Citation
Preparation, Characterization, and in Vitro Testing of Poly (Lactide-Co-Glycolide) and Dextran Magnetic Microspheres for in Vivo Applications

Material Information

Title:
Preparation, Characterization, and in Vitro Testing of Poly (Lactide-Co-Glycolide) and Dextran Magnetic Microspheres for in Vivo Applications
Creator:
LEAMY, PATRICK J. ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Dextrans ( jstor )
Diameters ( jstor )
Ferrofluids ( jstor )
Iron oxides ( jstor )
Magnetic fields ( jstor )
Magnetism ( jstor )
Magnets ( jstor )
Nanoparticles ( jstor )
Polymers ( jstor )
Solvents ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Patrick J. Leamy. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
7/1/2003
Resource Identifier:
434595866 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

PREPARATION, CHARACTERIZATION, AND IN VITRO TESTING OF POLY (LACTIDE-CO-GLYCOLIDE) AND DEXTRAN MAGNETIC MICROSPHERES FOR IN VIVO APPLICATIONS By PATRICK J. LEAMY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2003

PAGE 2

ACKNOWLEDGMENTS I would like to thank all of the students, faculty and staff of the Materials Science and Engineering department and the Biomedical Engineering program. The people I work with are very willing to share their expertise and facilities with me, and I appreciate their generosity in and out of the lab. More specifically I thank my graduate advisor, Dr. Christopher D. Batich, for his support, guidance, and for allowing me the opportunity to do research in polymeric biomaterials. I would also like to thank my committee members (Dr. Anuj Chauhan, Dr. Eugene P. Goldberg, Dr. Laurie Gower and Dr. Wolfgang Sigmund) for their time and support of this work. The help and encouragement of the Batich research group members were invaluable to me. This group includes Jon Paul Bullivant, Matthew Eadens, Nakato Kibuyaga, Bernd Liesenfeld, Dr. Gilberto Lunardi, Olajompo Maloye, Albina Mikhailova, John Rotella, Taili Thula, Tara Washington and Bradley Willenberg. Special thanks go to Dr. Swadeshmukul Santra for his advice, generosity, and for always making time for me and the other members of the group. I thank Jennifer Wrighton for carrying out the many administrative duties that make the research efforts possible. Additional thanks go to the MSE students who helped me along the way. Jae Young Choi helped me with the SQUID measurements. Brian Cuevas, Dr. Ahmad Hadba, Paul Martin and Joshua Stopek were always willing to offer advice, to help with ii

PAGE 3

lab work, and to lend equipment. In fact I still have the ultrasonic probe in case they were wondering what happened to it. Each summer for the past four years, a different engineering student from The Ecole de Mines in Douai France worked as an intern in our research group. All four of these students helped by working all or part of the summer with me. I thank Xavier Julle, Sophie Henri, Celine Guerville and Sebastien Berthier for their friendship and hard work. The Engineering Research Center was an invaluable resource for this project. I thank Dr. Kevin Powers for his advice and for giving generous access to the labs. Gilbert Brubaker helped me with particle size analysis and BET measurements, and Gary Scheiffele helped with FTIR. The Major Analytical Instrumentation Center (MAIC) provided valuable support to this work. I thank Dr. Luisa Amelia Dempere and Wayne Acre for their assistance. I would like to thank my parents for their love and support. I Thank my brother Michael for his love, friendship, and for encouraging me in my studies. Finally, I thank Karla for her love, understanding and encouragement. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS..................................................................................................ii LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT......................................................................................................................xii CHAPTER 1 INTRODUCTION........................................................................................................1 2 BACKGROUND..........................................................................................................6 Magnetic Materials.......................................................................................................6 Magnetism in Fine Particles.........................................................................................8 Biological Applications for Magnetic Particles..........................................................11 Magnetic Separations..........................................................................................11 Application of Magnetic Particles for Magnetic Resonance Imaging (MRI) Contrast Enhancement.....................................................................................13 Magnetic Particles for Drug Delivery.................................................................15 Polylactide, Polyglycolide and their Copolymers......................................................18 Synthesis of Polylactide, Polyglycolide and their Copolymers by Ring Opening Polymerization..................................................................................18 Microsphere Preparation for PLA, PLGA and PLA-PEG-PLA Tri-Block Polymers..........................................................................................................23 Clearance of PLA, PLGA and PLA-PEG-PLA Particles by the RES Following Intra-Venous administration...........................................................24 3 PREPARATION OF CHLOROFORM BASED FERROFLUIDS............................29 Methods......................................................................................................................30 Preparation...........................................................................................................30 X-ray Diffraction and Particle Size Estimate for Iron Oxide..............................34 SQUID Magnetometry........................................................................................35 Determination of Ferrofluid Product Yield.........................................................35 Size of Particles in Chloroform Ferrofluid..........................................................36 Results.........................................................................................................................36 Conclusions.................................................................................................................44 iv

PAGE 5

4 PREPARATION OF MAGNETIC PLGA MICROSPHERES USING A SOLVENT EVAPORATION/EXTRACTION METHOD........................................45 Methods......................................................................................................................46 Preparation of Oil Phase......................................................................................46 Microsphere Preparation.....................................................................................46 Microsphere Characterization.............................................................................47 Results.........................................................................................................................49 Conclusions.................................................................................................................60 5 DEXTRAN MICROSPHERES..................................................................................61 Methods......................................................................................................................61 Precipitation of Iron Oxide and Preparation of a Stable Ferrofluid....................61 Preparation of Methacrylated Dextran (Dex-MA)..............................................63 Adsorption of Dex-MA onto Iron Oxide in Ferrofluids......................................67 Preparation of Magnetic Dextran Microspheres..................................................68 Characterization of Magnetic Dextran Microspheres..........................................71 Results.........................................................................................................................72 Characterization of Ferrrofluid............................................................................72 Adsorption Studies..............................................................................................74 Preliminary Microsphere Synthesis Studies........................................................74 Microsphere Synthesis Studies Using PEG20K and Varying the Iron Oxide Content.............................................................................................................80 Conclusions.................................................................................................................87 6 IN VITRO MICROSPHERE RETENTION STUDIES.............................................90 Methods......................................................................................................................90 In Vitro Retention Studies...................................................................................90 Retention apparatus......................................................................................90 Retention experiment...................................................................................93 Quantifying % of microspheres retained using spectrophotometer.............94 In Vitro Retention Study Model..........................................................................97 Results and Discussion...............................................................................................99 Conclusions...............................................................................................................101 7 MRI CONTRAST EFFECT FOR MAGNETIC MICROSPHERES.......................103 Methods....................................................................................................................103 Preparation of MRI Standards with Different Concentrations of Microspheres..................................................................................................103 Magnetic Resonance Imaging (MRI) of Magnetic Microsphere Standards......104 Results and Discussion.............................................................................................106 Conclusions...............................................................................................................113 v

PAGE 6

8 OVERALL CONCLUSIONS AND FUTURE WORK...........................................118 Conclusions...............................................................................................................118 Future Work..............................................................................................................119 APPENDIX A CALCULATION OF THEORETICAL AMOUNT OF OLEIC ACID NEEDED FOR MONOLAYER COVERAGE OF IRON OXIDE PARTICLES.....................121 B SAMPLE CALCULATIONS FOR FLOW SPEED AT 50% RETENTION FOR PLGA MICROSPHERES USING EQUATION 6-8 (FOR 19 WEIGHT % IRON OXIDE MICROSPHERES AND MAGNET DISTANCE OF 4 MM)....................122 Calculation of V F (Volume Fraction Iron Oxide).....................................................122 Calculation of V y for 50% Retention........................................................................123 C SAMPLE CALCULATIONS FOR FLOW SPEED AT 50% RETENTION FOR DEXTRAN MICROSPHERES USING EQUATION 6-8 (FOR 35 WEIGHT % IRON OXIDE MICROSPHERES AND MAGNET DISTANCE OF 9 MM).........125 Calculation of V F (Volume Fraction Iron Oxide).....................................................125 Calculation of v y for 50% Retention.........................................................................126 LIST OF REFERENCES.................................................................................................127 BIOGRAPHICAL SKETCH...........................................................................................134 vi

PAGE 7

LIST OF TABLES Table page 2-1. Units for magnetic quantities.....................................................................................6 2-2. Superparamagnetic to single domain (SP-SD) and single domain to multi-domain (SD-MD) transition diameters for iron and selected iron minerals..........................11 3-1. List of reagents for preparation of chloroform based ferrofluids.............................30 3-2. Composition of ferrofluids based on weight of starting materials...........................33 3-3. Iron oxide particle diameter calculations using the Scherrer Formula.....................36 4-1 Reagent list for magnetic PLGA microspheres........................................................45 4-2. Composition of oil phase for microspheres..............................................................47 5-1. List of reagents for magnetic dextran microspheres................................................62 5-2. Composition of starting materials for each microsphere batch................................75 5-3. Summary of results for preliminary microsphere synthesis studies.........................80 5-4. Composition of starting materials for each microsphere group...............................81 5-5. Iron oxide content of dextran microspheres measured with SQUID.......................82 6-1. Flow velocity (v y ) for 50% retention: Experimental results and model...............100 7-1 List of reagents used for MRI sample preparation.................................................103 7-2 Concentration of microspheres and iron oxide in MRI samples............................105 7-3. Relaxation times (T 1 and T 2 ) for dextran and PLGA microspheres.......................117 7-4. Relaxivity values for dextran and PLGA microspheres.........................................117 vii

PAGE 8

LIST OF FIGURES Figure page 2-1. Domain structure response to an external magnetic field..........................................9 2-2. Hysteresis loop: Magnetic induction B vs. applied magnetic field H.......................9 2-3. Suspension crosslinking process..............................................................................16 2-4. Proposed mechanism for coordination polymerization of dilactones......................19 2-5. Polymerization of lactide with polyethylene glycol to form lactide-ethylene glycol-lactide triblock copolymer............................................................................20 2-6. In Vitro Degradation of Polyglycolide.....................................................................22 2-7. In Vitro Degradation of Poly (L-Lactide)................................................................22 2-8. Preparation of Dex-MA by reaction of glycidyl methacrylate with dextran............27 2-9. Crosslinking of dex-MA in water by free radical polymerization...........................28 2-10. H 2 O : PEG : Dex-MA phase diagrams..................................................................28 3-1. Preparation of chloroform ferrofluid........................................................................31 3-2. XRD data for bare iron oxide...................................................................................38 3-3. Yield for iron oxide suspensions in chloroform as a function of surfactant (oleate) concentration............................................................................................................41 3-4. Size of iron oxide agglomerates measured by PCS..................................................42 3-5. Effect of oleate concentration on agglomeration.....................................................42 3-6. Magnetic hysteresis curves for bare iron oxide........................................................43 3-7. Magnetic hysteresis curves for bare iron oxide compared to dried oleate coated product......................................................................................................................43 3-8. Measured oleate/iron oxide ratio vs. theoretical oleate/iron oxide ratio..................44 viii

PAGE 9

4-1. Microsphere preparation process.............................................................................48 4-2. PLGA microsphere size distributions......................................................................51 4-3. Diameter of microspheres with different iron oxide contents..................................52 4-4. Coefficient of variation for particle size distributions.............................................52 4-5. Scanning electron micrographs for plain microspheres...........................................53 4-6. Scanning electron micrographs for 17% iron oxide microspheres...........................54 4-7. Scanning electron micrographs for 20% iron oxide microspheres...........................55 4-8. Scanning electron micrographs for 25% iron oxide microspheres...........................56 4-9. Scanning electron micrographs for 33% iron oxide microspheres...........................57 4-10. Scanning electron micrographs for 50% iron oxide microspheres........................58 4-11. Iron oxide content in microspheres measured with SQUID..................................59 4-12. Loading efficiency of iron oxide for microspheres................................................59 5-1. Apparatus for preparation of dex-MA......................................................................65 5-2. Nuclear magnetic resonance spectra for dextran and dex-MA................................66 5-3. Magnetic separator...................................................................................................68 5-4. Calibration curve for rotation angle vs. [dex-MA]..................................................69 5-5. Magnetic dextran microsphere preparation apparatus..............................................70 5-6. Iron oxide X-ray diffraction pattern.........................................................................73 5-7. Magnetization of bare iron oxide as a function of applied magnetic field...............73 5-8. Adsorption studies for dex-MA onto iron oxide in ferrofluids................................74 5-9. Size distributions for dextran microspheres prepared with PEG 20K and PEG 35K..................................................................................................................77 5-10. Size distribution for dextran microspheres prepared with PEG 10K.....................77 5-11. Scanning electron micrograph for dextran microspheres prepared with 12% iron oxide and PEG10K.........................................................................................78 ix

PAGE 10

5-12. Scanning electron micrograph for dextran microspheres prepared with 12% iron oxide and PEG20K.........................................................................................78 5-13. Scanning electron micrograph for dextran microspheres prepared with 12% iron oxide and PEG35K.........................................................................................79 5-14. Scanning electron micrograph for dextran microspheres prepared with 29% iron oxide and PEG20K.........................................................................................79 5-15. Magnetometry (SQUID) for dextran microspheres prepared with PEG20K compared to bare iron oxide..................................................................................80 5-16. Scanning electron micrographs for 20% iron oxide microspheres........................83 5-17. Scanning electron micrographs for 30% iron oxide microspheres........................84 5-18. Scanning electron micrographs for 50% iron oxide microspheres........................85 5-19. Microsphere size distribution for representative samples......................................86 5-20. Volume mean microsphere diameters calculated from microsphere size distributions............................................................................................................86 5-21. Microsphere yield as a function of iron oxide content..........................................87 5-22. FTIR spectra for crosslinked dex-MA hydrogels..................................................88 5-23. FTIR spectra for 30% iron oxide microspheres.....................................................89 6-1. Retention apparatus with labeled components.........................................................91 6-2. Retention apparatus photograph...............................................................................92 6-3. Syringe pump calibration.........................................................................................92 6-4. Magnet holder..........................................................................................................93 6-5. Magnetic field as a function of distance from the magnet.......................................95 6-6. Magnetic field gradient as a function of distance from the magnet.........................95 6-7. Spectrophotometer calibration curve for PLGA microspheres................................96 6-8. Spectrophotometer calibration curve for dextran microspheres...............................96 6-9. Flow of a microsphere through tubing under a magnetic field................................97 6-10. Retention vs. flow velocity (v y ). Flow velocity plotted on log scale..................102 x

PAGE 11

6-11. Retention vs. flow velocity (v y ). Flow velocity plotted on linear scale..............102 7-1. Gels containing dextran microspheres...................................................................105 7-2. Magnetic Resonance T 1 images for microsphere standards...................................107 7-3. Magnetic Resonance T 1 images for control gels....................................................108 7-4. Magnetic resonance T 2 images for microsphere standards....................................109 7-5. Magnetic resonance T 2 images for control gels.....................................................110 7-6. Comparison of dextran microsphere standard T 1 and T 2 images...........................111 7-7. Comparison of PLGA microsphere standard T 1 and T 2 images.............................111 7-8. MRI signal intensity in the Z direction as a function of T R for dextran microspheres...........................................................................................................114 7-9. MRI signal intensity in the Z direction as a function of T R for PLGA microspheres...........................................................................................................114 7-10. MRI signal intensity in the XY plane as a function of T E for dextran microspheres........................................................................................................115 7-11. MRI signal intensity in the XY plane as a function of T E for PLGA microspheres........................................................................................................115 7-12. Reciprocal relaxation time (1/T 1 ) as a function of iron concentration.................116 7-13. Reciprocal relaxation time (1/T 2 ) as a function of iron concentration.................116 xi

PAGE 12

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PREPARATION, CHARACTERIZATION, AND IN VITRO TESTING OF POLY (LACTIDE-CO-GLYCOLIDE) AND DEXTRAN MAGNETIC MICROSPHERES FOR IN VIVO APPLICATIONS By Patrick Leamy May 2003 Chair: Christopher D. Batich Major Department: Materials Science and Engineering Many research groups are investigating degradable magnetic particles for magnetic resonance imaging (MRI) contrast agents and as carriers for magnetic drug guidance. These particles are composite materials with a degradable polymer matrix and iron oxide nanoparticles for magnetic properties. The degradable polymer matrix acts to provide colloidal stability and, for drug delivery applications, provides a reservoir for the storage and release of drugs. Natural polymers, like albumin and dextran, which degrade by the action of enzymes; have been used for the polymer matrix. Iron oxide nanoparticles are used for magnetic properties since they can be digested in vivo and have low toxicities. Polylactic acid (PLA) and its copolymers with polyglycolic acid (PLGA) are versatile polymers that degrade by simple hydrolysis without the aid of enzymes. Microspheres are easily formed using the solvent extraction/evaporation method and a wide range of drugs can be encapsulated in them. Magnetic PLGA microspheres suitable for iv applications were synthesized for the first time in this dissertation. This was xii

PAGE 13

accomplished by coating iron oxide nanoparticles with oleic acid to make them dispersible in the organic solvents used in the extraction/evaporation microsphere preparation method. In addition to the magnetic PLGA microspheres, a novel all-aqueous method for preparing crosslinked dextran magnetic microspheres was developed in this dissertation. This method uses free radical polymerization for crosslinking and does not require the use of flammable and harmful solvents. For efficient MRI contrast and magnetic drug guidance, maximized iron oxide content of microspheres is desirable. The two different microsphere preparation methods were optimized for iron oxide content. The effect of iron oxide content on microsphere size and morphology was studied. In addition, an in vitro circulation model was used to evaluate the ability of magnetic microspheres to be guided at physiologic blood flow velocities. The MRI contrast effect was studied as a function of microsphere concentration. xiii

PAGE 14

CHAPTER 1 INTRODUCTION Magnetic particles are increasingly used as carriers in a wide variety of applications. For example, they are routinely used for sorting cells and biomolecules. For this application, selective binding molecules to the cell or biomolecule of interest are coupled to the surface of magnetic particles. The cells or biomolecules adhere to the magnetic particles and are separated from solutions using a magnetic field. Degradable magnetic particles are being investigated for in vivo applications such as MRI contrast agents and magnetically guided drug delivery. In MRI applications, magnetic particles shorten the relaxation times (T 1 and T 2 ), thereby acting as negative contrast agents and darkening the MRI image. Magnetic particles create inhomogeneities in the magnetic field that accelerate spin-spin relaxation and are therefore particularly suited for contrast agents in T 2 weighted imaging. The magnetic particles are ordinarily injected intravenously (iv) and eventually are taken up by the RES system of the lymph nodes, liver and spleen where they are digested. Smaller particles remain in the circulation longer and are used for imaging the circulatory system. Larger particles (on the order of 1 m) are used for imaging the spleen and liver since these organs take them up within minutes of iv administration. The particles used for MRI contrast agents are usually iron oxide nanoparticles or agglomerates of nanoparticles because they have large surface and thus are readily digested in vivo. The nanoparticles are typically coated with natural polymers like dextran to give them colloidal stability. 1

PAGE 15

2 Another application for degradable magnetic carriers is guided drug delivery. For this application, drugs are entrapped in or bound to magnetic microspheres or nanoparticles. These carriers are injected iv or intrarterialy (ia); and a magnetic field is used to retain them in the tissue or organ of interest. Similar materials are used for magnetic drug delivery and MRI contrast applications. Microspheres are preferable to nanoparticles since the velocity of a magnetic particle in response to a magnetic field is proportional to the particle radius squared [1]. Microspheres are usually crosslinked natural polymers to allow for diffusion controlled drug delivery. They contain iron oxide nanoparticles since they are degradable. In addition, iron oxide nanoparticles less than 25 nm do not retain their magnetizations after removal of a magnetic field. The degradable polymers based on polylactic acid (PLA) and polyglycolic acid (PGA) are attractive biomaterials. They degrade by simple hydrolysis without the need for enzymes. Copolymers of PLA and PGA with each other (PLGA), and with other monomers results in polymers with a wide range of physical properties and degradation rates. Non-magnetic microspheres and nanoparticles are easily prepared. A wide variety of lipophilic drugs, hydrophilic drugs, and proteins can be encapsulated and released from them. Microspheres with long iv circulation times are prepared using polylactic acid-polyethylene glycol copolymers, or by coating PLA or PLGA microspheres with copolymers containing polyethylene glycol blocks. Microspheres and nanoparticles based on PLA and PGA are created by two methods. One is the solvent extraction/evaporation method. In this method, the polymer is dissolved in an organic solvent like chloroform or dichloromethane to form an oil phase. This solution is then dispersed in water that contains stabilizers such as polyvinyl

PAGE 16

3 alcohol or polyvinyl pyrrolidone. Stirring, or other means of agitation, is used to reduce the size of the oil phase droplets. The solvent is then removed from the oil phase by controlled extraction into the aqueous phase and in many cases evaporation from the aqueous phase. The second method for preparing particles based on PLA or PGA is the precipitation method. In this case, the polymer is dissolved in a water-soluble organic solvent such as acetone or acetonitrile. This solution is added to water. Because of the high miscibility of solvents, the organic solvent is quickly extracted into the water, and the polymer precipitates. Because of the rapid extraction and precipitation, this process ordinarily results in nanoparticles on the order of 100 nm. Although PLA and PLGA microspheres and nanospheres are widely used in the scientific literature, there are no small magnetic microspheres or nanospheres based on PLA or PGA. This is because of the difficulty in incorporating and suspending the hydrophilic iron oxide nanoparticles into the organic solvents used for microsphere preparation. In this dissertation PLGA microspheres containing iron oxide nanoparticles were prepared by coating the iron oxide nanoparticles with a surfactant that suspended the iron oxide in chloroform. A novel all-aqueous method for producing crosslinked dextran microspheres was recently developed by Stenekes et al. [2]. This method involved modifying dextran with methacrylate groups for crosslinking functionality. The methacrylate-modified dextran (dex-MA) was then dissolved in water and added to an aqueous PEG solution. The aqueous dex-MA phase is immiscible in the aqueous PEG phase. Dex-MA droplets were formed by vortexing. Crosslinking was accomplished by free radical initiated

PAGE 17

4 polymerization of the methacrylate groups. Proteins were loaded into the dextran microspheres by dissolving them along with the dex-MA. Proteins were released by enzymatic degradation of the microspheres. Microspheres were too large for iv applications, but smaller microspheres probably can be produced using stronger agitation. This microsphere preparation technique is attractive since it uses gentle chemistry for crosslinking that is less likely to alter the chemical structure of loaded drugs, or to alter the chemical structure or conformation of proteins. We modified this method to produce magnetic dextran microspheres of a suitable size for iv applications. For the aforementioned in vivo applications (and others), the magnetic particles must be degradable, must elicit minimal immune response, and must have high magnetic moments. The magnetic moment is a function of the magnetic material content. The aim of this dissertation was to produce magnetic microspheres from degradable materials with a controlled size and with optimized iron oxide contents. Magnetic PLGA microspheres were investigated since the polylactide and glycolide family of polymers can be tailored to deliver a wide range of drugs, can be formulated as long circulating microspheres, and degrade without the aid of enzymes. The microspheres were prepared by incorporating iron oxide that was made hydrophobic by coating with oleic acid. The all-aqueous method for producing dextran microspheres was modified by incorporating iron oxide nanoparticles to produce magnetic microspheres. This method was chosen since it has gentle chemistry and the microspheres can be washed without using harmful organic solvents. PLGA and dextran microspheres were optimized for iron oxide content and were characterized using SQUID magnetometry, scanning electron microscopy, laser particle

PAGE 18

5 size analysis, and other techniques. An in vitro flow system with an external magnetic field was used to test the ability of the microspheres to be retained by magnetic fields with physiologic flow speeds. The effectiveness of the microspheres as MRI contrast agents was also determined experimentally.

PAGE 19

CHAPTER 2 BACKGROUND Magnetic Materials In an external magnetic field (H), the internal magnetic field of a magnetic material increases because of the alignment of electron spins in the direction of the magnetic field. This increase in the internal magnetic field is the magnetization (M). Magnetic induction (B) measures the increase in magnetic field because of magnetization and is given in SI units (Equation 2-1) and cgs units (Equation 2-2). he permeability of free space ( 0 ) is equal to 4 x 10 -7 tesla-meters per ampere; 0 has no physical meaning and is only needed for SI units. Table 2-1 shows the units for each Equation. SI units: B = 0 H + 0 M = 0 (H + M) (2-1) cgs units: B = H + 4M (2-2) Table 2-1. Units for magnetic quantities Magnetic quantity SI units Cgs units B (magnetic induction) tesla (T) gauss (G) H (applied field) ampere/meter (A/m) oersted (Oe) M (magnetization) ampere/meter (A/m) gauss (G) or emu/cm 3 Numerical conversion factors: 1 A/m = 4 x 10 -3 Oe 1 T = 1 x 10 4 G = 10 4 /4 emu/cm 3 0 = 4 x 10-7 Tm/A Below the Curie temperature, ferromagentic and ferrimagnetic materials are divided into magnetic domains. The electron spins in each domain are aligned giving each domain a net magnetic moment or magnetization. When the material is 6

PAGE 20

7 demagnetized, the magnetic moments of the domains are randomly oriented. The magnet moments then cancel each other, and the material has no net magnetization on the macroscopic scale. When an external magnetic field is applied to a demagnetized magnetic material, the domains that are oriented in the direction of the field grow at the expense of the less favorably oriented domains. As the field becomes stronger, the domains begin to rotate and become aligned with the field. Figure 2-1 shows the magnetization (M) as a function of the applied magnetic field (H) and the response of magnetic domains. The plot of magnetic induction (B) vs. H would look nearly identical to M vs. H since the magnetization dominates the right hand side of Equation 2-1 for ferro and ferrimagnetic materials. Figure 2-2 plots of B vs. H for a de-magnetized ferro or ferrimagnetic material that is magnetized in the positive direction, then magnetized in the negative direction, and finally re-magnetized in the positive direction. The material is first magnetized to its saturation induction B s (point A). As the external field is decreased from B s , the curve is not retraced and at 0 applied field (point C) there is still an induction known as the remnant induction (B r ). This remnant induction is responsible for the magnetization in permanent magnets. As the field is increased in the negative direction from point C, the induction reaches 0 at point D and the material is demagnetized. The field necessary to demagnetize is known as the coercivity or coercive force H c . This is indicated as -H c in the Figure 2-2. As the negative applied field is increased further, the negative saturation induction (-B s ) is reached at point E. If the magnetization is then increased until point A is reached again, the negative remnant induction (-B r ) will be reached at point F and the coercivity (H c ) will be reached at point G. Further applications of the reverse and

PAGE 21

8 forward applied magnetic fields will retrace the path of ABDEFG. This graph is commonly known as a hysteresis loop. Magnetometers are used to generate experimental B vs. H data. We used a very sensitive magnetometer known as a Superconducting Quantum Interference Device (SQUID) for this study. The domain structure in a bulk magnetic material is determined primarily by two different energies, the exchange energy and the magnetostatic energy; with the most stable domain structure attained by minimizing the sum of these energies. The exchange energy is the internal potential energy because of the alignment of the magnetic dipoles. Exchange energy is minimized when all of the magnetic dipoles are aligned, (i.e. when all the electron spins are aligned). The magnetostatic energy is the external potential energy because of the magnetic field caused by alignment of the dipoles within domains. This energy increases with domain size since the external magnetic field increases with domain size. Therefore the domain size is determined by the interplay of the exchange and magnetostatic energies[3]. Magnetism in Fine Particles As the size of a magnetic material is reduced, a point is reached where the material or particle volume equals the domain size. At this point the particle is a single-domain particle. The particle size for a single-domain particle of a given material can be determined from magnetometry experiments (since B r and H c are maximized at the onset of single domain behavior). The B r and H c are reduced in multi-domain particles by the reversal of domain wall growth.

PAGE 22

9 M Rotation of domain mo m e n t s Growth of favorable domains Rotation of domain Continued Do m a in Rando m H Figure 2-1. Domain structure response to an external magnetic field. B E A G F D C -Hc Br Bs H Figure 2-2. Hysteresis loop: Magnetic induction B vs. applied magnetic field H.

PAGE 23

10 As the size of a single-domain particle is reduced, the domain naturally reduces in size. As the size is reduced further, the thermal energy (kT) overcomes the exchange energy and the particle can no longer sustain its magnetization for appreciable periods of time. Under a magnetic field, however, the particle can become magnetized, but no hysteresis is evident since the particle cannot maintain its magnetization in the absence of a magnetic field. These particles lose their magnetizations following removal of the magnetic field, and then B c and H c equal 0. This phenomenon is known as superparamagnetic behavior because (as with like paramagnetic behavior) there is no magnetic hysteresis. Table 2-2 shows the threshold sizes for superparamagnetism and single-domained particles based on experiments and on theoretical calculations for iron and iron containing minerals. The large difference in the single domain to multi-domain threshold size between hematite and maghemite shows how crystal structure has a dramatic effect on magnetic properties. The onset of superparamagnetism is particle size dependent, but there is an additional inverse temperature dependence since thermal energy is responsible for the phenomenon. The temperature below which a superparamagnetic particle becomes a single-domain particle is known as the blocking temperature. The blocking temperature for a given sample is determined by generating hysteresis curves as a function of temperature. The blocking temperature can be used to estimate particle size for a given material since blocking temperature has a negative dependence on particle size.

PAGE 24

11 Table 2-2. Superparamagnetic to single domain (SP-SD) and single domain to multi-domain (SD-MD) transition diameters for iron and selected iron minerals from the literature. Values are for particles at 20 C. Table reproduced from Dunlop 1981 [4]. Material SP-SD transition diameter (m) SD-MD transition diameter (m) Iron <0.008 *0.026 0.023 *0.017 Magnetite (Fe 3 O 4 ) 0.025-0.030 0.05-0.06 *0.08 Maghemite (-Fe 2 O 3 ) *0.06 Hematite (-Fe 2 O 3 ) 0.025-0.030 15 Pyrrhotite (Fe 7 S 8 ) 1.6 *theoretical value; otherwise values determined experimentally Biological Applications for Magnetic Particles Magnetic Separations Magnetic particles are commercially available for use in cell and biomolecule separations. Polymer microspheres that contain magnetite or maghemite nanoparticles are among the forms of magnetic particles used for these separations [5-9]. The most common polymer used for magnetic microspheres is polystyrene (available from Bangs labs, Fishers IN). Microspheres come surface modified with streptavidin for easy coupling with biotin. They are also available with proteins already coupled to the surface. Magnetic polymer microspheres are ideally monodisperse in size and are on the order of 1 to 5 m in diameter. One method for producing magnetic polymer microspheres is emulsion polymerization in the presence of surfactant stabilized magnetite nanoparticles [9]. Proper choice of surfactants entraps the magnetite in the polymer latex particles as they form. Another method for producing magnetic polymer microspheres (developed by Ugelstad), is to precipitate magnetite inside polymer microspheres after they are formed

PAGE 25

12 by emulsion or dispersion polymerizations [10]. The process involves producing hydrogel or nanoporous microspheres that contain functional groups like amines or carboxyl groups (which can chelate Fe 2+ and Fe 3+ ions). The microspheres are soaked in Fe 2+ and Fe 3+ solutions, which are retained in the microspheres by the chelating groups. A basic solution such as ammonia is then added to the suspension of iron containing microspheres. The base diffuses into the microsphere and reacts with the iron ions to precipitate magnetite. Magnetite is precipitated inside the microspheres since the chelating groups retain the iron ions and prevent them from diffusing into the bulk solution. Polymer microspheres with diameters of 0.5 to 5 m are produced by emulsion or dispersion polymerization. Larger microspheres up to 50 m can be prepared by a two-step polymerization. The first step is to create seed particles by emulsion polymerization. The seed particles are then washed and suspended in an aqueous solution. Monomer is then added to swell the particles and a second polymerization is conducted to produce the larger microspheres. In addition to magnetic microspheres, Ferrofluids [11-14] are used for cell and biomolecule separations. Ferrofluids are single crystals or small agglomerates of magnetite or maghemite nanoparticles that are stabilized by coating with water-soluble natural polymers like proteins or polysaccharides. They are produced by two methods. One is precipitation of magnetite by addition of a base to a ferrous and ferric salt solution, followed by adsorption of the water-soluble polymer. The other method is precipitation of magnetite in the presence of the polymer by adding base to a solution containing the ferrous and ferric salts and the polymer. In the former method the exact amount of polymer needed to coat the particles can be used and purification is not

PAGE 26

13 needed. The latter method requires excess polymer. Large agglomerates can form requiring purification by centrifugation and often by dialysis or HPLC. Although this method is more complex, the polymer adsorption is less reversible (i.e. the polymer is more firmly anchored to the particles). Ferrofluids are on the order of 10 to 250 nm in diameter. Receptors to cells or biomolecules are coupled to the ferrofluids so that they will bind to the biomolecule or cell of interest [15]. The iron oxide particles in both magnetic microspheres and ferrofluids are superparamagnetic and therefore do not sustain permanent magnetizations in the absence of a magnetic field. This is important to avoid aggregation by attraction of north and south poles in the absence of a magnetic field. Ferrofluids contain considerably smaller particles than magnetic microspheres and therefore have a higher surface area for a given concentration. They are therefore able to bind more biomolecules or cells. On the other hand, they require higher magnetic fields for separation since their higher surface area to volume ratio slows their movement through fluids. Application of Magnetic Particles for Magnetic Resonance Imaging (MRI) Contrast Enhancement Superparamagnetic magnetite nanoparticles are known to be effective MRI contrast agents in animals and humans. In MRI applications, magnetic particles shorten the relaxation times (T 1 and T 2 ), thereby acting as negative contrast agents and darkening the MRI image. Magnetic particles create inhomogeneities in the magnetic field that accelerate spin-spin relaxation and are therefore particularly suited for contrast agents in T 2 weighted imaging. Magnetite-based ferrofluids used for MRI contrast applications are produced by the same methods as ferrofluids for magnetic separations. They are most commonly

PAGE 27

14 stabilized using the polysaccharide dextran. The mean diameter of these agents varies from 10 nm to 1 m. [16-19] After iv injection, the nanoparticles eventually are collected in the liver and spleen. Because of their high surface areas, the nanoparticles are digested in the spleen and liver in a week or less. The toxicity is very low as evidenced by a 200 fold margin of safety between effective dose for liver contrast and LD 50 in mice for starch-coated magnetite particles delivered iv [20]. These particles’ most common application is imaging the liver and spleen. Magnetic particles injected iv are quickly taken up by the reticulo-endothelial system (RES) of the liver and spleen. Larger agglomerates (on the order of 0.25 to 1 m) are used because they are taken up faster than smaller particles [20]. Tumors in the liver have greatly reduced RES activity, and thus take up fewer magnetic particles. Therefore iron oxide nanoparticles have been used to identify liver tumors [21]. Similarly iron oxide nanoparticles delivered iv can detect metastases in the lymph nodes. Iron oxide nanoparticles are taken up by macrophages in healthy lymph nodes. Lymph nodes that are partially or completely colonized with tumor cells have a diminished level of macrophage activity and thus reduced uptake of nanoparticles [22-24]. Nanoparticles that are small enough to penetrate the capillary fenestra and interendothelial junctions can be used for receptor-directed MR imaging. Tumor imaging has been accomplished by coupling ferrofluids with monoclonal antibodies for carcinoembrionic antigen [25] and epidermal growth factors [26]. Iron oxide nanoparticles have also been used to enhance MR angiography for blood vessels like the aorta, inferior vena cava and portal vein [27].

PAGE 28

15 Finally iron oxide nanoparticles have been used for MR imaging of the coronary arteries [28], pulmonary vasculature [29], and lung and abdominal hemorrhages [30]. Magnetic Particles for Drug Delivery Another application for magnetic particles is magnetic drug delivery. In this application, drugs are chemically or physically bound to magnetic particles, injected iv or ia, and a magnetic field is used to retain the magnetic particles in the target tissue or organ. Fields are created using rare earth permanent magnets or electromagnets. Ferrofluids [31-34], crosslinked protein or polysaccharide microspheres [1, 35-44] and carbonyl iron particles [45] have been used. Ferrofluids are prepared by the methods discussed in the section on magnetic separation of biomolecules. Crosslinked protein or polysaccharide magnetic microspheres are most often formed by the suspension crosslinking method [1, 35-44]. This involves: preparing aqueous solutions of the polysaccharide or protein, adding superparamagnetic iron oxide, emulsifying the aqueous solution in an oil or organic solvent, followed by crosslinking to form microspheres. Figure 2-3 shows the suspension crosslinking process. Epichlorohydrin and cyanogen bromide are used to crosslink polysaccharides. Aldehydes such as formaldehyde and glutaraldehyde are used for proteins. The iron oxide is added as a 4 to 5 volume percent electrostatically stabilized ferrofluid. One problem with this method is that the iron oxide content is low since the ferrofluid is dilute and is added to a polymer solution. Another problem for drug-loaded microspheres is that crosslinking agents can react with drugs, thereby altering their structure and behavior. These magnetic microspheres have a relatively wide size distribution and average 1 to 2 m in diameter with the largest microspheres in a given distribution

PAGE 29

16 around 4 m. Therefore all of the microspheres can traverse the smallest capillaries (6 to 8 m) in diameter. Microspheres typically have lower iron oxide contents than do ferrofluids, but are more responsive to a magnetic field since the velocity of a magnetic particle in response to a magnetic field is proportional to the radius squared. Equation 2-3, which was developed by Senyei [1], shows this relationship for the cgs unit system. Polymer (aq) + Ferrofluid Oil + Surfactant B A Stirrer Crosslinker Figure 2-3. Suspension crosslinking process. A) Aqueous phase containing polymer and ferrofluid are added to oil phase. B) Small aqueous droplets are formed by agitation and droplets are crosslinked to form hardened microspheres. dxdHVF9r M 2v2x (2-3) Where: M = magnetization of magnetite (G) V F = volume fraction magnetite in particle r = particle radius (cm) dH/dx = magnetic field gradient (Oe/cm) = viscosity of fluid (poise) Arshady [46] developed Equation 2-4 for polymer particle size from suspension polymerization. Since the suspension crosslinking process is analogous to suspension polymerization, microsphere size is controlled by the same factors and Equation 2-4 can be applied.

PAGE 30

17 SmSdVCNDRDkd (2-4) Where d = mean microsphere size; k = parameters such as apparatus design, type of stirrer etc.; D V = diameter of vessel; D S = diameter of stirrer; R = volume ratio of the droplet phase to the suspension medium; N = stirring speed or power of mixing; d = viscosity of droplet phase; m = viscosity of suspension medium; = interfacial tension between the two immiscible phases; C s = stabilizer concentration. Carbonyl iron particles are prepared by milling iron metal with activated carbon. Particle size is a function of the milling process and results in wide particle size distributions. The largest particle in a given size distribution must be less than around 5 m since the smallest capillary beds are on the order of 6 to 8 m in diameter. Carbonyl iron particles have very high volume fractions of iron, and are larger than ferrofluid nanoparticles. Accordingly they respond very strongly to a magnetic field, but drug loading is low since drugs are adsorbed on the surface. In addition these particles are not degradable. The primary challenge to overcome in magnetically targeted drug therapy is generating enough force to retain magnetic particles in deep tissues of the body. Success in retaining magnetic carriers in animal and human experiments has been limited to areas close to the surface. These areas include the ears [34], tails [36, 38, 42, 44], limbs [31, 32] of small animals, or in tumors close to the skin in humans [33]. Targeting of deeper organs in a pig model was successful using carbonyl iron particles [45], but these particles have the aforementioned drawbacks. Improvements in magnetic carriers or in

PAGE 31

18 magnet technology are necessary to make magnetically targeted drug therapy applicable to human treatments. Polylactide, Polyglycolide and their Copolymers Polylactide and polyglycolide are the most widely used synthetic degradable biopolymers. They are popular since they have good mechanical properties and degrade to non-toxic metabolites: glycolic or lactic acid. Polylactide, polyglycolide and copolymers of the two find clinical use in degradable sutures and orthopedic pins and screws. Synthesis of Polylactide, Polyglycolide and their Copolymers by Ring Opening Polymerization Polylactide, polyglycolide, and their copolymers are routinely synthesized from lactide and glycolide cyclic dimers using a metal coordination catalyst and an alcohol initiator such as lauryl alcohol. Polylactide differs from polyglycolide in that R is a methyl group for polylactide and a hydrogen for polyglycolide. Stannous 2-ethyl hexanoate (stannous octoate) is a common catalyst. Figure 2-4 shows the ring opening polymerization mechanism proposed by Kissel [47]. The metal coordination catalyst activates the carbonyl group (1). The alcohol group of the initiator then reacts with the carbonyl group by nucleophilic attack (2) to form an unstable intermediate that is stabilized by ring opening of the ester bond and formation of an alcohol (3). The propagation proceeds by the same mechanism as initiation with the alcohol group of the growing polymer chain initiating ring opening. An important result of this polymerization is that the alcohol initiator is incorporated into the polymer. Polylactide-polyethylene glycol-polylactide triblock copolymers (PLA-PEG-PLA) are easily prepared by polymerizing lactide in the presence of polyethylene glycol using

PAGE 32

19 MOOOORR R'OH MOOORR-OOR' H+ OOOORRR'OH 123 nMOOOORR H+ OOOORRR'OH OOOORROHOOR'OORR MOOORR-OOOOORR'OR n R = CH3 or H B A Figure 2-4. Proposed mechanism for coordination polymerization of dilactones proposed by Kissel [47]. A) Initiation. B) Propagation

PAGE 33

20 Sn-Oct.nOOOOCH3H3C m HOOOOOOCH3OOCH3OCH3OCH3H + m n-l lHOOH ABC Figure 2-5. Polymerization of lactide (A) with polyethylene glycol (B) to form lactide-ethylene glycol-lactide triblock copolymer (C). the same metal catalysts as the polylactides and glycolides (see Figure 2-5) [48-51]. The OH groups at each end of PEG initiate polymerization of polylactide to form the triblock polymer. Polyethylene glycol is commonly referred to as polyethylene oxide (PEO), therefore these materials are also known as PLA-PEO-PLA triblock polymers. These materials exhibit phase segregation of the PLA and PEG blocks. Since the PEG blocks are water miscible and the PLA blocks are hydrophobic, they become physically crosslinked hydrogels when placed in water with equilibrium water content a function of the mole ratio of PEG to PLA [52]. For lactide, the carbon attached to the methyl group is chiral. Since there are two chiral carbons per lactide molecule, lactide exist in three enantiomers: L-lactide where both stereocenters have L configuration, D-lactide where both have D conformation, and DL-lactide where one stereocenter is L and the other D. Isotactic polylactide is synthesized using pure L-lactide or pure D-lactide. Since L-lactic acid is a common

PAGE 34

21 metabolite in the human body, and D-lactic acid is not found in the body; poly (D-lactide) is seldom synthesized. Poly (DL-lactide) is synthesized using DL-lactide, a mixture of D and L-lactide, or a mixture of the three stereoisomers. Poly (DL-lactide) is an amorphous polymer. Because of its isotactic stereochemistry, poly (L-lactide) is 35% crystalline while poly (DL-lactide) is amorphous. Poly (L-Lactide) has a higher modulus and tensile strength than the amorphous poly (DL-lactide). Similarly, the crystalline poly (L-lactide) degrades completely in vivo in 20 months to 5 years depending on polymer molecular weight and location, while poly (DL-lactide) degrades much faster; in 6 to 17 weeks [53]. Copolymers of glycolide and lactide (poly (lactide-co-glycolide)), abbreviated as PLGA, are amorphous and have similar mechanical properties and degradation rates as poly (DL-Lactide). Pure polyglycolide is very strong and stiff, yet degrades at similar rates as the poly (DL-lactides) and PLGA. Polyglycolide is highly crystalline with crystallinities between 35 and 70%. Figures 2-6 and 2-7 show degradation rates for polyglycolide and poly (L-Lactide).[53] Since their structures can be deduced by the direct condensation of lactic and glycolic acid, polylactide, polyglycolide, and poly (lactide-co-glycolide) are often referred to as polylactic acid, polglycolic acid and poly (lactic-co-glycolic acid). Though it is rare, synthesis of polylactic and glycolic acids can be achieved by direct condensation, but this results in low molecular weight polymer (on the order of 2,000 g/mole) with poor mechanical properties. Degradation rates increase with reduced molecular weight. This may be advantageous in some applications.[53]

PAGE 35

22 Figure 2-6. In Vitro Degradation of Polyglycolide: retained tensile strength vs. Time. Reproduced from “Chapter 1. Polyglycolide and Polylactide” D. E. Perrin and J. P. English. in Handbook of Biodegradable Polymers. [53] Figure 2-7. In Vitro Degradation of Poly (L-Lactide): retained tensile strength vs. Time. Reproduced from “Chapter 1. Polyglycolide and Polylactide” D. E. Perrin and J. P. English. in Handbook of Biodegradable Polymers. [53]

PAGE 36

23 Microsphere Preparation for PLA, PLGA and PLA-PEG-PLA Tri-Block Polymers Recent research has focused on the use of PLA, PLGA and PLA-PEG-PLA tri-block polymers as drug delivery matrices since sustained release of drugs can be achieved. Drug delivery matrices include monoliths and microspheres. Because of the relatively complex polymerization of these polymers, the familiar techniques of suspension, emulsion and dispersion polymerization are rarely attempted for production of microspheres. However, one group has achieved a ring opening dispersion polymerization of L-lactide to form poly(L-lactide) [54, 55]. This polymerization was performed in a heptane-dioxane solvent. Specialized surfactants were synthesized and used to stabilize the microspheres as they polymerized. Microspheres are routinely prepared using the solvent extraction/evaporation technique. This technique involves dissolution of polymer and drug in an organic solvent, suspension in aqueous solution (to form an oil in water (o/w) emulsion), and extraction/evaporation to form drug entrapped microspheres or nanospheres. The aqueous solution contains surfactants to stabilize the oil in water emulsion. The most common surfactant is 88% hydrolyzed polyvinyl alcohol (PVA). Chloroform and dichloromethane are widely used as the organic solvents since they dissolve PLGA and PLA and they are slightly soluble in water. The solubility of chloroform and dichloromethane in water are 0.8 and 1.73 mass % respectively at 25 C [56]. There are two different methods for removing the organic solvent in the solvent extraction/evaporation method: the extraction method and the extraction/evaporation method. In the extraction method, excess aqueous solution is added following droplet formation to extract the organic solvent from the oil phase. In the extraction/evaporation method, the emulsion is continuously stirred and sometimes heated to evaporate the

PAGE 37

24 weakly water-soluble organic from the aqueous solution. As the solvent evaporates from the water phase, solvent is extracted from the oil phase into the water phase. This process continues until the solvent is completely removed from the oil phase and hardened microspheres result. Since they are water soluble, proteins and polynucleotides are encapsulated using a water in oil in water ((w/o)/w) double emulsion. The aqueous solution is first dispersed in the organic polymer solution with the use of surfactants and sonication or homogenization. This primary emulsion is then emulsified in another aqueous solution (usually PVA) and the same suspension-solvent extraction process is used. Nanoparticles can be formed using high-energy sonication or homogenization in the o/w and the (w/o)/w technique [48]. A simpler method to form nanoparticles is by precipitation methods [49, 57, 58]. The drug of interest and the polymer are dissolved in a water miscible solvent like acetone or acetonitrile. This solution is then added to water. Nanoparticles form because of rapid diffusion of the solvent into the water and precipitation of the polymer/drug complex. Clearance of PLA, PLGA and PLA-PEG-PLA Particles by the RES Following Intra-Venous administration. The use of PLGA and PLA nanoparticles or small microspheres for iv applications was limited by their fast clearance by the RES. The surface chemistry of the particles plays a large role in the clearance rate. Poloxamer coated PLGA particles and particles made from PLA-PEG-PLA triblock polymers have increased circulation times compared to uncoated PLGA particles. Uncoated PLGA nanoparticles [57, 59, 60] are cleared in a few minutes while PLA-PEG-PLA triblock polymer nanoparticles have reported half lives of 6 hours [60]. Poloxamer coated PLGA nanoparticles have increased half lives of

PAGE 38

25 30 minutes compared to uncoated PLGA nanoparticles, but they are not as effective as the PLA-PEG-PLA triblock polymer nanoparticles at eluding the RES [57, 59]. This is probably because of desorption of the poloxamer from the PLGA surface with time. Dextran is a simple water-soluble polysaccharide manufactured by Leuconostoc mesenteroides and L. dextranicum (Lactobacteriaceae). Its structure is usually represented as a linear polymer, but some branching occurs. The native form of dextran has a high molecular weight near 5 X 10 8 g/mole. Dextran is depolymerized to yield a variety of molecular weights depending on the application. Similar to polyvinyl pyrrolidinone, dextran solutions can be used as a blood plasma extender for mass casualty situations. Dextran of between 50,000 and 100,000 g/mole is used for this application. Like many of the water-soluble polymers, crosslinked dextran can be used as a drug delivery matrix in whole or microsphere form. As mentioned earlier, dextran-coated magnetite (Fe 3 O 4 ) nanoparticles are finding use as Magnetic Resonance Imaging (MRI) contrast agents. The dextran adsorbs on to the particle surfaces and provides a steric barrier to prevent agglomeration of the nanoparticles. A method to crosslink dextran using free radical initiation was developed by Vandijkwolthuis [61]. This method involves less toxic reagents and in many cases causes less damage to encapsulated drugs. Dextran is ordinarily crosslinked using cyanogen bromide or epichlorohydrin catalyzed by a strong base. These reagents are highly toxic and the aggressive crosslinking chemistry can alter drugs for in situ loading applications. The method of Vandijkwolthuis involves modifying dextran with methacrylate groups to form methacrylated dextran (dex-MA). This is accomplished by reacting dextran with glycidyl methacrylate. This reaction is shown in Figure 2-8. Once the dex-MA is

PAGE 39

26 prepared and purified, it can be dissolved in water and crosslinked by free radical polymerization. Vandijkwolthuis used the water-soluble initiator system of KPS and tetramethyl ethylene diamine (TEMED). Figure 2-9 shows a schematic of the dex-MA crosslinking. A novel all-aqueous suspension crosslinking technique was developed for producing protein loaded dextran microspheres using dex-MA. The technique takes advantage of phase separation that occurs between polyethylene glycol (PEG) and dex-MA in aqueous solutions [2]. Dex-MA and PEG aqueous solutions were added together and phase separation into a dex-MA rich and PEG rich phase occurred. A vortexer was then used to form dextran droplets and the free radical initiators were added to crosslink by polymerization of methacrylate groups. Proteins were incorporated into microspheres by adding proteins to the dextran phase. Crosslink density, which affects degradation and protein release rate, was controlled by varying the degree of substitution of the methacrylate groups on dextran. Water content was controlled by the concentration of water in the dextran-PEG-water system. Additionally, the all-aqueous system made washing the microspheres easier since water was the only solvent needed to wash microspheres. In this dissertation, magnetic microspheres were prepared using this method by adsorbing dex-MA onto precipitated iron oxide particles, adding PEG to achieve phase separation and finally crosslinking by free radical initiation. Stenekes et al. [2] produced phase diagrams for the PEG, dex-MA, water system to determine the concentrations at which phase separation occurs and the composition of the phases (see Figure 2-10). The composition of each phase in the two-phase region can be determined using tie lines. Bulk hydrogels were also prepared to determine the effects of

PAGE 40

27 crosslink density on swelling of gels. Crosslink density was defined in terms of degree of substitution (DS) of methacrylate groups onto dextran. For example, DS of 5% means that there are 5 methacrylate groups for every 100 dextran repeat units. The swelling data on bulk hydrogels can be used to predict the swelling of microspheres. Stenekes et al. [2] found that dextran hydrogels with DS greater than 10% crosslinked in water do not swell further following polymerization when placed in a large volume of water. Therefore, microspheres using dextran with DS greater than 10% will retain their size following washing with water and their water contents can be predicted using the phase diagrams. O HO OH HO O HO OH HO O O O B A DMSO DMAP + O O O O HO OH O O HO OH HO O O O OH O O C Figure 2-8. Preparation of Dex-MA (C) by the reaction of glycidyl methacrylate (A) with dextran (B). Ring opening of epoxide with dextran OH groups is catalyzed by dimethyl amino pyridine (DMAP). DMSO is the solvent.

PAGE 41

28 Methacrylate group Dextran TEMED KPS Figure 2-9. Crosslinking of dex-MA in water by free radical polymerization of methacrylate groups. Potassium persulfate (KPS) and tetramethyl ethylene diamine (TEMED) are free radical initiators. Figure 2-10. H2O : PEG : Dex-MA phase diagrams. Left shows 1 and 2 phase regions. Right shows tie lines for polymer concentration in each phase for the 2-d 1Phase 2 Phase phase region. PEG weight average molecular weight = 10,000 g/mole. Dex-MA weight average molecular weight = 40,000 g/mole. Reproducefrom Stenekes et al. [2].

PAGE 42

CHAPTER 3 PREPARATION OF CHLOROFORM BASED FERROFLUIDS PLA, PLGA and other degradable polyester microspheres are normally produced by the solvent extraction/evaporation method. This involves dissolution of the polymer in a solvent such as chloroform or methylene chloride, followed by emulsification in an aqueous phase, and finally extraction of the chlorinated solvent in the aqueous phase. PLA or PLGA microspheres incorporating superparamagnetic iron oxide have not been produced. This may be because of the difficulty in suspending iron oxide, whose surface is hydrophilic, in the hydrophobic chlorinated solvents. The patent and scientific literature show that coating iron oxide with surfactants such as oleic acid enables suspension in various oils and organic solvents. This study shows that coating with oleic acid allows the iron oxide to be suspended in chloroform to yield ferrofluids that are stable for at least 3 months. The method of Robineau and Zins [62] was adapted to produce the ferrofluids. Their method used an oleate/iron oxide ratio of 0.547. This was calculated based on the weights of starting materials assuming pure Fe 3 O 4 was produced. In this study lower oleate/iron oxide ratios were used. The smallest amount of oleic acid, which yields a stable ferrofluid, was determined by varying the amount of oleic acid while keeping the weight of iron oxide constant. In this study, stable chloroform ferrofluids were produced with oleate/iron oxide ratio of approximately 0.16. Table 3-1 shows a list of reagents used in the study. 29

PAGE 43

30 Table 3-1. List of reagents for preparation of chloroform based ferrofluids. All reagents were used without further purification Material Source Description, catalog # Ferrous chloride tetrahydrate (FeCl 2 H 2 O) Aldrich 99%, # 22,02-9 Ferric chloride hexahydrate (FeCl 3 H 2 O) Aldrich 98%, #20,792-6 37% Hydrochloric acid solution (aq) Acros # 12463-0010 28 – 30% Ammonia Solution (aq) Aldrich # 32,014-5 Cyclohexane Aldrich 99% spectrophotometric grade, #15,474-1 Oleic acid Aldrich Tech grade 90%, #36,452-5 Chloroform Acros 99.8% HPLC grade, #61003-0040 Methanol Fisher Laboratory grade, #A411-4 Methods Preparation The method of Robineau and Zins [62] (used to prepare a suspension of iron oxide in cyclohexane solutions) was modified to form stable suspensions of iron oxide in chloroform. Figure 3-1a shows a schematic of the process. 2.03 grams of ferrous chloride tetrahydrate (FeCl 2 H 2 O), 4.88 grams of ferric chloride hexahydrate (FeCl 3 H 2 O), and 0.887 mL of 37% HCl were dissolved in 20 mL of DI water. An 8.3 mL volume of the 28-30% NH 4 OH solution was dissolved in 155 mL of DI water. This NH 4 OH solution was stirred in a 250 mL glass beaker at 350 RPM using a mechanical mixer fitted with a 4-bladed stainless steel stirrer. The ferric chloride/ferrous chloride/HCl solution was quickly added to the stirring ammonia solution to precipitate iron oxide. After stirring for 10 minutes, stirring was stopped and the iron oxide particles were collected using a 1” square NeFeB magnet at the bottom of the beaker. All but approximately 65 mL of the solution was decanted and stirring at 200 rpm was resumed. Oleic acid, dissolved in 11.14 mL cyclohexane, was added to the aqueous iron oxide slurry shortly after stirring resumed. The amount of oleic acid used for coating iron

PAGE 44

31 3. Stop stirring, add magnet to collect iron oxide 200 RPM cyclohexane/oleate solution Fe3O4 Fe3O4 Fe3O4 2. Stir 10 minutes 350 RPM 350 RPM Fe3O4 NH 4 OH Fe 2+ /Fe 3+ solution 1. Add Fe 2+ , Fe 3+ solution to N H4OH solution to precipitate iron oxide 5. Stir and add cyclohexane/oleate solution to coat iron oxide with oleate and extract into c y clohexane 4. Decant all but 65 mL of su p ernatan t Figure 3-1. Preparation of chloroform ferrofluid (continued on following page).

PAGE 45

32 Chloroform ferrofluid ferrrofluid ferrrofluid ferrrofluid methanol/aqueous p hase 9. dry overnight. Redisperse in chloroform 8. Pour ferrofluid into petri dish 8. Decant methanol/aqueous phase 7. Add magnet to collect ferrofluid droplets 200 RPM 6. Add methanol to reduce density of aqueous phase methanol aqueous phase Cyclohexane with oleate coated iron oxide (ferrofluid droplet) 10. Centrifuge and retain supernatant to remove large agglomerates Figure 3-1. Continued

PAGE 46

33 Table 3-2. Composition of ferrofluids based on weight of starting materials Sample Oleate (g) *Iron Oxide (g) *Oleate / iron oxide (w/w) **Iron oxide content of solid (w/w) 1A 0.6991 2.18 0.321 0.76 1B 0.6965 2.18 0.319 0.76 2A 0.5738 2.18 0.263 0.79 2B 0.581 2.18 0.267 0.79 3A 0.4513 2.18 0.207 0.83 3B 0.4624 2.18 0.212 0.83 4A 0.3420 2.18 0.157 0.86 4B 0.3640 2.18 0.167 0.86 5A 0.2212 2.18 0.101 0.91 5B 0.2258 2.18 0.104 0.91 *calculated based on stoichiometry of Fe 3 O 4 and weight of ferric and ferrous chloride used for precipitation of iron oxide. **calculated by dividing iron oxide weight by sum of oleate and iron oxide weight oxide was varied to determine the minimum amount of oleic acid necessary to yield a stable ferrofluid with a reasonable yield. Oleic acid concentrations varied from 0.10 to 0.32 w/w with respect to iron oxide. Table 3-2 shows the exact ratios of oleic acid to iron oxide used for the experiments. According to Robineau and Zins [62], the oleic acid should coat the iron oxide particles in the aqueous solution, making them hydrophobic, thereby drawing them into the cyclohexane solution. After 15 minutes of stirring, 55 mL of methyl alcohol was added to reduce the density of the aqueous phase and allow the cyclohexane-based ferrofluid to settle to the bottom of the beaker. The cyclohexane-based ferrofluid was then separated from the aqueous solution using the magnet placed at the bottom of the beaker. The aqueous phase was decanted and the cyclohexane ferrofluid was poured into a glass petri dish and dried in air overnight to remove the cyclohexane. The dried black product for each sample was then placed in a 50 mL polypropylene centrifuge tube and suspended in 35 mL of chloroform by shaking followed by sonication using a Branson model 2510 sonicating water bath for 5 minutes.

PAGE 47

34 The samples were shaken again and sonicated for 5 more minutes. Large agglomerates were removed from the chloroform-based ferrofluids by centrifuging at 5000 RPM for 10 minutes using a Beckman JS-21 centrifuge, followed by pipetting the supernatant. This centrifugation procedure was repeated once. X-ray Diffraction and Particle Size Estimate for Iron Oxide Three uncoated iron oxide samples were prepared by precipitating iron oxide in an identical manner to the oleic acid coated sample. Following precipitation, samples were centrifuged at 10,000 RPM using a Beckman JS-21 centrifuge and redispersed in DI water using the sonicating water bath and a vortexer. This washing procedure was repeated 2 more times. Finally one more centrifugation was performed and the supernatant was decanted. The samples were then frozen and lyophylized overnight using a Labconco model 4.5 lyophilizer. X-ray diffraction was performed on three separate samples. XRD powder diffraction scans were performed using a Philips model 3720 instrument with copper K X-rays. Scans were performed from 25 to 80 degrees 2 at a rate of 3 degrees per minute. The Scherrer formula (Equation 3-1) was used to estimate the diameter of the iron oxide particles based on X-ray peak broadening. A Vernier caliper was used to measure a portion of the 2 axis in mm from printed XRD patterns. A conversion factor for mm to degrees 2 was then determined by dividing the 2 value by the measurement in mm. The half maximum peak widths were then measured in mm on printed XRD patterns. The peak width at half maximum in mm was finally converted to degrees 2 by multiplying by the calibration factor.

PAGE 48

35 B cos B 0.9t (3-1) = wavelength of X-ray (1.54 for Cu K) ns 2 SQUIagnperconducting quantum interference device (SQUID) magnxide to L aliquot of the ferrofluid with an Eppen hours Where: t = crystal thickness (particle diameter) B = peak width at half maximum in radia B = Bragg angle of peak in degrees D Met o metry. A Quantum Design su etometer was used to measure hysteresis curves for bare iron oxide and iron ofrom dried chloroform suspensions. Three bare samples were prepared in an identical manner to the samples used for XRD. In addition to magnetic properties, the oleic acidiron oxide ratio could be determined by comparing the saturation magnetization (emu/g) of coated to bare iron oxide. Since purification of product by centrifugation was employed, the measured oleate to iron oxide ratio may be different than the value deduced from the weight of starting materials. Determination of Ferrofluid Product Yield. Yield was determined by removing a 2-m dorf brand micropipette. The suspensions were then dried in air for at least 2and the dried product was weighed. The concentration of the oleic acid/iron oxide could then be determined by dividing the dry weight by the aliquot volume. The amount of dry product in the entire sample was then determined by multiplying the aliquot dry weight by the ratio of the entire sample volume to the aliquot volume. The yield was then determined by dividing the dry product in the entire sample by the theoretical sample dry weight attributable to the iron oxide (assuming Fe3O4) and oleic acid.

PAGE 49

36 Size of Particles in Chloroform Ferrofluid. Some degree of agglomeration is expected from the chloroform ferrofluids. Particle size was measured using a Brookhaven model Zeta Plus photon correlation spectroscopy (PCS) instrument to determine the degree of agglomeration. Results XRD patterns for all three samples indicated magnetite, maghemite or a mixture of the two. Figure 3-2 shows an example XRD pattern. The crystal structures of magnetite and maghemite are similar and they have nearly identical patterns making it difficult to distinguish them. Based on chemical analysis, Robineau and Zins [62] reported a mixture of maghemite and magnetite; therefore, it would be expected that the iron oxide in this dissertation is a mixture of the two. The peaks at approximately 35.5, 57 and 63 degrees 2 were used for the estimation of particle diameter. The estimated diameter for each sample is simply the average of the estimates for the three peaks. Table 3-3 shows the data used for the particle size estimate using the Scherrer formula. The average particle size estimate for the three samples is 10.62 with a standard deviation of 0.82. Table 3-3. Iron oxide particle diameter calculations using the Scherrer formula Sample Bragg angle (degrees 2) Peak width (degrees 2) Diameter estimate (nm) Average diameter for 3 Bragg angles (nm) 35.6 0.78 10.7 57.2 0.75 12.0 Sample 1 62.7 0.86 10.9 11.2 35.5 0.73 11.4 57.1 0.92 9.9 Sample 2 62.7 0.80 11.6 11.0 35.6 0.86 9.74 57.1 0.95 9.51 Sample 3 62.8 0.95 9.81 9.7 Mean(standard deviation) diameter for samples 1,2 and 3: 10.6 (0.8)

PAGE 50

37 Figure 3-3 shows the results for yield as a function of oleic acid added during the chloroform ferrofluid preparation. As the oleate concentration decreases from 0.30 to 0.15, the yield slowly decreases. This is not surprising since the amount of stabilizing agent is decreased. After 0.15, the yield is quickly reduced. For minerals suspended in organic solutions with surfactants, a monolayer of surfactant around the crystals (or small agglomerates) is necessary for optimal colloidal stability. Below the monolayer concentration, portions of the iron oxide surface are exposed. Exposed particle surfaces tend to coalesce since the mineral-mineral interaction is energetically favorable to the solvent-mineral interaction. Appendix A shows sample calculations for the theoretical oleate concentration that yields a monolayer of surfactant assuming unagglomerated particles whose entire surface is available for adsorption of oleic acid. The calculation gives an oleate concentration (oleate/iron oxide ratio) of 0.25 for a monolayer of oleate. Since stable suspensions with high yields (above 50%) are possible below an oleate concentration of 0.25, it appears that a complete monolayer is not necessary for dispersion of the iron oxide particles. However, below an oleate concentration of 0.15, too much mineral surface area is available for agglomeration and yield is greatly reduced. Microcal Origin 5.0 was used to create a linear regression model for yield as a function of oleate concentration. The model was used to verify the linear dependence of yield on oleate concentration that appears to be present before the oleate concentration reaches 0.15. The line in Figure 3-3 represents the linear regression model for the data (not including the two lowest oleate concentrations). Equation 3-2 shows the least squares regression equation for the line; where Y equals % yield and [oleate] is oleate to iron oxide weight ratio.

PAGE 51

38 Y = 32 +144[oleate] (3-2) The slope (144) is highly statistically significant (p<0.00027). This means that the slope is not 0 and therefore there is at least some dependence of yield on oleate concentration. The coefficient of determination (r 2 ) is 0.91, indicating that 91% of the variability is because of the oleate concentration and 9% is because of experimental error. Therefore the linear regression model predicts a strong linear dependence of yield on oleate concentration Figure 3-2. XRD data for bare iron oxide Figure 3-4 shows the size of iron oxide in chloroform ferrofluids measured by photon correlation spectroscopy (PCS). Particle (agglomerate) size in the ferrofluids decreases as oleate concentration increases. Following precipitation, the iron oxide is very agglomerated, which is evident by the fast settling rate of the aqueous iron oxide

PAGE 52

39 slurry. The cyclohexane based ferrofluid and final chloroform ferrofluid are composed of a large portion of colloidaly stable material. The oleic acid therefore in addition to creating a hydrophobic iron oxide surface, can separate agglomerates without the aid of high energy mixing or high power sonication. As the oleate concentration is decreased, there is less oleate to coat iron oxide crystals, leading to bare iron oxide surfaces and increased agglomeration. This phenomenon is shown schematically in Figure 3-5. The particle size dependence on [oleate] appears to be linear. The line in Figure 3-4 represents the linear regression model. Equation 3-3 shows the least squares regression equation for the line, where D equals particle (agglomerate) diameter with units in nm. The slope -155 is statistically significant (p<.00011) and r 2 = 0.86 indicating a strong linear dependence of particle diameter on [oleate]. D = 73 155[oleate] (3-3) Figure 3-6 shows hysteresis curves for 3 different bare iron oxide samples. The samples were prepared in an identical manner. The bare iron oxide samples were precipitated as explained in the experimental section, but they were not coated with oleic acid. Following precipitation. the salts were removed by centrifugation and washing with DI water several times. These samples were then lyophilized overnight and stored in a vacuum dessicator until measurement. The saturation magnetizations at 5 T (50,000 oersted) were 65.5, 68.7, and 70.5 emu/g respectively for samples 1,2 and 3. The average of the three samples is 68.2 emu/g. The curves show very little if any hysteresis and are therefore superparamagnetic. Superparamagnetic behavior is expected since magnetite crystals are superparamagnetic at room temperature when they are below approximately 25 nm in diameter. Superparamagnetic particles are desirable to prevent agglomeration

PAGE 53

40 of microspheres in the absence of a magnetic field by attraction of permanent magnetic dipoles. Since large agglomerates were removed from samples by centrifugation, the actual ratio of oleate to iron oxide in the suspensions may vary from the value predicted from the weight of starting materials. The true ratio was estimated by comparing the saturation magnetization (emu/g) of dried suspensions to the saturation magnetization of uncoated iron oxide. Since the oleic acid is non-magnetic, the reduction in saturation magnetization can be entirely attributed to the oleic acid. The oleate/iron oxide weight ratio was determined using Equation 3-4. R ol = (M b M s )/M s (3-4) Where R ol = oleate/iron oxide ratio M b = saturation magnetization at 5 T for bare iron oxide (68.2 emu/g) M s = saturation magnetization for oleate/iron oxide complex dried from chloroform suspensions (emu/g) Figure 3-7 shows hysteresis curves for an oleate-coated sample (sample 1A) compared to bare iron oxide. Less data points were taken for the coated sample since the desired information is the saturation magnetization only. The reduction in saturation magnetization for the coated compared to the bare iron oxide is small meaning that the sample is primarily iron oxide by weight. Equation 3-4 was used to calculate the oleate/iron oxide ratio for all samples except samples 5A and 5B whose yield was very low. The results are plotted in Figure 3-8. The oleate concentration calculated using Equation 3-4 is about 30% less than expected from the weight of the starting materials. This may be because of inaccuracies in measuring the saturation magnetization of the bare iron oxide or the coated particles. One possible source of error for the bare iron oxide is that the bare surfaces are hydrophilic. The weight of sample measured therefore

PAGE 54

41 may have included water, which would reduce the saturation magnetization of bare iron oxide (M b ) and by Equation 3-4, increase the calculated value of R ol . But more importantly, the R ol in the final chloroform solution does decrease with decreasing addition in the starting composition, which is the desired result. The measured oleate concentration ([oleate] m ) appears to be linear with the theoretical oleate concentration ([oleate]). The line in Figure 3-8 represents the linear regression model. Equation 3-5 shows the least squares regression equation. The slope 0.72 is statistically significant (p<.00015) and r 2 = 0.92 indicating a strong linear dependence of particle size on [oleate]. [oleate] m = 0.72[oleate] (3-5) 0.060.080.100.120.140.160.180.200.220.240.260.280.300.320.340.3605101520253035404550556065707580 % Yieldoleate/iron oxide Figure 3-3. Yield for iron oxide suspensions in chloroform as a function of surfactant (oleate) concentration.

PAGE 55

42 0.100.150.200.250.300.3520253035404550556065 agglomerate size (nm)oleate/iron oxide ratio Figure 3-4. Size of iron oxide agglomerates measured by PCS A Oleic acid molecule Iron oxide crystal B Figure 3-5. Effect of oleate concentration on agglomeration. A) Unagglomerated iron oxide exists at high oleate concentrations. B) Stable small agglomerates with oleate excluded from agglomerate interior exist at lower oleic acid concentrations.

PAGE 56

43 -60000-40000-200000200004000060000-80-60-40-20020406080 sample 3 sample 2 sample 1emu/gOersted Figure 3-6. Magnetic hysteresis curves for bare iron oxide -60000-40000-200000200004000060000-80-60-40-20020406080 Bare iron oxide Oleate coated iron oxideemu/gOersted Figure 3-7. Magnetic hysteresis curves for bare iron oxide compared to dried oleate coated product from chloroform ferrofluids. Oleate coated iron oxide sample illustrated is sample 1A.

PAGE 57

44 0.140.160.180.200.220.240.260.280.300.320.340.100.120.140.160.180.200.220.240.26 oleate/iron oxide ratio (measured)oleate/iron oxide ratio (theoretical) Figure 3-8. Measured oleate/iron oxide ratio (measured with SQUID) vs. theoretical oleate/iron oxide ratio. Theoretical oleate/iron oxide ratio is based on weight of starting materials. Conclusions Chloroform ferrofluids were prepared with varying amounts of surfactant (oleic acid) and the optimal composition was determined to minimize surfactant. The lowest concentration of oleate at which a high yield of stable ferrofluid resulted was between 15% and 20% of the weight of iron oxide in the suspension. This correlates closely to the monolayer concentration of oleic acid that was calculated as 25% of the weight of iron oxide. Chapter 4 will describe the magnetic PLGA microspheres that were prepared using the ferrofluids.

PAGE 58

CHAPTER 4 PREPARATION OF MAGNETIC PLGA MICROSPHERES USING A SOLVENT EVAPORATION/EXTRACTION METHOD The chloroform ferrofluids discussed in chapter 3 enabled the production magnetic microspheres from polylactide-co-glycolide (PLGA) polymers using a solvent extraction/evaporation method. PLGA was dissolved in the chloroform-based ferrofluid to create an oil phase. This oil phase was then emulsified in an aqueous polyvinyl alcohol (PVA) solution. The chloroform solvent was finally removed to form hardened magnetic PLGA microspheres. Table 4-1 Reagent list for magnetic PLGA microspheres. All were used as received without further purification. Material Source Description, catalog # Chloroform Ferrofluid 5.64% w/v solids in chloroform Chloroform Acros 99.8% HPLC grade, #61003-0040 Poly (lactide-co-glycolide) (PLGA) Birmingham Polymers Inherent viscosity = 0.58 in hexafluoroisopropanol Poly vinyl alcohol (PVA) 88% hydrolyzed Janssen Chimica Molecular weight = 22,000 g/mole, #18-030-85 Triton X-100 Aldrich #23,472-9 Oleic acid Aldrich Tech grade 90%, #36,452-5 Methanol Fisher Laboratory grade, #A411-4 The chloroform ferrofluid was prepared in an identical manner to sample 3A from chapter 3. This ferrofluid was chosen since it had a high yield (around 60%) but a low concentration of oleic acid. The ferrofluid had a solid loading of 5.64%. This was determined by drying a known volume (2 mL) of ferrofluid, and weighing the dried 45

PAGE 59

46 product. The solid loading is simply the weight of the dried product divided by 2 mL. The solid loading in this case includes both the weights of the oleic acid and the iron oxide. From the SQUID results in Chapter 3, the weight of the iron oxide can be estimated by multiplying by 0.83 (the iron oxide content of solid in the ferrofluid). Conversely the weight of oleic acid can be estimated by multiplying by 0.17 Methods Preparation of Oil Phase PLGA was added to mixtures of chloroform ferrofluid and pure chloroform to form the oil phase. The ratio of PLGA to ferrofluid was varied to produce microspheres with different compositions. These mixtures were stirred for at least 2 hrs to allow the PLGA to dissolve completely. The solvent volume was held constant by making sure the ferrofluid volume plus the pure chloroform volume equaled 15 mL. The solid content (PLGA + oleic acid + iron oxide) was 1 gram for each experiment. Table 4-2 summarizes the 6 different microsphere compositions that were prepared. The oil phase compositions were prepared to yield microspheres with 0, 17%, 20%, 25%, 33% and 50% iron oxide by weight. Microsphere Preparation A 1.5% solution of PVA in DI water was prepared by adding 15 grams of PVA to 1 liter of DI water and stirring overnight. This solution was vacuum filtered once through 25 m filter paper and a second time through a 5 m paper to remove any undissolved PVA. The oil phase was poured into a 500 mL polyethylene bottle containing 200 mL of 1.5% PVA DI water solution. The solutions were then stirred for 1.5 minutes using a Kinematica GmbH model PT10/35 homogenizer set to level 6.5 to form the o/w emulsion. The o/w emulsion was then poured into a double wall beaker that was

PAGE 60

47 maintained at 45 C using a recirculating water bath. The chloroform was extracted/evaporated by stirring at 600 RPM with a mechanical stirrer for 3 hrs. The hardened microspheres were centrifuged at 4000 RPM for 10 minutes using a Beckman model J2-21 centrifuge. The supernatant was decanted and the microspheres were redispersed in 0.1% Triton-X-100 solution. This washing procedure was repeated 3 more times. Figure 4-2 is a schematic diagram showing the microsphere preparation process. Table 4-2. Composition of oil phase for microspheres (based on compositions of starting materials). 3 samples were prepared for each composition for a total of 18 samples. Sample ID 0% Fe 3 O 4 17% Fe 3 O 4 20% Fe 3 O 4 25% Fe 3 O 4 33% Fe 3 O 4 50% Fe 3 O 4 Chloroform 15.0 mL 11.5 mL 10.7 mL 9.7 mL 7.9 mL 4.4 mL Ferrofluid 0 mL 3.5 mL 4.3 mL 5.3 mL 7.1 mL 10.6 mL Iron oxide in ferrofluid 0 g 0.17 g 0.20 g 0.25 g 0.33 g 0.50 g Oleic acid in ferrofluid 0 g 0.03 g 0.04 g 0.05 g 0.07 g 0.10 g PLGA 1 g 0.80 g 0.76 g 0.70 g 0.60 g 0.40 g Microsphere Characterization Microsphere size was determined using a Coulter LS laser particle sizing instrument. Microspheres were measured in a 0.01% triton X-100 solution. Three 90 second runs were performed for each sample. A JEOL model 6335 field emission scanning electron microscope (SEM) was used to characterize the size and morphology of microspheres. The MMPS SQUID magnetometer was used to measure the saturation magnetization in emu/g of microspheres in order to determine the weight percent iron oxide in the microspheres, and the iron oxide loading efficiency. Following the first centrifugation of microspheres at 4,000 RPM using a Beckman model J2-21 centrifuge, the supernatant was not completely clear and was light brown to dark brown in color.

PAGE 61

48 C B A PLGA/chloroform/ferrofluid PVA (aq) j acketed beaker mechanical stirrer recirculating water bath set to 45 C Figure 4-1. Microsphere preparation process. A) Add Oil to water phase. B) Homogenize to form oil in water emulsion. C) Extract/evaporate chloroform at 45 C to form hardened microspheres

PAGE 62

49 The color became darker with increasing iron oxide contents. It is unclear whether or not this material left in the supernatant was composed of smaller microspheres or unincorporated magnetite. After the last centrifugation, the supernatant was poured off and the collected microspheres were freeze dried overnight using the Labconco model 4.5 freeze drier. Saturation magnetization at 5 Tesla was determined for small microsphere samples (about 60 mg). The percent iron oxide in the microspheres was then determined by comparing the saturation magnetization of microspheres to the saturation magnetization of bare iron oxide (68.2 emu/g) measured in chapter 2. Results Figure 4-2 shows particle size distributions for representative microsphere samples. With the exception of the 0% iron oxide microspheres, iron oxide content appears to have little effect on the size of microspheres. The existence of one primary peak with a wide range of larger sized microspheres indicates agglomeration in the 0% microspheres. Perhaps the surface chemistry of the unloaded microspheres makes them difficult to suspend in the 0.1% triton X-100. Figures 4-3 and 4-4 summarize the size data measured using the Coulter LS instrument. The volume average microsphere size was between 1.1 and 1.3 m for the iron oxide loaded spheres. The average size for the unloaded microspheres was significantly larger at 2.4 m, but judging from Figure 4-2 this increase was mainly because of agglomeration. From Figure 4-3, the coefficient of variation (CV) was also very similar for the iron oxide loaded microspheres. The 17% iron oxide sample appears narrower from the particle size distribution when looking at the primary peak. Upon closer examination, it is apparent that the small secondary peak is separated from the primary peak and therefore widens the overall distribution. The size ranges for the microspheres are ideal for iv administration since they are less than the

PAGE 63

50 size of the capillaries (about 5 m), but are not so small as to diminish their ability to be moved in a magnetic field. Finally, for the magnetic microspheres, the microsphere preparation method is very reproducible as evidenced by the small error bars for size and CV. Figures 4-5 through 4-10 show SEM images for the microspheres. The SEM confirms the suspicion that the broad size distribution for the unloaded microspheres is because of agglomeration, since no microspheres larger than 2 m are apparent in the micrographs. In addition, an agglomerate of about 5 m in size is evident in the upper left-hand corner of the 2000 X micrograph for the unloaded microspheres. The 17% and 20% iron oxide microspheres look very similar to the unloaded microspheres except that the surface is slightly rougher. This may be because of iron oxide nanoparticles near the surface of the particles. The microspheres with higher loadings look very irregular. Their appearance suggests a phase separation between the chloroform solvent and the polymer and/or iron oxide during extraction of the chloroform phase into the water phase. When simply drying the oil phase in air, phase separation into a dark colored (because of iron oxide) phase and a clear phase on a macroscopic scale was noted. Figures 4-10 and 4-11 show the loading of iron oxide in microspheres and loading efficiency as a function of iron oxide added to microspheres. Figure 4-10 shows that the iron oxide content in the microspheres measured by SQUID is slightly lower than the iron oxide content in the starting materials. Some iron oxide therefore appears to have left the oil phase and entered the water phase. But, for every composition, the loading efficiency exceeds 90% (see Figure 4-12).

PAGE 64

51 024681012024681012141618 O % Fe3O4 17 % Fe3O4 2O % Fe3O4 25 % Fe3O4 33 % Fe3O4 50 % Fe3O4volume percentmicrosphere diameter (m) Figure 4-2. PLGA microsphere size distributions with varying amounts of iron oxide.

PAGE 65

52 0.00.51.01.52.02.5 33 %50 %17 %20 %25 %0 %microsphere diameter (m)% iron oxide Figure 4-3. Diameter of microspheres with different iron oxide contents. Diameter is mean of particle size distribution from Coulter LS. Each bar represents the average of three sample mean diameters. Error bars are the standard error. 0 %17 %20 %25 %33 %50 %0.00.20.40.60.81.01.21.41.61.82.02.2 coeficient of variation% iron oxide Figure 4-4. Coefficient of variation for particle size distributions measured using the Coulter LS. Coefficient of variation is the standard deviation of the size distribution divided by the mean, and is therefore a measure of the breadth of the distribution.

PAGE 66

53 Figure 4-5. Scanning electron micrographs for plain microspheres

PAGE 67

54 Figure 4-6. Scanning electron micrographs for 17% iron oxide microspheres

PAGE 68

55 Figure 4-7. Scanning electron micrographs for 20% iron oxide microspheres

PAGE 69

56 Figure 4-8. Scanning electron micrographs for 25% iron oxide microspheres

PAGE 70

57 57 Figure 4-9. Scanning electron micrographs for 33% iron oxide microspheres

PAGE 71

58 58 Figure 4-10. Scanning electron micrographs for 50% iron oxide mi Figure 4-10. Scanning electron micrographs for 50% iron oxide microspheres

PAGE 72

59 1520253035404550101520253035404550 % iron oxide (measured with SQUID)% iron oxide (theoretical) Figure 4-11. Iron oxide content in microspheres measured with SQUID compared to iron oxide content in starting materials 05101520253035404550707580859095100 loading efficiency (%)% iron oxide (theoretical) Figure 4-12. Loading efficiency of iron oxide for microspheres with different iron oxide contents.

PAGE 73

60 Conclusions Magnetic PLGA microspheres of a suitable size for iv applications were prepared for the first time. This was accomplished by creating a hydrophobic surface on iron oxide nanoparticles by coating with oleic acid. The oleic acid allowed suspension of nanoparticles in an organic solvent so that the solvent evaporation/extraction process could be used to prepare microspheres. This process is simple and should be applicable to any polymer that is soluble in chloroform. Iron oxide content had little effect on microsphere size, but had dramatic effects on microsphere morphology. As the iron oxide concentration increased, the microsphere morphology became increasingly irregular. The SEM suggests a phase segregation of polymer and iron oxide from the chloroform solvent as the chloroform is extracted.

PAGE 74

CHAPTER 5 DEXTRAN MICROSPHERES The all-aqueous method of Stenekes [2] is attractive for microsphere preparation since it uses gentle chemistry and washing does not require the use of harmful organic solvents. As far as we know, microspheres small enough for iv applications have not been produced by this method. Magnetic microspheres have not been produced either. In this dissertation magnetic microspheres in a size range suitable for iv applications were produced for the first time. The effect of iron oxide content on microsphere size was determined and the microspheres were optimized for size and iron oxide content. Methods Precipitation of Iron Oxide and Preparation of a Stable Ferrofluid Table 5-1 shows a list of reagents used for all experiments. A stable aqueous ferrofluid was prepared by a method similar to the Massart Group [63]. The first step was to precipitate iron oxide. An iron chloride solution was prepared by adding 3.58 g of ferric chloride hexahydrate (FeCl 3 H 2 O), 1.32 g of ferrous chloride tetrahydrate (FeCl 2 H 2 O) and 6.67 mL of 1 M HCl to 13.33 mL of DI water in a glass beaker. The mixture was stirred with a magnetic stirrer for at least 10 minutes to dissolve the iron salts. An ammonia solution was prepared by adding 16.20 mL of 28-30% NH 4 OH to 167 mL of DI water in a 250 mL beaker. The ammonia solution was stirred at 350 RPM using a mechanical stirrer fitted with a stainless steel single bladed stirring paddle. The iron chloride solution was quickly added to the ammonia solution and a black product 61

PAGE 75

62 Table 5-1. List of reagents for magnetic dextran microspheres. All were used as received without further purification. Material Source Description, catalog # Ferrous chloride tetrahydrate (FeCl 2 H 2 O) Aldrich 99%, # 22,02-9 Ferric chloride hexahydrate (FeCl 3 H 2 O) Aldrich 98%, #20,792-6 37% Hydrochloric acid solution (aq) Acros # 12463-0010 28 – 30% Ammonia Solution (aq) Aldrich # 32,014-5 70% Nitric acid (aq) Ashland Molecular sieves Aldrich 3 angstroms, 4-8 mesh, #20,857-4 Dextran Polydex Weight average molecular weight = 36,600 g/mole 4-(dimethyl amino) pyridine Aldrich # 10,770-0 Glycidyl methacrylate Acros #16589-1000 Dimethyl Sulfoxide (DMSO) Acros 12779-0025 Polyethylene glycol Aldrich Weight average molecular weight = 10,000 g/mole, #30,902-8 Polyethylene glycol Polysciences Weight average molecular weight = 35,000 g/mole, #22568 Polyethylene glycol Polysciences Weight average molecular weight = 20,000 g/mole, #22569 N,N,N’,N’ tetramethyl ethylene diamine (TEMED) Aldrich 99.5%, # 41,101-9 Potassium persulfate (KPS) Aldrich 99%, #21,622-4 Potassium chloride (KCl) Aldrich #20800-500G immediately precipitated. Stirring was continued for 10 minutes, after which the black precipitate was collected at the bottom of the beaker using a 1” square neodinium iron boron magnet. The supernatant was decanted and the precipitate was redispersed in 1 M nitric acid. The slurry was poured into a 50 mL polypropylene centrifuge tube and the tube was rotated for 10 minutes at room temperature using a Robbins Scientific model 400 hybridization incubator. The sample was then centrifuged using an Adams Dynac centrifuge and the supernatant was decanted. The precipitate was redispersed in DI water and rotated at room temperature for 10 minutes. A 20 mL volume of 1 M nitric acid was then added to the tube to flocculate the iron oxide particles. The product was then centrifuged at 9000 RPM using a Beckman model J2-21 centrifuge. A stable ferrofluid

PAGE 76

63 was finally obtained by decanting the supernatant, adding 30 mL of DI water and vortexing. The concentration of each ferrofluid was determined by pipetting 2 mL of ferrofluid, drying the ferrofluid in air, and weighing the dried product. Typical concentrations were around 35 mg/mL. Photon correlation spectroscopy (PCS) was used to measure the particle size of iron oxide in the ferrofluids. A Brookhaven Zeta Plus instrument was used for PCS. Five 2-minute runs were used for each measurement. X-ray diffraction was used for phase identification of the iron oxide. Dry iron oxide powder samples were prepared from the ferrofluids for the X-ray diffraction measurements. NaOH (0.1 M) was added dropwise to the ferrofluids to flocculate them. The ferrofluids were centrifuged and washed several times with DI water and then freeze dried overnight. The dried ferrofluids were ground with a mortar and pestal. The ground powders were deposited on glass slides along with a colloidon/amyl acetate solution which when dried adhered the iron oxide to the slide. X-ray diffraction 2 diffractograms were generated using a Phillips APD 3720 diffractometer ran at 4 degrees 2 per minute using a step size of 0.02 degrees 2. Specific surface area (using the BET method) was determined for the dry iron oxide powders using a Quantichrome NOVA 1200. Nitrogen adsorption was determined using the multipoint BET method at the following N 2 pressures: P/P 0 = 0.05, 0.1, 0.15, 0.2, 0.25 and 0.3. Preparation of Methacrylated Dextran (Dex-MA) Figure 5-1 shows a schematic for the experimental apparatus used for preparation of methacrylated dextran (dex-MA). Dextran was modified with methacrylate groups to allow crosslinking by free radical initiated polymerization. Methacrylate modification of

PAGE 77

64 dextran was accomplished by reacting dextran with glycidyl methacrylate as described by the Vandijkwolthuis group [61]. Dimethyl amino pyridine (DMAP) and dextran were placed in a labconco model 4.5 freeze drier overnight to remove absorbed and adsorbed water. DMSO was dried overnight using molecular sieves. A 0.63 mL volume of DMSO was added to 7 g of dextran and 1.4 g DMAP in a sealed 125 mL 3 necked flask that was fitted with a gas bubbler. The mixture was agitated with a magnetic stirrer. Nitrogen gas dried through a Drie Rite column was flowed through the 125 mL 3 necked flask to provide an inert atmosphere during dissolution of the DMAP and dextran. Once the DMAP and the dextran dissolved, 856 l of glycidyl methacrylate (GMA) was injected with a syringe and needle through a rubber septum that sealed the 3-necked flask. Nitrogen gas flow was continued throughout the reaction. After 72 hours, 957 l of 37% HCl was injected into the flask to neutralize the DMAP and stop the reaction. The reaction mixture was then placed in 32 mm dialysis tubing membranes (Spectrum labs #132655) that were sealed with two 55 mm enclosures (Spectrum labs # 132737) on each end. Dialysis membranes were soaked in DI water for 10 minutes prior to use in order to hydrate them. The reaction mixture was dialyzed against 1 L of DI water for 14 days to purify the dex-MA by removing the DMAP, HCl, and any ungrafted GMA. The dialysate was changed daily. After dialysis, the purified dex-MA solution was poured on the inside of a 600 mL labconco freeze drying flask while chilling the outside of the flask with a dry ice/ethanol mixture. This resulted in a frozen solution on the inside of the flask that was readily dried overnight in the Labconco model 4.5 freeze dryer. Following freeze drying, the fluffy white product was stored in a freezer.

PAGE 78

65 N 2 gas ” Tygon tubing Dri-rite column Female luer lock needle Male luer lock adaptor Rubber septa 3 necked flask Magnetic stirrer Bubbler filled with DMSO Figure 5-1. Apparatus for preparation of dex-MA by reaction of dextran with glycidyl methacrylate A Varian 300 MHz nuclear magnetic resonance (NMR) spectrometer was used to determine the degree of methacrylate substitution following the reaction of glycidyl methacrylate and dextran. Figure 5-2 shows spectra for dextran and dex-MA. The peak at 5.0 ppm (labeled a) is characteristic of the proton bonded to carbon 1 of the dextran glucose repeat unit. The peak at 5.3 ppm (a’) is also for the proton bonded to carbon 1, but is shifted for glucose units with branching at carbon 3. The peaks at 5.8 and 6.3 ppm are characteristic of the protons of the CH 2 group of the methacrylate double bond and are labeled b. The peaks at 5.25 ppm are characteristic of the proton bonded to the carbon adjacent to the ester group of the glycidyl methacrylate and are labeled d. The peak at 1.95 ppm is characteristic of the protons in the methacrylate methyl group (c).

PAGE 79

66 The degree of substitution was determined by comparing the area of peak c to the sum of the areas of a and a’. The degree of substitution was calculated as 12% meaning that there are 12 methacrylate groups for every glucose unit. The reaction mixture contained 15 moles of glycidyl methacrylate per 100 moles of glucose. Therefore the yield was 12 divided by 15 which equals 80%. ’ O a O HO OH HO a & a’ O a’ a OH O O O b O d HO O a & a’ OH c O c b b a’ d Figure 5-2. Nuclear magnetic resonance spectra for dextran (top) and dex-MA (bottom).

PAGE 80

67 Adsorption of Dex-MA onto Iron Oxide in Ferrofluids Since the microsphere preparation method involves adsorbing dex-MA onto iron oxide in the ferrofluid, the adsorption of dex-MA onto iron oxide was determined as a function of dex-MA concentration. The ferrofluid used for adsorption studies was first diluted to 20 mg/mL by adding DI water. Solutions with varying ratios of dex-MA to iron oxide were then prepared. This was accomplished by adding 2 mL of the ferrofluid to 1.5 mL DI water containing varying amounts of dex-MA. The dex-MA concentration was varied from 0.92 to 84.4 mg. All samples were prepared in 4 mL glass vials. The samples were rotated overnight at 37 C in a Robbins Scientific model 400 hybridization incubator to allow the dex-MA to adsorb onto the iron oxide particles. The samples were then poured into 12 mm diameter glass culture tubes and placed in magnetic separators for 2 days. The magnetic separators consist of 4 magnets retained in a steel cylinder. An insert machined from Delrin is used to maintain the spacing between magnets. The magnets are 1” x 3/8” x 3/8” neodinium iron boron magnets from Magnet Sales and Manufacturing Inc., Culver City, CA. They are magnetized through the width of the magnet. The magnets are oriented so that the north poles of magnets 1 and 3 face each other while the south poles of magnets 2 and 4 face each other. After 2 days, the solutions were clear, meaning that all the magnetic particles were separated. The clear solutions were pipetted into separate vials. Since dex-MA is an optically active polymer, its concentration in solution is proportional to its angle of rotation in a polarimeter. A Perkin-Elmer model 241 polarimeter was used to measure the concentration of free dex-MA in solution after adsorption onto the iron oxide particles. First a calibration curve for rotation angle vs. concentration was prepared for dex-MA concentrations of 0.1, 0.5, 2, and 10 mg/mL dex-MA. The calibration curve is

PAGE 81

68 shown in Figure 5-4. Microcal Origin 5.0 software was used to create the linear regression model with the origin going through 0. The slope of the line is 0.183 and coefficient of determination (r 2 ) equals 1.0000. After the calibration curve was created, the concentration of free dex-MA was determined by measuring the rotation angle. The concentration of dex-MA adsorbed on the particles was determined by subtracting the concentration of dex-MA before adsorption from the free dex-MA concentration after adsorption. 4 3 2 1 Figure 5-3. Magnetic separator. Left shows magnet, Delrin insert and steel ring. Right shows assembled separator. Preparation of Magnetic Dextran Microspheres Microspheres were prepared using the all-aqueous phase separation technique. Aqueous PEG solutions were used as the continuous phase. PEG with molecular weights of 10,000, 20,000 and 35,000 weight average grams/mole were used. In all cases, the PEG solution was prepared by dissolving 30 grams of PEG in 64.4 mL of 0.22 M KCl in a 125 mL glass Erlenmeyer flask. The dispersed phase was prepared by adding a volume of the 35.2 mg/mL ferrofluid to 200 mg of dextran dissolved in DI water and rotating overnight at 37 C. The volume of the DI water in addition to the volume of ferrofluid

PAGE 82

69 was 4.6 mL for all samples. The iron oxide content of the microspheres was controlled by the volume ratio of the ferrofluid to the DI water. 02468100.000.250.500.751.001.251.501.752.00 rotation angle (degrees)[dex-MA]: mg/ml Figure 5-4. Calibration curve for rotation angle vs. [dex-MA]. The iron oxide/dex-MA phase was homogenized in the PEG continuous phase to form droplets of the iron oxide/dex-MA phase. A free radical initiator was then added to the system to polymerize the methacrylate groups and crosslink the dextran, thereby stabilizing the microspheres. The PEG solution contained in a 125 mL Erlenmeyer flask was placed in a jacketed beaker that was filled with water and maintained at 10 C with a Fisher Isotemp 9500 recirculating water bath to cool the solution during homogenization. The iron oxide/dex-MA phase was then poured into the PEG solution and stirred for 2 minutes using a Kinematica GmbH model PT10/35 homogenizer set to level 6.5. Nitrogen gas was used to purge the flask during homogenization. Following homogenization, 2 mL of a 20% N,N,N’,N’ tetramethyl ethylene diamine (TEMED)

PAGE 83

70 solution adjusted to pH 8.5 with HCl was added to the mixture while stirring with a glass stir rod. A 7.2 mL volume of 25 mg/mL potassium persulfate (KPS) solution was then added while stirring with a glass stir rod. The TEMED and the KPS are the free radical initiation system. After the initiators were added, each sample was poured into two 50 mL centrifuge tubes and rotated overnight at 37 C. Figure 5-5 shows the apparatus used for production of the magnetic dextran microspheres. N 2 gas Erlenmeyer flask j acketed beake r homogenizer recirculating water bath Figure 5-5. Magnetic dextran microsphere preparation apparatus. After reacting overnight, the microspheres were centrifuged and washed many times in DI water. Each sample was diluted 3 times with DI water to lower the viscosity of the PEG solution. The sample was then added to 6 50 mL centrifuge tubes and centrifuged at 9000 RPM for 10 minutes using a Beckman model J2-21 centrifuge. The supernatant was poured from each of the six tubes and the microspheres were redispersed in DI water and combined into 1 tube. They were then centrifuged at 5000 RPM for 10 minutes and redispersed in approximately 40 mL of DI water. This was repeated 2 more times to wash the microspheres.

PAGE 84

71 In addition to the microspheres, bulk hydrogels were prepared to study the polymerization crosslinking reaction. 50 mg of dex-MA were dissolved in 450 l of DI water in a plastic 1.5 mL centrifuge tube. A 9.9 l volume of the 20% TEMED solution and 2.5 mg of KPS powder were added to the centrifuge tube, which was vortexed to dissolve and mix the KPS and TEMED solution in the dex-MA solution. The solution was then added to a covered plastic petri dish and polymerized in an oven at 37 C for 1 hr. Following polymerization, the crosslinked gel was placed in a 50 mL centrifuge tube with 40 mL of DI water and rotated for 2 hrs. at 37 C to purify the gel. This purification step was repeated one more time using fresh DI water. The gel was dried in an oven at 45 C overnight. Characterization of Magnetic Dextran Microspheres As described in chapter 4 for the PLGA microspheres; particle size, iron oxide content and microscopy were performed on the magnetic dextran microspheres. Laser particle size analysis was used to determine the volume average size distributions and mean diameter of the microspheres. SEM was used for microsphere morphology and to qualitatively assess size. Iron oxide content was determined using SQUID magnetometry. Fourier transform infrared (FTIR) spectroscopy was used to verify the polymerization of the vinyl groups during the crosslinking of the microspheres and the bulk hydrogels. A Nicolet MAGNA 760 FTIR spectrometer in the diffuse reflectance mode was used for all spectra. Diffuse reflectance samples were prepared by grinding approximately 3 to 9 mg of sample along with 350 mg of potassium bromide (KBr).

PAGE 85

72 Microsphere yield was determined from microspheres suspended in 40 mL of DI water by first withdrawing 2 mL of suspended microspheres using an Eppendorf brand micropipette. The suspension was dried and the weight of the dry microspheres recorded. The weight of the microspheres in the entire sample was then determined by multiplying the recorded weight by 20 to obtain the weight of the entire 40 mL suspension. The yield was then calculated by dividing the weight of the entire sample by the theoretical weight from the weight of starting materials. Results Characterization of Ferrrofluid Figure 5-6 shows the X-ray powder diffraction pattern for the iron oxide from the ferrofluid. The peaks are consistent with magnetite or maghemite which both have similar magnetic properties. The BET surface area was 105 m 2 /g. Assuming spherical particles and a density of 5.25 g/m 2 for magnetite, a particle diameter of 10.9 nm is estimated from the surface area. Using the Scherrer formula, the particle diameter was estimated to be 7.6, which is in fairly close agreement with the BET particle diameter. The diameter measured in solution using PCS was 68 nm, which is an order of magnitude higher than the other 2 methods. This means that the iron oxide must exist as agglomerates in the ferrofluid. Figure 5-7 shows the results of the SQUID magnetometry. The curve shows that the iron oxide nanoparticles become magnetized at a low field and show little or no hysteresis. This means that they are superparamagnetic. The lack of hysteresis is expected since the iron oxide particles were measured to be less than the 25 to 30 nm diameter required for superparamagnetic magnetite [64].

PAGE 86

73 Figure 5-6. Iron oxide X-ray diffraction pattern. -20000-15000-10000-500005000100001500020000-80-60-40-20020406080 Magnetization (emu/g)H (Oersted) -1000-50005001000-80-60-40-20020406080 Figure 5-7. Magnetization of bare iron oxide as a function of applied magnetic field. Measurement temperature was 25 C.

PAGE 87

74 Adsorption Studies Figure 5-8 shows the results of the adsorption studies for dex-MA adsorption onto the iron oxide in the ferrofluid. The adsorption shows reversible Langmuir behavior where adsorbed polymer increases with polymer concentration in solution. Since the adsorption is reversible, the dex-MA does not adsorb strongly to the iron oxide nanoparticles. 01020304050607080024681012 mg iron oxide = 40solution volume = 3.5 mlDex-MA adsorbed (milligrams)Dex-MA in solution (milligrams) Figure 5-8. Adsorption studies for dex-MA onto iron oxide in ferrofluids. Preliminary Microsphere Synthesis Studies Preliminary microsphere synthesis studies were carried out to see the effects of PEG molecular weight and iron oxide content on microsphere size and morphology. Microspheres were made with 12% iron oxide by weight and PEG with three different molecular weights: 10,000 (PEG 10K) , 20,000 (PEG 20K), and 35,000 (PEG 35K)

PAGE 88

75 g/mole. In addition, microspheres were made with 29% iron oxide and PEG 20K only. Table 5-2 summarizes the formulations for the microspheres prepared. Table 5-2. Composition of starting materials for each microsphere batch. Composition of dex-MA/iron oxide dispersed phase Sample Dex-MA (mg) DI water (mL) 35.2 mg/mL ferrofluid (mL) 12% iron oxide* PEG10K** 200 3.83 0.77 12% iron oxide* PEG20K** 200 3.83 0.77 12% iron oxide* PEG35K** 200 3.83 0.77 29% iron oxide* PEG20K** 200 2.28 2.32 * % iron oxide is dry weight percent iron oxide in the dispersed phase ** continuous phase consists of 30 g PEG dissolved in 64.4 mL of 0.22 M KCl. Figures 5-9 and 5-10 show size distributions for the microspheres. The PEG10K sample was plotted in a separate figure because of its broad size distribution. Looking at the 12% iron oxide samples only, microsphere diameter decreases as the PEG molecular weight increases. This is because of the increased viscosity of the of the PEG polymer solution as PEG molecular weight is increased. This increased viscosity of the continuous phase leads to higher shear forces on the dex-MA/iron oxide droplets during homogenization and smaller dex-MA/iron oxide droplets. The PEG 10K sample has an extremely wide distribution of microsphere sizes. The low viscosity of the PEG10K solution may allow the dex-MA/iron oxide droplets to coalesce after they are formed. Another explanation of the large size distribution for the PEG 10K relates to the dex-MA/PEG phase diagrams generated by Stenekes [2]. The dex-MA/iron oxide phase originally is very dilute with approximately 4% dextran. According to the phase diagrams, the equilibrium dex-MA concentration in the droplets is around 40%. This

PAGE 89

76 means that with time, water diffuses out of the dex-MA/iron oxide phase, thereby concentrating the phase and increasing its viscosity. The diffusion of water should increase for the low molecular weight PEG. If significant diffusion occurs during homogenization, a broad size distribution results. Comparing 12 to 29% iron oxide samples with PEG 20K shows that microsphere diameter increases as iron oxide content increases. This size increase is because of the increase in viscosity of the dex-MA/iron oxide phase with iron oxide content. Figures 5-11 through 5-14 show SEM micrographs of the 4 different microsphere groups. The SEM verifies the particle size results, showing a large size distribution for the PEG 10K sample. Higher molecular weight PEG (PEG 20K and PEG 35K) results in smaller microspheres with narrower size distributions. Comparing Figure 5-9 to 5-10, the increase in size with iron oxide content is apparent. Figure 5-15 shows the SQUID magnetometry for bare iron oxide compared to the 12 and 29% iron oxide (weight based on starting materials) microspheres made with PEG 20K. The iron oxide content of the final microspheres was determined by comparing the saturation magnetization (emu/g) of the bare iron oxide to the microspheres. This data is shown in Table 5-3 along with a summary of the size data from the laser particle size analysis. The measured iron oxide content is higher than the nominal iron oxide content in all cases. This is probably because of the loss of dex-MA to the PEG phase during homogenization and while the crosslinking reaction occurs. Loss of dex-MA to the PEG phase is because of a small degree of solubility of dex-MA in PEG.

PAGE 90

77 0.00.51.01.50510152025 PEG 35K, 12 % iron oxide PEG 20K, 12 % iron oxide PEG 20K, 29 % iron oxideVolume %m Figure 5-9. Size distributions for dextran microspheres prepared with PEG 20K and PEG 35K. 0501001502002503000.00.51.01.52.02.53.0 PEG 10K, 12 % iron oxideVolume %m Figure 5-10. Size distribution for dextran microspheres prepared with PEG 10K.

PAGE 91

78 Figure 5-11. Scanning electron micrograph for dextran microspheres prepared with 12% iron oxide and PEG10K. Note broad size distribution Figure 5-12. Scanning electron micrograph for dextran microspheres prepared with 12% iron oxide and PEG20K.

PAGE 92

79 Figure 5-13. Scanning electron micrograph for dextran microspheres prepared with 12% iron oxide and PEG35K. Figure 5-14. Scanning electron micrograph for dextran microspheres prepared with 29% iron oxide and PEG20K

PAGE 93

80 -6.0x104-4.0x104-2.0x1040.02.0x1044.0x1046.0x104-80-60-40-20020406080 Iron Oxide Microspheres: 29% iron oxide Microspheres: 12% iron oxideMagnetization (emu/g)H (Oersted) Figure 5-15. Magnetometry (SQUID) for dextran microspheres prepared with PEG20K compared to bare iron oxide. Table 5-3. Summary of results for preliminary microsphere synthesis studies Sample Molecular weight of PEG (g/mole) Nominal iron oxide wt. % Measure iron oxide wt. % (SQUID) Mean diameter (m) 1 10,000 12 15 75 2 20,000 12 15 0.6 3 35,000 12 15 0.5 4 20,000 29 36 0.8 Microsphere Synthesis Studies Using PEG20K and Varying the Iron Oxide Content PEG 10K was not used for further studies since it resulted in an undesirable broad microsphere size distribution. Since PEG 20K and 35K yielded similar results, PEG 20K was chosen for the remaining microsphere synthesis studies. The effect of iron oxide content on microsphere size and morphology was studied. Microspheres with iron oxide contents of 20, 30 and 40% were prepared. Three samples were prepared at each iron oxide content. As in the preliminary studies, the volumes of both the dex-MA/iron oxide

PAGE 94

81 phase and the PEG solution were the same for all sample groups. Table 5-4 summarizes the formulations use for the three different sample groups. Table 5-4. Composition of starting materials for each microsphere group. Composition of dex-MA/iron oxide dispersed phase Sample Dex-MA (mg) DI water (mL) ferrofluid (mL) 20% iron oxide* 200 3.18 1.42 30% iron oxide * 200 2.19 2.41 40% iron oxide* 200 0.85 3.75 * % iron oxide is dry weight percent iron oxide in the dispersed phase. Continuous phase consists of 30 g PEG20K dissolved in 64.4 mL of 0.22 M KCl. Three samples were prepared at each iron oxide content. Figures 5-16 through 5-18 show SEM micrographs for the microspheres with three different iron oxide contents. The microspheres are relatively spherical and size increases with iron oxide content . The size increase is probably because of increased viscosity of the dex-MA/iron oxide phase with increased iron oxide content. Figure 5-19 shows the microsphere size distributions and Figure 5-20 shows the mean microsphere diameters as a function of iron oxide content. These data verify the SEM micrographs showing the increase in size with iron oxide content. The 40% iron oxide sample contains microspheres above 5 m in diameter and is therefore not suitable for iv applications. The 30% iron oxide microspheres are all below 4 m in diameter, making them suitable for iv applications. Microsphere yield is shown in Figure 5-21. Yield increases with iron oxide content. During centrifugation to wash the microspheres, the smallest microspheres are discarded with the supernatant. Since the higher iron oxide content microspheres have a diminished proportion of small microspheres, less microspheres are lost while washing. Table 5-5 shows the iron oxide content of the microspheres measured using SQUID. As in the preliminary experiments, measured iron oxide content is higher than predicted from the starting materials.

PAGE 95

82 Table 5-5. Iron oxide content of dextran microspheres measured with SQUID *Nominal % iron oxide **Measured % iron oxide 20 24.6 (0.2) 30 35.1 (0.3) 40 45.0 (0.4) * based on dry weight of starting materials ** Measured with SQUID. Each is average (standard error) of three samples Figure 5-22 shows FTIR spectra for dextran, dex-MA and the crosslinked hydrogel prepared by polymerizing the dex-MA methacrylate groups. The absorption near 1700 cm -1 for the dex-MA and the crosslinked hydrogel is for the carbonyl group of the methacrylate group. Any vinylidine stretch for the methacrylate group, which should be around 1650 cm -1 , is obscured by a large peak from the dextran. The strongest IR absorption for the vinylidine functional group is the CH 2 wag [65], which for methyl methacrylate occurs at around 813 cm -1 [61, 66]. This means that the disappearance of the absorption at 813 cm -1 can be used as evidence of crosslinking by polymerization of the dex-MA methacrylate groups. The peak at 813 is not evident in the full scale spectrum in Figure 5-22. However the absorption at 813 cm -1 for dex-MA is clearly evident in the expanded view of the higher wave number region. This peak disappears for the crosslinked hydrogel indicating that all or at least nearly all of the methacrylate groups are polymerized. Similarly, Figure 5-23 shows spectra for the 30% iron oxide microspheres compared to dex-MA. The peak at 813 disappears indicating that the methacrylate groups were polymerized for the microspheres.

PAGE 96

83 Figure 5-16. Scanning electron micrographs for 20% iron oxide microspheres

PAGE 97

84 Figure 5-17. Scanning electron micrographs for 30% iron oxide microspheres

PAGE 98

85 Figure 5-18. Scanning electron micrographs for 50% iron oxide microspheres

PAGE 99

86 024681012140246810 20 % Fe3O4 30 % Fe3O4 40 % Fe3O4Volume %m Figure 5-19. Microsphere size distribution for representative samples. 012345 30 % Fe3O440 % Fe3O420 % Fe3O4mean microsphere diameter (m)Sample Figure 5-20. Volume mean microsphere diameters calculated from microsphere size distributions. Each bar is the average of three samples. Error bars are the standard error.

PAGE 100

87 01020304050607080 40 % Fe3O430 % Fe3O420 % Fe3O4% YieldSample Figure 5-21. Microsphere yield as a function of iron oxide content. Each bar is the average of three samples. Error bars are the standard error. Conclusions Dextran microspheres were produced using the all-aqueous phase separation method. Microspheres suitable for iv applications and magnetic microspheres were produced by this method for the first time. Microsphere size increased with increasing iron oxide content. The diameter of microspheres with 35 weight percent iron oxide was less than 5 m for the entire size distribution, while many microspheres were greater than 5 m in diameter for the 45% iron oxide samples. The iron oxide content of the microspheres was higher than predicted from the weight of starting materials. This was most likely because of loss of dex-MA to the PEG phase because of a small degree of dex-MA miscibility in PEG. Finally FTIR revealed that the free radical initiated crosslinking reaction polymerized the majority of methacrylate groups.

PAGE 101

88 -0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 0.12 0.13 0.14 0.15 1000 2000 3000 4000 Wavenumbers (cm-1) A Log 1/R 0.110 0.115 0.120 0.125 0.130 0.135 0.140 0.145 0.150 0.155 0.160 0.165 0.170 0.175 0.180 0.185 600 800 1000 1200 Wavenumbers (cm-1) B Log 1/R Figure 5-22. FTIR spectra for crosslinked dex-MA hydrogels. A) Full Spectra. B) Expanded scale highlighting peak at 813 cm -1 . From bottom to top: dextran, dex-MA, and crosslinked dextran by polymerization of dex-MA

PAGE 102

89 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 0.12 0.13 0.14 0.15 0.16 0.17 0.18 0.19 L o 1000 2000 3000 4000Wavenumbers (cm-1) Log 1/R 0.14 0.15 0.16 0.17 0.18 0.19 0.20 0.21 0.22 0.23 0.24 0.25 0.26 0.27 0.28 0.29 0.30 0.31 0.32 1000 Wavenumbers (cm-1) B 140 0 120 0 600 800 A g (/ R) Log 1/R Figure 5-23. FTIR spectra for 30% iron oxide microspheres compared to dex-MA. A) Full spectra. B) Expanded view highlighting peak at 813 cm -1 . For each graph, top spectrum is 30% iron oxide microspheres; bottom is dex-MA.

PAGE 103

CHAPTER 6 IN VITRO MICROSPHERE RETENTION STUDIES An in vitro model consisting of a flow system and a rectangular neodinium iron boron magnet was used to determine the ability of the microspheres flowing in a liquid stream to be retained by a magnet. Arterial flow velocities are around 10 cm/s, and capillary flow velocities are approximately 0.05 cm/s [1]. Accordingly, various speeds between these two values were used in the experiment. Microspheres were easily retained at the capillary flow speeds, but were not retained at speeds approaching the arterial flow velocities. Methods In Vitro Retention Studies Retention apparatus Senyei et al. [1] used a flow system as a model for magnetic guidance of magnetic albumin microspheres. The system consisted of a syringe pump to control the microsphere flow, polyethylene tubing, and a magnet to retain the microspheres against the interior wall of the tubing. A similar apparatus, shown in Figures 6-1 and 6-2 was used for this dissertation. A Harvard model 975 mechanical syringe pump with a 60 mL plastic syringe was used to control the flow rate of a carrier solution. For the PLGA microspheres, the carrier solution was 0.1% w/v triton X-100 in DI water, and for the dextran microspheres the carrier solution was DI water. The flow rate was calibrated for each syringe pump speed setting by weighing the amount of water pumped with the 60 mL syringe for a certain time. Flow rate Q is equal to flow velocity V multiplied by 90

PAGE 104

91 cross sectional area A. The flow velocity V was therefore calculated by dividing the measured Q by the cross sectional area A of the 1/16” inner diameter tubing used for the experiment. Figure 6-3 shows a graph of the flow velocity for each syringe pump speed. Microsphere suspensions were injected into the injection port using a 1 mL syringe fitted with a 25 gauge needle. A ” by ” by 1” neodymium iron boron magnet obtained from Edmunds Optics was used to retain the microspheres within silicone rubber tubing. The magnet was magnetized through its thickness. The silicone tubing was purchased from Cole Parmer and has an inner diameter of 1/16” and outer diameter of 1/8”. A 15 mL polypropylene centrifuge tube was used to collect the microspheres not retained by the magnet. Figure 6-4 shows the magnet holder. The holder is made of aluminum, which is non-magnetic. The 2 screws on the end of the holder were used to control the distance between the magnet and the tubing. The 4 screws on the back of the holder are magnetic and were therefore used to hold the magnet in place. 15 mL centrifuge tube Magnet holder 60 mL syringe 3/16” OD tubing 3/16” OD tubing 1/8” ID tubing Injection port with rubb er Magnet 1 mL syringe for microsphere injection Syringe pump Figure 6-1. Retention apparatus with labeled components

PAGE 105

92 Figure 6-2. Retention apparatus photograph. 468101214161820220.010.1110 flow velocity (cm/s)syringe pump setting Figure 6-3. Syringe pump calibration: Flow velocity as a function of syringe pump setting.

PAGE 106

93 C B A Figure 6-4. Magnet holder. Scale in inches. Retention experiment For each experiment, the flow speed was selected on the syringe pump while the pump was stopped. A 200 l volume of microsphere suspension was injected through the septum in the injection port. The PLGA microspheres were the 20% iron oxide microspheres from chapter 4. The microspheres were suspended in 0.1% triton X-100 at a concentration of 12.5 mg/mL. The dextran microspheres were the 35% iron oxide microspheres from chapter 5 and had a concentration of 4.3 mg/mL. The pump was started and then stopped when 2 mL of solution was collected. The volume of solution in the tubing between the injection port and the end of silicone tubing was approximately 0.7 mL. Therefore all of the microspheres reached the magnetic field after 2 mL was pumped. The magnetic field and field gradient were changed by adjusting the distance of the magnet edge to the center of the silicone tubing. The magnetic field as a function of distance from the edge of the magnet was measured using a F. W. Bell model 5080 gauss meter with a transverse probe. The field was measured with the long axis of the probe perpendicular to the magnet edge and the width of the probe parallel to the magnet edge.

PAGE 107

94 Figure 6-5 shows the field (Oe) as a function of distance from the edge of the magnet. The field gradient (Oe/cm), shown in Figure 6-6, was calculated for each point. This was accomplished by subtracting the indicated field from the field measured 1 mm (0.1 cm) closer to the magnet and dividing by 1 mm (0.1 cm). Quantifying % of microspheres retained using spectrophotometer After each experiment, the microspheres retained in the silicone tubing were washed out of the tubing using 0.1% triton X-100 for the PLGA microspheres and DI water for the dextran microspheres. DI water for dextran microspheres or triton X-100 for PLGA were subsequently added to the washing solution (containing the retained microspheres) to adjust the solution to 10 mL. Similarly, the 2 mL of solution containing microspheres that had passed by the magnet was adjusted to 10 mL by addition of the appropriate solution. The concentrations of both microspheres retained and those that passed by the magnet were then determined with a spectrophotometer. First, standards of various concentrations were prepared by diluting the 12.5 mg/mL PLGA microspheres and the 4.3 mg/mL dextran microspheres. These standard solutions were then placed in cuvettes and their absorption measured using a Shimadzu UV-2401PC UV-vis spectrophotometer. Absorption was measured at 900 nm since this wavelength yielded the highest r 2 values: 0.99992 for PLGA microspheres and 0.99955 for dextran microspheres. Figures 6-7 and 6-8 show the spectrophotometer calibration curves for the two different microspheres. Once the calibration curves were established, the concentrations of the retained microspheres and the microspheres that passed through the magnetic field were determined by measuring their absorbance in the spectrophotometer.

PAGE 108

95 0510152025300500100015002000250030003500 magnetic field (Oe)distance from edge of magnet (mm) Figure 6-5. Magnetic field as a function of distance from the magnet. 05101520253002000400060008000 field gradient (Oe/cm)distance from edge of magnet (mm) Figure 6-6. Magnetic field gradient as a function of distance from the magnet.

PAGE 109

96 0.000.050.100.150.200.250.300.350.00.20.40.60.81.01.2 absorptionmg/ml Figure 6-7. Spectrophotometer calibration curve for PLGA microspheres 0.000.020.040.060.080.100.120.00.20.40.60.81.0 absorptionmg/ml Figure 6-8. Spectrophotometer calibration curve for dextran microspheres.

PAGE 110

97 In Vitro Retention Study Model Senyei et al. [1] developed a simple model to predict the % retention of magnetic microspheres as a function of tube diameter, magnetic field gradient, iron oxide magnetization and flow speed. This model was compared to the experimental results with albumin microspheres and the results were found to be similar. Figure 6.9 shows a schematic of a microsphere traveling through a tube that serves as a basis for the Senyei model. The derivation of the model is explained on the following pages. The model makes many assumptions. The first is monodisperse microspheres since it is based on calculations for a single microsphere of a given radius. The model assumes a constant magnetic field (H) and field gradient (dH/dx). Therefore the distance between the magnet and the center of the tube (x 0 in Figure 6-9) must be much greater than the tube radius, r. Since the model is based on single microsphere calculations, it assumes no agglomeration between particles under the magnetic field. The magnetic force on a magnetic microsphere because of a magnetic field is shown in Equation 6-1. M was determined for each value of H from the SQUID data. vy z y x w magnet x0 vx R R Figure 6-9. Flow of a microsphere through tubing under a magnetic field. v y is the velocity of the microsphere in the y direction because of the flowing liquid and v x is the velocity in the x direction because of the magnetic.

PAGE 111

98 The magnetic force is a function of the microsphere volume V, the magnetization M, the volume fraction magnetite V F , and the magnetic field gradient dH/dx as shown in Equation 6-1.[1, 67] dxdHVF M VFM (6-1) where: F M = magnetic force (erg/cm) V = volume of microsphere (cm 3 ) M = magnetization (G) V F = volume fraction magnetite in microsphere dH/dx = magnetic field gradient (Oe/cm) Assuming a spherical shape for a microsphere with radius, r: V = 4/3r 3 (6-2) Substituting Equation 6-2 into 6-1 yields Equation 6-3: dxdHVF M r F334M (6-3) For a microsphere to be retained against the silicone tubing wall, it must travel at a speed in the x direction (v x ) so that it reaches the silicone tubing wall before its speed in the y direction (v y ) carries it past the magnetic field. Therefore for every v y there is a minimum v x that must be generated from the force of the magnetic field on the magnetic microsphere. This force is the viscous drag force (F D ) of the water for the in vitro retention studies where is the fluid viscosity. F D = 6rv x (6-4) When the microspheres are moved by the magnetic field, a steady state velocity (v x ) is reached when the drag force (F D ) equals the magnetic force (F M ). Setting F M = F D gives the velocity of the microsphere (v x ) in the x direction:

PAGE 112

99 dxdHVF9r M 2v2x (6-5) During the experiment, microspheres are distributed throughout the diameter of the silicone tubing. It therefore can be assumed that the average distance the microsphere must travel in the x direction to reach the tubing wall is the radius of the tubing R. Dividing R by v x results in the time (t x ) to reach the tubing wall. dHdxVF2xr M 2R 9t (6-6) Assuming the magnetic field acts in planes perpendicular to the y axis along the width (w) of the magnet only, the time for the microsphere to travel in the y direction is calculated by dividing the width (w) of the magnet by the average velocity (v y ) of the liquid stream in the y direction. This yields Equation 6-7. yyvwt (6-7) The average microsphere located at a distance of R from the tubing wall is retained when t x = t y . Therefore when t x = t y , half of the microspheres will be retained. Therefore setting Equations 6-6 and 6-7 equal to each other and solving for v y gives the flow speed at which 50% of the microspheres are retained (Equation 6-8). dxdHVFR9r M w2v2y (6-8) Results and Discussion Figure 6-10 shows the results of the retention studies for the PLGA and dextran microspheres. Two different magnetic field conditions were used for each microsphere type by positioning the edge of the magnet 4 and 9 mm away from the center of the

PAGE 113

100 silicone tube. At 4 mm, the magnetic field strength was 1.05 kOe and the gradient was 2.7 kOe/cm. At 9 mm, the magnetic field strength was 0.41 kOe and the gradient was 0.73 kOe/cm. Both the microspheres showed similar results, which is not surprising since they have similar size and iron oxide contents. The velocity due to the magnetic field (v x : Equation 6-6) is strongly dependent on microsphere radius since v x is proportional to the radius squared. For the 1.05 kOe gradient, most microspheres are retained until a flow speed of 1 cm/s is reached. At flow speeds greater that 1 cm/s, the % retention drops as the smaller microspheres and/or the microspheres that are farther from the magnet do not have sufficient speed in the x direction to reach the inner wall of the tubing before passing through the magnetic field. For the 0.41 kOe gradient, the microspheres begin to escape the magnet to a considerable degree at flow speeds exceeding 0.2 cm/s. Figure 6-11 shows the same graph as Figure 6.10 except that the x axis is linear. The dashed lines are construction lines used to determine the flow velocity for 50% retention, which is shown in Table 6-1 along with the results of the model. Table 6-1. Flow velocity (v y ) for 50% retention: Experimental results and model Microsphere mm from magnet v y (experimental) for 50% retention (cm/s) v y (model) for 50% retention (cm/s) PLGA 9 0.75 0.22 PLGA 4 3.5 1.2 dextran 9 0.9 0.33 dextran 4 3.1 1.8 The model underestimates v y by a factor of 2 to 3 for all 4 experimental conditions. Appendix B shows sample calculations of v y for PLGA microspheres and Appendix C shows sample calculations of v y for dextran microspheres. One reason for the differences between the model and the experiment is that the model does not take into account the magnetic field gradient above and below the width of the magnet in the plane

PAGE 114

101 perpendicular to the y direction. Although the magnetic field and gradient are diminished outside of this area, the microspheres are still influenced by the magnetic field and are moved in the x direction. The second reason for the discrepancy is aggregation of the microspheres in the magnetic field. When the microspheres are magnetized in the magnetic field, their poles align with the magnetic field. The microspheres form linear aggregates in the direction of the field gradient as the microsphere north and south poles are attracted to each other. This behavior has been shown in mathematical models and in experiments [68, 69]. These linear aggregates have a larger v x than a single microsphere since their magnetic moment is the sum of the moments for all of the microspheres, while the viscous drag force is not increased to a large degree as the linear aggregates grow. In a retention study similar to the one in this thesis, Viroonchatapan et al. [70] found that retention of magnetic liposomes increases with increasing concentration. This is further proof of the role of aggregation in retention studies since aggregation kinetics increase with increasing microsphere concentration [68]. Conclusions An in vitro system was developed for testing the ability of microspheres to be retained by magnetic fields in a flow system for physiologic flow speeds. The retention vs. flow velocity for PLGA and dextran microspheres was very similar. The results were also similar to those of Senyei et al. for retention of magnetic albumin microspheres using the same field gradients as in this dissertation. The experimental results differ from the model since the model does not account for magnetic fields outside the width of the magnet or for microsphere aggregation induced by the magnetic field. Finally, microspheres were easily retained at capillary flow velocities, but were not retained at speeds approaching the arterial flow velocities.

PAGE 115

102 0.1110020406080100 H = 1.05 kOe, dH/dx = 2.70 kOe/cm: dextran H = 1.05 kOe, dH/dx = 2.70 kOe/cm: PLGA H = 0.41 kOe, dH/dx = 0.73 kOe/cm: dextran H = 0.41 kOe, dH/dx = 0.73 kOe/cm: PLGA% retainedflow velocity (cm/s) Figure 6-10. Retention vs. flow velocity (v y ). Flow velocity plotted on log scale. 12345678910020406080100 H = 1.05 kOe, dH/dx = 2.70 kOe/cm: dextran H = 1.05 kOe, dH/dx = 2.70 kOe/cm: PLGA H = 0.41 kOe, dH/dx = 0.73 kOe/cm: dextran H = 0.41 kOe, dH/dx = 0.73 kOe/cm: PLGA% retainedflow velocity (cm/s) Figure 6-11. Retention vs. flow velocity (v y ). Flow velocity plotted on linear scale. Dotted lines are construction lines used to determine v y for 50% retention.

PAGE 116

CHAPTER 7 MRI CONTRAST EFFECT FOR MAGNETIC MICROSPHERES The microsphere concentration necessary for MRI contrast was determined for 20 % iron oxide PLGA microspheres and 35% iron oxide dextran microspheres. Microspheres were encapsulated in alginate gels to prevent them from settling because of gravity. Gels with varying microsphere concentrations were imaged and the relaxation times, T 1 and T 2 , were determined. The gel encapsulation method was modified from a previous method [71], which featured a sodium polyphosphate (NaPP) sequesterant to slow gelation time. The sequesterant delayed the gelation for approximately 20 minutes so that the gel mixtures could be formulated and poured into tubes prior to gelation. Table 7-1 List of reagents used for MRI sample preparation. All were used as received without further purification. Material Source Description, catalog # 20 weight % iron oxide PLGA microspheres 12.5 mg/mL suspension in 0.1% Triton X-100 prepared as described in Chapter 4 35 weight % iron oxide dextran microspheres 4 mg/mL suspension in DI water prepared as described in Chapter 5. Sodium alginate Keltone Product ID: Keltone LV CR Sodium polyphosphate (NaPP) Aldrich # 30,555-3 Calcium sulfate anhydrous (CaSO 4 ) Aldrich Methods Preparation of MRI Standards with Different Concentrations of Microspheres. A 2% w/v sodium alginate stock solution was prepared by dissolving 10 grams of sodium alginate and 0.5 grams of sodium polyphosphate (NaPP) in 500 mL of DI water. A stock suspension of PLGA microspheres in 2% sodium alginate solution was prepared by adding 0.30 g of sodium alginate and 0.015 g sodium NaPP to a mixture of 9 mL of 103

PAGE 117

104 DI water and 6 mL of the 12.5 mg/mL PLGA microsphere suspension. Magnetic stirring was used to dissolve the sodium alginate and NaPP in the mixture and create a 5 mg/mL suspension of microspheres in 2% alginate solution. Similarly a 1.43 mg/mL stock suspension of dextran microspheres was prepared by mixing 300 mg of sodium alginate, 15 mg of NaPP, 9.64 mL of DI water and 5.36 mL of the 4 mg/mL dextran microsphere suspension. More dilute microsphere suspensions were prepared by diluting the microsphere stock suspensions with 2% alginate stock solution. A 2.5% w/v calcium sulfate suspension was prepared by mixing 0.5 grams of CaSO 4 with 20 mL of DI water. Gel samples were prepared in 20 mL glass test tubes. A plain alginate gel was prepared in the bottom half of the tube to serve as a reference for the MRI instrument. A gel with a desired concentration of magnetic microspheres was prepared in the top half of the tube. The plain alginate gels were prepared by slowly pipetting 1 mL of the calcium sulfate suspension into 10 mL of a plain sodium alginate solution while agitating with a Fisher Vortex Genie vortexer. Before significant gelation occurred, the calcium sulfate containing alginate solution was poured into the bottom of the 20 mL glass test tube. The plain samples were left to gel at room temperature for at least 20 minutes before the microsphere containing standards were added to them. After the alginate portion gelled, the microsphere containing standard portion of each test tube was prepared in an identical manner to the plain samples. Figure 7-1 shows a photograph of the dextran microsphere samples. Table 7-2 shows the concentration of microspheres and iron oxide for each gel sample. Magnetic Resonance Imaging (MRI) of Magnetic Microsphere Standards A 4.7 tesla MRI instrument was used for generation of T 1 and T 2 weighted MRI images. The T 1 images were obtained using a spin echo sequence with variable repetition

PAGE 118

105 times of 10 s, 5 s, 2.5 s, 1.25 s, 600ms, 300ms and 150 ms, and an echo time (T E ) of 10 ms. The T 2 images were acquired using a spin echo sequence with 14 echos starting at 6.5 ms and repeating at 6.5 ms intervals. The field of view for all experiments was 7 cm by 7cm. Table 7-2 Concentration of microspheres and iron oxide in MRI samples Sample ID Microsphere concentration (mg/mL) *Iron oxide concentration (mg/mL) PLGA 1 1 0.2 PLGA 2 0.1 0.02 PLGA 3 0.01 0.002 Dex 1 0.29 0.1 Dex 2 0.057 0.02 Dex 3 0.029 0.01 Dex 4 0.0057 0.002 Dex 5 0.0029 0.001 * determined by multiplying microsphere concentration by weight fraction iron oxide in microspheres. 2 cm Figure 7-1. Gels containing dextran microspheres. The left most sample was prepared from the stock microsphere suspension (1 mg/mL iron oxide) and was not used for MRI imaging. The remaining samples are ordered from left to right in decreasing concentration: dex 1, dex 2, dex 3, dex 4, and dex 5.

PAGE 119

106 Results and Discussion For ease of data collection, all of the samples were imaged together in a bundle of tubes. One slice was taken through the microsphere containing portion of the tubes, and one slice through the plain (control) alginate gel portion of the tubes. Figure 7-2 shows the T 1 weighted images of the microsphere containing slice for representative repetition times (T R ). Figure 7-3 shows the T 1 images for the control slice. The darkening of the microsphere standards with time shows how the MRI contrast increases with T R . Figure 7-4 shows the T 2 weighted images for the microsphere containing slice, while Figure 7-5 shows the T 2 images for the control slices. Similar to the T 1 images, contrast is enhanced for longer echo times (T E ) in the T 2 experiment. Figure 7-6 compares the T 1 and T 2 weighted dextran microsphere images for the conditions that gave the best contrast: T R = 10 s for the T 1 experiment, and T E = 84.5 ms for the T 2 experiment. The T 2 image is almost black for 0.02 mg/mL iron oxide and a medium gray for 0.01 mg/mL. The corresponding T 1 images are barely darker than the controls. From these results the dextran microspheres should be detectable, for T 2 weighted imaging, in tissues with iron oxide concentrations as low as 0.02 mg/mL. This corresponds to dextran microsphere concentrations of 0.057 mg/mL. Higher iron oxide concentrations approaching 0.1 mg/mL or dextran microsphere concentrations of 0.29 mg/mL would be needed for T 1 imaging. Figure 7-7 is the comparison of T 1 and T 2 weighted images for the PLGA microspheres. Qualitatively the PLGA results are identical to the dextran results with a large contrast effect at 0.02 mg/mL iron oxide (0.1 mg/mL PLGA microspheres) in T 2 images and reduced contrast effect for the T 1 images.

PAGE 120

107 Figure 7-2. Magnetic Resonance T 1 images for microsphere standards at T R = 150 ms, 600 ms, 2.5 s and 10 s.

PAGE 121

108 Figure 7-3. Magnetic Resonance T 1 images for control gels at T R = 150 ms, 600 ms, 2.5 s and 10 s.

PAGE 122

109 Figure 7-4. Magnetic resonance T 2 images for microsphere standards at T E = 6.5, 19.5, 45.5, and 84.5 ms.

PAGE 123

110 Figure 7-5. Magnetic resonance T 2 images for control gels at T E = 6.5, 19.5, 45.5, and 84.5 ms.

PAGE 124

111 Figure 7-6. Comparison of dextran microsphere standard T 1 and T 2 images. Repetition time T R = 10 s for T 1 images and T E equals 84.5 ms for T 2 images Figure 7-7. Comparison of PLGA microsphere standard T 1 and T 2 images. Repetition time T R = 10 s for T 1 images and T E equals 84.5 ms for T 2 images

PAGE 125

112 MRI signal as a function of T R graphs were created from T 1 images by sampling a 72 pixel area of each sample for the different T R values. Figures 7-8 and 7-9 show MRI signal as a function of T R for the dextran and PLGA microspheres. For MRI studies, T 1 is defined as the time for the MRI signal in the Z direction to return to 63% of its original value following a 90 rf pulse [72]. In this study, T 1 was determined accordingly for each sample from the curves fit to the MRI signal vs. T R data in Figures 7-8 and 7-9. Similarly, T 2 is defined as the time for the MRI signal in the XY plane to decay to 37% of its original value [72]. Figures 7-10 and 7-11 show MRI signal as a function of echo time (T E ). In this study, T 2 was determined accordingly for each sample from the curves fit to the MRI signal vs. T E data in Figures 7-10 and 7-11. Table 7-3 shows T 2 and T 1 for the different microsphere standards along with the T 2 and T 1 for the control gels. The T 1 and T 2 values are the average of 72 pixels for each microsphere containing sample. The control gel values are the average and standard deviation for the 8 control gels, which were calculated individually from 72 pixels. At Fe 3 O 4 concentrations of 0.02 mg/mL and above, both microspheres show an order of magnitude decrease in T 2 that is supported by the T 2 weighted images in Figures 7-6 and 7-7. The T 2 values for the highest concentration of PLGA and dextran microspheres could not be calculated because the decay occurred before the shortest T E of 6.5 ms. The reduction of T 1 for 0.02 mg/mL is modest compared to the T 2 reduction which is supported by the MRI images in Figures 7-6 and 7-7. The T 1 relaxivities for the PLGA and dextran microspheres were determined from the slope of the 1/T 1 vs. concentration of iron in mmole/l (see Figure 7-12). Similarly, the T 2 relaxivities were determined from the graph of 1/T 2 vs. mmole/l Fe 3 O 4 (see Figure

PAGE 126

113 7-13). Microcal Origin linear regression models were used to determine slopes from the data. Table 7-4 shows the relaxivity data. The relaxivities are similar for both the dextran and PLGA microspheres. The dextran microspheres have a higher iron oxide content, therefore the relaxivity in terms of weight of microspheres rather than moles iron would be higher for the dextran microspheres. Relaxivities for iron oxide contrast agents measured using a 0.47 tesla magnet are normally around 30 mmole -1 L s -1 for T 1 and 100 mmole -1 L s -1 for T 2 . The T 1 relaxivities are smaller than the typical values for iron oxide contrast agents. This may be because of the larger magnetic field (4.7 T) used for the experiment and/or the large size of the microspheres compared to typical ferrofluid MRI contrast agents. It is known that T 1 relaxivities have a stronger dependence on iron oxide nanoparticle distribution than T 2 relaxivities [73]. The T 2 relaxivities measured for the microspheres are similar to typical values for MRI ferrofluid contrast agents. Conclusions A method for determining MRI relaxivities from microspheres entrapped in alginate gels was developed. The results of the MRI studies show that the magnetic microspheres can be detected by T 2 weighted MRI imaging in low concentrations: 0.1 mg/mL for PLGA microspheres and 0.059 mg/mL for dextran microspheres. Higher microsphere concentrations are needed for detection in T 1 weighted imaging.

PAGE 127

114 02000400060008000100000.05.0x1041.0x1051.5x1052.0x1052.5x1053.0x1053.5x1054.0x1054.5x1055.0x1055.5x105 blank 0.001 mg/ml 0.002 mg/ml 0.01 mg.ml 0.02 mg/ml 0.1 mg/mlMRI signal (arbitrary units)Repetition time: TR (ms) Figure 7-8. MRI signal intensity in the Z direction as a function of T R for dextran microspheres. Legend shows concentration of Fe 3 O 4 in gels. The T 1 values were determined when MRI signal returned to 63% of maximum value. 02000400060008000100000.05.0x1041.0x1051.5x1052.0x1052.5x1053.0x1053.5x1054.0x1054.5x1055.0x1055.5x105 blank 0.002mg/ml 0.02 mg/ml 0.2 mg/mlMRI signal (arbitrary units)Repetition time: TR (ms) Figure 7-9. MRI signal intensity in the Z direction as a function of T R for PLGA microspheres. Legend shows concentration of Fe 3 O 4 in gels. The T 1 values were determined when MRI signal returned to 63% of maximum value.

PAGE 128

115 0204060801000.05.0x1041.0x1051.5x1052.0x1052.5x1053.0x1053.5x1054.0x1054.5x105 blank 0.001 mg/ml 0.002 mg/ml 0.01 mg.ml 0.02 mg/ml 0.1 mg/mlMRI signal (arbitrary units)Echo time: TE (ms) Figure 7-10. MRI signal intensity in the XY plane as a function of T E for dextran microspheres. Legend shows concentration of Fe 3 O 4 in gels. The T 2 values were determined when MRI signal decayed to 37% of maximum value. 0204060801000.05.0x1041.0x1051.5x1052.0x1052.5x1053.0x1053.5x1054.0x1054.5x105 blank 0.002 mg/ml 0.02 mg/ml 0.2 mg/mlMRI signal (arbitrary units)Echo time: TE (ms) Figure 7-11. MRI signal intensity in the XY plane as a function of T E for PLGA microspheres. Legend shows concentration of Fe 3 O 4 in gels. The T 2 values were determined when MRI signal decayed to 37% of maximum value.

PAGE 129

116 -0.20.00.20.40.60.81.01.21.41.61.82.02.22.42.62.80.00.51.01.52.02.53.03.54.04.55.05.5 PLGA dextran linear fit1/T1 (s-1)iron (mmoles/L) Figure 7-12. Reciprocal relaxation time (1/T 1 ) as a function of iron concentration. Graph used for determination of T 1 relaxivities. 0.000.050.100.150.200.250.300510152025303540 dextran PLGA linear fit1/T2 (s-1)iron (mmoles/L) Figure 7-13. Reciprocal relaxation time (1/T 2 ) as a function of iron concentration. Graph used for determination of T 2 relaxivities.

PAGE 130

117 Table 7-3. Relaxation times (T 1 and T 2 ) for dextran and PLGA microsphere standards. Relaxation time T 2 was not calculated for PLGA 1 and Dex 1 since T 2 was too short for measurement by method used in this study. sample [Fe 3 O 4 ] mg/mL T 1 : ms (stdev.) T 2 : ms (stdev.) PLGA 1 0.2 202 (83) PLGA 2 0.02 2100 (33) 32.2 (0.8) PLGA 3 0.002 2610 (57) 147 (7) Dex 1 0.1 453 (67) Dex 2 0.02 1700 (31) 26.5 (0.6) Dex 3 0.01 2401 (88) 51.2 (1.1) Dex 4 0.002 2700 (113) 140 (6) Dex 5 0.001 2980 (132) 193 (11) Control 0 2770 (230) 255 (11) Table 7-4. Relaxivity values for dextran and PLGA microspheres. microsphere T 1 relaxivity mmole -1 L s -1 T 2 relaxivity mmole -1 L s -1 PLGA 1.8 106 dextran 1.5 130

PAGE 131

CHAPTER 8 OVERALL CONCLUSIONS AND FUTURE WORK Conclusions Two different methods for producing magnetic microspheres using degradable materials were developed in this dissertation. The microspheres are in a size range suitable for iv applications. SQUID magnetometry showed that the microspheres contain superparamagnetic iron oxide nanoparticles, which become magnetized in the presence of a magnetic field, but demagnetize upon removal of the field. Both methods were optimized for iron oxide content. In addition to the novel preparation methods; methods for characterizing magnetic microspheres and measuring their properties were developed. These methods include SQUID for iron oxide content determination, MRI contrast studies, and an in vitro retention model. Magnetic PLGA microspheres of a suitable size for iv applications were prepared for the first time. This was accomplished by creating a hydrophobic surface on iron oxide nanoparticles by coating with oleic acid. The oleic acid allowed suspension of nanoparticles in an organic solvent so that the solvent evaporation/extraction process could be used to prepare microspheres. Regular shaped microspheres were produced with iron oxide loadings of nearly 20%. As the loading increased from 20% , the microspheres became increasingly irregular. This process is simple and should be applicable to any polymer that is soluble in chloroform. Dextran microspheres were produced using the all-aqueous phase separation method. Microspheres suitable for iv applications and magnetic microspheres were 118

PAGE 132

119 produced by this method for the first time. Microsphere size increased with increasing iron oxide content. The diameter of microspheres with 35 weight percent iron oxide was less than 5 m for the entire size distribution, while many microspheres were greater than 5 m in diameter for the 45% iron oxide samples. An in vitro system was developed for testing the ability of microspheres to be retained by magnetic fields in a flow system for physiologic flow speeds. The retention vs. flow velocity for PLGA and dextran microspheres was very similar. The experimental results differ from the model since the model does not account for magnetic fields outside the width of the magnet or for microsphere aggregation induced by the magnetic field. Finally, microspheres were easily retained at capillary flow velocities, but were not retained at speeds approaching the arterial flow velocities. A method for determining MRI relaxivities from microspheres entrapped in alginate gels was developed. The results of the MRI studies show that the magnetic microspheres can be detected by T 2 weighted MRI imaging in low concentrations: 0.1 mg/mL for PLGA microspheres and 0.059 mg/mL for dextran microspheres. Higher microsphere concentrations are needed for detection in T 1 weighted imaging. Future Work Further refinement of both microsphere processes could be examined. From the morphology of the PLGA microspheres, I suspect that a phase separation causes the irregular shape. Use of a different molecular weight of PLGA or different copolymers may reduce the irregular shapes when using higher iron oxide loadings. Use of a different solvent than chloroform or using mixed solvents may also alleviate the problem. The water content of the dextran phase, the dextran molecular weight and degree of

PAGE 133

120 methacrylate substitution was constant for the dextran microspheres. The effects of these parameter on size and iron oxide loading should be explored. Degradation studies on the microspheres would be interesting. The oleic acid may accelerate the degradation of PLGA or other polylactones since the degradation rate is increase with acid catalyzation. Enzymatic degradation of dextran microspheres as a function of degree of methacrylate substitution and iron oxide content should be performed. Finally, applications for the microspheres such as targeted drug delivery, and MRI contrast agents in animal models should be examined.

PAGE 134

APPENDIX A CALCULATION OF THEORETICAL AMOUNT OF OLEIC ACID NEEDED FOR MONOLAYER COVERAGE OF IRON OXIDE PARTICLES 1. Estimate specific surface area (SSA: m 2 /g) for iron oxide based on diameter determined from scherrer formula. SSA = specific surface area of iron oxide (m 2 /g) = density of iron oxide = 5.25 g/cm 3 D = iron oxide diameter = 10.6 nm SSA =6D6525106101001076393333.(/)..gcmmmcmmg 2. Determine grams of oleate per grams of iron oxide for a mono-layer based on parking area of oleate and SSA of iron oxide particles Parking area of 1 oleate molecule = 20 2 = 2 x 10 -19 m 2 [74] Molar mass of oleate = 282.47 g/mole 1076210160210282470252341922334....mgFeOmoleatemoleculemoleoleatemoleculesgoleatemolegoleategFeO 121

PAGE 135

APPENDIX B SAMPLE CALCULATIONS FOR FLOW SPEED AT 50% RETENTION FOR PLGA MICROSPHERES USING EQUATION 6-8 (FOR 19 WEIGHT % IRON OXIDE MICROSPHERES AND MAGNET DISTANCE OF 4 MM) dxdHR9r V M w2v2Fy (6-8) Calculation of V F (Volume Fraction Iron Oxide) Iron oxide from chloroform ferrofluids is 17% oleate by weight: esmicrospherin oleate % weight 89.3oxideiron % 83acid oleic % 17 oxideiron % weight 19 100% 19% iron oxide 3.89% oleic acid = 77.1 weight percent PLGA in microspheres The density of the microspheres ( m )can then be calculated using the densities and weight fractions of the three components of the microspheres using Equation B-1 ironironololPLGAPLGAwww1m (B-1) Where: w PLGA = weight fraction PLGA = 0.771 w ol = weight fraction oleate = 0.0389 w iron = weight fraction iron oxide = 0.19 PLGA = density PLGA = 1.34 g/cm 3 ol = density oleate = 0.895 g/cm 3 iron = density iron oxide = 5.17 g/cm 3 )/(17.519.0)/(895.00389.0)/(34.1771.01333mcmgcmgcmg =1.525 g/cm 3 esmicrospher of gram 1in cm 0.6556 g/cm 525.1133 122

PAGE 136

123 esmicrospher of gram 1in oxideiron cm 0.03675 g/cm .175oxideiron g 0.1933 6150.0 oxide)(iron V esmicrospher cm .65560oxideiron cm 0.03675F33 Calculation of V y for 50% Retention: Equation 6-8, (derived in chapter 6) was used to calculate the flow speed (v y ) necessary for 50% retention. dxdHR9r V M w2v2Fy (6-8) where: w = width of magnet (cm) M = magnetization (G) r = microsphere radius (cm) V F = volume fraction magnetite in microsphere dH/dx = magnetic field gradient (Oe/cm) = viscosity of water (poise) R = radius of silicone tubing (cm) cm 2.54 incm 2.54 inch 1 w G 2794 mAG 10 4 c m emumA 10 cmoxideiron g 5.17 emu/g 43 M3-333 (M = 43 emu/g was determined from SQUID data for bare iron oxide) cm 10 5.925 mcm 10 m 0.5925 r 5--4 V F = 0.0561 (see above calculations) dH/dx = 2700 Oe/cm (measured with gauss meter) = 9.00 10 -3 poise[75] cm 10 7.938 inchcm 2.54 inch 321 2R

PAGE 137

124 cmpoisecmOecmGcm 10 7.938 )10 9(9 2700 )10 (5.925 0.0561 2794 2.54 2v2-3-22-5y v y = 1.17 cm/s Flow speed for 50% retention using PLGA microspheres and magnet 4 mm from center of silicone tubing equals 1.17 cm/s. A check of the units is shown below scmg Oe G 2/12/1 scmgpoise scmsgscmgcmscmgcmscmgcmscmgcm 1 v22/12/122/12/1y

PAGE 138

APPENDIX C SAMPLE CALCULATIONS FOR FLOW SPEED AT 50% RETENTION FOR DEXTRAN MICROSPHERES USING EQUATION 6-8 (FOR 35 WEIGHT % IRON OXIDE MICROSPHERES AND MAGNET DISTANCE OF 9 MM) dxdHR9r V M w2v2Fy (6-8) Calculation of V F (Volume Fraction Iron Oxide) The microspheres are 35% by dry weight iron oxide. Therefore 1 gram of microspheres contains 0.35 grams of iron oxide. Dividing 0.35 grams of iron oxide by its density (5.17 g/cm 3 ) gives the volume of iron oxide in 1 gram of microspheres: 32310770.617.535.0cmcmgg Magnetic dextran microspheres are crosslinked hydrogels and therefore contain water. For calculation of the volume fraction iron oxide, the water content of the dextran polymer matrix is estimated to be 60 weight percent based on work be Stenekes et al [2]. The dextran content is therefore 40% by weight. The weight of crosslinked dextran hydrogel matrix (including water) in 1 gram of microspheres (by dry weight) can be determined from the weight of dextran (0.65 g) in 1 gram of dried microspheres and the dextran content of the hydrated crosslinked dextran matrix (40%) gdextrangdextrangdextrang625.14.0%40165.0 hydrated crosslinked dextran matrix The density of the hydrated crossliked dextran matrix should be similar to that of a 40% dextran solution. This density is 1.1764 g/cm 3 . The volume of the hydrated dextran matrix for 1 gram of dry material is calculated by dividing its weight by the density. 33381.1/1764.1625.1cmcmgg hydrated crosslinked dextran matrix The volume fraction (V F ) iron oxide is calculated from the volume iron oxide (6.770 10 -2 cm 3 ) and the volume of the hydrated crosslinked dextran matrix (1.381 cm 3 ) 125

PAGE 139

126 0467.0381.110770.610770.6V22F Calculation of v y for 50% Retention dxdHR9r V M w2v2Fy (6-8) cm 2.54 incm 2.54 inch 1 w G 1949 mAG 10 4 c m emumA 10 cmoxideiron g 5.17 emu/g 30 M3-333 (M = 30 emu/g was determined from SQUID data for bare iron oxide) cm 10 7.95 mcm 10 m 0.795 r 5--4 V F = 0.0467 (see above calculations) dH/dx = 730 Oe/cm (measured with gauss meter) = 9.00 10 -3 poise [75] cm 10 7.938 inchcm 2.54 inch 321 R2cmpoisecmOecmGcm 10 7.938 )10 9(9 730 )10 (7.95 0.0467 1949 2.54 2v2-3-22-5y v y = 0.332 cm/s Flow speed for 50% retention using dextran microspheres and magnet 9 mm from center of silicone tubing equals 0.332 cm/s

PAGE 140

LIST OF REFERENCES 1. Senyei, A., K. Widder, and G. Czerlinski, Magnetic Guidance of Drug-Carrying Microspheres. Journal of Applied Physics, 1978. 49(6): p. 3578-3583. 2. Stenekes, R.J.H., O. Franssen, E.M.G. van Bommel, D.J.A. Crommelin, and W.E. Hennink, The Preparation of Dextran Microspheres in an All-Aqueous System: Effect of the Formulation Parameters on Particle Characteristics. Pharmaceutical Research, 1998. 15(4): p. 557-561. 3. Smith, W.F., Principles of Materials Science and Engineering. 1990, New York: Mcgraw Hill. 649-653. 4. Dunlop, D.J., The Rock Magnetism of Fine Particles. Physics of the Earth and Planetary Interiors, 1981. 26(1-2): p. 1-26. 5. Janssen, W.E. and A.M. Rios, Non-Specific Cell Binding Characteristics of Para-Magnetic Polystyrene Microspheres Used for Antibody-Mediated Cell Selection. Journal of Immunological Methods, 1989. 121(2): p. 289-294. 6. Ugelstad, J., P.C. Mork, R. Schmid, T. Ellingsen, and A. Berge, Preparation and Biochemical and Biomedical Applications of New Monosized Polymer Particles. Polymer International, 1993. 30(2): p. 157-168. 7. Bahar, T. and S.S. Celebi, Immobilization of Glucoamylase on Magnetic Poly(styrene) Particles. Journal of Applied Polymer Science, 1999. 72(1): p. 69-73. 8. Chatterjee, J., Y. Haik, and C.J. Chen, Modification and Characterization of Polystyrene-Based Magnetic Microspheres and Comparison with Albumin-Based Magnetic Microspheres. Journal of Magnetism and Magnetic Materials, 2001. 225(1-2): p. 21-29. 9. Yanase, N., H. Noguchi, H. Asakura, and T. Suzuta, Preparation of Magnetic Latex Particles by Emulsion Polymerization of Styrene in the Presence of a Ferrofluid. Journal of Applied Polymer Science, 1993. 50: p. 765-776. 10. Ugelstad, J., Magnetic Polymer Particles and Process for the Preparation Thereof. 1987, SINTEF: USA: Patent # 4654267. 127

PAGE 141

128 11. Ramchand, C.N., P. Pande, P. Kopcansky, and R.V. Mehta, Application of Magnetic Fluids in Medicine and Biotechnology. Indian Journal of Pure & Applied Physics, 2001. 39(10): p. 683-686. 12. Roger, J., J.N. Pons, R. Massart, A. Halbreich, and J.C. Bacri, Some Biomedical Applications of Ferrofluids. European Physical Journal-Applied Physics, 1999. 5(3): p. 321-325. 13. Geldwerth, D., D. Helley, K. de Jong, D. Sabolovic, C. Sestier, J. Roger, J.N. Pons, J.M. Freyssinet, P.F. Devaux, and F.A. Kuypers, Detection of Phosphatidylserine Surface Exposure on Human Erythrocytes Using Annexin V-Ferrofluid. Biochemical and Biophysical Research Communications, 1999. 258(1): p. 199-203. 14. Sonti, S.V. and A. Bose, Cell-Separation Using Protein-a-Coated Magnetic Nanoclusters. Journal of Colloid and Interface Science, 1995. 170(2): p. 575-585. 15. Schutt, W., C. Gruttner, U. Hafeli, M. Zborowski, J. Teller, H. Putzar, and C. Schumichen, Applications of Magnetic Targeting in Diagnosis and Therapy Possibilities and Limitations: A Mini-Review. Hybridoma, 1997. 16(1): p. 109-117. 16. Weissleder, R., A.S. Lee, A.J. Fischman, P. Reimer, T. Shen, R. Wilkinson, R.J. Callahan, and T.J. Brady, Polyclonal Human Immunoglobulin-G Labeled with Polymeric IronOxide Antibody Mr Imaging. Radiology, 1991. 181(1): p. 245-249. 17. Kresse, M., S. Wagner, D. Pfefferer, R. Lawaczeck, V. Elste, and W. Semmler, Targeting of Ultrasmall Superparamagnetic Iron Oxide (USPIO) Particles to Tumor Cells In Vivo by Using Transferrin Receptor Pathways. Magnetic Resonance in Medicine, 1998. 40(2): p. 236-242. 18. Bachgansmo, T., Ferrimagnetic Susceptibility Contrast Agents. Acta Radiologica, 1993. 34: p. 1-&. 19. Weissleder, R., A.S. Lee, B.A. Khaw, T. Shen, and T.J. Brady, Antimyosin-Labeled Monocrystalline Iron-Oxide Allows Detection of Myocardial Infarct Mr Antibody Imaging. Radiology, 1992. 182(2): p. 381-385. 20. Fahlvik, A.K., E. Holtz, P. Leander, U. Schroder, and J. Klaveness, Magnetic Starch Microspheres, Efficacy and Elimination a New Organ-Specific Contrast Agent For Magnetic-Resonance-Imaging. Investigative Radiology, 1990. 25(2): p. 113-120. 21. Bartolozzi, C., R. Lencioni, F. Donati, and D. Cioni, Abdominal MR: Liver and Pancreas. European Radiology, 1999. 9(8): p. 1496-1512.

PAGE 142

129 22. Bellin, M.F., C. Roy, K. Kinkel, D. Thoumas, S. Zaim, D. Vanel, C. Tuchmann, F. Richard, D. Jacqmin, A. Delcourt, E. Challier, T. Lebret, and P. Cluzel, Lymph Node Metastases: Safety and Effectiveness of MR Imaging with Ultrasmall Superparamagnetic Iron Oxide Particles Initial Clinical experience. Radiology, 1998. 207(3): p. 799-808. 23. Anzai, Y., J.A. Brunberg, and R.B. Lufkin, Imaging of Nodal Metastases in the Head and Neck. Jmri-Journal of Magnetic Resonance Imaging, 1997. 7(5): p. 774-783. 24. Anzai, Y., K.E. Blackwell, S.L. Hirschowitz, J.W. Rogers, Y. Sato, W.T.C. Yuh, V.M. Runge, M.R. Morris, S.J. McLachlan, and R.B. Lufkin, Initial Clinical-Experience with Dextran-Coated Superparamagnetic Iron-Oxide for Detection of Lymph-Node Metastases in Patients with Head and Neck-Cancer. Radiology, 1994. 192(3): p. 709-715. 25. Tiefenauer, L.X., G. Kuhne, and R.Y. Andres, Antibody Magnetite Nanoparticles: in-Vitro Characterization of a Potential Tumor-Specific Contrast Agent for MagneticResonance-Imaging. Bioconjugate Chemistry, 1993. 4(5): p. 347-352. 26. Suwa, T., S. Ozawa, M. Ueda, N. Ando, and M. Kitajima, Magnetic Resonance Imaging of Esophageal Squamous Cell Carcinoma Using Magnetite Particles Coated with Anti-Epidermal Growth Factor Receptor Antibody. International Journal of Cancer, 1998. 75(4): p. 626-634. 27. MayoSmith, W.W., S. Saini, G. Slater, J.A. Kaufman, P. Sharma, and P.F. Hahn, NIR Contrast Material for Vascular Enhancement: Value of Superparamagnetic Iron Oxide. American Journal of Roentgenology, 1996. 166(1): p. 73-77. 28. Ahlstrom, K.H., L.O. Johansson, J.B. Rodenburg, A.S. Ragnarsson, P. Akeson, and A. Borseth, Pulmonary MR Angiography with Ultrasmall Superparamagnetic Iron Oxide Particles as a Blood Pool Agent and a Navigator Echo for Respiratory Gating: Pilot Study. Radiology, 1999. 211(3): p. 865-869. 29. Weishaupt, D., P.R. Hilfiker, M. Schmidt, and J.F. Debatin, Pulmonary Hemorrhage: Imaging With a New Magnetic Resonance Blood Pool Agent in Conjunction with Breathheld Three-Dimensional Magnetic Resonance Angiography. Cardiovascular and Interventional Radiology, 1999. 22(4): p. 321-325. 30. Hilfiker, P.R., G.G. Zimmermann-Paul, M. Schmidt, H.P. Klotz, G.M. Kacl, and J.F. Debatin, Intestinal and Peritoneal Bleeding: Detection With an Intravascular Contrast Agent and fast Three-Dimensional MR Imaging Preliminary Experience from an Experimental Study. Radiology, 1998. 209(3): p. 769-774. 31. Alexiou, C., W. Arnold, P. Hulin, R.J. Klein, H. Renz, F.G. Parak, C. Bergemann, and A.S. Lubbe, Magnetic Mitoxantrone Nanoparticle Detection by Histology, X

PAGE 143

130 ray and MRI After Magnetic Tumor Targeting. Journal of Magnetism and Magnetic Materials, 2001. 225(1-2): p. 187-193. 32. Alexiou, C., W. Arnold, R.J. Klein, F.G. Parak, P. Hulin, C. Bergemann, W. Erhardt, S. Wagenpfeil, and A.S. Lubbe, Locoregional Cancer Treatment with Magnetic Drug Targeting. Cancer Research, 2000. 60(23): p. 6641-6648. 33. Lubbe, A.S., C. Bergemann, H. Riess, F. Schriever, P. Reichardt, K. Possinger, M. Matthias, B. Dorken, F. Herrmann, R. Gurtler, P. Hohenberger, N. Haas, R. Sohr, B. Sander, A.J. Lemke, D. Ohlendorf, W. Huhnt, and D. Huhn, Clinical Experiences with Magnetic Drug targeting: A Phase I Study with 4'-Epidoxorubicin in 14 Patients with Advanced Solid Tumors. Cancer Research, 1996. 56(20): p. 4686-4693. 34. Lubbe, A.S., C. Bergemann, W. Huhnt, T. Fricke, H. Riess, J.W. Brock, and D. Huhn, Preclinical Experiences With Magnetic Drug Targeting: Tolerance and Efficacy. Cancer Research, 1996. 56(20): p. 4694-4701. 35. Driscoll, C.F., R.M. Morris, A.E. Senyei, K.J. Widder, and G.S. Heller, Magnetic Targeting of Microspheres in Blood-Flow. Microvascular Research, 1984. 27(3): p. 353-369. 36. Gupta, P.K. and C.T. Hung, Albumin Microspheres. V. Evaluation of Parameters Controlling the Efficacy of Magnetic Microspheres in the Targeted Delivery of Adriamycin in Rats. International Journal of Pharmaceutics, 1990. 59: p. 57-67. 37. Hassan, E.E., R.C. Parish, and J.M. Gallo, Optimized Formulation of Magnetic Chitosan Microspheres Containing the Anticancer Agent, Oxantrazole. Pharmaceutical Research, 1992. 9(3): p. 390-397. 38. Fruitwala, M.A. and N.M. Sanghavi, Site-Specific Drug Targeting with Fluorouracil Microspheres. Drug Delivery, 1996. 3: p. 5-8. 39. Mosbach, K. and U. Schroder, Preparation and Application of Magnetic Polymers For Targeting of Drugs. Febs Letters, 1979. 102(1): p. 112-116. 40. Pulfer, S.K. and J.M. Gallo, Targeting Magnetic Microspheres to Brain Tumors, in Scientific and Clinical Applications of Magnetic Carriers. 1997, Plenum Press: New York. p. 445-455. 41. Pulfer, S.K. and J.M. Gallo, Enhanced Brain Tumor Selectivity of Cationic Magnetic Polysaccharide Microspheres. Journal of Drug Targeting, 1998. 6(3): p. 215-227. 42. Widder, K.J., A.E. Senyei, and D.G. Scarpelli, Magnetic Microspheres: a Model System for Site Specific Drug Delivery In Vivo. Proceedings of the society for experimental biology and medicine, 1978. 58: p. 141-146.

PAGE 144

131 43. Widder, K., G. Flouret, and A. Senyei, Magnetic Microspheres: Synthesis of a Novel Parental Drug Carrier. Journal of Pharmaceutical Sciences, 1979. 68(1): p. 79-82. 44. Widder, K.J. and E. Senyei, Magnetic Microspheres: A Vehicle for Selective Drug Targeting. Pharmaceutical Therapeutics, 1983. 20: p. 377-395. 45. Goodwin, S., C. Peterson, C. Hoh, and C. Bittner, Targeting and Retention of Magnetic Targeted Carriers (MTCs) Enhancing Intra-Arterial Chemotherapy. Journal of Magnetism and Magnetic Materials, 1999. 194(1-3): p. 132-139. 46. Arshady, R., Suspension, emulsion and dispersion polymerization: A methodological survey. Colloid and Polymer Science, 1992. 270: p. 717-732. 47. Kissel, T., Z. Brich, S. Bantle, I. Lancranjan, F. Nimmerfall, and P. Vit, Parenteral Depot-Systems on the Basis of Biodegradable Polyesters. Journal of Controlled Release, 1991. 16(1-2): p. 27-41. 48. Labhasetwar, V., C.X. Song, W.R. Humphrey, R.J. Shebuski, and R.J. Levy, Arterial Uptake of Biodegradable Nanoparticles: Effect of Surface Modifications. Journal of Pharmaceutical Sciences, 1998. 87(10): p. 1229-1234. 49. Song, C.X., V. Labhasetwar, H. Murphy, X. Qu, W.R. Humphrey, R.J. Shebuski, and R.J. Levy, Formulation and Characterization of Biodegradable Nanoparticles for Intravascular Local Drug Delivery. Journal of Controlled Release, 1997. 43(2-3): p. 197-212. 50. Li, Y.X. and T. Kissel, Synthesis and Properties of Biodegradable Aba Triblock Copolymers Consisting of Poly(L-Lactic Acid) or Poly(L-LacticCo-Glycolic Acid) a-Blocks Attached to Central Poly(Oxyethylene) B-Blocks. Journal of Controlled Release, 1993. 27(3): p. 247-257. 51. Cohn, D. and H. Younes, Biodegradable Peo/Pla Block Copolymers. Journal of Biomedical Materials Research, 1988. 22(11): p. 993-1009. 52. Hu, D.S.G. and H.J. Liu, Structural-Analysis and Degradation Behavior in PolyethyleneGlycol Poly (L-Lactide) Copolymers. Journal of Applied Polymer Science, 1994. 51(3): p. 473-482. 53. Perrin, D.E. and J.P. English, Polyglycolide and Polylactide, in Handbook of Biodegradable Polymers, A.J. Domb, J. Kost, and M.W. Wiseman, Editors. 1997, Harwood Academic Publishers. p. 2-27. 54. Slomkowski, S., S. Sosnowski, and S. Gadzinowski, Microspheres by Polymerization of Lactides and -Caprolactone. Macromolecular Symposia, 1997. 123: p. 45-60.

PAGE 145

132 55. Sosnowski, S., M. Gadzinowski, and S. Slomkowski, Pol(L,L-lactide) Microspheres by Ring-Opening Polymerization. Macromolecules, 1996. 29: p. 4556-4564. 56. CRC Handbook of Chemistry and Physics, ed. D.R. Lide and H.P.R. Frederikse. 1997, New York: CRC Press. 8-87. 57. Redhead, H.M., S.S. Davis, and L. Illum, Drug Delivery in Poly(Lactide-co-Glycolide) Nanoparticles Surface Modified with Poloxamer 407 and Poloxamine 908: In Vitro Characterisation and In Vivo Evaluation. Journal of Controlled Release, 2001. 70(3): p. 353-363. 58. Govender, T., S. Stolnik, M.C. Garnett, L. Illum, and S.S. Davis, PLGA Nanoparticles Prepared by Nanoprecipitation: Drug Loading and Release Studies of a Water Soluble Drug. Journal of Controlled Release, 1999. 57(2): p. 171-185. 59. Dunn, S.E., A.G.A. Coombes, M.C. Garnett, S.S. Davis, M.C. Davies, and L. Illum, In Vitro Cell Interaction and In Vivo Biodistribution of Poly(Lactide-Co-Glycolide) Nanospheres Surface Modified by Poloxamer and Poloxamine Copolymers. Journal of Controlled Release, 1997. 44(1): p. 65-76. 60. Verrecchia, T., G. Spenlehauer, D.V. Bazile, A. Murrybrelier, Y. Archimbaud, and M. Veillard, Non-Stealth (Poly(Lactic Acid Albumin)) and Stealth (Poly(Lactic Acid-Polyethylene Glycol)) Nanoparticles as Injectable Drug Carriers. Journal of Controlled Release, 1995. 36(1-2): p. 49-61. 61. Vandijkwolthuis, W.N.E., O. Franssen, H. Talsma, M.J. Vansteenbergen, J.J.K. Vandenbosch, and W.E. Hennink, Synthesis, Characterization, and Polymerization of Glycidyl Methacrylate Derivatized Dextran. Macromolecules, 1995. 28(18): p. 6317-6322. 62. Robineau, M. and D. Zins, Surfactant-Coated Particles in Magnetic Fluids. Characterization and Study of Thermal Stability Under Inert Atmosphere. Annales De Chimie-Science Des Materiaux, 1995. 20(6): p. 327-333. 63. Massart, R., Magnetic Fluids and Process for Obtaining Them. 1982, Agence Nationale de Valorisation de la Rechere (ANVAR), Paris, France: United States Patent # 4,329,241. 64. Dunlop, D.J., Superparamagnetic and Single-Domain Threshold sizes in Magnetite. Journal of Geophysical Research, 1973. 78(11): p. 1780-1793. 65. Colthup, N.B., H.D. Daly, and S.D. Wiberly, Introduction To Infrared and Raman Spectroscopy. 1990, San Diego: Academic Press. 247-252. 66. Pouchert, C.J., The Aldrich Library of Infrared Spectra, Edition III. 1981: Aldrich Chemical Company. M5590-9.

PAGE 146

133 67. Craik, D., Magnetism: Principles and Applications. 1995, Chichester, England: John Wiley and Sons. p. 39. 68. Relle, S., S.B. Grant, and C. Tsouris, Diffusional Coagulation of Superparamagnetic Particles in the Presence of an External Magnetic Field. Physica A, 1999. 270(3-4): p. 427-443. 69. Helgesen, G., A.T. Skjeltorp, P.M. Mors, R. Botet, and R. Jullien, Aggregation of Magnetic Microspheres Experiments and Simulations. Physical Review Letters, 1988. 61(15): p. 1736-1739. 70. Viroonchatapan, E., M. Ueno, H. Sato, I. Adachi, H. Nagai, K. Tazawa, and I. Horikoshi, Preparation and Characterization of Dextran Magnetite-Incorporated Thermosensitive Liposomes: an Online Flow System for Quantifying Magnetic Responsiveness. Pharmaceutical Research, 1995. 12(8): p. 1176-1183. 71. Monshipouri, M. and R.R. Price, Emulsification Preparation of Calcium Alginate Beads in the Presence of Sequesterant. Journal of Microencapsulation, 1995. 12(3): p. 255-262. 72. Brown, M.A. and R.D. Semelka, MRI: Basic Principles and Applications. 1995, New York: John Wiley and Sons. pages 1-64. 73. Bowen, C.V., X.W. Zhang, G. Saab, P.J. Gareau, and B.K. Rutt, Application of the static dephasing regime theory to superparamagnetic iron-oxide loaded cells. Magnetic Resonance in Medicine, 2002. 48(1): p. 52-61. 74. Okon, E., D. Pouliquen, P. Okon, Z.V. Kovaleva, T.P. Stepanova, S.G. Lavit, B.N. Kudryavtsev, and P. Jallet, Biodegradation of Magnetite Dextran Nanoparticles in the Rat a Histologic and Biophysical Study. Laboratory Investigation, 1994. 71(6): p. 895-903. 75. CRC Handbook of Chemistry and Physics, ed. D.R. Lide and H.P.R. Frederikse. 1997, New York: CRC Press. 6-3.

PAGE 147

BIOGRAPHICAL SKETCH Patrick James Leamy, along with his twin brother Michael Joseph Leamy, was born in San Francisco, California to George Leamy and Ann Beardmore on June 10, 1971. He was raised in Glenmont, New York and studied mechanical engineering and biomedical engineering at Worcester Polytechnic Institute (WPI) in Worcester, Massachusetts. After receiving his Bachelor of Science degree in mechanical engineering from WPI in January 1994, Patrick studied ceramics and calcium phosphate biomaterials at Pennsylvania State University. After receiving his Master of Science degree in materials from Pennsylvania State in May 1997, Patrick worked as a materials engineer for Framatome Connectors in York, Pennsylvania. To further his research career in biomaterials, Patrick Leamy began pursuing a Doctor of Philosophy degree in the Materials Science and Engineering department at the University of Florida in January 1999. Patrick Leamy happily married Karla Gutierrez, now Karla Leamy, on September 21, 2002. 134