and each particular type of material require a specific inquiry.

This work is focused on the structure-(bio)activity relationship for four types of particlestwo

types of polymeric nano- and microparticles, and two types of metal (hydr)oxide nanoparticles. Here, the

bioactivity of polymeric particles is defined as their biocompatibility and cell-stimulating activity in articular

cartilage engineering, whereas for metal(hydr)oxide nanoparticles, it is defined as their antioxidant

properties and cytotoxcity, as well as the adsorption and dispersion properties.

1.1 Literature review and gaps in the current knowledge

1.1.1 Bio-activity of polysebacic anhydride microparticles used for in situ delivery in articular

cartilage engineering

This section outlines

2

(a) Challenges in articular cartilage engineering,

(b) Rationale for using polymeric microparticles for in situ delivery of bioactive molecules that address

these challenges,

(c) Different classes of polymeric biomaterials that are used for encapsulation and

(d) Rationale for the selection of polyanhydrides, physico-chemical and biocompatible and biodegradable

properties of polyanhydrides,

(a). Challenges in articular cartilage engineering

Articular cartilage is a connective tissue between the articular surfaces of joint bones which helps

reduce friction between the joints, absorbs shock, and enables the sliding movement of joints. The major

component of native human articular cartilage is water (~72% by weight). Other significant components

include collagens and glycosaminoglycans (GAGs) weighing ~66% and 18% of dry mass, respectively.3

There are four zones of cartilage (superficial, middle, deep, and calcified) and they differ biochemically as

well structurally from each other (See Figure 1.1 and Table 1.1).4 For example, collagen content is

observed to be maximum in the superficial zone and decreases as we go deeper, whereas aggrecan

content reaches a maximum in the deep zone and calcified zone has mineralized cartilage tissue.(Table

1.1) It is obvious that a multiphasic scaffold that can guide the chondrocytes to elaborate a extracellular

matrix mimicking the native cartilage zones is necessary.

Figure 1.1 A typical cross section of articular cartilage showing four zones4

3

Table 1.1 Characteristics of human articular cartilage using zonal representation4

The regeneration of damaged or diseased sites of articular cartilage either through medical

treatment or implantation of engineered tissue constructs seeded with chondrocytes has been particularly

challenging due to the avascular nature and poor self-repairing ability of the cartilage, stringent

requirements of biomechanical properties (compressive, tensile, and shear resistances) and associated

problems of integration with subchondral bone.5, 6

Because of this poor self-repair capacity of articular cartilage, any damage due to injury or

disease can lead to an inflammatory arthritic condition which results in chronic pain and dysfunction in the

joints of afflicted patients. One treatment method introduced to relieve the pain and swelling is the

implantation of autologous chondrocytes around the lesions.7 However, this method is temporary and has

several drawbacks including lack of full load-bearing capacity, generation of inferior quality fibrocartilage

compared to articular cartilage, and the need for repeated surgeries have been reported.8, 9

Zones Structure Composition Functions

Superficial Zone

Chondrocytes are flat and parallel to the surface

Highly polarized and self-assembled parallel collagen network

Smallest collagen fibrils diameter (20 nm)

Collagen: Type II

Proteoglycan: decorin (leucine-rich), biglycan

Lowest aggrecans

Surface zone of the cartilage in contact with synovial fluid

Secretes lubricin: essential for lubrication

Highest tensile properties

Middle zone

Cell density lower

Rounded cells

Collagen fibrils larger in diameter and arranged randomly

Proteoglycan: aggrecans in medium concentration

Extensive ECM

Compressive modulus higher than superficial zone

Deep zone Cell density lowest

Collagen fibrils largest in diameter (120 nm)

Proteoglycan: Maximum aggrecans concentration

Collagen content lowest

Highest compressive modulus

Calcified zone

Hypertrophic phenotype Collagen: Type X

Intermediate mechanical properties between uncalicified deep zone and subchondral bone

4

Several novel ideas and growth factor schemes are being researched and tested to address the

shortcoming of this autologous chondrocyte implantation method.6, 10

Namely, 1) the seeding of

chondrocytes into a resorbable scaffold and culture ex vivo so as to partially elaborate the tissue before

implanting the constructs in vivo, 2) use of autologous stem cells or allogenic sources of chondrocytes 3)

growth factors and their schemes of addition 4) bio-functional scaffolds with cell-adhesive and chondro-

inductive properties 4) mechano-induction of chondrocytes through external dynamic load schemes 5)

bioreactors design 6) in situ delivery of nutrients and growth factors.

Several hundred biomaterials, both natural and synthetic varieties, have been tested as scaffolds

for seeding cells. Of these materials, the class of materials that form hydrogel matrix possess distinct

advantages due to their highly interconnected porous structure (better suited for sustained transport of

nutrients and growth factors) and large volume fraction of water (>98) which provides space for

accumulation of the extra-cellular matrix elaborated by chondrocytes.

Of the natural materials, agarose, a polysaccharide that forms hydrogel is a very popular scaffold

material, as it is found to maintain the phenotype of chondrocytes as well as to possess beneficial

mechanical properties.11-14

The advantage of using agarose scaffolds has also been shown by the

engineering of the cartilage tissue that has the Young’s modulus similar to that of native tissues11, 15

and

has already been developed into a practical technology named Cartipatch, autologous chondrocyte

implantation therapy.16

Some of the challenges in the regeneration of articular cartilage include,

1) Mimicking native cartilage tissues in structure and functionality. In particular, Type II

collagen content and the dynamic modulus of regenerated tissue is lower than the

native tissue values

2) Overcoming the detrimental spatial gradient in the properties of engineered tissue due

to poor transport & diffusion of nutrients and growth factors

3) Minimizing the inflammatory response under both in vivo and in vitro conditions

4) Achieving smooth integration of regenerated cartilage with original cartilage as well as

subchondral bone

5

5) Minimizing cell and extracellular matrix leakages during culture period

6) Cuing in the chondrogenic differentiation of stem cells

7) Effect of the composition and functional groups of the scaffold on tissue generation is

only beginning to be understood

(b). Rationale for using polymeric microparticles for in situ delivery of bioactive molecules

One of the critical challenges listed above is to overcome the decrease in the nutrients diffusivity

over a period of time in a homogenous scaffold. Diffusivity decreases due to the barriers created by the

progressive matrix elaboration. The resulting scaffold has spatially-inhomogeneous mechanical

properties.17

In particular, the internal regions of the scaffold exhibit poorer mechanical properties. The

creation of macrochannels has been suggested as a solution to this diffusion limitation.17

An alternative

solution is to embed the nutrient-encapsulated polymeric micro- or nano-particles which can release the

nutrients in a desired timeframe through controlled degradation rates (Scheme 1.1).

Scheme 1.1 In situ delivery of bioactive molecules using encapsulated polymeric microparticles

embedded in a hydrogel matrix. The surface eroding polymeric microparticles allows release the bioactive

molecules in a uniform and sustained manner.

Encapsulated polymeric nano- and micro-spheres have been extensively researched for

controlled release drugs and bio-active molecules. These polymeric carriers have several advantages

including the ability to vary their molecular weight and thereby control the release rate, degradation rate,

6

and hydrophilicity. Additionally it is known that supplementation of growth factors such as TGF-β3, has to

be transient (only for initial 2 weeks) at least for in vitro culture of juvenile bovine chondrocytes.18

This

necessitates that the delivery device and the selected polymer should have such erosion and degradation

rate such that release of most of the bioactive molecules should happen within this time frame.

(c). Comparison of biomedical properties of natural and synthetic polymeric biomaterials

The biocompatible and functionalized polymeric materials are being investigated for several

applications as drug, genes or vaccine delivery agents, scaffolds as well as guides for tissue growth in

regenerative medicine, as temporary prosthesis to name few. 19

20

In particular, use of polymeric

biomaterials (both natural and synthetic varieties) in the regeneration of damaged or diseased tissue is

fast emerging as an attractive alternative over artificial prosthetic implants due to the long-term

biocompatibility and prospects of permanent cure without repeated surgeries. However, there are several

challenging requirements for a biomaterial before it can be applied in vivo,21

1) Biocompatibility: Not only the initial material but also its degradation products should be

biocompatible. The long-term biocompatibility of the polymer and degradation products needs to

be confirmed before applying in vivo. The biocompatibility is determined not only by its cellular

toxicity but also by the elicited inflammatory response, genotoxicity, modification of cellular

morphology and biochemical processes, as well as impact on the secretion of extracellular

matrices.

2) Biodegradability: Unlike artificial prosthetic implants, the biodegradable polymers offer the scope

for restoring the regenerated tissue to its native state through the degradation and resorption or

elimination of the polymers in the body. However, their rate of degradation should be matched

with that of growing tissue such that overall biofunctionality (eg. mechanical properties) of the

implant and the tissue is not lost during the course of healing. Moreover, both the initial and

degraded components of the biopolymers should be either metabolized or eliminated quickly in

the body.

7

3) The other properties of polymers that are important include, porosity, gellability, crystallinity, glass

transition temperature, hydrophilicity/hydrophobicity, bulk or surface erosion properties, pH –

response, surfactancy, simpler synthesis, purification and sterilization processes, ease of

functioanlization and shelf life. The choice of these properties that are important is dictated by the

particular needs of a given biomedical application.

The biocompatible and biodegradable polymers can be classified based on the source as natural or

synthetic polymers. Some examples of polymers derived from natural materials include collagen, gelatin,

agarose, alginate polysaccharides and derivatives thereof and examples of synthetic polymers include

anhydrides, (ortho-)esters, urethanes, amides, carbonates and co-polymers in various combinations. The

(dis-) advantages of both natural and synthetic polymers in biomedical applications are shown in Table

1.2.

Table 1.2 A comparison of properties of natural and synthetic polymers relevant to biomedical

applications

Property Natural polymers Synthetic polymers

Biocompatiblity Typically good Needs extensive testing

Biomimetic activity and presence of receptor ligands

Good Typically not biomimetic and requires conjugation of biological motifs for targeted bio-functionality

Degradation Proteolytic degradation and typically elicits benign cellular response

Hydrolytic and requires careful studies to understand cellular response

Immunogenicity & disease transmission

High risk and difficult to control Low risk

Batch-batch variations High low

Functionalization with biological motifs

Difficult Relatively easy

Tunable properties Molecular weight Bad Good Hydrophobicity Bad Good Gellability Good Good Porosity Good Bad Degradation rate Bad Good Ease of processing Bad Good

8

(d). Sebacic acid based polyanhydride polymer for delivery device in articular cartilage

engineering: Advantages and challenges

The important criteria in selecting the encapsulating polymer for tissue engineering applications

are the faster degradation rate and surface erosion characteristics. Specifically, in articular cartilage

engineering, the in situ release of bioactive molecules from the microparticles is preferred to be within 2-3

weeks for maximum benefit. Thus polymer degradation rate should be necessarily in the faster range.

The synthetic polymers were selected for this study because of the ability to tailor the degradation rate

and further scope for surface modification with biological motifs. A survey of literature of different classes

of synthetic polymer indicates the following typical decreasing order of hydrolytic degradable bonds

anhydrides> orthoesters > esters > amides. The polyanhydrides are a class of biodegradable and

biocompatible polymers that used for a uniform, sustained and short-term delivery of bioactive agents.

They typically have a hydrophobic backbone chains linked to each other by easily cleavable anhydride

bonds. The polyanhydrides were first synthesized in the beginning of 20th century and at present, several

dozens of new polyanhydrides have been synthesized and researched for various biomedical

applications. Apart from the claim that they are non-mutagenic and non-cytotoxic, they have several

advantages, including surface erosion properties and tunable degradation rates suitable for short-term

release. The high hydrophobicity of the polyanhydride-based microparticles excludes the contact

bioactive encapsulation from water or water-based oxidants thus protecting their bioactivity until release

through surface erosion. Thus, polyanhydrides have found applications in delivery of vaccines, adjuvants,

and proteins where the immunogenicity of these actives is preserved. In particular, polyanhydride-based

wafers have already obtained FDA clearance for delivery of drugs in the treatment of brain tumor.22

One of the most common component of anhydride (co-) polymer used for these purposes sebacic acid

with either acetyl or methacryl end groups. Several parameters control the degradation rate of

polyanhydrides, namely, molecular weight, hydrophobicity, pH, crystallinity, porosity, and surface area of

the polymeric device whereas, the thickness of the polymer layer determines the total time for

degradation.23-29

In particular, hydrophobicity of polyanhydrides has significant influence on the

degradation rate.25

By of incorporation of aromatic group (poly(carboxyphenoxy propane-co-sebacic

9

anhydride), P(CPP-SA)) in the anhydride copolymer, degradation rate was observed to decrease by

several orders of magnitude when compared simple poly(sebacic anhydride) (P(SA or PSA).

Interestingly, the in vivo degradation of polyanhydrides is found to be slower than the in vitro rate

due to decrease in water content in the former.30, 31

The polyanhydrides and their degradation products

are considered to be safe, non-cytotoxic,32-39

non-mutagenic40, 41

and non-carcinogenic39

under both in

vitro and in vivo conditions. However, MRI imaging of P(SA) & P(CPP-SA) tablets implanted in the neck of

Wister rats, showed muscle edema at around Day 7 but was resolved eventually.30, 31

Unlike aromatic

polyanhydrides, aliphatic ones, such as P(SA) are shown to be extensively metabolized by rat brain cells

using radio-labeling technique. In particular, 80% of the SA was metabolized and eliminated as CO2.42, 43

It was further speculated that SA-based degradation product can participate in β-oxidation metabolic

pathway producing acetyl-CoA, which is an important intermediate molecule.44

Other cells such as aorta

epithelial osteoblast-like, and smooth-muscle cells also did not show cytotoxic effects due to

polyanhydrides.39, 45

However, in the case of SA dominated copolymer, lack of cell adhesion was

observed at latter times, probably due to faster degradation rate this component.

Scheme 1.2 Hydrolysis of acetyl-endcapped poylanhydride polymer into polyacid-anhydride oligmers and

acetic acid. The red ellipsoid is a lipophilic backbone chain with or without anhydride interlinks

H2

O

10

In contrast, other studies,46

show that maximum tolerable concentration of polyanhydride (P(CPP-

SA)) is as low as 2.8 mg/mL, above which cytotoxic effects are observed. However, no mechanism for the

cytotoxic effects at higher concentration was elucidated. Moreover, there are no reported mitigation

routes for the cytotoxicity of this polymer.

Generally, several pathways can be hypothesized for the inflammatory, cytotoxic, genotoxic or altered

biochemical response of cells due to the presence and uptake of biomaterials above certain limits. These

include variation in local pH and osmolality at biomaterial-chondrocyte interfaces, cytoplasmic membrane

disruption, cellular uptake of the biomaterials and corresponding changes in the intracellular pH, Ca2+

levels, lipotoxicity, and induction of reactive oxygen or nitrogen species.47-49

It is also known that a

decrease or increase in pH from physiological value of 7.2 significantly decreases matrix synthesis. The

hydrolytic degradation of polyanhydrides in to fatty acids and corresponding uptake by cells can decrease

the intracellular pH. It has been noted that an increase in extracellular pH and corresponding increase in

the intracellular pH can cause increase Ca2+

, which in turn can initiate the signaling cascades for cellular

apoptosis. Similarly, it has been shown that changes in osmolality away from native values of 350-400

mOsm/kg can significantly alter the morphology of chondrocytes and hence their metabolism. The

potential cytotoxicity due to polyanhydride can also be caused by their intracellular lipotoxicity due to their

possible cellular uptake.49

In summary, the polymeric microparticles can be used for in situ delivery of growth factors and

nutrients to overcome the transport limitations of elaborated matrix. There are no studies that address the

decrease in nutrient transport using short-term in situ delivery in articular cartilage engineering and show

an improvement the tissue properties.

The polysebacic anhydride with surface eroding property and highly tunable degradation rate as

well as hydrophobicity was selected for this work to encapsulate bioactive materials such as TGF-β3 and

vitamin-C. The poly(sebacic anhydride), undergoes hydrolysis to form poly(sebacic acid) and both the

polymer and its degradation products are reported to be biocompatible. There are few investigations

using of copolymeric anhydrides containing smaller quantities of sebacic acid in tissue engineering, 50, 51

.

However, none of them have studied encapsulation and delivery of bioactives such as TGF-β3, vitamin C

11

or chondroitinase-ABC in cartilage regeneration within a time frame of delivery of 2-3 weeks. The

unaddressed challenge in delivering appropriate quantities of these bioactives in situ is the weight

fraction of polymeric microparticles needed for this purpose calls their established biocompatibility in to

question. A rough estimate for a sufficient in situ release of bioactive molecules such as TGF-β3 or

vitamin C for initial 2-3 weeks of culture indicates a 50-200 mg/mL of PSA-M in the casting melt.

However, there is no data in the literature that points towards biocompatibility of such high doses of PSA-

M with chondrocytes. As a consequence there are no studies exploring the potential mechanisms of

cytotoxicity of polyanhydrides. Furthermore, mitigation strategies for such toxicity are non-existent.

Without answering these questions, the utility of polyanhydrides (in particular polysebacic anhydride) as

delivery device in articular cartilage tissue engineering will be very limited.

1.1.2. Bioactivity of polyacrylate nanoparticles in articular cartilage tissue engineering:

Effects of functional groups

Many recent advances in the design of multi-functional biomaterials for tissue engineering are

geared towards the objective of manipulating the cell-bioscaffold interactions to control and direct the

tissue morphogenesis.52-58

In this section, functionalization of polymeric biomaterials with bioactive motifs,

their biomedical applications, and their relevance in articular cartilage engineering are reviewed.

Two important design features that can enable cell-scaffold interactions are, (i) functionalization

of scaffolds and (ii) engineering of nanoscale roughness and structures of the scaffolds. Polymeric

nanoparticles if fabricated with these design features have high potential to be used as a nanocomposite

material in bio-scaffolds. There are several examples in literature where functionalization of the substrate

has been shown to have markedly influenced the extracellular expression of cells as well as the

phenotypic expression of stem cells. For example, types I and X collagens, as well as aggrecan, are

found to be constitutively expressed by human MSCs on polystyrene dishes,59

whereas type X collagen

expression is inhibited by the presence of Nylon-6 and polypropylene.59

Polyethylene glycol (PEG)

12

tethered with decorin results in 10 times more Collagen type II and 2.5 times more glycosaminoglycan

(GAG) content. An increased collagen type II and GAG content is desired for better mechanical

properties. The KLER sequence, a binding site from decorin protein, is known to bind strongly to collagen

type II and is responsible for matrix organization, while RGD promotes the general survival of cells.60

A

poly(lactic-co-glycolic acid) (PLGA) scaffold, grafted with TGF-β and loaded with dexamethasone, has

showed enhanced type II collagen production.61

Crosslinked poly(ethylene) glycol-fumarate (PEGF)

copolymer has been functionalized with Gly-Arg-Gly-Asp (GRGD) peptide to enhance attachments of

cells.62

The interesting effects of functional groups of the substrate on the phenotypic expression of

human mesenchymal stem cells is observed in a study where, PEG functionalized with carboxylic,

phosphate and t-butyl groups yield cartilage-, bone-, and adipose-like tissues, respectively.63

This study

shows the importance of the functional groups in maintaining the phenotype of stem cells and also the

chemical constituents of their extracellular expression.

However, most studies have not addressed the effects of functional groups on their

biocompatibility with cells such as chondrocytes. The effects that arise due to the presence of the

functional groups, such as changes in the local pH and changes in the overall charge state of the

extracellular space and their impact on biocompatibility are not well studied. In particular, cartilage tissue

is already highly anionic due to the presence of sulfate-containing proteoglycans in extracellular matrix.

The additional presence of anionically functionalized polymer can alter not only the hydration state of the

scaffolds but also the local pH and thereby affecting their overall functions. The scaffold which is part of

the extracellular space during the tissue elaboration can also alter the cellular behavior by influencing the

inter-cellular signaling processes.64

In particular, 3D scaffold in conjunction with secreted matrix provides

not only a substrate for cellular attachment but can also potentially act as a repository for morphogenetic

protein storage and transport. It is known that extracellular matrix components such as proteoglycan and

collagen possess motifs that can bind to the morphogenetic proteins leading to their accumulation in the

interstitial space. This allows a crucial mechanism by which the cells communicate with each other on the

status of extracellular environment thereby eliciting appropriate responses. Only recently there have been

some attempts to engineer these functionalities onto the biomaterials that are used as scaffolds. 64

13

To summarize, functional groups present in scaffold are known to play an important role in

directing the cells into synthesizing matrix with different biochemical and biophysical characteristics. In

particular, the cell-scaffold interaction is considered crucial for tissue morphogenesis. To prevent articular

cartilage implant failure or rejection, it is important that the engineered tissue mimics the biochemical and

structural variations seen in the different zones of the native cartilage (Figure 1.1). There are several

studies directed towards this objective. However, most of these studies focus on the functionalized

scaffold-cell interaction, whereas, there are no studies which instead uses functionalized polymeric

nanoparticles embedded in an inert hydrogel scaffold to achieve similar results. This preserves the

advantages the inert scaffold (present in larger fraction) and functionalized polymeric nanoparticles

(present in smaller fraction) and allows a greater flexibility in the design of multiphasic scaffolds. Though

acrylate-based scaffolds have been well studied, the effects of various functional groups, such as amides,

phosphonates and sulfonates on the cellular viability as well as cellular metabolism have not been

explored.

The functional groups and the chemical composition of these scaffolds or nanoparticulate

systems themselves can cause (1) changes in the local pH (2) changes overall charge state of the

extracellular space (3) interfere in the intercellular signaling. The presence of additional anionically

functionalized polymer can create disequilibrium in the ionic state of extracellular matrix and thereby

affect the overall functions of chondrocytes. Thus, it is possible that presence of nanoparticles with high

anionic charge can potentially mimic biofunctions of sulfated proteoglycans and thus can either be

beneficial or harmful when present in the interstitial space. There are no studies which consider the effect

of surface charge of functionalized polymeric nanoparticles in clinically relevant hydrogel culture of

chondrocytes.

1.1.3 (Bio-)chemical activity of metal (hydr)oxide nanoparticles

Development of the nanomedical, bioanalytical and theranostic applications of metal oxide NPs

calls for a deeper understanding of the chemical, redox and dispersion properties of such NPs, as well as

their interactions with live cells. This knowledge is also required to assess the safety of the growing

amount of nanomaterials exposed to the human beings and environment, as well as to develop predictive

14

models of the fate and transport of the released nanomaterials in the environment. Finally, this knowledge

is needed to find out ways to mitigate the adverse effects induced by the NPs.

Cerium dioxide (CeO2) and ferric (hydr)oxide NPs are of special interest in this context. Ceria

nanoparticles (nanoceria) are an important technological nanomaterial which has found applications in

catalysis, sensing, and solid oxide fuel cells (SOFCs). In particular, nanoceria is a well-known component

of the modern three-way exhaust-gas catalyst, where it acts as an oxygen reservoir, releasing oxygen

under fuel-rich conditions, and absorbing oxygen under deficient conditions. In the nano-biomedicine,

CeO2 NPs has been demonstrated as a regenerating anti-inflammatory agent, scavenger of reactive-

oxygen & reactive nitrogen species and as anti-oxidants. This allows nanoceria to act as in situ redox

buffers of the oxidative stresses exerted on live cells by unfavorable conditions.

The interest in the adsorption and dispersion properties, as well as to the cytotoxicity of ferric

(hydr)oxide NPs stems from their application as pigments and MRI contrast agents, as well as from their

abundance in the environment, which significantly increases the frequency of the interaction of these NPs

with living organisms.

The following literature review presents the current paradigms of the cytotoxicity and biocompatibility of

metal oxide nanoparticles

Current paradigm of the cytotoxicity and biocompatibility of metal oxide nanoparticles

It is known that the presence of (nano)particles can exert oxidative stress on live cells. A way to

mitigate this stress is by intervention with antioxidants such as vitamins C and E or -carotene. These

substances, when present in low concentrations compared with those of an oxidizable substrate, can

significantly delay or prevent the substrate oxidation. The term ‘oxidizable substrate’ includes every type

of molecule found in vivo.65, 66

Reactive oxygen species (ROS) is a common name for oxygen-based radicals and

hydrogen peroxide (H2O2). Main ROS are superoxide anion (O2), hydroxyl radical (OH),

singlet oxygen (1O2), and hydroperoxyl HO2. Hydrogen peroxide, even though not a free radical,

is classified as a ROS because it acts as a conduit to transmit free radicals induced damage

15

across cell compartments and between cells.67 All the ROS are formed by partial reduction of

dioxygen O2. Due to the unpaired electrons, ROS are strong and nonspecific oxidants. Human

cells produce ROS as byproducts of their aerobic metabolic activity, in the course of normal

growth factor signaling, and specific enzymatic reactions.67, 68 Specifically, O2 is the final

electron acceptor for cytochrome-c oxidase, the terminal enzymatic component of a

mitochondrial enzymatic complex that catalyzes the four-electron reduction of O2 to H2O.

Superoxide anion and hydrogen peroxide are formed by one- and two-electron reductions of O2

during these (and other) electron transfer reactions, respectively. Although hydroxyl radical

formation can occur in several ways, by far the most important mechanism in vivo is likely to be

the transition metal catalyzed decomposition of superoxide and hydrogen peroxide.69

In the normal physiological state, cells protect themselves from the elevated concentrations of

ROS using antioxidant enzymes such as superoxide dismutase (which reduces O2

to H2O2), catalase,

and glutathione peroxidase (which reduces H2O2 to H2O) (Figure 1.2). However, due to endogenous or

environmental factors, the ROS–antioxidant balance can be broken in favor of the former, imparting

oxidative stress on the cells and eventually leading to the cell damage. In fact, ROS can react rapidly with

most biological molecules and damage lipids, protein, and DNA (Figure 1.3).68

In general, nanoparticles can exert oxidative stress on live cells by several mechanisms, including

catalytic redox activity, solubility, cellular uptake, damage to cell wall, hydrophilicity, etc. 2 In the case of

metal oxide nanoparticles, the main reasons have recently been identified as the catalytic redox activity

and solubility leading to the damage of biomolecules, disruption of the signaling pathways, and leaching

out of toxic cations.2, 70

In particular, more catalytic active Mn2O3, CoO, Cr2O3 are more cytotoxic than

less catalytically active TiO2 and Fe2O3. At the same time, the catalytic properties of some metal oxide

NPs such as CeO2 and Y2O3 make them biocompatible and even effective in scavenging oxygen free

radicals.71-73

Though there is an emerging speculative paradigm70

addressing the causes for such varied

biocompatibility among catalytically active metal oxide, it is still not well understood.

16

1.1.4 (Bio-)activity of engineered ceria oxide nanoparticles: Anti-inflammatory and reactive-

oxygen species scavenging properties

The goal of this study is to explore the potential of nanoceria as an anti-inflammatory agent for

treating damaged or diseased articular cartilage. Accordingly, this section is organized as follows: (a) the

structural, electronic and abiotic redox properties of CeO2 nanoparticles surveyed. (b) the rationale for

selection of nanoceria to combat inflammation in articular cartilage

(a). Structural, electronic, and abiotic redox properties of CeO2 nanoparticles

Cerium is the most abundant of the lanthanides, and forms compounds with two oxidation states,

+3 and +4, and hence is redox active. Cerium dioxide, CeO2, crystalizes in a lattice with the fluorite-like

(CaF2) structure (space group Fm3m). This structure (Figure 1.2) can be viewed as a face-centered cubic

(fcc) array of Ce (white) ions with O (red) ions residing in the tetrahedral holes.

Figure 1.2 The fcc cell of CeO2 with the fluorite structure. Ce is white, O is red.

In the reduced form, cerium oxide lattice has oxygen vacancies or defects which are

associated with the loss of oxygen in the reaction:

CeO2 CeO2-δ + δ/2O2. (1)

17

The deviation from stoichiometry can be as large as 0.5.74

Each released O atom leaves behind an

oxygen vacancy and creates two Ce3+

(Ce 4f1) cations by transferring electrons to two Ce

4+ cations (Ce

4f0). This property determines the oxygen storage capacity of ceria. The nonstoichiometric oxide CeO2-δ is

readily reoxidised back to CeO2 absorbing oxygen. In nanoceria, the reducibility, oxygen storage capacity,

and high ionic conductivity are remarkably enhanced. The redox properties of ceria arise from the

reversibility of the Ce3+

/Ce4+

redox couple with consequent formation/annihilation of surface defects

(oxygen vacancies).

An important property of ceria is its ionic conductivity: The defective fluorite structure allowa

oxygen conductivity of ceria through the hopping of oxygen ions to the vacant sites. The pure CeO2 is

paramagnetic, but acquires ferromagnetism upon partial reduction of Ce ions from 4+ to 3+ state. The

effect is likely to have close relationship with oxygen vacancies.75, 76

The nature of bonding in CeO2 and

Ce2O3 may be described as polarized ionic.77

The origin of the nanosize effect on the chemical and physical properties of ceria is under debate.

The effect can be due to the size-induced partial removal of oxygen atoms in the surface layer which is

paralleled by an increase of Ce3+

defects (bigger radius compared to Ce4+

) and leads to the lattice

expansion.78

Given that annealing at high temperature suppresses significantly the oxygen storage

property and catalytic activity of ceria, it is suggested that the interstitial oxygen ions associated with Ce3+

are the "active" species that provide necessary oxygen mobility crucial for the functioning of ceria as a

catalyst.74

In contrast, it has been shown79

that under ambient conditions, the oxidation state of ceria NPs

is always close to 4+, and there is no stabilization of Ce3+

in smaller ceria nanoparticles but rather

stronger structural disorder in nanoceria compared to bulk CeO2.79

As stated above, the redox catalytic activity of nanoceria that arises from the presence of both

Ce4+

and Ce3+

state. This redox activity forms the basis for their superoxide dismutase-like (SOD) and

catalase-like enzyme-mimetic scavenging ability (Figure 1.3). 72, 80

It has been proposed that SOD- and

catalase-mimetic activities arise due to oxidation of Ce3+

and reduction of Ce4+

states by radicals

18

respectively.72

Though the mechanism of auto-regeneration of nanoceria oxidation state is not clear, a

simple suspension in water81

or a PEG-coating82

were observed to restore the Ce3+

oxidation state.

Figure 1.3 A model of the reaction mechanism for the oxidation of hydrogen peroxide by nanoceria and

the regeneration via reduction by superoxide. An oxygen vacancy site on the nanoceria surface (1)

presents a 2Ce4+

binding site for H2O2 (2), after the release of protons and two-electron transfer to the two

cerium ions (3) oxygen is released from the now fully reduced oxygen vacancy site (4). Subsequently

superoxide can bind to this site (5) and after the transfer of a single electron from one Ce3+

, and uptake of

two protons from the solution, H2O2 is formed (6) and can be released. After repeating this reaction with a

second superoxide molecule (7) the oxygen vacancy site returns to the initial 2Ce4+

state (1). It is also

possible that the third Ce3+

indicated, which gives rise to the oxygen vacancy, could participate directly in

the reaction mechanism. The square Ce–O matrix is shown here only to illustrate the model and does not

correspond to the actual spatial arrangement of the atoms in the crystal structure. Reproduced from

Ref.72

(b). Rationale for selection of nanoceria to combat inflammation in articular cartilage

Some of the major problems encountered in cell-based tissue engineering repair strategies for

injured articular cartilage are due to the adverse inflammatory response to the grafts leading to its

19

rejection or reduced functionality and poor integration with host tissues.10, 83-85

Several strategies have

been tested to overcome these problems. For instance, it was shown that the constructs grown in vitro to

a partially or fully matured state and then implanted in vivo can withstand harsh catabolic environment of

the diseased region better than a freshly seeded or under-developed constructs.86, 87

Other strategies

include application of molecules combating inflammation such as dexamethasone and exogenous

crosslinkers.87-89

Alternatively, an in situ approach for combating inflammation that is clinically attractive through

incorporation of auto-regenerative biocompatible substances in to the implants capable of scavenging

reactive species. Such an in vivo approach has the benefits of speedy recovery and reduced in-patient

time since it allows in vivo maturation of grafts.

The cerium oxide nanoparticles has the potential for such an in vivo role since it has been shown

in vitro and in vivo to scavenge reactive oxygen and nitrogen species, combat inflammation, down-

regulate cytokine levels and confer cell-protection. The inhibition of nitrated proteins is particularly

significant since they are found to be one of the important causes for the rejection of grafts.90, 91

The

nanoceria has already been shown to protect several mammalian cell types including neural,92

retinal,93

hepatic,94

epithelial,73

cardiac,92

breast,95

and epidermal cells.96

Interestingly, viability and invasive

capacity of tumor cells are shown to decrease in the presence of ceria nanoparticles at concentrations

that are non-toxic to normal cells.97-100

It is suggested that due to the Warburg effect, tumor cells have

acidic pH which results in inhibition of catalase-like activity of nanoceria and accumulation of H2O2. This

accumulation selectively kills the tumor cells but not normal cells because at physiological pH both SOD-

and catalase-like activities of nanoceria are preserved. The effects of nanoceria in tissue engineering

have been reported recently101

where it was shown that nanoceria had improved both the osteoblastic

differentiation of h-MSC and collagen production.

In summary, the presence of nanoceria during the in vitro culture of chondrocytes can not only

enhance the collagen production but can also protect the elaborated extracellular matrix from degradation

20

due to exogenously added inflammatory chemical, interleukin-1α (IL-1α). However, no study has looked

at exploring the potential of nanoceria in articular cartilage engineering or in treatment of arthritic-like

conditions.

1.1.5 (Bio-)activity of ferric (hydr)oxide nanoparticles: adsorption, dispersion and

biocompatible properties

The goal of this study is to understand the size, morphology and surface dependence of

adsorption and dispersion properties of ferric (hydr)oxide NPs as well as their biocompatibility.

Accordingly, this section discusses the following, (a) the structural, electronic and abiotic redox properties

of hematite nanoparticles (b) the size, shape and morphology dependence of (bio-)activity of ferric

(hydr)oxide NPs

(a) Structural, electronic, and abiotic redox properties of hematite nanoparticles

Hematite (-Fe2O3) NPs have been used since ancient times as pigments and more recently

have found application as (photo)catalysts, sorbents, and sensors due to their special physicochemical

and electronic properties, low processing cost, high resistance to corrosion, and the environmental

compatibility.102-106

Hematite-like clusters are found in living organisms (as a core of ferritin) and, along

with other iron (hydr)oxides, are common in the environment (oceans, ground and surface waters, dust,

soils), where they control transport of elements and present a nutrient source for bacteria and

phytoplankton. 106

Hematite (-Fe2O3) adopts the corundum-type structure which has hexagonal symmetry with R ̅c

space group (Figure 1.4). This structure can be viewed as a hexagonal close-packed lattice of oxygen

atoms with two-thirds of the interstitial sites occupied by Fe atoms (Figure 1.4). The oxygen atoms occur

in layers along the z axis, three atoms per layer, within the hexagonal cell. Between each of these layers

there are two Fe atoms in a noncoplanar arrangement.

21

Figure 1.4 Crystal structure of α-Fe2O3. Red atoms are O; Blue atoms are Fe.

Hematite is a charge-transfer insulator (energy gap of ~2 eV), which exhibits interesting magnetic

properties. Below the Neel temperature TN = 953K, the stable phase is antiferromagnetic (space group

R3), with alternating (001) layers of spin-up and spin-down iron atoms in the hexagonal cell (Figure 1.4).

However, due to a slight canting of the spins, hematite is weakly ferromagnetic at room temperature.106

Hematite is the most stable ferric (hydr)oxide polymorph at bulk sizes and oxidizing conditions.

However, on the nanoscale, it becomes thermodynamically unstable relatively to several less stable

polymorphs,52,53

in common with corundum, rutile, zirconia, and other oxides. As the end-member of the

nanosize-induced phase transformation under hydrous conditions, goethite (-FeOOH) has been

suggested based on the XPS spectra.54

In contrast, vibrational spectroscopy data suggest that the end-

member is ferrihydrite (Fe10O14(OH)2+~H2O in the more ordered form107, 108

).52

The nanosize-induced structural changes of hematite are associated with changes in the

electronic properties such as an increase in ionicity and a decrease in electron affinity (opening of the

band gap).109

These electronic changes have been shown to be the driving force for the degradation of

the oxidative properties, as well as for the enhancement of the reductive dissolution of this mineral with

decreasing NP size.110, 111

The changes in the reactivity can be related to a decrease of the toxic effects

towards bacteria with decreasing NP size.112

Accordingly, 50-nm hematite NPs provoke an inflammatory

response in endothelial cells, 113

while smaller and less catalytically active114

30-nm NPs do not.115

22

The redox properties of hematite NPs in aqueous media are described by the electrochemical

mechanism.116

Figure 1.5 Conceptual scheme of electrochemical mechanism of redox reaction Ox2 + Red1 Ox1 +

Red2 in the dark.116

FL = Fermi level, VB = valence band, CB = conduction band, △s = gradient of

surface potential.

The term “electrochemical” means that the redox reaction consists of two spatially separated half-

reactionsanodic and cathodic.93-100

For the oxidation of a reductant (Red1) by O2

Red1 + 1/2O2 + 2H+ + 2e = Ox1 + H2O, (1)

117-124117-124the anodic reaction is the oxidation of Red1 by either ejected hole or by injection of electron

Red1 = Ox1 + 2e. (2)

The cathodic counterpart is the oxygen reduction reaction (ORR) by ejected electron

O2 + 4H+ + 4e

− 2H2O. (3)

However, at relatively low cathodic overpotentials, the “direct” four-electron reaction (3) has a

much higher kinetic barrier than a less thermodynamically favorable production of ROS which can be

schematized as

O2 + 2H+ + 2e

− H2O2. (4)

Thus, hematite NPs can oxidize biomacromolecules such as proteins, DNA and carbohydrates

and reduce the dissolved O2 molecules to generate ROS which will further affect the cells.

redox reaction in the dark

Anodic

reaction

Cathodic

reaction

23

(b). Size, shape and morphology dependence of (bio-)activity of ferric (hydr)oxide nanoparticles

Both the biotic and abiotic activity of ferric (hydr)oxide NPs is a complex function of their, size, shape,

crystallinity, solubility, porosity, hydrophilicity, catalytic activity, surface charge and aggregation state. In

the biotic case, the phenomenon becomes even more complicated by the profound role of the dynamics

of nano-bio interface including protein corona formation, favorable pathways for cellular uptake,

interference in cellular biochemical processes to name a few.2 Surfactants

125 and proteins

2, 126-129 are

known as modifiers of the (bio)chemical reactivity, stability, and availability of ferric (hydr)oxide NPs.

Specifically, these (bio)molecules can change the wettability, surface charge, and aggregation state of

NPs, as well as can dissolve-re-precipitate them.

Sorption and adsorption of fatty acids and size, shape and surface charge-dependent dispersion

phenomena

It is well recognized that the adsorption of organic (bio)molecules provides a way to manipulate

the nanotechnologically critical interfacial properties of nanoparticles (NPs) such as wettability, surface

selectivity towards specific moieties, surface dipole, surface charge, magnetization, and population of

surface states.130-132

Specifically, wettability determines the stability and foamability of nanoparticulate

dispersions, which need to be controlled in the engineering of nanocomposite materials (e.g., hybrid

materials, polymers, composite electrochemical coatings, ceramics, and paints), biomedical applications

of NPs, and processing of ultrafine minerals. Adjustment of surface dipole and charge, passivation of

surface states, and physical protection of the NPs (e.g., by making their surfaces hydrophobic or

impermeable for gases) are necessary design elements in the NP-based (opto-)electronics,133

while

magnetic properties are critical for the theragnostic applications.134

Although the NP uptake by a cell is a

rather complicated phenomenon,2, 135

it is affected by the surface charge and wettability of the NPs.136

In general, it is assumed that smaller or more porous nanoparticles are more reactive because of

a relatively higher fraction of surface and under-coordinated atoms as compared to the larger or less

porous counterparts, respectively.137-139

However, hematite seems to exhibit the opposite trend. Namely,

the cytotoxicity,112, 140

redox activity,116, 141

and bacterial activity142

decrease with decreasing NP size. This

24

size dependence is consistent with the size-driven variations of the electronic properties of the NPs.109

However, the effects of size can be masked by the effects of the morphology and synthesis history of the

NPs.143

There is also a general understanding that the adsorption properties of particles are determined

by their composition and stoichiometry as well as structural and electronic properties, which in turn can be

manipulated by nanoscaling the solid or varying its morphology (habit/texture, phase distribution, surface

porosity).1, 137, 143-146

At the same time, no study has systematically addressed and delineated the

contribution of NP size and morphology to the interaction of metal (hydr)oxide NPs with carboxylic groups

of organic molecules in aqueous solution. Distinguishing the effect of NP size from the effects of the NP

morphology and synthesis history is challenging problem. The main reason is the practical limitations to

independently vary the nanosize and the morphology of the NPs without employing capping agents

during the synthesis and annealing. Moreover, no data is available on the effect of electric polarization on

the adsorption properties of metal (hydr)oxide NPs in water, although this effect is established for metal

films interfaced with solid electrolytes (known as electrochemical promotion of catalysis)147

and for

semiconducting sulfides in water.120, 148, 149

Semiconducting (nano)particles can electrically be polarized by varying the redox potential of the

suspensions, through contact with other (semi)conducting particles with different work functions (the so-

called galvanic interactions)148, 150

or, as shown in our recent paper,151

upon aggregation or deposition on

different substrates. Hence, being implemented by one of these approaches, electric polarization

presents an important extrinsic control of the adsorption properties of NPs, which has escaped attention

of previous researchers.

Why ferric (hydr)oxides and fatty acids?

The choice of this model system is dictated by substantial interest in commercialization of green and cost-

effective materials and reagents. Ferric (hydr)oxide NPs are industrially used as sorbents,

(photo)catalysts, abrasives, polishing agents, pigments, and components of optoelectronic and

optomagnetic metamaterials,104-106

and are tested in alternative energy152, 153

(“using rust to create

fuels!”154

), sensorics,102, 103

magnetic resonance imaging (MRI)155

and magnetic hyperthermia treatment,

25

as well as in drug delivery.104, 156

These NPs are a natural component of the environment and the living

organisms.106

Long, straight chain monocarboxylic, or fatty acids and their salts (soaps) have long been used

for laundry and personal care since the ancient Egyptian period. They represent by far the highest volume

of all surfactants.157

These reagents are green in the sense that: They have low toxicity and high

biodegradability and can be obtained from renewable resources by the hydrolysis of the plant or animal

fat.158

In modern technology, fatty acids and their derivatives are used to modify surface and dispersion

properties of iron (hydr)oxides in the personal care, biomedical, micro- and opto-electronics, abrasive,

paint and plastics formulations as well as in the mineral processing, corrosion protection, lubrication, and

tar-sands extraction operations. The interaction of ferric (hydr)oxides with fatty acids is also related to

many processes in nature. In particular, being a component of the natural organic matter (NOM), aliphatic

carboxylic acids regulate the fate and transport of the soil components, contaminants, and different types

of bacteria and pathogens in the environment.159, 160

Sodium laurate (dodecanoate), CH3(CH2)10COONa, which is used in this study is classified as a

weak electrolyte.161

Its state in water is controlled by hydrolysis to lauric acid (pKa = 5.0), precipitation (3

mM at pH 7.1), and micellization.161, 162

Critical micellar concentration (CMC) for sodium laurate (SL)

depends on pH, being 1 and 20 mM at pH 7.1 and 10.0, respectively.161

A range from 25 to 42C has

been reported for the Krafft point.161

According to Refs.,163-165

fatty acids can form acid-soap dimers or

low-molecular weight sub-micellar aggregates, although others have argued against their existence.166-168

The formation of self-assembled monolayers (SAM) and bilayers of fatty acids on ferric

(hydr)oxide NPs directly relates to the NP applications in the areas listed above, as well as to

the corrosion protection of iron and steel surfaces given the nanoparticulate structure of rust.

One of the main difficulties in distinguishing between the size and morphology effects of hematite

NPs on their interaction with fatty acids is the controversy in the current knowledge on the adsorption

forms of fatty acids. In earlier macroscopic studies, it is suggested that the chemisorption and

physisorption dominate at pH > PZC and pH < PZC, respectively (based on argument that electrostatic

26

repulsion between the anionic surfactant and negatively-charged surfaces is the determining factor).169, 170

This assumption has been recently questioned using spectroscopic results171-173

and is shown that the

electrostatic repulsion can be compensated by the favorable formation of surfactant hemimicelles and

structured interfacial water (influenced by surfactants).

As discussed in Section 1.1.6 below, to resolve the structures of the adsorbed fatty acids,

spectroscopic data should be complemented by the theoretical modeling. However, to the best

knowledge of the author, density functional theory (DFT) modeling has not been performed for this

system. Also, the effects of NP size and morphology on the interaction of metal oxide NPs in general and

ferric (hydr)oxide in particular have not been addressed, despite their practical importance.

Taking the previous results of the size-dependence of electronic properties of ferric (hydr)oxide

NPs in to account,109

it is hypothesized that the adsorption capacity/affinity of fatty acids should decrease

with decreasing NP size, although effects due to morphology is unknown and may affect this trend. As a

consequence, the stability of the NP suspensions in solutions of fatty acids should also vary. A knowledge

of the size-dependent variation in the dispersion properties of NPs is critical for many applications

including electrophoretic deposition, dewatering of mineral tailings,174

polymer composites, water

purification, bioavailability, and the mitigation of the toxicity of NPs.

Ferric (hydr)oxide nanoparticle-Cell interaction

The interest to the cytotoxicity of ferric (hydr)oxide NPs stems from a high probability of their

contact with living beings. Such NPs are found in oceans, ground and surface waters, dust, and soils.

They are added as pigments to paints, plastics, rubber, building materials, food and pharmaceutical

products, and are among the most widely used catalysts. 106

Superparamagnetic Fe2O3 NPs have been

used for biomedical applications for over two decades as MRI contrast agents for the detection of liver

tumors, and also have been commercialized for the diagnosis and therapy of other specific diseases. 156,

175 Ferric (hydr)oxide NPs are present in the most living organisms, where iron is a biologically important

element as a core of ferritin (an intracellular protein that serves as an iron depot in a body). Since the

released ferric iron can participate in the normal iron metabolism, ferric (hydr)oxide NPs were initially

considered to be non-cytotoxic. At the same time, as a transition metal, iron is able to catalyze the

27

production of hydroxyl radicals (by a Fenton or Haber–Weiss reaction) and thus induce oxidative stress,

which is considered as the principal injury mechanism in cells 70

.

So far, a series of studies have been performed mostly for superparamagnetic - Fe2O3

(maghemite) NPs to address the concerns about the safety of ferric (hydr)oxide NPs. The results, which

are summarized in several reviews,135, 175-177

demonstrate that such NPs can exert drastic effects on the

cell’s wellbeing, and their impact depends on many parameters, such as the exposure concentration and

time, NP size, fraction of Fe2+

(the redox state of iron), surface functionalization, cell line. 176

However,

the great variety in types of the NPs, cells and incubation protocols used have rendered it impossible to

make any conclusions about the safety of magnetic iron oxide NPs.176

Moreover, all the previous works

consider the NP-cell interaction from the nanotoxicological and nanomedical perspectives, being focused

on the assessment of the level/manifestation of inflammatory response, if any, as well as on the fate and

transport of the NPs in the culture media. The results are mostly interpreted in terms of the amount of

leachable iron. At the same time, to our knowledge, no attempt has been made to relate the cytological

effects to the emerging physico-chemical paradigm,70

according to which, along with the acute solubility,

the redox activity is the determining factor of the bioactivity of metal oxide NPs. Moreover, the nanosize

effect on the bioactivity of uncoated ferric (hydr)oxide NPs has not been addressed systematically using a

multidisciplinary approach, in which the cell-NP interaction is considered from the combined perspective

of colloid chemistry, catalysis, cytology, and nanotoxicology.

The current knowledge on - Fe2O3 is in a similar situation, except that the number of

nanotoxicology studies on these NPs is much lower112, 113, 115, 140, 178-181

as compared to maghemite.

Specifically, unclear is the dependence of the cytotoxicity of hematite NPs on nanosize: 50-nm hematite

is genotoxic for human bronchial fibroblasts and epithelial cells, and its adverse effects are more

pronounced as compared to microsized hematite at the same weight doses.178

The 50-nm hematite also

indices functional neurotoxicity of brain cells.180

In contrast, low toxicity was observed for 12-nm hematite

towards human bronchial epithelial cells and rat alveolarmacrophage cells,70

as well as for 30-nm

hematite towards human lung epithelial cells. 115

.

28

Given that the abiotic activity of hematite NPs decreases with decreasing NP size,110

it is

hypothesized, based on the emerging paradigm,70

that the cytotoxicity of hematite normalized to the

surface area of the NPs should also increase with increasing size of the NPs

1.1.6. Challenges in modeling the adsorption of fatty acids at the metal oxide nanoparticle-

water interface

Macroscopic/thermodynamic adsorption measurements can provide a wealth of information on the

interaction of the carboxylate groups with metal oxide NPs. Specifically, using macroscopic methods, the

adsorption of fatty acids on iron (hydr)oxides and corresponding hydrophobicity were related to the

surface tension minimum of the surfactant.163, 182-185

The formation of highly surface active premicellar

species such as acid-soap dimers were suggested as the reason for this relation. However, the

“premicellar” model is inapplicable for hematite and C6-C18 homologues of saturated fatty acids.169

Moreover, this model does not account for the chemisorption of the fatty acid adsorption,182

and neglects

the variation of the oxide surface chemistry with size of the particles.184

A better mechanistic understanding and hence modelling of the adsorption process requires

knowledge of the surface speciation of fatty acids on ferric (hydr)oxides which is currently debated. Based

on the adsorption measurements, the fatty acids are suggested to preferentially physisorb on hematite at

pH < IEP but chemisorb at pH > IEP (IEP-isoelectric point).169, 170

Moreover, the chemisorption was

inferred based on electrostatic repulsion arguments for pH > IEP. However, there are several recent

studies that question this model by suggesting a thermodynamically favourable formation of hemimicelles

and strong hydrophobic interactions can overcome the electrostatic repulsion.173,183, 186

On the other hand,

oleate adsorbed on iron oxides were interpreted as chemisorption based on the conventional spectral

correlations of the splitting between the asCO2– and sCO2

– modes = asCO2

– sCO2

– in the FTIR

spectra and the speciation of the adsorbed carboxylate.171, 172

If so, it is unclear why the hydrophobicity

maximum of oleate-coated hematite is coincident with the highest solution activity of the premicellar

associates of fatty acids.

29

At the same time, resolving the surface speciation of fatty acids at the metal-oxide water interface

is challenging problem, given the several possible adsorption modes (Scheme 1.3) and inefficiency of X-

ray adsorption spectroscopy to study such systems.

Scheme 1.3 Cluster models of different laurate adsorption complexes. ISBB-Inner sphere bidentate

bridging; ISMM – Innersphere monodentate mononuclear; OS – Outersphere; Blue octahedra – Fe3+

; red

spheres – oxygen; gray spheres – carbon; green spheres – hydrogen.

The only remaining practical microscopic alternative is FTIR spectroscopy. However, in common

with all vibrational spectroscopies, FTIR spectroscopy is not directly sensitive to the atomic positions, but

rather to the symmetry and bond strengths. Additionally, the traditional correlation of to the binding type

of the adsorbed carboxylate group187-189

can be misleading because these correlations are established

for bulk coordination compounds and species formed at the solid-gas interface. They neglect strong

effects of the polarization of the anion in the double layer, the surface charge of the NP, H-bonding with

coadsorbed water, and the coupling of the carbonate modes with modes of coadsorbed water, to mention

a few (See Ref.190

for more details).

Thus, there is a need to further develop the DFT method to resolve the bonding type (outer vs.

inner sphere) and configuration of carboxylate groups at the metal oxide-water interface based on their in

situ FTIR spectra.

b) ISMM a) ISBB c) Chelating d) OS surface hydration

-shared

e) OS surface

hydration-separated

30

1.2. Goals and objectives

Based on the many critical unresolved issues formulated in the literature review sections, the following

goals and objectives are determined, and the hypotheses are proposed

Goal:1. Tailoring of the bioactivity of (poly)sebacic anhydride-based microparticle-delivery

systems for articular cartilage tissue engineering

Objective: To overcome the problem of the decrease in the diffusion of nutrients and growth factors due

to progressive elaboration of extracellular matrix in 3D agarose scaffolds with time.

Hypotheses: The problem can be solved using nutrient-encapsulated microparticles based on

(poly)sebacic anhydride polymer which can release the bioactive molecules in situ in a desired timeframe

controlled by the polymer degradation rate. This also has the advantage of a uniform, localized and in situ

delivery of the nutrients or growth factors.

On this basis, the following working hypotheses are formulated,

a. Due to the short term release faster degradation rates required for this objective, there may be an

upper limit for the amount of polyanhydrides-based microparticles in vitro above which cytotoxic

issues may arise (due to the accumulation of its degradation products). This is particularly important

since most of the reported studies that attest to the biocompatibility of polyanhydrides either have

longer degradation profile or have been for delivering adjuvants or cancer therapy.

b. It is hypothesized that serum-albumin, which naturally regulates the intracellular lipid content, can

be applied to mitigate the potential lipotoxicity of polyanhydrides by supplying an abundant quantity

bovine serum albumin in the culture medium.

Goal 2. Understanding the effects of functional groups of polyacrylate nanoparticles on their

bioactivity in articular cartilage tissue engineering

31

Objective: Synthesize functionalized acrylate-based polymeric nanoparticles and study their effects on

biochemical composition and mechanical properties of cartilage

Hypothesis: The surface charge of the polymer alone may play an important role in the chondrocyte-

scaffold interaction.

To test this hypothesis, the effects of acrylate-based nanoparticles with different charge characteristics

when introduced into agarose 3D scaffolds seeded with chondrocytes is studied.

Goal 3. Anti-inflammatory and scavenging (bio-)activity of engineered ceria oxide nanoparticles

Objective: The bioactivity of cerium oxide NP will be explored for their potential benefit in scavenging

inflammatory and oxidative chemicals that are relevant to in articular cartilage regenerative treatments.

Hypothesis: It is hypothesized that presence of nanoceria during the in vitro culture of chondrocytes will

not only enhance the collagen production but also protect the elaborated extracellular matrix from

degradation due to exogenously added interleukin-1α (IL-1α). I have selected IL-1α isoform for our study

since it was already shown that under both in vivo and in vitro condition they not only inhibit the

maturation but also damage the tissues.87, 89, 191

Goal 4. Size, shape and morphology dependence of (bio-)activity of ferric (hydr)oxide

nanoparticles-Interactions with chondrocytes and surfactants

Objectives:

1) To answer the still open questions about whether there are any effects of size, morphology,

and surface charge of the ferric (hydr)oxide NPs in their interaction with cells and surfactants? and if so

why?

2) To rationally control the dispersion and other interfacial properties of ferric (hydr)oxide NPs

using fatty acids

32

3) To understand the effect of size, morphology and surface charge of the ferric (hydr)oxide NPs

on the cellular viability and metabolism of chondrocytes in 3D culture

Hypothesis: The (bio-)activity of ferric (hydr)oxide NPs increases with increasing particle size

and depends on the semiconducting properties of hematite NPs. Thus, it is further hypothesized that the

adsorption density of fatty acids decreases with decreasing NP size, although variations in morphology

can affect this trend. As a consequence, the stability of the NP in fatty acid containing suspensions can

also vary. Additionally, it is also surmised that increase of the particle size of ferric (hydr)oxide NPs might

have larger impact on chondrocytes

Goal 5: To develop a predictive methodology for identification of adsorption forms using density

functional theory (DFT) simulations and FTIR spectroscopy

Objective: The overall objective is to develop a comprehensive methodology to predict the

adsorption forms of fatty acids using density functional theory (DFT) simulations by comparing

theoretically calculated vibration spectra to the experimental spectroscopic data.

Hypothesis: The inclusion of effects of ionic strength, pH and co-adsorption on the DFT

calculated spectra

33

Chapter 2 Materials and Methods

2.1 Materials

Chemicals required for synthesis of polymeric and hematite nano- and micro-particles were

obtained commercially and are of analysis or higher grade. The aqueous solutions were always made

from triple distilled water or nanopure water having less than 18 MΩ/m. Polymeric nanoparticles made up

of the acrylate backbone and different functional groups are synthesized by inverse microemulsion

method. The size of the particles varied between 70-300 nm. Poly(sebacic anhydride) microparticle

particles are synthesized using double-emulsions and solvent-evaporation method. The size of the

microparticles particles ranged between 2-20 microns. Hematite nanoparticles of average sizes 7, 9, 30,

38, and 120-nm are synthesized by forced hydrolysis. Hematite particles of size 150-nm and ferrihydrite

particles of 4-nm size are obtained commercially.

2.2 Tissue culture studies

Agarose-based hydrogel are used in this study as scaffold for culturing chondrocytes. The advantages of

using agarose are outlined in Section 1.1.1. In the first set of studies, the fabricated polymeric nano- and

micro-particles are embedded within the agarose matrix during the casting of the hydrogel scaffolds.

Aqueous melt of agarose of 4% by weight was prepared at 40 0C and mixed with chondrocytes and

polymeric particle suspensions such that final concentration of agarose is 2%. The polymeric particles are

incorporated at varying weight percentages. In the second set of studies, inorganic ceria nanoparticles

and hematite nanoparticles are incorporated at different concentrations. Control sets in both of these

studies are the agarose scaffolds that have neither nano- nor micro-particles. Chondrocytes harvested

from cartilage of bovine calves are passaged up to 4 generations and are added to the agarose melt

diluted with growth media such that the final concentration of chondrocytes in the suspensions is ~30

million cells/ml. These chondrocyte loaded suspensions which are at 40 0C are cast into rectangular mold

and cooled to 37 0C and maintained at this temperature overnight. After which the scaffolds were

punched into cylindrical constructs of ~ 4mm diameter and ~ 2.4 mm length. These constructs were

cultured in Dulbecco's modified essential medium supplemented with amino acids (0.5x minimal essential

34

amino acids, 1x nonessential amino acids), buffering agents [10 mM HEPES, 10 mM sodium bicarbonate,

10 mM TES, and 10 mM N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid], and antibiotics (100 U/ml

penicillin, 100 mg/ml streptomycin). The growth media was regularly replenished and 3-7 constructs from

each experimental and control groups were used for mechanical testing, biochemical and histological

analysis at specified time intervals during entire culture period. The mechanical properties studied include

equilibrium Young’s modulus and dynamic modulus in under confined compression, and biochemical

content studied include proteoglycan represented by glycasoaminoglycans (GAG), collagen and DNA

analyses.

Mechanical properties

A computer- controlled custom made mechanical tester capable of measuring equilibrium

Young’s modulus (EY) and dynamic moduli, G*, (at 0.1, 0.5 & 1 Hz) of constructs in unconfined

compression was used. The EY is defined as the stress generated by the construct under 10% strain and

no external load conditions. The 10% compressive strain is achieved by gradual application of external

load (strain rate of 0.05 %/s). Once the 10% strain is achieved the external load is removed and the

construct is allowed to equilibrate for 30 minutes and this equilibrium compressive stress is used for

calculating the Young’s modulus. The dynamic modulus is measured by imposing 2% sinusoidal strain at

specified frequencies on the above equilibrated constructs.

Cellular viability

The viability of chondrocytes were tested using live/dead cytotoxicity kit from Molecular Probes and

confocal laser scanning microscope from Olympus (Fluoview-100). The cell permeant calcein-AM

fluoresces green only in the cytoplasm of live cells, whereas the cell-impermeant ethidium bromide

fluoresces red when it complexes with the nucleus of compromised cells (dead cells).

Biochemical content

The constructs were dried by lyophilization for 48 hours and digested using Proteinase K for 16 hours at

56 oC.

192 The agarose is centrifuged and the supernatant collected for further analysis of DNA, GAG &

collagen. The concentration of GAG in digested samples is measured spectrophotometrically using 1,9-

35

dimethylmethylene blue dye-binding assay and by comparing with chondroitin-6-sulfate standards193, 194

The collagen present in the digested sample is calculated by measuring ortho-hydroxyproline (OHP)

content through a colorimetric procedure.195

The cellular viability is additionally inferred from the amount

of DNA present in the samples measured using PicoGreen assay. The biochemical contents (GAG,

collagen, and DNA) of constructs are normalized with either wet or dry weight of the constructs.

2.3 Characterization and physico-chemical experimental methods

The primary particle size and morphology are characterized using scanning (SEM) and

transmission (TEM) electron microscopies. The SEM was used for characterizing polymeric nano- micro-

particles used in this study. The SEM is best suited for polymeric particles because particles are larger in

size and can be coated with gold to enhance conductivity as well of contrast. The polymeric particles

were deposited on a carbon substrate and were used with or without a coating of Au-Pd layer. The SEM

images were obtained using a 4700-Hitachi SEM microscope with voltage set at 0.8 kV and current at 20

mA. The ceria and ferric (hydr)oxide NPs were imaged using a Jeol JEM 100CX Transmission Electron

Microscope (TEM) that is operated in the bright field mode at 100 keV. HRTEM images were obtained by

Evans Analytical Group (EAG) with a FEI Tecnai TF-20 FEG/TEM operated at 200kV in bright-field. The

use of TEM is justified since these nanoparticles are small and have poor conductivity. The sample was

prepared by dispersion of nanoparticles on to a lacey copper grid. The SAED aperture of the microscope

used for the TEM imaging was 150 nm, which did not allow selecting a single particle.

The X-ray diffraction (XRD) was performed to measure the overall crystallinity of the ferric

(hydr)oxide NPs. The XRD analysis was conducted using a Scintag Model X2 X-ray powder

diffractometer. A CuK( = 0.154 nm) radiation source operated at 45 kV and 35 mA is used. The scan step

size was 0.05 deg.

The impact of NPs size, morphology or surface charge on a given phenomenon can be plotted

based on either unit mass or unit surface area. The estimates of the surface area are made with a single-

point BET N2 adsorption isotherm (Monosorb, Quantachrome surface area analyzer). The samples are

36

degassed with N2 at 50 oC for 30 minutes before measurements are taken. Two or three replicate

measurements are performed, which gave results that agreed within less than 5%.

The measure of surface charge of NPs, zeta potential ξ, was characterized using electrokinetics

and salt titration method. These two methods provide information on the macroscopic basicity of the NPs

and characterize the point of zero charge and iso-electric point, respectively. Electrokinetic

measurements are carried out using a Zeta Sizer instrument, Nano-ZS from Malvern. Zeta potential (ξ)

measurements are performed within a pH range of 4–10 and at a NPs loading of 0.005–0.01 by vol%.

The concentration of the background electrolyte, NaNO3, is 0.01 M and the pH of each subsample is

adjusted using NaOH and/or HNO3. The pH-adjusted hematite NPs suspensions are allowed to

equilibrate for 24 h and pH is re-adjusted to initial values two hours before the actual measurement.

Since electrokinetic measurements performed did not exclude adsorption natural CO2, salt

titration technique was employed for determining true PZC of NPs.17, 18

Experiments are conducted in a

glove box continuously flushed with N2 gas at room temperature. Freshly decarbonated triple-distilled

water (TDW) is always used and the decarbonation is realized by boiling the TDW for several hours under

continuous N2 bubbling. The NPs suspensions in 0.0015M NaNO3 are decarbonated by N2 bubbling for a

day. For pH measurements, an Accumet combination double-junction Ag/AgCl reference pH electrode is

used under slow stirring condition of the suspensions. pH adjustments are carried out using 0.01M

solutions of either HNO3 (prepared in decarbonated TDW) or NaOH (prepared from a Dilut-It analytical

NaOH, Baker). The total surface area of NPs in the suspension is kept constant at 200 m2. These

suspensions are divided into 5 portions 40 ml each and pH is adjusted to selected values. After

stabilization of pH (0.5–1 h), a known quantity of NaNO3 is added to increase the ionic strength up to

0.1M of suspensions. Another 10-30 min is allowed for pH stabilization to a new pH value. The resulting

change in pH, ΔpH, is determined from the difference (inclusive of sign) between the final pH and the

initial pH. Point of zero charge is the pH at which pH = 0.

To delineate the adsorption modes, whether chemisorption or physisorption, x-ray photoelectron

spectroscopic (XPS) measurements were performed. The XPS spectra are collected with a Perkin-Elmer

PHI 5500 instrument using monochromatic AlK X-rays with pass energies of 17.6 eV at resolution of 0.9

37

eV at take-off angle of 45 at pressures of less than 1 10–8

Torr, calibrated using the Ag 3d peak. Scans of

X-ray induced Auger peaks are performed at 0.1 eV steps. All samples were prepared by spreading a thin

layer of an aqueous suspension of NP on a UHV metallic holder followed by air-drying.

The adsorption isotherms of laurate on ferric (hydr)oxide NPs were measured to understand the

size-dependent surface coverage and packing. Adsorption isotherms of sodium laurate were measured

using depletion method and total organic carbon analysis (TOC-5000A, Shimadzu). 0.15 g (0.2 g in the

duplicate experiments) of particles was suspended in 20 ml of aqueous solution with a 0.005 M NaCl ionic

strength and different concentration of dissolved sodium laurate surfactant at natural pH of 7.5-9.0.

Afterwards, pH of the suspensions was adjusted to 7 using HCl solution and the adsorption is allowed to

proceed for 2 hours under vigorous stirring condition. The suspensions were then centrifuged, the clear

supernatants are pipetted out for the total carbon analysis, while the fraction of the wet particles was used

for the contact angle measurements.

The molecular mechanisms of adsorption of laurate on ferric (hydr)oxide NPs was followed using

vibrational spectroscopy. The Fourier transform infrared spectroscopy-Attenuated total reflection (FTIR-

ATR) spectra were measured using a Perkin-Elmer Spectrum 100 FTIR spectrometer equipped with an

MCT detector. A horizontal ATR accessory consisting of a ZnSe internal reflection element (IRE) (10

internal reflections, angle of incidence 45°) and a cap on IRE to prevent evaporation of water was

employed. A particulate film of hematite NPs was deposited onto the ZnSe IRE. This film was air dried

and then rinsed several times with water to remove any detached particles. The plate with the NPs-

coated IRE was mounted on to the ATR accessory, then water at a preset pH was added to 3 ml IRE cell

and the system was equilibrated for at least 1 h with water being replaced manually every 10 minutes. In

the case of laurate, protein and sulfate adsorption experiments, the background water or electrolyte

solution was replaced with 3 ml of adsorbate solution and the sample spectrum was measured after a

selected equilibration time, both without and with refreshing the SL solution in the FTIR cell. The spectra

were collected for 300-400 scans at a resolution of 4 cm–1

and represented in the absorbance scale. As a

background, either the ATR spectrum of the clean ZnSe IRE covered by water of a selected pH or the

ZnSe IRE covered by the NP layer and the water was used. All the FTIR spectra were measured in

38

duplicate on the NPs synthesized in different batches and were reproducible in terms of the NP size-

induced regularities.

FTIR spectra of polymer samples were measured both as dry sample and as aqueous solution

using ATR setup. Background spectra of air and water were suitably subtracted.

Raman spectra were obtained using Horiba JobinYvon’s Aramis confocal Raman microscope.

Adsorbed protein on ferric (hydr)oxide nanoparticles and polymeric nanoparticles were measured using

785 nm and 532 cm lasers respectively.

2.4 Density Functional Theory (DFT) Methodology

DFT calculations are performed to develop a combined methodology with FTIR spectroscopy to

elucidate adsorption forms of laurate on a hematite surface. The exchange-correlation functional used is

Perdew–Burke–Ernzerhof (PBE) generalized gradient approximation (GGA)19

as implemented in DMol

code and available in Material Studio package (Accelrys Inc., San Diego, CA, USA). The numerical basis

set, namely, double numerical polarization (DNP), which is capable of accurate description of hydrogen

bonding, is employed. Full geometry optimization is performed without spatial or symmetry constraints.

Spin-polarized orbitals are utilized to account for the anti-ferromagnetic properties of hematite. A real

space cut off in the range of 3 – 4.3 Å is chosen for orbitals. Vibrational frequencies are estimated

through a two-point finite difference scheme of analytical forces. The computed frequencies did not

exhibit imaginary frequencies. They are not corrected for anharmonicity and are not scaled. The

calculations are performed both in the gas-phase approximation (at dielectric constant r of 1) and within

the continuum solvation model, COSMO,20, 21

at r of 12, 45, and 78 corresponding to the bulk hematite,

water-hematite interface22

and bulk water, respectively. The problem of modeling surfaces of metal oxide

NPs is generally complicated by a lack of knowledge of their surface structures, which are different from

the structures of their bulk counterparts. In particular, under oxidizing hydrous conditions, hematite NPs’

surface region undergoes partial transformation to the FH phase.23

However, the structure of FH is still

under debate24-26

which poses an essential obstacle for the interpretation and modeling of spectroscopic

and microscopic adsorption data on both hematite NPs and FH. Therefore, in the first approximation, to

elucidate the effect of co-adsorption as such, we model the hematite surface as a cluster consisting of

39

both corner-shared and edge shared multiple ferric octahedra extracted from the (012) hematite surface.

The surface charge of hematite, which depends on the pH of the system, is modeled using different

hydration states of the ferric octahedra.

2.5 Error Analysis of data

The standard error is calculated for all the reported data here using the below formula. The

standard error is an estimate of standard deviation of the mean based on sample population.

√

Where, σ is the standard deviation of the sample and n is the population size

The analysis of variance (ANOVA) was performed to test for the statistically significant difference

between different experimental sets. The ANOVA uses three or more classes of data for these sets. For

eg., To determine if any statistically significant difference exist between control and experimental

constructs, the data from four different mechanical tests - Young’s modulus and three dynamic moduli

(0.1, 0.5 and 1 Hz)) of the constructs were used. In particular, a two way ANOVA with replication was

performed and the statistical significant difference is reported as p-value. If p < 0.05, it implies that there

is a 95% chance that the compared data sets are different.

40

Chapter 3 Synthesis and characterization of

polymers and polymeric micro- and nano-

particles

In the first part of this chapter (Section 3.1-3.5), the synthesis and characterization results of

polymers and polymeric microparticles are presented. The details provided include, (1) synthesis of

polysebacic anhydride polymer using melt condensation method (2) functionalization of this polymer with

fluorescein label and vitamin C (3) characterization of polysebacic anhydride and its derivatives using

NMR and Raman spectroscopies (4) fabrication of polysebacic anhydride microparticles and

encapsulation of bioactive molecules using double emulsion method (5) characterization and release

kinetics of polysebacic anhydride microparticles using optical, electron, and confocal fluorescence

microscopies, as well as UV-Vis absorption spectroscopy.

The second part (Section 3.6) is devoted to the synthesis of acrylate-based functionalized nanoparticles

using inverse microemulsion and their characterization using light scattering, XPS, and SEM methods.

In the final part (Section 3.7), the pore size of agarose-hydrogel scaffold was modeled and determined as

function of agarose alone as well as in the presence polymeric microparticles.

3.1 Synthesis of polysebacic anhydride polymer using melt condensation method

Several methods have been reported for synthesizing polyanhydrides and polyanhydride-esters

including, 196, 197

a) Bulk melt condensation

(b) Ring opening polymerization

(c) Acyl chloride acid-base reaction

41

(d) Interfacial polymerization

Of these methods, melt condensation is the simplest and most common method for synthesis of both

aliphatic and aromatic polyanhydrides. Thus, the melt condensation method was applied for synthesis of

poly(sebacic anhydride) (PSA),with either acetyl and methacryl endcapped. Additionally, to functionalize

the polyanhydrides with L-ascorbic acid (Vitamin C), acyl chloride acid-base reaction was adapted.198, 199

Synthesis of polyanhydrides using melt condensation method involves two-steps (Scheme 3.1): (1)

de-hydration and acetylation of the diacid groups of monomer with acetic anhydride. (2) polymerization of

this precursor anhydride molecule in vacuum and at high temperature to increase both the molecular

weight and the hydrophobicity of polymer.

Scheme 3.1 Synthesis of polyanhydrides using melt condensation method

The poly(sebacic anhydride) (P(SA)) was synthesized using sebacic acid (SA) as a monomer.

(Scheme 3.1) The sebacic acid is a di-carboxylic acid connected by 8 methylene groups. A typical

synthesis procedure adopted is described below: 5 mg of recrystallized SA was mixed with 50 mL of

acetic anhydride and stirred for 20 minutes at 40 0C until a clear solution was obtained. The excess of

acetic anhydride was removed by vacuum evaporation at 70-120 0C. The temperature was further

(CH2)8 - Sebacic acid

(CH2)8

(CH2)8 (CH2)8 (CH2)8 (CH2)8

Acetic anhydride

42

increased and maintained at 150 0C under vacuum and constant stirring for 2 hours. The final product

was dissolved in a minimum amount of methylene chloride and recrystallized using a large excess of

ethyl ether and petroleum ether. The resulting suspension was centrifuged, dried at room temperature

under vacuum, and stored at 20 0C in tightly-capped vials. A typical yield of ~20% was observed after the

purification step and this lower value is due to the losses during the purification step.

3.2. Functionalization of polysebacic anhydride with fluorescein label and vitamin C

3.2.1 Functionalization with Fluorescein

The conjugation of PSA with a cell-permeant dye such as fluorescein will allow studies of cellular

monitoring and uptake of the polymer. For attaching fluorescein label to the PSA, methacrylic anhydride

was used instead of acetic anhydride during the first step in the melt condensation method. The

methacrylic endgroups in PSA was used to link the acrylate double bond in the fluorescein o-acrylate, a

fluorescein derivative. A radical initiated linking between methacrylic group of P(SA) and acrylic group of

fluorescein o-acrylate was carried out in an organic solvent (methylene chloride) at 40-50 0C. The

fluorescein o-acrylate of 1 wt% and two radical initiators 2, 2’ Azobis (2-methyl propionitrile) and 2, 2’

Azobis(2-(2-imidazolin-2-yl) propane] dihydrochloride were used. The reaction was carried out for ~30

minutes under stirring. The resulting polymer was purified by repeated dissolution in acetone and

recrystallization using hexane. The polymer was washed with water several times to remove any residual

unreacted fluorescein o-acrylate.

3.2.2 Functionalization with vitamin C (L-ascorbic acid)

The L-ascorbic acid (Vitamin C) is an essential nutrient for articular cartilage engineering. A local in situ

delivery of vitamin C can also be achieved by degradation of PSA chemically linked to vitaminC. To

functionalize the PSA with L-ascorbic acid (Vitamin-C), esterification procedure based on acid-base

reaction method was adapted. Typically, this method involves reaction between an acyl chloride (eg.,

sebacoyl chloride) and diacids (eg., sebacic acid) or hydroxy acids (eg., salicylic acid) in an appropriate

solvent such as tetrahydrofuran (THF). Pyridine is added not only to activate acyl chloride by formation

43

the acyl pyridinium ion but also to deprotonate the hydroxy- acid-groups. Thus formed precursor diacid

molecule can be dehydrated using acetic anhydride and further polymerized using melt condensation

under vacuum.

The L-ascorbic acid (Scheme 3.2) has several hydroxy groups that can form ester bonds with sebacoyl

groups. This process involves two steps (Scheme 3.2). In the first step, an equimolar concentration of

sebacic acid and sebacoyl chloride were reacted to form a diacid precursor molecule: A 15 mmol of

sebacoyl chloride was dissolved in 10 mL of tetra hydrofuran (THF) and was added drop wise into 40 mL

of THF containing 15 mmol of sebacic acid and 3.5 mL of pyridine. The mixture was stirred at room

temperature for 2 hours and poured over water at pH 2 (adjusted using HCl) to form an off-white

precipitate of oligomeric poly(sebacic acid). The precipitate was filtered and washed with water and dried

overnight under vacuum. In the second step, 1.5 grams of the above product, 375 mg of L-ascorbic acid,

and 365 mg of pyridine were dissolved in 100 mL of THF.550 mg of sebacoyl chloride dissolved in 1.5 mL

was added drop wise in the pyridine mixture. After two of stirring at room temperature, the mixture was

added to 500 mL of HCl solution at pH 2 resulting in an off white precipitate which was filtered and

washed with water and dried at room temperature under vacuum.

Scheme 3.2 Synthesis of L-ascorbic acid functionalized PSA polymer

44

3.3 Characterization of polysebacic anhydride and its derivatives using NMR and Raman

spectroscopies

The chemical identity of the products and the degree of polymerization were measured by NMR

and vibrational spectroscopy.

The NMR spectrum of sebacic acid is shown in Figure 3.1a, where the ratio of the area of

different proton peaks match with its stoichiometry. The protons and their chemical shifts are, a(8H)-1.25

ppm; b(4H)-1.48 ppm; c(4H)-2.18 ppm and e(2H)-11.94 ppm. The corresponding ratio of area are

8H:4H:4H:2H = 1:0.49:0.49:0.22.

If polymerization of sebacic acid into PSA is successful, several changes in peak positions and

area are expected. Namely, the peak at 11.94 ppm that corresponds to the protons of carboxylic acid

group should disappear, the methylene protons conjugated to the anhydride groups should have peaks at

around 2.45 and 2.55 ppm,200

and an additional peak due to methyl protons in the acetic anhydride end

groups should be observed at ~ 2.2 ppm. In fact, all these three requirements are confirmed by the NMR

spectra of PSA (Figure 3.1b). However, the peak at 2.2 ppm is convolution of methylene (conjugated to

anhydride groups) and methyl (from acetyl group) protons. The first peak at 2.17 ppm is due to the methyl

protons and the other at 2.16 ppm is due to the methylene protons. Such a fast degradation of PSA is

expected even in organic solvents.201

The areal ratio of methylene:methyl protons was calculated by a

curve fitting procedure and was found to be 0.7:0.3. Furthermore, the peak at 2.45 ppm is also a

combination of these methylene protons and the residual protons from the solvent d-DMSO.

Decomposition of this band into its components yields areal ratio of methylene:DMSO = 0.78:0.22. In

summary, the protons and their chemical shifts of PSA are, a(8H)-1.32 ppm; b(4H)-1.6 ppm; c(4H)-2.16,

2.45 and 2.55 ppm and d(6H)-2.17 ppm.

To obtain information on the degree of polymerization of the PSA achieved, the area of the

individual proton groups (Figure 4.1b) and their ratio are calculated. Assuming, n, is the number of

repeating units of sebacate monomer in PSA,

Protons in the polymer back bone:

45

Area of the 4 inner most methylene groups (‘a’) = nX8H = 1

Area of the 2 penultimate methylene groups (‘b’) = nX4H = 0.49

Area of the 2 outermost methylene groups (‘c’) = nX4H = 0.7*0.07 + 0.78X0.48 +0.09 = 0.51

Methyl protons in the end groups of PSA chain:

Area of methyl protons in the acetyl end groups (‘d’) = 6H = 0.3*0.07 = 0.021

To calculate n, a:d= nX8H:6H = 1/0.021; n/6 = 47.6; n ≈ 36

Therefore, number averaged molecular weight of the polymer, Mn, Mn ≈ 36X184+74 Da; Mn ≈ 6700 Da

Figure 3.1 NMR spectra of polymers on their precursors; (a) Sebacic acid, (b) Poly(sebacic anhydride)

(inset: expanded and curve fitted region from 2.18-2.48 ppm), (c) Ascorbic acid (d) Poly(Sebacic

anhydride)- Ascorbic acid.

The diagnostic carbonyl region in the Raman spectra of the PSA and their precursors is shown in

the Figure 3.2. In particular, there is only one carbonyl stretching band (νC=O) at 1642 cm-1

for carboxylic

acids. On the other hand, anhydrides have two carbonyl groups that are connected by an oxygen atom,

and hence exhibit two carbonyl bands. One band represents the symmetric stretching (νsC=O) of

46

carbonyls at 1808 cm-1

and the other is asymmetric stretching (νasC=O) at 1742 cm-1

. The symmetric

stretching band is always at higher frequency and is separated from the asymmetric by 50-80 cm-1

,

independent of their actual position.202

Thus the formation of poly(sebacic anhydride) was confirmed by

the disappearance of carboxylic band and appearance of these two diagnostic bands of anhydrides

1H NMR spectra of the PSA functionalized with L-ascorbate along with precursors were

measured (Figure 3.1c,d) to verify the reaction. The proton shifts at 4.98 and 4.02 ppm in the spectrum of

the functionalized polymer (Figure 4.1d) are not present for the original polymer implying successful

binding of ascorbic acid molecule.

Figure 3.2 Raman spectra of polymers on their precursors; Poly(sebacic anhydride) and Sebacic acid

3.4 Fabrication of polysebacic anhydride microparticles and encapsulation of bioactive

molecules using double emulsion method

Microparticles based on acetylated or methacrylated sebacic anhydride polymers (PSA) were

synthesized using a double emulsion method (Scheme 3.4). A double emulsion consisting of three

phases, internal water/oil/external water phases, and poly(vinyl) alcohol (PVA) as a emulsifier at the

interface between the external water and oil phases were employed. Dichloromethane (DCM) was used

as oil phase and an evaporative process was used for elimination of DCM leaving behind the polymer and

encapsulation of the bioactive cargo loaded in the internal water phase. Alternatively, bioactive solids

were directly encapsulated by using them as internal phase (internal solid/oil/external water) instead of

47

water. This overcomes the solubility imposed by water. Even though PVA is a biocompatible surfactant,

the DCM is a toxic organic solvent. Hence, even traces of DCM left behind in the microparticles can be

quite harmful. However, the use of DCM and other microparticle fabrication conditions adopted have been

shown to be harmful neither to cells nor to the bioactivity of encapsulated proteins.203, 204

Scheme 3.3 Synthesis of microparticles using double emulsion and solvent evaporation method

The microparticle fabrication procedure adopted in this study is given below. A 100 mg of PSA

was dissolved in the 1 mL of DCM and the bioactive internal phase either as aqueous solution or solid

phase was emulsified by applying ultra-sonication and homogenization for 60 seconds. The wt. % of the

internal phase is typical set at 50% by weight of PSA (50 mg). The emulsion was then added drop wise to

2mL of the external phase consisting of aqueous solution of an 1% emulsifier (PVA) under continuous

sonication and homogenization for 120 seconds more. The obtained double emulsion was then diluted

with 20 mL aqueous solution of 0.5 % PVA and stirred vigorously for 4 hours to evaporate the oil-phase

(DCM). A dense layer of polymer was left behind encapsulating the internal bioactive-phase thus forming

microparticles. The encapsulated microparticle suspensions were then centrifuged and washed with

48

water for three times. The microparticles were freeze dried and stored at 20 0C until their use. However,

these microparticles were utilized within 24 hours of their fabrication to minimize the effects of their

degradation products. Several variations of the polymers and bioactive internal phase have been utilized

for microparticle fabrication: the variations of the polymers include acetylated-PSA, methacrylated-PSA,

or fluorescein-functionalized PSA; the variations of the bioactive internal phases include aqueous solution

of L-Ascorbic acid 2-phosphate sesquimagnesium salt hydrate or their solid form, solutions of TGFβ-1,

TGF-β3, chondroitinase-ABC, amitriptyline, casein-FITC or glucose. The microparticles intended for the

delivery in the tissue constructs were fabricated under sterile conditions. The water was replaced with

DMEM solution at all stages of fabrication procedure.

3.5 Characterization and release kinetics of polysebacic anhydride microparticles using

optical, electron, and confocal fluorescence microscopies, as well as UV-Vis absorption

spectroscopy.

SEM image of microparticles fabricated using above procedure and encapsulated with casein-

FITC is shown in the Figure 3.3. The size distribution of microparticles extracted using image analysis

program ImageJ is shown in Figure 3.4. The average size of the particles was found to be ~1.5 μm. The

particles smaller than 1 μm were excluded from the analysis to avoid counting the image artifacts. The

dispersed state of the internal phase (aqueous L-ascorbic acid) in polymer was found using fluorescein-

functionalized PSA and confocal microscopy (Figure 3.5). The extent of dispersion of the internal phase

was clearly not uniform: some of the microparticles have one big droplet, whereas, others have several

well dispersed smaller droplets.

Figure 3.3 SEM images of PSA-microparticles encapsulated with casein-FITC after 24 hours in aqueous

suspension. All the images are from the same batch of microparticles but at different magnifications.

49

Figure 3.4 Particle size distribution of PSA-microparticles with casein-FITC after 24 hours in aqueous

suspension. The size distribution was obtained from SEM image analysis performed using ImageJ

program.

Figure 3.5 L-ascorbic acid loaded polyanhydride microparticles embedded in agarose scaffold. Green

color indicates the fluorescein-functionalized PSA polymeric shell. Black core is the encapsulated L-

ascorbic acid. Both the images are from the same microparticle sample

50

Release kinetics of PSA microparticles

Release of amitriptyline molecules from encapsulated-microparticles was studied to verify the

linearity of the release profile. Absorption of UV-light at the wavelength 239 nm by amitriptyline was used

to monitor its release. The concentration of the released amitriptyline molecules in microparticle

suspension was measured for every few days by centrifuging the suspension and extracting 5 mL of the

supernatant and was stirred to resuspend the microparticle particles. The encapsulation efficiency was

found to be ~50%. Within 2 weeks, ~60 % of encapsulated amitriptyline was released from the PSA

microparticles and a linear profile was observed (Figure 3.6). This shows that PSA-based polymeric

microparticles can be utilized to provide sustained linear delivery of enzymes or drugs with no initial burst

effect.

Figure 3.6 Release profile of amitriptyline drug from poly(sebacic anhydride)-based microparticles

3.6. Synthesis and characterization of acrylate nanoparticles

3.6.1. Ternary phase diagrams

Cross-linked polymeric nanoparticles were synthesized using the microemulsion technique. Nano-sized

water droplets stabilized in oil continuous phase by low hydrophilic-lipophilic balance (HLB) surfactant

51

such as Aerosol-OT (AOT) were used as a nano-container for synthesis of acrylate nanoparticles through

radical polymerization and cross-linking reactions. The inverse-microemulsion system of

water/AOT/hexane was used due to higher stability in the temperature range (40-60 0C) required for the

initiation of radical polymerization. The phase behavior was measured and phase diagrams were

constructed for water/AOT/hexane and (water+acrylic acid)/AOT/hexane systems at room temperature

(Figure 3.7a, b, respectively). Addition of acrylic acid monomer had in fact enlarged the stable region of

the inverse microemulsion phase. The phase diagram of (water+acrylic acid)/AOT/hexane system at 50

oC were constructed (Figure 3.8) and was found to be very similar to room temperature phase diagrams.

Two concentration points were selected for nanoparticle synthesis experiments (Figure 3.8), one of the

point is closer to the phase boundary of inverse microemulsion (phasepoint 1) and another point is well

inside this phase region (phasepoint 2).

3.6.2.Synthesis of poly(acrylic acid) nanoparticles and its functional derivatives

The poly(acrylic acid) nanoparticles were synthesized as shown in Figure 3.9a using bisacrylamide as a

crosslinker. The acrylic acid monomer was added to the water phase (~40 % by weight of water) along

with crosslinker (1% by weight monomer) and initiators (ammonium per sulfate and ascorbic acid, ~ 1%

each by weight monomer). The oil phase containing hexane was prepared by dissolving Aerosol-OT

surfactant (~50% by weight of oil phase). The water phase was added drop wise into hexane in a three-

necked flask under constant stirring condition. A clear inverse emulsion solution was noted which was

then flashed with ultra-pure N2 gas for 30 min. A reflux condenser was fitted and temperature was ramped

up to 50 – 700 C. The radical polymerization and crosslinking reactions were initiated due to the

increasing temperature and when the crosslinked nanoparticle particles were formed, the clear

microemulsion turned into a dull white emulsion system. The polymerization in the turbid emulsion was

then terminated (after ~ 30 minutes of maintaining at 50 – 700 C), cooled down to room temperature, and

the resulting polymeric nanoparticles were precipitated by addition of acetone.

52

Figure 3.7 Phase diagrams of (a) hexane/water/AOT (b) hexane /(water+acrylic acid)/AOT systems at

room temperature. Only inverse emulsion region is indicated by colors

The precipitate was filtered and washed repeatedly with hexane and acetone to remove the Aerosol-OT

surfactant and unreacted monomers. These polyacrylic acid nanoparticles were then freeze-dried for 2

days and stored at -20 0C. Polyacrylamide nanoparticles were synthesized using the same procedure and

chemicals except for the monomer which was changed from acrylic acid to acrylamide (Figure 3.9b). A

similar procedure was followed for polyacrylic acid nanoparticles functionalized with sulfonates, and

phosphonates, and amides by replacing only 10 % of the acrylic acid monomer with vinyl sulfonates, and

vinyl phosphonates, respectively (Figure 3.9c, d).

Figure 3.8 Phase diagram of hexane /(water+acrylic acid)/AOT systems at 50 C. Only inverse emulsion

region is indicated by the enclosed region.

(a) (b)

53

Figure 3.9 Synthesis of functionalized polyacrylate nanoparticles, (a) polyacrylic acid (b) polyacrylamide

(c) poly(acrylic acid-phosphonate) (d) poly(acrylic acid-sulfonate); Bisacrylamide is the crosslinker and

ascorbic acid is the initiator

3.6.3.Size determination using SEM and dynamic light scattering

The size of nanoparticles synthesized at the two phasepoints from Figure 3.8, were observed to

be widely different (Figure 3.10). In the case of phasepoint-2 (interior of inverse emulsion phase region),

the nanoparticles of smaller size were noted (70-nm to 300-nm) as seen in Figure 3.10b. However for the

phasepoint-1 (closer to the phase boundary), the size of nanoparticles increased by an order of

magnitude to become microparticles (Figure 3.10a). The size range of nanoparticles prepared using

phasepoint-2 were independently confirmed by dynamic light scattering measurements (Figure 3.11).

Figure 3.10 SEM micrographs of poly(acrylic acid) particles synthesized using concentrations at (a)

phasepoint-1 (b) phasepoint-2

54

Figure 3.11 Size of nanoparticles using dynamic light scattering technique, (a) poly(acrylic acid), 130-

340-nm (b) poly(acrylic acid-acrylamide(10%)), 100-320-nm (c) poly(acrylic acid-phosphonate(10%)),

140-260-nm (d) poly(acrylamide) 70-90-nm nanoparticles

3.6.4 Atomic composition using X-Ray photoelectron spectroscopy

X-ray photoelectron (XPS) spectroscopy were utilized to confirm the successful completion of

chain polymerization and crosslinking reactions. A sample XPS spectra of poly(acrylic acid) nanoparticles

was shown in Figure 3.12, where the CH2, C=O, and COO- peaks were identified to be the nanoparticle

peaks. A very small peak of C=C at 285 eV indicates that small amounts of monomers were left in the

solid unreacted.

3.6.5 Surface charge using electrophoretic mobility (Zeta potential) measurements

Zeta potentials of the nanoparticles were measured, and the isoelectric point was found to be <

pH 2 for all types of nanoparticle particles considered in this study (Figure 3.13). However, there is a clear

difference in pH dependence of zeta potential between poly -acrylamide, -phosphonate and -acrylic acid

nanoparticles. As a result, at pH ~7 (physiological pH ) acrylamide had lowest negative charge and

phosphonate nanoparticles have highest negative charge.

(a) (b)

(c) (d)

55

Figure 3.12 X-ray photoelectron spectroscopy of poly(acrylic acid) nanoparticle. Carbon 1S spectral

region is shown.

Figure 3.13 Zeta potential of poly(acrylamide), poly(acrylic acid-phosphonate), poly(acrylic acid)

nanoparticles

3.7 Pore-size determination of agarose-based hydrogel as a function of agarose

concentration and presence of polysebacic anhydride microparticles

It is known that pore-size of the agarose hydrogel varies with agarose concentration. To select

the optimum concentration of agarose in the agarose hydrogel scaffold, it can be suggested that the pore

size should be large enough for the easy transport of growth factors and nutrients (> 30 nm), but smaller

than 300 nm so as to retain secreted tissue components such as collagen fibrils. However, there is no

56

direct method to evaluate this quantity. This problem is very technically challenging, especially in the

presence of polymeric particles.

Turbidimetric measurements offer one of the fast and elegant method for measuring average

pore size of hydrogels formed by biopolymers.205, 206

The turbidity of an aqueous suspension of particles that are monodisperse and optically isotropic

was given by Doty and Steiner207

Turbidity, ( ) ( ) ( ) (3.1)

Where,

H(λ) is optical constant; Q(λ) is intraparticle dissipation factor

S(λ) is interparticle correction factor; M is molar mass of the particle

c is concentration of particles; λ is wavelength of incident light

The optical constant, H, depends on the refractive index of solvent (n0) and particle (n) as well as

wavelength of incident light (λ). The wavelength dependence of refractive indices of particles and medium

can be ignored in the first approximation. Then, the optical constant, H, can be approximated by the

following expression,208

( )

(

)

(3.2)

Where,

n0 and n are solvent and particles refractive indices, respectively; Na is Avogadro number

The intraparticle correction factor, Q(λ), represents integration of reduction in scattering intensity

due particle shape over all the viewing angles above acceptance angle, α. The value of Q varies from 0 to

57

1. For particles much smaller than wavelength of the incidents (D<<λ), there is no reduction in scattered

light intensity due to shape of particles. The particles are said to be in Raleigh regime and Q≈1.205

For

particle sizes equal to or greater than 5% of incident wavelength, Q deviates significantly from 1.205

The Q is defined for unpolarized light as follows,

( )

∫ ( )( )

(3.3)

Where, P(θ) is form factor and θ is scattering angle.

The form factor, P, can be either obtained by experiments through measuring scattered light at

various angles or calculated theoretically from shape-dependent expressions derived from pair-pair

correlations.209-211

The agarose chains in the hydrogel forms bundles of radii 0.9-10 nm212

allowing formation of

large pores of water.213

Thus, agarose can be approximated as Gaussian coils which allows defining the

form factor P as

( ) (

) ( ) (3.4)

Where, q = 4πn0 sin(θ/2)/λ;

ξ is correlation length ( ≈ average pore size)

N is apparent number of scattering units.

In the case of agarose fibers in water, N can be neglected since it is proportional to ξ/r0, the

monomer radii and the r0 << ξ as well as P0(q) can be assumed to be 1.208

Then equation 3.4 reduces to

( )

(3.5)

The final correction factor in the turbidity equation (3.1), the interparticle correction factor (S), can

be assumed to be 1 under the dilute particle concentration used.7

58

To obtain any structural information such as particle size or pore size using equation (3.1), it is

necessary to measure the absolute values of turbidity, which is non-trivial. The absorbance values

reported by most of the spectrophotometers are in arbitrary units. Thus

An alternative method that obviates the above problem is to measure the turbidity as a function of

wavelength. Thus equation (3.1) can be rewritten as,

( ) ( ) ( ) ( ) (3.6)

Differentiating with respect of log(λ),

( )

( )

( )

(3.7)

Using Eq. (3.2) into (3.7)

( )

(

)

(3.8)

The second and third term on the right hand side of the eq. (3.8) is evaluated to be 0.0248 and

0.0922 for wavelengths between 700 and 800 nm.214, 215

Thus the equation (3.7) can be written as,

( )

( )

(3.9)

Melik and Fogler216

has suggested that for extracting approximate particle sizes, the left hand

side of equation (3.9) (slope of turbidity and wavelength plotted in log-log scale) can be

considered as wavelength exponent (WLE).

( )

(3.10)

59

Using the procedure adopted by Aymard et al.,208

the dependence of wavelength exponent (WLE)

on pore size (ξ) for agarose-like material can be theoretically calculated using equations (3.2),

(3.4) and (3.10). The calculated WLE as a function of pore size is shown in Figure 3.14 for

different acceptance angles

Figure 3.14 Theoretical calculated wavelength exponent as a function of pore size for wavelength range

of 700-800 nm.

The experimental WLE from 700 to 800-nm is extracted from turbidity measurements as follows,

Turbidity, ( ) ( )

(3.11)

Where, l is the path length = 1 cm.

The WLE is obtained by fitting a slope of log-log plot of τ and λ from 700-800 nm

( ( )

)

(3.12)

60

Thus, agarose solutions were melted by increasing temperature to 95 0C. The absorbance of the

melt from 700-800 nm was measured immediately by placing them in UV-Vis spectrophotometer. The

absorbance was measured every 5 minutes for next 2 hours as the melt cools down to room temperature

and forms hydrogel network. The wavelength exponent (WLE) is extracted from the absorbance values

as mentioned before and pore size values were obtained using the established relationship (Figure 3.14).

The details of extracting WLE from the absorbance data is given in Appendix 2. The average pore size as

a function of cooling time is shown in Figure 3.15. The average pore size of agarose hydrogel at room

temperature as a function of agarose concentration is shown in Figure 3.16. It is noted that as the

agarose concentration was decreased from 4% to 1%, pore-size increased from 40-nm to 520–nm

(Figure 3.15).

Using this method, the variation of pore size with respect to agarose concentration was

empirically derived to be,

ξ = 521.95 C-1.833

where ξ is the pore size and C is the agarose concentration in w/v %.

When the polysebacic anhydride microparticles are introduced at 0.3% w/v to the agarose melt,

there was small an increase in pore size from 150-nm to 190-nm. This method failed to determine the

pore size when the microparticles concentration was greater than 0.3 %. This is due to the fact the

increased fraction of large microparticles themselves act as strong scatterers of the radiation apart from

the agarose fibers.

61

Figure 3.15 Pore size of agarose hydrogel as function of cooling time. The temperature at time t=0 is 95

0C

Figure 3.16 Effect of agarose concentration (C, % w/v) on the pore size (ξ, nm) of agarose hydrogel

ξ = 521.95 C-1.833

R² = 0.9998

0

100

200

300

400

500

600

0 2 4 6

Po

re s

ize,

ξ, n

m

Concentration of agarose, w/v %

62

Chapter 4 Synthesis and characterization of ferric

(hydr)oxide nanoparticles

This chapter presents the synthesis procedures used to obtained hematite nanoparticles of

different sizes, as well as the characterization of these nanoparticles using TEM, HRTEM, XRD,

Raman, XPS, acid-base titration, and zeta potential methods.

4.1 Synthesis of hematite nanoparticles

Two samples of hematite nanoparticles NPs (H7 and H9) are synthesized according to Ref.217

by slowly

dripping 60 mL of 1 M ferric nitrate (Sigma-Aldrich, 99.99% purity) solution into 750 mL of boiling TDW.

After the drip solution is consumed, the NP suspension is removed from the heat source. Difference

between H7 and H9 is that in the first case heating of the solution is stopped immediately after dripping is

started. H38 is synthesized according to Ref.217

by heating 8.08 g of ferric nitrate in 1 L of 0.002 HCl at

98C. This suspension is aged in the oven at 98°C for 7 d. In all cases, the precipitation by hydrolysis is

followed by cooling the suspensions overnight and dialysis for 2 weeks using singly distilled water until

the conductivity of the dialyzed water reaches that of the pure water. Sample H150 synthesized by the

annealing of ferric sulfate is purchased from Fisher. 2-line ferrihydrite (FH) is purchased from Alfa Aesar

as Iron(III) Oxide, 99.95% (metal basis), 3nm APS powder, surface area 250 m2/g. Star-shaped H400 are

synthesized by adapting the protocol of Ref.218

4.2 Characterization of ferric (hydr)oxide nanoparticles

Morphology and size of the NPs are characterized using TEM (Figure 4.1). The samples are labeled

according to mean NPs size as measured by averaging sizes of 10–15 NPs. FH NPs are strongly

agglomerated. Shape and size of the primary FH NPs cannot be resolved by the instrument used.

Reported crystal size of 2-line FH is about 2 nm,219, 220

while the extension range of local structural order

63

is as low as 4 Å.219

H7 and H9 hematite NPs exhibit average diameter of 7.0 0.5 and 9 1 nm,

respectively and a rounded irregular hexagonal shape as viewed along the optical axis of the microscope.

H30 have agglomerate-like irregular shape and average diameter of ~30 nm. These globules are porous,

which is confirmed by the fact that their geometrical surface area (GSA) is smaller than the BET surface

area value (Table 4.1). H38 NPs have well developed rhombohedral shape. Examination of the TEM

image does not reveal pores on the surface of H38 NPs and gives their average diameter of 38 5 nm.

The GSA value for H38 matches the BET surface area within the experimental error, confirming that H38

NPs are non-porous. H120 are rhombohedral, while H150 NPs are rounded polyhedral. Both H120 and

H150 are highly monodisperse and non-porous.

Additionally, the morphology of H38 was studied using HRTEM and selected area electron

diffraction (SAED) (Figure 4.2). In particular, it can also be noted that the faces of the H38 rhombohedra

are essentially flat, although the corners are slightly truncated. Furthermore, H38 NPs are enclosed by

rhombohedral facets of the {104} and {012} families. The morphology of the hematite NPs with the same

shape and HR-TEM pattern was characterized in detail in Ref.221

It was demonstrated that this pattern

corresponds to the (104) and (0-14) facets of rhombohedral hematite with zone axis of [-441]. At the

same time, the diffraction spots in the SAED pattern are attributed to the (012) and (014 ) planes of

rhombohedral hematite with the [100] zone axis. Aspect ratio of the NPs used in this study is similar

(ranging from 1 to 2).

The phase identity of hematite NPs and FH used in this study is verified using XRD (Figure 4.3a).

The diffractogram of FH is dominated by two broad lines characteristic of 2-line FH. There are also minor

peaks (eg. at 2θ=33) which point to the presence of a small admixture of hematite phase in FH. Except

for H7, all the hematite NPs show only hematite diffraction patterns, with no lines due to other impurity

phases. H7 too has hematite peaks but also has some admixture of 6-line ferrihydrite. For all hematite

NPs, the (110) and (300) lines are sharper than lines (104), (018), (202), and (024). It follows that

hematite NPs have either plate-like shape with an excess of the planes (near) perpendicular to the c axis

or better developed basal facets. The XRD lines gradually sharpen as NP size increases, which is a

signature of an increase in the average crystallinity size. The latter was quantified using the Scherrer

64

formula with a shape factor of 1 after removing the instrumental broadening by simple subtraction. The

instrumental broadening was taken as the line widths in the XRD pattern of micron-size natural hematite.

All the samples are tested for the presence of surface impurities using XPS. In addition to adventitious

carbon, only ferric iron and oxygen are detected on hematite NPs, while FH is found to contain Si at the

Si/Fe atomic ratio of 0.025. (See Ref.109

for more details on XPS and XRD analysis)

Comparative analysis of crystallinity of the NPs was performed using relative intensities of the

crystallinity-sensitive bands at 660 and 690 cm1 in the Raman spectra.

222 Intensity of both these bands

and hence crystallinity of the NPs increases with increasing NP size (FH < H9 < H38 < H150) (Figure

4.3b). UV-Vis spectroscopy revealed that the band-gap of the NP decreases in the same order.109

As

seen from Figure 4.4, PZC and the iso-electric point (IEP) of H9 and H38 are within the 8.5–9.5 range

typical for hematite NPs synthesized by the hydrolysis of a ferric nitrate/chloride solution and not

subjected to drying,106, 223, 224

although IEP of 9.2 of H38 is at the high limit of the range. H150 is more

acidic (PZC = 6.2 and IEP 7.0), which can be attributed to either a trace amount of acidic surface

impurities or the high-temperature route of the H150 synthesis that eliminates the most basic single-

cation coordinated hydroxyls. FH is characterized by IEP of 7.00.1, comparable with reported values of

7.0 (Ref. 225

), 7.2 (Ref. 226

), and 7.5 (Ref. 227

). H38 and H150 are the most basic and acidic among the

NPs, respectively. The macroscopic acid-base properties of H9 are intermediate, while FH is more acidic

than H9. Both PZC and IEP do not change monotonically with increasing the size of the NPs.

Figure 4.1 TEM images of hematite and FH NPs. NPs are labeled according to the average size as

measured by the analysis of TEM images.

65

Figure 4.2 Typical (a) HR-TEM image and (b) SAED pattern of 38-nm hematite (H38)

Figure 4.3 (a) XRD and (b) Raman spectra of dry hematite and ferrihydrite NPs. “T” marks the peak due

to tetrahedral defects, LO Eu is the peak activated by the lattice disorder.222

115

65

(0 ̅4)

5 nm

220 mm b

66

Figure 4.4 Effect of size of hematite nanoparticles on their point of zero charge and isoelectric point

As seen from Figure 4.17, values of the point of zero charge (PZC) and the iso-electric point (IEP) of

hematite NPs vary non-monotonically, with H38 and H150 are the most basic and acidic, respectively. FH

is characterized by IEP of 7.00.1, comparable with reported values of 7.0 (Ref. 225

), 7.2 (Ref. 226

), and

7.5 (Ref. 227

). The macroscopic acid-base properties of H9 are intermediate, while FH is more acidic than

H9.

67

Table 4.1 Morphological characteristics of hematite NPs

Sample

Average

Particle

Size

(TEM),

nm

Particle Habit

(TEM)

BET,

m2/g

GSA, m2/g

Approximation of shape,

heighta)

(h) in calculating

GSA

FH 2 b)

250 c)

–– ––

H9 9 Rounded

platelets 128.3 440/165

d)

Hexagonal platelets h = 1.5

nm/spheres

H38 38 Hexagons and

rhombohedra 37.0 40 Rhombohedra h = 26 nm

H150 150 Rounded

polyhedra 9.5 8.8

Rectangular prisms h = 80

nm

a) Sizes for hexagons, rhombohedra, and rectangles were taken based on the TEM images. Values

of h for H7 and H9 were approximated by the values provided by AFM for similarly synthesized

NPs,228

for H30–H60 h is taken as a mean coherence length in the [104] direction, while for H120

and H150 h is taken the average shortest dimension as measured by TEM. Size dependence of

the density of hematite NPs was obtained through linear interpolation between values for FH (3.96

g/cm3) for 2 nm NPs and a bulk value of 5.23 g/cm

3 for 120 nm NPs.

b) Cannot be resolved.

c)

Provided by manufacturer. e)

Second value is obtained approximating the NP shape by a sphere

with diameter taken as mean coherence length in the [110] direction;

68

Chapter 5 Bioactivity of polymeric particles in

articular cartilage engineering: In situ delivery

using poly(sebacic anhydride) microparticles and

functional group effects of poly(acrylate) based

nanoparticles

This chapter is divided into two parts: In the first part, the in situ delivery of bioactive molecules

using polysebacic anhydride microparticles (PSA-microparticles) is discussed. A cytotoxic limit for PSA-

microparticles is determined and the cytotoxicity is found to be due to the PSA and its degradation

products. The mechanism of cytotoxicity is deduced to be due to lipotoxicity of the PSA degradation

products. Several mitigation strategies for the PSA were tested and serum albumin treatment is found to

be quite effective. In the second part, the effects of functionalization of polyacrylate nanoparticles on the

cellular viability and metabolism are detailed. Finally, a summary of the results is presented.

5.1 Effects of in situ delivery of bioactive molecules on the chondrocytes cultured in 3D

scaffold

The objective of the first study is to overcome the problem of the decrease in the diffusion of

nutrients and growth factors due to progressive elaboration of extracellular matrix in 3D agarose scaffolds

with time. The polysebacic anhydride has high potential for fabricating microparticles with and

encapsulating critical nutrients (Vitamin C) or growth factors (TGF-β3) as well as for short-term delivery

(2—3 weeks) of the nutrients in tissue engineering applications. Thus, the goal of this study is to tailor of

the bioactivity of polysebacic anhydride microparticle-delivery systems that are cast together with

chondrocytes in a 3D hydrogel scaffold (Scheme 1.1).

69

5.1.2. Dose dependent cellular viability in the presence of poly(sebacic anhydride)

The dose-dependence of the biocompatibility of PSA-microparticles towards chondrocytes was

determined using a 3D agarose-scaffold.

Briefly, the samples were prepared as follows: the PSA-microparticles was encapsulated with

stabilized-vitamin C (Section 3.5) and cast along with chondrocytes in an agarose hydrogel slab. The final

concentrations of the components of slab were agarose-2%, chondrocytes-30 million/mL and PSA-

microparticles at 0.2%, 2% and 10% of w/v of the hydrogel scaffold. More details were provided in

Section 2.2. Each of these concentrations was cast and studied in separate batches and thus had its own

control set.

At 0.2 w/v % of PSA-microparticles, there is no difference in cellular viability as compared to

control (Figure 5.1). However, the viability decreases significantly for both 2% and 10% concentrations: At

the 10% polymer concentration, almost all the cells were dead, whereas 2% shows very little viable cells.

From visually inspecting the culture medium of the 2% and 10% concentrations, no metabolic activity was

observed throughout the culture period. Metabolic activity can be inferred from a decrease in pH due to

secreted acidic metabolic products (ex. lactic acid) which change the color of the phenol red indicator in

the culture medium from red to yellow. A comparison of the equilibrium Young’s modulus (EY) of these

constructs confirmed these findings (Figure 5.2). The y-axis of Figure 5.2 represents a normalized EY

value of microparticle-containing constructs with that of control constructs without microparticles. The x-

axis represents the concentration of polymeric microparticles (w/v %). These results together (figure 5.1

and 5.2) indicated that the presence of PSA polymeric microparticles even as low as 2% (20 mg/mL) was

detrimental to the cellular viability and their metabolism. This is a highly surprising result considering the

fact that majority of the literature considers the PSA to be highly biocompatible.

5.1.3 Mechanisms of the cytotoxicity of poly(sebacic anhydride) microparticles

Following the conclusion from the previous section that PSA has detrimental bioactivity above a

certain concentration, it is necessary to identify the mechanism of cytotoxicity for devising mitigation

strategies. As outlined in the Section 1.1.1(d), several reasons can account for the cytotoxicity of

70

Figure 5.1 Effect of polymer concentration on chondrocytes viability on day-1 in 3D agarose scaffolds.

PSA concentrations are (a) 0.2% (b) 2% and (c) 10% w/v. Green color indicates cytoplasm of the live

cells and red color indicates dead cells and probably microparticles

Figure 5.2 Equilibrium Young’s modulus of hydrogel constructs with different concentrations of

poly(sebacic anhydride) microparticles. The y-axis values are normalized by modulus of the

corresponding control constructs (without microparticles). The concentration of polymeric microparticles in

x-axis is represented as w/v% of hydrogel constructs

0

0.2

0.4

0.6

0.8

1

1.2

PSA-0.2% (day-15) PSA-2% (day-52) PSA-10%No

rma

lize

d e

qu

ilib

riu

m

Yo

un

g's

mo

du

lus

,

Poly(sebacic anhydride) microparticles, w/v %

Equilibrium Young's modulus vs. polymeric microparicles

increasing polymer concentration

71

poly(sebacic anhydride) microparticles. These include (a) shock to chondrocytes due to changes in local

pH; (b) cellular uptake or the damage to the cell wall by the polymer or its degradation products and the

generation of ROS due to the lipotoxicity of the degradation products. To test the hypotheses, (a) the

buffer capacity of the growth medium was increased to offset any local pH changes and (b) cellular

uptake was monitored using fluorescein-functionalized PSA-microparticles and confocal fluorescence

microscopy.

(a). Are cytotoxic effects of PSA-microparticles on chondrocytes due to local pH changes?

Testing the hypothesis through increased buffer capacity

A preliminary acellular experiment showed a sudden drop in pH of growth (CM) media from 7.5 to

6.8 when PSA microparticles were added at a concentration of 30 mg/mL. The pH decreased further to ~

6.3 within 5 hours. This implies a partial hydrolysis of anhydride bonds of polymeric shell of the

microparticles that produces carboxylic acid groups during the fabrication, freeze drying and storage

steps.

To verify if the pH drop due to PSA is the cause for observed cytotoxicity, the chondrocytes were cast

using CM growth media with increased buffer capacity by adding HEPES at 5mM concentration (for

justification of the selected HEPES concentration, see Appendix-A where the amount of buffer capacity

needed for offsetting the pH drop due to polymer degradation is estimated). Additionally, the carboxylic

groups of PSA were neutralized by pre-suspending the microparticles in 5 mL of growth medium (CM)

containing sodium bicarbonate. The particles were centrifuged and the medium was replenished. The

process was repeated until no visual drop in pH was observed. The microparticles were then

resuspended in a fresh growth media (with 5mM HEPES buffer) in appropriate quantity before mixing into

agarose melt. The microparticles were added at a concentration of 2% w/v. In this particular study,

microparticles were encapsulated with chondroitinase-ABC enzyme. The chondroitinase-ABC selectively

digests the GAG molecules and thus makes space for collagen accumulation. Even though the buffer

capacity was increased, as seen from Figure-5.3, cellular viability did not improve. In fact, right from the

beginning of the culture period, no change in the media color and hence no metabolic activity was

72

observed for the constructs with embedded microparticles (as opposed to the control). These results

prove that the local pH changes are not likely the cause of lack of metabolic activity.

(b). Are cytotoxic effects due to cellular uptake PSA-microparticles or its degradation products

and the resulting lipotoxicity? Confocal fluorescence studies

The main degradation product of polysebacic anhydride is polysebacic acid which mimics alkyl

fatty acids due to the presence of carboxylic acid groups and long hydrophobic chains. It is known from

the literature that excessive intracellular concentrations of alkyl fatty acids can cause abnormal levels of

ROS and can result in cellular death.49, 229

If the chondrocytes ingest the PSA-microparticles or its

degradation products, then a priori, the alkyl fatty acid-like concentrations inside cells can be expected to

increase thus causing lipotoxicity. To test if the cytotoxicity arises due to cellular uptake of PSA polymer,

a fluorescein label was attached the PSA as detailed in Section 3.2. This fluorescein-containing PSA

(fPSA) was then added to a 2D culture of chondrocytes aggregate and observed under confocal

fluorescence microscope for cellular uptake of fPSA (Figure 5.4). It can be seen that within 60 minutes

there was significant accumulation of fluorescein-labeled PSA in cells (Figure 5.4).

Figure 5.3 Effect of increased in buffer capacity (5mM HEPES) of growth medium on chondrocytes

viability on day-23 in 3D agarose scaffolds using Live/Dead cytotoxicity kit. The agarose constructs

contain (a) no PSA (control) (b) 2 % PSA microparticles. The microparticles are encapsulated with

chondroitinase-ABC. Green color indicates cytoplasm of the and red color indicates dead cells and

probably microparticles

73

Figure 5.4 Cellular uptake of fluoresceinated-PSA (F-PSA) by chondrocytes. F-PSA is added nearby a

chondrocytes aggregate at time t=0 and images were obtained at time, t= a) 5 min b) 20 min. c) 60 min.

Gradual increase in the fluorescence intensity is due to intracellular accumulation of F-PSA

5.1.4 Mitigation of the cytotoxicity of PSA on chondrocytes: positive impact of bovine serum

albumin (BSA)

Several strategies were applied to mitigate or reverse the cytotoxicity due to the PSA. The mitigation

strategies were tested on a 3D agarose systems seeded with chondrocytes and compared to control

scaffolds without the PSA microparticles. A summary of these studies and rationale for the selection of a

particular strategy are given below.

Initially, pH changes due to PSA uptake were thought to be the reason for cytotoxicity and hence the

buffer capacity of the medium was increased by increasing concentration of NaHCO3, Tris-HCl or HEPES

buffers. In the next set of experiments, osmolality increase due to PSA uptake was hypothesized to be

cause of cytotoxicity and hence trimethyl-amine oxide (TMAO), which is a known osmolyte, was used.

Additional tests were performed by decreasing the osmolality of growth media to offset any increase due

to PSA. None of these methods had any positive impact in mitigating the cytotoxicity due to PSA. Thus,

an alternative strategy was adopted to mitigate the toxicity, that is, to bind the degradation products of

PSA polymer in situ using proteins such as gelatin, which can additionally act as scaffold. Furthermore,

gelatin with amino acid functional group can act as buffer for pH shocks. For this purpose, a mixture of

agarose and gelatin scaffold was prepared and seeded with chondrocytes and PSA microparticles.

However, this strategy too did not help to reverse the toxic effects of PSA.

74

In fact, the natural pathway by which cells cope with lipotoxicity is through intracellular regulation of lipids

by serum albumin proteins.230, 231

That is, by maintaining a ratio of serum albumin/lipids of 0.7 - 2, cells to

combat adverse effects due to increase in fatty acid concentration.231

Thus, it can be hypothesized that

the presence of serum albumin can also combat the lipotoxicity, if any, of PSA degradation products. This

was tested by adding bovine-serum albumin both in the 3D constructs during casting and in culture

medium at 1 mg/mL concentration. It was found that BSA significantly reverses the cytotoxicity of these

anhydride polymers.

5.1.5 Delivery of TGF-β3 using PSA-microparticles in BSA-based system

The effect of uniform and sustained delivery of TGF-β3 using PSA-microparticles was tested

utilizing the above found mitigation strategy using BSA. For this purpose, the PSA microparticles were

fabricated with encapsulated TGF-β3. For a more consistent comparison with the control, only a fraction

of total TGF-β3 (one-fourth) needed for the first two weeks of culture period was delivered in situ using

encapsulated microparticles. The rest of TGF-β3 (three-fourth) was supplied externally through the

media. Given the encapsulation efficiency of ~50%, 1% w/v of microparticles is needed for delivering ¼ of

the TGF-β3 in situ of culture period of two weeks. The BSA protein was added both during casting

process and also in the growth medium throughout the culture period at a concentration of 1 mg/mL. Two

type of control scaffolds were cast, one without BSA and the other with BSA. The equilibrium Young’s

modulus and dynamic modulus at 0.1 are shown in Figure 5.5. It can be seen that the presence of BSA

has significantly mitigated the cytotoxicity of the PSA polymer and makes the PSA-microparticles effective

for short-term delivery. However, it is clear from presented results that biomedical applications using

polyanhydrides as a delivery device should not only incorporate BSA protein but also minimize

polyanhydride content by other means such as mixing polyanhydrides with slow degrading hydrophobic

polymers or reduce the polymer-to-bioactive molecules ratio.

75

Figure 5.5 Mechanical properties of control without BSA (BSA-), control with BSA (BSA+) and TGFβ3

microparticles-containing (ugel-TGF-β3-BSA+) constructs. (a) Equilibrium Young’s modulus, and dynamic

modulus at (b) 0.1 Hz. Using two way ANOVA analysis, there was no statistically significant difference

was found between Control BSA+ vs. ugel-TGF-β3-BSA+ for p< 0.05

To summarize the results,

(a) The PSA microparticles were found to be cytotoxic at a concentration of 20 mg/mL, but benign at

2 mg/mL.

(b) The mechanism of cytotoxicity is determined to be due to intracellular accumulation of PSA or its

degradation products and the resulting lipotoxicity. Changes in local pH due to polymer hydrolysis

was ruled out to have major impact on cellular viability

(c) Of the several mitigation strategies tested, bovine serum albumin treatment was found to offer the

best protection from adverse effects due to the polymer. Since it is known, in general, that serum

albumin is a natural regulator for intracellular lipids and prevents lipid overdose, the

a

b

76

chondroprotection of BSA against PSA polymer provides additional proof for its lipotoxicity apart

from its structural similarity to lipids

(d) The BSA-based mitigation strategy makes PSA microparticles effective in the short-term delivery

of TGF-β3 in articular cartilage engineering.

5.2 The effects of functionalization of polyacrylate nanoparticles on the cellular viability

and metabolism

In this section, the results of the effects of functionalized polyacrylate-based nanoparticulate system

(polyacrylic acid, polyacrylamide and polyphosphonate nanoparticles) embedded in a 3D hydrogel matrix

on the biochemical and biomechanical properties are presented. The results suggest that though these

nanoparticles are biocompatible, the negative surface charge on these nanoparticles correlates inversely

to the extracellular matrix synthesis and mechanical strength.

5.2.1 Sterilization tests:

Three nanoparticles, (1) polyacrylic acid (PAA), (2) polyacrylamide (PAM), polyacrylic acid-

phosphonate (PAAP), were kept under UV light for 20 minutes. These irradiated nanoparticles were

tested for sterility by suspending them in two separate media (water and tissue culture media with growth

factors). These were incubated at 37 oC in 5% CO2 environment for 4 days. The media remained clear

which indicated the absence of bacteria or other invasive microorganisms and success of sterilization

procedure. The next step was to culture chondrocytes in the presence of nanoparticles to find out its

biocompatibility.

5.2.2 Biocompatibility tests:

Three nanoparticles (PAA, PAM, PAAP) and a control (without nanoparticle) were used for these tests.

Chondrocytes obtained from calf-bovine source were plated on tissue culture flasks with or without the

nanoparticles at 10 μg/mL. The cell-to-nanoparticles concentration was initially maintained at 10,000

cells/μg. The cells were cultured for a week under these conditions viable cell density was counted. The

cell density for all the samples (control, PAA, PAM, PAAP) was found to be very similar (~ 10 million

77

cells). Thus it is concluded that nanoparticles are biocompatible with chondrocytes in a two dimensional

substrate at this concentration (10, 000 cells/μg). Following the biocompatibility studies, in vitro studies

using three-dimensional constructs of agarose hydrogels with embedded nanoparticles were conducted.

5.2.3 Three-dimensional tissue culture studies with acrylate nanoparticles

Agarose constructs embedded with three acrylate nanoparticles (PAA, PAM and PAAP) and a control

(without NPs) were used in this study. The juvenile bovine chondrocytes obtained from1st passage were

cast at 30 million cells/mL concentration. A total of four scaffolds, one control, and three nanoparticles-

containing (PAA, PAM, and PAAP) slabs were cast. The final w/v % of agarose and acrylate

nanoparticles were 2 and 0.2, respectively (see Section 2.2 for more details). Mechanical strength

(Equilibrium Young’s compressive modulus and dynamic modulus (at 0.1 Hz), cell viability (confocal

fluorescence microscopy), and biochemical analysis (glycosaminoglycan & DNA) were conducted on

constructs from day-0, day-14, day-28 and day-60 and are shown in Figure 5.6, 5.7 and 5.8. On day-3,

one of the experimental set (PAA) was lost due to infection and hence excluded from the study.

On the day-0, all the samples had the equilibrium compressive and dynamic moduli very similar to each

other (Figure 5.6). However, as time progressed, nanoparticles-containing scaffolds showed lower values

of the mechanical strength than the control. In the case of PAA-P nanoparticles, there was a very little

improvement in the mechanical strength up to day14 after which properties started degrading back to the

initial values. Though these moduli of the constructs containing PAM-nanoparticles were observed to

increase monotonically with time, the values were still lower than that of the control scaffolds.

The biochemical analysis provided interesting insights into the possible reason for lower values.

The glycosaminoglycan content of control on day-60 (Figure 5.7) was higher than both PAM and PAA-P-

containing constructs (similar to mechanical strength), whereas the DNA content per dry weight basis was

higher for PAM and PAA-P containing constructs when compared to control. This implies that, despite

having enough viable cells, the synthesis of extracellular matrix (GAG) was limited in the presence of the

acrylate nanoparticles. The confocal images (Figure 5.8) offer additional evidence for no difference in the

cellular viability between the control and nanoparticle scaffolds.

78

Figure 5.6 Mechanical properties of control, polyacrylamide containing (PAM) and polyacrylic acid-

Phosphonate (PAA-P) nanoparticles-containing constructs. (a) Equilibrium Young’s modulus, and (b)

dynamic modulus at 0.1 Hz.

These results indicate that the reason for the poor mechanical strength of the polymer

nanoparticles-containing scaffolds cannot be either cellular viability or pH shock. Even though the

nanoparticles are biocompatible, they may possess bioactivity detrimental to the synthesis of extracellular

matrix. It can be noted from the zeta potential of these nanoparticles at physiological pH ~7 (Figure 3.13)

that the higher the anionic charge of the polymer introduced externally into extracellular space the poorer

the matrix synthesis. In particular, anionic charge of the PAAP nanoparticles was greater than -45 mV

(Figure 3.13) and thus can disrupt the inter-cellular signaling process more effectively. It clear from these

results that chondrocytes has low tolerance for anionically charged polymers.

a

b

79

Figure 5.7 Total GAG & DNA content (in mg) of control, polyacrylamide containing (PAM) and polyacrylic

acid-Phosphonate (PAA-P) containing scaffolds

Figure 5.8 Live/Dead assay of scaffolds using fluorescence microscopy, for control, polyacrylamide

containing (PAM) and polyacrylic acid-Phosphonate (PAA-P) containing scaffolds on day0 and day33

0 10 20 30 40 50 60

0

10

20

30

40

50

60

70

80

90

GA

G/d

ry w

eig

ht

(w/w

)

Days

Control PAM PAAP

0 10 20 30 40 50 600.25

0.30

0.35

0.40

0.45

0.50

0.55

0.60

0.65

0.70

DN

A/d

ry w

eig

ht

w/w

Days

Control PAM PAAP

80

5.3 Conclusions:

In summary,

1) The PSA microparticles were found to be cytotoxic at a concentration of 20 mg/mL, but

benign at 2 mg/mL.

2) The mechanism of cytotoxicity is determined to be due to intracellular accumulation of PSA or

its degradation products and the resulting lipotoxicity. Changes in local pH due to polymer

hydrolysis was ruled out to have a major impact on cellular viability

3) Of the several mitigation strategies tested, bovine serum albumin treatment, was found to

offer the best protection from adverse effects due to the polymer. Since it is known, in

general, that serum albumin is a natural regulator for intracellular lipids lipids and prevents

lipid overdose, the chondroprotection of BSA against PSA provides additional proof for its

lipotoxicity

4) The BSA-based mitigation strategy makes PSA microparticles effective in the short-term

delivery of TGF-β3 in articular cartilage engineering.

5) The effect of functional groups present in poly-acrylate nanoparticles on cellular viability and

metabolism was studied. The polyacrylate nanoparticles are found to be biocompatibile .

6) Even though the nanoparticles are biocompatible, they possess a bioactivity which is

detrimental to the synthesis of extracellular matrix. This is confirmed by the poor mechanical

strength of the polymer nanoparticles-containing constructs when compared to control

constructs.

7) There is an inverse correlation between mechanical strength and the negative surface charge

of these nanoparticles. In particular, phosphonate nanoparticles containing constructs

exhibited the worst mechanical properties and is correlated to their high anionic charge (> -45

mV)

8) Disruption of the inter-cellular signaling process due to higher anionicity is proposed to be the

cause for poor matrix synthesis.

81

9) It clear from these results that chondrocytes have low tolerance for the extraneous anionically

charged polymers. Thus, further studies are needed to understand if the anionicity affects the

chondrogenic potential of stem cells too!

82

Chapter 6 Anti-inflammatory and anti-oxidant

bioactivity of ceria nanoparticles in articular

cartilage engineering

The interleukin-1α is one of the cytokines that cause inflammation at the damaged, diseased, or

implanted cartilage site, leading to tissue damage and cellular death. Here, the objective of this study is to

explore the chondroprotective potential of nanoceria as well as the applicability of nanoceria as an anti-

inflammatory agent against interleukin-1α.

This chapter is organized as follows: first section briefly details the experimental design and

parameters. The following two sections detail the effects of two different ways of adding nanoceria to a

3D-agarose based culture. The final section lists the main conclusion of this study.

6.1 Experimental design

The chondrocytes from articular cartilage were harvested from the knee joints of juvenile bovine.

The chondrocytes from the 1st passage were trypsinized and cast into a 2% wv agarose slab at 60M

cells/mL concentration with or without nanoceria at different concentrations.

Nanoceria were added into the system in two different modes:

In the first, nanoceria along with chondrocytes were seeded into the 3D hydrogel slab on day-0

but not added to the growth medium (Scheme 1a). The nanoceria was embedded at different

concentrations (10, 100, and 1000 μg/mL) during casting (day 0) whereas, the insult by IL-1α began on

day 16. The IL-1α was added to the growth medium at 10 ng/mL and replenished whenever the medium

was changed.

83

In the second, nanoceria was added only to the growth medium but not cast into the hydrogel

(Scheme 1b). The constructs were cultured for 21 days to obtain constructs with mechanical strength

suitable for IL-1α insults. On day 21, constructs were divided into 12 combinations of control and

experimental sets: combination of three IL-1α concentrations (0, 0.5 ng/mL, and 10ng/mL) and four

nanoceria concentration (0, 100, 1000, and 20000 μg/mL). In the experimental sets both nanoceria and

IL-1α were added simulataneously. One of the samples (nanoceria:1000 μg/mL and IL-1α: 10 ng/mL) was

lost due to infection.

Scheme 6.1 Two different ways of introducing nanoceria in a 3D-agarose based culture of chondrocytes.

(a) nanoceria is embedded in 3D agarose constructs during casting on day-0 (b) nanoceria was added in

to the growth medium along with interleukin from day-21

a

b

84

6.2 Effects of nanoceria embedded into the 3D-agarose scaffold on tissue growth as well

as combating inflammation due to interleukin-1α

On days 16 and 31, the equilibrium compressive Young’s modulus and dynamic moduli (at

0.1, 0.5, 1 Hz), as well as glycosaminoglycans (GAGs), collagen, and DNA content were measured.

Additionally, cellular viability was monitored using the Live/Dead cytotoxicity kit. The Young’s and

dynamic modulus at 0.1 Hzare reported in the Figure 6.1.

As can be seen from Figure 6.1, nanoceria at loadings of 100 and 1000 μg/mL improved

significantly the mechanical properties of the constructs in the absence of IL-1α. The direct contribution of

nanoceria to the mechanical strength can be easily ruled out because the increase in strength of

nanoceria-containing construts with time in relation to control was not constant.The nanoceria at 100 and

1000 μg/mL showed statistically significant difference in mechanical properties (Table 6.1). The presence

of nanoceria at 10ug/mL had minimal effect upon their properties and was not better than the control. The

biochemical analyses of these samples reveal an interesting pattern (Figure 6.2). Though DNA content

remianed the same across the sample sets, the GAG content and collagen content was higher for 1000

μg/mL nanoceria samples when compared to the rest. Thus, it can be concluded that nanoceria increases

the extracellular matrix production.

However, when IL-1α is present at a concentration of 10 ng/mL, nanoceria did not prevent

tissue degradation (Figure 6.1). As a result, the loss of mechanical strength was comparable to that of

control+IL-1α. The biochemical analysis revealed that both GAG and collagen content was reduced due

to the action of IL-1α across all the sample sets.

It should be noted, however, that the concentration of IL-1α (10 ng/mL) used in this study is very

high compared to its in vivo concentration (~500 times smaller than 10 ng/mL). As noted earlier, this

concentration was selected because it was found in previous studies to be appropriate for in vitro

experiments. Furthermore, the probablity of IL-1α coming into contact with cells may be higher than the

nanoceria getting to be internalized by the cells. The cellular viabilitiy on day 30 determined by Live/Dead

analysis remained the same across all the sets of samples (Figure 6.3). This fact provides additional

85

evidence (apart from DNA content (Figure6.2c)) that IL-1α was more active in degrading the matrix

degradation rather than being cytotoxic to chondrocytes .

Figure 6.1 Mechanical properties of control and nanoceria-containing constructs on day 16 and day 30.

Interleukin-1α was added from day 16 to 30. Nanoceria concentration in constructs are 10, 100 and 1000

μg/mL. Control constructs contain no nanoceria (a) Equilibrium Young’s modulus, (b) dynamic modulus at

0.1 Hz. The significance are *p<0.001 and **p<0.001

Figure 6.2 Biochemical composition of control and nanoceria-containing constructs on day 16 and day 30

as a fraction of wet weight. Interleukin-1α was added from day 16 to 30. Nanoceria concentrations in

constructs are 10, 100 and 1000 μg/mL. (a) Glycasoaminoglycan (GAG) (b) Collagen (c) DNA

86

Table 6.1 Two way ANOVA for all the mechanical properties (equilibrium Young’s modulus, dynamic

modulus at 0.1, 0.5, and 1 Hz) on day 30 of control and nanoceria-containing constructs without

interleukin-1α insults.

Figure 6.3 Live/dead analysis of control and nanoceria-containing constructs on day 30 using cytotoxicity

kit and confocal fluorescence microscopy. Green color indicates live and red color indicates dead cells.

Top row contains images of control, nanoceria-containing constructs at 10, 100, 1000 μg/mL

concentration without interleukin-1α insult. Bottom row: with interleukin-1α insult from day 16 to 30.

Samples ( only no IL-1α sets) P-valueⱡ

Control vs. nanoceria 10 μg/mL 0.5276

Control vs. nanoceria 100 μg/mL 0.0482

Control vs. nanoceria 1000 μg/mL 0.0004

nanoceria 10 μg/mL vs. nanoceria 100 μg/mL 0.0312

nanoceria 10 μg/mL vs. nanoceria 1000 μg/mL 0.0009

nanoceria 100 μg/mL vs. nanoceria 1000 μg/mL 0.1163

ⱡStatistically significant values (P<0.05) are shown in bold letters

87

6.3 Effects of nanoceria suspended in the growth medium on tissue growth as well as

combating inflammation due to interleukin-1α

In this study, instead of embedding nanoceria in the 3D-agarose scaffold, they were suspended

only in the growth medium. The rationale for the nanoceria addition to the growth medium was to

deactivate the inflammation by acting directly on IL-1α in the suspension medium itself. Several

conclusions can be drawn from the mechanical strength measured on day 34 (Figure 6.4). The anti-

inflammatory activity of nanoceria was observed at the highest concentration of both IL-1α and nanoceria

(Figure 6.4, nanoceria: 2000 μg/mL and IL-1α: 10 ng/mL). The Young’s modulus and dynamic moduli

were nearly twice higher than the control. The lower concentration of IL-1α (0.5 ng/mL) had negligible

effect on the constructs in vitro. In the absence of effects of IL-1α (0 ng/mL), there was a maximum

concentration (1000 μg/mL) of nanoceria at which mechanical properties were also maximum (see Figure

6.4). This again proves that nanoceria exerts postive effects on the mechnical and probably on

biochemical properties as well. The anti-inflammatory and chondroprotective actions of nanoceria were

also confirmed by GAG content measured on day 34 (Figure 6.4c). The GAG content of samples insulted

by 10 ng/mL of IL-1α in the presence of 20000 μg/mL nanoceria was higher than the control and other

concentrations. The maximum concentration of nanoceria for such positive effects can be between 1000

μg/mL and 20000 μg/mL. This maximum limit above which properties degrade can be explained as

follows.The reducing capacity of nanoceria, which is important for its ROS scavenging activity (Section

1.1.3), might become detrimental at high very particle concentrations. That is, a decrease in the oxygen

partial pressure of the growth medium due to high nanoceria concentration (20000 μg/mL) can be the

reason for drop in the mechanical strength in the absence of effects of IL-1α.

88

Figure 6.4 Mechanical and biochemical properties of constructs at day 34 as a function of nanoceria

concentration in the growth medium. (a) compressive Young’s modulus, (b) dynamic modulus at 0.1 Hz,

(c) glycosaminoglycan. Interleukin-1α is added from day 21 to 34 and its concentrations are given on the

top of the bars (0, 0.5, 10 ng/mL). For IL-1α 10ng/mL and 20000 μg/mL vs. 0 μg/mL *p<<0.00001

6.4 Conclusions

Nanoceria is found to combat inflammation due to interleukin-1α as well as enhances the mechanical

properties of cartilage through increased GAG and collagen production.

1) At concentrations up to 1000 μg/mL, nanoceria was biocompatible with chondrocytes.

2) The nanoceria has a beneficial impact on the biochemical as well as mechanical properties of cartilage

tissue grown in vitro.

3) An insult of the sample sets with interleukin-1α at 10 ng/mL for 12 days significantly degraded the

mechanical properties, GAG and collagen content.

4) The presence of nanoceria at 20000 μg/mL in the growth media was chondroprotective against the

inflammatory action of interleukin-1α and partially prevents matrix degradation.

89

Chapter 7 Dependence of the (bio)activity of

hematite NPs on their size, morphology, and

surface charge

The interactions of ferric (hydr)oxide nanoparticles (NPs) with chondrocytes and fatty acids are

reported. The impact of nanoparticle size and shape of ferric (hydr)oxide NPs on chondrocyte’s viability

and metabolism is reported in the first section. It was found that the mechanical strength of the NP-

containing engineered cartilage constructs increases with decreasing NP size.

In the second section, the interaction of these NPs with sodium dodecanoate (sodium laurate)

and the effects of size, morphology, and surface charge on this interaction is studied using results from

adsorption isotherms, zeta potential, contact angle, and FTIR measurements. Both the studies include

ferrihydrite (low crystalline hydrous ferric (hydr)oxide NPs of 2-5 nm size) for comparison as the end-

member the nanosize phase transformation of hematite. Contrary to the common correlation, the

adsorption capacity and affinity of hematite NPs decreases with decreasing particle size. The pattern with

which fatty acids fill the NP surface depends on the NP morphology. These results are rationalized taking

into account the nanosize induced changes in the catalytic and electron-accepting properties of hematite,

as well as the geometrical compatibility of the surface site arrangement and the attachment of carboxylate

groups.

7.1 Size, shape and surface-charge dependent bioactivity of hematite NPs on

chondrocytes

The goal of this study is to verify the hypothesis that the bioactivity of hematite NPs increases

with an increase in the NP size. This hypothesis is based on the previous results on the abiotic catalytic

activity of hematite NPs from our group.116

90

To test this hypothesis, three different sizes of hematite NPs (H9, H38 and H150) and 2-nm FH

were incorporated into the 3D-agarose hydrogel along with juvenile bovine chondrocytes at 30 million

cells/mL and cultured for 30 days. The NPs were added to the hydrogel in the weight range of 25-50

g/mL. A control hydrogel slab was cast at the same cell density but without NPs. The mechanical

properties and cellular viability were determined at the end of day-30 and day-71, respectively.

For comparison, the equilibrium Young’s modulus and the dynamic moduli are plotted (Figure

7.1) as a function of the loaded total surface area of the NPs, SAloaded. This was calculated as SAloaded =

SA m, where SA is the surface area per unit mass obtained using single point BET, m is the loaded

weight of the NPs per unit volume of the scaffold. As seen from Figure 7.1, the detrimental effect of the

hematite NPs on the mechanical properties of the grown tissue decreases with decreasing NP size, in

agreement with the initial hypothesis. The maximum and minimum effects on chondrocytes are for 150-

nm hematite and 2-nm FH, respectively. The following power law relationship between the decrease in

equilibrium Young’s modulus of the tissue constructs due to the presence of ferric (hydr)oxide

nanoparticles compared to control and the size of the nanoparticles is obtained (See Appendix C).

Normalized decrease in Young’s modulus = 0.048. (nanoparticle size)0.63

At the same time, no difference due to size in the cellular viability is observed among the NPs. In

fact, there is no difference in the viability between the control and the NP-exposed sets (Figure 7.2). This

implies that the NPs affect mostly the chondrocyte’s function of extracellular matrix production rather than

trigger cascades of cellular reactions leading to apoptosis. This finding matches with the previously

reported low or no toxicity (no DNA damage and no reactive oxygen species (ROS) production) of

hematite NPs toward cardiac, microvascular, endothelial and lung/bronchial epithelial cells.70, 115, 179, 181

Significant genotoxic effects have previously been observed for hematite only at loadings higher than 50

g/mL,178

that is, higher than the amount used in the present study.

To interpret this finding, our previous result of the nanosize effect on the oxidative catalytic

activity of hematite NPs may be recalled.116

Using batch kinetic measurements coupled with X-ray

photoelectron spectroscopy (XPS) as well as a new method that combines in situ FTIR spectroscopy and

ex situ XPS, the oxidative catalytic performance of hematite NPs was found to degrade with a decrease in

91

their size. It was demonstrated that these results can be accounted for by the nanosize-induced changes

in the electronic properties of ferric (hydr)oxides, along with the thermodynamic properties of the system,

given that the catalytic reaction proceeds through the electrochemical pathway. Briefly, as NP size

decreases, hematite NPs experience phase transformation to FH. These structural changes are

accompanied by an increase in the ionicity of the FeO bonds, as well as by opening of the

semiconductor band gap. The latter effect decreases the electron affinity of the NPs, and hence their

capacity to abstract electron from the reductant and to conduit it to a cathodic spot where O2 is reduced.

Therefore, the enhanced catalytic activity of larger NPs can lead to the oxidation of extracellular

biomacromolecules such as GAG and collagen, resulting in the lower strengths of the produced matrix.

A further non-invasive Raman spectroscopic study in future of the molecular composition and

structure of the extracellular matrix, as well as TEM studies of the NP localization in the scaffolds may

provide more detail on the effects observed.

92

Figure 7.1 Mechanical properties of control and feriic (hydr)oxide NPs-containing constructs on day

30.The concentration of the NPs is plotted as total surface area. (a) Equilibrium Young’s modulus,

dynamic modulus at (b) 1 Hz (c) 0.5 Hz (d) 0.1 Hz.

a

b

c

d

Control

Control

Control

93

Figure 7.2 Live/dead analysis of control and ferric (hydr)oxide NPs-containing constructs on day 72 using

cytotoxicity kit and confocal fluorescence microscopy. Green color indicates live cells, red color indicates

dead cells. Top row contains images of control, FH, and H9-containing constructs at 25 μg/mL

concentration and bottom row that of H38 and H150 containing-constructs at two different NP

concentrations (25, 50 μg/mL)

7.2 Size, shape and surface-charge dependent interaction of hematite NPs with sodium

laurate (dodecanoate)

This section is organized as follows: First, variation in the dispersion and adsorption properties of

laurate-coated ferric (hydr)oxide NPs is discussed including hydrophobicity and surface charge. Next,

spectroscopic data on the adsorption configuration of the fatty acid, as well as on the adsorption

mechanism and the chain packing order are presented.

7.2.1 Dispersion and adsorption properties of hematite NPs in the presence of sodium

laurate.

I found that at pH 7.2 and concentrations of sodium laurate (SL) lower than CMC (~1mM at pH

7.2),161

all the sizes ferric (hydr)oxide NPs flocculate and settle particles in 5mM NaCl aqueous solution.

In contrast, at SL concentrations higher than the precipitation limit (~ 2.5 mM at pH 7.2), H150 and H38

particles are well dispersed for at least two weeks, FH aggregates and settles, and H9 is partially

94

H150 FH H9 H38

dispersed (Figure 7.3). As compared to FH, hematite NPs are better dispersed: An increase in size of the

ferric (hydr)oxide NPs is accompanied by a regular improvement in their dispersion properties.

Figure 7.3 Suspensions of FH and hematite NPs at initial SL concentrations of 3 mM and ionic strength

of 5 mM NaCl.

To understand this difference in the NP dispersibility, the adsorption isotherms were measured for

the largest and smallest NPs (H150 and FH, respectively) which represent the well crystalline hematite

and FH end-members of the phase transformation sequence of hematite NPs in water.222

As depicted in

Figure 7.4, at concentrations lower than ~3.5 mM, 150-nm hematite has a higher sorption capacity.

However, at higher concentrations, FH abstracts the surfactant much more efficiently. The corresponding

part of the FH isotherm presents a vertical branch shifted to a lower equilibrium (residual) concentration.

Although the mechanism for such abstraction is not clear, it is hypothesized that FH coagulates with

surfactant precipitates, thereby occluding surfactant above its precipitation limit. Additionally, the effect

can be related to a higher solubility of FH as compared to coarser-sized hematite NPs,106

resulting in

precipitation of ferric laurate. Importantly, no such phenomenon is observed for H150 at the initial laurate

concentration up to ~7 mM. Instead, the adsorption isotherm consists of two steps and two plateaus. The

first plateau, which starts at ~0.5 mM, signifies the formation of the first monolayer. This monolayer has a

relatively high density of 4.3 0.1 molecules/nm2

and imparts superhydrophobicity to H150 (Figure 7.5).

This surface density is slightly below 4.8 molecules/nm2 typical for the tilted condensed phase of fatty

acids at the water-air interface.232

Hence, these particles are almost fully wrapped by the SAMs of laurate.

As the surfactant concentration is further increased, the second monolayer is assembled with the

95

surfactant head groups pointing out into the solution, as follows from a decrease in the contact angle

(Figure 7.5). The second plateau is caused by the pinning of the chemical potential of the surface by the

surfactant precipitation at 2.5 mM. In this concentration range, a second monolayer is already assembled,

as evidenced by the doubling of surface density and the evident onset of strong dispersion (Figure 7.6).

On this basis, it is concluded that the propensity to be more fully wrapped by a bilayer brings about a

better dispersibility of the oxide NPs.

To study the effects of NP size and morphology on the adsorption properties, the

adsorption isotherms were measured on all the four sizes at pH 7.2 and an initial SL concentration up to 1

mM (CMC). The last data point of the all the isotherms in Figure 7.7a corresponds to 1mM initial SL

concentration. As seen from Figure 7.7a, both the adsorption capacity and affinity of hematite NPs

decrease with decrease in size. The laurate surface densities at CMC are 4.2, 3.3, 2.6, and 2.1

molecules/nm2 on H150, H38, H9, and FH, respectively. A power law fit is obtained between the

monolayer coverage and the size of particles. Monolayer coverage values are normalized with maximum

coverage value of H150 (4.2 molecules/nm2), see appendix C for details: Normalized monolayer coverage

= 0.4 . (size)0.19

Figure 7.4 Adsorption isotherms of sodium laurate on FH (wine) and H150 (blue) NPs at pH ~ 7.2.

96

Figure 7.5 Contact angles of SL-coated H150 nanoparticulate film. Formal monolayer coverage of 4.3

molecules/nm2 was estimated from the adsorption isotherm. Hydrophobicity of 150-nm hematite

increases as the surface coverage of laurate increases up to 1 ML. At this point, the particulate film is

superhydrophobic (water droplets are repelled from the surface). With a further increase of surface

coverage, hydrophobicity gradually decreased, indicating that laurate is adsorbed with the head group

pointing out into the solution.

Figure 7.6 Effect of the initial concentration of laurate (written at the bottom as well as on the vials) on the

dispersion of H150 particles at pH 7.2

The electrophoretic data (Figure 7.7b) indicates that at SL concentration of at least up to 1mM

H150 has a higher negative zeta-potential than the other extreme in size, FH. Unfortunately, the visual

particle-tracking method in the electrophoretic technique failed for the H150 suspensions with initial

laurate concentrations above 1 mM, where the particles become dispersed and strongly scatter light.

Nevertheless, the available data confirm that the SAM on H150 has higher density, with the electrostatic

0mM 0.25 mM 0.75 mM 1 mM 3 mM 5 mM

0.25 mM

97

mechanism of stabilization of the NP suspensions by laurate. In contrast, a lower (by a factor of 2)

adsorption capacity of FH gives rise to their less negative charge which may have caused their

aggregation.

Figure 7.7 (a) Adsorption isotherms of sodium laurate on FH (wine), H9 (orange), H38 (red) and H150

(blue) NPs at pH ~ 7.2. The last data point corresponds to 1 mM initial laurate concentration (CMC of

laurate) (b) Zeta potential of H150 and FH as a function initial laurate concentration in 5mM NaCl

aqueous solution. Both electrophoretic (red and blue curves) and electroacoustic (green curve) results

are shown. Connecting lines are drawn for reading ease.

7.2.2 Speciation of carboxylic groups in the first monolayer.

To elucidate the effects of the size, morphology, and surface charge of the ferric (hydr)oxide NPs on the

speciation of the carboxylic groups and the packing order of adsorbed fatty acids, in situ FTIR-ATR

spectra were measured on the NPs interacted with a 1 mM solution of SL at pH 10.0 (Figure 7.8) and

7.15 (Figure 7.9) for 25 min. It was observed that at pH 10, FH does not adsorb laurate at initial SL

concentration up to 1 mM, while H9up to 0.5 mM. Three main adsorption forms are identified from the

fingerprint region of asymmetric (asCO2) and symmetric (sCO2

) stretching modes of the carboxylate

group (Scheme 7.1 and Figure 7.8 & 7.9). These are (1) Inner-sphere monodentate mononuclear

complex (ISMM): specific adsorption of one of the oxygen of laurate molecule with one of surface ferric

ion, whereas the second oxygen H-bonds with other surface hydroxyl or water. (2) Outer-sphere surface-

hydration complex (OS): non-specific adsorption of laurate on surface ferric ions mediated by a hydration

layer (3) Protonated outer-sphere surface-hydration complex (OS-H): non-specific adsorption of lauric

a b

98

acid on surface ferric ions mediated by a hydration layer. The details of such identifications using

combination DFT and spectroscopic data are provided in Chapter 9.

Figure 7.8 in situ FTIR-HATR spectra of FH and hematite NPs in contact with a 10–3

M solution of sodium

laurate at pH 10.00 ± 0.05 with no added electrolyte in spectral region of (a) COO and (b) CH

vibrational modes: (olive) H150, (red) H38, (blue) H9. Conditioning time is 2530 min.

Scheme 7.1 Cluster models of surface binding of fatty acid: (a) inner sphere monodentate mononuclear

(ISMM) carboxylate complex with H-bonded second oxygen (b) outer sphere (OS) surface-hydration

shared ion pair, and (c) OS fatty acid. Inner-sphere monodentate mononuclear complex: specific

adsorption of one of the oxygen of laurate molecule with one of surface ferric ion, whereas the second

oxygen H-bonds with other surface hydroxyl or water. Outer-sphere surface-hydration complex: non-

specific adsorption of laurate (or lauric acid) on surface ferric ions mediated by a hydration layer.

ISMM OS surface

hydration-shared

b a c

OS surface hydration-

shared protonated

99

Figure 7.9 in situ FTIR-HATR spectra of FH and hematite NPs in contact with a 10–3

M solution of sodium

laurate at pH 7.15 ± 0.05 with no added electrolyte in spectral region of (a) COO and (b) CH

vibrational modes: (olive) H150, (red) H38, (blue) H9, (magenta) FH. Conditioning time is 25 min.

Intensity of the spectra in the 12001800-cm–1

region is normalized by the intensity of the sCH2 band

a. Inner-sphere monodentate mononuclear complex (ISMM):

The main bands which dominate at ~1540 and ~1410 cm–1

are assignable, respectively, to the

asymmetric (asCO2) and symmetric (sCO2

) stretching modes of the carboxylate group of an ISMM

complex (Scheme 7.1).233

b. Outer-sphere hydration shared complex (OS):

The second asCO2/sCO2

pair at 1530/1425 cm–1

is attributed to an OS hydration-shared

complex.233

The positions and relative intensities of these two pairs of bands vary among the NPs.

Specifically, H38 and H150 are characterized by the lowest and highest positions of the sCO2 bands. It

is known from previous studies234, 23518,59

when there is an increase in the positive charge of the adsorbent

100

surface, the sCO2–

band of chemisorbed carboxylates decreases (Scheme 7.2). Indeed, as seen from

Figure 7.10, the red shift of the sCO2–

frequency can be correlated with an increase in PZC and hence

the surface charge of the NPs.

c. Protonated outer-sphere hydration shared complex (OS-H):

Another feature of the FTIR spectra measured at pH 7.15 (Figure 7.9a), which is missing at pH

10.0 (Figure 7.8a), is the C=O band at 1707 cm–1

of the OS complexes of lauric acid.233

For comparison

purposes, the spectra are normalized by peak intensity of the sCH2 band at ~2850 cm

-1. As seen from

Figure 7.7a, the C=O band is hardly distinguishable for FH and is maximal for H150. For H9 and H38, its

intensity is similar and intermediate (H150 > H38 ≈ H9 >> FH).

The Figure 7.9a also demonstrates a variation in the OS/ISMM intensity ratio for the sCO2

band

among the NPs: This ratio decreases in the following order FH H150 > H9 H38. As follows from the

effect of negative polarization on the intensity ratio (vide infra), it can be argued that the OS/ISMM ratio is

controlled by the adsorption capacity of the NPs towards the ISMM complexes (i.e., the concentration of

suitable strong adsorption sites).

Scheme 7.2 Stretching of sCO2– dipole of adsorbed carboxylate by positive charge of the NP surface.

Figure 7.10 Correlation between frequency of sCH2 band of ISMM laurate in the spectra shown in

Figure 7.9a and PZC of FH and hematite NPs measured in open air (Figure 7.7b)

6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5

1402

1404

1406

1408

1410

1412

H38

H9

FH

sC

OO

-

PZC, air

H150

101

7.2.3 Packing order of hydrocarbon chains.

The conformational order of the aliphatic chains is an critical parameter to be studied, as it is

directly related to the wetting properties of the SAM. The symmetric stretching vibrations of methylene

groups sCH2 at ~2850 cm1 is used for

assessing this information: This band shifts to lower

wavenumbers and narrows with increasing packing order of an ensemble of long-chain molecules.149

Then, as seen from Figures 7.8b and 7.9b, the conformational order of the aliphatic chains in the adlayers

formed at both pH 7.15 and pH 10.0, respectively, increases in the following order H38 < FH < H9 <

H150. This order correlates neither with the adsorption density of the surfactant (Figure 7.7), nor with the

composition of the ML (the fractions of the OS complexes and the lauric acid) (Figure 7.9a).

To get more details, the patterns of the monolayer filling with time were analyzed. As seen from

Figure 7.11a,c,d, except for the initial stage (the formation of ~0.030.1 ML), the sCH2 band does not

change its position when the surface coverage of laurate increases up to 1 ML on FH, H9, and H150,

suggesting the cluster-like pattern of the monolayer growth. In contrast, the sCH2 band continuously

shifts downwards in the case of H38 (Figure 7.11b), which is typical of the isotropic filling. Hence, H38

demonstrates distinct monolayer filling, which is reasonable to ascribe to the distinct rhombohedral

morphology of these NPs (Figure 4.1 and 4.2), as well as to their high capacity to form chemical bonds

(ISMM complexes) with carboxylic groups (Figure 7.9a). Due to the predominance of the ISMM

adsorption, it can be suggested that the headgroup-surface interaction is possibly stronger than the

lateral hydrophobic, dispersion, and H-bonding interaction between the assembled surfactant molecules.

The morphological homogeneity of H38 (which is enclosed by rhombohedral facets of similar surface

energy106

) is also expected to contribute to this phenomenon.

Analysis of the main crystallographic surfaces of hematite and typical chain-chain packing

distance of laurate confirms this interpretation:

In adsorbed monolayers, the chains can pack in four different subcells: orthorhombic, triclinic,

monoclinic or disordered pseudohexagonal subcells.236,149

In all these packing motifs listed, the chains

are arranged such that the chain-chain distance of 4.75.5 Å.237, 238

Given that only tilt angles between 0

and 30 have experimentally been observed for adsorbed monolayers,149, 237

the surface projected

distances of well packed chains are 4.76.3 Å.

102

As seen from Figure 7.12c,d, ferric cations are arranged on the (104) and (012) rhombohedral

hematite planes in a parallelogramic fashion, with the FeFe distances of 3.71 in the rows of ferric

octahedra and 5.04 Å between the rows. Because the distance of 3.71 Å is much shorter than the low

limit of ~4.7 Å of the FeFe range which supports ordering, the adsorption sites in the rows can be

occupied only partially, with the minimum FeFe distances of 7.42 Å. This distance of 7.42 Å is

incompatible with a conformational ordering of the fatty acids if they are chemisorbed on these sites.

Thus, the geometrical mismatch and the dominance of the chemisorbed complexes on H38 preclude a

dense packing of the SAM.

In contrast, cations are (pseudo)hexagonally distributed on the {101} and iron-depleted {001}

facets (Figure 7.12a,b) which enclose hematite platelets.239

Moreover the FeFe distances of 5.44 and

5.04 Å on these surfaces are within the region of 4.76.3 Å suitable for the chain ordering.

7.2.4 Controls of NP dispersion by fatty acids. Formation of a bilayer.

Although the stabilization of NPs in suspensions by fatty acid bilayers is a known phenomenon,240

the current study is the first to observe both the formation of such bilayers on metal (hydr)oxide NPs in

water and the dependence of this phenomenon on NP size. Namely, the fatty acids are found to better

disperse larger ferric (hydr)oxide NPs, independent of their morphology. The effect is related to a higher

adsorption capacity of the larger NPs: A denser formal monolayer attaches more molecules to the second

layer, because the bilayer formation by a surfactant at the solid-water interface is governed by the

hydrophobic and the London-van der Waals dispersion interactions.125

The formation of a bilayer imparts

the highest hydrophilicity and negative charge to the particles, which corresponds to the best conditions

of the electrostatic stabilization of the suspension.

Interestingly, the smallest particles, FH, have displayed strong ability to occlude laurate above its

precipitation limit. The mechanism of this effect is still unclear: It can tentatively be related to the

enhanced tendency of FH to coagulate with the surfactant precipitates due to the partial coating of FH by

a bilayer. Another reason can be a higher solubility of FH as compared to hematite.

103

Figure 7.11 Effect of surface coverage on the position of the sCH2 band of laurate adsorbed on (a)

H150, (b) H38, (c) H9 hematite, and (d) FH from 1 mM solution at pH 10.0 and 7.15. Surface coverage,

evaluated using the sCH2 peak intensity, increases from (red) ~ 0.030.1 to (magenta) 1.0 formal ML

which is assumed to be formed at the maximum interaction time. It is increased with increasing the

conditioning time which is indicated in the graph. The band intensity is normalized to the maximal value in

the series. Color codes for H150, H38, and H9 are the same. FH was conditioned only at pH 7.15

because it does not adsorb 1 mM laurate at pH 10.

2880 2870 2860 2850 2840 2830 28202880 2870 2860 2850 2840 2830 2820

pH10 5'

pH10 47'

pH10 65'

pH7.15 5'

pH 7.15 30'

pH7.15 50'

2880 2870 2860 2850 2840 2830 2820

2880 2870 2860 2850 2840 2830 2820

2'

8'

25'

35'

130'

2880 2870 2860 2850 2840 2830 2820

2'

8'

25'

35'

130'

2880 2870 2860 2850 2840 2830 2820

pH10 5'

pH10 47'

pH10 65'

pH7.15 5'

pH 7.15 30'

pH7.15 50'

H9

a

b

c

H150

H38 FH

d

Wavenumber, cm-1

Wavenumber, cm-1

Wavenumber, cm-1

Wavenumber, cm-1

Ab

so

rba

nce

, A

U

Ab

so

rba

nce

, A

U

Ab

so

rba

nce

, A

U

Ab

so

rba

nce

, A

U

surf

ace c

overa

ge

incre

ases up t

o 1

ML

104

Figure 7.12 Arrangement of surface cations on (001) basal, (101) hexagonal as well as on (012) and

(104) rhombohedral hematite surfaces.

7.2.5 Effect of negative polarization on the structure of the first monolayer.

To verify if the morphology-driven geometrical incompatibility and the high propensity to

chemisorb carboxylates are the main obstacles in the chain self-assembly, the effect of negative

polarization on the speciation and packing of the first monolayer of laurate on these NPs was studied. As

105

shown in a recent study,151

the surface charge of the metal oxide NPs can be changed by depositing

them on substrates with different work functions. Specifically, deposition on Ge renders negatively charge

to hematite NPs. As seen from Figure 7.13a, the negative polarization of H38 shifts the adsorbate

equilibrium from the ISMM to OS complexes and from laurate to lauric acid within the OS complexes. This

conclusion is based on an increase in the relative intensity of the OS sCO2– band at ~1420 cm1

and the

C=O band at 1707 cm–1

, respectively. The blue shift of both the OS and ISMM sCO2– bands reflects the

negative polarization of the NPs on Ge. Importantly, the change in the speciation of the carboxylate

groups is accompanied by an increase in the chain packing in the monolayer which is manifested by a

pronounced red shift and narrowing of the sCH2 band (Figure 7.13b). As a result, the chains become

even more ordered than in the SAM formed on H150 when the latter is deposited on ZnSe (Figure 7.14).

Hence, partial suppression of chemisorption can bring about a better chain ordering and higher

hydrophobicity of the fatty-acid capped NPs.

Figure 7.13 Effect of the substrate on in situ FTIR-HATR spectra of H38 hematite NPs in contact with a

10–3

M solution of sodium laurate at pH 7.15 ± 0.05 with no added electrolyte in spectral region of (a)

COO- and (b) CH vibrational modes: NPs were deposited on a (green) Ge and (red) ZnSe IRE.

Conditioning time is 25 min.

106

Figure 7.14 Comparison of in situ FTIR-HATR spectra in the CH region of (red) H38 deposited on Ge

and (violet) H150 deposited on ZnSe. The NPs were conditioned similarly with a 10–3

M solution of

sodium laurate at pH 7.15 ± 0.05 with no added electrolyte for 2530 min.

7.2.6 Speciation of adsorbed fatty acid. Effects of the oxide acidity.

Although the surface density and packing order of laurate is not affected by the intrinsic proton

surface charge of NPs, the speciation of laurate is affected. Namely, a higher OS/ISMM surface ratio of

laurate on H150 and FH which have lower proton surface charge (Figure 7.9a) is observed. Although a

further theoretical and thermodynamic study is required to understand this effect, it is suggested that the

chemisorbed complex is destabilized by negative proton surface charge due to the hindered ligand

exchange. In fact, a ligand exchange with surface hydroxyls has a higher energy penalty than with more

weakly chemisorbed water.241

The same effect is observed when the NPs are negatively polarized by Ge

(Figure 7.13 and 7.14). Therefore, both proton and electron charges control the OS/ISMM ratio.

Interestingly, in contrast to the OS/ISMM surface ratio, the fraction of the protonated OS

carboxylates does not depend on the macroscopic surface charge but on its size: It decreases as NP size

decreases (H150 > H38 H9 > FH). To rationalize this finding, It should be recalled that at pH 7.2 and 1

mM concentration, the surfactant exists in the bulk solution as acid-containing associates (either

aggregates165

or precipitates162

). The negligible amount of the OS lauric acid on FH suggests that the acid

-0.000

0.002

0.004

0.006

0.008

0.010

0.012

0.014

0.016

0.018

Ab

sorb

an

ce

2800 2850 2900 2950 3000

Wavenumbers (cm-1)

107

molecules are deprotonated upon physisorption on these NPs (Scheme 3), suggesting a higher basicity

of the negatively-charged oxygen sites on FH as compared to the larger NPs. On the other hand, the

surface fraction of the physisorbed acid increases with increasing surface coverage.233

Therefore, the net

result is a sum of the contribution of the intrinsic acid-base properties and their adsorption capacity of the

NPs.

Scheme 7.3 Deprotonation of lauric acid by basic surface sites upon adsorption.

7.2.7 Conditions for SAM formation. Effects of the NP size, morphology, and surface charge.

It is observed that only the largest (150 nm), well-crystalline particles have surface energy high

enough to form self-assembled monolayers (SAMs) upon interaction with fatty acids in water. Within the

set of the NPs used in this study, a decrease in nanosize causes a decrease in both the adsorption

affinity and capacity of the NPs (H150 > H38 > H9 > FH), suggesting that the NP morphology (H38) and

porosity (H9) are secondary for this particular system.

At a first glance, this finding is at odds with the conventional belief that smaller sizes having more

defects and cus cations should have larger adsorption affinity and capacity. At the same time, this size-

dependence agrees with the observed decrease in the surface energy of hematite with a decrease in NP

size242

which can be related to the nanosize-driven phase transformation of hematite into metastable

FH.222

The phase transformation of hematite NPs is accompanied by the lattice expansion50,51

and a

decrease in the average electron affinity.109

As a result, the surface area per adsorption site increases,

while their average affinity towards the HOMO electron of the carboxylate group decreases in the row

H150 > H38 > H9 > FH. The positions of H9 and FH in the adsorption trend are consistent with the

nanosize-induced phase transformation of hematite NPs into FH and the presence of FH-like defects on

108

the hematite NPs smaller than ~40nm.1 In line with this finding is the observed degradation of the

oxidative catalytic properties of the NPs in the same order.110

The monotonic size-dependence of the adsorption capacity towards the surfactant suggests that

this property does not depend on the macroscopic intrinsic proton surface charge (originated from the NP

acid-base properties and measured using a macroscopic method): In fact, FH and H150 are the least

positively charged NPs at pH ~7.2 (Figure 7.7b), but display the lowest and highest adsorption capacity

for fatty acids, respectively. This is contrary to the adsorption of small ions,243

short-chain

monocarboxylates,244

and aminoacids,245

the surface density of which strictly follows the variation in the

surface charge of the NPs. The lack of the effect in the case of fatty acids can be explained by the major

role of non-electrostatic (hydrophobic and H-bonding) interactions in determining the adsorption.233

At the

same time, surface charge affects the binding mode of the carboxylic group (see below).

However, the above size dependency does not hold for the packing order of the hydrocarbon

chains in the formal monolayer: The packing order decreases as H150 > H9 > FH > H38. The worst

packing order on H38 is somewhat surprising given the relatively high adsorption density on these NPs.

This effect is explained by the dominance of chemisorbed fatty acid on H38 (Figure 7.9a) and the

geometric incompatibility of the rhombohedral (104) facets of these NPs with self-assembly. Fortunately,

the morphological obstacle can be removed by negatively charging the NPs: The resulting upward shift of

the Fermi level of the NPs decreases their capacity to accept electrons151

and decreases the surface

population of the ISMM laurate. The increased spatial gap between remaining chemisorbed fatty acids is

now filled by OS complexes which have more freedom to rearrange laterally and promote better packing

and the formation of SAMs, in agreement with the general trend observed for alkanoate SAMs on

extended alumina surfaces.246-248

In the case of weak adsorption affinity (FH), although both the ISMM

and OS complexes are formed, their low adsorption density precludes their ordering. It is then expected

that the positive polarization of FH can SAMs can be formed on such NPs upon.

109

7.3 Conclusions

The interaction of ferric (hydr)oxide NPs with chondrocyte can be summarized as follows,,

1) The mechanical properties of the constructs embedded with these NPs degrade with increasing

NP size as follows: Control > FH (2-nm) > H9 (9-nm) > H38 (38-nm) >> H150 (150-nm).

2) The cellular viability remains similar between all these sample sets.

3) It is suggested that the size effect is driven by the catalytic activity of NPs which is responsible for

the reactive oxygen species (ROS) generation.

4) These results together with beneficial effects due to nanoceria imply that the bioactivity is directly

related to the redox potential and the catalytic redox mechanism (including either the deactivation

or generation of reactive oxygen species) of the metal oxide NPs toward redox active constituents

at the biointerface.

In the study of ferric (hydr)oxide NPs-laurate interactions, the first experimental evidence is obtained for

the dependence of the surface density, speciation, and packing order of adsorbed fatty acids on the

morphology, size (and, hence, crystallinity109

), and surface charge/electric polarization of metal

(hydr)oxide NPs. The conditions under which fatty acids form SAMs and bilayers on such NPs in water

are distinguished and the electric polarization of the NPs is demonstrated to be a powerful tool for

manipulating the interfacial properties of the NPs.

Specifically,

1) Laurate is adsorbed on hematite mainly as a mixture of chemisorbed (ISMM) and physisorbed (OS

hydration-shared) complexes, both at pH > PZC and pH < PZC;

2) The physisorbed laurate is partially protonated by a coadsorbed hydronium cation as either surface

coverage increases or solution pH decreases;

3) Independent of morphology, an increase in nanosize improves the adsorption capacity and affinity of

hematite NPs, as well as increases the fraction of the protonated physisorbed laurate, in agreement with

the nanosize-induced changes in the structural and electronic properties of the NPs.

4) In the case of hematite NPs with a similar hexagonal habit, an increase in nanosize improves the

packing order of laurate.

110

5) Independent of morphology, the dispersibility of the NPs by fatty acids in water degrades as NP size

decreases. This effect is attributed to the decrease in the adsorption capacity which ends up in the

formation of a loosely packed bilayer.

6) Morphology of the NPs is critical for the self-assembly if (i) the majority of the molecules are

chemisorbed (ISMM) and (ii) the morphology-determined surface arrangement of the adsorption sites is

geometrically incompatible with a surfactant ordering.

7) The surface ratio of the physisorbed/chemisorbed surfactant depends on the surface charge of NPs.

8) Negative polarization suppresses chemisorption. This effect can be used to remove the geometrical

constraints for self-assembly.

9) The dominance of the chemisorbed complexes results in the isotropic surface filling by fatty acids,

while an increased fraction of OS complexes favors the cluster-like pattern.

111

Chapter 8 A new approach to DFT modeling of the

adsorbed species at the metal (hydr)oxide NPs-

water interfaces

The first section of this chapter provides details of the computational quantum chemical method-

Density Functional Theory (DFT) used in this study. The second section details the new approach to

modeling the adsorption properties of metal oxides in aqueous media using DFT simulations and FTIR

spectroscopy. The third section presents the results obtained by applying this approach for adsorption of

fatty acids on hematite NPs and final section summarizes the results.

8.1 A combined DFT-FTIR approach to resolve the adsorption forms at the oxide-water

interface

The density functional theory (DFT) has become one of the most applied computational

techniques among the several available quantum chemical simulation methods for understanding

molecular processes in solids, fluids and interfaces. This is mainly due to its relatively low computational

cost, ease of implementation and conceptual simplicity which allows further innovations.

In quantum mechanics, all the properties of a system can be, in principle, obtained a-priori, if the

wavefunction (ψ) of the system is known. The ψ can be calculated using Schrodinger equation, under

time-independent and non-relativistic conditions,

[∑ (

( ))

∑ ( )] (8.1)

Where, m and N are mass and number of electrons, is reduced Planck’s constant, ν and U are

potential functions and E is energy of the system.

112

The first term of the left hand side represents kinetic energy term, second and third terms are

interaction energies between nuclei-electron and electron-electron, respectively. Several approximate

methods such as Born-Oppenheimer, Hatree-Fock were developed. All these electron wavefunction-

based methods have the major limitation of being computationally intensive and their application in

surface and interfacial problems are still relatively impractical.

At the same time, a useful theorem derived by Hohenberg and Kohn249

stated that all the ground

state properties of a system are a function of charge density functional, ρ. This implies that Schrodinger’s

equation can be solved for charge density (ρ) instead of multiple electronic wavefunctions, Ψel. The

Schrodinger equation (8.1) can be written in terms of ρ as,

( ) ( ) ( ) ( ) (8.2)

The electron-electron interaction energy, U(ρ), can be further divided into single particle

electrostatic term, UH, and many-body exchange-correlation energy (Exc). Furthermore, the electron

kinetic energy term can also be divided into non-interacting single particle term, Tint (ρ) and a correlated

interacting term which can be added to Exc term above.

Thus, the Schrodinger equation is simplified as follows,

( ) ( ) ( ) ( ) ( ) (8.3)

The non-interacting kinetic energy Tnon-int(φ) is a function of molecular orbital functions, φ. Using

Thomas-Fermi approximation, Tnon-int(φ) can be expressed as,

( )

∑ ∫

( ) ( ) (8.4)

The single electrostatic term between electrons is given

( )

∫ ∫ ( ) ( )

| | (8.5)

113

There are no exact functional forms for exchange-correlation term, Exc. Several approximate

functionals have been developed either based on local density approximation (LDA) or more generalized

gradient approximation (GGA).250

The LDA approach assumes a uniform Exc in the smallest volume

considered in the DFT scheme. This approach is inadequate where there are strong electron correlations

(for eq. transition metal ions). Thus, the GGA approach is superior since it includes the rate of spatial

variation of charge density (ρ) in the Exc functionals.250

[ ] ∫ ( ( )) ( )) (8.6)

However, the equation (8.3) still cannot be minimized for energy directly with respect to charge

density, ρ, alone because the kinetic energy term, Tint, is dependent on the molecular orbital functions (φ).

However, the molecular orbitals themselves are a function of charge density functional, φ(ρ). To

circumvent this problem, Kohn and Sham251

suggested solving the Schrodinger equation for an auxiliary

non-interacting system to obtain molecular orbitals (φ). These orbitals were further used to calculate

charge density function (ρ) and solve equation (8.3) using a self-consistent field procedure.250

The DFT

simulations, using above procedure, were performed using DMol code available in Material Studio

program from Accelrys Inc.

8.2 A combined DFT-FTIR approach to resolve the adsorption forms at the oxide-water

interface

As mentioned in Section 1.6.1, to extract the structures of adsorbed carboxylate groups from their

in situ FTIR spectra, the DFT approach should be further developed to circumvent the inherent limitations

of the direct one-to-one correspondence between the theoretical and experimental spectra.

114

One way to solve this problem is to search for the adsorption structure that provides the best

correspondence in the trends of the theoretical and experimental spectra upon variations of the solution

pH and protonation-deprotonation of the carboxylate group.241

Specifically, the following algorithm of the spectrum interpretation can be suggested for fatty acids

adsorption.

1) The number of adsorbed species is determined from the fingerprint region of the FTIR

spectrum by the correlation analysis of the spectrum dependence on the surfactant concentration,

adsorption time, pH, ionic strength, and the electric polarization of the metal oxide.

2) The adsorbed species can be divided into weakly adsorbed or strongly adsorbed through the

effects of ionic strength variation and polarization of the substrate on the spectra.

3) The protonation state of these species can be delineated by observing spectral change with

change in pH, and the absence/presence of the typical band of carboxylic groups at 1700-1730

cm-1

.

4) The impact of H/D exchange on the spectra allows to further classify these species into outer-

sphere (OS) & monodentate or bidentate due to different solvating properties of light and

heavy water.

In parallel to these experimental methods, DFT simulations can be used to verify independently

any of the proposed structures and its dependence of the above mentioned parameters.

a. Firstly, using DFT simulations several adsorption forms can be a priori postulated and

tested to ascertain if their formation is energetically favorable.

b. The impact of ionic strength can be modeled in DFT simulations by introducing counter

cation near the double layer and verify if the calculated spectra of the proposed structure

show similar dependencies.

c. Similarly, pH can be modeled in DFT by providing a charge to the surface through

changing the ratio of surface hydroxyls to the water (attached to the surface cations)

115

d. The impact of H/D exchange on the FTIR spectra can also be compared by replacing the

hydrogen atoms at the interface with deuterium atoms for the hypothesized adsorption

forms in DFT simulation.

e. Finally the site specificity data from the etching and TEM/HRTEM experiments can be

understood by simple geometric modeling cation-cation distances of the specific facets of

the substrates and the corresponding parking distance for adsorbates

Figure 8.1 A general predictive paradigm for identifying adsorption forms on metal (hydr)oxides using

density functional theory simulations and FTIR spectroscopy.

116

8.3 Identification of adsorption forms of fatty acid on ferric (hydr)oxide NPs

As a model for ferric (hydr)oxide NPs, hematite NP of 150-nm size (H150) was selected and for

fatty acid, sodium laurate (SL), CH3(CH2)10COONa, a commonly used anionic surfactant. In DFT

simulations, the hematite surface was modeled using a ferric trioctahedral cluster obtained from the (104)

hematite surface containing two corner-shared and one edge-shared ferric sites (Figure 8.2). The (104)

surface is one of the naturally available surfaces of hematite.106

A butanoate ion (C3H7COO) was used to

model laurate to reduce the calculation time, because there is no significant impact of the chain length on

the electrostatic or steric structure of the carboxyl head group.252

Spin-polarized orbitals were utilized to

account for the anti-ferromagnetic properties of hematite. A continuum solvation model, COSMO, with

dielectric constant, εr, of 12 was employed to model the interfacial species.241

A two point finite difference

scheme of analytical forces was employed to calculate vibrational frequencies of the model systems. The

vibrational modes were assigned through a visual inspection of the atomic displacement trajectories

created during the frequency calculations.

Figure 8.2 Ferric trioctahedral cluster extracted form hematite (104) surface. The cluster contains both

edge-shared and corner-shared oxygen ions. The surface terminal OH/H2O groups that are involved in

coordination with carboxylate group are shown in yellow spheres. (Iron – blue spheres; oxygen – large

red spheres; carbon – large gray sphere; hydrogen – small green spheres)

Iron

117

Scheme 8.1 Cluster models of different laurate adsorption complexes. Blue octahedra – Fe3+; red

spheres – oxygen; gray spheres – carbon; green spheres – hydrogen.

8.2.1 FTIR results

The FTIR spectrum of the monolayer of laurate (Figure 8.3) is characterized by three adsorption species

identified using the spectral asymmetric asCO2 and symmetric sCO2

stretching bands of the

carboxylate group and C=O mode of carboxylic group (–COOH). The asCO2 and symmetric sCO2

stretching bands of the two adsorbed carboxylate groups are at 1540 and 1410 cm–1

, and 1530 and 1425

cm–1

respectively. Whereas, the C=O mode of carboxylic group (–COOH) is present at 1707 cm–1

.

Based on the valence saturation argument, lauric acid can exist at the NP-water interface only as an OS

complex.

An interesting observation is the absence lauric acid at low surface coverage of the surfactant at

both pH 7.15 and 10.0 (Figure 8.4). It appears at pH 10.0 at coverage higher than 0.25 ML, which is

reached after ~1 h of the conditioning with 1 mM SL. Given laurate is present in the 1 mM solution at pH

10 exclusively as anion,161, 162

its protonation on the hematite surface can be attributed to the adsorption-

induced decrease in the local interfacial pH. This conclusion is supported by the DFT results described

below.

b) ISMM a) ISBB c) Chelating d) OS surface hydration

-shared

e) OS surface

hydration-separated

118

Figure 8.3 in situ FTIR-HATR spectra in region of (a) CO2 and (b) scisCH2

vibrational modes of (1)

hematite reacted for 25 min with a 10–3

M solution of sodium laurate at pH 7.15 ± 0.05 with no added

electrolyte. For comparison, shown are spectra of (2) aqueous 10–2

M sodium laurate at pH 10.2

measured with a Ge IRE, (3) solid sodium laurate solid, and (4) solid lauric acid. Baseline in the

16001660-cm-1

region of the adsorbed laurate was corrected by adding spectrum of water (water is

removed from the interface upon the adsorption). This figure is taken from Ref. 190

. Spectral regions of the

C=O and scisCH2 modes of the surfactant are shadowed. Band assignment is taken from Ref.253

119

Figure 8.4 in situ FTIR-HATR spectra of 150-nm hematite reacted successively with SL at (a) pH 7.15

and SL concentration of (black) 0.1 mM, (blue) 0.5 mM and (green) 1.0 mM and (b) pH 10.0 and SL

concentration of 1 mM. Time of the interaction is shown in the graphs. It was measured from the moment

of the introduction of a SL solution into the FTIR-ATR cell till the beginning of the spectrum scanning

(which takes 5 min). This figure is taken from Ref. 190

. The background spectrum was that of the

deposited particles pre-equilibrated with water of pH 7.15 for 1 hour. Spectra are shifted along the Y axis

for clarity.

Based on H/D exchange and pH response of FTIR spectra of adsorbed laurate,190

the following

adsorption forms can be tentatively assigned (1) the strongly adsorbed carboxylate species is assigned to

inner-sphere monodentate mononuclear complex (ISMM), (2) the weakly adsorbed carboxylate to outer-

sphere complex (OS) and (3) the adsorbed lauric acid to OS species.

8.2.2 DFT modeling of carboxylate binding to hematite surface

DFT calculations were performed to verify the foregoing interpretation of the vibrational

spectra as well as to determine the structure of the OS complex. In an earlier study of the

120

carbonate adsorption, I suggested a methodology which was based on analysis of stability of

different surface complexes and comparison of the theoretical and experimental frequencies of

these configurations with those of the dissolved species.241 However, the uncertainty in the

coordination of dissolved laurate (Section 1.1.6) hinders the direct application of this

methodology. Instead, a comparison of the stability of the OS hydration-separated and

hydration-shared complexes (Scheme 1) as well as the relationship between calculated

frequencies of these and ISMM complexes was performed.

The geometry optimization of the OS hydration-separated, OS hydration-shared, and

ISMM complexes were conducted at the ferric cluster charges q = 0 and q = +1. As opposed to

the OS hydration-shared complex, which was stable (Figure 8.5), the OS hydration-separated

counterpart was found to be unstable at both the cluster charges, moving away from the cluster.

Figure 8.5 Geometry optimized DFT models of a butanoate ion attached as ISMM and OS hydration-

shared complexes to a ferric trioctahedral cluster with charge q = 0 which represents the (104) hematite

surface. Numbers are the CO and CC and H bond lengths in Å Dashed lines show the surface

coordinating H bonds.

This finding can be attributed to the expected fact that two hydrating water layers better screen

the electrostatic attraction of the carboxylate group by the positively-charged surface as

ISMM

1.271 1.289

1.527 1.729

1.739

OS hydration-shared

1.517 1.275

1.286

1.85

1.685 1.862 1.788

Iron

121

compared to one layer. Moreover, similar to what was experimentally observed for the weakly

bound laurate, the sCO2 and asCO2

modes of the hydration-shared complexes are higher and

lower than that the ISMM complexes (Table 8.1), respectively. It follows that the OS complex

more likely adopts the hydration shared configuration.

Additionally, it was observed that at a charge of the ferric cluster q > +2 the OS

hydration-shared complex of the alkanoate was protonated by the water molecule that

coordinates the complex to a ferric cation. At the same time, protonation by the co-adsorbed

hydronium cation took place at a charge as low as q = 1. Therefore, the interfacial protonation

of the carboxylate group was more feasible by the co-adsorbed hydronium cation rather than by

the hematite surface. This result concurs with the relationship between the pKa values of the

involved functionalities. Namely, pKa = 5.0 of laurate is lower than IEP of hematite of 7.1 and

hence pKa the adsorbed water (see also Ref.254), but higher than pKa = 1.74 of hydronium. The

protonation by the co-adsorbed proton was further supported by the absence of protonated

laurate in the layer adsorbed from a 0.1 mM solution at IEP of hematite (Figure 5a).

The presented DFT results are consistent with co-adsorption of the OS hydration-shared and

ISMM complexes. They show that protonation of the hydration-shared OS complexes is more likely

caused by a co-adsorbed hydronium ion.

Table 8.1 Calculated vibrational frequencies of IS and OS complexes of butanoate

Configuration of

adsorbed butanoate

Charge on

ferric cluster

Frequencies (cm-1

)

sCO2 asCO2

ISMM 0 1383 1520

ISMM 1 1391 1521

OS hydration-shared 0 1391

1419

1499

OS hydration-shared 1 1382

1405

1480

122

The application of this paradigm for the prediction of laurate is summarized in Figure 8.6

Figure 8.6 The predictive methodology applied in this work for identifying adsorption forms of sodium

laurate on 150-nm hematite NPs using density functional theory simulations and FTIR spectroscopy

123

8.4 Conclusions

It is shown that the ambiguity of the conventional approach to employ DFT modeling for unraveling the

complex vibrational spectra of species at the oxide-water interface can be overcome for adsorbed fatty

acids by searching for the structure that provides the best correspondence of the trends of the theoretical

and experimental spectra upon variations of the solution pH and protonation-deprotonation of the

carboxylate group, instead of searching for the structure that provides the best one-to-one

correspondence between the theoretical and experimental spectra.

124

Chapter 9 Summary

Study 1.Polymeric nano- and micro-particles are being increasingly considered for delivery in

tissue engineering, though they have long attracted attention in other biomedical applications such as

cancer therapy or delivery of vaccines and adjuvants. Two of the important requirements for delivery in

tissue engineering of articular cartilage are the short-term and sustained release of bioactive molecules

such as vitamin C or growth factors.

To satisfy these criteria, polymeric microparticles were fabricated using polysebacic anhydride

(PSA) which has faster degradation rate and possess surface erosion profile. The PSA polymer was

synthesized using the polycondensation method. Its composition and structure were characterized and

confirmed using NMR and Raman spectroscopies. The PSA microparticles were fabricated and

encapsulated using the solvent evaporation based double emulsion method. The average size of the

microparticles was ~1.5 μm and had wide size distribution. A release of the drug (amitriptyline) from the

encapsulated microparticles was linear and sustained.

Although the PSA-based polymers are reported in literature to be highly biocompatible, their

biocompatibility and bioactivity are poorly understood at concentrations higher than 10mg/mL, which are

relevant for delivery in tissue engineering. Moreover, the transport properties of agarose hydrogel matrix

need to addressed and modelled so as to enhance the retention of secreted biomacromolecules. With

these main goals in mind, I (a) modeled and established a correlation between the pore size variation of

agarose hydrogel and agarose concentration; (b) found the optimum concentration of agarose using the

established model and evaluated the range of the applicability of the model when polymeric particles are

incorporated. (c) explored the biocompatibility of PSA microparticles with chondrocytes in a 3D hydrogel

scaffold.

To select the optimum concentration of agarose in the agarose hydrogel scaffold, one needs to

know the pore/void size. The pore size should be large enough for the easy transport of growth factors (>

30 nm), but smaller than 300 nm so as to retain tissue components such as collagen fibrils. However,

there is no direct method to evaluate this quantity. This problem is very technically challenging, especially

125

in the presence of polymeric particles. Through modeling the effect of the agarose concentration on the

turbidity, I established a correlation between the pore size and concentration of the agarose. In particular,

the correlation length between agarose fibers was obtained using the wavelength exponent method.

( ( )

)

( )

Using this method, the variation of pore size with respect to agarose concentration was

empirically derived as

ξ = 522 C-1.83

,

where ξ is the pore size, and C is the agarose concentration in w/v %. The optimal concentration of

agarose was found to be 2 % w/v. When the PSA microparticles are introduced at 0.3% w/v to the

agarose hydrogel, there was increase in pore size from 100-nm to 150-nm. The hydrogel pore size

determination using wavelength exponent method was found to be inadequate when the microparticles

concentration exceeded 0.3% w/v. This is due to the strong contribution to the scattering of radiation by

larger-sized microparticles or their aggregates, apart from the agarose fibers.

In the next step, the PSA microparticles were tested for their biocompatibility to chondrocytes and

their efficacy for delivery in articular cartilage. The PSA microparticles were found to be cytotoxic at a

concentration of 20 mg/mL, but benign at 2 mg/mL. I explained the cytotoxicity by the lipotoxicity resulting

from the intracellular accumulation of PSA or its degradation products, rather than by the changes in local

pH due to the polymer hydrolysis. It follows that the variables that collectively and uniquely determine the

bioactivity of PSA polymer are its degradation rate, functional groups and hydrophobicity but requires

further verification.

Among several strategies applied to mitigate the cytotoxicity of PSA, the administering of bovine

serum albumin (BSA) was found to be the most efficient. The chondro-protective action systemof BSA is

proportional to the ratio of BSA to the polymer. The BSA-based mitigation strategy makes PSA

microparticles effective in the short-term delivery of TGF-β3 in articular cartilage engineering.

Study 2. Functional groups of polymeric materials are known to be important in tissue

engineering, for example, for directing tissue morphogenesis. I addressed for the first time the effects of

126

functional groups on the bioactivity of polymeric nanoparticles in the tissue engineering of cartilage. For

this purpose, I synthesized polyacrylate-based nanoparticles and studied the effect of the

functionalization on the chondrocyte viability and metabolism in an agarose hydrogel. Since the published

protocols for synthesis of the polyacrylate nanoparticles using inverse microemulsion failed, the new

synthesis conditions were determined. For this purpose, a ternary phase diagram was constructed for the

water-in-hexane inverse microemulsion system stabilized by Aerosol-OT surfactant at room temperature.

It was found that the addition of a monomer such as acrylic acid increased the stable inverse emulsion

region. A similar phase diagram was constructed at 50 C, where the stable region remains more or less

similar to that at room temperature. On the basis of these diagrams, the functionalized polyacrylate

nanoparticles were obtained through radical polymerization of monomers (acrylic acid or its functional

derivatives) with crosslinkers. The size of nanoparticles was in the range of 70-300 nm, and surface

charge was characterized using the electrokinetic method.

These polyacrylate-based nanoparticles when introduced to chondrocytes in agarose hydrogel

are found to be biocompatible (2 mg/mL). However, they exhibit bioactivity that is detrimental to the

synthesis of extracellular matrix. There is an inverse correlation between mechanical strength and the

negative surface charge of these nanoparticles. The cause for the poor matrix synthesis is suggested to

be the disruption of the inter-cellular signaling process by higher anionicity. Thus, it is concluded that the

charge on the functional groups of the polymers has a significant effect the polymer bioactivity and can be

used to manipulate the cellular metabolism.

Study 3. Another control of the tissue growth can be the bioactivity of metal oxide nanoparticles. I

discovered, for the first time, a beneficial impact of nanoceria on the biochemical and mechanical

properties of articular cartilage tissue grown in vitro. Specifically, it was demonstrated that nanoceria is

biocompatible toward chondrocytes at concentrations at least up to 1000 μg/mL, while the constructs

cultured with embedded nanoceria at 100 and 1000 μg/mL have mechanical properties and collagen

content significantly better than the control. The mode of administering of nanoceria is also found to be

important in their efficacy to suppress the cartilage tissue damage caused by inflammatory agents such

as interleukin-1α (IL-1α). In particular, if nanoceria is just embedded into the constructs, it does not offer

127

any protection against IL-1α. However, significant chondroprotectivity against IL-1α and partial prevention

of the matrix degradation are observed when the nanoceria is suspended in the growth media at 20000

μg/mL. The found effects have potential in treating the inflammation of articular cartilage due to disease

or damage.

In contrast to nanoceria, ferric (hydr)oxide nanoparticles detrimentally affect the growth of

cartilage tissue. The cytotoxicity of these nanoparticles is unexpected given the literature data. It is shown

for the first time that the adverse effect increases with increasing nanoparticle size, following the trend

earlier observed for the catalytic activity of the nanoparticles. Specifically, the mechanical properties of

the tissue constructs embedded with hematite and ferrihydrite (FH) nanoparticles degrade with increasing

nanoparticle size as follows: Control > FH (2-nm) > H9 (9-nm) > H38 (38-nm) >> H150 (150-nm). On the

other hand, the cellular viability surprisingly remains similar among all these sample sets.

Correlating this result with the earlier observed trend in the catalytic activity116

and the decrease

in the adsorption activity (measured as normalized monolayer coverage of sodium laurate) found in Study

4 (see below), it can be concluded that both the catalytic redox and adsorption activities of metal oxide

nanoparticles contribute to their bioactivity. Moreover, given the beneficial effect of nanoceria, these

results imply that not only the catalytic redox activity per se, but also the catalytic redox mechanism

(including either the deactivation or generation of reactive oxygen species) controls the bioactivity of the

metal oxide at the biointerface. Thus, the emerging paradigm based on the overlap of the conduction

band energy with the redox potential range of redox active cellular components should be further

expanded with incorporation of the adsorption properties and the catalytic redox mechanism.

Study 4. This study was focused on the adsorption and dispersion properties of the metal

(hydr)oxide nanoparticles in the presence of fatty acids in aqueous media, as these properties can also

play an important role in the biochemical and environmental activity and bioavailability of the

nanoparticles. Specifically, I explored the effects of the size and morphology of ferric (hydr)oxide

nanoparticles in these phenomena. Using adsorption isotherm, electrokinetic, FTIR measurements as

well as DFT simulations, the first experimental evidence is obtained for the dependence of the surface

density, speciation, and packing order of adsorbed fatty acids on these properties. The conditions under

128

which fatty acids form self-assembled monolayers (SAMs) and bilayers on such nanoparticles in water

are distinguished. For the first time, the electric polarization of the nanoparticles is demonstrated to be a

powerful tool for manipulating the interfacial properties of the nanoparticles.

Specifically, it is shown that sodium laurate is adsorbed on hematite mainly as a mixture of

chemisorbed (inner sphere monodentate, ISMM) and physisorbed (outer sphere hydration-shared, OS)

complexes, both at pH > PZC and pH < PZC. The OS complex is partially protonated by a coadsorbed

hydronium cation as either surface coverage increases or solution pH decreases. Independent of

morphology, an increase in nanosize improves the adsorption capacity and affinity of hematite

nanoparticles, as well as increases the fraction of the protonated physisorbed (OS) alkanoate, in

agreement with the nanosize-induced changes in the structural and electronic properties of the

nanoparticles. This implies that the surface electronic and structural properties, which can be controlled

by size for metal oxide nanoparticles, offer important ways to manipulate adsorption properties.

Consequently, an increase in nanosize of hematite nanoparticles improves the packing order of laurate

and hence hydrophobicity of the nanoparticles provided a similar hexagonal habit of these nanoparticles.

It was also observed that, independent of morphology, the dispersibility of the nanoparticles by fatty acids

in water degrades as nanoparticle size decreases. This effect is attributed to the decrease in the

adsorption capacity which ends up in the formation of a loosely packed bilayer.

Morphology of the nanoparticles was found to be critical for the self-assembly if (i) the majority of

the molecules are chemisorbed (ISMM) and (ii) the morphology-determined surface arrangement of the

adsorption sites is geometrically incompatible with a surfactant ordering. The surface ratio of the

physisorbed/chemisorbed surfactant depends on the surface charge of nanoparticles. Negative

polarization suppresses chemisorption. This effect can be used to remove the geometrical constraints for

self-assembly. The dominance of the chemisorbed complexes results in the isotropic surface filling by

fatty acids, while the prevalence of OS complexes favors the cluster-like filling pattern.

A new approach to DFT modeling of the in situ FTIR spectra of adsorbed species was developed

to circumvent the intrinsic ambiguities of the conventional DFT modeling in interpreting vibrational

spectra, especially at the oxide-water interfaces. This new approach constitutes searching of the

129

adsorbed structure that provides the best correspondence of the trends of the theoretical and

experimental spectra upon variations of the solution pH, ionic strength and protonation-deprotonation of

the carboxylate group, instead of conventional approach of searching for the structure that provides the

best one-to-one correspondence between the theoretical and experimental spectra. This approach was

utilized to interpret the FTIR data in obtaining information on the adsorption forms and packing of fatty

acids on ferric (hydr)oxide nanoparticles.

Thus, the experimental, methodological and theoretical results obtained in these studies are an

important step toward developing a general paradigm predicting the (bio-)activity for both the polymeric

and metal oxide micro- and nano-particles. For the polymeric particles considered (polyanhydrides and

polyacrylates), the following factors are found to determine their bioactivity: degradation rate,

hydrophobicity, functional group and surface charge. In addition to above parameters, concentration of

these particles with respect to cell density is critical factor determining their (bio-)activity. In the case of

metal oxide nanoparticles, the catalytic redox properties, the redox mechanism, and the adsorption

properties have been suggested as the dominating factor controlling their cyto-protective and cytotoxic

properties, provided other physical parameters such as dissolution and aspect ratio are not critical.

Additionally, the DFT-based methodology to interpret the adsorption phenomena developed in this work

can further augment the possibility of formulating such a general paradigm for (bio-)activity.

130

Summary of main results

Effect of Material Significant physicochemical properties (Bio-)activity

Effect of polymers

a. Polyanhydrides

b. Polyacrylates

Degradation rate; hydrophobicity; carboxylic groups

Surface anionic charge

Lipotoxic effects

Decreased tissue synthesis

Effect of cerium oxide

nanoparticles

Mode of particles addition; redox properties Improved protein synthesis;

Protection against inflammation

Effect of hematite

nanoparticles

Nanoparticle size, morphology, redox and catalytic

properties.

Cytotoxicity due to size-

dependent catalytic activity

Size and morphology

dependent adsorption and

dispersion properties

131

Chapter 10 Suggested Future Studies

The following future studies are suggested based on the results of the present work.

1) To decrease the impact of lipotoxicity of polysebacic anhydride (PSA) due to the polymer

degradation, fabricate encapsulated microparticles using a physical composite of a slow

degrading polymer such as poly-caprolactone (in higher fractions) and the fast degrading PSA (in

lower fractions). This can create a bulk and surface erosion profile more effective for short-term

delivery.

2) To decrease the concentration of degrading PSA polymer, perform a study of the encapsulation

efficiency and release profile as a function of the thickness of the hydrophobic shell of the PSA

microparticles.

3) To enhance the cell-polymer interaction, functionalize the PSA polymer or polyacrylate

nanoparticles with either transforming growth factor (TFG-β3) or cell-adhering peptide sequence

Arginine-Glycine-Aspartic acid (RGD).

4) To improve the biocompatibility of polyacrylate nanoparticles (NPs) for chondrocytes, adapt the

stem cells to tolerate these NPs before their chondrogenic differentiation.

5) To increase the anti-inflammatory effect of nanoceria on chondrocytes, enhance nanoceria

uptake during the expansion stage of chondrocytes using mechanical motion or through

functionalization of nanoceria with cell-adhering RGD peptide sequence.

6) To protect tissue damage from oxidants, embed electron-rich nanomaterials such as zinc or

selenium nanoparticles in the tissue constructs and study its impact.

132

7) To rationally control the biocompatibility of ferric (hydr)oxide NPs, determine the effect of NP size

on their cellular uptake using electron microscopy and ICP atomic emission spectroscopy and

measure the impact of these NPs on the degradation of extracellular matrix components such as

GAG and collagen. Establish a direct correlation, if any, between degradation of matrix and

physicochemical properties of NPs

8) Explore the generality of the found nanosize effects on the adsorption and dispersion properties

of metal oxide NPs by studying other technologically important oxides that are thermodynamically

instable at the nanoscale such as corundum (-Al2O3) and rutile (TiO2)

9) To explain the size-dependence of the dispersion properties of (hydr)oxide NPs in the presence

of fatty acids in water, use DLVO or extended DLVO and flocculation models to quantify the

dispersibility.

10) To understand the (bio-)activity of ferric (hydr)oxide NPs, study the effect of size and morphology

on adsorption properties of proteins on ferric (hydr)oxide NPs. Establish any similarity to the (bio-

)activity ferric (hydr)oxide vis-a-vis chondrocytes and sodium laurate. This may enable to suggest

a general paradigm of the (bio-)activity for ferric (hydr)oxide NPs in particular or metal (hydr)oxide

in general

133

REFERENCES 1. Fernandez-Garcia, M.; Rodriguez, J. A., Metal oxide nanoparticles. In Nanomaterials: Inorganic and bioinorganic perspectives, Lukehart, C. M.; Scott, R. A., Eds. Wiley: Habooken, 2009; pp 453-474.

2. Nel, A. E.; Madler, L.; Velegol, D.; Xia, T.; Hoek, E. M. V.; Somasundaran, P.; Klaessig, F.; Castranova, V.; Thompson, M., Understanding biophysicochemical interactions at the nano-bio interface. Nature Materials 2009, 8, (7), 543-557.

3. Stockwell, R. A., Recent knowledge about adult articular-cartilage - a review. J. Bone Joint Surg.-Br. Vol. 1980, 62, (2), 267-267.

4. Poole, A. R.; Kojima, T.; Yasuda, T.; Mwale, F.; Kobayashi, M.; Laverty, S., Composition and structure of articular cartilage - a template for tissue repair. Clin. Orthop. Relat. Res. 2001, (391), S26-S33.

5. Vunjak-Novakovic, G.; Altman, G.; Horan, R.; Kaplan, D. L., Tissue engineering of ligaments. Annu Rev Biomed Eng 2004, 6, 131-56.

6. Hettrich, C. M.; Crawford, D.; Rodeo, S. A., Cartilage repair: Third-generation cell-based technologies--basic science, surgical techniques, clinical outcomes. Sports Med Arthrosc 2008, 16, (4), 230-5.

7. Brittberg, M.; Lindahl, A.; Nilsson, A.; Ohlsson, C.; Isaksson, O.; Peterson, L., Treatment of deep cartilage defects in the knee with autologous chondrocyte transplantation. New England Journal of Medicine 1994, 331, (14), 889-895.

8. Wood, J. J.; Malek, M. A.; Frassica, F. J.; Power, J. A.; Mohan, A. K.; Bloom, E. T.; Braun, M. M.; Cote, T. R., Autologous cultured chondrocytes: Adverse events reported to the united states food and drug administration. Journal of Bone and Joint Surgery-American Volume 2006, 88A, (3), 503-507.

9. Revell, C. M.; Athanasiou, K. A., Success rates and immunologic responses of autogenic, allogenic, and xenogenic treatments to repair articular cartilage defects. Tissue Engineering Part B-Reviews 2009, 15, (1), 1-15.

10. Hung, C. T.; Mauck, R. L.; Wang, C. C. B.; Lima, E. G.; Ateshian, G. A., A paradigm for functional tissue engineering of articular cartilage via applied physiologic deformational loading. Ann. Biomed. Eng. 2004, 32, (1), 35-49.

11. Mauck, R. L.; Soltz, M. A.; Wang, C. C. B.; Wong, D. D.; Chao, P. H. G.; Valhmu, W. B.; Hung, C. T.; Ateshian, G. A., Functional tissue engineering of articular cartilage through dynamic loading of chondrocyte-seeded agarose gels. Journal of Biomechanical Engineering-Transactions of the Asme 2000, 122, (3), 252-260.

12. Benya, P. D.; Shaffer, J. D., Dediffrentiated chondrocytes reexpress the differentiated collagen phenotype when cultured in agarose gels. Cell 1982, 30, (1), 215-224.

13. Buschmann, M. D.; Gluzband, Y. A.; Grodzinsky, A. J.; Kimura, J. H.; Hunziker, E. B., Chondrocytes in agarose culture synthesize a mechanically functional extracellular-matrix. J. Orth. Res. 1992, 10, (6), 745-758.

14. Buschmann, M. D.; Gluzband, Y. A.; Grodzinsky, A. J.; Hunziker, E. B., Mechanical compression modulates matrix biosynthesis in chondrocyte agarose culture. J. Cell Sci. 1995, 108, 1497-1508.

134

15. Lima, E. G.; Bian, L.; Ng, K. W.; Mauck, R. L.; Byers, B. A.; Tuan, R. S.; Ateshian, G. A.; Hung, C. T., The beneficial effect of delayed compressive loading on tissue-engineered cartilage constructs cultured with tgf-beta 3. Osteoarthritis and Cartilage 2007, 15, (9), 1025-1033.

16. Selmi, T. A. S.; Verdonk, P.; Chambat, P.; Dubrana, F.; Potel, J.-F.; Barnouin, L.; Neyret, P., Autologous chondrocyte implantation in a novel alginate-agarose hydrogel: Outcome at two years. J Bone Joint Surg Br 2008, 90-B, (5), 597-604.

17. Bian, L.; Angione, S. L.; Ng, K. W.; Lima, E. G.; Williams, D. Y.; Mao, D. Q.; Ateshian, G. A.; Hung, C. T., Influence of decreasing nutrient path length on the development of engineered cartilage. Osteoarthritis and Cartilage 2009, 17, (5), 677-685.

18. Ng, K. W.; O'Conor, C. J.; Kugler, L. E.; Cook, J. L.; Ateshian, G. A.; Hung, C. T., Transient supplementation of anabolic growth factors rapidly stimulates matrix synthesis in engineered cartilage. Annals of Biomedical Engineering 2011, 39, (10), 2491-2500.

19. Vert, M., Aliphatic polyesters: Great degradable polymers that cannot do everything. Biomacromolecules 2005, 6, (2), 538-546.

20. Nair, L. S.; Laurencin, C. T., Biodegradable polymers as biomaterials. Progress in Polymer Science 2007, 32, (8-9), 762-798.

21. Lloyd, A. W., Interfacial bioengineering to enhance surface biocompatibility. Medical device technology 2002, 13, (1), 18-21.

22. Dang, W.; Daviau, T.; Brem, H., Morphological characterization of polyanhydride biodegradable implant gliadel during in vitro and in vivo erosion using scanning electron microscopy. Pharmaceutical Research 1996, 13, (5), 683-91.

23. Akbari, H.; D'Emanuele, A.; Attwood, D., Effect of geometry on the erosion characteristics of polyanhydride matrices. International Journal of Pharmaceutics 1998, 160, (1), 83-89.

24. Domb, A. J.; Nudelman, R., In-vivo and in-vitro elimination of aliphatic polyanhydrides. Biomaterials 1995, 16, (4), 319-323.

25. Leong, K. W.; Brott, B. C.; Langer, R., Bioerodible polyanhydrides as drug-carrier matrices .1. Characterization, degradation, and release characteristics. Journal of Biomedical Materials Research 1985, 19, (8), 941-955.

26. Mathiowitz, E.; Kline, D.; Langer, R., Morphology of polyanhydride microsphere delivery systems. Scanning Microscopy 1990, 4, (2), 329-340.

27. Park, E. S.; Maniar, M.; Shah, J., Effects of model compounds with varying physicochemical properties on erosion of polyanhydride devices. Journal of Controlled Release 1996, 40, (1-2), 111-121.

28. Shieh, L.; Tamada, J.; Chen, I.; Pang, J.; Domb, A.; Langer, R., Erosion of a new family of biodegradable polyanhydrides. Journal of Biomedical Materials Research 1994, 28, (12), 1465-1475.

29. Tamada, J. A.; Langer, R., Erosion kinetics of hydrolytically degradable polymers. Proceedings of the National Academy of Sciences of the United States of America 1993, 90, (2), 552-556.

30. Mader, K.; Bacic, G.; Domb, A.; Elmalak, O.; Langer, R.; Swartz, H. M., Noninvasive in vivo monitoring of drug release and polymer erosion from biodegradable polymers by epr spectroscopy and nmr imaging. Journal of Pharmaceutical Sciences 1997, 86, (1), 126-134.

135

31. Mader, K.; Cremmilleux, Y.; Domb, A. J.; Dunn, J. F.; Swartz, H. M., In vitro in vivo comparison of drug release and polymer erosion from biodegradable p(fad-sa) polyanhydrides - a noninvasive approach by the combined use of electron paramagnetic resonance spectroscopy and nuclear magnetic resonance imaging. Pharmaceutical Research 1997, 14, (6), 820-826.

32. Laurencin, C.; Domb, A.; Morris, C.; Brown, V.; Chasin, M.; McConnell, R.; Lange, N.; Langer, R., Poly(anhydride) administration in high-does in vivo - studies of biocompatibility and toxicology. Journal of Biomedical Materials Research 1990, 24, (11), 1463-1481.

33. Attawia, M. A.; Herbert, K. M.; Uhrich, K. E.; Langer, R.; Laurencin, C. T., Proliferation, morphology, and protein expression by osteoblasts cultured on poly(anhydride-co-imides). Journal of Biomedical Materials Research 1999, 48, (3), 322-327.

34. Attawia, M. A.; Uhrich, K. E.; Botchwey, E.; Fan, M.; Langer, R.; Laurencin, C. T., Cytotoxicity testing of poly(anhydride-co-imides) for orthopedic applications. Journal of Biomedical Materials Research 1995, 29, (10), 1233-1240.

35. Domb, A. J., Synthesis and characterization of biodegradable aromatic anhydride coplymers. Macromolecules 1992, 25, (1), 12-17.

36. Domb, A. J.; Maniar, M., Absorbable biopolymers derived from dimer fatty acids. Journal of Polymer Science Part a-Polymer Chemistry 1993, 31, (5), 1275-1285.

37. Ibim, S. E. M.; Uhrich, K. E.; Attawia, M.; Shastri, V. R.; El-Amin, S. F.; Bronson, R.; Langer, R.; Laurencin, C. T., Preliminary in vivo report on the osteocompatibility of poly(anhydride-co-imides) evaluated in a tibial model. Journal of Biomedical Materials Research 1998, 43, (4), 374-379.

38. Ibim, S. M.; Uhrich, K. E.; Bronson, R.; El-Amin, S. F.; Langer, R. S.; Laurencin, C. T., Poly(anhydride-co-imides): In vivo biocompatibility in a rat model. Biomaterials 1998, 19, (10), 941-951.

39. Leong, K. W.; Damore, P.; Marletta, M.; Langer, R., Bioerodible polyanhydrides as drug-carrier matrices .2. Biocompatibility and chemical reactivity. Journal of Biomedical Materials Research 1986, 20, (1), 51-64.

40. Braun, A. G.; Buckner, C. A.; Emerson, D. J.; Nichinson, B. B., Quantitative correspondence between the in vivo and in vitro acitivity of teratogenic agents. Proceedings of the National Academy of Sciences of the United States of America-Biological Sciences 1982, 79, (6), 2056-2060.

41. Skopek, T. R.; Liber, H. L.; Krolewski, J. J.; Thilly, W. G., Quantitative forward mutation assay in salmonella-typhimurim using 8-asaguanine resistance as a genetic-marker. Proceedings of the National Academy of Sciences of the United States of America 1978, 75, (1), 410-414.

42. Domb, A. J.; Rock, M.; Perkin, C.; Yipchuck, G.; Broxup, B.; Villemure, J. G., Excretion of a radiolabeled anticancer biodegradable polymeric implant form the rabbit brain. Biomaterials 1995, 16, (14), 1069-1072.

43. Domb, A. J.; Rock, M.; Schwartz, J.; Perkin, C.; Yipchuk, G.; Broxup, B.; Villemure, J. G., Metabolic disposition and elimination studies of a radiolabeled biodegradable polymeric implant in the rat-brain. Biomaterials 1994, 15, (9), 681-688.

44. Katti, D. S.; Lakshmi, S.; Langer, R.; Laurencin, C. T., Toxicity, biodegradation and elimination of polyanhydrides. Advanced Drug Delivery Reviews 2002, 54, (7), 933-961.

136

45. Ross, R., Smooth muscle cell .2. Growth of smooth muscle in culture and formation of elastic fibers. Journal of Cell Biology 1971, 50, (1), 172-&.

46. Adler, A. F.; Petersen, L. K.; Wilson, J. H.; Torres, M. P.; Thorstenson, J. B.; Gardner, S. W.; Mallapragada, S. K.; Wannemuehler, M. J.; Narasimhan, B., High throughput cell-based screening of biodegradable polyanhydride libraries. Comb. Chem. High Throughput Screen 2009, 12, (7), 634-645.

47. Urban, J. P. G.; Hall, A. C.; Gehl, K. A., Regulation of matrix synthesis rates by the ionic and osmotic environment of articular chondrocytes. J. Cell. Physiol. 1993, 154, (2), 262-270.

48. Wu, M. H.; Urban, J. P. G.; Cui, Z. F.; Cui, Z.; Xu, X., Effect of extracellular ph on matrix synthesis by chondrocytes in 3d agarose gel. Biotechnology Progress 2007, 23, (2), 430-434.

49. Almaguel, F. G.; Liu, J. W.; Pacheco, F. J.; Casiano, C. A.; De Leon, M., Activation and reversal of lipotoxicity in pc12 and rat cortical cells following exposure to palmitic acid. J. Neurosci. Res. 2009, 87, (5), 1207-1218.

50. Kemppainen, J. M.; Hollister, S. J., Tailoring the mechanical properties of 3d-designed poly(glycerol sebacate) scaffolds for cartilage applications. Journal of Biomedical Materials Research Part A 2010, 94A, (1), 9-18.

51. Weiner, A. A.; Bock, E. A.; Gipson, M. E.; Shastri, V. P., Photocrosslinked anhydride systems for long-term protein release. Biomaterials 2008, 29, (15), 2400-2407.

52. Biondi, M.; Ungaro, F.; Quaglia, F.; Netti, P. A., Controlled drug delivery in tissue engineering. Adv. Drug Del. Rev. 2008, 60, (2), 229-242.

53. Goldberg, M.; Langer, R.; Jia, X. Q., Nanostructured materials for applications in drug delivery and tissue engineering. Journal of Biomaterials Science-Polymer Edition 2007, 18, (3), 241-268.

54. Jia, X. Q.; Kiick, K. L., Hybrid multicomponent hydrogels for tissue engineering. Macromol. Biosci. 2009, 9, (2), 140-156.

55. Kohane, D. S.; Langer, R., Polymeric biomaterials in tissue engineering. Pediatr. Res. 2008, 63, (5), 487-491.

56. Perez-Alvarez, L.; Saez-Matinez, V.; Hernaez, E.; Katime, I., Synthesis and characterization of ph-sensitive microgels by derivatization of npa-based reactive copolymers. Mater. Chem. Phys. 2008, 112, (2), 516-524.

57. Sahiner, N.; Jha, A. K.; Nguyen, D.; Jia, X. Q., Fabrication and characterization of cross-linkable hydrogel particles based on hyaluronic acid: Potential application in vocal fold regeneration. Journal of Biomaterials Science-Polymer Edition 2008, 19, (2), 223-243.

58. Ravi, N.; Aliyar, H. A.; Hamilton, P. D., Hydrogel nanocomposite as a synthetic intra-ocular lens capable of accommodation. 2005, 227, 191-201.

59. Nelea, V.; Luo, L.; Demers, C. N.; Antoniou, J.; Petit, A.; Lerouge, S.; Wertheimer, M. R.; Mwale, F., Selective inhibition of type x collagen expression in human mesenchymal stem cell differentiation on polymer substrates surface-modified by glow discharge plasma. Journal of Biomedical Materials Research Part A 2005, 75A, (1), 216-223.

60. Salinas, C. N.; Anseth, K. S., Decorin moieties tethered into peg networks induce chondrogenesis of human mesenchymal stem cells. J Biomed Mater Res A 2009, 90, (2), 456-64.

137

61. Park, K.; Cho, K. J.; Kim, J. J.; Kim, I. H.; Han, D. K., Functional plga scaffolds for chondrogenesis of bone-marrow-derived mesenchymal stem cells. Macromol. Biosci. 2009, 9, (3), 221-229.

62. Jo, S.; Shin, H.; Mikos, A. G., Modification of oligo(poly(ethylene glycol) fumarate) macromer with a grgd peptide for the preparation of functionalized polymer networks. Biomacromolecules 2001, 2, (1), 255-261.

63. Benoit, D. S. W.; Schwartz, M. P.; Durney, A. R.; Anseth, K. S., Small functional groups for controlled differentiation of hydrogel-encapsulated human mesenchymal stem cells. Nat Mater 2008, 7, (10), 816-823.

64. Swartz, M. A.; Fleury, M. E., Interstitial flow and its effects in soft tissues. In Annu rev biomed eng, 2007; Vol. 9, pp 229-256.

65. Halliwell, B.; Gutteridge, J. M. C., The definition and measurement of antioxidants in biological-systems. Free Radical Biol. Med. 1995, 18, (1), 125-126.

66. Gutteridge, J. M. C.; Halliwell, B., Free radicals and antioxidants in the year 2000 - a historical look to the future. In Reactive oxygen species: From radiation to molecular biology: A festschrift in honor of daniel l gilbert, Chiueh, C. C., Ed. 2000; Vol. 899, pp 136-147.

67. Young, I. S.; Woodside, J. V., Antioxidants in health and disease. J. Clin. Pathol. 2001, 54, (3), 176-186.

68. Thannickal, V. J.; Fanburg, B. L., Reactive oxygen species in cell signaling. American Journal of Physiology-Lung Cellular and Molecular Physiology 2000, 279, (6), L1005-L1028.

69. Stohs, S. J.; Bagchi, D., Oxidative mechanisms in the toxicity of metal-ions. Free Radical Biol. Med. 1995, 18, (2), 321-336.

70. Zhang, H.; Ji, Z.; Xia, T.; Meng, H.; Low-Kam, C.; Liu, R.; Pokhrel, S.; Lin, S.; Wang, X.; Liao, Y.-P.; Wang, M.; Li, L.; Rallo, R.; Damoiseaux, R.; Telesca, D.; Mädler, L.; Cohen, Y.; Zink, J. I.; Nel, A. E., Use of metal oxide nanoparticle band gap to develop a predictive paradigm for oxidative stress and acute pulmonary inflammation. Acs Nano 2012, 6, (5), 4349-4368.

71. Celardo, I.; De Nicola, M.; Mandoli, C.; Pedersen, J. Z.; Traversa, E.; Ghibelli, L., Ce3+ ions determine redox-dependent anti-apoptotic effect of cerium oxide nanoparticles. Acs Nano 2011, 5, (6), 4537-4549.

72. Celardo, I.; Pedersen, J. Z.; Traversa, E.; Ghibelli, L., Pharmacological potential of cerium oxide nanoparticles. Nanoscale 2011, 3, (4), 1411-1420.

73. Schubert, D.; Dargusch, R.; Raitano, J.; Chan, S. W., Cerium and yttrium oxide nanoparticles are neuroprotective. Biochem. Biophys. Res. Commun. 2006, 342, (1), 86-91.

74. Trovarelli, A., Catalytic properties of ceria and ceo2-containing materials. Catalysis Reviews-Science and Engineering 1996, 38, (4), 439-520.

75. Singhal, R. K.; Kumari, P.; Samariya, A.; Kumar, S.; Sharma, S. C.; Xing, Y. T.; Saitovitch, E. B., Role of electronic structure and oxygen defects in driving ferromagnetism in nondoped bulk ceo2. Appl. Phys. Lett. 2010, 97, (17).

76. Chen, S. Y.; Tsai, C. H.; Huang, M. Z.; Yan, D. C.; Huang, T. W.; Gloter, A.; Chen, C. L.; Lin, H. J.; Chen, C. T.; Dong, C. L., Concentration dependence of oxygen vacancy on the magnetism of ceo2 nanoparticles. J. Phys. Chem. C 2012, 116, (15), 8707-8713.

138

77. Skorodumova, N. V.; Ahuja, R.; Simak, S. I.; Abrikosov, I. A.; Johansson, B.; Lundqvist, B. I., Electronic, bonding, and optical properties of ceo(2) and ce(2)o(3) from first principles. Phys. Rev. B 2001, 64, (11).

78. Tsunekawa, S.; Sivamohan, R.; Ito, S.; Kasuya, A.; Fukuda, T., Structural study on monosize ceo2-x nano-particles. Nanostruct. Mater. 1999, 11, (1), 141-147.

79. Paun, C.; Safonova, O. V.; Szlachetko, J.; Abdala, P. M.; Nachtegaal, M.; Sa, J.; Kleymenov, E.; Cervellino, A.; Krumeich, F.; van Bokhoven, J. A., Polyhedral ceo2 nanoparticles: Size-dependent geometrical and electronic structure. J. Phys. Chem. C 2012, 116, (13), 7312-7317.

80. Karakoti, A.; Singh, S.; Dowding, J. M.; Seal, S.; Self, W. T., Redox-active radical scavenging nanomaterials. Chem. Soc. Rev. 2010, 39, (11), 4422-4432.

81. Das, M.; Patil, S.; Bhargava, N.; Kang, J. F.; Riedel, L. M.; Seal, S.; Hickman, J. J., Auto-catalytic ceria nanoparticles offer neuroprotection to adult rat spinal cord neurons. Biomaterials 2007, 28, (10), 1918-1925.

82. Karakoti, A. S.; Singh, S.; Kumar, A.; Malinska, M.; Kuchibhatla, S. V. N. T.; Wozniak, K.; Self, W. T.; Seal, S., Pegylated nanoceria as radical scavenger with tunable redox chemistry. J. Am. Chem. Soc. 2009, 131, (40), 14144-+.

83. Rotter, N.; Ung, F.; Roy, A. K.; Vacanti, M.; Eavey, R. D.; Vacanti, C. A.; Bonassar, L. J., Role for interleukin 1 alpha in the inhibition of chondrogenesis in autologous implants using polyglycolic acid-polylactic acid scaffolds. Tissue Eng. 2005, 11, (1-2), 192-200.

84. Capito, R. M.; Spector, M., Scaffold-based articular cartilage repair. Ieee Engineering in Medicine and Biology Magazine 2003, 22, (5), 42-50.

85. Habibovic, P.; Woodfield, T.; de Groot, K.; van Blitterswijk, C., Predictive value of in vitro and in vivo assays in bone and cartilage repair - what do they really tell us about the clinical performance? In Tissue eng., Fisher, J. P., Ed. Springer-Verlag Berlin: Berlin, 2006; Vol. 585, pp 327-360.

86. Francioli, S.; Cavallo, C.; Grigolo, B.; Martin, I.; Barbero, A., Engineered cartilage maturation regulates cytokine production and interleukin-1 beta response. Clin. Orthop. Relat. Res. 2011, 469, (10), 2773-2784.

87. Lima, E. G.; Tan, A. R.; Tai, T.; Bian, L.; Stoker, A. M.; Ateshian, G. A.; Cook, J. L.; Hung, C. T., Differences in interleukin-1 response between engineered and native cartilage. Tissue Engineering Part A 2008, 14, (10), 1721-1730.

88. Lima, E. G.; Tan, A. R.; Tai, T.; Marra, K. G.; DeFail, A.; Ateshian, G. A.; Hung, C. T., Genipin enhances the mechanical properties of tissue-engineered cartilage and protects against inflammatory degradation when used as a medium supplement. Journal of Biomedical Materials Research Part A 2009, 91A, (3), 692-700.

89. Tyler, J. A., Insulin-like growth factor-1 can decrease degradation and promote synthesis of proteoglycan in cartilage exposed to cytokines. Biochem. J 1989, 260, (2), 543-548.

90. Shiraishi, T.; DeMeester, S. R.; Worrall, N. K.; Ritter, J. H.; Misko, T. P.; Ferguson, T. B., Jr.; Cooper, J. D.; Patterson, G. A., Inhibition of inducible nitric oxide synthase ameliorates rat lung allograft rejection J Thorac Cardiovasc Surg 1995, 110, (5), 1449-1460.

91. McDermott, C. D.; Gavita, S. M.; Shennib, H.; Giaid, A., Immunohistochemical localization of nitric oxide synthase and the oxidant peroxynitrite in lung transplant recipients with obliterative bronchiolitis. Transplantation 1997, 64, (2), 270-4.

139

92. Niu, J. L.; Azfer, A.; Rogers, L. M.; Wang, X. H.; Kolattukudy, P. E., Cardioprotective effects of cerium oxide nanoparticles in a transgenic murine model of cardiomyopathy. Cardiovascular Research 2007, 73, (3), 549-559.

93. Hirst, S. M.; Karakoti, A. S.; Tyler, R. D.; Sriranganathan, N.; Seal, S.; Reilly, C. M., Anti-inflammatory properties of cerium oxide nanoparticles. Small 2009, 5, (24), 2848-2856.

94. Amin, K. A.; Hassan, M. S.; Awad, E. T.; Hashem, K. S., The protective effects of cerium oxide nanoparticles against hepatic oxidative damage induced by monocrotaline. International Journal of Nanomedicine 2011, 6, 143-149.

95. Clark, A.; Zhu, A. P.; Sun, K.; Petty, H. R., Cerium oxide and platinum nanoparticles protect cells from oxidant-mediated apoptosis. J. Nanopart. Res. 2011, 13, (10), 5547-5555.

96. Horie, M.; Nishio, K.; Kato, H.; Fujita, K.; Endoh, S.; Nakamura, A.; Miyauchi, A.; Kinugasa, S.; Yamamoto, K.; Niki, E.; Yoshida, Y.; Hagihara, Y.; Iwahashi, H., Cellular responses induced by cerium oxide nanoparticles: Induction of intracellular calcium level and oxidative stress on culture cells. J. Biochem. 2011, 150, (4), 461-471.

97. Alili, L.; Sack, M.; Karakoti, A. S.; Teuber, S.; Puschmann, K.; Hirst, S. M.; Reilly, C. M.; Zanger, K.; Stahl, W.; Das, S.; Seal, S.; Brenneisen, P., Combined cytotoxic and anti-invasive properties of redox-active nanoparticles in tumor-stroma interactions. Biomaterials 2011, 32, (11), 2918-2929.

98. Tarnuzzer, R. W.; Colon, J.; Patil, S.; Seal, S., Vacancy engineered ceria nanostructures for protection from radiation-induced cellular damage. Nano Lett. 2005, 5, (12), 2573-2577.

99. Colon, J.; Hsieh, N.; Ferguson, A.; Kupelian, P.; Seal, S.; Jenkins, D. W.; Baker, C. H., Cerium oxide nanoparticles protect gastrointestinal epithelium from radiation-induced damage by reduction of reactive oxygen species and upregulation of superoxide dismutase 2. Nanomedicine-Nanotechnology Biology and Medicine 2010, 6, (5), 698-705.

100. Babu, S.; Cho, J. H.; Dowding, J. M.; Heckert, E.; Komanski, C.; Das, S.; Colon, J.; Baker, C. H.; Bass, M.; Self, W. T.; Seal, S., Multicolored redox active upconverter cerium oxide nanoparticle for bio-imaging and therapeutics. Chem. Commun. 2010, 46, (37), 6915-6917.

101. Karakoti, A. S.; Tsigkou, O.; Yue, S.; Lee, P. D.; Stevens, M. M.; Jones, J. R.; Seal, S., Rare earth oxides as nanoadditives in 3-d nanocomposite scaffolds for bone regeneration. J. Mater. Chem. 2010, 20, (40), 8912-8919.

102. Li, X. L.; Wei, W. J.; Wang, S. Z.; Kuai, L.; Geng, B. Y., Single-crystalline alpha-fe(2)o(3) oblique nanoparallelepipeds: High-yield synthesis, growth mechanism and structure enhanced gas-sensing properties. Nanoscale 2011, 3, (2), 718-724.

103. Han, J. S.; Davey, D. E.; Mulcahy, D. E.; Yu, A. B., Effect of the ph value of the precipitation solution on the co sensitivity of alpha-fe2o3. Sensors and Actuators B-Chemical 1999, 61, (1-3), 83-91.

104. Zboril, R.; Mashlan, M.; Petridis, D., Iron(iii) oxides from thermal processes-synthesis, structural and magnetic properties, mossbauer spectroscopy characterization, and applications. Chem. Mater. 2002, 14, (3), 969-982.

105. Morup, S.; Madsen, D. E.; Frandsen, C.; Bahl, C. R. H.; Hansen, M. F., Experimental and theoretical studies of nanoparticles of antiferromagnetic materials. J. Phys. Condens. Matter 2007, 19, (21).

106. Cornell, R. M.; Schwertmann, U., The iron oxide: Structure, properties, reactions and uses. VCH: Weinheim, 1996.

140

107. Michel, F. M.; Ehm, L.; Antao, S. M.; Lee, P. L.; Chupas, P. J.; Liu, G.; Strongin, D. R.; Schoonen, M. A. A.; Phillips, B. L.; Parise, J. B., The structure of ferrihydrite, a nanocrystalline material. Science 2007, 316, (5832), 1726-1729.

108. Michel, F. M.; Barron, V.; Torrent, J.; Morales, M. P.; Serna, C. J.; Boily, J. F.; Liu, Q. S.; Ambrosini, A.; Cismasu, A. C.; Brown, G. E., Ordered ferrimagnetic form of ferrihydrite reveals links among structure, composition, and magnetism. Proceedings of the National Academy of Sciences of the United States of America 2010, 107, (7), 2787-2792.

109. Chernyshova, I. V.; Ponnurangam, S.; Somasundaran, P., On the origin of an unusual dependence of (bio)chemical reactivity of ferric hydroxides on nanoparticle size. PCCP 2010, 12, (42), 14045-14056.

110. Chernyshova, I. V.; Ponnurangam, S.; Somasundaran, P., Effect of nanosize on catalytic properties of ferric (hydr)oxides in water: Mechanistic insights. J. Catalysis 2011, 282 (1), 25–34.

111. Wang, Y.; Du, W. P.; Xu, Y. M., Effect of sintering temperature on the photocatalytic activities and stabilities of hematite and silica-dispersed hematite particles for organic degradation in aqueous suspensions. Langmuir 2009, 25, (5), 2895-2899.

112. Zhang, W.; Rittmann, B.; Chen, Y. S., Size effects on adsorption of hematite nanoparticles on e. Coli cells. Environ. Sci. Technol. 2011, 45, (6), 2172-2178.

113. Gojova, A.; Guo, B.; Kota, R. S.; Rutledge, J. C.; Kennedy, I. M.; Barakat, A. I., Induction of inflammation in vascular endothelial cells by metal oxide nanoparticles: Effect of particle composition. Environ. Health Perspect. 2007, 115, (3), 403-409.

114. Chernyshova, I. V.; Ponnurangam, S.; Somasundaran, P., Effect of nanosize on catalytic properties of ferric (hydr)oxides in water: Mechanistic insights. J. Catal. 2011, 282, (1), 25-34.

115. Karlsson, H. L.; Cronholm, P.; Gustafsson, J.; Moller, L., Copper oxide nanoparticles are highly toxic: A comparison between metal oxide nanoparticles and carbon nanotubes. Chem. Res. Toxicol. 2008, 21, (9), 1726-1732.

116. Chernyshova, I. V.; Ponnurangam, S.; Somasubdaran, P., Effect of nanosize on catalytic properties of ferric (hydr)oxides in water: Mechanistic insights. J. Catalysis 2011, 282 (1), 25–34.

117. Chernyshova, I. V., Anodic processes on a galena (pbs) electrode in the presence of n-butyl xanthate studied ftir-spectroelectrochemically. J. Phys. Chem. B 2001, 105, (34), 8185-8191.

118. Chernyshova, I. V., Anodic oxidation of galena (pbs) studied ftir-spectroelectrochemically. J. Phys. Chem. B 2001, 105, (34), 8178-8184.

119. Chernyshova, I. V., In situ study of the oxidation of galenite (natural pbs) in alkaline media by ftir spectroscopy: Anode processes in the absence of oxygen. Russ. J. Electrochem. 2001, 37, (6), 579-584.

120. Chernyshova, I. V., In situ ftir-spectroelectrochemical study of the anodic processes on a galena (pbs) electrode under open-air conditions in the absence and presence of n-butyl xanthate. Langmuir 2002, 18, (18), 6962-6968.

121. Chernyshova, I. V., An in situ ftir study of galena and pyrite oxidation in aqueous solution. J. Electroanal. Chem. 2003, 558, 83-98.

141

122. Chernyshova, I. V. In An in situ ftir spectroscopic study of the galena and pyrite oxidation in aqueous solution, Proceedings of the Electrochemical Society 2003-18, Princeton, 2003; Woods, R., Ed. Electrochemical Society: Princeton, 2003; pp 166-180.

123. Chernyshova, I. V., Pyrite oxidation mechanism in aqueous solutions: An in situ ftir study. Russ. J. Electrochem. 2004, 40, (1), 69-77.

124. Gerischer, H., Electrolytic decomposition and photo-decomposition of compound semiconductors in contact with electrolytes. Journal of Vacuum Science & Technology 1978, 15, (4), 1422-1428.

125. Rosen, M. J., Surfactants and interfacial phenomena. 3rd ed.; Wiley: Hoboken, 2004; p 444.

126. Nel, A.; Xia, T.; Madler, L.; Li, N., Toxic potential of materials at the nanolevel. Science 2006, 311, (5761), 622-627.

127. Lynch, I.; Dawson, K. A., Protein-nanoparticle interactions. Nano Today 2008, 3, (1-2), 40-47.

128. Dutta, D.; Sundaram, S. K.; Teeguarden, J. G.; Riley, B. J.; Fifield, L. S.; Jacobs, J. M.; Addleman, S. R.; Kaysen, G. A.; Moudgil, B. M.; Weber, T. J., Adsorbed proteins influence the biological activity and molecular targeting of nanomaterials. Toxicol. Sci. 2007, 100, (1), 303-315.

129. Lundqvist, M.; Stigler, J.; Elia, G.; Lynch, I.; Cedervall, T.; Dawson, K. A., Nanoparticle size and surface properties determine the protein corona with possible implications for biological impacts. Proceedings of the National Academy of Sciences of the United States of America 2008, 105, (38), 14265-14270.

130. Sperling, R. A.; Parak, W. J., Surface modification, functionalization and bioconjugation of colloidal inorganic nanoparticles. Philosophical Transactions of the Royal Society a-Mathematical Physical and Engineering Sciences 2010, 368, (1915), 1333-1383.

131. Caruso, F., Colloids and colloid assemblies: Synthesis, modification, organization and utilization of colloid particles. Wiley-VCH Verlag, : Weinheim, 2004.

132. Fendler, J. H., Self-assembled nanostructured materials. Chem. Mater. 1996, 8, (8), 1616-1624.

133. Boettcher, S. W.; Strandwitz, N. C.; Schierhorn, M.; Lock, N.; Lonergan, M. C.; Stucky, G. D., Tunable electronic interfaces between bulk semiconductors and ligand-stabilized nanoparticle assemblies. Nature Materials 2007, 6, (8), 592-596.

134. Gupta, A. K.; Gupta, M., Synthesis and surface engineering of iron oxide nanoparticles for biomedical applications. Biomaterials 2005, 26, (18), 3995-4021.

135. Soenen, S. J.; Rivera-Gil, P.; Montenegro, J. M.; Parak, W. J.; De Smedt, S. C.; Braeckmans, K., Cellular toxicity of inorganic nanoparticles: Common aspects and guidelines for improved nanotoxicity evaluation. Nano Today 2011, 6, (5), 446-465.

136. Yu, T.; Malugin, A.; Ghandehari, H., Impact of silica nanoparticle design on cellular toxicity and hemolytic activity. Acs Nano 2011, 5, (7), 5717-5728.

137. Auffan, M.; Rose, J.; Bottero, J. Y.; Lowry, G. V.; Jolivet, J. P.; Wiesner, M. R., Towards a definition of inorganic nanoparticles from an environmental, health and safety perspective. Nat. Nanotechnol. 2009, 4, (10), 634-641.

142

138. Calatayud, M.; Maldonado, L.; Minot, C., Reactivity of (tio2)(n) clusters (n=1-10): Probing gas-phase acidity and basicity properties. J. Phys. Chem. C 2008, 112, (41), 16087-16095.

139. Gilbert, B.; Banfield, J. F., Molecular-scale processes involving nanoparticulate minerals in biogeochemical systems. In Molecular geomicrobiology, 2005; Vol. 59, pp 109-155.

140. Wottrich, R.; Diabate, S.; Krug, H. F., Biological effects of ultrafine model particles in human macrophages and epithelial cells in mono- and co-culture. Int. J. Hyg. Environ. Health 2004, 207, (4), 353-361.

141. Cwiertny, D. M.; Handler, R. M.; Schaefer, M. V.; Grassian, V. H.; Scherer, M. M., Interpreting nanoscale size-effects in aggregated fe-oxide suspensions: Reaction of fe(ii) with goethite. Geochim. Cosmochim. Acta 2008, 72, (5), 1365-1380.

142. Bose, S.; Hochella, M. F.; Gorby, Y. A.; Kennedy, D. W.; McCready, D. E.; Madden, A. S.; Lower, B. H., Bioreduction of hematite nanoparticles by the dissimilatory iron reducing bacterium shewanella oneidensis mr-1. Geochim. Cosmochim. Acta 2009, 73, (4), 962-976.

143. Villalobos, M.; Cheney, M. A.; Alcaraz-Cienfuegos, J., Goethite surface reactivity: Ii. A microscopic site-density model that describes its surface area-normalized variability. J. Colloid Interface Sci. 2009, 336, (2), 412-422.

144. Mudunkotuwa, I. A.; Grassian, V. H., The devil is in the details (or the surface): Impact of surface structure and surface energetics on understanding the behavior of nanomaterials in the environment. J. Environ. Monit. 2011, 13, (5), 1135-1144.

145. Larcher, D.; Tarascon, J.-M., Oxide nanoparticles: Electrochemical performance. In Dekker encyclopedia of nanoscience and nanotechnology, Marcel Dekker, Inc.: New York, 2004.

146. Villalobos, M.; Perez-Gallegos, A., Goethite surface reactivity: A macroscopic investigation unifying proton, chromate, carbonate, and lead(ii) adsorption. J. Colloid Interface Sci. 2008, 326, (2), 307-323.

147. Vayenas, C. G.; Yentekakis, I. V.; Bebelis, S. I.; Neophytides, S. G., In-situ controlled promotion of catalyst surfaces via solid electrolytes - the nemca effect. Ber. Bunsen-Ges. Phys. Chem. Chem. Phys. 1995, 99, (11), 1393-1401.

148. Woods, R., In Modern aspects of electrochemistry, Bockris, J. O.; Conway, B. E.; White, R. E., Eds. Plenum Press: New York, 1996; pp 401-453.

149. Tolstoy, V. P.; Chernyshova, I. V.; Skryshevsky, V. A., Handbook of infrared spectroscopy of ultrathin films. Wiley: Hoboken, 2003; p 710.

150. Cruz, R.; Luna-Sanchez, R. M.; Lapidus, G. T.; Gonzalez, I.; Monroy, M., An experimental strategy to determine galvanic interactions affecting the reactivity of sulfide mineral concentrates. Hydrometallurgy 2005, 78, (3-4), 198-208.

151. Chernyshova, I. V.; Ponnurangam, S.; Somasundaran, P., Tailoring (bio)chemical activity of semiconducting nanoparticles: Critical role of deposition and aggregation. Journal of the American Chemical Society 2011, 133, (24), 9536-9544.

152. Klahr, B. M.; Hamann, T. W., Current and voltage limiting processes in thin film hematite electrodes. J. Phys. Chem. C 2011, 115, (16), 8393-8399.

153. Klahr, B. M.; Hamann, T. W., Voltage dependent photocurrent of thin film hematite electrodes. Appl. Phys. Lett. 2011, 99, (6).

143

154. Listorti, A.; Durrant, J.; Barber, J., Artificial photosynthesis: Solar to fuel. Nature Materials 2009, 8, (12), 929-U22.

155. Na, H. B.; Song, I. C.; Hyeon, T., Inorganic nanoparticles for mri contrast agents. Adv. Mater. 2009, 21, (21), 2133-2148.

156. Laurent, S.; Forge, D.; Port, M.; Roch, A.; Robic, C.; Elst, L. V.; Muller, R. N., Magnetic iron oxide nanoparticles: Synthesis, stabilization, vectorization, physicochemical characterizations, and biological applications. Chem. Rev. 2008, 108, (6), 2064-2110.

157. Farn, R. J., Chemistry and technology of surfactants. Blackwell Publishing Ltd: Oxford, 2007; p i-xix.

158. Jarek, E.; Jasinski, T.; Barzyk, W.; Warszynski, P., The ph regulated surface activity of alkanoic acids. Colloids and Surfaces a-Physicochemical and Engineering Aspects 2010, 354, (1-3), 188-196.

159. Gu, B. H.; Mehlhorn, T. L.; Liang, L. Y.; McCarthy, J. F., Competitive adsorption, displacement, and transport of organic matter on iron oxide .1. Competitive adsorption. Geochim. Cosmochim. Acta 1996, 60, (11), 1943-1950.

160. Gao, X. D.; Metge, D. W.; Ray, C.; Harvey, R. W.; Chorover, J., Surface complexation of carboxylate adheres cryptosporidium parvum oocysts to the hematite-water interface. Environ. Sci. Technol. 2009, 43, (19), 7423-7429.

161. Laskowski, J. S., Electrokinetic measurements in aqueous-solutions of weak electrolyte type surfactants. J. Colloid Interface Sci. 1993, 159, (2), 349-353.

162. Kralchevsky, P. A.; Danov, K. D.; Pishmanova, C. I.; Kralchevska, S. D.; Christov, N. C.; Ananthapadmanabhan, K. P.; Lips, A., Effect of the precipitation of neutral-soap, acid-soap, and alkanoic acid crystallites on the bulk ph and surface tension of soap solutions. Langmuir 2007, 23, (7), 3538-3553.

163. Somasundaran, P.; Ananthapadmanabhan, K. P.; Ivanov, I. B., Dimerization of oleate in aqueous-solutions. J. Colloid Interface Sci. 1984, 99, (1), 128-135.

164. Smith, R.; Tanford, C., Hydrophobicity of long-chain alkyl carboxylic-acids, as measured by their distribution between heptane and aqueous-solution. Proceedings of the National Academy of Sciences of the United States of America 1973, 70, (2), 289-293.

165. Kanicky, J. R.; Shah, D. O., Effect of premicellar aggregation on the pk(a) of fatty acid soap solutions. Langmuir 2003, 19, (6), 2034-2038.

166. Smith, S. W.; Anderson, B. D., Salt and mesophase formation in aqueous suspension of lauric acid. Pharm. Res. 1993, 10, (10), 1533-1543.

167. Lucassen, J., Hydrolysis and precipitates in carboxylate soap solutions. J. Phys. Chem. 1966, 70, (6), 1824-&.

168. Cistola, D. P.; Hamilton, J. A.; Jackson, D.; Small, D. M., Ionization and phase-behavior of fatty acids in water-application of the gibbs phase rule. Biochemistry 1988, 27, (6), 1881-1888.

169. Quast, K., Flotation of hematite using c6-c18 saturated fatty acids. Miner. Eng. 2006, 19, (6-8), 582-597.

170. Quast, K. B., A review of hematite flotation using 12-carbon chain collectors. Miner. Eng. 2000, 13, (13), 1361-1376.

144

171. Gong, W. Q.; Parentich, A.; Little, L. H.; Warren, L. J., Diffuse reflectance infrared fourier-transform spectroscopic study of the adsorption mechanis, of oleate on hematite. Colloids and Surfaces 1991, 60, 325-339.

172. Roonasi, P.; Yang, X. F.; Holmgren, A., Competition between sodium oleate and sodium silicate for a silicate/oleate modified magnetite surface studied by in situ atr-ftir spectroscopy. J. Colloid Interface Sci. 2010, 343, (2), 546-552.

173. Schrodle, S.; Richmond, G. L., Equilibrium and non-equilibrium kinetics of self-assembled surfactant monolayers: A vibrational sum-frequency study of dodecanoate at the fluorite-water interface. J. Am. Chem. Soc. 2008, 130, (15), 5072-5085.

174. McGuire, M. J.; Addai-Mensah, J.; Bremmell, K. E., Improved dewaterability of iron oxide dispersions: The role of polymeric flocculant type, ph and shear. Asia-Pac. J. Chem. Eng. 2008, 3, (1), 18-23.

175. Hong, S. C.; Lee, J. H.; Lee, J.; Kim, H. Y.; Park, J. Y.; Cho, J.; Lee, J.; Han, D. W., Subtle cytotoxicity and genotoxicity differences in superparamagnetic iron oxide nanoparticles coated with various functional groups. International Journal of Nanomedicine 2011, 6, 3219-3231.

176. Soenen, S. J. H.; De Cuyper, M., Assessing iron oxide nanoparticle toxicity in vitro: Current status and future prospects. Nanomedicine 2010, 5, (8), 1261-1275.

177. Pisanic, T. R., 2nd; Jin, S.; V.I., S., Iron oxide magnetic nanoparticle nanotoxicity: Incidence and mechanisms. In Nanotoxicity, Sahu, S.; Casciano, D., Eds. Wiley-VCH: Weinheim, 2009; pp 397-425.

178. Bhattacharya, K.; Hoffmann, E.; Schins, R. F. P.; Boertz, J.; Prantl, E. M.; Alink, G. M.; Byrne, H. J.; Kuhlbusch, T. A. J.; Rahman, Q.; Wiggers, H.; Schulz, C.; Dopp, E., Comparison of micro- and nanoscale fe+3-containing (hematite) particles for their toxicological properties in human lung cells in vitro. Toxicol. Sci. 2012, 126, (1), 173-182.

179. Freyria, F. S.; Bonelli, B.; Tomatis, M.; Ghiazza, M.; Gazzano, E.; Ghigo, D.; Garrone, E.; Fubini, B., Hematite nanoparticles larger than 90 nm show no sign of toxicity in terms of lactate dehydrogenase release, nitric oxide generation, apoptosis, and comet assay in murine alveolar macrophages and human lung epithelial cells. Chem. Res. Toxicol. 2012, 25, (4), 850-861.

180. Gramowski, A.; Flossdorf, J.; Bhattacharya, K.; Jonas, L.; Lantow, M.; Rahman, Q.; Schiffmann, D.; Weiss, D. G.; Dopp, E., Nanoparticles induce changes of the electrical activity of neuronal networks on microelectrode array neurochips. Environ. Health Perspect. 2010, 118, (10), 1363-1369.

181. Sun, J.; Wang, S. C.; Zhao, D.; Hun, F. H.; Weng, L.; Liu, H., Cytotoxicity, permeability, and inflammation of metal oxide nanoparticles in human cardiac microvascular endothelial cells cytotoxicity, permeability, and inflammation of metal oxide nanoparticles. Cell Biol. Toxicol. 2011, 27, (5), 333-342.

182. Ananthapadmanabhan, K. P.; Somasundaran, P., Oleate chemistry and hematite flotation. In Interfacial Phenomena in Mineral Processing, Engineering Foundation Conference Proceedings, Yarar, B.; Spottiswood, D. J., Eds. New York, 1981; pp pp. 207-227.

183. Kulkarni, R. D.; Somasundaran, P., Flotation chemistry of hematite-oleate system Colloids and Surfaces 1980, 1, (3-4), 387-405.

184. Shibata, J.; Fuerstenau, D. W., Flocculation and flotation characteristics of fine hematite with sodium oleate. Int. J. Miner. Process. 2003, 72, (1-4), 25-32.

145

185. Jung, R. F.; James, R. O.; Healy, T. W., Adsorption, precipitation, and electrokinetic processes in the iron oxide (goethite)--oleic acid--oleate system. Journal of Colloid and Interface Science 1987, 118, (2), 463-472.

186. Ofor, O., Oleate adsorption at a nigerian hematite-water interface - effect of concentration, temperature, and ph on adsorption density. J. Colloid Interface Sci. 1995, 174, (2), 345-350.

187. Deacon, G. B.; Huber, F.; Phillips, R. J., Diagnosis of the nature of carboxylate coordination form the direction of shifts of carbon oxygen strecthing frequencies. Inorganica Chimica Acta-Articles and Letters 1985, 104, (1), 41-45.

188. Deacon, G. B.; Phillips, R. J., Relationships between the carbon-oxygen stretching frequencies of carboxylato complexes and the type of carboxylate coordination Coord. Chem. Rev. 1980, 33, (3), 227-250.

189. Tackett, J. E., Ft-ir characterization of metal acetates in aqueous solution. Applied Spectroscopy 1989, 43, (3), 483-489.

190. Chernyshova, I. V.; Ponnurangam, S.; Somasundaran, P., Adsorption of fatty acids on iron (hydr)oxides from aqueous solutions. Langmuir : the ACS journal of surfaces and colloids 2011, 27, (16), 10007-18.

191. Benton, H. P.; Tyler, J. A., Inhibition of cartilage proteoglycan synthesis by interleukin-1. Biochem. Biophys. Res. Commun. 1988, 154, (1), 421-428.

192. Kelly, T. A. N.; Ng, K. W.; Wang, C. C. B.; Ateshian, G. A.; Hung, C. T., Spatial and temporal development of chondrocyte-seeded agarose constructs in free-swelling and dynamically loaded cultures. J. Biomech. 2006, 39, (8), 1489-1497.

193. Farndale, R. W.; Buttle, D. J.; Barrett, A. J., Improved quantitation and discrimination of sulfated glycosaminoglycans by use of dimethylmethylene blue. Biochimica Et Biophysica Acta 1986, 883, (2), 173-177.

194. Farndale, R. W.; Sayers, C. A.; Barrett, A. J., A direct spectrophotometric micro-assay for sulfated glycosaminoglycans in cartilage cultures. Connective Tissue Research 1982, 9, (4), 247-248.

195. Stegeman.H; Stalder, K., Determination of hydroxyproline. Clinica Chimica Acta 1967, 18, (2), 267-&.

196. Leong, K. W.; Simonte, V.; Langer, R., Synthesis of polyanhydrides - melt-polycondensation dehydrochlorination, and dehydrative coupling. Macromolecules 1987, 20, (4), 705-712.

197. Domb, A. J.; Amselem, S.; Shah, J.; Maniar, M., Polyanhydrides - synthesis and characterization. Advances in Polymer Science 1993, 107, 93-141.

198. Prudencio, A.; Schmeltzer, R. C.; Uhrich, K. E., Effect of linker structure on salicylic acid-derived poly(anhydride-esters). Macromolecules 2005, 38, (16), 6895-6901.

199. Schmeltzer, R. C.; Anastasiou, T. J.; Uhrich, K. E., Optimized synthesis of salicylate-based poly(anhydride-esters). Polymer Bulletin 2003, 49, (6), 441-448.

200. Domb, A. J.; Mathiowitz, E.; Ron, E.; Giannos, S.; Langer, R., Polyanhydrides .4. Unsaturated and cross-linked polyanhydrides. Journal of Polymer Science Part a-Polymer Chemistry 1991, 29, (4), 571-579.

146

201. Domb, A. J.; Langer, R., Solid-state and solution stability of poly(anhydrides) and poly(esters). Macromolecules 1989, 22, (5), 2117-2122.

202. Tudor, A. M.; Melia, C. D.; Davies, M. C.; Hendra, P. J.; Church, S.; Domb, A. J.; Langer, R., The application of fourier-transform raman spectroscopy to the analysis of poly(anhydride) homo-polymers and copolymers. Spectrochimica Acta Part a-Molecular and Biomolecular Spectroscopy 1991, 47, (9-10), 1335-1343.

203. Petersen, L. K.; Sackett, C. K.; Narasimhan, B., High-throughput analysis of protein stability in polyanhydride nanoparticles. Acta Biomaterialia 2010, 6, (10), 3873-3881.

204. Torres, M. P.; Wilson-Welder, J. H.; Lopac, S. K.; Phanse, Y.; Carrillo-Conde, B.; Ramer-Tait, A. E.; Bellaire, B. H.; Wannemuehler, M. J.; Narasimhan, B., Polyanhydride microparticles enhance dendritic cell antigen presentation and activation. Acta Biomaterialia 2011, 7, (7), 2857-2864.

205. Aymard, P.; Williams, M. A. K.; Clark, A. H.; Norton, I. T., A turbidimetric study of phase separating biopolymer mixtures during thermal ramping. Langmuir 2000, 16, (19), 7383-7391.

206. Janaky, N.; Jun-Ying, X.; Xiang-Yang, L., Determination of agarose gel pore size: Absorbance measurements vis a vis other techniques. Journal of Physics: Conference Series 2006, 28, (1), 83.

207. Doty, P.; Steiner, R. F., Light scattering and spectrophotometry of colloidal solutions. Journal of Chemical Physics 1950, 18, (9), 1211-1220.

208. Aymard, P.; Martin, D. R.; Plucknett, K.; Foster, T. J.; Clark, A. H.; Norton, I. T., Influence of thermal history on the structural and mechanical properties of agarose gels. Biopolymers 2001, 59, (3), 131-144.

209. Zimm, B. H., The scattering of light and the radial distribution function of high polymer solutions. Journal of Chemical Physics 1948, 16, (12), 1093-1099.

210. Zimm, B. H., Apparatus and methods for measurement and interpretation of the angular variation of light scattering - preliminary results on polystyrene solutions. Journal of Chemical Physics 1948, 16, (12), 1099-1116.

211. Horne, D. S., Determination of the fractal dimension using turbidimetric techniques. Application to aggregating protein systems. Faraday Discussions of the Chemical Society 1987, 83, 259-270.

212. Ramzi, M.; Rochas, C.; Guenet, J. M., Structure-properties relation for agarose thermoreversible gels in binary solvents. Macromolecules 1998, 31, (18), 6106-6111.

213. Serwer, P.; Hayes, S. J., Exclusion of spheres by agarose gels during agarose-gel electrophoresis - dependence on the sphere radius and the gels concentration. Anal. Biochem. 1986, 158, (1), 72-78.

214. Cancelli.A; Frontali, C.; Gratton, E., Dispersoin effect on turbidimetric size measurements. Biopolymers 1974, 13, (4), 735-743.

215. Cameriniotero, R. D.; Day, L. A., Wavelength dependence of turbidity of solutions of macromolecules. Biopolymers 1978, 17, (9), 2241-2249.

216. Melik, D. H.; Fogler, H. S., Turbidimetric determination of particle-size distributions of colloidal systems. Journal of Colloid and Interface Science 1983, 92, (1), 161-180.

217. Schwertmann, U., Cornell, R.M., Iron oxides in the laboratory: Preparation and

characterization. 2nd ed., ed.; Wiley-VCH Weinheim, New York, 2000.

147

218. Matijevic, E.; Scheiner, P., Ferric hydrous oxide sols .3. Preparation of uniform particles by hydrolysis of fe(iii)-chloride, fe(iii)-nitrate, and fe(iii)-perchlorate solutions. Journal of Colloid and Interface Science 1978, 63, (3), 509-524.

219. Carta, D.; Casula, M. F.; Corrias, A.; Falqui, A.; Navarra, G.; Pinna, G., Structural and magnetic characterization of synthetic ferrihydrite nanoparticles. Mater. Chem. Phys. 2009, 113, (1), 349-355.

220. Janney, D. E.; Cowley, J. M.; Buseck, P. R., Transmission electron microscopy of synthetic 2-and 6-line ferrihydrite. Clays Clay Miner. 2000, 48, (1), 111-119.

221. Rodriguez, R. D.; Demaille, D.; Lacaze, E.; Jupille, J.; Chaneac, C.; Jolivet, J. P., Rhombohedral shape of hematite nanocrystals synthesized via thermolysis of an additive-free ferric chloride solution. J. Phys. Chem. C 2007, 111, (45), 16866-16870.

222. Chernyshova, I. V.; Hochella, M. F.; Madden, A. S., Size-dependent structural transformations of hematite nanoparticles. 1. Phase transition. PCCP 2007, 9, (14), 1736-1750.

223. Kosmulski, M., The ph-dependent surface charging and the points of zero charge. J. Colloid Interface Sci. 2002, 253, (1), 77-87.

224. Kosmulski, M., Ph-dependent surface charging and points of zero charge iii. Update. J. Colloid Interface Sci. 2006, 298, (2), 730-741.

225. Smith, D. S.; Ferris, F. G., Specific surface chemical interactions between hydrous ferric oxide and iron-reducing bacteria determined using pk(a) spectra. J. Colloid Interface Sci. 2003, 266, (1), 60-67.

226. Su, C. M.; Suarez, D. L., In situ infrared speciation of absorbed carbonate on aluminum and iron oxide. Clays Clay Miner. 1997, 45, (6), 814-825.

227. Arai, Y.; Sparks, D. L.; Davis, J. A., Effects of dissolved carbonate on arsenate adsorption and surface speciation at the hematite-water interface. Environ. Sci. Technol. 2004, 38, (3), 817-824.

228. Madden, A. S.; Hochella, M. F., A test of geochemical reactivity as a function of mineral size: Manganese oxidation promoted by hematite nanoparticles. Geochim. Cosmochim. Acta 2005, 69, (2), 389-398.

229. Prasad, A.; Bloom, M. S.; Carpenter, D. O., Role of calcium and ros in cell death induced by polyunsaturated fatty acids in murine thymocytes. J. Cell. Physiol. 2010, 225, (3), 829-836.

230. Choi, J. K.; Ho, J.; Curry, S.; Qin, D. H.; Bittman, R.; Hamilton, J. A., Interactions of very long-chain saturated fatty acids with serum albumin. J. Lipid Res. 2002, 43, (7), 1000-1010.

231. Spector, A. A., Fatty-acid binding to plasma albumin. J. Lipid Res. 1975, 16, (3), 165-179.

232. Fainerman, V. B.; Miller, R.; Mohwald, H., General relationships of the adsorption behavior of surfactants at the water/air interface. J. Phys. Chem. B 2002, 106, (4), 809-819.

233. Chernyshova, I. V.; Ponnurangam, S.; Somasundaran, P., Adsorption of fatty acids on iron (hydr)oxides from aqueous solutions. Langmuir 2011, 27, (16), 10007-10018.

234. Berna, A.; Delgado, J. M.; Orts, J. M.; Rodes, A.; Feliu, J. M., Spectroelectrochemical study of the adsorption of acetate anions at gold single crystal and thin-film electrodes. Electrochim. Acta 2008, 53, (5), 2309-2321.

235. Iwasita, T.; Rodes, A.; Pastor, E., Vibrational spectroscopy of carbonate adsorbed on pt(111) and pt(110) single-crystal electrodes. J. Electroanal. Chem. 1995, 383, (1-2), 181-189.

148

236. Kitaigorodskii, A. I., Organic chemical crystallography. Consultans Bureau: New York, 1961.

237. Schwartz, D. K., Langmuir-blodgett film structure. Surf. Sci. Rep. 1997, 27, (7-8), 245-334.

238. Jacquemain, D.; Leveiller, F.; Weinbach, S. P.; Lahav, M.; Leiserowitz, L.; Kjaer, K.; Alsnielsen, J., Crystal-structures of self-aggregates of insoluble aliphatic molecules at the air-water-interface - an x-ray synchrotron study. Journal of the American Chemical Society 1991, 113, (20), 7684-7691.

239. de Leeuw, N. H.; Cooper, T. G., Surface simulation studies of the hydration of white rust fe(oh)(2), goethite alpha-feo(oh) and hematite alpha-fe(2)o(3). Geochimica Et Cosmochimica Acta 2007, 71, (7), 1655-1673.

240. Lan, Q.; Liu, C.; Yang, F.; Liu, S.; Xu, J.; Sun, D., Synthesis of bilayer oleic acid-coated fe3o4 nanoparticles and their application in ph-responsive pickering emulsions. J. Colloid Interface Sci. 2007, 310, (1), 260-269.

241. Ponnurangam, S.; Chernyshova, I. V.; Somasundaran, P., Effect of co-adsorption of electrolyte ions on the stability of inner-sphere complexes. J. Phys. Chem. C 2010, 114, (39), 16517–16524.

242. Navrotsky, A.; Mazeina, L.; Majzlan, J., Size-driven structural and thermodynamic complexity in iron oxides. Science 2008, 319, (5870), 1635-1638.

243. Lützenkirchen, J., ed., Surface complexation modeling. Elsevier: Amsterdam, 2006; Vol. 11., p 652.

244. Noren, K.; Persson, P., Adsorption of monocarboxylates at the water/goethite interface: The importance of hydrogen bonding. Geochim. Cosmochim. Acta 2007, 71, (23), 5717-5730.

245. Noren, K.; Loring, J. S.; Persson, P., Adsorption of alpha amino acids at the water/goethite interface. J. Colloid Interface Sci. 2008, 319, (2), 416-428.

246. Allara, D. L.; Nuzzo, R. G., Spontaneously organized molecular assemblies .1. Formation, dynamics, and physical-properties of normal-alkanoic acids adsorbed from solution on an oxidized aluminum surface. Langmuir 1985, 1, (1), 45-52.

247. Allara, D. L.; Nuzzo, R. G., Spontaneously organized molecular assemblies .2. Quantitative infrared sp ectroscopic determination of equilibrium structures of solution-adsorbed normal-alkanoic acids on an oxidized aluminum surface. Langmuir 1985, 1, (1), 52-66.

248. Lim, M. S.; Feng, K.; Chen, X. Q.; Wu, N. Q.; Raman, A.; Nightingale, J.; Gawalt, E. S.; Korakakis, D.; Hornak, L. A.; Timperman, A. T., Adsorption and desorption of stearic acid self-assembled monolayers on aluminum oxide. Langmuir 2007, 23, (5), 2444-2452.

249. Hohenberg, P.; Kohn, W., Inhomogeneous electron gas. Phys. Rev. B 1964, 136, (3B), B864-&.

250. Capelle, K., A bird’s-eye view of density-functional theory. arXiv:cond-mat; arxiv.org/pdf/cond-mat/0211443v5.pdf 2006, 0211443v5.

251. Kohn, W.; Sham, L. J., Self-consistent equations including exchange and correlation effects. Physical Review 1965, 140, (4A), 1133-&.

252. Rayne, S.; Forest, K., Theoretical studies on the pk(a) values of perfluoroalkyl carboxylic acids. Journal of Molecular Structure-Theochem 2010, 949, (1-3), 60-69.

149

253. Hayashi, S.; Umemura, J., Infrared spectroscopic evidence for coexistence of 2 molecular configurations in crystalline fatty-acids. Journal of Chemical Physics 1975, 63, (5), 1732-1740.

254. Venema, P.; Hiemstra, T.; Weidler, P. G.; van Riemsdijk, W. H., Intrinsic proton affinity of reactive surface groups of metal (hydr)oxides: Application to iron (hydr)oxides. Journal of Colloid and Interface Science 1998, 198, (2), 282-295.

255. Chiriac, V.; Balea, G., Buffer index and buffer capacity for a simple buffer solution. J. Chem. Educ. 1997, 74, (8), 937.

150

APPENDIX

A. Estimation of buffer capacity needed to offset the drop in pH due to polsebacic

anhydride degradation

It was observed from experiments that when polysebacic anhydride microparticles were added to

culture medium there is pH drop from 7.2 to ~6.3. Thus additional buffer capacity, other than bicarbonate

that is present in the growth medium, is needed to offset this drop.

The buffer capacity is defined as infinitesimal quantity of base needed to change the pH by

infinitesimal amount.255

( )

( ) (

[ ]

( [ ]) [ ] [ ]) - (1)

Where,

β is buffer capacity, B is moles of strong base added

A is moles of strong base added, C is the total concentration of the buffer

Ka is acid dissociation constant for buffer

Assuming the worst case of 50% prior degradation of polysebacic anhydride molecules into basic sebacic

acid molecules and subsequent release of two protons from each sebacic acid molecules when they are

suspended in growth medium. Molecular weight of sebacic acid is 202 g/mol

A 50 mg of polymer in 1mL of growth medium will then release = 25/202 ≈ 0.125 mmoles of sebacic acid.

Thus 0.25 mmoles/mL protons will be released. This corresponds to 0.25 M addition of strong acid.

Thus d(A) = 0.25M and let the change of pH be set 0.2 (from 7.2 to 7.0)

Therefore –d(A)/d(pH) = 1.25.

If the Tris-HCl buffer is added to maintain the pH then the pKa = 8.07

Solving the equation (1) for the needed total Tris concentration at pH 7.2,

C= (0.54 – [H+] –[OH

-])/0.10631 = 5.4 mM of Tris buffer

Thus ~ 5mM of Tris or buffers with similar pKa and pH range such as HEPES is even in the worst case of

pH changes.

151

B. Determination of wavelength exponent

The wavelength exponent experimental WLE from 700 to 800-nm is extracted from turbidity

measurements as follows,

Turbidity, ( ) ( )

Where, l is the path length = 1 cm.

The wavelength exponent (WLE) is obtained by fitting a slope of log-log plot of τ and λ from 700-800 nm

( ( )

)

For 4% agarose hydrogel

For 3% agarose hydrogel

152

For 2% agarose hydrogel

For 1% agarose hydrogel

For 2% agarose hydrogel+ 0.3% polysebacic anhydride microparticles

153

Table Pore size and wavelength exponent as function of agarose

concentration and cooling time (from 95 C at 0 mins to room temperature

by 2 hours)

4 % Agarose 3 % Agarose

Time (min)

Wavelength exponent

Correlation Length (nm)

Time (min)

Wavelength exponent

Correlation Length (nm)

22 -5.4 0 34 -4.32 0

26 -4.19 0 41 -3.67 34

34 -3.71 29 46 -3.5 41

40 -3.65 31.5 54 -3.35 53

46 -3.61 33.5 61 -3.3 59

52 -3.58 35 66 -3.27 62

59 -3.56 36 71 -3.25 63.5

63 -3.55 36.5 76 -3.24 64

71 -3.54 37 98 -3.21 68

77 -3.52 37.5 106 -3.2 69

89 -3.51 39 112 -3.2 69

102 -3.51 39 118 -3.2 69

108 -3.5 41 124 -3.18 66.6

117 -3.5 41

121 -3.5 41

154

2 % Agarose 1 % Agarose

Time (min)

Wavelength exponent

Correlation Length (nm)

Time (min)

Wavelength exponent

Correlation Length (nm)

48 -4.05 8 25 -5.63 0

53 -3.68 30 37 -4.96 0

58 -3.41 47 44 -4.09 0

66 -3.15 74.5 51 -3.52 39

73 -3.05 96 57 -3.21 63

78 -3 110 64 -2.98 104

83 -2.96 114 71 -2.84 156

88 -2.93 121 77 -2.72 225

94 -2.9 128.5 83 -2.65 347

101 -2.88 135.7 102 -2.54 400

106 -2.87 139.3 108 -2.51 503

111 -2.86 142.8 113 -2.5 505

120 -2.85 146.3 118 -2.49 517

124 -2.84 149.8 124 -2.47 540

155

2 % Agarose with 0.3 % polysebacic anhydride microparticles

Time (min)

Wavelength exponent

Correlation Length (nm)

21 -3.2 65

28 -2.97 100

33 -2.89 132

38 -2.85 151

43 -2.83 164

50 -2.81 175

56 -2.8 179

65 -2.79 184

73 -2.79 184

78 -2.78 190

156

y = 530.5e-37.47x R² = 1

y = 530.5e-195.3x R² = 1

y = 530e-416.3x R² = 0.3982

y = 530.5e-1735x R² = 0.9656

0

100

200

300

400

500

600

0 0.001 0.002 0.003 0.004 0.005 0.006 0.007

Eq

uilib

riu

m Y

ou

ng

's m

od

ulu

s,

EY,

kP

a

Surface areaof nanoparticles in suspension, m2/mL

FH

H9

H38

H150

C. Curve fitting of mechanical properties of engineered cartilage tissues and monolayer

coverage of sodium laurate as a function of size of ferric (hydr)oxide nanoparticles.

Mechanical properties of tissues

The dependence of Young’s modulus of cartilage tissues Figure 7.1a on the total surface area of

nanoparticles in the scaffold is modeled using exponential functions (Figure C1). Except for the H38, the

fitting error was negligible. The mechanical strength of tissue constructs at nanoparticle concentration of

0.002 m2/mL were obtained from the exponential function. The concentration was chosen arbitrarily. For

H38 particles, the experimentally measured value of Young’s modulus at 0.002 m2/mL was used. The

values are tabulated in Table C1. The detrimental catalytic activity of ferric (hydr)oxide is reported as

decrease in Young’s modulus from the control value (scaffolds without nanoparticles = 530.5 kPa). This

decrease is normalized with the decrease for H150 for comparison purposes.

Figure C1 Decrease of the Young’s modulus of tissue constructs as a function of surface area per of the

ferric (hydr)oxide nanoparticles per unit volume in the constructs. FH-4nm; H9-9nm; H38-38nm and

H150-150nm.

157

y = 0.048x0.63 R² = 0.8557

0

0.2

0.4

0.6

0.8

1

1.2

0 50 100 150 200

No

rmal

ize

d d

ecr

eas

e in

Yo

un

g's

mo

du

lus

of

tiss

ue

co

nst

ruct

s

Size of ferric (hydr)oxide nanoparticles, nm

Table C1 Decrease in Young’s modulus of tissue constructs due to detrimental catalytic activity of ferric

(hydr)oxide nanoparticle: effect of nanosize

Ferric (hydroxide) nanoparticles size, nm

Young's modulus values obtained at 0.002 m

2/mL

from exponential curves fitted on Figure 7.1a

Decrease in Young's modulus due to nanoparticles from control value (530.5 kPa)

Decrease in Young's modulus due to nanoparticles normalized with the maximum value (H150: 514 kPa)

4 492 39 0.08

9 359 172 0.33

38 300 231 0.45

150 16.5 514 1

A power law fit of the size of ferric (hydr)oxide nanoparticles and their bioactivity (represented by the

normalized decrease in Young’s modulus is shown in Figure C2.)

Catalytic bioactivity (normalized decrease in Young’s modulus) = 0.048. (size)0.63

Figure C2 Normalized decrease in Young’s modulus of tissue constructs as a function of size of the ferric

(hydr)oxide nanparticles present in the tissue constructs at concentration of 0.002 mg/mL

158

y = 0.42x0.19 R² = 0.9922

0

0.2

0.4

0.6

0.8

1

1.2

0 50 100 150 200

No

rmal

ize

d m

on

ola

yer

cove

rage

o

f so

diu

m la

ura

te

Size of ferric (hydr)oxide nanoparticles, nm

Size-dependent adsorption activity

The monolayer coverage of sodium laurate on ferric (hydr)oxide nanoparticles with respect size is given

in the Table C.2.

Table C.2 Normalized monolayer coverage of sodium laurate on ferric (hydr)oxide nanoparticles

Size of ferric (hydr)oxide nanoparticles, nm

Monolayer coverage of sodium laurate at pH 7.1, molecules/nm

2

Normalized monolayer coverage with the maximum value (H150: 4.2 molecules/nm

2)

4 2.1 0.5 9 2.6 0.619048 38 3.3 0.785714 150 4.2 1

A power law fit was obtained between size of the nanoparticles and the normalized values of monolayer

coverage (with that of maximum coverage value of 150 nm particles: 4.2 molecules/nm2).

Normalized monolayer coverage of sodium laurate, y = 0.4 . (size)0.19

Figure C3 Normalized monolayer coverage of laurate as a function of size of the ferric (hydr)oxide

nanoparticles