magnetizable implants for targeted drug delivery a thesis
TRANSCRIPT
Magnetizable Implants
for Targeted Drug Delivery
A Thesis
Submitted to the Faculty
of
Drexel University
by
Zachary Graham Forbes
in partial fulfillment of the
requirements for the degree
of
Doctor of Philosophy
May 2005
© Copyright 2005 Zachary Graham Forbes. All Rights Reserved.
ii
DEDICATIONS
This thesis is dedicated to my parents, David and Rosemary. They are truly my best friends. There has never been a moment in my life where I questioned their endless support of my studies and goals. The success of this work has brought them great joy, and anything I can do to make them proud makes it all worth while. My father is a model example of how modern medicine impacts lives. His determination and toughness have made me realize that if this invention can succeed, then I must make it succeed. Someone else’s dad may need it.
iii
ACKNOWLEDGEMENTS
My deepest gratitude must first be extended to my thesis advisor, Ken Barbee. He brought me to Drexel University into a challenging project studying angioplasty, and when my ideas that sprouted from that project evolved, and my new associations with Ben and Gary blossomed, he gave me the freedom and confidence to run with it. Ken has always made himself available to help me with my cell culture models, my experimental design, and to teach me how to get the most out of my graduate studies. He is a bioengineer in the truest form, with a vault of knowledge and skills I could only dream of possessing.
Ben Yellen is one of the finest people I’ve ever met in my life. What started out as an idea tossed around a chess board, has grown into publications, business plans, awards, magazine articles, and finally my Ph.D. thesis. Ben introduced me to his advisor, Gary Friedman, who took me in and welcomed me as his own. This technology stems from Gary’s ideas and expertise, and I am absolutely grateful that he has entrusted me with the responsibility of being the first to develop it experimentally. I must also thank the surgical all-star, Ari Brooks, for loving the work and taking a chance on me, as well as my other distinguished committee members, Dov Jaron and Fred Allen, for their support and input to this project.
I have been fortunate enough to work with brilliant students from both Ken and Gary’s labs. I first offer my thanks to Derek for developing the electroplating recipe and setup, putting in hours of work for the project, teaching me how to use the magnetometer, and laughing at my jokes. Thanks to Ondrej for putting together our magnetic field setup. Thanks to Greg for his hours of work on machining the flow chambers and keeping the electroplating setup in full function. Thanks to Devrim, Dihui, and Soonjin for all of the favors they’ve done for me, especially at the expensive of their own time and convenience. Thanks to Bahar for helping me prepare my samples at the last minute. I am also greatly in debt to Amir, Eric, and Paul, who were always there to advise my cell culture studies, and to get me a tube of whatever I needed on that given day.
I must finally acknowledge Banu Onaral, and the entire body of students, staff, and faculty of the School of Biomedical Engineering, Sciences, and Health Systems. Working in such a positive environment has made my life that much easier. I am proud to receive my highest degree from such a wonderful institution.
iv
TABLE OF CONTENTS
LIST OF TABLES.......................................................................................................... ix LIST OF FIGURES ..........................................................................................................x ABSTRACT.................................................................................................................. xvi CHAPTER 1: INTRODUCTION.....................................................................................1
1.1 THE DEMAND FOR LOCAL DRUG DELIVERY......................................1 1.2 STENTS AND CORONARY RESTENOSIS ................................................2
1.2.1 Stents................................................................................................2 1.2.2 Coronary Restenosis ........................................................................3 1.2.3 Current Treatment Options ..............................................................4
1.3 PREVIOUS METHODS OF MAGNETICALLY TARGETED DRUG DELIVERY.....................................................................................................5 1.4 PROPOSED DESIGNS FOR MAGNETIC TARGETING ...........................7
1.4.1 Embedded Particles in Vessel Walls................................................7
1.4.2 Magnetic Implants for Local Drug Delivery ...................................8
1.5 UNIQUENESS OF APPROACH...................................................................9
1.6 OBJECTIVES...............................................................................................11
1.7 THESIS ORGANIZATION..........................................................................12 CHAPTER 2: THEORETICAL MODELS OF MAGNETIC DRUG DELIVERY USING UNIFORM MAGNETIC FIELDS ..........................................................14
2.1 MAGNETIC CAPTURE BY PARTICLE SEEDING .................................14
2.1.1 Assumptions and Conditions .........................................................15
2.1.2 Governing Equations .....................................................................17
v
2.1.3 Results............................................................................................20
2.2 MAGNETIC CAPTURE BY MAGNETIZABLE IMPLANTS ..................21
2.2.1 Assumptions and Governing Equations.........................................21 2.2.2 Results............................................................................................24
2.3 DISCUSSION...............................................................................................25 CHAPTER 3: MATERIAL SELECTION, CHAMBER DESIGN, AND FLOW CHARACTERIZATION OF MAGNETIC PARTICLE CAPTURE .....36
3.1 INTRODUCTION ........................................................................................36
3.2 ELECTROPLATING OF SOFT MAGNETIC MATERIAL ONTO MODEL IMPLANTS...................................................................................................37
3.2.1 Model Implant Selection................................................................38
3.2.2 Electroplating Procedure................................................................39
3.2.3 Measurement of Magnetic Properties of Model Implants .............40
3.3 MATERIALS AND METHODS..................................................................41
3.3.1 Flow Chamber Design ...................................................................41
3.3.2 Material Selection ..........................................................................43
3.3.3 Magnetic Particle Selection ...........................................................44
3.3.4 Setup for Magnetic Capture in Parallel Plate Flow .......................45
3.3.5 First Magnetic Capture to 316L Stainless Steel Mesh...................47
3.3.6 Magnetic Capture of 2µm and 350nm Magnetic Particles ............47
3.3.7 Magnetic Capture across a Silicone Barrier...................................48
3.3.8 Magnetic Capture Accumulation Over Time Steps .......................48
3.3.9 Magnetic Capture in Porcine Blood...............................................49
3.3.10 Magnetic Capture of Robert Levy’s PLA Magnetic Particles .....50
vi
3.3.11 Magnetic Capture to 304 Stainless Steel Molded Mesh ..............50
3.3.12 Data Analysis ...............................................................................51
3.4 RESULTS .....................................................................................................51
3.4.1 First Magnetic Capture to 315L Stainless Steel Mesh...................51
3.4.2 Magnetic Capture of 2µm and 350nm Magnetic Particles ............53 3.4.3 Magnetic Capture Across a Silicone Barrier .................................53 3.4.4 Magnetic Capture Accumulation Over Time Steps .......................54 3.4.5 Magnetic Capture in Porcine Blood...............................................54 3.4.6 Magnetic Capture of Robert Levy’s PLA Magnetic Particles .......55 3.4.7 Magnetic Capture to 304 Stainless Steel Molded Mesh ................55
3.5 DISCUSSION...............................................................................................56
CHAPTER 4: PIPE FLOW CHARACTERIZATION AND QUANTITATIVE ANALYSIS OF MAGNETIC PARTICLE CAPTURE .........................75
4.1 INTRODUCTION ........................................................................................75 4.2 MATERIALS AND METHODS..................................................................76
4.2.1 Material Selection ..........................................................................76
4.2.2 Particle Selection ...........................................................................77
4.2.3 Pipe Flow Setup .............................................................................77
4.2.4 Molded 304 Stent Pipe Flow Experiments ....................................78
4.2.5 Compression Spring Pipe Flow Experiments ................................79
4.2.6 Alternating Gradient Magnetometer Analysis of Flow Capture Efficiency.......................................................................................80
4.3 RESULTS .....................................................................................................81
4.3.1 Molded 304 Stent Pipe Flow Experiments ....................................81
vii
4.3.2 Compression Spring Pipe Flow Experiments ................................81 4.4 DISCUSSION...............................................................................................82
CHAPTER 5: CELL CULTURE BIOCOMPATABILITY ANALYSIS OF MAGNETIZABLE IMPLANTS AND SUPERPARAMAGNETIC NANOPARTICLES.................................................................................95
5.1 INTRODUCTION ........................................................................................95
5.2 MATERIALS AND METHODS..................................................................96
5.2.1 Model Implant Selection and Preparation......................................96
5.2.2 Magnetic Particle Selection ...........................................................98
5.2.3 Culture of Endothelial Cells...........................................................98
5.2.4 Culture of BAECs on Magnetic Mesh ...........................................99
5.2.5 Internalization of Magnetic Particles into BAECs.......................100
5.2.6 Confocal Imaging and Magnetic Characterization of Magnetic BAECs ........................................................................................103
5.2.7 Shear Stimulation of Magnetic BAECs vs. Non-Magnetic BAECs .........................................................................................104
5.2.8 Static and Flow Capture of Magnetic BAECs by Magnetic Mesh.............................................................................................106
5.3 RESULTS ...................................................................................................108
5.3.1 Culture of BAECs on Magnetic Mesh .........................................108
5.3.2 Internalization of Magnetic Particles into BAECs.......................108
5.3.3 Confocal Imaging and Magnetic Characterization of Magnetic BAECs .........................................................................................109
5.3.4 Shear Stimulation of Magnetic BAECs vs. Non-Magnetic BAECs ........................................................................................ 110
5.3.5 Static and Flow Capture of Magnetic BAECs by Magnetic Mesh.............................................................................................111
viii
5.4 DISCUSSION.............................................................................................112
CHAPTER 6: CONCLUSIONS AND FUTURE WORK............................................127
6.1 SUMMARY OF PRINCIPLE FINDINGS.................................................127
6.2 ADVANTAGES AND RISKS ...................................................................131
6.3 FUTURE WORK........................................................................................134 LIST OF REFERENCES..............................................................................................136 VITA.............................................................................................................................141
ix
LIST OF TABLES
3.1 Mass-normalized saturation magnetization data for the three selected model
implant materials, in both plated and unplated forms.........................................62 4.1 Capture results for 5mm diameter 304 grade steel molded stent flow experiments
with 25mL doses at 1.0% by volume concentration of 350nm magnetic particles. Results are given in percentage of capture of the total particle dose, and the approximate number of captured particles. Alternating Gradient Magnetometer error is ±2.0%......................................................................................................87
4.2 This table presents experimental results for 302 stainless steel compression
springs that are unplated, and at CoNi plating heights of 2.5 and 5.5 µm, for dose concentrations of 0.5, 1.0, and 2.0% 350nm magnetic particles by volume in deionized water. Results are given in percentage of capture of the total particle dose, and the approximate number of captured particles. Alternating Gradient Magnetometer error is ±2.0%. ............................................................................90
x
LIST OF FIGURES
1.1 Conceptual image of a magnetically patterned stent capture injected magnetic
particle drug carries under the application of an external magnetic field...........13 2.1 An illustration of the trajectory of a single particle as it approaches an anchor is
shown (left); and the equilibrium shape of a chain of superparamagnetic beads aligned across the channel in varying flow conditions is shown (right).............27
2.2 (a) Trajectories of 1, 2.5, and 5µm diameter particles starting at 10%, 20%, and
30% of the channel height away from the wall that are attracted to a single anchor embedded in the wall. The maximum fluid velocity in the 50µm channel was taken to be 1mm/s velocity. (b) Trajectories of 1, 2.5, and 5µm diameter particles starting at 10%, 20%, and 30% of the channel height away from the wall that are attracted to a single anchor embedded in the wall. The maximum fluid velocity in the 50µm channel was taken to be 5mm/s velocity............................................28
2.3 (a) Trajectories of 1, 2.5, and 5µm diameter particles starting at 10%, 20%, and
30% of the channel height away from the wall that are attracted to multiple anchors embedded in the wall. The maximum fluid velocity in the 50µm channel was taken to be 1mm/s velocity. (b) Trajectories of 1, 2.5, and 5µm diameter particles starting at 10%, 20%, and 30% of the channel height away from the wall that are attracted to multiple anchors embedded in the wall. The maximum fluid velocity in the 50µm channel was taken to be 5mm/s velocity..........................29
2.4 The illustration provided in this figure depicts a simplified version of a magnetic
mesh, consisting of a 3 by 3 array of magnetic wires. The spots labeled 1, 2, and 3, indicate the starting positions of where the magnetic force profiles were taken in theoretical simulations ....................................................................................30
2.5 Major hysteresis loop and a reversal curve measured starting from the upper
portion of the major loop for the CoNi plated wire. Measurement was performed with a MicroMag Alternating Gradient Force Magnetometer (Princeton Measurements, NJ) .............................................................................................31
2.6 The graphical relationships depicted in this figure represent the force profiles at various applied magnetic fields on (a) 2µm beads, and (b) 370nm diameter beads. The series of lines indicate the force profiles taken at applied magnetic field strengths ranging from 0 to 300 G, in multiples of 50 G, on the beads starting at the surface of the intersection of two wires and moving vertically away from the mesh. The darkened line indicates the force profile taken for beads under an applied field of 150 G .........................................................................................32
xi
2.7 The graphical relationships depicted in this figure represent the force profiles at various applied magnetic fields on (a) 2µm beads, and (b) 370nm diameter beads. The series of lines indicate the force profiles taken at applied magnetic field strengths ranging from 0 to 300 G, in multiples of 50 G, on the beads starting at the surface of a single wire and moving vertically away from the mesh. The darkened line indicates the force profile taken for beads under an applied field of 150 G...................................................................................................................33
2.8 The graphical relationships depicted in this figure represent the force profiles at
various applied magnetic fields on (a) 2µm beads, and (b) 370nm diameter beads. The series of lines indicate the force profiles taken at applied magnetic field strengths ranging from 0 to 300 G, in multiples of 50 G, on the beads starting at a position in the space between the wires and moving vertically away from the mesh. The darkened line indicates the force profile taken for beads under an applied field of 150 G .........................................................................................34
2.9 This SEM image depicts the stainless steel mesh electroplated with Cobalt-Nickel
alloy that was the sole source of magnetic field gradients in experiments.........35 3.1 Diagram of electroplating setup for deposition of soft magnetic material atop of
stent-simulating materials ...................................................................................59 3.2 (a) SEM image of 316L Stainless Steel mesh. (b) Image of 304 Stainless Steel
Mesh. (c) 304 Steel mesh as seen in (b), heat molded and silver soldered into a stent-like tube (d) 302 Stainless Steel compression spring.................................60
3.3 (a) A front view of the Alternating Gradient Magnetometer. (b) A close view of
the mounted perpendicular probe, with the sample placed between the two magnets. (c) The perpendicular probe, in place on the sample mounting
device ..................................................................................................................61 3.4 (a) Overhead view diagram of the parallel plate flow chamber with magnetic
mesh fixed to the channel floor. (b) A side view of the chamber, cut down the middle diagram ...................................................................................................63
3.5 (a) Up close view of an actual machined and connected chamber. (b) The
complete experimental setup with pump, chamber, and catch beaker. The setup is aligned over a magnetization coil, and under a Leica microscope for imaging of captured particles ................................................................................................64
3.6 Conceptual sketch of the parallel plate flow chamber mounted above a magnetic
coil, with the flow delivery of magnetic particles...............................................65 3.7 Fluorescent image of 2 micron fluorescent magnetic particles captured at a 2%
concentration from stock, in a 60mL dose on a magnetically electroplated 316L Stainless Steel Mesh. ..........................................................................................66
xii
3.8 Scanning Electron Microscope images. (a) Captured 2µm magnetic particles captured on the struts of an electroplated 316L Stainless steel mesh, where particles align along the border of the plating discontinuity. (b) Chains of captured 2µm magnetic particles. (c) Shows well spread and densely captured 2µm magnetic particles all along the plated mesh surface..................................67
3.9 Presented here are results from 2 separate sets of experiments. 2µm and 350nm
diameter magnetic particles were captured onto the electroplated 316L Stainless Steel mesh when exposed to magnetic field under 15cm/s flow velocity at 1% concentration in DI water. Minimal capture was seen on unplated mesh where a 500 Gauss field was still applied ........................................................................68
3.10 Separate experiments on separate pieces of mesh, at the same concentration, show
an increase in capture of 350nm particles with the increase in dose volume from 10, 25, 50, up to 100mL doses at 1% concentration of magnetic particles. All experiments conducted at 15cm/s velocity in DI water ......................................69
3.11 Separate flow experiments where a silicone layer of thickness 100, 200, or 300µm
was cured atop of an electroplated 316L Stainless Steel mesh. experiments conducted at 15cm/s velocity, at a 1% concentration of 350nm particles in DI water....................................................................................................................70
3.12 Images from capture experiments at 15cm/s flow velocity, 1% concentration of
2µm Spherotech particles in porcine blood with added citrate anti-coagulent. A 50mL dose size with 500 Gauss uniform field was followed with 50mL DI water with field on for rinsing purposes .......................................................................71
3.13 Fluorescent images of captured nanoparticles manufactured by Robert Levy of
Children’s Hospital of Philadelphia. All experiments 15cm/s velocity, 1% concentration from stock solution, in DI water ..................................................72
3.14 Captured 2µm magnetic particles on an unplated piece of 304 Stainless Steel
mesh. (a) Flow is in the upward direction of the image. (b) Flow is applied towards the right of the image ............................................................................73
3.15 Densely captured 350nm particles on a 304 Stainless Steel mesh. (a) Flow is in
the upward direction of the image. (b) Flow is applied towards the right of the image...................................................................................................................74
4.1 (a) 304 stainless steel rolled stent. The stent is 5mm in diameter by 2cm in
length. The wire diameter is 150µm. (b) 302 stainless steel compression spring. The spring is 3mm in diameter by 2cm in length. The wire diameter is
355µm .................................................................................................................85
xiii
4.2 (a) Broad view of the pipe flow experimental setup, consisting of PVC tubing with an enclosed stent or spring, mounted within a uniform field, and connected to a flow pump. (b) Close up of magnetic spring in the pipe flow system between two magnetic coils. .............................................................................................86
4.3 (a) Shows the percentage of capture of 350nm magnetic particles by a 304 grade
steel molded stent against the normalized saturation magnetization data for unplated as well as 3.1 and 5.2µm CoNi plated samples. (b) Shows the capture numerically in 109 particles, against normalized saturation magnetization. Alternating Gradient Magnetometer error is ±2.0%...........................................88
4.4 5x objective fluorescent images of captured 350 nm fluorescent magnetic particles
on 302 stainless steel springs from pipe flow experiments. Top Images: INSIDE of Spring Bottom Images: OUTSIDE of spring .............................................89
4.5 (a) Shows the percentage of capture of 350 nm magnetic particles by an unplated
302 steel compression spring for each dose concentration. (b) Shows the capture numerically in 109 particles, for each dose concentration. Alternating Gradient Magnetometer error is ±2.0% .............................................................................91
4.6 (a) Shows the percentage of capture of 350 nm magnetic particles by a 2.5µm
CoNi plated 302 steel compression spring for each dose concentration. (b) Shows the capture numerically in 109 particles, for each dose concentration. Alternating Gradient Magnetometer error is ±2.0% ..............................................................92
4.7 (a) Shows the percentage of capture of 350 nm magnetic particles by a 5.5µm
CoNi plated 302 steel compression spring for each dose concentration. (b) Shows the capture numerically in 109 particles, for each dose concentration. Alternating Gradient Magnetometer error is ±2.0% ..............................................................93
4.8 (a) Shows the percentage of capture of 350 nm magnetic particles by a 302 steel
compression spring against the normalized saturation magnetization data for unplated as well as 2.5 and 5.5µm CoNi plated samples. Data is presented for each dose concentration. (b) Shows the capture numerically in 109 particles, against normalized saturation magnetization. Data is presented for each dose concentration. Alternating Gradient Magnetometer error is ±2.0% ...................94
5.1 Phase contrast images of BAECs grown on (a) CoNi electroplated 316 Stainless
Steel mesh (b) Unplated 316L Stainless Steel mesh, each with a thin layer of PDMS, a collagen substrate, and under the application of a 500 Gauss magnetic field for 24 hours. No substantial difference in growth or morphology was detected by inspection of these cultures............................................................115
5.2 BAECs with internalized 350nm diameter nile red polystyrene magnetic particles,
24 hours after particles were seeded into confluent wells. These cells are also referred to as “MBAECS” within this thesis ....................................................116
xiv
5.3 BAECs with internalized 350nm diameter nile red polystyrene magnetic particles, 1 hour after being passed and re-plated at a 1:6 splitting ratio .........................117
5.4 BAECs with internalized 350nm diameter nile red polystyrene magnetic
particles, 72 hours after being passed and re-plated at a 1:6 splitting ratio. Cells reached confluence earlier in the 3rd day, consistent with routine culture of BAECs split at the same ratio without internalized magnetic particles............118
5.5 60X fluorescent confocal image of fixed samples of BAECs with internalized
350nm diameter nile red polystyrene magnetic particles .................................119 5.6 Fluorescent confocal image of fixed samples of BAECs with internalized 350nm
diameter nile red polystyrene magnetic particles. 3D animations of segments along the Z-axis of the system, indicate that the particles tend to accumulate in clumps of 3-5 particles, in columns along the circumference of the membrane..........................................................................................................120
5.7 Another fluorescent confocal image of fixed samples of BAECs with internalized
350nm diameter nile red polystyrene magnetic particles. 3D animations of segments along the Z-axis of the system, indicate that the particles tend to mostly accumulate in clumps of 3-5 particles, in columns along the circumference of the membrane..........................................................................................................121
5.8 (a) Shows calcium response fluorescence versus time for a control sample of
BAECs and (b) magnetic particle labeled mBAECs during exposure to fluid shear stress. The two samples had an almost identical ratio of peak response to baseline, with the mBAECs having a second peak of higher amplitude and slower
decay .................................................................................................................122 5.9 BAECs with internalized 350nm diameter nile red polystyrene magnetic particles
30 minutes after being passed and magnetically delivered to a PDMS coated magnetically plated and unplated mesh. Specific attachment to the wire regions of the magnetic mesh can be seen, with very non-specific and disperse attachment on the unplated mesh .............................................................................................123
5.10 BAECs with internalized 350nm diameter nile red polystyrene magnetic particles
6 hours after being passed and magnetically delivered to a PDMS coated magnetically plated mesh. Remnants of the specific attachment to the wire regions of the magnetic mesh can still be seen, with cells beginning to migrate and spread across the surface ..............................................................................................124
5.11 BAECs with internalized 350nm diameter nile red polystyrene magnetic particles
24 hours after being passed and magnetically delivered to a PDMS coated magnetically plated mesh. Remnants of the specific attachment to the wire regions of the magnetic mesh can still be seen, but the cells have migrated, spread, and grown to near confluence..................................................................................125
xv
5.12 Captured BAECs with internalized 350nm diameter nile red polystyrene magnetic
particles immediately after being magnetically delivered at high flow rate in a parallel plate flow chamber to a 2 x 2 cm piece of magnetically plated mesh. An entire T75 flask of cells were labeled at approximately 103 particles per cell, and delivered at 15cm/s flow velocity .....................................................................126
xvi
ABSTRACT
MAGNETIZABLE IMPLANTS FOR TARGETED DRUG DELIVERY Zachary Graham Forbes Kenneth Barbee, Ph.D.
The capability to deliver high effective dosages to specific sites in the human body has
become the holy grail of drug delivery research. Drugs with proven effectiveness under
in vitro investigation often reach a major roadblock under in vivo testing due to a lack of
an effective delivery strategy. In addition, many clinical scenarios require delivery of
agents that are therapeutic at the desired delivery point, but otherwise systemically toxic.
This project proposes a method for targeted drug delivery by applying high magnetic
field gradients within the body to an injected superparamagnetic colloidal fluid carrying a
drug, with the aid of modest uniform magnetic field. The design involves patterning of
endovascular implants, such as coronary stents, with soft magnetic coatings capable of
applying high local magnetic field gradients within the body.
Examination of the feasibility of the design has been focused around the treatment of
coronary restenosis following angioplasty. Drug-eluting stents, which have debuted in
hospitals over the past two years, have thus far reduced restenosis rates to below 10%.
Our local drug delivery system is a viable alternative or enhancement to drug-eluting
stents, offering increased clinician control of dose size, the ability to treat a site
repeatedly, and a wide array of applications for treatment of other pathologies. The
theoretical models, parallel plate and pipe flow analysis, and cell culture models
presented give insight into the use of micron and sub-micron scale magnetic particles for
site-specific delivery of pharmaceuticals and magnetically labeled cells.
1
CHAPTER 1: INTRODUCTION
1.1 THE DEMAND FOR LOCAL DRUG DELIVERY
As the pharmaceutical industry continues to develop new and effective medications, the
need to develop efficient, minimally invasive delivery techniques is paramount. A
successful targeted drug delivery system will allow clinical usage of drugs not currently
accessible to physicians, as well as a more efficient means for delivering those already
available.
Popular cancer drugs have been found to have applications in many realms of clinical
medicine (Garas, 2001; Gershlick, et al, 2002; Hehrlein, et al., 2002). The best approach
for treating tumors and other localized medical defects is to administer drugs only at the
site of complication. By delivering the drug locally, the toxicity of the drug to the rest of
the body can be reduced while maintaining the desired therapeutic benefit at the site of
interest. Many exciting drugs developed by the pharmaceutical industry have shown
remarkable success during in vitro testing and animal trials, but have yielded undesirable
results in clinical trials due to systemic toxicity of the drug to other parts of the body.
Thus, the ability to deliver large concentrations of drugs only at the site of complication
is of major importance for both the pharmaceutical industry and for clinicians.
In most cases, however, the drug delivery vehicles have not been advanced to a state
where it is possible to deliver locally high concentrations of drugs with minimally
2
invasive techniques. This is especially true when repeat dosing is required. The
magnetic drug delivery system proposed herein overcomes many of these difficulties, and
provides a method for concentrating drugs at selected sites in the body with minimal
stress on the patient. The area of focus for the first application is the prevention of
coronary restenosis by adapting the system to coronary stents.
1.2 STENTS AND CORONARY RESTENOSIS
1.2.1 Stents
Stents are commonly used in a variety of biomedical applications. For example, stents
are routinely implanted in patients to keep blood vessels open in the coronary arteries, to
keep the esophagus from closing due to strictures of cancer, to keep the ureters open for
maintenance of kidney drainage, and to keep the bile duct open in patients with
pancreatic cancer. Such stents are usually inserted percutaneously under radiological
guidance. Stents comprise a tube made of metal or polymer, in a wide range of
physiologically appropriate diameters and lengths. Currently the most commonly used
materials for stents are 316L Stainless Steel or Nickel-Titanium (Nitinol) (Regar, 2001).
General stent design varies in the number of intersections and interstrut area, the in-strut
configuration, and the metal-to-artery ratio. The two different expansion principles for
stents are balloon-expansion and self-expansion, and the design types can be categorized
into five types: ring, tubular, multi-design, coil, and mesh. Stents have been routinely
3
used over the last 15 years in percutaneous transluminal coronary angioplasty (PTCA), a
procedure for the treatment of severe, symptomatic coronary stenosis (Garas, 2001).
1.2.2 Coronary Restenosis
In-stent restenosis (the re-closing of the vessel) remains a major limitation in coronary
stenting. Restenosis is generally considered a local vascular manifestation of the
biological response to injury. The injury as a result of catheter insertion consists of
denudation of the intima (endothelium) and stretching of the media (smooth muscle). The
wound-healing reaction consists of an inflammatory phase, a granulation phase, and a
remodeling phase. The inflammation is characterized by growth factor and platelet
activation, the granulation by smooth muscle cell and fibroblast migration and
proliferation into the injured area, and the remodeling phase by proteoglycan and
collagen synthesis, replacing early fibronectin as the major component of extracellular
matrix (Garas, 2001; Gershlick, et al, 2002; Hehrlein, et al., 2002).
Coronary stents comprise mechanical scaffolding that almost completely eliminates
recoil and remodeling. However, neo-intimal growth or proliferation is still a problem.
Neo-intimal proliferation occurs principally at the site of the primary lesion within the
first 6 months after implantation, a major checkpoint for patient health post-surgery
(Regar, 2001). Neo-intima forms during the first week after PTCA and the progress is
well under way after 4 weeks, with continued progression over the following months
(Hehrlein, 2002). This neo-intima is an accumulation of smooth muscle cells within a
proteoglycan matrix that narrows the previously enlarged lumen. Its formation is
4
triggered by a series of molecular events including leukocyte infiltration, platelet
activation, smooth muscle cell expansion, extracellular matrix elaboration, and re-
endothelialization (Regar, 2001).
1.2.3 Current Treatment Options
There are several treatment options in practice or under investigation for treating
restenosis. Local drug delivery provides limited systemic release, thereby reducing the
risk of systemic toxicity. Techniques for local drug delivery to arterial tissue that have
been described include, but are not limited to, direct coating of the stent with drug,
coating of the stent with a drug-containing biodegradable polymer, and hydrogel/drug
coating. Biodegradable stents have also been described (Garas, 2001; Gershlick, et al,
2002; Hehrlein, et al., 2002; Schwartz, 2002;)
Problems with these technologies, include damage to the polymer layer during
implantation (causing portions to sever, break off and cause clot formation), the inability
to deliver effective concentrations, one-time dosage limitations, and, in the case of
biodegradable stents, mechanical compromise. An additional concern with the polymer-
coated drug-eluting stents is limitation of the endothelialization necessary to cover the
stent and prevent the bare metal from coming in long contact with the blood, thereby
leading to clot formation (Schwartz, 2002).
In 2004, the FDA has approved TAXUS ™ Express 2™ Coronary Stent (Boston
Scientific, Boston, MA) which releases the drug paclitaxel, and in early 2003 the
5
CYPHER™ Sirolimus-eluting Coronary Stent (Cordis, Miami, FL) was approved. The
U.S. Food and Drug Administration (FDA) has announced that the agency is warning
physicians about blood clots and other side effects associated with the CYPHER™ stent,
manufactured by the Cordis Corporation, a subsidiary of Johnson & Johnson. The device
has been linked to more than 60 deaths and 290 cases of thrombosis as of late 2003
(Issued by the US Food and Drug Administration at http://www.fda.gov October 29,
2003). There have also been manufacturing recalls by Boston Scientific for
complications regarding their TAXUS ™ Express 2™ stent.
1.3 PREVIOUS METHODS OF MAGNETICALLY TARGETED DRUG
DELIVERY
Magnetic targeting of therapeutic agents to specific sites in the body enjoys certain
advantages over other drug delivery methods. One advantage is that magnetic colloids
can be injected into the bloodstream and guided to the targeted area with external
magnetic fields (Flores, 2002; Gallo, et al., 1997; Lübbe, et al., 2001; Mossbach, et al.,
1979; Rudge, et al., 2001). This technique, which requires only a simple injection, is far
less invasive than surgical methods for targeted drug delivery. Another advantage is that
particles in the magnetic fluid interact strongly with each other, which facilitates the
delivery of high concentrations of drug to targeted areas. Magnetic particles composed of
magnetite are well tolerated by the human body (Vyas, 2001). Also, magnetic fields are
well suited for biological applications as they are not screened by biological fluids and do
not interfere with most biological processes (Bell, et al., 1991; Buemi, et al., 2001;
Kirschvink, et al., (2) 1992; Schenck, 2000).
6
Previous attempts to use magnetic particles in these applications have relied on high
gradient magnetic fields produced by magnets external to the body to direct magnetic
particles to specific locations (Flores, 2002; Gallo, et al., 1997; Lübbe, et al., 2001;
Mossbach, et al., 1979; Rudge, et al., 2001). This limits the range of their applications.
The main disadvantage of this approach is that externally generated magnetic fields apply
relatively small and insufficiently local forces on micron and sub-micron sized particles.
As a result, there has been great interest in devising systems that produce strong and
highly localized field gradients in the interior of the body.
In addition to their frequent use as MRI contrast agents, magnetic liposomes,
microspheres, and ferrofluids have been studied and used for: drug delivery of
chemotherapeutics for cancer therapy (Babincova, et al., 1996-2001; Goodwin, et al.,
2001; Kato, T., 1983; Ovadia, et al., 1983; Ruuge, et al., 1993; Senyei, et al., 1978;
Widder, et al., 1979), embolization of tumors (Babincova, et al., 2001; Forbes, et al.,
2003; Minamimura, et al., 2000), localized diagnostic imaging (Illum, et al., 2001;
Päuser, et al., 1997; Tiefenauer, et al., 1996), targeted hyperthermia (Hilger, et al., 1997;
Minamimura, et al., 2000; Mitsumori, et al., 1996; Yanase, et al., 1998), and for the
separation, isolation, and twisting of cells (Moore, et al., 2000; Plavins, et al., 1993;
Wang, et al., 1992).
7
1.4 PROPOSED DESIGNS FOR MAGNETIC TARGETING
1.4.1 Embedded Particles in Vessel Walls
One possible approach to this problem is to use magnetic material embedded in blood
vessel walls as the source of strong localized magnetic field gradients at defined locations
in the body. The proposed design involves seeding magnetic particles onto blood vessel
walls at designated sites through specific receptor-ligand recognition (Forbes, et al.,
2003). Although this approach also uses external magnetic fields, these fields are uniform
and the intent is to magnetize the anchored particles to produce the field gradients rather
than to use the external field gradient itself to direct the particle.
Typical techniques for drug delivery involve saturation of site receptors with an
appropriate ligand chemically attached to a desired drug. In contrast with these
techniques, our method relies on saturating receptors with inert superparamagnetic
particle anchors coated with specific ligands for the site of interest (containing no drug).
High concentrations of anchors can be applied to saturate receptors without fear of
harmful side-effects. One can also imagine scenarios in which magnetic anchors can be
implanted in blood vessel walls through catheter based insertion methods. Once the
anchors or implants are in place, uniform magnetic fields may be used to attract an
injected drug-infused superparamagnetic colloidal fluid to the anchored particles, thereby
allowing high local concentrations of otherwise systemically toxic drugs to be captured at
the site of interest.
8
1.4.2 Magnetic Implants for Local Drug Delivery
The primary focus of this thesis is an implant-based drug delivery system, which operates
by placement of a magnetic implant (a stent or cylindrical endovascular implant) at
designated sites in the cardiovascular system and then attracting to the designated sites
injected doses of magnetically susceptible drugs. This is accomplished with the aid of a
modest and uniform, external magnetic field. Magnetic microspheres or nanospheres,
which can be designed to carry virtually any type of drug or medical agent, are attracted
to regions of the strongest magnetic field gradients. Consequently, our research is
focused on designing biocompatible implants that produce strong magnetic field
gradients near the surface of the implant so that sufficient doses of drug can be captured.
To maximize the magnetic field gradients, the implant is embedded in blood vessel walls
in order to be in proximal contact with the blood flow. One key advantage to this drug
delivery system is that the implant can be inserted by minimally invasive techniques such
as catheterization, as opposed to surgical techniques used to implant other drug delivery
devices. To accommodate more serious cases where open chest or otherwise more
invasive procedures are necessary, surgical placement of magnetized implants solely for
the purpose of targeted delivery of therapeutics may also be viable. A conceptual
drawing of the implant-based magnetic drug delivery system can be found in Figure 1.1.
The therapeutic agents that can be delivered to the implant by this method include
pharmaceutical drugs, radioactive polymers, an cells. For example, in the treatment of
coronary atherosclerosis, the ability to deliver growth-inhibiting drugs to the site of stent
implantation can greatly reduce restenosis (the in-growth of arterial tissue following
9
angioplasty). In addition, numerous low molecular weight polymer sphere formulations
have been examined for their ability to deliver poorly water soluble drugs (such as
paclitaxel, in the case of restenosis) by burst release, over a range of weight distributions
for particle content (Dhanikula, et al., 1999; Liggins, et al., 2001; Singla, et al., 2002;
Wang, et al., 2002). The proposed design is perfectly in league with current technology
in polymer science.
This drug delivery system could also offer significant benefits in the treatment of hepatic,
renal, pancreatic, prostate and other cancers where the ability to provide multiple doses is
of enormous importance. It may be possible to adapt currently used devices, such as the
FDA approved coronary stent, with a pattern of magnetic material to attract magnetic
drug to its surface. This approach is convenient for use in large vessels (i.e. vessels
larger than 3 mm diameter) that are easily accessible by catheterization.
1.5 UNIQUENESS OF APPROACH
Magnetic targeting of drugs to specific sites in the body enjoys certain advantages over
other drug delivery methods. Magnetic colloids can be injected into the bloodstream and
guided to the targeted area with external magnetic fields, which is far less invasive than
surgical methods for targeted drug delivery. Magnetic particles in fluids interact strongly
with each other, which facilitates the delivery of high concentrations of drug to targeted
areas. Magnetic particles composed of magnetite are well tolerated by the human body.
10
In addition, magnetic fields are well suited for biological applications as they are not
screened by biological fluids and do not interfere with most biological processes.
In most applications, however, the magnetic field gradients used to guide the magnetic
drug are generated externally (Widder, et al., 1982). The disadvantage to this approach is
that externally generated magnetic fields apply relatively small and insufficiently local
forces on micron-sized magnetic particles. Furthermore, the production of high gradient
external magnetic fields causes the drug to be pulled towards the surface of the body and
away from the site of interest. In cases where the treatment site is deep in the interior of
the body, this technique may not be a practical solution.
For these reasons, there has been great interest in devising systems that produce strong
and highly localized field gradients in the interior of the body. The use of internal
magnetic field gradients has been employed to create embolisms and controlled
blockages at the site of vascular defects, such as aneurysms and arterio-venous
malformations (AVMs) (Garibaldi, et al., 2001). These applications, which attempt to
block the flow through a blood vessel, are solutions for occluding cardiovascular defects
rather than for drug delivery.
Other inventors are also working on magnetic implants for drug delivery (Chen, et al.,
1999). Their methods, however, rely on large electromagnets or large rare-earth
permanent magnets to magnetize the drug and provide magnetic field gradients to capture
the drug. Since the magnetic field gradients due to large implanted magnets are far
11
weaker than those produced by implanted micron-sized particles or patterned magnetic
materials, their method is not an efficient system for capturing magnetic drug. In
addition, their methods require invasive surgical techniques to implant the magnetic
device in the tissue and often require external power supplies to energize the
electromagnet.
Our invention, by contrast, accomplishes the tasks of magnetizing the drug and providing
the magnetic field gradients separately. The magnetic field gradients are produced by the
magnetizable wires of the stent or implant, and the long-range fields for magnetizing the
drug-loaded particles are applied separately. By separating the two tasks, significantly
larger magnetic field gradients can be obtained without sacrificing the ability to
magnetize the drug. In addition, our methods insert the magnetic implant in the
cardiovascular system by minimally invasive catheter-based insertion methods. By
implanting the device in the cardiovascular system, strong magnetic field gradients are
created in close proximity to the blood flow, allowing for a more significant percentage
of magnetic drug to be captured.
1.6 OBJECTIVES
We have used mathematical models, a parallel plate flow chamber, and stainless steel
compression springs in pipe flow to guide our current designs to provide a theoretical
basis for rational magnetic implant design. How we have envisioned this technology will
take form and its final realization may vary significantly, but it is believed that regardless
12
of the application, the ability to produce magnetizable implants capable of capturing
therapeutically significant numbers of drug-infused magnetic particles would yield a wide
array of uses in clinical practice. While restenosis prevention, chemotherapy, and stem
cell delivery are areas of specific interest, this project has focused mainly on magnetically
capturing the magnetic particles at the site of interest, as well as magnetic manipulation
of cells, rather than on the actual release of any specific drug into tissue.
1.7 THESIS ORGANIZATION
This dissertation is presented as four separate manuscripts with a preceding introduction
(Chapter 1, present) presenting relative information on the applications and magnetic
drug delivery, as well as a conclusion (Chapter 6) which summarizes the primary findings
and future work needed for the rational design of a safe and effective magnetic drug
delivery system. Chapter 2 provides the previously published mathematical equations
and computational models used to legitimize an approach to magnetic drug delivery
based on uniform magnetic fields. In Chapter 3, strategies and methods for material
selection, electroplating, and parallel plate flow chamber design and implantation are
provided, followed by parallel plate experimental flow methods and qualitative results.
In Chapter 4, experimental flow methods and quantitative results from pipe flow are
presented. In Chapter 5, methods and results from cell culture models examining issues
of biocompatibility, magnetic particle internalization, and magnetic cell capture are
presented. Chapter 6 concludes with further discussion, implications of the results, and
plans for animal studies.
13
Figure 1.1: Conceptual image of a magnetically patterned stent capture injected magnetic particle drug carries under the application of an external magnetic field.
Uniform
magnetic field
External
Magnet
Magnetic Drug
Carriers
Magnetically
patterned insert
(stent)
Illustration of the magnetically targeted delivery
Uniform
magnetic field
External
Magnet
Magnetic Drug
Carriers
Magnetically
patterned insert
(stent)
Illustration of the magnetically targeted delivery
14
CHAPTER 2: THEORETICAL MODELS OF MAGNETIC DRUG DELIVERY
USING UNIFORM MAGNETIC FIELDS
To determine the feasibility of the idea of successful capture of magnetic colloid by an
implant in the cardiovascular system, multiple computational models were developed in
MATHCAD (Mathsoft, MA), C, and MATLAB (Mathworks, MA) software. The focus
was on limiting-case scenarios, to determine the latitude of the technology. The first
model designed was based around capture of particles in flow onto a placed ‘anchored’
particle. The mathematics and logic used for these models were the main influence on
the idea of a magnetically electroplated stent or implant, which are examined in the
second model. The idea of high, localized magnetic field gradients is present in both
designs, but accomplished in slightly different manners. The use of anchored particles
was only considered for micro-channels, but for the use of implants, larger vessels were
modeled. All necessary adjustments have been made and noted accordingly.
2.1 MAGNETIC CAPTURE BY PARTICLE SEEDING
Theoretical investigations of the trapping process have been conducted, focusing on the
likelihood of an anchored particle trapping a single magnetic particle in flow. The
purpose of the idealized theoretical models is to analyze limiting cases and determine
only order-of-magnitude approximations. These simulations are essential to further
experimental and theoretical modeling needed to determine the potential for targeting
magnetic microspheres.
15
2.1.1 Assumptions and Conditions
To reduce the number of dimensions considered, the channel was characterized by flow
between two infinite planes, also known as plane Poiseuille flow. The drag force on
each particle, which is the only hydrodynamic force considered in this model, is assumed
to behave consistently with the Stokes drag on an isolated sphere in uniform flow.
Hydrodynamic forces due to perturbation of the flow by the isolated particles and due to
hydrodynamic interactions between particles and walls are ignored. Analysis of these
additional hydrodynamic forces (sometimes known as the “lift forces”) has been the
subject of considerable attention in the past. It has been determined both experimentally
and theoretically that particles moving faster than local shear flow will experience
additional forces toward the walls of the vessel (Friedlander, et al; Segre, et al., 1962;
Schewe, et al., 1980). This aids the process of particle capture and, therefore is not taken
into account in the limiting case analysis.
Only Newtonian liquids were considered in the model. This is due to the fact that the
experimental model of particle trapping has been carried out mainly in such liquids.
Particle movements in non-newtonian liquids, such as blood, have also been the subject
of considerable attention in the past (Segre, et al., 1962). In such liquids, additional
phenomena that move particles towards the vessel walls may assist in trapping drug-
carrying particles.
16
In situations involving receptor-bound magnetic anchors, a strong uniform magnetic field
is applied perpendicularly to the channel walls. This saturates the magnetic moments of
both anchored and flowing particles in the same direction. The magnetic force between
any two particles is determined at arbitrary particle positions and compared with the
hydrodynamic forces.
Simple calculations of the drag and magnetic forces between particles at the channel
walls indicate that very large flow rates (significantly greater than those possible in
micro-vessels) will be required to dislodge isolated particles. Thus, the ability of an
anchor to trap flowing particles will depend mostly on whether the trajectories of flowing
particles terminate at the anchor sites.
A conceptual diagram of the particle capture trajectory as it approaches the anchor is
shown in Figure 2.1. Initially, it is assumed that the particle velocity is exactly equal to
the fluid velocity. As the particle approaches the anchor, the magnetic force of the
anchor will attract the particle, causing the particle’s velocity to increase. Lift forces
acting on the particles were neglected in this study since they are expected to be small
compared to the magnetic attraction. However, particles have been found experimentally
to migrate towards an equilibrium position of approximately 0.2 tube diameters away
from the wall (Schewe, et al., 1980). Therefore, we will consider the trajectories of
particles that migrate from around this position.
17
2.1.2 Governing Equations
All superparamagnetic beads are assumed to be uniformly magnetized. This first order
approximation is reasonable when the magnetic beads are not touching. In practice,
magnetic material inside each bead is encapsulated in glass or polymer and cannot come
into direct contact with magnetic material inside other beads. In the case of a linear
relationship between magnetization and magnetic field, the bead’s magnetic moment is
given by Equation 2-1.
HVmrr
3
3
+=
χ
χ Eq. (2-1)
where χ is the magnetic material susceptibility and V is the volume of the spherical
bead. By analogy, the nonlinear constitutive relation is given by the Equation 2-2
( )H
HHVfHFm r
rrrr
3
3)(
+==
χ
χ
Eq. (2-2)
Here f is an appropriately normalized scalar nonlinear function. For the purposes of the
calculations reported below a hard saturation model was assumed (linear region followed
by constant magnetization region) to allow comparison with completely linear models.
The magnetic field for the nth bead due to external sources and other beads is given by:
( )∑
=
−
⋅
+=
N
k kn
k
kn
knknk
nextn
r
m
r
rrm
HH
135
3
4
1rrrr
rr
π
Eq. (2-3)
where nextHr
is the field created by the uniform external field at the location of the nth
particle, knrr
is the vector from particle k to n and knr is the magnitude of this position
vector, and for k ≠ n .
18
The magnetic moments of the beads are found through a relaxation based iterative
method obtained by using Equation 2-3 in conjunction with Equation 2-4:
( ))( 111 −−− −−= jjjj HFmmmrrrr
λ Eq. (2-4)
where the relaxation constant λ (in our case 1≈λ ) was selected for fastest convergence.
Having found the dipole moments of all magnetic particles, the force on the nth particle
can be determined according to Equation 2-5:
( )nnn HmF
rrr∇⋅= 0µ Eq. (2-5)
The horizontal drag force on each particle is given as a function of position across the
channel in Equation 2-6:
)(6)( rarD νπηvr
= Eq. (2-6)
where a is the radius of the bead, η is the viscosity of the liquid, and νx(z) is the flow
velocity at the bead center in Poiseuille flow with the coordinate origin at the bottom of
the channel can be represented as in Equation 2-7:
)(4
)( 2
2
max zbzb
vzx −=ν
Eq. (2-7)
where νmax is the maximum flow rate at the center of the channel, and b is the height of
the channel.
For the particle capture analysis, the particle trajectory is calculated numerically by
analyzing the tangential velocity of the particle. The particle velocity in the x-direction is
19
taken to be that of the fluid for the given vertical position (as in Equation 2-7), whereas
the balance between the z-component of the magnetic force and the resisting drag is used
to find the velocity in the z-direction. The vertical position of the particle is accumulated
through small increments in the x-direction according to Equation 2-8:
xFm
zx
z ∆=∆πηαν6
Eq. (2-8)
where Fmz is the z-component of the magnetic force calculated in Equation 2-5. The x-
component of the magnetic force was neglected in this analysis.
A series of simulations model the trajectory of a single particle flowing by magnetic
anchors partially embedded in the wall of a channel. The model numerically simulates
the particle trajectory starting at a vertical position of 10%, 20%, or 30% of the channel
height away from the wall at a significant distance upstream of the anchored particle.
These simulations were performed for 1µm, 2.5µm, and 5µm diameter particles in either
1mm/s or 5mm/s fluid velocities in a 50µm diameter channel.
In the first set of simulations, the particle trajectory was analyzed as a result of the force
due to a single magnetic anchor embedded in the wall. In the second set of simulations,
the particle trajectory was analyzed as a result of the force due to multiple anchors
periodically spaced in the wall with a two-diameter period between the anchors. Results
for each can be seen on the Figure attachments.
20
2.1.3 Results
Figures 2.2(a) and 2.2(b) show the trajectories of the beads, at 1 and 5 mm/s velocities,
respectively. In both cases, the larger 5µm particles were captured from a starting
position of 10% or 20% of the channel height away from the wall upstream of the anchor,
but no particles were captured at 30% away from the wall. The 1 µm particle was not
captured at any starting position, while the 2.5µm particle was only captured at a starting
position of 10% away from the wall. In all of these graphs, the encircled area is
considered the “target region” where the anchor(s) are located and a successful capture is
denoted when the particle reaches the target region. The results indicate a large
dependence on the bead-to-channel diameter ratio rather than a dependence on fluid
velocity.
It is interesting to observe that the particles experience both repulsive and attractive
interactions with the anchor. This phenomenon is a result of biasing the moments of all
the beads in the same direction. In these circumstances, the particles will be repelled
weakly by the anchors when far away but attracted strongly when nearby. This
qualitative behavior can be observed in all graphs.
The trajectories in Figures 2.2(a) and 2.2(b) due to a single embedded anchor are
qualitatively similar to the trajectories in Figures 2.3(a) and 2.3(b) due to multiple
embedded anchors. The presence of additional anchors increases the chance of bead
capture, as demonstrated in Figure 2.4 where the 5µm bead can now be caught at all
starting positions. However, the periodicity in the anchor placement introduces an
21
additional chaotic effect on particle trajectory that is a result of competing repulsive and
attractive interactions between the anchors as the particle passes by.
2.2 MAGNETIC CAPTURE BY MAGNETIZABLE IMPLANTS
Modeling of magnetic capture by implants focused on ascertaining the feasibility of
capturing particles with magnetic implants in the presence of other competing forces.
Experiments to be discussed in further chapters have used a wire mesh that was
electroplated with Co/Ni alloy to produce the localized magnetic field gradients. With
the aid of magnetic measurements on the electroplated mesh performed by an Alternating
Gradient Magnetometer (AGM) at Princeton Measurements Corporation (Princeton, NJ),
theoretical models have attempted to describe the system of magnetic particle capture in
order to guide current and future experiments. These models were presented by
Benjamin B. Yellen, Ph.D., at the 5th Magnetic Carriers meeting in Lyon, France, in
Summer 2004, and were published by the authors in May 2005 (Yellen, et al., 2005).
2.2.1 Assumptions and Governing Equations
In these systems, the particle’s mass is so small that gravity can be neglected, so the
main competition against the magnetic force is hydrodynamic drag. In order to simplify
models, the mesh used was approximated as a planar 3 by 3 array of magnetic wires, as
shown in Figure 2.4.
This assumption is quite reasonable since the magnetic carriers will only feel the
22
magnetic gradients of the closest few wires in the implant. The wires are assumed to be
uniformly magnetized perpendicularly to their long axes by the presence of uniform
magnetic fields produced externally. These assumptions allow for an analytical solution
for each wire’s magnetic scalar potential to be obtained (Voltairas, et al., 2002). This
potential is given by Equation 2-9:
rr
wire
wire r
rr )cos(
2),(
θ
π
λθϕ =
Eq. (2-9)
The magnetic moment per unit length of the wire is denoted by wireλ where rr
is a vector
of magnitude, r , at an angle θ with the applied magnetic field. The measured magnetic
moment per unit length of the wire was found to relate linearly with the applied field
0Hr
in the range of 0 to 300 G, as mAmHwire ⋅⋅+⋅⋅= −− 526
0 108.6)103.1(rr
λ , as can be seen
from the hysteresis curve originating from the upper hysteresis branch in Figure 2.5.
Using the magnetic scalar potential, the magnetic fields produced by the wires can be
computed by taking its gradient.
Once the magnetic fields in the system are known, it is possible to calculate the magnetic
moments of the carriers and the magnetic forces applied to them. All magnetic carriers
are considered to be superparamagnetic beads that are uniformly magnetized, which is a
reasonable approximation for non-interacting beads. In the case of a linear relationship
between magnetization and magnetic field, the bead’s magnetic moment is given by
Equation 2-10:
23
HVmrr
χ= Eq. (2-10)
where 5.2≈χ is the effective bead susceptibility as measured by AGM, and V is the
volume of the spherical bead. Using the dipole moment calculated by the applied
magnetic fields, the force on the bead can be determined again using Equation 2-5:
( )HmFmag
rrr∇⋅= 0µ
Eq. (2-5)
The horizontal drag on each particle of radius a in a fluid of viscosity η is given as a
function of the velocity νv
of the bulk fluid by:
νπηvr
aFdrag 6= Eq. (2-6)
The velocity was assumed to be 15cm/s to approximate the flow conditions in the
coronary artery.
Using these equations, the ratio of the magnetic to hydrodynamic force on the
beads was computed in order to ascertain how strongly the mesh will compete with the
drag force on the magnetic carriers. This ratio represents only a gross approximation and
several interactions are ignored. For example, magnetic interactions between adjacent
wires are ignored and the interactions between nearby beads are neglected in the
magnetic force calculations. Only the vertical magnetic field and force components were
used in simulations. In addition, the hydrodynamic drag neglects wall effects associated
with Poiseuille flow by modeling the force as a constant magnitude proportional only to
the bulk fluid flow velocity. Under these assumptions, the following magnetic to
hydrodynamic force ratio is obtained from Equations 2-5, 2-6, and 2-10:
24
HHa
Rrr
∇⋅=νη
χ
9
2 2
Eq. (2-11)
It is believed that the above simplified model and the ratio represent the worst case
scenario, while still giving a general idea of how the forces scale and if bead capture is
possible. From Equation 2-11, it is clear that the force ratio scales with the square of the
bead’s radius. Therefore, it is predicted that larger magnetic beads can be captured more
easily, though administration of smaller beads is more desirable from a clinician’s
perspective. Hence there is an optimal bead size that can be deduced from these
equations that still allows for significant capture while minimizing the size of the bead.
2.2.2 Results
Force profiles in the vertical direction were computed for 3 different starting positions
above the mesh, and the results are shown in Figures 2.6 through 2.8. Figure 2.6(a)
shows the force profile on a 2µm bead and Figure 2.6(b) on a 370nm bead as it moves in
the vertical direction away from the intersection of two magnetic wires, which is depicted
as starting position (1) in the illustration in Figure 2.4. The series of curves in each graph
represents the force profiles as a function of uniformly applied magnetic fields set at
discrete values between 0 and 300 G in multiples of 50 G. The darkened line in each of
the graphs indicates the magnetic field strength of 150 G that was most typically used in
experimental testing. Figure 2.7 shows the same analysis for the beads moving vertically
away from the surface of a single wire, which is depicted as starting position (2) in Figure
2.4, and Figure 2.8 shows the same analysis for a bead moving vertically away from the
25
space in between the wires, which is depicted as starting position (3) in Figure 2.4.
Negative numbers indicate that the net magnetic force results in attraction towards the
mesh, whereas positive numbers indicate a net repulsion away from the mesh.
2.3 DISCUSSION
The strongest implications for the experimental models presented in the following
chapters come from the results of 2.2.2. An image of the wire mesh to be used in
experiments is shown in Figure 2.9. The typical wire diameter in the mesh is roughly
140µm with a period of approximately 500µm in between adjacent wires. These mesh
parameters were used in the simulations shown graphically in Figures 2.6 through 2.8.
The viscosity of water was used in these simulations. These results led us to believe that
only the beads in the vicinity of a few wire diameters of the mesh would be caught, with
the majority of beads would escape. Even though most material in simulations would not
be captured, the local accumulation of beads on the mesh can be still quite high. Since
the number of injected beads would be considered a low dose for the human body as a
whole, this capture method is consistent with the goal of delivering high dosage of
therapeutic agents to locally targeted sites within the body without administering
systemically toxic dosages to the rest of the body.
According to these results, the magnetic beads will be most strongly attracted to the
intersection of wires shown in Figure 2.6, since these regions produce the strongest
magnetic field gradients or have larger magnetic moments due to the woven structure.
26
Results also predict that magnetic beads are only weakly attracted to, and in some cases
even repelled from, the space in between the mesh struts as shown in Figure 2.5. As
expected, simulations predict that large 2µm beads are more easily captured by the mesh
than 370nm beads, since the ratio of magnetic to hydrodynamic force should scale with
the square of the bead’s radius. Based on these simulations, significant capture of only
the 2µm beads was anticipated.
These models serve as a guide to the likelihood of significant capture of magnetic
particles at high flow rates by creating local, high magnetic field gradients. The
predictions made have guided the selection of the suitable dimensions for experimental
flow analysis; most specifically the selection of appropriate starting particle sizes,
concentrations, flow channel dimensions, as well as material selections.
27
Figure 2.1: An illustration of the trajectory of a single particle as it approaches an anchor is shown (left); and the equilibrium shape of a chain of superparamagnetic beads aligned across the channel in varying flow conditions is shown (right).
28
(a)
(b)
Figure 2.2: (a) Trajectories of 1, 2.5, and 5µm diameter particles starting at 10%, 20%, and 30% of the channel height away from the wall that are attracted to a single anchor embedded in the wall. The maximum fluid velocity in the 50µm channel was taken to be 1mm/s velocity. (b) Trajectories of 1, 2.5, and 5µm diameter particles starting at 10%, 20%, and 30% of the channel height away from the wall that are attracted to a single anchor embedded in the wall. The maximum fluid velocity in the 50µm channel was taken to be 5mm/s velocity.
29
(a)
(b).
Figure 2.3: (a) Trajectories of 1, 2.5, and 5µm diameter particles starting at 10%, 20%, and 30% of the channel height away from the wall that are attracted to multiple anchors embedded in the wall. The maximum fluid velocity in the 50µm channel was taken to be 1mm/s velocity. (b) Trajectories of 1, 2.5, and 5µm diameter particles starting at 10%, 20%, and 30% of the channel height away from the wall that are attracted to multiple anchors embedded in the wall. The maximum fluid velocity in the 50µm channel was taken to be 5mm/s velocity.
30
Figure 2.4: The illustration provided in this figure depicts a simplified version of a magnetic mesh, consisting of a 3 by 3 array of magnetic wires. The spots labeled 1, 2, and 3, indicate the starting positions of where the magnetic force profiles were taken in theoretical simulations.
31
Figure 2.5: Major hysteresis loop and a reversal curve measured starting from the upper portion of the major loop for the CoNi plated wire. Measurement was performed with a MicroMag Alternating Gradient Force Magnetometer (Princeton Measurements, NJ).
32
(a)
(b)
Figure 2.6: The graphical relationships depicted in this figure represent the force profiles at various applied magnetic fields on (a) 2µm beads, and (b) 370nm diameter beads. The series of lines indicate the force profiles taken at applied magnetic field strengths ranging from 0 to 300 G, in multiples of 50 G, on the beads starting at the surface of the intersection of two wires and moving vertically away from the mesh. The darkened line indicates the force profile taken for beads under an applied field of 150 G.
33
(a)
(b)
Figure 2.7: The graphical relationships depicted in this figure represent the force profiles at various applied magnetic fields on (a) 2µm beads, and (b) 370nm diameter beads. The series of lines indicate the force profiles taken at applied magnetic field strengths ranging from 0 to 300 G, in multiples of 50 G, on the beads starting at the surface of a single wire and moving vertically away from the mesh. The darkened line indicates the force profile taken for beads under an applied field of 150 G.
34
(a)
(b)
Figure 2.8: The graphical relationships depicted in this figure represent the force profiles at various applied magnetic fields on (a) 2µm beads, and (b) 370nm diameter beads. The series of lines indicate the force profiles taken at applied magnetic field strengths ranging from 0 to 300 G, in multiples of 50 G, on the beads starting at a position in the space between the wires and moving vertically away from the mesh. The darkened line indicates the force profile taken for beads under an applied field of 150 G.
35
Figure 2.9: This SEM image depicts the stainless steel mesh electroplated with Cobalt-Nickel alloy that was the sole source of magnetic field gradients in experiments.
36
CHAPTER 3: MATERIAL SELECTION, CHAMBER DESIGN, AND FLOW
CHARACTERIZATION OF MAGNETIC PARTICLE CAPTURE
3.1 INTRODUCTION
In all magnetic drug delivery systems, the magnet serves two purposes; first, to
magnetize the drug, and secondly to provide magnetic field gradients to capture the drug.
Although large magnets held near or implanted in the body provide strong magnetic
fields to magnetize the drug, these large magnets inherently produce weak magnetic field
gradients. On the other hand, micron-sized magnets provide very strong magnetic field
gradients, however their fields are short-range and they cannot by themselves efficiently
magnetize the drug.
Our magnetic drug delivery system is novel in that the steps of magnetizing the drug and
providing the magnetic field gradients to capture the drug are accomplished separately.
By using micron-sized magnetic implants in combination with long-range magnetic
fields, it is possible to both magnetize the drug and create strong magnetic field gradients
for more efficient capture and localization of the drug.
The primary goal of the parallel plate flow experiments is to demonstrate the capability of
capturing magnetic particles onto a stainless steel mesh plated with soft magnetic
material with the aid of an externally applied uniform magnetic field. Due to the fact that
only so many experiments and scenarios can be examined, the model aided in narrowing
37
these down to a select few, but these experiments also aided in determining parameters
for qualitative capture analysis in pipe flow.
Of numerous conducted experiments, presented here is data from experiments each at the
same flow velocity of approximately 15cm/s (adapted to channel parameters and
volumetric flow output from the flow pump). The majority of these flow studies were
conducted in deionized water, although experiments have also been performed in culture
medium (DMEM + 10% Fetal Bovine Serum), phosphate buffered saline, and porcine
blood. Appropriate control experiments were also performed in order to demonstrate that
it is specifically the combination of applied external fields and plated magnetic features
that accommodates magnetic capture of nanoparticles in high flow rate.
This chapter examines the selected stent-simulating implant materials, the procedure by
which they were electroplated, the design and construction of a unique parallel plate flow
chamber, and flow capture of micron and nano-scale magnetic particles under various
conditions.
3.2 ELECTROPLATING OF SOFT MAGNETIC MATERIAL ONTO MODEL
IMPLANTS
Electroplating is a process of depositing a coating (commonly) of silver, gold, cobalt, or
nickel on an inferior metal, by means of electricity. The procedure for the technique used
was developed with the instruction of Derek Halverson (Department of Electrical
Engineering) and Gregory Fridman (School of Biomedical Engineering) at Drexel
38
University, and involved placement of the mesh in a borate solution containing
appropriate concentrations of cobalt and nickel ions, and following standard
electroplating procedure in the literature (Lowenheim, 1978). A diagram of the
electroplating setup can be found in Figure 3.1.
3.2.1 Model Implant Selection To simulate a stent-like surface as a plane instead of a circular cross-section, the first
material selected for electroplating was a woven, 316L stainless steel wire mesh (140µm
wire diameter, 400µm apertures), which can be seen in Figure 3.2 (a). This particular
material was selected due to its large strut spacing, and extremely low saturation
magnetization. As a result, a large difference in response to applied magnetic fields, and
subsequently in capture ability, can be compared between a virtual non-magnetic 316L
mesh and a CoNi electroplated 316L mesh.
The second stent-like material selected was a molded 304 grade stainless steel (150µm
wire diameter, 450µm apertures). This material has a more level, consistent surface, as
opposed to a woven geometry. In addition, it has stronger inherent magnetic properties
than 316L steel. An image of this material can be found in Figure 3.2(b). With the aid
of the Drexel Machine shop, stent-like tubes of this 304 steel were rolled by heating the
mesh, and sealing the roll using silver solder. Tubes were rolled to a 5mm external
diameter, 2cm in length. An image of a rolled tube can be seen in Figure 3.2(c). These
molded stents will be presented in Chapter 4.
39
The final material selected were industrial application 302 stainless steel compression
springs (2cm long spring, 3mm diameter, 355µm wire diameter). These springs are more
stent-like in geometry and flexibility, but their wire diameter is 3-5 times thicker than
struts in a typical stent. These springs are highly magnetic, and provide an upper bound
for examining different alloys and their inherent abilities to capture particles under the
application of external magnetic fields. An image of a 302 steel compression spring can
be seen in Figure 3.2(d). These springs will be presented in Chapter 4.
3.2.2 Electroplating Procedure
Before each electroplating session, a fresh bath was prepared. 100 mL of bath was
prepared of the following makeup: 0.45 M NiCl2, 0.65 M CoCL2, 30 g/dm-3 H3BO3 and
a trace of Saccharin (Sigma, MO).
The bath solution was placed in a 1L glass beaker, heated to 55ºC. A thermometer was
stabilized for constant temperature measurement. An air bubbler, connected to the house
air system, was fixed at the bottom of the beaker. A Princeton Applied Research 363
Potentiostat (Princeton Measurements, NJ) was used as the current controller for
electroplating. To the anode an industrial-grade sheet of cobalt (EMI, CA) 2 x 2 inches in
size was connected, and connected to the cathode was the piece of implant-simulating
material.
In order to maintain reproducibility, it was entirely necessary to account for how much
magnetic material was deposited on a given sample. While this can be calculated loosely
40
by ion concentrations in the bath, the dimensions of the sample, and the applied current,
some of these factors may be inconsistent. Plating height was determined by measuring
the mass of the sample before and after plating, using a digital balance with a resolution
of 10-5 grams. By using an air-bubbler, keeping the sample close to the anode, and
rotating the sample 180 degrees at the half time mark of each plating session, it is most
reasonable to assume near uniform plating. So assuming uniform plating, with
knowledge of the geometry and dimensions of each sample, electroplating height was
calculated.
The method for reproducing the same height in each session, was correlating the exact
sample size, ion concentrations, and plating time. This was accomplished with ease due
to the quality of the electroplating setup, and strict monitoring of these governing factors.
3.2.3 Measurement of Magnetic Properties of Model Implants
Magnetic properties of model implants were measured using a Princeton Measurements
MicroMag Alternating Gradient Magnetometer (AGM) (Princeton Measurements, NJ).
Images of the AGM can be found in Figure 3.3. The MicroMag AGM is a highly
sensitive instrument for detecting changes in the magnetic properties of materials
(Flanders, 1988).
5mm circular punch-outs of each mesh were made using an industrial holepunch, as well
as individual coils from the 302 compression spring. In order to normalize saturation
magnetization to mass, each sample was weighed and its mass documented. A piece of
41
mesh or spring was then mounted on the end of a cantilevered rod that incorporates a
piezoelectric sensor (the perpendicular probe was selected). A dc field then magnetizes
the sample while simultaneously subjecting it to a small alternating field gradient. This
gradient exerts an alternating force on the sample, which is proportional to the magnitude
of the field gradient and to the magnetic moment of the sample. The resulting deflection
of the rod is detected by the piezoelectric element. Computer software then generates
hysteresis curves and saturation magnetization data. Five samples of each material were
measured by AGM, and the normalized results averaged to obtain Ms per gram data. The
results can be found in Table 2.1.
3.3 MATERIALS AND METHODS
3.3.1 Flow Chamber Design
After consulting the literature, a parallel plate flow chamber (PPFC) was selected as the
best system for applying a laminar flow of particles to a magnetically patterned chip or
mesh contained within the chamber (Bakker, 2003). The PPFC to be used consists of
three main layers: a 4” glass wafer with a mounted sample of magnetic mesh, a layer of
Dow Corning Brand Sylgard poly(dimethyl siloxane) (PDMS) (Dow Corning, MI) with a
cut-out channel, and a 6mm thick abrasion-resistant cast acrylic coverpiece. The cover-
piece contains two reservoirs (flow in / flow out) that connect to the pump and the waste
receptacle, respectively. In addition there is one vacuum connection, which feeds a
milled out loop around the circumference of the piece. This top piece mainly serves to
42
assure correct parallel plate flow, and to seal the top of the channel firmly to the PDMS
layer.
The PDMS layer itself is glued atop of a glass wafer using a thinly applied micro-layer of
uncured PDMS, and is held to the top coverpiece by vacuum. This method for attaching
the PDMS layer has been tested for stability, endurance, leakage, as well as its ability to
withstand sterilization procedures and has been successful in all tests. At the end of the
channel, an approximately 2 x 2 cm magnetically electroplated stainless steel mesh is also
attached by an ultra-thin layer of PDMS at the end of the actual channel space. A
description of the preparation and design of the mesh is included in section 3.2. Diagrams
of overhead and side views of the flow chamber can be seen in the Figure 3.4. A
Fisherbrand variable flow chemical pump (Fisher, IL) is used to draw in solution
containing the magnetic particles and outputs the particles into the mouth of the PPFC.
Flow chamber output is released into a waste bottle. Photographs of the chamber and
flow chamber experimental setup can be seen in Figure 3.5.
Although in the model the variable parameter is often the vessel diameter, since we are
focusing here on a flow chamber scenario the characterization will be somewhat
different. We have centered the design of our flow chamber on an average sized
coronary artery of diameter 3mm. The parameters of the channel in the flow chamber are
Height: 2r, Width: 18r, Pattern Width 14r (2r clearance on each side), and Mesh Length:
2cm, where r is the radius of the average coronary artery: 1.5mm.
43
The goal in designing the pattern dimensions was to simulate an ‘unrolled’ stent (of
diameter 3mm and length 2cm) and to continue with the limiting-case scenario analysis,
channel height was set at 2r instead of r in order to have Vmax in the center of the channel.
Since each mesh/PDMS-gasket combination is single use, numerous other configurations
are possible, providing increased functionality to the design.
In addition, the pattern is aligned towards the end of the channel, to ensure many channel
diameters worth of distance from the entering reservoir to minimize entrance effects at
the pattern itself, and to hence obtain steady parallel plate flow. It is important to note
that while the actual channel itself is 4cm in length, the full cutout length in the PDMS is
6.5cm, allowing for space at either end of the channel at which the reservoir feeds or
empties the channel.
The reservoir dimensions are the same at the entrance and exit, and are sized at
approximately 5 times the cross-sectional area of the flow channel to ensure appropriate
flow parameters and speed in the channel itself. As calculated from resting blood flow
rates in the coronary artery of 1.25mL/sec, an approximately 15.0cm/s flow velocity for
the channel was used in the experiments (which can be increased or decreased without
modification of the flow chamber itself, and adjusted for different channel heights).
3.3.2 Material Selection
As described in 3.2, a woven stainless steel mesh was selected as the primary model
implant for parallel plate flow experiments. The 304 stainless steel mesh was also tested
44
in the flow chamber. The 302 stainless steel compression springs will be presented in
pipe flow experiments in Chapter 4.
3.3.3 Magnetic Particle Selection
Three different particle types were used in experiments. Particles used in the first two
series of flow experiments were commercially available superparamagnetic polystyrene
beads (Spherotech, IL). These beads, composed of 20% γ-Fe2O3 magnetite by weight
and labeled with nile red fluorescent pigment, had a nominal diameter of 2µm with
approximately 10% variance in size. Particles come in 2mL water solutions concentrated
at 1% w/v or 2.274 x 109 particles/mL.
Another sample of beads used from Spherotech, were also 20% γ-Fe2O3 magnetite by
weight and labeled with nile red fluorescent pigment, had a nominal diameter of 350nm
with approximately 10% variance in size. Particles come in 2mL water solutions
concentrated at 1% w/v or 4.8 x 1011 particles/mL.
In another set of experiments, a proprietary method by Children’s Hospital of
Philadelphia faculty members Dr. Robert Levy, M.D. and Michael Chorny, Ph.D., was
used to prepare fluorescent-labeled (BIODPY 564/570) biodegradable polylactic acid-
based particles loaded with magnetite. This population of nanoparticles had an average
diameter of 370nm with sizes spanning the range of 364 -380 nm. These nanoparticles
were found to have similar magnetic material properties ( 5.2≈χ ) as the commercially
available beads from Spherotech.
45
In developing this flow system, with no similar experimental models in the literature to
consult, experiments began at an extreme upper limit to determine simply the possibility
of magnetic capture of particles to a model implant fixed within a flow chamber. Due to
the size of the PVC tubing (6mm inner diameter, 30 cm in length tubes at flow input and
output) and dimensions of the chamber, a large volume of fluid input was necessary in
order to maintain continuity of the system while reducing the occurrence of air bubbles.
After cautiously evaluating the system at marginally high concentrations and larger dose
volumes, and deciding that particle capture could in fact be accomplished, working
concentrations and volumes were chosen. Eventually, these optimal parameters of a
25mL solution of particles at 1% particles by volume were selected for consistency and
resource management.
3.3.4 Setup for Magnetic Capture in Parallel Plate Flow
The mesh used in all experiments was 2 x 2 cm and was fixed at the end of the channel to
minimize the impact of entrance effects. Sections 3.3.5 through 3.3.11 detail specific
parameters for each different experiments, all of which were performed around the
following specifications.
The flow chamber was placed on the microscope stand just above a power supply
controlled magnetic coil. A Tesla coil controlled by a bipolar operational power
supply/amplifier (Kepco, NY) was used to generate the external magnetic field. In each
case, a Lakeshore Model 410 Gaussmeter and Probe (Lakeshore Cryotronics, OH) was
46
used to verify that a 500 Gauss uniform magnetic field was applied, in order to saturate
the moment of the magnetic particles as well as that of the soft-magnetically plated mesh.
The chamber was connected to vacuum, flow in, and flow out to the waste beaker. Six
millimeter diameter PVC tubing was used for all flow and vacuum connections. After
vacuum sealing the chamber, a small volume of the medium was introduced to prime the
channel and flush all air bubbles out.
At this point the entire dose of particle dilution (which in all cases, was prepared
immediately prior to experiments, and vortexed for 10 seconds for uniform suspension)
was introduced in one passage through the chamber at 550ml/minute to obtain a 15cm/s
flow velocity, followed by an equal volume of the medium without a concentration of
particles. The purpose of the flushing is to prevent any part of the particle solution from
being present at rest over the magnetic mesh, assuring all capture occurs under flow, and
due to magnetic forces. It must be noted that the external field remained on until the
flushing was concluded. A diagram of a flow experiment can be founded in Figure 3.6.
Theoretical models aided in selection of best starting points for particle size and channel
dimensions. Particle counts per manufactured container were used to calculate how many
particles are being introduced into the flow chamber. From extensive literature searches,
no experiments of this kind have ever been performed, so there is no prior data available
to supplement this proposal with exact figures of what concentrations of particles are
actually functional in a device such as ours. All experiments were conducted at least
twice for demonstration of reproducibility.
47
3.3.5 First Magnetic Capture to 316L Stainless Steel Mesh
The first flow experiments performed were setup as described in 3.3.4, and used a 2 x 2
cm piece of 316L Stainless Steel mesh electroplated with approximately 5µm of CoNi
alloy. As these experiments were initial proof of concept studies, a larger concentration
of 2% by volume in a larger dose volume of 60mL of deionized water was used, at 15
cm/s flow velocity, using 2µm diameter polystyrene magnetic particles. No control
experiments were performed during these first flow model design sessions.
3.3.6 Magnetic Capture of 2µm and 350nm Magnetic Particles
After numerous experiments were performed to validate the ability to capture magnetic
particles at high flow rate in the PPFC, concentration and dose size parameters of 1%
particle concentration in 25 mL were selected for examination of capture with other
limiting factors altered. It is important to emphasize that 1% concentration refers to
volume of particle solution drawn from 2mL stock samples provided by Spherotech,
relative to the total dose volume. Thus, a 25mL dose at 1% concentration contains 250µL
of particle stock, and 24.75mL of deionized water. While the 350nm particles are
obviously smaller in size and saturation magnetization, they are more concentrated in
stock than the 2µm particles, as noted in section 3.3.3.
In these experiments, two separate sets of experiments were performed for each particle
size. In each particle size group, an experiment was conducted with a 2 x 2 cm piece of
316 Stainless Steel mesh electroplated with 5µm of CoNi alloy fixed in the channel,
followed by another experiment where the fixed mesh was unplated. A 500 Gauss
48
uniform magnetic field was applied in each case. The purpose of these experiments is to
demonstrate that it is not simply the magnetic field that is capturing the particles, but
rather the combination of the soft magnetic coating with a uniform magnetic field.
3.3.7 Magnetic Capture across a Silicone Barrier
The next set of experiments served to simulate capture to a magnetic implant in the
presence of a layer of scar tissue, from moderate to extremely thick. In order to
accomplish this, layers of various thickness of PDMS were cured on top of the 316L
Stainless Steel mesh electroplated with 5µm of CoNi alloy. Each layer was cured in a
polystyrene petry dish at a 10:1 polymer to curing agent ratio, measured by digital
calipers and then fixed inside of the flow channel. The three thicknesses were 100, 200,
and 300µm, above the 150µm thick mesh. Flow experiments were performed at 1%
concentration in 25mL doses, at 15 cm/s flow velocity, using 350 nm diameter
fluorescent polystyrene magnetic particles.
3.3.8 Magnetic Capture Accumulation Over Time Steps
These experiments served to demonstrate that for a given concentration of 1% by volume
that capture would continue to increase over longer dose time. From the first experiments
performed, it has already been seen that extremely large concentrations (>5.0% by
volume) do in fact provide much larger amounts of capture, in increasing thick layers
upon the mesh. To evaluate accumulated capture over time, four separate experiments
were conducted at the same concentration, but at increasingly larger dose volumes. An
increase in dose volume can be seen as an increase of dose length with respect to time, as
49
a larger volume requires more time for complete delivery. It must be noted that a
completely new PPFC gasket bottom and a completely fresh piece of mesh was used in
each experiment. All experiments were conducted at 15cm/s flow velocity, using 350nm
diameter fluorescent polystyrene magnetic particles, with a 500 Gauss uniform magnetic
field applied. The three dose volumes were 10, 25, and 50mL of deionized water.
3.3.9 Magnetic Capture in Porcine Blood
In these experiments, the desired goal was to demonstrate the ability to capture magnetic
particles in a PPFC, while concentrated in porcine blood. Fresh porcine blood with
citrate anti-coagulant was obtained from Dr. David Wootton (Mechanical Engineering,
Drexel University), and used on the day of exsanguination. The parameters used for
these experiments vary slightly from those described above in other PPFC flow
experiments. Each experiment was conducted with a 2 x 2 cm piece of 316 Stainless Steel
mesh electroplated with 5µm of CoNi alloy fixed in the channel. Prior to introduction of
particles to the blood, the 50mL sample of blood was warmed to 37°C in a water bath. In
each flow experiment, 2um fluorescent polystyrene magnetic particles at 1%
concentration by volume in 50mL of porcine blood were delivered at a 15cm/s flow
velocity, with a 500 Gauss uniform magnetic field applied. It should be noted that this is
twice dose volume as used in flow experiments in a water medium. The field was left on
following the experiments, while a 50mL rinse of deionized water was passed through the
chamber. This was crucial to successful imaging of the mesh. Three separate
preparations of these experiments were performed not only for reproducibility, but also
for assurance that clear, characteristic images could be obtained for analysis.
50
3.3.10 Magnetic Capture of Robert Levy’s PLA Magnetic Particles
In order to demonstrate that the technology is adaptable, and not specifically tailored to
Spherotech’s commercially available non-degradable magnetic particles, these
experiments were performed using various sizes of polylactic acid magnetic nanoparticles
produced at the Children’s Hospital of Philadelphia, in the Cardiology Research group
lead by Robert Levy, M.D., and Michael Chorny, Ph.D. The makeup of the particles is
described in section 3.3.3. Experiments were conducted as described in section 3.3.4. In
each particle size group, an experiment was conducted with a 2 x 2 cm piece of 316
Stainless Steel mesh electroplated with 5µm of CoNi alloy fixed in the channel, at a 15
cm/s flow velocity, at a 1% by volume concentration in 25mL of deionized water. A 500
Gauss uniform magnetic field was applied in each case. Particles of diameters 170, 230,
and 350nm were used.
3.3.11 Magnetic Capture to 304 Stainless Steel Molded Mesh
The next flow experiments performed utilized plated and unplated 304 stainless steel
mesh, as seen in Figure 3.2. The mesh used was 2 x 2 cm and was fixed at the end of the
channel to minimize the impact of entrance effects. Two different experiments were
performed. In the first, 25mL of 2µm magnetic particles at 1% concentration by volume
in DI water was introduced to a piece of the mesh that was unplated. In the second,
25mL of 350nm magnetic particles at 1% concentration by volume dilutions in DI water
were introduced to a chamber containing a piece of mesh electroplated with a height of
3.1µm of CoNi alloy. Each set of experiments featured the application of a 500 Gauss
uniform field as described above.
51
3.3.12 Data Analysis
After each experiment was complete, all imaging was conducted fluorescently, in some
cases with an additional 75W incandescent light bulb used for backlighting. All results
presented within Chapter 3 are based solely on qualitative analysis.
The polystyrene magnetic particles have been stained by the manufacturer rhodamin for
imaging purposes. Quantitative analysis of all flow experiments was originally to be
performed by imaging under fluorescent microscope and subsequent analysis by
inspection, but due to field of view limitations governed by the geometry of the mesh and
particle size, the counting of individual particles or measurement average intensity of
fluorescence over particular areas were not reasonable options. For the purposes of the
experiments presented here, the limitations of the woven and molded structures directed
analysis to be purely qualitative with the intent of determining the capability of capture of
magnetic particles onto the magnetically plated mesh at various particle sizes under high
flow. Quantitative analysis of pipe flow experiments can be found in Chapter 4. Imaging
and visual analysis were performed using a Leica DM LFS microscope (Leica,
Germany), with Magnafire imaging software (Meyer, TX).
3.4 RESULTS
3.4.1 First Magnetic Capture to 316L Stainless Steel Mesh
As mentioned, the first flow experiments performed were conducted at larger dose
volumes (50-100mL), with larger particle sizes (≥ 2µm in diameter), and in many cases
52
concentrations higher than 1% by volume from particle stock. The purpose of these
adjustments was to look at the latitude of the technology’s capability by considering the
upper and lowermost bounds of magnetic capture. Figure 3.7 shows one of the first
experimental images obtained, in an experiment where 2µm magnetic particles at a 2%
concentration in DI water were densely captured at high flow velocity within the PPFC.
As is clear from inspection, particle capture was so high in this experiment that the
pattern of the particles upon the struts is literally mountainous. Most certainly this would
be a risk within the human body, due to the potential for macro-scale aggregates to break
off of the device, flowing downstream and potentially blocking a vessel.
As these experiments were continually scaled down, results as seen in Figure 3.8 were
obtained. Figure 3.8 shows three separate SEM images of a 316L Stainless Steel woven
mesh that has been removed from the flow chamber after experiments conducted at a
25mL dose at 1% concentration of 2µm diameter magnetic particles. Figure 3.8(a) shows
a portion of the mesh where the electroplating of CoNi alloy was discontinuous,
demonstrating specific attachment to magnetized portions of the device. This result
foreshadows the possibility of patterning “islands” of magnetic material along a device,
with the intent of specifically orienting capture at defined locations along struts. Figure
3.8(b) shows dense, well spread capture of these same particles, and provides a closer
look at the tendency of the particles to form chains even when magnetically captured
upon the mesh. Figure 3.8(c) shows a wider view of the captured particles densely
captured and well spread around the wires. These images demonstrate that not only is it
53
possible to capture these magnetic particles, but also that capture can be limited to a 1-2
particle thickness, an important factor in scalability and risk assessment.
3.4.2 Magnetic Capture of 2µm and 350nm Magnetic Particles
Figure 3.9 shows the results of experiments comparing plated and unplated 316L mesh,
both with an applied uniform magnetic field, and for the two particle sizes. These images
demonstrate that it is the combination of the applied field, and the soft magnetic coating
that facilitates magnetic capture of particles in high flow, not simply the applied field.
Since the soft magnetic coating is extremely weakly magnetic without an applied field to
saturate the moment of the material, it is insufficient to provide any moment to flowing
particles, let alone to apply magnetic forces to capture them. But when a 500 Gauss
uniform field was applied, saturating the moment of the CoNi plated layer, as well as the
moments of the particles, significant capture was obtained. These images provide basic,
but genuine, qualitative evidence that the two source method for magnetic particle
capture to magnetic implants is functional, and unique.
3.4.3 Magnetic Capture Across a Silicone Barrier
The ability to capture particles through a thick layer of tissue is of interest to this work, as
if the claim is to be made that an endovascular magnetic implant is capable of being
repeatedly dosed with magnetic drug at separate instances, potentially including doses
years after implantation, it must be demonstrated that the method is capable of
overcoming scar tissue. While it is difficult to test actual scar tissue, using a layer of
silicone to simulate large inhibiting layers was an effective alternative. Figure 3.10
54
shows images of 3 conditions (100, 200, 300µm layers) where magnetic particles were
captured using the two source method. As expected, sparser capture is seen as the
thickness increases, but capture nonetheless. This prompts the notion, in an in vivo
scenario, of simply increasing the concentration, dose volume, or potentially applied
magnetic field strength, in order to account for the decreased amount of distance-
dependent magnetic force that can be applied when scar tissue is present.
3.4.4 Magnetic Capture Accumulation Over Time Steps
In Figure 3.11, results demonstrate that by increasing the overall dose volume at the same
concentration (and consequentially an increase in the amount of time an implant is
exposed to a combination of applied magnetic field and magnetic particles on the first
pass), an increase in the total amount of captured particles is obtained. A steady increase
in capture was seen by inspection of resulting images, from the 10mL, 25mL, 50mL, and
100mL doses. This is simply one more step in demonstrating the scalability of this
method.
3.4.5 Magnetic Capture in Porcine Blood
While not a certain indicator of in vivo success, flow experiments performed in porcine
blood have shown that magnetic particle capture is not exclusive to deionized water.
Figure 3.12 shows two images of captured 2µm Spherotech particles atop of a 316L
Stainless Steel Mesh at 10x and 5x magnifications. Due to the size of chamber and flow
system, and the nonspecific adherence of blood to the surfaces of the system, it is highly
like that many particles bound to proteins in the blood were trapped in portions of the
55
system by surface forces. These experiments, by inspection, yielded slightly lower
capture than doses of the same concentration, but half the volume, in DI water flow
experiments. Because of the coating of the surfaces of the system with the blood, this is
not a concern. Rather, it reinforces the notion that all laboratory flow experiments
performed provide crucial starting points for feasibility and dose ranging in animal
studies.
3.4.6 Magnetic Capture of Robert Levy’s PLA Magnetic Particles
As expected, the PLA magnetic particles produced by Dr. Levy’s laboratory, were able to
be captured at 15 cm/s flow velocity to a magnetized mesh. Very few particles of 170
and 230nm diameters were captured, but significant capture comparable to that seen from
Spherotech 350nm particles were seen from the 370nm particles produced by Robert
Levy’s group. The smaller particles can indeed be made to be more magnetic, by the
addition of additional nano-crystals of magnetite, which should most certainly be
examined as a potential alternative to the 350nm size particles for which the system has
been most frequently tested. Flow results from each particle size can be found in Figure
3.13.
3.4.7 Magnetic Capture to 304 Stainless Steel Molded Mesh
The molded design of the 304 Stainless Steel mesh was intended to provide a more level
surface than the woven mesh used most frequently. It also maintains a much higher
saturation magnetization without an electroplated coating, than does 316L steel. Results
56
of the two experiments showed that a piece of unplated 304 steel mesh could capture
larger magnetic particles under the application of an applied field with no plated alloy
coating (Figure 3.14), and with the addition of a magnetic alloy coating, could capture
larger numbers of particles, at a much smaller size (Figure 3.15). Also, as can be seen in
Figure 3.15, the intersections of this mesh have a crater-like, semi-circle shaped dip. This
particular portion of the surface of the mesh may not be electroplating, or particles may
simply prefer to aggregate along the top of these craters, but slim to no capture is seen
within them. Otherwise, in a given plane of view, well dispersed capture can be seen.
3.5 DISCUSSION
As demonstrated by the Figure 3.9, significant capture of the 2µm diameter beads was
achieved by the magnetic mesh, while relatively few beads were captured by the non-
magnetic mesh. These results are consistent with theoretical models describing the
physics of particle capture. Not surprisingly, the largest concentration of captured
magnetic beads was located at the intersections between the magnetic wires where the
magnetic moment is expected to be maximal. This may be due to the woven structure,
which thickens the sample at the wire intersections, and as a result is elevated at those
intersections. Whether it is indeed an increased magnetic moment, or the woven
structure itself that causes this increase in capture at these locations, remains unclear.
Since the 304 Stainless Steel mesh was flat, and not woven, and larger accumulation
tended to occur at these intersections as well, logic dictates that the thickness of the wire
at these spots, and in turn a larger continuous portion of plated alloy, would cause a
57
higher magnetic moment and more capture. But it is also clear from examining each set
of experiments that capture can still be obtained along the full length of the wires.
Segmentation into islands may allow more control over aggregation, but much will
depend on the geometry of the device in a given application.
We were pleased to find that 350nm beads were also capable of being captured by the
magnetic mesh. Models predicted that the magnetic force at the surface of the mesh near
the intersection of two wires would be approximately 1/10 of the hydrodynamic drag
force, which means that the fluid forces would overpower the magnetic attraction of these
beads to the mesh. It was mentioned previously that these models represented the lower
limit of our estimates. Poiseuille flow profiles predict that the fluid velocity near the
walls of a channel should be significantly weaker than near the center of the channel.
The fluid velocity near the mesh could be an order of magnitude weaker than in bulk
flow, which would explain why beads were captured even though simulations predicted
otherwise. Again, bead capture was occurring primarily near the intersections between
wires in the mesh, which is consistent with theoretical simulations.
In conclusion, locally targeted drug delivery using two magnetic sources was
experimentally demonstrated as a method for optimizing the delivery of magnetic carriers
in high concentration to a magnetizable implant. Theoretical models had predicted that
the externally applied magnetic fields have a pronounced effect on particle capture.
Theory suggests that for clinical applications, magnetic fields in the range of 0.1 T should
be used so that the magnetic moments of both the particles and the implant are
58
magnetized to the point of saturation. Externally applied fields greater than 0.1 T will
have little additional effect on bead capture. Experimental results have demonstrated that
capturing superparamagnetic beads of both micrometer and sub-micrometer diameter at
reasonably high concentrations is possible in flow conditions consistent with the
dimensions and flow velocity occurring in the coronary artery in the human body. The
same experiments performed with unplated 316L mesh resulted in no significant capture,
indicating that the magnetization of the implant is responsible for providing the necessary
magnetic field gradients and forces to capture the injected beads. Flow experiments
performed with unplated 304 stainless steel mesh indicate that it may be possible to
produce a suitable magnetizable implant by varying the alloy content of a stent, as a
manufacturing alternative to a deposited soft magnetic coating.
59
Figure 3.1: Diagram of electroplating setup for deposition of soft magnetic material atop of stent-simulating materials.
60°C
Current Controlled Potentiostat Power supply
Bath on Hotplate
+ _
Borate solution concentrated with cobalt and nickel ions
Stainless steel mesh sheet, rolled mesh, or spring
Air tank to bubbler
Industrial Grade Purity Cobalt Anode
60
(a) (b)
(c) (d) Figure 3.2: (a) SEM image of 316L Stainless Steel mesh. (b) Image of 304 Stainless Steel Mesh. (c) 304 Steel mesh as seen in (b), heat molded and silver soldered into a stent-like tube (d) 302 Stainless Steel compression spring
61
(a) (b)
(c) Figure 3.3: (a) A front view of the Alternating Gradient Magnetometer. (b) A close view of the mounted perpendicular probe, with the sample placed between the two magnets. (c) The perpendicular probe, in place on the sample mounting device.
62
Table 3.1: Mass-normalized saturation magnetization data for the three selected model implant materials, in both plated and unplated forms.
Saturation Magnetization for Model Implants
Material Plating Height Ms per g
316L Woven Unplated 0.16emu
316L Woven 5.1µm 12.03 emu
304 Molded Unplated 2.97 emu
304 Molded 3.1µm 11.45 emu
304 Molded 5.2µm 17.22 emu
302 Spring Unplated 116.9 emu
302 Spring 2.5µm 140.4 emu
302 Spring 5.5µm 155.4 emu
63
Side view of parallel plate flow chamber, cut right down the middle
ReservoirsFlow In Flow Out
Flow Channel cut out of PDMS layer
Magnetic Pattern on wafer
PDMS PDMS
Vacuum
Glass Wafer
(a)
Flow In
Flow Out
Magnetic Mesh
Vacuum
Channel
Reservoirs
(b) Figure 3.4: (a) An overhead view diagram of the parallel plate flow chamber with magnetic mesh fixed to the channel floor. (b) A side view of the chamber, cut down the middle diagram.
64
(a)
(b)
Figure 3.5: (a) An up close view of an actual machined and connected chamber. (b) The complete experimental setup with pump, chamber, and catch beaker. The setup is aligned over a magnetization coil, and under a Leica microscope for imaging of captured particles.
65
Figure 3.6: Conceptual sketch of the parallel plate flow chamber mounted above a magnetic coil, with the flow delivery of magnetic particles.
Applied magnetic field ~ 500 G
FLOW IN FLOW OUT
66
Figure 3.7: Fluorescent image of 2 micron fluorescent magnetic particles captured at a 2% concentration from stock, in a 60mL dose on a magnetically electroplated 316L Stainless Steel Mesh.
67
(a) (b)
(c)
Figure 3.8: Scanning Electron Microscope images. (a) Captured 2µm magnetic particles captured on the struts of an electroplated 316L Stainless steel mesh, where particles align along the border of the plating discontinuity. (b) Chains of captured 2µm magnetic particles. (c) Shows well spread and densely captured 2µm magnetic particles all along the plated mesh surface.
68
Figure 3.9: Presented here are results from 2 separate sets of experiments. 2µm and 350nm diameter magnetic particles were captured onto the electroplated 316L Stainless Steel mesh when exposed to magnetic field under 15cm/s flow velocity at 1% concentration in DI water. Minimal capture was seen on unplated mesh where a 500 Gauss field was still applied.
Plated Mesh Unplated mesh
2µm
350nm
200 µm
200 µm 200 µm
200 µm
69
Figure 3.10: Separate experiments on separate pieces of mesh, at the same concentration, show an increase in capture of 350nm particles with the increase in dose volume from 10, 25, 50, up to 100mL doses at 1% concentration of magnetic particles. All experiments conducted at 15cm/s velocity in DI water.
10mL
100mL
25mL
50mL
200 µm
200 µm 200 µm
200 µm
70
Figure 3.11: Separate flow experiments where a silicone layer of thickness 100, 200, or 300µm was cured atop of an electroplated 316L Stainless Steel mesh. experiments conducted at 15cm/s velocity, at a 1% concentration of 350nm particles in DI water.
200µm 300µm
100µm 100µm
200 µm
200 µm 200 µm
200 µm
71
10x 5x
200 µm 200 µm
10x 5x
200 µm 200 µm200 µm
Figure 3.12: Images from capture experiments at 15cm/s flow velocity, 1% concentration of 2µm Spherotech particles in porcine blood with added citrate anti-coagulent. A 50mL dose size with 500 Gauss uniform field was followed with 50mL DI water with field on for rinsing purposes
72
Figure 3.13: Fluorescent images of captured nanoparticles manufactured by Robert Levy of Children’s Hospital of Philadelphia. All experiments 15cm/s velocity, 1% concentration from stock solution, in DI water.
170nm (10x) (10x)
350nm (10x)
350nm (20x)
230nm (10x)
200 µm 200 µm
200 µm 100 µm
73
200 µm 200 µm200 µm200 µm 200 µm200 µm
(a) (b)
Figure 3.14: Captured 2µm magnetic particles on an unplated piece of 304 Stainless Steel mesh. (a) Flow is in the upward direction of the image. (b) Flow is applied towards the right of the image.
74
(a) (b) Figure 3.15: Densely captured 350nm particles on a 304 Stainless Steel mesh. (a) Flow is in the upward direction of the image. (b) Flow is applied towards the right of the image.
10x 5x
200 µm 200 µm
75
CHAPTER 4: PIPE FLOW CHARACTERIZATION AND QUANTITATIVE
ANALYSIS OF MAGNETIC PARTICLE CAPTURE
4.1 INTRODUCTION
The work presented in Chapter 3 laid the foundation for further analysis in pipe flow.
The selection of the appropriate implant models and execution of proper electroplating
technique allowed for a suitable “target” for magnetic capture. The optimization of a
reusable and multi-functional parallel plate flow chamber provided an appropriate system
for determining an ideal working range for both particle size and concentration.
Before this magnetic drug delivery system can be tested in vivo, pipe flow experiments
are needed to obtain a stronger starting point for dose ranging. These experiments
examined the ability to capture nanoparticles in a method more closely simulating arterial
blood flow. Large and average size vessels are accounted for, as are differing geometries
and materials of the implant models used.
In addition to fluorescent imaging of the implant model materials after the flow
experiments are complete, percentage capture results have been computed based on
magnetic measurements of input versus output solution. This data gives insight into the
efficiency of magnetic capture by a magnetizable implant placed within a lumen.
76
4.2 MATERIALS AND METHODS
4.2.1 Material Selection
Two separate materials were selected for pipe flow experiments: 304 stainless steel large
diameter molded stents, and 302 stainless steel compression springs. Both of these
materials have a moderately high saturation magnetization before any soft magnetic
material is plated on them, allowing analysis of magnetic alloys to capture particles with
the aid of a external, uniform magnetic field.
With the aid of the Drexel Machine shop, stent-like tubes of the 304 grade steel were
rolled by heating the mesh, and sealing the roll using silver solder. Tubes were rolled to
a 5mm external diameter, 2cm in length. An image of a rolled tube can be seen in Figure
4.1(a). All electroplated stents were plated before molding, and according to the plating
protocol described in Chapter 3. Two separate plating heights of CoNi were electroplated
onto these stents: 3.1 and 5.2 µm. The geometry and malleability of these stents are quite
different from properties seen in commercially manufactured stents. But as this
technology is being considered for numerous applications where implants may not be
used for keeping a vessel open, but rather having drug delivery as its primary function,
it’s useful to examine a model with a significantly different geometry, as well as a larger
vessel diameter.
As described in Chapter 3, the final materials selected were industrial application 302
stainless steel compression springs (2cm long spring, 3mm outer diameter, 355µm wire
77
diameter) (Lee Spring, NY). These springs are more stent-like in geometry and
flexibility, but their wire diameter is 3-5 times thicker than struts in a typical stent. These
springs are highly magnetic, and provide an upper bound for examining different alloys
and their inherent abilities to capture particles under the application of external magnetic
fields. An image of a 302 steel compression spring can be seen in Figure 4.1(b). All
electroplated springs were plated according to the protocol described in Chapter 3. Two
separate plating heights of CoNi were deposited onto these springs: 2.5 and 5.5µm.
These compression springs have a much smaller surface area than the molded stents, and
may actually be an ideal starting design for manufacturing an endovascular implant that
has the sole function of capture magnetic drug.
4.2.2 Particle Selection
As described in Chapter 3, Spherotech magnetic particles were selected for all pipe flow
experiments. The sample of beads used from Spherotech, were 20% γ-Fe2O3 magnetite by
weight and labeled with nile red fluorescent pigment, had a nominal diameter of 350nm
with approximately 10% variance in size (Spherotech, IL). Particles come in 2mL water
solutions concentrated at 1% w/v or 4.8 x 1011 particles/mL.
4.2.3 Pipe Flow Setup
Two separate views of the pipe flow experimental setup can be found in Figure 4.2. The
setup consists of a 50mL beaker that supplies the flow pump with solution via a 20mm
long polyvinyl chloride (PVC) tube. The pump passes the particle dose through a 60mm
long PVC tube, in which the stent or spring was gently inserted within. The spring is
78
mounted between two Tesla coils controlled by a bipolar operational power
supply/amplifier (Kepco, NY) was used to generate the external magnetic field. In each
case, a Lakeshore Model 410 Gaussmeter and Probe (Lakeshore Cryotronics, OH) was
used to verify that a 500 Gauss uniform magnetic field was applied, in order to saturate
the moment of the magnetic particles as well as that of the stent or spring. The output of
the tubing was captured into a 50mL centrifuge tube for capture analysis.
A variable flow peristaltic pump (Fisher, IL) was used for these flow experiments. A
much lower volumetric flow rate was required to maintain a 15cm/s coronary flow
velocity, dependent on the diameter and length of the tubing, and averaged out at
approximately 1.25ml/s. The flow is not continuous, but rather pulsatile, unlike the flow
chamber experiments which required the use of a variable flow chemical pump for its
ability to deliver fluid at rates of up to 2L/min. Flow was calibrated before every
individual experiment.
4.2.4 Molded 304 Stent Pipe Flow Experiments
For these molded stent flow experiments, each were performed 3 times for each stent
group: unplated, and for the plating heights of 3.1 and 5.2µm of CoNi. For all flow
experiments using the 5mm molded stents, an approximately 5mm inside diameter PVC
tube was selected. The tubing was perfectly sized for firm placement of the device within
the system without deformation of the material. A 25mL dose solution of magnetic
particles in deionized water at 1% by volume particle concentration was selected as the
appropriate dose size, and was prepared and vortexed for 10 seconds, immediately before
79
delivery. Before experiments began, the magnetic field was turned on, and a 25mL
priming dose of DI water was passed through the system. The particle concentration was
then introduced, followed by a 25mL rinse with DI water in the continued presence of the
applied magnetic field. The 50mL sample of flow experiment output and rinse water was
kept in a 50mL centrifuge tube, and stored at 4ºC for analysis.
4.2.5 Compression Spring Pipe Flow Experiments
For these compression spring flow experiments, each were performed 3 times for each
spring group: unplated, and for the plating heights of 2.5 and 5.5 µm of CoNi. For all
flow experiments using the 3mm springs, an approximately 3 mm in diameter
polyethylene tube was selected. The tubing was perfectly sized for firm placement of the
device within the system without deformation of the material. Two different sets of
experiments were performed for each stent group: A 10mL dose solution of magnetic
particles in deionized water at 0.5%, 1%, or 2% by volume particle concentration was
selected as the appropriate doses, and was prepared and vortexed for 10 seconds,
immediately before delivery. The intent was to examine, quantitatively, the difference in
capture when particle concentration was double in the same volumetric dose size.
Before experiments began, the magnetic field was turned on, and a 10mL priming dose of
DI water was passed through the system. The particle concentration was then introduced,
followed by a 10mL rinse with DI water in the continued presence of the applied
magnetic field. The 20mL sample of flow experiment output and rinse water was kept in
a 50mL centrifuge tube, and stored at 4ºC for analysis.
80
4.2.6 Alternating Gradient Magnetometer Analysis of Flow Capture Efficiency
Magnetic properties of model implants were measured using a Princeton Measurements
MicroMag Alternating Gradient Magnetometer (AGM) (Princeton Measurements, NJ).
Images of the AGM can be found in Figure 3.3. The MicroMag AGM is a highly
sensitive instrument for detecting changes in the magnetic properties of materials.
Flow experiment output for each experiment had been priorly collected in a centrifuge
tube and refrigerated. Before each experiment had been performed, a control sample of
the input dose was prepared to compare to flow output results. 5mm diameter glass cover
slips were labeled appropriately for experimental parameters (dose concentration, plating
height). Each tube of flow output was then vortexed for 10 seconds to re-suspend the
particles, and a 5µL sample was added to each cover slip. The slips were allowed to dry
for one hour, and then brought to the clean room for AGM analysis.
One at a time, each cover slip was then mounted on the end of a cantilevered rod that
incorporates a piezoelectric sensor (the perpendicular probe was selected). A dc field
then magnetizes the sample while simultaneously subjecting it to a small alternating field
gradient. This gradient exerts an alternating force on the sample, which is proportional to
the magnitude of the field gradient and to the magnetic moment of the sample. The
resulting deflection of the rod is detected by the piezoelectric element. Computer
software then generates hysteresis curves and saturation magnetization data. Five output
samples of each flow experiment were measured, and then compared to the magnetic
81
measurement for the input sample for each experimental concentration. This particular
AGM has a ±2.0% error.
4.3 RESULTS
4.3.1 Molded 304 Stent Pipe Flow Experiments
These flow experiments were only performed for a particle concentration of 1.0% by
volume, but demonstrated that for a large diameter implant, at physiologically significant
flow velocity, magnetic nanoparticles can be captured. Table 4.1 provides the percentage
of capture for unplated, as well as CoNi plating heights of 3.1 and 5.2µm. Results are
given in percentage of capture of the total particle dose, and the approximate number of
captured particles. Figure 4.3 shows plots of these results against the normalized
saturation magnetization data presented for each sample (as seen in Chapter 3).
4.3.2 Compression Spring Pipe Flow Experiments
These flow experiments were performed for particle concentrations of 0.5, 1.0, and 2.0%
by volume, and demonstrated that for an average coronary artery scaled implant, at
physiologically significant flow velocity, magnetic nanoparticles can be captured. Table
4.2 provides the percentage of capture for unplated, as well as CoNi plating heights of 2.5
and 5.5 µm. Results are given in percentage of capture of the total particle dose, and the
approximate number of captured particles.
82
Figure 4.4 shows characteristic fluorescent images for each plating height for
experiments performed at a 1.0% by volume concentration of 350nm magnetic particles.
Images can be seen of the outside and inside of the spring. Due to the density of capture,
by inspection, it is difficult to differentiate between the three samples. Figures 4.5-4.7
provide bar graphs for each plating height that indicate percentage capture and numerical
capture, for each dose concentration group. In Figure 4.8, these results are combined
onto plots for percentage capture and numerical capture for each dose concentration,
against normalized saturation magnetization.
4.4 DISCUSSION
As demonstrated in Chapter 3, a stent model of 304 stainless steel is capable of
magnetically capturing significant numbers of 350nm magnetic particles without the
addition of a plated layer of soft magnetic alloy. Upon addition of a 3.1µm plating layer,
the saturation magnetization of the device is increased to four times that of an unplated
sample, but roughly only a doubling of capture occurs. Further plating to a 5.5µm height,
shows just a little over 10% increase in capture. These results indicate that the capturing
abilities of the applied magnetic forces are beginning to level off at the 17 emu/g
normalized saturation magnetization for this particular material and geometry.
It is not surprising that an unplated 302 stainless steel compression spring, with its high
saturation magnetization, is able to capture significant amounts of magnetic
nanoparticles. Due to the inherently high saturation magnetization of these springs, as
83
expected, very little benefit in terms of capture was obtained by the addition of the two
different plating heights of soft magnetic alloy. The most interesting relationships to be
found in these results is the variance between the three separate dose concentrations.
Starting with 0.5%, doubling the concentration to 1.0% lead to an approximately 20%
numerical increase in total captured particles. From 1.0% to 2.0%, a doubling of
concentration again led to an approximately 20% numerical increase in total captured
particles. But as the concentration of the dose increases, increases in saturation
magnetization of the material (due to plated layers of CoNi alloy) have a more significant
effect on total capture.
Even when considering potential error due to loss of particles to the tubing, or the
limitations of the AGM measurement procedure, even the lowest dose concentration of
0.5% provided extremely dense capture. Based on the free length and wire diameter of
these springs, they contain a total surface area of 1.93cm2. Idealizing and neglecting the
points of contact with the vessel wall, and if each particle is modeled as a cube with a
350nm side, 1.57 x 109 particles would be necessary to provide a uniform, single particle
layer coating. Capture by the unplated spring at 0.5% concentration, captured more than
three times that amount. From a theoretical standpoint, the effect of concentration is
unclear. The applied magnetic force is due to the net magnetization with respect to the
surrounding fluid, so at high concentrations of particles, the magnetization of an
individual particle would equal that of the surrounding fluid, making the force
approximately zero. So while that theory would mean that net magnetization would
decrease as you increase concentration, at higher concentrations there can be more chain
84
formation along the axis of the applied field. These chains, due to their large constituent
magnetic moment, are more easily captured. The combination of these two theories
could partially explain why a doubling of dose concentration only leads to a 20% increase
in capture.
Based on the theory that superparamagnetic particles will seek their minimum energy
state, and also that it is possible to obtain a uniformly captured single layer of particles in
vivo, these results are quite promising. If drug-loaded magnetic particles can be
crosslinked to contain as much as 40% by weight drug, 40% polymer, and 20%
magnetite, this data provides a large working scale for dose ranging. If a dose is to be
delivered intravenously, it will become heavily diluted, and dispersed throughout the
body. It would appear that if the dose is to be delivered by local arterial puncture or
catheter release, an extremely high efficiency can be obtained. Both are worth
investigating, and will only be truly be understood from in vivo trials.
In summary, for a range of 3 to 5mm in diameter, with different geometries, and for
normalized saturation magnetizations from 3 to 150 emu/g, magnetic particles were
successfully captured at significant numbers. While a material with such a high
saturation magnetization as 302 grade stainless steel could never be used for an
endovascular implant, it provided an opportunity to test the upper bounds of the full
latitude of the capture method. All results presented here further validate the hypothesis
that the method could be successful by altering an implant’s alloy content, or by the
addition of small, soft magnetic features.
85
(a) (b)
Figure 4.1: (a) 304 stainless steel rolled stent. The stent is 5mm in diameter by 2cm in length. The wire diameter is 150µm. (b) 302 stainless steel compression spring. The spring is 3mm in diameter by 2cm in length. The wire diameter is 355µm.
86
(a)
(b)
Figure 4.2: (a) Broad view of the pipe flow experimental setup, consisting of PVC tubing with an enclosed stent or spring, mounted within a uniform field, and connected to a flow pump. (b) Close up of magnetic spring in the pipe flow system between two magnetic coils.
87
Table 4.1: Capture results for 5mm diameter 304 grade steel molded stent flow experiments with 25mL doses at 1.0% by volume concentration of 350nm magnetic particles. Results are given in percentage of capture of the total particle dose, and the approximate number of captured particles. Alternating Gradient Magnetometer error is ±2.0%.
Capture Results for 304 Stainless Steel Molded Stents
Plating Height % Capture Numerical Capture (x 109)
Unplated 6.8 %
8.16
3.1µm
14.3 %
17.2
5.2µm 15.9 % 19.08
88
(a)
Percentage Capture per Saturation Magnetization - 304
Molded Stent
6.8
14.315.9
0
2
4
6
8
10
12
14
16
18
2.97 11.45 17.22
Normalized Ms (emu / g)
% 1.00%
(b)
Numerical Capture per Saturation Magnetization - 304
Molded Stent (in 109 Particles)
8.16
17.219.08
0
5
10
15
20
25
2.97 11.45 17.22
Normalized Ms (emu / g)
#
1.00%
Figure 4.3: (a) Shows the percentage of capture of 350nm magnetic particles by a 304 grade steel molded stent against the normalized saturation magnetization data for unplated as well as 3.1 and 5.2µm CoNi plated samples. (b) Shows the capture numerically in 109 particles, against normalized saturation magnetization. Alternating Gradient Magnetometer error is ±2.0%.
89
Figure 4.4: 5x objective fluorescent images of captured 350 nm fluorescent magnetic particles on 302 stainless steel springs from pipe flow experiments. Top Images: INSIDE of Spring Bottom Images: OUTSIDE of spring
Unplated 2.5 µm plated 5.5 µm plated
300 µm 300 µm
300 µm
300 µm
300 µm 300 µm
90
Table 4.2: This table presents experimental results for 302 stainless steel compression springs that are unplated, and at CoNi plating heights of 2.5 and 5.5 µm, for dose concentrations of 0.5, 1.0, and 2.0% 350nm magnetic particles by volume in deionized water. Results are given in percentage of capture of the total particle dose, and the approximate number of captured particles. Alternating Gradient Magnetometer error is ±2.0%.
Spring Flow Results – 0.5% Particle Concentration
Plating Height % Capture Numerical Capture (x 109)
Unplated 20.1 % 4.82
2.5µm
22.6 % 5.42
5.5µm 23.1 % 5.54
Spring Flow Results – 1.0 % Particle Concentration
Plating Height % Capture Numerical Capture (x 109)
Unplated 12.6 %
6.05
2.5µm
15.6 %
7.49
5.5µm 16.3 % 7.82
Spring Flow Results – 2.0% Particle Concentration
Plating Height % Capture Numerical Capture (x 109)
Unplated 7.6 %
7.3
2.5µm
11.3 %
10.8
5.5µm 13.5 % 12.7
91
(a)
Percentage Capture - Unplated Springs
20.1
12.6
7.6
0
5
10
15
20
25
0.5 1 2
Particle Dose Concentration
%
(b)
Numerical Capture (in 109 Particles) - Unplated Springs
4.82
6.05
7.3
0
1
2
3
4
5
6
7
8
0.5 1 2
Particle Dose Concentration
#
Figure 4.5: (a) Shows the percentage of capture of 350 nm magnetic particles by an unplated 302 steel compression spring for each dose concentration. (b) Shows the capture numerically in 109 particles, for each dose concentration. Alternating Gradient Magnetometer error is ±2.0%.
92
(a)
Percentage Capture - 2.5um Plated Springs
22.6
15.6
11.3
0
5
10
15
20
25
0.5 1 2
Particle Dose Concentration
%
(b)
Numerical Capture (in 109 Particles) - 2.5um Springs
5.42
7.49
10.8
0
2
4
6
8
10
12
0.5 1 2
Particle Dose Concentration
#
Figure 4.6: (a) Shows the percentage of capture of 350 nm magnetic particles by a 2.5µm CoNi plated 302 steel compression spring for each dose concentration. (b) Shows the capture numerically in 109 particles, for each dose concentration. Alternating Gradient Magnetometer error is ±2.0%.
93
(a)
Percentage Capture - 5.5um Plated Springs
23.1
16.3
13.5
0
5
10
15
20
25
0.5 1 2
Particle Dose Concentration
%
(b)
Numerical Capture (in 109 Particles) - 5.5um Springs
5.54
7.82
12.7
0
2
4
6
8
10
12
14
0.5 1 2
Particle Dose Concentration
#
Figure 4.7: (a) Shows the percentage of capture of 350 nm magnetic particles by a 5.5µm CoNi plated 302 steel compression spring for each dose concentration. (b) Shows the capture numerically in 109 particles, for each dose concentration. Alternating Gradient Magnetometer error is ±2.0%.
94
(a)
Percentage Capture per Saturation Magnetization - Springs
20.1
22.6 23.1
12.6
15.6 16.3
7.6
11.313.5
0
5
10
15
20
25
116.9 140.4 155.4
Normalized Ms (emu / g)
%
0.50%
1.00%
2.00%
(b)
Numerical Capture per Saturation Magnetization - Springs
4.825.42 5.546.05
7.49 7.827.3
10.8
12.7
0
2
4
6
8
10
12
14
116.9 140.4 155.4
Normalized Ms (emu / g)
#
0.50%
1.00%
2.00%
Figure 4.8: (a) Shows the percentage of capture of 350 nm magnetic particles by a 302 steel compression spring against the normalized saturation magnetization data for unplated as well as 2.5 and 5.5µm CoNi plated samples. Data is presented for each dose concentration. (b) Shows the capture numerically in 109 particles, against normalized saturation magnetization. Data is presented for each dose concentration. Alternating Gradient Magnetometer error is ±2.0%.
95
CHAPTER 5: CELL CULTURE BIOCOMPATABILITY ANALYSIS OF
MAGNETIZABLE IMPLANTS AND SUPERPARAMAGNETIC
NANOPARTICLES
5.1 INTRODUCTION
If the magnetic drug delivery system described within this thesis were to be completely
successful in capturing magnetic particles to implanted stents, and even in the therapeutic
transfer of drugs from carriers to tissue, proper risk analysis of numerous factors must
still be considered in the development of the system. For instance, if it was to be
accepted that a 1-10µm thick coating was to be deposited over the entire surface of a
stent (disregarding here the question of potential mechanical consequences of such a
layer), the issue of chemical reactivity of the CoNi alloy within the body comes
immediately into play. Numerous studies have indicated the toxicity of free Cobalt
within the body (Lonnemark, et al., 1989; Messer, et al., 2005). The coating must either
be heat treated to anneal the plated layer, or passivated by means of a metal, ceramic, or
polymer coating. The selection of material is a more of a manufacturing issue, but a
flexible non-degradable polymer, or a nanolayer of gold have arisen as acceptable
choices.
The effects of magnetic fields (both static and high gradient) on biological materials have
been analyzed extensively in the literature, with a variety of theories regarding safety
from the macro-scale to single cells (Bell, et al., 1991; Buemi, et al., 2001; Kirschvink, et
al., (2) 1992; Schenck, 2000), as well as the abilities of large (6T+) magnetic fields to
96
sediment erythrocytes (Iino, 1997; Sakhnini and Khuzaie, 2001; Watarai and Namba,
2002; Zborowski, et al., 2003). Additional studies have found magnetite and magnetite-
based composites to be extremely well tolerated by the body (Lonnemark, 1989). For the
modestly sized fields and magnetic features presented within the proposed design,
consultation of this literature increased confidence in low risks regarding these issues.
This chapter details the examination of any effects on the growth, morphology, or
behavior of endothelial cells due to the magnetic field gradients of the wires of an
electroplated 316L Stainless Steel mesh. The magnetic field gradients of the CoNi
coating rapidly decay after the removal of an applied magnetic field. Therefore,
endothelial cell cultures were investigated with and without constant application of an
external field to saturate the magnetization of the coating. Interestingly enough, one
original desire was to examine if cells would survive the delivery of large numbers of
magnetic particles to cultures, but were actually found to compartmentalize the particles
into the cell. As a result, investigations studying the ability for endothelial cells with
internalized magnetic particles to maintain normal growth, morphology, and behavior, as
well as to be captured magnetically to mesh surfaces were also performed.
5.2 MATERIALS AND METHODS
5.2.1 Model Implant Selection and Preparation
The implant material selected for biocompatibility studies was the woven, 316L stainless
steel wire mesh (140µm wire diameter, 400µm apertures) that can be seen in Figure
97
3.2(a). This particular material was selected due to its large strut spacing, and extremely
low magnetization. As a result, a large difference in response to applied magnetic fields,
and subsequently in capture ability, can be compared between a virtual non-magnetic
316L mesh and a CoNi electroplated 316L mesh.
Because of the likely chemical reactivity of the CoNi coating to cells and culture
medium, and in order to provide a “level playing field” on which cells could grow and
easily be studied, poly (dimethyl siloxane) (PDMS) (Dow Corning, MI) was used to
finely passivate the mesh. Depending on the desired cover slip size (0.25mm thick,
ranging from 5mm to 25mm in diameter in the below listed experiments), a volume of
PDMS and curing agent (10:1 polymer to curing agent mass ratio) was dropped onto the
glass cover slip. The slip was placed by vacuum upon a wafer spinner at the micro-
fabrication facilities of the Drexel University Clean Room, and was spun for an
appropriate period (5 to 25 seconds depending on the size of the slip) to obtain an even
layer of 150µm of PDMS.
A 5mm diameter circular punch-out of mesh (at 140µm thickness) was then dropped on
top of the PDMS layer, sinking into the polymer. The preparation was then cured under
vacuum at room temperature overnight, and measured with digital calipers to verify
height. This creates a roughly 10µm tall boundary layer between the mesh and the cell
culture surface, and while not identically physiologically relevant, provides an excellent
working model for studying endothelial cell cultures in close proximity to magnetic
98
gradients, as well as for endothelial cells with internalized magnetic particles in close
proximity to magnetic forces when an external field is applied.
5.2.2 Magnetic Particle Selection
As described in Chapter 3, Spherotech magnetic particles were selected for all pipe flow
experiments. The sample of beads used from Spherotech, were 20% γ-Fe2O3 magnetite
by weight and labeled with nile red fluorescent pigment, had a nominal diameter of
350nm with approximately 10% variance in size (Spherotech, IL). Particles come in
2mL water solutions concentrated at 1% w/v or 4.8 x 1011 particles/mL. These particles
have a carboxylate per nm2 of surface area, which can be used as a linker for peptides.
5.2.3 Culture of Endothelial Cells
Bovine aortic endothelial cells (BAECs) were selected as an appropriate culture model,
and were obtained from Dr. Keith Gooch, PhD (University of Pennsylvania, Department
of Bioengineering, Philadelphia, PA). These cells were previously isolated by standard
technique at the University of Pennsylvania in 2002. All experiments were be performed
at low-passages (<10). Cells were routinely cultured in low glucose DMEM (Sigma, MO)
supplemented with 10% Qualified Heat inactivated Fetal Bovine Serum (Sigma, MO) and
1% 2.5mM L-glutamine (Sigma, MO). After expansion of the culture using 100µg/mL
streptomycin (Sigma, MO), and 100U/mL penicillin (Sigma, MO) per 500mL batch of
medium, it was determined that BAECs could be cultured with ease without the use of
antibiotics or antimycotics, so they were removed from future batches of culture medium.
99
Cells were routinely cultured in 75cm2 flasks in a Fisherbrand cell culture incubator at
37˚C and 5.0 % CO2 (Fisher, IL). In preliminary experiments determining the
preparation of mesh for culturing, the BAECs were shown to grow poorly on silicone and
glass surfaces, and when these surfaces were treated with coated with a 1% Rat Tail Type
I Collagen (Sigma, MO), cells grew robustly, maintaining a growth rate and visible
morphology alike to BAECs cultured in T75 flasks.
5.2.4 Culture of BAECs on Magnetic Mesh
5mm punch-outs of magnetic and non-magnetic mesh were prepared, and treated with
PDMS as described above. Prior to use, each of the 6 cover slips (25mm diameter x
0.25mm thick, 3 with magnetic mesh, 3 with non-magnetic mesh) were washed with
Ivory soap and water, and rinsed thoroughly. Following cleaning, each mesh was placed
in a glass petri dish within a laminar flow hood, and soaked in 70% ethanol for 30
minutes. After 30 minutes, all ethanol was aspirated, and the cover slips were allowed to
dry for 20 minutes.
Two separate 6-well plates were obtained (one for magnetic mesh, the other for non-
magnetic mesh), and 3 slips of each experimental group were placed in each. A
preparation of 1% by volume Rat Tail Type I Collagen (Sigma, MO) in Phosphate
Buffered Saline (Sigma, MO) was used to add 150µL to each well. After 30 minutes, the
collagen solution was aspirated, the slips rinsed twice with an equal volume of PBS, and
allowed to dry for 10 minutes within the laminar flow hood.
100
BAECs, routinely cultured as described above, were then seeded to each cover slip at 1:2
split ratio (adjusted to cm2 growth area). 3cm long by 0.5cm tall by 0.5cm wide pieces of
neodymium permanent magnets were placed under each well, for both non-magnetic and
magnetic mesh, and separated by plastic spacers. These pieces applied an approximately
500 Gauss magnetic field at the center of the mesh as measured by a handheld
Gaussmeter (Lakeshore Cryotronics, OH). These magnets were kept in place, under the
wells, and placed in the cell culture incubator for 30 minutes while cells were allowed to
attach to the surface of the collagen-coated PDMS. 30 minutes was selected as a modest
period for which a magnetic drug delivery injection may be performed, saturating the
magnetic moment of the material for that period. At the conclusion of 30 minutes, the
magnets were removed and the samples were left in culture, and imaged by phase
contrast at 24 hours.
5.2.5 Internalization of Magnetic Particles into BAECs
Magnetic particle uptake into or chemical attachment on to cell cultures has been used as
a means of mechanically stressing cells (Wang, et al., 1993), and also considered as a
means for cellular localization (Consigny, 1999; Frank, 2004; Mertl, 1999), but has not
yet been successfully presented by a magnetic delivery method. When considering the
study of the biocompatibility of magnetic particles with cell cultures, and our prior
studies of surface labeling of cells with magnetic particles, it became of interest to study
the following:
1) Can 350nm diameter magnetic particles be delivered by vesicles into the cytoplasm or membranes of BAECs?
2) If so, can these cells survive, be split and re-seeded, and grow to confluence on their newly plated surfaces?
101
3) How and where are these particles orienting within the cell? 4) Do these magnetic BAECs maintain morphology and response to stimuli
consistent with unlabeled cells? 5) Can these magnetic cells be attracted to magnetic mesh, both statically and under
high flow rate? 6) Does this exposure to magnetic forces cause cell death?
These questions were approached by the experiments described within this and the
remaining experimental sections within this chapter.
On the day desired for delivering the particles to the cells, and according to published
protocol by researchers at Thomas Jefferson University in Philadelphia (Consigny, et al.,
1999), a simple method for weak surface attachment of Bovine Serum Albumin was
adapted to 350nm diameter magnetic particles from the 3-5µm diameter particles used by
the authors. As mentioned above, the particles maintain a carboxyl group per nm2 of
surface area. A 1% by volume solution of Bovine Serum Albumin (BSA) (Sigma, MO)
in Phosphate Buffered Saline (PBS) (Sigma, MO) was prepared according to the
literature.
Before labeling the magnetic particles with albumin, each sample of particles was rinsed
3 times with 70% ethanol, using neodymium permanent magnets for separation between
rinses. For every 1011 particles, 1mL of BSA/PBS dilution was added to the 15mL
centrifuge tube in which the particles were cleaned. Following the addition of the
solution, the sample was gently shaken on a gyrating shaker (Fisher, IL) for 1 hour. After
shaking, the particles were separated from the solution using a neodymium permanent
102
magnet, and re-suspended in appropriate volume of culture medium, and stored in a 37ºC
water bath (Fisher, IL).
The day before particle preparation, BAECs were routinely cultured as described above,
and seeded on 6 tissue culture treated cover slips (25mm diameter x 0.25mm thick) at a
1:2 split ratio (adjusted to cm2 growth area), and cultured for 1 day to reach confluence.
At confluence, all cover slips were seeded with magnetic particles at a 4 x 103 particle to
cell ratio, using particle preparations as described above. No magnets were placed
beneath the cover slips.
24 hours after particle seeding, each cover slip was rinsed three times with warmed
culture medium, and submerged in Earle’s Balanced Saline Solution (Sigma, MO)
supplemented with 10% Fetal Bovine Serum (Sigma, MO). Each cover slip was imaged
fluorescently to assess internalization and cell survival. After imaging, fresh, warm
culture medium was added back to each cover slip.
Following imaging, each cover slip was split at a 1:6 split ratio and re-seeded to fresh
cover slips. Cells were imaged at 1 hour, monitored for growth over 3 days by a standard
phase contrast Nikon microscope (Nikon, Japan), and imaged fluorescently as described
above, at 72 hours.
103
5.2.6 Confocal Imaging and Magnetic Characterization of Magnetic BAECs
Alternate BAEC samples were prepared as described above, but prepared on glass slides
containing 8 wells (0.69cm2 per well), and labeled with albumin-treated magnetic
particles as described above, but at a 2 x 103 particle per cell ratio, in order to obtain a
clear view of particle internalization and orientation, by confocal microscopy at 60 and
120X. After 24 hours, and particles had been uptaken by the cells, the well covers were
removed from the slide, and a drop of Vectorshield Mounting Medium with DAPI
(Vector Labs, CA) was applied to each of the 8 cultures, and a cover slip was fixed over
the sample with nail polish and allowed to set for 24 hours at 4ºC. Confocal imaging was
performed on a Leica TCS SP2 Confocal Microscope with Louise Bertrand of the
Department of Neurobiology and Anatomy at Drexel University College of Medicine.
Secondly, in order to estimate an average loading of particles per cell, the MicroMag
Alternating Gradient Magnetometer (Princeton Measurements, NJ) was used. BAECs
were routinely cultured as described above, and seeded onto nine 5mm diameter by
0.25mm thick glass cover slips resting in the bottom of wells in a 48 well plate. Each slip
was treated with Rat Tail Type I Collagen as described above. The slips remained in
culture for 24 hours until the cells had reached confluence. Cells were loaded with
particles by methods as described above, at a loading density of 4 x 103 particles per cell.
Approximately 8 x 103 BAECs can grow upon a 5mm diameter cover slip. After 24
hours, resting medium was aspirated, the slips were gently rinsed with culture medium
three times, top and bottom, following by a gentle alcohol swabbing of the bottom of slip,
to remove any particles that may have been attached. Each of slips was then measured for
104
saturation magnetization by AGM by methods as described in Chapter 4. The results
were then averaged, and based on the known quantity of cells per slip, used to give an
estimate of the average particle loading per cell.
5.2.7 Shear Stimulation of Magnetic BAECs vs. Non-Magnetic BAECs
In order to further examine the behavior of BAECs with uptaken magnetic particles,
experiments were performed side by side with controls of unlabeled BAECs.
Experiments studying calcium response in endothelial cells under shear stimulation have
been designed and performed regularly by Dihui Hong and Dr. Kenneth A. Barbee
(Drexel Universtiy School of Biomedical Engineering, Science, and Health Systems).
Dihui Hong was generous enough to perform these experiments on magnetic BAECs
alongside her non-magnetic BAECs, and provide analysis of the resulting calcium
response in each.
BAECs were routinely cultured as described above, seeded onto 6 tissue culture treated
cover slips (25mm diameter x 0.25mm thick), and cultured for 1 day to reach 90%
confluence. At 90% confluence, 3 cover slips were seeded with magnetic particles at a 4
x 103 particle to cell ratio, using particle preparations as described above, the other 3
cover slips.
Endothelial cells were loaded in the dark with fluo3 by incubation with 5uM Fluo3-
acetoxymethyl ester (Molecular Probes, Inc., OR) in Dulbecco’s phosphate buffered
105
saline (DPBS) (Sigma, MO) at pH 7.4, for 40 minutes at room temperature. The cover
slip was then rinsed three times with DPBS before experiments were performed.
For cell shearing experiments, a custom-built controlled cell-shearing device based on a
cone and plate configuration was mounted on the microscope stage and used to apply
precise mechanical loading conditions to the endothelial cells. For each experiment, a
25mm diameter cover slip was placed in the circular recess in the plate. Vacuum pressure
was applied to hold the cover slip and prevent motion during shear experiment. A volume
of 1.5ml DPBS was added to the well to fill the gap between the cone and plate. For the
mechanical loading period, endothelial cells were monitored for about 30 seconds under
static conditions prior to the onset of shear stress to establish the basal levels of calcium.
Then, the shear stress was ramped up linearly to 20dyn/cm2 over 0.1 seconds and
maintained at a steady level for 5 minutes.
Cell fluorescence was monitored and recorded using a Nikon Diaphot TE300 Eclipse
epifluorescent microscope (Optical Apparatus, Inc, PA) with a 20X objective. Fluo3 was
excited at 488nm and emitted fluorescence at wavelength of 515nm upon binding Ca2+.
The illumination was controlled by means of an electronic filter wheel (Lambda 10-2,
Sutter Instruments Co., CA). The emitted fluorescence passed through a barrier filter and
was detected by an intensified CCD digital camera unit (Vedio Scope International, Ltd.,
VA). The rate of image acquisition was 2 seconds per frame. Prior to the each stimulus,
images were recorded for at least 30 seconds. Axon Workbench image acquisition
106
software (Axon Instruments, Inc., CA) was used to acquire fluorescence images and
perform post-acquisition analysis.
5.2.8 Static and Flow Capture of Magnetic BAECs by Magnetic Mesh
Once it was determined that magnetic particles could be delivered to the inside of
BAECs, the logical next step was to examine if these magnetic BAECs could be attracted
to 316L mesh electroplated with CoNi, under the influence of a magnetic field. This was
first demonstrated statically, in sterile cell culture conditions, and then by high flow rate
experiments using the parallel plate flow chamber and methods as described in Chapter 3.
For static capture experiments, using the protocol for preparing mesh for cell culture
experiments as described above, 5mm punch-outs of mesh (3 magnetically plated, 3
unplated) were sealed with a thin layer of PDMS onto the center of 25mm diameter cover
slips, followed by cleaning, and coating with collagen. BAECs were routinely cultured in
6 well plates as previously described, and each well was loaded with magnetic particles at
a 4 x 103 particle per cell ratio. BAECs were allowed to remain in culture for 24 hours
for uptake of the magnetic particles.
On the following day, two 6 well plates were prepared, one containing the 3 cover slips
prepared with magnetic mesh discs, the other containing 3 cover slips with unplated mesh
discs. Neodymium pieces, as described above, were mounted beneath each well, and
measured to apply an approximately 500 Gauss magnetic field at the center of the mesh.
Magnetic BAECs were routinely split and seeded at a 1:2 split ratio to the cover slips and
107
allowed to remain in culture for 30 minutes under the applied field. At 30 minutes, the
magnets were removed, and the cover slips were imaged fluorescently. Imaging was also
performed at 6 hours, and 24 hours after cell seeding.
For magnetic cell flow experiments, BAECs were routinely cultured on a T75 flask. The
flask was loaded with magnetic particles by methods described above, at a seeding rate of
4 x 103 particles per cell. BAECs were allowed to remain in culture with the particles for
24 hours to internalize them into the cells. After setting up the flow chamber and
magnetic coil under the microscope (according to the methods in Chapter 3), the cells
were trypsinized routinely, and concentrated in 25mL of culture medium. The 500 Gauss
field was turned on, and the particles were delivered past a 2 x 2 cm piece of
magnetically plated 316L mesh at 15cm/s flow velocity, followed by a 25mL rinse with
Earle’s Balanced Saline Solution (EBSS) (Sigma, MO) supplemented with 10% Fetal
Bovine Serum (Sigma, MO). The field was left on during rinsing. Based on AGM results
estimating an average loading success of approximately 103 particles per cell, the
experiments delivered an equivalent number of particles as a concentration 0.25% by
volume in the medium. It should be noted that experiments performed in the flow studies
in Chapters 3 and 4, ranged from 0.5-2.0% by volume. Fluorescent imaging was
performed immediately after rinsing.
108
5.3 RESULTS
5.3.1 Culture of BAECs on Magnetic Mesh
BAEC cultures were successfully grown atop of PDMS/collagen coated magnetic mesh.
Figure 5.1 shows a pair of characteristic side by side images of BAECS cultured on a
magnetically plated mesh, and an unplated mesh. The cultures are virtually
indistinguishable. While the cells are not grown directly on top of the mesh itself,
experiments with a passivate oxide layer over the plated stents paired with fluorescently
stained cells is the next step in further evaluation of endothelialization of magnetically
plated wires in vitro.
5.3.2 Internalization of Magnetic Particles into BAECs
The scarcity of commercially available superparamagnetic particles in the sub-micron
diameter scale has resulted in most frequent use of much larger magnetic particles for
external labeling. From the performed cell culture studies, 350nm polystyrene magnetic
particles coated with albumin were compartmentalized and uptaken by BAEC cultures.
Figure 5.2 shows two images of a BAEC culture with internalized fluorescent magnetic
particles, 24 hours after particle seeding, indicating expected morphology by inspection.
Gentle rinsing with culture medium before imaging aided in ensuring minimal amounts
of free particles. BAECs internally labeled with these magnetic particles will be referred
to at times as mBAECS for the remainder of this chapter.
109
Fortunately, delivery of the particles to the cells did not appear to cause rampant (or any)
pattern of cell death. After imaging at 24 hours, the cells were split at a 1:6 ratio, and
imaged an hour after seeding. Figure 5.3 shows clear definition of internalized particles
around the entire circumference of BAECs, which attached and spread to culture surfaces
as expected from routine cultures, 1 hour after re-seeding. Figure 5.4 shows the same
culture, around 3 days later, reaching full confluence. From the fluorescent images in
Figure 5.4, it would seem that these internalized particles are being distributed to many
of, but not all daughter cells. Figures 5.2, 5.3, and 5.4, combined, indicate that BAECs
can survive, re-seed, and divide after the introduction of these magnetic particles to
cultures.
5.3.3 Confocal Imaging and Magnetic Characterization of Magnetic BAECs
The acquired confocal images provide a clearer view of magnetic particle orientation
inside BAECs, although three-dimensional animations and stacks (which cannot be
presented here) were even more useful. In Figure 5.5, shows a 60X fluorescent confocal
image of fixed samples of BAECs with uptaken 350nm diameter nile red polystyrene
magnetic particles. Particles were seeded more sparsely than attempts at saturation, for
less obstruction of views of their orientation. In Figure 5.6, a view at 120X of the BAEC
from the top left of Figure 5.5 can be seen. In Figure 5.7, a more densely loaded BAEC at
90X is seen. The 3D animations of the cell seen in Figure 5.7 captured by segments along
a rotation of the Z-axis of the system, and indicated that the particles were tending to
accumulate in clumps of 3-5 particles, in columns along the circumference of the
membrane. Very few instances of single particles are seen in these three image
110
sequences, but in many cases larger aggregations are found and difficult to count in terms
of the number of particles contained. The particles do appear to be compartmentalized,
and aggregating within the membrane and not within the cytoplasm. For the cell seen in
Figure 5.6, a stack movie of images acquired along the Z-axis of the system shows
consistent groupings of 3-5 particles, also oriented in columns along the circumference of
the membranes. The 350nm particle diameter appears to be quite optimal for vesicle
uptake. As no magnetic field was used in introducing these particles to cultures, any
clumping or grouping is likely due to surface forces or other chemical phenomena.
5.3.4 Shear Stimulation of Magnetic BAECs vs. Non-Magnetic BAECs
Figure 5.8 shows cell fluorescence over time as measured during cell shearing. The peak
ratio is computed from the amplitude of the peak response to the baseline. The selected
experimental results of cone and plate shearing of mBAECs and control BAECs showed
an amazingly similar calcium response with an almost identical peak ratio, although with
a larger second peak by the mBAECs. Minimal studies were performed simply for
validation of a reasonable cellular response to stimulus after internalization of particles,
which these experiments demonstrate. So in addition to being able to survive, re-plate,
and divide, mBAECs have maintained the ability to respond to shear stimulus. Analysis
of the mechanisms for these calcium responses is within the scope of the Ph.D. thesis
work of Dihui Hong and will not be presented here.
111
5.3.5 Static and Flow Capture of Magnetic BAECs by Magnetic Mesh
BAEC cultures with internalized magnetic particles (mBAECs) were successfully
delivered under static conditions with specificity to the surfaces of a CoNi electroplated
316L mesh. Figure 5.9 provides a selection of four images of samples 30 minutes after
seeding to the magnetic and control meshes, each under a constant exposure to a
magnetic field. Clearly the mBAECs have been delivered to the areas directly over the
wires of the magnetic mesh, while for the unplated mesh sample the cells are very
uniformly distributed across the entire surface. In each case the magnet can influence the
rate at which the cells attach to the surface, but without a doubt the magnetic coating of
the experimental mesh in combination with the field applied, caused those mBAECs to be
captured by the device. As seen in the flow experiments in Chapter 3, the cells did tend
to aggregate at the intersections of the mesh, consistent with those prior results.
At 6 hours after mBAECs were magnetically delivered to the magnetic mesh, Figure 5.10
shows these same samples after the field had been removed 5.5 hours before. Many
mBAECs had begun to migrate off of the wires into the surrounding area, indicating
many had survived the magnetic forces used to capture them to the mesh surfaces. At 24
hours, as seen in Figure 5.11, a great many of the mBAECs have migrated from the area
immediately over the wire into the surrounding growth area, and have nearly reached
confluence. Clearly, there are still large congregations of cells at the wire intersections,
and it is possible that if great numbers were brought down to these intersections that
some cells may have died from isolation from nutrients and proper gas exchange.
112
Cell flow chamber experiments at roughly 0.25% particle concentration (based on a
calculated average particle uptake per cell from AGM analysis) by volume demonstrated
the ability to capture mBAECs at high flow velocity in a parallel plate flow chamber. The
results of these studies can be found in Figure 5.12. It was expected that these cells could
be captured due to an increased susceptibility to magnetic forces due to large particle
concentrations in each cell. This indicates that it may be possible to magnetically deliver
cells of limited availability, such as stem cells, with increased accuracy and site
specificity to their intended destination. No studies of similar makeup have been found
from heavy literature searches, indicating the potential for an extremely innovative
method for the delivery of biologics or therapeutics.
5.4 DISCUSSION
The experiments presented in this chapter are very encouraging in their results. Not only
do vascular endothelial cells succeed in growth within microns of distance from
magnetically plated mesh under the influence of a field, but can be labeled with magnetic
particles, delivered by magnetic force to the surface of the mesh and survive. Attempts
by other laboratories have failed at well dispersed and equal loading of magnetic particles
into cultures, mainly due to the use of larger particles, but also from their use of applied
fields during particle seeding. By placing a permanent magnet underneath the culture
continuously during particle seeding to a culture, these particles may have an increased
tendency to form chains or large aggregates too large for delivery to the inside of a cell,
and may therefore remain on top of the culture. These large aggregates may even be
113
pulled through intracellular space to the bottom of the culture, beneath the cells, but if the
field is left in place throughout, one would expect minimal internalization for micron and
even large sub-micron particles.
In our experiments it cannot be assumed that every cell in the initial culture seeded with
particles contained them after the introduction to the culture, but it does seem quite likely
that many mBAECs distributed particles to daughter cells, and that most cells without
particles are daughter cells of the unlabeled cells from the first culture. Regardless,
certain cells in Figure 5.4 do seem to maintain the same density of fluorescence seen just
after initial particle introduction. In these cases, the question can be raised on whether or
not these mBAECs have undergone mitosis, or were saturated with particles to the point
of inhibition. In most cases, by inspection, the culture had a significantly reduced
fluorescence per cell, indicating particle transfer during cell division, or release of
particles into the surrounding medium during cell division, followed by eventual
reuptake.
These results are all quite promising and raise numerous further questions about
adaptation to other cell cultures, the use of smaller particles, and applications. The ability
to capture mBAECs under high flow opens the door for delivery of cells to implants
within the body, for the purpose of in vivo tissue engineering or wound healing, but also
the prospect of using cells simply as vehicles for drug delivery particles. When particles
themselves are introduced into blood flow, they will come in contact with numerous cells
and substances, and are particularly vulnerable to immune cell response. By loading
114
cells to saturation with magnetic particles, cells will have extremely large magnetic
susceptibilities, likely raising the capture efficiency significantly while reducing concerns
about immune response and blocking of capillaries. In many cases, autologous cell
sources could be utilized with ease. If a cell were used solely as a vehicle for drug-
loaded particles, its survival would be irrelevant as long as the carriers reached the target
site and could effectively transfer therapeutic agents upon arrival. This could be a much
more effective method for magnetically targeted drug delivery, and if it raised capture
efficiency could make intravenous injection a more viable option, rather than the current
“safe bets” of arterial puncture or catheterization.
115
(a) (b) Figure 5.1: Phase contrast images of BAECs grown on (a) CoNi electroplated 316 Stainless Steel mesh (b) Unplated 316L Stainless Steel mesh, each with a thin layer of PDMS, a collagen substrate, and under the application of a 500 Gauss magnetic field for 24 hours. No substantial difference in growth or morphology was detected by inspection of these cultures.
116
Figure 5.2: BAECs with internalized 350nm diameter nile red polystyrene magnetic particles, 24 hours after particles were seeded into confluent wells. These cells are also referred to as “MBAECS” within this thesis.
5X
20X
117
Figure 5.3: BAECs with internalized 350nm diameter nile red polystyrene magnetic particles, 1 hour after being passed and re-plated at a 1:6 splitting ratio.
20X
40X
118
Figure 5.4: BAECs with internalized 350nm diameter nile red polystyrene magnetic particles, 72 hours after being passed and re-plated at a 1:6 splitting ratio. Cells reached confluence earlier in the 3rd day, consistent with routine culture of BAECs split at the same ratio without internalized magnetic particles.
10X F
20X F
20X P
119
Figure 5.5: 60X fluorescent confocal image of fixed samples of BAECs with internalized 350nm diameter nile red polystyrene magnetic particles.
120
Figure 5.6: Fluorescent confocal image of fixed samples of BAECs with internalized 350nm diameter nile red polystyrene magnetic particles. 3D animations of segments along the Z-axis of the system, indicate that the particles tend to accumulate in clumps of 3-5 particles, in columns along the circumference of the membrane.
121
Figure 5.7: Another fluorescent confocal image of fixed samples of BAECs with internalized 350nm diameter nile red polystyrene magnetic particles. 3D animations of segments along the Z-axis of the system, indicate that the particles tend to mostly accumulate in clumps of 3-5 particles, in columns along the circumference of the membrane.
122
Control, shear stress 20dyn/cm2
Peak ratio: 4.10
(a)
0.00
1.00
2.00
3.00
4.00
5.00
6.00
0 30 60
Particle, shear stress 20dyn/cm2
Peak ratio: 4.05
(b)
0.00
1.00
2.00
3.00
4.00
5.00
6.00
0 30 60
Figure 5.8: (a) Shows calcium response fluorescence versus time for a control sample of BAECs and (b) magnetic particle labeled mBAECs during exposure to fluid shear stress. The two samples had an almost identical ratio of peak response to baseline, with the mBAECs having a second peak of higher amplitude and slower decay.
123
Figure 5.9: BAECs with internalized 350nm diameter nile red polystyrene magnetic particles 30 minutes after being passed and magnetically delivered to a PDMS coated magnetically plated and unplated mesh. Specific attachment to the wire regions of the magnetic mesh can be seen, with very non-specific and disperse attachment on the unplated mesh.
Plated Mesh Unplated Mesh
5X
10X
124
Figure 5.10: BAECs with internalized 350nm diameter nile red polystyrene magnetic particles 6 hours after being passed and magnetically delivered to a PDMS coated magnetically plated mesh. Remnants of the specific attachment to the wire regions of the magnetic mesh can still be seen, with cells beginning to migrate and spread across the surface.
5X 10X
125
Figure 5.11: BAECs with internalized 350nm diameter nile red polystyrene magnetic particles 24 hours after being passed and magnetically delivered to a PDMS coated magnetically plated mesh. Remnants of the specific attachment to the wire regions of the magnetic mesh can still be seen, but the cells have migrated, spread, and grown to near confluence.
5X 10X
126
10X 20X
Figure 5.12: Captured BAECs with internalized 350nm diameter nile red polystyrene magnetic particles immediately after being magnetically delivered at high flow rate in a parallel plate flow chamber to a 2 x 2 cm piece of magnetically plated mesh. An entire T75 flask of cells were labeled at approximately 103 particles per cell, and delivered at 15cm/s flow velocity.
127
CHAPTER 6: CONCLUSIONS AND FUTURE WORK
6.1 SUMMARY OF PRINCIPLE FINDINGS
This thesis has provided a proof of concept to the proposed magnetically targeted drug
delivery system. This system operates by first embedding magnetic implants, such as a
stent, at designated sites in the cardiovascular system, and then attracting injected doses
of magnetically susceptible drugs to those implant sites with the aid of a modest external
magnetic field. I have experimentally validated our claim that magnetic drug carriers are
attracted to regions of the strongest magnetic field gradients. Previous attempts to use
magnetic particles in these applications have relied on high gradient magnetic fields
produced by magnets external to the body to direct magnetic particles to specific
locations (Flores, 2002; Gallo, et al., 1997; Lübbe, et al., 2001; Mossbach, et al., 1979;
Rudge, et al., 2001). This limits the range of their applications. The main disadvantage of
this approach is that externally generated magnetic fields apply relatively small and
insufficiently local forces on micron and nano-scale magnetic particles. As a result, there
has been great interest in devising systems that produce strong and highly localized field
gradients in the interior of the body. The presented in vitro studies within this thesis
provide a clearer picture of our two source method for magnetic drug delivery, and its
scalability, biocompatibility, and functionality.
128
My research has focused on designing and analyzing biocompatible implants that
produce strong magnetic field gradients near the surface of the implant so that sufficient
doses of drug can be captured. Previously published computational models describing
the seeding of magnetic microspheres along the endothelium of smaller vessels were
presented. The results of these models provided the first evidence of the ability of local
magnetic field gradients to trap magnetic particles for the purpose of eliciting a
therapeutic response. Adaptation of these models to the analysis of a magnetizable steel
mesh provided insight for implant magnetization, external magnetic field parameters,
particle and vessel sizes, as well as preliminary dose concentrations for beginning in vitro
flow simulations.
Rational magnetic implant design began with the selection of stent-simulating materials
of different geometries, mesh sizes, and metallic content, suitable for in vitro flow
experimentation. Stainless steel materials ranging from 316 to 302 grade were chosen, in
grid-like mesh geometries, as well as in the form of a compression spring. A soft
magnetic alloy of Cobalt-Nickel was selected as a practical material for increasing the
saturation magnetization of the materials, while retaining a very low state of
magnetization in the absence of an externally applied magnetic field. An electroplating
setup was developed utilizing a cobalt anode, borate bath containing scaled
concentrations of cobalt and nickel, and controlled by a potentiostat. By combining the
use of very weakly magnetic materials (316L SS) and highly magnetic materials (302 SS)
with varied plating heights of soft magnetic alloy, flow experiments were able to examine
the scalability of magnetic capture over a range of saturation magnetizations.
129
Two flow systems were used to test the proposed design for magnetic particle capture to
the wires of model stent materials. The first system employed was a project specific
design of a parallel plate flow chamber (PPFC) adaptable to various channel heights and
capable of sustaining high volumetric flow rates needed to obtain physiologically
significant flow velocities. These experiments provided qualitative results from
fluorescent microscopy that validated model predictions that magnetic particles could be
captured using the proposed design. The system was validated for particle sizes ranging
from 130nm to 2µm in diameter, but optimized for 350nm and 2µm diameter
superparamagnetic particles stained with rhodamin for nile red fluorescence. These
particles were tested and validated for magnetic capture to both 316L and 304 steel
meshes, magnetic capture through layers of silicone in order to simulate scar tissue,
capture of particles concentrated in porcine blood, and for a range of dose concentrations
and volumes.
The second system for flow analysis of the proposed design utilizes three-dimensional
model implants placed within a pipe flow system. Large vessel scaled implants molded
from 304 grade steel into 5mm diameter tubes, as well as 3mm diameter 302 grade steel
compression springs were chosen as the materials for testing. These materials maintain a
much higher inherent saturation magnetization due to their alloy content, but were also
electroplated with various heights of soft magnetic alloy to examine capture over a range
of magnetic properties. Pipe flow analysis also allowed for a controlled method to
analyze uncaptured magnetic particles from a single dosage pass using a MicroMag
Alternating Gradient Magnetometer (AGM) (Princeton Measurements, NJ). AGM
130
analysis provided capture efficiency percentages, as well as insight into numerical
capture capabilities and its variance with dose concentration and material magnetization.
These results are paramount in the appropriate selection of dose ranges for the first
studies to test the feasibility of the proposed design in vivo.
The final area of examination for completing a thesis on a magnetically targeted drug
delivery system was indeed biocompatibility. While the risks of the use of Cobalt-Nickel
coatings and magnetite-based magnetic particles must be studied in vivo, prior literature
gave insight into their tissue tolerance. The goal of these studies was to examine any
effects on the growth, morphology, or behavior of endothelial cells due to the magnetic
field gradients of the wires of an electroplated mesh. An earlier idea was to examine if
cells would survive the delivery of a high concentration of magnetic particles to cultures,
but were actually found to compartmentalize the particles into the cell. As a result,
investigations studying the ability for endothelial cells with internalized magnetic
particles to maintain normal growth, morphology, and behavior, as well as to be captured
magnetically to mesh surfaces were also performed. The findings of this work were that
endothelial cultures not only appear to tolerate magnetic gradients when grown in the
presence of a magnetically plated mesh, but when delivered magnetic particles, could
compartmentalize and uptake them. These cultures were also found to attach, spread, and
divide at rates consistent with control cultures, and to respond to fluid shear stimulation.
Magnetic endothelial cells were also successfully delivered magnetically to the surface of
magnetic mesh, both in static culture as well as in high flow rate in the parallel plate flow
chamber. Significant numbers of these magnetic endothelial cells were able to survive
131
and grow, even after exposure to such potentially traumatic magnetic forces. These
results provide a starting point for the use of cells as delivery vehicles for magnetic drug,
as well as for delivery of stem cells and other cell types for innovative methods of
treatment.
6.2 ADVANTAGES AND RISKS
Many clinicians involved in discussions pertaining to this proposed design for this system
see the obvious benefits implied in its abilities. The original idea for this implant-based
method came from my study of a smooth muscle cell injury paradigm designed by Dr.
Kenneth A. Barbee. While learning about the progression of restenosis of coronary
arteries at the cellular level, I became aware of the development of drug-eluting stents by
the medical device industry. It occurred to me that while the risk for restenosing of the
coronary artery post-stenting would primarily occur within the first 6-12 months after
angioplasty, stents are still foreign bodies. Most recently, new studies in late 2004 have
shown that the even the oxide layers passivating the metallic elements of stents do not
remain perfectly continuous in the long term, and are believed to have effects on cellular
expression of adhesion molecules by endothelial cells. This is a reminder of the fact that
while different implants have been placed within the body well back into the 20th
Century, stents have been in widespread use only since the early 1990s. As they are
foreign bodies, and we know the body will always recognize them as such, there remains
a long-term threat for future inflammation, thrombus formation, and other complications.
132
In our current fast food culture with patients as young as their 30s receiving coronary
stents, if such patients make the appropriate lifestyle adjustments, they may otherwise
live to be 90 years old. What happens to these stents as the patient ages? It became
obvious to me that if one could safely adapt stents magnetically for the delivery of
magnetic drug, the ability to dose the site repeatedly over time would provide a clear
clinical advantage. Not only was repeatability advantageous, but if a high degree of
control over magnetic dosing could be built into the design, clinicians could have more
flexibility for treating on a patient to patient basis. The long shelf-life of controlled
release spheres and non-drug-coated stents, compared to the expensively sterilized and
briefly storable drug-eluting stents were other obvious benefits.
Both the scientific and business world questioned the need for another method for the
treatment of restenosis if drug-eluting stents were already obtaining significantly reduced
rates of restenosis in clinical tests and after market introduction. I then came to believe
that our proposed method for local drug delivery does not have to be disruptive, but
rather complementary; meaning that it can be paired with drug-eluting stent technology to
provide an enhancement for additional doses along the lifetime of the implant. Drug
eluting stents are limited by some of their problems: complications related to
implantation, cracking of the polymer layer, limited dose size, shelf life, and the fact that
they can only provide a single dose. For the other uses of stents (biliary, renal, brain),
pacemakers, and orthopedic implants, a dose at implantation may not be necessary,
making the addition of magnetic drug delivery functionality beneficial for prevention of
complications in the future as needed. We have also considered the use of endovascular
133
and extravascular implants for treatment of localized tumors. Implants in these cases
would have the sole or primary function of facilitating local magnetic drug delivery, and
could be implanted by catheter, or in cases where open chest surgery is already required
to excise tumors, implanted extravascularly over vessels or organs. This could provide a
future option for local chemotherapy at the same site should carcinomas be found to be
re-growing during remission.
Obviously the benefits of this system must not come at the cost of increased risk in other
arenas, such as chemical tolerance of a magnetic coating or final compositions of
polymer and magnetite crystals. The approach I have taken with this thesis utilized
previously FDA approved magnetic particle composites, as well as both soft magnetic
coatings and magnetic alloys in order to explore the range of manufacturing capabilities
that maintain the fundamental essence of the technology: controllable local delivery of
magnetic drug to the wires of a stent. While both soft magnetic coatings and varied alloy
composition appear to possess functionality for adapting implants to this magnetic drug
delivery system, it is possible their chemical effects, and responses to MRI will differ.
As biocompatibility surely represent a roadblock in clinical testing, having flexibility in
the design will make the system that much more attractive to industry. Regarding MRI,
technology is under development (Biophan, MA) which uses magnetic material to
enhance MRI safety and quality, which is promising to the possibility that there is a
happy medium between such enhancements and the point of magnetization of an implant
that would cause safety issues relative to movement or torquing of the implant. The
question of whether a technology such as this would be worth the loss of future MRI
134
procedures for the patient is a significant one. Development and in vivo testing of final
designs should heavily extend efforts to construct a magnetic drug delivery system that
will allow for safe and effective MRI procedures for patients receiving the implants.
6.3 FUTURE WORK
During the final months of writing this thesis, in vivo protocol was developed in order to
test the technology in large adult rats for simple validation of feasibility. This protocol
was developed and written under the expert tutelage of Dr. Ari D. Brooks, M.D., at the
Department of Surgery at Drexel University College of Medicine. These studies involve
implantation of 1mm diameter by 0.5mm long stainless steel compression springs into the
common iliac artery of the rat. The same 350nm diameter magnetic particles (Spherotech,
IL) were selected for use, as the particles do not need to be degradable for non-survival
surgeries simply testing feasibility. Surface chemistry has been developed to label these
particles with the gamma-emitting isotope Technetium-99m. Radiolabeling will provide
a means for gamma imaging as well as highly accurate gamma counting for quantifying
capture success and biodistribution. After the springs are implanted, the animals will
receive doses scaled from flow experiment results, under the influence of an externally
applied uniform magnetic field. These studies will examine the capabilities of both
intraarterial and intravenous injection.
Because of the nature of the technology, as an enhancement or system, rather than a
newly designed device structure, financial and material support from industry will be
135
important to advancing this work towards clinical testing. With that partnership, we can
work together for the selection of the best magnetization method for their stent or implant
design, in reaching manufacturability for a product for large animal testing. The costs of
such work make it important that all precautions are taken so that the Food and Drug
Administration will deem the results of an animal disease model legitimate enough to
justify human testing upon its successful completion.
Thus, the results of our rat studies will provide the needed information on dose ranging,
administration route, external field requirements, and biodistribution data crucial to
bringing the technology to a larger animal model. The care, complication, and cost of
large animal studies make these rat models that much more productive. The development
of an in vivo model with large rabbits or pot belly pigs will allow us to eventually study a
transgenic disease model, available for atherosclerosis for these two species. After
selection of an appropriate biodegradable magnetic carrier comparable to those used in
the design of the system, feasibility of magnetic capture must then be tested in that large
animal model. This will be followed by the analysis of the pharmacokinetics of the
system; mainly the successful transfer of molecules from captured particles to local
tissue, and the biodistribution of drug that does not reach the target site. The ability of
the vessel to appropriately endothelialize the implant must also be examined, and what
will happen to that implant when exposed to MRI after a significant implantation time.
Once these steps are complete, an appropriate transgenic model for the chosen species
can be performed to see if all of these pieces of the puzzle add up to the real treatment of
disease. If they do, the possibilities for this technology are as far as the eye can see.
136
LIST OF REFERENCES
Babincova, M., Babinec, P. Controlled drug delivery using magnetoliposomes. Cellular
& Molecular Biology Letters, vol. 2, pp.3-7, 1997. Babincova, M., Babinec, P., Bergemann, C. High-gradient magnetic capture of ferrofluids: implications for drug targeting and tumor embolization. Zeitschrift für
Naturforschung, vol. C, pp. 909-911, 2001. Babincova, M., Babinec, P. A controlled drug delivery system based on degradable magnetic polymers. Pharmazie, vol. 51, pp. 515-516, 1996. Bakker, D.P., et al. Comparison of Velocity Profiles for Different Flow Chamber Designs Used in Studies of Microbial Adhesion to Surfaces. App. Env. Microbiology, vol. 69(10), pp. 6280-6287, 2003. Bell GB, Marino A.A., Chesson A.L., Struve F.A. Human sensitivity to weak magnetic fields. The Lancet, 338: 1521-1522 (1991). Buemi, M., et al. Cell Proliferation/Cell Death Balance in Renal Cell Cultures after Exposure to a Static Magnetic Field. Nephron, vol. 86, pp. 269-273, 2001. Chen, et al. Internal magnetic device to enhance drug therapy. U.S. Patent 5921244, 1999. Consigny, P.M., Silverberg, D.A., Vitali, N.J. Use of Endothelial Cells Containing Superparamagnetic Microspheres to Improve Endothelial Cell Delivery to Arterial Surfaces after Angioplasty. J. Vasc. Interv. Rad., vol. 10(2), pp. 155-163, 1999. Dhanikula, A.B., Panchagnula, R. Localized paclitaxel delivery. Int. J. Pharmaceutics, vol. 183, pp. 85-100, 1999. Duch, M. Electrodeposited Co-Ni alloys for MEMS. J. Micromech.Microeng., vol 12, pp. 400-405, 2002. Flanders, P.J. An alternating-gradient magnetometer. J. Appl. Phys., vol. 63(8), pp. 3940-3945, 1988. Flores, G.A. In-vitro blockage of a simulated vascular system using magnetorheological fluids as a cancer therapy. Eur. Cells and Mater., vol.3, pp. 9-11, 2002. Forbes, Z.G., Yellen, B.B., Barbee, K.A., Friedman, G. An Approach to Targeted Drug Delivery Based on Uniform Magnetic Fields. IEEE Trans.s on Magn., vol. 39, pp. 3372-3377, 2003.
137
Frank, J.A., et al. Methods for magnetically labeling stem and other cells for detection by in vivo magnetic resonance imaging. Cytotherapy., vol. 6(6), pp. 621-5, 2004. Friedlaender, F.J. Particle Motion Near and Capture on Single Spheres in HGMS. IEEE
Trans. Magn., vol. Mag-17, no. 6, pp.2801-2803. Friedlaender, F.J. Particle Buildup on Single Spheres in HGMS. IEEE Trans. Magn., vol. Mag-17, no. 6, pp.2804-2806. Gallo, J.M, Häfeli, U. Correspondence re: A.S. Lübbe et al., Preclinical and clinical experiences with magnetic drug targeting. Cancer Res., vol. 57, pp. 3063-3064, 1997. Garas, S.M. Overview of therapies for prevention of restenosis after coronary interventions. Pharmacology and Therapeutics, vol. 92, pp. 165-78, 2001. Garibaldi, et al. Magnetic vascular defect treatment system. U.S. Patent 6315709, 2001. Gershlick, A.H. et al. Treating atherosclerosis: local drug delivery from laboratory studies to clinical trials. Atherosclerosis, vol. 160: 259-71, 2002. Goodwin, S.C., et al. Single-dose toxicity study of hepatic intra-arterial infusion of doxorubicin coupled to a novel magnetically targeted drug carrier. Toxicological
Sciences, vol. 60, pp. 177-183, 2001. Hanzlik, M., et al. Superparamagnetic magnetite in the upper beak tissue of homing pigeons. BioMetals, vol. 13, pp. 325-331, 2001. Hehrlein, C. et al. Drug-eluting stent: the ‘magic bullet’ for prevention of restenosis? Basic Res. Cardiol., vol. 97, pp. 417-23, 2002. Hilger, I., et al.. Evaluation of temperature increase with different amounts of magnetite in liver tissue samples. Invest. Radiol., vol. 32, pp. 705-712, 1997. Iino, M. Effects of a Homogenous Magnetic Field on Erythrocyte Sedimentation and Aggregation. Bioelectromagnetics, vol. 18, pp. 215-222, 1997. Illum, L., Church et al. Development of systems for targeting the regional lymph nodes for diagnostic imaging: In vivo behavior of colloidal PEG-coated magnetite nanospheres in the rat following interstitital administration. Pharmaceutical Research, vol. 18, pp. 640-645, 2001. Kato, T. Encapsulated drugs in targeted cancer therapy. In Bruck SD (Ed.). Controlled
drug delivery. CRC Press, Boca Raton, FL, pp. 190-240, 1983.
138
Kirschvink JL, Kobayashi-Kirschvink A, Diaz-Ricci JC, and Kirschvink SJ. Magnetite in human tissues: A mechanism for the biological effects of weak ELF magnetic fields. Bioelectromagnetics, Suppl. 1: 101-113 (1992). Kirschvink JL, Kobayashi-Kirschvink A, and Woodford BJ. Magnetite biomineralization in the human brain. Proc. Natl. Acad. Sci. USA, 89: 7683-7687 (1992). Liggins, R.T., Burt, H.M. Paclitaxel loaded poly(L-lactic acid) microspheres: properties of microspheres made with low molecular weight polymers. Int. J. Pharmaceutics, vol. 222, pp. 19-33, 2001. Lonnemark, M., et al. Effect of superparamagnetic particles as oral contrast medium at magnetic resonance imaging. A phase I clinical study. Acta Radiol., vol. 30(2), pp. 193-196, 1989. Loweinheim, F.A. Electroplating. New York: McGraw-Hill, 1978. Lübbe, A. S., Alexiou, C., Bergemann, C. Clinical applications of magnetic drug targeting. J. Surg. Res., vol. 95, pp. 200-206, 2001. Mertl, M. Magnetic Cells: Stuff or Legend? Science, vol. 283(5403), pp. 775, 1999. Messer, R.L., et al. Effect of vascular stent alloys on expression of cellular adhesion molecules by endothelial cells. J Long Term Eff Med Implants. vol. 15(1), pp. 39-47, 2005. Minamimura, T., et al. Tumor regression by inductive hyperthermia combined with hepatic embolization using dextran magnetite-incorporated microspheres in rats. Int. J.
Oncology, vol. 16, pp. 1153-1158, 2000. Mitsumori, M. et al. Targeted hyperthermia using dextran magnetite complex: A new treatment modality for liver tumors. Hepato-Gastroenterology, vol. 43, pp. 1431-1437, 1996. Moore, L.R., et al. The use of magnetite-doped polymeric microspheres in calibrating cell tracking velocimetry. J. Biochem. Biophys Methods, vol. 44, pp. 115-130, 2000. Mossbach, K., Schroder, U. Preparation and characterization of magnetic polymers for targeting of drugs. FEBS Letters, vol. 102, pp. 112-116, 1979. Myung, N.V., et al. Electrodeposited Hard Magnetic Thin Films for MEMS Applications. Proc. Electrochem. Soc., PV, pp. 2000-2029, 2000. Nakamura T, et al. Magneto-medicine: Biological aspects of ferromagnetic fine particles. J. Appl. Physiol., vol. 42, pp. 1320-1324, 1971.
139
Ovadia, H., et al. Magnetic microspheres as drug carriers: Factors influencing localization at different anatomical sites in the rats. Isr. J. Med. Sci, vol. 19, pp. 631-637, 1983. Päuser, S., et al. Liposome-encapsulated superparamagnetic iron oxide particles as markers in an MRI-guided search for tumor-specific drug carriers. Anti-Cancer Drug
Design, vol. 12, pp. 125-135, 1997. Plavins, J., Lauva, M. Study of colloidal magnetite-binding erythrocytes: Prospects for cell separation. J. Magn. Mag. Mat., vol. 122, pp. 349-353, 1993. Regar, E., et al. “Stent development and local drug delivery,” Br. Med.Bull.., vol. 59, pp. 227-248, 2001. Rudge, S., et al. Adsorption and desorption of chemotherapeutic drugs from a magnetically targeted carrier (MTC),” J Controlled Release, vol. 74, pp. 335-340, 2001. Ruuge, E.K., Rusetski, A.N. Magnetic fluids as drug carriers: Targeted transport of drugs by a magnetic field. J. Magn. Mag. Mat., vol. 122, pp. 335-339, 1993. Sakhnini, L., Khuzaie, R. Magnetic behavior of human erythrocytes at different hemoglobin states. Eur. Biophys. J., vol. 30, pp. 467-470, 2001. Schenck, J.F. Safety of Strong, Static Magnetic Fields. J. Mag Res. Imaging, vol. 12, pp. 2-19, 2000. Schewe, H., Takayasu, M., Friendlaender, F.J. Observation of Particle Trajectories in an HGMS Single-Wire System. IEEE Trans. Magn., vol .Mag-16, no.1, pp. 149-154, 1980. Schwartz, R.S., et al. Drug-Eluting Stents in Preclinical Studies: Recommended Evaluation From a Consensus Group. Circulation, vol.. 106, pp. 1867-73, 2002. Segre, G., Silberberg, A. Behaviour of macroscopic rigid spheres in Poiseuille flow Part 1 & 2. J. Fluid Mech., vol. 14, pp 115-157, 1962. Senyei, A., Widder, K., Czerlinski, G. Magnetic guidance of drug-carrying microspheres. J. Appl. Physiol., vol. 49, pp. 3578-3583, 1978. Singla, A.K., Garg, A., Aggarwal, D. Paclitaxel and its formulations. Int. J.
Pharmaceutics, vol. 235, pp. 179-192, 2002. Sousa, J.E., et al. Use of Rapamycin-Impregnated Stents in Coronary Arteries. Transplantation Proceedings, vol. 35(Suppl. 3A), pp. 165-170S, 2003.
140
Tiefenauer, L.X., et al. In vivo evaluation of magnetite nanoparticles for use as a tumor contrast agent in MRI. Mag. Res. Imag., vol. 14, pp. 391-402, 1996. Voltairas, P.A., Fotiadis, D.I., Michalis, L.K. Hydrodynamics of magnetic drug targeting. J. Biomech., vol. 35(6), pp. 813-821, 2002. Vyas, S. P., Singh, A., Sihorkar, V. Ligand-receptor-mediated drug delivery: an emerging paradigm in cellular drug targeting. Crit. Rev. Ther.Drug Carrier Syst., vol. 18(1), pp. 1-76, 2001. Wang, J., et al. Characterization of the initial burst release of a model peptide from poly(D, L,-lactide-co-glycolide) microspheres. J. Cont. Rel., vol. 82, pp. 289-307, 2002. Wang, N., Butler, J.P., Ingber, D.E. Mechanotransduction across the cell surface and through the cytoskeleton. Science, vol. 260(5111), pp. 1124-1127, 1993. Watarai, H., Namba, M. Capillary magnetophoresis of human blood cells and their magnetphoretic trapping in a flow system. J. Chromatography A, vol. 961, pp. 3-8, 2002. Widder, K.J., Senyei, A.E., Ranney, D.F. Magnetically responsive microspheres and other carriers for the biophysical targeting of antitumor agents. Advances in
Pharmacology and Chemotherapy, vol.16, pp. 213-271, 1979. Widder, et al. Method of delivering a therapeutic agent to a target capillary bed. U.S. Patent 4345588, 1982. Xia, Y., Whitesides, G.M. Soft Lithography. Annu. Rev. Mater. Sci., vol. 28, pp. 153-184, 1998. Yanase, M. Antitumor immunity induction by intracellular hyperthermia using magnetite cationic liposomes. Jap J. Canc. Res., vol. 89, pp. 775-782, 1998. Yellen, B.B., Forbes, Z.G., Halverson, D.S., Fridman, G., Barbee, K.A., Chorny, M., Levy, R., Friedman, G. Targeted Drug Delivery to Magnetic Implants for Therapeutic Applications. (to be published in Journal of Magnetism and Magnetic Materials May 2005). Zborowski, M., et al. Red Blood Cell Magnetophoresis. Biophysical Journal, vol. 24, pp. 2683-2645, 2003. Zhang, Y., Kohler, N., Zhang, M. Surface modification of superparamagnetic magnetite nanoparticles and their intracellular uptake. Biomaterials, vol. 23, pp. 1553-1561, 2002.
141
VITA
Zachary G. Forbes
Education PhD (pending) Biomedical Engineering, Drexel University, Philadelphia, PA, 2001-2005
M.S. Biomedical Engineering, Drexel University, Philadelphia, PA 2001-2003
B.S.E. Biomedical Engineering, Duke University, Durham, NC, 1997-2000
Professional Experience Drexel University, School of Biomedical Engineering, Cellular Biomechanics Laboratory
9/2001 - Present Graduate Research Assistant to Dr. Kenneth A. Barbee, Ph.D.
Duke University, Department of Biomedical Engineering, Topical Microbicide Development
12/1999 – 12/2000 Undergraduate Research Assistant to Dr. David F. Katz, Ph.D.
Awards 1. Drexel University GRID Fellow – “Magnetizable Implants for Targeted Drug Delivery” (2005) 2. Finalist – 1st Annual International Nanotechnology Business Idea Competition – Case Western
University and the Cleveland Clinic (2004)
3. First Place Paper and Presentation: “Locally targeted drug delivery to magnetic stents.” IEEE-
EMBS International Summer School & Symposium on Medical Devices and Biosensors, Hong
Kong, China (2004)
4. Winner of Phase I&II Drexel Baiada Center Business Plan Competition (2003)
5. Drexel University Calhoun Fellow (2001-2004)
Feature Articles “Stent and Deliver,” MIT’s Technology Review, pg. 18, August 31, 2004.
“Magnetic BioSystems,” In Vivo, p 17, November, 2003.
Selected Publications
1. Z. G. Forbes, B. B. Yellen, G. Friedman, K. A. Barbee. “An approach to targeted drug delivery
based on uniform magnetic fields.” IEEE Transactions on Magnetics, vol. 39(5), pp. 3372-3377
(2003).
2. B. B. Yellen, Z. G. Forbes, D. S. Halverson, G, Fridman, K. A. Barbee, M. Chorny, R. Levy, G.
Friedman, “Targeted Drug Delivery to Magnetic Implants for Therapeutic Applications,” Journal
of Magnetism and Magnetic Materials., vol. 293 (1), pp. 647-654 (2005).
