Magnetic Force Microscopy of Superparamagnetic Nanoparticles

for Biomedical Applications

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of

Philosophy in the Graduate School of The Ohio State University

Tanya M. Nocera, M.S.

Graduate Program in Biomedical Engineering

The Ohio State University

2013

Dissertation Committee:

Gunjan Agarwal, PhD, Advisor

Stephen Lee, PhD

Jessica Winter, PhD

Anil Pradhan, PhD

Tanya M. Nocera

2013

Abstract

In recent years, both synthetic as well as naturally occurring superparamagnetic nanoparticles (SPNs) have become increasingly important in biomedicine. For instance, iron deposits in many pathological tissues are known to contain an accumulation of the superparamagnetic protein, ferritin.

Additionally, man-made SPNs have found biomedical applications ranging from cell-tagging in vitro to contrast agents for in vivo diagnostic imaging. Despite the widespread use and occurrence of SPNs, detection and characterization of their magnetic properties, especially at the single-particle level and/or in biological samples, remains a challenge. Magnetic signals arising from SPNs can be complicated by factors such as spatial distribution, magnetic anisotropy, particle aggregation and magnetic dipolar interaction, thereby confounding their analysis.

Techniques that can detect SPNs at the single particle level are therefore highly desirable.

The goal of this thesis was to develop an analytical microscopy technique, namely magnetic force microscopy (MFM), to detect and spatially localize synthetic and natural SPNs for biomedical applications. We aimed to (1) increase

MFM sensitivity to detect SPNs at the single-particle level and (2) quantify and spatially localize iron-ligated proteins (ferritin) in vitro and in biological samples using MFM.

Two approaches were employed to improve MFM sensitivity. First, we showed how exploitation of magnetic anisotropy could produce a higher, more uniform MFM signal from single SPNs. Second, we showed how an increase in probe magnetic moment increased both the magnitude and range up to which the MFM signal could be detected from a single SPN.

We further showed how MFM could enable accurate quantitative estimation of ferritin content in ferritin-apoferritin mixtures. Finally, we demonstrated how MFM could be used to detect iron/ferritin in serum and animal tissue with spatial resolution and sensitivity surpassing that obtained using conventional biochemical assays.

We envisage these advancements will allow MFM to serve as a novel biosensing technique to spatially localize iron/ferritin in small aliquots of clinical samples (i.e. serum) and in tissue biopsies at the ultra-sensitive and ultrastructural level. We also discuss how future work incorporating our advancements could lead to the development of a novel indirect MFM technique, which could enable high-throughput analysis of SPNs for biomedical applications.

iii

Dedication

To my mother, Joni, who has always encouraged me to follow my heart and

reach for my dreams, and to my father, Kevin, who has given me practical

guidance and a level head to make those dreams come true.

To Jason, for keeping me grounded, for never letting me give up and for

“(un)quantifiably” loving me no matter what.

To Dr. Doros Petasis, for amplifying my love for science, teaching me to “build

character” and giving me confidence to pursue my Ph.D.

Acknowledgements

I would like to thank my advisor, Dr. Gunjan Agarwal, for calling me on

April Fool’s Day (and it not be a joke) to give me the opportunity to pursue my doctoral degree. This has been the most incredible and most challenging journey, and my scientific and academic successes would not have been possible without her continuous guidance and encouragement. I would also like to thank the members of my dissertation committee, Dr. Stephen Lee and Dr.

Jessica Winter, and my candidacy committee, Dr. Chris Hammel and Dr. Vish

Subramaniam. Their support, critiques and advice have truly made me a better scientist, and for that I am grateful.

A number of individuals have contributed their time and expertise to make the work in this dissertation possible. I would like to thank my co-authors Dr.

Christopher Murray and Dr. Jun Chen from the University of Pennsylvania for providing the synthetic Fe3O4 SPNs and SQUID/ZFC characterizations for my anisotropy and probe comparison studies. I am also grateful to Ed Calomeni and

Henk Colijn for the generous amount of time they took to teach me electron microscopy and for their thoughtful insights during discussions of my projects.

Thank you to BME undergraduate Yuzhi (Kevin) Zeng, who helped co-author and share in my enthusiasm for the ferritin MFM project. Kevin optimized ferritin/apoferritin sample preparation and assisted with MFM data collection. I would also like to thank Dr. Dana McTigue and Dr. Andrew Sauerbeck for their contributions to the work involving rat spinal cord tissue and serum. They generously provided the tissue/serum samples, performed Perls and immunohistochemistry studies and conducted serum ferritin experiments. I am also grateful to Dr. John Moreland from NIST for providing microfluidic silicon nitride membranes for our indirect MFM pilot studies.

To my past labmate, Dr. Angela Blissett and my current labmates, Jeff

Tonniges and David Yeung: your unending support, advice and breaks have kept me sane and helped me to push through. Thank you for your willingness to help me whenever you could and for making my experience much more enjoyable. I am especially grateful to my softball and Hamptons crew, broomball team and fellow graduate school friends; you have helped me to find many outlets for stress and the balance I desperately needed to face the challenges of grad school one week at a time. Thank you to my furry best friend, Tesla, for appearing on my doorstep two years ago and changing my life. Lastly, I would like to thank Panera Bread for providing bottomless coffee and tea, sometimes two or three meals in a day, free WIFI and kindness. I very well may have never finished writing my papers, proposals and this dissertation without your help.

Vita

August 1987 ...... Born,

Butler, PA

June 2005 ...... Butler Senior High School,

Butler, PA

May 2009 ...... B.S. Physics,

Allegheny College, Meadville, PA

December 2011 ...... M.S. Biomedical Engineering,

The Ohio State University, Columbus, OH

August 2009 to August 2012…………………...….AFM Core Laboratory Manager,

Davis Heart and Lung Research Institute, Columbus, OH

August 2012 to May 2013………………………..…..Graduate Teaching Associate,

Department of Biomedical Engineering,

The Ohio State University, Columbus, OH

June 2009 to Present………..……………………….Graduate Research Associate,

Department of Biomedical Engineering,

The Ohio State University, Columbus, OH

vii

Publications

Included in Dissertation Chapter 3:

Nocera, T.M., Chen, J., Murray, C.B., Agarwal, G., “Magnetic Anisotropy

Considerations in Magnetic Force Microscopy Studies of Single

Superparamagnetic Nanoparticles,” Nanotechnology 2012; 23 (49): 495704.

Included in Dissertation Chapter 4:

Nocera, T.M., Sauerbeck, A., Zeng Y., McTigue, D. M., Agarwal, G.,

“Quantification and Spatial Localization of Ferritin Using Magnetic Force

Microscopy,” Revised and resubmitted to Nanomedicine: Nanotechnology,

Biology and Medicine.

Additional Publications:

Stevenson, M.D., Pristine, H., Hogrebe, N.J., Nocera, T. M., Boehm, M., Reen,

R., Koelling, K., Agarwal, G., Saraang-Sieminski, A.L., Gooch, K.J., “Self- assembling Peptide Matrix that Independently Controls Stiffness and Binding

Site Density Supports the Formation of Microvascular Networks in 3D,” Acta

Biomaterialia 2013; 9(8): 7651-7661.

viii

Agarwal, G., and Nocera, T.M., "Atomic Force Microscopy (AFM)." In The

Nanobiotechnology Handbook. Ed.Yubing Xie. Boca Raton: CRC Press, 2012;

369-91.

Hilfiger, M.G., Chen, M., Brinzari,T.V., Nocera, T.M., Shatruk, M., Petasis, D.T.,

Musfeldt, J.L., Achim, C. Dunbar, K.R., “An Unprecedented Charge Transfer

Induced Spin Transition in an Fe–Os Cluster,” Angewandte Chemie International

Edition 2010; 49(8): 1410–1413.

Fields of Study

Major Field: Biomedical Engineering

Concentrations: Biomedical Micro/Nanotechnology,

Biomedical Imaging

Table of Contents

Abstract ...... ii

Dedication ...... iv

Acknowledgements ...... v

Vita ...... vii

List of Tables ...... xii

List of Figures ...... xiii

List of Abbreviations ...... xv

List of Symbols ...... xviii

Chapter 1: Superparamagnetic Nanoparticles in Biomedicine ...... 1

1.1 Introduction ...... 1

1.2 Superparamagnetic Nanoparticles (SPNs) ...... 3

1.3 Synthetic SPNs in Biomedicine ...... 11

1.4 Natural SPNs in Biomedicine ...... 16

1.5 Detection and Characterization of SPNs ...... 19

1.6 Limitations in Current SPN Detection Techniques ...... 24

Chapter 2: Magnetic Force Microscopy ...... 26

2.1 Introduction ...... 26

2.2 Magnetic Force Microscopy (MFM) ...... 26

2.3 Forces in MFM Probe-Sample Interactions ...... 29

2.4 MFM Probes ...... 32

2.5 MFM in Biomedicine: Present and Future ...... 33

2.6 Thesis Aims ...... 35

Chapter 3: Detection of Single SPNs Using MFM ...... 36

3.1 Aims and Rationale ...... 36

3.2 Materials and Methods ...... 37

3.3 Results ...... 43

3.4 Discussion ...... 55

Chapter 4: Quantification and Spatial Localization of Ferritin Using MFM . 59

4.1 Aims and Rationale ...... 59

4.2 Materials and Methods ...... 60

4.3 Results ...... 65

4.4 Discussion ...... 73

Chapter 5: Conclusions and Future ...... 76

5.1 Chapter Overview ...... 76

5.2 Summary of MFM Advancements ...... 76

5.3 Future Outlook: High-throughput MFM ...... 79

5.4 Conclusions: MFM in Biomedicine ...... 83

References ...... 85

Appendix A: Reprint Permissions ...... 116

List of Tables

CHAPTER 1

Table 1.1 Common Types of Magnetic Behaviors in Bulk Materials ...... 5

CHAPTER 2

Table 2.1 Properties of Commercial MFM Probes ...... 33

CHAPTER 3

Table 3.1 Medium Moment (MM) and High Moment (HM) Probe Parameters ... 54

xii

List of Figures

CHAPTER 1

Figure 1.1 Schematic of Magnetic Domains ...... 4

Figure 1.2 Magnetic Material Behaviors in an Applied Magnetic Field ...... 6

Figure 1.3 Superparamagnetic Nanoparticle (SPN) Isotropy and Anisotropy ...... 8

Figure 1.4 Effects of Temperature and Volume on Magnetic Behavior ...... 10

Figure 1.5 SPIONs for Multimodal Imaging and/or Multifunctional Purposes .... 12

Figure 1.6 Schematics of Apoferritin and Ferritin Proteins ...... 18

CHAPTER 2

Figure 2.1 Schematics of MFM Probe and its Paths for Image Acquisition ...... 27

Figure 2.2 Models for Calculation MFM Probe-Sample Interactions ...... 30

CHAPTER 3

Figure 3.1 TEM and MFM Sample Preparation and MFM Set-up for Randomly

Oriented and Pre-aligned SPN Samples ...... 40

Figure 3.2 Composition and Morphology of SPNs ...... 45

xiii

Figure 3.3 Magnetic Properties of SPN Ensembles ...... 46

Figure 3.4 Identification of Single SPNs for MFM Analysis ...... 48

Figure 3.5 MFM Phase Shift Analysis of Single SPNs ...... 51

Figure 3.6 MFM Analysis of Randomly Oriented and Pre-aligned SPNs ...... 52

Figure 3.7 MFM Phase Shifts of SPN Singlets Using High Moment and Medium

Moment Probes ...... 55

CHAPTER 4

Figure 4.1 Morphological Characterizations of Apoferritin and Ferritin ...... 66

Figure 4.2 Quantitative MFM Analysis of Apoferritin and Ferritin ...... 67

Figure 4.3 Quantitative MFM Analysis of Apoferritin/Ferritin Mixtures ...... 69

Figure 4.4 MFM of Endogenously Expressed Iron/Ferritin in Serum ...... 71

Figure 4.5 Detection of Endogenously Expressed Iron/Ferritin in Tissue ...... 72

Chapter 5

Figure 5.1 Schematic of Indirect Magnetic Force Microscopy ...... 80

Figure 5.2 Indirect MFM of Ferritin ...... 82

xiv

List of Abbreviations

AC alternating current

AFM atomic force microscopy

°C degrees Celsius

CT computed tomography

D diamagnetic

DI distilled

DMR diagnostic magnetic resonance

ED electron diffraction

EDS energy dispersive (x-ray) spectroscopy

EELS electron energy loss spectroscopy

ELISA enzyme-linked immunosorbent assay

EM electron microscopy emu electromagnetic unit

EPR enhanced permeability and retention eV electron volts

F ferromagnetic

FFT fast Fourier transform

FI ferrimagnetic g grams

G Gauss

HM high magnetic moment MFM probe

HRTEM high resolution transmission electron microscopy

Hz Hertz

ID-MFM indirect magnetic force microscopy

K Kelvin l liter m meter

MFM magnetic force microscopy

MM medium magnetic moment MFM probe

MRI magnetic resonance imaging n number

NIH National Institutes of Health

Oe Oersted

P paramagnetic

PET position emission tomography

RES reticuloendothelial system

SEM scanning electron microscopy

SP superparamagnetic

xvi

SPIONS superparamagnetic iron oxide nanoparticles

SPN superparamagnetic nanoparticle

SQUID superconducting quantum interference device

STEM scanning transmission electron microscope

T Tesla

TEM transmission electron microscopy

V Volts

ZFC zero-field cooling

xvii

List of Symbols

A Amplitude nanometers (nm)

B applied magnetic field Gauss (G) or Tesla (T)

BMFM B applied during MFM Gauss (G) or Tesla (T)

-1 -1 Bnet net magnetism Amperes meter (A m )

Bsample B applied during sample prep Gauss (G) or Tesla (T) c oxidation prevention thickness nanometers (nm)

ρ density Kilograms meters-3 (kg m-3) d diameter nanometers (nm)

DA areal density particles per unit area

δφ phase shift degrees (°)

F force Newtons (N)

K anisotropy constant Joules meters-3 (J m-3) k spring constant Newtons meter-1 (N m-1)

-1 -1 kB Boltzmann constant Joules Kelvin (J K )

-1 -1 keff effective spring constant Newtons meter (N m ) m Mass grams (g)

xviii

-1 μo permeability constant (Volts seconds) (Ampere meter) (Vs (Am)-1) -1 -1 Mp probe saturation magnetization Amperes meter (A m )

2 2 mp probe magnetic moment Amperes meters (A m )

-1 -1 Ms saturation magnetization Amperes meter (A m )

2 2 ms sample magnetic moment Amperes meters (A m )

-1 -1 MVS volume saturation magnetization electromagnetic unit gram (emu g )

Q quality factor unitless r SPN radius nanometers (nm)

R MFM probe radius nanometers (nm)

R1 uncoated probe radius nanometers (nm)

T Temperature Kelvin (K)

TB blocking temperature Kelvin (K)

τN Néel relaxation time seconds (s)

θ angle between easy axis and B degrees

V Volume nanometers3 (nm3)

3 3 Vc critical volume nanometers (nm )

ωD drive frequency kilohertz (kHz)

ωr resonance frequency kilohertz (kHz) z lift height nanometers (nm)

xix

CHAPTER 1

Superparamagnetic Nanoparticles in Biomedicine

1.1 Introduction

In recent years, both synthetic and naturally occurring superparamagnetic nanoparticles (SPNs) have become increasingly important in biomedicine. Due to their small size and unique magnetic properties (see section 1.2), man-made

SPNs are used for applications including tagging of cells in vitro and as contrast agents for in vivo diagnostic imaging (section 1.3). Further, certain iron-bound proteins such as ferritin have been identified as superparamagnetic in nature

(section 1.4). Ferritin levels serve as a biomarker for monitoring iron deficiency or iron overload. Because of the wide-spread occurrence and use of both man- made and naturally occurring SPNs in biological systems, it is critical to enable high-fidelity detection of SPNs. Such a capability will enable us to better understand the distribution, fate and concentration of SPNs in biological samples.

This knowledge could lead to improved design of SPNs for cell-tagging and a better understanding of iron homeostasis in vivo.

1.1.1 Significance

Current techniques to detect or characterize SPNs in biological systems are based on immuno-detection, iron detection, or magnetic detection (see section 1.5). These techniques typically provide an ensemble average of SPN properties and/or fail to detect or spatially resolve SPNs at the single particle level. We propose magnetic force microscopy (MFM) as a novel and simple solution; MFM can exploit the magnetic properties of SPNs and enable their detection in biological samples with high-sensitivity and spatial localization.

1.1.2 Thesis Aims

The overall goal of this thesis was to develop MFM for high fidelity detection of SPNs encountered in biological systems. Our central hypothesis was that MFM can reveal novel insights into the magnetic properties of SPNs and serve as an ex vivo tool to estimate iron content in biological samples. Two major aims were pursued:

(1) increase the sensitivity of MFM to detect and characterize SPNs at the

single particle level and

(2) quantify and spatially localize ferritin in vitro and in biological samples

using MFM.

These advancements reveal the potential of MFM as a novel bio-sensing tool for ex vivo detection of SPNs with very high sensitivity and spatial resolution.

1.1.3 Thesis Overview

In this chapter we discuss the properties, roles and composition of synthetic and naturally occurring SPNs, and review the current techniques for

SPN detection. In chapter 2 we provide technical details on the MFM technique.

Our results pertaining to aims 1 and 2 outlined above are described in chapters 3 and 4, respectively. In chapter 5 we conclude with the significance of our results and discuss the future potential of MFM for biomedical applications.

1.2 Superparamagnetic Nanoparticles (SPNs)

Magnetic materials are classified by their unique magnetic behaviors demonstrated in the absence or presence of an externally applied magnetic field,

B. They can be broadly categorized as diamagnetic, paramagnetic, ferromagnetic or ferrimagnetic (Table 1.1). The magnetic properties of the materials are governed in part by the existence of multiple magnetic domains, or areas in which the magnetic moments are parallel and thus the magnetization is regionally uniform (Figure 1.1). However, as the size of a material is reduced below 100nm, the same materials can exhibit significantly different magnetic properties. One of these properties, relevant in biomedicine, is superparamagnetism.

Superparamagnetism usually occurs in ferromagnetic or ferrimagnetic materials below a material-dependent critical volume, Vc, at which the particle possesses only one magnetic domain (Figure 1.1). At room temperature and in

3 the absence of an applied magnetic field, the single magnetic domain in the SPN will continuously rotate or flip directions. This results in a net magnetization, Bnet, of zero. Obtaining stable magnetization in a SPN at thermal equilibrium therefore requires the application of an external magnetic field along which the single magnetic domain can align.

Magnetic Domains

Magnetic Material SPN

Figure 1.1 Schematic of Magnetic Domains. Magnetic materials can possess multiple magnetic domains, or areas in which magnetic moments (red arrows) are regionally parallel. When the material is reduced below a critical size, it possesses only one magnetic domain. Single-domain particles (above a critical temperature) exhibit unique magnetic behaviors and are classified as superparamagnetic nanoparticles (SPNs).

Table 1.1: Common Types of Magnetic Behaviors in Bulk Materials Type of Magnetic Behavior for Common Magnetism Bulk Material, > 100nm Examples

B Super- Magnetic moments conductors, are random when Diamagnetic water, B = 0; copper, lead, B antiparallel to B Bnet net zinc

B Magnetic moments Aluminum, are random when Paramagnetic gold, B = 0; manganese B parallel to B Bnet net

B Permanent magnets; Iron, Nickel Magnetic moments Ferromagnetic Cobalt, align parallel to each Fe O other even in B = 0 2 3 Bnet

B Magnetic moments are antiparallel with Ferrites, Ferrimagnetic different magnitudes, Fe O 3 4 Bnet parallel to B Bnet

B is the externally applied magnetic field Bnet is the resulting net magnetism of material

SPNs are extensively used in biomedical applications due to their small size and unique magnetic properties. Like paramagnetic and diamagnetic materials, the magnetization of a SPN can be selectively “turned on” in the presence or “turned off” in the absence of an applied magnetic field, B. Similar to ferromagnetic materials, SPNs can also align perfectly along the direction of B, resulting in a large net magnetization, Bnet (Table 1.1 and Figure 1.2). Synthetic 5 ferromagnetic materials are less ideal for many biomedical applications because they express remnant magnetization even at B=0, which could lead to the formation of large, and therefore toxic, aggregates.

Figure 1.2 Magnetic Material Behaviors in an Applied Magnetic Field.

Ferromagnetic (F) materials have high magnetization (y-axis) along an applied magnetic field, B (x-axis), which is present even when B=0. Diamagnetic (D) materials have weak magnetization opposing B and paramagnetic (P) materials have weak magnetization along the direction of B; neither of these materials are magnetic in the absence of B. Superparamagnetic (SP) materials exhibit strong magnetization in an applied field, yet lack remnant magnetization when B =0.

Reprinted with permission from AANS and the Journal of Neurosurgery [1]. 6

The unique magnetic behaviors of SPNs are further explained in the following subsections with regard to factors such as magnetic anisotropy, Néel relaxation time and blocking temperature.

1.2.1 Magnetic Anisotropy

There are several types of anisotropy that may be associated with a SPN, resulting from its crystalline structure, particle shape, or stress. These anisotropies can individually or collectively induce magnetic anisotropy in the

SPN (Figure 1.3). In such a situation, the SPN possesses an easy axis, or an axis along which the magnetic moment prefers to align. In the presence of B, the magnetic anisotropy plays a critical role in determining the extent to which the

SPN magnetic moment can align with B.

A SPN that is free to rotate, such as those in a colloidal solution, can align its easy axis and in turn possess a stable magnetic moment along B. The ability of an immobilized SPN to align its magnetic moment with B, however, depends not only on the applied magnetic field strength but also on the angle, θ, between the SPN easy axis and the direction of B (See Figure 1.3D). This relationship can be demonstrated by [2],

ெ ஻ ߠ ൑ ܽݎܿݏ݅݊ ቀ ೞ ቁ, (Eq 1.1) ௠௔௫ ଶ௄௏ where ܯ௦ is the saturation magnetization of a SPN, ܭ is the material-dependent anisotropy constant, ܸ is the particle volume and ߠ௠௔௫ is the maximum possible angle at which the SPN magnetic moment can align along B. If ߠ൐ߠ௠௔௫ the

7 magnetic moment of an immobilized SPN will fail to align completely along B.

The shape and crystalline anisotropy of SPNs can be ascertained using techniques like electron microscopy [3,4] and x-ray diffraction [5]. However, very limited techniques exist to assess magnetic anisotropy in individual or an ensemble [6] of SPNs.

1 1 A 1 B 1 2 C 2 D

1 3 θ 3

Easy axis B Shape and Shape Crystalline Magnetic Crystalline Anisotropy Anisotropy Anisotropy Isotropy

Figure 1.3 Superparamagnetic Nanoparticle (SPN) Isotropy and Anisotropy

SPNs can be (A) isotropic in both shape and crystalline structure, or (B) anisotropic in shape and/or (C) anisotropic in crystalline structure. The dashed lines represent possible axes along which the SPN properties are either inherently identical (A) or different (B-C). Shape and/or crystalline structure

(and/or stress) on the SPN may result in (D) magnetic anisotropy, where the SPN magnetic moment (double arrow) preferentially aligns along an easy axis

(dashed gold line). The ability of the magnetic moment of a SPN to align with an external magnetic field, B, depends on both the strength of B and size of the angle, θ, between its easy axis and B.

1.2.2 Néel Relaxation Time

The Néel relaxation time ሺ߬ேሻ, or time it takes for the magnetic moment of a SPN to flip between its two anti-parallel directions along the easy axis, can be described using the Néel-Arrhenius model [7],

಼ೇ ൬ ൰ ೖ ೅ ߬ே ൌ߬௢݁ ಳ , (Eq 1.2) where ߬௢ is a material-dependent constant, K is the anisotropy constant, V is the particle volume, ݇஻ is the Boltzmann constant, and T is temperature. The relaxation time decreases exponentially as particle volume decreases and/or when temperature is increased. A magnetic nanoparticle below a critical volume

(Vc) and at room temperature may therefore have a negligible Néel relaxation time and in turn an unstable magnetic moment, classifying it as a SPN.

Néel relaxation time can be observed experimentally by aligning a sample’s magnetic moment(s) with an applied magnetic field, then turning off the applied field and measuring the time it takes for the magnetic moment(s) to relax back to their pre-magnetization state. This process is commonly observed using superconducting quantum interference device (SQUID) relaxometry or alternating current (AC) susceptometry for materials with ߬ே in milliseconds [8–10]. The ߬ே of

SPNs is too small to be measured experimentally and is therefore usually calculated based on known particle composition and volume. For example, the naturally occurring superparamagnetic protein, ferritin, with iron-core diameter of

~6nm has an estimated ߬ே~ 0.01ns [11].

1.2.3 Blocking Temperature

Blocking temperature (ܶ஻) is the critical temperature below which a SPN will transition from its superparamagnetic state (unstable magnetic moment) to its ferromagnetic state (stable magnetic moment). The ܶ஻ is dependent on the particle volume and the material’s anisotropy constant (ܭ). It is usually determined experimentally with SQUID magnetometry. Figure 1.4 summarizes the effects of temperature and volume on the magnetic behavior of a material.

F, FI SP

TB Temperature F, FI F Vc

Particle Volume

Figure 1.4 Effects of Temperature and Volume on Magnetic Behavior

A ferromagnetic (F) or ferrimagnetic (FI) material will become superparamagnetic

(SP) below its critical volume (ܸ௖) and above its blocking temperature (ܶ஻). The exact values of ܸ௖ and ܶ஻ are material-dependent.

1.3 Synthetic SPNs in Biomedicine

Synthetic SPNs employed in biomedicine are usually composed of Fe2O3 or Fe3O4, and therefore are typically referred to as superparamagnetic iron oxide nanoparticles (SPIONS) [12]. Aside from their magnetic properties, SPIONs are attractive because they can be cleared via ordinary iron metabolism processes involving the reticuloendothelial system (RES) [13]. It is common practice to coat

SPIONs with either inorganic materials like silica or gold or with organic polymers or peptides to minimize SPION aggregation and increase biocompatibility [14].

Depending on the application, SPIONs can range from 5nm to 50nm (iron core) with an outer coating thickness up to 160nm [12]. They may be functionalized with various labels for multimodal imaging applications and/or with various therapeutic agents to give them multifunctional purposes (Figure 1.5).

There are some synthetic SPIONs commercially available for biomedical applications. These include micron-sized Dynabeads® from Life Technologies and 50nm MACS MicroBeads from Miltenyi Biotec, both of which are marketed for applications in cell tagging and separation. The federal drug administration

(FDA) has also approved intravenously injectable colloid solutions of SPIONs for clinical use. Feraheme (or ferumoxytol), contains 17-31nm SPIONs to treat iron- deficient anemia in patients with chronic kidney disease [15]. There are also two

FDA-approved dextran-coated SPIONs for use as magnetic resonance imaging

(MRI) contrast agents, namely Feridex (or ferumoxides) and Resovist (or ferucarbotran). Feridex contains SPIONs with particle diameters between 120nm

11 and 180nm, while Resovist contains particle diameters of approximately 60nm

[16]. The following subsections describe the potential for synthetic SPNs in imaging and therapy-based biomedical applications.

MRI

Figure 1.5 SPIONs for Multimodal Imaging and/or Multifunctional Purposes.

Superparamagnetic iron oxide nanoparticles (SPIONs), typically Fe2O3 or Fe3O4, can be labeled with various agents for magnetic resonance imaging (MRI), computed tomography (CT) or fluorescence imaging and/or used to deliver drugs, siRNA or hyperthermia treatment to a targeted cell, tissue or tumor site.

Adapted and reprinted with permission from [17].

1.3.1 Methods for SPION Targeting and Delivery

In biomedicine, it is important to direct the SPN to a desired site for imaging and/or therapeutic applications. Cells in vitro can achieve non-specific

SPION uptake without the assistance of SPION coatings [18–20] and/or without targeting surface modifications [21–23]. Specific targeting is attained by functionalizing the SPION surface and/or by varying the SPION diameter and thickness of the SPION coating.

One of the most common methods for SPION delivery is to functionalize the SPION with antibodies that have a specific affinity for its intended target.

These immuno-labeled SPIONs have found several applications in cardiovascular and cancer imaging and therapy both in vitro and in vivo [24–31].

Aside from antibodies, SPIONs may also be labeled with ligands that are attracted to receptors on cell surfaces; these targets have included folate, biphosphate and integrin receptors [32–34].

Solid tumor tissues have distinct properties that can be exploited as another method for increasing SPION delivery in vivo [35]. These properties include an increase in vasculature that is leaky, defective and abnormal in structure, which permits SPIONs to accumulate in concentrations 10 to 50-fold greater than within healthy tissue [36]. Tumors also have poor lymphatic drainage and an impaired clearance mechanism, because of which the SPIONs can remain in the tumor region for prolonged periods. This exploitation of tumor- specific properties is also referred to the Enhanced Permeability and Retention

(EPR) effect.

Diameter is a critical component for ensuring EPR-based delivery of

SPIONs to a tumor site; it must be large enough to avoid uptake by healthy

13 tissue and clearance by the kidneys (> 5nm -10nm) yet smaller than the diameter limit for uptake by tumor tissue (up to 2μm depending on the tumor type) [37].

Many studies have shown optimal EPR effect to occur using particles with diameters between 100nm and 200nm [37–40], which can typically be achieved by controlling the SPION diameter and/or coating thickness.

1.3.2 SPIONs for Multimodal Biomedical Imaging

Magnetic resonance imaging (MRI) signal is rooted in the behavior of a sample’s magnetic moments as they are influenced by an applied magnetic field,

B [41]. A material with magnetic properties disrupts the homogeneity of B, which in turn changes the MRI signal intensity and produces image contrast at that specific location. Biological tissue is primarily diamagnetic and therefore minimally disrupts B [1]. This results in weak contrast, and in turn fuels the need for contrast agents of higher magnetic susceptibility. Paramagnetic materials such as gadolinium and copper have been most commonly used as MRI contrast agents in the clinic [1], however their magnetic susceptibilities are only slightly greater than tissue. The high susceptibility of SPIONs gives them potential to locally disrupt B on a scale much greater than their size [42]. This has led to a surge in their use as contrast agents for imaging tumors, tissue and even cells

[17,43–47].

Besides MRI, SPIONs are an excellent platform for providing contrast for multiple imaging modalities because of their intrinsic magnetic properties and

14 their ability to be coated or functionalized. Radionucleotides have been coupled to SPIONs for simultaneous position emission tomography (PET) and MRI for dual anatomical and functional imaging in vivo[48–51]. SPIONs labeled with gold nanoparticles or coated with a gold shell [52,53] have also served as dual contrast agents for computed tomography (CT) and MRI for enhanced anatomical imaging in vivo. A third multimodal imaging approach combines MRI with optical imaging. Here, the SPION is conjugated with a fluorescent dye or combined with fluorescent quantum dots [32,54–57] to simultaneously enhance the visualization of anatomical features (MRI) and molecular events like gene expression or protease activity (optical imaging).

1.3.3 Hyperthermia Therapy

Tumors are more susceptible to heat than healthy tissue, with survival rates rapidly decreasing above 42ᵒC [58]. Since SPIONs can be delivered to a tumor site either via surface functionalization or the EPR effect, they are being utilized for localized hyperthermia treatment confined specifically to the targeted tumor site. This application is based on the ability of the magnetic moments of

SPNs to be “turned on/off” and flipped by externally applied magnetic fields.

SPNs generate heat due to Néel relaxation processes when the direction of their magnetic moments is continuously flipped by a rapidly oscillating (AC) magnetic field [59]. Most in vivo hyperthermia studies target a SPION-induced tumor site temperature between 42ᵒC and 45ᵒC [22,23,58,59], however temperatures

15 approaching 60ᵒC have also been reported [59]. Hyperthermia therapy is a multifunctional application of SPIONs because it permits both magnetic-based imaging of the tumor site and therapeutic treatment.

1.3.4 Drug Delivery

Multifunctional SPIONs are being used to deliver drug or gene therapy to targeted areas. Common chemotherapy drugs have been loaded into the coating of SPIONs, conjugated to the surface of the SPION coating, co-encapsulated with SPIONs inside micelles or linked with SPIONs via electrostatic interactions

[55,60–62]. The SPION-drugs are then delivered to the tumor site via antibody targeting, receptor targeting or the EPR effect, which can reduce side effects and increases drug efficacy. Depending on how the drug is incorporated, it may be released via methods such as coating degradation or particle disassembly triggered by the naturally acidic tumor environment. siRNA has also been incorporated into or conjugated to the coating of SPIONs for applications in targeted gene therapy [63–66]. Since SPIONS are compatible with MRI imaging, the delivery and efficacy of these drugs can be visually monitored [60–62].

1.4 Natural SPNs in Biomedicine

At least three iron-storage proteins that occur naturally in human physiology are known to be superparamagnetic in nature. The most prevalent among these is ferritin, which is expressed intracellularly and secreted by cells

16 into serum and a variety of tissues [67]. Another protein, hemosiderin, is thought to be a variant of ferritin, but only participates in iron storage within cells [68]. The third natural SPN, hemozoin, consists of iron-oxide that has mineralized within malaria-infested mosquitoes. Hemozoin is found in the blood of individuals who have been infected with malaria [69]. The superparamagnetic properties of hemosiderin and hemozoin have not been as well-characterized as those of ferritin.

1.4.1 Ferritin and its Role in Pathology

Ferritin is a hollow, globular protein with outer shell diameter, d~12nm

[70]. It is widely expressed in a variety of tissues such as the brain, intestines, spleen, liver and bone marrow, and exists both as an intracellular and as a secreted protein. Ferritin aids in protecting the body from oxidative stress by reducing and sequestering iron from freely interacting with cells and tissue. It participates in iron homeostasis by binding and storing up to 4500 iron atoms as superparamagnetic ferrihydrite (Figure 1.6) [71,72]. Ferritin’s iron core has been characterized by its magnetic properties using techniques like superconducting quantum interference device (SQUID) magnetometry and Mössbauer spectroscopy [73].

Non-Iron-Ligated Protein, Iron-Ligated Protein, Apoferritin Ferritin Iron Atoms

Protein Proteinn shell, core, d~12nm d~8nm

Figure 1.6: Schematics of Apoferritin and Ferritin Proteins. The apoferritin and ferritin protein shells are d~12nm in diameter, with an inner core of diameter, d~8nm. Ferritin can bind up to 4500 iron atoms inside its core.

Serum ferritin concentration often directly correlates to total body iron content and is clinically used as a marker for detecting and monitoring iron- deficiency or iron-overload in patients. In certain pathologies, however, serum ferritin can be disproportional to body iron content. For instance, it has been speculated that elevated serum ferritin levels in inflammatory diseases like Still’s disease [74] or rheumatoid arthritis [75,76], and chronic kidney disease [77] may result from elevated apoferritin (non-iron-bound ferritin). It is understood that ferritin and apoferritin participate in separate cell signaling mechanisms [78].

Ferritin can often be found localized within iron deposits in solid tissues, especially in pathologies characterized by iron overload (i.e. hereditary hemochromatosis) [79,80]. Iron overload in tissues can also be a consequence of multiple blood transfusions or an excessively iron-rich diet. Iron/ferritin deposits have been found in atherosclerotic plaques [81], cirrhotic livers [82], and

18 in several neurological disorders [83–87] including Alzheimer’s, Parkinson’s and

Huntington’s diseases and multiple sclerosis. Additionally, iron/ferritin deposits have been observed from intracerebral hemorrhage following injury [88–90]. It is not completely understood whether increased iron/ferritin levels are a cause or an effect of many of these diseases [78].

1.4.2 Ferritin in Nanoparticle Synthesis

Aside from its physiological relevance, ferritin has also become of interest in nanoparticle synthesis for a variety of applications. In particular, nanoparticles can be engineered to encapsulate different types of atoms or compounds at controlled core diameters within the protein shell. Some examples of this include the synthesis of superconducting cadmium sulfide nanoparticles [44] for fluorescent labeling, the creation of magnesium, cobalt and copper nanoparticles with various conductive properties for applications in electronics [91,92], and the mineralization of titanium dioxide for environmental remediation, energy conversion, and photocatalysis applications [93]. There have also been numerous studies in which ferritin has been created with controlled iron content

[94–96].

1.5 Detection and Characterization of SPNs

The detection of both synthetic and naturally occurring SPNs is most commonly accomplished by exploiting their unique magnetic properties and/or

19 chemical composition. The prominent techniques currently employed for SPNs are summarized below.

1.5.1 Superconducting Quantum Interference Device (SQUID) Magnetometry

SQUID magnetometry is a well-established laboratory technique to characterize the properties of magnetic materials. SQUID can measure magnetic response of a sample in two ways: (1) at a fixed magnetic field applied over a range of temperatures (typically 2K to 350K+), or (2) at a fixed temperature over a range of applied magnetic fields (typically ± 5T). SQUID can detect magnetic moments as low as 10-8 emu, and therefore has the potential to characterize small aliquots of weakly magnetic biological samples. SQUID has been used to characterize synthetic SPN ensembles, including those discussed in section 1.3, and also naturally occurring SPNs like ferritin proteins [97,98]. The samples must be dried or lyophilized [99], or require high SPION concentrations when evaluated in liquid.

Recent advances have led to the development of a nanoSQUID [100], which has a sensitivity capable of detecting a single SPN [101,102]. Large-scale

SQUID magnetometers have also been considered in the clinic to measure tissue iron deposits in vivo much like with MRI [103]. The availability of both nanoSQUID and large-scale SQUID magnetometers is thus far very limited due to high costs [104,105].

1.5.2 Mössbauer Spectroscopy

Mössbauer spectroscopy is a technique that can measure subtle changes in energy. It is based on the Mössbauer effect, in which the nucleus of a material absorbs and emits gamma rays without losing energy [106]. This is achieved by using a gamma-ray source that is an isotope of the material being characterized.

The gamma-ray source is transmitted through a material at a range of velocities, thereby interacting with the atomic nuclei in the sample and altering the intensity of the gamma ray source. A detector measures the intensity of the transmitted beam, which is plotted as a function of beam velocity. Mössbauer spectroscopy can provide information regarding the material’s chemical and structural properties, in addition to its magnetic properties, at varying temperatures.

57Fe is the most common isotope for Mössbauer spectroscopy [107], thus making iron-containing material (i.e. SPIONs) amenable to analysis. Mössbauer spectroscopy has been used to characterize ferritin proteins in vitro [73] and ex vivo [108] in addition to a variety of SPIONs [109–112]. Like SQUID magnetometry, Mössbauer spectroscopy is typically limited to characterizing ensembles of SPNs and fails to resolve a single SPN.

1.5.3 Magnetic Biosensors

Magnetic biosensors exploit the magnetic properties of SPNs to detect their presence in biological systems. They can be broadly categorized into three groups. The first category includes magnetic relaxation switches, where SPNs

21 modulate the spin-spin relaxation time of neighboring water molecules [11,113].

This method forms the basis of diagnostic magnetic resonance (DMR) and MRI analysis [114,115], The second category includes magnetic susceptibility measurements [116] and magnetic particle relaxation sensors[117], where SPNs modulate the Brownian and Néel relaxation time. The third category includes giant magnetoresistive sensors [118], where SPNs bind to a sensor surface so that magnetic fields of the particles result in changes in the electrical current within the sensor. The sensitivity of these techniques can range from nanograms

(in vitro) to milligrams of iron (in vivo). However, all these approaches provide an ensemble average of the detected SPNs and fail to resolve the magnetic signal of individual SPNs.

1.5.4 Immuno-detection

Antibody-based detection, such as enzyme-linked immunosorbent assays

(ELISA) and immuno-fluorescence microscopy, can detect natural SPNs like ferritin or antibody-conjugated synthetic SPNs. Sandwich ELISA uses antibodies to capture and analyze small amounts of proteins (>0.5ng/ml). The captured protein is detected using a secondary antibody that can be measured with spectrophotometry. ELISA is typically performed on liquid samples (i.e. ferritin in serum [119]), and cannot provide information on spatial location, iron content or magnetic properties of the SPN. Immuno-fluorescence microscopy employs a primary or secondary antibody coupled with a fluorescent label to selectively bind

22 to the targeted SPN (ferritin or synthetic immuno-conjugated SPN) in cells or tissue sections [120–123]. The spatial resolution of this technique is limited by the resolution of the fluorescence microscope (typically > 200nm), which is greater than the size of a single SPN. Immuno-fluorescence also cannot measure iron content or the magnetic properties of the SPNs.

1.5.5 Biochemical Assays

Biochemical assays can analyze the iron content in SPNs. These include digesting the iron in synthetic SPNs [124] or chelating the bound iron from proteins using a ferrozine method [125]. The amount of iron is then measured spectrophotemetrically against a standard with known iron content. Such iron assays are used to measure iron content in biological fluids (e.g. serum) and in in vitro prepared solutions containing SPNs. The minimum amount of iron that can be detected is limited by the detection limit of the spectrophotometer (typically >

0.5 μg ml-1 of iron) [124]. These assays also cannot provide information on SPN spatial localization.

Another form of biochemical assay is the Perls Prussian Blue stain, which is commonly used to visualize iron content in tissue biopsy sections [126–130].

The tissue section is treated with a combination of hydrochloric acid and potassium ferrocyanide; the former releases the iron from binding proteins and the latter reacts with the iron to produce a visibly blue stain. The stained iron can then be spatially localized within the tissue using light microscopy. Like in

23 immuno-fluorescence, the resolution for this technique is limited to the resolution of the light microscope (>200nm). Perls stain also cannot offer information regarding the magnetic properties of SPNs.

1.5.6 Electron Microscopy

Electron microscopy (EM) techniques, like transmission electron microscopy (TEM) and scanning electron microscopy (SEM), utilizes the interaction of high-energy electrons with a sample to form an image. EM offers very high resolution (2-5nm), and can therefore characterize synthetic and naturally occurring SPNs at the single particle level [95,96,131–135]. Other EM- based techniques, including high-resolution TEM (HRTEM), electron energy loss spectroscopy (EELS), energy dispersive spectroscopy (EDS) and electron diffraction (ED), have revealed valuable information regarding the chemical composition and atomic structure of a variety of SPNs [95,136–142]. Although single SPNs can be spatially localized within a sample with EM, no direct information regarding a SPN’s magnetic properties can be provided.

1.6 Limitations in Current Techniques

With the exception of electron microscopy, all other techniques listed in section 1.5 are only capable of detecting or characterizing an ensemble of SPNs.

Several factors such as spatial distribution, magnetic anisotropy, particle aggregation and magnetic dipolar interactions can complicate the signals arising

24 from an ensemble of SPNs and thus confound their analysis. Electron microscopy fails to reveal the magnetic properties of SPNs. Therefore a highly sensitive magnetism-based technique, capable of detecting single SPNs and also revealing their spatial localization, would be desirable to enhance the design and analyze the occurrence of SPNs in biomedicine.

We propose a novel and simple approach that exploits the magnetic properties of SPNs and enables their high-sensitivity detection and spatial localization in biological samples. This proposed approach is based on magnetic force microscopy (MFM), and is described in detail in Chapter 2.

CHAPTER 2

Magnetic Force Microscopy

2.1 Introduction

Magnetic force microscopy (MFM) is an atomic force microscopy (AFM)- based imaging technique routinely used to characterize magnetic materials for solid-state devices. This chapter aims to (1) describe the principles behind the

MFM technique and (2) summarize the applications and potential of MFM to characterize superparamagnetic nanoparticles (SPNs).

2.2 Magnetic Force Microscopy (MFM)

In MFM, an AFM probe coated with a magnetic material is used to detect long-range magnetic interactions between the probe and the magnetic domain(s) within a sample (Figure 2.1A). The MFM probe travels across the sample in its first pass, during which the AFM topography is acquired. In the second pass, the probe traces the sample topography at a user-defined “lift height” (z) above the sample surface and the long-range magnetic interactions are recorded as an

MFM image (Figure 2.1B) [143]. At lower lift-heights, the probe-sample

26 interactions are dominated by short range van der Waals interactions. At greater lift heights, only long-range interactions (i.e. magnetic interactions) persist. Lift height selection is therefore critical for generating and analyzing MFM contrast.

A B MFM probe

MFM Signal

cantilever Lift height, z probe magnetic coating z Topographic Image

Figure 2.1 Schematics of MFM Probe and its Paths for Image Acquisition.

(A) The MFM probe is coated with a ferromagnetic material, and is attached at the end of a cantilever. (B) The MFM probe collects topographic images at lift height, z=0nm. The probe is then lifted to a user-defined lift height (z >0nm) above the sample surface, and MFM images are collected in the second trace.

MFM is most commonly performed in tapping or dynamic mode of AFM, in which the probe is oscillated near its resonant frequency as it scans the sample.

The MFM cantilever may therefore be thought of as a simple harmonic oscillator that scans above a magnetic sample with a certain amplitude and phase of oscillation. When the probe is positioned above a magnetic domain in the sample, the magnetic probe-sample interactions alter the forces experienced by the cantilever; this behavior is manifested in the MFM image.

In dynamic mode MFM imaging, the cantilever is oscillated or “driven” at

its resonance frequency, ߱௥, which is determined by the intrinsic spring constant

(k) and mass (m) of the cantilever [143]. This relationship can be described as,

௞ ߱ ൌ ට . (Eq 2.1) ௥ ௠

When the MFM probe interacts with the magnetic moments in the sample, the

డி cantilever experiences a vertical (z-direction) force gradient, or ܨԢ. In this case, డ௭

the cantilever, behaves as if it has a modified, or effective (݇௘௙௙), spring constant

given by [143],

݇௘௙௙ ൌ݇െܨԢ. (Eq 2.2)

If the interacting forces between probe and sample are attractiveሺ݇൐ܨԢ൐Ͳሻ,

the effective cantilever spring constant is decreased. Contrarily, a repulsive

interaction ሺܨԢ ൏ Ͳሻ yields an increase in effective spring constant. The magnetic

interaction-induced effective spring constant in turn modifies ߱௥ via Equation 2.1.

This shift in ߱௥ can be either directly measured as in frequency modulation AFM

[144] or indirectly by measuring the change in amplitude or phase of the

oscillating cantilever at a fixed drive frequency (߱஽), as in amplitude modulation

AFM [145].

Most commercial AFMs employ amplitude modulation. In this case, the

drive frequency is selected as the frequency where the amplitude, ܣ, of the

cantilever has the steepest slope, and is given by [146]:

ଵ ߱ ൌ߱ ൬ͳേ ൰, (Eq 2.3) ௥ ஽ ඥ଼ொ

where ܳ is the quality factor of the cantilever. Under this condition, the change in

amplitude, ߜܣ, at ߱஽ arising due to a change in resonance frequency is given by

ଶ஺ ொ ߜܣ ൌ ቀ ೚ ቁܨԢ (Eq 2.4) ଷξଷ௞

and the change in phase is given by

ொ ߜ߮ ൌ െ ܨԢ , (Eq 2.5) ௞

where ܣ௢ is the amplitude of the cantilever when ߱௥ ൌ߱஽.

The amplitude images acquired using the amplitude modulation method

tend to inherit non-magnetic signals such as topography artifacts and laser

interference streaking, which can sometimes appear stronger than the magnetic

signals [143]. These artifacts may be eliminated by using MFM probes with

thicker magnetic coatings, but at the cost of spatial resolution. Phase images

from the amplitude modulation methods produces clear, artifact-free MFM

images [143]. Probes with thinner magnetic coating can therefore be used. The

artifact-free phase detection method is therefore the more commonly used

method for acquiring and analyzing MFM images, and is available on most

commercial AFMs.

2.3 Forces in MFM Probe-Sample Interactions

There are multiple models one may use for measuring magnetic probe-

sample interactions. The most common are the extended charge and the point-

dipole models (Figure 2.2). Of these two models, the extended charge model

[147] provides a more accurate representation of the MFM probe magnetic

29 moment because it realistically models the magnetic coating as a continuous film distributed over the probe’s surface. This model, however, requires accurate determination of the MFM probe’s magnetic coating distribution and thickness, which can vary even between probes of the same type and from the same manufacturer. The point-dipole (or point-probe) model [147–150] simplifies calculations by exhibiting the magnetic moment of the probe as a single spherical point; this model is an adequate approximation and has been used in multiple

MFM studies [151–154].

MFM Probe A MFM Probe B

Magnetic Film Magnetic Pole Magnetic Interaction Magnetic Interaction

Magnetic Particle

Figure 2.2 Models for Calculating MFM Probe-Sample Magnetic Interactions

(A) The extended charge model represents the magnetization of the probe as a continuous thin film covering the probe surface. The entire film interacts with the magnetic particle. (B) The point-dipole model represents the probe’s magnetization as a single magnetic pole that interacts with the magnetic particle.

The point-dipole model simplifies the probe and the sample each to be a single dipole moment, and is therefore applicable for MFM imaging and analysis of SPNs. Since SPNs have a single magnetic domain and lack a stable magnetic moment at room temperature, an applied magnetic field is necessary to detect the SPNs. If the applied magnetic field aligns the SPN magnetic moment in the z- direction (perpendicular to the sample surface and parallel to the magnetic dipole moment of the probe), then the force, F, on the MFM probe by the SPN is given by,

ି଺௠ ௠ ܨൌ ೛ ೞ, (Eq 2.6) ௭ర where ݉௣ is the MFM probe’s magnetic moment, ݉௦ is the magnetic moment of

the SPN and ݖ is the distance, or lift height, between the probe and tip [146]. The derivative of Equation 2.6 yields the force gradient, ܨԢ, experienced by the MFM probe during image acquisition. This force gradient is what affects the phase or amplitude shift (Equations 2.4, 2.5) and generates contrast in the MFM image.

Quantitative magnetic characterization of single-domain magnetic nanoparticles has recently been developed [151,153–155]. Schreiber et al. have derived an equation to quantitatively determine the magnetic moment of SPNs using the phase detection method [155],

ఓ ଵଶగொ ଵ଼଴ ߜ߮ ൌ ೚ ݉ ݉ , (Eq 2.7) ೏ ఱ ௣ ௦ ସగ ௞ቀ ାோା௖ା௭ቁ గ మ where ߜ߮ is the phase shift (measured in degrees), ߤ௢ is the permeability constant, ܳ is the cantilever quality factor, ݇ is the cantilever spring constant, ݀ is the diameter of the SPN being detected, ܴ is the thickness of the cantilever’s 31 magnetic coating, ܿ is the thickness of the oxidation prevention layer coating the

magnetic layer of the probe, ݖ is the probe-sample separation distance (lift height), ݉௣ is the magnetic moment of the cantilever probe, and ݉௦ is the magnetic moment of the SPN. Cantilever properties such as ݇,݉௣,ܴǡ ܳ and ܿ are usually specified by the manufacturer. The diameter of the magnetic nanoparticle can be easily obtained from the AFM/MFM topography image or by using other imaging modalities like electron microscopy. The phase shift,ߜ߮, can also be directly measured from the MFM phase image. Based on these known parameters, the magnetic moment of the sample may be quantitatively determined.

2.4 MFM Probes

Selection of the MFM probe is critical for MFM sensitivity and spatial resolution. While the parameters k and Q of the cantilever govern its frequency response, the magnetic moment and coercivity of a probe is dependent on the type of magnetic material used to coat the MFM probe (typically Co-Cr). The magnetic moment of the probe can be increased by increasing the thickness of the magnetic coating; this can be at the expense of decreasing spatial resolution.

Table 2.1 summarizes examples of commercially available MFM probes and their manufacturer-specified properties.

Table 2.1 Properties of Commercial MFM Probes Magnetic Magnetic Spring Resonance Coerc- Probe Name, Moment Coating Constant Frequency ivity Manufacturer (emu) (N/m) (kHz) (Oe) ASYMFMLM, 3e-14 Co-Cr Asylum 2 70 <400 (low) (15nm) Research

MESPLM, 3e-14 Co-Cr 1 50 <400 Bruker (low) (proprietary)

NSC18, Unspecified Co-Cr 2.8 75 240-310 MikroMasch (medium) (60nm)

MESP, 1e-13 Co-Cr 2.8 75 400 Bruker (medium) (proprietary)

ASYMFM, 1e-13 Co-Cr Asylum 2.0 70 400 (medium) (50nm) Research

MESPHM, 3e-13 Co-Cr 2.8 75 400 Bruker (high) (proprietary)

ASYMFMHM, 3e-13 Co-Cr Asylum 2.0 70 575 (high) (100nm) Research

2.5 MFM in Biomedicine: Present and Future

MFM offers a very high spatial resolution of 2-5nm and unmatched sensitivity of ~10-17 emu, which makes it a promising approach for analyzing magnetic nanoparticles. In recent years, MFM has been used to characterize a variety of ferromagnetic particles [156–158] and nanoscale magnetic domains

[159–162]. Two recent reports have shown MFM to distinguish between

33 magnetic and non-magnetic synthetic nanoparticles. Using MFM, 50nm magnetite could be identified from 17nm gold or 100nm silica particles [153].

Others utilized custom-built instrumentation involving amplitude modulated and a frequency-controlled applied magnetic field during imaging to distinguish between clusters of superparamagnetic maghemite and diamagnetic gold nanoparticles, both with diameters < 20nm [3].

MFM has also been used for qualitative assessment of SPNs aggregates in biological systems in ambient air. These include detection of bio-conjugated magnetic nanoparticles in vitro [163], magnetic crystals naturally synthesized inside bacteria [164,165] and SPNs on cell surfaces [166–168]. Magnetic moments of synthetic SPNs [155] and ferritin proteins [152] have also been quantitatively estimated using MFM via the approach described in Equation 2.7.

The potential of MFM to detect and spatially localize single SPNs in biological systems, however, has not been completely harnessed. Detection of individual SPNs is challenging because a single SPN possesses a weak magnetic moment in addition to a small particle size and potential magnetic anisotropy. Additional considerations like improvement in MFM force sensitivity, sample preparation methods and data analysis are required to overcome these challenges. Once overcome, we envisage MFM can serve as a powerful tool to understand the complex magnetic properties of SPNs and serve as a biosensor to characterize nanoscale iron deposits in biological tissues.

2.6 Thesis Aims

The overall goal of this thesis was to develop MFM for enhanced detection of SPNs encountered in biological applications. Our central hypothesis was that

MFM can reveal novel insights into the magnetic properties of SPNs and serve as an ex vivo tool to estimate iron content in biological samples. In Aim (1) we developed methods to increase the sensitivity of MFM to detect monodisperse synthetic magnetite (Fe3O4) SPNs at the single particle level. In Aim (2) we developed methods to detect and quantify the iron-bound protein ferritin in vitro and in biological tissue (ex vivo). These advancements provide novel insights into the properties of SPNs and their distribution in vivo. We also discuss how our results can aid in the development of a novel indirect MFM technique, which has the potential for high-throughput detection of SPNs.

CHAPTER 3

Detection of Single Superparamagnetic Nanoparticles Using MFM

3.1 Aims and Rationale

The overall goal for this study was to achieve detection of superparamagnetic nanoparticles (SPNs) at the single particle level using magnetic force microscopy (MFM). We developed two new approaches to characterize single SPNs, (1) exploitation of SPN magnetic anisotropy to maximize SPN magnetic moment responses and (2) improvement of MFM sensitivity by increasing the magnetic moment of the MFM probe.

3.1.1 Magnetic Anisotropy of SPNs

Since SPNs do not possess a stable magnetic moment, an external magnetic field (BMFM), must be applied during MFM imaging to induce a stable magnetic moment in the SPNs and consequently enable their detection at room temperature. In all previous MFM studies it was assumed that magnetic moments of all SPNs aligned along the direction of BMFM. This assumption, however, fails to consider possible effects from magnetic anisotropy in SPNs, which could

36 hinder the ability of certain SPNs to align their magnetic moments with BMFM. In section 3.3.3 we demonstrate how magnetic anisotropy can be exploited to produce a more uniform and enhanced MFM signal from single SPNs.

3.1.2 Increasing MFM Probe Magnetic Moment

Previous MFM studies have detected SPN aggregates by using MFM probes with a standard (or medium) magnetic moment (MM). Recent advancements in MFM probe manufacturing have produced MFM probes with high magnetic moments (HM). These probes are reported to have a magnetic moment approximately three times greater than MM probes. Both probes are otherwise comparable in spring constant and resonance frequency. The use of

HM probes is expected to increase MFM sensitivity and enable detection of single SPNs. In section 3.3.4, we compare the performance of a standard MM probe to that of a HM probe, with particular emphasis on their abilities to detect single SPNs using MFM.

3.2 Materials and Methods

3.2.1 SPN Synthesis

To analyze single particles, we used oleic acid surfactant coated magnetite (Fe3O4) SPNs provided by Dr. Christopher B. Murray (University of

Pennsylvania). These SPNs were expected to be monodisperse and were suspended in a quick-drying solvent (hexane) to promote immediate nanoparticle

37 immobilization and minimize SPN aggregation that can occur during the solvent- drying stage. The Fe3O4 SPNs were synthesized by the Murray group using a slightly modified method developed by Hyeon et al [169].

The SPN samples were washed before sample preparation with the following method to minimize superlattice formation: 1ml of SPNs dispersed in hexane was mixed with 2ml of isopropanol (2-Propanol Certified ACS Plus,

Fischer Scientific) in a glass vial, centrifuged at 1507 g for 3 minutes, and the supernatant was decanted. Thereafter the SPN pellet was washed with 2ml of hexane, centrifuged and re-dispersed in fresh hexane. Suspension of SPNs in a quick-drying solvent (hexane) permitted immediate immobilization of SPNs on substrates and therefore minimized particle aggregation that can potentially occur during the solvent-drying stage.

3.2.2 Superconducting Quantum Interference Device (SQUID) Magnetometry

The magnetic properties of Fe3O4 SPN ensembles were performed by

Murray et al. on a commercial superconducting quantum interference device

(SQUID) magnetometer. Briefly, 50μl of the SPN solution (undiluted) was sealed in a glass tube. This tube was placed inside a SQUID magnetometer and measurements were performed from 15 K to 300 K. The magnetic moment of the sample at 300 K was recorded by varying the field from -40k Oe to 40 kOe. The zero-field cooling (ZFC) curve was acquired during the warming process with an applied magnetic field of 100 Oe.

3.2.3 TEM and MFM Sample Preparation

To evaluate the role of magnetic anisotropy in this study, SPNs were immobilized on a substrate using two different protocols, resulting in (a) randomly oriented and (b) pre-aligned samples (Figure 3.1). The pre-aligned samples were prepared by applying a magnetic field, Bsample, perpendicular to the substrate during the SPN immobilization and solvent evaporation stage. This was done to facilitate SPN easy axis alignment via Brownian motion [2] along the direction of Bsample.. Randomly oriented samples were immobilized in the absence of any external magnetic field.

For bright field TEM imaging, 10μl of SPN solution was immediately aliquoted onto formvar-copper TEM grids (Electron Microscopy Sciences) with or without the presence of a permanent magnet of field strength Bsample ~1000 G placed below the TEM grid, and allowed to dry for 10 minutes. For analytical

TEM studies, 10μl of SPN solution was aliquoted onto silicon nitride TEM grids

(SiMPore Inc.) and allowed to air dry. Thereafter the samples underwent 30 seconds of plasma cleaning (Fischione 1020 Plasma Cleaner, Fischione

Instruments) to remove oleic acid surfactant and minimize organic contamination due to the high energy electron beam required for single particle diffraction.

Identical samples were prepared on silicon nitride grids for high resolution TEM

(HRTEM).

Figure 3.1 TEM and MFM Sample Preparation and MFM Set-up for

Randomly Oriented and Pre-aligned SPN Samples SPNs were immobilized on the substrate in the (A) absence (randomly oriented) or (B) presence (pre- aligned) of a magnetic field, Bsample, applied perpendicular to the substrate (white block arrow). Bsample facilitates alignment of SPN easy axes (dotted lines) along itself. In the absence of any external magnetic field at room temperature, SPN magnetic moments are unstable and flip between the two antiparallel directions

(double arrows) along their easy axes. During MFM imaging (C,D), an external magnetic field, BMFM, is applied perpendicular to the sample substrate (white block arrows). The energy required to align the SPN magnetic moment (single black arrow) parallel to BMFM increases in proportion to sinθ, where θ is the angle between the SPN easy axis and BMFM. Reprinted with permission from [132]. 40

For MFM, 10-15μl of prepared solution was immediately aliquoted onto freshly cleaved mica substrate (ruby muscovite, S & J Trading) with or without the presence of a permanent magnet (Bsample ~ 1000 G) placed below the substrate and allowed to dry overnight. In the absence of Bsample, SPNs were free to immobilize on the mica substrate at any random crystalline orientation.

Application of Bsample facilitated SPN easy axis alignment along Bsample.

3.2.4 Analytical TEM

Bright field TEM studies were conducted at 80kV on a Zeiss EM 900 TEM

(Carl-Zeiss) and digital micrographs were captured on an Olympus SIS

Megaview III camera. SPN diameters were measured for n=100 particles from

TEM bright field images using ImageJ (NIH) software. Electron diffraction (ED) and energy-dispersive X-ray spectroscopy (EDS) were performed using a Tecnai

F20 field emission 200kV TEM / Scanning-TEM (STEM) and X-TWIN lens (FEI).

Diffraction on single SPNs required use of a microprobe STEM setting on the

Tecnai F20. High resolution TEM (HRTEM) was performed on a Titan3 80-300 probe-corrected monochromated (S)TEM and associated single particle diffraction patterns were obtained via fast Fourier transform (FFT) analysis using

ImageJ software (NIH).

3.2.5 MFM Studies

MFM studies were conducted on a Multimode AFM equipped with a

Nanoscope IIIa controller and Quadrex Extender (Digital Instruments). SPNs were imaged in the presence of an applied magnetic field, BMFM ~ 2000 G, emanating from the permanent magnet at the base of the JV scanner in the

Multimode AFM [170]. Samples on a thin, freshly cleaved mica substrate attached directly to the JV scanner with adhesive double-sided tape therefore experienced a magnetic field of BMFM ~ 2000G perpendicular to the mica substrate. MFM probes (medium moment MM probe NSC-18 from Mikromasch or high moment HM probe ASYMFMHM from Asylum Research) were magnetized with a permanent magnet for 2 minutes prior to use. The probes were auto-tuned to their nominal resonant frequencies with a 5% offset for main controls and 0% offset for interleave controls. Images were acquired in tapping mode with 512 lines per scan direction at a scan rate of 2 Hz. Topographical height images were acquired in the main scan while the phase images were acquired in the interleave lift mode at various lift heights. Images were flattened and quantitative analyses of particle diameter and phase were performed using the section analysis tools in Nanoscope Software, version 1.40 (Bruker Corp.).

For magnetic anisotropy analysis, a total of n=35 (n=10 high contrast, n=25 low contrast) particles in randomly oriented samples and n=24 particles in pre-aligned samples were selected with diameters corresponding to those of

42 single particles (as ascertained using TEM). The MFM phase shift was measured for all selected particles at each designated lift height.

3.3 Results

3.3.1 SPN Characterization

The composition and morphology of individual SPNs were ascertained using analytical transmission electron microscopy (TEM), including energy- dispersive X-ray spectroscopy (EDS), electron-diffraction (ED) and high- resolution TEM (HRTEM). Bright field TEM images (Figure 3.2A) show SPNs to be monodisperse spherical nanoparticles with a mode diameter of ~20 nm.

Energy dispersive spectroscopy (EDS) on individual SPNs confirmed the presence of iron and oxygen in all nanoparticles examined (Figure 3.2B). The silicon and nitrogen peaks observed in the EDS spectra were from the silicon nitride TEM grids used for EDS studies.

It is interesting to note that the bright field TEM images exhibited a mixture of contrast in nanoparticles, ranging from shades of light to dark gray. Such contrast differences have also been reported in magnetite [171] and FePt nanocrystals [4]. To examine whether the differences in SPN contrast were due to different crystalline orientations, we performed single particle electron diffraction (ED) analysis using STEM. Figure 3.2C shows that the SPNs exhibited a diverse range of diffraction patterns corresponding to different crystalline orientations. HRTEM studies (Figure 3.2D) further show that the

43 crystalline planes within individual SPNs exhibit various crystal lattice orientations. Magnetite (Fe3O4) has a cubic spinel structure with easy axis along the <111> direction [171]. However, no correlation between the crystalline orientation of the SPN and its corresponding contrast in bright field TEM images could be ascertained. We therefore envisage that the interactions of the electron beam with the SPN may be dependent on a multitude of factors including, but not limited to, crystalline structure and orientation, lattice defects, shape and stress anisotropy.

Magnetic properties of an ensemble of our SPNs were evaluated with

SQUID magnetometry and zero-field cooling (ZFC) experiments (Figure 3.3).

SQUID magnetometry results show the magnetic moment saturation, Ms, approaching 74emu g-1 at 300K. The magnetization hysteresis curve demonstrates a lack of remnant magnetization at zero fields (Figure 3.3A inset), confirming the superparamagnetic behavior of these nanoparticles at room temperature. The ZFC curve indicates the nanoparticle blocking temperature, ܶ஻, to be approximately 205K.

-1 The Ms value obtained in our studies (74 emu g ) is slightly higher than

-1 the Ms (60-65 emu g ) for 10 and 50 nm magnetite particles, as reported earlier

[171], and closer to that of bulk magnetite (~ 90 emu g-1). The particles used in this earlier study also exhibited a low level of coercivity at room temperature. It is likely that differences in the surfactant coating and/or surface structure of the

SPNs resulting from different synthesis techniques influence the inter-particle interactions and the resulting Ms and coercive behavior.

A B

C D

<011>

<223> <111> <111> 3 <001>

5nm <012> <112>

Figure 3.2 Composition and morphology of SPNs (A) Bright field TEM image shows SPNs to be monodisperse, spherical particles of diameter ~20nm. (B)

Energy dispersive spectroscopy (EDS) confirms the presence of iron (Fe) and oxygen (O) in SPNs. (C) Microprobe electron diffraction of individual SPNs show differences in diffraction patterns, indicating that SPNs have random crystalline orientation. (D) High resolution TEM (HRTEM) further shows the crystalline cubic structure (inset) of the SPNs, as well as three SPNs with three different crystal lattice orientations. Reprinted with permission from [132]. 45

A B

Figure 3.3 Magnetic Properties of SPN Ensembles. (A) Superconducting quantum interference device (SQUID) magnetometry M-H curve shows SPNs to have a saturated magnetic moment approaching 74 emu g-1. There is no remnant magnetism at zero fields (inset), confirming the SPNs to be superparamagnetic.

(B) The zero field cooling (ZFC) curve indicates SPN blocking temperature to be

TB = 205K. Reprinted with permission from [132].

3.3.2 Identification of Single SPNs for MFM Analysis

To evaluate the role of magnetic anisotropy in MFM studies, we analyzed two sample types as described in the methods, (a) randomly oriented and (b) pre-aligned samples (Figure 3.1). To ensure that our MFM analysis consisted of single SPNs, a size distribution analysis was conducted on MFM topography and bright field TEM images (Figure 3.4) of both randomly oriented and pre-aligned

SPN samples. MFM topographic images of randomly oriented samples yielded a mode apparent lateral diameter (d) of SPNs to be ~120nm. We utilized the geometric de-convolution method [172–174],

݀ൎͶξܴݎ, (Eq 3.1) to ascertain the real lateral diameter (2r) of the SPNs in our topographic images.

The average radius (R) of our MFM probe was ~90nm, as specified by the manufacturer (Mikromasch) and confirmed by us using scanning electron microscopy (data not shown). The particles’ real diameters calculated from the

MFM topography images are shown in Figure 3.4A inset. The mode values in

MFM and TEM images were used to identify SPN singlets for MFM analysis. We attribute occasional existence of larger particle diameter to SPN aggregation.

Pre-aligned samples exhibited a shift in particle diameters in both MFM and TEM images (Figures 3.4B and 3.4D). TEM images (Figure 3.4D) of pre- aligned samples revealed three populations: (i) isolated SPN singlets with diameters identical to the mode diameter in the randomly oriented TEM samples

(~20nm), (ii) larger ( > 100nm), less electron dense particles and (iii) SPN singlets situated inside of the larger, less electron dense particles. A similar feature was observed in pre-aligned MFM samples (Figure 3.4B), in which a fewer number of particles of type (i) were interspersed between particles of larger diameters of either particle type (ii) or (iii). We speculate the formation of particles with comparatively large lateral diameters in pre-aligned samples in TEM and

MFM images was due to magnetic field-facilitated SPN aggregation [154] and/or accumulation of free iron ions present in SPN solutions; these particles were excluded from phase analyses. In the following studies, only particles with de- convoluted lateral diameters of ~20nm were selected for MFM analysis.

A B

C D

Figure 3.4 Identification of Single SPNs for MFM Analysis. MFM topography images of (A) randomly oriented samples show nanoparticles with a mode de- convoluted diameter of 20nm (inset). In pre-aligned samples, a bi-modal distribution of particle sizes is observed in the (B) MFM topography image.

Similar observations were made using TEM for (C) randomly oriented and (D) pre-aligned samples. The large diameter particles observed in (B,D) are attributed to the accumulation of free iron ions in the SPN sample solution and/or

SPN aggregation facilitated by the application of Bsample; these particles were excluded from our MFM phase analysis. The individual particles in the pre- aligned MFM image are outlined in white. Reprinted with permission from [132].

3.3.3 Effects of Magnetic Anisotropy on MFM Signal of SPN Singlets

MFM was conducted on both randomly oriented and pre-aligned samples at increasing lift heights, z. Samples were imaged in the presence of a magnetic field, BMFM, applied perpendicular to the substrate to facilitate magnetic moment stabilization at room temperature. We predicted for randomly oriented samples that only single SPNs with small angles, θ, between easy axis and BMFM could successfully align their magnetic moments along BMFM (Figure 3.1C). For pre- aligned samples, however, we predicted that since the SPN easy axes should be aligned parallel to Bsample (which was parallel to BMFM), the majority of SPNs would stabilize their magnetic moments also along BMFM. We therefore expected that the pre-aligned samples would exhibit a greater number of SPNs with magnetic moments aligned parallel to BMFM during MFM imaging (Figure 3.1D).

MFM phase shift analysis of the randomly oriented samples (z = 15nm) revealed that although the selected SPNs had the same lateral diameter of

20nm, there were two distinct populations of phase shifts (Figure 3.5A). The low contrast population (first row) had both a positive and a negative phase shift component characteristic of a dipole moment not aligned parallel to BMFM. The average phase shift of the low contrast population (n=25) was measured to be

18.2 ± 4.0o from positive to negative peak. The high contrast particles (second row) were less abundant (n=10) in these randomly oriented samples and exhibited only a high positive phase shift (average 56.8 ± 12.8o at z = 15nm).

MFM analysis of pre-aligned samples at z = 15nm revealed a high positive phase shift of 69.6 ± 5.3o (Figure 3.5B) for all nanoparticle singlets (n=24).

These phase shift values were slightly higher, yet comparable to the high phase shifts observed in randomly oriented samples. The phase shift response from a

SPN is maximized only when the SPN magnetic moment is perfectly aligned along the direction of BMFM. It is likely that not all high-contrast particles in the randomly oriented samples had their magnetic moments completely aligned along BMFM. The application of Bsample to the pre-aligned samples, likely facilitated a more favorable orientation of the particle easy axes, resulting in a more complete alignment of their magnetic moments along BMFM. This is further supported by the larger spread in standard deviation (±12.8ᵒ) for the high- contrast particles in the randomly oriented samples compared to the standard deviation (± 5.3ᵒ) for the pre-aligned SPNs.

Phase shift analysis of singlets as a function of increasing lift height

(Figure 3.6A) for both randomly oriented and pre-aligned samples showed that while the randomly oriented samples exhibited two populations (low and high) of phase shifts, pre-aligned samples consistently showed only high phase shifts at each lift height. All singlets exhibited zero phase shift at z ≥ 30nm. The distribution of high and low contrast particles for each sample type was quantified

(Figure 3.6B), revealing only ~29% of the SPN singlets in the randomly oriented samples to have high phase contrast. However, high contrast was observed in

100% of SPN singlets in the pre-aligned samples. We attribute this observation

50 to the pre-aligned SPNs having a lower energy barrier for alignment along BMFM compared to randomly oriented SPNs. SPN easy axis pre-alignment therefore consistently maximized MFM phase shift from all individual SPNs in the sample.

A B

500nm d Low Contrast Oriented y ndomly High Randomly Oriented Oriented Randomly Ra Contrast

C d 40° ed

0° Pre-aligned Pre-aligned

Figure 3.5 MFM Phase Shift Analysis of Single SPNs. (A) MFM phase image of a randomly oriented sample exhibiting the coexistence of two populations of phase contrasts, low (white arrows) and high contrast (black arrows). (B)

Individual low and high contrast SPNs from the randomly oriented sample with corresponding phase shift profiles. (C) Pre-aligned samples exhibit single SPNs with high phase contrast only. The dimensions of all single SPN images are

250nm x 250nm. Reprinted with permission from [132].

A B

100%100 29% 100%

50%

0% Randomly Pre- oriented aligned low contrast high contrast

Figure 3.6 MFM Analysis of Randomly Oriented and Pre-Aligned SPNs.

(A) Phase contrasts plotted at increasing lift heights. In randomly oriented samples, low contrast SPNs (n=25) consistently produce lower phase contrast at each lift height as compared to the high contrast SPNs (n=10). All SPNs (n=24) from pre-aligned samples exhibit only high phase contrasts at each lift height.

Reprinted with permission from [132]. (B) Pre-alignment of SPN samples increases the percentage of high contrast particles from 29% (randomly oriented sample) to 100%.

3.3.4 Effects of MFM Probe Magnetic Moment on MFM Signal

As a second approach to increase MFM sensitivity for SPN detection, two types of MFM probes were tested and compared to determine whether an increase in probe magnetic moment could enhance the phase contrast of SPN singlets. These experiments were conducted on randomly oriented SPN samples

52 with applied magnetic field, BMFM, and at increasing lift heights, z. The plot in

Figure 3.7 shows the average phase shifts experimentally measured at each lift height using high magnetic moment (HM) and medium magnetic moment (MM) probes (solid black squares and circles, respectively). It is apparent that the HM probe yields phase shifts approximately larger in magnitude than the MM probe.

The HM probe also detects SPN singlets up to a lift height of z = 70nm, which is over two-fold greater than detected with the MM probe (maximum z=30nm).

We compared experimental to theoretical phase shift values for the SPNs

(diameter, d=20nm). First, the magnetic moment, ݉௦, of an individual SPN was calculated using the particle volume, ܸ [8],

݉௦ ൌܸൈܯ௩௦, (Eq 3.2) where ܯ௏௦ is the SPN volume saturation magnetization determined by

ܯ௩௦ ൌܯ௦ ൈߩ. (Eq 3.3)

Saturation magnetization, ܯ௦, was pre-determined using SQUID magnetometry

-1 (here, ܯ௦ = 74emu g ). ߩ is the density of the SPN’s bulk material. For Fe3O4 or magnetite, ߩ = 5240 kg m-3. The SPNs in this study therefore have a theoretical

-18 2 magnetic moment, ݉௦, of 1.6 x 10 Am , which could be used to calculate MFM phase shifts.

Theoretical phase shifts for the MM probes and the HM probes were calculated at each lift height, z, using the equation derived by Schreiber et al.

[155] and described in Chapter 2, equation 2.7,

ఓ ଵଶగொ ଵ଼଴ ߜ߮ ൌ ೚ ݉ ݉ . (Eq 3.4) ೏ ఱ ௣ ௦ ସగ ௞ቀ ାோା௖ା௭ቁ గ మ

The magnetic moment of each probe (݉௣) was estimated using a second equation provided by Schreiber et al. [155],

ସ ݉ ൌ ߨ൫ܴଷ െܴ ଷ൯ܯ , (Eq 3.5) ௣ ଷ ଵ ௣ where ܴଵ is the radius of the MFM probe before magnetic coating and ܯ௣ is the saturation magnetization of the probe’s magnetic coating (Co, 1.4 x 106 A m-1).

Probe-dependent variables and manufacturer specified nominal values are summarized in Table 3.1 for MM and HM probes. The theoretical phase shift values are plotted with the experimental values in Figure 3.7.

Table 3.1 Medium Moment (MM) and High Moment (HM) Probe Parameters Spring Magnetic Pre-coated Oxidation Magnetic Quality Constant Coating Probe Prevention moment Factor (k, N m-1) Radius Radius Coating (m , Am2) (Q) p (R, nm) (R1, nm) (c, nm) MM 150 2.5 70 10 20 2e-15 Probe

HM 150 2.0 90 10 20 5.9e-15 Probe

HM Probe:Probe: experimental HM Probe:Probe: theoreticaltheoretical MMMM PrProbe:obe: experimental MM Probe: theoretical

Figure 3.7 MFM Phase Shifts of SPN Singlets Using High Moment and

Medium Moment Probes. Experimental data (black points) shows the high moment (HM) probe detects SPNs with phase shifts nearly twice that of the medium magnetic moment (MM) probe at the same lift height. The lift height up to which the magnetic signal can be detected is also over twofold greater than the MM probe. Theoretical phase shift values are also shown (white points).

3.4 Discussion

We demonstrate here the effects of magnetic anisotropy and probe magnetic moment on the MFM phase shift signal obtained from single SPNs. In previous MFM studies by us [170] and others [152,167], the occurrence of particle agglomeration had resulted in a heterogeneous distribution of phase shifts most likely due to dipole-dipole interactions and random easy axis orientations in SPNs within the agglomerations. In this work, the use of

55 monodisperse, surfactant-coated SPNs in a quick-drying solvent (hexane) was critical to characterize the magnetic properties of single SPNs. Our findings indicate that both the exploitation of magnetic anisotropy and the use of HM probes enhance the MFM phase contrast of SPNs. The HM probe additionally increases the range up to which single SPNs can be detected as compared to a standard MM probe.

In our anisotropy studies, we elucidate how SPN samples with random orientations of easy axes can produce a large range of phase shifts, making quantification of SPN magnetic moments rather complicated. Magnetic anisotropy should therefore be taken into consideration as a possible source of

MFM phase shift inconsistency between SPNs of the same size and composition, and when imaged using MFM probes with similar properties. Easy axis alignment also creates a probe-SPN interaction that is either all repulsive (in our case), or all attractive depending on the relative polarity of the MFM probe magnetization and direction of BMFM. Controlling for how the SPN easy axis is oriented in a sample with respect to external magnetic fields could therefore maximize the magnetic response and simplify analysis of SPN magnetic moments using MFM.

Another mechanism that can give rise to a diversity in MFM phase contrast includes the presence of vortex domains, which have been reported for ferromagnetic nanoparticles with crystalline and morphological anisotropy [175].

It is unlikely that our MFM contrast is affected by vortex domains because our studies utilize single domain SPNs of uniform shape and size. Magnetic

56 properties of magnetite nanoparticles also depend on inter-particle dipolar interactions as well as intra-particle defects like antiphase boundaries, oxygen deficiency, and ionic disorder [171]. Since the inter-particle distances in our MFM images were typically > 500 nm, it is unlikely that dipolar interactions affected our

MFM phase contrast. However, the role of intra-particle heterogeneity cannot be completely ruled out. Techniques to quantitatively characterize the hysteresis loop and magnetic anisotropy of single nanoparticles would be required to completely understand the unique magnetic properties of SPNs.

While exploiting magnetic anisotropy did increase the MFM phase contrast of our SPNs, the lift height range remained the same for SPN singlets in all samples (z = 30nm) when using the standard MM probe. The employment of a

HM probe increased MFM sensitivity even further by not only enhancing the phase contrast of the SPNs but also by increasing the range up to which the SPN singlets could be detected (z = 70nm). The ability to increase the MFM working distance (lift height range) is especially important in biological applications where the samples tend to have a rough topography and/or the SPNs are not always positioned on the surface of the sample. Such circumstances could occur, for example, when imaging SPN-labeled cells or detecting iron-bound proteins in biological fluids and tissue. Additionally, we compared our experimental phase shift values to theoretical values calculated using an equation derived by

Schreiber et al [155]. Although experimental and theoretical phase values were in good agreement at low (<30nm) and high (>55nm) lift heights, there were some

57 discrepancies. This equation utilizes a point-dipole approximation of the MFM probe magnetic moment, which does not take into account potential variations in

MFM probe magnetic coating distribution and thickness. The theoretical calculations also represent phase shifts from purely magnetic probe-sample interactions, however experimental phase values could also include van der

Waals, electrostatic and capillary forces, especially at lower lift heights.

Experimental results show that the combination of using HM probes and exploiting the magnetic anisotropy of nanoparticles may enable the magnetic characterization of a more diverse set of samples with higher sensitivity

In summary, we demonstrate here how magnetic anisotropy can affect

MFM phase shift from single SPNs. Thus far magnetic anisotropy characterization was only possible on an ensemble of particles using techniques such as Mössbauer spectroscopy [6]. However, this technique is restricted to analysis on an ensemble of SPNs and is insensitive to variations between single nanoparticles. MFM can therefore serve as a unique tool to characterize magnetic anisotropy in single SPNs. We also show how increasing the magnetic moment of the MFM probe can increase the phase contrast and the lift height range up to which SPN singlets can be detected. Thus, we demonstrate how

MFM can have both the force sensitivity and spatial resolution necessary for characterizing magnetic nanoparticles at the single particle level.

CHAPTER 4

Quantification and Spatial Localization of Ferritin Using MFM

4.1 Aims and Rationale

The overall goal for this study was to employ magnetic force microscopy

(MFM) to ascertain the concentration and/or spatial distribution of iron ligated vs. non-ligated proteins. High magnetic moment (HM) MFM probes were used to increase MFM sensitivity and identify ferritin (vs. apoferritin) proteins in vitro and endogenously expressed in serum and tissue samples.

4.1.1 Quantification of Ferritin

The ability to quantify and distinguish ferritin from apoferritin and other iron-bound proteins is clinically important. AFM-based approaches have previously attempted to examine differences in conductivity of purified ferritin and apoferritin [176,177], however this would be difficult when trying to identify ferritin in biological fluids. Although there has been evidence that MFM can detect ferritin in vitro [152], its ability to quantify ferritin has not been adequately

59 investigated. We demonstrate how MFM can serve as a valuable tool to estimate ferritin content in vitro in ferritin/apoferritin mixtures and in animal serum.

4.1.2 Spatial Localization of Ferritin

Perls iron staining, which is typically used to localize ferritin, is limited to

~200nm in spatial resolution. We demonstrate how MFM can ultra-structurally localize iron/ferritin deposits in animal tissue. The high sensitivity characterization of iron deposits by MFM may help us detect and better understand early stages of iron overload.

4.2 Materials and Methods

Purified ferritin from equine spleen was utilized because its morphology and magnetic properties have been well-characterized [73,152,178,179]. In addition, its non-iron-bound counterpart, apoferritin (also from equine spleen) is readily available. MFM studies were performed using high magnetic moment

(HM) MFM probes based on our previous study (Chapter 3), which showed significant improvements in the magnitude of phase contrast and range of detection of SPN singlets as compared to standard medium moment (MM) probes [132]. For serum and tissue studies, a well characterized rodent model of spinal cord injury was used, which shows serum and tissue ferritin up-regulation post-injury [122].

4.2.1 Transmission Electron Microscopy (TEM)

Purified apoferritin and ferritin proteins (Sigma-Aldrich, St. Louis, MO,

USA) were from equine spleen. The proteins were diluted in milli-Q water to a final concentration of 1μg ml-1. Approximately 15μl of each protein solution was aliquoted onto copper-coated TEM grids and allowed to sit at room temperature for 10 minutes. The grids were then blotted and washed two times with milli-Q water and air-dried. For staining, ~ 5μl of 1% uranyl acetate was applied to the samples for thirty seconds; the grids were then blotted and allowed to air-dry overnight. Grids were imaged with a Zeiss 900 TEM operated at 80KeV (Carl-

Zeiss SMT). Digital micrographs were collected using an Olympus SIS Megaview

III camera. Core diameters and their standard deviations for ferritin (n=100) and shell diameters for both apoferritin and ferritin proteins were measured using

ImageJ software (NIH).

4.2.2 Magnetic Force Microscopy (MFM) of Purified Proteins

Apoferritin and ferritin stock solutions were diluted in milli-Q water, and then mixed in five different ratios (0%, 25% 50%, 75% and 100% ferritin) with a total protein concentration of 1μg ml-1. To reduce aggregation, the protein solutions were vortex mixed for 10-15 seconds prior to immobilization on substrates. Approximately 30μl of each protein solution was aliquoted onto thin, freshly cleaved mica substrates (Ruby muscovite, S&J Trading Inc.) and allowed to air-dry in ambient conditions overnight.

MFM imaging was performed on a Multimode atomic force microscope

(AFM) equipped with Nanoscope IIIa controller and Quadrex extender (Bruker,

Santa Barbara, CA, USA). Samples were attached directly to the base of the scanner with adhesive tape, and therefore experienced a magnetic field from the applied perpendicularly to the substrate [132,155] during imaging.

MFM was performed using high magnetic moment (HM) probes

(ASYMFM-HM, Asylum Research) that were pre-magnetized with a permanent magnet for 2 minute prior to use. MFM cantilevers were auto-tuned to the manufacturer specified resonance frequency (70 kHz) with a 5% offset for main controls and 0% offset for interleave controls. Height and phase images were obtained in tapping mode at a scan rate of 2Hz and at 512 lines/scan direction.

The oscillation amplitude of the probe was between 0.7nm and 3.7nm. Height images were collected in the first pass, and phase images were collected in interleave lift mode at increasing lift heights (z=0 to 50nm). At least three independent experiments per sample type were conducted.

Images were flattened and height and phase images for n=100 particles per sample type were analyzed with the section analysis tool in the NanoScope

Analysis software v. 1.40 (Bruker Corp.). Averages and standard deviations were reported for each sample type. Particles with a lateral width corresponding to the diameter of single apoferritin and ferritin proteins (13.2±1.1nm) were selected for phase shift analysis (n=50) and are referred to as ~13nm particles.

4.2.3 Protein and Iron Assays for Purified Proteins

Apoferritin and ferritin stock solutions were diluted in milli-Q water, and then mixed in five ratios (0%, 25% 50%, 75%, and 100% in ferritin) with a total protein concentration of 300μg ml-1. Protein concentrations were confirmed using the Bio-Rad DC Protein Assay at a wavelength of 750nm. The ferrozine method was used for the iron assay: 50μl of each protein solution was mixed in a glass vial with 100μl of 2M H2SO4 and 100μl of 0.5mM dihydroxyfumaric acid (both from Sigma-Aldrich) and incubated at room temperature for 30 minutes to strip the protein shells from their iron cores. Next, 200μl of 2.5M NaOAc (Sigma-

Aldrich) and 100μl of 1mM ferrozine (Fisher Scientific) were added to each vial and allowed to incubate at room temperature for another 30 minutes. Finally, each solution was diluted with 700μl of milli-Q water. A 1ml aliquot of each solution was then placed into a cuvette and measured with a DU730 Life Science

UV/Vis spectrophotometer (Beckman-Coulter). Absorbance was recorded at a wavelength of 561nm.

4.2.4 Animal Studies

All rat serum and spinal cord tissue were provided by our collaborators

(Dr. Dana M. McTigue, OSU). Briefly, spinal cord contusions were performed on adult female Sprague-Dawley rats (250g; Harlan) using standardized protocols as described elsewhere [122,180,181]. Serum samples and their corresponding serum ferritin levels (obtained using sandwich enzyme-linked immunosorbent

63 assay, ELISA) were provided for naïve and injured rats 21 days post injury.

Spinal cord tissue was provided for injured rats 64 days post injury. Adjacent tissue sections that had been stained with either Perls (iron) or immuno- fluorescence (ferritin) underwent light and fluorescence microscopy, respectively, to identify regions with (1) low iron/ferritin and (2) high iron/ferritin.

4.2.5 MFM of Serum and Spinal Cord Tissue

Approximately 25μl of undiluted naïve or 21 days post injury rat serum was aliquoted onto freshly cleaved mica. Samples were incubated at room temperatures for ten minutes, then blotted and allowed to dry overnight at 4ºC.

MFM imaging and analysis was performed as described for purified proteins in section 4.2.2. Height and phase images (z=0nm, 10nm and 30nm) of several regions totaling 100μm2 were collected on each sample type using a HM MFM probe. Quantitative phase analysis was performed at z=30nm and magnetic particle concentration was estimated from the MFM phase images to determine average magnetic particles per μm2.

MFM was also performed on unstained spinal cord sections from a 64 day post injury rat. Low and high iron/ferritin regions were located using morphological features and fiduciary markers corresponding to those in adjacent sections that had been stained with either Perls (iron) or immuno-fluorescence

(ferritin). MFM height and phase images (at z = 0nm, 10nm and 30nm) were collected as previously described (section 4.2.2) on the low and high iron/ferritin

64 regions using a HM MFM probe. Quantitative phase analysis was preformed on the z=30nm phase images.

4.3 Results

4.3.1 Characterization of Protein Morphologies

Apoferritin (0%) and ferritin (100%) morphologies were characterized using transmission electron microscopy (TEM) and topographic AFM imaging.

Ferritin iron cores were observable only in ferritin samples and not in apoferritin in TEM images (Figure 4.1). The insets in Figure 4.1A show the protein shells in both apoferritin (average diameter 13.2 ± 1.5nm) and ferritin (average diameter

12.4 ± 1.5nm), which were visible after the samples were stained with uranyl acetate. The iron cores in ferritin ranged from d = 4 to 7nm, with an average diameter of 5.1 ± 1.1nm; these measurements correspond well with earlier reports [95,96,131].

From AFM height images (Figure 4.2A), we found that the de-convolved

[132] lateral diameters of ferritin ranged from approximately 4nm to 45nm, while those of apoferritin ranged from 13nm to 45nm (Figure 4.2B). The presence of smaller particles (4nm - 7nm) in ferritin samples is likely due to the dislodging of some iron cores during sample preparation. Particles larger than 20nm were assumed to be protein agglomerates. Particles with d = 13.2nm ± 1.1nm (or d~13nm) were considered individual proteins; this diameter value is in good agreement with the typical outer shell diameter for ferritin (~12nm).

Figure 4.1 Morphological Characterizations of Apoferritin and Ferritin.

TEM of unstained (A) apoferritin and (B) ferritin samples show iron cores only in ferritin. Staining revealed protein shells (insets). [182].

4.3.2 Characterizing Ferritin Using MFM

MFM was conducted on ferritin or apoferritin samples with HM probes and at lift heights, z, ranging from 10 to 50nm above the substrates (Figure 4.2A).

Quantitative phase analysis was performed on particles with d~13nm

(ascertained using topographic AFM images), and plotted in Figure 4.2C.

Apoferritin particles primarily exhibited low phase (<10º); only 5% of apoferritin yielded high phase (45.4±23.0º) at z=10nm, which diminished rapidly with increasing z. Ferritin, however, consistently exhibited higher phase contrasts

(104.4±4.9º at z=10nm and 45.9±5.6º at 30nm) than apoferritin. 66

PhasePhaase A Height z = 10nm z = 30nm

45೔

0೔

Apoferritin 0.5μm

Ferritin

B MFM Diameter C MFM Phase

Frequency Phase (Degrees) Diameter, d (nm) Lift Height, z

Figure 4.2 MFM Analysis of Apoferritin and Ferritin (A) MFM images for apoferritin and ferritin. The white arrow points to a high contrast particle, which accounted for 5% of the apoferritin population. (B) Particle diameters from MFM height images. (C) Phase for particles (d~13nm) at increasing z.

4.3.3 Using MFM to Quantify Ferritin in Protein Mixtures

To test whether ferritin could be accurately distinguished from apoferritin when mixed in the same sample, MFM analysis was carried out on samples with varying proportions of the two proteins. Our results show that these mixed

67 samples exhibited a mixed population of particles with either low or high phase contrast (Figure 4.3A). Quantitative phase analysis of ~13nm particles was performed at z = 30nm for each mixed sample. The measured phase shifts were categorized as low contrast if < 10º and high contrast if > 70º (at z = 10 nm) or >

40º (at z = 30 nm). These low and high contrast phase values corresponded well with those of pure apoferritin (0%) and ferritin (100%), respectively (Figure

4.3B).

To quantify total ferritin vs. apoferritin content in a sample, we analyzed the percentage of high contrast particles in mixed samples at z=30nm (Figure

4.3C). High contrast analysis for d~13nm particles closely matched predicted ferritin content in each sample (R2=0.99). We additionally analyzed contrast distribution for all particles to determine whether MFM can distinguish ferritin from apoferritin irrespective of iron-core size or protein agglomeration. High contrast percentages for all particles also closely matched predicted ferritin content in each sample (R2=0.99).

We compared these results to spectrophotometric measurements of protein (Figure 4.3D) and iron (Figure 4.3E) content in samples with total protein concentration of 300μg ml-1 and increasing ferritin content. Protein concentrations, as expected, stayed the same while absorbance for iron concentration increased linearly with percentage of ferritin.

Figure 4.3 Quantitative MFM Analyses of Apoferritin/Ferritin Mixtures. (A)

MFM images of apoferritin/ferritin mixtures containing 1μg ml-1 total protein. (B)

MFM phase values for particles in various samples. Ferritin phase using HM

(shaded area) and MM probes [152] (dotted line) are indicated. (C) Percentage of high contrast particles in MFM phase images matched predicted ferritin content

(R2=0.99) irrespective of particle size. (D) Protein concentration and (E) Iron absorbance for apoferritin/ferritin mixtures with 300μg ml-1 total protein. 69

4.3.4 Detecting Endogenously Expressed Iron/Ferritin in Serum

To evaluate whether we could detect endogenously expressed iron/ferritin in serum samples, we performed MFM on serum from naïve and injured rats (21 days post spinal cord injury); these samples had a pre-determined serum ferritin concentration of 0.5 μg ml-1 (naïve) and 1.7 μg ml-1 (injured), respectively. MFM signal free of sample topography was detectable at z=30nm in the phase images of both sample types (Figure 4.4); these signals were negative, implying that the magnetic interactions between the sample and the MFM probe were repulsive.

Quantitative phase analysis (n=10 particles per sample type) yielded similar phase shifts of -15.7±5.1º for naïve and -13.9±5.9º for injured serum.

2 Areal density DA, or average number of particles per μm in phase images, was also calculated to determine whether there was a detectable difference in iron/ferritin content between naïve and injured serum when measured using MFM

(Figure 4.4C). A modest increase (1.4 times) was observed between the naïve sample (DA = 6.4) and the injured sample (DA = 9.1).

4.3.5 Detecting Endogenously Expressed Iron/Ferritin in Spinal Cord Tissue

The ability of MFM to detect endogenously expressed ferritin in biological samples was further tested on injured rat spinal cord tissue. Ferritin and iron content in these tissues were characterized using Perls staining and ferritin immuno-fluorescence (Figure 4.5A); MFM was subsequently performed on (1) low iron/ferritin and (2) high iron/ferritin regions (Figure 4.5B).

Spinal cord tissue topography diminished at z=30nm leaving only the phase contrast from magnetic components of the sample (Figure 4.5B). MFM phase images of low iron/ferritin region (1) revealed few areas with non-zero

MFM contrast (up to -65ᵒ at z=30nm). It is interesting to note these deposits were unresolvable with iron staining. High iron/ferritin region (2) exhibited even larger areas with high positive or negative phase contrast (±155ᵒ at z=30nm), suggesting iron/ferritin clusters with varying magnetic orientations.

Naïve Injured ELISA 2 A B on 1.51 ) -1 1 g ml μ ( 0.50 Height Concentration Concentration 0 0.25μm naïvee injured

MFM

20೔ ) 101 2 C m

μ 8 6 -20⁰ 4 Phase

z = 30nm 2 Areal Density

(particles per (particles per 0 naïvee injured Figure 4.4 MFM of Endogenously Expressed Iron/Ferritin in Serum.

(A) MFM height and phase images of serum from an uninjured (naïve) and spinal cord injured rat. (B) ELISA shows a 240% increase in serum ferritin in injured samples, (C) where MFM areal density only shows 42% increase.

Figure 4.5 Detection of Endogenously Expressed Iron/Ferritin in Tissue.

Low (1) and a high (2) iron/ferritin regions in an injured rat spinal cord tissue section visualized using (A) Perl’s (iron) and fluorescence (red: ferritin, blue: nuclei) imaging and (B) MFM height and phase imaging. Sample topography is no longer visible in phase images at lift height z=30nm [182].

4.4 Discussion

In this study, we demonstrate how MFM using a HM probe can accurately distinguish purified iron-ligated (ferritin) from non-ligated (apoferritin) proteins.

Our ferritin phase contrasts are approximately three times greater in magnitude than those reported earlier using MFM with medium moment probes [152]. This is in agreement with the increase in phase contrast observed in our previous studies (Chapter 3) where we compared medium and high magnetic moment

MFM probes on synthetic 20nm iron oxide superparamagnetic nanoparticles. The

HM probes therefore enabled us to enhance phase differences between ferritin and apoferritin as compared to the earlier report [152]. The HM probes used in this study therefore enabled us to amplify the MFM signal from ferritin.

Minor deviations in phase contrast could arise due to differences in size or chemical composition of iron cores of ferritin and/or asperities in MFM probes.

Like previous studies of horse spleen ferritin in vitro [152], our MFM phase contrasts were consistently positive. Endogenous rat ferritin, however, produced negative phase contrast in serum and mixed contrast in tissue. This effect could be due to differences in composition of iron-cores in horse spleen vs. rat ferritin.

Further, orientation of ferritin could be affected by factors such as pH, ionic strength and chemical composition of serum or tissue microenvironments. Unlike the synthetic iron-oxide superparamagnetic nanoparticles used in our previous studies, ferritin did not exhibit magnetic anisotropy, thus making MFM analysis relatively straightforward.

MFM could accurately quantify ferritin content in protein mixtures created in vitro. Relative amounts of ferritin in animal serum were also estimated using

MFM. Although sandwich ELISA showed that serum ferritin was over threefold higher in injured vs. naïve rats, there was only a modest increase in ferritin as ascertained using MFM, indicating an up-regulation of apoferritin. Serum also contains transferrin-bound and non-transferrin-bound iron [183,184]. However, these forms of iron consist of only a few iron atoms and are therefore undetectable using MFM. This enables us to directly equate MFM phase contrast to serum ferritin.

We additionally show how MFM can identify iron/ferritin deposits in mammalian tissue samples that were otherwise unresolvable using Perls (iron) and immuno-fluorescence (ferritin) staining. The resolution for detecting iron or ferritin stained samples is limited by the resolution of the microscope used for imaging. The minimal resolution for light and fluorescence microscopy is typically

200nm, which is over fifteen times greater than the resolution needed to detect a single ferritin protein (~13nm). MFM may therefore potentially serve as a novel technique to spatially localize endogenously expressed iron/ferritin at very low concentrations. Further, MFM may also be able to provide spatial information on the fate of synthetic superparamagnetic nanoparticles used in cell-tagging or drug delivery applications.

Finally, our experiments showed that a lift height of z=30nm produced purely magnetic MFM phase signal, which in turn permitted detection and spatial

74 localization of iron/ferritin in serum and tissue samples. Serum and tissue sample topography was present in phase images up to z=10nm, making it difficult to identify phase contrast arising from magnetic moments in the samples. We envisage phase contrast at z ≤ 10nm is confounded by factors like van der Waals interactions and/or differences in electrostatic interactions arising from serum or tissue composition.

In conclusion, we elucidate how MFM can quantify ferritin based on the magnetic signature of iron cores. Major advantages of MFM for analyzing biological samples include its high sensitivity and spatial resolution and the ability to analyze samples with very low volume and ferritin concentrations, typically non-detectable using conventional biochemical assays. Further work employing quantitative estimates of MFM probe magnetic moment and geometry may help quantify the iron content in purified and in endogenously expressed ferritin. MFM may therefore serve as a novel and complementary analytical tool to improve our understanding of iron homeostasis.

CHAPTER 5

Conclusions and Future Work

5.1 Chapter Overview

Here we summarize the advancements we have made with the magnetic force microscopy (MFM) technique for detection of superparamagnetic nanoparticles (SPNs). Our results and future work are discussed in context with the potential for MFM in biomedical applications.

5.2 Summary of MFM Advancements

This work has resulted in the following major contributions towards development of MFM for biomedical applications. Specifically, we have

(1) increased MFM sensitivity to detect single superparamagnetic

nanoparticles (SPNs) and

(2) detected iron-bound proteins in-vitro and in serum and tissue samples.

These contributions are summarized and discussed in the following subsections along with scope for future developments.

5.2.1 Increasing MFM Sensitivity to Detect Single SPNs

We had previously shown how the application of an external magnetic field could enable the detection and quantitative characterization of SPNs using

MFM [155]. This work was performed on SPN aggregates. One of our goals was to further improve the sensitivity of MFM to enable detection of SPNs at the single particle level. We identified two approaches to achieve this goal, (1) increase MFM sensitivity by exploiting SPN magnetic anisotropy and (2) increase

MFM sensitivity by increasing the magnetic moment of the MFM probe.

Previous MFM studies [152,155,167] on SPN aggregates had resulted in a heterogeneous distribution of phase shifts most likely due to dipole-dipole interactions and random easy axis orientations in SPNs within the agglomerations. In our work [132], we elucidated how monodispersed SPN samples with random orientations of easy axes could produce a large range of phase shifts, making quantification of SPN magnetic moments rather complicated. By pre-aligning the samples so that SPN easy axes aligned parallel to the magnetic moment of the MFM probe, we were able to produce a more uniform and enhanced MFM signal from single SPNs. Controlling for how SPN easy axes are oriented in a sample with respect to external magnetic fields can therefore maximize the magnetic response and simplify analysis of SPN magnetic moments using MFM.

In our second approach, we demonstrated how increasing the magnetic moment of the MFM probe could increase MFM sensitivity. A high magnetic

77 moment probe not only increased the phase signal over the medium moment probe at each lift height, but also increased the lift height at which a single SPN could be detected from z=30nm (medium moment probe) to z=70nm. Together these findings indicate that both the exploitation of magnetic anisotropy and the use of HM probes can significantly enhance MFM sensitive for SPNs detection.

5.2.2 Detecting Iron-bound Proteins in vitro and in Serum and Tissue Samples

We performed MFM using a high magnetic moment probe as a way to amplify the phase contrast of single purified ferritin proteins beyond those reported previously using medium magnetic moment probes [152]. This therefore enabled us to accurately distinguish ferritin from apoferritin in a variety of samples with mixed ferritin and apoferritin content. It also allowed us to detect iron/ferritin in serum and tissue samples from rats with increased serum ferritin and tissue ferritin content. This work further demonstrated how MFM could overcome the limitations of immuno-based (i.e. enzyme-linked immunosorbent assays, protein colorimetric assays, immunohistochemistry) and iron-based (i.e. iron-chelation and spectrophotometry, Perls staining) detection techniques to identify ferritin (from apoferritin) in vitro and in biological samples. Unlike the existing techniques, MFM offers the ability to spatially localize iron/ferritin at very low samples volumes (<50μl) and concentrations (1μg ml-1).

5.3 Future Outlook: High-throughput MFM

By increasing MFM sensitivity, we could surpass the sensitivity of other current detection techniques to accurately identify low concentrations of ferritin

(from apoferritin) in vitro and in tissue and serum samples. Our findings also serve as a foundation for the development of a novel indirect magnetic force microscopy (ID-MFM) technique applicable for high-throughput detection of

SPNs.

5.3.1 Indirect MFM (ID-MFM) Technique

The ID-MFM technique was conceived from a previous report [185] in which an MFM probe was used to manipulate micron-sized magnetic beads through a membrane. In ID-MFM, the sample is immobilized on Side A of an ultrathin (5nm to 50nm) membrane. The MFM probe is then scanned over Side B and magnetic interactions between the probe and sample are detected through the membrane (Figure 5.1). The ID-MFM technique would therefore have three major advantages over conventional MFM, especially for biological applications:

(1) MFM imaging at multiple lift-heights would not be required, (2) the probe would be physically separated from the sample and thus avoids becoming contaminated by the biological samples and (3) the sample could potentially be kept hydrated in a fluid environment without dampening the MFM cantilever.

High moment MFM Probe

air

Side B Ultrathin membrane Side A (10nm – 50nm)

Fluid or air MFM Signal Non-magnetic particle Magnetic particle

Figure 5.1 Schematic of Indirect Magnetic Force Microscopy.

A sample (in fluid or air) is placed on Side A of an ultrathin membrane. A high moment MFM probe is then scanned on Side B and MFM signals from the magnetic components of the sample are detected through the membrane, allowing the MFM probe to remain in air.

5.3.2 Silicon Nitride Membranes

ID-MFM requires a membrane that must be robust enough to withstand contact forces of the MFM probe, yet thin enough to enable detection of weak superparamagnetic nanoparticles through the membrane. We propose the use of silicon nitride membranes as the sample platform for the ID-MFM technique.

Silicon nitride membranes are a recent development in biological microscopy; they are transparent to light and electron optics and can be made as thin as

10nm, thus offering a competitive advantage over glass cover slips, which can only be fracture-free above a thickness of 100nm. In addition, silicon nitride 80 membranes are stable and fracture-resistant even in a fluid environment, making them attractive for biological applications.

Silicon nitride membranes also find applications in light, fluorescence and electron microscopies in dry [186–188] and fluidic [189–192] environments, which could permit multimodal imaging on the same samples undergoing ID-

MFM. Additional advantages of ID-MFM include a quicker imaging time since multiple lift heights are not required, and elimination of possible probe contamination that can occur when the probe tip comes in direct contact with soft and/or biological samples. Since ID-MFM may be performed on any commercially available MFM, we envisage the ID-MFM technique to be a low- cost, easy-to-use and high-throughput method for the nanoscale magnetic characterization of biological samples.

5.3.3 ID-MFM of Ferritin

As a feasibility experiment, we dried ferritin proteins onto Side A of a custom-made silicon nitride platform of thickness 30 nm obtained collaborator,

John Moreland from NIST (Figure 5.2A) High magnetic moment (HM) MFM probes were used to scan the sample from side B. As expected, contrast was not observable in the height images collected on Side B for either the membrane without (-) or the membrane with (+) ferritin. Phase contrast of 66.6 ± 7.8ᵒ was present only in the (+) ferritin sample (n=25 particles); these measurements were in good agreement with ferritin imaged at a lift height of z=30nm using

81 conventional MFM and a high moment probe (Figures 4.2 and 4.3B) [182], suggesting ID-MFM to have comparable sensitivity to conventional MFM.

Indirect MFM Height Phase A B (-) ferritin 20ᵒ

0ᵒ 250nm

C (+) ferritin

Figure 5.2 Indirect MFM of Ferritin A custom platform (A) consisting of a 30nm thick silicon nitride membrane suspended over a silicon support frame (purple), with six membrane windows (courtesy of John Moreland, NIST). Indirect MFM was performed from Side B of the membrane windows, and height and phase images were collected (B) without (-) and (C) with ferritin immobilized on Side A of the membrane. Phase contrast (66.6 ± 7.8ᵒ) was present only in the (+) ferritin sample.

5.4 Conclusions: MFM in Biomedicine

In order to fully understand how SPNs interact with or behave within biological systems, it is critical to ascertain high-fidelity characterization of SPN magnetic properties, aggregation, concentration and spatial distribution. This thesis elucidates how MFM can be used to detect SPNs with very high sensitivity and spatial resolution beyond that offered by current detection techniques. We have further demonstrated MFM’s applicability in biology by detecting low concentrations of naturally occurring SPNs in vitro and in serum and tissue samples without the need of stains or antibodies. We envisage our advancements will serve as a stepping stone toward future applications of MFM for synthetic and natural SPN characterization for a wide range of biomedical applications, including earlier detection of pathologies characterized by imbalanced iron homeostasis.

Although we have significantly improved the sensitivity of MFM and enabled detection of SPNs endogenously expressed in biological samples, there are still some remaining obstacles before MFM can reach its full potential for biomedical applications. These include:

(1) quantitative determination of iron content in samples,

(2) high-throughput imaging and

(3) SPN detection in fluid environments.

We suggest three future research focuses to overcome these obstacles. First, work dedicated to the quantitative approximation of MFM probe magnetic

83 moment and geometry may lead to a more quantitative estimation of SPN magnetic moments in biological samples. Second, future work aimed at developing the ID-MFM technique could enable high-throughput imaging. Studies focused on optimizing ID-MFM membrane thickness and window area will be critical to permit quick yet accurate detection of magnetic particles through the membrane. Third, future work to design and test a microfluidic platform for ID-

MFM could lead to highly sensitive detection and spatial localization of SPNs in a fluid environment.

By overcoming the aforementioned obstacles, MFM could serve as a powerful tool to understand the complex magnetic properties of synthetic and naturally occurring SPNs. We additionally foresee MFM as a novel biosensor to characterize nanoscale iron deposits in biological samples ex vivo. These include cells or tissue with endogenous SPN expression or those that have been targeted for cell-sorting, MRI contrast agents, hyperthermia therapy, or drug delivery. MFM could also find future applications as a novel “magnetic histology” technique, where antibody-conjugated SPNs targeted to specific cell or tissue sites could be spatially localized with resolution far surpassing that of current light and fluorescence-based immuno-detection techniques. Taken together, our results and future potentials establish MFM as a low-cost, highly-sensitive and easy-to-use solution for detecting SPNs for biomedical applications.

REFERENCES

1. Bronen, R. A. & Sze, G. Magnetic resonance imaging contrast agents:

theory and application to the central nervous system. Journal of

Neurosurgery 73, 820–839 (1990).

2. Nanoparticles: From Theory to Applications. (Wiley-VCH: Weinheim,

2004).

3. Block, S., Glockl, G., Weitschies, W. & Helm, C. A. Direct Visualization and

Identification of Biofunctionalized Nanoparticles using a Magnetic Atomic

Force Microscope. Nano Letters 11, 3587–3592 (2011).

4. Kang, S. S. et al. Microstructures and magnetic alignment of L1(0) FePt

nanoparticles. Journal of Applied Physics 101, 09J113 (2007).

5. Sun, S. et al. Monodisperse MFe2O4 (M = Fe, Co, Mn) nanoparticles.

Journal of the American Chemical Society 126, 273–279 (2004).

6. Kuswik, P. et al. Colloidal domain lithography for regularly arranged

artificial magnetic out-of-plane monodomains in Au/Co/Au layers.

Nanotechnology 22, 095302 (2011).

7. Tannous, C. & Gieraltowski, J. The Stoner-Wohlfarth model of

ferromagnetism. European Journal of Physics 29, 475–487 (2008).

8. Adolphi, N. L. et al. Characterization of single-core magnetite nanoparticles

for magnetic imaging by SQUID relaxometry. Physics in Medicine and

Biology 55, 5985–6003 (2010).

9. Kötitz, R., Weitschies, W., Trahms, L. & Semmler, W. Investigation of

Brownian and Néel relaxation in magnetic fluids. Journal of Magnetism and

Magnetic Materials 201, 102–104 (1999).

10. Kim, D. K., Zhang, Y., Voit, W., Rao, K. V. & Muhammed, M. Synthesis and

characterization of surfactant-coated superparamagnetic monodispersed

iron oxide nanoparticles. Journal of Magnetism and Magnetic Materials

225, 30–36 (2001).

11. Gossuin, Y., Gillis, P., Hocq, A., Vuong, Q. L. & Roch, A. Magnetic

resonance relaxation properties of superparamagnetic particles. Wiley

Interdisciplinary Reviews: Nanomedicine and Nanobiotechnology 1, 299–

310 (2009).

12. Mahmoudi, M., Laurent, S. & Journeay, W. S. Cell life cycle effects of bare

and coated superparamagnetic iron oxide nanoparticles. Toxic Effects of

Nanomaterials 53–65 (2012).

13. Liu, L. et al. Decreased reticuloendothelial system clearance and increased

blood half-life and immune cell labeling for nano- and micron-sized

superparamagnetic iron-oxide particles upon pre-treatment with Intralipid.

Biochimica et Biophysica Acta 1830, 3447–53 (2013).

14. Mahmoudi, M., Simchi, A. & Imani, M. Recent advances in surface

engineering of superparamagnetic iron oxide nanoparticles for biomedical

applications. Journal of the Iranian Chemical Society 7, S1–S27 (2010).

15. Lu, M., Cohen, M. H., Rieves, D. & Pazdur, R. FDA report: Ferumoxytol for

intravenous iron therapy in adult patients with chronic kidney disease.

American Journal of Hematology 85, 315–319 (2010).

16. Wang, Y.-X. J. Superparamagnetic iron oxide based MRI contrast agents:

Current status of clinical application. Quantitative Imaging in Medicine and

Surgery 1, 35–40 (2011).

17. Thomas, R., Park, I.-K. & Jeong, Y. Y. Magnetic iron oxide nanoparticles

for multimodal imaging and therapy of cancer. International Journal of

Molecular Sciences 14, 15910–15930 (2013).

18. Neri, M. et al. Efficient in vitro labeling of human neural precursor cells with

superparamagnetic iron oxide particles: relevance for in vivo cell tracking.

Stem Cells 26, 505–516 (2008).

19. Chen, C.-C. V et al. Simple SPION incubation as an efficient intracellular

labeling method for tracking neural progenitor cells using MRI. PloS one 8,

e56125 (2013).

20. Miyoshi, S. et al. Transfection of neuroprogenitor cells with iron

nanoparticles for magnetic resonance imaging tracking: cell viability,

differentiation, and intracellular localization. Molecular Imaging and Biology

7, 286–295 (2005).

21. Mettler, E. et al. Magnetic separation of encapsulated islet cells labeled

with superparamagnetic iron oxide nano particles. Xenotransplantation 20,

219–226 (2013).

22. Guardia, P. et al. Water-soluble iron oxide nanocubes with high values of

specific absorption rate for cancer cell hyperthermia treatment. ACS Nano

6, 3080–3091 (2012).

23. Zheng, S. et al. RGD-conjugated iron oxide magnetic nanoparticles for

magnetic resonance imaging contrast enhancement and hyperthermia.

Journal of Biomaterials Applications Epub, ahead of print (2013).

24. Weissleder, R., Lee, A. S., Khaw, B. A., Shen, T. & Brady, T. J.

Antimyosin-labeled monocrystalline iron oxide allows detection of

myocardial infarct: MR antibody imaging. Radiology 182, 381–385 (1992).

25. Tiefenauer, L. X., Kühne, G. & Andres, R. Y. Antibody-magnetite

nanoparticles: in vitro characterization of a potential tumor-specific contrast

agent for magnetic resonance imaging. Bioconjugate Chemistry 4, 347–

352

26. Suzuki, M. et al. Development of a target-directed magnetic resonance

contrast agent using monoclonal antibody-conjugated magnetic particles.

Brain Tumor Pathology 13, 127–132 (1996).

27. Suwa, T., Ozawa, S., Ueda, M., Ando, N. & Kitajima, M. Magnetic

resonance imaging of esophageal squamous cell carcinoma using

magnetite particles coated with anti-epidermal growth factor receptor

antibody. International Journal of Cancer 75, 626–634 (1998).

28. Shinkai, M. et al. Targeting hyperthermia for renal cell carcinoma using

human MN antigen-specific magnetoliposomes. Japanese Journal of

Cancer 92, 1138–1145 (2001).

29. Yang, L. et al. Single chain epidermal growth factor receptor antibody

conjugated nanoparticles for in vivo tumor targeting and imaging. Small 5,

235–243 (2009).

30. Natarajan, A. et al. NanoFerrite particle based radioimmunonanoparticles:

binding affinity and in vivo pharmacokinetics. Bioconjugate Chemistry 19,

1211–1218 (2008).

31. Zhou, Y. et al. Impact of single-chain Fv antibody fragment affinity on

nanoparticle targeting of epidermal growth factor receptor-expressing

tumor cells. Journal of Molecular Biology 371, 934–947 (2007).

32. Lee, C.-S., Chang, H. H., Bae, P.-K., Jung, J. & Chung, B. H. Bifunctional

nanoparticles constructed using one-pot encapsulation of a fluorescent

polymer and magnetic (Fe3O4) nanoparticles in a silica shell.

Macromolecular Bioscience 13, 321–331 (2013).

33. Panahifar, A., Mahmoudi, M. & Doschak, M. R. Synthesis and in vitro

evaluation of bone-seeking superparamagnetic iron oxide nanoparticles as

contrast agents for imaging bone metabolic activity. ACS Applied Materials

& Interfaces 5, 5219–5226 (2013).

34. Liu, C. et al. Specific targeting of angiogenesis in lung cancer with RGD-

conjugated ultrasmall superparamagnetic iron oxide particles using a 4.7T

magnetic resonance scanner. Chinese Medical Journal 126, 2242–2247

(2013).

35. Greish, K. Enhanced permeability and retention of macromolecular drugs

in solid tumors: a royal gate for targeted anticancer nanomedicines.

Journal of Drug Targeting 15, 457–464 (2008).

36. Iyer, A. K., Khaled, G., Fang, J. & Maeda, H. Exploiting the enhanced

permeability and retention effect for tumor targeting. Drug Discovery Today

11, 812–818 (2006).

37. Sohn, Y. S. & Jun, Y. J. Poly- and Cyclophosphazenes as Drug Carriers for

Anticancer Therapy. Polyphosphazenes for Biomedical Applications 249–

276 (2009).

38. Charrois, G. J. R. & Allen, T. M. Multiple injections of pegylated liposomal

Doxorubicin: pharmacokinetics and therapeutic activity. The Journal of

Pharmacology and Experimental Therapeutics 306, 1058–1067 (2003).

39. Hashizume, H. et al. Openings between defective endothelial cells explain

tumor vessel leakiness. The American Journal of Pathology 156, 1363–

1380 (2000).

40. Hobbs, S. K. et al. Regulation of transport pathways in tumor vessels: role

of tumor type and microenvironment. Proceedings of the National Academy

of Sciences of the United States of America 95, 4607–4612 (1998).

41. Cho, Z. H., Jones, J. P. & Singh, M. Foundations of Medical Imaging.

(Wiley: New York, 1993).

42. Wang, Y. X., Hussain, S. M. & Krestin, G. P. Superparamagnetic iron oxide

contrast agents: physicochemical characteristics and applications in MR

imaging. European Radiology 11, 2319–2331 (2001).

43. Bulte, J. W. et al. Magnetoferritin: characterization of a novel

superparamagnetic MR contrast agent. Journal of Magnetic Resonance

Imaging 4, 497–505 (1994).

44. Okuda, M., Iwahori, K., Yamashita, I. & Yoshimura, H. Fabrication of nickel

and chromium nanoparticles using the protein cage of apoferritin.

Biotechnology and Bioengineering 84, 187–194 (2003).

45. Na, H. Bin, Song, I. C. & Hyeon, T. Inorganic Nanoparticles for MRI

Contrast Agents. Advanced Materials 21, 2133–2148 (2009).

46. Weinstein, J. S. et al. Superparamagnetic iron oxide nanoparticles:

diagnostic magnetic resonance imaging and potential therapeutic

applications in neurooncology and central nervous system inflammatory

pathologies, a review. Journal of Cerebral Blood Flow and Metabolism 30,

15–35 (2010).

47. Shapiro, E. M. et al. MRI detection of single particles for cellular imaging.

Proceedings of the National Academy of Sciences of the United States of

America 101, 10901–10906 (2004).

48. Cherry, S. R. Multimodality imaging: beyond PET/CT and SPECT/CT.

Seminars in Nuclear Medicine 39, 348–353 (2009).

49. Yang, X. et al. cRGD-functionalized, DOX-conjugated, and 64Cu-labeled

superparamagnetic iron oxide nanoparticles for targeted anticancer drug

delivery and PET/MR imaging. Biomaterials 32, 4151–4160 (2011).

50. Lee, H.-Y. et al. PET/MRI dual-modality tumor imaging using arginine-

glycine-aspartic (RGD)-conjugated radiolabeled iron oxide nanoparticles.

Journal of Nuclear Medicine 49, 1371–1379 (2008).

51. Torres Martin de Rosales, R. et al. Synthesis of 64Cu(II)-

bis(dithiocarbamatebisphosphonate) and its conjugation with

superparamagnetic iron oxide nanoparticles: in vivo evaluation as dual-

modality PET-MRI agent. Angewandte Chemie (International Ed.) 50,

5509–5513 (2011).

52. Kim, D., Kim, J. W., Jeong, Y. Y. & Jon, S. Antibiofouling Polymer Coated

Gold@Iron Oxide Nanoparticle (GION) as a Dual Contrast Agent for CT

and MRI. Bulletin- Korean Chemical Society 30, 1855–1857 (2009).

53. Kim, D. et al. Amphiphilic polymer-coated hybrid nanoparticles as CT/MRI

dual contrast agents. Nanotechnology 22, 155101 (2011).

54. Wang, D., He, J., Rosenzweig, N. & Rosenzweig, Z. Superparamagnetic

Fe 2 O 3 Beads−CdSe/ZnS Quantum Dots Core−Shell Nanocomposite

Particles for Cell Separation. Nano Letters 4, 409–413 (2004).

55. Park, J.-H., Von Maltzahn, G., Ruoslahti, E., Bhatia, S. N. & Sailor, M. J.

Micellar hybrid nanoparticles for simultaneous magnetofluorescent imaging

and drug delivery. Angewandte Chemie (International Ed.) 47, 7284–7288

(2008).

56. Josephson, L., Kircher, M. F., Mahmood, U., Tang, Y. & Weissleder, R.

Near-infrared fluorescent nanoparticles as combined MR/optical imaging

probes. Bioconjugate chemistry 13, 554–60

57. He, X., Gao, J., Gambhir, S. S. & Cheng, Z. Near-infrared fluorescent

nanoprobes for cancer molecular imaging: status and challenges. Trends

in Molecular Medicine 16, 574–583 (2010).

58. Hayashi, K. et al. Superparamagnetic nanoparticle clusters for cancer

theranostics combining magnetic resonance imaging and hyperthermia

treatment. Theranostics 3, 366–376 (2013).

59. Laurent, S., Dutz, S., Häfeli, U. O. & Mahmoudi, M. Magnetic fluid

hyperthermia: Focus on superparamagnetic iron oxide nanoparticles.

Advances in Colloid and Interface Science 166, 8–23 (2011).

60. Santra, S., Kaittanis, C., Grimm, J. & Perez, J. M. Drug/dye-loaded,

multifunctional iron oxide nanoparticles for combined targeted cancer

therapy and dual optical/magnetic resonance imaging. Small 5, 1862–1868

(2009).

61. Quan, Q. et al. HSA coated iron oxide nanoparticles as drug delivery

vehicles for cancer therapy. Molecular Pharmaceutics 8, 1669–1676

(2011).

62. Yu, M. K. et al. Drug-loaded superparamagnetic iron oxide nanoparticles

for combined cancer imaging and therapy in vivo. Angewandte Chemie

(International Ed.) 47, 5362–5365 (2008).

63. Lim, J., Clements, M. A. & Dobson, J. Delivery of short interfering

ribonucleic acid-complexed magnetic nanoparticles in an oscillating field

occurs via caveolae-mediated endocytosis. PloS one 7, e51350 (2012).

64. Dobson, J. Gene therapy progress and prospects: magnetic nanoparticle-

based gene delivery. Gene Therapy 13, 283–287 (2006).

65. Waerzeggers, Y. et al. Methods to monitor gene therapy with molecular

imaging. Methods 48, 146–160 (2009).

66. Chen, Y. et al. Development of an MRI-visible nonviral vector for siRNA

delivery targeting gastric cancer. International Journal of Nanomedicine 7,

359–368 (2012).

67. Chasteen, N. D. & Harrison, P. M. Mineralization in Ferritin: An Efficient

Means of Iron Storage. Journal of Structural Biology 126, 182–194 (1999).

68. Wixom, R. L., Prutkin, L. & Munro, H. N. Hemosiderin: nature, formation,

and significance. International review of experimental pathology 22, 193–

225 (1980).

69. Wunderlich, J., Rohrbach, P. & Dalton, J. P. The malaria digestive vacuole.

Frontiers in Bioscience 4, 1424–1448 (2012).

70. Crichton, R. R. Ferritin. Metal Bonding in Proteins 67–134 (1973).

71. Kilcoyne, S. H. & Cywinski, R. Ferritin: a model superparamagnet. Journal

of Magnetism and Magnetic Materials 140-144, 1466–1467 (1995).

72. Jambor, J. L. & Dutrizac, J. E. Occurrence and constitution of natural and

synthetic ferrihydrite, a widespread iron oxyhydroxide. Chemical Review

98, 2549–2586 (1998).

73. Papaefthymiou, G. C. The Mössbauer and magnetic properties of ferritin

cores. Biochimica et Biophysica Acta 1800, 886–897 (2010).

74. Patel, S., Monemian, S., Khalid, A. & Dosik, H. Iron Deficiency Anemia in

Adult Onset Still’s Disease with a Serum Ferritin of 26,387 μg/L. Anemia

2011, 184748 (2011).

75. Smith, R. J., Davis, P., Thomson, A. B., Wadsworth, L. D. & Fackre, P.

Serum ferritin levels in anemia of rheumatoid arthritis. The Journal of

Rheumatology 4, 389–392 (1977).

76. Bari, M. A. et al. Assessment of anaemia in patients with rheumatoid

arthritis. Mymensingh Medical Journal 22, 248–254 (2013).

77. Fishbane, S., Kalantar-Zadeh, K. & Nissenson, A. R. Serum ferritin in

chronic kidney disease: recondsidering the upper limit of iron treatment.

Seminars in Dialysis 17, 336–341 (2004).

78. Torti, F. M. Regulation of ferritin genes and protein. Blood 99, 3505–3516

(2002).

79. Bothwell, T. H. & MacPhail, A. P. Hereditary hemochromatosis: etiologic,

pathologic, and clinical aspects. Seminars in Hematology 35, 55–71

(1998).

80. Kushner, J. P. Secondary Iron Overload. Hematology 2001, 47–61 (2001).

81. Sullivan, J. L. Iron in arterial plaque: modifiable risk factor for

atherosclerosis. Biochimica et Biophysica Acta 1790, 718–723 (2009).

82. Ganne-Carrie, N. et al. Liver iron is predictive of death in alcoholic

cirrhosis: a multivariate study of 229 consecutive patients with alcoholic

and/or hepatitis C virus cirrhosis: a prospective follow up study. GUT 46,

277–282 (2000).

83. Dexter, D. T. et al. Alterations in the levels of iron, ferritin and other trace

metals in Parkinson’s Disease and other neurodegenerative diseases

affecting the basal ganglia. Brain 114, 1953–1975 (1991).

84. Jellinger, K., Paulus, W., Grundke-Iqbal, I., Riederer, P. & Youdim, M. B. H.

Brain iron and ferritin in Parkinson’s and Alzheimer's diseases. Journal of

Neural Transmission 2, 327–340 (1990).

85. Mariani, S. et al. Effects of hemochromatosis and transferrin gene

mutations on peripheral iron dyshomeostasis in mild cognitive impairment

and Alzheimer’s and Parkinson's diseases. Frontiers in Aging

Neuroscience 5, 1–7 (2013).

86. Rouault, T. A. Iron metabolism in the CNS: implications for

neurodegenerative diseases. Nature Reviews. Neuroscience 14, 551–564

(2013).

87. Viles, J. H. Metal ions and amyloid fiber formation in neurodegenerative

diseases. Copper, zinc and iron in Alzheimer’s, Parkinson's and prion

diseases. Coordination Chemistry Reviews 256, 2271–2284 (2012).

88. Wagner, K. R., Sharp, F. R., Ardizzone, T. D., Lu, A. & Clark, J. F. Heme

and iron metabolism: role in cerebral hemorrhage. Journal of Cerebral

Blood Flow and Metabolism 23, 629–652 (2003).

89. Wu, J. et al. Iron and iron-handling proteins in the brain after intracerebral

hemorrhage. Stroke 34, 2964–2969 (2003).

90. Hua, Y. et al. Long-term effects of experimental intracerebral hemorrhage:

the role of iron. Journal of Neurosurgery 104, 305–312 (2006).

91. Rakshit, T. & Mukhopadhyay, R. Tuning band gap of holoferritin by metal

core reconstitution with Cu, Co, and Mn. Langmuir 27, 9681–9686 (2011).

92. Rakshit, T. & Mukhopadhyay, R. Solid-state electron transport in Mn-, Co-,

holo-, and Cu-ferritins: force-induced modulation is inversely linked to the

protein conductivity. Journal of Colloid and Interface Science 388, 282–292

(2012).

93. Amos, F. F., Cole, K. E., Meserole, R. L., Gaffney, J. P. & Valentine, A. M.

Titanium mineralization in ferritin: a room temperature nonphotochemical

preparation and biophysical characterization. Journal of Biological

Inorganic Chemistry 18, 145–152 (2013).

94. Wong, K. K. W. & Mann, S. Biomimetic synthesis of cadmium sulfide-

ferritin nanocomposites. Advanced Materials 8, 928–932 (1996).

95. Gálvez, N. et al. Comparative structural and chemical studies of ferritin

cores with gradual removal of their iron contents. Journal of the American

Chemical Society 130, 8062–8068 (2008).

96. Martínez-Pérez, M. J. et al. Size-dependent properties of magnetoferritin.

Nanotechnology 21, 465707 (2010).

97. Makhlouf, S. A., Parker, F. T. & Berkowitz, A. E. Magnetic hysteresis

anomalies in ferritin. Physical Review B 55, 14717–14720 (1997).

98. Allen, P., St Pierre, T. & Street, R. Magnetic interactions in native horse

spleen ferritin below the superparamagnetic blocking temperature. Journal

of Magnetism and Magnetic Materials 177, 1459–1460 (1998).

99. Johansson, C., Hanson, M., Pedersen, M. S. & Mørup, S. Magnetic

properties of magnetic liquids with iron-oxide particles — The influence of

anisotropy and interactions. Journal of Magnetism and Magnetic Materials

173, 5–14 (1997).

100

100. Foley, C. P. & Hilgenkamp, H. Why NanoSQUIDs are important: an

introduction to the focus issue. Superconductor Science and Technology

22, 064001 (2009).

101. Vohralik, P. F. & Lam, S. K. H. NanoSQUID detection of magnetization

from ferritin nanoparticles. Superconductor Science and Technology 22,

064007 (2009).

102. Bouchiat, V. Detection of magnetic moments using a nano-SQUID: limits of

resolution and sensitivity in near-field SQUID magnetometry.

Superconductor Science and Technology 22, 064002 (2009).

103. Canavese, C. et al. Validation of serum ferritin values by magnetic

susceptometry in predicting iron overload in dialysis patients. Kidney

International 65, 1091–1098 (2004).

104. Brittenham, G. M. & Badman, D. G. Noninvasive measurement of iron:

report of an NIDDK workshop. Blood 101, 15–19 (2003).

105. Maliken, B. D. et al. Room-temperature susceptometry predicts biopsy-

determined hepatic iron in patients with elevated serum ferritin. Annals of

Hepatology 11, 77–84 (2012).

106. Wertheim, G. K. Mossbauer Effect: Principles and Applications. 116

(1964).

101

107. McCammon, C. Mineral Physics & Crystallography: A Handbook of

Physical Constants. AGU Reference Shelf 2, 332–347 (1995).

108. Dubiel, S. M., Zablotna-Rypien, B. & Mackey, J. B. Magnetic properties of

human liver and brain ferritin. European Biophysics Journal 28, 263–267

(1999).

109. Mikhaylova, M. et al. Superparamagnetism of Magnetite

Nanoparticles: Dependence on Surface Modification. Langmuir 20, 2472–

2477 (2004).

110. Vijayakumar, R., Koltypin, Y., Felner, I. & Gedanken, A. Sonochemical

synthesis and characterization of pure nanometer-sized Fe3O4 particles.

Materials Science and Engineering: A 286, 101–105 (2000).

111. Lehlooh, A.-F. & Mahmood, S. H. Mössbauer spectroscopy of Fe3O4

ultrafine particles. Journal of Magnetism and Magnetic Materials 151, 163–

166 (1995).

112. Sorescu, M. et al. A Mössbauer study of the Verwey transition in cobalt-

doped magnetite. Journal of Magnetism and Magnetic Materials 246, 399–

403 (2002).

102

113. Haun, J. B., Yoon, T.-J., Lee, H. & Weissleder, R. Magnetic nanoparticle

biosensors. Wiley Interdisciplinary Reviews: Nanomedicine and

Nanobiotechnology 2, 291–304 (2010).

114. Anderson, L. Assessment of iron overload with T2* magnetic resonance

imaging. Progress of Cardiovasccular Diseases 54, 287–294 (2011).

115. Virtanen, J. et al. Iron overload: accuracy of in-phase and out-of-phase

MRI as a quick method to evaluate liver iron load in haematological

malignancies and chronic liver disease. British Journal of Radiology Epub,

e162–167 (2011).

116. Marinelli, M., Gianesin, B., Avignolo, C., Minganti, V. & Parodi, S. Iron

overload detection in rats by means of a susceptometer operating at room

temperature. Physics in Medicine and Biology 53, 6849–6860 (2008).

117. Perez, J. M., Josephson, L., O’Loughlin, T., Högemann, D. & Weissleder,

R. Magnetic relaxation switches capable of sensing molecular interactions.

Nature Biotechnology 20, 816–820 (2002).

118. Wang, S. X. & Li, G. L. G. Advances in Giant Magnetoresistance

Biosensors With Magnetic Nanoparticle Tags: Review and Outlook. IEEE

Transactions on Magnetics 44, 1687 – 1702 (2008).

103

119. Conradie, J. D. & Mbhele, B. E. Quantitation of serum ferritin by enzyme-

linked immunosorbent assay (ELISA). South African Medical Journal 57,

282–287 (1980).

120. Lee, J. C., Lee, S. S., Schlesinger, K. J. & Richter, G. W. Production of

ferritin by rat hepatoma cells in vitro. Demonstration of protein subunits and

ferritin by immunofluorescence. The American Journal of Pathology 80,

235–248 (1975).

121. Lee, J. C., Lee, S. S., Schlesinger, K. J. & Richter, G. W. Detection of

protein subunits of ferritin in situ in cells by immunofluorescence. The

American Journal of Pathology 75, 473–487 (1974).

122. Sauerbeck, A., Schonberg, D. L., Laws, J. L. & McTigue, D. M. Systemic

iron chelation results in limited functional and histological recovery after

traumatic spinal cord injury in rats. Experimental Neurology 248, 53–61

(2013).

123. Jasmin et al. Labeling stem cells with superparamagnetic iron oxide

nanoparticles: analysis of the labeling efficacy by microscopy and magnetic

resonance imaging. Methods in Molecular Biology 906, 239–252 (2012).

124. Reynolds, F., O’loughlin, T., Weissleder, R. & Josephson, L. Method of

determining nanoparticle core weight. Analytical Chemistry 77, 814–817

(2005).

104

125. Riemer, J., Hoepken, H. H., Czerwinska, H., Robinson, S. R. & Dringen, R.

Colorimetric ferrozine-based assay for the quantitation of iron in cultured

cells. Analytical Biochemistry 331, 370–375 (2004).

126. Uraz, S., Aygun, C., Sonsuz, A. & Ozbay, G. Serum iron levels and hepatic

iron overload in nonalcoholic steatohepatitis and chronic viral hepatitis.

Digestive Diseases and Sciences 50, 964–969 (2005).

127. Bonilla, S. et al. Neonatal iron overload and tissue siderosis due to

gestational alloimmune liver disease. Journal of Hepatology 56, 1351–1355

(2012).

128. Nicolas, G. et al. Lack of hepcidin gene expression and severe tissue iron

overload in upstream stimulatory factor 2 (USF2) knockout mice.

Proceedings of the National Academy of Sciences of the United States of

America 98, 8780–8785 (2001).

129. O’Glasser, A. Y. et al. Hepatic and cardiac iron overload among patients

with end-stage liver disease referred for liver transplantation. Clinical

Transplantation 24, 643–651

130. Kramer, J. H. et al. d-Propranolol protects against oxidative stress and

progressive cardiac dysfunction in iron overloaded rats. Canadian Journal

of Physiology and Pharmacology 90, 1257–1268 (2012).

105

131. Wong, K. K. W., Douglas, T., Gider, S., Awschalom, D. D. & Mann, S.

Biomimetic Synthesis and Characterization of Magnetic Proteins

(Magnetoferritin). Chemistry of Materials 10, 279–285 (1998).

132. Nocera, T. M., Chen, J., Murray, C. B. & Agarwal, G. Magnetic anisotropy

considerations in magnetic force microscopy studies of single

superparamagnetic nanoparticles. Nanotechnology 23, 495704 (2012).

133. Gupta, A. K. & Wells, S. Surface-Modified Superparamagnetic

Nanoparticles for Drug Delivery: Preparation, Characterization, and

Cytotoxicity Studies. IEEE Transactions on Nanobioscience 3, 66–73

(2004).

134. Pavon, L. F. et al. Ultrastructural characterization of CD133+ stem cells

bound to superparamagnetic nanoparticles: possible biotechnological

applications. Journal of Microscopy 231, 374–383 (2008).

135. Aryal, S. et al. Engineered magnetic hybrid nanoparticles with enhanced

relaxivity for tumor imaging. Biomaterials 34, 7725–7732 (2013).

136. Bordat, C. et al. Distribution of iron oxide nanoparticles in rat lymph nodes

studied using electron energy loss spectroscopy (EELS) and electron

spectroscopic imaging (ESI). Journal of Magnetic Resonance Imaging 12,

505–509 (2000).

106

137. Tang, W., Su, Y., Li, Q., Gao, S. & Shang, J. K. Well-dispersed, ultrasmall,

superparamagnetic magnesium ferrite nanocrystallites with controlled

hydrophilicity/hydrophobicity and high saturation magnetization. RSC

Advances 3, 13961 (2013).

138. Parracino, A. et al. Towards nanoscale biomedical devices in medicine:

biofunctional and spectroscopic characterization of superparamagnetic

nanoparticles. Journal of Fluorescence 21, 663–672 (2011).

139. Dormer, K. et al. Epithelial internalization of superparamagnetic

nanoparticles and response to external magnetic field. Biomaterials 26,

2061–2072 (2005).

140. Sun, S. & Zeng, H. Size-Controlled Synthesis of Magnetite Nanoparticles.

Journal of the American Chemical Society 124, 8204–8205 (2002).

141. Towe, K. Structural distinction between ferritin and iron-dextran (imferon).

An electron diffraction comparison. Journal of Biological Chemistry 256,

9377–9378 (1981).

142. Quintana, C. et al. Initial studies with high resolution TEM and electron

energy loss spectroscopy studies of ferritin cores extracted from brains of

patients with progressive supranuclear palsy and Alzheimer disease.

Cellular and Molecular Biology 46, 807–820 (2000).

107

143. Hartmann, U. Magnetic force microscopy. Annual Review of Materials

Science 29, 53–87 (1999).

144. Albrecht, T. R., Grütter, P., Horne, D. & Rugar, D. Frequency modulation

detection using high-Q cantilevers for enhanced force microscope

sensitivity. Journal of Applied Physics 69, 668–673 (1991).

145. Holscher, H. & Schwarz, U. Theory of amplitude modulation atomic force

microscopy with and without Q-Control. International Journal of Non-Linear

Mechanics 42, 608–625 (2007).

146. Agarwal, G. Characterization of Magnetic Nanomaterials Using Magnetic

Force Microscopy. Magnetic Nanomaterials for Life Sciences (2009).

147. McVitie, S., Ferrier, R. P., Scott, J., White, G. S. & Gallagher, A.

Quantitative field measurements from magnetic force microscope tips and

comparison with point and extended charge models. Journal of Applied

Physics 89, 3656 (2001).

148. Gomez, R. D. et al. Quantification of magnetic force microscopy images

using combined electrostatic and magnetostatic imaging. Journal of

Applied Physics 83, 6226 (1998).

149. Hartmann, U. The point dipole approximation in magnetic force

microscopy. Physics Letters A 137, 475–478 (1989).

108

150. Van Schendel, P. J. A., Hug, H. J., Stiefel, B., Martin, S. & Güntherodt, H.-

J. A method for the calibration of magnetic force microscopy tips. Journal

of Applied Physics 88, 435 (2000).

151. Sievers, S. et al. Quantitative measurement of the magnetic moment of

individual magnetic nanoparticles by magnetic force microscopy. Small 8,

2675–2679 (2012).

152. Hsieh, C. W., Zheng, B. & Hsieh, S. Ferritin protein imaging and detection

by magnetic force microscopy. Chemical Communications 46, 1655–1657

(2010).

153. Neves, C. S. et al. New insights into the use of magnetic force microscopy

to discriminate between magnetic and nonmagnetic nanoparticles.

Nanotechnology 21, 305706 (2010).

154. Torre, B. et al. Magnetic force microscopy and energy loss imaging of

superparamagnetic iron oxide nanoparticles. Scientific Reports 1, Epub

(2011).

155. Schreiber, S. et al. Magnetic force microscopy of superparamagnetic

nanoparticles. Small 4, 270–278 (2008).

156. Puntes, V. F., Gorostiza, P., Aruguete, D. M., Bastus, N. G. & Alivisatos, A.

P. Collective behaviour in two-dimensional cobalt nanoparticle assemblies

109

observed by magnetic force microscopy. Nature Materials 3, 263–268

(2004).

157. Sun, S. Monodisperse FePt Nanoparticles and Ferromagnetic FePt

Nanocrystal Superlattices. Science 287, 1989–1992 (2000).

158. Vallet, C. E. et al. Magnetic force microscopy of ferromagnetic

nanoparticles formed in Al[sub 2]O[sub 3] and SiO[sub 2] by ion

implantation. Journal of Applied Physics 92, 6200 (2002).

159. Hehn, M., Padovani, S., Ounadjela, K. & Bucher, J. Nanoscale magnetic

domain structures in epitaxial cobalt films. Physical Review B 54, 3428–

3433 (1996).

160. Hehn, M. et al. Nanoscale Magnetic Domains in Mesoscopic Magnets.

Science 272, 1782–1785 (1996).

161. Shinjo, T., Okuno, T., Hassdorf, R., Shigeto, K. & Ono, T. Magnetic Vortex

Core Observation in Circular Dots of Permalloy. Science 289, 930–932

(2000).

162. Chou, S. Y., Wei, M. S., Krauss, P. R. & Fischer, P. B. Single-domain

magnetic pillar array of 35 nm diameter and 65 Gbits/in.2 density for

ultrahigh density quantum magnetic storage. Journal of Applied Physics

76, 6673 (1994).

110

163. Arakaki, A. et al. Detection of biomolecular interaction between biotin and

streptavidin on a self-assembled monolayer using magnetic nanoparticles.

Biotechnology and Bioengineering 88, 543–546 (2004).

164. Proksch, R. B. et al. Magnetic Force Microscopy of the Submicron

Magnetic Assembly in a Magnetotactic Bacterium. Applied Physics Letters

66, 2582–2584 (1995).

165. Suzuki, H. et al. High-Resolution Magnetic Force Microscope Images of a

Magnetic Particle Chain Extracted from Magnetic Bacteria AMB-1.

Japanese Journal of Applied Physics 37, L1343–L1345 (1998).

166. Shen, H. et al. Magnetic force microscopy analysis of apoptosis of HL-60

cells induced by complex of antisense oligonucleotides and magnetic

nanoparticles. Biophysical Chemistry 122, 1–4 (2006).

167. Zhang, Y., Yang, M., Ozkan, M. & Ozkan, C. S. Magnetic force microscopy

of iron oxide nanoparticles and their cellular uptake. Biotechnology

progress 25, 923–8 (2003).

168. Moskalenko, A. V, Yarova, P. L., Gordeev, S. N. & Smirnov, S. V Single

protein molecule mapping with magnetic atomic force microscopy.

Biophysical Journal 98, 478–487 (2010).

111

169. Hyeon, T. et al. Ultra-large-scale syntheses of monodisperse nanocrystals.

Nature Materials 3, 891–895 (2004).

170. Schreiber, S. et al. Magnetic force microscopy of superparamagnetic

nanoparticles. Small 4, 270–278 (2008).

171. Goya, G. F., Berquo, T. S., Fonseca, F. C. & Morales, M. P. Static and

dynamic magnetic properties of spherical magnetite nanoparticles. Journal

of Applied Physics 94, 3520–3528 (2003).

172. Mott, D. et al. Size Determination of Nanoparticles Based on Tapping-

Mode Atomic Force Microscopy Measurements. Journal of Scanning Probe

Microscopy 3, 1–8 (2008).

173. Wang, Y. & Chen, X. Carbon nanotubes: a promising standard for

quantitatve evaluation of AFM tip apex geometry. Ultramicroscopy 107,

293–298 (2007).

174. Zenhausern, F., Adrian, M., Heggeler-Bordied, B. T., Eng, L. M. &

Descouts, P. DNA and RNA polymerase/DNA complex imaged by

scanning force microscopy: Influence of molecular-scale friction. Scanning

14, 212–217 (1992).

112

175. Cebollada, a. et al. Growth and magnetic characterization of Co

nanoparticles obtained by femtosecond pulsed laser deposition. Physical

Review B 79, 1–13 (2009).

176. Xu, D. G., Watt, G. D., Harb, J. N. & Davis, R. C. Electrical conductivity of

ferritin proteins by conductive AFM. Nano Letters 5, 571–577 (2005).

177. Ho, R.-H., Chen, Y.-H. & Wang, C. M. Surface differentiation of ferritin and

apoferritin with atomic force microscopic techniques. Colloids and

Surfaces. B, Biointerfaces 94, 231–235 (2012).

178. Tian, L., Cao, C., Liu, Q. & Pan, Y. Low-temperature magnetic properties of

horse spleen ferritin. Chinese Science Bulletin 55, 3174–3180 (2010).

179. Boča, R. et al. Deposits of iron oxides in the human spleen. Polyhedron 66,

65–69 (2013).

180. Almad, A. & McTigue, D. M. Chronic expression of PPAR-delta by

oligodendrocyte lineage cells in the injured rat spinal cord. The Journal of

Comparative Neurology 518, 785–799 (2010).

181. Almad, A., Lash, A. T., Wei, P., Lovett-Racke, A. E. & McTigue, D. M. The

PPAR alpha agonist gemfibrozil is an ineffective treatment for spinal cord

injured mice. Experimental Neurology 232, 309–317 (2011).

113

182. Nocera, T. M., Sauerbeck, A., Zeng, Y., McTigue, D. M. & Agarwal, G.

Quantification and Spatial Localization of Ferritin Using Magnetic Force

Microscopy. Revised and resubmitted to Nanomedicine: NMB (2013).

183. Gomme, P. T., McCann, K. B. & Bertolini, J. Transferrin: structure, function

and potential therapeutic actions. Drug Discovery Today 10, 267–73

(2005).

184. Brissot, P., Ropert, M., Le Lan, C. & Loréal, O. Non-transferrin bound iron:

a key role in iron overload and iron toxicity. Biochimica et Biophysica Acta

1820, 403–410 (2012).

185. Mirowski, E., Moreland, J., Zhang, A., Russek, S. E. & Donahue, M. J.

Manipulation and sorting of magnetic particles by a magnetic force

microscope on a microfluidic magnetic trap platform. Applied Physics

Letters 86, 243901 (2005).

186. Ramachandra, R., Demers, H. & De Jonge, N. Atomic-resolution scanning

transmission electron microscopy through 50-nm-thick silicon nitride

membranes. Applied Physics Letters 98, 93109 (2011).

187. Koh, A. L., McComb, D. W., Maier, S. A., Low, H. Y. & Yang, J. K. W. Sub-

10 nm patterning of gold nanostructures on silicon-nitride membranes for

plasmon mapping with electron energy-loss spectroscopy. Journal of

114

Vacuum Science & Technology B: Microelectronics and Nanometer

Structures 28, C6O45 (2010).

188. Carter, E. A. et al. Silicon nitride as a versatile growth substrate for

microspectroscopic imaging and mapping of individual cells. Molecular

bioSystems 6, 1316–22 (2010).

189. Mirsaidov, U. M., Zheng, H., Casana, Y. & Matsudaira, P. Imaging Protein

Structure in Water at 2.7 nm Resolution by Transmission Electron

Microscopy. Biophysical Journal 102, L15–L17 (2012).

190. Klein, K. L., Anderson, I. M. & De Jonge, N. Transmission electron

microscopy with a liquid flow cell. Journal of Microscopy 242, 117–123

(2011).

191. De Jonge, N., Peckys, D. B., Kremers, G. J. & Piston, D. W. Electron

microscopy of whole cells in liquid with nanometer resolution. Proceedings

of the National Academy of Sciences of the United States of America 106,

2159–2164 (2009).

192. Nishiyama, H. et al. Atmospheric scanning electron microscope observes

cells and tissues in open medium through silicon nitride film. Journal of

Structural Biology 169, 438–449 (2010).

115

APPENDIX A

Reprint Permissions

Figure 1.2 is reprinted with permission from Bronen et. al (1990):

Permission to reproduce the requested material is granted to Tanya Nocera for her thesis without charge by the copyright owner, AANS, provided that full acknowledgment is given to Journal of Neurosurgery.

Figure 1.5 is adapted, reprinted with permission from Thomas et al. (2013):

116

Figure 3.1, 3.2, 3.3, 3.4, 3.5 and 3.6a are reprinted with permission from Nocera

et al. (2012):

Dear Tanya,

Thank you for your email and for taking the time to seek this permission. When you assigned the copyright in your article to IOP, we granted back to you certain rights, including the right to include the article within any thesis or dissertation. Therefore, please go ahead and make what use you wish of the article. The only restriction is that if, at a later date, your thesis was to be published commercially, further permission would be required.

Please let me know if you have any further questions. In the meantime, I wish you the best of luck with the completion of your dissertation.

Kind regards,

Laura

Laura Sharples

Rights & Permissions Officer

IOP Publishing Ltd

Temple Circus, Temple Way, Bristol BS1 6HG

Figures 4.1, 4.2, 4.3, 4.4 and 4.5 are part of a manuscript that has been

revised and resubmitted to Nanomedicine: NBM (Nocera et al.).

117